Quadratic forms vs congruences

T. D. Noe noe at sspectra.com
Thu Apr 24 23:31:07 CEST 2008


similar to the  "15 theorem", "33 theorem", or "290 theorem" (see A030050-1
Return-Path: <wilson.d at anseri.com>
X-Ids: 165
From: "David W. Wilson" <wilson.d at anseri.com>
To: <seqfan at ext.jussieu.fr>
References: <200804231422.m3NEMSCQ015379 at prim.research.att.com>
In-Reply-To: <200804231422.m3NEMSCQ015379 at prim.research.att.com>
Subject: RE: 
Date: Thu, 24 Apr 2008 17:43:28 -0400
Message-ID: <000301c8a654$3be73150$b3b593f0$@d at anseri.com>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Office Outlook 12.0
Thread-Index: AcilTZRZDBCWF75VQnu6+Fh39WZcDQA8h2uQ
Content-Language: en-us
X-Scanned-By: MIMEDefang 2.36
X-Greylist: IP, sender and recipient auto-whitelisted, not delayed by milter-greylist-3.0 (shiva.jussieu.fr [134.157.0.165]); Thu, 24 Apr 2008 23:47:18 +0200 (CEST)
X-Virus-Scanned: ClamAV 0.92/6926/Thu Apr 24 20:23:40 2008 on shiva.jussieu.fr
X-Virus-Status: Clean
X-Miltered: at jchkmail.jussieu.fr with ID 4810FFE5.004 by Joe's j-chkmail (http : // j-chkmail dot ensmp dot fr)!
X-j-chkmail-Enveloppe: 4810FFE5.004/64.140.224.12/zorn.worldpath.net/zorn.worldpath.net/<wilson.d at anseri.com>
X-j-chkmail-Score: MSGID : 4810FFE5.004 on jchkmail.jussieu.fr : j-chkmail score : XX : R=. U=. O=# B=0.313 -> S=0.357
X-j-chkmail-Status: Unsure


Theorem: Let p be prime, x >= 0, y >= 0.  p = 4x^2-4xy+6y^2 <=> p == 7 (mod
24).

Part I:  p = 4x^2-4xy+6y^2 => p == 7 (mod 24).

Let p = 4x^2-4xy+6y^2.

Then p = a^2+6b^2 with a = 2x-y, b = y.

Thus p = (a+sqrt(-6)b)(a-sqrt(-6)b) is reducible in Q(sqrt(-6)).

Number theory tells us that rational (integer) prime p is reducible in a
quadratic field exactly when p has one of a fixed set of residues modulo the
determinant of the field (crossed fingers here). The determinant of
Q(sqrt(-6)) is 24, and we find empirically that p is reducible when p == 1
or 7 (mod 24). In summary,

    [1] p = a^2+6b^2  <=>  p == 1 or 7 (mod 24).

We can show that 2 is not of form a^2+6b^2, so p is odd => a is odd => y is
odd => b is odd. a and b odd rules out a^2+6b^2 == 1 (mod 24), so p == 7
(mod 24). []

Part II: p == 7 (mod 24) => p = 4x^2-4xy+6y^2

Let p == 7 (mod 24).

[1] gives p = a^2+6b^2.

p != 2, so p is odd => a is odd.

If b is even, then p = a^2+6b^2 != 7 (mod 24) so b is odd.

Since a^2+6b^2 = |a|^2+6|b|^2, so we can assume a >= 0, b >= 0.

a and b odd >= 0 gives integer (a+b)/2 >= 0.

Hence x = (a+b)/2, y = b are positive integers with p = 4x^2-4xy+6y^2. []

QED.









More information about the SeqFan mailing list