Batre yi = eo, - Sve are | Ae OSE 0 a ee el Oe a ee hp ee Pee 4 et oe cai r= aed ee ee 7 - i 7 = a JOURNAL OF MORPHOLOGY FOUNDED BY C. O. WHITMAN EDITED BY ig Sia LESION GE; Soo ye University of Illinois Urbana, IIl. WITH THE COLLABORATION OF Gary N. CALKINS EpwINn G. CONKLIN C. E. McCiune Columbia University Princeton University University of Pennsylvania W. M. WHEELER WILLIAM PATTEN Bussey Institution, Harvard University Dartmouth College VOLUME 26 1915 THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY PHILADELPHIA WAVERLY P By tHe Wituams & Wie ; Battrmore, Mp., \ CONTENTS 1915 No. 1. MARCH Pauutine H. DepErER. Oogenesis in Philosamia cynthia. Sixty-four figures PMT Pn Sie) rae oe i NRO ee eg aid VsreD o = 5 oss a Co, ETE poe Wituiam H. F. Appison anp J. L. Appieton, Jr. The structure and growth of the incisor teeth of the albino rat. Twenty-nine figures............... James G. Hucues, Jr. A peculiar structure in the electroplax of the star- Razer, Astroscopus guitatus. Vhree figures. ........ 051 o. que. ee alle W. Rees Bremner Rospertson. Chromosome studies. III. Inequalities and deficiencies in homologous chromosomes: their bearing upon synapsis and the loss of unit characters. Fourteen figures (three plates)............... No. 2. JUNE Grorce W. TannrEuTHER. The embryology of Bdellodrilus philadelphicus. went y-clxiemntioureseandseight plates........... cae. sete ane eee KE. E. Just. The morphology of normal fertilization in Platynereis megalops PT 9 2a) EA Posi Cs ON rom TEV 2) Sen a cD ORR IE we hice Boia Wn. A. Kepner AND J. R. Casu. Ciliated pits of Stenostoma. Tour figures G. Cart Huser. The development of the albino rat, Mus norvegicus albinus. I. From the pronuclear stage to the stage of mesoderm anlage; end of the first to the end of the ninth day. Thirty-two figures................... G. Cart Huser. The development of the albino rat, Mus norvegicus albinus. II. Abnormal ova: end of the first to the end of the ninth day. Ten figures. 3! MARIANNA VAN HERWERDEN. Comment on Miss Beckwith’s paper on ‘‘The genesis of the plasma-structure in Hydractinia echinata’’ and reply by LAINE. Sree SG pe eater ea i A a LY as PN ee Sd No. 3. SEPTEMBER EK. A. BAumaarTNER. The development of the hypophysis in Squalus Acan- cee OEE VEE IR PTGR 2 et 8.. fa). .d,s.0 0c s > sashes DR EE J. Frank Dantet. The anatomy of Heterodontus francisci. II. The endo- ekeletan. “Dhirsy-one fieures (eight plates) ........cud. 700. 34 Metaphase of second oocyte division. X 700. 35 Copulation of pronuclei; same egg as figure 27. x 700. 36 Post-oogonial nuclei with chromatin masses. The cells are connected by protoplasmic strands or tubes. 37 Longitudinal section through portion of an egg string of a larval ovary. x 400. a, Stage a, oogonial region; b, Stage e, synizesis; c, Stage f, post-synap- tic spireme; d, Stage h, dyad or tetrad chromosomes in nurse cells; e, eggs and nurse cells well differentiated. 38 and 39 Stage c; the chromatin masses are transformed into small irregular fragments which later assume a thread-like form. 40 Stage d; presynaptic leptotene. 41 and 42 Two groups of protoplasmic tubes with branches. Dial : OOGENESIS IN PHILOSAMIA PLATE 4 . PAULINE H. DEDERER 37 PLATE 5 EXPLANATION OF FIGURES 43 Stage e; synizesis. 44 to 47 Stage f; post-synaptic spireme of nurse cells. 48 and 49 Stage g; diplotene stage; 13 split rods. 50 to 52 Stage h; chromosomes begin to fragment. 53 and 54 Stage]; young nurse cells, showing tubes entering egg cell. 700. 55 A, nucleus of nurse cell with plasmosomes and chromatin granules; X 700. B, plasmosome enlarged. 56 Older nurse cells, with tubes. X 400. ' / PLATE 5 39 OOGENESIS IN PHILOSAMIA PAULINE H. DEDERER PLATE 6 EXPLANATION OF FIGURES 57 Portion of an egg string, showing an egg and two nurse cells. X 400. 58 and 59 Stage f; early and later post-synaptic stages of egg cell; spireme probably continuous. 60 Nucleus from an egg cell similar to the one shown in figure 57. 700. 61 Stage g; spireme disappearing in the later-growth period; plasmosomes of varying size and form. 62 Sheghtly older nucleus; all traces of the spireme have disappeared. 63 Same. X 850. ‘ 64 Nucleus from a nearly mature egg, showing dark crescentic region. The nuclear cavity is filled with granules. X 400. 40 PLATE 6 OOGENESIS IN PHILOSAMIA PAULINE H. DEDERER 41 rr - =a re af a) THE STRUCTURE AND GROWTH OF THE INCISOR TEETH OF THE ALBINO RAT WILLIAM H. F. ADDISON AND J. L. APPLETON, JR. From the Anatomical Laboratory of the University of Pennsylvania and The Wistar Institute of Anatomy, Philadelphia TWENTY-NINE FIGURES CONTENTS LUTE OCCT, Sots Sc ese, ated Gs eee na or em RNs ee dGP eC Oye 43 JISROTEO A! SUPE? 5 As SOk Seas lhe de touse eee a eer ner Pe EOE Ac UE) oe Se 44 Materielman cennne (MOOS reer tert ns occ cls ds Sisis via Se 0ee kh wns 2 oe ee ete 46 PTET OE OMNG ATTA 52 = FS ase ew le 4 vice vos we be oe oe ome eee ale aereenss OMEN 46 MimTferdescuipilonsoluhe IMCISOnS:......4.-+.. 0 s0- +e oe ete ose eee nee 52 Microscopic. structureor enamel and dentine........... 0.002. s0..0e+-ven deed 5d ewe a pine muna he INGISOLS 9.5 oe os 5c noes is oes gs oe + BE oe eo) SOE 59 Detailed description of development up to time of eruption................. 61 JE PUTS URUD) aT CHE "LrLOV ENS Hers 11M trl ee eR 8 Pn Aa nS 7 Simneecnnrapex Ol TOObM OY USEL 2. {0.05.06 0 bose aoa bes 24s aden So ee 81 Description of mature tooth and tooth-forming organs, in 5-month animal.. 83 RarerOMorowuhrotsInelsor Geeta. apace. coc. O: 4a esis. oe nea. en ee 88 Pe ROR Nir OL SHACISOTMEE UIs tae 8.3% le. aceelascic s wave sed )S aa, te 0 Wend Pee Oe ee 89 WSS CINTA DT eae env NE Bars Ss yeah esate ecb oY baud o.a aS Shes co RRA SS ASR RO 91 VASE Te euteUM CMC Ibe CMS sip ie cts, cic oSnsen 4, ciGngnsl-s aye e acct stes Re opera om ee 95 INTRODUCTION The incisor teeth of the Rodentia have long been regarded by the zoologist as having a high value for the understand- ing of many of the characteristics of this order. For instance, in 1888, Cope wrote “nearly all the peculiarities of the rodent dental system and manner of mastication are the mechanical consequences of an increase in length of the incisor teeth.” Tullberg (9899) gives the taxonomic position of the genus Mus, proceeding from the more general to the more specific group- ing: Rodentia, Simplicidentati, Sciurognathi, Myomorphi, My- oidei, Muriformes, Myodontes, Muridae, Murini, Mus. ) M =. There is only one set of teeth, and hence the dentition is mono- phyodont. The time of eruption of the various teeth extends over a period of 33 weeks. The incisors are the first to appear, viz., at 8 to 10 days after birth. The first and second molars erupt at about the 19th and 21st days respectively, and it is after this period that the young animals may be weaned and are able to maintain an independent existence, as far as food is concerned. The third molars are delayed until 2 weeks later and do not appear until about the 35th day. STRUCTURE AND GROWTH OF INCISOR TEETH AT “f iy he AN \ FR Oe OS A eg Fig. 1 Cranium of a 5-month albino rat. X 2. Fig. 2 Cranium of a 5-month albino rat, with the bony alveoli dissected away to show the entire length of the incisor teeth. X 2. The incisors are permanently-growing (or rootless) teeth, while the molars have a definite limited period of development and acquire roots. A wide diastema separates the incisors from the molars as may be seen by reference to figure 1. The incisors are strongly curved and Owen (’40-’45) has described the lower incisor as being the smaller segment of a larger circle, and the upper incisor as the larger segment of a smaller circle. In the lower incisor of the albino rat this statement needs a slight modification. For while the curvature of the upper in- 48 W. H.’F. ADDISON AND J. L. APPLETON, JR. cisor is in one plane only, the lower incisor is a portion of a flat- tened spiral, possessing a curve in three planes. The upper incisor is a segment of a true circle (at 5 months about 210°) and in cases of overgrowth it has often been known to complete the circle. In the case of the lower incisor, however, when we project it on the sagittal, frontal or coronal planes, it gives in each case a curve. It was the very evident curved projec- tion seen on the sagittal plane to which Owen referred. Con- sidering only this view, the lower incisor of a 5-month animal forms a segment of about four-fifths of a semicircle (140-145°). TABLE 1 23 41 10 15 5 5 10 DAYS | DAYS WEEKS WEEKS MONTHS MONTHS MONTHS mm. mm. mm. mm. mm. mim. mm. Naso-occipital length.............. 297 | 32.5 | 39 40 43 44 46.5 Do) LAS Galea sale Gy elope ~J = — = Ts o Inierzy gomatiey. 2.0.0 .c/2-) see ilse Upper diastema. . eel alerts Oo sone LO A ZED aie Upper Ee ee it iene PR oe IRS LS 1823 | 2023) | 23.35) 237 | 2022 Upper incisor—extra-alveolar lenothy 2.0: ere anche eae Byell || Bese ip “7 8.4: 8.7 | 9 93 Lower diastema. . i rE et On| D0) |) 16 Oat || 7 6.8 Lower saeiao “tata ence Ee aoe 18.1 | 20.7 | 25.5 | 26.4.4 29.41.29 9iotes Lower incisor—extra-alveolar lento thie fava is. cite Sask eee aera @5).| - 7 WO)ecaye |) TEL aah ete) ee 12.4 Measurements of the incisors and skulls of animals of differ- ent ages, were made as shown in table 1. The teeth were measured along their convex surfaces by means of silk thread wet with water, and applied to the object to be measured. The thread was then cut with scissors at the end of the object, straightened on paper and measured to tenths of millimeters. A consideration of table 1, shows in a definite way the pecul- iarities characteristic of the dentition, not only of the rat but of rodents in general. As is well known, these are the great development of the incisors, the wide diastema, and the con- sequent posterior position of the molar teeth as related to the rest of the skull. Cope (’88) wrote that he considered “the STRUCTURE AND GROWTH OF INCISOR TEETH 49 increase in the length of these teeth has been due to their con- tinued use, as believed by Ryder.” The effects of this increased elongation upon surrounding parts he described under several different headings, but reference will be made here only to one, viz., upon the shape of the glenoid cavity. “A peculiarity of the masticating apparatus is the lack of a postglenoid process, and the consequent freedom of the lower jaw to slide backward and forward in mastication. Appropriately to this motion, the condyle of the mandible is extended antero-posteriorly and the glenoid cavity is a longitudinal instead of a transverse groove.” Fig. 3 Thimble-shaped portion of the maxilla bone, in which the basal end of the upper incisor is located. X 2. The lower incisors are longer and more slender than the upper and extend far back in the mandible, beneath the lower molars, to near the sigmoid notch. The upper incisors are contained within the premaxilla and maxilla, the basal end occupying a thin-walled, thimble-shaped recess of bone (fig. 3) to be seen best in the disarticulated skull, and which is attached at only one limited region to the rest of the maxilla. In both upper and lower teeth, the intra-alveolar portions are longer than the extra-alveolar. When one compares the extra-alveolar lengths of the upper and lower teeth of the mature animal, the latter are always greater, and, as may be seen by reference to table 1, the difference in lengths becomes greater with increased age and size. In both upper and lower incisors the bone is so contoured around their imbedded portions that their course may be easily recognized. The basal end or foraminal apex of the lower 50 W. H. F. ADDISON AND J. L. APPLETON, JR. incisor forms on the outer aspect of the mandible a marked rounded projection, directed upwards and backwards beneath the coronoid process, and sometimes extending slightly posteriorly beneath the sigmoid notech.- Almost directly opposite this projection on the mesial aspect of the mandible is the inferior dental foramen. This projection marks the position of the grow- ing end of the formative organs of the incisor in the adult. In the new-born animal it is not present, nor at the end of the first month. By the age of 23 months it may be recognized, and thereafter it increases in prominence and constitutes a very evident feature of the bone. This region of the growing end of the tooth is protected by the zygomatic arch, and also by the overlying muscles. The course of the upper incisor may also be readily followed in the prepared skull. Laterally it is covered with a thin rounded layer of bone. Mesially it forms an elevated, distinct ridge projecting markedly into the nasal fossa. In the adult the position of its basal or growing end is not so prominent as that of the lower incisor. As these incisor teeth are an indispensable part of the rodents’ existence their importance demands pro- tection from traumatism which might injure their growing pulp. Here in the upper incisors, this protection is afforded by a flange of the maxilla running parallel to the lateral wall of the cranium, as shown in figure 1, as well as being encased in a separate thim- ble-shaped recess of bone (fig. 3), beneath, and separated by a narrow interval from, the outer layer of the maxilla. These details are in harmony with Cope’s idea (’88) of the influence of the incisors in moulding the general topography of the rodent skull. The diastema in the upper jaw is always longer than in the lower (fig. 1). By reference to table 1 it may be seen that in the mature animal the upper is nearly twice as long as the lower, but that in the younger stages the difference is not so great. The upper hair-covered lips are infolded into the diastema, dividing the oral cavity into an anterior and posterior com- partment. This arrangement probably prevents the débris and splinters of gnawing from entering the main oral cavity. STRUCTURE AND GROWTH OF INCISOR TEETH Sill The mandibular symphysis is formed of fibrous tissue and allows independent rotation of either ramus with its contained tooth. This lateral movement of the lower incisors appears to be under the control of the will of the animal. According to the observations of Jolyet and Chaker (’75) this mobility has a definite purpose in mastication. They observed a rapid alter- Fig. 4 Cross-sections of the (a) upper and (b) lower incisor teeth of a 5-month albino rat, taken near the alveolar margins. These show the arrangement of the enamel and the dentine, and the difference in contour of the enamel in the upper and lower teeth. The mesial surface cf each tooth is towards the right side. >< 15: nate separation and approximation of the tips of the lower in- cisors in the act of attempting to bite into a match or other slender object offered to the animal. At the same time the upper incisors were held stationary. Mention may be made here of a point of variation among the Rodentia in the relation of the angle of the lower jaw to the sheath of bone around the lower incisor. In the Myomorphi and Sciuromorphi the angle arises from the lower surface of the incisive sheath, while in Hystrix the angle arises entirely on the outer side. 52 W. H. F. ADDISON AND J. L. APPLETON, JR. Ryder (’77) suggested a classification of rodents based on the shape of their incisors as seen in cross-section. In some genera the diameter of the teeth is less from side to side, than in the antero-posterior direction, while in others the reverse condition is found. The present form belongs to the former group, as is shown in figure 4. From the consideration of many rodents, Ryder deduced the general principle, that where the incisors are thicker in the antero-posterior direction, the gnawing habit is greatly developed. MINUTE DESCRIPTION OF THE INCISORS Enamel and dentine make up the hard tooth substance, en- closing the pulp. Owen, in his ‘‘Odontography” (’40—’45, p. 399) said that there existed a general investment of cementum over the whole tooth structure. J. Tomes (50, p. 533) was not able to agree entirely but said that in most, if not in all, incisors of rodents cementum could be seen investing the posterior sur- face. In the rat, it is not apparent that there is any cementum at all. The enamel is usually colored with a pigment which is yellowish in the young but becomes orange-colored with age, and is usually more pronounced in the upper than in the lower incisors. At 13 days, there is as yet no color, but at 21 daysa slight tinge of yellow is perceptible in the uppers, but none in the lowers. At 25 days the uppers are distinctly yellow, and the lowers have now acquired a slight color. At 38 days, these colors have intensified, the uppers having more pigment than the lowers; and in the mature animal the same relation con- tinues, the uppers being orange-colored and the lowers yellow. The enamel is found principally on the labial side, and this accounts for the shape of the occlusal surface. For, the enamel being harder than the dentine, the latter is more easily worn away by the action of the opposing tooth, and the more resistant enamel remains as the cutting edge or point. The shape of the incisal end of the upper and lower teeth is different, being chisel- like (sealpriform) in the upper, and more rounded and narrower in the lower. The incisal line is also usually different in the STRUCTURE AND GROWTH OF INCISOR TEETH 5G upper and lower teeth. In the former, it is often slightly con- cave from side to side, while in the latter it is convex (fig. 5). As is shown in figures 1 and 5 the occlusal surface is an elon- gated concave area on the lingual aspect of the teeth, and in the living animal extends practically to the gingival margin. Due to the difference in the curve of the upper and lower teeth, the occlusal surface of the lower teeth is always longer than that of the upper, and in the mature animal it is usually found to be nearly twice as long. ' W/ | Fig. 5 Labial and lingual aspects of the extra-alveolar portions of the (a) upper and (b) lower incisors of a 5-month albino rat, showing the occlusal sur- faces and incisal edges of the teeth, and the outline of the bony alveolar mar- eins.) <2: It follows that because these teeth are constantly growing, the occlusal surfaces are constantly being worn away. As we shall see, when discussing the growth of the teeth, the elongated tem- poro-mandibular articulation is important, in allowing the teeth to have either the position pictured in figure 1 or to have the opposite relation, with the lower teeth outside of the upper. Thus the very important factor in the animal’s economy—the proper regulation of the length of the opposing incisors—is con- trolled by their own inter-action. 54 W. H. F. ADDISON AND J. L. APPLETON, JR. The pulp-chamber has the characteristic shape found in all permanently growing teeth, as is well seen, for instance, in the elephant’s incisor. Its cross-area is greatest at the basal end of the tooth, and gradually diminishes anteriorly. The pulp- chamber is found to extend in the tooth beyond the line of the gingivus, and very nearly to the occlusal surface. The shape Broke / E Ae ye A PA \ anterior end JUYIYL Fig. 6 Upper incisor of a 5-month albino rat (X< 5) and cross-sections of it at different points (X 8), to show the relative cross-area of the dentine and of the pulp chamber at these regions. The dotted line indicates the position of the margin of the alveolus. in cross-section of the pulp-chamber at different levels may be seen by reference to figure 6. The position of the filled-in pulp- chamber is usually well marked on the occlusal surfaces as a line (fig. 5). In weathered specimens of rats’ teeth from recent geological formations this last-formed part which fills in the pulp- chamber at the end of the tooth, is usually found to be lacking, and is evidently not of the same hardness as the surrounding parts of the tooth. STRUCTURE AND GROWTH OF INCISOR TEETH 55 MICROSCOPIC STRUCTURE OF ENAMEL AND DENTINE Sections of enamel show two layers; an outer thin and an inner thicker layer, as noted by Owen (’40-’45, p. 399). The enamel rods run in different directions in the two layers as fully described by J. Tomes in 1850. In the inner layer the enamel rods appear to run in two sets, obliquely to one another, while in the outer layer the rods are all parallel. The outer layer has also been called the fibrous layer, and in its superficial part is situated the yellow or orange pigment which gives the color to the enamel. Figures 7 and 8 show the arrangement of the enamel rods in the two layers. In the inner or plexiform layer, when exam- ined in cross section, the alternating series of enamel rods decus- sate, forming an angle varying between 70 and 90°. In longi- tudinal sections (fig. 26) these rods are slightly S-shaped, running outwards from the enamel-dentine surface at an angle of 50 to 54°, and inclining towards the anterior end of the tooth. Figure 8 is from a ground-section in which the enamel was broken during the process of preparation, and the broken edge shows distinctly the two sets of rods running at nearly right angles to each other. Under high magnification the rods are slightly notched. In cross-sections of the outer fibrous layer, the rods are paral- lel and form in the mid-line of the tooth an angle of 90° with the outer surface. As one proceeds away from the mid-line of the tooth, whether mesially or laterally, the general tendency of the long axis of the rods as they pass from the dentine junc- tion to the periphery, is to incline in the direction away from the mid-line of the tooth. The ameloblasts usually form an obtuse angle with the rods of the outer layer and seldom coincide in direction with them (fig. 7). In longitudinal sections the rods of the outer layer are not usually so distinctly seen as in cross-sections. In favorable longitudinal sections, however, they are seen to run quite obliquely, inclining towards the apex of the tooth, and forming an angle of 20 to 25° with the plane of the enamel-dentine junction. The pigment, as will be seen layer of = ameloblasts inner layer of enamel Fig. 7 Portion of cross-section of lower incisor with enamel-organ, pre- pared by the petrifaction method, showing the decussation of the enamel-rods in the inner or plexiform layer and their parallel arrangement in the outer or fibrous layer. X 350. Fig. 8 Small piece of enamel, showing the rods of the inner or plexiform layer running in two directions nearly at right angles to one another. X 350. 56 STRUCTURE AND GROWTH OF INCISOR TEETH ay below, is confined to the outermost part of the fibrous layer. There appears to be no Nasmyth’s membrane over the enamel, which means that there has been a complete transformation of the enamel matrix into enamel rods. The pigment extends - about two-thirds of the total length of the upper tooth, and about one-half of the total length of the lower tooth, and hence it follows that the deposition of enamel is completed within the basal third of the upper and the basal half of the lower tooth. By examining cross-sections of the tooth at different regions (fig. 6) it would seem that the full thickness of the enamel is attained within even a smaller area at the basal end of the tooth. The arrangement of the enamel over the labial aspect of the upper and lower teeth is shown in figure 4, drawn from cross- sections of the teeth of a 5-month animal. In both teeth the sections were made just posterior to the alveolar border. In both upper and lower teeth the enamel is thickest over the labial aspect, and is continued over the adjacent mesial and lateral surfaces. In both, the enamel is continued farther on the lateral than on the mesial surfaces, and relatively far- ther on the lateral surface in the lower than in the up- per tooth. In the upper tooth the enamel has a_ flattened external surface labially, while in the lower it has a rounded contour. In the upper there is a distinct labio-mesial and a labio-lateral angle, the enamel being somewhat thicker at the former. In the lower there is a labio-mesial angle, though less prominent than in the upper, and the labio-lateral angle is practically absent. In a 5-month animal the thickness of the enamel and its constituent layers was measured in the mid-line of the teeth, as follows: Upper Lowe Me M oO valet nicki esis cae Men bat eee cso vt aso teed ndteaee ae 100-110 140-150 Oudersilbromsela vers his el ate Ok oe 5 vs «id coe baoete oe 30-40 20-30 Pigmented portion of outer fibrous layer.............8-10-12 6-8 ihiner lesatiorta Tavern c2 jo. 0 64 cs vecinidy Jeialie lew see 70 120-125 It will be observed, however, in figure 4 that the enamel is not thickest in the mid-line of the upper tooth, but at the lateral and mesial angles. While the enamel of the upper tooth meas- 58 W. H. F. ADDISON AND J. L. APPLETON, JR. ures only 100 to 1104 in the mid-line, it measures 160 to 180u at the region of these angles, and is, therefore, thicker here than the enamel of the lower tooth. The increased thickness at the angles is principally in the inner plexiform layer, the other layer being increased only slightly or not at all. The outer fibrous layer is distinctly thicker in the uppers and has a slightly wider band of pigment in it superficially. This, no doubt, is the basis of the more deeply pigmented appearance of the labial surface of the upper as compared with the lower teeth. The dentine, unlike the enamel, grows continually thicker as one passes towards the outer end of the tooth. At the basal, erowing, end it begins as an extremely thin layer. The thick- ness at different points is seen in figure 6. As the dentine in- creases in thickness, the pulp-chamber is in consequence propor- tionately reduced. At the distal end there is no longer any pulp-chamber and the site of its previous position has been filled in by the formation of a kind of secondary dentine. C. Tomes (14) notes that ‘“‘in some rodents the final closure of the axial tract takes place almost by a continuance of the forma- tion of normal fine-tubed dentine, with very little secondary dentine of different structure, while in others there is a large area of dentine with vascular tracts in it.’’ In the rat there is relatively little of this secondary dentine. It is laid down in irregular trabeculae, with the pulp tissue, including blood- vessels, at first within it. At the exposed surface, however, it forms a continuous granular mass with apparently no soft tissues in it (fig. 27). The ordinary dentine of the tooth is quite typical in structure, with numerous parallel dentinal tubules, each having many fine lateral branches. The tubules are slightly sinuous, and the lateral branches anastomose with those of neighboring tubules. Sometimes a tubule sends off at an acute angle a branch nearly equal in diameter to the con- tinuation of the main tubule. This is usually in the dentine not covered by enamel. Where these large branches come off the diameter of the tubule is greater than elsewhere, measur- ing nearly 2u. Elsewhere the diameter varies from 1 to 1.7u. Slight differences may be seen between the tubules (a) in the dentine covered by enamel, and (b) in the dentine free from STRUCTURE AND GROWTH OF INCISOR TEETH 59 enamel. The tubules of the anterior region (a) of the dentine, covered by enamel, are more regularly parallel and have finer lateral branches than elsewhere. They also seem to taper slightly as one follows them towards the enamel. In the dentine not covered by enamel (b) the tubules are more sinuous and irregu- lar, the irregularities marking the position of origin of the larger lateral branches. In all parts at the periphery of the dentine the tubules end in a great number of very fine anastomosing arching branches. As a consequence of the smaller diameter of the little tubules here, a narrow zone at the periphery of the dentine has usually a more homogeneous appearance than has the remainder.. Towards the anterior end of the tooth, in the vicinity of the pulp-chamber, are vascular channels in the form of loops within the dentine. The tubules must necessarily take a curved course around these vascular channels, and thus the position of the vessels is more easily seen. In the dentinal tubules Mummery (12), Fritsch (’14) and others have demonstrated not only the processes of the odontoblasts, but also fine non-medullated nerve fibers. As to why the exposed den- tine on the lingual aspect of the teeth is insensitive, there are no definite observations to decide. A contributing factor may be the compression which the pulp tissues undergo at the an- terior end of the pulp-chamber, leading to the physiological cutting off of the nerve supply to the dentinal tubules. DEVELOPMENT OF THE INCISORS The times of the early stages of development of the incisors were seen as follows: 14-day fetus—slight thickening of oral epithelium 15-day fetus—distinct thickening and growth inwards of oral epithelium 16-day fetus—dental ledge and beginning of flask-shaped enamel organ 17-day fetus—dental papilla with crescentic enamel organ capping it 19-day fetus—both ameloblasts and odontoblasts differentiated new-born animal—enamel and dentine formation begun 8 to 10 days—eruption of the tooth Throughout life growth continues, and in the adult animal is on the average 2.2 mm. per week in the upper and 2.8 mm. per week in the lower incisor. JOURNAL OF MORPHOLOGY, VOL. 26, NO. 1 60 W. H. F. ADDISON AND J. L. APPLETON, JR. The structures to be described here, as in the development of the crowns of all teeth, are the enamel-organ with the amelo- blasts, and the dental papilla (which becomes the pulp-sub- stance) with the odontoblasts. There are two factors, however, which alter the usual history of the development of these struc- tures, and especially of the enamel-organ. First, in permanently growing teeth of which these are examples, all these struc- tures continue functional throughout life, so that the enamel- organ is also a persistent structure. The other factor and one correlated to some extent with the first, is that the enamel is formed on one side of the tooth only, and here only does the enamel-organ develop to its most highly differentiated functional condition. The history of the development and growth of the tooth may be conveniently considered in two stages (1) pre-eruptive, and (2) post-eruptive. The pre-eruptive stage extends from the 14th or 15th day of fetal life until eruption of the tooth takes place between the 8th and 10th post-natal days. Until near the time of birth there is no formation of enamel and den- tine, but from birth onwards these substances are laid down rapidly, so that at eruption, the teeth have their characteristic elongated narrow form. This pre-eruptive stage is characterized by the rapid elongation of the tooth-forming organs, and by the teeth attaining very similar relations to the other structures of the jaw which the imbedded portions of the erupted teeth pos- sess. Thus, the anlage of the lower incisor appears under the oral epithelium in the anterior region of the mandible, and grows continually backwards, until its growing end reaches the region beneath the developing molars. At this time the growing end presumably reaches a region which, by reason of its increasing calcification, offers resistance to further progress. The result of the ever-continuing mitotic division and cell growth at the basal end, is the pushing of the whole tooth and its forma- tive organs, in the opposite direction, and the consequent erup- tion of the tooth. During the latter half of this pre-eruptive stage, the anterior tip of the developing tooth structure is im- mediately beneath the oral epithelium, and remains at a fixed STRUCTURE AND GROWTH OF INCISOR TEETH 61 point, while the posterior end is continually growing back- wards and changing its relations. At eruption this condition changes, and the posterior extremity becomes practically a fixed point from which the whole tooth moves forward. That there is, however, a gradual change in the position of the pos- terior end of the tooth may be seen in figure 9. As the jaw grows, the entire tooth not only grows to keep the same general relative position, to surrounding structures, but it may be seen that the growing end progresses gradually posteriorly. In the full-grown animal this end occupies a distinct outpushing of the bone (fig. 1). During the post-eruptive period, which continues through- out life, this outward growth is continued at a regular rate, and at the same rate the outer end has to be worn away. This wear- ing-away process would soon result in the pulp becoming exposed were not the occlusal end of the pulp-chamber also being con- tinually filled in. As may be seen from figure 6 the dentine con- tinues to increase in thickness until near the end of the tooth. This means that the odontoblasts continue their regular func- tional activity until near the end of the tooth. However, the final filling-in of the pulp-chamber to form a continuous hard occlusal surface is accomplished by the deposition of a hard ma- trix between the pulp elements and by the probable calcification of the latter. The result is, that as the tooth is worn away, the soft pulp never becomes exposed. Although the pulp reaches very near to the end of the tooth, a hard substance always fills in the end of the pulp-chamber, and so protects the pulp beneath. DETAILED DESCRIPTION OF DEVELOPMENT UP TO THE TIME OF ERUPTION The anlage of the enamel-organ of each incisor arises as an epithelial ingrowth, distinct and separate from that for the molars. In frontal sections of the 14-day fetus, there are slight diffuse thickenings of the oral epithelium in the four positions, which represent the sites of the future tooth-formations. 62 W. H. F: ADDISON AND J. L. APPLETON, JR. 15 weeks 10 months Fig. 9 Series of mandibles of the albino rat at ages varying from birth to ten months, viewed from the lateral aspect. These show the changing relation of the basal end of the incisor to the rest of the mandible during this period. STRUCTURE AND GROWTH OF INCISOR TEETH 63 At 15 days these thickenings have become more definite, and in the lower jaws especially have begun to push into the underlying mesenchyme, and may be described as the dental ledges or dental laminae. At 16 days the ingrowths have continued to increase as broad masses of cells, pushing deeper into the underlying mesenchyme, and in the lower jaws the enamel organs may be distinguished as expanded structures, each connected by a slightly narrower mass of cells with the oral epithelium. In the upper jaws the differentiation of the enamel-organs from the remainder of the epithelial ingrowth is not so marked. At 17 days (fig. 10) the dental papillae are beginning, and the enamel-organs in both upper and lower jaws have a crescentic outline. In the enamel-organs there is already an indication of the differentiation into three layers. As seen in sagittal sec- tions, the papillae develop on the posterior side of the enamel- organs, thus foreshadowing the axis of growth of the tooth- forming organs in the antero-posterior direction. Eighteen-day fetus Series of frontal sections of 18-day lower jaws, show that the enamel-organs are growing over the dental papillae more rapidly on the labial and lingual surfaces than elsewhere, and extend more posteriorly on these surfaces. There are thus two pro- jections of the posterior margin of each enamel-organ as already noted by Meyerheim (’98). The labial process is broad and thin and extends more posteriorly than the lingual process, which is somewhat narrower and thicker. One may here remark, therefore, an early difference between the labial and lingual part of the enamel-organ. Other differences which will soon appear have not yet developed. Thus, the inner layer of the enamel-organ is made up of columnar elements which are still similar in all parts, both labially and lingually. In the dental papilla no columnar odontoblasts are yet seen. The enamel-organ remains connected with the surface epithe- lium by a broad band of epithelial cells. In the lower jaw, 64 W. H. F. ADDISON AND J. L. APPLETON, JR. immediately laterad to the line of junction of this stalk of the enamel-organ to the surface epithelium, the lip furrow is a de- pression, the plane of which is continued into the underlying mesenchyme by an ingrowth of surface epithelium several layers of cells in thickness. It is by the subsequent splitting of this epithelial layer into two, that the separation of the lip will be effected. Nineteen-day fetus At 19 days, the enamel-organ in the upper jaw (fig. 11) is erescentic in outline in sagittal section, and in the lower Jaw (fig. 12) is more elongated and conical in shape. At this age odontoblasts are first seen as columnar cells on the labial aspect of the mesodermal papilla. Three layers are recognizable in the enamel-organ, but the middle layer (enamel pulp), as has been also described by Sachse (94) for the mouse, is extremely thin, and therefore is not present in the great quantity typically seen in the development of rooted teeth. It appears as a more lightly stained zone between the inner and outer layers, and is thickest at the basal end of the enamel-organ as shown in figures 11 and 12. It averages about 20u in thickness and is made up of stellate cells loosely arranged. Already there is an indica- tion of a compact arrangement of the two or three rows of cells next the inner layer of the enamel-organ, which will result in the so-called stratum intermedium seen at later ages. This middle layer is also slightly more abundant at the anterior end in the region where the enamel-organ is continuous with the stalk which joins it to the oral epithelium. At this age the enamel-organs in the lower Jaws have a greater total length than those in the upper, and especially in the lower Fig. 10 Longitudinal section of upper jaw of 17-day fetus, showing tooth anlage of incisor, with the enamel-organ longer labially than lingually when measured from the point of junction of the stalk of the enamel-organ. The dental papilla is on the posterior aspect of the enamel-organ. % 70. Fig. 11 Longitudinal section of upper incisor anlage of 19-day fetus of al- bino rat, showing the crescentic outline of the enamel-organ, its greater length labially than lingually, and its thickened basal margin. X 70. Fig. 12 Longitudinal section of lower incisor anlage of 19-day fetus of albino rat, showing the conical outline of the enamel-organ and its greater length than in the upper jaw at the same age. X 70. STRUCTURE AND GROWTH OF INCISOR TEETH 65 66 W. H. F. ADDISON AND J. L. APPLETON, JR. jaws distinct differences may be made out between the oral and labial sides of the tooth-forming organs. These differences are: (1) The enamel-organ is longer labially than on the oral side. (2) The staining of the inner layer of the enamel-organ on the labial side is more intense, and here the cells are slightly longer than in other parts of the enamel-organ, measuring 24, in length and assuming the typical appearance of ameloblasts. Measurements show the similar cells on the lingual side to be about 20u in length. It is also to be noted at this age that the site of the most advanced cells which are differentiating to become ameloblasts is not at the apex of the enamel-organ, as is the case in the development of rooted teeth. For as one follows the cells of the labial side of the enamel-organ from the apex towards the base, while at the apex the cells are columnar they become longer as one goes posteriorly, and then towards the base of the enamel-organ diminish again. So that the site of most advanced differentiation here is a short distance pos- terior to the apex on the labial side. This is true also of the developing odontoblasts which are longest opposite the tallest ameloblasts. (3) The outer layer of the enamel-organ on the labial side is becoming slightly wavy in outline, and this denotes the begin- ning of the papillae, which form such a characteristic part of the mature functional enamel-organ (fig. 26). (4) The odontoblasts are seen only on the labial side of the dental papilla. Mitoses are abundant in all parts of the developing tissues. T'wenty-one-day fetus Thus the anlage of the rodent incisor begins in the usual way, and for a short time continues along the typical mammalian course. From 19 days onward, however, the differences which have already begun, become more distinct and definite. At 21 days (end of gestation) the enamel-organ has become more defi- nitely differentiated into a labial and a lingual region. Of the three constituent layers, the inner especially is strikingly different STRUCTURE AND GROWTH OF INCISOR TEETH 67 in these two parts. On the labial side at the anterior end, the organ has advanced to the condition where functional activity is beginning, while the oral side has remained stationary, or has actually retrogressed. Thus in the innermost layer on the labial side of the lower incisor, where the ameloblasts have begun to form enamel, these cells measure 30 to 34u in length, while the non-functional cells on the oral side of the innermost layer are low columnar or cubical in shape and measure only 12, in length (fig. 13). Comparison of these measurements with those at 19 days shows that the cells of the inner layer of the labial side of the enamel-organ have advanced in length from 24u to 30: or 34u, while the cells on the lingual side have decreased from 20 to 12u. There is, therefore, a primary tendency for the cells of the inner layer to develop equally in all parts, but very soon the non- enamel-forming cells of the lingual side begin to retrogress, while the functional cells of the labial side continue to grow. This constitutes another point of contrast with the development of the crowns of rooted teeth. For here in the 21-day fetus, when the enamel and dentine formation has just begun, these substances are thickest, not over the apex of the tooth-forming organs, as in the usual method, but at a short distance posterior to this point, on the labial surface. Thus, not only are the odon- toblasts and the ameloblasts first differentiated on the labial side, posterior to the apex, but at this region enamel and dentine formation is also evidently first begun. Over the apex of the dental papilla there is apparently a very thin outline of dentine deposited, but within this, in the tissues of the apex of the dental papilla, there is also beginning an irregu- lar formation of a hard matrix. Between the cells of the pulp, trabeculae of a bone-like material are appearing. As develop- ment proceeds this substance increases until the final result is, as seen in figure 20, that the primary apex of the tooth has a bone-like structure, consisting of cells imbedded in lacunae within a dense matrix. This has been called by Tomes (’04) ‘osteo-dentine.’ A similar difference between the labial and oral sides is noted in the cells on the margin of the dental papilla, which are be- 68 W. H. F. ADDISON AND J. L. APPLETON, JR. coming odontoblasts. In the basal half of the papilla (fig. 14), odontoblasts occur only on the labial side opposite the tall amel- oblasts, the peripheral cells of the other sides being still irregu- lar or cuboidal in shape. Farther forwards the odontoblasts are found also on the lateral and mesial surfaces of the dental papilla, but not on the lingual. In the apical one-fourth of the dental papilla odontoblasts occur all round, measuring 20 to 24 in length, and are engaged in the formation of dentine (fig. 13). The dentine is thickest on the labial side. In the region where enamel and dentine formation has begun no mitoses were seen in the formative ameloblastic and odonto- blastic cells, but posteriorly, where the deposition of enamel and dentine has not yet commenced, many mitoses occur in the layers of developing ameloblasts and odontoblasts, as well as elsewhere. The nearer one approaches the basal margin of the enamel-organ the more numerous are the mitoses and it is ap- parent that it is principally in this region that growth by addi- tion of new cells is taking place. One day old At the end of the first day of post-natal life, there has been great progress in the enamel and dentine formation, and the narrow, pointed outline of the tooth has been already laid down. In the upper jaw the teeth measure about 2.3 mm. in length and in the lower jaw about 3 mm. Definite changes in its relation to the oral epithelium have occurred also at the anterior end of the tooth. The original epithelial stalk connecting the enamel- organ with the oral epithelium has increased in size and the end Fig. 13 Cross-section of developing lower incisor of 21-day fetus of albino rat, nearer the anterior extremity of the tooth than figure 14. Shows the greater thickness of the labial side of the enamel organ, as compared with that of the other sides, and shows odontoblasts around the entire periphery of the pulp. Enamel and dentine formation has begun. X 110. Fig. 14 Cross-section of developing lower incisor of 21-day fetus of albino rat, posterior to the region shown in figure 18. No enamel or dentine yet formed at this point. Odontoblasts highest on the labial aspect, decreasing in height laterally but not yet differentiated as columnar elements on the lingual side. Enamel-organ thickest on the labial side. 110. STRUCTURE AND GROWTH OF INCISOR TEETH ite mew x4 ets: 69 70 W. H. F. ADDISON AND J. L. APPLETON, JR. of the tooth has apparently advanced somewhat into it. So this thick stratified layer of epithelium forms a close-fitting investment about the tooth apex, and is continuous posteriorly with the remainder of the enamel-organ. But in this epithelial cap there are no ameloblasts and consequently there can be no enamel over the osteodentine which forms the tip of the primi- tive tooth. This substance forming the tip of the unerupted tooth is a form of secondary dentine with its cells located in the lacunae of the matrix. Passing backwards, one comes to the ordinary dentine containing the vascular pulp with odonto- blasts situated at the periphery of the pulp-chamber in a regular manner. As the odontoblasts were first differentiated labially, and dentine formation began there before on the other side, the den- tine of the labial side is thicker than on the lingual side. Thus at a point about the middle of the entire tooth structure, the dentine measured 54u labially and 20u orally (fig. 15). Be- tween the odontoblasts are numerous fine capillary loops. At this region may also be seen the characteristic structure of the enamel-organ (fig. 15). This extends all around the tooth, but is much thicker on the labial side than elsewhere. This difference in thickness is seen in all the constituent layers. In the inner layer, the tall ameloblasts of the labial surface measure 40u, while the similarly situated cells on the other surfaces are cubical and measure only 10u. Comparing these with the previous stage described, it is seen that the cells on the labial surface have increased and those on the other surfaces have decreased. Of the middle layer on the labial side, the stratum intermedium is a distinet line of cuboidal cells, one to two rows in thickness, lying behind the ameloblasts. The other con- stituent—the original enamel pulp—is small in amount and is principally within the elevations of the outer layer, which form the beginning of the epithelial papillae. The cells of the outer layer, somewhat irregular in shape with round nuclei, are in a single row. Between the developing papillae (called by Sachse Stiitzpapillen) are numerous capillary blood-vessels. On _ the other surfaces, practically nothing remains of the middle layer, STRUCTURE AND GROWTH OF INCISOR TEETH fal although the outer layer still persists as a layer of flattened cells. Thus lingually the enamel-organ is represented by only two rows of cells—one representing the inner, the other the outer layer of the enamel-organ. . Fig. 15 Cross-section of developing lower incisor of 1-day albino rat, show- ing the great development of the ameloblasts on the labial side, and the thin- ness of the enamel-organ elsewhere. The space between the ameloblasts and the dentine is an artefect, and was formerly partly filled by the enamel, which has disappeared in the process of decalcification. In the layer of odontoblasts are seen the nuclei of the endothelial cells of the walls of capillaries. 110. Two days old Figure 16 shows a longitudinal section of a 2-day upper incisor. The epithelial enamel-organ is continuous over the whole tooth, but only shows its specialized functioning structure on the labial side. On the lingual side it is still intact and consists only of two rows of cuboidal or flattened epithelial cells. On the labial side, along the region where enamel has been formed (fig. 17) the Fig. 16 Longitudinal section of upper incisor of 2-day albino rat showing the enamel-organ continuous over the labial surface and terminating posteriorly in the thickened margin. X 18. Fig. 17 Small portion of preceding figure more highly magnified, to show the structure of the enamel-organ and the odontoblasts. a, outermost layer of enamel-organ and epithelial papillae; b, enamel pulp; c, stratum intermedium; d, layer of ameloblasts; e, layer of dentine; f, layer of odontoblasts. 175. 72 STRUCTURE AND GROWTH OF INCISOR TEETH ta ameloblasts measure about 40u. These are backed by two rows of darkly staining flattened cells composing the stratum inter- medium. Next to these is the looser arrangement of stellate cells, comparable to the enamel pulp of ordinary tooth develop- ment, but with much smaller spaces between the cells. This tissue is covered by the layer of cells constituting the outer layer of the enamel-organ, and the two together constitute the epithelial papillae. At the summit of each of these papillae the cells of the outer layer are grouped in a more compact manner. With higher magnification processes can be seen running from the ameloblasts into the developing enamel—the so-called enamel processes of Tomes. At the basal formative part of the enamel-organ the three original layers show distinctly. At the thickened basal margin of the enamel-organ, around its entire circumference, is a mass of rapidly dividing cells. As seen in figure 16 this thickened margin is more noticeable on the labial side. Its peripheral zone as seen in longitudinal sections is deeply staining and its cells, more or less columnar in shape, are compacted together. The interior, of more lightly stained appearance, is composed of oval or elongated cells, irregularly parallel, but more loosely arranged than the cells of the periphery. This region consti- tutes the site of origin of the cells of the ever-forming enamel- organ. From this pass forward the outer and inner layers, and between them, in larger quantity than is found more anteriorly, the tissue of the middle layer. This for a short distance is all enamel pulp and shows no differentiated layer of stratum inter- medium. In this formative region on the labial side, the inner layer consists of columnar cells, the future ameloblasts, in which many mitoses are seen. While the outer layer consists of cells which are columnar near the margin, a short distance anterior to this (150u) they change shape, first to cubical, then to flattened cuboidal. Between the two layers are cells representing the enamel pulp. At this region there are no papillae, although numerous blood-vessels are seen alongside the outer layer of the enamel-organ. About 0.5 mm. from the basal end this outer 74 W. H. F. ADDISON AND J. L. APPLETON, JR. layer of the enamel-organ becomes sinuous, and low papillae are being formed. On the lingual side, the structure of the basal end of the enamel- organ is similar, but somewhat simpler. Thus there are three layers at and near the basal margin, but soon, proceeding an- teriorly, these become reduced to two by the disappearance of the middle layer. The lingual side then consists of two rows of cuboidal or flattened cells, one constituting the outer and the other the inner layer of the enamel-organ in this situation. The dental papilla is made up of closely packed small stellate cells, with rounded nuclei. The mesenchymal cells which lie against the basal margin of the enamel-organ are rounded or irregular in shape, but within a short distance (0.5 mm.) anterior to this margin, the peripheral cells become first cubical and then columnar in shape. Where they are beginning to form dentine they measure 30 in length. From the odontoblasts processes enter the dentinal tubules of the dentine. The outer surfaces of the odontoblasts from which these processes arise show a distinct cuticular margin. Between the odontoblasts at short intervals capillaries form loops around the cells. These are evidently for the purpose of insuring an ample blood supply to these functionally active cells. Four days old By 4 days of age there has been continued growth, and deposi- tion of enamel and dentine. The upper incisor measures 3.6 ~ mm. in length and the lower 5 mm. The position of the apex of the tooth is in close relation to the oral epithelium (fig. 18). A thickened mass of epithelium, partly a derivative of the origi- nal stalk of the enamel-organ, and partly an ingrowth from the Fig. 18 Longitudinal section of upper incisor of 4-day albino rat, showing the increased curvature of the outline of the tooth and the relation of the apex of the tooth to the ingrowth of the oral epithelium. X 16. Fig. 19 Longitudinal section through basal end of labial side of enamel- organ of 4-day albino rat showing the region of the thickened margin. a, mar- gin composed of mass of proliferating cells; b, region where three layers are seen; c, region where stratum intermedium becomes differentiated from rest of middle layer. Anterior to the region of this figure the epithelial papillae appear and the ameloblasts begin to form enamel. X 80. 75 STRUCTURE AND GROWTH OF INCISOR TEETH 76 W. H. F. ADDISON AND J. F. APPLETON, JR. surface epithelium surrounding the tip of the tooth, is a prepara- tion for the eruption of the tooth, and will serve as a resistant ring of tissue through which the tooth will be pushed at eruption. It may be looked upon as a protective device, to prevent adjacent tissues from being carried out by the erupting tooth. The typical enamel-organ seen on the labial side does not cover the apex, for the tall columnar cells give place here, first to cubical and then to flat squamous epithelial cells, which form but a part of the thick mass of stratified epithelium, constituting the epithelial sheath over the end of the dentine. The other layers of the functioning enamel-organ also lose their identity at the region where the ameloblasts cease to have their char- acteristic elongated form. As maintained by von Brunn (’87) and Sachse (’94), there is no enamel apparent over the dentine at the apex of the tooth. The cells representing the enamel-organ on the lingual side can be traced forward for a short distance as a two-layered stratum. These cells are flattened, with oval nuclei. Beyond this point only a single regular row of cells is apparent, and about half way along the length of the tooth-structure, even this ceases to be definite, and apparently here the mesenchymal cells of the peridental tissues have grown between and scattered these cells. As a result of this activity of the mesenchymal cells in this region, the enamel-organ now ceases to exist as a com- plete conical investment of the tooth. Approaching the apex of the tooth on the lingual side, one finds the prolongation of the epithelial sheath as a thin layer of flattened cells which thickens as it passes forwards into the epithelial sheath. The basal formative end of the enamel-organ consists of a thickened band of tissue, as shown in figure 18, and under higher magnification in figure 19. This end is thicker on the labial side than elsewhere and it curves inwards, as seen in longitudinal sections, thus considerably diminishing the diameter of the entrance to the pulp-chamber. The extremity of this mass of tissue (fig. 19, a), constitutes a common origin for the several layers of the enamel-organ and contains many dividing cells. A short distance (0.1 to 0.2 mm.) from the extremity (fig. 19, b) STRUCTURE AND GROWTH OF INCISOR TEETH ai the cells form three layers, inner, middle and outer. The inner and outer layers, made up of columnar elements, stain more darkly than the middle layer, and the inner is thicker than the outer. The middle layer consists of elongated cells with oval nuclei, arranged for the most part with their long axes parallel to the surface of the enamel-organ. Frequent mitoses are also to be seen here, especially in the inner layer. In the region about 0.6 mm. anterior to this (fig. 19, c), where enamel formation has not yet begun, the innermost layer shows a single row of distinct tall columnar cells, the ameloblasts. The middle layer now shows two subdivisions (a) two or three layers of compacted flattened cells lying against the ameloblasts, and composing the stratum intermedium, and (b) a somewhat thicker stratum, lightly staining, of more loosely arranged cells, constituting the enamel pulp. The outermost layer is a single row of cubical cells, which form a straight continuous surface for the enamel-organ. Beyond this layer and in contact with it are numerous small blood-vessels. Passing still farther for- wards, the outermost layer becomes more sinuous in outline, and blood-vessels occupy the depressions between the elevations. This arrangement shows the beginning formation of the typical epithelial papillae. Seven days old At 7 days the tip of the tooth is in the oral epithelium (fig. 20), and ready for eruption, being separated from the outside by only a thin layer of superficial cornified epithelium. The epithe- lial tissues immediately about the apex of the tooth show the appearance of pressure atrophy. The cell boundaries are more indistinct than elsewhere, the tissue takes the acid stain deeply, and there is increased granularity—evidently degenerative effects due to the pressure of the advancing tooth. In the upper jaw, the basal end of the tooth in its backward growth has reached the region of the maxilla, into which it continues to grow, pushing before it a little pocket of thin bone. The average length of the upper teeth at this age is 5 mm., and of the lower teeth, 7 to 8 mm. Their pointed apices, and their comparatively slight curvature are shown in figure 24. 78 Wie Hin Eh. ADDISON AND Sits Ace PIG ON arin. Fig. 20 Osteodentine of apex of tooth of 7-day albino rat imbedded in the surface epithelium, showing cells in the lacunae in the matrix. » 175. ERUPTION OF TEETH Eight to ten days During the process of eruption (fig. 21), the tooth and its formative organs gradually move forward as a whole, and the apex of the dentine forming the anterior end of the tooth pierces the surface epithelium. This procedure is accompanied by new changes in the tooth-forming organs. For while the same process of cell-division continues at the basal end of the dental papilla and enamel-organ, these structures are subjected to Fig. 21 Longitudinal section of the upper tooth of an 8-day albino rat, show- ing the apex of the tooth piercing the surface epithelium. X 10. Fig. 22 Longitudinal section of the upper tooth of a 12-day albino rat, show- ing the increased size and curvature of the tooth, the basal end directed more towards the palatal surface and the progression of the apex of the tooth through the epithelium. X 10. Fig. 23 Longitudinal section of the upper incisor of a 26-day albino rat, showing the well-established occlusal surface, the approximation of the basal end towards the palatal surface, abundant blood-vessels in the pulp, and the position of the granular osteodentine filling in apex of the pulp chamber. X 10. 23 79 80 W. H. F. ADDISON AND J. L. APPLETON, JR. new conditions at the erupting end of the tooth. Before detail- ing these changes, it may be advisable to state, in a general way, the changing circumstances attendant upon eruption. Up to this time the anterior end of the tooth has been nearly stationary, but there has been continued growth backward of the posterior extremity. At this time the rate of progression forward is greatly increased, and the rate of progression backward much reduced. As suggested before, the process of eruption may depend largely upon the fact of increasing calcification in the bones, rendering them more resistant to the backward growth of the develop- ing tooth. Whatever may be the causal factors, from now on the tooth continues to grow out at a regular rate, through the development of new cells at the basal end of the formative organs, these cells in turn giving rise to the hard parts of the tooth. Within a few days after eruption, the use of the tooth involves the process of attrition by which, in spite of the regular rate of growth, the exposed length is kept nearly constant for any age. It is generally agreed that, by reason of the protoplasmic processes which extend into enamel and dentine from amelo- blasts and odontoblasts respectively, these cells must be car- ried along with the tooth as it moves. Thus, as there is con- stantly a regeneration of these cells at the basal end of the tooth, there must be an opposite process of some nature by which these cells are eventually lost at the apical end, when carried thither by the outward progress of the tooth. First we may follow the history of the ameloblasts in this locality. Before eruption, the enamel-organ is continuous with the stratified epithelium forming the sheath around the gingival margin, and this relation continues at and after eruption. As the tooth moves forward during eruption the ameloblasts must move along with it and, when those at the anterior end approach the gingival margin, they must either be held there, or be car- ried out on the enamel until detached. On examining longi- tudinal sections at 12 days (fig. 22) it is seen that the amelo- blasts, as they approach the gingival margin, become shorter and shorter, until, beneath the thickened sheath of epithelium STRUCTURE AND GROWTH OF INCISOR TEETH 81 forming the gingival margin they acquire a flattened form. As a continuation of these flattened cells next the tooth is seen, extending out into the space between the erupted tip of the tooth and the epithelial gingival margin, a thin layer of tis- sue, which must be looked upon as the portion of the enamel- organ which has been carried out during eruption. At later stages this same appearance occurs—a thin layer of flattened cells continuous with the enamel-organ lying in the space be- tween the tooth and the epithelium of the gingival border. It may be that some of the cells are added to the epithelium of the gingival margin, but the majority appear to be continually car- ried out, and eventually detached. The mesenchymal tissues of the pulp at the anterior end are little affected by the mere act of eruption and not until some days later when attrition begins, do we see definite changes. At eruption the anterior conical extremity of the tooth is formed of osteodentine, containing within its matrix the remains of scattered cells and blood-vessels. Immediately posterior to this begins the true fine-tubed dentine with a central pulp-cham- ber. The cells at the anterior end of the pulp-chamber are irregularly arranged, but following backwards one soon sees the odontoblasts in parallel arrangement at the periphery of the chamber. At 10 days, when the apex of the tooth has pierced the epithelium and is easily seen from the outside, the measurements of the upper and lower teeth are 7 and 11 mm. respectively. At 12 days, they have increased to 7.5 and 11.8 mm. CHANGES IN APEX OF TOOTH BY USE Already at 12 days, when one examines the exposed ends of the teeth, they show little pits, which have been caused by the pressure of the opposing teeth. At 14 days, the ends are flat- tened, and at 16 days, because of the increased obliquity of this flattened surface due to the wearing away of the lingual side of the dentine, they are acquiring a cutting edge. The length of these occlusal surfaces continues to increase so that by 19 or 21 days (fig. 24), they have nearly the appearance typically 82 W. H. F. ADDISON AND J. L. APPLETON, JR. seen in the fully developed teeth. The osteodentine of the tip of the tooth is softer than is true dentine, for when the young tooth is dried this end shrivels and darkens in color. This cap of osteodentine on the end of the tooth may be useful, as suggested by Sachse (’94), because of its softness, in allowing the early formation of the functional occlusal surface. When this soft substance begins to wear away the tissues of the pulp would soon become exposed were there not a provision for the filling in of the apex of the pulp-chamber. This is effected by the formation of an irregular hard matrix, which may also be called osteodentine, within the extremity of the pulp-chamber. As Fig. 24 Isolated upper and lower incisors of several ages of young albino rats. The pointed shape just before eruption is seen at 7 days. At 12 days, there is yet very slight change in the apices. At 21 days the occlusal surfaces are concave, and at 26 days they have nearly the typical mature appearance. x 2. the outer surface of the tooth wears away, this formation is constantly taking place a short distance from the occlusal surface. Thus in examining a longitudinal section of the tooth at an age when the process of attrition has begun, and the typical occlusal surface has been formed (e.g., 26 days, fig. 23), we find this form of secondary dentine or osteodentine filling in the distal extremity of the pulp-chamber. As one approaches the anterior end of the pulp-chamber, the pulp becomes more and more restricted and the blood-vessels appear congested. Pro- ceeding distally, the irregular matrix formation is seen between the cells and blood-vessels and finally near the occlusal worn surface is a granular mass of osteodentine with no circulating blood in it, but spaces are still seen containing the remains of the pulp elements. Here the living elements have disappeared, STRUCTURE AND GROWTH OF INCISOR TEETH 83 but by staining (e.g., with acid fuchsin), the remains of these may be made out in less calcified spots in the matrix. Evidently the odontoblasts and other tissues of the pulp which move with the dentine, become more and more compressed at the narrow- ing apex of the pulp-chamber, and finally there is this irregular deposit of secondary dentine between them, which serves to obliterate the pulp-chamber. As the tooth moves out, this process is constantly going on, just in advance of the occlusal surface, and keeps pace with the process of attrition. It is interesting to note the rate at which the teeth are in- creased in length during their formative period and prior to attrition. Upper Lower mm. mm. iL releinedralhal,. .. o's Sheba eaten eile ete am hag mn a ze 2.3 3 AMG SO CPR ee Rae os. Sin eenctr ne aa Gi bid ct aaa a 3.6 5 CLAN, Sig © | CLONER ER ENA ean Ine Jc ree teasers MR te thi stcegs dee a ae eee 5 7-8 ORG aiSe OG Merry ee cues cP ou sb rae Ane ai cee e Se Ti 11 Average growth 0.52 mm. and 0.88 mm per day As will be seen later, this exceeds the rate at which the mature tooth continues to grow out. DESCRIPTION OF MATURE TOOTH AND TOOTH-FORMING ORGANS, IN FIVE-MONTH ANIMAL In the mature tooth, the general relations are shown in figure 25, made from a photograph of a decalcified section of the upper tooth of a 5-month animal. The regular curved outline is seen, with the greater proportion of the length imbedded within the jaw, and only a small part projecting. The formative end lies within an investment of bone belonging to the maxilla. At this end the dentine is very thin and the pulp greatest in amount. As one goes forward, the dentine increases regularly in thickness while the pulp-chamber becomes smaller and smaller. The vacuolated appearance at the anterior end of the chamber is due to shrinkage of the pulp tissue during fixation. The enamel has been lost in the process of decalcification except over the basal third. Numerous blood-vessels are seen within the pulp. 4 "We epithelial papillae } ameloblasts jouter layer of enamel eeeensiiemineemiameamienmn ee inner layer of ename!| dentine Fig. 25 Longitudinal section of upper incisor of a 5-month albino rat. The letter a shows where the next illustration (fig. 26) is taken. X 6. Fig. 26 Small portion of the preceding, more highly magnified, to show the enamel-organ and the enamel, and the odontoblasts and dentine. X 185. 