1876.] Vaynations of the Barometer in India. 35 



From these results it appears that the maximum happened — 



h 



Earlier at Simla than Madras .... 7 times on an average of 1 6 

 „ ,, Singapore . . „ 16 



„ Madras than Simla . . 3 ,, „ 18 



Singapore. . 4 ,, „ 14 



„ Singapore than Madras 6 „ ' 16 ' 



Simla „ „ 4 „ „ 22 



The mean interval between the occurrence of a maximum at any two 

 of the stations is 16^*7. The maximum happened first at Simla 43-3 

 times per cent., first at Singapore 33-3 times percent., and first at 

 Madras 23*3 per cent. 



We find that the minimum happened — 



h 



Earlier at Simla than Madras .... 4 times on jui average of 4 



,, Singapore . . 4 „ ,, 8 



Madras than vSimla. ... 4 ,, 11 



Singapore ... 3 „ „ 8 



Singapore than Madras. 5 ,, ■ „ 8 



„ Simla . . 4 ,, „ 18 



The mean interval between the occurrence of the minima at two 

 stations is 9^*6 ; there was little difference in the number of cases first 

 at any station. 



On the whole the number of cases considered is not sufficiently great 

 to give any weight to the conclusion that the maximum or minimum 

 happens oftenest first at any of the stations*. 



There were four cases (6, 8, 14, and 16) in which the maxima 

 occurred at all the three stations within a space of less than 16 hoin-s. In 

 these cases the mean intervals betvreen the occurrence of the maxima — ^ 



At Simla and Madras .... = 6'2o hours, 



,, Singapore . . = 10*75 „ 



At Madras and Singapore .= 12-00 „ 



Or the mean interval for any two of the three stations was 9*7 hours. 



There were four cases (2, 7, 15, and 18) in Avhich the minima 

 occurred at all the stations within a space of 6 hours. In these cases 



the mean intervals were for 



Simla and Madras = 4*0 hours. 



„ Singapore . . . , = 3*4 „ 



Madras and Singapore . . . . = 3'9 „ 



* All the cases have been employed ; if we were to reject the cases for which the 

 times are not certain to a few hours (indicated by ?) the liaal results for the mean 

 interrals Avould be little affected; the mean inter\al fur tiie maxima and mimma 

 would be l.^'l and lO-,") hours respectively. 



