350 MR. S. M. JACOB ON 



This will be attempted with the help of the modern methods of treatment which 

 originating with Bravais and Quetelet owe their greatest development to the work of 

 Francis Galton and Karl Pearson. 



The application of these methods has not, so far as the author is aware, been 

 previously made to the data considered in this paper, or indeed to any similar data/ 

 so that it is hoped that, in spite of the fact that the statistics discussed are far from 

 being as complete or wide-reaching as is desirable, some at least of the results obtained 

 will be of value if only for purposes of comparison when fuller treatment becomes 

 possible. In any case, too, an advance will have been made by indicating what are 

 the more favourable lines on which to approach those problems whose practical solu- 

 tions appear to be of high utility. 



Some familiarity with the more elementary parts of the modern theory of statis- 

 tics has to be assumed for an understanding of the procedure to be adopted in this 

 paper, as it would unduly lengthen the argument to enter into a mathematical 

 explanation of all the processes employed, and such an explanation might well be 

 regarded as impertinent by those conversant with their use. 



At same time it is believed that even without a grasp of the ' exact ' meaning of 

 such terms as ' correlation ' , ' standard deviation ', c probable error ' and the like, the 

 reader may at least be able to obtain a qualitative idea of the nature of the results 

 reached, although the sooner such qualitative notions are replaced by definite quanti- 

 tative ones, the more hope there will be that statistical problems will receive adequate 

 statistical treatment. 



Similarly, the characteristic agricultural and metereological conditions which 

 prevail in India will be assumed to be known at any rate in outline. The different 

 modes of agriculture and the variations of climatic influence are of extreme import- 

 ance in their effect on the resultant harvest, but into these it is not proposed to 

 enter in detail as the statistics dealt with apply (except in the case of certain rainfall 

 data) only to a very limited area in the Panjab. 



It will be time to discuss the effect of special differences in agricultural operations 

 and physical circumstance when the applications of the present methods has been 

 made to many diverse and widely-separated tracts. Until such time it will not be 

 wise to try to generalise the formulae so as to make them applicable to other places 

 than those for which they are calculated. 



It is, however, quite certain that such generalisation is possible, but it must, of 

 course, be reached very gradually and by considering one by one all the contributory 

 causes. 



1 I have since been informed by Mr. G. Udny Yule, Honorary Secretary of the Royal Statistical Society, that the 

 similar problem for English crops has been treated by R. H. Hooker in the Journal of the Royal Statistical Society for 

 1907. Mr. Yule has kindly promised to send me a copy of this paper, but unfortunately I have not yet received it. , 



