Analysis of Heterogeneity in the Population 327 



general type of results arrived at in L. multitmniata, and in the same way with 

 materials from Tierra Blanca and Campeche the conditions in L. undecim- 

 lineata, which will be fairly illustrative of the method and principles. In this 

 work it would have been fortunate if the opportunity had been available to have 

 used cages in some of the habitats, to contain materials taken to that point from 

 some other, so that the tests would be made more accurate and much more 

 critical. One method employed was to find isolated locations in which the 

 species was not living, stock them with food, and then introduce the materials to 

 be tested. This was not so difficult in southern Mexico, where the topography 

 lends itself admirably to the needs of problems of this sort. Most fortunate 

 conditions of isolation were found ready or easily created in the southwest por- 

 tion of the valley of Mexico. At this point in the valley a huge surface lava- 

 flow has broken out of the sides of the volcano of Ajusco and flowed down over 

 the sides of the valley, covering the whole mountainside from Contreras and San 

 Angel on the northwest to the Rio San Buenaventura on the south and to 

 Tlalpam on the northeast and east. Although recent in geologic time, there 

 had been sufficient erosion and denudation to provide the region with a rather 

 well-developed but peculiar flora, and, most important, the forming of pockets of 

 soil in the much-broken surface of the flow that could be utilized as locations 

 for growing the food for these animals for the introduced colonies. In the 

 area as I found it there was no food for these animals and none of them was 

 found living in the area, due to the fact that the agencies of dissemination in the 

 valley were entirely against their introduction. The pedregal had become 

 inhabited by plants and animals from the higher slopes of the mountainside that 

 had been carried down by the wash of the storms. No roads traversed the 

 area, and only a few footpaths crossed the edges thereof, so that introductions 

 by transportation were trivial factors. The added character of the area that 

 it was of no use as a grazing-ground made it immune to the introduction of S. 

 rostratum, the food of these animals, by grazing cattle, one of the most common 

 and efficient means of disseminating the plant. By going into the area from 1 

 to 2 miles, conditions of complete isolation were obtained, and with consider- 

 able labor and trouble the proper locations were obtained. These had, within 

 certain limits, the same general set of conditions, which, located as they were, 

 were not greatly different from the location at Chapultepec. It was found in 

 practice that the dissemination from the colonies was slight, owing to the char- 

 acter of the bare, dry lava-flows about each of the locations chosen, so that if 

 any of the beetles did wander from the colony the chances were immensely 

 against their ever gaining the next one in the maze of pits, cracks, caverns, 

 bare, hot rock surfaces, and so on, that completely filled the area, so that the 

 random wandering of one of these animals would with large probability end in 

 elimination and not in reaching the next location, even when only a few hun- 

 dred yards away. So broken and bad was the surface, so cut by pits, huge, 

 deep cracks in the lava, that passage of the portion used, on either horse, mule, 

 or burro, was impossible, while foot travel was slow and tiresome. 



The first tests were begun in 1905, when 10 good locations were found, 

 located in the midst of the pedregal and isolated by distances varying from 300 

 yards to a mile apart. The majority of these locations were depressions in the 

 face of the flow into which denudation from above had washed accumulations 

 of soil, in some instances to a depth of several feet. The areas were mostly 



