HARVARD UNIVERSITY Library of the Museum of Comparative Zoology /IDemolrs of tbc fIDuseum of Comparative ZooIoqs AT HARVARD COLLEGE. Vol. XL. No. 9.., STUDIES FROM THE NEWPORT MARINE LABORATORY Communicated by Alexander Agassiz. XVI. THE DEVELOPMENT OF OSSEOUS FISHES. II. THE PRE-EMBRYONIC STAGES OF DEVELOPMENT. PART SECOND. — THE HI8TORY OF THE EGG: CLEAVAGE, FORMATION OF THE PERIBLAST, AND DEVELOPMENT OF THE GERM RING. BY ALEXANDER AGASSIZ and C. O. WHITMAN. WITH ELEVEN PLATES. CAMBRIDGE, U.S.A.: |)ciiUeC) for tbe /Duseum. April, 1915. riDemolrs of tbe /iDuseum of Comparative Zoology AT HARVARD COLLEGE. Vol. XL. No. 9. STUDIES FROM THE NEWPORT MARINE LABORATORY Communicated by Alex.\nder Agassiz. XVI. THE DEVELOPMENT OF OSSEOUS FISHES. II. THE PRE-EMBRYONIC STAGES OF DEVELOPMENT. PART SECOND. — THE HLSTORY OF THE EGG: CLEAVAGE, FORMATION OF THE PERIBLAST, AND DEVELOPMENT OF THE GERM RING. BY ALEXANDER AGASSIZ and C. O. WHITMAN. WITH ELEVEN PLATES. CAMBRIDGE, U.S.A.: prlnte& foe tbe /Duseum. April, 1915. PREFATORY NOTE. The study of young fishes was one of the early zoological interests of Mr. Agassiz, and he had published a number of papers ' previous to 1883 when the late Prof. C. O. Whitman joined the staff of the Museum of Comparative Zoology. During a part of the years 1883-1886, Professor Whitman worked upon the development of some pelagic fishes both at the Newport Laboratory and at the Museum, and some of the results of his studies were published in collaboration with Mr. Agassiz in the Proceedings of the American Academy and in the Memoirs of the Museum.' Though the eleven plates for another part of the Memoirs to be ' ' devoted to cleavage, formation of periblast and the development of the germ ring" were printed so long ago as October, 1885, the accompanying text was evidently not written. The plates, however, seem of sufficient importance to warrant their publica- tion, and this has been made possible through the kind interest of Prof. R. M. Strong of the University of Mississippi, formerly of the University of Chicago. Professor Strong carried out a thorough search among Professor Whitman's manuscripts for any records upon the embryology of fishes, and has most-kindly written a brief introduction. To Professor Strong and to Mrs-. C. O. Wliitman, who most willingly aided him, sincere acknowledgements are tendered. Samuel Henshaw. ' The development of flounders. i FUHF pi ,k / PLATE XXXVI. Fig. 1. Ctenolabras. One hour afler 32-ceU stage. [p6 = peribla-st). Out of a large number of (4) caps (20) of this stage, thi.s one alone .show.s decided evidence of the fii'st beginning of the endo- derm. The cap is two to three and four cells deep in the central portion, becoming two and then one cell thick at the edge. The upper layer of cells is not dotted, and the nuclei are colored, for sake of distinction, more deeply than those of underlying cells. There are considerable intercellular spaces as seen in undotted portions, which are probably due to the cells having been slightly disturbed, or to action of reagents. Numerous vacuole-like spaces (round) varying from size of nucleus to much smaller, are seen prevailingly in the ectoderm. Histological evidence of a differentiation of endoderm from edge of cap: — 1. The edge-cells are stained with carmine — ^not browned and the nuclei are all stained with carmine, with only faint traces of browning; while the remaining cells are more or less deeply browned and the nuclei also more browned, and less brilliant than those of endoderm. 