on the family of diophantine triples { k + 1, 4 k , 9 k + 3}
TRANSCRIPT
Periodica Mathematica Hungarica Vol. 58 (1 ), 2009, pp. 59–70DOI: 10.1007/s10998-009-9059-6
ON THE FAMILYOF DIOPHANTINE TRIPLES {k + 1, 4k, 9k + 3}
Bo He1 and Alain Togbe2
[Communicated by Attila Petho]
1Department of Mathematics
Key Laboratory of Numerical Simulation of Sichuan Province
Neijiang Normal University, Neijiang, Sichuan, 641112, P. R. China
E-mail: [email protected]
2Department of Mathematics, Purdue University North Central
1401 S. U.S. 421, Westville, IN 46391 USA
E-mail: [email protected]
(Received June 24, 2008; Accepted July 28, 2008)
Abstract
We prove that if k is a positive integer and d is a positive integer such
that the product of any two distinct elements of the set {k + 1, 4k, 9k + 3, d}
increased by 1 is a perfect square, then d = 144k3 + 192k2 + 76k + 8.
1. Introduction
A set of m distinct positive integers {a1, . . . , am} is called a Diophantine m-
tuple if aiaj + 1 is a perfect square. Diophantus studied sets of positive rationals
with the same property, particularly he found the set of four positive rationals{
116 , 33
16 , 174 , 105
16
}
. But the first Diophantine quadruple was found by Fermat. In
fact, Fermat proved that the set {1, 3, 8, 120} is a Diophantine quadruple. Moreover
Baker and Davenport [2] proved that the set {1, 3, 8, 120} cannot be extended to a
Diophantine quintuple. In general, let n be an integer. A set of m positive integers
{a1, . . . , am} is called a Diophantine m-tuple with the property D(n) or a D(n)−m-
tuple (or a Pn-set of size m), if aiaj + n is a perfect square. Authors also consider
n as a parametric expression. See for example [19].
The problem of extendability of Pn-sets is of big interest. See for examples
[1], [4]–[25], [27]. Several generalizations of the result of Baker and Davenport
are obtained. In 1997, Dujella [6] proved that the Diophantine triples of the form
Mathematics subject classification numbers: 11D09, 11D45, 11B37, 11J68, 11J86.
Key words and phrases: Diophantine m-tuple, Pell equation, Diophantine approximation,linear forms in logarithms.
0031-5303/2009/$20.00 Akademiai Kiado, Budapestc© Akademiai Kiado, Budapest Springer, Dordrecht
60 B. HE and A. TOGBE
{k − 1, k + 1, 4k}, for k ≥ 2, cannot be extended to a Diophantine quintuple.
The Baker–Davenport result corresponds to k = 2. In 1998, Dujella and Petho
[14] proved that the Diophantine pair {1, 3} cannot be extended to a Diophantine
quintuple. In 2008, Fujita [21] obtained a more general result by proving that the
Diophantine pairs {k − 1, k + 1}, for k ≥ 2 cannot be extended to a Diophantine
quintuple. A folklore conjecture is that there does not exist a Diophantine D(1)-
quintuple. In 2004, Dujella [11] proved that there are only finitely many Diophantine
D(1)-quintuples. A stronger version of this conjecture is the following.
Conjecture 1.1. If {a, b, c, d} is a Diophantine quadruple and d >
max{a, b, c}, then d = a + b + c + 2abc + 2√
(ab + 1)(ac + 1)(bc + 1).
A Diophantine quadruple {a, b, c, d} is regular if and only if (a+ b− c− d)2 =
4(ab + 1)(cd + 1). Therefore the quadruple in the above conjecture is regular.
The aim of this paper is to consider the Diophantine triple {k + 1, 4k, 9k + 3}and to prove the following result.
Theorem 1. If k is a positive integer and d is a positive integer such that
the product of any two distinct elements of the set
{k + 1, 4k, 9k + 3, d}
increased by 1 is a perfect square, then d = 144k3 + 192k2 + 76k + 8.
Remark. One can easily check that our quadruple {k + 1, 4k, 9k + 3, d} with
d = 144k3 + 192k2 + 76k + 8 is regular.
The organization of this paper is as follows. In Section 2, we recall some useful
results obtained by Dujella and adapt them to our case. We use a result due to
Bennett [3] on simultaneous approximations of algebraic numbers which are close
to 1 to get an upper bound for k. Finally, in Section 4, we prove Theorem 1 by
means of linear forms in logarithms and the Baker–Davenport reduction method.
