502 STUDIES IN TH£ POLLINATION ON INDIAN CROPS 



flowers in large numbers. If, however, the weather is damp, the pollen masses 

 decompose into a gelatinous substance, fertilizaticu does not take place and 

 the flowers fall. 



Fertilization. This crop is characterized by the great extent of its flower- 

 ing period and by the enormous number of flowers it produces. All are self- 

 pollinated before the flowers open but the great majority fall without setting 

 any seed. The expenditure of energy in flower formation is extraordinary. 

 Setting depends on the humidity of the air. Dry bright days favour fertili- 

 zation while dull damp weather causes the pollen grains to disintegrate and 

 the flowers to fall without forming seed. Under dry bright conditions setting 

 takes place freely under nets when insect visitors are excluded, thereby proving 

 that self-fertilization readily takes place. The fact that the flowers are cons- 

 tantly visited by bees suggests that cross-fertilization is also possible. 



Natural cross-fertilization. A considerable amount of work has been 

 carried out at Pusa on the occurrence of natural cross-fertilization in this crop. 

 In 1909, samples of arTiar seed were collected from Bihar, the United Provinces, 

 the Central Provinces and Bombay. The following year, 68 single plants 

 were selected and the seed of each was sown separately. Of these, 61 gave 

 rise to obviously mixed offspring while 7 (belonging to Bihar) appeared to 

 breed true. In 1911, a further selection of 72 single plants was made of which 

 26 were obtained from cultivators' fields in the neighbourhood of Pusa. These 

 were sown separately and definite splitting was observed in 65 of these cultures 

 in the following respects — habit, time of flowering, colour of flowers (pale yellow, 

 yellow, deep yellow, orange yellow, orange, reddish orange and all kinds of 

 intermediate tints), colour and shape of the pods and seeds. In 1914, another 

 set of 41 difierent plants was selected on the Dholi Estate. Twenty-three of 

 these cultures did not breed true. The above results prove that natural 

 crossing is common in the pigeon pea. As the self -pollinated flowers are 

 largely visited by bees, it is probable that these insects are mainly involved. 



To determine the precise extent of natural crossing, some observations 

 were made in 1912-13 on a pure culture of arhar (characterized by pale yellow 

 flowers and white seed) grown on a plot about half an acre in size next to a 

 similar area of the ordinary local ciop. The colour of the flowejs and seeds 

 of this pure culture made the detection of heterozygotes an easy matter. 

 One hundred and forty single plants which were true to flower and seed colour 

 were selected and sown separately the following year. Natural crossing was 

 found to have occurred in 91 of the cultures (40 as regards flower colour and 

 41 as regards seed colour), that is, in 65 per cent. An exact count of the number 



