II.—Spicule Formation in Alcyonium digitatum; with Remarks on the Histology

A. von Kolliker (1), who was one of the first to study Alcyonarian spicules, supplied in his ‘leones histologicæ’ (published in 1864) a full description and classification of their various forms and a detailed account of their structure, but did not investigate their development. According to Kolliker, an Alcyonarian spicule consists of an organic and an inorganic part, the former, constituting the bulk of the spicule, being crystalline in nature and with a complicated structure, the latter consisting merely of a cuticular spicule sheath. Kolliker thus denied the existence of one of the most conspicuous features of this type of spicule, viz. the large central organic axis; this, however, is not surprising since he also failed to observe the still more conspicuous fact of the enclosure of these spicules within a well-defined protoplasmic layer containing granules and one or more nuclei, and therefore of their intra-cellular origin. Nevertheless, many of Kolliker’s observations on the more intimate structure of the Alcyonarian spicule are correct, as instance, e. g. his statements as to the concentric lamellar structure of the spicule body when seen in transverse section, the breaking up of the spicule into smaller crystalline elements on treatment with weak acids, and the “axial “continuity of the processes of the spicule with the main trunk, resembling, as Bourne remarks, the origin of secondary roots in a plant.

G. von Koch (2) was the next investigator who contributed to our knowledge of Alcyonarian spicules. He ascertained to some extent, and for the first time, the development of the spicules in Clavularia prolifera, giving in his paper on the histology of this species highly-coloured figures of the earliest stages of spicule formation and of the subsequent growth. He rightly insisted on the endoplastic development of the spicules, showing that these not only originate in scleroblasts derived from the ectoderm, but remain conspicuously enveloped by the cell-substance throughout their existence. Koch’s figures show well the presence of nuclei—never exceeding two in number it is important to note—in the granular protoplasmic layer investing the spicule, and also incidentally prove that the change in shape of the spicule from the spherical to the elongated is strictly correlated with the division into two of the nucleus of the original mother-cell. It is also shown that spicules situated in different portions of the colony (and so subject to different environ-mental influences) are different in form—spicules situated near the surface being monaxon and unbranched, spicules more deeply situated being more or less branched and irregular.

Kowalevsky and Marion (3) in 1883 confirmed Koch s statements as to the intra-cellular growth of Alcyonarian spicules, although they gave no figures in support of their statements.

In 1892 K. C. M. Schneider (4), in a very short paper, and also without giving any figures to justify his statements, asserted that in development the scleroblast gradually assumed the form of the adult spicule, which latter was then produced by the calcification of the cell-substance. As Hickson (5) remarks, the appearance of developing spicules does not by any means justify this assertion, and it is, moreover, grossly improbable that the actual protoplasm should become converted into the crystalline substance of the skeleton.

In 1899 G. C. Bourne (6) made some observations on the formation of the spicules in Alcyonium digitatum and Gorgonia cavolinii preparatory to investigating the minute structure of the Madreporarian corallum, and, although only incidentally made, these observations of Bourne up to the present afford the most complete account of the subject extant.

In the following lines I shall occasionally have to refer to certain of the statements made by Bourne and less often by Hickson.

Bourne’s method of preparing microscopic specimens of Alcyonium digitatum for examination of the developing spicules was as follows:—“The specimens were killed in an expanded condition by rapid immersion in a ·5 per cent, solution of osmio acid in sea water, were thoroughly washed with distilled water and stained for twenty minutes with Ranvier’s picro-carmine. The expanded polyps were cut off close to their bases, placed in dilute glycerine, and laid open. The ectoderm and endoderm having been removed with a camel’s-hair brush, very satisfactory flat preparations were obtained illustrating the formation of the spicules in the lower moities of the exserted portions of the polyps.” My expe-rience of this method of making flat preparations of Alcyonium digitatum for the study of spicule formation is that it is by no means a satisfactory one—certainly not so satisfactory as that which I have adopted, and which I will shortly describe. For, in the first place, spicules do not occur to any great extent in the region of the retractile polyps—as, indeed, Bourne admits when he states that “one of the smallest sclerites which I was able to discover is shown …,” whereas in my own preparations I have been able to count such by the dozen—and hence there exists a difficulty in observing all the stages of spicule formation; secondly, what spicules do occur here are not very characteristic in shape of the greater number of the spicules found in the mass of the colony—there existing a tendency for spicules situated in the region of the polyps to be smooth and spindle-shaped—and hence the development of these others cannot from these data be assured.

