THE MELON ELY IN HAWAII. 47 



Puncture 1: Twelve unhatched and fresh-looking eggs in a newly made 

 cavity. No gall-like formation evident. 



Puncture 2: Nine dead first-stage larva' and 7 dead third-stage larvae; all in 

 ;ough hard gall about one-half inch in diameter. 



Punctures 3 and 4 (merged) : Eighteen living and one dead first-stage larvae 

 n a callous dark-brown spot beneath the puncture. Living larva' moderately 

 vctive but apparently dying and unable to get out of gall-like formation sur- 

 •ounding them. 



Punctures 5 and 6 (merged) : Eight living and 6 dead first-stage larvae in 

 conditions similar to those in punctures 3 and 4. 



Puncture 7: Four dead first-stage larvae dark brown in color. 



Puncture 8: Sixteen dead first-stage larvae dark brown in color. 



Puncture 9 : Seventeen dead first-stage larvae dark brown in color. 



The data just given would indicate, if considered alone, that the 

 nortality resulting from the gall-like formations might be of prac- 

 ;ical value; however, under Hawaiian conditions they are of neg- 

 igible value. In the field from which the watermelon examined was 

 ;aken, all the very young fruits were badly stung. The 3-inch speci- 

 men examined was one of the comparatively few which had resisted 

 rttack, and it would have been rendered worthless by the deformities 

 ieveloping about the punctures, even if it did not break down under 

 ;ontinued attack by the fly or through decay. 



Larvae die if forced to feed on the decay resulting from bacterial 

 )r fungous diseases of the host. While on occasion 90 per cent of all 

 larvae within a decaying fruit may be found dead, the duration of 

 ?gg and larval life is so short that mortality resulting from decay is 

 'datively small and unimportant. Mortality among larvae within 

 jean pods is occasionally due, apparently, to excessive dryness devel- 

 oping with the maturing of the pod. Excessive sun heat kills a few 

 arvae in fruits well exposed, and the common and abundant intro- 

 duced harvester ant Pheidole megacephala, Fab. is responsible for 

 ihe destruction of some larvae and adults. In climates colder than 

 the Hawaiian littoral climate mortality due to coldness will be 

 found to play a particularly active part in reducing the abundance 

 >f the melon fly. This, however, is not a factor in Hawaii. 



ARTIFICIAL CONTROL. 



No satisfactor}' artificial control measures have been successfully 

 ipplied in controlling the melon fly under Hawaiian conditions. The 

 mly means now employed in Hawaii to safeguard fruits is that of 

 protecting the young fruits with some type of covering until they 

 ire large enough to withstand attack. Trapping the adults and kill- 

 ing them by spraying~Tiave thus far given poor results, while killing 

 the immature stages by submergence in water, burial in soil, or by 

 jold-storage temperatures are methods to be applied in lessening the 

 abundance of the pest only after infestation has taken place within 

 the host fruit. If all growers of host plants would systematically 



