of a Diffraction grating. 231 
efficient of expansion, it is not necessary to know either the 
wave-length of the line upon which the measurements are 
made or the value of the grating space. 
Unfortunately there are considerable difficulties in the way 
of the direct application of this formula. It is not easy to 
contrive any plan by means of which the temperature of the 
grating can be varied by any considerable and known amount, 
and, at the same time, to insure that these changes in tempera- 
ture shall not be accompanied by even a slight shifting of the 
position of its plane in relation to the collimating telescope. 
To accomplish the first of these results, a box, measuring 
about 8 cm. long, 5 cm. wide, and 8 cm. deep, was cut out o 
a solid block of wood, and of this box the grating was made 
to serve as one of the longer sides. The box was made water- 
proof, and a thin layer of paint along the edges to which the 
metal, with the grating sealed in as one side, was first used, 
but it was promptly rejected on account of a tendency to shift 
changes in tem erature were much more likely to be accom- 
Panied by a shifting of the plane of the grating. 
