MENDELISM AND SELECTION 369 



entire germ-cell at the time of fertilisation. Prof. W. L. Tower, 

 in a series of important experiments testing the influence of 

 altered environmental conditions on the breeding of potato- 

 beetles (Leptinotarsa), found that the conditions surrounding 

 and incident upon the germ-cell at the time of fertilisation may 

 be to a very considerable extent responsible for the determina- 

 tion of the dominant character in the cross and largely responsible 

 for the variability of such characters (1910, p. 332). 



Mendelian experiments give us a vivid impression of the pos- 

 sibilities of variation. Crossing two races of silk-moth, one 

 with striped caterpillars and yellow cocoons, the other with 

 unstriped caterpillars and white cocoons, yields in the hybrid 

 generation (F 1 ) only forms with striped caterpillars and yellow 

 cocoons, these being the two dominant characters. But the 

 inbreeding of the hybrids yields in the next (F 2 ) generation, 

 four different combinations which we may briefly allude to as — 

 Yellow Striped, Yellow unstriped, white Striped, and white 

 unstriped. But if there had been 10 unit-characters instead 

 of four, there would have been a theoretical possibility of 1,024 

 combinations. In short, Mendelism enables us to understand 

 the origin of that kind of variation which consists of permutations 

 and combinations of already existing qualities. 



Mendelism in Relation to Selection. — The facts of Mendelism 

 are in several ways important in relation to natural selection : — 

 (1) The facts warrant us in believing in the possibility of the 

 particular evolution of unit characters while the rest of the 

 organism remains stable. (2) When a variation is, through 

 inherent stability or through inbreeding, prepotent — i.e. when 

 its possessors breed true inter se — we can understand how it is 

 that even crossing with variants having an antagonistic character 

 need not imply any diminution of the dominance of the character 

 in question. The inbreeding of the hybrids simply results in 

 the sifting out of the pure parental types. (3) Suppose Mendelian 

 phenomena to occur in a series of generations, and suppose 



24 



