1^1 
for Mkroscopkal and other purposes. 
to ascertain the power of giving distinct vision, which belonged 
to each separate colour of the spectrum, and though I had 
stated in my former paper, “ that it was difficult to discover 
“ any reason why one coloured medium should be preferred 
to another, provided each of them transmit equal quantitites 
“ of homogeneous light yet it was desirable to put this the- 
oretical opinion to the test of direct experiment. Sir William 
Herschel ^ had long ago investigated this point, in reference to 
the use of coloured media for solar observations, and had con- 
cluded that every colour (f the spectrum possessed the same 
power of giving distinct vision ; but his method of observation, 
which consisted in viewing through a miscroscope a Nail illumi- 
nated in succession with each of the colours of the prism, was 
by no means calculated to give definite results, and therefore 
left the question in all its uncertainty. 
In order to obtain precise indications, which were not capable 
of being misinterpreted, when applied to practical purposes, I 
formed a spectrum from a luminous disc, by means of a prism 
of a highly dispersing substance, and with a large refracting 
angle. I then examined this spectrum through a great variety 
of coloured media, both solid and fluid, and marked the size 
and shape of the image into which it was converted. The per- 
fection of this image, or its narrowness in the direction of the 
length of the spectrum, became a precise and unequivocal test 
of the fitness for distinct vision which belonged to the light out 
of which it was formed. 
By this method of observation, I found that a distinct image 
of the luminous disc could not be obtained either by produ- 
cing a blue or a green image, and that it was only in the red 
portion of the spectrum that such an effect was likely to be ob- 
tained. In the use of purple glasses, I observed that the middle 
portion of the red space was absorbed before the two extreme 
portions, so that instead of one red image there were two quite 
separate, and tolerably distinct. By increasing, however, the 
thickness of the plate, the most refrangible red image was absor- 
bed, and the least refrangible one left in a state of the most per- 
fect distinctness. Although I had now determined the part of 
Philosophical Transactions, 1800. 
