Silver staining methods have been applied to the nervous system of Velella, Two histologically distinct plexuses are described under the headings ‘open’ and ‘closed’ systems. The open system consists of neurones with fine processes which run for distances of up to 2 mm, retaining their independence ins pite of frequent contacts with other fibres. The fibres of the closed system are large and run together, forming a nerve-net in which neurofibrillar material from different neurones intermingles; it is provisionally to be regarded as a syncytium. A certain type of ‘fibre’ in this system is believed to arise secondarily by the drawing out of adhesion connexions into long strands. Free nerve-endings resembling growing-points occur in both systems. The two systems occur throughout the ectoderm, but in the invaginated ectoderm the open system is poorly developed. The functions of the two systems are not known, but the closed system is probably specialized for through-conduction.

Neuro-sensory cells occur in the external ectoderm, making contact with fibres of both open and closed systems. No specialized endings have been found in a muscular region examined. No nerve-rings or centres have been found. Nerves are sparsely distributed in the endoderm, but they lie independently of one another and of ectodermal nerve-fibres crossing the mesogloea between the invaginated and external ectoderm layers.

The nervous system of Velella was discovered by Chun (1881,1882). Conn and Beyer (1883) were the first to investigate the nerves of the related form Porpita. Of subsequent work, reviewed by Chun (1902), Schneider’s investigation (1892) is of particular importance. As Chun observes, ‘Velella provides brilliant demonstration material for study of the coelenterate nervous system’. In view of this, and since no further work has been done on Velella during the last 60 years, a reinvestigation with modern staining methods was undertaken, the results of which form the subject of this paper.

The confusion surrounding siphonophore morphology and taxonomic relations has largely been dispelled during the last thirty years. It is clear from the studies of Leloup (1929) and Garstang (1946) that the Chondrophora (Porpita, Velella, and Porpema} are not medusoid organisms in any sense but are really highly modified, free-living hydranths that can be homologized with sessile tubulariid hydroids such as Corymorpha. I have discussed this at greater length elsewhere (Mackie, 1959) and have endeavoured to trace behavioural as well as morphological homologies between Porpita and Corymorpha. Fig. 1 shows the correct relationships. The chondrophores are distinct from, or only very distantly related to, the true siphonophores. In the most authoritative recent account, Totton (1954) ranks the group as a separate order from the Siphonophora.

FIG. 1.

Diagrammatic sections comparing a hydroid polyp such as Corymorpha (A) with the young and adult stages of Velella (B and c). Broken lines connect the corresponding parts of A with those of B or c. Continuous lines connect the parts of B with what they become in c. Redrawn with modifications from Hardy (1956).

FIG. 1.

Diagrammatic sections comparing a hydroid polyp such as Corymorpha (A) with the young and adult stages of Velella (B and c). Broken lines connect the corresponding parts of A with those of B or c. Continuous lines connect the parts of B with what they become in c. Redrawn with modifications from Hardy (1956).

According to Totton (1954) there is probably only one species of Velella, the cosmopolitan V. velella L. The specimens of Velella used in this investigation were obtained from Naples, Villefranche-sur-Mer, Arrecife (Canary Islands), Smitwinkel’s Bay (Cape Province, Union of South Africa), and Long Beach, Vancouver Island. The bulk of the material used came from the South African collection, which was kindly sent me by Dr. N. A. H. Millard. The specimens chosen for examination were about an inch long. Velella grows to over 212 in. in length. The fixative used was Bouin made up in sea-water.

While the general distribution of nerves in a coelenterate can be made out with the aid of simple dyes or by use of the phase-contrast microscope, investigation of their fine structure requires special techniques. Of these, methylene-blue staining and osmic-acetic maceration have been most used in the past. Metallic impregnations have been less employed. However, silver methods are now available which give repeatable and consistent results, and have the further advantage over methylene-blue and maceration techniques that they start with properly fixed material and end with permanent preparations. Methylene blue rarely stains all nerve-elements at the same time and is hard to make permanent. The results of maceration methods are often open to doubt because of the possibility that degenerative changes have occurred in the cells. Even among specialists of the osmic-acetic technique there is some lack of agreement. Krasinska (1914), for instance, casts doubt on the value of Schneider’s results (1892), saying that he overmacerated or macerated badly. Schneider’s procedure is, however, still followed (Semal-Van Gansen, 1952). While it is true that methylene blue and osmic-acetic maceration are of value for special purposes, it is becoming clear that silver methods are generally to be preferred.

