110 On the Formation of the Tails of Comets. 



1. There appears to be no satisfactory reason to be assigned 

 why the force which expels the nebulous matter to the end of the 

 tail should not urge it still farther. Let us take the case of the 

 comet of last year. The extremity of its tail was, at one time, 

 at about the same distance from the sun as was the nucleus of 

 the great comet of 1811, a while after its perihelion passage; 

 when it had a bright tail more than a hundred millions miles in 

 length. The force that was sufficient to expel such a quantity 

 of nebulous matter from this latter comet, ought it not to have 

 driven still farther the much rarer matter at the remote parts of 

 the visible tail of the comet of 1843 ? A resisting medium might 

 give a limitation to the velocity of flow, but could not destroy it 

 altogether. 



2. In the case of the comet of last year, the repulsive force of 

 the sun, which, by the theory in question, is supposed to keep 

 the tail continually opposite to the sun, or nearly so, could not 

 have been of sufficient intensity to do this, without materially 

 deranging the orbit. If this cannot be proved to a mathematical 

 certainty, it can, at least, be rendered highly probable ; as I will 

 now proceed to show. 



If the head and tail of a comet revolve together as one mass, 

 it must be the centre of gravity of this mass that describes the 

 parabolic orbit ; and moreover, since the tail keeps continually 

 opposite to the sun, this mass must rotate about its centre of grav- 

 ity, at the same rate that the centre revolves around the sun. If 

 the tail be conceived to fall somewhat back of the line of the ra- 

 dius-vector prolonged, as it does in point of fact, then, it may 

 chance that the sun's repulsive actions upon the varying parts of 

 the comet may give a resultant, having its line of direction so 

 situated, behind the centre of gravity, as to tend to produce the 

 rotation required. If we could find the situation of this resultant, 

 for any assumed inclination of the tail to the radius-vector, pro- 

 duced beyond the orbit, as well as the expression for the moment 

 of inertia of the whole mass divided by the mass, then, as the 

 angular velocity of rotation would be known at each point of the 

 orbit, being the same as that of revolution, we might, by a well 

 known formula of mechanics, easily compute the velocity of 

 translation that would be given to the centre of gravity of the 

 rotating mass by a force of sufficient intensity to produce the 

 rotation. Afterwards it would be easy to find whether this 



