166 The Rev. T. R. Robinson's Description of an improved Ane7nometer. 



I found to fail, partly from tlie difficulty of eliminating the action of the wind, 

 but still more from the fact, that a carriage drags loith it a quantity of air, 

 so that for many feet from it the anemometer does not feel the full effect of 

 the motion. At low speeds, and on days of calm, I have got results which 

 agree with that given by other methods, but more frequently the discord- 

 ance destroys all confidence in it. The aerial log proposed by Sir W. S. 

 Harris in the report at Plymouth, could not be applied, on account of the 

 lofty position of my instrument ; but I tried one far more delicate, by ex- 

 ploding small charges of powder at it, while my assistant noted the time re- 

 quired by the little globes of smoke (which in dry weather are not dissipated 

 for many seconds) to traverse 150 feet. But the irregularity of the wind's 

 motion makes aU such trials unsatisfactory, and I got the most discordant 

 results, the reason of which was evident by watching the track of the smoke ; 

 it rose, descended, twisted in eddies, and even occasionally came back many 

 feet against a stronar breeze. But in addition it can only give the movement 

 of that one part of the current which it occupies, while the anemometer shows 

 those of all that pass it in the same time, which are essentially distinct. I may 

 add, that the impossibility of obtaining accurate measures of velocity by such 

 means, was long siace pointed out by Mr. Brice.* 



The plan which succeeded consists in applying the whirling apparatus to 

 carry the anemometer, as in the annexed figure. The anemometer has four hemi- 

 spheres; it is similar to 

 the actual one, and about 

 a fourth of its dimensions: 

 the distance AB is 45 "6 

 inches, and as the diame- 

 ter of the hemispheres is 

 only 3 inches, we may as- 

 sume the velocity of their 

 centres to represent that 

 of the wind. C is a coun- 

 terpoise. I found that in Fig. 2. 

 this case the rotation produced no important outward current. The machine 



* Phil. Tra ns., 1766. 