Hickson remarks that “Alcyonium digitatum is not a favourable form to take for the study of the development of the spicules, as it is a matter of very great difficulty to make a thin section of the surface of the colony before decalcification,” but, owing to the transparency of the mesoglœa, it is not necessary to cut thin sections for this purpose, as I think is proved by the fact that all the figures supplied in the plates accompanying the present paper were carefully drawn by means of a camera lucida from free-hand sections which measured anything from 10 μ to 50 μ in thickness. The method of preparing microscopic slides of Alcyonium digitatum¹ which I adopted was practically identical with Bourne’s as regards the fixation and staining of the specimens, but different as regards the region of the colony selected for examination and the manner of obtaining thin portions of it. Young colonies only were selected (an important point), and these were simply fixed by sudden immersion in 1 per cent, osmic; they were then thoroughly washed in distilled water, and deposited in either Ranvier’s or Weigert’s picro-carmine for three hours. When thus effectually stained the colonies were again washed and carefully graded up to absolute alcohol. The spicules offering too great an obstruction to microtome section-cutting, the colonies were transferred straight from the absolute alcohol to coco-butter, in which substance they were held whilst free-hand sections of them were made. These sections were then placed in xylol, and finally mounted in Canada balsam. By this method excellent preparations were obtained, the nuclei, cell-plasma, and organic axes appertaining to the numerous spicules being rendered very evident. I occasionally substituted paraffin-wax for the coco-butter, but the latter is preferable in many respects, and, moreover, is more economical as regards time.

Before proceediug to the description of the development of the spicules in Alcyonium digitatum, it will be as well to give a brief account of the general histology, if only merely in order to distinguish the spicule-forming cells or scleroblasts from neighbouring endoderm-cells, etc.—a distinction that has not always been made. The ectoderm, which forms the limiting layer of the mesoglœa “consists,” according to Hickson, “of a number of columnar, spindle-shaped, and flask-shaped cells … connected at their outer borders, but free from one another for the greater part of their course. At the base of the epithelium there are a few spherical interstitial cells of different sizes.” In the course of my observations of the spicules I came across a number of exceedingly small cells—some even smaller than the nuclei of other cells—which were situated in the peripheral layers of the mesoglcea. These cells, which I have represented in Pl. 17, fig. 20, were slightly granulated, and many contained a distinct though faintly-stained nucleus; in others, however, it was impossible to distinguish one. It is probable that these are the interstitial cells of Hickson, more especially since, owing to variations in size, it would be possible to trace cells inter-mediate in size and appearance between these and the scleroblasts they in all probability give rise to. Hickson says that “it is difficult to determine with certainty the origin of the cells that give rise to the spicules, but, for many reasons, I am inclined to agree with von Koch’s results on Gorgonia and Clavularia, and attribute them entirely to the ectoderm. Among the interstitial cells of this tissue one frequently finds large spherical cells which lie beneath their neighbours, and cells very similar to these may be seen isolated in the subjacent mesoglcea.” These scleroblastic mother-cells are represented in Pl. 17, fig. 19, and Pl. 16, fig. 1.

