(51 ) 
COLOUE AND CHEMICAL CONSTITUTION. 
Part YIII. — Fluorescence and its Laws. 
By James Moir, M.A., D.Sc, F.I.C. 
While engaged in investigating the absorption-spectra of uranium com- 
pounds (see Part YII), I noticed that the green fluorescence which they 
mostly exhibit had a very definite emission- spectrum of 4 or 5 nearly 
equally spaced bands of about the same breadth as the bands (in the blue) 
of the absorption- spectrum. This phenomenon has been known for a long 
time in the case of uranium glass, but I think it is generally believed that 
all uranium compounds give the same fluorescent spectrum, which is not 
the case. 
It is true that all aqueous uranium solutions, which by the way are devoid 
of fluorescence, if examined in a thick layer of dilute solution, give the same 
absorption- spectrum, this spectrum being due to the ion UOo"^"^. The 
differences are only shown when there is no ionisation, i. e. in the solid state 
or in non-aqueous solution — and in such case there is generally a fluorescence. 
The following are the positions of the centres of the bands in these 
emission-spectra of fluorescent uranium compounds : — 
(1) JJranyl nitrate crystal: \K 560, 535, 510 and 485 (vague). 
(2) JJranyl acetate crystal (yellow) : only one bright band, at about X 525. 
(3) Basic uranyl acetate (orange) : XX 560 (vague), and 536 and 511. 
(4) Autunite {calcium uranyl ^Jhosphate) : XX 575, 554, 529 and 508. 
(5) Uranyl Usulpliate (in HgSO.i.) = ^^5, 541, 514 and 492. 
(6) Uranyl metajpliosphate (in IIPO3 solid) : X X 590, 566, 542, 515 and 495. 
(7) Uranyl sodium metaphospliate (microcosmic bead): XX 602, 562, 543 
and 517. 
(8) Solid uranyl chloride had a fluorescence too faint to observe. 
As will be seen on examining these figures, the bands are equally spaced 
or rhythmic : probably there are always five of them, but some are too faint 
to observe, and it is not always the same one in the order which is the 
brightest. It is therefore rather difficult to correlate them, ])ut I have 
succeeded in finding a remarkable numerical relationship between the wave- 
length of any fluorescent band and that of the absorption-band which 
5 
