Recently the sandfly Phlebotomus guggisbergi was found to be a vector of Leishmania tropica causing cutaneous leishmaniasis in the Laikipia focus, Kenya, but extensive searches have shed no light on the identity of the rural reservoir host(s). In order to discover more about the biology of the vector, a host feeding preference study was conducted on wild sandflies in their natural cave environment over a 6-month period. Solid state Army miniature (SSAM) traps, without light bulb, were suspended over cages with potential hosts or an empty cage control. The animals tested included sheep, goat, dog, cat, hamster, rabbit, giant rat (Cricetomys gambianus), crested rat (Lophiomys imhausi) and rock hyrax (Procavia capensis), all of which (except hamsters) are normally found in the vicinity of the study site. Sandfly collections from traps baited with goat, sheep, cat, dog, rabbit, or wild rodent species were significantly higher than the control, whereas trap collections with hamster and rock hyrax were not significantly different from the control. Numbers of sandflies collected from the goat, sheep and cat were significantly greater than from the rabbit and rodents. The sex ratio also varied between collections: larger animals attracted a higher proportion of female P. guggisbergi than did the smaller animals (P > 0.05). Therefore greater emphasis should be placed on surveying larger animals to assess their status as reservoir hosts for L. tropica in Kenya.