CHAPTER I

L-FUNCTIONS AUTOMORPHIC FORMS ARITHMETIC

L-FUNCTIONS:

Suppose you are given an interesting sequence a (n), n

, of complex numbers that

you would like to investigate. The method of analytic number theory is to encode this
sequence in a generating function. There are several choices, and if some multiplicatively is
involved, one might consider the Dirichlet series.

a (n )
L( s ) s
n 1 n
(1.1.1)

a ( n ) n
Let

for all

,
n | a ( n) | x1

Or even only an average bound

s 1,

Then (1.1.1) converges absolutely and uniformly on compacta in

And thus defines a holomorphic function. We can hope that in this way we translate the
arithmetic of the sequence a (n) into analytic properties of the function L(s), and indeed there
is very often a remarkable interplay between arithmetic and analysis.
1) Let a(n)=1 for all n.

The Riemann function (. . -zeta function)

1
s
n 1 n

( s)

1 1
p
s
p

=
1

s 1,

The Euler product Show that

vanishing of

on the line

(s) 0 in

>1, and it is a classical fact that the non-

=1 is equivalent to the prime number theorem

:
(x) = {p <x/p prime} x/log x,
x

The
-function can be extended meromorphically to all of , and one has more
s

precisely an equivalence2

(S)0 in s >1- for some 0<1/2

x

dt
0 1x
log t
2

( x )

and all

for some 0< 1/2

This show a very precise translation of an arithmetic statement (distribution of prime

numbers) into an analytic statement

1 Let k/ be a number field and

Let a (n)={integral ideals a/Na=n}.

This gives the Dedekind -function

(s) =

n 1

a (n )
ns
1

(N
a

)s

=
Which has a simple pole at s=1
The analytic class number formula states
r

2 1 2 r2 Rh
w |D|

ress=1

k(s)=

where as usual r1,r2 are the number of real resp.

Pair of complex embedding of K into , R is the regulator, h is the class number,

w is the number of roots of unity in K and D is the discriminant

Let

be a primitive Dirichlet character to some large modulus q, this gives rise to a

Dirichlet L-function

(n )
s
n 1 n

L( s, )

Which again can be continued to an entire function in practice, one often encounters

n x ( n ),
character sums of the type

And one would expect that there is a lot of

L( s , )

cancellation in such a sum. the lindelof hypothesis for

Such that
L(1 / 2 it , ) ((1 | t |)q)
0
for all
(n) x1/2
n x

for all

and for all x > 0. Again there is an intimate connection between an

arithmetic statement.

If E/

is an elliptic curve, we known by mordells theorem that the set of rational points

on E is a finitely generated abelian group

E ( ) Z r Etor ( )

The rank r seems to be an elusive object, however, it is relatively simple to count points
on (the reduction of) E over finite fields, and we can define

aE ( p )

p 1 E ( Fp )
p

For a prime p. This can be extended in a more or less natural way to all integers,
and yields an L-function

L E ( s ) aE ( n )n s
Given that LE()exists, that the rank can be recovered from the Laurent expansion
at , namely ords=1/2LE(s)=r
Every decent L-function has a functional equation of the from
_

L( s )G ( s ) L(1 s )G (1 s )
(1.1.2)

Where

=1

and
d

s j

G ( s ) N s /2 s / 2
j 1

for some integer N

and some complex numbers 1,.. d. The complexity of an L-

function is measured by its analytic conductor

C c (t )
d

N (1 | t j |),
j 1

t s.

( 1.1.3)

Since we are assuming that L(s) converges absolutely in s>1,

We have L(s) < 1 in s=1+ the functional equation (1.1.2) and stilings formula

translate this into3 L(s) < C+

on

s= -

If we assume in addition that L is of finite order , which is always satisfied in

applications, then a standard argument shows
L( +it) < c(t) 4+

(1.1.4)

This is usually referred to as the convexity bound, and any exponent smaller than 4 is
called a subconvexity bound. If the generalized Riemann hypothesis holds for the L-function in
question, then 4 can be replaced with 0.

1.2 AUTOMORPHIC FORMS ON GL2:

Let G=PSL2(R) we have the lwasawa decomposition G=NAK where
5

 1 x

N
x R

1/2
y
A

y 0
y

1/2

cos
K

sin

sin

[0,

cos

R
The group K=PSO2

is a maximal compact subgroup of G

0 ( N )

Let
a b

PSL2 (Z) | c 0(mod N )

c d

Then G acts on L

/G

by the rignt regular representation, and we have a G-equivariant

decomposition.

L2

G .1 v H (t)dt

Into the constant functions, cuspidal irreducible representations (, V) and Eisenstein

series for the cusps a (that enter the picture because

is not compact).Each V decomposes

further according to the characters of K.

V V , q
q2 Z

and known that dim V , q 1 . The Left and right G-invariant Laplace operator

y 2 2 x 2 y yx 0

On each V as a scalar

in algebraic terms, this is the Casimir element (up to

normalization) of the universal enveloping algebra U(g). sometimes we need a variant of the

L2
space

For a character x of modulus dividing N.

.

L2
Let

denote the L -space of function G

2

a b
as f
g (d ) f ( g )
c d

for

that transform under

a b

c d

L2
In general there is no left action of G an

if F is

-invariant, then f(g) is only

g-1 g invariant.

However if =

subgroup of

(N) and g

PSL2(

), then g-1

g contains some finite index

(N)and using a suitable average, we can get back to

This is a special feature of group like

(N).

(N) (as opposed to arbitrary discrete subgroups

of G) and yields a family of naturally defined operators {Tn|n

} that forms a commutative

algebra, which also commutes with .

since is left G-invariant

Mostly for technical reasons we consider only the subspace L2 new ( \G) whose
irreducible representations are generated by so-called newforms, They do not come from

V L2new

subgroups with smaller index in PSL2( ). Then each operator Tn acts on each

V , q

( n )
a scalar

and a function

has a Fourier expansion.

as

 y1/2

0

( n )
w
|n|
n 0

eiq

xy 1/2 cos

y 1/2 sin

sin

cos

1 / 4 4 | n | y e( nx )
sgn(n)q/2,

e 2 ix w p,
Where as usual e(x) denotes the additive character

1,0,1

( n ) (n )
and

is the function

with

( n )
This is relevant for us, because the Eigen valves
information

Examples.1

carry often number theoretic

K Q

Let N be the discriminate of the field

and x the character of the

extension K/Q

Then there is a representation (, V) c L

G,

new

with

1 / 4

such that

(n )
{ideals

p
QK/N =n}

Let E/Q be an elliptic curve

Then there is a representation ( ; v)

L new

With

=0,

such that

p
(p)=(p+1=#E(Fp))/

T and G in any case for each representation (, V) c L new

we can define an L-function

(n )
ns
n 1

L , s

10

for all primes P

And we hope to learn more about eigen values by studying L (, s) from an analystic
point of view. If we work over a number field K/Q with class number h>1, the about setup is not

appropriate One could work with h copies of G modulo certain conjugates of

, but it is better

to work adelically for each place u of k

let Kv be the completion and Ou the ring of integers (if v; then Ov=Kv).Then the Adele
ring is the restricted product

kv

With respect to the sets Ov, with K embedded diagonally. There is a natural surjection from
Ax to the group of non-zero fractional ideals of K, and we often do not distinguish between an
idele and its image.Again GL2(A) acts by the right regular representation on L2(Ax GL2 (K)\G
L2(A)). where Ax is identified with the center of GL2(A).
This setting has no dependence on the level of the subgroup any more, since it treats

simultaneously all subgroup

o(c) with c an ideal in K.

If we want to make the level explicit, we define

k (c) k c p
p . finite

11

 GL2 ( fin )

for a non zero ideal c,

a b
x
k ( c ) :
GL2 ( K p ) | a , d O p , b 1, c p c, ad bc o p

c d

Where

With

the different of K.
The conductors of an irreducible representation (, V) contained in

L2 AxGL2 K \ GL2 A
is the smallest ideal Such that V contains a right K(c)-invariant vector
(which is automatically a newvector).
The Fourier expansion in the adelic setting reads
y x


1

rk x

y y X
When

ry
fin

N ry fin

fin

Ax
,

12

W ry rx ,

A,

a smooth vector in some cuspidal irreducible representation , N the norm,

ry fin

a product of function,

the standard additive character on A, and

depends only

on the fractional ideal represented by ryfin and is non-zero if this ideal is integral. A typical way is
twisting, and the simplest twist is by a character (that is by an Automorphic form on GL 1).
Let X: Kx\AxS1 be a conductor q, (q is the largest ideal such that x is trivial on finite

ideals

1|mod q), and define the twist of a representation on GL2 with X by

x ( g ) x (det g ) ( g )

This is another representation on GL2, and if the integral ideal a is co-prime to the
conductors of and x , then

x (a ) (a ) x (a )

If x has conductor q and has conductor c co-prime to q, then
conductor of the character enters quadratically.

13

conductor cq2, so the

1.3 SUBCONVEXITY FOR AUTOMORPHIC L-FUNCTIONS:

Theorem 1.3.1:
Let K be a totally real number field of degree d, an irreducible cuspidal

representation on GL2(A),

C C (t, )
is a conductor q and

L(s,) in the sense of (1.1.3)

Then
14

the analytic conductor of

L ,1 / 2 it C A N q

1/2

Where N denotes the norms, A is some obsolute constant, >0 is arbitrarily small
and

111

In general, and

More precisely the constant

1 8 if k
1
in the general case5 is

Where is the bound towards the ramanujan conjecture and

The convexity bound in this context is

L , 1 it = t , , N q
2

1 2

1
11

= 0 is also conjectured.

1
2

The first sub convex bound is this direction was

18
our bound
matches the quality,

where the case a Eisenstein series is treated

1 22

for K=Q.

Theorem 1.3.2:
We will only sketch briefly the ideas that go into the proof, it rests on the
following ingredients.
The amplification and an approximate functional equation
The spectral decomposition of Dirichlet series associated with a shifted convolution

sum which makes good use of

The kirillov model and Soboba norms
15

Let us look at the first point.

To start with, we have to find a way to work conveniently with the values

L ^ , 1 it
2

Since a priori they exist only by analytic continuation.

However, often a suitably truncated part of a divergent or conditionally convergent

series gives a good approximation of the quantity that one is interested in.
For L- function this can be made precise with an approximate functional equation, and in
fact the first about C terms of an L- function with analytic conductor C are a very good

approximation on the critical line

s 1 / 2

In other words for all practical purposes

L 1 , :
2

a a

Na : Nq

Na

(1.3.1)

and similarly for other points on the critical line,

where has to be understood in a very broad sense.

L s,
Now, we have an explicit description of

as finite sum, Let us try to exhibit

cancellation in such sums.

16

| sin x cos x | 2
Suppose you want to prove that

. There are containly many ways of

proving this.
Here is one: square the left hand side and add a spectrally useful non negative quantity

| sin x cos x |2 | sin x cos x |2

| sin 2 x cos 2 x 2sin x cos x sin 2 x cos 2 x 2sin x cos x |

| 2sin 2 x 2 cos 2 x |

| 2 sin 2 x cos 2 x

| 2 |

In a similar way, it is useful to embed an L-function into a family. Let us cut the equation
(1.3.1) into h pieces according to the ideal class of the ideal a. For simplicity we will only work
with the principal class.
We consider now the second moment.

| L( , 1 2 it ) |

17

Where is a family of characters containing x. For example the family of characters of

(0/q)x
These characters are in general not Hecke characters, because they may not be trivial

on units, and so strictly speaking the expression

L , 1 it
2

does not make sense as a

value of an Automorphic L-function.

It is typical in this context to consider such fake moments; after all, we are free to

add to our original quantity

L , 1 it
2

Here the expression

whatever we want as long as it is non-negative.

L , 1 it
2

is just a notation for a suitable Dirichlet

p p
polynomial whose coefficients behave roughly like

p
on principal ideals ( ).This

family is of size about Nq: so even.

