38 
Transactions of the Royal Society of Soidh Africa. 
(8) Mono-phenolphtJialein. — 
n >0U ^ 
from phenol and phthalaldehydic acid is yellowish and has the calculated 
A = 290 x factor C = \ 349. This agrees with my observation so far as such 
can be made, as the centre of the band is just below the limit of visibility, 
using sunlight. 
The number of substances in the monocyclic class is limited, and the 
remainder are, I believe, all coloured in or near the ultra-violet, and therefore 
beyond my observation. I therefore predict their colours so that some other 
observer possessing an adequate photographic apparatus can verify them. 
(9) Para-aminobenzhydrol— Calculated A = 290 x 1 135 x 1 14 = A375. 
(10) Para-a.minobenzyl alcohol. — Calculated A = 290 x 114 = A330f. 
(11) Para-dimethylaminobenzyl alcohol. — Calculated A. = 290 x 122 = 
A 354. 
(12) Para-oxyphthalide.- 
HO< >— CH, 
XX) — o 
probably has A = 290 x 1 06 = A 307 and so on for the corresponding 
amino- and dimethylammo -compounds. 
It is probably an important point that the parent-substance (whether 
regarded as quinomethane or as oxybenzylalcohol) has its absorption-band 
close to those observed for sodium phenate and paracresolate (Baly and 
Ewbank, ' J.C.S.; 1905, p. 1352), the former being A 290, and the latter, 
calculated to water from alcohol, about A 285. The difference is similar to 
that caused by a methyl- group, * so we can infer that in alkaline water 
HO-^ CH 3 and HO—/ \— CH. 2 OH 
have very nearly the same colour (absorption, speaking more strictly), viz. 
A 290. 
It would therefore seem that colour cannot be traced to any lower terms 
than this, namely, the presence of one benzene ring with one ionisable group, 
and (usually) another active atom. All the classes of colours except the 
azo-dyes and certain nitroso-derivatives are therefore accounted for in these 
terms, so it is probable that the conception of the physical cause of colour 
as the periodic orbit of an electron round a molecule is the correct one. 
The sinuous orbit of Part XII is not very likely, but the tautomeric pauses 
assumed therein are quite a probable feature of the orbit. 
It is to be noted that in the dicyclic compounds two rings, each carrying 
* Baly and Ewbank's diagram is too small to show the difference. 
