1 60 The Rev. T. R. Robinson's Description of an improved Anemometer. 



paratus, its motions round the vertical axis will not exactly correspond with the 

 oscillations of the wind ; and very trifling variations in the angle of the vanes 

 will make a great variation in their speed. The fourth condition excludes those 

 horizontal windmills wliich act by a moveable screen. Of the remainder, in one 

 class the vanes are made to turn during the revolution, so as to present a duni- 

 nished surface to the wind while returning against it ; these are objectionable, 

 because the necessary machinery is liable to derangement, and involves much 

 friction, which will vary during a long period of working, and change the space 

 unit. There remain then those only in which the vanes are curved, so as to be un- 

 equally resisted on their opposite surfaces. Of these, the most elegant in prin- 

 ciple and definite in action that I know, was suggested to me many years ago 

 by INIr. Edgeworth. Its vanes are hollow hemispheres, whose diameters coin- 

 cide with the arms that support them ; the action on their concave surfaces ex- 

 ceeds that on the convex so much, that the machine is capable of being used as 

 a motive power with considerable advantage ; its simplicity of form is such 

 that, without very great exactness of workmanship, similarity of action can be 

 attained ; and it combines great lightness with strength sufiicient to resist very 

 severe gales.* 



The relation between the velocity of its vanes and that of the wind can be 

 determined satisfactorily, in the actual state of hydrodynamics, only by experi- 

 ment. In this instance, however, the problem is so modified by the antagonism 

 of the returning vanes, that theory gives not merely the law which connects 

 them, but a close approximation to their ratio, and the correction due to 

 friction. 



Let AH be an arm of the machine, bearing the hemispheres AIB, DKH, and 

 revolving in the direction of the arrows, so that the velocity of their centres = v.f 



* In the gale of December 15, 1848 (the anemometer diagrams of wliicli are among the speci- 

 mens exhibited to the Academy), the space recorded during the hour 2* •3'' = 61 '5 miles, but during 

 2^ minutes it is = 4'27, which gives 102'5 miles per hour for the velocity of that squall. Short as 

 it was, it did much damage in the neighbourhood, but the instrument was unhurt. A still heavier 

 gust is recorded in the diagram of the cyclone of March 29, 1850, where the velocity is nearly 130 

 miles per hour for 3 minutes. 



f At least this point is assigned as the centre of effect by the common theory ; it may, per- 

 haps, be a little further out in the concave. 



