48 
Transactions of the Royal Society of South Africa. 
observed wave-lengths of the absorption-centres, viz. 572, 547, 525, 504, 485, 
468 and 451 (observed 572, 546i, 5241 508, -486, 469 and 454). The 
reflection spectrum of solid KMnO^ contains emission-bands at 577, 552, 
530, etc., which evidently correspond to the non-ionised KMn04 of slightly 
higher molecular weight or volume than the ion Mn04~ to which the ordinary 
absorption bands belong. 
lY. Spectra of Uranium Compounds. 
The uranyl compounds in solution exhibit a spectrum which is a sort of 
imitation or continuation of the permanganate spectrum, consisting of five 
bands equally spaced in the blue. For the UO2++ ion (thick layer of dilute 
solution) the positions of the five absorption-centres are X\ 473, 459, 446, 
433 and 420, the central one being strongest. The effect of '* loading" was 
studied by examining solid uranium salts by transmitted light, excluding 
fluorescence by means of a black cover. Uranyl acetate ( + 3Aq) gave the 
readings XX 476, 464, 445, 432 and 422. Basic uranyl acetate (orange) 
gave XX 487, 471, 459, 449 and 440. Hydrated uranyl chloride gave only 
AX 483 and 467. Hydrated uranyl nitrate gave XX 486, 469 and 455, which 
remained unchanged even after melting the crystals and heating for some 
time. 
Uranyl bisulphate in cone. HoSO^ (strong thin layer) and uranyl meta- 
phosphate in HPO3 both had the same four bands, at XX 492, 475, 458 and 
438 ; uranyl perchlorate in cone. HCIO4 had the bands at XX 488, 471, 456 
and 443 ; in the solid state this substance gave bands at about XX 506, 490 
and 470, and possibly a higher and a lower one ; finally, solid natural 
autunite (from Madagascar*) gave XX 514 faint, 600, 485, 470 and 456. 
The latter two show great loading effect as a result of their much greater 
molecular size, and, speaking generally, the change in position of the 
bands follows the molecular weight qualitatively though not proportionally. 
In fact, the change in wave-length appears all through to be roughly 
proportional to the 6th root of the molecular weight. Thus for autunite, as 
914 
compared with the ion UOgt, the ratio of the molecular weights is of 
514 500 485 470 , 
which the 6th root is 1*091 : now the five ratios and 
456 
have all an average of 1'089 within the experimental error. 
A similar state of affairs holds for uranyl perchlorate if we assume one 
mol. of Aq. and make the line X 506 correspond with the second line of 
UO-j^"^, viz. 459 : the ratios of all the lines agree with the 6th root of the 
ratio of the molecular weights. 
* Ca(TJO-3P04)2 . 8Aq. 
t Multiplied by 2. 
