MARINE BIOLOGICAL LABORATORY. n Received Accession No. cpC^.y ^ 2^ Given by //'ax,^L<^<*<^^i'*^^ i>(-<-^^ <^)y^^ Place, ^.X^^ d.J^, %*No book OP pamphler is to be removed ffom the Uab- opatopy ixiithout the pepmission of the Trustees. Committee on Publication Barton W. Evermann Chairman and Editor C. Hart Merriam Bernard R. Green Frank Baker C. F. Marvin PROCEEDINGS OF THE Washington Academy of Sciences Vol. VI 1904 WASHINGTON October, 1904-FEBRUARY, 1905 AFFILIATED SOCIETIES. Anthropological Society of Washington. Biological Society of Washington. Botanical Society of Washington. Chemical Society of Washington. Columbia Historical Society. Entomological Society of Washington. Geological Society of Washington. Medical Society of the District of Columbia. National Geographic Society. Philosophical Society of Washington. Society of American Foresters. Washington Society of the Arch^ological Institute of America. PHf 83 OF The N«w Eh* Printinq Company LANCASTER, Pa. CONTENTS. PAGE Contributions to the Knowledge of the Life History of Pinus with special reference to Sporogenesis, the Development of the Gametophytes and Fertilization; by Margaret C. Ferguson . i Studies of Variation in Insects; by Vernon L. Kellogg and Ruby G. Bell ^°3 Hopkins-Stanford Galapagos Expedition.-XVH. Shore Fishes of the Revillagigedo, CHpperton, Cocos and Galapagos Islands; by Robert Evans Snodgrass and Edmund Heller . • -333 Some interesting Beaver Dams in Colorado; by Edward R. Warren ^^^ Organization and Membership of the Washington Academy of Sciences ^^^ Index '^^^ ILLUSTRATIONS PLATES FACING PAGE I-I V. Microsporogenesis 156 V. Development of Pollen-grain 164 VI. Pollination and subsequent Phenomena 166 VII. Growth of the Pollen-tube 168 VIII-XI. Spermatogenesis 170 XII-XIII. Macrosporogenesis 178 XIV. Germination of Macrospore 182 XV. Female Prothallium 1S4 XVI-XVIII. Oogenesis 186 XIX-XXI. Fertilization 192 XXII-XXIII. Development of Proembryo 198 XXIV. Abnormalities 202 XXV. Map of Slate River, Colorado 438 XXVI. Map of Slate River, Colorado 438 XXVII-XXXIV. Thirteen photographs show^ing beaver dams, houses, etc., in Slate River Valley, Colorado.. 438 TEXT FIGURES PAGE I. Fore and hind wings of honey bee (drone) 214 2-5. Fore wing of honeybee (drone) 215 6-7. Hind wing of honey bee (drone) 216 8. Hind wings of honey bee (drone) 217 9. Fore and hind wings of honey bee (worker) 217 10. Fore wings of honey bee (drones) 219 11. Fore wing of honey bee (drone) 222 1 2-20. Frequency polygons of variation in wings of honey bees 223-330 21. Part of costal margin of hind wing of honey bee much magnified, to show hooks 231 22-33. Frequency polygons of variation in costal hooks of wings of honeybees 232-239 34. Fore and hind wings of male black ant 245 35-44. Frequency polygons of variation in costal hooks of wings of black ants 246-252 45. Wing of ant showing venation 255 46-48. Diagrams showing variation in elytral and pro- thoracic pattern of the convergent lady-bird 258-272 49. Frequency polygon of variation in prothoracic pat- tern in the convergent lady-bird 273 50. Diagram showing variations in elytral pattern of the flower beetle 274 51-53. Frequency polygons of variation in elytral pattern of the flower beetle 275-279 54; 55. Diagrams showing variation in the yellow jacket.. 284; 285 56. Leaf hopper, Typhlocyba comes 288 57. Diagram showing variation in pattern of prothorax of a flower-bug 291 58. Water boatman, Corisa sp 293 59. Parnassian butterfly, Parnassius smintheus 295 60. A\WQ.x\z?ci\ coc\i.xo?Lz\i^ Periplaneta americana 296 61. Red-legged locust, Melanoplus femur-rubruin... 301 62-69. Frequency polygons of variation in the red-legged locust 302-305 70. Seventeen-year locust. Cicada septendeclm 306 71-78. Frequency polygons of variation in Cicada sep- tendecitn 307-3 10 79. Antenna of scale insect, Ceroputo yuccce 311 80. Biting bird-louse, Z.z^^Mr?^5 cc/er 314 81. Fredaceous ground-beetle, Pterostichus sp 317 WASHINGTON ACADEMY OF SCIENCES OFFICERS FOR 1904 President Charles D. Walcott Vice-Presiden ts Fro7ti the Anthropological Society W. H. Holmes Archceological Society John W. Foster Biological Society Barton W. E vermann Botanical Society Frederick V. Coville Chemical Society C. E. Munroe Columbia Historical Society W J McGee Err '^-^olos^ical Society H. G. Dyar GeoL ciety G. K. Gilbert Medicu iety W. W. Johnston National eographic Society A. Graham Bell Philosophical Society Richard Rathbun Society of American Foresters Gifford Pinchot Secretary Treasurer Frank Baker Bernard R. Green Managers Class o/ igoj Class of i gob Class of igoj L. O. Howard P. W. Clarke Geo. M. Kober O. H. Tittmann C. W. Hayes Gifford Pinchot » Carroll D. Wright G. W. Littlehales F. A. Lucas PROCEEDINQS OF THE WASHINGTON ACADEMY OF SCIENCES Vol. VI, pp. I-203. Sept. io, 1904. CONTRIBUTIONS TO THE KNOWLEDGE OF THE LIFE HISTORY OF PINUS WITH SPECIAL REFERENCE TO SPOROGENESIS, THE DEVELOPMENT OF THE GAMETO- PHYTES AND FERTILIZATION. By Margaret C. Ferguson, Ph.D., Associate Professor of Botany, Wellesley Cot lege. Plates I-XXIV. v^ CONTENTS. Introduction. Purpose of the study 3 Historical notes 5 Methods. Collecting 12 Fixing 13 Staining. 15 Chapter I. Microsporogenesis. The microsporangium. The wall of the pollen- sac '. 17 The primitive archesporium 18 Tetrad-division. The definitive archesporium 20 The first nuclear division of the microspore-mother-cell 21 The second mitosis of the mother-cell 30 The problem of reduction 31 Development of the microspore. The formation of the spore-wall 34 The origin of the air-sacs and liljeration of the microspore .... 36 The growth of the microspore 37 Summary 39 Proc. Wash. Acad. Sci., September, 1904. (i) 2 MARGARET C. FERGUSON Chapter II. The male gametophyte. The development of the pollen-grain. The formation of the prothallial cells 41 The mature pollen-grain 44 Pollination. The ovule at the time of pollination 45 The pollen-chamber 4^ Development of the pollen-tube. The first period of growth. Germination of the pollen-grain 47 The division of the antheridial cell 4S The winter condition 50 The second period of growth. Renewed activities in the macrosporangium 50 Renewed activities in the male gametophyte 51 Division of the generative nucleus 53 Growth of the sperm-nuclei 62 Elongation of the pollen-tube 64 Summary 66 Chapter III. Macrosporogenesis. The female cone. The macrosporangium 7° Formation of the axial row. The macrospore-mother-cell 7' The first division of the mother-cell 7- The second division of the mother-cell 74 Significance of the tetrad-division within the ovule 76 Later history of the axial row. The fate of the upper cells 79 The growth of the macrospore 80 Summary 81 Chapter IV. The female gametophyte. Development of the prothallium. The first period of growth 82 The second period of growth 84 The so-called " spongy tissue." The first period of growtli 86 The second period of growth 88 The nature and functiftn of this tissue 89 Development of the archegonium. The early growth of the archegonium 90 The division of the central cell 95 History of the ventral canal-coll 97 Matination of the egg. The descent and growth of the egg-nucleus 99 Tlie " proteiil vacuoles" 103 Tlie receptive vacuole 109 Suinmarv 110 LIFE HISTORY OF PINUS 3 Chapter V. Fertilization and related piienonicna. Conjugation. The coming together of the gametoplijtes 113 The union of the sexual nuclei 114 The first division following fecundation. The prophases of the division 115 Later stages in the mitosis 118 The pro-embryo. Division of the two segmentation-nuclei 122 The four segmentation-nuclei 124 The development of cell-walls 126 Later mitoses in the formation of the pro-embrvo 127 The fate within the egg of the smaller sperm-nucleus, the stalk- cell, and the tube-nucleus 128 Summary 130 Appendix. Some abnormal conditions. Supernumerarj' nuclei in the male gametophjte 133 Unusual conditions in the female gametophjte 135 A peculiar method of conjugation 138 Note 139 List of Papers cited 142 Explanation of Plates 154 INTRODUCTION. There is no chapter in the annals of botanical science more fascinating than that which deals with the history of sexuality in plants. No definite date marks the discovery of the fact that plants, like animals, are male and female ; the idea was rather a growth, as is plainly shown by the writings of Aristotle, Theophrastus, Pliny and others of the early philoso- phers. The fact may, however, be said to have been estab- lished by Camerarius (1694) in his " De sexu Plantarum," but for many years after his time botanists found in this question merely a favorite subject for philosophical speculation. Their ideas remained vague and uncertain, no effort being made to confirm their theories either by observation or experimentation. It was not until near the middle of the last century that actual investigations were begun along this line. Amici (1830-1846) made certain interesting observations regarding the develop- ment of the pollen-tube and the origin of the embryo in several plants ; but the splendid series of embryological papers pub- lished by Hofmeister (1848-1867) first placed the science upon a sure foundation and marked a new era in the study of sexual 4 MARGARET C. FERGUSON reproduction in plants. Although the researches of Hofmeis- ter, Strasburger, Warming, Belajeff and others who have con- tributed to our knowledge of this subject, especially during the last decade, have disclosed many facts concerning the structure and development of the pollen-grain, of the ovule and of the embryo, our knowledge of certain phases of spermatogenesis and oogenesis is still ver}'' meager, and not a sufficiently large number of plants have been thoroughly investigated to admit of generalizations. The celebrated discoveries of Hirase, Ikeno and Webber, in 1897, gave a new incentive to this study, par- ticularly in connection with the Gymnosperms, and rendered it highly desirable that fertilization and associated phenomena should be worked out for other members of this group by the more modern methods of investigation. The present studies were begun in the fall of 1897 with the hope of adding somewhat to our knowledge of this subject. Incidentally, it seemed desirable to determine whether any ves- tiges of the bodies called blepharoplasts by Webber (1897^) still persist in the conifers. As a result of the past embryological studies, a vast number of facts pertaining to the life-history of the gametophytes in the higher plains has accumulated. While many of the conclusions reached are the outcome of serious direct investigations, others are based on the insufficient evi- dence found in a rather superficial study of a large number of plants. What we need to-day is not more facts regarding un- related plants, so much as a careful working out of the details of development in representative genera. This research is based primarily upon a study of Piiius Strobus, but nearly every observation recorded has been con- firmed for Phius rigida and P. austriaca^ and to a large extent, for P. montana var. uncinata and P. rcsinosa. The descrip- tions given may be understood to refer alike to the five species named above unless otherwise stated in the text. Nearly six hundred paraffin blocks with imbedded material have been made, and more than four thousand slides of serial sections have been stained and studied. Six hundred separate collec- tions of material would seem unnecessarily large if one were studying a plant like Nicotiana in which, according to Guig- LIFE HISTORY OF PINUS 5 nard (1902), fertilization follows in 2 hours after pollination, but in Phms, where almost 13 months intervene between these two processes, such a number is not excessive. While it is true in cytological studies, as elsewhere, that numbers, or mere mass work, do not signify excellence, it is equally true that the re- sults of investigations based upon a study of a limited amount of material are, at best, unsatisfactory, and, other things being equal, those conclusions will be most valuable which have been formulated after a careful observation of many specimens.' HISTORICAL NOTES. In the following brief summary of the literature dealing with the AbicttiiccB, only the more important papers have been noted, and the observations recorded by the various writers have been given without comment. The tetrad-division in the pollen-mother-cell of Pinns and Abies was studied in 1848 by Hofmeister. He stated that the pollen-mother-cells were already developed in the anthers at the end of November, two special daughter-cells were formed at the close of the first division in the spring, and the four cells resulting from the second division were found to lie either in one plane or at the corners of a tetrad. Three years later (185 1) Hofmeister published the results of his remarkable series of investigations in the higher cryptogams and conifers. He described and figured the pollen-grain in the AbietinecB as con- sisting of a cell-complex, noted the depression in the apex of the nucellus in Piniis at the time of pollination, and the single embryo-sac-mother-cell deep in the interior of the nucellus. It appeared that the pollen-grain rested some weeks upon the nucellus before the pollen-tube was emitted. After the germina- tion of the pollen-grain, the tube grew for several weeks and penetrated nearly to the point of union between integument and nucellus, but it might cease growth before so great a depth was reached. 1 This paper was given especial honorable mention on April 26, 1903, by the Association for Maintaining the American Women's Table at the Zoological Station at Naples and for Promoting Scientific Research by Women. I wish here to express my deep gratitude to Mrs. Ellen H. Richards, Miss Florence Gushing and other members of the above named association through whose generous efforts the publication of this paper in its present form has been made possible. 6 MARGARET C. FERGUSON He concluded that the embryo-sac remained for a long time as a single cell, its nucleus finally dissolving to be replaced by a number of free nuclei ; in a few days the sac was filled with long cells reaching to the middle ; at the beginning of winter, the walls of this transitory endosperm were greatly thickened ; in the spring, the thickened walls of the endosperm were absorbed and the cells liberated. Each primordial cell thus made free contained, somewhat later, three or four daughter-cells which were, in their turn, liberated by the disso- lution of the mother-wall. Thus the number of cells within the embryo-sac was greatly increased, the embryo-sac itself growing to more than twenty times its previous volume. The cells of the nucellus also multiplied rapidly except in the region previously penetrated by the pollen-tubes. In the middle of May, a layer of cells lined the embryo-sac, cell layers in- creased until they met in the center, then the corpuscula were differentiated. The corpuscula were always separated in the AbietinccB by one or more layers of cells, and the walls enclos- ing the corpuscula were thought to be channelled, thus afford- ing open communication with the surrounding cells. In Pinus from 3 to 5 corpuscula were developed in each ovule, and a corresponding number of funnel-shaped openings occurred in the upper part of the endosperm. When the pollen-tube reached the corpusculum it contained free spherical cells in its lower end. The tube either flattened itself out upon the corpusculum or penetrated a short distance into it. After fertilization the impregnated germinal vesicle increased in size, its nucleus dis- appeared, and soon a large daughter-cell was seen at the base of the corpusculum. By repeated divisions of this cell the pro- embryo was formed. In 1858 Hofmeister found the usual number of neck-cells in Pinus Strobus to be four, exceptionally three, five, or six, all lying in the same plane. He further demonstrated the vacu- olate character of the contents of the corpusculum during its development. These vacuoles disappeared before impregna- tion, being replaced by free cells — the germinal vesicles, or Keimblaschen. A pit was figured in the apex of the pollen- tube after its entrance into the corpusculum, but it was said that LIFE HISTORY OF PINUS 7 the tube remained closed until after the formation of the pro- embryo, when it was ruptured by mechanical means. The great abundance of starch in the pollen-tube of the Abictinece was also mentioned at this time. While the " Higher Crypto- gamia" appearing in 1862 was largely a translation of Hof- meister's earlier publications, it likewise presented many new observations. The fact was noted that in Pinus the integument surrounds the nucellus, leaving open above its apex a wide micropylar canal. In all the ConifercB^ after the embryo-sac was entirely filled with cellular tissue, certain cells near the micropylar end ceased dividing but increased markedly in size ; the other cells of the endosperm continued to multiply in num- ber, but remained comparatively small ; thus the corpuscula were differentiated. After the cutting off of the neck-cells in the AbictinecB^ additional cells were developed at the top of the endosperm, giving rise to the depressions referred to in 185 1. Scarcely a day intervened between the approach of the pollen- tube and the formation of a four-celled pro-embryo at the base of the corpusculum, and this occurred contemporaneously in all ovules of all trees growing under similar circumstances. The works of Strasburger on this subject have been more numerous and complete than those of any other investigator. It is extremely interesting to note how his interpretations have kept pace with the improvements in methods of research. In 1869 he traced the development of the endosperm from the free cells lining the embryo-sac to its maturity, and established the fact that shortly before fertilization the central cell divides to form the canal-cell and the egg-cell. He confirmed Hofmeister's observations regarding the channels in the upper part of the endosperm, and the presence of a closed pit at the apex of the pollen-tube ; but he did not observe the nuclei in the pollen-tube, and remarked that, inasmuch as the sexual organs touch in these plants, spermatozoids would be superfluous and were, in reality, not present. He added, however, that their place was taken by granular protoplasm and starch grains which exercised the same fertilizing effect on the egg as do spermatozoids. After fertilization four nuclei were detected at the base of the corpusculum and a division into a cross took place, these cells 8 MARGARET C. FERGUSON divided and were separated by cross- walls, the lower four di- vided again making three layers of four cells each, the middle layer then elongated pushing the lowest cells down into the endosperm. In Picea a fourth layer of cells was observed at the base of the central cell. In 1872 Strasburger stated that the canal-cell loosened itself from the &^^ and hung as a cap just beneath the neck- cells, at the same time the egg-nucleus increased in 'size and moved to the center of the corpusculum. He detected two cells in the pollen-tube of several Gymnosperms, but considered that such cells were extremely rare in the Ad ietmecs, as he had only once found one in this group. The shrunken remains of these cells were seen in the pollen-tube after fertilization. He be- lieved that the pit of the pollen-tube remained closed, and that the exchange-substance was apparently communicated by a vacuole between the apex of the pollen-tube and the egg- nucleus. After fertilization the central nucleus was dissolved, and, in " abnormal " cases, four new nuclei appeared in the central part of the egg, but there was strong evidence that these did not develop into an embryo. Six years later (1878), he observed one or more divisions in the pollen-grain shortly before pollination. The small cells resulting from these divi- sions were interpreted as rudimentary prothallium. Two large primordial cells were demonstrated in the pollen-tube of Pimis and Picca when the tube was just above the archegonium. Ac- cording to Strasburger's interpretation at that time, the nucleus in front was dissolved while the one behind entered the egg and fused with its nucleus. This was a great advance on his previous observations, but he still conceived of the pollen-tube as remaining closed, and fancied that the protoplasmic contents passed through the membrane directly while the starch was dis- solved before its transmission into the ^^^' He was now con- vinced that only a part of the contents of the pollen-tube was taken up by the egg-nucleus, the remaining portion uniting di- rectly with the egg-plasma ; but he was not certain whether the protoplasm active in fertilization came in as a formless mass or in the shape of a nucleus. Strasburger established the fact, in 1S79, '^^^^'^ ^^ ^^ ^^^ fore- LIFE HISTORY OF PINUS 9 most of the two sperm-nuclei in the pollen-tube which is instru- mental in effecting fertilization. He reported the presence of an axial row of three cells in Larix^ the lowest cell being the embryo-sac-mother-cell. The generalization was made that the prothallium arises in all the gymnosperms through free cell-division, all the free nuclei dividing at the same time. It was claimed that but a single endosperm was formed in the Abietine(B, that the primary nucleus of the embryo-sac remained undivided during the first year, and that the " transitory endo- sperm " of Hofmeister was in reality the freed cells of the nucel- lus which were destined to be absorbed. It was to these cells that the term spongy tissue was applied. In the following year (1880) Strasburger described and figured the mature archego- nium in Picea and discussed the early stages of endosperm for- mation in Pimis, but he gave little that was new at that time. It was in this same year that Sokolowa (1880) published the results of her researches in the development of the prothallium in the gymnosperms. Cell-walls were laid down between the nuclei imbedded in the peripheral layer of protoplasm, but no cell thus formed was furnished with a wall on its inner free side. These open cells were termed " alveoli." They grew in length until the middle of the embryo-sac was reached, then walls arose at the inner ends and the alveoli were closed ; cell divis- ions followed, and gradually the elongated alveoli gave place to many cells. Goroschankin (1880 and '83) reported that the protoplasm of the Qgg and of the sheath-cells was in immediate contact through pores in the separating membrane; he saw (1883^) the two sperm-nuclei pass into the egg in Pinus Pumilio^ and believed that both fused with its nucleus ; the great similarity which the spheres in the egg bear to nuclei was commented upon and he questioned the propriety of calling them vacuoles. Stras- burger (1884) confirmed Goroschankin's observations as to the passage of the two sperm-nuclei from the pollen-tube into the egg, but pointed out that only the one in advance fused with the egg-nucleus. As the protoplasmic contents of the central cell increased, the vacuoles decreased, and every transition could be traced between the large vacuoles and the meshes of the proto- Proc. Wash. Acad. Sci., July, 1904. lO MARGARET C. FERGUSON plasm filled with metaplasm. In the pines, a large vacuole often held several smaller ones. The egg-nucleus slowly filled itself with metaplasm during its descent to the center of the cell. Three successive divisions occurred in the large cell of the pollen-grain in Larix, the first two prothallial cells formed were small and soon disorganized, the third one increased greatly in size and divided to form the stalk- and the bod3'-cell. It was left for Belajeff (1891) to establish the true nature of the cell-complex found in the pollen-grain of the Gymnosperms. He demonstrated the fact that in Taxiis baccata the large nucleus of the pollen-grain is the vegetative or pollen-tube-nucleus, as in the Angiosperms, and that the sperm-nuclei arise b}' the division of one of the smaller cells of the pollen-grain, this smaller cell first dividing to form the stalk- and the generative cell. Strasburger (1892) showed that Belajeff 's observations on the structure of the pollen-grain and the development of the pollen- tube in Taxtis baccata were, in general, true for the other Gymnosperms. He described the mature pollen-grain in Piims as containing a large tube-cell, a small cell — the third prothallial or antheridial cell — and the remnants of the first two prothallial cells. Pollination was immediately followed by the germination of the pollen-grain, and the nucleus of the large cell wandered at once into the tube. The last formed prothallial cell remained in its place in the pollen-grain until the following spring, when it divided into the stalk- and the body-cell of the antheridium. The division of this cell was not studied, but Strasburger thought it took place at about the same time as the develop- ment of the archegonia. The pollen-grain of Picca was found to correspond exactly with that of Pinus excepting that the an- theridial cell divided while still within the anther. The sperm- cells in Pinus were seen in the apex of the pollen-tube ; the lower cell was the larger ; and each cell was almost entirely filled with its large, coarsely granular nucleus. At the tip of the pollen-tube, the stalk- and the tube-nucleus could no longer be distinguished one from the other. The sperm-nucleus was shown to be smaller than the egg-nucleus, but the two were alike in the amount of active nuclear substance ; and attention LIFE HISTORY OF PINUS II was called to the smallness of the first nuclear figure following fecundation in comparison with the size of the conjugating nuclei. The germ-nucleus divided in its original position in the egg, and the two nuclei passed towards the " organic" apex of the archegonium. Belajeff (1893) worked out the development of the pollen tube in Picea as a type of the AhietinccE. He found that the generative cell divided while still within the pollen-grain and gave rise to two sperm-cells which he figured as of the same size. Dixon (1894) traced the history of the pollen-grain and the pollen-tube in Piniis sylvcstris from the middle of April to the time of fertilization. He thought that the prothallial cell divided towards the end of April to form a small stalk-cell and a larger body-cell. The body-cell immediately divided into two cells of almost equal size — the male sexual cells. The sperm-cells moved into the pollen-tube followed by the nucleus of the stalk- cell. Pollen-tubes were found to branch freely while in the upper "brown" tissue of the nucellus but only one branch of each tube was continued through the lower part of the nucellus. He noted that the four nuclei, much of the protoplasm, and considerable of the starch of the pollen-tube passed into the oosphere. As a rule, eight chromosomes were found in the nuclei of the female gametophyte. In giving an account of some work done by his students on the Gymnosperms, Coulter (1897) reported that the work of Dixon " was largely confirmed in the minutest detail " ; and in 1900 he figured the pollen-tube "in pines," when just above the archegonium, showing two sperm-cells of equal size. Atkin_ son (1898) stated that the sperm-mother-cell in Pinus divided into two sperm-cells after having passed into the pollen-tube. Blackman's excellent treatise on fertilization and related phenomena in Pinus sylvestris was published in 1898. Man y details of development were most carefully worked out, but th e facts recorded are not enumerated here, since they will be duly considered in connection with the observations, as record ed in the body of this paper, that have been made by the writer on other species of pines. Since the appearance of Blackman's monograph, a considerable literature dealing with various stages 12 MARGARET C. FERGUSON of development in the gametophytes of the Abietinece has been published. The details of these investigations are familiar to all students of the subject. These papers will, therefore, be men- tioned at this point by title only ; they will be referred to again in the discussions which follow. Chamberlain (1899), Oogene- sis in Phiiis Laricio; Wuicizki (1899), Ueber die Befruchtung bei den Coniferen ; Arnoldi (1900), Beitrage zur Morphologic der Gymnospermen, IV; Juel (1900), Beitrage zur Kenntniss der Tetradentheilung ; Murrill (1900), The Development of the Archegonium and Fertilization in the Hemlock Spruce {Tsuga canadensis Carr.); Coulter and Chamberlain (1901), Morphology of the Spermatophytes ; Ishikawa (1901), Reduc- tion Division in Larix ; and the papers published by the writer in 1901.^ METHODS. Collecting. — On November 15, 1897, and each week there- after until December 25, cones of Pinus Strobtis, P. rigida, P. austriaca^ P. niontana var. tmcinata, and the staminate strobili of P. ausiriaca were collected. Material was brought in occas- sionally during the remainder of the winter. Pistillate cones of the species named, and also of P. resinosa, were collected once each week beginning with April i ; collections were made twice each week throughout the month of May, and three times a week during June. From June 10-30,. a period which was sure to cover fertilization, cones of Pinus Slj'obus were collected every day at about nine o'clock in the morning, and frequently again at four o'clock in the afternoon. Male cones were gath- ered, from those species in which they had appeared, at irregu- lar intervals during the early spring. From the first of May until the time of pollination, which varies by a number of days in the different species, staminate strobili were collected each day. During May and June the young female cones were gathered as well as the more mature ones of the previous year's growth. After July i, the older cones were no longer collected, but the young cones of Pinus Strobus^ P. rigida^ and P. ausiri- aca were collected once each week until November 15. Cones ' See "Note" at close of Appendix. LIFE HISTORY OF PINUS I3 of Pimis Strobus were again collected regularly, as described above, throughout the spring and early summer of 1899. Collec- tions of the staminate cones of Pinus Strobus and P. rigida were made during May and June 1901, and from May 15 to June 15 of the same year the young pistillate cones oi Pinus ris;ida were gathered daily. Material was obtained from different trees and different locali- ties. The practice of collecting all one's material from a single tree, as reported by Murrill (1900), Land (1902) and others, does not seem a safe one to follow, for certain peculiarities of develop- ment which are not characteristic of the species may appear in an individual. At the time of each collection, ovules were put up from several cones of each species, these cones being taken not from the tip of one branch but from different branches. The central portion only of the cone was used, the ovules at either extremity being more or less abortive. After collecting, the material was taken at once to the laboratory and preserved. The staminate cones and, in the early stages of development, the pistillate ones were fixed entire or cut into quarters longitu- dinally. Very soon the individual scales of the female cones were removed from the receptacle before fixing, and, when the scales were of sufficient size to admit of such manipu- lation, all superfluous parts were cut away, leaving the two tiny ovules still united by a small portion of the scale. With the renewal of growth in the spring, the ovules were removed from the scales and, as soon as it was feasible, a portion of the integument was cut away* from two or more sides of each ovule, thus bringing the fixing fluid into direct contact with the young gametophyte. For later stages, the endosperm was frequently removed from the integument, but such material did not prove to be as satisfactory as that in which the nucellar cap and a small portion of the coat were left in connection with the prothallium. Throughout the entire mechanical process of preparing material for the fixer, the most extreme care was used, as it was found that a very slight pres- sure was sufficient to cause distortions and thus to render the material worthless for cytological studies. Fixing. — The methods used in fixing and staining do not I/j. MARGARET C. FERGUSON differ materially from those generally employed in cytological work. The fixing fluids tested were chrome-osomo-acetic acid solution, chrome-acetic acid solution, corrosive sublimate in aqueous solution, absolute alcohol, and Carnoy's fluid. The first two were tried with variations in concentration and in length of time. The chrome-osomo-acetic acid solution giving by far the best results, the other fixers were entirely discarded. It was made up according to the following formula : Chromic acid crystals 1.3 gms. Osmic acid (in glass bulb) .5 gms. Glacial acetic acid 83 c.c. Distilled water 160.0 c.c. This solution used in one half strength and allowed to act for about 15 hours proved to be most excellent for fixing the pro- thallium at the time when it consists of a wall layer of proto- plasm containing numerous free nuclei. For the development of the pollen-grain and fertilization stages, it was most satisfac- tory when undiluted, and allowed to act for about 24 hours. If the fluid blackened at all, it was poured off after 2 or more hours and fresh added. After fixing, the material was washed in running water from 2 to 12 hours, but as a rule specimens were not kept in the running water longer than 6 hours. The very convenient piece of apparatus described by Durand ('99) was used for this process. Subsequent to washing, material was dehydrated in 8 grades of alcohol beginning with 15^ and ending with the absolute. It was not allowed to stand in the lower grades for more than 6 hours, and was rarely kept in the absolute alcohol longer than that time ; the latter was changed 3 times, once about every 2 hours, to insure perfect deh3^dration in as short a time as pos- sible. After material had been in 85^ alcohol for 12 hours, it was decolorized in a 35^ solution of hydrogen peroxide, made up in 95^ alcohol, for 24 hours. While material was always bleached in toto, it was frequently found necessary to decolor- ize again on the slide. After dehydration, material was brought gradually, through ascending grades, into pure cedar oil, xylol or chloroform. The best results were obtained with the cedar oil and it was far more commonly used than the others. If it LIFE HISTORY OF PINUS I5 was desirable to store material for a few days or weeks, pure cedar oil was found to be a much better medium than 75 ^ alcohol, which is commonly used for temporary storing of ma- terial. For the purpose of getting specimens into pure paraffin they were transferred to tiny wire-gauze baskets and carried successively into 25, 50 and 75^ paraffin in cedar oil, and finally into pure paraffin with a melting point of 54°, in which they were at last imbedded. This is a very convenient and eco- nomical method for getting material through the paraffin oven. The grades of cedar oil in paraffin can be kept in the bath a long time and used repeatedly with impunity, and material can be carried in the little baskets from bottle to bottle much more quickly and with less liability to injury than in an}^ other way with which I am familiar. At the time of fixing, a small piece of paper, bearing the number, in pencil, corresponding to the number of the entry in the record book, was placed in each bottle, remained with the material through all the changes which followed, and was finally imbedded in one corner of the paraffin block containing the specimens. Staining. — A Minot-Zimmermann revolving microtome was used in cutting the material. The sections varied in thickness from 4 to 13.6 microns, but by far the greater number were made 6.3 microns thick. They were fastened to the slide by means of albumen-fixative, and the slides were labelled with glass-ink. In preparing this ink, a paste was made of the best English vermilion in sodium silicate, and sufficient water was added to give the proper consistency for writing. Glass-pen- cils, Higgins' waterproof ink, both with and without collodion, and other methods for marking slides were tried ; but I have never found anything at all comparable, for excellence, with the glass-ink. When properly prepared it is not dissolved dur- ing the process of staining, but can be removed from slides or dishes, when desirable to do so, by heating in a strong solution of potash or in gold dust. As is usual in cytological studies, considerable experimenta- tion was necessary before satisfactory stains were obtained. Among the stains tested were Rosen's ('92) fuchsin and methy- lene-blue method ; the Ehrlich-Biondi-Heidenhain mixture, as 1 6 MARGARET C. FERGUSON prepared by Dr. G. Griibler ; Guignard's combination of methyl green, acid fuchsin, and orange G; Flemming's safranin- gentian-violet-orange combination ; and Heidenhain's iron- hasmatoxylin. The last two proved the most satisfactory and were almost exclusively used. The iron-hsematoxylin was fol- lowed by orange G, or, if it was desirable to stain cell-walls, by Bismarck brown. Iron-haematoxylin followed by Flem- ming's triple stain, or by gentian-violet and orange G, brought out the so-called kinoplasmic structures with great definiteness. The best differentiation was obtained with the iron-hgematoxy- lin by allowing the h£ematoxylin to act from 12 to 18 hours, decolorizing in iron-alum, and then washing in running tap- water from 2 to 6 hours. Flemming's triple stain was often used without the safranin with excellent results. Both anilin and aqueous solutions of gentian-violet were used. As a rule, a one-half percent, solution was employed, the slides remain- ing in it from 5 to 20 minutes. The achromatic figures in the divisions of the pollen-mother-cell, especially in Piniis Strobus, were, however, brought out with great difficulty with this stain. The best results were obtained for these stages by allowing the slides to stand from 24 to 48 hours in stender dishes of distilled water to which not more than 10 drops of a one percent, solution of gentian-violet had been added. Pimis sections take the orange with such avidity, that a fraction of a minute was in all cases a sufficiently long time to allow this stain to act. After washing out the superfluous gentian-violet and deh3'drating in absolute alcohol, differentiation was effected by dashing with clove oil. Bergamot oil was used for fixing and clearing, and I have found it expedient to pass the slides from bergamot oil to jars of xylol. They can remain in the xylol for hours, if desirable, without injury, and the xylol is so readily miscible with the balsam that the preparations become clear and more satisfactory for studying in a much shorter time than when car- ried directly to the balsam from the bergamot oil. LIFE HISTORY OF PINUS 1 7 CHAPTER I. MiCROSPOROGENESIS. THE MICROSPORANGIUM. The Wall of the Pollen-sac. — With the exception of Pinus Strobus, the staminate cones, in the pines which I have studied, make their appearance in October or November. I have searched repeatedly in the autumn for the male inflorescences of Pinus Sirobus but have never been able to find them until late April or early May of the following spring. If they are present at all before spring they can be scarcely more than potentially so, for they are not sufficiently devoloped to be detected in the field, nor by careful dissection in the laboratory. The structure of the microsporangium agrees perfectly with that usually described for the AbietinecB. The wall of the young pollen-sac consists of three or four layers of cells. The cells of the outer layer are nearly isodiametric, while those of the inner layers are smaller and more or less tabular in outline. Just within, and in immediate contact with the archesporium, is the ring of tapetal cells. In the early stages of development the wall-cells are rich in cytoplasm and contain nuclei which are large in proportion to the size of the cells. The microsporangium increases much in size in the spring, and by the time that the microspore-mother-cells are in the prophase of division, considerable change has occurred in the wall-cells of the pollen-sac. The outer layer has lost its nuclei and the cells have become filled with a homogeneously staining resinous sub- stance ; in Pinus Strobus this resinous deposit extends to the second layer of wall-cells as well ; the cells of the inner la3'ers have been considerably flattened out, and their cytoplasmic con- tent has become much reduced. When the pollen-grains are mature, all the wall-cells of the microsporangium, except the outermost layer, have disappeared. They have doubtless been absorbed, their substance contributing to the nutrition of the pollen-grains. The tapetum cannot be distinguished during the earlier stages of development from the other tissues. It is first clearly differ- l8 MARGARET C. FERGUSON entiated in the spring, when the mother-cells are in the early prophase of the heterotypic division. The mitoses leading to development of this layer have not been studied, but there are indications that it is formed from the outer layer of the sporog- enous tissue rather than, as usuall}'" described, from the inner layer of wall-cells. The microsporangium-wall, after the appearance of the tapetum, is composed, as before, of three or four layers of cells ; furthermore, the tapetum is always inti- mately associated with the sporogenous tissue, while it is fre quently found separated from the wall of the pollen-sac, probably as a result of imperfect fixation. The question as to the origin of this tissue in Pimis must, however, await further investiga- tion. During the later stages of division in the pollen-mother- cells, the tapetal cells increase much in size, their cytoplasm becomes very dense and each cell comes to have from one to three nuclei which have been observed in all stages of fusion. Karyokinetic figures have been frequently noted in the tapetal cells indicating that the nuclei of these cells divide mitotically, and the division conforms to the ordinary or typical method of mitosis. When the young microspores become free, these cells have attained to their greatest size, and show a diffuse reaction to stains. From this time they gradually diminish in size and finally disappear altogether. The nutritive function of this tissue is too well understood to require discussion here. The Primitive Archesporiuni. — With the exception of Pintis Strohus, the primitive archesporium is clearly differentiated in the autumn, but the mother-cells of the microspore do not arise until the latter part of April, and in Piims Strobus not until about three weeks later. In the younger stages of development, a superficial study shows no sharp demarcation between archesporium and wall, but a careful examination reveals certain differences by which the two can always be distinguished. The cells of the arche- sporium are larger, have larger nuclei, and denser cytoplasm than those of the wall. They are also polyhedral in outline while the wall-cells are somewhat tabular from the first, though not so markedly so as at a later period. During the winter, the nucleus of a primitive archesporal cell contains several nucleo- I.IFR HISTORY OF PINUS I9 lus-like bodies, of which as many as eleven have been counted in a single section of a nucleus, and a less number than seven is rarely found. The delicate but extensive nuclear reticulum is slightly chromatic and stains scarcely more strongly than the cytoplasm of the cell. Both cytoplasm and nuclear network stain diffusely with gentian-violet during this period of rest (fig- ^)- . . . In those species in which the microsporangia make their appearance in the autumn, the pollen-sacs remain small and the archesporial cells comparatively few in number until the following spring. Hofmeister ('48) found the mother-cells of the pollen-grains in the anthers of Pinus and Abies at the end of November, Belajeff ('94) observed the pollen-mother-cells of Larix in the spireme stage in October, and Coulter and Cham- berlain ('01) have recently figured the ' ' microsporangium of Pinus Laricio in the mother-cell stage in October." The sporogenous tissue, as they have illustrated it, bears a very strong resem- blance to that shown in fig. i of this paper. There is undoubted evidence that these are not pollen-mother-cells in the species of pines which I have studied. In the first place, the number of cells in a single anther in November is far less than the number of microspore-mother-cells which is eventually formed. As the microsporangium enlarges in the spring these cells not only increase in size but multiply in number. During the last of March and first of April karyokinetic figures, representing various stages of division, are seen in all preparations, and in all cases division is proceeding by the typical method character- istic of vegetative or somatic cells. In the latter part of April or first of May (for Pimis Strobus about the middle of May), typical division ceases, and, after a period of growth, the pro- phases characteristic of the heterotypical division are entered upon. The time at which the rest preparatory to the hetero- typic mitosis begins varies by about three weeks in the different species, and by ten or more days in the same species in different seasons. Had Coulter and Chamberlain examined microspo- rangia during the latter part of March they would doubtless have found typic divisions taking place in the archesporial tissue. 20 MARGARET C. FERGUSON TETRAD-DIVISION. The Definitive Archesj^orium. — During the period of " rest " preceding the heterot3^pic division, the microspore-mother-cell increases much in size, its nucleus becoming even larger than an entire cell of the primitive archesporium, as is readily seen by comparing figs, i and 2 with figs. 3 and 4. The walls en- closing the spore-mother-cells thicken considerably, and the cytoplasm assumes a fine, almost granular structure which, under high magnification, resolves itself into a delicate, close reticulum. At this stage, only three or four nucleoli are found within the nucleus, but this reduction in number may be only apparent, for the nucleus has enlarged to such an extent that no one section would be liable to contain as many of these structures as would a section of one of the smaller nuclei of the primitive archesporium. No attempt has been made to de- termine the exact number of nucleoli in the nuclei of the arche- sporium at any time in its history, as it is next to impossible to trace accurately the sections in the series of any given cell when each anther contains hundreds of archesporial cells all of which are practically alike in form, structure and staining capacity. As the nucleus of a pollen-mother-ceil enlarges, its reticu- lum becomes more open, the threads of the net gradually in- crease in thickness, the net-knots or karyosomes become more or less prominent, and numerous smaller granules are distrib- uted irregularly upon the linin. Many cross-threads are with- drawn but no true spireme is formed at this time (fig. 3). The thickening of the threads is more prominent in Pintis Strobiis than in the other species, the net-knots are more conspicuous, and a somewhat imperfect spireme arises, although here, too, many anastomosing threads still persist (fig. 4). A remarkable change has taken place in the attitude of the different elements of the cell towards stains. When the microspore-molher-cells are first formed both cytoplasm and nuclear net stain more or less diffusely with gentian-violet as in the primitive arclie- sporium, but, as growth proceeds, the cytoplasm ceases to react to chromatin dyes and takes the orange G with avidity. The nucleoli are colored far less deeply with the gentian-violet than LIFE HISTORY OF PINUS 21 formerly, and the nuclear reticulum takes the blue characteris- tic of chromatin. In this condition, the contracted state known as synapsis is entered upon. The First Nicclear Division of the Mic7'ospore-niother-ceU . — As soon as a microspore-mother-cell has attained full size, cer- tain changes in its nucleus indicate that the prophase of the first division has been initiated. The reticulum gradually draws together, its threads becoming thicker and the meshes smaller (figs. 5 and 6). Contraction continues until the network forms a compact mass at one side of the nucleus. During synapsis the nucleoli may be entirely confined within the contracted sphere or one or more may be partially extruded (fig. 7). Some of the nucleoli still stain deeply with the gentian-violet, but one or more usually take the plasma stain at this time and appear as yellow, porous, or spongy bodies. The same appear- ance has also been obtained with iron-hasmatoxylin followed by orange G. In Pinus rigida no appearance at all comparable with that known as synapsis is observed until April 21. In material pre- served on this date a few nuclei in all anthers show the begin- nings of contraction as illustrated for P. aiistriaca in fig. 5 and P. Strobus in fig. 6. On April 30 the nucleus of every mother- cell has reached the point of greatest condensation, its contents forming a somewhat spherical, deeply-staining mass at one side of the nuclear cavity — fig. 7 illustrates this stage for P. Strohiis. On May 2 some of the nuclei still retain this structure while others show various stages of recovery. Two days later. May 4, not a vestige of this condition remains, all the nuclei having passed on to more advanced stages in the mitosis. These dates have been given for Pinus rigida, but they would not differ materially in the other species, except that in P. Strobus corresponding phases in this division would occur about 3 weeks later. Synapsis is not universally recognized as a normal step in the heterotypical division. Guignard ('97), Mottier ('97), Schaffner ('01), and others still look upon it as an artifact caused by im- perfect fixation. On the other hand, Sargant ('97), Wiegand ('99), Duggar('99 and '00), Ernst ('01), Rosenberg ('01) among 2 2 MARGARET C. FERGUSON botanists, and man}- zoologists consider it a definite characteristic of the early prophase of the heterot3'pic mitosis, several of these investigators having noted it in their material before fixation. I have observed this stage in the fresh material in Pimis, and after carefully studying it in many permanent preparations, I see no reason why this condition, simply because it happens to be one of contraction of the nuclear substance, should be set down as abnormal. If this appearance were produced artificially why should there be transitional forms both in leading up to and in recovery from it? If it were the result of diffusion currents, as has been suggested, we should expect to find the nuclear substance in all the nuclei of a given anther carried or forced to the same side of the nuclear cavity, but such is not the case. It is doubt- less true, as indicated by Strasburger ('95), that many phenom- ena described as synapsis represent pathological conditions which do not occur under all circumstances, but it seems equally true that this condition of the nuclear substance represents, in some species at least, a characteristic stage in the heterotypic division. Although a contraction comparable with that of synapsis has been reported for somatic cells, I am not aware that anything like so marked an appearance has been described as a usual accompaniment of any but the heterotypical division. The exact significance of this phase is not well understood, but that it is intimately associated with a readjustment of the chromatic and nucleolar substances there can be little doubt. As the nucleus slowly recovers from synapsis, it soon becomes apparent that the reticular structure has been replaced by a broad, closely coiled band which stains more deeply than did the net- work prior to the contracted stage. The coils of the thread gradually open out until the nuclear cavity is filled with a spireme, which consists of a broad linin band, so irregularly studded with chromatin-granules that it has a much roughened, almost minutely echinulate, appearance. These granules soon collect into indefinitely outlined masses which remain connected by clear, faintly staining portions of the linin thread. The chro- matin-groups never assume the definite disk-like form figured by Mottier ('97) for Lilitun and IlcUcborus, and by Duggar (00) LIFE HISTORY OF PINUS 23 for Symflo car-pus^ but they remain always irregular and jagged in outline (figs. 8 and 9). Whether there is one continuous thread or more than one could not be determined with certainty, as the coil is at first very densely massed, and free ends might be obscured. When the loose skein fills the nuclear cavity more than one spireme can usually be detected, but the indica- tions are that this effect has been produced by the microtome knife. At certain places the coils of the spireme run together and appear to be more or less anastomosed. Such a point of contact alwa3's indicates the position of a nucleolus which has become almost obscured by the massing of the thread about it, figs. 9, 13 and 15. Not all the nucleoli are found thus associ- ated with the skein, but in those cases in which they are free from the coils of the nuclear thread their capacity for staining has generally been greatly reduced (figs. 9, 11 and 15). As soon as the chromatin-band has become loosely wound about the entire nuclear cavity, longitudinal splitting occurs, and the segmentation of the spireme becomes apparent (fig. 10), but transverse fission is not completed until the longitudinal division has taken place (fig. 11). The segments are long, coiled, and present various appearances. Whether they correspond in number to the number of chromosomes eventu- ally formed, I could not ascertain with any degree of certainty, since they are so long and closely intermingled in the nucleus (fig. 11). Most of those shown in figs. 12 and 12, «, were taken from sections through the edge of nuclei, and, while they rep- resent the looped and twisted condition of the chromatic seg- ments at this time, they have in many instances been cut during sectioning so that only a portion of most of the segments appears. From a study of many nuclei containing chromatic threads similar to these, it is evident that the looped figure has not been formed by the bending on itself of one of the longi- tudinal halves of a segment. There are no indications that the sister-halves of any portion of the nuclear band ever become entirely disassociated. They may separate widely at one or both extremities, but at some point along the thread, an inti- mate relation is permanently maintained. The loop arises, therefore, by the complete fusion of the sister-threads at one of 24 MARGARET C. FERGUSON their free ends (fig. 12, «, r, J, e). Even in such a late stage of fission as that represented in fig. 13 the sister threads can ahnost invariably be traced, but not always, as some are out of focus and others are doubtless in another section. The stages immediately following longitudinal splitting and segmentation of the nuclear spireme are somewhat different from any that I have seen described by other writers. So puzzling were they to me when the study of microsporogenesis was first undertaken in 1899 that a paper, partially prepared at that time, was laid aside until a larger experience with cell structures could be brought to bear upon this, which is to me at once one of the most intricate and interesting problems con- nected with the activities of the cell. As stated in the intro- duction, new material was collected in 1901 and fixed with great care. Many slides were subsequently prepared, and the phases in the tetrad-division were found to accord perfectly with those observed during the first period of study. The interpretation of the phenomena noted is, however, much more satisfactory now than formerly, although there is still much that is obscure. Sporogenesis has not been studied in Pimis montana var. un- ct7iata, but there is complete accord, except in such details as have already been mentioned, in the other four species. Longitudinal division is scarcely more than completed when the double skein begins to contract, the two halves of each seg- ment twisting upon each other to a greater or less degree and gradually fusing. As the segments contract the sister-halves may frequently become more or less twisted upon each other ; they may appear as parallel threads ; the half segments may separate at both ends, remaining united at the middle only ; or, having fused at both extremities, they may open out, forming rings (figs. 12 and 12, a). Fusion invariabl}^ begins first about those nucleoli which have still retained, although in a less degree than prior to S3'napsis, the power to react to chromatin-stains (fig. 13). Contraction and fusion continue until a coarse, more or less anastomosing structure is formed in which only traces of the earlier longitudinal division re- main evident (fig. 14, plate II), and a little later all signs of fission, both longitudinal and transverse, disappear (fig. 15). LIFE HISTORY OF PINUS As the thread thickens and broadens it becomes irregular in outline, the irregularities increase, those from neighboring por- tions of the threads meeting and fusing. Soon afterwards a transverse division again becomes apparent (tig. i6). The segments continuing to shorten and thicken gradually draw away from one another, finally remaining united only by delicate threads ; the connecting fibers are at last severed and the chromosomes lie free in the nuclear cavity. The usual number of segments formed is twelve, although thirteen, four- teen, and, in rare instances, as many as sixteen have been counted (figs. i6, 17, 18, a-c, and 20). The chromosomes thus arise from an incompletely reticu- lated structure rather than directly from the spireme. While this suggests the condition in magnolia w^here, as recently described by Andrews ('01), the chromosomes arise directly from the resting reticulum without the intervention of a spireme, it is, in matter of fact, very different. We have here not a nuclear reticulum in the ordinary acceptation of that term, but a somewhat reticulated structure formed by the anastomosing with each other, at certain points of contact, of adjacent por- tions of a previously longitudinally split spireme. As the chromosomes separate out almost every conceivable form may be found, not only the X's, Y's and V's of Belajeff, but rings, parallel rods, eights open and closed, L's, U's and irregular- shaped bodies (fig. 19, a-l). In my earlier study of this phenomenon, I supposed the chromosomes to be the equivalents of the long, coiled segments first formed, and with such an hypothesis the whole series of events following longitudinal fission was inexplicable. But after again considering not only such stages as those represented in figs. 10-17, but every transitional form connecting them, I am convinced that this assumption was incorrect and that each seg- ment consists, rather, of two distinct chromosomes standing side by side, each half of the double chromosome represent- ing two sister-segments which were formed by the earlier longi- tudinal fission but have now fused. If such be the origin of these chromosomes, and I no longer have any hesitancy in affirming that they have thus arisen, the phases following the Proc. Wash. Acad. Sci., July, 1904. 26 MARGARET C. FERGUSON longitudinal and transverse divisions of the skein are no longer unintelligible. The sister-threads formed by the longitudinal splitting not only unite again, but adjacent portions of the double threads draw together and become more or less fused, giving rise when transverse fission again becomes apparent to the one half number of chromosomes. The forms of the resultant chromosomes are exactly what would be expected from such an origin. In fig. i8, h^ for instance, adjacent portions of double segments have fused at the ends, trans- verse division has followed, and three chromosomes — parallel rods, a U, and a Y, are seen in the act of separation. When the component chromosomes have fused at both ends only, the ring, or, if a twist follows, the closed eight results ; if fusion has occurred at but one extremity the V, U, or open eight is formed ; if the segments remain attached at the middle point the X occurs ; when the constituents of the double chromosomes have united end to end and the bend has not taken place at the point of their union the L results and so on. The structure or composition of the X, Y and V forms of chromosomes as found in plants have been explained in much the same way as the above by Belajeff ('97 and '98), but he did not trace their development from the closed spireme and considered these three forms as the typical or characteristic ones whereas, in Pinus, the other forms named have been quite as frequently observed. When the chromosomes first become apparent, irregular fragments of the chromatic substance are frequently left at various points (fig. 17), but these are ultimately absorbed, doubtless being appro- priated by the growing chromosomes (fig. 20). ^ At the time when the chromosomes are being differentiated, they often appear as if pulling away from the nucleoli, and may be seen still connected with them b}-- delicate threads (figs. 18, a and c). The nucleoli now have a spongy or porous appearance and fail almost absolutely to take either nucleolar or chromatic stains. With the final separation of the chromosomes they dis- appear altogether. The history of these nucleoli from the primitive archesporium up to the time of their dissolution leads irresistably to the conclusion tiiat here, at least, there is a very ' See " Note " at close of Appendix. LIFE HISTORY OF PINUS 2*J intimate relation between nucleolar and chromatic substances. Whether the nucleoli are actual reservoirs of chromatin which is given out passively to the chromatic thread, or whether they are actively engaged in furnishing nourishment to the chromatic substance, I have not been able to determine, but, from certain observations to be described in a later chapter, I am inclined to consider them more than passive elements of the cell. Coordinate with the formation of the chromosomes the nuclear membrane resolves itself into a weft of threads which crowd into the nuclear cavity, together with delicate granular fibers from the cytoplasm. The latter are evidently formed by a re- arrangement of the granules of the cytoplasmic reticulum. Up to this time the cytoplasm has remained close meshed in the region of the nucleus but has become less dense at the periphery of the cell. As the nuclear membrane disappears, coarser reticulations arise in the cytoplasm and extend towards the nucleus, doubtless contributing to the forming spindle. When the achromatic figure is fully developed, the cytoplasm again becomes uniform in structure throughout the cell, but there seems to have been an actual loss in granular substance, the meshes of the network being much larger now than formerly (figs. 20 and 21). A few delicate fibers may be seen in the cytoplasm just before the dissolution of the nuclear membrane, but, although I have searched repeatedly for cytoplasmic phe- nomena such as that described by Mottier ('97 and '98), Duggar ('00), Juel ('00) and others, I have never been able to detect anything at all comparable with the structures figured by these authors. If they are present in Pinus, I have not been able to differentiate them with any of the stains used. The spindle is almost invariably tripolar in origin, but it may arise as a multipolar diarch. In either case, its ultimate form is that of a sharply pointed bipolar spindle (Figs. 21-24). Belajeff ('94) describes this spindle as many poled in origin in Larix, and Mottier ('97) makes the same statement for Pinus; but in the many thousands of karyokinetic figures observed for this division, I have never found one that showed more than three poles. A few scattering fibers have occasionally been seen to pass from all sides towards the nucleus but achromatic threads have not been found to converge at more than three points. 28 MARGARET C. FERGUSON As the spindle-tibers press into the nuclear cavity, the chro- mosomes take up their position at the equatorial plate. They are now verv regular in outline, apparently homogeneous, and the X. Y, V, O, etc., forms can still be clearly distinguished (hg. 24, plate III). Each segment is oriented with its longer axis perpendicular to the axis of the spindle, the free limbs ex- tending outward. The spindle-fibers are attached at one ex- tremity of the parallel rods, and ordinarily at or near the point of union of the constituents of the dual chromosomes. In the Y-shaped chromosomes the achromatic threads may become at- tached at the point where the two limbs become free or at the free end of the fused chromosomes, but, whatever the shape of a segment, the spindle-fibers are never attached at the extremi- ties of its free limbs. The line of cleavage at the equatorial plate is not such as to separate the two chromosomes but is rather such as to effect a longitudinal splitting, the two half chromosomes of each pair passing together to opposite poles. During metakinesis the daughter-chromosomes become very irregular in outline and in- crease much in size, the half chromosomes apparently exceeding in volume the undivided ones (figs. 25-2S). This augmentation of the segments maybe due to actual addition of new substance, but from the fact that in the telophase they are unquestionably smaller than in the late prophase, it is probable that this is merely an amplification without actual or permanent growth. The parts of the spireme separated during the longitudinal fis- sion following synapsis have so completely fused again that they are now disunited with difficulty. The appearance of the dividing chromosomes indicates that they are being subjected to great strain. Under this tension they are flattened out and rendered irregular in outline ; the irregularities result from the unequal stretching of the chromatic substance at different points, just as a poor rubber band when greatly extended be- comes more or less moniliform. The complete separation of the half chromosomes may sometimes be greatly delayed, when the stretched segments extend nearly the entire length of the spindle, the achromatic figure being almost obscured, in some instances, by the chromosomes (figs. 25, 26, 2S and 29). That LIFE HISTORY OF PINUS 29 these segments are actually flattened out is further shown by the fact that the arms which remain united and elongated stain much less deeply than do those which, having become free, have contracted to nearly their former length. This would seem to indicate that the chromatic spireme is a plastic or viscid body. Lloyd ('02) describes a similar action, though much less marked, in Crticianella. While the position of the retreat- ing half chromosomes is such as to give ordinarily the appear- ance of V's or U's, other figures occur with sufficient frequency to establish the reality of their persistence after the close of the metaphase of the division. This point will be considered more fully later. The achromatic figure increases but little in length as the chromosomes pass to the poles so that the movement here must be due in large measure to a pull exerted by the contracting fibers and not to any great extent to a push brought about by the growth of the central spindle. If the force which seems necessary to effect the separation of the half chromosomes is furnished by the achromatic fibers, we should expect to find the poles of the spindle firmly buttressed as described by Stras- burger ('00) for Larix ; but no strengthening fibers are devel- oped, and, although the apices of the spindle are usually inserted in the ectoplasm, they not infrequently end blindly in the cytoplasm. It is possible that the force exercised by the growing fibers of the central spindle just equalizes the counter force exerted by the mantle fibers in drawing the chromosomes to the poles, the equilibrium thus established giving rigidity and rendering a support for the poles unnecessary. By the time the pairs of daughter-chromosomes have reached the poles they have become much reduced in size and regular in contour (figs. 27 and 30). After the chromosomes reach the point where the daughter- nuclei are to arise, they do not at once fuse end to end to form a continuous spireme, but as the chromosomes lie side by side they lose their clear outline and gradually assume a diffuse reaction to stains. In this condition the halves of the longi- tudinally split pairs of chromosomes are doubtless fused, after which fusion the adjacent segments unite by their ends to 30 MARGARET C. FERGUSON form a coiled, somewhat moniliform thread (figs. 30-32). Immediately upon the formation of the skein a delicate nuclear membrane appears, the coils loosen somewhat and branch freely thus giving rise to a reticulum. Extensive growth follows and a large " resting" nucleus is formed (figs. 33 and 34). The nuclear net consists at first of delicate achro- matic linin threads bearing scattered chromatin-granules and uniting large irregularly branched chromatic portions. Distri- bution of the chromatin continues until there is a delicate linin reticulum with chromatin granules of varying sizes imbedded in it (figs. 33-35). These nuclei have the form of a plano- convex lens the flat side of each nucleus being perpendicular to the axis of the spindle and facing the other daughter-nucleus. It is obvious from the foregoing that a definite resting nucleus is formed in Pinus at the close of the heterotypic division. This accords with the recent observations on the formation of the microspore by Duggar ('99) in Bignonia, Strasburger ('01) and Gager ('02) in Asclc^ias and Andrews ('01) in Magnolia. A true nucleolus has not been observed in the daughter-nuclei. Contrary to the observations of Hofmeister ('51), no cell-wall is laid down and in only a very few instances has a slight thickening of the spindle fibers in the region of the cell-plate been observed. The Second Mitosis of the Mothei'-cell. — The resting daugh- ter-nuclei are scarcely more than established before the initial steps of the second division are instituted, as evidenced in the readjustment of the nuclear reticulum. The more delicate threads of the net are withdrawn, the nuclear membrane fades out, the chromatin loses its granular aspect and becomes evenly distributed upon the linin, and there issues forth a heavy, homo- geneous, deeply-staining band which is more or less coiled and branched (fig. 36). The chromatin-thread, which now lies free in the cytoplasm of the mother-cell, continues to thicken, the branches or cross fibers disappear, and in an almost incredi- bly short time, the delicate nuclear net has given place to a broad, somewhat spirally coiled skein (fig. 37). Achromatic threads arise in the cytoplasm forming a multi- polar diarch spindle. The fibers are not abundant and always LIFE HISTORY OF PINUS 3 1 arise in a plane perpendicular to the axis of the primary spindle. Harper (bo) makes the statement that in Larix^ where no cell- wall follows the first division of the pollen-mother-nucleus, the spindle-fibers of the primary mitosis are utilized in the formation of the spindle for the second division, lam unable to trace any such connection in the pollen-mother-cells of Pinus^ all traces of the first karyokinetic figure having been lost to view before the inception of the spindle for the second division. As the kinoplasmic fibers appear the chromatin-band forms a double row of loops extending across the spindle-threads in the plane of the equatorial plate. The longitudinal splitting is now clearly apparent. The loops continue to shorten, and in this position transverse fission occurs, segmentation almost always taking place at the outer free ends of the loops (figs. 38 and 39, plate IV). The sister-halves of each V- or U-shaped chromo- some entirely separate, undergo readjustment, and finally come to stand in a double row with their free ends in the line of the nu- clear plate and their angles towards their respective poles (figs. 38-41). The spindle-fibers become attached to the chromosomes at their point of bending, and the half chromosomes pass to the poles (figs. 42-43). The dissociation of the sister-halves of each segment is so complete before the beginning of the separation at the equatorial plate that the figure during metakinesis is such as to give the impression of whole chromosomes passing to the poles, but a study of the prophases of the division shows clearly that each represents the half of a double chromosome. In the telophase of the division the chromosomes unite end to end to form a spireme (fig. 44). The nuclear membrane appears, and the chromatic band branches, giving rise to the reticulum of the resting nucleus (figs. 44 and 45). The Probleju 0/ Reduction. — Here as in all studies of spore- formation at the present time the question of reduction demands consideration. As already indicated, the reduction in the num- ber of chromosomes takes place, as is the rule, during the so- called resting stage of the spore-mother-cell, the one half num- ber of chromosomes appearing in the prophase of the hetero- typical division. But the inquiry concerning the presence or absence of a qualitative reduction is not so easily answered. 