Studies in Spicule Formation.: I.—The Development and Structure of the Spicules in Sycons: with Remarks on the Conformation, Modes of Disposition and Evolution of Spicules in Calcareous Sponges generally

The present paper on Sycon spicules—forming the first of the series—was concluded long before I had decided to extend my researches to other groups, and in consequence the Theoretical Considerations forming Part II are to be received with caution. I retain these speculations because, crude as they may ultimately turn out to be, I am firmly convinced, judging from my later researches, that they contain a large element of truth.

Part I. Description of Spicule Formation in Sycons.

An account of the development of the spicules in Homor cœlous Calcarea (Ascons) having been published in January, 1898, ¹ a parallel inquiry into the conditions obtaining in the Heterocœla (Sycons) is desirable in order to establish the mode or modes of formation of these skeletal structures for the Calcarea as a whole. This is more especially necessary since Dr. O. Maas ² has recently attempted to describe the development of the Sycon spicules, and in so doing has unfortunately presented a very erroneous view of the facts. ³ For these reasons then, and at the suggestion of Prof. Minchin, I under-took to work out the histology of the Sycons, more particularly in reference to the spicules, in order to ascertain whether the very anomalous condition of things described by Maas had any actual existence or not. I say anomalous condition of things, since it was improbable à priori that such a fundamental character as spicule formation should materially differ in the two subdivisions of the Calcarea, and, as I shall show, cautious inquiry entirely disproved the supposition.

Before proceeding further, I should like to here acknowledge my great indebtedness to Prof. Minchin, who has so kindly afforded me his valuable assistance in the practical part of the work and supplied me with information and criticism in connection with the theory. With reference to this latter, however, I had better add that I alone am responsible for the speculations advanced in Part II of the present paper.

Maas’paper is based on the two species Sycon setosa and S. raphanus; the present account applies to Sycon coronata and S. ciliata, ¹ both common species on the English south coast. The Sycons examined by me were obtained from Plymouth, and were prepared as follows:—After detachment the specimens were quickly (to ensure full distension of the oscular rim) transferred to 1 per cent. Osmicacid, in which they remained for some few minutes, and then washed with several changes of distilled water; afterwards they were immersed in Ranvier’s (Weigert’s is as good) picro-carmine for three hours, again washed with changes of distilled water, and graded either into 60 per cent, glycerine for surface-view examination, or into absolute alcohol for section-cutting and surface-view examination.

These two species of Sycon are cylindrical in form, tapering at the two extremities—the basal or affixed, and the apical or free. From the base to near the apical extremity the body wall of the sponge possesses uniformly-distributed short diverticula or. chambers projecting from its sides, which diminish in size from below upwards, and are, as is well known, lined with the collar cells characteristic of sponges. Immediately above the region of these chambers is a short space also lined with the choanocytes, but distinguished from the rest of the body wall situated below by the absence of evaginations, so retaining the primitive asconoid condition. Succeeding the choanocytes again is a layer of flat epithelial cells which extends to the upper extremity of the sponge, where it is reflexed and continued as the exterior dermal coat. This internal layer of flat epithelial cells marks out the region of the oscular rim, the area which, in these sponges, is alone suitable for surface-view observations of the developing spicules. The substance of the body wall, bounded internally and externally by the gastral and dermal layers respectively, is gelatinous in consistency, and contains numerous free cells, the majority of which are the scleroblasts concerned in spicule formation.

In mounting the oscular rim for examination it is first severed off just below its origin, and the short cylinder thus obtained, after having been slit on one side, is laid out flat in glycerine (or balsam) on the surface of the slide, preferably with its gastral aspect uppermost, and covered.

Three primary forms of spicules occur—the simple monaxon and the compound triradiate and quadriradiate. Monaxons are present in greatest abundance at the apical extremity of the sponge, where they form the “brush “or circlet of long spines, but they also occur in the sponge wall generally, through which they protrude, and are especially noticeable at the extremities of the chambers where they form projecting tufts. In the region of the osculum, i. e. above the region of the diverticula, the monaxons all assume a more or less vertical disposition, but below this region the vertical disposition disappears with the development of the diverticula, and the arrangement becomes comparatively irregular.

The triradiate spicules are fairly uniformly distributed in the substance of the sponge wall (the three rays of each spicule lying in its curved plane, and therefore not protruding at the sides), and, like the monaxons, assume a regular and symmetrical disposition in the region of the oscular rim, i. e. above the region of the lateral chambers. This regular and symmetrical disposition of the triradiate spicule consists of one ray being vertical in position and situated next the base of the sponge, and the two companion rays in consequence (since the spicule is equiangular) lying towards the apex of the sponge at an inclination of 30° to the horizontal. The reason for this regular disposition of the triradiates will, as also in the case of the monaxons, be supplied later.

Triradiates situated in the “body” of the sponge are approximately both equiangular and equiradiate in form, but, corresponding to the unlike conditions to which they are subject, those situated in the extreme upper and lower regions of the sponge depart somewhat from this type. Triradiates, e. g. of adult Sycons situated in the oscular ¹ rim, have their paired rays the more depressed towards the horizontal the nearer they are situated to the upper extremity (text-fig. 1), and (to provide an explanation which will be more fully appreciated when Part II has been read) it seems probable that this depression is associated with the greater tendency of the wall to be invaginated in this region (see p. 265)—just as the upraised arms of a semaphore apparatus, joined horizontally by a spring, would incline the more to the horizontal the greater a weight borne by them. Generally speaking the “structural differentiation of the rays (in sagittal triradiates) is correlated with their position and function in the sponge” (Minchin ¹), and, as is doubtless the case elsewhere, the secondary forms just mentioned are, in all probability, determined in each individual instance by exposure of the formative (apical) cells, during their activity to the incident forces peculiar to the localisation of the spicule in the sponge —formative cells naturally being susceptible to all such influences. Indeed, the conformation of these apically-situated triradiates proves this susceptibility of the formative cell, for, at their central junction, the three rays are strictly equiangular, i. e. contain three equal angles—showing that the spicule would have assumed the equiangular triradiate form had it remained undisturbed,—and it is only the more distal portions of the three rays that become depressed towards the horizontal—this equally showing that some external cause must have exerted a disturbing influence during the later stages of growth of the spicule. It is true that in other sponges this depression of the paired rays in triradiates—the “alate “form—cannot always be attributed to a tendency to invagination of the sponge-wall, since mere contact with a lining membrane appears to be capable of producing the same effect. Nor is it a fact that in all sponges the triradiates situated at the edge of the osculum have depressed paired arms, since in many Ascons they protrude through the thin body-wall instead, so resembling the monaxons. But despite these exceptions I think it will be admitted (judging from the figure above provided) that the forms of the Sycon spicules (and these same modified forms occur in numerous other vertical and cylindrical sponges, both Ascons and Sycons) bear a relation to the incident forces I have briefly indicated, and, taken in conjunction with the general hypothesis to be elaborated below as to the cause of the symmetrical disposition of the spicules, little room for doubt as to this relationship can remain.

Another example of the modification of the ideal triradiate form due to environmental influence is supplied by the varying length of the vertical “posterior “rays of triradiates situated in different regions of the sponge. If text-fig. 1 above, e. g. be again examined, it will be observed that the posteriorrays of triradiates situated at the base of the sponge are longer than is consistent with the equiradiate type of spicule; and that, on the other hand, the posterior rays of triradiates situated at the extremity of the oscular rim are shorter. One possible explanation of this difference of length is provided, as before, by the fact that the sponge is constantly undergoing flexion (see p. 260). In the upper regions of the Sycon flexion attains its maximum, and hence there is, on this account, less room for elongation of the vertical posterior ray (as indicated in text-fig. 2), i. e. the posterior rays are here shorter for the same reason that the terminal segments of a crab’s claw are shorter than the more proximal segments, or caudal vertebrae than thoracic. Towards the base of the sponge, however, actual flexion is very small, and hence there is more room for elongation of the posterior rays; moreover, this elongation is here probably aided by the fact that the stresses due to flexion of the sponge are greatest in this region, and are, of course, borne entirely by the longitudinal element of the skeleton; in other words, it is quite possible that this mechanical stimulation exerts an influence on the formative cells, leading to their increased activity. Additional evidence of the foregoing explanation as to the cause of elongation of the posterior rays of basal triradiates is afforded in the case of other sponges. In the erect Leucosoleniidæ, and in those forms of Clathrinidæ “characterised by a more erect growth, such as Clathrina blanca and lacunosa, the posterior ray is indicated by its greater length, so that the triradiates become sagittal. In lacunosa this feature is carried to an extreme in the stalk, where a distinct peduncular skeleton is developed, composed partly of sagittal triradiates, partly of dinctinal monaxons,” Minchin). The triradiates in this last instance are of the “tuning-fork” type, and afford a good illustration of growth taking place in the direction of least resistance (in the ovoid body of lacunosa supported by the narrow stalk, on the other hand, the triradiates are regular). ¹

The quadriradiate spicules each consist of a triradiate basis, bearing on its gastral aspect an additional ray, which is apposed at the common junction of the three ray axes (or near it), and at right angles to them. These additional rays, so attached to a certain proportion of the triradiate spicules, are fairly uniformly distributed over the interior surface of the gastral cavity of the sponge, and project into this cavity with a slight upward inclination. It is not improbable that this upward inclination is directly due to the current of water ¹ which constantly flows up the gastral cavity and out of the oscular aperture. The gastral actinoblasts (formative cells) which deposit the substance of these gastral rays, as they are termed, must be influenced by such a current, and the inclination of the rays is perhaps the visible expression of this influence; at least there is no other assignable cause. The gastral rays probably possess no function, being, as will hereafter be explained, inevitable results of the architecture of the sponge wall.

