The Bionomics of South African Insects. 449 



at the beginning of the wet season is not sharp ; indeed, 

 he has a distinct recollection of seeing them flying together 

 at that season fairly frequently. " I believe," he writes 

 (1902), "that in some seasons one might take sesamtts in 

 every month of the year. Certainly at Gadzima, in 1895, 

 the true winter broods of sesamns lasted right up to the end 

 of December. In a dry spring, that is when the rains are 

 late in starting, butterHy life appears to be less abundant 

 and the emergence of the wet-season forms seems to be 

 retarded. On such occasions an actual break without 

 specimens might occur in such a comparatively unfavour- 

 able locality as Salisbury. But I believe that this would 

 be an unusual occurrence, and even if it happened in one 

 locality I doubt if it would necessarily take place every- 

 where at the same time ; for examj)le, in the moister parts 

 of the low veldt the succession of the broods would 

 probably continue unbroken. I am quite satisfied that 

 there are at least two or three broods of sesanms during 

 the winter months, that is if the condition of wild specimens 

 can be taken as any criterion. Food is much less plentiful 

 in the winter, bvit it is obtainable in quite sufficient 

 quantities to keep the species going. The change of 

 seasons from wet to dry is of rather a gradual character ; 

 the reverse change is more marked, but this depends a 

 good deal upon the total rainfall of the preceding year. 

 When this has been heavy, the ground retains a certain 

 amount of moisture throughout the winter, so that when 

 the frosts cease and the sun's heat increases in the spring, 

 a large number of the earlier plants spring up and flower 

 before a drop of rain has fallen. But after a succession of 

 dry years this does not take place, and, with possibly a 

 few exceptions, none of the plants come out in response to 

 the heat, but require the rains to bring them out. In 

 this latter case the change in conditions is very strongly 

 marked, much more so than during a wet cycle." 



The discussion of the possible nature of the environ- 

 mental stimulus, if any, is better deferred until after the 

 description and consideration of Mr. Marshall's experiments 

 in the next section. 



TRANS. ENT. SOC. LOND. 1002.— PART 111. (noV.) 30 



