Talk:Paper - The free-martin a study of the action of sex hormones in the foetal life of cattle (1917)

From Embryology
Revision as of 10:51, 27 May 2020 by Z8600021 (talk | contribs) (Created page with "5 The stimulation of the integumentary nerves of fishes by Light. Amer. Jour. Physiol., vol. 14, pp. 413-420. Pearse, a. S. 1910 The reactions of Amphibians to light. Proc....")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

5 The stimulation of the integumentary nerves of fishes by Light.

Amer. Jour. Physiol., vol. 14, pp. 413-420. Pearse, a. S. 1910 The reactions of Amphibians to light. Proc. Amer. Acad.

Sci., vol. 45, pp. 161-208. Reese, A. S. 1906 Observations on the reactions of Cryptobranchus and

Necturus to light and heat. Biol. Bull., vol. 11, pp. 93-99. Walion, a. C. 1916 The reactions of Paramoecium caudatum to light. Jour.

Anim. Behav., vol. 6, pp. 335-340.


THE FREE-MARTIN; A STUDY OF THE ACTION OF SEX HORMONES IN THE FOETAL LIFE OF CATTLE

FRANK R. LILLIE

Zoological Laboratory, University oj Chicago

TWENTY-NINE FIGURES

I. INTRODUCTION

In a preliminary paper (Lillie '16) the author has briefly discussed the theory of the sterihty which is the rule, subject to a few exceptions, in the female of two-sexed twins of cattle. In such females, commonly known as free-martins, the internal organs of reproduction are usually predominantly male in character, and the external organs are usually, at least, of the female type; there are however considerable variations as will appear. The conclusion was reached that the sterile free-martin is zygotically a female, modified by the sex hormones of the male twin, which circulate in both individuals during foetal life owing to secondary fusion of the chorions and anastomosis of the foetal circulation of the two individuals. The conditions thus established in this natural experiment enable us to draw far-reaching conclusions as to the origin and the differentiation of sex-characters in mammals. A somewhat extended treatment of the data, and analysis of the facts is therefore herewith presented.

Previous studies have dealt mainly wdth the anatomy of the free-martin and have furnished important, though incomplete data on this part of the subject. These will be considered in their appropriate connection. A study of L. J. Cole's ('16) supplies som€ important statistical data referred to beyond. We shall introduce the subject by a consideration of the various theoretical points of view from which the discussion must start. In considering these it will be important to define more precisely the condition with which we are dealing. The phenomenon of sterility of the female of different sexed twins is usually consid 371


372 FRANK R. LILLIE

ered to be peculiar to cattle;^ it has been stated to occur in sheep rarely, but I cannot discover on what evidence. There has also been a superstitition,- which reappears in the press

' According to E. J. Davies ('13) 'hermaphrodites' are common in certain breeds of goats, for instance Toggenburgs and Anglo-Swiss. He even estimates that in 1913 "Among the total birth of kids eligible for entry in the Herd Book it is believed that at least 2 per cent belonged to the class under discussion^ — and the proportion is probablj' much greater." This statement is understood from the context to apply to the Toggenburg breed. He goes on to say that "malformed kids have recently come (a) singly, (b) as twin with a normal male, (c) as twin with a normal female, (d) as one of triplets the normal kids being male and female, (e) as one of triplets the normal kids being both males." Assuming the malformation to mean genital abnormalities we are justified in explaining cases b, d and e on the same principle as the free-martin of rattle. As regards case a we can only say that this phenomenon is not uncommon in other mammals, whatever may be the explanation; it could be explained as due to two-sexed twin association in utero with early death and absorption of the male twin. Case C, an 'hermaphrodite' twinned with a normal female, is of unusual interest, for we have no evidence that this condition ever occurs in cattle; but it is to be expected theoretically in any case of blood community of male and female embryos in which the female is decidcly in advance of the male in development. The male would then be subject to action of the female sex hormones, and its development would tend to be intersexual. The case cited in illustration of this association seems to be well authenticated. Unfortunately the anatomy of the malformed individual was not studied, as a kid it appeared as a female externally except for the much enlarged phallus; but it grew to male size and even developed very strongly the characteristic smell of the male goat.

" I do not know how wide spread this superstition may be; but that it still exists is undoubted. It has been reported to me as a common belief in certain farming communities, and it has been gravely discussed in the correspondence of the Chicago Tribune as recently as March, 1916. Formerly the idea was more wide spread, and was even countenanced by medical men as reported by Dr. James Y. Simpson of Edinburg who quotes from Burn's "Principles of Midwifery" 1843, p. 236, "It is a popular opinion, and I do not know any instance to discountenance it, that if twins be of opposite sexes the female is sterile." Simpson ('44) made an investigation of the subject which may be regarded as disposing forever of the superstitition. He investigated the family history of 123 married women born twin to males of whom 112 had families and 11 had none; he foimd also that this proportion of childless inarriages was not greater than in the general population. Ninetj^-four of these cases in which he had complete histories had 409 children, an average of 4.2, w^hich was about the same as for the general population of that time and place. He concluded that females born co-twin with males are, when married, as likely to have as many childreh as other females belonging to the general community. There is no basis for belief in the sterility of such females; the superstition works cruel hardship to innocent people,

iii(l it cannot be too Htinngly stated thai there i.'^ no l)asis whatever for it.


ACTION OF SEX HORMONES IN FOETAL LIFE 373

from time to time, that it applies to human twins, but this is certainly not the case. In cattle, in about 87 per cent of apparently different-sexed t^\dns, as nearly as I can ascertain, the female is sterile; about 13 per cent are normally fertile (see data beyond p. 381). We must therefore distinguish sterile and fertile free-martins, and it is important to note that the fertile individuals are not known to be in any respect inferior to other females in respect of breeding. The sterile individuals have the external organs of a female, usually, but the internal organs of reproduction are more or less of the male type. Their general bodily appearance is more or less intermediate between a male and a female — ^it has been compared to that of an ox or spayed heifer — so that an experienced cattle man can usually distinguish them from normal heifers. The bull twin is always, normally fertile, and does not exhibit any anatomical pecuharities so far as is known.

It is essential to recognize the fact that the sterile free-martin condition is found only in association with a bull twin.^ The

3 Numan ('43) is, so far as I know, the only author who has questioned this; his study is by far the most extensive, and in many respects the most thorough, that has been made on the free-martin. The publication is exceedingly rare, but I have been able to study a copy from the library of the Smithsonian Institution, and as his conclusions are so often quoted from author to author, it seems worth while to give the evidence on which his dissenting statements are based. (D. Berry Hart has published an abstract of Numan' s paper (Hart '12). Numan states "The anomaly occurs not only in twins of different sex, but also in female and male pairs, though more rarely." In these cases he refers first to an individual about two years old, judged by external signs alone to be a sterile freemartin; the owner stated that it was born twin to another female, which however was sold shortly after birth about two years previously. This would appear to be slender evidence upon which to base a unique exception, for it was not positively certain that the individual was a sterile free-martin, nor could Numan know unquestionably that it was born twin to another female. It is known also that cystic degeneration of the ovaries may lead to extensive assumption of male secondary sex characters in the cow (Pearl and Surface, '15); the case may belong in this category. This was the only case he had on the female side, and no others have been recorded since. In the case of male pairs he cites also a single case which is of great interest, but wrongly interpreted by him. It was a case of twins one being a normal male and the other a sexually abnormal individual. He judged the abnormal individual to be male on account of the presence of testes in the groins, and malformation of the external parts. The scrotum was absent. This is almost certainly an extreme case of modification of the

THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 23, NO. 2


374 ■ FRANK R. LILLIE

association with a male in utero is, therefore, in some way a necessary condition of the phenomenon. It is also an invariable rule that the male tmn is normal; the reverse condition of a normal female with a defective male probably does not occur. All of my own cases were found in association with a normal bull twin; in the other possible twin combinations both individuals are normal. Of course genital abnormalities may occur apart from tmnning, so that the possibility exists that in a sufficiently large collection of twins an individual might be found with genital abnormaUties not due to its association with a twin. In this connection Numan ('44) notes that malformation of the sexual organs causing infertility had not yet been observed in single born heifers in his experience; but, in single born males, individuals with incomplete formation of the organs of reproduction had

female similar to one of my own cases (no. 44), and hence no exception; it is discussed in extenso, p. 413 of the present paper.

Numan also states, "In two-sexed twins the malformation is not confined exclusively to the heifer; but may also occur in the case of the bull, in which case the heifer is normal. However such examples appear to be very rare." This also refers to a single case viz., that of a two year old bull sent him by a veterinarian who stated that it had been born twin to a female. According to the veterinarian there was not the least abnormality discernible in the external organs of the heifer; the internal anatomy was not studied. The bull exhibited a hypospadiac condition; it possessed a split scrotum and the testes were in the abdominal cavity against the inguinal rings. Numan himself points out that such abnormalities are not rare in l)ulls born single. Since the actual condition of the female twin was not known, and the condition in the bull is a rather common anomaly the evidence is entirely inadequate to support the idea that the condition was due to twinning.

Numan's emphasis of these doubtful cases appears to be due in part to his fundamental objection to considering free-martins as hermaphrodites. He classified them therefore either as females or as males with defective organs of reproduction; in two-sexed twins the female was classified as male if the modification of the reproductive organs proceeded beyond a certain degree, thus establishing in his mind the occurrence of a profound genital anomaly in one individual of exceptional male pairs; this seemed to render probable a similar occurrence in the opposite case of female twins, which may seem to account for liis uncritical acceptance of the case cited above, and his equally uncritical interpretation of modification of the bull as due to twinning with a female.

The entire basis for Numan's statements concerning exceptions to the rule is untrustworthy, and the exceptions cannot be accepted seeing that they are supported by no other writers.


ACTION OF SEX HORMONES IN FOETAL LIFE 375

often been found. This observation is of interest in connection with the problem stated in footnote 5, p. 389.

The theoretical interpretation of the free-martin must be based on one of two assumptions, either (1) that it and its partner, are identical, i.e., monozygotic twins, or (2) that they are fraternal, i.e., dizygotic twins. Under the first assumption, as the sex of monozygotic twins is undoubtedly identical, the freemartin would necessarily be interpreted as a modified male. Under the second assumption it would be almost equally a matter of necessity to interpret the free-martin as a female, for there is no possibility that the association of two males in utero should cause the transformation of one of them into a free-martin. On the first assumption the explanation of the modification must be found in the twdnning process itself, i.e., in the division of the single zygote that ex hyp. formed the two twins. But on the second assumption no one has hitherto attempted to explain how the association of male and female in utero could lead to sterility of the female with a more or less pronounced male organization of the internal organs of reproduction, nor why certain females should escape the defect, nor why the phenomenon should be peculiar to cattle. It is, therefore, natural that the first assumption should have been the one followed in all previous theoretical interpretation of the free-martin, and that it formed the working hypothesis wdth which my own work began.

The first theoretical view that we shall consider is that of Spiegelberg ('61) who said of cattle twins: "If the twins are both female or of opposite sex, the organs of reproduction are as a rule well formed; if they are both male, it very frequently happens that one of them is an hermaphrodite." This conclusion was based on the examination of two pairs of different sexed twins in cattle ; he made anatomical examination of the free-martins and found one of them a normal female, but in the other the female internal organs of reproduction were mostly absent, and w^ere replaced by rudimentary seminal vesicles, rudimentary vasa deferentia, and a rudimentary gonad on one side which he interpreted as probably a small testis with more or less separated epididymis. The internal organs were accepted


376 FRANK R. LILLIE

as diagnostic of sex, and the case was interpreted as simple male transverse hermaphroditism .

D. Berry Hart ('10) also interprets the free-martin as a male, basing: his interpretation on a comparison of the anatomical descriptions of thirty cases given in the literature, and on an original histological examination of the gonads of John Hunter's specimens which hacf been in alcohol for one hundred and forty years.

The special fact that emerges is that all the sexual glands are testes in Hunter's cases, that adjacent structures are epididymis, and that in none of the sexual glands are ova present. The characteristic testicular tissue is in the form of tubuli semeniferi, and in only one are spermatozoa present. It seems to me, therefore, fully established that the free-martin, when the co-twin is a potent male, is a sterile male, and not a sterile female, i.e., they are identical male twins except in their genital tract and secondary sexual characters.