32 MARGARET C. FERGUSON With few exceptions, botanists of to-day follow the present lead of Strasburger and accept the view of a double longitudinal splitting of the chromosomes in the first division of the spore- mother-cell. According to this interpretation, reduction, in the sense in which Weismann uses the term, does not occur in plants. Among the exponents of a qualitative or true reduction in plants, Atkinson ('99), Belajeff ('97, '98), Calkins ('97), Ishikawa ('97, '01), and Schaffner ('97, '01) are almost alone to-day in not having retracted their earlier published conclusions regarding this subject. It has seemed best to record the details of the observations made in studying the tetrad-division in Piiiiis, before entering upon any discussion of the significance of the phenomena noted, but in so doing some reiteration is inevitable. Strasburger's statement that certain forms of chromosomes occurring in the anaphase of the heterotypic division are inex- plicable on any other assumption than that of a double longi- tudinal splitting is, doubtless, correct when those forms have been derived from V-shaped chromosomes. But, while it may be true that such figures are due to a double longitudinal fission when derived from other than V-shaped chromosomes, it is like- wise true that, in such cases, the phenomena are capable of rational explanation on other grounds. The V with the three arms, for instance, may result from the attachment of the spindle fibers at the middle point of a Y, the stem of the Y bending down as it moves to the poles (fig. 30, a, plate HI), and a double V might be derived in the same way from an X-shaped chromo- some (fig. 30, c). In fig. 26 the second chromosome on the left represents a Y opening out from its lower extremity, and the next chromosome shows parallel rods just separating. Occasionally an X or Y figure becomes apparent in the late anaphase of this division (figs. 28, 29). Such appearances are doubtless to be attributed to an early straightening out of the segments. If the constituents of the double chromosomes are disunited in this mitosis, then such chromosomes as those illustrated in figs. 28, d^ and 30, a, c^ and «?, might result from the more or less com- plete longitudinal fission of the sister-segments. Should this prove to be the case, and if my interpretation of the origin of LIFE HISTORY OF PINUS 33 these chromosomes is correct, then both a quantitative and a qualitative reduction of the chromosomes would occur in the first or heterotypic division, and whole chromosomes, each representing the half of a dual chromosome, would pass to opposite poles. I am aware that such a phenomenon has been described by Atkinson and a few others, but after long and care- ful study there does not seem to me the least doubt, that, in the case of the pines investigated, a longitudinal fission, and not a transverse one, occurs in this first mitosis ; and X-, Y-, and ring-shaped segments, as well as V's, pass to the poles, although, as Belajeff has pointed out, they usually, because of their posi- tion, have the form of V's in the anaphase of this divison. Most writers on sporogenesis, and especially those who are advocates of the true reduction, have not found a resting nucleus intervening between the heterotypic and the homotypic divisions. As already stated a resting nucleus is clearly demonstrated at this point in Pinus. The spireme formed from this nucleus shows signs of longitudinal division before segmentation, and, while lying at the equatorial plate, the two halves of each seg- ment separate entirely, in most instances at least, before their final orientation on the spindle. Now the question arises as to whether or no this homotypic division effects a qualitative reduc- tion. If the theory of the so-called " individuality of the chromo- somes " is without foundation then it certainly does not ; but, if the possibility of the complete rehabilitation of the chromosomes be accepted, a qualitative reduction very probably does occur. For under such conditions, the skein preceding the homotypic division would consist of the daughter-chromosomes, formed as a result of the heterotypic mitosis, fused end to end. These daughter-chromosomes, it will be remembered, arose by the longitudinal fission of a double chromosome and each, therefore, consists of a pair of half chromosomes. Thus the second, apparently longitudinal, splitting would effect the separation of the half chromosomes of each pair, rather than the longitudinal fission of a single chromosome. Reduction would thus take place in the true or Weismann's sense. Because of certain phenomena to be described in connection with the development of the pro-embryo, I am inclined to believe that the chromo- 34 MARGARET C. FERGUSON somes retain their individuality through succeeding cell-genera- tions. I am, therefore, disposed to regard the tetrad-division in Pinus as a true reducing division ; in this way only does the complicated process just described find satisfactory explanation. No positive statement can, however, be made either way, in connection with this division in Pinus, until we are in posses- sion of greater knowledge than at present of the origin and ulti- mate destiny of chromosomes. Guignard ('97) expresses the opinion that the regularity of the chromosomes in certain forms has been overestimated. Be that as it may, I am conscious that there is recorded in this paper a greater variation in the forms of the chromosomes than has been described in a single genus by other writers. It has been my purpose to note not only that which is in accordance, or at variance, with the observations of other investigators, but to give as faithful a record as possible of the conditions found in the preparations studied. And may we not yet find that here, in the divisions preceding spore-formation in plants, as in many other instances, there is greater variation in matters of detail than was formerly supposed to be the case? DEVELOPMENT OF THE MICROSPORE. The For7nat{on of the Sforc-zvall. — Hofmeister ('51) de- scribed four " special " cells, each with its own wall, within the pollen-mother-cell in the AhietinccB, and Juranyi ('72 and '82) devoted particular attention to the formation of the wall of the microspores in many Gymnosperms and Angiosperms. He described the development of a wall separating the two nuclei after the first division. This wall was soon absorbed and during the second division the entire cell was filled with connecting fibers stretching between the four nuclei. Delicate walls were then laid down between the nuclei giving rise to the four microspores. These dividing walls thickened and united with the inner wall of the spore-mother-cell : thus a portion of each spore-wall was formed from the inner mother-wall. After a period of rest the outer mother-wall was burst and the " pollen- cells " became free. If there is any recent literature of value on this subject, I have failed to find references to it. LIFE HISTORY OF PINUS 35 As already indicated, no wall separating the daughter-nuclei is formed at the close of the heterotypical division in Piniis. During the late telophase of the second mitosis in the microspore mother-cell, a readjustment of the spindle-fibers occurs giving rise to the complex figure that has been described as character- istic of spore-formation in many plants. The development of the archoplasmic structures connecting the nuclei of the tetrad is much less marked than in PodofhylliLm (Mottier '97) and in many other phanerogams (fig. 44). By the time the nuclei have reached the resting stage, a division has occurred in the cytoplasm giving rise to four cells which are surrounded by delicate clear walls. A prominent thickening of the wall of the spore-mother-cell takes place, and at the same time a thick wall, continuous with the inner portion of the mother-wall, appears between the daughter-cells. This wall frequently attains remarkable thickness. Whether it constitutes an inner wall, or is merely a thickening of the primary wall by the deposition of new material on its inner sur- face, I am unable to say. The outer, primary wall stains more deeply and is frequently seen separated from the inner broad portion (figs. 44-47). This inner wall, which is continuous with the broad walls separating the young microspores, stains deep yellow with orange G, if the orange is allowed to act from one to two minutes ; it appears a pale rose when treated with safranin, but fails altogether to stain with iron-haematoxylin. In a few instances, slight evidences of stratification have been observed, but ordinarily the wall appears perfectly homogene- ous, giving the impression of a liquid or viscid substance in which the spores are imbedded ; but the fact that it is often separated from the outer wall by a clear space, and also that it is left behind as a definitely outlined wall after the escape of the spores militates against the probability of its fluid nature. After the spores have grown for a certain period the mother- wall is ruptured and the spores are liberated. At this time the empty mother-cell with its four chambers is often met with (figs. 48, 49). In so far as I am aware, this permanent division of the mother-cell into four compartments by thick cellulose walls has 36 MARGARET C. FERGUSON not been previously described. A broad open space, repeatedly figured between the daughter-spores and the mother-wall, has been invariably attributed to shrinkage ; but it is probable that, in some cases at least, it represents this thickened wall which has failed to be differentiated with the stains used. Wiegand ('99) says that the spores of Potamogeton are as if imbedded in a ground mass of some viscid substance, but he does not figure it and makes no statement regarding the development of cell-walls between the microspores. Origin of the Air-sacs. — As soon as the young microspores have become enclosed, each within its own special chamber of the mother-cell, it is evident that a special wall has been de- veloped about each spore. This is doubtless secreted by its own cytoplasm and is not, as Juranyi thought, derived from the inner wall of the microspore-m other-cell. The spore-wall while still very delicate becomes differentiated into an inner and an outer layer corresponding to the intine and extine of the pollen- grain. The young microspores are characterized by the rela- tively large size of their nuclei, the nucleus filling almost the entire cell just prior to the discharge of the spores. The cyto- plasm which fills the remainder of the cell is in the form of a loose reticulum (figs. 46, 47). As time goes on the outer wall of the microspore expands at two points on opposite sides of the spore. A resistance is met with in the thick wall of the spore-mother-cell and the plastic inner wall of the microspore responding to this new pressure becomes indented along the surfaces corresponding to the ex- tended portions of the outer spore-wall. Thus a clear open space having in section the form of a biconvex lens is formed between the extine and the intine on either side of the microspore. These are the beginnings of the wings or air-sacs that are so conspicuous in the mature pollen-grain of the AbietinccB. Finally the pressure becomes so great that the mother-wall is ruptured and the spores are liberated (figs. 47, 48). Coulter and Chamberlain ('01) noted the fact that the wings make their appearance in Pinus Laricio while the microspores are still within the mother-cell, but they recorded no observations regard- ing the origin and development of these sacs. Strasburger and LIFE HISTORY OF PINUS 37 Hillhouse ('oo) consider that these bladder-like appendages con- sist of the outer part only of the extine, the extine having under- gone cleavage at these two points. In studying the develop- ment of these organs from their earliest beginnings, it appears to me that the line of cleavage lies rather between the two coats of the young spore. If it is not, then at the time that the micro- spore leaves the parent-cell, the intine has not been developed, or, if present, is so delicate that I have not been able to detect it (fig. 48). Growth of the Microspore. — After its escape from the mother-cell the microspore undergoes rapid growth, and the outer surface of the spore becomes beautifully marked by the formation of delicate, irregular ridges over the entire inner sur- face of the extine, except along that portion which connects the two wings on the concave or ventral side of the pollen-grain. It is at this point that the pollen-tube later makes its exit, and there is here no appreciable thickening of the spore- wall. These ridges continue to grow and extend inward forming a very pretty reticulated structure which is most distinctly apparent on the walls of the wings ; along the convex or dorsal side of the pollen-grain the reticulations are closer and the extine forms a broad, deeply staining layer (figs. 50-54, plate V). This irregular thickening of the extine is an admirable adaptation for securing strength with slight increase in weight. When the young microspore attains to its mature-size, a par- tial wall, extending along the back and for a longer or shorter distance down the sides of the spore, becomes apparent within the intine (fig. 54). It consists of a broad, homogeneous- appearing band which gives precisely the same staining reac- tions as the thick wall developed within the spore-mother-cell after the formation of the young microspores. These immature pollen-grains, after treatment with Flemming's triple combination or with the gentian-violet and orange G alone, afford the most brilliant effect that I have observed with these stains. The extine presents a very intense, clear blue, the inner homogeneous wall an equally vivid yellow, while the protoplasmic elements take the colors characteristic for these dyes. The fact that this third partial wall fails entirely to respond to some stains doubtless 38 MARGARET C. FERGUSON accounts for its haviug been overlooked by previous writers. It is not shown at all in the series of figures, recently published by Coulter and Chamberlain ('01), illustrating the development of the pollen-grain in Pinus Laricio. The various tests commonly used in determining the nature of the cell-wall have been applied to the young pollen-grains as well as to the special spore-mother-walls. These tests show that the outer wall of the pollen-grain is clearly of the nature of cutin, as has been demonstrated by Strasburger. Both the innermost wall of the microspore, and of the pollen-grain, as also the wall of the special spore-mother-cells, respond to the reaction for cellulose, but not in a very marked manner. If they are of the nature of cellulose there would seem to be an admixture of some other substance, but I have not succeeded in obtaining entirely satisfactory results regarding the nature of these inner, prominent walls. Tests thus far have been made with "fixed" material only; further experimentation along this line will be made when fresh material is at hand. During the season of growth, the nucleus of the microspore always remains close against the convex or dorsal side of the spore, occupying a central position along this wall. As is usual in cell-development, the microspore-cell attains full size before any mitoses occur within it, and there is never any fur- ther increase in the size of this cell after the inception of the first division. The fully developed microspore is, therefore, the exact counterpart, so far as size is concerned, of the mature pollen-grain. Compare fig. 54, plate V, with fig. 65, plate VI. During the development of the microspore, the cytoplasm which at first was uniformly distributed in a rather loose net work, becomes more closely reticulated and at the same time less abundant in proportion to the size of the cell. At the maturity of the spore the cytoplasm is largely distributed about the nucleus from which strands extend outward in a radial man- ner and end in the ectoplasm. In 1898 the microspores of Pinus Strobns were ready to leave the mother-cells on May 30, they had attained full size on June 7, and on June 10 the pollen- grains were fully mature. LIFE HISTORY OF PINUS 39 SUMMARY. In Finns rigida, P. austriaca and P. resinosa the primitive archesporium is well developed before the approach of winter, but the microspore-mother-cells do not arise until the end of the following April. The male inflorescence does not appear in Pinns Strobus, until the end of the April preceding pollination, and the definitive archesporium is differentiated in this species about the middle of May. The nuclei of the primitive arche- sporium are characterized by several deeply staining nucleoli and a fine, close-meshed reticulum which responds but slightly to chromatic dyes. The wall of the pollen-sac consists in all cases of from three to four layers of cells. The tapetum is not clearly distinguished until spring and there are indications that it may be derived from the outer layer of sporogenous tissue. The nuclei of this tissue multiply mitotically and the cells reach their maxi- mum size about the time when the microspores become free. At this period each cell has from one to three nuclei which pre- sent all stages of fusion. When the pollen-grains are mature the tapetum has entirely disappeared and the wall of the micro- sporangium consists of a single layer of cells, or at most of not more than two. Synapsis is recognized as a normal stage in the prophase of the heterotypical division in the pollen-mother-cell of Piniis. It is not preceded by a definite spireme, but a broad skein con- taining irregular masses of chromatin separated by clear portions of the linin thread issues from the contracted nuclear mass. The chromatic spireme splits longitudinally and breaks up by transverse fission into several segments. The loosely coiled, delicate threads resulting from the longitudinal division soon draw together and fuse, double threads also come into contact at various points and fuse more or less perfectly. These threads always anastomose most freely in the region of the nucleoli, some of which still stain deeply while others stain but faintly after synapsis. Fission occurs at various points in the now irregularly con- tracted and anastomosed threads, and the separate chromosomes, 40 MARGARET C. FERGUSON in the reduced number, become apparent. These segments are at first irregular and jagged in outHne showing distinctly the points at which each has separated from neighboring segments, but they gradually diminish in size and become more regular in contour. The chromosomes thus formed are in the form of X's, Y's, V's, U's, L's, parallel rods, rings, and indefinitely-shaped bodies. Each segment consists of two chromosomes fused side by side. The spindle-fibers arise both from the nuclear membrane and from the cyto-reticulum. The achromatic figure ma}'- originate as a multipolar polyarch of three poles or as a broad multipolar diarch spindle. At the close of the prophase of the heterotypic division the spindle has become sharply bi-polar and its extremities may be imbedded in the ectoplasm or they may end blindly in the cytoplasm. The chromosomes are separated at the equatorial plate with difficulty giving the appearance of a plastic substance under tension. Their separation may be so delayed that the daughter- chromosomes stretch from pole to pole. They ordinarily have the form of V's or U's during the anaphase of the mitosis, but other forms are not infrequent. The first division effects a longitudinal splitting of the chromosomes into daughter-seg- ments of the same form as the parents. A resting nucleus is established at the close of the first mitosis but the daughter-nuclei are not separated by a cell-wall. The daughter-reticulum soon gives rise to a more or less spirally coiled chromatic band which loops itself at the equatorial pl^te and splits longitudinally before segmentation. The chromosomes have the form of U's and are oriented at the equatorial plate in two rows with their free ends touching and the bent portion of each segment directed towards the poles, the complete fission of the segments having been completed before their migration to the poles begins. The writer inclines to the view that these are the half chromosomes of the daughter- pairs which were separated in the first division. If this hy- pothesis be correct, the homotypic mitosis in Piiius effects a true or qualitative reduction of the chromosomes. The wall of the microspore-mother-cell increases markedly in thickness and its protoplasmic contents is separated into four LIFE HISTORY OF PINUS 4I parts by prominent cross walls which are continuous with the inner portion of the mother-wall. The microspores are then developed each in its own particular chamber of the mother- cell. A double wall is quickly developed about each spore and the air-sacs become apparent while the spores are still within the mother-wall. They arise by the separation of the extine from the intine at two definite points on opposite sides of the spore. By the growth of the spore, and more especially by the expan- sion of the air-sacs, the spore-mother-wall is ruptured and the spores set free. Growth ensues, the extine becomes irregularly thickened on its inner surface except at the concave side of the spore, and a broad partial wall is laid down just within the intine and along the back and sides of the microspore. During the growth of this cell its nucleus maintains a position at the central point of its dorsal side. Before the germination of the microspore it attains to the full size of the mature pollen-grain. CHAPTER II. The Male Gametophyte. the development of the pollen-grain. Formation of the ProthaUial Cells. — So much confusion has arisen in the application of terms used to designate the various cells of the male gametophyte in Gymnosperms that it is desirable, if not almost necessary, that one should define at the outset the nomenclature adopted. Throughout this paper, the first two cells cut off from the larger cell are known respectively as the first and second prothallial cells, and the third small cell formed represents the antheridial or third prothallial cell. The large cell, so long as it continues to divide, is designated as the apical cell, but after division ceases in this cell it is referred to as the tube-cell and its nucleus constitutes the tube-nucleus. The antheridial cell divides to form the stalk-cell and the gen- erative cell, the latter giving rise to the binucleated sperm-cell. Proc. Wash. Acad. Sci., July, 1904. 42 MARGARET C. FERGUSON As soon as the microspore has reached maturity, there arises within its nucleus one of the most beautiful, homogeneous, loosely-looped and coiled spireme-bands that I have ever seen in any dividing nucleus (fig. 54)« The material studied showed every stage in the first division, and all succeeding mitoses which occur within the microspore, but they offer nothing especially instructive from a cytological point of view, since they conform to the typic method of division. I shall, there- fore, describe and figure only such phases as are of interest in tracing the development of the pollen-grain. It is interesting to note that in the late prophase of all the mitoses which occur in the development of the male gametophyte the achromatic figure presents a very characteristic appearance, being sharply monopolar at its outer or lower extremity and broadly multi- polar at the opposite end. It thus forms a fan-shaped body rather than one resembling a spindle. During the telophase it usually becomes bluntly bipolar, though the upper pole often remains to the last somewhat broader than the lower pole (figs. 55, 56 and 60, and plate V. A similar method of karyokinesis has been noted by Wiegand ('99) in the development of the pollen-grain in Potamogeton^ by Duggar ('00) in Symploca7'pus^ and by Coker ('02) in Podocarpus. This mode of division will be referred to again in connection with certain phases in the development of the female gametophyte. In all the divisions which occur within the wall of the micro- spore the nuclear substance is divided equally, the cytoplasm unequally. The nucleus of the first prothallial cell, however, never equals in size that of the apical cell and always stains more or less diffusely, thus showing signs of disintegration from the time of its organization (fig. 57). Fig. 58 shows one of the very largest and most nearly normal of all the prothallial cells observed. The nucleus of the apical cell enters the complete resting stage, instituting a definite network within the meshes of which one or more faintly staining nucleoli become apparent, but this reticulum at once resolves itself into a homogeneous, spireme exactly similar to the one first formed. When the nucleus of the apical cell has reached the spireme-stage of the second division, the first prothallial cell is invariably found LIFE HISTORY OF PINUS 43 pushed against the dorsal side of the spore-wall, not a vestige of its cytoplasm is left, and the nucleus has become greatly flattened, although there is still a faint suggestion of its former reticular character (fig. 59). When the telophase of the divi- sion is reached this nucleus has lost all traces of its former structure and persists only as a deeply staining, linear body lying against the spore-wall (fig. 60). During the following division it becomes scarcely more than a line so that it is fre- quently detected with difficulty. Coulter and Chamberlain ('01) figure this cell in Pinus Laricio as still projecting into the cytoplasm of the apical cell when that cell is in the telophase of the second division, but I have never found it in such a state of preservation at so late a date. The second prothallial cell is invariably smaller than the first, and during the third mitosis of the apical cell, which follows immediately the formation of the second prothallial cell, it exactly repeats the history of the first cell (figs. 61-63). The partial, broad, innermost wall, described in connection with the development of the microspore, persists throughout the entire history of the pollen-grain, and a comparatively broad wall, continuous with it and having exactly the same staining capacity, invests both the first and second prothallial cells as shown in figs. 57-63. The presence of the remnants of the prothallial cells imbedded apparently in the inner wall of the mature pollen-grain (fig. 63) was very perplexing before the histor}'^ of these cells was studied. But in tracing their develop- ment it is clearly demonstrated that the remnant of each cell is pushed back against the wall of the spore and remains perma- nently covered on its outer side by its own wall. That the remains of these cells come to lie nearer the intine than when first formed would again suggest the somewhat plastic nature of the partial or incomplete membrane against which the pro- thallial cells are pressed (figs. 57-64). These observations con- firm the statement of Strasburger, Noll, Schenck and Schimper ('97) that the two prothallial cells formed in the pollen-grain of the Gymnosperms are invested with cellulose-walls. Coulter and Chamberlain ('01) make no mention of the formation of walls in connection with the development of these cells in Pinus 44 MARGARET C. FERGUSON Lartcio, and Coker ('02) says that in Podocarjbus " as in other cases " no cellulose-wall is formed. The small cell cut off by the third and last division of the apical cell persists as a perma- nent feature of the mature pollen-grain. Its cytoplasm is dis- tinctly differentiated from that of the tube-cell, but no cellulose- wall has been observed in connection with this cell, its boundary being marked by scarcely more than a condensation of its periph- eral cytoplasm. The Mature Pollen-grain. — During the development of the male gametophyte the cytoplasm of the large cell gradually increases in amount, the vacuoles becoming smaller from the region of the nucleus outward, and finally disappearing alto- gether. The pollen-grain has the same size, form, and, so far as the wall is concerned, the same structure as the microspore just prior to its germination. The thick, innermost, partial wall described in connection with the microspore still persists as a very prominent characteristic of the mature pollen-grain. With the expansion of the wings, certain protoplasmic portions of the microspore-cell are left with no support except the delicate endo- spore ; it therefore seems probable that this broad, incomplete wall extending along the back and down the sides of the pollen- grain has been developed for the purpose of strengthening these weakened points in the spore-wall, and as an additional support to the dorsal side of the pollen-grain. But, while the wall of the mature pollen-grain is identical with that of the microspore, the essential or protoplasmic part of the spore has undergone marked changes, as we have already seen. One or two deeply staining lines, more often one than two in the mature pollen-grain, lie on the dorsal side of the pollen-grain apparently imbedded in its innermost wall. Extending from this wall at its middle point is a strongly convex cell, the antheridial cell, with delicately reticulated cytoplasm and a comparatively large nucleus. Just below and always in contact with this cell is the nucleus of the tube-cell. The cyto- plasm of the tube-cell is closely reticulated and slightly more dense than that of the antheridial cell. Imbedded in its cyto- plasm are numerous starch-grains. In this condition the pol- len-grain of Finns awaits pollination (tigs. 64, 65, plate VI). LIFE HISTORY OF PINUS 45 Starch-grains have been found in the large cell from an early date in the development of the pollen-grain, but they are more abundant after maturity is reached than at any previous time. According to Coker ('02) the pollen-grains of Podocarfiis con- tain large starch-grains from the beginning of the first division. With such variations in details as have been noted above, this description of the development of the pollen-grain in Pinus agrees with that given by Strasburger in 1892 and Coulter and Chamberlain in 1901. ' POLLINATION. The Ovule at the Time of Pollination. — In the vicinity of Cornell University, 42^° north latitude, the pollen-grains of Pinus Strolms are ready for dispersion late in May or early in June, but in the other species studied pollination takes place during the latter part of May. At this time the axis of the female cone elongates, thus separating the ovuliferous scales which now make an angle of about thirty-five degrees with the rachis. After pollination the fruit scales draw together and, according to Strasburger and Hillhouse ('00), their edges are consolidated by the ingrowth of papillae. The presence of two ovules at the base of each scale, each ovule with its apex extend- ing downwards, that is towards the base of the scale, and out- wards, is too familiar a fact to need more than a passing men- tion here. As pointed out by Hofmeister ('62) the integument is con- tinued above the nucellus into two long arms which curve out- ward before pollination and lead below to a wide mycropylar canal. The degree of development which the ovule has obtained at the time when the pollen-grains reach the nucellus is shown in fig. 66. Deep within the central portion of the ovule, at its chalazal end, a single cell is distinguished from the others by its greater size and larger nucleus, this is the macrospore ^ of Hof- meister ('51). The so-called "spongy "tissue of Strasburger is already well differentiated when pollination takes place (figs. 66^ plate VI, and 124, plate XII). Somewhat later the integu- 1 In 1901, I stated that, at the time of pollination, there was in the nucellus an axial row of cells. I know, now, that this condition has rarely been reached at so early a date, and should be noted as very exceptional rather than as normal. 46 MARGARET C. FERGUSON ment has closed over the pollen-grains and the macrospore mother-cell has divided giving rise to an axial row of cells the lowest of which becomes the functional macrospore (fig. 69, plate VI). The Pollen- cha7iib er . — The pollen-grains fall upon a scale and slip down to its base where they come into contact with the extended arms of the ovule. These prolongations of the integu- ment now straighten and partially draw together thus bringing the pollen-grains down into the wide micropylar canal (fig. 123, plate XII, and fig. 66, plate VI). The free limb of the integument is seen in section to consist, at this time, of three layers of cells. As soon as the pollen-grains have found their way into the lower portion of the micropylar canal and some, at least, have come into contact with the tip of the nucellus, the cells constituting the middle layer of the arms, at a point slightly above the apex of the nucellus, elongate rapidly. The bulge or protuberance thus formed extends inwards from all sides and meets, closing the opening above the pollen-grains (figs. 66 and 67). As soon as the opening has been closed and the pollen-grains secured, these elongated cells give rise by division to many smaller ones (fig. 68). By the rapid elongation of these cells the safety of the pollen-grains is as- sured in a very short time, and then cell multiplication follows leisurely. This very pretty mechanism by which the final clos- ing of the micropyle is effected has not been previously described for any Gymnosperm, unless it be noted in Shaw's ('96) state- ment, unaccompanied by figures, that the micropyle in Sequoia is closed by the radial elongation of the cells about it. The depression in the apex of the nucellus in the Abielinece at the time of pollination, described by Hofmeister in 185 1, and since noted by man}?^ writers, has, it seems to me, been greatly exaggerated so far as Pinus is concerned. The expression " cup-like depression " is not infrequent in literature, but, in so far as my observations go, saucer-like is as strong a term as one is justified in using (figs. 66, 67 and 69, plate VI, and 75, plate VII). At the time of pollination the upper concave portion of the nucellus terminates in a row of more or less elongated cells, which are not closely united at their free extremities, but LIFE HISTORY OF PINUS 47 stand up, as it were, like so many fingers to catch the pollen- grains ; they also serve to facilitate the entrance of the pollen- tubes into the tissue of the nucellus (fig. 75, plate VII). A little later this depression may become more prominent, both by the slight disintegration of some of the superficial cells of the nu- cellus, due to the action of the pollen-tubes, and by the incon- siderable growth, after pollination, of the peripheral layer of cells of the nucellar tip. The deep cup-like depression some- times observed is invariably the result of abnormal disintegration. The pollen-chamber in Pinus, then, consists of a space bounded on the bottom by the more or less concave upper surface of the nucellar tip, and arched above by the ingrowth of the free por- tion of the integument. Later a resinous substance is secreted which securely seals the opening by which the pollen-grains entered. DEVELOPMENT OF THE POLLEN-TUBE. THE FIRST PERIOD OF GROWTH. Germination of the Pollen-grain. — Germination of the pollen- grain follows immediately after pollination. Ovules of Pimis Strobus that were fixed on June 6, 1898, had not been pollinated, but on June 13 pollination had occurred and the pollen-tubes had been emitted ; similar evidence could be given for the other species studied, but exact data on this point are at hand ior Pinus rigida only. Dispersion of the pollen occurred in this species in the vicinity of Wellesley College in 1902 on May 27, and in material fixed two days later. May 29, the first stages of germi- nation are clearly evident. It is probable that the time is not longer in the other species. This confirms Strasburger's ('92) statement that germination takes place in Pinus at once after pollination. Hofmeister ('5 1) was doubtless unable to detect the early stages in the germination and hence was led to the con- clusion that pollination and germination were separated by several weeks in the AbiciinccB. The pollen-grain increases slightly in size, the ventral or concave portfon of the wall becomes convex, then bulges out, the exospore is ruptured, and the endospore is gradually pro- longed into a tube. Immediately upon the formation of the 48 MARGARET C. FERGUSON pollen-tube the tube-nucleus, as shown by Strasburger ('92) moves away from the antheridial cell and into the pollen-tube (figs. 75, 76, plate VII). According to Coulter and Chamber- lain ('01, page 92), the tube-nucleus does not enter the tube until the following April. That the tube-nucleus should at once loose its association with the antheridial cell and accompany the growing point of the pollen-tube is exactly what we should expect from what we know, through the investigations of Haberlandt ('87) and others, regarding the relation of the nucleus to growth ; and, also, judging from the standpoint of analogy, from the remarkable migrations of the tube-nucleus in order to be near the growing point of the pollen-tube in Cycas (Ikeno '98) and in Zamia (Webber '01). Division of the Antheridial Cell. — Strasburger ('92) described the antheridial cell in Pinus sylvestris as remaining unchanged until the archegonia are formed in the following spring. Dixon states that it divides about a month before fertilization, but from a careful reading of the text one is given the impression that this was an inference on his part rather than a demonstrated fact, as he did not study material that was preserved earlier than April 24 and did not find the karyokinetic figure for this division. And, in so far as I am aware, this mitosis has not been observed in Pinus. Strasburger describes and figures it in Picea while the pollen-grain is still within the anther.^ I have found great variation in the time at which the anther- idial cell divides, not only in different species but in the same species. It is rather interesting that Pinus Strohus, which invariably lags somewhat behind the other species in all other developmental phases studied, is remarkably precocious as regards this step. Figs. 78, 80, and 81 were all taken from material of Pinus Strohus which was collected and pre- served on August 4, 1898, barel}' two months after pollina- tion. In the same material, other pollen-grains were observed in which the division of the antheridial cell had not yet taken place ; but in material fixed somewhat later it was rarely found undivided. The division of this cell has not been observed in Pinus austriaca, but two cells have been found in the pollen- grain in the middle of November and in February, and in such ' See note at close of appendix. LIFE HISTORY OF PINUS 49 instances the tube-nucleus can invariably be detected in the pollen-tube. As pollen-grains containing but one cell were also observed in this species on these dates, it might be suggested that in the case of two cells the second prothal- lial cell had persisted. The two cells, however, are exactly similar to the stalk and the generative cell in their young con- dition, and I see no reason for considering that they are not these cells. On and after March 8 the antheridial cell of P. austriaca is almost never found undivided. This date is given for 1899; it would probably fluctuate in different years. Fig. 79 shows the prophase of this division in Finns rigida. Mi- totic figures for this species have been found from April 21 to May 13 of the same season. The division of the antheridial cell in Pinus resinosa has been observed but once, this division occurring on April 11. All that can be said at present regard- ing this mitosis in Pinus montatia var. uncinata is that the gener- ative cell and the stalk-cell are found as early as April 9. When they are formed has not been determined. In one preparation of Pinus Strobus two of the three pollen- tubes which have almost reached the prothallium are furnished with sperm- and stalk-cells, while in the third only the tube- nucleus is found. On the apex of the nucellus there is a pollen-grain which at this late date contains one cell, the antheridial cell, still undivided (fig. 73). The nucleus of this pollen-grain (fig. 74) is large, plump, and to all appearances perfectly normal, and it is possible, though scarcely probable, that it might still have divided. That one cannot trace a defi- nite connection between the pollen-tube containing only the tube-nucleus and this pollen-grain signifies little, for those who have studied the pollen-tube of Pinus know that it is the excep- tion rather than the rule when a given pollen-tube can be traced through the lacerated dead tissue of the upper portion of the nucellus to the pollen-grain from which it proceeded. Such a condition as that described is rarely met with at so late a date ; but occasionally during the summer and fall pollen-grains of Pinus Strobus are found in which no cell-division has taken place since pollination, although in the great majority of cases 50 MARGARET C. FERGUSON the stalk- and the generative cell have been formed before the middle of August. These observations indicate that, while the division of the antheridial cell takes place comparatively soon after the pollen- grain has germinated in Pimis Strohus^ and in some instances, at least, before the winter's rest in P. austriaca, it is deferred until the following spring in Pinus rigida and P. I'esinosa. Furthermore, the time during which this cell may divide in a given species may extend over several weeks, and in some cases the division may never take place at all. The Winter Condition. — A vertical section of an ovule of Pinus Strobtis collected on January 4 is represented in fig. 70, plate VI. The spongy tissue surrounds a cavity crossed by irregular strands of cytoplasm in which the free nuclei of the prothallium are imbedded. In this instance the prothallium has doubtless been displaced during fixation as it consists, normally, at this stage, of a uniform layer of cytoplasm surrounding the gametophytic vacuole and containing several nuclei. The stalk- and the generative cell are enclosed within the pollen- grain, and the tube-nucleus is near the apex of the irregularly branched pollen-tube. This pollen-tube is shown more highly magnified in fig. 83, plate VIII. At this time the pollen-tubes have penetrated the nucellus almost to the point at which it joins the free limb of the integument. The greatest depth to which the tubes may have grown is not indicated in the illustration, but this section was figured because it shows more clearly than any other section in the series the cells of the pollen-grain and the tube-nucleus. Other sections of the same ovule would have shown pollen-tubes which had pierced to a greater depth into the nucellus. The conditions of development as figured for Janu- ary coincide perfectly with those which exist during the latter part of October. THE SECOND PERIOD OF GROWTH. Renewed Activities in the Macrosporangitim. — Growth is very slow during the first period of development following pol- lination, but with the renewed activities of spring the ovule increases rapidly in size ; the central cavity of the nucellus LIFE HISTORY OF PINUS 5^ becomes greatly enlarged and is lined with the growing endo- sperm. The cells of the nucellar cap which are penetrated by the pollen-tubes during the previous season do not again become active, but remain as deeply staining, thick-walled, dead cells. The cells just beneath them, however, multiply rapidly, and become literally packed with large starch-grains. A few of the cells from this portion of the nucellar cap represented in fig. 73, plate VII, are shown more highly magnified in fig. 89, plate VIII. By the growth and increase of these cells, the dead top of the nucellus with its pollen-tubes is lifted far above the developing endosperm, so that the pollen-tubes, once so near their goal, are now removed from it by a considerable distance (figs. 70-72, plate VI). Renewed Activities in the Male Gametofhyte. — During the rapid development of the ovule in the spring, the pollen-tube increases little, if at all, in length, renewed activities in the male gametophyte being first indicated by a further development of the cells within the pollen-grain. The stalk-cell increases in size and its cytoplasm assumes a vacuolate character. The growth of the generative cell is still more marked, and its cytoplasm on the contrary becomes dense and deeply staining. (Compare fig. 83, January 4, with fig. 84, May 3, plate VIII.) In Pinus syhestris, as studied by Dixon ('94) and confirmed by Coulter ('97) in Pimis Laricio, the gen- erative cell divides while it is within the pollen-grain. In the species of pines which I have investigated, this division does not occur until the generative and the stalk-cell have entered the pollen-tube and the stalk-cell has passed below the gen- erative cell. As the generative cell increases in size it stretches out towards and into the neck of the pollen-tube, drawing after it the stalk-cell, or possibly being forced out by that cell, the two passing into the tube together. Dixon states that only the naked nucleus of the stalk-cell enters the polfen-tube, and in so far as I am aware, no writer has described the entrance of the entire stalk-cell into the pollen- tube in Pinus. The material which I have studied shows con- clusively that the nucleus does not '* slip out " of its cytoplasm (figs. 83-86). The entire cell can be identified in the tube and 52 MARGARET C. FERGUSON later in the egg- During the time that this cell is moving over the generative cell its C3^toplasm cannot always be differentiated from that of the latter ; but when once the stalk-cell has passed the generative cell, its nucleus surrounded by a sphere of very vacuolate cytoplasm, scarcely more than a peripheral la3'er, is again distinctly demonstrated (figs. 90 and 91). After pass- ing the generative nucleus, the stalk-cell ordinarily takes up a position between the generative cell and the tube-nucleus (fig. 92), but occasionally it may pass the tube-nucleus (fig. 93). This phenomenon is always accompanied by a great in- crease in the starch content of the pollen-tube, the tube being in some instances almost filled with starch in the region of the generative cell (fig. 91). When the generative cell leaves the pollen-grain, its nucleus is situated near the top of the cell, but the nucleus of this cell evidently moves faster than its cytoplasm, and at the time when the stalk-cell is passing over the generative nucleus this nucleus has come to lie at or below the center of its cell (fig. 84, 90 and 91). Shortly after this the generative nucleus is again observed at the uppermost part of its cytoplasm. During its passage into the tube, the generative cell increases much in size ; it has no definite cell-wall, and its cytoplasm forms a large, irregular tongue about the nucleus. This cytoplasm in no way suggests the alveolar structure of Butschli ('94) but is distinctly reticular, differing in appearance from the nuclear net only by its greater delicacy. This is shown more clearly at a somewhat later stage. The tube- and generative nuclei are now very similar in structure, though each is sufficiently characteristic to be readily recognized by one who is familiar with them. The tube-nucleus has one large, usually homogeneously staining nucleolus, rarely one or more smaller nucleoli, and it is furnished with a rather scanty, delicate reticulum which is apparently poor in chromatin. Either it is in a state of partial collapse, or, what is more prob- able, it is very hard to fix at this period in its history, for its outline is, as a rule, quite irregular at this time. The genera- tive nucleus has one large, hollow or vacuolate nucleolus, and commonly two smaller ones; its reticulum, though more abun- LIFE HISTORY OF PINUS 53 dant than that of the tube-nucleus, is still delicate and often shows a weak reaction to nuclear stains. The stalk-nucleus has a very decided individuality which it maintains throughout its entire history. It bears a strong resemblance from the first to the nuclei of the nucellar tissue ; rarely, if ever, contains a true nucleolus ; and its close-meshed reticulum is conspicuous for its comparatively large net-knots or karyosomes. Division of the Generative Nucleus. — Comparatively few students have occupied themselves with the growth of the pol- len-tube in the Abietincce, and no one, in so far as I have been able to determine, has described the cytological features attend- ing the formation of the sperm-nuclei in this group. Dixon ('94) describes this division in Pinus sylvestris as tak- ing place about a month before fertilization^ while the genera- tive cell is still ivithin the pollen-grain ; and Coulter ('97) states, as already mentioned, that in his study of Pinus Laricio he has been able to confirm Dixon's observations in the minutest detail. At this time, as pointed out by Dixon, the nuclear and cytological phenomena are very greatly obscured by the pres- ence in the pollen-tube of large quantities of starch (fig. 91). The starch, which resists the microtome knife and is therefore easily displaced by it, not infrequently falls out and carries away with it the free cells of the pollen-tube. The dead, deeply staining tissue of the nucellus, representing that portion of the nucellar cap which was penetrated by the pollen-tubes during the previous season, and in which the generative nucleus divides (fig. 72, plate VI) is also very troublesome. Furthermore the dense cytoplasm of the generative cell has a great affinity for stains, so that when the archegonia and other portions of the ovule are well stained, this cell often appears merely as a deeply stained mass showing no differentiation of parts. Considering the fact that I was led not only to expect this division to take place within the pollen-grain but to search for it some weeks earlier than it actually occurs in the species of pines studied, together with the difficulties of staining, it is not surprising that seven hundred slides of serial sections were made, which means that more than two thousand pollen-tubes were studied, before any definite clue was obtained as to the 54 MARGARET C. FERGUSON true sequence of events in the development of the pollen-tube. When once the mitotic figure was observed in the ^ollen-ttibcy scarcely more than a week before fertilization, and the fact noted that special staining was necessary in order to study this mitosis satisfactorily, further research was prosecuted with comparative ease. I find no authority in Dixon's paper for the statement recently made by Coulter and Chamberlain ('oi) which reads as follows: "The liberation and descent of the body cell into the tube," etc., " has recentl}?^ been described in detail by Dixon." What Dixon ('94) does affirm is this : " Verj' shortly after this it is found that the body-cell has broken free from the stalk-cell and has divided into two cells, which are almost equal in size. These cells are the male sexual cells. During this process the wall of the stalk-cell is ruptured and its nucleus follows the two cells resulting from the division of the body-cell which move into the pollen-tube." And throughout Dixon's paper there is no sentence that could be interpreted as implying that the body-cell ever passes into the pollen-tube before dividing to form the male sexual cells. After the generative cell has passed into the pollen-tube but while it is still in the upper dead portion of the nucellus, it gives rise to the sperm-nuclei by a division which presents some new and interesting features, although it resembles to a greater or less degree certain mitoses described by various cytologists ^ during the past few years. When the generative nucleus has again come to lie in the extreme upper portion of its cell, certain changes in the cyto- plasm indicate that division is being initiated. At some little distance below the nucleus the cytoplasm shows a finely granu- lar structure which is not at this stage dense nor deeply stain- ing. From this region irregular granular threads arise which extend outward tow^ards the periphery of the cell, those extend- ' Of the long list that might be mentioned I have noted only the following : Rosen ('95) in the root-tip of hyacinth; Ostcrhout ('97) in Equisetnm ; Swingle ('97) in Sp/iacelariacece ; Schaffner ('98) in root-tip of Allium Ccpa ; Mottier ('98) in the embryo-sac of Lilititn ; Fulmer ('98) in pine seedlings: Ilof ('98) in Ephedra and other plants; Nawaschin ('99') in Plasmodiophora ; Nemec ('98 and '99) in various plants; Strasburger ('00) in Vicia Faba ; Mottier ('00) in Dictyota; and Murrill ('00) in Tsuga. Of animal cytologists I mention but one, Hertwig, R. ('98) in Actinospkccrium. LIFE HISTORY OF PINUS 55 ing in the direction of the nucleus forming a hollow cone over its lower portion (fig. 94, plate VIII). Gradually the granular area increases in density and in staining capacity, at the same time drawing nearer to the nucleus which is separated from it by a hyaline court. Into this court delicate granular threads pass (fig. 95, plate IX). When these threads reach the nuclear mem- brane, the nucleus is forced so closely against the peripheral layer of cytoplasm that its wall is frequently indented on the upper side, while the condensation from which the so-called kinoplasmic threads arise withdraws, or is forced by the growth of the threads, further from the nucleus. A great number of delicate anastomosing threads now extend, in the form of a solid cone, from a point within the granular condensation up towards and against the nucleus. The outer threads of the cone pass over the lower portion of the nucleus and appear in sections of the cell as closely packed against either side of the nucleus. At the same time the entire cytoplasmic reticulum has assumed a more or less radial arrangement about the condensed area in which the spindle-fibers arose and from which some of the more delicate threads extend into the surrounding cytoplasm (fig. 96). Coordinately with these changes in the cytoplasm, the chro- matin of the nuclear net collects in spherical or irregular masses on the reticulum, and sooner or later gives rise to a broad spi- reme, along which the chromatic disks are distributed at regu- lar intervals (figs. 94-98). After the segregation of the chro- matin, there remains a delicate achromatic reticulum distributed throughout the nucleus. This reticulum is also granular like the chromatic network, but whether or not these granules rep- resent the oxychromatin-granules of Heidenhain ('93 and '94) I am unable to say. Webber ('01) has recently described and figured a similar achromatic network in the generative cell in Zamia. Whether the formation of the spireme precedes or fol- lows the penetration into the nuclear cavity of the achromatic threads seems to depend upon the length to which these threads attain. They may become very long when their entrance into the nucleus is delayed ; but more frequently a portion of the nuclear membrane gives way, and some of the achromatic 56 MARGARET C. FERGUSON fibers pass into the nuclear cavity before the spireme is estab- lished (fig. 100). Rarely, the nuclear membrane appears pushed in irregularly along its entire lower margin, as indicated in figs. 96 and 98 ; as a rule, however, there seems to be one deep, sharp indentation along one side of which the nuclear wall first gives way (figs. 99 and 100). With the initial steps in the disappearance of the nuclear membrane the nucleolus is either not apparent or, if still demonstrable, it stains but feebly. When the membrane disappears along the entire lower portion of the nucleus, the kinoplasmic threads press so closely against it that it can not be definitely demonstrated whether it passes into the cytoplasmic and the nuclear reticulum or becomes fib- rous and contributes to the formation of the achromatic threads (figs. loi and 102). The threads which have been packed so closely against the wall of the nucleus now press into the nuclear cavity and mingle with those which have entered from below. And the dense, granular, cytoplasmic area from which the threads diverge is gradually dissipated (fig. 103). With the disappearance of the wall along the lower part of the nucleus, the achromatic nuclear network seems to undergo a partial rearrangement. A portion of it is resolved into granu- lar threads of more or less regularity which, in general, assume a position parallel to the threads entering the nuclear cavity ; some of them become attached directly to the ends of these fibers, lose their granular appearance and doubtless contribute to the growth of the elongating spindle-threads. As the spindle-fibers proceed in their development across the nucleus the chromatic spireme collects in the region of the future equatorial plate, and becomes more or less massed together. At the same time it assumes an homogeneous aspect and gives rise by segmentation to the chromosomes (figs. 101-104). Some of the ingrowing spindle-threads may extend across the nucleus to the nuclear membrane, which is still present on the upper side of the nucleus, but by far the greater number unite some distance below this membrane to form several poles, thus giving rise to a diarch spindle which, like the karyokinetic figures occurring during the development of the pollen-grain is multi- polar at its upper extremity and unipolar, or nearly so, at its LIFE HISTORY OF PINUS 57 lower end. Gradually the poles of the upper portion draw together, while the spindle is somewhat shortened Dy the lower extremity of the threads being again resolved into granules. Finally a true bipolar diarch spindle is formed with the V-shaped chromosomes oriented at the equatorial plate. Each pole termi- nates in a slight granular condensation. The upper pole has never been observed to reach the nuclear membrane, but fre- quently coarse granular threads extend from the pole to the membrane of the nucleus, and apparently act as supports for the upper pole (fig. 105, plate X). These are evidently formed by a rearrangement of the linin reticulum. The nuclear mem- brane persists along the upper side of the nucleus until the late telophase of the division (figs. 101-103, plate IX, and 104-107, plate X). As the chromosomes pass to the poles the central spindle elongates, so that the daughter-nuclei are separated, as a rule, by a greater distance than the length of the original spindle. While this is characteristic of cell-division in general, it is occa- sionally much exaggerated here, the daughter-nuclei being apparently forced apart with considerable energy. The nucleus which occupies the position nearest to the micropylar end of the ovule often shows a deep indentation along its upper surface as if a resistance had been met with in the peripheral layer of cytoplasm (figs. 11 1, plate X, and 113, plate XI). Not infre- quently the upper nucleus is found almost entirely separated from the cytoplasm (fig. 112). This, however, maybe due to mechanical rupture during sectioning and staining. No cell-wall is ever formed, and in only one instance was a condensation of the spindle-threads in the region of the cell-plate observed (fig. no). The spindle may contract at or near its center during its dissolution, thus presenting the appearance of an hour-glass, or it may give rise to such a condition as that shown in fig. 113. These appearances, with various modifications, are not uncommon in this mitosis in Pinus. Hertwig ('98) describes and figures a very similar lengthening of the spindle-fibers in ActinosfhcBrium. He also finds that the elongating spindle finally bends along its median line so that the daughter-nuclei come to lie near together in very much the same way as that Proc. Wash. Acad. Sci., July, 1904. 58 MARGARET C. FERGUSON shown in fig. 113. I am unable to trace definitely the origin of this figure, but it is not improbable that it is caused by a con- traction of the cytoplasm resulting from the cessation of the force which effected the separation of the daughter-nuclei ; or it may be produced by the resistance which the peripheral layer of cytoplasm, along the outer surface of the upper nucleus, offers to the growing fibers, thereby forcing them back upon themselves as shown in the figure. When all traces of the spindle have disappeared, the two sperm-nuclei are surrounded by a common mass of cytoplasm, and there is never throughout the later history of this cell the least suggestion of a dividing wall. The mitosis just described seems to be unique as regards the origin and development of the achromatic spindle. Hertwig's ('98) fig. 3, plate V, illustrating an early stage in the division to form the first polar body in AclinosphcBriti7Ji, bears a striking resemblance to the prophase of this mitosis as illustrated in fig. 95, plate IX, of this paper ; but the origin of the figure shown by Hertwig, and the later history of the division are very dissimilar to that of the karyokinesis under consideration. The most exaggerated instances of asymmetry in spindle-formation which I have found recorded as occuring in plants is that described and figured by Nemec ('99^) in Solamim tuberosum, and more recently by Murrill ('00) in the division of the central cell in Tsuga. In both these instances the nucleus lies at one side of the cell, and the spindle-fibers are very much more prominent on the free side of the nucleus than on the side adjacent to the cell-wall. In another paper Nemec ('99^) shows by experimenta- tion that the form of the figure which gives rise to the extra- nuclear spindle depends upon external forces or conditions. In obedience to the law established by Haberlandt ('87) we should expect to find the generative nucleus in that part of its cell which is nearest the growing point of the pollen-tube, rather than at the end more remote from it, and it may be that its passage from the lower to the upper side of the cell is due to the fact that the forces, instrumental in effecting the division, first become active at a point below the nucleus, and exert a repelling action on it. But I have at present no adequate explanation or theory to offer LIFE HISTORY OF PINUS 59 regarding the position of this nucleus at the time of its division. Whether it is due to the origin of the karyokinetic figure, or whether the unusual method of division is attributable to the very eccentric position of the nucleus, I have not been able to determine. It is evident, hov^ever, that the position of the cfenerative nucleus at the time of its division is such that the spindle if extranuclear in origin must of necessity be unipolar, since there is no cytoplasm, or almost none, above the nucleus from which fibers could arise. The blending of the linin reticulum with the cytoplasmic network after the disappearance of the lower portion of the nuclear membrane, and the relation of certain portions of the achromatic nuclear reticulum to the ingrowing fibers are such as to suggest an intimate relation between these structures. That the spindle-fibers which originate in the cytoplasm and apparently grow by a differentiation of its network are later fed by the linin of the achromatic nuclear reticulum, there seems little room for doubt. In fact, all the phenomena connected with this division indicate that we are dealing, not with per- sistent cell-constituents, but with different manifestations of one and the same thing. In a word, we find no evidence here of the presence in the cell of a definite kinoplasmic substance. I am aware that these observations are directly opposed to the views of the students of the Bonn laboratory, and many others of the highest authority ; but the relations of nucleus, spindle, and cytoplasm, not only in this division but in those to be described in connection with fertilization, are such, it seems to me, as to render no other conclusion in the case of these divis- ions in Piniis possible. In 1895 Farmer arrived at a similar decision regarding the origin of the spindle in spore-formation in the Uepaiicc^, and Farmer and Williams ('98) in a study of Fitctis " do not regard the kinoplasm as a persistent proto- plasmic structure, but as forming the visible expression of a certain phase of protoplasmic activity." Hertwig ('98) expresses himself as opposed to the view of a special spindle-forming sub- stance in the protoplasm, while Wilson ('99 and '00) states that the astral rays " grow by a progressive differentiation out of the general cytoplasmic meshwork," and he finds in the echino- 6o MARGARET C. FERGUSON derm's egg " no ground for a specific kinoplasm." The term, however, is a convenient one and may be employed consistently, as suggested by JNIottier ('oo), by those who do not find in kino- plasm a morphological constituent of the cell, as descriptive of that portion or manifestation of the protoplasm which is active in spindle-formation. Nothing has been said regarding the nature of the granular, cytoplasmic condensation from which the achromatic spindle takes its origin. It never has a definite boundary, though it is often very clearly differentiated by its dense granular appear- ance and its strong affinity for stains ; but at certain stages in the division it may be inconspicuous or fail entirely of demonstration. Such a vast amount of literature has accumulated during the past decade regarding the nature and existence of the centrosome and the centrosphere that one feels inclnied to avoid the subject alto- gether. Yet the question may very properly be asked : Is this condensation which forms the center of a system of radiating fibers a centrosphere? It certainly is as clearly an attraction- sphere as some bodies which have been described as such ; but if we accept Wilson's ('oo) definition of the centrosphere, the body under consideration cannot be so denominated, as no cen- trosome has been observed at its center. More deeply staining granules may sometimes be present within the condensation, but these are not considered of any special significance as such granules may be found anvwhere in the cytoplasm. Karsten ('93) describes the nucleoli in Psilotum as passing out of the nucleus and assuming the role of centrosomes, and Strasburger ('00) considers that the nucleoli not only contribute material for the formation of kinoplasmic threads, but that they also make active the spindle-forming substance in the cytoplasm — in other words, they act as the kinetic centers of the cell. There seems to be no evidence that such is the case here, for the nucleoli, after the condensation has arisen and the spindle- threads have attained considerable length, are morphologically the same as they were before the inception of the spindle. Neniec ('99') remarks that in the higher plants, where the cen- trosome is not demonstrablv present, the entire nucleus may exercise llie function of the centrosome. The idea of a diffused LIFE HISTORY OF PINUS 6l centrosome in the cells of the higher plants was suggested by Guignard in 1897 and was again hinted at by Le Dantec in 1899. If we may accept Guignard's suggestion, then the kinetic center of the cell in the higher plants is no longer indicated by the presence of a definite organ, the centrosome, but the power of this organ has become dissipated throughout the entire cell. When that phase of cell-activity which has to do with spindle- formation comes into play, the points at which it is centered would naturally be indicated by a greater accumulation of the microsomes, and thus an aster of more or less definiteness would be formed, as when the individualized centrosome is present. In the division of the generative nucleus in Pmits, the position of the nucleus is such that the energy active in spindle-forma- tion must perforce, if external to the nucleus, be centered at some point below it. Such a centering of the activity would naturally result in an attraction-sphere of unusual prominence ; and there would be no occasion for its division since there is not sufficient space above the nucleus for the organization of kino- plasmic threads. When these studies were undertaken, it was thought that it would be interesting to determine whether any suggestions or remnants of a cilia-forming body (called blepharoplast by Webber in Zamid) still persist in the Conifers. Somewhat later, after the present research was begun, MacMillan ('98) pointed out the desirability of such a study both in Conifer<^ and Gnetales. I have seen no indication of a structure which might be regarded as a reduced blepharoplast, or as suggestive of a cilia-forming body of any sort in connection with the formation of the sperm- nuclei in Pinus. Inasmuch as spermatozoids do not exist here, such an organ, if present, must be functionless. But the cyto- plasmic radiations which accompany the division of the genera- tive nucleus in its early stages seem to differ in degree only from those found by Webber ('97) in the generative cell of Zamia. If we compare figs. 3 and 5 of Webber's paper with figs. 96 and 97, plate IX, of this paper, the question may be raised whether in this cytoplasmic figure we may not have still persisting in the cell the last vestiges of such an organ as that described by Webber. 