THE DEVELOPMENT OF THE SQUID.

The mode of development of the common Squid is essentially like that of Sepia and Grenacher's Cephalopod, but in minor points it is different from each of them, and in many features it seems to be intermediate between them, and although we have a number of figures, by various writers, of stages in the development of Loligo, there still remains a need for a continuous account of its history as a basis for comparison with other Cephalopods, and I hope this paper will furnish part of the necessary material for a general discussion of the subject.

The mantle, m, fig. 9, is now quite sharply defined, and its posterior edge, m, fig. 10, begins to overhang a little, arching over the rudimentary mantle cavity, which is thus seen to be formed in Loligo, as in most Gasteropods and Lamellibranchs, by the outgrowth of a flap of integument from the surface of the body. A reference to Grenacher's figures will show that the origin of the mantle cavity was quite different from this in the form which he studied. At a corresponding stage (see his figs. 6 and 7), the mantle area is much larger than in Loligo, and covers almost one half of the body of the embryo, and is more developed externally than it is in Loligo at the stage shown in our fig. 15. The mantle cavity, however, is barely indicated, and when it makes its appearance it is not formed, as it is in Loligo, by the outgrowth of a mantle-fold, but by the involution of the integument under the mantle, after the latter has spread over a considerable area of the body. The resemblance between the manner in which the mantle cavity is formed in most Gasteropods and Lamallibranchs, and the way it is formed in Loligo, seem to indicate that the latter presents the primitive mode of development among the Cephalopoda; and the way it is formed in Grenacher's species must, therefore, be regarded as a modification of the primitive mode which has been retained by Loligo.

It is interesting to notice that there are other groups of Mollusca, among which the mantle cavity is produced in more than one way. In Cyclas, Pisidium, and many other Lamellibranchs, a fold grows out from the body, and thus converts the space between the fold and the body into a mantle cavity precisely as in Loligo, but in Anodonta the shells are quite well advanced in their development before the mantle cavity is formed, and when this is produced it is formed by the retraction of the body wall into the space between the shells; and we have in Anodonta almost precisely tile same modification of the primitive history which Grenacher describes in the Cephalopoda. In the oyster again, we have tile mantle cavity produced by a process which is about midway between that met with in Cyclas, and that presented by Anodonta.

Projecting from under the overhanging edge of the mantle, on the posterior surface of the body, at the stage shown in figs. 9 and 10, are the rudimentary gills, g, one on each side of the middle line. They are formed as little papillae, or outgrowths from the surface of the body, and are covered with cilia. In position and method of formation they are very much like a single pair of the gill-tentacles of a Lamellibranch embryo, and their embryonic history would seem to indicate that they are to be regarded as greatly specialized gill filaments, rather than structures comparable to the entire gill of a Gasteropod or Lamellibranch.

The shell area in the middle of the mantle is now a deep pit, fig. 9, s, and its edges have begun to fold towards each other, as shown in fig. 10,, s. As Lankester has pointed out, it originates in exactly the same manner as the shell-gland of a Gasteropod or Lamellibranch, and is strictly homologous with these structures.

The ridge, a, of figs. 6 and 7, is now divided into three arms, fig. 9, a, upon each side of the body, about half-way between the mantle and the opposite pole of the egg, and therefore much nearer the mantle than at the stage of fig. 6. The yolk is now entirely surrounded by the blastoderm, and has departed still further from the regularly curved shape of fig. 1. The prolongations into the mantle and the eye-stalks are well defined, and the portion of the yolk contained within the body of the embryo, which is not quite half the whole, is separated by a well-marked constriction, just dorsal to the arms from the remainder, which is now nearly spherical. The thin layer of blastoderm which covers this external portion of the yolk is split into two layers, separated from each other by a cavity which is largest along the median plane of the body, and which is traversed by a few branched corpuscles, by the contraction of which, rhythmical waves of the outer layer are set in motion on the surface of the yolk. The fact that the mantle is dorsal to the row of arms at this period is worthy of notice.

The three pairs of arms, a, are much elongated, and begin to bend away from the surface of the yolk, which is now divided into three well-marked regions; the external yolk, y'; the portion within the head region and eye-stalks, y"; and the portion within the body and mantle, y"'. During its development the embryo has undergone an increase in size, and although the drawing is less enlarged, the embryo shown in fig. 12 is actually much larger than that shown in fig. 10. The external yolk sac shares in this growth, and is very much larger at a somewhat later stage than the whole egg was at the beginning of the process of development.

The points of difference between Loligo, Sepia, and Grenacher's embryo are more conspicuous at this stage than at any previous time.

Kolliker's figures 2O, 21, 23, of taf. 11, represent about the same stage as our fig. 12, and a comparison will show a remarkable difference in all the details of structure, while the general plan remains much the same.

The body of the embryo is very much smaller as compared with the external yolk, and is not separated from this by a constriction, but is flattened down upon it. The blastoderm does not yet cover the yolk, but only extends a short distance beyond the circle of arms. Five pairs of arms have made their appearance in the Sepia embryo and only three pairs are present in Loligo. The outer and inner siphon folds of each side have been united to each other from their first appearance in Sepia, but in Loligo they do not unite until after the two inner folds have united on the middle line to form the mouth of the siphon. The mantle and mantle-cavity of Sepia are about like those of Loligo. Grenacher's figs. 8 and 9, taf. 4O, are in about the same stage of development; and comparison will show that while this form is more like Loligo than it is like Sepia, the differences are still numerous and considerable.

The external yolk is very small, and hardly projects beyond the tips of the arms. Grenacher, p. 454 is inclined to regard it as entirely wanting, but his " stirntheil," in figs. 7 and 8, is clearly the same as the external yolk of Loligo, and this again is obviously the same as that of Sepia. His argument that since the internal yolk of Sepia is divided into three regions, three divisions must be found in the internal yolk of his form, and that the "stirntheil" being necessary to make up the three, must therefore belong to the internal yolk, cannot carry much weight; for at this period the internal yolk of Loligo is only divided into two regions, although a third afterwards becomes recognizable. As regards the external yolk, then, Sepia, Loligo, and Grenacher's embryo form a series, with Loligo as the intermediate form. Grenacher's embryo differs from Sepia, and agrees with Loligo, in having only three pairs of arms at this time. The eyestalks are about equally prominent in the three forms, and the period of greatest prominence is about the same in all; but while the prominence began to be conspicuous in Loligo and Sepia at about the same time, its development is retarded in Grenacher's embryo until some time after the siphon folds have all united. The outer and inner siphon folds unite with each other in Grenacher's embryo sometime before the two inner folds meet on the middle line; while in Loligo, the two inner folds not only meet but unite with each other before they join the outer folds. In this respect Grenacher's embryo is intermediate between Loligo and Sepia, since the two sets of folds are united with each other from the first in tile latter form. As it has four separate folds at one time, Grenacher's embryo would seem to resemble Loligo more than it. does Sepia so far as the format on of the siphon is concerned.

At the stages shown in figs. 8 and 9, the mantle of Grenacher's embryo covers almost half the body, and its surface is thickly set with chromatophores, while the mantle cavity is very shallow, and hardly extends beyond the edge of the mantle. In both Loligo and Sepia, the mantle is much smaller, the chromatopllores are wanting, and the mantle-cavity extends nearly or quite to the dorsal extremity of the body.

The fins of Sepia and of Grenacher's embryos make their appearance at about the same time, and are developed later than they are in Loligo.

The fact that three such closely related forms, belongin~ to the same group of (:ephalopods, differ from each other in the order of development of all organs of the body, except the otocysts, is a striking comment upon the assumption that the order of development of parts, as observed in a single species, can give any information as to their morphological importance, or as to the order in which they were acquired in the evolution of the group. We must notice, too, that the series which these three forms furnish when any one organ is taken as a basis of comparison, may be quite different from that given by another organ, and it is clear, without comment, that we have no information at present which will allow us to generalize upon such questions as to which of these forms recapitulates the phylogenetic record most perfectly. There are a number of interesting points, however, in which all three agree. In all, the halves of the siphon are separate at first, and all pass through a stage in which the siphon resembles that of a Tetrabranch. In all, the eyes are at first open pits, like those of the Nautilus, and in all they are at some time carried upon long stalks. In all there is a time when only four of the five pairs of arms are present.

The walls of the otocysts, er, have grown thin, and their cavities have greatly enlarged; the otoliths have made their appearance, and the two chambers have begun to move towards the median line, under the end of the siphon.

The external openings of the otocysts have become constricted to long, tortuous, ciliated ducts, which are not visible with a low power, and are not shown in the figure.

The eye-stalks, es, are of about the same relative length as in the last figure, but the yolk prominences which have filled them up to this time are now almost entirely withdrawn or assimilated, and the cavity of the eye-stalk is nearly filled by the ball of the eye, e, the optic ganglion, ga, and the white body, wb.

