OF MUTANT CHARACTERS. 209 



failing to show the character. Such a character can be used, though 

 very inefficiently, by considering those flies which do show the charac- 

 ter and disregarding the majority which do not. 



That gap is recessive was indicated by the mating carried out when 

 gap reappeared in still another experiment (May 27, 1913). A gap 

 (black arc) female out-crossed to a wild male gave only not-gap off- 

 spring (Culture II 38, + 9 51, X cf 63). On the other hand, another 

 such female out-crossed to a black male gave in Fi 38 gap to 65 not-gap 

 offspring (II 41). This seemed to indicate dominance, since the bkick 

 stock had been examined several times and no gaps had been found. 

 In addition, several of the gap males and females were bred together, 

 and these produced a stock which seemed pure, since in every fly at 

 least a weakening of the vein and color could be seen, while fully 90 

 per cent showed a distinct and large gap. The stock maintained this 

 condition when run without selection, while at the same time the 



Text-figure 76. — Gap venation, showing the break in the fourth longitudinal vein. 



regular black and black arc stocks failed to show gap when examined 

 occasionally. It now seemed probable that there might be some 

 special relation operative in the case of gap to account for the apparent 

 dominance when crossed to black or black-arc and the recessiveness 

 when crossed to wild. About a year later (July 1914) the cross of 

 gap black arc to black arc was repeated, whereupon the same domi- 

 nance reappeared in the case of two tests out of the three (table 56). 

 Some of these Fi gap males were in turn crossed back to black-arc 

 females from stock, with the result that only a very small percentage 

 of the offspring showed gap (table 56). IMost of these gaps were 

 females and all showed the character very weakly. This result was in 

 sharp contrast to the Fi result, where the gap cliaracter was well 

 developed, and appeared in fully half the flies, about equally in males 

 and females. 



A not-gap Fi black arc male similarly back-crossed to a black arc 

 female of stock failed to give any gap in 123 flies. This may indicate 

 a genetic difference between the gap and not-gap P^ flies. An F2 

 culture raised from a gap female and a not-gap male showed 35 strongly 



i 



