492 PROCEEDINGS OV THE THIRD ENTOMOLOGICAL MEETING 



may be considered to be connected with the greater purity of Affifi 

 as compared to Mitaffifi. 



Finally, the deviations observed nowhere reach three times the 

 standard deviation and may not have any real significance. 



In the next Table are recorded the percentages of lint for various 

 varieties for the period 1911-1917 which just covers the period during 

 which GelecJiia has risen from insignificance to importance. It can 

 here again be pointed out that none of the fluctuations shown is suffi- 

 ciently large to be significant. 



Finally, comparing the standard deviations for the annual varia- 

 tions in the percentage lint with the standard deviation found for the 

 samples composing the crop of a single field shown on Table XIII (com- 

 parison of sound and attacked samples) it will be found that the standard 

 deviation of the annual variation is nearly the same as the standard 

 deviation of the variation of the percentage hnt of various parts of a 

 single crop. 



The damage done to the crop can be calculated consequently from 

 the seed alone, or from seed cotton, or from the lint alone. Provided 

 that the sampling in the field is properly done, and tliis has through- 

 out been found to be the most difficult part of the work to get done 

 correctly, the rest is simj)le. It is necessary that a really representa- 

 tive sample be taken, boll by boll, each boll being wrapped separately 

 in paper. For this purpose all the ripe bolls of a large number of trees 

 are required. In the laboratory the bolls are picked apart and examined 

 for damaged seeds. The sound bolls are counted and then ginned 

 together, the damaged ones likewise. Before ginning, if the two jjiles 

 of seed cotton are weighed, we already obtain material for the estima- 

 tion of the damage. The loss can be calculated by finding the total 

 weight of all the seed cotton, and subtracting this from the weight 

 found by multiplying the average weight of one sound boll by the total 

 number of bolls used, sound or attacked. 



The loss itself is however composed of various items, which are 

 briefly : — 



(1) Reduced weight of soimd seeds developed in attacked bolls 



and consequent proportionate loss of seed and lint weight. 



(2) Lost substance of attacked seeds, and proportionate loss of 



Unt. 



(3) Total loss of some of the seeds in attacked bolls, and total loss 



of the lint they ought to produce. Material in which the 

 bolls have been carefully separated as above described, 

 from which all the seed has been carefuly counted and 

 weighed, can be tested for the proportion of loss due to each 