The arms have lengthened, and suckers have appeared upon the longest pair, a", and a new pair, a', have made their appearance upon the posterior or siphonal surface of the body. In Grenacher's embryo this pair of arms make their appearance as buds upon the next pair, cc", but in Loligo they are distinct from the first, and make their appearance as elevations of the integument upon the surface of the body.

The gills, g, have lengthened considerably, and are divided by constrictions into a series of enlargements, the dorsal one being much larger than the others, and becoming the branchial heart. The inner and lateral folds of the siphon have completely united with each other, and at the point of union the siphon is also united to the body wall, and the retractor muscle of the siphon; sin, now runs back to unite with the inner anterior surface of the mantle. The otocysts have almost met each other upon the median line, under the siphon, and their walls are now very thin. The eye-stalks, es, are most prominent at this stage, and soon begin to disappear.

A comparison of this figure with Grenacher's figure 10, which represents about the same stage of development, will show that certain organs are more accelerated in Loligo than in Grenacher's species, while others are retarded. The yolk sac of Grenacher's embryo has disappeared, while it is as large as ever in Loligo. The mantle is much further developed in Grenachrer's embryo, and covers the whole base of the siphon, which is almost entirely uncovered in Loligo, even when the animal is retracted into the mantle. The posterior or siphonal pair of arms, a', are well developed in Loligo, and just making their appearance in Grenacher's embryo as buds from the next pair of arms.

The mantle now covers about one-half the entire length of the embryo, exclusive of the yolk- sac, and the neck-cartilage, nc, has made its appearance, forming a support for the edge of the mantle, on the middle line of the anterior surface of the head. The posterior surface of the mantle is now pretty well covered with chromatophores, which at this stage possess remarkable power of expansion and contraction, and render the living embryo a very beautiful and wonderful sight under a low magnifying power. They are, as yet, entirely absent from the anterior surface of the mantle.

About this time small polygonal areolations, much like epithelial cells, begin to make their appearance on the posterior surface of the mantle, and soon spread over the whole mantle, except the middle line of the anterior surface, as shown in the figure. At a later stage, figs. 17 and 18 they cover tile head and arms as well as the mantle, and still later they make their appearance upon tile surface of the siphon.

Upon cursory examination, they resemble epithelial cells so much that they might readily be mistaken for them; but when more carefully examined with a high power, they are seen to be due to the presence of minute branching tubes, which, spreading over the surface of the body and inosculating, divide it up into small polygonal areas.

No fluid could be seen to circulate in them, but as they appear at about the same time with the larger blood-vessels of the surface of the body, they are probably the indications of a system of capillary vessels.

At the stage shown in fig. 15, the siphon has substantially its adult form, and is made up of two lateral chambers, si', which have been formed from the lateral siphon folds, and which open into the mantle-chamber, but have no external openings; and a single median chamber, si, on the posterior surface of the body, which has been formed by the union of the two inner siphon folds, and which opens into the mantle-chamber as well as externally.

At the point where the lateral chambers meet the median chamber, the wall of the siphon is united to the wall of the body, and the three chambers are thus shut off from communication with each other.

The animal is so perfectly transparent that the valve-like action of the two outer chambers can be perfectly seen, as their free inner edges are thrown out against the mantle so as to close it at each contraction, and the water, which passes in around the whole free edge of the mantle, is thus concentrated in the funnel-shaped middle chamber of the siphon.

At about this time the valve of the siphon, figure 15, v, is developed as a single unpaired flap, which arises from the posterior surface of the neck.

Considerable change has now taken place in the shape of that portion of the yolk which is contained in the head. It is reduced to a long narrow tube, y", which connects the portions contained in the body proper, y"', y"", with the external yolk sac, y'. The pulsatile space, x, between the outer wall and the surface of the yolk sac, is more plainly shown in this figure than in the preceding ones, although a profile view shows it with equal distinctness at earlier stages.

The figures of the later stages hardly call for much sufficiently well in themselves the gradual the form and structure of the adult.

Fig. 16 is a posterior view of an embryo a little older than the one shown in fig. 15. A large rounded prominence on each side of the bed marks the position of the eye-stalk, and the eyes are farther forward than they are in older specimens, but in other respects the form is very similar to that of the adult. The ink sac, i, has appeared, and is filled with ink, and tile tip of the free portion of the rectum is prolonged at its corners into the pair of ear-like anal valves.

Figs. 17,18, are copies of drawings made at Mr. Agassiz's laboratory at Newport during the summer of 1876, from specimens which were found swimming at the surface of the water of Narragansett Bay. As the arrangement of the chromatophores is like that of the other embryos, they undoubtedly belong to the same species and are given here to show the later stages of development. The external yolk-sac has almost disappeared, in fig. 17 and it is entirely wanting in fig. 18.

There are considerable individual variations in the arrangement of the chromatophores, but there are certain features which were observed in all the specimens, and which seem to be constant .

The first which make their appearance are dark brown in color, and are placed in a ring of six or seven (plate 2, fig. 14), around the edge of the mantle on the posterior surface. They are a little smaller, and somewhat more excitable than those which appear subsequently, and they can be readily recognized in the later stages shown in figs. 15, 16, 17, 18. They are soon followed by larger spots of the same dark brown color, scattered irregularly over the posterior surface of the mantle (fig. 16).

The next spots to appear are upon the arms, and are also dark brown. At first there are two upon the first or siphonal pair of arms' and three upon the second pair (fig. 16). A fourth soon appears upon the second arm, and these four remain conspicuous until quite a late stage of development (fig. 18). Three large brown spots now appear upon the posterior surface of the head (fig. 16), and they are soon followed by others (fig. 17).

A second set of spots, more deep-seated and of a bright orange color, soon make their appearance, and are much more constant in position than the brown ones. The first pair which appear are just in front of, or ventral to the eyes. They are soon followed by a single one on the middle line of the head, at the bases of the first pair of arms, and another single one on the middle line of the edge of the mantle. About the same time a pair appear dorsally to the eyes, and another pair on the edge of the mantle, near the sides.

Four small orange spots next appear upon the second pair of arms (fig. 17, a"), alternating with the four larger brown spots, and soon after a ring of six or eight orange spots appears on the mantle, dorsal to the ink bag. Two orange spots next appear upon the first pair of arms, figure 18, a', alternating with the brown spots.

I was not able to procure specimens in sufficient numbers to follow the development of the spots to any later stage than this, but they do not seem to develop regularly at the later stages.

In a general view of the facts which have been detailed, the most prominent and conspicuous feature seems to be the remarkable directness with which the embryo develops into the adult, and the total absence of anything like a metamorphosis. Everything which does not contribute to the formation of the adult animal has been dropped out of its life history. With the exception of the velum and the eye-stalks, there are no larval or rudimentary organs, and the whole process of development is wonderfully direct.

When we bear in mind that the Cephalopoda are almost the most highly specialized of Invertebrates, and that they must have had a long and complicated phylogenetic history, I think we must acknowledge that tile embryonic record has been simplified to a degree which is without a parallel in the animal kingdom, and it is hardly too much to say that the ontogenetic process furnishes us with no knowledge whatever of the phylogeny of the group.

The method of formation of the shell-area and of the shell; the mode of origin of the mantle and of the mantle-cavity, and the form and position of the gills of the Cephalopod embryo are all of them features which show closer relationship to a typical Gasteropod than could be inferred from the condition of these organs in the adult. They thus show the affinity of these groups to each other, and help us to a clearer understanding of tile homology between the organs of the Cephalopod, and those of a typical Mollusc, but this is about all. I do not see th at they furnish any basis whatever for speculation upon the origin of the Cephalopod, or give us the least information as to the manner in which its peculiar specializations of structure have been brought about.

In a word, the case is hardly too strongly put, by the statement that the developing Dibranch has so completely lost all ancestral features that no traces of them remain.

Fig. 20, plate 3, is the embryo of a fresh-water Pulmonate, in what I regard as a similar position. The foot, f, is below; the shell, ski, above; the mouth, m, the tentacles, t, and the head on the right, and the rectum, r, on the left. The figure therefore shows the right side of the body of the Gasteropod; and as the twist by which tile rectum and the mantle-chamber are subsequently carried around the right side of the body on to the anterior surface has not yet appeared, the embryo is bilaterally symmetrical, and the mouth, the anus, the foot, the mantle, and the shell occupy substantially the same positions, and bear about the same relations to each other, that they do in a Lamellibranch. A comparison with fig. 19, or with fig. 1O, will show that there is, at the same time, an essential similarity to a Cephalapod.

