which would so well explain all the facts of inheritance as to permit us to foretell them, it would have some weight as an argument in the discussion, because evolution and heredity are undoubtedly two portions of the same process. However, that much. cannot be said of the series of hypotheses worked out by Weismann to meet the objections of his critics; but as they enjoy a certain favour with a number of biologists, we are bound to examine them, and to see what bearing they may have upon our opinions about the inheritance of acquired characters.

All we can expect now from a hypothesis of heredity is that it should give us some plausible physiological explanation of two sets of facts: the reproduction in the offspring of the characters belonging to the species-that is, the maintenance of the ancestral type-and the appearance of new features, without which there would have been no evolution. All the hypotheses which were brought out during the last fifty years had in view this double purpose; but there are only two of them-Darwin's Pangenesis and Weismann's Germ-plasm-among which the suffrages of biologists are divided by this time..

Everyone knows more or less the Pangenesis hypothesis. Following Spencer's idea of 'physiological units,' Darwin suggested that all the cells of the body of a plant or an animal throw off during their life extremely minute living particles which are capable, like cells or spores, of multiplying by subdivision, and thus reproducing their mother-cells. Darwin described them as gemmules. Being extremely small, the gemmules wander through the body, passing through the membranes of the cells, and finally they collect in the reproductive cells of the individual, which thus contain representatives of all the cells and all the groups of cells of the whole organism. Every organ, every tissue, every bone, muscle, nerve, blood-vessel, and so on, has its representative gemmules in the reproductive cells. And when the time comes for these cells to reproduce a new being, they transmit to it the gemmules capable of reproducing all the features of the race, as also, to some extent, the modifications which the parents may have gone through during their own lifetime.

The germ cells of-let us say-a cart-horse would thus contain gemmules capable of reproducing all the typical organs and features of a cart-horse of a given race; and they would also bear traces of the changes which the individual horse had undergone during its life in consequence of a good or a bad food, overexertion, and so on.

The inheritance of both the racial and the individually ' acquired' characters is thus rendered more or less comprehensible. But with our present ignorance of the inner life of organic tissues, the difficulty is to conceive how the gemmules are carried

к к 2

from the spot where they originate to the reproductive cells. How can they reach them in due proportions, and why, later on, when the germ begins to develop into an embryo, do they enter into action in due succession-each group at the proper moment? These difficulties certainly are great; and yet, with all that, the Pangenesis hypothesis so well explains various aspects of heredity that Professor Delage, the author of an elaborate work on the subject, is quite right in saying that all the subsequent hypotheses which retained the idea of representative units' added nothing substantial to the explanation proposed by Darwin." Still, the fact is, that the Pangenesis hypothesis, including Brooks's attempt to improve it, has not yet met with much support, and, for some time at least, the hypothesis of Weismann-before he was compelled to introduce into it the hypothesis of germinal selectionrallied the majority of suffrages."

III

Weismann started from the idea that the inheritance of characters acquired by individuals under the direct action of surroundings is not needed for explaining evolution. Variation is not something coming from without it comes from within-from the organisms themselves, and it is regulated by natural selection, which, given the spontaneous variations of the germ-plasm, is sufficient to explain all the adaptations of the organisms to the conditions of their existence. However, this hypothesis so much runs against all the tendencies of modern empiric science, that I referred to the earlier writings of the Freiburg professor to find in them some indications as to its origin. If I am not wrong, an essay written by him in 1876, with the idea of reconciling Darwin's teachings with a teleological conception of evolution, contains such an indication."

Karl von Baer, in his criticism of Darwin's hypothesis of natural selection, had made the remark that the followers of this

13 Yves Delage and M. Goldsmith, Les théories de l'hérédité (Paris, 1909), pp. 113-115. The main work of Delage is L'hérédité et les grands problèmes de la biologie générale (2nd edition: Paris, 1903).

