The infectivity of 4 isolates of Metarhizium anisopliae to puparia of Ceratitis capitata treated as late third-instar larvae in unsterilized soil was investigated in the laboratory under controlled temperature and moisture. At 20-30 degrees C, mortality in puparia was highest at water potential of -0.1 and -0.01 mega Pascal (MPa) and lowest at water potential of -0.0055 and -0.0035 MPa in all the isolates. In wetter soil however, isolates ICIPE 20 and 60 caused significantly higher mortality than ICIPE 18 and 69. The survival of conidia in drier soil (-0.1 MPa) was not adversely affected at all temperatures. However, in wet soil (-0.0035 MPa) there was drastic reduction in colony counts in ICIPE 18 and 69 at 25 and 30 degrees C but conidial density in ICIPE 20 and 60 remained at the initial level at 14 days after inoculation at all temperatures. When ICIPE 20 was evaluated against three other fruit fly species (Ceratitis cosyra, Ceratitis rosa, and Ceratitis fasciventris), significant reduction in adult emergence and higher pupal mortality occurred in C. cosyra and C. fasciventris than in C. rosa at a combination of 15 and 20 degrees C and -0.1 and -0.0035 MPa. However, at higher temperature and the same moisture level, the isolates were equally pathogenic across the 3 species. It is probable that in addition to pathogen cycling and multiplication from dead infected insects in the soil, a balance between microbial degradation and replenishment of inoculum of virulent isolates occur through fluctuations in, and intricate interactions between temperature and moisture levels. This study is indicative of the potential of using isolate ICIPE 20 for soil inoculation against pupariating third-instar larva of fruit flies, thus providing a novel alternative to chemical soil application.