22 
Proceedings of the Royal Society of Edinburgh. [Sess. 
or minima ; the maximum for violet has apparently neither the same value 
nor does it occur at the same temperature in homogeneous ethyl tartrate 
as in ethyl tartrate in some solvent such as nitrobenzene, or in some related 
ester, either in the homogeneous state or in solution. 
The behaviour of the esters of tartaric acid at higher temperatures 
having thus been examined with some care, it seemed natural next to 
attempt to follow the trend of these T-R curves towards low tempera- 
tures, adopting the same method as before. In the following experiments 
a start has been made in this direction, but before passing on to our 
primary purpose we thought it worth while to investigate one rather 
striking instance of solvent effect connected with the high-temperature 
end of the diagram. It has been shown by Walden ( Ber ., 1905, 38, 371) 
that cinnamic aldehyde has a very marked effect upon the rotation of 
ethyl tartrate and of methyl malate, the mere numerical values being 
greater than those recorded for either of these active substances in any 
other solvent, on which account we were anxious to ascertain whether the 
behaviour in the former case fitted in with the views previously developed 
or not. 
A solution of ethyl tartrate in cinnamic aldehyde, p * = 9‘64, was there- 
fore prepared, and the rotation examined for six colours of light. Three 
of these, yellow, green, and violet, were obtained direct from a mercury 
arc lamp, whilst the other three, dark red fa), red fa), and blue, which 
correspond to three fainter lines in the mercury arc, were obtained from 
a Nernst lamp by a method which has beenjlescribed elsewhere ( J.C.S . , 
1916, 109, 1144). The wave-lengths of the light used are given on p. 30, 
along with the experimental data. 
It will be noticed that the rotations at low temperatures are great, 
and that as the temperature rises the rotation diminishes, which was in 
accordance with expectation. The T-R curves for this solution are shown 
in fig. 2, and are there contrasted with those for ethyl tartrate in quinoline, 
y> = 13‘601 (J.C.S. , 1916, 109, 1145, 1151). It will be noticed that the curves 
are, on the whole, somewhat similar, but that those for the cinnamic 
aldehyde solution lie higher on the diagram than those for the quinoline 
solution. In quinoline a minimum is apparent, but in cinnamic aldehyde 
the curves are only tending towards a minimum, which would lie at a 
distinctly higher temperature than could, meantime, be reached. Now it 
has been suggested in recent papers that the influence of a ' solvent should 
be measured, not so much by the actual value of the rotation in given 
circumstances, as by the effect which the solvent produces on the whole 
* p = grams active substance per 100 grams of solution. 
