CONTINUOUS ELECTRIC CALORIMETRY, 
105 
a short spiral of about 2‘5 centims. diameter. The vacuum jacket EF extends about 
13 centims. along the inflow and outflow tubes, and is provided with a side tube for 
exhausting. The inflow and outflow tubes are provided with two side openings, the 
smaller of which was intended for the inflow or outflow of mercury, and the larger 
for the leads conveying the electric current. In some of the earlier experiments the 
larger tubes were fitted with a pair of delicate mercury-thermometers for watching 
the progress of the experiment, and observing when the conditions became steady. 
But these mercury-thermometers were found to be of little or no use, and the large 
side-tubes were subsequently removed to facilitate the fitting of the calorimeter in 
a tubular form of water-jacket. 
(26.) Method of Determining the True Mean Temperature of Outflow. 
By far the most important point in this method of calorimetry is the device 
adopted for obtaining the true mean temperature of the outflowing liquid, and 
securing a definite measurement of the electrical watts expended in heating it. If 
a thermometer were merely inserted in the outflow-tube, leaving a free space all 
round for the circulation of the liquid, it is evident that the heated liquid would tend 
to flow in a stream along the top of the outflow-tube, and that the thermometer 
might indicate a temperature which had little or no relation to the mean temperature 
of the stream. It is easy to make an error of 20 per cent, in this manner, as I 
found in my preliminary experiments in the summer of 1896. A fairly uniform 
distribution of the flow might he secured by making the space between the 
thermometer and the outflow-tube very narrow. But this leads to another difficulty 
in the case of mercury. As the space is narrowed, the electrical resistance is 
increased, and an appreciable quantity of heat, which cannot be accurately estimated, 
is generated in the vicinity of the thermometers. 
Both the difficulties above mentioned were overcome in the mercury experiment 
by fitting the inflow and outflow tubes with soft iron cylinders, 6 centims. long, 
turned to fit the tubes, and bored to fit the thermometers. The soft iron had a 
conductivity about ten times that of mercury for both heat and electricity. The 
heat generated by the current in the immediate vicinity of the thermometer bulbs 
was so small that the watts might fairly be calculated from the difference of 
potential between the iron blocks at the middle points of the bulbs. The mercury 
stream was forced to circulate in a spiral screw thread of suitable dimensions cut in 
the outer surface of the blocks, which prevented the formation of stream-lines along 
one side of the tube, and secured uniformity of temperature throughout the cross 
section of the outflow-tube. The high conductivity of the iron also assisted in 
securing the same result. 
A precisely analogous device for averaging the outflow temperature was applied in 
the water calorimeter. The bulb of the thermometer was fitted with a copper sleeve 
VOL. CXCIX.—A. 
P 
