136 
DRS. L. MONR, W. RAMSAY, AND J. SHIELDS ON THE 
extracted from fully charged platinum hlack in vacuo at the ordinary temperature, 
whilst the other corresponds to the occlusion of the hydrogen, which can be removed 
from the platinum black in vacuo by raising its temperature from the ordinary 
temperature up to 184° C. In order to test the question wdiether the occlusion of 
different fractions of hydrogen corresponds to different heat changes, as has been 
stated by Favee and Berthelot, some additional experiments were made, the results 
of which are given in tabular form in Table 11. 
Vfhen Experiment 1. was over, the platinum black remained fully charged wdth 
hydrogen at 0° C. Before removing the experimental tube from the calorimeter for 
the purpose of re-exhausting it at 184°, an attempt was made to measure the heat 
ahsorhed on removing as much of the hydrogen as could be pumped off at 0° C., wdth 
the results given under Experiment la. As wdll be seen from the table, 2’51 cub. 
centims. of hydrogen were removed. In Experiment 16. the platinum black was 
again fully charged up with hydrogen. Tire results of these two experiments, namely, 
65’7 K ahsorhed per gram of hydrogen removed, and 63'7 K evolved per gram of 
hydrogen occluded, are sufficiently close to each other and to the mean value 6 9‘4 K, 
considering the probable errors which will be discussed immediately, to show that 
there is no difference between the thermal changes which take place when the two 
different fractions of hydrogen are occluded. In Experiment III. a small quantity of 
platinum black was employed, and it wms exhausted at 230° C. instead of at 184° C. 
before being placed in the calorimeter. Approximately, the same value was obtained 
for the heat evolved per gram of hydrogen occluded. 
The probable accuracy of the results may be estimated from the following con¬ 
siderations. The hydrogen w'as measured in a calibrated burette, H, divided into 
tenths of a cubic centimetre, which w'as read to one-hundredths of a cubic centimetre. 
The volume of hydrogen occluded (ahvays reduced to 0° C. and 760 millims.), given 
in the foregoing tables, represents the difference between the total hydrogen measured 
and that required to fill the experimental tube. The measurements are also liable to 
slight errors involved in reading temperature and pressure for the reduction of the 
gas volumes to standard conditions. The burette could be read with certainty to 
one-fiftieth of a cub. centim., but if we allow the error from all sources to be 
one-twentieth of a cub. centim., this would mean in Experiments I. and II. a possible 
error of about f per cent. ; wdiilst in Ta., 16., and III. the error might amount 
to 2-| per cent. The calorimeter deflection, which w-as observed on a scale etched 
on a plate-glass mirror, K, lying underneath the capillary tube, could be read accurately 
to OT of a millim. ; but, if we admit a total error of 0‘5 of a millnn. for plotting 
the curve and for any change of normal creepage during the experiment, then the 
error involved in the heat measurement may approach 1 per cent, in Experiments I. 
and II., and 5 per cent, in Experiments la., 16., and III. 
The total error in the most unfavourable circumstances in the first two experi- 