The procedures used in this investigation were based on Holmes’s buffered impregnating solution. This was followed by hydroquinone/sulphite reduction and gold toning as in Holmes’s original method (1947) or by development in Peters’s physical developer, which contains glycin (Peters, 1955). Both paraffin sections and thin strips of tissue treated whole were employed. The latter proved most useful although they were harder to stain evenly. A pH of 8·0 was found to be suitable for the impregnating solution. Experiments with temperature, dilution, pH, and concentration of silver ions in the physical developer were crucial in obtaining good results. Satisfactory development was obtained with a modified Peters solution made as follows:

sodium citrate, 0·1 M, 10 ml

Peters’s developer, stock solution, 10 ml

distilled water, 25 ml.

To the above add 0·1 M citric acid until a pH of 6·3 is reached. To 45 ml of the mixture add 12 drops (0·7 ml) of 1% AgNO3. Develop at 15° C. The physical developer gave a stronger, more contrasty picture of the nerves than did hydroquinone/sulphite reduction, but it was hard to get really finegrain development by this means, although many attempts were made with different mixtures, temperatures, &c.

Interpretation of silver preparations

It is true, as Pantin (1952) states, that structures other than nerves are often picked out by silver methods. Among the structures stained most sharply are flagella, discharged nematocyst filaments, centrioles, the membranes of nuclei, nucleoli, and cells, spindle fibres, and fusomal structures (mitotic spindle relics). The latter are frequently found connecting young cnidoblasts, as I have described for Physalia (Totton and Mackie, 1960), and it is possible that some authors may have mistaken them for nerves. On the whole, however, if the preparation is a good one and the pH of the impregnating solution has been correctly adjusted, there is never any risk of confusing these structures with nerves. Muscle-fibres usually stain slightly at pH 8·0, but they are unlikely to cause confusion even in the worst preparations. Filamentous, branching processes running from the endoderm into the mesogloea are common in Velella, as in many Hydrozoa, but their general topographical relations distinguish them from nerves. In hydrozoan medusae, strongly argentophil branching fibres occur in the mesogloea and these might cause confusion were it not for the fact that they lack nuclei. The mesogloea in Velella is completely unstained by silver methods. Dr. K. M. Rudall has shown that it gives the X-ray diffraction pattern of a collagen.

It is probably true that there is little risk of mis-identification of structures in silver preparations provided that the investigator is aware that other structures may stain and further attends to the pH of impregnation. Peters (1955) gives an interesting table showing how depth of staining in nerve-fibres, nuclei, cytoplasm, and connective tissue may vary with the pH.

Distribution of the nerves

Nerves are found in the ectoderm and mesogloea and around the endoderm canals. The nerves crossing the mesogloea run between the external ectoderm and the internal, invaginated ectoderm (fig. 1) which secretes the chitinous float material. In crossing the mesogloea they do not associate with the endoderm canals as far as can be seen but pass between them without making contact. The nerves described here as endodermal are very sparsely distributed and lie by themselves, without connexions with one another or with the ectodermal fibres crossing the mesogloea. The few examples seen have all been closely wound around the canals, and are thus distinguished from the fibres which cross the mesogloea uniting the ectodermal plexuses. The endoderm of the organs of ingestion has not been examined. Fig. 4, A shows the cell-body (cb) of a neurone lying in the mesogloea. Several nerve-fibres (nf) are shown. Filamentous endodermal processes (fep) are also seen. Flagella (f) project into the lumen of the canals (c). Symbiotic algae (zx) are found in many parts of the endoderm.

FIG. 4.

