In a large urban practice, all families with a child aged under five years were identified and, after suitable exclusions, 294 (80%) of the 369 families approached cooperated in the study. For each child, the number of parentinitiated first consultations in a one-year period was extracted from general practice records, and mothers were asked to provide additional information by postal questionnaire, which included a test of the mother's reaction to a series of hypothetical child ;illness situations'.The most important factors found to be associated with higher rates of consultation were increased urgency of the mother's reaction to the test situations and attendance of the child at a pre-school group. Children with low consultation rates tended to be younger and more likely to come from larger families; their mothers generally spent less time in regular employment outside the home and were more likely to be single parents. The effect of family size was independent of maternal age and no socioeconomic gradient was found. There was also no evidence that differences in frequency of consultation were related to particular diagnostic categories. The mothers' sources of advice in the postulated event of their child becoming unwell are also described.