Instrument for measuring the expansion of Solid Bodies. 325 



part, the different results of philosophers in some of their investiga- 

 tions on temperature, and as an example of this, the temperature of 

 of distilled water at its maximium density may be quoted. 



Although the flint glass experimented on by Biot and Arago, had 

 such a rate of expansion as to counteract almost exactly the increas- 

 ing rate of expansion in mercury and thus produce the effect of a 

 uniform expansion of the mercury, yet the flint glass manufactured 

 for thermometer tubes is not composed of the same proportions of 

 the materials in different manufactoi'ies and is of different densities 

 and rates of expansion, and hence it follows, that its rate of expansion 

 in most cases differs from that of mercury, and consequently, the 

 compensation will not be exact. 



Again, glass is highly elastic, and the increasing length of the col- 

 umn of mercury, causes the bulb, which is of thin glass, to increase 

 its capacity by the effect of hydrostatic pressure, and as all the bulbs 

 have not the same thickness in relation to their capacities, it follows, 

 that in different thermometers with even the same length of columns 

 of mercury, the effect of the hydrostatic pressure, in increasing the 

 capacities of the bulbs, would prevent them from being sti'ictly com- 

 parable. There are other sources of error too well known to require 

 notice. 



Of all the metallic thermometers, those of Borda and Breguet, are 

 perhaps least objectionable. The first, to be very accurate, requires 

 the bars to be very long and it then becomes cumbrous : the other 

 has too many sources of error to be regarded as an accurate instru- 

 ment. 



Air thermometers are the most accurate of all, yet they labor 

 under this objection, that while the air expands uniformly for equal 

 increments of temperature, the containing body expands in a slightly 

 increasing ratio as the temperature increases, thus enlarging its ca- 

 pacity, so that a less temperature is indicated than the true one. 

 This is true as well with the differential as with the other air ther- 

 mometers, but the inaccuracy is so slight as to render it inappreciable 

 at ordinary ranges of temperature. 



The published details of the measurement of the base lines of vari- 

 ous extensive trigonometrical surveys, of the measurement of the 

 arcs of the meridian, and of the establishing of standards of weight, 

 length and capacity, show how much labor, thought and science 

 have been employed in endeavoring to arrive at rigid accuracy. The 

 thermometer has been the principal stumbling block, in consequence 



