474 SCIENTIFIC RECORD FOR 1882. 



Michelson has also devised another form of thermometer of preat 

 sensitiveness, based on the high expansion coefficient of hard rubber. 

 A very thin strip of this substance is fastened to a similar strip of 

 brass 5<="i long, I™'" wide, and about O.l'"™ thick. Tlie lower end of the 

 compound strip is fixed, and on the ui)])er end is fixed a fine thread of 

 glass, the end of which is bent at right angles, and impinges upon a 

 light mirror of silvered glass suspended by a vertical silk filament. W heu 

 the temperature changes, the compound bar curves and the glass thread 

 rotates the mirror. The displacement is observed with a telescope and 

 scale. The whole is inclosed in a double box, to avoid air currents. In 

 another form of the instrument the curvature of the compound bar is 

 made to bring the end near a metallic rod immersed in liquid, thus 

 shortening the column and diminishing the resistance. (-/ Phys., April, 

 II, I, p. 183.) 



Govi has made use of the fact that the expansion coefficients of mer- 

 cury and hard rubber are very nearly equal, for the purpose of illustrit- 

 ing the effect of expansion on the bulbs of liquid thermometers. A 

 capillary tube of glass has a bulb of ebonite, the bulb and a part of the 

 stem being filled with mercury. No change is observed in the height 

 of the column when the temperature gradually changes. When tlie va- 

 riations are rapid the mercury descends with heat and rises with cold, 

 the original level being soon restored. The ebonite responds first, and 

 being a bad conductor, transmits its heat slowly to the mercury. {Na- 

 ture^ December, xxvii, p. 209.) 



Crafts has studied the conditions in mercurial thermometers which 

 produce a lowering of the zero point. He considers that the glass 

 after it has been softened in the blowing preserves for an indefinite 

 time at ordinary temperatures a C(>ndition of strain, as in the Prince 

 Rupert's drop. When the glass is reheated, the mobility of its ])articles 

 increases and a contraction takes place, the result of it being the more 

 effective in i)roportion as the original temperature is the more nearly 

 approached. In constructing the thermometer this contraction pro- 

 duces a I'.ermanent elevation of the zero point, which may amount even 

 to 26°. When heated to the boiling point of mercury, as in tilling, there 

 is a similar effect, but the depression rarely amounts to more than 2° 

 for lead glass. To remove this error, therefore, the author heats the 

 thermometers to the required high temperatures and then cools slowly. 

 {Gomptes Rendus, May, xciv, p. 1298.) Subsequently Crafts has given 

 the results of his comparison of fifteen mercury thenuometers with the 

 hydrogen thermometer, all immersed in the vapor of nai)hthalene. 

 Seven of these thermometers were made by Bandin and seven by Alver- 

 gniat. They were all of flint glass containing 18 i)er cent, of lead. One 

 was made of soda glass, by Miiller, of Bonn. These results are given 

 in tabular form, showing the correction required for every ten degrees 

 from lOOO to 330O. For comparison, the corrections for thermometers 

 made of Choisy le Boy glass and ordinary Paris glass, froui Eegnault's 



