DOMB’S NUMBERS AND RAMANUJAN-SATO TYPE

SERIES FOR 1/π

HENG HUAT CHAN, SONG HENG CHAN, AND ZHIGUO LIU

Abstract. In this article, we construct a general series for π1 . We

indicate that Ramanujan’s π1 −series are all special cases of this general
series and we end the paper with a new class of π1 −series. Our work is
motivated by series recently discovered by Takeshi Sato.

1. Introduction
In [14], S. Ramanujan recorded a total of 17 series for 1/π, one of which
is
∞ ¡ 1 ¢3
4 X (6k + 1) 2 k
(1.1) = ,
π (k!)3 4k
k=0

where
(a)k = (a)(a + 1) · · · (a + k − 1).
Ramanujan did not indicate how he arrived at these series but instead hinted
that some of these series belonged to what are now known, after the work
of B.C. Berndt, S. Bhargava and F.G. Garvan [2], as “the theories of elliptic
functions to alternative bases”.
Ramanujan’s series for 1/π were not extensively studied until around
1987. The Chudnovskys succeeded in extending Ramanujan’s list of series
and derived the following series which they used to compute over a billion
digits of π:
X ∞
6403203/2 (6k)! (13591409 + 545140134k)
(1.2) = 12 (−1)k .
π (k!)3 (3k)! 6403203k
k=0

For a recent discussion of the Chudnovskys’ series, see [3].

In [6], the Borweins provided rigorous proofs of Ramanujan’s series for
the first time and derived many new series for 1/π. Both the Borweins and

Keywords : Ramanujan-type series, theta functions, hypergeometric series, Clausen’s

identity, modular equations, Domb’s numbers, differential equations.
2000 Mathematics Subject Classifications : 11F11, 11F27, 11Y60, 33C20, 33C05, 05A10.
The first and third authors are funded by National University of Singapore Academic
Research Fund, Project Number R146000027112.
1
2 HENG HUAT CHAN, SONG HENG CHAN, AND ZHIGUO LIU

the Chudnovskys admitted that Clausen’s identity, namely, [6, p. 188]

µ µ ¶¶2 µ ¶
1 1
2 F1 a, b; a + b + ; z = 3 F2 2a, 2b, a + b; a + b + , 2a + 2b; z ,
2 2
plays an important role in their derivations of such series.
In this article, we will derive a general series for 1/π without using
Clausen’s identity. We will show that all the existing series for 1/π are
special cases of this general series. We will then specialize our series further
to derive new classes of series for 1/π. In particular, we will show that
X∞ X k µ ¶2 µ ¶ µ ¶ µ ¶k
8 k 2(k − j) 2j 1
(1.3) √ = (1 + 5k).
3π j k − j j 64
k=0 j=0

Our work is motivated by the following series discovered recently by Takeshi

Sato [12] 1:
√
1 15
√
π 120(4 5 − 9)
∞ X
X k µ ¶2 µ ¶2 µ ¶ Ã√ !12k
k k+j 1 3√ 5−1
= − 5+k .
j j 2 20 2
k=0 j=0

The above series shows that Clausen’s formula is probably not needed in
the derivation of Ramanujan’s series since there are no Clausen-type trans-
formation formula for series such as
X∞ X k µ ¶2 µ ¶2
k k+j
xk .
j j
k=0 j=0

2. A general series for 1/π

Suppose the function Z(q) satisfies the relation
√ √
(2.1) rZ(e−2π r/s ) = Z(e−2π/ rs ).
By differentiating (2.1) with respect to r and simplifying, we find that
√ √ √
rs √ 0 −2π/ rs ) √ 0 −2π r/s )
−2π/ rs Z (e −2π r/s Z (e
(2.2) =e √ + re √ ,
π Z(e−2π/ rs ) Z(e−2π r/s )
where
dZ(q)
Z 0 (q) = .
dq
Next, set
Z(q)
(2.3) MN (q) = .
Z(q N )
1The authors are grateful to S. Kanemitsu for sending the abstract of Sato’s talk which
contains several new series for 1/π.
 DOMB’S NUMBERS AND RAMANUJAN-SATO TYPE SERIES FOR 1/π 3

