778 PROFESSOR JOHN G. M'KENDRICK 



unequal thickness at each end to suit the tapering form of the mandril and it cannot 

 be put on in a reverse position, and, in the second, when the wax cylinder is in its 

 normal position a spiral groove is cut from left to right, and if it were reversed it would 

 pass from right to left, while the feeding screw of the machine which secures the motion 

 from left to right would still be moving in that direction. Having secured the reversal, 

 it is easy to show that words pronounced backwards to a phonograph moving in the 

 direction of left to right and then reproduced from a phonograph moving in the direction 

 of right to left are uttered normally and with considerable distinctness. 

 Thus the sounds — 



Fargonof 



become 



Fonograf 



Nofelet 



)> 



Telephon 



Fargelet 



)> 



Telegraf 



Pokselet 



)> 



Teleskop 



Arrubnide 



5! 



Edinburra 



Wogsalg 



)} 



Glasgow 



Elponitnatsnoc 



y> 



Constantinople 



Eetisrevinu 



)> 



Universitee. 



The difficulty, of course, is in emphasising the proper syllables. In pronouncing a 

 word backwards not only may the emphasis be placed on the wrong syllable but suppose 

 that in an emphasis one begins the sound crescendo and ends it off diminuendo, when 

 the reversal takes place we have the opposite, often giving a ludicrous effect. This is 

 Avell observed on listening to reversed music. Not only are all one's notions of the 

 relations of tones thrown into confusion, but as with many instruments the tone is sharply 

 and distinctly taken and then is allowed to weaken in intensity, the effect is produced 

 of tones beginning diminuendo and ending loudly and abruptly. Still it is interesting 

 to find that such tones as those of the vowel sounds, flute tones, cornet tones, and forks, 

 if spoken, sung, or played smoothly, and, as far as possible, with a uniform intensity, 

 come out with equal distinctness and, so far as I can observe, with very little if any 

 alteration in quality, whether they are taken normally or reversed. No experiment 

 could illustrate more strikingly the dependence of quality of tone on the resultant wave 

 forms imprinted on the cylinder of the phonograph. At the same time, I am not con- 

 vinced that with the means at our disposal we have yet obtained an absolutely accurate 

 reversal. This criticism was first suggested to me by Professor Tait some time ago, 

 and a careful scrutiny of the marks shows that he is right. The depression made on 

 the wax cylinder for a given vibration does not seem to be deepest in the centre, but 

 the maximum dip is nearer one end while there is a tail or " trail out " as the pressure is 

 removed. The direction of the trail out is opposed to that in which the cylinder is 

 travelling, and as the cylinder travels with great velocity, the trail out is quite appreci- 

 able. Were the phonograph a perfect machine, the maximum depth of the curves made 

 by compound tones or by harmonies would be in the centre, and the form of the curve, 

 corresponding to the periodic variations of air pressure, would depend on the sum of its 



