Gas-Burners, and on the Illuminating Power of the Gases. 25 
it will be observed, on kindling the gas, and gradually increasing its flow, with- 
out putting on the glass chimney, that the flame contracts by degrees at the 
top, and at length almost meets about an inch above the centre of the burner. 
In this state, the flame is yellow ; and if made longer, by increasing the flow 
of gas, it becomes brown, and smokes. The reason is obvious ; because, though 
the outer surface of the flame is properly burnt, the internal portion of the 
gas, decomposed by the heat, is not consumed. The air cannot penetrate be- 
tween the jets from the outside to the inside of the flame, and the central cur- 
rent is too feeble. In such a burner, it is necessary to enliven the combus- 
tion, by increasing the central current of air ; and this is done by means of the 
glass chimney. Now, the rapidity of the central current increases as the dis- 
tance between the flame and the chimney, or, in other words, the diameter of 
the chimney, diminishes. Accordingly, if the diameter be diminished j)y suc- 
cessive small portions, the brown smoking flame, described above, will increase 
in brightness, and the point at which it contracts will rise higher and higher, 
till at length the summit opens out, and the flame becomes quite cylindrical. 
If the experiment be pursued farther, by applying glasses still less in diame- 
ter, the intensity of the light goes on increasing, but, at the same time, the 
flame becomes shorter. Reasoning from the principles formerly laid down, we 
should presume, that, in relation to the expenditure, the greatest light will be 
emitted when the flame is~ fully opened out, and that the combustion cannot 
be farther enlivened without loss. This conclusion is quite conformable with 
the results of actual experiment. 
As an example of the different forms of glasses required for different de- 
scriptions of burners, we may mention those we found best fitted for the 
burners which were used in the series of experiments on the effect of approxi- 
mating the jet-holes, (p. 21.). The burner with 8 and that with 10 holes 
gave most light with a glass an inch and a half in diameter. When the num- 
ber of holes was increased to 15, it was requisite to lessen the diameter to 
an inch and two-tenths. The same glass answered pretty well when the holes 
were twenty in number. But when the number was twenty -five, the most 
favourable diameter was only an inch. 
The height to which the flame may be raised without smoking, or the ele- 
vation at which it gives most light relatively to the expenditure, differs very 
much in these different burners. In the burner with 8 holes, the maximum 
of light is at 4 inches; in that with 10 holes, at 3^ inches; in those with 15 
and 20 at 2| ; and in that with 25, at 2 inches only. It is singular, that all 
these flames emit nearly the same quantity of light, as may be seen by in- 
specting the numbers for each in the Table p. 21. Now, the flame of the last 
burner has only half the surface of that of the first ; and therefore, an equal 
surface of it gives twice as much light. Accordingly, the difference between 
them, as judged of directly by the eye is most striking. The flame of the 8-holed 
and 10-holed burner is streaked with blue, dull, and flickering. That of the 
15-holed and 20-holed burner is steady, smooth, without blue streaks, and of a 
peculiar sparkling star-like appearance. But that of the burner with 25 holes, 
which, for equal surfaces, gives a fifth more light than the preceding, is by far 
the most brilliant and beautiful light we have ever seen. It was formerly 
proved to be likewise the most economical. 
