Outcomes from large scale health surveys suggest women's morbidity and medical care utilization to be higher than men's. Survey findings have often been questioned on methodological grounds, a main criticism being the subjective nature of the data on which the alleged sex differences are based. Moreover, health differences vary with the type of illness (behavior) that is assessed. Instead of asking subjects about their illness experiences, we performed analyses on data derived from a continuous registration of morbidity as presented in four general practices in the Netherlands. Information was obtained on physician diagnoses of 4723 male and 4963 female patients of all ages enlisted in the four practices from 1984 to 1988. To gain insight in the nature of sex differences, four diagnostic categories were distinguished (sex specific conditions, trauma, symptoms without disease, and prevention and diagnostics). Data on mean number of diagnoses, sex ratio's and proportions of overall morbidity are presented for separate age-groups. In addition, the contribution of each diagnostic category to overall female excess morbidity was computed, for age-groups. Results show that over 40% of the significantly higher overall female morbidity is accounted for by gynaecological and obstetrical diagnoses, and more than a quarter could be explained by prevention and diagnostics. In contrast with the often alleged female excess of 'vague' or psychosomatic symptoms, these accounted for less than 20% of the overall sex difference.