298 
From this it is evident that an accuracy of one in four thousand 
to one in five thousand is attained at the lower pressures, while in 
the series of measurements made at higher pressures the accuracy 
reached is greater than one in ten thousand. 
§ 4. Calculation of B4. From the former of the two series con- 
tained in Table I (pressures varying from 1.1 to 0.46) B4 can be 
calculated. Instead of 24260 = 0,00067 it gives 
B4200 = 0,00074 so that O—C = 0,00007, 
in which only the fourth decimal is significant. In the second series 
the percentage error expected in B4 is too great to allow of a 
calculation of B4 itself. Only under more favourable circumstances 
could one count upon an accuracy of one in ten thousand or more 
in the values of pva; the error in pv, becomes greater at smaller 
pressures; in B4 it is magnified four or five times and at small 
densities the utmost value of the whole term Byd, for that series 
is 0,00026. In the meantime it may be remarked that a comparison 
of the positive differences found here between observation and cal- 
culation (+0,0013) with the corresponding positive difference in the 
first series seems to indicate a possible systematic error which makes 
its presence specially felt at the lower pressures °). 
In order to be able to compare the results obtained with others 
which just had in view the determination of the compressibility at 
ordinary temperature we must reduce the results to a common basis. 
Take first the measurements made by Lepuc’) at 16°C. and at 
pressures varying from 1 to 1.5 atmospheres. From the numbers 
which he obtains from his experiments after the incorporation of 
other data for the compressibility at O° C. we find to correspond 
with his result 
B 4200 = 0,0007 and therefore O—C = 0,0000. 
The figure last given does not necessarily lead to the conclusion 
that the Leiden determinations with the volumenometer are the less 
accurate. The degree of accuracy of Lepuc’s results is indicated by 
the fact that he goes only to the fourth decimal place (for CO, 
Cuappuis') and Lepve differ by 0,0002). And the pressures used by 
Leptc in this determination, which is accurate to 1 in 10000 were 
very much more favourable (the smallest density was twice as great 
as that of the first series of Table I) than those which are expe- 
1) Possibly a small constant error arising from a change in the correction for 
the capillary depression since the control measurement of Comm, N°, 121a. 
2) A. Lepuc, Recherches sur les gaz. 1898. 
