[PLASKETT-DELURY] DETERMINATION OF THE SOLAR ROTATION 113 
by means of the screws shown in Fig. 2 and the rounded projection back 
of the prisms on which they rock. 
The notched prism B’ is much more easily adjusted than the two 
prisms formerly employed for the same purpose. The idea of placing 
the two strips of spectra from one limb on each side of the strip from 
the other limb by means of the notched prism, we believe to be a new 
and important point, asit eliminates certain small errors of measurement 
which may occur in the method of having just two strips side by side, 
which has been employed by other investigators in this field. 
Over these prisms is placed a brass plate serving to protect them 
from dust and injury and also providing means of guiding the sun’s 
image centrally. This is done by aid of a circle of diameter 229 mm., 
slightly greater than the greatest diameter of the sun’s disc. In this 
circle are two little adjustable windows through which light from any 
desired parts, on or within the limb, may be admitted to the reflecting 
prisms. Under these windows are little cups to receive suitable filters 
when it is necessary to use such. The diameter of the circle which 
passes centrally through these windows is ruled on the plate, and parallel 
to this a tangent is ruled to the circle for determining the “East and 
West” line as mentioned above. To take a rotation plate, the sun’s 
image—so focussed that after passing through the prisms it comes to 
a focus at the slit—is guided centrally in the circle, and the ruled diameter 
is set at any desired inclination to the sun’s equator. 
As already stated, it was decided in 1906, to make the determina- 
tion of the Solar rotation one of the main researches for the coelostat 
telescope then designed. The shelter for this telescope was not com- 
pleted until the summer of 1908, and the completion of the solar spectro- 
graph was delayed until a somewhat later date. The plane grating used 
in this instrument was one of Michelson’s early attempts, having a ruled 
surface about 5 by 44 inches with 500 lines to the millimetre. Tests 
of this soon showed that the definition given was poor and that the 
grating had some curious focal properties! Preliminary measures of 
some rotation plates made with this grating showed, owing to the diffuse 
nature of the lines, high probable errors and it was felt that satisfactory 
results were impossible.” It was not until the spring of 1910 that a 
new grating, No. 55, ruled specially for this work, by Prof. Michelson 
was obtained. This had a ruled surface of 5 X 54 inches with 700 lines 
to the millimetre and gave very bright spectra with good definition. 
The astigmatism present was shown to be due to the lines of the ruling 
not being quite straight, and was entirely removed when about two- 
! De Lury, Jour. Roy. Astron. Soc., Can. V. 26-32. 
? De Lury, Report of the Chief Astronomer, 1909, 251-256. 
