ACTA ARITHMETICA

LXIII.4 (1993)

Rational approximations to π and some other numbers

by

Masayoshi Hata (Kyoto)

1. Introduction. In 1953 K. Mahler [12] gave a lower bound for

rational approximations to π by showing that
p
π− ≥ q −42
q
for any integers p, q with q ≥ 2. He also indicated that the exponent 42
can be replaced by 30 when q is greater than some integer q0 . This result
is based on the classical approximation formulae to the exponential and
logarithmic functions due to Hermite. Our aim is to improve the knowledge
of approximations √ by rational numbers of the classical constants of analysis
such as π and π/ 3. In general, we say that a real number γ has an
irrationality measure µ ≥ 2 provided that, for any ε > 0, there exists a
positive integer q0 (ε) satisfying
p
γ− ≥ q −µ−ε
q
for any integers p, q with q ≥ q0 (ε).
Later the exponent 30 mentioned above was improved to 20 by
M. Mignotte [13] by a refinement of Mahler’s method, and to 19.8899944 . . .
by G. V. Chudnovsky [3]. Indeed, Chudnovsky has succeeded in determining
the exact asymptotic behaviour of the complex integral
 k
1 R n!
(1.1) ewz dz
2πi C z(z − 1) . . . (z − n)

as n tends to infinity. Recently the author [11], modifying the above integral
(1.1), obtained a fairly good irrationality measure 13.394 for π.
On the other hand, F. Beukers [2] gave an elegant irrationality proof of
ζ(2) = π 2 /6 by considering the double (real) integral
336 M. Hata

R1 R1 L(x)(1 − y)n
(1.2) dx dy,
0 0
1 − xy
1
where L(x) = n!
(xn (1 − x)n )(n) is the Legendre polynomial. The inte-
gral (1.2) gives an irrationality measure 11.85078 . . . for π 2 and this has
been pared down to 10.02979 by R. Dvornicich and C. Viola [9], to 7.552
by E. A. Rukhadze [15], to 7.5252 by the author [10], and to 7.51 . . . by
D. V. Chudnovsky and G. V. Chudnovsky [6]. (The measures 7.552 and
7.51 . . . were announced without proofs.) Thus the last result due to Chud-
novsky and Chudnovsky brings an irrationality measure 15.02 . . . for π.
All the measures mentioned above are based on rational approximations
to π and its powers, though a single system of approximation to π is desired.
In this paper we consider the complex integral
R dz
(1.3) (F (a1 , a2 , a3 ; z))n
Γ
z

where
(z − a1 )2 (z − a2 )2 (z − a3 )2
(1.4) F (a1 , a2 , a3 ; z) =
z3
with non-zero distinct complex numbers a1 , a2 and a3 . By taking a1 = 1,
a2 = 2 and a3 = 1 + i, the integral (1.3) enables us to obtain the following
Theorem 1.1. For any ε > 0, there exists a positive integer H0 (ε) such
that
|p + qπ + r log 2| ≥ H −µ−ε
for any integers p, q, r with H ≡ max{|q|, |r|} ≥ H0 (ε), where the exponent
µ is given by
√
2 log α0 + 6 − π/ 3
µ=− √
2 log α1 + 6 − π/ 3
with
 √ 
10 √ 92 √
 
4jπ 1 23 69
αj = 18265 cos θ0 + + 6 and θ0 = arctan .
9 3 3 3 1209303
(Numerically one has µ = 7.016045 . . .)
In particular, we have
p π p
π− ≥ q −8.0161 and − ≥ q −8.0161
q log 2 q
for any integers p, q with q ≥ q0 . Hence the former remarkably improves the
earlier irrationality measures of π mentioned above.
As√ another application √ of the integral (1.3), by taking a1 = 1, a2 =
(1 + 3i)/2 and a3 = (3 + 3i)/4, we have the following
 Rational approximations 337

