The Effects of Fat-Soluble Vitamin A Deficiency on Reproduction in the Rat

In the course of a lengthy study of the fat-soluble vitamins which has been in progress in these laboratories for several years past, we have often observed that whereas reproduction seldom takes place in rats when they are being fed on a diet deficient in these essential factors it is soon re-established when the animals are once again given a normal diet.

This would not be surprising if the length of time on the deficient diet had been comparatively short, but it is indeed striking when the recovery follows a very long period of stunted growth and development.

In view of the results recently reported by us on the disturbances of sex-ratio in rats caused by a deficiency of the vitamin B (6), it becomes of interest to make a closer study of the influence of these other dietary components on reproduction.

It is well known that withholding the fat-soluble vitamins from the food of young animals results in a retardation or suspension of growth. This effect is usually accompanied by disturbances of the calcification of bones and teeth and by a more or less marked fall in the resistance of the organism to infection. Indeed, it is to the latter cause that one ascribes the high mortality from diseases of the respiratory tract that characterise animals fed on such defective rations.

Between the condition of the normally nourished animal and that of the rat which stops growing after four to seven weeks as a result of vitamin A starvation there is one state that is of particular interest to us. It is that of the animal receiving a relatively small supply of the fat-soluble vitamins and apparently enabled thereby to maintain fairly normal health, at any rate as regards resistance to the usually fatal diseases that kill off the more restricted controls, without being able to grow. Such animals may with care be maintained alive in a stunted condition for several months, certainly up to an age when had they been normally fed they would have arrived at sexual maturity and adult size. Even then, on being given normal food, they will grow and prove capable of producing and rearing young. The capacity for growth after a period of suppression due to deficiencies of one type or another in the diet has been studied by Osborne and Mendel (5) who disproved the contention of Minot (4), supported by Rubner, that the rate of growth depends on the degree of senescence, and that the capacity for growth finally disappears even when it is not called into use. The evidence provided by Osborne and Mendel supplemented the earlier and less convincing observations of Aron, in showing that rats can grow to adult size on a satisfactory diet after their growth had been suppressed by qualitative defects in their food supply for periods considerably longer than that normally required to arrive at maximum body weight.

Some interesting confirmations of this work are provided by our own observations, which are dealt with in the next section of this paper.

The material for this work was selected more or less at random from the large number of animals being employed in these laboratories for researches on the fat-soluble vitamins. Such animals were fed from an age of about 5 weeks (50 to 60 gms.) on a basal food mixture designed to be as free from the fat-soluble vitamins as possible.

Animals fed on this ration ceased growing after a short time and unless they received some of the missing dietary factors declined and died, frequently from pulmonary disease, in from six to eight weeks. Many of the rats used in this research received from time to time supplements that supplied the small amount of fat-soluble vitamins necessary to enable them to survive without showing appreciable growth. Such animals, in spite of their stunted condition, showed no obvious signs of ill-health, ate well and displayed marked activity, but no reproduction took place. In appearance they resembled younger normal rats of the same size. Numbers of these animals were killed and subjected to careful post-mortem examination. In general the organs appeared normal, apart from the bony structures, which usually showed defective ossification, and the reproductive organs which though showing no great histological abnormality appeared to be functionally deranged. This surprising normality from a histological point of view made it imperative to seek some other cause of sterility than primary absence of gametes, and to this end further animals were prepared on the deficiency diet for examination of the œstrous cycle and mating instincts.

Since this work involved mating deficiency animals with normals on a complete diet some difficulty was experienced in arranging the feeding. The normals, however, were maintained on a complete diet by individual administration of cod-liver oil.

Other animals, after being given full opportunity for breeding on the restricted diet, were given either a normal mixed ration of cereals, meat and fresh vegetables, or an artificial ration similar in composition to that described above but supplemented with 2 per cent. of cod-liver oil to supply the fat-soluble vitamins A and D.

The growth of these animals recuperated on diets complete for growth requirements is discussed in Section II. of this paper.

On arriving at adult size the animals whose growth had been retarded showed no appreciable difference from others which had had normal food throughout, except that whereas those recuperated on natural foodstuffs reproduced in a normal manner those given the basal diet supplemented by cod-liver oil showed irregular and unsatisfactory breeding. This is readily explained in the light of the researches of Evans and Bishop, for the artificial ration even when supplemented with cod-liver oil would not supply a sufficiency of the reproductive vitamin E which is contained in the natural diet.

For this reason we have not attempted an analysis of the breeding in the second group, but have restricted our attention to those animals that recuperated on the natural foods.

