A total of 83 patients with metastatic transitional cell carcinoma who had previously received no systemic therapy entered a randomized phase II evaluation of carboplatin and cis-dichloro-transdihydroxy-bis-isopropylamine platinum IV (CHIP), administered respectively at 400 and 270 mg./m.2 every 28 days. Among evaluable patients with measurable disease response rates were 3 of 22 (14%, 95% confidence interval 5 to 35%) for carboplatin and 4 of 25 (16%, 95% confidence interval 5 to 36%) for CHIP. Among 17 patients with evaluable but not measurable metastases (10 carboplatin and 7 CHIP recipients) there were no responses. Median survival for 64 evaluable patients was 4.8 months (5.0 months for carboplatin and 4.3 months for CHIP recipients). Independent factors prognostic for survival (p less than 0.01) were performance status (0 or 1 versus 2 or 3), liver metastases, prior radiation therapy and recent weight loss (p = 0.02). Multivariate analysis confirmed that a performance status of 2 or 3 and liver metastases were predictive of shorter survival. A total of 31% of the patients treated with carboplatin and 34% of those who received CHIP experienced severe or life-threatening myelosuppression. While the response rates with carboplatin and CHIP are modest, we believe that the characteristics of these agents indicate that they should be evaluated further.