A panel of 17 melanoma xenograft lines was exposed to various DNA reactive agents. Growth delay values were calculated from tumor volume responses of xenograft bearing mice and correlated with donor patient's response as determined by standard clinical criteria. Most donor patients were refractory to dacarbazine and received the combination of ifosfamide and cisplatin as secondary treatment. Clinical and preclinical responses correlated well for resistance in 26/27 instances, whereas sensitivity was observed consistently in only 9/12 instances. Incomplete cross-resistance among cisplatin, dacarbazine, dibromodulcitol, ifosfamide, methyl-CCNU, mitomycin C, and PHM has been observed after sequential drug exposures in clinical trials and was confirmed by simultaneous drug exposures in this panel of melanoma xenografts. Most xenograft lines revealed unique chemosensitivity patterns, which offer a therapeutic potential yet to be fully exploited in clinical therapy guided by predictive tests. Macromolecular DNA damage was compared in xenograft lines differing at least 100-fold in sensitivity to dacarbazine. Removal of DNA damage 24 hours after in vivo treatment with dacarbazine paralleled in vivo sensitivity. The preclinical and clinical data on responses to various DNA reactive drugs validate the xenograft model for therapeutic purposes. The model has found an additional application in analysis of the molecular pharmacology of response to in vivo drug treatment.