50 NATURAL CONTROL OF THE COTTON BOLL WEEVIL. 



their work. During July the rainfall amounted to 8.36 inches, anij)ly 

 justifying the remark, ^'Much rain for past five weeks," and fully 

 explaining the low mortality from heat or dr^^ing. 



The remaining localities given in Group B and those which have 

 been included in Group C show a continual decrease in total mortality. 

 The decrease is more evident from the effects of heat or drying and 

 from ant attack than it is from the attack of parasites. Thus in 

 Group B the average weighted percentage of mortality from the para- 

 site attack is 3.8, while in Group C it is 3.3 per cent. From ant work 

 in Group B there is a weighted average mortality of 26 per cent, while 

 in Group C it averages 8.7. Mortality from heat or dr^'ing in Group B 

 is 13 per cent, while in Group C it is but 7.3 per cent. It is evident 

 that the extent of ant and parasite attack will be directly influenced 

 more by the abundance of the species concerned than by the climatic 

 conditions under which they may work. It is impossible to determine 

 whether ants or parasites are particularly abundant* in any locality 

 except by determining the proportion of weevil stages which they 

 have destroyed, but the primary actual cause of their abundance may 

 de})end upon very different conditions. 



It is possible that the ants might be quite abundant and still attack 

 the weevil stages ])ut little if some other food su])ply were more easily 

 available. It is manifestly more difficult to determine why mortalit}' 

 is low in a field than it is to determine what are tlie effective factors 

 when it is liigh. From a comparison of Grou])s A and C in Table VIII 

 it would appear that in Group A the predominant ty})e of soil is sandy, 

 while in Group C the soil is heavier and probably would tend to give 

 a ranker growth of weed. This conclusion seems to be borne out by 

 the remarks given for Group C, which indicate a heavy growth and a 

 large degree of shade. This condition would decrease the efficiency 

 of sunshine, and it is c^uite lilvely also that the soil conditions in these 

 fields were not as favorable to the ants as in the more sandy locations. 



It has been generally considered that the weevil has done greater 

 damage in river bottoms than in upland fields. This has been held 

 to be due to the greater moisture of the ])ottom lands and the ranker 

 growth of the plants grown thereon. From the considerations which 

 have been mentioned it would appear that this idea is well founded, 

 but that the difference is due in a j^ossibh^ equal degree to the absence 

 of ants in the river bottoms. If this be true, then the importance of 

 strictly cultural methods in controlling the weevils in the bottom lands 

 may be even greater than it is on uplands, since as a general rule the 

 factors of natural control may be less effective in such locations. It 

 is evident, however, that there is no defmite line of division between 

 the localities in which natural control will be most effective and those 

 in which it will ])e least so. Onl}" by continuing observations for 

 several years will it be possible to determine at all definitely the class 



