Ramanujan type 1/π Approximation Formulas

Nikos Bagis
arXiv:1111.3139v1 [math.GM] 14 Nov 2011

Stenimahou 5 Edessa Pellas 58200, Greece.

e-mail: nikosbagis@hotmail.gr

Abstract
In this article we use theoretical and numerical methods to
evaluate in a closed-exact form the parameters of Ramanujan
type 1/π formulas.

keywords π-formulas; Ramanujan; elliptic functions; singular mod-

ulus; alternative modular bases; approximations; numerical methods

1 Introduction
We give the definitions of the Elliptic Integrals of the first and second kind
respectively (see [9],[4]):
π/2 π/2
dt
Z Z q
K(x) = q and E(x) = 1 − x2 sin2 (t)dt. (1)
0 1 − x2 sin2 (t) 0

In the notation of Mathematica we have

K(x) = EllipticK[x2 ] and E(x) = EllipticE[x2 ]. (2)

Also we have (see [9],[7]):

dK(k) E(k) K(k)
K̇(k) = = − . (3)
dk k(1 − k 2 ) k
The elliptic singular moduli is defined to be the solution of the equation:
√

K 1 − w2 √
= r. (4)
K(w)
In Mathematica is stated as
√
r 1/2
w = k = kr = k[r] = InverseEllipticNomeQ[e−π ] . (5)
′
The complementary modulus is given by kr2 = 1 − kr2 .
Also we will need the following relation of the elliptic alpha function (see [7]):
√
 
π E(kr )
a(r) = − r −1 . (6)
4K(kr )2 K(kr )

1
The Hypergeometric functions are defined by

m+1 Fm (a1 , a2 , . . . , am+1 ; b1 , b2 , . . . , bm ; z) :=

∞
X (a1 )n (a2 )n . . . (am+1 )n z n
= , for |z| < 1, (7)
n=0
(b1 )n (b2 )n . . . (bm )n n!

and (a)0 := 1, (a)n := a(a + 1)(a + 2) . . . (a + n − 1), for each positive integer n.

2 The construction of some 1/π and 1/π 2 formu-

las
It holds
1
√
4K 2


2 (1 − 1 − z)
 
1 1 1
φ1 (z) = 3 F2 , , ; 1, 1; z = . (8)
2 2 2 π2

Consider the following equation with respect to the function φ1 (z):

1 3
∞

X
2 n n g g
z (an + b) = ⇔ bφ1 (z) + azφ′1 (z) = .
n=1
(n!)3 π π
√
√
Set w = 1/2 1 − 1 − k 2 , 1 − 2w = 1 − z = kr′ .
But
g 4K(w)(aE(w) + (b + a(w − 1) − 2bw)k(w))
bφ1 (z) + azφ′1 (z) = ⇔g= ,
π π(1 − 2w)

hence
1 3
∞

X 2 n n
4 (w − w2 )n (an + b) =
n=1
(n!)3
√ √ √
4K( w) (aE( w) + (b − a + aw − 2bw − 2bw)K( w))
= .
π 2 (1 − 2w)
For w = kr we get
1 3
∞

X 2 n n
4 (kr kr′ )2n (an + b) =
n=1
(n!)3

4K(kr ) aE(kr ) + (b − a + akr2 − 2bw − 2bkr2 )K(kr )

= . (9)
π 2 (1 − 2kr2 )
Now using the formula for a(r), in the sense that

a(r)K(kr ) π
E(kr ) = K(kr ) − √ + √ , (10)
r 4K(kr ) r

2
for suitable values for a, b, c we get the following theorem:

Theorem 2.1
1 3
∞

X
2 n n √ √
1
4 (kr kr′ )2n r(1 − 2kr2 )n + a(r) − rkr2 = (11)
n=1
(n!)3 π

Example.
1 3
Γ2 98
∞

√ √


X
2 n n 4a 5a
(40 2 − 56) (an + b) = + √ + 4(−4a + 2a + 14b)
.
n=0
(n!)3 7π 7 2π 7πΓ2 58

