Eeactiojsts and Pkoducts in Interspecific Crosses 115 



in this test series is the formation of a new association of the V determiner, 

 giving a new race ia the series that breeds true. 



In the further breeding the new race (J) that comes out in the test and 

 its testing with the same reaction show that it is in no wise an exception to the 

 entire series, but that the test will, under the conditions of experiment, produce 

 the dissociation of the agents, restoring them in some of the gametes to their 

 original relations, and giving diversa as the result in lines in which it had not 

 been present for generations. This dissociation, in this instance, is easy to 

 accomplish if the temperature and humidity are increased above that previously 

 used and if the evaporation rate is low. 



There results from this testing and other operations two races that in appear- 

 ance are signaticollis, in reality are diversa, with the V determiner in different 

 positions in the gametic system, and that breed true without limit. 



This series of observations and experiments with crosses between two species 

 under conditions that give the reaction shown in plate 9 are interesting and 

 suggestive of many important considerations. In that portion of the F^ popu- 

 lation that was signaticoUis-like in aspect there were possibilities for misinter- 

 pretation and wrong conclusions and operations. The fact that it bred true, so 

 that in ordinary operations it would have passed as a pure extracted race of 

 signaticollis, that had come out as the result of one or the other of the parents 

 being " heterozygous " for some complement or factor present in homozygous 

 condition in the other, might easily lead to confusion, and the further test cross- 

 ings would have been interpreted as mutations in the extracted line. If the 

 operation had taken place in nature and had been unknown, and the stock had 

 come into our hands in the laboratory, the conclusion would be drawn that it 

 had " mutated," giving new " elementary species " or races. 



Aside from the potentialities for mistakes in the laboratory, the series is 

 suggestive of important roles in nature, both in the matter of rendering species 

 in nature complex and also in the role of serving as an actual motive force ia 

 evolution. It is not difficult to picture a series of events in nature, such that 

 introduced forms in crossing would, in the manner shown here, gametically 

 change the nature of the resident form without in any way altering its pheno- 

 typical aspect, so that it would pass as one form, " variable " perhaps, which at 

 some future time might mutate, producing quite new characteristics, which were 

 in reality only the reconstitution of an arrangement that had previously formed 

 a portion of the attributes of some intercrossed species. An apparently pure 

 signaticollis race, when crossed with another of known purity under the con- 

 ditions internal and external that have been stated, showed in the test in Fj an 

 array of 10 types in differing proportions. Some of these 10 are pure-breeding 

 when mated with likes and remain constant, even though they may be hetero- 

 zygous. Others are pure, as far as tests can determine, and others are hetero- 

 zygotes in the ordinary sense. The array might in nature well be the basis for 

 the origin of new groups or species. 



In many respects the action of the signaticollis race that comes out in F^ 

 reminds one of the fixed hybrids that are often described. I have had a number 

 of instances in my cultures of fixed pure-breeding heterozygous races, and all 

 appear as fixed hybrids in the ordinary tests that are applied. The fact that the 

 form comes true in breeding may be of importance, but it is not a reliable 