84 STRUCTURE AND GROWTH OF INCISOR TEETH 85 The enamel-organ is continuous over the convex labial sur- face of the imbedded portion of the tooth but is restricted to the most posterior region of the other surfaces, extending only 1 mm. forward from the basal margin. The enamel-organ differs in its structure in three regions of the labial side, and may be described separately in these three parts: (1) at the basal formative end, (2) near the gingival margin and (3) in the long intervening region. In (1) the enamel-organ is being constantly regenerated by the addition and growth of new cells. In (2) the enamel-organ is undergoing a retrograde proc- ess, while (3) represents the region where the enamel-organ is at its highest functional development, although its activity in increasing the thickness of the enamel is restricted, as noted before, to the basal third or less in the upper and to the basal half in the lower tooth. Considering first the region (3), as shown in figure 26, the enamel-organ is conspicuous by reason of its tall ameloblasts and the high, narrow papillae. The enamel-organ is described in three layers—inner, middle and outer. The inner layer consists of the ameloblasts, which measure about 40 wu in height, with nuclei situated towards the outer end of the cells. The middle is composed of two strata (a) stratum intermedium, and (b) enamel pulp. The stratum intermedium is formed of 1 or 2 rows of fairly regular cuboida cells resting upon the outer ends of the ameloblasts, but the enamel pulp is not now recognizable as a distinct layer and exists principally within the papillae. The outer layer of the enamel-organ consisted originally of a single layer of cells, but these are no longer regu- lar in form or arrangement. Together with the remains of the enamel pulp, the outer layer forms the papillary elevations, 60 to 70 » in height. These papillae are surrounded by an abun- dant capillary blood supply for the nourishment of the cells engaged in the formation of the enamel, and the purpose of the elevations is apparently to increase the surface area through which absorption may take place from the blood-stream. The enamel is in two layers (fig. 26), the rods while travers- ing the inner layer being very distinct, and inclining towards 86 W. H. F. ADDISON AND J. L. APPLETON, JR. the apex of the tooth at an angle of from 50 to 54° with the dentine surface. The continuations of these rods in the outer layer are not so distinctly seen, but the inclination, as made out in thin sections, is still greater towards the apex, forming an angle of from 20 to 25° with the plane of the surface of the dentine. As noted before in the study of enamel, the rods in the inner layer, when observed in cross-sections of the tooth, decussate at an angle of from 70 to 90°, but when they reach the outer layer all run parallel. The fact that the rods run in these various directions seems incontrovertible, but in the light of our present knowledge of enamel formation it is difhieult to understand how this condition is arrived at. If each amel- oblast is responsible for an enamel-rod, then it follows that because the alternate layers of rods are oblique to one another, the ameloblasts responsible for these series of rods must have changed their relative positions during the process of formation of these rods. No such phenomenon has been observed, or even suggested. The other possibility is that the matrix of the rods is formed in a regular manner, but that afterwards, before calcification is complete, the rods become re-arranged owing to pressure strains. The plane of direction of the rods is suggestive of the im- portance of the enamel-organ in the persistent growth. For always the general plane of the rods, as they leave the enamel- dentine junction, is towards the outer end of the tooth, as if the ameloblasts, while engaged in enamel-formation were al- ways held back by the enamel, in which their processes were imbedded. The basal formative end of the enamel-organ (region 1) in the adult animal corresponds very closely in structure with what has already been described for earlier ages, e.g., 4 days. This is the region where the enamel-organ is constantly being renewed, and it retains the same embryonal character at all stages of development. At the anterior end where the enamel-organ is continuous with the surface epithelium (region 2), a gradual transition oc- STRUCTURE AND GROWTH OF INCISOR TEETH 87 curs between the typical enamel-organ and the stratified squa- mous epithelium (fig. 27). As one follows the innermost layer of the enamel-organ forward, the cells become shorter, until they are cubical and finally flattened in shape. Here the other layers also lose their regular arrangement, and form, with the preceding, a thin layer of stratified cells. This layer can be followed directly into contact with the epithelium of the gingivus. Fig. 27 Longitudinal section of the apex of the tooth of a 5-month albino rat, showing at a the position of the outward prolongation of the remains of the enamel-organ, and at b the more granular osteodentine filling in the apex of the pulp chamber. X 10. The cells, however, do not lose their identity in the surface epithelium but remain separate as a thin layer lying against the enamel (fig. 27, a). This thin layer of epithelium, there- fore, represents the ultimate fate of the enamel-organ after it has completed its functional activity. It is being continually pushed out and its most anterior part must be continually being lost. 88 W. H. F. ADDISON AND J. L. APPLETON, JR. RATE OF GROWTH OF THE INCISOR TEETH Two methods were used for determining the rate of growth of the incisor teeth (a) cutting off one or more teeth at the gingi- val margin and (b) making marks upon the enamel. The results here given are based on the latter method, as giving more nearly the normal rate of growth. By means of a dental engine, the animal always having been anesthetized, a fine transverse notch was made on the enamel of the incisors a short distance from the gingival margin. The interval between this mark and the tip of the tooth was then measured. At the end of about a week the distance between these two limits was again taken, and the difference between the two measurements showed the amount of wearing away. ‘Two series of experiments were made by this method upon adult animals. In the first series six animals were used and in the second four animals, and meas- urements were made for several consecutive weeks. The long- est period that one individual was studied was six weeks. The two series gave very similar results. The lower tooth was always found to grow more rapidly than the upper. The upper tooth averaged 0.31 to 0.32 mm. per day, or 2.2 mm. per week, while the lower tooth averaged 0.4 mm. per day, or 2.8 mm. per week. No doubt there are many variations of the rate of growth under different circumstances, so that these figures must be taken as representing the average rate under one particular set of conditions. The food of these animals was the mixed diet now in use in the rat colony of The Wistar Institute. A short series of trials was made with animals kept in a large glass jar and given only soft food. In these animals the rate of wearing away corresponded very closely to that seen in animals which have also hard food and have the oppor- tunity of gnawing. In these, therefore, the interaction of the opposing teeth must have caused the attrition. For assistance with these experiments we wish to thank Dr. Stotsenburg, who greatly aided us in carrying out our ob- servations. STRUCTURE AND GROWTH OF INCISOR TEETH 89 It is interesting to compare these results with those obtained on the rabbit—the only other rodent which has been carefully studied in respect to the growth of its incisors. MacGillavry (75), using a young adult rabbit, made marks upon its lower incisors 2.6 mm. and 3 mm. from the tip. After five to seven days the marks had disappeared. Evidently the rate of growth was about 2.5 to 3 mm. per week. Noé (’02) used a rabbit which happened to possess overgrown teeth. The animal accidentally broke off the lower incisors in the bars of its cage, and Noé made observations upon the rate of their growing out. This he found to be .615 mm. per day, or 4.3 mm. per week. This is larger than MacGillavry’s results and may have been due to the unopposed growth and to the other abnormal conditions which may have been present in the formative organs. Using MacGillavry’s figures for comparison, it would seem that the lower teeth of the albino rat and of the rabbit grow out at about the same rate. OVERGROWTH OF INCISORS Examp!es of overgrowth of the incisors of rodents, especially in rabbits and hares, which were hunted as food, must have been observed from early times. In the older literature, they are referred to principally as curiosities, which have excited the interest of whoever has found them. Later the causes of the malformations were also considered. Thus Jenyns (’29), to cite only one observer, found several examples in wild rabbits, and has given a good illustration of the curved aspect of the teeth. He also clearly states the several causes which, in his opinion, may give rise to the condition. In addition to the one usually accepted at his time—accidental breaking off of one tooth—he considered also as causes (a) too soft food, (b) morbid or too rapid secretion of the osseous matter of the teeth, and (c) dislocation of one of the condyles. Wiedersheim (’02—’03) has reported a case occurring in a rat, where he found an associated assymmetry of the cranium. He is in doubt as to which was cause and which was effect—the overgrowth of the teeth or the assymmetry of the cranium. 90 WiiG AslG MIG MDIDIESOI, VND) As Wis ZNIPIBILIDAMOIN, dilate Fig. 28 Cranium of albino rat, showing the overgrown upper incisors recurv- ing to the left side. The left incisor passes to the outer side of the skull, while the apex of the right incisor has penetrated the bone of the maxilla in the region of the basal end of the left incisor. X 1. Fig. 29 Cranium of the same albino rat shown in the preceding figure, viewed from the right side. It shows the overgrown lower incisors recurving to the right side, and the cavity which the right incisor has worn in the palate bone. Figures 28 and 29 show a skull obtained some years ago from the rat colony of The Wistar Institute by Dr. Stotsenburg, and prepared in the Histological Laboratory by Miss E. F. Brooks. ‘The upper teeth curved to the left side of the head and the lower to the right side. As seen in figure 29, the right lower has penetrated through the bone of the palate into the STRUCTURE AND GROWTH OF INCISOR TEETH Of nasal chamber, while the right upper (fig. 28) has recurved and grown into the maxilla. In The Wistar Institute rat colony, at the time when the animals were fed on bread and milk, frequent examples of this and similar conditions were found, but now under a more varied mixed diet they practically never occur. Beretta (13) has recently made an analysis of these abnor- malities and has classified them in three groups. (1) Overgrowth of the upper and lower incisors through lack of an opposing tooth. (2) Overgrowth of the incisors of the upper and lower jaws through deviation of the jaws. , (3) Prognathism of the lower jaw, and as a result, over- growth of the incisor of the lower Jaw. In the present instance, diet seemed to be the controlling factor, probably by reason of its influence on the hardness of the bone of the alveoli from which the teeth grew out. SUMMARY The rate of growth of the upper and lower incisor teeth of Mus norvegicus albinus, in the mature animal, averages 2.2 and 2.8 mm. per week, or 12.5 cm. and 14.5 cm. per year, respectively. Growth is due primarily to the proliferation and growth of cells at the basal end of the enamel-organ, where new enamel- forming cells arise, and at the basal end of the dental papilla where new dentine-forming cells develop. The enamel-organ of the adult forms a narrow circular band around the basal end of the tooth, and extends forward from this on the labial side only. It coincides in its lateral bound- aries with the enamel, and extends along the entire imbedded portion of the tooth. Anteriorly, it comes in contact with the epithelium of the gingival margin, and is carried out continually as a narrow band of cells lying on the enamel, between the latter and the gingival epithelial tissue. The first indication of the anlage of the incisors appears in 14-day-old fetuses. In fetuses, 21 days of age (just before JOURNAL OF MORPHOLOGY, VOL. 26, NO. 1 92 W. H. F. ADDISON AND J. L. APPLETON, JR. birth), enamel and dentine formation is beginning. In animals 1 day old the upper and lower teeth measure 2.3 and 3 mm. At 8 to 10 days the teeth erupt, and at 10 days measure 7 and 11 mm. respectively. This period is therefore characterized by the rapid elongation of the teeth. The process of attrition begins within a few days after erup- tion, so that by 19 or 21 days of age, the typical occlusal surface is formed. Up to the time of eruption the anterior end or apex of the tooth is immediately under the oral epithelium, while the basal or growing end is continually progressing posteriorly. After eruption, the basal end becomes nearly stationary in position, while the whole tooth structure is continually moving forward. The extra-gingival length of the tooth is kept con- stant, however, by the attrition of the occlusal surface, either through use in gnawing or by the action of the opposing teeth. The histogenesis of the enamel-organ is practically com- pleted by the 4th day after birth, although it does not attain its final relations to the tooth as a whole, until after eruption. In the 18-day fetus the enamel-organ is similar in all parts, and the cells of the inner layer measure the same, both lingually and labially. From this period forwards, however, the labial por- tion continues to progress towards its fully differentiated func- tional structure, while the lingual portion retrogresses, until at 4 days after birth the latter is disrupted, by the ingrowth of the surrounding connective tissue. Contrasting the cells of the inner layer—the potential ameloblasts—on the labial and lingual sides, they are practically the same in the 18-day fetus, but at 19 days they are found to measure 24 and 20 uwrespectively. In the 21-day fetus, they measure 30 to 34 and 12 u, and 1 day after birth the true ameloblasts on the labial side have increased to 40 uw, while the non-functional cells of the lingual side are only 10 u in height. At 4 days, the latter cease to form a con- tinuous layer, by reason of the dispersion of the cells by the surrounding connective tissue, except at the basal formative region. Characteristic of the permanently-growing enamel-organ are the epithelial papillae, formed by the elevations of the outer STRUCTURE AND GROWTH OF INCISOR TEETH 93 layer of the enamel-organ, and the cells of the enamel pulp. Between these elevations are numerous capillaries which in- sure a rich blood supply to the enamel-forming cells. There are three layers in the functional enamel-organ—inner, middle and outer. The inner is constituted of the tall amelo- blasts, and the middle is made up of two divisions, (a) stratum intermedium and (b) enamel pulp. The latter unites with the single layer of cuboidal cells which compose the outer layer, to form the epithelial papillae (fig. 26). The apex of the primitive tooth is formed of a variety of secondary dentine—‘osteo-dentine’ of Tomes—which is softer than true dentine, and differs in its structural arrangement (fig. 20). After eruption, this terminal portion of osteodentine is soon worn away by attrition, and the typical occlusal surface is developed, as seen at 19 or 21 days. At 21 and 23 days the first two molars erupt in both upper and lower jaws, and from now on the animal is able to secure food for itself, and if neces- sary can maintain an independent existence. As the tooth continues to be worn away there is a provision for the continual filling in of the apex of the pulp-chamber by the formation of what may also be called osteodentine. This is a form of secondary dentine, containing, when first formed, cells and blood-vessels. This is always at a little distance, about 1 mm., from the occlusal surface, but as any part of the tooth, in its outward progression, approaches the occlusal sur- face, the soft elements disappear within the osteodentine, and the latter forms a hard continuous surface with the adjoining true dentine. The position of this osteodentine is marked as a line on the occlusal surface of the teeth (fig. 5). Prior to eruption there develops around the apex of the tooth, as it lies in contact with the surface epithelium, a thickened ring of stratified epithelium. This ring of tissue is pierced by the apex of the tooth at eruption, and would seem to have the function of serving as a resistant margin for the soft tissues, and of preventing other tissues being carried along with the erupting tooth. 94 W. H. F. ADDISON AND J. L. APPLETON, JR. The length of the teeth varies with the size of the cranium (table 1) so that the persistent growth is not only sufficient to offset the continual attrition, but also serves to keep the length of the teeth in a definite relation to the length of the skull, as the latter increases in size. The lower tooth is always longer than the upper, and this difference manifests itself even in the anlagen of these structures in the 19-day fetus (figs. 11 and 12). The contour of the enamel, as seen in cross-sections, is char- acteristically different in the upper and lower teeth (fig. 4). The enamel of the tooth is composed of two layers which are different in appearance. The enamel rods run in two sets which decussate with each other in the inner or plexiform layer, but they change their direction as they continue into the outer layer, so that in it they are all parallel. In longitudinal sections, the general direction of the rods (fig. 26), is to incline towards the apex of the tooth, as they run from the enamel-dentine boundary to the outer surface of the enamel. In conclusion, we wish to thank Professor Piersol for generous assistance in many ways, and Professor Donaldson for his con- stant interest in the study. We also wish to acknowledge the kind assistance of Mr. E. F. Faber in the preparation of the drawings. STRUCTURE AND GROWTH OF INCISOR TEETH 95 LITERATURE CITED Apuorr, P. 1898 Zur Entwicklungsgeschichte des Nagetiergebisses. Jena. Zeitschr. fiir Naturwissenschaft, Bd. 32, ss. 347-410. Beretta, A. 1913 La normala dentatura dei roditori in rapporto alle anomalie dentali in questi osservate. La Stomatologia, t. 10. Abstract in Deutsche Monatsschrift fiir Zahnheilkunde, April, s. 287. von Brunn, A. 1887 Ueber die Ausdehnung des Schmelzorganes und seine Bedeutung fiir die Zahnbildung. Arch. f. mikr. Anat., Bd. 29, ss. 367-383. BurckuarpT, R. 1906 In Hertwig’s Handbuch der Entwickelungslehre der Wirbeltiere, Bd. 2, Teil 1, Kapitel 4, ss. 349-456. ‘‘Die Entwickelungs- geschichte der Verknéckerungen des Integuments und. der Mund- hoéhle der Wirbeltiere.”’ Corr, E. D. 1888 The mechanical causes of the origin of the dentition of the Rodentia. Amer. Nat., vol. 22, pp. 3-11. Donaupson, H. H. 1912 The history and zoological position of the albino rat. Proceed. Acad. Nat. Sci., Philadelphia. Freunp, P. 1892 Beitrage zur Entwicklungsgeschichte der Zahnanlagen bei Nagethieren. Arch. f. mikr. Anat., Bd. 39, ss. 525-556. Fritscu, C. 1914 Untersuchungen tiber den Bau und die Innervierung des Dentins. Arch. fiir mikr. Anat., Bd. 84, ss. 307-20. Harat, 8. 1907 On the zoological position of the albino rat. Biol. Bull., vol. 12, pp. 266-273. Jenyns, L. 1829 Observations on a preternatural growth of the incisor teeth occasionally observed in certain of the Mammalia Rodentia. Lou- don’s Magazine Nat. Hist., London, vol. 2, pp. 134-137. JOLYET ET CHAKER 1875 De l’acte de ronger, étudié chez les rats. Comptes Rendus et Mémoires de la Soc. de Biol., pp. 73-74. Lowe, L. 1881 Beitrige zur Kenntniss des Zahnes und seiner Befestigungs- weise im Kiefer. Arch. f. mikr. Anat., Bd. 19, ss. 703-719. MacGituavey, T. H. 1875 Les dents incisives du Mus decumanus. Arch. Néerl. Sc. exact. et nat., Haarlem. MeryerueimM, M. 1898 Beitrige zur Kenntnis der Entwicklung der Schneide- zihne bei Mus decumanus. Dissertation. Leipzig. Mummery, J. H. 1912 On the distribution of the nerves of the dental pulp. Philos. Trans. Roy. Soc., London, vol. 202, B., pp. 337-349. Nok, J. 1902 Vitesse de croissance des incisives chez les Léporides. Comptes Rendus, hebd. des Séances et Mémoires de la Soc. de Biol., pp. 531-532. Oupet, J. E. 1823 Expériences sur l’accroissement continué et la reproduc- tion des dents chez les lapins. Jour. de Physiol. Expér. et Patholog., Tomes 3 et 4. Owen, R. 1840-45 Odontography. London. 96 W. H. F. ADDISON AND J. L. APPLETON, JR. REIcHERT, E. T., anpD Brown, A. P. 1910 The crystallography of hemo- globin. Pub. Carnegie Inst. of Washington. Rerzius, A. 1838 Bemerkungen itiber den inneren Bau der Zihne mit be- sonderer Riicksicht auf den in Zahn vorkommenden Réhrenbau. Miiller’s Archiv. Roetter, F. 1889 Ueber Entwicklung und Wachstum der Schneidezihne bei Mus musculus. Morphol. Jahrb., Bd. 15, ss. 457-477. Ryper, J. A. 1877 The significance of the diameters of the incisors in Roden- tia. Proc. Acad. Nat. Sci., Philadelphia, vol. 29, pp. 314-318. 1878 On the mechanical genesis of tooth-forms. Proc. Acad. Nat. Sci., Philadelphia, vol. 30, pp. 45-80. Sacuse, B.° 1894 Entwicklung der Schneidezihne bei Mus musculus. Disser- tation. Leipzig. Sracu, J. 1910 Die Ontogenie der Schneidezihne bei Lepus cuniculus. Ex- trait. Bul. d’Acad. Sc., Cracovie. Tomes, C. 8. 1914 A manual of dental anatomy; human and comparative. 7th Ed. Edited by H. W. Marett-Tims and A. Hopewell-Smith. Tomes, J. 1850 Structure of the dental tissues of the order Rodentia. Phil. Trans. Royal Society of London, pp. 529-567. TuLuBEerRG, T. 1898-99 Ueber das System der Nagethiere. Nova Acta Reg. Soc. Sc. Upsaliensis, Series 3, ss. 1-514. Weser, M. 1904 Die Siugetiere. WIEDERSHEIM, R. 1902-03 Ein abnormes Rattengebiss. Anat. Anz., Bd. 22, ss. 569-573. Wiuuiams, J. L. 1896 The formation and structure of dental enamel. Dental Cosmos, vol. 38. Woopwarpb, M. J. 1894 On the milk dentition of the Rodentia with a de- scription of a vestigial milk incisorinthe mouse (Mus musculus). Anat. Anz., Bd. 9, ss. 619-631. A PECULIAR STRUCTURE IN THE ELECTROPLAX OF THE STARGAZER, ASTROSCOPUS GUTTATUS JAMES G. HUGHES, JR. From the Histological Laboratory of Princeton University, U. S. A. THREE FIGURES The purpose of this paper is to determine the function and composition of the peculiar pointed fibers and long pointed rods lying in the electric layer of the electroplaxes of the stargazer, Astroscopus guttatus. Before proceeding with a discussion of these rods, a_ brief description of the electric organ of this fish (according to Dahl- eren)! will be given. The electric apparatus is composed of two organs, which form two vertical columns roughly oval in horizontal section, and placed behind and somewhat under each eye. Each organ extends from the peculiar bare spot on the top of the head down to the tissues which form the roof of the oval cavity; and is composed of about 200 thin layers of electric tissue, which extend horizontally all the way across the organ. These layers of tissue are flat, and always at the same distance from one another. Each layer contains about 20 electroplaxes, the outlines of which present a very irregular or scalloped appear- ance. The electric tissue in which the electroplaxes are im- bedded is in appearance a Jjelly-like or mucous-like tissue, usu- ally known as electric connective tissue, and which I have shown in the course of my work to be of the same composition as white fibrous connective tissue. The nerve and blood supply runs in the above tissue. The general form of a vertical section of an electroplax is shown by figure 1, which is a drawing of part of a section of a single electroplax. Each electroplax is composed of three principal layers, a nervous or electric layer which forms the upper surface, a 1 Anat. Anz., Bd. 29, S. 387, 1906. 97 98 JAMES G. HUGHES, JR. middle layer, and a lower or nutritive layer which along with the middle layer is evaginated into a large number of long papillae. All three layers are deeply marked with a dense series of fine striations, which are peculiar to the electroplaxes of several other fishes. The upper or electric surface is flat and smooth and receives the nerve endings. The current of electricity runs downward through the organ which produces it, and thus the nerve endings in accordance with Pacini’s law are found on the negative pole of the electroplax. Proceeding directly to the subject of this paper, we may say that one of the most interesting of the points noted in the elec- troplaxes, when properly fixed and stained with iron hematoxy- lin, is a series of rod-like or thread-like objects running hori- zontally in the electric layer, among, above and below the nuclei and without any apparent connection with them (figs. 1 and 2).? These rods are of various sizes and shapes, and in form are said to resemble the classic thunderbolts seen in the hand of rep- resentations of Jove. They usually taper slowly and branch extensively at one or both ends. Some of these branches some- times seem to be mere lines, while others are wide and heavily pointed; at their other ends the rods are usually rounded; this latter appearance may be due, however, to the cut ends of the rods, for as noted above they sometimes branch at both ends. Some are short and heavy in appearance while others are long and thread-like. Peculiar looping, twisting, or knot-like bends are sometimes found at points on the longer rods. The out- lines and contour of these rods are always smooth. Their size may vary from thick or thin rods of over 300 u in length down to small ones that do not exceed 1 uw. In those electroplaxes where the rods are few they sometimes lie parallel and point in a definite parallel direction, while in others where the rods are very numerous they do not seem to have any definite arrangement. In this latter condition the rods present a very Wavy appearance. Their form may be seen in figures 1 and 2, which are drawings of the electric layer of an electroplax when * All the figures are drawings of sections of electroplaxes of Astroscopus guttatus. ‘OOZI X «= AaAB] OLAQOOTO 9Y4 UT Burd] 1998 O1B SPOT SUMOYS ST ONSST} OATZOOUUOD OI1}O9TO OY} SU [JOM Sv xB[doIQd9]9 9[OYM oT, “xvidoryoeja ue jo yared jo uoTZdeS [BOTFIOA JT “BI 100 JAMES G. HUGHES, JR. Fig. 2 Horizontal section through an electroplax. Only the electric layer is seen; a, an extensively branching rod; b, fine branches of the above rod; c, a large characteristic loop in a rod; d, acut end of a rod; e, nuclei of the electric layer. XX 1200. stained with iron hematoxylin. These drawings show the electric layer in which the rods are found in horizontal section (fig. 2) and in vertical section (fig. 1). The purpose, function, and chemical composition of these rods have been previously unknown to histologists. In order to determine anything in respect to their function or purpose, a knowledge as to the class of organic substance to which they belong, whether muscle, connective tissue, nervous, or chitinous, ELECTROPLAX OF ASTROSCOPUS 101 and also a rough knowledge of their chemical composition is imperative. The contour and form of these rods as they appear under the microscope resemble both smooth muscle fibers and fibers of elastic connective tissue. The belief that the function of these rods was somewhat of the nature of support for the delicate substance of the electroplax, and the fact that their form resembled connective tissue fibers led the writer to take for one of his first hypotheses, that they were of some form of connective tissue, and to perform accordingly the following series of experiments. As the most logical and best way for determining the kind of connective tissue, if any, of which the rods might be composed, a number of stains used by other investigators to identify similar substances were applied and the results noted. Controls were used on known tissues. Before taking up the connective tissue stains, however, a description of the results from the iron hematoxylin staining is now noteworthy; the material used being fixed in pure corrosive sublimate. The jelly connective tissue stained a very light gray. The nutritive, striated, and electric layers stained a much darker gray. The nuclei in all the layers stained somewhat black. The pointed fibers, and rods found in the electric layer stained a deep black, thus being clearly differentiated from the surround- ing cytoplasm. In some electroplaxes they were very numer- ous, while in others the number was rather small. ' The connective tissue stains applied as follows: (1) Mallory’s connective tissue stain, using the modifica- tion given in Lewis’ ‘‘Text-book of histology.’’ White fibrous connective tissue should stain blue in this medium. Paraffin sections of the electroplaxes fixed in corrosive subli- mate were stained for 10 to 12 minutes in a 1 per cent aqueous solution of acid fuchsin. They were then transferred directly to a stain consisting of 0.05 grain of aniline blue (soluble in water) and 0.2 grain of orange G dissolved in 100 ce. of a 110 per cent aqueous solution of phospho-molybdic acid. In this they remained from 2 to 3 minutes. They were then rinsed in distilled water, dehydrated rapidly, cleared and mounted. 102 JAMES G. HUGHES, JR. A description of an electroplax as seen under the 2 mm. oil immersion lens is as follows: The white fibrous or electric con- nective tissue stained a light blue or purple. The electric nutri- tive and middle layers or the electroplax proper stained a red- dish purple, and the nuclei as a whole in all of the layers, stained somewhat lighter than their surrounding cytoplasm. The peculiar rods and fibers stained a deep red, and were thus clearly differentiated from the other elements of the elec- troplax. They were very numerous and as noted above their outlines were always smooth. Blood corpuscles lying in the jelly connective tissue, of which there were only a few, stained a brilliant red, much the same as the rods. The white fibrous or jelly electric connective tissue (be- tween the electroplaxes), as noted above, stained a light blue. The pointed fibers and rods stained a deep red. This would indicate therefore, that these rods are not composed of white fibrous connective tissue. It would seem also that they are not muscle for they stained a different shade of color from the rest of the electroplaxes, which I have found in the course of my work to stain much the same as muscle. (2) Ven Gieson’s connective tissue stain, in which white fibrous connective tissue should stain red: Paraffin sections of the electroplaxes fixed in corrosive sub- limate were stained for 4 to 5 minutes in a 1 per cent aqueous solution of hematoxylin. They were rinsed in distilled water and transferred to a stain consisting of a saturated aqueous solution of picrie acid containing 20 per cent acid-fuchsin. They remained in here 15 to 20 minutes and were then rinsed, cleared. and mounted. The white fibrous connective tissue layer stained pink. The three layers of the electroplax stained brown. The nuclei and rods stained the same color, that is, brown; and, while the nuclei could be seen with difficulty, the rods were scarcely visible owing probably to their similar refractive in- dex. The fact therefore that the rods did not stain the same color as the white fibrous connective tissue, which stained pink, indicates again that they are not composed of white fibrous connective tissue. By Van Gieson’s stain, therefore, I have ELECTROPLAX OF ASTROSCOPUS 103 confirmed the evidence as presented by Mallory’s stain in re- spect to the composition of the rods; that is, they are not white fibrous connective tissue. I now undertook to apply stains which were tests for elastic fibers, and the first under this head is Weigert’s resorcin-fuchsin stain in which connective tissue according to Weigert and other writers stains dark blue. The stain was made up as follows: One per cent of basic fuchsin and 2 per cent of resorcin were dissolved in water; 50 ec. of the solution were raised to the boiling point, and 25 cc. of liquor ferri sesquichlorate P.G. were added and the whole boiled with stirring from 3 to 5 min- utes; a precipitate was formed. After cooling, the liquid was filtered, and the precipitate which remained on the filter was boiled with 50 ec. of 95 per cent alcohol. It was then allowed to cool, filtered and the filtrate made up to 50 ce. with alcohol, and 1 ec. of hydrochloric acid added. Paraffin sections of the electroplaxes fixed in corrosive sub- limate were stained for 6 hours in the above alcoholic solution of the precipitate. They were then washed in 95 per cent alcohol, dehydrated quickly, cleared and mounted. We Win a NA al Bh 6 ( WN NLS, aaa Ie B ee. <— > <—. w 16) ] Sate ales DRED — ————_? FIT TTLLTLL DZD TT === = ee ae 141 Wy eo. ‘Oy > m.) < ot ] Pay M ; nyt : = ri ies / \ . y Bk | ae ' eet st ss . i 'S 4 THE EMBRYOLOGY OF BDELLODRILUS PHILADELPHICUS GEORGE W. TANNREUTHER Zoological Laboratory, University of Missouri TWENTY-SIX TEXT FIGURES AND EIGHT PLATES CONTENTS ]ianigeOde eC So odie oletete epee echo oece Oa eee ORR MEI dic.o rcs. ao do! 144 INSRHOHL IMHO doce reas es oo ce ete SEL aoa eine wter’s ocd on casts 146 PEEL Ol nen Mare lODMEMe fc steak e aeais e+ » fad vie s Se aus be ene meat 147 Rant CRP eMm ee TN eee a fer te tc os a 8G ag areleln hea teh re 149 i wesinanionuror cleavare Cells. oo: Ah. 0b... 6. sak .noe baibe esd are eee 149 Ply MeSIOMCleaVARE es. we won iste a8 6.0 cio a4 4 aay ee cece ge ene 149 DOM Men peTrOGFOtClEAVAGER naam ve. Ps cus Gard eleeu seers See tane clenats 184 (Gir Me is 2 ech arog Ey ORS see SCREEN, eee he Al 186 ACCOMM ALAMIVe SiG yO1iGiiheremt: fOPMS ss)... os « oh ed oe. ele eee See aes 187 Pee Minewtirs eS OMMAUODIASUMA tsar & sish sc sectors sao sect 1.9 coe aeons eennyeenotee att 190 Eu SSECOMCT GONNA OULAGiE . sare fc shee.2 Ses os «oh be Dae Rieter e Reece ts a 193 3. Variations in the method of mesodermal formation.................... 199 Se GCOsMeSO OLAS tee ae elie Bic 2 ur da ee eit ones tae eee ee 199 EO OEIG-OIESOOPAS beara. ales cealte Sei sick Gs cedie hates «oy Meena oe ee ee ae 199 PVAniahons mri ne source, OlemlvOC@erm: 2. 4- ae yee ees aes: 205 General adaptation and interpretation of cleavage...................+-.25. 207 iene hexcleavaceroins GelloGniluse. . oiaccc ati ocsot-tie omelets Gee be cis ees eee 208 rlitei nec lear ewOleOtMenrtOGIMlS! ce. cvoe x tele atest dae ees Geese eens ea 212 ene alles UUMMTT NEN Tay ee eae Meera scom in lieie ahs aise. scosa.at a checha haan ae es A Oueles aa vliccche sayesintelt Ss NG 213 LL di GLEDAUTONRS HOM UENG | Se aa eee OR ee a oe MLN epta eRe Sturn Sc a cis cers A 215 143 JOURNAL OF MORPHOLOGY, VOL. 26, NC. 2 JUNE, 1915 144 GEORGE W. TANNREUTHER INTRODUCTION The Discodrilidae are, in many ways, extremely favorable for the study of annelid development. The material can be readily obtained at almost any season of the year. The develop- ment of any one egg can be followed from the time of fertiliza- tion to its complete development. The smallness of the eggs and the chitinous-like cocoon are the most objectionable features to contend with. Notwithstanding these facts, this group of annelids has been almost completely neglected by investigators in the study of cell lineage. Moore, in his paper on “The anatomy of an American Dis- eodrilid (Bdellodrilus illuminatus),” refers incidently to his ‘embryological studies’ in the course of his investigations, but has published no account of the Discodrilid development. Salensky, in his paper on “The development of Branchio- bdella an European Discodrilid, parasitic on Astacus fluvitalis,”’ gives an account of the cleavage, axial relation, origin of the germ layers and the formation of the adult structures. But the points of chief importance are so inadequately described and imperfectly worked out, that his results have no special signifi- cance in the study of Discodrilid development. The development of Bdellodrilus philadelphicus, one of the Discodrilidae, has, so far as I can learn, never been worked out. It is in many respects a very important form, not only from the standpoint of development, but from its adult anatomical strue- ture as well. The more important points in the following paper may be briefly summarized as follows: 1. The cleavage is unequal and regular, but may be variable in some eggs. A very small cleavage cavity is present. The gastrula is formed by the epibolic process. The blastopore occupies a very small area on the ventral surface. Its point of closure corresponds to the median ventral side of the adult worm. 2. The early cleavage planes are definitely related to the future organs of the adult worm; i.e., every cleavage foreshadows EMBRYOLOGY OF BDELLODRILUS 145 the position of some definite future formation. The large macro- mere D after the formation of d* divides very unequally, the smaller cell becomes the entomere D and the larger cell becomes the mesomere d‘ (M). The entire mesoblast is derived from the large cell M after its equal cleavage. The primary meso- blasts M, M, or meso-teloblasts, are completely grown over by the derivatives of X and the cleavage cells of the third genera- tion of ectomeres. The descendants of the primary mesoblasts are differentiated into two distinct groups of cells. The first group becomes the dorsal mesoderm of the adult worm. The second the mesodermal germ bands, becomes the ventral and lateral mesoderm. The cells of the first group remain undif- ferentiated until late development. The latter becomes differ- entiated into muscle tissue much earlier than the former. 3. When completely formed, the germ bands consist of three distinct strata of cells: (a) An outer stratum, ectoblast from one to two cells thick, which is produced by the three generations of ectomeres. This layer persists as the definitive hypodermis and secretes the cuticle; (b) A middle stratum, which gives rise to the nervous system and the nephridia; (c) An inner stratum, mesoblast which gives rise to all of the mesodermal elements, blood vessels, septa, reproductive organs, etc. 4. The middle stratum is composed of eight distinct longitudi- nal rows of cells, which at first lie at the surface and form part of the general ectoderm (ectoblast), but afterwards become completely covered over by the ectoderm. There are four rows in each germ band, terminating at the posterior end in a large cell or teloblast. The inner or ventral neural row on each side gives rise to the corresponding half of the nervous system. The three remaining rows of cells (nephridial) on either side, give rise to the nephridia. The outer nephro-teloblast often proliferates but few cells. 5. The brain or cephalic ganglia take their origin from the extreme anterior ends of the neural rows and are distinctly in- dependent of the ectoderm. 6. The cleavage of the entomeres A, B, C, and D is con- tinued to the end without delay. The entire digestive tract, 146 GEORGE W. TANNREUTHER with the exception of the very short stomodaeum and procto- daeum is derived from the entomeres. The proctodaeum is on the dorsal side of the tenth segment. The stomodaeum is formed at the apical pole. The embryo is completely turned on itself, 1.e., the extreme anterior and posterior ends are in Immedi- ate contact. The outer or curved surface, becomes the ventral side of the future adult worm. NATURAL HISTORY The Discodrilids occupy rather a unique position in the annelid group. They resemble the Hirudinea in their parasitism, in their general shape, in the presence of an anterior and posterior sucker and in the existence of chitinous Jaws. The last char- acter is not found in any other oligochaet, but occurs in a large number of leeches. These facts, perhaps not important in them- selves, are indications of a very close relationship between the Discodrilids and the Hirudinea, a group which they approach, not merely in such habits as the formation of the cocoon in which the eggs are enclosed, but in many other points of internal and external structures. The fundamental differences between the two groups are not numerous and are not of such importance as has been assigned them by different writers. The Discodrilids are classified as a distinct family of the Oligochaeta. Bdellodrilus philadelphicus occurs very abundantly on Cam- barus virilis, especially in the early spring and summer months. A few may persist throughout the entire winter in their natural habitat. For convenience, the animal may be divided into three dis- tinct regions; the head (pharynx), the body proper, and the postericr sucker. The head is much broader than the anterior body segments. The head is composed of four distinct annuli, which perhaps represent distinct segments. The first or peri- stomal annulus is divided into very mobile dorsal and ventral lobes or lips, which exhibit slight median emarginations, but are otherwise entire. It has sensory hairs and papillae, which are common in this family. The fourth annulus is very narrow. EMBRYOLOGY OF BDELLODRILUS 147 The middle two appear as muscular rings. The chitinous jaws are triangular, the dorsal with a single tooth, the ventral jaw with a pair of smaller teeth. No lateral mucous glands which are very common in some of the species are present. The body proper consists of eight strongly bi-annulate somites or rings. The anterior somites are longer and broader than the posterior. When contracted, the minor annuli of the somites are telescoped within the major annuli. The fifth, sixth, and seventh somites are sexual. The first, second, third, fourth and eighth somites are nephridial. The spermatheca is broad, thin walled, and nearly cylindrical. The penis is carried to the exterior by the eversible bursa, into which its projecting end is received. There is a conspicuous prostate in addition to the large glandular sperm sac. These parasitic forms remain attached to the ventral surface of the host throughout their entire life history. The eggs are deposited on the ventral sur- face of the host, more abundantly where the water is kept in constant motion by the movement of the appendages. Each egg is enclosed in a distinct separate stalked cocoon. The base of the stalk is firmly attached to the host. The deposition of eggs occurs during the entire year, if the parasites be kept in aquaria at room temperature. In their natural habitat eggs are not deposited during the severe winter months. BRIEF OUTLINE OF DEVELOPMENT The cleavage of the ovum takes place with considerable pre- cision and regularity. Especially is one impressed with this striking phenomenon, after following the cleavage of many ova. The only perceptible variations being (a) slight differ- ences in the time at which the individual cells divide; (b) slight variations in the size of the same cells in different ova. The rate of cleavage varies somewhat with temperature. Occasion- ally all the cleavage cells of an individual egg are nearly equal and it is impossible to orient the embryo before the germ bands begin their formation. This, however, is an exception, rather than a usual occurrence. 148 GEORGE W. TANNREUTHER As development progresses the variations between individual embryos become less apparent and as far as can be recognized, do not affect the final result. The history of the cleavage is distinguished by three well marked periods, namely: oblique, transitional, and_ bilateral. In the first period, which extends to the twenty-four-cell stage, the germ layers are differentiated, and the parent cells, which give rise to the future organs are definitely marked out. The first cleavage is nearly transverse to the median longi- tudinal axis of the adult worm. The second cleavage plane occurs at an angle of forty-five degrees to the first. The third cleavage plane is horizontal and separates the four ectomeres above from the four macromeres below. Three generations of four ectomeres each are successively separated from the macromeres A, B, C and D. The first gen- eration of ectomeres (a!, b!, c! and d!'), are formed in a right handed direction. The second generation of ectomeres (a2, b?, ce? and d?), are formed in a left handed direction. The third generation of ectomeres (a’, b’, c? and d*), are formed in a right handed direction. From these twelve ectomeres the entire ectoderm is formed. The ectomere d? gives rise to all or nearly all the ectoderm of the trunk, to the nervous system and to the nephridia. The oblique type of cleavage is maintained in the division of macromeres. At the close of the oblique period the embryo consists of twenty-four cells (text fig. 6 and fig. 36). The rela- tion of the cleavage cells to the germ layers is as follows: A er macromeres..... CNUOGERM ee esi 4 ok ee Ile Re Sere mesomere....... MES OGE RITE ores eR re er ens foties lec) coy se é IGLOE tte ectoderm PAE Se Oe roe me ectomeres. shes i 1 (d?)..ectoderm, nervous system, nephridia Bilateral division now occurs in some of the ectomeres, while others may continue to divide obliquely. The transitional period shows both types of cleavage. Oblique cleavage per- sists in some of the cells until the fiftieth or more cell stage. In the third period, the cleavage becomes essentially bilateral EMBRYOLOGY OF BDELLODRILUS 149 in most of the cells and the teloblasts of the right and left halves of the embryo are formed. Bilateral symmetry now becomes definitely established and the animal increases in length very rapidly. CLEAVAGE 1. DESIGNATION OF CLEAVAGE CELLS In the designation of the cleavage cells, for the sake of uni- formity and convenience, I have for the most part adopted the system followed by Wilson in his ‘‘Cell lineage of Nereis,’ and Lillie in his study on ‘“‘The embryology of the Unionidae.” The first four cells (macromeres) are designated by the capital letters A, B, C and D. The generations of micromeres (ectomeres) by the small letters a, b, c, and d. The first index number indicates the generation to which the ectomere belongs. Thus al, b!.2 or c!1-2 or d! 4 all belong to the first generation; c?, b?,d?.3 belong to the second generation, etc. A, B, C and D corre- spond to the vegetal pole; a, b, c and d to the apical pole. When a cell divides the products receive the designation of the parent 2,1 cell with the addition of a further index number; thus b? i és Exceptions to this rule are made only in the case of special cells, which, for convenience, receive special designations: thus d? of the second generation of ectomeres becomes the ‘first soma- toblast’ and is designated by (X), and its small derivatives by x', x?, x3, etc.; d‘ the ‘second somatoblast’ is designated by (M). 2. TYPES OF CLEAVAGE a. The oblique period of cleavage: one io twenty-four cells First cleavage: The first cleavage occurs about five to ten hours after the deposition of the egg. The time varies somewhat with external conditions. The plane of division passes through the area where the polar bodies are formed (fig. 1) and divides the egg into two very unequal parts, AB and CD (text fig. 1 and fig. 2). The smaller of the two cells AB is anterior, and GEORGE W. TANNREUTHER 150 UL (sejqoin wx y in UMaPO PUD WNAC) uapopug. uy) -—_ : Wuaspo Pu | ir a Wap oar ite poe V (ett) oie p | ae | HEN sabpana),) T WIaVL EMBRYOLOGY OF BDELLODRILUS pot 3b As Fig. 1 Two-cell stage from apical pole view. Fig. 2. Early three-cell stage from apical pole Fig. 3 Four-cell stage from apical pole view. Fig. 4 Four-cell stage from vegetal pole. a-b, median longitudinal plane of future adult; /-1, first cleavage plane; 2-2 second cleavage plane; a, anterior; b, posterior; 7, right side; /, left side. the larger cell CD is posterior. The cleavage at first is very deep and the cells are rounded, but soon they begin to press against each other and flatten at their point of contact. Before the second cleavage begins the egg assumes its original ellipti- cal shape and the point of contact externally, between the two cells is represented by a mere line or shallow groove. No actual fusion of the two cells ever takes place; sections always show a distinct line of separation between them. The deutoplasm is equally distributed in both cells. The cytoplasm surrounding the nucleus contains very little yolk 52 GEORGE W. TANNREUTHER material. This makes it possible, not only to recognize the posi- tion of the nucleus, but to be able to make out the exact position of the cleavage spindle in the living egg. Second cleavage: The second cleavage is meridional and takes place at an angle of forty-five degrees to the median plane of the future adult. The two cells divide at different times (occa- sionally both cells divide simultaneously). These two cleavages taken together represent the second cleavage in otherannelids. CD divides first into two very unequal parts (text fig. 2 and figs. 3, 78). The division of AB is nearly equal (figs. 5 and 6). The largest cell, D, is posterior. B is anterior, inclined a little to the right. C is right (text figs. 3-4) and A is left with refer- ence to the median axis of the future worm. The large cell D has a tendency always to divide first. The exact formation of the four macromeres must be carefully worked out, and correctly understood, since their position largely determines the orientation of the future organs. For descriptive conveniences the region of the first generation of ectomeres will be considered as the upper or apical pole and the point directly opposite, as the lower or vegetal pole. The centers of the upper and lower poles of the dividing ovum, coin- cide with the median longitudinal plane of the adult worm. The poles however may be shifted somewhat anteriorly or pos- teriorly, with reference to the macromeres, more especially to D in the formation of d?. When viewed from the upper pole A and C are in contact, while B and D are separated. But when viewed from the ventral pole A and C are separated and B and D are in contact (figs. 6-7). This extensive cross furrow found at the vegetal pole is also present in forms like Nereis, Clepsine and Crepidula; while in those forms like Unio, in which the greatest mass of the four macromeres is concerned in the formation of ectoderm instead of endoderm, the cross furrow is greatest at the animal pole. These cross furrows (‘Brechungslinie’ of Rauber) have no special significance in the development of Bdellodrilus, as the cleavage of the macromeres is carried to the end, immediately after the three generations of ectomeres are formed. In those forms like Nereis, where the cross furrow DESCRIPTION OF PLATES All drawings were made with a camera lucida under a magnification of about 125 diameters. All whole amount drawings, with one or two exceptions, were made from the living egg. The variation in size of the surface views is due to a difference in the size of the eggs. The sections were not uniformly magnified. Stippling has been adopted for the sake of clearness. REFERENCE LETTERS a., anterior mi.a., Minor annulus bl., blastopore mj.a., Major annulus c., ciliated cells mo., mouth c.c., cleavage cavity p., posterior c.l., cerebral lobes pb', polar bodies coe., coelom ph., pharynx €.cp., egg capsule pr., proctodaeum ec., ectoderm so.mes., somatic mesoderm en., entoderm sp.mes., splanchnic mesoderm gn., ganglia st., stomodaeum mes., mesoderm v., ventral A, left macromere B, anterior macromere C, right macromere D, posterior macromere a!, b!, ce, d!, at!, ete., first generation of ectomeres a”, b?, c?, d?, a2-1-1, ete., second generation of ectomeres a3, b3, c3, d3, a3-1, etc., third generation of ectomeres at, b4, c*, d+, ete., fourth generation of micromeres X = d? first somatoblast X, X, right and left proteloblasts X@, neuroblast X(2), X(3), X4), nephroblasts M = d* second somatoblast m, secondary mesoblast x!_x’, small derivatives from X N, posterior end of nephric rows Ne, posterior end of neural rows nc, neural rows np, nephric rows 153 PLATE 1 EXPLANATION OF FIGURES 1 Surface view of an unsegmented ovum, to show the polar bodies and the cleavage nucleus. 2 Two-cell stage from the upper pole. 3 Two-cell stage; cell CD dividing. 4 Three-cell stage from the upper pole; division of CD is complete. 5 Same stage as preceding; cell AB dividing. 6 Four-cell stage from the upper pole; the cleavage spindle for the first ectomere forming. 7 Four-cell stage, ventral view. 8 Four-cell stage viewed from the left side. 9 Same as the preceding, viewed from the right side. 10 Four-cell stage from the upper pole, showing the formation of the first ectomere. 11 Five-cell stage from the upper pole, d! formed. 12 Stage showing the cleavage spindle of the second ectomere. 13 Six-cell stage from the upper pole; cleavage spindles for third and fourth ectomere forming. 14 Hight-cell stage from the upper pole; the macromeres are considerably flattened. 15 Same as the preceding, from the ventral pole; similar to the same view of the four-cell stage. 16 Eight-cell stage from the left side. 17 Same as the last showing the behavior of the macromere D in the forma- tion of the first somatoblast. 18 Stage a little later than the preceding. 19 Nine-cell stage from the left side after the formation of the first somato- blast. 20 Nine-cell stage turned a little to the left, so that all the cells are visible. 154 EMBRYOLOGY OF BDELLODRILUS PLATE 1 G. W. TANNREUTHER B Ly Kline, del. 156 GEORGE W. TANNREUTHER persists until late development, it serves as an unmistakable point of orientation. Figures 8 and 9 show the four-cell stage from the left and right sides. The dorsal ventral axis of A, B and C is about the same as that of D, but immediately after the formation of the first generation of ectomeres, the cells A, B and C shorten and become more rounded (fg. 16). In later stages of development these cells often become very much flattened and cause the developing embryo to appear unusually large, when viewed from the upper or lower poles. Third cleavage (eight-cell stage): In the formation of the first generation of ectomeres (d', c', b! and a'), each of the four macro- meres divide obliquely. The ectomere end of the cleavage spin- dle is uppermost. The macromere D divides first; d! is budded off from D towards the upper pole, in the direction of the hands of a watch (dexiotropic), (figs. 10-11). We have thus a five- cell stage. Each of the macromeres C, B and A next bud off a small cell towards the upper pole. These are not formed simultaneously, but in the invariable order ¢', b! and a'. Thus there occurs successively, a six, a seven and an eight-cell stage (figs. 11-14). In figure 13, an upper pole view, D and C have divided and A and B are preparing to divide. In both A and B the asters of the ectomere end of the spindles are visible. The position of the opposite end of the spindles are indicated by circles. This figure shows the oblique nature of the cleavage spindles. The spindle in A points to the space between A and B. The spindle in B points to the space between B and C. In figure 14, an eight-cell stage, the exact relation of the ectomeres and macromeres are shown as they normally appear from the apical pole. The position of the first generation of ectomeres is obvious. They suggest a possible rotation, after their forma- tion, through an angle of about forty-five degrees in the direc- tion of the hands of a watch. If actual rotation did occur there would be no difficulty in explaining their final position. But the fact that the cleavage spindles are oblique and the position of the completely divided nucleus can be definitely determined, before there is any indication of the cytoplasmic division of the EMBRYOLOGY OF BDELLODRILUS a parent cell in the formation of the daughter cells, suggest that the apparent rotation process is not mechanical or even pro- duced by pressure of the macromeres. The formation of d! in figure 10 shows how the process takes place; d! is budded off obliquely from the macromere D over the inner posterior edge of A and becomes partly imbedded in A. Its final position is determined by the direction of the cleavage spindle. This characteristic method in the formation of the ectomeres is quite a prevalent one. It occurs not only in the eggs of annelids, but in those of the molluses and polyclads as well. Fourth cleavage: A nine-cell stage is reached in Bdellodrilus by the division of the macromere D in an oblique direction. Figure 16, an eight-cell stage viewed from the left side, shows the position of the macromeres A, B and C with reference to the macromere D, before the formation of the ectomere d?. The large macromere D contains about two-thirds of the volume of the dividing ovum. In preparation for the formation of d?, D elongates in an oblique direction at an angle of about forty-five degrees to the horizontal plane of the developing embryo. The ventral anterior portion of D shifts forward beneath A, B and C (fig. 17). After the formation of d?, D takes a position directly beneath the first generation of ectomeres, and completely covers the inner ends of A, B and C (figs. 19-20). In some instances D is shifted more anterior and completely covers the ventral surface of the other macromeres (fig. 20); but in most cases, as in the nine-cell stage, D occupies the region of the ventral pole, directly beneath the first generation of ectomeres (figs. 22-23). The formation of d? is shown in figures 16 to 19. The division is equal in most cases. When unequal, d? is the larger cell. In figure 20, a nine-cell stage turned to the observers left so that all the cells are visible, A and B are preparing for the formation of a? and b?. In most the succeeding stages d?, the ‘first soma- toblast’ will be designated by the capital letter X. It is also the first cell of the second generation of ectomeres. The formation of a?, b? and c? is shown in figures 20, 21, 22 and 24 in side and top views respectively. The second generation of ectomeres, with the exception of d? (X), is about the same size as those of the PLATE, 2 EXPLANATION OF FIGURES 21 Nine-cell stage, viewed from the right side. 22 Same stage as the preceding, from the upper pole; spindles for the second generation of ectomere are forming. 23 The same stage from the ventral pole. 24 Twelve-cell stage from the upper pole; spindles for the third generation ectomeres are forming. 25 Thirteen-cell stage from the left side; the cell d? nearly formed. 26 Same stage as the preceding from the right side; the embryo is consider- ably elongated. 27 Fourteen-cell stage from the right side; X dividing to form x'. 28 Same stage as the preceding, ventral view. 29 Fifteen-cell stage from the upper pole; c* budded off symmetrical with d?. 30 Fifteen-cell stage from the left side; upper pole turned considerably to the left. 31 Same stage as the preceding from the right side; the small cell x! is drawn out between c* and X. : 32 Fifteen-cell stage, ventral view, as a transparent object, with the position of all the cells indicated; drawn from a prepared specimen, cleared in xylol. aan) O EMBRYOLOGY OF BDELLODRILUS PLATE 2 G. W. TANNREUTHER aoe. aa Kline and Tannreuther, del. 159 JOURNAL OF MORPHOLOGY, VOL. 26, NO. 2 160 GEORGE W. TANNREUTHER first. The order of their formation is d? (X), c?, b?, a2, the same as the first generation. Figure 24 illustrates the twelve-cell stage from the upper pole. The cells are somewhat flattened. The macromere D is located a little to the left of the median longitudinal plane, while X is symmetrical with reference to the median axis of the future worm. The cleavage spindles of the third generation of ectomeres form in an oblique direction. The thirteen-cell stage is reached by the formation of d*. The manner in which d? is formed, is rather unique when we take into consideration the size and position of D with reference to the other macromeres A, B and C (figs. 28-25). It is budded off from the outer surface of D and takes up a position symmetrical with c*?. The fourteen- cell stage is reached by the formation of x! (fig. 27), it is budded off from the median right side of X. Its final position is between X and C. Figure 28 shows the same stage as the preceding in ventral view, turned a little to the observer’s right. Immediately after the formation of x! the macromere C buds off c?, thus making a distinct fifteen-cell stage (fig. 29). The order of formation of the third generation of ectomeres is the same as the first and second. Figures 29, 30, 31 and 32 show the fifteen- cell stage in dorsal, left, right and ventral views respectively. In figure 33, a® and b* are formed and the first generation of ecto- meres are preparing to divide; d! and ec! divide first; at the same time x? is budded off from X, symmetrical with x!, between D and d’, thus making a twenty-cell stage (fig. 34). Figure 35 is a side view of a twenty-two-cell stage after. a! and b! have divided. This figure represents an anterior posterior elongation of the embryo, which is a very common occurrence. ‘The cells, taken as amass, are very plastic and may assume different shapes. This peculiarity is only secondary and has no special significance. The cells become more spherical before division and flatten out somewhat after the division is complete. The twenty-three-cell stage is reached by the formation of x® from the upper posterior side of X between ¢! and d! (fig. 36). The division of the first generation of ectomeres is unequal and radial rather than oblique. From the twenty- to the thirty- EMBRYOLOGY OF BDELLODRILUS 161 cell stage several types of cleavage are present; oblique, radial and bilateral. This period of variable cleavage will be desig- nated as the transitional period. b. The transitional period of cleavage: twenty- to thirty-cell stage After the formation of x® there is a short inactive period and in many of the developing embryos, the cleavage furrows become very indistinct. Cleavage is again initiated by the formation of d‘ from the large macromere D. The cleavage is oblique and very unequal (text fig. 5 and fig. 85). The smaller cell is almost completely hidden when first formed. It is budded off directly between A and B, near the ventral anterior surface (fig. 37). The smaller cell persists as D (ento- mere) and the larger cell d‘ or M becomes the ‘second somato- blast.’ After the formation of M the entire endoderm is contained within the entomeres A, B, C and D (figs. 37-88). The germ layers are now distinctly separated and the em- bryo at this stage of development is composed of twenty-four cells (text figs. 6-7). Nereis at the same period of differ- entiation, shows thirty-eight cells. Unio (Lillie) at the time of the separation of the germ layers contains thirty-two cells. This difference is due, in case of Bdellodrilus to the lagging of division in the cells of the upper pole. The composition of the embryo at the twenty-four-cell stage is as follows: Bmbomeresa.2..7 2.22. 0 Ney) Bi Orbe Be iy ete ae RRS Beware RAC <2 Sd 4 Hetomeress.- 4454.0. Orimsincenenatlom:. ....2 2% Sed) oes ene 8 Hetomeres ye 22.42) S (Olt MSCOMCl mys HMOle se nebo ancanacdodesodes 4 Befommeresie x 4.211..." Op aindeeenerAtlons... 0.02 s4e Se eee eee 4 Teron ul Mets ered pacer Mee al LL meee ar Res: HA is says dad in eee eee ene 1 EUCA eS OnIALOOLAStIGeRUVATI VES: «... as. os. id ste saat hese oaeinelae ees 3 24 Many of the cells during the transitional period have a defin- ite shape and if separated from the cell complex, they could be readily recognized. The embryo at this stage of develop- ment is somewhat spherical (figs. 36-38). Immediately after 162 GEORGE W. TANNREUTHER Fig. 5 Twenty-four-cell stage in ventral view, to show the division of the large macromere D. The larger of the two cells d‘ (M) becomes the ‘second somatoblast.’ The smaller cell D, becomes one of the four entomeres; D is scarcely visible from the exterior. Fig. 6 Same as the preceding, in an apical pole view. Fig. 7 Twenty-four-cell stage, ventral view, shows the bilateral division of M. D after the bilateral division of M becomes more distinct from surface view. Fig. 8 Horizontal section of an early embryo to show spindles in the forma- tion of x® from either proteloblast X, X. Fig. 9 Horizontal section little later than preceding, to show the small cell x!-x®; apical pole view. EMBRYOLOGY OF BDELLODRILUS 163 the establishment of the germ layers, the bilateral division of the ‘first and second somatoblasts’ occurs. The bilateral divi- sion of the ‘second somatoblast’ usually precedes that of the first; occasionally they divide simultaneously. c. The bilateral period of cleavage: twenty-five-cell stage The first bilateral cleavages occur in the first and second somatoblasts (text fig. 7 and figs. 37-43). The small super- ficial cells of the lower pole are derived from the second and third generation of ectomeres and from the derivatives of X. The arrangement of these cells with reference to the blastopore is shown in figure 42. The entomeres A, B, C and D are partly grown over by the other cells and the open space becomes the blastopore. It is bounded anteriorly and laterally by small cells from the second and third quartettes, and posteriorly by the primary mesoblasts M,M. Its hinder lip, which is formed by the primary mesoblasts, lies anterior to the center of the lower pole. The closure of the blastopore takes place by a con- vergence of the cells from all sides. The principal growth of cells is from in front backwards, formed by the derivatives of the second and third generation of ectomeres (figs. 42, 48, 51). The entomeres now divide very rapidly and the cells soon be- come smaller than those of the ectomeres, which grow over them (figs. 43, 51). 3. THE FIRST SOMATOBLAST The history of the ‘first somatoblast’ in Bdellodrilus is of considerable interest when considered from the standpoint of its origin and its derivatives. When first formed from the posterior macromere D, it contains one-third of the entire bulk of the developing embryo. As already described, it first buds off the small cell x' on the right, x? symmetrically on the left and a third cell, x°, on the median posterior upper side. These three small cells are symmetrical with reference to the median longitudinal axis. The fourth cleavage divides the somato- PLATE 3 EXPLANATION OF FIGURES 33 Seventeen-cell stage from the upper pole; spindles in the first generation of ectomeres forming. 34 Twenty-cell stage from the upper pole; cells c'!, d'~! and x? just formed. 35 Twenty-one-cell stage from the right side; b!!, new cell; this embryo is unusually elongated. 36 Twenty-three-cell stage from the upper pole; d' and x*, two new cells formed; embryo considerably flattened. 37 Twenty-four-cell stage from the ventral side; the unequal division of D has just occurred; D partially visible; the cleavage spindle of M forming. 38 Twenty-five-cell stage, ventral view; division of M complete and the spindle for the first bilateral division of X is forming. 39 Twenty-five-cell stage from the left side. 40 Twenty-five-cell stage from the upper pole; embryo is nearly spherical. 41 Twenty-six-cell stage, ventral view. 42 Same as the preceding, ventro-anterior view. 43 Twenty-nine-cell stage, same aspect as preceding; a‘, b* and c4 are the three new cells formed. 44 Forty-two-cell stage from the upper pole; increase in number of cells due to the rapid division of the ectomeres. 164 EMBRYOLOGY OF BDELLODRILUS G. W. TANNREUTHER 4.2 XK Kline and Tannreuther, del. 165 PLATE 3 PLATE 4 EXPLANATION OF FIGURES 45 Sixty-three-cell stage from the upper pole. 46 Same stage as the preceding, from the ventral pole; very few of the ecto- meres are visible. 47 Seventy-cell-stage, ventral view. Increase in number due to division of small cells. Consult table of cleavage. First nearly equal division of the pro- teloblasts, X, X. This division separates the neural and nephridial elements. 48 Seventy-two-cell stage, ventral view. 49 Little later than the preceding stage, to show x® and x’. 50 Embryo, ventral view, to show the first division of the nephroblasts; transverse axis greater than the longitudinal. 51 Embryo, ventral view, turned anteriorly to show the blastopore. 52 Embryo to show the lengthening of the anterios-posterior axis. The small cells, x®, x®, are good points to mark the orientation of the different figures. All figures on plate are similarly orientiated with reference to the right and left sides of the embryo. 53 Stage a little later than the preceding. 54 Embryo from upper pole, to show derivatives of x® between the nephro- blasts. 55 Embryo with upper surface turned posteriorly, to showthe rapid division of the ectoblast cells. : 56 and 57 Show that either nephroblast X(2) or X() may divide, to form the three nephroblasts on either side. 166 EMBRYOLOGY OF BDELLODRILUS PLATE 4 G. W. TANNREUTHER x pilot pean ant Kline and Tannreuther, del. 167 168 GEORGE W. TANNREUTHER blast into two equal parts, right and left (figs. 40-41). These two cells, for convenience in description, will be called the pos- terior right and left proteloblasts. At the fifth division each proteloblast buds off a small cell, x‘, external to x! and x? re- spectively (fig. 44). At the sixth division each of the protelo- blasts buds forth a small cell, x° on either side of x*, beneath the derivatives of c! and d! (text fig. 8). At the seventh division each of the proteloblasts buds off a small cell, x® on the ventro- anterior edge at the junction of the two cells (fig. 46). At the eighth division each proteloblast, on either side of the median plane of the embryo, divided into two equal parts (figs. 47, 48, 93). The four cells formed become the posterior telo- blasts, X® and X®, right and left of the median axis (fig. 47). X‘? without any further division, becomes the neuroblast on either side, and X° becomes the nephroblast on either side of the median axis (figs. 47-48). Next X®, right and left, divides equally, giving X and X® (figs. 50, 52). Next either X or X‘* divides equally. If X divide, which is the common occurrence, we get X® and X™. But if X” divide instead of X, the final result is the same. In either case, we get four teloblasts on each side (one neuroblast and three nephroblasts) as shown in figures 56 and 57. Next X® on either side divides very unequally and gives rise to x7 on the anterior ventral outer surface (fig. 49). The progeny of the ‘first somatoblast,’ when the teloblasts are completely formed, is twenty cells. Table 2 shows the derivatives of the ‘first somatoblast.’ TABLE 2 x 3 KEXEX? ay pee Gab Gap -el/ V U wa DKK (a) ee Eb Sse!) EE Proteloblast Proteloblast EMBRYOLOGY OF BDELLODRILUS 169 4. THE SECOND SOMATOBLAST Immediately after the formation of X from the posterior macromere D, d? is budded off from D (figs. 24-25). Next D divides very unequally in an oblique direction and gives rise to d‘ (M) the ‘second somatoblast,’ as previously described. Fig- ures 37 and 85 show the position of D and M after the cleavage of the macromere D. The twenty-five-cell stage is reached by the bilateral division of M (figs. 37-88). The cells M, M at first are a little to the left of the median plane, but later in course of development they become symmetrical tothe longi- tudinal axis of the adult worm. Scon after the bilateral division of M, the ‘second somato- blast,’ each cell M,M right and left buds off five or more small cells directly beneath the first generation of ectomeres (figs. 86-87). It is impossible to detect these cells except by means of sections, hence the uncertainty as to their exact number. They are characterized by their large nuclei with homogeneous staining chromatin and they contain but little yolk material. These small cells divide once or twice soon after their formation from the primary mesoblasts and then remain inactive until late embryological development, at least until after the germ bands are completely formed and the embryo has undergone considerable differentiation (as the formation of the lumen of the digestive system, etc.). These undifferentiated mesodermal cells occupy the region which becomes the central dorsal side of the embryo, at the point where the developing worm is completely turned on it- self (figs. 92, 98-99). The history of these cells can be readily followed through their different stages of development, so that there can be no question as to their exact origin and history. When the embryo begins to straighten, the progeny of these small cells extend toward either end and form the splanchnic and somatic mesoderm on the dorsal side of the worm. This secondary mesoderm later becomes continuous with the primary mesoderm, which forms directly from the mesoblasts M,M. 170 GEORGE W. TANNREUTHER 5. THE ENTOBLAST The formation of the entoblast in Bdellodrilus represents an unusual type of development among the annelid worms. The macromeres A, B, C and D, after the formation of d‘, give rise to the entire entoderm. D is greatly reduced after the forma- tion of d‘. The position of the entomeres is shown in figures 40 and 42. In figure 40 A, B and C appear rather large, be- cause of the flattened condition of the cells. In figure 42 (from the ventral pole) the cells are rounded and appear more normal. The position of the entomeres and their boundary cells are dis- tinectly shown. This figure shows more clearly the bulk of the entoderm, when compared with the mass of the entire egg. In figure 43 (a twenty-nine-cell stage) A, B and C have divided nearly equally. This division is considered by some investi- gators as the formation of the fourth generation of micromeres; d‘ of the D quadrant has formed earlier. Figures 44 and 45 (apical pole views) show the upper outer edge of the entodermal cells. In figure 46 (the same stage as preceding from the ven- tral pole) a very small part of the entodermal and ectodermal cells are visible. This figure shows the prominence of the four large cells, which later form the ten teloblasts. These four large cells, from their position, resemble the four large entomeres, which are so prominent in many other forms. These cells (X, X, M, M), according to Selensky, share equally in the formation of the germ layers, i.e., ectoderm, endoderm and meso- derm are produced by each of them. In forms like Clepsine, Crepidula and others, at a similar or later stage of development, the entomeres are very prominent and the ectomeres with the first and second scmatoblasts, form a cap of cells on their upper surface. In Bdellodrilus the con- ditions are different. The ectomeres and the entodermal cells form a cap of cells on the upper anterior surface of X, X, M and M. This difference is due to the prominence of the first and second somatoblasts, which constitute the greatest bulk of the embryo. At about the seventy-cell stage the ectodermal and endodermal cells are nearly uniform in size (figs. 47-49). In EMBRYOLOGY OF BDELLODRILUS V7 figure 51 (a little later stage) the blastopore is nearly closed. This early closure of the blastopore in Bdellodrilus, is due largely, to the ventro-anterior shifting of the macromere D over A, B and C in the formation of the somatoblasts (figs. 18, 85). The closure of the blastopore, in some of the annelids, occurs at a very late stage of development. In Clepsine the telo- blasts give rise to rows of cells, which pass anteriorly around the entomeres A, B and C beneath the edge of the blastodise or cap of cells. The blastodise with these rows of cell cover about half of the entomeres. By the downward growth of the blas- dodisc and the concrescence of the germ bands, the closure of the blastopore is completed. The closure of the blastopore in Clepsine occurs on the ventral side, nearer the anterior end. In Bdellodrilus, the germ bands are not formed until later and take no part in the closure of the blastopore. Text figures 10 to 13 show the position of the ectoderm, entoderm, and the first and second somatoblasts, at different stages in the closure of the blastopore. The region of closure is similar to that of Clepsine. At the time of the formation of the secondary mesoblast just beneath the first generation of ectomeres, the entire ento- derm is situated in the anterior half of the embryo. But soon after the formation of the m cells (text fig. 12 and fig. 86), the entodermal cells by a rapid proliferation extend posteriorly between the m cells and the primary mesoblast. During the formation of the primary mesoblast, the meso-teloblasts them- selves are carried posteriorly, ahead of the entoderm. The entoderm, thus becomes situated between the m cells or second- ary mesoderm above and the mesoblast bands or primary meso- blast below. The entoderm in reality never reaches the posterior limit of the meso-teloblasts, as shown in figures 98 and 99. The interior of the developing embryo, now consists of a solid mass of small entodermal cleavage cells (figs. 95-99), heavily laden with yolk. These cells are readily distinguished from the surrounding mesodermal cells, by their deeper cytoplasmic stain. Figure 99 (a vertical longitudinal section near the median plane) shows the pcesition of the entodermal cells in the embryo. a2 GEORGE W. TANNREUTHER As the embryo elongates, the entodermal cells increase in number. This process of growth is continued until the digestive tract is completely formed. Figure 99 shows the anterior and posterior limits of the digestive tract, which is formed from the four en- tomeres. The anterior end shows a distinct lumen, while the posterior end is yet a solid mass of cells. The entire digestive tract, except the insignificant stomodaeum and proctodaeum, is entodermal in origin. The proctodaeum is not formed until the time of hatching. It occurs on the dorsal side of the tenth segment. The embryo is completely turned on itself (fig. 99) and brings the anterior and posterior ends of the digestive tract in close proximity. The differentiation of the digestive tract begins anteriorly and progresses posteriorly. As growth con- tinues, the outer cells of the entodermal mass form an epithelial layer. At first the cells are somewhat flattened, but soon take a columnar position, and form the columnar epithelium of the digestive tract. The cells within soon lose their staining proper- ties, break up and serve as food for the developing embryo. The digestive tract in its course of development, passes through the following stages: the first stage is represented by the four entomeres A, B, C and D; the second stage by a solid mass of cleavage cells (the cell boundaries are often very indistinct) within the center of the embryo; the third stage by an elonga- tion of the entodermal mass as the larva lengthens, and by the establishment of a lumen; the fourth stage by a thin layer of flattened epithelium and later a columnar epithelium; the fifth stage, the cells within the epithelial layer serve as food; and sixth the formation of the stomodaeum and proctodaeum. GENERAL HISTORY OF THE GERM BANDS The term ‘germ bands’ has been variously interpreted by differ- ent investigators on cell lineage. The term germ bands, or the German. equivalent ‘Keim Steifen,’ is usually restricted to the strata derived from the teloblasts, the ectoblastic layer being excluded. It is held by others that the germ bands of annelids are purely mesoblastic. EMBRYOLOGY OF BDELLODRILUS 173 Balfour, Hatschek, Goette, Kowalevsky and many others made use of the term ‘mesoblastic bands’ as the equivalent of the germ bands. In Hirudinea, according to Whitman, the germ bands are composed of three distinct layers; the ectoblastic, mesoblastic and the neuroblastic elements. Wilson gave the same interpretation in his studies on ‘‘The embryology of Lum- bricus.” In Bdellodrilus the term ‘germ bands’ includes the three strata of cells as in Hirudinea and Lumbricus. 