The curves in Fig. 1 are typical of a large number collected during the course of this research and illustrate in a clear manner both the curious stunting that may occur on the diets that have been described, and the very sharp recovery of growth that follows the making good of the dietary deficiency responsible for the previous failure to grow. The case of Rat 5952 is interesting as an example of maintenance at almost constant body weight over a period of about four months. This rat finally died as a result of the deficiency at an age when had it been normally fed it should have weighed over 200 gms. and have reared litters of young. The recovery of the stunted Rats 6041 and 6093 illustrate the striking manner in which growth follows the correction of the dietary balance ; the rate of recovery of the former animal is distinctly greater than the normal rate of growth for male rats illustrated in the first curve.

At the end of the deficiency period when the animals had been on the insufficient diet for periods varying from two to six months, and when they were, therefore, at an age when breeding should normally have taken place quite freely, no litters had been produced, in spite of the sexes having been kept together. Considering the generally stunted condition of the animals this sterility was not surprising, but its exact nature was not apparent.

Complete sterility such as exhibited by these animals may be of three general types :—

• (a) Inability of one or both sexes to produce gametes.

• (b) Inability to effect fertilisation (lack of sex instinct or malformation of organs).

• (c) Inability on the part of the female to perform gestation or non-viability on the part of the fœtuses.

It was therefore endeavoured to determine the nature of the A deficiency sterility in the light of this classification. In order to investigate whether the first of these causes was operative 10 of the deficiency animals, 6 females and 4 males, were killed at the time when the rest were put on the recovery diet. The post-mortem macroscopic findings for each individual are recorded below.

Rat 5915. This rat was five months on “A” diet. Except for a large amount of fat the animal on dissection appeared fairly normal, as though a normal female adult had been reduced to a smaller scale. The ovary, however, appeared to be large, out of proportion to the rest, approaching that is, normal size. The vaginal orifice was open. Microscopically the ovary showed both corpora lutea and follicles while the uterus had a typical condition, except that an epithelioma was developing. The uterine glands were present in about normal numbers, though they appeared to be enlarged, probably owing to the tumour.

Rat 5958. This female was put on the deficiency diet when weighing 46 gms. From this figure the weight increased gradually to 95 gms. at the time when the post-mortem examination was made four months later. During the middle of this period the animal had very small amounts of cod-liver oil for six weeks. It was $512$ months old when killed, and appeared normal except for its small size. The body was well nourished, and large amounts of fat were present, especially on the mesentery and broad ligament. The uterus and ovaries appeared normal. On sectioning, both mature follicles and corpora lutea were found in the ovary, and the uterus was found to present a typical transverse section (Plate I., Fig: 1a).

Rat 5963. This rat weighed 52 gms. when put on the diet at 5 weeks old. In two months the weight had risen to 80 gms. and in four months to 122. During the middle of this period small amounts of codliver oil were given for six weeks. In this animal again a large amount of mesenteric fat was present, and the uterus was again surrounded with fat. The uterus itself was rather thin, but it appeared to be otherwise normal. On sectioning, the muscular coat was found to be very thin, but the stroma and glands had a normal appearance. At the time of dissection the ovary appeared to be normal, and this was confirmed when both corpora lutea and follicles (mostly immature, however) were found in the sections.

Rat 6008. This rat was four months on “A” with small amounts of cod-liver oil for six weeks. When killed it weighed 118 gms. Externally this rat again appeared fairly normal, though on a small scale. On opening the body cavity an enormous amount of fat was found. The uterus was small while the ovary was normal. On sectioning the latter was found to contain both follicles and corpora in large numbers; the former, however, was subnormal, having a poor muscular covering and few glands.

Rat 6119. This rat showed very little growth. When put on the deficiency it weighed 56 gms. One month later it weighed 57 gms., and it was then given a subcutaneous injection of the unsaponifiable matter from cod-liver oil in paraffin. Five weeks later at 14 weeks old, when weighing 63 gms. it was killed. The vaginal orifice was open, and the clitoris was abnormally large. There was much mesenteric fat, and the ovary was fairly large; but this was apparently caused by a large number of follicles, a diagnosis confirmed by sectioning (Plate II., Fig. 2). The remarkable feature of this animal, however, was the uterus. When the animal was opened some difficulty was experienced in finding the organ, and finally it was located as a mere thread amongst the fat in the body cavity. The microscopical appearance was no less striking. In Plate I., Fig. 1 b a section of this uterus is shown, together with that of Rat 5958, which is to all appearances normal. The two photographs have approximately the same magnification, and the difference is obvious. The uterus of Rat 6119, in addition to being on a pigmy scale, has no glands (actually one minute one may be seen in the illustration) and the muscle layer is undeveloped. It has, in fact, the appearance of the uterus of a young rat of about 5 to 6 weeks old. Although the ovary had no corpora lutea reference to the section figures shows at least one follicle almost in the act of rupturing, and the ovary generally as a whole has a normal sort of development. The discrepancy in this rat between the ovary and the uterus is remarkable, but it should be remembered that in this case deficiency, though short, was very severe.

