PHYSICS: C. BARUS 
433 
focus and the spectroscope is removed. They appear and widen with 
the washed image of the slit. They are quite strong and sharp through- 
out and gorgeously colored, the fields and shades figure 2 being nearly 
complementary in color. The spectrum fringes must be centered if the 
others are to occur. The former (fig. 3) are usually long ellipses or hy- 
perboHc with the major axis horizontal, while the corresponding new 
fringes (fig. 2) are hyperbolic with their major axis vertical. They are 
extremely sensitive to rotation of the micrometer mirror about a hori- 
zontal axis, rising or falling, but they soon vanish. When the microm- 
eter mirror M' is rotated around a vertical axis, an operation which 
separates the white sHt images in the telescope if originally coincident, 
the new fringes move bodily by displacement, from left to right, or the 
reverse, depending on the sign of the rotation, while they continually 
change their color scheme. When the design is thus displaced as a 
whole the individual fringes move as shown in figure 2. As a group the 
fringes closely resemble the lemniscates of a binaxial crystal in polar- 
ized light. The variation of the color scheme is probably the same 
since with homogeneous light (sodium) the design is in yellow and black. 
The pattern is not quite dichroic but appears so, red-green, blue-yellow 
combinations with intermediate violet-yellowish succeeding each other. 
In polarized Hght the figures are sharpened as a whole, but there is no 
discrimination. The pattern gradually vanishes with a wide slit, where- 
upon the achromatic fringes may be seen when the ocular is restored to 
the principal focal plane. 
