48 NATURAL CONTROL OF THE COTTON BOLL WEEVIL. 



shaded, while it was very dry and had cracked deeply. The tempera- 

 ture recortls at Corpus Christ i (Table ^'II) show far less range in 

 temperature than is found at most of the other locahties, wliich are 

 all situated farther from the coast. The mean maximum tempera- 

 ture, therefore, appears to be relative h' low, although the mean 

 average temperature is 1.3 degrees above the normal for that locality. 

 The weather records for May show that the temperature averages 1.3 

 degrees above normal, while during that month the rainfall was 1.78 

 inches below normal. June records show that the temperature 

 averaged 1.7 degrees above normal, while rainfall was 1.63 inches 

 below normal. July was also warm and dry, so that in this case the 

 40 per cent of mortality ascribed to heat or drying is very evidently 

 the result of the extreme drought which had prevailed during the 

 preceding eight or ten weeks. The high total mortality at Corpus 

 Christ i appears, therefore, to be largely a direct result of the extreme 

 drought in a ver}' open field, together ^\'ith the active assistance of the 

 ants. 



At Hallettsville during August occurred the highest mean maximum 

 temperatures found in am' locality' except Kerrville, although the 

 mean average temperature was only 1.1 degrees above normal. The 

 rainfall recorded was 1.54 inches below normal for the month. The 

 full eifect of tliis drought was probably not realized because of the fact 

 that during the month there were traces of rain upon eleven days, 

 when the precipitation was not sufficient to be measured. The cotton 

 was d^-ing in places and green in patches, so that forms must have been 

 largely exposed, and the high temperature and drought should have 

 produced. an even higher mortality than the 31.7 per cent found for 

 this factor. The effect is probably also modified by the fact that dur- 

 ing July nearly the normal amount of rain had fallen, so that the ground 

 during August was not nearly as dry as it was at Corpus Christi during 

 July. At Hallettsville, also, ants were responsible for more than the 

 average percentage of mortality. An explanation of the reasons for 

 the great difference shown in the effect of heat, especially between 

 Hallettsville and Kerrville, may be given at this point. A direct 

 comparison of the Weather Bureau records will show at a glance how 

 the influence of one important factor may be essentially aflected by 

 other related factors, and in this case it will afi'ord an ample explana- 

 tion of the mortality from heat, amounting to 31.7 per cent at 

 Hallettsville, with a mean maximum temperature of 95 degrees, and 

 only 3 per cent at Kerrville, with a similar temperature of 95.3 degrees. 



