CHEMISTRY: RICHARDS AND DAVIS 
55 
be set at the appropriate point, and at the moment of ignition M and N 
fully opened. The acid was then delivered automatically at a rate suit- 
able for the case in hand. 
Incomplete combustion. — In one of the earHer Harvard investigations 
the gases remaining in the bomb after the combustion of sugar were 
repeatedly analyzed for carbon monoxide or other volatile carbon com- 
pound, and none was found. ^ Nevertheless, with volatile liquids the 
danger is much greater, as is indicated by the careful work of Auwers, 
Roth and Eisenlohr.^ The final test is the actual analysis of the result- 
ing gases for combustible carbon compounds, especially carbon monoxide. 
This we have now carried out after many combustions, and can definitely 
report that with volatile Hquids the combustion is often somewhat 
incomplete. 
The arrangement of the apparatus was as follows: To the bomb, 
after each calorimetric combustion, was connected a train of tubes, 
as follows: 2 U- tubes filled with concentrated potassium hydroxide 
solution and glass pearls; a Liebig absorption tube containing saturated 
barium hydroxide solution; a drying tube filled with phosphorus pen- 
toxide; a tube of copper oxide heated to a dull red heat by electricity; 
a tube of phosphorus pentoxide, and finally either a Liebig's potash 
bulb or a spiral filled with a definite quantity of tenth normal barium 
hydroxide, which was afterwards titrated. 
In very many cases considerable quantities of carbon monoxide were 
thus found, often corresponding to errors of several tenths of a per cent 
in combustions carried out by the older methods, which seemed other- 
wise complete. This discovery entirely explains the discrepancy in 
previous results of many experimenters for benzene. Fortunately, it is 
only with volatile Hquids of this sort that any important incomplete- 
ness of combustion was found. 
On the other hand, the combustion was found to be fairly complete 
when volatile liquids were burned by the finally approved method of 
enclosing the liquids in very thin bulbs ignited by the ring of parafiin. 
The gases from a combustion of benzene were also tested for oxidiz- 
able carbon compound by bubbhng through a weak solution of alkaHne 
potassium permanganate, and for ozone by allowing them to inpinge 
on starch-iodide paper. Both of these tests gave negative results. 
Source and preparation of materials. — Napthalene was obtained from 
the Bureau of Standards; it had evidently been prepared by them with 
very great care, gave every evidence of having a high degree of purity, 
and serves admirably for standardizing calorimetric determinations. 
We found it much superior to sugar. 
