A Revision of the Current Theory of Echinoderm Embryology.
(Communicated by Professor L. R. Richardson.)
[Read before Wellington Branch, May 2, 1945; received by the Editor, May S, 1945; issued separately, September, 1945.]
|Material for Evidence||74|
|What is the “Typical” Echinoderm Development?||74|
|Phases of Development Analysed—|
|1. The Ovum||78|
|2. The Early Cleavages||82|
|3. The Blastula||84|
|4. Formation of the Gastrula||86|
|5. Critical Point in Development||88|
|6. Disappearance of the Larva from Ontogeny||88|
|7. Vitellaria, or “Yolk-Larva” Series||92|
|8. Ontogeny of the Coelom||93|
|9. Later Stages of Development||97|
|10. Viviparity as a Factor Producing Direct Development||97|
|11. Mechanism Which has Brought About Direct Development||98|
In October, 1846, Johannes Müller astonished the scientific world of his day by his discovery that a peculiar free-swimming marine animal which he had previously named Pluteus paradoxus was in reality the larval stage of an ophiuroid. This was the first demonstration that echinoderms undergo an indirect process of development, and the remarkable fact emerging from the observation was that these bottom-dwelling, radially symmetrical animals pass through a stage in their life-history in which they are free-swimming and bilaterally symmetrical. Müller's discovery, and his later researches in the development of the echinoderm groups, were destined afterwards to exert a great influence upon theories as to the origin and evolution of the echinoderms; but their immediate effect was mainly upon the science of embryology, stimulating research workers to identify and elucidate other free-swimming larval forms.
The subsequent history of echinoderm embryology is too complex to be capable of summary here, but in its main lines it can be expressed in a single statement. It has resolved itself into the simple and dogmatic doctrine that all typical echinoderms pass through a free-swimming larval stage. There are various adjuncts to this doctrine, the most notable being the widely accepted statement that the coelom of echinoderms is always an enterocoel. This paper is concerned with a critical examination of the evidence upon which the doctrine is based.**
[Footnote] ** Assistance in the cost of printing this paper was given from the Publications Fund of Victoria University College.
The problem under consideration may be briefly enunciated— “To what extent does direct development occur in echinoderms; what are its causes; by what processes does it take place; and what evidence does it provide of the evolutionary history of the echinoderms?”
In view of the more extensive nature of our knowledge of the subject in the Ophiuroidea, the present paper is predominantly concerned with that class, but wherever material is available comparisons are made with conditions holding in the other classes.
Material for Evidence.
The evidence upon which this investigation is based is as follows:—
(a) Our previous knowledge of the indirect, pelagic type of development. This is founded mainly on the researches of MacBride (1907) on Ophiothrix fragilis and the recent work of Narasimhamurti (1933) on Ophiocomina nigra.
(b) Our previous knowledge of the development of a moderately yolky ophiuroid egg. This is founded solely on Grave's work on Ophiura, brevispina (1900 and 1916).
(c) The recently published work of the writer on Kirk's ophiuroid, a form with very yolky eggs, and on Ophiomyxa brevirima, a viviparous form also with very yolky eggs (1940 a, 1941).
(d) The scattered body of data in the literature in regard to isolated stages of development of other ophiuroids—mainly referring to larval forms.
(e) The research presented elsewhere (Fell, 1945) on Amphipholis squamata, a form with moderately yolky eggs and of viviparous habit.
It will be seen from the above list that there is a fairly representative body of data available, ranging from non-yolky eggs to those with a very large proportion of yolk, and including cases of viviparity. In addition to these data, there is a considerable amount of scattered information on the other echinoderm classes, but in no ease presenting so complete a sequence as is now available from the Ophiuroidea.
The method adopted in this investigation is comparative and analytical—i.e., each of the principle phases of development is taken separately, and examined with reference to the forms showing indirect development, those showing intermediate characters, and those showing direct development.
What is the “Typical” Echinoderm Development?
Before commencing the study of the nature of direct development in echinoderms, it will be useful to examine the position in regard to development in general in this group. As pointed out in the introductory section of this paper, it has come to be generally assumed that all “typical” echinoderms have an indirect development; conseqently, whenever an echinoderm has been found which departs from this criterion, it has been glossed over in the text-
books with a remark to the effect that it is “not typical”—and hence unworthy of further consideration.
Examples of this attitude are abundant. Thus, commenting on the development of Antedon, MacBride (1914) states, “We may discount the characteristic features of the development of Antedon, which are obviously due to the yolky egg …”, etc. Antedon, it should be noted, is the only crinoid of which we have any embryological knowledge, yet, because it lacks a larva comparable to the Ophiopluteus or Bipinnaria it is automatically rejected as “atypical.” for, it is stated, a “typical” crinoid must undoubtedly possess a bilaterally symmetrical and non-yolky larva.
If echinoderms with yolky eggs—and consequently having direct development to a greater or less degree—are held to be atypical, then obviously a census of the echinoderms of whose development we have any evidence should reveal at once that directly developing forms constitute the great minority. It should be found that most echinoderms have an indirect development, coupled with a pronounced larval stage. Is this the true position? In order to obtain some information on this question I have selected three specimen faunas for special examination. These are the echinoderm faunas of Great Britain, of New Zealand, and of the Antarctic and Sub-Antarctic. In these we may expect to find reasonably typical samples of the world's echinoderms.
A survey of the literature shows that there are some 61 British echinoderms about which we have sufficient information to be able to deduce with a fair degree of certainty the type of development which they follow. This is a larger proportion of the entire fauna than holds in other regions, and is attributable to the fact that most British species are found over a large part of the northern European seas, where most research work on the subject has been done. Taking first those species whose development is either known to be shortened, or which may be regarded as fairly certain to be so, we have the following list:—
|Mesothuria intestinalis||Psolus phantapus|
|Bathyplotes natans||Leptosynapta minuta|
|Cucumaria frondosa||Labidoplax buski|
|Crinoidea. (Two species.)|
|Antedon bifida||Antedon petasus|
|Asteroidea. (Twelve species.)|
|Ctenodiscus crispatus||Solaster papposus|
|Psilaster andromeda||Solaster endeca|
|Pontaster tenuispinus||Henricia sanguinolente|
|Ceramaster granulatus||Brisingella coronata|
|Hippasterias phrygiana||Leptasterias mulleri|
|Asterina gibbosa||Pedicellaster typious|
|Ophiuroidea. (Four species.)|
|Asteronyx loveni||Ophiopus arcticus|
|Ophiomitrella clavigera||Amphipholis squamata|
|Echinoidea. (Two species.)|
|Poriocidaris purpurata||Neolampas rostellata|
Note.—The larvae of Amphiura, filiformis and of Ophiura affinis, although pelagic, lack the full development, and may be regarded as representing stages in the adoption of a shortened development, in addition to the four ophiuroids named above.
