510 A NEW THEORY OF THE ORIGIN OF SPECIES. 



which degrades the others in the concurrence, either direct or indirect, 

 which exists between animals and plants, in that sort of strnggle for 

 existence whose importance was perceived already by A. de Candolle 

 and Lyell, and which results in the disappearance of the vanquished 

 species and the effective triumph of the new one. 



It may be noted that natural selection is not a single hypothesis; it 

 is a linking together of three hypotheses. If we separate the links of 

 this chain we can show" that not one of them w^ill stand test. The first 

 hypothesis is that of the advantage in the struggle for existence which 

 is given to an animal by the possession of a small, adaptive variation; 

 the second is that of a preservation, by transmission, of this acquired 

 character; the third is the progress, always in the same direction, of 

 these profitable variations, which, accumulating, finally create a spe- 

 cific character. None of these hypotheses will support a searching- 

 examination. 



In the first place, as to the benefit of a small, adaptive variation, it 

 may be observed that it would be, in itself, too insignificant to give 

 rise to selection. Let us take for example the transformation of an 

 ungulate quadruped into a giraffe according to the Darwinian theory. 

 In this system an increase of some centimeters in the length of the 

 neck would be a favorable adaptive variation; it would allow the 

 animal, in case of famine, to browse upon the verdure of trees some 

 inches higher than his companions could. But with Mivart, Naegeli, 

 Delage, Osborn, Emery, Cuenot, and others, we may affirm that in 

 case of actual famine this advantage would amount to nothing and 

 would not assure the survival of its possessor. The individuals who 

 would die would be the youngest or the oldest, or, in a general way, 

 the feeblest. The variation must be considerable in amount in order 

 to constitute a real advantage and in order that tiie process of selec- 

 tion may be applied to it. 



The second hypothesis is, then, to imagine that this variation, 

 admitted, for the moment, as useful, may be preserved and trans- 

 mitted by generation. We have stated above what naturalists think 

 at the present time concerning the transmission of acquired characters. 

 The least that one can say is that it is \evy much controverted. 



The third h3q5othesis, grafted upon the first two, is the repetition of 

 the variation. Even if we disregard the objections made to the pre- 

 vious h3'potheses there are still others which present themselves here. 

 It is, indeed, necessary that the variation should continue to be pro- 

 duced in the same direction during a great number of generations in 

 order that it may be recognizable, since it is minute each time it occurs; 

 many additional elongations would be needed in order to produce the 

 neck of a giraffe from that of an ungulate. Lamarck, by placing the 

 cause of variation in external conditions, makes this continual addition 

 of effect plausible. The permanence, or better, the repetition of the 



