222 Messrs. Trowbridge and Duane on 



by several sparks in which the dots made by one spark- 

 terminal had the characteristics just described, whereas those 

 made by the other were quite regular. Following out 

 this hint, we found that the particular substances used for 

 the secondary spark-terminals had a large effect upon the 

 characteristics of the photographs. We tried spark- terminals 

 made of a number of different metals — tin, aluminium, mag- 

 nesium, fuse-metal, &c, and finally adopted cadmium as 

 productive of the best sparks. In the case of cadmium the 

 characteristics described are much less marked, and we have 

 succeeded even in producing a few sparks in which no differ- 

 ence in blackness could be detected between one dot and the 

 next. The photographs from cadmium terminals, too, are 

 far more distinct and far more easily measured than those 

 from terminals of any other metal that we tried. 



An interesting question arose here as to whether the dis- 

 tance between two successive dots would depend upon the 

 period of oscillation of the primary circuit if the secondary 

 were unaltered. To test this point the circuits were brought 

 into resonance and a photograph taken. The self-induction 

 of the primary circuit was increased by about 20 per cent, of 

 its value and a second photograph taken. In the first case 

 the distances between successive dots were all within 2 or 3 per 

 cent, of the average obtained by measuring over several dots 

 and dividing by the number of intervening spaces ; whereas 

 in the second case the measurements of some of the single 

 spaces were from 8 to 12 per cent, greater than before, the 

 average from long measurements being the same. This 

 indicates that the vibrations of the secondary circuit are not 

 necessarily perfectly regular, and at a distance apart fixed by 

 the character of the circuit, but are to be looked upon as a series 

 of pulses started travelling along the circuit and keeping at a 

 distance from each other that is determined by the exciter. 

 Owing to the fact that the damping of the" primary is much 

 greater than that of the secondary, the seventh and eighth 

 pulses started are too weak to obliterate the first and second 

 which have travelled the length of the circuit and back. We 

 should expect from this that the bolometer throws, which 

 measure the average length of the wave, would not indicate 

 a shifting of the node when the circuits are thrown slightly 

 out of resonance, but that the minimum throws would be 

 greater than when the circuits are exactly in resonance. 

 This, as is well known, is what happens. 



The improved sparks which the new arrangement of appa- 

 ratus and the use of cadmium as material for the spark- 

 terminals have enabled us to produce, have brought to light 

 another interesting fact, namely that, even when the best 



