% Wal Oe ‘ A Sts) bh one Ny Lay, JOURNAL OF MORPHOLOGY FouNDED By C. O. WHITMAN MANAGING EDITOR C. E. McCLUNG ASSOCIATE EDITORS Gary N. CALKINS J. S. KINGSLEY Columbia University University of Illinois E. G. CONKLIN M. F. GuyEr Princeton University University of Wisconsin C. A. Kororp F. R. LInuie University of California University of Chicago VOLUME 35 Wo. PaTTEN Dartmouth College W. M. WHEELER Bussey Institute, Harvard University J. T. PATTERSON University of Texas MARCH, JUNE, SEPTEMBER 1921 THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY » PHILADELPHIA CONTENTS Now i, MARCH PETER OKKELBERG. The early history of the germ cells in the brook lamprey, Entosphenus wilderi (Gage), up to and including the period of sex differen- tiation. Four text figures and twelve plates (seventy-eight figures)...... il R. Kupo. Studies on Microsporidia, with special reference to those parasitic in mosquitoes. One text figure and five plates (one hundred seventeen AU GERITES) Bia cayor ese teR IN omens oak sh enatinn eC oe sic a stata ale says she seeuaye ara ls cements Sates eaters 153 Hrirowo Iro. On the metamorphosis of the malpighian tubes of Bombyx MOVIN. cMUnbeCme tIPMGE Se vias es rd chs oe tS ected o Okie tink Weisel Mee are onemaets 195 R. W. Suuretpt. Observations on the cervical region of the spine in chelomeanseeMiviesti@unesss svsomankinsceieoen oclee soe mieriac aaa a, Ch To: 213 A. G. Poutman. The position and functional interpretation of the elastic ligaments in the middle-ear region of Gallus. Twelve figures............. 229 No. 2) JUNE JAMES RoLLIN SLONAKER. The development of the eye and its accessory parts in the English sparrow (Passer domesticus). Ten text figures and seven- Deensplabess (a @umes: Letom OS) Merits teases Weel Seg etoe bPscycuetskohe ois sxe! cv ated 263 W. Haroutp LereuH-SHarre. The comparative morphology of the secondary sexual characters of Elasmobranch fishes. II. The claspers, clasper siphons, and clasper glands. Fifteen text figures...................0.5-5- 359 Harry H. Cuaruton. The spermatogenesis of Lepisma domestica. Six platess Munety=tivelauUes)) tae css ias & «cakes mie Sass ole ccs seep eeeteesien «hee. + oa, siete 381 JAMES ErNEsT KinprEp. The chondrocranium of Syngnathus fuscus. Four- [AST aed Bea DNASE cde caer breld ni mao tete Cues cic Chee care eoonctand Bole cS treks ata Pree cnnES Ohare NAA A25 E. Eveanor Carotuers. Genetical behavior of heteromorphic homologous chromosomes of Circotettix (Orthoptera). Five plates (thirty-five figures). 457 No. 3. SEPTEMBER Gencuo Fusimura. Cytological studies on the internal secretory functions in the human placenta and decidua. One diagram and two double plates (TINS EISID-S UTA TES). Geena c) «oho eG ORES REMI os oipi clic coo Sloth alee oI IOle earn ’cidiare biti Orc 485 A. M. Reese. The structure and development of the integumental glands Gimnhe Crocodilias (Six plates: forty-six Geures)) 4.0... fon. 4. cise sists elevely oi 581 \GL45 JOURNAL OF MORPHOLOGY, VOL. 35, No. 1, MARCH, 1921 Resumen por el autor, Peter Okkelberg, Universidad de Michigan. Historia de las células germinales de la lamprea de arroyo, Entosphenus wilderi (Gage) hasta el periodo de diferenciacion. Las células germinativas se segregan de las restantes en una época muy temprana de la vida del animal, aun antes de haberse formado definitivamente las hojas germinales. Pueden recono- cerse por primera vez cuando el mesodermo se separa del endo- dermo (embrién de unas 191 horas de edad). Las células mas- culinas y femeninas definitivas no reconocen otro origen que el de las células germinales primordiales; las células germinales tampoco toman parte en la formacién de estructuras somaticas. Muchas de las células germinales degeneran y desaparecen en cada individuo. Durante el periodo de diferenciacién sexual las células ger- minales de cada glindula germinal son claramente de dos tipos: Unas que presentan marcada tendencia hacia una divisién continua (células catabdélicas) y otras que tienden a crecer (células anabdélicas). El autor considera a las primeras como posesoras de una potencialidad masculina, y a las segundas como femeninas. La proporcién relativa de células anabdlicas y cataboélicas determina si la larva ha de ser un macho o una hembra. Las observaciones del autor parecen justificar la con- clusion de que cada larva de esta especie lleva la potencialidad para producir los dos sexos, y que el sexo, por consiguiente, no se fija de modo irrevocable en el momento de la fecundaci6n. En el trabajo se describen las diversas estructuras nucleares y citopl4smicas y los cambios que sufren durante las diversas fases del desarrollo. Translation by José F. Nonidez Cornell Medical College, New York AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, DECEMBER 6 THE EARLY HISTORY OF THE GERM CELLS IN THE BROOK LAMPREY, ENTOSPHENUS WILDERI (GAGE), UP ‘TO AND INCLUDING THE PERIOD OF SEX DIFFERENTIATION PETER OKKELBERG Zoological Laboratory, University of Michigan FOUR TEXT FIGURES AND TWELVE PLATES (SEVENTY-EIGHT FIGURES) CONTENTS [GaKROCHICIIOINS 454 obo cahoots sudo unos su cben nooo. on Suen DOeH Od Usage ams 2 Spawning habits and life-cycle of Entosphenus wilderi..................-- 3 /Ne- VOR NVDINU OV? AEM OLS. Weck Gs eledinue dd ores Cho acecle one oe oo AG uasp ocr co oc 3 Ep SRE COLE pies Ata eta teehee hap roy PNM A eof st ed data oy ) = © eteRble el wie ie) als 39 CeePerodionsecond ary GkvIstomars 2.209 ne Sotietenils ois va.cs vga tae rene 41 f Observations on Entosphenuswilderi 32 222 Sina. . Slice cms a 41 2. Historical review and theoretical discussion............-.-++++: 45 3. Summary of the period of secondary division.................. 49 DiezeriOdsotese ximenentlatlOMiis ups ie ein tleieines aqcrer ete ciciele te eure) 50 IeGenens eSabement egy as orks toes Leber btn hte cictoitcate’s of & exahe.s sda « 50 2. Sex characters in the aaly brook lamprey. . See ee OO 3. Changes in the germ gland during the period it sex © differentiation 51 @, Changes duning mit OSIS;: Sse ts a ce javoe ae cicsiesie ets Saale ta so 52 6. The synapsis phase of the oocytes... .......... 0. cc neeeeee 55 c. History and fate of the germ cells during the period of sex difreremiraciony. yas ercmels ase eless's oj s1g ste silo ches oie =e! b= 67 d. Literature on the period of sex differentiation in the lamprey U2, e. Other cases of juvenile hermaphroditism among vertebrates 73 PP DISCMSSION AP ee Seca te ph ASS eialctahe i Sc aih ele afbiabe a wietabertiele wie. Saye 78 APPPresenuistatus Oimuhe:sex POD leMn meter icicles letlekel: -fe)= eter 84 1 i) PETER OKKELBERG 5. Discussion of the hermaphroditic condition found in the lamprey in connection with other sex phenomena not easily explained byRCuUrrent-theOnies's pfx. ssincers Seca Melia cies oe ro eee ee 88 a. Normal hermaphroditism): 00 cece enc ieee ee oe 88 b. Alternation of the hermaphroditic and dioecious condition 90 c. The.effect of delayed fertilization om sex .................. 92 d. Hermaphroditism and sex reversal due to external conditions 94 e. Hermaphroditism as a result of hybridization............. 98 ys Sporadic hermapiirodtiie’. 2 ac, cere a eee ee a oe ee eee 102 g. Hermaphroditism as a result of hormone action........... 103 h. Hermaphroditism as a result of parasitism................ 103 i. Sex in parthenogenetic animals’... /......./.055.52 0.2 e eee gq. Variation in RexmethOss is Scueiiceae neke one a ones sd e eeee 105 6. General conclusions in regard to the problem of sex determination 106 @eneralisummanysoiobsenvallonsee aeeneee eee eens oe eens 115 ConGlapipns by. 245 tobe a ee Stas is asics Asch Oe aoe OIE Eee 116 Literature, G1teds.., Jems. ce gis nic cisis deacon theicne emis NG EE oan arate ie See eee 118 INTRODUCTION Owing to their bearing on current theoretica! questions, cer- tain phases of the germ-cell cycle in vertebrates have received more attention than others. Maturation has been fully studied because it appears to involve a redistribution of parental heredi- tary factors. The small size and the brief maturation period of the spermatocytes make it possible to obtain all stages very easily on a single slide, whereas much time and material are needed to obtain the stages of maturation in the much larger oocytes. Since it usually is assumed that the maturation phe- nomena are practically identical in the two sexes, the male has been commonly selected for the study of this period because it presents lesser technical difficulties. A second phase on which attention has been focused in recent years, because of its bear- ing on the question of germ-plasm continuity, is that of the origin and early development of the germ cells. The period of | differentiation of the male and female sex cells from the pri- mordial germ cells has been relatively neglected, apparently because of the belief that in all animals sex is irrevocably deter- mined at or before fertilization through the agency of sex chro- mosomes. The result is that we have much literature on isolated periods in the germ-cell history of vertebrates, chiefly in the male sex, but few comprehensive accounts. : GERM-CELL HISTORY IN THE BROOK LAMPREY 3 The work of Brock (81), Schreiner (04), R. Hertwig (05, 06, ’07), and others has shown that in both sexes of certain species of vertebrates there is a tendency toward hermaphrodi- tism at the period of differentiation of the primordial germ cells— an indication that sex is not irrevocably determined at the time of fertilization. This fact and the lack of a complete account of the germ-cell cycle in any vertebrate led me to undertake a study of the whole history of the germ cells in both sexes in the American brook lamprey, Entosphenus wilderi (Gage), in which species there is a decided tendency toward a condition of juve- nile hermaphroditism. This seemed the more worth while because the lampreys and the hag-fishes are now elevated to a separate vertebrate class, and in this class very little work has been done on the germ-cell cycle. ‘I wish to express my sincere appreciation of the help and encouragement received from Prof. Jacob Reighard during the progress of the work. SPAWNING HABITS AND LIFE-CYCLE OF ENTOSPHENUS WILDERI A. Spawning habits Entosphenus wilderi is abundant about Ann Arbor in several streams tributary to the Huron River. All the material used in the present study was collected from Honey Creek, a small stream about four miles west of the city. On the average, the first lampreys appear on the spawning grounds at Ann Arbor about the 10th of April. Dean and Sumner (’97) report them spawning on the 16th of April in New York City. According to Gage (’93), they spawn between the Sth and 20th of May at Cayuga Lake, New York. ‘The time of spawning is undoubtedly dependent upon temperature as deter- mined by the progress of the season. The temperature of the water in Honey Creek, when the lampreys first appear in the spring, ranges from 13°C. to 143°C. The water is warmer down stream than farther up, and it is usually down stream that the lampreys are first observed. This, of course, may be due to the 4 PETER OKKELBERG fact that the temperature is higher in this part of the stream, but it may also be correlated with the fact that many larvae have been carried down stream during successive years of their life, so that the number ready to transform into adults is greater in the lower part of the stream. Evidence for this is the fact that older larvae are usually obtained from the lower part of the stream, while it is very seldom that any large or full-grown larvae are found in the upper part. I have found that the males appear on the spawning grounds before the females. Usually also only males are found in the nests early in the morning and late at night. That the males appear earlier in the season than the females was observed also by Young and Cole (’00), and, according to Surface (’97), the same is true of the lake lamprey. It is therefore necessary to collect the animals when they are spawning under optimum con- ditions in order to obtain reliable data concerning sex ratio. Dean and Sumner (’97) report more males than females in the proportion of five to one. It is easy to see how one might get such results from collections early in the season or at certain times of the day. Loman (712) found no such disproportion of sexes in the European brook lamprey and neither have I found in the American brook lamprey an excess of one sex over the other when spawning conditions were at their optimum. It is claimed by Loman (’12) that in the European brook lam- prey internal fertilization takes place. The same idea has been advanced by Ferry (’83) in regard to the marine lamprey. To test whether or not spawning females of Entosphenus wilderi contained spermatozoa, the urogenital sinus, as well as the pos- terior portion of the body cavity, of a large number of living speci- mens was examined. Inno case could any spermatozoa be found, nor did the eggs from the posterior part of the body cavity de- velop without the addition of sperm after being stripped. This shows clearly that, if internal fertilization occurs in this species, it must be only as a rare exception. The spermatozoa of the brook lamprey are motile for less than a minute after being shed into the water, but they are extremely active and are extruded simultaneously with the eggs. It is, therefore, inherently prob- GERM-CELL HISTORY IN THE BROOK LAMPREY 5) able that fertilization of nearly all the eggs is insured. In fact, all eggs collected from the stream after deposition are found to be developing. B. Infe-cycle The life-cycle of the lamprey may be divided into three main periods: the embryonic, the larval, and the adult (table 2). There is no sharp structural change between the first and second periods, but the first period may arbitrarily be considered as ending at the time of hatching and the larval period at the time of metamorphosis. The duration of the embryonic period de- pends on the temperature of the water during development. Balfour (81) estimated the length of the period to be from thirteen to twenty-one days in Petromyzon planeri. : ‘ UOT SEAT ; f£aspuocoeg gO poyaed $35G0572 6944000 ,JO poradg ygmoud e4qeq 7. ee a eg soyd -J0wey4aW “JO potasg 8694000 go potaed yqmoap AT aBg eseud stsdsufs ¢ r > SITOD wzeD (4) STB go woTzessusseg p deg5ey Jo poytied uots d wore soTenag SAT YTUTIEG ut STT?0 mae 944 go KaACYSTH ome a) Seve I@lLOd 3ST po £1048TH uo ‘48 uead , 8ou EIST g SR Fe 8TT9D wep -ita xeS JO poyaog|xes Rueaed eTVeu9y pus oTBH 30 uotT4yeztTezoeds go potdog YTAPV SS Ss ee ---- =>— eatey] — —- ----- +------ = @eTOLD PITT ! ‘ Spotded *% Ieek udstal aetok ygqanoy Jee PAzud avoX puooss aeox asatd OS A eT a vr rere faeces a = os eaeok uty STOLD 8ITT°T $aj0.i19 fig payuasaidat alD $7199 wah ayy, “jounwn ay7 fo aoa afr] ay 07 uorjnjas ut faudy) yooug ay} fo 81199 wmsab ay) fo huogsry ay? fo au2zino 6 WTd VL GERM-CELL HISTORY IN THE BROOK LAMPREY 13 of the table, in which case factors influencing the animal during development may be responsible for the resulting sex. The possibility is not excluded of sex being the result of the joint action of hereditary and external factors. Whether the germ cells of the larval lamprey eventually give rise to ova or to spermatozoa, their early history appears to be the same. Sooner or later some of the primordial germ cells transform into oocytes in practically all the larvae, irrespective of whether the larvae which bear them eventually become ma'es or females. In the present work the history of the germ cells has been studied up to a period when males and females can be distinguished by an examination of the germ glands. The scheme in table 2 presents in a graphic form the history of the germ cells in the lamprey in both males and females in rela- tion to the development of the body. As here shown, the life of the animal extends over a period of five years, and only a small part of the life-cycle is spent in the adult stage. The scheme also forms a basis for the terminology employed in the subsequent pages. B. Origin and early history of the germ cells up to the beginning of sex differentiation 1. Observations on Entosphenus wildert. During cleavage and gastrulation in the lamprey all cells are more or less laden with yolk and, if the germ cells are segregated at this time, no char- acter has been discovered by means of which they may be dis- tinguished from other cells. But after the mesoderm begins to separate from the entoderm certain large cells occur which may be identified as the primordial germ cells. a. Embryo 191 hours old (fig. 5). A camera-lucida drawing of an embryo 191 hours old (two or three days before hatching) ‘s shown in figure 5, and figure 11 shows a section along the line ab of figure 5. In this embryo the mesoderm has already sepa- rated from the entoderm cranially, but at the caudal end, i.e., in the region from which the section (fig. 11) was taken, there seems as yet to be no division line between the two layers. The mesentoderm in this region extends dorsad as two ridges, one 14 PETER OKKELBERG on either side of the nerve cord and notochord, and these are bounded externally by the ectoderm. The cells of the mesento- derm vary in size, are irregular in shape, and all of them are heavily laden with yolk. Several larger rounded cells stand out rather clearly among the more irregular surrounding cells. One of the most cranial of these is shown in figure 11. It lies directly under the ectoderm, in that part of the mesentoderm which later forms the mesoderm. Many other cells similar to this one are found farther caudad. That these are germ cells is shown by their later history, and the posterior region of the mesentoderm may therefore be considered a region for the proliferation of these cells. The shaded area in figure 5 shows in a diagrammatic way the relative position of the germ cells at this stage. b. Embryo 238 hours old (fig. 6). A sketch of an embryo 238 hours old, about one day before hatching, is shown in figure 6, and a section along the line ab is shown in figure 12 In the latter figure three spherical germ cells, structurally identical with those of the preceding stage, are seen in the lateral portions of the mesoderm which have recently separated from the ento- derm. Other germ cells may be recognized caudad of this sec- tion and also considerably farther craniad. The relative posi- tion of the cells in the whole embryo is shown by the shaded area in figure 6, and it will be seen that they extend much farther craniad than in the preceding stage. They are not arranged in regular groups in relation to the body somites, but form irregular bands, one on each side of the body. At this stage the dorsal part of the mesoderm has separated from the entoderm along its whole length, but its lateral plates are still continuous ventrally with the entoderm except along the cranial part of the germ-cell region, and craniad of it where a complete separation seems to have taken place, but even here one cannot be absolutely certain that the two germ layers are not continuous ventrally. Each lateral plate becomes thinner along the side of the body and terminates ventrally in a sharp edge, but the separation of this edge from the entoderm is not distinct. This makes it very difficult to determine whether the mesoderm extends ventrad at the expense of the entoderm or by independent growth. GERM-CELL HISTORY IN THE BROOK LAMPREY 15 In the caudal region of the body the germ cells are large and spherical and stand out clearly among the apparently smaller and more irregular-shaped cells of the yolk entoderm, as shown in figure 13, a section from the region cd of figure 6. Farther craniad the germ cells are more irregular in shape due to pres- sure from surrounding cells. No structural difference could be found between the germ cells and the large yolk-bearing cells of the entoderm, except that the former have more definite out- lines. It is probable that even this difference is the result of location rather than of any inherent difference in structure. This suggests the possibility that any of the yolk-bearing cells of the mesentoderm which are so situated that they have a chance to get into the mesoderm at the time it separates from the entoderm may become germ cells. Another probability is that the germ cells are segregated in an earlier stage and that many more are produced in the early development of the embryo than can be included in the mesoderm when it separates. In this case all of the germ cells which remain in the entoderm prob- ably degenerate in situ or are thrown off bodily into the lumen of the intestine. Later some evidence for this will be presented. c. Larva 274 hours old (fig. 7). The larva of this stage has just broken out of the egg membrane and the anterior portion of the body has straightened out, as shown in figure 7. The caudal region, however, which includes most of the yolk, still forms a right angle with the cranial region. The position of the germ cells from three different regions is shown in figures 14, 15, and 16, taken from the parts of the larvae indicated by the lines ab, cd, and ef, respectively. At this stage the mesoderm extends farther ventrad than in the preceding stage. The germ cells lie in the nephrotome region, either ventrad or latero- ventrad of the newly formed pronephric ducts. They are much more numerous than in the preceding stage, and some- times they lie so close together that in every section two or more cells are found. The absence of mitotic figures and the uniform size of the germ cells indicate that the increase in num- ber is not due to any division of the cells, but to the fact that more and more germ cells are being included in the mesoderm as JOURNAL OF MORPHOLOGY, VOL. 35, NO. 1 16 PETER OKKELBERG its separation from the entoderm extends caudad. ‘There is no indication at this or at any other stage that the germ cells are segmentally arranged. They form two bands which are sepa- rated caudally, but converge cranially. The most cranial cells, although older in the sense that they were first to be included in the mesoderm as it became separated from the entoderm, are apparently not different from the posterior cells which were included much later. d. Larva 2864 hours. This stage, shown in section in figure 18, is somewhat more advanced than the preceding. The mesoder- mic somites in the anterior region have become differentiated into a muscle plate, a dermal plate, and a sclerotome. The pronephric ducts lie in the regions laterad of the muscle plates. ~ Between them and the yolk entoderm, and sometimes indenting the latter, are the large yolk-laden germ cells. They are of the same size and structure as the cells of the preceding stage. Occasionally the cells are found in groups, but no mitosis has ever been observed in the germ cells of this period, and this makes it. probable that the cell aggregations are the result of a shght amount of migration or of several cells being separated from the entoderm at the same place. Due to a pressure from surrounding cells, many of the germ cells have lost their rounded appearance at this stage. The bands of germ cells of the two sides approach each other more closely cranially than in the preceding stage, but caudally they still he far apart. e. Larva 2993 hours (fig: 8). The larvae of this stage have increased considerably in length, but the caudal region is still loaded with yolk and remains perpendicular to the rest of the body (fig. 8). Figure 17 represents a section through the poste- rior part of the body, from the region indicated by the line ab in figure 8. In this section four germ cells are scattered along the lateral plates of the mesoderm. A section near the cranial end of the germ-cell area is shown in figure 19. This contains three germ cells which now have reached a position mediad of the pronephric ducts. Not only do the entoderm and mesoderm lie so close against each other that it is difficult in some places to see the line of separation, but the germ cells often lie in little GERM-CELL HISTORY IN THE BROOK LAMPREY 17 depressions in the entoderm formed by the pressure of the germ cells against it. The result is that in certain sections the germ cells appear to be still in the entoderm. Such sections of this late stage examined without knowledge of the previous his- tory of the cells might lead one to believe that they were migrat- ing from the entoderm into the mesoderm. But there is no reason for believing that such a belated migration takes place in the lamprey, for other sections show that practically all the meso- derm has separated from the entoderm at this stage and that all the germ cells which are destined to become functional now lie in the mesoderm. Their position is about the same as in the preceding stage. They have not yet reached the midline cranially, while caudally they are scattered along the lateral plates of the mesoderm so that at the very extreme caudal end they are still found near the midventral line of the gut entoderm. In later stages germ cells which lie in the lateral plates, far removed from their final destination, are often found in various stages of disintegration. It is also likely that many prospective germ cells never reach the mesoderm, but remain in the gut entoderm either to degenerate in situ or to be thrown off. f. Larva 320 hours old. A larva of this stage is considerably longer than that of the preceding stage. The caudal part of the body is still shghtly curved. Cranially the germ cells lie between the, dorsal aorta and the pronephric ducts; caudally they lie ven- trad or laterad of the ducts. The germ cells are in all respects similar to those of the preceding stage. g. Larva 3595 hours old (fig. 9). The body of a larva of this stage is almost straight (fig. 9). The germ cells, two of which are shown in figure 20 are nearer the middorsal line than before, have lost their rounded contours, and are flattened between the gut entoderm and the pronephric ducts, dorsal aorta, and inter- vening mesenchyme. The germ cells are still filled with yolk globules and there is no indication that mitosis is taking place. The nucleus is usually eccentric and contains two deeply staining nucleoli, besides scattered chromatin granules. Each germ cell is surrounded by a number of flattened mesoderm cells. Figure 20 shows one germ cell cut through the nucleus and another cut 18 PETER OKKELBERG along one side. The two germ-gland anlagen have not yet fused at any place along the midline, although they approach each other very closely at their cranial ends. h. Larva 3733 hours old. The germ-gland anlagen extend farther forward than in the preceding stages. Since the large, inert, yolk-bearing cells are very poorly adapted for independent migration, it is probable that their movement craniad as well as mediad is due, at least in part, to the mechanical shifting of the parts surrounding them. A coelom had formed in the cranial portion of the mesoderm and is also forming in the caudal region in front of the anal opening. In the middle portion, however, no body cavity is yet formed. The gut is loaded with yolk and is surrounded by mesoderm. In this mesoderm, on the doral side of the intestine, the yolk-laden germ cells occur, sometimes singly and sometimes in groups. Cranially, the two lateral germ-gland anlagen are well defined and practically come together. The cells are greatly flattened dorsoventrally by the pressure of the surrounding tissues. Caudally the germ cells are scattered along the whole lateral plate. In a cross-section from the caudal region slightly craniad of the anal opening, a large germ cell (fig. 21) was found lying in what may be considered the ventral mesentery. It is highly probable that a germ cell so situated will never become functional. 7. Larva 42934 hours old. The posterior cardinal veins have appeared at this stage and the germ cells Jie ventrad of these cranially (fig. 22). As the coelomic cavity is being formed by a splitting of the mesoderm, the germ cells become included in the somatic portion (fig. 23). j. Larva 4784 hours old (4:3 mm. long). At this stage the germ-gland anlagen have fused cranially (fig. 24); caudally they still remain apart. The cardinal veins have increased in size and now lie dorsomediad of the pronephric ducts. Cranially the mesenchymal tissue has increased greatly in amount in the region in which the germ cells are found. It fills a considerable space between the germ cells, the dorsal aorta, and the cardinal veins. The germ cells are still flattened against the gut ven- trally. They still retain their embryonic structure and are not dividing. GERM-CELL HISTORY IN THE BROOK LAMPREY 19 k. Larva 53883 hours (533 mm. long) (fig. 26). In the caudal region of the body the germ cells still retain their embryonic form. They are apparently not yet able to reach their final median posi- tion on account of the large amount of yolk in the entoderm of this region. Cranially, on the other hand, the yolk in the ento- derm is being absorbed so that more space is left for the germ cells, and in consequence they shift their position toward the midline. As a result, the two bands of germ cells are now arranged in the form of a V with the apex pointing craniad. With the release of pressure and with the assumption of a median position, the anterior cells begin to show signs of activity. The yolk globules in many cells have lost their sharp contours and often appear fragmented (fig. 25). Sometimes they are absent from certain parts of the cells. The cytoplasm, hitherto clear, now has a granular appearance. The chromatin material in the nucleus now stains more deeply, and often chromatin-like gran- ules are found in the cytoplasm surrounding the nucleus. It appears probable that there is at this time an active inter- exchange of material between the nucleus and the cytoplasm. No mitoses were observed, however, for a long time subsequent to this stage. © l. Larva 647 hours. At this stage the coelomic cavities have formed by a splitting of the lateral plates throughout their length, and the cavities of the two sides have fused and nearly the whole of the dorsal mesentery has disappeared. The germ cells are included in the somatic layer of the mesoderm. Most of them now lie along the dorsal midline directly below the dorsal aorta, but are spread out over a considerable area on each side of it. All have disappeared from the caudal region. No germinal fold is yet present. Some of the germ cells have lost the greater part of their yolk. None of the cells were found in mitosis. m. Larva 9023 hours. A few germ cells along the posterior portion of the body cavity still retain some yolk, but the great majority of them are now free from it. As compared with the mesodermal cells of the same region, they may be described as large spherical cells with large spherical nuclei each with two large nucleoli. No cells were found in mitosis. Sometimes 20 PETER OKKELBERG two or more cells are so grouped as to suggest that they are derived from one cell by division; but since no mitoses are ob- served, the grouping is probably the result of a migration of the cells. The larvae of this stage have begun to feed. An exami- nation of larvae between this and the former stageshows that they begin to feed when they are about 7 mm. long, although a great deal of yolk is still present in the intestinal wall. n. Larva 10 mm. long (June 22). This larva was obtained from the creek on June 22nd and is in the neighborhood of seventy days old. ‘The yolk is now all absorbed from the intes- tinal wall and the lumen of the digestive tract is full of diatoms and other organisms upon which the larvae feed. The germ- gland anlagen are in the middle two-thirds of the coelom, but are absent from its cranial and caudal parts. In later stages, when the cells begin to increase in number by division, their range is extended both craniad and caudad. The germ cells are irregularly distributed along the anlagen with no indica- tion of a segmental arrangement. They lie in the mesenchyme on the ventral side of the dorsal aorta and close against the peritoneum, which consists of very flat epithelial cells (fig. 29). Some of the germ cells may project slightly into the coelom, but these projecting cells do not yet form a continuous germ fold. Although the germ cells may lie against the peritoneum, they never form a part of it. They may be distinguished from the epithelial cells and other cells of the same region by their larger size and spherical shape; by their large spherical nuclei, each containing two large nucleoli, and by their clear transparent cytoplasm (fig. 30). Each is surrounded by flat epithelial cells which are similar to those forming the peritoneum. ~The germ cells in this stage are absolutely distinct from the cells of the soma, as they appear to be from the time when they are first recognized as germ cells. They have lost all their yolk, but no signs of mitosis could be found. At this stage the lumen of the intestine is very much enlarged and many of the cells from the walls of the intestine have been set free into the intestinal cavity (fig. 31). This is the case also in much earlier stages (larvae about 7 mm. long), and suggests GERM-CELL HISTORY IN THE BROOK LAMPREY Bik that the extruded cells are germ cells which have failed to reach the germ-gland anlage. o. Larva 20 mm. long. This larva is about four months old. The germ fold has now formed (fig. 27) and extends along the dor- sal wall of the coelom as a low longitudinal ridge. Some of the germ cells have migrated into the fold, but others are still in the mesenchyme above it. Their position is such that they are prac- tically surrounded by blood-vessels—the posterior cardinal veins laterally and the dorsal aorta above. Besides these vessels, a large number of smaller vessels permeate the tissues around the germ cells. The germ cells occur in groups and in some places long distances intervene between them so that the various groups do not form a continuous band. The germ cells are more numerous than in the preceding stage, and in ome cases two are found which apparently are surrounded by a single follicular membrane—an indication of recent division. In most of the cells at this stage, a distinct attraction sphere occurs. It con- sists of a mass of closely set granules and lies against one side of the nucleus. It may cover as much as one-third of the cireum- ference of the nucleus in each section through the middle of the cell (fig. 28). No distinct centrosome could be found. Besides the attraction sphere, there is in the cytoplasm a spindle-shaped body, the ‘vitelline body’ of King (’08), which is much smaller and is made up of coarser granules than the attraction sphere. In longitudinal section it appears oval, in cross-section, round. Judging from sections through various planes, it is shaped like a spindle which tapers abruptly at both ends. It may occur almost anywhere in the cytoplasm, sometimes near the nucleus but sometimes close under the cell membrane. The origin of this body could not be ascertained. The presence of a centro- sphere indicates the beginning of mitotic activity. The cells may now be considered as having passed out of the primary period of rest and entered the period of secondary division. There is yet no indication of sexual differentiation and, in the absence of any characters which distinguish the secondary sper- matogonia from the secondary oogonia, the germ cells may be regarded as still indifferent as to sex (table 2). iM), PETER OKKELBERG The germ cells lie against the peritoneum which covers the gland (fig. 30), but they are always separated from the coelom by peritoneal cells and they have never been found to form a part of the peritoneal epithelium. Epithelial cells may be seen, still in part included in the peritoneum, but with processes extending to the germ cells and forming a part of their follicles. These cells occur in all stages of detachment from the peritoneal epithelium and in all stages of inclusion in the follicular mem- branes of the germ cells. There is no evidence at this stage that the follicular cells are derived from any other source. Both in the peritoneum and in the follicles these cells are distinguishable from the germ cells and mesenchyme cells by their ovoid nuclei, each with one large plasmosome and several smaller chromatin nucleoli, and by their flattened form and indefinite contours. The mesenchyme cells are recognizable by their nearly spherical nuclei, and the germ cells by their large size and spherical nuclei, each with two large plasmosomes. It seems clear from their early history, from the fact that they are at no time seen to be included in the peritoneal epithelium, and from their distinguish- ing structural characters, that the germ cells are not derivatives of the peritoneal epithelium. It seems equally clear that the follicle cells are derived from this source. 2. Historical and critical. a. Invertebrates. A study of the early history of the germ cells in various species of invertebrates has disclosed the fact that they often are segregated during early cleavage. These primordial germ cells are at first distinguished either by the behavior of their chromosomes or by the presence of certain cytoplasmic inclusions. In vertebrates the primordial germ cells are usually not recognizable until the three germ layers are formed, although most investigators of the subject (Beard, Allen, Dodds, Nussbaum, King, Witschi, Rubaschkin, Swift, Tschaschkin, and others) believe they must have been segre- gated at a much earlier stage. The stanchest adherents of the theory of early segregation (Beard, Allen, Rubaschkin, Witschi Swift, and others) hold that all the definitive germ cells are derived from the primordial germ cells, a conclusion that the theory of the continuity of germ plasm naturally demands. GERM-CELL HISTORY IN THE BROOK LAMPREY 23 Others (Abramowicz, Bouin, Kuschakewitsch, Dustin, Firket), while admitting the presence of the primordial germ cells in the early embryo and the possibility that they give rise to definitive reproductive products, still think it probable, and even sup- ported by very strong evidence in some cases, that many of the definitive germ cells are derived from other elements which, strictly speaking, have formed a part of the soma. Opinions vary as to whether these other cells should be considered true somatic elements or simply another type of undifferentiated cells. Child (06) is convinced that in the cestode, Moniezia expansa, the germ cells develop from cells of the parenchymal syncytium which must be regarded as differentiated tissue cells. In his development of the ‘theory of dedifferentiation’ (15) he makes the following statement: In the tapeworm Moniezia, for example, the sex cells arise from the parenchyma, and apparently any parenchymal cells which he within the region involved in the production of sex cells may undergo dediffer- entiation and take part in the process. Even the large muscle cells may give rise to testes. . . . . In such cases the muscle fiber undergoes degeneration, the vacuoles disappear, and the nucleus begins to divide, apparently at first amitotically (pp. 331-832). C. W. Hargitt (06) thinks that the germ cells in Clava leptostyla arise in the entoderm, and that it is unlikely, though possible, that these cells may be undifferentiated. In Campanularia flexuosa, George T. Hargitt (13) has found that the egg cell arises in the entoderm by the transformation of single epithelial cells, or from the basal half of divided cells. He conc udes: “Therefore the egg cells have come from differentiated body cells (so-called) and there is no differentiation of the germ plasm in the sense that the germ cells are early differentiated and set aside and do not participate in the body functions” (p. 111). Max Jérgensen (’10) comes to the same conclusion for Sycon. He says: ‘‘Indessen zeigen mir meine Priparate dass auch eine Entstehung von Oogonien aus Mesodermzellen denkbar und morphologisch nachweisbar ist” (p. 169). b. Vertebrates. The earliest theory of the origin of germ cells in vertebrates is the ‘germinal epithelium theory,’ advanced by 24 PETER OKKELBERG Waldeyer in 1870. He found some large spherical cells in the coelomic epithelium on each side of the dorsal mesentery in the early chick embryo, and supposing that they were young stages of eggs he called them ‘ Ureier’; the epithelium in which they were found he called ‘Keimepithel’ (germinal epithelium). In 1875 Semper found that both ova and spermatozoa were derived from. the so-called Ureier. Waldeyer and his followers believed that these cells were derived directly from the cells of the germinal epithelium, and this idea is held by a few investigators at the present time. Since the development of the theory of ‘early segregation’ by Nussbaum (’80) many investigators have worked on the origin of the germ cells in vertebrates. It has been found in most cases in which the early history of the cells has been traced that they do not originate in the coelomic epithelium, among the cells of which they are later found, but that they attain this position after a migration from other parts of the embryo (Woods, Allen, Dodds, King, Witschi, and others). These same investigators have found that the germ cells in very early stages are usually located in the entoderm, from which they migrate to their definitive position in the coelomic epithelium. When first found, they are large yolk-bearing cells of the ento- derm and distinguished from the entoderm cells principally by their location. — From numerous investigations on the subject there seems to be no doubt about the existence of the so-called primordial germ cells which are segregated very early in the development of the embryo, but whether or not all, some, or any of the definitive germ cells are derived from these is a question about which there is very little agreement. Rubaschkin (’09, ’12) and others hold that the definitive reproductive cells in mammals are derived exclusively from primordial germ cells. Firket (’14) thinks it is possible that a few of the oogonia may be derived from the pri- mordial germ cells, but that most of them are derived from certain cells in the germinal epithelium, which he calls ‘gono- cytes secondaire.’ Kuschakewitsch (’10) believes that in Rana esculenta the oogonia are derived from cells in the germinal GERM-CELL HISTORY IN THE BROOK LAMPREY 25 epithelium which are descendants of the primordial germ cells; while the spermatogonia are developed from ‘Paragonien,’ or secondary germ cells, which take their origin in the axial mesen- chyme. Von Winiwarter and Sainmont (’09) regard the pri- mordial germ cells in mammals as only temporary structures, which later degenerate. The same conclusion has’ been reached by Kingery (’17). Von Berenberg-Gossler (714), from his work on Lacerta agilis, comes to the conclusion that the migration of the so-called primordial germ cells from the ento- derm is nothing but, ‘‘eine spite, sich nach lingere Zeit hinzie- hende Mesodermbildung aus dem Entoderm.”’ He thinks that these cells as well as other mesoderm cells, such as those of the coelomic epithelium, may give rise to the stem cells of the ova and spermatozoa. According to Gatenby (16), there is in the frog (Rana temporaria) and other amphibians an annual trans- formation of peritoneal cells into germ cells, so that in these forms there can be no talk of a continuity of the definitive germ cells and the primordial germ cells. Some investigators believe that they have found evidence that the primordial germ cells are segmentally arranged, being derived from segmental portions of the mesoderm (Riickert, Van Wijhe, Dustin, and others). This has been termed the ‘gonotome theory.’ It is an attempt to homologize the con- dition found in vertebrates with that found im Amphioxus in which the gonads are segmentally arranged from the begin- ning of their development. According to this theory, the germ cells are derived from mesodermal cells. Only a few investigators have followed the later history of the primordial germ cells and found that they actually give rise to definitive germ cells. Among these are Witschi (14) for the frog, King (’08) for the toad, and Swift (’14, ’16) for the chick. Many supporters of the theory of early segregation have studied only the early embryonic stages and have assumed that definitive germ cells originate from no other source than the primordial germ cells. From the above it is clear that there exists at the presen time a great diversity of opinion concerning the origin of the 26 PETER OKKELBERG TABLE 3 Outline of theories concerning the origin of germ cells in vertebrates. The various stages in the development of the individual are represented by five vertical columns. The dotted lines indicate the germ paths as conceived by the various theories. The different germ layers are represented by horizontal lines. The proximity of the germ paths to each of these lines indicates where the germ cells are supposed to take their origin and where they may be found during the different stages of development. In the sixth vertical column a brief summary of each theory ts given fish weil crt Stern, [ete | sane! aaitiel eal Blastoderm lcastruis | Aree Germ Layers bryojddult coo Dieter’ Ectoderm ee ey Theory of early segregation, Rubaschkin "12 and others, ( Mes-ontoderm poe Ectoderm fasoderm bea cl | --- Ectoderm eee Germinal epitheli- um theory and Gono- tome theory. Waldeyer "70, Rttckert '88, and others, PALA Ug Oogonia derived from primordial germ cells,sperm- atogonia from mesoderm cells (Paragonia), Kuschakewitsch "10. . La Primordial germ cells degenerate; secondary germ calls derived from mesoderm, Von Winiwarter and Sainmont '09, Primordial germ cells may form definitive germ cells but these may also come from mesoderm cells, Von Berenberg- Gossler '14, } Ectoderm Primordial germ cells may give rise to a few oogonia but most of the definitive germ cells are derived from mesoderm cells, Firket '14, Entoderm GERM-CELL HISTORY IN THE BROOK LAMPREY 27 reproductive cells in animals. ‘Table 3 presents a summary of the various theories relating to the origin of the germ cells in vertebrates. For one who adheres closely to the germ-plasm theory it is hard to conceive of the germ cells as coming from any other cells than the early segregated embryonic cells which have had no part in the building up of the body. Some who believe in the theory of early segregation maintain that, even when the germ cells appear to arise from so-called somatic elements, these are in the strict sense not somatic cells, but cells that have maintained their embryonic structure and have not specialized in any given direction. There is, of course, no direct evidence for this. In many forms, it is true, the germ cells seem to be segregated very late in the life of the individual. This is apparently true of annelids and flatworms among animals, and it seems to be true of all plants. There are no investigators of germ cells in verte- brates who maintain that they come from highly differentiated somatic cells, such as muscle cells, as observed by Child (’06) in Moniezia. The various theories concerning the origin of germ cells in vertebrates have now been stated. Below is a partial list of the most important contributions on the subject, each followed by a brief statement of the conclusions which the various investigators have reached. The references have been arranged chronologi- cally for the various groups of vertebrates. 28 PETER OKKELBERG AUTHOR YEAR | SPECIES RESULTS AND CONCLUSIONS Cyclostomes Goette 1890 | Petromyzon Germ cells derived from the meso- fluviatilis derm Wheeler 1899 | Petromyzon Germ cells derived from blastoderm planeri cells Beard 1902 | Petromyzon Germ cells are early segmentation planeri cells. Their number is 25-1 Elasmobranchs Semper 1875 | Plagiostomes | Germ cells are derived from the coe- lomic epithelium Balfour 1876 | Scyllium, Germ cells are probably derived from 1877 | Pristiurus the mesoderm. They may have been introduced from elsewhere Rickert 1888 | Pristiurus Germ cells are derived from segmen- tal mesoderm cells. Gonotome theory Van Wijhe 1889 | Seyllium, Germ cells are derived from segmen- Pristiurus tal mesoderm cells. Gonotome theory Beard 1900 | Raja batis, Germ cells are derived from early 1902 | Pristiurus segmentation cells Woods 1902 | Squalus acan- | Early segregated germ cells are first thias found in the entoderm Ganoids Allen 1909 | Amia Germ cells are segregated early and Lepidosteus are first found in the entoderm Teleosts Nussbaum 1880 | Trout Germ cells are segregated early and are first seen in the region of the germ gland, but they are not de- rived from the mesoderm MacLeod 1881 | Hippocampus, | Germ cells are derived from the ger- Belone minal epithelium Hoffmann 1886 | Salmon Germ cells are derived from peri- toneal cells Eigenmann 1891 | Micrometrus Germ cells are segmentation cells 1896 aggregatus from about the fifth generation Bohi 1904 | Trout, salmon | Germ cells are derived from cells of the coelomic epithelium GERM-CELL HISTORY IN THE BROOK LAMPREY 29 AUTHOR YEAR SPECIES RESULTS AND CONCLUSIONS T eleosts—Continued Federow 1907 | Salmo fario Germ cells are first found in the somatopleure and splanchnopleure Dodds 1910 | Lophius pisca-| Germ cells are first found in the pri- torius mary entoderm. They are early segmentation cells Bachmann 1914 | Amiurus nebu- | Germ cells are first found in the lat- losus eral plate of the mesoderm. They are early segregated cells Urodeles Dustin 1907 | Triton alpes- | Germ cells are derived from meso- tris derm cells (gonotome) Spehl and Polus 1912 | Axolotl The gérm cells are derived from mesoderm cells Schapitz 1912 | Amblystoma The germ cells are derived from mesoderm cells (gonotome) Abramowicz 1913 | Triton Primary germ cells are derived from the entoderm (early segregation cells). Secondary germ cells are derived from the mesoderm Anura Nussbaum 1880 | Rana fusca Germ cells are first found in the mesoderm, but are early segre- gated cells Bouin 1901 | Rana tempora-| Germ cells are derived from early ria segregated cells, and from peri- toneal and mesenchyme cells Allen 1907 | Rana pipiens | Germ cells are derived from early segregated cells. They are first found in the entoderm Dustin 1907. | Rana fusca Germ cells are derived from the lat- eral plate of the mesoderm (gono- tome) and from peritoneal cells Dustin 1907 | Bufo vulgaris | Germ cells are derived from the lat- eral plates of the mesoderm (gono- tome) and from peritoneal cells King 1908 | Bufo lentigino- | Germ cells are early segregated cells sus and are first found in the entoderm Kuschakewitsch. | 1910 | Rana esculenta | Primary germ cells are derived from early segregated cells. Secondary germ cells are derived from the peritoneal epithelium and axial mesenchyme 30 PETER OKKELBERG AUTHOR YEAR SPECIES RESULTS AND CONCLUSIONS Anura—Continued Champy 1913 | Rana tempora- | Germ cells are derived from segmen- ria tal mesoderm cells (gonotome) Witschi 1914 | Rana tempora- | Germ cells are derived from early ria segregated cells and are first found in the entoderm Gatenby 1916 | Rana tempora-|Germ cells originate periodically ria from peritoneal cells in adult frogs Reptiles Allen 1906 | Chrysemys Germ cells are first found in the ento- 1907 marginata derm and are derived from early 1911 segregated cells Jarvis 1908 | Phrynosoma Germ cells are first found in the ento- cornutum derm and are derived from early segregated cells Dustin 1910 | Chrysemys Primitive germ cells are derived from . entoderm cells. Secondary germ cells come from peritoneal cells Von Berenberg- 1914 | Lacerta agilis | So-called primordial germ cells are Gossler formed inthe entoderm. ‘They are probably not germ cells, but give rise to mesoderm cells. Germ cells are derived from mesoderm Aves Waldeyer 1870 | Chick Ova are formed from cells of the ger- minalepithelium. Spermatogonia come from cells of the wolffian duct epithelium Hoffmann 1892 | Twelve species | Germ cells are early segregation of birds cells Nussbaum 1901 | Chick Germ cells are first found in the splanchnopleure, but they are early segregated cells Rubaschkin 1907 | Chick, duck Germ cells are early segregation cells first found in the splanchno- pleure Tschaschin 1910 | Chick Germ cells are first found in the splanchnopleure. They are early segregated cells Von Berenberg- 1912 | Chick The so-called germ cells first found Gossler in the splanchnopleure may not be germ cells GERM-CELL HISTORY IN THE BROOK LAMPREY ot AUTHOR YEAR SPECIES RESULTS AND CONCLUSIONS Aves-—Continued Firket 1914 | Chick Primordial germ cells may give rise to definitive germ cells, but most of these are derived from the ger- minal epithelium Swift 1914 | Chick Germ cells are derived from cells in ; 1916 the germ wall entoderm. They : . are early segregated cells Mammals Allen 1904 | Pig, rabbit All functional germ cells are derived from the peritoneum Sainmont 1986 | Cat Primitive ova are present, but are not functional. Definitive germ cells are derived from epithelial cells Winiwarter and 1909 | Cat Primitive ova are present, but are Sainmont not functional. Definitive germ cells are derived from the germinal epithelium Rubasechkin 1908 | Cat, rabbit Germ cells are derived from early 1909 | Mole, porpoise segregated cells and are first found 1912 | Guinea pig in the entoderm. There is no sec- ondary origin of germ cells c. Review of work on the early history of the germ cells in lampreys. W. Miiller (’75) described the germ glands of young lamprey larvae as median, unpaired thickenings of the perito- neum situated between the bases of the mesonephric bodies and extending along the whole length of the body cavity. At this stage groups of germ cells were found, but sex could not be distinguished. Goette (90) found the reproductive cells in larvae of a much earlier stage, corresponding approximately to that represented by my figure 7. He observed that, while most of the cells in the mesodermal plates soon lost their yolk and began to divide, some of the cells retained their yolk and remained undivided. These cells were found in the mesodermal plates on both sides along their thickened median portions directly under or outside of the pronephric ducts and sometimes against the yolk entoderm, JOURNAL OF MORPHOLOGY, VOL. 35, NO. 1 32 PETER OKKELBERG so that it appeared as though they might belong to it. He says, however, ‘‘Eine genaue Untersuchung hat mich aber tiberzeugt dass es urspriingliche Mesoderm-elemente und nicht etwa vom Darmblatt her eingewanderte Zellen sind” (p. 53). Wheeler (’99) has given an excellent account of the early devel- opment of the germ cells in the lamprey (Petromyzon planeri). He recognized the germ cells in the posterior region of embryos as early as my figure 5, the stage in which they were first observed by me in Entosphenus wilderi. Wheeler found: “Just laterad to the myotomes a few very large rounded masses of yolk.” He described each mass as containing a nucleus and ‘“‘more or less distinctly marked off from the adjacent entoderm elements.” He says further, ‘‘These large masses are the primitive repro- ductive or sex cells. They can hardly be assigned to the meso- derm because their appearance and position are those of entoderm cells in this stage. Still they lie in a portion of the entoderm which becomes mesoderm with the more lateral extension of the latter layer.”’ Beard (’02), in an attempt to work out a numerical law for the primordial germ cells in animals, says that the number of cells should in each case be 2-1. His theory is that the blasto- derm in animals corresponds to the sporophyte in plants, and to it he applies the term ‘phorozoon.’ After a time one of its cells divides a definite number of times and forms the primordial germ cells. The number of divisions varies according to the species. One of these primordial germ cells is saerificed to form the embryo so that the actual number of germ cells remain- ing is in each case 22-1. In the case of the lamprey (Petromy- zon planeri) Beard finds that 2" = 32 and that therefore in this species the number of primordial germ cells is thirty-one. Interesting in connection with the description of the early history of the germ cells in the lamprey is an observation made by Kupffer (’90). In the early gastrula of Petromyzon planeri, he found, between the ectoderm and the entoderm in the region of the blastopore, certain cells which he called ‘teloblast cells.’ They were easily distinguished from the yolk cells adjoining them, but their origin was not observed. Kupffer thinks that GERM-CELL HISTORY IN THE BROOK LAMPREY 30 the mesoderm in this region develops at the expense of these cells. He says: Wenn aber das dorsale Mesoderm entstanden ist und bis zum Teloblast reicht, tritt es in dieselbe einige Verbindung damit wie der Neuralstrang und die Chorda, und nachdem die Segmentierung des Mesoderm bis zum Schwanzende fortgeschritten ist, ergiinzt sich der jewellig hinterste Abschnitt des Mesoderm durch Zellen die aus dem Teloblast stammen. Hatta (92, ’07) describes certain cells as budding off between the ectoderm and the entoderm in the region of the blastopore. These cells he calls the ‘peristomal mesoblast.’ He could not find any cells that corresponded to Kupffer’s teloblast cells. It is possible that Kupffer’s teloblast cells and Hatta’s peristomal mesoblast cells are identical and that they correspond to the large yolk-laden cells which later become included in the meso- derm and form the germ cells. 3. Discussion. a. Karly segregation. We have seen that the germ cells in Entosphenus wilderi may be traced to the Jarge yolk- bearing cells which at first are located in the mesentoderm. This is in agreement with the observations of Wheeler. The history of these cells, previous to their inclusion in the mesoderm is not known. They apparently lie among similar yolk-bearing cells belonging to the entoderm, and it is a question whether or not they are essentially different from these. The germ-plasm theory, as expressed by Weismann, demands a segregation of the germ cells at a very early stage, or their origin, at ‘east, from cells that have never taken any part in the formation of body tissues. In one sense all the yolk-bearing cells of the entoderm may be considered as undifferentiated cells, but only some of these cells which are included in the mesoderm become germ cells. Most of the mesoderm cells, however soon begin to divide, become smaller, lose their yolk, and form various tissues, while the large cells that become germ cells do not change in the least for a very long time. This indicates that they are endowed with certain qualities which distinguish them from the cells that become somatic. 34 PETER OKKELBERG My observations and those of Wheeler (’99) show that the germ cells appear first in the posterior region of the body, probably in a small area around the blastopore. None have been seen to separate from the entoderm very far craniad of this region, so that those found later in the region further craniad gain this position by some form of migration. In Entosphenus wilderi the peristomal and the paraxial regions of the mesoderm are con- tinuous and do not differ in structure or in origin as maintained by Hatta (92, ’07). The only distinction found between them is that the mesoblast of the paraxial region is delaminated earlier than that of the peristomal region, but only the peristomal mesoderm carries the germ cells in early stages. In Entosphenus wilderi no cells corresponding to Kupffer’s teloblast cells were found in this region. ‘These may have been either germ cells or, as he supposed, mesoderm cells. The condition in the lamprey is in favor of the theory of early segregation. The large yolk-laden cells that are at first found among the other yolk-bearing cells of the entoderm in the caudal region of the body become included in the peristomal mesoderm when it separates from the entoderm by a process of delamination. These cells retain their embryonic structure for a long time after the other elements of the mesoderm have become differentiated. Even after all their yolk has been used up, they remain as large conspicuous cells among the smaller somatic cells of the germ gland. Up to this point the history of these cells has been traced and the later history to a time when they begin to divide. Later it will be seen that there is in Ento- sphenus wilderi no evidence that any other cells take part in the formation of definitive germ cells. What evidence there is in other forms will now be considered. Two lines of evidence have been advanced for the secondary origin of germ cells from mesodermal elements. ‘The first is that transitional stages have been found between mesodermal cells and true germ cells (Abramowicz, Dustin, Firket, Gatenby, and others), but in most cases the figures which purport to repre- sent this transition are not convincing. The second line of evidence is that advanced by Kuschakewitsch (10). He found GERM-CELL HISTORY IN THE BROOK LAMPREY 35 that in the frog (Rana esculenta) all the embryos produced by eges in which fertilization had been delayed for a certain number of hours were males. He believes that normally oogonia are derived from primordial germ cells which are situated in the germinal epithelium, while the spermatogonia are derived from the axial mesenchyme. When fertilization is delayed no pri- mordial germ cells are produced and all the definitive germ cells come from mesodermal cells which he calls ‘Paragonien.’ These give rise only to spermatogonia, and the larvae are therefore males. Witschi (’14), who more recently has worked on the ‘development of the germ cells in Rana temporaria, finds that in this form there is no secondary origin of germ cells. He says: So scheinen alle Tatsachen dafiir zu sprechen, dass von ihrem friih- esten Erscheinen an, die Keimzellen als Gebilde spezifischer Natur zu betrachten sind, welche, wenigstens unter Bedingungen die von nor- malen nicht sehr abweichen, weder sich in somatische Elemente umwandeln, noch aus solehen durch Unwandlung entstehen kénnen. Among others who have followed the history of the germ cells and have found no evidence of secondary origin, may be men- tioned King (’08) for Bufo and Swift (’14, ’16) for the chick. b. Method of migration of the germ cells. Three opinions have been advanced concerning the method by which germ cells reach the germ-gland anlage in vertebrates: 1. Most investiga- tors are inclined to the belief that, by some sort of ameboid move- ment, there is an active migration of the germ cells from the ento- derm to the splanchnic mesoderm, and then through the mesentery into the germ-gland anlage. That such migration exists seems certain in forms in which the germ cells are so late in arriving at their final destination that they lose their yolk before leaving - the entoderm. 2. A second theory is, that in some of the lower forms the migration of the germ cells may be accounted for partly by a shift in the position of the somatic tissues around them. In this case the germ cells are considered passive elements which take little or no active part in the migration. It seem likely that in some forms migration is partly active and partly passive. This is probably true in the lamprey. No 36 PETER OKKELBERG ~ ameboid forms have been observed in the yolk-laden germ cells of the early embryo, and it is not likely that they are capable of independent movement, but it is probable that they are trans- ported to the germ-gland region by a shifting of the tissues dur- ing somatic differentiation. In later stages, after all the yolk has disappeared, the germ cells extend farther and farther craniad. In sections of these stages the individual germ cells always have a spherical shape and no pseudopodial processes have been seen to indicate that they migrate independently; yet it seems prob- able that in life there may be a slight amount of migration in this way. Individual germ cells and cysts are usually not in contact, so that it is a little difficult to conceive of these shifting to a more cranial position as a result of the pressure of the cells or cysts against one another as they increase in number. Since the follicle cells show ameboid processes, it may be that the germ cells are carried along by a movement of these. Some of the germ cells in the lamprey, as well as in other forms, never reach the germ gland, but remain in the entoderm or some other part of the body. Some of these cells may divide and form cell nests in other organs. Such cell nests have been found in the lamprey, both in the fat body and in the median and lateral portions of the mesonephros (figs. 32, 33). The fate of these cell nests is not known, but they probably degenerate. 3. A third method of migration appears to have been observed by Swift (14) in the chick embryo. He says that the large yolk-laden germ cells, which are first found in the germ-wall entoderm, are taken up by the blood-vessels and carried by the blood-stream to the germ-gland region. In fact, the germ cells may be carried to any part of the body, but it is only in the germ-gland region that they migrate out of the blood-vessels. Swift (16) has followed the later history of these cells and has come to the conclusion that they give rise to the definitive germ cells. Von Berenberg-Gossler (’14), who worked on the early germ cells in duck embryos, also found cells similar to those described by Swift, but he expresses a doubt as to whether or not they are germ cells. He says: ‘‘ Alles in allem bin ich der Ansicht, dass das ganze Verhalten dieser Zellen in héherem Grade davor warnt sie fiir Keimbahnzellen zu halten”’ (p. 261). GERM-CELL HISTORY IN THE BROOK LAMPREY al A study of the methods by means of which germ cells migrate from the entoderm into the mesoderm in the various groups of vertebrates, and of the time at which this migration takes place, shows that these groups may be arranged in an interesting phylogenetic series as represented in the following diagrams (text figures A, B, C, D). In the lamprey (text figure A) the germ cells are shown as being included in the mesoderm at the time when it becomes separated — from the entoderm. In Triton (text figure B), according to Abramowicz (713), the germ cells migrate into the mesoderm be- fore the dorsal mesentery is formed, but much later than in the case of the lamprey. In the frog, according to Allen and others, the germ cells are separated from the yolk entoderm at the time the dorsal mesentery is formed, as shown in text figure C. In each of the first two forms, therefore, the original germ-cell anlage is double, while in the latter it appears to be unpaired at first and to lie along the middorsal line above the gut entoderm. In the lamprey, although the original anlage is paired, the germ gland later becomes single by a fusion of the two parts on the ventral side of the dorsal aorta. In Triton the paired anlagen fuse, but later they become paired again, and in the frog the original unpaired anlage becomes paired. In reptiles and mam- mals the germ cells usually separate from the entoderm much later than in the other three forms, often after they have lost all their yolk. Then migration follows: first into the splanchnic mesoderm, then through the dorsal mesentery, and from there to the germ-gland anlage on each side (text figure D). c. Relation of germ cells to body cells. In the later embryonic stages of many forms, when the germ cells have reached their final destination in the region of the germ gland, they lie among the peritoneal cells covering the definitive germ-gland anlage and apparently form a part of the peritoneal membrane. This fact gave rise to the idea that the cells were derived from the peri- toneal cells, so that this portion of the peritoneum became known as the germinal epithelium. In the lamprey there is, strictly speaking, no germinal epithelium since the germ cells never form a part of the peritoneal membrane, but are independent elements PETER OKKELBERG 38 ’ ((9X0} 008 UOIYdIJoSsep 104) ‘speuIUET puv sopiydar ‘gq ‘sory ‘9 fu0qny Sg ‘Kordurey ‘Vy ‘“ulloposour oy} ur UoryISsod r10y4 SNOMVA JO S[[od ULES vy} poyyour ywYyM Aq SUIMOYS Sor10S dIyoUBS0;Ayd y q pu UIv 4B BOS SOAIGUID 0}B1G9}.10A ‘Vy somnsy 4x07, + Prog Ise ae iene GERM-CELL HISTORY IN THE BROOK LAMPREY 39 situated in the mesenchyme dorsal to the peritoneum. Some- times the germ cells may lie so pressed against the peritoneum that its cells become greatly flattened, but the continuity of the peritoneal epithelium is apparently never broken. I have found no evidence that any germ cells are formed out of peritoneal cells at any stage. d. Number of germ cells. No evidence was found that divi- sion takes place in the germ cells of the lamprey before the larva is about 20 mm. long. The germ cells are few in number, and, as the young larva continues to grow, they become more and more scattered along the whole length of the body cavity, so that in following a series of sections they are often found far apart. In a larva about 8 mm. long, in which the germ cells stood out with great clearness, thirty-six were counted. These were dis- tributed through 153 10-z sections or along 1.53 mm. of the body length, as shown in figure 10. The distribution of the cells by sections was as follows (table, p. 40): It will be seen that at this time no germ cells are found in the caudal region of the body for about 0.58 mm. in front of the anal opening, and cranially they do not extend any farther than to about the middle of the coelomic cavity. The number of germ cells found in this larva comes very near that found by Beard (’02) in larvae of Petromyzon planeri, but it can hardly be cited as supporting his theory, especially since it is known that many of the germ cells never reach the germ-gland region. It is highly doubtful if the number of germ cells segregated in the development of the embryo is constant, and it certainly seems probable that the number of germ cells that reach the germ gland must vary greatly in different individuals. It would be futile, therefore, to form any conclusions concerning the number of primordial germ cells from a count of those that reach the germ gland. 4. Conclusions and summary. Evidence presented supports the theory that germ cells are segregated very early in the develop- ment of Entosphenus wilderi. These primordial germ cells are. first found in the yolk entoderm and become included in the mesoderm when it separates from the entoderm. The germ cells 40 PETER OKKELBERG occur at first only in the posterior region of the body, but they gain a more cranial position with the growth and development of the larvae. The movement of the germ cells craniad is probably SECTION CRANIAD OF ANUS NUMBER OF GERM CELLS 58 61 68 76 80 — no eo) a Ne Ce Co Ce oo OO ia ° (=~ ise) ay (Je) [op a passive one in early stages and results from a shifting of the tissues during development. In later stages, after the cells have lost their yolk, migration is probably due partly to an active GERM-CELL HISTORY IN THE BROOK LAMPREY 41 movement of the cells. The original double anlage of the germ gland is reduced to a single median gland in later stages, due to a fusion of the two parts along the midline. This fusion begins at the cranial end of the gland. The germ cells which do not reach the germ gland probably degenerate. Some of them form extraregional cell nests, the fate of which is not known. ‘There is no division of the germ cells before the larvae have reached a length of about 2 cm. Long before this all the yolk has been absorbed. Granules staining like chromatin are found in the cytoplasm around the nucleus when the yolk begins to disappear in the cells. When the yolk has been absorbed, two cytoplasmic bodies are found, one a spindle-shaped, granular vitelline body, the other an astrosphere (yolk nucleus). The origin of the vitel- line body was not ascertained. The number of the primordial germ cells is small. In one count thirty-six were found. There is no indication during the early stages that the germ cells are derived from any other source than the early segregated cells. C. Period of secondary division After the primary period of rest, already described, during which the germ cells do not divide, they enter upon a period of division (table 2). In the following pages the changes are de- scribed which take place in the germ gland during this period, ~ which begins when the larva is about 20 mm. long. 1. Observations on Entosphenus wilderi. a. Larva 21.5 mm. long, August 23 (fig. 34). In this larva, somewhat more than four months old, the germ-gland fold is still interrupted. While many single germ cells are found and many cysts of less than eight cells, there are a few of from eight to sixteen cells, and the latter number is rarely exceeded in this stage. Most of the germ cells are in the resting stage, but a few are in various phases of mitosis. The gland is 2.5 mm. long and extends from a point near the middle of the mesonephros, about 2.5 mm. back of the cranial end of the coelom, to a point about 2 mm. in front of the anal opening. Its position, therefore, coincides about with the 42 PETER OKKELBERG middle third of the coelomic cavity and not with its whole length as in older larvae. Where the coelomic epithelium covers the germ gland it is thicker than elsewhere and its nuclei are much more numerous. These nuclei are ovoid in shape and are usually so oriented that their long axes are parallel to the long axis of the body and to one another. Each contains, besides a chromatin network, a large nucleolus, usually near its center (fig. 35). Many of the larger chromatin granules lie against the inner surface of the nuclear membrane. No membranes are visible between the cells of the peritoneum, but cell limits are indicated by the lessened thickness of the cytoplasm between the nuclei. All the cells of the germinal peritoneum are similar and in no case was the peritoneal membrane found to be interrupted so that the germ cells might be said to form a part of it. Evidence has already been presented to show that these two kinds of cells are of different origin and that there is no genetic relationship between them. That the follicle cells are derived from peritoneal cells, as in earlier stages, is indicated by the similarity in structure of their nuclei and by the fact that in various places peritoneal cells are migrating inward from the peritoneal covering of the gland to take part in the formation of follicles around the cysts and individual germ cells. When a germ cell or cyst lies against the peritoneum, the cells of the latter are usually heaped up around it in such a way as to suggest that they will eventually enclose it altogether. Some of the peritoneal cells also migrate into the interior of the gland without at first coming into rela- tion with germ cells. This usually happens where the peritoneal membrane is indented, in which case the cells separate from the apex of the indented portion (fig. 35). Whether or not all these cells finally attach themselves to germ cells and form follicles for them is not known, although in some cases they seem to become associated with the more deeply situated germ cells. For a time at least, they make up most of the stroma of the gland, although part of the latter enters the gland from the mesenchyme dorsal to it. The nuclei of these few mesen- chymal stroma cells may be distinguished from the peritoneal GERM-CELL HISTORY IN THE BROOK LAMPREY 43 and follicle cells by their smaller size, structure, and rounded form. The cytoplasm ramifies in all directions so as to form a loose parenchymatous structure (fig. 86). Blood-vessels enter the gland from above with the mesenchyme and fill practically the whole of some sections. The endothelium of these blood- vessels consists of the usual flattened cells containmg much flattened nuclei. The germ cells remain as distinct elements with more or less clear cytoplasm, usually without observable limiting mem- branes, and with large rounded nuclei. In each nucleus there are, as before, two deeply staining nucleoli and a distinct chro- matin network. One has no difficulty in distinguishing the germ cells from the somatic elements of the gland by their nuclear structure and the much greater size of both nuclei and cell bodies. No transitional stages were found between the germ cells and the somatic cells of the gland, and no indication of the transformation of somatic elements into germ cells or germ cells into somatic cells. The number of the germ cells is in- creased rapidly during this stage, but so many of them are in mitosis that the increase in their number may easily be accounted for without supposing that new germ cells are being formed out of surrounding tissue cells. b. Larva 25 mm. long (fig. 37). The structure of the germ gland in a larva of this stage is similar to that of the preceding stage, except that, owing to the enlargement of the gland without corresponding increase in the number of germ cells and cysts, these are more scattered than in the earlier stages. In many sec- tions not a single germ cell occurs. In some sections both germ cells and blood-vessels are absent and the peritoneum encloses only a loose stroma with greatly scattered nuclei. The gland is now suspended from the middorsal line of the coelom by a broad mesentery, in which germ cells are present, while some are still found even in the mesenchyme above it. The germ cells and cysts often lie against the peritoneum and new follicle cells are being formed continually by division and migra- tion of peritoneal cells. 