TRANSACTIONS OF SECTION A. 403 
2. The Variation of the Specific Heat of Mercury at High 
Temperatures. By Professor H. T. Barnes, D.Sc. 
In 1902 the author communicated to the British Association the 
preliminary results of an investigation of the variation of the specific 
heat of mercury with temperature between 0° and 100° C. Since then 
the work has been carried over a wider temperature interval, and it has 
been possible to extend the curve of variation towards the boiling-point 
of the mercury. 
In the older work the continuous method of calorimetry was used, 
including much of the apparatus employed by the author in his measure- 
ments of the specific heat of water. Thus the thermometers were’ the 
same, as well as the Clark cells and resistance standards. It should be 
possible, then, to express the specific heat of mercury with considerable 
accuracy quite independent of any values assumed for the electrical 
standards. 
The specific heat of mercury above 100° C. was also obtained by a 
continuous method, and therefore possessed the advantages of steady 
temperature conditions. 
A stream of mercury was heated while flowing through a fine steel tube 
in the vapour of some suitable organic substance possessing a steady 
boiling-point up to the desired temperature. It then passed directly into 
the calorimeter, which was provided with an inflow and outflow tube fitted 
with thermometers. Directly through the heart of the mercury passed a 
fine glass tube, through which a stream of water at room temperature 
flowed. Thermometers in the water gave the rise of temperature, which 
with the flow gave a measure of the heat taken out of the mercury. The 
temperature fall in the mercury, together with a knowledge of the flow, 
gave a measure of the cooling effect produced, and hence, after correcting 
for heat loss, of the specific heat. By this method the flows were arranged 
so that the mercury was seldom cooled more than 40° C., while the water 
was not warmed more than 10° or 15° C. Thus the average specific heat 
could be determined at the high temperature over a comparatively small 
interval of temperature, while the heat extracted by the water could 
be measured over a temperature interval in which the specific heat of the 
water was accurately known. The heat loss was determined by special 
experiments with no water flowing. This was found to be proportional to 
the mean temperature of the mercury in the calorimeter. Having deter- 
mined the heat loss for several mean temperatures between 100° and 200° 
the curve was plotted, and for each calorimeter the heat loss was read 
off for any intermediate mean temperature. Differential platinum 
thermometers were used to obtain the temperature, except in some of the 
experiments when sensitive mercury thermometers were placed in the water 
inflow and outflow. The readings of these mercury thermometers were 
reduced by direct comparison with a platinum thermometer. 
The author has recalculated all of his older observations, and these 
results have been reduced to the normal hydrogen scale. The results 
obtained by the later work have been reduced to the same scale. It was 
not possible to obtain such accuracy in the later work as characterised 
the observations by the electrical method at the lower temperatures. This 
was partly due to want of water jacketing and to the impossibility of 
airanging a vacuum jacket about the calorimeter. Lagging was employed 
of asbestos string, wound on carefully, and glass wool. The observations 
check out to one or two parts in 1,000, and it is unlikely that the curve 
is in error by more than that, if as much. All of the observations at 
the lower range taken by the electrical method with water and vacuum 
jackets are of much greater accuracy. It is very unlikely that they are 
in error by more than one or two parts in 10,000. 
DD2 
