The Role of the Pituitary and the Thyroid in the Development of Teleosts

External appearance undergoes a striking transformation at a certain stage in the life cycle of some teleostean fishes. The previous pigmentary pattern of the body is replaced by a brilliant silveriness. Examples are the change from the parr to the smolt condition of Salmo and from the ‘yellow eel’ to the silver phase of Anguilla. The physiological significance of such changes is not known or is imperfectly understood.

It has been assumed that the silvering of the eel is a ‘breeding’ dress. Of Salmo such an assumption—if less explicitly made—has been shown to be unfounded by recent work of Orton et al. (1938), Jones & Orton (1940) and Jones (1940). These authors have found that a substantial proportion of male salmon parr becomes sexually mature before the change to the silvery smolt dress.

The present communication records data resulting from experiments designed to explore the role of endocrine agencies in relation to the change to the silver so-called breeding dress of Anguilla vulgaris L. and Salmo salar L. In the course of these experiments it has become clear that serious misconceptions, to which it is desirable to draw attention, prevail with regard to sexual differentiation in the former species.

The accepted criterion of sex distinction is the organ of Syrski. In view of confusion regarding this structure, it may be recalled that Syrski’s (1874) work antedated modern methods of histology and modern concepts of sex. Syrski inferred the existence of a sex distinction from the fact that silver eels may be divided into two classes: (a) larger ones with gonads containing recognizable oocytes, (b) smaller ones with Tobed’, i.e. more coarsely laminated, gonads in which no visible oocytes can be detected. Petersen (1894) emphasizes that the two groups are relatively discontinuous with respect to size.

Implicitly Syrski took for granted that the change from yellow to silver corresponds to the putting on of a breeding dress, i.e. denotes pending sexual maturity, an assumption repeated explicitly by subsequent authors, e.g. Petersen (1894) himself.

The validity of the criteria used as a basis of sex distinction in Anguilla may be questioned, especially because records of indisputable, i.e. sexually mature, males are extremely few. A Syrski organ is by general admission a structure in which no ripe sexual elements are as yet recognizable, i.e. it is an undifferentiated gonad. The line of reasoning followed by most, if not all, workers who have sexed eels would appear to be as follows:

• (a) members of most species may be divided into two sexes;

• (b) these may be distinguished by their gonads after sexual maturity ;

• (c) the eel becomes silver at the time of sex differentiation when two types may be distinguished;

• (d) of these one type is irtdisputably female ;

• (e) so the other type is presumptively male.

In the absence of a detailed cytological study of the gonads of the eel over a long period this argument fails to carry conviction unless it is proved that the silvering of the eel represents the assumption of a breeding dress. The relative size discontinuity emphasized by Petersen (1894) is irrelevant, since both growth and metamorphosis are seasonal phenomena. In other words, there is necessarily at least the difference of a year’s growth between two eels that do not undergo metamorphosis in the same season. So if we assume that sex differentiation and metamorphosis commonly occur about the same time, we should expect a more or less marked discontinuity between smaller silver eels with an undifferentiated gonad and larger silver eels in which the gonad is recognizably of one or other sex.

In comparatively recent studies contemporary workers in eel fisheries, e.g. Tesch (1928) and Jespersen (1926), have evaluated the secondary sex characters of the eel with no more certain guidance than the Syrski organ to indicate which individuals are males. Tesch (1928) divided eighty eels 20–25 cm. long into two groups. One group was killed. All of these had Syrski gonads (i.e. were supposed to be males). The second group was kept in running water and fed on dead shrimps. Of these, twenty-one killed after one year’s captivity also had Syrski gonads. The remaining eels, twelve in number, proved after two years’ captivity to be females (30–45 cm. in length) with ‘true ovaries in which ova were developing’. Prof. Lancelot Hogben, in whose laboratory the present work was carried out, wrote to Dr Tesch asking whether it was a correct inference that the so-called males had spermatozoa. Dr Tesch replied that the males were recognized as such by their Syrski organs, and that these organs, as is well known, do not contain ripe gametes.

