# Popular Science Monthly/Volume 64/December 1903/Recent Theories in Regard to the Determination of Sex

(1903)
Recent Theories in Regard to the Determination of Sex by Thomas Hunt Morgan

THE

POPULAR SCIENCE

MONTHLY

DECEMBER, 1903.

 RECENT THEORIES IN REGARD TO THE DETERMINATION OF SEX.

By Professor T. H. MORGAN,

BRYN MAWR COLLEGE.

IT was long believed that the sex of the embryo is determined at a relatively late stage in its development, and therefore it seemed probable that external factors must decide whether the embryo is to become a male or a female individual. Many views have been held as to what these external factors are, and from time to time hopes have been held out that it might be possible to regulate, by artificial means, the sex of the developing embryo.

In the last few years opinion has begun to turn in the opposite direction, and several attempts have been made to prove that the sex of the embryo is determined in the egg. That this must be the case in man seemed to be indicated by the fact that 'identical twins' are always of the same sex. There can be little doubt that such twins come from the same egg, and the presumption is strongly in favor of the view that they represent the separated first two cells of the segmenting egg. These twin embryos are enclosed in the same chorion, which further indicates that they have come from one egg. The 'ordinary twins' of man are no more like each other than are any other two children born at different times. The pair of ordinary twins often consists of a male and a female. Since the embryos that give rise to ordinary twins are subjected to practically the same conditions during their uterine life, and are often, as has been said, a male and a female in a pair, it follows that in man the external conditions that affect the egg, after it has left the ovary or after it has been fertilized, do not determine the sex. A similar and even more remarkable fact is known in the case of the armadillo, Tatusa hyhrida of Paraguay. The eight to eleven young of each birth are always of the same sex. This occurs also, it is said, in another species, Tatusa novemcinta. In the latter case it was found by Jehring that all the embryos of one birth are enveloped in a common chorion, although each has its own separate placenta. It is probable that these embryos are the product of a single egg that has become separated during the early stages of segmentation into as many parts as there are embryos produced. That separated blastomeres or cells are capable of giving rise to whole embryos has been demonstrated experimentally in recent years for a number of animals.

The following discovery also bears on the same question. A hymenopterous insect, a chalcid bee, Encrytus fuscicollis, lays one or two eggs in the egg of a caterpillar that is to become the host. The egg of the parasite develops inside the body of the young caterpillar, not into a single embryo, as is the rule, but into a chain of embryos. As many as a hundred embryos may come from the same egg, all united in a common amnion. It has been observed that the bees that emerge from the same caterpillar are frequently of the same sex. Thus in twenty-one observations the progeny was in fourteen cases all of the same sex. In the remaining seven cases both males and females appeared. In the former it is probable that only a single egg had been laid in the egg of the butterfly, and in the latter more than one egg may have been deposited.

One of the earliest and most important of the recent memoirs that have attempted to show that the sex of the individual is determined in the egg is that of Cuénot.[1] This paper deserves first place not only because in point of time it precedes the others to be mentioned, but also because the author has undertaken a considerable number of important experiments that bear on the problem of the determination of sex.

It had been claimed that when young caterpillars are poorly nourished they give rise to a larger number of males, and conversely, when well nourished to a great majority of females. The experiment was first carried out by Landois, and later confirmed by Giard, Treat and Gentry. On the other hand, Riley found that starved caterpillars, as well as those abundantly supplied with nourishment, give both male and female individuals with no greater disproportion in numbers than ordinarily exists. Other observers have recorded similar results. Furthermore, a number of investigators have shown that the sex of the young insect is already determined at the time when it emerges from the egg and even some time before that event. Brocadello's observation is even more important. He discovered that there are large and small eggs laid by the silk-worm moth, and that the caterpillars emerging from the large eggs are, in the great majority of cases (88 to 95 per cent.), females, while those from the smaller eggs give a corresponding majority of males (88 to 92 per cent.). It is therefore clear that the difference in size corresponds to a difference in the sex of the embryos, and that with sufficient care it would be possible to separate the two sorts of eggs so completely that all of one kind would be females and of the other males. A similar result has been obtained by Joseph in another moth, Ocneria dispar. Cuénot states that he has been able to verify completely this statement of Joseph.

How can we explain the apparent contradiction between the results of Landois, Treat and Gentry, and those of Brocadello, Joseph and Cuénot? It is probable that in all cases the facts recorded are correct. Cuénot suggests that in the lots of caterpillars that were poorly nourished there was a large mortality of the young females, so that of those surviving there was a larger percentage of males. If we apply this same view to the case in which abundant feeding gives rise to more females we shall have to assume that here a larger percentage of males are eliminated, but it is not at all evident why this should be the case. Cuénot points out another possible source of error; namely, that in selecting the caterpillars for the experiment the larger ones may have been picked out to be given an abundant diet and the smaller ones for a meager diet. If this had been done consciously, or unconsciously, the results would not be due to the quality of the food, because the young caterpillars that were large were already females (having come from larger eggs), and those that were small were already males (having come from small eggs). It is clear, therefore, that all the early experiments must be repeated and the precaution taken to note the number of caterpillars at the beginning and at the end of the experiment, and the sex of those that die must, if possible, be determined. Care must also be taken that no selection of large and of small individuals takes place. Since, however, it has been so clearly shown by Brocadello and by Cuénot that large eggs become females and small eggs males, it will be desirable in repeating the experiments to take this fact into account, and to attempt to discover if the potentialities of the large and of the small eggs can be changed by external conditions. Here we have a new field for experimental work that will yield results of great interest. The lines are now so definitely drawn, and it is clearly so important to settle this question on as many forms as possible, that it is much to be hoped that in the near future many workers will turn their attention to this important subject.

