EFFECTS PRODUCED BY PINK BOLLWORM ON THE YIELD OF COTTON SEED. 305 



Provided that the samples at one's disposal are sufficiently large and sufficiently 

 accurately taken, the loss can be estimated in the following manner. 



A large number of sound bolls must be taken from the field for which the loss is 

 being estimated. The number of sound bolls being known, their total weight divided 

 by their number gives an average weight for sound bolls. The total yield in sound 

 and attacked bolls of a sufficient number of plants is further required to be known, 

 both as regards number of bolls and weight. This gives the average crop. The 

 theoretical true crop for that number of trees can be calculated by multiplying 

 the number of bolls found by the average weight of sound bolls. The difference of 

 the weight thus found from the weight actually observed is the probable actual loss. 



If the statements made in sections A, C and D are correct, the probable actual 

 loss as found above must also equal the sum of the following partial losses. 



(1) From number of seeds developed in sound bolls divided by the number of 



sound bolls is subtracted the total number of seeds developed in all 

 bolls of the crop divided by the total number of bolls forming the crop. 

 The remainder is multiplied by the number of bolls forming the total 

 crop. This gives the number of seeds which should have been produced, 

 but were not. Their average probable weight can be estimated by 

 multiplying them by the total weight of the seeds from the sound bolls 

 divided by the number of seeds found in the sound bolls. 



(2) From the weight of the seeds found in the sound bolls divided by the number 



of seeds set in the sound bolls is subtracted the weight of the sound seeds 



set in all the bolls forming the crop, divided by the number of sound 



seeds found in those bolls. The remainder multiplied by the number 



of sound seeds found in the entire crop gives the loss due to reduction 



of average weight of the sound seeds in consequence of Gelechia attack. 



(3) The weight of the attacked and double seeds is subtracted from the average 



weight of the same number of sound seeds from the sound bolls. This 



gives the probable actual loss due to seed substance being destroyed by 



the insect. 



The result obtained by adding to the actual weight of crop observed (1) the probable 



weight of the seeds which did not develop,(2) the weight calculated as the reduction 



in weight of sound seeds in consequence of Gelechia attack, and (3) the actual loss 



due to seed substance having been destroyed by the larvae, should give the same 



sum as was obtained by multiplying the entire number of bolls observed by the 



average observed weight of the sound bolls from the same sample. It is to be shown 



now, that where the samples are of sufficient size to be reliable these conditions are 



very nearly realised, thus proving the triple nature of the damage. It is for this 



purpose immaterial whether the weights used throughout are those of the seed 



alone, or of the seed plus lint, confirming what we have said about the small degree 



to which the percentage of lint is affected. 



In Table XVII the calculations sketched above have been made, and the results 

 can be briefly summarised as follows :— 



The average total weight found for 100 sound bolls in any one set of the samples 

 under examination very closely agrees with the average total weight found for 100 

 (C541) c 



