574 Mr. Arthur Stanley Eddington [March 26, 



to see through the twisting streamers and changing features to the 

 few simple forces that govern it all. There are, however, a few signs 

 of regularity in the photographs to which it seems most hopeful first 

 to turn. I would therefore draw your attention to the envelopes. 

 It is very difficult to reproduce these clearly on lantern slides, 

 although they are plain enough on the original negatives. The 

 envelopes are wreaths or veils thrown out towards the sun and flowing 

 away on each side. They are not like the streamers from the nucleus, 

 for they seem quite detached, forming an arch over the head. The 

 mode of formation may be illustrated by a well-known analogy. If 

 you have a fountain consisting of a large number of jets of water 

 in different directions, the limiting surface is a sort of dome in the 

 form of a paraboloid, which, when seen sideways, exactly imitates 

 the envelope of a comet. It is not merely a bounding surface 

 beyond which none of the water is projected ; the arch is thickened 

 along this surface. When the water is turned on fuller, the arch 

 rises ; if it is turned off gradually it sinks, but if it is turned off 

 suddenly the arch does not subside, but vanishes ; the water of 

 course subsides, but the thickening vanishes. 



It can hardly be doubted that the envelopes of a comet are 

 formed in this way ; the explosion, from which the envelope results, 

 throws out matter with fairly uniform speed in all directions, this 

 matter being under the influence of solar repulsion, just as in the 

 analogous case the water was under gravitation. By studying them 

 we can learn something of the explosions that produce them ; further 

 in them we are concerned with the general mass of fine particles, so 

 that the study of the rather exceptional knots and luminous patches is 

 supplemented ; and finally in them we have to deal with the repul- 

 sion of the particles, very shortly after they are projected, which is 

 of special importance in the light of the recent evidence that the 

 repulsion may cease to act. 



The best defined and most regular envelopes on the Greenwich 

 plates are those of October 27 ; the envelopes approach the parabohc 

 form so closely as to confirm our hypothesis as to their formation, 

 and to indicate that any disturbing forces are small. I give below 

 some measures of two of the envelopes shown on that night on plates 

 taken at various times. The first column shows the height of the 

 arch deduced from measures made at the apex, that is to say, by direct 

 measurement ; the second column from measures made of the direction 

 of the envelope near the ends of the latus rectum, and therefore de- 

 duced indirectly. We have thus two independent determinations of 

 the height. 



The figures (p. 575) show very characteristically the transitory 

 nature of the envelopes and of the explosions. The large one, for 

 instance, formed at about 8 h. ;30 m. (it was hardly formed in the first 

 photograph), and in the space of two hours subsided from its original 

 height of 70,000 miles to 40,000 miles ; that is the typical behaviour 



