Neoplastic and non-neoplastic mammalian urothelium was tested for angiogenesis activity in the rabbit iris assay. Tiny fragments of tissue were placed on a rabbit's iris and observed through a slit lamp stereoscope for as long as 4 days after implantation. Angiogenesis activity was demonstrated either by the appearance of new capillary growth or, following an injection of fluorescein, by a green fluorescence around an implant. Capillary proliferation was stimulated by 99 per cent of the specimens containing malignant urothelium, while only 9 per cent of normal urothelial specimens had this effect (p less than .001). Cells grown in vitro were similarly tested for angiogenesis by implanting 1 mm. fragments of collagen coated either with malignant or normal cells. A total of 92 per cent of the specimens covered with neoplastic cells induced angiogenesis, while only 15 per cent of the implants covered by normal cells caused this response (p less than .001). Benign reversible hyperplasia was induced in a rabbit bladder by systemic administration of cyclophosphamide, and an angiogenic response could not be detected in a small number of specimens. The results suggest that angiogenesis activity may be a reliable method for distinguishing between benign and malignant bladder tissue.