44. PETER OKKELBERG c. Larva 27.5 mm. long. Although the histological features of the gland have not changed from those of the preceding stage, this particular larva was selected for study because certain cytological structures in the germ cells are more clearly shown. One of these is the centrosphere, which differs from that of the 20-mm. larva, in which it was last referred to, in that it shows a distinct centrosome. ‘The vitelline body, as variable in position as in the 20-mm. larva, is visible. It is now more compact and stains black in sections treated with iron haematoxylin. In fay- orable sections the outline of each individual germ cell may be faintly seen in the cysts (fig. 38). The germ cells are in various phases of mitosis, but most of them are in the resting phase, as in previous stages. Those of a cyst do not divide simultaneously, but several cells within the same cyst may be in different phases of division, while others are in a resting condition (figs. 39 and 67). The cysts range from two to over thirty cells and there are a great number of isolated cells. d. Larva 30 mm. long. In most of the larvae of this stage the germ glands are similar to those of the preceding stage. In one (larva no. 622) a large number of germ cells were in mitosis. The products of the division of an isolated germ cell may become enclosed, each in its own follicle, or they may remain together as a part of a nest of cells enclosed in a common follicle. The follicle cells of such a nest often stretch into the spaces between the germ cells, while other follicle cells, recognizable by the form and structure of their nuclei, are detached from the follicular mem- brane and lie free among the enclosed germ cells. Follicle cells occur in all stages of the process of detaching themselves from the follicles and also from the peritoneum, and of penetrating between the germ cells of the nests (fig. 40). In this way, as the germ cells increase in number, old cell nests are broken up and new ones formed, so that in this larva very few nests remain which have more than eight cells. The many follicle cells neces- sary to enclose the increasing number of nests are probably pro- duced by proliferation of cells already in the follicles as well as of those in the peritoneal epithelium. If this be so, the mitoses must take place very rapidly, for dividing nuclei in either fol- GERM-CELL HISTORY IN THE BROOK LAMPREY 45 licle or peritoneal cells are rare. There is no indication that follicle cells are derived from germ cells. If this were the case, transitional stages should be found which might be detected by -nuclei of intermediate form and structure, but such have never been found. e. Larvae 34 to 35 mm. long. In the gonad of this stage many single germ cells, as well as many cysts, occur ranging in size from two cells to one hundred or over. The cysts are sur- rounded, as before, with follicular cells. In many of them the germ cells are in a resting condition, but in some they are in mitosis. Sometimes cell nests, identical with those of the germ gland, occur in abnormal positions. One of these from the fat body is shown in figure 32. Some of the germ cells at this period have entered the synapsis phase, while others have transformed into growing oocytes, the first visible sex-distinguishing character. 2. Historical renew and theoretical discussion. Wilhelm Miiller (75) studied the early stages of development in Petromyzon fluviatilis and P. planeri. He gives an account of the repro- ductive system in a larva 35 mm. long, but the earlier stages were apparently not studied by him. At this stage the sexes could not be distinguished. Oocytes were first observed in larvae 50 mm. long, and in larvae 65 mm. long the ovary and testis were fully differentiated. Lubosch (’03) found the first anlage of the sex gland in larvae of Petromyzon planeri 18 mm. long. He found two types of cells in the gland, the follicle cells and the germ cells, and he believed that the peritoneum was the source from which both were derived. In a slightly later stage he describes a germ ridge as being formed, into which connective tissue and blood- vessels descend from the region dorsal to the gonad. He ob- served that peritoneal cells migrate inward to form the follicles surrounding the germ cells. He found that the glands remained undifferentiated until the larvae are about 4 em. long when the ovaries could be distinguished from the male glands. t No further literature exists on this stage in the development of the germ gland in the lamprey. 46 PETER OKKELBERG - In the lamprey, according to evidence presented above, most of the somatic part of the germ gland comes from the perito- neum. This is certainly true of the follicular epithelium and a considerable portion of the interfollicular tissue. On the other hand, some mesenchyme is included in the gland fold as it is formed and some is carried in later by the blood-vessels. Whether or not the follicular cells take any active part in nourishing the germ cells in the early stages is not known. The only case of a true nurse cell observed in the germ gland of lower vertebrates in early stages is that described by Kuschakewitsch (10). He found what he considered to be nurse cells in frog larvae which developed from his so-called ‘Spatbefruchteten’ eggs. They were distinguished by greatly concentrated nuclei and by cytoplasm filled with granules which he thinks are chromidia. These nurse cells were not, according to him, themselves used as food for the neighboring germ cells, but served as a source of certain ferments which were useful in preparing nourishment for the cells. He says concerning them: ‘Es handelt sich aber um eine Sezernierung von Fermenten, welche es bewirken dass die in der Keimanlage zirkulierenden Nihrstoffe von den Ampul- lenelementen besser ausgeniitzt werden kénnen.”’ A proliferation of peritoneal cells to form follicle cells has been observed by other investigators in various groups of verte- brates other than the cyclostomes. In the frog it has been observed by Bouin (’01), Dustin (’07), Kuschakewitsch (’10), and others. Bouin believes that these cells take part in the formation of germ cells as well as follicle cells. Kuschakewitsch does not believe that they take part to any great extent in the formation of germ cells.’ Dustin states that he observed the transformation of ordinary peritoneal cells into germ cells in the turtle (Chrysemys). Allen (06), however, working on the same form, found no evidence of such transformation. In Triton, Abramowiez (13) observed the proliferation of peritoneal cells in the germ gland, and believes that these cells give rise to follicle cells and interfollicular tissue as well as to germ cells, and to the cells of the sex cords. King (’08) thinks that in Bufo lentiginosus the peritoneal cells give rise to all the elements of the sex gland with the exception of the germ cells. GERM-CELL HISTORY IN THE BROOK LAMPREY 47 The germ cells of the lamprey have not been seen to divide except by mitosis. The nuclei of the cells are always spherical and never have the lobulated appearance which seems to be characteristic of the nuclei of the germ cells of amphibians. Several investigators (Vom Rath, 791; Meves, 91; McGregor, 99) have recorded amitosis as taking place in the germ cells of the latter group, but it is possible that the appearance of ami- tosis comes from sections through lobulated nuclei. Recently Macklin (16) has made a study of apparent amitotic phenomena in heart cells of chick embryos growing in vitro, and has found that nuclei divide by bilateral and unilateral constriction, but that afterward the parts of the nuclei recombine and divide by normal mitosis. It may be that in some animals similar processes take place in the germ cells. In the lamprey the cells in the same cyst do not always divide simultaneously, though generally, if the cells in a cyst are divid- ing, most of them are in one phase or another of mitosis. At the same time the cells in neighboring cysts may all be in a resting stage. Since the germ cells do not multiply very rapidly, there is a considerable period of rest between successive simultaneous divisions in a cyst. The synchronous division of the cells of a cyst may be due to their close relationship, all being derived from one cell and having been subjected to similar environmental influences, or to some stimuli from outside sources, the effect of which is limited to a single cyst. King (’08) found that in Bufo the cells of a cyst did not divide simultaneously. According to Jorgensen (’10), this is also true in Proteus. Bouin (’01), how- ever, describes the cells of a cyst as dividing simultaneously in the frog. It seems likely that there may be considerable variation in this respect in different forms and even in the same species. Since, in the lamprey, most of the cysts are broken up from time to time by the inward migration of follicular cells, it is impossible to say whether or not there is a constant number of divisions of the indifferent germ cells. In some cases the cysts become very large and contain hundreds of cells, but it is not certain that only these came from one primordial germ cell. In most, if not in all cases, each primordial germ cell gives rise to many JOURNAL OF MORPHOLOGY, VOL. 35, NO. 1 48 PETER OKKELBERG cysts, probably to an indefinite number. In Bufo it appears that the cysts do not break up, for, according to King (’08), all the cells of a cyst are descendants of one primary oogonium in the female, and the cyst wall is formed out of the original follicle cells which surrounded the primary oogonium. Witschi (14) states that the number of cells formed from the primary oogo- nium in Rana is at least thirty-two in some cases, but often fewer, while in the testis the number is greater. In the turtle Clemmys, Munson (’04) found that the number of divisions was three, each cell thus giving rise to eight cells. It is impossible to distinguish oogonia and spermatogonia in the lamprey until the larva is about 35 mm. long. Before that time the cells in all larvae appear structurally alike and divide in a similar manner. The centrosphere with its centrosome corresponds to the yolk nucleus described by Lubosch (’04) in the larvae of lampreys about 4 cm. long. He describes it as an oval, clearly defined body of the same structure as the sur- rounding plasma, but not staining so deeply. Surrounding it, like a membrane, he found deeply staining granules. Lubosch believes that the yolk nucleus introduces the process of yolk- building; but this cannot be its function in very early stages (larvae up to 35 mm. in length) while the cells are still dividing and long before there is any formation of yolk. In these stages it probably functions in cell division in the germ cells as it does in other cells. During the subsequent growth period it may play some part in the process of yolk-building. The vitelline body is probably present in all indifferent germ cells. It seems to be a permanent element in the germ cells of the lamprey, but it was not possible to determine its origin. King (’08) believes that in Bufo it must be considered as a secretion product of the cytoplasm itself, but she thinks it not improbable that a fluid, possibly an enzyme, may pass from the nucleus into the cytoplasm and there cause the formation of the ' body, and that this enzyme, while in the nucleus, may be in the form of plasmosomes. Dodds (’10) has described a body in the eytoplasm of the early germ cells of Lophius and believes that it is a mass of plasmosome material that has been separated and GERM-CELL HISTORY IN THE BROOK LAMPREY 49 cast out of the nucleus. The act of extrusion was not observed. The body found by Dodds, however, has only a transitory exist- ence in the cytoplasm, so it cannot correspond to the vitelline body. Since the vitelline body is found to be a very prominent structure in the oocyte of the lamprey, a full discussion of it will be reserved for a future paper which will describe the growth period of the egg. Witschi (’04) found that sex could be distinguished in frog larvae before there was any differentiation of the germ cells, by the fact that in the larvae destined to become females the germ cells remained along the periphery of the gland, while in those destined to become males the cells migrated into the interior of the gland in very early stages. No differences of this kind have been found in the larvae of the lamprey during the indifferent period. Neither are any genital cavities formed in the gland by means of which the sexes may be distinguished pre- vious to germ-cell differentiation, as is the case in amphibians. 3. Summary of the period of secondary division. During the period of secondary division (larvae 20 mm. to 35 mm. long) the germ cells multiply by frequent mitoses. The resulting cells may remain together after division, enclosed by a common follicle, or they may become separated by the migration of follicle or peritoneal cells between them. The result is that in all glands both cysts and isolated germ cells are found. The germ cells are distinguished from the somatic elements of the _ gland by their size, the structure of their cytoplasm, and the form and structure of their nuclei. There is no indication that germ cells are derived from somatic cells or somatic from germ cells. Usually most of the cells of a cyst are found in one or another phase of mitosis at the same time, but some of the cells of the cyst may be in a resting stage while others are dividing. In the cytoplasm of the germ cells a centrosphere with a centro- some is often visible; also a vitelline body of unknown function and origin. Spermatogonia and oogonia cannot be distinguished during this period. Neither are there any other characters by means of which future males and females may be distinguished. 50 PETER OKKELBERG D. Period of sex differentiation 1. General statement. The period following that of secondary multiplication of the germ cells is characterized in dioecious ani- mals by the differentiation of male and female individuals. The latter may be distinguished from the former by the appearance in the germ gland of yolk-filled oocytes. ‘This phenomenon is accompanied or preceded by somatic changes in the germ gland and elsewhere. The somatic differentiation may take the form of changes in the structure of the germ gland, the appearance of accessory reproductive organs peculiar to one sex or the other, or the development of other secondary sexual characters. In the lamprey there are no secondary sex characters developed until after metamorphosis, before which the sexes cannot be dis- tinguished except by an examination of the germ gland, and in early stages even the germ gland does not form a criterion of the future sex of the animal. 2.. Sex characters in the adult brook lamprey. ‘The reproductive gland in the adult lamprey is unpaired and is suspended by a mesentery from the middorsal line. Previous to spawning, it fills practically the whole body cavity in both sexes. The surfaces of both the male and female glands are thrown into more or less oblique folds, more easily seen in the testis than in the ovary. The testis is made up of numerous cysts filled with spermatozoa and enclosed by follicle cells. It is supported by the mesorchium, from which connective tissue cords radiate into the body of the gland. In the ovary each ovum is enclosed by follicle cells as are the cysts in the testis. Connective-tissue cords, similar to those of the testis, radiate into the body of the gland from the mesovarium above. The number of ova varies in different females as also does their size. In the same animal, however, the ova do not vary greatly in size. ~ The cysts in the testis and the ova in the ovary may be con- sidered homologous structures. In the one case the germ cells have continued to divide, while in the other case they have stopped dividing early in the life of the animal and have entered upon a period of growth. The greatest amount of growth takes place in the female after metamorphosis, while in the male GERM-CELL HISTORY IN THE BROOK LAMPREY on metamorphosis is followed by a period of very rapid multiplica- tion of the germ cells. In both sexes the germ cells reach the outside through abdominal pores. These open in each sex into a common urogenital sinus which terminates externally in a urogenital papilla. There are only a few external sex-distinguishing characters. The urogenital papilla is short and wide in the female, with a large opening at the end, while in the male it is long and slender with a small opening at the tip (fig. 66). Both sexes have a lateral fold of skin on each side of the papilla. In the female the papilla is hidden by these folds, but in the male it extends beyond the folds as a prominent structure. In the female a small anal fin, connected with the caudal fin by a low ridge of skin, lies directly behind the urogenital papilla, while in the male the anal fin is absent or rudimentary. Loman (’12) states that in the European brook lamprey the anterior dorsal fin is always lower in the female than in the male, and Gage (’93) found the cranial end of the second dorsal fin always swollen in the female of the American brook lamprey. From the above description it will be seen that the reproductive organs in the lamprey are reduced to extreme simplicity. The sex glands and the accessory structures, by means of which the reproductive elements are extruded, are morphologically similar in the two sexes. This point is emphasized because of its bearing on the tendency of the animal toward hermaphroditism. 3. Changes in the germ gland during the period of sex differen- tiation. In larvae less than 35 mm. in length the majority of the germ glands are in an fndifferent condition as regards sex. During these stages the germ cells are scattered irregularly through the germ gland, either singly or in cysts; either in a resting condition or in various phases of mitosis, but they are all alike. This indifferent period is followed by one in which distinct sex characters appear in the germ glands through the develop- ment of oocytes found both in the synaptic and growth phases (tables 1 and 2). Whether the germ glands eventually become ovaries or testes, they all develop this female character, and the animals may therefore during this stage be considered interme- diate as to sex or hermaphroditic. 2 PETER OKKELBERG During this period (larvae 35 mm. to 70 mm. in length), therefore, one may find in any germ gland, germ cells and. cysts which are in all respects like those of the indifferent period, as well as germ cells, located singly or in cysts, which are in the various stages of the synaptic phase or which have entered the growth period. To avoid confusing the germ cells which are in the various — phases of mitosis with the cells that had already entered the prophases of maiosis, it seemed necessary to follow as carefully as possible the various steps in the two processes. On account of the small size of the cells and the large number of chromo- somes present, it was very difficult to get a complete history, but the main features were worked out. a. Changes during mitotic division. The resting germ cell. The resting cell (fig. 41) has a spherical nucleus in which the chromatin material is in small and large granules or masses, united by fine achromatic threads. Two rather large plasmosomes, at some distance from each other, are present in the nucleus. In cells stained in iron haematoxylin and afterward destained so long that the chromatin masses can no longer be seen, the plasmosomes retain the stain and stand out as two very distinct elements. The formed elements of the nucleus are surrounded by a clear homogeneous nuclear sap, and the whole nucleus is enclosed by a nuclear membrane. The amount of cytoplasm is rather small. An astrosphere with a centrosome lies along one side of the nucleus, although the centrosome is not always vis- ible. A vitelline body sometimes occurs. The cytoplasm is granular and, in cells, fixed in Meves’ solution and afterward stained in iron haematoxylin, mitochondria, in addition to finer protoplasmic granules occur. Prophase. During the first phases of mitosis (fig. 42), the amount of chromatin increases greatly. The original chromatin granules of the resting cell grow in size but remain united by linin threads, so that the whole gives the appearance of a network with conspicuous masses of chromatin at the crossings of the threads. The. two nucleoli remain distinct as before. GERM-CELL HISTORY IN THE BROOK LAMPREY 58 A little later (fig. 43) the chromatin masses become still more conspicuous, but they continue to be held together by the achromatic threads. The nucleoli are still large, but generally one is considerably smaller than the other. They are closer together than in the preceding stage. The chromatin granules or masses soon become separated from one another by the breaking of the achromatic threads which up to this time have been holding them together (fig. 44). Each mass is now clearly a short globular chromosome. It was. im- possible, however, to count the chromosomes with any degree of accuracy, either at this or any other stage. At this time only one nucleolus is visible. The nuclear membrane is still intact. There is some indication that the chromosomes are arranged in pairs. In the next stage (fig. 45) the nuclear membrane disappears and the chromosomes come together in a mass at about the equator of the cell. The two centrosomes occur at opposite poles, and achromatic spindle fibers extend from the centro- somes to the chromatin mass. The nucleolus is no longer visible and its fate is not known. A polar view of this stage is shown in figure 46. The chromosomes are so closely massed that it is difficult to distinguish each individual chromosome. Metaphase. Although numerous dividing cells were exam- ined, none was found in which the splitting of the chromosomes could be observed. It seems likely that this process takes place so rapidly that the chances of finding a cell in this stage are slight. Furthermore, the chromosomes are so closely massed together on the equatorial plate that their division would be difficult to observe. Anaphase. Figures 47 and 48 represent cells in early anaphase. The daughter chromosomes have already begun their migration to the opposite poles. Occasionally bodies having the appear- ance of chromosomes are found outside of the spindle in the cytoplasm of the cell, or they may be scattered along the spindle threads. Telophase. In this stage (figs. 49 and 50) the daughter chro- mosomes have separated from each other and form irregular 54 PETER OKKELBERG groups at opposite poles. The spindle is still distinct. The cell is constricted in the middle and the two daughter cells are about to separate. Sometimes a deeply staining strand or thread remains between the two chromatin masses after they are some distance apart (fig. 49). In a slightly later stage a distinct mid- body is seen along the line of separation (fig. 50). Soon after this the two daughter cells separate, a new nuclear membrane forms around each daughter nucleus, and the two nuclei are in a stage of reconstruction. The two daughter cells are smaller than the mother cell (fig. 51). Small chromatin masses reap- pear and are united by connecting achromatic threads. At least one nucleolus occurs in each cell. After division the cells enter a period of growth until they have assumed the structure and size of the mother cell. Discussion. There is nothing strikingly peculiar in common mitosis of the germ cells of the lamprey and only a few obser- vations need further comment. First, the process of division is probably not very rapid. This statement is based on the fact that the number of germ cells does not at any time increase very fast. From the number of germ cells in the mitotic phase at any one time, it is evident also that the karyokinetic period must be rather long, probably occupying several days. The chromosomes are more or less rounded and stand out most clearly in the middle prophase (fig.44). The number of chromo- somes is very large, and they often give the impression of being in pairs. This may mean one of two things: either that the chromosomes have already divided during the prophase stage before they have reached the equatorial plate or that the mater- nal and paternal chromosomes remain associated during the prereduction stages. The latter was found by Chubb (’06) to be the case during the multiplication period in the germ cells of Antedon. Stevens (’07, ’08) found that in Diptera a pairing of the chromosomes took place in germ cells far removed from the reduction stages and occurred in connection with each oogo- nial and spermatogonial division. Metz (’16) has reinvestigated the problem in Diptera, and from a study of about eighty species has come to the conclusion that in somatic, as well as in germ GERM-CELL HISTORY IN THE BROOK LAMPREY 515) cells, the chromosomes are associated in pairs. The paired con- dition persists throughout the various phases of cell division. An association of maternal and paternal chromosomes is appar- ently effected during early cleavages, and probably before the first cleavage, thus continuing from the fertilized egg to the adult stage. In Diptera the pairing is side by side (parasyndesis) and similar to synaptic pairing. The pairing, according to Metz, comes about through a physicochemical similarity of the homologous maternal and paternal chromosomes. In the lamprey the granules on the achromatic network of the resting nucleus form the centers for the reconstruction of the chromosomes. ‘There is little doubt that these granules actually represent chromosomes in all the various phases through which the cells pass; for this reason one may speak of a visible continuity of chromosomes from one cell generation to the next. Thus the individual chromosomes in the germ cells of the lamprey never lose their identity during mitosis. The origin, function, and fate of the plasmosomes remain obscure in the cells of the lamprey. In the resting cell there are usually two plasmosomes which appear approximately of the same size and which lie some distance apart. As the phe- nomena of prophase advance, they approach each other, and somewhat later only one is present. Whether a fusion of the two takes place or one dissolves and disappears at this time is not known. During late prophase the single plasmosome also disappears or at least is no longer distinguishable among the chromosomes. During the telophase a new plasmosome soon appears in each daughter cell and, as the growth of the cell progresses, a second plasmosome also appears. It is likely that the plasmosomes dissolve during cell division to be formed de novo in the daughter cells. b. Synapsis phase of the oocytes. After an indefinite number of divisions, some of the primordial germ cells (oogonia), which lie singly or in cysts, undergo a series of changes preliminary to the stage of actual growth. When the cells have entered this stage they are termed oocytes of the first order (table 1). 56 PETER OKKELBERG Very little was known concerning the processes that take place in the oocyte of any animal preliminary to its growth period, until von Winiwarter (’01) published his extensive observations of this period in the cat and in man. Based upon changes which take place in the nucleus, he divided the transition period from the oogonium to the oocyte into the following periods: 1) Noy- aux protobroques; 2) noyaux deutobroques; 3) noyaux leptoténes; 4) noyaux synaptenes; 5) noyaux pachyténes; 6) noyaux diplo- ténes, and, 7) noyaux dictyes. In a later publication by von Winiwarter and Sainmont (’09) another stage has been added between the first two, namely, noyaux poussiéroides. Other investigators have divided the period differently and have applied other terms to the different phases, but on the whole the processes taking place in all animals in which the synapsis phase has been studied seem to follow the general course outlined by von Winiwarter. In the lamprey the changes correspond in the main with those of other forms studied. An abbreviated list of the above stages has already been given (table 1). Nuclear changes. Early leptotene. After the last oogonial division the germ cell enters a period of rest, during which the chromatin of the nucleus becomes broken up into small particles (fig. 52). These are scattered throughout the whole substance of the nucleus so as to make its contents appear almost homo- geneous. ‘There are no very distinct chromatin bodies in the nucleus at this stage and the only stainable parts are a very fine network and two very distinct plasmosomes. This stage cor- responds to the stage in the germ cells of the cat described by von Winiwarter and Sainmont (’09) as ‘noyaux poussieroides,’ and to the stage in the germ-cell history in Proteus described by Jorgensen (710) as ‘erste Zerstaéubung.’ Sonnenbrodt (’08) studying this period in the germ cells of the chick, found that the chromatin at this stage was very small in amount, and he believes that the period succeeding the last oogonial division is devoted largely to the formation of new chro- matin. He says it ‘‘besteht in der Hauptsache in der chromatin Aufnahme oder richtiger Chromatinbildung.” In the lamprey, too, the chromatin network begins to reappear at a somewhat GERM-CELL HISTORY IN THE BROOK LAMPREY GY more advanced stage, but for a long time the chromatin remains in a finely divided condition. The nucleus appears to grow larger as the period progresses. The two nucleoli persist through- out the period. Bouin (01) studied this period in the oocyte of the frog, and called the cells of this stage ‘ovogonies de transition.’ He de- scribed ‘the nucleus as losing its membrane at this time so that there was a free communication between the nuclear and cyto- plasmic substances. This has not been confirmed for other forms and it is certainly not the case in the lamprey. Late leptotene stage. During this stage (figs. 53 and 54) the chromatin network becomes much more distinct. The whole nucleus is now filled with chromatin threads which cross one another in various ways. Irregular thickenings are found on the chromatin threads not only at their intersections, but in other parts as well. Synaptene stage. During this stage (fig. 55) the chromatin becomes massed together along one side of the nucleus in the form of an irregular tangle of rather thick, deeply staining threads, and forms what has been termed a ‘contraction figure.’ This is the stage described by von Winiwarter as the synaptene stage and by Maréchal (’04) as the ‘bouquet stage.’ On the side of the nucleus on which the chromatin is concentrated, the individual threads can no longer be distinguished; but in the clearer parts of the nucleus many of the ends of the chromatin threads extend out from the concentrated mass, sometimes as far as the nuclear membrane of the opposite side. Occasionally a nucleolus, which later appears to be lost, occurs during the early phases of this stage. It is possible that later, when it is not visible, it is hidden among the chromatin threads of the contraction figure. If in reality this be the case, it indicates a tendency of the chromatin to concentrate around the nucleolus during this stage; for other- wise, if the nucleolus be present, it should be found occasionally in the clearer portions of the nucleus. Some investigators have termed this period the synapsis stage, because in many forms the chromatin threads come together in pairs at this time. McClung (05) has called it the ‘synizesis stage,’ and this is a more appro- 58 PETER OKKELBERG priate term, since apparently synapsis does not always occur during this period. In the case of the lamprey it was not pos- sible to find any pairing of the chromatin threads at this time, although hundreds of cells were examined. Pachytene stage. In the pachytene stage (fig. 56), the chro- matin material again becomes uniformly distributed throughout the nucleus. It is now in the form of thick threads which appear more or less continuous in some places, but are generally broken up into segments. One large nucleolus appears. There is no indication that the chromatin threads are paired. Diplotene-dictyate stage (diakinese) (figs. 57, 58, 59, 60, and 62). In an oocyte somewhat farther advanced than the above, the whole chromatin network has become broken up into definite chromosomes, and the paired structure of every chromosome is very apparent, a condition which persists throughout the early part of the growth period and probably up to the time of matura- tion. It is very difficult, however, to follow the history of the chromosomes during the later growth period, since the nucleus becomes very large and the chromatin material may be scattered throughout its whole extent. ‘The nucleolus at this time is large and almost spherical. It has not been possible to observe any relation between the nucleolus and the chromosomes at this period of development. Many of the paired chromosomes le against the nuclear membrane, but they may occur also in vari- ous other parts of the nucleus. Very often they are arranged in the form of tetrads which are best seen along the nuclear mem- brane (fig. 61). There appears to be no regularity in the arrange- ment of the chromosomes. The dictyate stage (diakinese) con- tinues during the growth period of the oocyte. Cytoplasmic changes. The nuclei of the oogonia are sur- rounded by a small amount of granular cytoplasm. Often no visible cell boundaries are present, although favorable sections show that the cells do not form a syncytium, but are morpho- logically independent of one another, in spite of the fact that no true cell membrane is found. In cells, fixed in Meves’ solution or in other solutions that fix mitochondria, these occur in great numbers. They are usually granular and appear to be more GERM-CELL HISTORY IN THE BROOK LAMPREY 59 or less grouped. ‘There is no evidence that they are derived from the nucleus, so they may be considered as true cytoplasmic bodies. The mitochondria in the early oocytes of the lamprey are not essentially different from those found in the indifferent germ cells (figs. 41 and 52). During the progress of the early nuclear changes, up to the time of the synaptene stage, there is a gradual decrease in the amount of cytoplasm and a gradual disappearance of the mito- chondria. After the synaptene stage the amount of cytoplasm increases again, but no study was made of the mitochondria sub- sequent to the synapsis phases. In the undifferentiated germ cells there is an astrosphere near the nucleus, even in the resting cells. Sometimes a minute cen- trosome has been distinguished in the middle of the astrosphere. The astrosphere may be distinguished from the surrounding cyto- plasm by its more granular appearance. During the middle synapsis phases it seems to disappear with the decrease in the amount of cytoplasm, but it reappears in the early growth period and remains through this whole period as a very promi- nent structure (fig. 60). The vitelline body may be traced through the various phases of the synaptic period, and during the growth period it becomes a very conspicuous structure (figs. 60, 62). Discussion. The general history of the period has been out- lined above, but certain features require further discussion. The changes taking place in the cells during this period initiate the period of heterotypic or reduction division. This is a period through which all germ cells apparently must pass before they can become functional ova or spermatozoa. In the female cell these changes take place at a very early stage in the development of the animal, and in the ease of the lamprey they precede the maturation period proper by at least two or three years. Inthe male lamprey these changes occur much later in the life of the individual and usually precede the maturation divisions by only a very short time, probably not more than three or four months. I have made a study of the synaptic phases of the male germ cells in the lamprey and found the process of development to be 60 PETER OKKELBERG much like those of the female. In both sexes the germ cells presumably come out of the synaptic phase with the number of chromosomes reduced to one-half the number found in the pre- synaptic cells. The subsequent history of the cells, however, differs in the two sexes. In the female the cells grow to an enor- mous size by the accumulation of yolk, while in the male there is very little growth. . Aside from the apparent pairing of the chromosomes during the period of multiplication, no other instance of chromosome pairing was observed in the cells before they have reached the diplotene stage; but it cannot be said with certainty that a doub- ling does not take place before this stage during the synaptic phase. In some forms which have been investigated, synapsis seems to take place during the synaptene stage, or during the period of transition from the leptotene to the synaptene. Von Winiwarter (01) figures a pairing of the chromatin threads in the germ cells of the rabbit during the early synaptene. Von Winiwarter and Sainmont (’09) describe a similar condition in the cat. In Proteus also, according to Jérgensen (’10) the leptotene stage is followed by a stage which shows a double nature of the chromatin threads. In Bufo, King (’08) figures double chroma- tin threads for the first time after the synaptene stage. Maréchal (04) observed the double structure of the threads during the synaptene stage in Pristiurus and Secyllium and later (’07), in Ciona and Amphioxus. Janssens (’04) found that in Triton the reduced number of chromosomes, or chromatin filaments, appears ~ shortly after synizesis and that these filaments subsequently split longitudinally forming two sister threads which remain together. d’Hollander (’04) found a massing of chromatin (synizesis) in the oocyte of the hen before synapsis. The phenomenon of synizesis (McClung, ’05) has been found by various investigators to occur in the oocyte of invertebrates as well as in vertebrates; it seems to be a universal phenomenon of the early heterotypic prophase. Chubb (06) thinks that synapsis takes place in Antedon during the oogonial divisions, and that it is followed by still one more division. In Sycon Jorgensen (’10) thinks that the reduced number is present in GERM-CELL HISTORY IN THE BROOK LAMPREY 61 the oogonia. According to these investigators, synapsis may take place previous to synizesis, and the two phenomena proba- bly have nothing in common. Most investigators, however, agree that the double nature of the chromosomes is first visible during a late stage of the heterotypic prophases, but their inter- pretations of the doubling vary. Some consider it a suppressed mitosis (Hertwig, ’08; Matscheck, ’10, and others), while the majority of workers on germ cells look upon it as a pairing of parental chromosomes similar to that which takes place in the male germ cells previous to the maturation division. Very little light can be thrown upon this subject by the lamprey, since it was found impossible to count the chromosomes before or after synapsis. To all appearances, however, the chromosomes enter synizesis in the univalent condition. The bivalent nature of the chromosomes is not observable before the diplotene stage. The meaning of the ‘contraction figure’ has been variously interpreted. Some investigators consider it simply an artifact due to poor preservation (Janssens, 05; Jorgensen, 710, and others). Maréchal and Saedeler (’10) insist that it is not an artifact in Raja clavata. King (’08) has shown that it is a perfectly normal condition in the toad. In the lamprey it appears to be a normal phenomenon, and forms a stage in the morphological changes which take place in the oocyte at this time. In the same gland were cells in the contraction. phases, other cells in the various stages of the synapsis phase, normal resting cells, and cells in the different phases of mitosis. Degenerating cells also occurred in most glands, but no evidence was found to indicate that the contraction figure is a phase in the process of degeneration. Even in the same cyst, there are contraction figures side by side with resting cells and cells in other stages of the synapsis phase. It must be concluded, therefore, that the phenomena connected with synizesis in the lamprey are perfectly normal and due to some peculiar condition of the cells at this time—a condition the nature of which is not yet understood. Whether the contraction figure is normally formed around the nucleolus or on the side next to the centrosome could not be deter- mined. A body sometimes occurs in the cytoplasm near the 62 PETER OKKELBERG nucleus on the side toward which the contraction figure is formed, but this was taken to be the vitelline body since it may often be found in other parts of the cells. No distinct centrosome or attraction sphere could be found in the cells at this time. J6r- gensen (’10) figures a very distinct astrosphere during the bou- quet stage in Proteus, but he was not certain of the presence of a centrosome. He also found that during this period there was an extrusion of chromatin material from the nucleus into the cytoplasm on the side next to the astrosphere. This was not found to be the case in the lamprey, in which the nuclear mem- brane appears to be intact throughout the period of transforma- tion of the germ cells into growing oocytes. No centrosome was found by King (08) in the germ cells of the toad during the synapsis period, and she concludes that probably the egg cen- trosome disappears after the last oogonial division. Lams (’07), on the other hand, observed a centrosome in the germ cells of the frog during the bouquet stage. Although no centrosomes or astrospheres were found during the synapsis phase in the oocytes of the lamprey, they are not permanently lost, for they reappear somewhat later in the grow- ing oocyte. It is probable that special technique might make them visible also during the synapsis stages. J6rgensen holds that the centrosome is functional in connection with the con- vergence of the chromosomes along one side of the nucleus during the bouquet stage, with the radiations in the cytoplasm through the orientation of plasma inclusions, and with the formation of a permeable region in the nuclear membrane where chromatin bodies may be extruded from the nucleus. In the lamprey no extrusion of visible chromatin material from the nucleus at this stage has been observed, but there is ample evidence that such extrusion takes place in the dictyate stage, although there appears to be no special area of the nuclear membrane over which it occurs. It has been shown that two nucleoli are present in the germ cells of the lamprey during the multiplication period. During mitosis these are reduced to one, which also later disappears. In the resting cells after mitosis a single nucleolus appears and GERM-CELL HISTORY IN THE BROOK LAMPREY 63 shortly afterward a second. When the cells enter the synapsis phase, the nucleoli again disappear, apparently during synizesis. When the cells enter the growth phase, only one nucleolus appears in each cell. This remains during the whole period as a very prominent spherical structure. Whatever may be the function of the nucleoli, they are, as shown by their reaction to stains, true plasmosomes, and not composed of chromatin mate- rial. This is contrary to the idea of Lubosch (’03), who believes that in the lamprey and in other forms the chromatin material is stored in the nucleolus during the growth period of the egg. This view is based largely on the fact that the chromosomes seemingly disappear during the later stages of growth. Lubosch also thinks he has evidence that the maturation chromosomes are derived from the nucleolus. JI have found some evidence which indicates that the maturation chromosomes appear in the clearer portions of the nucleus and do not come from the nucleolus. Von Winiwarter and Sainmont (’09) found in the nucleus of the oogonia of the cat, at the time when the cells were preparing for mitosis, an elongated body which stained like chromatin. Often it had a horseshoe shape, and it was larger than the other bodies of the nucleus. During mitosis it divided, but much more slowly than the chromosomes. In the oocyte it was often attached to the plasmosome, but sometimes it was free. When dividing, it split longitudinally, and during the growth period it disappeared. The body was supposed by the authors to be a sex chromosome (monosome). Gutherz (12) found a similar body in the spermatocytes of the cat, but came to the conclusion, on account of its staiming reaction, that it was a true plasmosome “der einen Gestalt ein Heterochromosome in Herteropyknose vertauscht.” Gutherz doubts that the body observed by von Winiwarter and Sainmont was a true sex chromosome, since no differential stain was used by them. Furthermore, there should be two sex chromosomes present in the oocyte of the cat, if such bodies are present at all, since in this form the male is apparently heterozygotic with respect to sex. There is considerable danger of misinterpreting nuclear bodies. Many of the structures described in the germ cells of vertebrates JOURNAL OF MORPHOLOGY, VOL. 35, NO. 1 64 PETER OKKELBERG as sex chromosomes may be plasmosomes and have nothing to do with the determination of sex. Wilson (713) found a body in the spermatocytes of Pentatoma that simulated an accessory chromosome, and which he called a ‘chromatoid body.’ A similar body has been found by Wodsedalek (14) in the spermatocytes of the horse and by Bachhuber (16) in the spermatocytes of the rabbit. More recently, Swingle (’17) describes what he consid- ers to be the same kind of body in the spermatogonia of the frog (Rana pipiens and R. catesbiana). The body was in all these cases of cytoplasmic origin and was found with the chromosomes only during mitotic division. An examination of the various figures given of a so-called sex chromosome in vertebrates reveals a striking resemblance to plas- mosomes similar to those that are found in the early oocyte of the lamprey. Stevens (’11) describes such bodies in the spermato- cytes of the guinea-pig; two such bodies were found by Wodse- dalek in the spermatocyte of the pig; similar bodies were found by Levy (15) in the spermatocyte of the frog. Guyer (09, 716) has described such bodies in the spermatocytes, in the oocytes, as well as in the body cells of the fowl, and ascribes to them a sex-determining function. Finally, Jordan (’14) has found such a body in the germ cells of various mammals. Bohring and Pearl (’14) have studied the body found by Guyer in the domestic fowl, and have come to the conclusion that it is not a sex chromosome. At the present time the status of the sex chromosome in the germ cells of vertebrates is very uncertain. It is unfortunate that the subject has been studied almost exclusively in the male germ cells. Von Winiwarter’s account of such a body in the oocyte of the cat and Guyer’s description of the body in the female germ cells of the fowl seem to be about the only accounts dealing with the sex chromosomes in the female germ cells of vertebrates. It is generally assumed that when the spermatozoa in a species are dimorphic, the female of the same species produces eggs only of one kind. These correspond in their chromosome make-up to the male cell possessing the sex chromosome. All the eggs, therefore, in such species possess accessory chromosomes. On the GERM-CELL HISTORY IN THE BROOK LAMPREY 65 other hand, it is assumed that if the female produces two kinds of eggs which differ in their chromosomal structure by the pres- ence or absence of a sex chromosome or chromosome complex, the male must produce only one kind of spermatozoa. In this case one-half of the ova should correspond to the spermatozoa in their number of chromosomes. Guyer (716) says that in the fowl, where it has been shown experimentally that the female is heterozygous for sex, there are also two kinds of spermatozoa. He believes it is probable that only the spermatozoa containing the odd element become functional. In the lamprey no evidence has been found of the presence of an accessory chromosome in the oocyte during the synaptic phase, the growth period, or the maturation division stage. A search has also been made for this body in the spermatocytes during the various stages of maturation, but without success. If, as seems to be the case in some invertebrates and in Necturus among vertebrates (King,-’12), the sex chromosomes might be united with other chromosomes, it would be extremely difficult to find it in forms like the lamprey where the chromosomes are very small and numerous. Observations on the lamprey can neither affirm nor deny the existence of sex chromosomes which might be responsible for sex. It can only be said that such a body has not been found. Whether or not the assumption of the presence of such a body is necessary to account for sex in forms like the lam- prey where the sex potentialities are so equally balanced, is a question which will be discussed later. I have not found in the lamprey a transfer of visible material between the nucleus and the cytoplasm of the oocytes during the synaptic period, but have found that there is an intimate relation between the two parts of the cells. ‘The absolute amount of cytoplasm decreases greatly during this stage, and it is not until the nucleus enters the diplotene phase that the cytoplasm begins to grow again. All the energy of the cell seems to be devoted to nuclear changes in the early oocyte and to cytoplasmic changes during the growth period which follows synapsis. The mitochondria, which are abundant in the oogonia, disappear dur- ing the synaptic phases or, at least, can no longer be seen. In 66 PETER OKKELBERG this respect they seem to behave like zymogen granules in gland cells. In the resting gland cells the zymogen granules are very abundant, but in a cell which is secreting they decrease in num- ber and size, and may disappear entirely if activity continue. Most of the mitochondria in the germ cells of the lamprey are spherical, but occasionally rod-shaped ones may also occur. The theory that the early germ cells may be distinguished from the somatic cells by the shape of the mitochondria has been developed by Rubaschkin (10) for mammals, Tschaschkin (’10) for birds, and Aunap (’13) for fishes. These investigators think that the mitochondria of the germ cells are spherical, and that during the process of development of the embryo they become chain-like and finally rod-shaped in the differentiated tissue cells. The primitive character of the germ cells is, therefore, according to these investigators, indicated by the fact that they possess granular mitochondria after the other cells of the embryo show rod-shaped ones. ‘That this is a universal character of the early germ cells has been denied by von Berenberg-Gossler (’12) and others. Von Berenberg-Gossler found that in the individual germ ’ eells of the duck and the chick, the shape of the mitochondria may vary from granular and chain-shaped to rod-shaped. Firket (14) also found that in the germ cells of the chick, the shape of the mitochondria is not constant. In the oocyte of the lamprey the mitochondria may be found again after the beginning of the growth period. They are cyto- plasmic structures and not related to the chromidia which are so abundant in the growing oocytes of the various stages. This is in agreement with Schaxel (’10, ’11) and others who consider the chromidia to be of nuclear origin and the mitochodria to be of cytoplasmic derivation. Meves (’08), the first to study the mitochondria in embryonic cells, upheld the theory which had previously been advanced by Benda (’03) and others, that the mitochondria are bearers of cytoplasmic heredity. This theory has since been advocated by Duesberg (08, 710) and others. Those who adhere strictly to the chromosome theory of inheritance are opposed to it. Ac- cording to Cowdry (’16), the chemical nature of the mitochondria GERM-CELL HISTORY IN THE BROOK LAMPREY 67 seems to oppose the idea that they are individual constant bodies in the same sense that the chromosomes are considered to be so. It is more likely that they play an active réle in the meta- bolic activities of the cell and that they may vary in number, size, and shape as the activity of the cell varies. c. History and fate of the germ cells during the period of sex differentiation. This period (table 2) includes larvae from about 35 mm. to about 70 mm. long. Figure 68 is a cross-section through the germ gland of a larva 54 mm. long, from the middle of this period. It shows many large, growing oocytes, as well as many cysts. A comparison of sections from various larvae shows that the oocytes are formed from germ cells which have entered synapsis, either while isolated or while constituent ele- ments of small cysts, usually of less than eight cells. In the latter. case the cysts become broken up by the penetration into them of follicular cells, so that each of their contained cells, while still in the synaptic phase or in the early growth phase, becomes isolated within its own follicle. Thus, as shown in figure 68, - nearly all of the growing oocytes are sooner or later isolated cells. In addition to these cells the section contains numerous large cysts, and of these there are two kinds. In one kind thegerm cells are still indifferent, as shown by the fact that they have not entered synapsis, but are dividing by typical mitosis. Such a cyst containing two cells in mitosis is shown at the left of the figure. In the second kind of large cyst the cells have entered synapsis. Such cysts are shown to the right of the figure and are recognizable by the different behavior of the chromatin. In large cysts the cells which enter synapsis rarely become growing oocytes, but sooner or later degenerate, until finally they and their enclos- ing follicles break up and disappear. A longitudinal section of a gland from a larva 59 mm. long, in which two such degenerating cysts are found, is shown in figure 69. The cyst at deg.cy.2 is in a later stage of degeneration than the one at deg.cy.1. A detailed drawing of such a cyst from another section is given in figure 65. The cells of such a cyst rarely get beyond the synize- sis stage before degeneration sets in. Degeneration begins with a condensation of the chromatin into solid masses, as shown at 68 PETER OKKELBERG deg.g. in the figure. The cytoplasm disintegrates and finally the chromatin masses break up and the fragments are scattered through the whole cyst. Such fragments are shown at gf. Other stages occur showing the various steps up to the final dissolution of the cysts. No large cysts containing cells in synap- sis have been found to break up by the inward migration of fol- licle cells. The final fate of al! of them and of their enclosed cells is degeneration. Rarely, however, a single cell of such a cyst may enter the growth phase. A cyst containing such a cell is shown in figure 64. This contains a large oocyte among the smaller cells of the cyst. Whether such oocytes continue to grow and form functional ova is not known, but it is certain that nearly all of the growing oocytes are derived from isolated germ cells or from the breaking up of small cysts. Other sec- tions, showing both cysts and oocytes in about equal numbers, are given in figures 70 and 71 from larvae 55 mm. and 62 mm. long, respectively. In the middle of the period of sex differentiation, therefore, there are in the same gland undifferentiated germ cells that are still dividing; germ cells in the various prophases of heterotypic division; cells that have entered the period of growth, forming recognizable oocytes, and cells undergoing degenerative changes. The gland described above represents only one type of germ gland during this period—a type in which the number of growing oocytes and cysts is approximately the same. There are other glands with very few growing oocytes, and still others in which there are very few cysts. Figure 72 is a section through the germ gland of a larva 50 mm. long in which there are very few growing oocytes, and none of these are shown in the section. The gland is filled with cysts, some of which contain undifferen- tiated germ cells, while some contain cells in the synaptic phases. It is often found that if a gland seems free from growing oocytes when examined superficially, more careful search usually reveals at least a few of them scattered about through the gland. In one such gland only one growing oocyte could be found. Figure 73 is a section of the germ gland of a larva 71 mm. long. In this larva there are only a few growing oocytes, only one of GERM-CELL HISTORY IN THE BROOK LAMPREY 69 which is shown in the section. Numerous cysts are present, some with cells in synapsis, others with degenerating cells, and still others with undifferentiated cells. Figure 74 shows a cross-section of a gland from a larve 65 mm. long. In this larva the gland contains practically only growing oocytes, although a few individual cysts are scattered throughout the gland. Such cysts may contain only actively dividing germ cells or only cells in the various stages of degeneration. Larvae with such glands quite certainly become females. Another gland of this type from a larva 63 mm. long, is shown in figure 75. Figures 76 and 77 are two sections from larvae 50 mm. and 60 mm. long, respectively. In these the number of oocytes in the gland is greater than the number of cysts. Germ cells of the various types mentioned above are distrib- uted throughout the germ glands of this period, apparently without any regularity and without any relation to one another or to the somatic parts of the gland. All of the types may occur along the periphery of the gland or in the deeper portions. Cells may enter the synaptic phases whether they lhe singly or in cysts, and irrespective of the position they occupy in the gland. Fur- thermore, no difference has been observed between the follicular cells surrounding those germ cells which have entered the synaptic phases and those enclosing the resting and dividing cells. There is, then, no indication that the somatic environment has anything to do with the initiation of the synaptic phases. The fact, how- ever, that the cells of large cysts which have entered the synaptic phase usually degenerate indicates that the environment of the cells at this time may determine whether or not they shall form growing oocytes. In none of the germ glands do all the cells that are destined to form growing oocytes enter the growth phase simultaneously, and therefore, in the same gland one may find oocytes of all sizes, as well as cells in the various stages of synapsis. Figure 78 is a section of the germ gland of a larva 72.5 mm. long, in which are oocytes of various sizes. ‘There is no special limited period during the course of early development when the germ cells show a greater tendency than at other times to change into oocytes. 70 PETER OKKELBERG The change may take place in some cells when the larvae are less than 40 mm. long; while in other larvae no oocytes are found until a much later stage, in some cases not until the larvae are 70 mm. long. Few or many growing oocytes may form in the germ land in very early stages. In most cases they are formed before the larvae are 70 mm. long, at which stage the sex glands are either predominantly male or predominantly female. A few oocytes may enter the growth phase after the larva is 70 mm. long, especially in glands that are predominantly male. After that time the cysts and undifferentiated germ cells gradually disappear in larvae destined to become females, while in those that become males the oocytes which have reached a considerable size or which may form in the gland subsequent to this time remain in the gland up to the adult stage. One such cell from an adult testis is shown in figure 63. It has been found that in the majority of adult testes such undeveloped oocytes occur. Although the caudal portion of the germ gland remains smaller and less developed than the cranial portion, yet no difference has been found in the structure of the glands in the two regions. The tendency to form oocytes seems to be equally strong in the cranial and caudal portions of the gland. So, whether the germ glands eventually become ovaries or testes, they all develop oocytes, and this is an undoubted female character. Sometimes only a few oocytes are present, and again, with the exception of a few indifferent germ cells scattered through the gland (singly or in cysts), oocytes may make up the whole of it. Hundreds of glands from young larvae have been examined to ascertain whether or not there are, in addition to the oocytes, any other sex characters which might indicate that the larvae are predetermined to form one or the other kind of sexual indi- viduals, but none have been found. The germ glands vary some- what in different regions, the result of the presence of blood- vessels and the irregular distribution of the various somatic elements, but the limits of such variation of the somatic parts of one gland are not appreciably different from those of any other gland. The only basis, therefore, for considering a larva male or female is the relative number of cysts and oocytes present in the GERM-CELL HISTORY IN THE BROOK LAMPREY ral gland. If in older larvae of this period, the number of cysts is greater than the number of oocytes, the larva probably becomes a male; if the number of growing oocytes is greater than the number of cysts, it probably becomes a female. In younger larvae of the period this cannot be a reliable criterion, for in some cases oocytes do not apparently begin to form in great numbers until the larvae are over 5 em. long. Other larvae of the same size show evidence of having formed a large number of TABLE 4 Diagram showing the relative number of cysts and oocytes in various lamprey larvae during the period of sex differentiation a 90% | 80% | 70% | 60% | 50% | 40% | 30% | 20% | 10% | 0% to to to to to to to to to to 1004 90% | 80% 70% oy Oh t0g p08 20% | 10% Cy oy Cc is! Oa, 00000 900000 c 55 00 L oF 104 | 208 30% | 40% 50% 60% | 70% | 80%] 90% to to to to to to to to to to 10% | 20% | 30% | 40%] 504] 60%] 70%] 80%] 90%] 100% gy ¥ g g g b g g g g g oocytes in their early stages, but appear later to have formed cysts only. Other larvae show about equal tendencies to form oocytes and cysts during the whole intermediate period. Sh he relative number of cysts and oocytes found in the various larvae of this period may be represented in percentages from one to a hundred as shown in the above table (table 4), which should be compared with table 2 (page 12). The circles in the table represent ova, the small dots cysts. The spaces between the vertical lines may represent individual larvae, and the relative number of cysts and ova in the glands Te PETER OKKELBERG of each of these is indicated by the dots and open circles above and below the oblique linecd. The whole diagram may represent ten larvae of any size between 35 mm..and 70 mm. in length. If a large number of larvae of all sizes between 35 mm. and 70 mm. in length were sorted according to the number of cysts and — oocytes present, approximately the same number of them would fall within each of the vertical spaces. Those to the left of the line ab would be predominantly male and those to the right pre- dominantly female, as judged by the number of cysts and oocytes present in the germ gland. Those on the extreme left would be more strongly male and those on the extreme right more strongly female than those which would fall in the groups nearer the line ab. Although the future sex of the larvae cannot be determined for a long time after oocytes have appeared in the germ gland, there is apparently developed out of these indeterminate larvae an approximately equal number of males and females. As previously stated, it is very difficult to obtain exact data on the sex ratio in the adult lamprey, since the habits of the two sexes are so different; but it has been found that out of all the adults collected from year to year and at various times of the day as well as at various times during the breeding season, there is only a slight excess of males over females, the ratio being about 118 males to 100 females. Many of the larvae, therefore, which bear the unmistakable female character of possessing oocytes must later develop into males. d. Literature on the period of sex differentiation in the lamprey. W. Miller (75) found that in larvae of Petromyzon planeri 35 mm. long the germ cells were not yet differentiated, while in larvae 50 mm. long oocytes were found. In larvae 65 mm. long he found the ovary and the testis to be fully differentiated. ~ Lubosch (’03) found in the same species that sex differentiation takes place at the end of the first year. The youngest identified ovary was from a larva 40 mm. long. Lubosch was the first to call attention to the fact that hermaph- roditism is of common occurrence in the larvae of lampreys. He examined forty-nine germ glands from larvae of Petromyzon GERM-CELL HISTORY IN THE BROOK LAMPREY (3) planeri. Among these 10.3 per cent were undifferentiated, 16.3 per cent older undifferentiated (probably young testes), 48.9 per cent true ovaries, and 24.5 per cent mixed glands. He came to the conclusion that the greater part of the hermaphrodite or mixed glands ‘‘. . . wird als mannliche Anlage gelten kén- nen, in der eine kolosale, gleichsam atavistische Anlage von Hiern stattfindet.”” He states further: ‘‘Es ist anzunehmen, dass diese Kier spiter wiihrend oder nach Metamorphose der Riickbildung anheimfallen werden.”’ I published (’?14) a summary of some work on the sex glands of the American brook lamprey, Entosphenus wilderi. Fifty larvae ranging in length from 20 to 75 mm. were examined for sex. Out of these 46 per cent were regarded as female because the germ glands contained practically nothing but growing oocytes, 10 per cent were taken to be true males on account of the absence of any growing oocytes in the germ gland, and 44 per cent were considered intermediates, for the reason that both cysts and growing oocytes were found in the glands. Since in the adult stage males and females occur in nearly equal numbers and since undeveloped oocytes were found in the adult testes, the conclusion seemed warranted that the intermediates became males. Two other instances are recorded in which oocytes were found in the mature testis of the lamprey. Beard (’93) found in the testis of a specimen of Petromyzon planeri one well-marked oocyte in an individual follicle for every forty sections of 10 » thickness. Ward (’97) describes the occurrence of a single microscopic oocyte in the testis of an adult Petromyzon. Further discussion of juvenile hermaphroditism in lampreys will be found at the close of the next section following the discus- sion of like conditions in other vertebrates. e. Other cases of juvenile hermaphroditism among vertebrates. Cyclostomes. Cunningham (’86) found that in all specimens of Myxine glutinosa with very immature eggs, the caudal portion of the sex gland had the structure of a testis. In one specimen this testicular portion showed spermatogenesis and a number of spermatozoa. In all the sex glands of specimens with well- 74. PETER OKKELBERG developed ovarian eggs there was, with one exception, no testicu- lar portion. He concluded that in the young state nearly all females were hermaphroditic and that the testicular portion of the sex gland normally disappeared as the eggs became more mature. He believed that fertilization was normally effected by these hermaphrodites, since true males were so rare. Out of hundreds of specimens examined he succeeded in identifying only eight males. , Nansen (’87) also worked on Myxine, and came to the conclu- sion that all the animals are males up to a length of 32 to 33 em., after which they change sex and become females. He regarded this as a case of protandric hermaphroditism among vertebrates. Dean (’97), as a result of his study of Bdellostoma stouti, doubted the conclusion reached by Nansen, as also did Price (’96). It was not, however, until 1904 that Schreiner proved by the exam- ination of hundreds of specimens of Myxine that hermaphroditism in this form is of a juvenile character and that each animal ma- tures only one kind of germ cells. Schreiner divided the animals examined into three groups, namely, males, females, and sterile. In the males the testes occupy the posterior portion of the gonads, while the anterior portion may not develop at all or may contain ova arrested in their development and showing signs of degen- eration. In the female the ovary occupies the anterior portion of the gonad and is well developed, while the posterior or testicu- lar portion is sterile and only slightly developed. The sterile individuals were of two kinds: those that showed neither follicles nor eggs and those in which both ova and follicles were found. Ova were usually found in the testes of the males, ranging from only one in some specimens to a large number in others. Out of hundreds of specimens examined only nineteen males were found without ova. The evidence obtained by Schreiner and later confirmed by Cole (’05) shows that Myxine is not a protandric hermaphrodite, but a juvenile hermaphrodite like the brook lamprey. ‘This same conclusion has been reached by Conel (’17) for Bdellostoma as well as for Myxine.. Conel believes that in the males the sex gland may degenerate with age, and that this accounts for the sterile individuals found by Schreiner. GERM-CELL HISTORY IN THE BROOK LAMPREY WD Teleostei. Certain species of teleosts belonging to the families Sparidae and Serranidae are said to be normally hermaphroditic, and those of the latter family are even said to be self-fertilizing (Brock, ’81; Howes, ’91). Brock found in certain species of Sparidae that some of the young were hermaphroditic and others unisexual. In the former the ovarian elements did not mature and they became males, while the latter became females. I have shown that in the young lamprey the male part of the predomi- nantly female gonad soon disappears, or is at least in many cases not very conspicuous, so that the impression may be gained by superficial examination of certain individuals that they are pure females. In the male, however, the female character, namely, the presence of oocytes, persists in many cases even up to the adult stage. For this reason one might conclude from the examination of older larvae that the males alone are hermaphro- ditic or heterozygotic as to sex. This was the conclusion that I first reported (’14). From a more careful examination of earlier stages I have now found that there is essentially no difference between the males and the females with regard to the juvenile hermaphroditic condition and that the apparent purity of the older larval females is due to the fact that the male character in the form of cell nests does not persist for any length of time after the female character has become predominant. Out of the her- maphroditic larvae, therefore, both males and females develop, and not males only as was at first supposed. In bony fishes there are no recent researches on sex differentiation, and the problem of juvenile hermaphroditism in this group needs to be reinvestigated. It may be expected to occur in both sexes. Amphibia. It has long been known that in toads anterior to the true sex gland there is an organ (Bidder’s organ) which has the structure of a rudimentary ovary. In males it persists through- out life, but in females it disappears after the second year in all forms that have been studied with the exception of Bufo vulgaris, in which, according to Ognew (’06), it is retained throughout life. Certain cells in this organ are, in structure and development similar to true oocytes. Ognew states that the boundary between Bidder’s organ and the germ gland is often quite indefinite, so 76 PETER OKKELBERG that one merges more or less into the other. In many cases oocytes occur even in the center of the testes, and Cerruti (07) also found follicles which contained spermatozoa in Bidder’s organ. With age, however, the boundary between Bidder’s organ and the testis becomes more and more definite. Pfliiger (82), one of the early workers on the sex problem in frogs, came to the conclusion that in recently metamorphosed frogs there are three kinds of individuals, namely, males, females, and hermaphrodites. During development the hermaphrodites become definitive males and females, so that in the adult condi- tion the number of males and females is about equal. Pfliiger also found that in certain races of frogs there is a greater tendency toward juvenile hermaphroditism than in others. An examina- tion of recently metamorphosed frogs collected in nature from three different geographical regions gave the following results: PER CENT PER CENT NUMBER MALES FEMALES EXAMINED IRGyaud AN Pee ehh 66 ok ORE ee aaine aaae 35.0 65.0 228 UIDIREVOLI Rs eee ate eae, tah ame a com est ie Ee conse, (oe NB.73 86.8 459 Gr SCR Ewe ante erctele)s Cato e Berens Meee 47.2 52.8 500 Collections of adult frogs from the three regions showed that the number of males and females was approximately equal. The conclusion was reached that the young are often hermaphroditic and do not reveal the final sex condition of the animal. R. Hertwig (’05, ’06, ’07) also found that young frogs showed a tendency toward juvenile hermaphroditism. In two laboratory cultures, in which all the larvae were brought beyond metamor- phosis, he found forty-three females and eighteen males in one and forty-seven females and eight males in the other. He be- lieved that the females were pure and that only the males were hermaphroditic, and this form of hermaphroditism he termed rudimentary protogynic. Schmitt-Marcel (’08) made micro- scopic examinations of the sex glands of a large number of Rana temporaria in different stages after metamorphosis and con- cluded that all the intermediates or hermaphrodites became males. Kuschakewitsch (?10) worked on Rana esculenta and GERM-CELL HISTORY IN THE BROOK LAMPREY 77 came to the conclusion that, if intermediates were found in a culture, all the individuals were intermediates and consequently some developed into males and some into females. Witschi (14) found that individuals of Rana temporaria that had just metamorphosed showed all the intermediate conditions between pure males and pure females. He believed that the testis developed out of an ovary and not from an undifferentiated gland. No evidence was found that an ovary developed out of a testis; for, usually, when a germ gland had begun to develop in the male direction, the whole or at least a part of it became male. Witschi has advanced the idea that a germ cell in a frog becomes a male or a female cell according to the length of time it remains in the germinal epithelium. He found that when eggs developed under optimum temperatures (21°C.), no hermaphrodites appeared in the cultures; but when the eggs developed under low temper- atures (10° to 15°C.) and under high temperatures (27°C.), inter- mediates were formed. Witschi concludes that heat and cold increase the chances of an early migration of some of the germ cells out of the germinal epithelium into the sex cords, and that when a part of the germ gland has thus once differentiated in the male direction, the whole gland usually changes into a testis, although the germ cells which remain in the germinal epithelium transform into oocytes. Under this assumption there would nat- urally be more males formed at either extreme of temperature. In the same way Witschi assumes that overripeness of the egg at fertilization so alters the trophic condition of the whole organism that only males are produced. Abnormal temperatures and overripeness Witchi speaks of as ‘Ausenfaktoren,’ and claims that they may influence the sex of the organism. Under normal conditions they are absent and sex is then determined by ‘Erbfaktoren’ and ‘Innenfaktoren.’ The latter he speaks of as female determining: ‘‘Wenigstens spielen mi&nnchenbestimmende Innenfaktoren keine auffillige Rolle.” These factors are local conditions in the germ gland which retard the migration of the germ cells into the sex cords. The ‘Erbfaktoren,’ however, are supposed to be the chief sex 78 PETER OKKELBERG determining factors under normal conditions. . In attempting to explain how these factor operate, Witschi has adopted the inter- pretation of Goldschmidt (12) that the female is homogametic with respect to sex (MFMEF), while the male is heterogametic (MFME’). In these formulae the letters are ranked in value as follows: F>M, F>F’, and M>F’. Goldschmidt speaks of a variation in values of the factors as ‘Potenzgraden,’ which may be represented by figures; for example, M = 40, F = 60, ete. It is assumed that these potencies may vary for different fertilized eggs and that, in order that the resulting offspring may become a male or a female which is free from the characters of the opposite sex, the dominance of one sex tendency over the other must reach a certain epistatic minimum. If it falls below this minimum, intermediates are formed. Witschi found also, as had Pfluger, that different races of frogs varied as to the number of intermediates produced. Races of Rana temporaria from northern Germany (Ké6nigsberg) and from the Alps (Ursprungtal) differentiated early and only a few or no intermediates were produced. In middle Europe (Utrecht, Munich), however, intermediates were commonly formed. Wit- schi designates these races as differentiated races and undiffer- entiated races. Considerable space has been devoted to the results obtained by Witschi, for it is about the only experimental evidence we have that an external factor may influence the resulting sex of the individual. There are other cases in which the sex of an indi- vidual appears to be reversed by factors influencing the egg before fertilization, such as overripeness of the eggs, overwork in repro- duction, desiccation of the eggs before fertilization, etc. Some of these cases will be discussed later. f. Discussion. It has not been found practicable to test sex determination in the lamprey by experimental means, similar to those employed in the case of the frog, because of the length. of the larval period and the difficulty, under laboratory condi- tions, of rearing the larvae through the period of sex determination. Witschi thinks that in the early phylogenetic history of the frogs the germ cells were probably all of the same value as to sex and GERM-CELL HISTORY IN THE BROOK LAMPREY 79 that sex was determined wholly by external factors. In the lamprey it appears that the two sex potencies are almost balanced and that under normal conditions it is a matter of chance which sex develops, so that slight changes in the environment might suflice to throw the balance in favor of one or the other sex. It cannot be denied, however, that the eggs, from the time of fertilization, may show a greater tendency in favor of one sex than the other, as has been supposed by Witschi to be the case in normally developing eggs of the frog; but this inherited tend- ency may not be strong enough to prevent the formation of a series of intermediate individuals with glands ranging from those with no oocytes to those with no cysts. Under such conditions, it is not difficult to understand how a sex reversal might take place as a result of extraordinary external conditions. The more equally balanced the sex potencies are, the more easily a sex reversal might be effected. Granting that sex potencies may be inherited factors, it does not seem necessary to assume that one sex is homozygous for sex and the other heterozygous, as Witschi and others have assumed, who felt themselves obliged to bring the phenomena of the inher- itance of sex in line with those of the inheritance of mendelian characters. Before entering upon this question further, it will be necessary to summarize briefly the morphological evidences obtained from the present study in favor of a possibility of sex reversal in the lamprey. Following an earlier indifferent period, when no sex characters are present, there is an indeterminate period in the early larval life of the lamprey during which the future sex of the individual cannot be determined, in spite of the fact that the sex characters are present. During this indeterminate period, all of the germ glands develop oocytes in greater or less number, with the exception of possibly a few in which no oocytes are found. At the same time many germ cells in all the glands remain in an indifferent condition and are found either as individual cells or in smaller or larger cysts. Since the secondary sexual char- acters do not appear until later in the life of the animal, there is during this period no other sex distinguishing character than JOURNAL OF MORPHOLOGY, VOL. 35, NO. J 80 PETER OKKELBERG the presence or absence of oocytes. The presence of large cysts of indifferent germ cells has been taken to be a male charac- ter, the whole cyst being homologous to an oocyte with its fol- licle; but to a certain extent the male character remains obscure, since it is not until the animal approaches the sexually mature condition that the male germ cells can be identified as such. The presumably male germ cells of the larger cysts continue to divide until after metamorphosis, but the cells resulting from each divi- sion are, for a long time, not essentially different from the pri- mordial germ cells. The only secondary characters that distin- guish the adult male are the long, slender urogenital papilla and the absence of an anal fin, but these do not appear until after metamorphosis. Since this study does not involve the stages in which secondary sex characters are present, we are concerned only with the primary ones, namely, the presence of male or female sex cell. Only the latter are structurally recognizable in the stages studied, and they give us the only definite clue to the sex condition of the young larvae. Oocytes appear in practically all lamprey larvae; in normally dioecious species oocytes appear only in about one-half of the young, while in the other half only male cells appear. Whether or not male germ cells occur in all lamprey larvae cannot now be stated for reasons already presented. When a germ cell, however, shows no tendency to transform into an oocyte, but continues to divide and form cysts, it has been assumed that it is poten- tially more strongly male than female in constitution. The activ- ity of the germ cells may be along either of two lines. In the one case their tendency is toward growth, and in the other toward rapid multiplication.. This difference in the activity of the cells may be due to an inherent tendency in the cells themselves or to factors operating in the cell environment. If due to the former it must be admitted that all the germ cells of the same gland do not have the same make-up and that during the process of devel- opment each cell inherits a different constitution after each divi- sion. Some of the cells are endowed with a tendency toward rapid division and others toward an early cessation of division and entrance upon a period of growth. It may be due to an unequal GERM-CELL HISTORY IN THE BROOK LAMPREY 81 partition of chromatin material during cell mitoses or to an unequal distribution of cytoplasmic material during the early divisions of the cells. Just as the germ cells are set aside from the somatic cells in early stages of development by some differ- ences in their make-up, so the germ cells may also differ among themselves in their inherited structure after each division. There is direct proof that there is an unequal distribution of material among the cells in early stages of some animals, and that certain cells are destined to form certain parts of the body. These dif- ferences appear to be cytoplasmic in most cases; for example, in the case of Cynthia, in which (Conklin, ’05) there are several kinds of organ-forming substances which are unequally distributed among the cleavage cells. In later stages, however, there is no direct proof of an unequal division of the cells; so in the case of the germ cells of the lamprey it must remain an assumption that unequal division does take place; but this would explain the two types of behavior of the germ cells. Again the germ cells might all be assumed to have the same inherited structure, and yet they may develop along different lines on account of their different local environment in the gonad. In this case the behavior of the cells would be the result of factors or circumstances acting from without. These factors may be supposed to be differences in nutrition, the presence of various enzymes and toxins, differences in pressure, and various other factors operating in the germ gland of the animal. In this case it must be assumed that the germ cells are so constituted that they can respond to environmental factors in two different ways. Under certain conditions the cells will continue to divide, under others they will stop dividing and enter upon a period of growth. As yet we know too little about the physiology of the cell to be able to decide between the two possibilities. We know that in a form like the lamprey, whether it eventually becomes a male or a female, the two kinds of cells make their appearance, some with a tendency toward rapid division and some with a tendency for growth. Now, since the latter is an undoubted female quality and the former is supposedly a male quality, practically every individual must possess both male and female potencies. That 82 PETER OKKELBERG these potencies are practically in a balanced condition is seen from the fact that both male and female sex cells appear in the majority of larvae. Sometimes a larva is inclined more strongly toward the female side and at other times it leans toward the male side. In some cases it appears that a larva may fluctuate back and forth between the two extremes until finally one or the other sex condition takes the lead and sex reversal becomes more difficult. This is indicated by the fact that the sex glands of older larvae from the period of sex differentiation (larvae 50 mm. to 70 mm. long) often show that an earlier sex condition has been replaced by that of the opposite sex. After sex has become definitely established, one or the other sex potency becomes so strong that only unusual circumstances are able to reverse the condition. The elements in the body or in the germ gland, which have specialized in the opposite direction, stop developing and either degenerate or remain in an undeveloped condition during the whole lifetime of the animal. The cysts in the developing ovary contain small cells which soon degenerate and disappear so that the larva soon becomes apparently a pure female. In ‘the developing male gland the undeveloped oocytes remain, in many cases, even up to the adult stage; but often they degenerate in early stages, so that fragments of oocytes occur in the develop- ing testes. Out of the juvenile hermaphroditic condition, there- fore, both males and females eventually emerge. The condition in the lamprey is not essentially different from that in Myxine, except that in the latter the two kinds of germ cells develop in different parts of the gonad, while in the lamprey there is no segregation of the two kinds of cells. The whole gland is in fact hermaphroditic in the lamprey while in Myxine the anterior portion of the gonad is ovarian and the posterior portion testicular. In some individuals of Myxine there is a tend- ency for the two kinds of cells to be mixed. This is especially true on the border-line between the testicular and ovarian por- tions. There is also probably no essential difference between the condition in the frog and that in the lamprey, and an explanation of the phenomena in one case should hold for the other as well. Whether or not different races of lampreys show a greater or GERM-CELL HISTORY IN THE BROOK LAMPREY 83 less tendency toward juvenile hermaphroditism is not known, and no opportunity has yet been offered for an investigation of this question. What it is that keeps a larva with hermaphroditic tendencies from developing into a functional hermaphrodite is not known. It appears that when one of the sex tendencies takes the lead, it prevents the development of the structures characteristic of the other sex; or it may be that when one set of sex elements begins to degenerate, there are removed certain influences that have previously inhibited the development of the other set. Some- thing similar to this is seen in most true hermaphrodites, where only one set of sex cells develops at a time, so that the animal is either protandrous or protogenous. In this case, too, the devel- opment of one set of germ cells is antagonistic to the develop- ment of the other, but a reversal always takes place when one group is exhausted. This may be due to the fact that certain hormones are eliminated from the germ cells which have taken the lead in development, and that these are unfavorable to the develop- ment of the other set. As soon as the first set of cells has been eliminated, a reversal takes place and the opposite set develops. It may be looked upon as alternate periods of vigor and depression as far as the particular germ cells go. In bisexual animals with juvenile hermaphroditic tendencies, it may be supposed that the animals never recover from the state of depression relative to the opposite sex. Reviewing the case of the lamprey, the evidence seems to warrant the conclusion that sex is not irrevocably fixed at the time of fertilization; that the future sex of the animal is not definitely determined until the larva has reached a considerable size, and that sex is not the result of any unchangeable sex quality present in the egg at the time of fertilization, but is rather the outcome of a balanced sex potency which results in one or the other sex being formed, largely as a matter of chance under normal environmental conditions. It is possible that one sex potency may be stronger than the other from the beginning of development, and that even the germ cells themselves at the time of fertilization may be inclined in one or the other direc- 84 PETER OKKELBERG tion; but such an admission is not necessary for an explanation of what actually takes place. Practically the same conclusion has been reached by Shull (711) in the case of plants. He says: May not maleness and femaleness be thought of as alternative states which can be crudely analogized with the acidity and alkalinity of chemical solutions. . . . In some species the sexes appear to represent a much more strongly polarized (?) condition than in other species, and a transition from the characters of the one sex to those of the other is attained only with the greatest rarity, if at all; while in other species the sex conditions may be so nearly balanced or neutral that individuals are not absolutely determined in their sex relations by their genotypic nature. . . . With such a conception of sex, it also appears probable that sex may be influenced sometimes by external factors as well as by internal ones, and in this case the preponderance of one sex over the other, which has been observed in many animals and plants, need not be attributed alone to selective disorganization of germ cells, a selective fertilization or a selective death rate, but might conceivably be controlled to a certain extent by environmental condi- tions, acting at some particular ‘sensitive’ period in the ontogeny of the organism in question (pp. 363-364). 4. Present status of the sex problem. We may now ask whether or not the view expressed above can be brought into harmony with current opinions concerning sex determination. The gen- erally accepted view is that sex is established at the time of fer- tilization as a result of the presence or absence of so-called sex chromosomes in the fertilized egg. This view was first expressed by McClung in 1902. During the progress of his work on the maturation of the germ cells in insects, he found a certain body ~ in the spermatocytes which was interpreted as being a sex-deter- mining element. This body had been seen before by Henking (91), Montgomery (’98), and Paulmier (’99), but it had not been suspected that it might be a sex-determining factor. McClung’s statement concerning the function of the accessory chromosome as it was called, was asfollows:‘‘. . . . itis the bearer of those qualities which pertain to the male organism, primary among which is the faculty of producing sex cells that have the form of spermatozoa.” This interpretation was quite generally accepted. Previous to this time numerous theories had been advanced concerning the cause of the appearance of GERM-CELL HISTORY IN THE BROOK LAMPREY 85 two kinds of sexual individuals, each theory to be replaced by others, which further research found equally untenable. In- vestigators now began to search for this odd element in the sex cells of various species of animals. More and more forms, espe- cially among insects were found in which the odd element was present in the germ cells and in which it became distributed to half of the mature cells. It is natural that this discovery should have led to a qualita- tive explanation of sex. There apparently was something present in half of the male germ cells which, after fertilization, was responsible for the development of a male. This was McClung’s interpretation and this explanation was accepted by the majority of his immediate followers. The early work was done on the accessory chromosome of the male germ cells alone. When cytologists began investigations upon the chromosomal structure of the female germ cells (Wil- son, 05; Stevens, ’05, and others) it was found that this odd ele- ment was present there also, not singly but in duplicate. These two accessories were so distributed during maturation that every egg retained one, and consequently all the eggs were alike in their chromosomal structure. Theoretically, therefore, an egg which happened to be fertilized by a spermatozoon containing the accessory chromosome would give rise to a female, and not to a male as had been supposed to be the case. It became clear that the accessory chromosome could not be sex determining by virtue of any qualities it might possess, but rather that sex was due to a quantitative difference in the amount of the odd chromosomal material present in the fertilized egg. Certain studies in heredity have shown that some characters are sex-linked. The interpretation of this fact is that the factors for such characters are carried by the sex chromosome. It was discovered that the inheritance of sex-linked characters in forms like moths, butterflies, and birds was such as to necessitate the assumption that the ova in these forms rather than the sperma- tozoa were dimorphic in regard to the sex chromosome. Later it was discovered by Seiler (’14) that there are actually two kinds of eggs in the moth Phragmatobia fuliginosa. In the case of 86 PETER OKKELBERG birds the problem has not yet been cleared up. Guyer thinks he has evidence that the spermatozoa are dimorphic in this form, while the inheritance of sex-linked characters in birds points to the egg as being dimorphic. In his last paper on the subject, Guyer (’16) again emphasizes the fact of the presence of two kinds of spermatozoa in the common fowl, but admits the pos- sibility of only one kind being functional. If it be admitted that the eggs also are dimorphic, it would be difficult to explain why two kinds of cells should be produced in both sexes of the offspring. The assumption of a dimorphism of both spermatozoa and ova of the same species has been made before. Castle (’03) proposed a theory of this sort. Such a theory necessitates the further assumption of selective fertilization, for which there is apparently no direct evidence. In a recent paper by Stockard and Papanicolaou (716), dealing with the hereditary transmission of degeneracy and deformities in aleoholized guinea-pigs, a statement is made that the junior author is in possession of data which indicates that the female guinea-pig, as well as the male, shares in the determination of sex, and that in this species both ova and spermatozoa may be dimorphic. Previous to this, Papanicolaou published some of his results in Science (715), where he states that the sex of the guinea- pig is determined, sometimes by two and sometimes by three factors, depending upon whether or not the mother had previ- ously given birth to young. The three factors are: 1) The sex tendency of the father; 2) the sex tendency of the mother; 3) the change of sex tendencies in the female from litter to litter. If these observations prove to be correct, the sex potency of the fertilized egg is not determined by a sex chromosome, unless there be a selective fertilization that is subject to variation ac- cording to the physiological condition of the parent. An accessory chromosome has not been found in all forms studied. This does not, however, exclude the possibility of its being present, since it appears that it may often be united with some other chromosome. This seems to be the case in Ascaris megalocephala among invertebrates and Necturus maculosus GERM-CELL HISTORY IN THE BROOK LAMPREY 87 among vertebrates. If it occupy such a position, it might very easily escape observation. In plants no accessory chromosome has been found, except in Salamonia biflora in which Cardiff (06) describes one, but his interpretation has been doubted by Strasburger and others. It is in insects that the accessory chro- mosome has been studied with most care, and in this group it has also been found that the secondary sexual characters appar- ently develop independently of the sex glands. It has been shown by Kellogg (04), Meisenheimer (’09), Kopec (11), and Steche (12) that in moths, at least, the presence of a particular germ gland in the animal is not responsible for the development of the secondary sexual characters. In this case the primary and secondary sexual characters seem to develop in consequence of the presence in the developing embryo of a common factor or.set of factors which may be located in sex chromosomes. Many of the bodies that have been described as accessory or sex chromosomes are probably something else. Our knowledge concerning many of the cytoplasmic bodies in the cell is very limited, but it is known that some of them may occur among the chromosomes during mitosis. It will be recalled that Wilson (713) in his work on Pentatoma warned against mistaking a so-called chromatoid body in certain cells for a sex chromosome. Wod- sadelek (714) found a similar body in sex cells of the horse, and Bachhuber (’16) found it in the rabbit. There also seems to be a certain relation between the nucleolus of the cell and the acces- sory chromosome in certain cases. Goldsmith (’16) thinks he has found evidence in Pselliodes cinctus that the nucleolus is composed of both chromatic and achromatic material. The achromatic material he thinks is linin, or closely related to it in composition, while the chromatic part constitutes the sex chromosome. It must be admitted, however, that an accessory chromosome is undoubtedly present in the cells of a great number of forms, and that it may function as a sex determiner, at least in the absence of other factors. There is, however, good reason for believing that it forms only one link in a series of events that precede the development of sex. This conclusion has also been 88 PETER OKKELBERG reached by Doncaster (714) who says: ‘‘It seems evident that sex cannot depend on a chromosome alone for the chromosome must act by its relation with the cell-protoplasm and it is on this relation that sex determination depends.” This same proposi- tion is admitted by Morgan (’15). He says: “‘It is quite con- ceivable that one or more of these other factors might so change that the sex differentiation would become inoperative or even change so that these other factors themselves become the diferen- tiators that determine sex” (p. 95). He admits that the environ- ment is one of the important factors that enters into the develop- ment of every individual and that it is quite possible that it may turn the scale and determine sex. Loeb (’16) accepts the cytologi- calevidences for sex determination by sex chromosomes, but speaks also of a physiological basis of sex determination by specific sub- stances or internal secretions. He thinks it possible that the sex chromosomes may favor the formation of specific internal secretions which are responsible for the formation of sex charac- ters in the animal and that if it should be found ‘‘possible to modify secretions by outside conditions or to feed the body with certain as yet unknown specific substances the influence of the sex chromosome upon the determination of sex may be overcome” (p: 228). From these statements it will be seen that the possibility is admitted by some of the foremost investigators of the sex prob- lem that all germ cells carry the potentialities of both male and female, and that after fertilization the egg may be inclined in one or the other direction, but not so strongly that it excludes the possibility of a reversal in the other direction. ‘There seems to be at the present time a decided tendency away from the idea that the sex chromosomes carry absolute sex determiners. We are, therefore, no longer antagonistic to the idea that other sex factors may exist, either in the cell itself, in the developing organism which comes from the germ cell, or in the environment of the cell or organism. 5. Discussion of the hermaphroditic condition found in the lam- prey in connection with other sex phenomena, not easily explained by current theories. a. Normal hermaphroditism. The sex- GERM-CELL HISTORY IN THE BROOK LAMPREY 89 chromosome hypothesis does not offer a satisfactory explanation of normal hermaphroditism in animals and plants. Hermaphro- dites normally produce both male and female sex cells in the same individual, but in most cases the two kinds of cells are not ma- tured at the same time. Usually the male germ cells are rip- ened first, and in such cases the species is known as protandrie. When the eggs are matured first, the species is known as proto- genic, and when the male and female germ cells are produced at the same time, the condition is known as simultaneous hermaphro- ditism. ‘The latter condition is usually found in species with more or less widely separated male and female sex glands, but it may also appear in species in which an ovotestis is found, as, for example, in certain pulmonates. When the germ cells are ripened during successive seasons of the life-cycle of the animal, the condition may be called polycyclic. On the other hand, if the animal produces only one kind of germ cells during the early period of its life and the other kind of germ cells during the later period, the condition may be termed monocyclic hermaphrodit- ism. ‘The latter condition exists in Crepidula fornicata. Orton (09) has made a study of this form and has found that the indi- viduals associated in chains offer transitional series from maleness to femaleness both in primary and secondary sexual characters, beginning with a male in the young stage and ending with a female in the older stages. Three hundred and fifty chains were examined, and it was found that the individuals could be arranged as follows: 1) male; 2) male with rudimentary uterus; 3) hermaphrodite with small uterus; 4) hermaphrodite; 5) her- maphrodite with small penis; 6) female with rudimentary penis; 7) female. In monocyclic hermaphroditism it appears that with the aging of the animal its metabolism becomes antagonistic to the devel- opment of one or the other of the two kinds of sex cells. In the case of Crepidula, the metabolism of the young animal is favor- able to the development of the male sex cells, while the metabol- ism of the older animal is more favorable to the development of the female sex cells. In the case of polycyclic hermaphrodites, when the two kinds of germ cells are ripened in close succession, 90 PETER OKKELBERG the development of the second kind of germ cells may be the result of a changed metabolism. In some true hermaphrodites sex conditions seem to be dis- turbed at times so that true males and females appear. Accord- ing to Maupas (’00), the number of males per thousand females in various nematode worms may vary from 0.13 to 45. 6. Alternation of the hermaphroditic and the dioecious con- dition. A more’ complex case of hermaphroditism than those mentioned above is that found in the nematode worm Rhabdites (Rhabdonema) nigrovenosum, which is parasitic in the lungs of frogs. While in the lungs, the worms are hermaphroditic, but in the free-living state, which alternates with the parasitic, two sexes occur. The free-living worms again give rise to herma- phroditic parasitic offspring. This has been explained by Boveri (11) and Schleip (11) as being due to the disappearance of one kind of spermatozoa in the free-living males, so that upon fer- tilization only one kind of sexual individual is produced, namely, a female which again becomes parasitic. This female is capable of giving rise to both spermatozoa and eggs, both of which should have the same chromosomal make-up. During maturation three kinds of spermatids are produced, with five, six, and seven chromosomes, respectively. It is supposed that the last kind degenerates. Some evidence has been advanced for a chromosomal explana- tion of true hermaphroditism. Zarnik (11) thinks that in cer- tain hermaphroditic Pteropods, the female cells are of one kind only (homogametic) ; while the male cells are of two kinds (hetero- gametic), but that only one kind of male cells is functional, namely, the one corresponding in chromosomal make-up to that of the female cells. The offspring from such union should result in a female, but instead it develops into a hermaphrodite in which again half of the male germ cells degenerate. Kriiger (’12) found what she thinks is an accessory chromosome in the hermaphrodite Rhabdites aberrans. During spermato- genesis it becomes distributed equally among the spermatozoa, with the exception of a very few cases when it lags behind and is retained in one cell. Apparently in this species the spermatozoon GERM-CELL HISTORY IN THE BROOK LAMPREY 91 simply initiates development in the egg and the sperm nucleus degenerates. The parthenogenetically developing egg forms a hermaphrodite. Only in one case was a fusion of the male and the female nucleus observed, and Miss Kriiger assumed that the male nucleus in this case was one in which the accessory chro- mosome was lacking. This would give rise to a male of which there were a few formed. Demoll (712) thinks that in Helix pomatia, which is hermaphro- ditic, two kinds of spermatozoa result from the unequal distri- bution of the sex chromosome and that only one kind, that with the accessory, becomes functional. It appears from the above cases, namely, that of the gastro- pods, which are true hermaphrodites, and that of Rhabdites, in which the hermaphroditic condition alternates with the dioe- cious, that sex cannot be the result of the action of the sex chro- mosome alone, but that the activities and behavior of the sex chromosomes which results in their peculiar distribution must be due to some physiological activity in the cell which antedates the sex chromosomes, so that the latter are simply the final link in a series of processes which determine the sex potentiality of the cell. This possibility has been admitted by Schleip in the case of Rhabdites. He says: ‘‘Es scheint, dass die Entwick- lung mancher Keimzellen zu Spermatozyten statt zu Ovozyten zum Teil auf Ursachen beruht, die ausserhalb dieser Keimzellen liegen.” . . . ‘‘Diese fusseren Ursachen brauchen nicht ausserhalb desselben befinden; man kann sogar vielleicht daran denken, dass innere Secretion dabei eine Rolle spielt’’ (p. 128). Further on Schleip adds that external conditions may influence the development of the sex cells. He says: ‘‘Wie bei manchen Tieren dussere Bedingungen einen Hinfluss auf das Geschlecht der sich entwickelnden Tiere auszuiiben imstande zu sein schienen, so beeinflussen also dussere Bedingungen bei der zwittrigen Genera- tionen die Entwicklungsrichtung der Keimzellen.” He says further: “‘ Daher wird die Frage erlaubt sein, ob die verschiedene Chromosomenzahl tiberhaubt einen Einfluss auf die Gesch- lechtsbestimmung hat, und ob die Spermien nicht aus anderen Ursachen und in anderer Weise in minnliche und weibliche dif- 92 PETER OKKELBERG ferenziert sind und die verschiedenen Chromosomenzahl, die sie erhalten, nur die Folge davon ist.” Another line of investigation on hermaphroditism is the study of the segregation of the germ cells in the sex gland. In the case of Sagitta, Elpatewsky (’09, ’10) finds a body in the cyto- plasm of the cells during early cleavages, which he calls the ‘beson- dere Korper.’ This is retained by only one cell after each cleav- age up to the sixth, when it divides during mitosis and part of it. passes to each of the resulting cells. These two cells become the germ cells, and Elpatewsky believes that one becomes the fore- runner of the spermatozoa and the other of ova, and that the former gets a larger portion of the ‘besondere KG6rper.’ Ancel (03) has worked on the early development of the germ cells in Helix pomatia and thinks that three kinds of cells appear in the germ gland, spermatozoa, oocytes, and nurse cells. He thinks that the primordial germ cells become transformed into female and male elements according to whether or not the nurse cells are present at the time of transformation. Buresch (711) thinks that in Helix arbustorum, also, the fate of the mdifferent germ cell depends on its proximity to a nurse cell. The cases which have been cited, show that the differentiation of the germ cells into male and female cells has been interpreted as being due to nuclear differences in some cases, cytoplasmic differences in others, and to differences in the environment of the cells in still other cases. If we conceive of sex as a metabolic state rather than the result of definite sex factors, it is easy to see how any one of the above factors might result in a metabolic change which would throw the balance in favor of one or the other sex. It is unfortunate that so little work has been done on the history of the germ cells in hermaphroditic animals; for it is in these forms that one undoubtedly must look for valuable clues to the problem of sex determination. c. The effect of delayed fertilization on sex. An interesting case in which sex metabolism seems to be disturbed by outside factors is that of the frog in which delayed fertilization results in the development of the eggs into male individuals exclusively. It was found both by R. Hertwig (’05. ’06, ’07), and by Kuscha- GERM-CELL HISTORY IN THE BROOK LAMPREY 93 kewitsch (10) that the percentage of males increased with the length of time that fertilization was delayed. Kuschakewitsch found that when fertilization was delayed as much as eighty- nine hours all of the eggs developed into males. The mortality among all the eggs in the culture was about 4 per cent. In the ease of the frog, an accessory chromosome has been described, both by Levy (15) and by Swingle (17). Levy found twenty- five chromosomes in the male germ cells of Rana esculenta. During maturation division these were so distributed that half of the cells received twelve and the other half thirteen chromo- somes. The odd chromosome of the thirteen is the sex chromo- some. Levy believes that the accessory chromosome undoubt- edly has something to do with sex, but he thinks that it is not the only sex-determining factor. He says: ‘‘Man darf aber die Geschlechtschromosomen nicht als den geschlechtsbestimmenden Faktor bezeichnen, den sie sind nur die zuerst morphologisch erkennbaren Zeugen einer stattgefundenen sexuellen Differen- zierung.”’ | Swingle found the spermatogonial number of chromosomes to be twenty-five in Rana pipiens. He found some cases in which the sex chromosome divided during the second spermatocyte divi- sion instead of during the first, and one case in which the two parts of the X-body were unequal in size. He thinks that there may be some connection between the abnormality of chromatin distribution which results, presumably, in the production of three kinds of spermatozoa, and the fact that in certain strains of the species, males, females, and individuals possessing marked hermaphroditic tendencies occur. In the case of the frog it seems evident, both from the experi- ments of Hertwig and Kuschakewitsch on delayed fertilization and from those of Witschi (’14) on the effects of temperature on the sex of the animal, that the accessory chromosomes, known to be present, are not the sole sex determiners. Such a conclusion is not a condemnation of the sex-chromosome theory. If other factors also affect the sex of an individual, it shows that the sex chromosome is but one of many such factors which may bring about the same result. Temperature, for instance, may result 94. PETER OKKELBERG in reactions in the protoplasm of the cells or may so change the whole metabolism of the organism that the visible results might be quite different. An analogy may be drawn between the phe- nomena of sex and those of the red-flowered Primula which, ac- cording to Klebs (’03) becomes white when grown at high tempera- tures. In this case the two color potencies are present in the organism, and which one shall appear depends upon an external factor, namely, temperature. Similarly, sex potencies may assert themselves differently under various conditions, so that a reversal may take place, or intermediates be formed, such as are found in cyclostomes and amphibians, and possibly in many fishes as well. d. Hermaphroditism and sex reversal due to external condi- tions. Another interesting case showing the double sex poten- tiality of early larvae is recorded by Baltzer (14). He found in Bonellia viridis, the males of which live parasitic upon the females, that if the larvae have a chance to attach themselves to a female they become males, and if they do not succeed in becoming attached they form females. If they are allowed to attach themselves and are later removed, they become hermaph- rodites. In the attached larvae the sex determining substances are undoubtedly taken up from the host, since the female-deter- mining: substance seems to be stronger in the free-living state. Baltzer concludes that sex is partly predetermined and partly epigenetic, and that both sex tendencies are inherited but in dif- ferent degrees. He believes that the male tendency is stronger than the female. If this be so, we have here a case of sex rever- sal, providing the larva remains unattached. A case somewhat similar to that of Bonellia is that of the protandric hermaphrodite Crepidula plana, in which Gould (’17) finds that the develop- ment of the male phase is dependent upon the presence of a larger individual of the same species, but not necessarily a female. In the absence of a larger individual, the larva develops into a female, but the process of transformation in the female direction may be halted at any time, up to the period of formation of grow- ing oocytes, by bringing the animal into proximity with an older individual. Gould does not offer any explanation as to the nature GERM-CELL HISTORY IN THE BROOK LAMPREY 95 of the stimulus exercised by the older individual over the sex of the larva. Among plants there seem to be many cases which indicate that every individual possesses a double sex potentiality. It appears that in some of the lower types of plants which are nor- mally dioecious, the organs of the opposite sex can be made to appear on all the individuals under proper culture conditions; that is, the male plant will produce female organs and vice versa. Bordage (’98) cut back the apex of young male plants of Carica papaya just before the appearance of the first male flowers. Lat- eral branches arose below the cut, and these produced female flowers and fruit. Strasburger (00) found that the smut Ustil- ago violacea caused the dioecious plant Melandryum album to produce the opposite sex organs; that is, the male organs appeared on the female. The pistils remained undeveloped, while the normally rudimentary anthers grew large and produced pollen mother cells. Later, Strasburger ('09) came to the conclusion, from a consideration of many evidences, that sex determination in plants cannot be the result of mendelian segregation. He says: “Teh bin nach alledem der Ansicht dass alle Versuche, die Ge- schlechtsbestimmung getrenntgeschlechtlicher Organismen auf Mendelische Spaltungsregeln zuruckzufiihren, erfolglos bleiben werden” (p. 17). Strasburger also did not consider the so-called sex chromosomes as true chromosomes. He says: ‘‘Denn nicht nur zeigen sie ein eingeartigen Verhalten, sondern auch ihre Beseitigung aus den Geschlechtszellen ist mdéglich, was fiir Trager von Erbeinheiten nicht zulissig wiire.”’ Should they be proved to stand in some relation to sex, they might yet be indi- vidual linin bodies ‘‘die aber nicht Pangene fiihren, sondern der Aufnahme des tiber das Geschlecht bestimmenden Stoffes dienen’”’ (p: 22). It may be suggested in this connection that other bodies are present in the cell which may become unequally divided during mitosis. This is true of plasmosomes, which may not always dis- solve and become diffused throughout the cell before division. It may be equally true of mitochondria and other cytoplasmic bodies. Such an hypothesis has been advanced by Schaudin F JOURNAL OF MORPHOLOGY, VOL. 35, No. 1 96 PETER OKKELBERG (05) in the case of Protozoa. A normally functioning cell is regarded by him as a hermaphrodite which has the male and female qualities equally balanced. The differentiation which leads to the formation of gametes is due to inequalities of cell division which result in a more or less imperfect distribution of the qualities of the parent cell between the daughter cells, so that some cells may receive more male and others more female properties. The male cells show greater kinetic energy; the female cells greater trophic energy. The opposite tendencies accumulate in different cells which thus become one-sided in their vital activities. ‘The want of balance may reach a stage in which syngamy must take place or the cell dies. A similar idea was advanced above, in my discussion of the appearance of two kinds of germ cells in the sex glands of the lamprey. In this case, too, the development of the two kinds of cells in the same gland may be due to a disturbance in the metabolism of the cell during mitosis, which results in the develop- ment of a cell along either one or the other of two potential lines. It is conceivable also that there may be various grades of male and female potentialities in the germ cells thus formed, and that even in their mature condition some cells may be more strongly sexed than others. After fertilization, the same differences of sex potentialities may exist, and, in so far as no other factors are introduced to disturb the relative sex potentialities, the sex of the resultant animal may be said to be determined at the time of fertilization. Whether or not these differences in sex potentiality are the result of a variation in the chromosomal make-up of the cells is not certain. This suggestion appears contrary to certain known facts of sex-linked inheritance, which seem to require for their interpretation that the sex characters reside in the same chromo- some as the sex-linked character. It might be assumed equally well, however, that certain characters appear, only when asso- ciated with a certain kind of cell metabolism which may be pecu- liar to one or the other sex. This conception might also account for the exceptions to the inheritance of sex-linked characters which are difficult to explain by the chromosomal theory. GERM-CELL HISTORY IN THE BROOK LAMPREY 97 Some further examples of mixed sex among plants may be given. Among the flowering plants some species are hermaphro- ditic, others dioecious, and still others produce three kinds of individuals, namely, males, females, and hermaphrodites, as, for example, the sweet pea. In the strawberry three kinds of flow- ers are produced, staminate, pistillate, and perfect. Valleau (’16) has investigated the inheritance of sex in grapes, and his results are as follows: The wild grape develops two kinds of individuals, staminate and pistillate, and both possess flowers of the opposite sex In a suppressed condition. The grape, therefore, occupies an intermediate position between purely dioecious plants, lke the willow, and purely monoecious plants, like the apple. On individ- ual plants of the grape all gradations are sometimes found, from staminate to functionally hermaphroditic flowers, and sometimes only hermaphroditic flowers are produced. Certain clusters of the vine may be entirely staminate, while other clusters on the same vine contain all gradations from staminate to functionally perfect flowers. In the grape, therefore, both staminate and _pistillate vines carry the determiners for femaleness and maleness, respec- tively, but with one or the other partially suppressed. Valleau draws the conclusion that, if the chromosomes carry the deter- miners for sex, then in hermaphroditic plants the determiners for maleness and femaleness must be carried in the same chromosome. There are two possibilities, therefore, for the origin of functional hermaphrodites. The maleness may express itself fully in one of the chromosomes bearing the determiners for femaleness in a pistillate plant and femaleness may express itself similarly in staminate plants. Pritchard (716) discusses the change of sex in hemp. Hemp is dioecious, and the female plant is distinguished by its dense foli- age as well as by the production of female flowers. The male plants have very scanty foliage. The sex ratio is normally 1:1. Hermaphroditic individuals appear in small numbers, but they are of the female type and predominantly female in flower devel- opment. Disturbances in the plant’s physiological equilibrium were induced by the removal of flowers and of vegetative parts, as well as by the injection of various chemicals into the stem. 98 PETER OKKELBERG It was found that sex was alterable by removal of flowers. Re- moval of female flowers caused staminate flowers to appear, and the removal of staminate flowers resulted in the development of female flowers. Pritchard believes that the change is probably due to disturbances in nutrition. He concludes that maleness and femaleness are not always fixed characters, but frequently appear more like responses of the developing organism to external stimuli. He thinks that facts do not support the theory that sex is wholly a matter of zygotic constitution, but indicate that both males and females are partially hermaphroditic. Certain plants which, under normal conditions are true her- maphrodites, will, under other conditions, produce two kinds of sexual individuals. This is true of certain mosses and ferns which normally produce antheridia and archegonia on the same plant, but which, by being supplied with a certain kind of nourish- ment will produce only one or the other of the two kinds of germ cells. Again, it has been found that under certain conditions some dioecious plants may become monoecious. Wuist (’13) found that Onoclea struthiopteris, which is normally dioecious, could be induced to become monoecious under proper culture conditions, so that the male plant preduced female organs and the female plant produced male organs. Here, again, the appear- ance of the organs of the opposite sex is apparently due to the’ nutritional environment. e. Hermaphroditism as a result of hybridization. During the last few years some interesting facts have been brought out in connection with hybridization in animals and these seem to throw some light upon the sex problem. Goldschmidt (’16, ’17) found that by crossing European and Japanese races of the gypsy- moth many so-called gynandromorphs were produced. Different results were obtained if the material had a different race origin. The explanation of this seemed to be that the potency of the sex factors differed in different races. It will be seen that this case is somewhat similar to that of the frog, in which Hertwig and Witschi found a racial difference as regards the tendency toward juvenile hermaphroditism; but in the latter case the hermaphro- ditic condition was not retained up to the adult stage. GERM-CELL HISTORY IN THE BROOK LAMPREY 99 It was supposed by Goldschmidt that the sex potency varied with the geographical distribution of the moth, and for this reason it was decided to study the behavior of different local forms of the Japanese moths crossed inter se and with European moths. The result was that a great number of individuals were obtained, which, for the various crosses, showed all intermediate conditions between true males and true females; consequently, if maleness and femaleness are represented as the end points of a series, say one as zero and the other as one hundred, a given moth might be represented by twelve, thirty-five, forty-two, ete. These ani- mals do not represent a mixture of the primary and secondary characters of the two sexes, but a definite point between the two extremes, maleness and femaleness. Since the term gynandro- morphism applies only to individuals showing a mosaic of char- acters of both sexes, Goldschmidt discards this term; for in the moths the entire individual represents a definite quantitatively fixed point intermediate between the two sexes, and not a mixture of the characters of both sexes. Such sex intermediates he calls intersexes—female intersexes, if they are genetically female, but transferred to some stage toward maleness, and male intersexes, if they are genetically male, but transferred to some point in the opposite direction. Goldschmidt has succeeded in breeding every step from a normal female through the different intersexes to a normal male; also the steps starting with the normal males and passing through the male intersexes toward the female up to three-fourths of the way. Every single step can be produced by the right combination of races. The change in any given direction is through the secondary characters first and the prim- ary characters last. The explanation of the above condition appears to be that each sex possesses the potentiality of the other. In both sexes, irre- spective of the zygotic constitution, both anlagen may become patent; which one shall appear depends entirely upon the quan- titative relation between the two potentialities. Applying sym- bols and recognizing that the female is heterozygous for sex in moths, Goldschmidt makes use of the following formulae: FFMm = Female, FFMM = Male. The value of the sex factors 100 PETER OKKELBERG he speaks of as a potency or valency. Now, it may be assumed that in a certain case the female factorial set, FF, has a value of 80 units, and the male factor, M, a value of 60 units. The form- EP Mims Tae FF MM 80 60 * 80 60+ 60 = Male. In the first formula the female set overpowers the male set by twenty units, and in the second formula the male set overpowers the female set by forty. According to Gold- schmidt, two possibilities are open. Either the slightest prepon- derance of one over the other, say only one unit, is sufficient to determine the male or the female sex, or there is a necessary minimum of preponderance beyond which only one or the other sex appears. This minimum he speaks of as the epistatic mini- mum. If the epistatic minimum be twenty; then when FF —M is greater than twenty a female is produced, while if MM —FF is greater than twenty then a maleistheresult. The intermediate points represent the intersexes and, if they are heterozygous for M, they are intersexual females, but if they are homozygous for M, they are intersexual males. Definite races possess special potencies for the male sex factors. A cross of races of similar potencies gives normal offspring. Races of different potencies of the male factors give female intersexes in the F; generation if the mother belongs to a race of lower potency. The degree of intersexuality depends upon the differences in the potencies. Another interesting case which seems to show that sex may be disturbed by hydridization is that of the Norway rat when hybridized with the albino rat. King and Stotsenburg (’15) found a great excess of males among hybrid rats and came to the conclusion “. that hybridization alters the sex ratio by producing a marked increase in the relative proportion of males” (p. 110). Detlefson (14) on the other hand, found a marked preponderance of females, especially in the early hybrid genera- tions of the wild Brazilian guinea-pig and the common domestic guinea-pig. Riddle (16, and others) has conducted an important series of experiments on sex behavior in crosses between the various races of domestic pigeons. This work was begun by Professor ulae would then read as follows: GERM-CELL HISTORY IN THE BROOK LAMPREY 101 Whitman and, since his death, the experiments have been con- tinued by Doctor Riddle. Whitman found that, if certain dis- tantly related pigeons were mated, for example, individuals of different families, only male offspring resulted. If matings were made of individuals not so distinctly related, as, for example, between different genera, and to this was added the element of overwork in reproduction, males only were produced in the early part of the season and females only in the later part of the season. _ He also observed that at the transition period during the summer some pairs of eggs produced males and females, the first usually male and the second female. It was noticed, further that toward the end of the season the eggs were not quite able to hatch, and produced embryos of fewer and fewer days’ development. This led Whitman to conclude that the developmental energy is greatest in the male-producing season. = __ Riddle, in a long series of experiments, has been able to verify the results obtained by Professor Whitman. He has also dis- covered many more facts which tend to show that in pigeons there is a reversal of sex, and that under certain conditions male offspring are hatched from normally female-producing germ cells, and vice versa. In birds there should be, according to evidence obtained from experimental breeding, two kinds of eggs; one maleproducing the other femaleproducing. These two kinds should normally be produced in equal numbers. Riddle does not deny the exist- ence of a chromosomal difference in the eggs of birds. He admits that it has been definitely shown that in some species, at least, when bred under stable conditions, certain chromosomes are associated with sex; but he denies that the sole cause of sex lies in the sex chromosome and that sex is definitely fixed and non-reversible from the very beginning of development. Data — collected, he says, “‘strongly indicates that the basis of sex is a fluid, reversible process; that the basis of adult sexual difference is a quantitative rather than a qualitative thing.”’ In pigeons, therefore, it has been shown that eggs which nor- mally develop into males or into females can have their develop- mental energy so changed by the introduction of spermatozoa 102 PETER OKKELBERG from another species, or through overwork of the parent in repro- duction, that they produce individuals of the opposite sex. Some- times this sex reversal is not absolutely complete, for many of the. females showed different grades of masculinity in their sex be- havior. Females hatched from eggs laid earlier in the season were more masculine in their behavior than those of their own full sisters hatched later in the season; and a female hatched from the first egg of a clutch was more masculine than her sister hatched from the second egg of the clutch. Here there is, there- fore, a second form of intersexualism which does not show in the primary sex characters, but in the sex behavior. f. Sporadic hermaphroditism. Banta (’16) has published some observations on the appearance of sex intergrades in the partheno- genetic Phyllopod, Simocephalus vetulus. The culture was started from material collected in an outdoor pond and the propa- gation was continued in the laboratory. During the 131st gen- eration of parthenogenetic offspring, one of the strains suddenly produced a large percentage of males, together with some normal females, and a large number of sex intergrades. These inter- grades were either males, with one or more female secondary sex characters, or females, with one to several secondary male characters, together with some individuals which had hermaphro- ditic sex glands and showed various combinations of male and female secondary sex characters. The sex intergrades are of all possible sorts of combinations of secondary and primary sex characters. The highly male-like female intergrades produced few or no young, and males with one or more female secondary sex characters in nearly every case had incompletely developed reproductive organs. Banta succeeded in propagating female intergrades for sixteen generations with no apparent change in the ratio of the various forms and with no apparent tendency of the stock to lose vigor or become less prolific. An attempt has been made to classify the intergrades on the basis of sex characters, and no less than twenty classes are distinguished. At the ends of the scale are the normal males and females. . GERM-CELL HISTORY IN THE BROOK LAMPREY 103 Banta draws the conclusion from his observations that sex depends on environmental factors which influence the general physiological whole of the organism. In the intergrades the sexual balance has in some way been disturbed and the origin of this disturbance he considers a mutation. g. Hermaphroditism as a result of hormone action. Another interesting case of disturbed sexual condition is found in the so- called free-martin in cattle. Frank Lillie (16, 17) has made a study of this problem. Forty-one cases of twins were examined in utero and a classification made of them without a possibility of error. In fourteen cases both members were males, in six cases both were females, and in twenty-one cases the two were of opposite sex; 97.5 per cent were monochorial, but, in spite of this, nearly all were dizygotic as determined by the number of cor- pora lutea present. It was found that twins in cattle are nearly always the result of fertilization of an ovum from each ovary. As development proceeds, the developing embryos sink down into the median portion of the uterus and the blood-vessels anas- tomose in the chorion, so that it is possible to inject the blood- vessels of either foetus from the other. If both of the embryos are of the same sex, no harm results from the continuity of their circulations; but if of different sex, the reproductive system of the female is largely suppressed and certain male organs are devel- oped. This is interpreted as a case of hormone action which may be due to a more precocious development of the male hor- mone or to its natural dominance. In this case, therefore, a dis- tinction can be made between the effects of the sex-determining factors that are zygotic and those due to hormones. But the sex reversal is not complete and the result is the development of an intersexual individual. h. Hermaphroditism as the result of parasitism. That a par- tial reversal of sex may be induced by parasitism has been ob- served by Geoffrey Smith (710) in the case of the spider-crab, Inachus, when infected with the rhizocephalan, Sacculina. The males, under the influence of the parasite, are capable of assum- ing all the female secondary sex characters, and often even develop ova in the testes. In this case, however, the females do not seem 104 PETER OKKELBERG to develop toward the male line when infected with the parasite. The explanation offered by Smith is that the parasite causes the host to elaborate a yolk substance similar to that which is elab- orated in the ovaries during growth of the eggs. The apparent change of sex, therefore, is due to a change in the metabolism of the organism. It is clear that such a change could take place without the assumption that one sex is heterozygous for sex and the other homozygous, as Smith has assumed. If the change in the sexual condition be due to a change of metabolism in the direction that he has suggested, it follows that only the males should take on the characters of the opposite sex. 7. Sex in parthenogenetic animals. The determination of sex in parthenogenetic animals has been studied by various investi- gators. Woltereck (’11), who worked on Daphnia, came to the conclusion that there are, in each egg, competing sex substances, one kind becoming active at the maturation of the egg, while the other remains latent. In summing up his results he says: Die resultate meiner Versuche lassen sich nur verstehen, wenn wir in jedem Ei verschiedene konkurrierende Geschlechtssubstanzen annehmen, von denen jedesmal die eine aktiviert wird, wihrend die andere latent bleibt . . . . die Geschlechtssubstanzen selbst kénnen wir uns unter dem Bilde von (latenten) Profermenten und (aktivierten) Fermenten vorstellen. In the rotifer Hydatina senta, A. F. Shull (’12) found that it is decided, in the growth period of the parthenogenetic egg from which the female hatches, whether it is to be a female-producer or a male-producer, or, in other words, that sex is determined a generation in advance. In some later experiments upon this form it has been found by Shull and Ladoff (16) that an impor- tant factor involved in the production of male-producers is the amount of oxygen present in the culture, and that this probably acts by increasing the rate of the physiological processes taking place in the body. This conclusion is analogous to that arrived at by Riddle in connection with his experiments on sex in pigeons. Riddle says: ‘‘. . . the low-storage capacity of the male- producing eggs as compared with the high storage capacity of the female-producing eggs is therefore an index of higher oxidizing GERM-CELL HISTORY IN THE BROOK LAMPREY 105 capacity of the male-producing eggs as compared with the female- producing eggs.” More recently Whitney (19) has reinvestigated the problem relative to the effect of oxygen as male-producer in the rotifer, and has come to conclusions opposite to those of Shull, namely, that oxygen does not act as a factor in the production of male- producers. The question cannot be considered fully settled. j. Variations in sex ratio. A variation in sex ratio might indi- cate that sex is not irrevocably established at the time of fertili- zation. Jt has been found that, in certain dioecious plants, females are more commonly derived from seeds of one region and males from those of another region. This may be due to differ- ences in the metabolic activity of the two kinds of seeds, brought about, possibly, by differences in the conditions of the environ- ment under which they were raised. Montgomery (’08) found that there were 8.19 males for every female in a count of 41,749 spiderlings. Out of the total of 127 cocoons, only eight showed a male ratio of less than one. Out of the total number of eggs in the cocoons only 2871 failed to hatch, and even though all of these should be assumed to be female eggs, the ratio would not be appreciably altered and the results cannot, therefore, be due to selective survival. Exam- ples of this sort might be given by the score, and they are not easy to explain on the hypothesis that the chromosomes are the only and absolute sex determiners; for this hypothesis demands that there should be an equal number of males and females pro- duced. Pearl and Parshley (’13) have found that, in cattle, the sex of the offspring is somewhat dependent upon the time of coitus. Early in the heat the number of males to one hundred females was 98.4; in the middle of heat the ratio was 115.5, and late in heat it was 154.8. The conclusion is drawn that, granting the presence of an X-chromosome, the results may be interpreted by assuming that it is not a positive cause of sex differentiation, but rather an inhibitor of the development of male characters— two doses inhibits maleness, while one dose is insufficient. On this hypothesis it is assumed that the general conditions of metab- 106 PETER OKKELBERG olism in the germ cells might modify sex. The case is similar to that of delayed fertilization in the frog which results in the formation of an increased number of males. In fish cultures it is not rare to find an excess of males or of females. Woltereck (’08) has reported various records made by Thumm upon the sex ratio in fishes. In Jenynsia lineata broods were obtained of 68, 92, and 116 individuals without a single female. A female of Cichlasoma nigrofasciatum three vears old, bred to a male one year old, gave a progeny of 800 individuals, ‘of which not fifty were females. A female of the same species one year old, bred to a male two years old, (the same male as in the first case), gave 400 young, of which over 300 were females. To summarize: ‘‘aeltere starke Weibchen, verpaart mit jiingeren, daher schwicheren Minnchen, brachten in Nachzucht vorwie- gend Mannchen und umgekehrt.”’ It was also found that in viviparous ‘K6érpflingen’ the percentages of males were higher in the spring than later in the season, and that in the fall it was often very low. ‘This apparently corresponds with the results of Riddle on pigeons, where also the percentage of males is greater in the early part of the season. It does not seem possible that the results obtained by Thumm in the cases above are due to selective survival. 6G. General conclusions in regard to the problem of sex determina- tion. In the case of the lamprey it has been seen that, in young stages, a series of individuals may be arranged exhibiting all the intermediate forms between apparently pure females and appar- ently pure males. Pure in this sense is used to designate the individuals which possess no visible characters which normally distinguish the opposite sex. Out of the sex intermediates both males and females develop, so that in the adult condition only two kinds of individuals are found, functional males and females. The designation of sex intermediates among the young is based on the appearance in the germ gland of primary sex characters, namely, oocytes and cell nests. The presence of oocytes in the germ gland is unquestionably a female character, while the pres- ence of cysts may indicate a juvenile condition or a male charac- ter. The oocytes, in one case, and the well-developed cysts, in. GERM-CELL HISTORY IN THE BROOK LAMPREY 107 the other, may be considered homologous characters of the oppo- site sexes. Judging from the quantitative appearance of the characters in the germ glands of hundreds of larvae which have been care- fully studied, the conclusion is reached that every individual is a potential hermaphrodite possessing the sex qualities pertaining to either sex. It appears, however, that some individuals are more strongly inclined toward the male side, others more strongly toward the female side. Some, on the other hand, are appar- ently in a balanced condition as regards sex, and it would be im- possible to say whether they are more strongly male or more strongly female. This condition usually lasts only a short time, after which one sex takes the lead over the other. When, in the course of time, an individual has become more strongly male or female, the opposite sex character gradually disappears, or at least remains undeveloped; so that if it appear at all in the adult, it is in a rudimentary or degenerate condition. This is the case, for instance, with the oocytes which are in an undeveloped condi- tion in the adult testes. ‘The secondary sex characters which appear in the adult are probably not hereditary characters at all, but are formed as a result of the presence of special kinds of hormones produced by the testis or ovary. These secondary characters are not present until after metamorphosis, when one or the other appears, depending upon the form of germ gland present. As far as the primary sex characters are concerned, it appears that both male and female potencies exist in every individual from the beginning of development; that is, from the time that the egg is fertilized, and probably in both of the sex cells that are brought together in fertilization. These potencies then are transmitted from parents to offspring. It seems quite likely, however, that the two potencies are not always transferred from parent to offspring in equal strength, so that the two are not, in all cases, in a balanced condition from the very beginning. ‘This appears from the fact that during the stage of sex differentiation all kinds of variations are found as to the quantitative appearance of the male and female characters. Since all of the larvae develop 108 PETER OKKELBERG under practically identical conditions, it does not seem likely that these variations can be due to environmental factors. Many mendelian workers have found it convenient to assume that the appearance of the two sexes in approximately equal numbers in most animals is due to the fact that one sex is hetero- zygous for sex, while the other is homozygous for this character. This suggestion comes from the fact that the sex ratio corre- sponds with the ratio obtained when a first-generation hybrid is bred to a pure recessive. In this case half of the offspring will be pure recessives or homozygotic, while the other half will be hybrid or heterozygotic with the dominant character present. From the study of the sex chromosomes it seems to have been found that one sex may produce two kinds of germ cells which are visibly different, while the other sex produces only one kind. It appears further that it is sometimes one sex and sometimes the other that is heterozygous with regard to the sex character. It has been supposed, and is still maintained by a number of investigators, that the sex chromosomes are absolute sex deter- miners. The idea, however, that the chromosomes act quali- tatively has given way to the belief that the influence exercised by the sex chromosome is a quantitative one, and this concep- tion has paved the way for a better understanding of the sex phenomena in forms that exhibit hermaphroditic tendencies. The conception that all individuals carry the factors of the oppo- site sex in a latent condition will probably prove to be correct, and it may lead to a general acceptance of the theory that sex is not unalterably fixed at the time of fertilization. The primary difference between a male and a female in any species is not as great as one might conclude from the appearance of the adults of the two sexes. The first sexual changes usually take place in the germ gland; in the female some of the germ cells very early stop dividing and enter upon a period of growth, while in the male the germ cells continue to multiply for a long time. Since the primary difference between the female and the male is, that in one the germ cells enter early upon a period of growth, while in the other they continue to divide, it seems probable that GERM-CELL HISTORY IN THE BROOK LAMPREY 109 one or the other mode of development is the result of a difference in the body metabolism in the two kinds of individuals, and is not due to the inheritance of unalterable sex factors. If sex should prove to be the result of slight differences in metabolism, it would be easy to understand how a reversal of sex might take place under certain circumstances. It must be admitted that these metabolic differences might exist in the animals from the very beginning of development, and, in so far as they are transmitted by the parental germs which unite in fertilization, they may be said to be inherited. If the sex characters are to be compared with other so-called mendelian characters, we have to admit also the possibility of a quantitative variation in the latter, which seems to be contrary to the opinion of the majority of geneticists at the present time. From evidence already presented we are forced to the conclu- sion that in many dioecious forms, at least, every individual is a potential hermaphrodite, in so far as it carries the latent quali- ties of the opposite sex. Whether an animal develop into one or the other sex may depend upon an inheritable quantitative rela- tionship existing between the male and female potentialities in the fertilized egg. As has already been mentioned, there seems to be some evidence that in some forms it is one and in other forms it is the other of the two sex cells that unite in fertilization which is responsible for the quantitative difference of the sex factors. The theory advanced by Castle (’03), that both male and female cells are heterozygous with regard to sex, required the assump- tion that selective fertilization was necessary in order to bring about the observed results. This has been objected to on account of the lack of evidence that there is such a selection among the germ cells. Papanicolaou (715) and Stockard and Papanicolaou (16) have brought forward some evidence that selective fertili- zation might take place in the case of guinea-pigs, but the full data have not, as far as I know, been published. If it be found necessary to consider one sex heterozygous for sex and the other homozygous, the formula that appears most ap- plicable is that adopted by Goldschmidt, which has been ac- cepted with some modifications by Witschi and others. 