Remarks on the Tails of Halley’s Comet. 329 
tected a curvature on one side. ‘They will, however, scarcely dis- 
tinguish this emanation from some of those to which the name tail 
has been generally appropriated. 
If we call this cone a tail, it would be desirable to classify these 
appendages. Were there others, like the cone of Halley, traceable 
directly to the nucleus by observation alone, we might denominate 
them caude nuclei, in contradistinction to the ordinary appendages 
of the envelope. But let us see whether there are not physical 
considerations which should induce us to make this class more ex- 
tensive. As some luminous emanations are nearly opposite to the 
sun, and retain a position nearly constant in relation to the radius 
vector, and as others, which are usually much shorter, have different 
and variable positions in relation to the radius vector of the comet, 
but in all probability positions nearly or exactly constant in relation 
to the radius of its nucleus at the point of emanation, we might 
still distinguish the latter as the tails of the nucleus, (although in 
some instances, as in the comet of 1825, they are not directly trace- 
able to it by observation,) and the former as the caudz solares of 
comets, on account of their characteristic position and peculiar mode 
of development. In the former article, [ hazarded a conjecture as 
to the fact of a rotation of the nucleus of Halley, suggested by ob- 
served differences in the position of the cone on different days. 
This interesting fact, (er what for our present purpose is equivalent 
to it) may now be regarded as established by the observations of 
M. Bessel, not only on different days from the 2d to the 25th of 
October, but during 8 or 9 successive hours on the 12th, during 
which there was a rapid progressive motion from right to left. It is 
an interesting circumstance, that the angular opening of the cone, 
(‘about 90°’*) observed at Kénigsberg on the 2d of October, was 
the same as that observed at Schenectady on the 16th. ‘This tends 
strongly to confirm the identity of the cone at distant epochs, and 
affords some data for determining the period of rotation, as there 
are reasons for believing that most of the angular changes of the 
intermediate periods were not real, but resulted from variations in 
the obliquity of the axis of the cone in relation to the visual ray. 
If we assume three complete rotations with an equable motion be- 
tween the 2d and 16th of October, the period of each must have 
been 43 days. Now M. Bessel (from data which have not yet 
* Bib. Univ. Avril, 1836, p. 357. 
Vol. aan No, gq. 42 
