14 INSECT ENEMIES OF THE BOLL WEEVIL. 



The figures of 1902 are based on the total number of stages reared. 

 The data of 1903 are based on the total number of stages reared, but 

 include both stages in hanging and fallen forms. There were 654 

 stages in fallen forms of which 14, or 2.14 per cent, were parasit- 

 ized, and 165 stages in hanging forms of which 45, or 27.27 per cent, 

 were parasitized. The figures of 1905 were separated to show the 

 investigations of March and August because of the great difference 

 in the mortality from parasites in these two months. Between 1906 

 and 1909 the data represent all classes of infested material and all 

 infested regions. It will be noticed that for the last four years the 

 total insect control of the immature weevils has fluctuated between 



15 and 30 per cent. 



2. NATURE AND SOURCES OF THE MATERIAL EXAMINED. 



During the four years 1906-1909 examinations of mortality have 

 been made of material collected at 6 places in Arkansas, 26 in Louisi- 

 ana, 6 in Mississippi, 7 in Oklahoma, and 65 in Texas, making a total 

 of 110 places. These examinations are based upon 94,677 stages, 

 involving an individual examination of over 222,700 cotton forms 

 (squares, blooms, and bolls). Many other collections and exami- 

 nations were made, but because of incomplete records are excluded 

 from the accompanying tables. 



During the four years there has been the equivalent of examina- 

 tions in 176 localities, or an average of 44 localities per year. 



3. SEASONAL STUDIES OF lis ECT CONTROL, BY CLASS OF INFESTED 



MATERIAL. 



Very shortly after the work began in 1906 it became evident that 

 the activity of the parasites and other insect enemies of the weevil 

 was very different in squares and bolls, and in fallen or hanging 

 squares or bolls, and also that the highest control by parasites was 

 in hanging squares. 



An examination of the squares of various varieties of cotton plants 

 will show the observer that certain ones have a transverse attachment 

 of the pedicel to the stem. In all cases where this attachment is 

 perfectly transverse, the square when injured by any insect is caused 

 to drop because of the separation of the infested part from the main 

 stalls: by the growth of an absciss layer at the point of attachment. 

 (PI. I, fig. 1; fig. 3, a.) Certain other varieties indicate a long diag- 

 onal attachment to the stem. When these squares are injured, a 

 diagonal absciss layer is formed which runs down the stem from one- 

 half to three-fourths of an inch or even more. This layer is gen- 

 erally incomplete at the lower point and consequently the square 



