213 
as nearly as possible through the fifty points thus obtained 
I found the curve to take the course given in Fig. II. The 
numbers for the year ending September, 1876, treated in 
the same way, gave a similar curve, both curves slanting 
very decidedly in the opposite direction to the public-house 
curve, and would seem to indicate that the fewer public 
houses there are the greater on the average the number of 
apprehensions for druukenness. This method of examining 
the numbers seems justified in this respect, that grouping a 
large number of towns together will to some extent neutral- 
ise the effect of more or less stringent police regulations 
and other circumstances peculiar to individual towns. 
It would of course be absurd to infer from these curves 
that because there are more public-houses in certain towns , 
therefore there is less drunkenness. It would be far more 
likely that the fewness of public-houses, and the prevalence 
of drunkenness, arise from some common cause. It seemed 
possible, for instance, that when there was great prosperity 
the population would probably increase quickly, and through 
the difficulty in obtaining licenses, the public-houses would 
not keep pace with the increase of the population. At the 
same time, when there was great prosperity, there would be 
higher wages and more money to spend, and therefore more 
possibility of drunkenness. 
I therefore grouped the towns as before, in twenties, but 
this time taking the numbers showing the percentage in- 
crease in population during the years I86I-7I, and then 
drawing a curve nearly through the points thus obtained, it 
took the course shown in Fig. III. This, by altering the 
scale, might almost be made to coincide with the public- 
house curve in Fig. I. 
If the towns be now arranged in order of their percentage 
increase of population, during these ten years, from Canter- 
bury, Bath, and Ashton, with either no increase or a decrease 
to Huddersfield, with an increase of 217 per cent, the in- 
