Witli 11 Text-flgures.

During the progress of an investigation on the anatomy of three recently discovered species of Cephalodiscus (C. nigrescens, C. hodgsoni, and C. gilchristi), kindly entrusted to me for description by Prof. E. Ray Lankester, F.R.S., Director of the Natural History Museum, London, I noticed that in the development of the buds there is no torsion of the first and second plume-axes such as is described by Masterman as occurring in buds of C. dodecalophus,1 and that the last plumes do uot develop between the earlier plumes and the buccal shield, as recorded by the same writer, but on the side of those plumes remote from the buccal shield. His figures 64—68 are reproduced as textfig. 1 

I was unable to settle the point satisfactorily in time for the introduction of my results into the two papers describing the structure of the polypides and tubaria of the new species,2 but since the completion of those papers I have made an exhaustive study of the growth of the plumes in buds of Gephalodiscus hodgsoni, C. nigrescens, and C. gilchristi, and having made from the abundant material at my disposal a well-graded series of buds of each species, I was enabled to confirm my earlier conclusions.

I finally prepared a series of buds of the “Challenger” species, Gephalodiscus dodecalophus, with the result that I am satisfied that the buds of this species agree with those of the other three species. The torsion described by Masterman does not occur, but the grooves of the first and second pairs of plumes, developed in such a manner that they face ventrally towards the buccal shield, continue to face the shield. The last two pairs of plumes are not developed between the first two pairs and the shield, but on the dorsal side of them, i.e. on the side remote from the shield. The grooves of these last plumes (fifth and sixth pairs) are not directed towards the shield, but away from it.

Although the four series of plumes of buds were prepared for the purpose of deciding the points raised by Masterman, they serve to show a number of other interesting features, such as (1) the difference of plume development in the buds of the four species; (2) the relation which the last-formed pair of plumes bear to the edge of the post-oral lamella; and (3) the conversion, in the later stages of growth, of the line of the bases of the collar-outgrowths (plumes and postoral lamella) from a circle to a double crescent, or two incomplete ellipses. I have selected from each of the four series a number of preparations to illustrate these points; they are reproduced in text-figures 2, 3, 4, 6, 7, and 9, and are described under headings of the respective species.

In the case of the youngest stages the whole bud was mounted in diluted glycerine, and gold size was run round the edge of the cover-glass to keep it firmly in position, and to prevent the glycerine from accumulating dust. Most of the buds were dissected, the shield being first removed by tearing through its stalk by the aid of fine mounted needles, and then the collar region, with plumes and post-oral lamella (e.g. text-fig. 2, J), was removed by carefully manipulating the needles between these parts and the “body” of the bud. The three parts, shield, collar region, and “body” with its stalk, were then mounted on the same slide in dilute glycerine.

The relation which the collar region, as dissected out, bears to the whole bud will be made clear by reference to textfig. 5, A, a diagrammatic side view of a bud with three pairs of plumes, and text-fig. 10, a diagram of a longitudinal section of the adult polypide; in the latter figure the limit of the collai’ region is marked by two lines of heavy dashes.

The results here recorded and illustrated are based upon 135 preparations of buds made and mounted as above explained. In some cases (e.g. text-fig. 6, F) the three parts— “body,” collar appendages, and shield—were drawn separately on tracing paper, and the perfect bud was reconstructed by a superposing of these transparent sheets.

No staining fluids were used. The dissections were made under a Greenough binocular erecting microscope magnifying 20 and 40 diameters ; the drawings were made by the aid of a compound microscope with Zeiss apochromatic 8 mm. objective and No. 6 compensating ocular with eye-piece micrometer. The figures in text-figs. 2, 3, 4, 6, 7, and 9 are all reproduced to the same scale of enlargement, viz. 62 diameters.

In the earliest phase of development the bud is ovoid or pyriform (text-fig. 2, A) without signs of buccal shield or plumes. The shield and the first pair of plumes make their appearance simultaneously (B and c). The plumes are at first small mounds, later hemispherical (D), and still later longer than broad (E, F, etc.). The second pair of plumes appear on the external side of the first pair (D, 2), and in contact with them. The two plumes of the first pair are not in contact with one another at their bases. The third pair of plumes appear lateral to the second pair (E, 3), and in contact with them.,

The bud increases considerably in size before the fourth plumes appear; the buccal shield assumes its definitive shape by becoming flattened, and by the differentiation of the posterior lobe. In G and H the dotted line marks the position of the free edge of the posterior lobe. The fourth plumes appear when the bud has reached the stage of development represented in H. The structure which in G is marked po.l. might at first glance be taken for the developing fourth plume; it is the lateral part of the post-oral lamella.

