to the perpetual rearrangement or "neotaxis" of the germplasm.

As regards the crossing of old and stable species, it must be remembered that, whereas the determinants relating to the individual character may be different in every id of the germplasm, the determinants relating to the specific characters are identical in every id. Thus, when two individuals of such species cross, the homodynamic ids of the one meet the homodynamic ids of the other. It is obvious that, as all the ids are identical, the reduction of the plasm can produce no neotaxis, for every remingling will be a remingling of identical ids. The offspring of such first crosses must therefore be identical with one another as far as the specific determinants of the species are concerned; only secondary and individual differences can obtain among them.

In the crossing of individuals belonging to the same species, the specific determinants of the species play a secondary rôle. Differences in the individual characteristics do not exist merely between the two parental germ-plasms, but between the different ids of the germ-plasm; whereas in the case of species the characteristics of the two species are identical in all the ids. With regard to individual characters, the different combinations of ids, which are, as we have seen, the result of each successive reduction, can make up in every case a different sum of forces; so that either parent can produce germ-cells containing different organic arrangements, and, inter alia, either can produce in an active form arrangements which were latent in the parent, but which, owing to the chances of amphimixis and reduction, obtain a majority in the plasm of the offspring, and are therefore no longer latent.1 Or else the ids of one parent containing a specific

1 Latent characters are represented by a minority of determinants in the germ-plasm which are not strong enough to exert any influence on the development; but they form a reserve which, if the chances of amphimixis and reduction be favourable, may obtain a majority in the germplasm of the offspring, and thereby cause the latent character to manifest itself actively.

PREPONDERANCE OF A SINGLE PARENT

163

variation may be opposed by the homologous ids of the other parent containing another specific variation; and, in such a case, it may often happen that the biophors of the one parent are inferior in assimilating power to those of the other parent, so that the variation presented by the one parent will be entirely absent from the offspring.

The number of variations rendered possible by the neotaxis of the germ-plasm is, as we have seen, enormous. For instance, let us suppose the germ-plasm of the mother to consist of the homodynamic idant complex A1, and the germ-plasm of the father to be composed of a majority of the same idant complex A1, and likewise of a minority of ids containing variations A2 and a3. The mother's hereditary substance A1 will be counterbalanced by the hereditary substance A1 of the father, and the mingling of the two substances as a result of amphimixis can only result in A1 + A1-that is to say, in a summation of forces. But the paternal germ-plasm possess id complexes A2 and a3, which are not counterbalanced by similar id complexes in the maternal germ-plasm; and, as the paternal germ-plasm is composed of A1 + A2 + a3, whereas the maternal germ-plasm is limited to the variation A1, homodynamic with the paternal A1, the share of the father in determining the offspring will be preponderant. Of course, when we speak of the maternal germplasm as composed of a single variation A1, and the paternal germ-plasm as composed of the variations A1 + A2+a3, we are speaking of the germ-plasms as they are constituted after the reduction in their "amphimixiated" condition, so to speak; and we do not mean to imply that any combination is really as simple as this, which we give only to illustrate how the preponderance of one parent can be brought about.

The maternal and the paternal germ-cells contain only one half of the maternal or paternal hereditary substance, the other half having been lost through the reduction process. The more or less numerous eggs produced by the mother during the life

history do not contain the same id combinations, any more than do the innumerable sperm-cells produced by the male. The greater the number of chromosomes possessed by the species, the greater the number of possible combinations resulting from the perpetual rearrangement of the parental elements at each successive reduction division.1

1 A phenomenon which remains to be noticed is that of "return to the ancestral type," or atavism. This phenomenon is similar to that of the transmission of latent characters. Return to the ancestral type is almost always observable among young species. Among old and constant species it occurs very seldom, if at all. The explanation is that a young species may still possess a considerable minority of determinants of an older specific type in its germ-plasm; and the chances of reduction and amphimixis may bring about, even after numerous generations, a return to this type by the sudden "majorisation" of the minority of old determinants. In proportion as a species increases in age it also increases in stability, and returns to the ancestral type become less and less frequent. Such atavisms are aberrations which, being unsuited to the actual conditions of life of the species, do not propagate their kind. The experiments of breeders tend to show that six or eight generations, as a rule, are required to fix a new character definitely, and to prevent the accident of an atavistic return to the ancestral type. Atavism, it must be noted, may occur not only in the direct line of descent, but also in the collateral (vide Ribot, L'Hérédité psychologique, pp. 197 ff. Paris, 7th edition, 1905).