62 MARGARET C. FERGUSON The endosperm has become a solid mass of tissue at the time when the generative nucleus divides. The archegonia are still comparatively small and quite vacuolate and the central cell has not yet divided (fig. 72, plate VI). Grozvth of the Sperm-nuclei. — After the mitotic figure has entirely disappeared, the sperm-nuclei are separated by a con- siderable distance. The form assumed by the cytoplasm sur- rounding them seems to vary with the shape of the pollen-tube. Gradually the two nuclei approach each other until they come to lie in the extreme uppermost part of their cytoplasm (figs. 112, plate X, 117, 118, plate XI). There is now considerable differ- ence in their size. This inequality in size could be detected as far back as the formation of the daughter-nuclei (figs. 109, no, plate X). Belajeff ('91) was the first to figure and describe bi- nucleated sperm-cells in the Gymnosperms. Coulter and Cham- berlain ('01), page 94, cite Belajeff as having observed an unequal division of the generative cell in Taxtis^ the larger male cell func- tioning, the smaller one remaining in the tube. But if I translate the German correctly, what Belajeff says is that the nucleus of the generative cell divides forming two nuclei which are about one-half as large as the nucleus from which they were derived ; one nucleus becomes larger and occupies a central position in the plasma, the other nucleus is flattened and remains at the periphery of the cell on its upper side ; the flattened nucleus was never found surrounded by its own plasma, but in the same plasma with the spherical nucleus. This is exactly the condi- tion shown in Belajeff's figures, one of which is reproduced by Coulter and Chamberlain. Jager ('99), however, has shown two dissimilar sperm-cells in Taxus, the larger one in advance, but he finds that occasionally the nucleus of the smaller cell may exceed in volume that of the larger one. Jaccard ('94) found two sperm-nuclei of the same size in Ephedra both sur- rounded by the same mass of cytoplasm, and Coker ('02) has re- cently described the sperm-cell in Podocarpus as binucleated, the smaller nucleus being above the larger and " thrust almost out of the cell." No one, I believe, except the writer (1901''""'^), has recorded the presence of a single binucleated sperm- cell in the AbietinecB. In his earlier studies of the Gymno- LIFE HISTORY OF PINUS 63 sperms, Strasburger ('6g-^2) was unable to demonstrate, satis- factorily to himself, the character of the cells found in the pollen-tube in Pinus, and he has not recently investigated the male gametophyte in the Abietinece. Coulter ('97) described two sperm-cells which were of the same size until within the arche- gonium. Blackman ('98) stated that each sperm-nucleus was clearly seen in the pollen-tube surrounded by its own cytoplasm, but he did not figure them.^ Chamberlain ('99) figured the sperm-nuclei, in Pinus Lartcto, of equal size in the pollen-tube, and showed them lying together in the cytoplasm of the tube. Not having seen these cells within the archegonium before the conjugation of the sexual nuclei, he accepted Coulter's state- ment for the growth of one of them after their entrance into the egg. According to Coulter ('00) the " male cells in pines " are alike in size. The same figures are reproduced by Coulter and Chamberlain ('01). As stated by the writer in 1901, two sperm-cells have not been observed in any of the pines which I have studied ; but the sperm -nuclei, which are of unequal size from a very early date, remain, while in the pollen-tube, surrounded by a common cyto- plasmic body (figs. 109-112, plate X; 113-118, plate XI, and 119-120, plate XII). As Strasburger ('92) observed, the larger nucleus is always ahead, that is, on the side nearest the apex of the pollen-tube. The smaller nucleus remains close against the upper boundary of the cytoplasm, and suggests the condition in Cycas (Ikeno '98) and Ginkgo (Hirase '98), where the stalk- nucleus is forced entirely out of the cytoplasm surrounding the generative nucleus. In the case of the smaller sperm-nucleus in Pimis, the action is not carried to so great an extent. Webber ('01) has recently shown that such an interpretation as that re- corded above for Cycas and Ginkgo is not true as regards the stalk-nucleus in Zamia. One very interesting preparation which I have obtained shows the smaller sperm-nucleus in advance of the larger (fig. 114). Here it will be seen that the entire order of arrangement has been changed, the stalk-cell and the tube- nucleus being above the sperm-cell. But this abnormal arrange- ment is onl}'- apparent, for it was found that the ^^^ which had * See note at close of Appendix. 64 MARGARET C. FERGUSON been approached by this pollen-tube had already been fertilized, and the pollen-tube had turned aside and was passing up over the top of the endosperm, as if seeking for another egg. The position of the various elements of the pollen-tube is therefore normal, the larger sperm-nucleus being in reality in advance of the smaller. This suggests that, w^hen a pollen-tube has con- jugated with the egg, a substance may be secreted which repels other pullen-tubes, as has been described in case of spermato- zoids in the Bryophytes and Pteridophytes. The formation of the sperm-nuclei shows most beautifully the manner of the development of the nuclear reticulum. The chromosomes unite end to end, giving rise to a homogeneous, coiled band, before the nuclear membrane is formed. When the nuclear-wall has been differentiated, the coil expands about the periphery of the nucleus, while the band broadens, at the same time becoming irregularly jagged along its margins. These irregularities increase in length until finally those from adjacent threads meet and fuse, thus giving rise to the reticulum (figs. 107-110, plate X). When the sperm-nuclei have nearly or quite come into contact they have as a rule reached their ma- ture size. More than a year has now elapsed since pollination. Elongation of the Pollen-tube. — Up to this time the pollen- tube has elongated very slowly, having penetrated as yet little, if any, beyond the nucellar tissue of the previous year's growth. In this upper portion of the nucellar cap the tube may become very broad, or it may branch freely (figs. 71, 72, plate VI, and 83, 87, plate VIII). When the sperm-nuclei have attained their full size, the downward growth of the tube is exceedingly rapid, travelling in from eight to ten days more than twice the distance traversed during the entire preceding year. The path pursued during this rapid growth is comparatively straight and the tube is unbranched (fig. 73, plate VII). In Pinus Strobus^ P. rigida and P. mistriaca about ten days intervene between the division of the generative nucleus and fertilization ; in Pinus ino7itana unci^iata, the two processes are separated by an even shorter space of time. The sperm-nuclei which at first present a very beautiful, rather delicate reticulum (figs. 112, plate X, 117, plate XI), become LIFE HISTORY OF PINUS 65 more dense as the pollen-tube advances through the nucellus. Strasburger ('92) describes them as coarsely granular ; but, with a high power, the presence of a reticulum which is sometimes coarse and interrupted can invariably be made out in well pre- pared material. By the time that these nuclei have reached in their downward course the central portion of the nucellar cap they have usually become very dense in structure (figs. 115 and 116), and frequently stain intensely, though they may show at this time a weak reaction to dyes. The reticula of the two nuclei may present the same appearance, or they may differ as in the figures referred to above. The nucleolus, if it be present at this time, is usually obscured by the dense network. Arnold! ('00) described the sperm-nuclei in Cephaloiaxus as being grad- ually filled with metaplasm. I find no evidence of such a proc- ess in the development of these nuclei in Pinus. Archoplasmic areas similar to those figured by Chamberlain ('99) have been observed in connection with the sperm-nuclei, but as such granular accumulations may occur at any point in the cytoplasm of the sperm-cells no importance is attached to them. When the pollen-tube reaches the egg-, its apex is abundantly supplied with cytoplasm, in the upper part of which the tube- nucleus lies. The sperm-cell is just above with the stalk-cell still in contact with the lower portion of its cytoplasm (fig. 120, plate XII). Still higher up the tube may contain many starch- grains. There is never any doubt at this time as to the identity of the stalk-cell and the tube-nucleus in the material which I have studied. Yet Dixon ('94) states that they cannot be distin- guished, and Coulter ('97) describes them as having lost their original outline. As many as six pollen-tubes have been found making their way through the same nucellus, but, as a rule, not more than three pollen-tubes renew their growth during the second season, and frequently only two penetrate to the endosperm. The effect of the pollen-tubes upon the upper part of the nucellar tissue is very marked. The cells in the immediate vicinity of the branched pollen-tubes early lose their protoplasmic contents and their walls become crushed and broken. Those cells more 66 MARGARET C. FERGUSON remote from the tubes do not suffer so severely, and retain their protoplasm for a much longer time. Finally all the cells representing the first year's growth of the nucellar tip loose their content to a greater or less degree, and their cell-walls become thickened and dead. During the rapid growth of the pollen-tubes through that portion of the nucellar cap which develops the second season, the effect of the tubes on the sur- rounding tissue is less marked, though here, too, the cells with which they come into contact are crushed and destroyed (fig. 73, plate VII). I have made no physiological investigations regarding the action of these tubes on the tissue of the nucellus, but, judging from the disappearance of the starch in the cells just in advance of the tubes and the gradual disintegration of those cells, it seems very probable that the destruction of tissue is not due to mechanical reasons alone, but to the action of some ferment or digestive substance as well. Various views have been expressed concerning the action of the pollen-tube and the directive agent in its growth by Molisch ('93), Miyoshi ('94), Lidforss ('99) and others, but we are still far from a clear under- standing as to the controlling factor in the movement. The pollen-tube cannot be guided to the egg in Pimis by any peculiar attraction existing between the sexual cells, for it grows with normal rapidity when no sperm-cells are formed, and also when the archegonia are in a state of disintegration. SUMMARY. Upon the germination of the microspore, three divisions fol- low in rapid succession giving rise to the pollen-grain. At the close of the prophase of each division the karyokinetic figure is pointed at its lower extremity and very broad at the extremity in contact with the dorsal side of the young pollen-grain. The inner, incomplete, thick wall formed in the development of the microspore persists as a part of the mature pollen-grain. It probably serves as a strengthening layer, particularly at those points at which the wall has been weakened b}"^ the expansion of the exospore. When the telophase of the second division is reached the first prothallial cell has become flattened against the convex side of the spore-wall, its cytoplasm has been withdrawn, LIFE HISTORY OF PINUS 6'J andthe nucleus has lost all signs of its former structure remain- ing as a much flattened, deeply staining mass. At the close of the third division, the second prothallial cell has suffered a simi- lar fate. Both prothallial cells are furnished with cellulose-walls. In the mature pollen-grain the prothallial cells are usually represented by two broken, dark lines along the dorsal side of the pollen-grain, but all vestiges of the first cell may have dis- appeared. The antheridial cell projects from the convex side of the spore at its middle point, and the tube-nucleus is always directly below but in contact with the antheridial cell. Starch is found in the pollen-grain at maturity and during its develop- ment. Pollination takes place between 42° and 43° north latitude during the latter part of May or the first ten days in June. At this time the macrospore-mother-cell is distinctly visible in the center of the ovule, but slightly nearer its basal end. In the young ovule the free portion of the integument, above the tip of the nucellus, consists in cross-section of three layers of cells. After pollination the arms of the integument become erect, thus bringing the pollen-grains into the wide micropylar canal. Then the inner layer of cells just above the pollen- grains elongates rapidly, extending inwards and meeting at the center. The pollen-grains having thus been made secure, the elongated cells become divided into many small cells. It is felt that the pit in the apex of the ovule in Piniis has been ex- aggerated. There is rarely more than a slight concavity before pollination. Through the action of the pollen-tubes it may be somewhat deepened, but in normal conditions it does not become " cup-like." Two days after pollination, in Pinus rigida, the pollen-tubes have been emitted. In the other species germination has been shown to take place in less than a week after pollination, but more exact data have not been obtained for these species. As soon as the pollen-grain has germinated, the tube-nucleus severs its connection with the antheridial cell and moves into the elon- gating tube. The division of the antheridial cell takes place in Pinus Strobus during the first week in August. It sometimes divides 68 MARGARET C. FERGUSON during the summer and fall in P. aush'iaca^ but, as a rule, the division takes place in this species very early in March. This mitosis has been observed in P. resinosa during the second week of April, and in P. rigida from the middle of April to the middle of May. It is evident that this cell does not always divide at a definite and fixed time, but that in a given species the time during which it may divide extends over a considerable period. During the first season the pollen-tube grows very slowly, and it may be broad and irregular in outline or it may branch freely. Shortly before fertilization the generative cell, followed by the stalk-cell, moves into the pollen-tube. The stalk-cell soon passes the generative cell and takes up a position near the tube- nucleus. These changes and those immediately following are frequently much obscured by the presence in the pollen-tube of large quantities of starch. When the macrosporangium enters upon the winter's rest, the pollen-tubes have penetrated nearly to the line at which the in- tegument becomes free from the nucellus and the tube-nucleus maintains its position in the apex of the pollen-tube. The generative cell is never limited by a well-defined cell- wall, and consists at the time of its division of an irregular pro- toplasmic body in the upper part of which the nucleus lies. In the division of the generative nucleus the spindle is extra- uuclear arid unipolar in origin, a unique and heretofore unob- served method of division. The formation of the spindle indicates that the cytoplasmic network and the nuclear reticulum have essentially the same structure, and the spindle-fibers are apparently formed by a transformation of both. The nuclear membrane persists along the upper part of the nucleus until the earl}^ stages in the forma- tion of the daughter-nuclei. This division takes place a little more than a year after pollination and from a week to ten days before fertilization, nearly thirteen months elapsing between pol- lination and fertilization. Two sperm-cells are never formed, but the sperm-nuclei remain surrounded by a common mass of cytoplasm. An in- LIFE HISTORY OF PINUS 69 equality in the size of these nuclei is very early apparent, and becomes more pronounced as they reach maturity. The sperm- nuclei soon come to lie together in the upper part of their cyto- plasm and quickly attain their full size, the larger one being invariably in advance. The nuclear reticulum, at first delicate, soon becomes very dense, but there is no evidence of the pres- ence in these nuclei of a special metaplasmic substance. During the division of the generative nucleus the ovule in- creases much in size, and the nucellar cap becomes several times deeper than during the first season, thus carrying the upper portion of the nucellus with its pollen-tubes far above the endosperm. At the time when the sperm-nuclei come into contact, or nearly so, the pollen-tube has penetrated little, if at all, beyond the nucellar tissue of the first year's growth. Now, however, it again begins to elongate, and its downward course through the new nucellar tissue is extremely rapid. The destruction of the nucellar tissue through which the pollen-tubes travel, ap- parently results not only from mechanical disturbances, but from the entire dissolution of some of the cells through the action of a ferment. When just above the egg, the apex of the pollen-tube is filled with cytoplasm. The tube-nucleus lies in the upper part of the cytoplasm, and near it is seen the stalk-cell still in contact with the lower portion of the cytoplasm which surrounds the sperm- nuclei. The existence of the diffused centrosome is suggested in con- nection with the division of the generative nucleus, and there is a possibility that, in the prominent cytoplasmic figure from which the spindle takes its origin, we may have represented, in its vestigial state, the cilia-forming body found In the lower Gymnosperms. 70 MARGARET C. FERGUSON CHAPTER III. Macrosporogenesis. the female cone. The Macrosforangiuni. — During this investigation I have made no attempt to study the early development of the ovule except to note definitely the date of its origin. The pistillate strobili cannot be detected in Pimis Strobiis with the most careful examination until the last of April or the first of May. In the other species studied they are about one and one-half milli- meters long at the middle of March, and it is possible that in these species they were organized in the autumn, but I have not been able to find any evidence that such is the case. I have recently, November 25, 1902, attempted to discover the young cones of Pinus rigida and P. atcstri'aca, but, as formerly, the search was futile. I was led to look again for these strobili in the autumn by the recent statement of Coulter and Chamberlain ('01). On page 79 of their book on the morphology of the Gymnosperms, I find this sentence, based on a study of Pinus Lai'icio : ** In June the archegonia are ready for fertilization, which occurs about the first of July, at least twenty-one months after the first organization of the ovule." This by a very simple mathematical calculation places the " organization of the ovules " on October i. I have not only been unable to detect the pistillate cones before the approach of winter, but in tlie tin}^ cones of Piniis rigida^ P. aitslriaca and P. montana uncinata, fixed on March 14 there is not the least suggestion of ovules, the entire cone consisting in each case of a broad axis on the margin of which are slight elevations or papillaj — the beginnings of the bracts which subtend the ovuliferous scales (fig. 121, plate XII). The first indications of the ovules are found in these species about the last of April or the first of May. In material of Pinus Strobus fixed on May 31, 1898, the position of the ovule can be detected only by a slight bulge on the inner sur- face of the ovuliferous scale, the integument not 3'et having been differentiated. One week later, June 6, the ovule is LIFE HISTORY OF PINUS 7 1 found fully organized and nearly ready for the reception of the pollen-grains (ligs. 122, and 123). The evidence is conclusive that the ovules are not organized in the species of pines studied by the writer until about three weeks or less before pollination, and seven months later than in Pimis Laricio as recorded by Coulter and Chamberlain. This is the more surpris- ing when we consider that P. aiistriaca is at least a variety of P, Laricio, and, according to some authorities, it is a synonym for that species. It is not my purpose to enter into a discussion of the origin and cellular development of the female cone, nor yet of the homologies of its parts. These points have been fully investi- gated by Celakovsky, who has frequently published papers on this subject from 1879 to the present time, and the many theories advanced by different writers regarding these structures have recently been brought together and reviewed by Worsdell ('00). FORMATION OF THE AXIAL ROW. The Macrosfore-iuother-cell. — The origin of the sporog- enous tissue from a hypodermal cell or cells was described by Strasburger for several Gymnosperms in 1879, and this idea without further confirmation has come down to the present time. While this may be true for many Gymnosperms, and possibly for Pimis, I find no evidence, direct or indirect, that the macro- spore-mother-cell is derived from a hypodermal cell in the pines investigated. When the mother-cell is sufficiently differentiated to be distinguishable from the other cells of the surrounding tissue, it is found to lie deep within the nucellus ; and there are no rows or axial strands of cells lying above it to suggest its derivation from a hypodermal cell. On May 8, 1902, the ovules of Pinus rigida were sufficiently developed to show clearly the separation into nucellus and integument, and a like condition was found to exist in P. Sirobits on June 6, 1898. In both instances, so far as one is capable of determining, every cell of the nu- cellus is exactly like every other cell (fig. 123), and the same condition obtains in the other species at this time. One week later, as illustrated for Pinus rigida, the macrospore- mother-cell can first be distinguished, and the so-called spongy 72 MARGARET C. FERGUSON tissue is clearly differentiated about it (fig. 124). The mother- cell in this instance has relatively the same position in the ovule as that shown in fig. 66, plate VI, which was taken from an ovule collected twelve days later. If this cell be the direct de- scendant of a hypodermal cell, it has now become deep-seated by the addition of cells above it ; but there is nothing in the arrangement of the cells of the nucellus either before the appearance of the mother-cell or after it to denote such a course of development. The mother-cell is first detected by its larger size and b}- its failure to stain as deeply as do the other cells of the nucellus. In the first stages of growth the nucleus almost fills the cell (fig. 125), and its weakened capacity for staining is doubtless due to its rapid growth without a proportional increase in the amount of nuclear substance. The nucleus contains in this young stage a delicate reticulum with a varying number of larger and smaller net-knots, and from two to four small nucleoli, not differing materially, except in size and staining power, from the nuclei of the adjacent tissue. This cell in- creases considerably in size before its division so that it becomes very conspicuous in the nucellus, its reticulum taking the chro- matin-stains with greater avidity than at an earlier period. The season of growth for the macrospore-mother-cell may extend over about three weeks. The early stage shown in figs. 124 and 125 represent its size on May 15, 1902, and the spireme stage illustrated in fig. 126 indicates the condition of this cell on June 5 of the same year. First Division of the Macrosporc-moiher-ccU. — After the mother-cell has attained its full size, the reticulum of the resting nucleus gradually becomes more open, the chromatic granules become more prominent and there arises a beauti- ful, regularly moniliformed, more or less interrupted skein, but a true spireme is not formed until after synapsis (lig. 126). This somewhat branched thread is very delicate, the chro- matic discs are uniform in size and distributed upon the linin with great regularity. It is probable that these apparently homogeneous discs, which have doubtless been derived from the fusion of the smaller chromatic granules, would, under LIFE HISTORY OF PINUS 73 greater magnification, be resolved into slightl}' irregular and roughened bodies, as in the prophase of the heterotypical mitosis in the microspore-mother-cells, but with the powers of the micro- scope at my command, I have no evidence that such is the case. The phenomenon of synapsis is as marked here as in the primary mitosis of the microspore-mother-cell, but the contracted mass is less dense, probably because of the smaller size of the nucleus and the consequent diminution in nuclear substance (fig. 127, plate XIII). With the recovery from synapsis the linin thread is seen to have increased in thickness, and the chromatin- granules are irregularly distributed upon the continuous spireme, which gradually comes to fill the entire nuclear cavity with its open uninterrupted coils (figs. 128 and 129). The chro- matic substance again collects into definite areas of varying dimensions, which are united by clear portions of the linin- band, and the longitudinal splitting now becomes apparent. Condensation and segmentation follow, and the distinct chro- mosomes, in the reduced number, become evident (figs. 130, 132 and 133). The forms of the chromosomes are similar to those already described in connection with the division of the microspore-mother-cell (figs. 132-136). Because of the com- paratively small size of these nuclei, the steps by which the irregularly shaped chromosomes are derived could not be traced with the same degree of confidence as in the microspore-mother- cell ; but the entire phenomenon is such as to indicate very con- clusively that the process is practically the same in both. The spindle, at first a multipolar diarch, early becomes bi- polar and during metakinesis it is very sharply so. The poles do not reach the walls of the cell, but a few threads sometimes radiate from them and extend to the ectoplasm. There may be a slight granular .condensation in the neighborhood of the poles but it is never prominent and often does not appear at all. The chromatic segments become short and broad at the equa- torial plate, and their separation into daughter-chromosomes presents the figure characteristic of the heterotypic division. Unsplit ends of the chromosomes extend outward in the plane of the equatorial plate, thus giving rise to dark clumps of chro- matic substance along the median line (figs. 134-137). The Proc. Wash. Acad. Sci., July, 1904. 74 MARGARET C. FERGUSON passage of the one-half chromosomes to the poles has not been observed. Resting nuclei are formed during the telophase of the mitosis, and a cross wall divides the mother-cell into two compartments (fig. 138). From the foregoing it is evident that the first division which takes place in the macrospore-mother-cell is heterotypic in nature, and agrees in all essentials with the primary mitosis within the microspore-mother-cell. This is in accordance with the conclusions reached by all other investigators who have recently studied the tetrad divisions occurring within the ovules of various Phanerogams. Second Division of the Macrospore-juother-cell. — Beginning with the telophase of the first division considerable variation may occur in the subsequent steps in the formation of the axial row. A cell-plate is always formed between the daughter- nuclei though it may remain very delicate, consisting of little more than a condensation of the ectoplasm. The daughter- cells may be very similar in appearance, excepting that the lower one is usually the larger, and in such instances both nuclei enter the resting stage, presenting a clear, definite reticulum (figs. 138, and 141). More often, however, the lower cell is much larger than the upper one and the nucleus of the upper cell does not enter into the complete resting stage, but early shows signs of disintegration. The chromosomes may unite to form a spireme as usual, but development may then cease without the organization of a network, and the diffuse reaction of the nucleus to stains shows that disintegration has begun (figs. 139, 140). I have but a single preparation showing the second division of the macrospore-mother-cell, and I can therefore offer no con- clusions of any value regarding the nuclear phenomena accom- panying the mitosis. From this figure it appears that the spindle originates as a multipolar diarch as in the first division, and both nuclei in this instance are dividing at the same time. During the initiation of the spindle the chromosomes are short and thick, somewhat irregular in outline, and apparently in the forms of U's, V's and rings. The reduced number of chromo- somes occurs in both of the dividing nuclei (fig. 142). LIFE HISTORY OF PINUS 75 The state of disintegration referred to above is always con- fined to the upper of the two daughter-cells and never occurs in the lower one, except in those cases in which the whole ovule is undergoing destruction. The lower cell invariably divides again and the basal cell thus formed constitutes, in every instance observed, the functional macrospore. The lack of constancy in the division of the upper cell would naturally give rise to some axial rows of four cells and some of three, and this is exactly what we find (figs. 144, 145, plate XIV). Fig. 143 shows the second division of the lower cell just completed, and it is evi- dent from the structure and appearance of the uppermost nucleus that it would never have divided. In the axial row presented in fig. 144 some time has elapsed since the mitosis was completed, as evidenced by the increase in size of the lowest cell of the row. The upper of the two cells formed as a result of the first mitosis still remains undivided, and, moreover, it would not have divided later, judging both from its appearance and from the fact that the rapid growth of the initial cell of the female gametophyte would soon have been instrumental in effecting its obliteration. Juel ('00) finds that these cells do not divide simultaneously in Larix^ but he does not find the division completed in the lower cell before it begins in the upper one. In the single preparation showing the second division in the macrospore-mother-cell, both nuclei are dividing, and both are in the same stage of the prophase, but this does not necessarily mean that when both cells divide they always do so synchronously. This lack of uniformity in the number of cells in the axial row is not peculiar to Pinus ; it has been observed by many investigators in a large number of plants including both Gymnosperms and Angiosperms. Coulter and Chamberlain ('01) figure an axial row of four cells in Piniis Laricio, and, as above indicated, such an axial row is frequently met with in the species of pines which I have studied, but it is much more common in Piniis ausiriaca than in the other species (figs. 145, plate XIV, 142, plate XIII, and 261, plate XXIII). There is no doubt whatever, after a study of many preparations showing the axial row, that in the great majority of cases in Pinus Strobus and P. rigida the upper cell remains undivided and that the usual axial row in these species 76 MARGARET C. FERGUSON consists of three cells. The axial row represented in fig. 144, for instance, is a beautiful object, clearly and definitely differentiated from the surrounding tissue, 3^et there is not the least ground for supposing that the upper cell has ever divided. Such a figure as this represents the characteristic axial row in Pinus Strobiis and P. rtgida, while the axial row of four cells illus- trated in fig. 145 is typical for P. austriaca. This point has not been sufficiently studied in the two other species to admit of generalizations for them. The axial row, then, varies from three to four cells in the same species, but there is a tendency in some species to form three and in others to form four cells. Significance of the Tetrad DivisionWithin the Ovule. — We have observed that at a certain point in the development of the ovule in Pinus a centrally located cell becomes differentiated from those surrounding it by its greater size and the more vacuolate character of its cytoplasm. This cell after under- going a period of growth and rest gives rise to the reduced number of chromosomes by a peculiar method of division known as the heterotypical division, and this mitosis, as is characteris- tic in spore formation, is quickly followed by a second division, at least in the lower cell. The basal cell resulting from this last division passes through a season of growth extending over several weeks, as we shall shortly see, and finally, by repeated divisions, gives rise to the female gametophyte. The process of division is in all essentials exactly similar to that which takes place within the microspore-mother-cell, and results, as there, in spore-formation. Nuclear phenomena attending the early development of the female gametophyte have not been carefully investigated until comparatively recent times, but wherever studied the conclusion has been unhesitatingly drawn that in the ovule, as within the anther, a true spore-formation takes place. The essential character of a spore is, manifestly, not that it should have a certain arrangement relative to its sisters within the mother-wall, neither is the presence or absence of a wall of vital importance to its existence unless, indeed, the spore is to be disseminated. Rosenberg ('01) finds the pollen-grains to be filiform in Zostcra and arranged side by side ; Strasburger ('01) LIFE HISTORY OF PINUS 77 and Gager ('02) show that the descendants of a pollen-mother- cell in Asclcpt'as have a linear arrangement ; while Juel ('00) discovers that in the Cyperaccce three young pollen-grains or microspores abort and the fourth remains permanently within the microspore-mother-wall. Yet from the standpoint of origin alone, no one hesitates to call the young pollen-grains of these plants microspores. Juel ('00) affirms that the heterotypic divi- sion must be the criterion by which we decide whether or no we have a true tetrad-division, and he concludes that in Larix the embr3^o-sac-mother-cell is homologous with a spore or a micro- spore-mother-cell. Schniewind-Thies ('01) reaches the same conclusion for Angiosperms ; and Lloyd ('01) asserts that the division of the embryo-sac-mother-cell in the RubiacecB is a true tetrad-division, and the four resultant cells are spores. Other instances where similar conclusions have been reached might be cited, but the above is sufficient to demonstrate that the most recent studies along this line point conclusively to a normal spore-formation within the ovule, and do not confirm Campbell's ('02) statement that a true tetrad-division is usually absent in the ovule of spermatophytes. For many years botanists have been involved in a contention regarding the true nature of the embryo-sac in Phanerogams. A paper was published by Atkinson in 1901 reviewing the interpretations made by earlier writers and suggesting as a solution of the difficulty that spores, no longer being necessary in the higher plants, had dropped out of the cycle of develop- ment in these plants. That is, the female gametophyte arises in the higher plants without the intervention of spores. While the results of recent investigations do not serve to strengthen this view, the theory is a most interesting one and the paper has further served an excellent end in stimulating thought and research along this line. Mottier observed one instance in which the first division of the embryo-sac-mother-cell was homo- typic, or, if we use Strasburger's ('00) term adopted through- out this paper, typical, and the number of chromosomes was not reduced. Juel found the same to be true normally in Anten- naria alpina, a species of Antennaria in which the embryo develops parthenogenically. In both instances we have an 78 MARGARET C. FERGUSON illustration of development within the embryo-sac without the intervention of a spore, but these are apparently isolated and exceptional cases. The whole difficulty seems to me to lie in the fact that all along we have been endeavoring to make a morphological unit out of that which is primarily a physiological unit, and not necessarily a morphological one, although it may be so. It has been shown conclusively that in Larix and Pinus among the Gymnosperms a true macrospore is formed which germinates within the macrosporangium and gives rise to the female gamet- ophyte — both a morphological and a physiological unit. But as we advance to the Angiosperms there is a shortening of on- togeny in the female gametophyte, the most extreme case being represented by Lilitim. Mottier ('98) demonstrated the fact that the division of the embryo-sac-mother-cell in Lilhim is a true tetrad division and we cannot, therefore, it seems to me, escape the conclusion that the resultant four cells are spores. But once rid ourselves of the idea descended from Hofmeister, that the mother-cell of the embryo-sac is always a macrospore, and the product of its development, therefore, always a single gametophyte, and many difficulties vanish. Lloyd ('02), in his recent discussion of this subject, accepts the heterotypical divi- sion as the criterion for spore formation, and then explains the condition in Lilium, where the first four cells of the embrj'o- sac are spores, by " regarding the gametophyte as an individ- ual by coalescence.'" It appears to me not only more simple but more plausible to consider that we have here four gameto- phytes each reduced to two cells. The embryo-sac is still here as elsewhere (with the exception of parthenogenic plants), a physiological unit whose function is to give rise to a new plant through the sexual process, but it is morphologically a complex made up of several individuals. Whether all eight cells thus formed are considered as potential eggs is immaterial, practi- cally, but one retains the power to respond to the sperm-cell, though the others have been shown to be capable of fertiliza- tion in some instances. Ordinarly, however, they remain ster- ile and have come to have a vegetative or nutritive function only. All work together for one end and in that sense may LIFE HISTORY OF PINUS 79 make " an individual by coalescence," that is, they are physi- ologically one. This is not the place to enter into a detailed discussion of the homologies of the embryo-sac, but I believe that the suggestion herein made will form an interesting working basis, and it may bring us nearer to a true conception of these structures than we have yet attained. But whatever our opinion regarding the ele- ments within the embryo-sac, it is clear that we cannot longer use the terms macrospore and embryo-sac interchangeably as so many writers have done. We now know that a tetrad division may occur within the ovule and it has been shown that the embryo-sac may result from the germination of a single macro- spore, that it may be formed directly from the macrospore- mother-cell, or that it may have its origin in one of the daughter- cells formed as the result of the heterotypical division. In any case would it not be far less confusing if we should designate the multicellular bodies, developed within the macrosporangium and the microsporangium of the higher plants, as embryo-sac and pollen-grain, or female and male gametophyte, respectively, and should retain the terms macrospore and microspore for the true spores in their one-celled stage? LATER HISTORY OF THE AXIAL ROW. The Fate of the Uffer Cells. — Whether the number of cells in the axial row of Pinus be three or four the female gameto- phyte is always the product of the lowest cell. Very shortly after the second division is completed, the upper cells of the axial row give evidence of disintegration, while the basal cell increases much in size, its nucleus becoming very large. The nuclei of the four spores in Larix are very similar, Juel ('oo), fig. i8, but in Pinus the basal cell is markedly different from the others at a very early date (figs. 144, 145, plate XIV). The upper cells of the axial row gradually disintegrate, and are crowded to one side by the growth of the macrospore, remain- ing for a time as deeply staining, amorphous masses which finally disappear altogether (figs. 69, plate VI and 147, 148, plate XIV). Instances in which one of the upper cells of the axial row in Angiosperms becomes the functional macrospore 8o MARGARET C. FERGUSON are not rare. Campbell ('oo) has recorded such a condition in the AracccB, Lloyd ('oi) in certain Rubiacece, and Karsten ('02) in the JuglandacecR. But, so far as investigated, the sequence of events following the establishment of the axial row in the AbietinecB results in the obliteration of all but the lowest cell. I have avoided using the term " potential macrospore " in con- nection with the upper cells of the axial row, because the upper of the two cells first formed does not always divide and in such instances it cannot properly be designated as a spore since development ceased before spore formation was completed. Growth of the Macrospore. — Starch is sometimes found within the cells of the axial row, though never in such abundance as in the cells of the adjacent tissue (fig. 143). It may become very abundant within the macrospore during its period of growth, and is sometimes found pressed so closely against the nucleus as to actually produce indentations in its membrane (fig. 146). The reticulum of the nucleus of the functional spore is very scanty during its growth period, but later it presents the appear- ance of an ordinary resting nucleus. The cytoplasm, never abundant, forms at an early date a loose, granular network. Later the nucleus is connected with the ectoplasm by delicate strands which are gradually withdrawn into the peripheral cyto- plasm, until there is thus formed in the one-celled stage a definite layer of cytoplasm lining the wall of the macrospore, and inclosing a large central vacuole. The nucleus moves to one side of the cell, usually the upper side, imbeds itself in the cytoplasm and awaits further development (figs. 147, 148). The organization at so early a period of this definite peripheral layer of cytoplasm has not, I believe, been demonstrated for any of the other Gymnosperms. Finding the cavit}' containing the developing endosperm crossed b}^ irregular strands of cyto- plasm as illustrated in fig. 70, plate VI, I had the impression for a long time after these studies were begun, as stated in an earlier paper (1901''), that such a condition, as that described above for the resting macrospore, did not obtain until the beginning of the second period of growth. This layer of cyto- plasm is very easily displaced by the action of the fixing fluid, but with care it may be obtained in an apparently normal con- LIFE HISTORY OF PINUS 8 1 dition. I now have an abundance of preparations which show not only that the wall layer is instituted in the one-celled stage, but that it persists as long as free cell-formation continues in the endosperm. The only reference which I find regarding the establishment of the wall-layer of cytoplasm in any of the Gymnosperms is the following statement made by Coulter and Chamberlain ('oi), with reference to Pinus : " Probably when but two or three free nuclei have appeared the nuclei become imbedded in a parietal, cytoplasmic layer." SUMMARY. The female cones can be distinguished early in March, excepting in Pinus Strohiis where they do not appear until the very last of April. The ovules cannot be detected until about three weeks before pollination. There is no evidence that the macrospore-mother-cell arises from a hypodermal cell. When first differentiated it is cen- trally placed nearer the chalazal end of the ovule. The division of the macrospore-mother-cell is a true tetrad- division and the cell which gives rise to the female gametophyte is a true spore. Of the two cells formed as a result of the heterotypic division the lower one always divides again, the upper one may. An axial row of three cells seems to be the rule in Pinus Strohus and P. rigida, and one of four cells the rule in P. austriaca, though neither is constant in any of the species. The lowest cell of the axial row always becomes the functional macrospore. The two or three upper cells of the axial row begin to disin- tegrate very soon after they are formed and are finally absorbed by the enlarging macrospore. The lower cell passes through a long period of growth during which the cytoplasm is withdrawn from the central portion of the cell and forms a uniform layer lining the wall of the macro- spore. The nucleus moves towards the upper side of the cell and imbeds itself in the peripheral layer of cytoplasm. The suggestion is made that the embryo-sac may or may not be a morphological unit, but that it is essentially a physiological unit, existing for the purpose of sexual reproduction. Such a 82 MARGARET C. FERGUSON conception of the embryo-sac seems to the writer to form a more satisfactory basis for a rational explanation of the structure, or composition, and homologies of the embryo-sac than do any of the existing theories regarding the nature of this body. CHAPTER IV. The Female Gametophyte. development of the prothallium. The First Period of Growth. — We are indebted to Hof- meister ('51) for our first definite knowledge regarding the life history of the female gametophyte in the Gymnosperms. It is true some errors in observations were made, but they were inter- mingled with much that has stood the test of the most modern research. In 1879 Strasburger declared the " transitory endo- sperm " described by Hofmeister to be a fallacy, but he himself fell into quite as grave an error, though in the opposite direction, when he stated that the primary nucleus of the embryo-sac remained undivided during the first year, an observation since corrected by himself. As already stated, the young macrospore immediately organ- izes a peripheral layer of cytoplasm and passes through a period of growth which continues for six weeks or more. The degree of development which has been attained by P. austriaca on June 13, 1898, is shown in figs. 145 and 147 ; the first division of the macrospore-nucleus in this species occurred on July 29 of the same year, as illustrated in figs. 149 and 150. The germinating macrospore had now enlarged to such an extent that it was found necessary to reduce the scale of mag- nification at this point so that a comparison of the figures does not present, visually, the amount of growth which ensues between the organization of the macrospore and its first division. Pinus differs substantially in respect to the very marked growth of the macrospore before the first division of its nucleus from Larix where two nuclei are formed before there is any con- siderable increase in size of this cell (Juel ('00) plate xv, figs. LIFE HISTORY OF PINUS 83 18-20). The persistence of the potential megaspores in Larix at this time is also in very striking contrast to Pimcs, where the other cells of the axial row have become entirely absorbed before the germination of the macrospore occurs (figs. 147-149). The third division of the macrospore-mother-cell, or the first division of the macrospore-nucleus, takes place during the very last of July or the first of August in all the species studied, and is of the ordinary or typic method. It differs from the mitoses occurring in the vegetative tissue of the sporophyte only in presenting the one-half number of chromosomes (fig. 150). The daughter-nuclei may remain at one side of the pro- thallial cavity, but more frequently they pass to opposite sides as in the development of the embryo-sac in Angiosperms (fig. 151). The second mitosis follows rather quickly, and is already completed in Pinus Strobus on August 4 (fig. 152). Nuclear divisions follow until several free nuclei have been formed. The observations of Strasburger ('79), and of all later students of the Gymnosperms, upon the simultaneous division of the free nuclei of the endosperm have been confirmed. On October 12, 1898, sixteen nuclei were observed in the cytoplasmic layer, all being in the spireme stage of division. On October 15 of the same year sixteen nuclei, all presenting the equatorial plate- stage of mitosis were found in the cytoplasm of the prothallium, (figs. 153-155). The karyokinetic figure is sharply bipolar, each pole ends in a slight condensation of the cytoplasm, and the chromosomes are clearly of the reduced number. I find no evidence that any further divisions occur during the first period of growth and it is probable that the thirty-two nuclei which result from the division just described pass into the resting stage and remain inactive during the winter. But I have not examined a sufficiently large number of preparations with this point in mind to affirm that the prothallium of Pinus invariably enters upon its long period of rest in the thirty- two nucleated stage. The number may not be fixed even in the same species, but it is certain that it is never large. The pro- thallium, therefore, at the close of its first season of growth is a spherical body composed of an ectal layer of cytoplasm in which are imbedded, in many instances at least, thirty-two free 84 MARGARET C. FERGUSON nuclei. This thin cytoplasmic shell encloses a large central vacuole which is reported by Strasburger, Arnoldi and others to be filled with a fluid substance. I have made no observa- tions regarding the cell-sap of this large vacuole and can neither affirm nor den}-^ its presence. The Second Pei'iod of Growth. — It has been seen that the ovular development in Pinus is very slow during the period imme- diately subsequent to pollination, but with the renewal of growth in the spring development becomes much more rapid. Coor- dinately with the enlargement of the ovule already described, the endosperm cavity increases in size until it occupies almost the entire basal and central portions of the nucellus, presenting in longitudinal section the figure of an ellipse (fig. 71, plate VI). The thin peripheral layer of cytoplasm with its free nuclei persists until the latter part of May, and free nuclear division continues to take place within it until a large number of nuclei are formed. Jager ('99) estimated that there are 256 free nuclei formed in Taxus, and Hirase ('95) made the same observation in Ginkgo. The number is certainly much larger in Pinus. More than 500 free nuclei are present early in May and about 2,000 have been counted in Pinus Strobus at the time when the nnclei are being separated by the development of dividing walls. The free nuclei are considerably larger in surface view than the nuclei of the nucellar tissue, but in side view they often appear somewhat flattened. They have the structure of typical resting nuclei (figs. 156-159, plate XV). Each contains, almost invariably, two rather large nucleoli surrounded by clear areas. The reticulum is close and studded with irregular granules, but the net-knots are not so prominent as in the nuclei of the nucel- lus. They simulate very closely the nuclei of the sheath-cells at certain stages in the development of the archegonia. The cytoplasm in surface view presents a pseudo-alveolar structure consisting of a coarse, granular reticulum enclosing numerous vacuoles (fig. 156). During the late telophase in the division of the free nuclei of the prothallium the complicated karyokinetic figure characteristic of free nuclear division becomes very con- spicuous, and is evidently formed as a result of the rearrange- ment of the cyto-reticulum (fig. 159). LIFE HISTORY OF PINUS 85 At some time during the latter part of May in Pinus Strobus and about the middle of the month in the other species free nuclear division ceases and cell-walls are developed between the nuclei. The development of the prothallium from this point on was studied by Sokolowa ('80), and her observations have in general been confirmed by all more recent writers, with the exception of Jager ('99) in Taxus. I find the development of cell-walls in the prothallium of Pinus to agree perfectly in its early stages with that described by Sokolowa. Walls are formed perpendicular to the wall lining the prothallial cavity, thus each nucleus with its proper portion of the cytoplasm is separated from all the other nuclei. No wall is laid down on the inner sides of these cells, so that in radial section the cells appear as uncovered boxes, the opening extending towards the center of the prothallial cavity. In surface view the cells are more or less isodiametric, polygonal in outline and very uniform in size. A layer of densely reticulated cytoplasm surrounds each nucleus, and delicate strands radiate from it to the ectal layer of cytoplasm, thus giving a very different aspect to the cytoplasm than it had prior to the development of cell walls (figs. 160 and 161). Jager described the presence of walls on the inner face of these cells in Taxus when the cells were first organized, but other students have not confirmed his observa- tions. According to Sokolowa these cells grew inwards forming long open tubes which extended to the center without division, a wall was then formed at the inner end and the cells became divided by cross walls. To these long cells the name alveoli was applied. Only those from the sides extended clear to the center before being closed, those from the extremities becoming more or less wedge-shaped. Jaccard ('94) notes that \x\ Ephedra ^otcvq of the alveoli may divide before reaching the center, but many do not, while Arnold! ('99 and '01) finds that no division occurs in Sequoia until after the alveoli have met at the center and their ends have become closed by walls. The development sub- sequent to the formation of the open cells varies considerably in Pinus from that described by these writers for other Gymno- sperms. No cell has ever been observed to extend from the 86 MARGARET C. FERGUSON circumference to the center of the prothallial cavity. The cells are long, it is true, the walls delicate and wavy in outline, but a ring of tissue composed of longer or shorter cells is formed rather early in the inward growth of the prothallium. The cells of the innermost row always remain open on their outer free sides, their cytoplasm is more abundant than in the other cells of the prothallium and their nuclei invariably retain a position near the open side of the cells (fig. 162). As observed by Jaccard ('94), and Jager ('99), the nuclei of the prothallium cease to divide synchronously after individual cells have been organ- ized. When the center is reached the cells close and thus, one year after pollination, the endosperm becomes a solid mass of tissue. The prothallium grows rapidly after it has become a con- tinuous cellular body and in a few days it fills all the central and lower portion of the ovule. Above it is the prominent nucellar cap, while only a few cells of the nucellus remain along the sides separating the gametophyte from the integument (fig. 73, plate VII). Cell-divisions continue to take place, and the cytoplasm becomes more abundant, though the prothallial cells are never richly supplied with cytoplasm. Strasburger ('80), Jager ('99), and several more recent students have noted many nuclei in the endosperm cells. I have not observed multi- nucleated cells in the prothallium of Piniis up to the time when the suspensor has elongated and carried a several celled embr3^o to a considerable depth into the endosperm. Later stages than this have not been studied. There is often an appearance of more than one nucleus in a cell, but careful study never fails to demonstrate a delicate cell-wall between the nuclei. At an early stage in prothallial development the cell-walls are very delicate, scarcely more than condensations of the ectoplasm, so that they might easily be mistaken, in Pinus^ for strands of cytoplasm. Doubtless the cells become plurinucleated during a more advaned stage in embryo formation. THE SO-CALLED SPONGY TISSUE. The Fh'st Period of Growth. — When tiie macrospore- mother-cell first becomes apparent it is surrounded by a group LIFE HISTORY OF PINUS 87 of cells, three to five cells in thickness, which are more or less clearly delimited from the surrounding tissue by their slightly larger nuclei, their somewhat radial arrangement about the macrospore-mother-cell as a center, and, in some instances, by a rather indefinite and broken space which separates this group of centrally lying cells from the adjacent nucellar tissue (fig. 124, plate XII). At the close of the tetrad-division these cells have become much more conspicuous by the increase in the size of their nuclei, the somewhat greater density of their cytoplasm, and by the presence just exterior to them of an interrupted layer of tabular cells which are evidently undergoing disintegration. The disintegrating cells usually appear on one side first then at other points about equally distant from the young gametophyte (figs. 66, 69, plate VI ; 124, plate XII, 148, and plate XIV). It was to this tissue, immediately surrounding the young endo- sperm, together with the disintegrating cells just exterior to it, that Strasburger gave the name " spongy " tissue, and for convenience I shall use this term in speaking of it. Ovules are frequently found during the summer and fall which, so far as external appearances go, are perfectly normal, but, when prepared for study, reveal the fact that either the macrospore-mother-cell has never divided or the macrospore, if formed, has not developed. Such ovules do not renew their growth in the following spring. In those cases in which the development of the mother-cell or of the young gametophyte is arrested, very characteristic changes occur in the spongy tissue. These cells grow and become rich in cytoplasm even when the mother-cell does not divide, or when the macrospore fails to germinate. But after a time they, too, become inactive, their cytoplasm is gradually lost, their nuclei become dense and deeply staining, and their cell-walls are very greatly thick- ened (fig. 163, plate XV). This state of disintegration may enter in at any time during the first period of growth but it is more common before any divisions have occurred in the macro- spore. When the mother-cell fails to divide, the cells of the spongy tissue may grow until they almost equal it in size before showing signs of breaking down. In such instances they bear a very striking resemblance to the mother-cell, and might easily 88 MARGARET C. FERGUSON be taken by one not familiar with the history of this tissue for a group of macrospore-mother-cells (fig. i68, plate XVI). In fig. 148, plate XIV, the slightly reduced cytoplasm of the cells of the spongy tissue and the prominence of their cell-walls are sure evidences that pathological conditions have entered in, though all other parts of the ovule are still perfectly normal the process of disintegration having only just begun. Had this ovule been left in connection with the sporophyte for a longer time, the spongy tissue would undoubtedly have assumed later the character shown in fig. 163. It is this abnormal appearance which I believe led Hofmeister to conclude that there were two prothallia formed in the pines, one for each season of growth. Strasburger thought that Hof- meister mistook the normal spongy tissue for endosperm, and Coulter and Chamberlain have recently expressed the same view. Now the walls of the normal spongy tissue are never thickened but remain even less prominent than those of the nucellus. Hofmeister was surely too accurate a student of cells as cells to have fallen into such an error. It is a well- known fact that many ovules are organized in Pinus that never reach maturity and they are very frequently found in the autumn and late winter in the condition just described ; but with the renewed growth of the healthy ovules in the spring, these fail to develop farther and are soon detected by their smaller size. Shortly afterward they become brown and dead. Having found this thick-walled abnormal condition in the autumn and winter, and in the spring finding within the ovules then developing the large central cavity, it is not surprising that Hofmeister should have concluded that a thick-walled transitory endosperm was formed in the fall. The Second Period of Growth. — When growth is renewed in the spring the cells of the spongy tissue become organized for the first time into a definite zone from two to three cells thick which forms a hollow prolate spheroid immediately surrounding the endosperm, and limited on its outer surface by a thin stratum of disintegrating nucellar tissue. The cells and their nuclei are not only somewhat larger than those of the nucellus, but their most distinguishing characteristic is to be found in the greater LIFE HISTORY OF PINUS 89 density of their C3'toplasm^ which is almost identical with that of the prothallium, while the cells of the nucellus are scantily supplied with cytoplasm. These cells divide karyokinetically, and, as they increase in number, they press against the adjacent cells of the nucellus which become flattened against this con- stantly advancing tissue, and are absorbed, only to give place to other cells which meet a similar fate. Sometimes absorption seems to precede the outward march of the spongy tissue, so that this tissue is separated from the normal nucellus by a clear space made up of cells of the nucellus which have lost all their protoplasmic content, but which have not as yet suffered collapse (figs. 157, 158, plate XV). The parietal layer of cytoplasm which constitutes the endosperm remains always in closest con- tact with the inner surface of this tissue (fig. 71, plate VI). The cells of the spongy tissue are still prominent when the endosperm becomes a solid multicellular body. Soon after- wards, however, they show signs of disintegration, and at the time of fertilization they have, as a rule, entirely disappeared as cells, only the remnants of the cell-walls remaining. The spongy tissue is then represented by a deeply staining fibrous body of no definite structure which persists between the gametophyte and the nucellus (figs. 162, plate XV, and 72, plate VI; 73, plate VII). The Nature and Function of the Spongy Tissue. — The prominent character of the cells surrounding the prothallium in certain Gymnosperms has been commented upon, in a general way, by all students of the Ahieti^iece ; but, as was noted by the writer in 1900 and 1901 and confirmed by Coker in Taxodium^ 1902, the true nature and function of these cells seem to have escaped entirely the notice of previous writers, as they have in- variably been described as tissue showing evidence of breaking down. After a preliminary note regarding the nature of this tissue was sent to press in 1900, Lang ('00) described a similar layer of cells about the endosperm in Stangeria. He designated them as sporogenous cells and " possibly tapetal in nature." As recently stated (1903),^ these cells may possibly represent sporogenous tissue, each cell being a potential macrospore- 1 See note at end of Appendix. Proc. Wash. Acad. Sci., August, 1901. 90 MARGARET C. FERGUSON mother-cell, but there is no evidence from the standpoint of origin that such is the case in Pinus. They arise directly from a nucel- lus in which a few days before their appearance every cell was apparently like every other cell. This alone is not conclusive, as the functional macrospore-mother-cell has a similar origin, so far as one can see. But, what is more conclusive, the divisions in this tissue are according to the typic method and present the number of chromosomes characteristic of the sporophyte (figs. 164-167). If these cells were once, in some remote ancestor, sporogenous in nature, they have entirely lost their primitive function and have acquired a new and important function in connection with the development of the endosperm. This is not then a layer of disintegrating tissue, as described by all earlier students of the AbietinccB, but rather as alread}'^ noted by the writer (1901^) a definite zone of physiological tissue which is intimately connected with the nutrition of the young gametophyte. It doubtless not only passes on to the endosperm the nutrititive substances derived from the nucellus, but is itself active in the manufacture of food, as numerous starch grains are often found within its cells. It is probable, too, that it performs an important mechanical role in the wa}^ of protection. It not only forms a support for the prothallium in its multinu- cleated state, but gradually receding, it pushes before it, as it were, the tissue of the nucellus thus making room within for the growth of the delicate gametophyte. Though we now know that this is a far more important tissue than it was formerly thought to be, it does not seem to me wise to apply to it the name tapetum or to suggest a new name by which to designate it. Strasburger's term " spongy" tissue, although given when the nature of this tissue was not understood and being a misnomer so far as its structure and function are con- cerned, has obtained a wide usage in the literature of the Gym- nosperms, and should be retained, just as the term cell is still retained in all biological literature. DEVELOPMENT OF THE ARCHEGONIUM. The Ea7'ly Grozvih of the A?'chci>-o)iii(i)i. — The archegonia first become apparent during the latter part of May or the very first LIFE HISTORY OF PINUS 9I of June, the time varying somewhat with the species and with the season. The degree of development which the prothalHum has attained when the archegonia-initials make their appear- ance also varies not only in the different species but in the same species. The differentiation of the archegonia may be deferred until the prothallial cells have united to form a continuous tis- sue ; but it quite as frequently happens that, while there still remains a comparatively large, open space at the center of the prothallial cavity, certain cells at the micropylar end of the pro- thallium divide by periclinal walls more rapidly than do the other cells of the endosperm and become comparatively rich in cyto- plasm ; several of the superficial cells in this region do not so divide, but continue to grow, and are distinguished from the adjacent cells by their greater size, larger nuclei and more vacuolate cytoplasm. These are the initial cells of the arche- gonia (fig. 162, plate XV, and 169-171, plate XVI). In less than a week after an archegonium-rudiment has ap- peared, and while it is still quite inconspicuous, it divides, giving rise to a small upper cell, the mother-cell of the neck, and a large, lower cell which forms the venter of the archegonium (figs. 171, 172, plate XVI). The small cell immediately divides by an anticlinal wall, and the two cells thus formed divide by walls that are perpendicular to the first, the resulting four cells all lying in the same plane. These constitute what may be called the normal neck in Pimis StrohtLS (figs. 173, 177, 180). Con- siderable irregularity in the number and arrangement of the neck-cells has, however, been noted even within the same spe- cies. Frequently two of the four cells divide again, as figured by Strasburger for Piniis Strobiis in 1869, the six cells being arranged in a single layer (figs. 178, 183, plate XVI, and 212, plate XIX). Occasionally all four cells divide by anticlinal walls, the neck then consisting of eight cells, all of which lie in the same plane (figs. 179, plate XVI, and 213, plate XIX). In rare instances the four cells divide by periclinal walls, when the eight cells which compose the neck of the archegonium are dis- posed in two tiers of four cells each (fig. 187, plate XVII). This last represents the structure of the neck in Pinus sylvcstris as figured by Mottier ('92) and Blackman ('98), and it is evi- 92 MARGARET C. FERGUSON dently the usual condition in P. austi'iaca^ P. rigida and P. resmosa, but in these species, too, much variation obtains. Variation in the number of neck-cells seems to be of common occurrence in the Gymnosperms. It was first noticed by Hof- meister in 185 1 and has recently been discussed by Coulter and Chamberlain ('01). Murrill ('00) has figured considerable irregularity in the number and arrangement of these cells in Tsuga, while Coker ('02) shows a very marked variation in Podocar^tcs. At first the growth of the central cell is not followed by a corresponding increase in the amount of protoplasm, so that its cytoplasm early presents a very vacuolate appearance. There may be one large, irregular central vacuole, or delicate strands of cytoplasm may extend out from the nucleus to the ectoplasm, these strands meeting and fusing at irregular intervals to form vacuoles of various sizes. Thus a very beautiful pseudo-alveolar structure is presented. Webber ('01) describes the cytoplasm in the central cell in Zamia as representing at this time a foam struc- ture of great beauty. I have never observed in this or any cell in Pinus a cytoplasmic structure which, according to my inter- pretation, could be designated as a true alveolar or foam struc- ture in the sense in which Biitschli ('94) uses the term. As the central cell continues to enlarge its cytoplasm begins to develop more rapidly, many strands extending out into and across the vacuoles. Thus the size of the vacuoles is decreased while their number is greatly increased. The central vacuole, if present, may persist for a considerable time, or it may be re- placed at once by smaller vacuoles (figs. 172-175). Gradually the cytoplasm becomes more dense, and the vacuoles, receding from the periphery of the cell, especially from its base and sides, disappear last from its upper portion (figs. 176, 177). When the ventral canal-cell is cut off, the vacuoles have nearly or quite been replaced by a finely granular cytoplasmic reticulum in which a greater or less number of larger, more deeply staining granules are imbedded. These granules are frequently sur- rounded by a clear court into which the protoplasmic network has not extended. The number of the so-called proteid vacu- oles is usually small at this time (fig. 178). LIFE HISTORY OF PINUS 93 The nucleus of the central cell attains full size very soon after its formation. It has a delicate, more or less interrupted reticulum, and is characterized by a large vacuolate nucleus which invariably occupies a central position. One or two smaller nucleoli may also be present. This nucleus always remains close beneath the neck-cells, as is the case in other Gymnosperms, and, as a rule, is more or less concave on the side toward these cells (figs. 172-177,181-183). As Blackman has pointed out, the vacuolate nature of the cytoplasm renders this nucleus very liable to displacement during the early stages in the development of the archegonia, yet with well fixed ma- terial it is always found in its normal position. Hirase ('95) states that certain granules, which appear in the cytoplasm just beneath the nucleus of the central cell in Ginkgo^ have been derived from this nucleus or from its nucleolus. Ikeno ('98), also, describes the nucleus of this cell in Cyciis as giving out a granular substance during its growth period. No comparable phenomenon has been observed in connection with the nucleus of this cell in the species of pines which I have studied, but, as above stated, the nucleus quickly reaches its mature size and remains apparently unchanged until the inception of its division. Very early in the history of the archegonium, the cells imme- diately surrounding it become differentiated from the adjacent endosperm-cells by their more regular form, the greater density of their cytoplasm, and the increase in the size of their nuclei. Thus a distinct sheath is formed about the venter of the arche- gonium. This sheath usually consists of a single layer of cells. It is more conspicuous in Pinus resinosa than in the other species, and may become two cells broad at certain points, but even here it is never two layered to any considerable extent. The nuclei of these cells divide as the archegonium increases in size, the axes of the spindles being always parallel with that face of the cell which is adjacent to the ^^^' All the sheath-cells of a given archegonium have several times been observed in the same stage of mitosis, but this is very exceptional as these cells do not ordinarily divide simultaneously. The sheath-cells persist until after fertilization when they gradually lose their cytoplasm and resemble the other cells of the prothallium. 