From which a special case is

∞ 1 3

√
X
2 n
√ n 2 1 8+5 2
(40 2 − 56) (n + − √ ) = .
n=0
(n!)3 7 7 2 14π

Theorem 2.2
∞ (2)
X Bn √ √ 1
2
(kr )2n ( rkr′2 n + a(r) − rkr2 ) = . (12)
n=0
(n!) π

Proof.
We use the function
  √
1 1 2K( z)
φ2 (z) = 2 F1 , ; 1; z = . (13)
2 2 π
Then if #2
n
"    
X 1 1
Bn(2) := n


j
j=0
2 n 2 n−j

∞ n
" #2
zn X
   
X 1 1
φ22 (z) = ... = n


j , (14)
n=1
(n!)2 j=0 2 n 2 n−j

where
∞ (2)
X Bn n
cφ2 (z) + bzφ′2 (z) + az 2 φ′′2 (z) = 2
z (an2 + (b − a)n + c)
n=0
(n!)

Hence we get
∞ (2)
2 aE(kr ) + (2b − 2bkr2 − 4a + 6akr2 )E(kr )K(kr )


X Bn 2n 2
k (an +(b−a)n+c) = +
n=0
(n!)2 r π 2 (1 − kr2 )2

3
 2(3a − 2b + 2c + (−4a + 2b − 2c)kr2 )K(kr )
+ .
π 2 (1 − kr2 )
For a = 0, b = 1, c = (−kr2 + a(r)r−1/2 )kr′−2 , we get

Theorem 2.3 Set

n
"     #3
X 1 1
Bn(3) n


:= j , (15)
j=0
2 n 2 n−j

then an 1/π 2 formula is the following

∞ (3)
X Bn 3
3
(2kr kr′ )2n (n2 + (b(r) − 1)n + c(r)) = (16)
n=0
(n!) (1 − 2kr2 )2 rπ 2

where √ √ √
3ar + r − 6a(r)kr2 − 9 rkr2 + 12 rkr4
b(r) = √
r(1 − 2kr2 )2
and √
3a(r)2 − 6a(r) rkr2 − rkr2 + 4rkr4
c(r) =
r(1 − 2kr2 )2
Proof.
Set
2 1
√
!2
16K 2 2 (1− 1 − z)

1 1 1
φ3 (z) = 3 F2 , , ; 1, 1; z = ,
2 2 2 π2
then
∞ (3)
X Bn n
cφ3 (z) + bzφ′3 (z) + az 2 φ′′3 (z) = 3
z (an2 + (b − a)n + c)
n=0
(n!)

The left hand of the above equation is a function of E(x), K(x), and can eval-
uated when we set certain values to the parameters a, b, c.

Examples.
1)
1
√ =
1200(161 5 − 360)π 2
∞ (3) √ n
   
X Bn  2 521 5 521
51841 − 23184 5 n + 1− √ n+ − √ (17)
n=0
(n!)3 288 5 12 576 5

2)

4
 1/3 (B − 489B 1/3
√

191211325848427 1010784962625383717350772720 · 22/3 4 · 2 1 2
b(163) = − −
151931373056001 151931373056001 B − √489B 1/3 151931373056001


1 2

B1 = 5680848001702137216093843898647314524189

B2 = 76896989960589381643149203281167

−5839006481108705728 + 9529627071955041072 · b(163) − 4530513053635162884 · b(163)2 +

+668649972819460401 · b(163)3 = 0

√
1/3
14178679829869760 4 C1 − 489C2 6241484569597616793758909818952 · 22/3
c(163) = − −
1/3
24764813808128163 24764813808128163 √
24764813808128163 C − 489C 3 4

