Abstract
On the circle of radius R centred at the origin, consider a “thin” sector about the fixed line \(y = \alpha x\) with edges given by the lines \(y = (\alpha \pm \epsilon ) x\), where \(\epsilon = \epsilon _R \rightarrow 0\) as \( R \rightarrow \infty \). We establish an asymptotic count for \(S_{\alpha }(\epsilon ,R)\), the number of integer lattice points lying in such a sector. Our results depend both on the decay rate of \(\epsilon \) and on the rationality/irrationality type of \(\alpha \). In particular, we demonstrate that if \(\alpha \) is Diophantine, then \(S_{\alpha }(\epsilon ,R)\) is asymptotic to the area of the sector, so long as \(\epsilon R^{t} \rightarrow \infty \) for some \( t<2 \).
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
The Gauss circle problem is the problem of determining how many integer lattice points lie inside a circle, centred at the origin, with radius \(R\rightarrow \infty \). This classical problem dates back to Gauss, who employed a simple geometric argument to show that the number of such lattice points is equal to the area of the circle, up to an error term of size \(E\left( R\right) \le 2\sqrt{2}\pi R.\) In 1906, Sierpiński [8] improved the bound on the error term to \(E\left( R\right) =O(R^{2/3})\), and further incremental improvements have been subsequently made throughout the years. The current state-of-the-art bound, due to Bourgain and Watt [1], is that \(E\left( R\right) =O\left( R^{t+\varepsilon }\right) \) for any \(\varepsilon >0\), where \(t=517/824 \approx 0.6274\). It is famously conjectured that \(E\left( R\right) =O(R^{1/2+\varepsilon })\), for any \(\varepsilon >0\).
A natural related problem is to determine the number of lattice points S(R) inside a sector \(\text {Sect}(R)\) of a circle with radius \(R\rightarrow \infty \). For sectors with fixed open angle, Gauss’s argument can be easily extended to show that
where \(E\left( R\right) =O\left( R\right) .\) Nowak [7] (who, more generally, considered sectors in domains of the form \(\left\{ x^{n}+y^{n}\le R^{n}: x,y\ge 0 \right\} \) for any \(n\ge 2)\) showed that the error term can be improved when the slopes of the sector’s two respective edges are either rational or irrational of finite type (see Definition 1.2). Specifically, when both slopes are Diophantine (i.e. of type \(\eta =1+\varepsilon \) for any \( \varepsilon >0 \)), we have \( E(R)=O(R^{2/3-\delta }) \) for a certain (small) \( \delta >0 \). Under a suitable assumption on the irrationality type of the edges’ slopes, these results were further extended by Kuba [4] tso segments of even more general domains. An additional closely related problem, dating back to the work of Hardy and Littlewood [2, 3], concerns the number of lattice points in right-angled triangles. An asymptotic formula for this count − which plays an important role in the proofs of [4] and [7] − is obtained by applying Koksma’s inequality together with standard discrepancy estimates (see, e.g. [5, Theorem 3.2, p. 123 and Theorem 5.1, p. 143]).
In this paper we are interested in counting the number of lattice points, \(S_{\alpha }\left( \epsilon ,R\right) \), lying within a sector whose open angle shrinks as \(R\rightarrow \infty \). More explicitly, we consider a sector \(\text {Sect}_{\alpha ,\epsilon }\left( R\right) \) about the fixed line \(y=\alpha x\) with edges given by the lines \(y=\left( \alpha \pm \epsilon \right) x\), where now \(\epsilon =\epsilon _{R}\rightarrow 0\) as \(R\rightarrow \infty \). Our main goal is to establish an asymptotic formula for \(S_{\alpha }\left( \epsilon ,R\right) \) rather than to optimize the relevant error term. In contrast to the case of fixed sectors, our results depend only on the rationality/irrationality type of \(\alpha \), and not on the rationality/irrationality type of \(\alpha \pm \epsilon \), the slopes of the two edges. For this reason, the results of [4] and [7] are not applicable for our problem, and our argument proceeds in quite a different direction.
