6 BULLETIN 778, U..S. DEPARTMENT OF AGRICULTURE. 
All of the foregoing experiments were conducted in a commercial 
greenhouse, and each test represents a 50-foot bed of roses. For the 
most part the killing results of all six experiments were satisfactory. 
Unfortunately, however, the molasses served as a medium for the 
development of sooty: mold, and, moreover, where lead arsenate was 
used* an objectionable white deposit developed. The presence of 
either on cut flowers necessarily would reduce their value and in 
some instances would eliminate them from the market. As the rose 
midge usually is present at a season when the grower can ill afford 
to have objectionable deposits on his cut flowers, it is evident that 
any of the foregoing combinations will be unsatisfactory, unless some 
method of counteracting these objectionable features is developed. 
In addition to the foregoing, 1 part of nicotine sulphate (40 per 
cent nicotine) to 400 parts of water, with the addition of enough soap 
to produce suds, was tested. The results of this experiment were very 
unsatisfactory, fully 95 per cent of the larve being uninjured. 
To determine the value of tobacco dust in preventing the full- 
grown larve, or grubs, from entering the soil, the following cage 
experiments were conducted: (1) Soil around caged plants covered 
with dry tobacco dust; (2) same as former except that the dust was 
wet. Full-grown larve which were placed in the cage containing 
dry dust were active for 24 hours, but did not go below the surface, 
whereas the larve similarly placed in cage 2 were exceedingly active 
upon coming in contact with the wet dust, acting as if they were 
burned, and after from 5 to 8 hours they were all dead. All larve 
used in the check immediately entered the soil. 
Having determined a satisfactory method of preventing the 
entrance of the full-grown larve into the soil, all of the rose beds 
in the infested houses at Colgate, Md., were covered on October 12, 
1916, with tobacco dust averaging from one-fourth to one-half inch 
deep. To prevent the larve from entering the dirt walks of the 
houses, all walks were sprayed with 5 per cent kerosene emulsion. 
Simultaneously nightly fumigation with tobacco stems was inaugu- 
rated and continued until October 30, inclusive, and from that date 
until November 8 the houses were fumigated every other night. The 
object of this fumigation was to kill all adults before eggs were 
deposited. 
Although this control work was not undertaken until October 12, 
its effectiveness Was soon apparent, and by the latter part of October» 
it was very difficult to locate an infested bud. Not only was the 
midge under control, but the owner was enabled to bring on his fall 
crop earlier than was the case in 1915. On May 7, 1917, these houses 
were carefully examined, and only 6 larvee were located, 2 in the 
buds of the Hadley and 4 in the buds of the Radiance. All plants 
at this time were in excellent condition and gave promise of pro- 
