278 
MESSRS. THOMAS MARTIN LOWRY AND PERCY CORLETT AUSTIN 
of molecular rotations for sodium light, by means of which a comparison may be made 
with rotations interpolated from the earlier data of Thomsen (‘ Jour. Prakt. Chem.,’ 
1886, new series, vol. 34, p. 74) and of Patterson (‘ Trans. Chem. Soc.,’ 1904, vol. 85, 
p. 1120). 
Our measurements of the molecular rotations of the tartrates show a close general 
agreement with the requirements of Biot’s law, but with small deviations of a different 
type from those recorded by Krecke. Thus whilst his values for [M] A 2 always passed 
through a minimum in the central portion of the spectrum, ours either pass through a 
maximum as in Table VI., or rise progressively as the wave-length diminishes as in 
Tables VII., VIII. and IX. In the case of sodium tartrate (22 • 54 gr. of Na 2 H 4 C 4 0 6 .2H 2 0 
in 100 c.c. of solution, Table VI. (a)) the agreement of our numbers with the require¬ 
ments of Biot’s law is remarkably close ; thus the product [M] A 2 has the value 20 • 29 
in the visible red region at wave-lengths 6708 to 6438, and the value 20-31 in the extreme 
violet at wave-lengths 4072 to 4005, rising in the intermediate region to a shallow 
maximum 20-68 in the blue at wave-length 4678 ; if all the 22 readings are considered 
the average value of [M] A 2 is 20-40, the maximum errors are +0-28 and —0-23, and 
the average error is only +0-11 or 0-5 per cent. 
In view of the smallness of the deviations from Biot’s law, it seemed probable that 
a complete agreement might be obtained between the observed and calculated values 
by using a “ simple ” dispersion formula containing a second arbitrary constant. In 
each case, therefore, the constants of the “ simple ” formula [M] = Jc/{\ 2 —A 0 2 ) were 
calculated in the usual way from the rotations for the two dominant mercury lines 
Hg 5461 and Hg 4358 , as shown in Table XI. 
Special attention should be given to the values of the dispersion-ratio” a 4358 /a 546 i, 
shown in the last column of Table XI., which increases and diminishes with the 
magnitude of the “ dispersion-constant ” A y 2 . The latter constant being the square of 
a real quantity must always be positive, so that the smallest value for this dispersion- 
ratio is that given by Biot's law, where a l , 2 = 0, namely, 
_ (5461) 1 A70 
<*5461 " (4358)"' 
The ratios shown above come nearer to this min imum than in any case hitherto 
investigated, and in one of the solutions of sodium tartrate, where A 0 2 falls to 0 • 00032, 
this minimum is almost attained. Some of the lowest values previously recorded are 
shown in Table XII. 
These figures show that the dispersion-ratio rarely falls below 1 • 630 or the dispersion- 
constant below 0-018; in other words, the absorption-band which determines the 
position of the vertical asymptote of the dispersion-curve maybe pushed out so far as 
A = % 0-018 = 0-135/x or 1350 A.U, but it never goes much beyond this, even in 
the case of the most transparent of the substances quoted in Table XII. When dealing 