Figure J represents the collar region and its appendages (the plumes and post-oral lamella) of a bud a little older than that shown in figure H; the buccal shield has been dissected off from the one side, and the “body” and stalk from the other. The position of the mouth is indicated in the figure, but this region is always more or less damaged in making such a dissection as that described. The torn stalk of the buccal shield in this preparation—as also in the similar preparations shown in text-fig. 3, N, O, and R, and in other of the text-figures—occupies a position intermediate between the mouth and the bases of the first pair of plumes. In figure J the fourth plumes are hemispherical projections at the bases of the third plumes, touching these on the one side, and in continuity with the post-oral lamella on the other. Traces of pinnules are appearing on the first and second plume-axes, and the terminal end-bulbs of these same plumes are swelling out.

In K the fourth plumes are now longer than broad, and the fifth pair are making their appearance, L is a side view of a bud of about the same stage as K, as may be determined by comparing the development of the fourth and fifth plumes in the two buds. Pinnules are now appearing on the third plumes, and it will be noticed, in the case of the first, second, and third plumes, that the pinnules are not laterally placed on the plume-axes, but ventro-laterally, and the groove between the two rows of pinnules of each of the plumes in question opens ventrally, or ventro-laterally, towards the dorsal or stalked side of the shield.

The figures in text-fig. 3 continue the series shown in text-fig. 2. M is a little more advanced than L of text-fig. 2; the fifth plume is now hemispherical in shape; pinnules are appearing on the fourth plumes, and the end-bulbs are becoming differentiated. The end-bulbs of the first, second, and third plumes—the last concealed in the figure by the second plume—are showing refractive beads (r.6.), and their surface is no longer smooth, but warty; the pinnules are larger and more numerous than in L. The figure M, like shows how the ciliated groove between the two rows of pinnules of each of the plumes 1, 2, and 3 faces towards the dorsal surface of the buccal shield.

Figure N is a dissection made in the same manner as that shown in figure J, by the removal of the buccal shield on the one side of the collar and the “body” of the bud on the other. The preparation is viewed from the dorsal surface, so that the grooves of the plumes are facing away from the observer. The fifth plumes are seen to be continuous with the bases of the fourth plumes, and with the post-oral lamella; they are unequally developed on the right and left sides of the body.

In O the pinnules on the first plumes are now finger-like processes, those on the fourth are in the same stage of development as the pinnules of the first plumes were in the stage represented in figure M. There are twelve or thirteen pairs of pinnules on the plumes of the first pair, and about nine on the fourth plumes, but the numbers later become increased by the development of new pinnules at the basal end of each series. The fifth plumes are larger than in the last stage, and show an indication of pinnules, but as yet they have no end-bulbs. There is no sign of a sixth plume between the fifth plume and the post-oral lamella, o, like N, is a dorsal view, and the ciliated grooves of the plume-axes are not seen ; the grooves of the first and second plumes are facing ventrally, those of the third ventro-laterally, and those of the fourth postero-laterally.

Figure P shows an unequal development of the right and left plumes of the fifth pair. On the right side the plume has about eight pairs of pinnules and a moderately differentiated end-bulb, whereas on the left side the length of the plume is only a little greater than the width. The groove on the right-hand plume is facing posteriorly.

Figure Q shows the developing sixth plume, wedged in between the base of the fifth and the antero-lateral edge of the post-oral lamella;1  it is already pear-shaped, and with a narrowing base. The preparation here drawn is presenting its dorsal surface to the observer, and the fifth plume is showing its groove facing postero-dorsally.

In figure R the fifth plumes are larger and with better developed pinnules than in Q, but the sixth plumes are not more advanced; indeed, on the left side the sixth is less developed than it is in Q. It is worthy of special remark that, while in the case of the first and second plumes the development of the right and left plumes of each pair is regular and symmetrical, in the fifth and sixth plumes the one is very frequently more advanced than the other of its pair. The grooves of the fifth plumes should be facing posteriorly or postero-dorsally; by the pressure of the coverglass upon the preparation the plumes have twisted round somewhat. For the same reason the first plumes have their grooves facing to the left, while the third plume on the right side is twisted right round so as to present its grooved surface to the observer. The pinnules on the first plumes are now fairly long, and the pinnules of the third and fourth plumes are finger-like, and resemble those of the first plumes in the stage shown in figure os differs from R in that the sixth plumes are more advanced, the right hand one showing signs of pinnules. The fifth plume shows its groove facing posteriorly. The bud from which the preparation shown in figure s was made was a very late bud, so large, and with such great development of the visceral mass, bringing the base of the stalk from a posterior to a ventral position, that, if it was not still in counection with a full-grown polypide by means of its stalk, one might easily mistake it for an adult polypide.

Possibly in some cases the sixth plumes are permanently arrested in their development on one side, or on both sides, or they are not developed at all, for some polypides of Cephalodiscus hodgsoni of full size and with mature gonads have five pairs of plumes only. In the cases of those polypides which are found with eleven plumes the explanation may just as likely be that the sixth plume has failed to develop on one side as that one of the plumes has been lost by injury.

As regards the development of the pinnules on a plumeaxis, about five pairs appear simultaneously, and at the time of their origin they occupy nearly the whole, or, at all events, more than half of the length of the plume-axis behind the end-bulb. The later pinnules are developed a pair at a time at the basal end of the series.

