Algebraic Curves Lecture 23

The 23rd (probably?) lecture of algebraic curves by Karl Christ
  1. More extremal curves

More extremal curves

We return to the case that d>2rd > 2r. Assume that CPrC\subseteq \mathbb P^r is extremal, i.e.

  1. non-degenerate

  2. g(C)=π(d,r)g(C) = \pi(d,r), where π(d,r)\pi(d,r) is the bound from Castelnuova's bound.

Set HH to be a general hyperplane and let D=CHD = C\cap H as always. Then

α1=h0(C,OC(D))=r+1\begin{aligned} \alpha_1 = h^0(C, \mathcal O_C(D)) = r + 1 \end{aligned}

and on the level of quadrics

α2=h0(C,OC(2D))=(r+1)+(2r1)=3r.\begin{aligned} \alpha_2 = h^0(C, \mathcal O_C(2D)) = (r+1) + (2r - 1) = 3r. \end{aligned}

So for an extremal curve, I exactly know the dimension of the restriction of quadrics to CC: we've got the map

φ:H0(Pr,OPr(2))H0(C,OC(2))\begin{aligned} \varphi:H^0(\mathbb P^r, \mathcal O_{\mathbb P^r}(2)) \twoheadrightarrow H^0(C, \mathcal O_C(2)) \end{aligned}

where the domain has dimension (r+22)\binom{r+2}{2} and the codomain has dimension 3r3r, so

dim(kerφ)=(r+2)(r+1)23r=r23r+22=(r12).\begin{aligned} \operatorname{dim}(\operatorname{ker}\varphi) = \frac{(r+2)(r+1)}{2}-3r = \frac{r^2 - 3r + 2}{2} = \binom{r-1}{2}. \end{aligned}

This is the number of quadrics containing a rational normal scroll.

New questions: What is ρ=Qi\rho = \bigcap Q_i, where the QiQ_i are the (r12)\binom{r-1}{2} linearly independent quadrics containing CC?

Pick a general hyperplane HH and consider D=CHD = C\cap H. Then D=d2r+1|D| = d \geq 2r + 1. The points in DD are in linearly general position and impose 2r12r - 1 conditions on quadrics. By Castelnuova's 2r+32r + 3 lemma, these is a rational normal curve CDHC_D \subseteq H containing DD.

Claim: CDSC_D\subseteq S. Suppose there is QiQ_i such that CD⊈QiC_D \not\subseteq Q_i. Then

CD.Qi=2(r1)d=deg(D)\begin{aligned} C_D . Q_i = 2\cdot (r - 1) \geq d = \deg (D) \end{aligned}

because DCDQiD\subseteq C_D\cap Q_i. The count above is simply Bezout's theorem. This is a contradiction however, since d>2rd > 2r, so we can't have 2(r1)d2\cdot (r - 1) \geq d. Thus CD⊈QiC_D\not\subseteq Q_i implies that CdQiC_d\cap Q_i has dimension 00.

Conversely: CDC_D is cut out by (r12)\binom{r - 1}{2} quadrics in HPr1H\cong \mathbb P^{r-1} which implies SH=CDS\cap H = C_D. A rational normal curve in Pr\mathbb P^r is cut out by the 2×22\times 2 minors of a matrix of the form

(x0x1x2xr1x1x2x3xr),\begin{aligned} \begin{pmatrix} x_0 & x_1 & x_2 & \dots & x_{r-1} \\ x_1 & x_2 & x_3 & \dots & x_r \end{pmatrix}, \end{aligned}

of which there are (r2)\binom{r}{2}. Because CDC_D is a rational normal curve in HPr1H\cong \mathbb P^{r-1}, it is therefore cut out by (r12)\binom{r-1}{2} quadrics. We know already that CDC_D vanishes on the QiQ_i, of which there are (r12)\binom{r-1}{2}, so there are no degrees of freedom left and hence SHCDS\cap H \supset C_D. This gives us both inclusions and hence SH=CDS\cap H = C_D.

Summary: We get that Qi=S\bigcap Q_i = S gives a rational normal curve of degree r1r-1 when we intersect with a general hyperplane:

CD:=SH.\begin{aligned} C_D := S\cap H. \end{aligned}

This implies SS is a surface of degree r1r-1, and hence that CC lies on a rational normal scroll or on P2P5\mathbb P^2 \to \mathbb P^5 (the Veronese embedding [x:y:z][x2:y2:z2:xy:xz:yz:][x:y:z] \mapsto [x^2:y^2:z^2:xy:xz:yz:]).


Suppose S=Xa1,a2S = X_{a_1, a_2} with a10a_1\neq 0 is a smooth, rational normal scroll. We saw (or claimed without proof? Can't remember) last time that Pic(S)=L,H\operatorname{Pic}(S) = \langle L, H\rangle with L.L=0L.L = 0, L.H=1L.H = 1 and H.H=r1H.H = r-1.

Recall if you have a curve CC with bidegree (d1,d2)(d_1,d_2) then 2g2=KS.C+C.C2g - 2 = K_S. C + C.C and C=αL+βHC = \alpha L + \beta H. Then

KS=K1L+K2H.\begin{aligned} K_S = K_1\cdot L + K_2\cdot H. \end{aligned}

From LL you have

2=KSL+L2=K2\begin{aligned} -2 = K_S\cdot L + L^2 = K_2 \end{aligned}

and for HH

2=KSH+H2=K1+K2(r1)+(r1)=K1(r1).\begin{aligned} -2 = K_S H + H^2 = K_1 + K_2(r-1) + (r-1) = K_1 - (r-1). \end{aligned}

This implies K1=r3K_1 = r-3. Hence KS=(r3)L2HK_S = (r-3)L - 2H, 2g(C)2=KS.C+C.C2g(C) - 2 = K_S.C + C.C and so

g(C)=(α1)(2mα)2(r1)+ϵ(α1),\begin{aligned} g(C) = \frac{(\alpha - 1)(2m - \alpha)}{2}(r-1) + \epsilon\cdot (\alpha - 1), \end{aligned}

hence d1=m(r1)+ϵd-1 = m(r-1) + \epsilon. So gg is maximal for α=m+1\alpha = m+1, β=ϵ+2r\beta = \epsilon + 2 - r for any ϵ\epsilon or α=m\alpha = m and β=1\beta = 1 and ϵ=0\epsilon = 0. For these values, equality in Castelneuovo's bound is achieved.

Theorem: Let r3r\geq 3 and d2r+1d \geq 2r + 1. Then extremal curves of degree dd exist and are of the following forms:

  1. The image of a smooth degree kk curve in P2\mathbb P^2 under the Veronese embedding P2P5\mathbb P^2 \to \mathbb P^5 (so d=2kd = 2k and r=5r = 5)

  2. A non-singular member of mH+L|mH + L| on a rational normal scroll

  3. A non-singular member of the linear system (m+1)H(rϵ2)L|(m+1)H - (r - \epsilon - 2)L| on a rational normal scroll.

Cor: In case (3), CC admits a gm+11g_{m+1}^1 cut out by LL.
©Isaac Martin. Last modified: April 05, 2024.