EESTJLTS OF CROSSING TWO HEMIPTEROUS SPECIES. 467 



length, and no Mendelian segregation in the F2 generation. These facts, 

 together with further- studies on size-relations, followed by his striking 

 results in selection (Castle, '12 & '14), have led him to question certain 

 hypotheses which he believes are quite unnecessary adjuncts to Mendelisra. 



As Castle believes that both small and large variations are inherited, he 

 questions the mutation theory which claims " that only variations of some 

 size are inherited '\ 



As he believes in quantitative variations of unit characters, he questions 

 the multiple factor hypothesis. 



As he has demonstrated that quantitative variations can be increased by 

 selection, he questions the hypothesis which denies this possibility (genotype 

 theory). 



Of unit characters he says : " In my experience every unit character is 

 subject to quantitative variation, that is, its expression in the body varies, 

 and it is clear that these variations have a germinal basis because they are 

 inherited.^' 



" It is the substantial integrity of a quantitative variation from cell- 

 generation to cell-generation that constitutes the basis of Mendelism. All 

 else is imaginary."" (Castle, ^12 a.) 



Castle has arrived at his present conception of Mendelism through his 

 extended and thoroughly scientific experiments on quantitative variations ; 

 and the results of his experimental studies of size-differences appear to be so 

 completely in harmony with the results of our cross-breeding experiments 

 with Suschistus, that we shall give a brief summary of his observations before 

 we present the facts which seem to us to support them. 



His wide experience in experimental breeding in relation to size-characters 

 entitles him to speak with authority on this subject. 



His observations may be briefly stated as follows : — 



First. Dominance is absent in the Fi hybrid generation. 



Second. The Fi hybrids are intermediate in size. 



Third. Mendelian segregation is absent in the F2 hybrid generation. 



Fourth. There is increased variability in the Fo generation as compared 



with the Fi generation. 

 Fifth. The F3 generation, like the Fi, is intermediate in size. 

 Sixth. Both extremes in size of the original parents may be found in 



the F2 generation, but not in the Fi generation. 

 Seventh. Size-characters, in common with all characters (even those that 



Mendelize), show quantitative variation. 



We shall compare these observations with our own results, and when 

 making this comparison we shall consider first the intromittent organ, and 

 then the genital spot, in order to point out that the results from these two 

 exclusively male characters are identical. 



