200 PROF. O. REYNOLDS ON COMETARY PHENOMENA. 



there might be one or several shells of cloud at certain 

 distances from the central mass. These, under the action 

 of the sun's rays, would be illuminated, and afford that 

 striking appearance of several bands which is so often 

 seen. 



The effect of any repulsive or attractive force in the sun 

 acting on the vapour of the comet, but not on the central 

 mass, would cause a train or tail ; but the direction of this 

 tail would depend on the motion of the comet as well as on 

 the intensity of the force. 



It does not necessarily follow because the tail points 

 away from the sun that the force which repels it must 

 entirely overcome the effect of the sun's attraction ; but it 

 can be shown that even for those comets whose tails show 

 the most curvature the force must be comparable with the 

 sun's gravitation ; and to produce the straight tails which 

 comets sometimes possess, it would require a force of almost 

 infinite intensity, if it acted as gravity docs on ordinary 

 matter. It has been found by Professor Norton that if 

 the bent tail of Donati's comet were composed of ordinary 

 matter, leaving the comet under the action of a repulsive 

 force m the sun, this force must be between 1*5 and '39 

 of the sun's attraction, the matter in the leading edge 

 being repelled by the former, and that in the following by 

 the latter. 



If the force is electrical, its intensity will depend on the 

 charge of electricity in the vapour ; and if there are seven 

 streams of vapour variously charged, these would each form 

 a tail of distinct curvature ; so that the comet might have 

 several bent tails or even a fan — which are features not 

 uncommonly observed. This, then, affords an explanation 

 of the great variety of bent tails often seen with comets 

 and even of those pointing towards the sun ; for some of 

 the vapour might have a charge of the opposite electricity 

 to that in the sun, under which circumstance it would be 



