Aspirin 650 mg and metoclopramide 10 mg in an effervescent preparation (Migravess) were compared with effervescent aspirin 650 mg (Alka-Seltzer) and placebo for common migraine attacks with a double-blind cross-over design. One hundred and eighteen patients with common migraine were entered. Eighty-five patients completed all three forms of treatment, eleven completed two, and six completed one. Medicine was taken when patients were sure they had a migraine attack and not just interval headache. After each form of treatment, they mailed a report form to the investigators. Additional medication was allowed after 2 h and was taken for 79/95 placebo treated attacks, 63/92 Migravess treated attacks, and 51/86 aspirin treated attacks (p less than 0.01). Aspirin was significantly better than placebo for pain but not quite significant for nausea. Migravess was significantly better than placebo for pain and for nausea. There was no significant difference between aspirin and Migravess with regard to analgesic effectiveness (p = 0.33) or to antinausea effect (p = 0.18).