128 



On the Formation of the Tails of Comets. 



must be a curved line concave towards the regions of space which 

 the comet has left. Supposing the arc AC to be so small, or its 

 curvature to be so slight that it may be considered as a straight 

 line, and neglecting the change of the velocity in the orbit, Ga 

 will be parallel to AD, and Qb parallel to BE, whence RCa= 

 CSA, and RC6=CSB. Thus the line joining any particle with 



R 



the nucleus always makes an angle with the prolongation of the 

 radius-vector, equal to the motion in anomaly during the interval 

 that has elapsed since the particle left the head. It follows from 

 this that, if we suppose the velocity of the particles to be continu- 

 ally the same, and the motion in anomaly to be uniform, the de- 

 viations of the particles a and b from the line of the radius-vector 

 will be in the ratio of the distances Ca and Cb. But in point of 

 fact, the velocity increases with the distance, so that the curva- 

 ture of the tail will be less than on the supposition just made : 

 and, we may suppose, may after a certain time, come to be so 

 great, compared with the velocity in the orbit, as to make the 

 rest of the tail almost perfectly straight, as the greater part of it is 

 sometimes observed to be. A curvature may afterwards spring 

 up at the extremity in consequence of the nebulous matter being 

 there more retarded by the resistance of the ether which is believed 

 to pervade all space — this resistance having a greater effect than 

 before, because of the diminution of the sun's force, and, perhaps, 

 of a diminution in the density of the cometic matter itself. As 

 to the amount of the deviation of the tail from the line of the radi- 

 us-vector, it must depend upon the proportion between the velo- 

 cities of the particles, and the velocity of the head in its orbit : 