The mode of formation of these three forms of spicule will be described first: theoretical problems being reserved for separate consideration in Part II.

As a preliminary to the following description, it is well to mention, especially in view of Maas’ misinterpretations, that it is sometimes necessary to use a certain amount of discrimination in describing the spicule and the cells in connection with it. Without due care, it is at first possible to err in distinguishing between the various adjacent cells—epithelial cells, collar-cells, pore-cells, wandering-cells, and isolated future spicule-cells (scleroblasts), not to name cells belonging to other spicules than the one under’ observation—and the formative cells of a particular monaxon or triradiate, but with a little practice such mistakes cannot be made. Criteria of those cells which alone appertain to an individual spicule are as follows:—(a) position of cell—the cell always being in dose proximity to the ray (especially to be observed in regard to the vertical, i. e. to the focus) and (in viewing triradiates from the gastral surface) with the nucleus lying in the plane of, or below, the spicule if basal in position (i. e. central), and above if apical (i. e. at end of ray); (&) form of cell—such, as will be seen from the figures provided, always having a well-marked relationship to the ray, as regards the long axis of the cell and the cell-contour; (c) character of cell—scleroblasts inter alia possessing greater definiteness of form than epithelial cells, a larger nucleus and less granulation than the choanocytes, difference of outline from porecells, and more granulation and absence of refringency as regards amœbocytes.

Unless otherwise stated, the following account refers to Sycon coronata; however, a like version holds true for S. ciliata, save in certain minor particulars which are duly considered in their place.

In both the species of Sycon examined the monaxons vary considerably in thickness, as the accompanying text-figure shows; nevertheless, the process of deposition is the same for all, though it is possible, and even probable, that the large thick monaxons have a slightly different origin from that of the thin variety (Appendix B). Judging from the few instances that I have observed, the first indication of the production of the future monaxon from an isolated scleroblast (the “mother-cell “of the spicule, Pl. 13, fig. 1) is the enlargement of the scleroblast nucleus (fig. 2). This enlargement is the precursor to division, and in this manner there is produced at the outset the two-celled, i. e. bi-nucleated, condition of the calcoplasm, which, in Sycons, remains constant throughout the entire growth of the spicule (fig. 3). In this bi-nucleated scleroblast the nuclei next separate from each other, and a concomitant of this is the appearance in the cytoplasm of a pale streak (fig. 4 -, not always easily seen in Sycons, though doubtless always present) stretching from nucleus to nucleus; it is in the interior of this mould that the spicule itself is first deposited as a minute refringent needle. As will be seen, the nuclei do not retain this initial position at the extremities of the young spicule, but soon come to lie on one side, and this fact seems to me to indicate that the nucleus is relatively unimportant in the actual secretion of lime, or at least of no immediate importance.

In fig. 3a two cells have apparently associated to otherwise produce the bi-nucleated condition of the calcoplasm, and if, as seems likely, a monaxon spicule similarly results from this somewhat differently-constituted bi-nucleated mass, this spicule should exhibit some difference in external appearance from one produced in the manner above described. In Appendix B I have supplied some evidence in favour of the supposition that the thicker variety of monaxons is derived from the association of two mother-cells (whose nuclei do not undergo division), and the thin from the division of the single mother-cell.

In all monaxons, even the thinnest, the two extremities are dissimilar, the proximal end, or end embedded in the sponge substance, tapering very gradually to the point, and the distal end, or end protruded through the sponge-wall to the exterior, being thicker, or, in other words, terminating more abruptly, and, in very young forms, resembling a barb (figs. 5 and 6). This latter feature is doubtless correlated with the prolonged adherence of the apical cell now to be described.

In all early stages of monaxons, the two cells associated with the spicule are situated at its extremities (figs. 5, 6, and 7), but as growth proceeds, the distal cell, after remaining stationary for some time, migrates towards the proximal end (figs. 8—11) until, in the adult structure, it replaces the proximal cell in position, this latter having previously deserted the spicule, after constructing the greater portion of it (fig. 12). The fact that the proximal cell does take considerably the larger share in the formation of the spicule is evidenced by the constant absence of the distal cell on the part of the spicule exposed to the outer world (at least two-thirds of the entire length), this evidently implying that during the protrusion of the spicule through the body-wall, the work of secretion in lengthening the monaxon has been solely performed by a proximal agency. The originally-distal cell, having replaced the proximal cell, adheres to the proximal extremity of the spicule for some time, and finally also deserts. The function of this distal cell is that of secreting the thick distal extremity of the monaxon before mentioned, and also possibly adding a secondary layer of lime to the body of the spicule in the course of its migration proximally. ¹

It may be added that longitudinal sections of Sycon ciliata (Pl. 15, figs. 45—49) show that monaxons may originate in the bi-nucleated cytoplasm before the mother-cell has separated from the epithelium whence it is derived; however, this is not by any means always the case, since many of the youngest monaxons are to be found embedded in the middle of the wall-substance, and apparently not in any way connected with either the dermal or gastral epithelium (figs. 48 and 49). Still, monaxons are found thus medianly situated, which still retain a connection with the gastral wall by means of fine protoplasmic processes, as shown in fig. 47.

Trios of more or less approximated spicule-cells (fig. 13) are to be found in every sponge-preparation, and these incipient congeries must be regarded as the first observable stages in the development of the triradiate (and quadriradiate) spicules. That the aggregate of three cells (fig. 14)—the “trefoil” (Minchin)—is formed by the association of three cells, and not by the continued division of one, is evidenced (apart from other considerations discussed below, which leave no room for doubt upon the subject) by the fact that aggregates of three cells showing two nuclei of smaller size than the third are never found, whereas aggregates of four cells containing two nuclei and aggregates of five cells containing four nuclei, of less diameter than their companions, occur pretty frequently (figs. 15, 16), small size of nuclei denoting recent cell-division, each cell of the trefoil subsequently dividing to form the “sextet” soon to be described. It is well to mention that nuclei only in the same individual sponge can be thus compared as regards size, since this character appears to vary slightly in different specimens, and certainly does in the two species, the nuclei of S. ciliata being much smaller than those of S. coronata. Three cells thus approximate to constitute the trefoil, and the more advanced the stage, the more closely associated are the cells. It is difficult to say exactly what this association consists of —whether, as seems probable, the cell-edges actually fuse, or whether they merely come into contact—but, however this may be, the three cells come together in much the same way as three billiard-balls would, and consequently possess the same triradiate formation thus conspicuous from the very beginning. Each of these constituent cells in the mature trefoil then divides centripetally and approximately radially in a direction more or less inclined towards the gastral epithelium, so forming the sextet, or six-cell stage, in the development of the triradiate spicule (Pl. 13, fig. 17, and Pl. 14, fig. 31).

It must here be pointed out that in the figures given depth of coloration of the nuclei is an artifice adopted to indicate relative elevation (high focus), or, in other words, proximity to the gastral surface from which the preparations are viewed (cf. nuclei of apical and basal cells) -, the depth of coloration actually observable in the preparations denotes, as remarked below, functional activity, and bears no relation whatever to the position of the nuclei. The device must be carefully distinguished from the fact.

