268 



REPORT 1876. 



of the spaces (considered more or less circular) over which the wind blew 

 with the force of a " strong gale." They therefore are not the entire areas. 

 But, apart from this, owing to incomplete information, the radii are not known 

 for each day, and hence the areas are only rough approximations. There is 

 no doubt, however, that they increased from 1856 or 1857 to 1860, decreased 

 from 1860 to 1867, increased from 1867 to 1872, and then decreased to 

 1875. 



On the whole, there was a similar progression in the duration of the 

 cyclones, the smallest number of days being in 1856, 1857, 1867, and 1875, 

 and the greatest in 1861 and 1870, 



The total areas, i. e. the products of the mean area of each cyclone by the 

 number of days it lasted, increased from 1856 to 1861, decreased from 1861 

 to 1867, increased from 1807 to 1872, and then decreased to 1875. 



It is to be remarked, however, that the total areas for the years 1860-62 

 were much greater than those for the years 1870-72. This may be owing 

 partly to the radii for the latter years having been underestimated. On the 

 other hand, the number of cyclone-days in the years 1870-72 was somewhat 

 greater than in the years 1859-61. 



liainfall. 



A sufficient number of rainfall returns for the years 1873-75 have not yet 

 been obtained ; but the annual mean rainfalls at seventy-seven stations from 

 1854 to 1863, and at seventy-two stations from 1864 to 1872, are given in 

 Table II., in which all the rainfall observations at my disposal have been 

 used, except a few Prussian and Mauritius ones, which would not have affected 

 the general resnlts. 



The Table shows that, with hardly an exception, the sun-spots and rain- 

 fall were both above or both below their respective averages in the same 

 years. 



By taking the longer period 1843-72, and expressing the amouuts of rain- 

 fall and sun-spots in percentages, we get the following results : — 



