406 ANNUAL REPORTS OF DEPARTMENT OF AGRICULTURE. 



together for publication. A total of nine years' observation at 

 Tallulali have been accumulated, including both annual comparisons 

 and special studies to determine the relative importance of dijfferent 

 locations as regards weevil shelter. These records are particularly 

 important in connection with the Florida method of weevil control, 

 since this method is largely based on accurate hibernation records. 



The question of the effect of the use of poisons for boll-wtevil 

 control on adjacent bee colonies was investigated in the vicinity of 

 Tallulah. It was found that no important mortality of bees re- 

 sulted in any instance from the application of calcium arsenate to 

 cotton plants. It therefore appears that there is no reason to be- 

 lieve that bee colonies would be seriously affected under the con- 

 ditions existing at Tallulah. However, cotton does not appear to 

 be an important honey plant at Tallulah, and it is possible that 

 different results might be secured in other districts. 



The annual dispersion of the weevil was determined in the fall 

 of 1922 as usual. The summary showed that only 4.01 per cent of 

 the cotton crop of the United States is produced in the territory not 

 yet infested. 



The usual outbreak of the cotton leafworm in 1922 presented 

 several new problems of control, particularly in relation to the 

 insecticides to be used. Some special studies were conducted on this 

 point and are being continued. 



Tobacco insects. — The experiments begun last season in the 

 Burley district in Kentucky upon the control of the tobacco horn- 

 worm met with a most flattering reception. The Burley growers 

 knew but little about modern methods of controlling the hornworm, 

 and they are rapidly adopting the cheaper and more thorough 

 methods advised by the bureau. 



True wireworms cause an annual loss of from $1,000,000 to 

 $2,000,000 to the Burley tobacco growers. In a series of five care- 

 ful experiments the nitrobenzene-flavored bait reduced the infesta- 

 tion by as much as 50 to 60 per cent. This promising line of work 

 will be enlarged next season. 



A poisoned bait using nitrobenzene as an attractant was again 

 tested on sod webworms and found effective. 



The development of mule-drawn dusting machinery has been 

 continued. The tandem-wheel, two-row duster mentioned in the 

 last annual report is being made commercially. It supplies the 

 need of the larger acreage very well, but on account of its cost a 

 lighter and cheaper two-row machine has been devised for smaller 

 acreages and will be tested during the present season. 



An excellent piece of work has been accomplished at the Quincy, 

 Fla., laboratorj^ in fully working out the life-historv stages and 

 seasonal history of the tobacco flea-beetle and in completing a long 

 existent gap in seasonal history of the tobacco budworm. Pre- 

 liminary control Avork was instituted against the budworm in the 

 sun-tobacco district at Tifton, Ga., and a method of control adapted 

 to sun-tobacco conditions was partially worked out. 



Sugar cane, rice, and cactus insects. — ^^Vork has been continued 

 on the rearing and release of Habrohracon hrevicornis, a parasite of 

 the European corn borer which was found to attack the sugar-cane 

 moth borer. The parasite has not yet been recovered in the field. 



