114 THE ROYAL SOCIETY OF CANADA 
fifths of their length was occulted. The definition was improved by 
masking one end of the ruling to the extent of two inches, so that the 
ruled surface in use was reduced to about 2? X 34 inches. Even with 
this reduction in surface the spectra were so bright as only to require 
comparatively short exposures in the 2nd and 3rd orders. The principal 
difficulty remaining seemed to be a certain amount of diffused light 
which blocked up the lines, diminishing the contrast, and increasing the 
difficulty and probable error of measurement. 
Michelson very generously offered to furnish a new grating on 
account of these defects, and in the autumn of 1910, we received grating 
No. 43 with a ruled surface 64 2% inches having about 580 lines to the 
millimetre. This was the grating used by Gale and Lemon in their 
investigation on the mercury lines and the one used by Michelson in 
his laboratory, and is undoubtedly a very fine instrument. Buta test 
showed that, for the purpose required in this work, its smaller dispersion 
and lower brightness rendered it not so suitable as the former, which 
was consequently retained. 
Within the last month or so a plane grating, ruled by Anderson at 
Johns Hopkins, has been received and tested. This grating has a ruled 
surface 34 X 5 inches with 15,000 lines to the inch. This grating gives 
definition of about the same quality as the Michelson grating with 
freedom from diffused light, so that there is considerable more contrast 
between the lines and continuous spectrum and consequently greater 
ease and accuracy of measurement. Unfortunately, there is also 
astigmatism present, which is only satisfactorily removed when about 
three-fifths of the length of the lines is occulted. However, the whole 
width of the ruling can be utilized and it gives good resolving power, 
and is fairly bright in the third order. 
A large number of experimental and test plates have been made 
with these four gratings to determine the conditions under which the 
best results can be obtained. Considerable experimentation with dif- 
ferent makes and brands of plates has also been undertaken to determine 
the most suitable emulsion. The Seed Contrast Process seems the 
best for the work at 4 4250, while the Cramer Iso. Process has been 
selected for the region À 5600. A trial was made of the Seed Process 
stained with Erythrosin for work in the latter region, but although 
they are of somewhat finer grain than the Cramer, the trouble of staining, 
the poor keeping qualities, and the lack of uniformity in successive 
batches led to the choice of the commercially prepared orthochromatic 
plates. 
In the determination of the solar rotation by the measurements 
of line displacements at opposite limbs, the three most important and 
interesting points to be settled are:— 