When the corresponding surface of the body of the Cephalopod embryo, at the stage shown in plate 2, fig. 10, is carefully examined with the highest power which can be brought to bear upon such a large rounded opaque surface, a well defined line or groove, plate 3, fig. 19, v, will be found to run from the median line out upon the sides of the body, bending up towards the dorsal surface in such a way as to partially encircle the eyestalks; and the motion of floating particles shows the presence of cilia along the line.

- BEAUFORT, N. C., July 31, 1880.

	1830 ANNIVERSARY MEMOIRS OF THE BOSTON SOCIETY OF NATURAL HISTORY 1880 THE DEVELOPMENT OF THE SQUID. LOLIGO PEALII (LESUEUR). BY W. K. BROOKS. BOSTON PUBLISHED BY THE SOCIETY 1880.
fig 1	ALTHOUGH several authors have recorded important observations upon various points in the embryology of the Cephalopoda, our knowledge of the outlines of the process of development, as a whole, is almost entirely derived from the accounts which have been given by Kolliker and Grenacher. Kolliker's paper contains an elaborate and exhaustive account of the external changes which are undergone by the developing embryo of Sepia officinalis and Grenacher gives an equally valuable and complete history of the embryo of an unknown species of Decapod Cephalopod. Although these two forms of embryos are substantially alike in all essential particulars, they differ so greatly in all the details of the process of development that similar outline sketches of other Cephalopods are greatly needed, and until they are furnished we shall not be in a position to discuss the relation between the young forms of this group and the embryos of other molluscs, or to speculate upon the manner in which the peculiarly direct and complex form of development exhibited by the Cephalopoda has originated. The mode of development of the common Squid is essentially like that of Sepia and Grenacher's Cephalopod, but in minor points it is different from each of them, and in many features it seems to be intermediate between them, and although we have a number of figures, by various writers, of stages in the development of Loligo, there still remains a need for a continuous account of its history as a basis for comparison with other Cephalopods, and I hope this paper will furnish part of the necessary material for a general discussion of the subject. The eggs and embryos which are described were, with two exceptions, obtained at the Chesapeake Zoological Laboratory of the Johns Hopkins University, during the summer of 1879, by dredging in water five or six fathoms deep, in the lower part of Chesapeake Bay. They were collected through the aid of the steamer Lookout of the Maryland Fish Commission, under the command of Major T. B. Ferguson, of the United States Fish Commission, to whom I take pleasure in expressing my thanks for this and for much more valuable assistance which our party has received from him. Two of the figures (17 and 18, plate 3) are from embryos which were obtained, during the summer of 1876, at Mr. Agassiz's laboratory at Newport. The youngest egg which I observed is shown, magnified eighty diameters, in Plate 1, fig. 1. It is surrounded by a well-defined wall or egg shell a, outside which was the gelatinous matter of the egg capsule. The egg shell is transparent, elastic, and is reflected inwards a little at one end, around an opening m, which may probably be regarded as a micropyle. As the embryo grows, the space inside the egg shell enlarges, and the shell itself is stretched and pushed out, away from the embryo, as if this were hollowing out a place for itself in the surrounding gelatinous matter; but as both the outer and inner surfaces of the egg shell are well marked and clearly visible, there can be no doubt that it is a real membrane, and the albumen of the egg capsule must therefore either be absorbed through it or drawn in through the micropyle. One axis of the egg is nearly twice as long as the other; the micropyle is situated at one of the ends, and the two ends are alike in outline. The yolk y, is elongated in about the same proportion with the egg shell, and the long axis is nearly parallel to that of the egg shell, but its surface is not concentric with the inner surface of the latter, since one end of the yolk is much more rounded than the other. The pointed end is the one upon which tile blastoderm makes its appearance is invariably turned towards the micropyle of the egg shell. The space b, between the egg shell and the surface of the yolk, is filled by a clear, transparent, albuminous fluid, which does not, at first, present much difference from that outside the egg shell, but as the embryo develops, the albumen inside the shell gradually becomes liquid. A number of small, spherical, highly refractive particles were visible in this fluid, and they were especially abundant around the micropyle, and on the surface of the yolk. Owing to the spherical shape of the egg I was not able to use a magnifying power of sufficient strength to determine whether they were spermatozoa or not. The yolk is transparent, highly refractive, and with a low power it appears perfectly homogeneous, but with higher powers it is seen to be packed with oil drops of various sizes, with very faintly marked edges, and a refractive index about the same as that of the more fluid portions of the yolk. At the earliest stage which I found, the blastoderm c, was well developed, and was perched, like a cap, upon the pointed end of the yolk, which, as already stated, is the end which is nearest the micropyle. The cells of the blastoderm are not very well marked in a living egg, but when treated with borate of carmine fluid, to which a very small quantity of one-tenth per cent. solution of osmic acid has been added, they become very conspicuous. Figure 4, plate 1, represents the edge of the blastoderm of the egg shown in figure 1, after it has been thus treated. The centre of the germinal area is occupied by a number of small spherules which are irregularly spherical, and each of which contains a very large nucleus. As we pass from the centre of the cap towards the periphery, the spherules become larger, and at its growing edge they are replaced by large flattened pyramids b, b, which radiate out on all sides, upon the surface of the yolk a, and gradually pass into the surface of the yolk, without any distinct boundary at their outer ends.
fig 2	Careful examination shows that the segmentation spherules are pretty regularly arranged with reference to these pyramids. Just inside the broad inner ends of the pyramids there is a ring of large spherules, c, equal in number to the pyramids, and presenting every indication of having been just formed by the separation of the proximal end of each pyramid from the larger distal portion. Inside these there is a second ring of spherules, d', about half as large, and exactly twice as numerous as the first set, and so placed that a pair of the spherules of the second set are pretty nearly in a straight line with one of the first set and the base of a pyramid. Each pair of this set is obviously the product of the division into two of a spherule like those of the set c, formed by separation, somewhat earlier, from the end of a pyramid. Inside this set is another series, et", equal in number to the set d', and arranged like this set, in pairs along the radii which end in the pyramids. Inside there is another set d"', of the same kind, so that, as we pass inwards in the line of each pyramid continued, we have the following series: 1, the pyramids; 2, one large spherule, c; 3, two spherules, d'; 4, two spherules, d"; 5 two spherules, d"'. I was not able to trace the arrangement any farther on account of the spherical shape of the egg, but the last series, d"', was some distance from the centre of the blastoderm, which appeared to be occupied by somewhat smaller spherules than d"'. Although the earlier stages of segmentation were not observed, the phenomena presented at this stage are sufficient to show that, as Lankester has stated (Observations on the Development of tile Cephalopoda), the egg of Loligo undergoes substantially the same form of segmentation as that described by Kölliker in the case of Sepia. My figures l and 2 obviously represent an egg a little older than Kolliker's figure 6, taf. 1, (Entwickelungsgeschichte der Cephalopoden) as copied in Bronn's Klassen und Ordnungen des Thierreichs, Band III, taf. cxxiii, fig. 13. Lankester states that the segmentation is not as regular in Loligo as it is in Sepia, but the arrangement of tile spherules in our species seems to be perfectly regular, and the only important differences from Sepia are the greater relative size of the nuclei, and the separation of the radiating pyramids from each other by areas of unsegmented yolk, figure 2, a, a, a, in Loligo, while the pyramids are almost in contact with each other in Kolliker's figures, as copied by Bronn. In an optical section of the egg at this stage, the blastoderm, figure 2, c, is seen to be formed of a single layer of large spherules, resting directly upon the yolk, d, from which they are sharply separated in the centre, but less sharply at the edges. In the species which Lankester studied, the blastoderm at about the same stage as our figure 1, is several cells thick at its edge, (see Lankester's figure 1, x), but this is not the case in our species.
fig 3	Figure 3 is a somewhat older egg in the same position as figure 1; a, is the egg shell; b, the space occupied by the albumen; c, the blastoderm; d, the yolk; and m, the micropyle as before. The blastoderm now covers a considerable area at the formative pole of the egg, and its edge is marked by two parallel lines around the yolk. The outermost of these lines appears to be the growing edge of the layer of ectoderm, which has been formed by the subdivision and increase of the segmentation spherules of an earlier stage, and the second or inner line I believe to be the growing edge of the lower layer of the blastoderm. The surface of the yolk below the blastoderm and inside this second line is covered by a number of large, well-defined, nucleus-like bodies, the autoplasts of Lankester, which are strictly confined to the area inside the inner one of the two circular lines noticed, and in no case reach to or beyond the growing edge of tile blastoderm. The appearances indicate that these bodies are the nuclei of the lower layer or endoderm, which is in process of formation on the surface of the yolk under the outer layer or ectoderm, but covering a smaller area, so that the ectoderm projects beyond the lower layer a little around the entire growing edge. According to Lankester's observations (page 39 and plate 4, fig. 1), the autoplasts of the egg of Loligo cover the surface of the egg far beyond the growing edge of the blastoderm. He says that "when the cap of klastoplasts (ectoderm), has spread one-third over the egg, its marginal cells grow by a regular increase in size, and consequent fission, taking place equally all round tile margin. But before the superficial extension of the cap of klastoplasts has commenced, there appear in a deeper stratum of yolk pellucid nuclei, at first arranged in a circle around the cap of klastoplasts." This is not the case in our species, where the autoplasts appear inside of, instead of outside of, the cap of segmentation spherules. Lankester goes on to say, however, that he believes that in the eggs of Loligo there may be, according to season, an increase of nucleated cleavage segments, or, on the other hand, of these bodies_the autoplasts_they being reciprocally vicarious within small limits. If this is the case in one species at different seasons, it is not improbable that there may be still more difference between our species and the one which he studied; and to this difference may be due the lack of agreement between his observations and my own.