The next advance from the sextet stage is the deposition in three centres of small calcareous elongated masses, needlelike, as in the case of the monaxons, and radially disposed, so as to include three equal angles at the centre of the cluster of six cells (fig. 18 for S. coronata, fig. 32 for S. ciliata). These three at-first-separate needles constitute the rudiments of the future compound spicule, into which they develop by the further activity of the sextet cells, two of these being devoted to each of the three rays. Whether this deposition of three needles is preceded in each case, as in monaxons, by the formation of a vacuole or mould in the substance of the cell (bi-nucleated) is not easy to determine, though its probable occurrence is evidenced by a like condition occasionally to be seen in more advanced stages. A. further point of some importance concerning this initial appearance of the triradiate system of deposits is the fact that each of these small needles is formed more in connection with the inner cell (i. e. the cell situated towards the gastral surface and which afterwards becomes the apical cell) than with the outer (cf. monaxons). Though both cells of each of the three pairs composing the sextet are essentially concerned in the production of the monaxon (it being probable, as pointed out below, that the monaxon or elongated form is in all cases determined by the presence of the two cells), yet this, in all probability, is for the greater part actually secreted by the future apical cell. The reasons for this supposition (for actual observation is difficult) are, as Minchin remarks, supplied both by the relations subsequently assumed by the two cells to the spicule-ray, and by the fact (which I have myself observed) that the nucleus of the inner cell stains more deeply at the initial stage of secretion than its companion—a reaction corresponding to functional activity. One more feature worthy of notice is the great preponderance as regards size of one of the three primary deposits in the sextet of S. ciliata (fig. 32)—a preponderance very slight and not at first apparent in the case of S. coronata. This initially-larger needle generally, if not always, becomes the large vertically-disposed or “posterior ” ray of the adult triradiate spicule, although the young triradiate may not originally be so placed as to render this particular ray “posterior” in position from the first. Why one pair of sextet cells should thus form a needle so much larger than its two companions is a question at present not easy to answer.

Each of the three separate primary needles being thus deposited in connection with a pair of the sextet cells situated approximately on a radial line, and these three needles collectively having a triradiate disposition corresponding to that of the original trefoil, the next step in the growth of the compound spicule is the junction of the three needles centrally and their individual thickening (figs. 19 and 33). About this time also the two cells connected with each of the three rays become more easily distinguishable into (1) a basal cell, which is outer in position, i. e. nearer the dermal surface, and remains until a late stage of development at the base of the ray, where it is wholly active, with its two companions, in firmly cementing, at their central junction, the proximal portions of the three constituent rays of the spicule, and (2) an apical, or inner cell, which advances in the direction of the ray at its extremity, and is, indeed, chiefly concerned in its construction. As the figures show, the apical cell (more deeply coloured for the purpose) lies well this side of the spicule ray, and therefore next the gastral epithelium, whereas the basal cell is situated pretty well in the plane of the spicule and in one of the interspaces contained by the rays. In figs. 19c and 19d, and some others, are shown spicules which are very hollow in appearance, but this feature is, as I subsequently ascertained from Prof. Minchin, due to corrosion by traces of acid contained in the glycerine in which the spicule preparations were mounted, and is therefore purely artificial in origin. ¹ The three apical cells, having built up the three rays of the spicule to their full length, leave the spicule for the surrounding mesoglœa; also, about the same time, the basal cells, having effected the junction of these three rays at their bases, begin to travel up their respective rays, so following in the course of the apical cell (figs. 26—30), and, like this, ultimately deserting the spicule. It is possible that during its migration towards the extremity of the ray, the basal cell may supply a thin secondary coating of lime, but this I have never observed. ¹ The destination of the apical and basal cells after leaving the spicule is unknown, and I have not attempted to ascertain it.

It is worth while remarking that after the apical cell has separated from the basal cell, the two cells apparently carry on their work quite independently of each other, and, as we have seen, their respective functions are, after the initial deposits have been formed, essentially distinct in nature. This subsequent independence of activity, or, otherwise speaking, lack of co-operation, is well shown by a curious type of spicule found by Minchin ²in Leucosolenia com-plicata which he has termed “derelict.” In this type of spicule, whilst the basal cells have been fully active, the apical cells seem to have shirked their duty, and the result is a large nodular mass with three small irregular rays arising from it—a conformation not only clearly exhibiting the independence, but also the respective natures of the activities of the two sets of cells.

As before stated, quadriradiate spicules are simply triradiates plus an additional ray, which is attached to the common junction (or near it) of the triradiate basis on its gastral aspect, and at right angles to the plane in which the triradiate lies—the plane of the sponge wall. The derivation of the mother-cell of this additional ray—the gastral ray—I have not been able to discover as yet, owing to the complexity of structure of the body-wall introduced by the presence of the diverticula, but certain facts tend to the conclusion that the origin of the gastral actinoblast is the same as that found in the Ascons. In this group, as Minchin has shown in detail, the mother-cell is produced by the division of a pore-cell in the vicinity of the future quadriradiate (when this is situated in the body of the sponge; when the triradiate is situated in the oscular rim, the mothercell is supplied by one of the unspecialised cells of the epithelium in that region), the resulting scleroblast settling over the point at which the gastral ray is to be developed, and, without further division, secreting a calcareous mass which adheres to, or near to, the centre of the triradiate basis, and gradually assumes the form of the adult ray. Evidence for the assumption that a similar state of things occurs in the Sycons is afforded by the general similarity of spicule formation found in the two subdivisions, and by the fact that pore-cells are generally situated in the neighbour-hood of gastral rays.

The youngest stages of develepment of the gastral rays which I have observed in sections of Sycon coronata are represented in figs. 35 and 36, in which the mother-cell must be distinguished from the basal cell or cells belonging to the triradiate basis. It must be pointed out that the figures given of the succeeding stages of gastral ray development represent the formative cell or cells as being more or less limited in respect to the area of the ray over which the cellsubstance extends, and such they appear to be in the ordinary picro-carmine preparations, ¹ but if the sections be immersed in Kernschwarz for ten minutes or so, the cell plasma (in which granules are scarce) becomes stained, and is seen to more or less entirely envelop the gastral ray whatever’ the stage of development may be (fig. 38). In studying the figures of the gastral rays given, this must he borne in mind. The gastral ray attains a considerable size before its mothercell divides into two (figs. 39 and 40)—at least two thirds of the adult length. Division occurs about midway up the length of the ray, and the two cells so produced apparently tend to separate. The further destiny of the two cells I have been unable to trace, probably because no further developments occur, the two cells persistently adhering to the ray throughout the life of the sponge.

As implied, never more than two cells are associated with the development of the gastral ray in S. coronata and S. ciliata, but, as in Ascons, it is probable, nay certain, that this limit of cell divisibility is merely a specific feature, not necessarily holding for other species and genera in which the spicule attains larger dimensions. In fact, a comparative study of calcareous spicules shows beyond doubt that the number of formative cells produced from the original mothercell (or mother-cells) is, other things equal, strictly correlated with the maximum size of the spicule attained (from the standpoint of cause and effect, the order should obviously be reversed), and hence in the case of exceptionally large structures the number of actinoblasts is above the average.

One feature in which the formation of the gastral ray differs from the type of development presented by the monaxons and triradiates is the apparent production of the ray in a single cell, the second not being formed until a very advanced stage of growth, and then bearing a very different relation to the spicule compared with that of the basal cell of triradiates or the distal of monaxons; that is, the second gastral actinoblast is concerned with the further growth of the ray, and not with its initial production. As pointed out below, it is probable à priori that one or more of the basal cells of the triradiate base co-operates with the division product of the pore-cell in order to supply the initial stimulus necessary, under normal conditions of growth, to the production of a monaxon structure.

If spicule preparations be stained with Kernschwarz for about ten minutes, the limits of the cell-substance become clearly defined, and the relationship of the form of the cell to the spicule is thereby rendered more evident than it is under ordinary conditions (the sheath of the spicule remaining though the calcareous matter becomes destroyed). Such preparations reveal the fact that the cells of monaxons, the apical cells of triradiates, and the cells of gastral rays form cylinders enveloping a portion (in the case of the two former) or the whole (in the case of the gastral ray) of the length of the ray on all sides (Pl. 15, fig. 50, shows monaxon cells thus treated; Pl. 14, fig. 38, shows a gastral actinoblast only slightly stained). On the other hand, the basal cell of triradiates (fig. 51) is not cylindrical, in form, but simply adheres as an elongated mass to one side of the ray. The reason for this difference of conformation is one I shall presently point out (p. 276); at present I may again remark that this non-cylindrical disposition of the triradiate basal cell is perhaps responsible for the non-secretion of a secondary layer of lime in the course of its migration up the ray (p. 247). In the case of the cylindrical distal cell of monaxons, a secondary layer of lime is secreted during migration, as also previously mentioned. This simple method of defining the cell-limits just described effectually disposes of the idea that in all cases the spicule is entirely enveloped by the distended cell-substance (e. g. see Maas’ figures). If this limited extension of the cell-substance at first seems inadequate to account for the relatively large mass of the spicule secreted by it (as e. g. in the case of the apical cell of triradiates), it must be remembered that the secreted substance is wholly derived from the surrounding medium, and that the cell, like the familiar kettle on the fire, is only the secreting agency and will, if allowed sufficient time, deposit any amount of lime required, though, like the kettle, it becomes worn out in the end. This truth also renders it more easy to understand the causes of variation in size of calcareous spicules, also to be referred to presently.

