by the production of a dark patch. If, however, the air carried no corpuscles there would be no image on the sensitive film over C.
The experiment was continued for 11 hours, 120 litres of air having passed through in the time.
On development and measuring the resulting images the following figures were obtained :
Circle over cell A Opacity log. = 0'342; Opacity = 2*20. Circle over cell B Opacity log. = 0'178 ; Opacity = 1-51. Circle over cell C Opacity log. = 0*025 Opacity = O'll.
Katio B/A = 0'68.
The apparatus was slightly altered. Cells B and C were put nearer together, and a short wide brass tube replaced the turned up glass tube formerly connecting them. The experiment was continued for 48 hours, during which time about 500 litres of air passed through the apparatus.
This repetition, allowing longer time, gave measurements as fol- lows :
Circle over cell A Opacity log. = 0'964 ; Opacity = 9'20. Circle over cell B Opacity log. - 0-730 Opacity = 5-37.
Ratio B/A = 0-58.
In this experiment a slight but measurable darkening took place over cell C, shown by an opacity log. of 0'025 and an opacity of 0*11. This shows that some few corpuscles have been able to pass through the wide tube from B to C, and act on the photographic film.
It thus appears that a current of air passed over the surface of a radium compound carries with it a certain proportion of the corpuscles. This is proved by the diminished photographic action in the second cell, slightly confirmed by the evidence that some few of the corpuscles so carried away get to the sensitive film on cell C. Judging from our slender knowledge of the properties of free electrons, it is highly pro- bable that they will not easily turn a corner, but cling to the sides of the tube through which they are being led. On the other hand, the constant collisions with the atoms of air may reduce their initial mobility almost to a vanishing point before they have travelled along the tube between B and C, and then they would be carried along with the air.
The experiment was repeated, using a preparation of actinium (Uranium X). It was kept going for 72 hours, during which time 750 litres of air were drawn through the apparatus. On develop- ment and measurement the following results were obtained :
Circle over cell A Opacity log. = 0'99 ; Opacity = 978. Circle over cell B Opacity log. = 0'67 : Opacity = 4'68. Circle over cell C Opacity log. = 0*25 ; Opacity = 1'78.