2. The radial arrangement of protoplasm is much more accentuated in the endoderm-cells than in the ectoderm, and the cells are rather more coarsely granular. We have here the best kind of evidence that the endoderm arises late, from the edge of the cap — thus from what may be regarded as the vegetative portion of cap. The endoderm is in process of differentiation, only certain cells having advanced so far as to be entirely outside the cap and with the charac- teristic absence of distinct cell-boundaries. Four cells {en 1,2,3, 4) have advanced to the syncytial stage — representing four free nuclei in the thickened rim of the pellicle (ph), around which the radial hues are very clear and strong. Most of the nuclei appear to be in a condition that precedes the formation of [mitotic] figures. These four cells are wholly within the pellicle, and hence beneath the level of the outer ecto- derm-cells. The cells marked en are cells destined to become like en 1-4, but have not lost completely, except in limited portions, their definite boundaries. Many of them are not sharply defined against each other, or against the pellicle, with which they are continuous. ec en i is a cell still in the outer rim of ectoderm, but it loses the sharp outline and blends with the pellicle and with the endoderm-cell at its right. ec en 2 is another similar cell, which still preserves its outline, but which lies in same level as en 1 . ec en 3 still holds its position in the ring of cells en, but it is not more than very faintly and imperfectly delimited from the pellicle. We have here, then, the outer ring of cells of cap in process of becoming fused with the pel- licle, only four cells of which have become undoubted syncytial cells. Fig. 2. About one hour after 32-cell stage, from above, treated in same manner and on same (16 1 slide. Here the cells are about half the size of those in Fig. 3, indicating that a single divi- sion has occurred. Here are seen the same coarsely granular marginal cells in floor of cap, but we see that they have divided tangentially, so that we have here two cells where we before had only one. The outer row of cells is more coarsely granular than the inner row, and their nuclei are at a little distance from the dark margin of the cap proper. Now I am inclined to take this outer row of cells as representing the first formed endoderm- cells; still it is possible that the inner row is also endodermic. If the outer row alone is endo- derm, then in Fig. 3, we should have to say that the endoderm has not yet separated from cap, but simply appears as a more coarsely granular part of the marginal cells. In this case Fig. 2 would correspond in age to Fig. 1, which has also about twenty cells in the marginal endoderm. Fig. 3. Ctenolabrus. Portion of cap about 30 minutes after 32-cell stage. From below. (15) Stained with osmic only, [ec = ectoderm; en = endoderm; pb = periblast]. There are about twenty cells in the periphery of the cap, forming the marginal cells of its floor and projecting a little beyond the smoothly outlined superficial cells. These cells are every- where sharply defined except against the pellicle. Their boundary line against the pellicle is tolerably distinct but not smooth. It is more or less ragged as if the delimitation was not complete. An important feature of these cells, in contrast with the finely granular cells elsewhere, is their coarsely granular nature. They are more and more coarsely granular as we pass from the inner to the outer margin. The outer half of these cells is thinnest and more coarsely granular than the inner half. Notice also that the nuclei of these cells are in the first stages of division; i. e. they are elongated radially and are faintly striated with a darker dotted line in the middle zone. The achromatic spindle and stellate rays are not to be made out, the preparation not being favorable to making out fine details of nuclear structure. These cells fit into the other floor-cells which arc shrunken away from them at this point. Fig. 4. One hour 30 minutes after 32-cell [stage], [pb = periblast). Section near middle. (87) Here the marginal cells are very distinct, but owing to method of treatment the subgerminal plate is very indistinct. Cap two cells deep at margin (sometimes two sometimes only one), three cells deep in middle. All the central cells from top to bottom are blackened alikr. Fig. 5. Transverse section near middle of 64-cell [stage], (30 minutes after 32-eell) Os and CrOa 3 (85) da vs. The marginal cells are lighter and redder than the central, the latter being sUghtly more tinged with osmic. The cells are closely packed and the subgerminal plate everywhere