Some properties of the distribution of the numbers of points on elliptic curves over a finite prime field
Saiying He, James Mc Laughlin

TL;DR
This paper investigates the distribution of the number of points on elliptic curves over finite prime fields, providing explicit evaluations of sums involving the trace of Frobenius for various parameters and congruence classes.
Contribution
It offers new explicit formulas for sums of Frobenius traces on elliptic curves over finite fields, using elementary exponential sum techniques and quadratic form results.
Findings
Derived formulas for sums of Frobenius traces over finite fields.
Evaluated sums for specific congruence classes of primes.
Provided an explicit cubic sum formula for primes congruent to 5 mod 6.
Abstract
Let be a prime and for , let denote the elliptic curve over with equation . As usual define the trace of Frobenius by \begin{equation*} \#E_{a,b}(\mathbb{F}_{p}) = p+1 -a_{p,\,a,\,b}. \end{equation*} We use elementary facts about exponential sums and known results about binary quadratic forms over finite fields to evaluate the sums , , , and for primes in various congruence classes. As an example of our results, we prove the following: Let mod 6 be prime and let . Then \begin{equation*} \sum_{t=0}^{p-1}a_{p,\,t,\,b}^{3}=β¦
| 5 | 7 | 9 | 11 | |
| 5 | -275 | -2315 | -20195 | -179195 |
| 11 | -10901 | -358061 | -12030821 | -411625181 |
| 17 | -36737 | -1582913 | -68613377 | -3016710593 |
| 23 | 8257 | 2763745 | 304822657 | 27903893665 |
| 29 | -35699 | -396299 | 184745341 | 35260018501 |
| 41 | -654401 | -88683041 | -12260782721 | -1716248660321 |
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Some properties of the distribution of the numbers of
points on elliptic curves over a finite prime field
Saiying He
Trinity College
300 Summit Street, Hartford, CT 06106-3100
Β andΒ
J. Mc Laughlin
Mathematics Department
Trinity College
300 Summit Street, Hartford, CT 06106-3100
(Date: August 13th, 2004)
Abstract.
Let be a prime and for , let denote the elliptic curve over with equation . As usual define the trace of Frobenius by
[TABLE]
We use elementary facts about exponential sums and known results about binary quadratic forms over finite fields to evaluate the sums , , , and for primes in various congruence classes.
As an example of our results, we prove the following: Let mod 6 be prime and let . Then
[TABLE]
Key words and phrases:
Elliptic Curves, Finite Fields, Exponential sums
1991 Mathematics Subject Classification:
11G20, 11T23
1. Introduction
Let be a prime and let be the finite field of elements. For , let denote the elliptic curve over with equation . Denote by the set of β rational points on the curve and define the trace of Frobenius, , by the equation
[TABLE]
A simple counting argument makes it clear that
[TABLE]
where denotes the Legendre symbol. We recall some of the arithmetic properties of the distribution of . The following theorem is due to Hasse [4]:
Theorem 1**.**
The integer satisfies
[TABLE]
Since we wish to look at how varies as the coefficients and of the elliptic curve vary, it is convenient for our purposes to write for the elliptic curve as . The following result is well known (an easy consequence of the remarks on page 36 of [3], for example).
Proposition 1**.**
Let the function be defined by setting
[TABLE]
Then for each integer ,
[TABLE]
The following result can be found in [2] (page 57).
Proposition 2**.**
Define the function by setting
[TABLE]
Then for each integer ,
[TABLE]
The following result is also known ([3], page 37, for example).
Proposition 3**.**
Let be a quadratic non-residue modulo . Then
[TABLE]
To better understand the distribution of the it makes sense to study the moments. The j -invariant of the elliptic curve is defined by
[TABLE]
provided . Michel showed in [7] that if is a one-parameter family of elliptic curves with and polynomials in such that
[TABLE]
is non-constant, then
[TABLE]
In [2] Birch defined
[TABLE]
for integral , and proved
Theorem 2**.**
111 In [2], Birch incorrectly omitted the factor of in his statement of Theorem 2.
For ,
[TABLE]
where is Ramanujanβs -function.
Theorem 2 evaluates sums of the form in terms of and these results were derived by Birch as consequences of the Selberg trace formula .
In this present paper we instead use elementary facts about exponential sums and known results about binary quadratic forms over finite fields to evaluate the sums , , , and , for primes in particular congruence classes. In particular, we prove the following theorems.
Theorem 3**.**
Let be a prime, and . Then
(i) ,
(ii)**
This result is elementary but we prove it for the sake of completeness.
