Lvght and the volume of Gas consumed. ; 19 
In this series the lights increase in considerably higher ratio 
than is required by Farmer’s theorem, which demands 6°60 
cubic feet, corresponding to a four-fold ¢ consumption, while 
the actual consumption was 1:05 cubic feet less than the quan- 
tity required by the theorem. 
Experiment 3d.—The following series was obtained by 
another argand burner. 
Index, 062, = 3-72 cub. feet. = 1 light. 
i ‘OSI4 == -4°68 *: =. 
si ‘Ivvo = Boo ” — 
sos 200° Va = 
In this series the ratio is more nearly in accordance with 
the demands of the theorem, the intensity being still a little 
in excess of the squares of consumption (3°72 x 2 = 7:44 in 
place of 7-219). 
The gas employed in these comparisons had a candle power 
of about 14 candles. 
Experiment 4th.—Results obtained by a comparison of fish- 
tail burners, ratio as 4 and 9 feet respectively. 
A, index, 0750 = 46500 cub. feet. = : light, 
B. “1586 == 9819 —— 
In this sce medi the ratio falls but little hort i the de- 
mands of the theore 
ment 5th. fp oe of fish-tail burners. 
A. pees ‘O86. <> S16 hae oe . 85 
B. 1677 = 10°06 “ 
In this trial the departure from. the ie ee of the 
pee 3 is considerably greater than in any of the precedes - 
nts. But it appears that from some cause the 
of the squares does not hold with gas of the power ed 
these trials (14 candles), where the cous Ne rises above 9 
or falls much below 3 cubic feet. This is undoubtedly con- 
nected with the well recognized Rie that ARE is for pal gas 
a kind of burner and a volume of gas better calculated than — 
any pies to develop its maximum intensity. 
ment 6th.—This series was ae by Mr. Farmer to 
test lb a icect comparison the value of the new as contrasted 
with the old method of correction. Both trials were made 
upon the same gas, thesecond observation following Phd ie 
atter the first and with the same candle, and therefore-should 
give about the same candle power. 
32-7 grains. 
lst ieee = a = — 
Mean al Pp © 13°93 cand 