Cuénot's experiments on certain flies, belonging to four different genera, are of especial interest. In this group also it had been claimed by Lowne that individuals from the large maggots that have been well nourished are nearly always females, while those from small maggots, poorly nourished, are usually males. Cuénot first determined, in the three genera used in his experiments, that normally the number of males and of females is about the same. The results of his experiments, in which the maggots were well fed, were as follows—in Lucilia cæsar 49.27 per cent, of females; in Calliphora vomitoria 51.02 per cent, of females; and in Sarcophaga carnaria 51.62 per cent, of females. It is obvious that the presence of an abundance of food did not produce an excess of females. In another experiment in which the maggots received as small a quantity of food as possible there was great mortality and the pupæ were of diminutive size. The results were as follows—Lucilia cæsar 57.92 per cent, of females; Calliphora vomitoria 57.92 per cent, of females; and Curtonerva pabulorum 26 females and 17 males. It is even more evident from the results of this experiment that starving does not have the effect of producing an increase in the number of males. Several variations of these experiments were made, but the results were always the same. Cuénot also tried to find out if the amount of food taken by the individual during its growth has any effect on the kinds of eggs that are produced. The larvæ of Calliphora vomitoria were starved from their birth until they pupated. They gave rise to twelve males and five females, whose size was scarcely half that of the normal individuals. These dwarf flies, confined in a cage with sweetened water and meat, laid twenty times. The larvæ that hatched were kept in a well-nourished condition, and gave rise to 359 females and 353 males. The results show that the amount of food supplied to the young maggots had no effect upon the relative number of male and female eggs that they produced. It is true that these animals, when poorly nourished, gave rise to only a few eggs, but the relative number of eggs that became male or female remained the same.

Among the earliest experiments that were carried out to show whether the sex of the individual could be determined by external conditions were those of Born in 1881, and of Yung in 1883 and 1885. Born tried to show that more male frogs develop when the fluid containing the fertilizing spermatozoa is more concentrated, but this conclusion has been shown to be wrong. Born also fed the tadpoles of Rana temporaria on a rich diet consisting of water plants and of the flesh of frogs and of tadpoles. A large percentage of females developed which Born attributed to the abundance of food. It was shown, however, by Pflüger in the following year, 1882, that Born's conclusion was erroneous, because, even under normal conditions, female frogs are more numerous. Pflüger found that the normal proportions of females to males is often as high as five to one; and this corresponds also to the proportion sometimes obtained when tadpoles are reared from eggs artificially fertilized.

Yung found that females of Rana esculenta are twice as numerous as males, while Cuénot found, on the contrary, in a different locality that there were more males than females.[2] It is not known whether this disproportion in the sexes is due to the greater mortality of one sex, or whether there are more eggs of one kind than of the other. The results appear to indicate, however, that external conditions do not have a determining influence on sex, and it seems not improbable, although not completely established, that there may be greater mortality among the male tadpoles than among the females in some species and in certain localities.

Cuénot made a few experiments with the eggs of Rana temporaria. He points out that his results are open to the same grave criticism as are those of his predecessors in that he did not determine the sex of those that died. In one experiment in which the tadpoles were given an abundant supply of vegetable food they suffered greatly from crowding and from insufficient aeration of the water. Their development was retarded and they remained small. Of the 26 frogs that metamorphosed all were females. In another similar experiment there were 3 females and 4 males. In a third experiment the tadpoles were placed in a large aquarium supplied with cold, running water. No food was given, and the tails of the tadpoles were frequently amputated in order to prolong the larval period. There emerged 57 young frogs, of which 33 were females, 29 males, and one hermaphrodite.

In a fourth experiment the tadpoles were separated into three lots. The first were given a vegetarian diet; the second were given only animal food; and the third were put into a large aquarium whose floor was covered with mud, but no food was present. The tadpoles that died were no doubt eaten by their companions and thus a certain amount of food was probably obtained.

The first and the second lots developed at the same rate, but the tadpoles did not reach a large size owing to the small dimensions of the aquarium. They became frogs after two months. Those of the third lot on the contrary were retarded in their development; they remained small and began to die from hunger after the third month. They were then given animal food; they grew rapidly and metamorphosed a month later, i. e., four months after hatching. The frogs were of small size and showed signs of having been poorly nourished. Of the 35 survivors of this third lot 23 were females and 12 were males. Of the 36 survivors of the second lot (with animal diet) there were 14 females and 22 males. Of the 108 survivors of the first lot (vegetarian fed) there were 51 females and 57 males (and 8 not differentiated). The proportionate number of females to males in all the tadpoles reared in these experiments is not different from that which Cuénot determined in nature. He concludes from his results that the sex of the frog is not influenced by the external conditions (especially of food) to which the tadpoles are subjected.

