Transplants of human tumors in nude mice have shown a progressive increase during the past 15 years as an experimental model for cancer research. A variety of factors, including relatively fragile health, have been identified that require appropriate experimental controls if the investigator is obtain consistent results. Not all tumors grow in nude mice. The frequency of tumor 'take' varies according to tumor origin, tumor type, inoculation site, age and conditioning of the mouse host, and a variety of other factors. Manipulation of these variables has led to successful propagation of almost every known variety of human malignancy. Following transplant, changes in characteristics have been documented, but the frequency and degree of such changes remains uncertain. Tumor growth rate probably increases after transplantation, requiring great care in the interpretation of chemotherapy experiments, but biochemical characteristics may be more stable. The nude mouse offers great interest as a model for the in vivo study of metastasis, as a number of experimental variables, mainly immunological, have been shown to affect this process. Spontaneous tumors have been shown to arise in these animals, but the controversy over their frequency relative to the thymus-bearing background strain is unresolved. We conclude that the nude mouse/tumor xenograft model, while requiring meticulous experimental controls, is nevertheless an extremely useful tool for cancer research.