140 Further Experiments on the [Marcn, 
I should have premised, that to prevent the straps of gold and silver 
from curving within the cylinder, when heated, they were held flat, 
one on each side of the iron bar, by coils of copper wire at distances of 
six inches apart—these were not so tightly bound as to impede free 
motion longitudinally. 
' At this period of the experiments it was determined to anneal the 
gold and silver bars, to observe what difference would be caused in 
their rate of dilatation thereby, as well as what would be the perma- 
nent elongation due to this change of condition. 
' To effect the annealment of such long slips of metal in the most 
equable manner without endangering loss or accident, required certain 
precautions. They were laid upon a flat bar of wrought iron, supported 
at distances of a foot asunder by fire-bricks, as represented in fig. 4, 
Pl. vii. Their whole length was then enveloped in gobar, or cakes of 
cowdung, in the same manner as is practised in heating the felly of 
a wheel. The heat was thus gradually raised, until the whole length 
was uniformly of a glowing red. But, not to lose the opportunity 
which this experiment afforded of ascertaining the relative expansions 
of the three metals at this higher temperature, an iron stake had been 
firmly fixed in the ground at one end of the bars, against which all three 
were made to abut firmly: the other ends were connected by an inter- 
mediate brass rod (kept cold) with the nonius of a sliding scale placed 
on the ground ina line with the bars, so as to measure off their elonga-_ 
tion with great facility. The results, and the temperature by Farenheit’s 
thermometer founded on the assumption of an equal rate of expansion 
throughout the scale of each metal, were as follows. 
Absolute expansion in inches. Deduced temperature. 
The gold, placed uppermost, ........ 1.638 L787 | 
The silver, in the middle, .......... 2.008 1655 
Theron, undermest,) !. Se. - sek 1.240 1609 
That the upper position was much hotter than the lower was evident, 
nor does it seem surprising that the difference of temperature should have 
been so much as 180 degrees. No knowledge therefore could be gained 
on the point sought, namely, the relative ratios of expansion ; but the 
method of operating is itself capable of further application, and I hope 
hereafter to be able to pursue it to more conclusive results. 
The absolute elongation of the precious metals, by annealment, was. 
measured by placing them once more under the microscopes at the 
same temperature as before, (77°.2.) It was found to be much less than 
was calculated from the difference of specific gravity before and after 
annealment, shewing that the compression under the rollers was in the 
gold 20, in the silver 8, times greater in the transverse than in the 
longitudinal direction. The results were as follows: 