C1 = 5512985602111283751597893407219881834715037026

C2 = 101526256966667546381077303112958296550

C3 = 2756492801055641875798946703609940917357518513

C4 = 50763128483333773190538651556479148275

−24380823840878077184 + 13131020889593608594752 · c(163)−

−30513780896384581928640 · c(163)2 + 17765361127840243394169 · c(163)3 = 0

∞ (3)
X 4 n Bn A
3
′
(k163 k163 )2n (n2 + (b(163) − 1)n + c(163)) = 2 (18)
n=0
(n!) π

 √
1/3 √
1/3 
4 12660947754667 + 26680 A1 − 489A2 + 26680 A1 + 489A2
A=
8254937936042721

A1 = 106866398697613339845357037

A2 = 3136555671686449089

1 266933400 10005 √
1/3

′
y163 = (k163 k163 )2 = − 1/3
+ −1 + 557403 489
16 √ 2


−1 + 557403 489

2 3
−1 + 16408588290048048 · y163 − 768 · y163 + 4096 · y163 =0

Formula (18) gives about 17 digits per term and is a formula for 1/π 2 . For
r = 253 we have another such formula which gives 21 digits per term constructed
in the same way as (18).

5
3 The study of a non usual 1/π formula
The j invariant is given by (see [17]):
 16  8 !3
η(z/2) η(z)
j(z) = + 16 , (19)
η(z) η(z/2)
√
where z = −r, r-positive real and
∞
Y
η(z) = eπiz/12 1 − e2πinz


n=1

is the Dedekind eta function.

Also
1/12
η(z) kr
= . (20)
η(z/2) 1/6
2 kr
′1/6

From√[24] section 7, Theorem 7.4 and from [11] formula (5.8), when q = e2πiz ,
z = −r, r positive real, the modular j-invariant is also given by

Q3 (q)
j(z) = 1728 . (21)
Q3(q) − R2 (q)

where
∞ ∞
X nq n X n3 q n
P (q) = 1 − 24 , Q(q) = 1 + 240
n=1
1 − qn n=1
1 − qn
and ∞
X n5 q n
R(q) = 1 − 504 .
n=1
1 − qn
The function tr is given from
 
Qr 6
tr = Pr − √ , (22)
Rr π r

where
√ √ √
r r r
Pr = P (−e−π ) , Qr = Q(−e−π ) and Rr = R(−e−π ).

i) Using Theorems 3 and 4 of [25], relation (21) equivalently can be trans-

formed to
432
j(z) = . (23)
βr (1 − βr )
Also note that we have
√ 256(1 − kr2 + kr4 )3 432
j( −r) = jr = 4
= . (24)
(kr kr )
′ βr (1 − βr )

6
Hence with our method in [25] we can simplify the known results of [24] and
[11] using the function βr , which defined as the root of the equation:
1 5


2 F1 6 , 6 ; 1; 1 − w √
1 5

= r. (25)
F
2 1 6 6, ; 1; w

ii) Set now mr := kr2 and let a(r), E(x) be the elliptic alpha function and
the complete elliptic integral of the second kind respectively (see [7],[4]), then:
 
1 6
tr = P (q) − √ =
(1 − 2βr/4 )u2r/4 rπ
!
1 E(mr/4 ) 3π 2
= 3 − 2 + mr/4 − p Fr/4
(1 − 2βr/4 )u2r/4 K(mr/4 ) 4 r/4K(mr/4 )2
or
√6 a r


1 + mr/4 − r 4
tr = q . (26)
1 − mr/4 + m2r/4 (1 − 2βr/4 )

Hence from the above evaluations and the 1/π series in [6] and [11] we get the
next reformulation:

Theorem 3.1 If we define

Jr := 1728jr−1 = 4βr (1 − βr ) (27)

1 + kr2 − √3r a (r) 1/3

2jr σ (r) G8r
Tr := p = √ √ (28)
1 − kr2 + kr4 (1 − 2βr ) r jr − 1728
then
1 5 1
∞



3 X
6 n 6 n 2 n
√ √ = (Jr )n (6n + 1 − Tr ) (29)
π r 1 − Jr n=0
(n!)3

Note. The function Gr is the Weber invariant and

√
σ(r) = 2 r(1 + kr2 ) − 6a(r)

(see [7],[5] chapter 5).

