8 



The Mechanical Action of Radiation, 



reversed. With a larger difference of temperature the re- 

 pulsion would respond to a greater heat, becoming attrac- 

 tion again, however, if a certain limit of intensity was 

 passed. Such observations finally revealed the necessary 

 conditions. Repulsion in air occurs when : 1st. The tem- 

 perature of the interior of the instrument has been for 

 some time a little higher than that of the external air ; and 

 2d. The degree of heat applied is appropriate to this 

 difference of temperature. 



The next step was to carefully determine the place of 

 the zero of the needle scale. Call to mind the arrange- 

 ment. The slender needle is suspended horizontally by 

 two parallel silk fibres. The torsion of these fibres is a 

 delicate force tending to hold the needle in a certain posi- 

 tion. The place of the disk thus held at rest is the zero 

 of the scale. IsTow the swino-ino^ of the needle toward the 

 zero can indicate no positive action of any force other than 

 the torsion of the fibres. Its swinging away from zero, 

 however, can be due to nothing less than some positive 

 force which overcomes the torsion. Having determined 

 the place of the zero, repeated observation showed that the 

 repulsion was real, since the needle moved beyond its zero 

 in opposition to the torsion of its fibres. 



In explanation of this curious efiecti offer the following 

 suggestions. It is well known that a cooled surface will 

 cause a downward air-current in contact with it, and 

 again that heat being applied will reverse the direction of 

 this current. But this reversal cannot be affected instan- 

 taneously. The effect of a gentle heat will be 1st, to slowly 

 neutralize the cold and thus gradually diminish the down- 

 ward current and 2d, afterward to produce, as gradually, 

 a current upward. Now while the cold current is being 

 neutralized, the torsion of the fibres will carrj^ the disk to 

 its zero and when the upward current is established the 

 disk will be again wafted toward the warmed surface. 