It ^vill be observed that Hart accepts the conditions of the internal organs, and more especially of the gonad, as decisive criteria of sex. This raises a point that we shall discuss later on. Continuing, Hart then proposes a ' Mendelian' theory as follows : He distinguishes potent and non-potent elements in the genital tracts of both sexes, the latter being the undeveloped rudiments of the opposite sex; in the tmnning process of a male zygote he supposes we may either get identical male twins, or the potent and non-potent complex of the genital organs may be divided so that the potent part goes to the potent bull calf, the non-potent to the free-martin." More specifically, "A free-martin with a potent bull twin is the result of a division of a male zygote, so that the somatic determinants are unequally divided, the potent going to one twin, the potent bull, the non-potent genital determinants to the free-martin." He supposes the potent organs to be dominant in the Mendehan sense, the non-potent recessive.

The entire argument is based on the unsupported assumption, which it is quite possible to decide definitely by the facts, that the free-martin is co-zygotic with its male mate. I shall show immediately that this is not the case; so that it is hardly necessary to point out that if the gonad of the free-martin is a testis, as Hart


ACTION OF SEX HORMONES IN FOETAL LIFE 377

maintains, it can hardly be classed as a non-potent part of the genital tract, i.e., by definition undeveloped parts of the opposite sex, nor can the vasa deferentia of the free-martin be so classed; nor yet the external genital parts which are usually pure female. The theory moreover implies that as the free-martin receives the 'non-potent' genital parts, the bull twin must lack them, and must continue to propagate male individuals lacking them, for neither of which deductions is there the slightest evidence, or any attempt to produce evidence. Finally, there is not any attempt to explain why the twinning process should be attended' by such extraordinary results in cattle, and involve nothmg of the kind in other mammals. Hart notes that the theory implies the possibility of a similar defect in the twinning of a female zygote, and he refers to Numan's case cited before (footnote 2) as an example.

Bateson ('13, pp. 44-45) also attempts an explanation of the free-martin on the basis that it is co-zygotic with its twin.

For it is impossible to suppose that mere development in juxtaposition can produce a change of this character. It is conceivable that we should interpret it by reference to the phenomenon of gynandromorphism, seen occasionally in insects, and also in birds as a gi-eat rarity. In the gjiiandromorph one side of the body is male, the other female. A Ijullfinch for instance has been described with a sharp line of di\4sion down the breast between the red feathers of the cock on one side and the brown feathers of the hen on the other. In such cases neither side is sexually perfect. If the halves of such a gynandromorph came apart, perhaps one would be a free-martin.

The interpretations of Hart and Bateson are based on the theory that the free-martin and its twin are monozygotic, and they involve the conclusion that the free-martin is derived from a male zygote. They may be called anatomical interpretations because they are based exclusively on anatomical evidence. Cole has also come to the same conclusion on statistical grounds, which I discuss in the second part (pp. 380).

The literature on this subject is scanty, and it would serve no good purpose to continue with the few incidental citations that might still be made. Embryological e\ddence that would alone give a basis for correct interpretation has been entirely lacking;


378 FRANK R, LILLIE

my principal work has been to secure this evidence, which is herewith presented. The order of presentation is decided by the theoretical considerations, which obviously require that we should first of all determine whether we are dealing with a phenomenon of division of a single zygote, or with development of two zygotes in juxtaposition. Other considerations follow immediately from this, and the anatomical side is considered last. In conclusion the more fundamental theoretical questions come up for consideration.

II. ARE THE STERILE FREE-MARTIN AND ITS MATE MONOZYGOTIC

OR DIZYGOTIC?

The first question that confronts us, therefore, is whether the free-martin and its male twin are monozygotic or dizygotic? This question can be answered decisively only by the embryological data; contributing evidence may be furnished by study of the degree of resemblance of the twins to one another, and by a statistical study of the sex ratios in twin births in cattle.

1. The embryological evidence

There are two ways in mammals of deciding whether twins are monozygotic or dizygotic: 1) If there is a single chorion for both of the twins, this would usually be regarded as evidence of monozygotic origin. But a single chorion is not decisive evidence, because, though one would expect monozygotic twins to have a single chorion, yet it is theoretically possible that two separate chorions may fuse to form one. The monochorial condition may be primary or secondary. 2) The number of ova concerned in a pregnancy may be ascertained by the number of corpora lutea, which correspond accurately. If we should find two corpora lutea for all free-martin twins, the dizygotic origin would be proved, provided that the rule holds for cattle.

I have had the opportunity of examining a large number of twin pregnancies in cattle through the courtesy of Messrs. Swifl and Company of the Chicago Stockyards; and I wish to express my appreciation of their generous cooperation, without which this study could not have been made. The superintendent of


ACTION OF SEX HORMONES IN FOETAL LIFE 379

the cattle-house kept watch for uteri containing twins, and when they were not of too large size notice was telephoned to the department of Zoology of the University, and our collector, Mr. Adams, went over and brought the specimens to the laboratory. The collection has been going on for two years and a half, and 55 pairs of twins have been studied. For a long time most of the uteri were received with one or both ovaries missing, but recently special pains have been taken to secure the ovaries also attached to the uterus. A large proportion of the earlier records are therefore incomplete in this respect.

It was a great surprise to find that nearly all twins of cattle are monochorial; only two complete exceptions have appeared in 55 cases. The first case (no. 40) is of great theoretical interest, and will come up for detailed consideration later on. In a very few other cases the connection between the two halves was slight (see cases 8, 9, 10, 24, in table); generally it was broad and strong. At first I had expected to decide the question of mono- , zygotic or dizygotic condition by the monochorial or dichorial state; so that for a considerable period not much attention was paid to the question of the corpora lutea. Later attention was directed to this question and the unexpected result was reached that in all cases in which both ovaries were present each had a corpus luteum. The exact data are: 22 cases in which both ovaries were present; in all of these there was a corpus luteum in each ovary; 7 of these were cfcf,49 9, 10 cf^, and one too young for sex diagnosis; 11 cases in which only one ovary was present, 9 of which had the corpus luteum present, and 2 absent; the two latter were same-sexed, one pair of males, and one of females; they may have been monozygotic, but the missing ovary may have contained two corpora lutea in each case. In 22 cases both ovaries were missing or not recorded. If we consider only those cases, 22 in number,, in which both ovaries were present, there is no exception to the rule that cattle t\vins are dizygotic, using the corpus luteum as evidence for a separate zygote. This is a sufficiently large number to make it certain that the occurrence of monozygotic cattle twins is at least extremely rare. The free-martin condition cannot possibly be interpreted as a result of monozygotic twinning.


380 FRANK R. LILLIE

It may, however, be objected that we are relying: on a rule which has not been proved for cattle, viz: that in a single pregnancy only one corpus luteum is present. To settle this question examination was made of 81 uteri each containing a single calf; in 45 of these both ovaries were present, and in every case only a single corpus luteum was found; in 36 cases one ovary only was present and 18 of these contained a single corpus luteum each, the other 18 lacked a corpus luteum. There can, therefore, be no doubt that the rule holds for cattle. This conclusion necessitates the inference that the monochorial condition of cattle twins is secondary. In a later section we shall consider the question of the probable time and nature of the fusion of the two chorions.

The conclusion drawn from the embryological evidence that the heterosexual cattle twins are dizygotic is supported by two other important considerations, viz : lack of close resemblance between the bull and heifer of such pairs, and by the sex ratios of all cattle twins.

2. Concerning the degree of resemblance of such tunns

A careful study of this point remains to be made; but it is noteworthy that no one has recorded resemblances similar to those of identical twins in the case of the free-martin and its bull mate. My own somewhat limited observations lead me to the conclusion that the resemblances are no closer than ordinary fraternal likeness. One striking case was recently observed by me in pure bred Holstein-Friesian cattle in w^hich the freemartin was about half black and white in patches, and the male was almost entirely black except its forehead and legs below the knees.

3. Sex-ratio of cattle twins

On this point we have observations by Cole and myself. Cole ('16) finds in a study of records of 303 multiple births in cattle that there were 43 cases male twins, 165 cases two-sexed twins (male and female), 88 cases female twins, and 7 cases of triplets. This gives a ratio of about 1 d" d" : 4 cf 9 : 2 9 9 f or the twins, instead of the expected ratio of 1:2:1. Cole then states :


ACTION OF SEX HORMONES IN FOETAL LIFE 381

The expectation may be brought more nearly into harmony with the facts if it is assumed that in addition to ordinary fraternal (dizygotic) twins, there are numbers of 'identical' (monozygotic) twins of both sexes, and that while in the case of females these are both normal, in the case of a dividing male zygote, to form two individuals, in one of them the sexual organs remain in the undifferentiated stage, so that the animal superficially resembles a female and ordinarily is recorded as such although it is barren. The records for monozygotic twins accordingly go to increase the homosexual female and the heterosexual classes, while the homosexual male class in which part of them really belong, does not receive any increment.

Cole thus tentatively adopts the theory, which has been worked out most elaborately by D. Berry Hart, stated also by Bateson, and implied in Spiegelberg's analysis ('61), that the sterile freemartin is really a male co-zygotic with its mate.

Cole's figures represent the only statistical evidence that we have previously had on this subject. Let us follow his suggestion and take from the two-sexed class enough cases to make the male twins equal in number to the female pairs; this will be approximately one-fourth of the class, leaving the ratio 2:3:2 instead of 1:4:2. Which one of these is the more satisfactory sex ratio I leave others to determine; I msh only to point out the fatal objection, that according to the hypothesis, the females remaining in the two-sexed class are normal; in other words, on this hypothesis the ratio of normal free-martins (females co-twin with a bull) to sterile ones is 3:1; and the ratio would not be very different on any basis of division of the two-sexed class that wolild help out the sex ratio. Hitherto there have been no data from which the ratio of normal to sterile free-martins could be computed, and Cole furnishes none. I have records of 24 cases statistically homogeneous, 3 of which are normal and 21 abnormal. That is, the ratio of normal to sterile free-martins is 1 : 7 instead of 3:1.

This ratio is not more adverse to the normals than might be anticipated, for breeders' associations will not register freemartins until they are proved capable of breeding, and some breeders hardly believe in the existence of fertile free-martins so rare are they. *


382 FRANK R. LILLIE

My own records of 55 cases of bovine twins, all examined in utero, and their classification determined anatomically without the possibility of error, give 19cfcf':24cf9:ll 99 and 1 (no. 49) too young for determination. It will be observed that the sum of the one-sexed classes is 25 per cent greater than the two-sexed class; and the cf cf class is much larger than the 9 9 class instead of being equal to it, as it should be if males and females are produced in equal numbers in cattle. The material cannot be weighted statistically because every uterus containing tmns below a certain size from a certain slaughter house is sent to me for examination without being opened." Cole's material shows twice as many female as male pairs, and the two-sexed class is about one-third greater than the sum of the other two classes. I strongly suspect that it is weighted statistically; the possibility of this must be admitted, for the records are assembled from a great number of breeders. But, whether this is so or not, if we add the sterile free-martin pairs of my collection to the male side in accordance with Cole's suggestion, we get the ratio 40d'cf:3d^9:ll9 9, which is absurd. And if we take Cole's figures, divide his heterosexual class into pairs containing sterile females and pairs containing normal females according to the expectation, 7 of the former to 1 of the latter, and add the former to his male class, we get an almost equally absurd result (180 d" cT: 20 o^ 9 : 88 9 9). On the main question our statistical results are sufficiently alike to show that the free-martin must be interpreted as female.

Prof. Alexander Graham Bell has kindly furnished me with a catalogue of the lambs born from 1890 to 1914 in his well-known experiments on his Beinn Bhreagh Estate in Nova Scotia from which I have taken all the records of twin births, 139 in number;

  • The great preponderance of the d^cT over the 9 9 class in foetal cattle

twins of the sizes dealt with in this study appears to be real, though it must be admitted that the numbers are too small to make this quite certain. Cole's data on the other hand indicate a great preponderance of the 9 9 class over the d" cf class in cattle twins after birth. It may be t^at abortion, which is so frequent in cattle, is even more adverse to the males in the case of twins than in single births; it is conceivable that the difference is largely a question of viability, but other explanations are possible.


ACTION OF SEX HORMONES IN FOETAL LIFE 383

of these 38 were cf' cf , 67 cf 9 and 34 9 9 , thus an exceedingly close approximation of the expected 1:2:1 ratio. In his "Problems of Genetics" Bateson cites Bernadin (La Bergerie de Rambouillet, 1890, p. 100) as to the frequency of twin combinations in Merino Sheep, viz: 87 d" d" to 187 d" 9 to 83 9 9, which also approximates the expected 1:2:1 ratio. These statistics, therefore, also support the interpretation of the free-martin as female, for they show that the actual ratios of the distribution of sex among twins are as a matter of fact the expected ones in ungulates.