Theorem 1.2. For any ε > 0, there exists a positive integer q0 (ε) such
that
π p
√ − ≥ q −ξ−ε
3 q
for any integers p, q with q ≥ q0 (ε), where the exponent ξ is given by
√
2 log(β/54) + 6 − π/ 3
ξ =1+ √
log(β/3) − 6 + π/ 3
√ √ √
with β = 4529 + 3 u + v + 3 u − v, u = 92868417494 and v = 319599 5073.
(Numerically one has ξ = 4.601579 . . .)
√
Concerning the number π/ 3, K. Alladi and M. L. Robinson [1] ob-
tained an irrationality measure 8.30998 . . . by using the Legendre polynomi-
als. Later this was improved to 5.7926 by G. V. Chudnovsky [4], to 5.516 by
A. K. Dubitskas [8], to 5.0874 . . . by the author [10], and to 4.97 by G. Rhin
[14]. (The last measure 4.97 was announced without √ proof.) Therefore our
Theorem 1.2 improves all the measures for π/ 3 mentioned above.
The principal part of the integral (1.3), as n tends to +∞, can be easily
obtained by the saddle method originated in Riemann’s work, which is de-
scribed in detail in Dieudonné’s book [7]. This may be regarded as a natural
complex version of the well-known formula
 R1 1/n
lim ϕ(t)(f (t))n dt = max f (t)
n→∞ 0≤t≤1
0

for real-valued positive continuous functions ϕ(t) and f (t). Owing to this
method we do not need in this paper any discussion on recurrence relations
of polynomials. We note that our new exponents appearing in Theorems 1.1
and 1.2 contain the roots of some cubic equations with integral coefficients.
We prepare four lemmas in the next section. In particular, Lemma 2.2 is
an elementary arithmetical lemma concerning binomial coefficients, which
plays an important role in this paper. Then Theorems 1.1 and 1.2 will be
proved in Sections 3 and 4 respectively.

2. Preliminaries. First of all we need the following lemma concerning

linear independences over Q:
Lemma 2.1. Let m be a fixed non-negative integer. Let γ1 and γ2 be real
numbers satisfying
qn γ1 − pn = εn and qn γ2 − rn = δn
√
for some pn , qn , rn ∈ Z + i m Z for all n ≥ 1. Suppose that
1 1 1
lim log |qn | = σ , lim log |εn | = −τ , lim log |δn | = −τ ′
n→∞ n n→∞ n n→∞ n
338 M. Hata

for positive numbers σ, τ and τ ′ with τ ′ ≥ τ . Suppose further that there exist
infinitely many n’s satisfying δn /εn 6= ̺ for any rational number ̺. Then
the numbers 1, γ1 and γ2 are linearly independent over Q. More precisely,
for any ε > 0, there exists a positive integer H0 (ε) such that
|p + qγ1 + rγ2 | ≥ H −σ/τ −ε
for any integers p, q, r with H ≡ max{|q|, |r|} ≥ H0 (ε).
P r o o f. Obviously there exists an integer n0 such that qn 6= 0, εn 6= 0
and δn 6= 0 for all n ≥ n0 . Putting Λ = p + qγ1 + rγ2 for arbitrarily fixed
integers (p, q, r) 6= (0, 0, 0), we have
   
pn + εn rn + δn
Λ=p+q +r
qn qn
pqn + qpn + rrn qεn + rδn An ωn
= + ≡ + , say.
qn qn qn qn
We first show that the set Ω = {n ≥ 1 : An 6= 0} is infinite. Suppose, on
the contrary, that Ω is finite; that is, An = 0 for all n ≥ n1 for some integer
n1 ≥ n0 . Since |qn | ≥ 1 and ωn → 0 as n → ∞, we have Λ = 0. Hence
ωn = 0; so,
δn
q+r =0
εn
for all n ≥ n1 . Since (q, r) 6= (0, 0), we have r 6= 0; hence δn /εn = −q/r for
all n ≥ n1 , contrary to the hypothesis of the lemma.
For any ε > 0, one can define a sufficiently small δ ≡ δ(ε) ∈ (0, τ /6)
satisfying
σ+δ σ ε
< + .
τ − 3δ τ 2
Then there exists an integer n(ε) ≥ n0 such that
e(σ−δ)n ≤ |qn | ≤ e(σ+δ)n and max{|εn |, |δn |} ≤ e−(τ −δ)n
for all n ≥ n(ε). We now define
H0 (ε) = min{N ≥ 1 : 4N ≥ e(τ −δ)n(ε) and N ετ ≥ (4eσ+τ )2τ +4σ }
and
N (H) = min{N > n(ε) : 4H < e(τ −δ)N }
for all H ≥ H0 (ε). Since |ωn | ≤ 2He−(τ −δ)n for all n ≥ n(ε), we have
|ωn | < 1/2 for all n ≥ N (H). Therefore
An + ωn 1 − |ωn | 1 1
(2.1) |Λ| = ≥ > ≥ e−(σ+δ)n
qn |qn | 2|qn | 2
for any n √
∈ Ω with n ≥ N (H). (Note that |An | ≥ 1 if An 6= 0, since
An ∈ Z + i mZ.)
 Rational approximations 339

