It is regretted that in these two places Jhr. G. M. M. Alting von Geusau was wrongly described as Chairman of the Association of Dutch Life Offices (Nederlandse Vereniging ter Bevordering van het Levensverzekeringwezen).

Although this paper deals with the application of mathematical formulae to mortality data, it is not concerned with ‘graduation’ if that word is held to imply the fitting of a particular curve to a particular experience with the object of satisfying statistical tests. Nor is the paper concerned with the development of any philosophical theory of mortality. The experiments which it describes were undertaken in the hope of finding a standard type of curve which would give a good over-all representation of adult mortality in general. If such a curve could be found, an examination of the variations in its parameters might contribute something to an understanding, not of the nature of mortality itself, but of the differences—more particularly the secular differences—between one mortality experience and another.

The purpose of this paper is to provide a basis for a discussion on the treatment of proposals for life assurance on sub-standard lives. To assist in this discussion a body of mortality data in respect of the sub-standard lives of a particular office is presented. The data relate to ordinary life assurances issued in this country subject to a diminishing lien. In the light of these data certain suggestions are made regarding the relative financial importance of the extra mortality of impaired lives in modem conditions. Apart from this statistical material, there is little that is new or original in the paper, although it is hoped that the author's relatively simple outlook on the subject may be of general interest and of assistance to students. The paper is concerned with the actuarial and practical aspects of the subject rather than with the medical aspects.

When the employment of as a medium for first-difference interpolation was suggested in J.I.A. LXXII, 453, no mention was made of a corresponding medium for the amount of an annuity. It seemed that the matter might bear further investigation from this angle, and that it would be well to bring the allied continuous functions within the scope of the inquiry. These latter are simpler to deal with because they may be expressed in terms of only one variable (ηδ).