248 METEOROLOGY AND ALLIED SUBJECTS. 



of pure distilled water, in tlie absence of wMcli melted snow or raiu- 

 ;vater may be used; the thermometer is to be x>laced in a re(;eptacle, so 

 arranged that the reading of thermometers may be taken when the 

 whole is surrounded by an atmosphere of steam; and the tension of the 

 vai:>orthus surrounding must be very accurately known; for thermome- 

 ters whose ice point shows a strong depression after heating, the boiling- 

 point also shows a small depression up to one-tenth of a degree. In 

 order to be sure, it is necessary that the boihng-poiut determinations 

 should be repeated from time to time with the freezing-point determina- 

 tion between, until the maximum depression of the ice point has become 

 constant, when the boiling point will also become constant. In case the 

 I^lace of observation is changed, it is necessary to introduce a correction 

 for the variation of gravity. The International Committee of Weights 

 and Measures have recently determined to adopt as the unit of pressure 

 that of a weight of a column of mercury of normal density, temperature 

 0°, and height of 7G0 standard millimeters at the latitude of 45°, and at 

 the level of the sea. 



Pernet proposes, in connection with this, to call the temperature of 

 boiling water corresponding to this normal pressure, 100'^ C. (equal 

 2120 F.), and thus also to obtain a uniform unit for the measurements of 

 temperature, instead of accepting as the unit of absolute pressure two 

 diiferent mercurial columns in the laboratories at Kew, London, and 

 Paris. Every warming of the thermometer brings a new change in the 

 freezing and boiling points. The maximum depression which is attained 

 after many years of quiet and maiij' days in ice can be expressed by the 

 following formula: 



«- 1002 



where d is the maximum depression for given temperature (^), D the 

 maximum depression for 100° C, both computed for thermometers that 

 have remained a long time in the ice. In general the thermometeis 

 made of French glass show smaller depression than those made of Bo- 

 hemian glass. {Z. 0. G. M., Xiy, 1879, p. 134.) 



Thiesen remarks that the method of calibration of thermometers 

 taught by Professor Neumann has considerable advantages over that 

 of Bessel. Let 0, 1, 2 .... ?i be points equally distant from each 

 other on the thermometer scale whose corrections are to be directly 

 determined, (the so-called principal points,) and let A o, A i, ... A, > be 

 the corrections for these points. Let the corrections of the intervals 

 between the principal i^oints, the so-called principal intervals, be '?i, 

 t>2 . . . . '9n7 so that in general '9; = Ai — A^. The mercurial threads 

 to be used for calibration should now be so chosen that they as ac- 

 curately as possible include a whole number of principal intervals. If 

 now the upper end of such a thread be in neighborhood of a principal 

 y)oiut *', then will the lower end be opposite a principal point l\ The 

 volume of such a thread can be expressed by/i.^; its apx)arent length in 