Now we distinguish two cases as follows: (a) N (H) ∈ Ω or (b) N (H) 6∈ Ω.

In Case (a), it follows from the definition of N (H) that 4H ≥ e(τ −δ)(N (H)−1) ;
so,
eN (H) ≤ 23/τ eH 1/(τ −δ) .
Hence, putting n = N (H) in (2.1), we have
|Λ| ≥ 21 e−(σ+δ)N (H) ≥ (H0 (ε))−ε/2 H −σ/τ −ε/2 ≥ H −σ/τ −ε ,
as required.
In Case (b), let M (H) be the smallest integer satisfying M (H) > N (H)
and M (H) ∈ Ω. Since AM (H)−1 = 0, we have
ωM (H)−1
|Λ| = ≤ 2He−(σ+τ −2δ)(M (H)−1) ;
qM (H)−1
so,
 1/(σ+τ −2δ)
M (H) 2H
e ≤e .
|Λ|
Hence, putting n = M (H) in (2.1), we obtain
 (σ+δ)/(σ+τ −2δ)
1 −(σ+δ)M (H) 1 −(σ+δ) |Λ|
|Λ| ≥ e ≥ e ;
2 2 2H
thus,
|Λ| ≥ (4eσ+τ )−1−2σ/τ H −(σ+δ)/(τ −3δ)
≥ (H0 (ε))−ε/2 H −σ/τ −ε/2 ≥ H −σ/τ −ε .
This completes the proof.
R e m a r k 2.1. By the same argument as above, one can show that if
qn γ − pn = εn for some irrational number γ,
1 1
lim log |qn | = σ and lim sup log |εn | ≤ −τ ,
n→∞ n n→∞ n
then γ has an irrationality measure 1 + σ/τ . This slightly improves Propo-
sition 3.3 in [10], in which the author posed the unnecessary hypothesis
that γ is not a Liouville number. This remark will be used in the proof of
Theorem 1.2.
We next give an elementary arithmetical lemma concerning binomial
coefficients.
Lemma 2.2. There exists a positive integer Dn such that
   
Dn 2n 2n 2n
(2.2) ∈Z
j + k + l − 3n j k l
340 M. Hata

for all integers 0 ≤ j, k, l ≤ 2n with j + k + l 6= 3n and that

 
1 1 π 27
(2.3) lim log Dn = 6 − √ + log ≡ κ, say .
n→∞ n 2 3 16
(Numerically one has κ = 2.35472 . . .)
P r o o f. Let v(s, p) be the exponent of any prime factor p in the resolu-
tion of each integer 1 ≤ s ≤ 3n into its prime factors; that is,
Y
s= pv(s,p) .
p prime

Then clearly
   
log s log(3n)
v(s, p) ≤ ≤ .
log p log p
√
We next show that any prime number p > 3n belonging to the set
   
1 n 2
Tn = p prime : ≤ < ,
2 p 3
divides all the integers
   
2n 2n 2n
,
j k l
where (j, k, l) runs through the set Ξ(n, p) = {(j, k, l) : 0 ≤ j, k, l ≤ 2n and
j + k + l ≡ 3n (mod p)}. ({x} denotes the fractional part of x.) To see this,
it suffices to show that
           
