1903-4.] Mr Milne on a New Form of Spectrophotometer, 341 
naturally from the light source with its different rays inclined in 
various directions, there would be the following difficulty. Some 
of those rays which are inclined downwards would enter the tube 
above the liquid, but before leaving the tube, they would pass 
into the liquid, and so the emergent lower beam would consist 
only partially of rays that have passed through the whole length 
of the liquid. 
At first sight it might he supposed that this error is compen- 
sated by a similar addition from the lower to the upper beam, hut 
this is not the case, for, as will be seen on reflection, each beam 
would thus gain equal quantities of light, whereas, did complete 
compensation occur, the gains of the lower and of the upper beams 
respectively would bear a ratio to one another which is equal to 
the fraction of the total light, incident upon it, which is trans- 
mitted by the absorbing liquid. 
In reality too the number of rays passing from the lower to the 
upper beam within the absorption vessel is not equal to the 
number passing from the upper to the lower, because a large pro- 
portion of the former rays will he totally reflected down again at 
the surface of the liquid ; and consideration will show that this 
fact will make the error spoken of above still greater. 
There is also the further point that with non-parallel light and 
a long absorption tube the number of rays that pass out through 
the sides of the tube will be different for the upper and for the 
lower part of the tube, owing to the presence of the liquid in the 
latter. 
With a non-parallel beam not only do the two above noted 
difficulties arise, but there comes in the additional error that the 
source of light is in effect brought some distance nearer in the 
case of the beam that passes through the liquid, and hence the 
light intensity of that beam is increased, that of the other beam 
being left unchanged.* 
Even were the beam of light employed to be the cone of rays 
proceeding from a very small hole in an opaque screen placed 
immediately in front of the light source and at the level of the 
* In this connection see a paper entitled “On the Absorption Spectra of 
some Copper Salts in Aqueous Solution.” by Thomas Ewan, B.Sc., Ph.D., 
Phil. Mag. (5), No. 203, p. 331, April 1892. 