94 MARGARET C. FERGUSON Where adjacent archegonia crowd against each other these cells early become distorted and partially destroyed. It is often diffi- cult to demonstrate the presence of cross walls in the arche- gonium-sheath. Neither have I been able to satisfactorily demonstrate the presence of pores in the wall separating the sheath-cells from the egg. Hofmeister ('61-62), Goroschankin ('80, '81), Arnoldi ('00), and Coulter and Chamberlain ('01) all describe this wall in Pinus as thick and furnished with pores ; but if such is the case it is not apparent in my material. On the contrary the wall seems very thin and is scarcely differentiated from the ectoplasm. It may be that further search on my part will reveal both the " pits " and the " thickened wall," but thus far I have not detected either. No special attempt has been made to count the number of chromosomes in the nuclei of the various parts of the sporo- phyte and gametophyte, but whenever a nucleus was observed in which the chromosomes were particularly clear and distinct their number was always noted. In such cases twelve chromo- somes have invariably been counted in the nuclei of the sheath- cells. Chamberlain ('99) has found the same number in the corresponding cells of Pinus Laricio. The earl}^ development of the archegonium, as just described, agrees in the main with that given by Strasburger in 1878. As the archegonia grow the prothallium also continues to increase in size, several layers of cells being formed above the archegonia, except over their neck-cells. Here no prothallial tissue is laid down, so that there arises an opening in the endo- sperm leading from the neck-cells of each archegonium to the nucellar cap (figs. 177-180). The presence of funnel-shaped openings leading from the nucellus to the archegonia-necks in Pinus was noted by Hofmeister in 1851 and their origin was correctly described by him in 1862. In the last stages of pro- thallial development preceding fertilization, the sides of this tubular cavity often become very closely crowded together so that the passage is obscured. The number of archegonia in a single ovule varies in Pinus Strobjis, P. 7-igida and P. 7'csinosa from one to five, the usual number being three. In Pinus austriaca and P. montana var. LIFE HISTORY OF PINUS 95 uncinata the number is larger, averaging about five. As many as nine have been observed in a given prothaHium in Pinus montana var. uncinata. The form of the mature egg depends largely upon the number and arrangement of the archegonia. When there are not more than two or three, as is frequently the case in Pinus StrobuSy they may become almost spherical in outline. Division of the CenU'al Cell. — As the central cell prepares for division the cytoplasm between its nucleus and the neck- cells is apparently resolved into fine granules, and there is a more or less pronounced condensation of the cytoplasm about the lower side of the nucleus. At the same time the nucleolus disappears wholly or in part, the nuclear reticulum becomes more open and broken, and the chromatin collects or condenses at various places on the network (fig. 182). Soon a clear court, similar to that described by Hof ('98), Fulmer ('98), Nemec ('98 and '99), Strasburger ('00) and others, makes its appearance along the lower half of the nucleus. Inasmuch as this nucleus is pressed close against the neck-cells such a court does not arise along its upper side (figs. 183, 184). Delicate, granular threads cross this court and press against the nuclear membrane, while at the same time the upper and lower surfaces of the nucleus become irregularly indented (fig. 185, plate XVII). As the chromatin condenses to form the spireme, an achromatic network, as already described for the corresponding stage in the division of the generative nucleus in Pinus, becomes apparent in the nuclear cavity (figs. 182-185). When the spireme is fully established it presents a beautiful moniliform appearance, and the longitudinal splitting of the band becomes apparent at some points. The threads which arose earlier in the cj'toplasm seem at this time to have been again resolved into granules (fig. 186). Whether any of them enter the nuclear cavity and con- tribute to the formation of the achromatic spindle has not been definitely ascertained. The spindle, when formed, lies wholly within the area previously occupied by the nucleus. Webber ('01) finds the origin of the spindle in the division of the gener- ative cell in Zaniia to be intranuclear. Farmer and Williams ('96 and '98) ascribe such an origin to the spindles studied in 96 MARGARET C. FERGUSON the FucacecB^ and spindles of intranuclear origin have been described by others. But while the achromatic figure in the division of the central cell in Pinus comes to lie completely within the nucleus, I would not claim that it is wholly of nuclear origin ; if such were its source, the cytoplasmic activity in con- nection with this division would be inexplicable. The earliest stages in spindle-formation in this mitosis have not been ob- served as yet, but when the transitional steps between the phases represented in figs. 186 and 187 have been observed we shall doubtless find that the cytoplasm has had some part to play in the institution of the spindle. During the early metaphase of the division the nuclear membrane can still be distinguished, and clearly consists of a weft of threads (figs. 187, 188). I have not observed any phenomenon in the prophase of this mitosis at all comparable with the beautiful figure shown by Murrill ('00), as illustrative of the prophase of the division of the central cell in Tsuga, When the spindle arises, it is " multipolar in an axial plane " and thus corresponds, with slight variation, to the mitotic figure described by Duggar ('00) in the microspore of Symj^lo carpus Jxelidus, and by Wiegand ('99) in the microspore of Potamogcton foliosus. In Pinus^ however, the upper extremities of the threads do not at first unite into groups, but remain practically free, and are closely pressed against the neck-cells (fig. 187). The several poles, formed at the inner or lower extremity of the kar3^okinetic figure, soon draw together forming a single, very sharply defined pole ; or the fully developed spindle may remain more or less truncate at its lower end. Blackman describes this spindle as bluntly truncate at both extremities. I have fre- quently observed such a spindle during a late anaphase of the division, but this is only one of the various aspects which may be presented during metakinesis and later stages in this mitosis. The upper extremities of the achromatic spindle-fibers may never draw together at all ; they may unite to form two or more poles ; or they may give rise to one pole which may be blunt or very slender (figs. 190-194). But whatever form may be assumed by this spindle during the later stages in its develop- ment, there is always formed, at an early period, a diarch spindle LIFE HISTORY OF PINUS 97 which is muhipolar at one extremity and monopolar, or nearly so, at the other (figs. 187, 188). A similar figure is also organ- ized in the mitoses which occur in the development of the pol- len-grain, and at an early stage in the division of the generative nucleus in the pines, as already described in this paper ; and it is suggested that such a figure may be characteristic, at least in the higher plants, of those indirect divisions which result in the formation of nuclei or cells of unequal size. The chromosomes, when oriented at the nuclear plate, are in- variably in the form of U's or Vs. Blackman states that they are straight rods but he does not so figure them. The cell- plate, during the early stages in its formation, lies midway be- tween the developing nuclei, but when the daughter-nuclei are fully formed, the nucleus of the oosphere is, as a rule, farther removed from the cell-plate than is the nucleus of the ventral canal-cell. A prominent cell-plate is formed and the plane of cleavage separating the ventral canal-cell from the egg becomes evident in many instances before the disappearance of the spindle. As Chamberlain ('99) has shown, the lower portion of the spindle at this time is ordinarily convex, while the part within the ventral canal-cell is concave (figs. 195-197, plate XVII, and 200, 201, plate XVIII). I was able in several preparations similar to that illustrated in fig. 191 to count the number of chromosomes, and twelve or thirteen were found in both groups instead of eight as counted by Dixon ('94). The Ventral Canal-cell. — According to my observations, a definite wall, separating the canal-cell from the egg-cell, is always formed in Pinus. Coker has made the interesting observation that no wall is developed in Podocarpus, the nucleus of the ventral canal-cell lying free in the egg.^ As a rule the nucleus of the ventral canal-cell in Pinus does not present a normal appearance, but shows signs of disintegration very early in its history. It is doubtful, in some cases, if a nuclear mem- brane is ever formed, and there are probably instances in which fusion of the chromosomes never takes place at all. The kar- yokinetic structure shown in fig. 193 would very presumably ^ See note at close of Appendix. pS MARGARET C. FERGUSON give rise to such a nucleus, if we may so denominate it, as that illustrated in the ventral canal-cell of fig. 196 ; although Black- man, judging from such a figure as that portrayed in fig. 194, considers it impossible that the chromosomes of the ventral canal-cell should ever fail to fuse. The nuclear membrane, when present, very soon breaks down, and the chromatic sub- stance becomes scattered throughout the cell (figs. 198-202). This cell immediately preceding and at the time of fertiliza- tion ordinarily forms a deeply staining mass which lies just beneath the neck-cells and above, but in contact with, the egg (figs. 180, plate XVI, 202, plate XVIII, and 213, 2i5,plateXIX). Rare exceptions to the rapid disintegration of the canal-cell have been observed and will be described in the appendix to this paper. But in the study of several thousand archegonia of Pinus Strobus no instance has been found in which the nucleus of the egg and of the ventral canal-cell were similar in form. The nearest approach to a normal nucleus that has been observed in the ventral canal-cell of this species is that shown in fig. 197, plate XVII. Occasionally this cell is somewhat enlarged and is furnished with a rather scanty amount of cytoplasm in which distinct chromosomes, or chromatic figures of various forms are imbedded. Of the many variations that have been found to occur in the structure of the ventral canal-cell in the mature archegonium but two have been illustrated — figs, ipg and 199, plate XVIII. It is probable that in such instances a true nucleus has ever been formed if, indeed, the chromosomes have fused at all. The character of the cell at this time is such as to pre- clude the possibility that a division of this cell is being initiated. There seems to be a definite relation between the structure of the ventral canal-cell and the character of the upper part of the mitotic figure formed in the division of the central cell. This is plainly demonstrated by a comparison of figs. 190 to 197, plate XVII, and 200-202, plate XVIII. Figs. 190, 193, 196 and 202 represent an especially interesting series. The separation of the canal-cell from the cytoplasm of the oosphere, as Strasburger ('72) and Blackman ('98) have de- scribed in Pinus^ is, I believe, due to a shrinkage of the egg- cytoplasm caused by imperfect fixation ; and it is possible that a similar appearance in Cycas^ Ikeno ('98), has a like origin. LIFE HISTORY OF PINUS 99 MATURATION OF THE EGG. The Descent and Grozvth of the Egi^-nucletis. — The egg- nucleus is no sooner formed than it begins to increase in size, becoming greatly enlarged even before the disappearance of the spindle-fibers (figs. 196-202). As the nucleus moves toward the center of the oosphere, threads of more or less delicacy extend, in a radial manner, from its wall into the surrounding cytoplasm. These fibers are not equally well defined in all preparations, but, whatever the degree of their prominence, they are invarably more strongly differentiated about the upper side of the nucleus, and may extend from the nucleus to the top of the egg (figs. 202-204). As already stated, few, if any, vacuoles persist within the the venter of the archegonium at the time of the division of the central cell. Following their disappearance, there arise numer- ous spherical bodies, the so-called proteid vacuoles. Coordi- nate with the downward movement of the egg-nucleus, these bodies assume a position about the periphery of the oosphere, more especially at its base (the organic apex of Strasburger), and at its sides (figs. 179, 180, plate XVI, 214, plate XIX). Un- der a low power, the cytoplasm of the mature egg appears dense and finely granular; the ''proteid vacuoles" do not seem to differ materially from the protoplasm in which they are im- bedded ; and many deeply staining granules are scattered throughout the cell. With greater magnification, however, a very beautiful, granular reticulum becomes apparent. There is no suggestion of the alveolar structure described by Biitschli ('94). At times this reticulum is ever3^where crossed by short fibers which have no definite arrangment and are, apparently, not confined to any fixed period in the history of this cell (fig. 200). The spheres in the outer and basal portions of the cyto- plasm are resolved into very complex structures which, although they simulate the appearance of nuclei, could never be mistaken for such bodies by one familiar with cell-structures (figs. 202, 203.) No cytoplasmic radiations, similar to those described by Belajeff ('91) in Taxtis baccata, and by Dixon ('94) in Piniis sylvestrts, have been observed in connection with the fully lOO MARGARET C. FERGUSON developed egg-nucleus in any of the species of pines which I have studied. During the growth and downward movement of the egg- nucleus, it never presents, in Pimis Strohiis^ a definite network, such as is observed in the nucleus of the ordinary resting cell ; but it is characterized at a very early date by an open, inter- rupted reticulum, on which are arranged irregular granules of various sizes. This meshwork may be extremely delicate ; it may assume a heavy appearance ; or it may become very much interrupted and broken, many detached portions lying loose within the nuclear cavity (figs. 196, plate XVII, to 205, plate XVIII). The egg-nucleus of Pinus cmstriaca and P. montana var. unci- nata^ may frequently show from an early date a beautifully regu- lar reticulum (fig. 269, plate XXIV). Nucleoli have rarely been observed in this nucleus in Piniis Strobus during the first stages of its development (figs. 196, 199 and 200-201) ; but in Pintis aiis- triaca they occasionally arise very early (fig. 195). When the nucleus has attained considerable size, small, nucleolus-like bodies, containing a single central vacuole, appear in connection with the nuclear net ; and at the same time a slightly larger nu- cleolus is observed in the lower part of the nucleus, usually in connection with its membrane (fig. 202). As the nucleus con- tinues to grow, this nucleolus also increases in size, gradually becoming large and very vacuolate (figs. 203-205). When the egg-nucleus reaches maturity, it has attained huge dimensions, and its outline, depending on the form of the egg, is spherical or elliptical. The nucleolus, if demonstrable, is always found in the lower part of the nucleus ; and there are usually several smaller bodies, designated in this paper as sec- ondary nucleoli, scattered throughout the nucleus (fig. 205). These secondary nucleoli are invariabl}'^ found in connection with the reticulum, but, as Montgomery ('98) believed regarding apparently similar structures, they are probably caught in, not vitally united to it. They may be present in great abundance, or they may be entirely absent from the nucleus. The reticu- lum, on which the chromatic substance is disposed, presents numerous aspects, as already indicated in the description of this nucleus during its period of growth. Under ver}' high magni- LIFE HISTORY OI<^ PINUS lOI fication, it does not show, in normal conditions, a true granular structure ; but it may present a most delicate, interrupted, granular network ; or, it may consist of large, irregular, dif- fusely-staining masses which are united into an imperfect reticu- lum (figs. 206, «, and 206,0'). In the latter instance the chromatic granules are either too minute to be distinguished, or they have been dissolved in the linin ground-work. The linin, always very abundant in this nucleus, may form heavy hyaline cords, on which the chromatin is collected at irregular intervals (figs. 206, c, and 206, y") ; but it more often consists of less con- spicuous strands (figs. 206, b^ to 206, d). Great as are the vari- ations in the structure of this nucleus, its chromatin has always been found, in the species of pines which I have studied, to exist either in the form of irregular granules of varying sizes, or apparently dissolved in the linin. Such a resolving of the chromatin into nucleoli as that described by Chamberlain ('99) in Piniis Laricio and illustrated in his figs. 14 and 15 has not been observed in normal nuclei by the writer. Whether the various appearances presented by the egg-nucleus represent normal phases in its life history, or whether one is normal and the others are artifacts resulting from the action of fixing agents, is, of course, a mere matter of conjecture. But, inasmuch as these different aspects are characteristic of this nucleus during its period of growth, also after it has to all appearances reached maturity, and again at the time of its con- jugation with the sperm-nucleus, it seems reasonable to conclude that all are normal and correspond to definite physiological processes, which take place within the nucleus. Hertwig's ('98) interesting experiments on fed and unfed Aciinosfhcerhun are in point here. They seem to show conclusively that the structure of a nucleus varies with the character of the work which is being done by it. Strasburger ('84) described the nucleus of the oosphere in the AbietinecB as being densely filled with a granular substance which entirely obscured or masked the chromatin. This substance he called metaplasm, and virtually considered the nucleus a vacuole filled with a nuclear sap capable of taking up or elaborating this material. Ikeno found a similar substance in the sexual nuclei I02 MARGARET C. FERGUSON in Cycas in 1898 and more recently in Ginkgo ('01), and Arnoldi ('00) in Cephalotaxtis. Blackman ('99) devoted several para- graphs to a discussion of metaplasm, as it manifested itself in the egg-nucleus of Pinus sylvestris. He found that it was present in the ^^oung nucleus in the form of granules, but that it later united with the chromatin to form the nuclear reticulum. Chamberlain ('99) does not recognize the presence of this sub- stance in the egg-nucleus in Pinus Laricio ; and there is no evidence of its existence in the sexual nuclei of the species of pines which I have studied. According to Wilson ('99) "protoplasmic substances repre- sent the active, metaplasmic structures the passive elements " of the cell. During the development of the egg-nucleus in the species of pines which have formed the basis of these studies, there is never any deposit within the normal nucleus of a granu- lar substance ; but the linin, as already stated, becomes very abundant. Just what proportion of it is active in cell division, we are unable to say. Without doubt a large part of the linin merges into the cytoplasmic network during the first segmen- tation of the oosphere-nucleus, but even so, it can not be classi- fied with the passive elements of the cell. Blackman ('98) wrote: "The stage in which the nucleus is found in a position between the apex and the center of the egg is rarely met with" ; and Chamberlain ('99) stated " that in over three hundred preparations, less than a dozen " show early stages in the development of the egg-nucleus. During the course of these investigations upon the pines, about four thousand preparations, representing many thousand archegonia, have been studied, and no developmental stage has been more frequently met with than that by which the nucleus assumes its central posi- tion in the egg. Such an appearance as that illustrated by Chamberlain in his figs. 18 and 19 has been observed in both the young and the mature egg-nucleus, in the conjugating nuclei, and also in the various nuclei of the proembr^-o. They have been wholly disregarded in the present discussion of the matura- tion of the ^^^t for, in my material, these figures, and also Blackman's figure 11, would be interpreted as representing dis- integration stages. Every step has been repeatedly traced from LIFE HISTORY OF PINUS IO3 the ordinary nuclear reticulum, to nuclei which can scarcely be distinguished from the surrounding cytoplasm, and then to arche- gonia, which appear perfectly normal except that no nuclei can be demonstrated within them. It is a well known fact, already commented upon in this paper, that the number of seeds derived from a pine cone is very small in comparison with the number of ovules formed in the same cone. An examination of fresh material shows that development may cease at any point be- tween the early stages in the formation of the ovule and the last steps in the ripening of the seed. This cessation of growth effecting first individual cells does not at once become apparent, and so cannot be avoided, in its earliest stages, when one is putting up material for cytological work. Under such condi- tions, it is inevitable that, with a limited amount of material, the abnormal will be interpreted for the normal. The entire development of the archegonium in Pinus is passed through in about two weeks, probably not more than five days elapsing between the cutting off of the ventral canal-cell and fertilization. In Pinus montana var. uncinata these processes are apparently much more closely united in point of time, as the pollen-tube, in some cases, has reached the endosperm before the division of the central cell is complete (fig. 207, plate XIX). The P}'oteid Vacuoles. — The true nature of the proteid vacuoles is a subject which attracted my attention very early in the course of these investigations. There can be no doubt that there is an intimate relation between the sheath-cells of the archegonia in the pines and the substance of the egg, such as is believed to exist between the follicle-cells and the egg in ani- mals. But the exact nature of this connection in Pinus is not easily determined. I have rarely examined a preparation show- ing archegonia without studying the relation of the sheath-cells to the oosphere ; and yet no entirely satisfactory evidence, be- cause not demonstrable beyond a question, of the origin and nature of the so-called proteid vacuoles has been found. Hirase C9S) observed that the granules in the egg of Ginkgo were of nucleolar origin, being derived both from the nucleus of the central cell and from the nuclei of the sheath-cells. 104 MARGARET C. FERGUSON Arnoldi ('oo) found that substantially the same thing was true in Cephal ataxics . He was not able to detect the passage of the nucleoli from the sheath-cells into the ^%^i but, since these granules were present on both sides of the membrane of the egg-cell he accepted the fact of their transference. I have frequently seen a nucleolus partly without and partly within the nucleus of a sheath-cell ; but in no instance could I be sure that such a condition was not the result of mechanical displacement. Ikeno ('98) found direct evidence that the nutritive spheres in Cycas are of nuclear origin. But no such phenomena as he observed in Cycas occur in Pinus. Platner ('86) described the passage of the follicle-cells into the ovum in Helix^ and a few other such instances have been recorded in animals. Arnoldi ('00) has recently noted a most remarkable migration of whole nuclei from the sheath-cells into the ^^^ in several species of pines. He has observed, in a single series, as many as one hundred and fifty nuclei passing into the ovum. From the fact that Arnoldi writes ^'- Strobus"" in a parenthesis after Pinus Pence, I infer that he employs the terms as synonyms ; but I find no authority for such a usage, and cannot accept his con- clusions as holding good for Pinus Strohus. It does not seem possible that, in a careful examination of several thousand archegonia, so obvious a phenomenon as that described by Arnoldi could have escaped detection ; and I must, therefore, conclude that it does not take place in the species of pines which I have studied. I fully believe that the sheath-cells play an important role in the nutrition of the Q.gg ; but it is the method by which this is accomplished, as described by Arnoldi, that I cannot accept for the species of pines studied. Coulter and Chamberlain ('01) not only accept Arnoldi's observations for Pinus but describe a like phenomenon in Cycas. Basing their statement on the results of Ikeno's studies, they record, on page 22, the following surprising fact with reference to Cycas: "The contents of the jacket-cells, nuclei and all, now pass through the pores into the central cell." I find no authority for such a statement in Ikeno's paper. If I correctly translate the German, Ikeno describes neither the transmission of the nucleus nor of the cytoplasm from the sheath-cells into the egg, LIFE HISTORY OF PINUS IO5 but he does note a most interesting transfer of nuclear sub- stance, that is, a substance secreted by the imclci, from the nuclei of the sheath-cells into the cytoplasm of the Qgg. In the course of his discussion Ikeno says : " Bemerkenswerth ist es ferner, dass der Zellkern der Wandungszelle hiiufig sich der Centralzelle niihert und dort einen nach dem nachsten Plasma- faden gerichteten kurzen Schnabel bildet (fig. 6). In einen andern Fall beobachtete ich, das der Zellkern der Wandungs- zelle sich bis an die Cellulosemembran begiebt, welche an die Centralzelle angrenzt und mit dem ganzen Korper an diese sich anlegt (fig. 7, a, b). Offenbar sollen alle diese Vorgange den Uebergang des in diesen Zellkernen enthaltenen Stoffes nach der Centralzelle erleichtern." So far as I am aware then, Arnoldi is the only investigator who has observed the passage of entire nuclei into the egg in the Gymnosperms. Some interesting^ observations have been made during this study regarding the nature of the nucleolus of the egg-nucleus. As already indicated this nucleolus does not arise in Pinus Strobtis until the egg-nucleus has attained considerable size. It appears in the lower part of the nucleus as a minute, solid, spherical body ; during growth a small central vacuole appears, then other vacuoles, until, at maturity, it is completely filled with vacuoles of various sizes (figs. 202-205, plate XVIII). A limiting membrane is not always apparent in this nucleolus (fig. 208, plate XIX; but in some instances, there seems to be very strong evidence of such a membrane (figs. 205 and 209). In fig. 205 the nucleolar wall has been broken at one place and a vacuole, lying near the point of rupture, has been in- dented along its outer surface, thus becoming crescent shaped. Montgomery ('98) sounded a word of warning against inter- preting the peripheral stratum of the ground substance of the nucleolus as a wall layer ; and there is a possibility that, in the figures above referred to, what appears like a limiting mem- brane is only the outer unmodified portion of the nucleolus. The attitude of this nucleolus toward dyes varies much at different periods in its history. It may or may not take the safranin stain characteristic of Flemming's triple combination ; it may stain intensely with gentian-violet or iron haematoxylin Proc. Wash. Acad. Sci., August, 1904. I06 MARGARET C. FERGUSON (figs. 205 and 20S) ; it may show a weak reaction to these stains (fig. 209), or it may be absolutely unaffected by them, remaining as a hyaline or greenish yellow structure (fig. 210). When the nucleolus resists the action of d^^es, its nucleus is usually totally free of the secondary nucleoli, which have been described in connection with the maturation of the egg-nucleus, and the cytoplasm of the egg is studded, to an unusual degree, with large, deeply staining granules. But the nucleus containing a nucleolus which stains with avidit}^, generally contains, also, innumerable secondary nucleoli ; at the same time, there are comparatively few deeply staining granules in the cytoplasm of the egg. The position of the secondary nucleoli with reference to the primary nucleolus is frequently such as to indicate that the former originate in the latter (figs. 227, plate XX, and 208, plate XIX). The only observations which would militate against such an origin are the few cases found in which the secondary nucleoli seem to appear earlier than the primary nucleolus (fig. 195, plate XVII). It may be that, in these cases, the primary nucleolus has not yet become differentiated in structure from the secondary nucleoli, as would evidently be true in a stage slightly younger than that shown in fig. 202, plate XVIII ; or it may be true that the primary nucleolus is pres- ent, but fails, at this time, to stain. Floderus ('96) describes a somewhat similar origin of the paranuclei, in Tunicates, from the nucleolus proper. The nuclei of the cells surrounding the archegonia contain from three to five nucleoli, and one or more nucleolus-like struc- tures may be present in the cytoplasm of these cells. Each nucleolus is surrounded by a clear court which, as Zimmermann ('96) has pointed out, is evidently not an artifact. Debski ('97) opposes this view, however, and considers the clear court to be attributable to the shrinkage of the nucleolus, since he does not find it when material is treated with xylol instead of cedar oil. These nucleoli may be spherical, elliptical, irregular, or long and almost dumbbell-like in outline. The ordinary cells of the prothallium do not show nucleoli. If such bodies be present in these cells they are small and obscured by the nuclear reticulum. LIFE HISTORY OF PINUS IO7 At about the time of the cutting off of the ventral canal-cell man}' small nucleolus-like masses appear in the nuclei of the sheath-cells — twenty or more occurring in a single nucleus. When the egg has reached maturity, and during the later stages of its histor}', no nucleolus, or but one or two nucleoli, can be demonstrated in the nucleus of a sheath-cell. These nucleoli are no longer surrounded by a hyaline court, but are imbedded in the chromatic network. The nucleoli of the sheath-cells present the same attitude toward stains as does the nucleolus of the egg-nucleus. But ■while the nucleoli of the sheath-shells frequently stain but feebly they rarely fail entirely to stain. Similar color reactions have been observed in connection with the nucleoli, as already described, in the microspore-mother cell of Pinus. The occurrence of unstained nucleoli in the same nucleus in which others were deeply colored is common in the microspore-mother-cells especially at about the time of synapsis. I am aware that conclusions based upon staining reactions alone are not to be trusted, but when accompanied, as here, with other phenomena they may be highly significant. The nucleolus of the egg-nucleus and also the nucleoli of the sheath -cells in Pinus appear to represent active portions of the cell rather than inert masses of matter. Certain aspects presented by these nucleoli are surely suggestive of plastids. The uncolored framework of the egg-nucleolus reminds one very strongly of a chlorophyll body from which the pigment has been extracted. Yet we would not, in the present state of our knowledge, denominate them plastids. I believe, however, although the phenomena are not of such a nature as to admit of definite demonstration, that the nucleolus of the egg-nucleus, and also the nucleoli of the sheath-cells are actively engaged in the formation of a substance which in the egg-nucleus, at least, assumes the shape of secondar}^ nucleoli. These nucleoli be- come diffused throughout the nucleus, from which they pass, probably in solution, into the egg cytoplasm. Here they are again differentiated, and by a gradual development, give rise to the " proteid vacuoles " or nutritive spheres of the oosphere. It may be that the greater size of the egg-nucleus, in com- I08 MARGARET C. FERGUSON parison with that of the sperm-nucleus, is correlated with the physiological role, as above suggested, which it plays in the cell. We cannot, here, enter into a discussion of the volu- minous literature dealing with the origin, function, and destiny of the nucleoli ; but a few of the many views which have been advanced may be noted. Strasburger ('95, '97 and '00) expresses his conviction that nucleolar substance contributes to the formation of spindle- fibers. A similar view is held by Fairchild ('97), Harper ('97), Debsky ('97), and other students of the Bonn Laborator}', and by Nemec ('99), Farmer ('94) and others. Strasburger ('95) also sees indications of a connection between the nucleolus and the cell-plate and he has recently ('97 and '00) sought to show that the nucleoli make active the spindle-forming substance in the cytoplasm, or that they enhance the activity of the kino- plasm. Flemming ('82), Humphrey ('94), Zimmermann ('95), Sar- gant ('96 and '97), Duggar ('99), Mottier ('00), and many others believe that the nucleoli represent reserve supplies of chromatin. Dixon ('99) finds in them a vehicle of inheritance. Hirase ('98) thinks that they give rise to the attractive spheres ; and accord- ing to Karsten ('93), Lavdovvsky ('94) and Wilcox ('95) they are centrosomes. Rosen ('95) considers that the nucleoli are equal in dignity to the chromatin, that they have no connection with the centrosome and that they do not serve to nourish the chro- mosomes. Jordan ('93) states that " their function is almost certainly one of nutrition either concerned in the storage or elaboration of nutritive material " and believes that there is substantial reason for looking upon the nucleolus wherever found as concerned in one way or another with the active metabolism of the cell. Lukjanow ('88) and Macallum ('91) consider the nucleoli to be excretory organs which are intimately related to the nutritive spheres of the Qgg, these spheres arising through a process of deposition from the nucleolus. And Hacker ('93) observes that the nucleolus is a contractile vacuole which absorbs proteid substances : the absorbed materials under"■), secondary nucleolus {sy.ns.), receptive vacuole {r. v.). All the figures have been given their normal position, as nearly as it was possible to do so, on the plates. That is, they are so placed that the primary axis of the ovule would be parallel with the longer axis of the plates ; and the portion of a figure nearest to the micropylar end of the ovule is always towards the top of the plate. ( 154 ) PLATE I. Fig. I. Two cells of the primitive archesporium showing the -winter condition of this tissue. X 1,400. Pintis austriaca. December 20, 1897. 2. A cell from the primitive archesporium in the early spring. Many of the cells of the archesporium are undergoing division at this time. X 1,400. Pinus austriaca. March 14, 1898. 3. A cell of the definitive archesporium, the microspore-mother cell, just prior to the inception of its division. X 1,400. Pinus austriaca. April 27, 1S9S. 4. The same as fig. 3. X 1,400. Pinus Strobus. May 24, 1S98. 5. The microspore-mother-cell approaching synapsis before a definite spireme has been formed. X 1,400. Pinus austriaca. April 28, 1S98. 6. The same as fig. 5. X 1,400. Pinus Strobus. May 24, 1898. 7. Synapsis. X 1,400. Pinus Strobus. May 24, 1S98. 8. Recovery from synapsis, showing a continuous spireme. X 1,400. Pinus Strobus, May 24, 1S98. Material showing figs. 4 and 6 was col- lected from a different tree than that showing figs. 7 and 8, and the microspore-mother-cells were in a slightly different stage of division. 9. Complete recovery from synapsis. Chromatin in irregular granules, on a broad linin band. X 1,400. Pinus Strobus. 10. The longitudinal splitting and transverse segmentation of the spireme. Chromatin still distributed in irregular granules. X 1,400. Pinus Strobus. 1 1. Longitudinal splitting completed, but the sister segments do not become entirely disunited. Nucleoli still apparent. X 1,400. Pinus Strobus. 12. a-e. Portion through the edge of a nucleus showing the twisting of the' chromatic segments after longitudinal splitting. In most in- stances these are not entire segments but portions that have been severed by the microtome knife. The entire segments are very long and coiled at this time. X 1,400. Pinus Strobus. 13. Early stage in the condensation and fusion of the longitudinally divided spireme. Threads anastomosing in region of nucleoli. X 1,400. Pinus Strobus. (156) Proc. Wash. acao. Sci. Vol. Plate I. Oi 3^ w 0" 2 *- i^r'il/^ M- C. F., DEL. "'"H.'v#2^* *''•'• FERGUSON, -PINUS. MiCROSPOROGENESIS. HELIOTYPE CO., BOSTON. PLATE II. Fig. 14. A more advanced stage in contraction, showing that adjacent threads are anastoniosing and fusing. X 1,400. Pi?ius Strobus. 15. A still more advanced stage in the fusion of the threads. Practically all evidence of the earlier longitudinal fission has now disappeared. X 1,400. Pinus Strobus. 16. The chromosomes becoming apparent. X 1,400. Pinus Strobus. 17. Distinct chromosomes, in the one half or reduced number, arising from the contracted and more or less anastomosed skein. X 1-400. Pinus Strobus. 18. a-c. Final stages in the formation of the chromosomes, showing the separation of the segments from one another, and also the relation of some of them to the nucleolus. X 1,400. Pinus Strobus. 19. a-l. Various forms of chromosomes observed before the organization of the spindle. Each chromosome consists of two of the longitudinal split segments which were formed immediately subsequent to synapsis. X 1,400. Pinus Strobus. 20. The chromatic segments completely differentiated. The remnant of a nucleolus is still present, and the nuclear membrane is being resolved into threads. X 1,400. Pinus Strobus. 21. An early stage in spindle-formation, showing kinoplasmic threads entering from all directions but as yet no poles, or centers of radiations, have been established. Chromosomes are homogeneous in structure and regular in outline. X 1,400. Pinus Strobus. 22. The tripolar spindle. X 1,400. Pinus rigida. May 4, 1S9S. 23. The spindle has become nearly bipolar. X 1,400. Pinus rigida. (158) Proc. Wash. Acad. Sci. Vol. Plate II. M C. F., DEL. HEUOTYPE CO., BOSTON. FERGUSON, -PINUS. PLATE III. Fig. 24. The equatorial plate stage. Spindle definitely bipolar and reaching to the ectoplasm. X Ij400. Pinus rigida. 25, 26. The metaphase of the heterotypical division ; chromosomes irreg- ular in outline and apparently much larger than in the late prophase. X 1,400. Pinus Strobus. 27-29. Anaphase of the heterotvpical division. The longitudinal split- ting of the chromosomes has been very greatly delayed in some cases. Such an appearance as that shown in fig. 29 is frequently met with, the stretched arms of the daughter chromosomes extending nearly the entire length of the spindle. X 1,400. Pifius Strobus. 30. The chromosomes just after reaching the poles, as seen in looking down upon the end of the pole. X 1,400. Pinus Strobus. 31-34. Stages in the development of the daughter-nuclei. A definite resting nucleus is formed at the close of the heterotvpical division. X 1,400. Pinus Strobus. 35. A late telophase in the first division, the daughter-nuclei fully estab- lished. Delicate spindle threads still present, but no indication of a cell plate. The wall of the microspore-mother-cell is beginning to thicken centripetally. X 1,400. Pinus Strobus. 36-37. Stages in the formation of the spireme for the second division. X 1,400, Pinus Strobus. (160) Proc. Wash. Acad. Sci. Vol. Plate II .^*. 11 25 ■i 26 27 29 31 M. C. F., DEL. HEUOTYPE CO., BOSTON. PLATE IV. Fig. 38. Origin of the second spindle; the chromatic band looped in region of the future equatorial plate, and showing longitvidinal fission. Xi!400. Pinus rtgida. 39. Transverse segmentation is completed ; and the distinct chromosomes have become apparent at the equatorial plate of the multipolar diarch spindle. X 1,400. Ptnus Siyobtis. 40. Separation of the daughter-chromosomes of each pair formed by the transverse division shown in figure 39. X 1,400. Pinus Strobits. 41. Daughter-chromosomes arranged in two parallel rows at the equatorial plate. X 1,400. Pitius Strobus. 42. A late anaphase in the second division. X 1,400. Pinus Sirobus. 43. Early telophase of the second division. X Ij400. Pinus Strobus. 44. Late telophase of the tetrad division ; the chromosomes of each nu- cleus have fused to form a spireme, but the nuclear membrane is not yet developed ; rather faint cytoplasmic threads connect the four nu- clei; the centripetal thickening of the mother-wall becomes more apparent. X 1,400. Pinus rigida. 45. The tetrad division is completed and the 3'oung microspores are dis- tinctly differentiated, each surrounded by its own wall. Xit40O. Pitius rigida. May 10, 189S. 46. The four microspores are separated by very prominent walls which are continuous with the broad wall lining the original wall of the microspore-mother-cell ; the outer, original spore-mother-wall is sepa- rated at two points from the thick, more recently formed inner wall. X 1,400. Pinus austriaca. May 9, 1898. 47. Microspores still within the mother-wall and showing the beginnings of the wings or air-sacs. X 1,400. Pinus Strobus. May 30, 1898. 48. Rupture of the mother-wall and escape of the microspores. X Sio. Pinus Strobus. May 30, 1S98. (162) Proc. Wash. Acad. Sci. Vol. Plate IV. 41 ^^ ^€1 ft 42 \.,,, M. C. F., DEL. 39 40 FERGUSON, -PINUS. MICROS PnRnnFNF9l9 HELIOTYPE CO., BOSTON. PLATE V. Fig. 49. Empty wall of the microspore-mother-cell showing the compartments formerly occupied by the microspores. X Sio. Pinus Sirobus. 50-54. Stages in the growth of the microspore ; the inner, partial wall very apparent in the mature spore. Fig. 53 represents a section through the middle of a young microspore in a plane perpendicular to the wings. X Sio- Pinus Strobus. 54-55. Stages in the first division of the microspore-cell ; the spindle sharply pointed on the ventral side, broad on the dorsal side. X 810. Pinus Strobus. June 7, 1898. 56. Telophase in the first division of the microspore. X Sio. Pitius Strobus. 57. The resting stage following the first division of the microspore. X 810. Pinus Strobus. 58. The same as Fig. 57, but showing an exceptionally large prothal- lial cell. X 8io- Pinus Strobus. 59-60. Spireme-stage and early telophase in the division to cut off the second prothallial cell. X^io. Pinus Strobus. 61. The germinated microspore at the close of the second division, show- ing the first prothallial cell already in an advanced stage of disintegra- tion. X 810. Pi7ius Strobus. 62-63. Stages in the third division of the microspore, showing the rapid and almost complete obliteration of the first and second prothallial cells. Both prothallial cells are cut off from the apical cell by definite walls. X Sio. Pinus Strobus. ( 164) Proc. Wash. Acad. Sci. Vol. Plate v. M. C. F., DEL. FERGUSON, -PINUS. npvpi nPMCMT r\c Doi I iiM_,~i HELIOTYPE CO., BOSTON. PLATE VI . Figs. 64-65. Mature pollen-grains ; in fig. 64 the remnants of the two prothal- lial cells can be seen, while in fig. 65 all signs of the first cell have disappeared. X^io* Pinus Sirobtis. June 9, 1898. 66. Vertical section through an ovule immediately after pollination ; the macrospore-mother-cell is very conspicuous ; the upper portion of the free limb of the integument is shown to be three cells in thickness, there is a slight concavity in the apex of the nucellus ; macrospore- mother-cell {m.m.c), nucellar cap {nuc), micropyle {mic). X 46- Pinus rigida. May 27, 1902. 67. Vertical section through the upper part of an ovule showing pollen- chamber ; the middle layer of cells in the upper part of the free limb of the integument has elongated and closed the microcarpylar canal. X 46. Pinus rigida. June i, 1902. 68. A vertical section through the upper part of an ovule. The elongated cells noted in fig. 67 have become divided by the formation of cross walls into smaller cells. X 46. Pinus rigida. June 4, 1902. 69. A vertical section through an ovule some days after pollination. Axial row [a.r.). X 62. Pinus Strobus. June 17, 1S9S. 70. A vertical section of an ovule showing the winter condition. X62. Pinus Strobus. January 4, 1898. 71. A vertical section of an ovule soon after the second period of growth has begun. X62. Pifius Strobus. May 26, 1898. 72. A vertical section through the upper part of an ovule at the time of the division of the generative nucleus ; («ac.i), that portion of the nucellar cap which was developed during the first period of activity; {nuc. 2), that portion of the nucellar cap which constitutes the second year's growth ; o, disintegrating spongy tissue. X 62. Pinus Strobus. June 9, 1898. (166) Proc. Wash. Acad. Sci. Vol. Plate V|. ■'-V ' -^ 'i^^^Tjitf^' '' ■flL-xj' "• C. F., DEL PLATE VII. Fig. 73. A vertical section through the upper part of an ovule shortly before fer- tilization ; reconstructed from three adjacent sections of the series; o, last vestige of spongy tissue. X ^2. Pinus Strobus, June 15, 1898. 74. Pollen-grain from the nucellus of Fig. 73. The antheridial cell is still undivided. X 472. 75. A vertical section through the extreme upper portion of an ovule soon after pollination, showing the uppermost part of the nucellar cap, and a pollen-grain in the first stages of germination ; p.c, pollen-chamber. X472. Pinus Strobus. June 13, 1S9S. 76. A pollen-grain soon after germination. The tube-nucleus is moving into the pollen-tube. X 472. Phius Strobus. June 24, 1S9S. 77. A pollen-grain after the tube-nucleus has passed into the pollen-tube. X472. Pinus Strobus. July 15, 1898. 78. Spireme stage in the division of the antheridial cell. y(^i,^oo. Pinus rigida. April 27, 1898. 79-80. Stages in the division of the antheridial cell. Xi'400' Pinus Strobus. August 4, 1898. 81. A poUeh-grain after the antheridial cell has divided. X 472- Pinus Strobus. August 4, 1898. 82. The same at a later date, showing a slight increase in the size of the generative cell. X472. Pinus Strobus. October 7, 1898. ( 168 ) KROC. WASH. AUAU. OCl. VUL. M C. F., DEL. HELIOTYPE CO., BOSTON. FERGUSON -PINUS. GROWTH OF THE POLLEN-TUBE. Proc. Wash. Acad. Sci., September, 1904. PLATE VIII. Fig. 83. The pollen-tube which is shown in fig. 70, more highly magnified. X 472. Pinus Sirobus. January 4, 1899. 84. A pollen-grain and the upper portion of a pollen-tube, showing the stalk- and the generative-cell just before their passage into the pollen- tube. X472. Pinus austriaca. May 3, 1898. 85, 86. Later stages than the above, showing the passage of the generative- and the stalk-cell into the pollen-tube ; in fig. 86, the two cells are breaking loose from each other. X472. Pinus austriaca. May 10 and 17, 1898. 87. The male gametophyte at the time of the entrance into the tube of the generative- and the stalk-cell ; «./", a bit of the dead nucellar tissue. X472. Pinus St robus. June 9, 189S. 88. A pollen-grain after the generative and the stalk-cell have passed into the pollen-tube ; taken from the top of the nucellus of the ovule shown in fig. 72. X472. Pinus Strobus. June 9, 1S98. 89. A few of the cells from that portion of the nucellar cap marked nuc.2 in fig. 72. The cells are filled with starch grains. X 472- Pinus Strobus. June 9, 189S. 90-92. Portions of pollen-tubes showing successive stages in the passage of the stalk-cell over the generative cell, as also the presence of large quantities of starch in the pollen-tube. X472. Pinus resi7iosa. June 2, P. Strobus, May 24 ; P. rigida, June 8, 1898. 93. The generative cell, bearing on its surface both the tube-nucleus and the stalk-nucleus. In this instance the stalk-cell has passed beyond the tube-nucleus. X 472. Pinus resinosa. June 3, 1898. 94. The generative cell showing a very early stage in the formation of the spindle. The nucleus is in the extreme uppermost part of the cell. X 744. Pinus rigida. June 8. (170) Proc. Wash. Acad. Sci. Vol. Plate VIII. HEUOTYPE CO., BOSTON. FPDi^llQniM _DIMIIQ PLATE IX. Figs. 95-96. The generative cell in the early stages of its division, showing granular condensation and radial arrangement of cytoplasm. The spindle fibers arise in the cytoplasmic condensation and extend in the form of a cone to the nuclear membrane. X 744- Pinus rigida. June 8 and 10, 1898. 97. A cross-section through the generative cell during an early stage in its mitosis. The protoplasmic condensation is seen from below looking toward the nucleus. X 744- Finns austriaca. June 4, 1898. 98. A later stage in the division of the generative nucleus. X 744- Pinus austriaca. June 10, 1898. 99. The generative cell just before the disappearance of the lower portion of the nuclear membrane showing a single deep indentation on the lower side of the nucleus. X 74^- Pinus Strobus. June 9, 1898. 100. A stage in spindle-formation directly following that shown in fig. 99. The nuclear membrane has given way and the spindle fibers are enter- ing the nuclear cavity. The nucleolus is still distinctly visible. X 744- Pinus Strobus. June 10, 1898. 101. The gradual disappearance of the nuclear membrane and the extension of the spindle fibers across the nucleus. X 744- Pinus austriaca. June 7, 1898. 102-103. Further development of the spindle and the formation of the chromosomes. The marked condensation in the cytoplasm from which the spindle arose has almost entirely disappeared. X 744- Pinus austriaca. June 8, 1S98. (17O Proc. Wash. Acad. Sci. Vol. Plate IX. ;5S;>^WD>>v, (^4^ ^i. ) M. C. F., DEL. 102 FERGUSON, -PINUS. 03 HELIOTYPE CO., BOSTON. PLATE X. Figs. 104-106. Later stages in the development of the spindle showing the gradual drawing together of the outer extremities of the threads to form the upper pole of the spindle. The upper pole of the spindle does not reach the nuclear membrane, but in fig. 105 definite threads extend from the pole to the nuclear membrane. X 744- Fig. 104. Pinus rt'gida, June 13 ; the other figures, Pinus ausiriaca, June 9-10, 189S. 107. First stage in the development of the sperm-nuclei. X 744- Pinus Strobus. June 9, 1898. 108. The sperm-nuclei just after the formation of the nuclear membrane showing early stages in the development of the daughter-reticula. The lower nucleus is already slightly larger than the upper one. X 744- Pinus monta7ia uncinata. May 31, 1898. 109-112. Various stages in the growth of the sperm-nuclei. A cell plate is sometimes apparent as in fig. 1 10, but no dividing wall is ever formed. X 744- Fig. 112. Pinus Strobus, June 10; fig. 109, P. resinosa, June 15; fig. no, P. ausiriaca, June 10. Fig. in represents another sec- tion through the upper nucleus of fig. no, and shows how the upper of the sperm-nuclei is frequently indented along its outer surface. 1898. ('74) Proc. Wash. Acad. Sci. Vol. Plate X. ^-/^^ 't^ ■^. Wf 4y^ •■■■':. ■ M. C. F., DEL. FERGUSON, -PINUS. HELIOTYPE CO.. BOSTON. PLATE XI. Fig. 113. A peculiar figure sometimes observed in the late telophase of the division. X 744- Pinus austriaca. June 10, 1S98. 114. A pollen-tube in which the smaller sperm-nucleus appears to be in ad- vance of the larger. This pollen-tube, having approached an egg that had already been fertilized, has turned aside and is passing up over the endosperm so that the normal position of the cells appears exactly reversed; «.c., neck-cells of the archegonium. X 289. Pifius Strobus. June 20, 1898. 115-116. Cross-sections through the two sperm-nuclei after they have attained full size and have about reached, in their downward passage, the middle of the nucellar cap. X 744- Pinus Sirobus. June 15, 1S98. 117. The sperm-cell after all traces of the spindle have disappeared, but before the two nuclei have come together. X 472- Pinus Strobus. June 13, 1898. 118. The same after both nuclei have come to lie in the upper part of the cell. X 472. Pinus Sirobus. June 10, 1S98. ( 176 ) Proc. Wash. Acad. Sci. Vol. Plate XI. /,-, 13 ■■^^i'-- ^\y^' v-^-? ,.;^-^'- '■'■^■^^i^^mi ' 17 ^ st.c FERGUSON -PINUS. HELIOTYPE CO., BOSTON. PLATE XII. Fig. 119. The lower portion of a pollen-tube which has penetrated about two- thirds the length of the nucellar cap. X 472- Pinus Strobus. June 14, 1898. 120. The lower portion of a pollen-tube which is just pushing between the neck-cells of the archegonium. p, pit in apex of tube. X 472. Pinus Strobus. June 20, 1898. 121. A vertical section of a joung cone; the ovuliferous scales have not as jet been organized. X 57- Pinus austriaca. March 14, 1898. 122. Section of an ovuliferous scale showing the first indication of an ovule, m. ovule; 0.5, ovuliferous scale; b, bract. X 150. Pinus Strobus. May 31, 189S. 123. A vertical section of an ovule one week later than that shown in fig. 122. X 150. Pinus Strobus. June 6, 1898. 124. A very young macrospore-mother-cell showing differentiation of spongy tissue. X 394- Pinus rigida. May 15, 1902. 125. The macrospore-mother-cell from fig. 124 more highly magnified. X810. 126. A macrospore-mother-cell just prior to synapsis. X 810. Pinus Strobus. June 27, 1898. (17S) Proc. Wash. Acad. Sci. Vol. f"LAI E All. W C F., DEL. 26 HELIOTYPE CO., BOSTON. FERGUSON, -PINUS. MACROSPOROGENESIS. PLATE XIII. Fig. 127. The macrospore-mother-cell in synapsis. X Sio. Pinus austriaca. June 6, 1898. 128. The same in recovery from synapsis showing continuous skein. X 810. Pinus austriaca. 129-133. Stages leading to the organization of the chromosomes in the first or heterotypical division of the macrospore-mother-cell. X 810. Fig. 132, Pinus Strobus, the others, P. rigida. Fig. 131 illustrates an instance in the unusually early disappearance of the nuclear membrane. 134-137. Stages in the establishment of the spindle in the first division of the macrospore-mother-cell. The reduced or one half number of chromosomes appear in this mitosis. The spindle arises as a multi- polar diarch. X Sio. Fig. 137, Pinus rigida, the others, P. Strobus. 13S. Late telophase in the first division. A cell-wall is laid down and defi- nite resting nuclei are formed. X 810. Pinus Strobus. June 13, 1S99. 139-140. The close of the heterotypical division. Resting nuclei are formed but the upper resting nucleus in each case shows signs of disintegration and doubtless would not have divided. X 810. Fig. 139, Pinus austriaca, fig. 140, P. rigida. 141. The two daughter-cells formed by the first division of the macrospore- mother-cell. Both would doubtless have divided again. X 810. Pinus austriaca. 142. The second or homotypic division of the macrospore-mother-cell. The spindles are oblique and arise as multipolar diarchs. The ciiro- mosomes have the same form as those which arose on the first division of the macrospore-mother-cell. X810. Pinus austriaca. (180) Proc. Wash. Acad. Sci. Vol. Plate Xlll. 127 fSf Mr 33 f C?V.-> 129 ^ 135 4 I ,-^ 13 7^ /^« 138 r \ Avi\ _ c .'^ "V 134 ).4 ,../ 136 140 . , / M. C. F., DEL. FERGUSON, -PINUS. MACROSPOROGENESIS. HELIOTYPE CO., BOSTON. PLATE XIV. Figs. 143-144. Two axial rows of three cells each. The upper of the two daugh- ter-cells formed as a result of the heterotypical division has not divided in either case ; a few starch grains in the cells of the axial row and many large ones in the spongy tissue as shown in fig. 143. X Sio. Fig. 143. Pinus Strobus, fig. 144, P. rigida. 145. An axial row of four cells, reconstructed from serial sections. X Sio. Pinus austriaca. 146. A macrospore nucleus surrounded by large starch grains. X Sio- Pinus austriaca. 147. Growth of the functional macrosnore ; the peripheral layer of cyto- plasm already established ; the three vipper cells of the axial row almost destroyed; one large cell of the spongy tissue shown. X^io. Pinus austriaca. June 13, 189S. 148. An axial row of three cells ; the functional macrospore much enlarged, and the two upper cells in an advanced stage of disintegration ; the spongy tissue distinctly differentiated ; the cells along its outer sur- face more or less tabular in outline and many of them badly disorgan- ized. Pathological conditions have just entered in as shown by the reduced amount of cytoplasm in the cells of the spongy tissue and the slight thickening of their walls. X 234. P. rigida. June 24, 1902. 149. The first division of the macrospore-nucleus. X 234. Pinus Strobus. July 29, 1898. 150. The karyokinetic figure from the above more highly magnified ; the division conforms to the typic type and shows the one-half number of chromosomes. X^io. 151. The first two nuclei of the female gametophyte. X234. Pinus aus- triaca. July 29, 1898. 152. The four-nucleated stage of the female gametophyte. X234. Pinus Strobus. August 4, 1808. 153. One of the sixteen free nuclei of a female gametophyte, all sixteen nuclei being in the spireme stage of division. XS^O- Pinus Strobus. October 12, 1898. 154. A vertical section of the central portion of an ovule showing the spongy tissue and the prothallium with its nuclei, of which there are sixteen, all in the equatorial stage of division ; the prothallium has been somewhat displaced by the action of the fixing fluid. X 46- Pinus Strobus. October 12, 1898. 155. One of the spindles from the above more highly magnified. X 744- ( 182 ) Proc. Wash. Acad. Sci. Vol. Plate XIV. M. C. F,, DEL. FERGUSON, -PINUS. GERMINATION OF MACROSPORE. HELIOTYPE CO., BOSTON. PLATE XV. Fig. 156. Surface view of a bit of the prothallium showing two free nuclei and the vacuolate protoplasm surrounding them. X 744- Pinus Strobus. May 17, 1898. 157. A radial section through the lower portion of an ovule showing pro- thallium, spongy tissue, and normal nucellar tissue. X472- Pinus Strobus. May 26, 1899. 158. As fig. 157, except that the spongy tissue and the normal nucellar tissue are separated by a double layer of cells, belonging to the nu- cellus, which have lost their protoplasmic content but their walls have not yet collapsed. X 472. Pinus Strobus. May 26, 1899. 159. A bit of the prothallium in surface view showing the complex cytoplas- mic figure characteristic of the late telophase in free nuclear division. X472. Pinus austriaca. May 17, 1898. 160. Surface view of a portion of a prothallium immediately after the or- ganization of cell-walls separating the free nuclei. X 472- Pinus Strobus. May 26, 1899. i6i. A bit of the prothallium as seen in radial section just after cell-walls have arisen. The cells are open on their inner surfaces and the nuclei remain near the open sides. X 394- Pinus austriaca. May 20, 1898. 162. A prothallium still open at the center showing that true "alveoli" as described by Sokolowa are not present; the archegonia rudiments at the micropylar end ; the spongy tissues still prominent. X 62. Pinus austriaca. May 24, 189S. 163. A condition often found in the ovule. The macrospore-mother-cell has failed to develop and the walls of the spongy tissue have thickened and stain deeply. X 46. Pinus Strobus. 164-166. Figures illustrating karyokinesis in the spongy tissue. The method is typic with the number of chromosomes characteristic of the sporophvte. X 810. Pinus Strobus. ( 184 ) Proc. Wash. acad. Sci. Vol. Plate XV. a -J ^^ ||. sS^j s.f..' 161 ^^^) w>v ^•^ 158 f<^ '^-«/:- 159 HELIOTYPE CO.. BOSTON. FERGUSON -PINUS. CFMAl P PRnXHAI I lUM. PLATE XVI. Pittus Styobus unless otherwise indicated. Fig. 167. Telophase in the division of a cell of the spongy tissue. X Sio- 168. The macrospore and some of the cells of the spongy tissue in the first stages of disintegration and having the appearance of a group of sporogenous cells. (»?«c.) macrospore. X 96. Pinus austriaca. 169-175. Stages in the early development of the archegonium. The central cell remains close beneath the neck cells. The cytoplasm is very vacuolate. X 140. Fig. 169, May 26, 1890; fig. 171, May 31, 189S. 176-179. Later stages in the growth of the archegonium. The vacuoles gradually disappear and many proteid vacuoles arise in the cytoplasm. X 62. Fig. 178 collected June 15, 1S99. 180. Mature archegonium. The nucleus has assumed a central position in the cell ; the ventral canal-cell is in an advanced stage of disintegration ; the proteid vacuoles are distributed about the periphery especially along the basal portion of the egg, the receptive vacuole has appeared but has not yet assumed its mature or final shape. X 62. June 17, 1899. 181. Nucleus of the central cell shortly before its division. This nucleus is almost invariably concave on the side towards the neck cells. X472. 182-184. Prophases in the division of the central cell. X 472- Fig. 1S4, Pi?itiS austriaca. (186) Proc. Wash. acao. Sci. Vol.. Plate XVI. AS 77 1^ •. 1$ "^ "T"' ■ , ^ •■-■ ■%%i^ 179 181 "^^^-Cevi- .^ 178 HELIOTYPE CO., BOSTON. PLATE XVII. Piuus Siyobtis unless otherwise indicated. Figs. 185-186. Later stages in the prophase of the division of the central cell. X472. Fig- 186, Pinus austriaca. 187-188. Disappearance of the nuclear membrane and establishment of the achromatic spindle. The spindle now lies wholly within the area pre- viously occupied by the nucleus. X 472- 189. Cross-section of the nucleus of the central cell just as the chromo- somes are undergoing longitudinal splitting at the equatorial plate. X 472. 190-197. Separation of the half chromosomes and formation of the daughter- nuclei. X 472- Figs. 192 and 195, Pinus austriaca. These figures show some of the variations occurring in the mitotic figure for this division, and the corresponding variations in the structure of the nucleus of the ventral canal-cell. Figs. 190, 191, 193 and 196 are very interesting, showing how some at least of those ventral canal-cells in which no definite nucleus is organized have arisen. Figs. 192, 194 and 195 are also interesting as leading to the formation of a normal nucleus within the ventral canal-cell. It will be noted that this spindle is always monopolar at its lower extremity and usually broadly multi- polar at the opposite end. Fig. 192 is the only instance observed of a sharply bipolar spindle. The q^s, nucleus is larger from the very first than the nucleus of tlie ventral canal-cell. (1S8) Proc. Wash. acad. Sci. Vol. Plate XVII. 185 0% 92 <;.»^ ';;\^ ^K^J^^Vfe 190 193 li=:«.M; 187 -fi^^'?:-. ^ 94 196 '-% ^^--•v»»v,,>?v'-< PLATE XVIII. Pinus Strobus. Figs. 198-199. Some of the aspects presented by the ventral canal-cell. It is doubtful in both of these cases if any nucleus has ever been organized within the ventral canal-cell, and the chromosomes have not even fused to form a spireme. X 472- 200-202. Later history of the ventral canal-cell and early stages in the development of the egg-nucleus. The first indication of the primary nucleolus is seen on the lower side of the egg-nucleus in fig. 202, and the ventral canal-cell already shows marked signs of disintegration. X472. 203-204. Later stages in the downward movement and growth of the egg- nucleus showing growth of primary nucleolus. X 472. 305. Mature egg-nucleus. The primary nucleolus is very large and vacuo- late and several secondary nucleoli are scattered throughout the nu- cleus. The structure of this nucleus varies greatly. This one was selected not because it can be said to be any more typical than others, but because it represents an average rather than an extreme condition as to density of reticulum and number of secondary nucleoli. X 472- ( 190) Proc. Wash. Acad. Sci. Vol. Plate XVIII. >? r. 203 20? .^^; v. 204 M. C. f., DEL. HEUOTYPE CO., BOSTON. FERGUSON.-PINUS. PLATE XIX. Pinus Strobus unless otherwise indicated. Fig. 206, a-g. Portions of the reticulum from different mature egg-nuclei, showing some of the variations which may occur in the structure of this nucleus. X 1050. 207. Division of the central cell, showing also the lower portion of a pollen- tube which has already reached the endosperm. In this instance a very short time would have elapsed between the division of the central cell and fertilization. X 209. Pinus montatia uficinafa. 208. The primary nucleolus from a mature egg-nucleus with secondary nucleoli clustered about it and evidently formed by it. The primary nucleolus has a great affinity for stains at this time. X 1050- 209. The primary nucleolus of a mature egg-nucleus. This nucleolus shows a weak reaction towards dyes, and apparently has an outer, limiting membrane. X 1050. 210. The framework of a primary nucleolus from a mature egg-nucleus. This nucleolus has remained of a light greenish-yellow color after treatment with Flemming's triple stain. X 1050. 211. The upper part of an archegonium showing cavity, the receptive vacuole, formed in the cytoplasm just prior to fertilization. X i40' 212. The upper part of an archegonium just after the entrance into the eg^ of the elements from the pollen-tube. X 140- 213. A slightly later stage. The cytoplasm of the sperm-cell has already fused with the cytoplasm of the egg. X 140- 214. An entire archegonium showing the sexual nuclei in contact, and, above them, the various elements which have come into the egg from the pollen-tube. X 62. June 21, 1S9S. 215. The upper part of an archegonium in the same stage as the above. X 140- ( 192 ) Proc. Wash. Acad. Sci. Vol. Plate XIX. #'" w # . V^^"'-t •■>^v.- 206a 206b •" 206c «i> 206e 206f ^O .^r^ji^P ^ 206g ^ ■s: 21 I V (- /g^0 '^ 212 ' ©,'1 ®^«;r; ~\ •5' ' 'ttv: «> ^■^, 207 209 208 'j% £^ 210 ^ I, ••^'^ -■Q.._ .-' c.p.t. r~ 213 ■5^^^ ^1 c.p.t. sf.c. f.n. ... ^ -i/ n.». ^ ' ».c. t.n. ^ § 228 -f;>^^Vi. - ^' ^ ^ 226 .;>'i^^v5^;^V 1?^^':^^'; ".^r^ 229 227 HEUOTYPE CO., BOSTON. M. C. F., DEL. FERGUSON-PINUS. PLATE XXI. Pinus Strobus. Fig. 230. The spindle fibers have become more abundant and transverse segmen- tation of the spiremes has occurred at some points. X 472. 231. The spindle fully established having now assumed the form of a multipolar diarch ; the two chromatic spiremes still perfectly distinct. X472- 232. The two spiremes after segmentation; the two halves of the spindle seem to indicate the maternal and the paternal portions of the mitotic figure. X 472- 233. Early stage in the formation of the chromosomes. The chromatic elements still occur in two distinct groups, but position, alone, deter- mines which are maternal and which are paternal. The segments can not be structurally differentiated. X 472. 234. The chromosomes being oriented at the nuclear plate. The distinc- tion between paternal and maternal elements no longer evident. X472- 235. A cross-section through the nuclear plate just before the separation of the chromosomes ; twenty-four segments are distinctly shown. X472. 