1887.] H. F. Blanford— Rainfall of the Carnatic. 1 1 



nitude as to be of considerable economic importance, but differing in 

 different parts of India. 



I will first take the facts relating to Southern India. Instead of 

 dealing with the rainfall registers of one station only, a very feeble and 

 unsatisfactory basis for testing any question of the kind, I have taken 

 the rainfall registers of all the sudder and some sub- divisional stations 

 in the Carnatic, including under this designation the plain below the 

 eastern ghats extending from Tinnevelly to the mouth of the Kistna- 

 There are 38 stations representing 72,000 square miles, and the registers 

 extend over 22 years, two complete sun-spot cycles, viz., from 1864 to 1885. 

 On simply tabulating the average of each year it was obvious that 

 during the first 13 years with very slight irregularities the rainfall 

 varied very nearly with the sun-spots. Being already below the average 

 in 1864 it fell to a minimum in 1867, when it was 9'4 inches below the 

 average of 34 - 6 inches, these increased steadily year by year up to 1872 

 when it was at a maximum, being 11*5 inches above the average. With 

 a sudden fall to an average amount in 1873 and a recovery in the follow- 

 ing year to 7'3 inches above the average, it declined steadily for 3 more 

 years to another minimum in 1876, when it was 13*2 inches below the 

 average. After this it became more irregular, but with sundry oscilla- 

 tions it rose again to 11"6 inches above the average in 1884. The years 

 of minimum sun-spots were 1867 and 1878. The years of maximum 

 sun-spots 1870 and 1883 or the beginning of 1884. Thus the first mini- 

 mum year of rainfall was a year of minimum sun-spots, the second pre- 

 ceded the latter minimum by 2 years. The first maximum of rainfall was 

 3 years after the sun-spot maximum, the second coincided with it or fol- 

 lowed it by a few months only. These facts taken in conjunction with 

 the amplitude of the oscillation amounting to considerably more than 

 half the average rainfall of the province are striking enough. 



But it remained to ascertain how far this is a real cyclical oscilla- 

 tion, to compute the amount for each year as it would be determined by 

 such a cycle, and to compare this year by year with the amount actually 

 recorded. Only in this way, as General Strachey pointed out, could we 

 ascertain how much of the result is due to the operation of a law and 

 how much due to other and non-compatible agencies. 



By the application of Bessel's well known harmonic formula it is 

 quite feasible to ascertain from any given series of figures what, accord- 

 ing to the theory of errors, is the most probable value of any cyclical 

 oscillation underlying them, when the length of the cycle is approxi- 

 mately known. This is so in the present case. The cycle of the 

 suu-spots has been determined by Wolf to be 111 years. For conveni- 

 ence, as I have only two cycles to deal with, I have taken it at exactly 