Rat 6122. The history of this rat was very similar to that of Rat 6119. After two months on the diet it was actually lighter than when put on, in spite of subcutaneous injections of unsaponifiable matter from cod-liver oil The animal was peculiar in being nearly hairless. The vaginal orifice was open, and not very much fat was found in the body cavity. The ovary had a normal appearance and was found to contain corpora lutea and Graafian follicles. In this case again the uterus was noteworthy. The organ was turgid and distended with fluid, and had in fact the appearance figured by Long and Evans as that typical of œstrus in the rat. That this might be the case is supported by the condition of the ovary. The section figured (Plate II., Fig 3) shows portions of four mature follicles in the act of rupturing, and the condition of the corpora lutea suggests that they are in involution.

The results obtained from the individual dissections of females may now be summed up and discussed.

From the findings recorded above no doubt whatever can be entertained that the ovary develops and proceeds to its normal function of ovulation in spite of the general dwarfing of the animal due to fat-soluble vitamin deficiency. Out of the six ovaries examined all had either corpora lutea indicating a previous ovulation, or mature follicles, while four had both. Furthermore, the ovaries, though undersized for normal animals of 4 to 6 months old, appeared to be less dwarfed than the animal generally, that is to say were large in proportion to the size of the animal. This observation is confirmed by the fact that though the body weights of the recovered animals increased 200 to 300 per cent., the ovaries from these recovered animals were very little larger than those previously removed from the deficiency animals.

The interstitial function of the ovary appeared to have been uninterrupted. There is indeed no real reason why it should have been, since the general endocrine mechanism of the body seemed to be in order.

The findings recorded above dispose effectively of the possibility that non-production of ova might be the cause or a cause of the observed sterility.

The case of the uterus is less well cut. If the uterus of Rat 6119 had not been examined the generalisation might have been made that, like the ovary, the uterus suffers less from dwarfism as a result of the deficiency than the rest of the body. The uterus of Rat 6119 cannot, however, be reconciled with this generalisation. Not only is the uterus stunted, but an actual arrest of development seems to have taken place, which is remarkable, because the general stunting associated with A deficiency is strictly dwarfing, merely an arrest of growth, and is not an arrest of development. The uterus under discussion undoubtedly belongs to the latter category, and would be quite compatible with cessation of the endocrine function of the ovary at the age of about 6 weeks. Some such alternative explanation would be very convenient, as it would bring into line the A deficiency results ; but it is difficult to see how cessation of the endocrine function of the ovary at 6 weeks old can be reconciled with mature follicles at $312$ months old. The uterus of Rat 6119 remains, therefore, an anomaly, and in summarising it can only be said that the general appearance of the uteri of the deficiency animals was fairly normal and did not suggest any grave obstacle to implantation, especially as one uterus appeared to be in a state of œstrus. Since œstrus is usually coincident with the appearance of the mature Graafian follicles it may be supposed that the œstrous cycle was intact in these animals, provided that the general physiological condition of the rat was not incompatible with the production of the ordinary manifestations of the ovarian stimulus.

The question of gametogenesis in the male way may now be dealt with.

Rat 5918, male, 5 months on “A.” The testes were descended, and the scrotum was large, out of proportion to the size of the body. The mesentery was covered with fat, and the seminal vesicles were very small. On sectioning the testes were found to be fairly normal, except that some of the tubules contained cores of debris. Spermatozoa were present in all the tubes.

Rat 5919, male, 5 months on “A.” The testes were descended and were fairly large for the size of the animal. On sectioning the seminiferous tubules were found to contain spermatozoa, and were generally fairly normal, and, though some of the tubules appeared to have cores of necrotic matter,^* all of them contained spermatozoa.

Rat 5659, male, $612$ months on “A.” The hair on the underside was thin. The testes were descended and the penis was well developed. Dissection showed that mesenteric and perinephritic fat was plentiful The vasa deferentia were well developed. The testes showed much the same condition as was found in Rats 5918 and 5919 (Plate III., Fig. 4).