2n j 2n − j 2n k 2n − k
J= − − + − −
p p p p p p
     
2n l 2n − l
+ − − ≥1
p p p
√
for any (j, k, l) ∈ Ξ(n, p), since p > 3n. Let ω = {n/p}, θ = {j/p},
θ ′ = {k/p} and θ ′′ = {l/p}. Then obviously
J = [2ω] − [2ω − θ] + [2ω] − [2ω − θ ′ ] + [2ω] − [2ω − θ ′′ ]
= 3 − [2ω − θ] − [2ω − θ ′ ] − [2ω − θ ′′ ] ,
since [2ω] = 1. On the other hand, since 3n = j + k + l + pq for some integer
q and since [3ω] = 1, we have
{3ω} = {θ + θ ′ + θ ′′ } = 3ω − 1 .
Thus θ + θ ′ + θ ′′ = 3ω − 1 + r for some integer r ≥ 0; hence
[2ω − θ] + [2ω − θ ′ ] + [2ω − θ ′′ ] ≤ [6ω − θ − θ ′ − θ ′′ ] = [6ω − 3ω + 1 − r]
= 1 − r + [3ω] = 2 − r ≤ 2 .
Therefore we have J ≥ 1, as required.
 Rational approximations 341

We define
Y  log(3n)  Y Y
∆1 (n) = p log p , ∆2 (n) = p, ∆3 (n) = p
p prime
√ p prime p prime
p≤ 3n p≤3n p∈Tn

for all n ≥ 1. The above argument implies that the integer

∆1 (n)∆2 (n)
Dn =
∆3 (n)
satisfies (2.2) for any integers j, k, l ∈ [0, 2n] with j + k + l 6= 3n. Put
bi = 3/(3i + 2), ci = 2/(2i + 1) and
1 X
Si (n) = log p
n
p prime
p/n∈(bi ,ci ]

for i ≥ 0 for brevity. Then clearly

∞
1 X
log ∆3 (n) = Si (n)
n i=0

and it follows from the prime number theorem that Si (n) → ci −bi as n → ∞
for each i ≥ 0.
Thus
∞ ∞
1 X X
lim log ∆3 (n) = lim Si (n) = (ci − bi )
n→∞ n n→∞
i=0 i=0
Γ ′ (2/3) Γ ′ (1/2)
 
1 π 27
= − = √ − log ,
Γ (2/3) Γ (1/2) 2 3 16
where Γ (z) is the gamma function. (The interchange of the operations of
limit and summation can be easily justified by using the estimate
X
Si (n) ≤ const cL
i≥L

for any L ≥ 1, where const √ depends neither on L nor on n.) Since it can
be seen that log ∆1 (n) = O ( n) and log ∆2 (n) ∼ 3n as n tends to +∞, we
thus obtain the asymptotic behaviour (2.3). This completes the proof.
We now consider the function F (z) ≡ F (a1 , a2 , a3 ; z) defined in (1.4) for
non-zero distinct complex numbers a1 , a2 and a3 . It is easily seen that F ′ (z)
has three zero points ζ1 , ζ2 , ζ3 in the region C − {a1 , a2 , a3 }, which satisfy
the cubic equation
(2.4) z 3 + s2 z 2 + s1 z + s0 = 0
342 M. Hata

with s0 = a1 a2 a3 , s1 = −(a1 a2 + a2 a3 + a3 a1 )/3 and s2 = −(a1 + a2 + a3 )/3.

Such points ζj are called “saddles” of the surface Σ of |F (z)|. Using the
equalities
s21 − 3s0 s2
 
2
F (ζj ) = 4 (s2 − 3s1 )ζj + − 7s1 s2 − s0
ζj
for 1 ≤ j ≤ 3, one can easily calculate every elementary symmetric function
on F (ζ1 ), F (ζ2 ) and F (ζ3 ). We thus have
Lemma 2.3. Put t0 = s0 , t1 = s1 s2 and t2 = s32 +s31 /s0 for brevity. Then
we have the identity
3
Y
(z − F (ζj )) = z 3 + 4σ2 z 2 + 16σ1 z + 64σ0 ,
j=1

