the Unit of Electrical Resistance. 331 



Dr. Matthiessen concentrates his argument in the two theses — 



1st. That no true mercury unit has been issued. 



2ndly. That the units issued from time to time have not the 

 same resistance. 



He endeavours to justify both these assertions by the fact that, 

 in the determinations of the unit, I made use of the wrong specific 

 gravity of mercury; that two sets of resistance-coils shown in 

 the London Exhibition of 1882 differed 1*2 per cent, between 

 themselves ; that in the first determination of the mercury unit 

 differences of 1-6 per cent, appeared between the different tubes; 

 and that his determination of the unit differs 08 per cent, from 

 mine. 



In supporting his first assertion, Dr. Matthiessen loses sight 

 of the fact that the measure proposed by me consists of a defini- 

 tion, and is therefore absolute. That the resistance- coils and 

 standards which I have issuedfrom time to time correspond exactly 

 with this definition I have never asserted ; on the contrary, I 

 have repeatedly expressed the wish that, sooner or later, experi- 

 enced physicists would take the trouble to reproduce the measure 

 according to the definition with the best means at command. 

 Dr. Matthiessen would have been justified in his first assertion 

 if my definition had been doubtful, or if the method laid down 

 had been untrustworthy; but he has proved neither. I con- 

 cede, however, willingly that the specific gravity of mercury 

 which I admitted into the calculation is not correct. When (in 

 1858) I made experiments to ascertain if the mercury unit was 

 capable of being reproduced with sufficient exactness, I found 

 the number 13*557, and adopted it as correct, as the same value 

 had been found by direct comparison of the heights of mercury 

 and water in connected tubes. Unfortunately, in the reproduc- 

 tion of the unit made more recently with much greater care and 

 with improved instruments, the same specific gravity was re- 

 tained, and not that given by Regnault (13*596) and by Balfour 

 Stewart (13*594), of the truth of which there can be no question. 



According to this, the standards hitherto issued are 0*287 per 

 cent, too great ; and if we take the percentage increment of the 

 resistance of the German silver which I have employed between 

 0° C and 100° C. = 2*72, these standards will not represent the 

 unit at the temperature stated, but at 10°*5 higher. 



I am desirous of allowing due credit to Dr. Matthiessen for 

 having given occasion to this correction, which has notwithstand- 

 ing no influence whatever upon the value of the system. 



Dr. Matthiessen asserts in addition that the issues of coils 

 from time to time do not represent the same unit. 



The unit has been determined in my laboratory at three dif- 

 ferent times, each time with a nearer approximation to the true 



