690 
Proceedings of the Royal Society 
first fall would then be naturally explained by the solidification 
of the excess of lead contained in this alloy over that required 
to form the eutectic alloy. This solidification of the lead takes 
place at a temperature lower than its ordinary melting point, a 
result doubtless in some way connected with its being mixed with 
the alloy. If three constituents are present, as in fusible metal, 
the whole of one and part of the remaining two may combine to 
form a tri-eutectic alloy, then all that remains of one of the two 
still free may form an eutectic alloy with part of the other metal ; 
and lastly, there would be a residue of the final metal which w T as 
not wholly used up in these combinations. This would lead to a 
curve with three successive falls, which is exactly what we find in 
fusible metal. But, as we have already said, it is still premature 
to form a judgment upon the merits of this theory, especially as 
we hope, at no distant time, to be able to bring forward many 
additional facts, which may materially aid in the formation of a 
conclusion for or against the suggestions offered above. 
It should, perhaps, be mentioned that Dr Matthiessen made 
experiments on fusible metal about twenty years ago,* but 
apparently failed to observe the nature of the alteration of 
resistance. This may have been due to the fact that the form of 
apparatus he employed was not so suitable for detecting the 
character of the alteration. 
The authors, in concluding this account, desire to express their 
thanks to Mr D. Callander for his kind assistance in taking the 
temperature observations. 
A few words may be added in explanation of the accompanying 
tables and diagrams. 
In tables I. and II. are recorded the observations on lead and tin. 
Table III. gives observations on fusible metal, table IY. on the alloy 
SnPb 3 , and table Y. on Sn 3 Pb. 
In the experiments on lead and fusible metal the deflections 
and times alone are given, as the thermo-electric apparatus was not 
adjusted for reading temperatures so high and so low, respectively, 
as those at which these metals melt. In the other tables the first 
column records the time at which each observation was made, 
reckoned in seconds from the commencement of the experiment; 
* Phil. Mag., voL xiii. 4th series, p. 90. 