By differentiating (2.3) with respect to q, we find that

0 (q)
MN Z 0 (q) Z 0 (q N )
= − N q N −1 ,
MN (q) Z(q) Z(q N )
which implies that
0 (q)
MN Z 0 (q) Z 0 (q N )
(2.4) q =q − N qN .
MN (q) Z(q) Z(q N )
√
Substituting q = e−2π/ N s into (2.4) and using (2.1), we find that
(2.5) √ √ √
√
e−2π/ N s MN0 (e−2π/ N s ) √ Z 0 (e−2π/ N s ) √ Z 0 (e−2π N/s )
= e−2π/ N s √ −N e−2π N/s √ .
N Z(e −2π/ N s ) Z(e−2π N/s )
Solving (2.5) and (2.2) with r = N , we conclude that
√ √ √ √
Ns √ Z 0 (e−2π N/s ) e−2π/ N s MN
0 (e−2π/ N s )
−2π N/s
(2.6) = 2N e √ + .
π Z(e−2π N/s ) N

At this stage, we suppose that there exist functions X(q) and U (q) such
that
X∞
(2.7) Z(q) = Ak X k (q), Ak ∈ Q
k=0

and
dX(q)
(2.8) q = U (q)X(q)Z(q).
dq
Relation (2.3) and (2.7) also imply that MN may be expressed as a function
of two variables, namely,
(2.9) MN = MN (X(q), X(q N )).
We may also assume that X(q) and X(q N ) satisfies a polynomial relation
and hence,using (2.8), (2.1) and (2.7),we deduce that
(2.10)
√ √ ¯
e−2π/ N s M 0 (e−2π/ N s )
N 1 dX(q) dMN (X(q), X(q N )) ¯¯
= q ¯ −2π/√N s
N N dq dX(q) q=e
¯
1 √
−2π/ N s
√
−2π/ N s
√
−2π/ N s dMN (X(q), X(q N )) ¯¯
= U (e )X(e )Z(e ) ¯ −2π/√N s
N dX(q) q=e
√ ¯
√ √ dMN (X(q), X(q )) ¯ N ¯
= U (e−2π/ N s )X(e−2π/ N s )Z(e−2π N/s ) ¯ −2π/√N s
dX(q) q=e
¯ ∞ √
√ √ dMN (X(q), X(q N )) ¯¯ X
= U (e−2π/ N s )X(e−2π/ N s ) ¯ −2π/√N s A k X k −2π N/s
(e ).
dX(q) q=e k=0
4 HENG HUAT CHAN, SONG HENG CHAN, AND ZHIGUO LIU

This simplifies the second term on the right hand side of (2.6). Next, by
(2.7) and (2.8), we find that
∞
X
0
qZ (q) = kAk X k−1 (q)qX 0 (q)
k=0
∞
X
= kAk X k−1 (q)U (q)X(q)Z(q).
k=0
Hence,
X ∞
Z 0 (q)
(2.11) q = U (q) kAk X k (q).
Z(q)
k=0
√
−2π N/s
Substituting q = e into (2.11) and combining with (2.10) and (2.6),
we deduce the following result:
Theorem 2.1. Suppose Z(q) satisfies (2.1) and let MN (q), X(q), Ak and
U (q) be defined as in (2.3), (2.7) and (2.8). Then
r ∞
s 1 X
k
(2.12) = (bN k + aN ) Ak XN ,
N 2π
k=0
where
√ ¯
√
−2π/ N s X(e−2π/ N s)
dMN N ¯
¯
aN = U (e ) (X(q), X(q ))¯ ,
2N dX(q) √
q=e−2π/ N s
√
bN = U (e−2π N/s ),
and √
XN = X(e−2π N/s
).

We end this section by listing s, x, Z, U and Ak which correspond to

Ramanujan’s series for 1/π. In Table 1, x is given by the relation
X = 4x(1 − x),
∞
Y
f (−q) = (1 − q k ),
k=1
and
Q3
j(q) = 1728 ,
qf 24 (−q)
where
∞
X k3 qk
Q := 1 + 240 .
1 − qk
k=1
In the next section, we will illustrate how we derive the fourth class of series
without using Clausen’s identity. We will then give a new proof of (1.1).
 DOMB’S NUMBERS AND RAMANUJAN-SATO TYPE SERIES FOR 1/π 5

