EFFECT OF PRESSURE UPON ARC SPECTRA. 
121 
XI. The cross-beam was placed in position and (a) the grating was rotated until 
the pivot about which the centre of the plate turned was in the normal of the grating 
through its centre; (/3) the length of the cross-beam was made equal to the radius 
of curvature of the grating mirror. 
(a) was accomplished by the method described by Kayser of marking on a ground 
glass screen (a photographic plate which has been fixed for 10 minutes, without 
previous exposure, when it becomes almost transparent, 
is, after washing, a serviceable substitute) two small 
squares equidistant from a central line, and placing it 
in the camera with the central line directly above the 
pivot about which the camera box turns. A candle 
was held behind A, and its image observed with the 
magnifying glass in the square B while the grating was 
turned about its vertical axis by an assistant who operated the tine adjustment G. 
When the centres of the two crosses coincided, the central line lay in the normal to 
the grating. 
(/3) was effected by observing the image of A through the square B, as before, and 
racking the ^camera box backwards and forwards by means of the screw K, fig. 5, 
until all parallax disappeared and no motion of the eye caused the image of A to 
move within the square B. This determination was a matter of some difficulty and 
delicacy, and several independent observations were made by different observers. 
The adjustments of the mounting necessary for (/3) require the displacement of 
the centre of the plate from its true position above the pivot, but the distance 
through which the plate holder was moved was noted, and the cross-beam was 
lengthened or shortened by the amount requisite to restore the plate to its correct 
position. The telescopic adjustment allows this to be easily performed. 
XII. The slit was adjusted parallel to the lines of the grating. 
The slit was rotated in its own plane about its centre by turning the screws X 
fig. 7, and with the usual method of illuminating the slit many photographs were 
taken on a small plate placed immediately over the pivot P, fig. 5. The spectrum of 
the iron arc was employed, and that disposition of the slit giving the best definition 
was chosen. 
A simple test, due to Kayser, of the accuracy of this adjustment is to cover the 
middle ot the slit to make it equivalent to two short slits separated by an opaque 
interval; a photograph taken with this disposition will show two spectra whose lines 
are drawn out by the astigmatic property of the grating to nearly meet in the 
centre. When the slit is parallel to the rulings, the two portions of the slit give a 
continuous line, otherwise the two lines are slightly displaced. 
The fineness with which the ends of spectral lines (e.g., those of the iron arc) are 
pointed, and the precise similarity of their two ends, is also a good gauge of the 
accuracy of this adjustment. 
VOL. CCVIII.—A. 
R 