(plate), A, nerves crossing the mesogloea (explanation in text, p. 122). B, free ending of a closed-system fibre in the mesogloea. C, free ending of an open-system fibre. D, portion of the nervous system in the mantle ectoderm. Compare fig. 3. E, closed-system fibre in mesogloea, showing fibrils clumped. F, the same fibre as in E in another region, showing fibrils spread out. G, adhesion bridge in the closed system. H, portion of the nervous system from the invaginated ectoderm showing a fine fibre of the open system (o) and some freely ending branches of the closed system. J, multiple interconnexion in the closed system. K, teased ectoderm, showing epithelial cells and a damaged neuro-sensory cell. L, cell-body of neurone in the closed system. M, interconnexion in the closed system. N, free nerve-ending from the invaginated ectoderm. O, double junction between two fibres of the closed system.

FIG. 4.

(plate), A, nerves crossing the mesogloea (explanation in text, p. 122). B, free ending of a closed-system fibre in the mesogloea. C, free ending of an open-system fibre. D, portion of the nervous system in the mantle ectoderm. Compare fig. 3. E, closed-system fibre in mesogloea, showing fibrils clumped. F, the same fibre as in E in another region, showing fibrils spread out. G, adhesion bridge in the closed system. H, portion of the nervous system from the invaginated ectoderm showing a fine fibre of the open system (o) and some freely ending branches of the closed system. J, multiple interconnexion in the closed system. K, teased ectoderm, showing epithelial cells and a damaged neuro-sensory cell. L, cell-body of neurone in the closed system. M, interconnexion in the closed system. N, free nerve-ending from the invaginated ectoderm. O, double junction between two fibres of the closed system.

Chun (1881) states that there are no organized tracts or centres in the nervous system. This appears to be true. Later, however, (1902) he refers to a special region of the plexus in the mantle which he believes to be homologous with one of the marginal nerve-rings of medusae. I have examined this region carefully and find that the plexus is well developed but that it is not organized into anything comparable to the tight bundle of fibres comprising a medusoid marginal nerve-ring. On many other grounds we now know that attempts to homologize Velella with a medusa are valueless.

The nerves in the tentacles, reproductive appendages, and hypostome have not been studied closely, but their existence has been confirmed.

Components of the ectodermal nervous system

There are two histologically distinct systems of nerve-fibres and, in addition, neuro-sensory cells are present. The two systems of nerve-fibres will be described under the headings ‘closed’ and ‘open’ systems. Fig. 3 shows a portion of the nervous system in which all components are well represented.

FIG. 3.

A characteristic portion of the nervous system from the external ectoderm. Based on a camera lucida drawing.

FIG. 3.

A characteristic portion of the nervous system from the external ectoderm. Based on a camera lucida drawing.

The closed system

The fibres of this sytem are the most conspicuous nerveelements in the animal, and are the elements dealt with by the earlier authors. The word ‘closed’ is used here because the processes constituting the system run together in net-like configurations. The nerves running in the mesoglea and in the invaginated ectoderm nearly all belong to this system. In the external regions the open system of fibres is equally well or even better developed, and complicates the picture. For this reason, the fine structure of the closed system is best studied in the invaginated regions.

The main features of this system are already known from the descriptions of Chun and Schneider. Chun states that the neurones are mostly 3- or 4-polar but that 2- and multi-polar examples occur; that the fibres branch frequently and run in straight lines or strong curves for long distances; and that at dichotomies and intersections a dreieckige Verbindungsplatte (referred to as a triangular veil in the present paper) is often found. Schneider (1892, 1902) describes details of the neurofibrillar system in Velella. He finds that the finest neurofibrils (Elementarfibrilleri) run throughout the nerve-cell body and processes, forming a grid or reticulum (Zellgitter) in the cell-body, and run directly from cell to cell forming an intercellular grid (Elementargitter). Some of these elementary fibrils run across the nerve-net through side branches, avoiding the cell-bodies.

