Discussion

Ectoderm: Although some anemones possess longitudinal columnar muscle (in the form of epitheliomuscular cells), these are considered to be "primitive" in this respect (Stephenson, 1928) in comparison to those forms in which the longitudinal body musculature is restricted to the mesenteries. Le Thi (1968, unpublished M.Sc. Hons. Thesis) described epitheliomuscular cells from the ectoderm of the column of I. olivacea, but page 11 in the present study these were not seen. Stephenson (1928) states that the muscle fibres are absent from the column ectoderm in the majority of forms.

According to Batham and Pantin (1951) the absence of ectodermal longitudinal muscles in the column in favour of the longitudinal endodermal parietal muscle is an advantage in large-disked forms since the muscles acting on the disc (the parietals and the retractors) are part of the same system. Also, they point out that grave mechanical difficulties would arise during contraction in diameter and length of the column if the longitudinal musculature was in the form of a continuous endodermal sheet, because the two muscle layers (circular and longitudinal) would be forced to buckle at right angles to each other.

The supporting cells of the ectoderm do not appear to be the same as the typical ciliated cells first figured by the Hertwigs (1879) and since used as the "typical" anemone supporting cell (Hyman, 1940, p.374; Stephenson, 1928). While the appearance of sections could certainly lead to the conclusion that the ectoderm is ciliated, direct examination of the living animal, experiments with charcoal particles, and dissociation experiments do not support this view. An electron microscope study of the body wall of I. olivacea is being undertaken at present, and should reveal the fine structure of the projections. However, the present conclusion must be that the ectoderm is not ciliated.

Pantin (1942) describes the occurrence of numerous conical, refractile projections scattered over the tentacle ectoderm of most actinians. The appearance of these projections as figured by Pantin is similar to the appearance of the ectodermal projections in I. olivacea. Pantin supports the findings of earlier workers that the projections are compound structures of fused cilia and that they are probably sensory structures. No reference is made to the occurrence of these projections on other body surfaces. In I. olivacea the projections are present on the adhesive disc, column, and oral disc as well as the tentacles, and because of this it would seem best at present not to consider the projections as of the same "type" as those described by Pantin. If the projections in I. olivacea are in fact sensory structures, it is difficult to see why they should be present on the adhesive disc. The further question then arises as to the significance of the projections seen at the surface in I. olivacea. One explanation is that the tongue-like projections assist the animal to shed unwanted mucous coverings or surface debris. The numerous glands present in the ectoderm indicate that a great deal of secretion takes place onto the surface, perhaps forming a temporary "cuticle" of hardened slime (Stephenson, 1928, p. 23). The ectodermal projections would probably act as "spacers" between this "cuticle" and the ectodermal cells (Pl. 1, Fig. 3). The secreted layer would thus not be firmly applied to the ectodermal cells, and perhaps could be shed more easily when required.

The results of this study suggest that the usual classification of actinian gland cells as either "mucous" and "ganular" (Hyman, 1940) or "mucous" and "albumen' (Stephenson, 1928) is useful only in the broad sense. In the present study the terms "ganular" and albumen" would be applicable to protein secreting cells, while "mucous" would probably refer to polysaccharide secreting cells. It would seem best to eliminate as far as possible solely morphological terms such as "granular" when describing gland cells because granules in one cell which show similar morphology to page 12 those of another cell may show different histochemical activity, indicating different composition and function. In I. olivacea for example, both protein staining and polysaccharide staining granular cells occur.

page 13

Histochemical results (Tables 1 and 2) indicate that 6 different types of secretion product can be recognised in the ectoderm of I. olivacea. The proteinaceous secretions (types "A" and "B") can be distinguished by the DMAB-nitrate test for trptophan, and by enhanced affinity for colloidal iron of the type "A" secretion after pepsin digestion. Also, the "B" secretion is pyroninophilic and some of this pyroninophilia is due to RNA. The other secretion types are strongly PAS positive. Type "B2" is distinguished by its negative reaction with the Mowry colloidal iron test. Types "C" and "E" are best distinguished by the criterion of reduced PAS positivity of the type "E" secretion after pepsin digestion. Type "D" is distinguished by the small size of the granules and by the fact that the secretion has PAS positive as well as Mowry positive components which remain separate within the cell. It is evident that the gland cells can be classified broadly as being either proteinaceous (types "A" and "B") or mucopolysaccharide containing (types "B2", "C", "D" and "E"). This latter group can be subdivided according to their reactions (see Table 3).