Pigeons have also furnished some interesting facts in regard to problem of sex. From the time of Aristotle it has been recognized that of the two eggs laid in each batch one generally produces a male and the other a female. Nevertheless numerous exceptions have been recorded in which both individuals were of the same sex. Cuénot himself found in eight sets that in two instances there were two males; in two instances there were two females, and in five instances there was a male and a female.[3] It has been claimed moreover, and the tradition also goes back to the time of Aristotle, that the first egg laid gives rise to a male and the second to the female. Flourens confirmed this fact for eleven sets, and Cuénot found the same result. The meaning of this is obscure, for it may be that a male egg is first set free, or that the conditions to which the first egg that is laid is subjected are such that it becomes a male. The former interpretation may appear to be the more probable, but it is not conclusively established by the facts.

Although many statistics have been brought together in regard to the determination of sex in man and in other mammals there is no convincing evidence showing that external factors determine the sex of the embryo; and, as has been pointed out, there is strong evidence pointing in the opposite direction.

If we turn now to some of the lower animals we shall find that there are a few indisputable cases in which it has been shown that the sex of the individual is predetermined in the egg. It was discovered by Korschelt that two kinds of eggs are produced by a small worm, Dinophilus apatris, and that the larger eggs develop, after fertilization, into females and the smaller into males. The females are about 1.2 mm. long, while the males are only 0.04 mm. long. The males are degenerate in structure; they are less numerous than the females, and live only ten days, whereas the females live a month or more.

A similar difference in the size of the eggs that produce males and females is found in certain rotifers, in Hydatina senta for example. In this species there are three kinds of females distinguished by the different kinds of eggs that they lay. One lays large eggs which without fertilization produce females. Another lays small eggs, less rich in yolk than the last, and these eggs, also without being fertilized, produce males. A third kind of female produces the winter eggs, which are fertilized by the males and give rise to females. In this rotifer the sex of the egg is determined while the egg is still in the ovary, and Nussbaum has made the important discovery that the amount of nourishment taken by a young female, between the time of her emergence from the egg and the deposition of her first egg, determines which kind of eggs she will subsequently produce. If she has been well nourished in this interval she produces eggs that become females, but if poorly nourished she produces male eggs. After the eggs have been once formed no subsequent change of food or of temperature can alter the kind of eggs that are produced. It has not been determined why some females produce parthenogenetic eggs and other females winter eggs that are to be fertilized. Nussbaum thinks that the effect of an early union with a male, combined with insufficient nourishment during the first hours of free life, determine that winter eggs are to be produced.

Amongst crustaceans and insects there are several instances known in which the sex of the individual appears to be connected with certain kinds of eggs. The water fleas, or daphnids, produce during the summer small parthenogenetic eggs with a thin shell which develop into parthenogenetic females,[4] but under certain conditions males and females appear. The females produce large winter eggs which are fertilized and produce in the following year only female daphnids which start the parthenogenetic summer broods. The sex of the winter eggs is probably determined in the ovary, since the eggs show their characteristic structure before they are set free. The appearance of the male and female generation is supposed to be connected with the change in temperature, or more probably with a change in the amount of food. Under these conditions, as has just been said, eggs that produce males and females are formed. Here it would appear that an external condition determines the appearance of the male and of a different kind of female.

Similar facts are known for the aphids, or plant-lice. If conditions are favorable, i. e., if they are kept warm and have an abundance of succulent food, they continue indefinitely producing wingless parthenogenetic females. But if the food becomes scarce or dry, then winged males and females arise from the parthenogenetic eggs; these unite, and the fertilized winter eggs are laid. From these eggs the wingless parthenognetic females arise in the following spring.

The life history of Phyloxera vitifolii, which is parasitic on the roots of the grape, is as follows: A series of parthenogenetic wingless females succeed each other, until at the end of June the last generation of these produces, parthenogenetically, winged females that are capable of migrating. These also produce parthenogenetic eggs of two kinds, small ones from which winged males develop, and larger ones from which winged females arise.[5] Union of the sexes now takes place and each female lays one egg which gives rise in the following spring to the parthenogenetic wingless female that lives on the root of the grape vine.

We come now to the much discussed case of the hive bee. There are here three kinds of individuals: the queen which lays all the eggs; the workers, which are immature females and do not reproduce at all, and the drones or males which fertilize the eggs of the queen.

