Jt. Messrs. Richardson, Nicol, and Parnell on the 
served to measure the rate of escape of the hydrogen from 
the heated platinum tube* FF’. In making the measure- 
ments the tap in the by-pass A was shut off ; by opening it 
the parts to the right of A could be connected with the pump 
and generator, and so could be exhausted or filled with 
hydrogen. By raising or lowering the mercury in the 
burette C, the volume of the imprisoned hydrogen, and hence 
the rate of movement of the gauge B, could be varied at will. 
The volume of mercury in the burette could be kept constant, 
by means of the clip D. The platinum tube F FI’ was 18 ems. 
long, 2 mms. external diameter, and its walls were 0:1 mm. 
thick ; it was sealed into the rather wide glass tubes shown 
at F and FE’. The heating-current entered the tube by the 
brass clips GG’ which were soldered to stout copper rods 
dipping into the mercury cups K K’. The brass clips were 
not allowed to come into contact with the platinum tube but 
were separated from it by platinum-foil, in order to prevent 
the brass from alloying with the hot platinum tube. The 
current was supplied by 6 motor-car (secondary) cells, 
capable of discharging at 42 amperes, and could be regulated 
by means of an adjustable resistance consisting of two spirals 
of thick german-silver wire which could be made to slide up 
and down in two wide glass tubes containing mercury. 
It has been pointed out that it is very important that the 
platinum tube should not be strained by the expansion and 
contraction produced by heating and cooling. It was very 
easy to get rid of straining due to the leads by making their 
ends dip into mercury cups. The rigidity of the glass 
framework F’ ELF might also have been obviated by only 
sealing the platinum tube into glass at one end and closing 
the other end by fusion. It is evident, however, that such 
an arrangement places the experimenter absolutely at the 
mercy of the purity of his hydrogen ; for the hot platinum 
acts as a very efficient filter for the impurities, which are 
stopped and collect in the tube, out of which they have little 
chance of diffusing since there is always a current of gas 
from the rest of the apparatus towards the hot tube. It is 
quite easy to see that the percentage of foreign substances in 
the gas in the tube must be increased during the experiment 
in the ratio of the volume of gas which has passed out to the 
volume of the tube. It was thought that this difficulty 
would be got rid of to a very large extent, at any rate with 
comparatively wide tubes like the one used in the present 
experiments, by having the tube vertical and open to the 
* The tubes in these experiments were made of the pure platinum 
supplied for thermometric work by Messrs. Johnson, Matthey & Co. 
