My initial experiments were carried out in the Graham Laboratory at the Department of Chemistry, University College London), on June 25, 28, July 2, 5, 7, 19, 21, 27, 30, and August 2, 3, 9, 16, 20, 25, 27, in 2004. The atmospheric pressure recorded at the time of these experiments ranged from 1000mb to 1013mb. In my general discussion of results, I have not applied pressure corrections to temperature readings, since they would be quite small (maximum variation of 0.3°C) compared to the large variations observed due to other factors, and even compared to the routine fluctuations in any given experimental run.

The video footage shown were taken on the morning of 27 September 2007, when the atmospheric pressure I recorded was 1016mb.

I have used the following thermometers:

Two ordinary mercury thermometers, reading from 110°C to -10°C, graduated down to 1°C (provided by Crosby Medley). The specified immersion depth is 75mm on one, 76mm on the other.
One ordinary mercury thermometer, graduated up to 250°C (provided by Crosby Medley). This was only used to monitor the temperature of the graphite bath.
Two Beckmann thermometers, each graduated down to 0.01°C (provided by Andrea Sella). Each Beckmann thermometer has a very large reservoir of mercury, and a very long 6° scale. One can set the zero of the Beckmann thermometer wherever one wants, through a clever mechanism by which the mercury can be shifted around between the main reservoir at the bottom and an auxiliary reservoir at the top. Of the two Beckmann thermometers I used, one was calibrated to read 0° at about 99.65°C, the other one at 96.0°C. The precision in the location of the zero of the Beckmann thermometer cannot exceed that of the instrument against which it is calibrated; in this case it is the ordinary mercury thermometer, whose readings can only be trusted down to 0.1°C at best. All temperatures recorded down to the second decimal place (0.01°C) in this report were taken with the Beckmann thermometers. On three occasions the mercury column in the first Beckmann thermometer broke up, but I was able to make it continuous again and the calibration was re-confirmed afterwards. On the Beckmann thermometers, no specification of required depth of immersion can be found. However, the indication from various operations is that no significant difference is made as long as the bulb is fully immersed in the water.
A modern digital thermometer (Pt 100 type). This is an electrical thermometer, based on the temperature-dependence of the electrical resistance of metals; the instrument I used has a platinum resistance element at the tip of the probe. This particular instrument did not add greatly to precision, but it gave an independent confirmation for the zero-point calibration of the Beckmann thermometers. I had also hoped for a very quick reaction time from this instrument, but I was disappointed in this expectation (as can be see in the jumpiness of the readings in Experiment 1).

Calibration between the three types of thermometers: I compared the readings of these three thermometers by putting all of them in one and the same particular situation, which was immersion in steam emerging from water boiling in a particular metallic vessel. This was a stainless steel cylinder with a narrow opening (actually, a clean oil-dispenser with the spout removed), which was filled only part of the way with distilled water; when the water was boiled, the space above the water was filled with boiled-off steam. So we can see that the ordinary mercury thermometer and the digital thermometer agree more or less with each other, and the main Beckmann thermometer I used is set to read about 0.35° at 100°C.

Return to the Experiments section of the main page

Return to the start of the main page