See discussions, stats, and author profiles for this publication at: [Link]

net/publication/282403574

On the complete solution of the general quintic using Rogers-Ramanujan

continued fraction

Article · September 2015

Source: arXiv

CITATIONS READS

0 944

1 author:

Nikos Bagis
Aristotle University of Thessaloniki
90 PUBLICATIONS 160 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Elliptic functions & integrals View project

All content following this page was uploaded by Nikos Bagis on 05 August 2016.

The user has requested enhancement of the downloaded file.

 On the solution of the general quintic
using the Rogers-Ramanujan continued fraction

Nikos Bagis
Stenimahou 5 Edessa
Pella 58200, Greece
bagkis@[Link]

Keywords: Quintc equation; Solution; Ramanujan; Continued fraction;

Singular modulus; Algebraic Equations; Algebraic functions; Elliptic functions;
Modular equations;

Abstract

In this article we give solution of the general quintic equation by

means of the Rogers-Ramanujan continued fraction. More precisely
we express a root of the quintic as a known algebraic function of the
Rogers-Ramanujan continued fraction.

1 Introduction and Definitions

In the most general case the quintic equation read as

ax5 + bx4 + cx3 + dx2 + ex + f = 0, a 6= 0 (1)

According to Abel’s impossibility theorem (see [22]), this equation has no general
solution. With the word solution we mean a known algebraic function of its
coefficients. For example the quadratic equation

ax2 + bx + c = 0, a 6= 0 (2)

have two solutions √

−b ± b2 − 4ac
x1,2 = (3)
2a
The cubic equation
ax3 + bx2 + cx + d = 0, a 6= 0 (4)
admits three solutions, one necessary real and two other complex in general (see
Appendix for a very amusing way of solving (4) with the help of (2)).
The quartic equation

ax4 + bx3 + cx2 + dx + e = 0, a 6= 0 (5)

1
has solution also. For higher degrees, greater than four we can not construct a
general solution.
In theory of elliptic functions (see [20]) exist functions that take algebraic values.
These functions are solutions of algebraic polynomial equations that we can not
solve. For example, the null Jacobi theta functions
∞ ∞ ∞
X 2 X 2 X 2
ϑ2 (q) := q (n+1/2) , ϑ3 (q) := q n , ϑ4 (q) := (−1)n q n , (6)
n=−∞ n=−∞ n=−∞

where |q| < 1, are forming the elliptic singular modulus

ϑ22 (q) √
kr := 2 , q = e−π r , r > 0. (7)
ϑ3 (q)
The elliptic singular modulus take algebraic values forpall r-positive rationals.
√ √ √
A few values of kr are k1 = √12 , k2 = 2 − 1, k3 = 12 2 − 3, k4 = 3 − 2 2,
r
√ √ √ √ p √
k5 = 9+4 5−2 √
38+17 5
18+8 5
, k2/3 = −3 + 2 3 + 14 − 8 3, . . .. The complexity
of these algebraic numbers is getting bigger as we move to greater and more
involved r. An example is k6/7 , which is root of

1 − 6600x + 51988x2 − 6600x3 − 103926x4 + 6600x5 +

+51988x6 + 6600x7 + x8 = 0. (8)
This happens to be symmetrical octic, hence it is solvable.
For a list of values of kr one can see [21],[15],[18] and Wolfram pages in the Web.
We may also, knowing the value of kr proceed to higher values using solvable
modular equations. For example hold the following relations ([7] chapter 18,
pg.213-219) p
1 − 1 − kr2
k4r = p , ∀r > 0 (9)
1 + 1 − kr2
and ([7] chapter 19, Entry 5, pg.230)
p p
kr k9r + kr0 k9r 0 = 1, (10)
p
where kr0 = 1 − kr2 , r > 0, ((9) and (10) are the solvable modular equations
of degree’s 2 and 3 respectively of the singular modulus).
The elliptic singular modulus kr is also the solution x of the equation
1 1 2


2 F1 2 , 2 ; 1; 1 − x √
1 1

= r, (11)
2
2 F1 2 , 2 ; 1; x

where 2 F1 (a, b; c; x) is the known Gauss hypergeometric function (see [23]). In

particular
∞ 1 2


2 π/2
  X Z
1 1 2 2 n 2n 2 dφ
2 F1 , ; 1; x = 2
x = K(x) = q (12)
2 2 n=0
(n!) π π 0 1 − x2 sin2 (φ)

2
 The 5th degree modular equation which connects k25r and kr , (see [7]) is:

kr k25r + kr0 k25r

0
+ 25/3 (kr k25r kr0 k25r
0
)1/3 = 1 (13)

The problem of solving (13) and finding k25r reduces to that of solving what
Hermite called depressed equation (see [20]):

u6 − v 6 + 5u2 v 2 (u2 − v 2 ) + 4uv(1 − u4 v 4 ) = 0, (14)

1/4 1/4
where u = k25r and v = kr .
The depressed equation is a sextic equation, which in general is not solvable in
radicals and also very important for the solution of the quintic (1). With the
help of null theta functions we can obtain a closed form solution

ϑ22 (q 5 )
k25r = . (15)
ϑ23 (q 5 )

