of Edinburgh, Session 1883-84. 
697 
or to the simple dispersive action, as viewed by the unassisted eye, 
of no less than 1800 prisms of dense flint glass. 
The result of such an immense prismatic power on a bright con- 
tinuous kind of light, such as that of incandescent carbon, is to 
produce a coloured spectrum strip, virtually 120 feet long, or 
stretching all round the meeting hall of R. S. Ed. And the regions 
or places of its several successive colours will be the places also of 
any lines of the same colours which we may meet with in the bright 
line spectra we are about to observe. 
CH, or Carbo- Hydro gen, in Flame. 
Beginning with the compound gas CH, in its Coal-gas form, and 
burning in a blow-pipe in the open air — very nearly as described to 
this Society by Professor Swan in 1856 — the new instrument con- 
firms all his findings, with the addition of further details. 
There are, for example, 5 bands, widely separated from each 
other — the orange, the citron, the green, the blue, and the violet. 
Each of these bands is intense towards the red, vanishingly faint on 
the violet, side. Each of the first four bands, too, is made up of 
certain strong lines and much interstitial and following haze. 
But the new instrument further shows that each of those strong 
lines is double, and the haze is entirely resolvable into a far minuter 
class of lines or linelets \ excessively close on the red side, but con- 
tinually widening towards the violet. 
With each of those bands I measured between 80 and 90 of such 
linelets ; and only stopped then, not because any definite termination 
of them was reached, but because they had then become too broad, 
faint, and hazy to justify micrometrical measure being expended 
upon them. 
A grand constant was however thus obtained, of a useful character, 
for reference in certain disputed questions in spectroscopy. And if 
for sharpness of definition and precision of detail it left much to be 
desired, I endeavoured to supply that by subsequently employing 
as the illuminant, not flame of any kind, but the well-known electric 
spark of the induction coil. Even this, however, may be sometimes 
insufficient for definition purposes, unless appropriately used ; for 
The simple induction spark, tried on chloride of sodium vapour in 
the open air, though less hazy than flame, was of a crackly, uncertain 