110 ; PETER OKKELBERG In the case of the lamprey, it does not seem necessary to tie the question of sex up with chromosomal constitution. It is easy to conceive of every fertilized egg as being practically in a balanced condition as regards sex. Some may be more strongly inclined in one direction and some in the other, and, in so far as this is so, the sex characters may be considered as inherited. But if we look upon the development of one or the other sex as a result of metabolic differences, there is no necessary reason why these differences should be referred to the chromosomal make-up of the fertilized eggs. They may equally well be the result of cyto- plasmic differences in the eggs, and these may be present even before fertilization. If we consider sex from this standpoint, it is not difficult to understand how, as a matter of chance, there might be an equality of males and females when conditions of development are normal, and also to understand how, under extraordinary circumstances, sex might be altered in the develop- ing organism. A cytoplasmic inheritance of the female charac- ters (FF) has been suggested by Goldschmidt for the gypsy- moth. In the case of Sagitta, which is hermaphroditic, it is claimed by Elpatewsky that the development of a primordial germ cell into a male or a female cell is dependent upon the pro- portion of the cytoplasmic body, the so-called ‘besondere K6r- per,’ that each cell receives. Whether we consider the chromosomes or some other part of the cell as responsible for the determination of sex, we must, in the last analysis, think of sex determination as due to the rela- tion between two opposite potencies which are both present in the fertilized egg. In true hermaphrodites, in which male and female germ cells are matured simultaneously, the two potencies are in a state of equilibrium, so that the presence of one is not antagonistic to the other. In protandric and protogynous her- maphrodites the two sexual states seesaw back and forth so that each alternately replaces the other, while in a case like that of Crepidula, which under normal conditions is male in the young stage and female in the older stage, the male potency never reap- pears. In the two latter cases we may think of the antagonism between the two sexual states as the result of the action of cer- GERM-CELL HISTORY IN THE BROOK LAMPREY 111 tain hormones, secreted during the development of one or the other form of sex cell, and which inhibits the development of the other. When one set of sex cells is exhausted, the action of the hormones ceases and the other set of cells begins to develop. In dioecious animals and plants the two forces do not exert themselves in the same individual except as the result of unusual conditions. In the free-martin, for example, the female potency is in the lead from the beginning of development; but by the action of hormones circulating through the body of the embryo, the male factor asserts itself so that the female factor is partially suppressed, even to the extent ‘‘that a gonad with a primary female determination may form a structure which is morpho- logically a testis” (Lillie, 717, p. 468). When the spider-crab, Inachus, is infected with the parasite Sacculina, the males, para- sitically castrated, may show every degree of modification toward the female state, even to the appearance of ova in the remaining part of the testis. The females, however, are not transformed toward the male condition, and the conclusion is drawn by Geoffrey Smith that the male is heterozygous for sex and the female homozygous. Such a conclusion is hardly war- ranted, for the parasite does not seem to act simply by arresting the action of one sex potency, but by also elaborating certain materials which are favorable to female development. That this is so may be surmised from the fact that when immature females are infected, the effect is ‘‘to force them to assume pre- maturely adult female characteristics’? (Smith, 710). There is no reason, therefore, why a female, when infected, should be transformed toward the male side. In pigeons, sex seems to be a matter of metabolic difference, and a disturbance of the metabolic level may be brought about by hybridization as well as by overwork in reproduction so that a sex reversal is effected. In this case Riddle thinks he has dem- onstrated ‘‘that germs normally female-producing, have, under experiment, been made to develop males; and that germs which were prospectively male-producing have been made to form female adults’ (Riddle ’16, p. 410). In the case of the gypsy- moth, hybridization again seems to disturb the sex metabolism JOURNAL OF MORPHOLOGY, VOL. 35, NO. 1 +12 PETER OKKELBERG so that a prospective female develops male characters and a prospective male develops female characters. Whether or not complete reversal of sex has occurred in moths, does not appear from literature on the subject, although there are apparently cultures yielding nothing but males (Goldschmidt, ’17, p. 605). In Bonnelia the sex metabolism is disturbed by environmental conditions. This is also true in the case of Crepidula. In frogs, delayed fertilization determines the results, as possibly also does temperature to a certain extent. In rotifers the change may be the result of the amount of oxygen present: in the culture (Shull and Ladoff, 716; Shull, 718). Whitney, however, has obtained different results with rotifers. From the various cases of sex reversal in plants, mentioned above, it appears that in these cases the change is effected by disturbing in one way or another the normal conditions under which the plant lives. It seems to be amply proved that among dioecious animals and plants every individual carries the qualities of the opposite sex in a latent condition. This is a great step toward the solu- tion of the problem of sex determination; but it remains to explain why in some cases one potency asserts itself, while in other cases the other appears. Opinions on this question converge around the conception of a variableness in cell metabolism and the action of enzymes. Riddle thinks of male- and female-producing eggs, in the case of pigeons, as different in regard to their storage capacity—a less storage capacity pertains to the male—and a high storage capacity pertains to the female-producing germ. Riddle says: ‘‘The progressive increase in storage capacity of the eggs during the season—under overwork—is to be interpreted as a decrease in the oxidizing capacity of the same eggs.” This opinion is similar to that expressed by Shull in the case of rotifers. The metabolic capacity of the germ is, of course, reflected in the adults derived from them: We can easily see how, in forms like the lamprey, the storage capacity and the oxidizing capacity may so nearly balance each other that every larva may exhibit both tendencies in different parts of the body. This, as has been suggested above, may be due to slight inequalities in the cells resulting from division or to environmental factors of some sort GERM-CELL HISTORY IN THE BROOK LAMPREY ts influencing the germ cells in different parts of the gonad. Thus we may have, in the same gonad, certain cells with high oxidizing capacity which continue to divide and form cell nests, and other cells, with a high storage capacity, which enter the growth period very early and become oocytes. After the larva has become decidedly male or female in character, as evidenced by the pro- portion of cell nests and oocytes, it appears that the opposite sex tendency is in decline. This can be attributed to the pres- ence of sex-differentiating enzymes produced by the predominat- ing sex character. This results in the arrest of the development of the opposite sex character and often in its degeneracy. Goldschmidt has worked out a theory of enzyme action in connection with sex which seems to be in the right direction. He assumes that ‘‘in the fertilized egg the enzymes which govern the differentiation of the organism towards one of the two alter- natives, maleness and femaleness, are both present.’’ These hypo- thetical enzymes he speaks of as andrase and gynase. The dis- tribution of the sex chromosome “‘results in the formation of two kinds of fertilized ova, differing in the relative concentration of the two enzymes.”’ Since in ‘‘mixtures of different enzymes, every single one reacts independently, providing no interfering reaction product is formed,” a decision must be reached during differentiation of the organs as to whether they shall develop along the male or the female line. ‘‘This decision must be brought about by the action of the dominating enzyme.” ‘The more nearly the two enzymes approach each other in strength, the earlier do they show their double influence on the developing or- ganism. Such seems to be the case in the intersexual moths which show all degrees of intersexuality, from slight changes in the sec- ondary sexual characters, which are latest to be formed, to changes in the germ gland itself, which is the first sex organ to differenti- ate. The same idea may be applied to the lamprey, where various grades of intersexuality are found in the germ gland, and where the sex differentiating factor seems to operate early in some of the larvae and later in others. We are not, strictly speaking, concerned in the present work with the causes which underlie the development of the secondary 114 PETER OKKELBERG sexual characters or the accessory reproductive organs, since there are none in the larval stages of the lamprey. In insects the secondary sex characters appear to be unaffected by the pres- ence or absence of the sex glands, but in most other forms the appearance of the secondary sexual characters may be the result of certain hormones which are produced by the predominating primary sex elements. This appears to be the case in the lam- prey. Sooner or later in the life of the individual the male- or the female-producing enzyme, if such it be, takes the lead so that the action of the opposite enzyme is more and more sup- pressed. The male or the female germ gland, which develops as a consequence of the stronger enzyme, is capable of producing certain hormones which, both at and after metamorphosis, cause the secondary characters to appear. The above conception is not opposed to the theory that the so-called sex chromosomes are associated with the phenomena of sex in many cases. The evidence indicates, however, that they are only one link in a series of processes which result in sex determination, and that other factors may operate so a$ to change development, in spite of the presence of the sex chromo- some. ‘The physiological action of the sex chromosome may be fundamentally the same as that of other factors. Finally, if we think of sex as an hereditary character, as it seems we must, then it is amply demonstrated that here we have an hereditary character that can be modified by a variety of circumstances. Unless we assign the sex character to another category than other hereditary characters, we are forced to acknowledge the possibility that other hereditary characters are modifiable also. If this should prove true, it is possible that the idea that the sex character is changeable will be accepted with less reserve than heretofore. GERM-CELL HISTORY IN THE BROOK LAMPREY 115 GENERAL SUMMARY OF OBSERVATIONS A. Origin and early history of the germ cells 1. The germ cells are first recognizable in the American brook lamprey when the mesoderm separates from the entoderm, as large yolk-laden cells which become included in the mesoderm. Their history previous to this time could not be traced. Their large size, however, indicates that they are early segregated cells. 2. The number of germ cells that become included in the meso- derm is small. There is evidence that many of them never reach the germ-gland region. Some of these degenerate before dividing, others form cysts in other regions of the body, and the possibility is suggested that some of them may be extruded into the lumen of the intestine in early stagés. 3. During the early period of their history the germ cells shift from a lateral position in the mesoderm to a median posi- tion. The change in position is accredited to a shifting of the tissues surrounding the cells and, to a lesser extent, to. independ- ent migration. 4. The germ cells begin to lose their yolk when the larva is about 5.5 mm. long, and no yolk remains in the cells when the larva is 10 mm. long. They do not begin to divide until the larva is about 20 mm. long. 5. The germ cells may be distinguished from the somatic cells by their size, structure, and location. B. Period of secondary division 6. When the larva is about 20 mm. long the germ cells begin to divide by mitosis. 7. After each mitosis the germ cells either separate or remain together, forming cell nests. Peritoneal cells migrate in and form follicles around the individual cells and cysts. 8. An astrosphere is distinguishable in the germ cells of this stage. 9. A vitelline body is found in the cytoplasm. Its origin could not be ascertained. It becomes a very prominent structure in the growing oocyte of later stages. 116 PETER OKKELBERG 10. Numerous mitochondria are present in the cytoplasm of the germ cells in most phases of their history. 11. Two plasmosomes are present in the primordial germ cells but in the growing oocyte there is only one. 12. The period of secondary division lasts until the larva is about 35 mm. long. During this period the larva appears indif- ferent as to sex. C. Period of sex differentiation 13. The period of sex differentiation extends from the time the larvae are about 35 mm. in length until they are about 70 mm. long. In some larvae, however, sex differentiates much earlier than in others. During this period the sex of the larvae is indeterminate. The condition may be described as juvenile hermaphroditism. 14. A varying number of oocytes appear in practically all the glands during this period, so that a series of glands might be arranged possessing from 0 per cent to 100 per cent of odcytes. 15. The changes taking place in the oocytes during the synap- sis phase are described. 16. Numerous germ cells degenerate during this period. De- generation may take place during the synapsis phase, the growth phase, or the indifferent phase of the germ-cell history. 17. When sex is established the germ cells belonging to the. opposite sex disappear or remain in the gland in a rudimentary condition. CONCLUSIONS The following general conclusions may be drawn from the above study of the germ cells of the American brook lamprey: 1. The germ cells are segregated very early in the life of the animal even before the germ layers are definitely established. They are first recognizable when the mesoderm separates from the entoderm. 2. The definitive germ cells take their origin from no other source than the primordial germ cells and the germ cells take no GERM-CELL HISTORY IN THE BROOK LAMPREY 17 part in the formation of somatic structures. Numerous germ cells are produced which do not become functional, and these degenerate and disappear during the process of development. 3. The germ cells of each germ gland are usually of two kinds namely, those showing a tendency toward rapid division (kata- bolic) and those showing a tendency for growth (anabolic). The former are regarded as having a male, the latter a female poten- tiality. The relative proportion of anabolic and katabolic cells determines whether the larva becomes a male or a female indi- vidual. 4, Observations seem to warrant the conclusion that each larva of this species carries the potentiality of both sexes, and that sex, therefore, is not irrevocably fixed at fertilization. 118 PETER OKKELBERG LITERATURE CITED ABRAMOWICZ, HeLeNE 1913 Die Entwicklung der Gonadenanlage und Ent- stehung der Gonocyten bei Triton taeniatus (Schneid). Morph. Jahrb., Bd. 47. Atcock, R. 1899 On proteid digestion in Ammocoetes. Journ. Anat. Phys., vol. 33. ALLEN, B. M. 1904 The embryonic development of the ovary and testis of the mammals. Am. Jour. Anat., vol. 3. 1906 The origin of the sex-cells of Chrysemys. Anat. Anz., Bd. 29. 1907 a A statistical study of the sex cells of Chrysemys marginata. Anat. Anz., Bd. 30. 1907 b An important period in the history of the sex-cells of Rana pipiens. Anat. Anz., Bd. 381. 1909 The origin of ithe sex-cells of Amia and Lapidbsteus Anat. Rec., vol. 3. 1911 a The origin of the sex cells of Amia and Lepidosteus. Jour. Morph., vol. 22. 1911 b The origin of the sex cells in Chrysemys (a reply to A. Dustin). Anat. Anz., Bd. 39. ANcEL, P. 1903 Histogénése et structure de la glande hermaphrodite d’Helix pomatia (Linn.). Arch. Biol., T. 19. Aunap, E. von 1913 Ueber die Chondriosomen der Gonocyten bei Knochen- fischen. Anat. Anz., Bd. 44. . BacuuvuBer, L. J. 1916 The behavior of the accessory chromosomes and of the chromatoid body in the spermatogenesis of the rabbit. Biol. Bull., vol. 30. BacHMANN, FrepaM. 1914 The migration of the germ cells in Ameiurus nebu- losus. Biol. Bull., vol. 26. Baurour, F.M. 1877 The development of Elasmobranch fishes. Journ. Anat. Rhy stvol iiss i128: 1881 RR. soes biol Paris, I 79. 1916 a Sur la structure de la spore des Microsporidies. C. R. soe. biol. Paris, T. 79. Linton, E. 1901 Parasites of fishes of the Woods Hole region. Bull. U. 8. Fish Comm., vol. 19. Lutz, A., aND A. SpLeNDORE 1908 Ueber Pebrine und verwandte Mikrospor- idieny Centralbl. Bakt. u. Parasitenk. (I) Orig., Bd. 46. MARcHOUX, SALIMBENI, AND P. L. Stmonp 1903 La fievre jaune. Ann. |’inst. Pasteur ais 17%, Mavor, J. W. 1915 Studies of the Sporozoa of the fishes of the St. Andrew’s region. Annual Report Dep’t Marine and Fisheries, 47th Session, paper no. 39b. Mercier, L. 1908 Néoplasie du tissue adipeux chez des Blattes (Periplaneta orientalis L.) parasitées par une Microsporidie. Arch. Protist., Bd. 11. 1908 a Sur le développement et la structure des spores de Thelohania giardi Henneguy. C. R. acad. sci. Paris, T. 146. 1909 Contribution a l’étude de la sexualité chez les Myxosporidies et chez les Microsporidies. Mém. l’acad. roy. Belgique (class. sci.) Deuxiéme series, T. 2. MrAzex, A. 1899 Sporozoenstudien. II. Glugea lophii Doflein. Sitzgsber. kénig. b6hm. Gesell. Wissen. (Math.-naturw. Classe), 1899. Omort, J. 1912 Zur Kenntnis des Pebrine-Erreger, Nosema bombycis Nigeli. Arb. kaiserl. Gesundheitsamte, Bd. 40. PéREZ, C. 1905 Microsporidies parasites des crabes d’Arcachon. (Note pré- liminaire.) Bull. Stat. Biol. Arcachon, T. 8. PrerrFER, L. 1895 Die Protozoen als Krankheitserreger. Nachtrige. Sasaki, C. 1897 Report on the studies on Pebrine-disease among Nipponese silk-worms. Tokio. Scur6pvER, O. 1909 Thelohania chaetogastris, eine neue in Chaetogaster dia- phanus Gruith schmarotzende Microsporidienart. Arch. Protist., Bd. 14. 1914 Beitraige zur Kenntniseiniger Microsporidien. Zool. Anz., Bd. 43. ScuuperG, A. 1910 Ueber Mikrosporieien aus dem Hoden der Barbe und durch sie verursachte Hypertrophie der Kerne. Arb. kaiserl. Gesundheits., Bd. 33. ScuuBperG, A., AND C. Roprigurz 1915 Thelohania corethrae n. sp., eine neue Mikrosporidienart aus Corethra-larven. Arb. kaiserl. Gesund., Bd. 50. Srmonp, P. L. 1903 Note sur une Sporozoaire du genre Nosema, parasite du Stegomya fasciata. C. R. soc. biol. Paris, T. 55. Srempety, W. 1902 Ueber Thelohania miilleri L. Pfr. Zool. Jahrb. Abt. Anat., Bd. 16. 1909 Ueber Nosema bombycis Nigeli. Arch. Protist., Bd. 16. 182 R. KUDO STEPHENS, J. W. W., AnD S. R. CuristopuHers 1908 The practical study of malarial and other blood parasites. London. SrrickLanp, E. H. 1913 Further observations on the parasites of Simulium larvae. Jour. Morph., vol. 24. THELOHAN, P. 1894 Sur la presence d’une capsule a filament dans les spores des Microsporidies. C. R. aead. sei. Paris, T. 118. 1895 Recherches sur les Myxosporidies. Bull. sei. France et Belg., ae 26; WEISSENBERG, R. 1911 Beitriige zur Kenntnis von Glugea lophii Doflein. II. Ueber den Bau der Cysten und die Beziehungen zwischen Parasit und Wirtsgewebe. Sitz. Ber. Ges. naturf. Fr. Berlin. 1913 Beitrige zur Kenntnis des Zeugungskreis der Mikrosporidien, Glugea anomala Moniez und hertwigii Weissenberg. Arch. mikros. Anat., Bd. 82; Abt. IT. Wuits, G. F. 1919 Nosema-diseases. Bull. U. S. Dep’t Agric., no. 780. ZANDER, E. 1911 Handbuch der Bienenkunde. II. Krankheiten und Schad- linge der erwachsenen Bienen. Stuttgart. EXPLANATION OF FIGURES IN PLATES All the drawings have been made by means of Abbe’s drawing apparatus. Zeiss’ compensation oculars, 6, 8, 12, and 18 and homogenous oil-immersion objective 2 mm., were used. Abbreviations in the explanations of plates are as follows: G., Giemsa-staining, followed by acetone dehydration, by mounting in cedar oil; H., Heidenhain’s iron-hematoxylin staining; P.F., pressed mechanically and stained after Fontana’s method; S., section preparation; Sm., smear preparation. 183 PLATE 1 EXPLANATION OF FIGURES Thelohania magna Kudo 1 to 23. Stages in schizogony. land2 Young schizonts. S.G. X 2860. 3 to6 Nuclear division. 8.G. X 3500. 7to9 Successive stages of binary fission. 8.G. X 2360. 10 to 14 Stages in multiple division. S.G. X 2360. 15 to 23 Stages of binary fission of the second type. S.G. X 2360. 24 ‘situs = eles 196 IPAViipiebiankbubes in Larval 2... ook POMEM be. Cosel vee ence 196 ZeeVialpienianwolbesmm pep lp ali stage ss serine aeee avert 27 se recone oe 199 Si IMENT or onepay TAU OYSS), iil JOU. op emoawBeL oo odds 5 obo bu GuedoLooeae lao 200 Am Malpighian Gubes 1m IMaeO).5 «(5 2)..,.,<\ ise See ero etos aie < Seles le ae deiaes 203 SHUTOMTED Ac: 3h obese Hatt Aeieete Noh ee ae RS it 4 Een." a en Me as 205 INTRODUCTION The metamorphosis of the malpighian tubes of insects has been much studied by many authors: Karawaiew (98), Anglas (00), Peréz (10), in Hymenoptera; Karawaiew (’99), Deegener (00) in Coleoptera; Vaney (’02), Peréz (01) in Diptera; Cho- lodkovsky (87), Samson (’08), Hufnagel (?12), Ikeda (13) in Lepidoptera. The results so far published are widely divergent, especially for Lepidoptera. Cholodkovsky (’87) demonstrated the fact in Tineola biselliela that the malpighian tubes of the larva disappear gradually by histolysis, while the basal trunk elongates to form the imaginal tubes. A diametrically opposite view has been held by Samson (08), who states that in Heterogenea limacodes that the mal- pighian tubes of the larva directly transform into the imaginal ones, although they undergo certain histological alterations. Hufnagel (12) investigated the metamorphosis of the tubes in Hyponomeuta padella, and came to the conclusion that one part of the larval tubes alone persists and differentiates into the 195 JOURNAL OF MORPHOLOGY, VOL. 35, NO. 1 196 HIROWO ITO definitive organ, the remaining portion disappearing entirely. In Bombyx mori Ikeda (713) has described.a mode of develop- ment very similar to that described by Cholodkovsky. The present study was undertaken to ascertain definitely in what manner the metamorphosis of the malpighian tubes takes place in the silkworm, and I was fortunate enough to obtain some results which have not hitherto been recorded regard- ing the following points: a) the histolysis of the urinary bladder; b) the histolysis and histogenesis of the common stem, and c) the histolysis of the portion of the malphigian tubes enclosed in the walls of the rectum, ‘included portion.’ MATERIAL AND METHODS The material for. the present study was obtained from the univoltine race of Bombyx mori, known in Japan as ‘Awojiku.’ The larvae were fixed either in Perenyi’s fluid or in sublimate aleohol warmed to about 80°C. For the pupae various fixing reagents were used. The best results were obtained with picro- sulphuric acid and Perenyi’s fluid. The material was embedded in paraffin. The sections were cut from 5 to 10u in thickness and were stained with Delafield’s haematoxylin, picrofuchsin, and various other combinations of stains. OBSERVATIONS 1. Malpighian tubes in larvae The malpighian tubes are situated near the posterior part of the body, opening into the ventrolateral side of the alimentary canal just at the junction of the small intestine and the colon. The tubes are six in number, three on each side. Two tubes first unite into one and then the third one fuses with the other two. The common stem of the three tubes is connected with the urimary bladder. This organ opens directly into the alimentary canal. Each malpighian tube runs at first forward as far as the anterior edge of the sixth or seventh segment, along the dorso- lateral side of the stomach. Then it turns on itself and goes backward a little beyond the colon, where it makes irregular MALPIGHIAN TUBES OF BOMBYX MORI L. 197 econvolutions. Finally, the posterior portion of each tube enters the wall of the rectum and ends blindly after several convolutions. The outer side of the tubes is richly provided with tracheae. For the sake of convenience, four regions may be distinguished in describing in some detail the histological structure of the tubes: a) the urinary bladder; b) the common stem; c) the coelomic portion which hes freely in the body-cavity, and d) the portion enclosed in the rectal walls, the ‘included portion.’ a. The urinary bladder. The urinary bladder is ellipsoidal in form and measures about 0.95 mm. by 0.75 mm. in the fifth stage. Its epithelium is composed of large flat cells, the cell boundaries not always being clear. The cytoplasm, which stains a violet-pink with eosin, shows a reticular structure as in figure 3. The nuclei are round or oval, and differ greatly in size, accord- ing to the size of the cells. The nuclei are densely filled with coarse deeply staining chromatin granules. The inner wall of the epithelium is covered by a chitinous intima which is directly continuous with that of the intestine. There is no striated border, contrary to the description of Bordas (11). The intima is fairly thick and does not stain, with either eosin or haema- toxylin. The basement membrane is a fine transparent struc- tureless limiting membrane. In some sections it is not readily detected. The circular muscles are strongly developed, and the striations of the muscles are clearly defined, as is shown in figure 3. There are some longitudinal muscles interior to the circular ones. b. The common stem. The common stem is a narrow and short tube, measuring 0.52 mm. in length and 0.26 mm. in width. It does not differ in appearance from the urinary bladder as described above, excepting the smallness of the lumen, as is shown in figure 5. The numerous cells at the distal end of the common stem are very small and pressed closely together. These cells represent the imaginal ring (fig. 1, ir), although they never have a several layered appearance. The nuclei, which are round or oval, are densely chromatic. The chromatin granules are of the same color and size. During the larval stage the imaginal cells divide to elongate the coelomic portion. 198 HIROWO ITO c. The coelomic portion. The epithelium is composed of a single layer of two large polygonal cells. In cross-section, the cell boundaries are not well defined. The cytoplasm, which stains deeply with eosin, contains, near the striated border, many large vacuoles which are undoubtedly secretion products (fig. 4). Some faintly stained granular substance occurs in the vacuoles. The nuclei are irregularly branched, as shown in figure 1, and contain fine chromatin granules and nucleoli. The epithelium of this portion never has the intima, but instead is lined with a striated border, as has been observed by Schindler (78), Bordas (10, 711), Metalnikov (’08), and Veneziani (’03, Fig. 1 Surface view of urinary bladder (wb.) and common stem (cs.) and coe- lomic portion (cp.) from a total preparation. XX 20. 03 a, 704). The striated border is lightly stained and is marked by many very fine and closely set striations. The lumen differs in shape in different portions, and is filled with numerous granules and crystals of various shapes, such as cubical, octahedral, quadropyramidal, which are supposed to be made of a urate salt, so far as my examination goes. The epithelial cells have a very delicate transparent membrane which completely sur- rounds the tubes. Muscles are entirely lacking in this region. d. The included portion. In a cross-section of the rectum (fig. 6), lying inside the circular muscles, is a nucleated peritoneum, which is a double layer, as has been described by Metalnikov (08), Ishimori (16). Though it is very diffcult to make out the exact course of the malpighian tubes in the rectal wall, it can MALPIGHIAN TUBES OF BOMBYX MORI L. 199 be said that the more wavy it becomes, the larger folds in both the inner and outer spaces of the peritoneum. They thus sur- round almost completely the epithelium of the rectum. Both the cells and nuclei of the tubes in the rectal wall are very much smaller than those of the coelomic portion of the tubes. The epithelium is composed of very flat cells, especially in the outer space of the peritoneum. The cytoplasm is either homogeneous or granular in structure. The nuclei are also ramified, and densely chromatic, with small darkly stained chromatin granules. The striated border is not well developed in this region, but doubtless it is always present, at least in a feebly developed condition, even in the cases in which it is not visible. The lumen is nar- rower than that of the preceding region and contains some excreted granules which stain deeply with eosin. 2. Malpighian tubes in the prepupal stage The urinary bladder has become considerably reduced in diameter, measuring about 0.58 mm. by 0.25 mm. The cyto- plasm has become vacuolated, and is gradually disappearing. The nuclei of the component cells have united to form a mass which stains deeply with haematoxylin. The intima is now separated from the cells and occurs shrunken in the lumen. The basement membrane is clearly defined, owing to the vacuo- lation of the cytoplasm, and is granular in structure. The muscles are also on the way to histolysis. The sarcolemma shrinks and is attacked by the phagocytes, and at the same time the striation disappears, as is the case with histolysis of the intestinal muscles. Finally, the muscular substance completely disappears, leaving behind some nuclei here and there. In the common stem, just the same process of disintegration takes place as in the urinary bladder. The imaginal cells are greatly increased in number and size. The nucleus is round or ellipsoidal and contains many fine chromatin granules. The coelomic portion comes in contact with the alimentary canal, which by this time has become very much flattened and contracted, owing to the digestion and excretion of the contents. 200 HIROWO ITO The distal end of the tubes has become narrow and shortly prior to pupation is detached from the rectum. The cytoplasm is homogeneous or very finely granular, and vacuoles no longer occur. The irregularly ramified nuclei are slightly shrunken and contain fine chromatin granules. The striated border is very faint and feebly developed, as is characteristic of the included portion. The lumen of the posterior portion is filled with homogeneous substance, while the anterior portion contains coarse granules. The basement membrane has the same char- acter as in the preceding region. The most remarkable histological change takes place in the included portion of the malpighian tubes, in connection with the histolysis of the rectum. The cytoplasm is granular and con- tains some small vacuoles. The ramified nuclei are fragmented into pieces, which form masses containing deeply staining chro- matin in the form of a solid ball (fig. 7). Some massed nuclei have already been divided into globules. The lumen has com- pletely disappeared as a result of the contraction of the epithelial cells. The phagocytes attack the basement membrane which sometimes is found already destroyed here and there, as is shown in figure 7. Such disintegration of the included portion of the malpighian tubes has not been observed by any previous authors, 3. Malpighian tubes in pupae In a pupa not older than twenty hours the cytoplasm of the urinary bladder and the common stem has completely disap- peared. The massed nuclei are now fragmented into globules, which are gradually scattered in the body cavity, as is shown in figure 10. The intima is cast off from the lumen with the last ecdysis. There is no trace of the basement membrane, it being digested by the phagocytes. The muscles are also disorganizing by the attack of the phagocytes. The imaginal cells at the distal portion of the common stem increase in number by multi- plication and gradually take the place occupied by the destroyed epithelial cells. The coelomic portion of the tubes is detached from the ali- mentary canal and lies in convolutions among the fat-bodies. MALPIGHIAN TUBES OF BOMBYX MORI L. 201 The cytoplasm is granular and contains large vacuoles between the nuclei and the basement membrane. ‘The nuclei are ramified as in the larval stage. The striated border disappears, probably at the last ecdysis. The lumen is conspicuously reduced in diameter, due to the contraction of the cells. The basement membrane comes in close contact with the periphery of the epithelium, as in the larval stage. The included portion of the tube no longer retains its original shape, owing to the histolysis of the component cells. The fragmentation and dispersion of nuclear elements is more active, and the globules begin to migrate from the peripheral portion. Some globules are already being engulfed by the phagocytes (fig. 9). In a pupa two to three days old, the massed nuclei of the urinary bladder and the common stem have entirely disinte- grated into globules. The muscular substance has now com- pletely disappeared, being digested by the phagocytes, and the muscle nuclei alone are still intact. A great many granular spheres are found around the muscle nuclei. The coelomic portion now lies in the same position as in the preceding day. The cytoplasm is granular in the posterior portion and vacuolated in the anterior. The vacuoles gradually migrate toward the nucleus so as to come to lie between it and the surface of the cells lining the lumen. The ramified nuclei are central in position and contain many fine chromatin granules. Within the lumen are some granules and globules which are strongly eosinophilous. The basement membrane becomes gradually separated from the cells, as is shown in figure 8, and finally it is completely lifted up as a result of a considerable contraction of the cells. A great number of the phagocytes attack the basement membrane, as is shown in figure 11. This observation agrees well with that of Samson (’08), on the mal- pighian tubes of Heterogenea limacodes. Observing the destruc- tion of the basement membrane, both Cholodkovsky (’87) and Ikeda (’13) came to the conclusion that the malpighian tubes of the larva completely disappear. 202 HIROWO ITO The nuclear globules in the included portion have completely disappeared, probably being absorbed by the phagocytes. There are many granular spheres surrounding the newly formed rectum. From this observation we may safely conclude that the mal- pighian tubes in the wall of the larval rectum are at first com- pletely destroyed, and only secondarily do the phagocytes act in this destruction, as in the cases with the silk glands and the salivary glands. Since Samson (’08) has not studied the histol- ysis of the included portion, it is natural that he came to his conclusion. In a pupa four days old, the muscle nuclei of the urinary bladder and the common stem have wholly disappeared through the action of the phagocytes and the imaginal cells have increased greatly in number. ‘The coelomic portion hes embedded among the fat-bodies of the ventral region. The cytoplasm is homo- geneous or granular, and stains deeply with eosin. It is thought that the function of the malpighian tubes is arrested in four-day pupa. There are many chromatin granules in the ramified nuclei which stain somewhat faintly with haema- toxylin. The lumen is, moreover, reduced in caliber and takes very irregular shape, varying greatly in different portions. The basement membrane now vanishes as a result of the activity of the phagocytes. The latter becomes enlarged; their contents assume a granular appearance. In a pupa five to six days old, the imaginal cells gradually migrate to form the imaginal common stem which opens directly to the alimentary canal, as shown in figure 12, and the urinary bladder is not newly formed. The coelomic portion is in the same position as on the preceding day. The cytoplasm is granular and contains many vacuoles which are basal in position (fig. 13). The branches of the nuclei are extended and the chromatin granules become very distinct in outline. The lumen increases in caliber and contains some granules near the wall. These granules are the result of the secretory activity of the cells, and they indicate that the formation of the striated border has already set in. A very thin transparent basement membrane appears, which may be considered to be the secretion from the cells. MALPIGHIAN TUBES OF BOMBYX MORI L. 203 In a pupa seven to eight days old, the structures of the cells in the coelomic portion are almost the same as on the preceding day. The lumen is, however, increased in caliber and contains numerous granules (fig. 14). In a pupa nine to ten days old, the formation of the common stem is already completed. The cytoplasm is granular and the nuclei are round or ellipsoidal and a thin new intima lies on the surface of the cells facing the lumen. In the coelomic portion the cytoplasm contains very many large vacuoles which are situated between the nucleus and the striated border. The ramified nucleus is basal in position and contains some small, highly refractive nucleoli. The chromatin is in the form of small granules or rods, and appears to be arranged surrounding the nucleolus. The striated border is composed of many very fine and closely set striations. The lumen is very irregular in shape, especially in the anterior portion. The process of the excretion is very active, as indicated by the homo- geneous or granular substance filling the lumen and staining light blue with haematoxylin. There are some vacuoles in the excreted substance of the posterior portion. The transparent basement membrane has increased in thickness. In a pupa eleven to thirteen days old, the common stem and the coelomic portion remain in the same histological condition as described above. The excretory activity of the cells has increased, and the excreted substance stains deeply blue with haematoxylin as in the larval stage. It need hardly to be mentioned that the imago emerges on the fourteenth or fifteenth day after the pupation. 4. Malpighian tubes in the imago In the imago the common stem opens directly at the junction of the stomach and the intestine, owing to the histolysis of the latter. The common stem is a short narrow tube and measures 0.61 mm. in length and 0.17 mm. in width. The epithelium is composed of large flat cells. The cytoplasm is granular and strongly eosinophilous. The nuclei are round or oval. The 204 HIROWO ITO intima is a thin chitinous membrane and is directly continuous with that of the intestine. The basement membrane is a trans parent structureless membrane surrounding the tube. The coelomic portion, which is provided with many fine tracheal branches, is convoluted in the ventral portion of the abdomen and ends blindly near the rectum. The distal end of the tube is very fine, and measures about 0.05 mm. in diameter. ‘The epithelial cells are arranged in a somewhat peculiar fashion; they are too large to admit of their forming a smooth lining, the oy) Fig. 2 Surface view of coelomic portion from a total preparation. X 50. cell of one side projecting between two adjacent cells of the other, so that the entire tube looks knotty, as is shown in figure 2, and the individual cells are somewhat polygonal. The cyto- plasm is granular or finely alveolar, and contains a great many large vacuoles. The nucleus is stained deeply with haema- toxylin, and some nuclei have already undergone fragmentation. In such cases the chromatin comes together to form a mass. The striated border is definitely recognized by its fine and closely set striations, as is shown in figure 15. The lumen is irregular in section, and is filled with spheroidal blue granules and yellow MALPIGHIAN TUBES OF BOMBYX MORI L. 205 erystals, very probably a urate salt. The basement membrane is thin, structureless, and transparent. SUMMARY 1. The histolysis of the urmary bladder is accomplished in such a way that the basement membrane and the intima are separated from the epithelial cells, with the destruction of the cytoplasm, and then both the nuclei and the basement membrane are ingested and digested by the phagocytes. The intima is east off from the lumen at the pupation. The muscular sub- stance and the muscle nuclei are completely destroyed by the phagocytes. 2. The histolysis of the common stem proceeds in a way similar to that of the urinary bladder. The histogenesis of the imaginal common stem takes place by the gradual extension of the imaginal cells which lie at the distal portion of the larval common stem. 3. The coelomic portion of the malpighian tube passes directly from the larval into the imaginal stage, though it undergoes the dissolution of the striated border and basement membrane. The imaginal striated border and basement membrane are formed by the secretion of the epithelial cells. 4. The portion of the malpighian tubes enclosed within the wall of the rectum is completely destroyed, along with the histolysis of the rectum. The nuclei break up into globules which are taken up by the phagocytes. The basement mem- brane is attacked and destroyed by the phagocytes. 5. The function of the malpighian tubes seems to cease during certain periods of metamorphosis. 206 HIROWO ITO BIBLIOGRAPHY Anatas, J. 1900 Observations sur les métamorphoses internes de la gtepe et de l’abeille. Bull. Sci. France et Belgique, T. 34. Borpas, L. 1910 Considérations générales sur les tubes de Malpighi des larves de Lepidoptéres. C. R. Acad. Sci., T. 150. 1911 L’Appareil digestif et les tubes de Malpighi des larves de Lepi- doptéres. Ann. sci. Nat., Zool., T. 14. CuotopKovsky, N. 1887 Sur la morphologie de l’appareil urinaire des Lepi- doptéres. Arch. de Biol., T. 6. DeEGENER, P. 1900 Entwicklung der Mundwerkzeuge und des Darmkanals von Hydrophilus. Zeit. f. wiss. Zool., Bd. 68. HurnaGet, ApA 1912 Métamorphose des tubes de Malpighi de l Hyponomeuta padellasiie IC. Re Soc Biol 73: Ikepa, E. 1913 Kaiko no Hentai Ron. (In Japanese.) Isuimorti, N. 1916 Les tubes de Malpighi 4 la paroi du rectum du ver & Soie. Bull. de l’Association Séricicole du Japon, no. 19. KaraAwalEw, W. 1898 Die nachembryonale Entwicklung von Lasius flavus. Zeit. f. wiss. Zool., Bd. 64. 1899 Uber Anatomie und Metamorphose des Darmkanals der Larve von Anobium paniceum. Biol. Centralb., Bd. 19. Mé&ratnikov, 8. 1908 Recherches expérimentales sur les chenilles de Galleria mellonella. Arch. de Zool. Expérim., Serie 4, T. 8. PéREZ, Cu. 1901 Histolyse des tubes de Malpighi et des glandes sericigénes chez la Fourmi rousse. Bull. Soc. Ent. France. 1910 Recherches histologiques sur la metamorphose des Muscides (Calliphora erythrocephala Mg.). Arch. de Zool. Expérim., Serie 5 aL. 4: Samson, Karu. 1908 Uber das Verhalten der Vasa Malpighii und die excre- torische Funktion der Fettzellen wiihrend der Metamorphose von Heterogenea limacodes Hufn. Zool. Jahrb., Abt. Anat., Bd. 26. ScH1nDLER, E. 1878 Beitrige zur Kenntniss der malpighischen Gefisse der Insecten. Zeit. f. wiss. Zool., Bd. 30. Vaney, C. 1902 Contributions a l’étude des larves et des métamorphoses des Diptéres. Ann. de l’Université de Lyon, Nouv. Sér. 1, Sciences Med. Fasc. 9. VENEZIANI, A. 1903 Note sulla struttura istologica e sul meccanismo d’escre- zione dei tubi di Malpighi. Monitore Zool. Italiano, Anno 14. 1903 a Intorno al numero dei tubi di Malphigi negli insetti. Comu- nicazione Accad. Sci. Mediche e Nat., 17. Nov. 1904 Valore morfologico e fisiologico dei tubi Malphigiani. Redia, vol. 2. — PLATES 207 PLATE 1 EXPLANATION OF FIGURES 3 Cross-section of the urinary bladder of a larva in the fifth stage. 400. 4 Cross-section of the coelomic portion of a larva in the fifth stage. X 220. 5 Longitudinal section of the common stem (cs.) and the urinary bladder (ub.), showing the multiplication of the imaginal cells. cp., coelomic portion. X 90. 6 0Z10'0 | 0000 | GS T:T | OOLL T | S6ET'T | Ss T:T | OOLS'T | OSTS'T skup % 0Z10°0 | £060°0 | 6O'1:1 | #200°T | OFZ6°O | SZ T:T | OSST'T | 09860 skep €] 0z10'0 | 0090°0 | 9L' 1:1 | OOgO'T | OS88°O | 8Z'T:T | OOST'T | 0006'0 sXep ZI 0910'0 | OFS0°0 | ZT 1:1 | OFZ8°O | OFOL'O | ETT | 0ZE6'0 | 00z2°0 sep 1 0910'0 | Z&90'0 | 2T' 1:1 | 9208°0 | 0989°0 | TE’ TT | O8T6'0 | 02020 skep OT 0020°0 | Z¢20°0 | FI 1:1 | 9FE9°O | O9G¢°0 | 9E°T:T | OS8Z'0 | 0929°0 skup 6 0020'0 | 0090°0 | GI’ 1:1 | OFG9°O | OSFS'O | 9E'T:T | OFLL°O | O89S'0 SAO pe SEE Sl geaskaghes see pecuemomede 0020'0 | 0zG0°0 | 92° 1:1 | O86¢°0 | OSZF°O | ZF'T:T | OZOL°0 | OS6F'O skup FL 0020°0 | 2270°0 | Ge'1:1 | epee'o | OFGE°O | ZF TT | 00E9'0 | OFIF'O skup tL 0080'0 | 0290'0 | 88° 1:1 | O80g'0 | 0998°0 | FF°T:T | OZI9'O | O96E°0 sXup ¢ €820'0 | 06F0'O | 28° T:1 | ZhOP'O | 2662'0 | GF T:T | Zz0g'0 | OFZE-O skup > 1120'0 | 0LF0'0 | 92'T:T | O8EE"O | 98920 | 9:1 | OZEF'O | 002z'0 sXup 0&20'0 | OFFO'O | IT'T:T | O¢9z'0 | O9ez'0 | TE’ T:T | OTSE"0 | 06Sz°0 skep % OSTO'0 0zLZ°0 | —Z:1 | 002g°0 | 006z°0 iskep $F OFE0"0 OFOL'O | OS 1:1 | O88z°O | 0O9T"O PAOUMOUL tlllWeectatts acne. Jeon sn 090° 0 0960°0 | ZE'1:1 | 0zEs'0 | O9LZT'O simoy 89 ae 00FO'0 0&90'0 | ZLT:T | OFFT'O | O8ZI°0 soy $9 LUVd UVvt UAAV'T “LuVd UV'T| -OOLDNGT ee LUVd YALAW VIG SNUT AO NOIMAINV | -ONNV 40 |MALAN VIG LAG UV TOOL Ivia YALAWVId Be son ee 10 SSa NMOIHL TVIt UGLAWVIA -NaA'T AO -OLVOO" Tvitt TVIXV adv §SSaN MOIHL Iviu -OLVO0"E AWIM UALANWVIG | OL TVIXV | -OLVOAdDa AVLOL IVIXV -OLVOOT | OL IVIXV -OLVada IVIXV 40 O1LLVY TV LOL 40 OLLVH SUOJIWY JUL UL ALD SJUuamMa.nsDa py ‘abDzs npp ay) 07 Buruurhag ay; wouf sua? burdopaaap ey} fo szund snorsva ay? fo suoisuamep ay} Ur asvaiour aarssaiboid ayy Burnoyy 6 WTA VL DEVELOPMENT OF EYE OF SPARROW 285 may be due to individual variation, to different effects of the reagents, or, more likely, to possible differences in the plane of section. The column giving the thickness of the anterior layer shows, after a slight increase, a progressive thinning. It is interesting to note that with this reduction in thickness the nuclei become less and less conspicuous. They either disappear (which would explain the thinning) or are rendered transparent. The annular pad increases in thickness by the lengthening of its cells. These cells are arranged radially at first and remain so until after the beginning of the lenticular chamber (text-fig. 6, Le). Soon after hatching the cells begin to show globular projections from their inner ends as in the adult (Slonaker, 718). The accumulation of these globules seems to exert a pressure between the lenticular portion and the annular pad. By the second day after hatching the lenticular chamber begins to appear slightly posterior to the middle of the annular pad (text-fig. 6, U). This occurs at the line of junction of the two portions of the embryonic lens. Separation occurs here because it is the place of least resistance. The cells showing the most prominent globular projections are adjacent to the forming chamber. All of these cells are still straight and converge toward the center of the lens. On the fourth day after hatching (V) the cells of the annular pad have become longer and are beginning to curve posteriorly. The lenticular chamber has increased and more cells show globu- lar secretions or projections. The lenticular chamber is appar- ently formed by the accumulation of this secretion. Since, in its formation, it follows the line of junction of the two formative layers, it may be said to correspond, in position at least, to the original lens cavity. With increasing age these cells become longer and more bent. Their nuclei remain in the peripheral ends, leaving the inner portions of the cells filled with granules. The lenticular chamber increases in size and seems to lag behind as the center of the lens grows forward. This anterior growth of the central part may in part be the cause of the bending of the cells of the annular pad. 286 JAMES ROLLIN SLONAKER The cells in the thickest portion of the annular pad are longest. The nuclei of these cells are arranged in a double row in the outer ends near the lens capsule. From this region they become gradually shorter, both anteriorly and posteriorly, and _ are arranged in a single layer and finally disappear from view. Toward the posterior portion of the annular pad the cells become gradually shorter and less granular until they have completely lost their identity as annular pad cells, at about the posterior border of the lenticular space. The nuclei are conspicuous in this region. ie Releiavesnt(s, Ding iielivieae dl a REESE Pet iy oy] aie iid! Ws heb ita 1 ¥ ann oud Sumy - CR Hales daly» {scan gitar! iM atuhgaeh, aaa BGT vig! a 2 a“ tah ais, care ay rs , Fad i) a) at 17 whe . * ; Resumen por el autor, Harry H. Charlton, Osborn Zoological Laboratory, Yale University. La espermatogénesis de Lepisma domestica. Lepisma domestica, aunque es un insecto primitivo presenta an proceso complicado durante la espermatogénesis. En la espermatogonia existen 34 cromosomas, pero en vez de encontrar 17 en el espermatocito primario, existen 18, puesto que dos de los cromosomas espermatogoniales no se unen. En la division que sigue, estos dos idiocromosomas pasan indivisos a uno de los polos, de tal modo que los espermatocitos secundarios poseen 16 y 18 cromosomas, respectivamente. Los mismos ntmeros existen en las espermdtidas al separarse los idiocromosomas durante esta divisi6n. E] autor ha seguido paso a paso la historia del centrosoma y su persistencia indica que es una estructura permanente, aun cuando puede cambiar en apariencia. En la espermatogonia se presenta en forma de grdnulo esférico; en el espermatocito primario exhibe forma de V, y en la figura de divisién del esperma- tocito secundario aparece en forma de bastoncito. Cada esper- matida posee uno de estos centrosomas, y de él nace el filamento axial. Mediante rotacién de la célula el centrosoma viene a tomar una posicién terminal y mas tarde forma el acrosoma, mientras que el filamento axial viene a ponerse en contacto secundario con el segmento intermedio. El desarrollo de este ultimo 6rgano no es completamente claro, pero al parecer deriva del nebenkern. La porcién del filamento axial situada entre el acrosoma y el segmento intermedio se transforma en la mem- brana ondulatoria del espermatozoide. El autor ha prestado alguna atecién a las mitocondrias, los restos fusoriales, la forma- cién del nebenkern y los cambios que experimenta. Translation by José F. Nonidez Cornell Medical College, New York AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, MARCH 28 THE SPERMATOGENESIS OF LEPISMA DOMESTICA HARRY H. CHARLTON Osborn Zoological Laboratory, Yale University SIX PLATES (NINETY-FIVE FIGURES) CONTENTS Nei TLOCUUMGETOM ES © acetate ans oes cy ateteversities stats, shel Sete eases) a cuercddee ease 381 (Miatemlalianel amet Od semerMysens, extens sie once oct sips, feusuey Vel earn Neemareretere tee crates waar 382 (OVeYS ETA AER EVO OSL paepin aiecaoGe icin cp ena SRtOe CHERRIES ic te Gut Se Gia Grea tame) A 384 Bxtennaesexuale chan ctehsattscrticiericsicorice tree eri acid seen: 384 Mialegreprod chives vst embers cris. theiss Soleierea ae ere Ionut cioers elo aicee 384 SPCTEMAL OPO ain, aes eee cena aay ciess wses iis a8 sic: oie eRe R eRe OILED LN oe AEN 384 PRE Rr gEO WOH WETIOG eek meena a oie ceieallen co oia:s wis! a ssa, Mat Reema el ea a oct oe 386 Myettirstispermatocy bey. tacks toe ciste swe a oicls de T ee ees oot Darts. OOo Resting stage of the second spermatocyte......2 2228) seek. ciel. 2-58 bah 390 shewsecondematunahlonedivislonenss ayer e eee ee eee Erector 391 The centrosome in the spermatogonial and maturation divisions..-...... 392 MHSLS Perna cid awe yess ast lw asdpde ate crhce « Mees Metis hve Seem AN Meo ae ore mcias 393 IRHEESPETIMATOZO AM ener. SANE acre eyeeie eae ei eee Rede usher uconnene 398 IM Isto el non avo Lists) S B's cp rican eve serene hen eres eH NPR EG a eee IAI ae Sein ts/3' She neh 401 DFS CUGSTOME PRISE aera ola rast oa olel sien e & sipnei ch oats Aes AUER sc RUMEN apa enc ones och 401 Resting stage of the second spermatocyte.:.20) 0.2... 5-. se. cee ete oe 401 RhesrarOchroMmOsOmeste nice sake Pees <6 crate eee coe ee 402 SVN Ast SeaM ene LLOM ts 575 1 Mercins sucks ere opi eters las syaceleg oer Pam sat cae sasyote = 403 AM OYE! CGEIMITRONOIIOG. 9 oe Si pao He ne ae OCIA Go sb Coa ae UIC aD Gain c Sei cole ae 404 AB) OVE BENG OYSKOMINES 3 «tere ercae pH ORD ERCP ROE A Te eet S rene ore RE RAELNS ae Rane elas 405 ihe Mmiddlesmve cence eae enneh res sake le caoeol AS Ra Tae ieee eo ia ase 406 ComparisonsratheOrthoptenannss ass come oe Aare eisee cls oe oe 406 SUMMA TY 5 S5. oe ae ck TN A ATE TET Cie Seehatach er oniia"al tebe ang Gide 408 SID IEG TAT DY, sc crave oer cae Ney oie also < “seer aaete dia ct shatestoiensbar ane eys,cvore ave old apepete 409 INTRODUCTION Cytologists have long found the class Insecta to be a very fertile field for investigation, but the first order and, according to many, the most primitive, namely, the Thysanura, seems to have been neglected. According to Harvey (’16), only three papers dealing in any way.. with the cytology of any closely 381 382 HARRY H. CHARLTON related forms have been published. The writers, Claypole (’98), on Anurida maritima, Lecaillon (’01), on Orchesella villosa, and Willem (’00), on Podura aquatica, simply report isolated obser- vations which are necessarily incomplete and limited to the class Collembola. The Lepismatoidea have therefore never been made the sub- ject of a cytological study, and it was in the hope that a survey of this primitive form would throw some light on the present-day cytological problems that this investigation was undertaken. The completion of the study shows that, instead of the expected simplicity, the process actually is a complicated one, differing only here and there from that already described in other forms. These differences, however, are interesting and, together with the fact that it is the first cytological work in a new class of insects, warrant its presentation. The work was done at the Osborn Zoological Laboratory at the suggestion of Professor Petrunkevitch beginning in the fall of 1916. During 1917-19 it was practically suspended except for an occasional day or so at Columbia University. It gives me pleasure to express my thanks to Prof. E. B. Wilson for his kindness in giving me laboratory privileges at Columbia and to Prof. Frank R. Lillie for facilities accorded at the Marine Bio- logical Laboratory, Woods Hole, Massachusetts, durmg the summer of 1919. Most of all, I am indebted to Prof. Alexander Petrunkevitch, first, for suggesting the problem and later for his unfailing help and criticism. MATERIAL AND METHODS Lepisma domestica, commonly called the fire-brat, from its frequently observed habit of running apparently unharmed over hot stones in bakeries, belongs to the class Thysanura of the order Lepismatoidea. It is a fairly common insect in New Haven, and can be kept alive in the laboratory for a considerable period. My method has been to keep them in large glass Stender dishes without covers, since the insects cannot climb up a clean glass wall, and to provide them with a cereal such as corn flakes SPERMATOGENESIS OF LEPISMA DOMESTICA 383 to eat. For moisture I have kept shallow dishes filled with moist filter-paper in the Stenders. In spite of these precautions, in the course of a month or so the creatures begin to show a shrinkage of the abdominal region and soon die with the posterior region shrunken fully one-half. The insects were either killed in xylol or decapitated and the testes dissected out immediately in physiological salt solution. As soon as the body cavity was opened, some fixing fluid was introduced by means of a fine pipette. This renders the tissues more opaque and makes it easier to locate the gonads which are loosely surrounded by fat, as well as to cause better fixation. | For counting chromosomes, Bouin’s fluid at 388°C. proved the best, but for general fixation of the cytoplasm as well as of chromatic structures nothing equaled Flemming’s strong solu- tion. In addition, Hermann’s fluid, Benda’s Flemming, 10 per cent formalin, Allen’s modification of Bouin, Petrunkevitch’s fluid, and Kopsch were tried out and had their special uses. The testes dissected out, fixed, washed, and dehydrated, were imbedded in paraffin and cut into sections from 3 to 12 u thick. Sections of 7 » thickness were found to be very satisfactory for study, and in general this was the thickness used. The stain used generally throughout was Heidenhain’s iron haematoxylin without any counterstain. In addition, various counterstains were tried, and also, in an effort to get a selective stain for mitochondria, Benda’s alizarin-crystal violet method. A modification of Cajal’s silver-impregnation method by Hortega (16), especially recommended for conor was given an inadequate trial with but fair results. I have also examined a number of splended ities stained by the safranin-gentian-violet-orange G method, for which I am indebted to Dr. P. W. Whiting. 384 HARRY H. CHARLTON OBSERVATIONS External sexual characters During the spring months of March, April, and May, the adult insects are in the best condition for study, but since it takes some time, probably a year or more, to attain sexual maturity, the early stages may be studied at any time of year in young individuals. It is comparatively easy to recognize the sexes by their external appearance. The female (fig. 2) has a long median ovipositor extending posteriorly which is quite prominent in the living insect. There is nothing comparable to it in the male (fig. 1), for the penis which could not possibly be confused with the ovipositor is more often retracted and not in view. Male reproductive system The testes (fig. 3), of which there are three pairs on each side of the middorsal line, occupy, in mature individuals, a consider- able portion of the anterior two-thirds of the abdomen. ‘The testes lie parallel to each other, extending in a ventroposterior direction, and each is connected by a short duct with the vas deferens, which passes as a straight duct posteriorly where it enlarges to form the seminal vesicle. From the seminal vesicle a similar duct extends, which soon enlarges considerably and, after bending upon itself a couple of times, opens into the base of the penis. Spermatogonra I have made a long and careful search for primary spermato- gonia in the youngest material at my disposal, which consisted of insects only 2 or 3 mm. long, but have not been able definitely to identify them. It is therefore very probable that the primary spermatogonia occur only very early in the life-history. Mun- son (’06) has described an apical cell which produces early spermatogonia, but this, too, if present at all, would be found in exceedingly young individuals. SPERMATOGENESIS OF LEPISMA DOMESTICA 385 One does find cells in quite large numbers at the blind end of the testis, which differ from the ordinary spermatogonia in having large homogeneous nuclei with the chromatin condensed into a single dense mass and irregular in shape (fig. 28). I believe these to be immature Sertoli or nurse cells, for later on one finds such cells, only now they are larger, more elongated, and contain two to four chromatic bodies. Those at the region of the mature spermatozoa are much paler in color and may be wrinkled and twisted upon themselves, indicating perhaps degeneration. The spermatogonia occupy a considerable part of the blind end of a tubule of the mature insect during the winter and early spring, and can be easily recognized by their position and by the arrangement of the chromatin in the form of clumps attached to each other by linin threads and grouped around the periphery of the nucleus. This arrangement (fig. 4 is a surface view) is the most common and probably represents a resting condition. Although I have not been able to count these clumps of chro- matin, the number is easily seen to be more than the haploid, and each one probably represents a spermatogonial chromosome. The nucleus of the early spermatogonium is quite large and almost equal in size to the nucleus of the growth period. The two or three spermatogonial divisions reduce the nuclear and cell size by apparently not allowing time for growth between divisions (figs. 8 to 11). In prophase the chromosomes are long and bent upon themselves and irregularly scattered throughout the nucleus; later they are drawn into the metaphase plate as shown in figure 5. It is only in the larger and therefore the earlier spermatogonia that good counts of the chromosomes can be made. Figure 6 shows thirty-four chromosomes in a very clear metaphase plate. The chromosomes are of the curved-rod type, differmg con- siderably in size, but close observation fails to show any chromosome or group of chromosomes behaving in any way differently from its neighbors. ' In the telophase of the spermatogonial division the chromatin becomes granular and forms a more or less eccentric ring around 386 HARRY H. CHARLTON the nuclear wall (figs. 11 and 16). At a little later stage one commonly finds an irregular clump of chromatic material repre- senting apparently two spermatogonial chromosomes lying against the nuclear membrane and retaining the haematoxylin stain (figs. 13, 14, and 15). For these chromosomes I shall use the term idiochromosomes, the name given by Wilson (’05) and meaning ‘peculiar or distinctive chromosomes.’ In some presumably young cysts the spermatogonial cells are arranged in the form of a rosette, their median ends tapering toward a faintly marked open center and showing an archo- plasmic mass or sphere (figs. 4 and 18). Hegner (’14), Meves (97), and Shaffer (17), as well as others, have described similar structures and considered them spindle remains. In some cases they figure them as extending from cell to cell. While this is true immediately after division when the spindle remains are very definite (fig. 7), it is not possible later to see any continu- ation or connection with similar bodies in adjacent cells. In the spermatogonial region isolated cells are occasionally seen in division, but, strangely enough, the chromosomes are paired and look somewhat like tetrads (fig. 32). The cells themselves are much larger than the spermatogonia and contain but little cytoplasmic staining material in the form of a flaky mass at either end of the cell in which a dark-staining granule may be seen. If these represent division in the Sertoli cells, they are a very rare occurrence. In the older Sertoli cells I have occasionally seen evidence indicating division by amitosis. The growth period The stages of the growth period correspond fairly closely with the stages described by Wilson (12). After the telophase, chromosomes of the last spermatogonial division break up and form a granular ring just inside the nuclear wall, the chromatin arranges itself as previously described in the form of clumps located on the nuclear membrane (Wilson’s stage b, similar to fig. 13). In heavily destained material two of these are closely’ related, one of them being flattened against the nuclear mem- SPERMATOGENESIS OF LEPISMA DOMESTICA 387 brane and retaining the dark stain of the haematoxylin (fig. 15). In addition to the idiochromosomes, a similarly staining, small, spherical granule appears (fig. 13). The chromatin clumps now become granular and form an eccentric circle against the nuclear membrane, leaving an open center very much like the condition following the last division. The homogeneous granular border is at first deeply stained, but later loses its affinity for the haema- toxylin and appears pale in color (figs. 11 and 16). The idiochromosomes also seem to break up into unequal spherical bodies, three to eight in number, six being the more common number (fig. 16). In the clear central region the remains of the preceding spindle are quite apparent. Following this stage we have the reappearance of the idiochromosomes (fig. 17), and after that the entire nucleus appears granular, the central clear area disappearing and the two idiochromosomes stand out clearly (fig. 21). It has not been possible to see anything like an unraveling stage as described by Wilson (12) for stage c; the granular con- dition being directly followed by delicate threads (Wilson, stage d, fig. 22) which seem to push out and distort or break the nuclear wall. This is soon followed by the synizesis or con- traction stage. Here the threads are drawn closely together and are located more to one side of the nucleus, the plasmosome and idiochromosome thread often remain outside of the con- tracted mass, as shown in figure 23. Popoff (08), Gates (’08), and Whiting (17) look upon this as due to a rapid absorption of water by the nucleus; in other words, an osmotic effect; however, it has often been considered an artifact. Although at this stage of the growth period the spireme threads stain very intensely, making it difficult to trace the individual threads, it would look as though the filaments became arranged in the form of loops polarized with their free ends near the plasmosome and idiochromosome threads. Later on when the threads have thickened, this bouquet stage is much . more clearly seen (fig. 25). It has not been possible to see a side-by-side union of the spireme threads, the synapsis of Moore (’95), but the number 388 HARRY H. CHARLTON of filaments certainly is reduced and each one becomes much thicker. The threads now loosen up and occupy practically all the cell, the space between the nuclear membrane and the cell wall being quite small (fig. 24, Wilson (12), stage f). I have not been able to find the longitudinal splitting of the thread —a process which Wilson (’12) describes as taking place. There follows a period when it is hard to distinguish the threads as such (fig. 29, Wilson (12), stage g). Wilson calls it a net-like arrangement. The actual breaking of the threads or pachytene stage is not well exemplified in Lepisma, but stage g is soon followed by the clumping of. the chromatin into masses irregular in shape and joined together by linin threads (fig. 30). By a further condensation of these masses we get the prochromosomes. The formation of tetrads showing the quadrivalent condition of the autochromosomes is never apparent, neither is there any split indicating a parasynapsis. The idiochromosomes retain their form and staining reaction until the formation of the delicate filaments (stage d), when they break up and form threads which are darker in color than the other threads, and one may be seen in close relation to a small plasmosome (fig. 35 a). The idiochromosome threads are at first very long and may extend across the entire width of the cell. They appear somewhat beaded, just as is the case with the threads of the autochromosomes. During the later periods the threads show an end-to-end appo- sition, being joined by very fine linin fibers (fig. 35 m). The threads now become shorter and thicker assuming the U shape followed by the definite formation of loops with the plasmosome between them (figs. 20 and 35 j). Figures 35 i and h would seem to indicate that the limbs of the loop come together and become still more compact to form clumps lying against the nuclear membrane with the plasmosome still lying between them. For a considerable time the idiochromosome threads show a very clear inequality in that the thread nearest to the plasmosome is the longer (fig. 35 k). A second small plasmosome may be formed and lies to one side with no attachment to either idiochromosome thread (fig. SPERMATOGENESIS OF LEPISMA DOMESTICA 389 35eandf). Later I believe it fuses with the first, for the latter is seen to increase considerably in size and to show at times a double nature (fig. 35 g and k). A third body similar in shape and staining reaction to that seen in the spermatogonia becomes quite prominent at this time (figs. 29 and 30), due to a slight increase in size and to the appear- ance of a clear transparent area encircling it. Painter (14) describes in spiders similar small dark-staining spherical bodies, which he calls planosomes and which first make their appearance in the late spireme stage and which he was able to trace through the succeeding divisions. The planosomes, according to him, have spindle fibers, and would therefore be comparable to chromosomes, although as a rule they do not divide, but linger near the middle of the spindle and later go to one side. From his description and figures, this body is the same as the one found in Lepisma domestica, only I find it first in the resting stages of the spermatogonia, and have not been able to follow it beyond the prophase stage of the first maturation division. The first spermatocyte With the condensation of the chromatin segments ‘nto the prochromosomes, the nuclear membrane breaks down and two chromosomes located near the periphery are seen joined together by a more or less ribbon-like connection, forming a V-shaped structure. Within or near the arms of the V the plasmosome may be found (figs. 36 and 37). With the exception of the prophase figures in which the idiochromosomes stain more deeply, there is no essential difference in the staining reaction of the idiochromosomes and the autochromosomes; but to make the behavior of the idiochromosomes plain throughout the different stages of the first maturation division, they have been drawn in black, while only the outlines of the autochromosomes are shown (figs. 36, 37, 41, 42, 48, 44, and 45). The chromosomes arrange themselves on the spindle and in the metaphase plate (figs. 38, 39, and 40), the sex or idiochromo- somes are still connected and one pair of the chromosomes is a 390 HARRY H. CHARLTON little further beyond the metaphase plate, so that in plate view one pair of chromosomes can be seen to be at a different level (figs. 89 and 40). The side view shows how one limb of the V extends farther than the other. The metaphase plate (fig. 38), in which the idiochromosomes are located in the center and surrounded by a ring of chromo- somes, reminds one of the arrangement in some Hemiptera. There is little change in the position of these joined chromo- somes in the anaphase (fig. 41), except for a shortening of the connecting thread and possibly a slight movement of the whole toward the distal pole. Figures 42, 43, and 46 picture the telophase arrangement, the idiochromosomes going undivided to one pole. There are sixteen chromosomes plus the two idiochromosomes, or eighteen in all, in the first spermatocyte division. Side views have not been counted, owing to the great overlapping of the chromosomes. The plasmosome may be identified during the late prophase (fig. 37), but not definitely after the actual spindle formation. Bodies which are plainly not chromosomes are often seen in relation to the spindle, as the two equal bodies in figure 40, but whether these represent the divided plasmosome or are mitochondrial is not conclusive. Resting stage of second spermatocyte In the telophase of the first or early prophase of the second spermatocyte (fig. 46), the chromosomes are breaking up. Some appear unchanged, while others have swollen to a spherical shape and stain more diffusely. It is not possible to identify the idiochromosomes at this time, but a little later, when the resting nuclear stage is reached, the double nature of the idiochromo- somes is quite apparent as the nucleolus in one of the now divided cells (figs. 47, 48, 50, and 51). It is not possible to confuse these resting second spermatocytes with the early spermatids, because both nuclear and cell size is much larger. The relative sizes of first and second spermatocytes and spermatids are shown in figure 33, which was diagrammed from measurements of the length SPERMATOGENESIS OF LEPISMA DOMESTICA 391 and breadth of ten representative cells of each kind, and the average diameter of each cell and of the ten cells taken. In the second place, the chromatic nucleolus is distinctly double (fig. 47), while in the spermatid it is single and smaller (figs. 64, 66, and 68). During the growth and division period, spindle remains stand out quite clearly as one, more usually as two vesicles, formed probably from the central fibers and showing a granular con- densation in their interior (figs. 25 and 46). The formation of the resting stage and the subsequent prophase is a rapid one, as I have observed resting nuclei, prophase, and dividing second spermatocytes in the same cyst. Figures 50 and 51, resting and prophase stages, respectively, are from a slide not particularly well fixed, as the cells are somewhat swollen, but figure 51 is interesting in that it shows the formation of spindle fibers before the nuclear wall has broken down and in figure 50 the idiochromosomes still show their double structure. The resting nucleus, at first granular, breaks up into faintly staining irregular or fantastically shaped entities without any visible unraveling stage and condense quickly into the DEG: chromosomes (figs. 49 and 50). The second maturation division With the formation of the spindle for the second maturation division, two types of metaphase plates are seen: one (fig. 53) with eighteen chromosomes and another (fig. 56) with sixteen. In the latter case I have one perfect anaphase (fig. 59), in which both plates can be counted and both show sixteen chromosomes. It appears that the idiochromosomes are now equal in size and no longer show a connecting thread. In the first maturation division the idiochromosomes were distinctly unequal, but each tapered into a thread connecting it with the other. This thread often seemed ribbon-like, granular, and taking the iron haema- toxylin stain like the chromosomes. It seems to attain its maximum length at the metaphase of the first maturation division and to shorten a great deal by the 392 HARRY H. CHARLTON time the telophase is reached, and it would appear as though this thread were fused with the smaller idiochromosome so that they both appear equal in the metaphase of the second maturation division. Another factor in favor of this hypothesis is that the chromatic nucleolus of the resting stage shows a double structure with hardly any inequality. In the early anaphase (fig. 55) all the chromosomes show a longitudinal split near their centers, except two which represent, I believe, the divided idiochromosomes. The anaphase often shows the chromosomes arranged in the form of a ring (fig. 62). In figure 63 the chromosomes are at the poles and are beginning to form a nuclear membrane, but no change has taken place in the centrosomes. Figure 60, a late telophase, shows that one chromosome differs from the rest in being elliptical, while the others are V- or U-shaped and slender. A still later telophase is figured in figure 61, the chromatin now being massed at the poles. Two types of spermatids are formed, those with sixteen and eighteen chromosomes, respectively. The centrosome in the spermatogonial and maturation divisions In the archoplasmic mass or sphere representing the remains of the previous spindle one may occasionally see two dark gran- ules (fig. 18), which I take to be the divided centrosomes. In the division figures of the spermatogonia centrosomes are difficult of demonstration, but in a few slides I can make them out as definite single granules at the poles of the spindle (fig. 12). I have never seen astral rays or anything comparable to a centro- sphere at the time of division, but during the resting stages the centrosome is found in a granular sphere. From the division figure of the last spermatogonial mitosis. until shortly before the synaptic or contraction stage, the centro- some has not been traced, and when it does appear a considerable metamorphosis has taken place. At about the time when the fine spireme threads are being changed into loops, a granular mass can be made out at one end of the cell, and in this mass appear two short, stubby rods lying parallel to each other. SPERMATOGENESIS OF LEPISMA DOMESTICA 393 Later (figs. 20 and 26) the rods lengthen and show small granules at their ends. At first the two rods form an angle of 180°, but this angle is later decreased to 90° or less. Each rod now divides, but the halves remain attached by their granule ends, forming a pair of V-shaped centrosomes, each V representing a divided centrosome. This whole process is a rapid one, for all stages as well as the separation of the V’s for some distance may be seen in cells which show little change otherwise. The migration is about completed and the V’s nearly at the poles by the time the prophase condition is reached (fig. 30). During the succeeding division the apex of the V is directed toward the chromosomes, while its limbs touch the surface of the cell. The V may open considerably, nearly to a straight angle, so that a large part of the outer surface of the rods is in contact with the cell wall. The cells may also show a slight depression at the poles (fig. 39). The spindle fibers all lead to the centrosome region, but an actual attachment of the fibers to the centrosomes, while taken for granted, does not show clearly in sections. This V arrangement can be identified up to a late telophase of the primary spermatocyte, but I have not traced it through the resting stage of the second spermatocyte. Each second sperma- tocyte would receive one V, but when the rods reappear in the division figure they are divided, a single rod at either pole lying ‘against the inner surface of the cell wall and oriented parallel to each other, but at a slight angle with the cell axis. The division or separation of the V’s as well as their migration to opposite poles must take place during the resting period. The centrosome rod can be traced through every succeeding stage to the early spermatid, where it may be seen lying free in the cytoplasm (fig. 65). In exceptional cases, as in figure 54, the rods have granules at their ends, or we may find a number of granules or fragments and no rod, as in figure 58. The spermatid The young spermatid cell is considerably smaller than the resting stage of the second spermatocyte. The chromosomes clump together, form a nuclear membrane, and quickly break 394 HARRY H. CHARLTON up. The nucleus appears round in polar view, but oval if looked at from the side. Later the nucleus becomes spherical, the chromatin appearing finely granular and congregated at the boundaries of the nucleus leaving an open center (fig. 66). One- half the cells show an idiochromosome nucleolus which usually presents a spherical part extending into the nuclear cavity and a flattened area against the inner surface of the nuclear mem- brane, while the other cells do not possess an idiochromosome nucleolus. The methods of fixation and staining have a great deal to do vith the structures observed in the spermatid. When strong Flemming is used for fixing followed by Heidenhain’s iron haematoxylin, the cytoplasm of the early spermatid contains such a mass of intensely staining material that the nuclear membrane is made out only with difficulty. The same stain after Bouin’s fluid brings out the nucleus and centrosomes, but not the cell inclusions. At the very first, the cytoplasmic structures are somewhat loosely aggregated around the nucleus, but particularly between the nucleus and the last division plane. The centrosome can easily be followed from the telophase; located at first on the cell wall of the dividing second spermatocyte, it later moves inward, occupying the space between the cell wall and the nucleus (fig. 65). As it moves around to get between the nebenkern and the nucleus, it turns 90° and comes to lie with one end on the nuclear membrane and the other against or near the cell wall (figs. 66 and 67). The rod-shaped centrosome now frequently shows a granule or enlargement at the nuclear end. The nebenkern has meanwhile formed a broad ring of densely staining granular material in the center of which spindle remains of the last division appear and on either side two spherical bodies become visible (fig. 69), exactly like those seen in the two matura- tion divisions, and undoubtedly represent old spindles. In cross-section they appear as rings with their boundaries staming in varying degrees, often looking like crescents, and may possess a darker staining center. Looked at from the side, they take the form of rods with faintly stained material between them. SPERMATOGENESIS OF LEPISMA DOMESTICA 395 When the centrosome is in contact with the nuclear wall, one usually sees a granule at its inner end (fig. 66), and later a granule similar in size located near it on the membrane (fig. 67). This suggests the breaking away or division of the granule at the base of the centrosome. The centrosome may now change its position, being found in the region of the nebenkern or even at the opposite side, and shortly the delicate axial filament is seen pushing from the cell and occasionally carrying a small clump of cytoplasm with it, very much as has been described by Buder (715) in the Lepi- doptera and called by him ‘plasmaklumpchen.’ The single granule arising from the centrosome increases considerably in size and divides, giving rise to two granules which move apart and come to lie against the nuclear membrane and closely applied to it (figs. 65, 70, 71, 72, and 73). About this time or a little later a somewhat larger, round body condenses out of the nebenkern ring, as shown in figure 72. Outside of the breaking up of the idiochromosomes and a slight tendency to become pale and homogeneous, the nucleus remains the same during the above changes in the cytoplasmic inclusions. In the stage which follows, the delicate axial fila- ment is quite obvious and its outgrowth from the distal end of the rod centrosome is very clear. The rod has swung so that now it is in contact with the nucleus throughout its entire length and the thread is seen traversing the space between nucleus and cell wall (fig. 74). The nucleus contains numerous dark-staining granules. The spindle remains are prominent, their borders have increased in thickness, and now appear as irregular-shaped thick- walled vesicles. The nebenkern ring, cleared of the spindle remains and of the various granules as well as of several aggre- gations of granular mitochondria, now rounds itself up into an oval-shaped dense mass which later becomes round (fig. 76). No structure is at first apparent except a heavily stained granular body, but one soon sees a vacuolization of its border and we get the rosette nebenkern of many writers (fig. 74). Of the three granules already mentioned, those arising from the centrosome have either disappeared or have become so 396 HARRY H. CHARLTON closely adherent to the nuclear membrane as to seem a part of it, while on the other hand the other body appears slightly larger (figs. 77 and 78). The nucleus continues to stain darker, due to the enlargement of the chromatin granules, and these may become joined to each other and give the appearance of short threads (fig. 75). The vacuolization of the nebenkern continues at the expense of the central body which becomes smaller. The walls of the pe- ripheral vacuoles break down, the spaces becoming larger and larger, until there is but one vacuole, which may exceed even the nucleus in size, with a small heavily staining central part (figs. 80 and 81). From this period on, the axial filament is in close relation to the central body of the nebenkern, which in well-fixed material is now seen to be made up of a spireme-like thread. I have been able to follow it throughout the greater part of its course and I feel almost certain that it is a single continuous thread (fig. 80). The cell now begins to lengthen somewhat and the central part of the nucleus to stain heavily, the chromatin moving toward the nuclear center, leaving a clear transparent border (figs. 78 to 83). Unless one is fortunate with his fixation and staining, the central part of the nebenkern shows no structure, but appears as a glassy elliptical body suspended in the single large vacuole by means of the axial thread, but it can be seen very clearly that the tail filament never enters the central body, but comes to lie against it. The vacuole membrane lengthens out as it increases in size, while the central thread-like structure breaks up into several large and many small vesicles (figs. 81 and 82). There are also mitochondria-like structures located between the nebenkern and the nucleus as well as some distal to the nebenkern. The nebenkern membrane forms apparently the sheath of the axial thread, some cytoplasm forming clumps around the distal part of the thread, but the vesicles in large numbers fill the spaces between the spermatozoa as they increase in length. The middle-piece anlage enlarges, and by a turning of the nucleus, the axial filament comes to lie against it (figs. 85, 86, 88, and 90). SPERMATOGENESIS OF LEPISMA DOMESTICA 397 The body then flattens out against the nucleus and later elon- gates slightly (figs. 85 and 90). The nucleus lengthens, and as it does, the axial filament between the middlepiece and the centrosome does likewise. The centrosome, however, is now at the apex of the nucleus and will hereafter be considered as the acrosome. At the time when the nebenkern membrane and its vesicles have completely broken up and are only apparent as end products ensheathing the elongated tails or located between the filaments, the nucleus is still spherical, compact, and does not take the haematoxylin stain very well (fig. 82). It still has a clear area about it and some mitochondrial material about the middle- piece anlage, which is a quite prominent body located usually on the opposite side of the nucleus from the acrosome. The axial filament arising from the acrosome at the apex of the nucleus is bent backward and passes near the middle-piece anlage. The further changes are the loss of the clear ring about the nuclear chromatin by the spreading out of the chromatic material. The nucleus shows better staining qualities. The axial filament comes to lie nearer to the middle-piece and the latter may some- times show one or more bubbles or vesicles, which are possibly mitochondrial, in relation to it (figs. 84, 85, and 86). The nucleus now begins slowly to elongate and seems some- times to pull away from the acrosome, so that part of the latter body may project beyond the nucleus. The nucleus continues to lengthen, the chromatin to appear paler in color. The mid- dle-piece enlarges and elongates (figs. 88, 89, 90, and 91). The axial filament strand between the acrosome and the middle-piece remains applied against the nucleus and sometimes may show one or several splits in the thread, leaving an elliptical opening. In material stained for mitochondria, a cloud of granules seems to gather about the thread (fig. 87). Although it cannot be traced directly, I am of the opinion that this axial filament, which is loosely applied to the outer nuclear surface, becomes the undulating membrane of the mature spermatozoon. 398 HARRY H. CHARLTON Thompson (717, p. 267) suggests the formation of the undulat- ing membrane from a free flagellum in the Trypanosomes, as follows: “It is a plausible assumption to suppose that, as the flagellum waves about it comes to lie near and parallel to the body of the cell, and that the frill or undulating membrane is formed by the clear fluid protoplasm of the surface layer spring- ing up in a film to run up and along the flagellum, just as a soap- film would be formed in similar circumstances.’ Of course the axial filament in this case is located between the nucleus and the outer cell wall, but it seems a reasonable hypothesis to think of the axial filament as having become loose from the nucleus and as able to draw out the thin layer of cytoplasm some little distance from the nucleus forming the undulating membrane. The nucleus and middle-piece now become drawn out to considerable length, the acrosome decreases in size and we see a slight projection of the nucleus extending beyond the acrosome. The elongated nucleus stains darker and darker until no structure can be made out. From this point until the mature spermatozoa are reached I have not been able to make observations (figs. 93 and 94). The spermatozoa A study of the mature spermatozoa has been made by teasing the contents of the seminal vesicle in a minute quantity of physi- ological salt solution and either studying them alive in the solu- tion or fixing the teased material in osmic acid fumes, hot corrosive sublimate, Bouin or strong Flemming, and staining. The unstained living contents make an interesting study when examined by means of the dark-field microscope. In addition to the spermatozoa with their waves of movement extending from anterior to posterior end of the undulation membrane, there are a large number of small elliptical bodies performing active brownian movement. I thought at first that these bodies were the true spermatozoa and the others giant spermatozoa, but further study convinced me that this was not the case, for no tails could be found upon the small bodies, they stained only by plasma stains, and furthermore no stages in their development could be made out. SPERMATOGENESIS OF LEPISMA DOMESTICA 399 Occasionally the lumen of the vas deferens is partially filled with granules differing in size, and as the cells lining the vas deferens may show similar granules in their cytoplasm, I have considered the granular material of the lumen to be secretion products of the cells. Although the bodies present in the seminal vesicle are a little longer than broad and show a little difference in their size relations, yet I think they represent the secretion found in the vas deferens. Munson (’06) considers the epithelial cells of the vas deferens of the butterfly Papilio to have a secretory function, but unfortunately he does not figure or describe the process. An added fact of interest is that the bodies are transmitted during copulation and are found in the seminal receptaculum of the female, which suggests that they have some function yet unknown to us. To ascertain their nature, smears of the seminal vesicle were made, fixed in hot corrosive sublimate, and stained in orcein, safranin and orange G, Delafield’s haematoxylin and orange G, safranin and malachite green, safranin and bleu de Lyon, Dela- field’s haematoxylin and erythrosin. In every case it was found that the granules took only the cytoplasmic stains. In order to test the possibility of the granules’ being of a fatty nature, fresh smears were stained in sudan III, but the result was negative. Smears treated with ether showed also no effect of the latter on the granules. When the spermatozoa are examined in the fixed and stained condition (text fig. A and fig. 95), one finds a long chromatin staining thread ending in a transparent fine point, the acrosome having disappeared, and extending from near the apex a con- spicuous undulating membrane. The free edge seems formed of a little denser material and represents in all probability the proximal part of the axial filament, i.e., that part between the acrosome and the middle-piece. It is almost impossible to see just where this membrane leaves off distally as it gets narrower gradually, but I should say that about the anterior two-thirds of the spermatozoon is provided with the membrane. It is not possible to find any trace of the 400 HARRY H. CHARLTON middle-piece or to see where the nucleus leaves off and the tail filament begins, as the latter structure becomes finer and more transparent until it is almost impossible to see where it ends. BR is / [ es aS eee ace — Ley / : rae pate ; i | RED ree 3 oe ses Cf ya X fj PP , Cp = 2 SS ee era OE Text fig. A Mature spermatozoon of Lepisma domestica arranged from suc- cessive camera-lucida drawings. X 2800. I have noted a tendency for the tails of the living spermatozoa to stick together. The spermatozoa are very long, measuring from 400 to 660z. SPERMATOGENESIS OF LEPISMA DOMESTICA 401 Mitochondria Mitochondrial structures are present in the spermatogonia, but in very small numbers, and it is difficult to make them out. At the beginning of the growth period they appear clearly as a dark-staining crescent-shaped mass usually located at one end of the cell. This mass soon breaks up and forms some six to eight bodies differing from each other in shape and size (fig. 21), but retaining an almost constant number. These bodies take the stain intensely and appear as the most prominent structures during the entire growth period. As the cell increases in size, the mitochondria becomes so conspicuous, even by ordinary iron-haematoxylin staining, that the cytoplasm seems like a dark border about the nucleus, and in this dark-staining mito- chondrial matrix the clumps stand out clearly. During the first maturation division the ring of granular mitochondria encircles the entire spindle, while the larger bodies are scattered about the cytoplasm and are located near but not on the spindle. The mitochondrial material seems to divide equally, half the granular material as well as four clumps going to each cell (fig. 39). In the resting stage of the second sperma- tocyte the mitochondria forms a narrow dark-staining ring about the nucleus, but the clumps are no longer apparent (figs. 49,50, and 51). In figure 54 the granular mass is seen arranged about the spindle of the second division, while figure 61 indicates how they gather about the chromatin at the poles of the young spermatids. In the description of the spermatids the further history of the mitochondria has already been given. DISCUSSION The resting stage While in many animals no resting nucleus is formed following the first maturation division, the chromosomes of the telophase being quickly transformed into the prophases of the second division, there are quite a number of exceptions reported in the literature. 402 HARRY H. CHARLTON Murray (’98) finds a well-marked resting nucleus in the Pul- monates, Helix and Arion. McGill (04) found it to happen occasionally in the dragon-flies, and Painter (’14) describes it as occurring in the spiders with the accessory chromosome persisting as a nucleolus. Kingsbury (’01) finds in the salamander Desmognathus fusea, that a nuclear membrane is formed follow- ing the first maturation division, but that the chromosomes never lose their individuality. In Lepisma domestica the chromosomes, with the exception of the idiochromosomes, entirely break up and a nuclear mem- brane is formed. While it is undoubtedly of short duration, still the outward individuality of the autochromosomes is lost and the second division is preceded by their reformation. The 1diochromosomes Wilson (’09) divides the sexual differences of the chromosomes into five and possibly seven types. Lepisma domestica falls in line with his type IV in which ‘‘the male has a pair of idiochro- mosomes, half the spermatozoa receiving both and hence two more than the other half.” Only one form has been found which has this arrangement, the coreid species Syromastes marginalis L. This form was first described by Gross (’04) and again by Wilson (’09). The accessory chromosome arises by a synapsis of two spermato- gonial chromosomes which divide equationally in the first spermatocyte, but fail to divide in the second. Lepisma domestica differs in that the two spermatogonial chromosomes do not fuse, but remain separate and joined by a stout thread. They pass undivided to one pole in the first spermatocyte division, but separate in the second. Wilson’s prediction that the female of Syromastes would have two more chromosomes than the male, he afterward found to be the case. _ Reasoning in a similar manner, Lepisma domestica females should have thirty-six chromosomes, but unfortunately I have been unable to make any chromosome counts so far in the female. SPERMATOGENESIS OF LEPISMA DOMESTICA 403 Synapsis and reduction It has not been possible in the ordinary chromosomes or auto- chromosomes to see whether there is either a side-by-side union of the spireme threads, a parasynapsis, or an end-to-end conju- gation, a telosynapsis. It is clear, however, that the spireme threads in postsynaptic stages are much thicker and are present in fewer numbers. Whether they are half the leptotene number or not could not be made out. In the case of the idiochromosomes the conclusions are clearer. Each idiochromosome breaks up into a spireme thread and the two threads eventually unite end to end, one of them being attached to a large plasmosome. From these threads two chromosomes are formed by the condensation of the chromatin, but they still remain united by a thread which is probably linin in nature and along which, when the thread lengthens, the chromatin is drawn out. Synapsis, or a side-by-side conjugation, if it takes place at all, does so following the telophase of the first maturation division. That the idiochromosomes do come into a very close relation is shown by the longitudinal split apparent in the idichromosome nucleolus of the resting nuclei of the second spermatocyte (figs. 47 and 50). If, as is generally conceded, the spermatogonial chromosomes represent two groups, one of maternal and the other of paternal chromosomes, and the homologous pairs conjugate at synapsis then each of the idiochromosomes represents one spermatogonial chromosome. 100. ABBREVIATIONS br.I, branchial arch I po.pr., postorbital process c.c., copula communis po.pr., postorbital process ct.h., ceratohyal cartilage ps., parasphenoid ossification f.hyp., fenestra hypophyseos rect.m., recti eye muscles f.my.vent., fenestra myodomus ven- — styl.h., stylohyal cartilage tralis sym., symplectic cartilage hyom., hyomandibula taen.tect.med., taenia tectum medialis hyom.VII, ramus hyomandibularis faci- _ trab.cr., trabecula ecranii alis ven.j., vena jugularis hyp., hypophysis cerebri As the divergence of the trabeculae becomes greater, the parasphenoid lamella remains as a small spicule in the median part of the fenestra and is widely separated from the trabeculae cranil. The gasserian ganglion lies in a space between each CHONDROCRANIUM OF SYNGNATHUS FUSCUS 433 trabecula cranii and the postorbital process of that side. Swin- nerton indicates this relation of the trigeminal ganglion to the trabecula and the postorbital process in Gasterosteus. Gaupp calls this space in a 25-mm. Salmo the incisivum prooticum. Farther posterior, the prootic process of each side abuts against the lateral surface of the trabecula craniu (figs. 1, 2). The cartilage forming each of these structures retains its identity at the point of union. ‘The perichondria of the two cartilages form the line of separation between them. In this immediate region the trabeculae are flatter and more ovoid than they are more anteriorly. This is probably the region of junction between the trabeculae and the parachordal cartilages (fig. 6). A wide space, such as is found between the trabeculae cranii and the ventral surface of the brain in the fenestra myodomus ventralis region, is obliterated here by the presence of the hypophysis cerebri, so that the membrane enclosing the brain is continuous with the perichondria of the trabeculae crani (fig. 6). A foramen is present in the cranial wall just posterior to the prootic process. Through this foramen the jugular vein and the ramus hyomandibularis facialis pass (fig. 6). The posterior margin of the foramen is formed by the wall of the otic capsule. Posterior to the facialis foramen the parachordal cartilages are fused with the ventromesial margins of the otic capsules. Mesially, the parachordals are closer together than were the trabeculae farther anterior. The parasphenoid lamella forms the floor of the intervening fenestra, its roof is formed by fibrous connective tissue ventral to the posterior end of the hypophysis. Just posterior to this region the parasphenoid ends and the space between the parachordals is occupied by the anterior end of the notochord (fig. 7). The notochord is separated dorsally from the cavum crani by connective-tissue stroma connecting the perichondria of the parachordals. The space between the parachordals in which the notochord lies has been called the inter- parachordal fossa in Gasterosteus (Swinnerton), and the fenestra basicranii posterius in Salmo (Gaupp); the latter terminology is used in this paper. 434 JAMES ERNEST KINDRED The cartilage of this region of the parachordals has the same relation to the notochord in both Syngnathus and the 5.7-mm. Gasterosteus, but it is not as widely separated from the noto- chord as it is in a 19-mm. Amia (Kindred, 719). This indicates Fig. 7 Cross-section through the middle part of the otic capsules, 8-mm. Syngnathus. Semidiagrammatic. Camera lucida. X 100. Fig. 8 Cross-section through the occipital region, 8-mm Syngnathus. Semi- diagrammatic. Camera lucida. X 100. ABBREVIATIONS br.II, branchial arch IT pch., parachordal cartilage f.bcr.p., fenestra basicranii posterius sept.semi., septum semicircularis lat- for.[X-X, foramen for IX and X cra- eralis nial nerves ven.j., vena jugularis nch, notochord X, ganglion of vagus nerve ot.c., otic capsule a precocious growth of the cartilage in this region of Syngnathus. Laterally, the cartilage of each parachordal is confluent with the ventromesial margin of the otic capsule (fig. 7). The carti- lage cells of the parachordals are arranged concentrically, while those of the otic capsule are in a vertical row. CHONDROCRANIUM OF SYNGNATHUS FUSCUS 435 Farther posterior the notochord increases greatly in diameter and invades the cavum ecranii. The parachordals abut against the ventrolateral surfaces of the notochord, so that it really lies in a groove on the dorsomesial surfaces of the parachordals. The parachordals are separated from each other ventral to the notochord and there is no trace of a hypochordal bridge. In the region between the posterior parts of the otic capsules, the enlarged notochord lies between the maculi utriculi and the auditory ganglia adjacent to them ventrolaterally. The relation _ of the parachordals to the notochord remains as in the more anterior region. At the posterior end of the otic capsules the cartilage forming their walls fuses with the occipital processes which lie on the dorsolateral surfaces of the notochord dorsal to the parachordals (fig. 8). The parachordals retain their identity for a short distance posterior to this fusion; gradually, however, the cells of the parachordals become confluent with those of the occipital cartilages (fig. 8). A slender bridge of fibrous connective tissue connects the occipital processes with each other dorsal to the notochord, but their ventral margins are separated by the notochordal sheath (fig. 8). A canal extends ventroposteriorly from the floor of the otic capsule and opens to the exterior posteroventral to its wall. The mesial wall of this canal is formed by the occipital cartilage (fig. 8). The canal contains a part of the sacculus and the fibers of the glossopharyngeal and vagus nerves. A common canal for these two cranial nerves is also found in the 6.6-mm. Gasterosteus, so it may be stated that Syngnathus is more precocious in devel- opment in this part of the cranium than in the more anterior parts, since all of the parts so far noted have been comparable in state of development to the same parts in the 5.7-mm. Gasterosteus. The occipital masses lateral to the notochord persist for a short distance posteriorly, but do not meet dorsal to the brain to form an occipital arch (figs. 2, 8). They gradually diminish in extent, the ventrolateral parts disappearing first, as would be expected, since these represent the posterior ends of the parachordals. Finally, the cartilage on the dorsolateral surfaces 436 JAMES ERNEST KINDRED of the notochord is replaced by densely cellular masses of fibrous connective tissue. There are no postvagal nerves in this region and the first neural arch is relatively far distant from the posterior end of the occipital cartilages. The postorbital process of the otic capsule appears in the posterior part of the membranous orbital wall (fig. 5), and its anterior end does not extend as far dorsal in the wall as does its homologue in a 6.6-mm. Gasterosteus. In cross-section the postorbital process is very thin and flat, increasing in thickness posteriorly. At the same time it trends ventrally in the cranial wall (fig. 1). The jugular vein and the gasserian ganglion lie ventral to the postorbital process, and for this reason Swinnerton compared the postorbital process of Gasterosteus to a part of the alisphenoid cartilage of Salmo (Parker, ’72)—a comparison with which I am in full agreement as regards the postorbital process of Syngnathus, both from this relation and also from its connection with the otic capsule. Posteriorly, the dorsal end of the postorbital process becomes bulbous, the abductor hyomandibularis muscle having its origin on the lateral face. Posterior to the origin of this muscle the dorsal end of the hyomandibula articulates in a groove on the lateral face of the postorbital process (fig. 6). As noted above, the ventral end of the postorbital cartilage becomes the prootic process which trends ventrally and meets the trabecula cranii (figs. 1, 2). The ramus hyomandibularis facialis and the jugular vein leave the cranium posterior to the prootic process, the dorsal margin of the foramen being formed by the postorbital process (fig. 6). The adductor hyomandibularis muscle has its origin on the ventroposterior margin of this foramen. Similar topo- graphical relationships between the cartilage and the cranial nerves have been described for Gasterosteus, but the muscle relations have not been noted. In a 25-mm. Salmo (Gaupp), a band of cartilage (prifaciale basicapsulire Commissure) con- nects the postorbital process with the parachordal and is perfo- rated by three foramina—an anterior one for the jugular vein, a posterior one for the ramus hyomandibularis facialis, and a ventral one for the ramus palatinus facialis. CHONDROCRANIUM OF SYNGNATHUS FUSCUS 437 In Syngnathus the cartilage forming the dorsal margin of the foramen for the passage of the jugular vein and the ramus hyomandibularis facialis extends more dorsally in the wall than does the postorbital process (fig. 1). Posterior to the foramen the cartilage forms the wall of the otic capsule proper, since the membranous labyrinth appears between its wall and the brain (fig. 7). The large basicapsular fenestra in the wall of the otic capsule of a 5.7-mm. Gasterosteus and of a 13-mm. Salmo has been interpreted as the homologue of the fenestra ovalis in the otic capsule of Amphibia. This is represented in the wall of the otic capsule of Syngnathus by a minute opening in its ventral part, closed by membrane. _ Two cartilaginous projections from the mesial face of the capsular wall, connected by fibrous tissue with the membrane separating the cavum cranii from the cavum labyrinthii, are the primordia of the lateral and posterior septa semicircularia. The lateral septum is shown in figure 7. The posterior one is located a short distance posterior to this and more dorsal. The dorsomesial margins of the otic capsular walls are con- tinuous with a narrow median cartilage which lies in the fibrous connective tissue dorsal to the brain (fig. 2). This bar of earti- lage starts as a small point in the membranous roof of the cranium of the postorbital region in the same transverse plane as that in which the facialis foramen is located (fig. 6). It gradually becomes wider posteriorly and forms a triangular plate in the cranial roof (fig. 2). The posterolateral margins of this plate are confluent laterally with the mesial margins of the otic capsular walls. Such a plate of cartilage has been described, but not figured for a 25-mm. Salmo by Gaupp, and termed the taenia tectum medialis. An epiphysial bar such as is found in the cranial roof of Gasterosteus connected with the ectethmoid region by a pair of supraorbital cartilages is lacking in Syngna- thus. Hence in Syngnathus the taenia tectum medialis is the only cartilage which at this stage lies dorsal to the brain, since the occipital cartilages have not as yet met mesially. It may be considered as the remnant of a once solid cartilaginous synotic tectum which has become very much reduced during the phylo- genetic processes which gave rise to Syngnathus. 438 JAMES ERNEST KINDRED As already mentioned, the posterior wall of the otic capsules becomes confluent with the occipital processes and forms a canal on either side of the chondrocranium for the passage of the glossopharyngeal and vagus nerves. B. The visceral arches The first part of the primordial visceral skeleton to be con- sidered is a small median precranial cartilage, the rostral cartilage (figs. 1, 2, 3). Lying on the middorsal surface of the rostral process of the ethmoid plate, it has a relationship to the latter, comparable to that in other teleosts, as mentioned by Gaupp. This cartilage was not noted by MecMurrich in Syngnathus peckianus. Sagemehl (’91), in considering the rostral cartilage in other teleosts, has homologized it to the median synchondrosis which occurs between the distal ends of the palatoquadrate cartilages of Heptanchus. This homology is further borne out by the conditions in Syngnathus at this stage, because the rostral cartilage is connected by densely cellular connective tissue with the anterior ends of the palatine cartilages. The beginnings of the premaxillary ossifications are connected with the lateral surface of this rostral cartilage (fig. 3). It has no homologue in Gasterosteus. The anterior ends of the palatine cartilages (ethmopalatines, MeMurrich) are flat in cross-section, the mesial surface of each articulates with the lateral surface of the rostral process of the ethmoid cartilage and is connected by fibrous connective tissue with the latter and with the rostral cartilage (figs. 1, 3). Poste- rior to the region of articulation, each palatine cartilage tapers and trends ventrally, finally dwindling to a small point embedded in the embryonal connective tissue surrounding the ethmoid plate and the olfactory pit (fig. 1). It is important to note here that the palatine cartilage has a posterior fibrous connection with the dorso-anterior margin of the pterygoquadrate, rather than a cartilaginous connection as in Salmo or Gasterosteus. This condition in Syngnathus bears out the statement made by Swinnerton to the effect that the palatine cartilage does not arise CHONDROCRANIUM OF SYNGNATHUS FUSCUS 439 independently in any teleost, but must at least have a fibrous, if not a cartilaginous connection with the more posteriorly situated pterygoquadrate. Such a condition exists at this stage in Syngnathus, and whatever the later conditions may be, this fibrous connection is primary. In this respect the relations of the posterior end of the palatine cartilage resemble that of the 10-mm. Amiurus (Kindred) in which the posterior end of the palatine is connected with the anterior end of the pterygo- quadrate by a connective-tissue bridge. MeMurrich in his description of Syngnathus peckianus failed to recognize this relationship and stated that the ‘ethmopalatine’ was independent of the pterygoquadrate. The anterior relation of the palatine cartilage to the ethmoid plate is similar to that of Gasterosteus—a condition which Swinnerton calls acrartete. If it is assumed that the fibrous connection is the homologue of the intervening cartilage in Salmo, then the fibrous connection between the posterior end of the palatine and the anterior end of the pterygoquadrate which passes ventral to the ectethmoid cartilage in Syngnathus may represent a condition comparable to that in Salmo, where the posterior part of the palatine process of the palatoquadrate articulates with the ventral surface of the ectethmoid process. In the 8-mm. Syngnathus, the mandible is formed by the fused meckelian cartilages. They project for a short distance beyond the anterior margin of the dorsal part of the oral gape and form the axes of the shovel-like ventral portion (figs. 1, 2). The anterior end of each meckelian cartilage abuts against its fellow in the median line by a flat thickened surface. The cartilage cells on the abutting surfaces are very small, numerous, and arranged in a vertical row, separated from each other by the fused perichondria. Posteriorly, the cartilages diverge, become smaller in cross-section, and are connected with each other mesially for a short distance by a slender band of developing muscle tissue. The muscle tissue is replaced more posteriorly by embryonal connective tissue. Each meckelian cartilage is gradually compressed to form the coronoid process (fig. 1). A small notch on the anteroventral face of this region separates 440 JAMES ERNEST KINDRED off an angular process, just posterior to which the cartilage articulates with a groove on the dorsal surface of the quadrate portion of the pterygoquadrate. Meckel’s cartilage of the 8-mm. Syngnathus differs in several respects from that of the 6.6-mm. Gasterosteus. In the first place, the curve of the ventral surface of this cartilage in Syngna- thus is concave, while in Gasterosteus it is convex. There is a continuous gradation from the anterior part of the cartilage into the coronoid process in the former, while in the latter, the coro- noid process projects abruptly from the dorsal surface. In Syngnathus the posterior end of the cartilage projects dorgal to the quadrate—a condition not found in Gasterosteus. Of the angular notch and process of Syngnathus, the latter only is present in Gasterosteus. A rudimentary inferior labial cartilage is represented by a cellular mass which extends along the dorsal surface of the anterior part of Meckel’s cartilage and ends posteriorly in the mandibular fold. It is connected by a bridge of connective- tissue cells with the primordium of the maxillary bone which lies in the supramandibular fold connected dorsally with the lateral surface of the palatine cartilage by fibrous connective tissue. The quadrate portion of the pterygoquadrate cartilage starts anteriorly between the posterior end of Meckel’s cartilage and the distal end of the symplectic (fig. 1). A narrow bridge of cartilage connects it with the latter element, showing possibly a common origin for the cartilage in this region. In cross- section the quadrate is dumb-bell shaped and hes at right angles to the articular surface of Meckel’s cartilage. The posterior part of it projects dorsally as a flattened plate, connected with the posterior end of the palatine cartilage by a densely cellular strand of fibrous connective tissue and separated from the ethmoid cartilage by a fold of the oral cavity (fig. 4). The symplectic cartilage extends along the ventromesial surface of the pterygoquadrate. This cartilage ends posteriorly in the midregion of the orbit with a small posteriorly projecting proc- ess—the metapterygoid process (fig. 1). According to Swin- CHONDROCRANIUM OF SYNGNATHUS FUSCUS 44] nerton, the metapterygoid process of Syngnathus represents a stage in the reduction of an elongate metapterygoid process such as is found in other teleosts. The metapterygoid process does not have the intimate relation with hyomandibula which char- acterizes the metapterygoid process of Salmo and Amiurus. As already stated, the symplectic element extends ventral to the anterior margin of the pterygoquadrate and is confluent with it by means of a bridge of cartilage. Continuing posteriorly asa slender cartilaginous core, the symplectic extends mesial to the pterygoquadrate along the entire extent of the latter (figs. 1, 2). A ventral diverticulum of the oral cavity separates it from the copula communis posterior to the pyterygoquadrate. Histo- logically, it has a very heavy perichondrium and there are usually two or three cartilage cells in a cross-section. Posterior to the orbit, the symplectic is connected to the trabecula communis by several strands of embryonic muscle tissue. Finally it becomes confluent with the ventral end of the hyomandibula, no line of division being present between the two. A similar condition is met in the symplectic of Gasterosteus, but as yet the cartilaginous continuity between the distal end of the sym- plectic and the pterygoquadrate characteristic of Syngnathus has not been described. The great distance between the meta- pterygoid process of the pterygoquadrate and the, symplectic is to be noted in Syngnathus as compared to the intimate relation between the elongate metapterygoid and the symplectic in Gasterosteus and Belone (Swinnerton). The hyomandibula at this stage is a rectangular piece of cartilage which articulates at its dorsal end with the anterior fourth of the otic capsule and is confluent ventrally with the symplectic (fig. 1). Ventromesially, it is flattened for articu- lation with the stylohyal cartilage (fig. 6). The dorsal end of the hyomandibula is thin and rounded where it abuts against the otic capsular wall. The ventral portion is thickened (fig. 6). Near the anterior dorsal margin a small foramen is present for the passage of the ramus hyomandibularis facialis. The abductor hyomandibularis muscle is inserted on the anterior margin of this foramen. The opercular process projects from the posterior 449 JAMES ERNEST KINDRED margin of the dorsal part of the hyomandibula and the adductor hyomandibularis muscle is inserted on its mesial face. Here again certain differences are to be noted between the chondrocranium of Syngnathus and that of a 5.7-mm. Gastero- steus. The hyomandibula of the former is more elongate and rectangular on its vertical axis than is the hyomandibula of the latter. The relative amount of articular surface with the otic capsule is greater in Gasterosteus than in Syngnathus. The ventral portion of the hyomandibula of Syngnathus is thicker Fig. 9 Ventral view of hyoid and branchial arches, 8-mm. Syngnathus. Drawing made from wax model 230 times actual size of parts. Ratio of drawing to model, 1:3. ABBREVIATIONS br.I-IV, branchial arches I to IV hyp.h., hypohyal cartilage c.c., copula commun.s styl.h., stylohyal cartilage ct.h., ceratohyal cartilage sym., symplectic cartilage hyom., hyomandibula than the dorsal part, while in Gasterosteus the conditions are reversed. The distinct opercular process present in Syngnathus is lacking in Gasterosteus. The foramen for the passage of the ramus hyomandibularis facialis is the only constant feature of the body of the hyomandibula of these two forms. The copula communis is at this stage a delicate cylindrical bar of cartilage which forms the median articulating support for the hypohyal cartilages and the first and second branchial arches (fig. 1, 9). It is supported in the embryonal connective tissue ventral to the oral cavity (fig. 4). The cartilage is continuous and of uniform caliber in its anterior part, tapering posteriorly CHONDROCRANIUM OF SYNGNATHUS FUSCUS 443 as a core of procartilage cells, against which abut the third and fourth branchial arches. The copula communis of a 5.7-mm. Gasterosteus has approxi- mately the same relations to the hypohyals and the first and second branchial arches, but in addition it has extended as cartilage as far as the third branchial arch. There is also a small separate cartilage between the fourth pair of branchials. The independent wedge-shaped piece at the anterior end of the copula communis of Gasterosteus is lacking in Syngnathus. The branchial cartilages of Syngnathus have not curved dorsally at their distal ends in this stage as they have in the 5.7-mm. Gasterosteus, nor has the fifth branchial arch appeared. The pharyngobranchial plates present in Gasterosteus are represented in Syngnathus by a pair of procartilaginous masses ventral to the parachordals and connected with each other by a sheet of muscle. The hyoid elements are very well developed in Syngnathus. The stylohyal is a small broad plate between the ventral end of the hyomandibula and the dorsolateral surface of the cerato- hyal (fig. 9). This element, the stylohyal, lies closely ventral to the posterior margin of the hyomandibula, and not ventral to its anterior margin as does the corresponding element in Gasterosteus. The ceratohyal is very massive at this stage and extends anteriorly rather than directly mesial, as in Gasterosteus. It abuts against the ventrolateral margin of the elongate, hori- zontal hypohyal. This latter cartilage is very peculiar in its relation and extent. Instead of being a small wedge-shaped articular plate, as in Gasterosteus, it is an elongate rod of carti- lage which anteriorly extends beyond the end of the ceratohyal (fig. 9). The surface of articulation between the hypohyal and the copula communis is nearer to the posterior end of the hypo- hyal than to its anterior end. 444 JAMES ERNEST KINDRED THE CHONDROCRANIUM OF THE 12-MM. STAGE A. The neurocranium The anterior end of the ethmoid plate of the 12-mm. Syngna- thus is much broader and blunter than it is in the neurocranium of the 8-mm. stage. Its relations to the palatine and rostral cartilages are similar (fig. 11). The anterior end, however, is no longer turned dorsally, but has straightened out, carrying with it the attached elements. Posterior to the broad rostral process, the ethmoid plate gradually diminishes in size. It becomes wedge-shaped at first and then more triangular in cross- section. The dorsal part of the ethmoid is the thicker. The vomer primordium represented in the 8-mm. stage by a mass of cells ventral to the ethmoid plate is now an osseous lamella embedded in a mass of osteogenetic cells in the same region (fig. 12). ; The dorsal surface of the ethmoid plate is surmounted by a septum of fibrous-connective tissue. The relationship of this to the cartilage is the same as that of the osseous ridge which later develops in this place. This ridge of connective tissue is connected with the stroma enclosing the olfactory pits and is continuous posteriorly with the membranous interorbital septum. Histologically, the ethmoid cartilage is the’same as it is in the 8-mm. stage except for an increase in size and the presence of a thicker perichondrium. The ectethmoid cartilages, which in the 8-mm. embryo were separate, connected with each other by connective tissue only, have now fused with each other dorsomesially to form a horse- shoe-shaped mass of cartilage posterior to the olfactory pits (fig. 12). The posterior margin of this mass forms the anterior boundary of the orbits. The anterior end of the brain lies in a trough on the dorsum of this ectethmoid arch, the margins of the trough projecting posteriorly for a short distance in the mem- branous cranial wall (figs. 10, 11). The olfactory nerves, after leaving the olfactory pits, pass to the mesial margins of the sides of the arch and extend posteriorly within it (fig. 12). Thus at this stage the olfactory nerves have not been separated from taen.tect.med. en occ.arch Fig. 10 Lateral view of the chondrocranium and visceral arches of a larva of Syngnathus fuscus, 12 mm. long. Drawing made from wax model 150 times the actual size of the chondrocranium. Ratio of drawing to model, 1:3. Fig. 11 Dorsal view of the same model as shown in figure 10. ABBREVIATIONS br.I-V, branchial arches I to V. c.c., copula communis ct.h., ceratohyal cartilage ect., ectethmoid cartilage eth., ethmoid plate f.ber.p., fenestra basicranil posterius f.hyp., fenestra hypophyseos f.my.vent., fenestra myodomous ven- tralis f.m., foramen magnum f.VII, foramen for exit of ramus hyo- mandibularis facialis For.hyom.VIT, foramen in hyomandib- ula for ramus hyom.fac. hyom., hyomandibula hyp.h., hypohyal cartilage j.’.p., Junction rostro-palatinus Mk., Meckel’s cartilage mpt.pr., metapterygoid process nch., notochord occ.arch., occipital arch ot.c., otic capsule pal., palatine cartilage pch., parachordal cartilage ph.br., pharyngobranchial cartilage pt.qu., pterygoquadrate cartilage po.pr., postorbital process pr.pr., prootie process r.c., rostral cartilage rs.p., rostral process of the ethmoid styl.h., stylohyal cartilage sym., symplectic cartilage taen.tect.med., taenia tectum medialis trab.comm., trabecula communis trab.cr., trabecula cranii 445 446 JAMES ERNEST KINDRED each other by cartilage. They enter the brain at the posterior margin of the mesial dorsal part of the arch. It is significant to note here that the olfactory nerves are in the process of being passively enclosed in cartilage and do not actively fenestrate it. Fig. 12 Cross-section through the ectethmoid region, 12-mm. Syngnathus. Semidiagrammatic. Camera lucida. X 102. Fig. 13 Cross-section through the posterior end of the orbit, 12-mm. Syngna- thus. Semidiagrammatic. Camera lucida. X 66. ABBREVIATIONS c.c., copula communis olf.n., olfactory nerve ct.h., ceratohyal cartilage po.pr., postorbital process ect., ectethmoid cartilage ps., parasphenoid ossification eth., ethmoid plate rect.m., recti eye muscles f.my.vent., fenestra myodomus ven- sym., symplectic cartilage tralis trab.cr., trabecula cranii front., frontal ossification V, ganglion of trigeminal nerve hyom., hyomandibula ven.j., vena jugularis int.cart., internal carotid artery vo., vomer ossification The ethmoid plate lies ventromesial to the arch, connected with it by the fibrous connective-tissue ridge already mentioned, as extending along the dorsal surface of the ethmoid. In Salmo, Gasterosteus, and Amiurus, the ectethmoid processes are out- growths from the ethmoid cartilage. If the ectethmoid carti- CHONDROCRANIUM OF SYNGNATHUS FUSCUS 447 lages of Syngnathus are to be regarded as remnants of the olfactory capsules, the condition of the ectethmoids in the above forms indicates a more specialized relation. Therefore, I am regarding the relation of the ectethmoid cartilages of Syngnathus as primitive. There is no tegmen or cartilage dorsal to the brain in the ethmoid region, the tips of the frontal ossifications appearing as a pair of delicately spined osseous lamellae in the fibrous connective tissue enclosing the brain. Just posterior to the ectethmoid region the oblique eye muscles have their origin in the fibrous tissue of the interorbital septum and are not related to the cartilage of either the ethmoid plate or the ectethmoid cartilages. The trabecula communis is flatter and broader in cross-section than that part of the ethmoid plate immediately ventral to the ectethmoid arch, but it gradually diminishes, until at the anterior margin of the fenestra myodomus ventralis it is a very slender cylindrical bar. The parasphenoid lamella is attached to its ventral surface by connective tissue. In the region of the anterior margin of the fenestra myodomus ventralis, the internal carotid arteries, which more anteriorly were dorsal to the trabecula communis, now pass ventrally into the space between the trabeculae cranii (fig. 13). The recti eye muscles enter the fenestra myodomus ventralis dorsal to the internal carotid arteries between the trabeculae cranii and the fibrous connective tissue enclosing the brain. The hypo- physis cerebri lies posterior to this region, so that, in accordance with Allis’ criteria, this space between the anterior ends of the trabeculae cranii has been termed the fenestra myodomus ventralis, and not the fenestra hypophyseos. As in the younger stage, the parasphenoid ossification forms the floor of the fenestra myodomus ventralis and serves as a surface of attachment for the recti muscles. The anterior end of the hypophysis cerebri appears just posterior to this region, and the space occupied more anteriorly by the recti muscles and by the carotids is obliterated, the hypophysis pushing the meningeal tissue against the trabeculae cranii. This portion of the intertrabecular fenestra may be properly termed the fenestra hypophyseos. 448 JAMES ERNEST KINDRED As in the younger stage the prootic processes abut against the lateral surfaces of the trabeculae and are separated from them by their perichondria (figs. 10, 11). Except for a greater size of all parts concerned, the foramen for the passage of the ramus hyomandibularis facialis and the jugular vem bears the same relation to the trabeculae and prootic processes as in the 8-mm. stage. In this region the trabeculae pass insensibly into the parachordals. These immediately begin to thicken and are fused laterally with the ventromesial walls of the otic capsules (fig. 11). As the hypophysis tapers posteriorly the fenestra hypophyseos begins to narrow and the parasphenoid lamella dwindles to a mere spicule of bone forming its floor. Where the hypophysis ends the anterior tip of the notochord appears between the parachordals. It is enclosed in the fenestra basi- cranii posterius. The roof of this fenestra is formed by fibrous connective tissue confluent on each side with the dorsal peri- chondria of the parachordals. The floor is formed by fibrous connective tissue confluent with the ventral surface of the para- chordals. The notochord is free and is not enclosed in cartilage as is the intercranial notochord of the 6.6-mm. Gasterosteus. The notochord is, however, much more closely applied to the cartilage in Syngnathus than is the notochord of a 19-mm. Amia. The parachordal-occipital process fusion in conjunction with the posterior wall of the otic capsule forms as in the 8-mm. stage, the canal for the passage of the glossopharyngeal and vagus nerves. Ossification has not begun in this region. The postorbital process bears the same relation to the tri- geminal ganglion that it did in the younger stage, but a deeper notch occurs in the anterior margin of the prootic process poste- rior to the ganglion (figs. 11, 13). This notch is the homologue of the incisivum prooticum of Salmo. Asin the younger Syng- nathus the perichondria between the trabecula and the prootic process persist. Due to the lateral growth of the brain and an increase in size of the ramus hyomandibularis facialis and the jugular vein, the foramen through which these pass is more ventral in the wall CHONDROCRANIUM OF SYNGNATHUS FUSCUS 449 and larger than it is in the 8-mm. stage (fig. 11). The Jamella of the frontal ossification is continuous with the perichondrium of the cartilage forming the dorsal margin of the foramen. Just posterior to the foramen, the cartilage is more deeply gi ooved for articulation with the hyomandibula than it is in the younger stage. The articular surface does not reach as far posteriorly as that part of the otic capsule containing the membranous labyrinth (fig. 10). The membranous labyrinth has increased greatly in size since the 8-mm. stage and consequently expanded laterally, displacing to a more ventral position the cartilage which formed the lateral wall of the otic capsule in the younger stage. A thin lamella of cartilage surmounts the margin of the thicker ventral portion and forms the dorsolateral wall of the capsule in this region. The formation of the septa semicircularia has not gone farther than in the younger stage and are as yet mem- branous. The small fenestra basicapsularis present in the otic capsule of the 8-mm. stage persists. Toward the posterior portion of the cranium the mesiodorsal margins are continuous with the cartilage of the taenia tectum medialis (fig. 11). The latter is enclosed in an osseous lamella which represents the beginning of the supra-occipital ossification. Some of the carti- lage is in the process of resorption. The occipital processes which in the younger stages were separate from each other have fused mesially, so that an occipital arch is formed. The posterior margin of the arch projects for a short distance beyond the sides of the arch. The fontanelle left between the taenia tectum medialis and the occipital arch is closed by membrane. B. The visceral arches The rostral cartilage has the same relation to the dorsal surface of the ethmoid plate that it had in the younger stage, but the premaxillary lamellae which are represented in the younger stage by cellular masses have now ossified and extend laterally from the rostral piece, overlapping the anterior ends of the palatines. 450 JAMES ERNEST KINDRED The palatine cartilages have the same relation to the rostral process of the ethmoid as before, but are longer than in the younger stage and lie in a more horizontal plane—a condition due to the straightening out of the anterior end of the ethmoid (fig. 10). Posteriorly, the palatine cartilages have grown along the fibrous strand which in the younger stage connects them with the pterygoid portion of the pterygoquadrate plates. As a result of this growth, the palatine cartilages and the pterygo- quadrates are in closer proximity than they are in the younger stage (fig. 10). The mandibular symphysis is broader and thicker than in the 8-mm. stage and the meckelian cartilages diverge from it as a pair of cartilaginous rods, which thicken gradually at their posterior ends to form the coronoid processes (fig. 10). The mandible is now more typically teleostean than it is in the younger stage, the peculiar concavity noted on the ventral surface in the 8-mm. stage has been obliterated. The primordium of the dentary ossification appears around each Meckel’s cartilage as a single lamella, lateral to and separate from the cartilage. No teeth are present. As in Gasterosteus, the angular process projects posteriorly beyond the surface of articulation with the pterygoquadrate. The pterygoquadrate articulates with the dorsal surface of this portion of Meckel’s cartilage at an oblique angle. It becomes vertical posteriorly and there is no longer a projection of the posterior end of the coronoid process dorsal to it. The symplectic element extends ventral to the posterior mesial region of the pterygoquadrate and the metapterygoid process projecting from this region is longer than it is in the 8-mm. stage (fig. 10). Due to the straightening out of the mandible and the enlarging of the oral opening, the notch present on the ventro-anterior margin of the distal portion of each Meckel’s cartilage has disappeared. This whole region has been pushed farther ante- rior than it is in the 8-mm. stage, by growth and elongation of the symplectic element which has grown in concert with the ethmoid cartilage. The posterior end of Meckel’s cartilage no longer meets the distal end of the symplectic. CHONDROCRANIUM OF SYNGNATHUS FUSCUS 451 The mass of cells indicated in the 8-mm. stage as the homologue of the inferior labial cartilage of Gasterosteus has not changed. However, the cells which represented the center of formation of the maxillary ossification have deposited an elongate, vertical lamella. This lamella is connected dorsally with the premaxil- lary lamella anterior to the palatine articulation, and ventrally by a chain of cells with the dorsal tip of the dentary lamella. The maxillary lamella supports the lateral wall of the oral cavity between the end of the cranium and the palatine cartilage. The maxillary lamella bears no teeth. The metapterygoid process is longer than it is in the 8-mm. embryo, but even now does not have a relation to the symplectic like that of the larval Gasterosteus. As Swinnerton states, the metapterygoid process of Syngnathus represents a stage in the disappearance of such a structure in the teleosts. The posterior end of this process which in the 8-mm. stage lies posterior to the transverse plane of the ectethmoid cartilages now lies quite far anterior to them (fig. 10). . The distal end of the symplectic cartilage which in the 8-mm. stage extends the whole length of pterygoquadrate now reaches only to its posterior end. The cartilaginous connection between these two elements has disappeared. The elongation of the symplectic and the subsequent changes in the positional rela- tionships in that part of the visceral apparatus lying anterior to it have already been stated. Ossification has appeared in the form of a curved lamella external to and distinct from the perichondrium of the symplectic shaft. Posteriorly, the proxi- mal end of the symplectic gradually widens as it becomes con- fluent with the hyomandibula, so that the angle, which in the younger stage appears between the proximal end of the sym- plectic and the anterior margin of the hyomandibula, has been obliterated. As a result, it has much the same appearance as that of the 6.6-mm. Gasterosteus. The hyomandibula, instead of having the vertical position of the 8-mm. stage, is now directed anteriorly at its ventral end, so that the bulk of the cartilage lies anteroventral to the surface of articulation (fig. 10). The posterior extent of the hyoman- 452 JAMES ERNEST KINDRED dibula is not nearly as great as it is in Gasterosteus or Salmo. Although the whole ventral portion has swung anteriorly, the foramen for the passage of the ramus hyomandibularis facialis lies as before near the anterior margin of the dorsal head of the Fig. 14 Ventral view of hyoid and branchial arches, 12-mm. Syngnathus. Drawing made from wax model 150 times actual size. Ratio of drawing to model, eo ABBREVIATIONS basi-br.III-IV, basibranchial ceartil- hyom., hyomandibula ages III and IV hyp.h., hypohyal cartilage br.I-V, branchial arches I to V ph.br., pharyngobranchial cartilage c.c., copula communis styl.h., stylohyal cartilage ct.h., ceratohyal cartilage hyomandibula. As in the 8-mm. stage, the dorsal portion is relatively thinner than the ventral. The mesioventral surface of the hyomandibula is flattened for articulation with the interpolated stylohyal, the plane of articulation having been carried anteriorly by the shift in the axis of the hyomandibula (figs. 10, 14). The shift which has taken place in the hyomandibula is due to the part played by the latter in changing the size of the oral cavity in the process CHONDROCRANIUM OF SYNGNATHUS FUSCUS 453 of respiration, since Syngnathus at this stage is a free-swimming larva. An osseous lamella lies external to and separate from the ventral portion of the hyomandibula, enclosing a blood- vessel between it and the perichondrium (fig. 13). As in the 8-mm. stage, the stylohyal element is interpolated beween the dorsal end of the ceratohyal and the hyomandibula (fig. 14). The ceratohyal, which in the same stage (when the embryo is yet enclosed in the brood pouch of the parent) extends anteroposteriorly in an almost horizontal plane with the posterior end in the more ventral position, has now changed in position, so that it lies almost vertical (fig. 10). The dorsal end of each ceratohyal articulates with the stylohyal of that side, while the ventral end, formerly parallel with the hypohyal element, now lies ventroposterior to it. The hypohyal element has also undergone a change in position. Instead of lying in a horizontal plane parallel to the copula communis, it now occupies a vertical position, articulating on its mesiodorsal surface with the copula communis (fig. 14). The change in the positional relationships of these elements is due to the functional activity of these as the supporting cartilages of the oral cavity, which by its changes in shape draws water into the mouth and expels it over the gills. Since the gills were not developed in the 8-mm. stage, these parts had not shifted. The conditions of these elements in the 12-mm. stage of Syn- enathus are more nearly like those of the 6.6-mm. Gasterosteus, which indicates an earlier functional activity of these parts in the latter, correlated with the longer protected period of Syngnathus. The copula communis of the 12-mm. stage has retained the same relative length as before, but is more widely separated from the ventral surface of the cranium, because of the increase in the size of the oral cavity (figs. 10, 14). Posteriorly it extends beyond the second branchial arch after having been displaced ventrally between the hypohyal elements due to the shift in their position. Its posterior end is connected by a cord of procartilage cells with the independent basibranchial cartilages between the third and fourth branchial arches. 454 JAMES ERNEST KINDRED The first branchial arch has elongated and is turned up laterally to fuse with the pharyngobranchial cartilage which lies at the laterodorsal margin of the oral cavity (figs. 18, 14). The pha- ryngobranchial plate is also connected with the lateral end of the second branchial arch, ending just posterior to it (fig. 14). The third and fourth branchial arches have not as yet turned up laterally, but they are longer than they are in the 8-mm. stage. The fifth branchial arch has appeared as a pair of small cartilages lateral to the procartilage continuum of the basi- branchial plate of the fourth arch (fig. 14). SUMMARY AND CONCLUSIONS Under the classification of the teleosts by Gregory (’07), Syngnathus is placed in the order Lophobranchii, which with the order Hemibranchii (including Gasterosteus), are grouped into a superorder, the Thoracostraci. The important similarities and differences of the chondrocrania of these two forms are stated in the following summary. Both Gasterosteus and Syngnathus are alike in the following chondrocranial characters: the presence of an elongate ethmoid region; the acrartete articulation of the palatine cartilages; the incomplete cranial roof; the horizontal position of the trabeculae crani and the parachordal cartilages; the presence of two septa semicircularia in each otic capsule; a prootic process separating the trigeminal and facialis nerves from each other; an elongate fenestra basicranil posterius; a common foramen for the glosso- pharyngeal and vagus nerves between the otic capsule and the occipital arch; a small pterygoquadrate cartilage; a large inter- cranial notochord; a postorbital process at the anterior margin of each otic capsule. The above characters wherein Syngnathus resembles Gastero- steus indicate the direction in which the skulls of these two forms have become specialized. In addition, further specializations are to be noted in the chondrocranium of Syngnathus: 1. The modification of the anterior end of the ethmoid plate by turning dorsally in the 8-mm. stage. CHONDROCRANIUM OF SYNGNATHUS FUSCUS 455 2. The presence of a definitive fenestra mvodomus ventralis between the anterior ends of the trabeculae cranii. 3. The absence of cartilage on the dorsal and ventral surfaces of the intercranial notochord. 4. The presence of a taenia tectum medialis in the cranial roof. 5. The presence of a minute basicapsular fenestra in the ventral wall of each otic capsule. 6. The fibrous connection between the posterior end of the palatine cartilage and the dorsal end of the pterygoquadrate cartilage. 7. The presence of an elongate symplectic cartilage and a slender hyomandibula. 8. The great length of the trabecula communis and the ethmoid plate. 9. The change in position of the hyoid elements during development. 10. The presence of a reduced metapterygoid process. The chondrocranial characters of Syngnathus which may be regarded as primitive are: 1. The presence of a rostral cartilage which has been con- sidered by Sagemehl in other teleosts as the homologue of the median synchondrosis between the palatopterygoid cartilages of Heptanchus. 2. The development of the ectethmoid cartilages independ- ently of the ethmoid plate, which seems to the author to indicate that they are the remnants of an olfactory capsule which devel- oped independently. 3. The open communication between the cavum labyrinthii and the cavum cranii. 4. The absence of postvagal nerves in the occipital region. From the above preponderance of specialized characters, it may be concluded that even in its early stages of development the skull of Syngnathus is already highly specialized. 456 JAMES ERNEST KINDRED LITERATURE CITED Aus, E. P., Jr. 1919a The lips and nasal apertures in Gnathostome fishes. Jour. Morph., vol. 32, pp. 145-205. 1919 b The myodome and trigemino-facialis chamber of fishes and the corresponding cavities in higher vertebrates. Jour. Morph., vol. 32, pp. 207-326. Ganin, M. 1880 Ueber die Entwickelung des Kopfskeletts bei Knochenfischen (Rhodeus, Gasterosteus). Zool. Anz., Bd. 3, 8. 140-141. Gaupp, E. 1906 Die Entwickelung des Kopfskeletts. Handbuch d vergl. u. Entwickelungslehre d. Wirbeltiere von Oskar Hertwig, Bd. C, Teil 2. Grecory, W.K. 1907 The orders of teleostomous fishes. Ann. N. Y. Acad. Sci., 17:437-508. Kinprep, J. E. 1919 The skull of Amiurus. Ill. Biol. Monographs, vol. 5, pp. 7-121. McMovrricn, J. P. 1883 On the osteology and development of Syngnathus peckianus. Quart. Jour. Mic. Sei., vol. 23, pp. 623-652. Parker, W. K. 1873 On the structure and development of the skull of the salmon. Phil. Trans. Roy. Soe., London, vol. 163, pp. 95-145. SacemenL, M. 1891 Beitrige zur vergleichende Anatomie der Fische. IV. Das Cranium der Cyprinoiden. Morph. Jahrb., Bd. 17, S. 489-595. Scureiner, K. E. 1902 Einige Ergebnisse iiber den Bau und die Entwickelung der Occipitalregion von Amia and Lepidosteus. Zeit. wiss. Zool., Bd. 72, S. 467-524. SwinnerTon, H. H. 1902 A contribution to the morphology of the Teleostean head skeleton. Quart. Jour. Micros. Sci., vol. 45, pp. 503-593. Winstow, G. M. 1897 The chondrocranium in the Ichthyopsida. Tufts Col- lege Studies, vol. 1, pp.147-201. ee iia t c as He i ; i. ny om We . @: " welin osha Phat ae! Gi 061441 , #7 Ah | Yea miharsiattsthbe, : LEAL oy eu i aly Seis af walt keel Tee) xvas BOs avid A aA AP le me ail Aly ig witlanig f MO ata “anamaisines nly wet Papen! 