12 Mohr. 



realized that siuged possessed somatic peculiarities which made a 

 separation from forked easy. Later on the converse situation was met 

 with: a new sex-linked recessive occured which looked so strikingly 

 like singed that there seemed to be no doubt that the character was 

 due to the reappearance of singed as the result of a new mutation. How- 

 ever, wlien tested, this new mutation proved to be a new allelomorph 

 of forked. 



The origin of forked^. 



In selecting from the pure coral stock males to be used in a 

 mating, it was observed (Jan. 20, 1921) that one out of the 30 males 

 present had forked bristles (coral, eye color, w"" at 1.5, an allelomorph 

 of white). The bristle character looked entirely like ordinary forked 

 and in order to ascertain whether we were dealing with a recurrent 

 mutation in the forked locus, the coral forked male was double mated 

 to 2 Bar and to 2 eosin vermilion forked females from the corresponding 

 pure stocks. The culture ('2471, Jan. 31, 1921) gave in three days 

 7 Bar and 18 eosin-coral forked females; the fact that all the eosin- 

 coral females were forked proved that the new character, forked 2, was 

 identical with or allelomorphic to forked. Of the male offspring 7 were 

 Bar and 11 eosin vermilion forked, while 12 which also were eosin 

 vermilion showed a bristle and hair alteration which seemed in every 

 respect closely similar to singed (Fig. 3). 



It had been shown that double recessive singed forked flies are 

 indistinguishable from ordinary singed, and it was accordingly assumed 

 that one or both of the eosin vermilion forked mothers had in addition 

 to these genes contained a singed gene in one of their X-chromosomes 

 as a result of a reoccurence of the singed mutation in their ancestry. 

 The males which received this chromosome would be eosin singed ver- 

 milion forked, but somatically then- hairs and bristles would be singed. 

 Some of the e5*ceptioual males w^ere therefore crossed to females 

 heterozygous for singed, and it was expected that half of their daughters 

 would be singed. However, all their daughters were wild-type, though 

 the fact that half of their sons were singed proved that the females 

 used had been of the desired constitution. This result demonstrated 

 that the modification of the forked character in the exceptional males 

 could not be due to the ])resence of singed or to an allelomorph of 

 singed. This was confirmed through another test: eosin-coral foi'ked 

 females from 2471 had been back-crossed to their own fathers. In one 



