The larval feeding requirements and biology of the generalist predatory muscid hunter fly Coenosia attenuata Stein 1903 (Diptera: Muscidae) were investigated at 25°C. Larval C. attenuata were fed second-, third, and fourth-instar (L₂, L₃, and L₄) larvae of the fungus gnat Bradysia impatiens (Johannsen) (Diptera: Sciaridae) at variable rates to determine minimum and optimum numbers of these prey required for normal development. The proportion of C. attenuata larvae surviving to pupation differed significantly as a function of L₂ and L₃ prey numbers. When the number of prey/ d was increased from 10 to 15 L₂ and from 5 to 7 L₃ per day, the respective percentages of pupation increased from 0 to 77% and from 0 to 48%. In contrast, all numbers of L₄ prey (1–7 prey per d) supported pupation, and the pupation rate did not vary with prey number. At the highest prey numbers tested, mortalities of C. attenuata larvae fed L₂, L₃, and L₄ fungus gnat larvae were 7, 30, and 75%, respectively. The higher mortality of larvae fed L₄ prey was clearly the result of lethal wounds inflicted by the fungus gnat larvae in defensive strikes against the predators. At prey numbers supporting maximum rates of adult emergence, larval development required 12–14 d, and duration of the pupal stage was ≈10 d. C. attenuata larvae killed large numbers of prey during their development (means of up to 232 L₂,144 L₃, or 87 L₄ fungus gnats), and larvae provided with marginally inadequate numbers of prey survived for long periods (mean 14–22 d, maximum 34 d) before succumbing to apparent starvation. These are favorable attributes with respect to use of C. attenuata as a biological control agent, suggesting a strong potential to substantially impact high-density pest populations and to survive in low-density pest populations.