262 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1956 



the acceleration of small particles under the influence of radiation 

 pressure in the tremendous energy outburst of a supernova. He calcu- 

 lates that such particles can receive from a supernova outburst from 

 0.01 to 1.0 billion electron volts per nucleon (proton or neutron) of 

 each particle, within an interval of a few hours to a few weeks. 



It is also postulated that the particles are held within the galaxy by 

 an extensive magnetic field. As a result of collisions with atomic 

 nuclei in the galaxy, the particles which have escaped collisions with 

 stellar bodies break up ultimately into nucleons. The neutrons soon 

 change to protons because of their finite life expectancy, and the pro- 

 tons are lost finally by encounters with atomic nuclei or by striking 

 large bodies like the earth. An equilibrium condition is set up in 

 which the number of high-energy nucleons contributed to the galaxy 

 per unit time is equal to the number lost by the aforesaid processes. 

 By this means, a cosmic-ray intensity between 0.0001 and 0.01 of the 

 measured intensity is predicted, depending upon assumptions as to 

 the frequency of nuclear collisions in the galaxy. 



E. Fermi, in 1919, suggested an intriguing mechanism which may 

 be pictured in elementary fashion by thinking of a room containing 

 gas molecules and many hard steel spheres flying about and rebound- 

 ing from one another and from the walls of the room. It will be con- 

 venient to eliminate gravity temporarily during our meditations. 



According to well-understood principles of thermodynamics, the 

 spheres will, in the last analysis, lose their energies to the gas mole- 

 cules and will finally come to a state of equilibrium in which the aver- 

 age translational energy of each sphere will be the same as that of a gas 

 molecule. If the spheres are sizable, let us say 10 centimeters in radius, 

 their average velocity will then be very small compared with that of 

 the molecules. If, however, the spheres and the walls of the room are 

 perfectly elastic, and if the spheres have considerable velocities ini- 

 tially, it will be a very long time before they get to this final state of 

 equilibrium. 



Meanwhile, the spheres will seek another quasi-stationary equilibri- 

 um in which they have a velocity distribution among themselves which 

 is like that of the gas molecules, but with an average kinetic energy 

 of each sphere enormous compared with that of one of the molecules. 

 This kinetic energy will be approximately equal to the total original 

 energy of the spheres divided by their number. In other words, the 

 spheres will have a kind of macroscopic temperature of enormous 

 amount which diminishes extremely slowly to the final temperature 

 representative of the true equilibrium of both spheres and gas 

 molecules. 



The quasi-equilibrium of the spheres is not an accidental phenome- 

 non, but is an inevitable consequence of the laws of dynamics as 

 applied to the collisions between the spheres. 



