| 
1908.| Agents in the Oxidation of Amorphous Carbon. 247 
renewing the tests, that the air contained within the incubator was then 
maintained at a strictly uniform temperature. The galvanometer leads 
passing through the incubator wall obviated the necessity of opening and 
closing the door, and permitted the research-flasks to remain quite undis- 
turbed. 
The glass flasks designed for this experiment were constructed with 
double walls. The inner flask, with a capacity of about 300 cc., was 
prolonged into two tubes diametrically opposite to each other, the upper one 
serving for the introduction of the charcoal and the lower for the escape of 
any CO, that might be formed. This inner flask was surrounded by a 
similar larger flask to provide a vacuum, the intermediate space being 
exhausted of all air in order to prevent any radiation of heat. 
Two similar vacuum-flasks were employed. After washing with fuming 
nitric acid, to remove any organic matter and for the purpose of sterilisation, 
they were filled with freshly-heated charcoal, which was moistened with 
sterile distilled water until it had absorbed all the water possible, the 
surplus being allowed to escape by the lower tube. It is important to note 
that the flasks were filled from the same supply of distilled water and with 
the same quantity. Into each of these flasks there was then inserted a 
thin test-tube previously washed externally in nitric acid and sterilised over 
a Bunsen flame, and the flasks were plugged with sterile asbestos wool. 
Finally, the iron-nickel thermopile was introduced into the test-tubes. 
The flasks were then placed in the incubator, the connections made with 
the galvanometer, and the apparatus was complete. Observations were 
taken at frequent intervals. When first arranged, the unavoidable handling 
of the metallic junctions necessarily produced thermo-electric currents, but 
these died away when the apparatus was left undisturbed for two or three 
days. When on depressing the key no movement of the spot of light could 
be observed, the whole arrangement was considered to be perfectly reliable, 
and the experiment could be proceeded with. 
One of the flasks was then inoculated with the bacterium, and thereafter 
readings with the galvanometer were repeatedly taken. At first no 
deflection was apparent on depressing the key, but after two days a move- 
‘ment of the spot of lght could be observed, and by depressing the key 
synchronically with the swing of the mirror, a deflection of several divisions 
of the galvanometric scale could be registered. The thermo-electric current 
gradually increased until after an interval of six days the maximum deflec- 
tion of 38 divisions was attained, proving a definite rise of temperature in- 
the inoculated flask, while the sterile flask indicated no increase. 
To ascertain with certainty that the movement of the mirror was a correct 