It has long been believed that when an egg of the queen is fertilized it gives rise to a female (either queen or worker according to the kind of food given to the young maggot), but if not fertilized the egg gives rise to a male. It has been generally assumed, in accordance with this belief, that all the eggs are alike and will produce males if they are not fertilized, but females if they are fertilized. It is known moreover that the cells of the comb in which the queen deposits the eggs that are to become males are different from the worker cells, and this fact is generally interpreted to mean that the queen is capable of determining the sex of her offspring by allowing or preventing the fertilization of the egg. The sperm which was received by the young queen at the time when she left the hive with a swarm is stored up in a special sac or receptacle with muscular walls and an outlet that opens near the oviducts. It is generally assumed that the queen squeezes out the sperm when an egg that is to be fertilized is laid, but does not do so when a male is to be produced. Some writers have marveled at this wonderful power, that seems almost akin to intelligence, by which the queen determines 'at will' the sex of her offspring, but this may give an entirely exaggerated idea of what takes place, for the act may be a very simple reflex. It has been shown by Drory that if the queen is supplied with an artificial comb containing only drone cells she may be forced to lay in them fertilized eggs that become workers. Conversely, if supplied with worker cells only she will sometimes lay unfertilized eggs in them. This has been interpreted to mean that there are really two kinds of eggs that are laid by the queen, male and female, and that only the latter are capable, as a rule, of being fertilized. On this assumption we should be forced to conclude either that the queen can determine which kind of egg is to be laid and places it in its proper cell, or that she has a knowledge of which kind of egg she is about to lay next, and seeks the proper cell to deposit it in. There are, however, some further facts that show that the conditions may be more complicated than has been generally supposed.

It has been possible to introduce a virgin queen of an Italian stock into a hive containing workers and males of a German stock. These two kinds of bees are sufficiently different to be readily distinguished from each other. The Italian queen becomes fertilized by the German males. In consequence all the queens and workers that come from her eggs are hybrids, since they come from fertilized eggs, but the males or drones are nearly all of the same kind as the queen, which indicates that they have come from unfertilized eggs. Occasionally, however—and this is the point of special interest in the present connection—a few males appear that are hybrids, as Dzierzon long ago observed. Hence we must suppose that an egg has been fertilized, and despite this fact it has developed into a male. This conclusion may indicate, as Beard has recently claimed, that the sex of the egg must have been already determined, and was not altered by the accidental entrance of a spermatozoon.

In this connection it should be pointed out that Weismann and Petrunkewitsch found that out of 272 drone eggs that they studied there was one that had been fertilized. Whether it would have become a male or not, could not be determined; for it is said that the queen sometimes makes a mistake and deposits a worker egg in a drone cell. Indeed 'whole combs of drone cells may produce workers instead of drones.'

These are some of the principal facts that seem to show that the sex of the individual is predetermined in the egg. From the evidence Cuénot arrives at the following general conclusions: He thinks that in the great majority of animals the sex is determined in the egg and at latest when the egg is fertilized. In no instance, he claims; has it been shown that the sex of the individual can be determined later than fertilization. The classic examples, insects and frogs, in which it was supposed that external conditions acting on the later embryo determined the sex, have been shown to be capable of a different interpretation. It has been especially made clear, Cuénot claims, that a meager or an abundant supply of food has no influence on the determination of the sex of the embryo. He believes moreover that it is the egg and not the spermatozoon that determines the sex of the individual. In several insects, in Dinophilus, in pigeons, and in the winter eggs of aphids and of daphnids, this has been clearly shown to be the case. In other animals, as in the rotifers and in the social hymenoptera, the spermatozoon appears to have a determining influence. In the mammals the entrance of the spermatozoon may have only the same influence as that of the egg itself. It will be observed that there is a certain catholicity in these conclusions at which Cuénot arrives, and in the present uncertain state of our knowledge on many important points it is probably wiser not to take too narrow a point of view in regard to what factors determine the sex of the individual.

Strasburger[6] has recently arrived at a somewhat similar conclusion in regard to sex, basing his evidence mainly on certain observations and experiments in higher plants. He, too, concludes that the sex of the individual is determined in the egg, but he does not attempt to push the question further than this general statement.

Lenhossek[7] has also discussed in a more popular form the question of the determination of sex, and he likewise urges that the sex of the individual is determined in the egg. His discussion of the relative number of males and females born in the human race is particularly instructive, but it would carry us too far here to discuss the conclusions at which he arrives.

Born[8] has carried out a series of experiments with mice, and finds that the amount of food given to the parents produces no effect on the relative numbers of males and females born. He also finds that the age of the parents has no effect, nor has close interbreeding. He arrives at the conclusion that the sex of the higher animals and plants is determined in the egg.[9]

In striking contrast to the general conclusions of Cuénot, Strasburger, Lenhossék, and of Born there are two more recent theories in which an attempt has been made to describe in detail how the sex of the individual may be determined in the egg. Beard's paper,[10] published in 1902, may be said, in a sense, to take up the problem where it was left by Cuénot. He attempts to bring the problem of the differentiation of the sexes into connection with the recent work relating to the origin of the reproductive cells or gametes. Beard tries to show that there are not only two kinds of eggs, but also two kinds of spermatozoa that correspond to the two kinds of eggs. It is supposed by him, however, that the determination of sex rests entirely with the egg, and that the spermatozoa do not have any influence on sex-determination. It is assumed, moreover, that one of the two kinds of spermatozoa has lost its power of fertilizing the egg and in most cases has become degenerate.

In respect to the occurrence of the two kinds of spermatozoa, Beard brings together rather a heterogeneous collection of facts. It has been known for some time that in a few cases two kinds of spermatozoa are found. The oldest and now the most thoroughly studied example is that of the snail, Physa vivipera. In this animal there are hair-like spermatozoa that resemble the ordinary forms of spermatozoa, and also worm-like forms which are as numerous as the other kind. A remarkable fact has recently been discovered by Meves in regard to these spermatozoa. An unusual and probably degenerate process occurs in the formation of the worm-like spermatozoon, so that instead of containing the reduced number (seven) of chromosomes it contains but a single one. In another form, Pygæra, this second form of spermatozoon contains no chromatin material whatsoever, i. e., it is headless and presumably functionless as well.

