130 
MR, \Y. GEOFFREY DUFFIELD ON THE 
long period of rainy weather proving of great service in the prosecution of this 
research. 
5. To have a means of ascertaining whether the photographs taken were reliable, 
the comparison spectrum was invariably photographed by a divided exposure ; the 
spectrum of iron arc at atmospheric pressure was taken immediately before and also 
immediately after the pressure exposure, and the positions of these two spectra were 
carefully examined. Any disturbance of the apparatus or continuous change in 
temperature during the experiment could thus be discovered, and all plates in which 
the lines were not accurately coincident were rejected. 
7. The Behaviour of the Iron Arc under High Pressure. 
Under ordinary conditions the iron arc was maintained almost as easily as the 
carbon arc, but an increase of the pressure of the air surrounding it caused its 
management to become more difficult. The disturbed area on the positive pole from 
which the arc sprang was in constant motion over the surface of the turbulent 
molten mass, and luminous metallic vapour was constantly expelled in all directions. 
These convection currents rendered the arc very unstable, and at a pressure of only 
10 atmospheres it was difficult to maintain it for more than half a minute on account 
of the tendency of these flames to blow it out. 
At higher pressures the life of the arc became shorter, and at 50 atmospheres the 
exposure consisted of a series of flashes of not much more than a second’s duration, 
the maximum length of the arc being then about 2 millims. For the first pair of iron 
poles, between 300 and 400 flashes were required to affect a photographic plate, so it 
was necessary to prolong the process of striking the arc and withdrawing the poles 
for a period of from 40 minutes to one hour. Up to 50 atmospheres a workable scale 
of exposures was to open the shutter for the same number of minutes as the number 
of atmospheres employed, but for pressures between 50 and 100 atmospheres it was 
not necessary to increase the exposure beyond one hour. 
A second set of photographs, Set B, was taken with another pair of iron poles, and 
these proved much more satisfactory, as the arc lasted much longer at all pressures, 
at the highest pressure only 8 to 12 minutes exposure being required, and this 
rendered the plates less liable to fogging from stray light in the room. The two sets 
of photographs are subsequently referred to as Sets A and B. 
The specimens of iron bar were obtained from the same source in the two cases. 
The length of the exposure necessitated the use of a wider slit than would 
otherwise have been employed; this was decided upon experimentally by taking a 
series of photographs of the iron arc at ordinary pressures with different widths of 
slit and choosing that aperture at which the definition just remained good—the 
series showed that the definition decreased gradually to a point at which it suddenly 
became bad; a reasonable margin for accidental alterations was allowed, and the slit 