On the other hand in man there is a very significant and interesting departure from the expected ratio: Simpson ('44) collected statistics of 788 cases of human twins, the various sex combinations being d^ d^ 229, o^ 9 298, 9 9 261, thus very far removed from the 1:2:1 ratio. Nichols ('07) has made a very much larger collection of statistics with the following ratios d d^ 234,497, cf 9 264,098, 9 9 219,312. It is obvious that there is a very large disturbing factor here; this is almost certainly the factor of monozygotic twinning. As the two-sexed combination must be dizygotic, we may estimate the dizygotic d^ d^ and 9 9 combinations at one-half of the d^ 9 combination, on the slightly inaccurate basis of a 1 : 1 sex ratio of male and female zygotes. This would give 132,049 dizygotic d^ d^ and 9 9 t\vin pairs each, and the excess viz : 102,448 d^ d^ and 87,263 9 9 would represent the monozygotic twin couples. This is not very far from the proportion of monozygotic twin pairs among onesexed twin couples estimated by physical resemblance. If, then monozygotic twinning is the disturbing factor in the unexpected sex-distribution ratios of human t\Adns, we may argue from the fact that the ratios in cattle and sheep approach expectation that monozygotic twinning either does not occur, or is very rare in them, and this is confirmed by the embryological evidence. This matter is discussed more fully in Newman's book on twins ('17) to which, the reader is referred.


384 FRANK R. LILLIE

4- Discussion

The preceding considerations constitute an a;-gument that the free-martin is zygotically female, which may be summarized as follows: 1) The only basis on which it could be logically interpreted as male is that it is co-zygotic with its male mate, because it is impossible to suppose that the association of two males in utero should cause the transformation of one of them into a free-martin in a certain definite proportion of cases. But we have seen that the free-martin and its male mate arise from separate zygotes. From this point of view the free-martin must be interpreted as zygotically female. 2) The somatic resemblances between the free-martin and its mate are not of the order of identical twins. 3) The assumption that the free-martin is male leads to an absolutely incomprehensible sex-ratio, while the interpretation that it is female comes nearer fulfilling the expected sex-ratio. From this point of view also the free-martin is female.

The only argument that remains for its male nature rests on the anatomy of the internal organs of reproduction, which unquestionably are more or less of the male type. But, as the external genitals and the mammary gland are almost invariably of the female type, the argument from anatomy may be made to turn either way depending on what anatomical characters are recognized as diagnostic of sex. In a later section we shall go fully into the anatomical problems involved. Here it may suffice to say that the anatomical argument is necessarily inconclusive.

In what follows, therefore, we shall treat the free-martin as demonstrated to be zygotically female, and the question becomes how the association of a male and female in utero may so transform the female.


ACTION OF SEX HORMONES IN FOETAL LIFE 385

List of cattle tiriijs examined in iitero


10


11


12


35 cm. each

d" 23 cm. ^ 21.5 cm.

25.5 cm. 26.5 cm.

cf 24 cm.

^22.5 cm.

17 cm. each

d" 16.8 cm. ^ 16.3 cm.

About 20 cm.


d 26.5 cm.

(wt. 2 lb., 6

oz.) 9 23.3 cm.

(wt. 1 lb.,

6.5 oz.)

d 20 cm. 9 20 cm.


23 cm. each


17 em




16 cm




(f 30


cm.


(2


lb.,


10.5 oz.)


^ 28


cm.


(2


lbs.


13


.5


oz.)




Single

Single

Single

Single Single

Single

? (See note 1)


Single, narrow connection


Single, narrow connection

Nearly separated; united by narrow strand

Single


Single


MATERNAL OVARIES


Not observed

Not observed

Not observed

Not observed

Not observed

Only one ovary present. Contains corpus luteum

Both ovaries present. Corpus luteum in each

Ovaries absent


Ovaries absent


Ovaries absent


Ovaries absent


Ovaries absent


386


FRANK R. LILLIE


NUMBER


SEX


SIZE


CHORION


MATERN.\1, OV.VRIE9



6


9


1



13


1


cT 20 cm. (14


Single

Ovaries absent






oz.) <^ 20 cm. (14








oz.)




14


1



1


(f 27 cm. ^ 27 cm.


Single


Ovaries absent


15



2



17.5 cm. 18 cm.


Single


Both present. Corpus luteum in each


16


2




13 cm. 12.75 cm.


Single'


Ovaries absent


17


1



1


cri3.75 cm. ^^ 13.1 cm.


Single


One missing, other had 1 corpus luteum


IS


2




12.5 cm. 13 cm.


Single


Ovaries absent


19


1



1


cT 8 cm. ^7.5 cm.


Single


Ovaries absent


20


2




20.5 cm. 20 cm.


Single


One ovary absent. Other contains corpus luteum


21


1



1


cT 27.5 cm. ^26.5 cm.


Single


Ovaries absent


22


1



1


cf 24.5 cm.

^ ?


Single*


Both ovaries present. Corpus luteum in each


23


1



1


d" 18 cm. ^ 17.5 cm.


Single


One ovary absent. Other has 1 corpus luteum


24


2




21 cm. each


Single; narrow connection


Both ovaries present. Corpus luteum in each


25


2




11.2 cm. each


? Injured


One ovar\' absent. Corpus hiteum in other


20


1



1


o^ 12.5 cm. ^ 12.25 cm.


Single


One ovary absent. Other has 1 {!orpus luteum


ACTION OF SEX HORMONES IN FOETAL LIFE


387


N[MBER


.SEX


SIZE

42.5 cm. each


CHORION


MATERNAL OVARIES




6

2


9


^



27


Not examined


One ovary absent. Other has








1 corpus luteum



2S


2




17 cm. each


Single


One ovary absent. Other has








1 corpus luteum



29


2




15 cm. each


Single


Both ovaries present . luteum in each


Corpus


30


2




23 cm. 22 cm.


Single


One ovary absent. No corpus luteum in other


31



2



12 cm. each


Single


Both ovaries present, luteum in each


Corpus


S2


1



1


cf 18 cm. ^ 16.75 cm.


Single


Both ovaries present, luteum in each


Corpus


33


2




10.5 cm. each


Single


Both ovaries present . luteum in each


Corpus


34



2



25 cm. (about)


Single


Both ovaries present, lutevmi in each


Corpus


35


2




13 cm. each


Single


Both ovaries present, luteum in each


Corpus


36


1



1


d" 18 cm. 9^ 17.5 cm.


Single


Both ovaries present, luteum in each


Corpus


37


1



1


cT 16 cm. ^ 15.5 cm.


Single


Both ovaries present, luteum in each


Corpus


3S


1



1


cf 23.5 cm. ^ 22.5 cm.


Single


Both ovaries present, luteum in each


Corpus


39


2




18 cm. each


Single


Both ovaries present, luteum in each


Corpus


40


1


1



d^ 10.4 cm. 9 10.2 cm.


Two separate chorions


Both ovaries present, luteum in each


Corpus


41


1



1


cf 22.7 cm. ^ 21.8 cm.


Single


Both ovaries present, luteum in each


Corpus


388 .


FRANK R. LILLIE


XUMBER


SEX


SIZE


CHORION


MATERN.^L OVARIES



6


9


i


>.


42—


se


e


N


ote 5




43


2




10.75 cm.


Single


Not observed


44—


se


e


N


ote 5




45



2



34 cm. 31.25 cm.


Single


Both ovaries present. Corpus luteum in each


46


2




23 cm. (about)


Almost separate


Both ovaries present. Corpus luteum in each


47


1



1


6 22.75 ^ 22.25 cm.


Single


Both ovaries present. Corpus luteum in each


48



2



20 cm. (about)


Single


One ovary absent. No corpus luteum in other


49 50


1


?


1


1 .5 cm. each

d" 21 cm. ^ 21 cm.


Separate, but overlapping Single


Both ovaries present. Corpus

luteum in each Both ovaries present. Corpus

luteum in each


51


2




5 cm. each


Single


Both ovaries missing


52



2



18 cm. 19.5 cm.


Single


One ovary absent. Single corpus luteum in other


53



2



19 cm.


Single


Both ovaries absent


54


2




20.5 cm. 21 cm.


Single


Both ovaries present. Corpus luteum in each


55


2




13.75 cm.


Single


Both ovaries absent


56



2



12.5 cm.


Two chorions separate


Both ovaries absent


57


1



1


d" 19.25 cm. g' 18 cm.


Single


Both ovaries absent


ACTION OF SEX HORMONES IN FOETAL LIFE 389

^ Case 7 was received in my absence, and the entire uterus was placed in formalin; preservation of its contents was bad, and condition of chorion must be recorded as doubtful.

- Case 13 uterus injured by butcher; chorion cut in two.

^ Case 16 uterus injured bj' l)utcher; chorion cut in two.

^ Case 22 uterus injured liy l)utcher; chorion cut in two.

'•' Cases 42 and 44 are not included because they were selected heterosexual pairs taken after birth.

Til. THE TIME OF -FUSION OF THE TWIN CHORIONS AND THE DEVELOPMENT OF THE VASCULAR ANASTOMOSES BETWEEN THE TWINS

In order to form an estimate of the probable time of fusion of the twin chorions it is necessary to present a few data concernmg the development of the usual single chorion. Figure 1 shows the non-pregnant uterus of the cow partly dissected. It will be noted that the horns of the uterus open by constricted apertures into the small body. The blastodermic vesicle forms in the horn of the uterus on the same side as the ovary from which the ovum was derived, as I have observed in numerous cases. It grows out into a long strand-like sac extending both distally and centrally. The embryonic area forms near the. center in the sheep (Bonnet) and presumably also in the cow. The grow^th of the strand-like vesicle in length is extraordinarily rapid, and it soon enters the body of the uterus centrally, and penetrates into the opposite horn. By the time that the embryo is 10 mm. long the vesicle has extended completely through the body of the uterus and far into the other horn (two cases observed) ; the embryo is thus excentrically placed in the very long vesicle. The allantois forms later than the blastodermic vesicle; it grows from the embryo both centrally and distally, and ultimately completely fills the blastodermic vesicle and occludes its cavity. In the case of an embryo of 19 mm. length the allantois had passed from the horn of the uterus containing the embryo well into the body of the uterus. In another case of an embryo of 21 mm. length the allantois had extended through the entire horn of the uterus opposite to that containing the embryo.

I have one case of a twin pregnancy in the cow in which the embryos were only 15 mm. long (no. 49). Unfortunately the

THE JOUKNAL OF EXPERIMENTAL ZOOLOGY, VOL. 23, NO. 2


390 FRANK R. LILLIE

collector did not recognize the case as a twin pregnancy until after the uterus had been opened and one foetus removed, thus cutting the membranes; the other foetus was also removed in the same way, and when the uterus and specimens reached me for examination it was necessary to reconstruct the original condition from the parts. This w^as, however, successfully done, as the central end of the one chorion was found still in place in the body of the uterus and extending into both horns. In this one the chorion "wdth contained allantois had passed the body of the uterus. In the other the end of the chorion had been drawn out of the uterus with the foetus, but measurement showed that it also with the allantois contained had passed the body of the uterus. The two chorions were thus not fused at this stage, but they overlapped and were in closest apposition in the body of the uterus. The conditions precedent to fusion were thus fully established at this early stage long before sexual differentiation begins.

Comparison of these twins with the stages of 19 and 21 mm. described above indicates some variation in the degree of development of the chorion relative to the length of the embryo. But these stages demonstrate the possibility of fusion of twin chorions a considerable time before the stage of beginning sexdifferentiation, which I estimate at about 25 mm. Vascular anastomosis between the twins is possible as soon as the allantoes from the two sides meet, or even earlier, because after the allantois has once fused to the chorion the blood-vessels tend to spread out more or less in the chorion beyond the area of fusion.

Owing to the extreme difficulty of obtaining early stages of twin pregnancies in cows the next earliest stage that we have (no. 51) is a case of male twins 5 cm. long in which there is no evidence of the place of fusion of the twin chorions, and there is a perfect vascular anastomosis between the two sides. Fusion is already perfect and any overlapping parts have entirely disappeared. The next case is a two-sexed pair (no. 19) in which the male foetus was 80 mm. and the female 75 mm. long. The twin chorion was single with a broad connection provided with cotyledons between the two halves; no evidence of the place of


ACTION OF SEX HORMONES IN FOETAL LIFE 391

fusion of the twin chorions remained. The urinogenital system of the female, described in section 4, was already definitely of the sterile free-martin type. The inference is, therefore, that, fusion had taken place some time previously in order to account for the completeness of the fusion and the transformation of the reproductive system of the female.