69 ibbiiihirg ra: % veil OGRE Ber a ei oy cul aia Se o_o Rn ye Bal ci “at BUR: Gantaly he ta 2 eG ithe (0 don hi Be elas asi fo AOL ot) Tage Wie) oa eel sab wba | WAAL. Spivey, ule Waibaad iid (the Babe. Ah tor RRS egal estan ae ine ma A) i EL CY eR fie, a We if er. ee Cay) ABU TOD Fe Whi’ f shia hewn) bil. 4! ces va 38 wobeatesien fo ay are 7 ae eed ii Rite ward wii as ROI Se- . i =! Thane 4) fle ec fern ty rise Bere hace MCh, it +) my ey its ‘At gs mi { is uy): Rays “iy Bytes = 0 A a SB ond «iar, a nin 4 ee me Sevibiticy UL 8 & -GR a yeaa Talis non 9 or ait eh Merge teta(i/ 1 Miib ics LED) gendrarey 4 Aah (3 OY uid . # pe 4 SOTTIAG AT bO},.4) 8) Siena ob oe ' u ii Uy 4 rout te aoe Ab QUOT RR Hse Oho Lapiabal w Dey Be th | Sas es Wh pit’ yi ia ; ee te : Th aN ‘gee 7 i | ’ nse: ine ok hg j Bid. VC Vec Ane eae My ee eee f Haart ty ay heathen Sie ma tj MR aE Pa al) OW can i aie ot) . es a aay he ee it? et rh You - ‘Vy wih eee bee Pe Dito: eh tS arin 4: pvt fk i va WM Se Se ' j Resumen por la autora, Eleanor Carothers, University of Pennsylvania. El] comportamiento genético de los cromosomas heteromdorficos homodlogos de Circotettix (Ort6épteros). Circotettix verruculatus posee tres pares de cromosomas claramente diferenciables, y cuyos miembros pueden ser telo- miticos o atelomiticos. Un andalisis citolégico de cuarenta machos salvajes ha susministrado la frecuencia relativa de la apariciOn de los dos tipos de cromosomas homdélogos para cada uno de los tres pares. En 28 descendientes machos procedentes de cinco cruzamientos, en los cuales se conocian los complejos cromosOémicos de los padres, se presentan 56 homdlogos (2 x 28) para cada uno de los tres pares, 0 sea un total de 168 homdélogos (3 x 56) en todos ellos, que podrian haber variado. Han debido tener lugar 22 cambios morfolégicos en dicho numero de homé- logos si existe una reorganizacién durante la ontogenia que repita las condiciones de la especie. Por el contrario, ni uno solo de los descendientes digere en su constitucion cromosémica mis alla’ de los limites que deben anticiparse al existir una combinacién de los gametos de sus padres. Las razones de combinacién de los dos tipos de hom6- logos en cada uno de los tres pares son las que resultarian de la unién al azar de los gametos de los padres. De este modo, mediante identificacién actual de los homélogos de un par dado de cromosomas desde los padres a los descendientes y con la determinacion de sus relaciones de recombinacion, el paralelismo entre el comportamiento de los cromosomas y los fenédmenos mendelianos es completo. Translation by José F. Nonidez Cornell Medical College, New York AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, MAY 23 GENETICAL BEHAVIOR OF HETEROMORPHIC HOMOL- OGOUS CHROMOSOMES OF CIRCOTETTIX (ORTHOPTERA) E. ELEANOR CAROTHERS Department of Zoology, University of Pennsylvania FIVE PLATES (THIRTY-FIVE FIGURES) CONTENTS WIMCROCUETRON 0s eee eh ey kee». dic, ied tka eklanaes le eh AACR TAY aie ol. tee 457 Meaterialifarrd same tin ods Aepep pers. « apcPavevs ots ars: crak fer ool of RE RR Tere lobe ese Sere ance 460 WSSETV ATOMS Ys 3/choya tare SER ME aroha ol ks: 5 415, elon chee ner can od See eR ered Sos aoe coe tes 461 1. Description of general chromosomal conditions in the species........ 461 Gao sy MICA CONGUTOMS A Sere nena noc eteeya cher aee e P EE Soecaee 461 b. Methods of transformation of tetrad no. 7; E-shaped tetrads..... 462 copAnvocbadbimiil Giles. acs cues cs tcsvlee omer Hie ee eRe Meaeeneeke 464 d. Differential frequency of atelomitic condition of the three critical TD DUT SS eM oe TE Good 5 See boas OEE ERE MD eran aca 465 2. Description of chromosomal complexes involved in specific matings... 466 Gr Matin gan Omen olateel) os Arch aotica sie 6 soaCheee, Taal 466 Oe aNisrbinn CRT OMMl es aU ACO Meh tac... ouctth tacks cctseure co Syslaetor ers eect Ra 468 Ca vMerbiapenOsmoeprabe ac cor 2. cceeltiee «ote 2 otideraes Seton troiane eetiers 468 ay Matingmnou 4platesias A 2). c'. 4 (de adie. outed eae sa eiters ead eats ere 469 Cam VS CIM eT Oop y ONAL Oto cyate sy 0.x « kceabyerela 4 Stentor s feelers 469 TD SVSPERTICTEV(GL O08, 8's ARSENE es Lanta echo oe ROE eC oes AMR hy Pannen So iPS 470 I Drs SHEE AVI ia 206 ie faye oe ke ae ah ar aR ht i ht eR 473 I. INTRODUCTION “Science begins with naive, often mystic conceptions of its problems. It reaches its goal whenever it can replace its early”’ guessing by verifiable hypotheses and predictable results, Mor- gan (716). The facts herein presented have been so thoroughly guessed as to cause almost a feeling of apology for presenting them. Since Mendel’s laws became generally known in 1900 so many points have been brought out showing that the behavior of the chromosomes during maturation and fertilization pro- 457 458 E. ELEANOR CAROTHERS vides a mechanism for carrying these laws into effect that the number of scientific people who question the proposition that the chromosomes are the bearers of the hereditary factors has become almost negligible. Mendel’s first law in substance is, that of a pair of contrasting unit factors contributed by the male and female parents, respectively, one will dominate the other in the first generation, but that during the gametogenesis of these individuals each member of a pair of factors segregates into a different germ cell. This is known as the law of segregation. Mendel’s second law is that when a number of pairs of unit factors are present the pairs assort independently of each other. That is, while the first law describes the behavior of the individuals of a pair, the second applies to the relation of the pairs to each other. This is sometimes called the law of inde- pendent assortment of different pairs of unit factors. Mendel, of course, was dealing with unit characters from a purely genetical standpoint. Shortly afterwards, the work of the early cytologists led to the Roux-Weismann theory of heredity, which recognizes the chromosomes as the bearers of the heredi- tary factors. Van Beneden in 1883 reported that egg and sperm contribute equal amounts of chromatin to the new individual in Ascaris megalocephala. Sutton in 1902 first clearly showed that the chromosomes occur in a duplex size series and pointed out that the behavior of the chromosomes during maturation and fertilization is such as would be necessary for carrying out the laws of heredity discovered by Mendel.! For a demonstration, however, two things were necessary: First, to find some species in which homologous chromosomes (those which unite in synapsis) differ from each other and from the remainder of the complex in such a way that they can be certainly identified. Second, to be able to breed the organisms freely in captivity in order that the behavior of the unlike homo-. logues might be studied in both parents and offspring. 1 This paper is intended to be taken in connection with the one published in 1917, deciding as it does questions left open at that time. For this reason it is not as complete in itself as it might otherwise be desirable to make it. GENETICS OF HETEROMORPHIC CHROMOSOMES 459 The first of these conditions has already been satisfactorily met; in a former paper (Carothers, ’17), based on a study of the germ cells of several species of a certain group of short-horned grasshoppers, the following facts were established: 1. Given tetrads (seven out of eleven in Trimerotropis fallax) may be composed of morphologically dissimilar homologues. 2. When heteromorphic, the members of these pairs segregate during the first maturation division. This behavior parallels that of unit characters as established by Mendel’s first law. 3. The homologues of three of these pairs were traced and found to segregate at random in regard to the accessory chromo- some and consequently in relation to each other, and presumably to the members of other pairs, thus furnishing a physical mechan- ism such as would be necessary for the carrying out of Mendel’s second law. 4. The chromosomal constitution of approximately one hun- dred wild individuals was such as would be expected from a free union of gametes bearing these morphologically unlike homo- logues. The logical conclusion was that here, with almost diagram- matical clearness, we could trace the segregation of given homo- logues to the gametes and their recombination in the zygotes. At the same time, however, the alternative possibility was pointed out that in these species there might be a reorganization at the time of fertilization which would result in a shift in point of fiber attachment and a corresponding change in the morphology of the chromosomes, such that the offspring of a single pair would tend to give the range of variation of the species instead of the range possible from a combination of the gametes of their parents. The possibility that there might be a reorganization was rendered more probable since at about the same time McClung (17) encountered relations between the octad and hexad multiples of Hesperotettix viridis which seemed to indicate a reorganiza- tion of the chromosomal combination involved in the multiples at the time of fertilization. The two cases, as Dr. McClung pointed out, are not exactly comparable, since in H. viridis non- homologous chromosomes are involved and the mechanism may 460 E. ELEANOR CAROTHERS very well be different. The solution in both instances could only be obtained by a comparative analysis of the chromosomal con- stitution of given pairs of grasshoppers and of their offsprmg. In any case, in order to complete the parallelism between the be- havior of the chromosomes and Mendelian phenomena, it was necessary to trace the behavior of the heteromorphic chromo- somes from parents to offspring. II. MATERIAL AND METHODS Since all of the species used in my 1917 work occur only in the western half of the United States, the hope was then expressed that Circotettix verruculatus, whose range extends to the eastern part of the country, might be equally favorable. This expecta- tion was justified, and combined cytological and genetical work was undertaken on this species. The stock was obtained from near Manchester, New Hampshire, July 29, 1918. Collecting during the last week of July, one obtains both adults and nymphs.’ It is necessary to obtain the females as nymphs to insure their being virgin. Males and females were kept in separate cages until about September 1, when individual matings were made up. After eggs had been laid both parents were killed and the gonads fixed; the testes in strong Flemming, the ovaries in picro- formol-acetic. Of eighteen matings, one or the other of the parents died in six, and these cages were discarded; four of the twelve remaining cages contained no eggs, due perhaps to the exhaustion of the ovaries of the females before the matings were made up (some of the females had laid at least ten days previously). The remain- ing eight matings gave 138 offspring. Fifty-one was the largest number obtained from one pair. Eggs from three pods hatched from this mating. Twenty-eight male offspring from five of the matings have been. studied cytologically, and it is believed that they furnish sufficient evidence as to the point under consideration. 2 It is intended to give details of the rearing and postembryonic development of C. verruculatus in a separate paper. GENETICS OF HETEROMORPHIC CHROMOSOMES 461 III. OBSERVATIONS The principal object of this paper is not to trace the spermato- genesis of Circotettix verruculatus, but to show as graphically as possible the genetic relationship of the chromosomes of the male and female involved in particular matings to those of their offspring. The plates accordingly are made up with the first spermatocyte chromosomes used as the standard, the sperma- togonial complex and the somatic complexes of the females being rearranged so that supposedly homologous chromosomes? fall in vertical rows. iI GD dAIHt dod nba Gah onpo frets PLATES 3 AND 4 EXPLANATION OF FIGURES 15 to 28 Complexes involved in mating no. 14. 15 First spermatocyte of father; tetrad no. 7 is a heteromorphic atelomitic. 16 Somatic complex of mother; pairs no. 7 and no. 8 heteromorphic. 17 to 28 First spermatocyte complexes of twelve sons; tetrads no. 7 and no. 8 show the various combinations in different individuals to be expected from a ’ free union of the gametes of the parents. 478 » Got Ol 41 » Caoos pri ~ $97 Le %al ; E)tdoj*\t PLATE 3 iV x, ivy % % i ‘ é i = ¥ ; ak | : a F bade y { é PLATE 4 See description of preceding plate _ ; f ) ies fg y ™- , ea] 1 , £ 4 | ' ? wo} 1 ' X Mi) : it oor J ‘ * y ive ; i ; : ta ee ‘ a . ’ 5 ; %, i o i : ‘ } \ : j : . i ‘Bh ad f i i L ‘ Gq “ay , i ‘ wrt Fy } : ov ‘a i wi + ‘ GENETICS OF HETEROMORPHIC CHROMOSOMES E. ELEANOR CAROTHERS he MU On On 2Oer i (Gams, A eS Daboser(s ~ ppelelefe = PEELE af = GOGGLE ~ PORRPE +4 = $400+["+< PLATE 5 EXPLANATION OF FIGURES 29 to 35 Complexes involved in mating no. 17. 29 First spermatocyte complex of father. ' 30 to 35 First spermatocyte complexes of five sons; pair no. 1 must have been telomitic in the mother to give these results. 482 GENETICS OF HETEROMORPHIC CHROMOSOMES PLATE 5 E. ELEANOR CAROTHERS |2 I] 10 9 8 eo Oi a ee eS es al Ae atecdy » BGR0KITPE oF vig yt i Ae it ett ray CYTOLOGICAL STUDIES ON THE INTERNAL SECRETORY FUNCTIONS IN THE HUMAN PLACENTA AND DECIDUA GENCHO FUJIMURA Osaka Medical College, Osaka, Japan TWO DOUBLE PLATES (NINETY-SIX FIGURES) AND ONE TEXT FIGURE CONTENTS TET OONUC ELON: f o05-. 5 sc ORO Oa ois « Dose a aia bo eee TSE cae oeaeiae ae 486 Materials }andumethodsigmmmascpice sos sia ee aude k eden aaio Ooo ete eines 491 VEO WNEODSEL Va LION emer rare pare sicres: ceeoeos PEN cre tore CECE eben cee 493 ieaihesyncytiumecelilaventen.. . . c c.ssutctiide coma Or ere eect 494 Zr he Aian Wan geiCellsh yompatryed © «h-ch ents Ske yin reve a A Slane ee Rice olsnaecre 500 3y7 Lhe, stroma cellsyos valle 122 S.c)c54 SAR Doe, Uae oats eae aus be 502 reed PLAYS) ra wexen (0 AUT) WYeCel ENS as RU OLS Rear RMRE 4 ORB RPS ELTA i CME NE Me 504 j-eLhe epitheliumypotauhequtenine ol amare sissies otro ee ciatia ronicie 511 The histological structures and their functionary significance (some reflec- tonsPonpliteratume) yy Meets etl ete ees dc TS 5 tds 6 oe SRD, Bae be 514 The phenomena of internal secretion in various cells of the placenta and OCLC eee esha Peis aici ea) ct days os Heater eschews eich austere eas Miestale ie 530 1. The phenomena of internal secretion in the syneytium layer......... 530 2. The phenomena of secretion of the Langhans’ cells.................. 533 3. The phenomena of internal secretion in the stroma cells of villi..... 536 4. The phenomena of internal secretion of decidual cells............... 538 5. The phenomena of secretion in the uterine glandular cells at the time CO} gy ON ECEY eA OVE OK CAI) Ebon LO OL eee Oy SY Ot 2 Bee eP EE OR REE UO be 543 SUTRAT Yr. ae Pat eee APSE EL A Mac's ine, oS 5th SR a eine bal ee 544 The minute histological structure and phenomena of internal secretion in the uterine mucous membrane prior tO MENSES.............0 cece eee eee eee 550 1 Miy: Own ODSORV-A TIONS seat tices «dso 5 sR IS ee hate eee 550 qe therchangestonumhenmterstibialecell Shas reer. hes Oey eee 550 6, The chanvesofitineslancdular, cells: 455 erpa isa. ps avon ieee 552 2. Lhe phenomena of intemal secretion. -accnesiie sta oc oe ae sepa tise 553 WONCIISION | o's scp See Cee ale hc es Ni RENO RRC ae Rcit Pec oot On 505 DAVeLAGUTe’ C1Led)..ocs scart eee RECON os bs ce ee MME etd RUN eo Sua eels 562 485 JOURNAL OF MORPHOLOGY, VOL. 35, NO. 3 SEPTEMBER, 1921 486 GENCHO FUJIMURA INTRODUCTION In recent times, along with an increase of knowledge on internal secretion, it has been imagined by many authors that in the placenta and the decidua there should exist such a func- tion. According to the literature which I know, Lettule and Larrier (01) are those whose attention was first drawn to this subject. They detected in the syncytium layer a kind of gran- ular body termed ‘Plasmoidale Kugeln.’ This body they took to be a secretion of the placenta. Subsequently, Veit (02) attempted to trace the cause of eclampsia, and Behm (’03) the cause of morning-sickness to the function of the syncytium, and Bouchacourt (’03) observed an increase of lactation by using some placental preparations. Further, Halban, supported by his plentiful clinical observations, stated that the swelling of the nipples noticeable in a newborn child, the congestion and hyperplasy of the uterus, the hypertrophy of the mammary glands of a pregnant mother, and hypertrichosis are all probably due to his so-called ‘Reizstoff,’ which is thought to be a product of the placenta. Among other things, it seemed he tried to deduce the existence of the closest functionary corre)ation be- tween the placenta and the mammary glands. Since the result of Halban’s studies was published, the functions of internal secretions likely to exist in the placenta drew general attention, and thereafter studies of this subject followed quickly one after the other. In the present essay I have refrained from chronologically relating all the results of these researches, but instead, with a view to giving only a general idea of what . is known about this subject at present, I have confined myself to the summing up of all the points of the investigations made by the various authors up to now, and to dividing them into a few large sections along the lines of experimentation, biolog- ical chemistry, and histology. Now, in the early experimental investigations, it was the first attempt of the various authors to investigate chiefly the effect which the placenta has upon the mammary glands, and the methods employed were either to inject into the animal used as the subject of the experiment an extract of the placenta SECRETORY FUNCTIONS IN HUMAN PLACENTA 4S7 or to transplant the placenta, and the results in either case were for the most part positive (Feller, 09; Lederer and Pribram, 710; Aschner and Grigoriu, ’11; Cristea and Aschner, ’12; Basch, 712; G. Kawaida, 712; M. Dohi, 716). However, as it was incidentally discovered at the same time that the placenta had shown some strong special reaction upon the other organs, e.g., the vascular organs (Schickele,’12), the internal secretory glands (Fellner, 713; Colle, 713), and the uterus (Fellner, ’09; Okintschitz, ’14; Hermann, ’15), it naturally became difficult to assert that the ingredients of the placenta have such proper- ties as react upon the mammary glands alone. And, further, as it became known generally that such changes in the mammary glands as mentioned above were not only due to pregnancy or the placental ingredients, but were also caused by extracts of the various organs, such as the embryo (Mandle, ’05; Bayliss and Starling, 06; Foa, 710; Biedl and Konigstein, 710; Aschner and Grigoriu, 710), the ovary (Ott and Scott, 710; Aschner and Grigoriu, 711; Frank and Unger, 711; Hermann,’ 15; Y. Tani- guchi, 716), the pituitary body (Hofstitter, ’11), the pineal and thyreoid glands (Ott and Scott, ’11),the uterus while in child- bed and the mammary glands which are giving suck (Schoefer and Mackenzie, 711), the intestines, the testicles, the spleen, and the thymus (Kehrer, *15), the decidua (Gentili and Bina- ghi, ’17), as well as by a certain kind of chemical matter, such as ‘Lymphagoga’ (Aschner-and Grigoriu), and albumin, too (Frank and Unger, 711; Fraenkel, ’14), the theory of internal secretion in the placenta which had begun to be adopted gen- erally for a time began to lose its value by degrees. In the next place, along lines of biological chemistry, there are the researches made by such authors as Higuchi (’09), Hermann (715), and Harada (’16). These authors extracted from the placenta various kinds of chemical matter as its principal ingredients and subjected them to close examination, but they never mentioned a ‘hormone,’ which is always existent in the placenta. Thirdly, with regard to the histological studies of the placenta, many researches have been made in the subject from earlier 488 GENCHO FUJIMURA times, but a great majority of these researches were confined to the development or general construction of the placenta, and very few authors have given any special consideration to the internal secretion of this organ nor have they closely examin- ed the minute construction of the tissue elements or cells of the placenta, Already Ahlfeld (78) had noted the appearance of vacuoles in the syncytium layer, and subsequently many authors recognized it; however, regarding the thing itself and its physio- logical significance, nothing definite has yet been stated. Gotts- chalk (’90) deemed it a pathological product; Kossman (’92) took it for a kind of degeneration, while Langhans (’92) inter- preted it as ‘Leichenerscheinung.’ Further, regarding the glob- ules of fat in the same layer, demonstrations had already been made by Pela-Leusden (97) and Marschand (798). Bonnet (99) deemed it a nutritious matter taken up by the embryo, and in recent times authors have generally agreed in recogniz- ing it as nutritious matter of the embryo taken from the mother’s blood. Others, among them Halfbauer (’05), Costa (04 and 05), and Bondi (711), inferred that the appearance of this matter must be due to the active functions of the cells, by which it is assimilated and absorbed. On the other hand, however, Wolff (713), as per Letulle and Larrier as mentioned above, on com- paring the granular body of the syncytium layer with the secre- tory grains within the ordinary gland-cells, stated that both of them were a similar production. This theory was subsequently supported by Fraenkel, who was close to the deduction that there might be existent a certain secretory relation in the granu- lar body of the syncytium layer. In short, the histological studies of this subject up to now have been confined, as men- tioned above, to the construction and, consequently, the functions of the syncytium layer only, and there are even many defects in the studies and some resultant weakness in the point of the arguments, and no one can say that the opinions agree. Al- though there are some authors who recognized the internal secre- tory functions of the placenta, yet, since their arguments are based on a single part of the organ, viz., the syncytium layer, it must be stated that on the whole the basis of histological SECRETORY FUNCTIONS IN HUMAN PLACENTA 489 arguments for internal secretion in the placenta is very feeble indeed, and much more is still to be done to accomplish a consummate investigation of this subject. It should none the less be added that Fraenkel (14) was one of those who were strongly opposed to the various theories given above, and he denied the effects, which, it was believed, the placental func- tions have upon the mammary glands, on the ground that ex- actly the same phenomenon as takes place in the suckling of mammals was observed in the suckling of Monotremata and Marsupialia, neither of which has the placental formation. In addition to that, Fraenkel, going so far as to discuss the changes which the internal secretory glands and other organs have to undergo in consequence of the placental extract as was witnessed in the experimentations of the afore-mentioned authors, concluded that such changes were attributable to several ingredients, his so-called ‘Gift,’ contained in the extract. According to Fraenkel, who contradicted the theory which generally is in favor of the existence of internal secretion in the placenta, the internal secretory glands first respond to this ‘Gift,’ and, as a result, the other physical changes follow. In short, it may be said with regard to the existence of internal secretion in the placenta that, in spite of the numerous researches made along the lines of experimentation, biological chemistry, and histology, as I have mentioned above, the views expressed are so varied that no conclusion has been reached, and, therefore, regarding the real situation of this subject, it is at present not yet an estab- lished theory. Many authors have given their attention to the existence of an internal secretory function in the decidua, and have tried to demonstrate it, Starling (06), Sfameni, Gentili (13 and 714), Schottlander (14), Aschheim (15), Gentili and Binaghi (17) being among the number. Starling conducted experiments on the rabbit to find the effects which the juice extracted from the mucous membrane of the uterus while in pregnancy exerts upon the mammary glands, and the result was of a negative nature. Sfameni. for many years past had held that in the decidual cells there should be existent an internal secretory 490 GENCHO FUJIMURA function, the basis of his arguments being the striking resemblance from the view-point of morphology, between the decidual cells and those of the other internal secretory glands. In order to demon- strate this, Gentili, in the same laboratory as Sfameni, carried out experiments by the use of the decidual juice on the dog, the rabbit, and the frog, and it was found that, similar to the luteal cells of the ovary, the decidual cells have a special action upon both the circulatory and generative organs. Schéttlinder also, by assuming that it may be possible for the secretion of the glandular cells in the spongious layer of the decidua in the first month of pregnancy perhaps to enter the mother’s blood directly, recognized seemingly that the uterine glands at times have the significance of an internal secretory gland. Aschheim, in 1915, by discovering plenty of lipoid in the decidual cells, imagined the existence of a special function in these cells. Last- ly, Gentili and Binaghi, by conducting a microchemical experi- ment on the various tissue elements which form the decidua of a cow, demonstrated the existence of a kind of lipoid in the tissues, as reflexive of the correlation which it seems they believed exist- ing between a certain function, particularly an internal secretion of the decidual cells, and the lipoid. On a close examination of the result of researches made by these authors, it must be stated that even in those whose view was in favor of an internal secretion in the decidua the basis of arguments presented is in general as feeble as in the case of an internal secretion of the placenta. It is a long time since I began to feel interested in the secre- tory function of the placenta and the decidua. I was ever of the opinion that both the placenta and the decidua have a certain secretory function, and the experimental method I adopted in investigating this subject was entirely different from that employed by the foregoing authors. That is to say, I con- ducted a serological research into the effects which the alchoholic extract of the placenta and the decidua has upon the mother’s blood, and the result was the discovery in the serum during pregnancy of, not only the well-known Abderhalden’s ‘ Abbau- ferment,’ but also of such an antibody as has a property of the fixation of complement with ingredients of the extract SECRETORY FUNCTIONS IN HUMAN PLACENTA 491 referred to. And,asin ordinary cases, the'antigen which represents the phenomenon of fixation of complement is either a proteid or some such like matter, whereas in my case it is a substance which is soluble in alcohol, the latter must be deemed to be qualitatively different from the former. It makes me feel as if it were in order for me to presume that the antigens in my case are due tothe lipoid substances which are peculiar to (having some relation with the secretory process of) both the placenta and the decidua. Accord- ingly I have become convinced that, in spite of the various authors whose arguments I have enumerated above as denying the exist- ence of an internal secretory function in both the organs concerned, the fact must be the reverse. Needless to say, however, this argument is only along lines of reasoning, and, therefore, it must be stated that the reason why I have applied at this juncture the modern cytological methods, thus planning out a close histological investigation of the principal tissues and cells, partic- ularly the cell bodies of the placenta and the decidua, was because I was anxious to decide more clearly the right or wrong of this hypothesis. I have also, by adopting the same methods, exam- ined the minute structure of the principal cells as indicative of the change which the uterine mucous membrane undergoes prior to menstruation, which, being compared with that during preg- nancy, has enabled me to arrive at a certain conclusion, as will be noted later, respecting the physiological significance of the menstrual changes of the uterine mucous membrane. MATERIALS AND METHODS The materials for research were taken from a total of forty- three cases, of which twenty-five had to undergo artificial interruption during the first half of pregnancy because of the following diseases: 6 cases of morning-sickness of high degree, 12 cases of phthisis, 2 cases of laryngeal tuberculosis, 1 case each of consumption of the bowels, of peritoneal tuberculosis, and of a valvular disease of the heart, 2 cases of glucosuria. Of the remaining 18 cases, 14 had to undergo artificial inter- ruption during the second half of pregnancy because of the follow- 492 GENCHO FUJIMURA ing diseases: 1 case of beriberi, 5 cases of nephritis, 2 cases of a valvular disease of the heart, 1 case of albuminuric retinitis, 3 cases of eclampsia, 2 cases of placenta praevia. And the remaining 4 cases had to receive a Porro’s operation because of the following diseases: 1 case of mislaid transverse position and 3 cases of the narrow pelvis. And as to the time of pregnancy in these 48 cases, it should be noted that 4 cases were less than 1 month, 2 cases 1 month, 4 cases 2 months, 7 cases 3 months, 6 cases 4 months, 3 cases 5 months, 4 cases 6 months, 3 cases 7 months, 3 cases 8 months, 4 cases 9 months, and 3 cases 10 months. And difficult as it was to accurately determine the exact time of pregnancy in each case, the method I adopted in determining those less than one month was to take into account the size of the egg and the de- gree of development of the embryo and villi, and in those one month and upward, to consider the time of menstruation, the size of the uterus, the length and weight of the embryo, the length of the navel string, and the weight of the placenta, thus arriving at the approximate time of pregnancy. As regards the obtaining of the materials from the fetal pla- centa, it should be noted that in the early stage of pregnancy a few pieces of ordinary villi from the surface of the ovum were taken, while in a little more advanced a few small cuts of the chorion frondosum, and in a well-formed stage of the pla- centa a few small bits of the latter have been cut out. Now, as to the maternal part of the placenta, a part was taken from placenta already delivered and from that not grown up yet, in the former case, from the surface opposite the uterus of the placenta, and in the latter, the mucous membrane of the uterus concerned being scratched off quickly, as during an opera- tion it is quite easy for the operator to determine the insertion of the placenta. The tissues obtained in this way have always shown microscopically, besides the proper decidual tissue, the existence in them of parts of villi, syncytium cells, and the de- posit of fibrinoid material (‘kanalisiertes Fibrin’ of Langhans), all of which were proof that they were the materials I desired. In this connection it must be added that the decidual materials SECRETORY FUNCTIONS IN HUMAN PLACENTA 493 were always taken from the decidua vera only, at every stage of pregnancy, by means of scratching out. All these pieces of tissues were so taken that they should not exceed a cube 3 mm. in size, and immediately after cutting they were dipped in the newly prepared fixing solutions. For the latter I have used: 1, Altmann’s fluid of potassium bichromate and osmium tetroxide; 2, Flemming’s solution modified by Benda; 3, Flemming’s solu- tion modified by Meves; 4, Levi’s mixture of formol, osmic acid, potassium bichromate, and corosive sublimate; 5, Luna’s mixture of formol, potassium bichromate, and glacial acetic acid. Each of the pieces was then treated in the usual manner being imbedded in paraffin and cut in sections of 3to4u. And asa treatment prior to staining, Rubaschkin’s method has been applied in a great majority of cases. For staining I have employed: 1, Altmann’s method of acid fuchsin and picric acid; 2, Heidenhain’s iron-alum- haematoxylin, and, 3, Benda’s alizarin-crystal-violet method, ete. All the methods which I have adopted, the fixation with Levi’s mixture, and the staining with iron-alum-haematoxylin have been found to be comparatively successful. MY OWN OBSERVATIONS The chorionic villi, which are a main component of the placenta» consist of: 1) the syncytium cell-layer; 2) the Langhans’ cells’ and, 3) the ‘Stromazellen,’ being cells in the stroma of villi while the components of the decidua serotina and the decidua vera chiefly consist of, 4) the decidual cells and, 5) the epithe- lium of the uterine gland. In the present chapter I intend to give a detailed account, mainly from the histological standpoint, of these important tissue components, or cell groups, especially as they appear in the different stages of pregnancy. However, since the structure of all kinds of cells varies even in the same stage of pregnancy and in the same group of cells, it would be a difficult task indeed to explain clearly the correlations existing between the minute structural changes of these cells, or cellular groups, and their functional significance. Therefore, I have 494 GENCHO FUJIMURA selected and drawn those deemed the most representative of all the structural images of the afore-mentioned tissue-cells which have been widely and thoroughly examined, as will be seen from the series of figures on the plates the object being to give a general idea of the structural changes of these cells. The explanations of each of these figures will make the pith of this chapter, as by so doing I believe the description could be much simplified and its understanding facilitated as much as possible. Thus, I shall ap- preciate very much if the reader will constantly refer to those figures while reading. (1). The syncytium cell-layer (figs. 1 to 12) As is well known, the syncytium layer is the epithelium which covers the surface of the villi, and the lack of boundaries between the cells is a feature of this layer. And it is also well known that within this layer there are several kinds of nuclei, differing in size and form and scattered here and there, besides dark-colored granular bodies and a great number of vacuoles which occasion- ally appeared in it. Its surface is covered with a brush-like border. ‘This layer is generally well developed in the early stages of pregnancy, but in the second half of pregnancy it becomes thinner and looks much like an endothelium, so that, when examin- ing its minute structure, it will be necessary to do so before the fourth month of pregnancy. Figure 1 shows that part of the anchoring villi which extends deeply into the decidua, while all the other plates show the different parts of the surface of the ordinary villi. In figure 1 the syncytium presents a homogeneous proto- plasmic layer generally dark-colored and contains an extremely large quantity of plastosomes (mitochondria). They are of different shapes, but mostly are rod-shaped or bacteroid and of different lengths, the longer ones being slightly curved. They arrange themselves in groups rather than being equally distrib- uted over the layer, and in some places some of them point in the same direction, while others point in various directions, SECRETORY FUNCTIONS IN HUMAN PLACENTA 495 thus in general giving them something of a meshy arrangement. The nuclei are clear and have in themselves a more or less con- spicuous nuclear network, with one or two nucleoli. In figure 2 the syncytium layer closely resembles figure 1 in structure, though the plastosomes contained therein differ from one another in their shape, arrangement, and number. These two figures show the simplest structure of the syncytium layer. In figure 3 the plastosomes are generally faint and quite scattered; in the protoplasm there are some dark-colored gran- ular bodies of different sizes and a small number of vacuoles; the smaller granules are somewhat dark in color and are gener- ally found very near the surface, viz., the brush-like border, whereas the larger ones are light-colored and are found in other parts of the layer. The vacuoles are found close to the Lang- hans’ layer. The nuclei are irregular in shape, and besides the nuclear network there are one or two nuceoli. In figure 4 the plastosomes are generally found in the deeper layer, i.e., close to the Langhans’ layer, and they are comparatively small in number. On the contrary, however, plenty of dark or yellowish dark-colored granules occur conspicuously all over the layer, the deeper colored ones being generally superficial. Undoubtedly, granular bodies of this kind have grown up from the dark-colored granules in a con- spicuous manner such as I have shown in figure 3 above, and they are very frequently met with elsewhere in the other parts of the syncytium layer. And, moreover, granular bodies of a similar kind are found, as will be seen in the following state- ment, not only in the syncytium layer, but also commonly in other cell groups. These granular bodies are, of course, extreme- ly varied in their size, quantity, and color; however, since they have a common affinity to certain chemical and coloring matter, e.g. osmic acid, iron-alum-haematoxylin, and acid fuch- sin, etc., I have followed, for the sake of brevity, the precedents of many histologists in including all these granular bodies under the name of ‘lipoid’ granules. In this figure there is, moreover, only one vacuole close to the brush-like border. The nuclei are less conspicuous in their network, the chromatin forming 496 GENCHO FUJIMURA itself into a large number of lumps, each of nearly equal size. In figure 5 there are no plastosomes to be found, but the lipoid granules occur in extremely large quantities, and their sizes are nearly the same. They are found more or less in groups and are distributed all over the layer. At the same time the vacuoles make their appearance in a conspicuous manner, some- times on the surface, sometimes in the innermost part, and some- times in the middle, and they are about the same size as the lip- oid granules. As is well known, it is in general very difficult to stain the plastosomes every time, and, therefore, accurately to determine their existence and where they are entirely want- ing, as in this figure, it would be a very difficult task indeed. I have paid the closest attention to this, and have always selected for it the most excellent preparations for staining, with a view of doing away with all the possible defects in the technique of: staining. From the existence, in a very conspic- uous manner, of plastosomes in the neighboring tissues, entirely in contrast to this figure, I was prompted to conclude that it was well nigh necessary for me to assert the absolute lack of plastosomes in this part of the syncytium layer. Further, I may add that, as will be noted below, the same amount of-atten- tion has been given all the other cells where there are no plas- tosomes to be found, and I, on this score, am convinced that my observations concerning them are not erroneous. In figure 6, the surface is somewhat light-colored and is clear. In it there are found innumerable quantities of vacuoles which are nearly of the same size, besides a small number of lipoid granules. The innermost part, however, is of a comparatively dark color, and it likewise contains innumerable quantities of minor lipoid granules of about the same size, with very few vacuoles which are generally of small size. Some of the nuclei are oval, while others are irregular in their shape, and there is to be found a great number of chromatin granules which make their appearance in lumps of different sizes; the nuclear networks are usually less conspicuous. There are absolutely no plastosomes to be found. SECRETORY FUNCTIONS IN HUMAN PLACENTA 497 In figure 7 the syncytium layer is extremely thick, and it is difficult to demonstrate the plastosomes. Lipoid granules of extremely varied sizes are found in large quantities. These granules are irregularly arranged, and they tend more or less to occur in groups; certain lipoid granules make their appear- ance as contents of vacuoles, in which case the granules always have a clear halo around them, as if they constituted the nucleus of the vacuoles. Such images are frequently met with not only in the syncytium layer, but also in other cell groups and, since they are worth recognizing as very clearly indicating the relations which exist between the lipoid granules and the formation of vacuoles, the reader’s special attention is hereby drawn to this point (vide the left-hand side of this figure). The vacuoles are abundant, and their sizes and shapes are quite irregular and, as will be seen in the middle part of this figure, a great number of them are joined together at some places without boundaries being noticeable between them. From the existence of such images I am led to infer that an extraordinarily large vacuole is in general the outcome of minor vacuoles being agglutinated and joined to one another. In this way, it seems that the vacuoles which have grown up into tremendous sizes finally rupture toward the surface, as it will be seen in the present figure that such vacuoles open up and connect directly with the intervillous spaces, clearly supporting the interpretation that I have given above. In figure 8, it is in general extremely deficient in its character- istic dark-colored protoplasm, but then there are all over the layer numberless vacuoles of a very small size, which grow so close to one another that they have exactly the appearance of a beehive. Of these vacuoles some of the larger ones lie to- gether close to the surface, while others, connected with one another, mutually find their outlets to the surface through comparatively large openings. Because of these openings the surface of the syncytium layer, which is naturally level, becomes very uneven and irregular. The lipoid granules are very few, and no plastosomes are to be found. The nuclei are not only 498 GENCHO FUJIMURA few, but being shrunken are changed into homogeneous bodies _ of small size, the nucleoli alone making a prominent appearance (vide the right-hand side of the figure). In figure 9 the syncytium layer is comparatively thin, and there are comparatively few plastosomes to be found, being scattered here and there, and mostly rod-shaped. The lipoid granules are middle-sized and are not many in number, and a few of them stay at the center of the vacuoles as if they were the nuclei of the vacuoles. The vacuoles are pretty large, and they arrange themselves close to the Langhans’ cells. The nuclei have distinct borders and masses of chromatin arrange them- selves in rows, usually close to the nuclear membrane. In figure 10 the protoplasm is, as in a majority of cases, generally dark-colored, though not homogeneous altogether and, on a close examination, it is found that it contains a great number of vacuoles, which gives the protoplasm more or less a foamy appearance, though very indistinct as compared with other foamy structures. The plastosomes are mostly rod-shaped and are very distinctly stained. They appear generally in the upper layer, though some are noticeable in the innermost layer. There are no lipoid granules to be found and no nuclei of normal condi- tions are to be met with, but, on the contrary, there are some curious bodies, whose size is somewhat larger than the or- dinary nuclei and which are irregular in shape. Some of them are homogeneous and are quite dark-colored, while others being non-homogeneous consist of different parts which are either dark or light or somewhat clear, when stained: We come across such structures very often in other parts of the syncytium layer, but so far I have not been able to find out their original nature. In figure 11 the protoplasm shows numberless vacuoles as its constituents. The vacuoles are somewhat varied in their size, and with the exception of some which are full and stained, a greater portion of them, particularly those which are found on the surface, are more or less loose and somewhat flattened in shape. Between these vacuoles there are extremely large quantities of plastosomes, which are mostly rod- or granular- SECRETORY FUNCTIONS IN HUMAN PLACENTA 499 shaped and equally deep-colored as these which are found in figure 10. There are few lipoid granules to be found, and they exist for the most part in the superficial layer, scattered here and there. What is especially worth noting is that there are red blood corpuscles in the protoplasm between these vacuoles (vide the left-hand side of the figure). The nuclei are oval-shaped, and the chromatin is very peculiarly arranged, its outward appearance resembling the shape of a chrysanthemum. In the center there is a nucleolus, and it must be noted that a nuclear condition of this kind is generally very rare. In figure 12 the protoplasm is glutted with numberless vacu- oles, and it is for this reason remarkably foamy in appearance. The vacuoles are of two sizes, of which the smaller ones are mostly located in the deeper portion and make a somewhat continuous layer, though in other parts there are to be found some of these smaller-sized vacuoles also. The larger vacuoles are crowded together in the middle part of the layer, and some of them burst forth onto the surface, while others make their appearance in the innermost layer close to the Langhans’ cells. The plastosomes are quite uniform in shape with those illus- trated in the preceding figure, and they are all found in the pro- toplasm between the vacuoles. The lipoid granules are nearly of the same size and are found at several places, some of them with the halos distinctly described. The nuclei appear to be somewhat smaller in size, but there are no remarkable changes in their structure. These structural conditions in the syncytium layer which I have illustrated and described above can be detected at almost any time and place at the different periods from the first month of pregnancy to nearly the fourth, and, therefore, there is no doubt that these structural changes cannot be taken as a measure to tell the precise time of pregnancy. At no time after the fourth month of pregnancy can we detect the plastosomes. .The lipoid granules and vacuoles reach their maximum growth from the second to third month of pregnancy, and after the fourth month they gradually begin to decrease, entirely disappearing after the seventh month. The syncytium layer comes in sight in a 500 GENCHO FUJIMURA remarkable manner on the seventeenth or eighteenth day after pregnancy, reaches the maximum of growth at the end of the first month, and from the second to the third month it seems, although not very conspicuously, more or less reduced in thick- ness, but after the fourth month it suddenly becomes thinner, and simultaneously its structure becomes simplified, and in the seventh month it will altogether atrophy and remain simply in the form of a thin membrane like an endothelium. Moreover, in the last two months the layer will disappear in some places and will be noticeable only as a discontinuous thin cover. In other words, this layer, excepting in the early stage of pregnancy, always decays and goes out of existence in inverse proportion to the growth of the embryo, and this is the very point which should engage the careful consideration of those who are inter- ested in the functions of this layer. 2. The Langhans’ cells (figs. 13 to 26) The Langhans’ cells have in general distinctly clear bodies, and are distinctly bordered with a thin membrane (pericula?) on the surface. Their sizes, shapes, and structures are extremely varied; on examination of the smallest cells (figs. 13 to 16 and 7, 8 and 10) it will be found that they are either round like a ball or oval-shaped, with foamy nuclei of corresponding shapes inside. The structures of the cell bodies consist of quite struc- tureless stroma and a large quantity of plastosomes, of which the latter are rod-shaped in several lengths, and are usually dis- tributed all over the cells, though sometimes they crowd together on one side of the cells. There is occasionally a small oval- shaped body somewhat dark in color close to one side of the nucleus, which might possibly belong to Meves’ so-called ‘Centrotheca;’ it, however, lacks a centriole within (fig. 13). Within the nuclei there are generally one or two conspicuous nucleoli, and in most of the cases it is difficult to discern the nuclear network. In the large sized cell bodies (figs. 17 to 19 and 2, 3, 9, 11 and 12) we always find either a small quantity of lipoid granules or vacuoles. The lipoid granules are nearly the same in size, and, though small in number, they are scattered SECRETORY FUNCTIONS IN HUMAN PLACENTA 901 everywhere (figs. 17 and 11). The vacuoles are extremely va- ried in their size, quantity, and arrangement, and it is for this reason that the structures of the cell bodies have so many special features (figs. 18, 2,3,and9). Insuch vacuolar cells the rod-shaped plastosomes are for the most part short in length, and they are arranged either along the walls of the vacuoles or crowded together in the protoplasm between the vacuoles; however, sometimes it will be found that, as will be seen in figure 18, the vacuoles are chiefly placed in order on the outskirts of the cells, and the plastosomes accumulate in the center around the nucleus. Again, it will be found that, as in figures 19 and 12, both the lipoid granules and vacuoles of various sizes are simultaneously con- tained in the cell bodies, in which case the plastosomes are com- paratively small in number and are scattered here and there in the protoplasm between the lipoid granules and vacuoles. In the largest cells (figs. 20 and 21) the cell bodies are in gen- eral well filled with a great number of vacuoles of various sizes, and consequently the protoplasm is noticeable only around the nucleus and between the vacuoles. There are almost no lipoid granules, which, when present, are small and are very few in number; also the plastosomes decrease in quantity and are mostly found around the nucleus. In short, the smaller-sized cells have in general only the plastosomes as their material components, while the larger- sized ones still contain a number of lipoid granules and vacuoles and in the largest ones the cell bodies are commonly vacuolated in a high degree and the protoplasm decreases considerably in quantity, the plastosomes in general gradually decreasing in number as the cells grow in size, though it sometimes happens that it is very difficult to demonstrate them, regardless of the size of the cell bodies. Of figures 22 to 26 it will be observed that figure 22 shows the lipoid granules only, figure 23 chiefly the small vacuoles and a few lipoid granules; in figure 24 it is entirely the same as the former, however, with the distinctive feature that the vacuoles are remarkably large and have lipoid 502 GENCHO FUJIMURA granules of various sizes within; in figure 25 and 26 the bodies are vacuolated to the highest degree, and it is perfectly plain that the vacuoles which are extremely varied in size and shape mix together and grow larger, thus giving the cell bodies the appearance of a honey comb in a striking manner. The large and highly vacuolated cells such as are illustrated in figures 21, 25, and 26 are very frequently met with in the Langhans’ islets. The various structural images in the Langhans’ cells that I have described above make their appearance in a most remark- able manner from the.end of the first month of pregnancy to the end of the third month ,and in the fourth month, though the plastosomes are still existent in a remarkable degree, the lipoid granules and vacuolar formations are no linger conspicu- ous, and in the following months not only do the cells decrease suddenly, but also these materials component of the cell bodies disappear, though the cells in the Langhans’ islets retain those structures as long as they exist. 3. The stroma cells of villi (figs. 27 to 38) The smallest of the stroma cells of villi, as is illustrated by figure 27, are mostly ball-shaped and have the nucleus of a similar shape within. In the cell bodies there are plastosomes, mostly rod-shaped. Figures 28 and 29 are nearly the same as figure 27 in their shape, though the one contains in the cell body a somewhat large quantity of lipoid granules of different sizes, while the other has a small quantity of extremely small lipoid granules and an equally small quantity of small vacuoles. Figures 30 to 38 illustrate the cells arranged in their approximate order of size and, though their shape and structure appear ex- tremely varied at a glance, it will be found on close examination that, with the exception of figure 34, there is a structure which is common to nearly all the rest, the only difference between them being principally the size and number of vacuoles contained in the cell bodies. Generally speaking, the larger sized-cells have vacuoles which are naturally large in size and number, and the fact that large vacuoles are built up to some extent from SECRETORY FUNCTIONS IN HUMAN PLACENTA 503 the fusion of the smaller vacuoles which are connected with one another can be often proved in these stroma cells (fig. 38). The plastosomes are mostly rod-shaped and lie scattered in the protoplasm located between the vacuoles, though they some- times crowd together in a somewhat larger number in certain places (figs. 31, 32, 36, and 38). The lipoid granules are gener- ally few in number and are found between the vacuoles, though at times they are present within the small vacuoles (fig. 36). The smallest of these lipoid granules, at a glance, bears a close resemblance to the granular plastosomes; however, since they are mostly very distinctly bordered, besides being stained darker, it is easy to distinguish them from the former (figs. 36, 37, and 31). Figure 34 looks somewhat different from the vari- ous cells described above in that the cell is nearly spherical, with a remarkably large nucleus within, besides a small number of somewhat large vacuoles and plenty of lipoid granules in the cell body. These granules have various sizes, but are in general of middle size and a few of them are distinctly included in the vacuoles. The plastosomes are extremely limited in num- ber, and le scattered in the protoplasm between the vacuoles. It is very seldom that this kind of cells makes its appearance, and a great majority of cells in the stroma, as are chiefly illus- trated by figures 35 to 38, present a distinct vacuolar formation. The plastosomes in the stroma cells have in general a very strong staining power, and are therefore more easily detected than other cell groups. It is, moreover, very rare that the cells which have no plastosomes are met with, but in stroma cells the lipoid granules seldom appear. The stroma cells appear distinctly and are therefore most easily found during the period from the second ‘to sixth or seventh month of pregnancy. In the eighth month, usually, numberless capillary blood vessels suddenly grow and increase within the villi, so that it is impossible to examine the cells. With every possible method it was difficult to detect the cells, and, therefore, I am inclined to believe that the stroma cells suddenly cease to exist at this stage of pregnancy. 504 GENCHO FUJIMURA 4. The decidual cells (figs. 39 to 71) It is a generally well-known fact that decidual cells are divided into very many kinds according to shape, size, staining prop- erties, and structure; however, it has not as yet been definitely decided whether or not these kinds of cells should be reckoned as one and the same sort. Marschand (’04) first divided the decidual cells into two types according to the difference in size. Subsequently, Fraenkel (14), too, who studied them chiefly from the staining standpoint, set up a similar theory, and tried to divide them into his so-called acid cells (Eckersche Form) and the neutral cells (of large type); however, he himself and others had no doubt that theré was not only no distinct divis- ion between these two kinds of cells, but rather there was exist- ent an intermediate type of cells between them. Figures 39 to 69 illustrate the various kinds of decidual cells placed in order. Of these, cells such as in figure 39 are the smallest and are spherical with a nucleus of a corresponding form. The pro- toplasm is, as compared with the interstitial cells at the time ‘of non-pregnancy, remarkably large in quantity, and contains in it a large number of rod-shaped plastosomes. Figures 40 and 41 are a little bit larger than the former, and the one is spherical while the other is oval, both having a nucleus of nearly corresponding shapes. The protoplasm becomes still more abundant and the plastosomes are somewhat longer rods. It is worth our noting that both cells have a kind of boraer membrane already on the superficial layer of the cell bodies. Figure 42 demonstrates the first appearance of a few lipoid granules of various sizes within the cell bodies. Figure 43 illustrates a pear-shaped cell, which holds in the body a some- what large quantity of granules and a few vacuoles. A few plastosomes are to be found, and they make their appearance only on one side of the cell. Figure 44 resembles the former in shape somewhat, and contains in the cell body remarkably large lipoid granules, which, on a close examination, are found to have amore or less distinctly clear halo around each of them, as though they constituted the contents of vacuoles. The plas- SECRETORY FUNCTIONS IN HUMAN PLACENTA 505 tosomes are mostly rod-shaped, and they crowd together on one side’ of the nucleus, while on the other they appear in are- markably long, granular string (Fadenkoérner). In figures 45 to 48 the cells have exceedingly varied shapes, and the nucleus trends toward one side of the cell close to its superficial layer. The cell bodies are filled with numerous irregular-sized vacuoles, which present a more or less distinct foamy structure. The plastosomes are mostly short and rod-shaped and they are to be found in the proplotasm between the vacuoles. Some of them are arranged in along row along the walls of the vacuoles, while others are found in groups in a certain section. The lipoid gran- ules are, in general, small in number, and their sizes are irregu- lar, some of them finding themselves distinctly at the center of the vacuoles (e.g., fig. 46). In the various cells described above it will be observed that the nuclei are generally dark-colored, with indistinct nuclear network in most of the cases, though the nucleoli contained are conspicuous enough. Figure 49 is extremely different in its appearance from these cells. The cell body presents a vacuolar formation in high degree and the protoplasm proper can be demonstrated in a small amount only along the surface of the nucleus, at which place only a few rod-shaped plastosomes are found. The border membrane on the surface of the cell is very distinct and the nucleus is different from that in the various cells described previously in that it is clear and presents a large foamy body. The nuclear network is somewhat distinct, and, besides, there are conspicuous nucleoli. It must be generally stated that this kind of cell appears very seldom. In figure 50 the cell body presents an equally high vacuolar formation as in the former, and, in addition to that, the vacuolar walls entirely disappear in some places and so allow the inner spaces of the vacuoles to com- municate with one another, thus clearly indicating the evidence of the vacuoles having been agglutinated. A few more plas- tosomes than in the former are found in groups in these cells, and, besides, there is a small quantity of lipoid granules, mostly within the vacuoles. What is especially peculiar about this cell is that at the center of the cell body there appears a black- 506 GENCHO FUJIMURA colored homogeneous star-shaped lump, from the surface of which are shot forth a number of processes, which run over directly to the protoplasm between the vacuoles. The proper nucleus is hardly detected. In figure 51 the cell is longish, and the overabundance of plastosomes which are distributed densely almost all over the cell body is the feature of this cell. Between there, is a somewhat large number of vacuoles of various sizes, and no lipoid granules at all are to be found. In figures 52 to 54 the various cells illustrated are gradually larger than those described above, the cell bodies are filled with a large number of vacuoles and granules. The latter are ex- tremely irregular in size and density and are sometimes found as contents of the former. The plastosomes are mostly short rods, and especially in figure 52 they are somewhat abundant, and some are found massed along one side of the nucleus. In figures 53 and 54 they are comparatively fewer and lie scattered between the vacuoles. The thin layer on the surface of the cell is thicker and more distinct in this kind of cell, to such an extent that it almost reminds us of the ordinary cell membrane. In figure 53, as in figure 50, we notice a black-colored round-shaped > lump at the location of the nucleus; however, in this case, the surface of the lump is smooth and has no process. Comparatively few cells have such a black-colored lump, and, as in these cells it is always difficult to tell the whereabouts of a nucleus of normal condition, I am quite at a loss to know whether or not the dark lump described above should be deemed a modi- fication of the nucleus or regarded as that part of the protoplasm which is just adjacent to the nucleus which has, by reason of its staining properties, obscured the nucleus. This question, along with the stained lumps in the syncytium layer as illustrated in figure 10, constitutes a puzzle, and I have mentioned it here for the sake of future investigations. However, in view of the fact that numerous plastosomes, which are the important elements of a living cell, are always demonstrated in the cells concerned, while at the same time they present no noticeable regressive phenomena. I am rather inclined to believe that it is possible to attribute a certain functional significance to these unknown lumps. SECRETORY FUNCTIONS IN HUMAN PLACENTA 507 The cells illustrated in figures 55 to 57 are already exceedingly large, and they are, at a glance, recognized as Marschand’s so-called large-type of decidual cells. On the surface there is arather thick layer, which may be divided into two of which the inner one is thick and the outer thin, and they exactly remind us of the definite cell membrane. The cell bodies consist of plastosomes, lipoid granules, and structureless stroma. The plastosomes are mostly long rods or threads, some being more or less peculiarly curved and the quantity of plastosomes is variable. The lipoids differ also in point of size, density, quantity, ete. What is worth our noting is that there is no vacuolar formation to be found in these cells. The nuclei are large, clear, and foamy. The nuclear network is especially conspicuous in figures 56 and 57. Figures 58 to 61 show the definite form commonly belonging to the so-called decidual cells of large type. In these cells the cell membrane is remarkably thick, and the distinction between the inner and outer layers is always clear. What deserves our special attention at this juncture is that the outer layer contains, in most cases, a kind of granular body which is somewhat large and yet irregular in size and stained a dark color. The cell bodies consist of a large quantity of plastosomes and homogeneous stroma. ‘The plastosomes are mostly long rods, and they sometimes appear extremely elongated in the shape of threads (fig. 61). They are distributed equally all over the cell bodies, although they sometimes tend to appear more or less in groups. The plastosomes in these large cells have, in general, very slow staining properties, and, therefore, they are a very difficult subject to be dealt with from the tech- nical point of view. The nuclei are foamy and dark-colored, and the nuclear networks are indistinct. The cells illustrated in figures 62 to 69 differ from those de- scribed above, and they all lack the plastosomes. Even in the most excellent stained preparations these cells appear within the decidual tissues in small numbers, for the most part more or less in groups, scattered like islets, so that it is, as a matter of course, incomprehensible that here alone the plastosomes are hard to be demonstrated. However, as, in consideration of 508 GENCHO FUJIMURA their structure and shape, it is premature yet to decide posi- tively that there is a tendency for a retrogressive change among all these cells, I will here suggest provisionally that such a phenomenon is a sign of a certain functional period in the cells concerned. Now, at first, in figures 62 and 63 the cell bodies are comparatively dark, and within they contain a large quantity of vacuoles and lipoid granules of various colors; some of the vacuoles distinctly have a lipoid granule as their nucleus, while others, being placed in rows close to the cell membrane, present a peculiar image. The nuclei are clear and have a somewhat distinct nuclear network. In figure ‘64 the cell body is filled with numberless vacuoles of nearly the same size, and on one side of the nucleus accumulates a large quantity of proto- plasm, and, besides, there are a few deep yellowish-brown lipoid granules. Figure 65 is of nearly the same type as the former, but the lipoid granules contained are by far greater in quantity than the vacuoles. The nucleus is as clear as the for- mer, with conspicuous nuclear network. Figures 66 and 67 illustrate the cells whose bodies are filled with an exceedingly large quantity of vacuoles of various sizes, in consequence of which the protoplasm becomes comparatively scarce and faint and is mostly noticeable only around the nucleus. And, besides, there are some vacuoles which hold dark or deep yellowish-brown lipoid granules; also vacuoles and lipoids, whose size, quantity, and distribution are as varied as will be seen illustrated in the respective figures. The nuclei are generally dark and show extremely delicate network formations. Figures 42 to 48 and 51 may be compared, from their sizes and histological point of view, with that class of cells which is termed by many authors as ‘decidual cells of small type’ (or possibly Ecker’s type), while, on the contrary, figures 55 to 61 and 64 to 69 may be nothing but the so called ‘large-type’ or ordi- nary decidual cells (neutral cells). Further, the various cells illustrated by figures 49, 50, 52, 53, 54, 62, and 63, judged from their size and internal structure, should be deemed an in- termediate type which may intervene between the former two, since it is a very difficult task to determine to which one it SECRETORY FUNCTIONS IN HUMAN PLACENTA 509 should belong. Now, the result of my careful examination of the appearance and distribution of these cells as compared with the time of pregnancy has been that, in the material which was taken a fortnight after conception, this being the earliest I have on hand, the cellular ingredients of the decidua chiefly consist of the small-type cells described above. Ina little more advanced stage (i.e., about 17 or 18 days after pregnancy) the decidua shows also the appearance of a large quantity of the ‘interme- diate-type cells,’ while in the few days following (i.e., about 22 or 23 days after pregnancy) with the decidua already of definite growth, all kinds of cells, especially the large-type ones, can be detected in comparatively large quantities. One month after pregnancy the large and intermediate type cells appear as the principal ingredients of the decidua, while on the contrary the small-type cells retrograde gradually and are met with only in the interstitial tissue. Such condition is maintained until the end of the last month of pregnancy. In short, I conclude from the histological and histogenetical point of view, that the various kinds of cells described above all belong to one class, and con- sequently it follows that the division of decidual cells into large and small types, is faulty. In other words, the term ‘small-type decidual cells’ applies only to the cells which are still at the early stage of growth, while the ‘large-type cells’ are those in the last stage of growth. The time taken in such growth is, according to my observations at the earliest stage of pregnancy at least, comparatively small, thus the small-type cells being perfected into the ‘definite large-type cells’ in so short a time. The appearance of lipoid granules and vacuoles in the decid- ual cells is most remarkable from the end of first month to the second month of pregnancy, and they gradually decrease in the months following, though very infrequently they can be demon- strated even at the end of pregnancy. The plastosomes could no longer be demonstrated in any of the decidual cells after the seventh month of pregnancy. And, in the interstitium of the decidua, such extremely strange- looking productions as are illustrated by figure 70 may some- times be detected. They are stained dark and consist mostly 510 GENCHO FUJIMURA of a great number of granular bodies which are extremely varied in size. There are granular threads, which are the result of the granules being linked together, more or less curved large rod-shaped bodies of different lengths, and sometimes threads which are very long and often curved like screws, be- sides a large number of material ingredients, which, being mixed up among them, have shapes similar to them and yet are unstained and noticeable only as a shadow. ‘The former, which are susceptible to staining, are dyed deep red by Altmann’s method and deep black by iron-alum-haematoxylin. At a glance they look like plastosomes in their shape and staining, and yet in general excel the latter in size considerably. If aggregated, they may be quite easily detected under medium magnifica- tion. Moreover, the staining properties of the ingredients are much stronger than the plastosomes, and they bear a rather close resemblance to lipoids in their density and appearance. The product of this kind do not only possess exactly the same properties in shape and stain as the granular bodies in the cell membrane of large-type decidual cells to which I alluded above, but also indicate that there is often the closest relation between the two; i.e., within the cell membrane of the cells concerned there are, besides the granular bodies above referred to, often short granular threads or large rod-shaped bodies, both of which are similar to the substances in the interstitium described above. One end of the rod-shaped bodies sometimes enters deep into the cell membrane and swells into a club-like shape, while a greater part of the other end juts out into the interstitium, thus giving itself the appearance of passing into the interstitial product. I am not able to explain the original nature and functional significance of this kind of product, and yet, according to the afore-mentioned observations, I have no doubt whatever that in the formation of the product the large cell, and especially its cell membrane, plays an important part and, since such interstitial substances are demonstrated in large quantities in the adjacent blood vessels as are illustrated by figure 71, it may be inferred that they are ultimately absorbed in the vascular organs. The products of this kind are for the first time noticed at about the SECRETORY FUNCTIONS IN HUMAN PLACENTA ell second month of pregnancy, appear most conspicuously from the end of the second month to the third, decreasing gradually after that, and, though the decrease is considerable after the fifth month, they may yet be demonstrated until about the sixth month. Besides the above, there are detected in certain parts of the interstitium numberless filar productions which have various length, and are sometimes long like threads or fibers, of which the smaller and shorter ones sometimes bear a close resemblance to the plastosomes, while the others usually gather in great num- ber and often make a mass of fibrous bundles. This kind of product, so far as staining is concerned, is entirely similar to the interstitial productions described above, and yet it differs from the latter in that its shape is not so varied, its thickness is nearly always even, and, besides there is no special relation which is noticeable as existing between the products and neighboring cells. 