2. Preliminaries
Let r, s, t be positive integers defined by
ab + 1 = r2, ac + 1 = s2, bc + 1 = t2. (1)
In order to extend the Diophantine triple {a, b, c} to a Diophantine quadruple
{a, b, c, d}, we have to solve the system
ad + 1 = x2, bd + 1 = y2, cd + 1 = z2. (2)
ON THE FAMILY OF DIOPHANTINE TRIPLES {k + 1, 4k, 9k + 3} 61
Eliminating d, we obtain the following system of Pellian equations:
az2 − cx2 = a − c, (3)
bz2 − cy2 = b − c. (4)
By [9, Lemma 1], there exists a solution (z(i)0 , x
(i)0 ) of (3) such that z = v
(i)m , where
v(i)0 = z
(i)0 , v
(i)1 = sz
(i)0 + cx
(i)0 , v
(i)m+2 = 2sv
(i)m+1 − v(i)
m ,
and |z(i)0 | <
√
c√
c2√
a. Similarly, there exists a solution (z
(i)1 , y
(i)1 ) of (4) such that
z = w(j)n , where
w(i)0 = z
(j)1 , w
(j)1 = tz
(j)1 + cy
(j)1 , w
(j)n+2 = 2tw
(j)n+1 − w(j)
n ,
and |z(j)1 | <
√
c√
c
2√
b.
The initial terms z(i)0 and z
(j)1 are almost completely determined in the fol-
lowing lemma [11, Lemma 8].
Lemma 1.
(a) If the equation v2m = w2n has a solution, then z0 = z1. Furthermore, |z0| = 1
or |z0| = cr − st or |z0| < min{0.869a−5/14c9/14, 0.972b−0.3c0.7}.(b) If the equation v2m+1 = w2n has a solution, then |z0| = t, |z1| = cr − st and
z0z1 < 0.
(c) If the equation v2m = w2n+1 has a solution, then |z0| = cr − st, |z1| = s and
z0z1 < 0.
(d) If the equation v2m+1 = w2n+1 has a solution, then |z0| = t, |z1| = s and
z0z1 > 0.
In the present paper, we have
a = k + 1, b = 4k, c = 9k + 3,
and from (1) we obtain
r = 2k + 1, s = 3k + 2, t = 6k + 1.
One can easily check that
|z0| <
√
c√
c
2√
a<
√
3(9k + 3)
2≤ 4.3
√k
and
|z1| <
√
c√
c
2√
b<
√
2√
3(9k + 3)
4≤ 3.3
√k.
62 B. HE and A. TOGBE
If (b) or (d) holds, this implies |z0| = t = 6k + 1 < 4.3√
k. Thus this is impossible.
If (c) holds, similarly we obtain |z1| = s = 3k+2 < 3.3√
k, which is also impossible.
Therefore we only consider the case (a).
Since cr − st = 1, then the first and the second possibilities in (a) both imply
|z0| = 1. Let us define d0 = (z20 −1)/c in the third possibility of (a). If |z0| > 1, then
d0 < z20/c < a−5/7c9/7/c < c. Thus according to the proof of the above lemma in
[11], {a, b, c, d0} is an irregular Diophantine quadruple. Also by [11, Proposition 1],
if {a, b, c, d} is an irregular Diophantine quadruple with a < b < c < d, then
d > 2.695c3.5a2.5. Therefore we have c > 2.695 · c3.5a2.5. When k ≥ 1, we get a
contradiction.
In other words, we need to solve the system of Pellian equations
(k + 1)z2 − (9k + 3)x2 = −8k − 2, (5)
4kz2 − (9k + 3)y2 = −5k − 3, (6)
with x0 = y1 = 1 and z0 = z1 = ±1. This is equivalent to solving the equation
z = v2m = w2n. (7)
Let λ := z0 = z1, λ ∈ {−1, 1}. We have
v0 = λ, v1 = (9 + 3λ)k + 3 + 2λ, vm+2 = (6k + 4)vm+1 − vm, (8)
w0 = λ, w1 = (9 + 6λ)k + 3 + λ, wn+2 = (12k + 2)wn+1 − wn. (9)
For the relations of indices m and n, we have
Lemma 2. If v2m = w2n, then n ≤ m ≤ 2n.
Proof. By [9, Lemma 3], if vm = wn, then n − 1 ≤ m ≤ 2n + 1. In our even
case, we have 2n − 1 ≤ 2m ≤ 4n + 1. The result is obtained.