In the case of twin pregnancies in cattle, therefore, the two vesicles starting in opposite horns of the uterus mil meet in the body of the uterus before the 10 mm. stage; the allantoes of the two vesicles will not however meet until about the 15 mm. stage, and the opportunity for vascular anastomosis therefore dates from this time.

Bonnet ('89) describes a very early t"v\dn pregnancy of the sheep, which confirms in the strongest way my conclusion concerning the early time of fusion of twin chorions in ungulates. His description is of so much interest that I quote it entire, he describes a pair of sheep twins 6 mm. long, secured 18 days and 6 hours after copulation,

deren serosen Hiillen an den sich beriihrenden Enden auf eine Strecke von 6 cm. in einander eingestlilpt imd verklebt, aber noch nieht verwachsen waren. Sie Uessen sich vielmehr noch leicht auseinanderlosen. Beide Eier maassen zusammen vom freien Ende des einen bis zum freien Ende des anderen 35 cm. und waren in maximo 1.5 cm. weit. Die Ktirze der Eier ist eine imVergleiche zur Lange einzelner Eier in diesem Stadium auffallende; sie betrug bei einem 15, bei dem anderen 17 cm. Wahrscheinlich behindern sich die bald einander mit den Spitzen beriihrenden Eier einigermaassen in der sonst normalen Langenentwieklung.

It will be noted that in this case the ova met at their apices and invaginated one another, and that the stage of such union was only 6 mm. The sheep's uterus is of precisely the same type as the COW" ; fusion follows the union of the ova in the sheep as in the cow; but vascular anastomosis does not occur in the sheep, as I describe in detail later on, and for this reason the female of two-sexed t\\dns remains unaffected in the sheep.

We have already referred frequently to the vascular anastomoses between twin foetuses of the cow, and it is now time to describe the matter fully. The working hypothesis with which


392 FRANK R. LILLIE

the investigation began was that the free-martin and its twin were monozygotic, and it was not until after 27 cases had been examined that I was convinced that they were dizygotic. The real explanation of the phenomenon then for the first time became evident. No vascular injections were therefore made during the first part of the investigation, and the evidence for vascular anastomosis among these rests upon incidental observations, the significance of which was not realized at the time. Relatively few of the twin ova received thereafter were in a fit state for complete injections. Of the 28 cases involved injections were made only in seven cases; two of which will be described in detail below. But in 21 of the 28 cases vascular anastomosis could be satisfactorily demonstrated either in the uninjected chorions or in injections. Some of the uninjected cases were just as demonstrative as though they had been injected. In four cases of the remaining eight there was no anastomosis; one too young (no. 49) ; the second was a case of normal male twins in which the connection between the two chorions was merely a narrow band-like connection (no. 46) ; the third was a case of male twins \\ith entirely separate chorions (no. 56); the fourth was another case of completely separate chorions (no. 40) of the greatest theoretical interest because one foetus was male and the other a normal female. Finally there were three cases wdth inadequate records. Eliminating these three we have 25 cases, in 21 of which, including the three possible twin combinations, vascular anastomosis could be demonstrated and 4 in which it was absent (nos. 40, 46, 49, and 56).

This is not, however, the only evidence that more or less complete vascular anastomosis between the pairs is the rule in cattle twins. I can distinctly remember the continuity of the thickened chorionic band that carries the main arteries as the rule in the first 27 cases, and this was recorded in certain cases in my notes.

There cannot be the least doubt that in bovine twdns fusion of the chorions usually occurs and is followed by anastomosis of the blood vessels of the two sides, and that intermixture of the blood of the two foetuses results. Nor can it be doubted that


ACTION OF SEX HORMONES IN FOETAL LIFE 393

rarely this does not occur either because the chorions fail to fuse (cases 40 and 56) or because a slender connection is not vascularized (case 46). The significance of the exceptions is very great.

The nature and extent of the vascular connections may now be illustrated by a detailed study of two cases. 1. Case no. 33 Males, 10.5 cm. long (figs. 2 and 3). The entire arterial system of both chorions was injected from one umbilical artery of one partner; the mass easily passed the constriction between the two halves of the chorion, and penetrated even into the umbilical arteries of the other; every cotyledon was injected on both sides. The venous system was also injected from one of the umbilical veins of the same specimen; the injection mass also passed the constriction far into the chorion of the opposite side, but the blood present in the veins prevented as complete an injection of the veins as of the arteries.

The two umbilical arteries of each foetus have a cross connection at the distal end of the umbilical cord, so that an injection from one artery outwards flows both centrally and distally. The two veins lack such an anastomosis.

The arterial anastomosis (fig. 3). The main artery from the right of the drawing divides in three branches 1, 2, and 3 as it approaches the center. Branch 3 need not be farther considered as it does not anastomose with the opposite side. Branch 1 can be followed directly through into communication with the arterial system of the other side, branch 2 has a strong anastomosis with the through trunk 1-1 (at 1-2), but branches for the most part within its own venous territory. The side branches of the through trunk 1-1 are of considerable interest, inasmuch as some are oriented in the direction of the blood flow from the right, and others from the left. Thus following the trunk from the right the first branches that we meet are directed against the blood stream coming from this direction (la); immediately after passing the anastomosis {1-2) we meet a branch Ih, directed with the blood flow from the right; the next two branches, Ic and Id, are directed similarly, but the large branch le immediately beyond has the reverse orientation. If we suppose the blood


394 FRANK R. LILLIE

flow to come from the left leis directed with the current, Id, Ic, and lb against it, etc. The orientation of these branches seems to indicate an alternation in direction of flow in the main trunk, as is to be expected ^\dth a beating heart at each end of it. It is significant that this is the only place in the arterial system of the membranes where reversal of orientation of branches is found in the course of a single trunk.

The venous anastomoses. The venous anastomoses are two in number (4 and 6 in figure 3). The larger one 4 was on the opposite side from which the drawing was made, and is therefore represented as a broken line; 5 comes from the same main venous stem. There is no reversal of orientation of side branches.

The circulation. It is obvious that any arterial blood that is pumped by either foetus into the capillary system situated beyond about the line A-B will be taken up by the venous system to the other foetus. There must therefore be a constant interchange of blood between the two foetuses, which, considering the size of the arterial intercommunications, must be very considerable. The venous anastomoses are not significant for the intermingling of the two circulations. The direction of flow along the main arterial trunk (1-1) mil depend on the blood pressure on the two sides. If for any reason an excess of blood is received by one of the two foetuses, this will have a tendency to raise the blood pressure on that side and thus to equalize the distribution. There is of course the possibility that the beat may alternate on the two sides, but nothing is known of this, and the effect of such an arrangement would not be easily deduced.

2. Case no. 47. cT 22.75 cm. and $^22.25 cm. (fig. 4). Arteries injected yellow; veins blue. The injection was made first into an umbilical artery and vein of the male. The arterial injection flowed regularly into the opposite chorion and through to the free-martin; the venous injection also flowed into the opposite chorion, but not so freely. The injection was then completed from an umbilical artery and vein of the free-martin in order to fill the vessels on this side more completely.


ACTION OF SEX HORMONES IN FOETAL LIFE 395

The arterial anastomosis is a single strong vessel, the relationships of which are shown clearly in the figure and require no further description. The stage is much more advanced than the preceding case, and the cotyledons are much more developed. Most of the arterial branches are distributed directly to the cotyledons. The venous anastomosis is much less viable than the arterial; macroscopically it consists exclusively of a connection between the two veins of one cotyledon (2, fig. 4) one of which returns to the male side and the other to the side of the freemartin. This is the only cotyledon that appears to be connected with the umbilical veins of both sides; therefore any other venous anastomisis must be through the capillary circulation of the extra-cotyledonary chorion if it exists.

The circulation in this case must be according to the same principles as in the preceding: whenever the arterial pressure is higher on one side than the other blood must be distributed from the side of higher pressure to that of the lower pressure; it will thus reach the veins and the foetus of the opposite side; variations in pressure on the two sides must constantly occur, if there is any difference in the time of occurrence of systole and diastole of the twin hearts. The blood of the twins must therefore intermingle intimately, and internal secretions of either must reach the other.

These cases adequately illustrate the time and nature of the vascular anastomosis; we may therefore turn to the question of duration of the intermingling of the blood during foetal life.

We have seen that the vascular anastomosis probably begins at the stage of about 19-20 mm. The two cases we have considered in detail indicate a strengthening of the arterial anastomosis, and a weakening of the venous anastomosis after a certain stage as development proceeds. This is to be expected because the arterial flow is stronger and toward the center, whereas the venous flow is slower and away from the center. The circulation itself tends therefore to strengthen any primitive arterial connection, and to diminish relatively any venous connection. Moreover as development proceeds the cotyledons increase in size, and the intercotyledonary circulation in the chorion becomes


396 FRANK R. LILLIE

correspondingly reduced in a relative sense A\'ith the result that the prominent arteries and veins become exclusively cotyledonary with the single exception of the artery connecting the two sides; and any intercotyledonary venous connections become insignificant.

It is an important question whether this condition persists throughout foetal life, even though completely sterilizing effects on the female reproductive system are produced by the stage of 7.5 cm., as we shall see in more detail in another section. The question therefore relates to possible influences on later stages of the female reproductive organs, and on the somatic characters of both t\\ins. The latest stage that I have examined wdth reference to this question was a pair of female t\\ins, 35.3 cm, and 31.25 cm. in length respectively. The arterial connection was even stronger than in earlier stages proportional to the more advanced stage of development. There is no reason to suppose that the connection is interrupted until birth, but the actual observations have not been made.

Thus the available records indicate a growth of the arterial anastomosis throughout foetal life and a consequent duration of action of the male hormones up to the time of birth. The possibility exists that in certain cases the connection may be interrupted at different stages of development; but so far no such cases have appeared. In any event the decisive effects on the reproductive system of the female are determined very early and they are presumably irreversible in their character.

Triplets occur rarely in cattle, and cases of eA'en more young at a birth are on record. Unfortunately records of their breeding history appear to be very rare. The only one that I have been able to discover is given by Pearl ('12). In this case there were two females and one male. The females were kept until they were about three years old, but they never came in heat. They were then killed, and The man that dressed them said that they never would have bred. Neither uterus nor tubes were recognized, but the vagina apparently ended at its anterior end as a blind sac." Both were apparently sterile free-martins. The male was put in service and got good calves. We have here.


ACTION OF SEX HORMONES IN FOETAL LIFE 397

therefore, in all probability a case in which the circulations of the three indi\dduals anastomosed, and in which the male sterilized both females.

IV. THE HORMONE THEORY OF THE FREE-MARTIN

We may now proceed to a consideration of the argmnent for the hormone theory. In our previous considerations we have dwelt upon the separate zygotic origin of the free-martin, and the foetal vascular connections; it is obvious that these conditions suggest a hormone theory; but, before such a theory could be regarded as demonstrated an explanation of the existence of fertile freemartins would need to be offered, and the limitation of the phenomenon of sterility of the free-martin to cattle as a common occurrence would have to be explained; the possibility of the existence of sex hormones at such an early period of the foetal life would also need to be demonstrated, and reason for limitation of the sterilizing effect to the female is needed.

We shall consider first the fertile free-martin; three cases of a normal female twin to a male have been found in my 24 cases of bovine two-sexed foetal twins. These are readily explained a -priori on the hormone hypothesis on the supposition that they represent cases in which anastomosis of the foetal blood-vessels did not occur. It is important to notice that such cases are exceedingly crucial, for if we should find a case of two-sexed bovine twins in which foetal vascular anastomosis was absent, and in which the female was nevertheless a sterile free-martin, the hormone theory would have to be abandoned.

The first two cases of fertile free-martins were nos. 8 and 9 of my series (figs. 7 and 8) ; they were collected before the hormone theory was formed and the records are incomplete. In my notebook I had merely recorded that the connection between the two chorions of each pair was narrow; it was probably not vascular, but this cannot be certainly known, and these cases must be left out of consideration. Fortunately the third case, no 40, is a veritable experimentum crucis. In this case organic connection of the two chorions was entirely lacking. The central ends of the two chorions merely overlapped in the body of the uterus, and fell


398 FRANK R. LILLIE

apart when removed; injection of the chorion of the male showed its circulation to be entirely closed. Dissection of the female showed its reproductive system to be perfectly normal (fig. 6); sections of the gonad showed it to be an ovary (Chapin, '17); each maternal ovary had a corpus luteum in it. Even though this case stands alone, it is obvious that it fulfills all the conditions of a radical experiment; so that we can say that foetal vascular anastomosis of two-sexed twins involves the sterile condition of the female, and absence of such anastomosis its fertile condition.