If we assume that the Generalized lindelo f hypothesis in the sense

L , 1 it
2

we can abound the above sum only by Nq which after taking the square root just
recovers the convexity bound.

18

The problem here is that our family, although very convenient to work with, is quite
large.One could try to evaluate a fourth moment instead of a second moment, in which case one
could take the fourth root at the end, but our current analytic techniques are not strong enough to
estimate a fourth moment appropriately.

We consider

A() |

L , 1 it
2

(1.3.2)

Where A (w) is an amplifier that is large for w=x, and rather small otherwise.

In practice, A ( ) will be a short Dirichlet polynomial,

Example:

(p)( p)

N ( p): L

Where L is a parameter that we can optimize later.

Now we open the square and sum over . This show that we have to bound nontrivially
sums roughly of the form

19

n1 ,n2Ok nB
1n1 2n2 (mod q )

( n1 ) ( n2 )

(1.3.3)

p1 , p 2
where
n1 , n 2 ,...n

are of norm about L, and B is a box in minkowski space of the form

d

( Nq)

1 , 2 ,... d
with

the embeddings of K into R. we break this sum into pieces according to

the value of

q 1 n 1 p 2 p 2 1

(1.3.4)

The term q=0 is the diagonal term, and pretty straight forward to handle.
Let us now assume q0 Expressions of the type (8) with a summation condition of (9) are
usally called shifted convolution sums.

Dq ( s )

( n 1 ) ( n2 )

n1 n2 q

(n 1 n 2 ) s

s 1
a

nd proved in some cases an analytic continuation to some right half plane

20

 0
with

.
However, without good control of the size in s and q Our method, based on the Kirillov

model and Sobolev norms gives not only the analytic continuation with good growth estimates
,but also the pleasing structural insight that the series Dq(s) can be decomposed according to the
decomposition (5)

Theorem 1.3.3:
Let k>60 and q>0 be any integers, and (n) eigenvalues of any irreducible cuspidal
representation on GL2 of conductor 1. Then there exist holomorphic functions F in the strip
< s < 3/2
Such that

21

1 s
( n ) ( m)( nm) ( k 1)/2
q 2 ( q) F ( s ) d

s k 1
(n m)
m n q

s1
,

and

( s ) | d | s | 22

1 3
s
2
2

Where the integral is taken over the union of the discrete spectrum and the continuous
spectrum. Armed with theorem 1.3.2 (or rather a slight generalization thereof) it is relatively
straight forward to complete proof of the proof of theorem1.2.1

1.4APPLICATIONS:
Although at first right theorem 1 may seem as some purely analytic trickery, it has, in
accordance with the general philosophy of L-functions, interesting arithmetic applications. And
its extensions to number fields. More precisely,
22

Let

be a cuspidal representation on the double cover SL2, generated by a half

integral weight modular form satisfying some technical assumptions, and

the representation on

GL2 Given by theta correspondence (shimura lift) .

_

Then for squarefree m, the square of the m-th Fourier co-efficient of to

L m , 1
2

Where xm is the quadratic character corresponding to the extension K (m)/k

In this way theorem 1.2.1 yields the currently best know bounds for Fourier co-efficient of half
integral weight Hilbert modular forms.
One particular situation where such bounds are needed, are asymptotic formulae for the
number of representations of totally positive integers by ternary quadratic forms, Hilberts
eleventh problem asks more generally which integers are represented by a given n any quadratic
form Q over a number field K.
If Q is a binary form, it corresponds to some element in the class group of a quadratic
extension of k, If Q is indefinite at some Archimedean place, if Q is positive definite at every
Archimedean place and n4 again siegels mass formula and simple bounds for fourier coefficients of Hilbert modular forms give a complete answer (some care has to be taken in the case
n=4).The only remaining case of Q positive definite and n=3.
Theorem 1.4.1:

23

Let K be a totally real number field and let Q be a positive integral ternary quadratic
form over K. Then there is an ineffective constant C>0 such that every totally positive square

m Ok
free integer

with Nm C is represented integrally by Q if and only if it is integrally

represented over every completion of K.

The representation of non-square free integers is quite subtle, but in principle can again
be characterized by more involved local considerations Theorem 1.4.1 can be refined in various
ways and also made quantitative, which yields for example applications of the following type,
GauB proved in his Disquisitiones that a rational integer n can be written as a sum of three
squares if and if only it is not of the form 4 k(8m+7) , and if it is in addition not divisible by a
very high power of 2, the number of such representations is about

L(1,

)n which by siegels theorem is n+0(1)

Hence one may ask if all integers satisfying some natural congruence conditions can

still be written as a sum of three squares of numbers with certain restrictions example sums of
three squares of primes, or sums of three square square free numbers or sums of three squares of
smooth numbers etc.,

24

Theorem 1.4.2 :

Let n 3 (mod 24) 5+n, be sufficiently large, and let

=1/567.

Then n is the sum of three squares of integers with all their prime factors greater than

number of such representations exceeds

? n1/2

. The

in particular every such n is the sum of three

squares haveing at most 284 prime factors. Under the assumption of a subconvex bound as above
it is proved that a certain family of Heegner points and certain d-demensional subvarieties are
equidistributed on the Hilbert modular variety PSL2(Ok) |Hd.

for example If K=

, and D is the discriminate of an imaginary quordiratic field, then each

ideal class

a = (a,(b--D) ), say, in the class group of

X PSL2 ( Z) / H

Z (b D
Gives a Heegner point

(-D)

/(2a) in

If D, these points become denser and denser in X, and the above statement says that
they becomes actually equidistributed with respect to the standard measure y-2 dx dy on X.
In order to prove this, one has to sum the values of a test function at these heegner points, and by
a spectral decomposition one can assume that the test function is an eigenform the laplacian
This leads to certain weyl sums, which can be expressed as central values of twisted L-functions
using bounds for the L-value as in theorem1 one derives an equidistribution statement,

25

Finally we note that the subconvex bound in theorem 1.2.1 is a crucial input for certain
input for certain subconvex bounds of higher degree L-functions which in turn have other
arithmetic applications.

CHAPTER II

A NOTE ON THE SECOND MOMENT OF

AUTOMORPHIC L-FUNCTIONS
2.1 INTRODUCTION:
In this paper, we study the Wised second moment of the family of L-function
arising from S*2(q), the set of primitive Hecke eigenform of weight 2; Lever q (q prime).

f ( z ) s2* ( q)
For

f has a Fourier expansion

f ( z ) n 2 f (n )e(nz )
1

n 1

where the normalization is

Such that

f (1) 1

The L-function associated of f has an Euler product

f (n)
ns
n 1

L( f , s )

1 f (q)

qs

(h ) 1

1 f s 2 s
hpeime
h
h
hq

s1
The series is absolutely convergent when

all of . the functional equation for L(f, s) is

26

, and admits analytic continuation to

q
( f , s )
2

s 1 2 L( f , s )

f ( f ,1 s )
1

f q 2 f ( q)
where

2) We define the harmonic average as

Af

Af
f s * ( q ) 4 ( f , f )

where (f, g) is the petersson inner product on the space o(q)/H. we are interested in the
twisted second moment of this family of L-functions.
We define

s( p, q)

L f,1

f s 2* ( q )

( p)
2
2

Our main theorem is

Theorem 2.1.1:
Suppose q is prime and 0<p Cq for some fixed C < 1
Then we have
1
d ( p)
q
s ( p, q )
log 2 0 p 2 q 1
4r p
p

Remark 2.1:
The twisted harmonic fourth moment has been considered, where they gave an

p q1/9
asymptotic formula for the fourth power mean value provided that
Remark 2.2:
In a similar setting, the twisted second moment of the Automorphic L-function
arising form k(1), the set of newforms in Sk(1), where Sk(1) is the linear space of
holomorphic cusp forms of weight k. precisely they showed that for K>2

K 0 (mod2) and for any m1

We have

27

2
12
d (m )
f L f , 1 2 f (m) 2(1 i k )

k 1 f H k (1)
m

2 i k
m

1
log 2 m
0 l k/2 l

2 i k
h
d (h)d (h m) pk m m
hm

d (h)d (h m)q
k

where pk(x) and qk (x) are Hankel transforms of Bessel functions

pk ( x) y0 y dy
0

qk ( x)

2
_

k y x J
0

k 1

( y )dy

f (2) L( sym 2 ( f ),1) 1

3) Here the weight

L( sym 2 ( f ), s )
Where the symmetric square L-function

corresponding to f is defined by
f (n 2 )

L( sym 2 ( f ), s ) (2 s )
n 1

ns

In the context of Dirichlet L-functions,

Consider
1
M ( p, q ) *
* | L(1/ 2, ) |2 ( p)
(q) X (mod q )

* ( q)

Where * denotes summation over all primitive characters

(mod q), and

the number of primitive characters.This is the twisted second moment of Dirichlet Lfunction. Proved that there is a kind of reciprocity formula relating M (p, q) and M (-q, p)
When p and q are distinct prime integers.
p log q log pq
p
1
p
B
M ( p, q )
M ( q, p )
(log A)
0

q
q
q
p
2 q
pq
q
28

is

Where A and B are some explicit constants.This provides an asymptotic formula for

M ( p, q)

p M ( q, p)
q

pq

under the condition That

p q1
So that the asymptotic formula holds for
We now take p to be prime and, similarly as before, S(q, p) is defined as the harmonic

g ( q),
second moment, twisted by

of the family of L-function arising

g ( z) S ( p)
*
2

From

s ( q, p )

h L( g , 1 2 )

gs2* ( p )

log p
Thus as a trivial consequence of theorem 2.1.1 for p<q.
we have
1 1
2
q
s( p, q) p s( q, p)
log 2 0( p 2 q 2 )
q
4 p
p
This leads to an asymptotic formula for S (p,q) -p/q S (q, p) at least for p as large as

1
2

The results in the Dirichlet L-functions suggest that the asymptotic formula should

p q
hold
4

for any

However, our technique fails to extend the range to any power for that purpose,
we need more refined estimates for the off-diagonal terms of S(p, q) and S(q, p).

29

The intricate calculations seem to suggest that there is large cancellation between
these two expressions.The nature of this is not well understood.

2.2 PRELIMINARY LEMMAS:

We require some lemmas.
Lemma 2.2.1:
For m, n 1

30

 f ( m) f ( n)

mn
f 2
d
d /( m ,n )

( d , q ) 1

The next lemma is a particular case of Petersons trace formula

Lemma 2.2.2:
For m, n 1
We have

( m ) f ( n ) m ,n q

f s2* ( q )

Where

m, n is the kronecker symbol and

s(m, n; cq ) 4 mn
J1

cq
cq
c 1

J q ( m, n ) 2^

Here J,(x) is the Bessel function of order 1, and S(m,n;c) is the sum

s( m, n; c)

* e

a (mod c )

ma n a
c

Moreover, we have
1

J p ( m, n ) ( m, n, q) 2 ( mn)

31

The above estimate follows easily from the bound J,(x)<x

Lemma 2.2.3:
Let f: R+->C be a c function which vanished in the neigh bowrhood of 0 and is
rapielly decreasing at infinity then for C 1 and (a, c)=1

x
am
f
(
m
)

2
f ( x )dx
log

c
c
0

c d ( m )e
m 1

2 ^ d ( m ) e
m 1

a m 4 mx
y0 c f ( x)dx
c
0

4 d ( m)e
m 1

a m 4 mx
0 k0 c f ( x )dx
c

Lemma 2.2.4:
Let G(s) be an even entire function satisfying G(0)=1 and G has a double zero at

each

Furthermore let assume that G(s) < A, B(1+|s|)-A for any A > 0 in any strip

B s B

f s2* ( q)
then for

L f,1

d (n) f (n) 4 2 n
Wq

n
n 1
q

32

where

Wq ( x )

1
ds
G ( s) ( s 1) 2 (2s 1) x s

s j (1)
s

Here

(s) is defined by

q (s)

n s ( 1)

n 1
( n , q ) 1

Proof from lemma 2.1.1 we first note that

L( f , s) 2 (2s)
n 1

d ( n ) f ( n )
ns

( 1)

Consider
G ( s) ( f , s 1 ) 2 ds
1
2
A( f )

2 i (1)
s
q
2

s 1
Moving the line of integration to

and applying Cauchys theorem and the

functional equation,
We know that A (f) = L (f, )2 A(f)
Expanding A (f, s+1/2)2 in a Dirichlet series and integration term wise we obtain the
lemma for our purpose, Wq is basically a cut-off function indeed we have the following.

