18 patients with varying grades of astrocytoma had tumour resection using a new frameless, interactive sterotactic localizing wand. The system enables localizing information to be presented to the operator in multiple two-dimensional or three-dimensional displays in real time. In all cases the wand was used to help outline tumour boundaries in an attempt to resect solid tumour completely. Other uses included placing a minimal craniotomy or modifying an existing craniotomy in 17 (94%) patients, for intraoperative physiologic mapping in 2 (11%) and electrode placement in 1 (6%). All patients had a complete (98-100%) resection by postoperative MRI, and in 8 (44%) the tumour was removed from eloquent areas. Evaluation at approximately 6 weeks after surgery showed that 3 (16%) were improved, 14 (78%) were the same, and 1 (6%) was worse. On the second postoperative day no patient was better than their preoperative status, 10 (56%) were the same, and 8 (44%) patients were clinically worse. At 3 months the figures were 2 (11%), 11 (61%), and 4 (22%) respectively, and 1 patient had expired. The extent of resection has been shown to be an important prognostic factor in both low and high grade astrocytomas. This system allows accurate volumetric near en bloc resection and this was achieved in 17 (94%) cases. In one case of hippocampal tumour the tumour was removed in a piecemeal fashion. This system is an effective low cost alternative to frame volumetric systems and is associated with minimal morbidity at 6 weeks.