The sheep and other normally uniparous ruminants should furnish another test of the theory; for though twin births are fairly common in sheep the female of two-sexed pairs is usually norrhal. This is a matter of common experience among breeders, and is strikingly demonstrated by Prof. Alexander Graham Bell's well-known experiments (Bell '12) on the production of a multinippled race of sheep ; 36 per cent of the lambs born on Professor Bell's farm were twins; and in 1912, 60 per cent of the lambs born from three year old ewes were twins; the records show that the twin ewes are used commonly for breeding purpose, ^vhich would not be the case if any considerable percentage were sterile. The fact that there is no reference in this very careful series of experiments to sterility of ewes from two-sexed twins would also show that such a phenomenon must be at least very uncommon. On the other hand Bateson states that it sometimes occurs among sheep; though, on what authority, I do not know.

In response to a letter of inquiry Wm. John G. Davidson .who has had charge of the breeding operations at Dr. Bell's estate for a great nlany years writes:

I may saj that in all my experience in sheep breeding I have yet to find a case where lambs born twin to males have turned out sterile. In fact when lambs are born twin male and female if they have the desired qualifications required in the flock both lambs would be retained in the flock and I have not had the slightest trouble with either male or female being unfruitful. I know there is nothing in the free-martin theory in sheep breeding.

It was therefore very interesting to examine twin pregnancies of sheep with reference to the relations of their membranes. I


ACTION OF SEX HORMONES IN FOETAL LIFE 399

found in the four cases, that I examined, that the twins were dizygotic (in one case both corpora lutea were in one ovary) and that the membranes were fused in the body of the uterus as in cattle. But when injections were made, as was done in all four cases, it was found that the circulation of each individual was entirely closed; the injection mass could not be forced from one side to the other, either through the arteries, or through the veins. Figure 5 gives a faithful representation of one case; it will be observed that the arteries and veins of each side end in a central neutral zone that they do not cross; this zone is no doubt occupied by capillaries, and it is possible that these anastomose from the two sides, though it is uncertain. The other cases were similar, though in one of them a single centrally placed cotyledon received an artery from each side; each artery was accompanied by its own strong vein returning to the same side, which indicated that there was little, if any, intermixture of blood in the cotyledon; the starch injection masses, yellow on one side and red on the other were not forced through.

In the sheep we have, then, all the necessary conditions for the production of sterile free-martins except the actual vascular anastomosis. If the vascular anastomosis should also occur exceptionally, such a condition should be accompanied by sterility of the female in the case of heterosexual pairs. This lends probability to the assertion that this condition actually occurs occasionally in sheep.

The hormone theory thus gives a satisfactory explanation of the occurrence of occasional fertile free-martins in cattle as well as of the usual condition of sterility of the free-martin; and it fits the case of the sheep equally well. As regards other ruminants we have unfortunately almost no information. But I have been much interested to find that the famous discoverer of the circulation of the blood, William Harvey, in his "Exercitationes de Generatione Animalium 1651 has some statements on the subject of twin pregnancies in ruminants: thus in Ex. 69, p. 487 (Sydenham Society edition, translated by Willis), he says of the deer, if the conception be double, one in either horn (of the uterus), each sends its umbilical vessels to its own


400 FRANK R. LILLIE

horn alone; the embryo in the right horn deriving nourishment from the right part of the conception, that in the left from the left portion of the same." He made similar observations on the sheep, goat, and "other bisulcated animals" and notes that "in the dog, rabbit, hog, and other animals that produce a considerable number of young at a litter, the thing is otherwise. In these each foetus has two humors, these being severally surrounded with their proper membranes." So far as I know there are no other published observations on the foetal membranes of twins in ungulates from Harvey's time to the present with the exception of Bonnet's single case already referred to. Harvey's observations show that fusion of chorions is wide spread in twin pregnancies in ungulates; but he states definitely that in the deer the umbilical vessels of each foetus are distributed to its own side only, in which it resembles the sheep. A more careful examination of the female of two-sexed twin pairs in these animals would be of interest in order to determine the possible sporadic occurrence of sterility.

The theory requires that if the same condition of common circulation of the foetal blood were to occur in other mammals as in twins of cattle the sterile free-martin condition should occur there also. Now in multiparous mammals such conditions certainly do not occur commonly; for, if they did, the very numerous researches on their embryology would have brought them to light. In the pig one can find occasional, but rare, fusions of adjacent chorions, but I have never found any vascular connection. A number of mammalian groups could be at once excluded from consideration because the conditions of placentation are such as to prohibit chorionic fusion; in mammalian groups such as primates and many rodents in which the ovum becomes embedded in the uterine mucosa, there is of course an insuperable bar to early chorionic fusion. And in those mammalian groups in which the placenta is a highly localized organ as in the remainder of the rodents, the insectivores, carnivores, and edentates the circulation in the chorion outside of the placental area is so restricted that, even if chorionic fusion did occur, it is difficult to believe that the circulation of separate foetuses would intermingle to any great extent.


ACTION OF SEX HORMONES IN FOETAL LIFE 401

Fernandez ('15) has described fusion of dizygotic chorions in one of the armadilloes (Dasypus villosus). This form has usually two young at a birth, which may be one-sexed or two-sexed. The two chorionic vesicles are separate at first, but they gradually fuse and by the time that the embryos are 3.5 cm. long the fusion of the two vesicles is so intimate that they appear as a single one. As this is the usual condition it cannot be supposed that it is accompanied by genital abnormalities. The author does not consider the problem of vascular anastomosis; but it can hardly be supposed to occur.

Although the hormone theory invokes a cause of the utmost generality in mammals, it is obvious that the conditions leading to its intervention must be restricted to forms with a relatively diffuse placentation, i.e., ungulates for the most part, and among these to forms in which quite special conditions obtain. Such conditions are found only in normally uniparous ungulates in which the ovum grows to an extreme length very rapidly, so that the associated ova meet at an early stage which favors their organic union. Even then vascular anastomosis is not likely to occur to any considerable extent unless the development of the foetal cotyledons is relatively late, so as to be preceded by a condition of general vascularization of the chorion, before the highly specialized circulation of the cotyledons becomes dominant. Such is the condition in cattle. In sheep the development of the cotyledons appears to be more precocious; I would at least venture this suggestion, although based on relatively few observations, to explain the difference between cattle and sheep in this respect.

Although these considerations give a color of great probability to the hormone theory we have still to deal with two difficulties : — 1) it is evident that we must be dealing with specific sex hormones, for their influence is limited to sex characters so far as our present evidence goes; is there any other evidence for such early production of specific sex hormones available? 2) It is certain that the intermixture of blood of the twins must be reciprocal, but the effect is exclusively on the female; in what way can this be explained?


402 FRANK R. LILLIE

Both of these difficulties receive a satisfactory explanation in certain previously known facts not hitherto correlated with the phenomenona in question, and in certain new facts which are described by Miss Chapin ('17). They constitute concurrent evidence which appears to me to render the entire evidence perfectly conclusive.

The previously known facts to which I refer are (1) the early development of the interstitial tissue of the mammalian testes, from the very beginning of sex-differentiation, and (2) the fact that the differentiation of the ovary is later than that of the testis, inasmuch as in the female of manunals the first generation of ingrowths from the germinal epithelium, a complete homologue of the seminiferous tubules of the male, forms only the medulla of the ovary, and the ovarian cortex is formed from a distinct second generation of ingrowths. Thus (1) interstitial tissue of the testis is present at the time for which male hormones are postulated, and (2) the testis has a start over the ovary in this respect which results in the suppression of specific ovarian tissue from the beginning as shown by Miss Chapin' s study.

On the question of the embryonic origin of interstitial cells in the testes we have the excellent studies of Whitehead ('04) and Allen ('04) on the pig; Allen also deals to a certain extent with the interstitial cells of the ovary. Allen finds that sex differentiation is strikingly shown in the structure of the testis and ovary in embryos of 2.5 cm. length. The sexes cannot be sharply distinguished at 1.8 cm. length so that the initial stages of sexdifferentiation lie in between. Interstitial cells are present in both testis and ovary at the stage of 2.5 cm., but while they are very numerous in the testis at this stage they are very rare in the ovary. Whitehead finds that they appear in the testis of the pig embryo at 2.4 cm. Althoughr these authors do not note it, it is unquestionably significant that these cells appear, and exhibit the usual evidences of active secretion at the time of the onset of sex-differentiation. It is also significant that they appear first in the testis and more abundantly than in the ovary; this is of course correlated with the fact that the cortex of the ovary in which they appear is a later formation than the seminiferous tubules or their homologue in the female, the medullary cords.


ACTION OF SEX HORMONES IN FOETAL LIFE 403

As regards the origin of the interstitial cells of the ovary we have relatively few observations. Allen notes, as we have seen, that they are very rare in the ovary of the 2.5 cm. pig as compared with the testis; they are also very short-lived, disappearing at the stage of 4 cm. No interstitial cells are found in the rabbit ovary until the stage of 45 days after birth." They would thus appear to form a very inconspicuous and transitory feature of the embryonic ovary.' It is not clear from Allen's account whether they occur in the medulla or cortex of the embryonic ovary; a point which is of some significance, in connection with the following discussion.

As regards the testis \^Tiitehead notes

Leydig's cells (the interstitial cells of the testes) pass through two phases of growth, between which a phase of atrophy intervenes. Growth is very rapid from their appearance in the embryo 2.4 cm. long until the length of 3.5 cm. is reached. Tliis is followed by the phase of atrophy, during which the cells return almost to their first state of nearly naked nuclei (figs. 4 and 5). This process reaches its acme in the embryo 14 cm. long. Synchronous with it there is extensive growth of the seminal tubules, particularly in length, so that they are much convoluted, and the intertubular spaces are correspondingly narrowed (fig. 6).

The atrophy proceeds slowly from 3.5 to 14 cm., and is not at all marked at 5.5 cm.

In the embryo 20 cm. long the cells enter upon the second phase of growth, which attains its maximum in the pig of 28 cm., very near to term. Here the cells are enormously increased in number and size, so that they constitute the predominating feature of the microscopic picture.

It seems to me that these facts are of great significance, viz: the intensive growth of the interstitial cells of the testis at the time of most rapid sex differentiation, between 2.4 and 3.5 cm., the subsidence of their activity after embryonic sex differentiation is once attained, and the second phase of activity which probably leads on to the juvenile sex differentiation. It would thus appear that both phases of sex differentiation are covered by periods of intensive acti\dty on the part of the interstitial cells of the testis.


404 FRANK R. LILLIE

These data relate for the most part to the pig, but there is no doubt that we are justified in assuming that the processes in cattle would not differ essentially; and hence that at the time of sex-differentiation in cattle there is an active secretion of male sex hormones which pass into the blood, and thus in the case of twins by the vascular anastomosis reach the circulation of the other twin. At this time, if the other twin is a female, the cords of Pfliiger have hardly begun to form; no interstitial cells can therefore be present ; hence there can be no conflict of hormones. Now Miss Chapin's study of the embryonic gonad of the sterile free-martin shows that the cortex of the ovary does not develop in these animals. Hence no conflict of sex hormones arises; hence also, there can be no question of the male of two-sexed twins being influenced in its sexual development by its mate.

There is at least no escape from the conclusion that it is the circulation of the blood of the male twin in the female that accounts for the results. The probability of the presence of interstitial secretion of the testis in the blood of the male at the time of beginning sex differentiation and the limitation of the action of the male blood to the reproductive system of the female are the reasons for attributing the effect to sex hormones of the male. The possibility of course exists that the blood of the male in such foetal stages differs from that of the female in this specific respect owing to other causes than secretion of interstitial cells of the testis; but there seems to be no reason for making such an assumption .

In my preliminary paper I left open the question whether the invariable result of sterilization of the female at the expense of the male was due to more precocious development of the male hormones, or to a certain natural dominance of male over female hormones. It now appears from the results of more detailed investigation that the latter alternative probably does not arise.

The main assumptions that are involved up to the present point are (1) that cattle resemble pigs with reference to the early origin of interstitial tissue in the testis; (2) that such tissue has in foetal life properties similar to those that have been demonstrated in post foetal life by a considerable number of investi


ACTION OF SEX HORMONES IN FOETAL LIFE 405

gators. While it wdll no doubt be desirable to clear up these assumptions by farther observations and experiments, it is nevertheless true that the facts that we have presented form a solid basis for both assumptions, which are in no sense unsupported or forced.

