458 ON THE DETERMINATION OF THE TRUE STRENGTH 
on in the same direction from whatever quarter the wind may come, and increases 
in rapidity exactly in proportion to the strength of the wind. 
Mr Epcrworts had proved many years ago the increased effect of wind 
on a concave surface to a flat one, by taking a sheet of tin, and shewing that; 
when bent into so very concave a form as to present considerably less area than 
inits flat condition, to a current of wind blowing straight upon it, that still it ex- 
perienced a greater degree of pressure. But it was only very recently that he 
carried the principle farther into a practical form ; and the first notice of this is 
found in Dr Rozinson’s communication to the British Association at Southampton, ’ 
descriptive of the application to a Whewell’s anemometer of one Mr EpGEworTH’s 
horizontal vane wheels, where each vane consisted of a hemisphere, and the con- 
cavities being all turned in one direction, and experiencing one-third more resist- 
ance than the convex sides, the whole revolves from the concave as it were to 
the convex sides, at one-third the rate of the wind. 
This, then, seemed to be eminently the sort of anemometer for use on board- 
ship. 
No particulars of size or construction were given, only the important prin- 
ciple involved was mentioned ; and as the first instrument which I had made, did 
not prove sufficiently sensitive, I entered on a course of experiments to ascertain 
the best sizes of the machine, and shape of the floats; and being greatly assisted 
therein by the practical skill and ingenious methods of Mr Mixing, the artist em- 
ployed, was at last enabled to fix definitively on the instrument figured in the ac- 
companying plate as the one which combines the greatest number of advantages, 
and forms altogether the best standard, perhaps, that can be adopted. 
There are 4 hemispherical floats, 4 inches each in diameter, and 1 foot apart 
from centre to centre: an endless screw, on the axis of the vane-wheel, gives 
motion to a train of wheels and pinions, which serve to measure the number of 
revolutions made; 13 grain, in the centre of one of the floats, is sufficient to pro- 
duce motion. (See Plate XI.) 
If the instrument be made on too small a scale, then it will have to overcome so 
much larger a proportion of the friction of surface than a larger one, that the floats 
will not move at one-third the velocity of the wind ; but as much of this friction 
and resistance depends on the linear measure of the parts, while the motive power, 
which is as the area of the float, increases as the square of the dimensions, it is 
evident that increased velocity may be obtained by adopting a larger instrument. 
There is, however, a danger of passing the proper medium again in this way ; for, 
as some of the sources of friction increase according to the weight of the moving 
parts, or as the cube of the linear dimensions, these may soon surpass the motive 
power, which increases as the square only. There is another objection also to 
having a vane-wheel with much mass, or vis inertiw; for although we wish in 
this inquiry to get a mean, or the total sum, of all the little separate gusts of 