In order to get a gap principle between indices m, n and k, we recall the
following lemma:
Lemma 3. We have
v2m ≡ z0 + 2c(az0m2 + sx0m) (mod 8c2),
w2n ≡ z1 + 2c(bz1n2 + ty1n) (mod 8c2).
Proof. These congruences came from items (a) and (c) of [9, Lemma 4].
ON THE FAMILY OF DIOPHANTINE TRIPLES {k + 1, 4k, 9k + 3} 63
Using these results, we have
λam2 + sm ≡ λbn2 + tn (mod 4c). (10)
In our case, the congruence becomes
λ(k + 1)m2 + (3k + 2)m ≡ 4λkn2 + (6k + 1)n (mod 36k + 12) .
Multiplying the above congruence by 36λ, we obtain 24m2 + 36λm ≡ −48n2 −36λn (mod 36k + 12). Therefore,
2m2 + 4n2 + 3λm + 3λn ≡ 0 (mod 3k + 1). (11)
If m, n ≥ 2, then
2m2 + 4n2 + 3λm + 3λn ≥ 2m2 + 4n2 − 3m − 3n > 0.
Hence from (11) we obtain
2m2 + 4n2 + 3m + 3n ≥ 2m2 + 4n2 + 3λm + 3λn ≥ 3k + 1.
By Lemma 2, we know that m ≥ n. So we have 6m2 + 6m ≥ 3k + 1 > 3k. Thus we
get
Lemma 4. Assume that v2m = w2n with m, n ≥ 2. Then m ≥√
k/2 − 1.
Now we prove the following result.
Lemma 5. Let x, y, z be positive integer solutions of the system of Pellian
equations (5) and (6) such that z /∈ {1, 36k2 + 30k + 5}. Then
log(2z) > (√
2k − 2) log(6k).
Proof. Using (3) with z0 = ±1, x0 = 1 and v2m = w2n, we have
z =1
2√
a
(
(±√
a +√
c)(s +√
ac)2m + (±√
a −√
c)(s −√
ac)2m)
.
For example, one can refer to the proof of [9, Lemma 1]. Since z /∈ {1, 36k2+30k+5},
then m ≥ 2 and n ≥ 2. We get
2z >1√a(±
√a +
√c)(s +
√ac)2m −
√a +
√c√
a(s +√
ac)2m
≥(
− 1 +
√
c
a
)
(2√
ac)2m − 2√
c√a(2
√ac)2m
> (−1 +√
6)(2√
ac)2m − 1
8a2.5c1.5> (2
√ac)2m .
64 B. HE and A. TOGBE
Hence,
log(2z) > 2m log(2√
ac) > 2m log(2√
(k + 1)(9k + 3)) > 2m log(6k).
From Lemma 4, we obtain the result.
3. Application of a result of Bennett
In this section, we will use the following result of Bennett on simultaneous
approximations of algebraic numbers which are close to 1 to get an upper bound
for k in the system of Pellian equations (3) and (4).
Lemma 6. If ai, pi, q and N are integers for 0 ≤ i ≤ 2, with a0 < a1 < a2,
aj = 0 for some 0 ≤ j ≤ 2, q nonzero and N > M9, where
M = max0≤i≤2
{|ai|},
then we have
max0≤i≤2
{∣
∣
∣
∣
√
1 +ai
N− pi
q
∣
∣
∣
∣
}
> (130Nγ)−1q−µ, (12)
where
µ = 1 +log(33Nγ)
log (1.7N2Π0≤i<j≤2(ai − aj)−2)
and
γ =
(a2 − a0)2(a2 − a1)
2
2a2 − a0 − a1if a2 − a1 ≥ a1 − a0,
(a2 − a0)2(a1 − a0)
2
a1 + a2 − 2a0if a2 − a1 < a1 − a0.
Proof. See [3, Lemma 3.2].
Let us consider the numbers
θ1 =
√
1 +2
3k + 1and θ2 =
√
1 − 1
3k + 1.
The following result gives us some information on the approximations of the alge-
braic numbers θ1 and θ2.
Lemma 7. If k ≥ 171 and (x, y, z) 6= (1, 1, 1), then
∣
∣
∣
∣
θ1 − 6x
2z
∣
∣
∣
∣
< 4z−2 and
∣
∣
∣
∣
θ2 − 3y
2z
∣
∣
∣
∣
< z−2.