But this is not satisfactory, since finding values of the ϑ functions, is the same
as using the kr and (14) (we have no further representations of ϑ functions to
lead us in some general theory). That is, we again need to know from the value
of kr , the value of k25r .
In this article we give a parametrization for the pairs of solutions
√ {kr , k25r }
(see (47),(48) below) in the form {A1 (R(q)), A2 (R(q))}, q = e−π r , with A1 (x)
and A2 (x) known algebraic functions and where R(q) is the Rogers-Ramanujan
continued fraction (RRCF)

q 1/5 q 1 q 2 q 3
R(q) = . . . , |q| < 1. (16)
1+ 1+ 1+ 1+
In the question ”Why the singular modulus must be replaced by the (RRCF)
?” the answer is:
As like the singular modulus, the (RRCF) obeys modular equations. The 2nd,
3rd and 5th degree modular equations of RRCF are (see [19],[11]):

2
R q 2 − R (q) 2 2


2 2 = R(q)R q , (17)
R (q ) + R (q)
 
3 
2
3
R q 3 − R (q) 1 + R(q)R q 3 = 3R(q)2 R q 3 ,


(18)
 5 1 − 2R(q) + 4R(q)2 − 3R(q)3 + R(q)4
R q 1/5 = R(q) (19)
1 + 3R(q) + 4R(q)2 + 2R(q)3 + R(q)4
Observe that the 5th degree modular equation is solved already and we have a
very strong advantage against the elliptic singular modulus k√r . Moreover as we
will show knowing for a certain r the value of R(q), q = e−π r , r > 0 we know
all k25n r , n ∈ N (although in view of [25] we can extend N into Z).

3
2 Main Results
We begin constructing the functions A1 (x) and A2 (x).


We call T1 (x) the algebraic function that relates R q 2 and R(q). Set T4 (x) to


be the algebraic function that relates R q 1/5 and R(q). Hence
 
R q 2 = T1 (R(q)) , R q 1/5 = T4 (R(q))


(20)

If we know the value of (RRCF), then we can find the value of j-invariant from
Klein’s equation (see [24] and Wolfram pages ’Rogers Ramanujan Continued
fraction’):

3
R220 − 228R215 + 494R210 + 228R25 + 1
jr = j(q) = − 5 , (21)
R25 (R210 + 11R25 − 1)
where R2 = R(q 2 ) = T1 (R(q)). Hence we set T2 (x) such
jr = j(q) = T2 R(q 2 ) = T2 ◦ T1 ◦ R(q)


(22)
From the relation
256(kr2 + kr04 )3
jr = , (23)
(kr kr0 )4
we can set T3 (x) such that
kr = T3 (j(q)) = T3 ◦ T2 ◦ T1 ◦ R(q). (24)
Note here that (23) is solvable with respect to kr in radicals. In the method
described above one must be very careful which branch of the solution (of the
polynomial equation) is using.

Knowing kr and kr/25 , we can evaluate k25r (see [25]) and specify relations
of the form



k25r = Ψ kr , kr/25 and k25n r = Ψn kr , kr/25 , n ∈ N (25)
The next theorem describes in details the relations (25).

Theorem 1. (Bagis [25])

For all n = 1, 2, . . . we have
v v
u " s #24
u n
kr0 kr0 0
u
u u
2 Y
k25n r0 = t1/2 − 1/2 1 − 4 kr0 kr0 0
t P (j) 12 , (26)
j=1
kr0 /25 kr0 0 /25

where the function P is in radicals known function and is given from

 h i1/6 
∗ 6
P (x) = P [x] = U Q U [x] (27)

4
 P (n) (x) = (P ◦| .{z
. . ◦} P )(x) , P (1) (x) = P (x). (28)
n

The function Q is that of (29) and U , U ∗ are that of (32) and (33) below.

We give the analytic expansion of Q:

 1 2
5
−1 − e 5 y + e 5 y
Q(x) =  1 2 3 4 6 7 8 9
 (29)
e 5 y − e 5 y + 2e 5 y − 3e 5 y + 5ey + 3e 5 y + 2e 5 y + e 5 y + e 5 y

and  
11 + x
y = arcsinh . (30)
2
Note that (for no confusion), the function exp (rational · y) is algebraic function
of x.
Consider now the following equation
√
X2 5Y 1
= √ Y 3 − Y −3


√ − 2
(31)
5Y X 5
This equation is solvable in radicals with respect to Y and X also. One can find
the solution
s
5 25 X4 h(X)
Y = U (X) = − 2
+ 2
+ + (32)
3X 3X h(X) h(X) 3X 2

where  √ p 1/3
h(x) = −125 − 9x6 + 3 3 −125x6 − 22x12 − x18

The solution of (31) with respect to X is

s √
1 Y4 1 + 18Y 6 + Y 12
X = U ∗ (Y ) = − 2 + + . (33)
2Y 2 2Y 2

We can simplify Theorem 1 replacing it with an equivalent transformation.

