A prospective study was undertaken to identify the epidemiological factors associated with the development and spread of the hydropericardium syndrome in broiler flocks. Data were collected between July 1989 and April 1990 from 131 flocks in 105 broiler production units pertaining to their demography, management practices, prophylactic procedures and concurrent diseases. The incidence rate of the syndrome in the whole population was 46.6 per cent. There were significant associations between the incidence in a flock and visits by a poultry vaccination crew (P = 0.014), the number of flocks raised (P = 0.004) and the source of light and heat (P = 0.007). Flocks that had one or more visits by a poultry crew were 15 times more likely to be affected by the syndrome than flocks that had no such visits. Premises where one flock was raised were nearly three times more likely to be affected than premises where two flocks were raised, and the use of electricity as a source of light and heat entailed a much lower risk of hydropericardium syndrome than kerosene oil.