1918.] _ The Fifth Indian Science Congress. Ixxxv 
area, I have taken for consideration that of South-West Bengal, 
i.e. the Burdwan and Presidency Divisions. I will not trouble 
you with the details of the processes employed, but will put on 
the screen! the curves showing the factors which seem to hove 
most influence on the fever rate of any year. They are— 
(1) the population at the beginning of the year, 
(2) the cost of rice in the previous year, 
(3) the area under cultivation in the actual year, 
(4) the rainfall in the actual year. 
The table of relations between these is: 
| 
| ever |. Pope | tre | anes 
Fever --| +100 | +055 | —0-35 | —040 | — 0-20 
Population .. + 0°55 | +100 | —0°65 | 0-00 +0102 
Cost ofrice | oss fogs b 5 ie | 40°15? +010? 
Cultivated area | — 0°40 | 0°00 | + 0°15? | + 1:00 | + 0°40 
Rainfall “+ — 0°20 | +0102. +0102 | +040. + 100 
Of these the probable error due to pure chance is about 
(fever)=0-65 (population) +0°15 (cost of rice) — 0°35 
(cultivated area) — 0-15 (rainfall) : . 
ere by writing any quantity, such as (fever), within brackets 
Wwe mean that we take the departure of the fever rate from the 
of fever: this is ses of cultivat- 
i : presumably because the proces ; 
ing the land leave fewer pools long stagnant for mosquitoes to 
1 See Fig. IX. 
