208 
THE METHODS OF WIND MEASUEEMENT. 
work. There wwe several difficulties to overcome : one was the 
necessity for having the weights suspended to the main spindle, 
in order to avoid further friction, and since in a wind of twelve 
miles an hour the cups and spindle at the author’s disposal would 
only turn once in a second, this was a serious consideration ; 
another was the form of frame to be used so as to keep the 
apparatus weather-tight, and yet have a short spindle in order, as 
before, to avoid excessive friction ; a third difficulty was to 
devise an arrangement so that the motion of the pencil should 
be proportional to the change of speed of the cups — that is the 
ordinates of the curve described to the velocity of the wind. 
Figs. 1 and 2 shew two views of the experimental instrument 
in which these various points have been dealt with. Sensitive- 
ness has been attained by using fairly large weights and causing 
the sleeve carrying the pencil to move with the application of a 
very small force by having it very light and in contact with the 
lever by means of rollers. A motion of the pencil theoretically 
proportional to the change of speed was obtained by first 
calculating the position of the weights at various velocities, and 
then setting out the curve of the cams on which the rollers of 
the sleeve rest, the distance of the sleeve from zero (its highest 
point) being made proportional in each case to the calculated 
velocity. Fig. 3 shews some of the curves obtained on the 
night of November 2nd, 1881, on the roof of the author’s house, 
in which it is seen that the wind was rising during several 
hours, and at the same time its gusty nature is evident.^ A 
particular feature in this instrument is the mode of supporting 
I From the paper of Mr. Curtis, already referred to, it is clear 
that an instrument similar to this one in principle had been 
previously constructed by that gentleman, and very possibly by 
others. See Journal Meteorological Society, vol. vii., p. 212. Mr. 
Curtis, however, used a second spindle geared to the first, and even 
then found the action very sensitive. 