The material presents the problem whether the intermixture of other internal secretions o the twin individuals modifies any of their characters? It is obvious that modifications could be expected only in the case of differentiating characters, of which the most fundamental are those of sex; all other main features of organization are common to the sexes, and there is no evidence of individual or sexual differences in the hormones of a species except in the sexual hormones themselves. A negative answer would therefore be expected to this question in general. Individual variations of course exist, and it is an interesting question whether the blood community of foetal life tends to reduce such variations and to approximate their resemblances towards the correlation of variability of identical twins? We have seen reason to believe that this is not the case in regard to color, but the problem remains for future investigation.

According to the conceptions involved in this discussion, the de^'iations of the sterile free-martin from the female type are due to the action of the male sex-hormones. In order to appreciate the full extent of this action, it is necessary to understand minutely the anatomy of the sterile free-martin in comparison with normal individuals. This comparison is made in a detailed way for numerous cases in section V. The results there given will enable us to correlate the whole series of phenomena and to discuss in a very general way the phenomena of sex-differentiation in mammals.

In terminating the present section then we may repeat .our main steps up to this point: — (1) We have demonstrated the separate zygotic origin of the free-martin and its male twin, (2) We have studied the foetal vascular connections in twin pregnancies of cattle. (3) We have explained the existence of occasional fertile free-martins. (4) We have explained the reason for the usual Hmitation of sterility of the female of two-sexed

THU JOURNAL OF EXPEHIMENTAL ZOOLOGY, VOL. 23, NO. 2


406 - FRANK R, LILLIE

tT\'ins to cattle. (5) We have shown that there is a firm basis for the hormone theory in the known data concerning the time of origin of interstitial tissue in the testis and in the ovary. (6) We have also shown why the effect is limited to the female.

V. THE EMBRYONIC AND ADULT ANATOMY OF THE FREE-MARTIN

We can now proceed to a discussion of the anatomy of various stages of the free-martin with a firm basis for interpretation of the conditions. This will be considered under two heads, A. the gross anatomy, B. the microscopic anatomy. The second section has been worked up by Miss Catherine L. Chapin from my material, and is published as a separate paper immediately following this.

The gross anatomy

The mammalian embryo passes through a stage of complete sexual indifference anatomically, in which all the organs of both sexes are represented by embryonic rudiments in each individual; the anatomical sex differentiation that follows is due to progi'essive development of certain parts and degeneration of others in different Avays in the two sexes. This is such a commonplace in embryology that it seems unnecessary to describe the anatomy of the indifferent stage.

For comparison with the anatomy of the free-martins I have introduced figures of the normal female reproductive system of embryos 10.2 cm. long (fig. 6), 17 cm. long (fig. 9), 20 cm. long (fig. 7) and 23.3 cm. long (fig. 8), and of the normal male reproductive system of embryos 10.4 cm. long (fig. 10), 15.8 cm. long (fig. 11), and 26 cm. long (fig. 12). The free-martins will be described in the order from the youngest to the oldest. ' The written descriptions are given very briefly, and the reader should refer to the figures for details. The exact interpretation of the ducts is often difficult or impossible on the gross anatomical evidence alone; Miss Chapin's descriptions should be compared.

1. Case 19. cf 8 cm.; ^ 7.5 cm. in length; figure 13, urinogenital Septem of Female. X 4. (Cf. Miss Chapin's description of the histology of this specimen in the paper following.)


ACTION OF SEX HORMONES IN FOETAL LIFE 407

This is the youngest free-martin found ; its reproductive system is ah-eady definitely of the sterile free-martin type. The gonads are much smaller than in the normal female (cf. description by Miss Chapin '17, p. 455). Miss Chapin's observations described in the follo^^^ng paper demonstrate a complete absence of cortical tissue in the ovary; the medullary component is hypertrophied as compared with the normal, and an albuginea is present. The ducts also show a sifting in the male direction (Chapin, p. 459), the Miillerian duct being in process of degeneration. Thus already at this stage the characters of the sterile free-martin are definitely estabhshed, and it is proved that the primary effect of the male sex hormones is to prevent the formation of the ingrowths that normally form the cortex of the ovary. This is a matter of great significance, which is repeated in all the other cases.

History. Collected October 20, 1915. Maternal ovaries absent. These twins were contained in a single chorion with broad connection occupied by cotyledons between the two halves.

2. The next stage ( $" 12.5 cm., case 26) was used for histological work; it is described by Miss Chapin. No study of the gross anatomy.

.3. Case 17. d" 13.75 cm.; ^ 13.1 cm. in length; figure 14, urinogenital system of ^. X 4.

In this preparation the very minute size of the gonad strikes the eye sX once; (compare fig. 6, normal female 10.2 cm. long, and 9 normal female 17 cm. long). The Wolffian body has entirely degenerated. Phallus typically female. About midway between the gonad and the genital cord the inguina' fold arching over the umbilical arteries unites the ducts (urinogenital fold) to the lateral body-wall; this fold is the foundation for the round ligament in the female and is the site of formation of the gubernaculum in the male. The ducts are interpreted as Wolffian ducts, though they appear to be united in the posterior third of the genital cord; Miillerian ducts are not \dsible macroscopically.

History. Bovine twins cf and ^. Collected October 9, 1915. One maternal ovary missing ; the other had a single corpus


408 FRANK R. LILLIE

luteum in it. The chorion was single with a broad connection occupied by cotyledons between the two halves.

4. Case 37. d^ 16 cm.; ^ 15.5 cm., figure 15, urinogenital system of ^ . X 2.

This figure may be compared with figure 11 (normal male of 16.8 cm.) and figure 9 (normal female of 17 cm.). The exceedingly rudimentary condition of the gonads is at once evident, about two-fifths of the length and one-third of the breadth of the normal ovary. The ducts appear to be Wolffian ducts. An extraordinary feature of the anatomy which recurs in every ^ specimen of this or greater size is the development of a pair of gubernacula from the inguinal fold, which penetrate the musculature precisely as in the male (fig. 11) and teach the integument, extending well down to the groin. In the normal female the round ligament is definitely formed (fig. 9 A and B) in place of the gubernaculum, the horns of the uterus are definitely differentiated from the Fallopian tube, the vagina is enlarged.

History. Bovine twins d" and ^. Collected January 25, 1916. Both maternal ovaries present, each wdth a single corpus luteum. The uterus was damaged by a cut above the cervix and the two halves of the chorion severed. Examination showed that the chorion had been single with both arterial and venous anastomosis between the two halves. Cotyledons continuous from side to side.

5. Case 6. d 16.8 cm., <^ 16.3 cm., figure 16, urinogenital system of female, (cf. Miss Chapin's description of the histology of this specimen and interpretation of the ducts.)

For comparison normal female of 17 cm. (fig. 9) and normal male of 16.8 cm., twin to this specimen (fig. 11).

The very minute size of the gonads is again apparent. Gubernacula again take the place of the round ligament. The Miillerian ducts are undeveloped. Seminal vesicles, however, have not begun to form (cf. normal male, fig. 11) ; the contrast between the genital cord of the ^ and normal 9 is very striking.

History. Bovine twins. Collected March 4, 1915. Only one of the maternal ovaries was present; corpus luteum in it. Single chorion for the twins somewhat constricted. No record of vascular conditions.


ACTION OF SEX HORMONES IN FOETAL LIFE 409

6, 7, and 8. Cases 32, 23, and 36. 16.75, 17.5, and 17.5 cm. respectively may be mentioned together (figs. 17, 18, and 19). The drawings of these three specimens show essentially the same anatomical characteristics as the preceding stage; the gonads are very small in all, but somewhat larger in 32 (fig. 17) than in the others. They all exhibit gubernacula and the ducts resemble the male ducts much more than the female; the urinogenital sinus is intermediate in length between the female (fig. 9) and the male (fig. 10).

Histories. Case 32. Figure 17. Bovine twins. Male 18 cm., female 16.75 cm. Collected January 1, 1916. Both maternal ovaries present; corpus luteum in each. The uterus was injured and the single chorion cut in two; there was an arterial connection between the halves, and cotyledons continued through.

Case 23. Figure 18. Bovine twins. Male 18 cm., female 17.5 cm. Collected October 29, 1915. Only one maternal ovary present, containing single corpus luteum. The chorion was single; broad connection between the two halves; cotyledons in connecting part.

Case 36. Figure 19. Bovine twins. Male 18 cm., female 17.5 cm. Collected January 21, 1916. Both maternal ovaries present with a single corpus luteum in each. The chorion w^as single with broad connection occupied by cotyledons between the two halves. Arterial anastomosis present; veins interdigitate, and may communicate.

Summary of the preceding cases. The gross anatomical findings for the stages 7.5 to 17.5 cm. thus show that the effects (either direct or indirect) of the male hormones on the reproductive system of the female are (1) to inhibit the growth of the gonad, (2) to cause the formation of typical gubernacula in place of the round ligament, (3) to favor the development of the Wolffian ducts and inhibit the development of the Miillerian ducts. The urinogenital sinus is intermediate between the male and female condition and the phallus develops similarly to that of the female; moreover the disposition of the teats is female, differing definitely from that of the male (figs. 4, 12, 25 A) and no


410 FRANK E. LILLIE

trace of the scrotum which is well developed in males of corresponding size is to be found. In other words the derivatives of the urinogenital ridge (gonads, Wolffian and Miillerian ducts and ligaments) tend to develop in the male direction, but other parts of the urinogenital system of such affected females are less involved.

In later stages the conditions established at the beginning undergo farther development. Drawings of various stages from 21.5 to 28 cm. are shown in figures 20 to 27, and these will hardly require detailed description after what has gone before. A rather common anomaly is the evagination of the right gubernaculum into the body cavity in place of growing into the body-wall (figs. 20, 22, 25, and 27); it is difficult to assign a cause for this; inasmuch as it rarely happens on the left side (fig. 27) it is presumably associated with the asymmetry of the urinogenital complex which is particularly striking in the case of the kidneys (see all figures of dissections). The gonad remains exceedingly small in these stages, nor is there much change in other directions. Seixdnal vesicles, which occur frequently at least in the sterile free-martins after birth are not yet found though they are very evident in the male from 16.8 cm. on. In the normal male the testes are drawn into the saccus vaginalis at the 26 cm. stage (fig. 11); in the sterile free-martin this may happen (fig. 25 A and B), or the gonads may remain in the peritoneal cavity (fig. 26).

History of cases:

Figures 20 A and 20B. Case 2. Bovine twins. Male 23 cm., female 21.5 cm. Collected October 30, 1914. Chorion single with constricted connection occupied by cotyledons between the two sides. This being one of the earliest cases no record was made of vascular conditions in the chorion. No record of maternal ovaries.

Figure 21. Case 41. Bovine twins cf and <^ . Collected February 16, 1916. Male 22.7 cm., female 21.8 cm. Both maternal ovaries present, corpus luteum in each. The uterus had been cut above the cervix and membranes injured; the fusion of the two chorions was broad, and the cut ends of the arteries connecting the two were readily discovered.


ACTION OF SEX HORMONES IN FOETAL LIFE 411

Figure 22. Case 4. Bovine twins cf' and ^. Collected January 30, 1915. Male 24 cm., female 22.5 cm. Single chorion, constricted between the two foetuses. No other records.

Figure 23. Case 38. Bovine twins d" and?". Collected January 26, 1916. Male 23.5 cm., female 22.5 cm. Both maternal ovaries present, corpus luteum in each. The uterus was received cut off above the cervix severing the connecting part of the chorion. The large vessels run to cut ends; extensive vascular anastomosis between the two foetuses involved.

Figure 24. Case 22. Bovine tmns d" and ^. Collected October 26, 1915. Male 24.5 cm., female not measured. Both maternal ovaries present, corpus luteum in each. Uterus cut into destroying connection of the two halves of the chorion; the chorion was single and the maternal cotyledons occurred throughout the corpus uteri. Vascular conditions of chorion not observed.

Figure 25, A and B. Case 21. Bovine twins d and ^ . Collected October 20, 1915. Male 27.5 cm., female 26.5 cm. Maternal ovaries absent. The two halves of the chorion were broadly connected, and cotyledons and blood vessels cross over from side to side.

Figure 26. Case 14. Bovine twins cT and ^. Collected July 15, 1915. Male 27 cm. long; female 27 cm. long (measurements by assistant) . Maternal ovaries absent ; the two halves of the chorion were broadly connected with cotyledons and arteries crossing over.

Figure 27. Case 12. Bo\nne twins cf and ^. Collected October 2, 1915. Male 30 cm., female 28 cm. long. Maternal ovaries absent. The chorion was single, not much constricted between foetuses. Cotyledons form continuous series from side to side. No observations on vascular conditions of chorion.

