372 ADAPTIVE VARIATIONS. 



of life may lead to very considerable changes in 

 the form and structure of all or most of the organisms 

 exposed, even in one generation. Hence, if Darwin's 

 definition be accepted as it stands, we are compelled 

 to admit that variations may be definite. Suppos- 

 ing, however, it be taken to imply the cumulative, 

 and so hereditary, action of conditions of life acting for 

 several generations, then those who refuse to admit the 

 validity of the instances of such cumulative action ad- 

 duced in the last chapter might also refuse to admit 

 the existence of definite variations. But assuming the 

 former interpretation as the correct one, are we to 

 agree with lienslow * that "in nature variations are 

 always definite," or are we to follow Darwin in believ- 

 ing that variations are, as a rule, indefinite, and only 

 exceptionally definite ? Here, it seems to me, we are in 

 want of more exact definitions. Probably Henslow 

 would admit that the variations in the number of car- 

 pels in the common daisy, or of veins in the leaf of the 

 beech tree, or of stigmatic bands on the seed capsules 

 of the poppy, are governed by the laws of chance, or if 

 he did not, how could he account for the fact that the 

 frequencies of distribution of the respective numbers 

 are in accordance with the Law of Error? Clearly, in 

 such cases the variations must be indefinite. Suppos- 

 ing, however, that the distribution of the variations in 

 the length of the leaves of a plant grown upon land 

 occur according to the laws of chance, whilst that of the 

 leaves of the same species of plant grown in water also 

 follows these laws, but supposing also, that the average 

 length of the aquatic plant leaves is considerably 

 *" Original of Plant Structures," p. ix. 



