146 
POPULAR SCIENCE REVIEW. 
either single as in fig. 3, or double as in fig. 5, or to a com- 
pound prism of five pieces, as shown in fig. 6. 
But the use of small instruments such as those above de- 
scribed, must be looked upon merely as preliminary to the 
thorough investigation of the spectra of celestial objects by means 
of spectroscopes attached to telescopes of considerable power. It 
is only by the use of such instruments that the observer can 
hope to make important additions to our knowledge of celestial 
physics. 
In fig. 7 is pictured the kind of spectroscope made use of by 
Mr. Huggins in his long series of researches. 
It will be understood that the large telescope to which the 
spectroscope is attached serves the purpose of light-gatherer. 
The spectroscopes we have hitherto described have had to be 
limited in their dispersive action, in order that the light received 
from a star might not be altogether lost to the eye through 
dispersal. But now the image of a star formed by a large ob- 
ject glass (or mirror) is to be made the subject of observation, 
and therefore the spectroscope may exercise a much greater 
dispersive power. In fig. 7, c represents the tube of the spec- 
troscope which is inserted into the eye-tube of the large tele- 
scope in place of the ordinary eye-piece. (The tube c and the 
inner sliding tube b, are longer than shown.) The slit of the 
spectroscope is at cZ, and the spectroscope is so placed that d is 
exactly at the focus of the object-glass (or mirror). The sliding- 
tube b carries a cylindrical plano-convex lens, the effect of 
which is to change the image of a star from a point into a line. 
It is clear that the linear image thus formed at d will be at 
right angles to the axis of the cylinder. For, in a plane passing 
through the axis of the object-glass, and also through that of 
the cylinder, the rays will converge to the focus of the object- 
glass, but in a plane at right angles to this (through the axis 
of the object-glass) the rays will converge to a focus nearer the 
object-glass, owing to the convexity of the cylindrical lens, and 
these rays will pass at a sensible distance from the focus of the 
object-glass. This linear image must fall upon the slit, which, 
therefore, must be placed at right angles to the axis of the 
cylinder.* In ^Ir. Huggins’ instrument the length of the image 
was one-tenth of an inch, and the spectrum formed by the light 
from this image had a corresponding width; but, as seen by the 
small telescope, it appeared about half an inch wide, owing to 
• A little consideration will show that there is another linear image 
parallel to the axis of the cylinder, at the second focus named above — that 
is, nearer tlie object-glass. Mr. Huggins tried the effect of causing this 
image to fall on the slit, but the spectrum w^as not so good as by the other 
method. 
