APPENDIX 



105 



excursions and records caused by different voltages acting for equal 

 times, and by a given voltage acting for different times. 



The voltage actually indicated by any given curve or portion of 

 a curve is most readily ascertained by superposing the record upon 

 the record of a "normal curve," z>., of the curve described by a 

 known constant P.D. taken under similar conditions. The two 

 plates are slipped over each other, with abscissae kept parallel until 

 portions of the curve under examination coincide with portions of 

 the normal curve' at known voltage. Such coinciding parts are 

 equipotential, so that the value on the normal curve gives the value 

 on the curve under inspection. 



For the more minute analysis of electrometric curves, the student 

 should consult the papers of Burch, Einthoven, and Garten. In 

 many instances the calculated curve appears very different from the 

 original record of the curve described by the mercury column, e.g., 

 Fig. 66. 



FlG. 67. The primary circuit of the inductorium is completed through the pools 

 pp. The galvanometer is short-circuited through the other pair of pools G G. The 

 level of mercury in the lateral pool, and the length of the semicircular wirep, are 

 adjusted so that circuit is completed at G G (i.e., the galvanometer is short-circuited) 

 before circuit is completed at pp. On releasing the commutator cradle the weight, W, 

 lifts the two wires from the mercury, breaking first the primary circuit, and subse- 

 quently the galvanometer short-circuit. The time of transfer is thus kept constant. 



Keys. For the study of blaze-currents, occasional use has been 

 made of two special keys, modified from the well-known model of 

 the Pohl's commutator, for the purpose of automatically (i) breaking 



