l66 TIME 



age intervenes numerically in the estimation of the duration 

 of a year, the curve representing the value of one year for each 

 age can be drawn by simply plotting as ordinates the values 

 of the reciprocal of the ages plotted as abscissae. We have 

 stated that the year of a five-year-old child seems long to him 

 because it represents one-fifth of his existence, or o-20. To a 

 man of twenty, a year (one-tzventieth of his existence or 0-05) 

 will seem shorter in the same proportion as i to 4. For a man 

 of fifty it will be only one-fiftieth or 0-02. To the latter, time 

 will seem to flow ten times faster than to the child of five. 

 We thus obtain the solid curve of Fig. 30. It is a very simple 

 curve, an equilateral hyperbola, the branches of which are 

 asymptotic of the axes of co-ordinates and which answers to 



the elementary equation: xy=i or y=-'. Let us now super- 



X 



pose this curve on that of Fig. 28, p. 155. The values of the 

 abscissae (age) coincide. For the ordinates we must keep the 

 scales comparable; namely the figures, although different, 

 must be proportional to those of the preceding diagram. We 

 obtain the dotted curve of Fig. 30 and we notice at once a 

 similarity which, if not absolute, is nevertheless very remark- 

 able. There is no complete coincidence, but it is indisputable 

 that the aspect is nearly the same between the ages of fifteen 

 and sixty. We must not forget that, as we said on p. 154, our 

 experimental figures are doubtful before twenty and after 

 fifty years of age. 



Does this mean that the reactions registered by age inside of 

 us, faithful witnesses of the flight of years, evolve according to 

 a hyperbolic law? We cannot affirm this, for it would imply 

 the admission that no other process enters into play, and 

 nothing authorizes at present the formulation of such a postu- 

 late. We may, however, note that such curves are not infre- 

 quent in chemistry: the dissociation curve of Nernst (iso- 

 thermal dissociation) is precisely an equilateral hyperbola. 

 This curve gives the number of ions produced in a solution 

 by dissociation, as a function of the concentration in mole- 

 cules. It shows that the phenomenon decreases when the 



