[MORAN] RADIUM EMANATION FROM WATER 61 
unless prevented by the mercury. In that case, there is only one 
source of supply of gas left, viz., that through the branch tube, and 
this can be regulated by a pinch-cock. If the bath cools, the mercury 
falls, the supply of gas is increased, and hence the bath is raised in 
temperature and the mercury rises and shuts off the gas at the proper 
temperature. 
Procedure. The gasometer was first rinsed out several times 
with air before starting the experiment. Water was previously 
poured into the main cylinder in sufficient quantity to seal the open 
end of the open cylinder. Air was then pumped into the gasometer 
by the motor-pump to a height of about two feet, over which drop 
of the cylinder the pressure could be uniformly regulated. The experi- 
ment was started with a current of air equal to three bubbles per 
second, and with the temperature at 16-5° C. Readings were taken 
every few minutes, and the growth of the ionisation current traced. 
The maximum was reached in about three hours, and at the end of this 
period about twenty consecutive readings failed to show any variation 
exceeding about one-half of one percent. A graph accompanies this 
paper showing the rise curve for 16-5° C. The temperature of the 
bath was now raised to 60° C. as quickly as possible. It was found that 
the rate of discharge of the gold-leaf increased as the temperature 
was raised. Another curve was obtained similar to the first one, 
which also reached a maximum at the end of three hours. Fig. IT 
shows the change observed. 
The next step in the investigation was to heat up the bath to 
80° C. before bubbling, trace the changes in the rise curve for this 
temperature, and see if the maximum ionisation exceeded that for 
60° C. This was exactly what happened. The maximum reached 
was found to be proportionately greater by an amount which can be 
determined from the curve submitted. 
The warm water was then quickly siphoned off, and the water in 
the bath replaced by water at room temperature, or 16-5° C. The 
ionisation-current gradually fell off, until at the end of between 33 
and 4 hours it assumed the same values as obtained in the previous 
experiment for the same temperature. 
An attempt was then made to obtain a set of readings by bubbling 
the air through the solution at 100° C. A condenser was connected up 
with the previous apparatus for the purpose of condensing the steam 
which would be otherwise driven over and condensed in the connecting 
tubes. A Kjeldahl trap was also used to prevent any loss of the 
solution. The experiment, however, was found to be impracticable 
under the conditions, if a uniform current of air was to be maintained. 
It was found to be impossible to prevent the water from condensing in 
