Eeaotions and Peoducts in Inteespecifio Chosses 135 



when they may be removed to the usual conditions of the laboratory for the 

 remainder of the cycle, and then again treated as before in the mating of F^. 



The test has usually been applied to Pa and P3 fraternities of the suspected 

 strains and to lines whose form-index was that of the parental signaticollis 

 species. The results show that the recovery of the signaticollis larval characters 

 from the different tests is by no means found in every test, nor in any regular 

 proportions, and sometimes it is only yellow color, at others it is the meristic 

 spot system, in others both that are recovered, showing that the operation needs 

 further investigation. 



After the recovery of the larval signaticollis characters from the race, they 

 continue to breed as true and to be as accurately alternative in crossing as if 

 they came from pure signaticollis, and the end-result is that from one of these 

 tests it is possible to get certain lines with these constantly present as the larval 

 characters in materials otherwise undecimlineata. Ko indication of the posi- 

 tion and relations of these hidden groups of larval characters in the suspected 

 race that I care to discuss has been determined. However, the test, in com- 

 mon with the trimodal condition in the Pa fraternities in the suspected line, 

 shows conclusively that the line is in reality heterozygous and is in all respects a 

 suppressed heterozygous series, excepting that in this series the combination 

 that the signaticollis groups of agents have made are binding and broken with 

 great difficulty, so that only certain most precise conditions and interactions re- 

 sult in the recovery of the agents from the suspected complex. It would be of 

 interest to know how long this condition would remain in the line. I have 

 found them present in Pj in one series and in P, in another, but I do not know 

 how much longer they may remain. I do not know whether they play any other 

 r61e in their new situation, as there is no evidence of new characters having 

 come out in the line as a result of the contamination present. 



One other suggestive test has shown that the adult elytral-pattem system is 

 also present in the suspected race and has been recovered in the two tests that 

 have been made of this kind. It was found that when the test was made with 

 the same day conditions as in the first test series, but with the night conditions 

 having a much lower temperature, as low as 10° C, with saturated atmosphere 

 and no wind-movement and no evaporation, the number of individuals showing 

 the larval characters increased and that a few individuals arose showing the 

 signaticollis adult character. 



These adult signaticollis characters thus recovered do not change, but are per- 

 manent in the race from that time on, and in one test produced in F^ a pure- 

 breeding homozygous race that was signaticollis in its adult characters. I have 

 not carried this testing further, as the requisite apparatus has not been available, 

 and an ice-box will not serve for the nocturnal low-temperature portion of the 

 test conditions; its conditions are too unsanitary and are productive only of 

 failure. 



Collectively the results of the testing show that the undecimlineata type that 

 appears in Pj is in reality a fixed or masked heterozygote in which aU of the con- 

 trasting characters are present, but, in some way not discovered, are bound in 

 associations that are not broken excepting by certain specific and energetic 

 treatments. There is no doubt but that in the cross between signaticollis x 

 diversa and signaticollisxundecimlineata, when the Ac values are different. 