Theorem 4**.**
Let mod 6 be prime and let . Then
[TABLE]
Theorem 5**.**
Let be prime and let . Then
[TABLE]
Theorem 4 and Theorem 5 could be deduced from Theorem 2, but we believe it is of interest to give elementary proofs that do not use the Selberg trace formula.
Theorem 6**.**
Let mod 6 be prime and let . Then
[TABLE]
2. Proof of the Theorems
We introduce some standard notation. Define , so that
[TABLE]
Define
[TABLE]
Lemma 1**.**
Let denote the Legendre symbol, modulo . Then
[TABLE]
Proof.
See [1], Theorem 1.1.5 and Theorem 1.5.2. β
We will occasionally use the fact that if is a subset of ,
[TABLE]
We will also occasionally make use of some implications of the Law of Quadratic Reciprocity (see [5], page 53, for example).
Theorem 7**.**
*Let and be odd primes. Then
a)$$\left(\frac{-1}{p}\right)=(-1)^{(p-1)/2}.
b .
c .*
We now prove Theorems 3, 4, 5 and 6,
Theorem 3. Let be a prime, and . Then
i ,
ii *
Proof. (i) From (1.1) and (2.3), it follows that
[TABLE]
The inner sum over is zero unless , in which case it equals to . The left side therefore can be simplified to give
[TABLE]
The last equality follows from (2.3).
ii: From (1.1) and (2.3), it follows that
[TABLE]
The inner sum over is equal to [math], by (2.1), since .
The result at (ii) follows also, in the case of primes (mod 4), from the fact that . However, this is not the case for primes (mod 4). For example,
[TABLE]
The results in Theorem 3 are almost certainly known, although we have not been able to find a reference.
Theorem 4. Let mod 6 be prime and let . Then
[TABLE]
Proof. From (1.1) and (2.3) it follows that
[TABLE]
The inner sum over is zero, unless , in which case it equals . Thus
[TABLE]
Since the map is one-to-one on , when , the in the inner sum can be replaced by . Thus the inner sum is zero unless , in which case it equals . It follows that
[TABLE]
We have once again used (2.3) to compute the sums, noting that the sums above start with . The result now follows since for all primes .
Remarks: (1) It is clear that the results will remain true if is replaced by any function which is one-to-one on .
(2) It is more difficult to determine the values taken by for primes . This is principally because the map is not one-to-one on for these primes (so that (2.1) cannot be used so easily to simplify the summation), but also because the answer depends on which coset belongs to in .
Before proving the next theorem, it is necessary to recall a result about quadratic forms over finite fields. Let be a power of an odd prime and let denote the quadratic character on (so that if , an odd prime, then , the Legendre symbol). The function is defined on by
[TABLE]
Suppose
[TABLE]
is a quadratic form over , with associated matrix and let denote the determinant of ( is non-degenerate if ).
Theorem 8**.**
Let be a non-degenerate quadratic form over , odd, in an even number of indeterminates. Then for the number of solutions of the equation in is
[TABLE]
Proof.
See [6], pp 282β293. β
Theorem 5. Let be prime and let . Then
[TABLE]
Proof. Once again (1.1) and (2.3) give that
[TABLE]
The inner sum over is zero, unless , in which case it equals . Thus
[TABLE]
We have used the fact that for all primes . The inner sum over equals , unless one of the factors , equals [math], in which case the sum is . The equation has solutions and, by (2.6) with , , and , the equation has
[TABLE]
solutions. However, we need to be careful to avoid double counting and to examine when has a solution with . The equation will have two solutions if and none if . Hence the number of solutions to the equation is . Thus the number of solutions to is
[TABLE]
Thus
[TABLE]
The right side now simplifies to give the result.
Before proving Theorem 6, we need some preliminary lemmas.
Lemma 2**.**
Let mod 6 be prime. Then
[TABLE]
Proof.
If , the left side of (2.8) becomes
[TABLE]
The second equality follows since for the primes being considered, the third equality follows from (2.1) and the last equality follows from (2.4).
If , then the left side of (2.8) equals
[TABLE]
Now replace by and then by (justified by the same argument as above) and finally by to get this last sum equals
[TABLE]
We wish to extend the last sum to include . If we set on the right side of the last equation (and denote the resulting sum by βr.s.β) and sum over and we get that
[TABLE]
Replace by in the second sum above and then
[TABLE]
It follows that the left side of (2.9) equals
[TABLE]
and thus that the left side of (2.8) equals
[TABLE]
The second equality in (2.10) follows upon replacing by and then by . β
Lemma 3**.**
Let mod 6 be prime. Then
[TABLE]
where
[TABLE]
Proof.
