122 



lua Vallej, at about 200 feet elevation, January 14, 1907. A 

 large number of small pupae was in a mass together with the 

 remains of some larva upon which they had fed as parasites. 

 There was no means of determining what this host larva was ; 

 but it was supposed to be either an Odynerus or a Pison larva, 

 as there were Pison cocoons and remains of Odynerus nests 

 in the pieces of rotten wood (Breadfruit tree) in which this 

 mass of parasitic pupae was found. There were also cocoons 

 of what appeared to be a Megachile; also a beetle krva which 

 on rearing proved to be Ceresium simplex^ and also among the 

 numerous burrows was "frass" of some wood-boring Tineid 

 larva, all of which made it the more difficult to determine the 

 true host with certainty. 



The following week these pupae transformed to the adult 

 parasites.* They were nearly all females, but one male was 

 observed. After remaining together in a tube for a few days 

 after maturing, 12 females were removed to a tube containing 

 several larvae of the mud-dauber wasp {Soeliphron caemeiv- 

 tai-ius) ; several larvae of this wasp were also put in the tube 

 with the remaining parasites. Many females were observed 

 in the act of ovipositing in these larvae. In a week's time, 

 small larvae of the parasite were observed very numerous feed- 

 ing externally on the larvae of the wasp, and in about another 

 week they entirely consumed their hosts and began to 

 pupate. The first ones matured in 21 days from the 

 time eggs were first laid, A count made of the number of 

 parasites per host gave 448 for one (of which only 8 were 

 males), and 324 for another. In one instance a larva within 

 a cocoon was parasitized. ISTo opening could be found by which 

 an adult parasite could have entered. The eggs must have been 

 laid by piercing the cocoon with the ovipositor, in which case 

 perhaps the eggs were not deposited within the host, but on 

 the outside where they normally feed after hatching. Perhaps 

 this would be the normal method of laying eggs. 



After starting the above experiments in breeding this para- 

 site, some larvae of Odynerus nigripennds and Pison hospes 

 were obtained, and parasites admitted to them. They bred 

 upon these the same as upon the former host. In the case of 

 the Pison larvae, one without a cocoon remained unparasitized, 

 while two within cocoons were parasitized. 



* Being referred to Dr. Perkins for identification iie has described it as 

 a new species. 