Rat 6078, male, $212$ months on “A.” This animal bad had subcutaneous injections of unsaponifiable matter from cod-liver oil in oil. The testes were not descended. On dissection the animal had a well-nourished appearance, and an amount of body fat was present. The testes, however, had an undeveloped appearance and microscopic examination revealed a most peculiar condition of the tubules. Photographs of a section of the right testes are shown in Plate III., Fig. 5a and Plate IV., Fig. 5b. In general the section had the appearance of one from a testes of a 5 to 6-weeks old rat. In conjunction with the non-descent of the testes this would seem to point to another case of actual arrest of development. The amount of interstitial tissue, however, seems to be fairly normal. It should be noticed too, that the testes tubules have far from the exact miniature of the immature gonad, especially as regards the abnormal heavily stained bodies, which appear to be picnotic nuclei. It is quite possible, however, that the abnormal retention of the testes in the abdominal cavity accounts for the peculiar condition of the tubules. The undescended testis is usually degenerate as far as the spermatogenic function is concerned. If this view is correct the abnormality cannot be directly ascribed to vitamin A deficiency. This question of non-descent of the testes will be mentioned again later, but the four examinations recorded above warrant the provisional statement that the young testes proceeds to its function of spermatogenesis except when the secondary complication of non-descent appears.

It is quite clear, therefore, that primary sterility, ^† inability to produce gametes, was not the cause of the failure to breed of the animals on a diet deficient in fat-soluble vitamin. There thus remains the possibility of continued reabsorption of fætuses, and the possibility that fertilisation was never effected. The former may be considered first. Since the animals were so stunted it is scarcely conceivable that they could bear a litter to full time and, at first sight, this hypothesis seems not unlikely, especially as Evans and Bishop have shown that this result is commonly brought about by feeding on synthetic diets deficient in their reproductive factor. Reabsorption, however, would presumably not take place during the first week of pregnancy ; ^* and yet no signs of pregnancy were ever observed on dissection, and no indications of even the earliest stages of pregnancy were found in any of the uteri examined microscopically. It is possible, of course, that some abnormality of the uteri prohibited implantation, but no evidence of this was found except in the case of Rat 6119. By far the most probable explanation of the sterility is that although the gametes were produced, fertilisation was never accomplished.

Since no anatomical malformations which might account for this were found, it would seem that owing to debility of the animals, or owing to lack of sex instinct, coitus was not performed. This hypothesis agrees with what is known of the habits of these animals. Although coitus is extremely difficult to observe in the rat, yet a normal buck will usually take immediate interest in a strange doe. Such behaviour was not noticed in these “A” deficient animals. At the same time the amount of interstitial tissue in the gonads appears to be normal ; but there are three facts which suggest that the endocrine functions of the gonads might be disturbed :—

• (a) The seminal vesicles in the males were all very small.

• (b) In one case the testes were undescended.

• (c) In one female the uterus was infantile.

(a) may of course be explained alternatively on the grounds of the low condition of the animal. The importance of the second point depends on whether or not permanent cessation of the endocrine function of the ovary at 5 to 6 weeks old would be compatible with the production of mature ova at a later date, and in the absence of definite information a negative answer seems to be indicated. In the male in which the testes were undescended the azoospermia of the tubules cannot be considered to be very significant since this condition is usually found in the abnormally retained testes.

In the majority of these animals, therefore, there is no direct evidence that the endocrine function of the gonads was disturbed. We are thus lead to the conclusion that fertilisation was not effected owing to the debility of the animals nullifying their normal sex instincts, though even if fertilisation had been achieved a normal gestation would very probably have been impracticable. These animals, therefore, appeared to furnish a case of secondary sterility. The accuracy of this a priori deduction was fully demonstrated by the work described in the next section.

Since the histological examination of the gonads of the first group of animals suggested that the cause of failure to breed must be sought in functional rather than in morphological abnormality, it was thought advisable to investigate the æstrous cycle and mating instincts of these deficiency animals by means of Long and Evans vaginal smear technique (3) and by reciprocal matings with normal animals. The vaginal smear method of detection of the œstrous cycle depends on the fact that the uterine and ovarian cycles are paralleled by cyclic variation in the cellular contents of the vagina. Since the vaginal contents can readily be obtained for microscopical examination by means of a suitable spatula, this method is extremely convenient for working out the æstrous cycle in the intact animal.

As a result of Long and Evans’ classic research it is possible to work out the œstrous cycle of any rat by means of making smears of the vaginal contents. These authors, by means of dissections and histological examinations, have correlated the variation in the vaginal contents with the cyclic variation in the ovaries and uterus. They divide the complete cycle into five stages :—

Vaginal mucosa tends to be dry. Smears of epithelial cells only. Vulva a little swollen. In heat towards the end of the period. During this period the follicles are maturing and the uterus begins to undergo proœstrous congestion and distension.

