126 Mr. R. H. M. Bosanquet on the Theory of Sound. 



removed — say half an incli of water. The motion in the sur- 

 rounding air was explored first with a candle-flame, and then 

 with a pendulum formed by a piece of light fluff attached to a 

 thread. The latter means proved far the most sensitive. The 

 motion consisted of a forward blast occupying a cone of ver- 

 tical angle 20° to 30°, the outer air streaming toivards the cone 

 from all sides. The inclination of the stream to the axis of 

 the tube is approximately half that of the radius vector from 

 the end of the tube to the point — the inclination of the stream 

 being nil at points close to the tube itself, about 45° at points 

 in the plane of the mouth, and nearly at right angles to the 

 prolonged axis of the tube where the stream joins the outward 

 blast. The phenomenon is that familiar as the origin of the 

 steam blast, and in connexion with a well-known problem in 

 hydrodynamics. 



When the flame was placed close to the pipe, a short dis- 

 tance behind the mouth, it curled over and bent towards the 

 mouth in such a way that it was hardly possible to believe 

 that the air was not being drawn in. A tube w^as then used 

 disconnected from the bellows, and the air alternately drawn 

 in and blown out while the flame was in the above position ; 

 the movements of the flame, when suitably placed, were undis- 

 tinguishable in the two cases. It is therefore clear that we 

 have here a source of transformation, in the case of musical 

 pipes with open ends, by which the octave of the note of the 

 pipe must be formed in the surrounding air. The amount of 

 transformation due to this cause, however, is small ; for the 

 currents of this description are very faint. 



The case of a tube terminating in a plane or flange was 

 then examined. A brass tube about 2 feet long and 9 inches 

 in diameter was fitted with a flange 6 inches in diameter, 

 through a hole in the centre of which the end of the tube 

 passed. The other end of the tube fitted into the sound-board 

 of the acoustic bellows, by means of which currents of air 

 could be driven through the tube. Burning substances which 

 gave oft' dense smoke were placed on the flange ; and the course 

 of the smoke was watched apainst a dark backoround. The 

 issuing air diverged in a cone with vertical angle 20 to 30°. 



When the blast was strong enough, a steady horizontal cur- 

 rent was maintained along the surface of the flange directed 

 towards the issuing cone ; the smoke was carried in by this 

 current and mixed with the issuing air, which then formed a 

 well-defined cone. There was here a source of inaccuracy in 

 the upward tendency of the smoke ; there could, however, be 

 little doubt that the currents of the outer air preserved a direc- 

 tion nearly at right angles to the prolonged axis of the tube. 