The above formulas (27), (28) and (29) can be used for numerical and theoretical
evaluations.

Similarities of formula (29) and a fifth order base formula

7
From the identity
 √ 2  
1 5 1− 1−z 1 5 1
F
2 1 , ; 1; = F
3 2 , , ; 1, 1; z , (30)
6 6 2 6 6 2
and using the following relations found in [7]:
   
π 1 1 π 1 1
Ks (x) = 2 F1 − s, + s; 1; x2 and Es (x) = 2 F1 − − s, + s; 1; x2
2 2 2 2 2 2
(31)
2 k(1 − k 2 ) dKs (t)
Es = (1 − k )Ks + K̇s , K̇s (t) = (32)
1 + 2s dt
cos(πs) √
 
π Es (xr )
as (xr ) := − r −1 , (33)
4Ks (xr ) 1 + 2s Ks (xr )
with s = 1/3 one can get, (working as in Theorem 2.1) the following Ramanujan-
type 1/π formula:
 
1

5
1
a5 (r)
X∞
6 n 6 n 2 n
β r − √
r  3
3
(4βr (1 − βr ))n ) 3n − 5 = √ ,
n=0
(n!) 1 − 2βr 2π r(1 − 2βr )

√ (34)
where the function α5 (r) = a1/3 ( βr ) is algebraic for r ∈ Q∗+ .
The parameters and the corresponding function α5 (r) of (34) are those of fifth
singular moduli base theory. Also (34) in comparison with (29) gives the fol-
lowing theorem.

Theorem 3.2
p 1 + kr2 − 3a(r)r−1/2
10α5 (r)r−1/2 = 10a1/3 ( βr )r−1/2 = 1 + 8βr − p (35)
1 − kr2 + kr4
The above formula is for general evaluation of elliptic alpha function in the
fifth elliptic base.

Also from the cubic theory as in fifth, we have

   √ 2
1 2 1 1 2 1− 1−w
3 F2 , , ; 1, 1; w = 2 F1 , ; 1; (36)
3 3 2 3 3 2
we get
∞ 1

2

1

√
X
3 n 3 n 2 n 3
[4α3 (r) − 4α23 (r)]n (n − b) = √ (37)
n=0
(n!)3 2π r(1 − 2α3 (r))
 
1/3
4 α3 (r) − a1/6 [α3 (r)]r−1/2
b= (38)
3(1 − 2α3 (r))

8
4 Examples and Evaluations
1)For r = 2
27
J2 =
125
5
T2 =
14
and √ n 
1 5 1
∞


 
15 5 X 6 n 6 n 2 n 27 9
= 6n + (39)
14π n=0
(n!)3 125 14
2) For r = 4 we have
q
11 11 1 5 1 n 
∞


 
3 X 6 n 6 n 2 n 8 10
= 6n + (40)
14π n=0
(n!)3 1331 21

3) For r = 5 we have
1  √ 
T5 = 139 + 45 5
418
√

27 −1975 + 884 5
J5 =
33275
Hence p √
21650 + 5967 5
=
π
∞ 1

5

1

√ n
!
X
6 n 6 n 2 n −53325 + 23868 5 √ 
= 836n + 93 − 15 5 (41)
n=0
(n!)3 33275

4) For r = 8 we have

√ √ 
r 
k82 = 113 + 80 2 − 4 2 799 + 565 2

√  √ 2
  q 
a(8) = 2 10 + 7 2 1 − −2 + 2 2

Then q
5
√

15 2 84125 + 81432 2
=
9982π
∞ 1

5

1

√ !n √ !
X
6 n 6 n 2 n 5643000 − 3990168 2 3276 − 1125 2 + 29946n
=
n=0
(n!)3 1520875 4991
(42)
5) For r = 18 we have
√ √
k18 = (−7 + 5 2)(7 − 4 3)

