74 
3d series. 23 c on. generated in er 
10 intervals, 6.5" each 
whole time of the series 583° 
Athser. 14 oz. generated i in 727.5" 
12 intervals, 6.8” each 82.5 
whole time of the series 810 
5th ser. 12 oz. veneiated + in 536” 
13 intervals, 12.9’ each 169 
whole time of the series "05 
6th ser 
13 intervals, 10.8’ each 140.5 
whole time of the series 675 
7th ser. 17 0z. generated in is 
ch 
14 intervals, 10” eac 
. 12 02. generated in 534.5” 
On the Production of Steam from heated Iron. 
Sth series. 2}.0z. generated in 6 
16 intervals, 6” each 96 
whole time of the series 791 
9th ser. 21 0z. generated in 724’ 
16 intervals, 1 ll” each 176 
whole time of the series 900 
10th ser. 2} .0z. generated in er re 
17 intervalé, 1 16” each Q71 
whole time of the series 870 
llthser. 1,%,0z. generated in 613° 
2A in tervals, 7’ each 168 
whole time of the series 781 
12th ser. 113 oz. gener’d in Le -d 
28 intervals, 8.2” each 
aie 
whole time of the series 707, whole time of the series 1011 
As the water covered generally but a small part of the surface of 
the basin even at the commencement of the experiment, the heat in 
the latter terms of each series, must have been furnished to the watét 
more slowly than in the preceding terms, both on account of the 
diminution of difference between the metal and the liquid, and 
account of the necessity of depending on the conducting power 
the metal, to bring the heat from the exterior to the center of the 
basin. Hence we might expect to find the terms obeying some law 
of geometrical progression. If we examine the last seven or eight! 
experiments in each series, we shall clearly perceive such a progte® 
sion. Omitting the last of each column, as presenting anomalies 
obviously derived from the final disappearance of vaporization, 
the substitution of mere evaporation, we may divide the last number 
but one, by that which precedes it ; this latter, by the next precedings 
and so on, until we obtain five quotients. - These quotients will co” 
stitute the ratios of the series, at the particular points where the & 
periments took place. ‘The mean results for each series may then 
be obtained in the usual mode. But it will soon be perceived that if 
we extend the divisions beyond five or six, the ratio will be essentially 
varied in its character, and the series, in some instances, beco™ 
