340 Harriey and RaMAGE—Banded Plame-Spectra of Metals. 
which accompanies the two strong lines which are common to flame and spark 
spectra. The bands are apparently of two kinds: those in the visible spectrum 
are of the nature of flutings, but those im the ultra-violet are groups of distinct 
lines. ‘The intervals between the component lines are smaller as they approach 
the stronger edge of the band which is situated at the more refrangible end of the 
spectrum. ‘The flutings in the visible spectrum have not hitherto been resolved 
* into lines by the instrument at our command, and undoubtedly will require a much 
higher dispersion. 
We have found that the flame-spectra of copper and gold are even more 
beautiful than that of silver: the lines forming the bands are sharper and more 
widely separated than in the spectrum of silver, and the bands of this character 
are more numerous. ‘The spectra of copper and gold have a much closer similarity 
in constitution than that of silver to either of them, but the complete details of 
these spectra have not yet been quite fully examined. In the three spectra there 
is this character in common, all the bands are degraded towards the red. 
Banpbrep FLAME-SpecrRA OF CoprreR, SILVER, AND GOLD. 
Measurements of the Edges of the Bands. 
CoprrEr. SILVER. GoLp. 
ieee | Intensity of Band. eas Intensity of Band. eS a Intensity of Band. 
4689 Strong. 3637 Weak. | 4452 Strong. 
4649 Strong. 3584 Weak. 4339 Strong. 
4280 Very strong. 3546 Weak. 8975 Very strong. 
4005 Strong. 3519 Very weak. 3652 Very strong. 
3777 Weak. 3358 Strong. 3457 Strong. 
3330 | Strong. 
The following are measurements of some lines which have been observed in the 
spark-spectra of these metals :— 
(Eder & Valenta.) (Exner & Haschek.) 
Copper, . . 4649°31 sharp, 3777°17 diffuse. 4651°3, 40038'1, 3777°3. 
Silver, . . 3358°79 not sharp, 33829°84 sharp. 3639°7, 3581°5, 35190, 3329°3. 
Gold, . . 89768 diffuse. 3650-9 sharp. 3976-77, agra 3457-05. 
A careful comparison of spark- with flame-spectra has led to the conclusion that 
the measurements quoted do not represent lines in the banded flame-spectra, 
although near to them. 
The flame-spectra of copper and silver contain, in addition to the bands, two 
very strong lines in the same region of the spectrum, which appear also as 