14 Besides Spencer and Darwin, the same idea of self-multiplying representative particles was developed also by Francis Galton, the botaniste Nägeli and De Vries, the anatomist Kölliker, and Oscar Hertwig. But I am compelled to pass over these extremely interesting hypotheses. The general reader will find them analysed in Prof. J. Arthur Thomson's Heredity (London, 1908), and Delage and Goldsmith's Théories de l'évolution (Paris, 1909). For excellent critical analyses of Weismannism, see S. R. Romanes's An Examination of Weismannism (London, 1893), H. W. Conn's The Method of Evolution (New York, 1900), and Dr. Plate's Selectionsprinzip und Probleme der Artbildung (Leipzig, 1908).

15 A. Weismann, 'Ueber die letzten Ursachen der Transmutationen,' second essay in Studien zur Descendenztheorie (Leipzig, 1876), chapter Mechanismus und Teleologie,' pp. 314 seq.

hypothesis being brought more and more to deny 'all purpose in evolution-scientific recognition could be granted to the hypothesis only if it recognised a universal tendency towards reaching a certain purpose.16 To which Weismann, after having quoted these words with approval, as also those of Hartmann, who considered Darwinian evolution as a 'mad chaos of stupid and capricious forces,' added that it was necessary, indeed, 'to combine in a theory of evolution the teleological principle with a mechanical conception.' Formerly, he thought it impossible. Now, he was going to prove that it was unavoidable.

Baer (he wrote) is right, for the phenomena of organic and inorganic nature cannot possibly be imagined as a work of accident. They can be conceived only as a process directed in accordance with a certain great plan' (p. 315). And he came to this conclusion: 'We must not hesitate to recognise the existence of a force acting with a purpose (einer zweckthätigen Kraft); only we must not represent it as directly interfering in the mechanism of the universe: we must conceive it rather behind the mechanism, as a final cause.'

17

Causality and purpose-he wrote further on-by no means exclude one another: we see them combined by a watchmaker in a watch; and they are likewise combined in the universe by the Mechanic of the universe.' 'The apparent contradiction between teleology and mechanism can be conciliated,' and he concluded with these significative words which contain the key to his subsequent conceptions of heredity: Why should we not return to the idea of "matter endowed with a soul?'

18

This matter endowed with an immortal soul-this conciliation of determinism with teleology-found, I am inclined to think, its expression in Weismann's 'immortal' germ-plasm. This was evidently the 'teleological mechanism teleological mechanism' through which the Baconian and Darwinian mad chaos of capricious forces' was avoided in the universe.

To avoid Hartmann's reproach Weismann eliminated from his hypothesis of heredity the very possibility of Buffon's and Lamarck's factor of evolution-the direct action of the monde ambiant, and the effects of use and disuse. To deny these effects upon the individuals was impossible, they are evident; but their hereditary transmission, he maintained, was inconceivable, impossible. Upon this postulate, borrowed from his conception of evolution, he built up his hypothesis of heredity.

16 'Soll der Darwin'schen Hypothese wissenschaftliche Berechtigung zuerkannt werden, so wird sie sich dieser allgemeinen Zielstrebigkeit fügen lassen. Kann sie das nicht, so wird man ihr die Geltung zu versagen haben' (p. 315).

17 P. 324.

19 P. 327.

The germ-cells which serve to reproduce the individual are not derived, he maintains, like Darwin's gemmules, from the individual's body-cells and tissues. They represent a speck of 'immortal' matter, transmitted to the now living beings-unchanged in the main features of its chemical and molecular structure-from their remotest unicellular ancestors. Each individual receives it from its parents, and keeps it quite apart from its own organs, tissues, and body-cells (the somatic' cells), so as to transmit it, unaltered, to its offspring. These were, at least, Weismann's ideas when he brought out his hypothesis in 18831885.19

