264 
MESSRS. THOMAS MARTIN LOWRY AND PERCY CORLETT AUSTIN 
but the relative intensity of these rotations for the different simple rays, obeys the 
general law of this phenomenon, to which tartaric acid alone is a marked exception, at 
least among all the bodies which I have been able to study hitherto.” 
Preliminary observations appeared to confirm Biot's observations, since the dis¬ 
persions of the tartrates in aqueous solutions not only approximated very closely to the 
requirements of the simple dispersion formula, but, in the case of sodium tartrate, 
showed a surprisingly close agreement with the law of inverse squares. The conclusion 
that the rotatory dispersion of the tartrates is “ simple ” was, however, open to grave 
suspicion on account of the extreme smallness of the dispersion-constants, corresponding 
with absorption-bands not far removed from zero wave-length. Moreover, on calculating 
a simple dispersion formula for sodium tartrate from the data for Hg 5461 and Hg4359, 
selected from a particularly long series of readings (including 16 wave-lengths in the 
visual and 6 wave-lengths in the photographic region of the spectrum, instead of the 
short series of four or five wave-lengths which we have generally used in investigating 
cases of simple rotatory dispersion), we obtained an unmistakable series of positive 
differences from red to green, negative between green and violet, and positive again 
beyond the violet mercury line. These differences were observed in two solutions of 
different concentrations, and could not therefore be accidental. It was therefore clear 
that the curves for the metallic tartrates were not really simple, but complex, with a 
large positive and a small negative term, so that the anomalies would be pushed right 
out into the ultra-violet region of the spectrum where the solutions are too opaque for 
observation by ordinary methods. 
In the case of boric acid we were more fortunate, since when an excess of boric acid 
was added to tartaric acid the aqueous solution of boro-tartaric acid gave a simple 
dispersion curve, with a normal value, 0-0246, for the dispersion-constant A 0 2 . Boro- 
tartaric acid then appears to be a “ fixed ” derivative of tartaric acid, in which the acid 
has been locked up in one of its two alternative forms. A similar result was obtained 
with tartar emetic, which gives very large dextrorotations, but a perfectly simple 
dispersion, with a dispersion-constant A 0 2 — 0*0494. Boro-tartaric acid, which has 
many analogies amongst the polyhydric alcohols of the sugar group (see p. 252) is 
probably 
HO.B 
\ 
O—CH.CO.OH 
O—CH.CO.OH 
and it is possible that the simple character of its rotatory dispersion may be due to the 
bridge between the two asymmetric carbon atoms which is shown in this formula. 
The search for a “ fixed ” derivative of the elusive lsevorotatory modification of the 
acid proved even more difficult than in the case of the dextrorotatory component. 
The negative rotations discussed on pp. 280 to 284 of this paper are usually complex 
in their dispersion, but we were fortunate in discovering that the lsevorotatory solutions 
obtained by dissolving tartar emetic in an excess of alkali are not only comparable in 
