Improvement of infant health has been a key objective of public health and population programs in both developing and developed countries. Monitoring of such programs through the measurement of infant mortality has occupied the attention of health statisticians and demographers in many countries. As this paper demonstrates, measuring changes in this important indicator is not an easy task because of the problems encountered in the accurate counting of both births and of infant deaths. Although it is difficult to verify the responses in this household survey because of the anonymity provided to respondents, the results and discussion provide some very useful suggestions for improving the registration of births and infant deaths in Cameroon and by extension in other developing countries. Their most important conclusion perhaps is that "a mechanism of registration that uses medical institutions may substantially improve coverage and timeliness of registration."