fig 4 fig 5	The outlines of the surface cells are now very obscure until treated with re-agents, but under the action of carmine and osmic acid, they become very conspicuous as shown in plate 1, figure 5. The growing edge is now formed by a row of large polygonal nucleated cells, b, with well-marked outlines upon the sides which are turned towards the formative pole, but with faint outlines fading gradually into the unaltered yolk a, upon their distal sides. These cells are evidently growing by assimilation of the yolk substance, and are what remain of the pyramids shown in fig. 4. Inside these there is a row of smaller polygonal cells, which are arranged in pairs along the radii of the larger cells of the first row, as in figure 4, but the regularity of this arrangement is not quite so perfect as it was at the stage shown in that figure. Inside this there are concentric rows, d, of much smaller cells, which have a somewhat peculiar arrangement; they are usually irregular pentagons, but some of them have six or four sides; their shape is nearly triangular, one end being pointed and the other blunt, and they are arranged in pairs, with the blunt ends in contact. The nuclei, which are near the centres of the cells of the series b and c, are at the blunt ends of the cells of the rows d, so that the nuclei of each pair of cells are nearly in contact. This peculiar arrangement indicates very clearly that each pair of cells of the series d, has been formed from a cell like that in the series c, by fission along a line parallel to one of the original radii of segmentation. The condition of the surface cells of the blastoderm, at this stage of development, is thus seen to have been reached in a very simple manner, by cleavage in four ways, as follows: first the blastoderm is cut up by radial lines of cleavage into a number of radiating pyramids; second, the central ends of these pyramids are cut off by cleavage at right angles to the first; third, each of the cells thus formed is divided into two by a cleavage parallel to the first; and fourth, each of the resulting cells is divided into two by a cleavage parallel to the first and third.
fig 6	The next embryo which is figured, plate 1, fig. 6, has advanced very far beyond the stage shown in fig. 3, and the position of the body of the future Squid is now indicated. The blastoderm has grown down around the yolk, which is now almost entirely covered, except at a point y, directly opposite the formative pole, where the process of segmentation was initiated. The growing edge of the blastoderm is marked by a ridge b, which is ciliated. This stage of development corresponds pretty nearly to the one given by Grenacher (Zur Entwickelungsgeschichte der Cephalopoden) in his figure 3; and this figure has been copied and referred to as a trochosphere, and the line of cilia around the edge of the blastoderm has been spoken of as a rudimentary velum. Grenacher does not himself suggest any such comparison, which is certainly unwarranted and without basis, as there is no possibility of an homology between the molluscan velum and a ciliated line, which, if it be homologous with any part of the body of an ordinary mollusc, can be compared only with the gastrula-mouth or orifice of invagination, a portion of the body which has no connection with the velum in any known mollusc. At the stage shown in fig. 6, the embryo has become bilaterally symmetrical about a plane which passes through the long axis of the egg, and the blastoderm has become raised into a circular area, the mantle, in, on that end of the egg where segmentation began. The primitive condition of the mantle, as shown by the figure, will be seen to be quite different from its condition at the same stage in Grenacher's embryo, as shown in his fig. 3, in which the mantle area is indicated by the presence of numerous chromatophores, while the margins of the area are not at all marked, and the elevation from the general surface of the yolk is very slight. In our species, the chromatophores do not make their appearance until much later, and the margin of the mantle-area is well defined at the time of its first appearance. Judging from Bronn's copies of Kolliker's figures, the mantle of Sepia seems to make its appearance about as it does in Loligo, which it resembles much more than it does that of Grenacher's form. The rudimentary arms are indicated in fig. 6, by a slight ridge or projection, a, upon each side of the embryo, a little nearer the nutritive pole of the yolk than the formative pole. The outline of the yolk, represented by a heavy black line in the figure, is no longer regularly rounded, but begins to conform to the shape of the body of the embryo, a decided prominence projecting into the mantle-area; and a large, rounded, projecting eminence on each side of the body in the space between the arm a, and the mantle, m, marking the position of the future eye-stalk. The history of the later stages of development shows that the end of the egg which is uppermost in this figure, and which is the end where segmentation began, becomes what is usually spoken of as the posterior end of tile body of the adult, the extremity opposite the head. The surface which is shown in the figure, is that which is usually called ventral, the surface which carries the siphon. The lower end of the figure is that which is occupied by the head in the adult animal. Without entering into a discussion of the homology of the Cephalopod body, which I shall return to at the end of this paper, I shall, for the present,_accepting the views of Leuckart and Huxley, which I believe to be essentially correct, speak of that surface which is uppermost in fig. 6, as the dorsal surface; of the opposite end, as ventral; of the surface shown in this figure, as posterior; and of the surface opposite the siphon, as anterior. Fig. 6 is therefore a posterior, instead of a ventral, view.
fig 7 fig 8	Figure 7 is an opposite or anterior view of a somewhat older embryo. The mantle, m, is more elevated, and its margin is quite sharply defined, and nearly circular, when see in a surface view. In the centre of the mantle area there is a shallow circular depression, the shell area, s. The eye-stalk is much more prominent than it was at the previous stage, and the layer of integument which covers it is very much thickened, and the surface of the yolk at this point much further from the surface of the body than it was at the stage shown in fig. 6. Near the centre of the eye-stalk, a depression or invagination of the integument marks the position of tile developing eye; which originates, as has been well described by Lankester and Grenacher, by the involution of the ectoderm of the eye-stalk. Between the mantle and the eye-stalk on each side of the body, another elevation or ridge has made its appearance, the outer siphon-fold, sit Fig. 8 is a view of the left side of an embryo of about the same age as the one shown in fig. 7. As before, m is the mantle; e, the eyestalk; b the growing edge of the blastoderm; y,, the uncovered end of the yolk; and a, the arm-ridge. The next two figures, plate 1, fig. 9, and plate 2, fig. 10, are of especial interest, since they represent two views of an embryo which exhibits more clearly than any other which has been figured, the relation between the Cephalopoda and the ordinary Gasteropod Molluscs. The embryo seems to be in substantially the same stage of development as the one shown in Grenacher's figs. 6 and 7, but the differences between the two are very considerable, as well as very instructive.
fig 9	Kolliker's figs. 17, taf. 1, and 25, taf. 2, as copied in Bronn's " Klassen u. Ordnungen," also represent an embryo at about the same stage of development; but the Sepia embryo differs much more from the Loligo embryo than Grenacher's form does. The latter and Sepia are extreme forms, with Loligo intermediate between them in most respects, but with much closer resemblance to the first than to Sepia. Plate 11, fig. 1O, is a slightly oblique view of the anterior surface of an embryo which, like those shown in figs. 6 and 7, has its dorsal surface above, and which is drawn in such a position as to show rather more of the left side than of the right. Plate 1, fig. 9, is a foreshortened view of the same embryo, as seen from above, in such a position as to exhibit the dorsal and posterior surface. The mantle, m, fig. 9, is now quite sharply defined, and its posterior edge, m, fig. 10, begins to overhang a little, arching over the rudimentary mantle cavity, which is thus seen to be formed in Loligo, as in most Gasteropods and Lamellibranchs, by the outgrowth of a flap of integument from the surface of the body. A reference to Grenacher's figures will show that the origin of the mantle cavity was quite different from this in the form which he studied. At a corresponding stage (see his figs. 6 and 7), the mantle area is much larger than in Loligo, and covers almost one half of the body of the embryo, and is more developed externally than it is in Loligo at the stage shown in our fig. 15. The mantle cavity, however, is barely indicated, and when it makes its appearance it is not formed, as it is in Loligo, by the outgrowth of a mantle-fold, but by the involution of the integument under the mantle, after the latter has spread over a considerable area of the body. The resemblance between the manner in which the mantle cavity is formed in most Gasteropods and Lamallibranchs, and the way it is formed in Loligo, seem to indicate that the latter presents the primitive mode of development among the Cephalopoda; and the way it is formed in Grenacher's species must, therefore, be regarded as a modification of the primitive mode which has been retained by Loligo. It is interesting to notice that there are other groups of Mollusca, among which the mantle cavity is produced in more than one way. In Cyclas, Pisidium, and many other Lamellibranchs, a fold grows out from the body, and thus converts the space between the fold and the body into a mantle cavity precisely as in Loligo, but in Anodonta the shells are quite well advanced in their development before the mantle cavity is formed, and when this is produced it is formed by the retraction of the body wall into the space between the shells; and we have in Anodonta almost precisely tile same modification of the primitive history which Grenacher describes in the Cephalopoda. In the oyster again, we have tile mantle cavity produced by a process which is about midway between that met with in Cyclas, and that presented by Anodonta. Projecting from under the overhanging edge of the mantle, on the posterior surface of the body, at the stage shown in figs. 9 and 10, are the rudimentary gills, g, one on each side of the middle line. They are formed as little papillae, or outgrowths from the surface of the body, and are covered with cilia. In position and method of formation they are very much like a single pair of the gill-tentacles of a Lamellibranch embryo, and their embryonic history would seem to indicate that they are to be regarded as greatly specialized gill filaments, rather than structures comparable to the entire gill of a Gasteropod or Lamellibranch. The shell area in the middle of the mantle is now a deep pit, fig. 9, s, and its edges have begun to fold towards each other, as shown in fig. 10,, s. As Lankester has pointed out, it originates in exactly the same manner as the shell-gland of a Gasteropod or Lamellibranch, and is strictly homologous with these structures.