The gross anatomy after birth

The anatomy of the reproductive organs of a seven weeks old free-martin (case 44) born on my farm is shown in figure 28. Incidentally it may be noted that the twins were pure bred Holsteins, and that while the male was mostly black, the female


412 FRANK R. LILLIE

was about half white. This is an exceedingly interesting case, as it represents the most extreme transformation of the internal organs towards the male type which I have seen. When the body cavity was opened gonads appeared to be entirely absent; however the ducts were found to perforate the body wall; and farther dissection revealed the gonads lying in peritoneal sacs situated in the groin on each side between the skin and abdominal muscles. There is no difficulty in interpreting the parts (fig. 28) : the vulva is typical ; it leads into a short urinogenital sinus beyond which there is no trace of vagina, uterus, or tubes. Opening into the dorsal wall of the urinogenital sinus at its anterior end are two Wolffian ducts (vasa deferentia) ; laterally to them seminal vesicles. The Wolffian duct of the left side is much larger than that of the right. They run in a broad membranous septum corresponding to the broad ligament of the uterus and enter peritoneal evaginations, open to the body cavity, which perforate the abdominal muscles in the manner already described. When the peritoneal sacs are opened, as shown in the insert figures, figure 28, B and C, they are found to contain testis-like gonads, with an epididymys-like organ associated. It is an interesting fact that the smaller right gonad is associated with a smaller Wolffian duct and a smaller seminal vesicle ; this is the side on which disturbances in relation of parts is found in foetal stages. Sections of the testis show a superficial dense albuginea and entire absence of cortical , ovarian tissue. The interior is occupied by exceedingly wide branching tubes lined by a one-layered epithelium. Connective tissue is abundant between the tubes.

History. Figure 28. Case 44. Holstein-Friesian twins & and ^ born March 12, 1916. The female (free-martin) was killed and dissected April 29, 1916.

In the foetal stages of the free-martin, which we have considered, the gonad presents a rudimentary aspect- and seminal vesicles are not formed, although they appear in the males at much earlier stages (fig. 1). On the other hand, the gonads may attain a very considerable size after birth, and seminal vesicles appear usually to be present (cf. literature). It therefore fol


ACTION OF SEX HORMONES IN FOETAL LIFE 413

lows that there must be a belated growth of the gonads and formation of seminal vesicles in the later foetal stages which I have not had an opportunity to examine, in some cases at least.

The literature contains descriptions more or less complete of the anatomy of the reproductive system of about 30 free-martins. These have been tabulated by D. Berry Hart ('09-'10). All agree in one fundamental respect, viz: that the gonads never show the least evidence of possessing the structure of an ovary; Graafian follicles have never been described. In some cases the gonad is represented merely by clumps of fat, or is absent; in others it is rudimentary and situated in the body cavity, or may be of considerable size in the same situation, or finally may be found in the groin. Whenever present it presents more or less superficial resemblance to a testis, sometimes- an exceedingly close resemblance. Histologically also it resembles a testis. However it is probable that spermatozoa are never formed, and even that the earhest stages of spermatozoa are lacking in the tubules. All agree likewise in the very great reduction or complete absence of vagina, uterus, and tubes, and the presence of Wolffian ducts in a greater or lesser degree of development. However, it would be merely tedious to review the variations in detail.

The very interesting question presents itself how far the modification of the female reproductive organs towards the male side may be carried. I have already described the most extreme case that I have met (no. 44, fig. 28) ; but one of Numan's cases indicates that the transformation may proceed considerably farther. This case seems to be well authenticated in all respects, and on account of the rarity of Numan's publication I include a full description of it : a pair of calf twins was delivered by Caesarian section in 1832 at Maarssen in Holland, one was male and the other was supposed to be a free-martin. Numan bought the latter the following year; on examination a small opening was found in the perineum a short distance below the anus through which urine dropped, the urethra passed over the pubic symphysis and opened externally about two handsbreaths below the anus through an opening surrounded by an


414 FRANK R. LILLIE

apparent female vulva, within which appeared an imperforate glans penis; the vulva was provided ventrally with a tuft of hair as normally. The animal was slaughtered in 1835, and the anatomy of the reproductive system studied. Numan's figure is here reproduced (fig. 29). It mil be seen that the internal anatomy is quite similar to my ease 44 (fig. 28), but the external parts are also modified in this case to a very considerable extent in the male direction.

If this case is well authenticated, as it appears to be, we would have to conclude, contrary to the evidence from my cases that the external organs of reproduction of the female are also susceptible of modification in the male direction by the male hormones; it may be that this occurs in cases of exceptionall}^ earlj^ action of the male hormones. The embryological history of the external male and female organs shows that the sexual type is fixed very early.

It is of course by no means certain that the extreme possible, modification of the female reproductive system by male hormones will be found in free-martins, indeed when we consider the various exigencies of the admission of the male hormones to the female circulation under the conditions of production of the free-martin, it seems improbable that the optimum conditions for modification of the female reproductive system are ever realized in such cases : the onset of the action can not be much before morphological sex differentiation has begun in any case, and its intensity must certainly be of a low order quantitatively at first not only on account of the minuteness of the first vascular inter-connections, but also on account of the rudimentary character of the interstitial gland of the testis at first, if we suppose that the effective hormone is derived entirely from this source. Nevertheless search should certainly be made for more pronounced cases of transformation of the free-martin towards the male condition. But it is obvious that definitely controlled experimental investigation will be the only means of deciding where the real limit of action lies, whether or not the complete transformation of a female zygote of mammals into a male individual by hormone action is possible.


ACTION OF SEX HORMONES IN FOETAL LIFE 415

VI. GENERAL DISCUSSION

The precision and definiteness of the transformation of the specific organ system, and the undoubted character of the primary cause of the transformation appears to offer a more definite basis than we have hitherto had for the analysis of the origin and differentiation of some fundamental sex characteristics of mammals. At the present time there is a general recognition of the primary zygotic determination of sex in mammals ; we have some scattering data on sex limited inheritance, a great deal of miscellaneous information on the effects of castration, and a little on the effects of cross transplantation of gonads. There is also a large literature on hermaphroditism and pseudohermaphroditism of mammals. But these data are by no means capable of arrangement in any general scheme. They are on the whole exceedingly confusing in spite of the recent advances concerning the zygotic character of the initial sex impulse,, and the analyses of sex hormone action in post-foetal life.

The free-martin gives us additional evidence of considerable value concerning the problem of sex-determination and sexdifferentiation in mammals, especially in its suggestion that the course of embryonic sex-differentiation is largely determined by sex-hormones circulating in the blood. The evidential value of this case is, however, limited, in the first place by the fact that only the female is affected, — we have no information on the reverse situation — and in the second place by the fact that we cannot study separately the effect of early embryonic castration of the female, but only as it is modified by the simultaneous presence of male hormones. On the male side there is complete absence of information as to the effects of early embryonic castration and the possible effect of the presence of female hormones^

5 A curious problem however presents itself in this connection viz : how it happens that the sex-hormones of the mother do not affect the reproductive system of the unborn sons. Steinach's results on feminization of infantile male rats by castration and implantation of ovaries demonstrates the far reaching effects of female sex hormones in the male system ; and we can hardly doubt that even greater effects would result in embryonic life. It would appear probable, therefore, that the embryo is in some way protected from the sex hormones circulating in the mother's blood. Either there is cessation of production of sex


416 FRANK R. LILLIE

in the absence of male hormones. These facts must be kept clearly in mind as definitely fixing the very provisional character of such speculations as we may make.

The present standpoint for the analysis of sex-characters of mammals would have to include (1) primary zygotic determination of the male and female sex and (2) secondary differentiation of the sex characters, in which internal secretions play a very specific and fundamental role. The production of intermediate zygotic conditions is theoretically possible, since Goldschmidt has demonstrated all grades of inter sexuality in the gypsy moth depending on variable conditions of the gametes. But it is obvious that such conditions are not involved in the present case and that we have to consider only the secondary factors.

It follows from the data that the female zygote must contain factors for both sexes ; the primary determination of the female sex must therefore be due to dominance of the female factors over the male. If we think of this as a simple quantitative relation, as Goldschmidt ('16) has done, we can explain the intersexual condition of the free-martin as due to an acceleration or intensification of the male factors of the female zygote by the male hormones. The degree of the effect which is quite variable, as we have seen, would of course be subject to all quantitative variations of the hormone. Thus the case of the free-martin could come under the same general point of view as that of the intersexes of Lymantria according to Goldschmidt with the one exception that the quantitative differences between the male and female factors of the female zygote necessary for the differentiation of female characters, are reduced in the free-martin by internal secretions instead of by variations of potency of the

hormones during foetal life, or they are neutralized in some way, or the placenta is impervious to them. The first possibility does not seem consistent with our knowledge of the physiology of the mammalian ovary, or with the cytology of the organ during pregnancy; the second one offers no point of attack at present; the third seems the most probable a 'priori, and is no doubt susceptible of experimental analysis. ' But whatever explanation may prove to be correct it seems probable that disturbance of the equilibrium that protects the male from the sex hormones of the mother would result in malformations of the male sex characters to a degree commensurate with the extent of the disturbances. There is, therefore, here a possible explanation of the greater mortality among male • foetuses, and of certain types of intersexes (pseudo-hermaphroditism).


ACTION OF SEX HORMONES IN FOETAL LIFE 417

male factors in different varieties as in the intersexual hybrids of Lymantria.

The case of the free-martin shows that a gonad with a primary female determination may form a structure which is morphologically a testis (Chapin '17), through suppression of the cortex and overdevelopment of the medullary cords and urinogenital union under the influence of male sex-hormones. Lesser degrees of transformation are of course possible, so that it is certain that the gonad of a mammalian female zygote is capable of most, at least, of the series of transformations that may exist between an ovary and a testis. Whether the transformation in the male direction may proceed under such conditions to the production of true spermatocytes and spermatozoa is at least doubtful. Such elements have not hitherto been described for free-martins, if we except D. Berry Hart's statement concerning the gonads of Hunter freemartins, that "in only one are spermatozoa present." More than six words seem necessary to establish so important an exception.

Pick's investigation of ^true hermaphroditism' in man and mammals shows that in all bilateral cases the gonad is part ovary and part testis; but that whereas normal germ cells may arise in the ovarian part, there is no trace of spermatocytes or spermatozoa in the testicular part. Four such cases are described for the hog, and it would seem that the theoretical' possibility exists of explaining them as due to embryonic foetal anastomosis of blood-vessels between opposite sexes, by making the farther assumption that such anastomosis was only temporary and ceased after the transformation of part of the genital ridge. The interpretation of 'hermaphroditism in mammals is in any

  • The term 'hermaphroditism' has so many and various connotations that it

seems better to drop it from our vocabulary, so far as mammals are concerned, and to describe the conditions hitherto gathered under this head simply as intersexual. Some conception of the confusion that results from present methods of classification of these cases may be felt by stating that our case 44 (fig. 28), would be classified as Hermaphroditismus spurius masculinus externus (Neugebauer '08): 'spurius' because both kinds of gonads are not present; 'masculinus' because the gonads are apparently testes; 'externus' because the external organs are female. But the animal is female, not male, and it represents merely a certain stage in a series. The animal is not a sex-mosaic, e.g., male in front of a certain transverse level and female behind ('transverse hermaphroditism'); it is a step in a series of sex intergrades.


418 FRANK R. LILLIE

event a very difficult problem, and it seems worth while to indicate the hitherto unconsidered possibility of action of heterologic sex hormones, however brought about, in this connection.

Regarding other parts of the internal reproductive system we have seen that the free-martins exhibit a graded series of inhibition of the female ducts, and of development of the male ducts which may obviously correspond to variable time of onset, intensity, and perhaps duration of action, of the male sexhormones. The series extends nearly to the normal male limit in exceptional cases (no. 44, fig. 28). There is indicated a rough parallelism at least between the grade of transformation of the gonad and that of the remainder of the internal reproductive system. The external organs of reproduction are the least liable to modification, but they do not escape in all cases, and may even exhibit considerable transformation in the male direction if we can accept Numan's case described on p. 413 and illustrated in figure 29.