The fibres of the closed system vary in diameter between 1 μ and 5 μ. The fibrillar material is usually clumped into a darkly staining axis around which the cell-membrane can be made out as a thin crumpled structure (figs. 2, F; 4, E; 5); in certain places, however, the fibrillar material can be seen to be spread out across the whole breadth of the fibre (fig. 4, F, which is another region of the same process seen in fig. 4, E; both are in the mesogloea). The triangular veils formed at dichotomies (e.g. figs. 2, E; 4, L) appear to be folds of the cell-membrane drawn out between the dividing branches. The fibrillar material at dichotomies usually divides equally, the point of divergence of the two streams of fibrils being quite precise. The fibrils are visible in the cell-body, passing around the nucleus either symmetrically or not, depending on the relative thicknesses of the processes (fig. 4, L). Points of contact in the closed system present a variety of appearances (figs. 2, C-F; 4, J, M, O). Frequently membranous veils occur, uniting the associating fibres. Usually such junctions appear continuous as far as the membranes are concerned, and frequently fibrillar streams can be seen crossing the junction and mingling with the fibrils of other processes (fig. 4, j, M). Other junctions may be discontinuous, there being no apparent mingling of fibrils or evidence of membrane continuity. The question of these neurofibrils and their interrelations will be discussed below (p. 127). The concept of the synapse is hard to relate to this material because even where the fibrils of associating processes appear to remain independent (fig. 4, o), there may be no trace of discontinuity between the membranous parts of the processes. It is perhaps best to reserve the term for physiological purposes.

FIG. 2.

(plate). Closed system. In A the arrows trace a continuous pathway between neurones B and c. The points labelled B to G are shown at higher magnification in the insets on the left.

FIG. 2.

(plate). Closed system. In A the arrows trace a continuous pathway between neurones B and c. The points labelled B to G are shown at higher magnification in the insets on the left.

In many regions it is possible to trace a pathway from one neurone to another through a series of direct connexions in which there appears to be actual fibrillar continuity. This accords with Schneider’s findings. Such a pathway (shown by arrows) is plotted out in fig. 2, A, certain of the regions being shown enlarged (fig. 2, insets B-G). Fig. 2, A also shows two neurones in the bottom left-hand region which are probably in direct connexion with each other and, by different pathways, with G and thence with B. At F an intersection probably of a discontinuous character occurs with a fibre not involved in the direct pathway under consideration.

It will be seen from this figure that fibres associating with one another frequently lose their identity, or appear to do so. Only rarely do two fibres associate, run together side by side for some distance, and then diverge again, as in the open system. The result is a nerve-net in which it is impossible to say where a fibre ends. Although fig. 2, E is here called a dichotomy of a process originating from neurone G, it might be a junction between this process and one from the left, the latter originating elsewhere. The general shape of the structure and the relative thicknesses of various fibres in the locality, however, make this rather unlikely.

Connexions of a certain type, here called ‘adhesion bridges’, appear to form a distinct category in the closed system (figs. 2; 4, G). My interpretation of these structures is shown schematically in fig. 5. It is suggested that they arise by adhesion of the membranes of two processes during growth and that the processes are subsequently carried apart by general tissue growth, the adhesion bridge being pulled out into a thin strand. All conditions from short, broad webs to strands 200 μ long and less than 12μ thick are found. In some of these, fibrillar material appears to have been drawn out into the strand or to be growing out into it, but usually the bridges seem to be empty of fibrils. In fig. 4, G what appear to be fibrils in the veil region are probably creases in which silver has been precipitated, or lines of stress. However, here and there in the nerve-net transverse connexions are found which might represent adhesion bridges in which fibrillar continuity had been established.

FIG. 5.

Suggested origin of adhesion bridges. In A an advancing fibre makes contact with another fibre. In B the cell-membranes are shown forming a connecting web. In c and D general tissue growth carries the fibres apart, stretching out the adhesion bridge.

FIG. 5.

Suggested origin of adhesion bridges. In A an advancing fibre makes contact with another fibre. In B the cell-membranes are shown forming a connecting web. In c and D general tissue growth carries the fibres apart, stretching out the adhesion bridge.

Free endings in the closed system are either single (fig. 4, N), or branched (fig. 4, H). The free extremity is membranous and without fibrils. Free endings have been found in the mesogloea (fig. 4, B), which resemble growing tips. It is possible that all free endings in the system are growing points. No muscle fibres occur in the regions shown in the photographs.

The open system. The fibres of this system do not form a net but run independently throughout their courses despite frequent close juxtapositions with fibres of both systems. The fibre diameters in this system are 14 to 12μ. It is easy to distinguish them from the majority of closed-system fibres on this basis alone. They can be distinguished from adhesion bridges, which approach them in diameter, by their discontinuous associations.