236-23S. Some of the aspects presented by this mitotic figure during meta- kinesis. X472. 239. An anaphase of the mitosis. X472. 240. A late anaphase of the division ; the poles terminate in granular areas from which delicate threads extend into the cytoplasm ; some of the nucleolar substance from the egg-nucleus still persists. X472. 241. One end of the spindle in the same stage as the above ; the fibers which radiate from the polar region of the spindle are very abundant and stain deeply. X472. (196) Proc. Wash. Acad. Sci. Vol. .=j^fef 230 ti'f^Sf^lilS^P^ Plate XXI 237 * r^.v;.-'V'>.»ViiV'0:J ^fj^;^l^v'i'^iv^v\'i^ 232 231 '^^O ^C 233 238 i 234 235 \ k.\^ •f'5' ■'t:';. «'"^-' ^ ^ 236 239* ' «-<•>.. >i'"^ 241 240 C? O ^ ^ # HEtlOTYPE CO., BOSTON. PLATE XXII. Pinus Strobus. Fig. 242. One aspect presented by the karyokinetic figure in the telophase of this division. X 472- 243. The two segmentation-nuclei fully formed. X472. 244. One of the two segmentation-nuclei in an early prophase of divi- sion. X 472. 245-2463. Later stages in the second division, showing two chromatic spiremes. X472. 247. A still later stage. The two groups of chromosomes can still be made out. X472. 248. An entire archegonium showing the position of the two segmenta- tion-nuclei during division. The receptive vacuole has been distorted by the entrance of the contents of the pollen-tube. X 62. 249. An archegonium showing the original position of the four segmenta- tion-nuclei. X62. 250<7. The same after the nuclei have begun their downward movement. X62. 250^. A nucleus from 250a showing details of its structure and fibers in the surrounding cytoplasm. X 472. 251a. An archegonium after the nuclei have almost reached the base of the oosphere. X62. 2513. A portion of fig. 251^, showing details in nuclear structure, and fibers in the surrounding cytoplasm. X472. (198) Proc. Wash. Acad. Sci. Vol. Plate XXII. .»;y.r "W oo*r 'v;--W"'^* ■ 246a 242 ^,<.^:i^^^l'^ifi 243 S^ 246b 250a *^«'"' ■%s 247 ,^. *..i m 0 c 6 249 ?^. -<.t;i C|^y ii ^'^'^ a ^ -'Pi y. O r. 248 .*f^^^^^:: j'w^. ',. ^-^^ ^;^'.^' vi,":^ 'ii;> :?-!^ 250b M C F., DEL. '■-.-W ■/ 251a FERGUSON -PiNUS. DEVELOPMENT OF PROEMBRYO. 251b HELIOTYPE CO., BOSTON. PLATE XXIII. Pinus Strobus unless otherwise indicated. Fig. 252a. The lower part of an archegonium after the four nuclei have ar- ranged themselves at the "organic apex" of the oosphere. X62. 252^. A portion of the above ; the nucleus is in the early prophase of division ; the cytoplasm surrounding the nucleus has became dense and deeply staining. X472. 253a. The basal portion of an egg ; the four segmentation-nuclei are in the metaphase of the mitosis. X62. June 19, 1899. 2533. A part of the same showing details. X 472- 254a. A portion of a lower part of an oosphere after the formation of the eight nuclei of the proembryo. X 62. 2541^. A part of the above giving details. No cell-walls have as yet been formed, but there is a slight differentiation of the cytoplasm about each nucleus. X472. 255a. A somewhat later stage than fig. 254a. X 62. 2553. An enlarged portion of the above, showing cell-walls in the process of formation. X 472. 256. Vertical section through the base of an archegonium showing that the four nuclei of the upper tier of cells in the proembryo divide before any divisions occur in the four lower cells. X 96- Pinus austriaca. 257-25S. Figures occurring in the upper part of archegonia during the division of the segmentation-nuclei. These doubtless represent the smaller sperm-nucleus. X472. 259^-2593. Figures occurring in the upper part of an archegonium at the time of the second division following fertilization ; fig. 259^7 represents the tube-nucleus; the karyokinetic structure in fig. 259^, is the smaller sperm-nucleus, and just above it the stalk-cell is still distinctly visible. X472. 260. Two macrospore-mother-cells. X S30. Pinus rigida. June 7, 1902. 261. An axial row showing oblique wall between two of the spores. X 394- Pinus austriaca. June 13, 189S. 262. a. A section through a prothallium showing unusual origin of archegonia from cells several layers deep in the prothallium. X 75 ( 200 ) Proc. Wash. Acad. Sci. Vol. ^=ir.>« 252b 252a 254a 253a h^LAI t AAll \-- 1 ^^ / m ^.^*J7 253b fr^SsS-- „. V4- 255a 254b ' -^*%— "- - > I! 259b 256 M C F., DEL. HELIOTYPE CO., BOSTON. FERGUSON -PINUS. DEVELOPMENT OF PROEMBRYO. Proc. Wash. Acad. Sci., Sep., 19O4. PLATE XXIV. Fig. 262^. Another section through the same prothallium as that shown in 262/^i.>afV^^ 269 270 \ %. ^. 265 %■ ■ y *! "^ •<^'- 1 266 ^- ^i;.,. .'^^WW -t:^-^ \- 274 "^^^^272 HELIOTYPE CO., BOSTON. M C F.. DEL FERGUSON -PINUS. ABNORMALITIES. PROCEEDINQS OF THE WASHINGTON ACADEMY OF SCIENCES Vol. VI, pp. 203-332. [Figs. 1-81.] December 14, 1904. STUDIES OF VARIATION IN INSECTS. By Veknon L. Kellogg and Ruby G. Bell, OF Leland Stanford Junior University. CONTENTS. Introduction. Data of variation in t~Menty-four insect species. Variation in the venation and costal wing hooks in Apis mellifica (the honej bee), drones (males) and workers (females) 212 Variation in the venation of costal wing hooks of male black ants .... 244 Variation in the venation of Culex iiicidcns and Grabhamia curriei (mos- quitoes), males and females 253 Variation in the elvtral pattern and prothoracic pattern of Hippodamia couvergcns (the convergent lady-bird) 2^7 Variation in the elj'tral pattern of Diabrotica soror (the California Hower- eating Diabrotica^ . . 273 Variation in the abdominal pattern and face spots of Vespa sp. (yellow- jacket) . . . . 282 Variation in the prothoracic pattern of Tetiigonia sp. (leaf hopper) 287 Variation in the prothoracic pattern of a flower-bug 291 Variation in the prothoracic pattern of Corisa sp. (water-boatman) . . 293 Variation in number of eye-spots on the wings of Parnassius smintheus and Cwnonymp/ia galactinus [hutter^its) 295 Variation (regenerative) in number of tarsal segments of Periplaneta «;«, complete cells at i and 3 ; £ and /^, complete cell formed by meeting of s])urs I. 3- Fig. 9. Fore and hind wings of honey bee (worker) showing normal vena- tion. A lot of 300 workers from the same hive from which the above lot of drones was taken was examined for variation in the character of the venation of the wings. Of the 300 left fore wings : 288 show no variation (fig. 9) ; 2 show var. i incomplete ; 2l8 KELLOGG AND BELL I shows var. i complete; 8 show var. 3 "very slight" or "slight"; I shows a complete new cell interpolated. Of the 300 right fore wings : 293 show no variation (fig. 9) ; I shows var. i complete; 5 show var. 3 "very slight" or " slight" ; I shows var. 3 " fair." Of the 300 left hind wings, but one shows any variation, that being a case of " var. 2 slight." There is manifest a tendency toward the appearance of spurs indicated by the frequent angu- lation of the veins at points where the spurs occur in the drones, but that is the limit of the variation. Of the 300 right hind wings also but one shows a variation of the degree called slight that being also a case of "var. 2 slight." The same angula- tion of the veins is to be noted as in the left wings. Another lot of drones (48) and workers (300) from another hive (Italians) was examined for the same conditions of varia- tion in venation of the hind wings. Among the 300 workers no describable variation was found, simply the slight although manifest tendency to have the spur-bearing (in the drones) veins angulated thus affording traces of spurs. Among the 48 drones, on the contrary, the various spurs recorded for the pre- vious lot from laboratory hives were present in various degrees of length, and five cases of the interpolation of complete new cells were noted (see fig. 8). Another lot of 100 workers from still another hive (Germans) was examined, and no variation in the character of the vena- tion of the hind wings was found. Thus comparing the variation conditions of the wing vena- tion in drones and workers, it is to be said that practically no variation exists among the workers, while a frequent and in a few cases extreme variation (interpolation of new cells) occurs among the drones. The drones are males, and are -partheno- genetically produced; the workers are females (of arrested sex- ual development) and have bisexual parentage. To ascertain what difference, if any, existed in the amount of variation (in venation of wings) between bees exposed to the struggle for existence and bees not yet so exposed, a lot of 200 drones from a hive (Italians) in a garden on the Stanford Uni- versity campus taken from the capped brood cells at a time STUDIES OF VARIATION IN INSECTS 219 when they (the drones) were just at point of issuance (some half issued from the cells, some gnawing at the caps, some just ready to begin breaking out) was examined. Of the 200 left fore wings : 94 show no variation ; 42 show spur 2, 3 or 4 m " very slight," " slight," " fair," " good," or '* marked " degree, or two or three of these in combination ; 32 show var. i complete ; i shows var. i incomplete ; i shows an almost complete new cell ; 2 show a complete new cell ; 16 show a break (interruption) in some vein (usually the vein carrying spur 4) ; II show a marked deformation of the venation (the veins in the discal area broken into small pieces and all askew) to such an extent as to be really a mutilation of the wing (fig. 10). Fig. 10. Fore wings of honey bee (drones) showing extraordinary deforma- tion of venation. Of the 200 right fore wings : 95 show no variation ; 56 show spur 2, 3 or 4 in " very slight," " slight," " fair," " good " or " marked " degree or two or three of these in combination ; 23 show var. i complete ; i shows an almost complete new cell ; 15 show a break (interruption) in some vein; 12 show a marked deformation of the venation (the veins of the discal area being broken into small pieces and all askew) (fig. 10). Of the 200 left hind wings: 139 show no or only "very slight " variation ; 28 show at least one (sometimes more) of vars. I, 2, 3 or 4 in condition called " slight " ; 28 show at least 2 20 KELLOGG AND BELL one (sometimes more) of vars. i, 2, 3 or 4 in condition called " fair," " good," or " marked"; i shows an almost complete new cell ; 6 show a complete new cell ; i shows a break (inter- ruption) in a vein. Of the 200 right hind wings: 134 show no or only " very slight " variation ; 30 show at least one (sometimes more) of vars. I, 2, 3 or 4 in condition called " slight " ; 28 show at least one (sometimes more) of vars. i, 2, 3 or 4 in condition called "fair," " good " or " marked " ; i shows an almost complete new cell ; 6 show a complete new cell ; i shows a deformation (only in less degree) like that shown in certain of the fore wings. Also a lot of 54 workers just ready to leave their cells, were taken from the brood comb of a laboratory hive and examined for variation in the venation of the wings. There was found in no wing, fore or hind, right or left, a variation of the degree called " slight" (as used in discussing the previous lots), al- though a manifest tendency toward the appearance of a spur at point 2, reaching sometimes the condition of trace or " very slight "was apparent. Another lot of 25 workers from the cells and 50 workers just issued from cells and acting as nurses (not yet having left the hive at all) taken from a hive (Italians) near Stanford Univer- sity was examined for variation in the venation of the wings. In neither fore nor hind wings was any variation of the extent called " slight." Thus it is apparent that it is not the action of a life and death selection which accounts for the differences in the amount of variation in drones and workers. As the workers in their con- stant going and coming outside the hive, carrying heavy loads of pollen and exposed to any dangers which slow or imperfect flight might induce, as capture by birds and robber-flies, maybe fairly said to run much more risk in their life than the drones which make but a single brief flight each day (and that not every day), it might be thought or assumed that this strenuous life of the workers would tend to weed out by life and death selection every slight disadvantageous variation in the supporting skele- ton (the venation) of the wings, all important organs in this out- side life. But our workers from cells (not. yet exposed to selec- STUDIES OF VARIATION IN INSECTS 221 tion on basis of adult structures) show no more variation in wing venation than those which have been so exposed. It is also to be noted that our drones from cells show no more variation than the free flying ones except in one particular. That is in the presence among them, in the proportion of ii to 200, of certain individuals whose front wings have a wholly abnormal venation, of the condition of a sort of mutilation or monstrosity of such a character as to prevent the full unfolding of the wing and probably its use at all as a flight organ. As the fore wings are the chief factors in the flight of bees, the hind wings being attached to them by hooks, in large degree moved by them and obviously auxiliary and subordinate to them, this variation prob- ably produces a crippled individual quickly eradicated by nature or perhaps purposely by the keen-eyed and thrifty workers. At any rate we have never found such a crippled-winged drone in any lot of individuals taken in the " drone catcher" (which is a device fastened over the external opening of the hive and which snares any drone that voluntarily attempts to issue from the hive). For the rest of the variation the free-flying drones show practically no less, and of course no more, modification in venation than the ones taken from cells. Practically, thus, the same amount of variation exists in the venation of the wings among free-flying drones and workers as exists among individuals which have just acquired their imagi- nal structure, which seems to mean nothing less than that this variation as manifest, as considerable and in a few cases as extreme as it is, and as common (/. e., occurring in numerous individuals) as it is, is not sufficient to be of life and death value in the struggle for existence. In the exposure of the bees to the rigor of the factors which determine natural selection this varia- tion in the skeletal framework of the wings, organs obviously unusually important in the relation of the bees to the outside world, does not afford a handle for the selective action of these factors. It is of interest to note in passing the large importance at- tached by entomologists to the venation characters in the sys- tematic study of insects, especially those in which the venation, as in the Diptera, Lepidoptera, and Hymenoptera (to which 222 KELLOGG AND BELL order the bee belongs) has been specialized by reduction, /. c, where the veins have been reduced in number and degree of branching and inter-connection by cross-veins, so that the remaining vein framework is presumably in all of its details essential to the best performance of the wing's function, namely flight. This reliance on venation is not based on such theoreti- cal grounds however, but on the practical one of experience and wide observation. In many of these specialized insects the venation is fairly uniform throughout a whole family, while practically never are describable differences in venation ex- pected to be found within generic limits. In Comstock's Man- ual of Insects (the standard American systematic manual) the keys to the families of Diptera and Lepidoptera are nearly solely, and of Hymenoptera, largely based on venation. Thus variation in venation is to be looked on as important. Measurements of Parts of Wing-veins. — Wishing to be able to give an accurate quantitative expression to some features of the variation of the wing venation, we have measured certain parts of veins whose limits are accurately established by sub- tending cross-veins, or by the forking or branching of the veins themselves. In order to determine the relation of any variation in these measurements to the varying size of the whole wing this size, as indicated by the measured length and breadth (or when impossible to get the length owing to battered and broken wing tips by the width alone), has been ascertained for all of Fig. II. Fore wing of honeybee (drone) with letters indicating subtending points on various veins. the wings studied. While at first thought the dimensions of parts of the veins might be assumed to be directly related to, i. c, a simple function of, the size of the wings, yet an inspec- tion of the parts selected for measurement will suggest the pos- STUDIES OF VARIATION IN INSECTS 223 sibility and even probability (a condition actually confirmed by the statistical study) of a variation in these dimensions due lo the var3nng situation in the \vin (I, 2), (5, 4. 8), (6, 7). 2. L. broken. i5- L. (9, 3), 2, i, 8, 7, 6, 5, 4. R. 9, 3, 2, I, (4, 8), (5, 6, 7). R. 9> 3, 2, (i, 7, 8), (4, 5- 6). 3. L. 9, 3, 2, (I, 5, 8), 4, (6, 7). 16. L. 9, 3, 2, 1/(5, 8), 6, (4, 7). R. 9, 3, (I, 2), 8, (4, 5, 6, 7). R. 9, 3> (i^ 2, 8), (4, 5, 6, 7). 4. L. (8, 3), (i, 2, 7), (4, 5, 6). 17. L. broken. R. (8, 3), 2, (I, 7), 6, (4, 5). R. (9. 3), 2, (I, 8), (4, 5, 7). 5. L. 9, 3, 2, I, (5, 6, 7, 8), 4. 18. L. broken. R. 3, 9, 2, I, (7, 8), 5, 4, 6. R. broken. 6. L. (9, 3), (I, 2), S, (4, 5, 6, 7). 19- L. (9, 3), 2, I, 8, 4, 7, (6, 8). R. 3, 9, 2, (i, 5, 8), 4, 6, 7. I R. broken. 7. L. 9, 3, 2, I, (4, 7, 8), (5, 6). 20. L. 9, 3, 2, I, (7, 8), (4, 5, 6). R. 9, 3, 2, 1, 4, (5, 6, 7, 8). R. 3. 9. 2, I, 8, 7, (4, 6), 5. 8. L. broken. 21. L. 9, 3. 2, 8, (i, 4, 5, 7), 6. R. (9, 3), (I, 2), 8, (4, 5, 6, 7)- R- 9. 3. 2, (5, I, 8), (4, 7), 6. 9. L. broken. 22. L. 3, 8, i, 4, 2, 7, (5, 6). R. (8, 3), I, 5, 2, (4, 7). R. 3' 8, 4, (I, 2), (5, 6, 7). 10. L. 3, 9, (I, 2), 5, (4. 7> 8), 6. 23. L. 9, 3. 6, (i, 7- 8), 6, (4, 5). R. 8, 3, 4, (I, 2), 7, (5, 6). R. 9, 3. 2, I, (6, 7. 8), (4, 5). 11. L. (9, 3), (i, 2), (7, 8), (4, 5,6). 24. L. broken. R. 9, 3, 2, 5, 8, (4, 5, 6, 7). R. broken. 12. L. 9, 3, (I, 2), 8, 7, (4, 5, 6). 25. L. 3, 9, (I. 2), (4, 5. 6, 7, 8). R. 9, 3, I, (2, 8), (5, 7), (4. 6). R. 9. (2, 3), I, (4. 5, 8), 7. 6. 13. L. 8, 3, (2, 4), I, (6, 7). 5. R. (8, 3),4, 2, 7, (1,5. 6). In actual lengths the antennal segments varied as follows, (only the range given) : Seg. i, 4-7 ; seg. 2, 4-8 ; seg. 3, 6-1 1 ; seg. 4, 3-8; seg. 5, 3-5 ; seg. 6, 3-4 >^ ; seg. 7, 3-5 >< ; seg. STUDIES OF VARIATION IN INSECTS 313 8, 4-9; seg. 9, 7-9. The length numbers refer to units of the micrometer scale, not reduced to millimeter fractions. It is surely obvious, as has been mentioned before in connec- tion with the account of the variations in the number of tibial spines of Mclanoffhis fciniir-7'ubrii.m (p. 34) that systematic students of zoology must take into account the conditions of variation exhibited by the characters they choose for use in specific diagnoses. Students of variation may never supply biology with " a precise criterion of species " but they can very promptly supply the systematist with a precise criterion of the stable value of any specific character. Variation (Absent) in Number of Antennal Segments in Eleodes sp. and Vespa sp. — In a lot of 123 adult individuals of Eleodes sp., a darkling ground beetle, collected at various times and places near Stanford University, no variation in the number of segments composing the antennae was found. The normal number is eleven. The insect has a complete meta- morphosis. In a lot of 55 adult individuals of Ves^a sp., yellow-jacket, collected on one day in one place (feeding on refuse), on the Stanford University campus, no variation in the number of seg- ments composing the antennas was found. The normal num- ber of antennal segments is thirteen. The insect has a com- plete metamorphosis. The same individuals show considerable variation in pattern of abdominal markings (see p. 284). Variation in Number of Long Tactile Hairs on the Meta- thorax of Lipeurus celer and L. varius (Biting Bird Lice). — The biting bird lice, or Mallophaga, are external wingless parasites on birds and mammals, living their whole life on the body of their host or in some cases on the bodies of two or more hosts, migrating from parent to young in nesting time, from one sex to another in mating time, or from one companion to another in the case of crowding, gregarious species. Their migrations however, are practically limited to passing directly from one host to another when the hosts are in actual contact and the majority of parasitic individuals undoubtedly spend the whole life from egg to death on one host. Thus there exists a pro- nounced isolation of groups of individuals on individual hosts 3H KELLOGG AND BELL and necessarily much in-and-in breeding. It is a condition analogous to island life carried to an extreme of isolation of small groups of individuals of the same species. The meta- morphosis of the Mallophaga is incomplete. On the latero-posterior angles of the dorsum of the metathorax of all species of the genus Lifeurtis, parasitic on birds, are certain very long, spine-like hairs (fig. 80), probably tactile in function, whose number and arrangement vary in different species, but are presumably constant for any given species. The number of these hairs is used to some extent in distinguish- FiG. 80. The biting bird louse, Lipetinis ccler, and diagram of metathorax (enlarged) showing tactile hairs in lateral posterior angles. ing species of the genus. We have examined the variation in number of these hairs in 239 individuals of Li^eu7'us celer KelL, a parasite of the fulmars, Fulniarus glacialis vars. rodgcrsi and glupischka. These individuals were taken from 31 Fulniarus glacialis hosts, shot in the Bay of Monterey, California, in a time period of two weeks. The range in number of the tactile hairs is from 3 to 6, with 4 as the mode, the mode occurring in a large majority of cases. STUDIES OF VARIATION IN INSECTS 315 Out of the 239 individuals the hairs could be accurately deter- mined on the right side in 189 cases, on the left in 193, the fre- quencies being as follows : On the right side : 2 individuals have 3 hairs, 177 individuals have 4 hairs, 10 individuals have 5 hairs. On the left side : 3 individuals have 3 hairs, 185 individuals have 4 hairs, 4 individuals have 5 hairs, i individual has 6 hairs. The mean on the right side is 4.04, on the left side 4.01 ; the standard deviation on right side .248, on left side .238 ; the coefficient of variation on right side is 6.15, on left side 5.92. The few cases of variation from the mode, 4 hairs, were scattered in various groups of individuals, letting all the para- sites from an individual host constitute a group. That is, the variation was not associated with isolation. From one host were taken ten parasites of which one had 5 hairs on one side, one 3 hairs on one side and all the others 4 on each side ; from another host were taken twenty-six parasites, the number of hairs on each side of all the individuals being 4 except one case of 5 on one side, and so on. The twenty cases of varia- tion from the mode were distributed in ten groups of parasites, i. e.^ occurred in parasites from ten out of the thirty-one host individuals. The correlation conditions (degree of bilateral symmetry) of the variations are shown by the following summary (only 183 specimens out of the 239 had the hairs so intact that the exact number of hairs on both sides of the metathorax could be determined) : Of 183 individuals, 166 have 4 hairs (the mode) on each side, the remaining having hairs as follows : nine have 4 on one side and 5 on the other, five have 4 on one side and 3 on the other, two have 5 on each side, one has 4 on one side and 6 on the other, none has the combination of 3 and 5. In 72 specimens of Lipeiirus variiis taken from 29 specimens of Fulmar us glacialis vars. rodgersi diX\6. glupischka (Bay of Mon- terey, time period, two weeks), the range of variation in number of metathoracic tactile hairs is 3-5 with 4 as a very dominant mode, the frequencies being as follows : 3l6 KELLOGG AND BELL On the right side : 13 individuals have 3 hairs, 49 individuals have 4 hairs, 2 individuals have 5 hairs. On the left side : 14 individuals have 3 hairs, 48 individuals have 4 hairs, 2 individuals have 5 hairs. The mean on right side is 3.83, on left side 3.81 ; the standard deviation on right side .453, on left side .463 ; the coefficient of variation on right side is 11.98, on left side 12.16. From these figures it is seen that in this species of Li;peii7'iis the variates possessing three hairs greatly outnumber the vari- ates possessing five hairs. This might be interpreted to mean that the species is tending toward a form possessing three tac- tile hairs instead of four, or alternately that the present form with four hairs is developing (evolving) from an older form with three. The coefficients of variation for this character are much larger (twice as large) in this species than in L. ccler. Examining the distribution of the variations by isolated groups of individuals (each group from a single bird host) the thirtv-one variations are distributed among eleven groups out of the total twenty-nine groups. In no group in which there was a case of the occurrence of 5 hairs, was there an occurrence of 3 hairs (the largest group in which 5 hairs occurred contains only five individuals). In a group of twelve individuals three have 3 hairs on each side, and a fourth has 3 hairs on one side and 4 on the other, while the remaining eight have 4 hairs on each side ; in a group of ten individuals three have 3 hairs on each side, a fourth has 3 hairs on one side and 4 on the other, while the others have 4 hairs on each side. With regard to correlation between right and left side, forty four individuals of sixty two (in which the hairs on both sides could be accurately determined) have 4 hairs on each side ; twelve have 3 on each side, two have 4 on one side and 5 on the other, and one has 5 on each side. Variation in the Character of the Elytral Striae of Pterosti- chus sp. (Predaceous Ground Beetle). — The ground beetles of the genus PtcrostlcJuis have about ten fine longitudinal grooves on each elytron (Fig. 81). These grooves present manifold variation in their make-up, an}^ one appearing as a continuous line, as a broken line, as a forking or branching line, or coa- STUDIES OF VARIATION IN INSECTS 317 lescing with its left or right-hand neighbor. In a lot of 149 individuals collected under stones on a hillside near San Jose, on April 4, 1903, a number of classes were established on the basis of the character of the tw^o lines, one on each elytron, lying on either side of the median groove which indicates the suture of fusion of the inner margins of the two elytra. These lines are referred to as «' and (f and either may have a short branch called rt"' or a''^\ or simply «'* when referred to for both sides (fig. 81.) The beetles are insects of complete metamorphosis and their imaginal structural characters appear at once, on the issuance c^M^MA Fig. Si. The predaceous ground beetle, Pterostichus sp., and diagram of elytron (much enlarged), showing striae. of the imago (after the brief necessary expanding and drying of wings, appendages and body-wall) in definite and fixed condi- tion. Variations in these elytral striae are to be looked on as strictly congenital in character. The classes and their frequencies based on the variations in the character of lines a' and «"" in the lot of 149 individuals are as follows : Class A : 102 individuals have lines «' and a" separate on both elytra. Class B : 4 individuals have line d' continued to line a on both elytra. Proc. Wash. Acad. Sci., Dec, 1904. 3l8 KELLOGG AND BELL Class C : 5 individuals have line a broken off and a)' con- tinued on both elytra. Class D : 4 individuals have line «^ separate and line a broken, then continued below on both elytra. Class E : 6 individuals have line a''' continued to line a" and line «'* separate. Class F : 8 individuals have line a'^ continued to line «' and line a" separate. Class G : 7 individuals have line «"■* separate, line a^^ sepa- rate and line a^ broken, then continued below. Class H : 2 individuals have line a'''' separate, line a*" broken, then continued below and line a'^ separate. Class 1 : 3 individuals have line a'' broken, and «'"'' continued below, line «'* separate. Class J : 3 individuals have line rt*"* separate, line a'' broken, then continued below, line a"' continued to line «'. Class K : i individual has line a"^ continued and a cross line between «'"'' and a% a"' separate. Class L : 2 individuals have line a!' separate in both elytra. Class M : i individual has line a''^ separate at first, line a'' broken, then continued below, its end connected with a line below line a*"'', and line a"' separate. Class N : 2 individuals have line a''' separate, line a"' con- tinued below and connected by a cross line with a^ In the above series the classes have been established on the basis of one pair of lines or grooves. An attempt was made to separate the lot of individuals into classes based on the varia- tion manifest in all the lines, with the result that no tzuo indi- viduals could be put into the same class. The variation in these lines apparently exhausts the possibilities of combinations of various conditions in the members of the line series. That these minute, although to the trained and microscope-aided eye distinct, variations can be of life and death selective pattern- value (and any other function for these lines than the making of pattern is not apparent) seems inconceivable. But in numer- ous cases the presence or absence, and even the arrangement and character of these fine elytral striaj are characteristics diagnostically used by systematists in their keys to genera of STUDIES OF VARIATION IN INSECTS 319 beetles of the family Carabidas (to which family Pterostichus belongs). In the family Dytiscidas (predaceous water-beetles) the elytral striae are also used as classificatory characters. A curious female dimorphism exists in some Dytiscid species in which one form of female has the elytra deeply grooved, while the other form has smooth elytra. The significance of this dimorphism is not known. GENERAL RESULTS AND SIGNIFICANCE. Blastogenic and Acquired Variations. — The importance of distinguishing, in any study of variation considered as a factor in species-forming, between those variations in the animals under observation which are truly blastogenic and those which are, in part at least, acquired by reaction to some causative influence from without during the immature life (development) of the individuals is obvious. If acquired characters are non- heritable, then, while the rigor of selection among adults may and will take into account any variations, either blastogenic or acquired which may exist in these adults, the selected survivors will, in fact, tend to transmit, and thus retain in the species, only those variations which are blastogenic. If acquired char- acters are heritable then the importance, perhaps, although not the interest of such a distinction between the two categories of variation may be lessened. But undoubtedly a majority of working naturalists believe that the inheritance of acquired characters is yet unproved. With the importance of this distinction well in mind — indeed with the belief that variation study without this distinction in mind has not much claim to attention from biologists intent on discerning the factors in a method of evolution — we have tried to point out in this paper in the case of each insect or charac- teristic studied, the character, blastogenic or acquired, which the variations discussed possess. For example, the variations in the pattern of Diabrotica,^ Hippodamia and Vespa (insects of complete matamorphosis with all adult external structures • By reference to the table of contents the position in this paper of the par- ticular discussion of any variation referred to in this general part of the paper mav be found. 320 KELLOGG AND BELL never exposed in definitive unchangeable condition to outside influences) are blastogenic variations, as are also the structural variations in the character of the venation and the number of costal hooks in A^is and in the black ant that was studied but the variations in the pattern of the pronotum of Tettigonia^ CorisUy the capsid and flower bug, and the number of tibial spines in Cicada and Mclanopltis^ may be in part acquired, for in these latter cases the insects are exposed during their imma- ture life (development), with these color and structural charac- ters in formative condition, and to some extent in use, to the continuous influence of their environment. Given a criterion (either in the character of the variation, or the variable character itself, or in the life history of the animal showing the variations) that will enable one to distinguish be- tween strictly blastogenic variations and those which may be wholly or in part acquired — and we believe we have such an one and a particularly valid one, in the case of the insects with complete metamorphosis — and this criterion may be applied not only to an individual but to an hereditary series of individ- uals and the larger question — that Sh'cilfrage of modern biol- ogy— as to the transmission of acquired variations be approached through breeding. By comparison of series of generations whose immature life has been exposed to various (experimen- tally controlled and quantitatively determined) conditions of life, with series of generations whose immature life is exposed to only a single rigidly controlled set of life conditions may be determined not only the effect of varying conditions on the pro- duction of variations, but the heritability of these variations. The most satisfactory answer to the question of the hereditary transmission of acquired characters will come as the result of a quantitative (statistical) study of variations known to be blas- togenic compared with a similar study of variations known to be acquired, both studies to be made on complete series of in- dividuals bred under quantitatively determined life conditions. Such studies are certainly not impossible, with a criterion for the distinction between blastogenic and acquired variation once obtained. The character of this criterion, for insects, exists by reason of the effective difference in the course of the life his- STUDIES OF VARIATION IN INSECTS 32 1 tory as regards insects of complete metamorphosis, and has been ah-eady explained in the introduction. Continuous and Discontinttous Variation. — By continuous variations we mean to refer to those variations variously called fluctuating, individual, etc., which are present in any series of individuals of a species, and which cluster about the modal or most abundantly represented form of the species as would be expected from the law of error (law of probabilities). Although the extremes (at either end of the range) among these varia- tions may differ considerably, they are so connected with the mode, by such a nearly perfect series of gradatory or intermedi- ate steps, that a curve or polygon graphically expressing their frequency and range will usually (where the number of indi- viduals in the series examined is large enough to exhibit the actual conditions of variation in the species) correspond closely to the theoretical curve which may be plotted for the species on the basis of the law of error. Although Morgan (Evolution and Adaptation) objects to the use of ''continuous" as a de- scriptive name for these variations, on the ground that the word suggests persistence or continuity through successive genera- tions, it seems to us that the name is apt if " continuous" be taken to mean that the occurring variations in any (sufficiently large) set of individuals form a continuous series, the extremes being connected or immediately merging into each other by a series of small gradatory steps. By discontinuous variations we would mean, in contrast to continuous, such considerable and radical variations as have been variously called single varia- ions, sports, mutations, etc., that is, variations, which are not members of a gradatory series, do not group themselves in orderly manner about the modal species form according to the law of error, and although often not large are yet rarely so minute as those differences which distinguish the adjacent mem- bers in any series of individuals arranged on a basis of con- tinuous or fluctuating variation. Mutations, according to the usage of De Vries, our discontinuous variations may or may not be. Thus, all mutations might be called discontinuous variations, although not all discontinuous variations are neces- sarily De Vriesian mutations, that is, certain to breed true under varying conditions of environment. 322 KELLOGG AND BELL As a matter of fact not all continuous variation follows the law of error ; the curve or polygon of frequency is not infre- quently an asymmetrical one; '*skewness" prevails, or the curve may even be bi-modal. But nevertheless the " continuity " of the variations is unmistakable. In a sufficiently large series the extremes of the range are perfectly connected with the mode or modes and hence with each other by gradatory steps very small in size. Whatever the largeness of the difference be- tween the extremes, any two adjacent members of the series are hardly distinguishable. This gradual, insensible but yet effec- tive (as regards widely separated members of the series) kind of variation is most typically illustrated in cases of what Bateson calls " substantive" variation, that is, where the varying char- acteristic is one of pattern, of length, width, or bulk, of the curving of a vein or leg or spine. Excellent examples of this continuous substantive variation are presented by the abdominal and face patterns of Vesfa (see p. 284), the elytral pattern of Diabrotica (see p. 274), the prothoracic pattern of Corisa (seep. 293) and others. According to Bateson, variations in number of antennal and tarsal segments, number of spines, hairs or other processes, and often such numerical or, as called by him, meristic variations, must be looked on as different in kind from the substantive variations, — those capable of perfect mergence from one condition to another, in other words, practically incap- able of quantitative measurement. These meristic variations are called discontinuous by Bateson. Numerous typical ex- amples are included in (Our data. (See the accounts of the variation of the number of the costal wing-hooks in bees and ants, the number of tibial spines in the locust and the cicada, the number of metathoracic tactile hairs in biting bird-lice, etc.) But when one stops to consider the fact that in all these cases variation could hardly occur by any less steps than those of one hook or one spine or one hair, that a half hook or half antennal segment is inconceivable, some serious doubts as to the validity of Bateson's classification of variations as continuous and dis- continuous will certainly result. The doubt is strengthened by the difficulty of a clean classification presented by such cases as that of Jlippodainia convo'gcns. Here we have a substantive STUDIES OF VARIATION IN INSECTS 323 variation in pattern, appearing, however, in such a way as to demand numerical, /. c, meristic, expression. One speci- men has 9 el3^tral spots, another lo, another ii, and so on; the whole range is indeed from o to i8, with every number between represented, each by various combinations of spots. But it is conceivable, and indeed is really the case among our specimens, that these spots might be either of normal size, or of any lesser size down to the limits of visibility. Some of the spots are of the diameter of pin-points ; some of the pin-shaft and some pin-heads. There is perfect gradation or con- tinuity in this variation. But even in such cases as variations in spines and hairs, this gradation might exist ; and indeed does. Althoucrh in our consideration of the variation in the number of the tibial spines of the locust and the cicada and in the number of the tactile hairs of the bird-lice, we have referred to these variations only numerically, /. e., meristically, as a matter of fact there are obvious differences in the length, /. e., size, of the spines and hairs, so that it would be wholly fair to break down the unit differences and speak of differences by one- quarter, one-third and two-thirds of a spine. For the tibial spines of the locust we actually recorded the conditions in the form of fractions. But in the case of a hook or an antennal or a tarsal segment it is a unit or nothing. To our mind the dis- tinction between substantive and meristic variation is not at all equivalent to a distinction between continuous and discontinuous variation. It is a distinction between two categories of varia- tion only in that one category includes such conditions as per- mit more readily of extremely slight, nearly insensible, prac- tically unmeasurable differences, as those of pattern or shape or extent, while the other category includes particularly conditions in which any variation must of necessity be fairly obvious, and usually capable of numerical expression. But we believe, nevertheless, that discontinuous variations occur among insects and that examples of them are presented in the data referring to the species studied by us. For example the occurrence of interpolated, wholly new, and complete cells (determined by the presence of new cross veins or branches of longitudinal veins) in the fore and hind wings of drone honey- 324 KELLOGG AND BELL bees (p. 214) and the occurrence of the curious malformations of venation called by us " deformation" (p. 219) among drone bees must be looked on as sports or truly discontinuous varia- tions. The regular occurrence of a 4-segmented foot, per- fectly complele, functional in those numerous specimens of Blattidas (p. 296) in which natural regeneration has taken place, may be looked on as an example of discontinuous variation. Although no difference in tarsal segments less than that of one is conceivable, it is quite conceivable that the foot with one fewer than the normal number might be in such condition that it would be obviously a 5-segmented foot with one segment dropped out ; in other words that when compared with a normal 5-segmented foot it would appear to be a modification of such a foot with some one segment — and that readily determinable — wanting. But that condition is not at all what appears after the cockroach regenerates a foot. The new foot is onl}^ very little, if any, shorter than the normal 5-segmented foot; one cannot say that it is precisely this or that segment which is lost. It is a new kind of foot, apparently just as capable, as " fit," as useful as the 5-segmented kind. We have regularly occurring, in these cases of regeneration, the development of a wholly changed organ, similar as a whole to the old one, but different from it in all its parts, this difference not being one of incompleteness or serial addition or subtraction, but the difference of newness. It is the regenerative mutation of an organ ! The Rigor of Natural Selection and Determinate Variation. — The theory of determinate variation is based on the hypothe- sis that fluctuating variations are not in all cases, nor necessarily in any case, purely fortuitous and scattering but that because of some intrinsic or extrinsic influence they tend to occur along definite or determined lines. The need for the theory rests on the claimed inadequacy of slight fortuitous variation in offer- ing selection a suflicient " handle " for action. The greatest logical difficulty with the theory is tliat none of the influences, which are known or may be conceived of, is adequate to cause such an effect as that of producing persistent determinate vari- ations. In the case of any developing individual, determinate STUDIES OF VARIATION IN INSECTS 325 variation can be attained by controlling the environment (kind and quantity of food, degree of temperature, humidity, light, etc.), but if such variations (modifications) acquired during development are not inherited, there will be no advance gener- ation after generation along any certain line. There will be no cumulative effect of such determinate variation. The constant repetition of a certain environment on generation after genera- tion of a certain species would of course produce a constant repetition of certain individual modifications (orthoplacy), but we do not know as yet of any actual effect on the species of such persistent ontogenic variations. The need, however, for some such factor in species-forming as determinate variation is obvious and strongly felt. There are certainly few selectionists left who honestly believe that the minute fluctuating variations in pattern, in size, in curve of a vein, in length of a hair, etc., have that life and death value which is the sole sort of value that an *' advantageous varia- tion " must have to be a serviceable handle for the action of natural selection. As a matter of fact, no systematist will have escaped having had it distinctly impressed on him that he rec- ognizes differences in the pattern of lady-bird beetles, in the number of fin rays in fishes, in the branching of a vein in flies' wings, that no enemy, no agent of natural selection, can recog- nize, at least to the extent of pronouncing sentence of death (or not pronouncing it) on its basis. And further, no biologist really satisfies himself with the worn statement, "We must not presume to judge the value of these trivial, these microscopic differences, for we do not know all the complex interrelation and interaction of the organism and its environment." We do not, but we do know for many cases that such differences are actually not of life and death selective value, and reason- ing compels us to believe to a moral certainty that in other cases these fortuitous trivialities have similar lack of life and death importance. The case of the variation of the convergent lady-bird beetle, Hiffodamia convergens (p. 257 et seq.) is distinctly in point. In our account of this variation we have called attention to the suggestiveness, in its light on the rigor of the " struggle for existence" among individuals, of the fact that 326 KELLOGG AND BELL among several thousand individuals, gathered together to hiber- nate after an active life, having been exposed to the attacks of bird and insect enemies, to the rigors of climatic conditions and to the necessities of obtaining food (other smaller insects, as aphids, etc., caught alive), such a range of variation in pattern is found as enables us to describe (so that they may be actually readily distinguished by verbal description), eighty-four "aberrations" or pattern-variates : lady-birds with no spots, with one, with two, with three, with each of all the numbers up to and including eighteen distinct small black spots, the different numbers usually being represented by several different combinations of spots. Systematic entomologists describe Hippodamia convergens as a brown-red beetle with six black spots on each elytron and this description is true for most beetles of this species. But not at all for all ; nor even approximately for many. After a season of exposure to the struggle for existence, to the rigors of selec- tion, individuals with one spot, with six spots, with twelve spots, with eighteen, find themselves alive and healthy ; they come together to pass a quiet winter under the fallen oak leaves on a mountain side ready to mate miscellaneously in the spring and produce young of all manner of pattern (as far as number and arrangement of spots go) which young, whether twelve-spotted as they ought to be, or no-spotted, or eighteen-spotted as they may be, will apparently go safely through life despite the mal- evolent search of the all-powerful bug-a-boo. Rigor of Selection ! Directly touching the point, too, are our data of the variation of series of honey-bees collected from free-flying individuals after exposure as adults to the rigors of out-door life, as com- pared with the variation in the series of bees, adult, but col- lected just as issuing from the cells before being exposed as adults in any way to the external dangers of living. Series of both drones and workers representing both exposed and unex- dosed individuals were studied. The results of this examination are, put in one statement, that the variation among the exposed individuals is no less than that among the unexposed individ- uals. This means that these various mostly slight blastogenic variations (although in such important organs as the wings) which occur among bees at the time of their issuance as active, STUDIES OF VARIATION IN INSECTS 327 winged creatures, are not of sufficient advantage or disadvan- tage to the individuals to lead to a weeding out (by death) or saving of such varying individuals by immediate selective action. Whatever the rigor and danger of the out-doors bee life, these variations seem to be insufficient to cut any figure in the per- sistence or non-persistence of any individual in the face of this rigor. Still other cases in point are those revealed by our study of the variation in the pattern of various insects with incomplete metamorphosis, as the leaf-hopper Tettigonia sp. (p, 287), the Capsid flower-bug, (p. 291), the water-boatman Corisa sp. (p. 293) and the variation in structure of other insects with incom- plete matamorphosis, as the variation in number of tibial spines of the red-legged locust Melano^his fei-mtir-rubrtim (p. 301), in the periodical cicada. Cicada seftendecim (p. 306), etc. In all these cases variation of much range and variety is found in series of the adult individuals, during which a more or less pro- tracted post-embryonic development have been exposed to the struggle for existence, with their patterns and superficial struc- tural characteristics in practically the same conditions as found in the adult stage. This variation has existed for the most part, all through the exposed life of the individual^ and has had its chance to influence for weal or woe the fate of the individ- ual. How much influence have these variations exerted? Our case which most nearly seems to illustrate determinate variation is that of the variation of the flower-beetle, Diabrotica soror (p. 274 et seq.). Among a thousand individuals col- lected on the university campus in 1895, a certain condition of variation in the elytral pattern exists as represented graphic- ally by figure 53. In 1901 and 1902, other thousands col- lected from the same place and examined to determine the con- dition of the variation in this pattern show a distinctly different status, as illustrated in figures 51 and 52. (To be sure that a series of 1,000 individuals really reveals the conditions of this pattern variation, repeated series of 1,000 individuals each were examined and found practically identical.) The difference in the variation status between the 1895 lot and the 1901-1902 lots consists in the dominance in 1901-1902 of one of the two 328 KELLOGG AND BELL modal conditions found to exist in the species, which in 1895 was not the dominant one. There has been a marked change in seven years, not in the pattern itself but in the prevalence or dominance of one type of pattern/ Has the change been brought about by natural selection? Or is it the result of a determinate variation caused by we know not what intrinsic or extrinsic factors ? The variation in Diabrotica's elytral pattern is wholly comparable with the variation in Uippodaniia's elytral pattern. The fusion, partial or complete, of two adjacent spots produces Diabrotica's variety of pattern ; the suppression or addition, partial or complete, of various spots produces Hippo- damia's larger variety. But in Htppodatnia all the variety ex- ists among individuals after exposure, for practically all of the time that such exposure, will occur, to the rigor of selection among individuals. Shall any greater effectiveness be ascribed to this rigor in the case of Diabrotica than actually exists in the case of Hippodamia'^. When we straighten up after a careful microscopic examina- tion of the pattern of Diabrotica to determine its variation, we assure ourselves that no other enemy of these flower-beetles can be conceived to use such discrimination as ours. Does the fly- catcher swooping from its station on fence post or tree branch determine which of two heavily flying Diabroticas shall be its prey on the basis of " two middle spots on left elytron partially fused " in one and " these two spots not touching " in the other? To our minds the change in variation status, the dominance of one mode to-day which was the subordinate mode in 1895, is not due to the action of selection. We do not, indeed, hesitate to believe in those " unknown factors in evolution" which may produce among other results that condition of affairs best named " determinate variation." This variation is not necessarily to be conceived of as purposeful or even advantageous ; if by its cumulation it becomes a disadvantage of life and death value natural selection, which is after all a logical necessity and un- ^ Moreover, in a 1904 lot, consisting of the same number of individuals and collected from the same locality as the lot of 1895 and 1901-02, there is (vpithin 2 years) a marked rise in the percentage of the class which, being the mode in 1901-02, was in the minority in 1895. STUDIES OF VARIATION IN INSECTS 329 doubtedly an actual actively-regulative factor in species control, will take care of it. Variation in Parthenogenetically Produced Individuals and in those of Bisexual Parentage. — Among the " explanations " of variation, that emphasized by Weismann is the most con- spicuous. The admixture of the heredity-bearing germ-plasm of two individuals explains why there is variation, and has in- deed for chief raison d'etre (if " rejuvenescence " be not the primary reason) the production of those variations necessary for the grounding of the natural selection theory. Sex is indeed for the sake of variation ; variation is the result of amphimixis. As a matter of fact, parthenogenetically produced animals vary, and by casual inspection seem to vary in practically equal degree with those of biparental ancestry. That in certain in- stances they really do vary quite as much as do the progeny of two parents is shown by our statistical study of the variation in series of drone honey-bees (parthogenetically produced) as com- pared with the variation of the same organs in series of worker honey-bees, of the same maternal parentage as the drones, but having an added parent, with series of workers and of entirely distinct parentage. The organs examined for variation in these series of bees are the wings, organs used by both drones and workers and having no immediate relation either structurally or physiologically to the differentiation of these two castes or kinds of individuals of the honey-bee species. The workers are " incomplete " only in that most of them are infertile : in no other structural or physiological feature of their make-up are they less " complete " than the drones. They are indeed dis- tinctly the more specialized kind of individual of the two and according to one of the early Darwinian canons of variation might be expected on that account to vary more than the drones. But the drones are males and according to another commonly accepted belief, this is the explanation for a larger variation on their part, if such larger variation occurs. As a matter of fact it does. Reference to our account (p. 214, et seq.) shows that the drones in all the series studied show markedly more varia- tion in the venation of the wings (something the entomological systematists expect to see little of, as witness the constant use 330 , KELLOGG AND BELL of venational characters in keys to families and genera in almost all insect orders) than do the workers, while they show quite as much variation as the workers in the number of the hooks which hold the two wings together in flight. Both these char- acters, /. £?., wing-venation and wing-hooks, are not so-called " male characters " ; they are not to be compared with those secondary sexual characters such as ornamental or aggressive spines, horns, patterns, etc., which are the characteristics that give males their special name for ultra-variation. Moreover we have been able to compare the variation in identical organs, viz., wings in the male honey-bees (partheno- genetically produced) and in male ants (of bisexual parentage) and in all the characteristics studied the male bees varied more than the male ants. Our results stated simply are that in the case of one kind of insect, produced parthenogenetically, varia- tion is quite as pronounced as in two other insects of similar general character but of bisexual parentage. Variations in Males and Fejnales. — The common belief is that males vary more than females. We do not know on just what evidence, if any, this belief is based, except that derived from an examination of those ill-understood and superficial structures familiarly known as secondary sexual characters, such as ornamental combs, wattles, feather-tufts and color-pat- terns, and spines, horns and spurs for fighting. The only cases in our series of insects studied which have given us evi- dence touching this point are those of the honey-bees and mos- quitoes. In both these cases we have noted the variation in certain structural characters common to both males and females and distinctly not of secondary sexual character. The result is that we find a larger variation in wing characters in male honey-bees (drones) than in (infertile) female bees (workers), while in mosquitoes we find (using confessedly dangerously short series) that the females show a slightly larger variability. In the case of the bees, we have the complication of partheno- genetic birth for the males. Correlated Variation ; Bilateral Symmetry^ Metamerism^ Other Correlations, — Insects are bilaterally symmetrical and metameric animals. There are thus right and left and fore and STUDIES OF VARIATION IN INSECTS 331 aft structural correlations. Do the variations, continuous and discontinuous, show similar bilateral and metameric correlation ? Evidence regarding this question will be found on many pages in the present paper, right and left correlation, at least, having been considered and briefly discussed in connection with almost all of the various cases studied. And the evidence is curiously conflicting. For example in the male black ant in which was studied the variations of the venation and number of hooks, a close correlation in the variation conditons of right and left wings exists. On the other hand in the honey-bee the bilateral correlation of variation seems surprisingly small (see pp. 214- 222.) In the case of variations in pattern also there is no uniform- ity among the various cases studied. In Hiffodamia convergens (P- 257 et seq.) the two elytra show pattern-variations quite independently ; in Diabrotica soror (p. 274 et seq.) on the con- trary there seems to be a marked right and left correlation in the elytral pattern-variation. In the cases of the variation in number of tibial spines on the right and left hind tibiae of locusts (p. 301) and cicadas (p. 306) we have simply made a brief statement, in each case, of the actual conditions of corre- 'lation leaving the reader to draw his own conclusions. In the case of the variation in actual and relative length of the anten- nal segments of the scale insect, CeropUo yucccB (?) (p. 310) there is a surprising lack of correlation between the right and left antennae. We have not attempted to determine the mathematical ex- pression (coefficient of correlation) for any of the cases studied. The data presented, however, will enable any biometrician who sees an advantage in doing this, to do it. But without check- ing our results by the use of that method there seems, on the whole, to be a surprising lack of that fine degree of correlation in variation which we should expect to find existing — if we believe that the actual existing conditions of structure and pattern in these bilaterally symmetrical animals are an expres- sion of the result of the action of a rigorous natural selection. If one condition of pattern or structure is the most advantageous (of the many conditions which selection among a host of fluc- tuating variations could have established) surely this condition 332 KELLOGG AND BELL ought to be pretty closely similar on both sides of the insect. That as much bilateral variety as actuall}^ exists, in many of the species examined by us, should exist — a variety comparable in certain cases even with the degree of variety revealed by the comparison of considerable series of individuals — is a state of affairs that only confirms us in the belief that these innumerable small continuous variations, on which for so long the thorough- going selectionists have put their faith as the sufficient bases for natural selection's species-forming work, are clearly not com- petent to serve as such bases. If these " continuous " variations are the foundation stones of new species, some other agent than selection must be found or invoked to build several courses on them, to produce some cumulation of them, before natural selec- tion finds them of that life and death worth which is the prere- quisite for her potent interference. PROCEEDINGS WASHINGTON ACADEMY OF SCIENCES Vol.. VI, pp. 333-427. January 31, 1905. PAPERS FROM THE HOPKINS-STANFORD GALAPAGOS EXPEDITION, 1898-1899. XVII. SHORE FISHES OF THE REVILLAGIGEDO, CLIP- PERTON, COCOS AND GALAPAGOS ISLANDS. By Robert Evans Snodgrass and Edmund Heller. TABLE OF CONTENTS AND DISTRIBUTIONAL INDEX. R = Revillagigedo, CI = Clipperton, C = Cocos, G = Galapagos, (A) = American, (W) = Western Pacific, (C) =\vide ranging or cosmopolitan. Page. Introduction 333 Branchiostomidae 342 1. Braiichiostoma clotigatum. G, (A) 342 Galeidse 342 2. Galeocerdo tigriuus. G, (W) 342 3. Carckarias. galafagensis. G 343 4. Carcharias platyrhy7ichus. R,C1,(A) 344 5. TricBtiodon obesus. C, (W) 344 Sphyrnidae 345 6. Sphyrna tudes. G, (C) 345 RhinobatidK 345 7. Rhinobatus flnniccps. G, (A) 345 Dasyatidae 345 8. Dasyatis longa. G, (A) 345 Mobulidae 346 9. Majtfa birostris. G, (A) 346 Ophichthyida; 346 10. Myrichthys pa7itostigmius. R 346 11. Ophichthus trtserialis. G, (A) 347 Proc. Wash. Acad. Sci., January 31, 1905. (333) 334 SNODGRASS AND HELLER MursenidjE 347 12. Rabula mormorea. G 347 13. Gymnothorax pictus. R, (W) 347 14. Gym7iothorax chlevasfes. G 347 15. GymnotJiorax dovii. G, (A) .... 348 16. Murcena insularum. G 348 17. MurtEna le7itiginosa. G 348 18. Echidna jwcfurtia. R, (A) 348 Clupeidse 34S 19. Clupanodo?i libertatis. G, (A) 348 Chauliodontidse 348 20. Zalai-ges lucetius. G, (A) 348 Hemirhamphidce 349 21. Hyporkamphus roberti. G, (A) 349 22. Hetniramphus saltator. G, (A ) 350 23. EuleptorampJius lougirostris. G. (W) 350 ExoccEtidae 351 24. Rvolatitia microptera. G, (W) 351 25. Exocceiiis volitans. R, (C) 351 26. Exonautes speculiger. (C) 352 27. Cypsilurus xenopterus. R, (A) 352 28. Cypilurtis cyanopterus. G, (A) 352 Mugilidse 352 29. Mugtl cepkalus. G, (C) ■ • . . 352 30. Alugil thoburni. G, (A) 353 31. Mugil curcma. R, (A) 353 32. Mugtl seiosiis. R, (A) 353 33. Chcsnotnugil proboscideiis. R, (A) 354 34. ^iierimana /larettgiis. G, (A) 354 Sphjrjenidae 354 35. Spkyrcena idiastes. G 354 Holocentridae 354 36. Myriprisfis occidentalis. C, G, (A) 354 37. Myripristis clarionensis. R 356 38. Myripristis murdjan. C, G, (W) 356 39. Holotrachys lima. C,(W) 358 40. Holocetttrus siiborbitalis. R, C, G, (A) 360 Mullida; 360 41. Pseiidupcnens dentatus. R, (A) 360 Scombridae 360 42. Scomber Japotiicus. G, (C) 360 43. Gymnosarda pelamis. R, G, (C) 360 44. T/tiiintiis f/iynmis. G, (C) 361 45. Germo alaltuiga. (C) 361 46. Scombcromoriis sierra. G,(A) 361 Carangidie 362 47. Elagatis bipinnulatus. (C) 362 48. Decaptertis scombrinus. G 362 49. Trachurns symmetricus. G, (A) 363 SHORE FISHES OF GALAPAGOS ISLANDS 335 50. Trachurops crumenophihalnia. R, (A) 364 51. Zalocys siilbe. R 364 52. Caraiix cabaUus. G, (A) 364 53. Caranx margtnatiis. R, (A) 364 54. Caranx lattis. G, (C) 364 55. Caranx lugtibris. R, (C) 365 56. Caranx melamfygus. R, C, G, (W) 365 57. Caranx orthogravimus. R 365 Coryphienidie 365 58. Coryphcena hippurus. (C) 365 59. CoryphcEna equisetis. R,(C) 366 Nomeidse 366 60. Gobiomorus gronovii. (C) 3^^ Kuhliidae 3^6 61. KuJiUa tcenitira. R, C, G, (W) 366 Apogonichthyidie 367 62. Amia atradorsata. C, G 3^7 63. Amia atricauda. R 3^7 64. Galeagra paminelas. G 3^7 Serranidae 3^7 65. Epinepheltis analogus. R, G, (A) 367 66. Epinephehis labrtformis. R, CI, C, G, (A) 367 67. Dennaiolepis punctatus. R, C, G, (A) 368 68. Mycteroperca xenarcha. G, (A) 368 69. Alycteroperca olfax. C, G, (A) 3^^ 70. Alycteroperca ruberrima. G 37° 71. Cratinus agassizii. G 37° 72. Paralabrax albomaculatus. G 37° 73. Prionodes fasciatus. R, G, P, (A) 372 74. Prionodes stilbostigma. G 37^ 75. Paranthias furcifer. R, G, (A) 372 76. Pronotogrammics multifasciatus. R, C, (A) 373 77. Rypticus bicolor. G 373 Priacanthidae 373 78. Priacanthus criientaitts. R, C, G, (A) . 373 Lutianidse 374 79. Lutianus viridis. R, C, G, (Tres Marias Ids.) 374 So. Lutianus jordani. C, (A) 375 81. Lutianus argentiventris. C, G, (A) 375 82. Xenocysjessice. G 375 83. Xenichthys agassizi. G 37^ HsemuHdse 37^ 84. Anisotremus surinatnensis. G, (A) 37^ 85. Anisotremus interruptus. R, (A) 377 86. Atiisotremus scapularis. C, G, (A) 377 87. Orthopristis forbesi. G 377 88. Orthopristis lethopristis. G 37^ 89. Orthopristis chalceus. G, (A) 379 90. Orthopristis ca7itharinus. G 379 336 SNODGRASS AND HELLER Sparidse 379 91. Calamus taurinus. G, (A) 379 92. Arch osnrgus pour talesti. G 3S0 GerridiE 380 93. Eucinostomus dozvi. G, (A) 380 94. Xystaema ct)iereu?n. G, (A) 382 Kyphosidse 382 95. Doydixodon fremiiivillei. G 382 96. Kyphosus analogus. R, (A) 384 97. Kyphosus elega7is. R, C, G, (A) 384 98. Kyphosus lutesccus. R 384 Sciaenidae 384 99. Corvula eurymesops. G 384 100. Scicena perissa. G 385 loi. Umbriiia galapagoruin. G 385 Cirrhitidce 385 102. Cirrhitus rivulatus. R, G, (A) 385 Pomacentridae 385 103. Azurina eupalama. G 3S5 104. Pomacetilrus leucorus. R,C,G 387 105. Pomacentrus redemptus. R 3S9 106. Pomacentrus arcifrons. C, G 389 107. Nexilarius concolor. G, (A) 389 108. Abudefduf margitiatus. R, C, G, (A) 390 109. Microsfaihodon bairdii. R, G, (A) 390 no. AlicrospathodoH dorsalis. R, C, G, (A) 390 111. Nexilosus albcmarlcus. G 391 Labridse 391 112. Bodianus dtplotcenius. R, CI, C, G, (A) 391 113. Bodianus eclancheri. G 392 114. Pimelometop07i darxvinii. G 394 115. Halichceres nicholsi. R, G 395 116. Halichceres scllifer. R 395 117. Pseudojulis adustus. R 396 118. Pseudojulis notospilus. R, (A) 396 1 19. Thalassoma socorroeiise. R 396 120. Thalassoma grammaticum. R 396 121. Thalassoma vircns. R 396 Scaridse 397 122. Calotoinus xenodon. R 397 123. Callyodou noycsi. G 397 124. Callyodou perrico. G, (A) 397 Oplegnathida- 397 125. Oplcgnathus iiisigiic. G, (A) 397 Chsetodontidae 399 126. Forcipiger longirostris R, (W) 399 127. Chcetodon ttigrirostris. R, G, (A) 400 128. Holocanthus passer. C, G, (A) 401 129. Holocanthus clarionensis. R 401 SHORE FISHES OF GALAPAGOS ISLANDS 337 130. IlolocautJius iodoctis. G 4°^ Zancliche 4°^ 131. Zanclus caticscetis. C, R. G, (W) 402 Teuthididie 402 132. Cienockceius strt'gosus. C, (W) 4°^ 133. Hepatus triostegus. R, C, (W) 403 134. Hepatus crestonis. C, (A) 4°3 135. Hepatus aliala. R, CI, C. (W) 403 136. Xesurus fuuctatus. R, (A) 4°4 137. Xesicrus laticlavius. R, C, G 4°4 Balistidie 4o6 138. Batistes verres. R, CI, C, G, (A) 4°^ 139. Canthidermts atigulosus. C, (W) 4^7 140. Xa?itkichikys mento. R 4°^ 141. Melichthys bispinosns. R, C 4°° Monocanthidse 4°9 142. Cantherines sandivichensis. R, (W) 409 143. Osbeckia scripta. R, (W) 41° Ostraciidse 4^° 144. Ostracion leiitiginosttm. CI, G, (W) 410 145. Ostracion clippertonense. CI 4^" Tetraodontidae 4^2 146. Spheroides angusticeps. G, (A) 412 147. Spheroides lobatus. G, (A) 4'- 148. Spheroides a>inulatus. G, (A) 4'- 149. Tetraodon setosus. R, CI, C, G, (A) 4^3 Diodontidie 4^3 150. Diodon hystrix. R, G, (C) 4^3 151. Chilomycterus affinis. G, (W) 414 Scorpaenidje 4^4 152. Sebastopsis xyris. R, G, (A) 414 153. ScorpcBiia histrio. G, ( A) 4^5 154. Pontinus strigatus. G 4^5 Gobiidse 4^5 155. Eleotris tubular is. C 4^5 156. Cotylopus cocoensis. C 4^5 157. Zonogobius rhizophora. G 4'° 158. Zonogobius zebra. R, (A) 4^" 159. Odontogobius gilberti. G 4^" 160. Mapo soporator. R, CI, C, G, (C) 416 MalacanthidiE 4^7 161. Caulolatilus princeps. G, (A) 4^7 Dactyloscopidse 4^7 162. Myxodagnus opercularis. G, (A) 4-7 Batrachoididse 4^° 163. Porichthys tnargaritatus. G, (A) 418 Blenniidse 4^" 164. Dialommus fuscus. G 4^" 165. Emmnion brisiohe. G 4^° 338 SNODGRASS AND HELLER 166. Rnnula azalea. G 419 167. Alticus ailanticHS. G, (A) 419 168. Aliicus ckiostictus. R, (A) 419 169. Malococtenus zonogas.ter. G 420 170. Leptsoma jenkitisi. G 420 171. E7icheliophis jordant. G . . . 420 Ophidiida; 420 172. Chilara taylori. G, (A) 420 173. Otophidium indefatigabile. G, (A) 421 Brotulidae 421 174. Petrotyx hopkinsi. G 421 175. Eutyx diagrammtis. G 421 Triglidse 421 176. Prionotus miles. G . . " 421 Echeneididse 421 177. Bcheneis remora. G, (C) 421 Gobiesocidae 422 178. Gobiesox poecilophihalmus. G 422 179. Gobiesox adustus. R, (A) 422 180. Arbaciosa truncata. G 422 Pleuronectidic 422 181. Platophrys constclladis. G,(A) 422 182. Platophrys leopardinus. R, C, G, (A) 423 Soleidae 423 183. Symphurtts atramejitatus. G, (A) 423 Antennariidse 424 184. Antetinaritis tagtcs. G 424 INTRODUCTION. In the fish-fauna of these islands there is a very conspicuous element formed of species that belong to the islands of the western part of the Pacific. These are : Galeoccrdo trig inns. G. TrtcBiiodon obesus. C. JEul€pio7-ainphus longirostris. G. EvolantJa niicroptera. G. Afyripristis viitrdjan. C, G. Jlolotrachys lima. C. Kuhlia i Centura. R, C, G. Caranx tnclamfhygns. R, C, G. Fo7'cipiger longirostris. R. Zanclus caiiescens. R, C, G. Ctenochcptus strigosiis. C. Jlcpatiis tri osteons. R, C. shore: fishes of galapagos islands 339 Hefaius aliala. R, CI, C. Canthidcr^nis angidosus. C. Cantherines sandzuichensis. R. Osbeckia scri^ia. R. Ostracion Icntiginosum. C, G. Chihmyctcrus affinis. G. Of these i8 species on\y Galeocerdo ttgrmum^ Osbeckia serif ta and Chilomycterus affinis are known from the American main- land coast. There is a small number of species that are peculiar to the islands as a group, i. c, that occur at 2 or more of them but are not known elsewhere. These are : Atnia atradorsata. C, G. Ltttianus viridis. R, C, G (Tres Marias Ids.) Pomacentrus leucortis. R, C, G. Poniacentriis ai'cifrons. C, G. I chthy callus nicholsi. R, G. Melichthys bisfinostis. R, C. Xesurus laticlavitis. R, C, G. These may be termed Eastern Pacific Insular species. Two of them, Po7nacenirus leucorus and Xesurus laticlavius, may be regarded as diagnostic of the islands as a group. They are known from all the islands except CHpperton, but the Clipperton fishes are too little known to be here considered. Ltttianus viridis^ although occurring at all the islands, is known also from the Tres Marias Islands near the coast of Mexico and is, there- fore, likely to be taken also along the mainland shore. R. CI. Widely ranging species American species Western Pacific species Eastern Pacific Insular species Peculiar species Totals 9 'I 5 15 I 18 II 6 2 10 59 9 6 42 72 38 126 Amia atradorsata and Pomacentrus arcifrons are represented at the Revillagigedo Archipelago by the related species Amia atricauda and Pomacentrus rcdemptus. 