Ixxxvi Journal of the Asiatic Society of Bengal. [N.S., XIV, 
breed in. The+0-40 presumably means that when rainfall in 
any year is scanty the peasant does not cultivate as much land 
as when there is plenty of rain for his crops. Let us now look 
with some care at the—0-35 which informs us of a decrease 
in fever next year due to an increase in the cost of food this 
year. At first sight this result is surprising, for we should 
expect that less rice to eat this year would lower vitality and 
so increase fever next year, not decrease it. But the explana- 
efficient is very often treated as if it gave the direct effect of one 
factor on another, it frequently does nothing of the kind: 
always gives the total effect, direct and indirect. We cannot 
get the direct effect except by forming the regression equation 
of the type that we have just considered: and we may 
misleading results unless that equation contains all the facto™ 
which produce appreciable effects. It will be remember 
that there was a coefficient of —0-65 between the cost of ne 
in any year and the population at the end of it: this relation” 
ship which is shown on the screen! seems at first sight 8° close 
=o 0 suggest that the population must be in dire POV.” 
married and the trade activity that the young peo his” 
were as anxious to get wedded as a starving man 18 eee 9 
food, a state of things which does not fit with a declining ee s : 
Tate. As a matter of fact the high price in any year ari 
relationship of —0°35 with the increase of population in t 
year, but—0-65 with that in the previous year : * 
are shown on the screen. Thus, if we can trust our data, 
! See Fig. X. 
