204 CERTAIN LAWS OF VARIATION. 



the rate of growth at the time. To what extent, how- 

 ever, does a retardation or acceleration of growth pro- 

 duced at any particular period of development persist 

 to the adult stage? It is quite conceivable that such 

 an effect, produced in the earliest stages, may be en- 

 tirely compensated for by a subsequent variation of the 

 growth in a reverse direction, and leave no trace be- 

 hind. On the other hand, it is conceivable that it may 

 persist unchanged and undiminished. In all proba- 

 bility, however, it does neither the one or the other. 

 The initial effect gradually diminishes with progress of 

 development, but never entirely disappears. An effect, 

 once produced, can never be wholly eradicated. Upon 

 this point some of the data given by Minot are again 

 most pertinent. He calculated the arithmetical mean 

 error in the weights of more than a hundred guinea- 

 pigs, which were bred under as normal and favourable 

 conditions as possible, and weighed at short intervals 

 from the day of birth up to the end of their second 

 year. During the first fortnight of growth, the indi- 

 vidual weights were on an average found to vary from 

 the mean weight by 19.0 per cent, in male guinea- 

 pigs, and 17.2 per cent, in females. Between the 

 70th and 140th days this variability had sunk to 

 12.7 per cent, for males, and 14.7 per cent, for fe- 

 males. During the 145th to 215th days, it fell still 

 further to 7.5 per cent, for males, and 13.0 per 

 cent, for females. At later periods it remained con- 

 stant in females, but increased again somewhat in 

 males. Roughly speaking, therefore, one may say that 

 with progress in development from the earliest to the 

 latest stages, the variability becomes nearly halved. 



