262 



SCIENCE 



[N. S. Vol. XXIX. No. 737 



ception of a mutation. All opponents and 

 many adherents of de Vries seem to be able 

 to conceive a mutation only as the addition or 

 loss of a complete character and therefore a 

 wide jump. That such a change sometimes 

 takes place, the mutationist believes, and so 

 also does the Darwinian (or at least so also 

 did Darwin), but by far the larger number 

 of mutations are quantitative changes in 

 characters already possessed, i. e., simply the 

 production of new modes as centers for linear 

 fluctuation. The difference between fluctua- 

 tions and mutations is merely in their trans- 

 mission. 



Part II. — Causes of Variations — begins 

 with an admirable non-technical description 

 of cell division, followed by a development of 

 the actual theme, in which are taken up first, 

 internal causes which affect the individual, 

 and second, those which aflect the race. One 

 can not help but be struck with the logical 

 manner in which the author disposes of old, 

 popular myths, such as telegony and intra- 

 uterine influences, by appealing to the law of 

 error. It is also gratifying to know that he 

 has found by correspondence that practical 

 breeders, for example dog fanciers, seem to 

 be or/sTowing their past childlike faith in 

 these things. It is likely that this book will 

 do much toward changing such superstitions 

 as belief in maternal impressions in the next 

 generation of animal breeders. Modern biom- 

 etry as based on the theory of probability 

 seems to the reviewer to have its greatest 

 value to practical breeders in showing how 

 illogical it is to consider isolated cases as 

 proofs of a biological principle, and consid- 

 erable emphasis is given by the author to this 

 point. It is hardly likely that many present- 

 day farmers will study thoroughly so bulky 

 and technical a volume, although the author 

 hopes to reach them, but their sons who have 

 had some training in genetics in our agri- 

 cultural colleges will be given a broader and 

 more scientific point of view. 



A large portion of the remainder of Part 

 U. is devoted to external causes of variation. 

 This portion is almost wholly an abstract of 

 C. B. Davenport's "Experimental Morphol- 



ogy." Then, as a transitory introduction to 

 Part TIT, on " Transmission," follows a very 

 judicial treatment of the inheritance of ac- 

 quired characters. Here, again, the theory of 

 probability is utilized to show that instances 

 of apparent transmission of somatic modifica- 

 tions are, at least, doubtful — a judgment 

 which the most ardent Lamarekian could 

 hardly question. 



A considerable percentage of the remainder 

 of the part is devoted to pure biometry which 

 the author believes to be the coming method 

 for attacking genetic problems. The ele- 

 ments of biometry are explained with greater 

 clearness than in any other English publica- 

 tion, and as such are to be highly recom- 

 mended. The elaboration of the results of 

 Karl Pearson and his school, however, shows 

 too great a trust in Pearson's lame biology. 



It will be exceedingly unfortunate if there 

 is a promulgation of this idea of certain 

 biometricians, namely, that valuable biological 

 conclusions can be drawn where there have 

 been no biological premises, by the mere fact of 

 complex, mathematical treatment. As an ex- 

 ample of a biometrical explanation which ex- 

 plains nothing the following may be cited 

 (p. 53Y) : 



The principal function of selection, therefore, is 

 to alter the type, not to reduce variability, and 

 the facts here cited show the inherent impossi- 

 bility of " fixing " the type in the sense that indi- 

 viduals will not depart from it. But, on the 

 other hand, the same principle assures us that, 

 however much we improve by shifting the type, 

 there always remains sufficient variability for 

 further selection, and so long as variability re- 

 mains there is hope and possibility for still fur- 

 ther improvement} 



Compare this view with that of Johannsen. 

 (By the way, the epoch-making work of 

 Johannsen on " pure lines " is not mentioned.) 

 The latter considers the variation's curve of 

 a quantitative character of any general popu- 

 lation to be made up of numerous family 

 lines. Members of these family lines are true 

 to their own type, their fluctuations being due 

 to nutrition, etc., and not inherited. The role 

 of selection, therefore, is to isolate a family 



"Last clause italicized by E. M. E. 



