October 2 1, 1909] 



NATURE 



4S9 



the length of each line, from side to side, represents 

 in time 03 second. .Short descriptions of the tracings 

 are printed below the figures. 



The experiment was performed thus : — After care- 

 fully adusting the markers on the smoked paper, the 

 cylinder was allowed to rotate until it reached its 



made which was attached to the circular plate of 

 another gramophone, and having a circumference such 

 that the recorder traversed from 21 to 24 inches per 

 second. The two gramophones, one to play, the other 

 to carry the wooden drum, were driven at the same 

 speed. Tracings were thus obtained similar to the 



I 



Fn,. 2. — :5inaU port.-, n of tracing of the \Veslmin^ter cliiin;s. Tne tracing btf^ins at the left-hand corner at the bottom, runs along the line to the 

 right, then is cjniinued in the next line above, again at the loA'er left-hand corner, and so on. The length of time represented by the length of 

 one line fiom left to right is about one-third of a second. The tiacing> of the i/icoth of a second are seen at the top. The tracing is from the 

 portion of the record giving the strokes of twelve o'clock. Near the top little groups of waves indicate the beats, &:c., after the last stroke. 



maximum uniform speed. By closing a key, a time 

 record was taken. Then the gramophone was started; 

 the sound waves acted on the microphone, and the 

 little electromagnetic marker began to sing or play; 

 finally, by having my right hand at the top of the 



indentations or waves on the gramophone record. 

 Trouble arose, however, from the oscillations of the 

 gramophone plate (one of the conditions of the success 

 of the gramophone, in which all the arrangements are 

 more or less mobile), but this difficulty was easily 



t IG. 3. — Small portion o tracing of a nuuiber ol in^le vo.cts s nging the " Soldiers' Chorus " from I'an^t. From the La 

 limj relatio.iS, &c., arc the same as in Fig. 2. Observe the complicated form of the curves, 



im [he La .Scala Theatre, Milan. The 



screw, as shown in the figure, I was able slowly to 

 raise the platform, carrying the recorder, so as to 

 describe a long spiral line, about 135 feet in length, 

 from the bottom to the top of the cylinder. On 

 reaching the top, the experiment came to an end. To 



surmounted. Finally, I found that with my arrange- 

 ment it was not necessary to use the gramophone, as 

 it was easy to record the vibrations of a human voice 

 bv causing the singer or speaker to sing or speak 

 direct to the microphone. The arrangement is an 



Fig. 4. — Small portion of a tracing giving the vibrations of the voice of Signor Caruso in singing " Spirito Gentile." Time relations, &c.. same as in 

 Fig. 2. Observe the crescendo and diminuendo of certain notes, the high pitch of others, and the regularity in form of the waves. 



secure good results, great care had to be taken to 

 secure nicety of adjustment. It must be explained that 

 in the tracings so taken the recorder traverses 

 !2 inches per second. The waves in my tracings are 

 thus compressed laterally. To obtain waves at the 

 speed of the gramophone, I had a wooden cylinder 



NO. 2086, VOL. 81] 



excellent phonautograph. About five minutes are 

 occupied in taking a tracing, the average length of 

 which is 135 feet. 



An inspection of these tracings shows the wonderful 

 variety of pressures pouring in upon the ear as we 

 listen to music. Three or four or inore notes differ- 



