Primary Sex-Phases in Ostrea edulis.

Although the phenomena of sex-change in the adult European oyster (Ostrea edulis) have become better known following the work of Spärck (1925) and Orton (a series of papers), who have followed up the pioneer work of Davaine (1853) and Hoek (1883), no clear account is available of the development of the primary sex-phases in the young animal, comparable, for example, with the study made by Coe (1932) of these phases in Ostrea lu rida. It was in order to fill this gap in our knowledge that this study was undertaken. Spärck gives an account of the development of the early sex-phases of oysters from the Limfjord (Denmark), but he was dealing with the oyster at practically the northern limit of its range (if we except the special conditions in the Norwegian polls) where breeding is intermittent, and good spatfalls occur only in exceptionally warm summers. Further, Spärck had at his disposal a very limited number of young oysters, whose age was rather uncertain. It is, moreover, not likely that the rate of succession of sex-phases would be the same in the Limfjord as on the British and French beds, which in favourable years may be said to provide something approaching optimum conditions. The differences have in fact been found to be considerable, as Spärck himself anticipated.

As early as 1853 Davaine recorded the discovery of oysters in September which had settled that season and yet contained ripe spermatozoa; this precocity in the formation of sexproducts has been confirmed by Orton (1921). Later Gerbe (1876) found spawning females amongst one-year-old oysters from Arcachon. This was followed by the work of Dantan (1913) who found 6· per cent, with embryos or ripe eggs in a sample of one-year-old oysters from Arcachon and the Gulf of Morbihan. Orton (1922) also records that following an unusually hot summer, about 10 per cent, of the largest supposedly one-year-old oysters from the River Blackwate carried embryos. Dodd et al. (1937) have recently recorded a very small percentage carrying embryos in one-year-old oysters from the Essex beds ; in this instance, however, as in some others, the oysters were collected from the natural beds and their exact age was rather uncertain.

Hoek (1883) concluded that in the Eastern Scheldt oysters might bear larvae for the first time when two years old, and might also produce the initial crop of sperm at this age. Neither Hoek nor earlier Dutch investigators had proved that sperm was produced by one-year-old oysters, while Hoek was clearly of the opinion that oysters never spawned as females at this age, and only occasionally when two years old, the majority of the larvae being produced by oysters four or five years of age.

Spärck (1925), in the course of a more comprehensive study of the breeding of young oysters in the Limfjord (Denmark), came to the conclusion that on these northern beds oysters never became sexually mature in the summer in which they attached themselves, but that an oyster in its second summer might produce sperm, and in exceptionally favourable circumstances egg-development might also begin in the second summer. Normally the development of the first batch of eggs only begins in the animal’s third summer, and it is even an exception when actual spawning occurs at this time. Ordinarily the eggs are liberated in the fourth summer, when the oyster is three years old.

Orton (1936), in a general review of the subject, concludes (p. 87): ‘The baby oyster probably always develops first into a male, but after exceptionally fine summers, as that of 1921, the female phase may be reached in twelve months. This rapidity of development of the young female is more usual in France than here in England, where young oysters may not usually produce the first batch of eggs until their third or even fourth summer.’ This recent statement reflects the lack of definite information concerning the breeding of young oysters and the sequence of sex-phases, and provides the justification for this study.

In much of the published work on the breeding of young oysters the age of the material has been somewhat uncertain; oysters from the natural beds have been judged to be one, two, or three years old as a result of their size and appearance and only in a few cases has the age been known with certainty. Even Spärck (1925) used oysters collected from the natural beds, and although from a knowledge of the growth-rate in the Limfjord he was able to assess their age with some confidence, yet doubt must remain. I have been fortunate in having at my disposal a large quantity of oysters of known age, all of which had originally settled in the breeding tanks at Conway (see Cole, 1938). After the oysters had attached themselves, the tiles carrying the spat were transferred to our grounds in the Menai Straits. When approximately one year old the oysters were detached and placed in wire trays. There was available, therefore, an almost unlimited quantity of young oysters of known age which formed ideal material for a study of this kind.

To give a measure of reproductive capacity the weight of the soft parts of each oyster has been recorded. This weight was obtained as follows. The preserved oysters were removed from alcohol, the surface liquid dried off, and the oyster weighed. The weight of the soft parts probably gives a better standard for the comparison of the reproductive capacity of different individuals than the width of the shell which may vary widely among oysters with the same volume of meat.

Orton in his extensive series of studies of sex-change has relied almost entirely upon fresh smears of the gonad contents in determining the sex-phase of individual oysters. This method has certain advantages, notably speed, but suffers from several drawbacks. In the first place it is not certain that the contents of a smear are fully representative of the actual contents of the gonad. This difficulty was realized to some extent by Spärck (1925), who however employed the smear technique extensively, although he also sectioned a number of oysters ; on p. 32 he writes: ‘If the sex is determined by means of pressing out some of the contents of the generative organ on an object plate and then observing them under a microscope, the eggs will only be visible in case the egg-development is rather advanced, otherwise only the fully developed genital products, the spermatozoa, are detected. In several cases I have, by means of sections, stated a more or less advanced egg-development in individuals which had through such a preliminary examination been determined to be males proper.’

Furthermore it has been shown by Coe (1932) that in Ostrea lurid a parts of the gonad may differ considerably in development, and it is confirmed that such differences may be observed alsoin Ostrea edulis. In such cases only a series of sections cutting across numerous follicles will reveal these differences. Again the relation between the different sex-cells both as regards quantity and position, the condition of the nuclei and cytoplasm, and the extent of the follicles themselves, cannot be accurately determined by the smear technique.