Set p such that
p !!
∗ 6 p + 4 + p2
U (x) = −11 + 2 sinh 5 log (34)
2

then
11 + U ∗ (x)6
  
1
p = 2 sinh arcsinh (35)
5 2

5
and (29) becomes
(p − 1)5
Q(p) = (36)
11 + 6p + 6p2 + p3 + p4
Hence if s
3 (p − 1)5
s= (37)
11 + 6p + 6t2 + p3 + p4
√
then P1 = P (x) = Y , where Y is root of
Y 3 + 5s−1 Y 2 − sY − 1 = 0. (38)

Lemma 1.
Given kr0 and kr0 /25 , if
s !6
kr0 kr0 0
α = U∗ 12
(39)
kr0 /25 kr0 0 /25

and   
1 11 + α
p = 2 sinh arcsinh (40)
5 2
then s
3 (p − 1)5
s= (41)
11 + 6p + 6t2 + p3 + p4
Hence s !
kr0 kr0 0 p
P 12
= Y (s) (42)
kr0 /25 kr0 0 /25
where
Y 3 + 5s−1 Y 2 − sY − 1 = 0, Y = Y (s) (43)

From the above Lemma 1 we get the value of k25r knowing kr and kr/25 :

Proposition 1.
v v
u u s !12
u

u 1 1u (p − 1)5
k25r = Ψ kr , kr/25 = − 1 − 4(kr kr0 )2 · Y 3
t t
2 2 11 + 6p + 6p2 + p3 + p4
(44)

Example.
In case of r0 = 25, we need the value of
s
0
! √ !
k r 0
kr 0 r0 =25 5−1
P1 = P 12 = P
kr0 /25 kr0 0 /25 2

6
 √ √ √

6  √ 6 
But α = U ∗ 5−1
2 = −2 + 5, hence Q U ∗ 5−1
2 = Q −2 + 5 .
With √ √
x = a + b d = −2 + 5
we get s
1/2
 √ 1/6 6 (p − 1)5
s = Q −2 + 5 =
11 + 6p + 6p2 + p3 + p4
where √ !!
1 9+ 5
p = 2 sinh arcsinh
5 2
Hence we take the value of k625 :
v v
u u s !12
√
u
u1 1u 3 (p − 1)5
k625 = − 1 − (51841 − 23184 5) · Y
t t
2 2 11 + 6p + 6p2 + p3 + p4

For to make things more clear and simple we use the function m(q) which is
related with kr by √
m(q) = kr , q = e−π r (45)
and for complex values is
p
m(q) := k log(q)

2 , m0 (q) = 1 − m(q)2 . (46)
π

Proposition 2.
The solution of the depressed equation is

l = kr = (T3 ◦ T2 ◦ T1 ) (vr ) , s = kr/25 = (T3 ◦ T2 ◦ T4 ◦ T1 ) (vr ) (47)

and
k25r = Ψ(l, s) (48)
√
−π r
where vr = R(q), q = e , r > 0.

Example.
Suppose r = 16 following classical theory we have
√ √
q
k16 = 33 + 24 2 − 4 140 + 99 2

The value of k25·16 = k400 is symmetric and of degree 16, which is quite difficult
to evaluate. But it happens that
√ s √
−2π

1 + 5 5+ 5
R e =− +
2 2

7
hence
v16 = R e−4π = T1 R e−2π

and from (47) the values of l = k16 and s = k16/25 are known. From (44) we
find the value of k400 (in radicals).

Hermite has proved (see [20]) the next

Theorem 2. (Hermite)
If
x5 − x + a = 0 (49)
with
2(1 + m(q)2 )
a= √
4
p (50)
55 m(q)m0 (q)
then
Φ(τ )
x= √
4
(51)
2 53 m(q)1/4 m0 (q)
where
 1/4    τ + 16   
5 1/4


1/5 τ + 64
Φ(τ ) = m q +m q φ −φ ×
5 5
    
τ + 32 τ + 48
× φ −φ (52)
5 5
and s
1/4 ϑ2 (q)
φ(τ ) = m(q) = , q = eπiτ , Im(τ ) > 0.
ϑ3 (q)

In the above Hermite’s theorem we can write Φ equivalently as

  1/4    1/4 

1/4
Φ(τ ) = −4 m q 5 + m q 1/5 Re m −(−q)1/5 ×

"s p #
4 1− 1 − c21
×Im p , (53)
1 + 1 − c21
where
 Re(z), Im(z) are the real and imaginary part of z respectively and c1 =
√
1/5 
m − − q .
Therefore we have to evaluate
   
m −(−q)1/5 = m −(−1)1/5 q 1/5 (54)

8
A first result is
m(q)
m(−q) = p (55)
m(q)2 − 1
which means that we can define the function T5 such that
m(−q) = T5 (m(q)) . (56)
We will use
the theory of Ramanujan Quantities (see [26]) for the evaluation
of m (−q)1/5 .

If v = R(1, 2, 5; q) = R(1, 3, 5; q) = R(q) is the Rogers Ramanujan continued

fraction and u = u(q) = R(1, 3, 10; q), then we conjecture that
?
u3 − uv + u2 v 3 + v 4 = 0 (57)
Note. In general its an observation that different Ramanujan Quantities are
related with polynomial relations (see [10],[26]):
2
1 − H2 (1 − T )(3 + T )3

kr2 = 16H 2
= , (58)
1 + H2 (1 + T )(3 − T )3
√
where T = 1 − 8V 3 and V = V (q) = R(1, 3, 6; q) is the Ramanujan’s cubic
continued fraction. The quantity H = H(q) = R(1, 3, 8; q) is the Ramanujan-
Gollnitz-Gordon continued fraction.