In the long list of cases given by Beard in which two forms of spermatozoa have been described, there are several cases in which the two distinct forms appear to be always present and characteristic, as in the cases cited above; but he has also included some other cases in which giant spermatozoa occur, and some of these at least have been shown to be the result of a failure of the spermatocytes to divide. Until it can be shown that this failure to divide is usual and characteristic of one set of these spermatocyte cells the result may really have no bearing at all on Beard's contention.

Much more striking are the cases in which there is an accessory chromosome present in two of each of the four cells that develop from a single spermatogonial cell. The discoveries of McClung, Montgomery and Sutton in this connection indicate that there are two kinds of spermatozoa, and McClung has urged that this difference is connected with the determination of sex; but there is nothing more than the supposition that this may be so to go upon at present. In these cases, although the form of the spermatozoa is the same for the two kinds, there appears to be a difference in the amount of the chromatin material. It has not been shown that a difference of this kind would have any value in the determination of sex, and even if this were the case the results do not conform to the requirements of Beard's theory, as we shall see presently.

Beard calls attention to the fact that in nearly all the cases in which two kinds of spermatozoa have been described there is evidence of the degeneration of one of the two kinds. Prom this he draws the rather sweeping conclusion that throughout the animal kingdom one of the two forms of spermatozoa has become suppressed. He arrives at this conclusion in the face of an overwhelming body of evidence to the contrary, for in the great majority of forms all the spermatozoa that are formed develop in the same way and are, so far as we can see, capable of fertilizing the eggs.[11]

Beard's conclusions in regard to the determination of sex may be summarized as follows:

1. The sex of the individual is determined in the egg before fertilization.

2. The determination of sex probably takes place at the time of the reduction in the number of chromosomes.

3. Each egg and its two polar bodies are potentially of the same sex, either male or female.

4. A corresponding differentiation of the primary germ-cells takes place in the male. An early separation of the spermatogonial cells into male and female occurs. After this each cell may continue to divide, but remains of the sex that it has acquired in the differentiating division. Finally each of these cells produces four spermatozoa. This division is comparable to the one in the egg-series when the polar bodies are given off, so that each group of four spermatozoa corresponds to a female egg and its three female polar bodies, or to a male egg and its three male polar bodies; but in the cases of the spermatozoa the individuals are supposed to be without sexual qualities. It is the egg alone that determines the sex.

5. One set of these fourfold groups of spermatozoa Beard supposes to have become functionless, in the sense that even if it develops the spermatozoa have lost the power to fertilize the eggs. The other spermatozoa are functional so far as fertilizing the egg is involved, but, as stated above, take no part in the determination of sex.

Beard also advances certain views in regard to parthenogenesis. The sex of the individual that develops from a non-fertilized; i. e., from a parthenogenetic egg, is not in any sense a consequence of the non-fertilization of the egg. Whether the individual is a male or a female depends entirely upon whether a male or a female egg has been produced. Whenever we find long series of parthenogenetic females, as in the aphids, developing from and also producing parthenogenetic eggs, Beard supposes that only female eggs have been produced in the ovary, and that the male eggs, which have appeared in one at least of the first generations of the germ-cells 'must be either delayed in their ripening or suppressed.' Here we meet with a paradox that is so patent and touches such a fundamental point of Beard's theory that it is more than surprising that he has said nothing about it. If the female aphid develops from a female egg (the polar bodies of which are on Beard's theory also female), we can understand why in the next generation she must give rise to female eggs, but why should males ever be again produced? Since it has been established beyond question that these parthenogenetic females do produce both males and females at the end of the summer, the question is where have the male eggs come from?[12] Beard appears to take for granted that a female egg can give rise to cells that become male eggs. If so his theory can have very little if any value, since the entire conception on which it rests, namely, the separation of the male and the female eggs at one division, is rendered valueless, I think, by the assumption that after such a thing has once taken place a female cell may in the next generation give rise to male eggs. Furthermore Beard's assumption, that the separation of the male and the female eggs occurs at the time when the reduction in the number of the chromosomes takes place in the egg, is pure guess-work, and not very good guessing either, for certain recent work indicates that the reduction in the number of the chromosomes involves a process that can have no conceivable connection with the separation of the male from the female elements of the egg. On the whole it does not appear that Beard has offered a very convincing theory as to how the determination of the sex of the individual is accomplished.

Castle[13] also has recently advanced certain hypotheses in regard to the determination of sex. In certain superficial respects his view appears similar to that of Beard, but closer scrutiny shows that the two views are essentially different in many important points.

Castle assumes that there are two kinds of eggs, male and female, and two kinds of spermatozoa, male and female. He supposes that both kinds of spermatozoa are functional in the sense that each carries with it the possibility of determining the sex of the individual, and each spermatozoon is also capable of fertilizing an egg, but a male spermatozoon can fertilize only a female egg and a female spermatozoon a male egg. It is evident, therefore, that Castle's idea in regard to the spermatozoa is fundamentally different from that of Beard. Furthermore, Castle supposes that the separation of the male from the female qualities of each egg takes place at the time when the second polar body is extruded, and, in consequence, the egg and one of the polar bodies will be female and the other two polar bodies male, or if the egg remains female, one polar body will be female and the other two male. Similarly for the spermatozoa; two of each four (formed from the first spermatocyte) are female, and two are male.

