A mortality analysis has been carried out on a cohort of patients given Fowler's solution (potassium arsenite) for periods ranging from 2 weeks to 12 years between 1945 and 1969. An excess of fatal and non-fatal skin cancer was apparent, but there was no overall excess mortality from cancer. Further analyses by site of cancer, dose level, and time from first exposure are also presented. A subset of patients were examined in 1969-70 for the presence of arsenical keratoses, hyperpigmentation and skin cancer. About half the patients had one or more of these signs. Although the cancer mortality of this entire subgroup was similar to the expected value, all the cancer deaths occurred in patients with prior signs of arsenicism. These data suggest that while any excess of internal malignancy due to the use of Fowler's solution is small or non-existent, there may be a susceptible subgroup which can be identified from dermatological manifestations.