case of the mutilations and the experiments of Brown-Séquard). In truth the decisive experiment, verified and certain, is still lacking. As for the nonexperimental proofs, they are always subject to discussion and can never attain beyond second rank.

As it is necessary, however, to hold some opinion on a question so important that it dominates all the conceptions of evolution, we must decide for ourselves according to the relative weight which we attribute to the various arguments, and naturally each one of us holds the more strongly to his own view, the more this view allows for a greater degree of personal appreciation. Some, partisans of heredity, receive with pleasure experiments or observations, even mediocre, which seem to them to constitute proofs; others who reject it, hunt for weak points in the demonstrations, and usually find them. It is almost a matter of faith, of nationality. The great majority of French zoologists favor the affirmative, following Giard, Edmond Perrier, Le Dantec, F. Houssay, Delage, all more or less Lamarckists; the Americans, except some paleontologists, are nearly all for the negative. I am perfectly ready to admit that up to the present I have put myself on the negative side.

It is obvious that there are categories of acquired characters arising from different causes: (1) Mutilations; (2) the effects of parasitic diseases producing a general intoxication; (3) the action of the great natural factors, light, temperature, humidity, salinity, nourishment; (4) the effects of use or those of disuse; (5) the psychic acquisitions of training, of instruction. In my opinion, the negative demonstration, that of the nonheredity of acquired characters of one of these categories, is of value for that one only, and can not legitimately be extended to the others, for if there are certain acquisitions of the body which are not transmitted to the sexual cells, the bearers of the hereditary patrimony, it does not necessarily follow that the same is true for all. But, on the other hand, if there were an experiment which showed indisputably the heredity of a truly acquired character, it would be a strong probability for some other categories, for although we do not understand completely how an acquired bodily modification can add itself even in weakened form to the hereditary patrimony, if the fact were proved one single time, the argument of incomprehension would lose all of its force.

For the first category of acquired characters, it can be said that the answer is definite; since the critiques and the experiments of Weismann, many times repeated, no one believes any longer in the heredity of mutilations. Every-day observations confirm those of biologists and it is certain that the pseudo-examples of the transmission of mutilations that are quite often reported among domestic

animals and among human families are simple coincidences, which have no more interest and often no more authenticity than cases of supposed maternal impressions. The experiments of Brown-Sequard, badly done moreover, which concern the heredity of mutilations and that of physiological disorders following nervous mutilations, have been completely disproved by researches which inspire confidence, and there is nothing left of them.

Various authors have affirmed the heredity of acquired characters of the fifth category, as little likely as this seems. Here are some examples which Hachet-Souplet reports as demonstrative. A macaque monkey which he had taught, not without difficulty, to kill rats, gave birth to young who hunted rats marvelously; cats trained to respect mice had young which did not take mice, even when the distribution of their food was intentionally retarded; sparrows trained to draw a chain from a little well for six generations gave birth to young which, without training, were able to draw the same. chain. A dog had been trained to make rapid pirouettes to the left; a daughter of this dog, raised in the country, having no example before her eyes and having received no training, began by herself to make pirouettes to the left toward five or six months of age. All that is very astonishing. Although one can not criticize experiments which he has not followed, I am persuaded that there is a "hole" in these observations, due perhaps to the deceit of assistants, to a surreptitious training continued unobserved, etc., and I do not doubt that the heredity of acquisitions of training will go to join that of mutilations.

The third category concerns the factors of the medium; no one doubts their determining influence on the characters of animals and plants, and it is certain that when they have been made to vary experimentally up to the extreme limits compatible with life there often result notable modifications among beings which are submitted from youth to a change of environment. But the question is to know whether these modifications pass, even in a very attenuated degree, to the succeeding generation reared in the normal medium. If they do, we have the key to the formation of geographical races and of many adaptations; for it is a fact that the results of the action of the medium would necessarily be cumulative, and after a sufficient number of generations passed in the modifying medium, the species might be very notably transformed, perhaps even irreversibly; if they do not, the effects of the medium would be produced anew for each individual, without cumulation, and the influence of the environment on the body would no longer have any interest from the point

Hachet-Souplet, La genèse des instincts, étude expérimentale, Bibliothèque de Philos. scient., Flammarion, Paris, 1912. P. 239.

of view of evolution. Many experiments have been made with the primary factors, as Giard calls them. Hundreds of experiments, bearing especially on insects, have given negative results as far as the heredity of acquired characters was concerned. A few only (Standfuss, E. Fischer, Schröder) have had a feebly positive result. But these are not sufficiently convincing, the authors having worked with species naturally so variable that it has been necessary to ask ourselves whether the characters which they believed to be acquired did not exist in a latent state in certain individuals before the modification of the medium. The experiments of Kammerer are still more puzzling. From 1904 to 1911 he published a great number of researches demonstrating the considerable influence of the environment on the methods of reproduction of amphibians (Alytes obstetricans, Salamandra maculosa and atra), on the color of amphibians (Salamandra maculosa) and of reptiles (Lacerta), and bringing into evidence in the majority of cases the transmission of acquired characters. It is not too much to say that at first glance the results of Kammerer appear incredible. He certainly had at his disposal an exceptional installation at the Prater in Vienna (Biologische Versuchsanstalt der Akademie der Wissenschaften), but that he should have been able to accomplish even with these means breedings of such great difficulty and of such long duration is indeed surprising. From the very beginning his experiments seemed too successful, too demonstrative, and too extraordinary to merit confidence. Boulenger and Bateson have criticized them severely, and there have even been mentioned trickeries and substitutions of preparations. Hans Przibram, who also worked at the Prater, has been kind enough to tell me that nothing of this sort took place, but the suspicion is at least an indication that Kammerer's results have met with a general incredulity. Argument can not then be based on them until they have been confirmed by observers in other countries, something which has not yet occurred-in fact, quite the contrary.