1. INNER STRATUM OF THE GERM BANDS After the formation of the teloblasts, five on either side of the median axis (one neuroblast, one mesoblast and three nephro- blasts), the mesoblasts or meso-teloblasts are the first to begin the formation of the germ bands by a forward. proliferation of cells near the posterior lip of the blastopore (text figs. 12, 13, 17). The plane of division is nearly at right angles to the forma- tion of cells in the secondary mesoblast (text figs. 10, 18). The cells of the mesoblast bands are considerably smaller than the teloblasts from which they originate. They grow forward between the entoderm and the ectoderm and finally meet at the anterior end of the larva. As these bands grow forward they become several cells broad, but seldom more than two cells deep. Their differentiation begins anteriorly and progresses backward. The first cells of the mesoblast bands, when formed, are on the surface, but soon become covered by the ectodermal cells. As the mesoblast band extends forward below and around the entoderm, it forces its way to the extreme anterior end of the embryo beneath the ectoderm. It finally encloses the digestive tract on the ventral and lateral sides and becomes continuous with the secondary mesoblast on the dorsal side. The two mesoblast bands fuse first at the anterior end along the median, ventral side and subsequently with the dorsal second- ary mesoderm. In figure 99 (a longitudinal section) the meso- blast is differentiated into splanchnic and somatic layers, with the coelom between. The longitudinal muscles become differ- entiated before the circular. At the extreme posterior end the 174 GEORGE W. TANNREUTHER Figs. 10-13 Diagrammatic figures to show the ventro-posterior extension of the ectomeres, in the closure of the blastopore. Fig. 10 Thirty-three cell stage, taken a little to the right of the median plane. Fig. 11 Vertical section to the left of the median plane. Fig. 12 Vertical section of a ninety-cell stage. Fig. 138 Vertical section of an embryo at the time of closing of the blasto- pore. The mesoblast bands have just begun. The heavy stippling represent endoderm; the light stippling mesoderm, and the unstippled the ectoderm or ectomeres. m, secondary mesoblast; M@, meso- blasts; bl, blastopore; ble, point where the blastopore closes; X, derivatives of the ‘first somatoblast;’ mes, mesoblast bands. meso-teloblasts are represented by an undifferentiated mass of cells, which later give rise to the musculature of the last three segments of the worm, and are directly concerned in the forma- tion and movements of the posterior sucker. EMBRYOLOGY OF BDELLODRILUS ie 2. MIDDLE STRATUM OF THE GERM BANDS The middle stratum of the germ bands can readily be dis- tinguished while the embryo is still nearly spherical. Upon close examination it is seen that the ectoblast cells are arranged into four distinct rows, on either side of the median ventral axis (figs. 65-66). Each row terminates posteriorly in a large cell or teloblast. Text figures 15 and 17 and figure 58 show the early formation of these rows of cells. Sections of these various stages show that Fig. 14 Surface view from upper pole, to show the position of the ten telo- blasts. The meso-teloblasts or mesoblasts have budded off eight or ten cells in the formation of the mesoblast bands. Their position is indicated by dotted outline. The broken outline represents the region of the entoderm. The position of the ectoderm is indicated by a continuous line. Fig. 15 Third horizontal section from the top, passing through four of the large nephroblasts. The spindles represent the beginning of the first division in the formation of the middle germ band. The anterior end and the right side of the section are a little below the horizontal plane. Fig. 16 Seventh horizontal section from the top. It passes through the upper portion of the entoderm and the secondary mesoblast (m). JOURNAL OF MORPHOLOGY, VOL. 26, No. 2 176 GEORGE W. TANNREUTHER these rows of cells form a part of the general ectoderm, being partly covered here and there by adjoining cells. In later stages of development, these rows of cells become completely covered as they gradually sink beneath the surface, and thus come to lie between the mesoblast and the ectoderm or ectoblast. Fig. 17 Twelfth section from the top, to show the anterior extension of the mesoblast bands below and around the entoderm. Fig. 18 Nineteenth section from the top, to show the lower side of the meso- blast. The section passes below the entoderm. The spindles represent the beginning of the first division of the neuroblasts to form the neural rows. Heavy stippling represents entoderm; light stippling mesoderm and the unstippled portion the ectoderm. The sections of figures 15-18 were eight micra thick. There were 21 sections in all. a, anterior, and p, posterior, represent the respective ends of the cleavage cells, but not the future ends of the embryo; r, right; 1, left; ent, entoderm and ect, ecoderm. 3. OUTER STRATUM OF THE GERM BANDS This stratum forms the definitive ectoderm and needs no further description at this point of development. The embryo now elgongates very rapidly, and the general shape of the adult worm becomes recognizable. The telo- blasts become less and less distinct, until finally the cell rows terminate posteriorly in a group of small cells. The meso- teloblasts extend farther posteriorly than the neuro-teloblasts. New cells are always formed from the anterior surface of the teloblasts. There can be no question as to the origin of the germ bands from corresponding teloblasts, as their formation EMBRYOLOGY OF BDELLODRILUS ETe can be followed step by step. The mesoblastic and neuroblastic portions of the germ bands can be traced to the anterior end of the embryo. The meso-teloblasts are the last to disappear. They are distinct until after the formation of the stomodaeum and its connection with the pharynx. The concrescence of the germ bands begins anteriorly and progresses posteriorly. THE ECTODERM AND ITS PRODUCTS The three generations of ectomeres are given entirely to the formation of the ectoderm, which later becomes differentiated into the definitive hypodermis, with its glands, the cuticle and the anterior and posterior ends of the digestive tract. The ectoderm includes, in addition to the above, all of the telo- blasts, except the two larger and deeper ones which represent the mesoblasts. The reason for regarding the eight teloblasts and their derivatives as a part of the general ectoderm, is on account of their origin and position. In position, they are superficial at first and can not be distinguished from the general ectoderm, except by their arrangement in rows. Small cells are budded off from the teloblasts, which form the trunk ectoderm. In Clepsine these teloblasts are at first superficial at the posterior end of the embryo. In Lumbricus they are found directly in the general ectoderm, and beyond question form a part of it. 1. THE NERVOUS SYSTEM The nerve chain in Bdellodrilus first appears as a double row of cells, nearly uniform in size. Each row of cells originates from a single cell, the neuroblast. The neuroblasts, when first formed, are widely separated, but symmetrical to the median axis of the body. Figures 47 and 48 (in ventral view) show their position when first formed by an equal division of the protelo- blasts X and X. They take up their position on either side of the mesoblasts (figs. 47, 48, 93). When first formed the neuro- blasts are turned somewhat anteriorly as shown in the horizon- tal section of figure 93. This movement of the cells to their final position, independent of the former position of the cleavage 178 GEORGE W. TANNREUTHER spindle, is a common occurrence in Bdellodrilus. In some in- stances it is necessary to employ sections, in order to determine the origin of cells. The transverse axis of the embryo at this stage is often greater than the longitudinal (figs. 49-50). This condition persists for a brief period only, during the formation of the teloblasts. As the embryo increases in length the neuro- blasts are carried more and more posteriorly (figs. 56-57). In order to get a better understanding of the origin and orienta- tion of the neuroblasts—X right and X“” left—with reference to . the other teloblasts, the figures of plate 4 are so arranged that the left side of the developing embryo corresponds to the reader’s left. In figure 45 the upper pole is turned a little posteriorly, to show the upper outer edges of the entodermal cells. Figure 47 (from ventral pole) shows some of the ectodermal cells. The remaining figures are either turned forward or backward on their transverse axes. The ectomeres x*® and x® right and left serve as good points for orientation (figs. 46-53). After the formation of the teloblasts, bilateral symmetry is fully established. The meso-teloblasts, however in some instances, are still a little to the left of the median axis. This variation in the symmetry of the mesoblast does not in any way change the end result. In the early history of the germ band formation the teloblasts X and X™) are slightly separated, while X™ and X® are widely sepa- rated from the corresponding teloblast on the opposite side (figs. 56, 58). The neuroblasts and the nephroblasts begin their proliferation of cells to form the germ bands, about the same time (fig. 58). At this stage of development, the exact orienta- tion of the embryo is distinct. Since the embryo is completely turned on itself, the further use of the terms, apical and ventral poles, is significant only as being convenient in description. The mouth, as stated above, is formed in the center of the apical pole and the anus in close proximity on the dorsal side of the tenth segment. Figure 59 (upper pole view) shows the com- plete curvature of the embryo. The heavily shaded portion represents approximately, the boundary between the anterior and posterior ends. This figure shows that the teloblasts are coming more and more in a straight line. Since the two ends of EMBRYOLOGY OF BDELLODRILUS 179 the embryo are in immediate contact, it is impossible, except by longitudinal sections, to determine the exact point of separa- tion. The ectoderm of the anterior end of the embryo, which is derived from the three generations of ectomeres is continuous with the ectoderm derived from the ‘first somatoblast.’ The separation of the two ends of the embryo becomes recog- nizable in the early formation of the germ bands, as shown in figures 59 and 60. The posterior and ventral shifting of the neuroblasts (figs. 58-60) continues until all of the teloblasts are in a direct line. The small cells between the teloblasts are derived from the first somatoblast. In viewing the embryo from the upper pole (which now corresponds more to the anterior and posterior ends of the future animal) the germ bands extend laterally, downward and forward, being curved somewhat. pos- teriorly as they pass from the upper to the lower pole (fig. 59). The meso-teloblasts in figure 58 are still visible from the exte- rior. In figure 59 they are almost grown over, while in figure 60 they are completely covered. This is due to the posterior shifting of the neuroblasts and the growth of the ectomeres from above and below. In an embryo viewed from the right side (fig. 61, a little older than fig. 60), the position of the neural and nephric rows of the germ band are shown. As the rows extend anteriorly they are more difficult to distinguish from the ectoderm. The neural rows alone can be followed to the extreme anterior end. The posterior end of the embryo is widely blunt, while the anterior end is more rounded. The heavily shaded portion represents the point of separation be- tween the two ends. Figure 62 represents the same embryo from the upper pole, with the ends of the embryo rotated or turned a little posteriorly. - In figures 63 and 64 (from right and left sides respectively) the embryo is more elongated and the point of separation be- tween the two ends is more distinct. The neuroblasts are lag- ging in their posterior extension. Their position is median ventro-posterior, as shown in figure 65. Their concrescence is not yet complete at the posterior end. In the following stages of development the cells of the neural and nephric rows divide PLATE 5 EXPLANATION OF FIGURES 58 Embryo from upper pole, tilted a little to the right. The position of the ten teloblasts are shown; the small cells between the teloblasts on the surface are derived from x® and x’? on either side. 59 Same view as the preceding; the neuroblasts have migrated a little pos- teriorly and are approaching each other. 60 The ecto-teloblasts are nearly in a direct line; the germ bands have be- gun to form; the two primary mesoblasts M, M are no longer visible from the exterior; the transverse heavily shaded portion shows the approximate point of separation between the two ends. 61 Embryo viewed from the left side; the posterior end is extremely blunt. 62 Same embryo as preceding, from the upper pole (upper pole corresponds to the anterior and posterior ends). Shows very strikingly the close proximity of the two ends. 63-65 Represent the same embryo from the right, left and ventral sides respectively. The ectoderm which partially covers the germ bands is not shown. 66 Embryo from upper pole; bilateral symmetry is well marked; the telo- blasts are considerably reduced by the time they come in contact with their fellows on the opposite side. 180 EMBRYOLOGY OF BDELLODRILUS PLATE 5 G. W. TANNREUTHER BY BBS Pee cs ey: S ae Gap Aen Ge ae, 2s NC Oars eee, seg Smee < ewes i LY Ro) gag. (RU Pe sec Mapua peer tte’ , lige ieee aDOs Oley Ne Hare OAS i ees Os Sy < AS OAS rH dy ar Lest ca i oi ase aK oR — TiS Tannreuther, del. 181 PLATE 6 EXPLANATION OF FIGURES 67 Embryo turned slightly to the left to show the anterior and the posterior ends; the embryo at this stage begins to rotate within the egg membrane. 68 The same as the preceding from the ventral pole, turned a little to the left. 69 Embryo viewed from the right side; condition before the posterior end becomes drawn out or pointed. 70 Embryo from the upper pole; shows compressed condition of the two ends; at this stage the embryo rotates very rapidly. 71 Embryo viewed from the right side; the teloblasts are partially visible at the posterior end; the tapering of the posterior end is well marked. 72 Embryo to show the overlapping of the ends; indications of segments are visible anteriorly; the stomadaeum is distinct. 73 Unusual condition, where the two ends remain in immediate contact until after the form of the worm is distinct; this occurs in eggs with an unusually large cocoon. 74-6 Different stages in the final growth of the embryo. 77 Condition of embryo at the time of emergence from the egg. PLATE 6 G, W. TANNREUTHER 183 EMBRYOLOGY OF BDELLODRILUS Re ORS BSE Shy ae Doers spe Kline and Tannreuther, del. 184 GEORGE W. TANNREUTHER very rapidly and gradually become covered by the ectodermal cells as they sink beneath the surface (figs. 66-72), and form the middle stratum of the germ bands. In a nearly median longi- tudinal section (fig. 99), the relation of the parts are shown. The neural plate at different points shows the formation of ganglia. The anterior end of the section passes through the exact median plane and does not show any ganglia. The neuro- blasts still persist at the posterior end. The entire nervous system arises as two simple longitudinal rows of cells, and each row is produced by the continued pro- liferation of cells from a single cell, ‘the neuroblast.’ This proof is established by the study of surface preparations in connection with sections taken in different planes through the developing neural elements. The neural rows can be followed to the ex- treme anterior end, where they pass up around the pharynx and. give rise to the cerebral ganglia on either side (fig. 96) by a thickening of the anterior extremity of the corresponding neural rows. There are exactly four rows of cells in the middle stratum of each germ band. The outline of the rows can be easily seen in surface views of the living embryo (figs. 63-66). They are more distinctly marked at the posterior ends, and become less distinct anteriorly, which is due to the more ad- vanced development; that is, each row becomes double, then triple, ete. (figs. 68-69) and at the same time, its boundary lines become less distinct. . 2. THE EXCRETORY SYSTEM After the bilateral division of the ‘first somatoblast,’ each proteloblast contains the neural and nephric elements of their respective sides. According to Whitman, these two cells are called the neuro-nephroblasts. But when each _proteloblast X, X divides equally the neural and nephric elements become separated, X neural and X nephridial (figs. 47-48). The nephroblast X on either side next buds off a very small cell x’, which becomes ectodermal (fig. 49). Immediately after the for- mation of this small cell, X‘? divides nearly equally, and forms EMBRYOLOGY OF BDELLODRILUS 185 X and X on either side (figs. 50-54). Both cells are nephrid- ial. This fact perhaps is made more suggestive by the behavior of X° and X, Either of these cells may divide equally, but never both in the same embryo. In either case we have three teloblasts derived from the nephroblast X‘? on either side. The cells of the nephridial rows are somewhat smaller and narrower than those of the neural rows. In some cases the outer. nephridial row of cells is very short. In other embryos it is composed. of but one or two cells and its presence is hard to verify, suggesting a possible disappearance in the group. As stated above, the nephridia arise in connection with a con- tinuous nephric cord of ectoblastic origin, which forms a part of the middle stratum of the germ band and lies along side of the neural row. Each nephric cord terminates at the posterior end in three teloblasts. Thus the entire nephric cord of cells is formed by the continued division of the ‘nephroblasts,’ which agree precisely with the neuroblasts in structure, action and mode of origin. The nephric cord at first is composed of three rows of cells posteriorly, but passing forward the rows are no longer definitely separated and the nephric cord or plate con- sists of an irregular series of cells which extend anteriorly to the posterior end of the pharynx. The formation of the nephridia progresses from in front backwards and keeps pace with the formation of new segments in the embryo. The beginnings of a pair of nephridia are found in each of the main segments. Only two pairs of nephridia are retained in the adult worm. The details of the formation of these segmental organs have not been worked out. Berg considers the entire nephridia in Criodrilus as mesodermal in origin; Whitman held the extreme opposite view, that the entire nephridium was ectodermal in origin; while Wilson re- garded the nephridia as being part mesodermal and part ecto- dermal in origin. In Bdellodrilus the nephridia are ectodermal. The anterior pair occurs in the first, second, third and fourth body segments. The left nephridia of the anterior pair, extends from the first to the third segments inclusive, while the right extends from the second to the fourth segments inclusive. Both 186 GEORGE W. TANNREUTHER have a common opening on the dorsal side of the third segment. The funnel of the left occurs in the second and that of the right in the third segment. The posterior pair is found in the eighth segment. Each nephridium has a separate opening to the exterior on the dorsal side of the eighth segment. GROWTH The developing embryo does not increase appreciably in bulk until after the teloblasts are formed. Up to this period it is merely a division of the egg content into the various cell com- plexes. Even at this stage the increase in the long axis of the embryo is brought about by a decrease in the transverse diameter. Figures 50 and 55 show the transverse axis greater than the longi-. tudinal, while in figure 56 and 57 the longitudinal axis is greater, due more to a change in shape than to growth. The egg content is very plastic and when removed from the cocoon the egg mem- brane, in most cases, is not of sufficient tenacity to retain the embryo intact. The ten teloblasts are shown in figures 56 and 57. The first increase in length is due to the formation of the mesoblastic portion of the germ bands (text figs. 17-18). The neuroblastic and nephroblastic portions of the germ bands begin simultaneously after the meso-teloblasts have formed eight or ten cells (text figs. 15, 18 and fig. 58). Figures 58-71 show the various stages in the formation of the germ bands. Figure 71 is about the last stage when the germ bands can be detected externally. A longitudinal section of figure 71 near the median axis shows a differentiation of the germ bands into their incipient organs (fig. 99). From this point of development, growth is very rapid, and the embryo begins to rotate on its transverse axis. The movement is produced by the action of cilia, which occur on the large ectodermal cells in the median ventral half of the anterior end of the embryo (figs. 96-99). These cilia disappear before hatching, but the cells from which they are produced persist as a part of the ectoderm. The an- terior and posterior ends are no longer in immediate contact, as in figure 71, but begin to overlap. The ends of the embryo EMBRYOLOGY OF BDELLODRILUS 187 now take the position within the egg membrane of the least resistance to their further growth. Figure 74 shows the over- lapping of the ends. The stomodaeum is completely formed and the annuli of the pharynx are visible. Figure 73 shows an unusual condition in the position of the ends. At this stage of development the animal often turns on its longitudinal axis, largely on account of the action of the muscles, and, instead of the convex side being ventral, it now becomes dorsal. This rotation on its longitudinal axis has no significance, as has been thought by previous investigators, in the later stages of develop- ment. The animal is extremely plastic and may assume any position or shape, as shown in figures 74 and 76. Figures 77 shows the completely developed animal at the time of emergence from the cocoon. The number of the segments is distinct. This peculiar form of growth within the cocoon is merely adaptive. Occasionally, when the cocoon is of an unusual size, the develop- ing worm is less bent on itself. A COMPARATIVE STUDY OF DIFFERENT FORMS In following the cleavage cells of annelids, molluscs and poly- clades, one is impressed with the striking resemblances in their different stages of development. If this marked similarity alone were a sufficient criterion for a basis of classification, some of the most widely separated forms, when considered from the standpoint of their early development, would be grouped as closely related species. How can such resemblances in develop- ment be explained in animals which are so unlike in their late stages of growth? Are they merely the result of such mechanical principles as surface tension, alternation of cleavage, and pres- sure, or is the nature and structure of the protoplasm the com- mon cause? According to Driesch, ‘the striking similarity’ between the types of cleavage in annelids, molluscs and poly- clades does not appear startling and is easy to understand, since cleavage is of no systematic worth. However, the more recent investigators on cell lineage, according to Heath, look upon the early cleavage stages as something more than a mere 188 GEORGE W. TANNREUTHER manifestation of simple mechanical forces. Rather are the blastomeres the expression of the active intrinsic forces, which control development from the earliest stages unto the end. Gravity, surface tension, cohesion and pressure no doubt are effective, but not to the extent that they become the control- ling or coordinating agents in development. The early cleav- ages are aS important as those occurring in later life, and may even be considered more so. ‘‘Also the long continued resem- blances which exist in the development of these different forms from the earliest segmentation of the eggs are as fundamental and deep seated as are the homologies existing in the adults.” The number of these resemblances in the annelids and mol- luses is surprisingly great. In all forms accurately studied, the first three generations of ectomeres give rise to the entire ecto- derm. The mesoblast arises at the fourth division of the pos- terior macromere D. The remaining members of this quar- tette and the macromeres produce the entoderm. The division and position. of the cells up to the twenty-four or thirty-cell stage are identical in many different species. Beyond this point Wilson believes a divergence between the two classes ensues, and that development proceeds upon two entirely different lines. However, subsequent investigators have shown that the supposed differences are more superficial, and that the points of resem- blances become more numerous and extend throughout longer periods of development. Lillie (95) showed that points of resemblance existed in the lamellibranchs and the annelids, and that in both classes there is an essential similarity between the development of the ‘first somatoblast.’ In annelids this structure develops to a greater extent than in Unio, but the two have many points in common. Mead (’97) and Conklin (97) showed that the rosette series had the same origin and position in ahnelids and molluscs, and that in both it probably gave rise to the apical sense organ. According to Conklin, it also gave rise to the cerebral ganglia, while Mead considered this particular point doubtful. Further- more, Mead in his annelid studies demonstrated that the same cells in five different annelids gave rise to the entoderm; that EMBRYOLOGY OF BDELLODRILUS 189 the head kidney in Amphritrite and Nereis developed from the same cells. Conklin further states that the axial relation of all the blastomeres, with the possible exception of the macro- meres, are the same in both the annelids and molluscs, and that the larval mesoblast in Crepidula and Unio arises from the same group of ectodermal cells. Heath (’99) found that the prototroch in annelids and mol- luses was homologous, and that the twenty-two to twenty-five cells concerned have exactly the same origin, direction of cleavage, and destiny. Also that the remainder of the first quartette, forming the head vesicle with its rosette series and molluscan cross cells or intermediate girdle cells, has in all probability, the same fate in both. He found many other resemblances and concludes: Thus it is seen that not only in the origin and position of the various quartettes do resemblances appear, but that the early cleavage of these are in many cases cell for cell the same. In later stages close cell homologies cease, but the relation of the cell groups and their develop- ment in giving rise to larval or adult structures follow along much the same path. After passing these facts in review and considering the various structures in detail and modifications which they undergo, one fact presents itself with greatest clearness, that between Ischnochi- ton and the annelids the resemblances are moref undamental and closer than are the differences. For a more direct comparative study of Bdellodrilus with the annelids and molluscs, special references will be made to Clep- sine (Hirudinea), Dinophilus (Polychaete), and Unio (Lamelli- branch). In all these forms the first and second cleavages are meridional and divide the eggs into four unequal macromeres (text figs. 19-22). In Dinophilus C and D are approximately posterior and A and B are anterior. In the other three forms B is anterior, D posterior, C right and A left. In each case D is the largest cell; A, B and C are nearly equal; B is usually the smallest when variation occurs. The eight-cell stage has the same structure, and in all probability arises in the same manner in the four forms, the only apparent difference being the much greater relative size of the ectomeres in Dinophilus than in the three remaining forms. The first cleavage plane in Bdellodrilus 190 GEORGE W. TANNREUTHER Qe yall Fig. 19 Four-cell stage of Unio, upper pole (after Lillie). Fig. 20 Four-cell stage of Bdellodrilus, upper pole. Fig. 21 Four-cell stage of Dinophilus, upper pole (after Nelson). Fig. 22 Four-cell stage of Clepsine, upper pole (after Whitman). occurs at nearly right angles, while in Unio and Clepsine it is inchned at an angle of about forty-five degrees to the sagittal plane of the future adult. In Dinophilus the direction of the first cleavage is in doubt. The second cleavage plane in Unio, Clepsine and Bdellodrilus occurs at an angle of about forty- five degrees to the sagittal axis. The origin of the ectoderm, the entoderm and the mesoderm is approximately the same in each form. 1. THE FIRST SOMATOBLAST The first somatoblast in each instance is derived from the large posterior macromere D (text figs. 23-26). The cell d? (X) is extremely large and occupies a median posterior position. In Clepsine (Whitman) d? (X) is called the ‘neuro-nephroblast.’ EMBRYOLOGY OF BDELLODRILUS 191 It divides into two, four and finally eight large cells called the teloblasts; the middle stratum of the germ bands is derived from them. These eight teloblasts are arranged into two groups of four cells each. Each group, which later is composed of four rows of cells, produces the middle stratum of the germ band on the corresponding side. The inner row of each band lies ulti- mately near the median ventral plane and gives rise to the corresponding half of the nervous system. The adjoining rows —‘nephroblasts’—give rise to the nephridia. The derivatives of the outer row are still in doubt, but probably take part in the formation of the ectoderm. In Dinophilus (Nelson) d? (X) is formed by a laeotropic divi- sion of the macromere D (text fig. 25); D is much smaller than X. Immediately after the formation of X, x! is budded off to the right at a low level. Next x? is budded off to the left at a higher level than x'; x is next formed by a dexiotropic division from the dorsal side, a little to the left. Next X divides equally and produces X and X, right and left. These two large cells cor- respond to the proteloblasts in Bdellodrilus. Finally X on either side divides equally, and produces the two teloblasts on each side of the median plane. These four cells, according to Nel- son, correspond to the posterior teloblasts of Nereis. They also correspond to the neuroblasts and nephroblasts of Bdellodrilus. The division of X in Dinophilis and Nereis differs no more than do the corresponding divisions in Nereis and other annelids (Amphitrite, etc.). At the time of the closure of the blasto- pore in Dinophilus, the descendants of X are distributed dorsally and laterally to the posterior stem cells. In Neries the main bulk of the descendants of X lay on the vegetal side of the stem cells. In Unio (Lillie) the ‘first somatoblast’ X is formed by an un- equal division of D (text fig. 24) in a median posterior position; x! is budded off from X, just behind C on the vegetal pole; next x? is budded off from X symmetrically with x! on the right side, just posterior to d’; next x* is formed from X towards the apical pole, posterior to d'2; then x‘ is budded off from X an- teriorly, towards the vegetal pole. This division of X does JOURNAL OF MORPHOLOGY, VOL. 26, No. 2 192 GEORGE W. TANNREUTHER Fig. 23 Nine-cell stage of Clepsine, upper pole (after Whitman). Fig. 24 Nine-cell stage of Unio from behind (after Lillie). Fig. 25 Nine-cell stage of Dinophilus, left side (after Nelson). Fig. 26 Nine-cell stage of Bdellodrilus, left side. not occur in this manner in Dinophilus, Bdellodrilus or even in Nereis. The fourth division in the above three forms is equal and bilateral, while in Unio the fifth cleavage of X is the first bilateral division and forms X, X right and left. Next X, X on either side divides nearly equally and gives rise to the shell gland. These four cells might be regarded as the posterior telo- blasts, which occur in other forms, as in Nereis and Dinophilus. In Bdellodrilus X is formed by an equal division of the macro- mere D (text fig. 26), and takes a median posterior position. First x! is budded off from X to the right, posterior toC. Then x? is budded off to the left, symmetrical with x' and posterior to d?. Next x? is formed from the median dorsal anterior edge of X, between d! and c'. Now the first bilateral division of X takes place and forms the proteloblasts X, X, right and left. Each of the proteloblasts bud off x‘, x® and x® respectively. At the EMBRYOLOGY OF BDELLODRILUS 193 next division each proteloblast divides nearly equally, and gives rise to X™, neuroblast, and X, nephroblast, on each side of the median axis of the embryo. Next each nephroblast divides nearly equally and produces X‘ and X‘). Now a very inter- esting thing happens; either X° or X® divides (but never both in the same egg) and produces the three nephroblasts on each side, which are designated as X, X‘? and X™. In Clepsine only two of these teloblasts are concerned in the formation of the nephridia. The lateral teloblasts, as stated above, are prob- ably ectodermal. These four forms unquestionably show that there is a marked similarity in the cleavage of the ‘first somatoblast,’ not only in widely different individuals in the same group, but in individuals of widely separated groups. This comparison could be extended to other groups or forms, but the above will suffice for our purpose. 2. THE SECOND SOMATOBLAST In Clepsine, the ‘second somatoblast’ has rather a unique origin. It is formed at about the twelve-cell stage. D, after the formation of X, becomes directly the ‘second somatoblast’ or M. (These cells are differently designated by Whitman; D is represented by x, X by x! and the mesoblasts by x and xy.) M divides nearly equally and produces the right and left meso- blasts, from which the inner stratum of the germ bands is formed. In Bdellodrilus, M is formed by a very unequal division of the macromere D, at the twenty-four-cell stage (fig. 85). The larger cell or M is formed in front of X. It is inclined a little to the left of the median axis. The first division of M is equal, producing the mesoblasts, one on either side. These primary mesoblasts now bud off a number of small cells, directly be- neath d' and ec! (figs. 86-87). It is very difficult to make out the exact number of these small cells, since they are not visible externally. There are at least twelve formed, six on either side from each mesoblast. After this small group of secondary mesodermal cells are formed, the mesoblasts M, M, give rise to PLATE 7 EXPLANATION OF FIGURES 78 Two-cell stage, horizontal section; CD dividing. 79 Four-cell stage; horizontal section taken above the center of the egg. 80 Same as the preceding, with plane of section below center. 81 Nine-cell stage, parasagittal section, to right of median plane. 82 Horizontal section of a nine-cell stage, taken four sections from the top. Taken from embryo composed of 21 sections, each eight micra in thickness. 83 Same as the preceding; sixth section from top. 84 Same as figures 82 and 83; fifteenth section from top. 85 Parasagittal section of a twenty-four-cell stage; plane of section little to left of median axis. This figure shows the unequal division of the macromere D in the formation of the second somatoblast. 86 Thirty-three-cell stage. Parasagittal section to left of the median plane. Shows the formation of the secondary mesodermal cells (m cells) from the pri- mary mesoblasts. 87 About the same stages as the preceding, to show the distribution of the yolk in different cells. 88 Horizontal section of an eighteen-cell stage, fourth section from top. Series composed of 20 sections, each eight micra in thickness. 89 Same as the preceding; seventh section from top. 90 Taken from series same as figure 88; eighth section from top. 194 EMBRYOLOGY OF BDELLODRILUS PLATE 7 G. W. TANNREUTHER Tannreuther, del. PLATE 8 EXPLANATION OF FIGURES 91 Taken from the same series as figures 88 to 90; eighth section from the top. 92 Horizontal section of an eighteen-cell stage; sixth section from the top; shows the neuroblasts and nephroblasts. 93 Same as the preceding; tenth section from top; this figure shows the per- sistence of the cleavage spindles after the cell membranes are distinct. 94 Same as 92 and 93; fourteenth section from top; taken from a series of 20 section each eight micra thick. This figure shows the upper side of the macro- mere D wedged in between the other cells. 95 Transverse section of an embryo represented by figure 64; section taken at plane 2 — 2, or at a region corresponding to plane 2 — 2, of figure 98. 96 Transverse section of stage corresponding to figure 63; section taken at plane 2 — 2; figures 95 and 96 shows germ bands only partially covered by the ectoderm. 97 Transverse section of embryo represented in figure 69; section taken at level marked by line 2 — 2; here the germ bands are completely covered by the ectoderm. 98 Longitudinal section, near median line, of stage represented by figure 69. 99 Longitudinal section of an embryo represented by figure 71; plane of section near median line. 196 EMBRYOLOGY OF BDELLODRILUS PLATE 8 G. W. TANNREUTHER Se A Gi {Ojo} PN KOO NATE Boy] YO) ae sengt : gta! Dios yy, Beal {Oo} Score iss (oe pyoaeias FOOD of 24 a (= ie ef EI ae Softer NG} ees Pewh | CR, ee eae is Sais 0, oe iy nC) t Weg? @ neat : Ke aera Tannreuther, del. 197 198 GEORGE W. TANNREUTHER the inner stratum of the germ bands. The plane of division in the formation of the primary mesoblast is at right angles to that of the secondary. In Dinophilus, M is formed at the twenty-nine-cell stage, by an unequal division of D. M is much larger than D, as in Bdellodrilus, and is in front and below X, slightly to the left of the median plane. The division of M is now delayed until the seventy-two-cell stage, when two small cells are budded off anteriorly towards the vegetal pole, close to the line of junction of the two mesoblasts. At the next division two small cells are budded off, one on either side of the first pair. The follow- ing cleavages are teloblastic and produce the mesoblastic bands. The mesoblasts do not move into the cleavage cavity as in many other annelids, but remain on the surface until covered by the ectoderm. In Unio, at the thirty-two-cell stage, M is formed by a very unequal division of the macromere D. The first division of M is equal and bilateral. Their position is immediately behind the entomeres. The next division of the two mesoblasts is very unequal, two small cells, m, m, are budded off at the posterior lip of the blastopore. Later the mesoblasts are included within the segmentation cavity, where they take up their final position behind the archenteron and give rise to the definitive meso- blastic germ bands with lateral teloblasts. From the forms compared above it is very evident that there is a remarkable similarity with Bdellodrilus, not only in the early cleavage stages, but in the establishment of the germ bands as well. Thus cells having the same origin and lineage have the same final result in a wide series of forms (d‘ the mesoblasts). Again, cells of unlike origin have a different fate (first and second somatoblasts); or cells of a different origin may have the same fate (d‘ of annelids and the second and third generation of ecto- meres in polyclades). Then cells of the same origin may havea different fate (a2? in Unio and Bdellodrilus). These contradic- tions, however, are far less striking than the resemblances. The ‘first somatoblast’ in each of the above four forms gives rise to the ventral plate and all or nearly all of the trunk ectoderm, while EMBRYOLOGY OF BDELLODRILUS 199 the ‘second somatoblast’ produces the definitive mesodermal elements of the adult animal. 3. VARIATIONS IN THE METHOD OF MESODERMAL FORMATION All annelids and molluses which have been carefully studied show that the ectoderm arises from the three generations of ectomeres, the mesoderm from M and the entoderm from the remaining cells. There aré, however, a few minor variations in forms like Clepsine, Crepidula and Nereis. In polyclades the mesoderm is directly associated with the ectomeres. The second and third generations of ectomeres, as in Discocoelis, produce the mesoblast, and the macromeres the entoblast. In molluscs and annelids the mesoderm is more closely associated with the macromere D. There is considerable variation in the cleavage of D in the formation of the ‘second somatoblast;’ in some forms D is given over entirely to the mesoderm; in other forms it shares equally, or in still others it contributes but little to the mesoderm. Forms in which the mesoblast has two sources (a) Ecto-mesoblast and (b) coelo-mesoblast. In Thalassema, Torrey distinguished between ecto-mesoblast, from the ectoderm and. coelo-mesoblast from M. He states that the coelomeso- blast is present in two bands, each consisting of five sub-equal cells. These are closely applied to the body wall and lie in the usual position on each side of the neural rows, but are more widely separated than in most annelids. The ecto-mesoblast on the other hand, derived from the first and second quartette of ecto- meres, is present in great abundance and many of the cells have already undergone considerable differentiation in the formation of the muscles. He further states that the two mesoblast cells, M, M, are the last to sink in at gastrulation, instead of the first, as in the case where development is more direct (Nereis and Amphitrite). The two coelo-mesoblast bands of five cells each are shown to have the same origin and early history as the meso- blast bands in other annelids. The coelo-mesoblast, which is meagerly developed in the trochophore, is clearly correlated with 200 GEORGE W. TANNREUTHER the long duration of its free swimming and almost stationary larval existence. In all forms where there is a tropchophore stage of long duration, as is the case in all annelids with equal cleavage, the two coelo-mesoblast cells do not, in the early stages at least, bud like teloblasts. This is tue in Hydroides, some species of Polygordius, in Thalassema and Podarke. Many of the annelids and molluscs show that the so-called ecto-mesoblast (designated. as larval mesoblast or as mesenchyme by some authors) arises from certain ectodermal cleavage cells of the second or third quartette of ectomeres and is entirely independent of the coelo-mesoblast. In Thalassema (Torrey) ten. large ectodermal cells sink in from the ectoderm and. give rise to all the mesenchyme. Three of these cells are from the a, c and d quadrants of the third quartette and seven from the first quartette of ectomeres. The most important source of mesenchyme in Thalassema is from the three cells of the third quartette (3d2-2-2-1, 3c?121, and 3a?-22). The first two sink into the cleavage cavity, Just before gastrulation and at first le close to the two coelo-mesoblast cells. They soon migrate laterally and bud off simultaneously small cells toward the M cells. They divide like teloblasts, but in the reverse order to the ordinary direction. So close is the connection of these cells with the coelo-mesoblast that one would be certainly led to think that they formed a part of these bands, unless their cytogeny had been carefully followed. Similar conditions are described by Tread- well in Podarke obscura. The progeny of these two cells form almost the entire mesenchyme of the post throchal region and be- come differentiated for the most part into muscles of the diges- tive tract. The progeny of the other ectodermal cell migrates to the mid-ventral line. The ecto-mesoblast cells of the first quartette sink into the primary body cavity later than those of the third; their exact cell lineage has not been traced, but probably give rise to gut musculature. This mesoblast has commonly been considered as purely larval and transitory. In some instances it is possible to determine its exact origin, but in many others merely the general region from which it arises. EMBRYOLOGY OF BDELLODRILUS 201 During the last thirty years embryologists have differed in their conception of the origin of the mesoderm and. of its phylo- genetic significance. Hatschek (’78) was among the first to distinguish between. mesenchyme and mesoderm, but held, after studying the embryology of Polygordius, Echiurus and Eupo- matus, that these two morphologically different mesoblasts, arise from a common foundation. This same view was later put forth by the Hertwigs (’81) in their ‘Coelomtheorie,’ which, according to Meyer, has formed the foundation of all later work on mesoderm. Roule (’89 and ’94), Burger (’91 and 794), Frai- pont (’88), Hacker (’95), and others, have described the meso- blast as having a single origin. On the other hand, those who have studied the embryology of annelids and molluscs, con- sider the origin of the mesenchyme distinct from that of the meso- blast or coelo-mesoblast. This later view was first described by Kleinenberg (78 and ’86), and later by Whitman (’87), by Berg (’90), by Schimkewitsch (’94), by Meyer (01), by Torrey (03), and others. A larval mesoblast was first described by Lillie (’95) in Unio. It arises asymmetrically from the derivatives of a? and later mi- grates into the segmentation cavity, where it divides equally and becomes symmetrically arranged on either side of the mid- line. The derivatives of these two cells become metamorphosed into ‘myocytes’ and larval adductor muscles, which are functional only during larval life. Treadwell (97) regards both mesenchyme and mesoderm as morphologically the same tissue, the apparent difference in their mode of origin being of no significance. And, further, Wilson regards the larval mesoblast (ecto-mesoblast, because of its origin from the ectoderm) as a distinct tissue from that of the definitive mesoblast or ento-mesoblast, and states that it is homologous with the mesenchyme of the turbellarian ances- tors of the annelids, while the mesoblast from which the adult structures arise is phylogenetically younger and is represented prophetically in the ontogeny of such a form as Discocoelis (polyclade) by the peculiar lateral division of M, and states that the ecto-mesoblast and endo-mesoblast are phylogenetically of 202 GEORGE W. TANNREUTHER different origin. This same point was previously urged by Meyer. The condition, however, found by Wilson in Nereis and Lum- bricus does not indicate a hard and fast distinction between the two kinds of mesoblast. In Nereis, cells from the anterior end of the germ bands separate early and pass forward into the seg- mentation cavity where they give rise to the larval musculature. This corresponds exactly in structure and function with the larval mesoblast of Unio (Lillie) and Podarke (Treadwell). In Lumbricus the origin of the mesenchyme is similar to that in Nereis. These two kinds of larval mesenchyme have also been described by Eisig (98) as occurring in the same individual (Capitella, a polychaete annelid). In Thalassema and Podarke the larval mesenchyme arises directly from the ectoderm, while in Nereis and Lumbricus it arises from the anterior ends of the mesoblast bands. Accord- ing to Treadwell, no one has yet proven that no ‘mesenchyme’ arises from the germ bands in cases where a larval mesenchyme exists. If we accept Wilson’s view that mesenchyme and meso- derm are different phylogenetically, we must regard the two sets of larval mesenchyme which have the same structure and fune- tion, as non-homologous, or we must regard the mesenchyme and mesoderm as morphologically the same tissue and the dif- ference in their modes of origin as of no significance. Further- more, Wilson has pointed out that the trochophore, as it occurs at present, is more than a mere ancestral stage, for it contains in a concentrated form the anlage of the whole future body. According to Mead, the ectoderm behind the first septum in Amphitrite arises from a group of cells which surround the proc- todaeum of the young trochophore and are descended from a single cell, the ‘first somatoblast.’ The same is true of other trochophore forms. There is no need to assume phylogeneti- cally a new formation of ectoderm for the body as distinct from that of the head. Neither is there any necessity to assume a distinct phylogenetic origin of the larval mesoblast from that of the mesoderm. EMBRYOLOGY OF BDELLODRILUS 203 It is evident that in Nereis and Lumbricus, both kinds of mesoblast have the same origin, and simply shows a more complete concentration of the mesoderm than in Thalassema and Podarke, where the mesen- chyme is formed direct from the ectoblast. The mesoblast cells col- lected at the posterior end of the trochophore, which are derived from M, represent the mesoderm of the body. It is morphologically con- tinuous with that of the head, as in Nereis, and is concentrated at this point to provide for the elongation as new segments are formed. The difference in the concentration of the mesodermal elements, as to whether they have a single or double origin in no way interferes, as already pointed out, with the morphological unity of the tissue, and as to the source of its origin, whether from the ectoderm or from the endoderm phylogenetically, we are not able to say (Treadwell). Meyer (’01) in his study of the phylogenetic significance of the two kinds of mesoblast, gave a view directly opposed to that expressed by Treadwell. After an exhaustive review of the whole mesodermal question, he concludes that the great mass of evi- dence, both embryological and anatomical, points to the con- clusion that in annelids, at least, there are two entirely distinct forms of mesoblast, the ecto-mesoblast (primary mesoblast) and the coelo-mesoblast (secondary mesoblast). Of these two he considers the primary mesoblast to be phylogenetically the older, and as a rule, to be derived from the ectoderm. The coelo-mesoblast, on the other hand, is regarded as a later forma- tion, which has originated from the gonad cells. The formation of the ecto-mesoblast in annelids and molluscs from certain cells of the first, second, third and fourth generation of micromeres, can well be regarded as vestiges or survivals of the process which occurs in all four cells of the second and third quartettes of certain polyclads. The origin of the ecto-meso- blast from the ectoderm in annelids and molluscs, partially bridges the gap between them and the polyclads. In order to have a complete homology of the mesoderm in the polyclads, annelids and molluses, it is necessary to find a polyclad in which there is a double origin of the mesoderm. The development of the polyclad Leptoplana (Wilson) is the nearest representative to complete the homology. In Leptoplana only a part of the four quadrants of the second quartette contributes to the entire mesoderm, the typical condition in polyclads being that all of 204. GEORGE W. TANNREUTHER the second and third quartette is mesodermal. The behavior of d‘ in the polyclad Discocoelis, is also very suggestive. Here the division of d‘ is equal and gives rise to two symmetrically placed cells at the posterior end of the embryo, comparable to the primary mesoblasts found in annelids and molluscs. Some investigators have even suggested that these two posterior cells in the polyclads may give rise to the mesoblast bands in this particular group. This latter point, however, has never been verified. Table 3 shows that the first, second, third and fourth generation of micromeres, in a series of widely separated forms, may con- tribute to the formation of the mesoderm. TABLE 3 Ist GEN. | 2D GEN. 3D GEN. 4TH GEN. Annelids: | | Thalassema | part of a, b | none 1 cell each of | d* part mes. /and ¢ quad’s | a, b and -c | quadrants Bdellodrilus | none none none d‘ all mes. Molluses: | Unio none | ar? none dé all mes. | (larval) Crepidula none | ea ler ner none d‘ part mes. Physa none | none beme d‘ part mes. Podarke none | none eseeed2 d* part mes. | @3-2-1-2, (3:2-2-2 Polyclads: | Discocoelis none | all mes. allmes. | none Leptoplana none _part of each none none quadrant In case of the ecto-mesoblast a complete series could be ar- ranged, in which all of the cells of certain quartettes contribute to the mesoblast, to those forms in which only a small part of cer- tain quartettes is mesoblastic. Again in case of the coelo-meso- blast we have a wide range of variation, in which all of d* 1s meso- dermal, to those in which only a small part of d‘* is mesodermal. As far as records show, Capitella is the only annelid in which EMBRYOLOGY OF BDELLODRILUS 205 d‘ does not contribute to the coelo-mesoblast. Here then we have quite a unique series ranging from those forms where the entire mesoderm is ectodermal in origin, or where it is both ectodermal and entodermal, to those where it is entirely ento- dermal. From the above it is evident that the entire meso- blast of polyclads is derived from the ectomeres, and, if homol- ogies be any significance, it would be fair to conclude that this mesoblast is represented by the ecto-mesoblast in the annelids and the molluscs. The origin and development of the mesoblast in Bdellodrilus contributes but little to the phylogenetic significance cf the primary and secondary mesoblast. Here, beyond question, when considered from the standpoint of their origin, they are one and the same tissue. Both are formed directly from the primary mesoblasts. The secondary mesoblast cells are budded off from the two primary mesoblasts before the germ bands begin their development. Similar conditions are found in other forms, as in Lumbricus; here, however, the secondary mesoblast is formed later directly from the anterior ends of the mesoblastic germ bands. The difference is only in the point of time in their formation. In Bdellodrilus there can be no hard and fast dis- tinction made between the two kinds of mesoblast. Both must be considered as the same tissue phylogenetically. 4, VARIATIONS IN THE SOURCE OF THE ENTODERM In general, as stated above, the ectoderm originates from the three generations of ectomeres, the mesoderm from d‘, and the entoderm from the remaining cells. The origin of the three germ layers, however, depart somewhat from the above rule in some of the annelids and molluscs. In some species cells from the first, second and third quartettes contribute to the meso- derm; in others d‘ gives rise to entoderm as well as mesoderm. Tn all annelids and molluses, A, B and C, after the formation of the first three sets of ectomeres, are distinctly entodermal. The macromere D, after the formation of d*, is likewise entoder- mal. In some forms D is the same size as its fellows, in others 206 GEORGE W. TANNREUTHER it is reduced until it is little more than a mere nucleus, while in others it has completely disappeared as an entomere, and is given over entirely to the formation of mesoderm. In annelids, in a gradually decreasing series, D (Nereis) is the same size as the entoblast cells A, B and C. In Dinophilus it is about half the size of these cells. In Bdellodrilus D is little more than a mere nucleus; while in Clepsine D is given over entirely to the formation of the mesoderm. In molluscs it is a fairly common condition to find the entoblast cell D smaller than A, B and C, or even greatly reduced. In Crepidula it is very little reduced; in Unio it is more than half reduced, while in Ischnochiton, D is often little more than a mere nucleus. The second somatoblast, M, may contribute tc the formation of entoderm as well as mesoderm. In forms like Crepidula M is mostly entodermal. In Fiona (Casteel) the division of M in the formation of the entoderm is very similar to that in Crepi- dula. In Unio two small cells are budded off from M, which lie near the entoderm, and are probably concerned in the forma- tion of that layer. In some of the annelids the primary mesoblasts bud off small cells directly posterior to the macromeres. This number varies; in Nereis there are six to ten, and in Aricia there are but two. In many of the other annelids and also in some of the molluscs, where their cell lineage has been traced, it is found that these small cells give rise to entoderm. There are at least sixteen to twenty species of annelids and molluscs in which similar cells have been found (small cells from the primary mesoblasts.) Diverse accounts of their behavior and fate have been given by different investigators. Table 4 shows the fate of these small cells in a few of the annelids and molluscs. In the molluse Alpysia, according to Carazzi, each primary mesoblast buds off four small cells. Three of these are meso- blastic and one is entoblastic. This interesting condition might be considered as a transitional form or as a connecting link be- tween. those forms in which these small cells are entirely mesoder- mal and those in which they are entodermal. Again we could arrange a series of annelids and molluscs in which at one extreme EMBRYOLOGY OF BDELLODRILUS 207 TABLE 4 ENTODERM MESODERM NOT CERTAIN Crepidula | Amphitrite Dreissensia Nereis Arenicola Patella Podarke | Umbrella Spio Thalassema | Planorbis Serpulorbis Fiona Unio? Cyclas Ischnochiton | Limax | Aricia Physa fontanalis Physa hyponurum Aplysia the entoblast derived from M is greater in amount than the meso- derm, as found in Crepidula, and at the other extreme, where but two rudimentary cells of M are entoblastic, as in Aricia. According to Wilson, a series of this nature may indicate a gradual elimination of the entodermal element from the macro- mere D of the fourth quartette, and finally its complete trans- formation into the mesoblast. Kovalevksy (’71) suggested that this transformation shows quite forcibly that the mesoblast pole cells are to be regarded, phylogenetically, as derivatives of the archenteron, because of their close association with the posterior entoblast cell, D. The primary entoblasts, A, B, C and D, undergo but little change until late development in thcse forms which possess a larval stage, and may remain in this condition until after the trochophore is developed, or until after the blastopore is closed. In those individuals with a fetal type of development, they often remain distinct until after the germ bands are completely formed, as in Clepsine. GENERAL ADAPTATION AND INTERPRETATION OF CLEAVAGE The cleavage of eggs of widely separated forms exhibit unique resemblances. At certain stages of development these resem- blances exceed their differences. Is the persistence of these features due to the influence of ancestral inheritance, or are they due more to the adaptive conditions of their environment, to JOURNAL OF MORPHOLOGY, VOL. 26, No. 2 208 GEORGE W. TANNREUTHER meet the highest need of the developing animal? It has been demonstrated, again and again, in annelids, molluses and even in polyclads, that homologous cells of like generations give rise to like parts in the developing embryo and the adult. The occurrence of these conditions in such widely separated forms furnishes a very interesting and important phase in the study of cell lineage. The tendency has been rather to emphasize these resemblances, than to give special stress to the exact conditions which occur in any one species in its different stages of develop- ment. It is true, however, that the general form of cleavage may be inherited from a long series of ancestors, probably from some of the Turbellarian worms. But the problem of more direct importance in any one group is, why such variation in the size, form, direction and rate of cleavage? 1. IN THE CLEAVAGE OF BDELLODRILUS In Bdellodrilus we have a determinate type of cleavage, i.e., the fetal as well as the adult structures can be shown to have a definite or direct cell lineage, and can be traced back to the unsegmented egg. The structure of the ovum is quite homo- geneous, and at the time of maturation, the egg can be definitely oriented as to the future axis of the body. Before the first cleav- age is complete, the parts of the ovum which give rise to the differ- ent germ layers can be traced or ascertained with a fair degree of accuracy, 1.e., definitely localized parts which’ give rise to definite organs or structures. “Adaptation in cleavage can manifest itself only in three pos- sible ways or modes of cleavage variation, which are, as has been pointed out by Lillie, Mead, Conklin and others, the following: first differences in the rate of cleavage; second differences in the size; and third, differences in the direction of cleavage.”’ The general plan of cleavage in Bdellodrilus, is similar to that of other forms. The ectoderm is derived from the four basal cells, by three successive horizontally formed cleavages. The mesoderm from a fourth cleavage of the posterior macromere D and the entoderm from the remaining cells. The first cleavage EMBRYOLOGY OF BDELLODRILUS 209 in. Bdellodrilus is meridional and very unequal. In the two-cell stage the larger cell is posterior and the smaller cell anterior. The larger cell divides first and very unequally, while the smaller cell divides nearly equal (text figs. 1-3 and fig. 5). In the four- cell stage D is posterior, C right, A left and B anterior, inclined a little to the right. Thus it is very evident that the four-cell stage illustrates a difference in the rate of cleavage, a differ- ence in the size of the cells, and a difference in the direction of the cleavage. The significance of these variations may be em- phasized as follows: a. Difference in the rate of cleavage of cells If we compare a thirty-two-cell stage of Bdellodrilus with other forms or perhaps, better, with an ideal ovum, in which there is a uniform rate of cleavage in the formation of the cleav- age cells, a uniform size and a uniform direction of cleavage, a distinct variation occurs as shown in table 5. TABLE 5 CREPIDULA rer Bai | NEREIS ey in Soe Pon cane —_s ————— ee a | =i53 | ; | First generation of ectomeres............ 12 16rs.| 16 8 Second generation of ectomeres.......... 9 8 8 11 Third generation of ectomeres............ 5 4 t 4 “NL TEYETE! CPT OY ee noe a a 0 0 2 HEEL) DISSE SRP A a eS a ea 4 4 4 ih 32 32 32 32 It is evident that in the first generation of ectomeres Bdellodril- us departs very far from the ideal condition. The first generation contains eight cells instead of sixteen. This means that the cells have divided more slowly than in the ideal ovum. In an ideal ovum these cells form the prototroch and the entire region in front of it, with the apical plate in the center. Jn Bdellodrilus, this region is degenerate and no trace of the apical plate appears. This indicates an adaptive modification—eight cells instead of sixteen—due to a degenerate frontal region. In the second generation of ectomeres, the ideal number is 210 GEORGE W. TANNREUTHER eight, while in Bdellodrilus it is eleven. This increase above the ideal is due entirely tothe rapid succeeding divisions of one cell— d?, the first somatoblast. The other cells of the quartette have not divided, while d? has given rise to three new cells. Does the behavior of d? suggest any significance, or is it adaptive? From d? the ectoderm of the trunk region, the nephridia and the entire nervous system is derived; d? is not only the largest but the most actively dividing cell of the entire embryo; hence its rate of cleavage is well adapted to its resulting formations. _ The number of ectomeres in the third generation is the same as in an ideal ovum; d? however is often formed before a?*or b? of the second generation. This interesting phenomenon is due to the tendency of the basal cell, D, and its derivatives to divide more rapidly than those of A, B or C. The differences in the rate of cleavage in the first, second and third generation of ecto- meres, no doubt possess prospective significance, looking for- ward to the definitive parts. This may fairly be called adapta- tion in the rate of cleavage. In Bdellodrilus the more rapidly dividing cells do not form the first functioning parts. The cells of the first quartette are the first to function, in the production of cilia for the movement of the embryo. The variation in the rate of cleavage is not due to the varying conditions of the media, or the dividing ovum would be uniformly affected, as a whole. Nor is it due to the size of the individual cells, as the largest cells divide more rapidly. At the thirty-two-cell stage the ideal ovum contains four entodermal cells, while in Bdellodrilus there are seven. Here the cleavage is carried to the end without any resting stage of the four basal cells. This is due to the fact that the larva develops very rapidly and the entodermal cells must keep pace with the rapid development in order to reach their final position, Just where they are needed. b. Variation in the size of cells The relative sizes of the cells in the early cleavage of the eggs of Bdellodrilus are adapted to the later developing parts. EMBRYOLOGY OF BDELLODRILUS PAI The largest cell at the four-cell stage is D. Its first division is very unequal, and the smaller cell is less than one-tenth the size of the larger. It is the first ectomere of the first generation formed. The second division, in most instances, is equal; when unequal, the largest cell passes into the upper product, and forms d?, the first somatoblast. The third division is unequal, and d*, the smaller product, is again uppermost; and finally, the fourth division is very unequal and only a small portion re- mains as the macromere D. The greater bulk, d‘, becomes the second somatoblast. In each of the above instances the larger cells form a large part of the embryo and the adult, while the smaller cells, in-every instance, form a very insignificant portion. The unequal division in each instance is evidently adaptive, for the great bulk of the material passes into the two somatoblasts and gives rise to the muscular, nervous and excretory systems. c. Variation in the direction of cleavage Here only some of the special cleavages will be emphasized. The first division of the second somatoblast is equal, and each part forms equal parts of the mesoderm. Next, each primary mesoblast buds off five or six small cells beneath the first quar- tette of ectomeres. These small cells remain quiescent for a considerable period and later give rise to the dorsal mesoderm. Immediately after these small cells are formed, the mesoblast bands are begun by a forward proliferation of cells from the an- terior face of the primary mesoblasts. The plane of division is at right angles to that of the small cells. These bands extend forward between the ectoderm and the entoderm, and at the same time the entodermal cells extend posteriorly between the mesoblast bands and the group of small cells, thus separating the primary and secondary mesoderm. Here the direction of the cleavages place the .cells where they are later used in the formation of some special part, adapted for that particular region. . The first somatoblast buds off x! to the right, x? to the left, and x? median dorsal anterior. X now divides equally and 212 : GEORGE W. TANNREUTHER each proteloblast buds off a small cell, x‘, one to the right of xt and the other to the left of x2. Again each proteloblast buds off a small cell, x*, on either side of x*. At the next division each buds off a small cell, x®, on the ventral anterior edge. Later, x7 is budded off from each neuroblast on the ventral side. These small cells give rise to the trunk ectoderm and the larger cells to the nephridia and nervous system. Here again the cells are formed just where they are needed; the smaller on the exterior or outer surface while the larger remain within. 