If \(\epsilon \rightarrow 0\) at a rate slower than 1/R, then upon applying a geometric argument similar to that used in the Gauss circle problem, we find that \(S_{\alpha }\left( \epsilon ,R\right) \sim \text {Area}\left( \text {Sect}_{\alpha ,\epsilon }(R)\right) \) (see Theorem 1.1 below). To obtain an asymptotic count for more quickly shrinking sectors, we must apply an alternative method. First, we approximate \(S_{\alpha }\left( \epsilon ,R\right) \) by \(\Delta _{\alpha }(\epsilon ,R)\), the number of lattice point lying within a thin triangle whose two long edges lie along the lines \(y=\left( \alpha \pm \epsilon \right) x\). We then fix a rational number \(p/q \in {{\mathbb {Q}}}\) that well-approximates \(\alpha \), and compute \(\Delta _{\alpha }(\epsilon ,R)\) by summing the contributions from lattice points lying on a discrete collection of lines, each of which has rational slope p/q.
When \(\alpha \in {{\mathbb {Q}}}\) is rational, we obtain an asymptotic for \(S_{\alpha }\left( \epsilon ,R\right) \), regardless of how fast our sectors shrink. This is due to the fact that, in such a case, all the lattice points in \(\text {Sect}_{\alpha ,\epsilon }(R)\) lie precisely on the line \( y=\alpha x\) once \(\text {Sect}_{\alpha ,\epsilon }(R)\) is sufficiently thin. If \( \alpha \) is irrational of finite type \(\eta \), we obtain an asymptotic for \(S_{\alpha }\left( \epsilon ,R\right) \) under the assumption that \(\epsilon \) decays at a rate slower than \( 1/R^{1+1/\eta }\) (Theorem 1.3 below). Specifically, when \(\alpha \in {{\mathbb {R}}}\) is Diophantine, we obtain an asymptotic for \(S_{\alpha }\left( \epsilon ,R\right) \) under the assumption that \(\epsilon \rightarrow 0\) at a rate slower than \( 1/R^t\) for some \( t<2 \).
The behaviour of lattice points in even faster shrinking sectors about irrational slopes is a more subtle question. If \(\epsilon \) decays at a rate \(1/R^{1+1/\eta }\) or faster, the above method fails to produce an asymptotic count for \(S_{\alpha }\left( \epsilon ,R\right) \). However, if \(\epsilon \) shrinks sufficiently quickly, then the count once again becomes much simpler. Specifically, when \(\epsilon \) decays faster than \(1/R^{1+\eta }\), we may apply an elementary argument to prove that for sufficiently large R, \(\text {Sect}_{\alpha ,\epsilon }(R)\) contains no lattice points whatsoever (Proposition 1.7). A related question concerns the distribution of lattice points in a randomly chosen sector of width \(\epsilon \asymp 1/R^2\). This interesting question has been addressed by Marklof and Strömbergsson [6], who successfully applied tools from homogeneous dynamics to prove the existence of a (non-Poissonian) limiting distribution for the number of lattice points in such sectors.
1.1 Notation
Fix \(\alpha \in {{\mathbb {R}}}\), and consider the interval \(I_{\epsilon }(\alpha ):= \left( \alpha - \epsilon ,\alpha +\epsilon \right) \), for some \(\epsilon > 0\). Let
denote the sector of radius R with edges given by the lines \(y=\left( \alpha \pm \epsilon \right) x\), which has an open angle of size
In what follows, we view \(\epsilon = \epsilon _R \) as a function of R. Our main interest will be in thin sectors, i.e. when \( \epsilon \rightarrow 0 \) as \( R \rightarrow \infty \). Taylor expanding about \(\alpha \), we find that as \( \epsilon \rightarrow 0 \), the area of \(\text {Sect}_{\alpha , \epsilon }(R)\) is equal to
Let
count the number of integer lattice points in \(\text {Sect}_{\alpha ,\epsilon }(R)\).
We are interested in the value of \(S_{\alpha }(\epsilon ,R)\) in the limit as \(R \rightarrow \infty \). For example, we may consider the case \(\epsilon := R^{-\lambda }\) for some fixed \(\lambda \ge 0\). We then classify our sectors based on the decay rate of \(\epsilon \).
