Colour and Chemical Constitution. 
305 
molecule. From the analogy of the behaviour of fuchsiiie I should estimate 
that the ion of Doebner's violet in water would have X about 560. My 
own preparation has A 562. The calculated wave-length for the triamino- 
derivative of fuchsene, which is the ordinary dye parafuchsine, is \593 X 
(0-972) or \ 544. This agrees with direct observation in water. 
These two ratios can next be combined. The theoretical wave-length for 
oxyfuchsonimine or aminofuchsone is X 593 x 0-9657 x 0-972 or X 556^. 
Observed X 558. That for diaminofuchsone is X 593 x 0-9657 x (0-972)2 
or X 541. Observed X 541. That for dioxyfuchsonimine is X 593 x (0-9657)^ 
X 0-972 or X 537^. I have made this new substance, which is alternatively 
named paramino-benzaurine, and observed its wave-length to be X 538. 
The third ratio is the N-methyl-f actor, which has the value 1-0245. 
From this the colours of all the violet and green dyes of the series can 
be calculated. It will be found more convenient, however, to combine the 
second and third factors, giving (fourthly) a y-N-methylamino-factor, the 
value of which is practically unity (viz. 0-9965), and (fifthly) a y-N-di- 
methylamino-factor, the value of which is 1-021. 
Applying these, the theoretical colour of " fuchsonedimethylimine " 
(dimethylaminofuchsene) is X 593 x 1-021 or X 605|^. The only band, in 
water, of this substance is very broad, at about X460, and may be a harmonic, 
since f of 605 J is 454. The fraction -~f agrees exactly : see previous 
page. 
That of bis-dimethylaminofuchsene (" malachite green '') is X 593 x 
(1'021)" or X 619, agreeing with observation. That of *tris-dimethylamino- 
fuchsene ("crystal-violet") is calculated to be X 593 x (1-021)^ or X632. 
This agrees with observation in the green phase of crystal violet obtained by 
adding a trace of mineral acid to its solution (observed X 632, but in neutral 
solution A 595). t Hofmann's violet (pentamethylpararosaniline) is calculated 
to have wave-length X 593 x (1-021)^ x 0-9965 or X 616 ; this also agrees 
with the acid phase of this dye. It is to be noted that the commercial 
article is partly derived from orthotoluidine (methylated fuchsine or 
rosaniline, not parafuchsine) and therefore gives colours a shade higher up 
the scale. 
I have also made the remaining new member of this series, dimethyl- 
aminofuchsone (dimethylamino-oxofuchsene), by condensing paraoxybenzo- 
phenone with dimethylaniline and AlCl.^. This has a wave-length of X 586 
when neutral (violet, sparingly soluble), but develops a pink colour and two 
other bands, XX 500 and 530, when acid (isomers ?). The calculated colour is 
A 593 X 1-021 X 0*9657 or X 585. Another member of the series is " paraoxy- 
malachite-green," which is bis-dimethylaminofuchsone. Its calculated colour 
is that of the foregoing multiplied by 1-021 or X 597, which agrees with its 
* The correct Latin for thrice is ter but might be unintelligible, 
t The violet phase is discussed later on. 