What Tesch has therefore shown is that a Syrski organ may become an ovary in the normal course of events. This is not surprising, since the organ is an undifferentiated gonad and there is not the slightest reason to question the validity of his data. From his experiments Tesch draws the conclusion that the eel is a protandrous hermaphrodite. Since (a) all his animals presumptively had Syrski gonads at the start, (b) all of those kept in captivity for two years were recognizably, female in the end, all that his observations definitely prove is that the Syrski organ is not a sufficient criterion of maleness. These facts may be interpreted in three ways, any one of which is more plausible than the inference that the eel is a protandrous hermaphrodite :

• (1) The eel (like Cynips kollari, Rhodites rosae, Artemia sp. and many other arthropods) is a perpetually parthenogenetic form of which functional males rarely turn up.

• (2) The genetic mechanism of sex determination in eels is so nicely poised that (a) environmental changes which normally bias sex differentiation in favour of femaleness are all-important, (b) the production of a relatively minute proportion of mature males suffices to ensure fertilization in the localized breeding ground of the species.

• (3) Metabolic differences which antedate visible sex differentiation of the gonads lead to choice of different habitats by eels of different sex.

Of these (2) appears the most likely, though there is yet no conclusive evidence that either of the others are wrong, and it is difficult to see that such proof could be obtained without further research on the habits of the eel in its own spawning ground. So far we have no information about the spawning habits of the eel.

After the present work was begun an account of experiments by Tuzet & Fontaine (1937) was published. In 1933 these authors recorded active spermatogenesis after injection of human pregnancy urine into three eels which were already in the silver condition. They recorded better results (five specimens) in 1937 after similar treatment of silver eels 36–40 cm. long kept in darkness. One of these shed spermatozoa into the water. All these eels were caught in Loire and the Marais de Brière. This work suggests that the absence of light favours maturation of the gonads of male eels.

For the purposes of the present work on the problem of silvering we shall, therefore, regard the organ of Syrski not as an index of maleness, but simply as an undifferentiated gonad. The correctness of this view has been checked in every case by microscopic sections. As will be seen from what follows no indisputably male eel (i.e. one with gonads showing spermatogenesis) has been encountered, nor has spermatogenesis been induced during the whole of these experiments up to date.

The investigation began with the collection of data concerning the degree of development and cytological condition of the gonads of both yellow and silver eels, of various sizes. The eels used were caught in June 1937 from the River Dee at Aberdeen some two miles from its mouth.

The relation between body weight and gonad weight in twenty-five female eels is shown in Text-fig. 1, which shows that the weight of the gonads increases with the length of the eel. This appeared to be due solely to increase in the amount of fat deposited in these organs. The data show no significant difference with respect to relative gonad weight between silver and yellow eels of a given length. The gonads of all eels were sectioned and examined microscopically. The relevant findings were :

• (a) The gonads of all eels over 30 cm. long were recognizably female. No detectable difference was found between the appearance of the oocytes present in the gonads of yellow and silver eels of this size.

• (b) Smaller yellow eels had gonads containing no recognizable oocytes.

• (c) One smaller silver eel had an undifferentiated gonad.

Thus the preliminary sample supplied no justification for associating the change from yellow to silver with differentiation of sex or maturation of gonads.

Previous experiments on the effects of anterior lobe extracts, thyroid extracts and the urine of pregnant women had been carried out and were subsequently repeated. The activity of the urine and anterior lobe extracts used in all the experiments described was tested. Every extract and every specimen of urine gave a positive Hogben test (Crew, 1939; Landgrebe, 1939) with Xenopus. Each injection of thyroid and pituitary extracts represented the equivalent of 1 g. of fresh tissue. Each group of eels (six or twelve) was kept in a large black tank fitted with a closely fitting lid.

