October, '17] DUNN: cocoanut-tree caterpillar 485 



cially during a period of heavy rainfall or when they are exposed to 

 much dampness. The fall brood seems to suffer a heavier loss than 

 the spring brood, which may be accounted for by the fall brood be- 

 coming mature during the middle of the rainy season of the year and 

 there is a much heavier rainfall than when the spring brood are in 

 evidence. 



Many are found dead in their nests of this disease and are generally 

 found either in the bottom of the nests, or hanging from the sides 

 suspended by the second or third pair of prolegs and the anterior and 

 posterior ends of the body hanging downward in the shape of a horse- 

 shoe. Nearly all of the dead caterpillars become heavily coated with 

 this fungus a short time after death, while the chrysalids become very 

 hard on the inside from the heavy growth of the long filaments of the 

 fungi. 



Observations were made on one lot of about 1,000 caterpillars that 

 were collected and placed in a small screened house during the fall 

 season. This house is about 6 feet long by 3 feet wide, and 6| feet 

 high. Three sides are made of wire screening and the fourth of sheet 

 iron. The roof is made of corrugated iron with screening beneath, 

 which serves to keep the rain out. There is free ventilation in this 

 house and it is comparatively dry with the exception of the floor which 

 is of dirt and becomes quite damp during heavy rains. 



Of this lot of caterpillars that were under observation over 65 per 

 cent died as a result of this fungus invasion without changing to 

 chrysaHds, and about 50 per cent of those that did transform into 

 chrysalids died in the chrysalis stage. This is a much higher rate of 

 mortality than is likely to be found in a spring brood. It was noted 

 that nearly all of the chrysalids killed by this fungus that were dis- 

 sected and examined and found to contain females had died before 

 egg formation had begun to take place, and as the eggs can generally 

 be observed in a female after the sixth or seventh day, it would appear 

 that the chrysalids die within the first few days of the chrysalis period. 

 A few were found that contained eggs and these had evidently died 

 later. 



It is safe to believe that quite a large proportion of each season's 

 brood are destroyed by their natural enemies. 



Laboratory Observations of the Parasitism 



In order to discover whether the parasites infest B. isthmia in the 

 caterpillar or chrysalis stage considerable observations were carried 

 on at the Laboratory. Shultz"^ in speaking of the parasites says, "I 



^Shultz, Henry F., Proceedings of the Entomological Society of Washington, vol. 

 X, No. 1-2, March-June, 1908, p. 166. 