The bud of Cephalodiscus dodecalophus in its earliest phase (text-fig. 4, A) resembles that of C. hodgsoni except in being somewhat smaller. The first pair of plumes and the buccal shield begin to differentiate simultaneously (B and c). By the time the second plumes appear (E) the first pair are relatively longer than in C. hodgsoni. The third plumes develop before the second pair have grown very large (F and a).

In buds with three pairs of plumes present the size of the first pair varies considerably, as may be seen by reference to H and o. Although H is a larger bud than a, it is not much more advanced in its development; the second and third plumes are only slightly more developed than in G, and there is no sign of the fourth plumes. Yet the first plumes are very large in H, and have refractive beads in the terminal end-bulbs, whereas the corresponding plumes in G are no larger in proportion than in C. hodgsoni (cf. text-fig. 2, o).

In J the fourth plumes are appearing, and the gradation from the first plumes to the fourth is fairly uniform ; refractive beads are now visible on the second plumes. On comparing J of text-fig. 4 with j of text-fig. 2 it will be seen that in C. dodecalophus the lateral lobes of the post oral lamella are more prominent as free flaps than in C. hodgsoni.

Pinnules first appear in the stage represented in t, i. e. in the stage when the fourth plumes are appearing, or possibly a little earlier (H). They are at first .confined to the first pair of plumes, and they appear as five or six pairs of small eminences arising simultaneously; additional pairs develop in succession at the basal end of the series. In K there are seven pairs, and in L eleven or twelve pairs of pinnules on the plumes of the first pair. The pinnules on the second, third, and subsequent plumes arise later than those on the first, as might be expected. About four pairs of pinnules are just appearing on the second plumes in J.

Masterman,1  in figuring a bud about as advanced as o and H of text-fig. 4 of this paper, shows four pairs of long slender pinnules on the first plumes. There is nothing comparable with this in my preparations. My observations also fail to accord with those of Dr. Masterman in the matter of the development of the first few pinnules on the plumes. In figs. 69—74 of his Plate 4 he gives figures illustrating the growth of a plume—he does not say which. In fig. 71 there are no pinnules, in fig. 72 one pair, in fig. 73 the first pair have elongated considerably, and there is a new pair on their proximal side, in fig. 74 there are four pairs of pinnules shown. My own observations go to show that the first five or six pairs of pinnules arise simultaneously on the first plumes; on the second and third plumes the first four or five arise simultaneously, and on the fourth pair (text-fig. 4, K, 4) the first three or four. At a stage when any one plume has but four pairs of pinnules I have not seen the pinnules nearly so long and slender as those shown in Masterman’s fig. 74.

In L of text-fig. 4 the fifth plumes have attained considerable size,and exhibit about four pairs of pinnules. The pinnules on the first and second plumes are now long, except the youngest of the series near the base of the plume-axis. Endbulbs with refractive beads (r.b.) are present on the first four pairs of plume-axes. Both K and L are ventral views, and the groove between the two rows of pinnules of each plumeaxis is directed towards the observer.

It will be noticed that in J, K, and L the right and left plumes of the first pair are separated by a small interval, in which is situated a slight mound of the basal part of the stalk of the buccal shield, obliquely torn in the making of such preparations as those here figured; but the several plumes of the same side of the body are in close contact with one another at their bases, and the last developed plume is wedged in, with a certain amount of over-lapping, between the last plume but one and the antero-lateral edge of the post-oral lamella.

Except in distorted specimens the last developed plume lies slightly dorsal to the edge of the post-oral lamella, and lies slightly ventral to the last plume but one (K). It must be noted, however, that in stages such as that shown in i, and in later stages, the line of plume bases and base of the post-oral lamella does not lie in a plane, and a certain amount of distortion or crumpling is bound to result from any attempt at mounting the dissection between two pieces of glass. Note the wrinkle in the middle of the posterior edge of the post-oral lamella in L. In such a stage as is shown in J the line of plume-bases and attached edge of the post-oral lamella is nearly in one plane, a plane slightly oblique to the long axis of the bud. The fourth, fifth, and sixth plumes, however, are not developed in this plane, but more dorsally, and this is how it comes about that in the adult polypide the bases of the six plumes are set in a circle or ellipse in one oblique plane of the body, while the attached edge of the post-oral lamella lies in another oblique plane, which intersects the first plane at the points of contact of the sixth plumes with the edge of the post-oral lamella. This will be made clear, I think, by reference to text-fig. 5, and its accompanying legend. See also text-fig. 8, a diagram of Cephalodiscus nigrescens.