Taking now those species whose development is either known or believed to be of the indirect or pelagic type, we have the following:—
|Stichopus tremulus||Labidoplax digitata|
|Note.—Only one Auricularian larva has been taken in British waters—that of L. digitata.|
|Asteroidea. (Seven species.)|
|Astropecten irregularis||Asterias rubens|
|Luidia sarsi||Marthasterias glacialis|
|Porania pulvillus||Pedicellaster sex-radiatus|
|Ophiuroidea. (Twelve species.)|
|Ophiothrix fragilis||Ophiura texturata|
|Ophiocomina nigra||Ophiura sarsi|
|Ophiactis balli||Ophiura albida|
|Ophiopholis aculeata||Ophiura carnea|
|Amphiura chiajei||Ophiura affinis|
|Amphiura filiformis||Ophiocten sericeum|
|Echinoidea. (Twelve species.)|
|Cidaris cidaris||Strongylocentrotus droebachiensis|
|Psammeohinus miliaris||Echinocyamus pusillus|
|Echinus esoulentus||Brisaster fragilis|
|Echinus acutus||Spatangus purpureus|
|Paracentrotus lividus||Echinocardium cordatum|
|Sphaerechinus granularis||Brissopsis lyrifera|
Of ten holothurians about whose development we have any knowledge, only three form Auriculariae. The other seven species all have large yolky eggs and a more or less direct development. Two of these are known to have the special simplified “yolk-larva” which I have termed “Vitellaria” (see a later section of this paper) and the remainder may have absolutely direct devlopment.
Of the two crinoids whose development is known, both have large yolky eggs, and both form the same yolk-larva or Vitellaria.
Among the asteroids, only seven out of nineteen species have a pelagic development. The remaining twelve species all have large yolky eggs and probably direct development to a greater or less extent. In this class there is as yet no evidence of the existence of the “Vitellaria” larva.
In the ophiuroids, of the sixteen species about which there is any evidence, four have direct development as a result of viviparity or large yolky eggs. Of the remaining twelve species, not all have typical plutei, and stages in the loss of the arms are to be observed. In a species of Ophiura a Vitellaria is known.
In the echinoids there is far less tendency to have direct development. However, of fourteen species about which we have any informa-
tion two at least have large, yolky eggs and probable direct development. This development is likely to be of the type seen in Toxocidaris erythrogrammus of New South Wales. The remaining twelve species have pelagic plutei and indirect development.
Summarising these results, we find that in those British echinp-derms whose development is known either in whole or in part, or about which there is any evidence, the following table may be drawn up:—
|Group.||P.c. With Direct
|P.c. With Pelagic Un-
From this it is clear that the notion that echinoderms are typified by having an indirect development with pelagic larvae is without foundation. The theory probably originated from the fact that most laboratory work has been done upon the eggs and larvae of echinoids, which, as shown above, are alone notable for tending to have pelagic larvae.
New Zealand Echinoderms.
Turning now to the New Zealand echinoderm fauna, I select, the ophiuroids for particular mention, as I have had opportunities for personal study of these forms in recent years.
Mortensen showed in 1924 that at least six New Zealand ophiuroids are viviparouis. All of these lack a free-swimming larva and have a more or less direct development. In the case of Ophiomyxa brevirima direct evidence has been produced that this is so (Fell, 1940 a). In addition it has been shown that the oviparous form known as “Kirk's ophiuroid” has an absolutely direct development. I have also described the large, yolky egg of Pectinura maculata which indicates a direct development (1941). Other species provide some indication as to the nature of their development from the size of their eggs (see section of this paper on the egg).
The following ophiuroids are known, or are believed to have direct development:—
|Ophiomyxa brevirima||Pectinura gracilis|
|Amphiura magellanica||Pectinura maculata|
|Amphiura annulifera||Amphiocnida pilosa|
|Amphiura praefecta||Amphioplus basilicus|
|Amphipholis squamata||Kirk's ophiuroid|
|Pectinura cylindrica||Ophiozonoida picta|
In addition Mortensen (1924) has given reasons for believing that Amphiura hinemoae may have direct or shortened development.
New Zealand ophiuroids whose development is known or believed to be of the indirect pelagic type are:—
|Ophiocoma bollonsi||Amphiura amokurae|
|Ophiactis resiliens||Amphiura alba|
|Amphiura spinipes||Ophionereis fasciata|
The development of the remaining twenty-three species is unknown.
The above tables show that out of nineteen ophiuroids of New Zealand of whose development we have any knowledge at all, twelve—i.e., 63%, are believed to have direct development, and only seven, or 37%, to have indirect pelagic development. Thus, the proportion of ophiuroids believed to have direct development is much higher in New Zealand waters than in those of Britain (and Europe).
Antarctic and sub-Antarctic Ophiuroids.
Owing to the painstaking work of Mortensen (1936) in his description of the ophiuroids collected by the Discovery Expedition, we now have some knowledge of the breeding habits of many southern species, together with information in regard to the eggs. His researches have revealed that out of 56 ophiuroids from this region whose sexual character is known, no less than 31 are viviparous. This high proportion seems to be characteristic of southern seas, as it is not paralleled in Arctic forms.
These 31 ophiuroids must all have more or less shortened developments, without free larval stages. Details have been given above of species from Britain and New Zealand, and there is no occasion to quote any further lists of species here. The reader is referred to Mortensen's paper (1936). However, besides the 31 viviparous forms it appears that some eight species at least have large, yolky eggs and therefore probable direct development. Thus out of the 56 Antarctic species of whose development we have any information, it is highly probable that 39 species, or 70%, have direct development. Of the remainder, only in nine species has it been shown that there is likely to be a pelagic larval stage. Our knowledge of these southern forms must of necessity remain incomplete for a long time, but sufficient is known to show that the majority of them are likely to have direct development.
No doubt similar evidence of this kind could be provided for the echinoderm faunas of the rest of the world—with the possible exception of the tropical regions, where larval forms of all kinds tend to become exaggerated—but sufficient facts have been given to show that the supposition that a pelagic larval development is “typical” of echinoderms is founded on no evidence whatever. The fact is that no particular kind of development is “typical” for echinoderms in general, but that the kind of development followed depends on the particular conditions obtaining in each species. That these conditions can be analysed and classified will be shown in this paper.
The Phases of Development Analysed.
1. The Ovum.
In the characters of the ovum I believe that we have the clue to the factors controlling development in both the directly developing and indirectly developing forms. These factors can, I believe, be further traced to the cytoplasm of the ovum. But before proceeding to examine the evidence provided by the internal structure of the egg-cell, it will be useful to consider the egg-cell phase as a whole.
|The Inter-relationship Between Egg-size and Mode of Development in
|Group 1.—Small eggs, long indirect
development, with pelagic Ophioplutei.
|From 100 μ
|Group 2.—Eggs of intermediate size,
having a shortened development, and
with reduced larvae.
to 650 μ
|Group 3.—Eggs comparatively very
large, with development so shortened as
to be direct, without trace of a larval
If we take a series of ophiuroids ranging from forms with indirect, larval development to those with direct, non-larval development, and place their eggs in order of increasing volume, a remarkable fact at once becomes obvious (see Table II). It is found that the eggs of small volume fall into a group the common character of which is the possession of a long period of indirect development involving a pelagic Ophiopluteus stage and metamorphosis. The largest eggs form another group whose common character is the absence of indirect development, the loss of the larval stage, and no metamorphosis. Eggs of intermediate size form a third category exhibiting intermediate stages in the suppression of the larval development. This we can express as a simple and fundamental principle—“The degree of larval development of an embryo varies indirectly as the volume of the egg from which the embryo is derived.” Grave (1916) anticipated this “law,” but as no cases of completely direct development were known at that time, and the development of only two ophiuroids had been worked out, he was unable to express the fact in a full or convincing manner. His work will be further considered below.