9
 √ √ √
a(18) = −3057 + 2163 2 + 1764 3 − 1248 6
1 √
α6 = (68 − 27 6)
500
√

1 7 49982 + 4077 6
β18 = − √
2 10 5 989 + 54√6
3/2
√
637326171 − 260186472 6
J18 = (43)
453870144125
√
712075 + 49230 6
T18 = (44)
1074514
Hence we get the formula giving 8 digits per term:
(Note that the number of digits per term is determined by the value of Jr ,
approximately.) p √
5 23124123365 − 13274820 6
=
1074514π
∞ 1

5
1
√ !n
X
6 n 6 n 2 n 637326171 − 260186472 6
×
n=0
(n!)3 453870144125
√
!
9 40271 − 5470 6
× 6n + (45)
1074514

6) For r = 27 s
1 1 + 100 · 21/3 − 80 · 22/3
k27 = √
2 2 + 3 − 100 · 21/3 + 80 · 22/3
1 √
   
a(27) = 3 3 + 1 − 21/3
2
a(27) is obtained from [7] page 172.

56143116 + 157058640 · 21/3 − 160025472 · 22/3

J27 =
817400375
58871825 + 22512960 · 21/3 + 13208820 · 22/3
T27 =
132566687
Hence we get the 11 digits per term formula:
s
935 935
√ √
=
π 3 761257259 − 157058640 3 2 + 160025472 3 4

∞ 1

5

1

√ √ !n
X
6 n 6 n 2 n 56143116 + 157058640 3 2 − 160025472 3 4
= ×
n=0
(n!)3 817400375

10
 √ √
!
6 12282477 − 3752160 3 2 − 2201470 3 4
× 6n +
132566687

7) From the Wolfram pages ’Elliptic Lambda Function’ and ’Elliptic Singular
Value’ we have:  √ 6  √ 
k58 = −1 + 2 −99 + 13 58

and
1  √ √  √ 6  √ 
a(58) = −70 + 99 2 − 13 29 5 + 29 −444 + 99 29
64
Also using the cubic theta identities, (see [25] relations (2),(3),(4),(30)) we eval-
uate α174 numerically to 1500 digits and then β58 to 1500 digits accuracy. We
then apply the ’Recognize’ routine of Mathematica. The result is the minimum
polynomial of β58 (this can be done also from (19) and (23)):

1 − 1399837865393267000x + 79684665286353732299517000x2−

−159369327773031733812500000x3 + 79684663886515866906250000x4 = 0.
Solving this equation with respect to x we get the value of β58 in radicals. Thus
√
1399837865393267 − 259943365786104 29
J58 = (46)
39842331943257933453125
√

5 1684967251 + 24160612 29
T58 = (47)
10376469642
The result is the formula
q
5
√

5 87 13826969809210107 − 90211316 29
=
357809298π
∞ 1

5
1
√ !n
X
6 n 6 n 2 n 1399837865393267 − 259943365786104 29
n=0
(n!)3 39842331943257933453125
 
6117973 8628790
× − √ + 6n (48)
32528118 25557807 29

which gives 18 digits per term.

8) For r = 93 (see [7] pg.158), we have

√ !3
 √
3+1 √ √ 
σ(93) = 6G−6
93 15 93 + 13 31 + 201 3 + 217 .
2

11
From [5] chapter 34 we have
√ √
1/4 √
1/6
3 3 + 31 39 + 7 31
G93 =
21/3
also
√ 1 + kr2 σ(r)
a(r) = r −
3 6
2 2
G−24
93 = 4k93 (1 − k93 )
Hence
′ 2 1
(k93 k93 ) = √ √ √
224589314596 + 129666700800 3 + 40337431680 31 + 23288826960 93
−1
√
J93 = 119562334956358303022500 + 21474029280866147440000 31+
p √
+470106000 129368095019778762513344107725 + 23235195778655878514048710848 31
√ √ √
10559116299575 + 1317692448000 3 + 275805228680 31 − 81807235875 93
T93 = ,
15081520900138
where
−1
j93 = 1728J93
This result is a very flexible formula that gives about 24 digits per term.

