the Filar Anemometer. 83 



If the velocity of air normally to the wire is registered, 

 three wires at right angles to each other would register the 

 respective velocities in three cardinal directions. From this 

 the actual direction of the gust is at once obtainable as the 

 resultant of the three components. 



It is necessary, however, before a like deduction can be 

 accepted with confidence, to further elucidate StrouhaVs law 

 with data referring to the oblique passage of a wire through 

 the air. For just what will occur under these circumstances 

 cannot certainly be foreseen. It is also desirable to obtain 

 accurate data for the effect of temperature on the phenomenon, 

 an inquiry which Strouhal only carried far enough to appre- 

 hend that the pitch of a note, other things being equal, is 

 depressed with increasing temperature. 



From a theoretical side, the subject has recently been 

 elucidated by Lord Kelvin*. 



In constructing my whirling table I aimed at quantitative 

 decision on these questions. I therefore made a whirling 

 machine one end of the axis of which was pivoted to the 

 ceiling of a high room, and the other end rigidly attached to 

 the spindle of an ordinary whirling table about a metre above 

 the floor. The crank-wheel was placed at some distance, so 

 as to allow two horizontal arms, about 1*4 metre long and 

 1*9 metre apart in the same vertical plane, to rotate freely. 

 Clamp-screws at the end of these arms carried the wire to be 

 tested, vertically stretched and 1*9 metre long. 



The whole framework was made of gas-pipe and snugly 

 screwed together. It was therefore possible to move the 

 horizontal arms so that the vertical planes through them 

 would subtend a given (small) angle, and the (elongated) 

 wire be rotated through the air obliquely to the line of motion. 

 Due care had to be taken not to carry the obliquity too far. 



The customary electric brushes were added at the axis of 

 the whirling arm, and thus the rotation could be timed by a 

 chronograph pendulum of the simple kind sketched in a pre- 

 ceding paper f. Contact brushes for other electric apparatus 

 (ef. § 3) were also supplied. 



On trial, however, the whirling arm at the higher speeds 

 developed a weakness. It was found to yield seriously under 

 the strain of rotation, so that I did not feel sure of its con- 

 stants. I therefore abandoned further attempts at solving 

 the subsidiary questions just specified, for the present. 



3. A wire singing on the housetops, however, is as yet no 



* < Nature,' 1. pp. 524, 549, 573, 597 (1894); see particularly pp. 524 

 and 525, where the whistling- of a strong wind is discussed, 

 f Barus, American Journal, xlviii. p. 396 (1894). 



G2 



