120 
determination which exceed the error of measurement many times. 
Also when the liquids have not been previously expressly dried with 
a view to P,O,, but when they have only been repeatedly fractionated 
after preliminary drying with CaCl, (which was the case with the 
substances used by us) they show already a strong absorption of 
water-vapour. Especially for C,H,Cl, this appeared to be the case, 
but though in a less degree, benzene gave also greatly varying 
values. 
When the determinations are carried out with quite the same 
precautions as are observed for aqueous solutions, two successive 
determinations executed directly after each other, already in conse- 
quence of this source of error alone, yield greatly deviating values, 
which e.g. for C,H,Cl, can sometimes differ inter se no less than 
40 scalar divisions. In order to ascertain the influence of the water 
content on the interferometer reading a previously weighed solution 
of 5 mgr. water in 35 gr. of C,H,Cl, dried on phosphoric acid was 
measured against this same C,H,Cl,. The displacement was 85 scalar 
divisions. From this follows that the presence of 0.005 mgr. of water 
in 2 ce. of C,H,Cl, already causes an error of a scalar division. 
11. It is self-evident that with such sensitiveness the utmost 
care should be taken to prevent any contact of the liquid that is 
to be examined and water-vapour, if an accuracy is to be reached 
comparable with that in solutions in water. Therefore the liquids 
were always preserved over phosphoric acid which had been heated 
at 160° for some days. The storing 
bottles had the shape as indicated in 
fig. 3. The conveyance into the vessel 
took place with the aid of pipettes 
J, which had been cleaned with 
benzene, heated, and filled with dry air. 
These were filled from A by fasten- 
ing them with a rubber tube on to 
the tube B, and pressing in dry air 
J at F. The solutions were prepared by 
weighing in flasks B, which had also 
been previously rinsed with benzene, 
heated with evacuation, and filled 
with dry air. The small quantities of 
Fig. 3. C,H,Cl, were conveyed into them by the 
aid of a glass capillary. The solvent was directly pressed over from 
the storing bottle into B, after the tubes G and E had been connected 