The evidence on which Castle rests his assumption that there are two kinds of spermatozoa, as well as two kinds of eggs, is contained in the following statement: "That sex is borne by the egg is shown clearly by the case of parthenogenetic animals, which without the intervention of a male produce young of both sexes. That the spermatozoon also bears sex is manifest in the case of animals like the honey bee, for the egg of the bee, if unfertilized, invariably develops into a male, but if fertilized into a female." The finality of the conclusions drawn from these facts is by no means above question.

Perhaps the most distinctive part of Castle's paper is his attempt to apply the much-discussed Mendel's law to problems of sex-determination; an idea that had suggested itself to Bateson and Saunders, but had been rejected, because the 'distribution of sex among first crosses shows great disparity from the normal proportions.' Castle does not admit however the force of this objection.

A specific example may be the simplest way of illustrating Mendel's law and its application to sex as maintained by Castle. If a white mouse is crossed with a wild gray mouse all the offspring of this first cross will be gray like the wild mouse. The gray color of the gray mouse is said to be dominant and the white color (inherited from the other parent) does not appear, but is supposed to be present in a sort of latent condition. It is said to be recessive. If now these primary hybrid mice are interbred some of their young will be white and the rest gray in the proportion of one to three. If these white mice, when they become grown, are interbred their offspring will always be white as well as all their subsequent descendants. Some of the gray mice will also breed true, but the rest that are gray hybrids will, if interbred, give rise to some white and some gray in the proportion again of one to three. This is only a partial statement of Mendel's law, but will suffice for our present purposes.

The explanation that Mendel offered to account for the proportionate number of individuals that inherit the dominant and the recessive characters is very simple and is probably correct. As applied to our illustration of the mice it would be as follows: When the egg of the white mouse is fertilized by the spermatozoon of the gray mouse the fertilized egg and all the cells into which it divides contain chromatin material in the nucleus half from the white and half from the gray parent. The dark element dominates whenever the two are together, hence the first generation of hybrids are all dark. The cells of this primary hybrid that have gone into the reproductive organs (in the female into the ovary and in the male into the testis) are supposed to be at first like all the other cells of the body, and contain both white and dark elements.[14] But at some time in their later history, and presumably at the time when the egg sends off its polar bodies and at the time when the four spermatozoa are formed, a separation of the dark from the white elements occurs, so that two cells of the one kind and two of the other are formed. Thus the germ-cells are, as it were, purified, and consist of those that contain only white and of those that contain only dark elements. This is supposed to be the condition of the germ-cells in the ovary and in the testis of the primary hybrids.

Suppose now that these hybrids breed together, the white and the black spermatozoa will meet the white and the black eggs, and since it is a question of chance alone how they will come together, all possible combinations will be made. When a white germ-cell meets with a white one, a white individual results, and since it contains only white elements all its descendants will be white (if it is bred, of course, to white individuals). If a gray germ-cell meets a gray germ-cell, a gray individual will result, and all its purely bred descendants will be gray. If, however, a white and a gray germ-cell unite, the individual that develops will contain both elements in all its body-cells and, since the gray always dominates in such combinations, the individual will be gray, but will have the white as a recessive character that may crop out in subsequent generations. On the theory of chance combinations there will be twice as many of these gray-white individuals as of the white or of the pure gray. The series stands 2:1:1. Since in outward appearance all the gray-white mice are like the pure gray, we get three grays to every one white.

Let us now return to Castle's theory and see how he tries to make an application of Mendel's principle to sex. Just as there are two kinds of mice in our illustration, white and black, there are two kinds cf sexual individuals, males and females. It is now assumed that the germ-cells, when they reach their final divisions, separate their male from their female elements, giving pure male and pure female eggs, and pure male and pure female spermatozoa. If, as in the mice, all chance combinations of the germ-cells are possible, there will result three kinds of individuals in the proportion of 2:1:1. The first of these, that are twice as common as either of the other two, would be sex-ybrids. If we assume, as in the mice, that one character always dominates in such a combination, the male let us say, there would be twice as many males of this hybrid kind as there are individuals of either of the other two pure kinds, and since there are as many pure males as there are pure females, there would be in all three times as many males born as females. Since we know that there is no such disproportion of one sex to the other, it appears absurd to attempt to apply Mendel's law to the problem of sex. Castle is therefore obliged to make a further assumption to avoid this difficulty. He assumes that a male spermatozoon can fertilize only female eggs, and a female spermatozoon only male eggs. There is no evidence known at present supporting this assumption, but it must be admitted that it can not be disproved, however improbable it may appear. On this view every fertilized egg is a sex-hybrid, and may give rise to a male or to a female according to which element dominates. Thus we return once more to our original question as to what determines the sex of the individual. We shall see presently that Castle fails to meet this fundamental question.

