MR. CHARLES DE WATTEVILLE ON FLA.MK SPECTRA. 145 



E', one can see on the grating itself tin- luminous trace of the incident pencil of light. 

 The spark is then moved ahout until this trace occupies exactly the central part of 

 the surface of the grating. The burner is then lit and placed at B, between the 

 spark and the slit. The burner is then displaced laterally and rotated about a 



Fig. 4. 



vertical axis until the inverted image of the flame that one sees on looking at the 

 grating is as small as possible, this being the case when the series of small elementary 

 flames is in a straight line with the slit and at the same time the image is super- 

 posed on the luminous trace due to the spark. It only remains now to place two 

 lenses, one at L, between the burner and the slit, so as to illuminate the whole surface 

 of the grating, and the other at L', between the burner and the spark, in order to 

 concentrate the light of the spark at the centre of the burner. The distances are 

 so arranged that a small image of the spark is formed at the slit, so that the 

 whole surface of the grating is also illuminated by the light of the spark. The 

 height of the two lenses is adjusted until the spectrum of the spark falls exactly 

 in the centre of the field of view of the observing eye-piece. As to the height of 

 the burner, this is determined so as to obtain on the photographic plate the spectrum 

 of the two principal interesting regions of the flame, viz., the inner cone and the 

 flame properly so-called. A metallic wire is fixed on a level with the tips of the 

 small cones and the burner is raised or lowered until the dark line produced by this 

 wire across the spectrum of the flame is in the centre of the field of the eye-piece. 

 It is almost unnecessary to add that when the comparison spectrum was the arc 

 spectrum, the adjustment was made in exactly the same way by previously using the 

 two carbons as sparking electrodes. 



This procedure has always given a very satisfactory coincidence tatween the lines 

 of the two spectra, and has enabled me to avoid the accidental displacements which 

 were observed before these precautions were taken. 



On account of the extremely long time of exposure with the flame, the comparison 

 spectrum was obtained in the middle of this exposure, so as to establish a mean 

 position of the lines in the event of variations in the temperature of the apparatus 

 intervening to produce a widening of the lines of the flame. During the exposure for 

 the arc or the spark, the shutter of the dark slide was replaced by a screen provided 

 with a central aperture, 3 or 4 millims. wide, running along its whole length, an 

 exactly complementary screen to this being used during the exposure for the flame. 



VOL. CKJIV. A. \ 



