102 Journal of the Asiatic Society of Bengal. [March, 1910. 



stated on similar grounds that in Madras at least some natural 

 crossing occurs. 



All this evidence is opposed to Gammie's view, based on 

 observation at Poona, that cotton does not get cross-fertilised 

 in nature (Memoirs Dipt. Agric. ii, no. 2). Certainly Gammie's 

 view cannot be made a generalisation for India as a whole. 



In this connection it is now interesting to record, as an 

 apparently regular insect- visitor to cotton flowers in parts of 

 Berar, a bee, which by reason of its large hairy body is admir- 

 ably adapted for carrying pollen from flower to flower. This 

 insect is Megachile albifrons, Smith. It is, however, not very 

 active in moving from flower to flower. I first observed the bee 

 on the afternoon of September 22nd, 1909, sheltering in cotton 

 flowers during rain at Pingli, which is a village lying just under 

 the hills in the north-east corner of the Buldana district. By 

 a search through a field four more individuals were found. On 

 the next day between Pingli and Wasali, during the morning, 

 nine insects were seen , and in the afternoon at Wasali eight more. 



No observations were made on September 24th owing to 

 rain ; and on the 25th between showers no bees were seen on 

 the cotton flowers during a march from Jamod to Jalgaon ; 

 but on September 26th, four miles south of Jalgaon, two more 

 individuals were seen on cotton flowers. This makes a total of 

 24 individuals seen on cotton flowers in six days. On the 26th 

 I left the district. 



The common cotton of north-eastern Buldana is the yellow 

 flowered Jari (Gossypium neglectum, Tod., var. vera, Gammie). 

 Grown mixed with it, is a considerable quantity of white flowered 

 Jari (G. neglectum, Tod., var., rosea, Gammie) and a not incon- 

 siderable amount of Dharwar American cotton (G. hirsutum, 

 Mill.). No individuals of Megachile albifrons were seen on G. 

 neglectum, var. rosea, and only one on G. hirsutum: all the 

 others were on G. neglectum, var. vera. 



From the insects' point of view these three cottons may 

 be very unlike; for G. neglectum, var. vera, has large yellow 

 flowers which nod as a rule and open rather widely ; G. neglec- 

 tum, var. rosea, has smaller rose- white flowers which generally 

 ascend slightly from the horizontal and open but little ; G. 

 hirsutum has upright pale yellow flowers which open earlier 

 than the other two, and widely. 



Bani (G. indicum) has flowers exactly as G. neglectum, var. 

 vera, but not a single plant was noticed in north-eastern Bul- 

 dana, where the Megachile was seen, whereby to prove (as 

 might well be the case) that the bee would go from G. neglectum, 

 var. vera, to it. 



G. neglectum, var. vera, — Jari, — in north-eastern Buldana 

 spontaneously self-pollinates within an hour or two of opening ; 

 and the flowers are fertile to their own pollen : so that the bolls 

 set whether insects go to the flowers or not. 