The u−quantity satisfies Theorem 8 of [26], since 2|10 and (1, 2) = 1, (3, 2) =
1, we have
?
u |u∗ | = u2 = u q 2


(59)
where u∗ = u(−q).
Another evaluation, using Mathematica is
−v + (−1)1/5 v 0 − (−1)1/5 v 5 v 0 + 5(−1)2/5 v 4 v 02 − 10(−1)3/5 v 3 v 03 +
?
+5(−1)4/5 v 2 v 04 + vv 05 + v 6 v 05 − (−1)1/5 v 5 v 06 = 0 (60)
0
where v = |R(−q)| = v(−q).
Relation (60) follows with the same method as (57) (see Routine pages 8-9 of
[26]).

Lemma 2.
If R(q) is the Rogers Ramanujan continued fraction, then
?
R(1, 3, 10; q) = R(q)R q 2


(61)
Proof.(Partial)
It holds
∞
X qn X
log(R(a, b, p; q)) = Q log(q) − X(a, b, p; d)d, (62)
n=1
n
d|n

9
Multiplying with n and summing we get
∞ ∞ ∞
X (n) X X qn X
log(R(a, b, p; q)) = Q log(q) (n) − τ (d)(n/d) (63)
n=1
n n=1 n=1
n
d|n

We make use of the properties of the Dirichlet multiplication (see [3]).

Assume that exists arithmetic function X1 (n)n such that

X1 (n)n ∗ 1 = τ (n) ∗ (n) = X(n)n ∗ 1 ∗ (n),

then
X1 (n)n = (n) ∗ X(n)n (64)
Assuming the X(n) and X1 (n) correspond to Ramanujan Quantities we seek
(n) such that is always 0 after a few terms. An example is when X(n) =
X(1, 3, 5, n) and X1 (n) = X(1, 3, 10, n). Then
? ?
(n) = 1, if n = 1, 2 and (n) = 0, if n > 2 (65)

From the above, using (59),(63),(65), we have the partial proof of (61).

Lemma 3.
i) The function w = R(q)R2 (q 2 ) can always evaluated in radicals with respect
to v = R(q). The equation that relates the two functions is
p
(w − 1) R(q)w + R(q)3 (1 + w) = 0. (66)

Moreover
ii) The function v 0 = |R(−q)| can always evaluated in radical function of R(q).

Proof.
The proof of (i) follows from u√2 = R(1, 3, 10; q)2 = R(q)R(q)R2 (q 2 ), hence
w = u2 /R(q) = u2 /v, or u = wv. Setting this identity to (57) we get the
result.
The proof of (ii) follows from (61) and (57):

u(q 2 ) = u(q)|u(−q)| = R(q)R(q 2 )|R(−q)|R(q 2 ) = |R(−q)|w

Suppose the solution of (57) is u(q) = f1 (R(q)), then u(q 2 ) = f1 (R(q 2 )) =

|R(−q)|w, hence


f1 R q 2
|R(−q)| =
w
and the proof is complete.

From the above Lemmas we get the next

10
Proposition 3.

? f1 (T1 (R (q)))
R(−q) = (−1)1/5 2 = T6 (R(q)) (67)
R(q)T1 (R(q))
where f1 (x) is the solution u = R(1, 3, 10; q) = f1 (R(q)) of (57).

Note that the proof is not complete since we not prove (57),(59) and (65),
but we include it to show how we arrive to (67).

Proposition√ 4.
Let q = e−π r with r real 15 < r < 25, then
 
m −(−q)1/5 = T3 ◦ T2 ◦ T1 ◦ T4 ◦ T6 ◦ R(q) (68)

and  √ 1/5 
m − (− q) = T3 ◦ T2 ◦ T1 ◦ T4 ◦ T6 ◦ T7 ◦ R(q) (69)

Main Theorem. √
If 15 < r < 25, q = e−π r , evaluate t from (77) and (78), then
√
!2
(−1) K 1 − t2
r = ki (t) = k (t) = (70)
K(t)

and
t = m(q) = kr = T3 ◦ T2 ◦ T1 ◦ R(q). (71)
Also  
l = m q 1/5 = kr/25 = T3 ◦ T2 ◦ T4 ◦ T1 ◦ R(q) (72)

and if s √
∗ −1 x4 1 + 18x6 + x12
U (x) = + + (73)
2x2 2 2x2
6
√
 s
12 t 1 − t2 
α = U∗  √ (74)
l 1 − l2
  
1 11 + α
p = 2 sinh arcsinh (75)
5 2
s
(p − 1)5
s= 3 , (76)
11 + 6p + 6p2 + p3 + p4
the solution of
x5 − x + 0 a = 0, 0 = ±1 (77)

11
with
2(1 + t2 )
a= √
4
√√ (78)
5 5 t 1 − t2
is q 
8
p 1 1
−4 − 2 2 12
1 − 4t (1 − t )Y + T3 ◦ T2 ◦ T4 ◦ T1 ◦ R(q)
2 2
x= √4
√ ×
2 53 t1/4 1 − t2
"s p #
4 1 − 1 − c21
×Re [T3 ◦ T2 ◦ T1 ◦ T4 ◦ T6 ◦ R(q)] Im p , (79)
1 + 1 − c21
Y is the root of equation

Y 3 + 5s−1 Y 2 − sY − 1 = 0 (80)

and
c1 = T3 ◦ T2 ◦ T1 ◦ T4 ◦ T6 ◦ T7 ◦ R(q) (81)

