A series of important articles on the preparation and testing of standard thermometers have been communicated to the Zeitschrift für Instrtimentenkunde by Drs. Fernet, Jaeger, and Gumlich, of the Physikalisch-Technische Reichsanstalt. The selection of the best glass, the calibration of the thermometers, the determination of the coefficients of external and internal pressure, and the verification of the principal points are fully dealt with. One source of error in thermometers as usually constructed, lies in the fact of the bulbs being blown from the tubes. The vaporisation of certain constituents of the glass during this operation leads to a difference of chemical constitution between the stem and the bulb. This may be obviated by making the bulbs out of thin walled tubes of the same kind of glass, and welding them on to the stems. As regards the depression of the freezing point, it was found by Wiebe and Schott, of Jena, that glasses containing either sodium or potassium, but not both, showed this after-effect to the least extent. In order to render the reading of temperatures accurate to within O°˙'OO2, the length of a degree should not be less than 6 mm., and since the length of the stem cannot conveniently exceed 60 cm., the range of measurable temperature is practically limited to 100°. Stem thermometers without enamel backs or enclosing tubes were the only ones found suitable for first-class standards. When certain fixed points outside the scale were to be brought in, this was accomplished by widening out the tube above them. An equal linear division of the scale was adopted, this having great advantages over the more or less untrustworthy division by equal volumes. For calibration, threads of mercury of different lengths were cut off from the main portion and measured with micrometer microscopes, viewing them both through the face and the back of the stem. But the threads were not cut off by local heating, since that is apt to produce a permanent change of capacity. The small and almost microscopic bubble which remains in every thermometer was made use of. It was brought to the entrance of the bulb when the desired portion of the thread had been driven into the stem, and then a slight jerk sufficed to cut off the required length. To facilitate this operation, the bulb was narrowed to a neck at the entrance to the stem. As regards pressure, two factors had to be considered. The external atmospheric pressure, and the pressure of the liquid in which it is immersed, tend to compress the glass vessel and to produce an apparent elevation of temperature. The capillary pressure of the mercury, and its hydrostatic pressure, on the other hand, tend to widen the bulb and produce an apparent cooling. The first of these elements was investigated by exposing the thermometer to various high and low pressures in a glycerine bath, and the second by observing the readings when the thermometer stood horizontally and vertically respectively, at its highest measurable temperature. The capillary pressure was found to be too capricious to be accurately measured, but it is a negligible quantity. The coefficient of apparent expansion of mercury in the new Jena glass thermometer 16m was found to be 0˙0001571 between 0° and 100°.
Read full abstract