566
NOTICES OF COMMUNICATIONS TO THE
[555
where a is the order, A the class of the curve; a is the number, three times the
class + the number of cusps, or (what is the same thing) three times the order -t- the
number of inflexions.
2. On a Correspondence of Points and Lines in Sjmce. Report, 1870, p. 10.
Nine points in a plane may be the intersection of two (and therefore of an
infinite series of) cubic curves; say, that the nine points are an “ ennead ”: and
similarly nine lines through a point may be the intersection of two (and therefore of
an infinite series of) cubic cones; say, the nine lines are an ennead. Now, imagine
(in space) any 8 given points; taking a variable point P, and joining this with the
8 points, we have through P 8 lines, and there ig through P a ninth line completing
the ennead; this is said to be the corresponding line of P. We have thus to any
point P a single corresponding line through the point P; this is the correspondence
referred to in the heading, and which I would suggest as an interesting subject of
investigation to geometers. Observe that, considering the whole system of points in
space, the corresponding lines are a triple system of lines, not the -whole system of
lines in space. It is thus, not any line whatever, but only a line of the triple system,
which has on it a corresponding point. But as to this some explanation is necessary ;
for starting with an arbitrary line, and taking upon it a point P, it would seem
that P might be so determined that the given line and the lines from P to the
eight points should form an ennead,—that is, that the arbitrary line would have upon
it a corresponding point or points.
The question of the foregoing species of correspondence was suggested to me by
the consideration of a system of 10 points, such that joining any one whatever of
them with the remaining nine points, the nine lines thus obtained form an ennead; or,
say, that each of the 10 points is the “ enneadic centre ” of the remaining nine. I
have been led to such a system of 10 points by my researches upon Quartic Surfaces;
but I do not as yet understand the theory.
3. On the Number of Covariants of a Binary Quantic. Report, 1871, pp. 9, 10.
The author remarked [see 462] that it had been shown by Prof. Gordan that the
number of the covariants of a binary quantic of any order was finite, and, in particular,
that the numbers for the quintic and the sextic were 23 and 26 respectively. But the
demonstration is a very complicated one, and it can scarcely be doubted that a more
simple demonstration will be found. The question in its most simple form is as
follows: viz. instead of the covariants we substitute their leading coefficients, each of
which is a “ seminvariant ” satisfying a certain partial differential equation; say, the
quantic is (a, b, c,..., k\x, y) n , then the differential equation is (ad b + 2bd c ... + njdk) u = 0,
which qua equation with n +1 variables admits of n independent solutions: for
instance, if n= 3, the equation is (ad b + 2bd c + Scd d ) u = 0, and the solutions are a, ac — b 2 ,.