In addition to the epidermal cells, the interstitial cells, and the scleroblasts which they probably give rise to, and two or three other kinds of cells, which I shall mention, there exist, according to Hickson (and I can confirm his statements), in the peripheral portion of the colony, the endodermal canals, the “solid cords of endoderm,” and “isolated cells connected with one another and with the endoderm by fine anastomosing fibrils.” “The cords [see fig. 18] are, in some, fairly compact, resembling a canal in all respects except the presence of a lumen, but in others the cells are only loosely connected with one another, become elongated or star-shaped, giving off fine fibrils at their angles. There may be only a single row of oval or cubical cells, or in some cases the row may be drawn out into a chain of elongated spindle-shaped cells.” Bourne mistook these strands of endoderm cells for scleroblasts—“the scleroblasts have the form of irregularly polygonal, ovate, or amœbiform cells, varying very much in size and shape; they run in strands and patches through the mesoglcea at the bases of the expanded polyps, and may be found, though they are not easily studied, in the thickened mesoglœa of the cænenchyme” (6)—but the endoderm cells as a whole, besides often being multi-nucleated (which the scleroblasts never are until spicule-formation is fairly advanced, and then only two nuclei are present) are also considerably larger and much more irregular in shape than the true scleroblasts, which are approximately spherical. It is true that in many cases these endoderm cells possess large vacuoles, which, but for the absence of refringency, might be mistaken, at first glance, for young spicules, but, personally, I cannot call to mind having ever seen a real spicule contained in one of these irregular cells.

Bourne describes, in the paper before referred to, two other kinds of cells—one possessing refringent granules and the other containing the ovoid bodies—only the latter of which Hickson refers to. Bourne remarks that the cells of the first kind are “rather smaller than, but of similar shape to, the scleroblasts,” which, since Bourne reckons the endo-derm cells as scleroblasts, must therefore be somewhat irregular in form, as his figures indicate. These cells are said to be “filled with minute highly-refringent granules,” and their nucleus is “rarely to be seen, being hidden by the granules.” Bourne believes that “their function is to secrete the gelatinoid substance of the mesoglœa.” To some extent I can confirm Bourne’s statements, since I also have found peculiar cells containing very distinct minute granules and a faintly-staining nucleus, and which are “of similar’ shape to the scleroblasts.” The cells I recognise as scleroblasts are, however, spherical, and so are these cells containing the distinct granules; whether they are the same as the jelly-secreting cells of Bourne I cannot decide. I have shown three of these cells, which are somewhat bladder-like in appearance, in fig. 21.

Hickson states that the “nematocysts of Alcyonium are extremely small (0·0075 mm.) and all of one kind,” and describes those situated in the ectoderm of the tentacles. In addition to these nematocysts Hickson describes certain “oval bodies lying in and among the cells of the endodermic cords.” He adds that “they may be readily distinguished from the endoderm cells by their dark but homogeneous appearance,” and that “in one or two instances he succeeded in making out a somewhat irregular body in the centre, which maybe anucleus.” Hickson suggested that“they maybe some form of parasitic sporozoon.” Bourne also noticed these “ovoid bodies,” ¹ as he calls them, and stated quite correctly that “each … is surrounded by a protoplasmic sheath, and has a relatively large nucleus on one side of it,” and that they are not calcareous in nature. “They stain deeply with hæmatoxylin,” and (contrary to Bourne’s statement) also with picro-carmine, methylene blue, and some other stains. Bourne hazards a guess that these “ovoid bodies “are degenerate nematocysts. I also have observed these very conspicuous bodies, illustrations of which I have supplied in figs. 22 and 23. In one respect only can I improve upon the description of Bourne, and that is the observation of a structure distinguishing the cells containing “ovoid bodies,” situated at the edge from those situated more or less deeply in the mesoglœal substance. In the former I have observed in nearly every instance the presence of a distinct cnidocil, as shown in fig. 23, but I cannot detect this structure in those more internally situated, and it doubtless does not exist in their case. The presence of this cnidocil I think distinctly proves tthey become more and more separated he nematocyst nature of the “ovoid bodies ”of Alcyonium, which thus, according to Hickson’s account, possesses two kinds of nematocysts—one restricted more to the region of the tenta-cles, the other to the mass of the colony.

