418 WRITINGS OF JOSEPH HENRY. [1856 



beams in this case were merely in a single plane, perpen- 

 dicular to the mirror, and passing through the ear, while to 

 the focal point of the spherical mirror a solid cone of rays 

 was sent. 



The reflection from the cylindrical mirror forms what is 

 called a caustic in optics, while that from a cylindrical mir- 

 ror gives a true focus, or in other words collects the sounds 

 from all parts of the surface and conveys them to one point 

 of space. These facts furnish a ready explanation of the 

 confusion experienced in the Hall of Representatives, which 

 is surmounted by a dome, the under surface of which acts as 

 an immense concave mirror, reflecting to a focus every sound 

 which ascends to it, leaving other points of space deficient 

 in sonorous impulses. 



Water, and all liquids which offer great resistance to com- 

 pression, are good reflectors of sound. This may be shown 

 by the following experiment. When water is gradually 

 poured into an upright cylindrical vessel, over the mouth of 

 which a tuning-fork is vibrated, until it comes within a 

 certain distance of the mouth, it will reflect an echo in 

 unison with the vibration of the fork, and produce a loud 

 resonance. This result explains the fact, which had been 

 observed with some surprise, that the duration of the reson- 

 ance of a newly plastered room was not perceptibly less than 

 that of one which had been thoroughly dried. 



There is another principle of acoustics which has a bear- 

 ing on this subject. I allude to the refraction of sound. It 

 is well known that when a ray of sound passes from one 

 medium to another a change in velocity takes place, and 

 consequently a change in the direction or a refraction must 

 be produced. The amount of this can readily be calculated 

 where the relative velocities are known. In rooms heated 

 by furnaces, and in which streams of heated air pass up 

 between the audience and speaker, a confusion has been 

 supposed to be produced and distinct hearing interfered with 

 by this cause. Since the velocity of sound in air at 32° of 

 Fahrenheit has been found to been 1090 feet in a second, 

 and since the velocity increases 1'14 feet for every degree of 