33

Lemma 2.2.5:
The function Wq satisfies
Wq ( j ) ( x ) j, Nx N
for

x 1

j, N 0
and all

x jWq ( j ) ( x ) i, j | log x |

for

0 x 1

i j0
and all
and

1 log x log q
Wq ( x) 1

0N xN
q
2
q

for

0 x 1

and all

N 0

The implicit constants are independent of q

Proof:
The first extimate is a direct consequence of stirlings formula after differentiating

s N
under the integral sign and shifting the line of integration to
34

The only difference in the other two estimates is that one has to move the line of

s N
integration to

2.3 PROOF OF THEOREM 2.1.1:

From lemma 2.2.4 and lemma 2.2.2
We obtain

s ( p, q )

h L( f , 1 2 )
2

( p)

f s2* ( q )

where

L f,1

L f,1

d ( p) f ( p) 4 2 p
Wq

p
n 1
q

d ( p ) 4 2 p
Wq

p
q

where

35

n 1

( p)

n 1

( p ) 2R ( p, q)

s ( p, q ) 2

d ( p ) 4 2 p
Wq
2 R ( p, q )
p
q

4 2 n
d (n)
q (n, p )Wq

n
n 1
q

R ( p, q )

where
Using lemma2.2.5 the first term is

d ( p)
q
1
log 2 0( p 2 q 1 p1/2 q 1 )
4 p
p

Thus we are left to consider R(p, q)

We have

s( n, p; cq ) 4 np
J1

cq
cq
c 1

J p (n, p) 2

4 2 n
d (n) ( n, p; cq 4 np
R( p, q) 2
s cq J1 cq Wq q
n c 1
n 1

where

Using weils bound J,(x) < x, the contribution from the terms c qis

p 2q

(n, p)
n 1

36

4 2 n
d (c )
| 3
q cq c 2

d (n) | Wq

p 2 q 1

Thus we need to study

2
q

1
a p
*

c q c a (mod cq ) cq

an

cq

d (n)e
n 1

: R

4 np
cq

J1

Wq(
n

[0,1]

We fix a C function

4 2 n
)
q

( x) 0
which satisfies

for 0 x and

(n)

( x) 1
for x 1, and attach the weight

to the intermost sum.

Using Lemma2.3 This is equal to

_
4 ^ tp
4 2t
4
1
t
dt

s (o, p; cq ) log
J,
Wq
(t ) y k

2 2

q cq c
cq
cq
t
q
0

Where

4 2
q2

cq

d (n) s(o, p n; cq

n 1

4 tp
4 At
4 2t
dt
J
Wq

(
t
)

1
cq
cq
t
q

y0
0

(2.3.1)

37

and

8
q2

1
d (n) s (o, p n; cq

2
c q c n 1

4 tp
4 nt
4 2t
dt
J
Wq
(t )

t
q
cq
cq

k0
0

(2.3.2)

We will deal with Y and K in the next three lemmas.for the first sum, since S(o,p;cq)=M(q)

s (0, p, q; c
) and j,(x) <x

4 2t
1
p q 2 | Wq
| (log tcq)dt
q
c q c 1

2
1

p 2 q 2

Lemma 2.2.6:
For k defined as in 2.3.2
We have

38

k p 2 q ( q p ) 1

k p 2 q 1
and hence

p cq
Given that

for some fixed

c 1

Remark 2.3:
This is the only place where the condition p Cq for some constant C < 1 is used
Proof:
The integral involving Ko
k0 ( y ) y

2 y

Using

is

_
_
_2

4
^
nt
4
^
tp
4 ^ t
dt

k
J
Wq
0 0 cq 1 cq q (t ) t

cq
k0 ( y )J1

2 n 0

4 nt
cq

Where

39

p
c 2 qy 2
y
Wq

n
4n

c 2 q 2 y 2
dy

2
16 n

cqy
_2

4^

t
n

c2q2 y 2
_2

16 ^ n

cp 2 q1

y 2 e y dy

cq

ep 2 q1

e
n

n /2 cq

s(0, p n; cq) s(0,( p n ) q; c ) s(0, p n; q)

Thus as
_

| s (0, ( p n) q; c) | l /( p n, c) l

d (n )
e
n
n 1

k p 2 q 1
1

2 q2

l / ( p n, c) l
c

| s (0, p n; q) |
cq

d ( n )d ( p n )
e
n
n 1

p 2 q 1
1

40

2 q2

| s (0, p n; q) |

We break the sum over n according to whether q|(p +n) or q(p +n).
1

0( p 2 q 1 )
The contribution of the latter is

. That of the former is

d ( l ) d ( ql p )
e
ql p
l 1

p 2 q
1

ql p

2 q2

p 2 q ( q p ) 1 p 2 q 1

The lemma follow The case of y is more complicated as yo is an oscillating function for that we
need the following standard lemma.
Lemma 2.7 :
Let V0 and J be a positive integer. If f is a compactly supported Cfunction on

0
[y, 2y], and there exists
y j f ( i ) ( y ) j (1 y ) j
Such that

for o j J, then for any

>1
1

y p 2 q 1
We have
Proof:
We have
41

4 2
1
d (n)s(0, p n; cq) y(n)
q 2 c q c 2 n 1

(2.3.3)

where
_
_
_2

4
^
nt
4
^
tp
4 ^ t
dt

y ( n ) cq y0
J
Wq

(t )
1

cq
cq
t
0

(2.3.4)

k , k
We make a smooth dyadic partition of unity that

where each

k
supported C function on the dyadic interval [xk, 2xk]. moreover,

is a compactiy

x j (k j ) ( x ) 1
satisfies

for all

k
j0 we work on each

individually, but we write instead of and, accordingly, x rather than xk

by the change of variable x= 2t/cq

c2q2 x2
y (n) cq y0 (2 nx ) J1 ( c qx ) wq ( c qx ) (
)dx
4
0
2 2

2 2

We have

f ( x) J1 (2 px )Wq( 2c 2 qx 2 ) (
We define
42

c2q2 x2
)dx
4

2 x

| , 2 |

This is a C function compactiy supported on

cq

Where

We first treat the case x q.We note that this involves o (log q) dyadic intervals.
From lemma 2.5 We have
Xj W(j) (x) < j log q
for 1/q <x < 1 This together with the recurrence relation
(xv Jv (x))1 = xv jv-1(x)
Gives

Xj f(j) (x) <j (1 + px)j log q.

_

p 2 ^ n ,

(2.3.5)

We are in a position to apply lemma 2.7 to f with

and

The lemma yield for any positive integer J.

J

1 p
y ( n ) cq
log q
1

(2.3.6)

Later, we will break the sum over n in (2.3.3) in the following way

n 1

n p k

43

n p k

y 2 x / cq

Where k > 2 will be chosen later. The estimate( 2.3.6) will be used for
n k

need another estimate for the range

n k

We

For this we go back to( 2.3.4) using 1/0 (x)< 1+ |

log x| and J,(X) < x

px
(log q ) 2
cq

y (n )

(2.3.7)

To derive
We denote by Y1 and Y2 the corresponding splitted sums (Y=Y1+Y2). For the first sum,
using (2.3.7. We have
_2

Y1

4^
q

c
2

c q

d ( n ) s(0, p n; cq ) y ( n )

n k

p 2 Xq 3

n k

p 2 Xq 3

n q
2 X

d (n ) | s (0, p n; q) |
k

2 X l k c q
n
q

p 2 q 1

( p n ,c )

n q
2 X

44

1
3
c q c

d (n ) | s(0, p n; q) |

( p n ,c )

1
c3

d ( n)d ( p n )
k

| s (0, p n; q) |

p 2 q 2 k 5

(2.3.8)

for the second sum, we note that p < 1 in this range Using (2.3.6)

y (n ) x (log q)n

We have
Similliarly we can
_2

Y2

4^
q

c d (n)s(0, p n; cq) y (n)

c q

n k

d (n)
1
| s(0, p n; q ) | 2 l J
J
c q c l
n k n
( p n ;c )
2
d (n)
X ( J 1)/2 q J 2
| s (0, p n; q) | l J 1 C J 2
J
n
l / p n
2 x lk
n 2
c
n

X q 2

ql

q 1
n

d ( n )d ( p n )
n

( J 1) / k

| s (0, p n; q) | (2.3.9)

45

We choose

k 2 /2

and J large enough so that J/2-(J-1)/k>1,We hence obtain

Y1 p 2 q 1

Y2 p 2 q 1

and, since the sum over n in (2.3.9) converges,

for x > q ||| ly to (2.3.5), using the bound xj W(j) (x) < j x-2
We have
Xj f(j) (x) < j (1+ px)j q-2 (cx)-4

Lemma 2.7:
Then gives
J

1 p q 2
y (n ) cp
2
1 n x

For the range n p-k, a better bound than (2.3.7) in this case is

y (n)

px
q2
(log q) 2
cq
x

Since q2/x2 < 1 all the previous estimates remain valid, the one place where this is not
the case is the sum over n< (q2/2x)k in (2.3.8)

46

However this sum is void for x >q4/4 and the former estimate still words in the large
interval x q2/4. Also the quantity saved q2/x2 is sufficient to allow the sum over the dyadic
values of x involved to coverge.

CHAPTER III

THE FOURTH POWER MOMENT OF AUTOMORPHIC

L-FUNCTIONS FOR GL (2) OVER A SHORT INTERVAL
3.1 INTRODUCTION:
For the Riemann zeta function and Dirichlet L-functions, estimates for their power
moments on the critical line Re s=1/2 played central roles in analytic number theory. Classical
results on short intervals.

47

k k

1 4
it | dt K
2

where proved for =7/8 for any > 0. In this paper, we want to prove a similar result for
Automorphic L-functions attached to a certain holomorphic or maass cusp form g for 0(N).
To describe our results, we need bounds towards the Ramanujan conjecture for maass
forms. In terms of representation theory, let

be an Automorphic cuspidal representation of

( j )
GL2(QA) with unitary central character and local Hecke eigenvalues

( j )
(p) for p < and

() for p = ,
j = 1,2. Then bounds toward the

| ( j ) ( p ) | p
For p at which

| Re( ( j ) ( )) |

is unramified,

(3.1.1)

if is unramified at .

These bounds were proved for 0=1/4, for 0=1/5, for 0=1/9, and most recently for 0=7/64
The Automorphic L-functions we will consider are

g ( n )
ns
n 1

L ( s, g )

(1 g ( p ) p s p 2 s ) 1 (1 g ( p ) p s ) 1 ,
p/ N

48

p/N

where g is a holomorphic or maass cusp Hecke eigenform for 0(N) , and its twist by a real

primitive character

Q
modulo

N /Q:
with

g ( n ) ( n )
ns
n 1

L( s, g x )

(1 ( p )g ( p ) p s p 2 s ) 1
p /Q

for Re s > 1.
Q
We assume that

is odd and square-free, and x is the real, primitive character modulo

Q
, i.e., the Jacobi symbol, so that the twisted cusp form gx is a cups form for 0(N2). As point
p.1176,gx is primitive even if the Hecke eigenform g itself is not primitive. Our results,
L( s, g x )
nevertheless, are valid in other cases, as long as the twisted L-function

has a

standard functional equation in particular, our theorem below is valid for L(s,g) when N=1, we
will assume that g is self-contragredient. If g is holomorphic, we denote its weight by l. if g is
maass, we denote its Laplace eigenvalue by +l2.