where
σ0 = t30 − 12t20 t1 − 8t20 t2 + 30t0 t21 + 48t0 t1 t2
+ 16t0 t22 + 36t31 + 24t21 t2 + 9t41 /t0 ,
σ1 = 3t20 + 3t0 t1 + 20t0 t2 + 57t21 + 12t1 t2 + t31 /t0 ,
σ2 = 3t0 + 15t1 + t2 .
Finally, we recall the following result from Dieudonné’s book [7]:
Lemma 2.4. Let Γ be a smooth oriented path with a finite length, which
departs from aj and arrives at ak through a saddle ζl for some 1 ≤ j, k, l ≤ 3.
Put ℓz = {z + t : t ≥ 0}, the half-straight line parallel to the real axis.
Suppose that the saddle ζl is a simple root of (2.4) and suppose further
that Γ − {aj , ak } is contained in the simply connected region C0 ≡ C −
(ℓ0 ∪ ℓa1 ∪ ℓa2 ∪ ℓa3 ) and that Γ − {aj , ak , ζl } is contained in the open set
{z ∈ C0 : |F (z)| < |F (ζl )|}. Then the principal part of the integral
R dz
In = (F (z))n ,
Γ
z

as n tends to +∞, is given by

s
(F (ζl ))n F (ζl ) 1
In ∼ d0 √ with d0 = −2π · ;
n F ′′ (ζl ) ζl
in particular , we have
1
lim log |In | = log |F (ζl )| .
n→∞ n
Indeed, the above lemma is an immediate consequence of the result in
 Rational approximations 343

[7; Chapter IX], since the integral In can be expressed as

R
In = (−1)n g(z)enh(z) dz ,
Γ
P3
where g(z) = 1/z and h(z) = 2 j=1 Log(aj − z) − 3 Log(−z) are analytic
functions in C0 .

3. Proof of Theorem 1.1. Let Γz,w be a smooth oriented path de-

parting from z, arriving at w, and contained in C − {0}. We then consider
the integral
R dz
In (Γz,w ) ≡ (F (a1 , a2 , a3 ; z))n
Γz,w
z
2n X 2n
2n X    
X 2n 2n 2n R
= Bj,k,l z j+k+l−3n−1 dz,
j=0 k=0 l=0
j k l Γz,w

where
Bj,k,l = (−1)j+k+l a12n−j a22n−k a32n−l .
Therefore we have the identity
(3.1) In (Γz,w )
   
X Bj,k,l 2n 2n 2n
= (wj+k+l−3n − z j+k+l−3n )
j + k + l − 3n j k l
j+k+l6=3n
    R
X 2n 2n 2n dz
+ Bj,k,l
j k l Γ z
j+k+l=3n z,w

R dz
≡ un (z, w) + vn , say.
Γz,w
z

For the proof of Theorem 1.1 we choose a1 = 1, a2 = 2 and a3 = 1 + i.

Then it follows from Lemma 2.3 that wk ≡ 49 (i − 1)F (ζk ), 1 ≤ k ≤ 3, satisfy
the cubic equation
9w3 − 4968w2 + 862w − 36 = 0 .
This algebraic equation has three positive roots; indeed,
10 √
 
2kπ
wk = 109590 cos θ0 + + 184
9 3
√
23 69
for 1 ≤ k ≤ 3 where θ0 = 13 arctan 1209303


. We thus have |F (ζk )| =
2
√
9 2wk . Let Γ1,2 and Γ1,1+i be the paths illustrated in Figure 1 through
344 M. Hata

the saddles ζ1 and ζ2 respectively. Then clearly

R dz R dz 1 π
= log 2 and = log 2 + i .
Γ1,2
z Γ1,1+i
z 2 4

Fig. 1

Since
(−1)j+k+l Bj,k,l 2j+k+l−3n = 2j+l−n (1 + i)2n−l = 2j in (1 − i)l
and
(−1)j+k+l Bj,k,l (1 + i)j+k+l−3n = 22n−k (1 + i)j+k−n = (1 + i)j+n (1 − i)2n−k ,
it follows from Lemma 2.2 that both un (1, 2) and un (1, 1 + i) belong to the
set (Z + iZ)/Dn . Hence, putting
pn = −2iDn {un (1, 2) − 2un (1, 1 + i)}, qn = Dn vn , rn = −Dn un (1, 2) ,
we obtain
qn π − pn = 2iDn {In (Γ1,2 ) − 2In (Γ1,1+i )} ≡ εn
and
qn log 2 − rn = Dn In (Γ1,2 ) ≡ δn , say,
where pn , qn , rn ∈ Z + iZ. Then it follows from Lemmas 2.2 and 2.4 that
1
τ = − lim log |εn | = −κ − log |F (ζ2 )|
n→∞ n
 Rational approximations 345

and
1
τ ′ = − lim log |δn | = −κ − log |F (ζ1 )| ,
n→∞ n
since |F (ζ1 )| < |F (ζ2 )|. Moreover, the hypothesis on the sequence {δn /εn }
in Lemma 2.1 is clearly satisfied since τ ′ > τ .

