222 CHARLES ZELENY 



upwardly directed mutations in a high selection line while in a 

 low selection line the downwardly directed ones should be in the 

 majority. This view is supported by May's data, in which 

 eight of the nine mutations to full appearing in the bar-selection 

 experiments were in the high lines. The exceptional one 

 appeared during reverse selection from a low line. 



The present data obtained in the course of forty-two genera- 

 tions of selection from high facet number on the one hand and 

 low facet number of the other are given in figure 2. There is 

 clearly no significant difference between the high and low lines in 

 the frequency of mutation either to full eye or to ultra-bar eye. 

 There are four mutations to full in the high line and three in the 

 low line, and ultra-bar appears once in each line. A final con- 

 clusion on this point must of course await more extensive data, 

 but attention is called to the fact that the provisional conclusion 

 drawn here is supported by the findings regarding rate of muta- 

 tion in stocks derived from the selection lines as given in the 

 following paragraph. 



7. Does high or low facet number m a stock have an effect upon the 

 direction or frequency of mutation? High stocks were obtained 

 from the high selection lines and low stocks from the low selection 

 lines. Such stocks have remained high and low, respectively, 

 when bred without further selection. One of the very high stocks, 

 emarginate, arose by mutation during upward selection. 



It has been supposed by a number of students of evolution 

 that a high stock has a greater chance of reaching any particular 

 higher point than does a low stock, because the distance to be 

 traveled is less. As stated above, it has been demonstrated for 

 the bar series that within the bar locus there is no greater fre- 

 quency of the short than of the long jumps. The germinal 

 differences to be considered under the present heading are, how- 

 ever, not at the bar locus. As stated elsewhere (Zeleny, '18) 

 and demonstrated more fully in a paper on the effects of selection 

 which is to be published shortly, the high bar stocks discussed 

 here differ from the low ones in their accessory factors. As shown 

 in table 3, there is no significant difference either in direction or 

 infrequency of mutation between high and low bar stocks. The 



