I(j8 The Rev. T. R. Robinson's Description of an improved Anemometer. 



That this ratio holds for the large anemometer, as well as the small, is ex- 

 perimentally shown by placing the latter beside the other, and counting the 

 revolutions made by it during eighty-eight of that. Fourteen such trials give, 

 with a mean velocity of wind = 15'6 feet, the ratio of the revolutions =4'12, 

 the inverse of their dimensions being 4'29. The difference is due to the large one 

 being above the other, and therefore getting the wind more freely. 



As this relation does not depend on the elasticity of the impelling fluid, it 

 should hold when the instrument is acted on by a stream of water, with the 

 advantage of being much less affected by friction. I tried this in a large mill- 

 course near Armagh, placing the small instrument in the central part of the 

 current, where the velocity was found by floats = TGIS feet. I obtained, 



4 With four hemispheres, . . . ?« = 2 972. 



With two, = 3-208. 



With three, equidistant = 3 '041. 



The trial with two was made rather from curiosity than from any depend- 

 ance on the result which might be obtained, as, when passing the line of centres, 

 the impelling force is so slight, that any eddy will produce a disturbance of the 

 motion. It would not change the mean much, but I think should be rejected. 

 That of the other two is 3'006, still probably a trifle too large, as the four are 

 preferable on the same grounds to the three. 



From all this I think we are warranted in laying down this law, that in a 

 horizontal windmill of this description, the centres of the hemispheres move 

 with one-third of the wind's velocity, except so far as they are retarded by 

 friction. 



This principle once established, its application is easy. Plate IV., fig. 1, shows 

 the external appearance of my anemometer, as it stands on the flat roof of the 

 dwelling-house. Its frame consists of four uprights, 3' by 2', and l^ 4' long ; 

 6^ 5' asunder below, 2^ 4' above. They support the strong frame B, in which 

 a diagonal carries the bearings of the axles C and D. The part H is sheathed 

 and roofed with plank (the roof covered with painted canvass) ; and it forms 

 a very convenient room for the self-registering apparatus. The copper funnel 

 F is attached to each axle, to prevent the entrance of wet. The great height 

 of this frame is necessary to clear the dome of the west equatorial, which rises 



