and the Influence of Temperature on the Resistance of Metals. 199 



spent in cleaning the tube (in which operation the tube is liable 

 to be broken) and in purifying the mercury. 



5. Page 95 (2nd paper), M. Siemens gives a table, by which 

 he wishes to prove that he is able to reproduce resistances of 

 exactly the same values. He, however, only proves that he is 

 able to fill the same tubes with different mercury, and that their 

 resistances only vary 005 percent.; for he compared three un- 

 known resistances with two equal ones (when reduced to equal 

 lengths and diameters), and obtained very nearly the same values. 

 Now if, instead of taking normal tubes, called 3 and 7, he had 

 taken those called No. 1 and 4 (1st paper), would his results 

 have been the same ? No ; they would have varied 1*5 per cent. 

 (See his results given in Table I.) 



6. Page 96 (2nd paper), M. Siemens says, the statement 

 I made that the traces of foreign metals cause a decrement in the 

 conducting power of mercury, and not, as stated by Siemens, an 

 increment, is incorrect. In this M. Siemens is perfectly right. 

 I was misled by. the fact that when mercury is alloyed with 

 several per cent, of foreign metal, a smaller conducting power is 

 observed than the mean of the conducting powers of the relative 

 volumes of the metals employed ; and as in no case I had found 

 an increment in the conducting power of a metal when alloyed 

 with a trace of another, I concluded that traces (Ol or 02 per 

 cent.) of foreign metals would also cause a decrement in the con- 

 ducting power of mercury. 



As mercury behaves in this respect differently from the other 

 metals, instead of assuming, as I did in my paper on the con- 

 ducting power of alloys*, that the metals may be classed under 

 two heads, viz., — 



I. Those metals which, when alloyed with each other, conduct 



electricity in the ratio of their relative volumes ; 



II. Those metals which, when alloyed with one of the metals 



belonging to the first class or with one another, do not conduct 

 electricity in the ratio of their relative volumes, but always 

 in a lower degree than the mean of their volumes, — 

 we must now have three classes of metals, the third probably 

 being — 



Those metals which, when alloyed with very small per-centages 

 of another, have a greater conducting power, but when alloyed with 

 larger per-centages, have a lower conducting power than the mean 

 of their volumes. I am at present investigating how far this may 

 be true j and it will be very interesting to see whether pure 

 metals, such as bismuth, tin, &c, in a liquid state behave like 

 mercury; that is to say, if, when melted, traces of other metals be 

 added, an increment in the conductor will be observed. I also 

 * Phil. Trans. 1860, p. 161. 