Note. The function T3 will be each time one of T31 , T32 , T33 , T34 of paragraph
2.1 below

2.1 Formulas for T1 , T2 , T3 , T4 , T5 , T6 , T7

In this section we give closed form evaluation of algebraic functions Ti , i =
1, 2, . . . , 7.
√


x2 1 − i 3 −3x − x6
T1 (x) := − − √ √
1/3 +
3 6x 18x4 + x9 + 3 3 −x3 + 11x8 + x13
√
√ √
1/3
1 + i 3 18x4 + x9 + 3 3 −x3 + 11x8 + x13
+ (82)
6x

3
x20 − 228x15 + 494x10 + 228x5 + 1
T2 (x) := − 5 (83)
x5 (x10 + 11x5 − 1)

2
16T34 (x) − 8 =
q √
2/3 √
1/3
6i i + 3 D1 + 12D1 (−576 + x) + −6 − 6i 3 D1 (−1536 + x)x
− 1/3
6D1
(84)

16T33 (x)2 − 8 =

12
 q √
2/3 √
1/3
6i i + 3 D1 + 12D1 (−576 + x) + −6 − 6i 3 D1 (−1536 + x)x
1/3
(85)
6D1

r
1/3 √ q 2/3 1/3 2/3 1/3
24D1 + 3 −576D1 + D1 − 1536D1 x + D1 x + D1 x2
T32 (x) = √ 1/3
4 3D1
(86)

16T31 (x)2 − 8 =
q √
2/3 √
1/3
−6 − 6i 3 D1 + 12D1 (−576 + x) + 6i i + 3 D1 (−1536 + x)x
= 2/3
6D1
(87)
where √ p
D1 = 884736x − 2304x2 + x3 + 12288 3 1728x2 − x3 (88)

!1/5
x 1 − 2x + 4x2 − 3x3 + x4
T4 (x) := (89)
1 + 3x + 4x2 + 2x3 + x4

x
T5 (x) := √ (90)
x2 −1

f1 (T1 (x))
T6 (x) := (−1)1/5 (91)
xT1 (x)2
where
√


x3 1 − i 3 −3x − x6
f1 (x) := − + √ √
1/3 −
3 6 −18x4 − x9 + 3 3 −x3 + 11x8 + x13

1 √  √ p 1/3
− 1 + i 3 −18x4 − x9 + 3 3 −x3 + 11x8 + x13 (92)
6

√


1 1 + i 3 −1 + 3x5
T7 (x) := − 2 − √ √
1/3 +
3x 6x2 1 − 18x5 + 3 3 −x5 + 11x10 + x15
√
√ √
1/3
1 − i 3 1 − 18x5 + 3 3 −x5 + 11x10 + x15
+ (93)
6x2

13
3 Appendix
3.1 The Tschirnhausen transform for the quintic
For to reduce the general quintic

x5 + ax4 + bx3 + cx2 + dx + e = 0 : (eq)

in the Bring quintic form we use the following Tschirnahausen transforms

First use a quadratic transform

y = x2 + Ax + B (94)

with √
4a4 − 13a2 b + 15ac ± 5∆1
A= (95)
4a3 − 10ab
and √
5a2 b − 20b2 + 15ac ± 5∆1
B= , (96)
20a2 − 50b
where

∆1 = −3a4 b2 + 12a2 b3 + 8a5 c − 38a3 bc + 45a2 c2 + 16a4 d − 40a2 bd (97)

and write (eq) in the ”principal form”

x5 + A1 x2 + B1 x + C1 = 0, (98)

where

A1 = r = 3aAbB−3A2 bB−3b2 B−6a2 B 2 +6aAB 2 +12bB 2 −10B 3 −aA2 c+A3 c+

+Abc + 6aBc − 9ABc − c2 − 3aAd + 4A2 d + 2bd − 6Bd − 2ae + 5Ae (99)

B1 = t = −3aAbB 2 +3A2 bB 2 +3b2 B 2 +4a2 B 3 −4aAB 3 −8bB 3 +5B 4 +2aA2 Bc−

−2A3 Bc−2AbBc−6aB 2 c+9AB 2 c+2Bc2 −aA3 d+A4 d+A2 bd+6aABd−8A2 Bd−

−4bBd + 6B 2 d − Acd + d2 − 4aA2 e + 5A3 e + 3Abe + 4aBe − 10ABe − 2ce (100)

C1 = s = aAbB 3 −A2 bB 3 −b2 B 3 −a2 B 4 +aAB 4 +2bB 4 −B 5 −aA2 B 2 c+A3 B 2 c+

+AbB 2 c+2aB 3 c−3AB 3 c−B 2 c2 +aA3 Bd−A4 Bd−A2 bBd−3aAB 2 d+4A2 B 2 d+

+2bB 2 d − 2B 3 d + ABcd − Bd2 − aA4 e + A5 e + A3 be + 4aA2 Be − 5A3 Be−
−3AbBe − 2aB 2 e + 5AB 2 e − A2 ce + 2Bce + Ade − e2 (101)

14
Then again apply a Tschirnhausen transform (quartic)

y = x4 + kx3 + lx2 + mx + n, (102)

to reduce
x5 + rx2 + tx + s = 0 (103)
into the ”Bring-Jerrard” form

x 5 + A2 x + B 2 = 0 (104)

