122 
MR. W. GEOFFREY DUFFIELD ON THE 
XIII. The camera was adjusted so that the centre of the grating lay in the normal 
to the plate through its centre. 
The method adopted for accomplishing this was to take photographs of the iron 
arc spectrum in the usual way, on three small plates placed at the two ends and at 
the centre of the dark slide, between the exposure of each set the camera was 
turned slightly by means of the screw M, fig. 5, about the pivot P, and the disposition 
of the camera giving equally good definition for the ends as for the middle was 
selected. Ames describes a convenient method for obtaining a rough approximation : 
a piece of plate glass having been fixed to the face of the camera box, a candle is held 
on the girder near the grating, and the camera turned until the flame and its image 
come into line. 
XIV. The adjustment necessary to correct for changes of temperature was made. 
Changes of temperature prevent the spectrum from remaining in focus continually, 
since the focal length of the grating mirror and the length of the cross-beam are 
liable to alteration on this account. 
The first care has been, therefore, to keep the temperature of the grating room as 
even as possible, and after this to move the slit towards or away from the grating, 
according as the temperature rose or fell, the best position at each temperature being 
found by trial photographs. A scale denoting the correct positions of the slit for 
different temperatures was in this way constructed and has been found to work well; 
the necessary motion of the slit is approximately 1 millim. for one degree Centigrade. 
Since the expansions of the grating and the cross-beam are unequal in their effects on 
the focus, at some temperatures the best definition is rather better than that 
attainable at others. 
The expansion of the grating also affects the dispersion in the spectrum to an 
unsuspected extent. A simple calculation shows that SX/A is equal to a, the coefficient 
of linear expansion of the grating. For X = 5000 A.U., and a = 0'00002 for 
speculum metal, SX = 0T A.U. for a change in temperature of one degree Centigrade. 
The necessity for regulating the temperature of the grating room at once becomes 
apparent. 
I have pleasure in acknowledging the services of Mr. T. Foyds, B.Sc., who assisted 
me in making these adjustments. My indebtedness to the papers of Ames and 
Kayser has already been expressed ; to Professor Kayser 1 am also grateful for his 
criticism of photographs obtained with this grating. 
