284 Mr. T. C. Porter on a Method of 
discontinuous, each condensation and rarefaction being at- 
tended by a corresponding increase and diminution of the 
rate at which the gas and air are burning, and in many cases 
it is easy to convince oneself that this is really the case by 
viewing the flame in the rotating mirror, when the ap- 
pearance of the flame is similar to that of ‘the flame in ex- 
periments such as the chemical harmonicon or Koenig’s 
vowel flames ; indeed the explanation of the effect of the 
sounds upon the flames in those experiments, and in those to 
be described, is very similar. In the well-known experi- 
ments just mentioned, however, and in Tyndall’s and Ray- 
leigh’s work on sensitive flames, the flames themselves do not 
sensibly reinforce the sounds to which they are sensitive. 
In the chemical harmonicon, for example, in which the flame 
burning in an open tube may be made to give a loud note 
by a weak note of the same pitch sung outside the tube, it is 
the vibration of the air in the tube to which the loudness is 
due, and the note itself is chiefly determined by the dimen- 
sions of the tube. In my experiments, however, each sound- 
wave determines what may be called an explosion, which 
may give to the air a pulse of very much greater amplitude 
than that of ‘the corresponding sound-pulse, and the succes- 
sion of these explosions may thus reinforce the sound 
enormously. Three of the conditions for the maximum 
effect are obviously (@) that the mixture of gas and air shall 
be that for maximum force of explosion, (6) that each ex- 
plosion shall consume as great a mass of the explosive mixture 
as possible, and (c) that the form of the flame during the 
explosion shall be such as to spread the disturbance through 
the air as advantageously as possible. 
I have used for the source of the sounds an ordinary 
‘‘Home” Edison-Bell phonograph, with the “ reproducer ” 
sold to be used with the instrument. In this, as every one 
knows, the roughnesses of the “ record ’’? make a rod vibrate 
up and down, and these vibrations are communicated purely 
mechanically to a thin disk of glass or mica, which in turn 
transmits them to the air on the side of the disk remote from 
the rod; the aerial disturbances are then conducted by means 
of a tube, usually to a trumpet, but in the experiments here 
described the reinforcement of the sounds is obtained by the 
combustion of coal-gas and air, and the method in which I 
have obtained the best results will be next described. The 
arrangement can most easily be understood from fig. 1. 
R is the “record,” DD is the vibrating disk, T is the short 
metal tube on the upper side of the ‘ reproducer,” which is 
generally connected by a short piece of tubing with a trumpet. 