As noted by Coe (1932), who has based all his studies on the examination of serial sections, the method of ascertaining the state of the gonad by perforating the shell and withdrawing a little of the gonad contents is also unsatisfactory.

These considerations have led the writer to a belief that only by the examination of a series of sections across the body, cutting across numerous follicles, can a true picture be obtained of the development of the gonad and the precise relations and proportions of the cells contained in the follicles. Such a method is more laborious than the smear technique and does not permit of the examination of very large samples of oysters of any given age, but the amount of information gained per oyster examined is very much greater than by any other method.

One hundred and two oysters of the 1937 Conway crop were sectioned, samples being preserved at intervals of approximately one month from June to October 1938. When collecting a sample a selection was made of the various sizes, and particular attention was paid to the inclusion of some of the largest as it was rightly believed that these would be the most advanced in gonad development. The above samples of one-year-old oysters were supplemented by similar samples collected during 1939 from those trays containing oysters which had settled during 1938. To give information concerning the development of the gonad during the season of attachment, a certain number of the largest spat of the 1939 season were preserved during August, September, and October of that year. In addition oysters of the 1936 crop were preserved at monthly intervals from January 1938 to give information on sex-change during the second winter and third summer of the oyster’s life. Inclusive of all ages sections were prepared from about 180 individual oysters.

After trying other fixatives it was found that Susa (Heiden- hain) gave the best results, and all subsequent samples were fixed in this fluid. Sections were cut at 6-8μ. and stained with Ehrlich’s acid haematoxylin and eosin.

All the illustrations are from camera lucida drawings.

Before the development of the sex-products begins the gonad consists of a series of tubular ducts lying in the connective tissue between the epidermis and the digestive diverticula. The outer wall of each duct, i.e. the half nearest the surface, is lined with ciliated epithelium, interspersed with glandular cells, while the opposite inner wall, nearest the digestive diverticula, is lined by an irregularly double or triple layer of oogonia and spermatogonia. Even at this stage (Text-fig. 1), which may be called the indifferent or undeveloped phase (designated I), the spermatogonia and oogonia may be distinguished from each other. The nuclei of the oogonia are larger and more vesicular, with a prominent nucleolus and finely divided chromatin; the nuclei of the spermatogonia are smaller, more dense, with coarser chromatin and a less prominent nucleolus. It has been suggested by Orton (1927) that all gonocytes have the potentialities of developing into either ova or sperms, but this is against all theoretical conceptions and is not supported by histological evidence.

The indifferent phase (I) is followed by one in which spermato- genesis begins. The spermatogonia divide and form morulae of spermatocytes (Text-fig. 2) which by further division give rise ultimately to the ripe sperm balls. At this stage some of the oogonia begin to develop and become clearly recognizable. This stage may be called the young male phase (YM). The sperm morulae as they develop tend to occupy the lumen of the duct. Alveoli extending into the body at right angles to the surface now begin to appear. Text-fig. 3 shows a later YM phase with numerous developing sperm morulae filling up the lumen of the duct and follicle while others proliferate from the spermatogonia lining the walls of the follicle. The young oocytes are also becoming more prominent but as yet no ripe sperm balls are present.

The next phase is the functional male condition (M) in which the alveoli have extended further into the body while ripe and nearly ripe sperm balls fill the duct and the outer part of the lumen of the follicle (Text-fig. 4). In the deepest part of the follicle active proliferation of spermatogonia is still going on. Lining the walls of the follicle are numerous oocytes, some already very large. The development of the first female phase proceeds concurrently with the development and ripening of sperm morulae.

In the subsequent transitional phase (M-F) between the first male and female phases, ripe sperm morulae occupy the lumen of the duct and follicle, interspersed with others in all the later stages of development (Text-fig. 5), but the whole wall of the follicle, which has now begun to branch, is lined with developing oocytes.

By growth and enlargement of the oocytes and follicles the transitional M-F phase passes into the fully developed female phase (F). The follicles are packed with fully ripe eggs ready to be liberated (Text-fig. 6). Here and there between the oocytes and on the walls of the follicles may be seen residual groups of spermatocytes, while lining the walls at intervals are the groups of spermatogonia and oogonia which will form the basis of subsequent male and female phases. Text-fig. 7 shows a portion of the follicle wall of a ripe female and shows clearly that spermatogonia and oogonia are easily distinguishable although much compressed by the closely packed eggs. In the ripe female there may or may not be a few ripe sperm balls remaining in the ducts.

After the eggs are liberated proliferation of sperm morulae immediately begins and in the course of a few days ripe sperm morulae are again present. The early M₂ phase (Text-fig. 8) is easily recognizable on account of the large rather collapsed- looking follicles with sperm morulae at intervals along the walls. The walls of the follicle may still carry a few residual oocytes, particularly those which had not completed their growth when liberation occurred. The development of sperm morulae con- tinues and as they ripen they lie free in the lumen. On the walls of the follicles can be seen the oogonia which will later give rise to the second female phase.