2. ADAPTATION IN THE CLEAVAGE OF OTHER FORMS In following the variation of cleavage cells in annelids and molluses, special cells can be arranged in a complete series, from those of an almost insignificant size to an extremely large cell. In following these variations, step by step, we can not fail to be convinced that these variations are adaptive to the future needs and habits of the larva and of the adult animal. In forms with equal cleavage, the first somatoblast gives rise to the ectoderm of the trunk region. In Polydorgus, Po- darke, Hydroides, Eupomatus and others with equal cleavage, d? is the same in size as the cells of the other quadrants. Equal cleavage has been offered by some as due to a lack of differentia- tion in the early stages but in such forms as Podarke with equally cleavage, very early differentiation occurs, and the prominence of these early functioning parts varies according to the size of the initial cell from which they are formed. In forms with unequal cleavage, the first somatoblast differs in size from the remaining members of the same quartette. Beginning with Amphitrite, the relative size of d? increases suc- cessively in Chaetopterus, Arenicola, Nereis, Capitella, Aricia, Spio, Clepsine and Bdellodrilus. Those forms with equal cleav- age pass through a distinct trochophore stage and are charac- terized by an almost equatorial prototroch, a very large exum- brella, and with a very slow trunk development. In those with unequal cleavage, especially in the second generation of ecto- meres, there is a gradual decrease in the prominence of the EMBRYOLOGY OF BDELLODRILUS 213 trochophore to its approximate or complete disappearance; on the other hand, there is a gradual acceleration in the time of the trunk development, varying according to the increase in the relative size of the first somatoblast or X. Treadwell states that the extra amount cf material stored in the macromere D is in some way related to the amount of somatic and mesoblastic material needed in the future organism. This statement is true of the condition that occurs in such annelids as Bdellodrilus and Clepsine. GENERAL SUMMARY The undivided egg of Bdellodrilus philadelphicus is nearly oval. Its median longitudinal axis through the region of the polar bodies corresponds to the median axis of the future adult. The polar bodies occupy the region which later becomes the anterior end of the embryo. The first cleavage plane is nearly at right angles to the median axis of the resulting individual, and divides the egg into two very unequal parts. The second cleavage occurs at an angle of about forty-five degrees to the first. It divides the smaller cell nearly equally and the larger cell very unequally; the larger cell divides first. In a four-celled embryo the large cell D is posterior, B is anterior, inclined a little to the right; A left and, C right. The ectoderm is separated from the four macromeres by a series of three oblique cleavages. The first generation of ecto- meres is formed in a dexiotropic direction. The second genera- tion laeotropically and the third in a dexiotropic fashion. In the fourth generation of micromeres, d‘ is mesoblastic. The other cells of the fourth quartette, together with the four macromeres, form the entoderm. The cleavage of the ento- dermal cells is carried to the end without delay, in the forma- tion of the digestive tract, and the interior of the embryo be- comes a solid mass of entodermal cleavage cells, which later become differentiated into the epithelial portion of the alimen- tary canal. As the core of entodermal cells grows posteriorly, 214 ' GEORGE W. TANNREUTHER it separates the primary and secondary mesoderm. The ex- treme ends of the digestive tract are ectodermal. Among other annelids and also in molluscs, so far as is known, the entodermal cells are not broken up into cells but enter directly into the for- mation of the digestive tract. The posterior cell, d? (X), of the second generation of ecto- meres is the largest cell of the segmenting ovum. The deriva- tives of X are symmetrically placed with reference to the median plane of the future individual. The large cell, X, gives rise to the trunk ectoderm, the nervous and the excretory systems. The nervous system is derived from the two neuroblasts. The brain is formed from the extreme anterior end of the neural rows. The largest cell, d* (M), of the fourth generation of micro- meres, gives rise to the entire mesoderm. It is the first cell to divide in a bilaterally symmetrical manner. The primary mesoblast cells, M, M, bud off five or six small cells each, beneath the first quartette of ectomeres, which give rise to the secondary mesoderm on the dorsal side of the embryo. Immediately after these small cells are budded off, the primary mesoblasts, by a teloblastic proliferation of cells, produce the mesoblast bands. The embryo increases but little in bulk before the germ bands are formed. The embryo as a whole, during its early stages of development, is extremely plastic and may vary considerably in its transverse and longitudinal axes. The developing embryo is completely turned on itself, and the anterior and posterior ends are in immediate contact. The outer surface is ventral and the turned in portion is dorsal. This peculiarity of develop- ment is foreshadowed in the position taken by the early cleavage cells. At the beginning of the germ-band formation, the embryo begins to rotate on its transverse axis. This movement is due to the action of cilia, which are produced by the ectodermal cells on the median ventro-anterior end of the embryo; the ro- tation alternates. As growth continues within the cocoon, the ends of the em- bryo soon begin to overlap. The embryo may assume almost EMBRYOLOGY OF BDELLODRILUS 215 any position in the cocoon during its later stages of development. The embryo is completely developed before emergence; the tro- chophore stage is completely suppressed; the gastrulation is of the epibolic type. Columbia, Mo. November 3, 1914 LITERATURE CITED Batrour, F. M. 1893 Elements of embryology. London. Bere, R. S. 1888 Zur Bildungsgeschichte der Excretionsorgans bei Crido- drilus. Arb. a. d. zool. Inst. Wiirzburg, Bd. 7, Heft 3. BiaeLtow, A. M. 1899 Notes on the first cleavage of Lepas. Zool. Bull., vol. 2; no. 4. Bircer, O. 1891 Beitrage zur Entwicklungsgeschichte der Hirudineen. Zool. Jahrb., Bd. 4. Carazzi, D. 1900 L’embriologie dell’Aplysia limacina. Anat. Anz., Bd. 18. CasTEEL, B. D. 1904 The cell lineage and early development of Fiona marina, a nudibranchiate molluse. Proc. Ac. Nat. Sci., Philadelphia. CastLe, W. E. 1896 The early embryology of Ciona intestinalis Flemming (L). Bull. Mus. Comp. Zo6él. Harvard College, vol. 27, no. 7. Cuitp, C. M. 1900 The early development of Arenicola and Sternaspis. Arch. f. Entwick. der Organismen, Bd. 9, Heft 4. Conkuin, E. G. 1897 a The embryology of Crepidula. Jour. Morph., vol. 18. 1897 b Cleavage and differentiation. Biol. Lecture, Woods Hole. Eisic, H. 1898 Zur Entwickelung der Capitelliden. Mittheil. a. d. Zool. Stat. Neapel, Bd. 13. Frareont, J. 1887 Le genre Polygordius. Fauna u. Flora d. Golfes von Neapel, Monographie, tom. 24. GortE, A. 1882 Ueber die Entwickelung der Chaetopoden. Leipzig. Harscuex, B. 1885 Entwickelung der Trochophora von Eupomatus uncina- tus. Arb. zool. Inst. Wein, Bd. 6. Hearn, H. 1899 The development of Ischnochiton. Zool. Jahrb., Bd. 12, Heft. 4. Hertic, O. unpD R. 1881 Die Coelomtheorie. Jenaische Zeitschr., Bd. 14. KLEINENBERG, N. 1879 The development of the earthworm. Quart. Jour. Mier. Sci., vol. 19. Kowatevsky, A. 1871 Embryologische Studien an Wiirmern und Arthropoden. Mem. de |’Acad. Imp. de Sc. de St. Petersbourg, tom. 16, no. 12. Lanc, A. 1884 Die Polycladen des Golfes von Neapel. Flora und Fauna des Golfes von Neapel, Bd. 6. Lituiz, F. R. 1895 The embryology of Unionidae. Jour. Morph., vol. 10. 1899 Adaptation in cleavage. Biol. Lecture Woods Hole. 1901 The organization of the egg of Unio. Jour. Morph., vol. 17. Meap, A. D. 1897 The early development of marine annelids. Jour. Morph., vol. 13. 216 GEORGE W. TANNREUTHER Meyer, E. 1890 Die Abstammung der Anneliden. Biol. Centralb., Bd. 7. Moore, J. P. 1895 The anatomy of Bdellodrilus illuminatus, an American Discodrillid. Jour. Morph., vol. 10. Neuson, A. J. 1904 The early development of Dinophilus. Proc. Ac. Nat. Sci., Philadelphia. Route, L. 1899 Etudes sur le développément des annelides et en particulier d’un Oligochéte limicole marin (Enchytraeoides marioni). Ann. Sci. Nat., (7), tome 7. SaLensky, W. 1885 Développément de Branchiobdella. Arch. de _ Biol., tom. 6, fase. ip: 1h ScuimMKEwitscH, W. 1895 Zur Kentniss des Baues und des Entwickelung des Dinophilus vom weissen Meere. Zeit. wiss. Zool., Bd. 59. Torrey, J. C. 1903 The early embryology of Thalassema mellita (Conn). Ann. New York Acad. Sci., vol. 24, no. 3. TREADWELL, A. L. 1898 Equal and unequal cleavage in annelids. ‘Biol. Lecture Woods Hole. 1901 The cytogeny of Podarke obscura. Jour. Morph., vol. 17. Wuitman, C. O. 1878 The embryology of Clepsine. Quart. Jour. Micr. Sci., vol. 18. 1887 The germ layers in Clepsine. Jour. Morph., vol. 1. Witson, E. B. 1887 The germ bands of Lumbricus. Jour. Morph., vol. 1. 1889 The embryology of the earthworm. Jour. Morph., vol. 3. 1891 The origin of the mesoblast bands in annelids. Jour. Morph., vol. 4. 1892 The cell lineage of Nereis. Jour. Morph., vol. 6. Wierzesski, A. 1905 Embryologie von Physa fontinalis. Zeit. f. wiss. Zool., iBd-"'62: THE MORPHOLOGY OF NORMAL FERTILIZATION IN PLATYNEREIS MEGALOPS EK. E. JUST THREE PLATES (THIRTY FIGURES) 1. INTRODUCTION In a previous paper on Platynereis megalops, which described the egg-laying habits, it was stated that insemination takes place in the body cavity of the female and, further, that the eggs will not fertilize when inseminated in sea water. The present paper is a description of the normal fertilization proc- ess in Platynereis. An experimental analysis of fertilization in Platynereis appears elsewhere (Just, ’15). 2. NORMAL FERTILIZATION OF PLATYNEREIS The living egg. The egg of Platynereis is compressed and irregular in shape while in the body cavity. Those eggs which happen to be uninseminated when laid gradually round out in sea-water as almost perfect spheres equatorially, but with a rather shorter polar axis. The large, centrally placed, germinal vesicle is slightly elongated -in the polar direction. The largest eggs, fully rounded out, measure 180 to 200 uw. They are almost perfectly transparent, have an equatorial ring of oil drops, and a well marked transparent exoplasm or cortical layer of proto- plasm with very faint granules forming a delicate mesh. In short, the living egg closely resembles that of Nereis; it is larger (but cf. Wilson, ’92) not so deeply pigmented, and lacks the characteristic yolk spheres of the Nereis egg. A. Fertilization in the living egg We may consider the fertilization of the egg under the follow- ing heads: (1) insemination, (2) penetration of the sperm, and (3) copulation of the germ nuclei. 217 218 E. E. JUST 1. Insemination. In Platynereis insemination normally takes place in the body cavity ( Just, ’14). The eggs, when laid, have the sperm attached within a thin hull of jelly, the secre- tion of the cortical layer. If the worms be allowed to deposit eggs in India ink ground up in sea-water it can be proved satis- factorily that a hull of jelly, as in Nereis, envelops inseminated eggs. This jelly, absent in uninseminated eggs, is formed from the exoplasm of the egg as the result of stimulation through sperm attachment. In sea-water the zone between India ink particles and the vitelline membrane gradually widens, not so much because of the slow diffusion of the jelly from the egg, as because of the swelling of the extruded jelly. _ Insemination in some way brings about oviposition. The presence of the sperm in the female is a stimulus to egg laying; as in Nereis (see Lillie and Just) the presence of the sperm in the sea-water brings about the shedding of the eggs. The first result of the attachment of the sperm to the egg is jelly formation through cortical secretion, with the consequent formation of the perivitelline space; and this process must begin in the body cavity, since eggs have a thin jelly investment when laid. As in Nereis the vitelline membrane is preformed; the sperm does not cause ‘membrane formation.’ For twenty to thirty minutes after oviposition, the sperm remains external to the egg. During this time profound changes take place in the egg, many of which doubtless are to be inter- preted as changes incident to maturation, the mechanism of which is released with the breakdown of the’ cortical substance and the consequent formation of the perivitelline space. These changes: breakdown of the germinal vesicle, formation of the spindle, polar body formation, and cytoplasmic movements, are easily followed in the living egg. 2. Penetration. In Nereis a striking phenomenon of sperm attachment is the fertilization cone (Lillie, ’11, 712). In Platy- nereis no sharply defined cone is found. There are, however, cytoplasmic disturbances at the point of sperm entry. In polyspermic eggs the cytoplasm may form a low blunt protrusion, with as many as five spermatozoa attached to it, but this is not FERTILIZATION IN PLATYNEREIS MEGALOPS 219 a cone. Twenty-five minutes after oviposition a slender strand of protoplasm may be discerned across the perivitelline space and beneath the point of sperm entry; even this does not seem to be constant, but appears to be formed only in the animal hemisphere. This protoplasmic strand lies, first, in a radius of the egg, but gradually bends so that it now lies almost tan- gential to the egg. It is found after sperm entry. After thirty minutes the spermatozoon is engulfed. Often the formation of the sperm aster is discernible, the middle- piece and tail remaining outside. The maturation asters are always visible in the living egg. Mathews (’06) has called attention to the difference between the structure of the asters of living eggs and of fixed material. The difference is certainly striking in Platynereis. Instead of the short stiff astral fibres of chrom-osmic material or the long slender ones of mercuric fixation, one sees in the living egg beautiful broad rays sweep- ing through the cytoplasm. 3. Copulation of the germ nuclei. About fifty minutes after egg-laying, the germ nuclei copulate, the cleavage asters form, and at sixty minutes the egg divides unequally. The egg at this time exhibits a stratification of protoplasmic stuffs. Dur- ing maturation the cytoplasmic currents shift the materials. The equatorially placed oil drops, about eighteen in number, grad- ually become massed at the vegetative pole, the coarser (yolk) granules lie above these; at the clearer animal pole are the male and female pronuclel. Beneath the polar bodies the cytoplasm is most transparent. The asters are very distinct. One can- not get an adequate picture of these structures from sections. In the living egg they are incomparably clear; large broad rays which bear little resemblance to the short stiff fibres seen in the sections. The penetration path of the spermatozoon may often be followed, the copulation path always followed. The sperma- tozoon enters at any point of the egg and through this, as in Nereis (Just, ’12), the first cleavage plane passes along the copu- lation path of the germ nuclei. 220 E. E. JUST B. Observations on the sectioned egg Observations of the phenomena of fertilization in the living ege were supplemented with a study of sectioned material. Technique. Eggs were fixed in Meves’ fluid for thirty minutes, one hour, or twelve hours. Aceto-osmic-bichromate mixtures (Mathews,! 99; Bensley); Bouin’s fluid, modified by the addition of an equal volume of water; and Gilson’s mercuric-nitric mix- ture were likewise used. Although very destructive to the yolk and oil, the modified Bouin proved helpful in the study of cer- tain details in connection with sperm penetration. The difficulties of fixation, which are great in this egg, as in Nereis, may in large measure be overcome by the subsequent treatment. The following methods were used after fixation with Meves: (a) Clearing with double distilled anilin oil from 80 per cent alcohol. (b) Clearing in cedar oil from 95 per cent alcohol. (ec) Clearing in cedar oil from 95 per cent alcohol after treat- ment with glycerine (eggs put in 70 per cent alcohol plus an equal amount of glycerine). (d) Clearing in xylol from 95 per cent alcohol or from absolute alcohol. In all cases xylol was used before imbedding in paraffin or in paraffin with some admixture of Johnston’s rubber-asphalt mass. It was found that avoidance of absolute alcohol left the eggs less brittle and therefore less refractory in cutting. By far the most natural contours of both the Platynereis and the Nereis eggs are preserved through the use of aniline oil after 80 per cent—a clearing agent that I have used successfully for several years. Staining was with iron hematoxylin alone. Sections were cut four micra thick. Spermatozoa, after fixation, were studied for the most part unstained after the methods of Koltzoff, de Meyer, etc. The iodine mixture recommended by Mayer for Volvox proved in- 1From the legend of Mathews’ figures it appears that he used aceto-osmic- bichromate mixtures. - FERTILIZATION IN PLATYNEREIS MEGALOPS PPX valuable. For permanent preparations Bensley’s staining mix- tures were used. 1. Stages previous to the penetration of the sperm. The egg of Platynereis rivals in structure the beauty of the Nereis egg. A section of an uninseminated egg (fig. 1) teased out of the. female directly into Meves’ fluid gives many of the details. The cytoplasm is sharply marked off into two regions: the exoplasm made up of clear cortex and zone of oil and yolk and the deeply staining endoplasm. The outer portion of the exoplasm is a mesh of pale blue deli- cate fibrils, the alveoli of the cortical jelly. The outer limits of this cortical layer—slightly more dense than the deeper portions— is studded with black granules immediately below the vitelline membrane. The inner border of the cortex arises from a zone of closely-packed, deep-staining bodies, from which apparently the walls of the cortical alveoli project. Below this inner border is the region of oil drops which lies in the equatorial zone, among spherules which prove, from their later behavior, to be yolk spheres, although even in the best preparations, the fine granules of which they are composed tend to shrink from their spherical walls (cf. Lillie, ’11; figure of Nereis egg fixed in Fleming). These yolk spheres are evenly crowded against the deeply stained basal area of the cortex. Around the germinal vesicle and closely applied to it is the endoplasmic mass, made up of fine granules which take the stain very tenaciously. Its outer limits are uneven, encroaching on the area of oil drops and yolk spheres as blunt projections. ‘Seattered throughout the germinal vesicle, as in Nereis, are the chromosomes—fourteen tetrads. These lie among many black granules of varying size. Although an attempt has been made to study their number, distribution, etc., and to ascertain any constant characters, nothing now can be said further of them. These granules tend to be spherical and to grade down to minute bodies. The whole egg, therefore, exhibits a granular structure, both living and fixed, as Mathews some time since (’06) for echino- derm eggs and more recently Kite for some other eggs have shown. 222 E. E. JUST Lillie (06), too, in a most elaborate study on the egg of Chaetop- terus, has determined the granular structure of the cytoplasm. Vacuoles found in mercuric-nitrie or picro-acetic preparations are filled with yolk or oil in Meves’ preparations or in the living Platynereis egg (cf. Wilson, 798, on the cytoplasmic structure of eggs, including that of Nereis). The egg of Platynereis, as compared with that of Nereis fixed with the same methods, does not show so clearly the radial striation in the cortical layer or the homogeneous yolk spheres. Ten minutes after laying the germinal vesicle is breaking down and maturation asters, formed outside its wall, are pushing into its substance. The deepest of the cortical alveoli are often still unemptied; the whole process of jelly extrusion can easily be followed from its beginning in inseminated eggs. On one or between two of the apices of the wavy vitelline membrane the spermatozoon is found attached by its perforatorium. Sperm head, middle-piece, and tail are readily distinguished (fig. 2). Fifteen minutes after laying, the cortical jelly has been wholly extruded (fig. 3) and the first maturation spindle formed, with the chromosomes in late prophase. The endoplasm, with the extra-chromatin substance of the germinal vesicle, imbeds the spindle. Jn toto mounts of the egg at this stage, as is true of the Nereis egg, give no view of the spindle. One sees only a deeply stained core of substance which incloses the spindle. The egg is irregular in shape and the vitelline membrane is closely applied. The spermatozoon is visible on the membrane (figs. 3 and 4) above a group of granules similar to those more thinly scattered throughout the periphery of the egg. These granules are mark- edly like those described by Meves and are doubtless ‘mito- chondria;’ but in Platynereis they cannot possibly have the significance that Meves ascribes to them in the eggs of various forms. The granules appear massed beneath the point of sperm entry, but these masses assume no definite form. I have pur- posely figured those that give the nearest approach to cone formation (figs. 3, 4, and 5). A slender strand of cytoplasm may extend toward the membrane just below the perforatorium. FERTILIZATION IN PLATYNEREIS MEGALOPS 223 The granules in the region may appear as a disc, but never as a retracted cone, as in Nereis. The cortical breakdown has released the close application of the yolk spheres to the inner cortical margin; they are now irregularly spaced and among them lies the granular cytoplasm. The figures (2 to 5) also give good pictures of the spermatozoa. They are much like the living spermatozoon. The head is almost spherical, the perforatorium a large blunt cap; the mid- dle-piece and tail are often clearly defined. 2. Penetration of the spermatozoon. The penetration of the sperm head begins at twenty to twenty-six minutes after laying (cf. Nereis, forty-five minutes after insemination). The first maturation spindle, in the metaphase, is oriented in the polar plane of the egg; the inner endoplasmic mass which incloses the spindle is, at this stage, triangular in section; the outer aster of the spindle is near the apex of the triangle. The base of the triangle is less blunt than in previous stages and reaches farther outward along a radius of the egg. The various stages of pene- tration are shown in figures 6 to 18. The sperm substance enters the egg as a slender black thread, which gradually increases in size at its inner end. The sperm head, in my preparations, is usually homogeneously black, but often the external bulb is not so dark; or lighter areas appear along the entering thread; (particularly figs. 10,11, 12, and 14). Often, especially in sec- tions stained for twelve hours only, in stages just after the attach- ment of the perforatorium to the cytoplasm, the head appears, not as a homogeneous chromatin mass, but as a slightly differ- entiated body. One gains, therefore, the impression that the spermatozoon flows into the egg (cf. Koltzoff and Lillie, who, with different methods, find Nereid spermatozoa extremely ductile). Cytoplasmic changes due to sperm entry are clearly marked during the later stages of penetration; striae appear in the cyto- plasm around the entering spermatozoon, the area stains more deeply, and a projection from the endoplasmic mass reaches out toward the point of sperm entry (see figs. 14 to 17) (cf. on these points, Foot, Gardiner, Vedjovsky, Jenkinson, and Lillie, ’12). 294 E. E. JUST As the head is drawn into the egg, the inner bulb turns with its growth. Finally, the portion forming the external bulb is engulfed. ‘The middle piece and tail, asin Nereis, never enter the egg (fig. 17). They may often be found in sections out- side the membrane after penetration of the sperm head (see figures). I have never found the spermatozoon in the Nereis or in the Platynereis egg at the time or in the form figured by Wilson (96 and ’00). Does the sperm head rotate? I could not positively determine the rotation of the sperm head in the egg of Platynereis. In the first place, a definite cone organ, like that of Nereis, is lack- ing, and secondly, the middle-piece does not enter the egg. The history of the sperm penetration is known practically for every minute from entrance to pronuclear copulation. Meves’ fixation alone was not depended upon. The Bouin preparations gave results much like those of Bonnevie’s with picro-acetic mixtures on the Nereis eggs. While absolutely worthless for cytoplasmic detail, they were helpful in determining the struc- ture of the sperm nucleus after penetration. The evidence favors rotation; the turning of. the inner sperm bulb (fig. 17) and the position of the long axis of the sperm and aster as often found at right angles to the radius of the egg (fig. 22). The sperm aster does not arise until the nucleus is beyond the yolk region (figs. 14 to 22). Within the endoplasm, the aster once formed, quickly divides equally, but the amphiaster does not long retain its equal poles, for one sperm centrosome and its aster gradually dwindle in size. Rays arise between one or both of the sperm centrosomes and the inner centrosome of the matura- tion spindle, thus forming a secondary spindle. The sperm nucleus lies nearer the larger sperm centrosome (see figs. 20 to 25). 3. Copulation of the germ nucler. The egg chromosomes, after the formation of the second polar body, form fourteen chromo- some vesicles which fuse to establish the egg nucleus (figs. 26, 27), all vestiges of the egg aster disappearing. ‘The sperm nucleus enlarges as its asters become smaller. At the time of apposition, but one sperm aster is found (fig. 28). I believe that one sperm FERTILIZATION IN PLATYNEREIS MEGALOPS 225 aster begins to wane soon after the formation of the homody- namic amphiaster and finally disappears. One aster can always be found (fig. 29). The opposing nuclear membranes break down and one nucleus forms with the single sperm aster. Soon a small aster appears on the nuclear membrane (fig. 30), the nucleus breaks down, and the heterodynamic first cleavage spindle forms. 3. DISCUSSION The case of Nereis and. Platynereis, with respect to the en- trance cone offers an interesting parallel with that of Toxop- neustes and Arbacia (Wilson and Mathews). In both Nereis and Platynereis, however, the middle-piece is left outside the egg. The absence of cone-organ in Platynereis makes the ques- tion of rotation obscure, whereas in Nereis the evidence is indis- putable. Bonnevie (’08), in her paper on Nereis, has mentioned cer- tain cytological differences between the ‘large and small varie- ties’ of Nereis eggs. As indicated above, the time of sperm entry is earlier in Platynereis. It is also true that the polar bodies are formed earlier, the first cleavage is earlier, and the subsequent rhythms are faster, so that the larval stage is reached earlier. So far as both Nereis and Platynereis are concerned, the rdéle of the middle-piece or its contained centrosome, as the chief actor in fertilization, is wanting. There are spermatocytes with intra-nuclear centrosomes? (see Julin on Styleopsis). But if this hypothesis be postulated (cf. Packard in 1914) for Nereis sperm, this next step, as was pointed out by Lillie in 1912, should also be taken: the centrosome zradient must be quantitatively different from its base at the middle-piece to the tip of the sperm head: “If intra-nuclear centrosomes are the causes of the formation of the sperm aster, not only must they exist at every level, but also (that) they must decrease in size from the base to the apex of the sperm nucleus!”’ 2See also Hegner and Newman for intra-nuclear centrosomes in oocytes. 226 E. E. JUST According to Schaxel, the middle-piece does not enter the egg of echinoderms. Meves will not admit this for echinids and doubts that in Nereis the middle-piece is left outside the egg while denying the centrosome the chief part in fertilization. In Platynereis, as in Nereis, by diverse methods it can be shown that the middle-piece does not enter the egg. We are thus forced to conclude that, whatever its rdle, the middle piece in Platynereis can play no part, either in heredity or through a centrosome in the dynamics of fertilization. Marine Biological Laboratory Woods Hole, Mass. LITERATURE CITED BonNEvIE, Kristine. 1908. Chromosmenstudien II. Heterotypische Mitose als Reifungscharakter. Arch. fiir Zellforschung, Bd. 5. Brnstey, R. R. 1911. Studies on the pancreas of the guinea pig. Am. Jour. Anat., vol. 12. Foot, K. 1897. Origin of the cleavage centrosomes in Allolobophora. Jour. Morph., vol. 12. GARDINER, Ep. C. 1898. The growth of the ovum, formation of the polar bodies, and the fertilization in Polychoerus caudatus. Jour. Morph. vol. 15. Heraner, R. W. 1908. Intra-nuclear mitotic figure in primary oocytes of a copepod, Canthocampus. Biol. Bull., vol. 14. JENKINSON, J. W. 1904. Maturation and fertilization of the axolotl egg. Quar. Jour. Micros. Sci., vol. 48. Juin, J. 1893. Structure et dévéloppement des glandes sexuélles, ovogénése spermatogénése et fécondation chez Styleopsis grossularia. Bull. Se. de France et Belgique, 24. Just, E. E. 1912. Relation of the first cleavage plane to the entrance point of thesperm. Biol. Bull., vol. 22. 1914. Breeding habits of ile heteronereis form of Platynereis megalops. at Woods Hole, Mass. Biol. Bull., vol. 25. 1915. An experimental analysis of fertilization in Plaines mega- lops. Biol. Bull., vol. 28. Kirrt, G. L. 1918. Studies on the physical properties of protoplasm. Am. Jour. Physiol., vol. 32. 5 Koutzorr, N. K. 1909. Studien iiber die Bestalt der Zelle. II. Untersuchung- en iiber das Kopfskelett des tierschen Spermiums. Arch. fur Zell- forsch., Bd. 2. Linuiz, F. R. 1906. Observations and experiments concerning the elementary phenomena of embryonic development in Chaetopterus. Jour. Exp. Zool., 3, 1906. FERTILIZATION IN PLATYNEREIS MEGALOPS QF. Linuiz, F. R. 1911. Studies of fertilization in Nereis. I. The cortical changes in the egg. II. Partial fertilization. Jour. Morph., vol. 22., 1912. III. The morphology of the normal fertilization. IV. The fertilizing power of portions of the spermatozoon. Jour. Exp. Zool., vol. 12. Linu, F. R. anp Just, E. E. 1913. Breeding habits of the heteronereis form of Nereis limbata at Woods Hole, Mass. Biol. Bull., vol. 24. Matuews, A. P. 1899. The changes in the structure of the pancreas. Jour. Morph., vol. 11. 1906. A note on the structure of the living protoplasm of echinoderm eggs. Biol. Bull., vol. 11. 1907. A contribution to the chemistry of cell-division, maturation, -and fertilization. Am. Jour. Physiol., vol. 18, No. 1. Newman, H. H. 1912. Maturation of the armadillo egg. Biol. Bull., vol. 23. PackarpD, C. 1914. The effect of radium radiations on the fertilization of Nereis. Jour. Exp. Zool., vol. 16. VepJovsky, F. unpD Mrazex, A. 1903. Umbildung des Cytoplasma wihrend der Befruchtung und Zellteilung. Nach der Untersuchungen am Rhynchelmis-Ei. Arch. fiir mik. Anat., Bd. 62. Wiuson, E. B. 1892. The cell lineage of Nereis. Jour. Morph., vol. 6. 1897. Centrosome and middle piece in the fertilization of the sea- urchin egg. Science, vol. 5, No. 114. 1899. On the protoplasmic structure in the eggs of echinoderms and ‘some other animals. 1900. The cell in development and inheritance. The Macmillan Co. Witson, E. B. anp Maruews, A. P. 1895. Maturation fertilization, and po- larity of the echinoderm egg. Jour. Morph., vol. 10. JOURNAL OF MORPHOLOGY, VOL. 26, NO. 2 DESCRIPTION All figures were drawn with the camera lucida with Leitz 7: oil immersion objective and No. 5 ocular, except where otherwise stated. All figures from sections of inseminated eggs of Platynereis megalops. All sections from eggs killed in Meves’ fluid and stained in iron haematoxylin. PAB: EXPLANATION OF FIGURES 1. Section of an unfertilized ovocyte. The oil drops are a delicate brown, the granular yolk spheres very lightly stained. The cortex is intact. 2. Ten minutes after laying. The cortex is partially reduced. The head, middle-piece, and tail are clearly shown. 3 to 5. Fifteen minutes after laying. The granules are massed below the point of sperm attachment; the perforatorium is still attached to the membrane 6 to 8. Twenty minutes after laying. 6. The perforatorium is touching the cytoplasm. The granular mass has disappeared. 7. The perforatorium is in the cytoplasm. 8. A somewhat tangential section, showing the very beginning of penetration. 9 to 13. The penetration stages, twenty-five minutes after laying, mesophase of the first maturation division. ‘The figures show that there is no constant disposition of granules at the point of sperm entry—certainly nothing of the nature of a cone, as in Nereis. : FERTILIZATION IN PLATYNEREIS MEGALOPS PLATE 1 E. E. JUST PLATE, 2 EXPLANATION OF FIGURES 14, 15. Penetration stages, twenty-five minutes after laying. Note the turn- ing of the inner sperm bulb. 16 to 18. Twenty-seven minutes after laying. The middle piece is shown in fig. 17. In 16 and 18 the middle-piece was found in adjacent sections. 19 and 20. Thirty minutes after laying, telophase, first maturation division. The sperm head is still within the zone of oil drops and without an aster. 21 to 22. Thirty-two minutes after laying; early prophase, second maturation division. 21. The sperm head is at right angles to a radius of the egg, the aster forms around the granule at the tip of the sperm head. 22. Formation of sperm aster within the endoplasm. 23. Thirty-five minutes after laying. The amphiaster is in contact with the egg aster. The spermatozoon is in an adjacent section. 230 FERTILIZATION IN PLATYNEREIS MEGALOPS E. E. JUST PLATE 2 vr 18 19 Bae: eT | & cn . », lest \ nae ee ee f Se >, fy \ 5 ee Ls % Q 22 / 21 20 PLATE 3 EXPLANATION OF FIGURES 24 to 30; oc. 1, jy oil im. Later stages, showing marked inequality of sperm asters. Note relation of the spermatozoon to the larger aster. 25a and b. Forty minutes after laying. 26. Forty-six minutes after laying. The egg aster is degenerating. 27 to 29. Copulation stages. 27. The smaller sperm aster could not be found. Two egg and three sperm nuclear vesicles are shown. 28 to 29. Formation of the male and female nuclei. 30. Origin of the first cleavage spindle. 232 FERTILIZATION 1N PLATYNEREIS MEGALOPS PLATE 3 E. E. JUST CILIATED PITS OF STENOSTOMA WM. A. KEPNER AND J. R. CASH University of Virginia FOUR FIGURES The material for this paper was found upon glea and sediment deposited upon submerged leaves and twigs taken from both near the surface and the bottom of pools in the vicinity of the University. The animals are about | to 2 mm. long and 200 micra broad at the widest region. The body is oblong, spindle shaped, and is widest in the region of the mouth; it is grayish white and its epidermis, which contains rhabdites, is thickly covered with short cilia. The mouth is on the ventral side, about 150-200 micra posterior to the anteriorend. The urinary system consists of a single nephridium. The ciliated pits are by far the most striking organs of the body which are externally visible. These are small invagina- tions in the epidermis which are located lateral-dorsally about 100 micra from the anterior end of the body. Their shape is, in general, similar to small sacs, but their lateral walls are highly contractile so that the pit may be made to assume the form of a deep cut or that of a shallow, concave disc. The ciliated pits have a diameter of about 50 micra and a depth of about 40 micra. This paper is concerned with the minute anatomy of the ciliated pits and their development. In order to carry out such a study the animals were fixed in Flemming’s stronger fluid, which consists of 15 parts 1 per cent chromic acid, 4 parts 2 per cent osmic acid, 1 part glacial acetic acid; time of fixing 25 minutes. Chrome-aceto-formaldehyde, hot and cold solutions of aceto- sublimate, and Zenker’s fluid were all tried, but without success, 235 236 WM. A. KEPNER AND J. R. CASH either causing great distortion or disintegration of the animals. The worms were cut into sections, some three and some five micra thick and the sections stained with iron haematoxylin and counter stained with Bordeaux red. Macerations stained with such intra vitam stains as Wright’s stain and methylen blue were very valuable in corroborating results. HISTOLOGY, OF. THE: PIT The histology of the pit involves an understanding of the epidermis. The epidermal cells of the animal have the power to secrete a protective mucus-like substance. That such is the case can readily be seen by placing the animal in an abnor- mal solution of not too rapid killing power. The animal will at once enshroud itself in a thick sheath of protective mucus within which it swims around. Such a phenomenon will be more fully described in a later part of this paper. The pit is associated with a region of the central nervous system known as the ciliated pit-ganglion. The pit, as well as this ganglion, is a modified region of the general epidermis. The marginal walls of the pit are formed by cells transitional in structure between the general epidermis and the low cells at the bottom or fundus of the pit. As the invagination which forms the pit takes place this transition takes place until there is a layer of low, small epithelial cells lining the fundus of the pit. The boundaries of these fundus cells are less pronounced than the boundaries of the general epithelial cells and their nuclei less frequent. In certain regions of this lining of the fundus the few nuclei which are present are indefinitely placed, which fact suggests that there is no basement membrane. On the exterior of the body, lying close upon the fundus of the pit, is a homogeneous mass of mucus-like substance. The mar- ginal walls of the pit are thickly covered with cilia which appear to be longer than the cilia of the general body epithelium, but no cilia at all were found upon the low cells lining the fundus nor were any seen projecting above the homogeneous mass of mucus-like substance. CILIATED PITS OF STENOSTOMA 237 The ciliated pit-ganglion is by far the most conspicuous fea- ture of the pit. It is located just within. the body and lying around the base of the ciliated pit in a cup-like manner. The cells of this structure are only indistinctly separated from those of the dorsal ganglion or ‘brain’ by a few muscle fibers and have the same characteristic, granular nuclei as those of the ‘brain.’ Some of these ciliated pit-ganglion cells are seen to send processes through the epithelium of the fundus which les in contact with the homogeneous body of mucus-like substance. These we take to be sensory rods of highly special nature which enable the organ to detect very slight changes in its surrounding medium. These sensory rods are shown clearly, as dark blue structures, in intra vitam staining with Wright’s stain and in many of the regular sections. Figure 4 shows such a section. The fact that the ganglion cells arise from the epithelial cells which line the fundus of the pit also supports the idea that these processes are left behind by the cells as they migrate inward to enter into formation of the ciliated pit-ganglion. Rightful interpretation and appreciation of the above state- ments will only be obtained through a study of the origin of the ciliated pits. THE ORIGIN OF CILIATED PITS The origin of the ciliated pits can readily be studied in speci- mens which are dividing. The first appearance of the pits is seen in two sharp, abrupt depressions of the epidermis, one on each side of the animal. From the bottoms of the depressions (i.e., the region which will be the fundus of the new pit) the cilia disappear. Figure 1 shows an early pit in this stage of develop- ment. Ventral to this abruptly lowered region a crowded mass of mesenchymal cells is formed which represents the anlage of the ‘brain.’! 1 We are indebted to Prof. Bohmig, through the kindness of Prof. L. von Graff, for the following quotation from page 34 of O. and R. Hertwig’s Die Coelom- theorie, Jena, 1881. In regard to the Platyhelminthes they say: ‘‘In der Abteil- lung stammt wahrscheinlich der motorische Teil der Centralorgane des Nerven- systems im Anschlusse und die Muskulatur aus dem Mesenchym, der sensorielle Teil im Anschlusse an die Sinnesorgane aus dem Ektoderm.”’ 238 WM. A. KEPNER AND J. R. CASH Fig. 1. An early stage in the formation of the ciliated pit and its ganglion. (end.) Endoderm or wall of enteron. (g) General epidermis lowered at (a) to form the rudiment of the fundus of pit. Note absence of. cilia in this lowered region and migrating mitotic cells (me, m’c’). (mes) Mesodermal cells crowded about the forming ciliated pit-ganglion (g). > 1500. This sharp depression is already a rudimentary pit with its non-ciliated fundus and its ciliated marginal walls but lacks a ciliated pit-ganglion. At this time, about the fundus of the rudimentary pit mitoses arise which send into the mesenchymal space between the fundus and the anlage of the ‘brain,’ which has already been formed, a proliferation of cells which radiate from beneath the developing fundus of the pit. This mass of cells is the beginning of the ciliated pit-ganglion. Thus there are established at the outset two parts of the ciliated pit. a) The epithelium of the pit; b) The rudiments of a ciliated pit-ganglion. CILIATED PITS OF STENOSTOMA 239 The epithelium of the growing pit is extended as a region which is morphologically different from the general epidermis in that its cells are lower and are repeatedly dividing to yield additional cells to the formation of the ciliated pit-ganglion; also in that the cells which line the fundus of the pit have irregularly placed nu- clei, and have lost their cilia, while the cilia on the marginal cells have become longer than the cilia on the general body epithelium. These characters are shown in figures 2 and 3. Up until this stage in their development the cells of the pit- epithelium retain their power to elaborate rhabdites, as is illus- trated in figures 2 and 3. As the pit grows larger no rhabdites are to be found in the epithelium of its fundus. But before these rhabdites have to- tally disappeared theformation of a peculiar body isstarted, which in the mature pit is a highly refractive, homogeneous layer, which Ott (92)? has called the ‘homogeneous mass.’ The nature of this ‘homogeneous mass’ can best be arrived at by observation of the specimens during fixing. As. an animal lies in contact with the slide, if it be fixed by dropping the fixing fluid upon it, it will adhere to the slide on account of the protec- tive discharge thrown out by the cells of the general epithelium. To avoid such trouble it was necessary to apply the fixing fluid with a dash as the animal swam around in a small drop of water on the slide. Thus any adhesions which would injure the fixed specimen were avoided. The details of this trouble can be plainly observed under the binocular microscope. If an entire speci- men be treated with methylen blue, a protective blue sheath of mucus with imbedded rhabdites (stained deep blue), will be seen to be formed around the animal. If this sheath be removed from the specimen and the stain again applied the epidermal cells fail to respond and the protective sheath is not formed the second time. Thus it is evident that in an effort to protect itself the epider- mis not only discharges rhabdites but also a mucus which stains with methylen blue less deeply than the rhabdites. Now, since 2 Ott; Jour. Morph., vol. 7, 1892. 240 WM. A. KEPNER AND J. R. CASH in the fundus of the mature pit there are no rhabdites, it is sug- gested that, as the cells of the developing epithelium are physio- logically differentiated, they lose their power to elaborate rhab- dites and develop a greater capacity to secrete a permanent, refractive, mucus-like glea which protects the greatly exposed and extremely sensitive fundus of the pit. So we draw the con- clusion that the only difference between the mucus secreted by the cells of the general epithelium and the mucus which com- poses the ‘homogeneous mass’ of the ciliated pit is that the ‘homogeneous mass’ is permanent, perhaps more dense, and withstands the action of reagents better than does the temporary secretion of the general epithelial cells. We have made compari- son of these two substances by staining with methylin blue, in which case they stain alike, both staining in living specimens a rather dark blue as contrasted with the intensely dark blue of the rhabdites. So much for the development of the epithelium of the ciliated pit and its secretion product. DEVELOPMENT OF THE GANGLION OF THE CILIATED PIT As has been stated previously in this paper, beneath the form- ing ciliated pit there is a mass of cells which we take to be mesen- chymal in origin. This statement is made in abeyance since we are not concerned at present with the origin of the ‘brain.’ We have, however, been able to see that the ‘brain’ arises from these cells. But the important point which we endeavor to make is that the ciliated pit-ganglion has a distinct origin from the epidermis. With the earliest formation of the pit-depression at the sur- face of the body there occur mitoses in its epithelium which send into the mesenchymal region a number of cells which locate themselves between the epithelium of the fundus and the mesen- chymal cells which form the ‘brain,’ as shown in figure 2. There is, however, a distinction at the very outset between the cells of the ‘brain’ and those which are forming the ciliated pit-gan- glion, as shown by the above figure. This proliferation of cells arising from the epidermis continues to grow with the develop- ment of the superficial part of the pit. CILIATED PITS OF STENOSTOMA 241 Fig. 2. Later stage of formation of ciliated pit andits ganglion. Note widened fundus with its low epithelium that vet has rhabdites (rh). Mitoses (me) in region of fundus continue to be present. Ganglion (g) has enlarged. (mes) Mesoderm that develops into ‘brain’ and commissure. (end) Endoderm. x 1500. Throughout the growth of this ganglionic mass of cells their nuclei have a constant chromatin pattern. At the earliest ‘ and intermediate stages of the development of these cells their nuclei tend to be oval while their cytoplasmic bodies are more or less elongated, pyriform, or spindle shaped. Figures 1, 2 and 3. In the final stages of their development the cell bodies become less distinct until in the mature ganglion there is only a network of fibers or cytoplasmic strands supporting many sphe- roidal nuclei. Figure 4. 242 WM. A. KEPNER AND J. R. CASH Fig. 3. Later stage in formation of ciliated pit and ganglion. Note appear- ance of ‘homogeneous mass’ (hm), with rhabdites (rh) yet present in the fun- dus epithelium. Ciliated pit-ganglion (g) has now fused with the fibrous part of the ‘brain’ (br); (end) Constricted enteron. (c) Commissure of ‘brain’ forming. > 1500. The interesting outcome of this development is a sensory epithelium from which many cells have retreated, leaving be- hind a low, secreting epithelium through which they leave elon- gated processes of themselves. Figure 2. These processes are the sensory ends of the ganglionic cells which have been de- scribed. Thus we have the development of a ciliated pit whose marginal cells are covered with extremely long cilia which may protect the delicate fundus against impacts of external objects by practically closing the mouth of the pit to any particles of matter which might enter and in any way injure the sensitive ’ Tt cannot be definitely stated that all ot these ganglionic cells have such processes, since on account of the nature of the case, only a few such processes in each animal can be sectioned parallel to their axes. CILIATED PITS OF STENOSTOMA 243 Fig. 4. Ciliated pit with its ‘homogeneous mass’ (hm); fundus epithelium (fe); marginal epithelium with its cilia (mge), and pit-ganglion (g) well es- tablished; the latter receiving a bundle of fibres from ‘brain’ (br). (mus) Mus- cles. X 1500. base. Over the base is spread the ‘homogeneous mass’ elabo- rated by the epidermis of the fundus into which the sensory rods of the ciliated pit-ganglion cells extend and test the chem- ical nature of the water, conveying the sensations obtained to the pit-ganglion, which merges into the ‘brain.’ In these results we have been able to confirm the descrip- tion of Ott (92) so far as the general structure of the pits is concerned, but have not, however, been able to agree with his description of the fundus of the pit. He says ‘‘the cilia of the 244 WM. A. KEPNER AND J. R. CASH epithelial cells could be seen passing through the homogeneous mass,” and that—‘“‘the cilia on the small cells at the base of the pit are like those of the epithelial cells of the integument except that they are much longer. They range from 8 micra to 15 micra in length.’ We have found that the cilia on the marginal walls of the pit are longer than those on the general surface of the body but have found in no case any suggestion of cilia re- lated to the fundus of the pit or its ‘homogeneous mass.’ Accord- ing to Ott,® Landsberg says “‘that the bottom of each pit is covered with a thick layer of homogeneous substance which may be re- garded as mucus. Below this is a thin layer of ciliated epithelial cells whose cilia project through the homogeneous layer. Next to this is a much thicker layer which is made up of mostly pyri- form cells, although there are other histological elements scat- tered through it. Next to this layer is the ganglion which is connected with the nerve.’ This description is very much in accordance with our results with the exception of the statement that there are cilia on the low cells at the fundus of the pit. Ott says “There are three possible methods by which the three layers described by Landsberg might be produced: 1) By a division of the epithelial cells, 2) by a migration of cells from the brain ganglia to the walls of the pits, 3) by a migration outward of some of the epithelial cells to form a second outer layer. If a new layer of cells was formed by the first method we ought cer- tainly tofindnumerous spindles vertice to the surface in every de- veloping pit.’”® Now we have assumed the first mentioned method of formation of the ciliated pit ganglion mainly on account of the fact that we have seen these numerous spindles as — described by Ott (fig. 2). We find them in nearly every section that passes through the fundus and ganglion. In one other respect our conclusions differ from a previous description of the general structure of the pits by von Graff (13). He says “Das Nervensvstem besteht aus zwei lang- gestreckten Halften, deren jede durch eine swache EKinschniirung 4 Jour. Morph., vol. 7, p. 291, 1892. 5Same reference, p. 291. 6 Same reference, pp. 292-8. CILIATED PITS OF STENOSTOMA 245 in eine hintere, durch eine breite Kommisur verbundene und eine kleinere, vordere Partie zerfallt. Die letztere bildet die beiden Griibschenganglien, in welche sich die Wimpergriibshen einsenken.? (Das Tierreich, s. 20.) According to our observa- tions, the ciliated pit-ganglion does not arise from the ‘brain’ but arises independently from the epidermis. In conclusion, it is interesting to observe the striking parallel- ism presented by this organ in its function and mode of origin with the olfactory organ of a vertebrate so far as its function (1.e., its function in the fish) and its mode of origin is concerned. This organ functions as a tester of the chemical nature of the water which passes through or over it as do the olfactory organs of the fish. Moreover, this organ arises as a modified region or plate in the epidermis. Some cells of this plate sink beneath the base of the plate to form the ganglion of the pit. All this is closely analagous to the following description of the origin of the olfactory ganglion as given by Minot (’92).8 The ectodermal cells of the olfactory plate multiply, the kary- okinetic figures being found next to the outer or free surface of the layer; the cells thus produced assume the appearance of medullary neuroblasts and at four weeks are found migrating toward the mesen- chymal surface, so that the base of the layer of the olfactory ectoderm becomes crowded with nuclei; the protoplasm of these neuroblasts is collected on one side of the nucleus in a pointed mass; the cells now grow forth from the ectoderm and constitute the anlage ofthe ganglion between the ectoderm and the brain. (P. 637.) CONCLUSION The ciliated pit and its ganglion in this flatworm arise from the general epithelium in a manner closely analogous to the mode of origin of the olfactory epithelium and olfactory ganglion of the vertebrates. 7 Italics our own. 8 Minot; Human Embryology. THE DEVELOPMENT OF THE ALBINO RAT, MUS NORVEGICUS ALBINUS I. FROM THE PRONUCLEAR STAGE TO THE STAGE OF MESO- DERM ANLAGE;, END OF THE FIRST TO THE END OF THE NINTH DAY G. CARL HUBER From the Department of Anatomy, University of Michigan, and the Division of Embryology, Wistar Institute of Anatomy and Biology, Philadelphia THIRTY-TWO FIGURES LTO IOV So BO Ae eee uci ats eS ote cra tot a a Pn SPREE BA ES 2 ay 247 RUE OLA ANOS UMO GIS. ccc k ee A Shieh te kd hee iw oelere'es oS cehes DHOE A Laer eee 249) OvMlaniomematiration, and tertilivatiOn.......00-.+.2.:o.+scee os oe eee 253 PeeTEIE ARES RNs 2 AM Pe ok thee oc S gna o's OE ono a uaa eee 257 PERMA ee UO MNASELS CH. tye (tare ans = cases spstetees His da. o'd's see © Sule dana eA ORS 265 BARBI SEES SS ona win An i a AN Ce SO) ke 265 Am C CBS S CMA ener ats eRe ae OE ee ld aun nS wis 6 Sno Se ee 273 IC CMIMSI ARE Certs heron Sa NAR Or Ss th tal ate CR ae eee 275 URLOmIO=Cellyst ayer ne esen gsr state f staystsieises. 200. is, that the first two blastomeres of the mammalian ovum are equivalent in size and structure if the stage is observed soon after its formation. As above stated, the 2-cell stage of the mam- malian ovum extends through a relatively long period, probably about 24 hours. The two cells do not as a rule divide synchron- ously, the division of one preceding the other by some little time, resulting in a 3-cell stage. The cell to divide first increases slightly in size and presents a clearer protoplasm prior to its division. Ina2-cell stage, viewed in this phase of cytomorphosis, one of the cells appears slightly larger with clearer protoplasm than does the other cell, explaining the difference in size and structure observed by Van Beneden and by other observers who concur in his views. I am convinced that a difference in the size of the two cells may be accounted for by the plane of section in which they are cut, even though the’ nuclei of both cells are included in the section. In the figures of sections of the 2-cell 268 G. CARL HUBER stage of the mouse, given by Sobotta and Melissinos, the nuclei of the two cells lie in about their center and essentially in the same plane. In my own material of the 2-cell stage of the albino rat it is not unusual to find the nuclei of the respective cells nearer the opposite poles of the two cells than at their centers, as shown in C, of figure 1. In B of this figure, where the two nuclei appear as lying much nearer the center of the cells, they are in reality placed much as in C, as is shown by the reconstruction. Fig. 7 Model of the right oviduct of rat No. 59, 2 days. 10. Not quite the entire oviduct was available for reconstruction, the upper end of the uterine horn thus not shown in the figure. The position of the four 2-cell stages, each of which is outlined in a circle, found within the tube, is shown as if seen through a transparent wall. To determine the position of the segmented ovum in the 2-cell stage in the oviduct, reconstructions were made of two oviducts. In figure 7 is shown a reconstruction of the right oviduct of rat No. 59, killed two days after the beginning of insemination. In preparing the material for embedding, this oviduct was cut not quite at its insertion into the uterine horn. The portion of the oviduct reconstructed measures 2.29 cm. Nine major loops are shown. The four ova in the 2-cell stage found in this tube are situated in the sixth to the seventh loop at a distance of about 1.4 em. from the fimbriated end. This portion of the oviduct falls to segment three of Sobotta’s designa- tion. It is lined by non-ciliated epithelium resting on a mucosa with inconspicuous secondary folds, but presenting four or five characteristic major folds. This portion of the oviduct is closely DEVELOPMENT OF THE ALBINO RAT 269 applied to the outside of the ovarian capsule, and conspicuous in all of the figures of models of the oviducts here presented. The detail of the distribution of the ova in the tube is given in figure 8, a reconstruction under a higher magnification of the segment of the oviduct containing the ova. The lumen is exposed so that the character of the mucosal folds may be seen. The ova are spaced in a segment of the tube measuring 3 mm., and are Fig. 8 Model of the segment of the right oviduct of rat No. 59, 2 days, con- taining the four 2-cell stages as shown in figure 7. X 50. Note the absence of prominent folds in the mucosa. The segment presented in the reconstruction measures 3 mm. The four 2-cell stages contained in this tube are relatively widely spaced. in this case more widely separated than is usual for this stage. In figure 9, there is reproduced a reconstruction of the left ovi- duct of rat No. 62, killed 2 days, 22 hours after the beginning of insemination. This tube was also cut a little before its in- sertion into the uterine horn. The portion reconstructed meas- ures 2.45 cm. In it there are found five ova in the 2-cell stage, situated about 2 em. from the fimbriated end, and in the last loop of the third segment of the oviduct. The five ova are closely 270 G. CARL HUBER grouped between two opposing folds of the mucosa. Their general relations are shown in figure 10, a reconstruction under higher magnification of the segment of the oviduct containing Fig. 9 Model of the left oviduct of rat No. 62, 2 days, 22 hours. X10. Not quite the entire oviduct was available for reconstruction, thus the relative posi- tion of the upper end of the uterine horn is not shown in this figure. Fimbriated end and infundibulum removed in the drawing, so as to expose the underlying loops; their relative position is given in dotted outline. The position of five 2- cell stages, found within this tube, is given as if seen through a transparent wall. Fig. 10 Model of the segment of the left oviduct of rat No. 62, 2 days, 22 hours, containing the five 2-cell stages, the general position of which is shown in figure 9. 50. Note the compact grouping of the ova. the ova, cut so as to expose the lumen. At the magnification used it was not possible to reproduce in the model the exact shape of the several ova, their relative position is, however, correctly given. In all, ten oviducts, containing 40 ova in the 2-cell stage, are included in my series. Of these, two, as above given, were reconstructed by the Born method. The other eight were reconstructed graphically, beginning with the uterine DEVELOPMENT OF THE ALBINO RAT - 2k end of the tubes. In six of these, the ova are quite closely grouped as given in the reconstructions shown in figures 9 and 10. In the remaining two they were more widely spaced, about as shown in figures 7 and 8. In the oviducts taken from rats Nos. 58, 61, 62, killed respectively 2 days, 17 hours, 2 days, 18 hours, and 2 days, 22 hours, after insemination, the ova are found in a portion of the tube which corresponds very closely to that shown in the reconstruction presented in figure 9. In rat No. 60, killed 1 day, 18 hours after insemination, the ova are more widely spaced and are situated in a segment of the oviduct approximately one loop nearer the fimbriated end than that given in figure 7, a model of the oviduct of rat No. 59, killed two days after insemination. In one of the segmented ova of rat No. 60, the two blastomeres resulting from the first segmentation division are distinctly sep- arated by a space equal to about one-half of the diameter of each of the cells. No oolemma is discernible. The two separated cells appear normal in size, shape, and structure, as do also their nuclei. They lie free in a slightly distended portion of the lumen, and appear not to have been separated as a consequence of ma- nipulation. The possibility of each developing separately is suggested, and may be offered as a possible explanation of the occurrence of very small embryos now and then found among others showing normal development. King states that “On dissecting pregnant females (rats) one frequently finds one or more embryos that are much smaller than the rest. Whilein some instances such small embryos appear normal and are presumably either runts or embryos that have resulted from superfecunda- tion, in the majority of cases they are pathological, probably because of faulty implantation of the ovum.’ My own material contains pathologic ova and embryos in different stages of development. This portion of the material will be considered in Part II, where the possibility of the occurrence of half em- bryos will be discussed. As may have been seen, the 2-cell stage of the albino rat covers a period of somewhat more than 24 hours, extending from about the middle of the second day until toward the end of the third 272 G. CARL HUBER day after the beginning of insemination. During this period the segmented ova migrate in the oviduct for a distance equaling nearly half its length. The trustworthiness of the material, it would seem to me, is shown by the fact that in the shorter time stages the segmented ova are situated nearer the fimbriated end, while in the longer time stages they approach the region of the insertion of the oviduct into the uterine horn. This is clearly shown in the reconstructions shown in figures 7 and 8. A 3-cell stage was observed only twice: in one of eight ova contained in the oviducts of rat No. 58 (2 days, 17 hours) and in one of eleven ova found in the oviducts of rat No. 62 (2 days, Fig. 11 Two views of each of three models of 4-cell stages of the albino rat. Rat No. 50, 3 days, lhour. X 200. A, B, and C, gives aside view, A’, B’, and C’ a vertical view, of each of the three models. 4 22 hours). All the other ova found in these two animals were in the 2-cell stage. In the two 3-cell stages noted, the undivided blastomeres of each ovum presented a nucleus in mitosis; in one, in the monaster phase, in one, in the diaster phase. The divi- sion of the first two blastomeres, resulting in the 4-cell stage, it would appear, occurs in the albino rat toward the end of the third day. The material gathered at the beginning of the fourth day after insemination presents throughout a 4-cell stage. In D of figure 1 is shown reproduced one of the sections of a series of six sections including one of the ova in the 3-cell stage. Only one of the two cells resulting from the division of one of the first two blastomeres is included in the section; the cell in mitosis represents the undivided blastomere. DEVELOPMENT OF THE ALBINO RAT 21D 4-cell stage. The material includes the oviducts of two rats, Nos. 50 and 63, killed 3 days and 1 hour after the beginning of insemination, with twelve ova in the 4-cell stage. In figure 11, there are shown two views of each of the models obtained by reconstruction after the Born method, at a magnification of 1000, of the three 4-cell stages found in the oviducts of rat No. 50. The drawing of the reconstructions do not present the conventional figures of the 4-cell stage of the mammalian egg. In none of the twelve ova of this stage was the plane of section such as to include all of the four cells in one section.» Nearly all Fig. 12 Cross-section of right oviduct of rat No. 50, 3 days, 1 hour. X 100. This section contains two cells of a 4-cell stage of the albino rat, slightly com- pressed between the folds of the tubal mucosa. lie in a portion of the tube which presents a relatively narrow lumen, and appear as if slightly compressed between the folds of the mucosa. I am not disposed to regard this as a resultant of fixation, due to contraction at the time of fixation. In figure 12 is reproduced a cross section of the right oviduct of rat No. 50, passing through a 4-cell stage. It is evident that in shape the two cells included in the section, conform in the main to the form of the lumen, the mucosa appearing as slightly retracted to one side of the egg mass. This conformity in shape of cell mass to the form of the lumen I find quite general in my material showing segmentation stages of the albino rat, to some extent 274 G. CARL HUBER even in the 2-cell stage, more clearly shown in the 4-cell and later segmentation stages, as will appear from further reconstructions presented. It would seem to me reasonable to assume that these cell masses are of such plasticity that they are molded by the tubal mucosa rather than they would compress the mucosa and maintain an inherent form. A number of segmented ova in presumably the 6- and 8-cell stages were removed from oviducts by injection and studied in warm normal salt solution, in a liy- ing state. In the warm normal salt solution the morula masses Fig. 13 Model of right oviduct of rat No. 50, 3 days, 1 hour. X 10. A short segment of the upper end of the uterine horn, lower part of the figure, is included. The fimbriated end and a part of the infundibulum removed in the drawing so as to expose the underlying loops; their relative position is indicated in dotted outline. The position of the four ova in the 4-cell stage, at the beginning of the last loop of the oviduct, is shown as if seen through a transparent wall. presented a nearly spherical form, conforming to the conventional illustrations of the same. In none of the sections of fixed material of my series was this the case. The form of the cell mass, assumed by the segmenting mammalian ovum in early stages of segmenta- tion, therefore, seems to me a question more for academic dis- cussion than one of fundamental importance. The right ovi- duct of rat No. 50 (3 days, 1 hour) was reconstructed after the Born method. This model is reproduced in figure 13, and includes the uppermost end of the uterine horn. The oviduct DEVELOPMENT OF THE ALBINO RAT 215 measures 2.8 cm. and contains four ova in the 4-cell stage, situ- ated at the beginning of the last loop leading to the uterine horn, 2.25 em. from the fimbriated end, thus in the fourth seg- ment of the oviduct as of Sobotta’s designation. In figure 14 is reproduced a detailed reconstruction of the segment of the oviduct containing the ova, with the convex portion of the wall of this loop, as shown in figure 13, removed. The section re- produced in figure 12, passes through the lower of the three upper ova, shown in reconstruction in figure 14. In the figure of the reconstruction as also in that of the section, is shown the groove in which these three ova he. The other oviducts con- Fig. 14 Model of the segment of the right oviduct, rat No. 50, 3 days, 1 hour, containing the four ova in the 4-cell stage, the general position of which is shown in figure 13. The convex portion of the wall of the loop containing the ova is removed, so as to expose the lumen. taining 4-cell stages were reconstructed graphically, beginning with the uterine end. The position of the ova in each is essen- tially as given in the model reproduced in figure 13. 8-cell. stage. In rat No. 57, killed 3 days, 17 hours after the beginning of insemination, there are found in the left oviduct, six segmented ova in the 8-cell stage and one segmented ovum in the 11-cell stage. The right ovary and oviduct was injured in the process of embedding and could not be used for sectioning. The ova are spaced in the loop of the oviduct which terminates in the uterine horn. Six of the segmented ova were recon- structed, the seventh was not detected at the time the recon- structions were made. The six models obtained are reproduced DATAG) G. CARL HUBER in figure 15, two views of each model being shown. Five of the models, A to EK, show 8-cell stages. In F, there is figured an 1l-cell stage, three of the cells having completed the next following division. As may be seen from the figures, the form of these morula masses is not spherical but in the main slightly oval, with further irregularities better shown in the models than in the illustrations, due to the fact that the egg masses con- form to the shape of the lumen of the oviduct in the region in which they are found. The mucosa lining the segment of this Fig. 15 Models, obtained by reconstruction after the Born method, of 8- cell and 11-cell stages of the albino rat. Rat No. 57, 3 days, 17 hours. X 200. Two views of each model is presented. A-A’, to E, E’ are of models of 8-cell stages; F and F’ of a model of a 11-cell stage. oviduct containing the ova presents four quite regular longitudi- nal folds. In figure 16, there is presented a model of a detailed reconstruction of the segment of the oviduct containing the ova, their relative position in the tube and their relation to the major folds is clearly shown. One of these folds it was necessary to in part remove so as to bring to view in the drawing certain of the ova. In figure 17, there is reproduced a portion of one of the sections of the series from which the medel shown in figure 16 was made. The fold of the mucosa occupying the center ee a ee eee DEVELOPMENT OF THE ALBINO RAT DEE Fig. 16 Model of the segment of the oviduct, rat No. 57, 3 days, 17 hours, containing the ova shown in fig. 15. X 50. A portion of the wall of the oviduct and a part of the major folds of the mucosa are removed in the drawing so as to expose the contained ova. The relative position of the seven ova found in the tube is shown, as also the extent and character of the folds of the mucosa. The exact form of each of the several ova could not be reproduced in the model at the magnification used; their position is given correctly. bes Fig. 17 Camera lucida drawing of a portion of a section of the left oviduct of rat No. 57, 3 days, 17 hours. X 100. This section is of the series of sections from which the models shown in figures 15 and 16 were made. Sections of four 8-cell stages, as seen in a single section, are included. The close proximity of three of these ova, their relation to the tubal wall and mucosal folds is to be noted. 