State of Experimental Acoustics, 303 



Bellows of Precision, 



The only bellows of precision that I know of is that of 

 Cavaille Coll*: it is expensive; the complete machine costs 

 £80. I will state shortly the principles on which the obtain- 

 ing of steady wind from the feed depends ; it will appear that 

 it is not necessary to increase largely the cost of the ordinary 

 bellows. 



The simplest form of supply is one feeder driven by the up- 

 and-down motion of a handle, the stroke overcoming also the 

 weight of the feeder. This arrangement discharges a volume 

 of air into the reservoir with a uniform velocity, which begins 

 and ends suddenly : the weight of the feeder always causes a 

 shock at the beginning and end of the stroke. The area of the 

 feeder is generally equal to the area of the reservoir, so that 

 the velocity imparted by the stroke to the top board of the 

 reservoir is half that of the lift of the end of the feeder. This 

 is the worst form of supply ; it is common in very small organs. 

 In all cases the reservoir must be made with one inverted 

 and one direct rib. This is well understood by organ-builders. 

 The next best form of feed is that common in English 

 organs. Two equal feeders, each occupying half the area of 

 the reservoir, are moved in opposite directions by up-and-down 

 strokes of a handle. The feeders balance and the shock is ma- 

 terially lessened. The velocity of the upper board is a quarter 

 the lift of the end of the feeder. But the stroke is still generally 

 made w T ith uniform velocity, beginning and ending suddenly. 

 The next improvement is the application of a pair of cranks, 

 axle, and fly-wheel to the previous combination. In this case 

 there are two discharges for each turn of the fly-wheel, whose 

 velocities follow the pendulum-law. The velocity alternates 

 between zero and its maximum. Great smoothness is here 

 attainable so long as the speed is low; but at high speeds the 

 variation of the velocity of supply will be objectionable. 



This inconvenience may be to a great extent obviated by 

 using three feeders instead of two, the three feeders being 

 worked by three cranks on a shaft, set at angles of 120°. The 

 velocity of supply in this case has maxima and minima at in- 

 tervals of 30° ; but these maxima and minima are in the ratio 

 of 2 : V3, or 1 : 0*866 nearly. This variation is less than 

 that obtained by using four feeders, but it is more frequent. 

 The value, however, is so nearly uniform that it is not thought 

 that any considerable gain would result by increasing the 

 number. If more exact uniformity were desired, six feeders 

 at intervals of 60° would give exceedingly small variations of 



* Comptes Rendus, 1863, i. p. 339. < Nature/ xviii. p. 381. 



