HOWAED, HOWAED AND ABDUE EAHMAN 207 



These results of course do not prove that pure line selection in this crop 

 will never, under any circumstances, lead to improvement. They indicate 

 however that the method will bs attended with very considerable difficulties. 

 Continuous mass selection by the elimination of unthrifty plants vnW probably 

 give the best results. 



3. Java Indigo. 



The details relating to the flowering, pollination and to the occurrence 

 of natural cross-fertilization in Java indigo {Indigofera arrecta Hochst.) were 

 described and figured in 1915 in Bulletin No. 51 of the Agricultural Research 

 Institute, Pusa. The bearing of these facts on the improvement of 

 this crop was dealt with in considerable detail in Bulletin No. 67 published 

 in 1916. Very little self-fertilization takes place in the case of protected 

 flowers, the visits of bees being necessary for seed formation. The crop is 

 composed of a mass of heterozygotes, differing widely in habit, time of flowering, 

 root range and amount cf leaf surface. There is evidence for believing that 

 self -sterility occurs to some extent as there is a great falling off in vigour obser- 

 vable in the plants raised from self -fertilized seed. These circumstances and 

 the ease with which the land becomes contaminated with the seed of 

 previous cultures will render the improvement of Java indigo by the methods of 

 pure line selection a time-consuming and very difiicult undertaking. As is well 

 known, this crop produces hard seeds in abundance and at harvest time there 

 is a considerable amount of loss through the explosion of the pods. These 

 hard seeds germinate naturally a few at a time during the second half of the 

 monsoon and readily contaminate any culture sown on the same land unless 

 special precautions are taken. 



4. Sumatrana Indigo. 



■ The structure of the flower and the arrangements for pollination in the 

 case of Sumatrana indigo {Indigofera Sumatrana Gaertn.) clcsely follow 

 those of Java indigo. The anthers burst in the bud just before the flower opens 

 in the early morning. The flowers are visited during the morning by bees 

 {Apis florea and A. indica), and the well known phenomencn occurs cf the 

 upwards discharge cf pollen brought about by the explcsion of the flower 

 (Plate III). On November 11th, 1915, observations beginning at 7-15 a.m. 

 were made en two branches to determine the number of flowers which opened 

 and which were sprung during the course of a day. By 9 a.m., 13 flowers 

 were open of which 7 were sprung. At 5 p.m., the same day, a total of 32 



2 



