Avrnry— Photographs of Spark Spectra from the Rowland Spectrometer. 309 
spectrum could be focussed on each plate, and hence any general resemblance 
between the spectra of similar elements was not rendered obvious.* 
The ultra-violet spectra obtained with the quartz spectrograph, while showing 
the differences in the lines, gave a dispersion in the ultra-violet region which was 
sufficient for all practical purposes. So important were the characteristics of the 
lines considered to be that a large amount of labour was expended in giving a 
minute description to each one of them, as well as their positions and wave-length 
measurements. 
The only other observers who had given particular descriptions of the lines 
previously, and those only in the visible spectrum, were 'Thalén and Lecocq de 
Boisbaudran. 
Photographs of spectra render any detailed descriptions of the lines now 
unnecessary, but attention may be directed to the fact that methods of producing 
spark spectra latterly employed fail to render the characteristic features of several 
well-marked groups of elements. 
Other investigators have since published photographs of spark spectra, namely, 
Crew and Tatnall,t and F. McClean, m.a.£ (these range from D to H only), and 
Eder and Valenta.§ In the first-mentioned memoir a table of corrections is 
given for converting Hartley and Adeney’s measurements from Angstrém’s to 
Rowland’s scale. Eder and Valenta’s beautiful reproductions of spectra show 
that many of the features previously preserved are wanting in theirs. The same 
remark may be applied also to Exner and Haschek’s photographs of the spark 
spectra of the elements. In both series the mode of producing the sparks 
differed trom that which had been previously commonly resorted to, and 
concave gratings were employed. 
It would be a remarkable fact if the measurements published in 1884 did 
not require revision for purposes of physical research, since at that period the 
properties of Rowland’s concave gratings, which had so largely increased the 
accuracy of spectroscopic work, had not been discovered. 
Fifteen of the spectra were entirely new to science, and the wave-lengths 
assigned to the lines have for all practical purposes been proved sufficiently accurate 
and of great utility in chemical investigations of a diverse character. 
In view of the attention now being given, not only to the numerical relation- 
* Phil. Trans., vol. clxxxy., p. 63, 1884. 
} Phil. Mag. (5), vol. xxxvili., p. 379, 1894. 
{ “Comparative Photographic Spectra of the Sun and the Metals.” Monthly Notices of the Roy. 
Astron. Soc., vol, li., No. 1. 
§ ‘‘ Beitrage zur Spectralanalyse,’’ K. Akademie der Wissenschaften (Wien), 1892 to 1899, 
3 A 2 
