Population-based registration of all cancer cases diagnosed in Singapore began in 1968. While most of the basic demographic and clinical data collected are reasonably adequate and accurate, the item on occupation is quite unsatisfactory. Most notifying practitioners either do not state the patient's occupation or use general terms that do not indicate any particular environmental exposure. In spite of the difficulties, some notifications do contain specific information on the occupation. For the period 1968-1977, analyses on cancer incidence in occupational groups were confined to Chinese males aged 35-64. Due to the lack of relevant population data, relative risks were based on the Standardised Relative Proportional Risk (SRPR), standardised for age, dialect group and place of birth. Managers and clerical workers seem to have high SRPRs for cancer of the large bowel, but low SRPRs for cancer of the lung and oesophagus. High SRPRs for lung and oesophagus are seen in some groups of manual workers (e.g., bricklayers, carpenters, transport equipment operators and labourers not otherwise specified). There is also a high SRPR for skin cancer among farmers.