Thus, in Alcyonium digitatum, in addition to the scleroblasts, there exist in the mesoglœal substance, endoderm cells, the spherical “jelly-secreting” (Bourne) cells, the small interstitial cells and the nematocysts or “ovoid bodies,” all of which latter are more or less distinguishable from the former.

The scleroblasts, as already stated, are granular, more or less spherical cells situated at the periphery under the ectoderm, and probably derived, as Hickson suggests, from the interstitial cells of that layer. Further, “the spicules are far more numerous at the periphery than in the deeper parts of the colony. This suggests very forcibly that the spicules are only formed at the periphery, and that with the growth of the mesoglœa they become more and more separated from one another”—a suggestion I can amply confirm. The spicule first appears in the cytoplasm as a small spherical concretion (figs. 1 and 2), and remains approximately spherical in its further enlargement until the division into two of the nucleus (fig. 3).

This last statement, I am aware, is contrary to the account of early spicule formation given by Bourne, who supplies figures representing the young spicule as of very irregular shape long before nuclear division occurs. One possible explanation of this difference of form in the two cases is the difference of situation of the young spicules drawn by Bourne and of those drawn by myself—Bourne’s spicules being localised in a mobile portion of the colony (the bases of the polyps, where the mother-cell is obviously subject to disturbing influences which might very possibly lead to irregularity of form of the contained spicule) and mine in a quiescent. But, apart from this, some of Bourne’s figures (e. g. his fig. 4) certainly suggest the idea of his spicules having been subjected to rough treatment of an artificial kind, such as the stretching open of the polyp, and the brushing of its internal wall might by chance involve, but whether the fractured appearance of the spicules suggested by Bourne’s figures is due to this, or to the above-mentioned cause or to the mode of drawing, I am unable to say. Certainly the young spicules observed by me did not exhibit this broken irregular appearance.

As just stated, up to the division of the scleroblast nucleus, the form of the young spicule remains smooth and approximately spherical, but when this change in the nucleus occurs, the spicule becomes elongated and somewhat dumb-bell in shape (not dumb-bell in shape in the sense in which Hickson refers to adult spicules), i. e. thickened and rounded at the two extremities, and the two nuclei, which form centres for the aggregation of the cytoplasm, in general travel to its opposite ends. This simple dumb-bell then enlargens and its two extremities become amphicœlous by the development of a broad rim round the terminal surface of each “head” of the dumb-bell (figs. 3 and 4). Following on this again, the rim on the terminal surface becomes developed into two, three, four, or more processes, some, or occasionally all, of which afterwards become the main branches of the spicule; other smaller processes may also appear, and the spicule now assumes the form depicted in figs. 5 and 7. This stage thus reached forms in the vast majority of cases the basis of all future development, since it is a ground plan common to, and recognisable in, all spicules whatever shape they may ultimately assume (with the possible exception of the spindles presently to be described), and is, in fact, the starting point of the morphological differentiation of the spicules. In form it somewhat resembles, as Bourne suggests, a caudal vertebra, and like all preceding and succeeding stages is enclosed in a granular protoplasmic sheath containing two nuclei, which, in the vast majority of cases, are situated at the two extremities of the dumb-bell basis.

