Analysis of Heterogeneity in Complex Chakacters 347 



obtained from a pair of biotype 4, also from the Chapultepec colony, as wild 

 stock. In Fi this pair gave biotype 3 (A), biotype 4 (5), and biotype 5 (C). 

 In F2 two pairs of biotype 3 (A') were found that gave 100 per cent of this same 

 type and continued to do so through F4, F5, Fg, F^, Fg, and Fg, while 5 pairs 

 A" gave biotypes 3 and 4. Four pairs of biotype 3 {A"a) were found that came 

 true in F3 and subsequent generations when inbred, while the biotype 4 of F2, 

 only 1 pair gave progeny, and these were of biotype 4 and biotype 2. In F3 the 

 biotype 4 showed 3 pairs out of 11 breeding true in F4 and F^, but in F5 one pair 

 again gave biotype 3 in small numbers. In Fg 1 pair out of 12 of biotype 2 gave 

 all biotype 2. 



From the mating B in F,, 1 pair, all that gave progeny, continued to breed 

 true through F3, F'^, F-, and Fg. Mating C of biotype 5 was heterozygous, 5 

 pairs giving in F3 two classes : biotype 5 (C") and biotype 9 (C) . Matings in 

 Fg gave of biotype 9 only one pair with progeny, all of biotype 9, which continued 

 to breed true in F4, F5, Fg, and F^. Three matings of biotype 5 were found to 

 breed true C"a, and came true in F^, Fg, Fg, F^, Fg, and Fg, while 10 pairs did 

 not breed true in Fg, but were not analyzed further (C'b). 



One might continue this type of analysis without limit. From the natural 

 population of L. muHitwniata at any point in its geographic range, materials 

 taken from nature show this diversity of gametic constitution as regards this or 

 any other character that might be considered. It follows that each individual 

 may be different from every other in the population, even in a single character, 

 and any individual from nature may produce many geno-different gametes. It 

 has been clearly shown by Johannsen, Jennings, and others that in self-fertilized 

 or sexually reproduced forms any genotypic constitution that may be selected by 

 the observer can be made to continue in straight lines, but in sexually reproduc- 

 ing types it is these " biotypes," or groups of like gametic make-up, that are to 

 be looked for, isolated, and studied. 



A considerable number of writers has isolated genotypes and biotypes that 

 breed true and are asserted, especially by Jennings, to be incapable of modifica- 

 tion by cumulative selection. Jennings, in Paramcccmm, worked only with 

 clones, which are not exactly comparable to genotypes of self-fertilized organ- 

 isms, and not at all to biotypes of cross-fertilizing forms. 



In almost all organisms genotypes and clones will be absent and the biotype 

 will be the group mostly studied and employed. Johannsen's clear definition 

 that a biotype is a group of individuals possessing the same germinal constitu- 

 tion (identical genotypically) leaves no opportunity for misconception regard- 

 ing the nature of biotypes. A current idea among the supporters of the pure- 

 line hypothesis is that genotypically different gametes have no intergrades 

 between them, but are discontinuous. Woltereck, working with Daphnia, has 

 denied this, and others from less secure data have also criticised the view. 



Two questions seem of fundamental importance in this connection and 

 involve the validity of the entire conception of minor groups within the 

 " species." 



(1) Where in any instance does the limit lie with respect to the determi- 

 nation of a genotypical constitution ? I can with my materials, as shown in the 

 first part of this section, easily isolate a race that is homozygous in behavior, 

 absolutely so, which breeds true without limit to the one type, and is in every 