The open system is well developed in all external regions examined but is very sparsely developed in the invaginated parts. A fibre of the open system is shown in fig. 4, H (o). Sometimes the fibres in the invaginated region run at large in this way, but more often they cling to the closed-system fibres, twining around them like beans around a pole. Some of these fibres have been traced for distances between 1·5 and 2·0 mm. In the external ectoderm (fig. 4, D) the fibres show less tendency to twine around closed-system fibres, possibly because the layer is less flattened and they have more freedom of movement during growth. It is of interest that in the ectoderm below the disk, where nematocyst production adds greatly to the thickness of the layer, the fibres of both systems run freely, making few contacts with each other. The types of connexions found, however, are similar to those found elsewhere. The cell-bodies in the open system contain less cytoplasm than do those in the closed, and details of neurofibrillae have not been made out. The nuclei in the two systems are of comparable dimensions.

Free endings in the open system usually have a bladder-like form (figs. 3; 4, c). What such structures represent is discussed below (p. 127).

Neurosensory cells

Cells of this type occupy a sub-epithelial position, but do not lie as deep as the nerve-fibres. They have conical projections running up to the surface of the epithelium, from which sense hairs project. The latter are thick and short, measuring 2 to 6 μ. They colour strongly in silver preparations. The conical eminence contains granular material but there are no special structures associated with the base of the hair. A teased neuro-sensory cell is shown lying on its side in fig. 4, K. Some teased epithelial cells are also shown and it will be seen that these are wider at the free end than at the base in which the nucleus lies. The nerve-fibres run in the spaces between these narrow stems in the intact tissue. The free ends of the epithelial cells join one another, giving polygonal patterns when seen in surface view (fig. 2, A). The teased preparation (fig. 4, K) was made by stroking the surface of the tissue strip with a blunt needle just before covering with the coverslip. It will be seen that part of the fibre at the base of the neuro-sensory cell has been broken off by this procedure.

Other neuro-sensory cells are shown in fig. 3. The number of processes varies from one to five. Those with three seem most typical. The processes are generally shorter than 100 μ but some have been found as long as 150 μ. In diameter the processes compare with open-system fibres (measuring 114 to 112 μ), but unlike the latter they are often thicker and more densely staining near the cell-body than at their extremities. The processes either end freely or make contact with fibres of the open system or with those of the closed. In one preparation from the mantle region, of 28 processes examined, 6 ended on the open system and 14 on the closed, while 8 ended freely or could not be traced to their conclusions for some reason. In all cases where the junctions were clearly seen, they appeared to be of the discontinuous type. Many of the free endings are swellings, similar to those seen at the ends of the open-system fibres. Other endings are thin tapering structures. Here and there neuro-sensory cells with very short processes are found; these might be confused with interstitial cells (which are also present) were it not for their sense hairs. It is possible that these cells are undergoing differentiation into neuro-sensory cells.

With regard to distribution, neuro-sensory cells are present in all external regions examined, but are absent from the invaginated regions.

Neuro-muscular relationships

Ectodermal muscle in the adult Velella is confined to the crest which surmounts the float, to the mantle region and to the tentacles, reproductive appendages and hypostome (‘gastrozooid’). In the mantle region nerves run all over the muscle-layer and I have examined them for possible specializations. Contacts between nerves and muscle-fibres appear to be undifferentiated. No special endings or elaborations have been found. Other regions have not been closely examined.

Possible artifacts

In the closed system, as explained, fibrils are visible in the cell-bodies, at junctions and dichotomies, and often in the fibres themselves where they may be spread out or clumped into dense strands. Electron microscope studies of nerve structure in higher animals show that neurofibrils are genuine structures and are not fixation artifacts. However, it is not clear whether the fibrils seen in Bouin-fixed Velella nerve are actual neurofibrils, bundles of small, visible elementary fibrils (as Schneider maintains), or groups of submicroscopic fibrils aggregated by fixation. The methods employed in this investigation show the distribution of fibrillar material but do not give a very precise picture of the fibrils themselves and their interconnexions. From Schneider’s illustrations it would appear that he was able to make out far more detail by maceration methods. Possibly silver staining applied to material fixed in a mercuric chloride fixative would show the details more clearly.

