14:2 The Mechaj^ism of Evolution in Leptinotaksa 



These two groups are alike in reaction under ordinary conditions and unlike 

 in constitution, actually and in principle, but ordinarily would be rated as 

 homozygous, and so used in the further crossing of them. The first group ofPers 

 nothing of interest or difficulty and is the product of the usual metathesis of 

 groups of agents that takes place in crosses of the type where there are three 

 agents or groups of agents capable of dissociation and rearrangement. The lat- 

 ter group would in usual operations be interpreted by the assumption of some 

 agent in one parent not present in the other, one being heterozygous, the other 

 homozygous, therefore with the result of the production in F^ of the extracted 

 t}'pe. That this is not the case in these experiments is shown by the fact that 

 from all of these in F^ races it has been possible in varying degrees to recover 

 from them the other characters, which would not be possible were they true 

 homozygous extractives. There are thus available 10 homozygous-acting kinds 

 of substance, 8 of which are homogeneous in composition and action, 2 in action 

 only, and in addition there are the diverse kinds of heterozygous combinations. 



LEPTINOTARSA UNDECIMLINEATA X LEPTINOTARSA DIVERSA. 



Crossing of these two species is simple, and also serves to show that the reac- 

 tions are normal metatheses of the character groups involved. Between the 

 two there are the larval character groups of lipoid color, yellow or white, and the 

 spot system or its absence, while in the adult the elytral-pattern system of 

 agents, while not precisely alike, is with difficulty separated, so that it is not so 

 clearly a diagnostic character in the adult as they were in the former crosses. 

 However, the antennse in their coloration serve as a certain mark of recognition 

 for separating the two adult conditions. This is not in this type of cross dis- 

 sociated from the general species complex, as far as known, always appearing 

 in correlation with the form-index of the proper species. The basal joints in 

 L. undecimlineata are light in color, the distal are black; and in L. diversa 

 the entire antennse are black. This in common with the index serves well to 

 separate the adult conditions. In both the Ac determiner values are alike, so 

 that no derangement of the crossing reaction occurs, and the conditions 

 employed, unless in extreme ranges, do not change the array. Eeciprocal crosses 

 show identity of F^ results, giving yellow larvge with the spot system minutely 

 present or not visible, so variable in this manifestation that precise statement is 

 possible only for individual instances, due to the fact that slight variations in 

 the conditions of the medium produce oscillations of the manifestation of this 

 character. The adults are also an intermediate in all of the characters, as 

 antennse, index, pattern, and so on. The production of Fg from these shows 

 separation of the larvas in the second stage into white and yellow ; into an array 

 in the third stage that consists of 9 white without spots, 3 white with spots, 3 

 yellow without spots, 1 yellow with spots, and each of these give three classes of 

 adults that appear as one, but that can be separated on the basis of the antennse, 

 the form-index, and to some extent by the elytral pattern. There is no indica- 

 tion that the reaction is in this instance anything but of the most orderly tri- 

 hybrid type in all respects. 



With this combination I have not attempted many crosses when the Ac 

 determiner had divergent values experimentally produced, because of the 

 technical difficiilty that the separations of the adult types were slow and must be 



