610 
DR. T. S. HUMPIDGE ON THE ATOMIC WEIGHT OF GLHCINUM. 
experiment; this, however, was found to he too small to affect the results to any 
appreciable extent. The loss between two experiments, even after removing the 
substance, never exceeded 0T gramme, and a difference of 0‘05 gramme would only 
produce a much smaller error than a difference of 0°'01 in reading the bent thermo¬ 
meter. In order to still further test the evaporation, the calorimeter was left open, 
exposed to the air of a warm room for 16 hours, and then only lost 0'26 gramme. 
Provided, therefore, that the agitator or substance does not rise above the level of the 
liquid, the loss by evaporation during an experiment can be very safely neglected. 
A series of blank experiments were next made to determine whether the calorimeter 
gained heat during the time that it was open to receive the substance. It was found 
that the shortest time possible in which the whole series of operations could be 
performed—the calorimeter run under the heater, the substance introduced, and the 
calorimeter returned to its original position—was five seconds, and during this time 
no change could be noticed in the thermometer T 2 . If, therefore, the operations follow 
one another promptly the gain of heat to the calorimeter may be safely said to he 
within the unavoidable errors of reading. All readings were of course made through 
a telescope.'" 
The following is the course usually adopted in the experiments. The substance, 
weighed to milligrammes, is introduced into the heater from below, this part of the 
apparatus being removed from its stand for the purpose. The heater is then returned 
to its position, the wooden slider adjusted, and the two tubes connected with the 
boiler and waste pipe respectively. During the time that the substance is heating, 
about 70 grammes of turpentine are poured into the calorimeter and weighed to 
centigrammes. The calorimeter is then placed in its brass casing, the string attached 
to the agitator, the thermometer and electro-magnets adjusted, and then a large plate 
of glass placed in front of the whole apparatus. At the end of about an hour’s 
vigorous boiling the thermometer in the heater becomes stationary, starting with all 
cold, or in about three-quarters of an hour if the instrument has been previously used. 
This maximum temperature is either higher or lower than the boiling-point, according 
to the size of the waste pipe. In the experiments already made the waste pipe was 
generally small and the maximum temperature slightly higher than the boiling-point. 
Longer heating appears unnecessary, owing probably to the small quantities used. 
As soon as thermometer T x is constant, continuous readings of it and the other 
thermometer (T 2 ) are made at regular intervals of one minute. Thermometer will 
be quite constant, and if the liquid in the calorimeter is at a temperature within 1° ol 
that of the air, the other thermometer (T 2 ) should not oscillate through more than 
0 o, 01 in five minutes. At the end of a given minute T 2 is read for the last time, 
giving the initial temperature of the calorimeter ( t ), while r Ij, giving the temperature 
to which the substance was heated (T), has been read just before. The substance is 
* I should state that in using large weights, as with the 10 grammes of silver, there was sometimes a 
little splashing, though never exceeding 0'05 gramme. 
