SCIENCE. 



245 



sound, would result from a periodical communication 

 and abstraction of heat, and he says : " We may conclude, 

 I think, that there is, at present no reason for discarding 

 the obvious explanation that the sounds in question are 

 due to the bending of the plates under unequal heating." 

 (Nature, xxiii, p. 274). Mr. Preece, however, seeks to 

 prove that the sonorous effects cannot be explained upon 

 this supposition ; but his experimental proof is inade- 

 quate to support his conclusion. Mr. Preece expected 

 that if Lord Rayleigh's explanation was correct, the ex- 

 pansion and contraction of a thin strip under the influ- 

 ence of an intermittent beam could be caused to open 

 and close a galvanic circuit so as to produce a musical 

 tone from a telephone in the circuit. But this was an 

 inadequate way to test the point at issue, for Lord Ray- 

 leigh has shown (Proc. of Roy. Soc, 1877) that an aud- 

 ible sound can be produced by a vibration whose ampli- 

 tude is less than a ten-millionth of a centimetre, and cer- 

 tainly such a vibration as that would not have sufficed 

 to operate a " make-and-break contact " like that used 

 by Mr. Preece. The negative results obtained by him 

 cannot, therefore, be considered conclusive. 



The following experiments (devised by Mr. Tainter) 

 have given results decidedly more favorable to the theory 

 of Lord Rayleigh than to that of Mr. Preece : 



1. A strip (A), similar to that used in Mr. Preece's 

 experiment was attached firmly to the centre of an iron 

 diaphragm (B) as shown in Fig. 5, and was then pulled 

 taut at right angles to the plane of the diaphragm. 

 When the intermittent beam was focussed upon the 

 strip (A), a clear musical tone could be heard by apply- 

 ing the ear to the hearing tube (C). 



This seemed to indicate a rapid expansion and con- 

 traction of the substance under trial. 



But a vibration of the diaphragm (B) would also have 

 resulted if the thin strip (A) had acquired a to-and-fro 

 motion, due either to the direct impact of the beam or to 

 the expansion of the air in contact with the stiip. 



2. To test whether this had been the case an addition- 

 al strip (D) was attached by its central point only to the 

 strip under trial, and was then submitted to the action 

 of the beam, as shown in Fig. 6. 



It was presumed that if the vibration of the diaphragm 

 (B) had been due to a pushing .force acting on the strip 

 (A), that the addition of the strip (D) would not interfere 

 with the effect. But if, on the other hand, it had been 

 due to the longitudinal expansion and contraction of the 

 strip (A), the sound would cease, or at least be reduced. 

 The beam of light falling upon the strip (D) was now 

 interrupted as before by the rapid rotation of a per- 

 forated disk, which was allowed to come gradually to 

 rest. 



No sound was heard excepting at a certain speed of 

 rotation, when a feeble musical tone became audible. 



This result is confirmatory of the first. 



The audibility of the effect at a particular rate of in- 

 terruption suggests the explanation that the strip (D) had 

 a normal rate of vibration of its own. 



When the frequency of the interruption of the light 

 corresponded to this, the strip was probably thrown into 

 vibration after the manner of a tuning fork, in which 

 case a to-and-fro vibration would be propagated down 

 its stem or central support to the strip (A). 



This indirectly proves the value of the experiment. 



The list of solid substances that have been submitted 

 to experiment in my laboratory is too long to be quoted 

 here, and I shall merely say that we have not yet found 

 one solid body that has failed to become sonorous under 

 proper conditions of experiment.* 



* Carbon and thin microscope glass are mentioned in my Boston paper as 

 non-responsive, and powdered chlorate of potash in the communication to 

 the French Academy, (Comptes Rendns, vol. xlc, p. 595.) All these sub- 

 stances have since yielded sounds under more careful conditions of ex- 

 periment. 



EXPERIMENTS WITH LIQUIDS. 



The sounds produced by liquids are much more diffi- 

 cult to observe than those produced by solids. The high 

 absorptive power possessed by most liquids would lead 

 one to expect intense vibrations from the action of inter- 

 mittent light, but the number of sonorous liquids that 

 have so far been found is extremely limited, and the 

 sounds produced are so feeble as to be heard only by the 

 greatest attention and under the best circumstances of 

 experiment. In the experiments made in my laboratory 

 a very long test-tube was filled with the liquid under ex- 

 amination, and a flexible rubber-tube was slipped over 

 the mouth far enough down to prevent the possibility of 

 any light reaching the vapor above the surface. Pre- 

 cautions were also taken to prevent reflection from the 

 bottom of the test-tube. An intermittent beam of sunlight 

 was then focussed upon the liquid in the middle portion 

 of the test-tube by means of a lens of large diameter. 



RESULTS. 



Clear water No sound audible. 



Water discolored by ink Feeble sound. 



Mercury No sound heard. * 



Sulphuric ether* Feeble, but distinct sound. 



Ammonia " " " " 



Ammonio-sulphate of copper. " " 



Writing ink " " " " 



Indigo in sulphuric acid " " " " 



Chloride of copper* " " " " 



The liquids distinguished by an asterisk gave the best 

 sounds. 



Acoustic vibrations are always much enfeebled in 

 passing from liquids to gases, and it is probable that a 

 form of experiment may be devised which will yield better 

 results by communicating the vibrations of the liquid to 

 the ear through the medium of a solid rod. 



EXPERIMENTS WITH GASEOUS MATTER. 



On the 29th of No /ember, 1880, I had the pleasure of 

 showing to Prof. Tyndall in the laboratory of the Royal 

 Institution the experiments described in the letter to Mr. 

 Tainter from which I have quoted above, and Prof. 

 Tyndall at once expressed the opinion that the sounds 

 were due to rapid changes of temperature in the body 

 submitted to the action of the beam. Finding that no 

 experiments had been made at that time to test the 

 sonorous properties of different gases, he suggested filling 

 one test-tube with the vapor of sulphuric ether, (a good 

 absorbent of heat,) and another with the vapor of bi- 

 sulphide of carbon, (a poor absorbent,)and he predicted 

 that if any sound was heard it would be louder in the 

 former case than in the latter. 



The experiment was immediately made, and the result 

 verified the prediction. 



Since the publication of the memoirs of Rdntgen* and 

 Tyndallf we have repeated these experiments, and have 

 extended the inquiry to a number of other gaseous bodies, 

 obtaining in every case similar results to those noted in 

 the memoirs referred to. 



The vapors of the following substances were found to 

 be highly sonorous in the intermittent beam : Water 

 vapor, coal gas, sulphuric ether, alcohol, ammonia, amy- 

 lene, ethyl bromide, diethylamene, mercury, iodine, and 

 peroxide of nitrogen. The loudest sounds were obtained 

 from iodine and peroxide of nitrogen. 



I have now shown that sounds are produced by the 

 direct action of intermittent sunlight from substances in 

 every physical condition (solid, liquid, and gaseous), and 

 the probability is, therefore, very greatly increased that 

 sonorousness, under such circumstances, will be found to 

 be a universal property of matter. 



*Ann. der Phys. und Chem., 1881, No. 1, p. 155. 

 tProc. Roy. Soc, vol. xxxi, p. 307. 



