a1 From the George Williams Hooper Foundation, University of California, San Francisco, California
All fleas used in this study were collected in the field and except for Pulex irritans were cultured in the laboratory. The ten species studied were Xenopsylla cheopis, Nosopsyllus fasciatus, Orchopeas sexdentatus sexdentatus, Opisodasys-nesiotus, Megabothris abantis, Malaraeus telchinum, Diamanus montanus, Echidnophaga gallinacea, Pulex irritans and Oropsylla idahoensis.
All fleas transmitted in individual feeding studies with the exception of the latter two species, which transmitted en masse. Echidnophaga gallinacea could not be fed periodically as were the other fleas because of its tick-like feeding habits. Consequently, the vector efficiency obtained for this species is not strictly comparable to that found for the other species.
The transmission data obtained from individual flea feeding studies was analysed statistically to estimate the expected number of transmissions per flea of each species. These values are obtained as intervals which have a 90% probability of containing the true value. The true vector efficiency of Xenopsylla cheopis was found to be 0·660 ± 0·234 (expected transmissions per flea), that of Nosopsyllus fasciatus to be 0·213 ± 0·157, and that of Orchopeas sexdentatus sexdentatus to be 0·170 ± 0·138. Opisodasys nesiotus, Megabothris abantis, Malaraeus telchinum and Diamanus montanus transmitted very inefficiently.
Experimental evidence was obtained that different strains of a species of flea may differ markedly in their biological vector capacity. In contrast to results obtained in this study, Wheeler & Douglas found Diamanus montanus to be an exceptionally good vector; in their studies it proved to be an even more efficient vector than Xenopsylla cheopis. The strain of Diamanus montanus employed by them came from an area widely separated from that in which the strain used in the present studies was originally collected.
Since many blocked fleas did not transmit it is probable that the experimentally determined vector efficiencies are lower than they would be in nature, where the blocked flea has constant access to a host and hence greater opportunity to feed.
Attempts to determine the number of organisms regurgitated by a blocked flea during its attempt to feed did not prove entirely satisfactory, but gave an indication that the number may be at times from 11,000 to 24,000 organisms. The technique consisted of feeding a blocked flea on the shaved abdomen of a mouse (later on the ear), then immediately doing a biopsy on the area around and including the bite wound. This biopsy material was then finely ground and plated on a sensitive bacteriological medium. Some mice upon which biopsies were performed nevertheless contracted plague and died. This must lead to the conclusion that an infective flea may deposit organisms directly into the capillaries and that the primary stage of infection resulting from a flea bite is frequently bacteraemia.
(Received January 23 1947)
* Now with the Division of Entomology and Economic Zoology, University of Minnesota, St Paul, Minnesota.