There is one result that Castle cites, which he claims indicates that his assumption that the eggs may show a selective power towards certain of the spermatozoa is not unwarranted. He found some years ago in the ascidian, Ciona intestinalis, that the eggs of one individual can not be fertilized by the sperm from the same individual, except very rarely. This case is cited as indicating that successful fertilization depends upon unlikeness between the gametes that unite. I have repeated this experiment on Ciona and have confirmed in large part this result, but, unfortunately for the point of view, I found in other ascidians that this relation does not hold. In Molgula, for example, the eggs are perfectly fertile with sperm from the same individual. Furthermore, by making the sperm of Ciona more active by adding ether to the water, I have been able to make them, under certain conditions, fertilize all the eggs of the same individual. In the light of these facts I do not think the conditions in Ciona can be given the interpretation that Castle has applied to them.

There is another side of Castle's hypothesis that must be briefly referred to, since he suggests a way of meeting a difficulty that is fatal to Beard's theory. I refer to parthenogenetic development and to the production at the end of a parthenogenetic series of male and female individuals. Castle supposes that in parthenogenetic reproduction the female character dominates over the male, when the two are present together, and that when a separation of the sex-characters takes place it does so at the time of the formation of the second polar body in the egg, and probably at the corresponding state of development in the spermatozoon. There is a fact in this connection, the bearing of which Weismann was the first to fully appreciate, namely, that the parthenogenetic eggs of daphnids and of some rotifers give off only one polar body, while eggs that are to be fertilized give off two polar bodies. Castle suggests that the second polar body is the female gamete, hence when it is given off the egg must become a pure male if it develops. If this polar body should be retained in the egg the conditions are exactly the same as when a female spermatozoon enters a male egg. Hence, since the female element dominates in these animals when the two sexes meet, the individual must become a female. Since therefore such an egg carries the male element in a recessive form, this element may, if it becomes separated from its female associate, give rise to a male. In this way the theoretical difficulty referred to above is met.

Let us follow out a little further the applications of this view. It is probable in the honey-bee that all the eggs give off two polar bodies. Consequently unfertilized eggs must produce pure males. If they are fertilized by female spermatozoa they will give rise to females, and, on the hypothesis, only female spermatozoa can enter male eggs.

In rotifers and certain crustaceans only one polar body is given off, but since this is the first polar body it does not involve the question of sex. Consequently the parthenogenetic eggs in these forms are sex hybrids. If at any time the conditions change so that one or the other sex element dominates, males and females may arise. But why the female element should dominate in some eggs and the male in others is not explained, and thus we are in exactly the same predicament as we were before Castle's hypothesis was proposed.

One case of special difficulty should not pass unnoticed since Castle has made an interesting suggestion that appears to clear up a difficulty, provided the facts on which the conclusion rests are confirmed. The eggs of the honey-bee extrude two polar bodies, as we have said, and hence are purely male. It is assumed that these males must produce spermatozoa that are female. This is a necessary assumption, because the eggs of the bee having extruded their two polar bodies are purely male, but become female after they are fertilized. Therefore the spermatozoon must bring the female element into the egg. Castle tries to meet this difficulty of the formation of female spermatozoa in a purely male individual by reference to a recent observation of Petrunkewitsch, namely, that the reproductive organ of the male bee develops not from the egg itself, but from the second polar body which fusing with one of the first pair reenters the egg. This second polar body is, on Castle's theory, purely female, hence the spermatozoa must be female. The ingenuity of the explanation is admirable, and rescues the theory from a fatal objection, but of course even if Petrunkewitsch's results are accurate (and they are not above suspicion) it by no means follows that the spermatozoa that come from the second polar body are female, as Castle assumes. The facts in regard to the parthenogenesis, and in regard to the special case of the bee, may possibly be given a much simpler explanation than that which Castle applies to them. For instance, if in certain insects the addition of the chromatin material of a spermatozoon (or what amounts to the same thing the chromatin contained in a polar body) determines that the egg shall become a female, we can explain the results without complicating the problem by the assumption of male and female spermatozoa, and male and female eggs. Castle's theory appears needlessly complex, and the whole attempt to apply the Mendelian principle to the question of sex does not appear to have been very successful. The weakest side of the theory has already been spoken of, namely, that it fails to account for the very problem that a theory of sex should explain, namely, the problem of what it is that determines whether an egg that contains both potentialities becomes a male or a female.

The reaction that has set in against the old view, that the sex of the embryo could be determined at a relatively late stage in development, is no doubt in the right direction. It has been shown in several cases by recent discoveries that the sex of the embryo is already determined in the fertilized egg, and in other cases it appears to be determined even before fertilization, but this need not mean that there are male and female eggs, and male and female spermatozoa. We have just examined two recent theories that rest on assumptions of this kind and have found, in my opinion, that they are both unsatisfactory. Let us see whether it may not be possible to bring under one point of view the old and the new discoveries in regard to the determination of sex, and construct a hypothesis that does not involve the idea that there is separation of the primoidia of sex in the germ-cells.