278 G. CARL HUBER of the drawing, and greatly occluding the lumen, is the fold re- moved in the model. In this very fortunate secticn four of the morula masses are included; all are of the 8-cell stage and repre- sent in section the four ova which are placed closely together as seen in the model figured in figure 16. In figure 19, A, there is reproduced at higher magnification another of the sections of the series, including the right one of the three ova in close apposi- Fig. 18 Model of the left oviduct of rat No. 51, 4 days. X-10. A short segment of the upper end of the uterine horn was included in the reconstruction, lower end of the figure. The position of three of the morula masses, 12-cell to 16-cell stages, in the terminal part of the oviduct is to be noted, a further one is located in the upper part of the uterine horn. These are shown as if seen through a transparent wall. A fifth morula, situated in the uterine horn about 1.5 em. from the entrance of the oviduct, is not included in the figure. tion as seen in figure 17, showing six of the eight cells, each cut in the plane of its nucleus. In both of these figures (figs. 17 and 19) the morula masses, as seen in the sections drawn, present a quite regular oval outline. In succeeding sections, in which the mucosal fold and the wall of the oviduct approximate, the cross diameter of each of the four morula masses becomes greatly reduced, they appearing in the final sections of the series in which they are included as narrow, non-nucleated bands of protoplasm. DEVELOPMENT OF THE ALBINO RAT 279 This series, it seems to me, corroborates the statement previously made, that the detail of form of the living segmenting ova of certain mammals, while in transit through the oviduct, is in a great measure dependent on the configuration presented by the lumen of the oviduct in the particular region in which they are found. ; 12-cell to 16-cell stages. Rat No. 51, killed 4 days after the beginning of insemination, presents the end of the segmentation stages in the oviduct. In the genital tract of this animal there were found eight morula masses, five on the left side and three on c D Fig. 19 Sections of morula stages of the albino rat. % 200. A, 8-cell stage, rat No. 57, 3 days, 17 hours; six of the eight cells, each cut in the plane of its nucleus, are included in the section figured. B,C, and D, 12-cell to 16-cell stages, from right oviduct, rat No. 51, 4 days. the right side. Itis somewhat difficult to determine definitely the number of cells constituting each of the morula. The number appears to vary between 12 and 18, though nearly all of the moru- la masses-show certain nuclei in mitosis. The left oviduct with a short adjoining segment of the uterine horn was reconstructed. Slight tension was applied to the tissue prior to fixation, which accounts for the elongation of the proximal loop of the oviduct. The model is reproduced in figure 18. As is evident on study of this figure, three of the morula masses are situated in a por- tion of the oviduct just prior to its insertion in the uterine tube. These are closely grouped between folds of the mucosa. days | 7 | Early stage of blastodermic vesicle 56 | © days 5 Early stage of blastodermic vesicle Thus there are at hand 30 ova, showing late morula stages, the beginning of segmentation cavity formation and early stages of the blastodermic vesicle, falling in the latter half of the fifth day after the beginning of insemination. Jn all of the uteri from which this material was taken, the ova are spaced in the uterine horns about as in later stages of development; they lie free in the uterine lumen, are in the main ovoid in form, their long axis presenting no definite relation to the long axis of the uterine horn. In preparing this material for sectioning, 1t was the custom to cut an entire uterine horn into segments measur- ing about 1.0 em. to 1.5 em. in length. These segments were then embedded so as to admit cutting longitudinally and in a plane parallel to the plane of the mesometrium. Cut in this way, the majority of the ova were cut longitudinally or nearly so, others in an oblique plane, others again, crosswise. Since it DEVELOPMENT OF THE ALBINO RAT 287 is impossible to orient the ova prior to sectioning, the securing of desirable sections is a matter of chance. The difficulty is further enhanced by reason of the fact that owing to shrinkage as a result of the action of the fixing fluid, the ova in the vesicle stage are apt to be more or less folded, so that even though the plane of section may be that desired, the resultant sections lose in value by reason of this folding. It has been shown that in the albino rat, the ova pass from the oviduct to the uterine horn toward the end of the fourth day. During the first half of the fifth day, the migration of the ova from the oviduct to the uterine horn appears to be completed, so that by the second half of the fifth day the ova are spaced in the uterine horn about as after fixation to the uterine mucosa. As to the factor or factors which play a réle in the descent of the ova through the uterine horn and _ their fairly regular spacing, my own material gives no data; these changes occurring, apparently, during the first half of the fifth day, covering which my material is lacking. Widakowich, who has given especial study to these questions, presents the following considerations: In the downward migration of the ova in the uterine horn, it cannot be assumed that the ova are capable of active movement nor can their motion be ascribed to the action of gravity. While peristaltic action may play a part, it is diffi- cult to see how peristalsis could be so regulated as to space the ova fairly regularly within the uterine cavity. The presence of a ciliated epithelium in the human uterine cavity during the intermenstrual period suggested the presence of a ciliated epithe- lium in the uterine horn of the rat. After many preparations had been searched in vain for its presence, Widakowich found short cilia, not more than 2 » long in the epithelium lining the uterine cavity of a rat killed four days after copulation, and containing ova in the blastodermic vesicle stage. It would ap- pear, therefore, that the uterine epithelium of the rat presents a ciliary border for only a relatively short time, and that the transportation of the ova within the uterus is effected by the cilia. Mandl also found, his material however not including the rat, that cilia are present in many animals on the epithelium lining the uterus only at certain periods, and perhaps only relatively 288 G. CARL HUBER short periods. While the presence of cilia may explain the migration of the ova in the uterine tube, Widakowich can offer no conclusions as concerns the regulatory mechanism by means of which the ova are spaced at fairly regular intervals in the lumen of the uterus. In none of my sections of the uteri of albino rats, obtained during the fifth day after insemination, have I been able to note the presence of cilia on the uterine epithelium, even when sections were studied under the oil im- mersion. After reading the account of Widakowich, their presence was looked for in all pertinent stages, but without suecess. Especially in rat No. 50, in which the ova were pass- ing from the oviduct to the uterine horn was careful search made, but nothing like a distinct ciliary border, composed even of short cilia, was ascertained. In the left genital tract of this rat, as has been stated, three ova were found in the terminal part of the uterine end of the oviduct, one in the uterine lumen just distal to the mouth of the oviduct. and one a little over a centimeter from this opening. The latter was lodged in a shallow depression of the uterine mucosa, as is characteristic for stages lying free in the lumen. The question as to whether this ovum was permanently lodged is difficult to answer. If this is assumed, it is further necessary to assume that the other ova would need to pass it to reach the more distal parts of the uterine lumen. The literature contains no definite statements as concerns the reactions of the epithelium and mucosa of the uterus to the ova soon after their appearance in the uterine cavity. Widako- wich summarizes the views by stating that “It is generally stated, that so long as the ova lie free, the uterus shows no changes.’’ He himself notes that at this time the mucosa pre- sents evidence of marked new formation of capillaries. Burck- hard, who had at his disposal a large number of stages showing implantation of the ovum of the mouse, discusses at length the appearance presented by the uterus soon after the ova enter the same and the lodgment of the ova therein. This observer notes that in the non-gravid uterus of the mouse, the lumen lies more or less eccentric, and towards the mesometrial border. DEVELOPMENT OF THE ALBINO RAT 289 The lumen is not smooth, but presents numerous radially ar- ranged folds, certain of which are relatively deep. Essentially the same characteristics pertain to the mucosa of the uterus, soon after the beginning of gravidity. As the ova pass from the oviduct to the lumen of the uterus they become lodged in certain of the mucosal folds, and generally in certain of the deeper ones to be found along the anti-mesometrial border. I find Burckhard’s account of the form of the lumen of the uterine horn, of the structure of the mucosa in early stages of gravidity, and the lodgment of the ova, pertaining to the mouse, applies equally well to the albino rat. No reason can as yet be given as to why the ova are lodged in the mucosal folds in which they are found, and not in others. So far as may be ascertained from the sections, the particular mucosal folds in which the ova are found, do not differ in form and structure from neighbor- ing folds. It is possible that by reconstruction of the epithelial lining of the entire uterine horn in pertinent stages, certain characteristics of form and position might be revealed as pos- sessed by certain mucosal folds which make them especially favorable for the lodgment of the descending ova. Such re- constructions, however, have not been made. Burckhard states that in the mouse, about the middle of the fifth day, after the ova have been in the uterine cavity for a number of hours, there may be observed the first changes in the uterine wall. The changes consist primarily in a flattening of the uterine epithelium. In the immediate region where implantation is to occur, the lining epithelial cells present instead of a cylindric form, a cubic form. The area is sharply demarked from the surrounding epithelium, the transition of cubic to cylindric epithelium being marked by a sharp-lipped epithelial ledge. In my own material of the rat covering these stages, the uterine mucosa likewise presents shallow pits, in the immediate regions where the ova are lodged, lined by slightly flattened, cubic epithelium, very much as described by Burckhard for the mouse. Widakowich presents an excellent figure (fig. 2 of his contribution, rat four days after fertilization—‘Befruchtung’) showing clearly the relations of the ova to the uterine wall. In this rat, the 290 G. CARL HUBER uterine epithelium presented a ciliary border, present even in the shallow pit lodging the ovum sketched. He argues from this that the shallow depression and the flattening of the epithelium are not a result of pressure exerted by the vesicle, as thought by Sobotta and Melissinos, but must be due to an active change in the epithelium itself. The mucosa underlying the shallow pits presents at this stage no change of structure. I am thus in ac- cord with Widakowich when he states that he was not able to observe in the mucosa of the rat in the early stages of gravidity, the giant cells described by Disse as found in the uterine mucosa of Arvicola arvilis, in similar stages. The form presented by the ova of the albino rat, in the late morula stages and the early stages of blastodermic vesicle, is ovoid, as may be seen from the figures to be presented. Wida- kowich is inclined to believe that the form of the blastodermic vesicle of the rat is in a measure dependent on the general form of the space in which it is lodged. He figures two vesicles (figs. 1—2) one of which is nearly spherical, the other of distinctly oval form. Duval (figs. 73-83) presents vesicles having ovoid, trian- gular, and spherical forms. Christiani’s figures covering these stages, are too schematic to be of any value in drawing con- clusions. I fear Robinson’s account is based on imperfectly fixed material. He states that ‘toward the end of the fifth day, or the commencement of the sixth day, the longitudinal axis of the blastodermic vesicle is 125 » long. During the sixth day, that axis is diminished, first to 95 uw, and then to 64 u, after which it again increases, and at the commencement of the seventh day, it is 121 .”’ Neither Fraser nor Selenka describes nor fig- ures the stages here considered. In the mouse, according to the accounts of Melissinos, Burekhard, and Sobotta, the form of the blastodermic vesicle in early stages is spherical. The more specific consideration of my own material I shall introduce with a discussion of three stages taken from the uterus of rat No. 52, killed 4 days, 15 hours after the beginning of insemination. In A, of figure 20, there is reproduced the mid- dle one of seven sections of a late morula stage. This morula is of ovoid form, measuring 85 u in its long diameter, 54 u» in its DEVELOPMENT OF THE ALBINO RAT 291 broad diameter—that is, in plane of sections, and since it passes through seven sections of 10 thickness, measures approximately 70 w inits third dimension. It consists of 24 cells, as estimated by counting the nuclei contained in its several sections. The cells vary in size as well asin shape. The nuclei are for the main of spherical form, presenting one or several large nucleoli and fine chromatin granules. This morula is found within a fold of the mucosa, each side of which presents a slight depression, lined by slightly flattened epithelium. T his morula: mass lies Fig. 20 Sections of morula mass and early stages of blastodermic vesicle of the albino rat. > 200. A, B, C, rat No. &2, 4 days and 15 hours. D and E, rat No. 68. 4 days and 16 hours. A, late stage of morula; B, shows the very be- ginning of the formation of the segmentation cavity; C, D, E, early stages of blastodermic vesicle, in E, a distinct covering layer in the thicker portion or floor of the vesicle is evident. free in the lumen of the mucosal fold, and not in contact with the uterine epithelium. The outline and extent of the shallow depressions found in the opposing walls of the mucosal folds conform to shape and size of the morula mass contained, which appear as if slightly retracted as a result of fixation. In B, figure 20, is figured one of the sections of a series passing through a morula stage, comprising as estimated 30 cells and measuring 90 u, by 60 uw, by approximately 50 », in which the very beginning of the formation of the segmentation cavity is shown. Near one pole the outermost cells have separated slightly from the more deeply placed cells, so that an irregularly shaped 292 G. CARL HUBER cavity, eccentrically placed and passing clearly through two of a series of five sections of 10 » thickness, is evident. The small cleftlike cavity is bounded by the surrounding cells, the outline of which is distinct. So far as may be judged from the appear- ance noted as presented in the two sections in which this cavity is found, this arose as a single space and as a result of the separa- tion of the enclosing cells. In C, of figure 20, there is presented a slightly older stage showing the blastodermic vesicle formation and measuring 80 u, by 50 uw, by approximately 50 yu, comprising as is estimated, 34 to 36 cells. Unfortunately, the lower part of this vesicle is slightly folded as is shown in the lower left of the figure. The appearances presented in the sections are reproduced as faith- fully as could be. Owing to the folding, a portion of the thin wall is cut tangentially. The more darkly colored curved line represents in reality the outer boundary of this portion of the vesicle. The segmentation cavity in this vesicle is distinctly larger than that shown in B of this figure. In the section re- produced the segmentation cavity is bounded for the greater part by four somewhat flattened cells, the increase in the size of the cavity being accompanied, it would seem, by a flattening of the enclosing cells. In these three closely approximated stages, which, since they are taken from the same uterus are probably separated in time of development by only short intervals, the cells though varying in size and shape, show no essential or fundamental difference in structure, neither in cytoplasm nor nuclei; nor do they show any regularity in arrangement. Only few mitotic figures are to be observed; none in the morula mass shown in A, and but two in each of the other two stages, shown in B and C. Judging from these preparations, one would be led to con- clude that segmentation cavity formation in the albino rat is not associated nor accompanied by active cell proliferation. This point will be referred to again after the presentation of further material at hand. In slightly older stages of the blastodermic vesicle than here considered, the thicker portion of the vesicle is designated by Sobotta and others as its floor, which is directed DEVELOPMENT OF THE ALBINO RAT 293 toward the mesometrial border of the uterine horn, while its thinner portion is designated as its roof, directed toward the antimesometrial border. Therefore, in slightly older stages than thus far figured, the vesicle lies with its long axis approxi- mately at right angle to the long axis of the uterine horn. In further description of the blastodermic vesicle, I shall use the term ‘floor’ and ‘roof’ as here specified. In D and E of figure 20, there are reproduced typical sections of the two blastodermic vesicles taken from the uterus of rat No. 68, killed 4 days, 16 hours after insemination. Vesicle D measures 90 » by 30 u by approximately 60 u, and is of distinctly ovoid form and slightly compressed. This vesicle is found lying free in a long but narrow fold of the mucosa, both sides of which are slightly molded in conformity with the form of the vesicle. The long axis of the vesicle is still parallel to the long axis of the uterine horn. The roof of the vesicle appears as if slightly contracted, though when traced through the series of six sections it does not appear folded. The roof is composed of only a few cells, perhaps seven in all. The segmentation cavity presents a regu- lar outline. This vesicle supports the contention of Widakowich, that the form of the blastodermic vesicle of the rat is dependent in a measure on the form of the space in which it is found. Vesi- cle E, of figure 20, measuring 85 uw by 45 »w by approximately 40 uw, presents a roof that is slightly folded and shows an early stage in segmentation cavity formation. A figure of the vesicle is included since it represents more clearly than any other blas- todermic vesicle of the albino rat in my possession, a differentia- tion of a layer of surface cells in the mass constituting its floor. This is a question to be more fully considered in further discussion. In all the measurements of blastodermic vesicles thus far given, even in those given for the morula mass shown in A, figure 20, it is evident that one of the short diameters is appreciably shorter than the other. The vesicles are not only of ovoid form, but slightly flattened, so that even when not folded, the form of the vesicle as seen in section, even when cut parallel to the long axis of the respective vesicles, is dependent in a measure on the plane of the section, whether parallel to the longer or the shorter 294 G. CARL HUBER of the two cross diameters. This may be seen from the series of drawings made of a blastodermiec vesicle cut cross-wise, taken from the uterus of rat No. 68, from which were also taken the two vesicles shown in D and E of figure 20. This series of figures is shown in figure 21, in which are reproduced in serial order the seven successive cross sections into which the vesicle was cut. It measures 65 » by 38 uw by approximately 70 u, and is found at the bottom of a mucosal fold, found at the mesometrial border, and is resting with one side on the epithelial lining of a shallow pit, the other wall of this mucosal fold, also showing a shallow pit, is shghtly retracted. From a study of this series Fig. 21 A complete series of cross-sections of an early stage of blastodermic vesicle of the albino rat. > 200. Rat No. 68, 4 days and 16 hours. A to C, sections through roof of vesicle, showing segmentation cavity; D to G, sections through floor of vesicle. of sections, I feel certain that the plane of section is cross and not oblique to the long axis of the vesicle. The roof of this vesicle passes through three sections, A, B and C. The segmentation cavity has thus a depth of less than 30 u. The overlapping of the cells surrounding the segmentation cavity is to be noted, especially as seen in B of this figure. This arrangement of the cells may explain how the cavity may be enlarged without a material increase in the number of the enclosing cells—in part, by a flattening out of the cells, in part by a rearrangement of the relations of the cells. In the figures of the sections passing through the floor of this vesicle, D to G, attention is drawn to the size, form and relations of the cells and to the fact that there is no distinct covering layer. In this series of sections, there are DEVELOPMENT OF THE ALBINO RAT 295 shown in all 35 nuclei, two of which are in a late diaster phase. By excluding the nuclei that appear to be cut in two, appearing thus in two successive sections, I estimate that the blastodermic vesicle is made up of only about 30 to 32 cells. In figure 22, there are reproduced typical sections of four blastodermic vesicles taken from the uterus of rat No. 53, killed five days after the beginning of insemination. This uterus con- tained in all, eight vesicles, one of which was distinctly pathologic. The four vesicles selected for reproduction and discussion pre- sent each certain characteristics worthy of consideration. Vesi- cle A, which presents an early stage of segmentation cavity forma- Fig. 22 Sections of early stages of blastodermic vesicle of the albino rat, xX 200. Rat No. 53, 5 days. tion is of interest owing to the number of mitotic divisions it contains. As a rule, I have noticed only a few mitoses at this stage. In this vesicle, which measures 90 » by 55 » by approxi- mately 40 u, there are no less than five mitoses to be noted, three of which are in cells forming the roof of the small segmentation cavity, and are included in the section figured. This is the only vesicle in my possession in which an increase in the size of the segmentation cavity is accompanied by active mitoses in the cells bounding it. The vesicle lies free in the uterine lumen, one wall of which is only slightly pitted. In B of figure 22 is reproduced a section of a blastodermic vesicle, measuring 90 u by 55 uw by approximately 45 uw. It is evident that this vesicle JOURNAL OF MORPHOLOGY, VOL. 26, NO. 2 296 G. CARL HUBER passes through five sections of 10 » thickness, though one of the end sections, the fifth section of the series, seems to have fallen out during manipulations of staining and mounting, since the preceding, or fourth section does not quite complete the series. This vesicle lies free in the lumen of the uterus, and there is evi- dent only a shallow pit in the mucosa juxtaposed. In this vesicle the cells forming the roof of the segmentation cavity are relatively numerous, and are not markedly flattened, and in one an early mitotic phase is recognized. Here again cell prolifer- ation appears to have accompanied increase in size of segmenta- tion cavity. The vesicle shown in C of figure 22, measuring 130 u by 30 yu by approximately 40 u, lies free in a long, narrow fold of the uterine mucosa, in close proximity to a shallow mucosal pit, lined by cubic epithelium; the pit conforming in shape and extent to the form of the side of the vesicle presented to it. Therefore, it would seem that the form of the vesicle as seen in sections of the fixed material is essentially the same as that obtained in vivo. The two vesicles, typical sections of which are shown in B and C of this figure, are almost in identically the same phase of development, although their form as seen in sec- tions differs markedly. The plasticity of the living blastodermie vesicles is no doubt such that their form is in a great measure dependent on the shape of the mucosal fold in which they are lodged. In D of figure 22, there is reproduced a section of a blastodermie vesicle which points to a stage of development which is slightly more advanced than that shown in previous figures. The vesicle measures 100 » by 70 » by approximately 50 uw. The roof enclosing the segmentation cavity is slightly folded; a portion of its wall is thus cut tangentially, as shown in the lower left of the figure. The segmentation cavity is distinctly larger than that shown in the preceding figures, and is bounded by a relatively large number of cells, fourteen in that portion of the roof sketched in this figure, one of which is in a mitotic phase. The mass of cells constituting the floor appears as slightly compressed, in consequence of a slight intravesicular pressure which aided in the enlarging of the segmentation cavity. DEVELOPMENT OF THE ALBINO RAT 297 The cells forming the roof of the segmentation cavity do not appear so distinctly flattened as is the case in certain of the vesicles figured in figures 20 and 21. It would appear, there- fore, that at least two factors are operative in the increase of size of the segmentation cavity after its anlage—a flattening out and consequent increase of the exposed surfaces of the enclosing cells, and secondly, a cell proliferation; and it would appear that both of these factors may be operative from the time of the beginning of segmentation cavity formation. Early stages in the blastodermic vesicle formation in the albino rat have been previously described by Robinson, Christi- ani, Duval, and Widakowich; Selenka’s youngest stage is slightly older than any discussed by me. My own observations are wholly in accord with those of Widakowich in so far as his account covers early stages of blastodermic vesicle formation. He dis- cusses and figures, however, only two vesicles, obtained four days after fertilization—‘Befruchtung,’ in each of which the segmentation cavity presents a smooth and regular outline and is of appreciable size. The observations of the other observers who have considered these stages will be discussed in connection with a very brief presentation of much more comprehensive observations on the mouse in similar stages of development. Of these latter, those of Sobotta (’03) are based on abundant and apparently well fixed material. Sobotta begins his discussion with the consideration of three ova taken from the same mouse, the second half of the fourth day after fertilization, each of which shows beginning of segmentation cavity formation, one of which was cut in longitudinal axis and is figured in his figure 1. This ovum is interpreted as showing that the segmentation cavity arises not as a single space, but as a number of disconnected spaces, which later become confluent and form a single space. A similar observation was made by Van Beneden on the bat, a fact which Sobotta uses to support his contention that the mouse ova studied by him were of normal structure. Melis- sinos gives a number of figures showing early stages in the forma- tion of the segmentation cavity in the mouse. His figures 21 and 22 (66 hours) are not unlike my own figures shown in B of 298 G. CARL HUBER figure 20 and A of figure 22. According to this observer, the segmentation cavity arises as a single space, due to an accumula- tion of fluid secreted by the cells of the morula. This secretion is evidenced by globule-like droplets which are shown in his drawings as adhering to certain of the cells bounding the seg- mentation cavity. In my own preparations of the albino rat, I find no evidence which would lead to the supposition that the segmentation cavity does not begin as a single space nor do I find any evidence of secretory globules as described by Melis- sinos. Selenka has described quite fully two blastodermic vesicles of the mouse, lying free in the uterine lumen. His account of their structure, supported by two somewhat dia- gramimatic figures, is as follows: The wall of the vesicle is formed by a layer of covering cells—‘ Deckzellen’—constituting a cover- ing layer—‘‘Deckschicht or Rauber’s layer.’’ The space en- closed by this layer, for about one-third to one-half of its extent, contains the ‘formative cells,’ for the remainder it contains fluid. The covering cells and formative cells are said to be sep- arated by a sharp line. The formative cells are in all parts separable into two fundamental germ layers. An inner layer, bordering the cavity and constituting the entoderm, is said to be composed of cells possessing more deeply staining nuclei, irregular outline, with tongue-like processes which extend into the cavity, and a granular protoplasm; further, of cells which are more clear, more peripherally placed, and which constitute the ectoderm. Each of these fundamental germ layers con- sists of a single layer of irregularly formed cells. According to the observations of Selenka, therefore, the floor of the vesicle consists of three layers of cells; an outer covering layer—‘Deck- schicht’ or ‘Rauber’s layer’—an inner layer of entodermal cells, and an intermediate layer of ectodermal cells. Jenkinson’s account reads as follows: ‘““There are present (1) an outer layer, one cell deep, of trophoblast, which is continuous over (2) an inner mass which becomes differentiated into the embryonic epiblast and the hypoblast, and which is quite distinct from the overlying trophoblast, as my specimens invariably show.” In Jenkinson’s figures 1 and 2, giving early stages of the blasto- dermic vesicle, there is not shown a differentiation of the inner DEVELOPMENT OF THE ALBINO RAT 299 mass into ectodermal and entodermal cells; the outer layer, covering layer, Rauber’s cells, or trophoblast, is clearly differ- entiated from the inner mass by a distinct space. Duval has recognized in early stages of blastodermic vesicle formation of the mouse and rat, in the thicker part of the vesicle, entodermal and ectodermal cells. The former are of irregular form, pos- sess granular protoplasm and are said to possess the property of ameboid movement. The remaining cells are recognized as ectoderm; a distinct covering layer is not recognized. In Christiani’s figures (rat), which are, however, so diagrammatic as to be of little value and are evidently drawn from poorly fixed material, entodermal cells, ectodermal cells, and covering cells may be recognized as per legends. Melissinos (mouse), while not describing definitely a peripheral or covering layer, states that outer cells of the thicker pole, like the cells enclosing the segmentation cavity, stain less deeply than do the more centrally placed cells. In earlier stages of vesicle formation, neither in figures nor in text as given by this observer, do I find reference to a differentiation into ectodermal and entodermal cells. The observations of Selenka, Duval, Robinson, Jenkin- son, and others, bearing on the structure of the blastodermic vesicle of the mouse and the rat in early stages of development have been so thoroughly and critically reviewed by Sobotta that an extended discussion has here been deemed unnecessary. It may here suffice to say that while Sobotta’s observations were made and his discussions based on ova obtained from the mouse, my own observations made on the albino rat are in agreement with his and support the contention that in early stages of blas- todermic vesicle formation a differentiation of the thicker part or the floor of the vesicle into a covering, Rauber’s cell, or tropho- blast layer, and a further differentiation into ectodermal and entodermal cells, is not to be made: we differ in our accounts of the beginning of the formation of the segmentation cavity. An outer or covering layer is suggested in certain of my own prep- arations, most clearly in that sketched in E of figure 20. How- ever, a uniform difference in structure and reaction to staining reagents of the outer layer of cells is not present in my own preparations. None of my own preparations gives evidence of 300 G. CARL HUBER such an early differentiation of ectoderm and entoderm as given by Selenka, Duval, Christiani, and others. Cells of irregular outline with tongue-like projections, such as figured by Selenka and Duval I have not observed. The cells constituting the floor or the thick part of the vesicle all present essentially the same structure, while the segmentation cavity, as soon as it presents appreciable size, shows a smooth and regular outline. In figures 1 and 2, of Widakowich’s communication, excellent figures of early stages of blastodermic vesicles of the albino rat, there is presented no evidence of a trophoblast layer nor a dif- ferentiation of ectodermal and entodermal cells. My own figures, 20 to 22, were drawn with the aid of camera lucida at a magnification of 1000 diameters and with the use of an intense Welsbach light. They are reduced five times in reproduction. With the exceptions of cell outlines, which as sketched do not in the preparations fall in the same optical plane, and aresketched more sharply than is perhaps warranted, the figures portray quite accurately the structural appearances presented, so far as may be with the use of a single color. BLASTODERMIC VESICLE, BLASTOCYST, OR GERMINAL VESICLE The material on hand is listed in table 6. TABLE 6 RECORD NUMBER AGE NUMBER OF VESICLES 75 5 days, 15 hours 6 91 5 days, 16 hours 2 88 5 days, 21 hours Sant 89 5 days, 21 hours 6 73 6 days 10 74 | 6 days 5 99 6 days 6 106 6 days 10 104 6 days 6 Total 58 During the sixth day, the blastodermic vesicle of the albino rat increases In size relatively rapidly. The greater portion of its wall is, at this stage, composed of a single layer of flattened cells. The vesicles are not as yet attached to the uterine wall, though the uter- DEVELOPMENT OF THE ALBINO RAT 301 ine mucosa shows a distinct reaction to their presence. Localized thickenings of the uterine mucosa, sufficient to cause localized swellings of the uterine tube, indicating the position of the ova, are evident. I have experienced more difficulty in successfully fixing the vesicles during this stage than any of the earlier or later stages studied. Although my material contains 58 vesicles of the stage under consideration, none of them may be regarded as being well fixed, and the majority of them are so folded as a result of contraction during fixation that they are of little value as objects for especial study. That the vesicles are still un- attached to the uterine wall is readily determined by the fact that the shrivelled vesicles are found lying free in the depres- sions of the uterine mucosa, lined by a low cubic epithelium, intact throughout, and retaining its normal relation to the mucosa. The molding in these mucosal depressions no doubt gives the size of the respective vesicles as in vivo. It is not my purpose at this time to consider more than super- ficially the changes affecting the uterine mucosa during ovum implantation in the albino rat. It is hoped that this may be the subject of a future communication. It is the purpose in the present communication to confine consideration to the develop- ment of the ovum itself. Many of the observations recorded by Burckhard on the implantation of the ovum of the mouse and the formation of the decidua, I find equally adapted to similar phenomena in the albino rat. Differences are to be observed in certain details which it is not the purpose to enter into here. Grosser gives a number of excellent figures (67 to 70, and 112 to 116) showing implantation and decidua formation in the albino rat; to these the interested reader is referred for the present. The thickening of the mucosa affects primarily its antimesometrial portion. During this process of thickening, the mucosal fold in which the ovum primarily finds lodgment, becomes deepened and converted into a funnel shaped crypt communicating with the uterine lumen, and surrounded by the ‘Hibuckel,’ or oval fold. Burekhard’s schematic figures (text figures 2 to 4) may be consulted to make the phenomenon intelligible. 302 G. CARL HUBER In figure 23, there are reproduced representative sections of five blastodermic vesicles falling to the end of the sixth day after insemination. None of these five vesicles can be regarded as well fixed. All show a certain amount of distortion, much more evident were the entire series of each of the respective vesicles shown. The form of the blastodermic vesicle of the albino rat at this stage of development, as indicated by the molding of the uterine mucosa, is ellipsoid. Their size as in vivo, when distended and of regular outline, again as indicated by the molding of the uterine mucosa, is slightly larger than would be Fig. 23 Sections of blastodermic vesicles or blastocysts of the albino rat. X 200. A and C, rat No. 99, 6 days; B, D, E, rat No. 100, 6 days. y.ent., yolk entoderm; p.ent., parietal layer of entoderm; p.ect., parietal or transitory ectoderm. supposed from the drawings presented. By reason of this dis- tortion, exact measurements of size cannot be given. In A of figure 23, there is reproduced that portion of one of the sections of a blastocyst (rat No. 99, 6 days) which passes through its floor; the thin roof of this vesicle was so folded that its inelu- sion in the drawing was deemed undesirable. However, its floor or the germinal disc, seems to have retained its normal form and structure, presenting when traced through the series a regular concavo-convex, discoidal form. It consists in the main of three layers of cells of polyhedral type; toward the border of the disc, of two layers of somewhat flattened cells, the peripheral layer being continuous with the single layer of cells DEVELOPMENT OF THE ALBINO RAT 303 forming the roof of the vesicle, not shown in the figure, and known as the parietal or transitory ectoderm. In the floor or germ disc, there is evident a single layer of cells bordering the segmentation cavity or blastocele and possessing a more gran- ular protoplasm, which stains a little more intensely in Congo red. Their differentiation and characteristic reaction to stain- ing agents is at this stage of development not quite so distinct as in slightly older stages. This layer of cells, similar to that described by Sobotta for the blastodermic vesicle of the mouse in essentially the same stage of development, he has termed the yolk entoderm, ‘Dotter entoderm,’ a designation which is here followed. In the more superficial layer or layers of cells no characteristic differentiation is observed. In no portion of the floor of this vesicle was a distinct covering or trophoblast layer recognized. In the vesicle, a section of which is reproduced in B of this figure (rat No. 100, 6 days), the floor or germ disc presents essentially the same structure as that shown in A. The vesicle shown under B, was also folded, especially its roof, which was drawn to one side and was thus not cut through its entire length in the section figured. Furthermore, the section chosen for drawing does not pass quite through the center of the germ disc, but a little nearer to one of its edges, which probably accounts for the fact that there is recognized for the greater part only a single layer of cells, superimposed over the yolk endoderm, which layer is continuous with the parietal or transitory ecto- derm forming the roof of the vesicle. The cells forming the yolk entoderm constitute a single layer and are quite distinctly differentiated; one of the cells shows a mitotic phase. The roof of the vesicle formed by the parietal or transitory ectoderm, is composed of a single layer of flattened cells with flattened nuclei, the form and structure of which is more correctly shown in the right half of the roof wall, which in the section is cut transversely, while the left half, owing to the folding, is shown as cut obliquely. In C of figure 23 (rat No. 99, 6 days), there is shown a greatly compressed blastodermic vesicle, taken from a series of cross sections of the uterine horn. In this figure there is reproduced the fifth of a series of 10 sections of 10 » thickness; therefore, 304 G. CARL HUBER the third dimension of the vesicle is approximately 100 u. It is evident that had this vesicle been cut in a favorable plane at right angles to the present series, or parallel to the mesometrial plane, its form would have approached that of a circle. IJ have in my possession one vesicle of this stage of devlopment, similarly compressed, cut parallel to the plane of compression, in which almost the entire roof falls within a single section of 10 » thick- ness. The structure of the vesicle shown in C is very similar to that shown in A and B of this figure. The normal form of this vesicle is quite readily reconstructed from a study of the series of sections into which it has been cut. The cells of the yolk entoderm are evident. The parietal or transitory ectoderm constituting the roof consists of a single layer of much flattened cells, with relatively few nuclei, having, as seen in cross section, a long ovoid form, which, when seen in surface view present a regular, nearly circular outline (see lowermost nucleus in the figure). In similarly compressed vesicles cut parallel to the plane of compression, the germ disc may appear as consisting of three to four layers of cells. In an imaginary section passing in a plane at right angles to that figured in C, and having perhaps a slight obliquity, the germ disc would appear as if much thicker than that shown in A and B of this figure. Such sections may readily lead to false conclusions. It seems evident from a study of the material at my disposal that during the sixth day after the beginning of insemination in the albino rat, the blastodermic vesicle or blastocyst, which has its anlage in the latter part of the fifth day, enlarges relatively rapidly; this largely owing to a distension of the segmentation cavity or blastocele. This enlargement is accompanied by a flattening and extension of the enclosing roof cells and by a re- arrangement of the cells of the floor, which is reduced in thick- ness to a discoidal area, the germinal dise or germ area, forming about one-fifth to one-sixth of the wall of the vesicle and con- sisting of two or three layers of cells. During the rearrangement of the cells which constitute the floor of the vesicle, those adjacent to the segmentation cavity or blastocele differentiate to form the anlage of the yolk entoderm. The remaining cells of the ger- DEVELOPMENT OF THE ALBINO RAT 305 minal disc, having all essentially the same structure, are of irregular polyhedral form and are mutually compressed. To designate them as a distinct germ layer at this stage seems inappropriate. A differentiation into a layer of covering cells and a layer of formative ectoderm (Selenka) is not to be made. Active cell proliferation as evidenced by mitotic figures does not appear to accompany this enlargement of the vesicle. This phenomenon seems rather to be accomplished by a rearrange- ment of the cells constituting its floor, however, primarily by an extension and consequent flattening of the cells forming the roof of the vesicle. A similar stage is shown for the mouse by Sobotta (’03) in his figures 3, 4, and perhaps 5, of mouse vesi- cles from the fifth day after fertilization—‘Befruchtung’. So- botta had at his disposal much more perfectly fixed vesicles than my material contains. The structure of these vesicles as given by this observer, both as depicted in figures and text, is very similar to the presentation given by me. He also recognizes in this stage the anlage of the yolk entoderm. Figure 30, ac- companying the account of Melissinos (mouse, 84 hours) presents a similar stage, although he figures fairly distinctly a layer of covering cells, which if I read him correctly, however, is of only transitory existence. None of the figures given by Robinson and Jenkinson is comparable with figures A, B, C, of figure 23 of this account. In D, of figure 23 (rat No. 100, 6 days) there is reproduced a section of a blastodermic vesicle which on superficial study presents a somewhat later stage of development than those shown in A to C, of this figure. It is, however, only very slightly older than the three vesicles discussed. Vesicle D, cut in good longitudinal direction, is in reality much more folded than ap- pears from the section figured. Its floor or germ disc is com- pressed in a plane parallel to that of the plane of section, so that the germinal disc is cut obliquely and not transversely, and thus appears thicker in the section than it in reality is. 200. A, rat No. 81, 7 days, 22 hours; B, rat No. 96, 8 days; C, rat No. 94, 8 days, after insemination; ect.pl., ectoplacental cone or Triger; p.ect., parietal or transitory ectoderm; ez.ect., extraembryonic ectoderm; ect.ves., ecto- dermal vesicle, with wall composed of primary embryonic ectoderm, at + junc- tion with the extraembryonic ectoderm; a.met.pr., antimesometrial portion of proamniotic cavity; met.pr., mesometrial portion of proamniotic cavity; pr.c., proamniotic cavity; v.ent., visceral entoderm; pr.emb.ent., primary embryonic entoderm. 325 326 G. CARL HUBER portion of the proamniotic cavity, developing in the extra- embryonic ectoderm, is of greater dimension. Two relatively large spaces, bordered by a single layer of cells of the extra- embryonic ectoderm, are to be observed. At the junction of the extraembryonic ectoderm and the ectodermal vesicle of primary embryonic ectoderm a further space of triangular out- line may be seen. The primary embryonic ectoderm is ar- ranged in the form of an oval-shaped vesicle, forming the anti- mesometrial end of the egg-cylinder. Its wall is relatively thin at the region of its apposition to the extraembryonic ectoderm, just below the triangular space above mentioned. This ecto- dermal vesicle is peculiar in that its cavity contains the re- mains of four cells. A study of the series of sections shows that these cells do not represent the crest of a fold of the wall of this vesicle, since they are not nearly so distinct in preceding and succeeding sections. It may only be conjectured that during the rearrangement of the cells of the ectodermal node, resulting in the formation of the ectodermal vesicle, certain of the cells became separated from the wall and remained free in the cavity. The primary embryonic ectoderm, forming the wall of the ectodermal vesicle is readily differentiated from the extraembryonic ectoderm, both by the fairly sharp definition of the ectodermal vesicle and by reason of the fact that its cells stain somewhat more deeply than do the cells of the extraem- bryonic ectoderm, as also the cells of the visceral entoderm. In the egg-cylinder shown under B of figure 27 (rat No. 96, 8 days) the antimesometrial portion of the proamniotic cavity, developing in the ectodermal node, and the mesometrial portion of the proamniotic cavity, developing as several discrete spaces in the extraembryonic ectoderm, have in part joined to form a single proamniotic cavity. The mesometrial portion of this cavity is still bridged by a septum of extraembryonic ectodermal cells, closing off a relatively large space found in its mesometrial portion. With the junction of the antimesometrial and the mesometrial portions of the proamniotic cavity, the primary embryonic ectoderm and the extraembryonic ectoderm become a continuous layer, the line of union of the two portions, however, remains evident and is readily recognized in all the egg-cylinders DEVELOPMENT OF THE ALBINO RAT Bak of this and older stages, a question which will receive further consideration in following pages. In C of figure 27 (rat No. 94,-8 days) the proamniotic cavity forms a continuous, single space. The figure presented is drawn from two sections; its greater portion, to the base of the ecto- placental cone from one section, the ectoplacental cone from another section. The junction of the membranous wall of the vesicle to the base of the ectoplacental cone, in the two sections used for the figure, was superimposed under camera lucida in joining the portions drawn from the two sections. It is be- lieved that the drawing as presented gives correctly dimen- sions and relations of the different parts of this vesicle. The wall of the antimesometrial portion of the single proamniotic cavity is formed by the primary embryonic ectoderm, the cells of which are for the main of irregular columnar shape, with alternately placed nuclei. These cells are in active proliferation, as is evidenced by numerous mitoses. The wall of the meso- metrial end of the proamniotic cavity is formed of a single layer of cells of the extraembryonic ectoderm; these cells are of quite regular shape with nuclei placed in about the same plane. They stain less deeply than do the cells of the primary embryonic ectoderm. In this egg-cylinder (C, fig. 27) the proamniotic cavity does not extend so near the base of the ectoplacental cone as in a number of other preparations in my possession, showing about the same stage of development; in certain of these, the proamniotic cavity extends to near the mesometrial end of the egg-cylinder. A more definite characterization of the different parts of the egg vesicle of the albino rat at the stage of development shown in C, figure 27, end of the 8th day, seems desirable, and in doing so I shall use the terminology used by Sobotta and Widakowich. The vesicle under consideration has reached a length of 0.65 mm., and a width of 0.12 mm. Somewhat more than one- fourth of its length consists of ectoplacental cone or Trager. The cavity enclosed is derived from the cavity of the blasto- dermic vesicle with germ disc, the blastocele, and is termed by Sobotta and Widakowich the ‘Dottersackhéhle’ or yolk-sac cavity. This cavity is bounded by a thin structureless mem- JOURNAL OF MORPHOLOGY, VOL. 26, NO. 2 328 G. CARL HUBER brane derived from the parietal or transitory ectoderm and the scattered cells forming the parietal layer of entoderm. This membrane is continuous with the base of the ectoplacental cone and presents scattered flattened cells on its inner surface. I have designated this thin membrane with cells on the inner surface as the parietal or transitory ectoderm (Kolster’s feinfaserige Haut). The egg-cylinder which extends to the antimesometrial end of the yolk-sae cavity, encloses the proamniotic cavity, the antimesometrial portion of which is walled by primary embryonic ectoderm, its mesometrial portion by extraembryonic ecto- derm, the two forming a continuous layer, with line of union of the two types of ectoderm evident. The uncleaved extra- embryonic ectoderm is continuous with the base of the ecto- placental cone. The egg-cylinder is surrounded by a single layer of cells of the visceral entoderm, differentiated so as to consist of a portion which surrounds the antimesometrial end of the egg-cylinder in relation with the primary embryonic ec- toderm; the cells of this portion being of a rather thick pavement type, constituting the primary embryonic entoderm, and fur- ther a portion which covers the sides of the egg-cylinder, with cells of a columnar type, showing special cytomorphosis. The egg-vesicles and egg-cylinders of the stage of development under consideration and for somewhat older stages show no bilateral symmetry so far as can be discerned by study under the micro- scope. In longitudinal sections of egg-cylinders, cut respectively in two different planes, at right angles to each other, no differ- ence in form, relation and structure of different parts can be observed. Selenka, Kupffer, Duval, and Sobotta have pre- viously called attention to this fact and shown that longitudinal sections of egg-cylinders may be obtained no matter whether the sections are cut parallel to the plane of the mesometrium, thus parallel to the long axis of the uterus, or at right angles to this plane. The want of bilateral symmetry is also evident in cross sections of the egg-cylinder, as may be seen from the series of sections presented in figure 28 (rat No. 27, 7 days, 17 hours). The cross-cut egg-cylinder, from several sections of which these figures were drawn, represents a stage of develop- DEVELOPMENT OF THE ALBINO RAT 329 ment very similar to that of the egg-cylinders shown in longitudi- nal section in figure 26. Widakowich, after discussing very briefly the mode of develop- ment of the egg-cylinder, discusses and figures an egg-cylinder of the albino rat, obtained 6? days after the last coitus. His figure 3 corresponds in stage of development very closely to that shown by me in A of figure 27. In his figures, there is pre- sented an egg-cylinder showing the anlage of the mesometrial Fig. 28 A series of cross sections at different levels of an egg-cylinder of the albino rat after the anlage of the antimesometrial portion of the proamniotic cavity. XX 200. Rat No. 27, 7 days, 17 hours, after insemination. The sec- tions selected for the several levels drawn, A to D, are as follows: A, middle of ectoplacental cone; B and C, through extraembryonic ectodermal portion of egg-cylinder, just below junction with ectoplacental cone (B), and just above ectodermal vesicle (C); D, through middle of ectodermal vesicle. Compare with B, figure 26, a longitudinal section of an egg-cylinder of the same stage of development; p.ect., parietal or transitory ectoderm; ez.ect., extraembryonic ectoderm; pr.emb.ect., primary embryonic ectoderm of the ectodermal vesicle; v.ent., visceral entoderm; pr.emb.ent., primary embryonic entoderm; a.met.pr., antimesometrial portion of proamniotic cavity. portion of the proamniotic cavity. Emphasis is given to the fact that in the antimesometrial portion of the egg-cylinder, there may be recognized the primary embryonic ectoderm. His own words with reference to this point read as follows: Der Schnitt zeigt nun sehr deutlich, dass sich die Zellen, die die antimesometrale Hohle so begrenzen, dass die alte Kugel-oder Eiform dieses Teiles noch zu erkennen ist—das primire embryonale Ectoderm —intensiver farben wie die Zellen des mesometralen Abschnittes oder die des Ectoplacentarconus—das extraembryonale Ectoderm. Die Kerne zeigen keinerlei Unterschied in der Farbung, wohl aber das Plasma, dass im antimesometralen Teile von dichterer Structur zu sein scheint. This description corresponds very closely to that given by me for a similar stage. The differentiation of these two kinds of ectoderm was also recognized by Robinson, who states: 330 G. CARL HUBER The epiblastic cylinder is closed at its distal end, the trophoblastic at its proximal, and the open ends of the two cylinders are in close apposi- tion, but not indistinguishably fused, for the character of each por- tion of the ectoderm, after treatment with carmine, is still quite dis- tinctive; the protoplasm of the trophoblast being tinged much more faintly than that of the epiblast. Selenka, on the other hand, who has recognized in his ‘Ekto- dermblase’ with ‘Markamnionhéhle’ a distinctive structure, believes this to blend completely with the Trager. Since his account with reference to this point has influenced later workers, I may be permitted to quote him in the original. Referring to the ‘Ektodermblase’ with ‘Markamnionhohle,’ he states: Dieser Ektodermkeim, welcher von dem vorriickenden Tragerzap- fen anfinglich sehr wohl abgegrenzt ist, indem beiderlei Gebilde sich in Folge der convexen Kriimmung ihrer einander zugekehrten Flachen sozusagen nur in einem Punkte beriihren, fliesst endlich mit dem Traiger vollstandig zusammen, und zwar bei der Waldmaus_ bevor, bei der Ratte und Hausmaus aber nachdem die Markamnionhéhle enstanden war. That the proamniotic cavity of the egg-cylinder of the albino rat has its anlage in two distinct cavities, the one developing in the ectodermal node in the antimesometrial portion of the egg- cylinder, which is the first to develop; the other in the meso- metrial portion in the extraembryonic ectoderm, was recognized by Selenka (fig. 30, plate 14, E, Markamnionhohle, E’, falsche Amnionhohle), Duval (fig. 100,) Robinson, and Widakowich (fig. 3). Corresponding stages of egg-cylinder development as presented by me in figures 26 and 27, for the albino rat, are shown by Sobotta (’02), for the mouse in his figures 12 to 14 and text figures a to f. On comparison of my figures with Sobotta’s, it becomes evident that the egg-cylinder of the rat is much longer and more slender than that of the mouse. According to the account of Sobotta, the egg-cylinder of the mouse, soon after its anlage, shows by reason of a distinct transverse furrow a division into two parts, an antimesometrial portion of globular form, surrounded by a visceral layer of entoderm, corresponding to what I have designated as the ectodermal node; and a meso- metrial portion which early shows the anlage of a proamniotic DEVELOPMENT OF THE ALBINO RAT aol cavity. A lumen is obtained in the antimesometrial portion later than in the mesometrial portion. As development proceeds, this sharp demarkation of antimesometrial and mesometrial portion is gradually lost. This, as stated in his own words, reads: Sehen wir von dem die (der Keimhohle zugekehrte) Oberfliche des Cylinders iiberziehenden Dotterentoderm zuniichst ab, so sieht man, dass die Furche, welche die oben erwihnten mesometralen und antimesometralen Abschnitte in Stadium der Fig. 11 u. 12 trennte, jetzt wieder wenig deutlich ist. Es bahnt sich eine Verschmelzung beider Abschnitte wiederum an, was man am leichtesten daraus er- sieht, dass bald (Fig. 14) beide Abschnitte ein gemeinsames Lumen erhalten. With the formation of a continuous proamniotic cavity, this is bordered by a single layer of ‘ectodermal cells,’ with al- ternately placed nuclei. The cells are described as being the same throughout; neither in text nor figure does Sobotta differ- entiate between ectodermal cells derived from the antimesome- trial portion of the egg-cylinder and those derived from the mesometrial portion. Melissinos also recognizes antimesome- trial and mesometrial portions in the development of the egg- cylinder of the mouse, in his figure 34. According to this ob- server, the antimesometrial portion of the proamniotic cavity is the first to appear; later it appears in the mesometrial por- tion, the two cavities joining as development proceeds. The parts of the ectoderm derived from these two portions may be recognized, however, after a single proamniotic cavity has developed. This Melissinos states in the following words: ‘““Trotz aller Vereinigung der beiden Hohlungen bleibt die Unterscheidung . des normals abgesonderten antimesomtralen Abschnittes von dem mesometralen immer leicht zu machen, sel es durch eine klare Grenzlinie oder durch eine an der Peripherie des visceralen Dotterblattes befindliche Furche.” The account of Melissinos is More in agreement with the presentations as observed in the albino rat than is that of Sobotta. Selenka, Sobotta, and Melissinos recognize three different regions of constriction to which significance is given, in the egg-cylinder of the mouse. As stated by Sobotta, the first con- . Say G. CARL HUBER striction is in the region of the original furrow which demarks the antimesometrial and the mesometrial portions of the egg- cylinder, the region of the primary amniotic fold; the second where the mesometrial cavity ends; and the third where the original blastodermic cavity reaches its mesometrial end. The three folds recognized by Melissinos, are characterized by the specificity of the ectoderm. Since his statement concerning this point is somewhat involved, I find it necessary to use his own words; they read as follows, referring to these folds he states: Der eine derselben a liegt antimesometral und ist der bekannte erste kugelf6érmige Buckel (EKktoderm) mit den langlichen, cylinder- pyramidalen oder polygonal-pyramidalen Zellen; der zweite 6b legt in der Mitte und besteht aus kubisch-polygonalen Zellen, und der dritte Buckel c, aus polygonalen Zellen bestehend, legt mesometral und ist von dem mittleren durch Einschniirung, von der Basis des Ectoplacentarconus aber durch die bekannte Urfurche des Eicylinders getrennt, in der sich das viscerale Dotterblatt zum parietalen Dotter- blatt umbiegt. So far as I am able to determine, the account of Melissinos agrees with that given by Sobotta, as concerns the folds of the egg-cylinder of the mouse. Selenka’s account need not receive special consideration. In well-fixed egg-cylinders of the albino rat no such folds are recognized. At the line of junction of the primary embryonic ectoderm and the extraembryonic ectoderm, a slight infolding of the layers, variable in degree, is recognized. Other foldings of the wall of the egg-cylinder I have regarded as accidental and not of special significance. Therefore, I am wholly in accord with Widakowich, who has also discussed this question with reference to the albino rat and has described the low fold in the region of the junction of the primary embryonic ectoderm and extraembryonic ectoderm. Referring to that fold, he states: “Dass war die einzige konstante, bald stirker, bald schwiicher ausgeprigte Einschniirung der Proamnionhohle.”’ Sobotta deserves credit for having described fully the differ- entiation and cytomorphosis of the cells of the visceral entoderm of the egg-cylinder, and since his observations on this point apply in the main to the albino rat, they may at this time be given DEVELOPMENT OF THE ALBINO RAT gaa consideration. During the early stages of egg-cylinder differ- entiation and anlage of the proamniotic cavity, the layer of viscer- al entoderm differentiates into a portion which is in relation with the primary embryonic ectoderm of the antimesometrial portion of the egg-cylinder, in which region the cells of the en- toderm are first of short cubic shape, later of the pavement type; this portion may be regarded as forming the primary embryonic entoderm, since it forms the greater part of the entoderm of the embryo. The greater part of the visceral entoderm, that which surrounds the sides of the mesometrial portions of the egg-cylin- der, consisting of extraembryonic ectoderm, differentiates into cells of the columnar type. In this latter portion, with the formation of a continuous proamniotic cavity, the entodermal cells undergo characteristic cytomorphosis. In them, as stated by Sobotta, there may be recognized three main zones: (1) a basal zone with denser protoplasm containing the nucleus; (2) a middle zone with markedly vacuolated protoplasm; (3) an outer zone in which hemoglobin granules are recognized, the latter zone staining deeply in eosin. These three zones in the cells of the visceral entoderm in the region of the extraembryonic ectoderm of the egg-cylinder may be recognized in figures 26 and 27, not so clearly as in Sobotta’s colored figures, particularly his figure 17 (03) and figure 8 (11). However, I am able to follow closely his description in my own preparations of a somewhat older stage than thus far figured. It is Sobotta’s contention that in the extravasated blood surrounding the egg vesicle, in close apposition to its thin outer wall, there may be observed many red. blood cells which, though presenting normal form, show a distinctly granular content. These granules stain deeply in eosin and are in shape, size, and reaction to stain very similar to granules found in the peripheral part of the cells of the visceral entoderm. On the outer surface of the thin wall of the vesicle; on its inner surface; in the cells lining this; in the yolk sac cavity; and on the outer surface of the cells of the visceral entoderm, similar granules are found. These appearances are interpreted as showing an absorption of maternal hemoglobin by the ento- dermal cells of the mesometrial portion of the egg-cylinder. aoe G. CARL HUBER Sobotta’s statement concerning this point, which, owing to its importance, I quote in full, reads as follows: Man wird diese mikroskopisch erkennbaren Verhiltnisse nicht an- ders deuten kénnen als in folgender Weise: Die Himoglobinschollen, die durch die éussere Wand des Dottersackes in die Dottersackhéhle gelangt sind, werden von der Oberflaiche des zylindrischen, die ganze Seitenfliche des Eizylinders itiberziehenden visceralen Dottersack- epithels aus resorbiert und zwar geschieht das in der Weise, dass die Himoglobinschollen ziinachst als solehe in der Zelle selbst eintreten, dann aber im vacuolisierten Teil der Zelle gleichsam verdaut werden, wobei die einzelnen kleinen Schollen vorher zu grésseren Tropfen zusam- men-fliessen scheinen. My own observations on the albino rat as concerns this phe- nomenon, more particularly as concerns the structure of the cells of the visceral entoderm in the region of the extraembryonic ectoderm, corroborate Sobotta in many particulars. This question will beagain and more fully considered in a contemplated later publication dealing with the implantation and decidua formation in the albino rat. It could not be considered now without a discussion of the changes involved in the development of the decidua, a question which I am not prepared to consider fully now. It may be stated, however, that judging from my own preparations and the figures of Grosser, the extravasation of blood into the egg chamber is not nearly so extensive in the albino rat as is shown in the figures of Sobotta for the mouse. The thin membrane which surrounds the yolk-sac cavity, which I have designated as the parietal or transitory ectoderm, is derived in development from the parietal or transitory ecto- derm, and the relatively few parietal entodermal cells, as de- scribed and figured for younger stages. At the stage of egg-cylin- der development under consideration—with continuous pro- amniotic cavity—this structure appears as a thin, practically homogeneous membrane with scattered, flattened nucleated cells on its inner surface. Sobctta regards these cells as derived from the parietal entoderm, the cells of the parietal ectoderm having disappeared. As concerns this, I am unable to speak with certainty, since the Congo red solution used as a double stain is not particularly favorable in differentially coloring these DEVELOPMENT OF THE ALBINO RAT 335 cells. However, I am disposed to regard these flattened cells as derived from the parietal ectoderm. The parietal entodermal cells are never numerous in the rat, and mitotic figures are sel- dom observed in them. With the extension of the vesicle with the enlargement of the blastocele, the cells of the parietal or transitory ectoderm become attenuated until they appear for the greater part as a thin cuticular membrane, and I am dis- posed to regard the flattened nucleated masses of protoplasm lining the inner surface of this membrane as derived from the cells of the parietal ectoderm. Much attention has been given to certain large cells which are found in close relation with the outer surface of this thin mem- brane. These cells, generally referred to as giant cells (Riesen- zellen) were, by Duval, Sobotta (earlier publications) and Gros- ser thought to be of embryonic origin and derived from the cells of the parietal ectoderm. Selenka, Disse, Kolster, Melis- sinos, Pujiula, Widakowich, and later Sobotta (11) regard them as derived from the maternal tissue and as representing differ- entiated decidual cells. It is not my purpose to consider more fully these cells in the present communication, since they are by me not regarded as of embryonic origin. My own observa- tions as concerns them agree in the main with those of Wida- kowich, who, in the albino rat has followed their origin from decidual cells. Since not of embryonic origin, they have been disregarded in making the figures. I have previously, in connection with a discussion of the structure of vesicle C, figure 24, alluded to the fact that the cells of the ectoplacental cene as also the cells of the parietal or transitory ectoderm have a phagocytic action for maternal blood cells. This Sobotta has also observed for the mouse, in which he is confirmed by Kolster who has further shown that the cells of the ectoplacental cone also take up fat particles. Withthe ingestion of maternal blood cells by the cells of the ectoplacental cone, more particularly, with the absorption of hemoglobin by the entodermal cells of the mesometrial portion of the egg- eylinder, a period of rapid growth of the egg vesicle is initiated. To this Sobotta has called attention for the mouse; the same 336 G. CARL HUBER is evident in the albino rat. Indeed, Sobotta presents the far-reaching conclusion that the explanation of the phenomenon of germ layer inversion or entypy of the germ layers is to be found in the dearth of food supply of the ovum in the stages preceding the formation of more definite relations between the ova or germ vesicles with the decidua. It is thought by this observer that the inversion of the germ disc has for its purpose the increase of the absorptive surface of the visceral or yolk sac entodermal epithelium, which as a differentiated layer comes to surround nearly the whole of the egg-cylinder on comple- tion of the inversion, and is thus increased in extent and brought in relatively close relation with the maternal blood lacunae surrounding the egg vesicle. LATE STAGES IN EGG-CYLINDER DIFFERENTIATION AND THE ANLAGE OF THE MESODERM In the rat series there are found 24 egg cylinders showing the stages of development considered in this section; certain of them are cut longitudinally and others cross-wise. For the special consideration of egg-cylinder formation just prior to the anlage of the mesoderm, I present two egg-cylinders obtained during the latter half of the ninth day after insemination; one of these was cut longitudinally, the other in favorable cross- section. The egg-cylinder shown in figure 29, rat No. 40, 8 days, 17 hours after insemination, seems unusually well fixed, as evidenced by its symmetrical outline, and is cut in a very favorable plane. The sections are from a series cut at right an- gles to the long axis of the uterine horn. The decidual crypts lodging the egg-cylinders of this stage are by this time nearly completely separated from the lumen of the uterus, and are surrounded by a well-developed decidua. Extravasated mater- nal blood nearly surrounds such egg-cylinders. Fig. 29 Longitudinal, sagittal section of egg-cylinder of the albino rat show- ing the final mesoderm-free stage. > 200. Rat No. 40, 8 days, 17 hours, after insemination; ect.pl., ectoplacental cone or Trager; p.ect., parietal or transitory ectoderm; pr.emb.ect., primary embryonic ectoderm; ew.ect., extraembryonic ectoderm; pr.c., proamniotic cavity; v.ent., visceral entoderm, absorptive for maternal hemoglobin, cells showing the three zones described by Sobotta; pr.emb.ent., primary embryonic entoderm. iy BS €ct pl. ——-€x. ect, *¢ bent --- Prem 337 338 G. CARL HUBER The egg-cylinder shown in figure 29 presents a total length of 1.15 mm., a width of approximately 0.18 mm. The ectopla- cental cone presents a length of 0.4 mm. and of the proamniotie cavity, 0.6 mm., of which 0.2 mm. falls to the antimesometrial portion lined by primary embryonic ectoderm. This egg-eylin- der differs only in shape and size from that shown in C of figure 27, obtained 8 days after insemination. The primary embryonic and extraembryonic ectoderm lining or enclosing the proam- niotic cavity are readily differentiated. The primary embryonic ectoderm, derived from the ectodermal node, constitutes a pseu- dostratified epithelium, composed of relatively long columnar cells, with nuclei radially placed with reference to the lumen of the proamniotic cavity, and shows active cell division, no less than 12 mitotic figures occurring in the section figured. The protoplasm of its cells stains distinctly deeper than does that of the cells of the extraembryonic ectoderm. The cells of the latter are of cubic, short columnar, or polyhedral shape, ar- ranged in a single or double layer, with no definite arrangement of the long axes of its nuclei. It is, therefore, possible readily to distinguish—by reason of shape and size of cells, relative posi- tion of nuclei, reaction to stain of protoplasm—between the cells of the primary embryonic and extraembryonic ectoderm, and to determine the sharp line of Junction at which the two types of cells form a continuous layer, a fact which will receive further consideration in dealing with the anlage of the mesoderm as observed in slightly more advanced stages. At the meso- metrial end of the proamniotic cavity, the cells of the extraem- bryonic ectoderm become continuous with the cells at the base of the ectoplacental cone; in the region of this junction, active mitosis are often to be observed. In this egg-cylinder the visceral entoderm may readily be differentiated into two portions. The portion which surrounds the primary embryonic ectoderm to nearly the region of its junction with the extraembryonic ecto- derm, censists of a single layer of broad, flattened cells which assume a cubic or short columnar shape as the mesometrial border of the primary embryonic ectoderm is approached. This portion of the visceral entoderm we have designated as DEVELOPMENT OF THE ALBINO RAT 339 the primary embryonic entoderm. The portion of the visceral entoderm surrounding the sides of the egg-cylinder in the region of the extraembryonic ectoderm, to near the base of the ecto- placental cone, consists of a single layer of columnar cells, regu- larly arranged and presenting the three zones described by So- botta. In this stage of egg-cylinder development of the albino rat, the absorption of hemoglobin granules derived from maternal blood cells, first shown for the mouse by Sobotta and Kolster, may be readily made out. In preparations stained in hematoxylin — and Congo red, in and on the outer zone of the visceral entodermal cells there may be observed granules staining deeply in the Congo red, presenting the color reaction of hemoglobin. In the mid- dle zone of these cells the protoplasm is distinctly vacuolated, while the inner zone, containing the nuclei, presents a denser protoplasm. The transitory or parietal ectoderm consists of a homogeneous membrane, closely adherent to the maternal de- cidua, especially along the sides of the egg-cylinder. This layer presents scattered nucleated protoplasmic masses of spindle or dome shape on its inner surface, the relations and distribution of which may be clearly seen in the figure. Attention needs yet be drawn to the ectoplacental cone of the egg-cylinder. Its relation to the maternal decidua is very intimate, so that in places, owing to blood extravasations, it is difficult to differentiate between embryonic and maternal tissue. Many of the cells of the ectoplacental cone present a vacuolated protoplasm, the vacuoles enclosing maternal blood cells. Therefore, they are distinctly phagocytic. Sobotta has also observed and described this for the mouse. Referring to a slightly older stage after the anlage of the mesoderm, his own words read as follows: Weiterhin sehen wir im Stadium der Fig. 5 auch eine starke Ver- langerung und Vergrésserung des Ectoplacentarconus, an dem im meso- metralen Teile jetzt Vacuolen auftreten, die in spateren Stadien regel- massig gefunden werden und zwar erfillt mit miitterlichen Blutex- travasaten. Die Ebrnaihrung des Embryo mit miitterlichem Himo- slobin * * * * ist jetzt im vollen Gang. Absorption of maternal hemoglobin by the cells of the ecto- placental cone appears to be established at a relatively earlier period in the rat than in the mouse. 