SYNOPSISThe McGilI‐Melzack Pain Questionnaire (MMPQ), comprised primarily of adjectives descriptive of pain, was administered to 100 patients seeking treatment for headache. Diagnostic classification of patients into migraine or muscle contraction headache groups was conducted by a screening neurologist using information other than pain description. Reliability determinations were made following independent diagnosis by two other neurologists using a headache pain history and symptom form (HPHSF) devoid of pain adjectives. Comparison of 30 subjects HPHSF's revealed that the screening neurologist's diagnosis and that of two other neurologists, (t = .87,p<.01) and (t = .48,p<.05), indicate high agreement between three physicians in headache diagnosis devoid of pain description. Results show that migraine headache patients report significantly more affective words (t99= 3.89,p<.001) than do muscle contraction headache sufferers; however, no significant differences existed between the groups' use of sensory or evaluative works. Migraine sufferers, while not reporting their headache as more severe “generally,” did report more intense pain when recalling headache at its “worst” (t99= 2.69,p<.01) and at its “least” (t99= 1.74,p<.05) compared to the muscle contraction headache 'group. Discriminant analyses were conducted on one‐half of the sample to determine diagnostic group membership on the basis of pain description alone. Findings revealed that group membership could be predicted at a 90% rate (x21.45= 33.06,p<.001). A cross validation on the second half of the sample confirmed these findings (x21.48= 13.08,p<.05) suggesting that the MMPQ is of value in headache diagnosis. Differences between electromyographic studies and headache pain report are discussed as well as suggestions concerning modification of the MMPQ for head