( 83 ) 
In this connection we may recall that this law, regarded as holding 
with close approximation, had long ago drawn attention to the 
uranyl salts in particular as being, from the point of view of optical 
phenomena, of remarkable molecular structure. This formed the 
starting point of the first experiments which were undertaken with 
a view to discovering rays analogous to those which had been 
observed in vacuum tubes. a ) They led to the discovery of radioac¬ 
tivity. 
§ 6. Character and nature of the phosphorescence spectra of the 
uranyl salts. 
We have seen what the law is that governs the succession of the 
homologous bands in the various groups. Is, now, the disposition of 
the bands in the safrie group governed by a particular law? In the 
most intense portion of the green of the spectrum the bands are 
V numerous, especially at the temperature of liquid hydrogen, forming 
a system in themselves. But regarding the bands at the ends of the 
spectrum, we find, especially in the last group in the red, that they 
become simpler, and in many cases there remains only a Succession 
of bands lying close together and showing marked similarity with 
the groups of bands in the channeled spectra of gases (nitrogen, 
carbon) when these are observed with a weak dispersion. In the least 
refrangible groups in the spectrum of uranyl sulphate (fig. 10) this is 
particularly well marked. This group consists of a strongly developed 
head towards the side of the small wave-lengths followed by seven 
or eight bands at regular distances from each other, and decreasing 
regularly in intensity as they go further way from the head. Each 
of these bands is a little asymmetrical, for the side towards the 
smaller wave-lengths is more sharply defined than the other. The 
blue group in the spectrum of the same salt (fig. 11) is also a 
channeled spectrum, but it is not so extended and seems to be formed 
out of various series of overlapping bands. 
The bands are not sufficiently fine and especially they are not 
sufficiently numerous for us to examine if the successive bands in 
the same group and the initial bands in the various groups follow 
Deslandres’ law. Nevertheless this law receives as much support as 
might be expected especially from the groups in the orange and red 
in the case of the simple sulphate. 
It seems to us, therefore, that these spectra are channeled spectra 
of the same character as the band spectra of gases. 
*) Henri Becquerel, Gompt. Rend. GXXII, p. 420. (1896). 