Upon changing the order of summation slightly, we get that
[TABLE]
If , the inner double sum over and is zero, unless , if which case it equals and the right side of (2.11) equals
[TABLE]
By similar reasoning, if , the right side of (2.11) also equals
[TABLE]
Thus
[TABLE]
where the last equality follows upon replacing by and by . The inner sum over and is zero unless
[TABLE]
in which case the inner sum is . We distinguish the cases and . If , then necessarily and the sum on the right side of (2.12) becomes
[TABLE]
If then
[TABLE]
and since , we exclude and . After substituting for in the sum in the final expression in (2.12), we find that
[TABLE]
where
[TABLE]
β
Lemma 4**.**
Let mod 6 be prime and let be as defined in Lemma 3. Then
[TABLE]
Proof.
Clearly we can remove the restrictions , and freely. If we set , we have that
[TABLE]
The last equality follows from (2.4). Thus
[TABLE]
If is set equal to 0 in the sum above we get
[TABLE]
If is set equal to -1 in this sum we get
[TABLE]
Thus
[TABLE]
If we set in this latest sum we get
[TABLE]
If we set in this sum we get
[TABLE]
Thus
[TABLE]
β
Lemma 5**.**
Let mod be prime. Then
[TABLE]
Proof.
If is replaced by and then is replaced by , the value of the double sum above does not change. Thus
[TABLE]
We evaluate the inner sum using (2.3).
[TABLE]
The next-to-last equality follows since the sum over in the previous expression is 0, unless , in which case this sum is . The sum over equals if and equals otherwise. Hence the sum at (2.16) equals
[TABLE]
the last equality following from the remark after (2.7). β
Corollary 1**.**
Let and be as defined in Lemma 3. Then
[TABLE]
Proof.
Lemmas 4 and 5 and the fact that if give (i). Lemma 3 and part (i) give (ii). β
Theorem 6. Let mod 6 be prime and let . Then
[TABLE]
Proof. Let be a generator of . It is a simple matter to show, using (1.1), that
[TABLE]
Thus the statement at (2.17) is equivalent to the statement
[TABLE]
Let denote the left side of (2.18). From (1.1) and (2.3) it follows that
[TABLE]
The inner sum is zero, unless in , in which case it equals . Upon letting , replacing by and by , we get that
[TABLE]
which was what needed to be shown, by (2.18). The second equality above follows from Lemma 2. Above is as defined in Lemma 3 and in the next-to-last equality we used Corollary 1, part (ii). In the last equality we used once again the fact that .
3. Concluding Remarks
Let mod 6 be prime, and be an odd positive integer. Define
[TABLE]
(It is not difficult to show that the right side is independent of )
By Theorem 6
[TABLE]
We have not been able to determine for (We do not consider even , since a formula for each even can be derived from Birchβs work in [2]). We conclude with a table of values of and small primes mod 6, with the hope of encouraging others to work on this problem.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1] Berndt, Bruce C.; Evans, Ronald J.; Williams, Kenneth S. Gauss and Jacobi sums . Canadian Mathematical Society Series of Monographs and Advanced Texts. A Wiley-Interscience Publication. John Wiley & Sons, Inc., New York, 1998. xii+583 pp.
- 2[2] Birch, B. J. How the number of points of an elliptic curve over a fixed prime field varies. J. London Math. Soc. 43 1968 57β60.
- 3[3] Blake, I. F.; Seroussi, G.; Smart, N. P. Elliptic curves in cryptography . Reprint of the 1999 original. London Mathematical Society Lecture Note Series, 265. Cambridge University Press, Cambridge, 2000. xvi+204 pp.
- 4[4] Hasse, H. Beweis des Analogons der Riemannschen Vermutung fΓΌr die Artinschen und F. K. Schmidtschen Kongruenzzetafunktionen in gewissen eliptischen FΓ€llen . VorlΓ€ufige Mitteilung, Nachr. Ges. Wiss. GΓΆttingen I, Math.-phys. Kl. Fachgr. I Math. Nr. 42 (1933), 253-262
- 5[5] Ireland, Kenneth; Rosen, Michael A classical introduction to modern number theory . Second edition. Graduate Texts in Mathematics, 84. Springer-Verlag, New York, 1990. xiv+389 pp.
- 6[6] Lidl, Rudolf; Niederreiter, Harald Finite fields . With a foreword by P. M. Cohn. Second edition. Encyclopedia of Mathematics and its Applications, 20. Cambridge University Press, Cambridge, 1997. xiv+755 pp.
- 7[7] Michel, Philippe Rang moyen de familles de courbes elliptiques et lois de Sato-Tate. Monatsh. Math. 120 (1995), no. 2, 127β136.
- 8[8] Miller, Steven J. One- and two-level densities for rational families of elliptic curves: evidence for the underlying group symmetries. Compos. Math. 140 (2004), no. 4, 952β992
