The indirect method of calculating infant mortality rates was first developed by Brass, and was further modified by other scholars such as Trussell, Sullivan, and Feeney. Although those 3 methods are derived from the same method (Brass), they require different assumptions, and may sometimes give different results. Accordingly, it is necessary to select the 1 method that is most suitable for the available data as well as for the Indonesian population in order to produce a single estimate. For these reasons, this paper first presents the estimates of mortality rates by applying the Brass, Sullivan, and Trussell methods to the available data from the 1971 Population Census, the 1976 Intercensal Population Survey, the 1979 National Socioeconomic Survey, and the 1980 Population Census. It is reported (in Table 1 of the original paper) that the relative differences of the estimates are about 1-3%. This is true for all 4 data sets. Given this small difference, it is concluded that any method may give similar results. Trussell's method is then used in further discussions. Applying the method to the data resulting from the 1980 Population Census results in an estimate of infant mortality of 107/1000 live births with a life expectancy at birth of 52 years. These figures, however, refer to 1977. The corresponding figures for urban and rural areas are 86 and 113, respectively. Large differences in infant mortality rates are also reported among provinces ranging from 62, the lowest for Yogyakarta, to the highest, 187, for West Nusa Tenggara. The pattern of provincial differences in infant mortality rates seems to resemble the development differences. There has been a decline in infant mortality rates in Indonesia. This is clearly indicated and supported by the estimates based on the 4 sets of data. The paces of the declines are not the same, however. The 1971 Population Census which gives the trends for the 1960s indicates a less rapid decline than the 1 resulting from the 1980 Population Census and the 1979 National Socioeconomic Survey. Whether the mortality rates will decline more rapidly in the coming decades requires further research. In comparing the estimates of infant mortality rates resulting from the 1971 and 1980 Population Censuses and assuming an equal annual change, an annual reduction of 3.2% is reported. If data from the 1980 Population Census only are employed, the annual reduction rate is estimated at 1.6%. Given these figures, it is argued that the target of an infant mortality rate of 45 in the year 2000 will be hardly met unless there is a rapid improvement in the health of the population. Additionally, more attention should be given to the health of the population in the provinces with high mortality rates such as those of outer Java. (author's modified)