PRESIDENTIAL ADDRESS. 4.75 
possible therefore that the luminosity might be ascribed likewise to the metal 
separated in the ionic state. Experimental investigation undertaken with a 
view to elicit information on this subject seemed to favour the view that the 
metal was reduced by chemical processes, and that it glowed in the un-ionised 
condition. KEvidence seemed to point to the conclusion that, for example, when 
common salt is introduced into the flame of cval-gas the sodium chloride yields 
sodium by the conjoint action of steam and reducing gases; when liberation of the 
metal was prevented by adding a large quantity of hydrochloric acid to the flame 
the glow disappeared, but the conductivity was not always diminished. The fact 
that sodium salts, including the chloride, impart their characteristic glow to the 
flame of cyanogen and to other flames in which water is absent leads to some 
difficulty in finding a chemical explanation, and it must be admitted that a direct 
thermal dissociation of an alkaline halide or oxide is not out of the question. The 
interval of detachment of the metallic atom may be exceedirgly brief, but it must 
be remembered that even so short a time as the interval between the molecular 
encounters inagasat a high temperatureis still sufficient for the emission of thousands 
of undisturbed characteristic vibrations. The experiments to which I have alluded 
have been followed up with great industry and success by Professor H. A. Wilson, 
who has added much to our knowledge of the electrical condition of the flames 
containing vaporised salts; but the question of the condition of the luminous gas is 
still far from being settled. Very interesting and important investigations have 
been carried out by Lenard, who has shown that the stream of luminous vapour 
produced from a sodium salt in a Bunsen flame is deflected in an electric field in 
such a way as to indicate that the vapour is positively charged; but he gives 
reason for believing that the charged condition is intermittent with the neutral 
condition. The lines in the spectrum of an alkali metal are divisible, as is well 
known, into distinct groups or series, in each of which the oscillation frequencies 
corresponding with the lines are in a definite mathematical relationship. The prin- 
cipal series, which include the lines scen individually as such in ordinary flame 
spectra, are, according to Lenard, due to the electrically neutral atoms. In a 
salted spirit flame, and in other flames of low temperature where only lines of the 
principal series are represented, the stream of luminous gas does not behave in an 
electric field as if it were charged. In the flame of coal-gas burnt in a Bunsen 
burner the salt-vapour gives, in addition to the distinct lines of the principal 
series, diffuse bands of luminosity on the dark background, which, according to 
Lenard, represent the undeveloped subordinate series; and it is the atoms emitting 
these series that are deflected in the electric field. It is inferred, therefore, that the 
light in a salted Bunsen flame comes from different groups of centres of emission— 
the principal series from the neutral atom, and the lines of the first, second, and 
third subordinate series from. atoms which have lost respectively one, two, and 
three electrons. Lenard goes further, and shows that the salt-vapour in a Bunsen 
flame, as in the flame of the electric arc, emits these different kinds of radiation 
from different structural regions; thus the vapour at the edge of the flame is 
electrically neutral and gives only the lines of a principal series. 
The negative electricity in a salted flame would, according to Lenard, be dis- 
embodied, and recent experiments by Gold confirm the view that the negative 
carrier in flames is a free electron. In connection with this subject I ought to 
allude to an investigation by Tufts, which seems to throw some doubt on the 
conclusions which were drawn from the experiments made by Professor Wilson, 
Dr. Dawson, and myself; and I must also mention an important contribution to the 
subject recently made by Professor Hartley, in which considerable light is thrown 
upon the chemical changes undergone by compounds of the alkaline earth metals 
when they are introduced into flames, and upon the relation of these changes to the 
various spectral features. I am afraid, however, that it would be wearisome if I were 
to prolong this summary, and [I must be content to leave it without doing justice 
to those who are engaged upon the work. The subject is obviously one of funda- 
mental importance in relation to spectrum analysis, and my own slight connection 
with it has only strengthened my opinion that there is still a great deal connected 
with the genesis of spectra that requires the attention of the chemist even more 