Fig. 2

On the other hand,

   
X 2n 2n 2n 1 R dz
vn = Bj,k,l = (F (z))n ,
j k l 2πi C
z
j+k+l=3n

where C = Γ1,1+i ∪ Γ1+i,1 is a closed oriented curve enclosing the origin and
Γ1+i,1 is the path illustrated in Figure 2 through the saddle ζ3 . Hence
1
σ = lim log |qn | = κ + log |F (ζ3 )| ,
n→∞ n
since |F (ζ3 )| > |F (ζ2 )|. Therefore, by Lemma 2.1, the numbers 1, log 2 and
π have a linear independence measure
σ κ + log |F (ζ3 )|
µ= =− .
τ κ + log |F (ζ2 )|
This completes the proof.

4. Proof of Theorem 1.2. We take a1 = 1, a2 = λ, a3 = λ′ =√(1+λ)/2

with λ = eπi/3 . Then it follows from Lemma 2.3 that wk = 3 3iF (ζk ),
346 M. Hata

1 ≤ k ≤ 3, satisfy the cubic equation

18w3 − 4529w2 + 81w − 1 = 0 .
Indeed, this algebraic equation has one positive root
1 √ √ β
w3 = (4529 + 3 u + v + 3 u − v) ≡ , say,
54 54
√
with u = 92868417494 and v = 319599 5073, and two complex roots
w1 , w2 = w 1 . We thus have
√ √
3 3β 1
|F (ζ3 )| = w3 = and |F (ζ1 )| = |F (ζ2 )| = √
9 486 3 β
since w1 w2 w3 = 1/18. Let Γ1,λ be the path illustrated in Figure 3 through
the saddles ζ1 and ζ2 . Then clearly
R dz π
= Log λ = i
Γ1,λ
z 3

and
√
j+k+l 2n−k 2n−l l−2n Z + i 3Z
(−1) Bj,k,l = λ (1 + λ) 2 ∈ .
22n+1

Fig. 3

Hence, putting
√
qn = 22n+1 Dn vn and pn = 22n+1 3iDn un (1, λ) ,
 Rational approximations 347

we obtain, from (3.1),

π √
qn √ − pn = −22n+1 3iDn In (Γ1,λ ) ≡ εn , say,
3
√
where pn , qn ∈ Z + i 3 Z. Then it follows from Lemmas 2.2 and 2.4 that
1 1
lim sup log |εn | ≤ log 4 + κ + lim sup log(|In (Γ1,λ′ )| + |In (Γλ′ ,λ )|)
n→∞ n n→∞ n
≤ log 4 + κ + log |F (ζ1 )| ≡ −τ, say,
and that
1
σ = lim log |qn | = log 4 + κ + log |F (ζ3 )|
n→∞ n
since
1 R dz
vn = (F (z))n
2πi C
z

where C = Γ1,λ ∪ Γλ,1 is a closed curve enclosing the origin and Γλ,1 is the
path illustrated√in Figure 4 through the saddle ζ3 . Therefore, by Remark 2.1,
the number π/ 3 has an irrationality measure
σ log 4 + κ + log |F (ζ3 )|
ξ =1+ =1− .
τ log 4 + κ + log |F (ζ1 )|
This completes the proof of Theorem 1.2.