The length of time elapsing between the first and second female phases depends upon the season of the year and the feeding conditions. If the liberation of the first batch of eggs occurs early in the season and conditions are good, the second male phase is accompanied by the development of the oocytes of the second female phase. Such oysters reach an advanced transitional M₂-F₂ phase by the time development stops at the approach of winter and mature in the F₂ phase early in the following summer. If the liberation of the first batch of eggs occurs late in the season or if feeding conditions are bad it will be followed by a shortened and weak second male phase which will then pass into a resting phase without sperm morulae or large oocytes. On occasion the post-spawning male phase may be suppressed altogether. This resting phase if entered upon at the end of the season persists throughout the winter and the second female phase (F₂) is not reached until fairly late in the following summer, being preceded by the M₂ phase in the early part of the season. The F₂ phase is followed immediately by the M₃ phase and thereafter alternation of sex-phases follows regularly, one female phase and one male phase being completed each year on the average on British beds.

During 1939 the settlement of oyster spat began in the Conway tanks on June 20, the earliest date recorded during the twenty years that the oyster-breeding experiments have been in progress. As a result of this early attachment the 1939 spat reached an unusually large size (ca. 3 cm.) during the 1939 growing season and a good opportunity was provided for examining gonad development during the season of attachment. In October a sample of the largest spat obtainable was preserved and a few of the heaviest individuals were sectioned. In all those examined the gonad was still in the indifferent undeveloped condition without sperm morulae, the gonads consisting of simple tubules lined on one side with spermatogonia and oogonia (as in Text-fig. 1). It should be noted however that, although this spat set very early in 1939 and water temperatures were high in August, the growing season was very much shortened by the occurrence oi unusually cold weather in late September and early October.

As noted earlier only two examples are known of oysters which contained ripe spermatozoa in the gonad in the season in which they settled. In one case (Davaine, 1853) the specimens were obtained from the warmer French beds, and in the other (Orton, 1921) the discovery was made in December, following an unusually long -warm summer. The time of year alone shows that the season was quite exceptional since spermatogenesis does not normally occur in the winter months (see p. 342).

We may therefore conclude that it is only in quite exceptionally favourable circumstances that oysters on British beds will experience their first functional male phase during the season in which they attach themselves. On the French beds, however, suitable conditions for early gonad development may occur more frequently.

(1) 1937 Spat in 1938.

The gonads of 102 oysters of this age group were sectioned and examined. The results of this examination are briefly recorded in Tables 1 to 4. It will be seen that the June and July samples consisted in the main of very small spat, the largest weighing only 0·57 gm. The August sample, on the other hand, included oysters up to 2·0 gm., while the smallest oyster sectioned weighed 0·47 gm. This marked difference resulted from the inclusion in the August sample of a good number of the largest spat available, whereas during July in the absence of the waiter no special effort was made to collect the larger spat; further, growth at this time was very rapid. During September the prevalence of high winds made many of the trays containing these one-year-old oysters inaccessible and it was not possible to obtain such a good selection of the larger sizes as had been obtained a month earlier. In October the sample corresponded to that obtained in August. The year 1938, in which these samples were collected, was noteworthy for the mildness of the autumn and the length of the growing season, which extended well into November. This season was therefore very favourable from the point of gonad development, which, judging from Table 4, continued practically to the end of October in the majority of cases.

Of the oysters examined in June and July (Table 1) very few had reached the functional male stage and only one (July 0·57 gm.) contained more than one or two ripe sperm balls. It is noticeable that this particular oyster is the heaviest in the two samples. In the few functional males the development of oocytes had already begun. The bulk of the oysters in the June and July samples were in the indifferent (I) or young male (YM) phases. It is quite evident that in all the oysters examined during these two months, which showed any development beyond the indifferent condition, the initial sex-phase would be a period of maleness. The August sample, as noted above, was much more widely representative of the population than that collected in July, which did not include the largest spat available, and a glance at Table 2 reveals the marked advance in gonad development.

The August collection comprises individuals in all stages of development from the indifferent juvenile phase to the fully- developed female ready to spawn, but there are no post-femalespawning phases. Generally speaking there is a marked correlation between the weight of the soft parts and the development of the gonads. Of the 32 oysters examined, 16 were predominantly female, including 6 ripe or nearly ripe females. The remainder was made up of 14 individuals in the transitional M-F phase, 7 functional males and 5 indifferent or early male phases. As might be expected the greater part of the undeveloped or early male phases occurred among the oysters of low bodyweight, whereas the heaviest oysters were predominantly female. The oysters of medium weight exhibited a great variety of sex-phases. In all functional males the oocytes show a considerable degree of development and a few ripe sperm balls are to be found even in the most advanced females. The liberation of the male sex-products therefore continues practically up to the time of female-spawning.

For reasons already explained the sample preserved in September was not so representative as that procured in August and in particular contained few of the largest oysters (compare body-weights in Tables 2 and 3). For this reason the number which had actually spawned as females among those collected in September was smaller than expected. Only two oysters had reached the second male phase (M₂) and in one of these the production of sperm balls was already declining and the tubules reverting to a resting condition; the other is illustrated in Text-fig. 8. The sample includes four ripe females several of which contain no sperm morulae at all. Inasmuch as the weather was mild throughout October 1938 it is probable that all these oysters would liberate eggs. The remainder of the sample includes individuals in all stages of development up to the transitional M-F phase and also one individual with no recognizable gonad in a long series of sections.

The October sample (Table 4) is in many ways the most interesting and informative of the whole series. It contains oysters varying in weight of soft parts from 0·51-2·66 gm., i.e. it is representative of all sizes. The stages of development shown by these oysters include all phases between the young male and the transitional stage between the second male and second female phases. Seven individuals exhibit post-femalespawning phases while there are several ripe females, but it is not certain that these -would have liberated eggs, although water temperatures were still high. Table 4 includes also a number of oysters which have passed into clearly defined resting conditions. These resting phases will be described later along with other phases occurring in the winter months.