Table 1

s x Z U Ak

à s ! ¡1¢ ¡1¢ ¡5¢

1 1728 p 2 k 6 k 6 k
1 1− 1− Q 1 − 2x
2 j(q) (k!)3
à µ ¶24 !−1 s µ ¶24 ¡1¢ ¡1¢ ¡3¢
1 f (−q) f 8 (−q) f (−q 2 ) 2 k 4 k 4 k
2 1+ 1 + 64q 1 − 2x
64q f (−q 2 ) f 4 (−q 2 ) f (−q) (k!)3
à µ ¶12 !−1 à ∞
!2 ¡1¢ ¡1¢ ¡2¢
1 f (−q) X
m2 +mn+n2 2 k 3 k 3 k
3 1+ q 1 − 2x
27q f (−q 3 ) m,n=−∞
(k!)3
à µ ¶8 !−1 à ∞
!2 ¡ 1 ¢3
1 f (−q) X
m2 +n2 2 k
4 1+ q 1 − 2x
16q f (−q 4 ) m,n=−∞
(k!)3

3. Ramanujan’s series for 1/π associated with the Jacobian

theta functions
Let Z(q) = ϕ4 (q) where
∞
X 2
ϕ(q) = qk .
k=−∞

Note that Z(q) satisfies the transformation formula [1, p. 43, Entry 27(i)]
√ √
(3.1) N Z(e−π N
) = Z(e−π/ N
),
and hence, s = 4.
2
It is known that if
ψ 4 (q 2 )
(3.2) x = x(q) = 16q ,
ϕ4 (q)
with
∞
X
ψ(q) = q k(k+1)/2 ,
k=0

2Our x given here is equal to the x given in the Table in Section 2.

6 HENG HUAT CHAN, SONG HENG CHAN, AND ZHIGUO LIU

then z = z(q) = ϕ2 (q) and x satisfy the differential equation [15, p.54,
(3.83)]
d2 z dz z
(3.3) x(1 − x)
+ (1 − 2x) − = 0.
dx2 dx 4
√
Substituting z = Z and X = 4x(1 − x) into (3.3), we deduce that3
µ ¶
3 1 3
(3.4) ϑX (Z) = X ϑX + (Z),
2
where the operator
d
ϑX = X .
dX
If
∞
X
Z= Ak X k ,
k=0
then (3.4) yields a recurrence relation satisfied by Ak and we deduce imme-
diately that
¡ 1 ¢3
2 k
Ak = .
(k!)3
In other words, we find that
µ ¶
1 1 1
(3.5) Z = 3 F2 , , ; 1, 1; X .
2 2 2
It is also known that [1, p. 120, Entry 9 (i)]
dX
(3.6) q = ZX(1 − 2x),
dq
which means that
U (q) = 1 − 2x(q).
By (3.6), we find that if Y = X(q N ) and y = x(q N ), then
dX 1 X 1 − 2x
(3.7) = MN .
dY N Y 1 − 2y
Squaring (3.7), we deduce that
µ ¶2
2 Y 2 (1 − Y )
2 dX
(3.8) MN =N .
X 2 (1 − X) dY
dMN
From the above relation, can be computed explicitly once we compute
dX
the modular relation satisfied by X and Y .
We now give an explicit series for 1/π. First, from (3.1) and [1, p. 43,
Entry 27 (ii)], we find that
αϕ4 (e−α ) = βϕ4 (e−β ),
3The calculations are tedious but straightforward.
 DOMB’S NUMBERS AND RAMANUJAN-SATO TYPE SERIES FOR 1/π 7

and
16αe−α ψ 4 (e−2α ) = βϕ4 (−e−β ),
if αβ = π 2 . By (3.2) and the above relations,
√
√ ϕ4 (−e−π N)
(3.9) x(e−π/ N
)= √ .
ϕ4 (e−π N )

Using the famous Jacobi identity [1, p. 40, Entry 25 (vii)],

ϕ4 (q) − ϕ4 (−q) = 16qψ 4 (q 2 ),

we conclude that
√ √
(3.10) x(e−π/ N
) = 1 − x(e−π N
).

Now, let N = 3, then from Ramanujan’s modular equation [1, p. 230, Entry]

(x(q)x(q 3 ))1/4 + ((1 − x(q))(1 − x(q 3 )))1/4 = 1,

and (3.10), we find that

√ 1
(3.11) X(e−π/ 3
)= .
4
Solving (3.11) for x, we deduce that
√
√
−π 3 1 3
(3.12) x(e )= − .
2 4
4
Next, one can show that X = X(q) and Y = X(q 3 ) satisfy the relation

− 4096XY + 4608X 2 Y + 4608XY 2 − 900X 3 Y + 28422X 2 Y 2

− 900XY 3 + 4608X 3 Y 2 + 4608X 2 Y 3 − 4096X 3 Y 3 + X 4 + Y 4 = 0.

