Ee ACTIONS AND PeODUCTS IN INTERSPECIFIC CROSSES 111 



enables it to be recovered from the signaticolUs-like races that arise in F^ as an 

 extracted stock. 



Another test was made with a signaticollis stock that had the Ac determiner 

 with a value of Ac^^. This was a pure native stock that had been in the labora- 

 tory for 15 generations and had from the tenth generation onward been made, 

 by external conditions, to attain the value of Ac^^. This was done by placing it 

 for breeding under conditions in which the temperature ranges were those of 

 the native habitat at Cuernavaca, but in which the water-relation was much 

 more rigorous, especially in the evaporation in the daylight hours. Under these 

 conditions two or at the most three generations are all that is needed to attain 

 the full development of the faster value of Ac. It was found that when a race 

 of pure signaticollis, with the value Ac^^, is crossed with the suspected race of 

 signaticollis, having the diversa form-index, and having the value Ac^°, under 

 the conditions used to develop the Ac^^ value in the pure signaticollis, the result 

 is to produce an F^ array that is signaticollis in general aspect, but whose index 

 is intermediate between the two, on the heterozygous mode, and these when 

 inbred showed in F, a variable number of diversa types, with several others, 

 showing that V, the elytral stripe determiner, has been returned to its normal 

 relations, resulting in the production of gametes that are pure diversa in all 

 respects, and these when combined with likes produce pure diversa types. 

 The^e were found to breed true in subsequent generations, to be pure diversa 

 stock in every way. This reaction became the test upon which this complex 

 and apparently unconformable series was finally elucidated, and is referred to 

 in the remainder of this account of the crosses between the species signaticollis 

 and diversa as the test reaction. 



This reaction shows that in the gametes of the extracted signaticollis appear- 

 ing line with diversa index there is still present the proper agent, V, which, in 

 certain relations of interaction, is able to produce the typical diversa elytral 

 pattern; but that in other relations it is not, but produces the condition of a 

 diversa form, with the signaticollis appearance, and the possible development of 

 pigmentation in irregular spreading areas in connection with the system of 

 elytral punctations. There is present in the elytron the agents needed to pro- 

 duce pigment, and it is constantly active in the production of pigment in the 

 pits that form the elytral punctation series. The T^ determiner is an agent that 

 acts, as far as known, as a pattern-determiner in conjunction with the general 

 pattern-factor of the elytron. For the purposes of the series of crosses between 

 signaticollis and diversa, only the Ac determiner, Math its values represented by 

 the exponent accompanying it, the elytral pattern-determiner Elp, the deter- 

 miner for the punctation system {Pu) upon the elytron, and the pattern- 

 determiner for elytral longitudinal stripes (F). These are agents of specificity 

 in the gamete as is shown by the ability to replace the Pu determiner of signa- 

 ticollis by that from some other source, as also the F determiner of diversa can 

 be replaced by another agent occupying precisely the same relations in this 

 system and productive of its specific type of resultant product. Or, in other 

 operations, the entire group may be dissociated from the gametic complex and 

 transferred entire to a new association. In the figures and plates I have shown 

 these agents with bonds between them, which represent no more than the direc- 

 tions of interaction, or the combinations of action that experience with these 