(2) 1938 Spat in 1939.

This spat settled late in 1938 and was consequently small when the 1938 growing season ended. The effect of this late settlement was still evident during 1939 as the oysters were not so large as the 1937 crop had been at the same age. Further, although it was very hot during August 1939, an unusually cold spell during September and October brought the growing season to an abrupt close fully a month earlier than in 1938, when the autumn was very mild. The contrast between the two seasons is clearly brought out in the weights reached by the individual oysters.

Samples of the largest 1938 oysters obtainable during August, September, and October 1939‘were sectioned, and the condition of the gonads oi these oysters is shown in Table 5.

Of the twenty-nine oysters sectioned only two had spawned as females. Both of these had also functioned in the M₂ phase. Of the rest of the population, a few had reached the transitional M-F phase, but in no instance were the oocytes larger than halfgrown. Many oysters had produced some sperm, but practically all had passed into a resting condition by mid-September. Many of those classed as indifferent (I) in September and October may also have produced a little sperm earlier in the season, since it is almost impossible to differentiate between dormant males which have produced a little sperm and those oysters whose gonads have not developed beyond the indifferent condition.

Regarded solely from the point of view of conditions for gonad development the seasons 1938 and 1939 lay at opposite poles. In 1938 about a third of the population of one-year-old oysters produced broods of larvae; in 1939 only one or two of the heaviest individuals. In both years the bulk of the population matured in the first male phase, although this phase was not completed in the majority of cases during 1939. Despite the poor gonad development during 1939 it is highly probable that practically all of the 1938 crop would spawn as females at two years old, albeit late in the season, since in most cases the first male phase was not completed in 1939 and would be resumed early in 1940. A comparison of Tables 1 and 3 shows that under favourable conditions egg-development occupies only about two months so that the bulk of the 1938 crop would spawn for the first time as females in August or September of 1940. This population of two-year-old oysters would therefore show a considerable difference in spawning time when compared with the 1937 crop in 1939. Since a great many of the 1937 oysters wintered in the transitional M-F phase at one year old there would be in this population a high proportion of early summer spawners, while those spawning late in the season would be functioning as females for the second time. It is probable, therefore, that conditions during the first year or eighteen months of growth may control the spawning behaviour of a population of oysters for several years, until the cumulative influence of climatic conditions in successive seasons overcomes the effect of a favourable or unfavourable season in the first year of life.

It is surprising to find that a significant proportion of the one-year-old oysters may be expected to produce larvae on British beds in favourable seasons, for a study of the literature leads one to believe that female spawning at this age is quite exceptional (see Orton, 1937). Significant numbers of one-year- old female spawners have only been recorded previously on the French beds at Arcachon and on the Brittany coast. Dantan (1913) recorded, however, only 6-7 per cent, with nearly ripe eggs in the gonad or carrying larvae in a single sample from Arcachon examined in August, although a further 19 per cent, were recorded as having possibly emitted their genital products. Gerbe (1876) on the other hand, who examined samples of one- year-old oysters from Arcachon and Brittany at intervals from June 15 to July 31, found 35 with young, 127 ripe females, 189 males, 6 hermaphroditic, and 78 doubtful, which may have emitted spawn, in a total of 435. As Orton (1936) points out it would be quite wrong to assume that this population contained $35+127435×100$ per cent, of females, since many of the so-called ripe females would reappear in successive samples, as these were taken at roughly five-day intervals. Orton (1936) estimates the proportion of females in Gerbe’s material at about 15 per cent. Had Gerbe continued his observations into August and September it is highly probable that he would have found many more spawning females, and it seems likely, therefore, that on these warmer French beds quite high percentages of females occur among one-year-old oysters. It is almost certain that most of the so-called males of Gerbe were transitional M-F stages.

Orton also points out that it is important that Gerbe does not state that his material was selected at random, and the possibility cannot be excluded that he obtained such a high proportion of spawning females early in the season by selecting extra large oysters. Orton (1922) in a selected sample of large 1921 spat found about 10 per cent, with young in August 1922, but notes that the percentage with young in the whole population of one-year-old oysters was probably very small. Similarly Dodd et al. (1937) found rather less than 5 per cent, with either embryos, ripe, or nearly ripe eggs among selected large one-year-old oysters from the river Blackwater in 1936, but found only one mature female among 276 of the largest obtainable one-year-old oysters from the river Roach. In this ease, however, there is some doubt as to the exact age of the material.

In the warmer Norwegian oyster polls where water temperatures of 20°-30° C. are the rule in the summer (Gaarder and Spärck, 1932), Helland-Hansen (1907) is of the opinion that the whole stock of oysters breeds in one season, and, since the growth of spat is phenomenally rapid under these conditions, it is highly probable that some female functioning occurs at the age of one year. This is confirmed by Gaarder and Bjerkan (1934) who, although they give no information concerning the. percentage spawning, state that a one-year-old oyster produces a brood of about 100,000 larvae, which shows that oysters of this age with larvae do at least occur occasionally.

