The Embryology of the Viviparous Ophiuroid Amphipholis equamata Delle Chiaje.*^*

By H. Barraclough Fell, M.Sc., Ph.D.,
Victoria University College, Wellington.

that the problem of direct development in ophiuroids should be investigated. I have also to thank Dr. Fabius Gross, of Edinburgh, for advice on the use of Erdschreiber culture medium for excised embryos, and also the Directors of Marine Laboratories at Millport and Plymouth for supplies of living and preserved material, to supplement that collected in the Firth of Forth and at Island Bay, New Zealand. The terminology employed for the germ layers is that suggested by Professor E. S. Goodrich, of Oxford, and employed by the writer in previous papers. Points of interest are discussed as they arise during the course of the paper, instead of leaving a general dicussion to the end.

Nomenclature.

From examination of the literature, it appears that Amphipholis squamata has received no less than eighteen different names during the last one hundred and thirty years. Leach (1815) first described the species under the name Ophiura elegans. In 1828 it received the specific name which it now bears, being named Asterias squamata by Delle Chiaje. In succeeding years it became Ophiura neglecta—under which name it is recorded by Forbes (1841) as occurring at Newhaven “in the Frith of Forth.” In 1842 it received the name Ophiolepis squamata, and under this designation the first embryological papers on the species appeared, by Milne-Edwards, Krohn, and Schultze (see below). In 1861 Sars gave the ophiuroid the name Amphiura squamata, which it was destined to bear for the remainder of the century. More recently it has been separated from the genus Amphiura, and it therefore reverts to the generic name Amphipholis used by Ljungman in 1872, still retaining the specific name squamata of Delle Chiaje. Under this name it appears in Mortensen's “Echino-derms of the British Isles” (1927). It was during the period when it was known as Amphiura squamata that most previous work was done on its development. The other names the species has borne are not of importance for the present purpose, for they have not been used in any embryological papers.

Historical Summary.

With the close of the nineteenth century research on the subject came to an end, leaving a series of accounts so conflicting not only with what was then known of other ophiuroids, but also with one another, that MacBride in 1914 made only a passing reference to the species in his “Text-book of Embryology.” He quotes, however, Ludwig's work on the development of the skeleton, and his own work on the late development of the ovoid gland and related structures; but as the findings of the previous workers, such as Apostolides (1882) and Russo (1891) differed widely from what had been found for Ophiothrix fragilis, he regarded their statements as “improbable in the highest degree.”

In order to bring out more clearly the questions most in dispute, it is preferable to give a comparative table of previous results under subject-headings, rather than a purely chronological list of workers and their findings. This is set out as follows:—

1. Viviparity.

2. Hermaphroditism.

3. Embryonic Attachment.

4. Gastrulation.

5. Larval Oesophagus.

6. Larval Anus.

7. Rudimentary Larval Skeleton.

8. The Coelomic Vesicles.

9. Perivisceral Coelom.

10. Epineural and Perihaemal Coloms.

11. Nervous System.

• Cuenot stated that the nervous system arises through the inturning of gutters of ectoderm from the ventral surface of the arms.

• Russo, on the other hand, stated that the nervous system takes its origin from “four yellow cells” seen by him in an unstained living embryo on either side of the stomodaeum.

The Problem.

My recent research has been concerned with elucidating three principal questions. These are:—

• (1) What is the true nature of the development of Amphipholis squamata, and, in particular, what is the mode of origin of the coelom in the species?'

• (2) To what extent has viviparity affected development, and in what ways has it done so?

• (3) What are the causes of direct development in ophiuroids, and how have they acted?

This paper is concerned only with the first of these problems, namely, elucidating the real mode of development of Amphipholis squamata. The second and third questions are considered in a separate paper (Fell, 1945) in which are utilised some of the results given here.

Technique.

The methods used in this work are for the greater part the same as described in an earlier paper (Fell, 1941) and need not be repeated. The embryo of Amphipholis contains yolk material, rendering it both opaque and brittle, but not so markedly so as in Kirk's ophiuroid. It is essential not to leave the egg more than a few minutes in each of the xylol and paraffin baths (two minutes in each bath was generally found enough, using strengths of wax to xylol of 25%, 50%, 75%, and 100% wax). The opacity of the earlier stages makes whole mounts impracticable, while unfixed and living material is practically useless. The inaccurate results of previous workers are undoubtedly largely attributable to their use of living material or of imperfectly fixed preparations. It was found that the yolk granules of the egg and embryo of Amphipholis are far less basiphilic than those of Kirk's ophiuroid, and were very much smaller. Consequently ordinary staining methods for nuclei and cytoplasm could be employed, and it was unnecessary to adopt the special modified staining technique described in my previous paper.

The viviparous habit of the species, however, presented a problem not encountered in my previous work, and a number of special methods had to be devised. To obtain the embryos for study it was necessary to discover an anaesthetic for the parent of a nature such as to leave the delicate embryo unharmed by any convulsive contractions of the maternal tissues. Chloral hydrate, as used in 5% concentration in sea water in my previous work proved quite unsuitable. Chloroform and ether also produced violent shock, and nicotin was found to bring on an intense muscular rigor. Finally it was found that a solution of 2½–5% of menthol in sterile sea water produced a gentle anaesthesia, followed afterwards by complete recovery if not prolonged.

Owing to the viviparity it is not possible to observe the developmental process taking place within the bursa. Previous workers relied upon isolated stages excised from the parent and examined individually, a method which is only satisfactory if a large number of intermediate stages is available.

in Gross's paper (1937). One modification of Gross's technique was made in view of the supposed intolerance by echinoderms of fresh water; to the final filtrate of soil extract was added the equivalent amount of dissolved sea salts from an equal volume of sea water. However, little advantage seemed to result from this addition, and other experiments have since convinced me that Amphipholis is by no means so sensitive to change in the density of the sea water as echinoderms are generally supposed to be. Some specimens were allowed to remain in a culture jar from which the sea water very slowly evaporated over a period of ten weeks. At the end of that period the sea water was highly hypertonic, having a salinity of 58 per mil. when titrated, and yet several specimens, both adults and newly-born embryos, remained alive till the end of the period.

Adults kept alive in aquarium tanks and Petri dishes in the laboratory were fed upon a diet of diatoms (Skeletonema) which were inoculated into the sea water at intervals.

As before, polarised light was used for the examination of the developing skeleton. For the purpose of decalcifying the embryos before imbedding, the new method of Wilks (1938) was used and gave very satisfactory results. By the use of sodium hexa-meta-phosphate calcareous structures may be removed from delicate tissues without any evolution of carbon dioxide or other gases; thus the method is of the greatest advantage in embryology, insofar as it removes all possibility of artificial cavities being produced.

