604 PEOFESSOE STOKES ON THE LONG SPECTEUM OF ELECTEIC LIGHT. 
triangles vertically opposite, and having their common vertex at P, their lengths lying 
in the plane of refraction. The invisible spectrum is in fact made up of two such pairs 
of triangles corresponding to the two tips respectively, as may be readily seen when the 
electrodes are not too close. At a distance from P at which the length of the primary 
focal line becomes equal to that of the image of the spark, the two lines which are the 
images, for rays of the refrangibility answering to that distance, of the tips of the elec- 
trodes meet in the middle of the spectrum, and beyond that distance they overlap, so that 
a line appears to run across the spectrum, though it relates to rays which emanated only 
from the immediate neighbourhood of the tips of the electrodes, as may be seen by turning 
the prism till that part of the spectrum is at a minimum deviation, and focusing afresh. 
Besides the bright lines, evidently due to metals, which have been mentioned, other 
weaker light is perceptible, too faint for precise observation. A portion of this is pro- 
bably due to the air. 
The chief part of the visible spectrum as seen by projection appears plainly to belong 
to the air ; for the lines stretch across the interval separating the electrodes, while the 
lines belonging to the metals extend but a little way, even in the visible spectrum, and 
the former reappear when the electrodes are changed. With some metals, however, 
lin es belonging to the metal appear in the visible spectrum which are comparable in 
strength with the invisible lines of high refrangibility ; but in general it is rather 
remarkable how poor is the visible spectrum, and even the invisible region for a good 
distance beyond, compared with the part of the spectrum of still higher refrangibility, 
with respect to strong lines characteristic of the metal. 
I have lately adopted a mode of laying down positions in the invisible spectrum which 
is extremely simple and convenient, and yields results agreeing well with one another. 
It might be applied to the formation of maps of the metallic lines ; but this is unneces- 
sary, as the subject has been worked out by Dr. Miller. It is still useful, however, for 
laying down the positions of bands of absorption, being more convenient and exact than 
estimating their place with reference to the known metallic lines. 
The method is as follows. The quartz prism is placed on a block, raising it to a con- 
venient height above a long drawing-board, to which the block is screwed, and is fixed 
at pleasure by a screw pressing upon it from above. The lens is fixed in a blackened 
board screwed edgeways to the dravring-board near the prism, so as to be ready to receive 
the rays of all refrangibilities after refraction through the prism. The focal length 
of the lens actually used was about 12 inches, and its diameter 1|- inch. A convenient 
distance of the spark from the prism having been selected (I chose 30 inches), the 
drawing-board was turned round till it attained such a position that, on placing the 
prism in the position of minimum deviation for the middle of the long spectrum, the 
rays belonging to that part fell perpendicularly, or nearly so, on the lens, which had 
previously been placed so that this should be a convenient position relatively to the 
drawing-board. The prism was then fixed by its screw, and to mark the angle of inci- 
dence a pin was placed at the edge of the shadow of one of the blocks. On account of 