Fig. 4
348 M. Hata

5. Remarks. Using Lemma 2.1 one √ can also obtain a linear inde-
pendence measure for the numbers 1, π/ 3 and log(3/4) from the integral
considered in Section 4. However, it can be seen that the case in which
a1 = 3/2, a2 = 2 and a3 = 1 + λ with λ = eπi/3 gives a better linear
independence measure µ for such numbers. Indeed, we have
log 8 + κ + log |x3 |
µ=− = 7.813765 . . . ,
log 8 + κ + log |x2 |
√
where wk = −(1/ 3)e−πi/6 xk , 1 ≤ k ≤ 3, are three positive roots of the
cubic equation 8748w3 −1621134w2 +3547w−1 = 0 satisfying w1 < w2 < w3 .
The hypothesis on {δn /εn } in Lemma 2.1 is clearly satisfied since τ ′ > τ .
(We must multiply both sides of (3.1) by the factor 23n+1 Dn .) In particular,
π p
√ − ≥ q −8.8138
3 log(3/4) q
for any integer p and for sufficiently large integer q.
′
Similarly√ one can obtain a linear independence measure µ for the num-
bers 1, π/ 3 and log 3 by taking a1 = 1, a2 = λ and a3 = 1 + λ. In
fact,
1
log 3 + κ + log |y3 |
µ′ = − 21 = 8.385245 . . . ,
2 log 3 + κ + log |y1 |
√
where wk = 43 3iyk , 1 ≤ k ≤ 3, are three roots of the cubic equation
9w3 − 1187w2 − 61w − 1 = 0 satisfying w3 > 0 and w1 = w2 . (We must
multiply both sides of (3.1) by the factor 2(1 + λ)n Dn .) In this case we have
|y1 | = |y2 | (that is, τ ′ = τ ); so it follows from Lemma 2.4 that
δn
= d∗ einν + o(1)
εn
as n tends to +∞, where d∗ =
6 0 is some complex constant and ν = arg w1 −
arg w2 satisfies 0 < |ν| < π. Hence the sequence {δn /εn } satisfies the
hypothesis of Lemma 2.1, as required. In particular,
π p
√ − ≥ q −9.3853
3 log 3 q
for any integer p and for sufficiently large q.

References

[1] K. A l l a d i and M. L. R o b i n s o n, Legendre polynomials and irrationality, J. Reine

Angew. Math. 318 (1980), 137–155.
[2] F. B e u k e r s, A note on the irrationality of ζ(2) and ζ(3), Bull. London Math. Soc.
11 (1979), 268–272.
 Rational approximations 349

[3] G. V. C h u d n o v s k y, Hermite–Padé approximations to exponential functions and

elementary estimates of the measure of irrationality of π, in: Lecture Notes in Math.
925, Springer, Berlin 1982, 299–322.
[4] —, Recurrences, Padé approximations and their applications, in: Lecture Notes in
Pure and Appl. Math. 92, Dekker, New York 1984, 215–238.
[5] D. V. C h u d n o v s k y and G. V. C h u d n o v s k y, Padé and rational approximations
to systems of functions and their arithmetic applications, in: Lecture Notes in Math.
1052, Springer, Berlin 1984, 37–84.
[6] —, —, Transcendental methods and theta-functions, in: Proc. Sympos. Pure Math.
49 (2), Amer. Math. Soc., 1989, 167–232.
[7] J. D i e u d o n né, Calcul infinitésimal , Hermann,
√ Paris 1968.
[8] A. K. D u b i t s k a s, Approximation of π/ 3 by rational fractions, Vestnik Moskov.
Univ. Ser. I Mat. Mekh. 1987 (6), 73–76 (in Russian).
[9] R. D v o r n i c i c h and C. V i o l a, Some remarks on Beukers’ integrals, in: Colloq.
Math. Soc. János Bolyai 51, Budapest 1987, 637–657.
[10] M. H a t a, Legendre type polynomials and irrationality measures, J. Reine Angew.
Math. 407 (1990), 99–125.
[11] —, A lower bound for rational approximations to π, J. Number Theory, to appear.
[12] K. M a h l e r, On the approximation of π, Nederl. Akad. Wetensch. Proc. Ser. A 56
(1953), 30–42.
[13] M. M i g n o t t e, Approximations rationnelles de π et quelques autres nombres, Bull.
Soc. Math. France Mém. 37 (1974), 121–132.
[14] G. R h i n, Approximants de Padé et mesures effectives d’irrationalité, in: Progr. in
Math. 71, Birkhäuser, 1987, 155–164.
[15] E. A. R u k h a d z e, A lower bound for the approximation of ln 2 by rational numbers,
Vestnik Moskov. Univ. Ser. I Mat. Mekh. 1987 (6), 25–29 (in Russian).

INSTITUTE OF MATHEMATICS
YOSHIDA COLLEGE
KYOTO UNIVERSITY
KYOTO 606, JAPAN

Received on 3.3.1992
and in revised form on 27.10.1992 (2239)