The practice records were an inadequate source of information except when the certificate of death had been issued by the practice or a hospital report had been received. Autopsies were done only for deaths reported to the coroner and for some which occurred in public hospitals; in most cases the cause of death recorded by the doctor could not be verified. The lack of accurate information recorded about the cause of death in a patient who did not die in a hospital suggests that the doctor's clinical interest dissipates once death has occurred. Notes about the death were found in some records of relatives of the deceased but were not sufficient to substantiate the suggestion that the doctor had other demands at the time in supporting the bereaved. It was not certain that every death in the practice population was known to the practice; this could be overcome by the use of a computer which could identify inactive records and thus enable further enquiry. Malignant disease was the most commonly recorded cause of death. It is possible that most of the 27 deaths of unknown cause were due to myocardial infarction (which would parallel the causes recorded for all Australia) because they were unexpected. Three quarters of those who died were older than 65 years. A doctor was in attendance at 63 per cent of all deaths in the practice; the major determinant of this involvement was the place of death.(ABSTRACT TRUNCATED AT 250 WORDS)