Masterman has stated1  that the first three pairs of plumes arise with their grooved faces directed towards the buccal shield, but later rotate upon their axes so that the grooves face away from the shield.; further, that the last three pairs of plumes arise between the first three pairs and the shield and have their grooves directed towards the shield (see textfig. 1 of this paper). I am in a position to confirm Harmer’s opinion1  that this statement is without foundation, and that Masterman’s earlier enumeration of the plumes2  is the correct one. In Cephalodiscus dodecalophus, as in C. hodgsoni, the grooves of the first and second pairs of plumes, developed in such a position that they face the buccal shield, continue to face the shield, and the last two pairs of plumes are not developed between the first two pairs and the shield, but on the side of the first two plumes remote from the shield, i. e. to their dorsal side; and. the grooves of the last-formed plumes are not directed towards the shield, but away from it. The grooves of the third and fourth plumes face somewhat laterally.

The six pairs of plume-bases are set on an ellipse, and the grooves of the twelve plumes face away from the foci of the ellipse. The six grooves of the right side lead down into the space between the shield and the post-oral lamella on the right side of the mouth; similarly the six grooves of the left side lead into the left side of the mouth (see text-fig. 8, C. nigrescens).

While in buds of Cephalodiscus hodgsoni and Cephalodiscus dodecalophus the first pair of plumes begin to appear at about the same time as the buccal shield, in those of Cephalodiscus nigrescens the buccal shield is well differentiated, and with a well-defined posterior lobe (text-fig. 6, A) before there is any sign of plume development. When the first pair of plumes make their appearance (B, 1) they are small as compared with the bud as a whole, and when the second plumes begin to develop (c, 2) the buccal shield is oe remarkably large size. Compare text-fig. 6, c, with text-fig. 2, D; and text-fig. 4, E.

The third plumes appear to the outer side of the second, and are in contact with them at their bases (D, 3). The two plumes of the first pair are widely separated from one another (D, 1). When in buds of Cephalodiscus hodgsoni and Cephalodiscus dodecalophus the third plumes make their appearance (text-fig. 2, E, 3, and text-fig. 4, G, 3) they are directed strictly laterally, whereas in Cephalodiscus nigrescens (text-fig. 6, D and E, 3) they point more anteriorly than laterally. This doubtless is connected with the relatively late origin of the plumes, and the relatively small size of the plumes during the early stages of their development.

A, B, and D illustrate a peculiarity of many buds and. adults of Cephalodiscus nigrescens, the bendingforward of the posterior lobe of the buccal shield. The occurrence of the posterior lobe in this position may be due to an exceptional contraction of the muscles of the shield brought about by the irritating properties of the formalin solution in which the animals were killed, but whether this be so or not, it indicates considerable mobility of the organ in question in normal conditions of existence.

Text-fig. 6, F, shows the dorsal view of a bud in which the fourth pair of plumes are making their appearance. The “body” of the bud here figured was placed so far forward as compared with the stalk of the shield that the plumes, instead of being set on the anterior edge of the “body” as in D, lie in a deep groove between the “body” of the bud and the dorsal wall of the buccal shield. The plumes are drawn in a dotted line to signify that they would be seen in the positions they occupy in the figure if the “body” were transparent; as a matter of fact the “body” is so black that the employment of the usual clarifying reagents fails to make it transparent.. The bud in question was drawn as a whole, with no plumes visible; then the “body” was dissected off and the characters and positions of the plumes noted. This relation of the “body” and shield whereby the plumes are not visible except by dissection is a common one; it is only exceptionally that the “body” is so much drawn back as to expose the developing plumes as completely as in text-fig. 6, D, and in fig. 66 of Plate 7 of the ‘” Discovery” Expedition Reports/ vol. ii, 1907.

The post-oral lamella is not well defined until after the fourth plumes have developed. In text-fig. 7, G, the two lateral flaps are complete, but the free edge connecting the two flaps behind the mouth is not yet entire, and so in making such a dissection as is shown in G and H this part is left with a ragged edge (H, r.e.). The plumes of the first three pairs are elongating and are swelling out so that the base of each appears comparatively narrow (G and H). The fourth plume in G is set between the base of the third plume and the anterior part of the free edge of the post-oral lamella, and so when the fifth plume appears (H, 5) it would seem as though that plume must have arisen from the edge of the post-oral lamella itself. Similarly also with the sixth and seventh plumes.

I have not met with a bud showing the early development of the sixth plumes, so that there is a greater interval than I could have wished between H, with fifth plumes appearing, and J, with seventh plumes appearing.

Up to the time when the fifth plumes are beginning to develop, the size of the plumes diminishes from the first to the fifth (H), but when the seventh is making its appearance (j) the plumes first formed have failed to maintain their initial superiority in size, and the first five pairs of plumes are nearly of equal size.

