436 Journal of Agricultural Research voi.xvm, No. s 



and after inoculation the insects were placed in sterile boxes in the 

 manner described above. 



Microscopic examination showed that the liquid held quantities of 

 conidia in suspension. Many conidia therefore must have lodged on 

 the insects, though the liquid did not spread in an even pellicle over 

 the body but gathered in little droplets. Control larvae were sprayed 

 with sterile water only and were subsequently removed to sterile boxes. 



On May 5, 10 larvae of Feltia sp. were inoculated. On May 14, i 

 died from an unknown cause; on May 21,1 died of Sorosporella; and 

 by June 16, 8 were alive. In the control experiment of 10 larvae, 2 

 died on May 29 of unknown causes, and the remaining 8 were alive on 

 June 1 6 at the close of the test. 



Seventy-five army worms were inoculated on August 6, and 25 were 

 held as controls. Three of the infected insects died on August 10, but 

 the parasite was not found in them. On August 13, 3 of the inoculated 

 insects, as well as 2 of the control larvae, died from Metarhizium; 18 of 

 the inoculated specimens had died from unknown causes; and 6 of the 

 control larvae were dead. On August 17, the parasite was recovered 

 from 5 larvae. On August 20, 22 of the infected ones emerged; 8 were 

 dead from Sorosporella, and 8 from other causes. Finally on August 23 

 the remaining 7 of the inoculated army worms were dead from unrecog- 

 nized causes, and 17 of the control larvae emerged as adults. 



Several other tests of a similar nature were conducted, but it is not 

 advisable to consider them here. The foregoing examples are sufficient 

 to show what may be expected of this type of inoculation. The average 

 mortality of all the tests, however, was somewhat below 30 per cent; 

 and the experiments as a whole serve to corroborate an opinion already 

 formed by the writer in connection with other entomogenous fungi, to 

 the effect that the spray method is of less value in artificially inoculating 

 insects than the direct contact method. 



The reason for this is not apparent, because hundreds of conidia must 

 have lodged upon the infected insects. Furthermore, the conidia germi- 

 nate freely in water and when sprayed upon insects they were apparently 

 in a suitable position to insure infection. 



FEEDING METHOD 



In order to introduce the fungus into the alimentary tract of cutworms, 

 a conidial agar paste was smeared over clover leaves, which the cater- 

 pillars were allowed to eat. But first the leaves were cut into small 

 portions not larger than 5 square millimeters in order to reduce to a 

 minimum the possibility of contact between the leaves and the external 

 parts of the insect's body. Such leaf fragments were then placed in 

 sterilized Petri dishes, into which were also inserted fresh larv^ae, one to 

 each dish. Many of the larvae were hungry, for they had purposely been 

 unfed for several hours ; and the process of eating was followed in many 



