The following notes are an attempt to show
(1) How select and aggregate rates of mortality can be found from insurance office data by the methods at present used to find rates of mortality from censuses and the records of deaths among the general population, and
(2)How the census method can be modified in the case of a prolonged investigation to give a continuous mortality investigation.
We have also tried to deal with a few of the objections to the method, and some of the practical points that arise in its application.