The same relationship between egg-size and development appears to hold good in the other four classes of living echinoderms, but in no case is there as yet so complete a series as is now known from the Ophiuroidea.
In my previous paper on the development of Kirk's ophiuroid (1941) it was suggested that the immediate ontogenetic factor causing the direct development and unexpected method of development of the organs might be the presence of yolk material in the tissues of the developing embryo. The yolk material, it was suggested, might act as a retarding agent, being inert and lifeless. This interpretation of the effect of yolk material accords well with the facts revealed by the sequence of egg types shown in Table II, for in all the cases it is the larger eggs—and consequently those with the greater amount of nutritive material—which have direct development.
In the literature it has been customary to speak of eggs as being “yolky” or “very yolky” or “not yolky”—a method of description which is necessarily inexact. It seemed to the writer that a more precise definition of these arbitrary terms would be of use in determining more accurately the influence exerted by the yolk. Accordingly the following method of quantitative estimation of yolk content in minute eggs was devised, and proved susceptible to mathematical treatment. Sections of known thickness were cut through the egg of an ophiuroid whose yolk content it was desired to estimate. Haematoxylin was used to stain the yolk granules deep black. Then, by using a squared eyepiece the average number of granules in an area of known size was f oxind. Knowing the thickness of the section, it was then possible to calculate the number of granules in a given volume of egg. Next the average diameter of the yolk granules was estimated, and, combining this result with the former, it was possible to estimate approximately the actual volume of yolk material in a given volume of egg. In practice the yolk was estimated as a percentage of 100 cubic microns of egg material. Knowing the average diameter of the egg, it was then possible to determine the absolute volume of yolk material present in the entire egg.
A calculation made in this way shows that in Kirk's ophiuroid there exists in the egg approximately 5.3 × 106 cubic microns of yolk material.
A similar estimation for Amphipholis squamata, a form with only “moderately yolky” eggs, yields a figure of circ. 8.8 × 104 cub. μ. Thus we have a much more clear-cut picture of the relative “yolkiness” of these two types of egg, and can now proceed to compare the yolk value with the other variables.
Tabulating the yolk value against the diameter of the egg, as in Table III, it is seen that in these examples yolk content increases with size of egg. This, of course, was merely what has usually been assumed hitherto, though without actual demonstration.
|Diameter of egg.||100μ||500μ|
|Volume of yolk in egg.||8.8 × 104 cub.μ||5.3 × 106 cub. μ|
|Cytopiasm expressed as
percentage of egg
Having noted that the yolk increases with increasing diameter of the egg, we can now consider the case in regard to the cytoplasm. In this matter the literature provides no information. The tendency, however, has been to suppose that in large, yolky eggs the cytoplasm becomes proportionately reduced, remaining more or less constant in amount while the yolk increases. This attitude is maintained by Grave (1916) who also states that it is impossible to measure the ratio of yolk to cytoplasm. In his latter statement he is, of course,
incorrect, since it is possible to make such, measurements by the method described above. In the calculation illustrated, at the stage when the yolk-volume per 100 cubic microns was estimated, the resultant is also automatically an expression of the percentage content of yolk of the whole egg. By a simple subtraction, therefore, we obtain an expression of the percentage content of cytoplasm—ignoring the nucleus in both measurements. Taking now the cytoplasm value as a percentage of egg-volume, it is clear from Table III that comparatively little change in the ratio of cytoplasm to yolk has taken place in the transition from smaller egg-size to larger. This means that as the amount of yolk has increased, so also has the amount of cytoplasm, keeping pace each with the other. The older vague methods of describing “yolky” and “less yolky” eggs failed to bring out this fact, and hence the mistaken notion that the yolk increased and not the cytoplasm. It is therefore necessary to modify the conclusion expressed in my earlier paper on Kirk's ophiurpid (1941) in regard to the influence of increasing yolk-mass, by adding that the effect is associated with a parallel increase in the cytoplasm.
Finally, if we combine into one diagrammatic graph the three associated conditions we have been considering in this section of the paper—namely, increasing egg size, increasing yolk content, and shortening of the development—we obtain a convincing demonstration of how the different types of development are interrelated (see Fig. 1). This graph shows that at one end of the scale we have such types as Ophiothrix fragilis and Ophiocomina nigra, with small egg's and small amount of yolk material. Next comes the group with an increase in the amount of yolk, together with a shortening in the development. This group includes Amphipholis squamata and Ophiura brevispina. In the case of the latter species, the increase in yolk is associated with increase in egg-size, together with more marked shortening of the development. At the other end of the scale comes the group in which a very much greater increase in yolk material has occurred, together with cytoplasm increase, the eggs have become progressively larger, and the development so shortened as to be direct. In this group are such species as Ophiopus arcticus, Kirk's ophiuroid, Ophiomyxa brevirima and (probably) Pectinura maculata.
From these facts it is reasonable to conclude that the factor most important in causing direct development is intimately related with an increase in the size of the egg and an associated increase in the amounts of cytoplasm, and its product—the yolk.
Whether other factors may have operated in bringing about direct development will be considered later in the paper, as also will the mechanism through which the factors may have acted.
2. The Early Cleavages.
In ophiuroids with small, comparatively non-yolky eggs the plan of cleavage followed is of the usual echinoid type—i.e., the first cleavage is equal, as also is the second, and both are vertical. The third cleavage is horizontal, and the upper and lower quartets are equal. The fourth cleavage is again vertical and more or less equal, and after that the cleavages occur fairly evenly over the whole embryo to produce a morula. The latter rapidly gives place to a regular blastula. These stages are represented in Figs. 2 to 6.
In the case of the moderately yolky eggs of ophiuroids, Grave (1916) has described the cleavages of the egg of Ophiura brevispina. His account shows that no difference is to be observed from the plan followed in the non-yolky group. He concludes, therefore, that the early stages of development in this species have been disturbed little if any by the increase in the yolk mass that has occurred.
Coming now to the yolky group having very large eggs, we have such forms as Ophiomyxa brevirima and Kirk's ophiuroid. The type of cleavage followed here has been described by the writer (1941) in the latter species. Cleavage in this case follows a somewhat different pattern to that seen in the two preceding groups, but the alteration nevertheless is not very profound. The first cleavage (Fig. 7) is unaltered.* The same applies to the second cleavage
[Footnote] * In about 30% of cases, however, an abnormal cleavage occurred in which the first two blastomeres were unequal, the larger undergoing two divisions to form a normal four-cell stage.
Figs. 2–11.—Diagrams illustrating the effect of increasing yolk mass on the cleavage of the egg.
Figs. 2–6.—Cleavage in ophiuroids with small or moderate yolk mass (e.g., Ophiothrix, Ophiocomina, Ophiura).
Figs. 7–11.—Corresponding stages in cleavage of Kirk's ophiuroid, which has heavily yolked eggs.
