[ 155 ] 
YII. The Sensitiveness of the Retina to Light and Colour. 
Bjj Captain Y/, de W. Abney, 
C.B., {late) R.E., D.C.L., F.R.S. 
Eeceived Maj 10,—-Read June 3, 1897. 
1. New Form of Apparatus for Reducing the Intensity of Light. 
In “ Colour Photometry,” Part III. (‘ Phil. Trans.,’ A, 1892), a description was given of 
the apparatus employed for estimating the intensity of light of any colours which just 
failed to cause any sensation on the retina. In the following research a modified form 
of apparatus has been employed, and experience has shown it to be equally as accurate 
as and more convenient in many ways than that formerly used. When rotating sectors 
are employed with less than 4° of aperture, small errors in the reading of the graduated 
arc cause appreciable errors in the result. Hence, the more nearly the zero reading is 
approached, the less trustworthy are the determinations of the diminution of the total 
intensity, and this uncertainty affects all observations in which the light is reduced 
below 4 ^ of its initial amount. The sectors have also the disadvantage of not being 
noiseless, and of requiring an electro or other motor to work them. With the 
extinction box, described in Part III., the difficulty of being able to reduce the 
light to, say xooFo of its initial intensity, was overcome by using diaphragms 
of varying aperture in front of the ground glass, which was practically the source 
of illumination. It is obvious that any instrument which would allow a similar 
decrease in illumination without the intervention of a diaphragm would be advan¬ 
tageous, more particularly if it were noiseless. In the estimation of star magnitudes 
by extinction, a wedge has been used from time to time by various investigators, 
and in 1870 I myself used one for the purpose of measuring the intensities of 
the electric light. As a rule, these w^edges are of dark green glass wdiich have a 
fairly high exponential coefiicient of absorption. The fact that it has a dominant 
colour at once indicates that for the comparison of spectrum colours by extinction 
such a wedge should be avoided. Thanks to the kindness of Dr. Grossmann, of 
Liverpool, I had put into my possession a pair of black glass wedges, and with 
these I hoped to avoid the difficulty caused by tlie colour of the green glass 
wedges, but after mounting them and preparing to use them, it was found that 
the material exhibited bands of absorption in several parts of the S23ectrum; after 
graduating them for each spectrum colour, the results were so unsatisfactory that 
X 2 9.12.97 
