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



for experience in reproducing the mercury unit had taught him 

 the great difficulty of such ' operations ? I think Mr. Sabine's 

 sentence that I refer to would have been more correct if thus 

 written: — Following this method, every electrician may, with 

 perfect apparatus, &c, and with a very great deal of trouble, by 

 taking some weeks for his determinations, together with the use 

 of all precautions to ensure normal results, as well as employing 

 proper constants and coefficients for deducing and correcting the 

 values so obtained, make himself a standard measure. 



In the second place, we are told that the unit has been re- 

 produced twenty-one times, and that, allowing for the unfor- 

 tunate misrepresentation of the measure by individual errors of 

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

 agreement between them all is greater than could be guaranteed 

 between any two single measurements with different measuring- 

 apparatus. 



The first tubes show,asIhave elsewhere pointed out, a maximum 

 difference between the calculated and observed resistances of 1*6 

 per cent., and this differenee is said to be caused by errors in 

 the measuring-apparatus. Assuming this to be the case, it 

 proves that the precautions to be adopted in adjusting the appa- 

 ratus must be very great. For if in Dr. Siemens's hands, with his 

 skill and with such good apparatus as is described in his paper, 

 such errors can occur in his endeavours to reproduce this unit, 

 how much more are errors to be feared from others who have not 

 such good apparatus at their command ! As to the agreement 

 between the other reproductions being more accurate than any 

 two measurements with different apparatus, the following facts 

 will constitute a good answer. 



1. The Messrs. Siemens' s reproductions, they state, agree 

 within 0*05 per cent. Two coils* were compared with each 

 other some years ago (the one German silver, the other gold 

 silver), and now, with different apparatus and by a different ob- 

 server, they were again compared : the ratio of their resistances 

 was found absolutely the same. 



2. The Messrs. Siemens have issued an 1864 mercury unit. 

 Now four different copies of this have been compared with four 

 different copies of the B. A. unit with four different measuring- 

 apparatus, and by four different observers f, and the ratios of the 

 resistances, as determined by the four observers, do not vary 

 between each other more than 0*02 per cent. 



In the third place, we are told that, by direct production, 

 much greater accuracy has been attained than by any other 

 method of determination or copying. 



* See Report of Electrical Standard Committee, 1864. 



t See Report on the B. A. Unit, by F. Jenkin, Proc. Roy. Soc. No. 74. 