Pinnules arise rather late. In text-fig. 7, J, the seventh plumes have already appeared, and yet the pinnules of the first plumes are not more than hemispherical projections. On the plumes of the first pair the first ten or twelve pairs of pinnules arise simultaneously, the latexones are added at the basal end of the series. In Cephalodiscus hodgsoni (text-fig. 3, o) and Cephalodiscus dodecalophus (textfig. 4, 1) the pinnules on the first plumes are already fingerlike projections at the time when pinnules first make their appearance on the fifth plumes, and before the axes of the last plumes (the sixth) have shown signs of development. In Cephalodiscus nigrescens, however (text-fig. 7, J), pinnules begin to appear simultaneously on the first five pairs of plumes, and are very little advanced at a time when the last plumes (the seventh in this species) have already come into existence. The bud from which fig. j was drawn is of the same size and general appearance as that represented in fig. 67 of Plate 7 of the ‘Antarctic “Discovery” Expedition Report on Cephalodiscus/ 1907.

In text-fig. 7, J, the first five plumes have elongated considerably since the stage shown in H, and the axes are less bluntly tipped. The interval between the right and left plumes of the first pair is not relatively, but absolutely less than in earlier stages—all these figures are drawn to the same scale of magnification. The exigencies of pictorial delineation demand that the structures under consideration be drawn flat; it is well to bear in mind, therefore, that, while the structures represented near the median line of the figure are on the antero-ventral surface of the “body” of the bud, just above the buccal shield, the lateral parts (plumes 6 and 7, and lateral parts of the post-oral lamella) are set high up the sides of the “body.” The figure if cut out and bent back would give approximately the correct relation of the parts.

In the adult the right and left seventh plumes are as close together as are the right and left first plumes. The line of plume-bases in the adult, when viewed from the front, is an ellipse, incomplete on the upper side; and the line of the attached edge of the post-oral lamella is another ellipse similarly incomplete. The two ellipses join where they are incomplete (see text-fig. 8). The parts seen are, of course, at very different distances from the observer; the base of the fifth plume is in the middle distance, the base of the first plume is nearest to the observer, and the mid-ventral part of the post-oral lamella farthest away. The stalk of the buccal shield and the mouth lie in the middle distance between the two ellipses, and the anus dorsal to both (compare text-fig. 8 and text-fig. 5, D).

The seventh plumes of the adult, and in some polypides the sixth plumes also, do not arise directly from the “body” of the polypide, but are processes of a paired lophophoral upgrowth, which is connected with the “body” at the base of the fifth plume and the more ventral parts of the collaroutgrowths (see “Cephalodiscus,” ‘“Discovery” Expedition Reports/ 1907, p. 31, last paragraph).

Late buds of Cephalodiscus nigrescens are extremely rare in the.material at my disposal. Possibly after the stage of development represented in figure j has been reached the buds migrate. Whatever be the explanation the fact remains that from an abundant supply of buds of various stages I have only discovered one which shows pinnules in a further stage of development than that drawn in figure J. The bud is figured in the ‘Report on Cephalodiscus’ (“Discovery” Expedition, PI. 7, fig. 68). It has fourteen plumes, of approximately the same size and development, standing forward in a bunch, parallel with one another; the stalk of the bud is no longer terminal, but projects from the ventral surface of the “body.” Two of the plumes of this bud are shown in text-fig. 7, x and L; K is a nearly dorsal view, n a nearly lateral view. The plume-axis is very massive and pigmented, and has a smooth hemispherical extremity ; some of the other plumes of this bud, however, have the ends less rounded, more like those of J. There are about seventeen or twenty pairs of pinnules, and the middle ones of the series are the longest.

Cephalodiscus gilchristi resembles Cephalodiscus nigrescens, and differs from Cephalodiscus dodecalophus and Cephalodiscus hodgsoni, in the tardy development of the plumes in comparison with the rapid growth of the buccal shield. The bud, at first pear-shaped (text-fig. 9, A), soon becomes differentiated into a buccal shield and a “body,” with stalk (B). Even at this early stage of development the posterior lobe of the buccal shield is well defined (B, p.l.). A pair of mounds appear on the right and left sides of the “body,” near its junction with the middle of the shield (c and D), and these develop into the first pair of plumes. It is to be noted that these mounds are strictly lateral structures, and do not project forward as do the first plumes of the other three species under consideration during the early stages of growth.

The second pair of plumes arise postero-ventrally to the bases of the first (E and F, 2), and in a dorsal view such as that shown are partially concealed by the first pair. When the third plumes begin to develop (o, 3), the first pair project antero-laterally instead of strictly laterally, possibly owing to the pressure put upon them by the second and third plumes, which arise posteriorly to their bases. Each of the six plumes now present is wider at the middle than at its base.

When the fourth plumes make their appearance (H, 4) the length of the first and second plumes is more than twice their width at the base; and before there is any sign of the fifth plumes pinnules begin to differentiate upon the first two pairs of plumes (J, 1 and 2). The first plumes are now directed more anteriorly than laterally, and the post-oral lamella is clearly distinguishable (J, po.l.).

At a time when the fifth plumes are just appearing (K, 5) the first plumes point well forward and not laterally, the fourth plumes are twice as long as broad, and pinnules are making their appearance on the third plumes. It would seem that the fifth plumes grow very slowly, for in the specimen figured in text-fig. 9, L, the fifth plume on the one side is not much longer than broad, and on the other side it is even less developed, and yet the pinnules on the first pair of plumes are digitate processes, twice as long as broad, and pinnules have already made their appearance on the fourth plumes. The first four pairs of plume-axes are now almost of the same length.