5 Neat Examples with Mathematica and Sim-

plicity
The class number h(−d), d ∈ N of the equivalent quadratic forms is given by
d−1  
w(d) X −d
h(−d) = − n, (49)
2d n=1 n

n


where w(3) = 6, w(4) = 4 else w(d) = 2. m , is the Jacobi symbol. Observe
that h(−163) = 1 (see [17]). For small values of h(−d) we have greater possibility
to evaluate Jd and Td in radicals.
The simplest way to evaluate the parameters J163 and T163 is again with
Mathematica.

The general algorithm is: √

i) Set r = d and k[r] = InverseEllipticNomeQ[e−π r ]1/2 , then we can evalu-
ate βr and jr from relations (19) and (23). Hence we get the value of Jr as in
section 4 example 7.
ii) For the evaluation of Tr we will need the value of a(r) which is given from
(see [7]):
√
 
π E
a(r) = − r −1 . (50)
4K 2 K

12
This in Mathematica is given from

√ EllipticE[k[r]2 ]
 
π
a(r) = 2 2
− r −1 (51)
4EllipticK[k[r] ] EllipticK[k[r]2 ]

Hence taking the package

<< NumberTheory‘Recognize‘

and
Recognize[N [J163 , 1500], 16, x]
Recognize[N [T163 , 1500], 16, x]
we get two equations. After solving them we get if r ∈ N (here r = 163), the
values of the parameters Jr and Tr in algebraic-closed forms. The results are
the π formulas.

1) We have that J163 is root of

−64 + 2552810853189232588558727380998000x − 2198253790246041723377943360187500x2 +

+224451422498574115473590775022822688001953125x3 = 0
hence
√
−1/3 √
1/3
C1 − C2 −A1 + 489B1 + 30591288 −A1 + 489B1
J163 = 4
10792555251621895860488211571345343375

A1 = 12737965652562547164590026038483234248161827096523072256574968383
B1 = 229038073182066825378006485964950394558349727761749294205546402325349
C1 = 8808429913332498766352891
C2 = 902206261147132595923169636910570558029813352485594880

From Jr = 4βr (1 − βr ), we get the value of βr and hence

√
1/3 √
1/3
12948195754365757115 + 8 A2 − B2 489 + 8 A2 + B2 489
T163 = 5
83470787671093501833
where
A2 = 3802386862487392962897493239274992371253057854289262
B2 = 3865464212119923579732688315287754932290919450

The above parameters give 32 digits per term

2) Another evaluation is taking d = r = 253:

13
 √ p √
A1 − A2 11 + 31990140 A3 − A4 11
J253 =
A5

A1 = 2804365789259959094417576921792857440357087269234369
A2 = 845548099807651569627713349319558464492321957799872
A3 = 1433462642401972199773341051748172965440271797713951
6818782945906676740858207407330990565
A4 = 43220524871261259540733172862370537466134334936322822
33926553935879770457716659641968088
A5 = 1066755353338783886372226117351012749877681799897625
and
p √ √
1875 B1 − B2 11 + 3847208393012364625 + 752271279708923520 11
T253 =
6969874104047710086

B1 = 213216899528167866600672118125
B2 = 60533150139616794053500831192

The above parameters give 41 digits per term.

Conclusion
We have given a way of how we can construct a very large number of Ramanu-
jan’s type 1/π formulas. It is true that in most cases, from r = 1 to 100 (or
higher), using Mathematica program, such formulas are very simple, as long as
h(−d) remains small and the parameters are solutions of solvable polynomial
equations.

References

[1]: M.Abramowitz and I.A.Stegun: Handbook of Mathematical Functions.

