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



it self-registering.* In the third class, of which Lind's is the type, the pres- 

 sure of the wind is measured by the column of water, or some other fluid, which 

 it is able to support in an inverted siphon. 



All these are liable to the following objections. First, wind fluctuates, both 

 in velocity and direction, to an extent of which I had no conception till I en- 

 tered on these researches. Instead of being a uniform flow of air, it may be 

 likened to an assemblage of filaments moving with very unequal speed, and 

 contorted in every direction ; being in fact analogous to a river in flood, but 

 with its eddies and counter-currents considerably exaggerated. Now, assuming 



2dV dP 

 the common equation F' — mP, we find —r^ = -p-, or the relative variations of 



pressure are twice as great as those of velocity; a record of the latter will 

 therefore, be far less irregular. But the evil goes further ; for in the fluctua- 

 tions both of pressure and direction, the inertia of the moving parts of the 

 anemometer carries them far beyond the point of balance, and makes the mea- 

 sure of pressure inaccurate, partly by exaggerating the amount of its changes, 

 partly by the surface which receives the wind's impulse being at times wrongly 

 placed with respect to its direction. The magnitude of this cause of error may 

 be appreciated from these two facts, that I have seen Lind (the only pressure- 

 gauge which I possess) range in a few seconds from to 2 '6 inches ; and that in 

 some winds a free vane will oscillate through arcs even of 120". 



2. The velocity can only be deduced from the pressure by experiment : if 

 the relation between them be constant, this necessity is of little importance ; 

 but the fact is the reverse. In the case of the windmill-vanes, we have no in- 

 formation ; and it is evident that the law which connects these variables must 

 be very complex, in consequence of the wind which glances off the anterior 

 surface modifying the minus pressure. In Bouguer's instrument it is com- 

 monly assumed that the pressure is the weight of the column whose height is 

 that due to the velocity : there is, however, experimental ground for believing 

 that it is nearly twice as great.f The excess is caused by the minus pressure, 



* An anemometer of this kind, acting against a series of weights instead of a spring, was long 

 used by the late Mr. Kihwan, and is described in vol. si. of the Academy's Transactions. 



I See D'AuBnissQN, Hydraulique, p. 295. From De Buat's investigations it is not unlikely 

 that the velocities deduced from the records of Osslee's gauge are about one-third too great. 



