Analysis of Heterogeneity in the Population 335 



remained upon the plants to make a fairly satisfactory analysis of the popula- 

 tion, while the conditions at Seibacabecera were more satisfactory in this 

 respect. In each some three or more generations had passed since the intro- 

 duction, and at each location there was the probability that some of the native 

 population entered the introduced colony and bred therewith. In spite of this, 

 in the changed locations involving considerable climatic differences and the 

 most probable interbreeding with native stock, the census made in the loca- 

 tions showed that in the time elapsed and with the combined action of the 

 slightly changed conditions and the mixing with local conditions, the two 

 introductions had changed but little if any in the period of the test. I did not 

 see the Tierra Blanca location again until the winter of 1911-12, when the 

 adverse conditions had driven two-thirds of the population into the ground for 

 the dry season ; but of these that remained, or that were recovered from beneath 

 the debris on the surface, it was evident that a considerable mixing or environ- 

 mental effect had been produced. This was indicated by the abundant presence 

 in the population of biotypes 4, 5, and 9, that were only represented by traces 

 in the previous census and that were wanting in the original location. The 

 companion test location has not been visited since 1909, owing to the disturbed 

 political condition of the country. 



The conditions observed in the Tierra Blanca test location are certainly not 

 the product of the external conditions, because conditions in the pattern-system, 

 represented by biotypes 4, 5, 6a, and 9 are the known product of the pres- 

 ence or absence in the gametic system of exact agents, and at no point in all of 

 my experiments or observations is there the least evidence that conditions in the 

 environment, especially so little different in the two locations, could or had ever 

 produced these changes in the pattern of any group. It is certain, therefore, that 

 the change in the condition of the test colony, which was not and could not be 

 perfectly isolated, had resulted from the incoming of native materials and the 

 incorporation of the missing agents into the introduced group to produce 

 the absent biotypes. So likewise other near locations in the region were show- 

 ing the effects of the introduction by the increased numbers of the biotypes 

 most characteristic of the Campeche stocks. These crude operations in nature 

 show the permanence of the condition in the pattern in the different locations, 

 regardless of the conditions of the new location, provided that the location does 

 not have conditions that are intense enough to act as agents in directly pro- 

 ducing germinal changes. In instances of this sort the action of the intro- 

 duced colony is always different from that seen in this series of tests. 



At one time or another, to different degrees, all of the stocks that I have used 

 in these investigations have been tested in this manner, and in no instance 

 have I found that when the test was made from the point of view of the consti- 

 tution of the character did the condition change when the habitat was changed 

 within reasonable limits, which is a very different ending from the condition 

 found when the permanency was tested in terms of the statistical value of pig- 

 mented area present, when the conditions were not permanent in any instance 

 tested. This is due, of course, to the fact that the area, or the measured bio- 

 metrical character in general, is not the character, or a character, in the 

 individual or the race. 



In the materials that I have u^ed it was necessary to establish the permanent 

 nature of the differences that are shown in restricted locations in nature for two 