The triangular veils occurring at dichotomies and intersections are also described in the early accounts; similar veils occur in Porpita material fixed in Baker’s formaldehyde-calcium made isotonic with the sea-water, and comparable structures are found in a number of other nervous systems, e.g. the atrial nervous system of amphioxus (Boeke, 1935; Bone, 1958). It is unlikely that they are artifacts.

The bladders at the tips of the free endings of neuro-sensory cells and in the open system resemble growing points in frog neuroblasts described by Harrison (fig. 10, c in Kappers, Huber, and Crosby, 1936). However, McConnell (1932) actually observed such bladders forming at free nerveendings in Hydra as an artifact during methylene-blue staining.

Beading or vésiculation of nerve-fibres is usually believed to be a degeneration artifact, although according to Meyer (1955) it may occur in normal healthy nerve-fibres as well as in damaged ones. Beading is occasionally seen in Velella nerve-fibres.

In some preparations which were probably stretched or pinched during dissection densely staining striae occur in the substance of the fibres and at junctions. These evidently represent slight tears or folds in or on which the silver has collected. The nerves are probably the least elastic structures in the fixed-tissue strips and they frequently show damage where other structures are intact.

Contact and continuity

Work with the electron microscope over the last decade has made it appear unlikely that there is continuity of neurones in those parts of the central nervous system or retina of higher vertebrates where cases of continuity had previously been believed by many to exist (Bielschowsky, 1928). The nature of the interconnexions in the gut plexuses is, however, still in doubt. Reiser (1959) says: ‘It is perhaps still possible that nervous tissue is not built according to one single principle but combines within it two regionally differentiated structural planes, the neuronal and the syncytial.’

In the case of Invertebrata the same holds true, but the evidence for the syncytial type of neurone system, particularly in certain giant-fibre systems, is very much stronger. Young (1939) has given the clearest possible evidence for the syncytial character of the giant fibres in Loligo. Both in annelids and in arthropods the fusion of neurones to form giant axons is known to occur. Syncytial connexions in neurone systems other than giant-fibre systems have also been described. A famous case is the enteric plexus of the leech (Apathy, cited and discussed by Bielschowsky, 1928); another is the atrial epithelial nervous system of Amphioxus (Boeke, 1935; Bone, 1958). We now have the first evidence for the simultaneous existence of a contact-system and a continuitysystem in a coelenterate. In the open system of Velella the connexions are discontinuous and are of the type described as en passant. In the closed system the neurones form a syncytium. The obvious resemblance to the double nervous system in, say, a squid is enhanced by the large size of the fibres in the closed system. They are, in fact, giant fibres in the sense defined by Nicol (1948, p. 291).

Elsewhere in the Coelenterata continuity between neurones has sometimes been reported. Workers on Hydra agree that the ectodermal nervous system is of the continuous type (Schneider, 1890; Hadzi, 1909; McConnell, 1932; Semal-Van Gansen, 1952). In hydrozoan medusae Hertwig and Hertwig (1878) found that the subumbrellar plexus in Carmarina is continuous but that in other forms the junctions were of the contact type. In Scyphozoa and Anthozoa there have been conflicting opinions, but most recent accounts favour contact (Bozler, 1927; Woollard and Harpman, 1939; Pantin, 1952). Pantin’s photos of the contact-system in Metridium make interesting comparison with my fig. 2 of the closed system in Velella. Although similar techniques were used in the two cases, the characters of the two nervous systems are obviously very different. It is not merely a question of different interpretations of a similar histological picture.

A completely satisfactory demonstration of continuity in the closed system of Velella would require experimental investigation of the living tissue and examination by electron microscopy. This would be necessary because of the remote possibility that the fibres in the system are composite structures, each formed by the interdigitation of minute terminal branches from separate processes instead of by confluence of axoplasm as would be the case in a true syncytium. Until such investigations are made, a slight element of doubt will remain. One test of a syncytium—synchronized mitosis—cannot be applied here because the neurones appear to be post-mitotic, although provided with centrioles.

