HYDROGEN AND OXYGEN IN CONTACT WITH HOT SURFACES. 
11 
at any given moment is indicated by the mercurial gauge M, fused into the horizontal 
tube KK at a point near to the globe A. 
It is necessary for the complete success of a given series of experiments that the 
rate of circulation shall be maintained strictly uniform for, if need be, many days 
together. This was secured by automatically regulating The suction of the water 
pump employed to work the Sprengel circulating pump. 
The suction of the water pump, applied at o, raises mercury from the trough H, a 
short distance up the tube f which just rests on the surface of the mercury. The 
mercury is broken up into short threads by the entrance of air (filtered through 
cotton-wool) through the horizontal arm of the wider T-piece g , the vertical branch 
ot which reaches to the bottom ot the trough H. The force is now sufficient to drive 
the short threads up to the top of the tube f whence they fall into the reservoir k. 
The circulation of the gases in the apparatus can be adjusted to any desired rate, 
within considerable limits, by means of the stop-cocks m, n, which serve to regulate 
the flow of mercury down the full tube of the circulating pump.* 
The capacity of the whole apparatus is about 1500 cub. centims., that of the 
combustion tube only about 75 to 125 cub. centims., according to the closeness with 
which it is packed with the catalysing material. In considering the experimental 
results, it is useful to remember that the dimensions of the apparatus are such that a 
fall in pressure of, say, x millimetres during a given time interval corresponds to the 
combination of as nearly as possible 2x cubic centimetres of the reacting gases 
(2H 2 + 0 2 ).f 
Sampling of the Gases during an Experiment. 
It was frequently desirable, and sometimes necessary, to check the pressure records 
by analysing samples of the gas cut off at regular intervals during an experiment. 
In order to do this without altering the pressure in the apparatus, or interfering with 
the course of the reaction, the device of “ bye-pass ” sampling tubes shown in 
Diagram I., fig. 3, was employed. When required, these tubes were fused into the 
circulation apparatus, between the worm E and the circulating pump F, at points 
indicated by the dotted lines pp. At the beginning of an experiment the taps ss at 
the ends of each of the tubes were opened, and the taps tt in the “ bye-passes” closed. 
It was thus only necessary to close the taps ss at each end of one of the tubes and 
* H lie rate of circulation, which is, of course, strictly uniform throughout a given series of experiments, 
can be varied in different series, so that the gases can be made to complete a circuit in from 30 to about 
120 minutes. Such abbreviated expressions as “rate of circulation = 1 in 30 minutes,” often used 
throughout this paper, mean a rate such that the gases completed the circuit once every 30 minutes. 
t For the purposes of this research two circulation apparatuses were installed in the laboratory. This 
not only effected a considerable economy of time, but it also permitted of simultaneous examinations of 
the action of two different surfaces. This plan was so advantageous that during part of the research a 
third apparatus was requisitioned. 
c 2 
