204 
DR. H. T. BARNES ON THE CAPACITY FOR HEAT OF WATER 
connecting tube was fused together on the pump so that the vacuum chamber was 
permanently and hermetically sealed. The jacket was carefully heated during 
exhaustion so as to drive off water-vapour and occluded gas from the glass. 
Three other calorimeters have been made by Eimer and Amend of New York, 
since September, 1898. The vacuum jackets of these were all exhausted by them 
while heating the calorimeter in an asbestos oven. These calorimeters were of 
a similar design to the earlier ones, except in having only one side tube at the inflow 
end, and in having fine-bore flow tubes of different sizes, ranging from 2 millims. to 
a little over 3 millims. One of the calorimeters was prepared with phosphorus 
pentoxide in the vacuum chamber, but this proved to be rather a drawback than 
otherwise, because of the greater capacity for heat of the calorimeter introduced by 
the P 3 0 5 . It was essential to have the jacket very perfectly exhausted to avoid the 
heat-loss due to convection currents of air, which acted in such a way as to make the 
radiation loss appear large and uncertain. Small losses, however, from conduction 
and convection in the residual vapours in the jacket produced no errors on account of 
the steady temperature conditions during an experiment. The radiation loss would 
depend on the state of the glass surface, but would apparently be increased at the 
lower temperatures, after the calorimeter had been maintained for several hours 
during an experiment at one of the higher points. It was impossible always to count 
on the constancy of the heat-loss from one experiment to another, even with one 
calorimeter at the same temperature, as it depended so much on the past history. 
The change in heat-loss occasioned by the driving-off of occluded gases and vapours 
from the glass when the calorimeter was at a high temperature took place so slowly 
that, during an experiment extending over several hours, no measurable alteration in 
it could be noticed. This same effect of a slight change in heat-loss from time to 
time was also noted in the calorimeter used for the mercury experiments. 
In fig. 9 is given a cross-section of the calorimeter and interior fittings in place, in 
the water-jacket. The thermometer bulbs are shown included in their glass cases. 
These cases were about 9 millims. in. diameter, and extended the full length and 
a little beyond the ends of the outflow and inflow tubes. The ends of the thermo¬ 
meter cases over the bulbs were enclosed in thin copper cylinders, gold-plated, about 
10 centims. long. These copper cylinders served the double purpose of preventing 
generation of heat in the vicinity of the thermometer bulbs by the electric heating 
current, and of helping to equalize the temperature of the water around the bulbs. 
The heating current was conveyed to the copper cylinders by eight No. 12 copper 
wires at each end, which were soldered into slits cut for them in the ends. The 
cylinders were made with closed ends, in one of which a hole was made for soldering 
in the platinum heating wire, and in the other a special screw clutch for catching the 
wire after the cylinders were put in place in the calorimeter. Sections of the 
cylinders are given in fig. 10. 
The central heating wire for the fine-bore tube was made in three ways; either 