fig10	The eye stalks, es, figs. 9 and 10, are now very conspicuous projections from the sides of the dorsal end of the embryo, and the eye-invaginations are well developed, and their external openings much smaller than the inner portion of the invaginated sac. In fig. 10, the left eye is shown in a surface view, and the right eye in profile. The lateral folds of the siphon, fig. 10, sí,', are a little more definite than they were during the previous stage and the two inner siphon-folds, fig. 9, si,, have made their appearance on the posterior surface of the body. They are not only separate from each other, but widely separated at this time from the lateral folds. Opposite the outer ends of the inner siphon-folds the ears, fig. 9, er, , have made their appearance, as two almost spherical invaginations of the integument, communicating with the exterior by large openings. Nearly opposite the inner folds of the siphon, on the median line of the anterior surface of the body, the mouth, mo fig. 10, is now visible as a blind sac with a large opening pointing downwards. On each side of the mouth, there is a very faintly-marked undulating line or fold of the surface of the body, v, which runs from a point near the corner of the mouth, out into the eye-stalk, crossing the outer ventral edge of the eye. The line makes two well-marked dorsal undulations, and one ventral one, the latter near the middle of the line. Any small particles which are floating in the vicinity of this line are thrown into active motion in the direction of the arrows, thus showing the presence of cilia too small to be visible by any magnifying power which can be used. The position of this line, its relation to the mouth and eye-stalk, and the presence of cilia along it, all indicate that it is to be regarded as a rudimentary velum. The ridge, a, of figs. 6 and 7, is now divided into three arms, fig. 9, a, upon each side of the body, about half-way between the mantle and the opposite pole of the egg, and therefore much nearer the mantle than at the stage of fig. 6. The yolk is now entirely surrounded by the blastoderm, and has departed still further from the regularly curved shape of fig. 1. The prolongations into the mantle and the eye-stalks are well defined, and the portion of the yolk contained within the body of the embryo, which is not quite half the whole, is separated by a well-marked constriction, just dorsal to the arms from the remainder, which is now nearly spherical. The thin layer of blastoderm which covers this external portion of the yolk is split into two layers, separated from each other by a cavity which is largest along the median plane of the body, and which is traversed by a few branched corpuscles, by the contraction of which, rhythmical waves of the outer layer are set in motion on the surface of the yolk. The fact that the mantle is dorsal to the row of arms at this period is worthy of notice.
fig11	The next figure, plate 2, fig. 11, is a view of the posterior surface of a somewhat older embryo, represented, like the preceding ones, with its dorsal surface above. The mantle, m, now overhangs the body considerably at the sides, as well as posteriorly, and the portion of the yolk which projects into it is more sharply marked off than before, and is drawn out to a point at the dorsal end. The eye-stalks, es, and their yolk protuberances, are much more prominent, and the constriction which separates the body from the external yolk is much more marked. The three pairs of arms are a little larger than before, and a cavity is visible in each of them. The inner siphon folds, si, have lengthened, and their outer ends now point towards the outer folds, si', from which, however, they are still widely separated. The most important differences between this and the preceding stage are differences of proportion and relative size, which are sufficiently well shown in the drawings, and do not call for description
fig12	Plate 2, fig. 12, is a posterior view of an older embryo, figured with its dorsal surface below instead of above, in order to facilitate comparison with the figures which follow, and with the adult animal. The mantle, m, has extended its edge sufficiently to form a very well defined mantle- cavity, within which the bases of the gill tentacles, g, are now contained. The tail fins,f, have made their appearance upon the dorsal surface of the mantle, and the rectum, re, is now present as a raised, longitudinal, hollow rod, upon the median line of the posterior surface between the gills. The two inner siphon folds, si, have met upon the middle of the body, and their free edges have bent towards each other to form the opening of the siphon; but they have not yet united with each other, and the siphon has the characteristics of that of the adult Nautilus. The inner folds are still separated from the outer ones, si, si', but the latter have begun to bend around upon the posterior surface of the body. The eye-stalks, es, are now extremely prominent and conspicuous, and the yolk protuberances no longer entirely fill them, but have begun to decrease in size, thus leaving between the eye and the yolk a space in which the optic ganglion has made its appearance. The three pairs of arms, a, are much elongated, and begin to bend away from the surface of the yolk, which is now divided into three well-marked regions; the external yolk, y'; the portion within the head region and eye-stalks, y"; and the portion within the body and mantle, y"'. During its development the embryo has undergone an increase in size, and although the drawing is less enlarged, the embryo shown in fig. 12 is actually much larger than that shown in fig. 10. The external yolk sac shares in this growth, and is very much larger at a somewhat later stage than the whole egg was at the beginning of the process of development. The points of difference between Loligo, Sepia, and Grenacher's embryo are more conspicuous at this stage than at any previous time. Kolliker's figures 2O, 21, 23, of taf. 11, represent about the same stage as our fig. 12, and a comparison will show a remarkable difference in all the details of structure, while the general plan remains much the same. The body of the embryo is very much smaller as compared with the external yolk, and is not separated from this by a constriction, but is flattened down upon it. The blastoderm does not yet cover the yolk, but only extends a short distance beyond the circle of arms. Five pairs of arms have made their appearance in the Sepia embryo and only three pairs are present in Loligo. The outer and inner siphon folds of each side have been united to each other from their first appearance in Sepia, but in Loligo they do not unite until after the two inner folds have united on the middle line to form the mouth of the siphon. The mantle and mantle-cavity of Sepia are about like those of Loligo. Grenacher's figs. 8 and 9, taf. 4O, are in about the same stage of development; and comparison will show that while this form is more like Loligo than it is like Sepia, the differences are still numerous and considerable. The external yolk is very small, and hardly projects beyond the tips of the arms. Grenacher, p. 454 is inclined to regard it as entirely wanting, but his " stirntheil," in figs. 7 and 8, is clearly the same as the external yolk of Loligo, and this again is obviously the same as that of Sepia. His argument that since the internal yolk of Sepia is divided into three regions, three divisions must be found in the internal yolk of his form, and that the "stirntheil" being necessary to make up the three, must therefore belong to the internal yolk, cannot carry much weight; for at this period the internal yolk of Loligo is only divided into two regions, although a third afterwards becomes recognizable. As regards the external yolk, then, Sepia, Loligo, and Grenacher's embryo form a series, with Loligo as the intermediate form. Grenacher's embryo differs from Sepia, and agrees with Loligo, in having only three pairs of arms at this time. The eyestalks are about equally prominent in the three forms, and the period of greatest prominence is about the same in all; but while the prominence began to be conspicuous in Loligo and Sepia at about the same time, its development is retarded in Grenacher's embryo until some time after the siphon folds have all united. The outer and inner siphon folds unite with each other in Grenacher's embryo sometime before the two inner folds meet on the middle line; while in Loligo, the two inner folds not only meet but unite with each other before they join the outer folds. In this respect Grenacher's embryo is intermediate between Loligo and Sepia, since the two sets of folds are united with each other from the first in tile latter form. As it has four separate folds at one time, Grenacher's embryo would seem to resemble Loligo more than it. does Sepia so far as the format on of the siphon is concerned. At the stages shown in figs. 8 and 9, the mantle of Grenacher's embryo covers almost half the body, and its surface is thickly set with chromatophores, while the mantle cavity is very shallow, and hardly extends beyond the edge of the mantle. In both Loligo and Sepia, the mantle is much smaller, the chromatopllores are wanting, and the mantle-cavity extends nearly or quite to the dorsal extremity of the body. The fins of Sepia and of Grenacher's embryos make their appearance at about the same time, and are developed later than they are in Loligo. The fact that three such closely related forms, belongin~ to the same group of (:ephalopods, differ from each other in the order of development of all organs of the body, except the otocysts, is a striking comment upon the assumption that the order of development of parts, as observed in a single species, can give any information as to their morphological importance, or as to the order in which they were acquired in the evolution of the group. We must notice, too, that the series which these three forms furnish when any one organ is taken as a basis of comparison, may be quite different from that given by another organ, and it is clear, without comment, that we have no information at present which will allow us to generalize upon such questions as to which of these forms recapitulates the phylogenetic record most perfectly. There are a number of interesting points, however, in which all three agree. In all, the halves of the siphon are separate at first, and all pass through a stage in which the siphon resembles that of a Tetrabranch. In all, the eyes are at first open pits, like those of the Nautilus, and in all they are at some time carried upon long stalks. In all there is a time when only four of the five pairs of arms are present.