Remark
Our results may be easily extended to more general sectors about the line \( y=\alpha x\). In particular, we note that Theorem 1.3 continues to hold when counting lattice points in any sector of the form
where, say, \(\epsilon _{1} \asymp \epsilon _{2} \asymp \epsilon \). Consequently, one may alternatively consider a sector centred about the angle \(\Phi :=\tan ^{-1}(\alpha )\) with radius R and open angle \(\theta \asymp \epsilon \); and express the resulting lattice point count in terms of the properties of \(\tan {\Phi }\) and the decay rate of \(\theta \) without any alterations to Theorem 1.3. Nonetheless, we have chosen to formulate our results in terms of slopes, rather than angles, in order to simplify the exposition, and because our analysis naturally depends upon the Diophantine properties of the slope \(\alpha \).
1.2 “Slowly” shrinking sectors
Suppose first that \( \epsilon \) is either fixed or decays slower than 1/R, in the sense that \(\epsilon R \rightarrow \infty \) in the limit as \(R \rightarrow \infty \) (e.g. \(0 \le \lambda < 1\)). Upon refining the elementary geometric argument of the O(R) bound for the error term in the Gauss circle problem, we obtain the following result, which yields an asymptotic count for the number of lattice points in such slowly shrinking sectors:
Theorem 1.1
Fix \(\alpha \in \mathbb {R}\), and assume that \( \epsilon R \rightarrow \infty \) as \( R \rightarrow \infty \). Then
1.3 “Quickly" shrinking sectors
In our investigation of more quickly shrinking sectors, our results depend heavily upon the rationality/irrationality type of \(\alpha \), defined as follows:
Definition 1.2
We say that an irrational \(\alpha \in {{\mathbb {R}}}\) is of finite type \(\eta \), if there exists a constant \(c=c(\alpha )>0\) such that
for all integers pairs \((p,q) \in {{\mathbb {Z}}}\times {{\mathbb {Z}}}_{>0}\).
Note that for irrational \(\alpha \in {{\mathbb {R}}}\) of type \(\eta \) we necessarily have \(\eta \ge 1\) by Dirichlet’s theorem. We say that \(\alpha \) is Diophantine if \(\alpha \in {{\mathbb {R}}}\) is irrational of type \(\eta = 1+\varepsilon \) for every \(\varepsilon > 0\). It is well-known that almost all \(\alpha \in {{\mathbb {R}}}\) are Diophantine (Khinchin’s theorem), and every algebraic number is Diophantine (Roth’s theorem).
1.3.1 Irrational slopes
For irrational \(\alpha \in {{\mathbb {R}}}\), our main result is as follows:
Theorem 1.3
Let \(\alpha \in {{\mathbb {R}}}\) be irrational of finite type \(\eta \), and assume that \( \epsilon \rightarrow 0 \), as well as that \( \epsilon R^{1+1/\eta } \rightarrow \infty \) as \( R\rightarrow \infty \). Then
in the limit as \(R \rightarrow \infty \).
The conditions \( \epsilon \rightarrow 0 \) and \( \epsilon R^{1+1/\eta } \rightarrow \infty \) (e.g. \( 0< \lambda < 1+1/\eta )\) consequently guarantee the asymptotic
In particular, if \(\alpha \) is Diophantine, then (1.2) holds whenever \( \epsilon \rightarrow 0 \) and \(\epsilon R^{t} \rightarrow \infty \) for some \( t<2 \) (e.g. \(0< \lambda < 2\)). Note furthermore that \(\text {Area}(\text {Sect}_{\alpha , \epsilon }(R))\) grows if and only if \(\epsilon R^{2} \rightarrow \infty \), and thus our results in such a case are essentially optimal (and “strictly" so whenever \(\alpha \in {{\mathbb {R}}}\) is a badly approximable irrational, i.e. irrational of type \(\eta = 1\)).