THEOREM 3.1.1:

49

Let g be a fixed self-contragredient holomorphic or maass Hecke eigenform for

Q
0(N), and x a real, primitive character mod
K L

N /Q:
with

then.

1
| L( it , g |4 dt , N , g ,Q ( KL )1
2

for L=K1-1/(4+20)+ . Here 0 is given by bounds toward the Ramanujan conjecture in(3.1.1) and we
can take 0=7/64 with 1-1/(4+20)=103/135.
A Subconvexity bound for L(s,g) in the t aspect was deduced by good for holomorphic
cusp form g
5
1
1

it , g = g (1 | t | 3 log 6 (2 | t |).
2

(3.1.2)

The goal of the present paper is not an improvement to this Subconvexity bound of
L (s.g). by a standard argument though, our theorem 3.1.1 implies
1
1

it , g = g (1 | t |) 21/(168 )
2

119

(1 | t |)

270

(3.1.3)

One can only get a fourth power moment bound of (K4/3L) + , not as good as our
Theorem 3.1.1 Rankin-Selberg L-functions L(s,f g) were proved as the Laplace eigenvalue of the

50

Maass cusp form f goes to , where g is a fixed holomorphic or Maass cusp form. While the
exponent (3+20)/4+ not hold because of a gap is the paper did prove a Subconvexity bound

it, f g = N ,t , g , k (15 20)/16

2

(3.1.4)

As pointed out in the first sentence was improved to a better bound O(k1-1/8+40)+ ).
L(1 / 2, f g )
To express
in terms of spectral decomposition of f and g by an approximate
L( s, f g )
functional equation. Therefore the central value of

is essentially expressed as a

spectral sum of g with Kloosterman sums as coefficients. An application of bounds for

shifted convolution sums of Fourier coefficients of g
What we will proceed in this paper is to consider the continuous spectrum of the Laplacian in
place of the Maass form f.
The author would like to express thanks to Chaohua Jia, Jianya Liu, and Wenzhi
Luo for informative discussions, and to the referee for detailed, helpful comments and
suggestions. Thanks are also due to the Obermann Center for Advanced Studies, the
University of Iowa, for support service and nice academic environment.

51

3.2 THE APPROXIMATE FUNCTIONAL EQUATION:

L( s, g )
we know that the twisted L-function
Q
primitive character modulo

is entire, where x is a real,

L( s , g )

N /Q
with

.note that

is indeed the L-function

attached to a twisted cusp form gx it is

gx ( z) y

n
n 1

( k 1)/2

x ( n )g ( n )e( nz ),

(3.2.1)

When g is holomorphic, and

g x ( z) y

( n ) ( n ) K
n 0

il

(2 | n | y )e(n )

(3.2.2)

When g is Maass
In any case, denote by

( s, g ) L ( s, g ) L( s, g )
The complete L-function, where
2

L ( s, g ) R ( s gx ( j ))
j 1

R ( s ) s /2 ( s / 2).
with
Here gx(1) and gx(2) are complex numbers associated to gx at . P.1188, our twisted
cusp form gx satisfies the standard functional equation

52

 ( s, g ) ( s, g x ) (1 s, g )

(3.2.3)

(s,gx)=T(gx)Q -sgx.

Where

Here Qgx>0 is the conductor of gx and T(gx)

T ( g x )T ( g x ) Qgx

Cx satisfies

. since

g is self-contragredient and x is real, we have T(gx)2=Qgx.

We actually want a functional equation for a product of two such L-functions:
L( s ir, g ) L( s ir, g ) ( s ) L(1 s ir, g ) L(1 s ir, g ),
Where according to (3.2.3)
2

( s ) T ( g x ) 2 Qgx2 s

R (1 s ir gx ( j )) R (1 s ir gx ( j ))
R ( s ir gx ( j )) R ( s ir gx ( j ))

j 1

Qgx

2 s 1

j 1

(1 s ir gx ( j )) / 2) (1 s ir gx ( j )) / 2)
( s ir gx ( j )) / 2)( s ir gx ( j )) / 2)

4 2

( s )

Qx

( gx (1) r ) 2 ( gx (2) r ) 2

2 s 1

(1 r ( s ))

r
where the error term

(s)<<(1+|s|)3/(1+|r|). We will consider the case of large |r| with fixed

g; hence r(s) is asymptotically (Qgxr2/4 2))1-2s

L( s ir, g ) L( s ir, g )
Again, we can express the central value of

1
1

ir, g L
ir, g
2
2

53

as

1
ds
1
1

X s L s ir , g L
s ir , g G ( s )

i Re s 2
s
2
2

1
1
X sr ( s ) L( s ir, g )
2
2
Re s 2

O (|

ds
1

s ir , g G ( s ) |),
s
2

XL

Where

Qgx
4

( gx (1) r 2 ) 2 ( gx (2) r 2 ) 2 .

Here G(s) is an analytic function in -B Re s B for a fixed B > 0 satisfying

G (0) =1,
G(s) =G(-s),
|G(s)|<<(1+|s|)-A
For a fixed large constant A. We note that x is real (and positive), because of our
assumption on g being self-contragredient. We may shift the contour in the integral of the big

r
s

1+

O term in (3.2.4) to Re s= + . This way X << r . Recall that

(1/2 s) << (1+|s|)3/(1+|

r|).
Moreover,

1
1

s ir , g L
s ir, g = , g 1
2
2

as r, for Re s=1/2+ , because its Dirichlet series is absolutely convergent.

All these show that the big O term in (3.2.4) is << r .
To compute the main term in (3.2.4), we expand
L(1 / 2 s ir, g ) L(1 / 2 s ir, g )
into its Dirichlet series. For Re s>1/2,
we have

54

s ir, g L
s ir, g
2
2

g ( m )g ( n )
1/2 s ir 1/2 s ir
n

(mn ) m

m ,n 1

(3.2.5)
As g is a Hecke eigenform, we have

g ( m ) g ( n )

d |( m ,n )

mn
2
d

apply this to the right side of (3.2.5) and set m= ad, n=bd. Then

s ir , g L
s ir , g
2
2

(abd 2 )

a ,b , d 1

(ab)

g
1/2 s irb1/2 s ir

d 12 s

( n ) g ( n )
d ir (n )
n1/2 s
n 1

L(1 2s, 2 )

(3.2.6)
where
a
d s (n)
ab |n| b

Q
We remark that in (3.2.6) the series is actually taken over n which are relatively prime to
. Consider the Eisenstein series for any, fixed cusp a of =0(N) defined by

Ea ( z , s )

a \

55

(Im a1 z ) s

for Re s > 1 and by analytic continuation for all s

a SL(2, R)
while

C. Here is the stability group of a

a1 a a .

a a
is given by

.and

This Eisenstein series is an

eigenfunction of the Hecke operators

Tn Ea ( z, s ) a (n, s ) Ea ( z, s ),

If (n, N) =1 for any n relatively prime to N, (n, s)=ds-1/2(n) (3.2.6) we get

1

L s ir , g L
s ir , g
2
2

( n )g (n )a (n,1 / 2 ir )
n1/2 s
n 1

L(1 2 s, 2 )

(3.2.7)

for Re s>
Substituting (3.2.7) into the integral of the main term in (3.2.4)
we get
1
1

L ir, g L ir , g
2
2

Res 2

( nm )2 g (n )a (n ,1 / 2 ir )
ds
s

X G ( s ) O( r )

2 1/2 s
( nm )
s
m ,n 1

( nm 2 )g ( n)a ( n,1 / 2 ir ) 1
nm 2
2
G
(
s
)

2 i Res 2
m n
m ,n 1
X

denote

56

ds
)O ( r )
s

V ( y)

1
ds
G( s) y s

2 i Re s 2
s

Lim V(y) = 1 and V(y) <<B (1+|y|)-B because of our choice of the function G(s). Therefore

1
1

ir, g L ir, g
2
2

1 m X 1/ 2
O ( r ),

2 ( m ) ( n )g ( n )a ( n,1 / 2 ir ) nm 2
V

m n1
n
X

(3.2.8)

Because the outer series is negligible if taken over m > X1/2+.

3.3 AVERAGING AND THE KUZNETSOV TRACE FORMULA:

According to (3.2.8) estimation of the central value of our L-function is reduced to
estimation of

n
SY ( g , r ) (n )g (n )a (n,1 / 2 ir ) H
Y
n
57

for fixed g, where H is a fixed smooth function of compact support contained in (1,2).
To prove our results on short intervals, let L be a number which satisfies K LK/4
for large K. let h(t) be an even analytic function in |Im t| satisfying h(n)(t) << (1+|t|)-N for
any N > 0 in this region. Thus h is a Schwartz function on R. we also assume that h(t) 0 for
real t. For example, we may simply take h(t) = 1/cosh(t). Denote

N ( s ) (1 p s ) 1
p/ N

We want to estimate
2
K r
K r
2 | N (1 2ir ) |
I K ,L h

h
|
S
(
g
,
r
)
|
dr
Y

L
| (1 2ir ) |2
L

(3.3.1)
_

n _ m
H
Y Y

( n ) ( m ) g ( n ) g ( m ) H
m ,n

K r
K r
X h
h

L
L

R
_

X a (n,1 / 2 ir ) ( m,1 / 2 ir )

| N (1 2ir ) |2
dr.
| (1 2ir ) |2

We apply the Kuznetsov trace formula to the integral on the right side of (3.3.1)
_
n _ m
( n ) ( m ) g ( m ) H H
Y Y
m ,n

58

(3.3.2)

_
K k j
K k j
X h

(
n
)

fj
f j (m)

L
L
fj

n _ m
H
Y Y

( n ) ( m ) g ( n ) g ( m ) H
m ,n

(3.3.3)

K r
K r
X h
h

L
L

R
| N (1 2ir ) |2
X a ( n,1 / 2 ir )a ( m,1 / 2 ir )
| (1 2ir ) |2
_

n _ m
H
Y Y

( n ) ( m ) g ( n ) g ( m ) H
m ,n

(3.3.4)

n, m
K r
K r
tanh( r ) h
h
rdr

R
L

n _ m
H
Y Y

2i (n ) (m )g (n ) g (m ) H
m ,n

(3.3.5)

4 nm
S ( n , m; c )
J
2
ir
R c
c
c 1

rdr
K r
K r
X h
.
h

L cosh( r )

Here in (3.3.2) fj are Hecke eigenforms, with Laplace eigenvalues 1/4+k2j and Fourier

f ( n ),
j

coefficients

which form an orthonormal basis of the space of Maass cusp forms for

0(N), while in (3.3.5) S(n,m;c) is the classical Kloosterman sum.

59

More precisely, (3.3.4)+(3.3.5) << LKY 1+ for L=K1-1/(4+20)+ . Since (3.3.2) and (3.3.3)
are both positive, this implies that (3.3.3), i.e., I K,L, is bounded by O(LKY1+ ) for the same L. By

(1 2ir ) = log(1 | r |),

this estimate of I K,L, implies that

K L

| SY ( g , r ) |2dr = LKY 1

(3.3.6)

For the above L. Now we can go back to the fourth power moment of
L(1 / 2 ir, g ).
since g is self-contragredient and x is real, we have from (3.2.8) that

K L

K L

L ir, g dr
2

K L

1
1

L
ir, g L
ir , g
2
2

dr

1 m X 1/ 2

2 ( m) (n )g (n )a (n,1 / 2 ir ) nm 2
V

dr.
m n 1
n
X

Here we can take X = K2 and get

K L

L ir, g dr
2

1
= 2
K

K L

(n) (n ) (n,1 / 2 ir )
K

n 1

60

1m K 1

(m )
2

V ( nm 2 / K 2 )
nm 2 / K 2

dr

Now we apply a smooth dyadic subdivision to

2 (m)

V ( nm 2 / K 2 )
nm 2 / K 2

1 m K 1

By dividing the interval [1,K1+ ] and covering, with overlapping, each subinterval by
a smooth, nonnegative function of compact support. The total number of subinterval is so
that
K L

L ir , g dr
2

log K
=
K2

K L

max
1 B K 2

n
( n )a (n,1 / 2 ir ) H 2 dr.

