The collection of maternity history data at the time of registration of a current birth would seem to be a profitable use of the existence of registration. However, such data present unique problems which have not previously been recognised. These problems relate to the biases caused by the sample not being a random one of all women but one of proven fecund women only, all of whom are at the end of a birth interval. The present work is based on birth notification data collected in the Solomon Islands over a period of nine years. The thesis develops techniques which can be applied to such data to derive conventional demographic indices of fertility and childhood mortality. Established techniques are adapted to take account of the nature of the sample and innovative techniques introduced in this study. The value of each technique developed is discussed and the results compared, where appropriate, both with those from the other techniques applied and also with census results. The birth notification data also provide a rare opportunity to monitor, in a demographic manner, the progress of the concurrent malaria eradication programme in the Solomon Islands. With the islands grouped according to their malarial status, birth characteristics and childhood mortality indices are compared in malarious and non-malarious areas. The effect of malaria eradication is reflected, sometimes dramatically, in some of these indices. With the successful derivation of demographic indices from such data as shown in this thesis, the collection of maternity history data at the time of registration of a current birth is found to be a valuable and viable method of data collection. In addition, for the Solomon Islands, malaria eradication could be monitored with these data.