As to the structure of the germ-plasm, Weismann's hypothesis also differs from Darwin's Pangenesis in an essential point. The plasm of the germs is not composed of gemmules capable of reproducing by subdivision the cells from which they originate. They only contain Anlagen, i.e. predispositions,' 'tendencies' of a molecular or chemical nature) to produce such and such cells, tissues, and organs. When the germ begins to develop into an embryo, its physiological units only determine the production of cells, tissues, and organs of a definite character; this is why Weismann gave them the name of determinants. There is a group of determinants for each new line of growth in the embryo originated by each new subdivision of its cells for the outer skin and the inner membranes-for each portion of the body, including such tiny local peculiarities as, for instance, a tuft of grey hair inherited in a family. Besides, there is a strong hierarchy among Weismann's determinants. There are determinants of each arm which govern the development of the arm in the embryo, and there are subordinate determinants for marshalling the development of each finger, each nail, each muscle, etc., of the arm and the hand. The development of the embryo thus

19 The whole foundation' of the Pangenesis hypothesis must be abandoned : 'it is impossible for the germ-cell to be, as it were, an extract of the whole body.' (Essays on Heredity, 2nd edition, 1891, i. 169.) Heredity is due to the transference from one generation to the next of part of its germ-plasm, containing the very same formative predispositions as were contained in the germplasm of the parents. The germ-cells may be contrasted with the rest of the body; as their development shows, a marked antithesis exists between the substance of the undying reproductive cells and that of the perishable body cells.' (Essays, i. 73-74.) There are no conceivable means' by which modifications produced in the body-cells by external agencies, or by use and disuse, could be conveyed to the cells of the germ-plasm (i. 172). Weismann did not assert that the germ-plasm is totally uninfluenced by forces residing in the organism.

'The nutrition and growth of the individual must exercise some influence upon its germ-cells'; but he was disposed to think that the influence of nutrition upon the germ-cells must be very slight, and that it may possibly leave the molecular structure of the germ-plasm absolutely untouched.' 'In fact,' he added, 'up to the present time, it has never been proved that any changes in general nutrition can modify the molecular structure of the germ-plasm.' (Essays, i. 172-173.)

reminds one of the mobilisation of an army, of which the determinants are the officers and sub-officers organising its different parts.

IV

Most of the processes of heredity lie unfortunately in a domain which still remains invisible for our best microscopes. We are reduced, therefore, to build up purely hypothetical, more or less probable, mental representations of these processes, and in the impossibility of verifying which of our mental images is nearest to reality, it would not have been worth while discussing them for years. However, even such representations may have a scientific value. It matters little whether we represent to ourselves heredity as a transmission of self-multiplying 'gemmules,' or of 'predispositions' and 'determinants.' But if it were proved, or only rendered very probable, that the germ-cells cannot receive from the body-cells impressions which affect them in the same sense as the body-cells are affected by environment, then an important argument would have been furnished against the hereditary transmission of acquired characters. This is the raison d'être of the interest taken by our best biologists in the hypothesis of heredity advocated by Weismann. The question at the bottom of this interest is the desire to know: Was modern physiology right when it maintained (to use Cuvier's words) that all the organs of an animal represent a unified system, of which all parts act and react upon each other, so that no modification can take place in one of them without producing analogous modifications in all others?' Or, as Houssay puts it in La forme et la vie: Were Darwin and Lamarck right when they conceived that an individual's inheritance is a resultant from all the external forces that had acted upon its ancestors-of all the variations, ancient and modern, they underwent? In a subsequent essay we shall see what data the direct experiments made during the last five-andtwenty years have given us to solve the problem. Now let us see first how the question stands from the theoretical point of view : what our present knowledge of the processes of heredity-so far as it goes-has to say about this question? Let us, then, cast a glance upon the main points won during the discussions upon the complicated framework of Weismann's hypotheses.

To begin with, it was proved that it was wrong to oppose the body-cells to the germ-cells, or reproductive cells. 20 Germ-plasm is not limited to the germ-cells only. On the contrary, it is contained, in a more or less advanced state of specialisation, in the nuclei of all the cells. Consequently, the idea of a sort of

20 Weismann had to recognise it in 1885. (Essays, i. 209-211.)