Urothelial tumors develop along two distinctive phenotypic pathways (super®cial papillary non-invasive tumors versus¯at carcinoma in situ lesions), with markedly di erent biological behavior and prognosis. Although multiple genetic alterations have been identi®ed in human bladder cancer, their cause ± e ect relationship with the two pathways has not been ®rmly established. Using a urothelium-speci®c promoter of the uroplakin II gene, we showed earlier in transgenic mice that the urothelial expression of SV40T antigen, which inactivates p53 and pRb, induced carcinoma in situ and invasive and metastatic bladder cancer. In striking contrast, we demonstrate here that the urothelial expression of an activated Ha-ras in transgenic mice caused urothelial hyperplasia and super®cial papillary non-invasive bladder tumors. These results provide strong, direct experimental evidence that the two phenotypical pathways of bladder tumorigenesis are caused by distinctive genetic defects. Our results indicate that Ha-ras activation can induce urothelial proliferation in vivo; and that urothelial hyperplasia is a precursor of low-grade, super®cial papillary bladder tumors. Our transgenic models provide unique opportunities to study the detailed molecular events underlying di erent types of bladder neoplasms, and can serve as useful preclinical models for evaluating the in vivo e cacy of preventive and therapeutic agents that act on various signaling pathways in bladder cancer.