132 Hartiey—An Investigation of the Connexion between 
are in the flame; they are wider apart, and individual bands are wider, as the 
vapour-densities of these elements are greater. They do not shift appreciably in 
relation to their atomic weights towards the red or less refrangible rays, and 
probably therefore have not an atomic or molecular origin. It appears as if they 
originate from the mass of vapour in the flame, and are dependent apparently 
also on the vapour-pressure. The exact cause of the bands has not yet 
been definitely ascertained, but they are being submitted to a more searching 
examination. That they do not arise from the atomic constitution of the vapours 
is, I think, shown by a somewhat similar group of bands being photographed 
from the flame produced when ether-vapour mixed with hydrogen is burnt in 
the oxyhydrogen blow-pipe.* 
+Second.—There are diffuse bands degraded on one side, but sharp and intense 
on the other. Such a band has been photographed in the red rays of potassium, 
but as a rule it is too feeble to show this distinctive character. There are also 
bands which extend on either side of a line, such as the broadening of the red 
line of lithium, and the yellow lines of sodium. These also depend on the 
volume of vapour in the flame. 
Third.—Bands of a very definite character, such as those of calcium, 
strontium, and barium, not always degraded on one side or the other. These 
bands contain rays that are more intense than those on either side of them, and 
have the appearance of lines not well defined. They become intensified, sharper, 
and narrower by supplying the flame with hydrochloric acid; and, by the 
action of the spark, lines are developed from them. ‘The bands become broadened 
and intensified by increasing the quantity of the element in the flame. 
* The photograph showing these bands is at present in the Franco-British Exhibition. 
} With reference to the bands in the blue in the spectrum of silver, the conditions under which they 
have recently been photographed seemed to render it possible that they are due to interference ; but as they 
were photographed in 1890 by a method which excludes this possibility, and yet renders the band sharper 
and more distinct, such a view is untenable. In a friendly communication for which I am indebted to 
Professor E. T. Whittaker, F.R.S., the opinion was expressed that instead of interference it might be 
possible to suggest ‘‘ absorption, resonance, and mutual interaction of vibrating systems.’’ In connexion 
with this subject, a publication by Professor R. W. Wood has recently appeared on Resonance Spectra 
of Sodium Vapour (Phil. Mag., 1908 [vi], 15, 581-601), in which an extremely complicated absorption 
spectrum has been described, showing about 6000 lines in the bluish-green. When illuminated by 
white light, the vapour becomes fluorescent, and gives an emission spectrum which is an exact counterpart 
of the absorption spectrum. From this it would appear probable that very small layers of vapour of the 
alkali metals would suffice to give some of these bands, due to absorption or to resonance, since the 
vapour is brilliantly illuminated by white light; but absorption could not be applied to the appearance 
of similar bands when nothing but ether vapour is burnt in the flame. 
{ Compare Spectra 1, 2, and 3 of fig. 1, and Nos. 1, 2, 3, 4, 5, and 6 of fig 2, Plate xx. Scient. Proc. 
Roy. Dublin Soc., vol. xi. (N.S.), p. 237, 1907. -- 
