E. J. BUTLER 31 



quite possible to remove all but broken fragments and plough these in early 

 so that they have time to decompose, as in Experiment X. 



As the disease does not occur naturally in the neighbourhood of Pusa, 

 only small plot experiments under very complete control have been practicable 

 there and no field trials on a large scale could be made. Within the infected 

 tract, field experiments and demonstrations have been hampered for want of 

 trained staff ; still something has been done to encourage clean cultivation 

 amongst the ryots and, incidentally, certain difficulties in particular classes of 

 land have been revealed. 



The j&rst field experiments * were started at Begumganj in Noakhali 

 District in 1912. Three plots were selected, in one of which the stubble was 

 burned on March 10th, 1912, in another it was burned on the same day and 

 lime added at the rate of 30 maunds per acre a few days later, and in the third 

 liming alone was tried, the stubble having been already ploughed in. The 

 fields were then ploughed and the usual mixed crop of broadcast aus and 

 aman (bajal) was sown in the limed plots while unmixed aman was broadcasted 

 in the other. Early in August I visited the plots and found the aus ripe and 

 perfectly healthy. Ufra first appeared in the unmixed aman in October, 

 nearly a month after it was virulent in the surrounding fields. The latter were 

 totally destroyed while the experimental plot gave a moderate yield. No 

 attempt to check infection from the adjoining fields was practicable and the 

 indications pointed to this as the source of the disease. 



In 1913 a further experiment was made at Begumganj, the stubble being 

 burned in two duplicate 1-acre plots of infected land after harvest in 

 December, 1912, and then well ploughed and harrowed. Each plot was 

 divided into 4 equal plots at sowing time (March 1st, 1913). One plot was 

 sown with broadcasted aman, a second with the usual broadcasted aus and 

 aman (bajal) mixture, a third with broadcasted aus which was followed by 

 transplanted aman, and the fourth with jute similarly followed by transplanted 

 aman. The three last plots escaped ufra, while there was some damage to the 

 first. The yields were at the rate of 14 maunds per acre in the first plot 

 (taking the mean of the two duplicates), 25 (13 aus + 12 aman) in the second, 

 34 (14 aus +20 aman) in the third, and 12 maunds jute with 10 maunds rice 

 in the fourth. Hence the damage cannot have been very great even in the 

 attacked plot, of which only about one-tenth of the area was affected. 



* All the field experiments in this section were arranged by the Bengal Department 

 of Agriculture, to test the recommendations made by the writer in his previous paper and 

 from time to time since. They were carried out under the supervision of Mr. G- P- Hector, 

 Economic Botanist of that Department, in consultation with the writer- 



