46 MEMOIRS OF THE NATIONAL ACADEMY OP SCIENCES. 



of the cup. This adjustment has to be made with great care. The spark which at each second 

 passes between the i)latinum tip and the mercury rapidly oxidizes the latter, and the mercury 

 must be renewed at each experiment. (2) The paper on the cylinder must be smooth, thin, but 

 not glazed. It required many days of experimenting before I succeeded in getting a paper which 

 gave the results I sought. The best paper is a very thin printing paper with a smooth, un glazed 

 surface. (3) The style on the fork must be very light and elastic. The best for this purpose is 

 made out of thin hard-rolled copper or aluminium. (4) The spark given by the inductorium must 

 be of the character already described. If the discharge of the inductorium be not composed of a 

 single sparic, and its impress on the paper a minute circular spot bisected by the trace of the fork, it 

 will be useless to expect accurate results from this method. 



To reach these conditions cost much time, and it may be interesting to describe some of the 

 variations in the character of the discharge of an inductorium when excited by various strengths 

 of current, and when condensers of vai-ious areas are or are not in the secondary circuit. 



The flash of an inductorium appears composed of a single discharge ; but only in certain 

 conditions is it really composed of a single spark. If the discharge be obtained through the style 

 on the fork with a current traversing the coil of a strength approaching that used in the usual 

 electrical experiments with it, several flexures of the trace of the fork will be obliterated by the 

 discharge deflagTating the carbon on the paper. This effect is produced by a multiplicity of dis- 

 charges, following each other with such rapidity, and of such strength, as to denude the paper of 

 carbon, to some extent, on either side of the trace. The breadth of carbon removed and the fuzzy 

 character of the contour of these traces give them the appearance of caterpillars. 



To obtain an analysis of these complex actions I devised the following method of experi- 

 menting, shown in Fig. 2: A revolving brass cylinder, similar to the one used in our appa- 

 ratus just described, was covered with thin printing paper, and the latter was well blackened by 

 rotating the cylinder over burning camphor. The paper was then removed from the cylinder and 

 cut into disks of about 15 centimeters in diameter. When one of these disks is revolved about its 

 center with a velocity of about 20 times per second, it is rendered very flat by centrifugal action. 

 It can then be brought between points or balls, even when the latter are separated by no more 

 than .75 millimeter. When in this position the discharge between the points or balls perforates 

 the disk and leaves a permanent record of its character, of the duration of the whole discharge, 

 and of the intervals separating its constituent flashes .and sparks. To obtain the time of rotation 

 I presented momentarily to the rotating disk a delicate point attached to the prong of a vibrating 

 fork of known period of vibration. The axis of the sinuous trace thus made by the fork is traced 

 by a needle point applied to the rotating paper disk. Drawing radii through symmetrical intersec- 

 tions of this axis on the sinuous line, we divide the disk ott' into known fractions of time. The 

 disk is now removed from the rotating apparatus and the carbon is fixed by floating the disk for a 

 moment on dilute spirit varnish. When the disk is dry it is centered on a divided circle provided 

 with a low -power micrometer microscope, and the duration of the whole discharge and the inter- 

 vals of time separating its components can be determined to the oiioo of a second. I here give 

 three typical experiments with this method, which will show the characteristics of the discharge 

 of an inductoriiim : 



1. Discharge of a large inductorium (striking distance, 45 centimeters between 

 brass points) between platinum points 1 millimeter apart. no jar or condenser 

 in secondary circuit. 



The platinum electrodes were neatly rounded and formed on wire 1% millimeter in diameter. 

 After the discharge through the rotating disk nothing was visible on it except a short arc formed 

 of minute, thickly-set white dots; but on holding the disk between the eye and the light, it was 

 found to be perforated with 33 clean, round holes, \nth the carbon undisturbed around their edges. 

 The portion of the discharge which makes these holes lasts „-3 second, and the holes are separated 

 by intervals which gradually decrease in size toward the end of the discharge, so that the last 

 spark-holes are separated about one-half of the distance separating the holes made at the begin- 

 ning of the discharge. The average interval between the spark-holes is 7-^ second. After this 



