466 
Proceedings of the Royal Society of Edinburgh. [Sess. 
Remarks : — 
(1) (a) and (d) give the value of the original volume determined in two 
ways. 
(2) Comparing (1) with (8) we find a decrease of *023 g. in the weight of 
the dilatometer. The average decrease in weight after four experiments 
was T25 g. 
(3) Weight (6) is a rough check that no mercury has escaped in the 
course of the manipulation. 
(4) A glance at the table of temperatures shows a gradual decrease in 
the temperature with time. This can be explained on two grounds. 
Firstly, in all the observations but the first, the observer has the advantage 
of knowing the approximate position of contact and can regulate the 
change of temperature accordingly. In the first case, the temperature is 
invariably rising more or less rapidly when contact is made. Secondly, 
the gradual conduction of heat through the mercury from KB to EF will 
evidently tend to lower the temperature of contact. 
In the table of temperatures given above, the average of the six 
observations for a definite pressure is taken as the temperature of 
observation. This is justifiable on the grounds that it counterbalances the 
effects of the two factors just mentioned, and, besides, gives fairly consistent 
results. That the latter is true can be seen from the following four 
experiments, in which the dilatometer after one experiment was allowed 
to stand under pressure till the following day, when the observations were 
repeated. 
Experiment 
1 . 
2 . 
3. 
4. 
Average temperature for 1st day . . . 
383 
508 
476 
583 
,, ,, 5 ) 2 nd ,, 
380 
516 
478 
586 
We come now to the question of treatment of observations obtained. 
The first question which arises is — What temperature are we to ascribe to 
the liquid in the bulb when the thermoelement shows a given temperature ? 
To aid in estimating this temperature, I made a series of observations with 
two similar thermoelements placed at varying heights in the heating oven. 
The results are given in the subjoined Table, the distances being measured 
from the lower surface of the plug S. 