Set

∆5 = −27r6 s2 + 256r2 s5 + 108r7 t − 1600r3 s3 t + 2250r4 st2 + 3125r2 t4 , (105)

then we will have

√
−27r3 s + 400s2 t − 375rt2 − 3 5∆5
k= (106)
54r4 − 320s3 + 600rst
√
18r3 s2 − 45r4 t − 250rst2 + 2s 5∆5
l= (107)
r (27r4 − 160s3 + 300rst)
√
135r4 s − 1280s4 + 3600rs2 t − 1125r2 t2 − 9r 5∆5
n= (108)
10 (27r4 − 160s3 + 300rst)
and m is root of a 3rd degree equation
√ √ √
3 M1 + N1 5∆5 2 M2 + N2 5∆5 M3 + N3 5∆5
rX + X + X+ (109)
S1 S2 S3
where
M1 = 162r7 − 1104r3 s3 + 2295r4 st + 1875r2 t3 , (110)

N1 = 16s2 − 15rt, (111)

S1 = 54r5 − 320rs3 + 600r2 st. (112)

M2 = 4374r12 − 59859r8 s3 + 205440r4 s6 + 124902r9 st − 869760r5 s4 t+

+102400rs7 t + 946350r6 s2 t2 − 688000r2 s5 t2 + 81000r7 t3 + 510000r3 s3 t3 +

+1040625r4 st4 + 1250000rs2 t5 , (113)

N2 = −783r5 s2 + 5120rs5 + 486r6 t − 12960r2 s3 t + 4725r3 st2 − 10000s2 t3 (114)

2
S2 = 2r 27r4 − 160s3 + 300rst . (115)

15
 M3 = −196830r17 + 4056885r13 s3 − 27812160r9 s6 + 61649920r5 s9 +

+13107200rs12 − 8496495r14 st + 117564615r10 s4 t − 400708800r6 s7 t−

−73728000r2 s10 t − 126918900r11 s2 t2 + 909441000r7 s5 t2 − 35200000r3 s8 t2 −
−4829625r12 t3 − 651358125r8 s3 t3 + 426800000r4 s6 t3 − 133650000r9 st4 +
+130500000r5 s4 t4 + 160000000rs7 t4 − 1065234375r6 s2 t5 + 125000000r2 s5 t5 −
−10546875r7 t6 − 828125000r3 s3 t6 − 263671875r4 st7 , (116)

N3 = 50301r10 s2 − 675648r6 s5 + 2298880r2 s8 − 22599r11 t + 1658475r7 s3 t−

−10376000r3 s6 t − 516375r8 st2 + 13329000r4 s4 t2 − 640000s7 t2 −

−2362500r5 s2 t3 − 4600000rs5 t3 − 84375r6 t4 +
+13375000r2 s3 t4 − 2109375r3 st5 (117)

S3 = 10r(27r4 − 160s3 + 300rst)3 (118)

And

A2 = 5n4 − 2m3 nr + 9lmn2 r − 12kn3 r + 2l3 nr2 − 6klmnr2 + 6m2 nr2 + 9k 2 n2 r2 −

−9ln2 r2 − 2k 3 nr3 + 6klnr3 − 6mnr3 + 2nr4 + 5lm3 s − 10l2 mns − 10km2 ns+
+15kln2 s + 15mn2 s − 2l4 rs + 6kl2 mrs + 3k 2 m2 rs − 9lm2 rs − 14k 2 lnrs+
+16l2 nrs + 2kmnrs − 12n2 rs + 2k 3 lr2 s − 6kl2 r2 s + 6lmr2 s + 2knr2 s−
−2lr3 s + 5k 2 l2 s2 − 5l3 s2 − 5k 3 ms2 − 5klms2 + 5m2 s2 + 10k 2 ns2 +
+10lns2 − 2k 3 rs2 + 4klrs2 − 7mrs2 + 2r2 s2 + 5ks3 + m4 t − 8lm2 nt+
+6l2 n2 t + 12kmn2 t − 16n3 t − l3 mrt + 3klm2 rt − 3m3 rt + 2kl2 nrt−
−10k 2 mnrt + 4lmnrt + 15kn2 rt + k 3 mr2 t − 3klmr2 t + 3m2 r2 t−
−2k 2 nr2 t + 4lnr2 t − mr3 t + kl3 st − 7k 2 lmst + 3l2 mst + 13km2 st+
+6k 3 nst − 4klnst − 22mnst − 3k 4 rst + 11k 2 lrst − 4l2 rst − 10kmrst+
+16nrst + 3kr2 st + k 2 s2 t − 6ls2 t + l4 t2 − 4kl2 mt2 + 2k 2 m2 t2 +
+4lm2 t2 + 8k 2 lnt2 − 8l2 nt2 − 16kmnt2 + 18n2 t2 − k 3 lrt2 + 3kl2 rt2 +
+k 2 mrt2 − 5lmrt2 − 2knrt2 + lr2 t2 + k 3 st2 − 3klst2 + 7mst2 − 4rst2 +
+k 4 t3 − 4k 2 lt3 + 2l2 t3 + 4kmt3 − 8nt3 − krt3 + t4 , (119)

