CH. xix.] INSTABILITY OF ATOM 183 



as being determined by the character of the composite 

 radiator; while the occurrence of the radiation itself is 

 limited to the periods during which the displacement 

 by chemical clash or collision continues effective : a 

 period in each individual case probably very short, 

 but rapidly renewed in the aggregate by the com- 

 bination or collision of other molecules. 



Radiation due to high temperature may be caused 

 in a somewhat similarly accidental and violent fashion, 

 not occurring at all appreciably during the intervals 

 of peace or regular motion. 



Instability of an Atom. 



The principle on which instability of an atom is 

 to be expected, on the electrical theory, at certain 

 critical points in its history, may, according to 

 J. J. Thomson, depend upon the fact that a rotat- 

 ing ring of electrons is only stable so long as their 

 angular velocity exceeds a certain critical value. 

 Three electrons indeed are stable even when stationary, 

 though they become more stable by rotation ; but 

 four corpuscles in one plane, at the corners of a 



square, are not stable unless they are rotating with an 



// N e 2 \ 

 angular velocity greater than /I *325 J^L or '57c 



as it is called in the next chapter. Whenever their 

 speed falls below that value, they adjust themselves 

 at the corners of a tetrahedron or triangular pyra- 

 mid, tumbling over into the new position with a 

 rapid collapse, analogous to the tumbling over of 

 a top when its rate of spin has fallen below a 

 certain critical value. 



With a greater rate of rotation, five corpuscles in 

 one plane may be stable ; but with six in a ring, 