fig13	Passing now to the next stage in the development of Loligo, plate 2, fig. 13, is a view of the posterior surface of an embryo somewhat older than in fig. 12. The external yolk sac, y, has grown so much larger that only a small part of it is shown in this and the next three figures. The mantle, m, has grown so much that the gills, g, and the rectum are nearly contained in the mantle- cavity. A constriction across the base of each gill has separated the branchial heart, h, from the gill proper. The inner folds, si, of the siphon, have united with each other to form the closed siphon tube, and the inner and outer folds, si, si', have met and are uniting with each other. The walls of the otocysts, er, have grown thin, and their cavities have greatly enlarged; the otoliths have made their appearance, and the two chambers have begun to move towards the median line, under the end of the siphon. The external openings of the otocysts have become constricted to long, tortuous, ciliated ducts, which are not visible with a low power, and are not shown in the figure. The eye-stalks, es, are of about the same relative length as in the last figure, but the yolk prominences which have filled them up to this time are now almost entirely withdrawn or assimilated, and the cavity of the eye-stalk is nearly filled by the ball of the eye, e, the optic ganglion, ga, and the white body, wb. The arms have lengthened, and suckers have appeared upon the longest pair, a", and a new pair, a', have made their appearance upon the posterior or siphonal surface of the body. In Grenacher's embryo this pair of arms make their appearance as buds upon the next pair, cc", but in Loligo they are distinct from the first, and make their appearance as elevations of the integument upon the surface of the body. The yolk is now divided into four well-marked regions, the external yolk sac, y', which is still nearly spherical; the head yolk, y", which is pretty nearly cylindrical, and which passes gradually into the external yolk sac; the body-yolk, y"', much smaller the head-yolk, and separated from it abruptly by a well marked change of outline; and the little mass of yolk, y"", at the dorsal end of the body constricted off from y"', by a deep groove.
fig14	Fig. 14 plate 2, represents a view of the posterior surface of a somewhat as seen from the left side. The mantle is now large and bowl-shaped, and covers the greater part of the body dorsal to the eye-stalks. Chromatophores now begin to make their appearance around the posterior side of the edge of the mantle, and those which first appear are of a dark brown color. The gills, g, have lengthened considerably, and are divided by constrictions into a series of enlargements, the dorsal one being much larger than the others, and becoming the branchial heart. The inner and lateral folds of the siphon have completely united with each other, and at the point of union the siphon is also united to the body wall, and the retractor muscle of the siphon; sin, now runs back to unite with the inner anterior surface of the mantle. The otocysts have almost met each other upon the median line, under the siphon, and their walls are now very thin. The eye-stalks, es, are most prominent at this stage, and soon begin to disappear. A comparison of this figure with Grenacher's figure 10, which represents about the same stage of development, will show that certain organs are more accelerated in Loligo than in Grenacher's species, while others are retarded. The yolk sac of Grenacher's embryo has disappeared, while it is as large as ever in Loligo. The mantle is much further developed in Grenachrer's embryo, and covers the whole base of the siphon, which is almost entirely uncovered in Loligo, even when the animal is retracted into the mantle. The posterior or siphonal pair of arms, a', are well developed in Loligo, and just making their appearance in Grenacher's embryo as buds from the next pair of arms.
fig15 fig16	The embryo shown, from the right side, in the next figure, plate 3, fig. 15, has assumed the general form of the adult, and the eye-stalks have almost disappeared, although, as shown in a posterior view, fig. 16, the eyes are very prominent still, and are directed more towards the ventral surface than they are in the adult. The mantle now covers about one-half the entire length of the embryo, exclusive of the yolk- sac, and the neck-cartilage, nc, has made its appearance, forming a support for the edge of the mantle, on the middle line of the anterior surface of the head. The posterior surface of the mantle is now pretty well covered with chromatophores, which at this stage possess remarkable power of expansion and contraction, and render the living embryo a very beautiful and wonderful sight under a low magnifying power. They are, as yet, entirely absent from the anterior surface of the mantle. About this time small polygonal areolations, much like epithelial cells, begin to make their appearance on the posterior surface of the mantle, and soon spread over the whole mantle, except the middle line of the anterior surface, as shown in the figure. At a later stage, figs. 17 and 18 they cover tile head and arms as well as the mantle, and still later they make their appearance upon tile surface of the siphon. Upon cursory examination, they resemble epithelial cells so much that they might readily be mistaken for them; but when more carefully examined with a high power, they are seen to be due to the presence of minute branching tubes, which, spreading over the surface of the body and inosculating, divide it up into small polygonal areas. No fluid could be seen to circulate in them, but as they appear at about the same time with the larger blood-vessels of the surface of the body, they are probably the indications of a system of capillary vessels.
fig17 fig18	The course of the larger blood-vessels on the posterior face of the mantle is shown, at a somewhat later stage, in fig. 17. A large vessel will be seen to enter the mantle on the median line near the dorsal end of the body. This is the pallial artery from the systemic heart. Passing forwards, it divides into three branches; a pair of large ones, and a median unpaired smaller one. The latter runs forwards, nearly to the lower edge of the mantle, and divides up into a number of smaller branches. The two larger branches diverge, and running out towards the free edge of the mantle, give rise, on their inner edges, to a number of irregular branches, and on their outer edges to a number of nearly parallel trunks, which communicate with a pair of large venous trunks, each of which receives a smaller trunk from the median tract of the mantle, and then, bending around the side of the body, runs inwards to open into the larger vena cava, from which the blood passes into the branchial heart, and is conveyed to the gills. The branchial hearts appear at quite an early stage of development, but the systemic heart is not developed until about the stage shown in fig. 16. During the later stages of development, and in the adult also, the small size of the gills is no doubt compensated for, to a great degree, by the aeration of the blood while it is passing through the system of vessels near the exposed surface of the mantle. At the stage shown in fig. 15, the siphon has substantially its adult form, and is made up of two lateral chambers, si', which have been formed from the lateral siphon folds, and which open into the mantle-chamber, but have no external openings; and a single median chamber, si, on the posterior surface of the body, which has been formed by the union of the two inner siphon folds, and which opens into the mantle-chamber as well as externally. At the point where the lateral chambers meet the median chamber, the wall of the siphon is united to the wall of the body, and the three chambers are thus shut off from communication with each other. The animal is so perfectly transparent that the valve-like action of the two outer chambers can be perfectly seen, as their free inner edges are thrown out against the mantle so as to close it at each contraction, and the water, which passes in around the whole free edge of the mantle, is thus concentrated in the funnel-shaped middle chamber of the siphon. At about this time the valve of the siphon, figure 15, v, is developed as a single unpaired flap, which arises from the posterior surface of the neck. Considerable change has now taken place in the shape of that portion of the yolk which is contained in the head. It is reduced to a long narrow tube, y", which connects the portions contained in the body proper, y"', y"", with the external yolk sac, y'. The pulsatile space, x, between the outer wall and the surface of the yolk sac, is more plainly shown in this figure than in the preceding ones, although a profile view shows it with equal distinctness at earlier stages. The figures of the later stages hardly call for much sufficiently well in themselves the gradual the form and structure of the adult. Fig. 16 is a posterior view of an embryo a little older than the one shown in fig. 15. A large rounded prominence on each side of the bed marks the position of the eye-stalk, and the eyes are farther forward than they are in older specimens, but in other respects the form is very similar to that of the adult. The ink sac, i, has appeared, and is filled with ink, and tile tip of the free portion of the rectum is prolonged at its corners into the pair of ear-like anal valves. Figs. 17,18, are copies of drawings made at Mr. Agassiz's laboratory at Newport during the summer of 1876, from specimens