16
 B2 = −n5 + m3 n2 r − 3lmn3 r + 3kn4 r − l3 n2 r2 + 3klmn2 r2 − 3m2 n2 r2 −

−3k 2 n3 r2 + 3ln3 r2 + k 3 n2 r3 − 3kln2 r3 + 3mn2 r3 − n2 r4 + m5 s−

−5lm3 ns + 5l2 mn2 s + 5km2 n2 s − 5kln3 s − 5mn3 s − l3 m2 rs+
+3klm3 rs − 3m4 rs + 2l4 nrs − 6kl2 mnrs − 3k 2 m2 nrs + 9lm2 nrs+
+7k 2 ln2 rs − 8l2 n2 rs − kmn2 rs + 4n3 rs + k 3 m2 r2 s − 3klm2 r2 s+
+3m3 r2 s − 2k 3 lnr2 s + 6kl2 nr2 s − 6lmnr2 s − kn2 r2 s − m2 r3 s+
+2lnr3 s − l5 s2 + 5kl3 ms2 − 5k 2 lm2 s2 − 5l2 m2 s2 + 5km3 s2 −
−5k 2 l2 ns2 + 5l3 ns2 + 5k 3 mns2 + 5klmns2 − 5m2 ns2 − 5k 2 n2 s2 −
−5ln2 s2 + k 3 l2 rs2 − 3kl3 rs2 − 2k 4 mrs2 + 6k 2 lmrs2 + 3l2 mrs2 −
−7km2 rs2 + 2k 3 nrs2 − 4klnrs2 + 7mnrs2 − l2 r2 s2 + 2kmr2 s2 −
−2nr2 s2 + k 5 s3 − 5k 3 ls3 + 5kl2 s3 + 5k 2 ms3 − 5lms3 − 5kns3 − k 2 rs3 +
+2lrs3 − s4 − m4 nt + 4lm2 n2 t − 2l2 n3 t − 4kmn3 t + 4n4 t + l3 mnrt−
−3klm2 nrt + 3m3 nrt − kl2 n2 rt + 5k 2 mn2 rt − 2lmn2 rt − 5kn3 rt−
−k 3 mnr2 t + 3klmnr2 t − 3m2 nr2 t + k 2 n2 r2 t − 2ln2 r2 t + mnr3 t+
+l4 mst − 4kl2 m2 st + 2k 2 m3 st + 4lm3 st − kl3 nst + 7k 2 lmnst−
−3l2 mnst − 13km2 nst − 3k 3 n2 st + 2kln2 st + 11mn2 st − k 3 lmrst+
+3kl2 mrst + k 2 m2 rst − 5lm2 rst + 3k 4 nrst − 11k 2 lnrst+
+4l2 nrst + 10kmnrst − 8n2 rst + lmr2 st − 3knr2 st − k 4 ls2 t+
+4k 2 l2 s2 t − 2l3 s2 t + k 3 ms2 t − 7klms2 t + 3m2 s2 t − k 2 ns2 t+
+6lns2 t + klrs2 t − 3mrs2 t + ks3 t − l4 nt2 + 4kl2 mnt2 − 2k 2 m2 nt2 −
−4lm2 nt2 − 4k 2 ln2 t2 + 4l2 n2 t2 + 8kmn2 t2 − 6n3 t2 + k 3 lnrt2 −
−3kl2 nrt2 − k 2 mnrt2 + 5lmnrt2 + kn2 rt2 − lnr2 t2 + k 4 mst2 −
−4k 2 lmst2 + 2l2 mst2 + 4km2 st2 − k 3 nst2 + 3klnst2 − 7mnst2 −
−kmrst2 + 4nrst2 − ls2 t2 − k 4 nt3 + 4k 2 lnt3 − 2l2 nt3 −
−4kmnt3 + 4n2 t3 + knrt3 + mst3 − nt4 (120)

Lastly with a simple change of scaling we can reduce (104) further into the
”Bring quintic” form
x5 + x + t = 0 (121)

17
3.2 The Bring radicals hypergeometric function
The Bring quintic (121) is solvable (not necessary in radicals), using the argu-
ments of Lambert and Euler, (see [5] pg.306-307). Since it has the form

aqxp + xq = 1, (122)

the next theorem solves (122) and hence (121) and (1) in terms of hypergeo-
metric functions.

Theorem 3.2 (Lambert-Euler)

The equation (122) admits root x such that
∞
n X Γ({n + pk}/q)(−qa)k
xn = , n = 1, 2, 3, . . . (123)
q Γ({n + pk}/q − k + 1)k!
k=0

where Γ(x) is Euler’s the Gamma function.

As a result of Theorem 3.2 we can define the hypergeometric function

3125t4
    
1 2 3 4 1 3 5
BR(t) = −t · 4 F3 , , , ; , , ;− . (124)
5 5 5 5 2 4 4 256

Then the solution of (121) will be BR(t). Hence the solution of (104) is
!
p
4 B2
x = A2 · BR 5/4
. (125)
A2

For the solution x of (104) also hold the next nested radicals expansion
s r
√
5
q
5 5
x = −B − A −B − A −B − A 5 −B − . . . (126)

3.3 The solvability of RRCF

The same holds for the sextic equation (see [16]):

b2
+ by + ay 2 = cy 5/3 (127)
20a
Set
Y = Y (q) = Yr = R(q)−5 − 11 − R(q)5 , (128)
then (127) has solution

b √
y= Y (q 2 ) = , q = e−π r , r > 0, (129)
250a

18
 3
where r can evaluated from the coefficients, using the relation jr = 250 ac2 b .
The j-invariant can be expressed in terms of the Ramanujan-Dedekind eta func-
tion as " 16  8 # 3
−1/24 η(τ ) 1/24 η(2τ )
jr = q + 16 q . (130)
η(2τ ) η(τ )
The Ramanujan-Dedekind eta function is given from
∞
Y
η(τ ) = (1 − q n ) , q = eπiτ , Im(τ ) > 0 (131)
n=1

and special values can be found from the relation (see [9],[21])

28/3 −1/3 √
η(τ )8 = 4
q (kr )2/3 (kr0 )8/3 K(kr )4 , q = e−π r , r > 0 (132)
π
But most appropriate is to use (23) directly.