Cornified cells only are found in the vagina and the animal is in heat. The Graafian follicles complete their maturation, while the uterus reaches its greatest distension and then begins to retrogress.

Cornified material is very abundant in the vagina, but the animal is not in heat. Ovulation occurs during this stage.

Vaginal contents consist of cornified cells and leucocytes. Young corpora lutea develop in the ovary. Ova in Fallopian tubes.

This stage is the dioestrous interval. Vaginal smear of leucocytes and epithelial cells. Uterine epithelium undergoes regeneration.

The duration of the first stage is about twelve hours, of the second and third stages combined twenty-seven hours, of the fourth stage six hours, while the diœstrous interval extends over about sixty hours. The complete cycle in the unmated cycle thus extends over about five days. The comparatively short duration of the first four of these periods makes it clear that very frequent examinations are necessary to detect each stage of the cycle (Long and Evans’ original observations were made at 3-hourly intervals); but for the sole purpose of ascertaining whether or not an animal is exhibiting cyclic activity such frequent examinations are not necessary, and for present purposes examinations every morning was considered to be sufficiently accurate.

For comparison with the vitamin A deficiency rats certain other features of the normal rat need to be mentioned. Since the experimental animals were put on the deficiency diet just after weaning time they reached the age of maturity while still on the deficiency diet, and notice was taken of the first signs of sexual maturity. Long and Evans state that the averages of their very numerous observations were as follows :—

As far as our own somewhat meagre observations go these data appear to apply to our own stock animals, but a word of caution may be said here with reference to assuming that such kind of data obtained on animals in one continent necessarily apply to a quite different strain of animals kept under different conditions in a different country. One very striking example of discrepancy may be mentioned. Long and Evans state that in their rats the average period of gestation is just short of twenty-two days, and this figure was originally assumed to apply to our own stock. Recent investigation, however, has shown that in our stock twenty-four days normally elapse between the discovery of the vaginal plug and the finding of the litter, though some few cases of twenty-three days have been found.

The animals intended for the œstrous cycle work were treated essentially like the batch discussed in the previous section, with the exception that the small amounts of vitamin given to maintain life were reduced to the absolute minimum and as a result many of the animals died at an early age. Nevertheless six females were reared to an age when they would normally have been mature and smearing was performed on five of these. In the case of the sixth the vaginal orifice for some unknown reason failed to open. The individual records of these does are as follows:—

Rat 7585. This doe was put on to the deficiency diet when weighing 56 gms. at 5 weeks old. At 10 weeks old the weight was steady at about 75 gms. and the vaginal orifice opened at 11 weeks. At 3 months old she was mated with a normal male, but the mating proved quite sterile. Vaginal smears were resorted to a month later and were continued for twenty-four days, at the end of which period the animal was in very low condition and was killed. Except for a slight gain following an administration of a small dose of vitamin A at about the time the smearing began the weight remained constant, between 70 to 80 gms. from 6 weeks old until its death at 5 months old. During the twenty-four days of vaginal examination three complete cycles and two incomplete cycles were observed. The three complete cycles averaged four to six days in length, but the last unfinished cycle stretched over nine days. The completed cycles were fairly typical vaginal cycles, and would have been indicative of normal periodic œstrus in a normal animal. In spite of this manifestation of æstrus activity, however, no copulation took place, although the animal was mated all the time with a normal buck who was known to be fertile.

It is clear that as far as what may be termed the histological aspects of the æstrous cycle are concerned this animal had normal cycles, in spite of the fact that it was less than half the size of a normal animal of the same age. Autopsy revealed conditions identical with those of the first batch of animals which were used for post-mortem. The body cavity was very full of fat, but the organs bad a healthy look. The ovaries were large for the size of the animal. Microscopically, the ovaries had a perfectly normal appearance, large numbers of follicles in various stages of maturation being present, as well as a number of retrogressing corpora lutea. The uterus showed a well-developed condition, the muscular layer and the glands being well pronounced.

Rat 7561. This animal was put on the deficiency diet when 57 gms. (5 weeks old), by the end of a month it had reached 90 gms., and it was kept pretty constant at this weight until the fatal decline which would have ended, in death set in some ten weeks later. The vaginal orifice opened during the eleventh week, i.e. about the normal time. During the last month vaginal smears were taken and three complete, though irregular, cycles were observed; but although mated with a normal male the whole time no copulation was observed.