It is au interesting fact that in all succeeding stages of spicule-formation—in all the varied and complicated forms which adult spicules assume—only two nuclei are present. It is important to insist upon this point since Bourne states that “in older and more complicated spicules I have [he has] counted three or four nuclei.” Personally, in all the hundreds of spicules which I have observed I have only once, possibly twice, chanced upon one possessing more than two nuclei; in one case the spicule was at the “caudal vertebra” stage of development, and of the ordinary form, although perhaps with some extra processes developed, and possessed as many as six nuclei, all apparently contained within the protoplasmic in-vestment of the spicule; in another case the spicule possibly possessed four nuclei, although of this I am not so certain. In both of these cases I may of course have been mistaken—may have misjudged adjacent free scleroblasts as cells appertaining to the spicule,—but in the former example, which I most carefully observed and drew, I do not think I was; in any case the rarity of such multinucleated spicules quite disproves Bourne’s statement that such are of common occurrence. The fact that Bourne mistook endoderm cells for scleroblasts¹—“the scleroblasts are often coenocytes containing two, three, or more nuclei”—was probably the source of his erroneous assumption on this head, despite the admission that “the nucleus apparently divides when the sclerite has attained a certain size,” and the erroneous statement that “this division is repeated as growth [of the spicule] continues.” I may here point out in this connection that von Koch also figures two nuclei as the maximum number attained in the formation of the comparatively simple spicules of Clavularia prolifera. Why only two nuclei are concerned in the formation of these Alcyonarian spicules I cannot at present say, and I have therefore no à priori objection to the existence of multinucleated spicules, and am quite ready to accept provisionally, e. g. Bourne’s other statement that, so far as he has been able to ascertain, the “scale-like spicules in Primnoa and Plumarella are formed by several cells, or at least by a comparatively large cœnocytial investment containing many nuclei; “all I at present contend is that, strange as the fact may appear, the huge spicules of Alcyonium digitatum never, in the ordinary course of things, possess more than two nuclei embedded in the wall of the protoplasmic sac which envelops them.

Starting from the “caudal vertebra” stage, different spicules develop in different directions assuming unlike forms. Nearly, if not all, the different forms are derived from the “caudal vertebra” condition by the special development either (1) of the large processes (see text-fig. 0 below) derived, as just described, from the rims of the amphicoelous extremities (some of the minor processes occasionally replace these, however, as the figures show), i. e. of some or all of the four angles or corners of the spicule basis when this is viewed from a lateral aspect, or (2) of one or both of its two ends, or (3) of the two sets of processes combined. In fig. 8, e. g. two of the diagonally opposite angles have become specially developed, and similarly in fig. 9, though here, owing to the spicule lying edge-on, the two processes resemble the elongated ends of the spicule and not its corners; in fig. 10 two of the angles on the same side of the spicule, but not in the same plane; in fig. 11 two of the angles at the same end of the spicule and the opposite extremity, and similarly in fig. 12; in fig. 13 three of the four angles; in fig. 14 all of the four angles, two more so than the opposite pair. It will be seen from the figures that in the growth of these angles or ends, the actual prolongation may not, as before pointed out, be derived from the original “corner “or end, but from a process developed to one or the other side of it. In some cases also the rounded extremity of the simple dumb-bell divides into (i. e. the rim develops into) many more than three or four regions, and the resulting spicule is then more complicated at its extremities, but this is rare in A., digitatum.

In every case I failed to observe any relationship between the position of the nuclei and the development of certain processes, and it seems certain that after the first elongation of the spicule the nuclei play no further part in determining its form.

On observing a section of Alcyonium digitatum containing spicules, preferably under a low power of the microscope (say × 500 diameters), one cannot fail to notice that each of the larger spicules is contained in a distinct cavity apparently formed in the substance of the mesoglœa by, and during the growth of, the spicule itself, since the thickened edges of the mesoglœal substance constituting the walls of this cavity are distinctly supported by the spicular processes —like tent canvas on a pole (see figs. 16 and 17),—and hence must have been displaced by their formation. This simple fact¹ serves to throw some light upon the physical consistency of the mesogloeal substance, and possibly has a bearing on the problem as to the causes of the various forms which spicules assume (see further).

The figures of the spicules provided in the plates illustrate well another fact of importance, and that is the large amount of organic matter contained in the substance of the spicules. Fig. 6 shows in optical section the concentric structure (probably indicating periodic growth) of the spicule axis, which stains pink with picro-carmine, and it is owing to the presence of this axis that all the spicules appear pink in colour. The width of the axis, as compared with the width of the spicule, is variable in different spicules, it being noticeably small, e. g. in the lancet-shaped or monaxon spicules (fig. 15).

