486 rROCEEDTXGS OF THE TIIIltD E.VTOMOT.OGICAL MEETING 



An examination made at Tel el Kebir tliis year has brought out 

 that after the last picking there are approximately 4 times as many 

 bolls on the trees, as on the ground and about 1 boll per tree ; as each 

 boll may contain up to six or seven worms, it is obvious that any work 

 done in destroying bolls must be of use. 



The damage done by the Pink Bollworm has been studied very 

 carefully during the last three years. A prehminary paper on this 

 subject was pubhshed early in 1916, since when an enormous mass of 

 data has been obtained. It is intended to publish these data elsewhere, 

 in extenso, but the main results may be summarized here. 



It was already indicated in 1916 that the average weight of sound 

 seeds that develop in attacked bolls is lowered on account of the attack. 

 It has now been shown that the average weight decreases in proportion 

 as the average intensity of the attack increases. The correlation between 

 the percentage of sound seed in a sample and the average weight of a 

 sound seed in that sample is as high as 0-976 ±0-008. Higher correla- 

 tion could not be expected. (Plate 87.) 



This falling in weight of sound seeds developing in attacked bolls 

 has further been shown to be due not to a general reduction in weight 

 of all the seeds but to an increase in the percentage of the smaller weight 

 seeds at the «ost of the heavier grades. This has been worked out by 

 weighing the seeds from a long series of bolls, each seed separately 



This year we have obtained the further indication that the sound 

 seeds developed in sound locks of attacked bolls are inferior in weight 

 to sound seeds developed in perfectly sound bolls. 



As in 1916 we had stated that there were strong indications that 

 seed developing in attacked bolls loses some of its power of germina- 

 tion, this problem was followed up to see how far the loss of germina- 

 tion is proportional to intensity of attack. (Plate 88.) 



The correlation between the proportion of sound seed in a sample 

 and the germination of that sound seed has been found to be 0-925 

 ^0-011, which again is very liigh. 



This problem has been pursued further, and by germination of seeds 

 carefully weighed singly and grouped in weight groups at 10 milligram 

 intervals, a correlation of 0-958 ^^0-014 has been found between weight 

 of seed and intensity of germination. 



It has further been found that attacked bolls do not form so many 

 seeds (sound -plus attacked) as sound bolls. The reason for this is that, 

 where the attack takes place early in the life of a boll, entire seeds are 

 liable to be totally destroyed without leaving traces. The correlation 

 between intensity of attack and reduction in number of seeds was found 

 to be 0-889±0-063. 



