522 
Proceedings of the Royal Society of Edinburgh. [Sess. 
The original intensity is I 0 , the final intensity is I, c is the concentration of 
the solution in gramme-molecules per litre, d the length of path traversed 
in centimetres, and A the molecular extinction coefficient. A is, of course, a 
function of A, the wave-length of the light in question. It may also vary 
with c, the concentration of the solution. 
In these Proceedings * I have investigated two alleged phenomena con- 
nected with absorption, and have found no evidence in favour of the exist- 
ence of either of them. In another recent paper , f using all data at present 
obtainable, I showed that the vibrating particles which correspond to the 
absorption bands of anilin colouring matters have masses of the same order 
of magnitude as the mass of an electron, whereas in the case of inorganic 
salts we are possibly dealing with ions. In order to go on with that work 
it is necessary to get more data, hence the present series of investigations. 
After considering the literature of the subject, I concluded that the most 
promising way of measuring A in the infra-red was by means of a Rubens 
linear thermopile and Du Bois Rubens ironclad galvanometer, and accord- 
ingly a galvanometer and thermopiles were obtained with the aid of a grant 
from the Carnegie Trust for the Universities of Scotland. Each thermopile 
has twenty iron-constantan junctions on a length of two centimetres, and 
has a resistance of about four ohms. The solder at the junctions is hammered 
into the form of discs of about one millimetre diameter. By placing a 
screen with a slit in it in front of the thermopile, its effective width can be 
reduced and readings taken closer together in the spectrum. The galvano- 
meter has a total resistance of 10*5 ohms, and when the light suspension 
system was used the sensitiveness was 3 x 10~ 10 amps./half mm. at one metre 
distance for a period of five seconds. 
A Boys radio-micrometer was borrowed from the Glasgow and West 
of Scotland Technical College and tested against the thermopile and 
galvanometer. The readings agreed very well. The surface for receiving 
radiation in the radio-micrometer is, however, rectangular, measuring 2x3 
sq. mm. ; hence it is not so suitable for use in the spectrum. Apart from 
that, it is not so sensitive. Boys’ own instrument^ gave a deflection of 
0*9 cm. per sq. mm. of sensitive surface for one candle, candle and scale 
being at a distance of one metre, and was about three times as sensitive 
as the radio-micrometer I used. For a period of five seconds with candle 
* “On a Question in Absorption Spectroscopy,” Proc. Boy. Soc. Edin., xxix. p. 68 (1908). 
“A Negative Attempt to detect Fluorescence Absorption,” Proc . Roy. Soc. Edin., xxix. p. 401 
(1909). 
t “ On tbe Mechanism of the Absorption Spectra of Solutions,” Proc. Roy. Soc., 82 A, 
p. 606 (1909). 
X Phil. Trans., 180 A, p. 159 (1889). 