Samples of the Conway 1936 crop of oysters preserved in May and June 1938, and sectioned, give a clear picture of the sex-condition of the population at the commencement of the third summer, i.e. at the age of two years. Of the oysters collected on May 29,1938, fifteen were sectioned ; particulars of the sex-phases of these oysters are given in Table 6. The high proportion with abundant large oocytes in the gonad is quite striking. It will be seen from the table that of these fifteen oysters, selected at random from a large number of two-year-old oysters, no less than five were in a very advanced state of female development, while of the remainder one was a young female with no spermatogenesis, two were good males with small oocytes, while one contained a moderate number of large oocytes but also numerous ripe and developing sperm morulae and may be described as a true hermaphrodite. With the exception of the two males all of these oysters would certainly spawn as females during the season and the majority during the early part of the season, while even in the two males oocyte development was already beginning in May and would probably have been completed before the season ended.

The above sample leads one to believe that this population of two-year-old oysters would be nearly 100 per cent, female functioning during 1938 and this impression is very definitely confirmed by a further sample procured from the same lot of two-year-old oysters a month later. Particulars of the gonad conditions in these oysters are set out in Table 6. The ten sectioned comprised five ripe females, two nearly ripe females, one late transitional stage with very advanced egg development, one normal transitional stage, and one functional hermaphrodite (Text-fig. 9). This population would clearly be 100 per cent, female functioning notwithstanding that no actual spawning had taken place during June. It should be noted, however, that, following a warm May, water temperatures dropped considerably and were low throughout June. Inasmuch as five of the ten oysters sectioned in June were fully ripe females, ready to spawn, it is possible that some of these oysters, after completing their post-spawning male phase, might even spawn again as females in the same season if water temperatures remained high during September and October, as was, in fact, the case.

It is possible to deduce, with reasonable certainty, the phase reached at the end of their second summer by each of the above twenty-five two-year-old oysters. The results obtained can be compared with those obtained from the two groups of one-year- old oysters. Nine oysters which had reached, or practically reached, the F₂ phase must have spawned as females (F phase) in 1937, when one year old, and later entered the M₂ phase ; in some the latter phase must have been practically completed. One other classed as F₂ might possibly be a very good F phase although the gonads are more voluminous than is usual in this phase. The two oysters in functional hermaphroditic phases are assumed to have spawned incompletely in the F phase at the close of the 1937 season. With the exception of the two good males, all the others reached the transitional stage between the first female and male phases, while in many this transitional phase was practically completed, with the result that they matured in the F phase the following spring without any further spermatogenesis. This population of one-year-old oysters, therefore, reached much the same average stage of development during 1937 as the 19.37 oysters attained during 1938 (see Table 4), with about one-third spawning as females in the summer following the season of attachment.

In order to provide information concerning the condition of the gonads during the winter months, oysters of the 1936 crop were preserved during January, February, March, and April 1938. Sections w’ere prepared from twenty-six individuals and these, when considered in conjunction with the samples of one-year-old oysters examined in October, give a clear picture of the sex-phases occurring during the winter. In addition, when considered in relation to the 1936 oysters examined in May and June 1938, they give further information on the course of gonad development during the spring and early summer in two-year-old oysters. Brief details of the gonad conditions in these wintering oysters are given in Table 7. It will be seen that individuals may be found during the winter in a great variety of sex-phases, but developing sperm morulae do not occur from January to March inclusive, but only in April when gonad development had recommenced ; on the other hand, ripe sperm morulae occur in abundance in some oysters. Spermatogenesis, therefore, is suppressed during the winter and where ripe sperm morulae are present they have been carried over from the previous season. Egg development is also arrested, and the stage at which an individual passes the winter is conditioned by the phase which it reaches at the close of the previous season.

All oysters in an indifferent or young male phase at the approach of winter pass the winter in an indifferent condition with only oogonia and spermatogonia in the follicles. Any developing sperm morulae left at the end of the season are broken down and absorbed by phagocytes. These oysters mature in the male phase early in the following summer, spermatogenesis commencing in April when the water temperature reaches about 10° C. and feeding begins. By the end of May spermatogenesis is well advanced and ripe sperm morulae are present in abundance in the gonad (see Table 6).

Those oysters in the first functional male phase and those in transitional M-F phases towards the close of the season have abundant developing oocytes in the gonads and may apparently pass the winter in either of two distinct conditions. In the first group all sperm morulae ripen and are liberated before winter begins ; in the second the sperm morulae ripen, but are retained throughout the winter. During the winter the former have gonads lined with numerous medium-sized oocytes interspersed with groups of spermatogonia (Text-fig. 10) but without any sperm morulae. It is highly probable that some of these oysters mature as pure females in the following summer (several apparent examples occur in Table 6), spermatogenesis being suppressed while the eggs develop. In others there is a burst of spermatogenesis in the lumina of the follicles during April and May and these sperm balls are liberated at the outset of the season, being quickly followed by the spawning of the eggs. The oysters comprising the second group may have gonads lined with oocytes up to about three-quarter-grown size but also containing numerous ripe sperm morulae. These sperm morulae appear to be put into cold storage, as it were, during the winter months and are liberated early in the following summer just before the eggs become ripe. It is not usual to find oysters with ripe or nearly ripe eggs during the winter months for the largest oocytes are generally about three- quarters grown ; from this we may infer either that individuals with eggs that are nearly ripe towards the close of a season do actually liberate these eggs even in October or November, or that absorption of large eggs takes place during the winter. It is well knowm that on occasion odd w’hite- or black-sick oysters may be found during the late autumn or early winter months and most workers could quote such examples. In this connexion it should be noted also that Orton (1936) has shown that eggs may be shed straight into the sea without incubation at the close of unfavourable seasons.