K / B
n 1

The sum inside the absolute value signs is indeed Sk2/B(g, r). By (3.3.6), the maximu
Contribution is from B = 1:

K L

log K
1

L ir, g dr =
K2
2

K L

log K
LK ( K 2 )1 = ( KL)1
2
K

for L = K1-1/(4+20)+ . This completes the proof of Theorem 3.3.1.

61

| S K 2 ( g , r ) |2dr

CHAPTER IV

MOMENTS OF PRODUCTS OF AUTOMORPHIC

L-FUNCTIONS
4.1 INTRODUCTION:
An important problem in analytic number theory is understand the behavior of Lfunction on the critical line and at the central point. The Langlands program predicts that the
most general L-functions are attached to Automorphic representations of GL(n) over a
number field and further conjectures that these L-functions should be expressible as products
of the Riemann zeta-functions and L-functions attached to cuspidal Automorphic
representations of GL(m) over the rationals. In this paper, we investigate the moments of
such products on the critical line. We also prove estimates for moments of Dedekind zeta-

functions,

k(s), of Galois extensions K over

. In general, unless Gal(K/

) is solvable, it

is not known that

(s) can be written as a product of Automorphic L-functions (though the

Langlands reciprocity conjecture predicts that is the case).

An L-function is called primitive if it does not factor as a product of L-function of
( s ) 1 / 2

smaller degree. Given a primitive L-function, L(s, ), normalized so that

C (k , )
critical line, it has been conjectured that there exist constants

62

such that

is the

for any

k 0

as T

L it ,
2

2k

dt : C ( k , )T (log T ) k

(4.1.1)

L ( s, ) ( s )
, The case

, the Riemann zeta-function, has received

the most attention. The conjecture in (4.1.1) has only been established in a few cases and
only for small values of k. for degree one L-functions, the Riemann zeta-function and
Dirichlet L-functions, the conjecture is known to hold when k is 1 or 2. For degree two Lfunction, many cases of the conjecture have been established when k=1. For higher degree Lfunctions, and for higher values of k, the conjecture seems to be beyond the scope of current
techniques.
It is expected that the values of distinct primitive L-functions on the critical line

L( s, 1 ),..., L( s, r )
are uncorrelated. Therefore, given r distinct primitive L-functions,

( s ) 1 / 2
normalized so that
k1,,kr >0 we have
T

L it, 1
2

is the critical line, one might conjecture that for any

2 k1

1
...... L it , r
2

2 kr

k 2 ,...,kr2
dt : C k , T log T 1

(4.1.2)

For some constant

C k ,

as T

where

k ( k1 ,..., k r )

( 1 ,..., r )

and

the case where k1,,kr are natural numbers. Using classical methods, the asymptotic
63

in

formula in (4.1.2) can be established for products of two Dirichlet L-functions in the case

when

L( s, 2 ) L( s, 2 )

k1 k r 1, L( s, 1 ) L( s, 1 ),
and

where

2
and

are distinct

primitive Dirichlet characters. It seems that there are no other cases where the asymptotic
formula in (4.1.2) has been established. The conjectural order of magnitude of the moment in
(4.1.2) is consistent with the observation that the logarithms of distinct primitive L-functions

it , 1
2

log L
on the critical line,

independent if

1 2

it , 2
2

log L
and

, are (essentially) statistically

as t varies under the assumption of Selbergs orthogonallity

L s, 1
1

conjectures for the Dirichlet series coefficients of

L s, 2
and

. This statistical

independence can be made precise;

i

Moments of Automorphic L-functions.:

In this paper, in support of the conjecture in (4.1.2), we prove the following
mean-value estimate for arbitrary products of primitive Automorphic L-functions.

Theorem 4.1.1:

64

L( s, 1 ),..., L( s, r )
Let

be L-functions attached to distinct irreducible cuspidal

j , ofGL(m j )
Automorphic representations,

over

each with unitary central character, and

if max m j 4
1 j r

assume that these L-functions satisfy the generalized Riemann hypothesis. Then,
we have
T

L it , 1
2

2 k1 ...

k1 ,..., k r 0
for any

and every

depends on

L
it , 1
2

1 ,..., r , k1 ,..., kr

2 kr

dt = T (log T )k1 .... kr

(4.1.3)

when T is sufficiently large. The implied constant in (4.1.3)

if max m j 5
1 j r

and .

, then the inequality in (4.1.3)holds under the

additional assumption of Hypothesis H .

Some of the standard properties of the L-functions described in Theorem 4.1.1 are
reviewed in 2. Observe that the upper bound in Theorem 4.1.1 is nearly as the conjectured
asymptotic formula in (4.1.2) This conjecture is implied by the Ramanujan-Petersson conjecture
and is known to hold for L-functions attached to irreducible cuspidel automorphic

GL( m)
representations on

over

if

m4

65

T log T

k2

1
= k it
2
0

= k , T log T

2k

dt

k 2

Hold for any k > 0 and every > 0 Assuming the Riemann hypothesis.
In the case r = 1, combining the result of Theorem 4.1.1 with the work of we know that

T log T

k2

= k ,

L it ,
2

= ,k , T log T

For any k > 0 and every > 0 where

2k

dt

k 2

(4.1.4)

is a self-contragredient irreducible cuspidal Automorphic

representations of GL(m) over

under the assumption of the generalized Riemann hypothesis

L ( s, )

and the Ramanujan-petersson conjecture for

. As mentioned above, the upper bound

holds under weaker assumptions and for more general L-functions.

L ( s, ) ( s )
Let

it is possible to replace the in Theorem 4.1.1 by a quantity which is

o(1 / log log log T );

moreover, an analogue of Theorem 4.1.1 for products of derivatives of L-

66

functions can be proved. There are a couple of aspects which make the proof of Theorem 4.1.1
different than the proof of the analogous result for the Riemann zeta-function. Assuming the
Riemann hypothesis, an inequality for the real part of the logarithm of the Riesmann zeta-

( s ),
function is derived which depends only on the primes. In the case of

one can handle the

contribution of the primes powers relatively easily. then we could similarly derive an inequality
involving only the primes for the real part of the logarithms of the L-functions in Theorem 4.1.1
In order to circumvent these additional assumptions, we must estimate the contribution from the
prime powers in a different way. Ideas similar to these were used fir degree two L-functions.
Finally we remark that, assuming the generalized Riemann hypothesis and the RamanujanPetersson conjecture,
= T (log T ) k
shown that the integral in (4.1.4) is

0k
k satisfying

2
m

if

is self -contragredient for any fixed

? T (log T ) k
Moreover, lower bounds for the integral in (4.1.4) which are

for any positive rational number k have been established assuming unproven bounds toward the
Ramanujan-Petersson conjecture but without assuming the generalized Riemann hypothesis. In
the case m = 2.
After proving our main results, assuming the Riemann hypothesis,
show that

67

1
0 2 it

2k

dt = k T log T

k2

By combining the ideas and results of the present paper with Harpers
techniques, it may be possible to prove a version of Theorem 4.1.1 with = 0 we
are investigating this possibility.
ii

Moments of Dedekind zeta-functons:

Ok
Let K be an algebraic number field, and let

denote its ring of integers. The

K ( s ),
Dedekind zeta-function,

is defined by

K ( s)
a 0

Na

1
s

p
Np

( s ) 1,

68

(4.1.5)

OK ,
where the sum runs over the nonzero ideals a of

OK ,
the product runs over the prime ideals p of

and

N N k /
denotes the absolute norm on K. it is known that the Dedekind zeta-function factors as a

product of Artin L-functions. If K is Galois extension of

then

k ( s ) L( s, ) x (1)
x

(4.1.6)

Where the product is over the irreducible characters

(1)
x

Gal ( K / )
of

and

Gal K /

K :

(4.1.7)

L( s, ) L( s, )
The Langlands reciprocity conjecture implies that each

Automorphic representation

of

(1) m

GL( m)
over

where

Galosi extensions K over

for an irreducible cuspidal

69

. By (4.1.2),(4.1.6)and(4.1.7), for

2k

1
0 K 2 it

for any k > 0 as T

dt : C (k , K )T (log T )

K :

k2

(4.1.8)

. Here C(k, K) is a constant depending on k and the number field K.

The conjectural asymptotic formula in(4.1.8)is known to hold when k = 1 for the Dedekind zeta-functions

of quadratic extensions of

Let d be a fundamental discriminant, and let

shown that

6
1
2
1
0 K 2 it dt : 2 L(1, d )
p|d

as T

L( s , d )

K ( s ) ( s ) L ( s, d )
using the factorization

1
T log 2 T
p

, where

is the Dirichlet L-

d
function associated to

, the Kronecker symbol of d. also in support of (4.1.8) for finite

Galois extensions K over

shown that the inequality

T

1
0 K 2 it

2k

dt

? T (log T ) K : k

70


Holds for any rational number k > 0 as T

Using the following mean-value estimate

for Dedekind zeta-functions is a consequence of Theorem 4.1.1

Corollary 4.1.2:

k ( s)

Let K be a finite solvable Galois extension of

, and let

be the associated

k ( s)
Dedekind zeta-function. Then, assuming the generalized Riemann hypothesis for

we have
T

1
0 k 2 it

2k

dt = K ,k , T (log T )[ K : ]k

for any k, > 0 when T is sufficiently large.

Proof:

If K is a finite solvable Galois extension of

shown that
r

k ( s ) L( s, j )
j 1

71

kj

(4.1.9)

j
where the

are irreducible cuspidal Automorphic representations of the appropriate

degree over

k12 ... kr2 [ K : ].

kj

and the exponents

are natural numbers satisfying

L( s, j )

K ( s)
Since

satisfies the Ramanujan-Petersson conjecture,factor

satisfies this

conjecture. Hence, Hypothesis H holds for each L-function in the product (4.1.9) and thus
Theorem 4.1.1 implies that
T

L it ,
j
0
2

j 1
r

kj 2 k

( k1 ... kr ) k
dt = T (log T )

T (log T )[ K : ]k

(4.1.10)

(K / )
The corollary now follows from(4.1.9) and (4.1.10)The condition that Gal

be a

solvable group can be removed with a little more work. In 5, we sketch how to modify the proof
of Theorem 4.1.1 to prove the following mean-value estimate.
Theorem 4.1.3:

Let K be a finite Galois extension of

k ( s)
hypothesis for

, we have

72

. Then, assuming the generalized Riemann

1
0 K 2 it

2k

dt

= K ,k , T (log T )[ K : ]k

for any k, > 0 when T is sufficiently large.

Unlike the proof of Corollary (4.1.2) our proof of Theorem 4.1.3 does not rely on a

K ( s)
factorization of
iii

. Into Automorphic L-functions.

Coefficients of zeta and L-functions:

In short intervals.
rK (n )
Let K be a number field and let
denote the number of ideals in K of norm
n. then, by(4.1.5)
we get

rK ( n )
,
s
n 1 n

K ( s)
( s ) 1

When K is a Galois extension of

, we can use Theorem 4.1.3 and a technique to study

rK ( n )
the distribution of

in short intervals assuming the generalized Riemann hypothesis for

K ( s)
. In order to state our result, we know that

Re s K ( s ) lim ( s 1) K ( s )
s 1

73

2 r1 (2 ) r2 hR
w D

r1
When

(4.1.11)

r2
is the number of real embeddings of K,

is the number of pairs of complex

embeddings , h is the class number of K, R is the regulator, w is the number of roots of unity in
K, and D = |dk| is the absolute value of the discriminant. Landaus classical mean-value

rK (n )
estimate for the arithmetic function

is

rK (n)
n x

2 r1 (2 ) r2 hR
x 0 x12/([ K : ]1)
w D

We prove the following conditional estimate for the variance of the arithmetic function

rK (n )
in short intervals.

