The Nature and Criteria of Senescence 



Some real animal populations decline in an approximately 

 logarithmic manner. The 'potential immortality' of individuals 

 in a population following such a path of decline, an entirely 

 meaningless phrase which has caused much philosophical agita- 

 tion in the past, is not more significant as a practical issue than 

 the 'potential 5 meeting of any pair of railway metals at infinity. 

 No population of organisms which is subject to a constant 

 overall death-rate contains 'potentially' immortal individuals. 

 The only advantage which a non-senescent organism possesses 

 over senescent forms is that the odds in favour of its death 

 within a fixed period remain constant instead of shortening 

 with the passage of time. 



The human survival curve, in societies possessing developed 

 medical services and a high standard of living, is intermediate 

 between the rectangular and log-linear contours, but approaches 

 the rectangular, with an initial decline due to infant mortality. 

 Figs. 1-3 and 7 show, first, the comparative curves of mortality 

 for populations in the present century living under different 

 conditions of economic and climatic advantage, and second, the 

 change in form of the life- table for North German populations 

 between 1787 and 1800. Many life- tables for populations 

 before the advent of scientific medicine are given by Dublin 

 (1949). The significance of technical and economic privilege is 

 nowhere more evident than in the study of life-tables. The 

 effects of public health upon the life- table are expressed rather 

 in making it approach more closely to the rectangular shape 

 than in prolonging the preinflectional part of the rectangle. In 

 very many organisms, and in man under bad social and medical 

 conditions, the infant mortality is so large as to obscure all sub- 

 sequent trends, the curve coming to imitate Pearl's fourth, 

 inverse rectangular, type (Fig. 6). The terminal increase in 

 liability to die may also be masked by cyclical variations in 

 mortality associated with breeding or wintering, but the pre- 

 sence of such an increase remains an essential requisite for the 

 demonstration of senescence in an organism. In many senescent 

 populations, such as the sheep and cavies in Figs. 11 and 12, 

 the survival curve in adult life is not so much rectangular as 

 arith-linear, a constant number of individuals dying during each 

 unit of time, the mortality necessarily decreasing as the supply 



21 



