
TO A THERMO-ELECTRIC DIAGRAM. 139 
alluded to. As already stated, this is a question involving the actual specific 
heat of electricity in each metal; not the difference of the specific heats in any 
two metals, which is all that my experiments furnish. 
Subject to these remarks we have the following table of the values of 4, 
whose contents are represented graphically in Plate VII., and where the unit of 
electro-motive force employed is nearly 10~° of a Grove’s cell. The tangents 
of the inclinations of the lines in the plate may be reduced to the corresponding 
numerical values of / in terms of a GROVE’s cell by the factor 4 x 107°. 
Fe — 00247 Ca + 00218 
Steel — ‘00171 Zn + 00122 
M — -00000 Ag 4 00076 
Pt. Ir (No. 1) — 00028 Au + 00052 
Pt. Ir (No. 2) — 00068 Cu + 00048 
Pt. Ir (No. 3) — -00032 Pb ‘00000 
N — ‘00000 Sn + 00028 
Pt (soft) — ‘00056 Al + 00020 
Pt. Ni — 00056 Pd — 00182 
Pt. (hard)  — ‘00038 Ni(to175°C.)  — 00260 
Mg — 00048 Ni (250°—310°C.) + 01225 
Arg — 00260 Ni (from 340°C.) — :00260 
Plate VIII. shows directly the galvanometric indications of circuits including 
various iron and steel wires ; one of which is a ribbon of pure iron, prepared by 
Dr MarruiessEn, kindly put at my disposal by Dr Russet. The other speci- 
mens of iron consist of two from the ordinary stock in my laboratory, and 
a third (probably, from its position so close to that of Dr MATTHIESSEN, very 
pure) which I owe to Sir R. Curistison, who has used portions of it for 
chemical testing for more than thirty years. It was, therefore, prepared at a 
time when more care was employed to secure purity than in the present day. 
The circuits were completed by the platinum alloy called N above, whose line 
is nearly parallel to that of lead, but a little above it. The temperature scale 
is the temporary one given by the galvanometric indications of the two platinum 
alloys M and N. Their lines are drawn as almost exactly parallel in Plate 
VII., but they intersect at some temperature about a white heat; so that to 
reduce the diagram to something roughly corresponding to absolute tempera- 
tures, the whole must be extended parallel to the temperature axis, and in 
ratios continually increasing for higher ranges of temperature. The experimental 
work on which this diagram is based has been performed almost entirely by 
Messrs Knorr and Situ, and its general accuracy may be estimated by the 
smoothness of the curves obtained: particularly as all the observed points 
which do not lie exactly on the curves have been inserted in the diagram. 
The points of contrary flexure in these curves correspond to the points of 
change of sign of specific heat of electricity in the specimens of iron and steel, 
