370 Miscellaneous, intelligence. 



in the oven ; introduce shot or sand into it till it almost sinks in 

 water, seal it hermetically, and draw out one part of the neck 

 until not more than a line in diameter, ihat part being about an 

 inch in length ; fasten a small basin on the top of the neck with 

 wax, and then, putting the instrument in water of a certain tem- 

 perature, 40° F. for instance, put weights in the cup till the 

 surface of the water is at the middle of the narrow part of the 

 neck ; then lay the instrument aside for some days, or better 

 still, some weeks or months, and after that time, again immerse 

 it in the same water at the same temperature and pressure, and 

 with the same weight ; the instrument will now sink lower than 

 before, in consequence of its diminished bulk from gradual con- 

 traction of the glass. 



It was found that, although the effect was greatest after the 

 glass had been rendered soft by heat, yet that it occurred also 

 when the elevation of temperature had not extended nearly to the 

 softening of the glass, and indeed more or less upon every rise 

 of temperature. We have referred to an illustration of this at 

 p. 160 of our last Number. Hence two kinds of irregularity in 

 thermometers arise from the same cause. The one is mani- 

 fested soon after the formation of the instrument, increases to 

 a certain degree, and then remains stationary : this may be 

 rectified by elevating the scale of the instrument the required 

 quantity. The other takes place at every change of temperature ; 

 it is small and scarcely perceptible, with small changes of tem- 

 perature, but by considerable changes becomes very evident 

 and important. 



Singular consequences sometimes result from the influence 

 of these changes. If two liquids be taken of different tem- 

 peratures, a greater difference will be found between them, 

 by trying the hot fluid, and then the cold fluid by the same 

 thermometer, than what will appear to exist by trying the cold 

 fluid first. Again, if a new thermometer be graduated by an 

 old one preserved as a standard, although it may be made to 

 agree with it, yet, after a while, the two will not accord ; and 

 if two old thermometers be taken that do agree, and the one be 

 heated whilst the other remains unused, they will no longer in- 

 dicate the same temperatures. 



The reason now becomes evident, why alcohol thermometers 

 are so much less affected in this manner, than those filled with 

 mercury. Alcohol expands several times more than mercury, 

 so that an instrument constructed with it having a tube of the 

 same diameter, and degrees of the same size, will require a 

 bulb several times less than if mercury had been used. Hence, 

 as the elevation is in proportion to the capacity of the bulb, 

 independent of the liquid it contains, the alcohol thermometer 

 will exhibit a much smaller effect than the mercurial instrument. 



