158 
DR. H. T. BARNES ON THE CAPACITY FOR HEAT OF WATER 
If h depends only on dd, then for different values of dd, for the same mean 
temperature and flow, we have 
4‘2 Q 8 fl- h — A = constant. 
This relation should hold provided the temperature coefficient of both 8 and h is 
linear. A small variation from lineality can, however, be safely neglected. 
If we vary the flow and keep the rise of temperature constant, then we have in the 
equation already given 
4-2 Q 8 + h = A. 
The value of A for different values of Q will vary in proportion to 4 - 2 Q 8, but the 
variation will be a linear one, provided we are not neglecting any term on the left- 
hand side varying inversely as the flow. If 8 = 0, then 
7 EC - JQ dd 
h = — w- 
for any value of Q. 
In Section 5 the experimental proof of these considerations is given, and it is 
shown that within wide limits of flow it holds with great accuracy. For very small 
values of Q the conduction becomes measurable, but the limits chosen in the present 
series of experiments are seen to hold for the higher temperatures as perfectly as for 
the lower. 
Sec. 3 .—Measurement of Fundamental Constants. 
Owing to the importance attached to the measurement of the different constants 
in the general equation of the method, it is proposed to treat each one separately in 
this section, dividing them up under the two heads electrical and thermal. In the 
first we have the Clark cell, standard resistance and potential measurements, and in 
the second the measurement of temperature, weight and time. 
The general plan of the electrical connections is given in fig. 3. A large 4-cell 
accumulator, of 200 ampere-hours each, supplied the steady heating current to the 
calorimeter through the resistance and rheostat. Potential terminals were taken 
from the calorimeter and resistance, and from two Clark cells in series, to the paraffin 
block, where they were placed in mercury cups cut in the solid paraffin. Wires 
leading from two holes, placed equi-distant from the other cups in the block, were 
carried to the potentiometer shown to the left, and included a galvanometer in the 
circuit. By interchanging two connections, the Clark cell, calorimeter or resistance 
could be connected through the galvanometer to the potentiometer. By altering the 
rheostat or connecting a smaller number of cells, the heating current could be 