which were found swimming at the surface of the water of Narragansett Bay. As the arrangement of the chromatophores is like that of the other embryos, they undoubtedly belong to the same species and are given here to show the later stages of development. The external yolk-sac has almost disappeared, in fig. 17 and it is entirely wanting in fig. 18. There are considerable individual variations in the arrangement of the chromatophores, but there are certain features which were observed in all the specimens, and which seem to be constant . The first which make their appearance are dark brown in color, and are placed in a ring of six or seven (plate 2, fig. 14), around the edge of the mantle on the posterior surface. They are a little smaller, and somewhat more excitable than those which appear subsequently, and they can be readily recognized in the later stages shown in figs. 15, 16, 17, 18. They are soon followed by larger spots of the same dark brown color, scattered irregularly over the posterior surface of the mantle (fig. 16). The next spots to appear are upon the arms, and are also dark brown. At first there are two upon the first or siphonal pair of arms' and three upon the second pair (fig. 16). A fourth soon appears upon the second arm, and these four remain conspicuous until quite a late stage of development (fig. 18). Three large brown spots now appear upon the posterior surface of the head (fig. 16), and they are soon followed by others (fig. 17). A second set of spots, more deep-seated and of a bright orange color, soon make their appearance, and are much more constant in position than the brown ones. The first pair which appear are just in front of, or ventral to the eyes. They are soon followed by a single one on the middle line of the head, at the bases of the first pair of arms, and another single one on the middle line of the edge of the mantle. About the same time a pair appear dorsally to the eyes, and another pair on the edge of the mantle, near the sides. Four small orange spots next appear upon the second pair of arms (fig. 17, a"), alternating with the four larger brown spots, and soon after a ring of six or eight orange spots appears on the mantle, dorsal to the ink bag. Two orange spots next appear upon the first pair of arms, figure 18, a', alternating with the brown spots. I was not able to procure specimens in sufficient numbers to follow the development of the spots to any later stage than this, but they do not seem to develop regularly at the later stages.
	Theoretical Discussion on the Observations In a general view of the facts which have been detailed, the most prominent and conspicuous feature seems to be the remarkable directness with which the embryo develops into the adult, and the total absence of anything like a metamorphosis. Everything which does not contribute to the formation of the adult animal has been dropped out of its life history. With the exception of the velum and the eye-stalks, there are no larval or rudimentary organs, and the whole process of development is wonderfully direct. When we bear in mind that the Cephalopoda are almost the most highly specialized of Invertebrates, and that they must have had a long and complicated phylogenetic history, I think we must acknowledge that tile embryonic record has been simplified to a degree which is without a parallel in the animal kingdom, and it is hardly too much to say that the ontogenetic process furnishes us with no knowledge whatever of the phylogeny of the group. The method of formation of the shell-area and of the shell; the mode of origin of the mantle and of the mantle-cavity, and the form and position of the gills of the Cephalopod embryo are all of them features which show closer relationship to a typical Gasteropod than could be inferred from the condition of these organs in the adult. They thus show the affinity of these groups to each other, and help us to a clearer understanding of tile homology between the organs of the Cephalopod, and those of a typical Mollusc, but this is about all. I do not see th at they furnish any basis whatever for speculation upon the origin of the Cephalopod, or give us the least information as to the manner in which its peculiar specializations of structure have been brought about. We, undoubtedly, have an ancestral feature in the embryonic condition of the siphon at the time v hen the siphon folds are separated from each other upon the median line, but in the present state of our knowledge it furnishes a very scanty basis for generalization. It is true that there is a resemblance between the adult condition of this organ in the Tetrabranch and its embryonic condition in the Dibranch, and the evidence therefore shows that in this respect, the former group is more embryonic, or less specialized, than the latter, but it certainly is not sufficient to warrant the conclusion that the one group has been derived from the other, rather than from a common ancestral form. I think that this is true of the development of tile eye as well as of that of the siphon. The history of this organ shows, as Grenacher has pointed out, that it is essentially similar to the ordinary molluscan eye, and that it is less specialized in the Tetrabranch than it is in the Dibranch, but the resemblance is not such as to indicate that the Tetrabranchs are the ancestors of the Dibranchs. It is just such an embryonic resemblance as we should expect to find, if both are the descendants of an unspecialized form, and while it is not inconsistent with the idea that this common ancestry is very remote indeed, neither would it conflict with the view that the divergence of the two groups was comparatively recent. The arms of the Dibranch embryo are hardly, if at all, more like those of a Tetrabranch, than those of the adult, and although the shell is at first external, it is, in this respect, no more like the Nautilus shell than any other molluscan shell is. In a word, the case is hardly too strongly put, by the statement that the developing Dibranch has so completely lost all ancestral features that no traces of them remain. In another point of view our embryo is more suggestive. Although it furnishes us no basis for phylogenetic speculations, it does furnish a safe ground for the discussion of cephalopod homology. Homology, as I take it, is the resemblance which is due to common ancestry, while phylogeny is the study of the steps by which specializations of structure have been acquired. It is quite conceivable that a form of life should exhibit its relationship to a remote ancestor, without any indications of the manner in which the divergence from this ancestor has been brought about. This seems to be the case with the Cephalopoda. While the Squid embryo fails to give us any information as to the way in which a typical Mollusc has been modified in order to convert it into a Cephalopod, or as to the transformations through which it has passed in reaching its present form, it nevertheless clearly shows the fundamental similarity of type which subsists between it and the other Mollusca. The precise relation between the organs and regions of the body of a Cephalopod and those of a more typical Mollusc, have always been as obscure as the close natural affinity of these groups is obvious, and the obscurity has been increased by the disposition to regard the various Mollusca as modifications of a highly specialized architype, uniting in itself the characteristics of all its derivations. This tendency still retains a considerable influence over the minds of morphologists. although it is now perfectly obvious that the architype, or proto-mollusc, must have been an unspecialized rather than a high]y specialized form, and that the various existing molluscs must have been derived from this primitive form by gradual specialization. The only basis for an homology of the Mollusca, in the absence of a phylo~enetic record, and of transitional forms, is therefore to be sought in the comparison of early stages in the development of individuals of the different groups, at a time when the specializations of structure, characteristic of the adult forms, have not yet made their appearance. In his paper on the development of an unknown Cephalopod, Grenacher has made such a comparison, and has attempted to furnish a sound basis for the discussion of tile homology between tile Cephalopoda and other Mollusca. As the result of this comparison, he reaches conclusions which, being the product of observation rather than of speculation, are much more valuable and natural than any which have been advocated by previous writers, although I believe them to be only partially correct. Although his embryological record is very complete, it unfortunately lacks the stages which are of the greatest importance as a basis for generalization, and I have been so fortunate as to fill this gap by finding embryos which exhibit general molluscan characteristics unobscured by the presence of the distinctive features of the Cephalopod. For comparison with a Gasteropod embryo I give in fig. 19 of plate 3, a diagramatic side view of an embryo at the same stage as that shown in plate 2, fig. 10, but with the rectum and anus represented as they are found at a somewhat later stage. The figure, like fig. 10, has the head and the yolk sac y, below; the so-called posterior end above; the surface upon which the siphon is to be developed at the left, and the so-called dorsal surface at the right. Fig. 20, plate 3, is the embryo of a fresh-water Pulmonate, in what I regard as a similar position. The foot, f, is below; the shell, ski, above; the mouth, m, the tentacles, t, and the head on the right, and the rectum, r, on the left. The figure therefore shows the right side of the body of the Gasteropod; and as the twist by which tile rectum and the mantle-chamber are subsequently carried around the right side of the body on to the anterior surface has not yet appeared, the embryo is bilaterally symmetrical, and the mouth, the anus, the foot, the mantle, and the shell occupy substantially the same positions, and bear about the same relations to each other, that they do in a Lamellibranch. A comparison with fig. 19, or with fig. 1O, will show that there is, at the same time, an essential similarity to a Cephalapod. In the Gasteropod, fig. 20, or in a Lamellibranch, we have the bilaterally symmetrical shell, ski, resting like a cap upon the dorsal surface of the body, and surrounded by a reflected ridge of integument, the margin of the shell area, sa. As Ray Lankester has pointed out, the embryonic shell of a Cephalopod is very similar, and in the Squid, fig. 19, we have the external shell, ski, surrounded by the reflected rim of integument, sa, exactly as in the Pulmonate. Running around the shell and shell area in both Cephalopod and Gasteropod, is a second ridge of integument, the margin of the mantle, ma, ma, which already projects a little from the posterior, ventral