APPENDIX I

TERATOLOGICAL HEREDITY

THE peculiar mode of inheritance which is known as teratological may seem at first sight to contradict in certain respects the theory of the nontransmission of somatic modifications. We are not aware that the subject has been discussed, either in the works of Weismann or in those of his followers. It may be useful, therefore, to devote a little space to a consideration of the facts. Teratological inheritance involves the transmission of certain structural malformations, and may be treated under the heading of morphological pathology.

It would be a mistake, however, to suppose that these pathological malformations, which manifest themselves in the body of the organism, and are indubitably hereditary, are exclusively somatic.

It is safe to say that such malformations, which often present the appearance of an atavistic return to an ancestral type, are never to be found except on what the French call un terrain morbide. They are the somatic expressions of an organic disease which affects the germ-plasm as much as the soma.

Illustrations of the transmission of teratological malformations are numerous. We need only cite a few examples, as this is not a medical treatise. The symmetrical atrophy of the parietal bone of the cranium may be hereditary. The same may be said of anomalies in the development of the arteries, the pathogenic results of which are well known. The congenital narrowing of the pulmonary artery is a factor predisposing to pulmonary tuberculosis. Atresia of the aorta has a considerable influence on the pathology of the heart; chlorosis, that spontaneous anæmic condition which manifests itself at puberty, is sure to develop on ground which is well prepared by the abnormally small size of the heart. Aplasia of the arteries, insufficient development of the uterus and ovaries, and chlorosis, can be hereditary. Consumption is favoured by a number of malformations of the thorax ; and all these abnormalities have the result of diminishing the respiratory capacity of the patient, of exaggerating the already insufficient activity of the upper part of the lung; they therefore predispose the patient to microbic infection. It is generally admitted that patients suffering from thoracic malformations are more exposed than others to bronchial pneumonia. The inheritance of appendicitis has been

recorded several times, and Féré inclines to the belief that this is due to the transmission of an abnormal anatomical peculiarity. The inheritance of arthritism is, according to some authorities, referable to a predisposition to proliferation on the part of the connective tissue, combined with greater irritability, which renders the tissue less capable of resisting. Gigantism may also be hereditary. Hypospadism, abnormalities of the testicle, absence of the uterus and ovaries, can all be transmitted. The same applies to obesity; and hæmophilia, which is accompanied by anomalies in the development of the heart, of the capillaries, and of the arteries, is essentially hereditary. Finally, so well-known a malformation as hare-lip is often transmitted.

We need not cite further instances. The medical world is practically unanimous as to the hereditary transmission of teratological malformations, and it is probable that familiarity with this transmission, coupled with inadequate acquaintance with the theory of Weismann, has led the medical profession to stand aloof from his theory of evolution; whereas the majority of non-medical biologists have adopted it. It is, perhaps, well to see whether Weismann's theory can be reconciled with the fact of the transmission of teratological characters; for a theory can only survive if it is in harmony with the facts; and it cannot be denied that the facts of teratological heredity do at first sight present a difficulty.

Dr. Charles Féré, of Bicêtre, has collected a store of facts relating to teratological heredity in general, and to the relations between teratological malformations and neuropathic diseases in particular.1 Accepting the facts, which, he states, are based on a vast number of authorities, we would submit the following considerations.

It might be maintained that the general pathological condition of the germ-plasm of the parents, a condition which is determined by various causes, predisposes the offspring to teratological malformations. In this case the latter are not themselves inherited; what is inherited is the general pathological condition which serves as their constitutional basis. The teratological phenomena themselves are not hereditary, but congenital -that is to say, they are produced by some accident which occurred after the act of procreation, during the intra-uterine period. It is well to note the difference between the terms "hereditary" and "congenital." Those characteristics alone are hereditary in the scientific sense of the word which are directly transmitted from the germ-plasm of the parents to the germplasm of the offspring; congenital characteristics are those which are acquired by the offspring during embryonic development, and which

1 Ch. Féré, La Famille névropathique: Théorie tératologique de l'hérédité et de la prédisposition morbides, et de la dégénérescence, pp. 152-191 (Paris, Alcan, 1898). Vide also Isidore Geoffroy-Saint-Hilaire, Histoire générale et particulière des Anomalies de l'Organisation. 3 vols. Paris, 1833–37.