Distribution.

Amphipholis squamata is believed to be the only ophiuroid with a world-wide distribution. It appears to vary but little, and in examples taken from the coastal rock-pools of Cook Strait, New Zealand, I have found no obvious difference in form or habit from those inhabiting the waters of Britain, twelve thousand miles distant.

Now, the existence during development of a free-living larval stage in many marine animals has been used as a means of explaining the wide distribution of some species. It is argued by this theory that while in the free-living planktonic stage a species may be carried over distances impossible for it to cover when in the adult, bottom-dwelling stage. In a communication to me, my friend, Mr. J. E. G. Raymont, of Edinburgh University, has pointed out the discrepancy between this theory and the distribution and life-history of Amphipholis; for although Amphipholis squamata has the most widespread distribution of any ophiuroid, yet it lacks a free-living larval stage, while other ophiuroids with a more restricted distribution possess well-developed larvae. The adult Amphipholis squamata is a typically littoral form, never extending below 125 fathoms (Mortensen, 1927).

In view of its strictly littoral distribution, the presence of the species in New Zealand is of the greatest interest, seeing that as it lacks a pelagic stage it can only have reached that country by way of shallow seas which no longer exist in the area. It occurs in South America, and Mortensen (1924) records it from the Sub-antarctic Auckland Islands. This distribution argues strongly in favour of the theory held by many biologists of the former existence of a land-

bridge or island chain, between New Zealand and South America, via the Antarctic continent. In the same way certain earthworms have been shown by Benham to be shared in common by New Zealand, the Sub-antarctic islands, and South America; also the floras of New Zealand and South America have remarkable points of similarity.

Further discussion would be irrelevant to the present paper, but it seemed advisable to mention the problem owing to the bearing on it of the mode of development of Amphipholis.

Breeding Habits.

There is good reason to believe that Amphipholis squamata, unlike most other ophiuroids, breeds throughout the year, a constant stream of eggs being produced one by one and fertilised. Thus at any time of year embryos may be obtained at various stages of development, though the earliest stages are of course very rare, owing to the comparative rapidity of the first cleavages, and the fact that only one egg is formed at a time. The difficulties in obtaining information on these early cleavages are thus closely comparable to those encountered in the mammalia.

Garstang recorded that A. squamata breeds at Plymouth, England, from May to September. However, specimens forwarded to me from Plymouth in February, 1940, during an exceptionally cold winter, proved to be carrying embryos at various stages, so that there is no reason to believe that the species has a restricted breeding period at that locality. Specimens from the Firth of Clyde area indicate that breeding continues uninterruptedly throughout the year on the West coast of Scotland. In the Forth area, on the other hand, somewhat different results were obtained, probably as a result of the more severe winter conditions of the North Sea. Thus, from January to March, 1940, no specimens at all could be obtained, and it was not till June of that year that the species was again common. However, whenever specimens have been obtainable in the Forth area, they have always been pregnant, so that it would seem that so long as the weather conditions are suitable for the life of the adults, these will breed.

In Forth specimens it is unusual to find more than five of the ten bursae pregnant at any one time, and each bursa seldom contains more than one embryo. In specimens from the Clyde, on the other hand, not only are the adults themselves much larger than examples from the Forth, but they also carry a greater number of embryos, and the latter are often much more advanced. Thus it seems that on the milder Atlantic coast the embryos are carried for a longer period before being born. This is a peculiar fact, as it is more usual for brood protection to be more highly developed in unfavourable conditions. In one particularly prolific specimen, forwarded from Millport, the following embryos were extracted: in one bursa were an embryo having 8–9 arm-segments, an embryo with 3–4 arm-segments, and an embryo at the “pentagon” stage; in another bursa were two embryos, one with 10–12 arm-segments, and one with 3–4 arm-segments; in a third bursa were two embryos with 3–4 arm-segments; one bursa was empty; and the remaining six bursae each contained one embryo, ranging in development from 3–10 arm-segments.

In regard to its habits in other parts of the world there is less information. Fewkes (1887) has recorded it as breeding in August and September at Newport, U.S.A., but he does not make it clear if these are the only months. Bernasconi (1928) records it as breeding at Neochea, Argentina. In New Zealand I have taken breeding specimens from July to March, and have little doubt that it breeds there also in the remaining months when I had no occasion to go collecting. These observations were made at Island Bay, on the northern shore of Cook Strait.

Fig. 1.—The genital organs of Amphipholis squamataa.
A dissection of an interradius, with the left bursa opened to show the contained embryo. The latter has almost reached the end of its intrabursal development, and is oriented in the characteristic inverted position with the arms directed toward the genital cleft.
Ab., abradial genital plate; Ad., adradial genital plate; B., bursa; E., embryo; O., ovary; T., testis; Stom., stomach.

The Reproductive Organs.

The morphology of the reproductive organs of Amphipholis squamata follows the general pattern of the whole genus, and the arrangement of the gonads can scarcely be distinguished from the condition described by Mortensen for A. tenuispina, A. japonica, and A. sabrina (Mortensen, 1920). As, however, the relationship of the gonads to the bursae is of great importance to understanding the subsequent history of the embryo, it is necessary to give here a somewhat more complete account.

There are ten bursae, so distributed that there are two in each of the five interradii. Each opens to the exterior through the narrow genital clefts, which lie on the under surface of the disc, one on either side of the base of each arm. Each cleft is bordered by two skeletal plates, the abradial and adradial genital plates respectively (Fig. 1). Attached to the dorsal (aboral) surface of each adradial genital plate is a single testis, thus making ten testes in all; the ten ovaries are each similarly related to an abradial genital plate. Occasionally a two-lobed testis is seen, or, less frequently, two testes are found attached to a single adradial plate. Occasionally, also, the ovaries are found to be secondarily increased in number in this way. The ovaries occupy a somewhat more peripheral position than do the testes, and thus when a series of vertical sections is being examined, the testes of any particular sector are seen cut across in sections nearer the centre than those in which the ovaries are cut.

