Ixxxviii Journal of the Asiatic Society of Bengal. [N.S., XIV, 
Here the probable errors vary from 0:07 to 0:1 and I find 
(fever) = 0°5(rain of same year)—0°15(rain of previous 
year) + 0 35 (prices). 
o the direct effect of rain in the previous year appears 
practically negligible, and its effect is brought about through a 
tics you can prove anything and that the method is not 
practical. But it appears to me obvious that we have evety 
right to be misled if we trust a man’s application of statisties 
to a problem, say, in economics, if he is not familiar both with 
statistical methods and with the principles of economics. If 
, uild a big bridge you call in a man who understands the 
calculation of stresses and strains as well as the methods of 
ndling steel: and only when you have done that have youé 
right to expect success. Further, although he makes calewla- 
tions you don’t dismiss him for being unpractical, because 
everybody knows that mechanical problems have been solved 
by mathematical processes for a century or more. But by yout 
self-restraint you do not forfeit the right, which long custom 
has given to all of us, to call any method unpractical either if 
it is not purely rule of thumb, or if it is in the slightest degree 
difficult to understand, or, most important of all, ifit ismeW. 
As Huxley remarked, the practical man is the man who praé 
tises the errors of his forefathers. 
_ Lhope that statistical methods may before long be recog: 
nized as essential for efficiency for the following reasons. F irst, 
a table of data covering, say, fifty years gives any intelligent 
man the same advantage as if he had carefully watched the 
conditions for fifty years and had a perfect memory : secont’s’ e 
employing a draughtsman to plot these data will su, : 
relationships which will distinguish direct from indirect effect 
and could be got in no other way. 
In conclusion I would urge on those of you who have ae 
control over the collection of statistics to exercise all the ~ i 
within your power in order to secure accuracy, and so by iy 
ing up invaluable materials for their use to deserve the gratl 
tude of posterity. 