I have never observed any horny sheath enclosing the spicule, such as that, e. g. figured by von Koch in the case of the spicules of Clavularia prolifera, and since, up to the time when the spicule attains its maximum size, the whole surface-area of the spicule is having in a varying degree fresh calcareous matter deposited upon it by the investing protoplasmic layer, I am inclined to think it does not possess one—at least not one of any considerable thickness.

With regard to the position of the nuclei on the more adult spicules, there is only one remark to make, and that is that in general one nucleus is to be found at one extremity of the “caudal vertebra” body of the spicule, and one on a process at the opposite end, but this position of the nuclei is not an absolutely constant one, and is probably of little or no significance.

The intimate structure of the Alcyonarian spicule has been investigated by Bourne and others, and since I have nothing to add to their accounts, there is no need for me to here describe this.

Like most, if not all, calcareous spicules which originate in connection with an isolated more or less spherical mass of protoplasm containing one nucleus—a single cell—the Alcyonarian spicules at their first appearance are approximately spherical in form. Further, as is also universally the case, bi-division of the nucleus of this cell¹ containing the spherical sclerite is followed by an elongation of the spicule more or less in the direction of nuclear division. And, in Alcyonium at least, it is not difficult to see why this should be so, for the scleroblast resembling all other isolated cells in its mode of fission—the mass of the cytoplasm becoming concentrated at two centres (each with its nucleus), and therefore attenuated in the region situated between these,—and lime salts being deposited most freely where the bulk of the cytoplasm is greatest, the sclerite must obviously, under these conditions, not only, like the cytoplasm, become elongated, but also, like the cytoplasm, become dumb-bell shaped, as we know it does (text-fig. A). So far, then, the Alcyonarian spicule follows the normal course of development. But, as before stated, beyond the dumb-bell stage of growth, the position of the two nuclei does not seem to exert any influence on the form gradually assumed by the spicule—processes being developed on all sides quite irrespective of the two thickenings of the general protoplasmic investment containing these bodies,—and the further morphogenesis appears to be solely related to external factors. In what manner, however, environmental conditions produce the various forms characteristic of adult Alcyonarian spicules is a question somewhat difficult to answer, and at present I can only supply a suggestion or two towards the solution of this problem.

Two general principles at once present themselves for consideration, the first of which is that growing spicules situated in a mass of mesoglœal substance far removed from any limiting surface must, owing to the proximity of other spicules, endodermal canals, and other heterogeneities of constitution of the surrounding medium, necessarily be subject to an aggregate of influences which tend to produce irregularity of form; and the second is that the extension of a growing body into a surrounding resistant medium is most easily effected by the protrusion of more or less acute processes which, in virtue of their acuteness, are best able to cleave a passage. In connection with this last principle, it may be pointed out that the form of the Alcyonarian spicule certainly suggests the idea that the minor processes (doubtless due to that localised activity of the protoplasm which remains un-absorbed in the prolongation of the main branches) may be attributable to the same cause as say the lobose pseudopodia of a rhizopod—the protoplasmic investment of the spicule being comparable to the ectosarc in which the protrusive movement originates, and the deposited calcareous matter to the stream of endosarcal granules which follows in its train,—and in all probability there does exist an affinity between the two kinds of emergencies.

As has already been described, the extremities of the young dumb-bell spicule quickly become amphicœlous in character, or, in other words, a broad rim usually develops around the hemispherical “head “of each dumb-bell rendering the terminal surface more or less concave (fig. 3), and a few minor processes developing at the “corners” of the spicule, the entire structure comes to resemble, as also before remarked, a caudal vertebra. A possible explanation of this change from the simple dumb-bell to the amphicœlous condition is to be found in the fact stated above, viz. that the consistency of the mesoglœal substance is largely solid, and not liquid in nature (figs. 16 and 17), from which it follows that in the elongation of the binucleated scleroblast and its contained sclerite, the resistant pressure offered by the mesoglœa is greatest at the extreme ends of the spicule, and therefore less towards the sides (i. e. the pressure is not the same at all points as would be the case were the medium liquid), and seeing that, at the same time, the mass of the protoplasm is situated towards the extremities (being pushed somewhat to the sides however by the terminal pressure), further deposition of the calcareous substance must, under these two conditions, take place at the sides and towards the ends of the spicule, i. e. the terminal surfaces will become amphicœlous (see text-fig. B).