(see Fig. 8). When, however, the third cleavage is reached, two quartets are formed which show a marked differentiation into micro-meres and macromeres (Fig. 9). This fact indicates that a prelocalisation must have occurred during the four-cell stage, or even earlier. As shown in my paper (1941), the macromeres are destined to form mes-endoderm, and are ventral (aboral) in position. The micromeres form the ectoderm and its derivatives, and in addition contribute to the mes-endoderm. This will be further discussed in the section of this paper dealing with gastrulation. In the cleavage immediately following the third, the division of the blastomeres becomes irregular and “out of step”. Finally, a morula with very turgid blastomeres is formed (see Fig. 11).
The changes to be observed in the plan of cleavage of this very yolky form are undoubtedly to be correlated with the great increase in the yolk material. It is interesting to note that, despite the comparatively enormous quantity of yolk present in Kirk's ophiuroid as compared with Ophiothrix or Ophiocomina, the plan of segmentation is really but slightly altered. One might, for instance, have expected to find some tendency to adopt partial cleavage, or to form a blastoderm, as has occurred in other groups where increase in the
yolk mass has occurred. Indeed, such a supposition was put forward by Dr. Th. Mortensen in a private communication to the writer; but so far no evidence of such alteration in development has been found.
It would appear that the reason for this retention of the primitive type of division may be correlated with the fact already shown in this paper that the cytoplasm mass has increased in step with the increase in the yolk mass. Again, we may consider that the earliest stages of development are always the most rigid, and the last to be affected by factors producing ontogenetic change. This seems to be indisputable, for, as will be shown below, the later stages come to be tremendously altered in the heavily yolked forms.
One further effect of the increased yolk-mass in Kirk's ophiuroid remains to be noted. This concerns the alteration in the rate of cleavage. In my account of the development of this species, a table was given showing the progress of cleavage during the first twenty-three hours following the first cleavage. From this it was clear that a pronounced retardation had occurred, for whereas other ophiuroids with non-yolky or moderately yolky eggs reach the blastula stage at the end of the first 24 hours, Kirk's form is at that time only at the morula stage. Not till the end of the second day is the blastula completely formed, while the gastrula stage cannot be regarded as complete till the fifth day. On the other hand, other ophiuroids (Ophiothrix and Ophiura) have formed the gastrula at the thirty-sixth hour. Ophiocomina, as described by Narasimhamurti (1933) is slightly slower, forming the gastrula at from the fortieth to the forty-eighth hour.
This retardation is illustrated in graphical form in Fig. 12. As the alteration in developmental rate is associated with increase in the yolk-mass, we can justifiably correlate the two conditions. That the presence of a large quantity of inert nutritive material should have such a delaying action on the ontogenetic processes is not surprising.
3. The Blastula.
Passing now to the blastula stage, we begin to find that changes in the developmental sequence become more pronounced, and furthermore, these changes not only affect the heavily yolked eggs, but become extended to the moderately yolked type.
In the non-yolky type, exemplified by Ophiothrix, MacBride (1907) states that there is a free-swimming blastula with ciliated cells. These form a cell-layer one-deep and surround the large, central blastocoel. From the vegetal pole mesenchyme cells are budded off (see Fig. 13).
Figs. 13–15.—Diagiams illustrating changes produced in the blastula by increasing yolk mass. Fig. 13, Ophiothrix, a non-yolky type (after Mac-Bride, 1907); Fig. 14, Ophiura, a moderately yolked type (after Grave, 1916); Fig. 15, Kirk's ophiuroid, a heavily yolked type (Fell, 1941).
In the moderately yolky type Ophiura, as described by Grave (1916), shows a thick-walled blastula with somewhat reduced blastocoel. The disposition of the nuclei shows that there is a tendency to form a blastula wall more than one cell deep—and in the later gastrula stage this tendency receives its full expression according to the earlier paper of Grave (1900) where he refers to the wall of the gastrula being more than one-cell deep. From the vegetal pole the mesenchyme in a reduced form bulges into the blastocoel, reducing its cross-section to a major segment of a circle.
Finally, in the heavily-yolked type we see in Kirk's ophiuroid the tendencies already weakly expressed in Ophiura reaching their fullest extent. The wall of the blastula is several cells thick. The blastocoel is reduced to a small cavity crescentic in vertical section owing to the bulging upward of the macromeres—obviously homologous with the mesenchyme producing cells of the other two forms.
The macromeres and micromeres occupy respectively the vegetal and animal poles as in the morula.
From this sequence we observe that with increasing yolk-mass the walls of the blastula become successively thicker, steadily reducing the blastocoel to a mere vestige in the animal hemisphere. The mesenchyme fails to separate as such but remains as a great bulging mass projecting upward into the blastocoel. As we see in the development that follows immediately upon this stage, the reduction of the blastocoel has a profound effect upon the process of gastrulation
4. Formation of the Gastrula.
As the ontogenetic process advances the modifications produced in the larger and more yolky forms become ever greater. We thus find that the classical concept of evagination in the echinoderms becomes inadequate; for the evagination doctrine presupposes a static organogeny, whereas in actual fact organogeny is as dynamic and susceptible to moulding influences as any other biological process. Evolution, in other words, can act upon early stages of development as upon later stages. Embryological processes can be altered in the same way as can the adult products of these processes. This theory of the ability of evolution to act upon embryonic forms—so convincingly set out in de Beer's essay on “Embryology and Evolution” (1930)—receives strong support from the whole of the evidence provided by the yolky-egged ophiuroids.
Thus it is that in considering together the various types of gastrulation process met with in the Ophiuroidea we are able to detect the operation of a unidirectional evolutionary force, the strength of the force operating in direct proportion to the size of the yolk-mass. whereas the doctrine of unchangeable recapitulation during development becomes meaningless in the light of such a process as is observable here, the view which envisages embryonic forms as essentially plastic and subject to a greater or less degree of modification—according to circumstance—provides, in the view of the writer, the only possible explanation of the facts observed.
The traditional concept of gastrulation in the echinodermata supposed that it takes place by invagination from one pole—-the vegetal one. This is indeed “true in the dwarf-egged group, such as Ophiothrix and Echinus. But, without any logical reason, it was then immediately assumed that such a process must also occur in all echinoderms.
The first voices raised against this rigid hypothesis were those of Apostolides (1882) and Russo (1891), who both claimed that the endoderm was formed by delamination in the embryo of Amphipholis (Amphiura) squamata. As it happens, their claim was based on a misinterpretation—as shown in the accompanying account of the development of that species—but the importance of their observation is that it demonstrated that gastrulation in Amphipholis certainly did not occur by means of simple invagination, for such a process when it occurs is unmistakable. However, their claim received no recognition, although it was never actually disproved. MacBride, in fact, dismisses the whole of Russo's account as “improbable in the highest degree,” and omits it from his account of the echinoderms
in his “Text-book of Embryology” (1914). The very fact of such an attitude is itself a demonstration of the inadequacy of the Recapitulation Theory—which of course provided no explanation of the peculiar development of Amphipholis.
Grave (1900) gave reasons for believing that the endoderm in Ophiura arose not by the invagination of a hollow archenteron, but as a solid inpushing which later became hollowed out to form an archenteron. Certain stages were missing from Grave's material, however, and advantage was taken of this fact to throw doubts upon the accuracy of his work (Bather 1901, and MacBride 1907). In his later work, however, MacBride (1914) quotes Grave's account without comment. As is seen from the facts quoted below, Grave's account receives strong support from independent evidence derived from the study of Kirk's ophiuroid.
