486 JOURNAL OF ECONOMIC ENTOMOLOGY [Vol. 5 



These were simpty covered with cheese cloth through which the wind 

 could blow. Two boxes were placed in each of three windows, and 

 twenty caterpillars, were placed in each with an abundance of red oak 

 leaves. To the windward of one box a little bag of material was hung 

 prepared according to Reiff's receipt. The cheese cloth lining the 

 windward side of another box was smeared with the juices of disin- 

 tegrated caterpillars. The leaves in two boxes were smeared with the 

 same juices and the two remaining boxes were kept as controls. 



Twenty-one caterpillars died of the disease in the two boxes contain- 

 ing leaves smeared with the virus and one died in each of the control 

 boxes. Two died in the box to the windward of which the bag was 

 hung and three in the box lined with the smeared cheese cloth. In the 

 experiment with the bag one died the day after the experiment was 

 begun, so in all probability it had already been infected before being 

 placed in the box, for one day is entirely too short an incubation period. 

 In this same box four died of tachinid parasitism; all the rest trans- 

 formed. In the controls three died of tachinid parasitism and the rest 

 transformed. In the boxes in which the leaves were smeared four died 

 of tachinid parasitism and the rest transformed. 



While these experiments are not very extensive they, nevertheless, 

 indicate that the wind is an unimportant factor in the rapid spread 

 of the disease and that infection in nature occurs when caterpillars 

 feed on leaves soiled by the juices of dead individuals. During the 

 height of the season we found leaves everywhere with brown spots on 

 them and, on making smears of such spots with some sterile water, 

 polyhedral bodies were found in great abundance. 



A very striking phenomenon observed in all of our experiments was 

 the large number of moths obtained from caterpillars infected several 

 times, both with the Berkefeld filtrate and with the unfiltered virus. 

 This seems to us to be very suggestive of immunity and agrees with 

 observations in the field where, in a given locality, the disease raged 

 for several weeks and yet moths were later seen in abundance. They 

 may have been fortunate enough to escape infection, but this seems 

 improbable in a heavily infested area and was entirely impossible in 

 the laboratory where thej^ were infected several times. 



SUMMAEY 



Our experiments and observations may be summ^ed up as follows : 



1. The presence of polyhedral bodies in the blood corpuscles may be 

 useful in diagnosing the health of nun moth caterpillars, but this test 

 cannot be used for gipsy moth caterpillars with any degree of certainty. 



2. The virus of the wilt disease is filterable with difficulty. 



3. Such a filtrate is free from bacteria and polyhedral bodies. 



