460 
MR. W. CROOKES OK THE SUPPOSED 
extremities of this spiral, the brass tube E can be heated. The upper end of the tube 
is clamped at A, and the lower end rests in a socket, F, capable of a little lateral 
adjustment by a screw 7 . By this means the brass tube (the pillar) can be adjusted in 
respect to its distance from the cylinder D. Both E and D are lampblacked. 
Fig. 4. 
This apparatus is almost identical with one described and figured in my paper 
“On Repulsion resulting from Radiation” # (Part II., pars. 99, 100), only there the 
cylinder was of magnesium, and the platinum spiral was bare. With my old apparatus 
the noteworthy fact w 7 as ascertained that the incandescent spiral attracted the 
suspended cylinder to a moderate extent at normal atmospheric pressure; the attrac¬ 
tion diminished to a minimum between a tension of 50 mm. and 150 mm., then rose 
as the pressure diminished, until, at a tension of 1‘15 mm., the attraction w r as nearly 
four times what it was in dense air. Above this exhaustion the attraction suddenly 
dropped and changed to repulsion, and at the best vacuum I could get the repulsion 
w 7 as nearly thirteen times stronger than the attraction in air. 
With the present apparatus I was able to verify the broad phenomena formerly 
* ‘ Phil. Trans.,’ 1875, pp. 528-532. 