ON THE FAMILY OF DIOPHANTINE TRIPLES {k + 1, 4k, 9k + 3} 65
Proof. Using (5) and (6), one can verify that
∣
∣
∣
∣
θ1 − 6x
2z
∣
∣
∣
∣
=
√3∣
∣z√
k + 1 − x√
9k + 3∣
∣
z√
3k + 1=
√3(8k + 2)
z√
3k + 1∣
∣z√
k + 1 + x√
9k + 3∣
∣
<
√3(8k + 2)
2z2√
(3k + 1)(k + 1)< 4z−2,
and
∣
∣
∣
∣
θ2 − 3y
2z
∣
∣
∣
∣
=
√3∣
∣
∣z√
4k − y√
9k + 3∣
∣
∣
2z√
3k + 1=
√3(5k + 3)
2z√
3k + 1∣
∣
∣z√
4k + y√
9k + 3∣
∣
∣
<
√3(5k + 3)
4z2√
4k(3k + 1)< z−2.
This completes the proof of the lemma.
The next result gives us the information on d.
Proposition 3.1. If k ≥ 2535 and if the set {k + 1, 4k, 9k + 3, d} is a
Diophantine quadruple, then d has to be 144k3 + 192k2 + 76k + 8.
Proof. If d satisfies the condition, then z2 = cd + 1 = (9k + 3)d + 1. Since
d > 1, we have z 6= 1. And d 6= 144k3 +192k2 +76k+8, we have z 6= 36k2 +30k+5.
Then Lemma 5 implies
log(2z) > (√
2k − 2) log(6k). (13)
Now we apply Lemma 6. We take a0 = −1, a1 = 0, a2 = 2, M = 2, q =
2z, p0 = 3y, p1 = q, p2 = 6x and N = 3k + 1. When k ≥ 171, the condition
N > M9 holds. Therefore, by Lemma 6 and Lemma 7, we have
4z−2 > (130(3k + 1)γ)−1(2z)−µ, (14)
where γ = 7.2 and
µ = 1 +log(33 · 7.2 · (3k + 1))
log(1.7 · (3k + 1)2/36)< 1 +
log(3k + 1) + 5.471
2 log(3k + 1) − 3.052
< 1.5 +3.499
log(3k + 1) − 1.526.
One can check that if k > 1900, then µ < 2. In fact,
2 − µ > 0.5 − 3.499
log(3k + 1) − 1.526.
66 B. HE and A. TOGBE
From (14) we have
(2z)2−µ < 16 · 130 · 7.2(3k + 1).
It follows that
log(2z) < log(14976(3k + 1))/(2 − µ). (15)
Combining (13) and (15), we obtain
√2k − 2 <
log(14976(3k + 1))
(2 − µ) log(6k)<
log(14976(3k + 1))
(0.5 − 3.499log(3k+1)−1.526 ) log(6k)
. (16)
If k ≥ 2535, we get a contradiction.
4. Proof of Theorem 1
In this section, we need to consider the remaining cases, i.e., 1 ≤ k ≤ 2534.
We will use a theorem on lower bounds to linear forms in logarithms to get an upper
bound for m.
Let
α1 = s +√
ac and α2 = t +√
bc.
Solving equations (3) and (4), we have
v2m =1
2√
a
(
(z0
√a + x0
√c)α2m
1 + (z0
√a + x0
√c)α−2m
1
)
and
w2n =1
2√
b
(
(z1
√b + y1
√c)α2n
2 + (z1
√b + y1
√c)α−2n
2
)
respectively. Notice x0 = y1 = 1 and z0 = z1 = λ, λ ∈ {−1, 1}. If z = v2m = w2n
with m, n 6= 0, we have (see [11, formula (60)])
0 < Λ := 2m log α1 − 2n log α2 + log α3 <8ac
3α−4m
1 (17)
where
α3 =
√b(
√c + λ
√a)
√a(
√c + λ
√b)
.
It follows that
log |Λ| < −4m log α1 + log(8ac
3
)
< (2 − 4m) log(2√
ac)
< (2 − 4m) log(3k + 3).(18)
We recall the following result due to Matveev [26].