There was a short 4-day cycle sandwiched in between two long 9-day ones, but the æstrous cornification was well marked in all three instances. The first two cycles occurred at the end of the period during which the weight remained constant, but during the third cycle the final decline in weight set in. A small dose of vitamin A was given on the seventh day of this cycle, the condition of the animal being very bad; but little improvement was shown, and soon after the fourth cycle began the animal was killed At the time of killing the rat was 5 months old and weighed just under 80 gms., this weight being only two-fifths of the weight of normal animals of that age (about :200 gms.). Nevertheless, the evidence clearly points to cyclic activity on the part of the reproductive organs, but as in the case of Rat 7585 psychic manifestations were entirely absent. Histological examination of the ovaries showed that both corpora lutea and follicles were present.

Rat A2. This rat was put on the deficiency diet at 5 weeks old when weighing 50 gms. At end of the first month on the diet it weighed 60 gms. and the weight was kept pretty constant, between 60 to 70 gms. for the remainder of the period of observation. The vaginal orifice opened at 10 weeks old and shortly after it was mated with a normal male and vaginal smears were commenced. The end of one and the beginning of the next cycle was observed but the animal then died suddenly. Examination of the ovaries (so far as their post-mortem condition permitted) showed that mature follicles were present but no corpora lutea. The uterus, however, was poorly developed, both the number of glands and the muscular layer being deficient.

Rat A6. This rat weighed 53 gms. when dieting was started. After a month of deficiency diet the weight had risen to 70 gms. and no variation of more than 2 gms. from this figure was recorded before it was killed when 15 weeks old. The vaginal orifice opened during the 11th week. During the 12th week it was mated with a deficiency buck (Rat 7592) and vaginal smear examinations were started In the course of the twenty-seven days during which these examinations were made the complete cycles and two portions of cycles were observed, the two complete ones lasting eight and seven days respectively. As with the other females, however, no copulation took place, and this rat therefore agrees perfectly with the previous cases in having a normal cyclic activity of the reproductive organs without the psychic complement.

On sectioning it was found that the ovaries contained large corpora lutea, as would be expected from the œstrous cycle history given above, but only semimature follicles were present. The uterus was not fully developed, but was far from being as immature as in the case of Rat A₂. The ovary is shown in Plate IV., Fig. 6.

Rat A7. This was the last of the does which survived deficiency diet for any appreciable length of time in the second batch. It weighed 50 gms. at the time when it was put on the diet and was the same age (5 weeks) as the preceding ones. The weight became steady between 65 and 70 gms. after the first month, and the vaginal orifice opened during the eleventh week. At three months old it was mated with a deficiency buck (Rat 7488) and vaginal smears were started. After twenty-three days of observation, however, an acute attack of diarrhœa set in and a day later the animal was killed. During these twenty-four days one very long and one fairly long cycle were observed, but in neither case did copulation occur. Autopsy showed no very abnormal feature beyond general dwarfing-the general appearance was as already described. On sectioning the ovaries were found to contain large mature follicles, a finding in keeping with the fact that the animal was killed during what the vaginal smears indicated to be an early stage of œstrus. The uterus had a fairly normal appearance.

The histological examination of the first batch of deficiency males (discussed in Section III.), showed it to be highly improbable that the endocrine function of the gonads was disturbed by vitamin A deficiency, and the further cases to be discussed support this view.

Rat 7488. This rat weighed 47 gms. when first put on the diet at about 5 weeks old, and the weight was eventually stabilised between 90 to 110. Young male rats start to copulate when weighing about 150 gms. at some 3 months old. It weighed 100 gms. at $412$ months old (when a normal male rat weighs about 230 gms.), and was then mated with a normal female which was known to be fertile. This female was examined daily for two weeks, but no signs of copulation were observed, although three œstrous periods were experienced during the time. The male was then mated with Rat A7 (described above), but again no copulation took place. When the male was dissected at $512$ months old the testes were found to be very much smaller than normal for a rat of that age, but the size was not greatly out of proportion to the weight of the deficiency animal. Microscopical examination of the testes revealed nothing to suggest that the buck would have been infertile had copulation occurred.

Rat 7594 At weaning this rat was well grown and when put on the deficiency diet at 5 weeks old its weight was 51 gms. On this diet (with occasional doses of cod-liver oil) the weight increased slowly but steadily up to 100 gms. at $412$ months old and became steady at about this figure. It was then mated with a normal doe which was known to be fertile, but although the doe had six œstrous periods during the time no copulation took place. At $512$ months old it was killed during a bad attack of diarrhœa. Microscopically the testes, though descended, were much smaller than in the case of normal adults, and were somewhat discoloured. Microscopically, it was found that spermatogenesis was taking place, but the number of spermatozoa present was sub-normal.

