474 



NATURE 



[March 14, 1895 



(" Sensations of 



tones by resonators is smill and dubious.' 

 Tone," translated by Ellis, p. 157.) 



Now this reinforcement by resonators ha^ been allojether 

 denied by mo;t of those wh'> ha»e taken an interest in the 

 matter, while, if an exception is allowed, it is in favour of the 

 beats cf a disturbed unison, the observed effects being ascribed 

 to the beats, and no: to the difference tone. 



Some writers make no exception whatever in their denial of 

 the objective reality of what may be broadly termed secondary 

 tones. Thus Mr. Bjsanquet, who made a most careful series 

 of experiments some fourteen years ago, stated that " the 

 ordinary first difference tone ... is not capable of exciting 

 a resonator. ... In short, thi difT.-rence tone of Helmholtz 

 ... as ordinarily heard, is not objective in its character." 

 \Proc. Phys. So:, iv. 1S81, p. 233.) 



Prof. Preyer, too, using very sensitive tuning-forks, found 

 that the differential tone given by two f jrks did not alTect a 

 fork the frequency of which corresponded with its own, except 

 in cases where the difference tone was itself a partial of one of 

 the forks. 



It must be remembered that the assertions of Helmholtz as 

 to the experimental proof of the objective nature of the tones 

 were made with reference to those instruments which he regarded 

 as most likely to produce objective notes, viz. the siren and the 

 harmonium, and that, therefore, experiments with forks hardly 

 affect his position. 



Let us now try with the siren whether it is possible to confirm 

 or to disprove the validity of his views. 



For this purpose the rather bulky apparatus which you see 

 before you has been constructed. I should hardly have been 

 able to realise the idei embodied in it, at all events in time to 

 show it to you this evening, if I had not been favourably 

 situated in two respects. In the fi.st place, I have hid the 

 zealou; co-operation of one of my assistant^, Mr. Edwin Edser, 

 who has not only made all the parts of the apparatus that re- 

 quired to be newly made, but has thrown himself into the 

 investigation with the utmost energy, working at it late and 

 early, and making many valuable suggestions and improve- 

 ments. In our joint work we have been helped by some ol my 

 senior students, and notably by Messrs. Cullen and Forsyth. 

 In the second place, I have had at my disposal the magnificent 

 collection of acoustical apparatus in the National Museum at 

 Soath Kensington, some of which I am alloAed, by the kind- 

 ness of the Department of Science and .\rt, to bring here this 

 evening. 



wave-length of light, the path of the ray which falls upon it is 

 shortened by a whole wavelet gih, and the position of each band 

 is shifted to that previously held by its neij;hlx>ur. If the fork 

 vibrates wiih an amplitude of this almost infinitesmal amount, 

 the bands will disappear, or will alternately appearand disappear 

 according to circumstances. The fork may therefore be used 

 to detect by resonance the presence of vibrations, the frequency 

 of which is 64 per second. 



.1 firiori, there were two difficulties of opposite kinds which 

 made it doubtful whether the fork would be an efficient weapon 

 for the purpose for which it was to be used. 



In the first place it would feel tremors of any sort, and it was 

 doubtful whether it would be possible to discriminate between 

 mere shakes and the vibiations which were to be studied. This 

 difficulty has been very largely overcome. 



The table m which the app.-iraius stands rests on india- 

 rubber. On the table are a pair ol lihrary steps ; these support 

 two pieces of wood, which are heavily weighted and rest on 

 india-rubber b.iUs. From these two beams hang steel wires, 

 which carry india-rubber door-fasteners, and the-e in turn sup- 

 port two rods on which the paving-stone is placed. By this 

 alternation of elastic and of heavy bodies we can make the 

 bands absolutely steady, unless the disturbances are violent 

 The quiet movements necessary for wo king the apparatus, the 

 blow ing of the bellows, and the like, priduce no effect. On the 

 other hand, the shutting of a door in a distant part of the build- 

 ing, the rumble of a catt in the street, will cause the bands to 

 disappear. A great deal of the work on which we rely has 

 been done at South Kensington between midnight and three 

 o'clock in the morning. Trustworthy observations have indeed 

 heen made at other times, but it is only in the still small hours 

 thit the apparatus is at its best. 



The second doubt was of a different kind. It was certain 

 that the instrument would be more or less shaken ; it was not 

 quite certain whether the fork would respond to vIbVations of 

 the given period. It is e.isy to set a tuning-fork in vibration 

 by resonance when it is mounted on a sounding ho'i, but in that 

 case the vibrations of the enclosed mass of air .are communicated 

 through the box to the fork. When the stalk of the fork is 

 held rigidly, a tuning fork is notoriously difficult to excite by 

 resonance. This objection is, of course, to some extent counter- 

 balanced by the extraordinary sensitiveness of the means of 

 detecting the vibrations, but it is necessary to supplement this 

 by other devices. The instrument used is a siren (s). In front 

 of it i< placed a hollow wooden pyramid, the narrow end of 



-o. 



The essential part of the appiratus. Fig. 1, is a tuning-fork, K, | 

 to one prong of which is attached a mirror, M, and to the other . 

 a square of thin wood, strengthened by ribs, which is of the same 

 weigh: as the mirror. The fork thus loaded has been compared 1 

 with one of Konig's large standards by means of Lissajous' figures. 

 Its frequency djcn not differ from 64 complete vibrations per 

 second by more than on; vibration in two minutes. The shank 

 is supported by a mass of leid, which in turn is placed upon a 

 paving-tlone. Upon thii stone ako rest the other mirrors 

 necessary for p-oducing Michelson's interference band*. The 

 mirror, M,, i« silvered ?.'. thinly th.ii hall ilie light which falls upon 

 it is reflected, and half is transmitted. 



A ray proceeding from the lantern, i., will be divided at M,, into 

 two, which follow the paths 1 M,Mj\i,iiand i,M|MM,1) respectively. 

 Interference bands are thus produccl, which can be projected on 

 to a screen, so ai to be rendered visible to a large audience. 



If the prong of the tuning-fork mives through the eighly- 

 th jusaodth of an inch, that is, through a distance equal to a half 



NO. 1324 VOL. 51] 



which IS near to, and is of the same area as the wooden plate 

 attached to the tuning fork. This serves to collect the waves 

 of sound, and to concentrate them on the fork. Behind the 

 siren is a large resonator by Koaig, timed to respond to 64 

 vibrations per^iecond. 



In some respects the apparatus requires careful handling. Of 

 course if you blow down the collecting con: the fork may be 

 disturbed, and sometimes a particular note of the siren appears 

 to affect the fork for no very obvious reason. Probably ttie 

 resonance of the air in the cone, or the vibrations of the wooden 

 disk, may at times be the causes of such effects. We have, how- 

 ever, found that whatever they may be due to, they differ in 

 appearance from those proluced by vibrations synchronous with 

 the periodic time of the fork, and they can in general be got rid 

 of by a very .slight readjustment of the appataius. The fact 

 that our main conclusions do not depend on any such nicety, is 

 proved by the fact that the instrument has been se; up twice in 

 the laboratory, and once in the lecture-room in the College. In 