Figs 16–18.—Effect on gastrulation of increasing yolk mass. Fig. 16, gastrulation in Ophiothrix (after MacBride); Fig. 17, first stage in gastrulatiou of Kirk's ophiuroid. Fig. 18, second (epibolic) stage of gastrulation in Kirk's ophiuroid. Figs. 17 and 18 simplified from Fell (1941).
In the latter species gastrulation takes place by means of two processes. First, as shown in Fig. 17, there is a slight inpushing of the macromeres from the vegetal pole. This results in the complete obliteration of the small blastocoel. The macromeres which were pushed in remain a solid mass of cells without any cavity. So far the process is similar to that described by Grave in Ophiura A second process commences now and involves extensive epiboly
of the micromeres, which migrate towards a central point on the vegetal hemisphere and then turn inwards (see Fig. 18). The point at which the inwandering of these micromeres takes place obviously is homologous with the blastopore, and a temporary small depression at that region is all that represents the archenteron. At a very much Jater stage a second cavity appears and extends as an excavation tip through the solid endoderm mass to form the definitive enteron.
To sum up, the effect of increasing yolk-mass upon the process of gastrulation has been, first, to modify invagination till it takes the form of a solid inpushing of cells. At a later stage an excavation in this mass produces the definitive enteron. This modification applies both to the moderately yolky type (Ophiura) and to the heavily yolked type (Kirk's ophiuroid). In the ease of the latter, however, the modification proceeds to a further stage, for the small blastoeoel makes it mechanically impossible to invaginate all the endoderm, and hence a secondary process of epibolic inwandering of the micromeres takes place, surrounding and enclosing the whole of the vegetal hemisphere.
5. The Critical Point in Development.
We have now reached the critical point in the development at which the forms with direct development diverge from the forms with a larval stage or with a vestige of a larva. Whereas the former proceed to adopt radial symmetry immediately after the conclusion of gastrulation. the latter begin to assume bilateral symmetry, and retain it for a greater or less period till it is finally obliterated by radial symmetry. At first sight it would seem that there is a complete hiatus separating the two types of development—but in the view of the writer, this is illusory. An important question that this paper seeks to elucidate is how it is that direct development has been brought about. The point is discussed in the following section.
6. The Disappearance of the Larva from Ontogeny
One of the proofs of the evolution theory cites the existence of “chains” or “series” of species showing, by small gradations, a progressive directional change. If the theory of receding metamorphosis outlined below be true, then we can reasonably expect to find some evidence of the existence of such a series of forms illustrating stages of the process. A survey of ophiuroid larval forms shows that such a chain does indeed exist.
We may take as the starting point of the bioseries a fully developed larval form such as the Ophiopluteus of Ophiura albida. Here there are four pairs of larval arms each supported by slender calcite skeletal rods (see Fig. 19a). The four pairs of arms are—the anterolateral, the postoral, the posterodorsal and the postero-lateral. There is a mouth communicating by the stomodaeal oesophagus to the stomach, from which a short intestine leads to the anus. To the left lies the five-lobed hydrocoel.
A second term in the bioseries is illustrated by the larva of Amphiura filiformis (Fig. 19b), in which the posterodorsal arms have disappeared, and at the same time the postoral pair have become reduced in size. The other features remain unaltered.
Fig. 19.—Sequence of Ophioplutei showing successive stages in the reduction und loss of the larva. a, first stage—-e.g., Ophiura, albida; b, second stage—e.g., Amphiura filiformis; c, third stage— e.g., Ophiura affinis; d, fourth stage—e.g., Ophiopluteus claparédei; e, fifth stage, Amphipholis squamata; f. final stage, lurva absent and development direct—e.g., Kirk's ophiuroid. The number above the larval arms indicate the order in which these organs disappear from the larva.
Next we may select the larva believed to belong to Ophiura affinis (Fig. 19c), in which the postoral arms have disappeared altogether, and the antcrolateral pair also. Thus the last pair to disappear will be the posterolateral.
In the larva known as Ophiopluteus clajtaredei (the parent species being unknown) we have the stage in which the posterolaterals have indeed disappeared. That the posterolaterals should be the last to go is of particular interest, because in the metamorphosis of Ophiothrix fragilis it is this pair which is the last to be lost. This peculiar armless larva was taken by Claparede (1863) swimming on the surface of the sea off the coast of Normandy. In his description
of the animal he confuses the hydrocoel with a developing young ophiuroid, owing to its five lobes, but his accurate rendering of the organ in the figure makes clear its true nature. He records that there is a mouth opening, but apparently the anus has disappeared for he does not figure or mention it. At the aboral end are at either side two spicules, clearly vestiges of the skeletal rods of the arms. At the aboral apex, and on each of the two pairs of projecting “shoulders” are tufts of cilia. There can be no doubt that these shoulders represent the two main regions whence the arms arise in the fully developed species. But the most significant fact recorded in his account is the fact that the larva was so opaque (“undurchsichtig”) that the internal organs were somewhat obscured. Now this opacity undoubtedly indicates the presence of yolk in the tissues, and here we have the first indication of the point in the series at which increasing yolk began to be of importance. The reduction in the alimentary canal is complementary to the presence of yolk (Fig. 19d).
A further step in this sequence brings us to such forms as Amphipholis squamata where the reduction has proceeded so far as to obliterate all traces of arm roots and cilia, leaving only a vestigial pair of skeletal meshes, recognisable as vestiges of the arm rods by their position and inclination, but having lost the slender spicular form. The mouth opening has now disappeared, the alimentary canal being thus vestigial. The yolk mass has increased so as to make the larva quite opaque until artificially cleared (Fig. 19e).
The endpoint in the series is represented by such forms as Kirk's ophiuroid, in which the larval stage has disappeared entirely from development (Fig. 19f).
For this process of shifting backwards of the time at which radial form is assumed I suggest the term “Recession of Metamorphosis”.
A further indication that this has actually occurred is provided by the hydrocoel. As is well known, the time at which metamorphosis first commences is indicated by the behaviour of the hydrocoel. For this organ moves from its position on the left side of the gut and begins to encircle the oesophagus, its five lobes becoming the five radial canals of the adult. As recession of metamorphosis proceeds, the encirclement takes place relatively earlier in the life cycle, till in the directly developing form it appears right from the beginning as a canal encircling the future oesophagus (Fell, 1941).
This process of earlier and earlier metamorphosis is in reality a species of neoteny—the larva becoming “adult,” as it were, at successively earlier and consequently undeveloped stages. A general diagram (Fig. 20) illustrates graphically the effect of recession of metamorphosis in ophiuroids.
It might be argued that the sequence of larval forms described above represents not a regression but a progression. This view would regard the fully developed Ophiopluteus as being a later evolutionary product derived from the simpler forms of larva, such as is found in
Amphipholis squamata. The following reasons seem to make such a view untenable:—
(1) The presence of what is obviously the representative of the larval arm-skeleton in the armless larvae of A. squamata and Ophiopluteus claparedei points strongly to a loss of the arms in these two species. It is very improbable that the skeletal organs of the arms would arise in evolution earlier than the arms themselves. Therefore the larval skeleton in these two species is to be regarded as vestigial, not rudimentary.
