694 
function of time, is also the distribution of temperature that prevails 
at a definite moment as function of place in the axis of the tube. 
This distribution of the temperature moves, as it were without 
changing with uniform velocity along the thermo-element, and the 
temperatures, existing simultaneously at the different points of the 
tube, are successively observed. 
In order to be able to derive the difference of temperature between 
the joint and the water in the thermostat, from the deviation of the 
galvanometer, the amount of the deviation corresponding with one 
degree, must also be known. For this purpose a current has been 
sent through the galvanometer circuit at the beginning and the end 
of every observation by an electromotive force of known value 
(2 «x 10-4 Volts) for some seconds. The deviation given by this 
current has also been registered. 
When the electromotive force of the thermo-element for one 
degree of temperature difference is known, the temperature may be 
derived by comparison of the deviations during the solidification and 
that which the known electromotive force has caused. A determi- 
nation of the electromotive force with thermo-sheet tin of the kind 
out of which the thermo-elements used had been cut, had as result 
that it amounted to 41.3 >< 10-® Volts’). 
As in these observations rapid variations of the temperature must 
be registered, the registering drum must rotate quickly. This being 
difficult to achieve with a clock-work provided with a balance, the 
balance was replaced by a flying-pinion. The objection to this way 
of propulsion, however, is that the movement of the registering 
drum is not uniform. For the determination of the temperature as 
function of the time from the temperature curves obtained, it was 
therefore necessary to place time signals on every curve. They 
were obtained as follows: a resistance was placed in the circuit 
of the Nernst lamp, which sends its light to the mirror of the gal- 
vanometer, and then in the registering drum. The extremities of 
the resistance were connected with a clock work, which every 10 
sec. effected a contact between them momentarily. In consequence 
of the diminished resistance the intensity of the lamp increased 
every 10 sec. for a short time; hence the registered line shows 
slightly thickened parts, which recur at intervals corresponding with 
this period. This method has the advantage that there are no gaps 
in the registered curve. 
1) This result is in perfect harmony with what others have found for unrolled 
material. 