It is possible that the 6 "secretion types" (types "A"-"E") may represent fewer than 6 gland cell types, for some of the secretions may be earlier or later stages in the secretion cycle of one particular cell type. Secretion types "C" and "E" for instance, are similar histochemically and sometimes morphologically. However, until the secretion cycle of the cells in question are known, it would seem best to accept the results as indicating 6 different gland cell types, while at the same time realizing that the classification may have to be altered in the future.

The local modifications to the ectoderm which have been described have been called verruco-cinclides (Le Thi, 1968 unpublished M.Sc. Hons. Thesis). They have the appearance of imperforate cinclides (Stephenson, 1928) in that each is an ectodermal concavity with only a thin layer of mesogloea beneath it. Stephenson states that these structures are usually formed by greater development of one cellular layer than of the other. In I. olivacea they would appear to have been formed by proliferation of ectodermal supporting cells together with a reduction in thickness of the mesogloea beneath the cells. According to Stephenson cinclides probably function as "safety valves", and always are associated with water currents in some way. The imperforate cinclides can rupture neatly when needed, allowing fine jets of water to escape from the coelenteron during rapid contraction of the animal.

Mesogloea: The few tests applied in this study indicate that the mesogloea has certain similarities with the fibrous connective tissue of higher animals. First, both cells and fibres are present in a matrix; second, the fibres show staining reactions characteristic of collagen. Indeed Hyman (1940, p.281) refers to this type of mesogloea as a fibrous connective tissue often containing layers of fibres coursing in different directions among which are scattered amoebocytes and connective tissue cells.

Chapman (1953) attaches some importance to the supporting tissues of coelenterates because of the possibility that they could shed some light on the origin and composition of connective tissues in general. He presents evidence to show that the bulk of the connective tissue protein material of coelenterates examined by him conforms in histological page 14 appearance and physical and chemical properties to the collagen of vertebrates. The histological examination of an anemone Calliactis parasitica by Chapman yielded results very similar to those obtained for I. olivacea in the present study. In both cases the thick fibrous layer which constitutes the mesogloea stains with aniline blue in Azan and Mallory techniques, does not stain for elastin, does not stain metachromatically with toluidine blue (without sulphation) and contains no fat. I. olivacea mesogloea fibres gave a positive result for reticulin in sections, but it is not stated whether Chapman attempted to test for reticulin in Calliactis. The fibres of I. olivacea stain moderately strongly in the PAS test, indicating the presence of mucopolysaccharides. This result is frequently obtained with connective tissues from coelenterates (Chapman, 1966). It is shown by Chapman (1953) that in Calliactis the mesogloea fibres are arranged in sheets parallel to the surface and to fibres at 45° to the long axis of the animal. It is possible that fibres in the mesogloea of I. olivacea approximate this condition more closely than has been observed in this study. Most specimens used were not able to be fixed in as extended a state as is desirable for such mesogloeal studies. Chapman also shows, by the use of tangential sections, that in the inner and outer mesogloeal layers the fibres form a lattice structure similar to the pattern seen in a woven fabric. The suggestion is made that the fibres owe their particular orientation to the forces of muscles and environment which are exerted throughout the life of the anemone and not to any organized method of secretion by individual cells. Baitsell (1925) thought it possible that the orientation of developing fibres in chick connective tissue is due to stresses set up in the tissue by migrating mesodermal cells or other forces.

