[PLASKETT-DELURY] THE SOLAR ROTATION 25 
difficulties in stellar spectroscopy, are not however of much’ moment 
here for, owing to the short and simultaneous exposures on opposite 
limbs, temperature changes will have no appreciable effect, and there 
can be no flexure when the spectrograph is stationary during the 
exposure. It may not be amiss to repeat here the four essential pre- 
cautions for accurate observations given in the previous paper. 
(a) The emulsion on the photographic plate must be exactly in the 
focus of the spectrum. 
(6) The illumination of the grating from the opposite limbs of the 
sun must be similar and uniform. 
(c) The solar definition must be good, the image steady, and the 
sky free from haze. 
(d) Care must be taken that the reflecting prisms receive light from 
the desired latitudes. 
6. Precautions a and 6, conditions inside the spectrograph, to 
which may be added the avoidance of undue heating of the slit jaws, 
are very necessary to prevent systematic displacements of the lines 
as a whole introducing corresponding errors in the velocity values. If 
either a or b are exactly fulfilled an approximate realization of the 
other should be sufficient; but, as it is practically impossible either 
to get or keep the plate at the exact focus or to have absolutely equal 
and uniform illumination of the lens and grating from the opposite 
limbs, the only safe procedure is to fulfil both conditions as closely as 
possible. Consequently the plate focus was determined frequently 
both by the definition test and, as a check, by the Hartmann method 
of extra-focal exposures. It was found that the field both in the A5600 
and in the 4250 region was curved, concave to the lens, about 2.5 
mm. longer at the centre than at the ends of a plate 30 em. long and 
inclined about 1°, in opposite directions for the two regions, to the 
normal to the axis. The illumination of lens and grating was tested 
before and after each plate, which consisted usually of seven spectra 
one of each of the six latitudes from 0° to 75°, and one of the pole. 
This was done by opening the slit wide enough to allow a visible image 
of the illuminated concave mirror to be projected on the diaphragmed 
front surface of the collimating objective. If this image was not 
central for both systems of prisms it was easily made so by the adjust- 
ing screws provided. It was found frequently that a slight change in 
position of the overlapping images occurred during the time the seven 
exposures were made, but never sufficient (since the image is consider- 
ably larger than the used portion of the grating) to prevent uniform 
illumination. This change of adjustment of the prisms must be due 
to the heating produced by the sun’s rays and to minimise this effect, 
the heating of the slit jaws, and the distortion of the coelostat, secondary 
Sec. III., 1912. 3. 