340 G. CARL HUBER The egg-cylinder presented in figure 29 constitutes the final mesoderm-free stage, the final stage in which no distinct bilateral- ity may be determined. I assume that the egg-cylinder pre- sented in the figure is cut in the sagittal plane. This assumption is based on the fact that the primary embryonic ectoderm ex- tends slightly farther toward the mesometrial pole on the one side than on the other. In good frontal sections one side of the egg-cylinder in this stage of development should present a mir- ror picture of the other side. The side on which the primary embryonic ectoderm extends farther toward the mesometrial pole, the left in the figure, is regarded as containing the caudal end of the future embryo. In the primary embryonic ectoderm of this region, it is believed, will develop the primitive streak and groove, and thus the anlage of the mesoderm. Not in all the egg-cylinders of this stage of development found in my series can the caudal end of the future embryonic area be postulated prior to the anlage of the mesoderm, and in cross-sections no such differentiation can be made. The proamniotic cavity of the ege-cylinder shown in figure 29 presents a regular and nearly smooth contour, not divisible into regions such as described for a similar stage for the mouse by Selenka, Melissinos, and Sobotta. A very slight constriction is to be observed only in the region where the primary embryonic and extraembryonic ectoderm are joined in a continuous layer. I am thus wholly in accord with Widakowich, who in describing a similar stage in one of his preparations, states: ‘‘Das war die einzige konstante, bald stirker, bald schwiicher ausgeprigte.Einschniirung der Proamnionhohle,” as previously quoted. A series of figures of critical regions taken from a series of cross-sections of an egg-cylinder of a stage nearly identical with that shown in figure 29, though of a shghtly smaller egg-cylinder, is given in figure 30, rat No. 42, 8 days, 16 hours, after insemina- tion. The sections chosen for the several drawings, A to D, are from the following regions, as may be ascertained by compari- son with figure 29; A, through about the middle of the ectoplacen- tal cone; B, through the proamniotic cavity just below its meso- metrial end; C, through the proamniotic cavity just above the DEVELOPMENT OF THE ALBINO RAT 341 region of the junction of the primary embryonic and extraem- bryonic ectoderm; D, a little above the middle of the antimeso- metrial portion of the proamniotic cavity. The levels of the Fig. 30° Four figures from a series of cross sections of an egg-cylinder of the albino rat in the stage of development shown in figure 29. > 200. Rat No. 42, 8 days, 16 hours after insemination. The levels at which the several sections drawn were taken is approximately indicated by the several crosses found to the left of figure 29. A, middle of ectoplacental cone; B, ectoplacental end of the proamniotic cavity; C, just above level of junction of the primary embryonic and extraembryonic ectoderm; a little above the middle of primary embryonic ectoderm. The want of any definite bilateral symmetry of albino rat egg-cylinders of this stage of development is shown by this series of sections; p.ect., parietal or transitory ectoderm; ez.ect., extraembryonic ectoderm, surrounding mesometrial portion of proamniotic cav- ity; pr.emb.ect., primary embryonic ectoderm; v.ent., visceral entoderm; p.emb. ent., primary embryonic entoderm; pr.c., proamniotic cavity. 342 G. CARL HUBER several sections drawn in figure 30 is approximately indicated by the several crosses found to the left of the egg cylinder drawn in figure 29. In A of figure 30, there may be observed a vacuolization of the protoplasm of the more peripherally placed cells of the ecto- placental cone, the vacuoles enclosing maternal blood cells. The more centrally placed cells of this ectoplacental cone show a tendency to concentric arrangement. Figures B and C present structural appearances nearly identical. The egg-cylinder is bounded by the thin layer of parietal or transitory ectoderm having scattered masses of nucleated protoplasm on its inner surface. This membrane of apparently homogeneous structure stains sharply in well fixed preparations and may be readily discerned. The cells of the visceral entoderm, somewhat taller in the section taken nearer the antimesometrial pole (C), present clearly the three zones to which attention has been drawn. The cells of the extraembryonic ectoderm bounding the mesometrial portion of the proamniotic cavity, are of cubic, short columnar, or poly- hedral form disposed in single or double layer, presenting relative- ly lightly staining protoplasm. In D of figure 30, the cells form- ing the primary embryonic ectoderm are of distinct columnar shape, with relatively deeply staining protoplasm and nuclei arranged nearly in a single layer except for such as show mitotic phases. The cells of the primary embryonic entoderm are of a broad, pavement type for a greater part of the circumference, and may be contrasted with the cells of the visceral entoderm shown in B and C of the figure; the latter are absorptive cells, the former not. This series of figures, more especially B, C, and D, show clearly the absence of bilaterality in the egg-cylinders of the albino rat at this stage of development. The slight com- pression observed in this egg-cylinder, as shown in the figures, I regard as not of moment. Fig. 31 Longitudinal sagittal section of egg-cylinder of the albino rat show- ing anlage of the mesoderm. X 200. Rat No. 34, 8 days, 18 hours, after insem- ination; ect.pl., ectoplacental cone or Trager; p.ect., parietal or transitory ectoderm; pr.emb.ect., primary embryonic ectoderm; ex.ect., extraembryonic ectoderm; pr.emb.ent., primary embryonic entoderm; mes., mesoderm in anlage; pr.c., proamniotic cavity; v.ent., visceral entoderm. v. ent___ y Si JOURNAL OF MORPHOLOGY, VOL. 26, No. 2 344 G. CARL HUBER Grosser has figured in his figures 68 and 114, an egg-cylinder of the albino rat which measures nearly 2 mm. in length. The age of this is given as 83 days. So far as may be determined from his figures, the preparation is not described in his text, the age, size, form, and structure of the egg cylinder shown in figure 29 and Grosser’s figures 68 and 114, are very similar. In Grosser’s figures, I see no evidence of his having differentiated between primary embryonic and extraembryonic ectoderm, while the reference letters for ectoderm and entoderm are reversed. Selen- ka’s figure 31, plate 45, may be of a similar stage. This figure is, however, too diagrammatic to admit of close study. No differ- ence is shown in the shape and structure of the cells bounding the two parts of the proamniotic cavity. Christiani’s figure 39 may be of the same stage, but is too schematically drawn. Figure 4 of the article of Widakowich is of a slightly older stage and presents only a part of the egg-cylinder; it is recorded as about 62 days old. The stage under consideration is not figured by Widakowich, although his text description corresponds closely with what has been here presented. The next stage and the one with which this communication is to be completed is one of importance since it is characterized by the anlage of the mesoderm. My own observations may be introduced with the consideration of an egg-cylinder, a section of which is presented in figure 31, rat No. 34, 8 days, 17 hours, after insemination. This was cut in the sagittal plane and measures 1.1 mm. by 0.2 mm., of which 0.4 mm. fall to the ectoplacental cone. This egg-cylinder is almost an exact dupli- cate, both in size and form, of that figured in figure 29 of the same age. In the egg-cylinder shown in figure 31, however, there may be observed, to one side, in the region of the junction of the primary embryonic and extraembryonic ectoderm, and between primary embryonic ectoderm and entoderm, a small group of cells which lie in close relation to the ectoderm and constitute early mesodermal cells. The sections of this series pass not exactly parallel to the mid-sagittal plane throughout the whole extent of the egg-cylinder; especially is this true of its an- timesometrial portion, in the region of the primary embryonic DEVELOPMENT OF THE ALBINO RAT 345 ectoderm. This portion in the section figured, passes a little to one side of the mid-sagittal plane. The two sections preced- ing the one figured enclose the mid-sagittal plane, and in them, the group of cells found between primary embryonic ectoderm and entoderm are in closer relation to the ectodermal layer and at all points distinctly separated from the entoderm. They are regarded as having wandered from the primary embryonic ectoderm to the place they oceupy, a fact which is more easily ascertained in cross sections of a similar stage, as will appear from further discussion. From a study of very slightly older stages it can be determined that this region constitutes the primitive streak region of the future embryonic area. It is not my purpose at this time and in this communication to give es- pecial consideration to the much discussed question of the origin of the mesoderm in Mammalia. In the rat, this question is complicated by the question of the anlage of the amniotic fold, which separates the proamniotic cavity into amniotic cavity proper and the ectoplacental cavity, the development of which will be considered in a projected contribution. In anticipation of this second publication, however, the following facts may here receive consideration. Widakowich presents in his figure 4, giving only the antimesometrial end of an egg- cylinder obtained the latter part of the 7th day, the anlage of the mesoderm as observed by him. This figure and my own figure 31 present almost identical relations, his figure showing only three mesodermal cells between primary embryonic ectoderm and entoderm. His own words concerning the anlage of the mesoderm in the albino rat, with which I find myself in full accord, except as to the age of the egg-cylinder, read as follows: Das erste auftreten des Mesoderms beobachtete ich an Keimen vom Ende des 7 Tages. Die ersten Mesodermzellen liegen im Bereiche der vom mesometralen Ende des stirker fairbbaren primiren embryo- nalen Ectoderm gebildeten Falte. Es kommt hier eine ganz bestimmte Stelle in Betracht, die dort liegt, wo sich spiter das hintere Ende des Primitivstreifens befindet. There is, however, wide divergence of the views of authors as concerns the anlage of the mesoderm in the rat and mouse. 346 G. CARL HUBER Selenka, it would seem, in part at least, interpreted correctly the development of the mesoderm in the rat, although a stage showing its anlage was not observed. Duval believes that the mesoderm has origin from a thickened part of the entoderm, probably in the region of the anterior portion of the future em- bryonic area; the primitive streak was not recognized. Christi- ani’s figures 45 and 47, transverse sections of the egg-cylinder from the eighth day, give correctly the relative position of the mesoderm with reference to the primitive streak; however, they show stages some little time after the anlage of the mesoderm. According to Robinson, in the early part of the eighth day the cavities of the epiblast (primary embryonic ectoderm) and of the trophoblast (extraembryonic ectoderm) meet and fuse to form a hollow cylinder, the proamniotic cavity. He states that ‘“‘For a time the united cavities of the epiblast and tropho- blast increase in size, together with the general growth of the ovum, and this increase continues until in the latter part of the eighth day the mesoblast appears around the margin of the epiblast where it is in apposition with the trophoblast.”’ Robin- son was able to differentiate between the primary embryonic ectoderm (epiblast) and the extraembryonic ectoderm (tropho- blast) and his figure 14 (plate 23-24), though schematic, shows that he recognized the positions of the anlage of the mesoderm correctly, as also its derivation from the primary embryonic ectoderm. The observations of Melissinos, bearing on the an- lage of the mesoderm have been critically reviewed by both Widakowich and Sobotta, and I am wholly in accord with their views when they state that no credence can be given these ob- servations since it is clear that Melissinos has confused sagittal and frontal sections in such a way as to make his observations of no value. According to Melissinos, the mesoderm arises from the outer surface of the middle fold. of the egg cylinder, in the region of its union with the antimesometrial ectodermal fold; it is certain that it does not arise from the part of the egg- cylinder that has differentiated from the primary embryonic ectoderm; but, if I interpret him correctly, from the extra- embryonic portion of the ectoderm. That Melissinos did not DEVELOPMENT OF THE ALBINO RAT 347 have before him the stages showing the anlage of the mesoderm seems clear. Sobotta’s (11) observations, mouse material, deserve fuller consideration. In interpreting his results, I am mindful of the fact that he was unable to locate the line of union between primary embryonic and extraembryonic ectoderm, as can readily be done in suitable rat material, as has previously been shown by Robinson and Widakowich, and to which atten- tion has constantly been drawn in this communication. I am unable to state from personal observation whether in the white mouse these two types of ectoderm which form the lining of the proamniotic cavity, can be differentiated on ascertaining the right technical method. Sobotta’s material seems well fixed. If not, it would seem to me difficult to determine. definitely the exact place of origin of the mesodermal cells, whether extra- embryonic or embryonic. Sobotta recognized the anlage of the mesoderm in the mouse during the last hours of the seventh day or first hours of the eighth day. This is said to appear at the caudal end of the future embryo as a group of loosely ar- ranged cells lying between the inner and outer layers of the ege-cylinder. At the place where the mesodermal cells arise from the inner layer of the egg-cylinder, there is developed a fold, recognized as the caudal amniotic fold (‘““Schwanzfalte des Amnios’’). After discussing these observations at length, Sobotta concludes as follows: Was die Deutung dieser friihen Stadien der Mesodermbildung in der Keimblase der Maus anlangt, so handelt es sich hier nicht um die Bildung des embryonalen Mesoderms, die erst mit der eigentlichen Gastrulation spater einsetzt, sondern um Entstehung ausserembryo- nalen Mesoderms, besonder des Teils des mittleren Keimblattes, dass bei der Bildung der primaren Eihéute, Amnios und Chorion in Betracht kommt und des den ausserembryonalen Teil der Leibeshéhle, das Exocoelom auskliedet, der Héhle, die eben Amnios und Chorion von- einander trennt. Es erfolgt also, um einen kurzen Ausdruck zu gebrau- chen, die Bildung des Amniosmesoderms. An embryonic anlage is said not to exist at this stage; this is recognized only after the development of the primitive streak. It is not my purpose to enter fully into a discussion of this im- portant question in this communication. This would involve 348 G. CARL HUBER consideration of older stages, and the making of a number of reconstructions, which it is not contemplated to consider now. It must suffice to state at this time that in the albino rat, as shown by Widakowich and here shown by me, it is possible to delineate clearly the primary embryonic ectoderm and to show that the first evidence of the mesoderm is found antimesometrial to the future amniotic fold and in the region of the future primi- tive streak; therefore is mesoderm which I would regard as peristomal mesoderm in the sense of C. Rabl, reference to which is made by Sobotta in his discussion of this question. It may be that the rat offers more suitable material for the elucidation of this question than is to be found in the mouse. In the albino rat, the anlage of the mescderm is from the sagittal portion of the caudal region of the primary embryonic ectederm, the caudal part of the future primitive streak and antimesometrial to the amniotic fold. Sobotta gives very favorable consideration to the observations of Widakowich, touching this question, which he regards as “Bei weitem die beste Darstellung des Gegenstan- des.’”’ My own observations fully confirm those of Widakowich. These questions will receive fuller consideration in a later pub- lication dealing with the embryology of the albino rat, carrying the development from the time of the anlage of the amniotic fold to the stage of embryo form, the material fcr which is at hand. In figure 32 are shown cross-sections of the antimesometrial portion of three egg-cylinders in the region of the developing mesoderm. Sections drawn in A and B, were taken respectively from egg-cylinders obtained from the same uterus as was the one shown in sagittal section in figure 31, rat No. 34, 8 days, 17 hours, after insemination; C, from rat No. 41, 8 days, 16 hours, after insemination. It is very probable that the series from which A of this figure was drawn, is not cut in exactly the cross plane. .0..< 200. 200. Rat No. 94, 8 days after the beginning of insemina- tion. ect.pl., ectoplacental cone or Triiger; v.ent., visceral entoderm; met.pr., mesometrial portion of the proamniotic cavity; p.ect., parietal or transitory ectoderm; pr.emb.ent., primary embryonic entoderm; ect.n., ectodermal node; a.met.pr., 1mperfectly developed antimesometrial portion of proamniotic cavity; ex.ect., extraembryonic ectoderm. ectoderm, in structure and relation to decidual crypt, are to be regarded as of normal development. The visceral entoderm, surrounding the extraembryonic ectodermal portion of the egg- cylinder, is of normal structure, showing the three zones evidenc- 378 G. CARL HUBER ing its absorptive function. The extraembryonic ectoderm, enclosing the mesometrial portion of the proamniotic cavity, presents normal structure and relations of cells. The only ab- normality observed is in the region of the ectodermal node, the anlage of the ectodermal vesicle with the enclosed antimesometrial portion of the proamniotie cavity. With this stage of develop- ment of the egg-cylinder (see figs. 26 and 27, Part I) the ectoder- mal node presents a well formed cavity, surrounded by the cells of the primary embryonic ectoderm, radially arranged. In the egg-cylinder under discussion (fig. 8) there is distinctly a retar- dation in the development of the ectodermal vesicle with full differentiation of the primary embryonic ectoderm. An imper- fectly developed antimesometrial portion of the proamniotic cavity is evident. This small cavity, indistinctly bounded, ex- tends obliquely through several sections of the ectodermal node, and contains amorphous granular detritus, which in the prepa- rations is stained by Congo red. The cells destined to form the: primary embryonic ectoderm show no definite arrangement, especially as concerns the more centrally placed cells of the node. Since the primary embryonic ectoderm is the anlage for the ectoderm of the embryo, an arrest in its differentiation would of necessity profoundly affect further development of the embryo. Antimesometrial to the ectodermal node (just above it in the figure) there is found a small vesicle the walls of which are not distinctly delimited and composed of extraem- bryonic ectodermal cells, surrounding a small, completely bounded cavity. J am not prepared to say whether this small vesicle is to be regarded as developing from cells of the extraembryonic ectoderm, or from a displaced, accessory ectodermal node, in which a discrete portion of the proamniotic cavity has developed. If the latter, the possibility of a double anlage for the embryonic ectoderm is to be considered. My interpretation of this egg- cylinder as showing a retardation of the development of the ectodermal node and differentiation of the primary embryonic ectoderm, is confirmed from a study of a slightly older stage show- ing essentially the same condition. This ovum is presented in figure 9, and is taken from rat No. 41, 8 days, 16 hours, after the Fig. 9 Egg-cylinder of albino rat, in which the antimesometrial and mesometrial portions of the proamniotic cavity have failed to unite to form a single or definite proamniotic cavity. X 200. Rat No. 41, 8 days, 16 hours, after the beginning of insemination. ect.pl., ectoplacental cone or “Triger; p.ect., parietal or transitory ecto- derm; v.ent., visceral entoderm; met.pr., mesometrial portion of the proamniotic cavity; ex.ect., ex- traembryonic ectoderm; a.met.pr., antimesometrial portion of the proamniotic cavity; pr.emb.ect., primary embryonic ectoderm; +, region at which, in normal de- velopment, by the end of the eighth and beginning of the ninth day, the two portions of the pro- amniotic cavity would have united to form a single space, the definite proamniotic cavity. a Mas yrrs\ (9) % GaN or (os ae ve ‘o ey 5 4 a OD pe ip : im a) aS ee Om at RIO AO 08 O@ S EOS POS sala, ae iA ~ [Es —(_pr,emb, ect. 380 G. CARL HUBER beginning of insemination. The uterus of this rat contains eight egg-cylinders, all of which, except the one here figured, show normal structure, though presenting quite different stages of development. One of these, cut serially in cross-section, is figured in C, figure 32, Part I, as showing anlage of mesoderm with primitive streak and groove. Two of the other egg-cylinders show the anlage of the mesoderm, two others show late pre- mesoderm stages of the egg-cylinder, the remaining egg-cylinders are less fully developed, one showing a development which may be compared to B of figure 26, Part I, thus a much younger stage. By the end of the eighth day and with the early hours of the ninth day after the beginning of insemination in the albino rat, the two parts of the proamniotic cavity, which de- velop discretely, have joined to form a single space (C, fig. 27, Part I). The egg-cylinder shown in figure 9, presents normal development in all parts, except that there is as yet no union of the two parts of the proamniotic cavity. This egg-cylinder is most favorably cut, in longitudinal direction; the plane of see- tion being almost parallel to the mid-sagittal plane. This egg- cylinder, therefore, is easily followed through the several sections of the series into which it was cut. The irregularity of outline of the ectodermal vesicle, lower right of figure, it is believed, is not due to fixation shrinkage. Judging from size and structural differentiation of this egg-cylinder, union of the antimesometrial and mesometrial portions of the proamniotic cavity should have been completed before this stage of development was reached, with the primary embryonic ectoderm and the extraembryonic ectoderm forming a continuous layer, as shown in figure 29, Part I. The folding of the wall of the antimesometrial portion of the egg-cylinder, lower right of figure, evident in nearly all of the sections of the series, is regarded as indicating an abnormal growth of the primary embryonic ectodermal cells composing the wall of the ectodermal vesicle, as a result of retarded extension of the antimesometrial portion of the proamniotic cavity, perhaps an adjustment to meet the altered mechanical stress resulting from abnormal development. The condition here seen, it would seem, 1s foreshadowed in the egg-cylinder shown in figure 8. pr.emb. ect,— ; \ae Pr. em». ent. —Qe Fig. 10 Two egg cylinders of the albino rat found within the same decidual crypt, with in part common ectoplacental cone. X 150. Rat No. 87, 9 days after the beginning of insemination. ect.pl., ectoplacental cone or Triiger; p.ect., parietal or transitory ectoderm; v.ent., visceral entoderm; ex.ect., extra- embryonic ectoderm; pr.c.; proamniotic cavity; pr.emb.ect., primary embryonic ectoderm; pr.emb.ent., primary embryonic entoderm; mes., mesoderm. 381 382 G. CARL HUBER The causes operative in this retardation of development and differentiation of the ectodermal vesicle and primary embryonic ectoderm, | have been unable to determine. They would ap- pear to be inherent in the egg-cylinder, since ectoplacental cone and visceral entoderm, so far as may be determined from a study of sections, appear to have functioned normally, in furnishing the necessary embryotroph in the form of maternal hemoglobin, as is normal for egg-cylinders of the albino rat of this stage of development: TWO EGG-CYLINDERS IN ONE DECIDUAL CRYPT The ova portrayed in figure 10 present a condition which must be regarded as exceedingly rare, since it represents the only instance of this condition observed in the extended series of preparations of the various stages of the development of the albino rat from the end of the first to the end of the ninth day after insemination, in my possession. This preparation is from rat No. 87, 9 days after the beginning of insemination. The uterus of this rat con- tained, other than the preparation here considered, six egg- cylinders of normal development, all showing a stage which is slightly older than that shown in figure 31, Part I, in that the mesoderm shows further development than is shown in that figure. In the preparation here figured there are found two ege-cylinders enclosed within the same decidual crypt. This figure, which is drawn by combining the drawings made from two sections, is reproduced at a magnification of 150 diameters, while all of the other figures portraying sections of ova, both in Part IT and in Part IT of this communication, are reproduced at a mag- nification of 200 diameters. This should be borne in mind when comparing this figure with the others. In figure 10, the lower portion of the large egg-cylinder to the level of the lower end of the smaller one was drawn from one section, while the remainder of the figure was drawn from the fourth following one. The adjustment was made by overlapping in the camera lucida drawing (xX 600) the sharp mesometrial border of the primary embryonic ectoderm of the larger egg-cylinder. Scarcely any PATHOLOGIC OVA, ALBINO RAT 383 adjustment was found necessary, none of the right wall of the larger egg-cylinder, and only very slightly so of its left wall. The slight deviation from the longitudinal axis of the larger egg- cylinder made the procedure desirable. It is thought that the figure as presented gives correctly the size of the respective egg- cylinders, and in all essentials, their relations; the greater part of the figure having been drawn from one section. Both of the egg-cylinders reveal normal structure for the stages of develop- ment attained. The larger one is cut in the coronal plane, as is readily determined by the distribution of the mesoderm, one side representing a mirror picture of the other. The direction of sec- tion in the smaller egg-cylinder, except that it is longitudinal, is not to be determined, since before the anlage of the mesoderm, a bilateral symmetry cannot be recognized in sections. Since these two egg-cylinders are in all essentials of normal form and structure, and since their structure is clearly brought out in the figure, an extended description of them at this place seems un- called for. For respective stages the reader is referred to Part I. Attention may be drawn, however, to the fact that the visceral entoderm on the contiguous surfaces of the two egg-cylinders is less fully differentiated, and shows less absorption of the ma- ternal hemogiobin than is seen on the exposed or free surfaces, this, no doubt, for mechanical reasons. Further, that in the region where the two egg-cylinders are in contact, the parietal ectoderm of each can be traced as a distinct layer to the bases of the respective ectoplacental cones, showing that each developed from a separate ovum. The ectoplacental cones are for a short distance distinct. In tracing the sections through the series the impression is gained that the ectoplacental cone of one of the egg- cylinders overlaps that of the other in such a way that in the plane of the sections obtained, one seems continuous with the other, as represented in the figure. The boundary between the two is not distinct, and it would seem that as a result of pres- sure, partial fusion of the two had taken place. The presence of two egg-cylinders, enclosed within a single decidual crypt, as shown in this figure, with one of them having much smaller size and representing a younger stage of development, I believe is 384 G. CARL HUBER not to be explained on the supposition of superfecundation or superfoetation. The record for this rat does not show insemi- nation on successive days. At The Wistar Institute, after all of the supposedly successful matings of albino rats, the females rats are caged apart from the males. The smaller egg-cylinder, though appreciably smaller, is in stage of development separated from the other by a time interval of perhaps less than 24 hours. It presents a stage of development which is comparable to C of figure 27 (8 days) and except for size, to the one figured in figure 29 (8 days, 17 hours) of Part I. It is believed that in this case both ova were seminated at about the same time, and pro- ceeded through normal segmentation and that on reaching the lumen of the uterus during the fifth day they became lodged in close proximity in the same mucosal fold. With the development of the decidual crypts, both became enclosed within the same erypt, at perhaps slightly different levels. In further develop- ment one blastodermic vesicle dominated the other and from about the seventh day on, one developed and differentiated more rapidly than the other. Had development continued, two distinct embryos, with separate amniotic cavities, attached to the same placenta, would have been formed, with one embryo large and more fully developed than the other. From mere difference in size and of development of embryos in the same litter it is not warranted to postulate superfecundation nor super- foetation. I am of the opinion that usually when two morula masses are lodged in close pres: in the same mucosal fold, one or the other degenerates (fig. 2, A) and that the normal development of both, as in the preparation shown in figure 10, is of very rare occurrence. CONCLUSIONS A study of the abnormal or pathologic ova met with in the ex- tended series of preparations covering the first ten days of the development of the albino rat, enables grouping them in two main classes: a. Such in which all of the ova of a given rat show, or are associated with, abnormal development. PATHOLOGIC OVA, ALBINO RAT 389 b. Such in which a single abnormal or pathologic ovum is found in the same uterus along with an average number of normally developed ova. When all the ova in a given uterus show abnormality, the presumption seems warranted that the underlying cause of the abnormality is to be sought in an altered or pathologic condition of the uterine mucosa. In the instances observed, the presence of maternal blood with many phagocytic leucocytes was noted in the lumen of the uterus, adhering to and surrounding the ova. From the study of sections of the uteri of an appreciable number of albino rats, in which insemination and supposedly semination seemed normal, but in which on complete serial sectioning of the uterine tubes no ova were found, but in the lumen of the uterine tubes of which the presence of maternal blood and phagocytic leucocytes was noted, the conclusion seems warranted that death and complete absorption of ova, after a given stage of nor- mal development has been reached, may occur. In such eases, one may with propriety speak of faulty implantation, due to altered or pathologic condition of the uterine mucosa, even in cases where no actual implantation would have occurred in cor- responding normal stages. In the two rats (Nos. 91 and 104) in which this condition was observed, the decidual crypts were shallow and not developed to the extent normal for the respec- tive stages, evidencing the abnormal condition of the mucosa. In cases in which a single abnormal or pathologic ovum is found in the uterus along with several normal ova, the pre- sumption seems justified that the underlying cause responsible for the abnormal development is to be sought in the ovum itself, and not in its environs. Abnormal developmental stages, interpreted as due to irregu- lar or retarded segmentation, irregular or abnormal segmenta- tion cavity formation, and retarded development of the ecto- dermal node and primary embryonic ectoderm, where only a single ovum shows abnormal development in a uterus contain- ing the average number of ova presenting normal development, are difficult to explain on the assumption that extraneous in- fluences affecting a single ovum are operative. Practically all 386 G. CARL HUBER of the abnormal ova of the class described, and especially is this true for older stages, present normal relations to the uterine mucosa and the walls of the decidual crypt after implantation, and so far as may be determined by structure, give evidence of normal absorption of maternal henaoglobin in stages in which such absorption is pertinent. It may be argued that a single ovum may be less favorably placed in relation to embryotroph or pabulum, and as a result of unfavorable nutrition, develop abnormally. This is difficult to conceive for stages in which the ova lie free in the lumen of the uterus, namely, to about the be- ginning of the seventh day after the beginning of insemination, when embryotroph or pabulum must be relatively evenly dis- tributed. The presumption, it would seem to me, in such cases is in favor of regarding the primary cause of the abnormal de- velopment as inherent in the ovum. Separation of the first two blastomeres and the presence of two egg-cylinders in a single decidual crypt are regarded as chance findings and as of rare occurrence, since each was met with only once in the material at hand. LITERATURE CITED Literature on pathologic ova of the albino rat is lacking. For the literature of all but the more recent work, dealing with comparative experimental tera- tology, the bibliographies accompanying the chapters of O. and R. Hertwig may be consulted; for that dealing with the pathology of human ova, the bib- liographies accompanying the contributions of F. P. Mall may be consulted. Hertwia, O. 1906 Missbildune und Mehrfachbildung, die durch Stérung des ersten Entwicklungsprozesse hervorgerufen werden. Hertwig’s Hand- buch der vergleichenden und experimentellen Entwickelungslehre der Wirbeltiere, Bd. 1, Part 1; Fischer, Jena. Hertwie, R. 1906 Der Furchungsprozess. Hertwig’s Handbuch, Bd. 1, Part 1. Matt, F. P. 1900 Welch Festschrift, Johns Hopkins Hospital Reports, vol. 9. 1903 Vaughan Festschrift, Contributions to medical science, G. Wahr, Ann Arbor. D 1908 40. For abbreviations, see figure 8. > RCRTUS AR 404 03 4 DEVELOPMENT OF THE HYPOPHYSIS 407 The position of the adult hypophysis has been described by Sterzi (’09) and others. The long tongue-like anterior lobe lies on the median ventral wall of the inferior lobes of the brain. Its ante- rior end extends forward almost to the optic chiasma. The supe- rior part is placed posterior to this and ina more dorsal plane. It extends asfar caudally as the saccus vasculosus. The inferior sacs of Squalus do not extend as far ventrally as has been de- seribed for other selachians. But they are ventral to the superior lobe and extend farther caudalward (fig. 10). From their middle connection a slender canal joins them to the ventral side of the caudal end of the anterior lobe. 2. Early development of the hypophysis Recent work on the development of the hypophysis in elasmo- branchs shows that it arises at an earlier period than was formerly believed. Hoffmann (’96) stated that the position of the future hypophysis is well marked in Acanthias embryos of 15 somites but that there is no indication of an evagination even in 8 mm. (50 somites) and 10 mm. embryos. Haller (’96) began his de- scription of the hypophysis in Mustelus in embryos 22 mm. long. At that time, the hypophysis is already a distinct outpouching. More recently Johnston (’09) briefly described the earliest for- mation of the hypophysis in Acanthias. In an embryo of 24 somites the ectoderm from which the hypophysis develops is readily recognized. He stated that a short anterior lobe, de- veloping later, extends toward the optic chiasma. Also, that the posterior part crowds between the brain and the median mass connecting the premandibular somites. Scammon (711) men- tioned a thickened hypophyseal plate in a 5.2 mm. embryo and a beginning evagination in a 6.2 mm. embryo (50-51 somites). A median sagittal section of an Acanthias embryo 8 mm. in length? is shown in figure 2. The anterior superior end is par- tially insinuated between the brain and the median mass con- necting the premandibular somites, as had been noted by Johns- ton (09) in about the same stage. During the time that the 3In the description of the figures ‘H. E. C.’ has reference to the embryos of the Harvard Embryological Collection used for this study. JOURNAL OF MORPHOLOGY, VOL. 26, NO. 3 408 E. A. BAUMGARTNER anterior part of the hypophyseal evagination is growing between the brain and the median mass connecting the premandibular somites, the posterior portion forms a more or less prominent ridge caudal to the connecting mass of the somites. A model of such an embryo shows as an evagination, the end of which is grooved transversely by the median mass connecting the pre- mandibular somites (fig. 2). This groove may be quite promi- nent even in 11 to 12 mm. embryos and evidences of it are usually to be found at that time. In 8 mm. embryos as well as in younger ones the anterior arm of the hypophyseal outpouching shows differentiation as far forward as the postoptic recess (fig. 2). Scammon (’11) stated that the notochord comes into contact with the early hypo- physeal outpouching, and Sterzi (712) has shown such a contact in Mustelus of 8 mm. (fig. 449). No such contact has been observed in this study of Acanthias. A mid-sagittal section of a 15 mm. embryo shows that the superior end of the hypophysis has now grown well between the brain and the median mass connecting the premandibular somites. The evagination is more marked and the anterior arm is longer than in younger stages (fig. 3). A model of the hypophysis of an Acanthias embryo 19 mm. in length is shown in figure 11; this view is taken from the left lateral side. The hypophysis at this stage is an anteriorly and dorsally directed outpouching, concave on its ventral surface where it lies in close relation to the diencephalon (fig. 18). Its superior anterior end extends to the infundibular recess. The opening from the pharynx into the hypophysis is small. A sagittal section at about the median line (fig. 4) shows that the thickened anterior (ventral) wall of the hypophysis extends forward as far as the postoptic groove as Johnston (09) has figured it. Fig. 11 Left lateral view of a reconstruction of the hypophysis of a 19 mm. embryo. .X< 130. Mo, lining of mouth; H, hypophysis. Fig. 12 Left lateral view of a reconstruction of the hypophysis of a 21 mm. embryo. X 100. a, anlage of anterior end of hypophysis; H, hypophysis; IL, anlage of the inferior lobes; Mo, lining of mouth. \ ' = a \ —_ Sey 410 E. A. BAUMGARTNER Several changes have taken place in the hypophysis of a 21 mm. embryo. It is still concave, both laterally and dorso-ventrally in its ventro-anterior surface (fig. 12). The thickened anterior wall of the hypophysis, reaching almost to the preoptic groove, is now distinctly evaginated. Scammon (’11) mentioned this closing off of the anterior part in a 20.6 mm. embryo and Sterzi (12) described the formation of this ‘rostral diverticulum’ in 20 to 24 mm. embryos. The lateral side of the anterior out- pouching is sharply demarcated by the formation of the stalk connecting the hypophysis to the mouth (fig. 19). The anterior endatthis stage is almost half as wide as the posterior, from which most of the hypophysis is developed. The opening from the mouth into the early anlage is located as before, but is smaller now. 100. For abbreviations, see figure 14. the exception of a cone-shaped mass of cells connected with the oral epithelium, and an irregular area in the floor of the hypophy- sis which represents the remains of the former attachment (fig. 8). The furrow uniting the two inferior lobes across the middle line is more marked, and the superior lobe also is promi- nent. The anterior end of the hypophysis now extends forward and downward. The straightening out of the.-head bend in the development of the embryo has probably helped to bring about this change. More marked changes have taken place in a 48 mm. embryo (fig. 16). The anterior end of the anterior lobe is wider than in earler stages. A short and narrow middle part connects this portion with the caudal extremity which is considerably: wider (fig. 23). The inferior lobes are attached to the, now, ventral (caudal) side of this part. The ridge connecting the two infe- rior lobes has become very pronounced but still opens widely into DEVELOPMENT OF THE HYPOPHYSIS 415 the anterior lobe (fig. 16). The furrows separating the inferior lobes from the anterior are much deeper and wider. The in- ferior lobes have enlarged in their dorso-ventral and in their transverse diameters, and they extend laterally beyond the pos- terior extremity of the anterior lobe. Marked development has taken place in the superior lobe. The lateral furrows sepa- rating it from the anterior one are deeper, and the cranio-caudal length of the lobe has increased so that there is a projection cau- dally beyond the anterior lobe. The antero-lateral ends of the superior lobe have grown forward. The median connection of the inferior lobes is constricted from the posterior (ventral) part of the anterior lobe in a 50 mm. embryo. The inferior lobes are directly ventral to the superior lobe. There remains a short slender tube in the mid-line con- necting the inferior lobes to the anterior (fig. 9). The duct connecting them to the anterior lobe extends almost straight anteriorly. A median sagittal section of an 86 mm. embryo (fig. 27, G) shows an increase in the length of the hypophysis. The inferior lobes lie more caudally and the duct joining them to the anterior lobe is longer. In a 95 mm. embryo the anterior lobe has increased greatly in length (fig. 17). A median ventral sulcus has appeared and the anterior third of this lobe is quite wide. A middle narrow portion, almost circular in cross section, connects the anterior extremity to a wider posterior end (fig. 24). The caudal extrem- ity is connected dorsally with the superior lobe. The inferior lobes are continuous across the median line. The connection between the inferior lobes and the anterior one is a small tube which extends almost straight forward to join the inferior sur- face of the caudal end of the anterior lobe just below where this opens into the superior one (fig. 10). The inferior lobes have enlarged in their cranio-caudal axis. The lateral parts of the superior lobe have increased in their cranio-caudal diameters and extend forward beyond the median part. The latter has grown caudalward and lies just dorsal to the tube joining the inferior and anterior lobes. 416 j <<) =_s SL E. A. BAUMGARTNER N De \ IIE RAIS NWI Bee < 75. C, interhypophys- 7 Left lateral view of a reconstruction of the hypophysis of a 95 mm. embryo. Fig. 1 eal canal; S, median ventral sulcus; other abbreviations as in figure 14. DEVELOPMENT OF THE HYPOPHYSIS 417 4. The hypophysis of the adult A detailed description of the pup stage has been given (p. 400) as typical of the morphology of the adult condition (fig. 1). A drawing of a dissection of the hypophysis of an adult will show the position and relation of some of the parts more clearly (fig. 26). The anterior end extends almost to the optic chiasma, as noted by Sterzi and others. From a study of the models of the ‘pup’ stage and of sections of adults it appears that the middle part of the anterior lobe is more than a mere constriction separating a larger rostral from a smaller caudal part as Sterzi (09) described. In the pup this middle part is only a little shorter than either extremity, while in the adults it is much shorter. It is, however, distinctly marked. In some cases there are cystic outgrowths from the floor of this part; in others | the walls and floor are quite regular and therefore are distinctly different from either extremity. The middle part is smaller than either extremity in its dorso-ventral and lateral diameters in pups (figs. 24, 25) and in adults, but the changes in diameter from either end to the middle part in adults takes place gradually and the parts are not so definitely marked, except in those cases in which no glandular outgrowths occur from the middle region which is much more prominent in sections and in wax reconstruc- tions than is shown in dissections. The superior lobe projects some distance laterally on the ventral side of the vascular sac. It is difficult to make out the lateral limits of this part in the dissected specimen, as it seems to be continuous with the ventral surface of the vascular sac (fig. 26). In transverse sections, however, the lateral extent of the wings of the superior lobe is clear. "The superior lobe is convex dorsally and closely at- tached to the saccus vasculosus. The latter dips down on either side of the caudal end of the anterior lobe and so partially sepa- rates the lateral wings of the superior from the anterior lobe. The inferior lobes no longer resemble the sac-like structures with cystic outpouchings of the pup stage. Both superior and inferior walls have developed a mass of tubular-like glands. None of these was observed extending cranialward, as described by ILD in Y | 7 \ ——— ———————— Se ee 2 Nin \ \ Pall t\\\ Py iN J FER ee, a ghee Aa R 7 ) WW, a ~ me fr \ Mig xc RK \ ij IN \\\ Ay / i ow 1) { vi NS te ») Vy a 4 AC) a yy, a ST —— nnn (| nM CA an i => 24 Figs. 18-20 Antero-ventral views of the reconstructions shown in figures 11 to 13. Figure 18, x 50; figures 19, 20, X 40. oe, Figs. 21-23 Anterior views of the reconstruction shown in figures 14 to 16. x 40. Figs. 24-25 Ventral views of the reconstructions shown in figures 17 and 1. Figure 24, x 40; figure 25, X 25. For abbreviations, see figure 1. 418 DEVELOPMENT OF THE HYPOPHYSIS 419 Fig. 26 Drawing of a dissection of the hypophysis of an adult Acanthias from ventral side. X 20. AJL, anterior lobe; B, inferior lobes of the brain; C, inter- hypophyseal canal; JL, inferior lobes of the hypophysis; SL, superior lobe of the hypophysis; VS, vascular sac. Haller in Mustelus. From their median connection a long slender tube extends to. the caudal end of the floor of the anterior lobe. As seen in figure 26, the median connecting portion of the infe- rior lobes may be large. In some specimens this part, as well as the lobes, shows a mass of tubules. The caudal ends of the inferior lobes are surrounded by cartilage. The shifting of the hypophysis with reference to its position in the body and the development of different parts is brought out very clearly in figure 27. Sewertzoff (99), in his studies 420 E. A. BAUMGARTNER on the development of the skull, made use of such a figure to show the interrelation of development of the brain and skull. He made an outline drawing of the skull and brain of an embryo, choosing an arbitrary magnification. He then made drawings of different sized embryos of such a magnification that points, corresponding to two arbitrarily chosen in the first drawing, would coincide. In figure 27 of this paper, the same scheme was adopted. The magnification of the first drawing was such as to avoid as much as possible the confusion of lines. The points chosen were the extreme anterior end of the notochord and the axis of the notochord at the level of the first spino-occipital nerve. All the other drawings were then made so that these two points— the extreme anterior end of the notochord and the axis of the notochord at the level of the first spino-occipital nerve—should coincide with those of the first drawing. The outlines of the hypophysis were then drawn, using a line between the two points as a base. The objections to such a drawing are readily apparent. There are, of course, individual variations in the embryos. Also, these points are probably continually changing during develop- ment. For a comparative study, however, the variations can be no great objection and the points chosen are probably as relia- ble as any. A series of embryos from 11.5 mm. in length to the pup were drawn in this manner (fig. 27). One can see at a glance in all these stages the relative position of the hypophysis with reference to the anterior end of the notochord. Also, as was pointed out in the description of the different embryos, what is first the dorsal wall becomes in later stages the ventral, while the ventral or anterior becomes dorsal. The early superior end of the evagination shifts more and more caudalward with reference to the rest of the hypophysis, until, in the pup, the superior lobe, which develops from the superior dorsal end, is caudal in position. The inferior lobes, which develop from the sides of the superior end and are on the same horizontal plane as the superior position, take a position ventral to the superior lobe in the late embryo and adult. The furrows separating these inferior lobes, described as appearing in the 20 to 22 mm. embryos 421 DEVELOPMENT OF THE HYPOPHYSIS ‘dAdoU [ByTd19900 -ourds 4sigy a} JO [AV] oY} YV pAOYyDO}OU ay} JO STXB oY} 0} pAOYoo}OU oY} Jo pus OTIAVUT oy} Wody UABIp ouly oseq ‘q{-y ‘dnd ‘77 Sw gg ‘4H fu ge ‘yy {suru OF ‘g {-wU 7e ‘Gg S‘uTUT FE ‘9 suru gt fg fosaquie ‘urUt GIL ‘py ‘sokaquio serpyuBoy jo stsAydoddy ayy Jo suoroos [vy4y13¥s ULTpout Jo SUIMBIpP oyIsoduloy YZ ‘BIT xX 422 E. A. BAUMGARTNER on the ventral (anterior) side of the hypophysis, are later (50 mm. embryos) on the dorsal side. The anterior lobe, first directed almost vertically, grows ventrally and later extends more and more anteriorly until it is directed horizontally (ecranio- caudally). The comparative growth of the different portions is also made clear. The anterior lobe comprises all of the original outpouch- ing and also the anterior tongue which evaginates later. The increase in length of the anterior lobe, particularly its anterior extremity,is marked. The inferior lobes, developing from the sides of the posterior portion of the early evagination, become continuous across the posterior side (34 mm. embryos) and finally constrict entirely except for a short duct connected with the anterior lobe. The inferior lobes increase greatly in size, but, in the adult, are largest transversely. The superior lobe, de- veloped from the superior dorsal part, spreads far out trans- versely and later enlarges in its dorso-ventral axis. The posterior extremity of the anterior lobe is then developed from the first outpouching. 450. R, roof of anterior lobe; Med, medial connection of in- ferior lobes; P, pigment granules. DEVELOPMENT OF THE HYPOPHYSIS 433 Fig. 33 Sagittal section of the floor of the anterior end and of one of the tub- ules of the anterior lobe of a pup. X 400. between these are spherical shaped nuclei. He has described these in the roof of the superior lobe. These become more evident in 40 and 50 mm. embryos. In an 86 mm. embryo, the nuclei of the roof are to some extent irregular in arrangement. The cytoplasm takes orange G quite readily and hence is some- what acidophilic. In the pup, the floor is well formed and is now the thicker wall of the anterior lobe. As shown in figure 33, it is composed of four or five layers of nuclei. The oval nuclei are crowded close to the inner free surface. There is, how- ever, a narrow outer cytoplasmic zone which is very granular. The character of the cells of the glandular outgrowths is shown in figure 33. Here, too, there is a narrower outer rim of granular eytoplasm. A wider inner zone of non-granular cytoplasm bounds the lumina of the tubules. The nuclei, as in the floor, are oval. Usually two or three layers form the walls. The nuclei have a distinct chromatin network with occasional denser- staining chromatin masses resembling nucleoli. 434 E. A. BAUMGARTNER b. Inferior lobes. 'The development of the inferior lobes as constrictions of the lateral walls of the early hypophyseal out- pouching has been described above. The point where these lobes will grow together across the median line is indicated in figures 5 to 10. The character of the cells forming this part differs from the rest of the hypophyseal outpouching as early as the 21 mm. stage (fig. 32). The cytoplasm here stains less densely than that of the rest of the hypophysis. The nuclei are distinctly spherical in shape and have a very scant chromatin network. A part of this floor, immediately posterior to the hypophyseal stalk, contains a considerable amount of a granular yellowish pigment. Both Miller (71) and Hoffmann (’96) have noted the Fig. 34 Sagittal section of the floor of the hypophysis near the median plane, showing the part which connects the inferior lobes. (H.E.C. 362). 400. pigment in the stalk of the hypophysis. The nuclei become more oval in older embryos. This is well shown in figure 34 which is a mid-sagittal section of the region which later forms the con- nection between the inferior sacs. The nuclei here are very irregularly placed. Extending caudally from the stalk are several nuclei flattened along the inner free surface. The cytoplasm stains very lightly. The pigment masses are- numerous. Cau- dally there is a sudden transition to columnar cells of. the kind found in the wall of the anterior lobe. The floor in this region is as just described until the 48 to 50 mm. stages when the infe- rior sacs are completely constricted from the anterior one. ‘The pigment and the flattened nuclei are found only near the median line. Both are still present in a 41 mm. embryo. In the inferior sacs proper, however, which are formed at the lateral sides, the cells are similar to those of the floor of the anterior lobe. The outer, narrower zone of cytoplasm, as well DEVELOPMENT OF THE HYPOPHYSIS 435 Fig. 35 Sagittal section of a portion of the inferior lobe of the hypophysis of apup. X 400. as the inner, wider one, stains lightly in a 50 mm. embryo. In an 86 mm. embryo the outer zone has fine granules. The cells are slightly acidophilic. The nuclei are elongately oval in out- line as in the anterior lobe. In the pup many outpouchings indicate the beginning glandular development. The upper wall is three or four cell-layers thick (fig. 35). There is a wide, inner, clear-staining cytoplasmic zone. The outer narrower rim is slightly granular. The nuclei are oval in outline, as in the anterior lobe. Numerous densely staining chromatin masses are to be seen. The lower wall or floor is much thinner. It is composed of only one or two layers of cells and has a narrow, inner, cytoplasmic rim. This zone, as in the roof, stains very lightly and is non-granular. The outer rim is, however, quite granular. The nuclei are like those in the roof but contain a somewhat denser chromatin network. c. Supervor lobe. Ina21 mm. embryo the wall of the superior end of the hypophyseal anlage is thickest where the superior lobe later develops (fig. 36). The nuclei are large and oval and have a light chromatin network. In a 50 mm. embryo (fig. 37) the wall in this region is considerably thicker than in the 21 mm. embryo because of an increase in the number of cell layers. There 436 E. A. BAUMGARTNER Fig. 36 Sagittal section of the superior end of the hypophysis of a 21 mm. embryo. (H.E.C. 1493). x 450. Fig. 37 Sagittal section of the superior lobe of a 50 mm. embryo. (H.I.C. Series 444). x 450. is an outer cytoplasmic zone which is non-granular and stains lightly. The nuclei are smaller than in the 21 mm. embryo and contain less chromatin. As Sterzi has stated, there are some spherical nuclei and some more slender oval nuclei, but no regular arrangement of these, such as Sterzi described, was observed. In a 95 mm. embryo, the roof of the superior lobe has increased in thickness (fig. 38). The nuclei are oval and contain a denser network of chromatin than is found in 50 mm. embryos. Along the periphery many of the nuclei are spherical. There is an inner zone of cytoplasm as in the 50 and 21 mm. stages, which is DEVELOPMENT OF THE HYPOPHYSIS 437 Fig. 38 Transverse section of the superior lobe of a95 mm. embryo. (H.E.C. Series 1882). x 450. Fig. 39 Sagittal section of the superior lobe of the hypophysis of a pup showing one of the cell columns and a portion of the roof. > 400. non-granular. Fewer mitotic figures are to be found at this time. In the pup the roof is a little thicker than in the pre- ceding stage. Above the roof there are many cell columns which are outgrowths from the roof proper (fig. 39). The relations: of cell columns and roof to each other are shown in figure 39. The roof has a narrow outer zone of granular cytoplasm. Its nuclei are oval and have a light chromatin network. The cell columns come into close contact with the overlying floor of the vascular sac (fig. 39). Numerous capillaries and a loose connective tissue fill the spaces between the cell columns and between the roof of the hypophysis and the floor of the vascular sac. The columns have an outer, granular cytoplasmic zone. ‘The cells are acidophilic. The nuclei are spherical in outline, have a light chromatin network and usually one or two larger chromatin masses or nucleoli. Only a brief statement can be given at this time concerning the development of the glandular columns of the superior lobe. In the region of the superior lobe in all embryos up to the 50 mm. stage, the nuclei are elongate-oval in outline and are arranged in two or three layers (fig. 36). In a 50 mm. embryo some of the nuclei at the periphery are spherical (fig. 37). It is possible 438 E£. A. BAUMGARTNER that the elongate-oval nuclei in this region in the younger embryos are changed into spherical ones. In the pup there are numerous columns extending dorsally, which consist of a central group of spherical, light-staining nuclei and a peripheral zone of cyto- plasm. Until some material between the 95 mm. embryos and the pup stage is studied, it must remain a question how these cell columns are formed. - It is possible that the cells with spheri- cal nuclei arrange themselves in groups and these groups then evaginate from the roof. The scarcity of these groups in the 95 mm. embryos and in all late embryonic stages argues against such a possibility, although Sterzi’s observations on the presence of groups of spherical nuclei lying between masses of cells with oval nuclei should be taken into consideration. Numerous cyst-like outpouchings are present in some adults in the anterior part of the superior lobe, or, rather, between this and the roof of the caudal extremity of the anterior lobe. Some of these outpouchings show areas of cells similar to those forming the columns of the superior lobe, interspersed with areas of cells like those of the anterior lobe. The areas of cells resembling those of the superior lobe may form the entire wall or may lie on a basement of cells resembling those forming the anterior lobe which line the cavity. This would indicate that the regions of the anterior and superior lobes are not sharply separated, or, that the cells of this region which still resemble the embryonic condition change into cell columns of the superior lobe. This need not imply, however, that the cells of the anterior and infe- rior lobes are of a more embryonic type, although they may be more primitive phylogenetically. 3. Development of the interhypophyseal canal The formation of the ridge connecting the inferior lobes on the dorsal (posterior) side of the hypophyseal anlage had been described. The character of the epithelium in this region in a 21 mm. embryo, as stated above, differs from that of other parts. A ridge is prominent in a 34 mm. embryo. In a 40 mm. embryo the groove on the inside of this ridge is marked (fig. 8). Ina DEVELOPMENT OF THE HYPOPHYSIS 439 -48 mm. embryo the ridge is very distinct and the connection of the lumen of this part with the lumen of the anterior lobe has become constricted. The constriction forms the narrowed con- nection of the inferior sacs to the anterior lobe. As previously stated, the growth of the furrows separating the inferior sacs from the lateral sides of the anterior lobe is well marked at this time. A sagittal section of the hypophysis of a 50 mm. embryo shows a short interhypophyseal canal (fig. 9). In a 95 mm. embryo the canal has lengthened considerably. The walls are composed of one or two layers of low columnar cells. In the pup the canal is longer than in the embryos but the diameter is about the same as In younger specimens. In: the adult the canal has increased - in length and diameter and is attached in the floor of the anterior TABLE 4 Showing the size of the interhypophyseal canal SIZE OF SPECIMEN ane oo, ec Pathe este CANAL Embryo of 50:mm....0...5..... . 0.13 0.042 OD sma swisee es ote: 0.36 0.048 Teta kee AE See 0.44 bs 0.048 BROUMRS S ccs MAE Boch atts, Gt, os 0.68 0.080 lobe at or near its caudal end. Its other attachment is near the dorsal side of the connection between the inferior lobes. There are no tubular outgrowths from it such as Haller found in Mus- telus. Table 4 shows the size of the canals at different stages. It will be seen that there is a continual increase in the length of the canal. The diameter in the older specimen is somewhat greater than in the 50 mm. embryo, although there is no great change. A distinct lumen is present. In the median line there is a well defined layer of connective tissue extending from the tip of the parachordal plate (fig. 9) forward to the floor of the anterior lobe of the hypophysis. This layer develops rapidly after the canal is formed and surrounds it. It is well defined in a 50 mm. embryo, but becomes thicker in the adult. This layer separates the inferior lobes from the rest of the hypophysis and JOURNAL OF MORPHOLOGY, VOL. 26, No. 3 440 E. A. BAUMGARTNER makes their dissection difficult. The cells and fibers in it are arranged concentrically around the canal. In the adult, the layer of connective tissue is still prominent and extends the entire length of the canal. Fig. 40 Transverse section of the hypophyseal stalk of a 28 mm. embryo. (H.E.C. Series 1357). X 350. Fig. 41 Transverse section of the hypophyseal stalk of a 33 mm. embryo. (H:E.C. Series 307). X< 350. ” 4. Development of the hypophyseal stalk The stalk connecting the hypophysis with the mouth is formed in 22 to 24 mm. embryos. It is present in one 21 mm. embryo which shows the anterior end constricting from the mouth. In these stages it is oval in cross-section, the lumen is very large, and its walls are formed of a layer of low columnar cells, the nuclei of which are somewhat elongated and contain considerable chromatin. In the posterior (superior) margin are found many yellow pigment granules, as has been described by Hoffmann (86). These granules are found within the cell, sometimes apparently in the nucleus (fig. 32). As Hoffmann stated, they are first seen in the bucco-pharyngeal membrane, but later occur also in the wall of the stalk. Miller (71) had described them in the stalk in Acanthias embryos of 30 mm. length. In 28 mm. embryos the lumen in the stalk is small. The wall consists of a double layer of epithelial cells, the outer of which DEVELOPMENT OF THE HYPOPHYSIS 441 Fig. 42 Transverse section of the hypophyseal stalk of a 37 mm. prabeya (H.E.C. Series 363). X 350. Fig. 43 Transverse section of the hypophyseal stalk of a 41.5 mm. embryo. (H.E.C. Series 369). xX 350. is low columnar in type (fig. 40); the inner is irregular and the nuclei are oval in outline but not regularly placed. Figure 40 is of a section through approximately the middle of the stalk. The lumen and stalk are larger where it connects with the hy- pophysis and with the mouth, both ends being funnel-shaped and joined by a narrower middle part. The connective tissue around the stalk is mesenchymal in character. Immediately around the stalk the cells are concentrically arranged. This character is more pronounced in later stages. In a 33 mm. embryo the two ends of the stalk are funnel- shaped, as before, and contain a lumen, while the middle part of the stalk is made up of a mass of cells in which no lumen is present (fig. 41). The outer layer of cells, though still definite at the ends, is no longer so in the middle part. The nuclei are more spherical in shape and contain a denser chromatin network. The arrangement of the mesenchymal cells around the stalk is concentric. In a 37 mm. embryo the stalk is greatly reduced in size. The nuclei are massed in the center and are surrounded by a densely staining cytoplasm (fig. 42). Some pigment is scattered through- out the stalk, as in all of the specimens. The concentric arrange- ment of connective tissue cells is more marked. In some embryos 40 mm. in length no remnant of the stalk is found. In a 41.5 mm. embryo a small elongated densely stain- 442 E. A. BAUMGARTNER ing stalk is present (fig. 43). What apparently is the remains of the nuclear mass is surrounded by a narrow densely-staining cytoplasmic rim. The chromatin network has disappeared, but larger masses of densely staining chromatin are to be seen. The concentric arrangement of the connective tissue cells is apparent, but not so marked as in the younger stages. In another specimen, 40 mm. in length, a strand of connective tissue, extend- ing from a funnel-shaped mass of epithelial cells—continuous with the epithelial lining of the mouth—to the base of the hy- pophysis, indicates the position of the degenerated stalk (fig. 8). The hypophyseal attachment of the stalk is anterior to the groove connecting the inferior lobes. Soon after this time the cartilages at the base of the brain become continuous across the mid line. I wish to thank Dr. R. E. Scammon for his many helpful suggestions throughout this work. Thanks are also due to Dr. R. J. Terry for his kindly interest during its completion. SUMMARY 1. The terms ‘anterior lobe,’ ‘inferior lobes’ and ‘superior lobe’ have been used for the several parts of the hypophysis of Acanthias. 2. Rathke’s pouch forms the posterior part of the anterior lobe. The later evagination of the ectoderm, anterior to this, forms the middle portion and the anterior extremity of the an- terior lobe. 3. The inferior lobes develop from the lateral sides of the pos- terior extremity of the anterior lobe, i.e., from the lateral sides of Rathke’s pouch. 4. The superior lobe develops from the caudal (superior) end of the hypophyseal anlage. 5. In the course of development the hypophysis shifts in posi- tion about 145 degrees, so that the upper wall becomes the floor and the ventral (anterior) surface the roof. DEVELOPMENT OF THE HYPOPHYSIS 443 6. There is glandular growth from the roof of the superior lobe and the inferior lobes, as has been described by Sterzi and others, and from the floor of only the anterior and posterior extremity of the anterior lobe in all adults. 7. The cells of both anterior and inferior lobes are acidophilic in character. 8. The cell columns of the superior lobe are solid as Sterzi described them. 9. Frequently the anterior and inferior lobes stain more densely than does the superior lobe. In these cases, it is the nuclei which take the darker stain. In general the anterior and inferior lobes may be considered the chromophilic ones. 10. Spaces containing some colloid-like secretion are present in the superior lobe. A similar secretion is present in the lumina of the tubules of the anterior and inferior lobes and in the large main lumen. 444 E A. BAUMGARTNER ADDENDUM After the completion of the present work a paper by M. W. Woerdeman: ‘‘Vergleichende Ontogenie der Hypophysis”’ ap- peared (Arch. f. mikr. Anat., Bd. 86). This investigator figured Rathke’s pouch in an 8 mm. Torpedo embryo. In somewhat older embryos (12-15 mm.) the region where Rathke’s pouch opens into the mouth evaginates and in still later embryos a region anterior to this is constricted from the mouth. The hypophysis then consists of a small Rathke’s pouch somewhat constricted from an anterior (ventral) ‘Mittelraum’ and anterior to the latter, the ‘Vorraum.’ The middle division, in 20 mm. embryos, divides by a circular constriction into a dorsal and a ventral part. In this way the ventral sacs are formed. The hypophyseal stalk now opens into the ventral sacs, in which observation Woerdeman agrees with that of Gentes and Sten- dell. According to Woerdemann’s comparison of the parts of the hypophysis with those described by Stendell, Rathke’s ~ pouch is homologous with the superior lobe (table 1, p. 400) and the ‘Mittelraum’ and ‘Vorraum’ are homologous with the anterior lobe. The ventral sacs and lateral lobuli which he described are probably homologous with the inferior lobes. In Squalus I have described Rathke’s pouch, or the early anlage of the hypophysis as giving rise to the caudal extremity of the anterior lobe. A later evagination ventral to this gives rise to the middle portion and anterior extremity of the anterior lobe. The secondary evagination is early recognized, as the epithelium here is thickened (page 408). The opening from the mouth to the early hypophyseal anlage, or Rathke’s pouch, secondarily comes to open in the later evagination (page 410) of which there is only one in Squalus. The inferior lobes and the superior lobe develop from the early hypophyseal anlage in Squalus, as has been described (pp. 410-11). The hypophyseal stalk is not constricted from the anterior lobe with the developing ventral lobes but remains connected with the caudal wall of the middle portion of the anterior lobe until it disappears. DEVELOPMENT OF THE HYPOPHYSIS 445 BIBLIOGRAPHY Argsu, M. 1914 L’ipofisi in Chimaera monstrosa L. Anat. Anz. Bd. 47. Bautrour, F.M. 1874 A preliminary account of the development of the elasmo- branch fishes. Quart. Jour. Micr. Sc., vol. 14. 1878 A monograph on the development of elasmobranch fishes. London. Burckuarpt, R. 1907 Das Zentral-nervensystem der Selachier als Grund- lage fiir eine Phylogenie des Vertebratenhirns. I Teil: Einleitung und Scymnus lichia. Abh. d. Kais. Leop.—Carol. Deutschen Akad. d. Naturf., Bd. 73. 1911 Das Zentral-nervensystem der Selachier. II Teil. Die iibriger Palaioselachier. Nova Acta. Abh. des Kais. Leop.—Carol. Deutschen Akad. d. Naturf., Bd. 94. Curarual, G. 1898 Di un organo epitheliale situato al dinanzi della ipofisi e di altri punti relativi allo sviluppo della regione ipofisaria in embrioni di Torpedo ocellata. Monit. Zool. Ital., vol. 9. EpineerR, L. 1888 Untersuchungen itiber die vergleichende Anatomie des Gehirns. I Das Vordenhirn. Frankfort a. M. 1892 Untersuchungen iiber die vergleichende Anatomie des Gehirns. II Das Zwischenhirn. Frankfort a M.; also, Anat. Anz., Bd. 7. GENTES, L. 1906 Signification choréidienne du sac vasculaire. Compt. rend. Soc. Biol., T. 60; also Réun. Biol., Bordeaux. 1907 Récherchés sur ’hypophyse et la sac vasculaire des vertébrés. Travaux des Laborat. Bordeaux. Soc. Sci. d’Areachon Sta. biol:, Bia 1908 a Développement comparée de la glande infundibulaire et des plexus choroides dorsaux chez la torpille. Compt. rend. Soc. Biol. Paris, T. 64. 1908 b Sur le développement des lobes inférieurs chez les Selaciens. Compt. rend. Soc. Biol. Paris, T. 64. 1908 ec Les lobes lateraux de Vhypophyse de Torpedo marmorata Risso; développement du sac inférieur de cette hypophyse. Compt. rend. Soc. Biol. Paris, T. 64. 1908 d Développement et évolution de l’hypencephale et de l’hypo- physe de Torpedo marmorata Risso. Trav. Soc. Sci. Arcachon, StseAoOlemele LIE Hauer, B. 1896 Untersuchungen iiber die Hypophysis und die Infundibular- organe. Morph. Jahrb., Bd. 25. Herring, P. T. 1908 A contribution to the comparative physiology of the pituitary body. Quart. Jour. Expt. Phys., vol. 1. Horrmann, C. K. 1896 Beitrige zur Entwicklungsgeschichte der Selachii. Morph. Jahrb., Bd. 24. Jounston, J. B. 1909 The morphology of the forebrain vesicle in vertebrates. Jour. Comp. Neur., vol. 19, no. 5. Joris, H. 1908 Le lobe postérieur de la glande pituitaire. Mem. d. l’Acad. de Méd. d. Belgique, T. 19. 1909 Le glande neuro-hypophysaire. Compt. rend. d. |’Assoc. des Anat., 11 Réun. ® 446 E. A. BAUMGARTNER von Kuprrer, C. 1894 Die Deutung des Hirnanhangs. Sitzungsber. d. Ges. - f{. Morph. u. Phys. in Miinchen. von Mrxuucuo-Mactay, N. 1868 Beitrag zur vergleichenden Anatomie des Gehirnes. Jena. Zeitsch., Bd. 4. 1870 Beitrige zur vergleichenden Neurologie der Wirbethiere. Das Gehirn der Selachier. Leipzig. Miuter, W. 1871 Ueber Entwickelung und Bau der Hypophysis und der Processus infundibuli cerebri. Jena. Zeitsch. f. Med. u. Naturwiss., Bd. 6. Nisuikawa, T. 1899 Notes on some embryos of Chlamydoselachus anguineus Garm. Annotationes zool. japonenses, vol. 2. Pettit, A. 1906 Sur Vhypophyse de Centroscymnus caelolepis Boc. et. Cap. Compt. Rend. de la Soc. de Biol., T. 61. Raspu-RiickHarD, H. 1880 Das gegenseitige Verhiltnis der Chorda, Hypophysis und des mittleren Schidelbalkens bei Haifisch embryonen, nebst Bemerkungen iiber die Deutung der einzelnen Theile des Fischgehirns. Morph. Jahrb., Bd. 6. ; ReEIcHERT, H. 1877 Ueber das vordere Ende der Chorda dorsalis bei friihzeiti- gen Haifisch Embryonen (Acanthias vulgaris). Abh. d. kéngl. Akad. Wiss. Berlin. Rouon, J. V. 1879 Das Zentralorgane des Nervensystems der Selachier. Denkschr. k. Akad. Wiss. Wien., math. naturw. K1. Rossi, U. 1902 Soprailobilateralidellaipofisi. Arch. Ital. Anat. Embr., vol. 1. SanpgErs, A. 1886 Contribution to the anatomy of the central nervous system in vertebrate animals. Phil. Trans. Roy. Soc. London, vol. 177. Sassp, H. F.A. 1886 Bydrage tot de kennis von de ontwikkelung en beteekenis der Hypophysis cerebri. Diss. Utrecht. Scammon, R. E. 1911 Normal plate of the development of Squalus acanthias. Normt. z. Entwicklungsgeschichte d. Wirbeltiere, H. 12. Sepewick, A. 1892 Notes on elasmobranch development. Quart. Jour. Micr. Soc., vol. 33. Sewertzorr, A. N. 1899 Die Entwicklung des Selachier Schidels. Festsch. zum 70. Geburtstag von. C. v. Kupffer, Jena. STENDELL, W. 1913 Zur vergleichenden Anatomie und Histologie der Hypo- physis cerebri. Arch. f. mikr. Anat., Bd. 82.. Sterzi, G. 1904 Intorno alla struttura dell’ ipofisi nei Vertebrati. Atti dell’ Acad. sc. veneto- trentino- istriano, vol. 1. 1909 Il sistema nervosa centrale dei Vertebrati, vol. 2, libro 1, P. 1; Padova. 1912 Il sistema nervosa centrale dei Vertebrati, vol. 2, libro 1, P. 2; Padova. Titney, F. 1911 Contribution to the study of the hypophysis cerebri with especial reference to its comparative histology. Memoirs of The Wistar Institute of Anatomy and Biology, no. 2. Viautt, F. 1876 Recherches histologiques sur la structure des centres nerveaux des Plagiostomes. Arch. d. Zool. Expérimentale, T. 5. ZinGuER, H. E. 1908 Ein Embryo von Chlamydoselachus anguineus (Garm). Anat. Anz., Bd. 33. THE ANATOMY OF HETERODONTUS FRANCISCI II. THE ENDOSKELETON! J. FRANK DANIEL From the Department of Zoology, University of California THIRTY-ONE FIGURES (EIGHT PLATES) CONTENTS MEET CHIOTMROAT RINE Ae one eM eh ae ark me ts Oa M5 5 ale eT ee 447 Deelvevendoskeletorew «csr kc ts wens seed fia ols cle) oidDov wae Se Ae een 448 Ite ASHE SIR IE HETOR GA Ieee. Rotate etal ghar i en ae oe ai Fane wil ike 449 Meape TY pes coll Penne Mn EMSC ee ais a a Gace cad oes Re Gee 449 Pea CRO ESATO NSS ME MISE. iS Src AMS wale. cavcid walle Same UC eee 449 Dalhommecendleskcletoliccg a man < Jadels 2502s Gl 4 2a oe Dee. 455 ee es ss Pe Me CP NUUMINTT hee? ee A ah on aa. 20 0.0 03 oo Ube ANCE 466 We SEHG re Ulam BRCLEGUGID £850 0A 5. 2 Gin deed eae oe ve le Hd a ane ea ee 470 fhe seclebouret tect girdles Pk. ces el SO eee 470 Pa RGURCCUOTHE -PAPEIOCK s CORNER, Load roi e i las « wilde ghee DA 470 pret news CIECUC dn tidang . fart Gut AE Piss

