310 Mr WHEWELL-S RESULTS OF OBSERVATIONS 



registering both the quantity and direction of the wind; and however 

 imperfect its construction may yet be, it must give some approximation 

 to the quantity which it is our object to measure, and must thus afford 

 the means of a better estimate of the mean direction for a year (or 

 for any other time) than has hitherto been possible. 



It is obvious that the mode of obtaining the mean direction of the 

 wind for any time would be to resolve each partial wind into its com- 

 ponent parts E. and W. and N. and S. The sum of all the west com- 

 ponents, subtracting the east elements, give the effective west wind ; 

 and the sum of all the south elements, subtracting the north elements, 

 give the effective south wind. The magnitude and proportion of these 

 two effective winds compoimded will give the magnitude and direction 

 of the effective wind, between west and south, which belongs to the 

 whole time. And the same may be said of any other cardinal points. 



The reduction of any wind to these cardinal directions is of course 

 to be performed by considering it as the hypotenuse of a right-angled 

 triangle, and here the multipliers by which the reduction is to be per- 

 formed are easily found. We may take fractions which are sufficiently 

 accurate, and yet simple enough to be easily used. Thus the inter- 

 cardinal winds, NE, SE, SW, NW, are reduced to the cardinal direc- 



tions N, S, E, W, by multiplying by — . The subordinate winds 



NNE, ENE, ESE, SSE, SSW, WSW, WNW, NNW, are reduced 



12 4 



to the cardinal directions by the multipliers — and — ; thus a wind 



NNE 65, is equivalent to N 60 and E 26. The oblique winds N by E, 



2 

 &c. might be reduced in the same manner by the multipliers — and 



^— or 1 . But these last I have not used. 



100 100 



I annex calculations made in this manner for the months of January, 

 February and March, 1837; in which I have resolved the days into 

 periods during which a certain group of neighbouring winds were pre- 

 valent. Thus from January 1 to 14, the prevailing winds were SSW, 

 SW and WSW ; from January 26 to 30 they were NE, ENE and E. 



