138 The Mechanism of Evolution in Leptinotaksa 



in the form-index ; that from these can be reared pure lines of the form-index ; 

 but this is possibly only by isolation and inbreeding. During the same time 

 that these were in progress certain mass-culture lines were carried, with the 

 result that a test of the form-index that I had made in the fourth generation 

 showed only a monomodal curve, and the same was true in the sixth, which I 

 also had tested. From the fourth I was able, by mating likes, to produce by Fg 

 much-restricted lines in the modes, so that there had taken place in the mass- 

 breeding a mixing with the result of obscuring the real conditions present, but 

 out of which I could isolate pure lines with respect to the form-index. 



The production of this undecimlvneata line is essentially the production of a 

 good species in every sense. It is strong, well able to survive widely ranging con- 

 ditions ; the fact of its slow breeding at once isolates it from crossing back with 

 the rest of its relatives in the production of Fg ; its low power of dissociability 

 with respect to its contained potentialities, the fact that it is produced in num- 

 bers of both sexes in the one operation, shows how in nature a new group might 

 easily arise that would thereafter become a permanent element in the fauna of 

 the location. Any of the usual tests of breeding would in natural materials 

 have shown only uniformity, when in reality they were heterozygous in origin 

 and in character, and by appropriate means the fact of this can be shown. It is 

 not known how long the signaticollis groups of agents would remain in the race 

 unseen, nor is it known what part they may be playing in the complex as it is 

 seen, but as far as one is entitled to an opinion it seems to me that they are most 

 probably present in some association that ties them with other agents, the 

 totality acting as a unit system in reaction, but in which a portion of the group 

 is relatively unimportant and may be recovered without altering the aspect of 

 the series. 



That the reproductive cycle is also in reality heterozygous is shown by the 

 experiences in two series — one in Fg, the other in F^ in another line entirely 

 independent of the first, in which, after subjecting the lines to severe conditions, 

 there came out of each a line that showed the normal rhythm with some changes. 

 I was not able to duplicate these experiences in later tests, and, therefore, am 

 not disposed to base more than the suggestion upon them that the experience 

 shows that in this character of the reproductive cycle the usual double brooded 

 condition could be recovered if we knew how to dissociate or analyze the present 

 combinations. To all intents and purposes the line is a stable typical species, 

 and the series shows how such might easily arise in nature. In ordinary terms 

 this is a fixed hybrid, and while it is fixed in one sense, it is by no means 

 irretrievably so altered, nor is the condition present in any manner at variance 

 with the principles of factorial composition and evolution, but as far as I am 

 able to discover are entirely a confirmation of them and an extension of their 

 operation into a type of operation that may well have played a frequent and 

 important role in the production of new specific groups in nature. But any 

 such group, pure as it might come in the usual breeding operations, has present 

 within it the associated agents ready, on the realization of the right conditions, 

 within and without, for the production of new characters or races by sudden 

 changes. 



Three other main types of reaction have been found in this series, which, 

 while in first appearance entirely at discord with any of the usual considerations 

 in crossing, are nevertheless entirely in harmony, as far as tested with the 



