ON THERMO-ELECTRIC CURRENTS IN CIRCUITS OF ONE METAL. 177 
Heat,’ that if the condition of metal at a certain temperature depended ex- 
clusively on that temperature, no distribution or movement of heat could 
possibly give rise to a current of electricity in a circuit of one metal; never- 
theless I find, as above stated, that in a circuit of one metal wire a current 
is maintained for five minutes at a time, gradually vanishing to nothing when 
the two ends of the homogeneous wire have been for some time in contact, but 
recommencing if one wire is cooled for a minute and then again applied to 
the hot one. One explanation of this might be that the condition of the 
wires does not solely depend on their temperature, but is influenced to a con- 
siderable extent by the time during which they have remained at that tem- 
perature. Nor is this a gratuitous assumption: Dr. Matthiessen has proved 
that wires of several metals do not attain a constant conducting power until 
they have been kept for some time at a constant temperature; he finds that 
the conducting power of bismuth increases, while that of tellurium decreases 
when kept for a time at 100°. Quite similarly, some metals may rise and 
some may fall in the thermo-electric scale after being heated for some time, a 
supposition which is necessary to account for the metallic contact currents by 
the theory I suggest. 
Another possible explanation of the metallic contact currents may be found 
in a partial hardening on the one side and annealing on the other, caused 
by the sudden contact of the hot and cold metal. If this be so, the current 
between annealed and unannealed wires of the same metal would correspond 
with the contact current between two homogencous wires, in a way which it 
does not seem to do. 
1 am, however, now engaged in investigating this subject, and hope before 
next year to be able to give facts which may decide whether either of these 
theories is tenable. There is great difficulty in forming any conclusion from 
experiments hitherto made, inasmuch as none of the observers, except 
Dr. Matthiessen, have used chemically pure metal, and it is found that the 
electrical properties of a metal are affected to an extraordinary degree by the 
presence of impurities in very small quantities. 
Explanation of the Table. 
“The names of the metals of which the loops were made are entered at the 
side and top of the Table. The experiments made with each combination are 
entered in the subdivision at the intersection of the horizontal and vertical 
columns corresponding to the two metals. The metals named at the top 
formed the right-hand loop, those at the side the left-hand loop. The arrows 
show the direction of the current across the joint. The first entry in each 
subdivision shows the deflection observed when the right-hand metal was 
heated and the wires held loosely together. The second entry shows the 
ee when the same metal was heated but the wires drawn tightly 
together. 
The third entry gives the maximum deflection, and the direction of the 
current, when the middle of the joint is gradually heated and the two wires 
held tightly together. 
The fourth entry (where given) shows the maximum deflection from a 
current in the opposite direction when greater heat was applied. The two 
entries show the common well-known metallic thermo-electric effects. 
The first entry shows the new loose-contact effect. The second entry shows 
an uncertain combined effect of metallic and imperfect contact effects. 
a example will perhaps make this clearer, When copper and iron were 
* N 