As the relationship of the gonads to the bursa and neighbouring organs can best be understood by means of vertical sections, two examples are illustrated (Figs. 2A and 2B). In Fig. 2A is seen part of a vertical section through an interradius, including a bursa. The ovary of this bursa is seen to the right, on the abradial border of the genital cleft. To the left of the bursa is the basal part of the neighbouring arm, cut in vertical transverse section. The part of the bursa which is cut across does not show any portion of a contained embryo, but the presence of the sinuses in its wall indicate that there is an embryo present in another part of the organ (see section dealing with the bursa, and nutrition of the embryo, below). The ovary, which contains a small number of oogonia in various stages of maturity, is ensheathed by a thin wall one cell in thickness. That side of the ovary toward the bursa is closely adpressed against the latter, so that it would not be difficult for a ripe egg to be liberated through the bursa directly into the lumen. Further details in regard to the ovary are given in the section dealing with the ovum and oocyte.

In Fig. 2B is seen a similar vertical section from the same series, but differing in that it has been cut somewhat further in toward the centre of the disc, and thus the ovary is absent from the abradial side, while the testis is visible on the adradial border. There is also seen the testis belonging to the interradius on the other side of the arm. As the bursa is cut more centrally, an arm of a contained embryo is seen cut in transverse section. The testis, like the ovary, is very simple, comprising only a spherical capsule one or two cells thick containing loosely-arranged spermatogonia. Each of the latter is a spherical cell with a very large and deeply-staining nucleus. It is

notable that the condition of the gonads in this species is so extremely primitive, and it is to be contrasted with the complex reproductive glands with gonoducts which have recently been described by Smith (1940) for Ophiothrix fragilis.

Fig. 2.—Vertical sections showing the relation between the bursa, ovary, testis and related organs.
Stom., stomach: Perit., peritoneum; Coel., general coelom; Con Tis., connective tissue; D., lower surface of disc; Gen.Clft., genital cleft: Arm, portion of arm bordering the genital cleft; N.Rad., radial nerver; Perb.C., perihaemal coelom; Amb.C., ambulacral canal; Mus.L., longitudinal muscle fibres of arm; Burs., bursa: Emb., embryo; Ad.H., adradial horn of bursa; Ov., ovary; Test., testis.

Viviparity of Amphipholis.

The fact that Amphipholis squamata, together with some other echinoderms, is viviparous raises an important question which appears almost to have escaped notice hitherto. It is a belief very common among embryological workers, and one frequently referred to in current papers (e.g., Just, 1928), that the coelomic fluids of echinoderms exert a strong toxic or inhibitory effect upon the ova. This, it is stated, is the mechanism by which embryos are prevented from developing within the body of the parent. In descriptions of methods of artificial fertilisation of echinoderm eggs it is stated with emphasis that eggs will not develop when excised from the gonad and gonoducts unless all traces of coelomic fluids are carefully excluded. The point arises here of how it is possible for embryos of Amphipholis squamata to develop within the thin-walled bursae of the parents; for it is improbable that traces of coelomic fluids would not diffuse through the bursal wall when it becomes thin and stretched by the growing embryo; and, in any case, as shown below, there are strong reasons for believing that a fluid is actually secreted by the bursal wall on to the embryo.

Furthermore, a number of cases have been recorded during the last forty years in which viviparity has occurred in species normally oviparous—and in which, according to the theory of “coelomic toxins,” internal development is not possible. It is a notable fact that in these cases of viviparous habit in oviparous forms, the condition is associated with, and can be fairly certainly regarded as caused by, a hermaphrodite state of the reproductive organs. The following records are known to me of hermaphroditism occurring abnormally in echinoderms: Asterina gibbosa (Cuenot, 1898); Asterina batheri (Ohshima, 1929); Asterias glacialis (Delage, 1902, and Buchner, 1911); Sphaerechinus granularis (Viguier, 1900); Strongylocentrotus droebachiensis (Gadd, 1907); Paracentrotus (or Strongylocentrotus) lividus (Herlant, 1918; Gray, 1921; Drzewina and Bohn, 1924; Paspaleff, 1927); Arbacia pustulosa (Gray, 1921); Arbacia sp. (Heilbrunn, 1929); Echinocardium cordatum (Paspaleff, 1927; Moore, 1935); and Echinus esculentus (Moore, 1932). The tendency to develop hermaphroditism, as shown by the above list, is thus not confined to one or two examples. Of the instances given above, the hermaphroditism is described as functional in six cases. Thus, in the specimen of Echinocardium cordatum examined by Moore (1935) there were larvae and eggs in the gonoducts, eggs having been fertilised by the animal's own spermatozoa.

It is thus quite clear that the supposed inability of echinoderm eggs to develop in proximity to the body fluids does not agree with the facts as observed in nature, and the belief rests entirely upon some data obtained from artificial fertilisations made in the laboratory. All the facts brought together above are in support of a claim made by Bogucki (1930), who also challenged the truth of the theory. He, however, approached the question from the results of experimental cultures of excised echinoderm eggs, artificially fertilised. Bogucki claimed that inhibition of development was not caused by the body fluids of the parent, but happened only if the gut was

ruptured. He concluded therefore that this organ is the source of the property incorrectly attributed to the coelomic fluids, As two independent lines of investigation lead to the same result, it seems that there can no longer be any reason to believe in the toxicity of echinoderm body fluids, and the case of Amphipholis squamata ceases to appear anomalous.

The Bursa.

A vertical section through the non-pregnant bursa shows that the lumen of the organ is very small, while the bursal wall is many cells in thickness, and contains few if any blood-sinuses of any importance. At the aboral (dorsal) apex of the organ there is visible a downwardly projecting cone of tissue appearing in section triangular. This may be termed the “median cone.” This gives the lumen an appearance of bifurcating, when it is viewed in section. As this subdivision of the lumen is of importance during the later growth of the embryo, it is convenient to give distinctive names to the two portions separated by the median cone. The portion of lumen directed adradially, i.e., towards the testis, may be termed the “adradial horn,” and the part directed towards the ovary the “abradial horn.”

In the earliest stage of pregnancy the embryo, as might be expected, is found occupying the abradial horn of the bursa, not far below the ovary. After the germ layers have been differentiated the embryo becomes closely invested by the wall of the abradial horn, and at one point a fusion of the parental and embryonic tissues occurs (see later section of this paper). The region of the embryo to be affected in this way is aboral in relation to its own axes of symmetry. The adoral pole of the embryo becomes directed toward the centre of the lumen of the bursa. At the same time numerous blood sinuses begin to become visible in the wall of the bursa, especially in the median cone. At the point where the embryo has become attached to the bursal wall, the latter grows out into a stalk, and thus the embryo now comes to lie suspended in the bursal cavity. As the embryo increases in size the wall of the bursa stretches, till the median cone completely disappears as such, its tissues having contributed to the expanding wall. The embryo thus occupies the entire lumen, and this condition holds during the rest of the development.

