102 CROSSING. 



And this is especially true of the new dominant characters. 

 New dominant characters, resulting trom a cross new in the 

 sense of not being observed before, and not being expected as 

 a simple combination of two known characters, should be very 

 common in such diversified material as these flies. We find 

 no data in the litera ure on cases of this kind. And the absence 

 of such data, observed in connection with the frequency with 

 which dominant "mutations" are described, makes it more 

 than probable that most, if not all these "mutations" are due 

 to novel combinations of genes already studied, and included 

 in the material. The method of firmly associating particular 

 genes with particular characters must almost inevitably bring 

 about this result. If an author has studied the mutual inde- 

 pendance of genes A, B and C, he may not feel sure that a 

 novel dominant character suddenly cropping up is due to 

 a new gene D, before he knows that this new character is not 

 a very common result in individuals carrying A and B, or A 

 and C, but if one works with genes named after definite char- 

 acters, he may think it absurd even to conceive. The possibil- 

 ity, that a novel character fitly called "pushpin," may result 

 from a genotj^e including genes "duckfoot", "gold" and 

 "sticky" but lacking "chestnut," and instead of making all 

 the possible combinations of all his genes studied so far, he will 

 contend himself by christening a new gene "pushpin" after the 

 novel character of the fly found in the "duckfoot" stockbottle. 



So long as the possibility exists of naming one ^ind the same 

 gene six times after characters it helps to produce in different 

 combinations of other genes, so long is it unneccessary to re- 

 gard the production of striking dominant novelties in this 

 material as evidence for progressive mutation in the sense of 

 de Vries. 



It must be understood of course that the tendency among 

 some of the American authors is to use the term "mutation" 

 as synonymous to "novelty", in the sense (rf Burbank, rather 

 than in that of de Vries. In those cases however, it should 

 be made perfectly clear, that the term is used in this way. 