Differentiating the above relation with respect to Y and substituting the

result into (3.8), we obtain an expression of M3 in terms of X and Y .
Differentiating the resulting relation with respect to X and using (3.11) and
(3.12), we conclude that
√ √
3 3
a3 = and b3 = .
12 2
This gives Ramanujan’s series (1.1).

4This modular equation can be proved using [1, p. 231, Entry 5(xii)]
8 HENG HUAT CHAN, SONG HENG CHAN, AND ZHIGUO LIU

4. A new class of series for 1/π

From the last section, we find that in order to derive a new class of series
different from that of Ramanujan, one needs to seek new pair of functions
X and Z. In this section, we will construct a new class of series for 1/π
and derive (1.3). Series similar to (1.3) where all the coefficients in the sum
are rational are very rare. It turns out that Ramanujan nearly exhausted
all these series in [14]. The Chudnovskys 5 and the Borweins 6 constructed
several such series for 1/π associated with the first and second type of series
in Table 1, respectively. Recently, Chan, W.C. Liaw and V. Tan added a
total of six such series [8] associated essentially to the third type of series in
Table 1.
Let
(f (−q)f (−q 3 ))4
Z=
(f (−q 2 )f (−q 6 ))2
and
µ ¶6
f (−q 2 )f (−q 6 )
X=q .
f (−q)f (−q 3 )
From our discussion in Section 3, we need to determine U (q) and Ak . It
follows from [4, Theorem 10.6]

dX p
(4.1) q = ZX (4X + 1)(16X + 1)
dq
that
p
(4.2) U (q) = (4X + 1)(16X + 1).
To determine Ak , we need to establish the differential equation satisfied
by Z and X.
Let
1 1 f 12 (−q)
(4.3) =1+ .
t 27q f 12 (−q 3 )
The differential equation satisfied by
à ∞
!2
X
m2 +mn+n2
u = 4t(1 − t) and f = q
m,n=−∞

is
µ ¶µ ¶µ ¶
1 1 2
(4.4) ϑ3u f = u ϑu + ϑu + ϑu + (f ).
2 3 3

5The Chudnovskys’ series (1.2) yields the fastest convergent series for 1/π which in-
volves only rational coefficients.
6
See [5, p. 145] for the two series missed by Ramanujan.
 DOMB’S NUMBERS AND RAMANUJAN-SATO TYPE SERIES FOR 1/π 9

The proof of (4.4) is similar to that of (3.4) but instead of using (3.3), we
use the differential equation [7, p. 200]
d2 z dz 2
t(1 − t) 2
+ (1 − 2t) − z = 0,
dt dt 9
where z 2 = f. Incidently (4.4) is the differential equation needed to deter-
mine Ak in the third type of series in Table 1. Our method of deriving a
differential equation satisfied by Z and X is to first establish the identities

(4.5) f = Z(1 + 16X)

and
108X
(4.6) u= .
(1 + 16X)3
Using (4.5) and (4.6), we conclude that
(4.7)
(1 + 16X)2 ϑX Z + 16X(1 + 16X)Z
ϑu f = ,
1 − 32X
(4.8)
1 + 16X
ϑ2u f =
(1 − 32X)3
× {(1 − 768X 2 − 8192X 3 )ϑ2X Z + (80X + 256X 2 − 16384X 3 )ϑX Z
+ (16X + 512X 2 − 8192X 3 )Z}

and
(4.9)
1 + 16X
ϑ3u f =
(1 − 32X)5
× {(1 − 16X − 1280X 2 + 4096X 3 + 524288X 4 + 4194304X 5 )ϑ3X Z
+ (192X − 1536X 2 − 147456X 3 − 393216X 4 + 12582912X 5 )ϑ2X Z
+ (96X + 8448X 2 − 12288X 3 − 1179648X 4 + 12582912X 5 )ϑX Z
+ (16X + 2560X 2 + 24576X 3 − 524288X 4 + 4194304X 5 )Z}.
Substituting (4.7)–(4.9) into (4.4) and simplifying, we immediately obtain
the differential equation
(4.10) (1 + 20X + 64X 2 )ϑ3X Z + (192X 2 + 30X)ϑ2X Z
+ (192X 2 + 18X)ϑX Z + (64X 2 + 4X)Z = 0,
d
where ϑX = X . For an alternative proof of (4.10), see [16].
dX
10 HENG HUAT CHAN, SONG HENG CHAN, AND ZHIGUO LIU

