1455 
was not attributable to a change of radiation in the radium tube, 
but to some other cause. 
Various test experiments seemed to show that it was caused by 
change in the power of absorption of the screens, due to their con- 
traction at low temperatures. It was therefore necessary to make 
use of heavier and thicker screens, to make sure that we only 
worked with the most penetrating rays, which are less susceptible 
to phenomena of this kind. After the radinm tube had been inclosed 
in a copper tube of 1 mm. thickness, we found that the decrease 
of enrrent when the liquid air was poured in was reduced to 0.1°/.. 
The decrease was completed in 10 minutes. Three successive expe- 
riments gave this result. 
We found that we could make the circumstances even more 
favourable, by changing the arrangement of the apparatus in such 
a way that the screens in which the absorption of the rays took 
place were not cooled at all. In order-to do this, the radium tube 
was once more put into the aluminium tube of 0.3 mm., while the 
central tube D, of the chamber D, was replaced by a brass tube 
of 2 mm. wall thickness. The decrease of the current became by 
this means less than 1 in 1000. This arrangement was used in the 
final experiments. 
Final Experiments. 
The experiments were made in Leiden from July 20th to 25th1911. 
The ionisation current in the main ionisation chamber was 1100, 
expressed in arbitrary units (about 10 electrostatic units). The strength 
-of the compensation current was so regulated that it was a little 
larger. The difference was at most 20 units, about 2°/, therefore. 
The rays used for the experiments were gamma rays. 
We were“ able to make two experiments with liquid hydrogen. 
[In the experiments the cold ionisation chamber, as we said above, 
was filled with dry gaseous hydrogen, and by this we made sure 
-that no deposit could come on to the radium tube. 
In the first experiment the current of originally 10.9 units, attained 
the value of 14.7 units after the pouring in of the liquid hydrogen, 
which took 15 minutes. This change corresponds to a change in the 
main current of 0.34 °/,.. In the second experiment the current 
measured had a strength of 18.3 units, and was very constant, the 
irregularities measured dufing an hour were less than 1/10000 of 
the main current. After the liquid hydrogen had been poured into 
the apparatus, measurements which agreed very well with each 
other gave for the value of the current during half an hour 18.5 