. oss eee eee 5 vols. (4) Altmann’s solution Bichromate. of potash, 5 penmwenteh.., 02 0%... 3... a eee eae 1 vol. Osmic) acid, 2 percent: Vaase epee ee Cs. J oe 1 vol. (5) Meves’ solution Osmie acid, 2 per cent. ve see nei - toes: ss) ee ee ee 100 ce. Chromic aerd. «ici! <0 1a OS «foe so ei ee ee eee O.oiee IN Cres hehehe soe ak pce ee roe ic titne i. ott eae ee 1 gram. STUDIES ON GERM CELLS 497 The general structure of the ovary and ovarioles was obtained from in toto preparations and from thick sections. Detailed studies were made of sections 4u thick and stained in the follow- ing ways: (1) Heidenhain’s iron hematoxylin; (2) Rubaschkin’s mitochondrial method;? (3) Altmann’s acid fuchsin; (4) Benda’s crystal violet; and (5) Delafield’s hematoxylin. These methods of fixation and of staining were selected in order that both cytoplasmic and nuclear bodies could be studied. The ovaries of insects consist of a number of tubes, the ovari- oles, which are attached at the anterior end by means of ter- minal threads and open at the posterior end into the oviduct. The variations in the structures of the ovarioles are due princi- pally to the presence or absence of nurse cells and the distribution of these when present. Certain ovaries, for example, those of the Orthoptera and Aptera, are not provided with nurse cells. In others the nurse cells may remain within a terminal chamber and supply the growing oocytes through a nutritive strand, as in certain Hemiptera and Coleoptera; or a rather definite number of nurse cells may become separated from the terminal chamber and accompany each oocyte, as in the Neuroptera, Hymenoptera, Diptera, and Lepidoptera. The result of the last named method of nutrition is the formation of ovarioles which resemble rows of beads. In the bee each oocyte is accompanied by a group of nurse cells. An outline of a single‘ovariole is shown in figure 1. The terminal filament (¢) consists of a row of long slender cells which extend entirely across the filament. Following this is a region occupying about half of the entire ovariole which is character- ized by rosettes of cells (r). These cells are apparently all alike * Rubaschkin fixes tissues in Meves’ modification of Flemming’s solution for one or two days. Sections are treated as follows: Potassium permanganate, } per cent..............2.eeeeees 1 minute Wash in water . Oxalic acid and potassium sulphate, } per cent............. 1 minute Peps) 10k TUT Ort See Oe arse eon a'e + oe vie = aes eels 15 minutes Beeric aluthy AGnermcemte y Soap eis ties sc « «oo «Lae ee 24 hours Wievsert a Memiapomplin .. oom. kee. slepss..s 6. as. 50a s age eee 2-3 days Differentiate in ferric alum, 2 per cent 498 ROBERT W. HEGNER and those in a single rosette have descended from a single mother cell which may be called the oogonial mother cell. The actual differentiation of the oocytes and nurse cells occurs in a much shorter part of the ovariole (d). After the oocytes are definitely established, they move down the ovariole, become arranged in a single row (0) and are gradually separated from each other by groups of nurse cells (n) which le in nurse chambers just above them. The elements within ovarioles of different ages differ, of course, both in their stage of development ‘and in their distribution. The three kinds of cells within the ovaries of insects are the oocytes, nurse cells, and epithelial cells. These three kinds of cells arise differently in different groups of insects. Thus the nurse cells and epithelial cells in the paedogenetic fly, Miastor (Kahle ’08; Hegner 714 a) are of mesodermal origin and the germ cells give rise only to oocytes. In the Hymenopteraon the other hand Korschelt (’86) in Bombus, Paulcke (’01) in Apis, and Marshall (’07) in Polistes agree that the three cellular ele- ments within the ovaries arise from one sort of cells, the germ cells. I have been unable to determine the origin of the epithe- lial cells in the ovarioles of the bee because of the lack of young ovaries, but that nurse cells and oocytes arise from oogonia there can be no doubt. Part of the rosette region of an ovariole is shown in figure 2. Two kinds of cells are present, (1) Those that make up the rosettes (r); and (2) the epithelial cells (e) among the rosettes. The ground substance within the ovariole in this region appears to be a loose cytoplasmic reticulum containing a few scattered nuclei. These nuclei are rather irregular in shape, and contain a clear matrix in which may be seen one or two large chromatin masses and a very delicate reticulum. No epithelial cell bound- aries could be observed in this part of the ovariole and it seems probable that the rosettes are imbedded in a syncytium. There seems to be no regular arrangement of the rosettes; they do not crowd one another, but the cells in each are closely united, hence it is a very simple matter to distinguish the separate rosettes in an ovariole even with low magnification. It seems strange STUDIES ON GERM CELLS 499 because of this perfect distinctness that Paulcke (’01) failed to observe these rosettes. The evidence for the statement that all of the cells in a single rosette have descended from a single mother cell is irrefutable. In figure 2 the cells of the rosette to which the guide line (r) extends are grouped about a branching strand which stains black in iron hematoxylin. A similar rosette is shown enlarged in figure 3. One branch of the black strand extends into the cytoplasm of each cell. These strands consist of the spindle fibers remaining after previous mitotic divisions, and, as will be pointed out later, such strands are not uncommon in either the ovaries or the testes of insects. A section through one end of a rosette at right angles to that shown in figure 3 is illustrated in figure 4. The spindle remains form a sort of axis about which the strands from the most recent divisions are radially arranged. The entire rosette is therefore oblong and may be sectioned longitudinally or transversely. The number of cells in each of the rosettes figured is sixteen, indicating that four divisions had occurred since the oogonial mother cell was established. No evidence was obtained which indicated the presence of amitotis in these ovarioles, and very few mitotic division figures were observed. Those that were found were invariably restricted to the cells in single rosettes (fig. 5), thus indicating that the cells in a rosette divide synchronously. A ¢ritical examination of both the cytoplasm and nuclei of the cells in the rosettes failed to reveal any constant differences among the cells of any particular rosette. Giardina (’01) dis- covered a difference in the nuclei of certain rosette cells in the ovarioles of Dytiscus, and Kern (712) has reported a difference in the cytoplasm of similar cells in Carabus, but no such dis- tinguishing marks were found in the bee. This indicates that all of the cells at this stage in the oogenesis of the bee are prob- ably potentially alike. At any rate no visible differences were discovered in material fixed and stained so as to bring out to the best possible advantage both nuclear and cytoplasmic bodies. The rosette zone in the ovariole is followed by the zone of differentiation (fig. 1, d). Certain of the cells increase in size 500 ROBERT W. HEGNER and are recognizable as oocytes (fig. 6, 0). This is brought about by an increase in the amount of cytoplasm and by the enlarge- ment of the nucleus. The arrangement of the chromatin within the nucleus changes during this differentiation; that of the nurse cells (fig. 6, n) retains the condition characteristic of the rosette stage (fig. 2), whereas in the newly formed oocytes the chromatin forms threads which are scattered about irregularly within the nucleus (fig. 6, 0). The connecting strands, so noticeable in the rosettes (fig. 3), either disappear at this time or lose their stain- ing capacity since they are apparently absent from this stage on. Nevertheless it is very easy to determine which cells have de- scended from a single mother cell since a dark double ring re- mains where the strands passed from one cell to another (fig. 7). These rings are quite conspicuous but were completely over- looked by Paulcke (’01). The change from the rosette zone to the zone of differentenaen _in the ovariole of the bee is an abrupt one—a fact which makes a study of the differentiation of the oocyte difficult, since no intermediate stages can be studied unless material in just the proper condition is obtained. In several cases which will be described later, investigators have found that a single rosette gives rise to one oocyte and a group of nurse cells. This is cer- tainly not true in the bee, since the oocytes in the zone of differ- entiation are much too numerous, compared with the number of rosettes, and many instances were observed of two or ‘more oocytes which had been directly connected by. spindle remains as indicated by the presence of double rings between them (fig. 7). If all of the cells in a single rosette are potentially alike the question arises, what causes some of the cells to become oocytes and others nurse cells? Three explanations have occurred to me: (1) There may be differential changes during the mitotic divisions in rosette formation as in Dytiscus (Giardina ’01) resulting in one or more cells (oocytes) which differ in content from the others (nurse cells). No Visible changes of this sort were observed. (2) The polarity of the rosette may influence the cells in such a way that those near the center of the ovariole and closest to the zone of differentiation tend to develop into STUDIES ON GERM CELLS 501 oocytes. (3) Those cells of the rosette which reach the zone of differentiation first are stimulated to become oocytes and by their growth and differentiation prevent the other cells of the rosette from similar changes. It would be futile to argue on the basis of known facts in favor of any of these hypotheses. The arrangement of the oocytes and nurse cells within the ovariole resulting in a linear series of oocytes which alternate with groups of nurse cells takes place a short distance back of the zone of differentiation (fig. 1,). Paulcke (’01) has satisfactor- ily described and figured the formation of the epithelium around the oocytes and the structure of the nurse chamber, but, as stated above, he failed to see the intercellular rings which indicate the descent of the cells concerned. A group of nurse cells about an oocyte is shown in figure 8. This oocyte is connected with at least three nurse cells. One of the nurse cells (a) lies below the oocyte in the ovariole; since this is never true in later stages it is probable that such a cell would either degenerate or become separated from the oocyte and forced over to one side. This has evidently happened in the case of the oocyte illustrated in figure 9, since a ring is present here at the lower end (a), but it does not connect the oocyte with a nurse cell. The relation between the oocyte and its accompanying nurse cells is shown in figure 10. All of the nurse cells are not included, since this is a camera drawing of a section. It illustrates, however, the way in which the nurse cells form into rows converging toward the oocyte. The descent of the cells within the zone of differentiation would -be impossible to determine if it were not for the presence of the rings between them. These rings continue to connect the nurse cells with the oocyte, even in late stages in the growth of the latter (fig. 11) and many of them may also persist between the nurse cells after the nurse chamber is fully formed, as in the stage illustrated in figure 12. Kern (’12) also finds these rings con- necting the nurse cells with the oocytes of Carabus, and claims that nutritive material passes through them during the growth of the egg. As soon as the oocytes are differentiated, numerous granules of various sizes appear within their cytoplasm; in the earlier 502 ROBERT W. HEGNER stages these lie mostly near the nucleus (figs. 7-8), but later (fig. 9) become scattered throughout the cytoplasm. These granules stain best in iron hematoxylin after fixation in Meves’ modi‘ cation of Flemming’s solution. No evidence was obtained that they are of nuclear origin, although their early position near the nucleus indicates that they may have arisen in this way; or if not directly from the nucleus, at least through its influence. On the other hand, their sudden appearance within the cytoplasm indicates that they are cytoplasmic bodies which have resulted either from the aggregation of smaller pre-existing bodies of a similar nature or from the synthesis of other substances under the stimulus of the metabolic processes set up at the inauguration of the growth period. Duesburg (08) has recognized granules in the peripheral layer of cytoplasm in the full grown egg of the bee, especially near the nucleus in the thickened area which Petrunkewitsch (’01) has called the ‘Richtungsplasma,’ and considers them to be mitochondrial in nature. It seems probable that the bodies we have observed are the ‘mitochondria’ of Duesberg at an earlier stage. Paulcke (’01) failed to observe them. Discussion. The differentiation of the cellular elements in the ovaries of insects and the relations of the oocytes to the nurse cells has interested students of histology and cytology for three quarters of a century. Mayer, as early as 1849, ex- pressed the opinion that the nurse cells are abortive eggs. The connections between them and the oocytes were observed by Huxley (’58) in oviparous aphids, and were considered by him a nutritive canal for the conduction of food material from the nurse cells to the growing ege—a conclusion concurred in by Lubbock (60) and Claus (’64). Balbiani (’70), however, proved this ‘nutritive canal’ to be a protoplasmic strand, but, as Wielowiejski (85) has pointed out, he was in error when he stated that the terminal chambers of the ovarioles of aphids contain a large central cell which gives rise to both the oocytes and nurse cells (abortive eggs). He nevertheless established the fact of a pro-_ toplasmic cellular bridge between these two kinds of cells. STUDIES ON GERM CELLS 503 Protoplasmic bridges between the cells of Metazoa are not uncommon and may exist in all tissues. As a rule, they are delicate strands which pass through pores in the cell walls. The cellular elements in a syncytium, such as occurs during the cleavage of the insect egg, must be even more closely united physiologically, since here the cytoplasm forms a continuous network. Cellular bridges similar to those described above in the queen bee, have been observed in the germ glands of a number of other animals, especially insects,, but mostly dur- ing spermatogenesis. Thus Platner (’86) found in Lepidoptera that often two neighboring spermatocytes, and sometimes three, were connected by intercellular ligaments which were attached to an intracytoplasmic body in each cell. The latter were considered ‘Nebenkerne.’ Similar conditions were dis- covered by Prenant (’88), Zimmerman (’91) and Lee (’95) in the male germ cells of Gastropoda. Lee, in his work on Helix, recognized the true origin of the intercellular bridges and their significance. They were found to be the remains of the spindle fibers following a mitotic division. The term ‘pont fusorial’ was applied by Lee to the bridge itself and ‘moignons fusoriaux’ to the ramification of the fibers within the cytoplasm of the cells. Similar intercellular ligaments were observed by Henneguy (’96) in the seminal cells of Caloptenus; by Erlanger (96, ’97) in both the testes and ovaries of the earthworm; by Wagner (’96) in the male germ cells of spiders; by Meves (’97) in both the testes and ovaries of the salamander; by Giardina (01), Debaisieux (’09) and Giinthert (’10) in the ovaries of Dytis- cus; by Marshall (’07) in the ovarioles of Polistes; by Kern (’12) in the ovarioles of Carabus; by Govaerts (’13) in the ovarioles of Carabus and Cicindela; by Maziarski (’13) in the ovarioles of Vespa; and by Hegner (14 a) in the testes of Leptinotarsa. By far the most interesting results are those obtained by Giardina and confirmed by Debaisieux (’09) and Giinthert (710). Giardina proved conclusively that a single oogonium in the ovary of Dytiscus undergoes four divisions, thus producing sixteen cells, one of which is the oocyte and the remaining fifteen nurse cells. The processes of differentiation in this genus are partic- 504 ROBERT W. HEGNER ularly interesting, because they include a separation of the chro- matin of the mother cell into two masses. One of these masses of chromatin forms an ‘anello cromatico;’ the other gives rise to forty chromosomes which divide equally, half of each passing to each daughter cell. The chromatic ring remains undivided and becomes situated entirely in one of the daughter cells. At each of the three succeeding divisions the chromatic ring is segregated entirely in one cell; this cell is the oocyte, whereas the other fifteen which have a common origin with it are nurse cells. Since the publication of Giardina’s observations many ‘investi- gators have attempted to discover similar visible differentiations in the ovaries of other insects, but without much success. Thus Govaerts (’13) made detailed studies of beetles of the genera Carabus, Cicindela, and Trichisoma but was unable to find anything resembling the chromatic ring which occurs in Dytiscus. He found however that the spindle fibers (‘residu fusorial’) persist after the daughter cells are formed during the differential divisions, just as they do in Dytiscus, and that a definite polarity is marked by the position of these spindle remains. The con- clusion is reached that something more fundamental than the unequal division of chromatic elements is responsible for the differential divisions and decided in favor of a ‘polarite pre- differentielle.’ No explanation is offered, however, as to the origin of this polarization. A brief account of the oogenesis in carabid beetles has also been published by Kern (’12), who finds that during the differ- ential mitoses, the oocyte mother-cell may be distinguished by the presence of certain intracytoplasmic granules which he describes as follows: Befinden sich die Zellen der Zellrosetten in Teilung, so findet man mitunter in einer Zelle neben der Teilungsfigur eine Anhaéufung von fiirbbaren Koérnchen, dhnlich denjenigen, die in spateren Stadien in der jungen Eizelle im Cytoplasma gefunden werden. Es liegt nahe, an einen Diminutionsvorgang, ibnlich demjenigen, welchen Giardina bei Dytiscus beschrieben hat, oder auch an einen Vergleich mit den Ectosomen bei Cyclops zu denken; doch gelang es mir bisher nicht, alle Einzelheiten festzustellen. Die K6rnchen im Cytoplasma junger Hizellen werden nach und nach aufgelost. ' STUDIES ON GERM CELLS 505 The origin of these granules was not determined, and although Kern is inclined to consider them similar to the chromatic-ring substance in Dytiscus, there is a possibility that they may be mitochondrial in nature or may consist of some other cytoplasmic material. The presence of intercellular bridges is important, since it makes it possible to determine the relationship of the groups of cells in the ovarioles. But in the queen bee these bridges do not persist to any considerable extent after the zone of differentiation has been reached. Here, however, as shown in figures 7 to 12, there are well defined rings between the cells which indicate their relationship. It might be argued that these rings may arise where two cells happen to come into contact, if it were not for the fact that all stages between the fully developed bridges and the presence of clearly defined rings have been observed. These are no doubt the persisting mid-bodies or ‘Zwischenkorper’ which remain between the cells after division. They have been noted especially by Giardina (’01) in Dytiscus; by Marshall (07) in Polistes; by Kern (12) in Carabus; and by Maziarski (13) in Vespa. Summary of Part I. 1. Four rather definite regions may be recognized in the ovariole of the queen honey bee (fig. 1): (a) the terminal filament; (b) a rosette region; (c) a zone of differ- entiation; and (d) the posterior part in which the oocytes are arranged in a linear series and separated from each other by groups of nurse cells. _ 2. The rosette region is filled with rosette-like groups of cells, each group consisting of the descendants of a single mother oogonium. The cells of a rosette are united by strands which are the persisting spindle fibers from earlier mitoses (fig. 3). The cells in a rosette divide synchronously (fig. 5). 3. Oocytes and nurse cells are both derived from the oogonia. Their differentiation occurs in the zone of differentiation (fig. 1, d). One or more cells of each rosette enlarges and becomes an oocyte, whereas the others retain more of their earlier characteristics and become nurse cells. Although the strands which connected the cells in a rosette disappear, the descendants of a single oogonium 506 ROBERT W. HEGNER may still be determined, because of the presence of deeply stain- ing rings between the cells (figs. 7-12). 4. The causes of differentiation could not be definitely de- termined, but several hypotheses are mentioned (p. 500). 5. Granules appear near the nucleus of oocytes shortly after their differentiation. Later they become distributed through- out the egg cytoplasm. These granules appear to be mito- chondrial in nature and to arise from, or under the influence of the nucleus. II. THE BACTERIA-LIKE RODS AND SECONDARY NUCLEI IN THE OOCYTES OF CAMPONOTUS HERCULEANUS VAR. PENNSYLVANICA DEG. The important contributions by Blochmann (’84, ’86) upon the growth of the oocytes in ants seem to be the only reports that have ever been made on this subject. Blochmann dis- covered two very interesting facts regarding these oocytes: (1) the presence of rod-shaped bodies almost completely: filling the growing egg which he considered symbiotic bacteria, and (2) the formation of nuclear-like bodies around the oocyte nucleus. Recently Tanquary (’13) has described, in the freshly laid eggs of the carpenter ant, a body which he calls a cleavage nucleus, but which resembles very closely bodies that have been dis- covered in the eggs of other animals and to which I have applied the term keimbahn- or germ-line determinants. The obser- vations recorded in the following pages were made in order to trace the genesis of the eggs of ants with special reference to the origin, distribution, and fate of the bacteria-like bodies, nuclear- like bodies, and the germ-line determinants. The material used for these studies consisted of the ovaries of the carpenter ant, Camponotus herculeanus var. pennsyl- vanica DeGeer. A large number of virgin queens were obtained from a dying apple tree on April 3, 1914, and some of them were kept alive until June 9, 1914. The ovaries were dissected out in Ringer’s solution and immediately fixed in the same manner as were those of the honey bee (page 496). Ovaries were pre- served at intervals of a few days during the period of two STUDIES ON GERM CELLS 507 months. In this way oocytes in all stages of growth were ob- tained up to almost the period of deposition. Sections were cut and stained as in the queen bee. The ovaries of the carpenter ant resemble those of the queen bee in general structure and the ovarioles are likewise similar. The youngest ovaries obtained had already passed the period when the oocytes and nurse cells are differentiated, so there was . ho opportunity to study the events that occur during this differ- entiation. Four regions may be distinguished in the ovarioles as shown in figure 13. There is a terminal filament (t) at the anterior end. This is followed by a region which we may call the terminal chamber (t.c.) containing oocytes, nurse cells, and epithelial cells without any special arrangement. The next part of the ovariole is short and contains oocytes which have grown considerably but have not yet taken a position in the axis of the tubule. This we may call the first zone of growth (g). The rest of the ovariole consists of a linear series of oocytes (0) each with its accompanying group of nurse cells (n). Each oocyte is larger than the one anterior to it and the nurse cells gradually become grouped into a definite nurse chamber (n.c.). The bacteria-like bodies discovered by Blochmann are present only in the last described zone. The first signs of nuclear-like bodies around the oocyte nucleus also appear here. For the sake of convenience oocytes in the various stages which need to be referred to have been drawn in outline and to scale as shown in figures 14 and 15. The posterior end of the terminal filament (¢) and anterior end of the terminal chamber (t.c.) are shown in outline in figure 16. The cells of the terminal filament are long and slender and extend entirely across it. One is shown enlarged in figure 17. Within the terminal chamber are two kinds of cells, oocytes and nurse cells. The oocytes, as indicated in figure 18, are the youngest to be found in the ovarioles at this time and I have regarded them as Stage A (fig. 14). The cell walls of the nurse cells are not very distinct. Their nuclei (fig. 19) are much smaller than those of the oocytes and contain a single irregular mass of chromatin granules. The structure of the oocytes and nurse cells is similar throughout the entire terminal chamber. 508 ROBERT W. HEGNER The terminal chamber is separated from the first zone of growth (fig. 20) by what appears to be a distinct membrane (m). The condition of all of the oocytes is similar throughout this zone (Stage B, fig. 14). The oocytes have grown considerably and their nuclei (fig. 21) contain a few clumps of chromatin granules lying near the nuclear membrane. Outside of the nucleus (fig. 21) is a layer of darkly staining substance which resembles chromatin in some respects and may represent chromatin which has passed through the nuclear membrane into the cytoplasm. The nurse cells now have definite cell walls (fig. 22) and are also characterized by a layer of darkly staining material lying around the nucleus. Among the oocytes and nurse cells are a few epithelial cells (fig. 23); these have no definite cell walls, and their nuclei are rather irregular in shape and contain a single mass of chromatin. Whether or not the first zone of growth is definitely separated from the remaining part of the ovariole could not be determined with certainty, but its limit is conspicuously marked by the abrupt appearance of the bacteria-like bodies of Blochmann. This is indicated in figure 24, which shows.the posterior portion of the first zone of growth and the anterior part of the rest of the ovari- ole. In the upper part of this figure is a single oocyte in Stage B and a number of nurse cells. These are apparently embedded in a loose reticulum of cytoplasm. Further down the ovariole the spaces surrounding the nurse cells and epithelial cell nuclei are filled with more or less wavy rods which Blochmann con- sidered symbiotic bacteria. These rods extend throughout the ovariole in all directions, being represented by distinct spherical granules where cut across. From this point on, the oocytes are arranged in a linear row in the central axis of the ovariole (figs. 13 and 25). The cyto- plasm of the oocytes increases rapidly in amount, but the nuclei enlarge very little. The nurse cells (fig. 25, n) become arranged more or less definitely into rows which radiate toward the upper end of the oocyte. Those nurse cells closest to the oocyte in- crease more rapidly in size than do the others. Compare, for example, that lettered a in figure 25 with its companions, and STUDIES ON GERM CELLS 509 those accompanying the upper oocyte with those of the lower oocyte. Surrounding the oocytes, nurse cells, and epithelial cell nuclei are the groups of bacteria-like bodies. The transition of the oocyte from Stage C (fig. 14, C; fig. 25, C.) to Stage D (fig. 14, D, fig. 26) is accompanied by an invasion of the oocyte cytoplasm by the bacteria-like rods. Some of these rods form almost perfect circles, resulting in what at first sight appear to be vacuoles. Some of the epithelial-cell nuclei are in very close contact with the oocyte but these were not observed actually within the oocyte cytoplasm. The principal difference between an oocyte in Stage D (fig. 26) and one in Stage E (fig. 14, #; fig. 27) is the sudden appearance of nuclear-like bodies around the nucleus, which I shall call secondary nuclei. The nucleus itself is about equal in size to that of the preceding stage (fig. 26). The chromatin, which in younger oocytes (figs. 24-26) has gradually migrated from the periphery toward the center of the nucleus where it formed an irregular clump, has again become scattered, being represented by a few smaller and widely separated masses. In the illustra- tion (fig. 27) three secondary nuclei are shown lying below but in contact with the oocyte nucleus. These likewise contain a delicate reticulum and from one to three chromatin masses. No intermediate stages between the nucleus of Stage D (fig. 26) and that of Stage E (fig. 27) were discovered, and it was thus impossible to determine with certainty the origin of these second- ‘ary nuclei. If, however, the oocyte nucleus continued to in- crease in size at the same rate as indicated in Stage C (fig. 25) and in Stage D (fig. 26) it would be about the size of that in figure 27 after having given rise to the secondary nuclei by the method of budding or in some other way. This subject will be discussed more in detail later. During the interval between Stage E (fig. 14, E#; fig. 27) and Stage F (fig. 14, F, fig. 28) the oocyte enlarges until it extends almost across the ovariole, and the epithelial cell nuclei become arranged in a single layer around it, forming a follicle. At this time (fig. 28) the cytoplasm of the oocyte and that surrounding the nurse cells and epithelial-cell nuclei is crowded full of the 510 ROBERT W. HEGNER bacteria-like rods. The secondary nuclei also increase in number around the oocyte nucleus; the.nucleus itself does not increase in size. Both the oocyte nucleus and the secondary nuclei are sometimes irregular in shape, a condition that may be due to the effects of fixation, or that may represent a stage in budding or in amitotic nuclear division (page 518). The next phase of the growth period (Stage G, fig. 14, G, fig. 29) witnesses the lengthening of the oocyte and the further arrange- ment of the nurse cells to form a compact group, which becomes surrounded by epithelial cells, thus producing a definite nurse chamber. The bacteria-like bodies increase in number as the oocyte grows and continue to fill it completely with bundles of rods. The secondary nuclei near the oocyte nucleus also in- crease slightly in number. Shortly after this condition is reached the oocyte is invaded just beneath the nurse chamber by an influx of cytoplasm elaborated by the nurse cells (fig. 30, c). This cytoplasm is © free from the bacteria-like bodies and it seems very probable that it either forces these rods out of its path or else dissolves those which it encounters. There is evidence that, from this stage on, the number of bacteria-like rods does not increase, the rods gradually lose their compact grouping and become further separated from one another, the spaces between them probably being occupied by the cytoplasm added to the oocyte by the nurse cells. The oocyte nucleus by this time (fig. 30) is com- pletely surrounded by secondary nuclei from which it differs in appearance. The secondary nuclei contain a rather dense reticulum and one or several large chromatin granules, whereas the oocyte nucleus is very irregular in shape and contains a delicate reticulum which causes it to appear clearer. The irregu- lar shape of the oocyte nucleus is probably due to the pressure upon it of the secondary nuclei which surround it. Its decrease in size is also noticeable and one cannot but suspect that this decrease is directly related to the increase in the number of secondary nuclei. A transverse section through an oocyte near the nurse chamber is shown in figure 31. STUDIES ON GERM CELLS 511 The nurse chamber is now completely formed (fig. 13, n.c.). The nurse cells are still free from the bacteria-like rods and their nuclei, as pointed out by Blochmann (’86), possess very thick membranes (fig. 30, ). Part of one of these nuclei greatly en- larged is shown in figure 32. The membrane contains, in a homogeneous matrix, a number of vacuoles and a great many granules of various sizes which appear in material fixed and stained by a number of different methods. Their reactions all indicate that they are chromatic in nature and their position suggests that they may have migrated into the membrane from inside of the nucleus and are on their way into the cytoplasm. It could not be definitely determined, however, whether this is a true case of chromatin emission or simply a condition due to the action of the fixing solutions used. A further increase in the amount of cytoplasm within the oocyte is evident when Stage H (fig. 15, H; fig. 33) is reached. Here an opening (a) is present in the follicle connecting the oocyte directly with the nurse chamber. The small plug of cytoplasm filling this channel is no doubt homologous with the nutritive string present in the ovarioles of insects whose oocytes are not accompanied by a group of nurse cells, but are connected with the terminal chamber by a protoplasmic thread. In this stage the oocyte nucleus (0) is still closely pressed by the secondary nuclei (s) surrounding it and the entire group lies within the cytoplasmic zone. Such a group is shown enlarged in figure 34, in which the oocyte nucleus may be distinguished from the secondary nuclei by its irregular shape, central position, and clearness. The succeeding stages in the growth of the oocyte (fig. 15, J, J, K, L; figs. 35-39) are characterized by a decrease in the number of bacteria-like rods, by the formation of yolk globules, and by the increase in number and the scattering of the secondary nuclei. Part of a section through an oocyte of Stage I (fig. 15, [) is shown in figure 35 which represents a portion extending from a point midway between the two poles out to the middle of the oocyte. Just within the follicular epithelium (e) is the suggestion of a clear layer (k) which later becomes the ‘Keimhautblastem.’ The JOURNAL OF MORPHOLOGY, VOL. 26, NO. 3 512 ROBERT W. HEGNER black spherical bodies are yolk globules (y) which appear to originate near the periphery and gradually to migrate into the central region. The bacteria-like rods are still present but they are widely scattered and faintly staining. By the time the next stage is reached (fig. 15, J; fig. 36) the bacteria-like rods have completely disappeared everywhere except near the periphery, around the lower part of the oocyte. According to Blochmann (’86) they are still present in this region after the eggs are laid, and they are also mentioned by Tanquary (13) in the freshly deposited eggs. The latest oocyte studied by the writer is Stage L (fig. 15, Z), which is considerably younger than the fully grown egg. 110. 2 Part of the rosette region of an ovariole; 7, a rosette, the cells of which are held together by deeply staining strands; e, an epithelial cell nucleus. 650. 3 1250. 10 An outline showing the arrangement of an oocyte and its accompanying nurse cells. XX 1250. 11 Part of arather old oocyte, o, still connected with nurse cells, n, by means of rings, e, epithelian cell. > 1250. 12 An outline of an older oocyte showing the rings between the nurse cells and oocyte and between neighboring nurse cells. 480. STUDIES ON GERM CELLS PLATE 2 ROBERT W. HEGNER R. W. HEGNER, del. 539 PLATE 3 EXPLANATION OF FIGURES Camponotus herculeanus var. pennsylvanica 13 Outline of an ovariole showing the terminal filament, ¢, terminal chamber, i.c., first zone of growth, g, and later growth zone containing oocytes, 0, and nurse cells, n and nc. X 170. 14 Outlines of oocytes in Stages A toG. & 110. 15 Outlines of oocytes in StagesHtoL. X 110. 16 Outline of part of the terminal filament, ¢, and terminal chamber, tc. The numbers 17, 18 and 19 refer to cells shown enlarged in figures 17, 18 and 19. x 620. 17 A single cell from the terminal filament. X 3300. 18 An oocyte from the terminal chamber. X 3300. 19 A nurse cell nucleus from the terminal chamber. > 3300. 540 STUDIES ON GERM CELLS PLATE 3 ROBERT W. HEGNER R. W. HEGNER, del. PLATE 4 EXPLANATION OF FIGURES Campanotus herculeanus var. pennsylvanica 20 Outline of the first zone of differentiation, showing the membrane, m, separating it from the terminal chamber, and the oocytes, 0, nurse cells, n, and epithelial cells, e. The numbers 21, 22 and 23 refer to cells shown enlarged in figures 21, 22 and 23. X 620. 21 An oocyte from the first zone of growth. X 3300. 22 A nurse cell from the first zone of growth. 3300. 23 An epithelial cell nucleus from the first zone of growth. X 3300. 24 The posterior portion of the first zone of growth and the anterior portion of the rest of the ovariole containing bacteria-like rods. e, epithelial cell nucleus; n, nurse cell; 0, oocyte. X 1250. 25 Part of an ovariole showing two oocytes, C; and Cs, in Stage C. mn, nurse cell. XX 1250. 26 AnoocyteinStage D. The bacteria-like rods have invaded the cytoplasm of the oocyte. X 1250. PLATE 4 HEGNER STUDIES ON GERM CELLS ROBERT W. R. W. HEGNER, del. JOURNAL OF MORPHOLOGY, VOL. 26, NO. 3 PLATE 5 EXPLANATION OF FIGURES Camponotus herculeanus var. pennsylvanica 27 An oocyte in Stage E showing three secondary nuclei, s, near the oocyte nucleus, 0. e, epithelial cell nucleus; 7, nurse cell. XX 1250. 28 An oocyte in Stage F. Lettering as in figure 27. X 620. 29 An oocyte in Stage G. Lettering as in figure 27. X 620. 30 Part of an oocyte and two nurse cells, n. Cytoplasm, c, elaborated by the nurse cells is present near the nurse chamber. 0, oocyte nucleus; s, secondary nuclei. X 480. 31 Transverse section through the anterior end of an oocyte. X 620. 32 Part of the nucleus of a nurse cell showing vacuoles and deeply staining granules in the thick nuclear membrane. X 3300. 544 STUDIES ON GERM CELLS PLATE 5 PLATE 6 EXPLANATION OF FIGURES Camponotus herculeanus var. pennsylvanica 33 Part of an oocyte in Stage H showing its connection, a, with the nurse chamber; 0, oocyte nucleus; s, secondary nuclei. X 480. 34 An oocyte nucleus surrounded by secondary nuclei from an oocyte in Stage H. X 1250. 35 Part of an oocyte in Stage I. e, follicular epithelium; k, ‘Keimhautblas- tem’; y, yolk globules. > 480. 36 Part of an oocyte in Stage J. X 430. 37. The anterior part of an oocyte showing the breaking up of the group of secondary nuclei, s. a, connection with nurse chamber. c, cytoplasm; e, fol- licular epithelium; y, yolk globules. 430. . 38 1250. 48 Outline of an oocyte in Stage C. X 1250. 49 Outline of an oocyte in Stage D. First appearance of the germ-line- determinant near the posterior end. X 1250. 50 Outline of an oocyte in Stage E. X 1250. 51 Outline of an oocyte in Stage F. Single chromosomes are present. x 1250. 52 Outline of an oocyte in Stage G. The chromosomes have united near their ends to form pairs. X 1250. 53 Outline of an oocyte in StageH. ‘The pairs of chromosomes are arranged in a parallel series. 1250. 54 Outline of an oocyte in Stage I. X 1250. ou or i=) PLATE 8 STUDIES ON GERM CELLS GNER ROEERT W HE R. W. HEGNER, del. 551 By) 56 o7 58 59. 60 61 PLATE 9 EXPLANATION OF FIGURES Copidosoma gelechiae An oocyte in Stage A (see fig. 46). > 3300. An oocyte in Stage B (see fig. 47). > 3800. An oocyte in Stage C (see fig. 48). X 3300. The anterior portion of an oocyte in Stage G (see fig. 52). X 3300. The nucleus of an oocyte in Stage H (see fig. 53). 3300. The nucleus of a slightly older oocyte. > 3300. A transverse section through a nucleus in a similar condition. X 3300. 62 to 65 Successive stages in the condensation of a spindle like that shown in figure 60. X 800. ou On bo STUDIES ON GERM CELLS PLATE 9 ROBERT W. HEGNER Bi R. W. HEGNER, del. PLATE, 10 EXPLANATION OF FIGURES 66 Ageniaspis; the anterior portion of an oocyte showing the arrangement of the chromosomes on the spindle (after Martin). 67 Ageniaspis; a later stage showing the mass of chromatin resulting from the condensation of the chromosomes (after Martin). 68 Ageniaspis; the first maturation division of the egg (after Martin). 69 Copidosoma; a young oocyte showing a group of granules near the poste- rior end of the nucleus (after Silvestri). 70 Ageniaspis; a young oocyte containing a cloud of granules in the poste- rior portion and a larger body, the ‘nucleolus’ (after Martin). 71-74 Apanteles. - 71 A young oocyte surrounded by epithelial cells, e, and accompanied by nurse cells, n. > 1900. 72 An older oocyte. X 1900. 73 An older oocyte showing the first appearance of the germ-line-determi- nant. X 1900. 74 Astill older oocyte. X 1900. 554 STUDIES ON GERM CELLS FLATE 16 ROBERT W. HEGNER R. W. HEGNER, del. PATE 11 EXPLANATION OF FIGURES Apanteles 75 An oocyte containing many secondary nuclei. X 850. 76 Part of the oocyte shown in figure 75. X 3300. 77 Anolder oocyte showing the parallel arrangement of chromosomes. 850. 78 Nucleus enlarged from figure 77. X 3300. 79 A later stage in the history of the nucleus. X 3300. 80-82 Successive stages in the history of the germ-line-determinant. X 1900. 596 PLATE 11 STUDIES ON GERM CELLS ROBERT W. HEGNER del. R. W. HEGNER, PEATE: 12 EXPLANATION OF FIGURES Andricus punctatus 83 An egg ready to be laid. X 430. 84-86 Nuclei showing stages in the condensation of the chromosomes. x 1900. 87 The chromatin mass resulting from the condensation of the chromosomes. < 1900. 508 STUDIES ON GERM CELLS ROBERT W. HEGNER PLATE 12 R. W. HEGNER, del. PLATE 18 EXPLANATION OF FIGURES 88-93 Diastrophus nebulosus. 88 An egg ready to be laid. X 430. 89-90 Longitudinal sections through the nucleus of such an egg. X 1900. 91 Transverse section through a nucleus in the same stage. 1900. 92 The germ-line-determinant near the posterior end. X 1900. 93 The mitotic division of a follicular epithelial cell. X 1900. 94-97 Rhodites ignota. 94 An egg ready to be laid. X 430. 95 Part of an oocyte showing stages in the formation of secondary nuclei. 1900. 96 An older oocyte showing the arrangement of secondary nuclei. X 620. 97 Two secondary nuclei much enlarged. X 2500. 560 STUDIES ON GERM CELLS PLATE 13 ROBERT W. HEGNER R. W. HEGNER del. 561 i s ’ 1 ‘ ; ; ral P \ - 0 x é ~ \ % . i { , \* A . hay ‘ . ' : a ( 1% ' ( = a THE HOMOLOGIES OF THE HYOMANDIBULA OF THE GNATHOSTOME FISHES EDWARD PHELPS ALLIS, JR. Menton, France ONE FIGURE Work that I have under way on the cranial anatomy of Chla- mydoselachus has led to certain conclusions regarding the hyo- mandubula which, if correct, are of considerable morphological importance. They are based on the assumption, which seems practically established by my work, that the dorsal ends of all of the so-called inner cartilaginous bars of all of the visceral arches, in all of the gnathostome fishes, always lie ventral to the vena jugularis, and that when parts of the cartilaginous bars of the adult fish articulate or fuse with the neurocranium. dorsal to that vein, those parts are derived either from the external cartilaginous bars of the arches, or from interarcual cartilages developed in, or in relation to, the dorsal interarcual ligaments. The terms inner and external cartilaginous bars, or arches, are - here used with the significance commonly given to them, but my work tends decidedly to confirm Dohrn’s conclusion that the so-called inner cartilaginous bars of the branchial arches of the enathostome fishes, and not the external ones, are the homologues of the cartilaginous branchial arches of the Cyclostomata. The cartilages which form the so-called external arches of the gnathostome fishes are commonly called the extrabranchials in whatever arch they may be found. Parker (’76) limited the use of this term to the external cartilages of the gill-bearing, branchial and hyal arches, employing the term extraviscerals to designate, collectively, these cartilages and certain others, the so-called la- bial cartilages of his descriptions, which he considered to be their serial homologues in the prehyal arches. The use of the term 563 JOURNAL OF MORPHOLOGY, VOL. 26, No. 4 DECEMBER, 1915 564 EDWARD PHELPS ALLIS, JR. extravisceral, thus employed, has been properly objected to, but as the term visceral is currently applied, not only to all of these arches themselves, but also to their inner cartilaginous bars, there would seem to be no good reason for not applying it also to the external cartilaginous bars.. But as, with equal reason, the branchial and branchiostegal rays would then have to be collectively called the visceral and viscerostegal rays, I retain the term extrabranchial for these cartilages in whatever arch they may be found. Branchial, hyal and visceral I employ as proposed by Gaupp (’05). The vena jugularis, which seems to have had a markedly im- portant. influence on the development of the definitive carti- laginous bars of the prebranchial visceral arches, is not always, in the different orders of fishes, formed by the fusion of identical sections of the venae cardinalis anterior and capitis lateralis of embryos, the former vein being the primitive one and being said to lie ventro-mesial to the roots of all of the cranial nerves, while the latter one is of secondary formation and is said to lie dorso-lateral to those roots (Hochstetter ’06). The definitive vein also has, in certain fishes, different relations to the hyoman- dibula, as I have quite recently had especially called to my at- tention, lying dorso-external to the hyomandibula in Chlamy- doselachus but ventro-internal to that element in Amia and tele- osts (Allis 714 b, p. 235). Wishing to know if these differences in the relations of the vein to the cranial nerves and to the - hyomandibula were in any way related to each other, I have had the relations of the vein to the trigeminus and posttrigeminus nerves traced in a certain number of fishes. In the Selachii and Batoidei the definitive vena jugularis is said by Hochstetter (’06) to be formed entirely by the vena capi- tis lateralis and to agree in this with the definitive vein in Tro- pidonotus and the Mammalia. In Mustelus (probably laevis) I find the vein lying dorsal to all of the components of the nerves here under consideration, excepting only the latero-sensory fi- bers which, in their exit from the cranium and in their peripheral distribution, are associated with the nervi trigeminus, glossopha- ryngeus and vagus. The vein lies ventral to all of these latero- HYOMANDIBULA OF THE GNATHOSTOME FISHES 565 sensory nerves, but it lies dorsal to the latero-sensory fibers that issue with and as a part of the nervus hyomandibularis facialis. In the Teleostomi, Hochstetter does not give either the method of development or the composition of the definitive vena jugu- laris. I find this vein, on one side of the head of a 43 mm. em- bryo of Amia calva, running posteriorly ventral to the ganglion on the main root of the nervus trigeminus and then upward between that ganglion and the ganglion on the root of the nervus facialis; then posteriorly dorsal to the latter ganglion, and dorsal also to the latero-sensory fibers which issue with the nervus hyomandibularis facialis, but ventral to the latero-sensory fibers which issue with the nervus trigeminus; then downward between the nervi facialis and glossopharyngeus; and then posteriorly ventral to the latter nerve and the vagus: the first section of the vena capitis lateralis that is formed in this fish thus correspond- ing to the one said by Hochstetter to be first formed in reptiles. On the other side of the head of this embryo of Amia the vena jugularis had a similar course, but a large branch of it passed dor- sal to the nervus glossopharyngeus and then downward between that nerve and the vagus to fall again into the main vein, this thus showing a second section of the vena capitis lateralis in process of formation. In one adult specimen of Amia I find the vein running ventral to the nervus trigeminus, dorsal to the nervi facialis and glossopharyngeus but ventral or lateral, and hence morphologically ventral, to the latero-sensory nerves which issue with the nervi trigeminus and glossopharyngeus, and then ven- tral to the nervus vagus. The definitive vena jugularis of Amia is thus formed by the trigeminus and vagus sections of the vena cardinalis anterior and the facialis and glossopharyngeus sec- tions of the vena capitis lateralis, and it corresponds, not only to the second stage in the formation of the definitive vein in reptiles, as given by Hochstetter, but also to the definitive vein (vena petroso-lateralis) of amphibians as given by Driiner (’04). The conditions in the one embryo examined would seem to show that no other sections of the vena capitis lateralis are ever formed in Amia. 566 EDWARD PHELPS ALLIS, JR. In an 80 mm. specimen of Lepidosteus osseus and a 141 mm. specimen of Polyodon spathula the vena jugularis has the same relations to the several cranial nerves that it has in Amia, and I find similar conditions in young specimens of Hiodon tergisus, Scorpaena scrofa and Cottus aspera, and in the adults of Scom- bresox saurus, Gadus merlangus and Trigla hirundo. In a 57 mm. specimen of Catostoma occidentalis and a 40 mm. specimen of Gastrosteus aculeatus the glossopharyngeus section of the vein was the only one which lay dorsal to the related nerve. In an adult Cyprinus carpio and a 48 mm. specimen of Carassius auratus the facialis section of the nerve alone had this position. In embryos and the adult of Ameiurus nebulosus the vein lay ventral to all of the four nerves here under consideration, the entire vena cardinalis anterior thus here persisting as the defini- tive vein; Ameiurus, and hence probably all of the Siluridae, thus differing from other teleosts in this respect as well as in the arrangement of the pseudobranchial and carotid arteries, the innervation of the recti muscles of the eye-ball, and the condition of the myodome (Allis, ’08, 709). In Ceratodus embryos of Semon’s Stage 45, Greil (713, figs. 2-3, pl. 51) shows the vena jugularis, called by him the capitis lateralis, lying ventral to the nervus trigeminus and dorsal to the nervi facialis, glossopharyngeus and vagus, but, as in the Plagiostomi and Teleostomi, ventral to the latero-sensory nerves which issue with the nervi trigeminus, glossopharyngeus and vagus. In Stage 48 of this fish (l.c. fig. 2, pl. 55)- the vein appar- ently still lies ventral to the nervus trigeminus and the same latero-sensory nerves, but it here lies ventral also to the nervus facialis, though still dorsal to the glossopharyngeus and vagus. This must, accordingly, be a less advanced stage than that shown in the embryo of Stage 45, unless it be that the vem change, a second time, its relation to the nervus facialis. But however this may be, it is evident that the definitive vein in this fish is of the plagiostoman and reptilian (Lacerta) type rather than the teleostean or amphibian. One further feature of this vein may here be mentioned. The large orbital venous sinus, found so well developed in the Plagi- HYOMANDIBULA OF THE GNATHOSTOME FISHES 567 ostomi, surrounds the nervus opticus, and the nervus opthalmi- cus profundus traverses it in its course through the orbit, thus in a measure being also surrounded by it. This sinus would thus seem to correspond to that stage in the development of the posterior sections of the vena capitis lateralis which Hochstetter (1. e., p. 183) describes as a ‘Veneninsel’ surrounding the related segmental nerve. A crescentic sinus is found, in the Plagios- tomi, similarly related to the nervus olfactoriuss These several variations in the relations of the vena jugularisto the cranial nerves, while they emphasize the facts that the Pla- giostomi form a group wholly apart from the other gnathostome fishes and that the Siluridae are similarly grouped apart from the other Teleostei, do not present any features which indicate that they have in any way influenced the development of the hyomandibula, as will be later evident. They can accordingly be neglected in the present discussion. ‘They however justify the use of the term vena jugularis rather than either of the terms vena cardinalis anterior or vena capitis lateralis, both of which are frequently used, and the terms vena petroso-lateralis or petrosa lateralis, introduced by Driiner (’01) for the correspond- ing vein in the Urodela and accepted by Kingsbury and Reed (09) as eminently appropriate for those animals, might, at present, be confusing if applied to fishes. In Chlamydoselachus the pharyngobranchials lie imbedded in what has, in one of three specimens examined, quite markedly the appearance of a continuous sheet of muscle fibers which is in process of differentiation into the Mm. interarcuales dorsales I of Vetter’s (74) descriptions of other selachians. This muscle- sheet lies immediately internal (dorsal) to the linmg membrane of the branchial cavity, extends from the hyal to the most pos- terior branchial arch, and, in the one specimen above referred to, its mesial edge lay everywhere slightly mesial to the dorso- mesial ends of the pharyngobranchials. The muscle fibers all run -postero-mesially, the muscle-sheet being considered, for con- venience of description, to run from in front posteriorly, and as the pharyngobranchials do not extend entirely across the sheet they have the appearance of being intercalated obliquely in the Fig.1 Ventral view of the roof of the branchial chamber of Chlamydoselachus anguineus after the lining membrane of the chamber has been removed, showing the pharyngobranchials, efferent branchial arteries and interarcuales dorsales muscles in natural position. ABBREVIATIONS cc, common carotid artery Coe, constrictor of the cesophagus ea.l, efferent branchial artery of first branchial arch ea.lI, efferent branchial artery of sec- ond branchial arch EB II, epibranchial of second branch- chial arch EPB.VI, epi-pharyngobranchial of sixth branchial arch HMD, hyomandibula Iad. IV, M.interarcualis dorsalis be- tween arches [V-V Iad.hy, M.interarcualis dorsalis be- tween arches hy-I lda, lateral dorsal aorta lmh, ligamentum mandibulo-hyoideum n, cut ends of nerves to tissues of roof . of branchial chamber PB.I, pharyngobranchial of first bran- chial arch PB.IV, pharyngobranchial of fourth branchial arch Rabd, retractor arcuum branchialium dorsalis Ssp, subspinalis muscle tiad, ligamentous sheet formed by ten- dons of Mm. interarcuales dorsales 568 HYOMANDIBULA OF THE GNATHOSTOME FISHES 569 course of the muscle fibers. Those fibers that have their ori- gins from the dorso-mesial (proximal) portion of a pharyngo- branchial, pass mesial to the dorso-mesial (proximal) end of the next posterior pharyngobranchial and are inserted on a series of long and slender tendons which run posteriorly, nearly parallel to the vertebral column. These tendons lie in and form part of a thin but strong sheet of ligamentous tissue which extends posteriorly considerably beyond the branchial chamber and is there attached to the tough fascia covering the ventral surface of the dorsal muscles of the trunk. The ligamentous sheet here lies between the trunk muscles and the constrictor of the oeso- phagus, but separated from the latter muscle by a median V- shaped muscle which has its origin from the ventral surface of the vertebral column and is inserted, on either side, on the closely adjoining dorso-mesial (proximal) ends of the fifth pharyn- gobranchial and the so-called sixth epibranchial (Garman, ’85), and in part also on the fourth pharyngobranchial. It is con- tinuous, along the greater part of its lateral edge, with the con- strictor of the oesophagus, and is evidently in process of differ- entiation from that muscle. I have not yet determined its in- nervation, but its position and its relation to the constrictor are such that it seems quite unquestionable that. it is innervated, - as that muscle is, by branches of the nervus vagus. This muscle, and not one or more of the intercuales dorsales muscles, is then quite certainly the homologue of the retractor arcuum, branchi- alium of Amia and teleosts, and Edgeworth’s (11) conclusion that this retractor muscle of Amia and teleosts is derived from trunk myotomes is probably in error. If this be so, one more of the instances frequently cited in support of the view that muscles are subject to radical changes in the manner of their innervation is apparently disposed of. The mesial edge of the ligamentous sheet above described is strongly attached, throughout its entire length, mainly to the ventro-lateral corner of the vertebral column, immediately lat- eral to a shallow median groove which lodges the lateral dorsal aorta, but in part, also, to the ventral surface of a tough mem- brane, continuous with the ligamentous sheet, which extends 570 EDWARD PHELPS ALLIS, JR. from this line of attachment across the median groove to the corresponding line on the opposite side of the head, thus enclos- ing and protecting the dorsal aorta. That portion of the muscle-sheet which lies anterior to the first pharyngobranchial, forms about one-half the length of the sheet, and, there being no pharyngohyal, the muscle fibers here all have their origins from a stout ligamentous band extending from the dorsal (proximal) end of the hyomandibula to the dorsal (proxi- mal) end of the first epibranchial, the band gradually diminish- ing in thickness and consistence toward its hind end. The pos- terior fibers of this part of the muscle-sheet have their insertions, as do the fibers of the interbranchial portions of the sheet, on the anterior edge of the next posterior pharyngobranchial, which is, in this case, the first pharyngobranchial. A small bundle of fibers immediately anterior to these posterior ones then passes mesial to the dorso-mesial end of the first pharyngobranchial, and its fibers, becoming tendinous, join and form part of the ligamen- tous sheet above described. The remaining fibers of this anterior portion of the muscle, which fibers form the larger part of it, are inserted in part directly on the ventro-lateral corner of the anterior portion of the vertebral column and in part on an anterior extension of the median subaortal membrane, the latter fibers reaching the median line of the head and there being in contact with their fellows of the opposite side. In the other two specimens that were examined the conditions differed from those above described only in that the muscle- sheet was narrower, and that the pharyngobranchials, excepting the first, extended entirely across it. Those bundles of fibers that, in the one specimen, passed mesial to the dorso-mesial ends of the pharyngobranchials were accordingly wanting in the two other specimens, excepting only the bundle that had its origin from the first pharyngobranchial. The related ligamentous sheet nevertheless existed in these two specimens as in the other one, and extended the full length of the branchial region, but it had unfortunately been cut and partly dissected away before atten- tion was called to it and its mesial and posterior attachments could not be determined. Those attachments were however HYOMANDIBULA OF THE GNATHOSTOME FISHES 571 quite unquestionably the same as in the specimen first described. In all three specimens the anterior portion of the muscle-sheet was wider and stronger than the posterior portions, and it was here, only, that the muscle fibers reached the middle line of the body. In the specimen first described a small and quite distinctly separate bundle of muscle fibers had their origins from the ventro- lateral edge of the hind end of the neurocranium, and, running directly posteriorly, were inserted on the anterior edge of the large muscle-sheet close to its mesial end. This little bundle of fibers was not found in the other two specimens, but in them certain of the anterior fibers of the large muscle-sheet were them- selves inserted on the ventro-lateral edge of the hind end of the neurocranium. This large muscle-sheet, considered as a whole, is thus at- tached, both anteriorly and posteriorly, to fixed, or relatively fixed (hyomandibula) points, and hence can not act, as a whole, either as a protractor or a retractor of the branchial arches. It can however act to draw the pharyngobranchials closer together and hence have either a protractor or retractor action on indi- vidual arches. It must also have a levator action on all the arches, for it is strongly attached to the neural axis. Thehyo- branchial portion of the sheet, the portion that lies between the hyal and first branchial arches, is certainly simply an anterior member of the Mm. interarcuales dorsales I of Vetter’s descrip- tions of other selachians, and it would seem as if the little anterior bundle of fibers found in one of the three specimens might be a remnant of a prehyal, or hyomandibular, portion of the sheet. The hyobranchial portion of the sheet would seem to be the homo- logue of the subspinalis muscle of Vetter’s and Marion’s (’05) descriptions of Acanthias, which has its origin on the hind end of the neurocranium, and its insertion on the dorso-mesial end of the first pharyngobranchial, while the small hyomandibular por- tion, found in one specimen, would seem to be the homologue of the subspinalis muscle of Vetter’s descriptions of Heptanchus, which has its origin on the ventral surface of the hind end of 572 EDWARD PHELPS ALLIS, JR. the neurocranium and its insertion on the ventral surface of the anterior portion of the vertebral column. The antero-mesial edge of the muscle-sheet is practically paral- lel with the pharyngobranchials, and at once suggests that a pharyngohyal must primarily have existed there. There are, however, no special tissues that seem to represent rudiments of that element. Attached to the interarcual ligament which gives origin to this part of the muscle, near its anterior end, there was, in one specimen, a small and delicate piece of cartilage, which is apparently one of the interarcual cartilages, to be later de- scribed, and not a rudimentary pharyngohyal. The efferent branchial arteries, in all three specimens, perfor- ated the median subaortal membrane and ran antero-laterally across the ventral surface of the muscle-sheet to its lateral edge, each artery reaching the edge immediately anterior to the pharyngobranchial of the arch to which it belonged and slightly mesial to the distal end of that pharyngobranchial. The artery then ran outward across the anterior edge of the pharyngo- branchial, crossed the dorsal (external) surface of that cartilage and so, having passed dorsal (external) to the related dorsal interarcual ligament, reached the external surface of the epibranchial of its arch. The vena jugularis lay everywhere dorsal to the sheet of muscular and ligamentous tissues and hence dorsal also to the pharyngobranchials. The pharyngobranchials of Chlamydoselachus see lie in a sheet of muscular and ligamentous tissues which, although it lies immediately internal (dorsal) to the lining membrane of the branchial chamber, is separated from that lining membrane by the efferent branchial arteries. The latter arteries therefore lie, in a part of their course, ventral (internal) to the pharyngobran- chials, and as these latter cartilages are universally considered to lie primarily internal to the arteries it seems quite certain that, in Chlamydoselachus, the dorso-mesial ends of the pharyn- gobranchials became secondarily attached to the muscle-sheet, which primarily lay dorsal to them as well as to the efferent branchial arteries, and were in consequence lifted upward dorsal . both to the latter arteries and to the dorsal aorta. If this HYOMANDIBULA OF THE GNATHOSTOME FISHES 573 muscle-sheet were then to abort, the pharyngobranchials still retaining their attachment to the neural axis, the dorso-mesial ends of those cartilages would still lie dorsal both to the efferent branchial arteries and the dorsal aorta. But, if the attachment to the neural axis had not been acquired, or had been lost, those ends of the pharyngobranchials would lie ventral to the arteries, and morphologically ventral also to the dorsal aorta. In either case both the pharyngobranchials and the muscle- sheet, if it persisted, would lie definitely ventral to the vena jugularis, their relations to that vein not having been in any way disturbed. And this is what I find in the very unsatisfactory descriptions of other fishes that I have at my disposal. In Stegostoma, Luther (’09) says that the epibranchials of the first and second branchial arches come into contact with the neurocranium, ventro-mesial to the hyomandibular articulation, and hence, as will be later shown, certainly ventral to the vena jugularis. The relations to the lateral dorsal aorta are not given. In Ceratodus, Krawetz (10) says that the epibranchial of the second branchial arch often comes into contact with the auditory capsule, and reference to Greil’s (’13) figures will show that the point of contact must certainly lie dorso-lateral to the arteria carotis interna, which is the anterior prolongation of the lateral dorsal aorta, and ventral to the vena jugularis. In Amia, I (Allis, ’97, fig. 61, pl. 36) found the dorsal (proximal) end of the pharyngobranchial of the first branchial arch attached to the neurocranium dorsal to the common carotid artery (lateral dor- sal aorta) and ventral to the vena jugularis, the pharyngobran- chials of the other branchial arches all lying ventral to the aorta as well as to the vena jugularis. In Scomber (Allis, ’03) the dor- sal end of the pharyngobranchial of the first branchial arch is also in contact with and attached to the neurocranium dorsal to the lateral dorsal aorta and ventral to the vena jugularis, and, as the dorsal end of the first pharyngobranchial of certain of the Clu- peidae (Ridewood ’04) comes into contact with the neurocra- nium immediately ventral to the trigemino-facialis chamber, through which the vena jugularis undoubtedly passes as it does in Amia and Scomber, the pharyngobranchial must there have . 