In all of the buds of Cephalodiscus gilchristi, except the very young ones, the surface of the “body” is closely studded over with refractive beads similar to those which occur on the end-bulbs of the plumes of Cephalodiscus dodecalophus and Cephalodiscus hodgsoni. In the present species the refractive beads occurring on the plumes are negligible; the beads occur mostly on the general surface of the body, and, in the adult particularly, on the upper surface of the anterior margin of the buccal shield (see description of Cephalodiscus gilchristi, ‘Marine Investigations in South Africa/ vol. 4, 1906, p. 184). The refractive beads are not shown in text-fig. 9.

Buds in which the fifth plumes are more advanced than in L, and buds in which the sixth plumes are developing I have been unable to find. This is not without significance when taken in conjunction with the fact that in the material of Cephalodiscus nigrescens I found only one late bud showing the plumes more advanced than is represented in text-fig. 7, J, and with the fact that in Cephalodiscus hodgsoni the later stages in plume development, showing the development of the sixth plumes, are not rare. The species gilchristi and nigrescens have tubaria in which the polypides are isolated, and live in separate tubular cavities, which open individually on the surface, and do not communicate the one with the other. The species hodgsoni, like dodecalophus, has a tubarium in which the polypides live in a common cavity, which opens by several ostia to the exterior. In my report on Cephalodiscus in the ‘“Discovery” Expedition Reports’ (pp. 23, 24) I ventured to suggest that, in the species of the former kind (i. e. species of the sub-genus Idiothecia) the late buds severed their connection with the parental stolon, and wandered over the surface of the colony in order to settle down in some convenient situation, usually the apex of a branch, and to secrete tubes of their own. The fact now recorded, namely, the inability to discover in the material of Cephalodiscus. gilchristi any individuals intermediate in growth between buds with small fifth plumes and full-grown polypides with well-developed sixth plumes, supports that suggestion, for the late buds, while developing their fifth and sixth plumes, would be migratory forms on the surface of the colony, and would be brushed off while the specimen was in the trawl. In the species of the sub-genus Demiothecia (e. g. Cephalodiscus hodgsoni and 0. dodecalophus) the late buds would complete their development within the common cavity of the tubarium, and while being drawn up in the trawl, and while undergoing fixation in preservative fluids, would not be more likely to be lost than the younger buds and the adults of the colony.

Harmer in his recent ‘Monograph on the Pterobranchia of the “Siboga” Expedition’ (p. 30) lays stress on the relations of the plumes (or “arms”) to the post-oral lamella (or “operculum”) in Cephalodiscus. He points out that “in Balanoglossus the anterior margin of the collar forms ai, projecting fold encircling the base of the proboscis-stalk. The ventral half of this fold may be regarded as constituting a lower lip, while the dorsal part is connected, in the middle line, with the anterior neuropore. In Cephalodiscus the neuropore is not represented, and the collar forms no projection in the median dorsal line above the base of the proboscis. Except for this interval the whole of the anterior margin of the collar forms a strongly-developed fold, split up dorsally to form the arms, and ventrally constituting the operculum.”

My observations on the development of the plumes and post-oral lamella of buds of Cephalodiscus entirely bear out this contention. While, however, Harmer finds it necessary to insist that the post-oral lamella is a derivative of the anterior edge of the collar, and not of its posterior edge as he formerly supposed,’ the point does not appear to me to be of special consequence. The post-oral lamella is really a pair of ventro-lateral flaps united together behind the mouth, the connection being but a mere hollow ridge. The ridge is certainly connected with the hinder edge of the collar (textfig. 10, po.l.), and the middle part of each lateral flap with the middle of the length of the collar, i.e. not with the front edge nor the hind edge; in the region of the collar canal, which is set on the postero-dorsal edge of the collar, about midway between the dorsal and ventral surfaces, the base of the flap is near the front edge. But the relations of the plumes to the post-oral lamella can be clearly understood in spite of this; and it is a very significant circumstance, in connection with the view that the post-oral lamella and the plumes belong to one and the same system, that in the development of the buds of Cep halo discus each new plume arises between the last-formed plume and the end of the post-oral lamella; one might almost say that each new plume is differentiated from that part of the post-oral lamella immediately in contact with the last-formed plume. The post-oral lamella may thus be regarded as composed of postero-ventral plumes which fail to differentiate as separate plumes.