in contact with cells except at one point where it "is broken. This is probably due to the con- traction caused by the chromic acid. Fig fj. One hour after 32-cell = fig. 1 in age). Transverse [section] near middle. (Os and Mk. (86) 3 days), [he = blastocoele; pb = periblast). The periblast is continuous under cap in most places. The marginal cells are distinct in color. In all caps of this age, I find the central cells smaller than marginal and from two to three cells deep; the marginal are from one to two deep. pb * / «. ^ h_ 7 k pi PLATE XXXVII. PLATE XXXVII. Fig. 1. Ctenolabrus from inner surface. Os i% 15 minutes, Merk. 3 day.s. 2 hours after 32-celI (8) stage. In the wreath are about twenty-seven cells. Protoplasm radiate around some, no nuclei in division. Cap = two to four cells deep; smaller in central, larger in periphery. Round cells are seen in the central portion. en 1, 2, 3 = delimited. en 4 not delimited against pellicle. en = syncytial endoderm-cells. Endoderm colored red, cap brown. For distinction, nuclei of en are red, nuclei of cap brown (Indian red). Most of en lie under, not outside, the margin of the cap. In a number of cases the nuclei are half under cap, half exposed, [ph = periblast]. Fig. 2. Two hours after 32-celi (stage). 42 sections .00.5 mm. thick. |pb = periblast]. (88) The first section cuts edge of cap. Dividing line between marginal cells does not cut quite through the periblast. The twenty-first section through middle. Marginal cells small, but well colored red, while the central cells are all brown. The twenty-fourth section shows an inner cell that is reddish, but less so than the marginal cell. The twenty-seventh section shows no marginal cell. The thirtieth section shows red cell above the rnarginal cell. These facts are to be explained by supposing that the marginal cells are constantly dividing and adding cells to the cap, which soon after division become like the other cells of the cap, taking brown color from osmic acid. At this stage the subgerminal plate is pretty complete everywhere. It is the upper portions of the marginal cells that are constantly cut off and added to the cap, until at length the basal portion ceases to have any connection, and its nuclei spread in all directions. Fig. 3. Ctenolabrus. Two hours after 32-cell stage. \ph = periblast]. A portion of cap from (9) above. The endodermal wreath is here very distinct and contrasts with the ectoderm every- where in the staining. There are ten paii'S of cells and an older one, en '. Three of the pairs [en^, en-, en^"\ are tangential and the rest are oblique or radially placed. In en 1 the division is so nearly completed that only the most faint traces of interzonal filaments are discernible. The right hand cell of the pair is in plane above the left, and [is] more sharjily defined but colored the same. en 2 shows last end of division with filaments still visible. A light line between cells indi- cates plane of division which is not completed. Both cells continuous with pellicle {ph). en 3, [The number does not appear on the plates, but the cells are evidently those between en 2 and en /,]. Division completed, inner cell highest, and best defined, but colored lighter than rest of ectoderm. en 4, inner cell also highest but below the highest of pair en 3. Division completed and outline clear in both. en 5, same as in 4. en 6, both in same plane — faint traces of spindle. en 7, division completed. en S, faint spindle-cells not sharply outlined except against the floor-cells of cap. en 9, outlined inner cell lighter than ectoderm and appears to be paired with the outer longer cell. en to, division not completed, pretty well defined but flowing into pellicle. All the inner cells of the pairs, except in en 3, lie in the floor of the cap, and in most cases abut against surrounding cells. The paired nature of these cells is most evident. Most of the irmer cells are covered by ectoderm but wherever they are uncovered, they are colored like endoderm. iJ I-/ Mi %# v'^iimnjn ;ei PLATE XXXVIII PLATE XXXVIII. Fig. 1. Cap of Ctciiolabrus. [pb = periblast). Osmic 15m. (3^%), Merkel 3ds. Same age a.s [in (2) Fig.] 2, but not quite so far advanced, i. e. the nuclei of [marginal] row are somewhat elongated, and in a few cases faint asters are seen at the poles of the nuclei, but no striation is visible in the nucleus. The entoblastic wreath is characteristically granular and thins out rather abruptly into the pellicle, just beyond the nuclei. The cells on both sui-faces of cap were distinct. The cell- boundaries around these [wreath] nuclei were visible here and there, but very faint as shown in figure. Now and then a nucleus of a second inner row — just under edge of blastoderm — is seen. Here also the nuclei are mostly elongated in a radial direction. Color: — Wreath and pellicle colored slightly with carmine, but the cap browned with osmic. The contrast between granular pink wreath and light brown cap is striking here as in all cases thus treated. Fig. 2. Cap of Ctenolabrus. [pb = periblast). Seen from outer surface three hours after the .32- (1) cell stage is reached. To the cap adheres the entoblastic wreath, which is thickest at the edge of cap, gradually thinning out towards the periphery, which is ragged in consequence of being broken from the pellicle. This wreath is stained faintly with carmine (borax alcohol) while the cap is browned by osmic acid and not stained by carmine. The wreath reaches under the cap for a short distance but not to the central parts of cap. There is a single row of nuclei, radially placed (some tangential and oblique), and all in process of division, mostly showing a well-marked nuclear plate in different stages of division, and asters (rather faint). The cell outlines about the nuclei are not visible, except in a few instances. Most of the nuclei of the ectoderm are round; one is seen in division, and beneath it is a nucleus of the endoderm-wreath in division. Although there is, broadly speaking, only one row of nuclei visible, an inner row makes itself apparent in several places and now aud then a nucleus (dividing) may be seen just under the edge of the cap. ^•^.H* :^A ^4?\** <^ ^•, V — '« ' -®-^ PLATE XXXIX. PLATE XXXIX. Fig. 1. Ctouolabrus, \mdei- (inner) .surface. Treated as Figs. 1 and 2, [Plate XXXVIII), four (3.) hours after 32-cell stage. One hour after Figs. 1 and 2 (Phite XXXVIII). In this cap the nuclei of the endoderrn are strongly colored and very well defined. One can see by the nuclei that the endoderm extends for only a short distance under the cap. The majority of the nuclei are outside the cap. Fig. 2. [Seven sections]. Three liours after 32-cell stage, [be = blastocoele; pb = periblast]. (89) (Os, Mk. 3ds). Second section shows oidy one or two outlined cells, the outlines being indistinctly marked. The third section shows two well-outlined cells; the right is sharply marked inferiorly. There is a partial outline of a dark cell of the caj). The fourth section shows the first appearance of the sub-germinal plate. All the red cells are shaded by dotting. The sixth section shows the subgerminal plate thinner. The sixteenth section shows on the left a red cell that looks as if it was to enter into cap. The twenty-sixth section is near middle. Here sometimes no nucleus is seen in periblast, at other times one or even two. Subgerminal plate very thin, wavy in outlines as in section sixteen. The twenty-eighth section shows an inner cell in process of splitting off, possibly destined to become one of the cap cells. The thirty-third section showed a similar case. I think it is possible that cells are still added to the cap, l.)ut that this process is nearly con- cluded, so that the periblast as a cell layer may be now considered established. Fig. 3. [Later stage. 6c = blastocoele; ep = epidermis; p6 = periblast]. Section near middle. (90) The periblast is very thin, vanishing or nearly so near the center of field. Epidermis well marked off. On left is a single periblast cell that looks as if it was a cap-cell. The fact that the periblast becomes so very thin indicates that cells are added to the cap up to about the time the rfng begins to form. The nuclei of the periblast are still confined to the thickening beneath the margin of cap; on the left [i. e. lower end of figure] a single nucleus is somewhat advanced from the margin towards the center, but this is exceptional in these sections. \l. AND WHIV J 5 •' .