One hundred twenty-eight alcoholic men were assigned randomly to receive either a regular dose of disulfiram (250 mg), a pharmacologically inactive dose (1 mg), or no disulfiram. There were no statistically significant differences among the three treatment groups in total abstinence, percentage of drinking days, days worked, family stability (living with same relative), or percent of scheduled appointments kept. However, 21% of those who received the regular dose of disulfiram and 25% who received the pharmacologically inactive dose remained abstinent, whereas only 12% of those who received no disulfiram did so. These results indicate that disulfiram may be of limited value in the treatment of alcoholism, fear of the disulfiram-ethanol reaction is important in preventing drinking, and patients wiling to take disulfiram are more likely to be abstinent if given the drug. We also found that complete abstinence correlated significantly with compliance and obtaining employment.