(2) The presence of a closed and non-functional alimentary canal in the larva of Amphipholis squamata is unlikely to be a primitive condition. An alimentary canal is required for digestion of food only by non-yolky larvae, such as those of Ophiothrix fragilis where it becomes necessary to swim and obtain food at an early stage. The closed alimentary canal of A. squamata is more likely to be a vestige of a formerly functional organ which has become physiologically unnecessary as a result of the presence of yolk.
(3) The simplified larvae, and the species which lack larvae, develop from yolky eggs of large size. It is unlikely that large yolky eggs are more primitive than small, non-yolky ones.
7. The Vitellaria or “Yolk-Larva” Series.
The bioseries we have just considered covers most of the ophiuroids. There remains, however, a peculiar divergent series which can be correlated with no larval series at all, but which nevertheless involves free-swimming larval forms with unmistakable characters in common. The amazing fact which emerges from a consideration of the literature is that this peculiar larval form is shared in common by no less than three of the echinoderm classes. Furthermore, it is invariably associated with a yolkmass in the egg. For this distinctive larval form I propose the term “Vitellaria”—or “yolk-larva”; it can only be considered as an independent sequence.
The general characters which distinguish this larval type are as follows:—The body is simply organised, having no pairs of larval arms, or other projecting organs. It is cylindrical or barrel-shaped, and opaque owing to the presence of yolk-material in the tissues. It is free-swimming, and is provided with rings of cilia. These bands are variable in number, but their general disposition is the same. There may be a larger or smaller tuft of cilia at the anterior end. They are commonly deeply pigmented. In Figs. 21–24 four larvae of this type are shown together; their common pattern is obvious, though they are drawn from three different classes of echinoderms. Indeed, these larvae, though belonging to Holothurians, Ophiuroids and Crinoids, have far more in common than have many of the Plutei or Bipinnariae. The Bipinnaria of Asterias, for example, is widely different from the great Bipinnaria of Luidia.
Hitherto in the literature these various forms have been considered separately. The Vitellaria is found among the Crinoids, and from that group no other larva is as yet known. Of Ophiuroidea, species of Ophiura are known to possess the larva in both America
Figs. 21–24.—Examples of echinoderm “yolk-larvae.” Figs. 21 and 22, of Holothuroidea (Cucumaria and Labidoplax); Fig. 23, of Ophiuroidea. (Ophiura brevispina); Fig. 24, of Crinoidea (Antedon bifida). Fig. 21, after Selenka; Fig. 22, after Dawydoff; Fig. 23, after Grave; Fig. 24, after Seeliger.
and Europe. It is the same as “The Worm-like Larva” of Muller, a fact first pointed out by Grave. Among the Holothuroidea, some species possess only the Vitellaria (“Barrel-shaped larva” or “pupa”), while others possess an Auricularia which later becomes a Vitellaria.
In view of the diverse internal organogeny it is impossible to regard the Vitellaria as representing any ancestral type. It seems to be a generalised yolk form developed independently by the various classes under some special circumstances of which we have no information. The circumstances, whatever they may be, are undoubtedly connected with the yolk mass common to them all, but why the larva should assume such a typical annulated form in each case is a problem. It is evidently an example of convergent evolution affecting the larval forms without changing the adult—that is, without leaving its impress upon phylogeny. The process by which the Vitellaria is developed is therefore yet another example from the echinoderms of the mechanism termed by de Beer “Clandestine Evolution”.
8. Ontogeny of the Coelom.
In parallel with the doctrine that gastrulation always takes place by invagination, there has grown a second doctrine equally
dogmatic that the coelom of echinoderms is always formed enterocoelously from pouches nipped off from the archenteron. Russo (1891) again was the first to state that the perivisceral coelom arose in Amphipholis squamata by a process of splitting in mesenchyme. It is instructive to note the way in which his results came to be ignored and forgotten. MacBride (1892) in his own paper on the very late stages of development of the same species makes no mention of the earlier development whatever. He does, however, mention that “. the coelom in Amphiura squamata is represented at first by a mass of mesenchyme; and as this condition of things is certainly not primitive, I do not think that even if reliable results as to the development of the cavities originating before the coelom clears were obtainable, they would be of much phylogenetic importance.” The reasoning here is very confused. First, there is no proof that the mode of origin of the coelom in Amphipholis is not primitive—for, as shown in my analysis at the beginning of this paper, there are as many directly developing echinoderms known to science as ones with indirect development. Secondly, MacBride expresses his opinion that any reliable results as to the development of the cavities originating before the coelom clears (italics mine) would not be of much phylogenetic importance. Here is a complete reversal in his attitude, for the only cavities which originate in the embryo before the coelom “clears” are the vestigial enterocoelous pouches! If, then, MacBride considers these vestigial enterocoels of little phylogenetic importance, why does he also regard the schizocoelous coelom as “certainly not primitive”?
However, it is clear from MacBride's next paper (1907) that he once more reversed his position, for he now proceeds to quote Metschnikoff's account of the enterocoels in Amphipholis, and uses this account to disprove Russo's later one. Still no actual reinvestigation has taken place of the development of Amphipholis to confirm or disprove Russo's account.
Finally, in MacBride's “Text-book of Embryology” (1914) Russo's results are rejected as “improbable in the highest degree”—though still no reinvestigation of Russo's work had been done. Russo's conclusions in regard to the origin of the coelom in Amphipholis squamata remained forgotten till, in the couse of my work on Kirk's ophiuroid, I was amazed to find the coelom originating as a schizocoel in a mass of mesenchyme. Such a mode of origin was of course foreign to embryological theory as taught in the text-books. Eventually I was able to obtain photostats of Russo's paper from the British Museum, and then the full similarity between his account and the conditions independently observed by myself became evident. My paper on the development of Amphipholis squamata contains the results of a reinvestigation of his work, and confirmation of his description of the origin of the schizocoel in mesenchyme (Fell, 1945).
Having thus described the rather belated recognition of a schizocoel in ophiuroid development, we can now pass to a comparison of the modes of origin of the coelomic cavities in the various types of development (see Figs. 25–27).
Figs. 25 and 26.—The enterocoels in Ophiothrix and vesicles of Amphipholis. In Ophiothrix (25) all parts of the coelom arise as enterocoelous vesicles. In Amphipholis (26) the number of vesicles is reduced, and none of them is an enterocoelous pouch. Only the hydrocoel survives, the other vesicles degenerating and contributing to the general mesenchyme, in which the remainder of the coelomic structures is later formed schizocoelously. Oes., oesophageal sac; Stom., stomach; Mesench., mesenchyme; Lft., Rt., Hydr., left and right hydrocoels; Lft., Rt., A. Coel., left and right anterior coelomic vesicles; Lft., Rt., P. Coel., left and right posterior coelomic vesicles.
In Ophiothrix, representing the non-yolky type, MacBride's description records that both right and left enterocoels form. The left divides into anterior and posterior parts, and soon afterwards the right does so also. From the posterior end of the left anterior coelom arises the hydrocoel. The wall of the left posterior coelom forms the arms, and its lumen the general coelom. The right coeloms become vestigial. MacBride also claimed that the right hydrocoel rudiments sometimes developed the five-lobe pattern, as does the left, but Narasimhamurti has since shown that this was a misinterpretation.