• (a) The first experiment was started in 1937, when eighteen yellow eels 39–46001. long were divided into three groups of six. Each group of eels was kept in aerated sea water. The water was changed twice a week. The animals in the first group were injected weekly with an extract of ox anterior lobe of the pituitary. Those in the second group received a weekly injection of an extract of ox thyroid. The third group were kept as controls. The experiment was continued for three months. The eels were then killed and examined. There was no difference between the three groups in the external appearance of either the animals or their gonads. Histological examination showed no difference between the gonads of the three groups. All of them were immature.

• (b) After Tuzet & Fontaine (1937) had obtained positive results in experiments on the eeLafter injection of human pregnancy urine, it was necessary to repeat this preliminary experiment, using both pregnancy urine and anterior lobe extracts. In a second experiment, undertaken in February 1938, eighteen yellow (44–48 cm. long) and eighteen silver (48–51 cm. long) eels were divided into three groups, each containing six yellow and six silver. The animals were kept under the same conditions as before. The eels of the first group were injected bi-weekly with an extract of ox anterior lobe of the pituitary. The eels in the second group were injected twice a week with 2 c.c. of pregnancy urine. The third group were kept as controls. The animals were killed after two months and no difference between the condition of the gonads in the two groups was detectable.

• (c) A third experiment began in 1939. Four batches of twelve yellow eels (30–35 cm. long) were used. No silver eels of this size could be obtained at this time from the Dee. They were kept in running fresh water in darkness in black tanks fitted with lids. The animals in the first batch received an injection of extract of fresh pig anterior lobe tissue once a week. Those in the second batch were injected weekly with an extract of fresh ox anterior lobe tissue. The eels in the third batch were injected weekly with 2 c.c. of pregnancy urine. The fourth batch were kept as controls. The experiment was continued for ten weeks. The animals were then killed and examined. There was no observable difference between the batches either with respect to the external appearance of the eels or of their gonads. Histological examination showed that the gonads of all the animals were immature. The gonads of the experimental eels did not differ from those of the controls.

• (d) Previous to the work recorded in this publication the author had found that the ovaries of Bufo vulgaris do not respond to injections of extracts similar to those used in the preceding experiments. However, the gonads of this toad can be matured and ovulation can occur after pituitary homeo implantation. A further attempt was made, therefore, to ripen the gonads of yellow eels by implanting freshly dissected eel pituitaries. Over a period of three weeks, eighteen glands were implanted into one animal. After a further six weeks, examination of the implantation site showed that most of the implanted glands had established themselves. Microscopical examination showed no change in the condition of the gonads.

A similar experiment was then performed on salmon with the assistance of Dr C. A. Wingfield, who is carrying out experiments on the effect of temperature and of mineral constituents of the water on the rate of growth of this species (S. salar). Six salmon parr, obtained from Thurso in March 1939 as yearlings, were used in these experiments. They were divided into groups of two fish and set out in running fresh water in three small white sinks. The fish were fed daily with liver. Injections began when the animals were about 20 months old. The first group were injected twice weekly with the equivalent of 0·5 g. ox anterior lobe pituitary tissue. The second group were injected with an extract of ox thyroid equivalent to 0·5 g. fresh tissue twice a week, and the third group remained as controls.

All six fish had characteristic parr markings at the beginning of the experiment. After one month those fish injected with thyroid showed marked silveriness and those in the group injected with anterior lobe extract showed slight silveriness. After two months the fish injected with thyroid were quite silver and the parr markings had completely disappeared. The control fish had not changed in appearance (Pl. 1). Fish injected with anterior lobe tissue assumed a silver condition more slowly and in addition acquired the typical green coloration generally taken to be an indication of the onset of maturity. Both groups had completely lost their parr markings and were of the same appearance as smolts. After treatment the six fish were killed and weighed. Their gonads were also weighed before fixation in Bouin’s fluid. The gonad body-weight ratio of the three groups did not differ significantly; and histological examination showed that gonads of all of them were immature. This confirms what the work of Orton and others has shown already. That is to say, the silvering of salmon is not of itself a sign of the onset of sexual maturity.

Crude extracts of anterior lobe tissue were used in this experiment and contained, among other autacoids, both gonadotropic and thyrotropic hormones. The silveriness which followed injection of this extract may have been due to stimulation of the animal’s own thyroid.