ON THE FAMILY OF DIOPHANTINE TRIPLES {k + 1, 4k, 9k + 3} 67
Lemma 8. Denote by α1, . . . , αl algebraic numbers, not 0 or 1, by log α1,
. . . , log αl their logarithms, by D the degree over Q of the number field K =
Q(α1, . . . , αl), and by b1, . . . , bn rational integers. Define B = max{|b1|, . . . , |bl|},and Ai = max{Dh(αi), | log αi|, 0.16} (1 ≤ i ≤ l), where h(α) denotes the absolute
logarithmic Weil height of α. Assume that the number
Λ = b1 log α1 + · · · + bn log αl
does not vanish. Then
|Λ| ≥ exp{−C(l, κ)D2A1 · · ·Al log(eD) log(eB)},
where κ = 1 if K ⊂ R and κ = 2 otherwise and
C(l, κ) = min
{
1
κ
(
1
2el
)κ
30l+3l3.5, 26l+20
}
.
Now, we apply the lemma with l = 3 and κ = 1 for
Λ = 2m log α1 − 2n log α2 + log α3. (19)
Here we take D = 4, b1 = 2m, b2 = 2n, b3 = 1. Notice that α1 = s +√
ac is a root
of X2 − 2sX + 1 = 0, α2 = t +√
bc is a root of X2 − 2tX + 1 = 0, and α3 is a root
ofa2(c − b)2x4 + 4a2b(c − b)x3 + 2ab(3ab − ac − bc − c2)x2
+ 4a2b(c − b)x + b2(c − a)2 = 0.
Thus we have
h(α1) =1
2log α1, h(α2) =
1
2log α2,
and
h(α3) =1
4
(
log(a2(c − b)2) + 4 log(
√
c/a + λ√
c/b + λ
)
)
<1
2log(ac) +
√
c/a −√
c/b√
c/b − 1< log(
√
(k + 1)(9k + 3)) + 3
< log(k + 1) + log 3 + 3 < log(k + 1) + 4.1 .
Therefore, we take A1 = 2 log(6k + 4), A2 = 2 log(12k + 2), and A3 = 4 log(k + 1) +
16.4. Using Matveev’s result we have
log |Λ| > −8.492 ·1013 · log(6k + 4) · log(12k + 2) ·(log(k + 1) + 4.1) · log(2em). (20)
Combining (18) and (20), we obtain
m
log(2em)< 2.124 · 1013 · log(6k + 4) · log(12k + 2) · (log(k + 1) + 13.1)
log(3k + 3).
68 B. HE and A. TOGBE
By Proposition 3.1, we only need to consider the cases k < 2535. It follows
m
log(2em)< 4.95 · 1015.
This implies m < 2.1 · 1017.
In order to deal with the remaining cases 1 ≤ k ≤ 2534, we will use a Diophan-
tine approximation algorithm, the so-called Baker–Davenport reduction method.
The following lemma is a slight modification of the original version of the Baker–
Davenport reduction method. (See [14, Lemma 5a]).
Lemma 9. Assume that M is a positive integer. Let P/Q be the convergent
of the continued fraction expansion of κ such that Q > 6M and let
η = ‖µ′Q‖ − M · ‖κQ‖,
where ‖ · ‖ denotes the distance from the nearest integer. If η > 0, then there is no
solution of the inequality
0 < mκ − n + µ′ < AB−m
in integers m and n with
log (AQ/η)
log B≤ m ≤ M.
We apply Lemma 9 with
κ =log α1
log α2, µ′ =
log α3
2 log α2, A =
4(k + 1)(3k + 1)
log α2, B = α4
1
and M = 2.1 · 1017.
In one minute all the computations were done. The use of the second con-
vergent was needed in 29 and 32 cases when λ = 1 and −1 respectively. We have
m ≤ 2 except for 1 ≤ k ≤ 9 and k = 16 for λ = ±1. In all these particular cases,
we obtained m ≤ 5. We took again M = 5 in these 20 cases, then we got m ≤ 2. If
m = 2 is a solution of (7), then by Lemma 2 we have 2 ≤ n ≤ 4 and from Lemma 4
we get k ≤ 18. It is easy to check that there is no integer k satisfy the conditions
in (7). Then m must be 1 (m = n = 0 gives the trivial solution d = 0).
Combining this and Proposition 3.1, we have m = n = 1 in equation (7).
When λ = v0 = w0 = −1, v2 = w2 = 36k2+30k+5 gives d = 144k3+192k2+76k+8.
This completes the proof of Theorem 1.
ON THE FAMILY OF DIOPHANTINE TRIPLES {k + 1, 4k, 9k + 3} 69
Acknowledgments
The authors express their gratitude to the anonymous referee for constructive
suggestions to improve an earlier draft of this paper. The first author is supported
by the Natural Science Foundation of China (No. 10872085). The second author is
partially supported by Purdue University North Central.
