1569 
B—=tn.4a0?{1—1.0667 (hv)? + 0.1741 (Av)? — 
— 0.4738 (Av)? + 0.6252 (hv)? — 0.2360 (hv)® + 0.1355 (hv)? — 
— 0.1019 (hv)§ + 0.05934 (Av)? — 0.08579 (Av)? + 0.01910 (Av)? — 
— 0.00993 (hv)? + [0.0048 (hv)'* — 0.0022 (hv)4] 4.6 (1) 
The coefficients as far as that of (Av)? were calculated with the 
aid of the formulae given in Suppl. N°. 39a. Those of (Av)? and 
(hv)'* are only preliminary values, and are therefore placed between 
square brackets. As a matter of fact we found in calculating the 
preceding coefficients that the ratio of the whole coefficient to the 
sum of the two largest terms of which it is built up according to 
equation (11) of Suppl. N°. 39a, was only changing very little. For 
the coefficients of (hr)* to (hv)? this ratio amounts to: 1.287, 1.289, 
1.275, 1.277 respectively. Accordingly for the coefficients of (hv)'* 
and (hv)* we assumed a ratio of 1.275. The calculation of the sum 
of the two largest terms, which are furnished by the first two terms 
of equation (11) of Suppl. N°. 39a, may be shortened appreciably 
by making use of the formula : 
1 
6(n—-1) 
[A,"| being derived from [A‚*-!] with the aid of equation (15) of 
Suppl. N°. 39a. 
By the addition of the new terms equation (20) of Suppl. N°. 39a 
[42 Bo] = (APIs aie le 48) 
becomes: 
B= Be 1 - 0,8539 Ce + 0,03327 ta —0,05215 ts + 
+- 0,03964 f—5 —0,00862 r—§ + 0,00285 ¢—7 — 
“nv. inv. ib. 
— 0,00123 ¢-8 4 0.000414 1-9 —0.000147 t—10 + 
inv. mr.) ) 
(¢nv.) 
4- 0,0000442 r—11—0,0000182 e-12 + 
(unv.) nv.) 
+ [0,0000037 t—'8_0,0000010 alken (5) 
me.) anw.) 5 
The last term in (3) becomes=7/,, 6, for tiny — 0.49. Down 
to that temperature we may put the aceuracy at 1°/, of B. 
To table I of Suppl. N°. 39a we can now add the following 
figures: (see table p. 1570). 
In calculating %/s, for the doublets equation (21) of Suppl. 
N°. 39a was supplemented by the term —0,0000123 tin) between 
the {}. 
It appears therefore, that the values of 8/,, for quadruplets and 
for doublets begin to diverge more and more below 0,75 Tire 
