C'ONI'KOI, \\\ I1I:AT, ants, AM) I'AIIASITKS. 25 



a general thiiii: (lie oinori::onre hole made l)y the w (M>vil ( PI. 1 1 1, fi«j:. 1 ) 

 is more regular in outhne and soniewiiat hiiLrei- in area than arc the 

 entranee holes made hy the ants. Upon opcMiing the xinaic caicfnllv 

 it will he found that if a weevil has nMichod lull dcvclopnicnt therein 

 and emerged, there will he found pnvstMit three delinile sii^ns whicii are 

 not })resent in eases whert^ ants ha\'e destroyed the wcexil staire. One 

 of these signs is the delicate white skin which is shed 1)\ a pupa 

 when the weevil transforms from that stag(» to its adult condition. 

 (1^1. in. lig. r><i.) This is nsually so shriveled and twisted as to haxc 

 no reseiuhlance whatever to the outlines of a living pupa and is fre- 

 (piently more or less hidden in the dehris which const it ut(\s the second 

 sign. This is a rather ahundant nniss of {in(» |)articles which hav<» 

 heen torn from the scpuire by the weevil in cutting its emergence hole. 

 These are not eaten hy the weevil l)ut are left within its cell (PI. Ill, 

 figs. 8 and r>h). The third sign of adidt activity consists of a nuniher 

 of particles of white excrement which are almost invariahly deposited 

 hy the adult weevil before its emergence from the s([uare (PI. Ill, 

 tigs. 3 and 5c). These three signs are absent in cases where ants 

 have entered and destroyed the weevil stages (PI. Ill, fig. 4). 



The value of a study based upon statistical data becomes easily 

 apparent to one engaged in a thorough study of such a subject as this. 

 Without it there is a strong ])robability that the general im})ressions 

 formed may not give proper credit to the influence of the various 

 factors. Conclusions whicli are based upon the total efl'ects of the 

 various factors in a large series of observations must necessarily be 

 more reliable. This nuist be especially true of the general average of 

 results fi'om a considerable number of localities and under a consider- 

 able variety of geological and climatic conditions. 



The fu'st division in the data as obtained has been secured by se])a- 

 rating the cotton fruit into several classes, each of which may contain 

 weevil stages which have been exposed to similar conditions. But 

 foiu' of these divisions have been retained in the arrangement of the 

 data shown in Table III. These are, (a) hanging, dried bolls; (6) 

 hanging, dried squares; (c) fallen bolls; (d) fallen squares. .Vs a rule 

 infested squares and small bolls are shed by the })lant in from seven to 

 twelve days after the weevil attack. Many small bolls, especially, 

 are shed normally, even though they may have suflVred no injmy 

 from insect attack. This happens most commonly w- it bin a few days 

 after the withering of the flower and ])efore the young boll has made 

 any growth. It may residt from a faihu*e in fertilization of the flower 

 or the inability of the plant to sustain an excessive load of fruit. 

 With both squares and small bolls, the shedding is accomj)lished 

 naturally, just as is the shedding of leaves in the autunm, by the for- 

 mation of an absciss layer of corky cells which cuts off the fibrovascu- 

 lar bundles through which the sap is supplied, thus destroying the 

 vital connection between the bud or boll and the j)lant branch. The 



