204 The Mechanism of Evolution in Leptinotaksa 



trail of Tuxtepec; also typical savannah but more open and park -like, fewer 

 trees and many annual and shrubby perennials, and about 300 feet above sea- 

 level. At San Marcos, 600 feet above sea-level, the location was about 1.5 miles 

 west of the station on the eastern edge of a rainy season lake. The area was 

 savannah, but with tree islands, poorly developed drainage, abundantly watered 

 and covered with short grass and low herbaceous plants. 



At all three places the annual cycle of seasons is the same in type but differs in 

 details. Eecording thermometers that were secreted in the open and read on 

 returning gave the highest and lowest temperature that had elapsed since the 

 last observation. Tierra Blanca gave the highest temperature, 107° F. in the 

 shade. May 1906, and the lowest record obtained there was 54° F. in December 

 1905, during a " norther." San Marcos was always cooler, the extreme records 

 being 97°F. in May 1906, and 45°F. in February 1906. El Hule was more 

 moist and always had fairly high humidities at all seasons, while Tierra Blanca 

 had the least rain and the lowest humidities. 



The plan adopted was to go to each location at proper seasons and there make 

 determinations of the value of at least 500 individuals. Owing to the peculiar 

 localized character of these colonies, none, or at least only a few, were removed 

 because it early developed that artificial removal soon altered the tone of the local 

 group under observation. To have removed 500 would often have taken nearly 

 the entire population of the place. Usually one or two natives were employed 

 to gather the specimens, while I determined and recorded the values. To facil- 

 itate this, scales of areas were made as a basis of comparison, and the material 

 seriated quantitatively, and also camera outlines were made and measured by 

 planimeter later. Either method gave the same results, the latter more detailed, 

 but no more accurate or significant. Variations in shape were recorded in 

 terms of type of variation. Two sets of observations were obtained, therefore, 

 at each location, and for each period of observation a quantitative measure and 

 an expression of the frequency of different types of distribution of the pigment 

 and the variation directions of the spot. As a rule, two to four days at each 

 place were necessary to obtain the requisite number of records. No specimens 

 were returned to nature until after the records were completed. 



The results obtained in the years 1905, 1906,. 1907, 1908, and 1909 are pre- 

 sented in the condensed form of tables. 



Quantitatively, the San Marcos gi'oup showed (table 23) constantly in all 

 of the ten determinations the least amount of pigment, but a considerable varia- 

 tion range, while the Tierra Blanca group gave constantly the largest amounts 

 of pigment, and also considerable range of variation ; the El Hule determinations 

 gave the greatest variability and the least constancy in values. In all it is 

 clear that the members of the population responded as a mass to the conditions 

 of life, in that the polygon moves as a whole plus and minus in different deter- 

 minations. From quantitative determination the conclusion might be drawn 

 that quantity varied determinately in this character, but it does not give any 

 basis for a discussion of its action in evolution until it is shown whether the 

 variation is permanent or can be made so. Statistics can not determine this, 

 but they might give the basis upon which to base experimental operations. The 

 variation differences found in these locations may be due to different external 

 conditions, to localized racial characters resulting from partial isolation, and 