Table 7 includes two oysters with gonads containing a mass of broken-down oocytes and loose yolk mixed with innumerable phagocytes and shows that absorption of large eggs does occasionally take place during the winter months. There is of course the possibility that these two oysters were in a pathological condition or that the eggs in the gonad had been retained after very incomplete spawning at the end of the previous season. In the latter case, however, one would have expected them to have disappeared before January.

Oysters which spawn as females for the first time in their second summer may also experience the second male phase towards the end of that summer, but, unless the oocytes of the second female phase have already begun their development, such oysters pass the winter in a dormant male phase with large empty follicles lined with spermatogonia and oogonia, the former being- frequently very numerous. These individuals mature early in the following summer as good males (M₂) with very abundant sperm morulae and later pass into the second female phase.

The wintering phases described above agree very well with those described by Spärck (1925), the only one of the earlier workers to have studied them in any detail.

The examination of samples of one-year-old oysters leaves no doubt as to the universal prevalence of protandry in oysters. AH oysters experience an initial functional male phase, although this phase may be brief and the quantity of ripe sperm morulae produced may be small. This initial male phase is rapidly succeeded by the first female phase, the oocytes beginning to develop before the first sperm morulae are ripe. Growth of the oocytes proceeds while the first male phase reaches its peak and a few sperm morulae may often be found in the gonad even when the eggs are ripe. Orton (1927) gives convincing proof of the universal occurrence of a functional male phase immediately following upon female-spawning in older oysters and one-year-old oysters are no exception to this rule. Following upon the spawning of the first batch of eggs the second male phase is rapidly assumed, being easily recognized in its early stages by the comparatively large almost empty follicles with sperm morulae at intervals along the walls, and by the frequent occurrence of small numbers of retained ova, these latter being usually unripe. Orton (1927) shows that the post-spawning male phase continues for two or three months, but states that egg development does not begin until sperm development has been completed. Amplifying this belief he states that the different sex-phases are mostly clear-cut. This view is however incorrect, for many oysters in the M₂-F₂ transitional stage contain abundant developing sperm morulae as well as very numerous oocytes (e.g. 2·37 gm., Table 6, and 2·40 gm., Table 7), while, as stated above, in one-year-old oysters egg-development is normally well advanced by the time the first spem morulae are ripe. Further, practically all authors agree (see Spärck, 1925) that some ripe sperm morulae are nearly always to be found in the follicles and ducts of oysters in the female phase shortly before spawning, if not actually at the time of spawning. Since egg-deyelopment occupies nearly two months under the most favourable conditions, spermatogenesis must continue for the greater part of the developmental period of the eggs. Towards the close of the season, however, a weak post-spawning male phase may occur without the oogonia of the next female phase showing any development.

The so-called ‘hermaphroditic’ phases of Orton (1926), which he regards as forming a special, rather anomalous, category, are without doubt normal transitional M_x-F_x stages. The reason why these phases have not been more frequently recognized and viewed in their proper perspective by Orton is to be found in the limitations of the method of determining sex-phases by examining fresh smears of the gonad contents. If the follicles are fined with small oocytes tightly adhering to the walls, as in the earlier transitional stages, the chances that these oocytes will show up in a fresh smear of the gonad contents are very small. Orton, himself, states (1927, p. 981) that young ova do not become recognizable in fresh tissues until they reach a size of 40-50μ, i.e. not until their diameter is about half that of ripe eggs, or roughly the size of the largest shown in Text-fig. 4. Therefore all the early transitional phases pass unrecognized and are classed as males in Orton’s analyses, which probably explains why he finds such high percentages of ‘pure males’ early in the season. In samples of oysters mainly four or five years old from the rivers Fal and Blackwater in June 1926 Orton records approximately 50 per cent, of pure males. As will be seen from Table 6, among two-year-old oysters it is usual for all to carry abundant developing eggs in the follicles at the beginning of the breeding season, and pure males without eggs do not exist at this time of the year. There are further no reasons for supposing that the sequence of sex-phases in older oysters differs from that in two-year-olds, in fact there are excellent reasons for assuming that a similar state of affairs prevails. It is probable that the majority, if not the whole, of these ‘pure males’ of Orton would on sectioning prove to be early transitional phases. It is significant that in the particular populations referred to above Orton records only about 12 per cent, of mixed sexes. Pure males, i.e. individuals without any female gonocytes other than oogonia, do occur towards the close of the season when, after complete female spawning, a weak male phase may occur if temperature and food conditions are not sufficiently good to initiate egg development. A similar pure male phase, preceding a resting condition, may also occur during the summer when environmental conditions are definitely unfavourable, but rarely, if ever, occurs in June after the spring feeding. Orton’s suggestion that there may be two completely distinct types of males obtains no support from the work here described.

The manner in which true hermaphroditic forms, such as those recorded in Table 6, are produced is still obscure. In these individuals the gonads contain both large practically ripe oocytes as well as very abundant ripe and developing sperm morulae. Oocytes and sperm balls are rather evenly distributed throughout the gonad and from the large size of the eggs and the abundant early sperm morulae one would be inclined to deduce that the eggs would become ripe, and would presumably be liberated, while the gonads were still actively sperm-producing. In a normal transitional M_x-F_x stage containing eggs as large as in the hermaphroditic example figured (p. 340), one would expect a condition similar to that shown in Text-fig. 5, where spermatogenesis is confined to the lumen of the follicle and unripe sperm morulae are not numerous, while the walls of the follicle are completely lined with large oocytes. It is not easy to see how such functional hermaphroditic phases arise, especially as they have been found only at the beginning of the season, but it is interesting to note that a similar phase occurs occasionally in Ostrea lurida (Coe, 1931), also at the beginning of the season. No explanation of their origin in this species is advanced by Coe. It may be that they are derived from individuals which spawned incompletely at the end of the previous season leaving behind in the gonad a fair number of unripe eggs which were retained throughout the winter. If so the hermaphroditic phases in Table 6 should be looked upon as good M₂ phases with large numbers of retained eggs properly belonging to the F phase.