The Oöcyte and ovum.

As the sexual products of Amphipholis are never at any time very abundant, it is not possible to give any detailed account of oogenesis, but some general observations may be included here. None of the earlier workers on the embryology of the ophiuroid studied these stages with the single exception of Russo (1891). His account, however, is very unsatisfactory, and cannot be seriously considered. He states, for example, that the ovum has a diameter of 4μ, of which 2μ are occupied by the germinal vesicle. Any ophiuroid having so minute an egg would indeed be remarkable, as it would be some 4,000 times smaller than the smallest known echinoderm egg (that of Toxopneustes variagatus, of 65μ diameter). Russo probably observed very immature oogonia. Some interesting comparisons, however, can be made with observations made on other echinoderm eggs. A short summary is given here before noting my own observations for Amphipholis.

to mask the nuclei. As the ovary possesses no gonoduct, it can only be supposed that the ovum enters the bursa through a rupture temporarily formed in the bursal wall. Unlike other ophiuroids, which have large gonads, it cannot be supposed here that pressure of the growing sex-cells ruptures the bursal wall, thus liberating the genital products; but as the portion of wall neighbouring the ovary is of no great thickness, it would not require more than a slight cytolytic action of the egg to bring about the perforation of the bursa.

If the egg of Amphipholis squamata is compared with those of other ophiuroids, it is seen to occupy an intermediate place in a series which ranges from small eggs with comparatively little yolk and having a long, indirect and pelagic development to such very heavily yolked types of egg, of large size, and with a shortened, direct and non-pelagic development. The importance of yolk content and egg-size, and their relationship to the type of development followed, has been discussed in detail in a separate paper (Fell, 1945).

Fertilisation.

Fertilisation has not been observed. Attempts to fertilise artificially such few ova as were obtained did not succeed. It is not known for certain whether the animal is self-fertilised, but it is a significant fact that so many viviparous ophiuroids are also hermaphrodite, a condition which suggests strongly that self-fertilisation accompanies viviparity in ophiuroids. This deduction, however, does not necessarily follow,. for in the New Zealand ophiuroid Ophiomyxa brevirima the sexes are distinct, though the animal is viviparous. This would indicate that copulation of some kind occurs in that species. The nature of the copulation required by these forms, the eggs of which cannot come in contact with the spermatozoa after extrusion from the body of the parent, may be deduced perhaps from the observations of Mortensen (1933) on the East African form Amphicyclus androphorus. Mortensen observed that in this species the female carries the male, which is very much smaller, on her oral side, so disposed that the ventral (oral) side of each animal is adpressed to the other. This type of copulation has also been seen to occur in Ophiosphaera insignis and Ophiodaphne materna. With the exception of the asteroid Archaster, in which a male-female superposition occurs (Ohshima and Ikeda, 1934), copulation is quite unknown elsewhere in the echinoderms. Now, none of the three ophiuroids quoted above is viviparous, so the conditions are not strictly comparable to those obtaining in Amphipholis squamata, but till such time as self-fertilisation has actually been observed, copulation must still be regarded as a possibility. The fact that hermaphroditism in the oligochaetes does not dispense with the need for cross-fertilisation gives added reason for caution. It would be necessary to rear young Amphipholis in isolated tanks from an early ontogenetic stage, before the ripening of the sex-organs, to establish whether reproduction can occur by self-fertilisation. It should be noted here, however, that during the last two years when adults have been kept under observation in aquaria, no case of copulation has been observed, though it was seen that the animals tended to congregate in one part of the tank.

that some of his statements, especially those in regard to a schizocoelous origin of the coelom, were in fact supported by observation of the development of Kirk's ophiuroid. But in regard to his claim that the endoderm arose by delamination in Amphipholis, I was unable then to comment, because such a process does not occur in Kirk's ophiuroid. It was partly with the intention to examine his claim for delamination that the present research was undertaken. As has now been shown, in actual fact the origin of the germ layers and mode of gastrulation probably differ only in a minor degree from what have been described for Kirk's ophiuroid—and in both species the mode of gastrulation can be seen to be merely a peculiar modified form of invagination, produced by the effect of yolk material. There seems to be no possibility whatever of delamination taking place. There can be little doubt that these mistaken interpretations of the gastrula were in large measure the result of the unsatisfactory histological methods employed by earlier workers; it is not surprising that the opaque appearance of gastrula in unfixed, or improperly fixed, and cleared material led these workers to confuse it with the superficially similar gastrula of a coelenterate.

Origin of the Hydrocoel and Coelomic Vesicles.

Sections through somewhat older embryos indicate that the mesenchyme mass occupying the aboral region of the embryo continues to increase, the cells proliferating particularly along that side of the embryo which is destined to be the right side of the larva. At the same time the mesenchyme cells in this right anterior region begin to round up into two well-defined masses, at first solid. These come to bulge forwards, gradually pushing the oesophageal mass over

Figure 7.
Ect., ectoderm; Oes., oesophageal sac; Arch., archenteron; Ant.P;, anterior pole; Hydr., hydrocoel; Lft.P.Coel., left posterior coelom; Mesch., mesenchyme.
Note: as the section is viewed from the (future) ventral aspect, right and left are reversed.

their greatest development simultaneously, for while the coelomic vesicles are disappearing the vestigial larval skeleton and the hydrocoel are growing.

Contrary to certain inaccurate figures and statements made in the papers of Metschnikoff (1869), Russo (1891) and others, and unfortunately reproduced in subsequent literature, there is no opening to the exterior either via an oesophagus or anus. Both the archenteron and the oesophageal sac are still closed vesicles, neither communicating with the other.

The Vestigial Pluteus.

The embryo is gradually growing larger, there being a notable increase in the amount of the mesenchyme, especially at the aboral (posterior) pole. This tissue comes to surround completely both the oesophagus and hydrocoel as well as the archenteron, filling up the space between these organs and the ectoderm. At the same time the mesenchyme cells become more compacted together, so as to eliminate the intercellular spaces originally present. Fig. 11 is an oblique

Fig. 11.—Oblique section of later larva, showing five-lobed hydrocoe.
Arch., archenteron; Hydr., hydrocoel; Rud.Amb.C., rudiment of ambulacral canal; Oes., oesophageal sac; Mesch., mesenchyme; Ect., ectoderm.

section through the larva, so directed as to pass through all lobes of the hydrocoel—these having now reached the maximum number of five; for no further differentiation of the organ occurs till the pluteus stage has been superseded. Each lobe is a rudiment of a future radial ambulacral canal, and at present they are placed in linear series in an antero-posteral direction on the left side of the oesophagus. The cells of the archenteron—or stomach as it may now be termed—have multiplied to form a wall two cells deep. In Fig. 11 the section, being oblique so as to show all lobes of the hydrocoel, does not pass through the centre of the stomach, and so only a small part of it is seen in the section. The wall tissue of the oesophagus

• (5) The only part of the original coelomic sacs to survive is the well-developed hydrocoel. This is interesting in view of the conditions found in Kirk's ophiuroid, where the hydrocoel is the part of the definitive coelom which is first brought into being. Here also the hydrocoel is the first part of the definitive coelom to form, there being no general coelom present.