5. The epithelium of the uterine gland (figs. 72 to 83) As is generally well known, the uterine gland undergoes a cer- tain morphological change at the early stage of pregnancy, and in my previous treatise I have drawn attention especially to the fact that the glandular cells also show a morphological change at such a time. Here I will observe and describe more thoroughly the changes of the cells concerned. Figure 72 shows a glandular cell which is commonly noticed at the early stage of pregnancy and which is already remarkably thickened and somewhat round. The shape of the nucleus for the most part corresponds to that of the cell. On top of the cell there are traces of cilia. The cell body contains many slender and rod-shaped plastosomes, which latter chiefly gather closely against and surround the nucleus. In figures 73 and 74 the cell grows larger, and the plastosomes are demonstrated only in the upper half, while in the lower half which contains the nucleus, none of them are found. At this section of the cell there are plenty of lipoid granules, which are of about equal size and are stained a bright yellowish-brown color, and on top of both cells there are still the traces of the cilia. In figure Bly GENCHO FUJIMURA 75 the cell is remarkably elongated, and within are contained a great number of lipoid granules. The nucleus is oval with a nuclear network distinctly observed; from this period on no more traces of cilia are to be found. The various cells illus- trated by figures 76 to 78 gradually increase in their size and, since their swelling is remarkable, especially in the upper half, it is usual that this part of the cell naturally juts far out into the lumen, though the lower half, being comparatively narrow, is closely united with the basement membrane. Within the cell body are contained a great number of both lipoid granules and vacuoles, of which the vacuoles in figure 76 are as yet small and few and they chiefly occupy positions in the upper half of the cell body, whereas in figures 77 and 78 the vacuoles enlarge tremendously and fully occupy the upper half, in conse- quence of which cell bodies have the appearance of a honey comb in a high degree, and the protoplasm remains only as a thin wall which separates the vacuoles. The lipoid granules in the latter two kinds of cells chiefly crowd together at the base of the cell bodies, and only a small portion of them are left behind as contents of the vacuoles. The afore-mentioned three cells each present conspicuous nuclear network and nucleoli. And in the various cells in figures 75 to 78 no plastosomes are to be detected. Figure 79 illustrates a large oval-shaped cell, which has a similar-shaped large alveolar nucleus. The cell body, because of the vacuolar formations of various sizes, presents the image of a conspicuous honey comb, while the plastosomes lie scattered somewhat plentifully in the interstice between the vacuoles. Though there are some extremely small vacuoles in jet black, no ordinary lipoid granules in coarse grains are to be found. The contents of the nucleus are nearly homogeneous, and the characteristic nuclear network is not found, but within the nucleus there are two nucleoli. Figure 80 is an extremely irregular-shaped cell, with the nucleus of a corresponding shape. The structure of the cell body is nearly the same as that in the preceding figure, and the vacuoles are somewhat plentiful, but the plastosomes are very few, and, besides, there SECRETORY FUNCTIONS IN HUMAN PLACENTA 513 are but few yellowish lipoid granules. The nucleus has a con- spicuous network. As the change of the glandular cells reaches a high degree, the cells leave the basement membrane in large numbers and are isolated in the glandular lumen. Such cells I call tempo- rarily the ‘desquamated cells’ in contrast to the cells which continue to settle on the basement membrane or remain fixed on the walls. Figures 81 to 83 illustrate my so-called desquamated cells, in the inside of whose cell bodies there are many vacuoles and a small number of extremely small lipoid granules which are stained in black. At a glance they look like the ‘wall-fixed’ cells, and yet on a careful comparison we find that there is a more or less remarkable difference between them. That is to say, in the desquamated cells the cell bodies are in general somewhat turbid, and also the vacuoles look somewhat withered and decayed. No plastosomes are to be found, and, besides, the change of the nucleus is remarkable, as will be seen in figure 81. Here it is converted into a dark- colored and homogeneous lump provided with a very irregular contour, having within a few small nucleole-like bodies. In figure 82, as before, the nucleus is simply a homogeneous and dark-colored lump; in figure 83 it has swollen somewhat re- markably, and the nuclear network is extraordinarily conspic- uous. In short, the desquamated cells have undoubtedly began a retrogressive degeneration already, and from the existence of many broken pieces of cells which are always found in the gland- ular lumen it can be simultaneously demonstrated that the des- quamated cells are doomed to break up and perish at this sec- tion sooner or later. The various changes undergone by the glandular cells described above are seen in a remarkable degree al- ready on about the seventeenth or eighteenth day of pregnancy in the decidua serotina, while in the decidua vera it is a little bit later and the changes are noticed to about the same degree as the former only toward the end of the first month of pregnancy. Both reach the maximal changes at about the end of the first month of of pregnancy, and on the days following they gradually pass into the so-called desquamated cells and break up and perish as such. 514 GENCHO FUJIMURA The series of changes described above have a more or less difference of time between the decidua serotina and the decidua vera, viz., in the former they may be followed up vigorously up to the end of the second month of pregnancy, though in the third month they suddenly decrease, whereas in the latter such changes may be demonstrated even one month later. THE HISTOLOGICAL STRUCTURES AND THEIR FUNCTIONAL SIGNIFICANCE (SOME REFLECTIONS ON LITERATURE) In the preceding chapter I have given a somewhat minute account of the delicate histological structures of the various important cell groups which exist in the placenta and the de- cidua vera at different periods of pregnancy. Now, on a peru- sal of the observations given therein, it will be found that as components which are common to the various cell groups there are 1) plastosomes, 2) lipoid granules, and 3) vacuoles. Asa mat- ter of course the degree of the appearance of the three kinds of components and their distribution in which they are present vary infinitely as the kind of cells differs or according to each individual cell. However, that which exercises the most im- portant influence over the shape and formation of the cell is chiefly the lipoid granules and vacuoles, of which the latter often appear in a very great number and occupy the whole body of the cell, thus giving the cell a highly foamy appearance or a honey comb structure. The cell which has fallen into such high- ly vacuolar formations makes one feel, at a glance, that it has presented a phenomenon of collapse due to the regression of the cell concerned, as some observers are apt to conclude quite hastily. However, as, on a careful examination of it, it is found that, in spite of the high degree of changes shown by the cell, there are demonstrated for the most part within the cell body the plastosomes which are deemed an important active element in the functions, and also in consideration of the normal struc- ture of the nucleus and of the fully stained conditions of the cell body, there is no doubt as to the cells being alive. And even if it should be conceded for a while that the cell having the fully stained vacuoles as described above is the indication of a kind SECRETORY FUNCTIONS IN HUMAN PLACENTA 515 of regression, who could explain the reason why this kind of ‘regressive’ cell actually appears in so high a degree within the tissues of the placenta and the decidua at a certain period of pregnancy, and especially in the first half of pregnancy when the tissues should grow in a most vigorous manner? There- fore, I am confident that not only is it wrong to deem such vacuolar formations a death phenomenon of the cell, but also they should rather be taken for a quite significant phenomenon which shows a certain function of the cell concerned. And, as regards the actual existence of such a function, I am in- clined to assert from their closest resemblance to the structures of many other glandular cells, as a result of my histological ob- servations, that a secretory function is existent in these kinds of cells. However, as the problem is of such a provisional character I will, for the present, reject all hasty assertions, but instead will consult literature widely and make general reference to the previous interpretations of many authors on the struc- tures and secretory processes of the various glandular cells in the organs in which secretory functions are definitely known, so that the most deliberate considerations can be given my histological observations and their functional significance, which, being compared and discussed under impartial criticism, it is hoped, will help toward making the original nature of the func- tions clear. Since the relations between glandular histology and secre- tory functions were early dwelt upon by R. Heidenhain (’68, "75, ’80) with his penetrating eyes, the subject has aroused the interest of many excellent physiologists and histologists, and the studies of the subject have since followed so quickly one after the other, that it would be difficult to enumerate them here. The following are the principal authors who have studied the subject, and the materials chosen by them for investigations were chiefly pancreas, salivary glands, gastric glands, lacrimal glands, skin glands, and pelvic glands, all of which are usually known as representative glands in many kinds of animals classed above the amphibians: Schultze, Fr. E., ’64 and ’67; Langhans, ’69; Pfliiger, ’71; Schwalbe, ’71; Ebner, ’73; Nussbaum, ’78 to ’82; 516 GENCHO FUJIMURA Ebstein and Griitzner, ’°74; Lavdowsky, ’76; Langley, ’79 to ’89; Mathews, ’80; Klein, ’82; Biedermann, ’82 and ’86; Flemming, ’82; Kiithne and Lea, ’82; Nicoglu, 93; Altmann, ’94; Galeotti, ’95; Krause, 795 and ’97; Miiller, 96 and ’98; Solger, ’96; Zimmer- mann, 798; Held, 99; Maximow, ’01; Noll, ’01; Fleischer, ’04; Heidenhain, ’07; Babkin, Rubaschkin, and Ssawitsch, ’09. I will not go to the trouble of giving a detailed account of each of the results of these research works, but will confine myself to summarising the main points of their investigations respecting the structure and secretory phenomena of glandular cells, which are most essential to my studies, and refer to the original works for details. In the first place, the structure of the glandular cells having a duct, or externally secreting glands, greatly differs, as is gener- ally well known, according as they are serous or mucous. The serous cells have an exceedingly large number of granular bodies, and consequently their characteristic is that they are generally dark. These granular bodies are generally known as Cl. Bernard’s secreting granules. The relation which the latter has to secreting functions has been generally recognized by many an interesting research work since that of R. Heidenhain, and it will be noted that Heidenhain, having observed a kind of slender thread-like structure which exists close to the base- ment membrane, of the glandular cell of a dog’s pancreas, for the first time drew general attention to a peculiar sort of organic structure which is existent in the glandular cell. This kind of thread-like structure was demonstrated also in the salivary ducts and convoluted uriniferous tubules in after years, and it cannot be anything but M. Heidenhain’s so-called ‘Basal- filamente’ or our plastosomes. Below I wish to give a general outline of the changes appear- ing in these organic tissue elements of the glandular cells which will follow the secretory functions. Now, at first, it is universally agreed by fa author that secretory granules increase or decrease according to secretory functions. According to the result of the close examination with respect to such correlation, conducted chiefly while in a fresh SECRETORY FUNCTIONS IN HUMAN PLACENTA aL7 condition, of the pancreas of the rabbit, the oesophagus glands of the frog, and the gastric glands of the water lizard, by Langley (79 and ’89), whose exposition of the correlation is best authen- ticated, it will be noted that during suspense of their func- tions glandular cells are generally glutted with secretory granules, whereas as secretion begins the granules gradually decrease and disappear, in consequence of which in each cell will appear a sharp demarcation between the homogeneous wide outer layer and the remaining granular inner layer. This observation of Langley has been repeatedly proved by many other authors in many other glands, and its validity except ina small number of cases, has been recognized by nearly everybody. For convenience, I defer to a later section a minute explanation of the functional significance of the thread-like production which is found in the basal part of glandular cells. Even in mucous cells it has been universally acknowledged by many authors since F. E. Schultze (64) that there are gran- ular bodies in large numbers at a fixed period of secretion, and at the time when secretion is very high these granules gradually disappear asin theserous cells (Langley, ’79; Biedermann,’82). It seems therefore that in the forms of secretion formation mucous cells agree, for the most part with the serous cells. However, the reason why both differ widely in their structure is that in the latter the secretion is, for the most part, speedily drained into the glandular lumen at the time of secretory function, whereas in the mucous cells it is formed within the cell bodies, besides being stored ‘up there for a certain period, thus giving the cells a peculiar structure and a characteristic clearness. In short, in the aforementioned two kinds of glandular cells, the large quantity of secretory granules always to be seen in both during suspense of secreting functions disappears by degrees as the functioning begins. In view of this fact, we have no doubt whatever that, at the time of secretion, the granules always play an indispensable part as the mother-ground for the secretions and, since secretory granules are generally deemed a solid production, according to the investigations of many authors JOURNAL OF MORPHOLOGY, VOL. 35, NO. 3 518 GENCHO FUJIMURA while the secretions are mostly fluid, it follows that it would be no great error to take it that, viewed simply from the histological standpoint, the so-called secretion, after all, means, the liquefac- tion of secretory granules. However, the modes of liquefaction are, according to my view, so varied that they should by no means be dealt with in one and the same manner, but rather there are, roughly speaking, several forms, such as are given below, according to the kinds of glandular cells, or according as the cause which prompts secretion differs. First form. This is observed in certain mucous cells. The secretion is brought into being simply by the melting down and growth of secretory granules which have developed in a fixed degree, and, different from many other glandular cells in which the secretory granules are mostly preproducts of secretion, the granules here in this case show the same reaction as secre- tion (that is, mucin) from the beginning of their appearance and find themselves already identical with the secretion as early as they appear. This fact was demonstrated by M. Heiden- - hain in the goblet cells of the intestine of the salamander, and it is, in general, difficult to tell definitely the time of liquefac- tion in what is covered by this form. Second form. This is observed in many mucous ne The granules, while in the first stage, appear as a certain preproduct (mucigen), and undergo a chemical change simultaneously as they have developed to a certain degree, and are changed into the ordinary secretion (i.e., mucin). Mention has been made to this effect by Biedermann (the mucous cells of the frog); M. Heidenhain and Nicoglu, 93 and ’98 (the skin glands of a salamander); Altmann, ’94 (the submaxillary glands of a cat); Maximow, ’01 (Glandula retrolingualis of a hedgehog). Third form. This is often seen in ordinary serous cells. As the functions begin, the secretory granules begin to liquefy gradually on the periphery and, although it appears that the granules are imbedded for a certain period in the secretion al- ready formed, they finally change thoroughly into secretion, and then appear as simple vacuoles within the cell bodies. The clear halos around the granules referred to by Langley (734), SECRETORY FUNCTIONS IN HUMAN PLACENTA 519 Corlier (’96) and Maximow (’01) were it seems, deemed by these authors an essential ingredient in the formation of gland- ular cells; however, M. Heidenhain with Nicolas (’92) attrib- uted one part of them rather to artificial production, while the other was brought into being in the form it appears in, possibly be- cause the granules were reduced in size in connection with secreting functions. And in recent times the researches re- garding pancreas cells, conducted by Babkin, Rubaschkin, and Ssawitsch (’09), have proved the above mentioned views of Heidenhain with more force and accuracy; that is, in the opinions of these authors, clear halos around the granules are, after all, nothing else but secreted matter which appears as a liquefied produét of the granular substance. And that such an image should be a sign of an important period in secretion formation within the cell bodies will be clear in the statement made by the three authors mentioned above on the changes which zymogene granules undergo at the period of a pancreas secretion: ‘‘ Wir erhalten den Eindruck, als ob das (Zymogene) Kornchen sich allmiahlich auflést, in dem es sich immer mehr und mehr verkleinert und in ein kleines verwandelt”’ (p. 92). Fourth form. According to the views of Babkin, Rubaschin, and Ssawitsch above referred to, this presents an appearance some- what like a modification of the third form. That is to say, many secretory granules together with the intergranular sub- stance concerned form themselves into somewhat large masses, which latter slowly begin to liquefy from the periphery, and are then gradually changed entirely into secretion in rather large drops, to be drained out into the glandular lumen. Com- pared with the third form, in which each individual granule becomes liquefied separately, this type differs from the former in that a number of granules coagulated together in lumps are transformed gradually into drops of secretion. This form is distinctly demonstrated at a time when the function of pan- creas cells is very greatly increased, especially by means of stimulating the vagus nerves or of soap infusion in the duo- denum, it being characteristic of the secretion at such a time that the latter is very dense, and is rich in albumen and fer- 520 GENCHO FUJIMURA ments. And the three authors mentioned above are of the opinion that the several bodies which are revealed by this secreting form may be compared with the small bodies which are often known as ‘Nebenkerne,’ ‘parasome,’ ‘corpusculus paranuclaires,’ and ‘noyau accessoir’ in the pancreas cells of the lower vertebrates. Fifth form. This points to a case in which the substance of secretory granules changes its quality at a certain period, be- comes soluble, and is only dissolved in the watery part of the secretion alone, as was instanced by Babkin, Rubaschkin, and Ssawitsch, in the experiment of the pancreas secretion after pouring acid in the duodenum. As a proof thereof, the secretion always shows in this case the same coloring réaction as do the zymogene granules. The three forms (the third to the fifth) described above are together met with in the pancreas cells, and it deserves special attention that according as the cause differs for the rise of se- cretion, even in one and the same glandular organ, there is a difference in the forms of secretion, and consequently in the nature of the secretion produced. Sixth form. Certain secretory granules sometimes have two extreme developments. As an instance, M. Heidenhain observed the skin glands of a salamander and he found that, as the development of the granule reaches a certain degree, one part of it is liquefied and passes into a viscous secretion, while the other goes on growing in size, and is perfected into the char- acteristic poison grains. Seventh form. A certain organic structure is presented in the secreting granules; at the beginning it is not different from ordinary cases, and yet as a certain development is accomplished, every granule is divided into a crescent-shaped section (Kapuze) which is stainable, and into an unstainable section (Triiger) within the crescent-shaped section, thus forming the ‘Halbmondképerchen’ so termed by M. Heidenhain. This latter as it eTOWS in size is considerably enlarged, especially in the Trager, and consequently the Kapuze is flattened gradually and is only stuck like a plate on its one side. Then the Traiger SECRETORY FUNCTIONS IN HUMAN PLACENTA SpA is at last dissolved altogether and is changed into secretion, while at the same time the Kapuze is also condensed, and forms itself into secondary granules, later on to be drained out along with the secretion, thus bringing about an entirely granule free condition of the cells. This form of secretion was first detected and examined closely by M. Heidenhain in the pelvic gland of a salamander and in the lacrimal gland. Subsequently, Nicolas (92), in a man’s lacrimal gland, Held (’99), in a rabbit’s sub- maxillary gland, and Fleischer (’04), in a cow’s lacrimal gland, observed and proved the appearance of the crescent-shaped small bodies described above. The above does not cover all the forms of granular liquefi- cation at the time of glandular secretion, and yet it will be quite clear how varied the modes of liquefication are. Now, I will discuss a little further the origin, and therefore the manner in which supply is made of secretory granules, this problem being the second in importance with respect to glandular secretion. With regard to the origin of secretory granules, the problem itself is a very difficult one indeed, but it should be noted that, along with the sudden increase of studies on plastosomes in re- cent years, many authors have attempted a solution of this troublesome question. Below I give a general account of the result of these researches. Generally speaking, every author equally agrees in the argument that every glandular cell always has within the cell body plasto- somes as constant ingredients and, though the quantity and arrangement of the latter arecertainly varied, it is the usual condi- tion that while the serous cells are in a state of rest the numerous plastosomes are chiefly arranged in rows along the long axis of the cells. These plastosomes, however, generally decrease in quantity considerably during the period of secretion, especially when the cells are filled, and they are then found largely, close to the basement membrane of the cell, while near the lumen, they are either entirely absent or, if they are found, they are between the granules in a very small number. And in mucous cells, in contrast to serous cells, the plastosomes are found largely Gay GENCHO FUJIMURA near the nucleus, though some lie scattered in the protoplasm between the secretory granules, all being arranged in an irregular order. With regard to the functional significance of the plasto- somes contained in these glandular cells, there are various theories, such as the theory of secreting mechanism by Benda (03, p. 780), that of water-secreting apparatus by M. Heiden- hain (’07), and that of prop system by Bruntz (’08); however, these are either the historic ones, having been refuted experimentally by Meves and Regaud (’08) already, or mere assumptions. What is believed by a great majority of authors at present is that plastosomes are very closely related to the formation of secretory granules. And the first man who published his views with regard to this was Altmann (’94), who thought that his so-called ‘vegetative Faiden’—a greater part of which agrees with our plastosomes of to-day—being split up in small pieces are formed into granular bodies in large numbers and make the beginning of secretory granules. Quite recently the researches of _Laguesse (99 a, b, ’05, 711), Regaud (’09), Regaud and Mawas (09), Hoven (’10, ’11, ’12), Champy (11), Schultze (11), have followed one after the other, and, though their observations may differ a little from one another, either in trifling points or in the form of description, they none the less fall entirely into line with Altmann’s view that the formation of secretory granules has its origin in plastosomes. And, moreover, the fact that, as is generally acknowledged, the number of plastosomes in gland- ular cells always increases or decreases according to the changes of secreting functions is nothing if not forcibly proving the valid- ity of such a theory. However, this very theory is not without objections. M. Heidenhain, Mislawsky (11), and Levi (12) are those who are opposed to it. Especially M. Heidenhain, taking his stand © on the ‘Protomeren-Theorie’ which he set forth, states that secretory granules should have their origin in extremely faint and extraordinarily small bodies in the protoplasm which are beyond our sight; these bodies growing up and increasing should gradually develop into the smallest granules—his SECRETORY FUNCTIONS IN HUMAN PLACENTA 523 ‘Primirgranula’—which are seen microscopically to gradually bring about the growth of definite secretory granules. Thus it seems he denies the histogenetical correlation between plasto- somes and secretory granules. His theory, though most dex- terously proposed on the most profound proofs, is after all a mere hypothesis. Besides, in consideration of the following quotation, it will be quite clear how he holds the views that his so-called ‘Primirgranula’ have a close relation with the genuine cytomicrosomes in their formation, and that the latter do correspond to a section of his ‘Basalfilamente,’ and, further, that the filaments are identical with Flemming’s Mitom: Ich bin daher mit Solger der Meinung, dass auch in den serésen Driisenzellen die Basalfilamente Teil eines Fadensystems sind, welches nach der Bezeichnungsweise Flemming’s dem sogennanten Cytomitom zugehoéren wiirde. . . . Nicht ganz ausschliessen darf man jedoch zur Zeit die Moglichkeit, dass die Driisengranula eventuell von den genuinen Plasmamikrosomen sich ableiten, welche nach unseren allge- meinsten Erfahrungen immer innerhalb der ‘Protoplasmafilamente liegen, also verdichtete Teile feiner Fidchen sind. . . . . Da nun die Kérperchen beiderlei. Art (Primérgranulis und genuine Mikrosomen) morphologisch schwierig unterscheidbar sind, so kann eine verwand- schaftliche Beziehung zwischen beiden zur Zeit wenigstens nicht abge- lehnt werden, besonders da wir tiber die positive Bedeutung der genuinen Plasmamikrosomen noch sehr im Unklaren sind (p. 390). And since it has already been clearly proved by the researches of Meves (07) that Flemming’s Cytomitom agrees with our plastosomes, it would appear that M. Heidenhain does not only maintain the view that his ‘Basalfilamente’ exists as a water- secreting apparatus as mentioned above, but at the same time also supports, instead of disproving, the argument which re- gards basal filaments as being the matrix for the formation of secretory granules, as Altmann and others do. As regards Mislawsky’s view, it will be noted that he, in line with Levi, takes the ground that as between plastosomes and secretory granules there were no conspicuous conditions to be detected in support of the existence of a formative relation between the two. However, anent Mislawsky, it will be noted that the invalid- ity of such a view has been exhaustively pointed out by a Belgian cytologist, Duesberg, in comments noted for their profundity 524. GENCHO FUJIMURA and thoroughness. The following phrase is there employed by him: ‘‘Ein negatives Resultat beweist nichts gegen eine Reihe positiver Resultat und man kann nur schliessen, dass Mislawsky die Bilder nicht beobachtet hat, die seine Gegner unter Augen hatten”’ (p. 785). This should incidentally prove to be a pertinent comment on the theory of the excellent Italian author, Levi. If we summarize the observations of the various authors respecting the structures and functions of the glandular cells described above, we may enumerate as constituents of the glanduar cells, 1) plastosomes, 2) secretory granules, and, 3) secretions (vacuoles), besides the protoplasmic stroma proper, and it would be superfluous to state that of these constituents secretory granules have a directly important relation to the glandular secreting functions. The secretions are nothing else than a liquefication or modified product of the latter and there is no alternative but to assume that the secretory granules, once lost at secretion, take their matrix mostly from the plastosomes, which latter,’ being split up and separated in small pieces, gradually meet the deficit so caused. Thus, the real state of glandular secreting functions is already a thing which can be largely followed up and made clear by means of minute histological investigations to-day. Also the histological studies of the ductless glands—internally secreting cells—have become very active of late years, and, especially among those which have been carried on by the methods of plastosomic study, we may enumerate the thyreoid and para- thyreoid glands, suprarenal capsules (chiefly its cortex), Lang- hans’ islets of the pancreas, and the ovary. Of these organs, the ovary has been used more often than the rest as the object of study, and consequently, comparatively speaking, much is known about it. I will therefore give a general summary of the observations of the various investigators respecting this organ, consider its structure and the relations of its internal secretory functions, and then glance at the structures of the other organs, thus contributing toward a histological account of internal secretion in general. SECRETORY FUNCTIONS IN HUMAN PLACENTA yas. That the corpus luteum is a kind of internally secreting organ (from the histological point of view) was advocated by Prenan- as early as 1898 and by Born in 1900. Subsequently, the argu- ment has been advanced with more certainty by a close histological study conducted by Regaud and Policard (01), Fr. Cohn (’03), Mulon (09), Athias (11), Van der Stricht (712), Tsukaguchi (’12, 713), and Levi (’13). And the main ground of this argument lies in nothing but that the luteal cells con- tain as formative ingredients of the cell bodies plastosomes and lipoid granules, and often demonstrate a very large quantity of vacuoles which may be deemed their secreted matter, and also that the close correlation between these ingredients func- tionally is plainly apparent in the same manner as it is seen in external secretory cells. Of the various ingredients mentioned above, lipoid granules certainly have the most important signifi- cance in internal secretion; however, since the latter is not only characteristic of luteal cells, but is also widespread among the other kinds of internal secreting cells, Tsukaguchi has already argued that it would be proper, in view of its functions, rather to compare them with the secreting granules of ordinary glandular cells. In the next place, Cohn first regarded the vacuoles in a rab- bit as a secreted matter of the luteal cell, and he saw them simply asa dissolved product of lipoids; subsequently Tsukaguchi also studied the same animal, and likewise deemed them a modified product of lipoids. However, vacuoles differ exceedingly in the degree of their development in all kinds of animals. For in- stance, Van der Stricht did not notice any vacuoles in the luteal cell of a bat, and subsequently Levi’s observations regarding the same animal nearly agreed with those of Van der Stricht. In the guinea-pig, however, Levi noticed a small num- ber of vacuoles, which he attributed to a retrogressive phenom- enon of the cell concerned, and he stated that this phenom- enon, by means of a chemical change of lipoids following it, could cause possibly the substance of the latter to be dissolved by the benzol and xylol which had been used by him as clearing media. In short, whenever the vacuoles 526 GENCHO FUJIMURA appear, the lipoid granules first make their appearance as their forerunners, and the former pass into the latter in suc- cession, as has been entirely agreed upon by many observers. With regard to the origin of lipoid granules, the Zoyas (91) pointed out that there was a certain quantitative relation between granules and plastosomes, and this fact was acknowl- edged by Mulon, Athias, and Tsukaguchi, though Levi alone tried to deny it, as in the case of external secretory granules. With regard to the secreting phenomena of luteal cells, Van der Stricht argued that in the order Chiroptera it was possible to divide them into two forms, viz., serous secretion and lipoid secretion. By serous secretion is meant that the follicular epithelium, being the antecedent of the luteal cells, secretes liquor folliculi, and, according to him, the Graafian follicle, after rupturing, still keeps on secreting in this manner for a fixed period, it being characteristic of this secretion that in the peripheral part of protoplasm of the young luteal cell some serous infiltration appears, and for that reason gives that part -generally a somewhat transparent appearance. Already at this period a large quantity of small lipoid granules appears within the cell bodies; however, since these granules do not as yet per- form any functions, he called this stage the serous secreting period of the luteal cells; then, as the ovum settles, the corpus luteum has already reached its highest degree of development, and he termed it the second stage of corpus luteum formation. At this stage the cell body is already filled with numberless lipoid granules, which, undergoing a chemical change, cause the cell body to keep on discharging secreted matter until the end of pregnancy. This is what the lipoid secretion means. ‘This theory seems to have been afterward accepted in the main also by Levi. However, the condition is entirely different in the Rodentia, and especially in the rabbit. According to the obser- vations of Cohn and Tsukaguchi, lipoids chiefly appear in a high degree only in the early part of pregnancy, then disappear rather speedily. In the second half of pregnancy the lipoids change into vacuoles almost as transparent as water, the cell bodies pre- sent a highly distinctive vacuolar structure, such condition being SECRETORY FUNCTIONS IN HUMAN PLACENTA 527 still more conspicuous toward the last period of pregnancy. Now comparing this with what is observed in the Chiroptera mentioned above, this period of lipoid secretion, so-called by Van der Stricht, in the rabbit passes away in a comparatively short time, and then slowly passes to the stage of the .character- istic vacuolar image; thus, it appears, it presents a certain stage of secretion which is peculiar to itself. In short, it is in- teresting to note that, in consideration of the changes in the forms of secretion in external secretory cells, the structure, and therefore the form in the secreting functions even in the same luteal cells, differ according as the class of animals differs. That the interstitial cells of the ovary possess the same inter- nally secreting functions as luteal cells, by reason of the close resemblance which they bear to the latter in shape and structure, has been universally acknowledged in the researches made, in various kinds of animals, including the Rodentia, Chirop- tera, and Carnivora, by Regaud and Policard (’01), Limon (’02, 03), Fr. Cohn (03), Regaud and Dubreuil (06), Mulon (11), Athias) (11, 712), Vaneder Sticht (12), Tsukaguchi (12, 713), Levi (’13). Now, according to these observers, the interstitial cells also contain, as do the luteal cells and many other glandular cells, important constituents, such as plastosomes, lipoid gran- ules, and vacuoles. The vacuoles are especially very plentiful, and they, for the most part, present a minute, delicate, and peculiar vacuolar image, the protoplasm proper being barely noticeable around the nucleus. The lipoid granules, as com- pared with the luteal cells, are generally small in quantity and, moverover, it 1s sometimes difficult to detect them. It is customary for the lipoid granules more or less to increase in quantity at the time of pregnancy. It has been equally acknowledged by many observers that the lipoids of the inter- stitial cells generally appear for a comparatively short period, commonly fade away in color and change in quality speedily, thus gradually passing into the vacuolar substance. And then, many authors, except Levi, have proved and recognized, even in this case, that plastosomes have a direct relation in the formation of lipoids. 528 GENCHO FUJIMURA Entering into details, Mulon stated that in the rabbit the plastosomes, granular in form at first, change into minute siderophil granules, and then diffuse siderophil substance, and the lipoids will be formed from the latter. Athias (12), who examined a newborn bat, also seconded the argument of Mulon for the most part, and argued that plastosomes produce the lipoid first at their center and accumulate it there, and in sup- port of his argument he stated that occasionally the cortex, which has the same staining properties as plastosomes, could be detected around the lipoid granules. Tsukaguchi certified to an intermediate type of granules between the plastosomes and the lipoid in the young interstitial cells, thus arguing that the granular plastosomes develop and grow in size directly into the lipoid granules, as are seen in the case of the luteal cells. In short, the various investigators have not as yet come to an agreement in their views as to the correlations between the two, and yet it has been universally acknowledged by them that as the cells grow and increase and the lipoids or vacuoles appear plentifully, the plastosomes decrease in quantity in inverse ratio. With regard to the organs other than the ovary, Mulon (10 a,b) closely examined the suprarenal capsules of a guinea-pig and rabbit, and stated that by the conglutination of the gran- ular plastosomes was produced directly a substance (possibly our lipoids) which has affinity for osmic acid (osmophil) or iron- alum-haematoxylin (siderophil), and is introductory to the formation of a vacuolar secreting matter to follow. Celestino da Costa (07), Champy (’09), and Colson (’10), also having experi- mented on the cortical cells of the suprarenal of a cat, guinea- pig, toad (Bombinator) and bat, have proved the same fact as above. Bobeau (’11) also argued that when the parathyreoid glandular cells of a horse form a certain effective product the plastosomes should play an important part. And, besides, Engel (09), Mawas (711), and Schultze (’11), each demonstrated the plastosomes in the thyreoid and parathyreoid glands of a man, a rabbit, and a frog, and, according to Duesberg, it was stated that the plastosomes could be detected in the Langhans’ cells SECRETORY FUNCTIONS IN HUMAN PLACENTA 529 of the pancreas. It should none the less be stated that a ma- jority of the facts enumerated above are merely preliminary. Many future investigations must be looked forward to for a further enlightenment of the secretion of these organs. In summarizing the histological knowledge we have at pres- ent of the internal secretory cells described above, almost every cell has, as its constant ingredients, plastosomes, lipoids, and vacuoles. Now, the plastosomes are not regular either in their shape or arrangement while the lpoids are not quite regu- lar in their size, quantity, and color, but the smaller ones bear a resemblance to granular plastosomes, while the larger ones, agglutinating with one another, form larger fatty droplets. The contents of the vacuoles probably consist of watery, transparent droplets, which are separated from one another by a very thin partition wall. The cell body should present a more or less conspicuous alveolar structure in pro- portion to the quantity of vacuoles contained. The quan- titative correlation by which these formative ingredients are connected to one another is not free from considerable varia- tions. Occasionally only one or two of them exist to the abso- lute exclusion of all the rest. It must be very often the case that such phenomenon may be partly due to the difference in the order of animals chosen for the subject of study and partly to the functional relation of the cell concerned. In short, it may be said that the various formative ingredients described above as being seen in internal secretory cells constitute equally necessary constant elements, the same as with external secre- tory cells, as has been minutely dwelt upon previously. Now, it is a marvelous sight indeed to compare histologically these two kinds of cells, and to look at the perfect agreement, not only in their structures, but also in the various histological changes which follow their functions. On this score, I am con- vinced that should there be a structure like the two kinds of cells described above, besides the functional changes which very nearly correspond to the above, it would be no error to bring such cells under the category of glandular cells, regard- less of the existence of a duct in them. And, if we look at the 530 GENCHO FUJIMURA various cells of the placenta and the decidua, we will find that all of them are well furnished with various conditions which mark the glandular cells above referred to, and that naturally we should assert the existence of secretory processes in them also. In the following section, I will give a detailed account of the secretory phenomena of the various cell groups. THE PHENOMENA OF INTERNAL SECRETION IN VARIOUS CELLS OF THE PLACENTA AND DECIDUA It is too plain to need argument that in all cases the real state of the life processes of any cell cannot be made clear unless all the phenomena in the living condition of the cell concerned be followed up closely with the microscope. However, since it is certainly difficult to attain such an object by the histologi- cal method employed by us at the present time, by noting the phenomena of internal secretion it is possible to denote each of the extremely varied structural images obtained from the prep- arations fixed and stained as indicating a certain period in the phenomena of secretion; to compare carefully and consider the correlations between the different periods, and thus to infer the whole of the process of secretion. Therefore, I must ask the reader to take this point into consideration. 1. The phenomena of internal secretion in the syncytium layer Now, at first, in figures 1 and 2 only plastosomes are present, and lipoid granules and vacuoles are entirely absent, so that we may conclude that this shows the early stage at which secre- tion is not yet in appearance or when secretion is at rest. In the next place, in figures 3 and 4, more or less lipoids appear, and vacuoles are either barely found at one part or not formed as yet; the lipoids, in general first appear on the superficial layer, at which place they tend to grow up and increase gradu- ally. The plastosomes have decreased considerably in quantity especially in figure 4. The last two figures may be taken as the early stage of secretion in the syncytium layer, and as it passed to the subsequent stage, a large number of vacuoles, viz., secre- SECRETORY FUNCTIONS IN HUMAN PLACENTA 531 tions, appear besides the lipoids. The vacuoles later on keep increasing, while, on the contrary, the lipoids decrease in quantity, and, moreover, of these vacuoles those which are near one another unite into vacuoles of various sizes, and it will be seen that the surface vacuoles of the latter ultimately rupture and open upon the surface of the syncytium layer. The process of secretion described above may be followed in figures 5 to 7. Should this process continue and reach its highest degree, such structural images as are shown in figure 8 would be probably brought about in the end! What is pecu- liar in these various stages when secretion is very high is that plastosomes are detected. In the next place, what is shown in figure 9 is taken, similarly to what is in figures 3 and 4, for a comparatively early stage of secretion. However, the former differs more or less from the latter in that already a somewhat large quantity of vacuoles is noticeable in it. And, especially, the vacuoles are chiefly arranged close to the Langhans’ cells. Figure 10 demonstrates on one side a large quantity of plasto- somes and on the other a region where no plastosomes are to to be found. This region presents, as mention has been made already, a foamy structure which can be detected only by care- ful attention. Figures 11 and 12 each clearly demonstrate plas- tosomes, besides a large quantity of vacuoles and a small quan- tity of lipoids. That is to say, in figures 9 to 12 it is always easy to demonstrate plastosomes at different periods of secretion, and therefore the various periods of secretion shown therein may be differentiated from those given in figures 5 to 8, though it is not easy to decide what correlations these stages have peri- odically between themselves in point of secreting functions. However, according to the various images described above, it is deemed practicable in the main to arrive at the following presumptions with respect to the phenomena of secretion of this cell layer: a. At a time when function of secretion has not yet begun, the chief ingredient of this layer is plastosomes, which are found in a very large quantity; however, as the function begins, they suddenly decrease in quantity, and at a certain period, 532 GENCHO FUJIMURA especially when lipoid granules are exceedingly plentiful, they entirely disappear. It has already been explained how this lack of plastosomes is not the result of the want of skill in the making of preparations. b. As the function begins, lipoid granules first appear con- spicuously, especially at a place which is close to the super- ficial layer, and then vacuoles appear. Sometimes both can appear simultaneously at a comparatively early stage (fig. 9). c. The direct relations between the lipoid formation and plastosomes could not be ascertained in my preparations. How- ever, the origin of lipoid granules was found in an extremely small granular body, and on the other hand it is clear that there is a tendency for the plastosomes either to decrease more or less in quantity as function of secretion increases or to en- tirely disappear. d. The transparent halos which are sometimes found around the lipoids give the latter an appearance of being the contents of vacuoles, and by reason of such a condition we are led to be- lieve that a certain intimate relation exists between lipoids and the vacuolar formation (figs. 7, 9, and 12). e. However, we cannot as yet positively state whether or not all vacuoles without exception have an intimate relation with lipoid granules such as is described above. For instance, I cannot definitely declare whether the formation of the extremely delicate foamy image such as is seen in a part of figures 8, 10, and 12 has been the result of the lipoid granules while in the earliest stage of lipoid formation, that is, while as yet in a stage in which they are exceedingly small, having changed their quality, and caused such small vacuoles to grow in groups, or whether as Mulon quoted before, has observed with respect to the ovarian interstitial cell, at the period of his ‘diffused siderophil substance’ (though it is still beyond my power to prove that such a period could appear in the syncytium layer), the substance has, instead of forming lipoids directly assumed a vacuolar formation. It is for future investigations to solve such a question. SECRETORY FUNCTIONS IN HUMAN PLACENTA 533 f. All vacuoles gradually unite, and it appears that they possibly make a kind of canal system running irregularly and crosswise within the syncytium layer, and some of them distinctly open their mouths in various places on the surface of the syncytium layer, thus it is apparent that their contents, their secretions, are drained out into the maternal vessel of the inter- villous spaces. And, as is seen in figure 11, the blood corpus- cles which are noticeable within the syncytium layer may be rightly taken for the mother’s corpuscles which have accidentally gone into the canal system mentioned above, which would in turn prove the existence of the latter. g. The secreting function is at work from the beginning of pregnancy to the end of the fourth month, though it is most active in the second and third months. 2. The phenomena of secretion of the Langhans’ cells The protoplasm in the Langhans’ cells which are still small and should be deemed comparatively young contains plastosomes only (figs. 13 to 16); however, as the cells reach a certain size the lipoid granules appear (figs. 4, 11, 12, 17, and 22). The latter have developed from extremely small granular bodies such as are visible in figures 9 and 21, and their number is not very large; the vacuoles appear in a very conspicuous manner, and it seems that these have also grown into what they are compara- tively speedily from a very minute form, it being clear that the larger of them have been brought into being by the agglu- tination and joining together of some of the vacuoles. And, while the secreting process is in progress, it seems that lipoids play a directly essential part; how they change their quality and liquefy from the periphery, and thus gradually pass into vacuoles as in the syncytium layer, can be very clearly seen in figures 12, 23, and 24. The loss of lipoids which keeps going incessantly by such vacuolation is made good by fresh forma- tions elsewhere, and thus, the same process being repeated, the vacuoles go on growing in number and size simultaneously as the cell appears more and more to increase in its capacity, 534 GENCHO FUJIMURA though the preparations I have are still very poor to prove this positively. However, the fact that, as mentioned above, lipoids have their origin in very small granular bodies and that plastosomes considerably decrease in quantity as the cell grows in size and therefore its functions are promoted (please re- fer to figs. 13 to 21), cannot but be taken for having proved that plastosomes, owing to their quantitative relations, take part in the lipoid formation and, if this deduction be practicable. it would follow that the large group of plastosomes which makes its appearance in an especially limited section, .as is seen in figures 3 and 9, should not be without significance for the new growth and supply of lipoids. In short, the phe- nomena of secretion of Langhans’ cells, in general, are not very active and yet the cells in the Langhans’ islets are somewhat different, and it seems the functions are very active in this part, so much so as to make it a feature of these cells that they present a very large and highly vacuolar formation (figs. 21, 22, and 26). ; Since the Langhans’ cells always have on the surface a com- paratively conspicuous border membrane, there is no alterna- tive for the contents of vacuoles, viz., secretions, but to pass through this membrane and be drained into the villous tissues, and they are therefore possibly bound to be ultimately absorbed on the side of the embryo. However, since the part like the Lang- hans’ islets where the functions are necessarily very active is, as is well known either mostly wrapped up in the decidual tissue or exists within the intervillous spaces, floating directly in the mother’s blood, it is possible that the secretions coming from such a place are absorbed on the mother’s side. More- over, the large number of vacuoles which is found in a part where the Langhans’ cell layer comes in touch with the syncy- tium cell layer as in figures 9 and 12, judged from the position they occupy, has been temporarily denoted by me as forming secretions of the syncytium layer and so described in the previous paragraph; however, I am afraid nobody can say for certain that it is so. If we suppose that the secretions are brought forth by the Langhans’ cells, who may say they will SECRETORY FUNCTIONS IN HUMAN PLACENTA 535 not come out into the intervillous spaces together with those of the syncytium layer? In a word, since the Langhans’ layer is entirely closed against the mother’s body by the syncytium layer in the early stage of pregnancy, it would follow that the secretions are entirely in the service of the embryo, but after that it is probable that a part of them are taken also by the mother’s body. The function of secretion of the Langhans’ cells, just like the syncytium layer, is active in the main almost from the first stage to the end of the fourth month of pregnancy, though in the second and third months it is very active, suddenly subsiding with the fifth month. In the Langhans’ islet it con- tinues still longer and commonly gradually subsides after the sixth or seventh month. The epithelium of villi is located between the circulation of the mother’s body and that of the embryo. and it is for this rea- son presumed that it must be an organ which takes nutrition for the embryo, as has been commonly held in literature, but that such is a groundless assumption must be quite clear from my histological observations given above. And, besides, there are some important reasons which prove the utter fallacy of this theory. It is in the first to third months of pregnancy that the growth of the epithelium of villi is most active. If the functions of the alimentary organ for the embryo be assign- ed to it, the epithelium of the villi should go on developing most vigorously, but the fact is quite the reverse, and it retrogrades and becomes thin in the second half of pregnancy, and accord- ingly the decline of functions is brought about. This is one of the absurdities. And in the eighth month of pregnancy, when the embryo calls for a still greater increase in the supply of its nutrition the capillary blood vessels of villi increases suddenly, as was mentioned above, and early in this stage the epithe- lium of villi becomes remarkably regressive and falls into decay, so that the embryonal circulation of the villi is separated from that of the mother only by a thin membrane like endo- thelium, undoubtedly it being quite easy for both to allow the interchange of materials between them. In other words, the 536 GENCHO FUJIMURA particular organ which is needed for the absorption of nutrition for the embryo first makes its appearance in a perfect condition only after the epithelium of villi retrogrades and becomes thin. On this score I am led to believe that the epithelium of villi is simply an organ of internal secretion, and that the ground is extremely weak for the argument, which treats it as an organ to take nutrition for the embryo, as has been generally conjec- tured in the past. 3. The phenomena of internal secretion in the stroma cells of villa The smallest of the stroma cells of villi is simply a ball- shaped cell which is comparatively rich in protoplasm, and within the cell body there is a large quantity of plastosomes (fig. 27), but presently a somewhat large quantity of lipoid gran- ules or vacuoles having various sizes appears within the cell body (figs. 28 and 29); and: subsequently, as the cell grows in size the chief ingredients of the cell body will be plastosomes and vacuoles, while the appearance of the lipoids is not very distinct. The image such as is seen in figure 34 is very seldom met with. On the contrary, however, the vacuoles may be deemed the almost constant ingredients of each cell, and especially as the cell developes and grows in size they increase the more in size and quantity, and present a highly foamy structure which is characteristic of this kind of cell. Now, if we consider the correlations between the different con- stituents mentioned above, it will be found first that in these cells the plastosomes are stained comparatively easily, and are therefore very distinctly detected in each cell; and as regards its quantitative relations, it will be noted that there is not the least tendency in the plastosomes to decrease in quantity, even though the functions increase and the cells enlarge, as was seen in the epithelium of villi described above. On the contrary, the plastosomes crowd together in large numbers in the various protoplasmic sections of cell bodies, and they present the ap- pearance of a new growth and multiplication in the sections concerned (figs. 30, 31, 32, 33, 36, and 38). The only excep- tion is that when the quantity of lipoids contained in the cell SECRETORY FUNCTIONS IN HUMAN PLACENTA 5037 body is remarkably large, the plastosomes decrease more or less remarkably in quantity (figs. 28 and 34). The lipoid granules, as mentioned above, do not appear in very large numbers and, since they arise from very small granular bodies, it is very difficult to clearly discriminate the latter from the ordinary granular plastosomes (figs. 28, 29, 31, 36, and 37). Therefore, in consideration of this fact and of the quantitative relations between lipoids and plastosomes as described above, I am in- clined to trace the mother-ground of the lpoid formation in the plastosomes. Then, with regard to the vacuolar forma- tion, we may infer from the conspicuous halos which often ap- pear around the lipoids, or from a phenomenon in which the lipoid often occupies the position of a nucleus within the vac- uole (figs. 29, 34, and 36), as in the case of the epithelium of villi as described above, that the vacuole should of necessity be the liquefied product of a lipoid. In short, it may be stated that the secreting phenomena of these cells, if looked at from the histological view-point, are very simple indeed, and _ plas- tosomes first bring forth lipoids, which latter in turn change into vacuoles, and the reason why the lipoids are comparatively scant is that the period of their appearance is exceedingly short. And, while perhaps on one hand the contents of vacuoles, viz., secretions, are gradually drained out of the cell body, on the other the protoplasm and therefore the plastosomes bring about a prospective new growth and multiplication, presum- ably to provide for the materials of the next secretion, and in this manner the afore-mentioned process, as simply it may be, is repeated and follows in succession. At different times and in different places, to make a secondary or tertiary secreting process within a cell all the time, the cell develops and grows in size gradually, and its structure therefore becomes extremely complicated, and in this way I suppose that, even in one and the same.cell body, the various periods of the phenomena of secretion make a simultaneous appearance according to the ingredients contained. This is a mere hypothesis of mine, and yet since it was early refuted by M. Heidenhain that the secretory granules of all kinds start their individual function 538 GENCHO FUJIMURA separately as a small independent ‘organel’ within the cell, this hypothesis of mine should not be taken exception to. And moreover, the increase and mass of protoplasm or plastosomes and the lipoids at different phases, all of which could be demon- strated in every part of this cell at any time, if sought for an interpretation of their significance, will each provide a mate- rial to substantiate the hypothesis mentioned above. In this manner, this cell, while promoting its secreting functions on one hand, grows in size more and more, and such lke rela- tions could be recognized more or less even in the Langhans’ cell. The functions of secretion in the stroma cells of villi begin at the end of the first month of pregnancy and _ continue actively until about the seventh month, though they are most active from the second month to the sixth. And though in the eighth month it appears that they suddenly subside, it will be found at all times that it is difficult clearly to follow the destiny of each of the cells, since it is interfered with by the strong increase of the capillary blood vessels of villi at this stage, as mentioned already. Since no special duct was detected, it is difficult to tell how the secretions are removed, other than by attributing it to osmose, as in the case of the Langhans’ cells, and by predestining the secretions to be absorbed by the embryo. 4. The phenomena of internal secretion of decidual cells If we first look at the decidual cells of the smaller type (figs. 39 to 51), we find that the chief components of the cell body in the youngest are plastosomes (figs. 39 to 41), next appear lipoid granules (figs. 42, 43, and 44), then follow vacuoles, it thus be- ing customary for the great majority of decidual cells of smaller type to contain many vacuoles and more or less lipoids besides plastosomes. The plastosomes sometimes decrease more or less in quantity in inverse ratio to the lipoids or vacuoles (figs. 43 and 49), but more often is it difficult to discern such relation, and, besides, many are plastosomes which either form a con- spicuous group in some part of the cell body or considerably SECRETORY FUNCTIONS IN HUMAN PLACENTA 539 increase in quantity (figs. 44, 48, 50, 51, and 52). The lipoid granules arise at the beginning in a very small granular body, whence they grow up to a certain degree, when clear halos ap- pear around them, and the manner in which they directly participate in the formation of vacuoles (figs. 44, 46, 52, and 53) is the same as what is observed in the epithelium and the stroma cells of villi. And sometimes it occurs that the vacuolar for- mation appears equally at a time within one and the same cell body, and as a result the foamy image of high degree, such as is illustrated by figure 49, is brought into being, but this is rather rare, and in most cases the vacuoles vary in their sizes. And, besides, it is customary for the vacuolar formation in most cases to contain at the same time groups of plastosomes or lipoids of different sizes. In short, the smaller-type decid- ual cells entirely agree with the stroma cells of villi in their structure, and, therefore, there is no need for argument that their secreting phenomena should be dealt with in the same manner as the latter. On this score, I will not go to the redund- ant trouble of touching upon the secreting process of the smaller type decidual cells here, but will confine myself to the brief statement that the function is repeatedly performed by the same methods as in the stroma cells of villi. If we take a glance at the figures in the plate, it will be quite clear that the smaller-typed decidual cells, repeatedly perform- ing as they do the functions as described above, develop and increase in size more and more, and passing through the various intermediate types (figs. 52 to 54, 62 and 63) grad- ually, as I mentioned in the previous chapter, pass into the larger-type decidual cells to attain the height of their growth. Therefore, the demarcation between the large and small types in the decidual cells is, after all, due to the difference in the de- gree of growth of the same kind of cells, and the smallness of the cell should be taken for an indication of comparative infancy, while the largeness of the cell shows that it has attained the region of perfection in its growth. The large-type decidual cells may be divided into two kinds with respect to structure. One represents the kind of cells whose 540 GENCHO FUJIMURA body is protoplasmic, and commonly has a large number of plastosomes, besides more or less lipoids which are often discernible, though vacuoles are almost absent (figs. 55 to 61). The other represents those cells whose body presents a highly vacuolar image, whereas the protoplasm considerably decreases except around the nucleus, while no plastosomes are: to be found. The lipoids contained are irregular in their quan- tity, but more or less of them are always existent (figs. 64 to 69). A great majority of the commonly so-called decidual cells belong to the former, while a comparatively small number is represented by the latter. In the former class the structure of the cell is entirely different from the small-type decidual cells and my ‘intermediate type,’ so that along lines of histology there is no indication of the existence of the process of secretion, and although. lipoids exist in small numbers, their quantity quickly decreases and they go out of existence as the cell body grows up in size, so that it would be in order to denote the lipoids rather as persistent bodies bequeathed from a period of their growth, and consequently it follows that it would be no great error to conclude that at this period a secretion, such as was notice- able at the period that preceded it, either considerably declines or entirely disappears. However, in the various cells which belong to the latter class, the afore-mentioned secretory func- tions are developed to the extreme throughout all their growth, and there is an appearance which points to the utter exhaus- tion of plastosomes on account of these functions. From the scantiness of materials, it is difficult to determine the destiny of this kind of cells; whether the cell body ultimately breaks up and decays or is absorbed or whether after throwing out the secretions, the plastosomes again increase or are re- plenished, and thus it slowly passes into the former class of cells; however, I at least am confident that it would be prema- ture to assert that the various periods illustrated stand for a direct indication of retrogression or decay. In short, in the larger-typed decidual cells, it is possible clearly to observe in ' a portion of them the same process of secretion as in the small SECRETORY FUNCTIONS IN HUMAN PLACENTA 541 type cells, whereas in the largest number there is almost no sign of such a function, which fact is worth much attention. The large-type decidual cells, as aforesaid, no longer present the ordinary phenomena of secretion for the most part, and yet, at a certian period, dark-stained coarse granular bodies of ir- regular sizes often make their appearance within the bordering membrane of the periphery (figs. 55, 59 to 61); various material products having the same staining properties are often de- tected within the interstitium (figs. 70 and 71), which makes one feel that there is a certain formative relation between the two. And, besides, similar products often filling up the blood vessels around them, give the appearance of being absorbed in the vascular organs (fig. 71). Such peculiar products hay- ing been originally observed in the fixed preparations, it follows that they might be an artificial product, a result of the fixatives, and yet from the observations mentioned above it is not diffi- cult to conclude that a certain material which corresponds to them is prepared, perhaps by some special function of the cell mem- brane of the cell concerned, and is sent forth in the direction of the interstitium. And should this supposition prove correct, it would follow that these two kinds of large-type decidual cells are functionally quite independent of one another, though they are genetically of the same origin. Looking on the whole of the functions of the decidual cells from the histological point of view given above, I am led to be- lieve that they may be roughly divided into three periods, ac- cording to the course of their development. The first period is seen in all the small-typed decidual cells, the intermediate type so termed by me, and in a portion of the large-typed cells, here the secreting functions are distinctly performed in the same manner as in the stroma cells of villi. In the second period, possibly by the functions of the cell membrane, a certain product is prepared, to be sent forth in the direction of the interstitium. ‘The third period begins after the sixth month of pregnancy, when the cell body in general shrinks considerably, ‘and no plastosomes are to be discovered, besides no particular tissue structures from which inference may be made of the functions performed are to be recognized. And, moreover, cells 542 GENCHO FUJIMURA at this period experience the rise of embryonal pressure (the inner pressure of the uterus) as the time of pregnancy elapses, in conse- quence of which they are remarkably flattened and afterwards present the appearance of flattened epithelium. And, as regards the destiny of decidual cells, it seems that it has been argued in the past that they retrograde and perish by fatty degenera- tion or coagulative necrosis (Klein); however, as a matter of fact, I have not as yet discovered such a change. All kinds of decidual cells perfect their growth comparatively rapidly early in the beginning of pregnancy, viz., in about three weeks after pregnancy, and after that only a quantitative increase or decrease of the various cells occurs. Consequently, the degree of growth of the cells cannot be the sole measure of the time of pregnancy. However, judging from their quantitative rela- tions, it is not difficult to arrive by way of inference at the approximate period of pregnancy; that is to say, the small type cells appear from about the second week of pregnancy to the end of the first month, and the intermediate-type cells from above the seventeenth or eighteenth day to the end of the second month, in both cases in exceedingly large numbers, while the large type cells appear throughout the whole remaining period begin- ning about the twenty-second or twenty-third day of pregnancy, and yet it will be noted that these large cells are at the height of their activity during the period from the end of the first month to the end of the third month of pregnancy, and while in the fourth month the functions are still pretty high, they consider- ably decline in the months to follow, and in the seventh month and after it is very seldom that such functions are clearly noticeable. The phenomena of secretion of cells in the decidua serotina in the first half of pregnancy are nearly the same as in the decidua vera as described above. In the second half, especially after the eighth month, giant-cells grow in large numbers, and somewhat remarkable changes take place, even histologically in the ordinary sense of the term, and, therefore, I have examined the subject with an especially keen interest; however, the absence of good materials, coupled with the difficulty in stain- ing them has hindered me in making excellent preparations, SECRETORY FUNCTIONS IN HUMAN PLACENTA 543 and the functions of cells at this period are therefore set aside for future investigations. 5. The phenomena of secretion in the uterine glandular cells at the tume of pregnancy The epithelium of the uterine gland undergoes a remarkable change in the early part of pregnancy, viz., on about the seven- teenth or eighteenth day after conception (figs. 72 to 83). Now, if we consider its phenomena of secretion, we shall find that, even in this cell, lipoid granules first appear, and then vacuoles are formed. In the stained preparations, lipoid granules appear assembled and are accumulated especially near the basal part of the cell body and show a remarkably clear yellowish-brown color (figs. 73 to 78). The latter often appearing as contents of vacuoles (figs. 77 and 78), it would probably appear that the vacuoles are a modified product of the lipoids, just the same as in the other cases. In this way the lipoids gradually change into vacuoles, the cell grows in size and presents a highly honeycomb structure (figs. 77 to 80). The plastosomes either decrease in quantity or become very difficult to discover as the functions of secretion increase in activity. However, I have not been able to make clear the formative relations between the plastosomes and lipoid granules. Be that as it may, it happens that, with the increase of the function of secretion and the growth of the cells, the latter gradually move over toward the comparatively enlarged glandular lumen, and, at last, leaving the wall of the glandular tubule, are entirely free within the glandular lumen. The characteristics of these desquamated cells are that either the cell body shows a highly vacuolar formation or that the vacuoles being somewhat reduced in quantity, the protoplasm becomes dark and turbid, and no plastosomes are to be found. The condition of the nucleus is also exceedingly abnormal (figs. 81 to 83). How the various cells of this kind are broken up by degrees and added to the large quantity of fragments filling up the glandular lumen can be observed and followed with a great certainty. The various changes of the uterine glandular cells as described 544 GENCHO FUJIMURA above have, as was mentioned in my own observations, a definite relation to the time of pregnancy, and accordingly the rise and fall of the functions of secretion of these cells also act upon it; that is to say that, in the first month and the first half of the second month of pregnancy, the functions are at the height of their activity, and they subside considerably from the beginning of the third month, the subsidence being by far the greater in the fourth month, and in the fifth month they seem to come to a standstill, it being no longer possible to demon- strate the function of secretion in the months that follow, viz., in the second half of pregnancy. In the next place, the afore- mentioned functions of the glandular cells, as compared with the decidua vera, appear more speedily and in a still higher degree in the decidua serotina, and fail accordingly earlier than in the former, and everybody easily recognizes that the secre- tions of the glandular cell and its broken-up matter both accu- mulate in the glandular lumen for a certain period, though some consideration should be given the question as to how they are removed or absorbed. It is said that, according to what has been written on this subject, the placental formation commences from the second month and is perfected in the fourth month and that the decidua reflexa and decidua vera are agglutinated in the fifth month. Should this opinion be true, it would follow that the secretions of the uterus, looked at from the periodic relations of secretion, are for the most part drained into the uterine cavity, and take a part in the forma- tion of the so-called uterine milk. However, according to my own experience, it appears that the placental formation and the adhesion of the decidua reflexa take part in an earlier part of pregnancy. Therefore, I am inclined to believe that a part of the secretions and detritus of the uterine glands, at least in a little advanced period of pregnancy, are naturally absorbed by the mother on account of the closure of the ducts. SUMMARY It is to be observed that the syncytium layer, Langhans’ cells, stroma cells of villi, decidual cells, and uterine glandular cells, all of which constitute the chief tissue elements of the placenta SECRETORY FUNCTIONS IN HUMAN PLACENTA 545 and decidua, each contained plastosomes as a constant in- gredient of its protoplasm, and that a majority have at the same time a certain quantity of either lipoid granules or vacuoles, or of both, and, consequently the minutes histological structure of these cell groups bears a close resemblance to that of both the internally and externally secreting cells. And, more- over, these main components of protoplasm or cell body are, according to their functions, as closely correlated to one another as they are in the glandular cells. Now, taking a general sur- vey of this correlation, it was found in my study that the plas- tosomes, being the first constituent, appear for the most to be the matrix of lipoid granules from which the latter rise, and as ia proof of this argument, I will cite the stroma cells of villi, in which the correlation between the two is very closely shown. We can notice it somewhat in the Langhans’ and decidual cells, and if we closely examine the manner in which the lipoids appear in these cells, it will be found that they always rise from gran- ular bodies which are very small and strongly siderophil. I believe that these may be rightly compared with the so-called ‘Primargranulis’ which Heidenhain found in the common glandular cells, and even though they appear very small, they do appear as a perceptible body. There is no evidence to be found of their appearing as slowly growing and increasing, as Heidenhain assumes to be the case, from an infinitely small body which is hardly seen microscopically until they enter the vision of a microscope. Rather is it found that some of them bear a close resemblance to the granular plastosomes in their size and staining properties, clearly indicative of images running over between the two (figs. 28, 29, 31, 36, and 42), which will account for my argument that plastosomes should be deemed the matrix of the lipoid formation. And, for the second rea- son, I will give the fact that the plastosomes, either being con- siderably reduced in their quantity or having gone out of exist- ence, as the lipoid formation progresses, are scarcely detected. Such is the fact which is often noticed in all the cell groups other than the stroma cells of villi, and a part or the whole of the plastosomes cannot but be seen as having been consumed 546 GENCHO FUJIMURA or exhausted in the formation of lipoids. However, in the Lang- hans’ cells, the stroma cells of villi, and in the decidual cells sometimes, when the functions have advanced considerably, the plastosomes not only show no sign of their decrease, but also increase and present more or less conspicuous groups in a limited section of the cell body. This apparently contradicts the statement given above, but, practically, the reverse is the case. It is probable that the plastosomes consumed partly by the functions performed, are increased and replenished, provid- ing for the repetition of secondary and tertiary functions; by such an assumption the significance of the increase of plastosomes in these cases will be made naturally clear, so that the various images described above do support with more force, instead of contradicting, the theory mentioned above. And then, the lipoid granules growing and enlarging, as they do, from the very small granular bodies described above, change more or less in quality at the same time, and their color becomes somewhat faint, and, moreover, in certain cells, as for instance in a part of the epithelium of the uterine gland and the large- type decidual cells, they sometimes appear as granules havy- ing a very clear yellowish-brown color. At any rate, when they reach a certain degree of development, these lipoid granules create more or less conspicuous halos around themselves, which gives them the appearance of the contents of vacuoles. Such appearances are very commonly noticed in all the cell groups I have examined, and I cannot help recalling to mind the obser- vations made by Babkin, Rubaschkin, and Ssawitsch respect- ing pancreatic cells as cited before. Therefore, I believe that this appearance has a very great significance in the secretion formation, in the same way as the phenomena of secretion of the pancreas as interpreted by these three observers just referred to does. In other words, the lipoids may be compared with the secretory granules of ordinary cells, and they like the latter are slowly liquefied, in accordance with the third one of the various forms of glandular secretion (liquefaction) described above, and pass over to the secretions of a vacuolar shape. There- fore, the vacuoles are, after all, nothing else but a modified SECRETORY FUNCTIONS IN HUMAN PLACENTA 547 product of lipoids, and the contents should possibly be secretions as transparent as water. According to the arguments given above, the various cell- groups of the placenta and decidua entirely agree with the ob- servations of the glandular cells not only in their structure, but also in the histological changes that follow their functions, and, therefore, there is no room for doubt that secretion should be existent in them also. And it will be briefly stated, con- cerning their secreting phenomena, that probably lipoid gran- ules rise directly from plastosomes, and then. the former, grow- ing in size, slowly change to the vacuoles, viz., secretions, and are thus thrown out of the cell body at times. If looked at from the standpoint of their secretion formation these cells, for the most part, closely resemble the external secretory cells, but viewed with regard to their inner structure in which they keep secretions within their own bodies for a comparatively long period and thus for the most part present a more or less conspic- uous foamy image, they should be rather compared with the various internal secretory cells, which are observed in the ovary and the cortex of suprarenals. The principles of the phenomena of secretion, as aforesaid, look very simple indeed, and yet these phenomena do not make their appearance in one and the same cell necessarily at the same time. On the contrary, it is cus- tomary that within different parts of the same cell body the various stages of phenomena appear, one after the other, in conse- quence of which the structure of each individual cell becomes comparatively complex and diverse. Each individual cell, while repeatedly performing its secreting functions in this manner, gradually increase in its size, and it is customary for the cell to grow considerably as it reaches the height of secretion. Even the syncytium layer whose cell border is indistinct, is generally very thick at the height of secretion, and the gradual increase in the size of the cell along with the rise of its secreting functions in this manner may be partly due to the accumulated assemblage of the secretions, though at the same time it cannot be denied that the rise of secreting functions is attended by the increase of the protoplasm and the growth of the nucleus. 