18 GEOLOGY. 



were driven apart by the explosive force of vapors formed at their 

 points of impact, as suggested by Darwin. As most bodies are neither 

 perfectly elastic nor perfectly inelastic, an intermediate result would 

 usually follow. In exceptional cases, the velocities might be low 

 enough to permit inelastic bodies to weld together, but apparently the 

 chances of disruption, or of elastic rebound, would be many times 

 greater than those of welding. Even in the case of welding, the result 

 is merely a larger meteorite in lieu of two smaller ones, and no progress 

 is made in the starting of a meteoritic swarm. Besides, the enlarged 

 meteorite is subject to disruption by succeeding violent collisions. 



The possibilities of parallel movement. — Apparently the only 

 working chance of starting a swarm of meteorites by attraction, under 

 these adverse conditions, lies in the exceptional case of meteorites 

 moving in nearly parallel directions, at nearly the same speed, and 

 in courses near one another. In this case the motions of the meteorites 

 only antagonize their mutual attractions to the extent of such small 

 differences of kinetic energy as may arise from their slight differences 

 of velocity and directions of motion. Under extremely favorable con- 

 ditions of this kind, two meteorites might come into mutual gravita- 

 tive control and revolve about their common center of gravity. Then 

 a third one might join them under like conditions, and so on. But the 

 plane of revolution of the third meteorite might chance to correspond 

 with that established by the pair it joined, or it might not. So also 

 its direction of revolution might or might not be the same. It is there- 

 fore extremely unlikely that the planes of revolution of any considerable 

 number of meteorites, coming thus together, would be even approxi- 

 mately identical, or that the directions of their revolutions would be 

 coincident, and hence opposite and cross revolutions would result, 

 with obvious liability to collisions, so that, in the end, the swarm would 

 perhaps develop into a quasi-gaseous condition. 



A possible alternative, if the material were sufficiently inelastic, 

 might be the mastering and killing out of the movements of the minority 

 of the group by the majority through a long series of collisions, and 

 the development thereby of a common direction of revolution, in which 

 case the system would take on the orbital, rather than the gaseous type, 

 and fall under the plane tesimal class. 



The danger of dispersion. — It must be noted, however, that the 

 conditions for starting the growth of such a swarm — nearly parallel 