Theorem 4.1.4:

74


Let K be a finite Galois extension of

. Let y=y(x) be a positive and increasing

function such that y and y/x 0 as x . Then, assuming the generalized Riemann

K ( s)
hypothesis for

, we have

1
x

2x

2r1 (2 ) r2 hR
r
(
n
)

y dx
K
w D
= y (log X )[ K : ]
x n x y

for > 0 when X is sufficiently large. Here the implies constant depends on K and .

K ( s ),
Assuming the generalized Riemann hypothesis for

x n x y

rK (n ) :

from Theorem4.1.4 that

2r1 (2 )r2 hR
y
w D

y / (log x )[ K : ]

For almost all x if we choose y to be a function of x satisfying

y/
x

but

as

Using Theorem 4.1.1 we can similarly study the behavior of coefficients of products of
Automorphic L-functions in short intervals. To state the results in this situation, we first

introduce some notation. For

k 0

k1 ,...., k r ,
an integer and

75

L( s ) ( s )

L( s,
j 1

kj

Let

Be an (Automorphic)L-function. Here we are assuming that the L-functions

L( s , j ) ( s )

L( s, 1 ),..., L( s, r )
are as in Theorem 4.1.1.and that

for all

L( s )
distinguish between the case

1 j r

is entire, and the case

k o

, where

s 1

. We
( s ) 1
. For

we get

aL (n)
, if k 0
s
n 1 n
L( s )
bL (n )

, if k

n 1 n s

The pole of the generating function when

k 1

. For

x0

, we define

x s
RL ( x ) Re s L( s )
s 1
s

RL ( x ) 0
Note that

if

k 0

, that
r

RL ( x ) x L(1, j )

kj

j 1

If
76

k 1

, and that

RL ( x )

x(log x ) k 1 r
k
L(1, j ) j o( x (log x ) k 2 )

( k 1) ! j 1
If k 2.

Assuming to show that

1
x

1
x

x x n x y

x xn x y

aL ( n ) dx

= y (log X ) k1 ...k r

2x

2x

bL ( n ) ( RL ( x y ) RL ( x ) dx

and

For > 0 when X is sufficiently large. Here y is any function satisfying the conditions

1 ,..., r , k , k1 ,..., k r
in Theorem 4.1.4 and the implied constants depend on
iii

and .

Quadratic twists of Automorphic L-functions:

L( s , )
Let

be an L-function attached to a self-contragredient irreducible

cuspidal Automorphic representation

GL( m )
on

over

(we assume the L-function

is self-dual so that the central values is real.) conjectured that the family of quadratic

77

L( s , )
twists of

has either symplectic or orthogonal symmetry corresponding to

L( s , , ^ 2 )
whether or not the symmetric square L-function

L( s, , ^ )
2

( ) 1
We get

has a pole at s = 1

if

( ) 1
does not have a pole at s = 1 and set

if

L ( s, , ^ )
2

has a pole at s = 1. Then for each k > 0 it has been conjectured that there

C b (k , ) 0
are constants

|d | x

such that
k

L , d : C b ( k , ) X (log X ) k ( k ( ))/ 2
2

as X . Here the superscript indicates that the sums run over fundamental discriminants

d , d
denotes the corresponding primitive quadratic Dirichlet character, and (as before)we
have normalized so that s = is the central point. In the case of quadratic Dirichlet L
functions and L-functions of quadratic twists of a fixed elliptic curve E over

|d | x

, show that

, d = X (log X ) k ( k 1)/2
2

78

(4.1.12)

|d | x

L , E d = X (log X )k ( k 1) / 2
2

(4.1.13)

and
For every k > 0 and any > 0 assuming the generalized Riemann hypothesis for the

relevant L-functions. (Note that in the first example the L-functions have ( ) = -1, and in

the second case the L-functions have ( ) = 1.) we generalize these results and, in analogy
with our Theorem 4.1.1 we prove the following result for central values of quadratic twists of
arbitrary products of Automorphic L-functions.Theorem 1.5 Let d denote a fundamental

d
discriminant, and let

be a primitive quadratic Dirichlet character of conductor |d|. Let

L( s, 1 ),..., L( s, r )
be L-functions attached to distinct self-contragredient irreducible

j
cuspidal Automorphic representations,

GL(m j )
, of

over

each with unitary central

L( s, 1 d ),..., L( s, r d )
character, and assume that the twisted L-functions

satisfy the

max m j 2,
1 j r

generalized Riemann hypothesis. Then if

we have

79

|d | X

kr

1
1

L , 1 d ... L , r
d
2

= X (log X ) k1 ( k1 (1 ))/2...kr ( kr ( r ))/ 2

(4.1.14)

For any k1,,kr > 0 and every > 0 when X is sufficiently large. Here the
superscript b indicates that the sum is restricted to fundamental discriminants, and the

max m j 3,

1 ,..., r , k1 ,..., k r ,
implied constant depends on

1 j r

and . If

then the inequality in

(4.1.14) holds under the additions assumptions of Hypothesis H and Hypothesis E .

We now give two example which are consequences biquadratic extensions of

Let d1 and d2 be coprime fundamental discriminants, and let

K d1 ,d2

d1 , d 2

be the

K d1 ,d2
corresponding biquadratic number field. Then the Dedekind zeta-function of

factors as

Kd ,d 2 ( s ) ( s ) L( s, d 2 ) L( s, d1 d 2 ),
1

L( s, E / K d1 ,d 2 )

and similarly, given an elliptic E over

, the Hasse-Weil L-function

K d1 ,d2
over

factors as

80

of E

L( s, E / K d1 ,d 2 ) L( s, E ) L( s, E d1 ) L( s, E d 2 ) L( s, E d 2 )

Using Theorem 4.1.5 we can estimate moments of

1
Kd1 ,d2
2

, E / Kd1 ,d2
2

L
and

by

averaging over two sets of fundamental discriminants. (we note that under the assumption of
the generalized Riemann hypothesis for these zeta and L-functions, these central values are
non-negative real numbers.) In particular, we have
k

d1d2 | X
( d1 , d2 ) 1

Kd1 ,d2

1
3k ( k 1)/2 1
= X (log X )
2

(4.1.15)

and
k

L , E / K d1 ,d2 = X (log X )3k ( k 1)/2 1

d1d 2 | X
2

( d ,d ) 1
b

(4.1.16)
for any > 0. Here the superscript b indicates the sum runs over two
sets fundamental discriminants,d1 and d2. When k = 1, as X

1
b a (d1 , d 2 ) Kd ,d
1 2
d1 ,d2odd
2

d d
F 1 2 : cX log 4 X
X

81

For a constant c > 0,where F is a smooth compactly supported test function satisfying

a d1 , d 2 1

d1 , d 2 1
if

and is (on average) small otherwise.Since the condition

d1 , d 2 1

d1 ,d 2 d1 ,d 2
implies that

, and
k

d1d 2 | X
( d1 ,d 2 ) 1

Kd1 ,d2

( ) 1

1
1

2
2

for degree one L-function,

b L
, d
|d1 | X
2 1

|d 2 | X /|d1|
( d ,d ) 1
1

L , d1
2

= X (log X ) k ( k 1) b
|d1 | X
| d1 |

1
, d
2 2

1
dd
L ,
2 12

= X (log X )3k ( k 1)/2 1

This proves that the estimate in (4.1.15) follows from Theorem 4.1.5
Imply that L(s ,E) and its quadratic twists correspond to L-functions attached to irreducible

( ) 1

cuspidal Automorphic representations of GL(2) over

. Moreover, we have

for each

of these L-functions. Therefore, under the conditions of Theorem 4.1.5 we similarly have
k

1
, E / K d1 ,d2 L , E
2

b L
d d | X
1 2

( d1 ,d 2 ) 1

b L
, E d1
|d1 | X
2

82

b L
, E d 2

|d 2 | X / d1 |
2

( d ,d ) 1
1

1
, E d1 d2
2

L , E d1
2

= X (log X ) k ( k 1) b
|d1 | X
| d1 |

= X (log X ) 3k ( k 1)/21

By two more applications of (4.1.14) and summation by parts. This shown that the estimate in
(4.1.16) also follows from Theorem 4.1.5

83

4.2. PROPERTIES OF AUTOMORPHIC L-FUNCTIONS:

Let

bean irreducible cuspidal Automorphic representation of GL(m) over

with

( s) 1,
unitary central character. For

we let

a (n)
L ( s, ) s
n
n 1
p

Be the global L-function attached to

j ( p )

1 ps
j 1

(4.2.1)

. Then L(s, ) is either the Riemann zeta-function or

L(s, ) analytically continues to an entire function of order 1 satisfying a functional equation of

the form

( s, ) : N s /2 ( s, ) L( s, )

(1 s, ),

Where N is a natural number, |

(1 s, ) ( s, )

|=

84

, and the gamma factor

( s, ) ( s j )
j 1

( s ) s /2 ( s / 2)
Here

, and the

are complex numbers. Logarithmically differentiating the

Euler product,
we define

L
d
( s, ) : log L( s, )
L
ds

(1l ( p ) ... ml ( p)) log p

pl s
p l ,l 1

( n )
ns
n 1

(4.2.2)

( p ) a ( p )

( s )
for

> 1. We note that

log p.

Let

be a primitive Dirichlet character modulo q satisfying (q, N) = 1, and define

85

a (n ) ( n )
ns
n 1

L( s, ) :

j ( p) ( p)

p
j 1

( s )
For

> 1 Then

L( s, ) :

d
log L( s, )
ds

(n ) ( n )
,
ns
n 1

L( s , )

( s ) 1
When

. For q > 1, function

continues to an entire function of order 1

and satisfies a functional equation of the form

( s, ) : ( q m N ) s /2 ( s, ) L( s, )

, (1 s, )

| , | 1, ( s, ) ( s, )

where

, and the gamma factor

m

( s, ) ( s j , )
j 1

86

 j,
For complex number

.The generalized Riemann hypothesis states that all the zeros

( s, f )

( s, f )
of the completed L-functions,

and

( s ) 1 / 2
, are on the critical line

. We

always indicate the L-functions for which we are assuming this hypothesis. The Ramanujan-

| j ( p ) | 1

j ( p)
Petersson conjecture states that the Euler coefficients

in (4.2.1) satisfy

for all

but a finite number of primes p. In general, this conjecture is open. Towards the RamanujanPetersson conjecture, show that
2

| j ( p ) | p1/21/( m 1)

For all p. it follows that

2

| (n ) | m (n )n1/21/( m 1)

( n )
Where

p j, j 1

( n )
is the Von Mandgoldt function, defined by

(4.2.2)

= log p if

otherwise. The bound in (4.2.2) is crucial to our proofs of Theorems 4.1.1 and 4.1.5

Hypothesis H:

87

( n )
and

=0

Let j 2 be fixed, and let

be an irreducible cuspidal automorphic representationof GL(m)

over

. Then we have

| ( p j ) |2
p p j

Hypothesis H is known to hold for automorphic L-functions of small degree.

Theorem 4.2.1:
Hypothesis H is true for m 4
Proof:
The case m = 1 is trivial, the case m = 2 follows from the work the case m = 3 and the case
m=4

L( s, ' ),

L( s, )
Given distinct automorphic L-functions

and

we need to understand the

correlation of their Dirichlet series coefficients averaged over the primes.

Selbergs Orthogonality Conjectures. Let

and

'

be two irreducible unitary cuspidal automorphic

representations of GL(m) and GL(m) over

, respectively, and let x 3. Then

a ( p )a ' ( p)
( p) ' ( p )

p
p log 2 p
p x
p x

88

 log log x o(1), if ',

if '.
o(1),

The following result allows us to use Selbergsorthogonality conjectures in the proofs of

Theorems 4.1.1 and 4.1.5

Theorem 4.2.2:

Let

and

'

be two irreducible unitary cuspidal automorphic representation of GL(m) and

GL(m) over

L( s, ' ),

L( s , )

, respectively. If

and

satisfy Hypothesis H, then the coefficients of

these L-functions satisfy Selbergs orthogonality conjectures. In particular, Selbergs orthogonality

conjectures hold if max (m, m) 4
Proof:

This was proved in the special case where at least one of

or

'

is self-contragredient in full

generality
In order to prove Theorem 4.1.5 we need to understand the behavior of the Dirichlet series

L( s, )
coefficients of automorphic L-functions averaged over the squares of primes. Let

function attached to a self-contragredient irreducible cuspidal automorphic representation

over
89

be an L-

of GL(m)

(i.e.

