631 
of Edinburgh, Session 1874 - 75 . 
wires could be at any instant compared. One great drawback in 
all these experiments was the oxidation of the iron wire. In order 
to get rid of this to some extent, an entirely new arrangement was 
devised, in which the heating of the wires was effected by the same 
current which measured the resistance ; but the results obtained by 
this method were far from satisfactory, owing to the many practical 
difficulties which were continually cropping up. These experi- 
ments were conducted during March and April of 1874. 
In the following June, experiments similar to those of the third 
series above mentioned were made with an iron wire and two 
platinum-iridium alloys — the same which are called M and N in 
the thermo-electric diagram. The resistances of M were compared 
with those of iron, and readings as nearly simultaneous as possible 
were taken of the deflections due to the M-N thermo-electric junc- 
tion in the manner described above. Immediately upon the com- 
pletion of this experiment a triple junction was set up of M, N, 
and the iron wire already used. The currents due to the Fe-M and 
M-N junctions were then compared, and, from the curve obtained 
(see diagram, Fig. VI.), which shows the usual parabolic charac- 
ters at and near the neutral points, the iron line was laid down with 
reference to N (Fig. VII.). The features of this line, taken in 
connection with the M-N deflections observed in the “ resistance ” 
experiment, conclusively prove that the bend of the iron line in the 
thermo-electric diagram occurs at almost exactly the temperature 
d It 
at which the sudden change in the otherwise nearly uniform 
of the same iron wire is observable. 
In the diagrams of the various experiments, all observed points 
which do not lie exactly on the curves traced have been inserted. 
