949]
on halphen’s characteristic n.
469
In fact, starting from an existing curve, say the complete intersection of two
given surfaces of the orders ¡ju, v respectively, we have, as is known,
d = /jLv, 2h = fiv (/jb — 1) (v — 1):
we find also, as will appear, n = (/j, — 1) (v — 1): hence also nd = 2h, viz. the cone of
the order n through the h nodal lines meets the cone of the order d in these lines
counting each twice, and in no other lines. I remark also that h is = \n (n + 3) if
¡i + v = 4 ; viz. we have here two nodal lines lying in a plane; but if /a + v > 4, then
h is greater than \n (n + 3), viz. in this case the nodal lines are not h lines taken
at pleasure, but they are lines subject to the condition of lying in a cone of the
order n. But this is not all; the h nodal lines lie not only in this cone of lowest
order n, but also in cones of the orders n+1, n+ 2, ..., n + (y, + v — 2) respectively:
I do not for the moment attempt to determine the number of independent conditions
which are hereby imposed upon the h lines.
The last-mentioned theorem constitutes in fact the geometrical interpretation of
results contained in Jacobi’s paper “ De eliminatione variabilis e duabus aequationibus
algebraicis,” Grelle, t. xv. (1836), pp. 101—124, [Ges. Werke, t. ili. pp. 295—320].
Consider the two equations
U = {*§%, y, z, wY = 0 ; V = (*$«, y, z, w) v = 0;
representing surfaces of the orders y, v respectively; since the form of the equations
is quite arbitrary, we may without loss of generality assume that the point
(.x, y, z) = (0, 0, 0)
is an arbitrary point in space; and this being so, we find the equation of the cone,
vertex this point, which passes through the curve of intersection of the two surfaces
by the mere elimination of w between the two equations. As the reasoning is exactly
the same for a particular case, I write for convenience y = 3, v = 4, and consider the
two equations
AqU? + A-iVfi -)- A^w + A 3 = 0,
I? 0 w 4 + B Y w z + B 2 w 2 + B 3 w + B 4 = 0,
where the suffixes show the degrees in regard to (x, y, z), viz. A 0 , B 0 are mere
constants, A 1} B } are linear functions (*$/», y, z) 1 , A.,, B 2 quadric functions (*$#, y, zf,
and so on. Multiplying the first equation successively by 1, w, w 2 , w z , and the second
equation successively by 1, w, w 2 , we have 7 equations from which to eliminate 1,
w, w 2 , w z , w 4 , w 5 , w 6 , and the result is
Ao,
A,
A,
A 3
A,
A u
A,
A,
^-Oj
Aj,
A 2 ,
A 3 ,
A 1}
A 2)
A 3 ,
Bo,
B,,
B 2 ,
B 3 ,
B 4
B 0 ,
B 1}
B 2 ,
B 3 ,
B 4 ,
B 1}
B 2 ,
B 3 ,
Bi,