Refraction of Sound by the Atmosphere. 53 1 



tion above it. During a wind the air moves faster above than 

 below, therefore sound moving against the wind moves faster below 

 than above, the effect of which is to refract or turn the sound up- 

 wards ; so that the " rays " of sound, which would otherwise move 

 horizontally along the ground, actually move upwards in circular or 

 more nearly hyperbolic paths, and thus, if there is sufficient dis- 

 tance, pass over the observer's head. This explanation was pro- 

 pounded by Prof. Stokes in 1857, but was discovered indepen- 

 dently by the author. 



The paper then contains the description of experiments made 

 with a vie w to establish this explanation, and from which it appears 

 that : — 



1. The velocity of wind over grass differs by one half at eleva- 

 tions of 1 and 8 feet, and by somewhat less over snow. 



2. "When there is no wind, sound proceeding over a rough sur- 

 face is destroyed at the surface, and is thus less intense below than 

 above. 



3. That sounds proceeding against the wind are lifted up off the 

 ground, and hence the range is diminished at low elevations ; but 

 that the sound is not destroyed, and may be heard from positions 

 sufficiently elevated with even greater distinctness than at the same 

 distances with the wind. 



4. That sounds proceeding with the wind are brought down to 

 the ground in such a manner as to counterbalance the effect of the 

 rough surface (2) ; and hence, contrary to the experiments of Dela- 

 roche, the range at the ground is greater with the wind than at 

 right angles to its direction, or where there is no wind. 



On one occasion it, was found that the sound could be heard 360 

 yards with the wind at all elevations, whereas it could be heard 

 only 200 yards at right angles to the wind, standing up; and, against 

 the wind, it was lost at 30 yards at the ground, 70 yards standing 

 up, and at 160 yards at an elevation of 30 feet, although it could 

 be heard distinctly at this latter point a few feet higher. 



As might be expected, the effect of raising the bell was to extend 

 its range to windward, to even a greater extent than was obtained 

 by an equal elevation of the observer. 



These results agree so well with what might be expected from 

 the theory as to place its truth and completeness beyond question. 



It is thus argued that, since the wind raises the sound so that it 

 cannot be heard at the ground, by causing it to move faster below 

 than above, any other cause which produces such a difference in 

 velocity will lift the sound in the same way ; and therefore that an 

 upward diminution in the temperature of the air must produce this 

 effect ; for every degree of temperature between 32° and 70° adds 

 nearly one foot per second to the velocity of sound. Mr. Glaisher's 

 balloon observations * show that when the sun is shining with a 

 clear sky, the variation from the surface is 1° for every hundred 

 feet, and that with a cloudy sky o, 5, or half what it is with a clear 

 * Brit. Assoc. Eeport, 1862, p. 462. 

 2 M 2 



