454 MR. A. FOWLER ON THE SPECTRUM OF MAGNESIUM HYDRIDE. 



intensity in the region about 4780, and can be followed to the violet head at 4371. 

 The spectrum, in fact, shows an almost unbroken succession of lines at very short 

 intervals throughout the entire visible spectrum, but the lines are mostly faint, except 

 in the definite flutings to which reference has been made. 



The ultra-violet spectrum, of which no measurements have yet been undertaken, 

 also exhibits a great number of lines distributed throughout its whole extent. There 

 is a well-marked fluting, very similar to the one in the green, with its head about 

 2430, and fading off on the more refrangible side. There is also an interesting series 

 of double lines in the region about 2940 to 3100, in which the lines close up towards 

 the side of shorter wave-length, whereas the somewhat similar series which can be 

 traced in the green and yellow-green flutings converge towards the red. 



Another feature of the spectrum, which cannot be very clearly seen in the 

 reproductions, is possibly of some importance. On photographs which have received 

 sufficient exposure, there are patches of unresolved continuous background, sometimes 

 terminated by strong lines, but often by lines which are very faint and indistinct. 

 Intervening places in the spectrum, varying in width from about 0'3 tenth-metre 

 upwards, remain conspicuously clear in the negatives, and have the appearance of 

 bright lines or narrow bright bands. 



Similar patches of " dark ground" have been noted by KAYSER and RUNGE in their 

 account of the 3883 fluting of cyanogen,* and the same appearance may be seen in 

 the flutings of titanium oxide and other substances. This appearance, however, does 

 not seem to be special to fluted spectra, as regions of dark ground and exceptionally 

 clear spaces are also found in negatives of the crowded ultra-violet arc spectrum of 

 iron. In the latter case, the darker places are probably due to groups of closely 

 adjacent faint lines, or to an underlying banded spectrum, and the darker places in 

 fluted spectra may similarly be due to groups of faint lines. Attempts to indicate 

 the positions of the dark and bright places in the magnesium hydride spectrum were 

 abandoned, on account of the difficulty of dealing with differences depending upon the 

 dispersion employed and upon differences of exposure. 



Determination of Wave-lengths. 



The catalogue of wave-lengths which appears at the end of the present paper has 

 been based entirely upon photographs with iron arc comparisons, taken with the 

 concave grating of 10 feet radius. The new standard wave-lengths determined by 

 FABRY and BuissoNf were adopted for the iron lines, and as many as possible of these 

 lines were included in each section of the spectrum measured, so as to furnish a check 

 on the general accuracy of the work. In order to eliminate local systematic errors as 

 far as possible, each of the three flutings to which the catalogue refers was at least 

 once completely measured with the same setting of the plate. 



* ' Abh. K. Preuss. Akad. Wiss.,' Berlin, 1889. 

 t ' Astrophys. Jour.,' vol. 28, p. 195 (1908). 