In the adult the food-grooves of the axes of the first and second pairs of plumes face ventrally, i.e. towards the dorsal wall of the front part of the shield; the grooves of the third and fourth pairs face laterally, and those of the fifth and sixth pairs dorsally or dorso-laterally. Tracing these grooves basally one finds that those of the right-hand plumes bend round and converge to the right side of the mouth, and those of the plumes of the left side of the body to the left side of the mouth, the food current being in all probability guided into the mouth and prevented from escaping by the free edge of the post-oral lamella being at the time in close contact all round with the posterior flap of the buccal shield. The lateral lobe of the post-oral lamella thus keeps the food current distinct from the exhalent current through the gill slit and from the water passing in or out of the collar canal, both the gill slit and the collar pore being situated posterior to the lateral lobe of the post-oral lamella.

If Phoronis is justly to be regarded as a member of the Hemichordata, and the tentacles are collar-structures comparable with the plume-axes of Cephalodiscus, one must admit, with Harmer,1 that what is in Cephalodiscus known as the post-oral lamella is in Phoronis produced into a row of tentacles instead of a paired flap. On this admission there is no great difficulty in effecting a comparison between the plume-systems of the two.

The line of the bases of the tentacles of Phoronis runs in the double spiral well shown in Benham’s figure of the end view of the animal with the tentacles cut short.2  This figure, if inverted, is closely comparable with text-fig. 8 of this paper. Below the mouth is a row of tentacles, corresponding with the post-oral lamella of Cephalodiscus, continued on each side to the centre of the spiral, which point is equivalent to the uppermost limit of the post-oral lamella in text-fig. 8. Here it becomes continuous with the tentacles of the supraoral series, as in Cephalodiscus the extremities of the base of the post-oral lamella are continuous with the terminal members of the series of plumes. The notch in the supraoral series of tentacles, marked x in Benham’s figure, is the equivalent of the space between the right and left plumes of the first pair in Cephalodiscus.

Not only has this correspondence between plumes and postoral lamella of Cephalodiscus with the two series of tentacles of Phoronis been indicated by Harmer, but that author also directs attention1  to a still more remarkable similarity between the processes of the lophophore and the lower lip of the Sipunculoid Phymosoma, described by Shipley,2  and the plume-axes and post-oral lamella of Cephalodiscus. The endsof the lower lip of Phymosoma are continuous with the ends of the series of finger-like processes of the lophophore, and Shipley’s figure 32 on Plate 4 of his paper bears a close similarity to my text-figure 8. It may be not without significance that, when viewed from the front, the nephridiopores of Phoronis and of Phymosoma occupy the same positions with regard to the anus and other parts as do the gonadial apertures of Cephalodiscus.

While the position of the epistome of Phoronis does not militate against the view here favoured of an affinity between Phoronis and Cephalodiscus, it is to be noted that, according to the observations of de Selys Longchamps,3  the pre-oral lobe of the larva Actinotrocha, which is homologous with the proboscis of Balanoglossus and the buccal shield of Cephalodiscus, disappears during the metamorphosis, and does not persist as the epistome of the adult Phoronis.

Although the occurrence of the five divisions of the coelom in Actinotrocha, corresponding with the proboscis cavity, two collar cavities and two trunk cavities, claimed by Masterman,4  has been contested by Goodrich,5  Ikeda,6  and de Selys Longchamps,7  these last three writers agree that there is a division of the cœlom into a pre-septal and a postseptal part, the septum in question following the course of the tentacles of the larva. It is still possible, therefore, that the pre-septal cavity may represent the pair of collar cavities of Balanoglossus and Cephalodiscus.1  The obliquity of the ccolomio septum of the Actinotrocha is exactly similar to the obliquity of the boundary between the collar cavities and trunk cavities of Cephalodiscus (sec the upper line of dashes in text-fig. 10), the dorsal part being much more anterior than the ventral.

A comparison of such published figures of buds of Rhabdopleura as show developing plumes, with the buds of Cephalodiscus described in the earlier part of this paper, confirms the general impression that the two plumes of the adult Rhabdopleura are equivalent to the first pair of plumes of Cephalodiscus. The plumes of Rhabdopleura arise as a pair of digit-like processes (see text-fig. 11), close together, and dorsal to the shield; later on the pinnules develop in pairs, and in the adult the pinnules are directed ventrally towards the shield, exactly as are the pinnules of the plumes of the first pair in Cephalodiscus. The second and later plumes of Cephalodiscus are not represented in Rhabdopleura. Besides the figures of buds of Rhabdopleura published by Lankester and reproduced here as textfig. 11, those by Allman 1  and Schepotieff 2  may be consulted with advantage.

1. The torsion of the axes of the first and second plumes of the buds of Cephalodiscus described by Masterman does not take place. The grooved faces of the axes of these plumes are in the first instance directed towards the dorsal face of the buccal shield, and they maintain this relation through life.

2. The last two pairs of plumes do not arise between the first two pairs of plumes and the buccal shield, as described by Masterman, but they arise on the dorsal side of those plumes, the side remote from the shield. Their grooves are directed, not towards the shield, but away from it.