Note that Theorem 1.3 gives a better error term than (1.1) whenever \( \epsilon = o(R^{-1/2}) \) and \( \epsilon R \rightarrow \infty \) (e.g. \( 1/2<\lambda <1 \)). Upon comparing the error terms in Theorem 1.1 and Theorem 1.3 we obtain the following corollary:
Corollary 1.4
Let \(\alpha \in {{\mathbb {R}}}\) be irrational of finite type \(\eta \), and let \( \epsilon = R^{-\lambda } \). Then in the limit as \(R \rightarrow \infty \),
In particular, when \(\alpha \in {{\mathbb {R}}}\) is Diophantine, Corollary 1.4 yields
1.3.2 Rational slopes
For rational \(\alpha \in {{\mathbb {Q}}}\), we obtain the following result:
Theorem 1.5
Fix \(\alpha = p/q \in {{\mathbb {Q}}}\), where \(q>0\) and \((p,q)=1\). Then in the limit as \(R \rightarrow \infty \), we have
where \(\{x\}:=x - \lfloor x \rfloor \) denotes the fractional part of x.
When \(\epsilon = o(R^{-1})\) (e.g. \(\lambda > 1\)), Theorem 1.5 simplifies to
In this case, \(S_{\alpha }(\epsilon ,R)\) is no longer asymptotic to \(\text {Area}(\text {Sect}_{\alpha , \epsilon }(R))\), and the only points that contribute to \(S_{\alpha }(\epsilon ,R)\) are those which lie precisely on the line \(y = \alpha x\).
When \(\epsilon \rightarrow 0\) and \(\epsilon R \rightarrow \infty \) (e.g. \(0< \lambda < 1\)), Theorem 1.5 yields
where
is a bounded function of R. In particular, as in the case of irrational slopes, if \(\epsilon = o(R^{-1/2})\) and \(\epsilon R \rightarrow \infty \) (e.g. \(1/2< \lambda < 1\)), then (1.5) yields a more precise count than (1.1). The following corollary summarizes the above analysis in the case \( \epsilon = R^{-\lambda } \):
Corollary 1.6
Let \(\alpha = p/q \in {{\mathbb {Q}}}\), where \(q>0\) and \((p,q)=1\), and let \( \epsilon = R^{-\lambda } \). Then in the limit as \(R \rightarrow \infty \), we have
Finally, we consider the case \(\epsilon \asymp R^{-1}\) (e.g. \(\epsilon R = c\), for some \(c \in {{\mathbb {R}}}_{> 0})\). Then Theorem 1.5 yields
where
In particular, whenever \(\epsilon < \frac{\sqrt{p^{2}+q^{2}}}{q^{2} R}\), the only points which contribute to \(S_{\alpha }(\epsilon ,R)\) are those which lie precisely on the line \(y = \alpha x\), and we find that
We moreover note that \(S_{\alpha }(\epsilon ,R)\) is asymptotic to \(\text {Area}(\text {Sect}_{\alpha , \epsilon }(R))\) if and only if \(\gamma = \frac{\epsilon q^{2}R}{p^{2}+q^{2}}\), i.e. if and only if \( \epsilon \) is an integer multiple of \(\frac{\sqrt{p^{2}+q^{2}}}{q^{2} R}\).
1.4 “Very quickly” shrinking sectors
While in the range \(R^{-1-\eta } \ll \epsilon \ll R^{-(1+1/\eta )}\) we are unable to obtain an asymptotic formula for \(S_{\alpha }(\epsilon ,R)\), for sectors that shrink even more quickly the situation becomes rather trivial. Specifically, whenever \(\epsilon = o(R^{-1-\eta }) \) (e.g. \(\lambda >1+ \eta \)), we show that \(S_{\alpha }(\epsilon ,R) =0\) for sufficiently large R:
Proposition 1.7
Let \(\alpha \in {{\mathbb {R}}}\) be irrational of finite type \(\eta \), and suppose that \(\epsilon = o(R^{-1-\eta }) \). Then there exists \(R_{0} > 0\) such that for all \(R > R_{0}\),
In particular, if \(\alpha \) is a Diophantine irrational, then for sufficiently large R, \(S_{\alpha }(\epsilon ,R) = 0\) whenever \(\epsilon = o(R^{-t})\) for some \( t>2 \) (e.g. \(\lambda > 2\)).
