700 PROFESSOR THOMSON ON THE ELECTRO-DYNAMIC QUALITIES OF METALS. 
wire much below the temperature of melting ice, when the proportion of platinum to 
iron in the bundle is sufficiently increased (the limit, of course, being the neutral 
point of copper to the platinum itself. See below, §§ 83, 84). 
83. A certain specimen of platinum wire in my possession, when tested by such 
elevations of temperature as could be produced by the hand, was found to lie in the 
thermo-electric series, on the other side of copper from the position in which platinum 
is placed in all statements of the thermo-electric qualities of metals previously 
published. That is to say, when connected by copper electrodes with the circuit of 
a galvanometer, and when heated at one junction up to ten or twenty degrees above 
the atmospheric temperature, a current set from copper to platinum through hot. On 
raising the temperature of the hot junction towards the boiling-point of water, the 
strength of the current began to diminish ; came to a stop when a temperature I suppose 
little above that of boiling water was reached ; and set in the reverse direction with 
increasing strength when the temperature of the hot junction was further raised, the 
other junction being kept all the time at the atmospheric temperature. I afterwards 
found that this specimen of platinum wire (referred to under the designation P, in 
what follows) became neutral to ordinary copper wire at the temperature 64° Cent. 
84. Of two other specimens of platinum wire which I tried with copper, one 
(marked P2) gave indications of a neutral point about the zero of Fahrenheit’s scale, 
but the other (P3) remained, for the lowest temperatures I reached, always on the same 
side of copper as that on which platinum appears, at ordinary and at high tempera- 
tures, generally to lie. When these three platinum wires were tried with one another 
thermo-electrically, they gave, as was to be expected, the mutual thermo-electric 
indications of different metals lying in the order Bismuth, P 3 , Pg, Pi, Iron, Antimony. 
They retained all the same qualities after being heated to redness ; and in a great 
many experiments performed upon them, in which I have found them extremely con- 
venient as thermo-electric standards, have exhibited perfect constancy in their 
thermo-electric bearings. 
I have not yet discovered on what their differences depend, but in all probability 
it is on the degrees to which they are alloyed with other metals. 
85. The fact of copper changing in the thermo-electric series from below the 
position of the platinum specimen Pg to above that of iron, when the temperature is 
raised from —30° or — 20 ° Cent, to 300°, proves that every metal which lies between 
P 2 and iron for any intermediate temperature, must become neutral to either P 2 or 
copper, or iron at some temperature between these limits. Now nearly all the 
common metals, for instance, lead, tin, brass, zinc, silver, cadmium, gold, lie between 
platinum and iron in the thermo-electric series at ordinary temperatures, and no 
doubt many of the rarer metals (I have found aluminium to lie between P 3 and P 2 ) 
are to be ranked within the same limits. Hence at temperatures easily reached and 
tested, neutral points may be looked for with the certainty of finding them, between 
each of those metals and one or other, if not several, of the metals and metallic 
