Dr. Matthiessen on the so-called Mercury Unit. 367 



agrees with the last value ; and as the true value for the resist- 

 ance of the mercury unit, as defined by Messrs. Siemens, we may 

 take 0*961 B. A. units, a value differing from their 1864 issue 

 by about 05 per cent., and when corrected for specific gravity, by 

 about 0*8 per cent. 



Now why do these differences exist ? Are we not led to think 

 from the papers written by these gentlemen, and others working 

 in their laboratory, that the reproduction of the mercury unit is 

 the most simple thing possible ? I will quote what one of them 

 says on the subject. Mr. R. Sabine states*, " Following this 

 method, every electrician may inexpensively and with little trouble 

 make himself a standard measure.. ..The mercury unit has there- 

 fore been reproduced inDr.Siemens's laboratory twenty-one times, 

 six times in the first determination, five times in the second, and 

 ten times in the present. A.nd, allowing for the unfortunate 

 misrepresentation of the measure by individual errors of the 

 measuring-apparatus used in the comparison of the first tubes, 

 the agreement between them all is greater than could be gua- 

 ranteed between any two single electrical measurements with 

 different measuring-apparatus. . ..From the foregoing results it 

 follows that, by the method of direct production proposed by 

 Dr. Siemens, much greater exactness has been attained than 

 by means of any other of the methods of determination or copy- 

 ing.^ 



I cannot allow the above passage to pass unnoticed, for it will 

 be apt to mislead many who have not worked in this field of 

 research. In the first place, we are told that this method of 

 reproducing resistances is inexpensive and causes very little 

 trouble. It may be inexpensive to those who possess first-rate 

 apparatus for normal weighings or measurements, normal thermo- 

 meters, &c; for with first-rate apparatus only normal results can 

 be obtained. Then, as to the trouble; does anybody for a mo- 

 ment suppose that a standard measure can be reproduced with 

 a little trouble ? Would it not take any observer weeks to re- 

 produce with accuracy a resistance which can be relied on ? 

 must not such an observer check and recheck every determi- 

 nation he makes? must not he be sure all his instruments are 

 graduated accurately ? and he can only be sure of this if they 

 are carefully tested by himself, or by perfectly reliable persons ; 

 and, what is of very great importance, must he not be sure that 

 he deals with an absolutely pure metal ? Did not Mr. C. W. Sie- 

 mens himself state at the meeting of the Royal Society when 

 remarks had been made upon Mr. Jenkin's paper on the B. A. 

 unit, that he was loth to give up the use of the mercury unit, as 

 he was afraid the B. A. unit could not be accurately reproduced, 



vol. xxv. p. 172 (1863). 



