50 
BRITISH ASSOCIATION.— 1835. 
yet taking place in a slight degree with one of only one twentieth of 
an inch diameter. 
If the conical air-tube be not inserted into the burner, but merely 
held close to its base or lower aperture, no retraction takes place, 
the flame is merely curtailed, and the combustion rendered more 
perfect; and the same result takes place when a tube equal in dia¬ 
meter to the internal part of the burner is used, in which case it is 
obvious none of the flame could retract. 
To the perfect production of the foregoing effects it is necessary 
that the apertures for the gas in the burners be of a much smaller 
size and more numerous than usual. When the axis of the conical 
air-tube is parallel with that of the burner, the direction of each 
separate jet of flame from the holes in the burner is also parallel to 
the same while the air-tube and burner are respectively concentric; 
but if, while they remain concentric, the axis of the air-tube be in¬ 
clined to that of the burner, a far more singular effect ensues: each 
separate jet of flame now in retracting describes a spiral round the 
internal surface of the burner, making from one third to perhaps 
one half a revolution. 
If the conical air-tube, while still inclined as above, be now brought 
into contact with that side of the burner towards which it is inclined, 
the obliquity of the spiral is much lessened; but the flame is so 
much retracted at the side of the burner opposite the air-tube that it 
makes its appearance out at the lower end of the burner. The same 
effects are produced whether the burners are vertically, up or down, 
or horizontal, or inclined at various angles, subject to merely the 
disturbances produced by the ascent of the neighbouring currents 
of heated air. 
The effects do not seem to depend upon difference of temperature 
between the current of air and the flame, as no change is produced 
by heating the former to upwards of 600° Fahrenheit, neither does 
the angle of the cone seem to be very essential, except it be so great 
as to nearly stop the aperture of the burner. A cylindrical tube 
answers equally well with a cone, but an inverted cone, that is, a 
tube terminating with an enlargement, will not produce the effects. 
Tubes of various other forms produce corresponding variations 
of the principal phenomena. A large flat disc, with an aperture 
just large enough to admit the burner, placed close to its perforated 
extremity, so as to prevent the passage of external currents parallel 
to the internal current of air, does not change the effects. 
The retraction is considerably lessened, however, by stopping up 
the space at the lower end of the burner, between it and the air- 
tube, but is not wholly destroyed. 
Another singular fact connected with these, remains to be men¬ 
tioned : if a glass or copper tube, of about three eighths of an inch 
greater diameter than that of the burner externally, be placed over 
it, the same sonorous effect is produced as in the well-known ex¬ 
periment of the combustion of pure hydrogen, but much louder; 
indeed, the copper tube used, which was eighteen inches long and 
