MAPPING THE LEAST REFRANGIBLE END OF THE SOLAR SPECTRUM. 659 
A collimator (Plate 30, fig. 1), 8 feet in length, was constructed after my design by 
Mr. Hilger, and instead of attaching a collimating lens to it, a collimating m irror 
(silver on glass) of about 8 feet focus was substituted. 
The slit S is inch long, and was mounted at the side of the tube as shown. 
The beam of light was directed full on to the slit by the heliostat, and passing through 
the slit falls on M, a plane and adjustable mirror (also silver on glass). It is then 
reflected down the tube, and falls on A, the concave mirror already referred to, and 
the rays pass up the tube in a parallel beam. 
At B they fall on a reflection grating of 17,600 lines to the inch, and about if inch 
in width'"; the dispersed beam falls on another concave mirror, B, of about 3 feet focus, 
and the image of the spectrum is formed on P, the photographic plate, which is 
attached to a camera in the usual manner. I may mention that a quartz, and also a 
rock-salt lens of 8-foot focal length is provided for this collimating tube, the mirror 
A and slit being movable. The reasons for adopting the long collimator and the mirror 
may be mentioned here, since at first sight there appears to be but little reason for using 
them. There is no doubt but that glass absorbs the ultra-red rays to an appreciable 
extent, and I need only refer to the experiments of Sir William IIerschel, Tyndall, 
Lamansky, and others in support of this statement. My primary object was there¬ 
fore to prevent any serious loss by only allowing the beam of light to traverse as small 
a thickness of this material as I possibly could. That the rays are not totally absorbed 
there is proof in the diagram of the prismatic spectrum, which I attach to this com¬ 
munication, the prisms employed being of the glass I have already referred to. By 
adopting the long collimator the beam of direct light reflected from the heliostat and 
falling on the grating was a circular patch of about f inch diameter, and this had 
traversed no glass except any that might be placed in front of the slit. There was 
but little loss of light, since to give the same definition the slit could be opened five 
times wider than with the shorter collimator, with which a condensing lens could 
be employed. There was another reason for adopting the system of reflection in the 
camera and also in the collimator, viz. : that as the foci for all rays are coincident, 
by focussing any visible portion of the spectrum of the second order on the plate, 
and then cutting this off by a suitable medium, the absorption lines in the dark rays 
beyond the red of the first order would be equally well defined. It will also be seen 
that by this plan it was possible to obtain a photograph of the second order on the 
same plate as one of the first order, by covering up the top half of the plate for the 
one, and the bottom half for the other. This is of prime importance in making a 
determination of the wave lengths of the invisible portions, and it has been utilised 
for settling the approximate values which are given in the map appended (Plate 31). 
I have no hesitation in recommending' that a mirror in the camera should be sub- 
stituted for the ordinary lens. It gives most excellent definition, and can be used to 
* Ruled by Mr. Chapman, of New York, with Mr. Rutherfurd’s ruling machine. 