• (6) The mesenchyme is very greatly developed, a condition also paralleled in Kirk's ophiuroid. In the present instance, as in Kirk's ophiuroid, the mesenchyme is destined to play a most important part in the subsequent differentiation of the coelom of the adult.

Before passing on to describe the metamorphosis, it is convenient at this point to consider in greater detail the larval skeleton, the embryonic attachment, and the mode of nutrition.

The Larval Skeleton.

As previously noted, the larval skeleton originates as two small meshes of calcareous material situated towards the aboral pole of the larva. Although I have not been able to observe a specimen showing the structure at the moment of their earliest appearance, there can be little doubt that here, as in the case of nearly all other skeletal plates of ophiuroids, the rudiment is at first a triradiate spicule in either case (see Fewkes, 1887; Woodland, 1907; and Fell, 1941). In my paper on Kirk's ophiuroid (1941) a description was given of a new method of observing the development of the echinoderm skeleton by means of polarised light. The same method has been employed in the present research, giving similar results. Here also each skeletal plate consists of a single crystal of calcite, shining with a bright golden light under crossed nicols, the remaining tissues being darkened. Extinction takes place along two axes which are at right angles. It was found here also that the longer and shorter morphological axes of the skeletal plates correspond approximately with the crystalline axes of the plates, as indicated by extinction. Thus in Amphipholis, as in Kirk's ophiuroid, there is a correspondence between molecular orientation and the anatomical orientation of skeletal plates. It is an extraordinary fact that protoplasm should have the power of causing calcite crystals to be precipitated with their molecular chains aligned along predetermined axes. It is perhaps to be compared with the property of mica of causing crystals of other substances to be formed in parallel series on its surface, a suggestion for which I am indebted to Dr. C. A. Beevers, of Edinburgh. This property of mica depends on the molecular orientation of its surfaces; possibly therefore the surface layers of the protoplasm of the skeletogenous cells may act in some similar way.

An average of eight readings gave the angle of inclination of the axes of the skeletal plates of the pluteus as approximately 41°. Comparative study shows that as a general rule each of the two plates has one long, slender branch extending laterally on either side of the stomach, and sometimes two of these branches may occur, one longer than the other. These clearly are vestiges of the slender supporting rods which in the typical Ophiopluteus extend beyond the central body-mass and out into the arms.

The Embryonic Attachment (“Nabelschnur”).

This structure, as already indicated, is an outgrowth of the wall of the bursa, from the position at which the embryo at first becomes adpressed to it, after liberation of the egg from the ovary—that is, from the abradial wall. With the growth of the embryo the portion of the bursal wall immediately related to it begins to grow out, to form an elongate stalk-like organ with the embryo at its distal extremity. One is reminded of the analogous “placenta” by which the gametophyte of an angiosperm plant is attached to the wall of the seed-capsule formed by the parent sporophyte. In the figure of the fully-developed larva (Fig. 12) this organ is also seen at its greatest development. In section its structure proves to be very simple (see Fig. 10, Emb.At.), comprising only a tissue of undifferentiated cells. These are somewhat spindle-shaped, elongate and staining only lightly with cytoplasm dyes. There are no sinuses or other vascular structures present in the organ, and there is little but superficial resemblance to an umbilical cord. After the assumption by the embryo of radial symmetry (see below), the attachment survives for a while, but becomes shorter again. It is then attached to one of the interradii of the young star, towards the dorsal side (Fig. 20). After the completion of the “pentagon” stage of the young ophiuroid, the attachment atrophies, the embryo breaking off and lying freely in the lumen of the bursa. A young free embryo is sometimes found in the bursa with the stump of the embryonic attachment still projecting from one interradius.

Nutrition of the Embryo.

The earliest workers on Amphipholis squamata (Krohn, Schultze, Metschnikoff) had noted the existence of the embryonic attachment, which was named “Nabelschnur” in reference to a supposed nutritive function analogous to that of the umbilical cord of a mammalian embryo. Russo (1891), as already mentioned, denied the existence of the organ, saying that the embryo merely adhered to the parent by a “pocketing” (insaccatura) of the bursal wall. Elsewhere in his paper he says that it is held in place by a “kind of cement” (una specie de cemento). This substance is said to be derived from the degeneration of the epithelium of the bursa. Needless to say, these statements are absurd and untrue; it is difficult to understand how such an error of observation could have been made. As will now be seen, these errors were but the forerunners of far worse confusion.

When Russo denied the existence of the umbilical cord described by earlier workers, he thereby deprived the embryo (as he described it) of what had hitherto been regarded as its organ of nutrition. This led him to develop what can only be described as a fantastic hypothesis, though Russo himself actually set it down as an observed fact. He states that certain cells degenerate from the bursal epithelium, and the embryo devours these, drawing them into the stomach by means of contractions of the oesophagus. It can only be observed that (a) Russo did not explain the methods by which he was enabled to observe the embryo feeding within the bursa

brought about in the bursa must be attributed to the direct contact stimulation provided by the embryo itself.

The Metamorphosis.

The description of the development of the embryo has been brought up to the stage of the fully-formed larva. The transformation of this simplified, but nevertheless easily recognisable pluteus with bilateral symmetry into the radially symmetrical ophiuroid can now be described.

Fig. 14.—Larva viewed from the left side, at the commencement of metamorphosis.
Emb.At., embryonic attachment; Lft.Skel.Pl., left skeletal plate of the larva; Stom., stomach; Hydr., hydrocoel; Oes., oesophageal sac; Mesench., mesenchyme; Skel.Rud.Ad., triradiate skeletal rudiment of the adult; Rt.Skel.Pl., right skeletal plate of the larva.

