12 



Hitherto we have only been dealing with the percentage of green 

 bolls affected. It is clear that percentages cannot afford any evidence 

 of direct increase or decrease of the pest, unless something is known 

 of the actual numbers of green bolls on the cotton plants at the time 

 of each ooservation. 



For example, it is possible that with a rapidly increasing number 

 of bolls per tree, a fall in the percentage of attacked bolls may mean : 

 (1) a reduction in the number of worms ; (2) no reduction and no 

 increase of the number of worms ; (3) an actual increase in the number 

 of worms. With the number of green bolls stationary, rise or fall of 

 the percentage of worms evidently corresponds to increase or decrease 

 in the actual Gelechia population. Witli the number of bolls per tree 

 falling, decrease of the percentage or stability of the percentage will 

 both mean decrease of the Gelechia population, whilst a rise in the 

 percentage may mean decrease, stability, or increase of Gelechia. 



As no record has been kept this year of the average number of 

 green bolls on cotton plants, it has been necessary to approach the 

 problem of the weekly average number of green bolls per unit (one 

 plant or any other multiple) by an indirect method. 



In connection with the cotton breeding a very large number of 

 flower counts have been made by Mr. Gr. C. Bolland, Botanist to the 

 Ministry of Agriculture. From these flower counts it is possible to 

 obtain figures to show the weekly average number of green bolls per 

 tree. The method adopted in calculation has been as follows. The 

 average number of flowers observed per day in any week was multi- 

 plied by seven to give the weekly average. From this twenty per cent 

 is deducted to allow for "boll- shedding. The figure obtained is supposed 

 to represent the number of bolls set during that week. Allowing that 

 it requires six weeks for a boll to mature, by placing the figures 

 representing each weekly average in the column for the six following 

 weeks one obtains for each week of the growing season six figures 

 which, when added, show the number of green bolls on the tree during 

 each week. As in the material from which my percentages were made 

 no bolls from to 8 days old were used, the number of flowers was 

 not entered for the week immediately following the flowering. The 

 totals used were consequently weekly average numbers of green bolls 

 of one week or older present on the plants. For the last week of 

 September, the curve is based on the total number of green bolls 