The fundamental determining factor in these events is undoubtedly the male sex hormones as has been argued previously (p. 396) ; but the entire causal nexus is by no means clear. We do not know what the results of embryonic castration of the female might be in itself, and hence we are unable to assert definitely in just what positive ways the male hormones act on the female zygote, because the earliest determinable result of such action is the suppression of the ovarian cortex, which must be regarded as practically equivalent to castration. This action at least is due to the male hormones ; how much of the subsequent events is due to mere absence of ovarian tissue, and how much to positive action of male sex-hormones is more or less problematical. It is well-known that spayed females of certain birds and mammals tend to develop male characters; heifers with cystic degeneration of ovary also develop certain male characteristics (Pearl and Surface, '15), so that we must admit in principle the possibility that much of the male development in the free-martin is due to the lack of inhibitions normally furnished by the ovary.

It is also probable that the various parts of the reproductive system have other means of correlation, and act and react on one another in various ways. Certain indications of this are


ACTION OF SEX HORMONES IN FOETAL LIFE 419

seen in lateral variations, as for instance in figure 28 where a large gonad on one side is associated wdth a large Wolffian duct, and seminal vesicle, and a much smaller one on the other side mth a correspondingly smaller duct and vesicle.

When, therefore, we attribute the free-martin condition to the male hormones we only mean to assert that they are the primary cause, and not that they are the decisive factors in each member of the series of events.

It follows from this discussion that sex in mammals cannot be diagnosed by the character of the gonads alone because a testis-bearing indi\'idual may develop from a female zygote. The unexpected result is reached that the external genitalia and the mammary gland are more reliable criteria of the female sex than the internal parts.

We have no comparable data on the male side, and we can only speculate as to the transformation that would be produced in the male reproductive system by the action of female hormones beginning before sexual differentiation. Theoretically we would have to assume that the male zygote contains female as well as male factors, but the male zygote may not be capable of such extensive transformation as the female, owing to the embryological fact that the male gonad never forms normally any homologue of the cords of Pflliger in the female, i.e. of the ovarian cortex, whereas the female does form the homologue of the seminiferous tubules before the cords of Pfliiger begin to arise. In the case of the free-martin we do not find that male hormones cause the development of any structure which is not represented embryologically in the normal female; the hormones act in this case merely by inhibition or stimulation of normal embryonic rudiments. If this should hold as a principle also on the male side we could not expect that the transformation of a testis into an ovary should ever occur, although suppression of complete testicular development would probably happen. But apart from this admittedly uncertain principle there is no reason for assuming that the possibility of the male zygote for acquisition of female characters may be any less than the reverse case.

For their proper evaluation the results concerning the freemartin should be associated with the other studies indicating that


420 FRANK R. LILLIE

zygotic determination of sex is not irreversible predestination, but a quantitative overbalance. in the direction of one sex on the other. I refer more especially to the studies of Richard Hertwig ('06), Kuschakewitsch ('14), of Whitman-Riddle on pigeons (Riddle '16), and especially the very convincing demonstration of reversibility of sex in the gypsy moth by Goldschmidt ('16). These studies demonstrate that sex differentiation is controlable within variable limits in certain groups, to which mammals must be added as a result of the present study. To determine the means, limits and subsequent results of such control is now one of the most important tasks of biology.

LITERATURE CITED

Allen, B. M. 1904 Embryonic development of ovary and testis of mammals. Am. Jour. Anat., vol. 3, pp. 89-146.

Allnatt, R. H. 1836 On hermaphroditism in the mammalia; case of the freemartin. London Medical Gazette, vol. 18, pp. 528-530.

Baltzer, F. 1914 Die Bestimmung und der Dimorphismus des Geschlechtes bei Bonellia. Sitzungsb. d. phys.-Med. Ges. zu Wiirzburg.

Bateson, William 1913 Problems of genetics (see pp. 44-45). Yale University Press.

Bell, Alexander Graham 1904 Sheep Catalogue of Beinn Bhreagh, Victoria County, Nova Scotia. Washington, D. C. Also mimeographed supplement to same to 1914.

19C4 The multi-nippled sheep of Beinn Bhreagh. Science, N. S., vol. 19, pp. 767-768.

1912 Sheep-breeding experiments on Beinn Bhreagh. Science, N. S., vol. 36, pp. 378-384.

Bergschicker, Ad. 1912 Die Miillerschen und Wolffschen Gauge und die Bildung des weiblichen Genitaltractus beim Rind. Arch. Anat. und . Phys. Anat. Abth., pp. 1-54.

Bonnet, R. 1889 Beitrage zur Embryologie der Wiederkiiuer, gewonnen am Schafei. 2. Vom Auftreten der ersten Ursegmente bis zur Bildung der Extremitatstiimmeln. Archiv. f. Anat. u. Physiol., pp. 1-106, Taf. 1-6.

Chapin, Catharine C. 1917 A microscopic study of the reproductive system of foetal free-martins. Jour. Exp. Zool., vol. 23, pp. 453-482.

Cole, Leon J. 1916 Twinning in cattle with special reference to the freemartin. Science N. S., vol. 43, p. 177.

Cribb, J. J. 1823 An inquiry into certain opinions which exist relative to the procreative powers of women who arc twins, the socius in utero, or co-twin being a male. Lon. Med. Rep., vol. 20, pp. 213-216.

Davies, C. J. 1913 Caprine free martins. The Veterinary Journal, N. S. vol. 20, pp. 62-70.


ACTION OF SEX HORMONES IN FOETAL LIFE 421

Fernandez, Miguel 1915 Ueber eiiiige Entwicklungsstadion des Paludo

(Dasypus villosus) und ihre Beziehuiig zum Problem der spezifischen

Polyembryonie des Genus Tatusia. Anat. Anz., Bd. 48, pp. 305-327. Free-jNIartin, The 18S7 Origin of term discussed. British Med. Journ.,

London, pp. 1, 47, 93, 141, 187. Geddes and Thompson 1901 Evolution of sex. 2 ed. London. Walter Scott. GoLDSCHMiDT, R. 1916 Experimental intersexuality and the Sex Problem.

American Naturalist, vol. 50, no. 600, pp. 705-718. (References to

earlier papers of the author here.) GuRLT 1833 Lehrb. der path. Anat. der Haussaugethieren. Tab. 20, cf. Bd.

2, p. 186; Tab. 22. Hart, D. Berry 1908-09 Mendelian action on differentiated sex. Abstr. Proc.

Roy. Soc, Edinburg, July 1909. in extenso. Edin. Obstetr. Trans.

1910 The structure of the reproductive organs in the free-martin,

with a theory of the significance of the abnormalitj". Proc. R. Soc.

Edinburgh, vol. 30, pp. 230-241.

1911-12 Numan, the veterinarian and comparative anatomist of

Utrecht; a forgotten observer on the free-martin. Tr. Edinburgh

Obstetr. Soc, vol. 37, pp. 89-129. Harvey, William 1651 Anatomical exercises on the generation of animals.

Translated l>y Robert Willis, London, Sydenham Society, 1847. Hering 1851 Repert. der Thierheilk, 12 Jahrg., p. 107. Hertwig, Richard 1906 Weitere Untersuchungen fiber das Sexualitatsprob lem. Verhand. d. Deutsch. Zool. Ges., pp. 90-112. Hunter, John 1786 Account of the free-martin: Observations on certain parts

of the animal economy. London, sold at no. 13 Castle St., Leicester

Sq., pp. 45-68. (See also Atlas attached to Palmer's Ed. of Hunter's

Works.) King, Helen Dean 1911 Studies on sex determination in Amphil)ians. IV.

The effects of external factors, acting before or during the time of

fertilization, on the sex ratio of Bufo lentiginosus. Biol. Bull., vol.

20, pp. 205-235. Kuschakewitsch, Sergius 1910 Die Entwicklungsgeschichte der Keim drusen von Rana esculenta. Ein Beitrag zum Sexualitatsproblem.

Festschrift zum sechzigsten Geburtstag Richard Hertwigs Bd. 2, pp.

61-224. LiLLiE, Fr.^nk R. 1916 The theory of the free-martin. Science, N. S., vol. 43,

pp. 611-613. MoNELL, G. C. 1846 Free-martin. X. Y. Journal of Medicine, vol. 6, pp.

83-88. MooRE, J. 1852 Fecundity of a free-martin. Med. Exam. Phila., N. S., 8,

p. 82. Neugeb.\uer, Franz Ludwig 1908 Hermaphroditismus beim Menschen.

Klinkhardt, Leipzig, pp. VII 748. Newman, H. H. 1917 The Inology of twins. University of Chicago Press. Numan 1843 Verhandeling over de onvruchtbaren runderen bekend onder den

naanmvan Kweenen in verband tot sommige andere Dieren mit mis vormde Geslachtsdeelcn mit 23 platen; Utrecht, W. van der Monde,

pp. 40, 85.

THE JOURNAL OF EXPERIMENf-^L ZOOLOGY, VOL. 23, NO. 2


422 FRANK R. LILLIE

NuMAN 1844 Translation without the illustrations: "Memoire sur les vaches

steriles" etc. Jour, veterinaire et agricole de Belgique, T. 3. Parona Degli organi riproditori d'una vaccatoro, o free-martin degli inglesi;

descrizione e considerazione. 8°. Atti d. Soc. Ital. di. Sc. Nat., v. 19. Pearl, R. 1912 A case of triplet calves, with some general considerations

regarding multiple gestation in normally uniparous animals. An.

Report of the Maine Agr. Exp. Station for 1912, pp. 259-282. Pearl, R. and Surface, Frank M. 1915 Sex studies VII. On the assumption

of male secondary characters by a cow with cystic degeneration of the

ovaries. An. Report of the Maine Agr. Exp. Station for 1915, pp.

65-80. Pick, Ludwig 1914 Ueber den wahren Hermaphroditismus des Menschen und

der Saugethiere. Arch. f. mikr. Anat., II Abth., Bd. 84, pp. 119-242. Riddle, Oscar 1916 Sex control and known correlations in pigeons. American

Naturalist, vol. 50, pp. 385-410 (references to earlier papers of the

author here). Rueff 1851 Repertor d. Thierheilkunde von Hering, 12 Jahrg., p. 103. Sanson, A. 1880, 1881 Sur la Sterilite des genisses jumelles des taureaux.

Bull. Soc. Centr. de med. vet., 4 Ser., T. 3, p. 132; T. 4, p. 103. Scarpa 1845 Opere varie del cav. A. Scarpa, etc. Vannoni, Firenze, pp. 175 179. See also Mem. della Societa Italiana, T. 2, p. 846. Schmalz, R. 1911 Die Structur der Geschlechtsorgane der Haussaugethiere,

mit anatomischen Bemerkungen. Berlin, 388 pp. See also in Ellen berger's Hand]>uch der Histologie. Simpson, Sir J. Y. 1836 'Hermaphroditism.' Todd's Cyclopaedia of Anat.

and Physiol., vol. 2. (Some free-martins briefly described.)

1844 On the alleged infecundity of females born co-twin with males;

with some notes on the average proportion of marriages without issue

in general society. Edinburgh Medical and Surgical Journal, vol. 61,

pp. 107-119. Spiegelberg, Otto 1861 Ueber die Verkiimmerung der Genitalien bei (ange blich) verschieden geschlechtlichen Zwillingskalbern. Zeitschr. f.

rat. Med. Ill Reihe, B. 11, pp. 120-131. Welles, J. 1831 Free-martin. Boston M. & S. J., 3, p. 238. Steinach, E. 1910 Geschlechtstrieb und echt sekundare Geschlechtsmerk male als Folge der innersekretorischen Funktion der Keimdrlisen. I.

Praexistente und echt sekundare Geschlechtsmerkmale. II. Ueber

die Entstehung des Umklammerungsreflexes bei Froschen. III. Entw.

der vollen Mannlichkeit in funktioneller und somatischer Beziehung

bei Siiugern als Sonderwirkung des innercn Hodensekretes. Zentral blatt f. Physiol., Bd. 24, S. 551.

1912 Wilkiirliche Umwandlung von Saugctier-Mannchen in Tiere mit ansgepriigt weiblichen Geschlechtscharakteren und weiblicher Psyche. Pfhigers Arch. f. d. ges. Physiol., Bd. 144, pp. 71-108.

1913 Feminierung von Mannchen und Maskulierung von Weibchen. Zentralblatt f. Physiol., Bd. 27, no. 14, S. 717-723.

Whitehead, R. H. 1904 The embryonic development of the interstitial cells of Leydig. Am. Jour. Anat., vol. 3, pp. 167-182.


ACTION OF SEX HORMONES IN FOETAL LIFE


423



Fig. 1" Cow's uterus strongly distended with formalin; ventral view; parts cut out. X f. 1, left ovary; 2, right fallopian tube; 3, left horn of uterus cut open; 4, opening from right horn into body of the uterus; 5, body of the uterus; 6, cervix; 7, vagina.


Drawings made by Kenji Toda.