60 NATURAL CONTROL OF THE COTTON BOLL WEEVIL. 



A casual examination of the data given in Table XT will be suffi- 

 cient to show that considerable variation exists in the mortality pro- 

 duced by each factor in the same class of forms in even the most 

 closely associated localities. This seems to prove the influence of 

 local conditions affecting the mortality produced by the different 

 factors. From several considerations this would seem to be an 

 encouraging fact, indicating that it may be possible to change unfa- 

 vorable conditions or to take advantage of those which are most 

 favorable in regard to environment, cultivation, soil, etc. It is 

 noticeable that in all classes of forms, in practically all localities, a 

 large proportion of the mortality from heat occurs during the larval 

 stage. In southern Texas alone does the larval mortality in hanging 

 dried squares fall below that in any other class of forms. In that 

 case three-fourths of all larvae found in fallen squares were killed by 

 heat, while in the hanging squares slightly more than one-half were 

 thus killed. As a general average for the entire area about 53 per 

 •cent of the larvae, 18 per cent of the pupae, and 6 per cent of the 

 adults had been killed *by heat. These percentages are very nearly 

 in the ratio of nine, three, and one. It appears that the nearer the 

 weevil stage approaches maturity the less susceptible is it to the 

 adverse influence of heat or drying. This shows how important it is 

 that the spacing of the plants be such as w T ill submit the largest pos- 

 sible proportion of fallen forms to the direct influence of sunshine. 

 Whatever spacing may have been found advisable where weevils 

 were not present, there can now be no question about the general 

 soundness of the recommendation for increasing the space allowed 

 each plant wherever the weevil is present in abundance. It is true 

 that the efficiency of this factor under similar cultural conditions may 

 vary widely during different seasons, and during different portions of 

 the same season, but in any case the data given w T ould seem suffi- 

 ciently reliable to prove that under nearly all conditions, with the 

 weevil present, w T ide spacing is advantageous. 



It was found impracticable to determine the effect of ant attack 

 upon the various stages of the weevil. It may be stated that, in 

 general, parasites were found to attack the weevil during the last few 

 days of the larval stage. A considerable number of parasitized pupae 

 were found, and in a very few cases the weevil had become adult 

 before death was caused by the parasite. As the parasite causes 

 the death of its host very quickly after beginning its attack, the time 

 required for the egg stage of the parasite should be taken into con- 

 sideration in determining the time of attack, but it has been impos- 

 sible, as yet, to determine the egg stage for any of the parasites of 

 the weevil. 



MORTALITY IN EACH CLASS OF FORMS FOR EACH SECTION. 



Some additional points in the comparison of the general results 

 shown in Table XI may be more conveniently shown by such arrange- 

 ment of the totals and average percentages as is given in Table XII. 