As usual in the ophiuroids, metamorphosis seems to be initiated by a change in the position and shape of the hydrocoel. The organ begins to curve round the oesophagus (Fig. 14) so that the five lobes originally placed in linear series on the left side now begin to take up a position such that each lobe begins to point outwards, the lobes being at equally spaced intervals about the oesophageal sac. Not all stages of this process have been obtained, but there is no reason to believe that there are any important differences from what has already been described for Ophiothrix fragilis by MacBride (1907) and Ophiocomina nigra by Narasimhamurti (1933). No doubt the change in the shape and relations of the hydrocoel is brought about by the liberation of a growth substance, as first recorded in echinoplutei by Huxley (1928). The ring canal becomes established about the oesophagus when the two extremities of the hydrocoel meet

and fuse on the right side of the oesophagus, whilst the five lobes already present form the five radial ambulacral canals. Nothing was found comparable with the peculiar process of rotation through 360° of the hydrocoel of Ophiura brevispina, as described by Grave (1900). While the hydrocoel is encircling the oesophagus, there occurs a flattening of the embryo, shortening the antero-posteral axis and causing the embryo to become more spherical in form. The anterior surface of the embryo becomes at the same time rather flattened.

It thus happens that the anterior (or apparent anterior) hemisphere of the larva becomes directly transformed into the ventral half of the young ophiuroid, and the posterior hemisphere becomes dorsal. The correspondence of the regions, however, is not an exact one, for the actual posterior pole of the larva is found after metamorphosis to have become dorso-lateral, being midway between the central point of the dorsal surface of the star and the periphery of one interradius. The only landmark of value which remains unaltered during the metamorphosis is the embryonic attachment—which, it will be recalled, is united to the posterior pole, thereby providing a useful index to the position of that region. This organ therefore comes to be dorsal and interradial in the young attached ophiuroid. The interradius to which the embryonic attachment is joined is always either of the two interradii which are furthest from the sector subsequently occupied by the madreporic canal (sec Fig. 20).

Fig. 15.—Vertical section through a newly-metamorphosed star, still attached to the parent.
Pod, podium; Amb.C., ambulacral canal; Mesch., mesenchyme; Coel.Sp., intercellular schizocoelous splits forming the perivisceral coelom; Stom., stomach; Emb.At., embryonic attachment; Ect., ectoderm; Amb.R., ambulacral ring; Oes.Rud., oesophageal rudiment; Mo.Rud., mouth rudiment.

Origin of the Perivisceral Coelom.

Soon after the embryo has assumed radial symmetry a change is observed in the texture of the mesenchyme, closely similar to that which has already been described for Kirk's ophiuroid (Fell, 1941). Between some of the cells of this tissue a process of splitting begins to take place, producing a number of intercellular cavities which are at first small. Serial sections reveal that there is another similarity to what occurs in Kirk's ophiuroid, in that this process of splitting begins in the outermost zone of mesenchyme, extending only subsequently to the more dorsal and central region. Thus, the mesenchyme immediately above the stomach is the last to be affected in this way. In Fig. 15 is illustrated a vertical section through the central region of an embryo in which the schizocoelous splitting is still at an early stage. There are a number of isolated cavities in the mesenchyme, and only in the outer and more lateral regions have they extended so as to link up with neighbouring clefts. Continuation of the process results in the fusion of all the spaces to produce the general body cavity, or perivisceral coelom. A vertical cross-section of one of the incipient arms of a young star, showing the earliest stage in the schizocoelous excavation of the perivisceral coelom, is seen in

Fig. 16.—Vertical section through the rudiment of the arm of a young star in which the coelom is appearing as splits between the cells of the mesenchyme.
Amb.C., ambulacral canal; Coel.Sp., schizocoelous intercellular splits forming the coelom; Ect., ectoderm; Mesench., mesenchyme; Pod., podium.

Fig. 16. Further extension of the splitting, with consequent linking up of the spaces produced, causes the coelom to take on what is essentially its adult relationship with the neighbouring organs. Its lining is at first irregular (Fig. 17), with cells and cytoplasmic bridges projecting into the lumen; but soon a coelomic epithelium is differentiated from the bounding mesenchyme cells, so that the perivisceral coelom takes on a more regular form (Fig. 18).

At this point it is proper to reiterate that one of the chief purposes of this study was to re-examine the statement made by Russo that the general coelom arises in Amphipholis squamata by splitting in mesenchyme (Russo, 1891). It can now be seen that (1) Russo's claim in this regard is completely confirmed, and that he thus is the original discoverer of this mode of origin of the perivisceral coelom in an echinoderm; and (2) the mode of origin of the perivisceral coelom in A. squamata differs but little from what has

been described for Kirk's ophiuroid. It now remains for future research to show whether the coelom can arise by schizocoelous development in any other groups of the echinoderms. Further discussion of the point is reserved for consideration in connection with the problem of direct development among the echinoderms in general in a separate paper (Fell, 1945).

Fig. 17.—Vertical section of an elder embryo showing complete differentiation of the coelom.
Epineur.C., epineural coelom; N.R., nerve-ring; N.Fib., nerve fibres; D.Gang., V.Gang. dorsal and ventral ganglia; Perihaem. perihaemal coelom; Amb.C., ambulacral canal; Amb.R., ambulacral ring; Coel., general coelom; Stom., stomach; Perit., peritoneum; Mes., mesoderm; Ect., ectoderm; Perist.C., peristomial coelom; Phar., pharynx; Buc.T., buccal tentacle; R., radius; I.R., interradius.

One notable difference from Kirk's ophiuroid concerns the relative times of appearance of the mouth opening. Kirk's ophiuroid is already free-living, with a completed mouth opening leading into the stomach before any perivisceral coelom is formed. In A. squamata, on the other hand, when the perivisceral coelom has been formed, the alimentary canal is still very rudimentary, the buccal region being as yet unpierced.

Nervous System.

Russo (1891) claimed to have recognised the nervous system at its earliest appearance as “four yellow cells” said to be attached to the ectoderm and stomodaeum soon after the formation of the latter. The cells were stated by him to be recognisable by their distinctive colour. This observation, according to his description, was made on living embryos without any use of sections. His figures represent the four cells in a highly diagrammatic manner, situated two on either side of the stomodaeum. His account has been quoted by Ludwig

reduced. Thus are formed the rudiments of the ventral ganglion mass of the first segment of the arm in each radius, and the ventral ganglia of the nerve-ring. Immediately dorsal to these ganglion cells there develops from them a zone of fibres which send out slips to the various body tissues. Dorsal again to these fibres there develops a narrow zone of ganglion cells, similar to those of the ventral ganglion, but forming a nervous sheet only one cell deep. Subsequently this dorsal ganglion sheet comes itself to be overlaid by the perihaemal sinuses (see below). Later growth of the arms outward results only in a fuller development, and, with segmentation of the arms, serial reduplication of the regions of the radial nerves already described. In the nerve-ring a similar differentiation into ventral ganglia, intermediate fibre-zone, and dorsal ganglion sheet also occurs.