1.5 Structure of paper
The remainder of this paper is structured as follows. In Sect. 2 we apply a simple geometric argument to compute \(S_{\alpha }(\epsilon ,R)\) in the case that \(\epsilon \rightarrow 0\) at a rate slower than 1/R. In Sect. 3 we approximate \(S_{\alpha }(\epsilon ,R)\) by \(\Delta _{\alpha }(\epsilon ,R)\), i.e. by the number of lattice points in a triangle whose two long edges lie along the lines \(y=\left( \alpha \pm \epsilon \right) x\). In Sect. 4 we then apply this approximation to compute \(S_{\alpha }(\epsilon ,R)\) when \(\alpha \in {{\mathbb {R}}}\) is irrational of finite type; and in Sect. 5 we address the case when \(\alpha \in {{\mathbb {Q}}}\) is rational. Finally, in Sect. 6, we address the case in which \(\text {Sect}_{\alpha ,\epsilon }(R)\) shrinks “very quickly", i.e. when \(\epsilon \rightarrow 0\) at a rate faster than \(1/R^{1+\eta }\).
2 Lattice points in slowly shrinking sectors
In this section we provide a proof of Theorem 1.1, namely a count for \(S_{\alpha }(\epsilon ,R)\) when \(\epsilon R \rightarrow \infty \) as \(R \rightarrow \infty \). The proof is an easy adaptation of the elementary geometric argument applied in the classical Gauss circle problem. As evidenced by the proof, this argument remains valid for slowly shrinking sectors.
Proof of Theorem 1.1
For each \(z \in {{\mathbb {Z}}}^{2} \cap \text {Sect}_{\alpha ,\epsilon }(R)\), let \(\square _{z}\) denote a square-box of unit area, centred at the point z. Then
i.e. \(S_{\alpha }(\epsilon ,R)\) is equal to the area formed by the union of such boxes. Note, moreover, that if \(w \in \square _{z}\) for some \(z \in {{\mathbb {Z}}}^{2} \cap \text {Sect}_{\alpha ,\epsilon }(R)\), then
i.e. the distance between w and \(\text {Sect}_{\alpha ,\epsilon }(R)\) is bounded by \(\sqrt{2}/2\). We therefore define a wider sector, \(\text {Sect}^{+}_{\alpha ,\epsilon }(R')\), with the same open angle and direction as \(\text {Sect}_{\alpha ,\epsilon }(R)\), but extended by a distance of \(\sqrt{2}/2\) on all sides, so that
To construct \(\text {Sect}^{+}_{\alpha ,\epsilon }(R')\) explicitly, we expand \(\text {Sect}_{\alpha ,\epsilon }(R)\) by drawing parallel lines distanced \(d=\sqrt{2}/2\) away from each of its two respective straight edges. Let x denote the distance between their point of intersection and the origin. Note that
from which we obtain
We therefore set the radius of our desired sector, \(\text {Sect}^{+}_{\alpha ,\epsilon }(R')\), to be equal to
which yields
upon noting that \( \theta \asymp \epsilon \), so that \( \theta ^{-1}=o(R) \). Thus
To obtain a lower bound for \(S_{\alpha }(\epsilon ,R)\), we similarly construct a sector, denoted by \(\text {Sect}^{-}_{\alpha ,\epsilon }(R'')\), with the same open angle and direction as \(\text {Sect}_{\alpha ,\epsilon }(R)\), but now shrunk by a distance of \(\sqrt{2}/2\) on all sides, of radius
which we note is clearly possible since \( \theta ^{-1} = o(R) \). Any point \(w \in \text {Sect}^{-}_{\alpha ,\epsilon }(R'')\) is within a distance of at most \(\sqrt{2}/2\) from some lattice point z, which, by construction, must lie in \(\text {Sect}_{\alpha ,\epsilon }(R)\). It follows that
Using a similar analysis to that above, we find that
and therefore
Combining (2.1) and (2.2) we conclude that
as desired. \(\square \)
3 Approximating sectors by triangles
In this section we approximate \(S_{\alpha }(\epsilon ,R)\) by considering lattice points in a triangle, namely the summation
We have the following lemma:
Lemma 3.1
Assume that \(\epsilon \rightarrow 0\). Then
In particular, if \(\epsilon = O(R^{-1})\), then
Proof
