An abundant literature provides informations upon the prognosis of supratentorial low grade (grade II) astrocytomas, but the series are quite heterogeneous in terms of clinical material, neuropathological evaluation and statistical methods of analysis. So, outcome, prognostic factors and therapeutical indications are poorly defined. A retrospective study of 141 adults patients suffering ordinary low grade astrocytoma diagnosed in our unit between 1978 and 1988 was conducted. A follow up of 5 years or more, since neuropathological diagnosis, was mandatory for inclusion. Endpoint of statistical analysis was duration of survival. Results were expressed after uni and multivariates analysis. Clinical and morphological features of our series were comparable to those previously reported in the literature. Median survival time was 52 months. Considering age at diagnosis, survival curve analysis showed highly significant differences (p < 0.0001) and established three prognostic classes of age (before 50, between 50 and 60, and after 60 years of age). Comparisons of survival curves showed significant statistical differences (p < 0.0001) according to pretreatment functional status. Analysis of a clinical condition using a functional scale is more powerful, from a statistical point of view, than an individual analysis of each constituting parameters. In this series a good correlation was found between functional status and age at diagnosis. Neurological deficit was more correlated to age than to tumor location. Multivariate analysis, using the Cox model, defined some parameters acting independently on duration of survival: fronto-parietal location (p < 0.0001), pretherapeutical functional status (p < 0.001), age at diagnosis (p = 0.001), deepseated or multicentric tumor, mass effect (p < 0.01), frontal location (p < 0.02), total surgical removal (p < 0.05). Non significant parameters were: radiotherapy, time before diagnosis, age of onset, sex, contrast enhancement on CT scan. Our results are in agreement with previously reported data confirmed literature data concerning prognostic factors characteristics of the host (age), characteristics of the tumor (location) and expressing the influence of the tumor on the host (functional status). Our series documented prognostic clinical forms of these tumors (function of age, of performance status, and of tumor location). Our results confirm the relative inefficacy of therapeutic weapons, considering the population as a whole, on duration of survival. Based on our results and on recently published data, a decision tree analysis was proposed for management of grade II astrocytomas.(ABSTRACT TRUNCATED AT 400 WORDS)