* 1^ J> :J m :C'CS '/'* -^S^ 3S» %":" '-W^ O" .® o €) 3'4 PLATE XL. PLATE XL. Fig. 1. P.[araHchthys] oblongus. [be = blastocoele; ep = epidermis; pb = periblast]. 61 sec- (92) tioiis. This figure is the twenty-fifth [of sixty-five sections] and is hke the middle ones. The periblast is thin but extends farther under the cap than in Fig. 2. Near the center the layer fades out and is not recognizable, but four nuclei are very distinct, showing that the layer exists though thin and perhaps not fully differentiated. The nuclei extend under the entire cap although the ring has scarcelj' begim. This shows that the periblast has already begun to expand as an independent layer and has therefore probably ceased to contribute directly to the cap. F"ig. 2. P. oblongus. [be = blastocoele; ep = epidermis; pb = periblast). Three hours after (91) 32-cell stage. Four of sixty nearly longitudinal sections. Left end = embryonic region. The second section does not reach cap; third touches cap. The fourth section shows some of cap, and periblast thinner at middle. The eleventh section shows periblast still thinner in middle. Nuclei arc more numerous at left end all through these sections. I think the pcriblastic nuclei are more numerous under region of embryo than elsewhere. On the seventeenth section the periblast is scarcely visible at center beneath cap. On the sixteenth two nuclei were found near the middle of the subgenninal plate, although the plate is scarcely traceable. One nucleus was seen near middle of the nineteenth section and one in twenty-fourth. The periblast is somewhat less in bulk on the middle sections than on the thirty-eighth which I have drawn. Periblast is here about same in quantity as in Fig. 2, Plate XXXIX. (Three hours after 32-cell). AcASSiz AXD Whitman Fi; k.=:s .-^^ v-/-; PLATE XLI PLATE XLI. Fig. 1. (P. oblongus) irom above. Very early stage of ring, [pt = periblast]. The ring has just (96) become well defined everywhere and the embryonic fold is seen somewhat bulging centripetally. The ring is nearly even in width everywhere, except for being a little wider as it enters the embryonic region. Width of ring = .02 mm.; in embryonic region = .04 mm. Cap almost perfectly circular. Diam. of cap = .55 mm. The thick wall (wreath) of periblast is not yet fully covered so that about one row of nuclei is seen all around. The number of nuclei could not be accurately determined, so they were put in at random guided by what was known of other cases. The outlines of only superficial cells (epiblast) are seen, but nuclei of deeper cells are given. The epiblast is seen in profile along the margin. The posterior half of the cap is thicker than the anterior, and this is shown by heavier dots and lines. The cap should appear convex — but the lithographer could [not] be instructed to shade it properly. The posterior [lower half of figure] half is the embryonic area, the anterior [upper] the pre- embryonic area. The ring is now only .02 mm. [in width], later it becomes .12-. 15 mm. in width. Fig. 2. [Section of] cap before appearance of ring, cleavage cavity shallow. (97) Fig. 3. [Section] just before rmg. [6c = blastocoele; pb = periblast]. Embryonic area thicker (98) than pre-embryonic area. (Large yolk cells supplied). Fig. 4, [Section, be = blastocoele; pb = periblast]. Blastoderm thinned out much in anterior (99) half, where for some distance the lower layer is only one cell deep. Towards the anterior edge the under layer thickens. PLATE XLII. Fig. 1. P.ohlongus. Sections drawn at intervals. Young ring .stage cut transversely into sixty- (94) three sections (.0075 mm.). [6c = blastocoele; ep = epidermis; p6 = periblast). Diagram 12 = longitudinal median section, constructed from the transverse sections. .0075 mm. X280 equals apparent thickness of each section. This amounts to 2.1 mm. 2.1 X 63 gives apparent length of the cap, which is 132.3 mm. Measuring off 132.3 mm., I divided it ^ into sixty-three equal parts, and then by vertical measurements of each section constructed the diagram. Thickness of upper layer in the sections. (Apparent of magnified 280). 