Analysis of Heterogeneity in Some Simplest Characters 233 



in figure 33A. No decided influence of the uniform conditions in restricting 

 the range of the index was found, the range at the end being about the same as 

 at the start. 



Second, there was the possibility of " impurities " in the constitution of the 

 strain. I went over the living stock carefully and over such of the preserved 



Fig. 33. — Diagrammatic representation of results obtained in the attempt to 

 breed pure lines as regards form-index in L. multitceniata. (bee text.) 



material as I had, and found in this " line " a suggestion upon which I went to 

 work. In all of the biotypical lines of this multilineata series the spots a and a' 

 are parallel with the median axis of the body and are also moved close thereto. 

 I found in this strain that these two spots were in many of the specimens some- 

 what divergent at the anterior end and were, on the whole, more widely separated 

 16 



