308 DR. L. H. GOUGH. 



Part of the data collected in this experiment has been utilised in a previous 

 chapter, when studying the damage done to the crop. This material has in addition 

 been studied for the progressive and comparative frequence of attack on buds, 

 flowers and bolls during the season. It will be found tabulated in Table XVIII. 



It has been found that buds and flowers are not normally atacked, until or 

 unless the supply of bolls has become insufficient for the needs of the Gelechia 

 population. This occurred at Gemmaiza in 1917, chiefly from September onwards. 

 It commenced (if we ignore a few sporadic records) at the time when the pink 

 bollworm had nearly reached its highest numbers and continued when, owing to 

 decrease in the number of bolls set, the highest percentage of attacked bolls was 

 registered. It is worthy of remark that it was during the same period that the 

 most frequent and significant figures for multiple infestation of bolls were obtained. 

 (Under multiple infestation I here refer to records showing the observation both 

 of larvae and of damage done during the previous occupation by larvae that had 

 left the boll—" traces.") 



In point of selection of food, bolls are greatly preferred to buds or flowers. 

 Larvae recorded from flowers cannot mature in a single flower, and probably have 

 previously fed in a bud, or will have to wander to another flower, bud or boll for food. 

 The greatest interest however centres in the records for the green bolls. It 

 must be remembered that the bolls figuring in each week's records may be from less 

 than one to as much as eight or nine weeks old. Further, in the table under 

 consideration " attacked " bolls denote bolls in which living worms were actually 

 found. No attempt was made in 1917 either to judge the age of the bolls, or to count 

 the number of larvae per boll. It is, however, a matter of general knowledge that 

 an attacked boll frequently holds more than one larva, and that the larvae present 

 may be of different ages, derived from different layings. Eggs being laid in groups of 

 about six, one would a priori expect to find more than one larva in each attacked boll. 

 Attention was however given to recording the traces of previous occupation of a boll 

 by larvae which had left it, and in many cases it was found that a boll containing 

 larvae had previously been attacked by others. Traces of previous occupation begin 

 to be recorded about one week after the first larvae had been seen, and there is 

 a very strong correlation between the attack of one week and the " traces " of the 

 next week, especially for the period of eight weeks commencing 7th July and ending 

 31st August and commencing 14th July and ending 7th September. This period 

 is se'ected for remark, as after that date part of the bolls showing " traces " would 

 have ripened and no longer be recorded. (The correlation was found to be 0-968 

 ± - 015.) For the whole period over which our observations extend, the attacked 

 bolls are about double the number of those showing " traces," doubtless on account 

 of part of the latter class disappearing unrecorded on the ripening of the bolls. 



It further seems fairly reasonable to infer from the table that the infestation 

 of the bolls takes place principally during the last two or three weeks before ripening. 

 Should the attack on the average take place much before that time one would expect 

 to find a much greater number of bolls showing traces of attack than is actually found. 

 Late in the season very many of the bolls contain living larvae during the last 

 week before ripening, many of these persisting unchanged as resting larvae in the 

 seeds of the ripe bolls or of the picked and stored seed-cotton. 