As regards the secondary migration of the basal cells of triradiates, and the distal cells of monaxons when their power of excretion is exhausted, there is nothing more remarkable in the phenomenon than in the re-assumption of locomotive powers by an amoeba or infusorian after feeding or being otherwise engaged, and the spicule ray evidently serves as a guiding path: the stimulus to movement in the latter case is possibly the same as that in the former. It may also be pointed out as a possibly significant fact that the gastral rays, which are alone directly immersed in the surrounding medium, are alone among spicule rays wholly enveloped by the cell-substanee; on calling to mind that the external portions of protruding monaxons never possess a cell on their surface, it seems possible that a connection exists between these two phenomena.

Part II. The Spicules of Calcareous Sponges in general; Theoretical Considerations.

Although owing to lack of information with regard to the chemical and physical aspects of lime secretion, it is as yet difficult to definitely account for many minor features of spicule formation, it is yet possible to indicate the main features of the process and such it is now desirable to do if we wish to attain to a true conception of the evolution of the calcareous skeleton of sponges.

The first general and obvious condition essential to the deposition of lime—the first law of spicule formation—is the proximity of the cell-substance to the area over which calcareous matter is being secreted. Illustrations of this condition have already been supplied in the above account of Sycon spicule formation, as e. g. by the respective positions of the basal and apical cells of triradiates corresponding to the thickened centre and elongated rays of the spicule, by the thickened distal end and secondary coating of monaxons corresponding respectively to the stationary condition and migration of the distal cell, etc., etc.

Other calcareous sponges afford like evidence. Thus the “derelicts” of Leucosolenia complicata before mentioned, the clubbed extremities of the triradiates of Clathrina clathrus (correlated, as Minchin shows, with the prolonged adherence of the apical cell), the gastral ray spikelets in Clathrina cerebrum corresponding to the fragmentationof the actinoblast nucleus, etc., etc., all illustrate the same law.

It must be observed, however, that although it is necesary for the surface which is receiving fresh deposits of lime to be covered by a layer of “calcoplasm,” yet the fact that the mass of protoplasm containing the nucleus is necessarily situated to one side of the growing ray does not affect the symmetry of deposition, as the figures of the Sycon monaxons show, and as is elsewhere abundantly illustrated. ¹ The process of spicule growth may, in fact, be compared in this particular with the building of a jetty by a multitude of labourers who, for a given reason, have moored a boat containing the provisions, timber, stores, etc., to one side. The one-sided disposition of the boat and stores relatively to the jetty evidently will not interfere with the bilaterally-symmetrical growth of the latter for the sole reason that the ship and stores do not constitute the building agency—are not engaged in the distribution of the added material, but are solely concerned with the nutrition of the building elements and the supply of, material for that which is built.

A second essential condition to the deposition of lime in any quantity in Calcarea (i. e. not taking into account the minute granules of lime often found in single scleroblasts) seems to be the co-operation of two dermally-derived cells, the deposition in every case (as may be inferred from the converse of the law just enunciated, viz. that where the bulk of the calcoplasm is situated there lime will be deposited) assuming an elongated form. Thus isolated monaxons are either formed as above described, on the occurrence of nuclear division in a single cell, i. e. on the separation of the substance of the cell into two distinct masses at opposite poles, or, as there is reason to believe (Appendix B), on the association of two cells—in either case two masses of cell-substance with their contained nuclei being distinguishable. As also already described, it is apparently necessary that each of the constituent cells of the trefoil should divide before the three monaxons composing the triradiate can be deposited. And like evidence is perhaps afforded by the divided-off small nuclei (i.e. cells) of the gastral rays of Clathrina cerebrum, each of which doubtless “co-operates “with the mothernucleus in order to produce a spikelet. At first sight the formation of the gastral ray appears to be an exception to the rule, but, seeing that there is no evidence to the contrary, it is legitimate to suppose that each gastral actinoblast “cooperates “with one or more of the basal cells of the triradiate system to produce the ray, and is thus conformable. That this is the case is evidenced by other considerations about to be discussed.

Assuming the truth of these two laws—the necessity of the proximity of the cell-substance to the site of lime secretion, and, in Calcarea, the necessity of the presence of two masses of dermally-derived cell-substance, between which the young spicule is deposited—it is possible to consistently explain the existence of the three kinds of spicules characteristic of calcareous sponges, showing not only why the three kinds of spicule occur, but also why other kinds do not.

Immigration of dermal cells into the median gelatinous substance of the sponge wall mostly occurs in those portions of the sponge which are in course of rapid growth, as, e. g., in the oscular rim, and in the diverticula of immature sponges. As already implied these isolated dermal cells or scleroblasts are incapable of depositing lime in appreciable quantity whilst in the uni-nucleated condition. In the majority of Clathrinidæ and some other sponges, and also in the very young stages of many sponges which possess monaxons at a later stage of development, the stimulus to lime secretion, whatever may be its nature, is not even called into existence when the sclero-blastic basis of the future spicule is bi-nucleated (and binucleated scleroblasts and two-cell associations are to be found), but in the majority of Leucosoleniidæ and Sycons monaxon spicules are produced either when the nucleus of a single scleroblast divides, or when two scleroblasts associate together, or on both occasions (see p. 273). As already indicated, the bi-nucleated, i. e. two-massed, scleroblastic basis necessarily produces a monaxon structure under such conditions owing to the elongation of the secreting layer of calco-plasm involved in the bi-polar redisposition of the mass of the cell-substance, i. e. the monaxon form is directly related to the conformation of the secreting agency.

When three scleroblasts associate together, the conditions as regards secretion are somewhat more complex. It is evident à priori that a monaxon cannot be formed between any two of the three cells, since the presence of the third (the potency of which is equal to that of either of the other two) must exercise a disturbing influence; in other words, three approximately equal secretory centres being present and grouped about a common centre, the deposition of calcareous matter must be symmetrical with regard to all. Why calcareous matter is not deposited at the centre of the trefoil, so fulfilling this last obligation, it is impossible as yet to say. Nor is it possible to supply a definite answer to the question as to why it is that a triangular system of three monaxons is not produced; it can only be pointed out that we possess no evidence that a nucleus can, in Calcarea, stimulate secretion in two places at once, and that the trefoil stage tends to show that the nucleus does not possess such a capacity. In’ actuality, deposition does not occur until each constituent cell of the trefoil has divided, and then the three monaxons produced inevitably tend, from the initial triradiate construction of the trefoil, to form a triradiate system. As in the case of the monaxon, the triradiate form is directly related to the conformation of the secreting agency. I am aware that this interpretation of the form of the triradiate is disputed, but until it is clearly shown how, e. g. surface tension can by itself “lead to the growth of three actines inclined at angles of 120° to each other” (Solias), or how pore-distribution can effect the same result when the pores are absent (as in sponge larvæ), I must adhere to the explanation I have provided. A simple explanation which, as will be seen, simultaneously explains the conformation of all three types of spicule—monaxon, triradiate, and quadriradiate—has, I think, something to recommend it. ¹

If the supposition hitherto made, viz. that the association of scleroblasts in twos and threes is largely, if not entirely, fortuitous, or, at most, only due to those influences which lead to conjugation and syzygy in Protozoa, be legitimate, then it follows that higher associations must also occur, and, if such be the case, there must be an explanation of the fact that four- and five-rayed spicules are rarely, if ever, met with. ¹ On mere grounds of probability these higher associations must be few in number, and in observations of the sponge-wall they would probably be passed over as stages in the formation of sextets, but I have no doubt that systematic search on a large scale would reveal their existence. ² Why these higher associations do not result in multi-rayed spicules can be explained as follows:—Suppose four cells to associate, then, as in the case of the trefoil, and for the same reasons, neither a central concretion nor a square of monaxons would be deposited. But, assuming that the four cells are not quite symmetrically placed about a common centre (a most improbable occurrence), there is no reason why the two nuclei most closely approximated should not, in virtue of their greater proximity, produce a monaxon, since, unlike what occurs in the case of the triradiate, the third nucleus is prevented from exercising a disturbing influence owing to the presence of a fourth nucleus, which is able to “saturate” its “affinity,” so to speak. In other words, granted the asymmetry of disposition, the four cells would pair off into two monaxon groups, and the potency, or-”affinity” of each cell would be satisfied. And similarly with an association of five scleroblasts, which, if it occurred, would probably resolve itself into a triradiate and a monaxon group. It will thus be seen that all these higher associations of scleroblasts differ from the trefoil in that in the former the “affinity” of each of the constituent cells can be immediately satisfied, whereas in the latter such is impossible, each of the three cells having to undergo division before secretion can occur. This hypothesis of the “saturation “of cell “affinities “thus not only readily explains the existence of the three kinds of spicules, but also shows reason for the sole existence of these three forms.