548 GENCHO FUJIMURA There is, of course, a certain limit to the growth of each cell, but there is something exceptional about the decidual cell. It rises from certain extraordinarily small spherical cells within the proper mucous membrane of uterus, and yet it grows so very rapidly and becomes enormous in size that the classi- fication between the large and small types in the ordinary de- cidual cells, if dealt with according to their genesis, should be any- thing but significant. For these two mutually run over to one another through the intervention of the intermediate type, and no sharp demarcation exists between them. There- fore, these two kinds, histogenetically, belong to exactly the same kind of cell, and they only differ in that one is still young in its growth while the other has already perfected its growth. However, it must be noted here particularly that the two pre- sent an entirely different appearance histologically, and, there- fore, in all probability, along lines of their physiological functions. In other words, the decidual cell entirely changes its structure and functions along with the perfection of its growth. That is to say, the decidual cell which has perfected its growth no more demonstrates within its body any important tissue in- gredients, except plastosomes; however, it seems that probably, at this period, the cell prepares, by means of the special action of a very strongly developed cell membrane, certain secretions, and sends them forth into the interstitium. At any rate, the cell passing through this stage gradually withers and becomes smaller. The secretions, while at the height of their formation, are conglutinated with one another, produce in abundance vacu- oles of various sizes and shapes, and will show a high beehive structure. And the way of their removal and absorption, if in the syncytium layer, will be, undoubtedly, by rupture, sooner or later, toward the intervillous spaces, and thus they will be absorbed in the mother’s blood, while in the various other cells, there is no knowing but that the secretions are thrown out by osmosis, and the secretions of the Langhans’ cells and the stroma cells of villi should, as a matter of course, be absorbed on the side of embryo, with the exception of those, which, finding their out- lets in the Langhans’ islets, are probably taken in by the mother’s SECRETORY FUNCTIONS IN HUMAN PLACENTA 549 body. Both the large and small types of decidual cells cer- tainly belong to the mother’s side, and the secreted or broken- up matter of the uterine glandular cells is at first probably drained into the uterine cavity, to be absorbed by the mother’s side. And, on comparing the relations between the secretion of these various cells, and the time of pregnancy we find that, in general, the secretion is at its height in the first half of preg- nancy, and especially in the early part of that period, whereas in the second half of pregnancy it generally declines consid- erably, it being possible to demonstrate it only for a certain period in the stroma cells of villi, the cells of Langhans’ islets, and in the decidual cells. Below I will give this correlation with a diagram. In short, it may be deduced that all the important tissue ele- ments of the placenta and decidua, if looked at from the histo- logical view-point, perform secreting functions. Pending fur- ther investigations in all possible directions, it would be diffi- cult to tell what significance these secretions thrown out of the various cell groups have physiologically; however, since it is evident that almost all of their secretions are internally rejected and are taken in either by the mother’s or by the fetal side, it makes one feel that, either by the codperation of certain ‘hor- mones’ which should of necessity be contained in each kind of secretions or by their contending actions, both the mother and the fetus would enjoy an extremely delicate and _ special physiological action. If that is so, it follows that the placenta should contain a great variety of ‘hormones,’ and the kind and quantity of ‘hormones’ contained should naturally differ according to the period of pregnaney and the kind of tissues, it being quite clear from the following chart that, speaking generally, those that are found in the early part of pregnancy should be comparatively numerous in kind and in abundance. On the contrary, however, I could not find any important secretions in the placenta which is well ripened. This is the ‘point to which I should like to call attention for the deliber- ate consideration of all observers who are interested in the. placental poison. JOURNAL OF MORPHOLOGY, VOL. 35, NO. 3 550 GENCHO FUJIMURA THE MINUTE HISTOLOGICAL STRUCTURE AND PHENOMENA OF INTERNAL SECRETION IN THE UTERINE MUCOUS MEMBRANE PRIOR TO MENSES 1. My own observations a. The changes of the interstitial cells (figs. 84 to 91). The one shown in figure 84 is an interstitial cell in normal condition, it is very small and ball-shaped, and the cell body as com- pared with its nucleus exceedingly small, containing within a certain quantity of plastosomes which are largely rod-shaped. A diagram showing the correlation between the secretion and the months of pregnancy in the various cell groups of the human placenta and decidua MONTHS OF PREGNANCY ORGANS CELL GROUPS aa : SS a syncytium layer Langhans’ cells Langhans’ islets oe a 3 & a S oO o 8 [any Stroma cells of villi Small and intermediate In the decidua n serotina 4 ra) Forma ——_—____| In the decidua qm ~~ > White: Absorbed by mother. Remarks: \ Black: Absorbed by fetus. Striated: Thrown out of mother’s body. Decidual and Placenta uterina decidua vera Glandular cells The nucleus is extremely clear and contains a nuclear network and conspicuous nucleoli. Figure 85 is remarkably larger than the former and is oval. The protoplasm increases in quantity and so does the nucleus. Within the cell body there are no plastosomes to be found, but, on the contrary, there are, in large numbers, strongly black-stained and almost equally shaped lipoid granules. The cell illustrated by figure 86 is filled by large numbers of vacuoles within the cell body, and very little is protoplasm proper. The plastosomes, being rod-shaped, for SECRETORY FUNCTIONS IN HUMAN PLACENTA oi the most part lie scattered in the partition walls between the vacuoles. Besides, there are, in large numbers, strongly black- stained and almost equally shaped lipoid granules. The cell illustrated by figure 86 is filled with large numbers of vacuoles within the cell body, and there is little protoplasm proper. The plastosomes, being rod-shaped, for the most part lie scattered in the partition walls between the vacuoles. Besides, there are, in various parts of the cell body, a few extremely small lipoid granules, small in numbers. The nucleus is somewhat dark and the nuclear network is indistinct. In this cell and those that are enumerated below there is a somewhat distinct border membrane on the surface. In figure 87 both the cell body and nucleus are. oval and, though the structure of the cell body is similar in general to the former, the lipoid granules appear in a somewhat larger quan- tity, in some cases existing as the contents of a vacuole. Now, the vacuoles grow larger than in the former in general. The plastosomes are comparatively few. In figures 88 and 89 both the cell body and nucleus are somewhat dark in color. Within there are vacuoles which appear in comparatively small num- bers. ‘The plastosomes in the one are somewhat larger in quan- tity and are distributed all over, while in the other they are com- paratively smaller in number and are confined to a certain section. Both demonstrate more or less lipoid granules of various sizes. In figure 89 some of the lipoid granules contained are light-colored, and it is extremely remarkable to find the manner in which they present themselves as contents of vacuoles. Figure 90 illustrates changes of a very high degree, and the cell body is filled up with remarkably large numbers of vacuoles of different sizes, while the plastosomes lie scattered, in somewhat large numbers, in the partition walls of the vacu- oles. The lipoid granules are extremely few in number and are very small in size, while the nucleus is grown in size consid- erably, is clear and has nuclear network and nucleole, both of which are distinct. Figure 91 also shows nearly the same struc- ture as the former, and yet its vacuoles being agglutinated with one another in large numbers, form large and irregular-shaped 52, GENCHO FUJIMURA On cavities, in consequence of which the cell body appears as though it were on the verge of destruction. There are no plasto- somes to be detected, though the nucleus appears in a still full and stained condition, and both the nuclear network and nu- cleoli are conspicuous. b. The changes of the glandular cells (figs. 92 to 96). Figure 92 illustrates, for the sake of comparison, a normal glandular cell, the nucleus is remarkably long and occupies the middle part of the cell body, so that the cell body is divided into the upper and basal parts, each being filled up with num- berless plastosomes. The cilia are somewhat short and thick and are not altogether normal. Figures 93 to 96 illustrate the changes which take place prior to menses. Figure 93, as com- pared with normal conditions, is remarkably larger and_ its nucleus, being relatively small, lies rather inclined to the base of the cell, while the plastosomes, being chiefly short and rod- shaped, largely lie scattered between the nucleus and the top of the cell, it being a peculiarity of this cell that there are large numbers of yellowish-brown lipoid granules assembled at its basal part. Besides, there are in another part of the cell a few deep-back lipoid granules, and, again, in this cell there are extremely large numbers of vacuoles nearly of an equal size, crowding together close to the top, viz., the cilial layer of the cell body, though some vacuoles arrange themselves along the surface of the nucleus in the deeper part of the cell. In the cell illustrated by figure 94, the upper part of the cell is clear, in general, because of the particularly conspicuous vacuolar formations, whereas the common protoplasm is accu- mulated more or less in the basal two-thirds of the cell, viz., around the nucleus, in which part vacuoles are also de- tected, though they are for the most part very small. Besides, in this protoplasmic part there are extremely large numbers of plastosomes, which arrange themselves and crowd to- gether in various directions. Again, yellowish-brown lipoid gran- ules are found in comparatively small numbers in the basal part of the cell, while deep-black lipoid granules, small in size and numbers, lie scattered in the upper part of the cell. The SECRETORY FUNCTIONS IN HUMAN PLACENTA 553 nucleus is relatively clear, and its nuclear network is indistinct. It is easy to find the traces of cilia in figures 93 and 94. In fig- ure 95 there are absolutely no cilia to be found, and the upper third of the cell is remarkably clear and is formed by somewhat large numbers of vacuoles, whose partition walls, having dis- appeared in part, give them the form of very irregularly shaped inner spaces. The lower two-thirds of the cell consist of re- markably dark protoplasm, and has in the middle a somewhat large nucleus. Within the protoplasm there are numberless vacuoles of a small size and comparatively small numbers of plastosomes. Both the nuclear network and nucleoli are very conspicuous. And this kind of cell is to be noticed in greatest numbers during the changes of the glandular epithelium. The cell shown in figure 96 is very weak in staining properties, both in its cell body and nucleus, and its minute structure is by no means ascertained. This kind of cell is very seldom seen, and may probably belong to the regressive type. 2. The phenomena of internal secretion The so-called menstrual decidual cells are extremely varied in their shape and size, and yet, if looked at from the minute histological structure of the cell body, it will be noted that plastosomes, lipoid granules, and vacuoles constitute their chief components. The manner in which the latter, probably follow- ing the functions of the cell, correlate with one another may be easily recognized as being in extreme agreement with what is in the small-type decidual cell during pregnancy, and conse- quently, there is no room for doubt that the functions of the cells concerned are performed in the same manner as the latter. Therefore, not only am I inclined positively to assert the exist- ence of internally secreting functions even in the menstrual decidual cells, but also I believe that the origin of these cells is found in the interstitial cells proper of the uterine mucous membrane, from which origin, gradually with the rise of the function of secretion, a remarkable development and increase of the nucleus and cell body such as is described above are 004 GENCHO FUJIMURA brought about, the relation in this case being in exact coincidence with the growth of the pregnant decidual cells. These facts taken into consideration, I am convinced that the two kinds of decidual cells (menstrual and pregnant) described above have the same origin, and yet the cells being influenced by the physiological conditions sometimes develop into the men- strual decidual cells, and sometimes, being advanced further, run over to the pregnant decidual cells. And, on taking a glance at the epithelial changes of the uter- ine gland, we find that, as in the ordinary glandular cells, plas- tosomes, lipoid granules, and a large number of vacuoles, which last may be deemed a modified product of lipoid granules, are contained therein. The vacuoles grow in size gradually and are finally fused and present a honeycomb structure, especially on the surface of the cell, and then after losing the cilia, the cells assume the appearance of goblet cells which have their secreted matter accumulated chiefly on the surface. Along with such changes, it will be noted, on the other hand, that _plastosomes and lipoid granules gradually diminish and dis- appear, and it appears that part of those cells which show changes in a high degree die and perish. In short, these structural changes cannot but clearly indicate the fact that these cells per- form functions which are similar to the ordinary glandular cells. And, on comparing these changes with those experienced in the glandular cells during pregnancy, we find that the backward- ness in the degree of the appearance of lipoid granules and vacuoles occurring in these cells makes one feel as though a decided difference would exist between the two, however true it may be that no radically great difference exists between them. With regard to the periodic changes of the uterine mucous membrane, there have been many researches, such as the investigations of Hitschman, Adler, and Schroder (07). Though a universally well-known fact, and yet confined chiefly to the shape of the glandular tubules and the epithelium, very few observers have so far paid attention to the functional significance of the so-called menstrual decidual cells which are produced by the evolution of the interstitial cells, with the ex- SECRETORY FUNCTIONS IN HUMAN PLACENTA 900 ception of Asada, who has quite recently demonstrated the ex- istence of fat within the cells concerned, and inferred only that this fat is not a degeneration product and must have some rela- tion to the functions of the mucous membrane. However, ac- cording to my observations mentioned above, it is easy to clearly recognize that, according to their structure, these cells also have secreting functions as in the case of the ordinary decidual cells. And, looked at from the histological view-point, I do not hesitiate conclusively to pronounce that this function declines and terminates immediately upon the beginning of the menses. From want of suitable materials on hand, I am not able to make a definite statement as to what destiny should befall these cells; however, I quite agree with the observations of other inves- tigators in that they suddenly diminish and perish with the arrival of the menses. And, since it is doubtless true that the secretions of these cells are absorbed in the mother’s body, it should be a matter of special interest to consider the several clinical symptoms which present themselves fre- quently at menstruation, in the light of this fact for the expla- nation of their causative relations. On the contrary, the changes of the uterine glandular epithelium, if compared at the time of pregnancy are remarkably small, and as we can easily assert that its secretions are thrown out of the body, there is certainly no need for argument that it is impracticable to attach an internal secretory significance to the glandular cells; there- fore, I am inclined to believe that this sort of periodic changes of glandular epithelium should be recongized as a mere prelim- inary behavior which is antecedent to pregnancy, and that by far the greater significance, rather theoretically than function- ally, should be attached to it. CONCLUSION 1. The epithelium and stroma cells of villi, decidual cells, and uterine glandular cells, all of which constitute the chief tissue elements of the placenta and decidua, if subjected to the closest cytological investigations, show within the cell bodies, and common to them all, the formative constituents, such as 556 GENCHO FUJIMURA plastosomes, lipoid granules, and vacuoles. These constit- uents, along with the functions of the cells, mutually show the requisite correlation with which they are connected with one another. Specifically: a. The plastosomes, though for the most part rod-shaped, are either long or short, but occasionally they are granular, chain-like, or filar in their shapes. Their quantity generally more or less diminishes along with the progress of the secreting functions. b. The lipoid granules are extremely varied in their shape, quantity, and in color (in the stained preparations), according to cell or the group to which the cell belongs or perhaps in ac- cordance with the difference in the period of functions. In the earliest stage of their appearance they are always granular- shaped of extremely small size, and sometimes it is difficult to distinguish them from the granular-shaped plastosomes (‘plas- tochondrin’), insomuch so that it suggests that the plasto- somes may exist in a direct formative participation as matrix of the -lipoid granules. And this connection is most conspicuously demonstrated in the Langhans’ cells, the stroma cells of villi, and in the decidual cells, and even in other cells it is quite easy to recognize it, because the plastosomes tend to diminish more or less in inverse proportion to the increase in the quan- tity of the lipoids. c. The vacuoles are probably nothing but the lipoids gradually liquefied and increased into what they are. And, with the rise of functions, they keep increasing in numbers and, as a higher degree of activity is attained, the vacuoles grow in size, and part of them by degrees become agglutined with one another, so that at last the cell body presents in its entirety a highly foamy image, being composed of numberless vacuoles of various sizes. The various cell groups mentioned above, if looked at from their minute structure as well as the changes in the formative components, which latter probably have an intimate connec- tion with their functions, bear a close resemblance to the ordi- nary classical glandular cells (pancreas, salivary and lacrimal SECRETORY FUNCTIONS IN HUMAN PLACENTA 557 glands) and the important internal secretory cells (luteal and interstitial cells of ovary, the cortical cells of suprarenals), and there exists no radically great difference between the two. That is to say, suppose we now take lipoid granules for secre- tory granules and vacuoles for secretions, and naturally these cell groups in placenta and decidua should come under the same category as glandular cells, and there would be no doubt what- ever that the former have certain secreting functions in themselves. 2. The secreting phenomena of placental and decidual cells, with only the exception of the large-type decidual cells, gener- ally present themselves as in the case of the ordinary glandular cells, with the changes which commonly appear in the structure of the cell bodies and almost under the same form. Now, looked at from the histological viewpoint, the secretions prob- ably rise from the ‘plastochondrin,’ and then first passing through the period of minor granules which corresponds to Heidenhain’s ‘Primargranulis,’ they gradually grow in size and form into the ordinary lipoid granules, which latter, being liquefied continuously, change directly to the secretions (vacuoles). And, in this matter, it seems that the series of histological changes ordinarily even in the same cell body take place at different times and in different regions, so that the changes make their appearance in repetition secondarily, thirdly, and so on, which fact is responsible for the intricacy of struc- ture which sometimes occurs in certain cells. 3. The methods of discharging secretions, if in the syncy- tium layer, are that the vacuoles finally rupturing themselves in several parts of the superficial layer cause their contents—se- cretions—to escape directly into the intervillous spaces in a striking manner, though in the other cell groups the secretions for the most part cannot but be recognized as passing out by ‘osmose.’ And, of all the secretions, it should be noted that those which come from the syncytium layer, decidual cells, uterine glandular cells (a part) and also probably from the Langhans’ islets are absorbed by the mother’s body, while those which pass from the ordinary Langhans’ cells and the stroma cells of villi are absorbed in the fetal side. 558 GENCHO FUJIMURA 4. As regards the relation between the secreting functions and the time of pregnancy: a. The secreting functions of the syncytium layer may be demonstrated from the beginning of pregnancy to about the end of the fourth month, and yet it is in the second and third months that they are most active. b. The secreting functions of the Langhans’cells are almost entirely the same as in the sycytium layer. It is in the Lang- hans’ islets alone that they last somewhat longer, it being pos- sible to demonstrate cells which have secretions in them up to the fifth or seventh month, and naturally it can be imagined that the functions continue up to that time. c. The secreting functions of the stroma cells of villi begin at about the end of the first month of pregnancy, and keep quite active up to about the seventh month, though from the sec- ond to the sixth month they are at their height. However, it should be noted with care that in the eighth month these ~ cells suddenly diminish remarkably and perish, in consequence of which the functions also will drop promptly at this period. d. The decidual cells are entirely different in their appearance falling in the classification into large and small types, as it is well known. That is to say, in the so-called small-type cells. the secreting conditions pretty well agree with those in the other cells. This kind of cells appears already quite active on about the seventeenth or eighteenth day after conception, and nearly at the end of the first month of pregnancy its growth and, consequently, its functions reach their climax. There- after, as the large-type decidual cells appear, the small-type cells suddenly diminish in quantity, and in consequence it ap- pears that the functions also drop quickly, though even up to the seventh month of pregnancy it is able to eee demonstrate the existence of the functions. Then, in the so-called large-type decidual cells, for the most part few, are the structures of the cell body by which the exist- ence of secreting functions may be proved; however, in its strongly developed cell membrane a certain substance is formed, probably by a peculiar faculty of its own, and in this manner SECRETORY FUNCTIONS IN HUMAN PLACENTA 559 there occurs a material formation which may be deemed a secreted matter which is excreted by the cell body. The large-type decidual cells are remarkable in their ap- pearance by the end of the first month of pregnancy, though in the second month they appear to reach their climax, and in the following third or fourth months, they diminish in their size. And, the afore-mentioned secreting phenomenon which is peculiar to these cells, begins in the second month, appears most remarkably in the third month, and may be demonstrated up to about the sixth month, though in the seventh month and after it is no longer possible to observe it. In general, the large-type cells retrograde and decay remarkably in the second half of pregnancy, though at the end of pregnancy it is still able to find them, and, moreover, at this period there are some few cells which do contain a small quantity of lipoids. The functions of the glandular epithelium are most active at the end of the first month of pregnancy, begin to drop considerably from the beginning of the third month, in the fourth month the decline is greater, and in the fifth month, it appears, they almost come to a standstill. In general, the functions make their appearance somewhat earlier in the decidua serotina than in the decidua vera, and accordingly they stop earlier in the former than in the latter. The secretions are thrown out into the uterine cavity probably only in the earliest period of pregnancy, and later as the openings of the glandular tubules are closed by the placental formation and by the adhesion of the decidua vera and decidua reflexa, the secretions, along with the detrital matters of the degenerated glandular cells are, of necessity, absorbed by the mother. 5. Since it is possible that the secretions of the various kinds of cell groups mentioned above are, for the most part, absorbed, either by the mother or by the fetus, as in the case of internal secretions, everybody will easily assent to the supposition that, like the secretions of many internal secretory glands, each of them contains a certain hormone, and should this be the case, it may be said that each of the two organs concerned is assur- edly a producer of hormones of various sorts and kinds, and is 560 GENCHO FUJIMURA also a reservoir for them; and the kinds of hormones and the pro- portion of their mixture as contents of these organs should have important bearings upon the time of pregnancy and the part of the organs concerned which is taken as material for investigation. And, in the first half of pregnancy, it is possi- ble to show quite a variety of hormones, whereas in a well- ripe placenta it is almost impossible to demonstrate their existence. 6. It has been believed by several authors that the epithe- lium of villi probably serves as an organ by which nutrition is taken to the embryo; however, histologically it is impossible to find any ground for such argument. 7. The various cells described above usually increase in size more and more as their secreting functions progress. This fact is most remarkably noticed in the decidual cells. The large-type cell is, after all, nothing but the small-type cell grown up; its growth being gradual along with the progress of its functions and with its largest size it has perfected its develop- ment. Therefore, though at a glance it would seem that these two kinds of cells are entirely different from one another, yet they have the same origin, and originally they are the cells of the same kind. However, the sharp demarcation which exists between the two functionally should deserve our atten-_ tion; that is to say, the decidual cell performs the conversion of its functions along with the perfection of its growth. 8. The foregoing conclusions would, at a glance, seem to contradict the work conducted by several authors up to now whose conclusion it was almost entirely to deny the internal secreting functions of the placenta and decidua; but the main reason for this is the fact that the materials employed for investigation by these. authors have been for the most part ma- ture organs, for in these there is almost no proof of any secreting phenomena, being existent, and they have been therefore taken at most unfavorable times as materials to help us attain our aim. Therefore, in the future, should anyone desire to try his hand in this sort of research, it would be necessary for him by all means to take materials while yet in the early SECRETORY FUNCTIONS IN HUMAN PLACENTA 561 part of pregnancy. And, even in this case secretions of several kinds, even if they come from cells of one and the same origin, would possibly be, by no means, similar in quality, but rather in the organ concerned, there would be existent various kinds of substances produced from the various cell groups whichform the organ. And, in case that there is a certain hormone action in these substances, it would follow that, during a certain period of pregnancy, the hormones will act upon both the mother and the fetus in diverse and complex manners. 9. The histological changes which the interstitial cells of the uterine mucous membrane and glandular cells undergo prior to menses resemble, in general, the changes which take place at the beginning of pregnancy, though they are by far the weaker. Therefore, even in that case, these two kinds of cells, looked at from their histological structure, have in common to them- selves, secreting functions, to whose existence we may positively assert. And, the secretions, if in the interstitial cells, are un- doubtedly absorbed internally, as in the case of the small-type decidual cells while in pregnancy, and should thereby bring about the various clinical symptoms which are experienced dur- ing menses. The glandular cells differ from the former, and the secre- tions have probably no endocrine nature and are immediately thrown out to the outside, viz., into the uterine cavity, so that it would be difficult to attach to them an important physiological signficance, such as hormone action. Rather, it would be fit to interpret such periodical changes of these cells as preliminary phenomenon of the coming pregnancy. 10. The interstitial cells of the uterus, prior to menses, are developed into the so-called menstrual decidual cells, which in point of structure, distinctly reminds us of the decidual cells of pregnancy. For this reason, it would be in order for us to trace the origin of the latter, as of the former, to the interstitial cells of the uterus. 562 GENCHO FUJIMURA In conclusion, my whole-hearted gratitude goes to Prof. R. Tsukaguchi for the kind and sincere leadership and revision which he has given me at the present work, and to Professor Ogata, chief of our gynecological department, for the valuable materials for research and for the immense assistance he has given me. July, 1920. LITERATURE CITED AHLFELD, F. 1878 Beschreibung eines sehr kleinen menschlichen Eies. Archiv f. Gyn., Bd. 13, H. 2. : ALTMANN, R. 1894 Die Elementarorganismen und ihre Beziehungen zu den Zellen. 2. Aufl., Leipzig. Asapa, Kk. 1918 The correlation between the fat contained in the uterine mucous membrane and the periodical changes of the latter. Journal of the Gynecological Association of Japan, vol. 13, no. 11 (Japanese). Ascot1 1902 Passirt Eiweiss die placentare Scheidewand? Zeitschr. f. physiol. Chemie, Bd. 36. Arntas, M. 1911 Observations cytologiques sur l’ovaire des Mammiferes. I. Les cellules interstitielles de l’ovaire chez le cobaye (foetus 4 terme et nouveau-né). Anat. Anz., Bd. 39. 1912 L’appareil mitochondrial des cellules interstitielles de l’ovaire du Murin. C. R. de la soc. de Biol., T. 73, N. 32, cited from Jahres- berichte iiber die Fortschritte der Anatomie und Entwicklungs- geschichte. AscuHeIM, S. 1915 Zur Histologie der Uterusschleimhaut. Zeitschr. f. Geb. und Gyne, Bde 17, Hes, AsScHNER, B., UND Grigoriu, Cur. 1911 Placenta, Foetus und Keimdriise in ihrer Wirkung auf die Milchsekretion. Archiv f. Gyn., Bd. 94. BaBkINn, B. P., RusascHKIn, W. T., uNpD Ssawitscn, W. W. 1909 Uber die -morphologischen Verinderungen der Pankreas-zellen unter der Wir- kung verschiedenartiger Reize. Archiv f. mikr. Anat., Bd. 74. Bascn 1912 Placenta, Foetus und Ovarium in ihrer Beziehung zur experi- mentellen Milchauslésung. Archiv f. Gyn., Bd. 96, H. 1. Bayuiss, W. M., unp Srariine, E. H. 1906 Die chemische Koordination der Funktionen des Kérpers. Ergebn. d. Phys. Beum, C. 1903 Uber Hyperemesis gravidarum mit Aufstellung einer neuen Intoxicationstheorie vom Wesen der Kranheit. Archiv f. Gyn., Bd. 69. Benpa 1903 Die Mitochondria. Ergebn. d. Anat. und Entwgesch., Bd. 12. BERNARD, CL. 1856 Mémoire sur le pancréas et le réle du sue pancréatique. Paris, after M. Heidenhain. SECRETORY FUNCTIONS IN HUMAN PLACENTA 563 BIEDERMANN, W. 1882 Uber morphologische Veriinderungen der Zungendriisen des Frosches bei Reizung der Driisennerven. Wiener Sitzungsber., quoted from Heidenhain. 1886 Zur Histologie und Physiologie der Schleimsecretion. Wien. Akad. Sitzungsber., Bd. 94, Abt. 3, after M. Heidenhain. BIEDL UND KOGnIGsTEIN 1910 Zeitschr. f. exp. Path. und Ther., Bd. 8, H. 2, after Aschner. Boseau, G. 1911 Mitochondries et lipoides dans les glandules parathyroides du cheval. C. R. de. l’Assoc. des Anat. Paris, after Duesberg. 1911 Recherches cytologiques sur les glandules parathyroides du cheval. Journal de |’Anat. et de la Phys. norm. et Path., T. 47, cited from Duesberg. Bonp1 1911 Uber das Fett inder Placenta. Archiv f. Gyn., Bd. 93. Bonnet 1899 Embryotrophe. Deutsche med. Wochenschr., N. 45. Boucuacourt 1903 C.R. Soc. Biol., Bd. 54, cited from Aschner and Grigoriu. Bruntrz, L. 1908 Sur la contingence de la bordure en brosse et la signification probable des batonnets de la cellule. C. R. Acad. des Se. de Paris, T. 147, cited from Duesberg. Caruier, W. 1896 On the pancreas of the hedgehog during hibernation. Jour- nal of Anat. and Phys., vol. 30, no. 3, cited from Heidenhain. Cuampy, Cur. 1909 Apropos des mitochondries des cellules glandulaires et des cellules renales. C. R. Soc. Biol. Paris, T. 66. N. 4, cited from Duesberg. 1911 Recherches sur l’absorption intestinale et le role des mitochon- dries dans l’absorption et la sécrétion. Archiv d’ Anat.micr., T. 13, cited from Duesberg. Coun, F. 1903 Zur Histologie und Histogenese des Corpus luteum und des interstitiellen Ovarialgewebe. Archiv f. mikr. Anat., Bd. 62. Cotte 1913 Degliosped, N. 38, from the Liepmann’s Handbuch. Cotson, R. 1911 Histogénése et structure de la capsule surrénale adulte. Archiv de Biol., T. 25, cited from Duesberg. 1911 Recherches sur l’absorption intestinale et le role des mito- chondries dans l’absorption et la secrétion. Archiv d’Anat. micr., T. 13, cited from Duesberg. Costa, R. 1904 Uber den Durchgang der Fette von der Mutter zum Foetus. Ann. di ost. e gin. Nr. 7, from the Journal of the Gynecological Asso- ciation of Japan, vol. 14, no. 3 (Japanese). Costa, A. 1908 Celestino da, 07, Nutes sur le noyau des cellules grandulaires a sécrétion interne. Bull. Soc. Portugaise Sc. Nat., vol. 1, from the Jahresberichte der Anatomie und Entwicklungsgeschichte. 1911 Notes sur le chondriome des cellules de la capsule surrenale. Bull. Soc. Portugaise Sc. Nat. T. V, from the Centralblatt fiir normale Anatomie und Mikrotechnik, Bd. 9, 1912. CRISTIA AND ASCHNER 1912 Rivista de chir. cited from Liepmann’s Handbuch. Dour, M. 1916 An experimental research on the milk-secretion (report 1). The Transactions of the Kinki-Provincial Gynecological Association, no. 2 (Japanese). DusrssperG 1911 Ergebnisse der Anatomie und Entwicklungsgeschichte, Bd. 203 He2. 564 GENCHO FUJIMURA Esner, V. v. 1873 Die acinésen Driisen der Zunge und ihre Beziehungen zu den Geschmacksorganen. Graz: Lenschner & Lubensky, cited from Heidenhain. Epstein, W., ANDGRUTZNER, P. 1874 Uber PepsinbildungimMagen. Pfliigers Archiv, Bd. 8. EncEL, E. A. 1899 Sui prosessi secretori nelle cellule delle paratiroidi dell’ uomo. 1 Taf, Internat. Monatsschr. Anat. und Physiol., Bd. 26, cited from Duesberg. FreLuner, O. 1909 Uber die Titigkeit des Ovariums in der Schwangerschaft. Archiv f. Gyn., Bd. 87. 1913 Experimentelle Untersuchungen iiber die Wirkung von Gewebs- extrakten aus der Placenta und der weiblichen Sexualorganen auf das Genitale. Archiv f. Gyn., Bd. 100. FieiscHEer, B. 1904 Beitrage zur Histologie der Tranendriise und zur Lehre von den Sekretgranula. Habilitationsschr., Anat. Heft, Abt. 1, H. 78. FLEMMING, W. 1882 Zellsubstanz, Kern und Zellteilung, from Duesberg. Foa 1910 Archiv di Fisiol., vol. 5, from Aschner. FRAENKEL, L. 1914 Quoted in Liepmann’s Handbuch der Frauenheilkunde, Bd. 3. Frank, P. T., AnD UNcGmR, A. 1911 An experimental study of the causes which produce the growth of the mammary gland. Archiv of Intern. Med., vol. 7, cited in Liepmann’s Handbuch der Frauenheilkunde, Bd. 3. Fusimura, G. 1917 A contribution to the histological studies of the uterine mucous membrane and decidua. Journal of the Tokyo Medical Association, vol. 31, no. 14 (Japanese). Fusimura, G., AND Htrosz, T. 1917 A biological study of the decidua. Jour- nal of the Gynecological Association of Japan, vol. 12, no. 1 (Japanese). 1917 Uber die Komplementbindung des der Schwangern entnom- menen Blutserums mit dem X-Placentarbestandteile. Journal of the Tokyo Medical Association, vol. 31, no. 24 (Japanese). Gaueorti, G. 1895 Uber die Granulationeninden Zellen. Internat. Monats- schr. f. Anat. und Phys., Bd. 12, after Duesberg. GentitI, A. 1917 Annali di ostetricia e di gin. Feb., from the Journal of the Gynecological Association of Japan, vol. 12, no. 5 (Japanese). 1914 Uber die innere Sekretion der Decidua in Hinblik auf die Arbeit von J. Schéttlinder. Zur Theorie der Abderhalden’schen Schwangerschaftsreactionen, u. s. w. Zentralbl. f. Gyn., Nr. 33. GENTILI, A., AND BinaGcui, R. 1917 Annali di ostetricia e di gin. July, from the Journal of the Gynecological Association of Japan, vol. 12, no. 7 (Japanese). GortscHaLk, 8. 1890 Beitrag zur Entwicklungsgeschichte der menschlichen Placenta. Archiv f. Gyn., Bd. 37. Haupan, J. 1904 Schwangerschaftsreactionen der fétalen Organe und ihre puerperale Involution. Zeitschr. f. Gyn., Bd. 53. 1905 Die innere Secretion von Ovarium und Placenta und ihre Bedeut- ung fiir die Funktion der Milchdriise. Archiv f. Gyn., Bd. 75. 1906 Uber ein bisher nicht beobachtetes Schwangerschafts-symptom (Hypertrichosis graviditatis). W.K.W., Nr. 1. SECRETORY FUNCTIONS IN HUMAN PLACENTA 565 Harapa, T. 1916 A study with regard to the chemical components of the placenta. “Report of the Kinki-Provincial Gynecological Association, no. 2 (Japanese). HeIDENHAIN, R. 1868 Beitrige zur Lehre von der Speichelabsonderung. Studien des Physiol. Inst. zu Breslau, 4. Heft, quoted from Heidenhain. 1875 Beitrige zur Kenntniss des Pancreas. Pfliigers Archiv, Bd. 10, quoted from Heidenhain. 1880 Hermann’s Handbuch der Physiologie. HeEIDENHAIN, M. 1907 Plasma und Zelle. Erste Lieferung. Fischer, Jena. Heutp, H. 1899 Beobachtungen am tierischen Protoplasma. Archiv Anat. und Phys., anat. Abt., Bd. 99, H. 5/6. Hermann, E. 1914 Uber eine wirksame Substanz im Eierstocke und in der Placenta. Monatsschr. f. Geb. und Gyn., Bd. 41, H. 1. Hieucut, 8. 1909 Ein Beitrag zur chemischen Zusammensetzung der Placenta. Biochem. Zeitschr., Bd. 22. HITzMANN UND ApLER 1907 Die Lehre von der Endometritis. Zentralb. f. Gyn., Bd. 66. 1907 Uber den normalen Bau der Uterusmucosa und ihre Entziindung. Verh. deutsch. Gynikologenkongr., Dresden. Horsaver 1905 Biologie der menschlichen Placenta. Braumiiller, Wien und Leipzig, cited in Liepmann’s Handbuch. Horstatrer, R. 1911 Pituitrin. Diskussionsbemerkung., W. K. W., No. 27. Hoven, H. 1910 Contribution 4 1’étude du fonctionnement des cellules glandu- laires. Don role du chondriome dans la Sécrétion. (Note prélimi- naire.) Anat. Anz., Bd. 37. 1911 Du réle du chondriome dans |’élaboration des produits de Séerétion de la glande mammaire. Anat. Anz., Bd. 39. 1911 Démonstration de préparations tendant 4 démontrer le réle des chondriosomes dans la formation des grains de Sécrétion. C. R. Assoc. Anat., Paris, cited by Diisberg. Kawaipa, G. 1912 An experiment in the milk-secretion. Journal of the Gynecological Association of Japan, vol. 7, no. 2 (Japanese). Kenrer 1915 Zentralb. f. Gyn., Nr. 51, from the Journal of the Gynecological Association of Japan, vol. 12, no. 9 (Japanese). Kuern, E. 1882 On the lymphatic system and the minute structure of the sali- vary glands and pancreas. Quart. Jour. of Micr. Sc., vol. 22, cited by M. Heidenhain. KossMAnn 1892 Zur Histologie de Chordonzotten des Menschen. Festch. z. 70 Geburtstag R. Leukarts, Leipzig, from Winckel’s Handbuch der Gebiirtschiilfe, Bd. 1, H.1. Krause, R. 1895 Zur Histologie der Speicheldriisen. Die Speicheldriisen des Igels. Archiv f. mikr. Anat., Bd. 45. LaacuesseE, E. 1899a Origine du zymogéne. C.R. Soc. Biol., cited by Heiden- hain. ; 1899 b Corpuscules paranucléaires (parasomes), filaments basaux et zymogéne dans les cellules sécrétantes (pancréas, sous-maxillaire). Cing. Soc. de Biol., cited by Heidenhain. 566 GENCHO FUJIMURA LacuEssE, E. 1905 Le pancréas premiére partie, la glande exocrine. Revue générale d’Histologie, cited by Heidenhain. 1911 Ergastoplasma et chondriome dans les cellules sécrét-antes sereuses. Bibl. Anat., T. 21, cited by Duesberg. LANGERHANS, P. 1869 Beitrige zur mikroskopischen Anatomie der Bauchspei- cheldriise. Inaugural-Dissertation, Berlin, cited by Heidenhain LanaHans 1892 Uber die Zellschicht des menschlichen Chorion. Festschr. f. Henle. Beitr. z. Anat. und Embryol., Bonn, quoted in. Winckel’s Handbuch. Laneey, J. N. 1879 On the changes in serous glands during secretion. Jour- nal of Physiol., vol. 2 (Proceedings of the Royal Society, no. 198), cited by Heidenhain. 1881 On the histology and physiology of pepsin-forming glands. Philosophical Transaction, Port. 3, cited by Heidenhain. 1882 On the histology of the chief-cells. Journal of Physiology, vol. 3, cited by Heidenhain. 1884 On the structure of secretory cells and on the changes which take place in them during secretion. Intern. Monatsschr. f. An. und Phys., Bd. 1, cited by Heidenhain. 1889 On the histology of the mucous salivary glands, and on the behaviour of their mucous constituents. Journal of Physiology, vol. 10, cited by Heidenhain. Lavpowsky, M. 1876 Zur feineren Anatomie und Physiologie der Speichel- driisen, insbesondere der Orbitaldriise. Archiv f. mikr. Anat., Bd. 13. LEDERER, R., UND PrziBrRaM, E. 1910 Experimenteller Beitrag zur Frage tiber die Beziehungen zwischen Placenta und Brustdriisenfunktion. P. A., Bd. 4, after Aschner. LETULLE AND LarRIER 1901 Die sekretorische Funktion der Placenta. Rev. d. Gyn., Bd. 2, after Fellner. LeuspEN, P. 1897 Uber die serotinalen Riesenzellen und ihre Beziehungen zur Regeneration der epithelialen Elemente des Uterus an den Placenta. Zeitschr. f. Gyn., Bd. 36, H. 1. Levi, G. 1912a Condriosomi nelle cellule secernenti. Anat. Anz., Bd. 42, Nr. 22/23. 1913 Note citologiche sulle cellule somatiche dell’ ovaio dei Mam- miferi. Archiv f. Zellforsch., Bd. XI. Limon, M. 1902 Etude histologique et histogénetique de la glande interstitielle de lVovaire. Archiv d’anat. micr., T. 5, fase. 2, Sept. Observation sur l’état de la glande interstitielle dans les ovaires transplantes. J.D. P. P., T. 6, p. 864, quoted from Diisberg. 1903 Etude histologique et histogenétique de la glande interstitielle de Vovaire. Archiv d’ Anat. miecr., T. 5, quoted from Diisberg. Los, W., unp Hiaucut, 8S. 1909 Zur Kenntniss der Placentaencyme. Biochem. Zeitschr., Bd. 22. Manpt 1905 Wr. klin. Wochenschr., N.3 u. 4, quoted from Aschner. Marscnanp 1898 Mikrosk. Priparate von zwei friihzeitigen menschlichen Kiern und einer Decidua. Sitz. d. Gesellsch. z. Beford. d. ges. Natur- wiss. zu Marburg, Nr. 7, out of the Jahresberichte der Anatomie und Entwicklungsgeschichte. SECRETORY FUNCTIONS IN HUMAN PLACENTA 567 Matuews, C. 1880 On the chemistry of the cytological staining. Journal of ‘Physiolog., vol. 1, cited from Heidenhain. Mawas, J. 1911 Sur la structure du protoplasme des cellules epitheliales du corps thyroide de quelques Mammiferes. Se chondriosome et les phénoménes de sécrétion. Bibl. anat., T. 21, cited from Duesberg. Maxtmow, A. 1901 Beitrige zur Histologie und Physiologie der Speicheldriisen. Archiv mikr. Anat., Bd. 58, H. 1. Meves, Fr. 1907 Die Chondriokonten in ihrem Verhitnis zur Filarmasse Flemmings. Anat. Anz., Bd. 31. 1908 Die Chondriosomen als Triger erblicher Anlagen. Cytologische Studien am Hithnerembryo. Archiv f. mik. Anat., Bd. 72. Mistawsky, A. N. 1911 Beitrige zur Morphologie der Driisenzelle. Uber das Chondriom der Pankreaszelle einiger Nager. Vorl. Mitteil. Anat. Anz., Bd. 39, N. 19/20. Miuter, L. R. 1896 Beitrige zur Histologie der normalen und erkrankten Schilddriise. Ziegler’s Beitriige zur path. An. u. zur allg. Path., Bd. 19. Mixture, Erik. 1898 Driisenstudien II. Zeitschr. f. wiss. Zool., Bd. 64, quoted by Heidenhain. Muvton, P. 1909 Etudes sur l’ovaire du cobaye: Sur un corps jaune kystique formé aux dépens d’un ovisaec non déhisce. Archiv d’anat. micr., T. 11, no. 1, quoted from Diisberg. 1910 a La méthode des mitochondries (de Benda) appliquée a la corticale surrénale du cobaye. C.R. Soc. Biol., quoted from Diisberg. 1910 b Sur les mitochondries de la surrénale (substance corticale, conche graisseuse, cobaye). C. R. Soc. Biol. quoted from Diisberg. 1911 Démonstration de préparations dela corticale surrénale (lapin), du corps jaune (brebis) et du la glande interstitielle de l’ovaire (lapin). C. R. Assoc. Anat. Paris, quoted from Diisberg. 1911 Un processus de Sécrétion interne dans la corticale surrénale. C. R. Soe. Biol. Paris, T. 70, n. 15, quoted from Diisberg. 1911 Mitochondries des cellules de la corticale surrénale du lapin, du corps jaune de la brebis et de la glande interstitielle de l’ovaire de la lapin. Assoc. Anat. Congr. Paris, quoted from Diisberg. Nicoaiu, P. 1893 Uber die Hautdriisen der Amphibien. Zeitschr. f. wiss. Zool., Bd. 56, H. 3, cited by Heidenhain. Nicouas, A. 1892 Contribution a l’étude de cellules glandulaires. Il. Le protoplasma des éléments des glandes albumineuses (lacrymale et parotide). Archiv de Physiolog. norm. et Pathol. 5° série, T. 4, année 24, cited by Heidenhain. Nout, A. 1901 Morphol. Verinderungen der Trinendriisen bei der Sekretion. Archiv f. mikr. Anat., Bd. 58 Nussspaum, M. 1877-1882 Uber den Bau und die Titigkeit der Drisen. Archiv f. mikr. Anat., Bd. 13, 15, 16 u. 21. Orr, I., anp Scort, J. C. 1910 The galactogogue action of the thymus and corpus luteum. Proc. Soc. for Exp. Brol. and Med., vol. 7, cited from Bied1’s Innere Secretion, Bd. 13. 568 GENCHO FUJIMURA Ort, I., anp Scott, J. C. 1911 Note on the galactogogue action of the thy- mus, corpus luteum and the pineal body. Monthly Cyclop., vol.25, Feb., cited from Biedl’s Innere Secretion, Bd. 13. 1912 Die Wirkung des Corpus luteum und der Glandula pinealis. Monthly Cyclopedia and Med. Bull., cited from Fellner. Oxrtntscuitz 1914 Uber die gegenseitigen Beziehungen einiger Driisen mit innerer Secretion. Archiv f. Gyn., Bd. 102. Pretiicer, E. 1871 Zum Nachweis der Nervenendigunge in der acinésen Driisen und der Leber. Archiv f. Physiologie. PRENANT, A. 1898 From Prenant, Bouin u. Mallard’s Traité d’Histologie, Te: 2,p. Mili REGAUD, CL., ET PotiKARD, A. 1901 Notes histologiquessurl’ovaire des Mam- miféres. C.R. de 1]’Association des Anat. Lyon, cited from Levi. ReGaupD ET Dusrevit 1906 Recherches sur les cellules interstitielles de l’ovaire chez le Lapin. Bibl. anat., T. 26, cited from Diisberg. Reaaup, Cu. 1908 Sur les mitochondries des cellules ciliées du tube urinaire. Ont-elles une relation avec la fonction motrice de ces cellules. C. R. Soe. Biol., cited from Diisberg. 1909 Attribution auz ‘formations mitochondriales’ de la fonction générale d’extraction et de fixation électives, exercée par les cellules vivantes sur les substances dissontes dans le milieu ambiant. C. R. Soc. Biol., cited from Diisberg. ReGaup, Cu., eT Mawas, J. 1909 Sur les mitochondries des glandes salivaires chez les Mammiféres. C.R. Soc. Biol., cited from Diisberg. 1909 Ergastoplasme et mitochondries dans les cellules de la glande sousmaxillaire de ’homme. C. R. Soc. Biol., cited from Diisberg. 1909 Sur la structure du protoplasme (ergastoplasme, mitochondries, grains de sérégation) dans les cellules séro-zymogénes des acini et dans les cellules des canaux excréteurs de quelques glandes salivaires des Mammiféres. C.R. Assoc. Anat. Nancy, cited from Diisberg. ScuarFrer, E. A., AnD Mackenzin, K. 1911 The action of animal extracts on milk secretion. Proc. Roy. Soc., vol. 84, ser. B, cited from Biedl’s Innere Secretion, Bd. 15. ScHICKELE, G. 1912 Zur Lehre von der inneren Secretion der Placenta. Bio- chem. Zeitschr., Bd. 38, H. 3 u. 4. ScHOTTLANDER, J. 1914 Zur Theorie der Abderhalden’schen Schwanger- schaftsreaction, sowie Anmerkungen iiber die innere Sekretion des weiblichen Genitales. Zentralb. f. Gyn., Nr. 12. Scuuttzp, Fr. E. 1864 Quoted by Heidenhain. 1867 Epithelund Driisenzelen. Archiv f. mik. Anat., Bd. 3. Scuuttze O. 1911 Uber die Genese der Granula in den Driisenzellen. Anat. Anz., Bd. 38, N. 10/11. ScuowaLBy,G. 1871 Beitrige zur Kenntnis der Driisen in den Darmwandungen, ete. Archiv f. mikr. Anat., Bd. 8. Soteer, B. 1896 Uberden feineren Bau der Gland. submaxillaris des Menschen mit besonderer Beriicksichtigung der Driisengranula. Festschrift fir C. Gegenbauer. SECRETORY FUNCTIONS IN HUMAN PLACENTA 569 Srricut,O.vanpER 1912 Surle processus del’exerétion des glandes endocrines: Le corps jaune et la glande interstitielle de Vovaire. Archiv Biol., T. 27, N. 4, cited from Levi and from Centrbl. norm. Anat. Taniaucui, Y. 1916 A biological study of the ovarian functions. The Chinzei- Provincial Med. Journal, no. 167 (Japanese). TsuKaaucuI, R. 1912 The 20th Report of the Anatomie Association, Okayama (Japanese). 1912 The trend of the granules-researches in recent times. Journal of the Osaka Medical Association, vol. 12, no. 3 (Japanese). Verr 1902 Uber Albuminurie in der Schwangerschaft. Ber. klin. Wochenschr., Nro22ue233 WINcKEL, V. F. 1903 Handbuch der Geburtshilfe, Bd. 1, H. 1. Wo.rr, A. 1913 Oxydasenreaktioninder Placenta. Monatsschr. f. Gyn., Bd. 37. ZIMMERMANN, K. W. 1898 Beitrige zur Kenntnis einiger Driisen und Hpi- thelien. Archiv f. mikr. Anat., Bd. 52. Zoya, L.& R. 1891 Interno ai plastiduli fuesinofili (bioblasti dell’ Altmann). Mem. del R. Ist. Lomb. de sci. e lett., vol. 12, cl delle Se. mat. e nat., cited from Heidenhain. 1891 Uber die fuchsinophilen Plastidulen. Archiv f. Anat. u. Phys., Anat. Abt. gi 4 L y Gok ; iy », 3 ! : NV : 4 if va, uF) i a 7 ‘ ; i 7 : Fi Bt 4 i lees | ' : y \ i “ a ian colnet ane rw Wiha Vaginal hte iad rt a ‘Le aa @ = i hs Adee ke Tes ohh pat! po i “6 Jas ee iatt, ; : f f ‘ . f ee a iH hah ph Ole NOR We ) w) ae =e, * F oe af 7 : . eo é Po yt @ ahrcs on ; oa: ‘AM eke: f 1 i a } ih pote _ ca@ G4 , ; (reo ihe a ae 7 ; ; , ™ ON 4 penny a Ry 7 4 ar aN f Fi — i i : oe . - . _ , te : AT -_ ; ee it 5 . 2 Ab : - we LU . } ,' &. k . iyo * > i i . 4 - ® , i J ' * Ww ‘ * ‘ ‘ j y ; 7 7 ‘ - , i ‘ ' ‘ = i ‘ ‘ i . » oll a : 1 at F i e F a a it : a ' D BE S Big Q r r § A - * ae EXPLANATION OF PLATES All the figures given on plates have been drawn at the height of the object stage, by Abbe’s apparatus, under the same magnifying power: Zeiss’ apochromat homogene immersion 3 mm., compensations okular. 12, tube length 160 mm. The various figures have all been drawn from the preparations fixed by Levi’s solution and stained by Heidenhain’s iron-alum-haematoxylin, with only the exception of figure 7, which has been reproduced from the preparations by Alt- mann’s method, after changing the color. PLATE 1 EXPLANATION OF FIGURES . 1toi2 Illustrate the syncytium layer and a part of the Langhans’ cells. SECRETORY FUNCTIONS IN HUMAN PLACENTA GENCHO FUJIMURA — PLATE 1 PRESS WORK ty PRED'K GOEN ¥ PEATE 2s", as aed hy . 1 EXPLANATION OF FIGURES bg 13 to26 Illustrate the Langhans’ cells. m 27 to38 The stroma cells in the chorion villi. es 39 to 69 The decidual cells. ie: 70 The peculiar-looking product which makes its appearance ‘* the membrane and interstitium of the large-type decidual cells. 71 The above-mentioned product filling up the blood vessels. 72to83 The glandular epithelium cee 84t091 The interstitial cells prior to menses. 92 to 96 The glandular epithelial cells Brie yr to menses oy v SECRETORY FUNCTIONS IN HUMAN PLACENTA GENCHO FUJIMURA PLATE 2 Resumen por el autor, Albert M. Reese. La estructura y desarrollo de las glindulas tegumentarias de los Crocodilia. El presente trabajo versa principalmente sobre Alligator mississipiensis, suplementado por algunas observaciones Ileva- das a cabo sobre el caiman, Caiman sp. de la Guyana Inglesa. El autor discute el desarrollo y la estructura adulta de tres series de glindulas, a saber: Las pequefias glindulas dorsales que constituyen dos filas debajo de ciertas escamas dorsales desde la mitad de la regién cervical, proximamente, hasta la regién de la cloaca; las glandulas mandibulares, las cuales poseen un ori- ficio bastante grande situado ventralmente en cada lado de la mandibula; y las gl4ndulas cloacales, un par de grandes cuer- pos ovales que se abren en la cloaca. Las glandulas man- dibulares y cloacales sirven para la secrecién de una substancia azmizclosa y probablemente son més activas durante la época de la cria. La funcién de las numerosas gldindulas dorsales, mas pequenas, es problematica. Translation by José F. Nonidez Cornell Medical College, New York AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, SEPTEMBER 26 THE STRUCTURE AND DEVELOPMENT OF THE INTEGUMENTAL GLANDS OF THE CROCODILIA A. M. REESE Zoological Laboratory of the University of West Virginia SIX PLATES (FORTY-THREE FIGURES) MATERIAL The greater part of the material upon which the embryological portion of this paper was done was collected by the author some years ago in Georgia and Florida, under a research grant from the Smithsonian Institution; several other papers have al- ready been published with this material as the foundation. The caiman material here used was collected by the author in British Guiana during the summer of 1919, under a grant from the Carnegie Institution of Washington; the species of caiman that laid the eggs could not be determined. ) The embryos were fixed in various ways (in sublimate-acetic more than in any other fluid), were stained mostly with borax ecarmine and Lyons blue, and were cut transversely, sagittally, and frontally. The structures to be described are the dorsal glands and the musk glands. THE DORSAL GLANDS The dorsal glands of the alligator extend from the cloacal region to about the midcervical region. As seen in figure 1, they lie beneath the anteromesial corner of the second row of scales from the middorsal line. When a piece of the skin of a young animal, preserved in fluid, is viewed by transmitted light, the glands appear in the form of circular areas, as seen in figure 1, which represents two rows of dorsal scales from the back of a 200-mm. alligator, in the region of the pelvic legs. 581 JOURNAL OF MORPHOLOGY, VOL. 35, NO. 3 582 A. M. REESE The structure of the adult glands will be taken up after a discussion of their development. The gland was first seen in an embryo of about 25-mm. crown- rump length (fig. 2) as a rounded thickening and invagination of the lower layer, stratum germinativum (fig. 3 and fig. 4, s). As may be seen in figure 3, the gland lies somewhat above the level of the dorsal side of the spinal cord, sc, and is relatively of large size. It extends nearly through the dermis or corium, c, to the outer edge of the underlying muscle, m. Under moderately high magnification the epidermis, at this stage, is seen to consist of two layers, a very thin, outer layer or periderm (Krause) (fig. 4, ) and a much thicker, inner layer or stratum germinativum, s. Beneath the latter lies the very thick and fairly compact and homogeneous layer, the corium, c. The outer layer, p, exaggerated in distinctness in figure 4, is very thin and contains flattened nuclei. The inner layer, s, is of somewhat variable thickness and consists of one or two layers of polyhedral or cuboidal cells; the nuclei of these cells are oval or spherical and frequently lie at the base of the cells. The gland, g, at this stage consists, as has been said, of a thickened invagination of the stratum germinativum. This thickening consists (fig. 4, g) of a mass of indistinctly outlined polyhedral cells, with large, spherical or oval nuclei. At the fundus or inner end of the gland the cells are more closely compacted than in the region nearer the surface. A wide invag- ination, d, of both layers of the epidermis represents the position of the future wide-mouthed duct of the embryonic gland. The next stage represented in figure. 5, is taken from an em- bryo of a total length of a little over 7 em. The epidermis and corium are here about as in the preceding stage, but the gland, g, is much more elongated and now projects into the underlying muscular tissue, m, as a bottle-shaped mass of cells, connected with the superficial epidermis by a rather narrow neck. The open invagination of the epidermis, d, is here deeper and narrower than in the preceding stage. The cells of the gland have about the same appearance as before, but they are distinctly less compactly arranged in the INTEGUMENTAL GLANDS OF CROCODILIA 583 center of the gland than nearer the periphery, thus foreshadowing the condition seen in the mature gland. The cells are small, compared to the size of their nuclei, and of indistinct outlines. In pushing down into the underlying muscular tissue the gland has carried with it a thin surrounding layer of the corium as a sort of connective-tissue sheath or capsule, cp. The next embryo studied (fig. 6) had a total length of about 13 cm. The gland here (fig. 7) is distinctly flask-shaped, the neck of the flask forming the wide duct, d, which is plugged with a fairly compact mass of cells. The more bulging side of the fundus is toward the median plane of the animal, to the right in figure 7. The fundus of the gland, g, projects well down into the underlying muscular tissue, m, and is surrounded by a thick, though loose, connective-tissue capsule, cp. The epidermis, ep, contains numerous pigment granules that mask somewhat the cell details. It extends down into the duct to about the region where the enlargement to form the gland proper begins; here it gradually becomes thinner and finally merges into the cell mass of the gland. As seen in figures 7 and 8 there are irregular spaces among the cells of both gland proper and duct. The cell details at this stage are very difficult to determine with accuracy. Figure 8 represents a section through the fundus of the gland, six sections anterior to the one shown in figure 7. The gland, which is now beginning to approach the adult structure, consists of a mass of irregular, often elongated, cells, with oval nuclei. Around the periphery is a more or less distinct row of nuclei, belonging to a sort of layer of basilar cells, to be noted in a later stage. In the irregular central cavity or lumen may usually be seen a few scattered cells; whether these are artificially torn off or are regularly desquamated from the under- lying cells itis not easy to determine. The lumen, lu, if it may be so called, being so irregular in outline, it is not possible to say just how many layers of cells make up the wall of the gland. In some places the lumen extends to within one or two cells of the periphery; in other regions there are six or more irregular layers of cells. Further discussion of these cells will be 584 A. M. REESE deferred to later stages; they are evidently not yet functioning, and would not be expected to function this long before hatching. The dorsal gland of a 15-em. embryo is shown in figure 9, as drawn with a camera under medium magnification. The epidermis, ep, here exhibits two well-defined areas; the above- mentioned periderm, which now shows clearly, under high powers, as a fibrillar or scaly structure, and the stratum germi- nativum, s, from which the gland has been developed, as described above. In the stratum germinativum, as as well as in the immediately subjacent corium, c, may be seen numerous brownish pigment bodies of irregular shape and size. In the region of the duct, d, which lies in the longitudinal groove between two scales, the corium is considerably thickened. In the particular section here shown the opening of the gland to the surface, is not shown and in fact, this gland did not show any opening at all, the periderm being uninterrupted. In other series a break in the periderm was visible and an irregular opening through the thickened area of the stratum germinativum could be made out. ‘There is, however, quite a sharp line of demarkation between the closely packed mass of cells of the stratum germinativum, that forms the apparent duct of the gland, and the more loosely arranged, less deeply staining cells of the gland proper. While the gland proper is now several times larger than it was at the preceding stage, the duct is relatively and actually considerably narrower and shorter. The gland, as may be seen in figures 1 and 9, is circular in out- line and is flattened until its dorsoventral thickness is about one-half its diameter. It lies almost entirely below the corium, c, and is surrounded by a thin connective-tissue capsule, cp. It consists of a fairly compact peripheral mass of cells, which become looser and more scattered towards the center (fig. 9). These cells are finely granular and have such indistinct walls that their boundaries can only with great difficulty be determined; they contain oval or round nuclei. Around the periphery of the gland, next to what may be called the basement membrane, is a fairly distinct row of nuclei, representing an indistinct layer of basilar cells, (figs. 9 and 10, 6); passing from these cells towards INTEGUMENTAL GLANDS OF CROCODILIA 585 the center or lumen of the gland, lu, the nuclei become more and more scattered and the cell walls less distinct, until the cells are quite indistinguishable and appear as an irregular, granular mass with an occasional nucleus, nu. Figure 11 represents a portion of the wall of the dorsal gland of a l-meter alligator, as seen under a magnification of 270 diameters. The glands, which are here about 3 mm. in diameter, were dissected from the skin, and hence do not show the duct. The structure is quite similar to that of the stink gland of the turtle, Terrapene odorata, as figured by Dahlgren and Kepner (708). The capsule, cp, consists of a loose layer of elastic fibers, ct, with numerous nuclei, surrounded by a muscular layer, m, of varying thickness, which is made up of two more or less distinct layers of involuntary fibers, running in different directions. In the gland proper is seen a great mass of irregular and spheri- eal cells, while the intercellular spaces and lumen (if it may be so called) are filled with the granular secretion. The peripheral cells, just beneath the capsule, form a narrow but usually fairly distinct layer, the basilar cells, 6, mentioned in the preceding stage. The basilar cells are small and clear, their large oval or spherical nuclei filling a large part of the cells. It is probably by the division of these cells that the other cells, with their contained secretion, are produced. ‘There is never more than one layer of basilar cells, and even in this layer the cells are rather scattered, though always close to the capsule. Next to the basal cells is a generally fairly distinct layer of much larger cells, gs, gs’, presumably those last formed by division of the basal cells. In some regions this second layer consists of swollen spherical cells, gs, with granular contents and peripherally located nuclei; in other places these cells are distinctly columnar or cuboidal in shape, gs’, due, possibly, merely to crowding. In these cells may occasionally be seen a spherical clear area, probably an oil droplet, od. The main body of the gland is made of a scattered mass of cells, secretion, degenerate nuclei, ete. Some of these cells, 0, seem 586 A. M. REESE to be filled with a fat or oil; they are clear and spherical, with a much flattened nucleus just beneath the cell membrane. A majority of the cells, 0’, while spherical in shape, still contain a greater or less amount of granular protoplasm, in which the nucleus lies. While an occasional cell, gs’’, may be seen in which the entire content is granular; most of these are indistinct in outline and of moderate size. A number of irregular cells, with little or no contents may be seen, cs, which have the appearance of being empty cell membranes; these may be the remains of cells from which the granular or the fatty secretion has been emptied. Numerous nuclei, nw, may be seen scattered among the cells; since they are not shrunken, but are usually well formed, they may be those that have been extruded in the breaking down of the granular rather than of the oil cells. What the function of these dorsal glands may be it is diffi- cult to surmise. No odor was detected in connection with them such as evident with the submandibular and cloacal musk glands. Their small size and wide distribution over the dorsum of the animal might indicate that they are of use in keeping the scales in good condition, making them comparable to the oil glands in the skin of mammals. Possibly in the living animal the dorsal glands may have an odor and may function as accessory musk glands, though, as as stated above, no odor has ever been noticed by the present writer. THE MUSK GLANDS According to Gadow (01): All the recent crocodilia possess two pairs of skin-glands, both secreting musk. One pair is situated on the throat, on the inner side of the right and left half of the lower jaw. The opening of the gland, visible from below ...., is slit-like, and leads into a pocket, which in large speci- mens is the size of a walnut; the bag is filled with a smeary pale brown- ish substance, a concentrated essence of musk, much prized by natives. The secretion is most active during the rutting time, when the glands are partly everted. My young Crocodiles and Alligators often turned them inside-out, like the finger of a glove, when they were taken up and held by force. The other pair lies within the lips of the cloacal slit, and is not visible from the outside. The use of these strongly scen ted organs, INTEGUMENTAL GLANDS OF CROCODILIA 587 which are possessed by both sexes, is obviously hedonic. The sexes are probably able to follow and find each other, thanks to the streak of scented water left behind each individual. CLOACAL MUSK GLANDS As in the skunk andother animals, the alligator possesses a pair of well-developed glands, one on each side of the cloaca, into which they open. These glands, like the dorsal glands, are developed from the lower layer of the epidermis, and are first seen in embryos of slightly more than 7 cm. total length (fig. 12). In these embryos the penis (fig. 12, pe) is a large, thick organ projecting markedly from the cloaca to the exterior. As seen in figures 12 and 13, the glands are thickenings and slight invaginations, cg, of the epidermis in the lateral walls of the cloaca, cl, rather nearer the surface than the bottom of the cloaca; they are at some distance posterior ‘to the opening of the rectum into cloaca. The epi- dermis, ep, is somewhat thicker in the cloaca than over the general body surface, consisting in the former region of two or three layers of cells instead of the single layer, beneath the periderm, seen over the body. The gland now consists of the above mentioned thickening of the lower layer of epidermis (fig. 18, cg) which is slightly invaginated and consists of about six or eight layers of cells, somewhat more loosely arranged near the surface than around the periphery or deeper part of the thickening. ‘The periderm, p, in some sections seemed interrupted at the point of invagination, but it is seen to extend to the very bottom of the duct in later stages, this interruption was probably an artifact or an optical illusion. Opposite this point of origin of the gland the penis, pe, showed a slight thickening or bulge of its wall. Figure 14 shows the cloacal region at a slightly later stage of development. The gland, cg, is somewhat more elongated in shape and exhibits a lumen of considerable depth, slightly bi- furcated at its bottom. The relation of parts will be understood by comparing this figure with figure 12. This embryo was about 9 em. long, from tip of snout to end of tail, measured along the greater curvature of the body. The next stage is an embryo of ~ 588 A. M. REESE about 15 em. total length. The cloacal glands (fig. 15, cg) are considerably more developed than in the preceding stage; the gland on the right. side is here seen cut through its median region; the other through the opening of the duct, d, into the cloaca, cl. The body of the gland is elliptical in shape, being somewhat greater along an anteroposterior diameter than along the lateral or the dorsoventral diameters.