6 Crkiav, Measurement of Air Velocities. 



in diameter, and the outlet 9 in. in diameter. An air circuit 

 was made up of an isosceles triangle of 9 in. tinned iron 

 piping. At the junction of the two equal sides was placed 

 the fan, drawing from one side and delivering into the 

 other, the length of these sides being 29 ft. The length of 

 the base of the triangle was 33 ft. 8 in., and this was joined 

 to the sides by two bends (forming the angles of the base) 

 each 10 ft. in diameter. It is evident that somewhere near 

 the middle point of the base of such a circuit the true pres- 

 sure must be atmospheric, and accordingly a slot, 7/16 in. 

 wide and 2 ft. long, was provided at the middle of the base 

 and fitted with a sliding cover so that various gauges could 

 be inserted and moved for a distance of ift. on either side 

 of the central point. It was argued then that any gauge 

 reading zero at one point, and with one velocity, should 

 read zero at all velocities at that point if there were no 

 induction. 



This argument is not quite sound unless the whole 

 air circuit be truly symmetrical, which in practice can 

 never be the case. If, however, the circuit can be adjusted 

 so that under any circumstances an increase of velocity 

 tends to move the zero point towards the outlet of the 

 fan, then induction and zero-movement are of the same 

 sign with increasing velocity ; and consequently that tip 

 which with increasing velocity shows the smaller change 

 of reading will be the least affected by induction. 



It is easy by slides at the inlet and outlet of the fan 

 to arrange the conditions under which the zero will move 

 slightly as required above, and this was done for all the 

 experiments. 



Instruments. It will be clear that for such measure- 

 ments manometers were required of a far more delicate 

 nature than those used by Heenan and Gilbert. The 

 apparatus selected was a Krell micromanometer having two 



