THE FUEL OF THE SUN. 33 



ing from it, would be subject to disturbances which would 

 continue until one of two things occurred. " Its tangential 

 force might become so far neutralized and its orbit so much 

 elongated, that finally its perihelion distance should not 

 exceed the solar radius, when it would finish its course by 

 returning to the sun. On the other hand, its tangential 

 velocity might be increased by heavy pulls from Jupiter, 

 when slowly turning its aphelion path, and be similarly in- 

 fluenced by friendly jerks in crossing the orbits of the in- 

 ferior planets ; and thus its orbit might be widened, until it 

 ceased periodically to cross the path of any of the planets 

 by establishing itself in an orbit constantly intermediate 

 between any two. Having once settled into such a path, 

 it would remain there with comparative stability and per- 

 manency. If I am right in this view of the dynamical his- 

 tory of these older ejections, all the long elliptical paths of 

 zodiacal particles, meteorites, or asteroids, would thus in 

 the course of ages become eliminated, and the remaining 

 orbits "would be of planetary rather than cometary propor- 

 tions." 



A little reflection on the above-stated laws of dissociation 

 will show that the maximum violence of hydrogen explo- 

 sion will not occur at the birth of the ejections, but after- 

 wards, when the dissociated gases have been already hurled 

 beyond the sphere of restraining vapors. If my explana- 

 tion is correct, the typical form of a solar prominence should 

 be that of a spreading tree with a tall stem. At first the 

 least resistance to radiation and consequent explosive com- 

 bination must be in the vertical direction, as this will afford 

 the shortest line that can be drawn through the thickness 

 of the surrounding jacket of resisting vapor; but when 

 raised above this envelope, the dissociated gases, cooled by 

 their own expansion and comparatively free to radiate in 

 all directions except downwards, will explode laterally as 

 well as vertically, and thus spread out into a head. My 

 theoretical prominence will be, in short, a monster rocket 

 proceeding steadily upwards to a certain extent, and then 

 gradually bursting and projecting its missiles in every di- 

 rection from the vertical to the absolutely horizontal. 

 Should the latter acquire a velocity of about 300 miles per 