574 EDWARD PHELPS ALLIS, JR. the same relations to the vein and artery that it has in Scomber. In Ammocoetes the dorsal ends of the branchial bars have fused with the neural axis, and Favaro (’08) shows the dorsal aorta lying ventral to them, and the vena jugularis and its anterior prolongations—the venae jugularis dorsalis and capitis lateralis —dorsal to them; and I find similar relations in a single specimen of Petromyzon that I have had examined. These so-called exter- nal cartilaginous arches of these fishes thus have the same rela- tions to these two important blood vessels that the so-called inner cartilaginous arches of the gnathostome fishes have. These relations of the so-called inner cartilaginous bars of the branchial arches to the aorta and vena jugularis are thus quite unquestionably not only a common feature but also a fundi- mental characteristic of all fishes, and, as it seems unquestionable that the prebranchial arches of the gnathostome fishes were pri- marily similar to the branchial ones, they must have been a pri- mary characteristic of those arches also. These relations have, in fact, persisted in both the hyal and mandibular arches of all the Plagiostomi, so far as I can find described and as will be later shown, and because of this, and also because all apparent deviations from the rule in the prebranchial arches: of other gna- thostome fishes can be fully explained by the assumption of the association, or fusion, with the inner cartilaginous bars of those arches, of the related extrabranchial or interarcual cartilages, I assume, as stated in the opening paragraph of this paper, that they have persisted and are invariable in all gnathostome fishes in so far at least as the vena jugularis is concerned, which is the important consideration in this discussion. But, before con- sidering these prebranchial arches, the extrabranchial and inter- arcual cartilages of the branchial region must be considered. The extrabranchials are generally considered to be peculiar to the Plagiostomi and to be specially modified dorsal and ventral ones of the branchial rays of that arch in the diaphragm of which they lie (K. Furbringer ’03). They are said to differ radically from the other branchial rays in being attached, respectively, to the pharyngeal and hypal, instead of to the epal and ceratal ele- ments of the inner arches, and this attachment is to those ele- HYOMANDIBULA OF THE GNATHOSTOME FISHES 52D ments of the arch next posterior to the one in the diaphragm of which they lie instead of to the elements of their own arch. Dohrn (’84, p. 119) says that they never articulate with the ele- ments to which they are attached, which doubtless means that their bases do not reach and, come into contact with those ele- ments, but Fiirbringer (1. ¢., p. 428) calls attention to the fact that, while this is true of most of the Plagiostomi, it is not true of Laemargus. Fiirbringer does not say with which element of the arch the extrabranchial of Laemargus comes into contact, but Gegenbaur’s (’98, fig. 270) reproduction of White’s figure of a branchial arch of this fish shows the contact with the dorsal end of the epibranchial. But however this may be, the impor- tant consideration in this discussion is that the conditions in Lae- margus definitely show that the extrabranchials can come into contact with some element of the inner cartilaginous bar of the arch next posterior to the one to,which they belong; and if they be modified branchial rays, as is generally considered to be the case, it is evident that they have potentially the possibility of coming into contact with some element of their own arch and of there fusing with it, as the branchial rays of Torpedo actually do. According as an extrabranchial is attached to an element of its own arch or to one of the next posterior arch, its relations to the nerve and artery of the arch to which it is so attached change from posterior and superior in the former case to anterior and inferior in the latter, this at once suggesting the suprapharyn- gobranchials and infrapharyngobranchials of van Wijhe’s (’82) descriptions of ganoids and Polypterus; and it was the evident suggestion of the extrabranchials representing one or the other of these two elements, both of which were considered by van Wijhe to belong to and form part of the inner cartilaginous bar of a complete and normal branchial arch, that led me to undertake the present study. And it is perhaps needless to say that had I known how far it would lead me I should not have undertaken it with the limited material at my disposal. The extrabranchials are said to be developed either in sup- porting and protecting relation to the branchial diaphragms, as are the other branchial rays (Dohrn ’84), or as a special adapta- 576 EDWARD PHELPS ALLIS, JR.. tion of certain of those rays to the purpose of keeping the bran- chial clefts open and so preventing a possible stoppage of the flow of water through them (Firbringer ’03, p. 435). According to Gegenbaur (’72, pp. 164-166) they are archaic and very variable structures, inherited from the common ancestor of the Selachi and Cyclostomata and now in process of reduction and disappear- ance, and, as rudiments, only, of the dorsal ones were frequently found by him when the ventral ones were well developed, Gegen- baur concluded that the dorsal ones were subject to reduction before the ventral ones. Furbringer (03, p. 432), on the contrary, concludes that the dorsal extrabranchials, although almost always less strongly developed than the ventral ones, persist longer. Gegenbaur says that both dorsal and ventral extrabranchials are wanting in most of the Batoidei, for excepting rudiments, only, in Rhynchobatus and Trygon, he failed to find them in any of these fishes. Parker (’76), also, says that extrabranchials are wanting in the Batoidei, so far as he can make out, and he only describes ventral ones in Seyllium canicula. Ridewood (’97) has however since called attention to the fact that both dorsal and ventral extrabranchials were described by Rathke, in 1832, in Raia. -aquila and Seyllium canicula, and Foote (’97) has described them in Raia erinacea, Raia radiata, Torpedo ocellata and Trygon pastinaca. . In Chlamydoselachus I find the extrabranchials only slightly de- veloped, as Fiirbringer (’03) has already described them. In Mustelus (probably laevis) I find the bases of the dorsal extrabranchials expanded into relatively large plates which lie directly against the dorso-lateral portion of the large venous sinus formed by the branchial portion of the vena jugularis, im- bedded in the connective tissues surrounding that vein and evi- dently specialized, if not developed, in protective relation to the vein. They do not reach the inner branchial cartilages by a considerable interval, but they are connected with them by the connective tissues in which they are imbedded and by a narrow band-like muscle related to each extrabranchial. No special liga- ments were found binding or connecting the two sets of carti- lages. The enlarged bases of the ventral extrabranchials lie HYOMANDIBULA OF THE GNATHOSTOME FISHES ont ventro-mesial to the vena jugularis inferior, separated from that vein by certain muscles of the region but still in protective rela- tion to it. The vein here lies between the base of the extrabran- chial and the base of the adjacent ventral ray of the branchial series, definitely separating the one from ‘the other; which must be a secondary adaptation if the extrabranchials are simply modified branchial rays. In both Heptanchus cinereus and Acanthias blainvillii the bases of the dorsal extrabranchials lie, as they do in Mustelus, against the lateral wall of the vena jugularis, but in a single speci- men of Cestracion that I have, the dorsal ends of the dorsal extrabranchials pass dorso-mesial to the vena jugularis and al- most reach the pharyngobranchials mesial to that vein, thus lying against the dorso-mesial surface of the vein. The rela- _ tions of the dorsal extrabranchials of Cestracion to the vena jugu- laris are thus similar to those of the ventral extrabranchials of Mustelus to the vena jugularis inferior. The enlarged bases of these dorsal extrabranchials of Cestracion all touch each other and are bound together by connective tissue, but they have not fused into a longitudinal bar such as Gegenbaur (’72) describes in his specimen of this fish. In Raia radiata I find the cartilages described by Foote (97) as dorsal extrabranchials much as she describes them, but they are less completely fused with the expanded outer ends of the branchial rays. Like Gaupp (’05, p. 897), I should have been in- clined to consider this cartilage simply as a plate resulting from the fusion of the outer ends of the branchial rays, were it not that the dorsal (proximal) end of the’ cartilage presents two points, one of which is bound by ligament to the epibranchial of its own arch and the other to the pharyngobranchial of the next poster- ior arch. This dorsal end of the cartilage lies, as ‘do the bases of the extrabranchials in the Selachii above described, against the lateral wall of the vena jugularis. The anterior cartilage of the series lies in the hyal arch. No other Plagiostomi or other fishes were examined in this connection, but in the dissections of the few fishes above referred to, the dorsal interarcual cartilages to which reference has several 578 EDWARD PHELPS ALLIS, JR. times been made were found developed in, or in relation to, the series of dorsal interarcual ligaments which extend from one branchial arch to the next, at or near the points where the pharyn- gobranchials and epibranchials articulate with each other. In the one specimen of Chlamydoselachus examined there were none of these cartilages excepting only the one already re- ferred to which lies in the ligament that extends from the hyal to the first branchial arch. In one specimen of Torpedo ocellata there was an interarcual cartilage related to each of the interarcual ligaments, and there was also one in a ligament which extends from the first branchial arch to the ventral portion of the hind end of the neurocranium; this last ligament evidently being the one that primarily extended from the first branchial to the hyal arch. In one specimen of Raia radiata there was, on one side of the head, aninterarcual cartilage related to the ligament which extends from the first to the second branchial arch, this cartilage lying close against the hind edges of the pharyngobranchial and epi- branchial of the first arch. On the other side of the head a pro- cess on the hind edge of the first pharyngobranchial corresponded exactly, in position, to the independent cartilage on the other side, and similar processes were found, on both sides of the head, on the pharyngobranchials of the more posterior arches. These processes thus apparently represent a series of interarcual carti- lages that have each fused with the pharyngobranchial of the next anterior arch. On one side of the head of this specimen a small and independent bit of cartilage was found at the dorsal end of the epihyal, apparently representing the dorsal interar- cual that lies between the hyal and first branchial arches. Simi- lar cartilages are said by Gegenbaur (’72, p. 175) to be found in most of the Batoidei, and he considered each of them to repre- sent a pharyngohyal that had been segmented off from the dor- sal end of the secondary epihyal of the fish, the primary epihyal being included in the hyomanidubula and the secondary epihyal being formed by secondary segmentation from the dorsal end of the ceratohyal. HYOMANDIBULA OF THE GNATHOSTOME FISHES 579 In one specimen of Mustelus (probaby laevis) a small carti- lage was found attached to the anterior edges of the articulating ends of the pharyngobranchial and epibranchial of the first bran- chial arch, this cartilage thus apparently being the interarcual cartilage that lies between the hyal and first branchial arches, and if so it is important in that it shows that these cartilages can be related either to the anterior or posterior one of the two arches between which they lie. No other independent interar- cual cartilage was found on either side of the head, but there was a somewhat corresponding process on the anterior edge of the | dorsal end of each epibranchial, similar to the processes shown by Parker (’76) in Seyllium canicula. These interarcual cartilages and the related ligaments of se- lachians thus have exactly the relations to the branchial clefts and inner cartilaginous arches that the ‘epitremal’ longitudinal processes of the cartilaginous branchial arches of Ammocoetes and Petromyzon have (Gaupp ’06), and if the cartilaginous arches themselves of these fishes are homologous, as seems so probable, it would seem as if the interarcual and epitremal cartilages must also be homologous, notwithstanding that the interarcual liga- ments of selachians lie internal to the efferent branchial arteries while the epitremal processes of Ammocoetes, as shown by Fa- varo (08), lie external to those arteries. The conditions found in the fishes above described, notwith- standing the limited number that were examined, seem. to war- rant the conclusion that the extrabranchials of the Plagiostomi —whether dorsal or ventral, and whatever their origin—have had their basal portions either developed or specialized in pro- tective relation to the related vena jugularis. And as these ex- trabranchial cartilages are presumably archaic structures, and as both dorsal and ventral ones are said to be found, in all the Plagiostomi, related to the hyal as well as to the branchial arches, there seems no reason to doubt, not only that they were developed also in the prehyal arch or arches of those fishes but also in the corresponding arches of the common ancestor of all the gnathos- tome fishes, and that, accordingly, rudiments or modifications of them should be found in the Teleostomi and Dipneusti. In JouRNAL or MorpHouogy, Vou. 26, No. 4, 580 EDWARD PHELPS ALLIS, JR. these latter fishes I have made no special search for these ecarti- lages, but it seems to me practically unquestionable that, as al- ready stated, they are represented in the suprapharyngobran- chials of van Wijhe’s (’82) descriptions of ganoids and Polyp- terus. These suprapharyngobranchials are shown by van Wijhe, sometimes as independent cartilages, sometimes fused with the infrapharyngobranchials, and sometimes fused with the epibran- chials of their respective arches. In Polyodon van Wijhe did not find any of them, but Bridge (79, p. 709) found, in each of the first two branchial arches of this fish, a little cartilage, de- ' seribed by him as a short pointed cartilaginous ray, directed up- ward and backward from the upper posterior angle of the related epibranchial; these so-called rays thus certainly being strictly simi- lar to the cartilages described by van Wijhe as suprapharyngo- branchials in Acipenser. The relations of the suprapharyngo- branchials to the vena jugularis are not given either by van Wijhe or Bridge, but it would seem as if they must lie lateral to that vein. If so they are certainly extrabranchials, and it is evident that all the various forms of suprapharyngobranchials described by van Wijhe would arise by simple adaptations and fusions of such cartilages with one or the other of the inner cartilages of its arch, or with that cartilage together with the related interarcual. In Amia there are, furthermore, indications that certain of the ven- tral extrabranchials have been preserved, for in this fish there are ventro-mesial processes on the third and fourth hypobranchials, shown by both van Wijhe (1. c.) and myself (’97; fig. 50, pl. 33), which le ventro-mesial to the vena jugularis inferior and hence in the same protective relation to that vein that the bases of the ventral extrabranchials have in the Plagiostomi. ~ bs SF =< pane pa MX he cana bans? " mm ANNAN | | | cn) — | } \\N ill I TLL NAS Z| G UR — 548 ere se il 1 fy ; Hn: iE i} Fig. 53 Pholadidea ovoidea; b, bay of mantle; fo mantle; g, gill; m, part of mantle thrown forward; p, posterior palp. Fig. 54a Zirfaea gabbi; aa, anterior adductor; es, excurrent tube of siphon; f, foot sihpon; pa, posterior adductor; pp, posterior palp. Fig. 54b Zirfaea gabbi, posterior end of body; septum between siphon tubes. , minute foot opening through ; ec, epibranchial chamber; (39, gills; gb, bases of gills; 7s, incurrent tube of gb, bases of gills; ms, musele CILIARY MECHANISMS OF LAMELLIBRANCHS 679 Palps. The ciliation of the thick, heavy palps presents the usual features. In figure 48 a part of the anterior palp of the left side has been removed (at p) and exposes a part of the folded surface of the posterior palp, and the primary course across folds toward the mouth. The ventral edge, with its backwardly directed current, is very narrow. The outer faces, or those with- out folds, are ciliated, and direct material backward to the tips. Mantle. Posteriorly and below the line of gill attachment (ga), material is moved to a line lying parallel with the mantle edge, and then along this to a point y, between which and the point 2, it is carried over the edge. Between y and z on the edge, there is a feeble ciliation directed outward over the edge. Pos- terior to the adductor muscle (a) and above the line of gill attach- ment, there is a feeble ciliation upward and backward, over the edge above the point z. There is here also a strong current of water from the gills, which takes the same course, and aids in sweeping outward feces from the rectum (r). The mantle and the greater part of the visceral mass wall are fused on the left side. On the right, there is a space (e) anterior to the adductor, where mantle and epibranchial chambers communicate. The mantle wall here, and also that of the visceral mass, possess a ciliation directed upward. Figure 49 represents the mantle ciliation of O. lurida of the Pacific. At the point x is a narrow line over which collections pass from the mantle edge. Pholadidea penita Conrad Specimens were taken from San Diego Bay, California. The form is a true borer; it possesses an unpaired plate dorsal to the hinge. Gills. Cilia currents on all lamellae are to the free edges of the demibranchs and forward (fig. 50). There is a current forward in the angle between demibranchs. Palps. The lateral extension of these organs is great. The ventral margins are well defined, and those on the dorsal edges are exceptionally wide. The currents of the apposed faces are 680 JAMES i. KELLOGG represented in figures 51 and 52. In the former the general current on the folds is seen to be obliquely forward and markedly downward across them. Between the folds are dorsally directed currents which seem always to be exposed, not simply by the parting of folds in response to the stimulus of large quantities Fig. 55 Zirfaea gabbi; ciliation of palps and visceral mass; f, foot; bg, gill bases; m, mantle wall; p, posterior palp; vms, visceral mass. Fig. 56 Zirfaea gabbi; section across siphon; es, excurrent tube of siphon cut open dorsally; g, gills; 7s, incurrent tube of siphon. Fig. 57 Barnea costata; b, bay of mantle; cm, collecting membrane; es, ex- current tube of siphon; f, foot; ga, line of gill attachment; 7s, incurrent tube of siphon; m, mantle wall; vms, visceral mass; x, point at which material from visceral mass passes to collecting membrane. CILIARY MECHANISMS OF LAMELLIBRANCHS 681 of material, as in Schizotherus and others. The result is, as shown in figure 5”, that large and small quantities of material alike are usually carried in a broken, zigzag course to the lateral portion of the oral groove, much as indicated by the long feathered arrow. Small masses in the lateral groove pass on toward the mouth, while larger quantities are invariably caught up out of it by cilia of the ventral margin, and conducted to the palp tip. Visceral mass. ‘The ventral portion of the visceral mass is shown in figure 50 (vms). The foot has disappeared—there are no remains of foot muscles—in mature individuals, though a small foot opening persists far forward in the fused mantle edge (fo). The erystalline style shows through the thin epidermis at the point of the mass. The walls of the upper portion of the mass are ciliated, and conduct material to its posterior side, where it is cast off into the mantle chamber in the usual manner. Mantle. In the same figure, the mantle is shown to present no unusual features. Pholadidea ovoidea Conrad Specimens were taken from burrows about four inches deep in extremely hard clay, between tide lines on Puget Sound. Ana- tomically the form is very similar to P. penita, but unimportant differences are shown in figure 53. The lateral parts of the palps are enormously enlarged, especially the dorsal margins, which are fused extensively with the mantle (m) in the case of the anterior, and with the visceral mass in the posterior palps. The visceral mass is relatively more extensive than in the pre- ceding species, and of a different form ventralward. The foot is entirely absent, the sexual mass entirely filling the foot-like projection of the visceral mass; but a minute foot pore (fo) opens through the mantle edge far forward. There is a small forwardly projecting fold on the left mantle wall under the incurrent siphon. Whatever it may be, it has no relation to the ciliary mechanism. The ciliation of organs is like that of P. penita except on the palps, where the general current is directly across the folds 682 JAMES L. KELLOGG cern £ fake toward the lateral oral groove; and no dorsally or ventrally directed currents were found between folds. Only two speci- imens were available for examination, however, and it may be that such tracts exist. Zirfaea gabbi Tryon The shell of one specimen of this Pholad taken from Puget Sound measured 12 cm. or nearly five inches in length. The animals were dug with a pick from blue clay, nearly as hard as rock, and the burrows of larger individuals were about 20 inches in depth. Siphon and gill development here, and in some other members of the family, are very extraordinary, as shown in figure 54. The mantle chamber cannot be distinguished from the incurrent siphon tube, the whole mantle being extended poste- riorly without the usual constriction of the siphon base. Instead of ending posteriorly at the base of siphon tubes, as is usual, the gills extend backward, in an uncontracted siphon, for nearly two-thirds of its extent, attaining an actual length, in large indi- viduals, of a foot or more. The tubes of the distal third of the siphon are separated by a muscular septum (ms), the dividing wall of the remainder being formed by the gills which are united on the median line and joined laterally to the siphon wall, as shown in figure 56. In this figure the excurrent siphon is cut and opened dorsally. Possibly the distal region only should be regarded as true siphon, but a strong argument could be formu- lated against such a view. Mantle and siphon walls are of great thickness. There is a piston-like foot which may in some way aid in boring; but it may be that the great muscular develop- ment of the siphon enables the animal to impart a drilling motion to the anterior part of the body when the structure fits tightly in the burrow. It would be interesting to know how, during its growth, this and other borers manage to deepen and enlarge their burrows. Gills. On all lamellae of these enormously extensive col- ‘ectors, currents trend ventrally to the margins, where material is received into grooves. Apparently the conduct of material CILIARY MECHANISMS OF LAMELLIBRANCHS 683 over the distance of a foot—as some of it must be carried—in the midst of the rushing stream of the incurrent siphon, would be attended by so much uncertainty in an open groove, that com- pletely covered passages have been developed on the edges of the demibranchs, which assure the delivery of the gill collections to the palps. When material from the lamella reaches the closed groove, its walls part, admit the collection, and close over it, if its volume is not too great. Palps. These organs are relatively large, and bear large folds. The anterior palp is not shown in figure 54. In figure 55, the mantle has been cut ventrally, the right side being lifted to show the attachment of the anterior palp to it, as well as the short line of union between posterior palp (p) and visceral mass below the line of gill attachment (gb). The ciliation of folds and margins is beautifully clear, and is found to be essentially like that of the Schizotherus palp, except that no dorsally directed currents were observed anywhere on the crests of the folds. Visceral mass. Posteriorly the visceral mass is extended into a point, just below which its collections are cast into the mantle chamber, as shown in figure 55. Mantle. No case has been observed among lamellibranchs, in which the inner walls of the siphon tubes were ciliated. There is in Zirfaea a downwardly directed ciliation on the mantle walls opposite the visceral mass, that trends posteriorly on its ventral surface. There is no bay beneath the siphon base, as in other forms, and apparently no point where a collection is made. Material simply gathers on the ventral mantle wall, and is no doubt discharged through the ventral siphon from time to time, as a result of the contraction of the adductor muscles. The mantle streams, in the forms examined, were very feeble. Barnea costata Lin This species, studied on the coast of Louisiana near the mouth of the Calcasieu river, is remarkable in having developed an organ, the function of which is to extend and perfect the ciliary mechanism for freeing the body of undesirable foreign matter. 684. JAMES L. KELLOGG Excepting this organ, the ciliation of the body conforms, in general, to that usually found in bivalves. Proceeding at once to a description of what may be called for lack of a better name, the collecting membrane of the visceral mass, it will be noticed that the thin, filmy structure (fig. 57, cm) is an outgrowth of the posterior wall of the body (vwms), and projects into the mantle chamber. It may be retracted close to its base, or extended backward for a great distance, when fully performing its functions being thrust into the incurrent siphon tube. Because of its great power of extension, it does not seem impossible that its end may sometimes be projected entirely through the siphon tube. It tapers from base to tip, and has the form of an inverted trough. In the figure, cilia currents on the extensive collecting area of the visceral mass, below the line of gill attachment (ga), converge on each side of the body to the base of the membrane, on its ventral edges, at the points x, right and left. An attempt has also been made to show that the outer or upper convex surface of the membrane is ciliated, and carries material to the same regions. Here all collections pass to the under, concave, surface of the collecting membrane, and are carried swiftly backward, directly in the face of the incoming stream of water, to the tip of the membrane, as represented by dotted arrows. Here they are held—the tip usually resting on the siphon wall—until a ball of considerable size has been collected. This, and the mantle collection at b are discharged on contraction of the adductor muscles. Specimens in which this membrane was found, were taken from waters which frequently were almost thick with silt for many hours at a time, but it was observed that mud was never so abundant as to cause the creatures to withdraw their siphons, and so prevent its entrance to the mantle chamber. This was clearly seen at a stage of the ebb tide when only a film of water covered the flats in which individuals were numerous. The presumption is that, under such conditions, the ordinary means of removing mud were insufficient, and that this collecting membrane was developed as an aid. At any rate, that mud CILIARY MECHANISMS OF LAMELLIBRANCHS 685 removal is its function, is not to be questioned. What seems to be an homologous structure, was found in B. pacifica. Barnea pacifica Dall. This beautiful form was studied at San Diego Bay, and was taken from stiff clay, the burrows being four or five inches in depth. In relation of siphon to body, it much resembles Zir- faea, previously described. As in that form, the gills extend posteriorly far into the siphon (fig. 58), there forming the septum between incurrent and excurrent tubes. Gills. Currents on all lamellae are downward to demibranch margins, where, as in Zirfaea, the groove leading to the palps possesses walls so high that they arch over and convert it into a closed tube, opening to receive and discharge collections (fig. 59). There are anteriorly directed streams between the bases of the demibranchs. Palps. Here, again, these organs are relatively extensive. The ciliation of folds and margins is precisely similar to that of Pholadidea penita, except that here a very narrow tract, lying along the upper ends of the folds, lashes toward the oral groove, but is so narrow that collections from the dorsal margin cross it on the way to the folds, with little influence from it. It is safe to say that there are no ventrally directed tracts on the palp folds. Visceral mass. Figure 60 represents the visceral mass from the right side, after mantle, palps, and gill have been removed. Ventralward there is a disk which contains some muscle tissue, but the sexual mass extends to the disk surface, and no foot remains, unless a few scattered muscles are the vestiges of it. Figure 61 is a ventral view of the visceral mass. In the last three figures there is shown a thin, nearly transparent struc- ture (cm) with a slight concavity on its ventral surface, which is evidently the collecting membrane so greatly developed in B. costata. When a comparison is made with figure 57, the course taken here by cilia currents is very suggestive. It will -be observed that these currents are directed to a region on the side, near the middle of the body, whence they course to the 686 JAMES L. KELLOGG ventral wall, and then posteriorly until they reach the end of the collecting membrane. Without extending itself, this membrane lies well within the base of the incurrent siphon, because of the Fig. 58 Barnea pacifica; g, position of gills; s, siphon. Fig. 59 Barnea pacifiea; em, collecting membrane; ig, inner demibranch of gill; m, portion of mantle thrown forward; og, outer demibranch of gill; pp, posterior palp; vms, visceral mass. Fig. 60 Barnea pacifica; ciliation of visceral mass; letters as in figure 59 Fig. 61 Barnea pacifica; ventral view of visceral mass ciliation. Fig. 62 Barnea pacifica; ciliation of mantle edges seen from above; b, bay of mantle; fo, foot opening of mantle; me, fused mantle edges; v, vortex. CILIARY MECHANISMS OF LAMELLIBRANCHS 687 backward prolongation of the visceral mass. Whether col- lections accumulate on its ventral surface or not, could not be determined. Because the course of cilia currents over the entire visceral mass is so much like that of B. costata, it would seem that there is here a much degenerated collecting membrane, which formerly may have been much more extensive. Mantle. The walls of the mantle chamber, unlike those of Zirfaea, may be distinguished from siphon walls, because they are very thin, while the siphon walls are muscular. Currents are directed ventralward from the sides to the fused ventral margins. The details of currents in this region are shown in figure 62 which represents a view of the fused ventral man- tle edges seen from above. The foot opening (fo), through which the disk-hke end of the visceral mass may protrude slightly, is still large. Some of the material brought from the side walls of the mantle, pauses, and may collect in small balls at two points (v) which suggest the vortices of Mya. There seems to be no whirling movement here, and collections are sooner or later moved on to the thickened and fused edges. At several places, as shown by the arrows, they pass on to a median, broad, slightly depressed tract (me) and are carried swiftly backward to a point, b, corresponding in position to the bay at the base of the incurrent siphon, which has been shown to be so generally present on the lamellibranch mantle. Unio complanatus Solander Gills. Specimens were from a lake in western Massachusetts. On both lamellae of the outer demibranch (fig. 63, og) collections are carried to the base, where, on each side, they are moved anteriorly, along narrow lines, to the distal oral groove, just in front of, and parallel with, the anterior margin of the inner demibranch (fig. 64, dg). Unless the material is very small in amount, it touches the anterior wall of this groove made by the mantle (m), and is lifted out and carried away by the mantle cilia. On both lamellae of the inner demibranch, material is carried to the free margin, and then forward in its groove to a 688 JAMES L. KELLOGG ( ae: 2 CILIARY MECHANISMS OF LAMELLIBRANCHS 689 point just posterior to the palps (below the reference letters dm) where it collects, if not at once transferred to the apposed palp faces. Collections are also brought downward to this point, along the demibranch edge lying parallel with the distal oral groove. It is probable that in all bivalves, the palps may be withdrawn from contact with this demibranch edge, so that accumulations on the margin become great enough to fall of their weight to the mantle or visceral mass walls. Palps. Unio is one of those forms in which the dorsal edges of the lateral extensions of the palps are united so as to bring the sets of folds close together, leaving but a very narrow tract (the lateral oral groove) between them (fig. 64, lg). Where the distal joins the lateral part of the oral groove, the dorsal margins make a short trough (to the left of reference letters ap). The ventral palp margins (vm) in the Unios and Anodons are narrow, as are the folds, and observation of their ciliation is difficult. The wide dorsal margin of the posterior palp receives much material collected by the visceral mass in its vicinity; that of the anterior, in the same way, receiving mantle collections on its inner, and also on its outer face from contact with the mantle. These points are not well shown in the figure. The outer faces of both palps are ciliated, and bear material to the free dorsal margins, where it comes over on to the inner faces. None of it is moved to the folds, but all is carried to the free tips and thrown off. The general ciliation on palp folds is nearly straight across them, forward. The details of the ciliation on the folds are Fig. 63 Unio complanatus; es, excurrent opening of siphon; f, foot; 7g, inner demibranch of gill; 7s, incurrent opening of siphon; 0g, outer demibranch of gill; p, palps. Fig. 64 Unio complanatus; ciliation of palp region; aa, anterior adductor; ap, anterior palp; dg, distal oral groove; dm, dorsal palp membrane; ig, inner demibranch of gill; lg, lateral oral groove; m, mantle, og, demibranch of gill; pp, posterior palp; vm, ventral palp margin; vms, visceral mass. Fig. 65 Unio complanatus; detail of ciliation of palp folds; fd, fd, fd, palp folds; vm, ventral palp margin; x, direction of ciliation across folds; y, separation of folds to expose deep cilia tract; z, course of particles over deep tract. Fig. 66 Anodon sp.; mantle ciliation; es, excurrent opening of siphon; ‘s, incurrent opening of siphon; m, mantle wall; p, palp. 690 JAMES L, KELLOGG shown in figure 65, which represents the ends of several folds (fd,fd) near the ventral margin (vm). While the lower ends of the folds project for some distance ventralward, they are raised (iy N = oe eeu — = NASUBUOUE rn pas) Saar Sen —— -% =) | iC 17, UCE {—— HOTTY ean C= | \ ee {\ | \ ee Oo) © Fig. 67 Yoldia limatula; ap, anterior palp; f, foot; g, gill; gm, suspending membrane of gill; lg, extension of lateral oral groove; pa, palp appendage; s, siphon; st, siphonal tentacle; vms, visceral mass. Fig. 68 Yoldia limatula; ciliation of gill and suspending membrane; cu, line of ciliary unions; gm, suspending membrane of gill; mp, modified plates of gills; pw, waste from palps; vec, ventral ciliated canal of gill. Fig. 69 Yoldia limatula; ciliation of ventral surface of gill; Letters as in figure 68. CILIARY MECHANISMS OF LAMELLIBRANCHS 691 above the margin, not touching and interfering with its current. The folds lie over anteriorly—toward the right of the figure— each covering a part of the base of the one in front of it, in the usual manner. The line of arrows x represents the general course taken by particles across the folds toward the mouth. On each fold, close to the edge of the fold behind it (as just above the letter y) there are very narrow tracts, the cilia on which direct currents dorsalward toward the lateral oral groove. These have some influence on material crossing the folds, tending to keep it away from the ventral margin, on which it would be ear- ried away from the mouth. When much material appears on the folds, they lift here and there (y), so that it is caught quickly and certainly by cilia deeper in the grooves, and, as represented by the dotted arrows, is carried to the outgoing stream on the ventral margin, precisely as in Schizothreus, Cardium, and others. Mantle. Figure 66 represents the ciliation of an Anodon, which is in most respects similar to that of Unio. In both, material is carried out over the margin below the lower siphonal opening (7s), as is usual in forms not completely buried. Yoldia limatula If this much discussed genus is properly placed among the most archaic of living lamellibranchs, this representative of it certainly possesses the most extraordinarily complex set of ciliary mechanisms observed in the group. It lives in soft mud, and, as shown by the stomach contents, allows relatively large quantities of indigestible material to pass into the digestive tract, though the relative volume of diatoms found, is very much greater than in Macoma. Yoldia does not, by any means, feed simply on what it is able to digest out of the mud of the bottom in which it lives, for diatom shells would not in that case form so large a part of the stomach contents; and several very elabor- ate and effective mechanisms exist, the function of which is to clear the organs exposed in the mantle chamber of objectionable material brought by the incurrent stream. In order to understand the operation of these mechanisms, it is necessary to have a clear notion of the relative positions of JOURNAL OF MORPHOLOGY, VOL. 26, No. 4. 692 JAMES L. KELLOGG CILIARY MECHANISMS OF LAMELLIBRANCHS 693 several organs bearing ciliated surfaces. In figure 67 the reference letters ap are placed on the anterior or outer palp, the edge of the posterior, which it overlies, appearing beyond its back margin. Arising from about the middle of the dorsal margin of the outer palp, is a long appendage (pa) which may be extended beyond the shell margin for a distance greater than the length of the shell. Above and behind the palps hang the gills (g) that are suspended from above by a membrane (gm). It is with the relative positions of gills and palps that we are here con- cerned. The two palps, on either side of the body, are united along their dorsal edges, upward from the point where the ap- pendage arises from the anterior or outer, and along this line of union is a deep groove, an extension of the lateral oral groove, which lies hidden between the palps below the region of the vis- ceral mass (vms). This extension of the groove lies beneath the reference letters lg in figure 67 and is shown in figure 70, at lg. It is often placed in such a position that it touches the lower surface of the anterior end of the gill, though in the first figure, the gill is raised somewhat above it, and in the second (fig. 70) this entire dorsal region of the palps is pulled downward, their free ends being spread apart in order to expose all of their cili- ated surfaces. In its normal position, then, this extension of the oral groove (lg) lies against, or close to, the ventral surface of the gill, the reason being, as will be shown, that, from time to time it receives the gill collections. The anatomy of the gill, first described in detail by the writer, is well known, as it has been studied by several observers, but one important structural modification has been overlooked. I know of but one brief reference to it—that found in Drew’s paper on Yoldia (’99) the reference being :—‘‘ With the exception Fig. 70 Yoldia limatula; detail of palp ciliation; dm, dm, dorsal margins of palps; g, gill; gm, suspending membrane of gill; 1g, extension of lateral oral groove; paa, palp appendage; vm, ventral margin of palp; vms, visceral mass; x, notch in groove passing to mouth. Fig. 71 Yoldia limatula; ciliation of outer surface of palp appendage. Fig. 72 Yoldia limatula; ciliation of mouth region; ap, anterior palp; lg, lateral oral groove; m, mantle; mo, mouth; pg, proximal oral groove; pp, posterior palp; vm, vm, ventral margins of palp; vms, visceral mass. 694 JAMES L. KELLOGG of a few of the plates at the extreme anterior ends of the gills, which are sometimes much distorted and swollen |italics mine], all of the plates are alike in shape and structure.” The modified plates which Drew saw are always present, are two in number (the third and fourth in the series, though if this may possibly vary, such variation has not been seen) and are always found only on that side of the gill nearest the median plane of the body. They are modified to dispose of gill collections which are not to be sent to the palps, being one more of the several anatomical structures that are developed to aid the ciliary mechanisms. Figure 68 represents the anterior end of the right gill seen from the inner side, so as to show the modified plates (mp). Along the sides of the plates is the series of ciliated unions (cu) that I have described elsewhere. On the ventral surface of the gill, and between the two series of plates, is a broad ciliated canal (vcc). The edges of the plates above cu are not ciliated, but below this line is a full ciliation of the plate edges which drives water upward between them (see the long feathered arrow) into the epibranchial space, and carries particles brought to them very swiftly down to the ventral canal. Here the collections are carried forward, but are halted momentarily about opposite the fifteenth plate by a narrow, backwardly directed tract lying along the bases of the inner plates (see also fig. 69, a ventral view). Very small amounts seem to pass this point without interference. It is possible that the halt is made here in order to facilitate the transfer of material to the palps, the oral groove of which, at times, lies against this region of the gill. At any rate, if this transfer is not made, the material, after revolving a few times, continues on toward the modified plates. These, like all parts of gills and palps, as described by Drew and myself, are capable of extensive movements. They are never in close contact (they possess no ciliated disks on their apposed faces) but may spread wide apart. Their apposed faces are richly ciliated, and material brought by the ventral canal is seized and quickly forced through between them into the epibranchial space. Here, on the wall of the suspending membrane (gm), and aided, perhaps, by the outgoing stream of CILIARY MECHANISMS OF LAMELLIBRANCHS 695 water, they leave the body through the excurrent siphon. In no other gills that I know of except those of Monia, and Pecten, are there special means of conducting undesirable material to outgoing tracts. In other cases, collections may be of such volume that the marginal groove cannot hold them, and they fall into the mantle chamber of their own weight. It is the function of other gills simply to collect, and pass collections on to the palps on which it is determined whether they shall be con- tinued on to the mouth, or to an outgoing tract; but here the gill possesses its own outgoing tract, which must inevitably be used unless contact is effected between gill and palp. Palps. The huge lateral extensions of these organs are for some distance suspended from the overhanging digestive gland (fig. 67, vms). Behind the line of this attachment, and as far toward their free ends as the origin of the appendage, their dorsal margins continue to be united, and on the outside along the line of this union is the groove (fig. 70, lg) the function of which ap- parently is to receive the gill coliections. This groove, so far as I know, has never previously been seen. From its position in reference to the ventral surface of the gill, from the direction of its cilia-currents, and from the fact that its margins may convert it into a nearly closed tube, or open so fully as to expose it, I am convinced that it has been constructed to receive gill collections, though I have no record of having seen the transfer actually made. Gill collections and the palps. Drew, from his study of the gills and palps, concluded that the former were not food col- lectors, as Mitsukuri (’81) had done, also. He says: Experiments were tried to determine, if possible, the part taken by the gills in the collection of food. . . . No definite results were reached, but they were not observed actively engaged in collecting food. Considering the remarkable activity of the palps as collectors of food, such activity for the gills seems rather unnecessary, and it would also seem that the pumping action of the gills would seriously interfere with their normally performing such a function. Drew had not seen the extension of what I have called the lateral oral groove, but I am puzzled to know how the tremendous 696 JAMES *L. KELUOGG activity of the gill in collecting and moving forward suspended particles brought to it in the water, could have escaped his notice, the whole process being precisely like the food collection of other lamellibranch gills. The pumping action of the gills does not disturb small collections, and there is no reason for assuming that it would interfere with the transfer of food from gill to palp. Collections of the palp appendage. In the quotation above, Drew refers to his own observation of collections made by the palp appendages, which I have confirmed many times. ‘This structure (fig. 70, pa) is a trough, the convex surface normally opening ventralward when it is extended from the body, and allowed to rest on the surface of the mud. Here, as Drew de- scribes it, great quantities of material, “foraminifers, ostrocods, and even small lamellibranchs and gastropods, together with the smaller forms and mud, are passed along the groove, finally between the palps, and so on into the mouth.”’ The relation of this groove of the appendages to other ciliated tracts—the oral, and several outgoing streams—is shown at the point x in figure 70. The letter is placed close to a narrow notch, bounded on either side by the outgoing tracts of the dorsal margins (dm). The only passage to the mouth from this entire region is through the notch (indicated by the bent arrow) and when in scores of trials the dorsal tracts were spread apart as far as possible, no large amount of material was ever allowed to pass through it but was always seized by these outgoing tracts. Always, when large quantities of material were brought to this gateway, along the groove of the appendage, as well as along the lateral oral groove, they were halted and turned on to outgoing tracts, especially those of the dorsal margins (dm). The dorsal margin of the posterior palp is very wide, and is bent into a broad, shallow trough which is continuous with the groove (Ig) above it. From the notch « there are also outgoing tracts on the sides of the groove of the appendage, as shown in figures 70 and 71, which carry away a small amount of material. Very narrow streams of material, on the other hand, were able to enter between the inner palp faces, and were often found collected there in considerable quantity. It seems as if Drew must have CILIARY MECHANISMS OF LAMELLIBRANCHS 697 assumed, and not actually observed, that these extensive ap- pendage collections passed to, and entered, the mouth. Yet it may be said that if they do not, but are usually cast away, itis hard to understand why they are made at all. Nevertheless, in very many observations, extensive collections by the append- ages were invariably conducted out of the body on outgoing tracts. It has been shown that large accumulations, brought to the gateway x by appendage or oral groove, are taken up by the dorsal margin tracts and cast into the mantle chamber from the palp tips. There is still another and very remarkable outgoing tract that. may be brought into operation at this point and throughout the groove (lg) shown in figure 70. The anterior wall of this groove is thick, rounded, and capable of no movement. Collections are brought over it from the outer surface of the anterior or outer palp. The posterior lip of the groove, continu- ing out on the dorsal margin, is very thin and flexible, and on its outer face (to the left of reference letters, Ig) is a strong cilia current directed upward. Its course is easily followed, and its loads are seen to be transferred to the inner face of the suspending membrane of the gill (above reference letters gm in figs. 68 and 70) Here this waste matter from the palp passes into the epibranchial chamber of the gills, and leaves the body through the excurrent siphon. This flexible groove margin receives its material chiefly from the groove, by rolling over into it in response to the stimulus of a large amount of material, a process similar to the cleaning action of the ventral margins of such palps as those of Schizo- _therus or Mytilus, where, by an upward folding, or a spiral twist- ing of the entire organ, they are brought in contact with the folds. Palp folds. The palp folds are numerous, and are continued forward very much nearer to the mouth than is usual (as shown in the view of the mouth region represented in fig. 72) so that the proximal oral groove (pg) is very short. It is note- worthy that the palp folds nearest the mouth are the smallest, and that they become very large near the palp tip (fig. 70), the reverse usually being the case. The chief current is directly across the folds toward the mouth, while in the grooves, and 698 JAMES L. KELLOGG apparently always exposed, are narrow, dorsally directed tracts which tend to work material upward to the extremely narrow lateral oral groove (lg, fig. 72). No ventrally directed tracts were seen. Mouth region. Figure 72 represents a very interesting cili- ation. Here the anterior palps (ap) have been folded forward, and the posterior (pp) backward. Instead of forming continu- ous lips in front of and behind the mouth (m) as is usually the case, there is nearly a complete break on the median line, the ventral edge of the anterior fusing with the mantle, which here stretches across between mouth and anterior adductor, about midway between letters m and mo. The ventral margins of the posterior palps, on the other hand, independently enter the mouth vtself, and extend for some distance into its cavity. In most lamellibranchs, material may approach the mouth only along the very narrow line of the oral groove (lg and pg) for near the mouth this is bounded by outgoing tracts; but here it may enter all along the sides of the opening. But even after actually entering, it may still be caught up by outgoing tracts, and ultimately be discharged from the body, and this invariably happens in experiment, when more than a very small volume of material attempts to enter at a time. This is repeated again and again on tracts leading to the mouth in nearly all lamelli- branchs. If material is to enter the digestive tract, it must be conducted to it in narrow streams, and a little at a time. The outgoing streams from the mouth opening are three in number. Qne is on the median line anteriorly, and leads, right and left, to the mantle wall. Another, corresponding to it, is a broad tract from the posterior mouth edge, and leads out on to the visceral mass. The third is found on the ventral margins of the posterior palps, and extends into the mouth itself. Here, however, the outgoing streams appear not to be exposed directly to the mouth cavity, but are on those sides of the palp edges which are next to the wall. They can be exposed, and have been seen to bear out material that had fairly entered the mouth. Mantle. There is a general ciliation of mantle walls, with no distinct tracts, collections being carried ventralward and back- CILIARY MECHANISMS OF LAMELLIBRANCHS 699 ward, and leaving the mantle chamber at the point, just below the siphon, where the palp tentacles are extruded. SUMMARY Numerous anatomical features, not previously noticed, are shown in the figures. Unique organs in the group of the lamellibranchs, are the immerse vascular fold arising from the visceral mass of Schizo- therus (fig. 1), and the collecting membrane of Barnea costata and B. pacifica, also arising from the visceral mass (figs. 57 and 60). The former is connected with the palps, the latter confined to the posterior wall of the visceral mass, and they are probably not homologous structures. The direction of cilia currents is quite different in the two cases. The organ in Barnea is de- veloped to aid and perfect the outgoing cilia mechanism, and apparently ean have no other function. The fold of Schizotherus is an immense blood reservoir, and its cilia to some degree aid the outgoing system. The direction of the beat of cilia is never changed. Cilia streams on the surfaces of organs exposed to the water are divided into two systems, namely, those leading to the mouth, and those bearing material outside the body. The operation of each, in Schizotherus, is described in detail. The numerous currents of all gills belong to the ingoing system, except, among the forms examined, Yoldia, Monia and Pecten, in which the gills possess certain outgoing tracts which are en- tirely different each from the others. Each of these ciliations is complex and extremely efficient. The palps, which have attained their present size, form, and position on account of the function of their cilia, are shown to exercise general control over the two ciliary systems, determin- ing whether or not collected material shall enter the digestive tract. The cilia tracts of the palps are shown to have greater complexity and to be more important than those of other organs, and this account deals most fully with their positions and actiy- ities. Practically nothing has heretofore been published con- 700 JAMES L. KELLOGG cerning them. Streams on outer palp surfaces usually pass over the dorsal margins to the inner or applied faces. Those on the inner faces vary greatly in different genera, as will be seen by a comparison of the various figures. In several forms, parts of the visceral mass collections are passed to the palps. In all cases, the greater part, if not all, is cast into the mantle chamber. Though there is greater variation in the direction taken by streams on the mantle tracts than on other organs, all belong to the outgoing system, except in a few cases in which small amounts are passed on to the anterior palps. There is no selection or separation of food organisms from other water-borne particles. Volume alone determines whether the collected foreign matter that reaches the palps, shall proceed to the mouth, or shall be sent from the body on outgoing tracts. The gills of Yoldia and Pecten also have the power—by two entirely different mechan- isms—of directing their collections on to outgoing tracts when the volume of these is sufficient. A lamellibranch is able to feed only when waters are compara- tively clear—when diatoms are brought to the gill surfaces a few at a time. In muddy waters, all suspended particles, of whatever nature, are led to outgoing tracts. An exception is found in the sand-eating genus Macoma. All ciliated surfaces produce mucus, which appears locally in response to the stimulus afforded by the touch of foreign particles. Its amount is always that necessary to entangle the stimulating particles. Long continued stimulation of any surface may cause relatively enormous quantities of the secretion to be discharged. In several genera, the mantle develops, on each side, folds lying parallel with the edges, posteriorly (Schizotherus and others). These may enclose a waste canal where waste matter accumulates, until carried out of the incurrent siphon or siphonal opening, by a sudden reversal of its stream. A waste canal is made necessary by the presence of a siphon membrane which under certain conditions throws the entering water stream on to CILIARY MECHANISMS OF LAMELLIBRANCHS 701 the ventral wall of the mantle space, as described for several forms. There is much difference in individuals of the same species in the strength or activity of any tract, even where external con- ditions apparently have been the same. No marked variations in the positions of tracts were noticed. The aid of the nervous system and of muscles is frequently necessary to the operation of cilia tracts, as in the exposure of hidden outgoing tracts by the spreading apart of palp folds, the apposition of ventral margin and folds, of palps and gills, gills and mantle, and gills and visceral mass. Besides this there are violent expelling contortions of gills, palps, and mantle edge, sudden contractions of adductor muscles, and other similar movements. BIBLIOGRAPHY Drew, G. A. 1899 Yoldialimatula. Memoirs Biol. Lab. Johns Hopkins Univ., No. 4. Kewioaa, J. L. 1892 A contribution to our knowledge of the morphology of the lamellibranchiate mollusks. Bull. U. 8. Fish Commis., No. 10. 1900 The ciliary mechanisms in the branchial chamber of the Pele- cypoda. Science, vol. 11. 1903 Feeding habits and growth of Venus mercenaria. Bull. N. Y. State Museum, No. 71. 1910 Shellfish industries. Henry Holt, New York. Mirsuxkur!, K. 1881 On the structure and significance of some aberrant forms of lamellibranchiate gills. Quar. Jour. Mier. Sci., vol. 21. Orton, J. H. 1912 The mode of feeding of Crepidula, with an account of the current producing mechanism in the mantle cavity, and some remarks on the mode of feeding in Gastropods and Lamellibranchs. Jour. Marine Biol. Assoc., vol. 9. Srenra, M. 1901 Uber eine bei Lamellibranchiaten beobachtete untere Riick- stromung sowie iiber die Wimperrinne des Mantels von Pinna. Zool. Anz., Bd. 11; 1903 Zur Kenntnis der Strémungen im Mantelraume der Lamelli- branchiaten. Arbeit. a. d. zool. Inst. Wien, Bd. 14. THE MORPHOLOGY OF THE FRONTAL APPENDAGE OF THE MALE IN THE PHYLLOPOD CRUSTACEAN THAMNOCEPHALUS PLATYURUS PACKARD ARTHUR T. EVANS NINE FIGURES INTRODUCTION In 1874 Dr. L. Watson collected near Ellis, Kansas, a singular Phyllopod with a peculiar frontal appendage on the head of the male, belonging to the family Branchiopodidae. These and other specimens collected by Doctor Watson were forwarded to Prof. A. 8. Packard, who described (’77) this particular species as Thamnocephalus platyurus. Since that date only two collections of the species have been recorded, one by Cockerell in 1912 from Montelair, Colorado, and one by Pearse in 1912 from La Junta, Colorado. The material studied by the writer was obtained by Dr. Max M. Ellis at St. Vrain, Colorado, on May 31, 1912. Thamnocephalus platyurus is an inhabitant of the stagnant plains-pools of western Kansas and eastern Colorado. These pools are usually very muddy and distinctly alkaline, except immediately after a heavy rain. Since the pools are not perma- nent-and are formed entirely by the collection of surface water, they are subject to a rapid concentration by evaporation. They appear in early spring, dry up during the summer, and again appear with the rains of autumn. The specimens taken at Mont- clair were found in a pool about 15 feet in diameter, the water area of which had been reduced to about 9 feet by evaporation. This pool varied from a few inches to about 8 feet in depth. About the shore was rubbish and various plains plants, as well as an incrustation of alkali. The specimens studied by the writer were collected in a pool filling the bottom of a ‘draw’ about 40 feet long and 3 or 4 feet deep near its center. The 703 704 A. T. EVANS water was very muddy and a broad incrustation of alkali ex- tended back several feet from the shore of the pool. In this pool the animals frequented the deeper parts, even the bottom. Occasionally they were seen gliding through the water at a depth of 2 or 3 inches where they were easily observed. Thamno- cephalus swims like Branchipus, with the dorsal surface of the body down, the head acting as the prow of a boat, the long thoracic appendages stroking regularly and rapidly. The ver- tical position in the water was easily changed by the elevation or depression of the broad, flat tail. Associated with Tham- nocephalus in this habitat were found Ambylstoma tigrinum, Bufo cognatus, tadpoles, snails, dipterous larvae of several kinds and beetle larvae; as well as large numbers of Phyllopods of the genera Apus, Streptocephalus and Estheria. Thamnocephalus platyurus is very similar in size and shape to Branchipus, except for the peculiar frontal appendage of the male and the broad, depressed, fin-like tail present in both sexes. The end of the tail is not forked, as in many of the other Phyl- lopods like Streptocephalus texanus, but is evenly rounded except for a slight but distinct notch in the extreme posterior part. The thickened posterior end of the abdomen occupies the central third of the tail, while the outer, lateral, portions, which comprise the rest of the tail, are thin and membranous. Each second antenna of the male is composed of a short, thick, fleshy portion, which forms the basal joint; and an outer, recurved, distal portion, composed of a more or less horny tissue. This outer distal joint is quite rigid, and when not in use, it is carried like a recurved tusk, backward under the body. All traces of an inner branch of the second antenna are lacking. In the female the second pair of antennae are long and oar-shaped, and near the distal end broaden laterally, rather abruptly, Just before the terminal point is reached. When not in use, the female carries these second antennae folded back under the head and thorax. In color these animals vary from a transparent to a milky white, specimens preserved in alcohol becoming more or less flesh-colored. The intestine is visible through the middle of the rather thick body, as a dark brown tube colored by the con- FRONTAL APPENDAGE—-PHYLLOPOD CRUSTACEAN 705 tained food material. The tail of the female bears a large, cres- centic spot on each side of the alimentary canal. These spots, which vary from flame scarlet along their posterior margin to light orange yellow (Ridgway’s ‘color standards,’ plates 2 and 3) anteriorly, cover from one-half to three-fourths of the posterior lobe of the tail. Similar spots of a smaller size and a lighter eolor occur on the tail of the male. The eighteen specimens used in this study varied in length from 15 to 38 mm. FRONTAL APPENDAGE Macroscopic structure The frontal appendage, which is present only in the male, projects in the middle line from the front of the head, directly between the bases of the second antennae. It is from one-half to three-fourths as long as the body proper, and at its base is slightly larger in diameter than the basal joint of the second antenna. The basal one-third of the frontal appendage consists of a single trunk, which then divides into two branches, each of which further subdivides into two parts about midway between their junction with the main trunk and their tips (fig. 1). These secondary branches may or may not be further subdivided, the amount of subdivision and the resulting number of tips de- pending upon the maturity of the individual. In a male 38 mm. long, the outer of the two secondary branches bore four sub- branches, of which the two proximal ones were forked; and the inner branch, three. Just below the division of each of the two large branches a small spur occurs on the inner surface of each branch (fig. 1, D). This arrangement of the different branches of the frontal appendage gives the whole a somewhat arch-like appearance when the appendage is correctly spread, since the inner branches leave the main branches at an angle and are directed upward and forward. In the water the animal opens the appendage suddenly and, when seen from above, it appears to be more or less flattened. When not in use or extended, the frontal appendage is closely folded under the anterior end of the body between the bases of the antennae, the smaller tips Sy =) op) A. T. EVANS Fig. 1 Dorsal view of frontal appendage and top of head of Thamnocephalus platyurus; A, trunk of the frontal appendage; B, main branch of the trunk; (, upper portion of the main branch with its mesial sub-branches; D, inner spur of the main trunk; #, Prinicpal branch of the main trunk; this branch is directed mesio-dorsally, and with the corresponding branch from the opposite side, forms the dorsal arched portion of the appendage; F’, forked tips of the mesial branches of the outer portion of the main branch; G, first antenna; H, eye. being more or less coiled and the main trunk closed up much like the blade of a jack-knife. The appendage is covered with a thin, flexible layer of chitin resembling a transparent membrane. The tip of each little branch bears a claw-like structure, formed by the rather sudden narrowing of the internal tissues, this nar- rowed portion tapering to a rather sharp point, the chitinous covering following its contour closely. The whole appendage is more or less roughened by numerous, small, conical projections over its surface. Examination of a number of the smaller and probably immature males showed that the number of branches of the principal branches varied greatly, but they are quite constant in the larger specimens, which would seem to indicate FRONTAL APPENDAGE—PHYLLOPOD CRUSTACEAN 707 that the smaller subdivisions of the appendage are added as the animal becomes larger (table 1). In both the young and old specimens it was found that the appendage was slightly flattened on the ventral side, which may be due, however to its contact with other parts while it is closed up under the head. TABLE 1 Showing the variation in the number of terminal branches and the apparent corre- lation of number with size, that is, age of the animal ARRANGEMENT OF ARRANGEMENT OF BRANCHES ON Cl Sen anita 5 BODY LENGTH NO. OF SPECIMEN IN MM. ey : o. de- No. de- e No. de- No. de- veloped veloping No. forked veloped veloping ame veto. ean 15 2 1 0 2 1 Dag: Migs Soa ere 20 2 1 1 3 0 Oe Sa tt Se 21 3 1 2 3 0 AAAS a cui ee se 24 3 1 2 3 0 CD AARON RR erate 38 4 1 2 3 0 1 See figure 1. Microscopic structure For the study of the minute anatomy of the frontal appendage a complete series of cross-sections, beginning with the outer tip and continuing well back into the body of the animal were made in celloidin. This series was used primarily for the reconstruction work. Thin sections of various parts of the head and frontal appendage were made in paraffin. These were stained with iron hematoxylin or safranin-gentian violet. Thick sections were stained with eosin or borax carmine. A typical cross-section (fig. 2) of a frontal appendage shows the trunk to be composed of several definite areas. % A. T. EVANS BIBLIOGRAPHY CockERELL, T.D. A. 1912 The fauna of Boulder County. Univ. Colo. Studies, vol. 9. DauHLGREN, U., AND Kempner, W. A. 1908 Principles of animal histology. New York. LANKESTER, EH. Ray 1909 A treatise of zoology. Part VIII. London. Packarp, A. S. 1877 Descriptions of new Phyllopod Crustaceans from the West. Geol. and Biol. Surv. of the Ter., Hayden Bull., vol. 3. 1878 A monograph of the Phyllopod Crustaceans of North America with remarks on the order Phyllocardia. Geol. and Biol. Surv. of the Ter. of Wyoming and Idaho, 12, Part I. Pearse, A. S. 1912 Notes on the Phyllopod Crustacea. 14th Rept. Mich. Acad. Sci. Ripaway, Roperr 1912 Color standards and nomenclature. Washington. SUBJECT AND AUTHOR INDEX BNORMAL OVA: end of the first to the end of the ninth day. The develop- ment of the albino rat, Mus norvegicus albinus. II ADDISON, Wituram H. F. and Appieton, J. L., Jr. The structure and growth of the incisor teeth of the albino rat............ ALLIS, EpWARD PHEtps, JR. The homologies of the hyomandibula of the gnathostome (NaS }s 1a seo ek eR AE Oe ee i a ge 5 Anatomy. Ciliary mechanisms of lamelli- branchs with descriptions of............. Anatomy of Heterodontus francisci. II. The endoskeleton., “Whe! ssc. ses eee ee Appendage of the male in the Phyllopod crus- tacean Thamnocephalus platyurus Pack- ard. The morphology of the frontal.... 7 APPLETON, J. L., Jk., ADDISON, WILLIAM H. F. and. The structure and growth of the in- cisor teeth of the albino rat.............. Astroscopus guttatus. A peculiar structure in the electroplax of the star-gazer,...... AUMGARTNER, E. A. The develop- ee of the hypophysis i in Squalus acan- Bdellodritas phiiadelphieus ‘The embryology Beckertirs ‘paper on “The genesis of the plasma-structure in Hydractinia echin- ata,’”’ and reply by Miss Beckwith. Com- Ment OnVMIsnees beeen eret aes ces oct e ASH, J. R.j Kepner, Wn. A. and. Cili- ated pits of Stenostoma................. y Cells. IV. Protoplasmic differentiation in the oocytes of certain Hymenoptera. Studies ODN EER yee cre ee tae es tee Characters. Chromosome studies. III. In- equalities and deficiencies in homologous chromosomes: their bearing upon synap- Bisrand: the llossioiuniiaee meee on ern a. Chromosome studies. III. Inequalities and deficiencies in homologous chromosomes: their bearing upon synapsis and the loss of Wat ChAarachersie a. eeeeee me emetic & Chromosomes: their bearing upon synapsis and the loss of unit characters. Chromo- some studies. III. Inequalities and de- ficiencies in homologous..............-...- Ciliary mechanisms of lamellibranchs with descriptions of anatomy.................. Ciliated pits of Stenostoma.................. ANIEL, J. Frank. The anatomy of Heterodontus francisci. II. The endo- Skeletons My seid ae ee: LeprererR, Pautine H. Oogenesis in Philo- Samia cyMphiareen 1.7 atm nave nee oe Deficiencies in homologous chromosomes: their bearing upon synapsis and the loss of unit characters. Chromosome studies. Tievinegualiinesiands:.snseeeeeeecse a... Development of the albino rat, Mus norvegi- eus albinus. I. From the pronuclear stage to the stage of mesoderm anlage; cue of the first to the end of the ninth day. 359 391 143 109 109 109 625 235 109 Development of the hypophysis in Squalus sacanthias;, “Rhettbs-sssenee eee ee Differentiation in the oocytes of certain Hy- menoptera. Studies on germ cells. IV. Profoplasmiciscn one nee eee CHINATA,” and reply by Miss Beck- with. Comment on Miss Beckwith’s paper on ‘‘The genesis of the plasma- structure in Hydractinia................. ‘ Electroplax of the star-gazer, Astroscopus guttatus. A peculiar structure in the.. Hiabrvolosy of Bdellodrilus philadelphicus. Endoskeleton. tus francisci. Evans, A. T. The morphology of the frontal appendage of the male in the Phyllopod crustacean Thamnocephalus platyurus Pa ckandmian ke sans stare ccm coe enon etee ERTILIZATION in Platynereis lops. Fishes. The homologies of the hyomandibula ofthe enathostomes: sc. cen oo een teens Frontal appendage of the male in the Phyllo- pod crustacean Thamnocephalus platyu- mega- rus Packard. The morphology of the... 7 AZER, Astroscopus guttatus. A pecu- liar structure in the electroplax of the Fa] Ih ih te 8 tis ak SOCIO RICA Oe CAS Germ cells. IV. Protoplasmie differentia- tion in the oocytes of certain Hymenop- PETA,» ISHUGIES OM yacr nc cnt rata soe ee Gnathostome fishes. The homologies of the hyomeandilbulajon bhesssss sess seis menace EGNER, Ropert W. Studies on germ cells. IV. Protoplasmie differentia- tion in the oocytes of certain Hymen- OPUCOLA tase sieyessttae Aculeae Cakes cutee ie poaeenetrenyaie HERWERDEN, MARIANNA VAN. Comment on Miss Beckwith’s paper on ‘‘The gene- sis of the plasma-structure in Hydrac- tinia echinata,’’ and reply by Miss Beck- at Di. eas ae cp ee eles pace nol teens imate & Heterodontus franciseci. II. The endoskele- tons he sanatomy, iole-eeer eee ee teas Homologies of the hyomandibula of the enathostome fishes. The.. Huser, G. Cart. The development ‘of the albino rat, Mus norvegicus albinus. I. From the pronuclear stage to the stage of mesoderm anlage; end of the first to the endioftbhemnthidayecn oem eee 2 Huser, G. Cart. The development of the albino rat, Mus norvegicus albinus. II. Abnormal ova; end of the first to the end Of the minthedaye sae eee peer ete ae Huaeues, James G., Jr. A peculiar struc- ture in the electroplax of the star-gazer, AStroscopus) Puttatuss. joc ce elects Hydractinia echinata,’’ and reply by Miss Beckwith. Comment on Miss Beckwith’s paper on ‘‘The genesis of the plasma- Structure wi ter isco won cece G nee mca The morphology of normal..... : 495 309 97 387 720 Hymenoptera. Studies on germ cells. IV. Protoplasmic differentiation in the oocytes Of (CeTtalM.2 aoe cceeninetcia aera lelels ware eres 495 Hyomandibula of the gnathostome fishes. The vhomologiesjotahersee eee eee aoe Hypophysis in Squalus acanthias. The de- velopmentiolthenpcee arts scnecaeencn eee NCISOR teeth of the albino rat. The structure and growth of the............. Inequalities and deficiencies in homologous chromosomes: their bearing upon synapsis and the loss of unit characters. Chromo- some: studies.) (UM teae cc nea ce cee cease je E. E. The morphology of normal fertilization in Platynereis megalops... . ELLOGG, James L. Ciliary mechan- isms of lamellibranchs with descriptions OL*ANA LOWY paca oi ciovs nase Wrolsl sadist ree: Kepner, Wm. A. and Casu, J. R. Ciliated Pits of StenvOstomMane.-e-e.-- ose ener | pecs eerie with descriptions of anatomy. Ciliary mechanisms of..... ECHANISMS of lamellibranchs with descriptions of anatomy. Ciliary.... Morphology of normal fertilization in Platy- nereispmeralops. Whe s)-c...2..+. 52-1 : OCYTES of certain Hymenoptera. Studies on germ cells. IV. Protoplas- mie differentiation in the.............-. Oogenesis in Philosamia cynthia............. Ova: end of the first to the end of the ninth day. The development of the albino rat, Mus norvegicus albinus. II. Ab- naLO) BABY Us ECG CO Defoe otha arhceraerherein rere HILOSAMIA eynthia. Oogenesis in.... Phyllopod crustacean Thamnocephalus platyurus Packard. The morphology of the frontal appendage of the male in the.. 70: INDEX Pitsiofistenostoma. (‘Ciliated!:s...2..)-.ceeer Platynereis megalops. The morphology of MOUMA LertilazatlOM iM. sain ose eeeene AT, Mus norvegicus albinus. I. From the pronuclear stage to the stage of mesoderm anlage; end of the first to the end of the ninth day. The develop- mien ofthe ‘albino. «2.7m ee eee Rat, Mus norvegicus albinus. IJ. Abnor- mal ova: the end of the first to the end of the ninth day. The development of the AI DINO oss sscye, dts Ne ee CI EE eee Rat. The structure and growth of the incisor teeth of the albino; 95. eeseneee teas Rogpertson, W. ReEES BREMNER. Chromo- some studies. III. Inequalities and de- ficiencies in homologous chromosomes: their bearing upon synapsis and the loss obnmnit characters=-nce see eee QUALUS acanthias. The development of the hy pophysis imesse sees eee ete ene Star-gazer, Astroscopus guttatus. A peculiar structure in the electreplax of the........ Stenostoma. Ciliated pits of...............- Synapsis and the loss of unit characters. Ctromosome studies. III. Inequalities and deficiencies in homologous chromo- somes: their bearing upon............... ANNREUTHER, Georce W. The em- bryology of Bdellodrilus philadelphicus. Teeth of the albino rat. The structure and PTroOwthiot the wncisoresc: ee. eee eee Thamnocephalus platyurus Packard. The morphology of the frontal appendage of the male in the Phyllopod crustacean. ... NIT characters. Chromosome studies. III. Inequalities and deficiencies in homologous chromosomes: their bear- ing upon synapsis and the loss of........ » Posed aa xi re j Mii