V 



V 



162 Tlie Rev. T. R. Robinson's Description of an improved Anemometer. 



As their mean force vanishes when the condition of permanent rotation is at- 

 tained, if we equate it to cypher, we deduce 



This shows that if we neglect the term introduced by friction, the ratio of the 

 velocities V and v depends on the ratio of a and a' alone, being independent 

 of their absolute magnitudes and also of v. It is, therefore, independent of the 

 speed of the wind and the size of the machine. 



Calling this ratio m, and making the instrument register mv = V, the true 

 velocity of the wind = V+ u, u being the correction due to friction, we have 

 from (1) 



^(V+uy-~(a + a'){V + n)-2v'(a' + b')-f=0, 



^' V'--(n + a') V -2v'(a'+b')^0: 



2i TT 



whence 



\ m-K \a — a J } a — a ^ ' 



the positive root of which may be tabulated for a series of values of V. 



The constants of these equations must be given by experiment, and it is 

 not easy to obtain them satisfactorily, especially the most important of them, a 

 and a'. But for the unsteadiness of the wind,* both in force and direction, 

 we might attach hemispheres to some weighing apparatus, with the '•concave 

 and convex surfaces turned to the wind, and thus obtain absolute measures of 

 them. This, however, coidd only be done by connecting the two with a pair 

 of registers like those of Ossler's instrument, which would give the mean pres- 

 sure for a considerable period ; and such an apparatus is not at my command. 

 As, however, m depends on their ratio only, I found a method, which, though 

 disturbed by the same cause, is tolerably successful. Two hemispheres, similar 



* In illustration of this I may mention, that having placed two hemispheres on the arm, so 

 that both concaves faced the -wind (when, of course, they might be expected to remain in equi- 

 librium), they oscillated with considerable force through arcs of 90^; the distance between their 

 centres was 48-5 inches. 



