282 DR CARGILL G. KNOTT ON 
permanent on restoration to the original temperature, provided the surface 
has remained the same chemically. Any permanent alteration in the density 
of the condensed gases presupposes, then, a chemical change on the surface ; 
and if there be no such permanent alteration, or if it be insufficient to account 
for the observed phenomena, the last resource still seems to be chemical change, 
to which accordingly we look as the only efficient cause, whether directly or 
indirectly, of the changes observed in the mutual electrical relations of metals. 
This hypothesis is also supported by the known phenomena of time-variation of 
metal surfaces in both their chemical and electrical relations. The electrically 
negative character of unpolished iron, copper, zinc, tin, aluminium, &c., to 
polished iron, copper, zinc, tin, aluminium, &c., is generally attributed to surface 
oxidation ; probably, then, the electrically negative character of polished and 
heated iron, copper, zinc, and tin, to polished but unheated iron, copper, zinc, 
and tin, is to be referred to a similar cause. If so, then the above experiments 
lead to the result that for these metals at least, there is for every temperature 
a definite surface condition which no amount of polishing can alter—a surface 
condition produced most probably by a film of oxide or other similar compound 
over the metallic surface by the action of atmospheric air ; and that, further, the 
surface change due to change of temperature is a direct function of that tem- 
perature-change. This surface state forms within the first few seconds after 
polishing, perhaps instantaneously, and thereafter no appreciable change ensues 
till several minutes have elapsed, when the inevitable time-variation of the 
surface, as depicted in the curves of fig. 1, begins to show itself. Hence it 
would appear that at ordinary temperatures a chemically pure surface of these - 
four metals in air is an impossibility; and that the same holds true for other 
metals, even for the so-called non-oxidisable, is a not improbable surmise. In 
this connection it should be remarked that to the eye there was no appreciable 
alteration of surface, no dimming of the bright metallic polish, even after the 
lapse of several minutes. 
The experiments which form the subject of this thesis were made in the 
Physical Laboratory of Edinburgh, during the summer session of 1878, and the 
winter session 1878-79. The apparatus was, for the most part, lent me by 
Professor Tart, whom I here thank for the kindly interest he has evinced in 
my work, and the ever ready advice with which he has aided me. 
Added, May 1881.—As it was just possible in the above experiments that 
the variations of potential observed might be affected by changes in the capacity 
of the condenser, further experiments were made in which any such alteration 
in capacity might be effectively eliminated. The two opposed surfaces of the 
condenser, brought to within a millimetre distance of each other, were put into 
metallic contact by means of external wires. In this way, after the method o 