I limit myself to this preamble, not intending to criticize one by one the various facts presented in botany or in zoology as proofs of the heredity of acquired characters. There are many of them, but they have not withstood criticism since none of them carry conviction."

Quite recently the question has taken on a new aspect with the very remarkable work of F. Guyer and E. A. Smith." It is known that

Bateson, Problems of Genetics, 2d edition, New Haven, 1916, pp. 199-211. Boulenger, Experiments on color-changes of the spotted salamander, etc. Proc. Zool. Soc. London, 1921, p. 99.

See Weismann, Essais sur l'Hérédité, 1892, and Vorträge über Descendenztheorie, Jena, 1902; Bateson, Problems of Genetics, 1916; Cuénot, La genèse des espèces animales, 2d édit., Alcan, Paris, 1921.

Guyer and Smith, Studies on cytolysins. Journ. Exp. Zool., vol. XXXI, 1920, p. 171. problems, Amer. Natur., vol. LV, 1921, p. 97.

II. Transmission of induced eye-defects, Guyer, Immune sera and certain biological

The germ would be a microcosm containing all the different colloids of the adult organism; if the experiment of Guyer and Smith is valid, I do not see how we can escape that conclusion. It contains, moreover, nothing contradictory to the Mendelian conception, for the chemical determinant or the representative colloid can very well be homologous with the Mendelian factor. In fact, Guyer and Smith show that the germinal element affected by the crystallolysin acts very nearly like a recessive Mendelian factor. When males or females with abnormal eyes are crossed with healthy individuals, not treated, coming from other regions, the progeny always appear with normal eyes (dominance); but the young are heterozygotes and contain in a dominated state the defective eye character, for if these individuals of normal appearance are crossed with rabbits with defective eyes, there appears this time among their progeny a more or less large number of young with abnormal eyes. Two individuals with defective eyes (dominated homozygotes), bred together, should produce according to the hypothesis of a simple and typical Mendelian factor, only young presenting the anomaly. If it is not always so, there is at least a striking majority with opaque or reduced lenses among their progeny. The evidence is still insufficient to see clearly into the genetics of the character, but in the main, leaving out of the question details which will probably be cleared up later, the characters normal-eye and lysis-eye form an allelomorphic pair with dominance of the former.

Can it be said that the experiment of Guyer and Smith is entirely satisfactory? Far from it; an evil fate decrees that the proofs of the heredity of acquired characters never present that completeness of evidence which irresistibly entails conviction and leaves the mind. at rest. Crystallolysin has a truly disconcerting action; it does not act on the lenses of the pregnant mother, a first anomaly, which, it is true, has received a more or less good explanation; it acts capriciously on the crystallines in course of development of the fetuses (9 times in 61), and according to the hypothesis, on the crystallinian determinants of the germinal cells of the same fetuses. We should then expect that it would also affect the germinal cells in the ovaries of the mother, as well as those of the fetuses. But this is not the case. The doe rabbits which survive the injections of anticrystalline serum and which have produced young with abnormal eyes are served several times by the same normal males after the serum treatments have been stopped (for how long a time is not known) and not a trace of ocular malformation is visible in their numerous progeny. There is in that fact something entirely incomprehensible, unless we admit that the female germinal cells can be modified only at a particular stage, that of the period of multiplication of the ovogonia,

or else of synapsis. Now it is known that among the mammals these phenomena take place in the embryo, and that the adult has only ovocytes which grow slowly and unequally and come successively to maturity. We have a means of verifying this subsidiary hypothesis, for adult males present all phases of spermatogenesis, from spermatogonia to spermatozoid. By injecting anticrystalline serum in males, their sexual cells should be affected at the sensitive stage, and, mated later with normal females, they should transmit the anomaly, which would be visible only among their grandchildren. It is a crucial experiment; I do not see that Guyer and Smith have tried it.

It would be better, in order to simplify the experiment, to use not merely the hen as the source of crystallolysin, but rather the rabbit itself, in this way not introducing the foreign factor, the hen serum. However, Guyer has recently succeeded in obtaining a young rabbit with two defective eyes by injecting several times in the normal mother crushed crystalline lens of rabbit before she was with young and during her gestation. A rabbit can, then, form a crystallolysin as efficacious for the rabbit as that produced by a foreign species.

Thus we have been led to admit, in order to interpret the results of Guyer and Smith, that the crystalline lens has some representation (it matters little how we understand the word and the thing) in the germinal cells, since the specific antibody acts upon them specifically. Now, the crystalline is indeed the last organ which we should expect to see represented in the sexual elements. Beautiful researches conducted, it is true, on fish and batrachians tend to lead us to consider the crystalline lens as the epigenetic organ par excellence, as a reaction of the epidermis in contact with or in the vicinity of the optic cup; when the latter comes in contact with the epidermis at any point whatever (abdomen, mouth, dorsal face, etc.), it develops a crystallinian thickening. It is not necessary, then, to believe that the crystalline is "represented " directly in the hereditary patrimony, since it is enough that the optic cup be represented in order that there may be formation of crystalline. This is, however, only a difficulty.

Let us come back to acquired characters. The experiments of Guyer and Smith, while admitting that they should be confirmed in fact and interpretation, will permit us to accept henceforth the heredity of characters acquired under the influence of the great general factors of the medium. At the same time that these produce their more or less adaptive effect on certain tissues of the

See Werber, Critical notes on the present state of the lens problem, Biol. Bull., vol. XXXIV, 1918, p. 219. Fessler, Zur Entwicklungsmechanik des Auges, Archiv. für Entw. der Org., vol. XLVI, 1920, p. 169.