It remains to establish (4.5) and (4.6). From the definition (4.3), we find
that
1 1 27
= 2 + P6 + 6,
u 27 P
where
f 2 (−q)
P = q −1/6 2 .
f (−q 3 )
Identity (4.6) then follows immediately from [4, Lemma 10.3, (10.4)]
µ ¶
3 27 2 (1 + 16X)3
P + 3 = .
P X
To prove (4.5), recall from [7, p. 200] that
du √
(4.11) q = f u 1 − u.
dq
This is the analogue of (3.6). From (4.6), we find that
du 1 − 32X
(4.12) = 108 .
dX (1 + 16X)4
Combining (4.12), (4.11), (4.6) and (4.1), we conclude that
f = Z(1 + 16X),
which is (4.5).
Now, from (4.10), we conclude that if
∞
X
Z= Ak X k ,
k=0
then Ak satisfies the difference equation
(4.13) k 3 Ak + 2(2k − 1)(5k 2 − 5k + 2)Ak−1 + 64(k − 1)3 Ak−2 = 0.
A check with N. Sloane’s Online sequences [13] shows that
X k µ ¶2 µ ¶µ ¶
k k 2(k − j) 2j
(4.14) Ak = (−1) .
j k−j j
j=0

This sequence appears to be first discovered by C. Domb [9]. The fact that
this sequence satisfies the recurrence (4.13) can be verified using the method
of creative telescoping implemented in D. Zeilberger’s MAPLE programme
EKHAD [11, Chapter 7]. More precisely, Zeilberger’s programme produces
the function
R(k, j) =
− (12k 3 + 62k 2 + 104k + 56 − 26k 2 j − 89kj − 74j + 18kj 2 + 30j 2 − 4j 3 )j 3
4(k + 1)2 (2k − 2j + 1)
× ,
(k + 1 − j)3 (k + 2 − j)3
 DOMB’S NUMBERS AND RAMANUJAN-SATO TYPE SERIES FOR 1/π 11

which satisfies the relation

(4.15)
(k + 2)3 f (k + 2, j) + 2(2(k + 2) − 1)(5(k + 2)2 − 5(k + 2) + 2)f (k + 1, j)
+ 64(k + 1)3 f (k, j) = R(k, j + 1)f (k, j + 1) − R(k, j)f (k, j),
where
µ ¶2 µ ¶µ ¶
k
k 2(k − j) 2j
f (k, j) = (−1) .
j k−j j
Summing (4.15) for j from 0 to k + 2 completes the proof that Ak satisfies
(4.13).
We now turn to the proof of (1.3). First, we let Z = Z(−q). Then from
the transformation formulas [1, Entry 27(iii), (iv)]
√ p
(4.16) e−α/12 αf (−e−2α ) = e−β/12 βf (−e−2β )

and
√ p
(4.17) e−α/24 αf (e−α ) = e−β/24 βf (e−β ),
where αβ = π 2 , we deduce that

√ √
(4.18) Z(e−π/ 3N
) = N Z(e−π N/3
).
This shows that s = 12. We may now apply Theorem 2.1 with the functions
Z and X = −X(−q). Note that
∞
X
Z= Ak (−X)k ,
k=0

and hence our corresponding

k µ ¶2 µ
X ¶µ ¶
k 2(k − j) 2j
Ak = (−1)k Ak = .
j k−j j
j=0

From (4.2), we have

p
U(q) = (1 − 4X)(1 − 16X).
In addition, if
Z(q)
MN = ,
Z(q N )
then the analogue of (3.8) is
2
µ ¶2
2Y (1 − 4Y)(1 − 16Y) dX
(4.19) M2N =N 2
,
X (1 − 4X)(1 − 16X) dY
where Y = X(−q N ).
12 HENG HUAT CHAN, SONG HENG CHAN, AND ZHIGUO LIU

