RADIATION OF ENERGY. 353 



the bi-spherical oscillator, and an unlimited number of others. It is of course 

 possible that the action of all oscillators may be reduced to that of the electric 

 doublet, or to that of the electron vibrating about a mean position, but it aeems 

 to me to be very improbable. 



Of especial interest are a class of little known oscillators consisting of closed 



conductors, the disturbance taking place within the space eneloeed by the 

 conductor, and not exterior to it. 1 Of such a character are the electrical oscil- 

 lations within tubes which have been investigated by Larmor, 1 Lang, 1 Ra\ leigh, 4 

 Drude, 6 Weber, 6 Kalahne 7 and others. The theory developed by Web r gives 

 the wave-length of the fundamental vibration within a straight tube as equal 

 to 3.415 times the radius. For the closed conductor we may have a straight tube 

 closed at the ends, a closed circular tube, or a cavity of any shape completely 

 closed by a conductor. If the electrical disturbance takes place entirely v thin 

 the cavity of the conductor, and there is no disturbance outside of t he conductor, 

 we shall have what may be called an interior system. Systems oscillating 

 electrically may be divided into interior and exterior systems, according as the 

 disturbance takes place within a conducting surface or outside of it. 



An electrical system may pass by gradual steps from a closed into an open 

 system. For instance, if we have a straight conducting tube closed at the ends, 

 and an oscillation taking place within it, we may imagine that a small hole is 

 made in the walls of the tube. Speaking in the language of the Maxwell theory 

 we may then say that a small part of the radiant energy within the tube will pass 

 out of it, and it will no longer be a closed system. The fraction of the total 

 energy issuing can be made as large as is desired by increasing the size of the hole. 



If the walls of a closed system are perfectly conducting, after th< state of 

 radiation has once been set up and a steady state has been reached it will con- 

 tinue for ever if the volume and shape of the closed surface remain unchanged. 



It is remarkable that if the walls of a closed system are perfectly conducting, 

 and it is at rest, there will be no way in which a human being who is outside 



of the system can detect the existence of the electrical oscillations within the 

 system. 



On the other hand, Hasenohrl 8 has shown that it follows as a result of Max- 

 well's theory that if such a system is set into motion the radiant energy, E, 

 within it will produce an apparent increase of mass equal to 8/3 E\& grams. 



When I say that an electrical disturbance takes place at a point in space, I simply mean that if an 

 electric charge is placed at that point it will be acted 



disturban 



Lond. Math 



can be no disturbance 



by a force. I do not mean to imply that any 

 In fact, I would say, that if there is no physical 



Sitz. Wien. Akad., II (a), 104, 980. 1895: 105. 253. 1896: Wied. Ann.. 57. 430, 1896 



Wied 



Mag., 43, 125-132 



1 Wied. Ann., 8, 743, 1902. 

 Wied. Ann., 18, 92-127, 1905; 19, 80-115, 1906 

 Annaten der Physik, Bd. 15, pp. 344-370, 1904. 



JOURN. ACAD. NAT. SCI.PHILA.. VOL. XV. 





I 



