SCIENCE AND PRACTICE. 339 



side, the conducting powers of the alloys are almost iden- 

 tically the same. In the reproduction of measures by this 

 system, this fact is of the utmost importance, inasmuch as 

 a trifling error made in the proportions of the constituent 

 metals does not materially affect the conducting power of 

 the alloy. The necessity of procuring absolutely pure metals 

 is also, in some measure, obviated by taking an alloy whose 

 conducting power would probably be very little altered by 

 small impurities, which would very considerably modify the 

 conducting power of a pure metal. There are, unquestionably, 

 difficulties in the way of reproduction by means of solid 

 metals, amongst which the principal are, annealing the speci- 

 mens to the same degree of hardness, drawing the metal 

 into wires of uniform thickness, and, perhaps the greatest 

 of all, the difficulty of soldering the ends of the wire to 

 thick connections without altering its resistance ; and to all 

 these difficulties, therefore, the system proposed by Dr. 

 Matthiessen is subject. Indeed, it is difficult to find any 

 system which is not burdened with very many difficulties and 

 sources of error. 



The Committee of the British Association have proved un- 

 deniably that at this date resistances can be copied with much 

 greater accuracy than they can by any system yet proposed 

 be reproduced. The reproduction of an unit, when once the 

 unit has been fixed upon, is therefore to be avoided as much 

 as possible. To reproduce any standard measure, other 

 measures are indispensable ; and unless these measures agree 

 with each other strictly whenever and wherever the unit is 

 wanted to be reproduced, it can never give twice the same 

 result, however much care and conscience are employed in 

 manipulation. This has been amply proved by the different 

 values of the mercury unit, which Matthiessen arrived at in 

 London, and the writer in Berlin, and those which Weber 

 arrived at in Grottingen, and Thomson in Glasgow, in deter- 

 mining the absolute unit. 



For this reason the Committee of the British Associa- 

 tion have done well in principle in fixing upon a material 

 measure as an unit, and in distributing copies of it ; but in 



z 2 