L( s, ) L( s, )
). The Rankin-Selberg L-function

factors as product of the

symmetric and exterior square L-functions

:

L( s, ) L( s, , 2 ). L( s, , 2 )

and has a simple pole at s = 1. This pole must be carried by one of the factors on the right-hand side. We

L( s, , 2 )
denote the order of the pole of

(1 ( )) / 2
as

(p
p x

. we known that

) : ( ) x

(4.2.3)

as x . We use this estimate, in a different form, in 6.

The proof of Theorem 1.5 also requires an assumption on the coefficients of the L-functions which is
stronger than Hypothesis H.
Hypothesis E:

Let j 2 be a fixed integer, and let

be an irreducible cuspidal automorphic representation

of GL(m) over

. Then we have

| ( p 2 j ) |
p p j

Note that Hypothesis E only applies to even powers of primes, and the power of the prime in the denominator differs from the corresponding exponent in Hypothesis H. Hypothesis E, though stronger than
90

Hypothesis H, is still considerably weaker than the Ramanujan-Petersson conjecture. Indeed, it would

| j ( p ) | p1/4v
follow if the Euler product coefficients in (4.2.1) satisfied a bound of the form

v > 0.

max m j 3
1 j r

Such a bound trivially holds when m = 1 and when m = 2. when

4.3 LEMMA:
Lemma 4.3 :

Let T be large, x 2, and let

and j be natural numbers satisfying

complex numbers b(p) we have

1
T

2l

2T

p jx

b( p )
dt
=
l
!

p j ( it )

Where j is fixed and the sum runs over the primes p

Proof.

91

| b( p ) |2
p 2 j
p j x

xl T j

. Then for any

The case j = 1 for any s

, write
l

b( p )
s
p y p

y ,l ( n )
Where

= 0 unless n is the product of

n yl

y ,l ( n )
,
ns

(not necessarily distinct) primes, all less than or equal to

y. thus, we have
2T

p y

2T

2l

b( p )
dt
p j ( it )
2

(n)
l n jy(,l jit ) dt
n y
2 jT

jT

(n)
l n jy(,l jit ) du
n y

yl T
Where in the last step we have made the variable change u = jt. If
2T

p y

, then implies that

2l

b( p )
dt
p j ( it )

2 jT jT
=
j
=T
n yl

n yl

y ,l ( n )
n 2 j

y ,l ( n )
n 2 j

By modifying the combinatorial argument appearing in the proof it follows thart

92

n yl

y, l (n)
= l !
2 j
n

| b( p) |2

p 2 j
p y

Combining estimates, the lemma follows.

4.4 PROOF OF THEOREM 4.1.1:

In this section, we state and prove a value distribution result for a linear combination

L( s, 1 ),..., L( s, r )
of distinct primitive L-functions and use this to deduce Theorem 4.1.1Let
r distinct primitive L-functions(as in Theorem 4.1.1) of degrees m1,,mr, respectively,
93

be

V max m12 1,..., mr2 1 ,

Let

B k1m1 ... k r mr 1.

(4.4.1)

and let
Define the set

A(T ,V ) t [T , 2T ] : k1 log L it , 1 ... k1 log L

it , r V
2

w ( k12 ... k r2 ) log log T

and the quantity

Note that
2T

L it , 1
2

2 k1

........ L
it , r
2

2 kr

dt exp(2V ) d meas ( A(T ,V ))

2 exp(2V ) meas( A(T ,V ))dV .

(4.4.2)

To prove Theorem 4.1.1 it suffices to estimate the measure of A(T , V) for all V 3 when T is
large. Note that the definitions of A(T, V) and W depend on our choices of k1,,kr,
We prove estimates for the size of A(T ,V) using Lemmas 4.3 The contribution to the size of A(T,
V) coming from the primes in the sum on the right-hand side of the inequlity and the

contribution from the prime powers pj with j >

is estimated triviall. More care is necessary to

94

handle the contribution H. this allows us to circumvent using the Ramanujan-Petersson

conjecture.
As might be expected, the proof of Theorem 4.1.1 relies on understanding the
correlations between coefficients of distinct automorphic L-functions. The key ingredient to the
proof of the proposition below (and hence Theorem 4.1.1) is the fact that the Selberg
orthogonality conjectures imply that

p z

k1 1 ( p ) ... k r r ( p )
2

p log p

k12 ...k r2 log log z o(1)

(4.4.3)

as z , which can be seen by expanding the square on the left hand side of (4.4.3)

p z

k1 1 ( p ) ... k r r ( p )
p log 2 p

k12 ...k r2 log log z o(1)

(4.4.3)

95

Proposition 4.4.1 :

L( s, 1 ),..., L( s, r )
Let

be L-functions attached to distinct irreducible cuspidal

j
automorphic representations,

of GL(mj) over

with unitary central character, and assume

max m j 4
1 j r

that these L-functions satisfy the generalized Riemann hypothesis. If

or each of the

W V
L-functions satisfies Hypothesis H, then the following inequalities hold. If

meas

if

w
1
V
w log w
2
B
2B2

V2
V
4
( A(T ,V )) = T
exp 1
w
4 w log w
w

, we have

2
V2
V
7 B 2V
meas( A(T ,V )) = T
exp
1
;
w

4 w log w
w

and if

1
w log w V
2B2

, we have
96

w
B2

we have

meas ( A(T ,V )) = T exp

V log V
2
129 B

for any k1,,kr > 0 when T is sufficiently large.

Proof:
L( s , )
let
be a primitive L-function of degree m. Choosing z=(log T)1- and

1
2

, it follows from (4.2.2) that

(1 0 )m log T
1
1

log L it , m(log T )1
o

2(1 ) log log T

2

(1 ) log T

3m log T
4 log log T

For sufficiently large T. therefore, we see that

3( k1m1 ... k r mr ) log T
1

k1 log L it , 1 ... k r log L

it , r
4
log log T
2

When T is large. Recalling the definition of B in (4.41) we assume that

3( B 1) log T
w V
4 log log T
While proving the proposition. Note that B > q(a fact that is useful when deriving the
estimates below ).
Define a parameter A as

B
w
if w V 2
2 log w,
B

1
w
A
w log w, if 2 V 1 2 w log w,
B
2B
2 BV
B,
1
ifV
w log w,

2B2

and let

x T A/V

and

z x1/log logT

. Choosing

1/ 2

97

known that

it , 1 ... k r log L
it , r
2

k1 log L

S1 (t ) S1* (t )

2 j V

S j (t )

3( B 1) V
0(1),
4
A

(4.4.4)

( k1 1 ( p ) ... k r r ( p)) log( x / p )

,
1

it
log x
p z
2 log x
p
log p

S1 (t )
Where

S1* (t )

(k1 1 ( p ) ... k r r ( p )) log( x / p )

,
1

it
log x
z p x
2 log x
p
log p

and
(k1 1 ( p ) ... kr r ( p j )) log( x / p j )
S j (t )
,
1
j(
it )
log x
p j x
j
2 log x
p
log p
2 j .
for

the coefficient bound in (4.2.2)implies that the error term in (4.4.4) is

O(1)since

s (t ) =
j V

p jx

k1 1 ( p j ) ... kr r ( p j )
jp j /2 log p

= 1.
Let

7( B 1)
( B 1)V

*
V1 : V 1
,V1 V j :
8A
8VA

2 j
for

A(T ,V ),
. Note that if t

then at least one of the following inequalities holds:

S (t ) V1*or S j (t ) j
*
1

98

For some j = 1,2,

and define
N j (T ,V j ) : meas t [T , 2T ] :| S (t ) | V j

N j (T ,V j )

N1* (T ,V1* )

For j = 1,2,, and define

similarly, then we can bound

N1* (T ,V1* )
and

using Lemma 4.3since Chebyshevs inequality limplies that

2T

N j (T ,V j ) (V j )

2 l

| S (t ) |

2l

dt

N j (T ,V1* ) (V1* ) 2 l

2T

| S (t ) |

2l

*
1

dt

l
For every non-negative integer
N1 (T ,V1 )
zl T
l
Let us first estimate
. Letting be any natural number such that
,Lemma 4.3

and(4.4.3) imply that

2T

S1 ( t ) dt = T l !
2l

p z

2
k1 1 ( p ) ... k r r ( p )

p log 2 p

= T l !(( k12 ... kr2 ) log log z O (1)) l

= T l !(( k12 ... kr2 ) log log T ) l
l ( k12 ... k r2 ) log log T
= T !

l w
=T l
e

Thus we have

l w
N1 (T ,V1 ) = T l
2
eV1

99

(4.4.5)

V
We consider separately the two cases where

V 2
w

w2
B4

and V >

w2
B4

in the first case, we choose

in (4.4.5) and find that

N1 (T ,V1 ) = T

w2
B4
In the case, where V >

We choose

V2
V
exp 1
w
w

=[10V] in (4.4.5) and find that

N1 (T ,V1 ) = T exp( 4V logV )

Hence
N1 (T ,V1 ) = T

V 2
V
exp 1 T exp( 4V logV )
w
w

(4.4.6)

for all V
2T

2l

S1* (t ) dt = T l !

z p x

2
k1 1 ( p ) ... k r r ( p)

p log2 p

= T l !(( k12 ... k r2 ) log log x log log z O (1)) l

= T ( l ( k12 ... kr2 ) log log log T O (1)) l
= T (2l ( k12 ... k r2 ) log log log T ) l

When T is large. Choosing

V
l
A

, we have that

100

2l

8A
V log V
(2l ( k12 ... k r2 ) log log log T ) l = T exp
(4.4.7 )

2A

( B 1)V

N1* (T ,V1* ) = T

N j (T ,V j )
Finally, we find an upper bound for

2 j
for each

for

x1/ j T

, Lemma 4.3

and Hypothesis H imply that

2T

2l

S j (t ) dt = T l !

p j x

2
k1 1 ( p j ) ... k r r ( p j )

j 2 p j log 2 p

= T (l C j ( k12 ... kr2 ))l

For each fixed j, where Cj is a constant (depending on j).

Let

Cmax max C j
2 j V

2 j
Be an absolute constant. Then for every
2T

we have

2l

S j (t ) dt = T ( l Cmax ( k12 ... k r2 )) l

2T

2l

S1* (t ) dt

Comparing this upper bound to the upper bound for

V logV
N j (T ,V j ) = T exp

2A

, we conclude that
(4.4.8)

2 j
For each

. The proposition now follows by combining the estimates in

(4.4.6),(4.4.7)and (4.4.8). We now use Proposition (4.4.1) and (4.4.2) to prove Theorem1.1
Proof of Theorem 4.1.1 Proposition 4.4.1 implies that

101

v 2

T
(log
T
)
exp

,
256w
w

if 3 V
,
meas ( A(T ,V )) =
B2
T (log T ) exp 4v ,
256 w

2 if V
.

B2

Inserting these bounds into (4.4.2)and estimating the range V < 3 trivially, we know that
2T

L it, 1
2

2 k1 ...

L
it , r
2

2 kr

dt

= T (log T ) e w 256 w
2

= T (log T ) k1 ... Kr
Theorem 4.1.1 follows by summing this estimate over the dyadic intervals

T T
T T T
2 , T , 4 , 2 , 8 , 4 ,...

