Non-target risk posed by an African parasitoid, Bracon celer Szépligeti (Hymenoptera: Braconidae), was assessed for a classical biological control program against olive fruit fly, Bactrocera oleae (Rossi) (Diptera: Tephritidae: Dacinae), in California, USA. Behavioral and reproductive responses to non-target tephritid species were tested with beneficial (Chaetorellia succinea [Costa] and Parafreutreta regalis Munro) (Tephritidae: Tephritinae) and native (Rhagoletis fausta [Osten Sacken]) (Tephritidae: Trypetinae) fruit fly species in successive no-choice and choice experiments under close confinement in quarantine. Non-target host-plant substrates exposed to B. celer were yellow-starthistle flower heads containing C. succinea, Cape ivy stem galls containing P. regalis, and bitter-cherry fruit containing R. fausta. The parasitoid probed all three infested non-target plant substrates, but significantly less than olives infested with B. oleae. It produced offspring from P. regalis in Cape ivy stem galls, but appeared unable to penetrate yellow-starthistle flower heads with its ovipositor. Bracon celer killed some B. oleae and R. fausta larvae without parasitism. Reproduction on P. regalis indicates that B. celer has a broad physiological host range, which, combined with the parasitoid's acceptance of all three host-plant substrates, indicates a strong potential to negatively impact non-target species. Although physical and temporal barriers to host attack may reduce risk to most non-target tephritids by B. celer in California, the parasitoid should not be released due to its risk of harming the beneficial P. regalis. Release of P. regalis is still under consideration, however, and final risk assessment should depend on whether the fly proves useful for weed control.