Eeactions and Products in Intekspecific Ceosses 133 



Thus far only one test agent has been found that serves to even partly show 

 the composition and nature of this stable heterozygote, namely, the pure 

 nndecimlineata. The suspected race appears as undecimlineata in all respects, 

 and the test of its heterozygous nature would be to recover from it in some 

 manner the signaticoUis character, either in whole or in part. When pure 

 undecimlineata and the suspected type are crossed under the ordinary breeding 

 conditions of the laboratory nothing results, the two breeding with perfect 

 freedom and no separation of the types appearing ; and perhaps the only reason 

 for so persistant pursuit of the race is the constant presence in the suspected 

 type of the trimodal condition in an otherwise perfectly uniform fraternity and 

 series of fraternities. 



By applying the same test to the fraternities that I used in the cross of diversa 

 X signaticoUis it was found that the Fg fraternities were trimodal, and that the 

 modes of the compound curve were essentially the same as the parental modes, 

 with the intermediate mode that of the heterozygous type. It was further 

 found that breeding from the modal conditions of the extreme modes at once 

 gave constant lines so far as the form-index was concerned, and that this was 

 in value the same as the parental modal values in pedigreed lines. Only meas- 

 urements of fraternities were of any use in this analysis, as fraternities are the 

 only homogeneous materials that we have to work upon. In figure 9 I have 

 shown the results of a biometric analysis of the suspected race and the products 

 of breeding in line from the modal conditions. 



In many respects this would by many be considered competent proof of the 

 heterozygous nature of the suspected line, but tlie final proof is to recover the 

 suspected contaminating agents from the line. In this I have, to the present 

 time, been only partly successful, due in the main to the lack of the requisite 

 facilities for control and production of the necessary external conditions. 



I have found that when the suspected race is crossed with a pure line of 

 undecimlineata under conditions to be shortly described, there resulted in the 

 progeny of such a cross the recovery of signaticoUis larval characters — i. e., 

 the yellow larval color and the meristic pattern system of the third stage. 

 These could only have come from the suspected race, as they are known not to 

 be present in the test materials used — i. e., pure L. undecimlineata. 



When the test race is crossed with the race to be tested under the following 

 conditions, the Fg fraternities show signaticoUis larval characters in widely 

 varying proportions in about 50 per cent of the trials. The combination of 

 conditions that I have employed are those of a desert complex, with high daily 

 temperature and intense desiccation, and low night temperatures, with satur- 

 ation, or nearly so. The average conditions were, in the tests that have been 

 made: 



Table 16a. 