Systematists commonly recognize silver sea trout and brown river trout as varieties of the same species (5. trutta), though in fact relatively little is known about the life history of the former. Concerning the latter it is definitely known that no transition to a silvery condition accompanies sexual maturity or occurs at any stage of the life cycle in its fresh water habitat. During the summer of 1940 the writer was able to obtain a batch of brown river trout of the same hatching from Mr R. M. Neill of this laboratory.

Three groups of four fish 18 months old were kept together in a large aquarium in running fresh water. The fish in each group were distinguished by clipping the dorsal fin and the tail fin respectively in each of two groups. The fish in the first group were given bi-weekly injections of ox anterior lobe extract equivalent to 1 g. of fresh tissue. Those in the second group received thyroid injections equivalent to 1 g. of fresh tissue. The third group were kept as controls.

Within one month the trout which had thyroid injections appeared silvery and outwardly resembled ‘sea trout’ of similar size. Even after two months those fish injected with pituitary extracts did not differ in appearance from the controls. At the end of two months the trout were killed and the gonads examined. There was no difference with respect to the weight or appearance of the gonads between the three groups.

These results indicate (a) that there is no direct organic link between silveriness and sexual maturity; (b) that ‘silvering’ is brought about by direct action of the thyroid extract ; (c) that the crude pituitary extract injected into the first group of fish was not sufficient to stimulate the animal’s own thyroid. It is not surprising that greater stimulation would be necessary in the river trout than in salmon as silvering does not naturally occur in the former.

In view of the fragmentary state of our knowledge concerning the genetic relationship of the sea trout and the brown river trout, the experiments last recorded raise issues which would merit further enquiry before publication in circumstances other than the present. Owing to the uncertainty of enjoying further opportunities for immediate investigation, it is only possible to indicate some of the issues raised.

Since systematists do not agree among themselves about the definition of a species, the mere fact that the two types of trout are placed in the same specific category is unimportant from the experimental standpoint. These experiments seem to show that the difference between them is associated with functional activity of the thyroid gland. If we admit this conclusion several possibilities arise :

• (1) Since the iodine content of sea water is very much higher than the iodine content of fresh water, we might first assume that any trout supplied with sufficient iodine would undergo silvering and that whether it becomes a sea trout is an accident of habitat, i.e. depends on whether or not it gets into an estuary.

• (2) We may supplement the same initial assumption with the qualification that some stocks genetically differentiated from other stocks in this respect migrate down-stream as they approach maturity, so that only certain stocks get into the situation in which the thyroid is able to accumulate sufficient iodine to manufaoture the threshold thyreoglobulin for silvering.

• (3) We may reject the assumption common to (1) and (2) and assume that different trout stocks differ with respect to the iodine threshold of the thyroid gland. So that only those with a more sensitive thyroid are capable of becoming silver under natural conditions.

In other words, we may assume that the thyroid of some brown trout is incapable of discharging its active product owing to low pituitary activity—a possibility excluded by the negative effects of pituitary injection.

Of the three possibilities enumerated above, the third is excluded by the circumstance that brown trout are not found at sea. Furthermore, it does not seem likely that the one sample chosen for these experiments would be composed exclusively of individuals capable of becoming sea trout. We are left only with (1) and (2); and it would be a comparatively easy matter to exclude (1) by bringing the iodine concentration of fresh water in which brown trout are kept up to the level of what is found in sea water. Should the results prove to be negative, the second possibility stated above seems to be the correct one. If so, river trout and sea trout constitute two morphologically differentiated and discrete breeding units. So according to any intelligible definition of the word ‘species ‘they are specifically distinct. Further work will show whether this is true.

This investigation was made possible by a personal research grant from the Halley Stewart Trust. A grant from the Rockefeller Foundation to Prof. Hogben defrayed the cost of materials. The author is indebted to Prof. Lancelot Hogben and to Mr R. M. Neill for valuable advice and criticism of the manuscript.