104 
PROFESSOR HUGH U CALLENDAR ON 
modifications into the arrangement as described by Gouy, the most important of which 
was to make the relay put another lamp in series, instead of breaking the circuit. 
This prevented the destructive sparking at the break, which was a serious matter 
with a 100 candle-power lamp. Of course, the greater part of the work still fell on 
the constant-pressure gas-supply, which was mainly responsible for the excellence of 
the results at high temperatures. 
(25.) Preliminary Experiments on the Specific Heat of Mercury. 
Our first experiments by the steady-flow electric method of calorimetry were made 
on the specific heat of mercury, as it presented fewer experimental difficulties than 
water. Since mercury is itself a conductor of suitable specific resistance, it was 
unnecessary to insert a heating conductor in the fine flow-tube, which greatly 
simplified the fitting together of the apparatus. The conducting properties ol 
mercury were also utilised in the design of an electrical device for maintaining 
a constant head so as to secure uniformity of flow. The level of the mercury in the 
reservoir was regulated by a platinum wire contact which actuated an electro-magnet 
Fig. 6. Diagram of Mercury Calorimeter. 
compressing a small rubber tube which supplied mercury to the reservoir. As soon 
as the level fell below the platinum point, the contact was broken, the armature 
released, and more mercury supplied. The inflow was arranged to keep the mercury 
near the platinum point in perpetual agitation, so that there was no sticking or 
hunting of the regulator. The whole arrangement was fitted to a wooden bracket 
belonging to a Geissler pump, which could be hung up on the wall at different levels 
when it was required to alter the flow of mercury in a given ratio. The flow was 
regulated and steadied by passing the mercury through fine glass tubes immersed in 
a tank of water before entering the calorimeter. In measuring the flow of mercury, 
the time of switching over was automatically recorded on an electric chronograph by 
the momentary contact of the mercury thread with an edge of platinum foil, which 
diverted the flow from one beaker to another. 
The design of the calorimeter itself will be readily understood from the accom¬ 
panying diagram (fig. 6). The inflow and outflow tubes AB and CD are exactly 
similar, about 2 centims. internal diameter and 25 centims. long. They are 
connected by the fine flow-tube BC, of 1 millim. bore and 1 metre long, coiled in 
