70 Prof. L. Lorenz on the Determination of 



the British Association Committee, a difference which actually 

 amounts to 8 per cent., points to errors which cannot be ascribed 

 to accidental faults of observation, but must rather be ascribed to 

 an imperfect theory. It is here to be observed that the experiments 

 were made with induction-currents of varying intensity ; yet we 

 may at present undoubtedly state, both from theoretical and ex- 

 perimental investigations, that we know only the elements of 

 the theory of variable induction-currents, and that the results 

 can only be regarded as a first approximation. Hence, notwith- 

 standing the great care with which these measurements have 

 been made, we must not ascribe too great a weight to them. 



The determination of the electrical resistance by means of the 

 disengagement of heat which a constant current produces in a 

 conductor is theoretically far simpler and more certain than the 

 induction method in the manner in which this has hitherto been 

 applied. Fortunately we possess a large series of experiments 

 carefully executed and calculated by Quintus Icilius*. This 

 physicist determined the heat which a given intensity of cur- 

 rent disengages in a second in various copper and platinum 

 wires, the electrical resistance of which was determined by means 

 of a standard measured by Weber in absolute units. If V de- 

 notes the number of relative units of heat (1 mgrm. water 1° C.) 

 which are developed in a second in a Siemens' s unit of resistance 

 by the current-intensity S, we obtain by these experiments a 

 determination of the constant a. in the equation 



Y=*s~x 1*0257 xlO 10 , 



if with Quintus Icilius we use Weber's determination of electrical 

 resistance ; while with the former signification of A as the abso- 

 lute equivalent of work for the relative unit of heat (1 mgrm. 

 water 1° C), and of s as the absolute value of Siemens' s unit 

 of resistance, we have 



AV= 5 2 S. 



From the two equations we have 



S = *Ax 1-0257 xlO 10 . 



In the experiments in question three several liquids were used 

 in the calorimeter, namely water, alcohol, and oil of turpentine. 

 The former liquid had the advantage over the two others, that it 

 gave the quantity of heat directly in the units chosen ; on the 

 other hand, a small error cannot be avoided, owing to the greater 

 conductivity of water for electricity, by which the disengagement 

 of heat, and therewith also the constant a, must be found rather 

 too small. Owing to its volatility, the experiments with alcohol 



* Pogg. Ann. vol. ci. p. 69, 



