THE KINETIC THEORY OF GASES. 281 



although even imperfectly. Some such picture may be formed by 

 thinking of the motions of the players in a game of football. At a 

 critical point in the game the players are running, some this way, some 

 tbat; one has picked up the ball and is running with it, followed by 

 two or three others; while players from the opposite side are slanting 

 toward him, intent upon a collision. The backs are at rest, perhaps; 

 bnt, on the approach of the ball to the goal, they quicken into activity, 

 and the throng of human molecules is turned and pursues an opposite 

 course. The failure of this analogy to represent what is believed to 

 occur in a gas is that the players' motion is directed and has a pur- 

 pose; that they do not move in straight lines, but in any curves which 

 may suit their purpose; and that they do not, as two billiard balls do, 

 communicate their rates of motion one to the other by collision. But, 

 making such reservations, some idea may be gained of the encounters 

 of molecules by the encounters in a football held. 



In considering averages, it is clear that there must be a practical 

 limit on both sides of the mean. If a man throws dice, he may turn up 

 sixes thrice in succession, or some greater number of times, by chance; 

 but it is clear he will not go on throwing sixes forever, though there is 

 no absolute reason why he should not. Similarly, in thinking of the 

 rates of motion of molecules, there will be a practical limit of rate at 

 which any one molecule will move. It is unlikely that any one mole- 

 cule will cease to move for any appreciable time; and it is unlikely, 

 too, that any one molecule will develop any exceptionally rapid veloc- 

 ity, say twenty times the mean. Still, such events may conceivably 

 occur; they will, however, be very infrequent. 



Those gases which are light, and whose molecules have a high 

 intrinsic average rate of motion, will, in the nature of things, contain 

 some molecules which happen to be moving at a high speed; and neces- 

 sarily will contain more such than a gas of higher density, the average 

 rate of motion of whose molecules is slower. It may happen that 

 molecules of each kind, of gas with low as well as of gas of high density, 

 may possess such exceptionally high velocity at the confines of our 

 atmosphere, where there are comparatively few gaseous molecules 

 altogether; and it may also happen that these molecules are moving 

 in a direction more or less nearly perpendicular to the surface of the 

 planet, and it may also happen that such molecules suffer no collisions 

 in their vertical path; if these events all happen, the molecules will 

 escape. But as, on the doctrine of chances, there are more molecules 

 of light gas endowed with such exceptionally high velocity than there 

 are of heavy gas, more molecules of the former will escape away from 

 •the neighborhood of the planet and enter free space as independent 

 entities than of the latter. 



Such a process, prolonged over ages, will ultimately remove from 

 the atmosphere of a planet all gases possessing less than a certain mini- 

 mum density. 



