436 Proceedings of the Royal Society of Edinburgh. [Sess. 
XXVII. — A Common Thermometric Error in the Determination 
of Boiling-Points under Reduced Pressure. By Alexander 
Smith and Alan W. C. Menzies. 
(MS. received April 13, 1910. Read May 16, 1910.) 
(Abstract.) 
When the bulb of a thermometer is enclosed in an evacuated vessel, the 
dilatation of the bulb introduces a considerable error in the temperature 
readings. This fact may be well known, but in the literature of boiling- 
point and vapour-pressure determinations we have observed no reference to 
it, and no corrections on account of it.* Yet, except in the roughest work, 
this effect cannot be ignored. Thus, a test carried out with eleven thermo- 
meters showed that when the pressure round the bulb was lowered from 
748 mm. to 20 mm. and thermal equilibrium with the bath had been 
recovered, the readings were from 0T0° to 0’17° lower. In all but one case, 
when there was a slight permanent dilatation, the change was constant and 
was a linear function of the change in pressure. The change bore no 
relation to the sizes of the bulbs. The thickness of the glass varied con- 
siderably, but could not, of course, be measured. 
Apparently, excepting in the roughest work, every thermometer to be 
used in vacuum distillation, or for measuring vapour pressures by methods 
which involve reducing the pressure round the bulb, should have its 
constant determined separately. The absence of corrections on this account 
vitiates many of the published data. 
* Since the above was written, a single, inconspicuous instance has come to our notice. 
Chicago, February 1910. 
{Issued separately June 16, 1910.) 