Now, according to the second of the two above stated principles, further growth of the spicule substance, i. e. further protrusion, of the protoplasmic investment into the surrounding medium, will take place at that portion of the spicule surface possessing the greatest angularity or rather convexity, i. e. at the rim of the amphicœlous extremity, or in other words, at any or all of the four “corners” always visible when the spicule is viewed from a side aspect. If the external medium were absolutely homogeneous in constitution, extension of the spicule substance would, under such conditions, take place uniformly at the rim of the spicule extremity, and a hollow cone-shaped structure would be produced, but in Alcyonium digitatum and the vast majority of other Alcyonaria, the surrounding mesoglœa is not by any means homogeneous in nature, and hence the extension of the amphicoelous extremities of the spicule is not uniform; on the contrary, the “head” of the spicule, or rather rim, becomes divided into two, three, four, or more regions producing the large processes in every case observable (figs. 4, 5, and 7, and text-fig. C above). Any spicule at this stage seen sideways from any aspect will thus present the appearance of the ordinary “caudal vertebra,” with its concave-headed dumb-bell basis, and more or less well-developed processes projecting from two, three, or all of its four angles according to the number of these developed, and to the side of the spicule presented to observation.

From the “caudal vertebra” stage onwards, it apparently depends entirely on the influence of the heterogeneous constitution of the surrounding medium which of the angles or corners shall develop into the elongated branches proceeding from the main body or basis of the spicule, and which shall remain inconspicuous. In some cases indeed, the extremities of the dumb-bell elongate, but only very occasionally in deeply-situated spicules. Adjacent spicules and other structures must exercise an influence on a developing spicule, and therefore determine to some extent at least its mode of growth, and seeing that the extreme irregularity of form found among the numerous spicules (“no two spicules in the field of the microscope are alike,” Hickson) is alone congruous with the view that such is related to the irregularity of the surrounding conditions—different conditions necessarily obtaining in the case of each spicule—this view is doubtless the correct one.

The alternative view that the furcations of the Alcyonium spicule are to be accounted for in the same manner as, e. g., the dichotomy of the lower Cryptogams, seems to me unlikely since, in this latter case, the branching is mostly related to food-supply—a large surface area being essential to the efficient nourishment of a large bulk—but in the case of the spicule the internal mass does not require nourishment and hence the increase of surface area of the cell-substance cannot be related to that end. I regard the cell-substance investing the spicule as unavoidably secreting lime—the cell “cannot help it “—and the faster it secretes the better is it nourished, and therefore able to secrete, since additional secretion of lime implies further increase of surface area, and this (although the scleroblast is not a plant) is favourable to the nourishment of the cytoplasm. However, as there is a limit to the divisibility of a Paramæcium, however well nourished,¹ so there is a limit to the growth, and therefore secretory power of the enormously-distended scleroblast with its two nuclei, which, judging from appearances in some of the full-sized spicules I have seen, like the Paramæcium, eventually degenerates.

