Colour and Chemical Constitution. 
227 
following data may be noted about the behaviour of these new halogenated 
phthaleins when dissolved in concentrated HgSO^ and spectroscoped. Under 
these conditions phenolphthalein gives a band with centre at about A 499 
(orange solution). The solutions of the halogen-derivatives are of various 
shades of salmon-pink and have the following central wave-lengths : 
2 monochloro-compound . . . . .A 502 
2-6 dichloro- „ ..... A 505 
2-6 dibromo. ..... X 506 
2-6 dibromo- 2' methyl compound . . . A 511 
2-6 dibromo-3' methyl „ . . . . A 521 
2-6-2'-6' tetrachloro „ . . . . A 530 
Although the formula for the halogenated colours is merely empirical — 
in the sense of lacking for the present a physical basis of explanation — it 
is obviously worth while seeing to what extent it applies to derivatives of 
phenolphthalein and fluorescein which are not halogenated. For this 
purpose such groups as methyl and isopropyl may be taken to have mole- 
cular numbers made up of the atomic numbers (1 for hydrogen and 6 for 
carbon) — viz. 9 and 25 respectively. On this basis ortho-monomethyl- 
phenolphthalein should have a frequency of 18'05 (1 — O'OllSo), corresponding 
to wave-length A 561, and the observed wave-length is \ 562 (the other name 
of the substance being phenolorthocresolphthalein). In the same way 
ortho-dimethyl-phenolphthalein (o-cresolphthalein) is calculated to have a 
frequency of 18*05 (l-"02366), corresponding to wave-length X 568, observa- 
tion giving X 570. 
If, further, we make the assumption (adumbrated in Part I of this series) 
that the meta-position has tivice the effect of the ortho-position, the calculated 
value for phenolmetacresolphthalein is X 568, and that for metacresol- 
phthalein (metadimethyl-phenolphthalein) is X 582. Observation gave 
X 569 and X 584 respectively. 
In the same way phenol-thyniolphthalein, which has one metamethyl and 
one orthoisopropyl group, should have the frequency 18'05 (1 — 2 x '01183 - 
•01242), correspoxiding to X 575 ; observation gave X 578. Thymolphthalein 
should similarly give 18*05 (1 — "0722), corresponding to x 597, which agrees 
precisely with observation. Again, phenol-carvacrolphthalein, which has 
one orthomethyl and one metaisopropyl group^ should have the frequency 
18*05 (1 — •01183 — 2 X '01242), corresponding to X 575^ ; observation gave 
X 580. So for carvacrolphthalein the calculated value is X 598 and observa- 
tion X 601. 
It is evident that the law is approximately true for all these varied and 
rather complicated derivatives, but equally evident that the agreement is 
not good enough and that the formula requires a small correction to make 
