Colour and Chemical Constitution. 
7 
phenolphthalein having its band at A. 554 (frequency 18-05), and aurine 
having its band at A 530 (frequency 18'88). In the former case the change 
■caused by the extra — OH group is insignificant, the frequency having only 
risen by 0*2 ; but in the case of aurine the difference is very great, the 
frequency being raised by 1*6 or nearly 10 per cent. This abnormal 
behaviour of aurine is doubtless connected with its ahnost perfect chemical 
symmetry as compared with all the other compounds which I deal with : 
as anhydride of C(OH) (CgH^OH).^, it has always three paths by which it 
may become quinonoid in the form C{Q^llfi)~pllo. It may be noted that 
concentrated alkali raises the band of aurine to A 539. 
Again, phenolphthalein may be looked on as the o-carboxylic acid of 
benzaurine, whence we infer that this — COoH group causes a change of fre- 
quency of 18-05-1 7- 55 or 0-5, from which the theoretical frequency of the 
parent substance fuchsone can be deduced to be about 17-3 — viz. that of 
monoxydiphenylphthalide less 0-5. Thus, the passage from fuchsone to oxy- 
fuchsone (benzaurine), with destruction of the symmetry, involves a rise of 
only 0*2 in the frequency, whereas the entrance of a second OH group, giving 
aurine and restoring the symmetry, leads to a rise of 1*6 in the frequency. 
This theoretical figure for alkaline fuchsone is confirmed by the fact that it 
gives a frequency of 22-0 when dissolved in HoSO^; this is 1*4 above the 
value for monoxydiphenylphthalide, and benzaurine is found to give a fre- 
quency in HgSO^p which is 1-5 above that of phenolphthalein. 
To complete the series I have prepared a small quantity of the unknown 
aurine- o-carboxylic acid (which is also the 4'-oxy-derivative of phenolphtha- 
lein) by condensing 4'-oxyphthalic acid with phenol. It was found to give 
a pink solution in caustic alkali, possessing a band at A 542, exactly half-way 
between those of aurine and phenolphthalein. 
Study of the Substituted Phenolphthaleins. 
These were chosen for investigation on account of their easy synthesis 
rendering it possible to make a large number of derivatives ; but it was also 
found that their absorption-bands are remarkably definite as a rule, thus 
rendering accurate measurements of their centres possible, which is generally 
not the case with the commoner dye-stuffs, particulary those of the azo-class 
which have been hitherto investigated. The centre of the band can generally 
be estimated within one unit of wave-length (in 10 metres) in the green, 
but with less accuracy at the red end ; consequently the frequency values in 
Tables I and II can, as a rule, be trusted to within about 0-05 unit — i. e. to 
an extent considerably finer than that to which the eye can appreciate differ- 
ences of shade of colour. 
A set of 50 phthaleins has been prepared and examined, and the results 
are classified in two tables. Table I gives the results for the simple deriva- 
