To discover whether the beneficial effects observed when Drosophila C virus (DCVc) infects Drosophila larvae should be considered as a specific case or, as a general rule, the DCVc-Drosophila association was studied in different environments (3 host-populations, 2 temperatures, 2 viral dose). We measured pre-adult mortality, developmental time and number of ovarioles per fly. In all experiments, and, in all contaminated populations, DCVc could be considered, either a parasite since it increased pre-adult mortality, or a symbiont since it decreased developmental time and increased mean number of ovarioles. The viral dose necessary for the appearance of beneficial effects varied with population, temperature and life-cycle parameters. A sufficient number of viral particles appears necessary at a specific point in the host life-cycle for a beneficial DCVc effect on the host. DCVc does not affect the variance of any of these parameters. The observed changes in DCVc-contaminated populations were thus not a response to the selection pressure of DCVc. Two arguments support the association as an example of the ''cohabitation-without-struggle model'' and will be discussed.