Rat 7592. The early history of this animal was very similar to that of Rat 7594, the weight reaching 100 gms. at 4 months old. At this time the testes were undescended. The weight appeared to have become fairly steady when a bad attack of diarrhœa set in, and in order to give the animal a chance to survive cod-liver oil was administered over a period of a week. This resulted in a rapid increase in weight up to 115 to 120 gms., a figure which was maintained subsequently. This increase in growth and improvement in the general condition resulted in the testes becoming descended. It was mated with a normal doe when weighing 115 gms., and two days later it was found to have copulated, but no pregnancy ensued. Immediately after this it was mated with Rat A6 (described above) for three weeks but no copulation took place. Since the male did not appreciably decrease in weight during the time of mating with A6, and since A6 experienced the histological features of œstrus, the failure to copulate during this mating must be put down to absence of psychic impulse on the part of A6.

On dissection male 7592 was found to have only small testes, and on sectioning the seminiferous tubules were found to show precisely the same condition as found in the case of Rat 6078. Since “A” deficiency does not cause degeneration of the descended testes, it seems highly probable that this condition of the tubules must be put down to the non-descent of the testes, rather than to the “A” deficiency. It is also worthy of note that in Rat 7592 the interstitial tissue shows relative augmentation, which is also a condition normally associated with failure to descend to the scrotum. The degeneration of the tubules provides ample explanation of the failure of the normal doe to become pregnant after copulation with this buck.

The condition of the tubules of Rats 5659 and 7592 make it clear that to the general conclusion that “A” deficiency does not inhibit spermatogenesis should be added the proviso that where “A” deficiency results in non-descent of the testes such inhibition may be indirectly brought about.

The individual records given in the preceding paragraphs of this section make it clear that in addition to having normal ovaries with normal follicles and corpora lutea (indicating ovulation) the deficiency females also show the histological manifestation of the œstrous cycle, a finding which was originally foreshadowed by the œstrous appearance of the uterus of one of the female rats of the first batch (Rat 6122). The occurrence of this cyclic activity in the generative organs of rats kept at a weight corresponding to 6 to 7 weeks of age is remarkable, especially in view of the complete cessation of growth activity in the body as a whole, and suggests that the periodic hypertrophy of the internal genitalia is independent of the presence of the vitamin essential for normal growth in the body tissues generally, or else that the genital organs had a preferential call on such supplies of the vitamin as were available. In addition it should be remembered that the external sign of the onset of sexual maturity (the opening of the vaginal orifice) appeared at the normal age. The whole evidence seems to show that in these dwarfed animals the female genital organs develop and proceed to their cyclic activity according to the normal time schedule, quite independent of the general growth of the body. The solitary piece of evidence against this view is the infantile state of the uterus of Rat 6119, and this solitary instance of arrested development is of no great consequence when considered in relation to the numerous cases of complete development.

Since, however, no copulation occurred even with normal adult males it is clear that the complete phenomena of the œstrous cycle do not occur in these deficient females,^* the willingness of the female to receive the male apparently being inhibited. Since the ovarian hormone responsible for the cyclic changes in the uterus and vagina is also responsible for the psychic manifestations of æstrus (i.e. willingness to receive the male) as shown by Allen and Doisey, these phenomena should recur together given a suitable basis for the hormone to work on. Since animals so thoroughly debilitated as these under discussion must be most unfavourable subjects for the production of psychic excitement of any kind, it may, we think, be confidently postulated that the failure of the deficiency does to copulate with normal males is due to the debilitated condition of the deficiency’ females inhibiting the willingness to copulate, which in normal animals is correlated with the cyclic hypertrophy of the uterus and vagina.

This failure of the does to copulate is in itself sufficient cause of the sterility of matings of deficiency animals with each other, but the cases discussed in paragraph (c) of this section seem to show that males of any but a very mild deficiency will also not copulate. That is to say, if the dwarfing of the animal due to vitamin deficiency is at all pronounced copulation will not take place with normal females. Rats 7488 and 7594 which were only moderately dwarfed between them allowed nine œstrous periods to pass. in normal fertile does without copulating, so it may reasonably be supposed that they had no inclination to copulate. Since the accessory organs (penis, seminal vesicles, vas deferens) were fairly well developed it is necessary to suppose that the internal secretory function of the testes was intact,^* so here again it must be imagined that the debility of the animal had inhibited the usual psychic manifestations of the sex hormone. ^†

This conclusion is supported by a further point. Rat 7592 progressed from a moderately dwarfed condition (when it would presumably not have copulated) to a merely slightly dwarfed state (when it did copulate) in little over two weeks, and it is much easier to suppose that the general physical condition of the animal improved sufficiently during this time to make possible the manifestation of the sex instinct than it is to imagine that the mechanism of testis secretion was set up, or at least set going, during the time. It is possible, therefore, to sum up the cause of the sterility resulting from vitamin A deficiency as disinclination of both sexes to copulate due to inhibition of the sex instincts by the physical debility of the animals. In addition, this conclusion effectively disposes of the possibility that the sterility might be due to prenatal death and reabsorption of fætuses.