From a consideration of the action of muscles, and of the forces acting on the mesogloea during muscular contraction, Batham and Pantin (1951) argued that the part of the mesogloea in immediate contact with the muscle layer must have different properties to the rest of the mesogloea. They were proved correct by Grimstone et al (1958), who showed that the muscles are, in fact, attached to a specialized layer of mesogloea which they designated a basement membrane, and which is composed of amorphous material. The present study indicates a similar specialized region of mesogloea immediately beneath the ectoderm. It seems reasonable to expect that here also the electron microscope will show the presence of a basement membrane composed of similar amorphous material quite separate from the fibres of the bulk of the mesogloea. The ectoderm is, after all, an epithelium (as was the epitheliomuscular layer studied by Batham and Pantin, 1951; and Grimstone et al, 1958), and epithelia characteristically are attached to underlying connective tissues by a more or less typical basement membrane (Fawcett, 1966).

The origin of the fibres of the mesogloea is unknown (Hyman, 1940). In vertebrates it seems clear that the fibres of connective tissue arise in the ground substance, outside the cell (Baitsell, 1925; Wolbach, 1933). It has been shown that fibrous chemical compounds can be produced in vitro by mixing solutions of proteins with solutions of hexosamine sulphonic acids (Meyer, Palmer, and Smyth, 1937), and as Chapman suggests it is conceivable that the conversion of a homogenous protein-aceous matrix could be brought about by the secretion into it of a hexosamine sulphonic carbohydrate or similar carbohydrate by the cells page 15 of the tissue (Chapman, 1953). In I. olivacea the mesogloeal cells are variable in shape, some being elongate and narrow, and others being rounded. They appear to be of the one type, but cannot be classified as amoeboid since they were not observed to move. Hyman (1940) does not state by what criterion the mesogloeal cells are classified into "amoeboid" and "connective tissue" types. These cells in I. olivacea do not give the appearance of secretory cells, although some contain PAS positive and naphthol yellow S staining granules, and they have been shown to contain RNA (Table 2; Pl. 2, Fig. 3, arrows). But the most characteristic component of these cells would seem to be lipid, which is present as small globules throughout the cytoplasm of most of them. Moreover, in some actinians, e.g. Edwardsia callimorpha, the mesogloea although fibrous is devoid of cells (Chapman, 1953) so that the presence of cells in the mesogloeal ground matrix cannot be essential for the secretion of the fibres. In those species containing both cells and fibres in the mesogloea Chapman found the cells to be in no way connected to the fibres, but Batham (1960) has published an electron micrograph showing an amoebocyte of the anemone Metridium canum in intimate contact with a banded mesogloeal fibre.

It has been suggested (Pantin, private communication to G. Chapman cited in Chapman, 1953) that the cells of the mesogloea are present in species which can absorb the mesogloea during starvation, and absent in those which cannot. This thought is restated by Robson (1957) who suggests that the passage of materials such as dissolved food and excretory products between endoderm and ectoderm, for example, and the reversible changes in the mesogloea which accompany growth, or regression during starvation, could perhaps be mediated by enzymes from these cells. She sees it possible that the mesogloeal cells form a physiological system throughout the body of the sea anemone, functioning in a continuous transport medium supplied by mesogloeal and subepithelial fluid.

The staining of "granules" in the mesogloeal cells with reduced methylene blue is interesting in the light of the above comments. Unfortunately it cannot be decided whether the staining is due to an affinity of pre-existing structures in the cells (such as granules, vacuoles, or cell organelles) for methylene blue, or whether it is due to other causes such as active engulfment of dye particles by mesogloeal cells. The former view seems probable as the property of "vital dyes" (especially netural red and methylene blue) in colouring cell valuoles and cisternae is well known (Baker, 1958). It is likely that the blue stained "granules" which appear to lie free between mesologeal fibres are in fact present in fine cytoplasmic extensions of mesogloeal cells. Although the origin of the mesogloea is not clear, there is little doubt as to its fundamental function. It is the base to which muscles are attached, and together with the muscles controls the deformation of the body wall (Batham and Pantin, 1951) by virtue of its visco-elastic properties. In addition Chapman (1966) states that it is tempting to look upon the mesogloea as a reserve of material on which the animal can draw in times of starvation.