340 SNODGRASS AND HELI.ER Omitting the 2 deep-sea Galapagos species, Galcagra -pani- melas and Pontinus strigattis, of whose distribution nothing is known, the derivation of the fauna of each island or group of islands may be indicated by the tabulation of species on page 339- Of the Revillagigedo fauna, as shown by this table, 12.5 ^ is composed of widely ranging species ; slightly less than 50 fo is American; 11 /o is Polynesian; about 7 fo is Eastern Pacific Insular, and nearly 21 ^ is peculiar. The peculiar Clipperton species, Ostracion clip-pei'tonense is very closely related to O. canmrtiin of the Hawaiian Islands. Hence, of the 9 species known from this island, 4 are of Western Pacific orijjin. Of the Cocos fauna a little less than 50 Jo belongs to the American mainland ; about 30 Jo is Polynesian ; a little less than 16 Jo is Eastern Pacific Insular ; and about 5 ^ is peculiar. In the Galapagos fauna about 8 Jo is composed of widely ranging species ; slightly less than 47 Jo is American continental ; 7 ^ is Polynesian ; 5 ^ is Eastern Pacific Insular ; and 33 Jo is peculiar. It is interesting to note that the American faunal element of Cocos Island is much more closely related to that of the Gala- pagos Archipelago than to that of the Revillagigedo Archipel- ago. There are 5 species of this class that occur at the Gala- pagos but not at the Rivillagigedo Islands, while there are no species that occur at Cocos and the Revillagigedo islands and not at the Galapagos. Furthermore, Amia alradorsata and Poinacent7'us acrifrons are peculiar to Cocos and the Galapagos islands. Mclichthys bispinosus is peculiar to Cocos and the Revillagigedo islands, but is very closel}^ related to the Polyne- sian species J/, radula. The Western Pacific fauna of Cocos is about equally distributed between the Galapagos and the Revillagigedo archipelagos. The Galapagos Islands lie in the cold Peruvian current flowing northwest from Cape Horn, while both Cocos and the Revillagigedos lie in the warm equatorial and counter equatorial currents. Of the 56 Galapagos species that belong to the American mainland, only 17 occur also at the Revillagigedo Islands. SHORE FISHES OF GALAPAGOS ISLANDS 34I Only 4 of these mainland forms are South American. Hence, one set of Central American species populates the Revillagigedo Islands and another set the Galapagos Islands, the 2 sets inter- mingling along the mainland. The fish-faunas of the different islands of the Galapagos Archipelago differ greatly in the num- ber of species and the relative number of individuals of the same species found at each, and in many cases different parts of the same island show differences of an equal degree. This paper is the first report on the Cocos and Clipperton shore fishes, and the majority of the Galapagos species listed have not heretofore been recorded from these islands. Many imperfectly known species are redescribed, and numerous color notes are given taken from freshly captured specimens. Two species are described as new. Twenty-three other new species were described in Paper XV of the Hopkins-Stanford Gala- pagos series.^ This paper, though in part compiled, is based mainly on specimens now in the ichthyological collection of Stanford University. The Revillagigedo specimens were col- lected by Dr. C. H. Gilbert during the Albatross expedition of 1889, and by Mr. R. C. McGregor in 1897. The Galapagos, Cocos and Clipperton specimens were nearly all collected by the authors during 1898 and 1899. The zoological sequence adopted is that used by Jordan and Evermann in their Fishes of North and Middle America. Measurements of length are given in millimeters ; other meas- urements are in hundredths of the length to the end of the caudal vertebrae, except where stated otherwise. The authors express their obligation to Dr. David Starr Jor- dan for assistance in the identification of species, and especi- ally for help in the determination of synonymy ; and to Dr. Charles Henry Gilbert for invaluable aid in procuring the fish colle;cting equipment with which the expedition was provided, and, later, for assistance while working with the material ob- tained. ^ Proc. Wash. Acad. Sci., Vol. V, 1903 (Sept. 12), 1S9-229. Proc. Wash. Acad. Sci., January, 1905. 342 SNODGRASS AND HELLER Family BRANCHIOSTOMID.E. I. BRANCHIOSTOMA ELONGATUM Sundevall. Branchiostoma elongatum Sundevall, Vet. Akad. Forh. 1853, 147, Chin- chas Islands. — Jordan & Evermann, Fishes North and Mid. Amer., I, note, 4, 1896. — Steindachner, Fauna Chilensis, 334, 1898 (Cav- ancha Bay, Iquique). Range. — Coast of Peru, Chinchas Islands, Galapagos Islands. Not hitherto reported from the Galapagos ; 13 specimens, the largest 20 mm. long, dredged in about 10 fathoms on a bottom of fine black sand in Tagus Cove, Albemarle. Family GALEID^. 2. GALEOCERDO TIGRINUS Miiller & Henle. Galeocerdo tigrinus Muller & Henle, Plagiostomen, 59, 1838. —Jordan & Evermann, Fishes North and Mid. Amer., i, 32, 1896. Range. — India, Australia, Polynesia, Galapagos Islands, west coast of tropical America. Individuals of this species frequently seen in Tagus Cove, Albe- marle and in the straits between Albemarle and Narboro. Generally solitary or associated with Carcharias galapagensis. Easily distin- guished from the latter by the greater size and by the vertical stripes on the sides of the body. We have the jaws and skin of the head of one specimen 9 feet long (2,700 mm.), and all that we saw were of about this length. The upper lobe of the caudal fin is about one-fourth of the total length, and is considerably shorter than the space between the doi'sal fins. Color: above, grayish, spotted with obscure dusky which runs into vertical bars on the sides of the body; below whitish. Snout to first dorsal 33 mm. ; snout to second dorsal yS ; snout to pec- toral 26 ; snout to ventral 44 ; snout to anal 80 ; upper lobe of caudal 25 ; first dorsal to second dorsal 33 ; pectoral 15. Snout short, length from tip to front of mouth about i^ in width of mouth. Nostril 3 in snout. A groove along the base of each jaw, the two continuous around the angle, upper about twice length of lower. Nostril with large triang- ular flap on inner half of anterior margin. Teeth all of same shape, being flat and triangular with a deep notch on outer side — a notch so large as to give the tooth a bilobed appear- ance, the lobes being one basal and the other apical, both directed out- wardly ; free edge of basal lobe roundly convex with coarse serrations ; SHORE FISHES OF GALAPAGOS ISLANDS 343 apical lobe acute, directed outward and upward or backward, serrated on both sides, the serrations largest on inner side near base ; teeth ar- ranged in several series ; a single row in front vertical, the others hori- zontal ; back of each anterior vertical tooth is a longitudinal row of hori- zontal overlapping teeth ; five well developed teeth in each longitudinal row mesially, but laterally decreasing gradually to one; teeth of lower jaw smaller than those of upper and in fewer longitudinal rows. 3. CARCHARIAS GALAPAGENSIS Snodgrass & Heller, new species. Eulamia lamiella, Jordan & Bollman, Proc. U. S. Nat. Mus. 1889, 179 (Chatham Island); not of Jordan & Gilbert. Eulamia {Platypodon) platyf/iynckus Gilbert (in part), Proc. U. S. Nat. Mus. 1 89 1, 543", Galapagos Islands. Carcharhinus platyrhyiichiis, Jordan & Evermann (in part), Fishes North and Mid. Amer., i, 36, 1896. Type. — No. 12334, Stanford Univ. Mus. Diagnosis. — The same in every respect as Carcharias platyrhyn- chus (Gilbert) except that the fins are at all ages of uniform coloration with the body, being never margined with white. Range. — Galapagos Islands. Description of the Type (embryo, 650 mm. long). — Length from tip of snout to front of mouth less than width of mouth by one-half diameter of eye; length from angle of mouth to symphysis of lower jaw equal to length of snout from mouth less one-half diameter of eye; distance between outer ends of nostrils a little less than width of mouth ; eye a little less than one-fourth width of mouth ; base of pectoral about 3 in its own length ; base of first dorsal \\ in its height; ventrals as long as base of dorsal, equal to length of snout from mouth ; base of anal 2 in entire length of ventral, equal to base of second dorsal ; height \ greater than that of second dorsal ; lower lobe of caudal a little less than \ of upper lobe; two gill-slits above front of base of pectoral. The proportions differ somewhat in different sized specimens. In one 550 mm. long some of the above measurements are as follows: Length from snout to mouth equal to width of mouth ; length from angle of mouth to pectoral a little greater than width of mouth ; length from angle of mouth to symphysis of lower jaw less than length from symphysis to snout by | diameter of eye ; distance between outer ends of nostrils equals width of mouth ; lower lobe of caudal | of upper lobe. Extremely abundant about the Galapagos Islands, especially about Wenman and Culpepper and between Albemarle and Narboro. The 344 SNODGRASS AND HELLER adults average 6 to 8 feet in length. We examined a large number of them, several hundred being taken aboard the schooner, and we saw probably thousands in the water. None of them had the fins marked with white. They feed on fish and are probably dangerous enemies of the young fur-seals and sea-lions of the Galapagos Islands, for they closely patrol the shores about the seal rookeries. We often found in their stomachs pieces of sea-lions, but they may have been feeding on the carcasses left by the sealers. MEASUREMENTS OF CarcJiarias galapagensts . No. Stanford University Museum. Length to base of caudal fin in mm. Snout to pectoral Snout to first dorsal First dorsal to second dorsal Upper lobe of caudal Base of pectoral to base of ventral..., Pectoral Height of first dorsal 12326 12325 470 553 650 32 37 27 44 ^l 43 30 28 31 37 40 35 30 34 21 21 26 22 II 16 13 12324 4. CARCH ARIAS PLATYRHYNCHUS (Gilbert). Etilamia {Platypodon) platyrhynchus Gilbert (in part), Proc. U. S. Nat. Mus. 1 89 1, 543, Clarion and Socorro islands and Magdalena Bay, Lower California. Carcharhinus platyrhyiichus, Jordan & Evermann (in part). Fishes North and Mid. Amer., i, 36, 1896, Range. — Coast of Lower California (Magdalena Bay) ; Revilla- gigedo Islands; Clipperton Island. The collection contains one specimen from Clarion Island and one taken at sea between Clarion and Clipperton islands, 13° 12' N. ; 111° 45' W. The Clarion specimen, which is about 3^ feet long, has the dorsal and pectoral fins tipped and posteriori)' bordered with white, and all others seen in the water about the island were similarly marked. The specimen taken near Clipperton Island has the mar- ginal parts of the fins pale. In the original description of the species Dr. Gilbert assigns this fin coloration to the largest specimens only. 5. TRI^NODON OBESUS (Riippell). Carcharias obesus Ruppell, Neue Wirbel., Fisch., 64, pi. 18, fig. 2, 1837. Trianodo7i obesus, Muller & Henle, 55, pi. 20. — Dumeril, Hlasmobr., 386. — GiJXTHER, Cat., VIII, 383, 1870. Range. — Red Sea, Indian Ocean, New Hebrides, Cocos Island. One specimen, about 5 feet long, taken at Cocos Island. This is the only record of the species from the Eastern Pacific. SHORE FISHES OF GALAPAGOS ISLANDS 345 Snout very short, 3.7 in width of mouth; angle of mouth to sym- physis of lower jaw 1.33 in width of mouth; width of nostril 2.4 in snout; eye longitudinally elongate-oval ; teeth in several series in each jaw, the outer ones most nearly erect, but all inclined backward, es- pecially the inner ones, which, in the upper jaw, are almost hori- zontal ; all tricuspid, having a long slender median cusp and a much smaller one on each side at base; nostrils with a double flap on inner half of anterior edge forming a sort of tubular appendage ; posterior gill-slit over the base of pectoral ; no grooves about the mouth. Length to base of caudal fin 911 mm.; snout to first dorsal 36; first dorsal to second dorsal 39; second dorsal to caudal 10; upper lobe of caudal 31 ; pectoral 20; base of first dorsal li in its height; base of second dorsal equal to height; height of anal li in height of second dorsal ; ventral i| in pectoral ; lower lobe of caudal 2 in upper lobe. Color: dark uniform slate above, below livid-yellowish slate; tip of first dorsal and of upper lobe of caudal creamy white. Family SPHYRNID^E. 6. SPHYRNA TUDES (Cuvier). Zygana hides Cuvier in Valenciennes, Mem. Mus., ix, 225, 1822, Nice. Sphyrna hides, Jordan & Evermann, Fishes North and Mid. Amer., i, 44 1896. Range. — Tropical parts of the ocean in general. We saw several small individuals of a Sphyrna^ probably 6'. tudes^ in Tagus Cove, Albemarle, but we were not able to secure any specimens. Family RHINOBATID^. 7. RFIINOBATUS PLANICEPS Carman. Rhhiobatiis planiceps Carman, Bull. Mus, Comp. Zool., vi, 168, 1880, Peru ; Calapagos. — Jordan & Evermann, Fishes North and Mid. Amer., i, 64, 1896. Range. — Coast of Peru ; Galapagos Islands. Numerous skates were seen about the Galapagos Islands, some of which may have been this species, but we did not obtain any specimens of it. Reported from the Galapagos Islands by the Hasslar expedition. Family DASYATIDiE. 8. DASYATIS LONGA (Carman). T?ygo7i longa Carman, Bull. Mus. Comp. Zool., vi, 170, 1880, Acapulco ; Panama. — Jordan & Evermann, Fishes North and Mid. Amer i 8;' i8q6. ' ' ^' 346 SNODGRASS AND HELLER Range. — Gulf of California to Panama ; Galapagos Islands. One specimen taken at Mangrove Point on the east coast of Narboro. Individuals numerous in the shallow sandy-bottomed lagoons of the mangrove swamps at Mangrove Point, Narboro. Length to root of tail, 460 mm. ; tail 287 mm. (apparently not en- tire) ; width of disc 534 mm. Caudal spine very slightly less than distance from tip of snout to mouth, about two-thirds longer than middle of mouth ; five papillae in mouth, the median 3 large and conspicuous, the lateral ones small ; an elongate patch of spine-like tubercles on middle of back ; a median series of similar tubercles be- ginning a little back of central dorsal patch and extending along back and tail to caudal spine ; a short longitudinal series of similar tuber- cles on each side of central dorsal patch. Family MOBULIDiE. 9. MANTA BIROSTRIS (Walbaum). Raia birostris Artedi, Piscium, 535, 1792. Mafifa birostris, Jordan & Evermann, Fishes North and Mid. Amer., i, 92, 1896. Range. — Both shores of tropical and subtropical America. We frequently saw, amongst the islands of the Galapagos Archi- pelago, enormous rays probably belonging to this species, but no specimens were obtained. [Family SILURID^. NETUMA INSULARUM Flora Hartley Greene. Tachystirus elatturus (var. ?), Jordan & Bollman, Proc. U. S. Nat. Mus. 1889, 179, Gulf of Panama. Netiima insularuni Flora Hartley Greene in Gilbert, Proc. U. S. Nat. Mus. 1896, 439, Galapagos Islands. — Jordan & Evermann, Fishes North and Mid. Amer., in, Addenda, 2770, 1898. The specimen from which this species was described was taken by the Albatross., in 1S8S, in the Gulf of Panama. The subsequent reference of the species to the Galapagos Islands is a mistake.] Family OPHICHTHYID.E. 10. MYRICHTHYS PANTOSTIGMIUS Jordan & McGregor. Myrichthys pantostigmius Jordan & McGrecor in Jordan & Evermann, Fishes North and Mid. Amer., in. Addenda, 2802, 1898, Clarion Island. — Jordan & McGregor, Rep. U. S. Fish. Comm. for 1898 (1899), 274, pi. 4 (Clarion Island). SHORE FISHES OF GALAPAGOS ISLANDS 347 Range. — Clarion Island, Revillagigedo Archipelago. (Collected by Mr. R. C. McGregor.) II. OPHICHTHUS TRISERIALIS (Kaup). Mtiranopsis triserialis Kpmv, Apodes, 12, 1856, Pacific. Ophichthys r^^z/^r Jordan & Bollman, Proc. U. S. Nat. Mus. 1889, 155 and 180, Charles Island. Ophichihus triserialis, Gilbert, Proc. U. S. Nat. Mus. 1889, 450 (Chatham Island). ^Jordan & Evermann, Fishes North and Mid. Amer., i, 384, i; Range. — West coast of tropical America, Galapagos Islands. Known from Charles and Chatham islands of the Galapagos Archi- pelago, it having been taken at both places by the Albatross. Family MUR^NID^. 12. RABULA MARMOREA (Valenciennes). Murcenophis martnoreusYx-LY.nci'Eyi^KS, Voy. Venus, Zool., 347, pi. 10, fig. I, 1855, Galapagos Archipelago. Rabula viarmorea, Jordan & Evermann, Fishes North and Mid. Amer., i, 391, 1896. Range. — Galapagos Islands. "A doubtful species." (Jordan & Evermann.) Reported only by the Venus. 13. GYMNOTHORAX PICTUS (Ahl). Muranapicta Ahl, De Muraena et Ophichtho, vi, 8, pi. 2, fig. 2, 1789. Lycodontis piciiis, Jordan & Evermann, Fishes North and Mid. Amer., iii, Addenda, 2805, 1898. —Jordan & McGregor, Rep. U. S. Fish Comm. for 1898 (1899), 274. Range. — Western Pacific, and Revillagigedo Islands. This is a common species of the East Indies and has been obtained at Clarion Island, Revillagigedo Archipelago, but nowhere else in the eastern Pacific. 14. GYMNOTHORAX CHLEVASTES ( Jordan & Gilbert). Sidera chlevasies Jordan & Gilbert, Proc. U. S. Nat. Mus. 1883, 208, Galapagos Islands. Lycodontis chlevastes, Jordan & Evermann, Fishes North and Mid. Amer., I, 398, 1896. Range. — Galapagos Archipelago. Known only from the Galapagos Islands, where one specimen was taken by the Albatross. 348 SNODGRASS AND HELLER 15. GYMNOTHORAX DOVII (Giinther). Mumna dovu GiJ^THER, Cat., viii, 103, 1870, Panama. Lycodotitis dovii, Jordan & Evermann, Fishes North and Mid. Amer. , i, 397, 1896 (" Gulf of California to Galapagos "). Range. — West coast of tropical Anierica ; Galapagos Archipelago. 16. MUR^NA INSULARUM Jordan & Davis. Murcena insularum Jordan & Davis, Apodal Fishes, 609, 1892, Chatham Island. — Jordan & Evermann, Fishes North and Mid. Amer., i, 400. Range. — Galapagos Islands. We have one specimen 390 mm. long collected on a rocky beach near Iguana Cove, Albemarle Island. 17. MUR^NA LENTIGINOSUM Jenyns. Murcsna lentiginosa Jenyns, Voy. Beagle, Zool., 143, 1842, Galapagos Islands. — Jordan & Evermann, Fishes North and Mid. Amer., i, 402, 1896. Range. — Galapagos Islands. One specimen from Turtle Point Reef, near Tagus Cove, Albemarle 18. ECHIDNA NOCTURNA (Cope). Pcecilophis nocturnus Cope, U. S. Geol. Surv. Mont., 484, 1871, Rio Grande, Costa Rica. Echidna nocturna, Jordan & Evermann, Fishes North and Mid. Amer. , i, 402, 1896. — Jordan & McGregor, Rept. U. S. Fish. Comm. 1898, 275 (Clarion and Socorro islands). Range. — Costa Rica; Cape San Lucas; Clarion and Socorro islands, Revillagigedo Archipelago. Family CLUPEIDiE. 19. CLUPANODON LIBERTATIS (Giinther). Meletta libertatis QVi^T\iY.\\, Proc. Zool. Soc. Lond. 1866, 603, Libertad, Cen- tral America. Opisthoncma libertate, Jordan & Evermann, Fishes North and Mid. Amer., I, 433, 1896. Range. — Pacific coast of Mexico and Central America ; Galapagos Islands. We obtained this species at Seymour Island near Indefatigable and in Wreck Bay, Chatham Island. At the latter place it was found in immense schools. Family CHAULIODONTID^. 20. ZALARGES LUCETIUS (Garman). Matiroliciis lucefius GxKUW, Mem. Mus. Comp. Zool., xxiv, Rep. Kxpl. U. S. S. Albatross during 1 89 1, xxvi, Fishes, 242, pi. J, fig. 2, 1899, Alba- tross Station 3428, at 21° 36^ 30'^ N., 106° 25^ W. in 238 fathoms. SHORE FISHES OP^ GALAPAGOS ISLANDS 349 Ra>ige. — Panamic region of the Eastern Pacific, ranging vertically from 100 to 2,000 fathoms. One mutilated specimen taken from the stomach of a Thutiniis caught a few degrees north of the Galapagos Archipelago. It agrees with Garman's description of Matirolicus hiccthis^ but differs from the figure in the possession of an adipose fin. A few patches of thin cycloid scales are present on the caudal peduncle and hack. The species is close to Zalarges ni7nbarius Jordan & Williams, differing from it in the larger head, the shorter and deeper body, and in the larger ventral photophores, which are crowded and juxtaposed. We have examined the type of Z. nimbarius and find that it possesses a well developed adipose fin and a few large cycloid scales. Neither of these characters is shown in the figure of the type. The fins are not well preserved, but the anal has apparently 15 rays, as in Z. hccetius. On the strength of the absence of pseudobranchias and of the pres- ence of scales, we have placed this genus in the Chauliodontidos. It is closely related to the Maurolicidae, and it is doubtful whether these 2 families are really distinct. MEASUREMENTS OF ZttlaVgeS luCCthlS. Length in mm 31 Head 31 Depth 23 Eje 10 Snout 10 Maxillary 23 Interorbital width 05 Pectoral 15 Base of anal 19 Caudal 22 Depth of caudal peduncle 09 Length of caudal peduncle 12 Family HEMIRHAMPHID^. 21. HYPORHAMPHUS ROBERTI (Cuvier & Valenciennes). Hemirhamphus 7-oberti Cuvier & Valenciennes, Hist. Nat. Poiss., xxi, 24, 1886, Cayenne. — Jordan & Bollman, Proc. U. S. Nat. Mus., xii, 1899, 180 (James Island). Hyporhamphits roberti, Jordan & Evermann, Fishes North and Mid. Amer. , I, 721, 1896. Range. — Both coasts of tropical America ; Galapagos Islands. Taken at James Island by the Albatross. 350 SNODGRASS AND HELLER 22. HEMIRHAMPHUS SALTATOR Gilbert & Starks. Hemirhamfhus balao, Jordan, Proc. U. S. Nat. Mus., viii, 1885, 370 (Pan- ama); not of Le Sueur, He7nirhainphus 5a/Ai/c;r Gilbert & Starks, Mem. Cal. Acad. Sci., iv, 1904, 53, pi. IX, fig. 16, Panama. Range. — Panama ; Galapagos Islands. One specimen taken between Albemarle and Narboro islands near Tagus Cove, Albemarle. It was secured by Captaiij^ W. P. Noyes, who stated that a school of the Hemiramphids was pursued by some porpoises past the boat in which he was rowing and that this one leaped into the boat. Length 480 mm.; entire head in total length 37; depth in total length 12; lower jaw beyond tip of upper in total length 20; length of body from tip of upper jaw 382 mm. ; depth in length without lower jaw 15; head in length from tip of upper jaw 22; pectoral in length without lower jaw 16; last ray of ventral in head without lower jaw 45 ; pectoral in head without lower jaw 74; eye in head without lower jaw 19; interorbital space in head without lower jaw 23. D. 14; A. 11; P. 11; teeth of the upper jaw simple, conical ; those of the lower jaw tricuspid. We have also 8 young individuals of a Hemlrhaiuph7is^ about 70 mm. long, seined in the surf on the beach north of Tagus Cove, Albemarle, which apparently belong to this species. They are marked by a black lateral band from base of pectoral to middle of caudal peduncle; by 2 black dorsal lines on the back, one on each side of the median line ; and by numerous transverse black lines cross- ing the back, not reaching laterally the lateral bands but broken into 3 segments by the longitudinal dorsal lines; lower jaw about 4.5 in the total length. 23. EULEPTORIIAMPHUS LONGIROSTRIS (Cuvier). Hcmirha7nphus lottgirostris Cuvier, Regne Animal, Ed. 2, Vol. 2, 286, 1829, Pondicherry ; ibid.. 111. Poiss., pi. 98. — Cuvier & Valenciennes, Hist. Nat. Poiss., XIX, 52. — Gunther, Cat., vi, 276. ? Eideptorha^nphus brcvoortiCiWA., Proc. Acad. Nat. Sci. Phila. 1859, 131, no locality. Euleptorhainphus lonoirostris, Jenkins, Bull. U. S. Fish Comm. 1902, 434 (Hawaiian Islands). Range. — East Indies; Hawaiian Islands; Galagapos Islands. Two specimens taken at Hood Island, the first record of the species from the Eastern Pacific. We have compared them with specimens from the Hawaiian Islands. SHORE FISHES OF GALAPAGOS ISLANDS 351 MEASUREMENTS AND FIN RAYS OF Eideftorham^hiis loiigi- rostrts. No. Stanford University Museum Total length in mm Length without lower jaw in mm Lower jaw beyond upper jaw Head Depth Pectoral Eye: Head Number of dorsal rays Number of anal rays 12333 420 427 305 301 37 4' 15 15 8 5 29 29 30 29 24 23 22 21 We have also numerous young examples of a long-finned Hemiram- phid, apparently a EuleptorhatTiphus and perhaps Euleptorhamphus longirostris^ taken from the stomach of a horse mackerel ( TJninjiiis thyn?tus) about 7° 26' N. ; 100° 26' W., in December; and from the stomach of an ocean bonito ( Gymnosarda pela7nis) at about 4° 30' N. ; 87° W., in July. During July these young individuals were extremely abimdant between Clipperton and Cocos islands, but we saw no adults. Those collected are about 75 mm. long; depth about 7; sides flat and parallel; pectorals 3|^; beak very short, about 9 in body, but perhaps broken in all, entirely gone in many. Family EXOCCETID^. 24. EVOLANTIA MICROPTERA (Cuvier & Valenciennes). Exoccetus microptc7-us Cuvier & Valenciennes, Hist. Nat. Poiss., xix, 92, 1846. Evolantia microptera. Heller & Snodgrass, Proc. Wash. Acad. Sci., v, 1903, 189. — Jenkins, Bull. U. S. Fish Comm. 1902 (1903), 434 (Ha- waiian Islands). Range. — East Indies ; Hawaiian Islands ; Eastern Pacific north of Galapagos Islands. One specimen, 150 mm. long, taken at 4° N., 90° W., July 12, in warm water about 250 miles north of the Galapagos Islands. 25. EXOCCETUS VOLITANS Linnaeus. Exoccetus volitanslA'H'tiiEVS, Syst. Nat., Ed. x, 316, 1758. — Jordan & Ever- MANN, Fishes North and Mid. Amer., iii, Addenda, 2835 ^"^ 2836, 1898. Exoccetus evolans Linn/EUS, Syst. Nat., Ed. xii, 521, 1766 (based on Gronow); Giinther, Cat., vi, 282, 1866. Halocypselus evola7is, Jordan & Evermann, Fishes North and Mid. Amer., I, 729, 1896. Range. — Widely distributed in the tropics. 352 SNODGRASS AND HELLER Numerous specimens taken in the warm water between 4° N. and 21° N. and 90° W. and 116° W. Comparatively scarce in the cooler water of the Humboldt current immediately to the south of this region, and none seen about the Galapagos Islands. 26. EXONAUTES SPECULIGER (Cuvier & Valenciennes). Exocceius speculiger Cuvier & Valenciennes, Hist. Nat. Poiss., xix, 93, 1846. — GuNTHER, Cat., VI, 287, 1866. Exoccvtus volitans, Jordan & Evermann, Fishes North and Mid. Amer. , i, 734. 1896. Exonautes speculiger, Jordan & Evermann, Fishes North and Mid. Amer,, III, Addenda, 2835 ^"d 2836, 1898. Range. — Widely spread in the tropics. One specimen taken at about 17° 23' N. ; 114° 9' W., southwest from the Revillagigedo Islands. 27. CYPSILURUS XENOPTERUS (Gilbert). ExoccEtus xejtopteriis Gilbert, Proc. U. S. Nat. Mus. 1890, 58, Clarion Island. — Jordan & Evermann, Fishes North and Mid. Amer., i, 738, 1896. Exonautes xenopterus, Jordan & McGregor, Rept. U. S. Fish Comm. 1898, 275 (Clarion Island ; Moro Hermosa, Lower California). Cypsilurus xenopterus, Jordan & Evermann, Fishes North and Mid. Amer., Ill, Addenda, 2836, 1898. Range. — Eastern tropical Pacific. This species was first obtained by Dr. C. H. Gilbert from the stom- ach of a booby {Sula) at Clarion Island, and later by jSIr, R. C. jSIc- Gregor at the same locality and off Moro Hermosa, Lower California. 28. CYPSILURUS CYANOPTERUS (Cuvier & Valenciennes.) Exoccvtus cyanopterus Cuvier & Valenciennes, Hist. Nat. Poiss., xix, 98, 1846, Bahia, Rio de Janeiro. — Gunther, Cat., vi, 294. — Jordan & BoLLMAN, Proc. U. S. Nat. Mus. 1889, 180 (James Island). — Jordan & Evermann, Fishes North and Mid. Amer., i, 739, 1896. Exoccetus albidactylus Gill, Proc. Acad. Nat. Sci. Phila. 1863, 167, Carib- bean Sea. Cypsilurus cyanopterus, Jordan &« Evermann, Check-list Fishes, 323, 1896. — Jordan & Evermann, Fishes North and Mid. Amer., iii. Addenda, 2836, 1898. Range. — Caribbean Sea, coast of Brazil; Galapagos Islands. Taken at James Island by the Albatross (iSS7-'SS). Family MUGILID^. 29. MUGIL CEPHALUS Linnreus. Mugil cepltalus Linnaeus, Syst. Nat., Ed. x, 316, 1758, luirope. — Jordan & BoLLMAN, Proc. U. S. Nat. Mus. 1889, 180 (Chatham Island, Hood Is- land). — Jordan & Evermann, Fishes North and Mid. Amer,, i, 811, 1896. SHORE FISHES OF GALAPAGOS ISLANDS 353 Range. — Cosmopolitan. This widely distributed species was taken by the Albatross (1SS7- 'SS) at Chatham and Hood islands, Galapagos Archipelago. 30. MUGIL TIIOBURXI Jordan & Starks. Mugilt hoburni ]oKi:>A^ & Starks in Jordan & Evermann, Fishes North and Mid. Amer., i, 812, 1896, Galapagos Islands, Range. — Pacific coast of tropical America ; Galapagos Islands. Specimens taken in Tagiis Cove and Elizabeth Bay, Albemarle ; and in shallow lagoons of the mangrove swamps on the east coast of Narboro. We have numerous young individuals taken in a seine from the surf on a beach north of Tagus Cove, Albemarle. These are certainly the young of JMugil thoburni\ they have the same number of fin rays, and the black base of the pectoral, characteristic of the adult, is con- spicuously present in all. The head, however, is very strikinglv dif- ferent in shape from that of the adult, being compressed and deep rather than wide and depressed ; but the specimens present a perfect gradation in this respect from the smallest to the largest. In a speci- men 27 mm. long, the interorbital space is .22 of the length of the head; in a specimen 38 mm. long .25 ; in one 62 mm. long .31 ; and in an adult 159 mm. long .38. The ridges on the rows of scales are present on specimens 40 mm. long, but on specimens smaller than this they are not yet developed. The adipose eye-lids are slightly de- veloped in specimens 65 mm. long; the smaller ones have the eye- lids simple. 31. MUGIL CUREMA Cuvier & Valenciennes. Mugil cnreDta Cuvier & Valenciennes, Hist. Nat. Poiss., ix, 87, 1836, Brazil, Martinique, Cuba. — Jordan & Evermann, Fishes North and Mid. Amer., i, 813, 1896. — Jordan & McGregor, Rep. U. S. Fish Comm. 1898, 275 (Socorro Island). Range. — Common along both shores of tropical America, but known from only Socorro of the eastern Pacific islands. 32. MUGIL SETOSUS Gilbert. Mugil setosus Gilbert, Proc. U. S. Nat. Mus. 1891, 549, Clarion Island. — Jordan & Evermann, Fishes North and Mid. Amer., i, 815, 1896. Range. — Known from Clarion Island, and from the American mainland at Mazatlan. 354 SNODGRASS AND HELLER 33. CH^NOMUGIL PROBOSCIDEUS (Giinther). Mugil proboscideus GvyiTH'ER, Cat., in, 459, 1861, Cordova Island, Chcsnomugil proboscideus, Jordan & Evermann, Fishes North and Mid. Amer., i, 816, 1896. — Jordan & McGregor, Rep. U. S. Fish Comm. for 1898 (1899), 275 (Socorro Island). Range. — West coast of tropical America; Socorro Island. 34. QUERIMANA HARENGUS (Gunther). Myxus harengus Gunther, Cat., iii, 467, 1861, Pacific coast of Central America. Querima7ia harengus, Jordan & Evermann, Fishes North and Mid. Amer., I, 817. Range. — West coast of America from Mazatlan to Peru ; Galapa- gos Islands. Seven specimens from Tagus Cove, Albemarle, the first reported from the Galapagos. The largest is 38 mm. long. Family SPHYRiENIDiE. 35. SPHYR^NA IDIASTES Heller & Snodgrass. Sphyrcena idiastes Heller & Snodgrass, Proc. Wash. Acad. Sci., v, 1903, 190, pi. II, Seymour Island. Range. — Galapagos Islands. Two adults, one from Seymour the other from the north coast of Narboro ; several young from Tagus Cove, Albemarle, and from Hood. Family HOLOCENTRID.E. 36. MYRIPRISTIS OCCIDENTALIS Gill. Myripristis occideiitalis Gill, Proc. Acad. Nat. Sci. Phila. 1863, 87, Cape San Lucas. — Jordan & Evermann, Fishes North and Mid. Amer., i, 847, 1898. Range. — Cape San Lucas to Panama ; Cocos Island ; Galapagos Islands. Abundant at Cocos Island and at Galapagos Archipelago ; not know^n from the Revillagigedo Islands. Over 50 specimens taken at Duncan, Seymour, Barrington, Hood and Tower islands. We have compared these specimens with specimens of J/, occidentalis in the Stanford University collection from Panama. Since the species has not been well described we give the following Description of a typical specimen. — Head 3 ; depth 3.5 ; eye 2.5 in head; interorbital width 4.3 ; snout 5 ; maxillary 1.75. D. X-I, 14; A. IV, 13; scales 3-36-6. SHORE FISHES OF GALAPAGOS ISLANDS 355 Body elliptical, compressed posteriorly, dorsal and ventral outlines nearly equal ; head flattened above at interorbital region ; snout short and blunt; opercle with a small spine at angle ; suborbital, preopercle, both limbs of opercle, interopercle, lower edge of subopercle, shoulder- girdle and occiput with their edges serrate ; mouth oblique ; lower jaw included ; maxillary extending to vertical from posterior border of eye, posterior margin without serrations; supplemental maxillary mod- erate; teeth in villiform bands in both jaws, in a diamond-shaped patch on the vomer and in club-shaped patches on the palatines; gill- rakers lo -f 20, long, length slightly less than diameter of pupil; eye large, diameter 2 in snout. Dorsal fin divided, first part of 10 spines, separated from the second part by an interval equal to half interorbital width ; third to fifth spines longest; first spine slender, equal to sixth; last spine united with soft dorsal ; first soft rays longest, equaling longest spines ; anal spines shorter than height of soft anal ; first spine short, the third and fourth longest and of equal length, but the third much thickened ; soft anal similar to soft dorsal and of same height ; pectoral pointed, of 15 rays, the upper ones longest, 1.5 in head; ventral pointed, 1.6 in head; caudal rather deeply forked, the lobes equal and pointed, twice the diameter of the eye. Scales sharply serrated, those on the lateral line with enlarged ser- rations mesially. Cheek and opercle scaled, the former with 4 vertical rows ; lateral line continuous, parallel with contour of back. Color in life. — Cardinal on sides of body, becoming darker olive- red on snout and before spinous dorsal ; belly and throat lighter and more silvery-red; sides with faint longitudinal dusky stripes produced by the over-lapping of the scales ; fins red like the sides of the body, and without dark bands or spots. MEASUREMENTS OF Myripristis occidetiialis. No. Stanford University Museum Length in mm Head Depth Pectoral Ventral Eye Interorbital width Snout Maxillary Longest dorsal spine Longest anal spine 12317 152 34 41 22 21 14 7 7 20 16 12 12318 120 35 42 25 22 14 7 22 18 H 12319 131 34 38 24 21 15 8 7 21 iS 13 12320 124 34 39 24 21 IS 8 7 20 17 13 356 SNODGRASS AND HELLER 37. MYRIPRISTIS CLARIONENSIS Gilbert. Myripristis clarionensis Gilbert, Proc. U. S. Nat. Mus. 1896, 441, pi. 69, Clarion Island. — Jordan & Evermann, Fishes North and Mid. Amer., Ill, Addenda, 2842, 1898. —Jordan & McGregor, Rep. U. S. Fish Comm. 1898 (1899), 275 (Clarion Island, Socorro Island). Range. — Clarion and Socorro islands, Revillagigedo Archipelago. 38. MYRIPRISTIS MURDJAN (Forskal). Sciana miwdjan Forskal, Desc. Anim., 48, 1775, Red Sea. Myripristis im(7'djan, Ruppell, Atlas Nordl. Afr. , 86, pi. 23, fig. 2; Giinther, Fische der Siidsee, 11, 92, pi. 61, 1873-75. — Day, Fishes of India, 170, 1878-88, Supplement, 788, pi. 41, fig. 2. —Jenkins, Bull. U. S. Fish Comm. for 1902 (1903), 440 (Hawaiian Islands). Rafzge. — East coast of Africa ; Red Sea ; India ; Malay Archi- pelago ; Polynesia; Hawaiian Islands; Cocos Island; Galapagos Islands. We have numerous large specimens taken at Cocos Island, and one specimen taken at Duncan Island of the Galapagos Archipelago, the only record of the species from the Western Pacific. The descriptions of the species vary somewhat in regard to the coloration of the fins, and the development of the external teeth of the anterior end of the mandible and of the angles of the premaxillaries. Since the species is not elsewhere described in American literature we give the following description based on Cocos specimens : Description of a typical specimen. — Length 350 mm. Dorsal profile ascending from snout to nape at an angle of about 45°, here slightly angulated, running backward at greater inclination to front of spinous dorsal ; from here to front of soft dorsal almost horizontal, then somewhat abruptly curved downward to caudal peduncle ; both dorsal and ventral outlines of peduncle concave; ventral profile with about same amount of convexity as dorsal, but more evenly curved — the belly being less angulated than the back; mouth very oblique; snout transversely truncate, so that the upper jaw presents a wide, straight, anterior margin, causing each premaxillary to present laterally a prominent angle where the transverse and longitudinal parts meet; upper jaw considerably exceeded by tip of lower projecting part 2 in length of snout, 4 in horizontal diameter of eye (varying much in different specimens) ; maxillary reaching vertical from posterior border of eye; supplemental maxillary large and projecting a little beyond the posterior end of the maxillary ; angle of opercle with a short blunt spine; preopercle notched just above angle, interopercle notched near lower entl ; opercle, preopercle, subopercle, interopercle SHORE FISHES OF GALAPAGOS ISLANDS 357 below, notch, and posterior lower edge of maxillary serrate; gill- rakers 12 4-21, a little greater than 3 in eye. Angle of each premaxillary and each side of symphysis of lower jaw with an elevation bearing about i3 (number variable in differ- ent specimens) hard, papillar, tooth-like projections, forming in adults a very conspicuous character; teeth of inside of mouth small, in villi- form bands in both jaws, in a triangular patch on the vomer and in elongated club-shaped patches on the palatines; an outer series of enlarged blunt conical teeth in each jaw similar in shape and appear- ance to those on the exterior of the jaws. Dorsal fin with 1 1 spines, third and fourth longest, deeply emargi- nate before the last; soft dorsal elevated, somewhat falcate, much higher than the spinous dorsal, 14 rays, the anterior longest; anal spines 4, the first very short, the third thickened but shorter than the fourth, fourth longest; soft anal similar to the soft dorsal; caudal forked, the lobes equal and bluntly rounded ; pectoral rather pointed, rays 15, the upper longest ; ventral pointed, slightly longer than the max- illary ; scales large, sharply serrate, 2^-30-6; lateral line nearly hori- zontal anteriorly, with a slight upward curve on the caudal peduncle. Variations. — The external teeth of both jaws vary greatly in indi- viduals of different ages, being much smaller and fewer in numbers in the young than in adults. In a specimen 125 mm. long they are scarcely conspicuous on the upper jaw, and on the lower only 3 small teeth are present on each side. In some adults all 4 sets are large and prominent, while in others of the same size they are much smaller. The degree to which the lower jaw projects beyond the upper varies with the development of the external teeth. MEASUREMENTS OF Myn'prtstts murdjan. Length in mm Head Depth Pectoral Ventral Eye Interorbital width Maxillary Longest dorsal spine Longest dorsal ray... Longest anal ray Longest anal ray , 141 205 210 220 35 35 35 35 44 43 43 43 26 24 24 24 22 22 21 22 14 13 13 13 7 7 7 7 22 20 22 20 1 15 IS 14 15 21 19 17 17 14 13 11 10 21 19 iS 19 225 34 39 25 23 13 7 21 13 19 u 20 Color 171 life. — Above bright cherry-red, lighter on sides, fading to silvery-red on belly; opercular flap and axil of dorsal dusky olive- Proc. Wash. Acad. Sci., January, 1905. 358 SNODGRASS AND HELLER brown (dusky in alcohol) ; spines of dorsal like back, membranes yellowish, with bluish base ; soft dorsal and anal cherry-red with first ray white-edged, a black blotch at tip of first rays; caudal fin like back, with upper and lower rays white-edged, a dark blotch just within the tip of each lobe; pectoral like side; ventrals lighter red with the first ray white-edged; iris silvery and cardinal. The specific identification here given Is based on a direct comparison of the Cocos and Galapagos specimens with Hawaiian specimens. 39. HOLOTRACHYS LIMA (Cuvier & Valenciennes). Myripristis lime (Zv^ii¥.^ & Valenciennes, Hist. Nat. Poiss., vii, 493, He de France. — Cuvier, Regne animal, iii, Poiss., pi. 14, fig. 2. Myripristis lima, Gunther, Cat., I, 28, 1859. — Kner & Steindachner, Sitz. Akad. Wissen. Wien, Liv, 375, pi. i, fig. i, 1866 (Samoa). — Gun- ther, Fische der Siidsee, 93, pi. 63, fig. A, 1875 (Mauridus, Samoan, Hawaiian and Kingsmill islands). Holotrachys lima, Jenkins, Bull. U. S. Fish Comm., xxii, 1902 (1903), 439 (Honolulu). Ra7ige. — Isle of France, Mauritius, Samoan, Gilbert, Hawaiian and Cocos islands. Two specimens of this species secured in Chatham Bay at Cocos Island, the first taken in American waters. Description. — Head 2\ ; depth 2 ; eye 4 in head ; snout 4! ; inter- orbital 5^; maxillary i|; D. XII, 15; A. IV, 11; scales 5-40-9. Body ovoid, not much compressed ; profile roundly convex from snout to front of dorsal, from here to front of soft dorsal slightly descending posteriorly, then abruptly curved downward to caudal jDcduncle, mak- ing nearly a right angle with the dorsal edge of the latter; ventral profile similar to dorsal; mouth large, slightly oblique; lower jaw included and armed at symphysis with a prominent knob; upper q(\^q, of premaxillary on level with lower margin of eye ; angle of opercle armed with 2 short, rather stout, spines, of about equal length, 3 in eye; these spines followed below angle by smaller spines on mar- gins of opercle and subopercle ; preopercle with a short spine at angle; interopercle with a very small spine at its angle; subopercle, preopercle, opercle, occipit, interopercle, and shoulder-girdle with their edges serrate: lower edge of maxillary and of subopercle entire; gillrakers 8 -f 12, of moderate length, those at angle as long as the gill-filaments; teeth small, villiform in bands in both jaws, a trans- verse oval patch on vomer and club-shaped patches on palatines. Dorsal fin deeply emarginate before soft part ; spinous part begin- ning above base of pectoral, margin rounded, fourth and fifth spines highest, 3 in head, slightly less than height of soft dorsal ; spines het- SHORE FISHES OF GALAPAGOS ISLANDS 359 eracanthous ami depressible in a groove; soft dorsal higher and more convex than spinous dorsal, longest ray 24 in head; first anal spine very short, scarcely projecting beyond its sheath; third considerably enlarged, much stouter than fourth which it equals in length ; soft anal rounded and similar in shape to soft dorsal, but exceeding it slightly in height; caudal forked, the lobes rounded and equal; scales at base of caudal, above and below, forming sharp serrations; pectoral broad, somewhat rounded, of 17 rays of which the upper ones are longest. Scales closely imbricate, and armed on their posterior edges with long, slender, spine-like serrations, the bases of which project on the scales as low ridges ; spines of scales on posterior part of body longer than those of anterior parts, and almost obscuring the scales themselves ; caudal fin scaled on basal half, otherwise fin-membranes naked ; lateral line continuous and extending the entire length of body, concurrent with back anteriorly, less curved posteriorly. Variations. — The other specimen differs from the one just described in having 2 spines on the angle of the preopercle, longer serrations on the scales of the posterior part of the body, fewer (38) scales on the lateral line, and serrations on the posterior edge of the supplemental maxillary. Color in life. — Above bright cardinal-red, lower part of sides and belly silvery-red ; iris and fin-rays like the back, the membranes of the latter yellowish ; ventrals silvery-red like the belly. MEASUREMENTS, FIN RAYS AND SCALES OF HolotvachyS lima. Locality. Length in mm Depth Head Pectoral Ventral Eye Interorbital Snout Number of dorsal spines Number of second dorsal rays Number of second anal rays... Scales on lateral line Cocos Island. Hawaiian Islands. 112 107 107 105 52 39 47 38 36 42 36 20 22 21 21 17 21 20 20 9 10 10 II 7 9 7 ID 7 8 7 8 XII XII XII XII 15 15 15 15 II 12 12 II 40 43 44 42 The 2 Cocos specimens are greatly swollen, to which fact is prob- ably due the difference in depth between the Cocos and Hawaiian specimens. 360 SNODGRASS AND HELLER 40. HOLOCENTRUS SUBORBITALTS Gill. Holoceiitrtwi suborbitale Gill, Proc. Acad. Nat. Sci. Phila. 1863, 86, Cape San Lucas, Holocetitrus siiborbitalis, Jordan & Evermanx, Fishes North and Mid. Amer., I, 850, 1896. — Jordan & McGregor, Rep. U. S. Fish Comm. for 1898 (1899), 275 (Clarion and Socorro islands.) Range. — Cape San Lucas to Panama; Revillagigedo, Cocos and Galapagos islands. A very abundant species at Cocos Island but not so abundant at the Galapagos Islands. At the latter place taken at Albemarle, Charles and Tower. Three young individuals, ^5 mm. in length, secured in tidepools at Charles are scarcely different from the adults. The opercular spines are shorter and the coloration is more greenish. Family MULLID.E. 41. PSEUDUPENEUS DENTATUS (Gill). Upeneus dentatus Gill, Proc. Acad. Nat. Sci. Phila. 1862, 256, Cape San Lucas. — Jordan & Evermann, Fishes North and Mid. Amer., i, 859, 1896. — Jordan & McGregor, Rep. U. S. Fish. Comm. for 1898 (1899), 275 (Clarion Island). Range. — Cape San Lucas ; La Paz ; Tres Marias Islands ; Clarion Island. Family SCOMBRIDiE. 42. SCOMBER JAPONICUS Houttuyn. Scomber japonicus Houttuyn, Verh. nit. Holland. Mattsch. der Weet., xx, 2, 1872, Japan. Scomber colias Gmelin, Syst. Nat., Ed. xiii, 1329, 1788, Sardinia. — Jordan & Evermann, Fishes North and Mid. Amer., i, 866, 1896. — Aiujott, Proc. Acad. Nat. Sci. Phila. 1899, 344 (I'eru). Range. — Cosmopolitan. Four specimens taken in Tagus Cove, Albemarle, Galapagos Archipelago. 43. GYMNOSARDA PELAMIS (Linnccus). Scomber pelamis Linnaeus, Syst. Nat., Ed. x, 297, 1758. Gymnosarda pclamis, Jordan & Evermann, Fishes North and Mid. Amer., i, 868, 1896. Range. — Intertropical. Two specimens from Wennian Island, Galapagos Archipelago. The species is abundant in the wann cur- rents north of the equator as far north as the Revillagigedo Islands. Both this species and the horse mackerel ( Thurniiis t/iynnus) were SHORE FISHES OF GALAPAGOS ISLANDS 361 specially numerous from about 3° N. to 1 1 ° N., between Cocos and Clipperton islands. We observed them here for about 3 weeks during July in great numbers, the ocean appeared to be everywhere filled with them. They fed principally on a young Hemirhamphid which greatly exceeded its enemies in numbers and on all sides both fugitives and pursuers were to be seen leaping from the water. The bonitos and horse mackerels often throw themselves a considerable distance through the air. 44. THUNNUS THYNNUS (Linnaeus). Scojuber thyjimis Linnaeus, Syst. Nat., Ed. x, 297, 1758, Europe. Thunnus tkvnnus, Jordan & Evermann, Fishes North and Mid. Amer., i, 870, i8'96. Range. — Tropical and temperate seas. Two specimens from Wenman Island and one N. E. off Point Albemarle, Albemarle Island, Galapagos Archipelago. Common in warm water from about 3° N. to 11° N. (See under Gyninosarda pelamis.^ This fish is not uncommon in the Humboldt current about the Galapagos Islands, but does not occur in these colder waters in any such abundance as in the warmer currents a little farther north. Nearly all the individuals that we saw were each about 2 feet long. 45. GERMO ALALUNGA (Gmelin). Sconibcr alalu7iga Gmelin, Syst. Nat., i, 1330, 1788, Sardinia. Germo a/altmga, Jordan & Evermann, Fishes North and Mid. Amer., i, 871, 1896. Range. — Tropical and subtropical seas in general. We obtained one specimen of this species during August, at about 26° 10' N., 123° 25' W., but the species was not observed south of here where Gymnosarda pelaniis and Thunnus thynnus were so abundant, nor did we meet w^ith it about Clarion, Clipperton, Cocos or the Galapagos islands. 46. SCOMBEROMORUS SIERRA Jordan & Starks. Scomberomorus sierra Jordan & Starks in Jordan, Proc. Cal. Acad. Sci., 2d series, 428, 1895, Mazatlan. — Jordan & Evermann, Fishes North and Mid. Amer., i, 874, 1896. Range. — Pacific coast of tropical America; Galapagos Islands. One specimen taken in Tagus Cove, Albemarle. 362 SNODGRASS AND HELLER Family CARANGID.E. 47. ELAGATIS BIPINNULATUS (Qimy & Gaimard). Seriola bipinnulata OuOY & Gaimard, Voy. Uranie, Zool. , i, 363, pi. 61, fig. 3, 1824, Keeling Islands. Elagatis bipinnidatiis, Jordan & Evermann, Fishes North and Mid. Amer,, I, 906. Range. — Intertropical. One specimen taken in July between the Galapagos and Clipperton islands, 4° 30' N. ; 97° 30' W. Length 7S5 mm. 48. DECAPTURUS SCOMBRINUS (Valenciennes). Caranx scombrmus Valenciennes, Voyage de la Venus, 332, pi. 7, fig. i, 1846, Galapagos Islands. Decapterns hypodus, Jordan & Bollman, Proc. U. S. Nat. Mus. 1889, 180. Decapteriis sco7nbrinns, Jordan & Evermann, Fishes North and Mid. Amer., I, 908, 1896. Range. — Galapagos Islands. Three adult examples of this species, two taken at Tagus Cove, Albemarle, and the other at Hood Island ; and numerous young from Chatham Island. Since the only description extant of this species is that of Valenci- ennes we give the following description based on our specimens: Profile regularly fusiform ; caudal peduncle wide, depressed ; cau- dal fin deeply forked. Head 3|; depth \\. D. VIII-I, 37-1 ( = IX, 28) ; A. II— I, 23-1. Maxillary 3 in head, reaching front of eye; tip of lower jaw very slightly projecting; eye 4 in head, somewhat narrower than interorbital space, with concentric adipose e3elid be- fore, and one behind ; first dorsal spine about -^ of second, shorter than the third which is longest; the eighth very short; the ninth, detached from the others and united with the soft dorsal, \ length of first soft ray and of fifth spine; anterior rays of soft dorsal longest; anal with two detached spines, the first equal to one half of interorbital space, longer than the second ; the third attached to soft part of anal fin, slightly longer than the first; soft anal similar to soft dorsal; upper jaw almost toothless, there being only a few minute teeth in the front of the jaw; lower jaw with a series of larger, but very small teeth; vomer and palatines toothless ; a median longitudinal patch of small teeth on the tongue. Head and body everywhere scaly except before eyes, anterior five eighths of lateral line convex dorsallv, posterior part straight; its entire length marked by a row of prominent scales, each with a trifid tubule; scales of posterior part of lateral line enlarged, each bearing SHORK P-ISHKS OF f;ALAPAGOS ISLANDS 363 a rid^e-like spine, these increasing in size posteriorly, forming on caudal peduncle a high sharp lateral crest; spirje-bearing scales not more than thirty in number ('30 in one specimen, 28 in another). Posterior ones lapping over one another, forming a high hard longi- tudinal ridge along the side of the caudal peduncle; in front of the caudal peduncle they decrease regularly in size, and the most anterior is a mere rudimentary projection on the posterior margin of the scale the most anterior of the spine-bearing scales scarcely at all enlarged; posteriorly they increase in size with increase in the size of the spines. Valenciennes represents about 40 enlarged plates on the straight posterior part of the lateral line ; his figure, however, does not show the spines of the scales. Jordan and Evermann {Fishes N. & M. A.^ /, -pp. Q07, Q08 ) have grouped Decapterus sco7nbrinus with D. puJictatus, the 2 being characterized by having 40 or more enlarged plates on the lateral line. D. hypodus Gill, has 30 enlarged plates on the lateral line and is otherwise very similar to D. scombrinus, but it has been described by Gill {Proc. Acad. Nat. Sci. Phila. 1862, 26) and by Jordan and Gilbert {Proc. U. S. Nat. Mus. 1882., Js8) as being very closely related to D. macarellus. There can be no doubt, however, that hypodus and scombrlnus are very closely related and it is not improbable that hypodus (1S62) may prove to be a synonym of scombrlnus (1846) . D. hypodus is recorded only from Cape San Lucas. Color of fresh specimen. — Above, light olive-green, fading to silvery-white on the sides; fins dusky; upper half of caudal olive- yellow; dusky below; a dark semilunar spot on posterior margin of opercle above posterior angle. MEASUREMENTS, FIN RAYS AND SCALES OF Decaj}teruS SCOm- brinus. Length in mm 304 l^ Depth 23 22 Head 27 2S Pectoral 24 22 Eye ° , 7 Number of dorsal spines VIII-I VIII-I Number of second dorsal rays 33"' 33-' Number of second anal rays 26-1 27-1 Number of spine-bearing scales of lateral line 30 28 49. TRACHURUS SYMMETRICU.S (Ayres). Caranx symmetricus Avres, Proc. Cal. Acad. Sci., i, 1855, 62, San Francisco, Caranx {Trachurus) cuvieri Steindachner, Ichthyol. Beitr., 11, 16, 1875, Talcahuano, Callao, Juan Femandes, Galapagos. 364 SNODGRASS AND HELLER Trachurus pichirattcs^ Jordan & Evermann, Fishes North and Mid. Amer., i, 909, 1896; not of Bowdich. Range. — West coast of America; Galapagos Islands. Known from the Galapagos only through Steindachner's report. 50. TRACHUROPS CRUMENOPHTHALMA (Bloch). Scomber crumenophthalmus '^I.OCH, Ichthyol., 343, 1793, Guinea. Trachurops crumenophthalmiis, Jordan & Evermann, Fishes North and Mid. Amer., i, 911, 1896. — Jordan & McGregor, Rep. U. S. Fish Comm. 1898, 276 (Socorro Island). Range. — Atlantic and Pacific coasts of tropical America; West Indies ; west coast of Mexico ; Socorro Island, Revillagigedo Archi- pelago. 51. ZALOCYS STILBE Jordan & McGregor. Zalocys stilbe Jordan & McGregor, Rep. U. S. Fish Comm. 1898, 277, pi. 5, Clarion Island. — Jordan & Evermann, Fishes North and Mid. Amer., iii, Addenda, 2848, 1898. Range. — Clarion Island, Revillagigedo Archipelago. One specimen taken by Mr. R. C. McGregor. 52. CARANX CABALLUS (Giinther). Caranx cabalhis Gunther, Fishes Cen. Amer., 431, 1869, Panama. — Jor- dan & Evermann, Fishes North and Mid. Amer., i, 921, 1896. Range. — Pacific coast of tropical America ; Galapagos Islands. One specimen from James Island. A common species of the main- land coast. 53. CARANX MARGINATUS (Gill). Caranx marginatus Gill, Proc. Acad. Nat. Sci. Phila. 1866, 166, Panama. — Jordan & Evermann, Fishes North and Mid. Amer., i, 923, 1896. — Jordan & McGregor, Rep. U. S. Fish Comm. for 1898, 277 (Socorro Island). Range. — Mazatlan ; Panama; Socorro Island, Revillagigedo Archipelago. 54. CARANX LATUS (Agassiz). Caranx laius Agassiz, Pise. Bras., 105, 1829, Brazil. — Jordan & Bollman, Proc. U. S. Nat. Mus. 1889, 180 (Panama ; Chatham Island). — Jordan & Evermann, Fishes North and Mid. Amer., i, 923, 1896. Range. — Intertropical. Chatham Island, Galapagos Archipelago, taken only by the Alba- tross (1SS7-XSSS). SHORE FISHES OF GALAPAGOS ISLANDS 365 55. CARANX LUGUBRIS (Poey). Cara7ix lugubris PoEY, Memorias Hist. Nat. de Cuba, 11, 222, i860. — Jor- dan & EvERMANN, Fishes North and Mid, Amer. , i, 924. Range. — Intertropical. One specimen taken at Clarion Island, but the species was not met with at either Cocos or the Galapagos Islands. 56. CARANX MELAMPYGUS (Cuvier& Valenciennes). Caranx 7nela)npygus CvwEK & Valenciennes, Hist. Nat. Poiss., ix, 116, 1833, East Indies. — Jordan & Evermann, Fishes North and Mid. Amer., i, 925, 1896. Range. — Tropical parts of the Pacific Ocean. Obtained at Cocos Island, where it occurs in large numbers ; and one specimen secured at Tower Island, Galapagos Archipelago. 57. CARANX ORTHOGRAMMUS (Jordan & Gilbert). Caranx orihogravwms Jordan & Gilbert, Proc. U. S. Mus. 1881, 226, Clarion Island. Carangoides orthogrammus, Jordan & Evermann, Fishes North and Mid. Amer., i, 928. — Jordan & McGregor, Rep. U. S. Fish Comm. for 1898, 277 (Clarion and Socorro islands). Range. — Revillagigedo Archipelago. Known from Clarion and Socorro islands. Perhaps identical with Cara7ix ferdau, as suggested by Jordan and Evermann, and of general tropical Pacific distribution. Family CORYPHiENID^. 58. CORYPH^NA HIPPURUS LinnjEus. Coryphcena hippurus Linnaeus, Syst. Nat., Ed. x, 261, 1758, open seas. — Jordan & Evermann, Fishes North and Mid. Amer., i, 952, 1896. Range. — Pelagic : Atlantic and Pacific. One adult, S90 mm. long, taken at 13° 12' N. ; 111° 45' W., be- tween the Revillagigedo Archipelago and Clipperton Island ; also an- other of about the same size captured near Cocos Island, 4° 32' N. ; 89° 17' W. There is in the Stanford University collection a young Corypkcefia^ 185 mm. long, from Clarion Island, which is probably this species, although it is too young to show the characters of either hippurus or equisetis. There are about 50 rays in the dorsal and 22 in the anal. The profile before the eyes is convex but scarcely prominently ele- vated. 366 SNODGRASS AND HELLER Our adult example has the profile of the head rising almost vertical from the snout, forming a large, rounded prominent angle above a point midway betw^een snout and nostril, from which it runs steeply upward and backward to front of dorsal fin, which latter arises above posterior margin of eye ; maxillary reaching halfway from pupil to posterior rim of orbit. Head4|-; depth 5^ ; D. 56; A. I, 25. The color of the specimen when fresh was as follows : Above greenish-golden, below yellow, everywhere except on belly and throat spotted with purple spots as large as a pea and placed about one inch apart; head yellowish; iris black with a golden blotch on the ball above the pupil ; pectoral fins purplish above, yellow beneath and with tips purple ; ventrals yellow below, deep green above ; dorsal purplish, spotted with violet ; anal yellow, with a row of purple spots ; caudal greenish. 59. CORYPH^NA EQUISETIS Linnaius. CoryphcE7ia equisetis Linn^us, Syst. Nat., Ed. x, 261, 1758. — Jordax & EvERMAXN, Fishes North and Mid. Amer., i, 953, 1896. — Jordan & McGregor, Rep. U. S. Fish Comm. for 1898 (1899), 276 (San Bene- dicto Island). This species, like the last, is one of wide tropical distribution. Jor- dan and McGregor report one specimen from San Benedicto Island. Their specimen, however, is a young one, 9 inches long, and it may be doubted that it belongs to this species rather than to C hippurzis. Family NOMEID^. 60. GOBIOMORUS GRONOVII (Gmelin). Gobms gronovii, Gmelin, Syst. Nat., xiii, 1205, 1788, Tropical America. Nof/ietts gronovii, Jordan tS: Evermann, Fishes North and Mid. Amer., i, 949, 1896. Range. — Tropical parts of the Atlantic and Pacific. We have a few small specimens, about 35 mm. in length, taken with Portuguese men-of-war {P/iysalia) at about 7° 26' N. ; 100° 36' W. Family KUHLIIDiE. 61. KUHLIA T^NIURA (Cuv. & Val.). Duh's iccniiira Cuvier & Valenciennes, Hist. Nat. Poiss., iii, 114, 1829. Kuhlia argc Jordan & Bollman, Proc. U. S. Nat. Mus. 1889, 159, Chat- ham Island. — Jordan & McGregor, Rep. U. S. Fish Comm. for 1898 (1899), 277 (Clarion and Socorro islands). — Jordan & Evermann, Fishes North and Mid. Amer., i, 1014, 1896. Kuhlia tccniura, Boulenger, Cat. Fishes Brit. Mus., 2d Ed. i, 39, 1895. Range. — Indian Ocean, East Indies, Polynesia, Revillagigedo, Cocos and Galapagos islands. SHORE FISHES OF GALAPAGOS ISLANDS 367 This species was taken by the Albatross (iSSy-'SS) at Chatham Island and later (1888-89) '^^ Clarion Island. Mr. R. C. McGregor obtained it at both Clarion and Socorro islands. Finally we secured specimens of it at Cocos Island. It has not yet, however, been reported from the mainland coast. Family APOGONICHTHYIDiE. 62. AMIA ATRADORSATA Heller & Snodgrass. Apogoii atrado7-satus Heller & Snodgrass, Proc. Wash. Acad. Sci., v, 1903, 192, pi. Ill, Charles Island. Range. — Cocos Island; Galapagos Archipelago. Very common about the Galapagos Islands where secured at Chat- ham, Barrington, Charles, Seymour, Duncan, Hood, James, Albe- marle, Narboro and Tower. Distinguished from A. atricauda Jordan & McGregor by having the second dorsal tipped with black. 63. AMIA ATRICAUDA Jordan & McGregor. Apogon atricaudits Jordan & McGregor, Rep. U. S. Fish Comm. for 1898 (1899), 277, Socorro, Clarion and San Benedicto islands. — Jordan & EvERMANN, Fishes North and Mid. Amer., iii, Addenda, 2853, 1898. Range. — Revillagigedo Islands. 64. GALE AGRA PAMMELAS Heller & Snodgrass. Galeagra pamme/as Heller & Snodgrass, Proc. Wash. Acad. Sci., v, 1903, 193, pi. IV, Wenman Island. One specimen secured near Wenman Island, Galapagos. A deep sea form. Family SERRANIDiE. 6$. EPINEPHELUS ANALOGUS Gill. Epinephelus analogies Gill, Proc. Acad. Nat. Sci. Phila. 1863, 163, Panama. — Jordan & Bollman, Proc. U. S. Nat. Mus. 1889, 181 (Charles Is- land).— Jordan & Evermann, Fishes North and Mid. Amer., i, 11 52, 1896. Range. — Pacific coast of tropical America ; Galapagos and Re- villagigedo islands. This species was taken by the Albatross at Charles Island, and by Mr. R. C. McGregor at Clarion Island. (i6. EPINEPHELUS LABRIFORMIS (Jenyns). Serramis labriformis Jenyns, Zool. Voy. of Beagle, Fishes, 8, pi. 3, 1840, Galapagos Islands. 368 SNODGRASS AND HELLER Epinephehis labriformis, Jordan & Evermann, Fishes North and Mid. Amer. , I, 1155, 1896. Range. — Pacific coast of tropical America from Cape San Lucas to Ecuador ; Socorro, Clipperton, Cocos and the Galapagos islands. We have 9 specimens from Albemarle Island, Tagus Cove and Elizabeth Bay, Seymour Island, Cocos Island and Clipperton Island. It has been taken by the Albatross also at Charles and Indefatigable islands of the Galapagos group. 67. DERMATOLEPIS PUNCTATUS Gill. Dermatolepis pwictattcs Gill, Proc. Acad. Nat. Sci. Phila. 1861, 54, and 250, 1862, Cape San Lucas. — Jordan & Evermann, Fishes North and Mid. Amer., i, 1169, 1896. Epinephehis dermatolepis, Boulenger, Cat. Fishes Brit. Mus., 2d Ed. i, 256, 1895. Range. — Cape San Lucas, west coast of Mexico, the Venados, Revillagigedo, Cocos and Galapagos islands. We have specimens of this species from Clarion, Cocos, and the Galapagos islands. Specimens of different sizes vary somevs^hat in appearance. In small specimens about 30 mm. long, the tips of the ventrals reach slightly beyond the anus. In large specimens the anus is far behind the tips of the ventrals. A specimen from Barrington Island, 520 mm. in length, has the anus 26 mm. back of the ventrals, while one from Clarion Island 650 mm. long has a space of 83 mm. between the anus and the tips of the ventrals. In some the maxillary reaches considerably beyond the eye, in others it ends slightly hi front of the eye. The spots vary from simple circular black areas to irregular dark blotches surrounded by broad white marginal fields each having the same outline as the dark spot it encloses. (>%. MYCTEROPERCA XENARCHA Jordan. Mycteroperca xc7iarcha Jordan, Proc. Acad. Nat. Sci. Phila. 1887, 387, James Island, Galapagos Archipelago. — Jordan & Evermann, Fishes N. and Mid. Amer., i, 1180, 1896. — Ahuott, Proc. Acad. Nat. Sci. Phila. 1899, 348 (Peru). Epinephehis xenarchus, Boulenger, Cat. Fishes Brit. Mus., 2d Ed., i, 266. Range. — Vanados Islands, off the west coast of Mexico; Payta, Peru ; Galapagos Islands. Taken at the Galapagos only by the Hasslar Expedition. 69. MYCTEROPERCA OLFAX (Jenyns). Serraniis olfax Jenyns, Zool. of Beagle, Fishes, 9, pi. 4, 1840, Galapagos Archipelago. SHORE FISHES OF GALAPAGOS ISLANDS 369 Epinephelus olfax, Boulenger, Cat. Fishes Brit. Mus. , 2cl Ed., i, 263. Myctcropen'a olfax, Jordan & Everm.\nx, Fishes North and Mid. Amer., I, 1183, 1896. Range. — Panama, Cocos Island and the Galapagos Islands. Our specimens, about 30 in number, are from Cocos, Albemarle, Tagus Cove and Elizabeth Bay, Narboro, Duncan, Barrington, Wen- man and Culpepper. About Albemarle, Narboro, Wenman, and Culpepper the species is extremely abundant, but very rare about the southern and eastern Galapagos Islands. It is an excellent food-fish, the individuals often associate in large schools, and are easily taken with a hook and line. The typical coloration of the species is as follows: above dark oli- vaceous brown spotted with purplish and lighter brown ; sides of head same ; belly grayish-brown ; maxillary and lower jaw lighter olive- brown; fins dusky ; iris golden with brown mottlings. Seven specimens 65 to 82 mm. in length and one 133 mm. long, are colored plain dark brown, lighter below, with the fins dusky, the soft dorsal and anal and the caudal pale-edged, there being no trace of spots on the body. These specimens are from shallow sandy lagoons at jNIangrove Point, Narboro. Two specimens, 165 mm, long, and one 195 mm. long, from Elizabeth Bay, Albemarle, have the body covered with faint circular dark brown spots most distinct on the paler ventral half of the body. These were taken also in shallow sandy lagoons. Another, from much deeper water at Tagus Cove, 190 mm. long, is spotted above and below, although the spots on the dorsal half are obscured by the dark brown color. Adults again lose the spots with age, and become of a plain brown color, but the age at which the spots disappear varies, large specimens being often very distinctly spotted. Specimens 65 and 82 mm. in length have the posterior nostril no larger than the anterior, and the anterior nostril provided with a mem- branous tube. In specimens 230 mm. long the posterior nostril is the size of the anterior. In specimens 350 mm. long the posterior nostril has almost twice the dorso-ventral diameter of the anterior and is much elongated in the same direction. Specimens 4S0 mm. in length have the anterior nostril about two-fifths as wide dorso-ventrally as the pos- terior, the latter being somewhat semilunar, embracing the anterior nostril. The very small specimens have the tips of the ventral fins reaching only slightly past the anus, being the same in this respect as specimens a foot long. Beyond this size the ventrals lengthen more rapidly than the body. 370 SNODGRASS AND HELLER 70. MYCTEROPERCA RUBERRIMA (Jordan & Bollman). Mycteroperca olfax ruberriina Jordan & Bollman in Jordan & Eigenmann, Review Serrattidcs, Bull. U. S. Fish Comm., viii, for 1888, 367. — Jor- dan & EvERMANN, Fishes North and Mid. Amer., i, 1183, 1896. Range. — Galapagos Archipelago. Associating with the brown M. olfax., yellow individuals are fre- quently met with exactly resembling the others in all respects except color. It is hence doubtful whether these individuals represent a dis- tinct species or whether they form a chromatic variety of M. olfax. We have such specimens from Wenman, Culpepper, Albemarle, Tagus Cove, and Duncan islands. They are colored in life as follows : Sides of body chrome-yellow, back orange, lower parts lighter lemon- yellow ; head orange-yellow; fin-membranes chrome-yellow; iris in some specimens orange, in others carmine. The specimen from which Jordan and Bollman described their variety Mycte7-operca olfax ruberrima (named ruberritna because supposed by them to have been red in life), is one of these yellow individaals. We have no evidence indicating whether it is a deep- water variety of M. olfax or not. No very young yellow examples nor any of an intermediate or mixed coloration were seen. Along the shore the yellow individuals are not nearly so common as the brown M. olfax., but where the latter was most abundant there occurred the greatest number of M. ruberrima. The type of M. ruberiina was taken by the Albatross at Abingdon Island, Galapagos, and similar specimens have not been found outside of the Archipelago. 71. CRATINUS AGASSIZII Steindachner. Craiinus agassizii Steindachner, Ichth. Beitr., vii, 19, 1878, Galapagos Islands. — Jordan & Evermann, Fishes North and Mid. Amer., i, 1 189, 1896. Serranus agassizi, Boulenger, Cat. Fishes Brit. Mus., 2d Ed., i, 282, 1895. Range. — Galapagos Archipelago. Known only from the Galapagos Islands. It is not a common fish. We have 7 specimens from Tagus Cove, Albermarle, Narboro, Bar- rington, and Seymour. 72. PARALABRAX ALBOMACULATUS (Jenyns). Serranus albomaciilatKs Ji:nvns, Zool. Beagle, Fishes, 3, pi. 2, 1840, Gala- pagos Archipelago. Paralahrax alboiitaciclatus, Jordan & Bollman, Proc. U. S. Nat. Mus. 1S89, 181 (Albemarle Island, Charles Island). SHORE FISHES OF GALAPAGOS ISLANDS 371 St'rra7i7is htdneralis Boulengkr, Cat. Fishes Brit. Mus., i, 2d Ed., 278, 1895. Paralabm.x humcralis, Jordan & Evekmann, Fishes North and Mid. Amer., I, 1 196, 1896. Paralabrax albomandatiis, Abbott, Proc, Acad. Nat. Sci. Phiia. 1899, 348. Range. — Galapagos Archipelago. If this species is distinct from Paralabrax hu7neralis^ it is known only from the Galapagos Islands. We have numerous specimens from Tagus Cove, Albemarle, and from Harrington. The synonymy given above shows that this species has been by some authors regarded as distinct, while others have included it in P. hmneralis. Abbott (1S99), comparing some young examples from Callao, Peru, with young examples from the Galapagos Islands in the Stanford University collection, concluded that the 2 should be retained as separate species. We have examined these same speci- mens, together with the numerous adults that we obtained at the Galapagos Islands, and likewise cannot regard the specimens from the 2 localities as the same species, although if adult material from the mainland were at hand this might be possible. MEASUREMENTS AND FIN RAYS OF : No. Stanford University Museum. Length in mm Head Depth Eye I n terorbi tal Width of base of pectoral Number of pectoral ravs Number of dorsal spines Number of second dorsal rajs Number of second anal rajs.... Paralabrax albomaculatus from Galapagos Islands. 374 196 39 29 6 5^ 17 X IS 7 39 32 8 5^ 6 17 X 13 7 7 12265 145 375 365 39 42 40 30 3i 31 H 6 5i ■; 8 H .