Assume \(\alpha > 0\), as the proof for the cases \(\alpha = 0\) and \(\alpha < 0\) follow similarly. Suppose \((m,n) \in S_{\alpha }(\epsilon ,R)\). Then \(m^{2}+n^{2}\le R^{2}\) and \(n> m(\alpha -\epsilon ) > 0\) (which holds for sufficiently small \( \epsilon \)) together imply
i.e. that
We may therefore write
where
and
Let us first estimate the size of \(S^{2}_{\alpha }(\epsilon ,R)\). Note that if \(m^{2}+n^{2}> R^{2}\) and \(m(\alpha -\epsilon )< n < m(\alpha +\epsilon )\), then \(m^{2}(1+(\alpha +\epsilon )^{2})>R^{2}\), and therefore \(m > R/\sqrt{1+(\alpha +\epsilon )^{2}}\). Moreover, since the length of the interval \((m(\alpha -\epsilon ),m(\alpha +\epsilon ))\) is \(2\,m \epsilon \le 2R \epsilon \), we find that, for any \(m \in {{\mathbb {N}}}\), there exist at most \(O(1+R\epsilon )\) integers \(n \in {{\mathbb {Z}}}\) such that
Thus
Note furthermore that
It follows that
Hence
from which we obtain that
Next, we wish to show that
Indeed, note that
and that each summand in (3.1) is O\((1+R\epsilon )\). It follows that
as desired. \(\square \)
4 Sectors about irrational slopes
In this section we provide a proof of Theorem 1.3, namely a count for \(S_{\alpha }(\epsilon ,R)\) when \(\alpha \in {{\mathbb {R}}}\) is irrational of finite type.
Let \(\alpha \in {{\mathbb {R}}}\) be irrational. For any rational \(p/q \in {{\mathbb {Q}}}\), we define \(\delta :=\alpha - p/q\). For the purposes of this proof, we will moreover assume that \(|\delta |< \epsilon /2\), which, in particular, implies that \( \delta -\epsilon <0 \) and \( \epsilon +\delta >0 \). We then write
Let \(d= n q -m p\), so that
Together with the conditions on m, this implies that
When \(d > 0\), the condition on m is equivalent to
while when \(d < 0\), the condition is then
Partitioning with respect to d, we then write
with
where \( {\bar{p}} \) denotes the inverse of p modulo q. Upon recalling that
we see that
Similarly, we compute
Finally, we note that
It then follows from (4.1) that
4.1 Choosing an appropriate convergent
Suppose \(\alpha \in {{\mathbb {R}}}\) is irrational of finite type \(\eta \), and let \(\{p_{i}/q_{i}\}_{i=1}^{\infty }\) denote the sequence of convergents to the continued fraction of \(\alpha \). Upon choosing an appropriate pair \(\{p_{i}/q_{i}\}\), we are able to proceed with a proof of Theorem 1.3:
Proof of Theorem 1.3
For any \(X:= X(R)\), there exists a unique i such that \(q_{i} \le X < q_{i+1}\). There moreover exists a \(c=c(\alpha )> 0\) such that
Hence
which further implies that
In other words, there exists a constant \(C > 0\) such that
Let \( p=p_i\) and \(q=q_i.\) By (4.4), it follows that
To ensure that \(|\delta | < \epsilon /2\), we choose X such that
namely, we subject X to the restriction
To optimize our error term, we seek a choice of X, subject to the restriction (4.7), which minimizes the value of
Note first that by (4.5) and (4.6),
which in turn implies that
and therefore that
Similarly, since (4.8) implies \(q^{-1} \le \epsilon X/2,\) we find that
Next, since \(q \le X\), it follows that
and finally it similarly follows from (4.8) that
By (4.9), (4.10), (4.11), and (4.12), we have that
We thus choose the minimal possible value for X, namely \( X \asymp \epsilon ^{-\frac{1}{1+1/\eta }}\), which is moreover o(R) by the assumption that \( \epsilon R^{1+1/\eta } \rightarrow \infty \). In particular,
By (4.3) and (4.13), we conclude that
Theorem 1.3 now follows directly from (4.14) and Lemma 3.1. \(\square \)
5 Sectors about rational slopes
In this section we provide a proof of Theorem 1.5, namely a count for \(S_{\alpha }(\epsilon ,R)\) when \(\alpha \in {{\mathbb {Q}}}\). The proof proceeds similarly to that of Theorem 1.3, upon setting \(\delta = 0\):
