567 



which the particles of air are displaced. Instead of this, phi- 

 losophers had observed the pressure, which is only useful as a 

 means of giving the velocity. To this most of the anemome- 

 ters on record were destined. They may be reduced to three 

 classes. The first, extending from Hooke's to that which the 

 President has recently constructed, consist of windmill vanes, 

 made to face the wind by a vane or other contrivance, and act- 

 ing against some graduated resistance which measures the pres- 

 sure; the second, invented 100 years ago by the celebrated 

 Bouguer, long used by a former illustrious President, Mr. Kir- 

 wan, and recently re-invented and brought into extensive use 

 by Mr. Ossler, consists of a square plane, which, when exposed 

 to the wind, compresses a spiral spring; and lastly, those which 

 like Lind's, measure the pressure by the column of fluid 

 which it can support. All these are open to the following ob- 

 jections : — First, their indications are most iiTCgular. Wind is 

 not a uniform rush of aii* ; it is irregular to a degree which 

 he could scarcely credit when he began his experiments. A 

 river in flood, with its rapids, counter-currents, and eddies, 

 gives but a faint idea of it ; it may be likened to a bundle of 

 filaments moving Avith all possible motions and contortions. 

 Under such circumstances the pressure varies excessively. He 

 had seen Lind in three or four seconds range from to 3 inches, 

 and, after long watching, could form no guess as to the true 

 measure. In fact, the relative variations of pressure are twice 

 those of velocity. And this innate source of irregvdarity is 

 increased by the veering of the wind acting now full and now 

 obliquely. It is also exaggerated by the inertia of the moving 

 parts, which carries them beyond the place of balance. Se- 

 condly ; in none of them can the relation between pressure and 

 velocity be determined but by trial. In most, if not all, that 

 relation is not constant. He insisted on this, because velocities 

 deduced in the common way, from Ossler's gauge, were often 

 one-third too great. Thirdly ; it was of the utmost import- 

 ance that these instruments should be self-registering. Now he 



VOL. IV. 2 T 



