226 
quantities of bichromate of potash, increasing from 1 to 30, 
the first members of the series would be easily arranged 
towards the middle of the series, the number of successful 
discriminations would diminish and few would be able to 
detect differences in the higher members of the series ; for 
an increment of the salt which produces a considerable dif- 
ference in intensity when a small quantity of the salt is 
present produces a small difference when the quantity is 
large. This would follow from the approximate formula 
cl i c 
used, for - - „- 2 (t = depth, Q — quantity of colouring 
matter, c — constant. So that for a given small difference 
in the quantity the difference in the depth would be nearly 
proportional to the inverse square of the quantity. Solu- 
tions wfftose quantities of colouring matter are so similar as 
to be beyond discrimination by colorimetry may be again 
brought within its range by the following device : in each 
of the two cylinders destroy by some appropriate method 
an equal quantity of the colouring matter ; then their differ- 
ences will be increased relatively to the quantities of colour- 
ing matter remaining, and may be sufficiently sensible to be 
estimated. When I suggested this method of colorimetry 
I had not seen the article on Light by Sir J ohn Herschel in 
the Encyclopaedia Metropolitans He there gives a prob- 
able formula for the intensity of light which has passed 
through an absorbing medium. His reasoning is as follows. 
Let the intensity of the incident light be unity, and its 
intensity on emerging from a unit thickness of the medium, 
k ; then its intensity at incidence on a second layer will be 
k x and its intensity at emergence k 2 ; hence if t be the thick- 
ness of the medium the intensity will be finally k L , and if 
the original intensity were a the final intensity would be 
ak l . Now, light consists of several species; hence if the 
composition of the incident light be cq a n + a in + &c. the 
composition of the emergent light will be a v k\ -f 
4 
ii“'n 