<;* 6 6 16 17 17 X X X 14 7 12266 12267 12268 14 7 290 39 31 6i 1\ 6i 17 X 14 7 372 38 26 5\ 7 6 17 X 14 7 Paralabrax humeralis from Callao, Peru. 11958 11958 11943 11943 178 28 5^ 5i ^h 18 X 13 7 175 41 27 6 5^ Si 18 X 13 7 148 40 29 6 Si 5i 18 X 13 7 IS7 +0 28 6 6 S\ 19 X 13 7 The above figures show that the Galapagos specimens, representing Paralabrax albomaculatus^ compared with specimens of approxi- mately the same size from Peru, that may be taken to represent the typical P. humeralis of Cuvier & Valenciennes (their specimen being from Chile) have a somewhat larger eye, a wider interorbital space, a wider base for the pectoral fin, fewer pectoral rays, and a greater number of dorsal rays. The figures show also that in the adult of P. albomaculatus the interorbital space is proportionally much wider and the eye smaller than in the young. Other characters of the species are as follows : Snout rather pointed, 372 SNODGRASS AND HELLER length of snout from nostril 3^ in head; mouth not very oblique ; maxillary 2^ in head; posterior edge of preopercle finely serrated, the serrations coarser on the lower limb, first dorsal spine shortest, a little more than i of eye; second f longer than first; third longest, about equal to maxillary; the following spines rapidly shorter to seventh, beyond which they are about equal ; caudal peduncle about twice eye in depth ; posterior margin of caudal fin slightly lunate, upper angle a little prolonged, the lower rounded. Color. — Above dark olive-brown ; belly grayish ; several large gray- ish spots along sides (a very characteristic mark of the species, absent in P. humeralis) ; opercles bronze-yellow, lower lip light yellow ; pec- torals bronze above, dusky beneath ; dorsal spines dusky, membrane saffron ; caudal grayish dusky, angles with orange tips ; posterior rays of anal orange-tipped, the rest grayish dusky ; ventrals pale dusky with rays orange-tipped. 73. PRIONODES FASCIATUS Jenyns. Prionodes fasciatus Jenyns, Voy. Beagle, Fishes, 46, 1840, Chatham Island. — Jordan & Evermann, Fishes North and Mid. Amer., i, 121 3, 1896. Serranus psittacimcs, Boulenger, Cat. Fishes Brit. Mus., 2d Ed., 295. Range. — Pacific Coast of Mexico, Panama, Revillagigedo Islands, Galapagos Islands (Chatham, Hood, Indefatigable, Albemarle). 74. PRIONODES STILBOSTIGMA Jordan & Bollman. Prionodes stilbostigma Jordan & Bollman, Proc. U. S. Nat. Mus. 1889, 158 (Galapagos Archipelago). — Jordan & Evermann, Fishes North and Mid. Amer., i, 12 16, 1896. Raitge. — Galapagos Archipelago. One specimen dredged in 45 fathoms by the Albatross (iSSy-'SS) near Hood Island, 0° 50' S. ; 89° 36' W. 75. PARANTHIAS FURCIFER (Cuvier & Valenciennes). &rra;n/.y /«;r//"6'r CuviER & Valenciennes, Hist. Nat. Poiss., 11, 264, 1828, Brazil. Serranus coloniis Valenciennes, Voyage Venus, ZooL, 300, pi. 2, lig. i, 1846, Galapagos Islands. Parantliias fiircifcr, Boulenger, Cat. Fishes, Brit. Mus., 2d Ed., I, 273. — Jordan & Evermann, Fishes North and Mid. Amer., i, 1221, 1896. Range. — Both coasts of tropical America; Revillagigedo and Galapagos islands. This fish is very abundant about tlic Galapagos Islands. We have specimens from Tagus Cove, Albemarle; James; Charles; Seymour; SHORE FISHES OF GALAPAGOS ISLANDS 373 Chatham; Duncan, and Hood. At Cocos it was not seen. It is known from Clarion, Socorro and San Benedicto ishmds of the ReviHagigedo Archipelago. 76. PRONOTOGRAMMUS MULTIFASCIATUS Gill. Pro7iotogrammus imdtifasciattis Gill, Proc. Acad. Nat. Sci. Phila. 1863, 81, Cape San Lucas. — Jordan & Evermann, Fishes North and Mid. Amer., I, 1226, 1896. — Jordan & McGregor, Rep. U. S. Fish Comm. for 1898, 278 (Clarion Island). Anthias multifasciatus, Boulenger, Cat. Fishes Brit. Mus., 2d Ed., i, 323, 1895. — Garman, Mem. Mus. Comp. Zool., xxiv. Rep. Expl. U. S. S. Albatross ^^xx\v\.^ 1891, xxvi, Fishes, 47, 1899. Range. — Cape San Lucas, Lower California ; ReviHagigedo and Cocos islands. Taken by Mr. R. C. McGregor at Clarion Island, and dredged near Cocos Island by the Albatross (Agassiz Expedition, 1891) in 66 fathoms. 77. RYPTICUS BICOLOR (Valenciennes). Smecticus bicolor Valenciennes, Voyage de la Venus, Poissons, 307, pi. i, fig. 2, 1855, Galapagos Islands. Rypticus bicolor, Jordan & Evermann, Fishes North and Mid. Amer., i, 1 23 1, 1896. Range. — Galapagos Archipelago. This species is known only from the description and figure of Va- lenciennes, whose specimen was taken at the Galapagos Islands by the Venus. We did not meet with the species. Family PRIACANTHID^. 78. PRIACANTHUS CRUENTATUS (Lacepede). Labriis cruentatus Lacepede, Hist. Nat. Poiss., in, 522, 1800, Martinique. Priacanthiis Carolines, Jordan & Evermann, Fishes North and Mid. Amer., Ill, Addenda, 2858, 1898. — Jordan & McGregor, Rep. U. S. Fish Comm. for 1898 (1899), 278 (Socorro and Clarion islands). Priacanthus cruentatus, Jordan & Evermann, Fishes North and Mid. Amer., I, 1238, 1896. Range. — West Indies; ReviHagigedo, Cocos and Galapagos islands. We have 14 specimens of a Priacanthus taken at the Galapagos and Cocos islands, which we cannot distinguish from the West In- dian species, P. cruentatus. It was most abundant at Cocos Island where most of our specimens were obtained, but we have i speci- men from Tagus Cove, Albemarle, and 2 from Barrington of the Galapagos. We have further compared these specimens with those Proc. Wash. Acad. Sci., January, 1905. 374 SNODGRASS AND HELLER from the Revillagigedo Archipelago identified by Jordan and McGre- gor as Priacanthus caroliniis^ and can find no constant difference among the specimens from the 3 localities. The length of the preopercular spine varies, in some it reaches the edge of the opercle and in others it does not. In a specimen in the Stanford University collection from Jamaica, 190 mm. long, the spine is short, not reaching the edge of the opercle, while in a speci- men from Bahia 93 mm. in length, the tip of the spine projects be- yond the opercle. The depth also varies, the smaller specimens being the more slender and duplicating the proportions of the Jamaica speci- men (see below). Color in life. — Bright metallic red, darkening to crimson on back ; head and belly silvery-red ; membrane of spinous dorsal blotched with olive and red; soft dorsal and ventrals red ; anal crimson with faint dusky bars, black-tipped ; pectorals pale red ; lips and snout olive- red; iris red. (Barrington specimen.) MEASUREMENTS, FIN RAYS AND SCALES OF Priacailt/lUS cruentatus. Locality. No. Stanford University Museum Length in mm Depth Head Eye Pectoral Ventral Number of dorsal spines Number of second dorsal rays Number of second anal rays Scales on lateral line tj s u 0 a (A 0 0 0 'ol CI •c 0 0 e .0 u 0 0 < >— > 12327 12328 12329 12330 4816 245 240 215 200 195 40 39 37 37 37 33 33 35 33 34 13 13 14 13 14 18 18 19 18 19 20 21 21 19 22 X X X X X 13 13 13 13 13 14 14 14 14 14 63 67 60 62 62 1596 92 43 37 15 20 24 X 13 14 66 79- Family LUTIANIDiE. LUTIANUS VIRIDIS (Valenciennes). Diacope viridis Valenciennes, Voyage de la Venus, 303, pi. i, fig. 2 (poor representation of the species), 1845, Galapagos Islands. Evoplitis viridis, Jordan & Evermann, Fishes North and Mid. Amer. , ii„ 1246, 1898. Range. — Tres Marias, Revillagigedo, Cocos and Galapagos islands. SHORE FISHES OF GALAPAGOS ISLANDS 375 We have 50 specimens taken at James, Tower, Banington and Seymoui- islands of the Galapagos Archipelago and at Cocos Island. The species was described by Valenciennes from specimens taken at the Galapagos Islands and has since been taken at the Tres Marias and Revillagigedo islands. We found the species very abundant at James, Seymour and Cocos, but about Tagus Cove, Albemarle, where most species were numer- ous, we saw only a very few and were not able to secure any specimens. 80. LUTIANUS JORDANI (Gilbert). Neomcenis jordani Gilbert in Jordan & Evermann, Fishes North and Mid. Amer., 11, 1251, 1898, Panama. Lutianus jordani, Gilbert & Starks, Mem. Cal. Acad. Sci., iv, 1904, 102. Range. — Panama, Cocos Island. Seventeen specimens from Cocos Island. They are about 230 mm. in length. The vomerine teeth form a diamond-shaped patch of which the posterior sides are considerably longer than the anterior. 81. LUTIANUS ARGENTIVENTRIS (Peters). Mesoprion argentive?itris Peters, Berlin. Monatsbr. 1869, 704, Mazatlan. Neomcenis argentivenfris, Jordan & Evermann, Fishes North & Mid. Amer., II, 1260, 1898. Lutianus argentivenfris, Gilbert & Starks, Mem. Cal. Acad. Sci., iv, 1904, 103. Rajige. — Pacific coast of tropical America ; Cocos and Galapagos islands. Six small specimens from Chatham, about 55 mm. long, have a dis- tinct bright blue subocular band reaching from below anterior part of eye almost to posterior angle of opercle, and a dark postocular band reaching from eye as far as the beginning of lateral line. Specimens in the Stanford University collection from Guaymas, 130 mm. long, have the subocular band somewhat broken posteriorly while the upper one is obsolete. In specimens up to this size the eye is on a line with the snout and the angle of the opercle, but in specimens 400 mm. in length, the eye is considerably above a line connecting these 2 points. The subocular band in specimens of this size is either absent or apparently represented by a row of spots which bends downward in front of the eye and extends forward toward the snout on a lower level. 82. XENOCYS JESSIE Jordan & Bollman. Xenocys jessice Jordan & Bollman, Proc. U. S. Nat. Mus. 1889, 160, Charles Island. — Jordan & Evermann, Fishes North and Mid. Amer., 11, 1285, 1898. 376 SNODGRASS AND HELLER Range. — Galapagos Archipelago. Known only from the Galapagos Islands. Our specimens from Tagus Cove, Turtle Point and Elizabeth Bay, Albemarle ; Tower ; South Seymour and Mangrove Point, Narboro. The species is gen- erally found in large numbers when met with at all, forming schools of great size. We did not often obtain specimens, but sometimes with I discharge of dynamite we killed many hundreds of them. 83. XENICHTHYS AGASSIZI Steindachner. Xenichthys agassizi Steindachner, Ichth. Beitrage, in, 6, 1875, Galapagos Islands. — Jordan & Evermann, Fishes North and Mid. Amer., 11, 1287, 1898. Range. — Galapagos Archipelago. This fish, like the last, which it resembles, is known only from the Galapagos Archipelago. It is rather rare. We obtained specimens only at Tagus Cove, Albemarle, and at Mangrove Point on the east side of Narboro. Family H^MULID^. 84. ANISOTREMUS SURINAMENSIS (Bloch). Lutjanus surinamensis 'Q'LOCU, IchthyoL, 253, 1791. Surinam. Anisotremus bilmeatus, Jordan & Bollman, Proc. U. S. Nat. Mus. 1889, 181 (Hood and Indefatigable islands). Anisotreimis biterruptiis (in part), Jordan & Evermann, Fishes North and Mid. Amer., 11, 1319, 1898. Anisotremus surinamensis, Jordan & Evermann, Fishes North and Mid. Amer., 11, 1318, 1898. Range. — Eastern shore of tropical America; Galapagos Islands. The 2 species, AnisotreiJiiis stirinamensis and A. ititerruptjis^ are distinguishable from each other by scarcely any other character than the presence in A. stcrinamensis generally of 9 rows of scales in an oblique series from the first dorsal spine to the lateral line, and the presence generally of only J in A. mte}-rupttis. The i species in- habits the eastern shores of tropical America and the other the western shores and the Revillagigedo Islands. The Galapagos specimens very curiously appear to belong to the former species rather than to the latter, the number of scales in an oblique row in the majority of our specimens being 9, in a few 8. This same fact is stated by Jordan and Evermann as being true of their specimens. Tlie}-^, however re- gard the Galapagos specimens as belonging to the west coast species. Our specimens are from Tagus Cove and Elizabeth Bay, Albe- marle ; Charles and Seymour. The species was taken by the Albatross also at Hood and Indefatigable. SHORE FISHES OF GALAPAGOS ISLANDS 377 Color in life. — Above silvery dusky, shading into golden about caudal peduncle; mouth whitish; pectoral dusky at base, fin amber; ventrals olive ; caudal olive w^ith the rays tipped with orange. 85. ANISOTREMUS INTERRUPTUS (Gill). Genytreimis interruptus Gill, Proc. Acad. Nat. Sci. Phila. 1861, 256, Cape San Lucas. A7iisotrevius interruptus, Jordan & McGregor, Rep. U. S. Fish Comm. for 1898 (1899), 278 (Clarion and Socorro islands). — Jordan & Evermann, Fishes North and Mid. Amer., 11, 13 19, 1898. Range. — Magdalena Bay to Panama ; Clarion and Socorro islands of the Revillagigedo Archipelago. We have compared Revillagigedo specimens with the Galapagos specimens of Anisotreinus sjirinamensis and find that the 2 are cer- tainly different in the respect stated under A. surinainensis. 86. ANISOTREMUS SCAPULARIS (Tschudi). Pristipoma scapulars Tscnvxii, Fauna Peruana, 12, 1844, Huacho. Aiiisotremus scapularis, Jordan & Evermann, Fishes North and Mid. Amer., II, 1320. Range. — Mazatlan (.^), coast of Peru, Cocos and Galapagos islands. We have 27 specimens of this species taken at the Galapagos Islands and one from Cocos Island. The Galapagos specimens are from Tagus Cove, Albemarle; James; Duncan; and Hood. This is the first record of the species from the Galapagos Archipelago. The specimens differ from 2 of Anisotremus scapularis in the Stanford University collection from Callao, Peru, merely in being of a general darker shade of color on the upper half, and in having the black of the axis and base of the pectoral not so dark and less definitely outlined. Color in life of a typical Galapagos specimen. — Above silvery greenish-blue, below white; fins dark bluish-gray; a few broad indistinct dusky vertical bands on sides ; preopercles and snout pur- plish ; iris white or silvery. The bands on the sides are not very definite markings and disappear soon after the fish is taken from the water. 87. ORTHOPRISTIS FORBESI Jordan & Starks. Orthopristis forbcsi Jordan & Starks in Gilbert, Proc. U. S. Nat. Mus. 1896 (1897), 443, Albemarle Island. — Jordan & Evermann, Fishes North and Mid. Amer., 11, 1336, 1898. 378 SNODGRASS AND HELLER Range. — Galapagos Archipelago. Described from 2 specimens taken at Albemarle. We have lo specimens from Chatham and one from Charles. Orthopristis forbesi., according to Jordan and Starks and Jordan and Evermann, differs from the other American species of the genus principally in having smaller scales and a greater number of them — 80 to 95 in longitudinal series. Our specimens, how^ever, all about a foot long, do not agree in this respect with the type description — the number of scales varying from 66 to 78 {^6., 6'j^ 68, 70, 72, 73, 74, 75, 76, 78). The dorsal rays are XII, 15; XII, 16; XIII, 15; anal rays III, 11 ; III, 12. Color in life. — Above dusky brownish; belly and sides grayish with faint purplish and bluish-gi-een iridescence ; several faint dusky vertical bars on sides, disappearing soon after death ; lips grayish- white; membrane of spinous dorsal livid bluish, spines blue-green; membrane of soft dorsal dusky ; pectoral pale dusky ; caudal dark with pale border posteriorly; opercular flap dark brown; iris brown. A specimen in the Stanford University collection, taken by the Albatross at Albemarle, has faint spots on the membranes of the soft dorsal as described by Jordan and Starks for the type. None of our specimens shows these spots. 88. ORTHOPRISTIS LETHOPRISTIS Jordan & Fesler. Orthopristis lethopristis Jordan & P^esler, Proc. Acad. Nat. Sci. Phila. 1889, 36, Galapagos Islands. — Jordan & Evermann, Fishes North and IMid. Amer., 11, 1340, 1898. Range. — Galapagos Archipelago. Heretofore only the type of this species known. We have 4 speci- mens, the longest 330 mm. in length, from Tagus Cove, Albemarle, and Duncan. All of our specimens have a slight serration on the bony posterior margin of the preopercle, being not " strictly entire" as described by Jordan and Evermann. Each has a wide band of scales on the mem- brane of the soft dorsal and of the anal back of each ray reaching to the margin of the fin. MEASUREMENTS, FIN RAYS AND SCALES OF Orthop>7'istis Ictho- prtstis. Length in mm Depth Dorsal spines Second dorsal rays. Second anal rajs. Scales on lateral line 1 70 265 270 265 35 34 36 XII XII XIII 15 16 15 11 8 II 70 64 69 320 32 XII 16 II 69 SHORE FISHES OF GALAPAGOS ISLANDS 379 Color in life. — Above grayish-olive, bluish iridescence before dorsal fin, greenish on sides of back ; sides dusky grayish ; belly lighter brownish-gray ; center of each scale dusky brown, the spots forming longitudinal streaks following the rows of scales, most dis- tinct above the lateral line; snout and sides of head dusky-olive with greenish iridescence; opercular flap dark brown or black; lips livid grayish; iris golden; fins dusky, spines bluish, rays of soft dorsal and of anal greenish; caudal dark at tip, lighter olive in middle; pec- torals and ventrals dark ; their rays dark bluish-gray. 89. ORTHOPRISTIS CHALCEUS (Giinther). Prisfipoma chalcciim Gunther, Proc. Zool. Soc. Lend. 1864, 146, Panama. Orthopristis chalceus, Jordan & Bollman, Proc. U. S. Nat. Mus. 1889, 181 (Albemarle, Chatham, Charles islands). — Evermann & Jenkins, Proc. U. S. Nat. Mus., XIV., 1891, 149 (Guaymas). — Jordan & Evermann, Fishes North and Mid. Amer., 11, 1337, 1898. Range. — Pacific coast of tropical America; Galapagos Archi- pelago. Taken at Charles, Chatham, and Albemarle by the Albatross. 90. ORTHOPRISTIS CANTHARINUS (Jenyns). Pristipoma cantharinum Jenyns, Zool. of the Beagle, Fishes, 49, 1842, Gala- pagos Islands. Orthopristis cantha7'inus, Jordan & Evermann, Fishes North and Mid. Amer., II, 1339, 1898. Range. — Galapagos Archipelago. Obtained by Darwin and by the Hasslar Expedition. Family SPARID^. 91. CALAMUS TAURINUS (Jenyns). Chrysophrys taurina Jenyns, Zool. of Beagle, Fishes, 56, pi. vii, fig. 12, 1842, Galapagos Islands. Clirysophrys cyanopfcra Valenciennes, Voyage Venus, v, pi. 4, fig. 2, 1846, Charles Island. Calamus tauriiius, Jordan & Bollman, Proc. U. S. Nat. Mus. 1889, 181 (Charles Island). — Jordan & Evermann, Fishes North and Mid. Amer., 11, 1354, 1898. Range. — Peru ; Galapagos Archipelago. We have 10 specimens from Mangrove Point, Narboro, and from Seymour. Not common anywhere. A number were taken with hook and line in about 15 fathoms off the west shore of the southern Seymour Island, but only i or 2 specimens taken with dynamite in shallow water. 38o SNODGRASS AND HELLER 92. ARCHOSARGUS POURTALESII (Steindachner). Sargus ^orfa/t^sn Stki'SDACHSER, Fische Afrikas, 39, 1881, Galapagos Islands. Archosargtcs pourtalesii, Jordan & Bollmax, Proc. U. S. Nat. Mus. 1889, 181 (Chatham Island). — Jordan & Evermann, Fishes North and Mid. Amer., 11, 1360, 1898. Range. — Galapagos Archipelago. We have 10 specimens all taken at Elizabeth Bay, Albemarle. The Albatross obtained one specimen at Chatham. Body very deep, much compressed, dorsal profile evenly convex from snout to base of caudal peduncle, a rounded elevation before eyes; mouth much below the longitvidinal axis of the body, lower profile evenly rounded but less convex than dorsal profile; pectoral long and narrow, base as wide as eye, upper rays longest ; caudal deeply forked ; dorsal spines all short, those back of the third, of equal length, the first 3 shorter, the first shortest, the second and third of equal length, the second much thicker than the third ; enlarged incisors in front of jaws f, wide, flat, constricted at base; smaller molar teeth on sides of jaws. Color in life. — Above and on sides metallic green-blue; sides with 7 longitudinal gold stripes, a black spot above base of pectoral; snout mottled olive and brown, sides of head silvery-copper; underparts white; iris golden-brown; pectorals light dusky; ventrals white, dusky at tips; dorsal spines dark brown, membrane of spinous dorsal amber, with bluish area basally back of each spine; anal brownish, spines amber, membrane and rays pale ; caudal dusky. MEASUREMENTS, FIN RAYS AND SCALES OY ArchosargllS f0U7-- talesii. Length in mm 154 Depth 46 Head 1 26 Pectoral 40 Ventral 19 Eye 19 Number of dorsal spines XIII Number of second dorsal rays 10 Number of second anal rajs 10 Scales on lateral line I 46 159 47 29 40 20 16 XII II 10 44 147 50 28 44 21 17 XII II 10 46 144 50 29 43 20 16 XIII 10 10 49 121 52 28 42 21 XIII 10 12 44 133 46 28 41 20 29 XIII 10 10 45 Family GERRID^. 93. EUCINOSTOMUS DOWI (Gill). Diapteriis doivit Gill, Proc. Acad. Nat. Sci. Phila. 1863, 162, Panama. Gerres dmni, Steindachner, Ichthyologische Beitriige (iv), Sitzb. der k. Akad. Wissensch., vol. lxxii, part I, Dec, 1875, 13 (Callao, Peru ; Galapagos Islands). SHORE FISHES OF GALAPAGOS ISLANDS S^I Gerrcsdowl, Evermann & Mkek. Proc. Acad. Nat. Sci. ^'hil'-^- > f 86. 259. Eucmostomus dowi, Jordan & Evermann, Fishes North and Mid. Amcr.. 11. 1367, 1898. i?a;zo-^. _ Panama, Callao, Peru ; Galapagos Islands. One^'specimen taken at Chatham Island which differs somewhat from descriptions of E. dowi also from specimens of the same in the Stanford University collection from Panama, in being less deep, m having the ventral profile of the body almost straight, and in having a more angulated profile from the snout to the dorsal. These are charac- ters, however, subject to much variation. This one specimen is all that we saw, but Steindachner states that the species is present in great numbers about the shores of the Galapagos Islands. Xystcema cinerus is in external appearance extremely similar to Eucinostomus doivi, and is a common Galapagos fish. Hence it might be possible that Steindachner mistook this species for the other, for only on dissecting out the second interh^mal bone would one suspect the two to be dif- ferent if the specimens were mixed together and not examined care- fully. Since, however, the characters by which our specimen differs from the mainland specimens of Eucinostomus dowi are such that one specimen would not suffice for the determining of a species we simply give the following description of it. More material must be obtained to show whether the Galapagos form is E. dozui or an unde- scribed form : Length 160 mm.; depth 3; head 3*; eye 2f in head; pectoral very slightly longer than head; ventral 2i in head; depth of caudal peduncle 3! in head; snout from eye 3! in head ; first dorsal spine If in head; third anal spine 3! in head; upper lobe of caudal li m head; dorsal IX, 10; anal III, 7. . r 1 General shape somewhat elongate, mouth placed below longitudinal axis of body, oblique ; profile of head straight from tip of snout to top of supraoccipital crest, from here to front of dorsal straight forming an obtuse angle with the part in front; lower profile of body straight and horizontal from first anal spine to isthmus of gill-opemng; men- tal profile slightly concave; first, second and third dorsal spines high- est, equal, those back of the third graduated decreasingly to the last; first soft ray abruptlv higher, the following successively shorter to the last; first soft ray of anal a little longer than the third spine, about same length as first soft dorsal ray, the following rays graduated to last, which equals last dorsal ray; caudal deeply forked, the upper lobe somewhat the longer; nostrils unequal, the anterior the smaller; maxillary reaching the front of the orbit, exposed part elongate ovate, 382 SNODGRASS AND HELLER the forward end acute ; preopercle entire, its upper limb inclined very slightly back of perpendicular; space between orbits equal to verti- cal diameter of eye; premaxillary groove about 5 in interorbital space; eye elliptical, the longer axis longitudinal; teeth very small, a very narrow band along sides of jaws, a larger group in front of each jaw; snout, premaxillary grooves, preorbitals, jaws and chin naked, rest of head scaled ; scales all large on body, especially below lateral line ; ridges along middle of scales forming conspicuous longi- tudinal series on sides of body, 10 below lateral line; dorsal and anal each w^ith rather high membranous sheath at base, that of dorsal with scales indistinct except posteriorly, that of anal densely scaled through- out ; lateral line gently and regularly curved on the body, straight on the caudal peduncle; crossed by 47 rows of scales. Color. — Plain silvery, dorsal fin punctate with minute spots of black pigment. 94. XYST^MA CINEREUM (Walbaum). Trudiis cinereus peltaitts Catzsby, Nat. Hist. Carolinas, 1731, Bahamas. Mtigil cineretisV\[A\,^x\]U, Artedi Piscium, 228, 1792, Bahamas. Gerres cinereus, Jordan & Bollman, Proc. U. S. Nat. Mus. 1889, 181 (Chat- ham Island). Xystcema cinereum, Jordan & Evermann, Fishes North and Mid. Amer., 11, 1372, 1898. Range. — Both shores of tropical America ; Galapagos Islands. We have specimens of this species from Narboro ; Elizabeth Bay and Tagus Cove, Albemarle; Chatham; and some very small ones from a salt lake in an old crater near Tagus Cove, now entirely shut off from the ocean. Family KYPHOSIDiE. 95. DOYDIXODON FREMINVILLEI Valenciennes. Doydixodon fre7nin7rillei Valenciennes, Voyage de la Venus, v, 323, pi. 5, 1855, Galapagos Islands. ? Doydixodon fasciatum Kner & Steindachner, Neue Fische aus Mus. Godeffroy, Sitzb. d. k. Akad. d. Wissensch. Wien, liv, VX. i, 3, 1866, pi. II, fig. 2, Iquique, Peru. Doydixodon freminvillci, Jordan & Evermann, Irishes North and Mid. Amer., II, 1382, 1898. — Abbott, Proc. Acad. Nat. Sci. Phila. 1889, 351. Valenciennes first described and figured this species from a specimen taken at the Galapagos Islands by the Venus. Since then, 1S55, very little more has been known of the species. In 1866 Kner and Stein- dachner described a Doydixodon from Iquique, Peru, which they named fasciatum. Their description and figure are, however, from a very small specimen having broad vertical bands on the sides, and it SHORE FISHES OF GALAPAGOS ISLANDS 383 may be probable that it is a young individual of Doydixodon frcinhi- villei^ for the other differences between it and the adults of the latter species are shown by different-aged specimens in our collection. We have about 50 specimens from Tagus Cove and Iguana Cove, Albemarle ; from the east side and Mangrove Point of Narboro, James ; Duncan ; Chatham ; Charles ; and Seymour. About Tagus Cove the species was very abundant, occurring in large schools along the shore in shallow water and feeding at the surface. Description of a tyfica,l speci7neii. — General appearance thick, deep and heavy ; head especially large, being wider than the body, profile of snout straight and steep, forming a prominent obtuse angle with the profile from the eyes to the front of the dorsal. Lips thick; premaxillary and maxillary both thick, the latter deeply concealed be- neath the suborbital ; distal part of premaxillary rudimentary being cartilaginous and fibrous, the bony part forming only front of jaw ; a strong process from near outer end of posterior margin hooks upward around the lower margin of maxillary. Teeth only in front of jaws, similar, in several rows; each tooth consisting of 2 parts, one short, fiat, vertical, forming the cutting part of the tooth, the other elongate, slender and horizontal, forming a right angle with the first and directed backward from it to its insertion posteriorly ; each nos- tril somewhat tubular, the 2 equal, placed below the center of the pupil ; preopercle entire, its angle rounded, the upper limb inclined a little backward ; opercle with a short, wide, flat spine. .Spinous dorsal very low, spines short and thick, depressed in a groove, alter- nating, one more to right, next more to left, etc. ; twelfth longest, |^ greater than eye ; last closely united to the soft dorsal ; soft dorsal compared with spinous dorsal rather high, evenly rounded, the middle rays longest ; anal spines short, thick and evenly graduated ; soft anal higher than soft doi'sal, acutely angulated, the fourth and fifth rays longest; both fins much thickened at their bases; dorsal XII, 15; anal III, 15 ; caudal short, wide, posterior margin lunate, the upper lobe a little the longer, somewhat shorter than the head ; pectoral wide, reaching to near tip of ventral ; lateral line rather high, concurrent with the back; scales large, cycloid, 51 in lateral line; head naked except occipital region, supra-opercular region and median part of preopercle ; membranes of soft dorsal, soft anal, pectoral and caudal with small scales on their margins. The description quoted by Jordan and Evermann {Fishes of North and Middle A?nerica^ II, p. 1382) for Doydixodon is from Giinther, but belongs to another genus. The teeth are not " tricuspid." 384 SNODGRASS AND HELLER MEASUREMENTS, FIN RAYS AND SCALES OF DoydlXodotl frcminvillei. No. Stanford University Museum. Length in mm Depth Head Eye Pectoral Ventral Number of first dorsal spines. Number of second dorsal rajs. Number of second anal rays.. Scales in lateral line 12308 12309 12310 12311 310 290 333 305 47 4b 43 47 32 31 31 32 iS 18 18 18 74 76 75 70 57 62 56 56 12 12 12 12 18 15 16 15 12 12 12 12 54 55 51 52 34S 45 32 17 77 61 12 16 12 51 96. KYPHOSUS ANALOGUS (Gill). Pijnelepterus analogus Gill, Proc. Acad. Nat. Sci. Phila. 1862, 245, Cape San Lucas. Kyphosus analogus, Jordan & Evermann, Fishes North and Mid. Amer., 11, 1385, 1898. — Jordan & McGregor, Rep. U. S. Fish Comm. 1898, 278 (Clarion and Socorro islands). Range. — Pacific coast of Mexico, Panama, Clarion and Socorro islands. 97. KYPHOSUS ELEGANS (Peters). Pimeleptertts elegans Peters, Berliner Monatsberichte, 707, 1869, Mazatlan. Kyphosus elegans, Jordan & Evermann, Fishes North and Mid. Amer., 11, 1387, 1898. Range. — West coast of Me.xico ; Revillagigedo, Cocos and Gala pagos islands. One adult of this species taken at Cocos Island, and one young one taken at Alangrove Point, Narboro. The species is known from both Clarion and Socorro islands of the Revillagigedo Archipelago. 98. KYPHOSUS LUTESCENS (Jordan & Gilbert). Pimeleptertts hctesccns Jordan & Gilhert, Proc. U. S. Nat. Mus. 1881, 229, Socorro Island. Kyphosus hitcscens, Jordan & Evermann, Fishes North and Mid. Amer., 11, 1388, 1898. —Jordan & McGregor, Rep. U. S. Fish Comm. 1898, 278 (Clarion Island). Range. — Revillagigedo Archipelago. Family SCI^ENID^. 99. CORVULA EURYMESOPS Heller & Snodgrass. Cofvula ciDymcsflps Heller & Snodgrass, Proc. Wash. Acad. Sci., v, 1903, 195, Charles Island. Range. — Galapagos Archipelago. SHORE FISHES OF GALAPAGOS ISLANDS 385 100. SCI^NA PERISSA Heller & Snodgrass. Sciana perissa Heller & Snodgrass, Proc. Wash. Acad. Sci., v, 1903, 197, Tagus Cove, Albemarle Island. Range. — Galapagos Archipelago. 101. UMBRINA GALAPAGORUM Steindachner. Utnbrina galapagontm Steindachner, Ichth. Beitr., vii, 20, 1888, James Is- land, Galapagos. — Jordan & Evermann, Fishes North and Mid. Amer., II, 1468, 1898. Range. — Galapagos Archipelago. Known only from Steindachner's specimens. Family CIRRHITIDiE. T02. CIRRHITUS RIVULATUS Valenciennes. Cirrhites rivtclaius Valenciennes, Voyage de la Venus, Poiss., 309, pi. 3, fig. I, 1855, Galapagos Islands. — Jordan & McGregor, Rep. U. S. Fish Comm. for 1898, 283 (Clarion and Socorro islands). — Jordan & Evermann, Fishes North and Mid. Amer., 11, 1491, 1898. Range. — Pacific coast of tropical America ; Revillagigedo and Galapagos islands. This species was taken first at the Galapagos Islands by the Venus., and has since been found at Clarion and Socorro islands and along the Pacific Coast of Tropical America as far north as Cape San Lucas. We have one specimen, about 12 inches long, taken in a tide-pool by Captain W. P. Noyes at Tower Island. Depth of body 2f ; D. X, 11; A. Ill, 6. Family POMACENTRID^. 103. AZURINA EUPALAMA Heller & Snodgrass. Azurina eiipalama Heller & Snodgrass, Proc. Wash. Acad. Sci., v, 1903, 198, pi. V, Hood Island. Range. — Galapagos Archipelago (Charles and Hood). Genus POMACENTRUS. The following is the history of the Pacific American species of Po- mace7itrus. In 1S45, Tschudi i^Faujia Peruana, Ic/itky., 77) described the species Pomacentrus latifrons., from Huacho. We have not seen this species but according to Tschudi's description it must differ from all the other Pacific-American forms in having the dorsal rays XIII, 20, instead of XII, 15. 386 SNODGRASS AND HELLER In 1862, Gill {Proc. Ac. Nat. Sci. Phila. 1862, 148) described 2 species of Pojiiacentrus from Cape San Lucas — P. rectifrcemim and P. Jlavilat7is., the type of each being an immature specimen. Later in the same year, Gvinther {Cat.., IV, 27, 1862) published, from Gill's manuscript, the description of a third species, Pomoccntrus analiguttata Gill. This species Gill himself in the following year retracted. In 1863, Gill {Proc. Acad. Nat. Sci. Phila. 1863, 215) redescribed P. rectifrcenum and P. Jlavilatus. In these descriptions, based on a greater number of specimens, the difference described between the types of the 2 species is much less prominent. Giinther, in 1866 {Fishes of Central America, 445) regarded Gill's 2 types, Pomacentrus rectifrcenum and P. fiavilatus, as simply 2 color forms of the same species, P. rectifrcemun. In 1 89 1, Gilbert {Proc. U. S. Nat. Mus. 1891, 554) described the species P. leucorus, distinguished from the other species by having the pectorals tipped with orange (white in alcohol) and posteriorly bordered with white. The type specimens of this species are from the Revillagigedo Archipelago, but the species is abundant also at both Cocos and the Galapagos islands. Jordan, 1896 {Fishes of Sinaloa, Proc. Cal. Acad. Sci., 2d Series, Vol. V, 1895 (1896), 474, desci"ibed as Fzipomacentrus favilatus, a very young example from Mazatlan, about |- inch in length. In 1899 Jordan and McGregor {Fishes of Revillagigedo Archi- pelago and neighboring Islands, Rep. U. S. Fish Comtn.for i8g8, 282, 1899) recorded from Clarion and Socorro islands, 6 specimens of a Pomacentrus that they identified as P. rectrifr(E?zum, and 12 others that they identified as adults of P. flavilatus. We have examined this Revillagigedo material and find that, excluding P. leucorus with orange-tipped pectorals, it includes only one species — not 2 as identified by Jordan and McGregor. Further- more, we have examined both young and adult examples from Panama which are certainly P. rectifrccmi??t Gill. The Revillagigedo adults are not the same species as the Panama adults, and the young are neither P. rectifra^mim Gill nor P. favilatns Gill. Hence, we have described the Revillagigedo form as a distinct species — Poma- centrus redemptus {Proc. Wash. Acad. Sci., v, 1903, 200, pi. vi). At Cocos and the Galapagos islands there occur 2 species of Pomacefitrus — one is P. leucorus, the other we have described as P. arcifrons {ibid, 202, pi. vii). In 1904, Gilbert and Starks (Mem. Cal. Acad. Sci., iv, 141, pi. xxi) described the species P. gilli hom Panama. SHORE FISHES OF GALAPAGOS ISLANDS 387 Adults that are surely P. Jlavilatus have, thus far never been re- ported. If P. gilli Gilbert & Starks is not the adult of P. Jlavilatus Gill, it is probably likely that P. rcctifrccftzim Gill and P. Jlavilatus Gill are the same species, as Giinther concluded. There are, therefore, 6 species of Pomacentrus in the Eastern Pacific — P. latijrojis Tschudi along the coast of South America ; P. rectijrcc7mm Gill and P. gilli Gilbert & Starks along the coast of Mexico and Central America ; P. leucorus Gilbert at the Revillagigedo, Cocos and Galapagos islands; P. redemptus Heller and Snodgrass at the Revillagigedo Islands ; and P. arcijrotis Heller and Snodgrass at Cocos and the Galapagos Islands. 104. POMACENTRUS LEUCORUS Gilbert. Potnacenfrus leucorus Gilbert, Proc. U. S. Nat. Mus. 1891, 554, Socorro Island. Eupomacentrtis leucorus, Jordan & McGregor, Rep. U. S. Fish Comm. for 1898 (1899), 282 (Socorro, Clarion, San Benedicto islands). — Jordan & EvERMANN, Fishes North and Mid. Amer., 11, 1551, 1898. Raitge. — Revillagigedo, Cocos and Galapagos islands. This is the first record of the species from Cocos and the Galapagos Islands. Somewhat rare at the former locality, extremely abundant at the latter, w^here, together v^^ith P. arcijrons^ probably outnumbering any other species of the islands. The young are hard to obtain since they live amongst the rocks at the bottom of shallow water and scarcely ever float when killed with dynamite. Their bright colors, however, make them conspicuous objects and, when dead, they can be picked up with a long-handled spear. Diagnosis. — Easily distinguished from all the other species by the white and orange tip of the pectoral ; forehead evenly retreating from snout to front of dorsal, and preopercle narrow, both as in P. redejjiptus ; serrations of opercle minute; lips dark, same color as head; young without blue mark on scales, posterior half of caudal peduncle pale. Color in lije oj a typical adtilt. — Above reddish-brown, belly and sides pale dusky ; lips and snout livid grayish ; iris purple, a golden spot above pupil; opercle with bluish black spot at upper margin; dorsal and anal like sides, with black tips ; caudal black ; pectoral light brown at base, black beyond, posteriorly white bordered with tip orange, caudal peduncle paler than rest of body ; the white posterior border and orange tip of the pectoral forming a conspicuous mark of the species. The orange tip is not mentioned by Gilbert whose de- scription is from alcoholic material. Color in lije oJ a typical young example. — Above bright coppery- 388 SNODGRASS AND HELLER red ; sides bluish-black ; belly bluish-gray ; tip of snout dark blue ; chin like belly; cheeks coppery; opercle black-bordered; iris silvery- dusky with an inner red ring; spinous dorsal with spines cardinal, membrane dusky ; rays of soft dorsal blackish, purple spot at base, membrane black ; a large ocellus on side of soft dorsal with red margin and dark blue center; pectorals, ventrals and anal dark, the last with blue spots at the base posteriorly ; caudal lighter dusky, peduncle very light; pectoral with light olive spot at tip. The young of different ages show the following changes : I. (Specimen 54 mm. long.) General color like that of adult. Posterior part of caudal peduncle pale yellowish ; a large round black spot on base of anterior part of soft dorsal ; a large quadrate white spot (orange in life?) at tip of pectoral, ventrally dark color extending between white spot and edge of fin. II. (Specimen 38 mm. long.) Color same as last with exception of white on pectoral, which here forms a large oval spot, having longer diameter longitudinal, very slightly in front of tip of fin. III. (Specimen 31 mm. long.) White spot on pectoral very small, considerably in front of tip of fin, same shape as in last ; spot on base of anterior part of soft dorsal conspicuously ocelliform, having a wide pale border around the dark center (this character, however, probably, simply faded from others), the whole encroaching much on the side of the back. MEASUREMENTS, FIN RAYS AND SCALES OF PomacentruS Icu- corus. No. Stanford University Museum Length in mm Depth Head Pectoral Ventral Eje Interorbital Preorbital Number of dorsal spines Number of second dorsal rays Number of second anal rays Scale rows Scales on lateral line 12269 12270 I227I 118 107 105 104 51 48 52 53 30 30 31 32 27 25 27 28 29 28 29 32 27 28 30 28 30 30 30 31 17 14 13 14 XII XII XII XII 15 15 15 15 13 13 13 13 27 26 25 25 20 20 20 20 12272 103 52 31 27 32 27 30 14 XII 15 13 26 Description of adult examples. — Almost identical in every respect, except color, with Pomacentrus redeviptus^ the only differences between specimens of equal length being as follows : Upper profile of SHORE FISHES OF GALAPAGOS ISLANDS 389 head conspicuously not so steep, in P. leiicorus evenly retreating from snout to front of spinous dorsal or but gently curved, never steep and bulging in front or almost vertical before eyes ; serrations on preoper- cle and suborbital much finer, especially those on the preopercle ; pre- orbital less deep, being about |^ of eye while in P. 7-edempttis it is almost as deep as eye ; fins and profile of body same in the 2 species. 105. POMACENTRUS REDEMPTUS Heller & Snodgrass. Etipomacentrtis rectifranum, Jordan & McGregor, Rep. U. S. Fish Comm. for 1898 (1899), 282 (Clarion and Socorro islands). Eupomacenfrus flavilatus, Jordan & McGregor, ibid., 282 (Clarion, Socorro and San Benedicto islands). Poinaceiitriis redemptus Heller & Snodgrass, Proc. Wash. Acad. Sci., v, 1903, 200, pi. VI, Clarion Island. Range. — Revillagigedo Archipelago. Similar in shape to Pomacentrus je^ikinsi Jordan & Evermann' of the Hawaiian Islands. Differs from this species in having the subor- bital serrated and in having (in alcohol) a yellow color diffused over the caudal peduncle and the posterior part of the body. 106. POMACENTRUS ARCIFRONS Heller & Snodgrass. Pomacentrtts arcifrons Heller & Snodgrass, Proc. Wash. Acad. Sci., v, 1903, 202, pi. VII, Barrington Island. Range. — Cocos and Galapagos islands. A very abundant species; easily recognized by the strongly arcuate profile of the forehead and top of head, and by the bright orange color of the lips. 107. NEXILARIUS CONCOLOR (Gill). Euchistodus coticolor Q\\A., Proc. Acad. Nat. Phila. 1862, 145, lanama. Nexilarins concolor, Jordan & Evermann, Fishes North and Mid. Amer., 11, 1559, 1898. Range. — West coast of tropical America and Galapagos Archi- pelago. We have one specimen of this species taken at Elizabeth Bay, Al- bemarle Island, Galapagos. Length, 137 mm. Color hi life. — Above dark brownish-olive ; sides with 4 broad, vertical bars of dark brown, a yellowish tinge on areas between the bars ; a straw-yellow blotch behind pectoral ; dorsal and anal fins tipped with bright greenish-blue; ventrals bright green; other fins dusky. ' Rep. U. S. Fish Comm. for 1903 (1903), 189. Proc. Wash. Acad. Sci., January, 1905. 390 SNODGRASS AND HELLER io8. ABUDEFDUF MARGINATUS (Bloch). Chcetodon marghia/us Bloch, Ichthyol., vi, 73, pi. 207, 1788, Martinique. Abudefihif saxatilis, Jordan & Evermann, Fishes North and Mid. Amer., II, 1561, 1898. — Jordan & McGregor, Rep. U. S. Fish Comm. for 1898, 282 (Clarion Island). Glyphisodon saxatilis, Gilbert & Starks, Mem. Cal. Acad. Sci., iv, 1904, 143 (Panama). Range. — Both coasts of tropical America ; West Indies ; Revil- lagigedo, Cocos and Galapagos islands. Very similar to Hawaiian specimens recorded by Jenkins ^ and by Jordan and Evermann ^ as Abudefduf abdominalis (C. & V.) Differs from this species in having the band on the caudal peduncle more dis- tinct and in lacking the spots on the base of the soft dorsal and the anal opposite the ends of this band. We have specimens from Charles Island and from Tagus Cove, Turtle Point, Elizabeth Bay and Iguana Cove, Albemarle. A small specimen 16 mm. long has the soft dorsal, soft anal, caudal and pectoral fins very pale, strongly contrasting with the other parts. Specimens larger than this, up to 25 mm., have these fins pale but darker than in the smaller ones. The vertical bars are present on speci- mens as small as 16 mm. in length. Specimens 10 mm. long, how- ever, have no stripes at all. 109. MICROSPATHODON BAIRDII (Gill). Pomacenfrns bairdii Gill, Proc. Acad. Nat. Sci, Phila. 1862, 149, Cape San Lucas. Microspathodoii bairdii, Jordan"'& Evermann, Fishes North and Mid. Amer. , II, 1566, 1898. Range. — Pacific coast of Mexico and Central America; Revil- lagigedo and Galapagos islands. We have 4 large specimens of this species, the largest 270 mm. long, taken at Seymour, Charles and Hood islands. Color in life. — Above dark brown with purplish iridescence; throat and belly lighter brownish; fins like back; pectorals and ven- trals with distinct livid spots on membranes; other fins with the spots very faint ; iris purple. no. MICROSPATHODON DORSALIS (Gill). Hypsypops dorsalis Gill, Proc. Acad. Nat. Sci. Phila. 1862, 147, Cape San Lucas. ' Bull. U. S. Fish. Comm. for 1902 (1903), 458. 2 Hawaiian Report. SHORE FISHES OF GALAPAGOS ISLANDS 39I Microspaihodon dorsalis, Jordan & Evermann, Fishes North and Mid. Amer., 11, 1568, 1898. Range. — Pacific coast of Mexico and Central America; Revilla- gigedo, Cocos and Galapagos islands. Of this species we have 17 specimens, the largest 200 mm. long, from Seymour, Barrington, Duncan and Charles islands and from Cocos Island. This species, as the last, has before this been known only from the Pacific coast of Mexico and Panama and from the Re- villagigedo Islands. Color in life. — Above and on sides bluish-black; belly light gray- ish-blue; fins like back, edged with lavender. III. NEXILOSUS ALBEMARLEUS Heller & Snodgrass. Nexilosus albemarleus Heller & Snodgrass, Proc. Wash. Acad. Sci., v, 1903, 204, pi. VIII, Tagus Cove, Albemarle Island. Range. — Galapagos Archipelago. This species, constituting the genus Nexilosus., differs from the genus Hypsypops in the same way that the genus Nexilarius Gilbert differs from Abudefduf., viz., in having the suborbital fused with the cheek. We have several specimens from Tagus Cove, Elizabeth Bay and Iguana Cove, Albemarle. Family LABRID^. 112. BODIANUS DIPLOT^NIUS (Gill). • Harpe diplotania Gill, Proc. Acad. Nat. Sci. Phila. 1862, 140, fe7nale. Cape San Lucas. — Jordan & McGregor, Rep. U. S. Comm. for 1898, 181 (Clarion and Socorro islands). — Jordan & Evermann, Fishes North and Mid. Amer., 11, 1582, 1898. Range. — Cape San Lucas to Panama, Revillagigedo, CHpperton, Cocos and Galapagos islands. We have 26 specimens from Cocos, CHpperton and the Galapagos islands — at the last locality from Tagus Cove and Elizabeth Bay, Albemarle ; James ; Charles ; Duncan ; Seymour ; Barrington ; Hood and Bindloe. Color of adult male iti life. — Head lake-red, grayish on snout ; tip of under lip yellow ; lower jaw pinkish ; dorsum slate-greenish ; dor- sal spines green-blue, purple at tip ; soft dorsal at base greenish, pur- ple at tip, posterior rays cherry-red ; sides of body and belly light brownish-purple; outer rays of tail purple, inner with greenish base, tips maroon ; chrome spot on side above pectoral ; pectoral blue on 392 SNODGRASS AND HELLER upper rays, rest bright red, upper angle black-tipped ; outer rays of ventrals purple, inner green, maroon-tipped ; iris golden and red. Color of adult female in life. — Above olive, scales on sides with lake-red centers and olive borders ; 2 dark brown stripes on sides, 2 spots of same color on caudal peduncle, an olive maxillary stripe; belly lighter lake-red, a narrow faint purplish-olive border on scales ; snout mottled pinkish and olive; iris whitish with inner yellow ring; first and second dorsal spines bluish with red tips, others lake-red, membrane Indian-red ; last rays of soft dorsal orange, lighter at tips ; anal like dorsal ; caudal orange ; peduncle olive yellow ; pectoral light lake-red ; spine of ventral purplish-red, rays with crimson tips, fin lake and olive at base. The males of different ages vary much in shape and in contour of the fins. The smallest specimens, 270 mm. long, much resemble in shape the females. The snout is pointed, the swelling before the eye very slight and the angle of the caudal but little produced ; profile gradually rising from snout to first dorsal except for slight rise before eyes; soft dorsal and anal prolonged as far as base of caudal ; tips of pectorals each with a large prominent black spot just as in large speci- mens. As age increases the elevation before the eye becomes larger and the angles of the caudal and soft anal and dorsal increase. The largest specimens, 400 mm. long, have the snout very blunt, a large thick swelling before the eye having, in some specimens, the front surface very slightly receding. The soft anal and soft dorsal pro- longed posteriorly into long streamers reaching considerably past the median rays of the caudal, that of the anal generally longer than that of the dorsal. The caudal angles are also greatly produced and are tapering. All of our Galapagos specimens differ from a specimen from Clarion Island in having the snout blunter and the hump on the face smaller. Also the flap on the lower lip is smaller in most of our specimens, but this character is very variable and is not dependent on age — in some its width is but little more than half the diameter of the eye, while in others it is wider than the eye. 113. BODIANUS ECLANCHERI (Valenciennes). Cossyphus eclancheri Y A\.¥.'S.c\v.\ti\LS, Voyage de la Venus, Zool., 340, Poiss., pi. 8, fig. 2, plates 1846, text 1855, (ialapagos Islands. Harpe eclaiicheri, Jordan & Evekmann, Fishes North and Mid. Amer. , 11, 1583, 1898, ibid. Check-list, 412. Range. — Galapagos Archipelago. SHORE FISHES OF GALArAGOS ISLANDS 393 Found very common about Tagus Cove, Albemarle, but rare else- where about the islands. Obtained at Albemarle, Charles, Chatham and Barrington. Heretofore not reported since Valenciennes's descrip- tion of the type in 1S55. Depth i\ to 2|^. Hence Valenciennes's figure with a depth some- what more than 3 is not correct. Head 3 ; D. XII, 10. The males and females apparently do not differ much and both present the same variations. All of our specimens have the hump on the face larger than is shown in Valenciennes's figure and the snout is not so sharp, the dorsal profile of the head being steeper. Description of a typical specitnen. — Profile of snout and lower jaw forming a large angle with each other, very obtuse and symmetric- ally rounded; hump before and above eye, its anterior face almost vertical; profile from summit of hump to nape a little concave; soft dorsal and anal (in all cases) prolonged behind, but not reaching beyond base of caudal rays; caudal truncate, the upper angle slightly produced, the lower rounded. (Both angles of caudal not equally produced as in Valenciennes's figure.) Outer rays of ventral long, slightly shorter than pectoral, latter fin wide ; eye small, 5 to 7 in head. Color in life. — Scarcely any 2 specimens are colored alike. The color of Valenciennes's figure is only one phase of the coloration of the species and is not typical. Some specimens are of a uniform dusky color; others are entirely pale-colored or have a few irregularly scattered blotches of black. A specimen of this sort was colored as follows : color above reddish-orange, purplish on sides of head, lighter below ; pectoral with blackish spot at base, above this dusky-grayish ; anal spines black; ventral spine black; caudal grayish, black on middle rays. Between these 2 extremes are all degrees of mixture of black and orange. In one specimen the entire fish is black except the head and a longitudinal band on upper half of tail, these parts being pale orange. Others have the head and most of the tail orange, with this color also running ventrally from each toward middle of belly. In others the paler color encroaches on the sides so that only the back and upper parts of the sides are black. In still others only a small amount of black is left — this being on the back about the base of the dorsal fin. Here is where the black appears always to make either its last resistance against the encroachment of the orange, or its beginning in replacing the orange. Our specimens are all of the same size and we have no way of knowing whether the black replaces the orange or vice versa .^ or whether the coloration changes at all with growth or is permanent throughout life. The position of the 2 394 SNODGRASS AND HELLER colors, however, would seem to indicate that one color replaces the other during the life of the individual. All grades from black to orange found at the same time of the year and at the same place. 114. PIMELOMETOPON DARWINII (Jenyns). Cossyphus darwhiii JK^Y'ns, Voy. Beagle, Fishes, 100, pi. 20, 1842, Chatham Island, Galapagos. Za<5rz^5 «/^r Valenciennes, Voy. de la Venus, ZooL, Poiss., 338, pi. 8, fig. I, plates 1846, text 1855, Galapagos Islands. Pitneloinetopon darwinii, Jordan & Evermann, Fishes North and Mid. Amer. , II, 1586, 1898. Range. — Galapagos Archipelago. Of this species we have 8 specimens from Tagus Cove and Eliza- beth Bay, Albemarle Island. Description of an adult male : Length 420 mm. ; head 3 ; depth 3 ; eye 7 in head ; profile of head from eye to tip of snout very steep, rising at an angle of 45 degrees ; lower jaw large, thick, regularly rounded in ventral profile, so that snout and lower jaw are very blunt ; profile of head above eye to first dorsal spine gently rising, frontal hump present, but although large, not forming a prominent and abrupt swelling; body deepest through middle of pectoral; back of this regularly and symmetrically decreasing in depth to middle of caudal peduncle, back of middle of caudal peduncle enlarging again slightly ; dorsal XII, 10; third to sixth rays of soft dorsal elongated, reaching to posterior third of caudal peduncle; anal III, similar in shape to soft dorsal; caudal lunate, upper rays the more produced (not equal as in Valencienne's figure) ; pectoral wide ; outer rays of ventral pro- duced, about equal to pectoral, i-| in head; flap on lower lip small, 1^ of eye in depth. Color of male in life. — Above purplish-gray, below grayish-green ; a large golden blotch on side behind opercle above pectoral ; tip of lower jaw white ; fins light grayish olive. All of the males have the large yellow spot above the pectoral, and the character remains well on specimens kept in alcohol, so that it forms a very good distinguishing mark of the species. The females are much smaller than the males, the snout is pointed, the soft dorsal and anal fins and the ventrals are not prolonged, and the caudal lobes are generally equal (one has the U2:)per longer). Colo7- oj" females in life. — Following are the color descriptions of 2 females: (i) Ridge of back dark, sides reddish-purple, belly grayish ; upper surface of snout blue-gray, chin whitish ; vertical fins like back, ventrals like belly; iris purple. (2) Above light brownish SHORE FISHES OF GALAPAGOS ISLANDS 395 ridge of back grayish-lavender, pinkish tinge on sides, below whitish ; ventrals and anal light olive; caudal brownish; iris green with inner ring of red. 115. HALICHCERES NICHOLSI (Jordan & Gilbert). Platyglossus nichohi Jordan & Gilbert, Proc. U. S. Nat. Mus. 1881, 231, Socorro Island. — Jordan & Bollman, Proc. U. S. Nat. Mus. 1889, 182 (Charles Island, Galapagos). Iridio nicholsi, Jordan & Evermann, Fishes North and Mid. Amer., 11, 1591, 1898 ; ibid., Check-list, 412. Range. — Revillagigedo and Galapagos archipelagos. One specimen, 2S0 mm. long, from Charles Island. Depth a little greater than 3 ; head, to end of opercular flap, a little less than 3; D. IX, 13; A. Ill, 12; ventrals and pectorals about equal, \^ in hand; first anal spine very small, concealed by the skin; posterior border of caudal straight, angles not at all produced ; lateral line with tubes on 27 scales; bent downward over 4 scales on the 19th oblique series of scales, and then again running on 5 others, so that it crosses only 24 oblique series ; anterior canines i— i ; upper ones straight, directed forward, outward and downward; middle lower ones at symphysis of jaw, straight, projecting upward and forward be- tween upper ones ; outer lower ones projecting upward, outward and forward and then curving backward ; two small posterior canines in the upper jaw, each conical, directed downward. Color in life. — Above light olive, belly pale grayish, back darker than sides ; a blackish blotch behind head, brick-red below it ; sides of head with bright blue spots about eye, lavender spots below level of mouth ; lips pinkish ; lower jaw barred with straw-yellow ; pec- toral bluish-black, base light olive with blue blotches ; three chrome- yellow blotches on sides above pectoral ; spine and border of first ray of ventral blue, rest of first ray hazel-brown, posteriorly bordered with blue, the other rays pale, transparent ; dorsal dusky brick-red near margin, at base spotted with green, just above base blue, edge of fin light blue; membrane of caudal dusk)'-olive, spotted with blue, pos- terior border of fin red. 116. HALICHCERES SELLIFER (Gilbert). Halicliceres sellifer Gli^B^KV, Proc. U. S. Nat. Mus. 1890, 67, Clarion Island. Iridio sellifer, Jordan & Evermann, Fishes North and Mid. Amer., 11, 1592, 1898. Range. — Revillagigedo Archipelago. 3q6 snodgrass and heller Only the type is known, from Clarion Island of the Revillagigedo Archipelago. Collected by the Albatross in 1SS9. 117. PSEUDOJULIS ADUSTUS Gilbert. PsettdoJuHs adustus Gi'LBY.KT, Proc. U. S. Nat. Mus. 1890, 66, Socorro Island. Julidio adustus, Jordan & Evermann, Fishes North and Mid. Amer. 11, 1602, 1898. Range. — Revillagigedo Archipelago. Taken by the Albatross at Clarion Island, in 1889. No specimens recorded since. 118. PSEUDOJULIS NOTOSPILUS Giinther. Pseudojidis notospihis Gunther, Proc. Zool. Soc. Lond. 1864, 26, Panama. Julidio nofospilus, Jordan & McGregor, Rep. U. S. Fish Comm. for 1898 (1899), 283 (Clarion Island). — Jordan & Evermann, Fishes North and Mid. Amer., 11, 1603, 1898. Range. — Pacific coast of Mexico, Panama and Clarion Island of the Revillagigedo Archipelago. Taken at the last locality in 1897 by Mr. R. C. McGregor. 119. THALASSOMA SOCORROENSE Gilbert. Thalassoma socorroeiise Gilbert, Proc. U. S. Nat. Mus. 1890, 69, Socorro Island. Chloric hthys socoj'roensis, Jordan & Evermann, Fishes North and Mid. Amer., 11, 1607, 1898. Range. — Revillagigedo Archipelago. Known only from Socorro Island, the type taken by the Albatross in 1889, and no other specimens obtained since. 120. THALASSOMA GRAMMATICUM Gilbert. Thalassoma grammaticuin Gilbert, Proc. U. S. Nat. Mus. 1890, 68, Socorro and Clarion islands. — Jordan & McGregor, Rep. U. S. Fish Comm. 1898, 283 (Clarion Island). Chlorichthys grammaticus, Jordan & Evermann, Fishes North and Mid. Amer., 11, 1610, 1898. Range. — Revillagigedo Archipelago. Collected in 1S89 at Clarion and Socorro islands by the Albatross., and in 1897 at Clarion by Mr. R. C. McGregor. 131. THALASSOMA VIRENS Gilbert. Thalassoma virens Gilbert, Proc. U. Nat. Mus. 1890, 68, Socorro Island. Chlorichthys viretis, Jordan & Evermann, Fishes North and Mid. Amer., 11, 1 610, 1898. Range. — Revillagigedo Archipelago. SHORE FISHES OF GALAPAGOS ISLANDS 397 Taken at Socorro Island by the Albatross in 1SS9 and not reported since. Family SCARIDiE. 122. CALOTOMUS XENODON Gilbert. Calotomus .r^;w.f/ra(rw«/?^«£-/a/w« Bloch & Schneider, Syst. Ichthy., 501, 1801. — Jenyns, Voy. Beagle, Fishes, 158, 1842. — Guntiier, Cat., viii, 261. Range. — Tropical Pacific in general; Hawaiian Islands, Clipper- ton Island, Galapagos Islands. One specimen from Hood of the Galapagos Islands and one from Clipperton Island, the first record of the species from the Eastern Pacific. 145. OSTRACION CLIPPERTONENSE Snodgrass & Heller, new species. Ratige. — Clipperton Island. Type in Stanford University Museum ; Clipperton Island. SHORE FISHES OF GALAPAGOS ISLANDS 4II Diag7iosis. — Differs from Ostracion caimirian Jenkins' of the Hawaiian Islands only in having the spots of the back black instead of white. Description of the type. — Head 3I in length; depth 2\\ eye z\ in head; interorbital width i-^-; snout i^ in head. D. 9; A. 9; P. 10. Length 113 mm. Profile of front of head steep, forming a prominent angle before front of eye, back of which gently ascending to occiput; body deepest through base of pectoral, tapering regularly posteriorly with gently rounded outlines; angles of carapace rounded; no spines anywhere; sides with slight convexity; gill-slit not very oblique, 3^ in interorbital "^vidth, a groove extending forward from its lower end toward the snout ; eye a little greater than length of gill-slit, 2i in head ; mouth small ; lips large, concealing the teeth ; teeth brown, 10 in each jaw, in a single series. Dorsal fin short, highest anteriorly, longest rays \\ in head; anal similar to dorsal but not so high and more rounded ; pectoral rounded, third and fourth rays longest, i^ in head; caudal fin truncate, the angles rounded, 5 in length. Scutes of carapace covered with small tubercles, arranged most thickly in center of each; I3 scutes in a line between eye and caudal peduncle, 13 in a longitudinal series on belly, 10 in a transverse series. Color (in alcohol), back and sides blackish, belly pale greenish yellow; snout anteriorly, large spot below eye, and broad interorbital band, yellow ; dusky color of head before interorbital band darker than that back of it ; caudal peduncle brown, with a few rather large white spots ; fins pale yellowish with dark-brown bases ; basal half of caudal fin brown, the brown with concave crescentic posterior margin ; carapace spotted above with small round black spots, these spots situ- ated on center of each scute except on occiput where the position is generally eccentric ; a few whitish spots on sides. Proportional measurements : Head 28 ; depth 36 ; interorbital width 24; eye 11 ; snout 24; gill-slit 10; longest dorsal ray 15 ; longest anal ray 13; length of pectoral 21; median caudal rays 23; length of dorsal edge of caudal peduncle 14; depth of caudal peduncle 10. We have only i specimen from Clipperton Island. The Hawaiian specimens of O. camurum differ in shape from our specimen of O. clipperto?iense on\y in having the sides concave instead of slightly convex; the latter may be swollen. The pattern of the ground-color is exactly the same in all, the inter- 'BuU. U. S. Fish Comm. 1899 (1901), 397, fig. 9 (Honolulu). Type, No. 49697, U. S. Nat. Mus. Coll. Dr. O. P. Jenkins. 412 SNODGRASS AND HELLER orbital band and the contrast of the darker color before it with the paler color behind it are characteristic marks. The only difference is in the maculation. The Hawaiian specimens all have white spots instead of black ones. In some, these spots are confined to the back of the carapace as are the black ones in the Clipperton specimen, the sides being plain. Others are spotted above and on the sides. In 2 the spots on the side are ocelli, having (in alcohol) a white center and a wide black marginal ring. One has the spots of the upper parts of the sides black like those on the back of the Clipperton specimen. The latter may, of course, prove to be simply a color variation of O. ca?nur2im. Family TETRAODONTIDiE. 146. SPHEROIDES ANGUSTICEPS (Jenyns). Tt'irodofi angusticeps Jenyns, Voyage Beagle, Fishes, 154, pi. 28, 1842, Gala- pagos Islands. Spheroides angusticeps, Jordan & Bollman, Proc. U. S. Nat. Mus., xii, 1889 (1890), 183 (Chatham Island; Charles Island; Panama). — Jordan & EvERMANN, Fishes North and Mid. Amer. , 11, 1731, 1898. Range. — Pacific coast of tropical America, Galapagos Islands. We did not meet with this species, but it was obtained by Darwin and by the Albatross at the Galapagos Archipelago. 147. SPHEROIDES LOBATUS (Steindachner). Canthigastcr lobatus Steindachner, Ichthyol. Notizen, x, 18, pi. 5, fig. 3, 1870, Altata. SpJicroides lobatus, Jordan & Evermann, Fishes North and IMid. Amer., 11, 1731, 1898. Range. — Pacific coast of tropical America, from the Gulf of Cali- fornia to Panama ; Galapagos Islands. There is one small specimen of this species in the Stanford Univer- sity collection from the Galapagos Archipelago collected by the Albatross, but the only Spheroides that we obtained is a\ & Williams, Proc. U. S. Nat. Mus., xix, 1896, 457. (Referred by mistake to the Galapagos Archipelago.) The type of Paralicht/iys woolmani was taken in the Bay of Panama and was at first doubtfully referred to P. adspersjis^ but later described as P. woolmani and erroneously recorded as taken at the Galapagos Archipelago. Known only from the type.'\ Family SOLEID.^. 183. SYMPHURUS ATRAMENTATUS Jordan & Bollman. Syinphurus atramentatus Jordan & Bollman, Proc. U. S. Nat. Mus. 1889, 177, Coast of Colombia. — Jordan & Evermann, Fishes North and Mid. Amer., in, 2706, 1898. Range. — Gulf of Panama and Galapagos Archipelago. One specimen, 68 mm. in length, taken in 14 fathoms at Tagus Cove, Albemarle Island. This specimen differs slightly from the description of Panama specimens. Head 5 in length; depth 3^; D. 100; A. 75; scales 103. Coloration, light brown with 7 faint darker cross bars becoming darker posteriorly ; dorsal and anal fins dark- barred for their entire length, the bars confluent on the caudal fin. 424 SNODGRASS AND HELLER Family ANTENNARIIDiE. 