Functional organization in double nervous systems

The behaviour of Velella is not sufficiently well known for an explanation of the functions of the two nervous systems to be possible at this stage. In the related form Porpita the behaviour is better known and there is a strong likelihood that a double nervous system is present as in Velella. An investigation begun in 1954 on the nervous system of Porpita was held up for lack of material. In my thesis (1956) I described continuous and discontinuous junctions in the Porpita nervous system and further referred to ‘association neurones’ which are small neurones with fine processes, apparently associated with the main part of the nervous system in some specialized way. I now think that these association neurones probably correspond to the open system of Velella, although better preparations will be needed before this can be confirmed. The behaviour of Porpita (Mackie, 1959) includes local responses in the organs of ingestion, co-ordinated tentacular movements, and curvature of the mantle, the latter being a righting reaction, possibly of a geotropic nature. The co-ordinated tentacular movements are of two sorts, one involving adoral flexions, the other aboral. Both are synchronized and, to the naked eye, appear to be through-conducted. Should through-conduction prove to occur in Velella as well, the closed system discussed in this paper would immediately come under suspicion as the probable conducting mechanism.

A system of giant fibres as well as a diffuse system occurs in scyphomedusae (Horridge, 1956a, b), the former conducting the contraction-wave during swimming and the latter being concerned in a more generalized set of functions. The giant fibres in these forms do not, however, form a continuum like the closed system in Velella. In the hydrozoan medusa Geryonia there is evidence of two conduction systems. Radial responses (between margin and manubrium) can occur at the same time as circular responses (regular swimming contractions) (Horridge, 1955). Two histologically distinct nervous systems have not been described in this form, however. It seems worth suggesting as a possible alternative to a double nervous system that in hydrozoan medusae the subumbrellar nervous system is concerned with radial responses only, the impulses for swimming contractions being conducted by the musclesheet itself. One reason for suggesting this is that in the nectophores of several siphonophores that I have examined I have found no trace of a subumbrellar nerve-plexus, although the two marginal nerve-rings characteristic of medusae are well developed. The implication is that in the absence of a manubrium and radial muscle, the subumbrellar nerve-plexus is not needed. There is also the small medusa Eucopella which lacks a manubrium and is said to be without a subumbrellar nerve-plexus (von Lendenfeld, 1883). It appears, too, that in certain Hydromedusae investigated (Krasifiska, 1914, Taf. VII, 4), where radial muscle overlies circular, the nerves are associated with the radial muscle, lying superficially, and are thus separated from the circular muscle. The puzzling absence of nerves from the velum of Hydromedusae could be explained on similar grounds: that is to say that conduction across the entire striated muscle-sheet is myoid. There is therefore no definite evidence yet for a double nervous system (as opposed to two conduction systems) in any hydrozoan outside the Chondrophora.

Structural resemblances between nerve and muscle

Reference to myoid conduction raises the whole question of fundamental resemblances between nerve and muscle. It will only be possible to touch on one aspect of this question here, that of structural similarities.

Kleinenberg regarded the epithelio-muscular cells of Hydra as ‘neuromuscular’ cells in which the free end of the cell is sensory and the base muscular; sensory-cells, nervous-cells, and muscle-cells in more specialized tissues were viewed as arising in ontogeny or phylogeny from such precursors, by differentiation of the appropriate portions. Whether or not this conception is held to have any value, it is possible to see a family resemblance, as it were, between the anastomosing muscle sheets of Hydra, Velella, &c., and the nervenets also found in these forms. In both muscle- and nerve-nets the tissue originates from simple, separate cells which later send out processes in which fibrils develop. These processes grow together, forming the net in which the fibrils from adjacent cells associate intimately. The nerve-net can perhaps be likened to a muscle ‘net’ in which the cells are wider apart, where intermingling of fibres is concentrated at certain points, and where the fibrils are much smaller and have different physiological properties.

The expenses for this work were met by grants from the General Research Fund of the University of Alberta and the National Research Council of Canada. The author wishes to record his appreciation for this support. Miss Lubow Stangret kindly prepared fig. 1.

