40 inheritance and Evolution in Orthoptera IT 



of the iiiili\ idiuils iiiated troiu thein was so great that tlie eorrectness 

 of the separations was not adecjuately j)roven by further breeding. 



Note. The mortality on the whole due to fungus and other disease, lack of proper 

 care, etc., was very great. However, recently, considerable improvement has been made. 

 Of the 974 matings made in this experiment only 470 produced .young that could be 

 recorded. Of the 43,914 young transferred from the mating jars only 21,686 (those used 

 in this paper) were recorded. 



The nature of this material permits us to observe the difference in 

 behaviour between the multiply alleloniorphic characters, ullelumorphic 

 to each other, never to absences, and the character @, allelomorphic only 

 to its absence, never to anything. Apjilication of this conception may be 

 made outside the Paratettix material ; the case of inhei-itance of combs 

 in fowls, for instance. Allowing P to represent the factor for pea, S the 

 fiictor for single, each allelomorphic to the other, and <I> the factor, 

 allelomorphic only to its absence, which, when present, modifies SS 

 to make the two kinds of rose, and PP and PS to make the four 

 kinds of walnut, we have the same situation that is produced by B, G, 

 or any other two multiple allelomorphs and in Paratettix. We then 

 secure the same ratios, and exact parallelism in all respects. When 

 Mr Bateson (2), p. 63, describes a cross, rose x single, with the resulting 

 dominance of rose in F^ and the usual 3 rose ; 1 single in F.,, he really 

 shows a cross of single homozygous for the modifying factor <t> and jjure 

 single {S<^8<t> x 8S) which produces SS<i>. Then when these are inbred 

 there results in F.,, 



SS SS'* SiS^ ,S*S<I> X SS Parents 



12 1 I ^"i 

 . , .S'«* 



13 I 



"* SS* X SSi> 



\ 



I I I 



SS S.S'* S^S* F-i 



12 1 



1 3 