References
[1] J. ARKIN, V. E. HOGGATT and E. G. STRAUSS, On Euler’s solution of a problem of
Diophantus, Fibonacci Quart., 17 (1979), 333–339.
[2] A. BAKER and H. DAVENPORT, The equations 3x2 − 2 = y2 and 8x2 − 7 = z2,
Quart. J. Math. Oxford Ser. (2), 20 (1969), 129–137.
[3] M. A. BENNETT, On the number of solutions of simultaneous Pell equations, J.
Reine Angew. Math., 498 (1998), 173–199.
[4] Y. BUGEAUD, A. DUJELLA and M. MIGNOTTE, On the family of Diophantine triples
{k − 1, k + 1, 16k3 − 4k}, Glasgow Math. J., 49 (2007), 333–344.
[5] A. DUJELLA, The problem of Diophantus and Davenport for Gaussian integers, Glas.
Mat. Ser. III, 32 (1997), 1–10.
[6] A. DUJELLA, The problem of the extension of a parametric family of Diophantine
triples, Publ. Math. Debrecen, 51 (1997), 311–322.
[7] A. DUJELLA, A proof of the Hoggatt–Bergum conjecture, Proc. Amer. Math. Soc.,
127 (1999), 1999–2005.
[8] A. DUJELLA, Diophantine m-tuples and elliptic curves, J. Theor. Nombres Bor-
deaux, 13 (2001), 111–124.
[9] A. DUJELLA, An absolute bound for the size of Diophantine m-tuples, J. Number
Theory, 89 (2001), 126–150.
[10] A. DUJELLA, On the size of Diophantine m-tuples, Math. Proc. Cambridge Philos.
Soc., 132 (2002), 23–33.
[11] A. DUJELLA, There are only finitely many Diophantine quintuples, J. Reine Angew.
Math., 566 (2004), 183–214.
[12] A. DUJELLA, A. FILIPIN and C. FUCHS, Effective solution of the D(−1)-quadruple
conjecture, Acta Arith., 128 (2007), 319–338.
[13] A. DUJELLA and C. FUCHS, Complete solution of a problem of Diophantus and Euler,
J. London Math. Soc., 71 (2005), 33–52.
[14] A. DUJELLA and A. PETHO, A generalization of a theorem of Baker and Davenport,
Quart. J. Math. Oxford Ser. (2), 49 (1998), 291–306.
[15] A. DUJELLA and A. PETHO, Integer points on a family of elliptic curves, Publ. Math.
Debrecen, 56 (2000), 321–335.
[16] A. FILIPIN, Non-extendibility of D(−1)-triples of the form {1, 10, c}, Internat. J.
Math. Math. Sci., 35 (2005), 2217–2226.
[17] A. FILIPIN, There does not exist a D(4)-sextuple, J. Number Theory, 128 (2008),
1555–1565.
[18] A. FILIPIN, On the size of sets in which xy + 4 is always a square, Rocky Mountain
J. Math., to appear.
70 B. HE and A. TOGBE
[19] Y. FUJITA, The non-extensibility of D(4k)-triples {1, 4k(k − 1), 4k2 + 1}, Glasnik
Mat., 61 (2006), 205–216.
[20] Y. FUJITA, The extensibility of D(−1)-triples {1, b, c}, Publ. Math. Debrecen, 70
(2007), 103–117.
[21] Y. FUJITA, The extensibility of Diophantine pairs {k−1, k+1}, J. Number Theory,
128 (2008), 322–353.
[22] Y. FUJITA, The Hoggatt-Bergum conjecture on D(−1)-triples {F2k+1, F2k+3,
F2k+5} and integer points on the attached elliptic curves, Rocky Mountain J. Math.,
to appear.
[23] Y. FUJITA, Any Diophantine quintuple contains a regular Diophantine quadruple,
preprint.
[24] Y. FUJITA, The unique representation d = 4k(k2 −1) in D(4)-quadruples {k −2, k +
2, 4k, d}, preprint.
[25] Y. FUJITA, The number of Diophantine quintuples, Math. Commun., 11 (2006),
69–81.
[26] E. M. MATVEEV, An explicit lower bound for a homogeneous rational linear form in
logarithms of algebraic numbers II, Izv. Math., 64 (2000), 1217–1269.
[27] R. TAMURA, Non-extendibility of D(−1)-triples {1, b, c}, preprint.