It will perhaps be generally admitted that few of the practical problems which have from time to be faced by the Medical Officer of Health in his capacity of statistician present more difficulty than is often involved by the estimation of populations. This problem may be said naturally to present itself in three different forms, according as the estimate is required for a past, a present, or a future date. Practically, however, there are but two primary divisions of the problem, the one relating to dates since which a census has been taken (and its results published), and the other to dates subsequent to that of the last published census. This distinction is pointed out by Dr Cressy Wilbur in the “Thirty-second Annual Report (for 1898) on the Vital Statistics of the State of Michigan.” He there refers to the first class of cases as estimates of inter-censal, and the second as estimates of post-censal populations.