surface of the body of the Cephalopod, to form the outer wall of the rudimentary mantle chamber. On the middle line of the posterior surface of the body, just below the mantle-ridge, in both Cephalopod and Pulmonate is tile rectum, r, with its external opening, the anus; raised from the surface of the body upon the anal papilla in the Cephalopod but other wise alike in the two forms. On each side of the rectum of the Cephalopod is a single tentacular gill, g, underneath the overhanging edge of the mantle. There are no correspondin~ structures in the Pulmonate, and in the Prosobranch the gills do not make their appearance until the symmetry of the body has been distorted by the twisting of the visceral mass into a spiral, but the gill filaments of a Lamellibranch are exactly similar in form, structure and position to the gills of the Squid embryo. On the anterior surface of the body, the surface which is usually called dorsal in the Cephalopod, and which is on the right in both figures, we have first, the thickened mantle ridge, ma, which forms the angle between the dorsal and tile anterior surface. Next we have in the Pulmonate and in most Gasteropods the large pulsatile " neck region," h. In the Cephalopod this region is short, not pulsatile, and it forms the " back " of the animal. Below the neck region of the Pulmonate a tentacle, t, is developed on each side of the median line of the body, and the eyes subsequently make their appearance upon these tentacles. On each side of the corresponding region of tile body of Squid embryo. the projecting eye-stalk, t, carries the rudimentary eye upon its rounded extremity, and there does not seem to be any doubt that the sensory tentacles of the Gasteropod, and the eye-stalks of the Cephalopod are strictly homologous. In tile Pulmonate tile rudimentary velum, v, is marked by a line of granular ciliated cells, which crosses the middle line of the body just below the tentacles, and then running out upon the sides, bends up towards the dorsal surface, in such a way as to almost encircle the tentacles. When the corresponding surface of the body of the Cephalopod embryo, at the stage shown in plate 2, fig. 10, is carefully examined with the highest power which can be brought to bear upon such a large rounded opaque surface, a well defined line or groove, plate 3, fig. 19, v, will be found to run from the median line out upon the sides of the body, bending up towards the dorsal surface in such a way as to partially encircle the eyestalks; and the motion of floating particles shows the presence of cilia along the line. Immediately below it at the point where it approaches the median line of the body, the mouth, m, is situated, in tile Cephalopod as well as in the Pulmonate, and the fact that the ciliated line bears exactly the same relation to the mouth and to the sensory tentacles in the Cephalopod that it bears in the Pulmonate seems to show beyond doubt that it is the same structure, a rudimentary velum, in both cases. The mantle; the mantle-cavity; the shell area; the shell; the rectum and anus; the sensory tentacles; the velum, and the mouth are thus seen to be so similar in position, relations, mode of development and function, in these two Molluscs, as to leave no doubt of their homology. When used as a basis for comparison these various features furnish us with a sufficient number of points of orientation to assure us that the Cephalopod embryo must be placed as it is in plate 3, fig. 19, and in plate 2, figs. 10 and 11, in order to be in a position which is homologous with that of the Gasteropod embryo, shown in fig. 20. The views which were advanced upon the morphology of the Cephalopoda, nearly thirty years ago, by Huxley, are therefore essentially correct in outline, although I shall now give my reasons for opposing certain of the homologies advocated by him. There are few morphological questions upon which more conflicting views have been expressed than those of the various writers who have discussed the homology of the siphon and arms of the Cephalopod, and the equivalent, in this group, of the Gasteropod foot. The history of opinion upon this subject has been treated at length by Grenacher, and more recently by Von Jhering (Vergl. Anat. des Nervensystems und Phylogenie der Mollusken, pp. 269- 281), and I may therefore enter at once upon an examination of the morphological aspect of the question without first reviewing its historical side. The molluscan foot, fig. 20, f, is a median unpaired structure, on the ventral surface of the body, between the mouth and the anus. In the Pulmonate, and in many other Gasteropod embryos, a large sinus-space, c, separates the integument of the foot from the endoderm and its derivatives. This space contains blood corpuscles, and as the integument is rythmically contractile the embryonic foot is a circulatory organ. A glance at fig. 19, or fig. 10, will show that the only unpaired structure on the median line of the ventral surface of the body of the Cephalopod embryo is the large external yolk-sac, y, and to this, if anywhere, we must look for the homologue of the Gasteropod foot. When the Cephalopod embryo is seen in a profile view, fig. 19, or fig. 15 the integument f, of the yolk-sac will be found to be separated from the yolk by a space, c, and as the integument is rythmically contractile, the fluid which fills this space is kept in constant motion. Physiologically then, as well as in its position, the yolk-sac of the Squid resembles the foot of the Gasteropod, and I think we must conclude that, as a locomotor organ, tile foot of the Cephalopod has been suppressed by the great development of a food-yolk at the point where it should be found. The arms of the Squid make their appearance as little protrusions, fig. 19, a, a, a, arranged in pairs around the neck or constriction which separates the external yolk-sac from the body proper. As they are, at first, ventral to the mouth, and as a true velum is present, they cannot be regarded as a velum, and as they are paired instead of median, they are not homologous with the median unpaired foot. They are paired outgrowths from the foot region, and may perhaps be regarded as the equivalents of the cephaloconi of Clio, but there does not seem to be any evidence that they have been produced by the modification of any part of the body of a typical Gasteropod, and they are undoubtedly structures which have been acquired by the Pteropods and Cephalopods, after these diverged from the common ancestral form which united them to the Gasteropod stem. The siphon originates as two pairs of folds, s and s', of the integument of the lateral walls of the body, and if we regard these four folds as homologous with the epipodial folds of a Gasteropod, the arms must be regarded as independently acquired structures. If we regard the arms as modifications of the epipodial folds we must consider the four siphon folds as independently acquired structures, and as we have nothing whatever to furnish us with a test, nothing seems to be gained by the uncertain homology of either the arms or the siphon, with any part of the body of a typical Gasteropod. It seems certain that the common ancestor of the Gasteropods and Cephalopods must have been an unspecialized form, rather than a highly complex archetype, and if this is the case we cannot expect any valuable results to follow from the attempt to compare any part of the body of a Cephalopod with structures which, like the epipodial folds, are not common to the Gasteropoda, but somewhat exceptional; which, when they are found at all, as in Aplysia, are not rudimentary but functional; making their appearance very late in the history of the individual instead of early, and presenting every indication of recent acquisition While we owe a great debt to Huxley's paper on the Morphology of the Cephalous Mollusca, for the demonstration of the general relations between the Cephalopod body and that of a Gasteropod, I think that confusion has resulted from his attempt to carry the homology into the details of Cephalopod structure. The growth of opinion upon the homology of the siphon and arms may be stated briefly as follows: Huxley regarded the arms as the true foot, and the siphon as the epipodial folds. Grenacher shows that as the foot is an unpaired structure, it cannot be homologous with the arms, and he follows Loven (Beitrage zur Kenutniss der Entwicklung der Mollusca Acephala Lamellibranchiata; reprint of 1879, p. 33), in regarding these as a modified velum. The foot he believes to be wanting, and the siphon he regards as the epipodium. Von Jhering opposes Grenacher's view that the arms are homologous with the velum, and points out that the cephaloconi of Clio, which are undoubtedly homologous with the arms, are certainly not homologous with the velum, since an embryonic velum appears in the young Pteropod, and then disappears without forming any part of the adult body. He also calls attention to the fact that in the Gasteropods, even when the velum is cut up into tentacles, these do not persist or become converted into any part of the adult body. He says, on p. 269, that the arms are tentacular appendages to the body, and does not appear to regard them as the equivalent of any part of the Gasteropod. The innervation of the siphon from the pedal ganglion leads him to regard this as the foot, and he concludes that the valve of the siphon is the true foot or protopodium and the two lateral folds pteropodia. He conjectures that Grenacher's two inner folds unite - and give rise to the valve, but this is opposed to Grenacher's account, as well as to my own observations, and there can be no doubt that the two inner folds form the tube of the siphon and the two lateral folds its lateral chambers. The valve appears quite late, fig. 15, v, as an outgrowth from the inner wall of the siphon tube, and there is nothing in its history which gives any reason to believe that it has ever had any other function than that of a valve to the siphon-tube. Von Jherring's homology rests upon the assumption that similarity in the method of innervation implies similarity of origin, but the bilateral character of the siphon seems to be an objection to its homology with the foot, even if it were in the right position upon the body, and if it is a new structure, as I believe, the origin of its nerves cannot have any profound morphological significance. - BEAUFORT, N. C., July 31, 1880.
	EXPLANATION OF THE PLATES. All the figures except plate 1, figs. 2, 4, and 5 arid plate 3, figs. 19 and 20 were drawn from living specimens, swimming without restraint within the uncompressed egg, or, after their escape from the egg, in a sufficient quantity of water to allow perfect freedom of movement. As the use of a camera was thus rendered out of the question, the drawings are not all upon the same scale. Most of them were made, however, with an amplification of about eighty diameters.