Fungi belonging to the order of Entomophthorales are found to induce host epizootics under natural conditions, restraining many agroforestry pest populations below economic damage level. For decades, researchers have focused on exploring these valuable resources with respect to their application in pest biocontrol programs, but the process has proven to be difficult. A poor understanding of the fungal epizootic mechanism and survival strategies in the nature, including the mechanism of in vivo resting spore formation, is one of the main obstacles hindering this effort. Entomophthorales can actively eject numerous conidia to infect nearby hosts, thus transmitting within the host populations. When the host population decreases dramatically, the fungi tend to form thick-walled spores (resting spores, RS) within the bodies of cadavers. Once the conditions improve, RS germinate to restart the infection cycle and regulate the host population. Thus, the RS formation is an important step for fungal survival and the subsequent initiation of the epizootic. Previous reports indicated that many biotic and abiotic factors influence in vivo RS formation, including temperature, humidity, light duration, host age, nutritive conditions, and inoculum levels. However, no consistent conclusion has been formed regarding the vital factors, as contradictory results have been reported for different pathogen-host systems. Temperature and inoculum levels were recently found to exert dominant effects on the formation of Pandora nouryi RS in Myzus persicae inhabiting fields in cold climates. To investigate the main factors influencing in vivo RS formation of the aphid-specific pathogen Conidiobolus obscurus, we performed a series of multi-concentration bioassays of C. obscurus against the host M. persicae under five different temperature treatments (10 ℃, 15 ℃, 20 ℃, 24 ℃, and 28 ℃). The results of the bioassays showed that aphid mortality increased with increasing conidial concentrations used to inoculate the aphids (inoculum level) under each temperature condition. The data were well fitted to the time-concentration-mortality model, and the lowest calculated median lethal concentration (LC₅₀) with 95% confidence intervals was estimated to be 13.6 (11-16.8) conidia/mm² on day 4 at 28 ℃. Within the crushed cadaver bodies, proportion of the RS formation significantly increased with increased temperature, reaching a peak of 0.66 at 28 ℃. The proportion also increased with inoculum levels at temperatures from 15-24 ℃. Thus, C. obscurus tends to form RS more readily at higher temperatures and inoculum levels. We explain this phenomenon as follows: the rise of temperature to high levels signals the upcoming variation in the host population to C. obscurus, because high temperatures always cause host diapause and population collapse in the fields. For survival in the successive period of host shortage, the fungi tend to form thick-walled RS in host cadavers. Further observation of other entomophthoralean fungi and their hosts is required to verify this prediction. The present results also showed that C. obscurus possesses high infectivity under higher temperature conditions and is thus suitable for aphid control in thefields.