The staining reactions of the differentiating nerves are very characteristic, especially when stained with Mallory Triple stain. Before differentiation all nervous elements are evenly stained with acid Fuchsin (cytoplasm) and Orange G (nuclei). Once the three main zones are laid down, the nuclei enlarge and become more prominent, taking up the orange dye more deeply. Most stains do not affect the fibre-zone, but Aniline Blue, which is present in Mallory, gives it a pale greyish-blue tinge, thereby making it contrast strongly with the strongly nucleated ganglionic tissue on either side.

The probable ectodermal origin of the nerve rudiments from an outer plastic region of the stomodaeum, even though it is deeply imbedded in mesenchyme, is of interest, the more so now that it is realised that nerve cells do not necessarily arise from ectoderm.

Peristomial Coelom, Perihaemal Sinuses and Epineural Sinuses.

The regions of the coelom so far described are the hydrocoel and the perivisceral coelom. The differentiation of the remaining parts can now be dealt with. While the process of splitting which gave rise to the perivisceral coelom was occurring, a portion of the schizo-coel thus formed remained distinct from the general coelom. This is the rudiment of the peristomial coelom (or peripharyngeal coelom), which is derived from a series of small splits about the most ventral limb of the stomach, dorsal and lateral to the oesophageal mass (Figs. 17 and 18). As in the case of the perivisceral coelom, it has at first a very irregular form, with its bounding cells projecting into the lumen; but subsequently a lining epithelium is differentiated.

The perihaemal and epineural sinuses are the last coelomic structures to appear, and they arise simultaneously, or almost so. As in the case of Kirk's ophiuroid (Fell, 1941), they form in a very simple manner, as splits above and below the nerve-cords and nerve-ring, never achieving any great degree of differentiation. In much older forms a fairly well-marked lining epithelium is developed in the case of the perihaemal sinuses above the nerve-ring and radial nerve-cords; but so far as I have observed in my sections of adult individuals, the epineural sinuses remain throughout life as little more than intercellular splits without epithelia. Indeed, as Delage and Herouard (1903) have pointed out, the whole epineural system is considerably reduced in Amphipholis squamata. The fact is

perhaps connected with the small size of the species, and is to be correlated with the simplification of the reproductive glands and other organs.

It is an interesting fact that the perihaemal and epineural sinuses, though comparatively unimportant subdivisions of the coelom, were the structures for which a schizocoelous development was first recorded. The discoverer was Hamann, in 1889, whose results were afterwards quoted in Bronn's Thier-Reichs (1901). Dawydoff (1901) also found that the epineural sinuses have a schizo-coelous origin in the arm of an ophiuroid which is regenerating lost radii. Yet, despite the evidence produced by these workers, Delage and Hérouard (1903) rejected Hamann's account of the schizocoelous origin of the structures, and considered Dawydoff's results of little importance, though they produced no evidence of their own to the contrary. The eminence which Dawydoff has since attained in the realm of invertebrate embryology is perhaps the most suitable reply to the dogmatic attitude which has unwarrantably been adopted in the instance quoted. MacBride (1914) similarly rejected as “improbable in the highest degree” the notion of a schizocoelous development of the coelom in echinoderms, though, like Delage and Hérouard, he produced no evidence of personal research to confirm or disprove Hamann and Dawydoff. Thus the matter came to be almost forgotten, and no mention is to be found of a schizocoelous origin of the echinoderm coelom in current text-books on embryology*.^*

Atrophy of the Embryonic Attachment.

At about this stage in development, when the outline of the embryo is roughly a pentagon, the embryonic attachment becomes completely atrophied, and the embryo thus comes to lie freely in the bursa. From this stage onward the embryo is usually found oriented with its ventral (oral) surface uppermost. The significance, if there be any, of this orientation is not clear, but it is noteworthy that Sladen (1889) quotes Wyville Thomson as having observed a similar condition in the Antarctic asteroid Leptoptychaster kerguelensis. Little is known of the embryology of this species, but Thomson records that the young stars develop within the oviducts, with their oral faces turned uppermost. After birth they adhere for a while in the re-entrant angles between the rays of the parent, still with the cral surface upward.

Oesophageal and buccal Cavities.

When the original oesophageal sac became transformed into a mass of undifferentiated cells, the change not only paved the way to the development of the nervous system, but also initiated a further step in the differentiation of the alimentary canal. Hitherto the stomach has remained a blind sac, having at no stage any anal opening, nor any connection with the exterior via the oesophageal sac. There now occur two convergent cone-shaped excavations into the oesophageal mass. One of these occurs upwards from the outer

ectoderm of the mid-ventral surface. As will be seen from the sections shown in Figs. 17 and 18, there has already existed here for some time a hollowing out which is obviously the forerunner of a buccal cavity. Figs. 19A and 19B illustrate the final stages in the process, when the second excavation occurs, downwards from the central-most part of the stomach. In this way there forms first a blind pouch from the stomach towards the buccal cavity (Fig. 19A), and then finally the oesophageal passage is pierced, and for the first time the alimentary canal is brought into being (Fig. 19B).

Figs. 19a (above) and 19b (below).—Later stages in the piercing of the mouth.
Amb.C., ambulacral canal; N.R., nerve-ring; N.Rad., radial nerve; Per.C., peristomial coelom; Phar., pharynx; Coel., coelom; Stom., stomach.

Later Changes.

From now onward the growth of the embryo is closely similar to what has already been described for Kirk's ophiuroid (1941). Soon after passing the “Asterina” stage, the arms begin to show the first signs of segmentation. The ambulacral system extends by outward growth into the arms. There arise the first set of podia, then the buccal tentacles between the first podia and the mouth, and then the second and third, etc., sets of podia are developed centrifugally.

These later changes have been described for Kirk's ophiuroid, and thus need not be repeated here in detail. The skeletal system develops within the mesenchymatous tissues in the disc and arms in the way already described by Ludwig (1881), Fewkes (1887), and Fell. (1941). As in Kirk's ophiuroid the madrepore is situated on an oral shield, at first near the periphery of the disc, but afterwards carried downwards on to the ventral surface when the oral shields take their place among the mouth ossicles. There remains thus only to describe the birth of the young ophiuroid.

