42 THE VARIATION OF ANIMALS IN NATURE 



scarcity of oxygen, and it is a fair inference that the original 

 ' mutants ' were produced by the unhygienic conditions of the 

 culture practised in China, in which oxygen-starvation in 

 particular was inevitable. We need not detail the particular 

 action on the growing embryos of the oxygen-starvation, its 

 specific effect on particular structures (to which Tornier 

 devoted some very interesting study), nor Tornier's special 

 theory of ' plasma-weakness,' which he held to be ultimately 

 responsible for the malformations in question. What is not 

 apparent from Tornier's and Milewski's experiments is that 

 specific malformations produced under observation by a 

 verifiable environmental factor are regularly transmitted to 

 the offspring. It seems quite clear from some of Milewski's 

 experiments (MacBride, I.e. p. 8) that embryos of one of 

 the ' mutant ' types (the Ryukin) born in conditions of oxygen- 

 starvation, but reared in oxygenated water, give a high per- 

 centage of ' Ryukin ' types (80 per cent.). It is nevertheless 

 by no means apparent how far the experimental animals were 

 genetically pure. If the regular causation of the abnormal 

 condition by specific environmental factors is established, and 

 if the original abnormal breeds were indeed produced by this 

 cause, we might be disposed to admit that the inheritance of 

 the character in control conditions suggests that an induced 

 modification had acquired some degree of stability. But it is 

 very uncertain how far we can eliminate an original selection 

 in the production of the ' fancy ' types. The relative in- 

 stability of these forms under experimental conditions makes 

 it very difficult to judge the value of this work. 



We may sum up this survey of the experimental work by 

 concluding that there is a small amount of evidence that 

 induced modifications of certain types may be inherited. We 

 shall defer any further discussion until we have considered 

 the circumstantial evidence. 



(c) Circumstantial Evidence. — There is a large body of 

 observations and suggestions for consideration under this 

 head. There are two principal groups which are available 

 for examination. 



(A) The effects of use and disuse. — So much evidence (of a 

 sort) is available from human heredity that the effects of use 

 and disuse are not inherited that it might seem superfluous 

 to discuss this question. Nevertheless the matter cannot be 