Bielschowsky
,
M.
,
1928
.
In Handbuch der mikroscopischen Anatomie des Menschen, ed. von Mollendorff, 4 (1), 119
.
Boeke
,
J.
,
1935
.
Quart. J. micr. Sci
.,
77
,
623
.
Bone
,
Q.
,
1958
.
Ibid
.,
99
,
243
.
Bozler
,
E.
,
1927
.
Z. Zellforsch
.,
5
,
244
.
Chun
,
C.
,
1881
.
Zool. Anz
.,
4
,
107
.
Chun
,
C.
1882
.
Ibid
.,
5
,
400
.
Chun
,
C.
1902
.
In Klassen und Ordnungen des Thier-Reichs, ed. Bronn
,
2
(
2
),
339
.
Conn
,
H. W.
, and
Beyer
,
H. G.
,
1883
.
Stud. Biol. Lab. Johns Hopkins Univ
.,
2
,
433
.
Garstang
,
W.
,
1946
.
Quart. J. micr. Sci
.,
87
,
103
.
Hadzi
,
J.
,
1909
.
Arb. zool. Inst. Univ. Wien
,
17
,
225
.
Hardy
,
A. C.
,
1956
.
The open sea. Its natural history: the world, of planckton. London (Collins)
.
Hertwig
,
O.
, and
Hertwig
,
R.
,
1878
.
Das Nervensystem und die Sinnesorgane der Medusen. Leipzig (Vogel)
.
Holmes
,
W.
,
1947
.
In Recent advances in clinical pathology, ed. Dyke. London (Churchill)
.
Horridge
,
G. A.
,
1955
.
J. exp. Biol
.,
32
,
555
.
Horridge
,
G. A.
1956a
.
Quart. J. micr. Sci
.,
97
,
59
.
Horridge
,
G. A.
1956b
.
J. exp. Biol
.,
33
,
366
.
Kappers
,
C. U. A.
,
Huber
,
G. C.
, and
Crosby
,
E. C.
,
1936
.
The comparative anatomy of the nervous system of vertebrates, including man. New York (Macmillan)
.
Krasinska
,
S. V.
,
1914
.
Z. wiss. Zool
.,
109
,
256
.
Leloup
,
E.
,
1929
.
Arch. Biol. Paris
,
39
,
397
.
Lendenfeld
,
R. von
,
1883
.
Z. wiss. Zool
.,
38
,
497
.
Mackie
,
G. O.
,
1956
.
Thesis for the degree of D.Phil. Oxford
.
Mackie
,
G. O.
1959
.
Trans, roy. Soc. Canada, 53 (3), 7
.
McConnell
,
C. H.
,
1932
.
Quart. J. micr. Sci
.,
75
,
495
.
Nicol
,
J. A. C.
,
1948
.
Quart. Rev. Biol
.,
23
,
291
.
Pantin
,
C. F. A.
,
1952
.
Proc. roy. Soc. B
,
140
,
147
.
Peters
,
A.
,
1955
.
Quart. J. micr. Sci
.,
96
,
323
.
Reiser
,
K. A.
,
1959
.
In Handbuch der mikroscopischen Anatomie des Menschen, ed. von
Môllendorff, 4 (1) Supplement
,
185
.
Schneider
,
K. C.
,
1890
.
Arch. mikr. Anat
.,
35
,
321
.
Schneider
,
K. C.
1892
.
Jena. Z. Naturw
.,
27
,
379
.
Schneider
,
K. C.
1902
.
Lehrbuch der vergleichenden Histologie der Tiere. Jena (Fischer)
.
Semal-Van Gansen
,
P.
,
1952
.
Bull. Acad. Belg. Cl. Sci
.,
38
,
718
.
Totton
,
A. K.
,
1954
.
‘Discovery’ Rep
.
27
,
1
.
Totton
,
A. K.
and
Mackie
,
G. O.
,
1960
.
‘Discovery’ Rep. (in press)
.
Woollard
,
H. H.
, and
Harpman
,
J. A.
,
1939
.
J. Anat. Lond
.,
73
,
559
.
Young
,
J. Z.
,
1939
.
Phil. Trans, roy. Soc. B
,
229
,
465
.