Fig. 20.—Ventral view of an attached “Pentagon.”
Buc.Te., buccal tentacle; Mad.C., madreporic canal; Madp., madrepore; Amb.C., ambulacral canal; Pod.1, first podium; Mes., mesoderm mass; Jaw, developing jaw; Amb.R., ambulacral ring; Emb.At., embryonic attachment.

Birth of the Young Ophiuroid.

As the arms increase in length they are folded above the disc, twisted in a spiral fashion. The mode of folding of the arms causes the young animal to have a roughly spherical form, thus occupying the smallest possible amount of space within the body of the parent. As mentioned before, the embryo usually lies in an inverted position, with its ventral surface turned towards the dorsal aspect of the parent. In Fig. 1 is shown a dissection of an interradius of a pregnant adult, with an embryo oriented in this manner within the bursa.

The actual process of birth was observed on one occasion. The embryo uncoiled its arms so that they became directed downwards to the genital cleft. Then, solely by its own muscular efforts, the young animal crept out of the bursa in such a manner that it emerged with its arms directed away from the nearby arm of its parent. The process occupied over three hours, but may have been delayed by the unfavourably bright conditions on the microscope stage (strong

light, as recorded in a previous paper, having a narcotic effect on young ophiuroids). When free from the parent, the young ophiuroid, which had about fifteen arm segments, fell to the bottom and commenced an independent existence. The young forms are at first colourless, except for the pinkish disc, but after ten days or so a greyish pigmentation begins to appear. The newly-born ophiuroids, and also those which have been free-living for some time after birth, make great use of the podia in progression, these organs having in early life a true tube-foot function. Thus young ophiuroids are enabled to climb vertical glass surfaces, and to a certain limited extent the capacity is retained by the adult. These facts go to confirm the observations of Smith (1937), and also of some earlier workers, on the function of the tube-feet of ophiuroids. It is still commonly stated in text-books that the podia of ophiuroids have only a sensory function—with possible use as respiratory organs; such statements are based on inadequate observation.

Summary.

References.

Apostolides, N. C., 1882. Anatomie et Developpement des Ophiures. Arch. Zool. Expl., 10.

Balfour, F. M., 1901. Echinodermata (Introduction). Zool. Reo. (for year 1900).

Bernasconi, Irene, 1928. Una Ofiura Vivapara de Neochea. An. del Mus. Nac. Hist. Nat., Buenos Aires, 34, p. 145.

Bogucki, M., 1930. A propos de la prétendue action inhibitrice due liquide coelomique sur la membranogénèse et sur la segmentation des oeufs d'oursin. Protoplasma, 11.

Boveri, Th., 1901. Die Polaritaet von Ovocyte, Ei und Larve des Strongylo-centrotus lividus. Zool. Jahrb. Anat. Ont., 14.

Buchner, P., 1911. Ueber Hermaphrodite Seesterne. Zool. Anzeiger, Bd. 38.

Carpenter, P. H., 1887. On the development of the apical plates in Amphiura squamata. Quart. Journ. Mic. Sc., 27.

Cuenot, L., 1891. Etudes Morphologiques sur les Echinodermes. Arch. de Biol., Beneden, t. 11.

— 1898. L'hermaphroditisme protandrique d'Asterina gibbosa et ses variations suivant les localites. Zool. Anzeiger, Bd. 21, p. 273.

Dawydoff, C., 1901. Beitrage zur Kentniss der Regenerationserscheinungen bei den Ophiuren. Zeit. f. Wiss. Zool., 69, 2.

— 1928. Traité d'Embryologie Comparée, Paris (Masson).

Delage, Y., 1902. Quelques Experiences et Observations sur les Astéries. Arch. Zool. Exp. et Gen., t. 10.

— and Herouard, E., 1903. Les Echinodermes. Traité de Zoologie Concréte, t. 3, Paris.

Delle Chiaje, S., 1828. Memorie sulla storia e notomia degli animali senza vertebre del regno Napoli, 3, p. 77.

Drzewina, A., and Bohn, G., 1924. Un nouveau cas d'hermaphroditisme chez l'oursin Strongylocentrotus lividus. Comp. Rend. Acad. Sc. Paris, t. 178.

Fell, H. Barraclough, 1940a. Origin of the Vertebrate Coelom. Nature, 145, p. 906.

— 1940b. Culture in vitro of the excised embryo of an ophiuroid. Nature, 146, p. 173.

— 1941. The Direct Development of a New Zealand Ophiuroid. Quart. Journ. Mic. Sc., 82, 3.

— 1945. A Revision of the Current Theory of Echinoderm Embryology. Trans. Royal Soc. N.Z., 75, 2.

Fewkes, J. W., 1887. On the Development of the Calcareous Plates of Amphiura squamata. Bull. Mus. Comp. Anat. Harvard, 13.

Forbes, Edward, 1881. A History of British Starfishes, London.

Gadd, G., 1907. Ein Fall von Hermaphroditismus bei dem Strongylocentrotus droebachiensis? O. F. Muel. Zool. Anz., Bd. 31, p. 635.

Garstang, W., 1931. Plymouth Marine Fauna, 2nd. Ed. (record of breeding of Amphipholis squamata).

Grave, C., 1900. Ophiura brevispina. Mem. Biol. Lab. Johns Hopkins, 4, 5.

— 1914. A solution of the problem of yolk manipulation by Ophiura. Science, New York.

— 1916. Ophiura brevispina, II. Journ. Morph. Philad., 27.

Gray, J., 1921. Note on the true and apparent hermaphroditism in Sea Urchins. Proc. Camb. Phil. Soc., 20, 4.

Gross, F., 1937. Notes on the Culture of some Marine Plankton Organisms. Journ. Mar. Biol. Assn., 21, 2.

Hamann, Otto, 1889. Anatomie und Histologie der Ophiuren und Crinoiden. Jen. Zeitsch. Naturw., Bd. 23.

Harvey, L. A., 1931. Studies in Echinoderm Oogenesis. Proc. Roy. Soc., B, 107.

Heilbrunn, L. V., 1929. Hermaphroditism in Arbacia. Science, New York, 69, p. 1790.

Herlant, M., 1918. Un cas d'hermaphroditisme complet et fonctionnel chez Paracentrotus lividus. Arch. de Zool., t. 57, p. 28.

Huxley, J. S., 1928. Experimentally induced metamorphosis in Echinus. Amer. Nat., 62, p. 363.

— 1928. Artificially induced metamorphosis in the Echinoderms. Nature, 121.