944 
EEV. T. E. EOBINSON ON SPECTEA OE ELECTEIC LIOST. 
or the moon (and very sharp), is over-corrected for G and the rays beyond it : as no pro- 
vision was made for changing the distance of its object-glass from the system of wires, 
1 at first had some difficulty from parallax at that end of the spectrum, till such an adjust- 
ment was applied. The eyepiece (positive, magnifying nine times) was also not achro- 
matic, and had to be constantly focused. It was replaced by a microscope’s objective 
about 0-4 inch focus, magnifying 14 with the prism. A higher power, and even a 
larger object-glass, avails little in comparison of an increase of the prism’s dispersion. 
Thirdly, I found the cross of spider’s lines useless, except for the brightest lines. In 
the spectra of rarefied gases, which are very faint, it is difficult to see them, and I sub- 
stituted the point of a fine needle, carefully ground to be a sharp wedge. One can 
estimate very nicely the equality of the tongues of light on each side. These changes 
were not made till about fifty spectra had been measured, which, therefore, are not as 
well determined as the rest. 
2. The collimator certainly possesses advantages over the simple slit which Fkacx- 
HOFER and his predecessors used in studying the spectrum. It requires no correction 
for parallax, secures from any accidental shift of the theodolite, and brings the observer 
close to his work; but it has the great defect of diminishing the light. Unless the sht 
be very narrow, it is impossible to distinguish close and fine lines. I find that with the 
instrument which I am describing, and the Merz prism, I cannot see D double if the 
slit subtend more than 72", which corresponds to a width of inch. If it be 3' (the 
opening used by Pluckee) I cannot see any of Fraunhofer’s lines, and the finer parts 
of electric spectra are lost. Obviously the quantity of light must diminish with the sht, 
and the evil is made greater by the necessity of keeping the latter at some distance from 
the discharge. If it be nearer than 1*5 inch, or at most 1 inch, the inductive action of 
the spark charges the theodolite, and the observer, on applying his eye to the telescope, 
gets a stream of pungent sparks anything but pleasant. But in order that the whole 
object-glass may be illuminated, we must have this distance less than •’ A 
A. cosecant/ y “‘T 
and /‘being the aperture and focus. This limit in the present instance =0‘03 inch, fiom 
which it is obvious that a comparatively small portion of the light can reach the object- 
glass. I was led to this discussion by a fact which at first startled me a little. "SATien 
obtaining the deviations of Fraunhofer’s lines, I was surprised to find that H was not 
visible to me, nor any line beyond h; though in 1838, with the same prism, but with a 
slit in the shutter 15 feet distant, I saw several beyond K. I concluded that my eyes 
had become insensible to rays of shortwave-length (in analogy to Wollaston’s inaudible 
sounds), or that their humours had undergone some change by which they absorbed 
that part of the spectrum ; but never suspected the collimator. However, last year, while 
examining a very fine Munich grating belonging to Mr. Stokes, I was surprised to find 
that it showed me the missing lines perfectly : here the aperture was a slit at 18 feet. 
He was so kind as to entrust me with the grating ; and on my return home I found 
that with the collimator it behaved no better than the prisms had done, but that with 
