Proof
../
Throughout, q is an odd prime power, μ∈Fq is a non-square
and, for t,u∈Fq,
P(t)=(t,t2,0),C(u)=(0,μu2,u),A={P(t)}∪{C(u)}.
Size. P(t)=C(u) forces t=0 and u=0, so the two parabolas share
only the origin P(0)=C(0)=0 and ∣A∣=2q−1.
A pair lemma. For s,r∈Fq the conditions t+t′=s and
t2+t′2=r determine the unordered pair {t,t′}: they give
2tt′=s2−r, so t and t′ are the two roots of
X2−sX+(s2−r)/2, and 2 is invertible.
Strong Sidon. Every pair of elements of A, doubles included, has a
sum of one of the forms
P(t)+P(t′)=(t+t′,t2+t′2,0),C(u)+C(u′)=(0,μ(u2+u′2),u+u′),P(t)+C(u)=(t,t2+μu2,u),
where a pair containing the origin may be read in more than one form. Let
σ=(σ1,σ2,σ3) be such a sum. We show that σ
determines the pair.
- If σ1=0 and σ3=0, the pair is not inside P
(third coordinate 0) and not inside C (first coordinate 0). So it
is {P(t),C(u)} with t=σ1 and u=σ3.
- If σ3=0, a pair meeting C∖{0} in some C(u) must
be {C(u),C(−u)}, since P(t)+C(u) has third coordinate u=0.
Otherwise the pair lies in P and the pair lemma recovers it from
(σ1,σ2). A pair {C(u),C(−u)} is recovered from
σ2=2μu2. The two cases cannot share a sum: that would need
t+t′=0 and 2t2=2μu2 with u=0, so μ=(t/u)2 would be a
square.
- If σ1=0 and σ3=0, a pair meeting P∖{0} in
some P(t) would have first coordinate t=0 (with a point of C) or
third coordinate 0 (with a point of P). So the pair lies in C, and
the pair lemma applied to u+u′=σ3, u2+u′2=σ2/μ
recovers it.
The cases are exhaustive and disjoint, so distinct pairs have distinct
sums.
Maximality. Let x=(x1,x2,x3)∈/A. We give a,b,c∈A with
c∈/{a,b} and x=a+b−c.
- x1=0. Put σ=(x2−μx32)/x1, t=(σ+x1)/2 and
w=(σ−x1)/2, and take a=C(x3), b=P(t), c=P(w). Then
a+b−c=(t−w, μx32+(t−w)(t+w), x3)=x. Here c=b because
t−w=x1=0, and c=a would force w=x3=0 and x=P(t)∈A.
- x1=0 and x3=0. Put σ=x2/(μx3), u=(σ+x3)/2
and v=(σ−x3)/2, and take a=P(0), b=C(u), c=C(v). Then
a+b−c=(0, μ(u−v)(u+v), u−v)=x. Here c=b because u=v,
and c=a=0 would force v=0 and x=C(u)∈A.
- x1=x3=0, so x2=0. Put t′=−x2/2 and take a=P(1),
b=P(t′), c=P(1+t′). Then a+b−c=(0, −2t′, 0)=x. Here c=a would
need t′=0, and c=b would need 1=0.
In each case a+b=x+c, and the pairs {a,b} and {x,c} differ
because x∈/A. Hence A∪{x} is not strong Sidon, and A is
maximal. ■
Remarks. The midpoint condition 2x∈A+A is never needed. The proof
uses only that 2 is invertible and that μ is not a square, so the
same set works over every finite field of odd characteristic, and through
Fq3≅(Z/p)3k it gives maximal Sidon sets of size
2pk−1 in elementary abelian p-groups of rank 3k.