After the animals described in Section III. had been killed and examined, a number of the remaining rats were allowed to recover on a full natural diet. The growth curves of the animals during the recovery period have been described in an earlier section. When a normal breeding size had been attained, after two to three months on natural food, the animals were mated. The matings were made in three ways:—

• (a) Recuperated females with normal bucks.

• (b) Recuperated males with normal does.

• (c) Recuperated does with recuperated bucks.

Of the recovered animals mated, two, one male and one female, proved sterile even when mated with rats of the other sex which were known to be fertile. Since thirty-two recovered animals in all were used, this is not an abnormal proportion, and it is clear that recovery from the deficiency must have been quite complete. The breeding records are summed up in the following table :—

The normal matings in the colony gave approximate equality in the number of the two sexes produced, so that of these three groups only the recovered male crossed with recovered female gives a sex-ratio significantly abnormal. The relations of the sex-ratio given by the different types of matings to each other are as follows :—

From this table two conclusions may be drawn:—

• (1) That the difference between the percentages of groups A and B (recovered males x recovered females and recovered females x normal males) tends to be significant.

• (2) That the percentages for groups B, C, and D are definitely not significantly different from each other.

Superficially this is a somewhat curious result, since whether the males or females were responsible for the lowering of the male percentage in the first group, it would be expected that one or other of the reciprocal crosses with normals would show the same effect. Actually, however, the discrepancy is not insoluble. In a case of previous malnutrition of this kind it is not unreasonable to look towards prenatal mortality as a potential source of disturbance of the sex-ratio,^* and one would expect this factor to operate more strongly in the case of the first litters produced after deficiency than in those resulting from matings made when the animals had had some chance of establishing their normal reproductive function. It may be said, therefore, that the expectation would be for the first litters to have a low male percentage and a low litter size. Now of the three types of matings, recovered with recovered, recovered male with normal female, and normal male with recovered female, the matings with both parents recovered from the deficiency were the first to be made, so that most of the litters (actually 15 out of the 24) of this class were first litters. Actually, also, the litter size (6.2) in this class is below the normal for the colony. Taking first litters only in this class the litter size was only 6.0, and the male percentage of these first litters stands at 40.0. These findings seem to emphasise the point.

It appears possible, therefore, to draw the following tentative conclusions relative to the influence of A deficiency during early life on fertility and the sex-ratio after recovery :—

• (1) That when the adult weight is reached after recovery, neither the sex-ratio nor the size of litter is affected by the previous deficiency as far as the number or nature of the gametes themselves is concerned.

• (2) That both the male percentage and the size of litter may, however, be lowered by the increased mortality which accompanies first gestations following recovery.

Following the breeding experiments recorded above the animals were dissected. Microscopically the reproductive organs were quite normal, and this was confirmed by microscopical examination in the case of the female.

The ovaries examined were indistinguishable from those of normal animals. The ovaries of Rat 6003, which had three litters after recovery, had a large number of corpora lutea, some of which were in involution. Numbers of immature follicles were also present. The ovaries of Rat 6117, a section of which is shown in Plate V., Fig. 7, presented much the same condition, except that the follicles were rather larger. This rat had had one Jitter after recovery on natural diet. The organs of the three other females examined were so normal as to need no description. All had corpora lutea and follicles in various stages of development.

The testes of the recovered males were not so normal histologically as the ovaries, but in every case mature spermatozoa were present. Most sections showed a few tubules with a core of debris similar to that found in the tubules of the animals dissected at the end of the deficiency period. This condition is markedly reminiscent of certain stages in the development of the degeneration caused by deficiency of vitamin B (6). Plate V., Fig. 8 shows both normal and abnormal tubules in the testis of Rat 5966, an animal which had fathered two litters after recovery.

Since all the recovered males and all but one of the recovered females were fertile, it is clear that the sterilising effects of vitamin A deficiency are in no way permanent, and this supports the conclusion already arrived at, that the sterility while on deficiency diets is purely secondary and physiological.

The expenses of this research were defrayed mainly from a grant from the Medical Research Council (J. C. D.), but also in part from a grant from the Royal Society (A. S. P.).