232 The Mechanism of Evolution in Leptinotaesa 



cially in the direction of a higher index, which was away from the condition in 

 the species as a whole. It would have been of interest to have followed out this 

 suggested tendency in the upper group, but space in my laboratory was at a 

 premium, so that the entire culture at the beginning of 1909 became the victim 

 of a struggle for space. In any event the culture had shown clearly the main 

 point at issue, namely, the presence in the population of minute differences, 

 fixed in the germinal constitution of the stock and capable of being isolated, and 

 held as distinguishable groups when the proper refinement in method was 

 employed, showing clearly that the " fluctuations " are not all uninheritable 

 in these " biotypic " lines. 



It must be remembered that these lines were not "genotypes," and as 

 " biotypic " lines are not fully comparable with the " genotypical lines " with 

 which Johannsen, Shull, East, and others have worked, and concerning which 

 so many of the recent qualifications have dealt. However, as most narrowly 

 delimited lines of biotypical breeding, they are good examples and do show, in 

 these lines, that the problem of the " fluctuations " is by no means as simple as 

 it might seem from some rather emphatic statements as to the limitation and 

 nature thereof. 



The issue might be raised that the original group was not a true biotype at 

 all, but a complex assemblage, and this may be true ; but where is the limit to 

 be drawn? With the ordinary methods in use the first group satisfied the 

 requirements fully, and remained as a constant assemblage with the customary 

 range in the index after once it was established without further attention or 

 selection on the part of the breeder. If I had gone further no one would for 

 a moment have even thought of the group being anything but a simple one. 

 Nor do I know how complex it was in any instance, nor how many of these 

 minor strains I could have isolated from the mass of a single population, with 

 time and refinement of method. In all of the pure-line work the same seems to 

 be true, that with refinement of method the lines may well become finer and 

 finer, less and less remotely separated, and lines only in the sense that our 

 methods of analysis and pure breeding hold them as such, and in a most 

 unnatural condition. 



The third possibility remains, that some of the " fluctuations " may have been 

 due to loosely combined " impurities " in the constitution of the lines, which 

 would in this instance be in the nature of substances present as "factors," 

 which, not being in the proper place and relation in the system, or lacking the 

 proper complement, gave no direct resultant of their own in the reactions of the 

 other " factors " of the mass, but by their presence served to disturb the normal 

 workings of the system and thereby produce distortion of the action and deforma- 

 tion of the resultant " character." It is well known that in the non-organized 

 bodies, this type of deformation is common ; indeed is the most common one in 

 nature, and there is no a priori reason why the same may not well be true in 

 organized matter, although not so easy to test or to certainly determine. 



In the culture 10 g VII (BAF), and again in the next generation of the same 

 line, as shown in table 32, a wide range in the index is shown, and the same 

 method that produced the restriction of the range in the other cultures had 

 failed in this one, so I tested it for two possible causes of the condition found. 



First, four pairs were put into controlled conditions and carried through four 

 generations with the results shown in table 33, and the graphic representation 