The production of a gastral ray on an already-formed triradiate basis is a phenomenon of a like order to the above. If in the oscular rim the central portion of a triradiate closely situated to the gastral layer comes into proximity to one of the unspecialised cells composing the actively-growing epithelium of that region, then, two dermally-derived cells (one in the epithelium and one of the basal cells in an interspace of the triradiate—unless all three of the latter co-operate) being brought into apposition, the conditions essential to the production of a monaxon structure are fulfilled, and a monaxon disposed at right angles to the plane of the triradiate will be produced. It is true that the basal cells of the triradiate are already engaged in lime deposition, and hence are not free to co-operate with the future gastral actinoblast in the same degree that an isolated scleroblast is able to, but, as is shown by the basal cells of all young triradiates, the mere presence of a cell is sufficient to stimulate another to active work (the one in the meantime remaining passive so far as secretion is concerned), and hence the basal cells of the triradiate can well fulfil this condition in the formation of a quadriradiate spicule.

But this last assumption naturally suggests a further question. If an unspecialised cell of the internal oscular epithelium is able, when brought into proximity with the basal cells of a triradiate spicule, to forcibly compel these to co-operate (forcibly, since a change or lapse of function is induced) and share in the production of a monaxon spicule, how is it that isolated scleroblasts do not by similar means produce adventitious rays on the opposite side of the triradiates ? Isolated scleroblasts must often come into the vicinity of triradiate actinoblasts, and hence, on the above assumption, might have been expected to initiate deposition under such circumstances. The answer to this question is afforded by the constancy with which the apposition of the two cells is maintained in the case of the gastral ray. This constancy of immediate apposition obviously results from the disposition of the triradiate with regard to the gastral wall in which the future gastral actino blast is situated, and is evidently not present in the case of an isolated moving scleroblast situated on the dermal side of the triradiate. In the former case persistent maintenance of the apposition forcibly induces the co-operation of the triradiate basal cells; in the latter case the conditions do not permit of such coercion, and in this distinction doubtless lies the explanation of the difference of results in the two cases.

This forced co-operation between dermal cells, one or other (or both) of which is previously engaged in another function, is still more notably illustrated by the induced division of a pore cell, situated in the body-region of the sponge (below the oscular rim), to provide the gastral actinoblast for a triradiate in its vicinity. Pore-cells are dermally-derived, and hence it happens that those pore-cells which happen to be situated in the neighbourhood of a triradiate are placed under conditions similar to those of the unspecialised cell of the oscular epithelium. The pore-cell itself being functionally specialised, and necessarily bearing a one-sided position with respect to the centre of the triradiate with its three basal cells (in its three angles), its division is a necessity for the end to be attained. Although it is difficult to conceive how the division of the pore-cell is forcibly induced by the proximity of the basal cells of the triradiate, yet that such is the case can hardly be doubted.

To sum up: a spherical isolated scleroblast gives rise to a spherical sclerite (especially well seen in Alcyonaria and Echinoderms); an elongated bi-polar (bi-nucleated) scleroblast, according to the same laws of spicule-formation, gives rise to a monaxon; similarly, a trio of cells ultimately produces a triradiate structure; in short, the form of the spicule is evidently related to the form assumed by the secreting calcoplasm. Further, as the foregoing statements also prove, maintained apposition of two dermally-derived cells is in Oalcarea essential to, and therefore the constant percursor of, the production of a monaxon spicule of appreciable mass, and since this maintained apposition of cells occurs, apart from the instances just supplied, at only one situation in the ordinary calcareous sponge, it is only rational to attribute the gastral ray which is there produced to this cause. Granting two simple and easily verifiable propositions respecting the rationale of spicule-formation, it is thus possible to enunciate a theory consistent with the facts, or, at least, such as are at present known.

Before considering the possible causes of the secondary forms assumed by spicules and other “features “of lime secretion, I will first discuss the disposition of the spicules in Calcarea, since the former will by this arrangement be more readily comprehended.

TO ensure due comprehension of the explanations about to be given in connection with the several modes of disposition which the elements of the sponge skeleton assume under different conditions of life, it will be necessary to first briefly consider the sponge organism in its relation to the environment, and to this end we may select as a convenient form of sponge either of the two species of Sycon, the spicules of which have been already described.

Sycons situated in shallow water which is often in motion, or planted upon rocks exposed to the action of falling waves, are in either case subjected to incident forces of considerable magnitude, and it will be readily understood that, were it not for the presence of the solid supporting structures contained within the sponge-wall, the organism could not attain to any considerable size, owing to the fragile nature of its semi-liquid gelatinous substance. Hence it follows that all the stresses to which the vertical sponge cylinder is subjected are borne by the contained spicules, and these inevitably react to the forces incident upon them.

The elongated hollow cylinder of which the Sycon ¹ consists can be affected in two ways by the motion of the surrounding medium; thus, being attached by the slender base, it can either (a) bend vertically as a whole (just as a tree is swayed by the wind), or (b) the wall of the cylinder can be invagi-nated upon itself, so tending to obliterate the gastral cavity (text-fig. 4). This latter reaction of the sponge is obviously but another phase of the former, since invagination of the wall is merely a flexion of one half of the sponge relative to the other; nevertheless, the two reactions must be distinguished, since they constitute two separate factors in the disposition of the spicules.

To ensure that neither of these reactions of the sponge shall become excessive, i. e. detrimental, it is necessary that means of support shall be developed, ¹ in order to preserve to some extent the vertical position of the sponge, and to maintain the appropriate distension of the gastral cavity. A support to protect the sponge-wall from undue vertical swaying is evidently furnished by a vertically-disposed skeleton, and likewise to maintain distension of the gastral cavity, there is needed a skeleton disposed in a horizontal manner, since flexion in either direction is resisted by skeletal material, the length of which is placed at right angles to the direction of stress in the plane in which flexion occurs. Hence the sponge skeleton must, under these conditions, be constituted of both vertical and horizontal elements. Both of these elements are supplied by the numerous triradiate spicules contained within the sponge-wall, for it inevitably follows from their conformation that if one ray be vertically disposed, then the two companion rays will lie in lines only deviating from the horizontal by an inclination of 30°, and hence the three rays practically constitute two axes, respectively lying in the required vertical and horizontal directions.

With regard to the two other forms of spicules—the monaxons and gastral rays—these probably do not in general exert a skeletal function, though the former lend considerable support to the oscular rim. The principal function of the monaxons is doubtless protective in nature, ¹ preserving the sponge from the attacks of other organisms by covering its surface with a multitude of sharp spear-heads. The gastral rays, as before mentioned, are doubtless functionless.

The triradiates can be affected in two ways by the pressures incident upon the sponge wall. If (a) the triradiates be situated in those portions of the wall which are in line with the direction of the incident force (in the plane of flexion of the sponge), then each individual spicule is acted upon by the incident pressure (transmitted through the gelatinous matrix) at right angles to the plane in which the three constituent rays lie, either from the side adjacent to the force or from the side opposite (text-fig. 5, A). If, on the other hand ()3), the triradiates be situated in those portions of the wall which are laterally placed with regard to the direction of the incident force, then each individual spicule is acted upon in the same plane as that in which it lies, i. e. laterally (B). As will shortly be shown—and as is, indeed, self-evident—pressures acting at right angles to the plane of the spicule have much more effect in determining the position of the spicule than pressures acting laterally in that plane, and, in consequence, when pressures are alternately incident upon the spicules in both these directions, only the effects of the former need be taken into account.

Recognising these facts, it is now possible to inquire as to the causes which have brought about the several modes of disposition of the spicules in Calcarea. And first we will consider the mode of disposition found in those sponges which are cylindrical in form and possessed of a thin wall, which are mobile about a fixed base, which possess an oscular aperture at the distal free extremity of the cylinder, and which are constantly flexed by the movements of the surrounding water (many Sycons, some Clathrinidæ and some Leucosoleniidæ). This disposition, which we may term the oscular disposition, is best observed in the Homocœla, owing to the absence of diverticula of the body-wall. In all sponges characterised as above, the triradiates are situated in the sponge-wall in such a manner that in each spicule one ray points towards the base of the sponge (in erect forms vertically downwards), whilst the two companion rays necessarily lie towards the apex (i. e. incline upwards in erect sponges at an angle of 30° to the horizontal), the whole spicule thus being symmetrically disposed with regard to the long axis of the sponge-body (p. 269, K).