3. The evidence derived from a study of the buds of Cephalodiscus bears out the contention of Harmer that the series of plumes and post-oral lamella are continuous. The plume-axes develop in pairs successively, the median pair first, then a pair lateral to these, and so on. When the post-oral lamella appears, usually at the time when the third oi’ fourth pair of plume-axes are beginning to develop, its edge is in contact with the last-developed pair of plume-axes. The fifth plume-axis appears between the fourth plume-axis and the end of the margin of the post-oral lamella, the sixth between the fifth and the end of the post-oral lamella, and, in the case of Cephalodiscus nigrescens, in which there are seven pairs of plumes in the adult, the seventh arises between the sixth and the end of the margin of the postoral lamella.

4. The line of plume-bases in the adult, when viewed from the front, is an ellipse, incomplete on the upper side (textfig. 8) ; and the line of the attached edge of. the post-oral lamella is another ellipse, similarly incomplete. The two ellipses join where they are incomplete.

5. Separate accounts are given in this paper of the plumedevelopment in buds of Cephalodiscus hodgsoni, C. dodecalophus, 0. nigrescens, and C. gilchristi.

6. The plumes develop relatively later in buds of 0. nigrescens and 0. gilchristi than those of C. hodgsoni and C. dodecalophus.

EXPLANATION or THE ABBREVIATIONS EMPLOYED IN TEXT-EIGURES

2, 3, 4, 6, 7, and 9.

1, 2, 3, 4, 5, 6, 7. The first, second—seventh pairs of plumes of the buds. b. The “body” of the bud. b.s. Front lobe or main portion of the buccal shield, m. Mouth, or position of mouth so far as can be ascertained in the dissected preparations, this region being much disturbed during the dissection. p. Pinnules, p.a. Plume-axis. p.l. Posterior lobe of buccal shield, po.l. Post-oral lamella, r.b. Refractive beads in the swollen ends of the plumeaxes in C. hodgsoni and 0. dodecalophus. r.e. Ragged edge caused by dissection of the part figured from the rest of the bud. r.l. Red line of the buccal shield, s. Stalk of the bud.

1

‘Trans. Roy. Soc. Edin.,’’ vol. xxxix, part 3, 1898, p. 521.

2

“Discovery”Expedition Reports,’’vol. ii, 1007, and ‘Marine Investigations in South Africa,’ vol. iv, 1906.

1

The close relation obtaining between the base of the last-formed plume and the edge of the post-oral lamella was, I believe, not known until it was pointed out by Harmer two years ago (‘Pterobranchja of the “Siboga” Expedition,’ 1905, p. 36),

1

‘Trans. Roy. Soc. Edin.,’ vol. xxxix, 1898, pl. 2, fig. 24.

1

‘Trans. Roy. Soc. Edin.,’ vol. xxxix, 1898, p. 521.

1

‘Pterobranchia of the “Siboga” Expedition,’ 1905, pp. 36, 37.

2

‘Quart. Journ. Mic. Sei.,’ vol. 40, 1S97, p. 346, pl. 26, fig. 36.

1

’;“Challenger” Reports/ vol. xx, part 62, 1887, p. 43.

1

It is to be noted tbat Harmer orientates the polypide of Cephalodiscus in such a way that the antero-posterior axi3 of the body is a line passing from the middle of the buccal shield through the central nerve mass, and ending on the rectal side of the body a little below the anus, so that the visceral mass is regarded as a ventral downgrowth (’ Pterobranchia of the “Siboga” Expedition,’ 1905, p. 23). In the present paper the ordinary orientation is adopted, namely, plume-apices anterior, stolon and face of shield ventral, intestine dorsal, rounded end of visceral mass posterior.

1

‘Pterobranchia of the “Siboga” Expedition,’ 1905, pp. 116,117.

2

‘Quart. Journ. Mic. Sci.,’vol. 30, 1889, pi. 10, fig. 7.

1

L. c., p. 119.

2

‘Quart. Journ. Mic. Sci.,’ vol. 31, 1890, figs. 1 and 2.

5

“Développement Postembryonnaire et Affinités des Phoronis,” ‘Mém. Classe Sci. Acad. Roy. Belgique,’ i, 1904, p. 73.

4

‘Proc. Roy. Soc. Edinb.,’ vol. xxi, 1896, pp. 62, 63, and 130; also ‘Quart. Journ. Mic. Sci.,’ vol. 43, 1900, p. 395.

5

‘Quart. Journ. Mic. Sci.,’ vol. 47, 1904, p. 111.

6

‘Journ. Coll. Sci. Imp. Univ. Tokyo,’ xiii, part 4, 1901, pp. 540 et scq.

7

Loc. cit., pp. 15,16.

1

See Fowler,‘Fcslschr. 70tcn Geburtstage R. Lcuckarts,’ 1892, p. 297, and Harmer, ‘Pterobranchia of the “Siboga” Expedition,’1905, p. 116.

1

‘Quart. Journ. Mie. Sei.,’ vol. 9, 1869, pl. 8, figs. 7 and 8.

2

‘Zool. Anzeiger,’ vol. xxviii, 1905, p. 802, fig. 6.