The Mature and Criteria of Senescence 



atory population can receive as detailed medical attention as 

 civilized man, but if such were possible, the life-table of mice 

 might then approach that for Western European man in 

 1953. 



Organisms which undergo senescence, as judged by the life- 

 table, also exhibit specific age, meaning an age at death which is 

 characteristic of the species when living under conditions ap- 

 proaching Bodenheimer's 'physiological' conditions. 



Some of the limitations of the statistical definition of sen- 

 escence have recently been re-stated by Medawar (1952). It is 

 obvious that any survival curve can be simulated by judicious, 

 or injudicious, choice of material. Tables based only on age at 

 death, a single arbitrary event, are open to serious criticism if 

 they are used as indices of a continuous process of declining 

 vitality. The shape of such a curve is a measure of many things, 

 including the genetic homogeneity of the sample. The incidence 

 of various risks itself varies between age groups: the statistical 

 appearance of senescence would, for example, be found in the 

 life- table of any population offish which was subject to frequent 

 fishing with a net of fixed mesh size. Selective predation cer- 

 tainly produces effects of this kind of nature. The increased 

 force of mortality among men of military age during a war is 

 not a manifestation of senescence. On the other hand, some 

 causes of mortality, such as cancer (Rutgers, 1953), have a 

 curve of incidence which parallels the total curve of mortality. 

 In employing the force of mortality as an index of senescence it 

 is essential, as we have seen, to exclude so far as possible external 

 factors which are not of random incidence in relation to age, 

 yet this cannot be done with strict logical consistency. In the 

 case of human life-tables, large secular changes in cause and 

 incidence of death may occur within an individual life-span, 

 while constitutional differences in rate of senescence between 

 individuals ensure that the genetical composition of the sur- 

 vivors at, say, age 60, is not representative of the whole cohort 

 under study. These sources of error are, in fact, capable of 

 avoidance or correction for most practical purposes, but they 

 must always be recognized in inferring senescence from any 

 life-table. 



Since there is no direct way of measuring the liability of an 



25 



