14 
Transactions of the JRoyal Society of South Africa. 
the fuclisoue family can be investigated in HgSOj^ solution which are not 
soluble in alkali — e/g. fluorane. Table III gives some of the observed 
results. 
The author has been able to trace the very simple law which governs the 
depression of the colour by H^SO^ as compared with alkali. The expression 
y = ^(a' — 4-55), in which x is the frequency corresponding to alkali solution 
and y is the frequency corresponding to HoSO^ solution for the same 
phthalein, is found to fit all the direct derivatives of phenolphthalein, and^ 
with the trifling modification to ?/ = ^ (.v— 4-7), also fits the whole of the 
fluorescein series. This means that when the values for y and x actually 
observed are plotted against one another in the case of all the 50 phthaleins 
and fluoresceins, all the 50 dots lie approximately on one line. The fact 
3 
that the quotient is - seems to indicate that, as the vibration is 1^ times as 
fast in HoSO^, there must be three phenyl rings vibrating in H2S0j^ solution 
as against two only in alkaline solution. Extraordinary as this may seem, 
there appears to be no other possD^le explanation of the raising of the fre- 
quency by loading with HoSO^^. 
It is, however, just possible that the bands in H^SO^ are not the bands 
of alkaline solution raised in frequency, but are secondary bands (originally 
in the ultra-violet) lowered in frequency by loading with HoSO^. Such 
secondary bands in alkaline solution, if they exist, would probably have 
either one-half or one-third of the wave-length of the visible bands, so that 
the above-discovered law of a linear relationship betw^een HoSO.^ and alkali 
solutions would still hold true. 
The constant 4*55 (or 4-7) probably expresses the effect of the loading 
with HgSO^, regarded merely as increasing each of the molecular weights 
by 98. 
