20 THE FALL OF THE BIRTH-RATE 



values of Tail's coefficient, so calculated, for the 

 census years, and also (for control) the values of the 

 standardised legitimate birth-rates as calculated by 

 the method of Sir A. Newsholme and Dr Stevenson. 

 Calling the respective values in 1871 1000 we can 

 compare the results of these two methods. It will be 

 seen that they never diverge by as much as i per cent, 

 from each other. The greatest difference occurs in 

 1 88 1 and is 6 in the last digit, or 0-6 per cent. For 

 practical purposes they lead to the same results : and 

 what they show is this : 



In 1 88 1 the birth-rate gave a greater fall than the 

 fertility : the indication of the rate on married women 

 15-45 was about right. 



In 1891 the birth-rate, considered as a measure of 

 fertility, again gave too low a figure: the indication 

 of the rate on married women was also rather low. 



In 1901 the figure given by the birth-rate was about 

 right, that given by the rate on married women much 

 too low. 



In 1911 the figures given both by the birth-rate 

 and by the rate on married women were too low. 



It will be seen then that not even the proportion of 

 births to married women under 45 years of age can 

 be entirely trusted as a measure of fertility. In the 

 years immediately preceding the war it tended to give 

 a slightly exaggerated idea of the fall : some standard- 

 isation based on the age-distribution of the married 

 women is essential. 



We may take it, then, that the fall in fertility of 

 married women between 1871 and 1911 was some 

 28 per cent., and that up to 1911 at all events the fall 

 was taking place at accelerating speed. Reckoning the 

 percentage in each case on the value of Tait's co- 

 efficient at the beginning of the period, the fall in each 

 decade was as follows : 