In Amphipholis, representing the moderately yolky type with vestigial larva, right and left enterocoels form. The right member disappears and the left gives rise only to the hydrocoel, and occasionally to a small posterior pouch which disappears later. The definitive general coelom arises by a process of splitting in an extensive mass of mesenchyme.
In Kirk's ophiuroid, where the embryo is heavily yolked and the larval stage completely omitted, both hydrocoel and general coelom
and its derivatives arise by splitting in mesenchyme. The hydrocoel is the first to form, and the embryo is free-living for some days before any general coelom commences to form.
In Ophiomyxa brevirima a closely similar condition holds, the coelom being preceded by an extensive zone of mesenchyme (Fell, 1940 b).
Fig. 27.—Schizocoelous origin of the coelom in a directly developing ophiuroid (Kirk's ophiuroid). A, the “rosette” stage, a solid gastrula assuming radial symmetry, and with budding podia; B, later stage, with the hydrocoel developing as a system of splits in the mesenchyme; C, general coelom developing by splitting in mesenchyme at a later stage than B. All represented in vertical section. Simplified from Fell (1941). Ect., ectoderm; Mes-end., mes-endoderm; Pod., podium; End., endoderm; Mo., mouth; Hydr., hydrocoel; Coel., general coelom; Mesench., mesenchyme.
To sum up, we note that as the yolk mass increases so does the tendency to form an enterocoel decrease. The hydrocoel lingers on longest as an enterocoel, but in the heavily yolked types it, too, arises by splitting, and is thus a schizocoel.
As noted in my paper on Amphipholis, the structures termed for convenience “enterocoels”—as they are clearly homologous with the enterocoels of Ophiothrix—are in fact formed by a splitting in small, solid masses of tissue formed on either side of the archenteron. Thus we have a series leading from the enteric pouches with preformed internal cavity, through enteric tissue-masses which later acquire an internal cavity, to finally the forms where the whole definitive coelom arises by an extensive process of splitting in a mesenchyme mass.
Whether it is the schizoeoelous or enterocoelous method of coelom formation that is the more primitive is a matter of personal opinion. In the view of the writer it is the enterocoelous method which is secondary. My reasons for this were set out in a previous paper (Fell, 1940a), and need not be elaborated here.
9. Later stages in Development.
With the foundation of the coelom and its derivatives the principle phases of early development are completed. The later stages in development are fundamentally similar in all the types. In the yolky forms the yolk spherules gradually disappear from the peripheral tissues to become localised in the mesendoderm region, where they are finally absorbed. Skeletal formation is essentially the same in yolky and non-yolky forms.
10. Viviparity as a Factor Producing Direct Development.
It is clear that viviparity must act as a factor producing direct development for the simple reason that it automatically effaces the pelagic stage from the ontogeny. It had been my hope that my study of the development of Amphipholis squamata would provide material for understanding the effects of viviparity. Unfortunately, however, Amphipholis, as already noted, has a yolky egg, and practically all the alterations in development noted are attributable to this factor. The only other viviparous ophiuroid of whose development we have much information is Ophiomyxa brevirima, of which I hope to publish a further paper at a later date. Here again the embryo proves to be extremely yolky and has a schizocoelous coelom (Fell, 1941). Indeed, from such facts as we possess at present, it seems that a yolky egg is a condition frequently associated with viviparity in ophiuroids.
One result of viviparity, however, is certain. This is, that the young sea-star is enabled to pursue its development to a much more advanced stage in the viviparous forms, before emerging into the outer world and seeking its food itself. Thus, in Amphipholis squamata, the newly born star has arms with ten to fourteen armsegments, whereas the newly metamorphosed star of oviparous forms is still in the unsegmented “Asterina” condition.
By means of experimental culture in vitro of explanted embryos of Amphipholis squamata it was shown that the embryo is unable to develop on its own yolk material alone (Fell, 1940 b). Certain substances must be added to the culture medium to enable normal development to proceed. This result taken in conjunction with anatomical evidence led to the conclusion that the developing embryos in Amphipholis are nourished by a secretion from the wall of the bursa. Here, then, is one undoubted effect of viviparity upon development in Ophiuroidea—that of prolonging embryonic life to a later stage by the secretion of nutritive substances. This effect is well illustrated by the following data:—Ophiocomina nigra, a non-yolky form with pelagic Ophiopluteus, has a functional alimentary canal by the third day, and continues to take in food for the remaining 35 days of larval life, before metamorphosis. A similar food-gathering larval period is found in Ophiothrix fragilis, which also has a functional
alimentary canal by the third day, and metamorphoses at about the twenty-sixth day. In the case of Kirk's ophiuroid, which is well provided with yolk, food is not taken in till the stage of about six arm-segments (i.e., when about twelve weeks old). In Amphipholis squamata, where the viviparous condition is found, the young ophiuroid does not seek food for itself till it has reached the stage of having about twelve arm-segments (i.e., when probably about six months old, on the supposition that it develops at a similar rate to other ophiuroids).
Beyond this conclusion there is insufficient evidence available as yet of the effect upon development in echinoderms of viviparity. Whatever its influence, the effect of increasing yolk-mass has apparently brought about the major portion of the modifications in development.
11. The Mechanism Which Has Brought About Direct Movement.
In the foregoing analysis of the developmental phases of ophiuroids it has been shown that the principle agent tending to produce direct development has been an increasing yolk-mass. One striking fact is the uniformity of the changes produced in diverse groups of ophiuroids—ranging from one of the most primitive species such as Ophiomyxa brevirima to the more specialised types. The modification of the gastrulation process, the suppression of the larva, and the increase in mesenchyme in which a schizocoel comes to form are some of the salient features of the yolky-egged forms.
The truly remarkable parallels which are observable in the embryonic stages of these various unrelated ophiuroids which have been influenced by accumulating yolk suggest very strongly that there exists some common principle which has operated in a common way upon these diverse forms. Before suggesting a possible answer to the problem of what this principle may be, it will be convenient to summarise the main stages in the sequence of forms showing reduction and loss of the larval stage.
We see in this sequence a progressive reduction in the size and number of the larval arms followed by reduction and final loss of the arm skeleton. Then comes reduction of the eoelomic pouches on either side, with a corresponding increase in the mesenchyme mass in which the coelom alternatively develops by splitting. Finally there is complete loss of bilateral symmetry, loss of the larval stage, loss of the vestigial enterocoels, and both general coelom and hydrocoel develop by splitting in the mesenchyme mass, which becomes increasingly important.
Now, in this progressive retreat inwards of the bilaterally disposed elements of the larval body with simultaneous recession of the radial symmetry into the earlier stages, we have a process suggesting alterations in the axial gradients of the embryo, these alterations being progressive along the series. Considering the modifications of development in this light, the following explanation may be put forward to explain the nature of the mechanism involved.
If we suppose that in the course of the evolutionary history of these larval forms changes began to take place in the metabolism along the axial gradients on either side of the body of the larva,
the changes being in the nature of inhibitions, then a reduction and final loss of the organs on either side would occur—larval arms, skeletal rods, coelomic pouches; in other words, the bilateral symmetry would become gradually submerged, and the point at which radial symmetry was assumed would undergo a recession towards the gastrula stage. This is precisely the sequence of events which, it has already been suggested, has actually occurred. Thus, an inhibitory influence acting upon the earlier stages of development is sufficient to bring about entire loss of the larval stage.
