188G.] The Clark Cell a Standard of Electromotive Force. . 79 



lines in this case are luminous on a fainter continuous background 

 and are not particularly sharp, whilst the reflection spectrum consists 

 of black lines sharply defined on a continuous spectrum. 



The spectrum emitted by incandescent erbia is shown in fig. 4. 



Fig. 5 shows the characteristic lines in the spark spectrum of 

 erbium, taken from a concentrated acid solution of erbic chloride, 

 with a Leyden jar in a shunt circuit. 



I have thought it advisable to give these five spectra of erbium, as 

 they show how entirely different the phosphorescent spectrum is to 

 any other spectrum given by this element. 



IV. " On the Clark Cell as a Standard of Electromotive Force." 

 By the Lord Rayleigh, M.A., D.C.L., Sec. U.S. Received 

 January 7, 1886. 



(Abstract.) 



This paper, supplementary to that " On the Electrochemical Equi- 

 valent of Silver, and on the Absolute Electromotive Force of Clark 

 Cells,"* gives the further history of the cells there spoken of, and 

 discusses the relative advantages of various modes of preparation. 

 The greatest errors arise from the liquid failing to be saturated with 

 zinc sulphate, in which case the electromotive force is too high. The 

 opposite error of super- saturation is met with in certain cases, 

 especially when the cells have been heated during or after charging. 

 Experiments are detailed describing how cells originally supersa- 

 turated have been corrected, and how in others the electromotive 

 force has been reduced by the occurrence of supers aturation conse- 

 quent on heating. If these errors be avoided, as may easily be done ; 

 if the mercury be pure (preferably distilled in vacuo), and if either 

 the paste be originally neutralised (with zinc carbonate), or a few 

 weeks be allowed to elapse (during which the solution is supposed to 

 neutralise itself), the electromotive force appears to be trustworthy 

 to toVo P ai "t. This conclusion is founded upon the comparison of 

 a large number of cells prepared by the author and by other physicists, 

 including Dr. Alder Wright, Mr. M. Evans, Dr. Fleming, Professor 

 Forbes, and Mr. Threlfall. 



As regards temperature coefficient, no important variation has been 

 discovered in saturated cells, whether prepared by the author or by 

 others. In all cases we may take with abundant accuracy for ordinary 

 applications — 



E=r43o{l- 0-00077(^-15 o )} ; 



the temperature being reckoned in centigrade degrees. For purposes 

 of great delicacy it is advisable to protect the standards from large 

 * "Phil. Trans.," vol. 175, 18S4. 



