A computer file of all Scottish hospital admissions in the period 1968-1976 was searched to identify the 2,505 children (aged less than 19 years) with a diagnosis of diabetes. The average annual incidence of the disease (based on first hospital admission) was estimated to be 13.8 per 100,000 children aged less than 19 years (boys 14.4 per 100,000; girls 13.2 per 100,000). The highest incidence, 20.0 per 100,000 was in the age group 10-14 years and the lowest 7.1 per 100,000 in those aged less than 5 years. It is estimated that during the study period there was an 80% increase in the annual incidence of juvenile diabetes, from about 10 per 100,000 in 1968 to about 18 per 100,000 in 1976. First admission rates showed seasonal variations for those aged 5 years or more, with peaks in October/November and January/February. Marked variation was found in the incidence rates in the different counties of Scotland. The central lowlands which includes the cities of Edinburgh and Glasgow was an area of low incidence. There appeared to be an inverse correlation between the incidence rate in each county and population density. In Glasgow, there was an inverse association between the incidence rate in each city ward and the average number of persons per room. There was no evidence of space-clustering of the disease in different years within the parishes (rural districts) of each county and there was no convincing evidence that the variation in the incidence of diabetes between parishes in the same county was more than might have been expected to arise by chance. The observations are compatible with the disease having a viral aetiology but it is difficult to explain the striking rise in incidence over the study period on this basis.