370 MR. S. M. JACOB ON 



The distribution of rainfall and crop with the regression lines is shown in the 

 Fig 15- 



§5. Correlation of the matured crop with other than the six-monthly periods 

 dealt with previously. 



The beneficial effect of heavy rainfall in September on the Rabi harvest, and its 

 frequently adverse influence on the Kharif crop, has already been referred to. 

 It seemed desirable, therefore, to exclude it from computation in the rainfall correlated 

 with the latter harvest and to include it in the period to be correlated with the 

 former. 



I have taken only the Tahsils Sialkot and Zafarwal which have the heaviest rain- 

 fall and whose correlations appeared to be probably the nearest to the truth. 



The following results were found : — 



Table showing the correlations of the Kharif matured area with the total rainfall 

 from April to August, both inclusive, and of the Rabi matured area with the total 

 rainfall from September to March, both inclusive, for unirrigated lands. 



Kharif. Rabi. 



Total unirrigated areas of Sialkot Tahsil .. -4161 + -125 -8281 + -047 



(•4693 + -ii8) (6309 + -093) 



Total unirrigated area of Zafarwal Tahsil .. -5045 + -H2 7215 + -074 



(•5174-Kiio) (-6532 + -086) 



Total unirrigated area of 30 selected villages -3939 + -127 '7550 + '065 

 in Zafarwal. 



(•4402 + -122) (7207 + -072) 



For purposes of comparison the correlations for the rainfall taken in each of the 

 six-monthly periods, April — September, October — March, are entered in round brackets 

 below the corresponding correlations. 



The results are of considerable interest, seeing that in spite of the large probable 

 errors, they all tend to corroborate the existence of the effects noted. The increase 

 in the Rabi coefficient is particularly marked, and the change in the Kharif 

 coefficient though small is consistently a diminution. 



It would, of course, be totally erroneous to argue from these figures to the effect 

 that all rain in September, whatever its amount and distribution, is beneficial to both 

 the autumn and to the spring crop. All that can be said is that rain, of the 

 quantity and mode of distribution that occurred during the period 1887-8 to 

 1906-7, appears on the whole to have done some slight good to the former crop 

 and considerable good to the latter. 



As a matter of fact, though it is difficult to see how rainfall in Septem- 

 ber, unless absolutely overwhelming, could harm the Rabi harvest, the case of 

 September rain causing harm to the Kharif crop is quite easy of conception. The 

 above results show that the contradictory of this latter proposition is more generally 

 true. 



