330 RemarJcs on the Tails of Halley^s Comet. 



been received here) fixes the period of rotation or (as he is dispo- 

 sed to regard it) of oscillation, at 4/^ days, which is nearly the 

 same. Not knowing the hour of M. Bessel's observation, (mine 

 was 7, P. M.) nor the exact angle of the cone as estimated by him 

 on the 2d, I make no correction for difference of hour or of longi- 

 tude. My estimate of the angle on the 16th, may not have been 

 so exact as to render this important.* Indeed, M. Bessel found 

 that from the 2d to the 25th, the discrepancy between his observa- 

 tions and his oscillatory hypothesis was considerable, " but scarcely 

 exceeded, except for a single day, the limits of uncertainty of which 

 observations of this kind are susceptible."! He found the discrep- 

 ancy still greater with the hypothesis " of an uniform motion of re- 

 volution of the axis of the sector around the straight line drawn 

 from the comet to the sun. "J I regret that his important memoir is 

 not at present accessible, as it might afford a reason which does not 

 occur to me, for the assumption of this particular position of the 

 axis of rotation. Perhaps a different position might satisfy the con- 

 ditions imposed by the observations. If we assume the fact of a 

 rotation, then there are originally three unknown elements, viz. its 

 period, the direction of its axis, and the actual angle between it and 

 the axis of the cone. The first element may be determined by the 

 position and angular opening of the cone at different epochs ; either 

 of the latter two may be variously assumed, and their possible com- 

 binations are infinitely numerous. The actual and complete revo- 

 lutions observed by Mr. Dunlop in the short tails of the comet of 

 1825, <§> seem to favor the hypothesis of rotation in the nucleus of 

 Halley. But either this or the oscillatory hypothesis of M. Bes- 

 sel, will if established, (and it is scarcely possible to reject both,) 

 justify the foregoing distinction in regard to the two classes of com- 

 etary tails. Those of one class assume different and variable posi- 

 tions with respect to the radius vector of the comet, and are pro- 

 jected from certain invariable parts of the nucleus, either in conse- 

 quence of the volatilization of matter from an unequally volatile 

 surface, or of its expulsion through orifices by the agency of an in- 

 terior explosive force, or they are formed by some auroral action in 



* This however being an angle of 90°, could be much more exactly estimated 

 by the eye than the other angles. 

 1-Bib. Univ. p. 359. ?Id. 



§See plates and description in Edinburgh Journal, Vol. VI, p. 84. 