Endoderm: The epitheliomuscular cells of cnidarian coelenterates are considered characteristic of this group. They were first described by Kleinenberg (1872, cited in Robson, 1957) in Hydra and have since been page 16 found in other orders. As early as 1879 the Hertwig brothers showed that the muscle fibres in the endoderm of actinians were part of the endodermal cells themselves. The form of epitheliomusclar cells varies between the different orders in such characters as the number of myonemes present and the shape of the epithelial part of the cell, and their distribution throughout the body layers also varies. In Hydra for example, they are present in ectoderm and endoderm. The endoderm cells are phagocytic, and the ectodermal cells each have several muscle fibres and lack flagella (Goodrich, 1942). In I. olivacea, however, these cells are absent from ectoderm.

The endoderm of the column of I. olivacea is very similar to that from the mesentery of Metridium senile described by Robson (1957). In both cases the epithelial part of the cell is connected by a fine protoplasmic "foot" or "extension" to the muscle fibre part which lies on the mesogloea and forms (along with other muscle fibres) a muscle field (Batham and Pantin, 1951). As Robson points out, because the epithelial part of these cells is much wider than the part connecting with the muscle fibre, the connecting "stems' of these epitheliomuscular cells must traverse a cavity. In life this cavity, or space, would be filled with fluid, which thus would form a continuous thin layer between the epithelium and the muscle field, and probably would function in several ways. First, it would act as a hydrostatic layer when the muscle contracts, forcing the epithelial cell parts to become much longer and more slender (see section on endoderm under Results). The epithelium thus follows closely any contraction of the muscle field. When the muscle fibres relax the epithelium returns smoothly to its original height. Relaxation is nearly always slower than contraction, and as Robson notes, is probably affected by viscosity of the protoplasm and mesogloea rather than by pressure changes in the subepithelial fluid. Second, Robson suggests that the subepithelial fluid is part of the general internal medium of sea anemone tissues, and must function to integrate local biochemical processes because diffusion could proceed readily in such a fluid layer. Later work (Batham, Pantin and Robson, 1960) has shown that the fluid provides the immediate environment for the neurites of nerve cells, as well as for all epithelial elements such as mucus cells and nematocysts (Robson, 1957).

Robson (1957) looks to the Turbellaria, Annelida, and vertebrates in search of an epithelial muscular structure which could show a functional parallel to the epitheliomuscular system of Metridium. She concludes that while true musculoepithelium is almost entirely confined to the cnidarian coelenterates, analogous tissues may occur in higher animals wherever a flexible epithelium covers unstriated muscle. Two examples can now be given from the Turbellaria which would seem to approach the situation described in Metridium by Robson, and in I. olivacea, more fully than the Turbellarian example used by Robson. These are the polyclad Polycelis nigra studied by Skaer (1961) and the triclad Palombiella stephensoni studied by Wineera (1971). In both of these flatworms no true epitheliomuscular cells are present, but sections of the body wall show that fluid filled spaces are present between the basal region of the cells of the ciliated epidermis. Skaer remarks that in Polycelis these spaces seem to form a ramifying system above the basement membrane. In both animals these spaces probably serve a hydrostatic function, in view of page 17 the earlier study by Robson (1957), and taking into account the remarkable plasticity of the epidermis in these turbellarians. The morphology of these flatworms differs from that in Metridium and I. olivacea in that the muscles are not part of epithelial cells but are separated from them by a thick "basement membrane". However, a functional parallel does seem evident. Whether these fluid filled spaces in the flatworms also function to integrate local biochemical processes, as they may do in Metridium is not known, and elucidation of this problem will have to await more extensive studies on these animals.

The Hertwigs (1879) recognised that the condition as seen in the epitheliomuscular cells of actinians could be modified by stages in which the muscle fibre was developed at the expense of the rest of the cell. This is seen in the sphincter muscles of many actinians, and in, for example, the ectodermal tentacular muscles of I. olivacea (Pl. 2, Fig. 5), where the cell body is reduced to a small mass of cytoplasm surrounding the nucleus, and is connected by a narrow bridge of cytoplasm to the well developed muscle fibre. In these cases the cell body no longer performs an epithelial, or covering, function. The morphological similarity of a muscle cell taken from the tentacle of I. olivacea to that from the body wall of Palombiella stephensoni is striking (see Plate 2, Wineera, 1971). It is proposed to comment on this similarity at a later date.