7
Now let N = 5. We have
(4.20)
Φ(X, Y) = X6 + Y6 + 30XY2 + 30X2 Y − 285XY3 − 285X3 Y + 970X4 Y
+ 970XY4 − 990XY5 − 990X5 Y − 930X2 Y3 − 930X3 Y2 − 25665X2 Y4
− 25665X4 Y2 + 62080X2 Y5 + 62080X5 Y2 + 92860X3 Y3 − 59520X3 Y4
− 59520X4 Y3 − 1167360X3 Y5 − 1167360X5 Y3 + 7864320X4 Y5
+ 7864320X5 Y4 − 16777216X5 Y5 − XY + 815X2 Y2 + 3338240X4 Y4 = 0.
√
When q = e−π/ 15 ,
then from (4.16) and (4.17), we find that
√ √ √
X(e−π/ 15 ) = X(e−π 5/3 ) = Y(e−π/ 15 ).
√
This shows that X(e−π/ 15 ) is a root of the polynomial Φ(T, T ) where Φ is
given by (4.20). This polynomial factors as
−T 2 (T − 1)(64T − 1)(64T 2 − 11T + 1)(64T 2 + 1)(1 + 8T )2 ,
and a quick check shows that
√ 1
(4.21) X(e−π/ 15
)= .
64
dX
Next, from (4.20), we can compute and substitute the result into
dY
(4.19). Differentiating the result with respect to X and using (4.21) and
(4.18), we conclude that
¯
dMN ¯¯
(4.22) = −128.
dX ¯q=e−π/√15

With these values, we conclude that

√
3 3 5
an = √ and bn = .
8 5 8
Substituting these values into Theorem 2.1 yields (1.3).

Acknowledgements. The authors wish to thank Professor D. Zeilberger for

his wonderful program which allows us to verify that the Domb’s sequence
satisfies the difference equation (4.13). The first author wishes to thank P.
Hammond from the University of Sussex for bringing his attention to [10].
This work is partly motivated by this excellent article by M. Kontsevich and
D. Zagier.

7This can be verified using Ramanujan’s modular equation of degree 5 in the theory of
signature 3 [2].
 DOMB’S NUMBERS AND RAMANUJAN-SATO TYPE SERIES FOR 1/π 13

References
[1] B. C. Berndt, Ramanujan’s Notebooks, Part III, Springer-Verlag, New York, 1991.
[2] B.C. Berndt, S. Bhargava and F.G. Garvan, Ramanujan’s theories of elliptic functions
to alternative bases, Trans. Amer. Math. Soc., 347 (1995), 4163-4244.
[3] B.C. Berndt and H.H. Chan, Eisenstein series and approximations to π, Illinois J.
Math., 45, no. 1 (2001), 75-90.
[4] B.C. Berndt, H.H. Chan and S.S. Huang, Incomplete elliptic integrals in Ramanujan’s
Lost Notebook, Contemp. Math., 254 (2000), 79–126.
[5] B. C. Berndt, H. H. Chan and W. C. Liaw, On Ramanujan’s quartic theory of elliptic
functions, J. Number Theory, 88 (2001), 129-156.
[6] J.M. Borwein and P.B. Borwein, Pi and the AGM, Wiley, New York, 1987.
[7] H.H. Chan, On Ramanujan’s cubic transformation formula for 2 F1 ( 13 , 32 ; 1; z), Math.
Proc. Camb. Phil. Soc., 124, (1998), 193-204.
[8] H.H. Chan, W.C. Liaw and V. Tan, Ramanujan’s class invariant λn and a new class
of series for 1/π, J. Lond. Math. Soc. (2), 64, (2001), 93-106.
[9] C. Domb, On the theory of cooperative phenomena in crystals, Advances in Phys., 9
(1960), 149-361.
[10] M. Kontsevich, D. Zagier, Periods, Math. Unlimited – 2001 and beyond, 771-808,
Springer, Berlin, 2001.
[11] M. Petkovsek, H.S. Wilf and D. Zeilberger, A = B, A K Peters, U.S.A., 1996.
[12] T. Sato, Apéry numbers and Ramanujan’s series for 1/π, Abstract of a talk presented
at the Annual meeting of the Mathematical Society of Japan, March 28-31, 2002.
[13] N.J.A. Sloane, http://www.research-att.com/ njas/sequences/index.html.
[14] S. Ramanujan, Modular equations and approximations to π, Quart. J. Math. (Oxford),
45 (1914), 350-372.
[15] K. Venkatachaliengar, Development of elliptic functions according to Ramanujan
(1988) Madurai Kamaraj University, Madurai.
[16] Y.F. Yang, On differential equations satisfied by modular forms, preprint.

National University of Singapore, Department of Mathematics, 2 Science

Drive 2, Singapore 117543
E-mail address: chanhh@math.nus.edu.sg

University of Illinois at Urbana-Champaign, Department of Mathematics,

1409 West Green Street, IL61801, U.S.A.
E-mail address: songchan@math.uiuc.edu

East China Normal University, Department of Mathematics, Shanghai 200062,

P.R. China
E-mail address: zgliu@math.ecnu.edu.cn