An oscular aperture being present, it is only necessary to consider the forces incident upon the spicules at right angles to the plane in which they lie (text-fig. 5, A). It is obvious that, were it not for the presence of the transversely-disposed rays of the triradiates, the sponge-wall would be invaginated to a smaller or greater extent whenever the sponge was affected by motion of the surrounding water, since the vertical element of the skeleton (vertical rays of triradiates and vertical monaxons) is not adapted to resist flexion in this direction. Hence invagination of the thin wall is resisted by the two upper arms of the triradiates, i. e. the portion of the sponge-wall adjacent to the paired arms of each triradiate tends to “bulge” through the space subtended by them, when flexion of the sponge occurs (see text-fig. 6, F, below), and the triradiates being numerous and irregularly distributed (i. e. not arranged in vertical rows), invagination of the thin sponge-wall is almost entirely prevented. This resistance offered by the paired rays of each individual triradiate is, as already implied, the means whereby the symmetrical disposition of the spicule is brought about. For if we suppose that a triradiate is not symmetrically placed with regard to the long axis of the sponge-body (as in C or E), then it will be evident that on flexion of the sponge next occurring in the appropriate direction, the spicule will at once be “righted,” for the arm that is more inclined towards the vertical will be influenced by the pressure on the sponge wall sooner than the lower arm, and hence the spicule will be rotated about its centre until the two arms are similarly disposed with respect to the incident force (D). This process is represented in the above diagram. If a cylinder of paper be taken, and one upper side pushed inwardly, it can easily be understood that it would tend to “right “an asymmetrically-disposed forked structure, between the arms of which the invagination occurred (F). Even if the young triradiate be so initially placed that the vertical ray points apically (towards the osculum), such a symmetrical position would not be maintained, owing to the flexion of the sponge not always taking place in an exactly vertical plane (speaking of vertical sponges); and, moreover, if the weight of the spicule be a factor in its disposition, there is still more reason for the change from a relatively unstable to a relatively stable position, such as would obviously be the case here. ¹ This supposition of a downward pressure (greater probably in its effect at one period of the growth of the spicule than another, except in the case of the apically-situated spicules of adult sponges) being brought to bear on the paired arms of the triradiates found in these sponges is confirmed when we observe that the depression towards the horizontal of these paired arms is the more marked the nearer the spicule is situated towards the apex of the (adult) sponge where flexion is greatest, the pressure on the sponge-wall having determined throughout the whole period of its activity the direction of growth of the apical actinoblast (see above, p. 236).

There is a second mode of disposition of the triradiate spicules which is typically found in the blind (without an oscular aperture) free elongated diverticula of the genus Leucosolenia which at first sight appears anomalous and antagonistic to the explanations just provided for the case of the oscular arrangement of the spicules. This second mode of disposition, which we may term the non-oscular, was first pointed out by Minchin. In this arrangement the spicules are placed in an almost exactly opposite manner to that just described, i. e. the “vertical “or longitudinally-disposed ray tends to point towards the apex of the horizontal diverticulum, and the paired rays therefore tend to lie next the base. As before, the triradiates are more or less symmetrically disposed (far less so than in the oscular arrangement) with regard to the long axis of the sponge body, but their position is reversed [see text-fig. 8 (L) below, p. 269].

It is clear that, owing to the horizontal position of the diverticula, considerations as to the weight of the spicules being a possible factor in their disposition must be rejected. Again, owing to the absence of an osculum, invagination of the sponge-wall is also out of the question since there is no ready exit at hand from which the contained water may be expelled, the pores being too minute to allow of ready exit. The diverticulum, in fact, here resembles a water-cushion, and pressures tending to invaginate the wall are entirely resisted by the bulk of the contained water, and not by the paired arms of the triradiates. Since pressures incident on the spicules at right angles to the plane in which they lie are here non-existent, or at least ineffective as regards the production of the oscular mode of disposition—the body of the diverticulum being wholly uninvaginable—it is evident that the only pressures which can affect the triradiates are those which are incident laterally, for although uninvaginable the diverticulum is yet freely flexible about its base. The triradiates are affected by these pressures when they are situated more or less laterally with regard to the forces incident on the sponge (see text-fig. 5, B), and the triradiates are by them caused to assume the non-oscular mode of disposition by adopting, as before, a position of equilibrium with regard to these incident forces. Thus if a young spicule be initially disposed as in G, text-fig. 7, it will be evident that the tendency of lateral pressure (exerted on flexion of the sponge) from, say, the right side is to produce rotation of the spicule about its centre (the force impinging upon β long before it can reach λ) the arms β and λ rotating to the left and right respectively (H). If pressure be exerted on the left, rotation will occur in the opposite direction (I). Also, if the triradiate be initially disposed as in J, rotation will similarly take place.

The larger the spicule grows the greater tendency will there be for it to assume the position of equilibrium, and this position of equilibrium is evidently attained when the triradiate is in the position shown in either H or I, for when so placed, forces from neither side possess any tendency to produce rotation (the moments of the forces on and A about the centre of the triradiate being then equal). Whether the position of equilibrium be attained by rotation of the spicule to the left or to the right evidently depends as to whether a pressure sufficient to produce rotation of the spicule into the position of equilibrium first arrives from the right or the left, when the spicule has attained a sufficient size to be so influenced by the pressures on the sponge.

The spicule having attained the position of equilibrium it appears that the ray situated nearest the apex of the diverticulum (β and a in H and I) tends to lengthen somewhat, so giving the triradiate a sagittal appearance. The causes of this lengthening and the consequent tendency of the spicule to assume a more symmetrical position with regard to the long axis of the diverticulum are doubtless the same as those concerned in the lengthening of the posterior rays of the basal triradiates of Sycons (see p. 237 above), and the vertical disposition of monaxons mentioned below.

And now observe a striking confirmation of the above contention that the mode of disposition assumed by the triradiates depends, other things equal, on the presence or absence of an osculum. After the Leucosolenia diverticulum has attained a certain length, an osculum is formed, and, as a result of this, the young triradiates in the vicinity of the osculum immediately assume the oscular arrangement (see M in text-fig. 8 below). This fact seems to me clearly to prove that the disposition of the spicules is due to the direct action of the environment and not to inheritance.

From what has been said hitherto, it would logically follow that in sponges not vertically disposed and not flexible on a slender basis, no definite arrangement of the spicules would occur. It remains to be pointed out as strong confirmation of the above general theory that such is actually found to be the case—that in those sponges which are either not subject to the pressures resulting from motion of the surrounding water or whose conformation is not such as to cause the spicules to be influenced by these pressures (as e. g. the numerous non-erect encrusting forms of the Clatlirinidæ), there is no regular disposition of the spicules, and that the same is the case in the very young forms of those sponges which are erect in the adult condition, in which, before either the osculum or the sponge-wall is formed, the spicules are not only irregular in disposition, but also irregular in form, all of which facts (except the last, of course) might be anticipated on the above hypothesis.

The more or less vertical position of the monaxon spicules in Sycons and other erect Calcarea can be explained in a manner similar to that adopted in the case of the triradiates of terminally-closed Leucosolenia diverticula. In brief, with the exception of those few initially disposed in an exactly transverse direction (and such are found), all young monaxons more or less inclined to the vertical will tend to be righted by the lateral pressures to which they are subjected, as the following diagram suggests (text-fig. 9). In addition to this cause, invagination of the wall will also tend to cause the monaxons to assume a vertical disposition, since (with the exception again of those few initially disposed in an exactly transverse direction) it is only when they are so disposed that they will offer least resistance to the invagination of the wall; and all structures tend to place themselves in that position which enables them to offer least resistance to an incident force. Invagination of the sponge-wall, in fact, will act in the same manner as the lateral pressures above named. The protrusion of the monaxons on the sides of the sponge, or at the margin of the osculum or apex of a blind diverticulum, is an inevitable result of their elongated form, disposition and place of origin, and the thinness of the body-wall. Their protrusion at the sides of the sponge is, perhaps, also, in large part, due to the possible reflexing of the wall-substance at the margin of the osculum in elongation of the sponge cylinder during growth.

A few of the more conspicuous secondary features characterising the triradiate spicules found in the different groups of the Calcarea, and their possible causes must be briefly discussed.

Spicules which develop, under undisturbed conditions, in the homogeneous substance of a wall of narrow breadth, assume au approximately ideal triradiate form, i. e. equi-angular, equi-radiate, and with the rays perfectly straight and “finished.” Such are to be found in the majority of the non-vertical encrusting Clathrinid sponges. Spicules, on the other hand, which develop in the body-wall of the erect Leucosoleniidæ and Sycons, i.e. under disturbed conditions (since these sponges are constantly flexed to and fro by the motion of the surrounding water), are not so regular in form as those just instanced, the rays not being perfectly straight, and in many cases (above described for Sycons) deviating more or less from the equi-radiate type. In other words, the irregular “sagittal “form of spicule “is correlated with the more erect growth of” the Leucosoleniidæ and certain Clathrinidæ, and the finished regular “symmetrical “form of spicule in the encrusting Clathrinidæ “is doubtless correlated with the reticular form and growth of the sponges themselves” (Minchin).