Spärck (1925) remarks that when a stock of adult oysters is investigated during the breeding season one cannot help being struck by the preponderance of male phases. This observation was also made by Da vaine (1853) and Hoek (1883). It is nevertheless incorrect and is due mainly to the technique employed in determining the sex, viz. by the examination of a smear of the fresh gonad contents. As males proper are reckoned all those individuals yielding abundant sperm morulae, but a great many of these are actually in an advanced M_x-F_x transitional stage, for sperm morulae are generally present in the gonad up to within a short time of spawning (see Table 3).

Orton (1927) notes that exceptions to the general rule that post-spawning females with embryos in the mantle cavity carry sperm morulae in the gonad, occur at the close of the season. Such individuals have passed directly into a resting phase with only spermatogonia and oogonia in the gonad, and will pass the winter in the same condition, maturing early in the following season as good males with, however, abundant small oocytes lining the follicles.

It is not known to what extent breakdown and absorption of large eggs may occur during the winter, for the two oysters in which such a condition was found may have been pathological. In both, the follicles were full of degenerating eggs and loose yolk mixed with phagocytes. Phagocytes usually begin to collect in and around the follicles when the eggs are ripe and play an important part in the clearing up and reorganization of the gonads following female-spawning. They are also frequently abundant in the gonads towards the approach of winter and during the wintering period itself, and there is good evidence to show that when early post-spawning male phases revert to the indifferent condition without sperm balls at the approach of winter, phagocytes play an important part in breaking down and absorbing the sperm morulae remaining in the gonad.

Coe (1931) has noted that in Ostrea lurida the individual tubules of the gonad may differ considerably in sex-phase, particularly in young oysters. A similar phenomenon also occurs in Ostrea edulis. The gonad tubules nearest the hinge, and therefore most remote from the gonoduct aperture, are always the least advanced in development and may be in a state of active spermatogenesis with very small oocytes while the rest of the gonad is full of practically ripe eggs. Such backward tubules do not usually form a large part of the gonad and appear in their proper perspective in a series of sections. Similarly when the follicles are full of ripe eggs ready to be spawned, and are free of sperm morulae, the large gonoducts may still contain numerous sperm morulae.

Orton in his study (1936) of spawning in the rivers Fal and Blackwater during 1927 arrived at a figure for the total of functional females by determining the percentage carrying coloured larvae in large weekly samples throughout the season and aggregating these weekly percentages. The proportion of this category in the population was 54·9 per cent, in the Fal and 59·3 per cent, in the Blackwater. This method is subject to error only if coloured larvae are on occasion retained more than a week, or if eggs may be liberated into the mantle cavity and reach the black-sick stage and be again liberated all in less than one week. Errors due to these two possibilities will be slight and will offset each other. The only other possible source of error in Orton’s observations lies in the fact that for two periods of three weeks and two weeks, in August and October- November respectively, he was unable to obtain samples from the river Blackwater. Judging from earlier figures, had samples been obtained during these two periods they might have added another 3-5 per cent, to the total of oysters with coloured larvae. We may therefore regard it as established that in 1927 the oysters which actually liberated larvae comprised about 60 per cent, of the population in both rivers. It has not been sufficiently emphasized, however, that the summer of 1927 was abnormally wet and cold and therefore distinctly unfavourable for the production of high percentages of spawning females. In his review of the weather during 1927 Orton quotes the following passage from the Meteorological Office report for the year. The general conditions are described as ‘a wet year with a dull and wet summer, … the outstanding features being the persistent rain, thunderstorms, floods and lack of sunshine during the summer months and early Autumn, June-September inclusive, and the excessive wetness of the year as a whole. To find a summer which in respect of persistent wetness and lack of sunshine can compare with that of 1927 it is necessary to go back to the black year of 1879.’ It is clear from this statement that the summer was an unfavourable one for oyster breeding and that the percentages of spawning females actually realized in the rivers Fal and Blackwater in such a year are likely to approach the minimum rather than the average. This point was not fully brought out by Orton in his conclusions for he states that ‘these facts prove that reversion to the female phase must occur on the average every other year, and that a small proportion revert more quickly than every other year’. Rather, we may say, that the figures given prove that in an unfavourable year during which conditions were so bad as to cause extensive abortion of embryos and the liberation of unfertilized eggs on the Fal, the proportion of the population which actually liberated normal larvae was approximately 60 per cent., and we may therefore anticipate much higher percentages under more favourable conditions.

