The results of posterior fossa explorations for trigeminal neuralgia over the period 1980-1990 in 58 patients and in 59 procedures were studied retrospectively. In 51 procedures vascular compression was treated by microvascular decompression (MVD). In the absence of such a compression, partial sensory rhizotomy (PSR) was performed in 5 cases and only adhesiolysis of thickened arachnoidea in 2 cases. In one additional case the procedure was terminated prematurely due to the development of cerebellar edema. Two months postoperatively a good or fair result was obtained in 80% of the procedures. At long-term follow-up (mean 77.3 months, range 8-146 months), a good or fair result was maintained in 71%. There was no significant difference in outcome between the MVD group and the other procedures, or between the three groups formed according to the type of vascular compression. However, in the group of 10 patients with a history of a procedure affecting the trigeminal ganglion or nerve root the result was worse. In the group of 41 MVD patients rendered free of pain at 2 months postoperatively, 8 patients perceived a recurrence. The annual recurrence rate was calculated to be 2.6%. There was no mortality in this series but the morbidity rate was 22% including 1.7% persistent neurological deficit. Compared to the literature results of percutaneous controlled differential thermocoagulation (PCDT), the recurrence and failure rates in the present series appear to be more or less the same. As neither of the two is an unequivocally more effective treatment, we are of the opinion that the less invasive procedure should be preferred.(ABSTRACT TRUNCATED AT 250 WORDS)