Aprnry— Photographs of Spark Spectra from the Rowland Spectrometer. 339 
of air may be, the lines and bands of oxygen and nitrogen do not obscure or 
suppress those of the elements which constitute the electrodes. The stream of 
vapour of greater density than air fills up the track of the spark and excludes the 
gases of the atmosphere. 
In such cases the electrodes give an imperfect air spectrum, and such imper- 
fections are characteristic of certain metals. In measuring air lines some of the 
metals were found to be more suitable than others. Metallic lines also suppress 
neighbouring air lines, and owing to their greater emissive power are easily 
distinguished. 
Experimental Details, 
The present series of spectra form a continuation of the work above quoted. 
They have been photographed from the Rowland’s concave grating, which has 
recently been mounted in the Royal University, Dublin. The grating has a focal 
length of 21:5 feet, and a ruled space of six inches, with 14,438 lines to the 
inch. 
The mounting of the spectrometer in the Royal University has already been 
described by the author, in conjunction with Mr. J. Carson, 4.R.c.sc.1., in the 
Scientific Proceedings of the Royal Dublin Society, 1898.* 
A current of ten ampéres, and a ten-inch coil by Apps, were employed, with 
a condenser 144 square inches in area, for ‘‘ sparking.” 
The specimens of the metals used as electrodes were, it is believed, of a high 
degree of purity. 
The specimen of iron was prepared from pure crystallized potassium ferro- 
cyanide in the same manner as that employed in the investigation by Hartley 
of the oxyhydrogen flame spectrat; also of the arc spectra by Hartley and 
Ramage, for comparison with the solar spectrum when measuring the lines of 
gallium.t 
The specimens of cobalt and nickel were portions of those prepared by 
Dr. W. J. Russell, rr.s., in his researches upon the atomic weights of those 
metals. 
Specimens of ruthenium, rhodium, palladium, osmium, iridium, and platinum 
were very kindly presented by Messrs. Matthey and Johnson for the purpose of 
this research. 
The chromium spectrum was obtained by sparking a saturated solution of 
potassium chromate between platinum electrodes. 
* See also Phil. Mag., vol. xlvi., p. 223, 1898. 
} Phil. Trans., vol. clxxxy., p. 161, 1894. 
{ Sci. Trans. Roy. Dublin Soc., vol. vii., ser. 2, 1898. 
