KENT. — THE SPECTRA OF TITANIUM, ZINC, AND IRON. 283 



separates the dark spectrometer room from the light room iu which are 

 placed the spark and arc. The entire mount, including channel beams, 

 steel tracks, trucks, truss-rod, draw-bars, and all possible parts of grating 

 and camera box, is carefully wrapped in several layers of paper or cloth 

 to facilitate obtaining constant temperature conditions. The steam pipes 

 in the room are protected thoroughly by asbestos covering or several 

 layers of paper. 



In adjusting the instrument the slit and grating rulings were made 

 parallel by the use of a plumb bob. The coincidence of the normals to 

 the grating and to the plane of the camera box was determined by 

 using an illuminated slit 1 mm. wide and adjusting until the imnge was 

 horizontally coincident with the object. The ordinary method of super- 

 posing the image of a candle flame upon the flame itself was not regarded 

 as sufficiently delicate. 



Duration and Method of Exposure. 



In general the exposures were short, ranging from three to ten 

 minutes, and all the quantitative work of the investigation was carried 



Figure A. 



X ■ ■ ■•. . . ■ . ■ ■■. • 



-NX-, ' " 



s\ 



: _^ , 



^ 



Figure C. 



Figure B. 



on when temperature conditions were most favorable, the change in 

 temperature as given by one or more thermometers whose bulbs were 

 placed underneath the cover of the mount usually being absolutely un- 

 readable, and never more than a half of a tenth of a degree Centigrade. 

 Moreover, the method of expose was such that pseudo-shifts, clue either 

 to mechanical jar or temperature change, must have been eliminated. 

 The camera shutter was arranged as shown in the following cuts. 

 Usually the arc was exposed first, its spectrum being placed along the 

 centre of the photographic plate — this by use of the slit as given in 

 Figure A ; then the spark image was thrown upon the slit and its 