3.4 The Cubic Equation

Suppose we have to solve the equation

x3 = 3ax + b (133)
We assume that the solution is of the form
√3
√
3
x= A+ B (134)

Then √ √ √  √
x3 = A + B + 3 AB
3 3 3 3
A + B = A + B + 3 ABx

Hence we have to solve

A + B = b and AB = a3 (135)

which reduces clearly to quadratic equation

X 2 − bX + a3 = 0 (136)

References

[1]: [Link]. ’The summation of series of hyperbolic functions’. SIAM J.

Math. Ana.10.192, 1979.
[2]: [Link]. ’Number Theory’. Dover Publications, New York, 1994.
[3]: [Link]. ’Introduction to Analytic Number Theory. Springer Verlang,
New York, Berlin, Heidelberg, Tokyo, 1974.

19
 [4]: [Link] and [Link]. ’Handbook of Mathematical Functions’.
Dover Publications, 1972.
[5]: [Link]. ’Ramanujan‘s Notebooks Part I’. Springer Verlag, New
York, 1985.
[6]: [Link]. ’Ramanujan‘s Notebooks Part II’. Springer Verlag,
New York, 1989.
[7]: [Link]. ’Ramanujan’s Notebooks Part III’. Springer Verlag,
New York, 1991.
[8]: [Link], Sen-Shan Huang, Jaebum Sohn and Seung Hwan Son.
’Some Theorems on the Rogers-Ramanujan Continued Fraction in Ramanujan
Lost Notebook’. Trans. Amer. Math. Soc. 352 (2000), 2157-2177
[9]: [Link] and [Link]. ’A course on Modern Analysis’. Cam-
bridge U.P., 1927.
[10]: Nikos Bagis. ’The complete evaluation of Rogers Ramanujan and other
continued fractions with elliptic functions’. arXiv:1008.1304v1 [[Link]], 2010.
[11]: Michael Trott. ’Modular Equations of the Rogers-Ramanujan Contin-
ued Fraction’. page stored in the Web.
[12]: Nikos Bagis. ’The First Derivative of Ramanujan Cubic Fraction’.
arXiv:1103.5346v1 [[Link]], 2011.
[13]: Nikos Bagis. ’Parametric Evaluations of the Rogers-Ramanujan Con-
tinued Fraction’. IJMMS, Vol. 2011.
[14]: Nikos Bagis. ’The w-modular function and the evaluation of Rogers-
Ramanujan continued fraction’. International Journal of Pure and Applied
Mathematics. Vol 84, No 1, 2013, 159-169.
[15]: [Link] and [Link]. ’Pi and the AGM’. John Wiley and
Sons, Inc. New York, Chichester, Brisbane, Toronto, Singapore, 1987.
[16]: Nikos Bagis. ’On a General Sextic Equation Solved by the Rogers-
Ramanujan Continued Fraction’. arXiv:1111.6023v2 [[Link]], 2012.
[17]: Soon-Yi Kang. ’Ramanujan’s formulas for the explicit evaluation of the
Rogers-Ramanujan continued fraction and theta functions’. Acta Arithmetica.
XC.1, (1999).
[18]: [Link]. ’Ramanujan Notebooks Part V’. Springer Verlag, New
York (1998).
[19]: [Link], Heng Huat Chan, Sen-Shan Huang, Soon-Yi Kang,
Jaebum Sohn and Seung Hwan Son. ’The Rogers-Ramanujan Continued Frac-
tion’. J. Comput. Appl. Math., 105 (1999), 9-24.
[20]: [Link], [Link]. ’Elliptic Functions’. Cambridge Univer-
sity Press. (2006)
[21]: J.M. Borwein, M.L. Glasser, R.C. McPhedran, J.G. Wan, I.J. Zucker.
’Lattice Sums Then and Now’. Cambridge University Press. New York, (2013).
[22]: Heinrich Dorrie. ’100 Great Problems of Elementary Mathematics
(Their History and Solution)’. Dover Publications, inc. New York. (1965)
[23]: N.N. Lebedev. ’Special Functions and their Applications’. Dover Pub.
New York. (1972)
[24]: W. Duke. ’Continued fractions and Modular functions’. Bull. Amer.
Math. Soc. (N.S.), 42 (2005), 137-162.

20
 [25]: Nikos Bagis. ’Evaluation of Fifth Degree Elliptic Singular Moduli’.
arXiv:1202.6246v1. (2012)
[26]: Nikos Bagis. ’Generalizations of Ramanujan’s Continued Fractions’.
arXiv:11072393v2 [[Link]] 7 Aug 2012.

21

View publication stats