Keactions and Pkoducts in Interspecific Ceosses 167 



the decemlineata form-index is present, yellow where that of multitceniata is 

 present, and complicated when the individual is heterozygous. 



The reaction in this cross is plainly of a dihybrid type in which the gametic 

 systems represent one unit in the reaction, with all of the non-dissociated agents 

 and characters proper to each, and the two head-pronotal patterns the other pair 

 of characters. In reality these are the dissociated agents, although in this 

 instance each is complex, even though acting as a unit in the reaction. I have 

 not the requisite data to attempt a statement of the ratios present, owing to the 

 necessity of testing out many more pairs of the heterozygous individuals, whose 

 composition can not be approximated by inspection, and whose study is further 

 made complicated by their great sensitiveness to conditions in the medium 

 which act to cause extreme fluctuations that obscure their true nature. The 

 existence of the four pure-breeding lines are, however, all of the data needed 

 here, in that they show fully the interchange of the head pronotum-pattern 

 producing groups as units in toto, and their combination with the other parental 

 gametic system. I have only been interested in the fact of interchange and the 

 number of possible stable combinations that are derived and the type of reac- 

 tion. Crosses so plainly dihybrid in type are not worth the effort to work out 

 the actual composition of the heterozygous types present, and especially so as 

 there has been no indication that there was anything unusual or interesting in 

 them. 



In this cross precisely the same species bases are used, the only difference 

 being the presence in multitceniata of the agents productive of the melanothorax 

 type of head and pronotum coloration. The type used was not the recurrent 

 mutant melanothorax, but only the transfer of the head pronotum group of 

 pattern agents into the biotype 7 complex, displacing the corresponding com- 

 plex, giving the type used. The cross had only the one difference present, that 

 of the two pattern-agent groups mentioned and the specific non-dissociated base. 

 In the first instance the crossing showed only uniform monohybrid reactions, 

 the two systems remaining intact; in the second the two remained intact, but 

 with the exception of the one interchange between the two groups of agents 

 observed. Why this should be so can not at present be decided, but it is clear 

 that the reaction is not the product of a chance shifting of independent agents, 

 but is more probably the interaction of the two gametic systems upon the basis 

 of their nature and conditions of reaction, these determining the extent and 

 character of the dissociated and interchanged agents or groups of agents. 



From this point it is possible to make all sorts of complications in the cross- 

 ings of these biotypic lines, simply by adding to one or the other agent or groups 

 of agents by the appropriate methods. Thus, for example, if to the same two 

 stocks utilized in the preceding experiment there be added to the L. multitceniata 

 line the red elytral color, the complications are greatly increased in the array 

 of pure-breeding types found in F,. 



The red may be derived from any source, but the most convenient one that I 

 have found is the geographic variety variabilis. This is readily reduced in cul- 

 ture to a homozygous-acting red race, and when crossed onto the combination of 

 multitceniata and melanothorax types gives a race that is pure for all of the 

 characters, is homozygous in action; and this, when it is crossed with the 

 decemlineata race used, further complicates the Fo array by the production of 

 yellow, orange, and red in pure-breeding races, thus adding to the possibilities 