Dover Publications. (1972).
[2]: B.C.Berndt: Ramanujan‘s Notebooks Part I. Springer Verlag, New York.
(1985).
[3]: B.C.Berndt: Ramanujan‘s Notebooks Part II. Springer Verlag, New
York. (1989).
[4]: B.C.Berndt: Ramanujan‘s Notebooks Part III. Springer Verlag, New
York. (1991).
[5]: B.C. Berndt: Ramanujan’s Notebooks Part V. Springer Verlag, New
York, Inc. (1998)
[6]: Bruce C. Berndt and Heng Huat Chan: Ramanujan and the Modular
j-Invariant. Canad. Math. Bull. Vol.42(4), (1999). pp.427-440.
[7]: J.M. Borwein and P.B. Borwein: Pi and the AGM. John Wiley and
Sons, Inc. New York, Chichester, Brisbane, Toronto, Singapore. (1987).

14
 [8]: I.S. Gradshteyn and I.M. Ryzhik: Table of Integrals, Series and Prod-
ucts. Academic Press. (1980).
[9]: E.T. Whittaker and G.N. Watson: A course on Modern Analysis. Cam-
bridge U.P. (1927)
[10]: I.J. Zucker: The summation of series of hyperbolic functions. SIAM J.
Math. Ana.10.192. (1979)
[11]: Bruce.C. Berndt and Heng Huat Chan: Eisenstein Series and Approx-
imations to π. Page stored in the Web.
[12]: S. Ramanujan: Modular equations and approximations to π. Quart.
J. Math.(Oxford). 45, 350-372. (1914).
[13]: S. Chowla: Series for 1/K and 1/K 2. J. Lond. Math. Soc. 3, 9-12.
(1928)
[14]: N.D. Baruah, B.C. Berndt and H.H. Chan: Ramanujan’s series for 1/π:
A survey. American Mathematical Monthly 116, 567-587. (2009)
[15]: T. Apostol: Modular Functions and Dirichlet Series in Number Theory.
Springer
[16]: Bruce.C. Berndt, S. Bhargava and F.G. Garvan: Ramanujan’s Theo-
ries of Elliptic Functions to Alternative Bases. Transactions of the American
Mathematical Society. 347, 4163-4244. (1995)
[17]: D. Broadhurst: Solutions by radicals at Singular Values kN from New
Class Invariants for N ≡ 3 mod 8’. arXiv:0807.2976 (math-phy).
[18]: J.V. Armitage W.F. Eberlein: Elliptic Functions. Cambridge Univer-
sity Press. (2006)
[19]: N.D. Baruah, B.C. Berndt: Eisenstein series and Ramanujan-type se-
ries for 1/π. Ramanujan J.23. (2010) 17-44
[20] N.D. Baruah, B.C. Berndt: Ramanujan series for 1/π arising from his
cubic and quartic theories of elliptic functions. J. Math. Anal. Appl. 341.
(2008) 357-371
[21]: B.C. Berndt: Ramanujan’s theory of Theta-functions. In Theta func-
tions: from the classical to the modern Editor: Maruti Ram Murty, American
Mathematical Society. 1993
[22]: J.M. Borwein and P.B. Borwein: A cubic counterpart of Jacobi’s iden-
tity and the AGM. Transactions of the American Mathematical Society, 323,
No.2, (Feb 1991), 691-701
[23]: Habib Muzaffar and Kenneth S. Williams: Evaluation of Complete
Elliptic Integrals of the first kind at Singular Moduli. Taiwanese Journal of
Mathematics, Vol. 10, No. 6, pp 1633-1660, December 2006
[24]: Bruce C. Berndt and Aa Ja Yee: Ramanujans Contributions to Eisen-
stein Series, Especially in his Lost Notebook. (page stored in the Web).
[25]: Nikos Bagis: Eisenstein Series, Alternative Modular Bases and Ap-
proximations of 1/π. arXiv:1011.3496 (2010)

15