242 Scientific Proceedings, Royal Dublin Society. 
the body of the flame; the bands are therefore of uniform bril- 
liancy and width. But, above all, the bands may be photographed, 
so that with the same photographic plates and the same exposure 
a similar spectrum is obtained, which can be measured by applying 
an ivory scale divided into hundredths of an inch, or fourths of a 
millimetre ; and measurements may be repeated and corrected. 
Hye-observations record the average effect of brilliancy and 
intensity of lines and bands ; while photographs are a record of the 
aggregate effect over a given period of time. All difficulties 
arising out of inequality in sensitiveness of the prepared film to 
different colours are now overcome by the use of Wratten and Wain- 
wright’s panchromatic plates (see Plate XX., fig. 2, Potassium). 
The examples of flame-spectra of the calcium, strontium, and barium 
group show that, with a constant exposure, the width of the bands 
increases with the quantity of substance in the flame; with a 
constant quantity of substance and varying exposures, the width 
and intensity of the bands increase with the exposure. With 
certain elements the bands are widened and intensified more on 
the less refrangible side; with others, on the more refrangible. 
This explains what has been remarked by Von der Seipen,} 
namely, that, between his measurements of the bands of metallic 
tin and mine, there is a large though constant difference in the 
wave-length values; and he attributes this to the old normal wave- 
lengths of Angstrom being used. The difference, however, 
between the two sets of measurements amounts to from 4 to 7 
Angstrom units, but over the same range of spectrum the 
maximum difference between Hartley and Adeney’s wave-lengths 
(1884) and Rowland’s (1893)? is, at most, + 1:1 A°. unit, the mini- 
mum being + 0:4, and the average something less than + 0°8. 
There is no doubt that my spectrum was photographed from 
a much larger quantity of material; and the exposure was also 
longer; and therefore the bands were broader and more intense. 
1 Uber das Flammenspektrum des Zinns,”’ Zeitsch. £. wissensch. Photographie, 
1907, Bd. v., pp. 69-85. 
2 J. F. Eder, ‘‘ Beitrage zur Spectralanalyse,” K. K. Akad. Wissensch., Vienna, 
1893, Bd. Ix., p. 13. 