1. It has been shown in a few cases that two kinds of eggs are produced which become male and female individuals, in some cases with, in others without, fertilization. It may be erroneous to conclude from these facts that the eggs themselves are male and female in the sense that the elements (primoidia) that determine the sex of the embryo have become separated and confined to male or to female eggs. In a case like that of the silk-worm, where a graded series exists, the size of the egg appears to be the determining factor in respect to which sex develops, not that the female sex-elements are found only in the large eggs, and the male elements in the small eggs. It seems more reasonable to assume on the contrary that both elements are present in all kinds of eggs. In other cases other factors than that of size determine which sex develops.

In regard to the two forms of spermatozoa that have been found in a few species, there is no evidence that one sort contains only the primoidia of a male individual and the other kind those of the female. In those arthropods in which an accessory chromosome has been found we have no evidence to show that this chromosome is the male or the female element, and so long as we know nothing at all in respect to the conditions in the egg it is useless to speculate further on these cases. 2. There is experimental evidence pointing to the conclusion that factors, external to the egg itself, may determine in some species what kind of eggs will be produced, as in Hydatina, and in the aphids where a change in the food causes the appearance of males and females. In bees the addition of the chromatin of the spermatozoon appears as a rule to determine that the egg gives rise to a female.

In other cases it appears that the addition of the chromatin in one of the polar bodies may accomplish the same result. Here the relation may be purely a quantitative one. In other animals the addition of the spermatozoon to the egg is not, it appears, the factor that determines the sex.

3. It is known in bees and in butterflies that individuals sometimes appear that are male on one side of the body and female on the other side. The explanation of this peculiarity may be found in the unusual way in which the nucleus of the fertilized egg is divided. If, for instance, all or most of the chromatin brought in by the spermatozoon should be carried into one of the first formed cells along with half of the chromatin of the egg-nucleus, then all the cells that descend from this cell may develop female characters, and all those from the other, male characters. This need not mean that the spermatozoon has brought into the egg female characters that dominate in all the cells in which it is contained, but only that those cells that contain more of the chromatin differentiate their female characters, and all those cells that contain less chromatin differentiate their male characters only.

4. Having discovered that the sex is already determined in the unfertilized egg in some cases, and in others that it is connected with the process of fertilization, the question at once suggests itself whether the determining influence comes from the nucleus or from the cytoplasm. At present we have no conclusive evidence pointing in either direction. That the quantity of the nuclear material may be important seems probable in the case of the bee. That the size of the egg, which is due to a greater amount of cytoplasm, may be a factor in the result seems in other cases to be important, but so long as we do not know what relation the nucleus bears to the cytoplasm in these forms we can not decide as to the meaning of greater volume as a sex determinant. If, as seems highly probable, identical twins come from halves of the same egg, then since the pairs may be of either sex it seems to follow that the absolute size of the egg is not a factor. Whether in these cases the relative amount of chromatin in the nucleus enters into the problem remains to be shown.

It should be pointed out that while we must suppose that the influences in the embryo that control the development of one or of the other sex reside, or have resided, in the nucleus of the egg, this is a different question from that as to whether the nucleus or the cytoplasm of the egg determines which of the two possibilities that are potentially present in the nucleus shall be awakened.

5. Our general conclusion is that while recent theories have done good service in directing attention to the early determination of sex in the egg, those of them which have attempted to connect this conclusion with the assumption of the separation of male from female primoidia in the germ-cells have failed to establish their point of view. The egg, as far as sex is concerned, appears to be in a sort of balanced state, and the conditions to which it is exposed, even when it is not fully formed, may determine which sex it will produce. It may be a futile attempt to try to discover any one influence that has a deciding influence for all kinds of eggs. Here, as elsewhere in organic nature, different stimuli may determine in different species which of the possibilities that exist shall become realized.

1. Bulletin Scientifique de la France et de la Belgique, XXXII., October, 1899.
2. Boulenger also found in Pelodytes, Pelobates and Bufo an excess of adult males.
3. There is here probably a misprint since 2 ${\displaystyle +}$ 2 ${\displaystyle +}$ 5 equal nine.
4. Lenssen claims that the parthenogenetic female eggs do not give off a polar body, and that the male eggs give off only a single polar body. Whether this difference may have any relation to the sex of the individual will be discussed later.
5. The small eggs appear to be laid by the smaller winged individuals and the larger eggs by the larger individuals.
6. Biologisches Centralblatt, XX., 1900.
7. 'Das Problem der geschlechtsbestimmenden Ursachen,' Jena, 1903.
8. Sitz. phys. med. Gesell. Würzhurg, 1902.
9. Born points out that while it has been shown experimentally that in hermaphroditic animals and plants, especially in the lower groups, it is possible to cause one or the other kind of sexual organs to develop or to be suppressed, that in the higher forms, at least in those in which the sexes are separated, it has been found that the sex can not be changed by external factors.
10. Zoologische Jahrbücher, XVI., October, 1902.
11. Meves has recently found in a male bee that two kinds of spermatids are formed by an unequal division of the spermatocyte. The smaller of the two, although it begins to undergo changes similar to those which in the larger one produce a spermatid, appears to be arrested in its development before the change is completed. Mittheilungen Verein Schlesw.-Holst. Aerzte, XI., Mai, 1903.
12. The same paradox appears wherever a female contains male eggs.
13. Bulletin Museum Compt. Zoology at Harvard College, January, 1903.
14. More accurately, elements corresponding to the white and black colors.