





( 



calling 



f 

 av'- 



- 



^1= 



-- m" 



a. 



m' 



F 





1 



a 



m"- 



m''' 



a 



I 



i 



Robinson — On the Cup Anemoineter. 441 



Having three such, equations with different values of / and v, 

 we can, of course, find the three co-efficients with much greater pre- 

 cision. It is desirable that V should be unvaried for the three obser- 

 vations, to avoid its influence on the result. Here, also, taking V as 

 large as possible, several such triplets may be obtained ; and if they 

 all give the same results the co-efficients are independent of v. 



Should a be shown by experiment to be constant, the co-efficients 



By.. . . 



-, -, are easily obtained ; for two of the above equations give 

 a a 



■xir 



^,nn_^l^>_l (Yil.) 



a av'' 



If this work should show that additional powers of V or v must 

 be introduced into (II.), tn will not be constant, and the instrument 

 should be constructed to show not mv, but v ; and a table of V should 

 be computed with v as argument which would serve for all Anemo- 

 meters of the same type. 



Having fully examined this Anemometer, others should similarly be 

 tried, varying in the size of cups and arms ; and this will show whether 

 a and y are exactly as the area of the cups, whether the eddies caused 

 by one cup interfere with the motion of that which follows it, and 

 whether there be any maximum relation between the cups and arms, 

 and how y varies with the latter. 



All this, however, rests on the assumption that the V given by the 

 whirling machine acts similarly to an equal one of real wind. I 

 think that most of the defects of M. Dohrandt's apparatus may be 

 avoided, except the difference between the action of a quiescent and a 

 moving fluid. Yet if this were determined in water for a cup ex- 

 posed perpendicularly and at 45, for the concave and convex surfaces, 

 it is highly probable that we should be able to reduce with sufficient 

 precision the results of the machine to those of the wind. Should 

 any further experiments like those of Mr. Proude, already referred to, 

 be undertaken, I hope this question will not be lost sight of. 



There are two other differences ; the Anemometer must produce 

 L-ddies in the air which in the case of wind are swept away from the 

 instrument ; but when it travels in confined space they may con- 

 tinue till in the course of its revolution it returns into them. It is not 

 easy to predict their effect or ascertain how long they last. The 

 circular current established in M. Dohrandt's work will probably (if 

 produced at all) be insignificant in a freer space and better constructed 

 apparatus. It may be best examined by stopping the whirl, and at 

 the instant projecting smoke into the track of the Anemometer. Useful 

 inf oiTaation on this point might be obtained by attaching to the revolv- 

 ing arm a small rectangle of sheet iron inside the Anemometer so as 

 to increase its resistance and observing what change this produces on v. 



K. I. A. PROC, SEE. II., TOL. 11., SCIBKCK. 3 A 