The massiveness and irregularity of the majority of the spicules in Alcyonium digitatum is due to the fact already suggested that all older spicules are situated in mesoglœal substance far removed from any limiting layer, the colony of A. digitatum being, as is well known, exceedingly massive in form and not racemose. When, however, spicules have existed for the greater part of their development in close proximity to the external ectoderm or to the endodermal layer of a gastral canal, irregularity of form largely disappears and the spicules assume the form of a simple manaxon either with very minute processes or without any (fig. 15). As Hickson says, “there may be seen a certain number of long unbranched lancet-shaped spicules covered with irregular tuberosities. They vary very much in shape, some being like a thick pin (without its head) others lancet- or spindle-shaped, and others again slightly curved like a boomerang. These long unbranched spicules occur chiefly in the tentacles and disc of the polyps [where the mesoglcea is evidently of narrow width]. They do occur in other parts of the colony [in proximity to the endoderm canals], but I do not remember to have seen any dumb-bell-shaped spicules in the extensible portion of the polyps.” Again, in the other English species, Alcyonium glomeratum, which is much more racemose in form than A. digitatum, and in which consequently the mesoglœa does not form such a solid mass enabling the spicules to be distantly situated from any limiting layer, “the majority of the spicules are elongated needles and spindles, and there is an entire absence of the small dumb-bell-shaped forms, very few Ks. and crosses [i. e. both resembling figs. 13 and 14], and there are several club-shaped forms [essentially monaxon spicules with fairly conspicuous processes and thickened towards one end, which occur in the most deeply situated portions of the mesoglœa], which I have never seen in any preparation of A. digitatum.” “It is true that many of the elongated and spindle-shaped spicules of A. glomeratum are almost exactly the same shape as the spicules of the tentacles and disc of the polyps of A. digitatum, but the clubs are peculiar to it, the dumb-bell [not my developmental dumb-bell] absent, and the Ks. and crosses very rare.” Further evidence in support of this general conclusion that the nearer a spicule is situated to a limiting layer the more regular will be its form, is supplied by von Koch in his account of the spicules in Clavularia proliféra already referred to, where he gives figures showing that spicules situated near the surface are quite regular in form (stick-like), but that spicules situated in deeper layers of the mesoglœa become irregularly branched, but never massive and tree-like, as in A. digitatum, since the colony of Cl. prolifera is very racemose in form. Other Alcyonaria supply like evidence, as may be seen, e. g., by glancing through the plates in the ‘Challenger Report ‘on the Alcyonaria. It is also worthy of note that when as in Calcareous Sponges and certain Holothuria the body-wall is very thin, and the contained spicules therefore necessarily in close apposition to both external and internal epithelia, these are in general remarkable for their definiteness of form.

Illustrating Mr. Woodland’s paper, “Studies in Spicule Formation,” II.

Figs. 1—7 and 18—22 magnified about 950 diameters.

Figs. 8—15 magnified about 475 diameters.

Figs. 16 and 17 magnified about 450 diameters.

PLATE 16.

FIG. 1 shows scleroblasts containing spicules at their first appearance.

FIG. 2 shows young spicules, in which the pink-staining central organic core has become distinguishable from the peripheral layer.

FIG. 3.—The scleroblast nucleus has here divided, and the young spicule has in consequence assumed an elongated form. The nuclei, it will be noticed, tend to be situated at the opposite extremities of the sclerite.

FIG. 4.—The extremities of 1 he spicule are becoming divided up into processes.

FIG. 5 shows the typical “caudal vertebra” form of spicule.

FIG. 6.—A young spicule placed end-on. The concentric lamellar structure of the organic axis shows well.

FIG. 7.—An older stage of growth.

FIGS. 8—12.—Various forms of adult spicules.

PLATE 17.

FIGS. 13 and 14.—Various forms of adult spicules.

FIG. 15.—Monaxon unbranched spicules found in the vicinity of limiting layers.

FIGS. 16 and 17 show the mesoglœal cavities in which the spicules are situated.

FIG. 18.—An “endodermal cord” with several multinucleated cells.

FIG. 19.—Scleroblasts.

FIG. 20.—Interstitial cells.

FIG. 21.—Spherical “jelly-secreting” (Bourne) cells.

FIG. 22.—Nematocysts (“ovoid bodies”) from the mesoglcea situated some distance under the ectoderm.

FIG. 23.—Nematocysts from the surface of the colony showing cnidocils.