Telephone as an Oscillograph. 911 



registration with over-lapped vibrations, when the damping 

 of the galvanometer-coil is too small. 



I have studied the effect of these three causes separately. 

 The galvanometer was in each case so damped (by introducing 

 a fitting resistance in the second coil-winding) that the 

 aperiodic case was obtained. With very slow speed of the 

 reo'isterino- drum a sudden current variation was then 

 correctly reproduced. A sudden change of the light-intensity 

 on the thermopile then gives, owing to its slowness, a slow 

 change in the registered curve. The vastness of this effect 

 is seen in fig. 2 a, where the light-intensity was suddenly 

 changed with a time-interval of \ minute. 



Fiar. 2. 



Finally, the effect of the breadth of the light-line is seen 

 in fig. 2 b. On the plate there were three black sections with 

 sharp edges. The correct curve would then have the form 

 of straight lines with right angles. The actual breadth of 

 the light-line can be calculated from this curve. 



In reality, this distortion is not so great since such sudden 

 changes are impossible. It may also be helped by diminishing 

 the image, or by drawing the photographic plate with 

 greater speed. 



3. The Damping of the Membrane. 



The damping factor of the system is of the greatest im- 

 portance in the application of a movable system as oscillo- 

 graphic recorder. As shown by the theories of the ordinary 

 string oscillograph (see the excellent paper on this subject 

 by H. Busch, Phys. Zs. 1912), the best results are obtained 

 with a damping factor of 23' 14 (the rate of two consecutive 

 amplitudes). This value, however, is deduced under the 

 supposition, that the system vibrates as a simple swinging 

 point, i. e., that it has only one single free vibration. As for 

 the telephone-membrane, we know that there are a series of 

 possible, free vibrations, so that the above theory is not 

 precisely applicable in this case. In my paper in Annalen I 

 have indicated a value of 1*14 as the damping factor of the 



