THE INADEQUACY OF EXPERIMENT 171 



is not alternative, which tends to be stable because it involves 

 the acquisition of a new unit 1 by the germ-plasm, but which is 

 necessarily stable only when the individual possessing it mates 

 with a similar individual. 2 



287. A little thought renders it evident that none of the 

 problems raised by the Mendelian and Mutation hypotheses can 

 be solved on the data furnished by experiment alone. Experi- 

 mental evidence lacks minuteness and precision. It is, besides, far 

 too fragmentary. Thus experiments which cover only a few 

 generations are absurdly inadequate to decide the questions 

 whether or not mutations are permanent and whether evolution 

 can or cannot be founded on the continued selection of fluctuations. 

 If, however, we turn to the more ample and conclusive evidence 

 afforded by the life-history of species, we see at once that of the 

 two deductions that evolution is founded on mutations and that 

 mutations are permanent, one must be wrong, for, with mere 

 cessation of selection, every character (e.g. eyes and limbs) tends 

 to retrogress till it disappears. The statement that mutations are 

 immutable is a very good example of the mistaken judgments 

 which tend to follow too exclusive a reliance on the meagre results 

 obtainable by a single method of inquiry. The observer is con- 

 stantly under the temptation to draw inferences larger than are 

 warranted by the scanty data at his command. It is interesting 

 to note that, since mutations may be small, what Darwin and his 

 followers call variations (changes " which have an internal origin 

 in the hereditary substance itself"), some members of the experi- 

 mental school term mutations ; and what the former call acquire- 

 ments (due to "varying conditions of food-supply, temperature, 

 density, moisture, light," etc.), the latter term variations. Except 

 that we are told that mutations do not blend or retrogress, nothing 

 more is done than to change the names and conduct the argument 

 on that basis. 



288. Again, the question as to whether the larger variations 

 (i.e. the mutations of de Vries) are more permanent than smaller 

 variations (i.e. fluctuations and the small mutations of Punnett), 

 and whether the inheritance of small mutations tends to be 

 alternative, cannot be decided by the mere experimental observer. 

 He has no special knowledge of the lesser variations, for the study 



1 " We may assume that these units are represented in the hereditary substance 

 of the cell-nucleus by definite bodies of too small a size to be seen, but constituting 

 together the chromosomes." Op. cit., p. 306. 



2 Op. cit., Lecture IX. 