Spicules, again, which develop in a mass of sponge-jelly, and which are therefore not in close proximity to two parallel surfaces, are also irregular in form; though, in these larval spicules, there necessarily exists an initial tendency to assume the triradiate form, yet, owing to the absence of the structures and hence forces which, in the adult sponge, ensure the symmetrical form of the spicule, this is not preserved in the ensuing growth. To speak in more detail, the thin unilaminar wall of such sponges as the Clatlirinidæ must be continually subject to pressures incident perpendicularly to its surface, and, granted the presence of a young spicule of regular triradiate form, this wall must tend to be invaginated in each of the three areas situated between the rays to an equal extent, which means that a groove is formed in line with each ray of the spicule along which the apical cell must tend to travel, as being the path of least resistance ¹ (see text-fig. 10 above). And, moreover, besides the constancy of direction of the transmitted pressures which affect the spicule of the adult sponge as contrasted with the indefiniteness of the conditions prevailing in the sponge larva, it must be remembered that the volume of the gelatinous matrix in which the spicule is embedded is considerably reduced in the adult sponge, and hence there is less space for the deviations from the typical triradiate form to occur in—the slight heterogeneities of constitution of the jelly cannot produce an appreciable effect. There is thus not only in the adult sponge a positive set of conditions tending to produce the regular triradiate structure, but also conditions which tend to negative asymmetry of form. ³

Again, spicules vary enormously as regards size, and these differences are as evident in the same sponge as in different individuals. The factors responsible for the size of a given type of spicule seem to be (a) the constitutional efficiency of the scleroblast; (b) the initial number of scleroblasts concerned in the production of the spicule; (c) the amount of fission which the actinoblast undergoes during the formation of the spicule; (d) the character of the region of the sponge in which the spicule is situated; and (e) the nature of the external environment, this partly depending upon the situation of the sponge in regard to its immediate surroundings, and partly upon the distribution of the species. Illustrations of (a) are to be found in every individual sponge, and in the different species and genera. As an illustration of (b) the difference between the two kinds of monaxons in Sycon coronata and ciliata already described may be instanced. Illustrations of (c) are to be found in the gastral rays of Clathrina cerebrum and contorta, in the former of which the gastral rays each possess two cells, and in the latter four cells, and in the huge monaxons of the majority of the Clathrinidæ, some of the more adult specimens of which possessing as many as five actinoblasts (Minchin); however, the precise development of these monaxons has yet to be determined. An illustration of (d) is to be found in such a genus as Heteropegma, in which the spicules situated in the cortical and medullary regions of the sponge differ largely as regards size, which difference in all probability solely results from the unlikeness of structural characters distinguishing these two regions of the sponge. ¹ In any given instance, the question as to how many of these factors are concerned, and in what proportion each has contributed to the result, can at present only be answered in a very general manner.

Another subject for consideration is the relative numbers of monaxons and triradiates present in different sponges. As yet I have not sufficient data to draw any general conclusions, and for the present I will only suggest a possible solution to the problem raised by the scarcity of monaxons found in many Clathrinidæ, as e. g. Cl. contorta, in which the monaxons are very large. As pointed out before, associations of two cells are more likely to occur than associations of three, and hence, merely on grounds of probability, the scarcity of monaxons in those sponges in which they occur at all is remarkable. Again, as previously remarked, sponges differ as regards the facility with which secretion occurs. In some sponges (as e. g. the Sycons above described) secretion occurs on division of a single scleroblast; in others, it is necessary that two cells should unite (as is apparently the case with Cl. contorta); and in others, no less than three cells must unite before secretion can take place (as in the majority of Clathrinidæ). Now, if we suppose that in Cl. contorta monaxons can be produced by union of two cells, but that the stimulus to secretion is very feeble, i. e. the union of two cells only just suffices to ultimately initiate secretion, then it is easily understood how in such a case it is that such a very few monaxons are produced—the majority of the associations of two cells becoming associations of three cells in the long interval which elapses before the binary association can produce a monaxon. Such is a possible explanation of this and other like phenomena; whetheir it is the true one I cannot undertake to say. The fact, however, that, so far as I have been able to discover, in no calcareous sponge do there co-exist small monaxons —monaxons presumably produced by division of a single scleroblast—with triradiates, but either small and large monaxons with triradiates, or solely large monaxons with triradiates, or triradiates alone, or monaxons alone, is suggestive and confirmatory of the idea just stated. Since, if a single binucleated scleroblast can produce a spicule, then the associations of two and three scleroblasts can; if secretion, however, only occurs on the associations of two and three scleroblasts, then only large monaxons and triradiates will result; if further, secretion only occurs on union of three scleroblasts, then only triradiates will be produced; if finally, secretion occurs very easily, then only monaxons will result. Why the two mother-scleroblasts which produce Clathrinid monaxons undergo division to such an extent as to give rise to such huge structures, it is impossible as yet to say.

One more explanation may be provided. It has been seen in Part I of this paper that the inner cell (next the gastral layer) of each of the three pairs of cells constituting the sextet is chiefly concerned in the formation of the ray, and that the outer cell of each such pair is basal in position, and with its two companions constructs the central portion of the triradiate system. Why should there be this particular division of labour ? The answer is doubtless largely to be found in the fact that the cells constituting the trefoil are from the first central in position, and closely adherent to each other, and therefore tend to maintain this disposition, whereas the three division-products of these are naturally more peripherally situated; are not adherent to each other, and hence are more adapted for centripetal migration. As to the question why the more peripherally situated divisionproducts of the trefoil should incline gastrally rather than dermally, it can only be pointed out that in all probability the sextet, like the rest of the organism, is influenced by the pressures on the sponge-wall. The sextet will, for example, always tend to be so placed in the sponge-wall as that the plane of the trefoil cells shall be parallel to that of the wall on account of the pressures transmitted through the jelly of the wall from both sides, ^{1 2} and seeing that the trefoil portion of the sextet constitutes the “body” of the cell-cluster, it is evident that all peripheral appendages (such as the future apical cells) will tend to swing to that side of it which is the less exposed to incident forces, i. e. gastrally—less exposed, since the pressures derived from the dermal surface of the sponge-wall exceed in intensity those derived from the gastral surface on account of a shield being provided by the opposite wall in the latter case. As the spicule becomes formed, the apical cell for the same reason will always incline gastrally at the extremity of the ray, the cell tending to rotate about its extremity (text-fig. 11), and this is probably the cause of the pyramidal conformation of the triradiate adapting it to the curvature of the sponge-wall. ³ Similarly the basal cells will also slide into the interspaces of the rays, in which position they are, as the figures show, normally found. In support of this conclusion may be named the fact that occasionally two cells are found in one interspace, both having slipped into the same retreat (figs. 23 and 24).

Finally, a word as to the relations between the apical and basal cells in the triradiate spicule. At first, as we have seen, each apical cell acts in conjunction with its respective basal cell in order to produce a monaxon structure, but eventually this co-operation ceases, and the basal cell commences to secrete spicular substance round the base of the ray in order to consolidate it centrally. Now the three rays once affixed centrally, the shifting of the basal cell from the base of the ray (and in young stages the basal cell is, as forming part of the trefoil, always more or less radial in position) into an adjacent interspace (which, being independent, it is at liberty to do) above noticed, is rather advantageous than otherwise, but, at the same time, it involves the loss of that cylindrical disposition of the cell-substance relative to the ray that commonly obtains; hence the fact cited above (p. 249), that the basal cells of triradiates alone do not possess this conformation—the cell in forming a deposit to one side of the ray loses connection with the other and does not regain its cylindrical disposition when secondarily migrating.

Many other problems remain to be solved, but the solution of these, as of those above discussed, can only be satisfactorily accomplished by means of a comprehensive survey of the vast array of facts presented by the structure and bionomics of the numerous genera and species of calcareous sponges known to exist.

The theory here adopted as to the origin of calcareous spicules has already been implied in the foregoing; in other words, the phylogenetic process is here held to be, in all essentials, identical with the ontogenetic process. It will tend to still further enforce the general argument if I here give a brief resume of the above discussion from the present point of view.

