584 MR EDMUND J. MILLS’S 
near to g, and otherwise adjusting the mercury as was necessary. About ten 
‘“‘readings” on the average were taken to find a “‘ mean reading,” and from this 
a “single comparison” was deduced. Altogether thirty-three “ single compari- 
sons” were made. It was considered advisable to employ two glass helices, 
whose respective capacities were (1) 71°7 cubic centims. and (2) 62:0 cubic 
centims. The correction for the barometer was made at Kew; its readings 
were reduced by ScoumaAcHER’s tables. The calculations, which were made in 
accordance with the well-known formula,* 
ig, v 1+,’ pe 
fe h+ ay Hite — eli ae a 
my) 1+’ piteey | 
a[ i wy Site Thaagt He 
and verified, proved extremely tedious. They were all based on a determina- 
tion of the coefficient of expansion of air, which can of course be readily 
determined by this instrument from observations at 0° and at the boiling-point 
of water. As a mean of four closely accordant determinations, the value 
a= ‘0036772 was adopted. The results obtained by Macnus, REGNAvLT, and 
JOLLY range from ‘0036678 to -0036645, the higher number having much greater 
weight. The use of a coefficient determined with the apparatus actually em- 
ployed in making the thermometer comparisons, destroys a number of small 
errors which might otherwise have an appreciable influence on the numbers 
obtained. The numerical results of the comparisons were plotted out on paper, 
the values of At (=Therm. 2— Air therm.) being taken along the ordinate, of 
therm. 2 along the abscissa. Points 1 and 2 were joined and bisected, 
points 2 and 3 also joined and bisected, &c.; the bisection points were then 
similarly treated, and the new lines thus een again bisected, &c. ; finally, a 
freehand curve was drawn through the points last arrived at. The following 
data were then taken from this curve— 
i 
Therm. 2. At. 
114-00 066 
65°61 123 
31:88 198 
These numbers lead to the equation 
At='012901¢— 0002493777 + 7000001239323, 
which gives At='036 (instead of 000) when ¢=100°. Of the three constants 
on the right hand side, the third, depending on the higher temperatures, is by 
* Wiiner, Lehrbuch der Experimentalphysik, iii. 104. 
