1903-4.] Mr Milne on a.Neio Form of Spectrophotometer. 353 
marked disparity in the absorption coefficients for the ordinary 
and the extraordinary rays exists ; and, further, any difference in 
the intensity of the transmitted beams after passing through the 
Wollaston prism could, in any case, only be a small one, because 
the rays of each beam are transmitted for the length of approxi- 
mately half their path through the crystal as ordinary (extra- 
ordinary) rays, and for the remaining half as extraordinary 
(ordinary) rays. 
Another possible source of asymmetric error lies in the fact 
that the rays from any, the same point of the image p (fig. 4) may, 
after passing through the Wollaston prism, diverge to a different 
extent from the rays from the corresponding point in the other 
image q, after they have passed through the Wollaston prism. 
Were this the case, and were the difference sufficiently marked* 
the eye would see the strip due to the less divergent beam to 
sensibly greater advantage as regards intensity than the strip due 
to the other beam. And indeed the two images themselves, 
because they are formed inside the Wollaston prism, may not 
correspond in brightness to the original beams, for the rays of the 
two beams respectively may be converged to a different extent on 
entering the prism. 
Any such errors, however, did they exist could be at least very 
approximately got rid of as follows. The light absorption of any 
liquid for any particular wave length would be twice measured* 
once with the Wollaston prism emitting the upper beam as the 
ordinary ray, and then with the Wollaston prism turned upside 
down and emitting the same beam as the extraordinary ray. The 
mean of these two measurements would give the true absorption 
very nearly. 
The model instrument which has been made, while it shows 
the general soundness of the principles involved, is not capable of 
measurements of the accuracy required to definitely settle this 
question. All that can be said in the circumstances is that no 
such discrepancy can be seen with the present apparatus. 
In the note of last year the use of the instrument for Murphy’s 
method of mapping the visual intensity of a spectrum was pointed 
out, and it only needs to be said that the necessary adjustments 
of the apparatus are those described in that communication in the 
PROC. ROY. SOC. EDIN. — VOL. XXV. 23 