Orton has, moreover, in an earlier paper (1926) provided evidence strongly suggesting that 100 per cent, female functioning occurs on the same beds in the river Fal in warm seasons. He shows that samples of oysters taken at monthly intervals during the weeks after full moon in July, August, September, and October 1925 included respectively 33, 18, 22, and 1·5 per cent, carrying embryos. Other samples taken midway between these dates in July and August included a further 28 per cent, carrying embryos, giving a total of over 100 per cent, female spawners during 1925. This fact is noted by Orton and he discusses the possibility that the high percentage might be the result of examining distinct populations from the different beds in the river Fal, that is to say populations exhibiting distinct differences in spawning time, but concludes that these differences are slight and may be ignored. It is just possible that marked differences in age between oysters in the various samples may have destroyed to some extent the homogeneity of the material, but it is very unlikely ; this possibility is not discussed by Orton. In a later paper (1936) Orton uses some of the results obtained in 1925 to prove that at least 50 per cent, of the population spawned as females during 1925, but does not refer to the fact that the figures given seem to prove that 100 per cent, female spawning occurred during that season if the population was homogen ecus as regards spawning behaviour. Presumably he is no longer satisfied that this condition was fulfilled. Nevertheless the results definitely indicate a very high percentage of female spawners since one sample alone contained 33 per cent, carrying embryos.

Again, the same author (1927) in the course of a study of the phases following female-spawning gives particulars (Table IX, p. 1034) of twenty-one oysters known to be female-functioning in one year which were also found to be female-functioning in the next. He gives also particulars of two individuals which contained shelled larvae on July 29,1926, and again had shelled larvae in the mantle cavity on September 29, 1926, and also of two which were found carrying shelled larvae on July 21,1926, and were again white- and grey-sick respectively on September 30,1926. These examples prove that in years when the autumn is mild oysters may spawn as females twice in the same season. The only other explanation of these facts would be that these four oysters had all spawned partially as females in July and had successfully spawned the remainder of the eggs in September. This explanation, which is unlikely to be correct in all four cases, is rejected by Orton. In the face of these examples and those quoted earlier it is remarkable that Orton should conclude that on the average oysters spawn as females only every other year..

To the present writer these results obtained by Orton from the study of populations of adult oysters seem to confirm to some extent the results obtained in this study of a strictly homogeneous population of two-year-old oysters and, taken together, seem to warrant the conclusion that in warm summers on British beds all the oysters two years old or older may be expected to mature as both males and females.

Since there is no evidence that a resting stage occurs during the summer except under definitely unfavourable conditions, oysters which spawn as females early in the year should reach at least the transitional M_x-F_x phase by the end of the summer and would again mature as females in the year following, thus functioning regularly as females each season. Those oysters maturing in the second male phase at the beginning of their third season, after a feeding period of about three months in the spring, complete this phase early in the season and spawn as females in the late summer, passing the winter in the dormant male condition and repeating the cycle in the following year. Thus in general each oyster should complete a male and a female phase during the course of the summer, and female spawners will fall roughly into two groups, early summer spawners and late summer spawners. Some of the data given by Orton (1926) tend to show that such a grouping is in fact evident on British beds in favourable seasons. The group of early summer spawners usually comprises considerably more than half the population as one can see from Table 6, which includes about 75 per cent, early spawners among two-year-old oysters. The reason for this can be deduced from what has been said about wintering conditions. In those oysters which winter in the transitional M_x-F_x phase, oocyte development recommences in spring as soon as the water temperature exceeds about 10° C., being accompanied in many cases by a recrudescence of spermatogenesis. Sperm morulae are liberated by such oysters very early in the season and the functional male phase is brief, being rapidly succeeded by the spawning of the crop of eggs. In some cases the period of spermatogenesis may be omitted altogether, such individuals appearing as fully developed pure females (see Table 6) early in the season. The bulk of the sperm morulae liberated early in the season is produced by oysters which were at the beginning of a post-spawning male phase when winter set in and passed the winter in a dormant male condition without large oocytes in the gonad. As the percentage of transitional phases during the winter months is always high we have here the explanation of the fact already noted by several observers (Hoek, Spärck, Orton, &c.) that spawning females are most numerous at the beginning of the season.

As I have already suggested elsewhere (Cole, 1941), it is probable that the quantity of eggs produced by early summer spawners considerably exceeds that produced by late summer spawners, for, while in the former the whole of the reserves accumulated during the spring feeding are available for the maturation of the first crop of eggs, in the latter a part of these reserves have already been utilized to mature a crop of sperm balls in the early part of the season.

It is clear that the length of the breeding season and the quality of the feeding conditions will affect the sequence of sexphases. An unusually shortened season with a cold autumn may prevent the maturation and spawning of the autumn group, while poor feeding conditions during the summer may induce a resting phase and have the same result. Most British oyster beds provide something approaching optimum conditions in warm seasons and there is little doubt from the evidence discussed herein that the normal figure for the proportion of female functioning oysters should be 100 per cent., from the age of two years onwards, and not 50 per cent, as has been frequently suggested. Indeed it is possible to go further and say that unless 100 per cent, female functioning is achieved in practice among all stocks of oysters two years old or older it cannot be maintained that favourable environmental conditions are being provided on the beds. It will be conceded I think that the principal biological advantage of sex-change lies in the possibility of achieving 100 per cent, female functioning each season. There is no reason why this figure should not be exceeded in years when the breeding season is exceptionally prolonged by the occurrence of a mild autumn, and Orton has actually provided particulars (detailed earlier) of four oysters which were known to have spawned as females in July and again in late September. In the very closely related species Ostrea lurid a from the Pacific coast of America, Hopkins (1937) has shown that it is by no means unusual for individuals to spawn twice in a season as females, such a second spawning sometimes involving as much as 50 per cent, of the population, while about 75 per cent, female functioning is the lowest proportion recorded.