The efficiency of two national record systems for malformation monitoring has been compared. During 1973 and 1974, two such registers were run in parallel in Sweden. One is based on specific report cards which are sent in when a malformed infant is born, the other is a computerised analysis of routine birth record diagnoses. Both systems lose a similar, and significant, proportion of malformed infants. The quality of diagnosis is lower in the system of routine birth records than in the system based on specific notifications--especially in infants with multiple malformations or with reduction malformations of the limbs. The specific report system is the only one suitable for monitoring of malformations; it is also somewhat quicker than the other system.