1 84. ANTENNARIUS TAGUS Heller & Snodgrass. Antennarius tagus Heller «& Snodgrass, Proc. Wash. Acad. Sci., v, 1903, 226, pi. XX, Galapagos. Range. — Galapagos Archipelago (Tagus Cove, Albemarle). LISTS OF SPECIES KNOWN FROM EACH ISLAND OR GROUP OF ISLANDS. I. REVILLAGIGIDO ARCHIPELAGO. Carcharias platyrhynchus, Gyninothorax pictus^ Exocoetiis volitans^ Mugil cure77ia^ Mugil setosus^ Myripristis clarionensis. Gymnosarda pelatnis,, Coryphcena equisetis^ Zalocys stilbe., Caranx marginatus^ Caranx melampygus^ Ainia atricaudus^ Epi7tepheliis analogus^ Dermatolepis punctatus,, Paranthias fzircifer^ Priacanthtis crice?tiatus^ Anisotremus interruphis^ Kyphosus elega77s, Cirrhitus rivulaius^ Pomacentrus leucorus., Mlcrospathodon dorsalis^ Bodiamis diplotcenius, JFIalichccres sellifer^ Pseiidojulis 7iotospilus^ Thalasso77ia gra77i77iatic7i77i^ Caloto7nus xe7iodo7i^ Ch(ctodo7i 7iigrirost7'is^ Zaticlus ca7tesce7ts, Hcpatiis aliala^ XesiD'iis pjuictatiis^ Myrichthys pantostig777ius^ Echid7ia 7toct7^r7ta Exo7iautes xenopte7'US^ Ch<^tio77iugil proboscideus^ Holocentrus suborbitalis^ Pseud7ipe7ie7is dentatus^ Co7yphce7ia kippurus^ Trachurops cru77ie7iophthalina., Ca7'a7zx orthog7'a77i77ius^ Ca7'a7zx latus^ CaruTtx lugubris^ Kuhlia t<^7ii7ira^ Epi7iephelus labrifo7-77iis^ Prionodes fasciatiis^ P7'onotograi777777is 77mltij'asciat7is^ L7itia7tus vi7'idis^ Kyphos7is analogzis^ Kyphosiis littescois^ Abudefd7if 77ia7gi7iatus^ PomaccTttrus rede77iptus^ j\[ic7'ospatJiodo7t bai7'di^ Halichccres 7ticholsi^ Pseudojulls adustus.y Thalasso77ia socor7-oense^ Thalasso77ia virens, Forcipiger longlrostris^ Holoca7ith7is clarifl7ze7isis^ Hepatus t/'iostcgus^ Xesurus laticlavius., Balistes vc/'7-cs. SHORE FISHES OF GALAPAGOS ISLANDS 425 Osbeckia scrip ta,, Xanthichthys mento^ Tetraodon setosus^ Sebastopsis xyris^ Zonogobius zebra^ AItic2is chiosticttis^ Melichthys bispinosus^ Cantherines sandivichiensis^ Diodon hystrix^ Mapo soporator^ Gobiesox adzistus^ Platophrys leopardinus. II. CLIPPERTON ISLAND. Car char ias platyr/iync/nis, Bodianus diplotcenius, Ballstes verres^ Ostracioii cUppertonense, Mapo soporator. III. Triccfiodon obesus^ Myripristis murdjan^ Holocentrus siiborbitalis^ Kuhlia tceniura, Epinephelus lab7'iformis^ Mycteroperca olfax^ Priacanthus cruentatus^ Lutianus Jordani., Anisotremus scapularis^ Poitiacentrus leucorus^ Abiidefduf marginatus^ Bodianus diplotcenius^ Zanclus canescens^ Hep at us triostegus, Hepatus aliala^ Balistes verres^ Melichthys bispinosus^ Eleotris tubularis^ Mapo soporator, Epinephelus labriforfuis^ Hepatus aliala^ Ostracion lentiginosuvi^ Tetraodon setosus, COCOS ISLAND. Holotrachys lijna, Myripristis occidentalism Caranx tnelampygus^ Amia adradorsatus^ Dermatolepis punctatus^ Pronotogrammus multifasciatus^ Lutianus viridis^ Lutianus argentiventris, Kyphosus elegans^ Pomacentrus arcifroJis^ Microspathodon dorsalis^ Holocanthus passer ^ Ctenochcettis strigosus^ Hepatus crestonis^ Xesurus laticlavius^ Cattthidermis angulosus^ Tetraodon setosus^ Cotylopus cocoensis, Platophrys leopardinus. IV. GALAPAGOS ARCHIPELAGO. Branchiostoma elongatuni^ Carcharias galapagensis^ Rhinobatus planiceps^ Manta birostris^ Rabula marjnorea^ Gymnothorax dovii^ Galeocerdo rayneri^ Sphyrna tudes^ Dasyatis longa^ Ophichthys triserialis^ Gymnothorax chlevastes, AJurcetia insularum^ 426 SNODGRASS AND HELLER Murcena lentiginosa.i Zalarges lucetlus, Hemiratnphus saltator, Evolantia i7iicroptera^ Mugil cephalus^ ^ueritnana harengjis^ Myripristis occidentalism Holocentrus suborbitalis^ Gymnosarda pelamis, Sco?nberomorus sierra^ Trachurtis syfnmetricus,, Caranx latus^ Kuhlia tceniura^ Galeagra pammelas^ Epinephelus labriformis^ Mycteroperca xenarcha^ Mycteroperca ruberrima^ Paralabrax albo^naculatus^ Prionodes stilbostigma ^ Rypticus bicolor^ Lutianus viridis^ Xenocys jessice., Anisotremus surinamensis^ Orthopristis foi'b esi^ Orthopristis chalceus^ Calamus taurinus^ Eucinosto7nus doivi^ Doydixodon freminvillei^ Corvula euryniesops^ Uinbrina galapagorum^ Azurina eupalama^ Pomaccntriis arcifrons^ Abudefduf inarginatus^ Microspathodon dorsalis^ Bodianus diplotcetzius^ Pimelojnetopon darzvinii^ Callyodon noyesi^ OplcgnatJius insigne^ IlolocantJius passer ^ Zanclus canescens Balistes verres^ Clupanodon libertatis^ Ilyporhamphus roberti^ Euleptorhamphus longirostris^ Cypsilurus cyanopterus^ Alugil thoburni^ Sphyrcena idiastes., Myripristis murdjan^ Sco?nber colias, Thunnus thynnus^ Decapterus scojnbrinus^ Caranx caballus^ Caranx vielampygus, Amia atradbrsatzis^ Epinephelus analogus^ Dermatolepis punctatus^ Mycteroperca olfax^ Cratinus agassizi, Prionodes fasciatus^ Para?ithias fiircifer^ Priacanthus crzientatus, Lutianus argent iveritr is y Xenichthys agassizi^ Anisotrefnus scopzilaris, Orthopristis lethopristis^ Orthopristis cantharinuSy Archosargus pourtalesii^ Xystccffia citterezim^ Kyphosus elegans, Scicena perissa^ Cirrhitus rivulatzis^ Poinacentrus leiicorus^ Nexilarius concolor^ Microspathodo ft ba irdii,, Nexilosus albemarleus^ Bodianus eclancheri^ Halichccrcs nicholsi^ Callyodon perrico^ Chcctodon fiigrirostris^ Holocanthus iodocus, Xesurus laticlavius^ Ostra cio n Ic « t ip^in osu w , SHORE FISHES OF GALAPAGOS ISLANDS 427 Spheroides angusticeps^ Spheroides annulatus^ Diodon hystrix^ Sebastopsis xyris^ Pontinus strigatus^ Mapo soporator^ Caulolatilus princeps, Porichthys marga ritatus^ E7nJ7inion bristolcc^ Alticus atlanticus^ Lepisoma jenkinsi^ C hilar a taylori^ Petrotyx hop k ins i^ Prionottis tniles^ Gobiesox poecilophthalniiis ^ Platophrys constellaUis^ Sytnfhurus atramentatus^ Spheroides lobatus^ Tetraodoti setosjis^ Chilomycter7is ajjinis^ Scorpcena histrio^ Zonogobius rhizophora^ • Zonogobius gilberii, Myxodagnus operciilaris^ Dialoiumus fuscus^ Runula azalea^ Malacoctenus zonogaster^ Encheliophis jordani^ Otophiditun indefatigabile^ Eutyx diagrammus^ Echeneis reftiora^ Arbaciosa truncata^ Platophrys leopardinus^ Antennarius tagus. RROCKEDDINGS OF THE WASHINGTON ACADEMY OF SCIENCES Vol. VI, pp. 429-437. [Plates xxv-xxxiv.] February 28, 1905. SOME INTERESTING BEAVER DAMS IN COLORADO. By Edward R. Warren. Slate River, the stream on which were situated the beaver dams and other structures which form the subject of this paper, is in Gunnison county, Colorado. It is a clear mountain stream, flowing in a southeasterly direction past the town of Crested Butte, a place of about 1,000 inhabitants, and the location of a coal mine and coke ovens employing some 300 men. The altitude is 8,900 feet. It heads about 10 miles above Crested Butte, and is fed by tributary streams. That portion immediately north of the town flows with a very crooked winding course through an almost level bottom, thickly overgrown with clumps of willow bushes. For several years I had noticed with interest old and new dams and one or two houses at various places along the river and in some of its tributary sloughs, and in 1902 my attention was par- ticularly called to some new work farther down the river than any I had seen before. This work may have been begun in 1901, as, on account of other matters, I had paid little attention to the beaver that season, but the freshness of much of the con- struction showed that it had largely been done in 1902. After a preliminary examination it was evident that I had found something quite different and much more extensive than any beaver work I had seen before, and that it was so compli- cated that the quickest way to unravel it would be to take transit and stadia rod and survey it, which was done. Besides sur- Proc. Wash. Acad. Sci., February, 1905. ( 429 ) 430 WARREN veying the extensive new work nearest the town, I carried the survey up the river quite a distance, as shown on the map (Plate xxv). In making the survey I did not have time to go as far as I wished. There were some interesting things still farther up, both in the river and in some of the sloughs which should have been included. At low water the river varies in width from 20 to 40 feet. During high water, at the time of melting snows in spring and early summer, it is much higher and through this level ground must often overflow its banks. The area covered by the accom- panying maps (Plates xxv and xxvi) comprises many sloughs and channels, varying in width from i to 15 feet, and averaging perhaps from 2 to 4 feet. No attempt was made to trace out all the channels, only those most closely connected with the beavers' work being shown. Considering the area covered by the larger scale map (Plate xxv) first, it will be noted that with the exception of the 2 dams across the river, all the work done is to the south of the stream, and consists, besides houses, of many dams, most of them small, though I or 2 are of considerable length. Altogether they flood an area some 1,500 feet from east to west which contains about 17 acres. The reason there are no works north of the river is that the beavers completed their work there many years ago, raising practically all the land on that side above high water mark and converting it into a meadow, now much overgrown with willows. The lower dam across the river was 76 feet in a straight line from end to end, but it is curved twice (Plates xxv and xxvii) and does not reach all the way to the left bank. Below this dam, and setting out diagonally into the stream from the left bank, is a short dam, 20 feet long, so that the water flows around the end of the long upper dam and then around the free end of the short lower one, which thus throws the main body of water back to the right bank where it would naturally flow if undis- turbed by obstructions. The explanation of these curves and dams is that at this place the right bank is higher and the water deeper than on the opposite side, so that the downward curve at the southerly end of the long dam creates an eddy and back SOME INTERESTING BEAVER DAMS IN COI.ORADO 43 1 water there above the dam, and throws the swifter current to the other side in the shallow water. Then the lower dam, turning most of the water back to the south side, keeps the water deep under that bank. These 2 dams backed the water up for more than 200 feet, and the pond thus formed was apparently inhabited or used by quite a colony of beavers. On the south bank were several trails leading back from the river, and 75 feet above the dam was a lot of willow brush stored in the water, the large ends of the branches stuck into or against the bank and the small ends lying out into the pond (PI. xxix, fig. i). These branches were anywhere from 3 to 7 or 8 feet long, and as they extended along the bank for a distance of 100 feet or more in water 4 feet or more deep, and were laid from the bottom to above the surface, it is easily seen that the animals had done much work in getting their winter's supply on hand. Near the upper end of this brush was a house or lodge, or at least I supposed it to be such, though it was somewhat different from any others that I have seen, being in a clump of willows (the willows here are not trees but bunches or clumps of small stems growing from one or more roots, sometimes to a height of 12 feet or more). Among the stems of this clump of willows the mud and sticks seem to have been piled up to a height of 3 feet and the mud well plastered dow^n. It stood a few feet back from the stream, but between the latter and the house was a little channel and there was one on either side, and these channels all showed signs of constant use by beavers, as the water was so shallow that in going through, the animals' bodies left clear indications of their passage on the soft bottom. Returning now to the dams, or rather to the slough a short distance below them, which was 4 feet wide, it will be seen (PI. XXV) that there is one dam across the slough and then farther along a dam beside it. This latter was at a low place in the bank and prevented an overflow to the east, but a little farther south was an overflow into a small ditch across which the beaver had thrown a small dam hardly 2 feet long. This last looked to me like a waste of labor for I could not see any- thing gained by it. This marked the limits of the flooded land here. 43: WARREN It should be understood that nowhere in any of the flooded limits was all the land under water as there were many patches and bumps higher than their surroundings and thus showing above the water, so that the expression "flood limits" simply means that the water was backed up to those limits but not necessarily covering from sight all the included land. To the southwest of the end of the long dam is a small one across a slough which would naturally flow southeasterly enter- ing the river below the main dam, but which was thus turned and made to flow in the opposite direction and emptied into the river above the dam. West of the small dam is another, begin- ning across a slough flowing northeasterly and running out to the northwest so as to back the water up to the southwest. The animals here, as in many other places, took advantage of the fact that many of the willow clumps form little hummocks higher than the other ground and by building in between these hummocks construct quite a long, though not ver}^ high, dam without much trouble. To the south again, farther up this same slough, across it and alongside of it, is a series of dams constructed in much the same manner by connecting the hummocks. The dam or embankment alongside the slough served to keep it from over- flowing its bank when the dam across backed up the water. This brings us to the flood limits of this series. The second dam across the river (PI. xxviii, fig. i) is about 350 feet in an air-line from the first, but along the windings of the stream, 3 times that distance. Much of the land between the two, except in the big bend of the river, was under water. This dam is nearly 100 feet long, and extends entirely across the stream, but the water flows around the north end, and 20 feet beyond that end is a small dam across the overflow, which looks like labor thrown away, for it appears to serve no practical pur- pose. It does not seem as if it could affect the depth of the water in the pond above. The water was backed up here for nearly 500 feet. Some 200 feet above the dam, on the right bank, is a lodge, about 7 by 10 feet on the ground and 4 feet above the ground, con- structed, in tlie usual manner, of sticks and mud. No brush SOME INTERESTING BEAVER DAMS IN COLORADO 433 was stored in this pond, though many sticks peeled of their bark were lying in the water showing that beavers had been feeding there. Some 170 feet above the dam a slough 15 feet wide enters the pond from the southwest. Across this, 100 feet back from the river, is a ver}' substantial dam, 20 feet long and 8 feet wide, apparently constructed ver}'- largely of mud (PI. xxviii, fig. 2) ; it also appears to be an old construction, though some new work had lately been done on it. Then 130 feet above is another across the same slough. This is new work, and ex- tends 50 feet west of the slough to cut off the overflow there, and then turns and by a series running from hummock to hum- mock crosses another smaller slough which flows northwest. Also, as the map shows (pi. xxvi) its course is southeast and then southwest until it connects with a dam made many years ago and now all overgrown with grass. This old dam crosses the slough to the west of the. junction with the new, but that part the animals did not use. But they did use the part running off to the southeast, patching it wherever necessary. West of this dam is another old one not now used, discernible as a grassy line hardly raised above the surrounding level. The other 2 long dams to the west are constructed in much the same manner as the preceding, as well as the shorter ones, and much land was flooded by them. It seems rather difficult to understand fully the object of all these ponds and dams, for there is only i other house besides the 2 already mentioned, and the land is so flat and level that the beavers would have to tunnel a very long distance before the end of the burrow would be in dry ground, as even outside the flooded land the surface rises so little that the water level is only a short distance underneath. Probably the main idea was to give the inhabitants territory with deep water about which they might move under the ice and snow during the long win- ter. But this theory does not account for such a large flooded district, as I doubt if, outside the sloughs and ditches, there was very much water deep enough for a beaver to swim in under the ice. There were plenty of muskrats about the place who could use it, however. It may be that the beaver, having no way of 434 WARREN surveying the ground and finding out the levels and grades, had to work it out on the ground by actual construction. Some other things that I have known them to do have given me the impression that they do work that way sometimes — the short dam across the overflow around the end of the second dam, for instance. The third house is a very large one, 17 by 22 feet on the ground, of an oval shape, as the photograph shows (pi. xxix, fig. 2) and 12 feet along the ridge on top. It is the largest house I have ever seen. The construction of the 2 large dams across the stream, and also of the various small ones, is practically the same. All are built of willow stems and branches and mud, no large trees being used, for none was to be had. They are substantially constructed and the long ones are nearly 3 feet wide on top. Farther up the river are dams in which the beaver used spruce sticks, brought down the mountain sides by snow slides and left where they would float down the river when the snow melted. Some of the sticks so used are 4 or 5 inches through. But in the dams I am now describing there is nothing of that kind. The builders depended entirely on the willows, the largest of which would hardly exceed an inch in diameter. But they so mat and weave willows together — if the latter expression be allowable, though not to be taken absolutely literally — and mix and plaster them so thoroughly with mud that they are very solid. The construction of a similar dam is shown in the picture of an old dam which was cut through in order to drain the pond (PI. xxxii, fig. 2). Nearly opposite the large house is part of an old dam extending half way across the stream (PI. xxvi). This is quite a massive structure; how old, I do not know. Above this dam, as far as the limits of my survey, there are occasional bunches of brush in the deep holes, also many sticks with the bark peeled off, showing that the beavers were scattered all along. As the banks are somewhat higher there was opportunity for burrowing. At a number of places I also noted the animals' slides and landing places. Across the Baxter and Davis ditch, a small irrigating ditch 3 or 4 feet wide, beavers SOME INTERESTING BEAVER DAMS IN COLORADO 435 had built a dam and formed a small pond. This ditch was dug in 1902 which thus fixes the date of that dam. The 2 ponds shown at the northwest corner of the map (PL xxvi) are rather old. I saw them in 1900, but when the survey was made there was fresh beaver sign all about. The 3 trails between the lower pond and river were within a space of 80 feet and showed signs of regular use. About three-quarters of a mile above these last workings are 2 series of 3 dams each, and all along the stream are many signs of the animals in the way of brush cuttings and landing places. The lower of these 2 series had i dam extending part way across the stream, and 50 feet above it are the other 2, nearly opposite each other, i on each side of the stream, but not coming out far enough to meet. There is a bend in the river here and it is quite wide. One of these 2 dams had many sticks in the lower side pointing down stream, while the other had many stones banked against its upstream side. The other series of 3 is a few hundred feet above (PL xxxii, fig. i). The upper dam of the set is on the easterly side of the river and extends about half way across, 25 feet below and on the opposite side is the middle dam, also extending half way across the stream, which is here 40 or 50 feet wide ; 65 feet below that, and on the same side of the river, is the third dam, like the other 2 reaching half way across. Twenty-five feet below the third dam and on that same side is a small one running out into the stream about 8 feet. The upper dam backed up quite a lot of water but it is not any great depth as the river has too much fall. The other 2 dams did not, I think, deepen the water very much, but created still water more to the beavers' tastes. All the dams in this part of the river contain more or less spruce sticks which had been brought to the river from the steep hill- sides on the west by snow slides during the winters. Another interesting set of workings, though long since aban- doned, is in a slough on the east side of the river valley a short distance outside the map limits. Here the dam has been cut through by a ranchman to drain the pond so that everything is exposed. The dam was 105 feet long, the outlet, which was the slough that furnished the water, was 42 feet from the east 436 WARREN end. The pond had been 600 feet long and was at least 100 feet wide at the upper end. Now there is only a small crooked stream meandering through the site, with a little water still re- tained by the dam, at the lower end. The dam at the outlet, where it had been cut, was 10 feet wide at the bottom and 4 feet deep. There is a lodge 25 feet beyond the west end of the dam and a little north of it (the dam was across the southeast end of the pond). I cut the house open so as to expose a cross section from north to south (PL xxxii, fig. 2). Before cutting, the house measured 10 feet from east to west and it was 8 feet wide across the section. The cavity or chamber was 2 feet wide and extended back 4^ feet. It measured 10 inches high but I think the roof had undoubtedly settled a little. The thickness of the roof over the chamber was 22 inches. The thickness of the walls on either side of the chamber was 3 feet. The floor was just above the water level and there was a bed of slough grass inside. The entrance came into the chamber from under a mass of brush which lay to the east of the house. This brush was willow stems and twigs which had been cut by the beavers and extended out from the house for a space of 15 feet. Then came a space of 7 feet without any brush, and then another brush heap 20 feet long. These 2 lots of brush varied in width from 6 to 10 feet. The one farthest from the house extended into the water (of which there was a little here) beyond the point to which I measured. The willows were an inch or less in diameter. They had been there so long that the bark was falling off, but I could not see any signs that the beavers had ever used them. Near the upper end of this pond was a muskrat house, built up against a willow bush, and was 3 feet in diameter and 2^^ feet high. The entrance was at the bottom and was 7 in. wide by 8 in. high. The structure was built of bunches of moss and slough grass. A runway still showed coming from the entrance out into the pond (PL xxxiii, fig. i). The beavers are found more or less all alongf the stream from the point where this description begins to the little mining camp of Pittsburg, 7 miles above. Above that place the stream flows through a narrow rocky gulch entirely unadapted to the ani- mal's habits. SOME INTERESTING BEAVER DAMS IN COLORADO 437 Below Crested Butte I have not seen any recent beaver sign, but there are old beaver meadows along the stream, and the old grass-grown dams may be seen occasionally. The Colorado law against the taking of beaver is quite strict and though perhaps not as well enforced as it might be, is still sufficient to deter people from molesting them so close to a town. A fence post in the fence shown on the map (PI. xxv) as cross- ing the river below the lower long dam was cut down by beavers, and the owners of the land in setting a new one were obliged to protect it with sheet iron. The post was a green aspen, and set close to the river bank. The tree shown in the photograph (PI. xxxiv) was cut by beaver on another stream some 15 miles northwesterly from Crested Butte. There were 3 trees at this place cut by the animals within a short distance of each other, all close to the creek bank, and of the 3 one fell toward the water, one away from it and the third lay parallel with the bank. In examining these trees to see if there was any special reason for their being felled in this way, it seemed to me as if they had been attacked from the most convenient side and allowed to fall as they would. I know that many claim that beavers fall their trees in any direc- tion they wish; and perhaps they do, but I must confess I am inclined to doubt it. At least 2 of the above-mentioned trees would have been much handier to get at to cut up if they had been dropped in another direction and there was nothing to prevent that so far as obstructions were concerned, and these are not the only instances of the kind I have seen. Proc. Wash. Acad. Sci., February, 1905. A Proc. Wash. ACAD. Sci. vou.VI. ROC. Wash. Acad Sci. Vol. VI. Plate XXVII. !(.. I. Slate River Vallev. The liea\er da^l^ are hevoiui the fence and are oliscureii h\ tlie willow: riiotos E. K. w. IG. 2. Beaver dams across Shite River, from north end. Tiie end of the short dam helow can be seen. Proc. Wash. Acad. Sci. Vol. VI PLATE XXVIII. Fig. I. The second dam crosses Slate River from below. The house shows indistinctly near the center of the picture. Fig. 2. \'er\ wide mud and stick dam ucioss .slouj^h. />/!.. (.)s /■:. K. w. pRoc. Wash. Acad. Sci. Vol. VI. Plate XXIX. Fig. I. Biiisli cut aiul >t()rcd in iioiul h\- licavcr--. Slate River, Colorado. ^i ' !^l^^l ^^^^^^^^^^^^^^^^^BC ''x^l^^^^^^^^^^l HH > 1/ 1 /7/..r.,, A'. /;. ir. Proc. Wash. Acad. Sci. Vol. VI. Plate XXX Fig. I. Bea\er poiul before draining, on Shite River, Coloracio. .'-.--** • , ._.iS^. Fig. I. Old dam and brii>h piles in the old beaver pond whieh has been drainetl. -^<'..M K^^ymF jr<-. ./,'> ^ Fig. I. Three beaver dams on dilate l -^ " 1905 450 Fifth '>' " t. .k Treasurer, 1903 452 Sixth '' •' '' " '' 1904 455 Seventh " -^ " '' "■ 1905 456 Officers and Committees 1 903 459 Officers and Committees 1 904 46 1 Officers and Committees 1905 463 List of Members 1 904-5 465 Proc. Wash. Acad. Sci., February, iQO.s. (439) 44© WASHINGTON ACADEMY OF SCIENCES INTRODUCTION. In this brochure has been collected material relating to the organization and present membership of the Academy. It in- cludes the Act of Incorporation, the By-Laws, and Rules re- specting Publication as now in force, with lists of officers and committees for 1903, 1904 and 1905. All reports of the Secre- tary and Treasurer, not previously published, are also included. These reports show the activities, growth and work of the Academy. The list of members gives simply the name and address of each member, and has been corrected to January 21, 1905. ACT OF INCORPORATION aai ACT OF INCORPORATION We, the undersigned, persons of full age and citizens of the United States, and a majority being citizens of the District of Columbia, pur- suant to and in conformity with sections 545 to 553, inclusive, of the Revised Statutes of the United States relating to the District of Colum- bia, as amended by an Act of Congress entitled "An Act to amend the Revised Statutes of the United States relating to the District of Columbia and for other purposes," approved April 23, 18S4, hereby associate ourselves together as a society or body corporate and certify in writing: I. That the name of the society is the Washington Academy of Sciences. ^ 2. That the term for which it is organized is nine hundred and ninety-nine years. 3. That its particular business and objects are the promotion of science, with power : a. To acquire, hold, and convey real estate and other property and to establish general and special funds. d. To hold meetings. c. To publish and distribute documents. d. To conduct lectures. e. To conduct, endow, or assist investigation in any depart- ment of science, y. To acquire and maintain a library. ^. And, in general, to transact any business pertinent to an academy of sciences. 4. That the affairs, funds, and proj^erty of the corporation shall be in general charge of a Board of Managers, the number of whose mem- bers for the first year shall be nineteen, all of whom shall be chosen from among the members of the Academy. Witness our hands and seals this iSth day of February, 1S9S: (Signed) J. R. Eastman J. W. Powell F. W. Clarke Geo. M. Sternberg G. K. Gilbert H. N. Stokes Arnold Hague Charles D. Walcott L. O. Howard Lester F. Ward W J McGee Frank Baker C. Hart Merriam Bernard R. Green 442 WASHINGTON ACADEMY OF SCIENCES District of Columbia, to vjlt : I, John D. McChesney, a Notary Public in and for the District aforesaid, do hereby certify that J. R. Eastman, F. W. Clarke, G. K. Gilbert, Arnold Hague, L. O. Howard, C. Hart Merriam, J. W. Powell, Geo. M. Sternberg, H. N. Stokes, Chas. D. Walcott, Lester F. Ward, W J McGee, Frank Baker and Bernard R. Green, parties to a certain Certificate of Incorporation, bearing date on the iSth day of February, 1898, and hereto annexed, personally appeared before me in the District aforesaid, the said J. R. Eastman, F. W. Clarke, G. K. Gilbert, Arnold Hague, L. O. Howard, C. Hart Merriam, J. W. Powell, Geo. M. Sternberg, H. N. Stokes, Chas. D. Walcott, Lester F. Ward, W J McGee, Frank Baker and Bernard Green, being per- sonally known to me as the persons who severally made and signed the said certificate and acknowledged the same to be their certificate, act and deed. Given under my hand and notarial seal, this i8th day of February, 1898. Jno. D. McChesney, [seal.] Notary Public. Endorsed as follows : 9:20 A. M. Received for record February 21, 1898. Recorded in liber 8, fol. 207 et seq., Acts of Incorporation for Dist. of Col. H. P. Cheatham, Recorder. BY-LAWS 443 BY-LAWS. (Ill force February 17, 1905.) Article I. — Mcjnbcrs. Sec. I. The Washington Academy of Sciences shall comprise 4 classes of members, as follows : Resident mcfnbers^ non-resident jne??ibers, honorary ineinbers and patrons. Sec. 2. Resident^ non-residetit and honorary jnembers shall be persons who by reason of original research or scientific attainment are deemed eligible to these classes. Resident members shall be chosen from the Affiliated Scientific Societies of Washington. Non-resident members shall be chosen from localities outside the District of Colum- bia, and honorary members may be residents of any country. Patrons shall be persons who have given to the Academy not less than $1,000, or its equivalent in property. On removing to the District of Colum- bia and acquiring membership in one of the Affiliated Societies non- resident members shall thereby become resident members. Resident members, non-resident members, and patrons who have Keen mem- bers, and they only, shall be entitled to vote. The annual dues of resident and non-resident members shall be five dollars; honorary members and patrons shall pay no dues. Members whose dues are in arrears for more than one year shall be dropped from the roll of the Academy, unless the Board of Managers shall otherwise determine. Sec. 3. Nominations for membership shall be endorsed by at least 3 Members of the Academy, who shall present in writing a statement of the qualifications of the nominee, with a list of his more important publications; and such nominations shall be referred to the Board of Managers for consideration. Article II. — Officers. Sec. I. The officers of the Academy shall be chosen from the resi- dent members, and shall be a President, one Vice-President from each of the affiliated societies, a Secretary, and a Treasurer, whose terms of office shall be i year, and 9 Managers, grouped in 3 classes of 3 each, whose terms of office shall be 3 years. Collectively they shall constitute the Board of Managers. Sec. 2. The President and the Treasurer are authorized to assign securities belonging to the Academy and indorse financial and legal papers necessary for the uses of the Academy. Sec. 3. The Board of Managers shall transact all business of the Academy not otherwise provided for, and shall have power to fill vacancies in its own membership until the next annual election. Va- cancies in the office of Vice-President shall be filled on nomination by the appropriate affiliated societies. 444 WASHINGTON ACADEMY OF SCIENCES Article III. — Meetings. Sec. I. The Annual Meeting shall be held on the third Thursday of January each year. At this meeting the reports of the Secretary, Treasurer, and Auditing Committee shall be presented and officers for the ensuing year shall be elected. Sec. 2. Other meetings shall be held at such time and place as the Board of Managers may determine. Sec. 3. Twenty resident members of the Academy shall constitute a quorum for the transaction of business. Article IV. — Committees. Sec. I. The Board of Managers may appoint such standing and special committees as it deems necessary. Sec. 2. The President shall appoint in advance of the annual meet- ing an Auditing Committee consisting of 3 persons, none of whom are officers, to audit the accounts of the Treasurer. Article V. — Elections. Sec. I. At each annual meeting there shall be elected by ballot a President, a Secretary, a Treasurer, and 3 Managers, who shall serve until the close of the meeting at which their successors are chosen. A majority of the votes cast shall be necessary to elect. Members whose dues are in arrears for one year shall not be entitled to vote or be eligible for any office in the Academy. Sec. 2. Resident members shall be elected by the members of the Academy, and three-fourths of the votes cast shall be necessary to elect. An election shall be void if the person elected does not within 3 months thereafter pay his annual dues or satisfactorily explain to the Board of Managers his failure to do so. Sec. 3. Non-resident members, honorary members, and patrons shall be elected by the Board of Managers, and three-fourths of the votes cast shall be necessary to elect. The Board shall have power to determine and change the status of resident members to non-resident. Article VI. — Cooper atio7i. Sec. I . The Academy may act as a federal head of the Affiliated Scientific Societies of Washington, with power to conduct joint meet- ings, publish a joint directory and joint notices of meetings, and take action in any matter of common interest to the affiliated societies: Provided It shall not have power to incur for or in the name of one or more of these societies any expense or lir.bility not previously authorized by said society or societies. Sec. 2. The term '■ affiliated societies ' shall be held to cover the Anthropological, Biological, Chemical, Entomological, National Geographic, Geological, Medical, and Philosophical Societies and BY-LAWS ^.45 such Others as may be hereafter added by a majority vote of the mem- bers of the Academy, the vote being taken by correspondence.' Sec. 3. One Vice-President may be nominated by each affiliated society from the members of the Academy, subject to election by a majority vote at a meeting of the Academy. Sec. 4. Any affiliated society may nominate candidates for mem- bership in the Academy. Article VII. — Amendments. These By-Laws may be amended in the following manner: Written notice of proposed change, signed by at least three resident members, may be presented at any meeting of the Academy. Such notice shall be referred to the Board of Managers for consideration and recommendation. The Board of Managers shall consider the proposed change and return it to the Academy for action, with such amendment or recommendation as it deems wise. A two-thirds vote of the members voting shall be necessary to adoption. 1 The following have been added to the list of ' affiliated societies ' under the authority of this section : The Botanical Society of Washington. The Columbia Historical Societj'. The Washington Society of the Archaeological Institute of America. The Society of American Foresters. 446 WASHINGTON ACADEMY OF SCIENCES AFFILIATED SOCIETIES. Anthropological Society of Washington. Biological Society of Washington. Botanical Society of Washington. Chemical Society of Washington. Columbia Historical Society. Entomological Society of Washington. Geological Society of Washington. Medical Society of the District of Columbia. National Geographic Society. Philosophical Society of Washington. Society of American Foresters. Washington Society of the Arch^ological Institute of America. RULES RESPECTING PUBLICATIONS 447 RULES RESPECTING PUBLICATION. 1. The Proceedings of the Washington Academy of Sci- ences shall be issued in dated brochures, paged consecutively for the volume. 2. A brochure may comprise one or more papers, according to length, at the option of the Committee on Publication. 3. The date on each brochure shall be that of actual publication, which shall be one day later than the date of delivery by the printer to the Committee. 4. Each brochure shall be distributed on the date of its publication. Copies shall be sent to all members of the Academy, to subscribers and to a library and exchange list approved by the Board of Managers. 5. At the close of each volume, which shall coincide as nearly as possible with the calendar year, a brochure comprising the title page, contents and index of the volume shall be issued. 6. The regular edition shall consist of twelve hundred copies. 7. Contributors to the Proceedings must be members of the Academy, or of one of the Affiliated Societies ; provided, however, that in exceptional cases the Board, by a three-fourths vote of the members voting at a stated meeting, may accept for publication papers contributed by non-members of the Affiliated organizations. 8. Papers offered for publication shall be delivered to the Chair- man of the Committee on Publication, who shall be Editor of the Proceedings. The Editor shall submit to the Committee an estimate of cost, and shall see that papers are promptly examined in such man- ner as the Board of Managers may direct. 9. Manuscript submitted for publication must be in form, as well as in substance, ready for the printer. It must be complete as to text and illustrations, must be perfectly legible (preferably typewritten), and must be preceded by a brief table of contents. 10. The Academy shall not be responsible for the cost of revising manuscripts or illustrations. The cost of proof corrections due to alterations made by the author shall be charged to him. 11. Papers accepted for publication in the Academy's Proceed- ings must not be previously published elsewhere except by consent of the Board of Managers of the Academy. 448 WASHINGTON ACADEMY OF SCIENCES 12. Authors' separates shall not differ in any particular from the regular edition. Any desired number may be ordered in advance through the Committee on Publication, at the expense of the author, and at a rate of cost agreed upon by the Committee and the printer. 13. Authors shall receive, free of cost, thirty copies of their papers. FIFTH ANNUAL REPORT OF THE SECRETARY, 1903. To THE Washington Academy of Sciences : Gentlemen: In compliance with the By-Laws the Secretary has the honor to submit this, the fifth annual report, which covers the period from January 17, 1902, to Januar}' 15, 1903. During the year the Academy held i business meeting, presided at the annual addresses of 2 of the affiliated societies, and arranged for i lecture in honor of the National Academy of Sciences which was de- livered by Professor Charles F. Chandler of Columbia University, New York City. The Board of Managers held 11 meetings for the transaction of business. Application was made during the year for admission to the group of affiliated societies by the Washington Society of the Archreological Institute of America and the Botanical Society of ^Vashington. The claims of these bodies for recognition of this nature were carefully considered by the Board and it was finally determined to submit the matter to the Academy by a con"espondence vote as required by the By-Laws. This was done with the following result : For the admission of the Archaeological Society 214 Against the same 3 For the admission of the Botanical Society 219 Against the same i These societies were accordingly admitted. The statistics of membership are as follows : Patrotis. At date of last report 8 Elected during the year o 8 SIXTH ANNUAL REPORT OF THE SECRETARY 449 ResideJtt JMciubers. At date of last report 14S Elected during the year 14 162 Deceased , 1:5 Resigned 3 Dropped for non-payment of dues 6 Transferred to non-resident class i 15 147 No7i-reside7it Members. At date of last report 152 Elected during the year 25 Transferred from resident class i 178 Deceased 4 Resigned 6 10 168 323 Counted twice I Total membership 322 Respectfully submitted, Frank Baker, Secretary. January 15, 1903. SIXTH ANNUAL REPORT OF THE SECRETARY, 1904. To THE Washington Academy of Sciences : Gentlemen : In compliance with the By-Laws the Secretary has the honor to submit this, his si'xt/i annual report, covering the period from January 15, 1903, to January 21, 1904. During this time 5 meetings of the academy have been held desig- nated as follows : Annual meeting January 15, 1903 Meeting to hear the address of the retiring Presidentof the Anthropological Society February 3, 1903 Meeting commemorative of "The Organiza- tion and Endowment of Research" April 16, 1903 Business meeting April 24, 1903 Vol.V of the Proceedings has been nearly completed and the finished brochures distributed to members. The Directory of the Academy 450 WASHINGTON ACADEMY OF SCIENCES and Affiliated Societies for 1903 was also printed and distributed. The Committee on Publication suffered severely from the loss of its able and efficient Chairman, Mr. Marcus Baker. The Board of Managers held 9 meetings for the transaction of the current business of the Academy. The Academy has had some significant losses by death during the present year, as follows : Resident Members, A. B. Richardson, R. U. Goode, and Marcus Baker; non-resident members, Henry B. Hill and R. H. Thurston. The statistics of membership are as follows : Patrons. At date of last report 8 Elected during the year . . . o 8 Resident members. At date of last report 1 47 Elected during the year 20 167 Deceased 4 Transferred to non-resident class 3 6 161 JVon-resident members. At date of last report 16S Elected during the year 6 Transferred from resident class 3 176 Deceased 2 Resigned 9 11 165 334 Counted twice i Total membership 333 Respectfully submitted, Frank Baker, Secretary. January 21, 1904. SEVENTH ANNUAL REPORT OF THE SECRETARY, 1905. To THE Washington Academy of Sciences : Gentlemen: In compliance with the By-Laws the Secretary has the honor to submit this, his sevent/i annual report of the operations of the Academy, covering the period from January 21, 1904, to January 19, 1 905 . SEVENTH ANNUAL REPORT OF THE SECRETARY 45 1 During this period the following meetings of the Academy have been held : 23d. Meeting January 3j, 1904 — Annual for election of officers, etc. 24th. Meeting March 39, 1904 — For hearing remarks by Mr. T. C. Chamberlin, of the University of Chicago, on the Planetesimal Hypothesis, together with discussion. 25th. Meeting January 5, 1905 — For a discussion of the Use of Copper Salts in the Purification of Water Supplies. The Board of Managers has held 10 meetings for the transaction of the current business of the academy. It has, during this time, en- deavored to establish a closer relationship between the resident and non-resident members of the academy by inducing the latter to pre- sent at public meetings important results relating to their special lines of work. The Proceedings of the Academy have been regularly published in the form of separate brochures. Vol. VI is now nearly completed. Arrangements have been made to prepare and publish a revised edi- tion of the Directory of the Academy and Affiliated Societies. The labor of the preparation and editing of the Proceedings and other publications of the Academy is so considerable that it has been deemed advisable to carry out the suggestion made by the Academy at the last annual meeting and employ a paid editor for this work. Mr. Barton W. Evermann, the Chairman of the Committee on Publi- cation, has been designated as Editor and has performed the services to the eminent satisfaction of the Board. A scheme for the wider distribution of the publications of the Academy has been inaugurated, and loi universities, libraries and other institutions have been added to the exchange list. Early in the year the Society of American Foresters applied for admission into the group of Affiliated Societies. In accordance with Article VI, Sec. 2, of the By-Laws, a vote on the subject of the admission was taken by correspondence with the following result : Voting members of Academy 31S Necessary to a choice 160 In favor of admission 23S Against admission 3 Declined to vote i Others not voting 76 The Society of American Foresters was accordingly admitted to the group of affiliated societies. 452 WASHINGTON ACADEMY OF SCIENCES The Board having received a communication from the Secretary General of the International Botanical Congress to be held in Vienna, June 12 to 1 8, 1905, requesting the Academy to appoint a delegate to that Congress, appointed accordingly Mr. Frederick V. CoviJle to represent the Academy at that meeting. The Academy has suffered the following losses by death during the year : W. B. Powell died February 6. E. A. de Schweinitz died February 15. H. L. Marindin died March 25. A. Lindenkohl died June 22. The statistics of membership at this date are as follows: Patro7is. At date of last report 8 Elected during the year o 8 Resident members. At date of last report 161 Elected during the year 2 Restored to membership ..T i 164 Deceased 3 Resigned 3 6 15S No7i-7-esident me?nbe7's. At date of last report 165 Elected during the year 2 167 Resigned S 8 159 325 Counted twice i Total inembership January 19, 1905 324 Respectfully submitted, Frank Baker, Secreia7y. January 19, 1905. FIFTH ANNUAL REPORT OF THE TREASURER, 1903. To THE Washington Academy of Sciences : The Treasurer has the honor to submit the following annual report of receipts, disbursements, and funds in his hands for the year from January 11, 1902, to January 7, 1903, when the account was closed and balanced. FIFTH ANNUAL REPORT OF THE TREASURER 453 The receipts during the year were as follows : Dues of resident members, 1S99 $ 10.00 Dues of resident members, 1900 20.00 Dues of resident members, 1901 110.00 Dues of resident members, 1903 1,290.00 $1,430.00 Dues of non-resident members, 1901 $ 25.00 Dues of non-resident members, 1902 735 -oo Dues of non-resident members, 1903 55-oo Dues of non-resident members, 1904 5.00 820. 00 Sale of publications 120.4^ Payments on the principal of note of Episcopal Eye, Ear and Throat Hospital 2,000.00 Interest on deposits and investments 374-58 Bequests from Dr. S. C. Busey : Cash $r,o6S.oo Note of Robert B. Tj^ler 1,250.00 I share of Union Trust and Storage Co.'s stock 104.00 1 share Colonial Fire Insurance Co.'s stock 100.00 2 shares Capital Traction Co.'s stock 344.75 9 shares Washington Sanitary Improvement Co.'s stock 90.00 1 1 shares Columbia Title Insurance Co.'s stock 50.88 1 1 shares People's Fire Insurance Co.'s stock 66.00 2 shares stock of the Scheutzen Park Land and Bldg. Ass'n 88.00 3,061.63 Sale of stocks from estate of Dr. S. C. Busey, as follows : II shares People's Fire Insurance Co $ 66.00 ir shares Columbia Title Insurance Co 49-50 2 shares Capital Traction Co 250.00 I share Union Trust and Storage Co 107.00 472.50 $8,279.16 454 WASHINGTON ACADEMY OF SCIENCES The amounts and objects of the expenditures were as follows : Paid on a/c of expenses incurred in previous year, 1901 : Publications $ 162.07 Paid on a/c of expenses of the past year, 1902 : Secretary's and President's offices 95*56 Offices of Treasurer and Finance Committee 42.50 Rent of Cosmos Club Hall for Academy meetings ... 6.00 Presidential addresses 15-95 Lectures i H-59 Publications i»543-37 Investments : 800 shares stock of Washington Sani- tary Improvement Co 8,000.00 Investment to balance receipt of stocks and mortgage notedabove 1,993.63 $11,973.67 Statement of Account. Balanced from last annual statement $ 6,725.68 Receipts during the year 8,279.16 To be accounted for $15,004.84 Disbursements during the year 11 ,973.67 Balance on hand $3,031.17 The funds are on deposit with the American Security and Trust Co., drawing 2 per cent, interest. There is but one outstanding bill against the Academy known to the Treasurer. This amounts to $16.00. The assets of the Academy are as follows : Cash, as per above account $ 3,031.17 Note of Robert B. Tyler 1,250.00 809 shares stock of Washington Sanitary Improve- ment Co 8,090.00 1 share stock Colonial Fire Insurance Co 100.00 2 shares stock the Scheutzen Park Land and Building Association, par value $100.00, actual value doubt- ful, say $44.00 88.00 $12,559-17 Respectfully submitted, Bernard R. Green, January 15, 1903. Treasurer. SIXTH ANNUAL REPORT OF THE TREASURER 455 SIXTH ANNUAL REPORT OF THE TREASURER, 1904. To THE Washington Academy of Sciences : The Treasurer has the honor to submit the following annual report of receipts, disbursements, and funds in his hands for the year from January 7, 1903, to January 15, 1904, when the account was closed and balanced. The receipts during the year were as follows : Dues of resident members, 1901 $ 30.00 Dues of resident members, 1902 120.00 Dues of resident members, 1903 730.00 $ 880.00 Dues of non-resident members, 1901 10.00 Dues of non-resident members, 1902 30.00 Dues of non-resident members, 1903 665.00 Dues of non-resident members, 1904 5.00 710.00 Sale of publications $ S4.67 Principal of note of Robert B. Tyler 1,250.00 Interest on deposits and investments 5^2.40 Bequests from estate of Dr. S. C. Busey 2S544 Total receipts $3^722-5^ The amounts and objects of the expenditures were as follows : Paid on account of expenses incurred in previous year, 1902 : Secretary's and President's offices $ 42. So Rent of Cosmos Club hall for Academy meetings 3.00 Publications ^^4-45 210.25 Paid on account of expenses of the past year, 1903 : Secretary's and President's offices $ 23.75 Offices of Treasurer and Finance Committee 74-25 Publication of Joint Directory 332-4^ Receptions 40-oo Meetings 125.00 Check for dues declined by bank-drawer deceased 5.00 Publications ^^^^2.88 Total disbursements $1 18 1 3- 54 Proc. Wash. Acad. Sci., February, 1905. 456 WASHINGTON ACADEMY OF SCIENCES Statement of Account. Balance from last annual statement $3,031.17 Receipts daring the year, as above 3,723.51 To be accounted for $6,753.68 Disbursements during the year, as above 1,813,54 Balance on hand $4,940.14 These funds are on deposit with the American Security & Trust Co., drawing 2 per cent, interest. There are no important outstanding bills against the Academy known to the Treasurer, but it is expected that the Committee on Publications has incurred expenses on the current volume of Proceed- ings, being Vol. V, 1903, to the extent of probably $1,300, the limit authorized by the Board of Management, for which bills are expected soon to be presented. With this consideration the net assets of the Academy are as follows : Cash on deposit in bank $ 3,640.14 809 shares stock of Washington Sanitary Improvement Co. 8,090.00 1 share stock of Colonial Fire Insurance Co 100.00 2 shares stock of The Scheutzen Park Land & Building Association, par value $100, actual value doubtful, say $44 88.00 $11,918.14 Certificates of the above stock are in the safe deposit vaults of the Union Trust & Storage Co. Respectfully submitted, Bernard R. Green, Treaszirer. January 15, 1904. SEVENTH ANNUAL REPORT OF THE TREASURER, 1905. To THE Washington Academy of Sciences : The Treasurer has the honor to submit the following annual report of receipts, disbursements and funds in his hands for the year from January 15, 1904, to January 16, 1905, when the account was closed and balanced. The receipts during the year were as follows : SEVENTH ANNUAL REPORT OF THE TREASURER 457 Dues of resident members, 1901 $ 20 Dues of resident members, 1902 40 Dues of resident members, 1903 35 Dues of resident members, 1904 660 $ 755.00 Dues of non-resident members, 1901 $ 5 Dues of non-resident members, 1902 10 Dues of non-resident members, 1903 45 Dues of non-resident members, 1904 695.10 Dues of non-resident members, 1905 10 765.10 Sale of publications 102. S5 Interest on deposits and investments 514.30 Bequest from estate of Dr. S. C. Busey 30.00 Contribution towards publication of paper on the Genus Pinus , 600.00 Total receipts $2,767.25 The amounts and objects of the expenditures were as follows: Paid on account of expenses incurred in previous year, 1903 : Publications $ 726.65 Paid on account of expenses of past year, 1904 : Secretary's and President's offices 62.44 Offices of Treasurer and Finance Committee 54-75 Meetings ^ 03 • 73 Publications i ,430.46 Total disbursements $2,378.03 Statement of Account. Balance from last annual statement $4,940.14 Receipts during the year, as above 2,767.25 To be accounted for $7,707.39 Disbursements during the year, as above 2,378.03 Balance on hand $5,329.36 These funds are on deposit with the American Security and Trust Company, drawing 2 per cent, interest. Of this balance $3,470 belongs to the permanent fund and awaits investment, which the Treasurer has been authorized by the Board of Managers to make, subject to its approval. He expects shortly to make a good 5 per cent, loan, for which the opportunities lately have not been good. 458 WASHINGTON ACADEMY OF SCIENCES The only outstanding bills within the knowledge of the Treasurer are three, amounting to $353.32, one being for publications, amount- ing to $312.12, and the remainder for clerical services, etc., for the Secretary and Treasurer, amounting to $41.20. Besides these the probable cost of binding the Proceedings, Vol. VI of 1904, will be $150.00. Taking account of these items the assets of the Academy are at this date, as follows : Cash on deposit in bank $ 4,826.04 809 shares stock of the Washington Sanitary Improvement Company 8,090.00 1 share stock of Colonial Fire Insurance Co 100.00 2 shares stock ^ of the Scheutzen Park Land and Building Association, par value $100, actual value doubtful, say $44.00 88.00 $13,104.04 Respectfully submitted, Bernard R. Green, Treasti7'er. 1 Legacy from the Dr. Busey estate. OFP'ICERS AND COMMITTEES I9O3 459 WASHINGTON ACADEMY OF SCIENCES OFFICERS ELECTED JANUARY 15, 1903 President Charles D. Walcott Vice-Presidents F>-o7i2 the Atithropological Society Alice C. Fletcher Archccological Society John W. Foster Biological Society Barton W. E verm ann Botanical Society Albert F. Woods Che7nical Society C. E. Munroe Columbia Historical Society W J McGee Entomological Society W. H. Ashmead Geological Society G. K. Gilbert Medical Society S. S. Adams National Geografhic Society A. Graham Bell Philosophical Society J. Howard Gore Secrexary Treasurer Frank Baker Bernard R. Green Managers Class of igo4 Class of igoj Class of igo6 Marcus Baker L. O. Howard F. W. Clarke George M. Kober O. H. Tittmann Whitman Cross George M. Sternberg Carroll D. Wright C. Hart Merriam standing Committees — igo 3 Committee on Meetings George M. Sternberg Whitman Cross W. H. Holmes Frederick V. Coville S. S. Adams Committee on Building George M. Kober E. M. Gallaudet W. H. Holmes Arnold Hague David T. Day Committee on Publication Marcus Baker C. Hart Merriam Bernard R. Green Frank Baker Barton W. Evermann Committee on Functions G. K. Gilbert Frank H. Bigelow Richard Rathbun 460 WASHINGTON ACADEMY OF SCIENCES Committee on Finance Bernard R. Green L. O. Howard H. G. Ogden D. S. Lamb C. E. MUNROE Committee on Rules Carroll D. Wright J. Howard Gore A. F. Woods Coinmiitee ott Membership F. W. Clarke George AI. Kober C. Hart Merriam Committee on Relationsjo Other O rga niza tions Charles D. Walcott Carroll D. Wright C. Hart Merriam O. H. Tittmann OFFICERS AND COMMITTEES — I9O4 461 WASHINGTON ACADEMY OF SCIENCES OFFICERS ELECTED JANUARY 21. 1904 President Charles D. Walcott Vice-President Frotn the Anthropological Society W. H. Holmes Archceological Society Jon^ W. Foster Biological Society Barton W. Evermann Botanical Society Frederick V. Coville Chemical Society C. E. Munroe Colujnbia Historical Society W J McGee Entomological Society H . G . D y ar Geological Society G. K. Gilbert Medical Society \V. W. Johnston National Geographic Society A. Graham Bell Philosophical Society Richard Rathbun Society of Americati Foresters. ...Giffokd Pinchot Secretary Treasurer Frank Baker Bernard R. Green Managers Class of I go^ Class of i gob Class of i gay L. O. Howard F. W. Clarke Geo. M. Kober O. H. Tittmann C. W. Hayes Gifford Pinchot Carroll D. Wright G. W. Littlehales F. A. Lucas standing Committees— 1904 Committee on Meetings C. W. Hayes D. S. Lamb F. A. Lucas Whitman Cross L. A. Bauer Committee on Publication Barton W. Evermann C. Hart Merriam Bernard R. Green Frank Baker C. F. Marvin Committee on Building John W. Foster O. H. Tittmann Gifford Pinchot Willis L. ]Moore Arnold Hague Committee on Functions G. K. Gilbert G. W. Littlehales C. E. Munroe 462 WASHINGTON ACADEMY OF SCIENCES Committee on Finance Bernard R. Green L. O. Howard E. M. Gallaudet C. W. Richardson Swan M. Burnett Committee on Rules Carroll D. Wright J. Howard Gore Miss Alice C. Fletcher Committee on Membership F. W. Clarke Samuel S. Adams C. Hart Merriam Committee on Relations to Other Orga^iizations Charles D. Walcott Carroll D. Wright C. Hart Merriam O. H. TiTTMANN OFFICERS AND COMMITTEES I9O5 463 WASHINGTON ACADEMY OF SCIENCES OFFICERS ELECTED JANUARY 19. 1905 President Charles D. Walcott Vice-Presideuta Frojn the Afithropological Society D. S. Lamb Archceological Society John W. Foster Biological Society V. H. Knowlton Botanical Society J. N. Rose Chemical Society F. W. Clarke Columbia Historical Society A. R. Spofkord Entomological Society H. G. Dyar Foresters Society Gifford Pinchot Geographic Society A. Graham Bell Geological Society G. K. Gilbert Medical Society S. S . Adams Philosophical Society C W. Littlehales Secretary TreasxtTer Frank Baker Bernard R. Green Managers Class of igob Class of 1907 Class of igo8 C. W. Hayes Geo. M. Kober L. O. Howard L. A. Bauer Frederick V. Coville O. H. Tittmann C. Hart Merriam J. S. Diller Barton \V. Evermann Standing Committees — X905 Committee on Rules Committee on Functions G. W. Littlehales G. K. Gilbert J. Howard Gore O. H. Tittmann Richard Rathbun F. W. Clarke Committee on Publication Committee on Meetings Barton W. Evermann C. W. Hays C. Hart Merriam D- S. Lamb Frank Baker Frederick V. Coville C. F. Marvin Whitman Cross J. S. Diller L. A. Bauer 464 WASHINGTON ACADEMY OF SCIENCES Commiltee on Buildings O. H. TiTTMANN GiFFORD PiNCHOT Arnold Hague G. L. Magruder J. G. Hagen Cofntniflee on Membership Geo. M. Kober Willis L. Moore H. C. Dyar Committee on Finance Swan M. Burnett Bernard R. Green E. M. Gallaudet C. E. MUNROE Robert Fletcher Committee on Relations of Academy to Other Organizations Charles D. Walcott A. Graham Bell J. N. Rose F. W. True E. W. Nelson MEMBERS 465 MEMBERS OF THE WASHINGTON ACADEMY OF SCIENCES January 21. 1905 Patrons Mr. Edward Henry Harriman, Union Pacific Bldg., 120 Broadway, New York City. Mrs. Phoebe Apperson Hearst, 1400 New Hampshire Ave., Washington, D. C. Mrs. Henry Lee Higginson, 191 Commonwealth Ave., Boston, Mass. Mrs. Gardiner Greene Hubbard, 1328 Connecticut Ave., Washington, D. C. Mr. Henry Cleveland Perkins, 1701 Connecticut Ave., Washington, D. C. Mr. GifEord Pinchot, Department of Agriculture., Washington, D. C. Mr. James Wallace Pinchot, 1615 Rhode Island Avenue, Washington, D. C. Mr. Thomas Francis Walsh, 2020 Massachusetts Ave., Washington, D. C. Resident Members Abbe, Cleveland, 2017 I Street, N. W. Acker, Geo. N., 913 i6th Street, N. W. Adams, Henry, 1603 H Street, N. W. Adams, S. S., I Dupont Circle. Adler Cyrus, Smithsonian Institution. Ashmead, William H., National Museum. Bailey, Vernon, Department of Agriculture. Baker, Frank, Zoological Park. Bauer, L. A., Coast Survey. Becker, G. F., Geological Survey. Behrend, Edwin, 1 214 K Street, N. W. Bell, Alexander Graham, 1331 Connecticut Ave., N. W. Bermann, I. S. L., Pennsylvania Ave. and Washington Circle. Bigelow, Frank H., Weather Bureau. Bovee, J. Wesley, 1404 H Street, N. W. Bo wen, W. S., 1228 i6th Street, N. W. Brooks, Alfred H., Geological Survey. Bryan, Joseph H., 1644 Connecticut Ave., N. W. Burnett, Swan M., 916 Farragut Square, N. W. Carr, W. P., 1418 L Street, N. W. Carroll, James J., Army Medical Museum. 466 WASHINGTON ACADEMY OF SCIENCES Chesnut, V. K., State Agricultural College, Bozeman, Mont. Clarke, F. "W., Geological Survey. Cook, G. Wythe, 3 Thomas Circle, N. W. Cook, O. F., Department of Agriculture. Coquillet, D. W., National Museum. Coville, Frederick V., Department of Agriculture. Cross, Wliitinan, Geological Survey. Cushman, Allerton S., Department of Agriculture. Darton, N. H., Geological .Survey. Davis, A. P., Geological Survey. Day, Arthur L., Geological Survey. Day, David T., Geological Survey. Diller, J. S., Geological .Survey. Dyar, Harrison G., National Museum. Eichelberger, William S., Naval Observatory. Eimbeck, William, Coast Survey. Eldridge, George H., Geological .Survey. Emmons, S. F., Geological Survey. Evermann, Barton W., Bureau of Fisheries. Fisher, A. K., Department of Agriculture. Fletcher, Alice C, 214 I St .Street, S. E. Fletcher, Robert, The Portland. Flint, James M., * Stoneleigh Court. Foster, John W., 1323 i8th Street, N. W. French, William B., 506 East Capitol Street. Fry, Henry D., 1601 Connecticut Ave., N. W. Gallaudet, E. M., I Kendall Green, N. E. Gannett, Henry, Geological Survey. Gilbert, G. K., Geological Survey. Gill, Theodore, Smithsonian Institution. Girty, George H., Geological Survey. Gore, J. Howard, George Washington University. Green, Bernard B,., Library of Congress. Greene, Edward L., National Museum. Hagen, J. G., Georgetown College Observatory. Hague, Arnold, Geological Survey. Harris, R. A., Coast Survey. Harris, W. T., 1360 Yale .Street, N. W. Hay, W. P., 1925 14th Street, N. W. Hayes, C. W., Geological Survey. Hickling, D. P., 1304 Rhode Island Ave., N. W. Hill, Robert T., 1738 Q .Street, N. W. MEMBERS 467 Hodge, F. W., Smithsonian Institution. Holmes, W. H., Bureau of American Ethnology. Hopkins, A. D., Bureau of Entomologry. Howard, L. O., Department of Agriculture. Hyde, John, Department of Agriculture. Johnson, Jos. Taber, 926 17th Street, N. W. Jung, F. A. R., 1229 Connecticut Ave., N. W. Easson, John A., 1726 I Street, N. W. Keith, Arthur, Geological Survey. Kendall, William C, Bureau of Fisheries. Kerr, James, 1711 H Street, N. W. King, A. F, A., 1315 Massachusetts Ave., N. W. Knowlton, F. H., National Museum. Kober, Geo. M., 1600 T Street, N. W. > Kiibel, Stephen J., Geological Survey. Lamb, D. S., The Cumberland. Langley, S. P., Smithsonian Institution. Lindgren, W., Geological Survey. Littlehales, G. "W., Hydrographic Office, Navy Depart- ment. Lucas, F. A., Brooklyn Institute of Arts and Sci- ences, Brooklyn, N. Y. McArdle, Thos. E., 1 1 20 i6th Street, N. W. McGee, W J, 1901 Baltimore .Street, N. W. Magruder, G. L., 4 Jackson Place, N. W. Marlatt, C. L., Department of Agriculture. Marvin, C. F., Weather Bureau. Matthews, Washington, 1262 New Hampshire Ave, N. W. Melville, Geo. W., 620 N. i8th St., Philadelphia, Pa. Merriam, C. Hart, 1919 i6th Street, N. W. Metzerott, John H., nioF Street, N. W. Miller, Gerrit S., Jr., National Museum. Mooney, James, Bureau of Ethnology. Moore, Willis L., Weather Bureau. Moran, John F., 2426 Pennsylvania Ave., N. W. Morgan, James Dudley, 919 15th Street, N. W. Mosman, A. T., Coast Survey. Munroe, Chas. E., George Washington University. Nelson, E. W., Department of Agriculture. ISTewell, F. H., Geological Survej'. Nordhoff-Jung, Sofie A., 1229 Connecticut Ave., N. W. Ogden, H. G., Coast Survey. Osgood, Wilfred H., Department of Agriculture. Palmer, T. S., Department of Agriculture. 468 WASHINGTON ACADEMY OF SCIENCES Parker, Edward W., Geological Survey. Paul, H. M., 2015 Kalorama Ave., N. W. Pieters, A. J., Department of Agriculture. Pinchot, Gifford, Department of Agriculture. Pratt, J. F., Coast Survey. Putnam, G. K.., Coast Survey. Bansome, Frederick L., Geological Survey. Rathbun, Mary J., National Museum, Rathbun, Richard, Smithsonian Institution. Reyburn, Robert, 714 13th Street, N. W. Ricliardson, Chas. W., 1317 Connecticut Ave., N. W. Richardson, Harriet, 1864 Wyoming Ave., N. W. Richmond, C. W., Smithsonian Institution. Ridgway, Robert, Smithsonian Institution. Rose, J. N., National Museum. Ruffin, Sterling, 1023 Vermont Ave., N. W. Salmon, D. E., Department of Agriculture. Schwarz, E. A., Department of Agriculture. See, T. J. J., Mare Island Navy Yard, Vallejo, Calif. Shands, A. R., 1319 New York Ave., N. W. Shute, D. K., 1719 De .Sales Street, N. W. Smith, George O., Geological Sur^-ey. Smith, Hugh M., Bureau of Fisheries. Spencer, Arthur C, Geological .Survey. Spofford, A. R., Library of Congress. Spurr, Josiah E., Geological Survey. Stanton, T. W., National Museum. Stejneger, L., National Museum. Sternberg, Geo. M., 2144 California Ave., N. W. Stevenson, Mrs. Matilda C, 1307 F Street, N. W. Sudworth, George B., Department of Agriculture. Tittmann, O. H., Coast Survey. True, F. W., National Museum. Tweedy, Frank, Geological Survey. TJlrich, Edward 0., Geological Survey. Van Rensselaer, John, 2 Thomas Circle, N. W. Vaughan, George T., 1718 I Street, N. W. Vaughan, Thomas Wayland, Smithsonian Institution. Walcott, Chas. D., Geological Survey. Ward, Lester F., National Museum. Warder, R. B., Howard Universit}-. Wead, Chas. K., Patent Office. MEMBERS 469 Webber, Herbert J., Department of Ag^riculture. Weed, Walter H., Geological Survey. Wells, Walter A., 1 1 33 I4tli St., N. W. White, David, National Museum. Wiley, H. W., Department of Agriculture. Willis, Bailey, Geological Survey. Wilmer, W. H., 1610 I Street, N. W. Wilson, H. M., Geological Survey. Woods, A. F., Department of Agriculture. Woodward, Wm. C, 464 lyouisiana Ave., N. W. Wright, Carroll D., Clark University, Worcester, Mass. Wyman, Walter, 3 B Street, S. E. Non-Resident Members Adams, Frank D., McGill University, Montreal, Canada. Allen, J. A., American Museum of Natural His- tory, New York City. Andrews, Ethan A., Johns Hopkins University, Baltimore, Md. Atherton, George W., State College, Penn. Atkinson, Geo. F., Cornell University, Ithaca, N. Y. At water, W. O., Wesleyan University, Middletown, Conn. Bancroft, W. D., Cornell University, Ithaca, N. Y. Bangs, Outram, 240 Beacon Street, Boston, Mass. Baskerville, Charles, College of the City of New York, New York City. Batchelder, Charles Foster, Boston Society of Natural History, Berkeley Street, Boston, Mass. Bessey, C. E,, University of Nebraska, Lincoln, Neb. Birnie, Rogers, N. Y. Arsenal, Governor's Island. New York City. Bixby, W. H., U. S. Engineers' Mich. Office, Detroit, Blunt, Stanhope E., Rock Island Arsenal, Illinois. Branner, John C, Stanford University, Calif. Brewster, William, 145 Brattle Street, Cambridge, Mass. Bumpus, H. C, American Museum of Natural His- tory, New York City. Campbell, Douglas H., Stanford University, California. Cattell, J. McKeen, Garrison-on-Hudson, N. Y. Chamberlin, T. C, University of Chicago, Chicago, 111. Chandler, C. F,, Columbia University, New York City. Chapman, Frank M., American Museum of Natural His- tory, New York City. Clark, Wm. B., Johns Hopkins University, Baltimore, Md. Clarke, John M., State Hall, Albany, N. Y. Cohen, S. Solis, 1525 Walnut Street, Philadelphia, Pa. 470 WASHINGTON ACADEMY OF SCIENCES Coleman, Arthur P., University of Toronto, Toronto, Ontario. Conner, P. S., 215 W. Ninth Street, Cincinnati, Ohio. Coulter, Jolin M., University of Chicago, Chicago, 111. Crafts, J. M., 59 Marlborough Street, Boston, Mass. Crozier, "Wm., Ordnance Office, War Department. Washington, D. C. Dabney, Charles W., University of Cincinnati, Cincinnati, Ohio. Dana, E. S., 24 Hillhouse Avenue, New Haven, Connecticut. Davenport, Chas. B., Cold Spring Harbor, N. Y. Dike, Samuel W., 113 Hancock Street, Auburndale, Mass. Dolbear, A. E., Tufts College; Mass. Dudley, C. B., Drawer 334, Altoona, Penn. Eastman, J. E.., Andover, New Hampshire. Eigenmann, Carl H., 650 Atwater Ave., Bloomington, Ind. Eliot, Charles W., Harvard University, Cambridge, Mass. Elliot, Daniel G., Field Columbian Museum, Chicago, Illinois. Ely, B. T., University of Wisconsin, Madison, Wisconsin. Emerson, B. K., Amherst College, Amherst, Mass. Evans, Alex. W., 2 Hillhouse Ave., New Haven, Conn. Fairchild, Herman L., University of Rochester, Rochester N. Y. Farnam, H. W., 43 Hillhouse Ave., New Haven, Conn. Farrand, Livingston, Columbia University, New York City. Faxon, Walter, Museum of Comparative Zoology, Cambridge Mass. Fisher, Irving, Yale University, New Haven, Conn. Fitz, R. H., 18 Arlington Street, Boston, Mass. Forbes, S. A., University of Illinois, Urbana, 111. Forchheimer, Frederick, The Ortig, 4th and Sycamore Sts., Cincinnati, Ohio. Frankforter, G. B., Minneapolis, Minn. Freer, Paul C, Box 580, Manila, P. I. Gage, Simon H., Cornell University, Ithaca, N. Y. Garman, Samuel, Museum of Comparative Zoology, Cambridge, Mass. Garrison, George P., 2600 White's Ave., Austin, Texas. Gilman, Daniel C, 614 Park Ave., Baltimore, Md. Gomberg, Moses, University of Michigan, Ann Arbor, Mich, Goodale, Geo. L., Harvard University, Cambridge, Mass. Graves, Henry S., 68 Trumbull vStreet.New Haven, Conn. Gray, Thomas, Rose Polytechnic Institute, Terre Haute, Ind. 0 MEMBERS 471 Hall, Asaph, Jr., Detroit Observatory, Ann Arbor, Mich. Halsted, Geo. B., Gambier, Ohio. Halsted, Wm. S., 1 201 Eutaw Street, Baltimore, Md. Hp-ner, William R., Chicago University,fChicago, 111. Heath, Harold, Stanford University,'California. Hilgard, E. "W., Berkeley, California. Hitchcock, C. H., Dartmouth College, Hanover, N. H. Holmes, J. A., Chapel Hill, N. C. Howe, Herbert A., Chamberlin Observatory, University Park, Colorado. Howe, James Lewis, Washington and Lee University, Lex- ington, Va. Hussey, WilliamL J., Lick Observatory, Mt. Hamilton, California. Iddings, J. P., University of Chicago, Chicago, 111. Jacobi, Abraham, no West 34th Street, New York City. Janeway, E. G., 36 West 40th Street, New York City. Jenkins, Oliver P., Stanford University, California. Jordan, David Starr, Stanford University, California. Kahlenberg, Louis, 234 Lathrop Street, IMadison, Wis. Keen, W. W., 1729 Chestnut Street, Philadelphia, Penna. Kelley, Howard A., 1418 Eutaw Place, Baltimore, Md. Kellogg, Vernon L., vStauford University, California. Kemp, James P., 211 West 139th Street, New York City. Kingsley, J. S., Tufts College, College Hill, Mass. Klotz, Otto J., Department of the Interior, Ottawa, Canada. Lawson, Andrew C, Univer.sity of California, Berkeley, California. Long, J. H., 2421 Dearborn Street, Chicago, 111. Macbride, Thomas H., Iowa City, Iowa. Macfarlane, Alexander, Gowrie Grove, Chatham, Ontario. Mallet, J. W., University of Virginia, Charlotts- ville, Va. Mead, A. D., Brown University, Providence, R. I. Meek, S. E., Field Columbian Museum, Chicago, Illinois. Miller, W. Lash, 50 St. Alban Street, Toronto, Canada. Moore, Clarence B., 1321 Locust Street, Philadelphia, Pa. Morley, E. W., Adelbert College, Cleveland, Ohio. Morse, H. N., Johns Hopkins University, Baltimore, Md. Miinsterberg, Hugo, 7 Ware Street, Cambridge, Mass. Muir, John, Martinez, California. North, S. N. D., 1414 2ist Street, N. W., Washington, D. C. 472 WASHINGTON ACADEMY OF SCIENCES Northrop, Cyrus, University of Minnesota, Minneap- olis, Minn. Nutting, C. C, University of Iowa, Iowa City, Iowa. Ortmann, A, E., Carnegie Museum, Shenley Park, Pittsburg, Pa. Osbom, H. F., American Museum of Natural His tory. New York City. Parker, Geo. H., Harvard University, Cambridge, Mass. Peirce, Benj. Osgood, Jefferson Physical Laboratory, Cam- bridge, Mass. Pickering, Edward C, Hars'ard University Observatory, Cambridge, Mass. Pirsson, L. V., Yale University, New Haven, Conn. Pritchett, H. S., Massachusetts Institute of Tech- nology, Boston, Mass. Putnam, F. W., Peabody Museum, Cambridge, Mass. Raymond, C. W., Room E 8, Army Building, 39 White- hall Street, New York City. Raid, H. F., Johns Hopkins University, Balti- more, Md. Remsen, Ira, Johns Hopkins University, Balti- more, Md. Richards, Mrs. Ellen H., Massachusetts Institute of Tech- nology, Boston, Mass. Richards, Theodore W., 15 Pollen .Street, Cambridge, Mass. Richardson, Clifford, 122 East 34th Street, New York City. Rising, "W. B., University of California, Berkeley, Calif. Ritter, Wm. E., University of California, Berkeley, Calif. Robinson, Benj. L., Gray Herbarium, Cambridge, Mass. Rotch, A. Lawrence, Blue Hill Observatory, Hyde Park, Mass. Roth, Filibert, University of Michigan, Ann Arbor, Mich. Russell, H. Ii., Madison, Wis. Rutherford, E., McGill University, Montreal, Can- ada. Salisbury, Rollin D., University of Chicago, Chicago, 111. Sargent, C. S., Arnold Arboretum, Jamaica Plais, Mass. Scott, W. B., Princeton, N. J. Scripture, E. W., \''ale University, New Haven, Conn. Searle, Geo. M., Catholic University, Brookland. D. C. Seligmann, E. R., Columbia University, New Y'ork City. Senn, Nicholas, 532 Dearborn Street, Chicago, 111. Setchell, W. A., University of California, Berkeley, Calif. Shattuck, Frederick C, 135 Marlborough Street, Boston, Mass. Smith, John B., Agricultural Experiment Station, New Brunswick, N. J. Solly, S. E., 2 North Cascade Ave., Colorado Springs, Colo. MEMBERS 473 Stearns, R. E. C, White, Israel C, 1025 East i8th Street, Los Angeles, Morgantown, W. Vri. ^^^^^- Whitfield, R. P., Stengel, Alfred, American Museum of Natural His- 181 1 Spruce Street, Philadelphia, torj'. New York City. ^^""- Willcox, W. F., Stevenson, J. J., Cornell University, Ithaca, N. Y. University Heights, New York City. Williams, H. S., Stockton, Chas. G., Yale University, New Haven, Conn. 436 Franklin Street, Buffalo, N. Y. Williams, Talcott, Stone, Ormond, 916 Pine Street, Philadelphia, Penn. University of Virginia, Charlottes- willoughby, William F., ville, Va. „ , t^ . tt San Juan, Porto Rico. Sutton, B. S., _-.., T o ' . ' . , Wilson, J. C, 341 Sixth Ave., Pittsburg, Penn. „,, , ^ ^^ ^ t^i -i j 1 t.- -o ■^^ ' ^' 1509 Walnut Street, Philadelphia, Pa. Thayer, William S., _^., ,„ ._ ^ ' ' Wilson, W. P., 3 West Franklin Street, Baltimore, 233 South 4th Street, Philadelphia, Pa. Tourney, J. W., ^°°'*' ^- ^^ •^ ' ^., „ 815 St. Paul Street, Baltimore, Md. 459 Prospect Street, New Haven, Conn. Woodward, R. S., Townsend, Charles H., Carnegie Instituton, Washington. New York Aquarium, Battery Park, New York City. Woodworth, William McMichael, Van Hise, Chas. R., Museum of Comparative Zoology, . ,, ,. Cambridge, Mass. University of Wisconsin, Madison, Wis. Worcester, Dean C, Wadsworth, F. L. O., ^^^"^^^' ^- ^• Allegheny Observatory, Allegheny, Wright, Arthur W., Penn. 73 York Square, New Haven, Conn. Wells, D. Collins, Ziwet, Alexander, Dartmouth College, Hanover, N. H. University of Michigan, Ann Arbor Wheeler, Wm. M., ^^'^^■ American Museum of Natural His- tory, New York City. Summary Patrons 8 Resident Members 159 Non-Resident Members i59 Counted twice 1 Total 325 INDEX. Note.— New names in black-face type, synonyms in italics. abdominalis, Abudefduf 390 Abudefduf abdominalis 390 marginatus 390 saxalilis 390 Aca)ithurus aliala 403 strigosns 402 adspersus, Paralichthys 423 adustus, Gobiesox 422 Julidio 396 Pseudojulis396 affiuis, Chilomycterus 414 agassizi, Serranus 370 Xenichthys 376 agassizii, Cratinus 370 alalunga, Germo 361 Sco7nber 2,(>i albemarleus, Nexilosus 391 albidactylus, Exocceius 352 albomaculatus, Paralabrax 370 Serranus 370 aliala, Acatilhurus 405 Hepatus 403 Teuthis 403 Alticus atlanticus 419 chiostictus 419 Alutera script a 410 Amia atradorsata 367 atricauda 367 analogus, Epinephelus 367 Kyphosus 384 Pimelepterus 384 Angelichthys iodocus 402 angulosus, Batistes 407 Canthiderniis 407 angusticeps, Spheroides 412 Tetodron 412 Anisotrermis bilineatus 376 interruptus 377 scalpularis 377 surinamensis 376 annulatus, Spheroides 412 Tetrodon 412 Antennariidce 424 Antennarius tagus 424 Anihias mulifasciaius 373 Apogon atradorsatus 367 atricaudus 367 Apogonichthyidae 367 Arbaciosa truncata 422 arcifrons, Pomacentrus 389 Archosargus pourtalesii 380 arge, Kuhlia 366 argentiventris, Lutianus 375 Mesoprion 375 Neomccnis 375 arundelii, Gobius 416 asper, Labrus 394 atlanticus, Alticus 419 Rupiscartes 419 Salarias 419 atradorsata, Amia 367 atradorsatus, Apogon 367 atricauda, Amia 367 atricaudjis, Apogon 367 atramentatus, Symphurus 423 azalea, Runula 419 Azurina eupalama 385 bairdii, Microspathodon 390 Pomacentrus 390 baloa, Hetnirhamphus 2,50 Balistes angulosus 407 longissimus 407 longus 407 men to 408 sandwichiensis 409 script us 410 verres 406 Balistidae 406 Batrachoidida? 418 Batrachus margaritatus 418 Beaver Dams in Colorado, Some interest- ing 429 Bell, Ruby G. 203 bicolor, Rypticus 373 Smecticus 373 bilineatus, Afiisotremus 376 bipintiulata, Seriola 362 bipinnulatus, Hlagatis 362 birostris, Manta 346 Raia 346 bispinosus, ]Melichtbys 408 Blenniidae 418 Bodianus diplotsenius 391 echlancheri 392 Branchiostoma elongatum 342 Branchiostomidtc 342 brevoorti, Euleptorhamphus 350 bristolae, Emmnion 418 Brotulidte 421 caballns, Caranx 364 Calamus taurinus 379 cat i for ni en sis, Chilomycterus 414 Callj'odon noyesi 397 perrico 397 Calotomus xenodon 397 camuruni, Ostracion 411 canescens, Cluctodon 402 Zanclus 402 canthariuum, Pristipoma 397 cantharinus, Orthopnstis 379 (475) 476 INDEX Cantherines carola: 409 7iasulus 409 sandwichiensis 409 Canthidermis angulosus 407 Canthigaster lobatiis 412 capistratus, Pachynaihus 406 Carangridae 362 Carangoides oj-thogrammus 365 Caranx cabal 1 us 364 ferdau 365 latus 364 lugubris 365 inarginatus 364 melampygus 365 orthogrammus 365 scombrinus 362 svmmetricus 363 (Trachjirus) aivieri 363 Carcliarias galapagensis 343 obesjis 344 platyrhynchus 344 Carcharhinus platyrhynchus 343 ; 344 carolcB, Cantherines i,