I'm interested in measuring the distribution of an ordered set of points relative to the Zariski topology. Possibly this is a standard idea (with different terminology?) in algebraic geometry. Any pointers to the literature would be appreciated. Here's what I have in mind.

Let $X\subset\textbf{P}^N$ be a projective variety defined over an algebraically closed field $k$. For any finite set of points of $X$, define the $X$-degree of the set to be the minimal degree of a homogeneous polynomial $F\in k[X_0,\ldots,X_n]$ that does not vanish identically on $X$, but vanishes at all of the points in the set. (In other words, the minimal degree of a hypersurface that goes through the points, but does not contain $X$.)

Now take a sequence of points $S=(P_1,P_2,P_3,\ldots)\subset X$, and for each $n\ge1$, let $S[n]=\{P_1,P_2,\ldots,P_n\}$. Intuitively, if $\deg_X S[n]$ grows quickly, then the set $S$ is well distributed relative to the Zariski topology. A rough guess is that one might define equidistribution by the condition $$ \lim_{n\to\infty} \frac{\log\deg_XS[n]}{\log n} = \frac{1}{\dim X}. $$ (It is not hard to check that the limsup of the left-hand side is at most $\frac{1}{\dim X}$.)

[Edit: value of limit fixed as suggested by JSE]


1 Answer 1


First of all, I think the limit above would like to be 1/(dim X) rather than 1/N -- given d^N points on X, there will indeed be a degree-d polynomial passing through them, but this polynomial might well vanish on all of X (e.g. consider the case where X is a line in P^N -- in order to pass through d points on the line and not vanish on the line, your polynomial had better have degree d.)

I would say this is closer to a weak notion of "general position" than of "equidistribution" -- for points on the plane, aren't you asking for some coarse version of "not too many of the first few points on a line, not too many of the first few points on a conic, ..... etc?"

Anyway, as for where this kind of thing appears in the literature, maybe the argument of Heath-Brown about uniform bounds for points on curves? (Huayi Chen has some appealing-looking recent papers extending this approach to more general varieties.) You could describe these results as saying that if you have a long list of rational points whose height grows very slowly, they DO tend to line up on unexpectedly low-degree hypersurfaces; so in your context, this would be the opposite of the usual situation in which slow growth of height enforces equidistribution!

  • $\begingroup$ @JSE: Thanks, Jordan. I'll have to think about the limit, probably my "back of the envelope" calculation was completely wrong. I agree this is a notion of general position, but maybe it's reasonable to say that "uniformly in general position" is a good way to define "equidistributed in the Zariski topology". (I don't know of another def for the latter.) Good thought about Heath-Brown et.al., although in principle, the points in S need not have small height, nor indeed do they have to be algebraic. But in my application, they are algebraic, and maybe they have bounded height. $\endgroup$ Jan 22, 2011 at 3:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.