Here, as in all theories, a useful test is the inductive one. What evidence have we that such a process can take place? Is there experimental evidence? The answer is, I believe, supplied already by the work of Child (1916). In the course of experimental work upon the plutei and earlier embryos of echinoids this worker was able to show that a wide range of simple chemical substances could cause an inhibitory effect very similar to that which has been postulated above. He was able to produce sea-urchin plutei showing successive degrees of reduction and obliteration of antero-posterior, medio-lateral and apico-posterior differences through inhibitions of the axial metabolic gradients. In extreme cases his larvae resembled somewhat the peculiar armless larvae already referred to in this paper. Child drew from his work the deduction that the larval forms of echinoderms may have been evolved by increases in the metabolism along the metabolic gradients, producing the outgrowths of the arms, etc. Combining his results with the sequence of forms described in this paper, we can similarly add that the larval forms of echinoderms may have been secondarily lost again through inhibitions in the same axial gradients. The inhibitory substance may have been developed in close association with the production of the yolk material, or it may be that the yolk material itself provided the inhibitory influence.
A further parallel remains. Child found that the inhibitory effects were differential—acting to a variable degree on various tissues. Mesenchyme he found was comparatively unaffected, and consequently underwent a great increase in his larvae at the expense of the bilateral organs which became reduced. Now this is also closely similar to the effect noted in the transition from indirect development to the direct type. With reduction and disappearance of the larva there has been a corresponding increase in the amount and importance of the mesenchyme.
Therefore, in the view of the writer, there is good reason to believe that direct development in echinoderms has been produced by an inhibitory influence upon the axial metabolic gradients of the larva. The inhibitory influence is associated with an increasing yolk-mass, and has manifested itself through a steady recession of the metamorphosis towards the gastrula stage.
1. Direct development in the Echinoderma is no less typical than development of the indirect, larval type. Of the five living classes of echinoderms, only the Echinoidea are characterised by being predominantly of the indirectly developing type.
2. Increase in volume of the ovum is directly related to increase of the cytoplasm and its product, the yolk-mass. With increasing egg-size and amount of yolk, there has arisen a steadily increasing tendency to have direct development.
3. The increase of yolk has not greatly modified the process of cleavage, as segmentation in all forms is total. There is, however, a tendency to form micromeres and macromeres with increasing yolk-mass.
4. With increasing yolk-mass the wall of the blastula becomes steadily thicker and the blastocoel becomes reduced to a vestige in the animal hemisphere. The mesenchyme fails to separate as such but projects in a solid mass into the blastocoel.
5. The effect on gastrulation of increasing yolk-mass has been, first, to reduce invagination to a solid inpushing of cells; and, secondly, to bring about a subsequent epibolic inwandering of micromeres to contribute to the mesendoderm. The archenteron becomes vestigial and the definitive enteron is later excavated in the solid endoderm by a process of splitting.
6. A succession of stages in the reduction of the Ophiopluteus makes it probable that by a recession of the metamorphosis towards the gastrula stage, the larval period has been shortened, and finally lost altogether.
7. The recession of metamorphosis has resulted in the extreme case in radial symmetry being adopted immediately after the completion of gastrulation. This occurs in Kirk's ophiuroid.
8. A collateral and independent evolution has been followed by certain echinoderms with yolky eggs, resulting in the formation by convergent evolution, of a special yolk-larva, termed in this paper the “Vitellaria”. This process has taken place in the Holothuroidea, Crinoidea and Ophiuroidea, independently in each case. The Vitellaria is characterised by its cylindrical form, opacity due to yolk material present in the tissues, the complete absence of larval appendages, and the development of transverse rings of cilia.
9. With increasing yolk-mass there has occurred a reduction and loss of the enterocoels, with a corresponding increase in the amount of the mesenchyme: in this latter tissue the coelomic cavities arise by schizocoelous intercellular splitting. The hydrocoel is the last enterocoel to remain as such.
10. Viviparity appears to have been only a secondary factor in producing direct development. It has chiefly acted through prolonging embryonic development by enabling the embryo to obtain nourishment from the parent.
11. The suggestion is made that the mechanism of direct development has been an inhibitory influence upon the axial metabolic gradients of the larva. The inhibitory influence must be closely related to the presence of a yolk-mass and it has manifested itself through a steady recession of the metamorphosis toward the gastrula stage.
Apostolides, N. C., 1882. Anatomie et Developpement des Ophiures. Arch. Zool. Expl., 10.
Bather, F. A., 1901. Echinodermata (Introduction). Zool. Rec. (for year 1900).
Child, C. M., 1916. Axial Susceptibility Gradients in the Early Development of the Sea Urchin. Biol. Bull. Woods Hole, 30.
Claparede, A. R., 1863. Beobachtet ueber Anatomie und Entwiklungs-' geschichte wirbelloser Thieré, an dem Kuest. v. Normandie Angest., Leipzig (Engelmann).
Dawydoff, C., 1928. Traité d'Embryologie Comparée. Paris (Masson).
de Beer, G. R., 1930. Embryology and Evolution, Oxford University Press.
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. Probable Direct Development in Some New Zealand Ophiuroids. Trans. Roy. Soc. N.Z., 71, 1.
— 1941. The Direct Development of a New Zealand Ophiuroid. Quart. Journ. Mic. Sc., 82, 3.
— 1945. The Embryology of the Viviparous Ophiuroid Amphipholis squamata Delle Chiaje, Trans. Roy. Soc. N.Z., 75.
Garstang, W., 1931. Plymouth Marine Fauna, 2nd Ed.
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.
Huxley, J. S., 1928. Experimentally Induced Metamorphosis in Echinus. Amer. Nat., 62, p. 363.
— 1928. Artificially Induced Metamorphosis in the Echinoderms, Nature, 121.
Ludwig, H. and Hamann, O., 1901. Schlangensterne. Bronns Thier-Reichs, Bd. 2, Ab. 3.
Macbride, E. W., 1892. Development of the Genital Organs, Ovoid Gland, Axial, and Aboral Sinuses in Amphiura squamata. Quart. Journ. Mic. Sc., 34.
— 1907. Development of Ophiothrix fragilis. Quart. Journ. Mic. Sc., 63.
— 1914. Text-book of Embryology, Invertebrates. London (Macmillan).
— 1918. Development of Echinocardium cordtum. Quart. Journ. Mic. Sc., 63.
Mortensen, Th., 1924. Echinoderms of New Zealand and the Auckland-Campbell Islands; Ophiuroidea. Vd. Medd. Dansk. Naturh. Foren., 93.
— 1936. Report on Echinoidea and Ophiuroidea. Discovery Reports, 12 (Cambridge).
Mueller, Joh., 1846. Ueber die Larven und Metamorphose der Ophiuren und Seeigel. K. Akad. Wiss. Berlin.
Narasimhamurti, N., 1933. The Development of Ophiocoma nigra, Quart. Journ. Mic. Sc., 76.
Russo, A., 1891. Embriologia dell Amphiura squamata, Atti Accad. Fis. Mat. Napoli, 5, 2.