In a case-control study of 401 breast cancer cases and 625 age-matched controls, the possible etiologic role of hair dye usage was investigated. No association was found between hair dye use and breast cancer (odds ratio = 1.02; 95% confidence interval = 0.77-1.32) or any factors affecting dosage, frequency, duration, type, or color of hair dye. Important confounders of hair dye use included religion and smoking status. When stratified by 10-year age groups, there was no appreciable trend in risk status between case and referent populations. No dose-response relationship (frequency X duration X shade X type of dye used) was found, even after controlling for factors affecting hair dye use. A multiple logistic regression analysis supported the absence of this relationship.