April, '16] BACK AND PEMBERTON: CERATITIS PARASITISM 309 



and not from the ground. One specimen of D. fullmvayi was reared 

 from larvte collected in coffee during Januaiy, or about one month after 

 the first liberation of this parasite in Kona. The most interesting 

 development, however, in the coffee section seems to be the gradual 

 change taking place in the ratio of control exerted by the 0. humilis 

 and D. iryoni shown by the March, June and September data. This 

 increase in effectiveness of tryoni seems to be taking place without 

 producing an increase in the total parasitism, but at the expense of 

 0. humilis. The percentages of parasitism among larvae developing 

 in kamani nuts (Table II) and other host fruits (Table III) show fewer 

 negative results than those recorded for 1914. D. fullawayi appears 

 to be very efficient in parasitizing larvae in coffee cherries in Honolulu^ 

 and bids fair to outstrip the earlier introduced 0. humilis. From larvae 

 reared from one lot of kamani nuts all four parasites are reared. From 

 the general observations of the writers, it would appear that T. giffardi 

 may prove a most valuable parasite in supplementing the good done 

 by humilis, tryoni, and fullawayi. While all four species of parasites 

 attack only the medium and well grown larvae, humilis, tryoni and 

 fullawayi are most active in parasitizing the mature larva while the 

 host fruits are still attached to the trees. Although Mr. E. M. Ehr- 

 horn has observed one specimen of 0. humilis hovering over fallen 

 Strawberry Guavas (Psidium cattleyanum) , the writers have never ob- 

 served humilis, tryoni or fullawayi attempting to oviposit in larvae 

 within fallen fruits. On the other hand, specimens of T. giffardi have 

 been taken from the channels in the pulp of well decayed kamani nuts 

 made by fruit-fly larvae, and observations in the laboratory have 

 proved that the adult female Tetrastichus will enter kamani nuts 

 through breaks in the pulp and attack larvae. As many as seventeen 

 punctures have been counted in a single larvae from which were dis- 

 sected forty-one Tetrastichus eggs. Laboratory data have shown that 

 the heaviest parasitism, especially during the warmer seasons of the 

 year, when larval development and emergence is rapid, is to be found 

 among the larvae emerging during the first one or two days after the 

 host fruits have been gathered. Since humilis, tryoni and fullawayi 

 do not oviposit, or at most but slightly, in larvae in fallen fruits, while 

 Tetrastichus does to a much greater extent, it is to be expected that 

 giffardi will be able to parasitize, as they become mature, those 

 larvae that were either unhatched or very young when the host fruit 

 fell from the tree. In this connection, it is interesting to note that the 

 highest percentages of parasitism due to Tetrastichus are shown in 

 Table II, developing among larvae bred from kamani nuts — a fruit 

 always gathered from the ground. 



