16 BULLETIN 66, UNITED STATES NATIONAL MUSEUM. 



temporary host, by their own volition. Saunders (1853) proved that 

 they have remarkable saltatorial power, which power, together with 

 the aid of the sucker-like pads of the tarsi, will enable them to reach 

 an insect winch may pause but a second at the flower. In such 

 manner there may possibly be several temporary hosts in some cases 

 before the triungulinid reaches the nest. In connection with the 

 part the maternal host fulfills in the destiny of the parasite, Newport 

 (1851) makes the significant remark that the bees live long enough 

 to perpetuate the enemies of their race and to serve the young Strep- 

 siptera as a means of transportation to their hosts. 



The triungulinids may be forcibly removed by the bee, as was 

 observed by Smith (1859) with regard to Andrena trimmerana. He 

 says: "She was covered with hundreds of larvae of Stylo ps and her 

 brushing and excitement were caused by efforts, on her part, to free 

 herself from the annoyance which the host of larvae evidently occa- 

 sioned her." 



The writer can now add another link in the chain of evidence 

 which has been produced to strengthen the flower-transfer hypothesis. 

 On July 21, 1906, at Dallas, Texas, Mr. W. A. Hooker collected from 

 Helenium tenuifolium, a Leionotus verus Cresson containing two 

 female parasites which had produced triungulinids, and also an Aga- 

 postemon texanus bearing two triungulinids of the identical species on 

 the last ventral segment. Thus it may be seen that the triungulinids 

 are actually left at the host flowers and also that many climb upon 

 the wrong host, just as the writer formerly showed to be the case 

 with the triungulinids of Mgodites solidaginis. 



Perez (1886) suggests that in seeking shelter overnight or in stormy 

 periods in the nests of other individuals, the maternal host or an 

 intermediate carrier may leave triungulinids in position to find per- 

 manent hosts. 



In case of the social and solitary wasps the places frequented are 

 little puddles of water, posts, and often flowers. In the case of Hq- 

 moptera the triungulinids are very likely scattered throughout the grass 

 over which the young leaf-hoppers are feeding. With Gryllotalpa the 

 agency of transfer must be some damp locality where these insects 

 congregate. The actual transfer agency must at present be a matter 

 of conjecture in all such insects. Suffice it that either the maternal 

 hosts or some temporary hosts transfer the young to the places where 

 they can find their permanent hosts. 



The triungulinid immediately attacks its host when it finds the 

 proper one. Saunders (1853) found that Xenos larva? would not 

 attack soft Polistes pupae nor eggs, but that they would burrow 

 immediately into a larva, being completely inside in a space of three 

 hours. He found that triungulinids of Hylechthrus were unable to 

 effect an entrance through the pellicle of Polistes larvae. 



