MUTATION. 155 



mutation as synon37mous with biotype. They call "mutation" 

 each individual with a new genotj^e, no matter how produced, 

 for instance they write about the selection of the best mutants 

 in the descendants of a cross. It is very probable that this 

 confusion is primarily due to the fact that de Vries, in re- 

 viewing the plant-breeding work of Mr. Burbank, assumed that 

 very many of Burbank's novelties had originated by mutation, 

 notwithstanding the fact that on Burbank's evidence they had 

 originated as descendants of hybrids. 



The fact,' that a family of animals suddenly produces a. new 

 recessive form, although it has never before during a long 

 series of controlled generations produced thils novelty, may 

 not be spoken of as mutation without further proof. We know 

 that during all the time that this recessive form has not been 

 produced, a number of the animals may well have been hetero- 

 zygous for the distinguishing gene, and the chance mating of 

 two heterozygotes may be responsible for the birth of the 

 n&fi recessive form. To be sure of the occurence of a mutation 

 in animals, it is necessary to show that an individual which is 

 proved to be homozygous for a certain gene, has nevertheless 

 produced at least one germ-cell which did not contain it. We 

 have done this durselves in mice. 



In one case we were deahiig with a family of animals which 

 all possessed the gene which we called G, and, which was known 

 to produce the difference between black and agouti. The ani- 

 mals were strictly inbred. In every generation brothers were 

 mated to sisters. This circumstance made it subsequently 

 possible to make probable, that a real spontaneous loss of 

 a gene was the cause of the production of the animals without 

 G. 



We found, that the two parents of the mutants were both 

 heterozygous for G, and that of the two grand-parents, one 

 was pure, GG, and the other impure for this factor, Gg. The 

 great-grand-parents, when tested, proved to be both homozy- 

 gous for G, GG. The male gave 34 young, all having G, when 

 mated to gg females, and the female gave 27 young with G, 



