54 
PHYSICS: C. BARUS 
If a direct vision prism or grating is placed in front of the telescope, the 
spectrum is seen to be crossed by intense black lines, very nearly parallel 
and horizontal, but actually diverging from blue to red symmetrically up and 
down from the horizontal central black line. It is not necessary here, that 
the slit be fine. In fact it may be several millimeters broad without destroy- 
ing these spectrum fringes, if essentially horizontal. 
If the linear phenomenon of reversed spectra, coinciding on a certain line 
of the spectrum is wanted, a prism may be inserted between m and m' suit- 
ably adjusted. These fringes then also appear at once and may be put in 
any color, at pleasure, on rotating any mirror, say iV", on a vertical axis. Ro- 
tation on a horizontal^ axis enlarges. 
Finally if separate plate glass compensators are placed in the paired beams 
85 and 21, for instance, and rotated around a horizontal axis independently, 
the fringes may be displaced up or down the slit image for the purpose of 
measurement. A double offset air compensator consisting of 3 right-angled 
V-mirrors with their corresponding faces parallel, the central V-mirror mova- 
ble on a micrometer (described in my paper on gravitation) is particularly 
available. Such a compensator would be placed normally to the rays 85 and 
21, for instance, to give them lateral path length. In these cases the spectro- 
telescope may also be used, where the strong bands register the displacement 
in any wave length. Since the sht may be broad, there is here also a great 
abundance of light. 
The rays 85, 21, may be made of almost any reasonable length, and dis- 
tance apart, if the mirrors N, N\ m, m', are broad. To secure greater length, 
the mirrors mm' (rigidly connected) may be moved at pleb,sure in the direction 
58, without disturbing the fringes, good slides presupposed. The rays 85 
and 21, 34 and 76 may be separated, if x is the available breadth of mirror 
to an extent x cos 45°, by moving the rigid system NN' in the direction 85. 
If either of the mirrors m or m' is separated (as for instance at a in figure 
6), the part may be placed on a micrometer; but the apparatus would not then 
be selfadjusting; for the parts will not in general be coplanar. 