Granting the presence of numerous isolated scleroblasts in the substance of the sponge-wall (generally in regions of vigorous growth), the capacity of their cytoplasm to secrete lime, and the more or less fortuitous association of these scleroblasts in twos, threes, and higher aggregates, and, as shown above, it is not only possible to show why monaxons, triradiates, and quadriradiates have been produced, but also why, in the normal course of things, spicules of other forms have not. I have further attempted to show that the modes of disposition and secondary forms of the spicules are inevitable results of environmental influences operating during the course of each individual ontogeny. Thus, e. g. as regards the mode of disposition which I have termed “oscular,” we have initially in the case of each sponge an irregular arrangement of young spicules, which, during growth, becomes a regular and symmetrical arrangement, and since the rotation of solid structures implies a mechanical cause, and since this cause must be uniformly dispositioned with regard to the sponge as a whole (since the spicules are regularly distributed round the whole circumference of the sponge), there is, on this account alone, an à priori probability in favour of the conclusion that the motion of the surrounding medium so functions. In the case of localised secondary forms of the triradiates this probability is even stronger since it is inconceivable that each scleroblastic “determinant ”in the sponge ovum should be guided to its appropriate situation in the adult organism in order to there produce, or aid to produce, a spicule of a particular conformation; i. e. it seems impossible to attribute the production of these secondary forms of triradiates to inheritance. The direct action of the environment in determining the disposition of spicules is particularly well shown in the Leucosolenia diverticula mentioned above, where, owing to a change in the architecture of the sponge body, and in consequence to a different mode of action of the environmental forces, the disposition of the triradiates immediately becomes altered.

Prof. Minchin, to whom I am indebted for so many of my facts, has, in the paper before referred to, put forward an ingenuous hypothesis respecting the evolution of triradiate spicules, contending that such have arisen, through natural selection, by the apposition and fusion of primitively separate monaxons, their extremities being brought together by the agency of pore-distribution. And a somewhat similar idea has previously been expressed by Schulze, who, after asserting that “in the angle between any two rays (of the triradiates) a pore is situated “(which is only true to a certain extent for some Clathrinidæ—in all other calcareous sponges no such relationship between the pores and the triradiates being observable at any stage of growth), contends that the equiangularity of the triradiate is “to be explained in much, the same manner as in the honeycomb.”

I exceedingly regret having to differ from high authority on these matters, but, for the present, I must adhere to my own explanations, if only for the reason that they present to my mind more definite conceptions of the causes involved. I think it highly improbable that survival of the fittest can have had much to do either with the modes of disposition or with the primary or secondary forms which the spicules assume.

A summary of the statements made by Maas in his account of the development of Sycon spicules is here given (from ‘Zoological Record’ for 1901, by Minchin). The misleading nature of these statements will be rendered most evident by appending to them the necessary corrigents.

1. Each monaxon arises in a single mother-cell—which is not accurately true, since no trace of the spicule occurs until the mother-cell has constricted into two portions, each possessing a nucleus.

2. There is never more than one cell on the smaller spicules of this class (monaxon)—whereas, in actuality, there are never less than two, however small.

3. The large monaxons have numerous formative cells upon them, not derived from division of the mother-cell, but from the dermal layer de novo—which is demonstrably incorrect, the largest monaxons (in Sycon ciliata and S. coronata) never possessing more than two formative cells, and if, as in some Ascons, they possess four, they are, judging by analogy, probably all derived from the original mother-cell (or cells).

4. The triradiates also arise each in a single mother-cell as a concretion, but at a later stage they bear several formative cells—a statement which misrepresents the facts about as far as it is possible to do, the triradiates being derived (as Minchin has described in the case of the Ascons, and as is shown above in the case of the Sycons) from three mother-cells which have associated together, and built up by their six division products. Additional formative cells from other sources never occur.

I may add that I treated several specimens of Sycon ciliata with absolute alcohol and ammonium carmine—the principal method adopted by Maas; needless to say, the results were the same as those obtained by me in my other preparations. Indeed, results obtained by this method are somewhat better than those obtained with picro-carmine, the clear elongated “vacuole “or mould in the substance of the apical cell at the extremity of the spicule ray, e. g. being made more evident.

I find it very difficult on any supposition to account for the statements and figures of spicule formation supplied by Maas.

I have provided two additional figures (43 and 44) which I think bear out the conclusion, based on à priori grounds, that some monaxons in these Sycons—presumably the larger kind—originate from two mother-cells. If figs. 43 and 44 be compared with figs. 3 and 5, it will, I think, be admitted that there are here presented two distinct modes of origin of the monaxon—the bi-division of a single scleroblast and the apposition of two being well distinguished by the respective cell forms. It may be objected that, as in the triradiates and monaxons already described, the original mother-cell always divides, the non-division of the two mother-cells of the large monaxons would be anomalous. But the objection has little weight, for, since “the number of formative cells produced from the original mother-cell (or mother-cells) is strictly dependent on the maximum size of the spicule attained,” it is evident that, if the adult size of the spicule does not demand it, i. e. (more logically speaking) if there exists no stimulus to division, then the two mother-cells will not divide. In cases where division of the mother-cell does occur, there is always a good reason for it. Thus, in the case of monaxons produced from a single bi-nucleated scleroblast, and in the case of triradiates produced from the trefoil, the division of the one or more scleroblasts concerned is, under the conditions, a necessity for the production of the spicule. And in many Ascons, the huge monaxons possess four or more actinoblasts—there exists a stimulus (whatever may be its nature) causing division of the scleroblasts initially concerned in the production of the monaxon.

Again, it is obviously probable à priori (as mentioned in the text) that if three scleroblasts associate, two also should associate, and in this latter case there will, under the conditions, inevitably result a monaxon structure. But if monaxons can either be produced from a single binucleated scleroblast or from an association of two scleroblasts, then, since in the latter case the initial capital is twice as great as that in the former, a difference in the size of the two classes of monaxons might be anticipated. This is the case in the Sycons, as text-fig. 1 shows. Although, of course, there exists a certain range of variation in size in each type of spicule, yet the thin and the thick monaxons are tolerably distinct from each other.

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

All figures magnified 1000 diameters and drawn with camera lucida.

FIG. 1.—Two ordinary scleroblasts.

FIG. 2.—The same with the nucleus enlarged.

FIG. 3.—Division of the scleroblast nucleus; in a the binucleated condition has probably been produced by an association of two cells, and not by a division of one; in c the cell-substance is more elongated than usual.

FIG. 4 shows the pale streak or “mould” in the cytoplasm.

FIGS. 5—12 illustrate the further development of the monaxon spicule. Fig. 7, b, is slightly abnormal in its shape. The white streak is also seen in Figs. 11, b and c.

FIG. 13.—Groups of three scleroblasts in S. coronata probably about to associate to form the trefoil.

FIG. 14.—The trefoil stage.

FIG. 15.—Division of one of the trefoil cells—first stage in formation of sextet.

FIG. 16.—Division of two of the trefoil cells—second stage in formation of sextet.

FIG. 17.—The sextet stage, a presents the more typical appearance of this stage.

FIG. 18.—The first appearance of the young triradiate spicule’—the three granules or rods being quite separate.

FIG. 19.—The junction of the three needles accomplished; thickening of the individual needles has also occurred, c and d show the hollow appearance due to corrosion by acid.

FIGS. 20—24 illustrate the further development of the triradiate spicule in S. coronata. Figs. 21 and 22 are quadriradiates, the gastral ray possessing one actinoblast. In both Figs. 23 and 24 two of the basal cells have slipped into one interspace—an unusual occurrence.

FIGS. 25—30 continue to illustrate the further development of the triradiate spicule in S. coronata.. Fig. 25 is a quadriradiate.

FIGS. 31—34.—Young stages of triradiate spicule formation in S. ciliata. The superior size of one of the three needles is manifest from the first.

FIGS. 35—40 show stages of gastral ray (quadriradiate) development, as seen in longitudinal and transverse sections of the sponge-wall. In

Figs. 35—38 the gastral ray possesses only one actinoblast. In

Figs. 39 and 40 the gastral actinoblast has divided. In Fig. 38 the cytoplasm has been slightly stained with Kernschwarz to exhibit its full extent.

FIGS. 41 and 42 show well the secondary thickening of the monaxon spicule due to the migration proximally of the distal scleroblast. In Fig. 41 is again well seen the pale “mould “in the calcoplasm in line with the extremity of the spicule.

FIGS. 43 and 44 furnish evidence for the supposition that some monaxons arise from the association of two mother-cells. See also Fig. 3, a.

FIGS. 45—49 are figures of longitudinal sections of the sponge-wall. Fig. 45 shows the division of a scleroblast before it has immigrated from the gastral epithelium.

Figs. 46, 48, and 49 show monaxon spicules being produced under the same conditions. In Fig. 47 the monaxon actinoblasts still remain attached to the gastral epithelium by fine protoplasmic processes, though otherwise separated. In Figs. 48 and 49, on the other hand, are shown monaxons which show no connection with the gastral epithelium.

FIG. 50 shows scleroblasts of monaxons stained with Kernschwarz to exhibit their cylindrical conformation; only the spicule sheath is seen.

FIG. 51 shows the migrating basal cells of triradiates similarly stained to show their non-cylindrical conformation.