104 The Mechanism of Evolution in Leptinotaksa 



In the natural habitat at Cuernavaca, the growing-season, from mid-June to 

 mid-September, 90 days, allowed only the requisite time in the year for the 

 species to complete its two annual generations and prepare for the oncoming 

 of the dry season in September. In the habitat of L. diversa at Orizaba the 

 growing-season was longer, from early June to the late autumn or even into 

 January, giving the species ample time to undergo its development before the 

 onset of the dry season. In the laboratory the average growing-season con- 

 ditions of the habitats were maintained throughout, so that there was no pressure 

 to hasten ontogeny in the case of either. The uniform result under laboratory 

 conditions is to slow down the rate of development of L. signaticollis and to 

 slightly hasten that of L. diversa, both coming to have the same rate of onto- 

 genetic progression. 



It M^ould seem, therefore, that this condition in the stocks on introduction 

 from nature might be productive of the difference foimd in the crossing on 

 introduction and after establishment. If this were true, it then ought to be 

 possible to produce in experiment, from races that were giving the monohybrid 

 behavior, the conditions shown in plate 9. It ought also to be possible to dis- 

 cover something of the mechanism of the production of the difference, especially 

 in the production in F^ of the " pure-breeding " types. 



To eliminate the possibility that the condition was the product of a local 

 difference, in the spring of 1907 I brought to the laboratory from Atlixco, State 

 of Puebla, Mexico, fresh materials of L. signaticollis Stal and of L. diversa n. sp. 

 from Guadalajara, State of Jalisco. These I obtained in late June in the larval 

 condition, and they were taken to the laboratory as pupte and emerged, the 

 sexes being immediately separated. A generation fresh from the field, repre- 

 senting random materials, but in which the virginity of the mates was absolutely 

 certain, was available and at once crossed, with the result that the different 

 crosses of the signaticollis X diversa gave uniformly the same result as had been 

 obtained many times then and since from stocks from the two other locations. 

 These I continued breeding and crossing at random through the seasons of 

 1907, 1908, 1909, and 1910, with the results shown in table 8, showing in all 

 respects the same behavior that was found in all of the other stocks. This result 

 made it certain that the condition was a general one and not due to special 

 local conditions. 



In the years 1908, 1909, and 1910 the same test was made with new stocks 

 from the Rancho Basoco and Cerro Borrega locations, with the same results, 

 showing close agreement in the progressive change in the rate of ontogeny, and 

 with this a similar change in the F^ array presented, which was as far as the 

 tests were carried. These tests were made at the time that otlier portions of this 

 investigation were in progress, as checks to make certain that the conditions and 

 the results that were found and being obtained in the two main sources of 

 material were not vitiated by the use of only one source of supply, in which 

 something unique might be present and productive of the complications found. 



Analysis of the conditions in the habitat of L. diversa at Orizaba and L. sig- 

 naticollis at Cuernavaca, as far as data were available, showed during the grow- 

 ing season that essentially the same temperature conditions were found in both 

 habitats, with differences in ranges that were not great. Relative humidity as 

 ordinarily determined showed nothing of interest, while the rainfall at Orizaba 



