432 
ME. J. T. BOTTOMLEY ON THERMAL 
a check on the current galvanometers, the absolute values of their readings were 
frequently determined by electrolysis ; and this was done specially after the chief series 
of experiments, and before calculation of the several portions of the results tabulated 
and shown in the curves given with this Paper. In such determinations the galvano¬ 
meter was not removed from its place, nor was any alteration made in the soldered 
connections, with the exception of opening the main branch of the bridge at the 
point c to admit the electrolytic cell. By this means local disturbance (if any) of the 
current galvanometer due to electrodes was taken into account ; and this method of 
ascertaining the absolute value of the indications of the current galvanometer was 
satisfactory because of the complicated nature of the circuits.. Influence from the 
electrodes was done away with as far as possible by laying side by side, insulated 
with paraffined paper, the electrodes coming to and going from any particular point; 
but a certain amount of complication was unavoidable, due regard being given to 
convenience of working, on account of the great number and variety of the pieces 
of apparatus used in these experiments, and on account of the necessity of placing 
the tube which contained the radiation wire, and which was attached to the 
Sprengel pump, at some little distance from the testing parts of the Wheatstone’s 
bridge. 
The radiation wire, in the present experiments of platinum, was stretched from end 
to end of a long straight tube, and this tube was kept permanently connected to a 
five-fall Sprengel pump. In my earlier experiments, including in particular one 
long and very complete series of measurements,* the platinum wire was contained in 
a glass tube, being sealed into the glass at the ends. I also tried platinum spirals in 
glass globes similar to the globes of incandescent electric lamps, although these 
spirals had the disadvantage of having the sides of the turns radiating towards each 
other. A glass envelope surrounding the radiation wire is, however, extremely 
unsatisfactory. It is impossible to tell with a material like glass, of low thermal 
conductivity and nearly opaque to heat radiations of low refrangibility, at what 
temperature the inner skin may be of a tube half an inch in internal diameter, 
with walls pg-inch thick, and containing a red-hot wire stretched at its centre from 
end to end. And it is still more difficult to form an opinion as to the amount of heat 
that may be returned by reflection and radiation together to the radiation wire. I 
shall have occasion to refer to this matter somewhat later. 
My reason for using glass envelopes in my earlier experiments was simply that I 
was unable to find a method of stretching 1 a wire in a metal tube with one end at least 
of the wire insulated, and connecting such a tube in a satisfactory manner with the 
glass exhausting tube of the Sprengel pump. Very moderate experience in the 
production of high vacuums shows the futility of attempting to work with connections 
of india-rubber, vulcanite, or the ordinary cements and varnishes. 
I have, however, been fortunate enough to find a cement which, when properly 
# ‘ British Association Report,’ 1885, and ‘ Nature,’ 1885. 
