334 The Mechanism of Evolution in Lbptinotaesa 



made upon the portions of a natural uniform species, variable to be true, and 

 complex in its make-up, and for that very reason probably a good one upon 

 which to test out some of these points, showing as do the results that in a uni- 

 form but variable species (in the taxonomic and faunistic sense) the materials 

 taken at different points over its natural range are not of necessity identical in 

 their germinal constitution, but that the conditions differ in neighboring placed 

 habitats, and that these gametic habitudinal differences are permanent. 



Similar tests were made with less satisfaction, owing to practical difficulties 

 in the years 1905 to 1907 in some of the steep-sided and well-isolated ravines 

 that are found in the valley-head above Maltrata. These ravines are the loca- 

 tion of small streams falling precipitously over the edge of the Mexican 

 Plateau to the level of the Maltrata Valley. The location was much more 

 difficult to operate in, and the results were not so interesting, owing to the all 

 too frequent damage to some of the locations by heavy torrential rains that 

 come in the rainy season, often washing everything before them in some of 

 these ravines, and one can not predict when one of these will strike the 

 chosen spot. 



The outcome of the tests that were made was in all respects identical with 

 those on the pedregal in the valley of Mexico. Of most interest were the 

 results of the series from Chapultepec, Chalcicomula, and Tlalnepantla, none 

 of which changed in any respect during the time that they were in the test loca- 

 tions, giving uniformity of action in all. 



One point is of practical importance in this connection, namely, in placing 

 these colonies for testing the condition in the population, it must be as certain 

 as possible that the location chosen for the test shall not be such m its conditions 

 that they will serve as a modifying agent and so give confusion in the determina- 

 tions. I used all of the precautions that were possible to avoid this, and seem 

 to have been most fortunate in my selections, not one of the test-series locations 

 showing anything in the way of changes in the population that could in any 

 way be regarded as a direct modifying action of the environment. 



TESTS WITH L. UNDEaMUNEATA. 



The populations of this species at Tierra Blanca and San Marcos are appar- 

 ently not sufficiently different to provide much in the way of contrasting condi- 

 tions for testing, so that the permanency of the condition in the pattern was 

 tested in this species by taking some of the population from the Tierra Blanca 

 to the region of Campeehe, and from a point near Campeche to the Tierra 

 Blanca in 1908. The population in the two locations apparently differs per- 

 manently in several respects. The chief difference between the two is the 

 absence in the Campeche location, near Seibacabecera of biotypes 4, 5, and 9, 

 with the preponderance of the groups 6a, 7, 13, 14, and 15, making a population 

 in which there was little breaking in the continuity of the array. The Tierra 

 Blanca colony, on the other hand, shows the 6a, 13, 14, and 15 series weak, and 

 in any given generation separated by large gaps. The stocks were inter- 

 changed in May 1908, and each put in a location not in contact with the original 

 colony, but isolated from it as well as the conditions in each location would per- 

 mit. The samples introduced into each location were fairly random. I did 

 not see the colony again until the winter of 1909-10, when a portion of the 

 colony at Tierra Blanca was probably in the groimd in aestivation, but enough 



