OF MUTANT CHARACTERS. 175 



At that period, it is true, methods of keeping records were poor as 

 compared with present standards, and errors were all too frequent. 

 Against the supposition that this particular mistake was made is the 

 internal evidence that the proportion of vermilion purple vestigials 

 in the questioned culture resembled that in the other l^ack-cross 

 cultures and is larger than that in any of the rightly labeled F2 cultures. 

 Again, that this culture should be a back-cross test of the female rather 

 than of the male would require a double error — i.e., as to the sex of 

 both parents— and this error would probably have been detected at the 

 time of the transference of the parents to fresh culture-bottles, espe- 

 cially since these parents w^ere transferred to a third culture-bottle. 

 The suggestion has been made that "maroon," a third-chromosome 

 recessive eye-color resembhng purple very closely, had been present, 

 probably only in heterozygous form, in the vermilion purple vestigial 

 stock, and that the introduction of maroon through the vermilion- 

 purple vestigial parent at the Pi and again at the back-cross mating 

 would account for the cross-over class taken to be vermilion purple 

 in the progeny. Such an explanation fails to account for the comple- 

 mentary class of exceptions, the few but carefully attested vestigials 

 that were not-purple. Several other suggestions have been made, 

 and while it seems highly probable in the light of the more recent 

 work that these apparent cross-overs were really due to error in the 

 conduction of the experiment or to unknown properties of the stocks 

 used, none of the suggested escapes from the alternative that there 

 really had been crossing-over in this particular Fi male have solved 

 all the difficulties. 



If these were true cross-overs, it is still possible that their production 

 should have no relation to the mechanism by which crossing-over is 

 ordinarily effected. Thus, Muller (1916) reported a case of crossing- 

 over in the back-cross test of a certain Fi male from the mating of 

 truncate to black. However, all of the gametes of this particular Fi 

 male proved to be cross-overs, so that crossing-over must have occurred, 

 once for all, in an early cell of the embryo, and, as usual, no crossing- 

 over whatever occurred during spermatogenesis. The spermatozoa, all 

 of which were descended from this embryonic cross-over cell, simply 

 inherited the cross-over combination. In the case of purple \'estigial 

 a like explanation would apply, except that in this case the crossing- 

 over occurred in a somewhat later stage of the embryo and in conse- 

 quence only a part of the spermatogonial cells carried the cross-over 

 combination, and only sperm descended from these particular cells 

 produced cross-over progeny. 



That somatic crossing-over has httle analogy to the ordinary type 

 is proved by a similar case of embryonic crossing-over in the female, 

 which was followed by crossing-over of the ordinary type. A nrnting 