102

4.5 SKETCH OF THE PROOF OF THEOREM 4.1.3:

We now sketch how to modify the proof of Theorem 4. 1.1 to deduce Theorem

K ( s)

4.1.3 throughout this section, let K be a finite extension of

, and let

be the associated

Dedekind zeta-function.
2T

1
K it
2

2k

dt 2 exp(2V )meas( A(T ,V ))dV

(4.5.1)

Where

1
A(T ,V ) t [T , 2T ] : k log K it V
2

K ( s)
In order to bound the measure of A(T,V), we need analogues of (4.4.3) for
( s) 1,
For

define

K
d
(n)
( s ) : log K ( s ) K s
K
ds
n
n 1

K ( s)
Since

satisfies the Ramanujan-Petersson conjecture,

we have
| K ( n ) | [ K : ] ( n )

0 0.4912...
Lemma 4.5.1 Let

denote the unique positive real number satisfying

e 0 0 02 / 2

K ( s)
.then, assuming the generalized Riemann hypothesis for

0 log x / 2
and log

x2

103

, for all

we have

1
it
2
n x

(n)

log K

K
1

it
2 log x

log n

1
(1 ) [ K : ]log T
O

2
log x
log x

For T t 2T and T sufficiently large, where the implied constant in the error term
depends only on K.
Proof:
This is a consequence of Theorem 4. 2.1
The analogue of (4.4.3) follows from the Chebotarev density theorem
Lemma 4.5.2:

Let K be a finite Galois extension of , and let p denote a rational prime. Then
K ( p)2 : [ K : ]1
p x

px

as x , and in particular

K ( p )2
: [ K : ]log log x

p
p x

(4.5.2)

Proof:
Let (p)denote the principal ideal in OK generated by p. then
( p ) 1e1....err
where the Pi are the distinct prime ideals in OK lying above p with norm pfi It follows that
r

e f
i 1

i i

[ K : ].

e1 ... er 1
If p is unramified in K, then

. Since K is Galois, all the pi lying above

f1 ... f r f ,
p are conjugate thus

say. Therefore, for unramified primes p, we know that

rK ( p ) 0
if and only if f = 1

rK ( p ) [ K : ]

r [K : ]
In this case, p completely splits,

, and hence

unramified primes p,
we have
104

. That is, for

 [ K : ],
0,

rK ( p )

if and only if p splits completely,

Otherwise

Since there are only a finite number of ramified primes, it follows that

r ( p) r ( p)
2

p x

p x

punramified

[ K : ] o(1)
2

px

psplitscompletely

On the other hand, the Chebotarev density theorem implies that

1:

p x
psplirs
completely

1
1
[ K : ] p x

as x thus

r ( p)
p x

: [ K : ]1
p x

Proving the first assertion of the lemma. Using this estimate, the prime number theorem
and partial summation imply(4,5,2) completing the proof of the lemma.
We now indicate how to prove Theorem 4.1.3

105

[K : ]
Choosing w = k

[K : ]
log logT,B = K

+1,

= 2(since the Ramanujan-

K ( s)
Petersson conjecture holds for

), and A as before, a straightforward modification of

the analysis in the previous section implies that

v 2

T
(log
T
)
exp

w ,
256 w

if 3 V
,
meas ( A(T ,V )) =
,
B2
2
T (log T ) exp 4v
B 2 ifV 256w ,

B2
Inserting these bounds into (4.5.1) we know the Theorem 4.1.3. In order to prove
Theorem 4.1.3 it is not necessary to derive an asymptotic formula for the sum in(4.5.2) An
[K : ]
upper bound of

log log x + O(1) for the sum in (4.5.2) would be sufficient and is
0 rk ( p ) [ K : ],

more easily derived. For instance, since

we see that
2
rk ( p )
rK ( p )
[ K : ]
[ K : ]log log x O (1)

p
p
p x
pc
By Landaus prime ideal theorem.

106

4.6. SKETCH OF THE PROOF OF THEOREM 4.1.5:

We now sketchhow tomodify the proof of Theorem 4. 1.1 to deduceTheorem 4.1.5
in this case, the starting point is the observation that
k

1
1

L , 1 d ... L , 1 d
|d | x
2

kr

exp V

k1 ( 1 ) ... k r ( r )

log log X N ( X ,V )dv,

2

(6.6.1)

where N(X,V) denoted the number of fundamental discriminants d with |d| X such that

k ( ) ... k r ( r )
1

, 1 d ... L , 1 d V 1 1
log log X (6.6.2)
2
2

k1 log L

we
can bound N(X,V) with following analogues of Lemma 4.3

Lemma4.6.1:
L( s , )
Let

representation

be an L-function attached to an irreducible cuspidal automorphic

on GL(m) over

and let d be a fundamental discriminant. Let

107

e 0 0 02 / 2

0 0.4912....
denote the unique positive real number satisfying

. Then,

0 log x / 2
assuming the generalized Riemann hypothesis for
1

, d
2

log L

and log x 2, we have

1
(n ) d (n ) log x / n (1 ) m log | d |

O
1
,

log n
log x
2
log x
n x
log x
2 log x
n

Where the implied constant depends only on

Proof:
This follows from Theorem 4.2.1
We now indicate how to prove Theorem 4.1.5 The primeary difference between the
proof of this theorem and the proof of Theorem 4.1.1 is how we handle the contribution from
the prime powers. By (4.2.3),the contribution from the prime squares to the inequality
( p 2 ) d ( p 2 )
( )
:
log log x,

2
1
2
p2 x
2
log x
p
log p
Giving rise to the extra term on the right hand side of (4.6.2) in the definition of
N(X , V) in this way, the squares of primes contribute to our bounds for the size of these
moments.In contrast to the proof of Theorem 4.1.1 we must handle prime powers pj with j >

d ( p j ) d ( p)
2 differently depending on whether j is odd or even. When j is odd,

and

hence Hypothesis H in a manner analogous to the analysis in 4. If j is even, them

p ' d

d ( p j ) 1
for

and therefore we cannot average over discriminats to estimate their

108

contribution. Instead, we use Hypothesis E to show that the contribution of these primes to
Lemma 4.6.1 is O(1)

( k12 ... kr2 )log log x

With these changes, choosing B, , and A as in 4 and w =

relatively straightforward modification of the proof of proposition 4.4.1 gives

v 2

X
(log
X
)
exp

2 w if 3 V 512 w

,
B2
N ( X ,V ) =
X (log X ) exp 4V
512 w

ifV

B
B2
Theorem 4.1.5 now follow ]s by inserting these bounds into (4.6.1)

4.7 PROFF OF THE THEOREM4.1.4

109

,a

We follow the proof who studied the distribution of primes in short intervals using

( s)
upper bounds for moments of the logarithmic derivative of

near the critical line. For k

a finite Galois extension of

Let

,
r2

2 r1 (2 ) hR
cK
, S ( x ) rK ( n ),
w D
n x
and
S0 ( x )

1
lim( S ( x ) S ( x ))
2 0

So that S(x)=S0(x) for almost all x. perrons formula implies that

2 i
1
xs
S0 ( x )

(
s
)
ds
K
2 i 2i
s

K (s)
Assuming the generalized Riemann hypothesis (GRH)for
( s )
left from

, we move the contour

( s )
= 2 to

= passing over a pole of the integrand at s = 1 and no other

K ( s)
singularities. Here we are implicitly using the generalized Lindelof hypothesis for
t-aspect (which follows from GRH) to justify the contour shift. Thus by the residue
calculation in (4.1.11) and a variable change,
we have
1/2i
1
xs
S0 ( x ) cK x

(
s
)
ds
K
2 i 1/2i
s

1
K 2 it

110

x
1 dt
it
2
1
it
2

in

x erk

Applying this formula twice with the values

S0 ( e k r ) S0 (e r ) cK ( e K 1) e r
1

r /2
e
2

and

x er

, it follows that

1 it
k 2
1

e 1 irt

it
K 2 1 e dt,
it

Giving a Fourier transform relation for all

s0 ( x ) s( x )

since

almost

everywhere, we have

Observing that the integrand on the left-hand side is even and letting

x e r , X T 2,
2X

dT
1
e
1
1
S (e k r ) S ( e r ) cK (e k 1)e r

it
K

dt ,
r
2
1
e
2
it

1
it
k 2

and

eK 1 1 / T
2

, we know that
2

x
x dx
x dx

x
S x S ( x ) cK
S x S ( x ) cK
2
T
T x
T
T x2

0
2

1
1
K it
0
2

l 0

=
l 0

111

1
it
2

1
l
(2 T )2

In follows from this Theorem 4.1.3

1
e it 1
2
dt
1
it
2
K

(2l 1 1) T

1
it
K 2

1
dt
1
it
2
2

1
K it dt.
2

1
X2

2X

x
x

S x S ( x ) cK dx
T
T

1
(log T )[ K : ]
l
l 0 2 T

(log T )[ K : ]
=
T
for any > 0. Theorem 4.1.4 now follows by choosing y = x/T.

BIBLIOGRAPHY

1. E.M. Baruch, Z.Maco, Central value ofautomorphic L-functions, Geom Funct.

Anal.12(2007),333-384.8
2. J. Bernstein A.Reznikov, Sobolev norms of automorphic functionals, Int. Math. Res.
Not. 2002,2155-2174.6

112

3. V. Blomer and J. Brudern, A three squares theorem with almost primes, Bull. Lond.
Math. Soc 37(2005), 507-513,8
4. V. Blomer, G. Haroos, Hybrid boundfor twisted L-functions, J. Reine Angew. Math. 621
(2008), 53-79. 5,6
5. V. Blomer, G. Harros, twisted L-functions over number fields, and Hilberts eleventh
problem, preprint. 57
6. J. Brudern, E Fouvry, Lagranges four squares theorem with almost prime variables, J.
Reine Angew. Math. 454 (1994), 59-96.8
7. D. Cax, Primes of the form x2+my2, john wiley & Sons, New york, 1989, 8.
8. W. Duke, Hyperbolic equidistribution problems and half-integeral weight Maass forms,
invert, math 92(1988),73-90.9
9. A. Good, Beitrage zur Theorie der Dirichletreihen, die Spitzenformen zugeordnet sind, J.
Number Theory 13 (1981), 18-65.7
10. G. Haroos, Uniform approximate functional equation for principal L-functions, Int,
Math. Res. Not. 2002, 923-932; Erratum, ibid. 2004, 959-960.6
11. E.T. Whittaker and G.N. Watson, A course of modern analysis, 4th edition, Cambridge
Univesity Press, 1927.4
12. A.O.L.Atkin and Wen-Ching W. Li,Twists of newforms and pseudo-eigenvalyues fo
w-operaors, Invent. Math.,48(1978), 221-243
13. L.B.Conrey and H.Iwaniec , The cubic moment of central values of automorphic Lfunctions, Ann Math., 151(2000), 1175-1216
14. W. Duke, J.B.Friedlander, and H. Iwaniec, Bounds for automorphic L-functions,
Invent. Math.,112(1993),1-8.
15. W. Duke, J.B.Friedlander, and H. Iwaniec, Bounds for automorphic L-functions, II,
Invent. Math.,115(1994),219-239;erratum,ibid. 140(2000),227-242.
16. W. Duke, J.B.Friedlander, and H. Iwaniec, Bounds for automorphic L-functions,
III,Invent. Math.,143(2001),221-248.
17. A. Good, The square mean of Dirichlet series associated with cusp forms, Math0ematika,29(1982),278-295
18. D.R.Heath-Brown, The fourth powe moment of the Riemannzeta-function, Proc.
London Math. Sec., 38(1979),385-422
113

19. A. Ivic, The Riemann zeta-function the theory of the Riemann zeta-function with
applications, John Wiley & Sons, New York 1985.
20. H.Kim, Functorialty for the exterior square of GL4 and the symmetric fourth of GL2, J.
Ame. Math. Sec., 16(2003),139-183
21. H.Iwaniec, Fourier coefficients of cusp forms and the Riemann zeta function, Seminaire
de Theorie des Nombres de Bordeaux, Annee 1979-1980, expose n
18,1-36

114