8th section = 15 mm 10" = 16 " 12" = 15 " 13" = 15 " 14" = 14 " 15" = 15 " 16" = 13 " 17" =1 14 " 18" = 13 " 19" = 14 " 20" = 14 " 21st = 13 " 22 d = 13 " 23rd = 13 " 2ith = 12 " 25" = 12 " 26" = 13 " 27th section = 11.5mm. 28 " = 11 29 " = 10 30 " = 10 31st = 10 32d = 10 33rd " = 10 34th = 9 35 " = 8 " 36-50th " = 8 53d = 10 54th = 9.5 " 55 " = 9 56 " = 10 59 " = 10 60 " = '8 61st = 8 The second section takes in many epiblast cells, which are seen partly in surface owing to the convexity of the cap near margin. Nearly all the cells are in Home atage of diinsion throughout all the sections. Notice that the nuclei of the periblast are few in section two, more in three, most in five and are fewer towards the middle of cap (37th). They are more numerous beneath the embryonic fold than elsewhere under the ring. The lower layer becomes distinct only in section five, where we see not only a line of division, l)nt notice that the nuclei retreat from this line somewhat. The lower layer is thickest (3-4 cells deep) in the embryonic fold; elsewhere the ring where it is well begun is 1-2 cells deep. The ring is seen as a true infolding in the thirty-second section and especially in the thirty-fifth. In some points the ring is scarcely begun, i. e. there is no distinct ring. The axial portion of the ring ceases with the tenth section, being thus less than ^ of the entire length of the caj). The upper layer is thickest over the embryonic apical fold, being here from 4-6-ce!ls deep. This thick area thins off grad\ially beyond the twenty-fifth section {i. e. between twenty-fifth and thirty-fifth). From the thirty-fifth onward the diminution in thickening continues but in a much less marked degree, dwindling down to about half the thickness of the embryonic fold. The thickness of the upper layer is nearly uniform in all transverse sections from margin to margin. The upper layer is composed of an epidermis and a deeper neural layer (neural at least in its apical region). The epiblast is composed of flattened cells that abut against the highest angle of the periblast. In no case does the ring appear to arise by an infolding of this laj'er; on the contrary, there is sometimes a small space left between the marginal epiblast cell and the nearest cell of the ring (so in 27th). The deeper neural layer, however, bends into the infolding lower layer as .shown in the twenty-seventh, thirty-second, thirty-fifth, and thirty-seventh. The nuclei may be seen in all stages of division in the ring region. The number of cells in a section of the ring varies from two (32nd) to four (27th, 37th) . In the floor, the ring is one cell deep (except in apical region — see above) or sometimes two cells deep (most often at the inner edge). Later it is plainly one cell deep. The cleanage-canly begins with the tenth section and ends with the sixtieth. The periblast is thicker under the apex than at the opposite point. Section sixty is much thinner than section five (both are at the boundary of the cleavage- cavity or at a point where the lower layer begins to be distinct from the upper layer). The periblast plainly does not contribute elements to the ring; it is more difficult to decide about the axial portion. In some sections, notably in the axial portion, I find cells sometimes more or less delimited in the periblast (3rd, 5th, 6th, 8th). This delimitation does not, how- ever, show that periblast cells pass into the cap, as such are visible at certain stages of the nuclear transformations. I am inclined to think — not certain — that the periblast does not enter into the ring at any point. AgASSIZ. AMD \'. PL .«.!!, 5* ««1E7 ^^r fib 81k w> 6* pb soa ' -V 3"'' a a * t * jp 12. 3e ss 3r fl 2 'a J* ^ 9 to \ V V \ \ \ \ \ \ NOV 9 1983 'lOO CAMBRIDGE STREET CHARLESTOWN, MASS. Harvard MCZ Library I II iiiiiiiNii II I !