Proof of Theorem 1.5
Recall that \(\alpha = p/q\), where \((p,q)=1\). Note that
Let \(d = nq -mp\), and note that \(|d|< mq\epsilon \) implies
as well as that
Partitioning with respect to d, we then write
where \( {\bar{p}} \) denotes the inverse of p modulo q (in particular, if \(\epsilon q^{2} R/\sqrt{p^{2}+q^{2}}< 1\), then the only contribution to \(\Delta _{\alpha }(\epsilon ,R)\) comes from the term \(d=0\), i.e. points \((m,n) \in {{\mathbb {Z}}}^{2}\) lying precisely on the line \(y = \alpha x\)). Upon setting
and recalling that
it follows that
We furthermore note that
and similarly that
Combining the above expressions we see that
and the desired result now follows from Lemma 3.1. \(\square \)
6 Very quickly shrinking sectors
Finally, in this section we provide a proof of Proposition 1.7, namely that when \(\alpha \in {{\mathbb {R}}}\) is irrational of finite type \(\eta \) and \(\epsilon = o(R^{-1-\eta })\), we find that \(S_{\alpha }(\epsilon , R)=0\) for sufficiently large R:
Proof of Proposition 1.7
Since \( \alpha \) is of finite type \( \eta \), there exists a constant \(c=c(\alpha )>0\) such that for all \((p,q) \in {{\mathbb {Z}}}\times {{\mathbb {Z}}}_{>0}\),
Take \(R_{0}\) sufficiently large such that for any \( R>R_0 \) we have
Then for any \(R > R_{0}\), and any \((p,q) \in {{\mathbb {Z}}}\times {{\mathbb {Z}}}_{>0}\) with \(0 < q \le R\), we find that
It follows that for all \(R > R_{0}\) we have \(S_{\alpha }(\epsilon ,R) = 0\), as desired. \(\square \)
Data availability
Data sharing is not applicable to this article as no datasets were generated or analysed during the current study.
References
Bourgain, J., Watt, N.: Decoupling for perturbed cones and the mean square of \(|\zeta (\frac{1}{2}+it)|\). Int. Math. Res. Not. IMRN 17, 5219–5296 (2018)
Hardy, G.H., Littlewood, J.E.: Some problems of diophantine approximation: the lattice-points of a right-angled triangle. Proc. London Math. Soc. 2(1), 15–6 (1921)
Hardy, G. H., Littlewood, J. E. Some problems of Diophantine approximation: The lattice-points of a right-angled triangle. (Second memoir.). Abh. Math. Sem. Univ. Hamburg 1 (1922), no. 1, 211-248
Kuba, G.: On the number of lattice points in certain planar segments. Math. Slovaca 53(2), 173–187 (2003)
Kuipers, L., Niederreiter, H.: Uniform Distribution of Sequences Pure Appl. Math. Wiley-Intersci. Publ, John Wiley and Sons, New York-London-Sydney-Toronto (1974)
Marklof, J., Strömbergsson, A.: The distribution of free path lengths in the periodic Lorentz gas and related lattice point problems. Ann. of Math. 2(3), 1949–2033 (2010)
Nowak, W.G.: Über die Anzahl der Gitterpunkte in verallgemeinerten Kreissektoren. Monatsh. Math. 87(4), 297–307 (1979)
Sierpiński, W.: O pewnem zagadnieniu z rachunku funkcyj asymptotycznych. Prace Mat. Fiz. 17, 77–118 (1906)
Acknowledgements
We thank Zeév Rudnick and Andreas Strömbergsson for helpful discussions and comments, and the anonymous referee for useful comments and a careful read of the manuscript. This research was supported by the ISRAEL SCIENCE FOUNDATION (Grant No. 1881/20), and the first author was funded by a Zuckerman Post Doctoral Fellowship.
Funding
Open access funding provided by University of Haifa.
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by Tim Browning.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Waxman, E., Yesha, N. On the number of lattice points in thin sectors. Monatsh Math (2024). https://doi.org/10.1007/s00605-024-01983-x
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s00605-024-01983-x