Annals of Mathematics

An Elementary Proof of the Prime-Number Theorem

Author(s): Atle Selberg
Source: Annals of Mathematics, Second Series, Vol. 50, No. 2 (Apr., 1949), pp. 305-313
Published by: Annals of Mathematics
Stable URL: http://www.jstor.org/stable/1969455 .
Accessed: 27/09/2014 21:17

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

Annals of Mathematics is collaborating with JSTOR to digitize, preserve and extend access to Annals of
Mathematics.

http://www.jstor.org

This content downloaded from 188.44.67.235 on Sat, 27 Sep 2014 21:17:42 PM

All use subject to JSTOR Terms and Conditions
ANNALS OF MATHEMATICS
Vol. 50, No. 2, April,1949

AN ELEMENTARY PROOF OF THE PRIME-NUMBER THEOREM

ATLE SELBERG
(Received October14, 1948)

1. Introduction
In thispaperwillbe givena newproofoftheprime-number theorem,whichis
elementary in thesensethatit usespractically
no analysis,exceptthesimplest
propertiesofthelogarithm.
We shallprovetheprime-number theoremin theform

(1.1) .lim =1(x) 1

Z__00 X

whereforx > 0, tQ(x)is defined

as usualby
(1.2) #(X) = E logp,

theprimes.
p denoting
The basic new thingin the proofis a certainassymptotic
formula(2.8),
whichmaybe written

(1.3) t(x) logx + logpa () =2x logx + O(x).

Fromthisformula thereare severalwaysto deducetheprime-number theorem.

The wayI present??2-4ofthispaper,is chosenbecauseit seemsat thepresent
to be themostdirectand mostelementary way.1 But forcompleteness it has
to be mentioned thatthiswas not myfirstproof. The originalproofwas in
factratherdifferent,and made use of the following resultby P. Erdos,that
positivefixednumber6, thereexista K(a) > 0 and an xo = x(a)
foran arbitrary,
suchthatforx > xo, therearemorethan
K(a) Xlog x
fromx to x + Ax.
primesin theinterval
My first thenotations
proofthenranas follows:Introducing

lim. t() = a, fim,() -A

- x x;
result
onecan easilydeducefrom(1.3),usingthewell-known

(1.4) E logP =
,<? X
logx + 0(1),

1
Because it avoids the conceptof lowerand upperlimit. It is in fact easy to modify
the proofin a fewplaces so as to avoid the conceptof limit at all, of course (1.1) would
thenhave to be stated differently.
305

This content downloaded from 188.44.67.235 on Sat, 27 Sep 2014 21:17:42 PM

All use subject to JSTOR Terms and Conditions
306 ATLE SELBERG

that
(1.5) a+ A = 2.
Next,takinga largex, with
(x) ax + o(x),
form
one can deducefrom(1.3) in themodified

(1.6) (d(x) -ax) logx + E logp ( () -A x= 0(),

that,fora fixedpositivenumber6,onehas

(1.7) s6(p > (A _ )x

setofprimes? x with
exceptforan exceptional

log P O(lo_ X).

Alsoone easilydeducesthatthereexistsan x' in therange-V1< x' < x, wit

6 (x') = Ax' + o(x').
andx' insteadofx,onededucesthe
Againfrom(1.6) witha andA interchanged,

(1.8) * (J;)< (a + 6) -,

set ofprimes? x' with

exceptforan exceptional

lop a
FromErd6s'resultitis thenpossibletoshowthatonecanchoseprimesp and p'
to anyoftheexceptional
notbelonging sets,with

x< x/ < (l + a) x.
P P P
Thenwe getfrom(1.7) and (1.8) that
X
(A- O < < (mu) < (a+6) < (a+ X

so that
A < (a + 6)(1 + 6).
or making6 tendto zero
A a.

This content downloaded from 188.44.67.235 on Sat, 27 Sep 2014 21:17:42 PM

All use subject to JSTOR Terms and Conditions
 THE PRIME-NUMBER THEOREM 307

Hencesincealso A ? a and a + A = 2 we havea = A = 1, whichprovesour

theorem.
Erd6s'resultwas obtainedwithoutknowledge of mywork,exceptthatit is
based on myformula(2.8); and afterI had theotherpartsoftheaboveproof.
His proofcontainsideas relatedto thosein the above proof,at whichrelated
ideashe had arrivedindependently.
The methodcan be appliedalso to moregeneralproblems.For instance
one can provesome theorems provedby analyticalmeansby Beurling,but
theresultsare notquiteas sharpas Beurlings.2Alsoone can provetheprime-
number theorem forarithmetic onehas thentouse inadditionideas
progressions,
and resultsfrommypreviouspaperon Dirichlets theorem.
Ofknownresultswe use frequently besides(1.4) also itsconsequence
(1.9) t (x) = 0(x).
Throughoutthe paper p, q and r denoteprimenumbers.A(n) denotes
M6bius'number-theoretic function,r(n) denotesthe numberof divisorsof n.
The letterc willbe usedto denoteabsoluteconstants,
and K to denoteabsolute
positiveconstants.Some of the moretrivialestimations are not carriedout
butleftto thereader.
2. Proofofthebasicformulas
whenx is a positivenumber
We write, andd a positiveinteger,

(2.1) Xd = kd.. = log d'

jut(d)

and ifn is a positiveinteger,

(2.2) On= =n,= = Ed/n Xd

Thenwe have
flog2
x, forn=1,
(2.3)
(2*3) On
- Jlog
)2logplogq,
forn== pa, aa
p logX2/p,forn paql,
>1
> > 1,

LO,forall othern.
The firstthreeof thesestatements followreadilyfrom(2.2) and (2.1), the
fourth n square-
is easilyprovedby induction.Clearlyit is enoughto consider
free,thenifn = PlP2 * Pk ,
On,, = On/rPi,z - On1Pkl/Pk

partof(2.3) follows.
Fromthistheremaining
2 A. BEURLING: Analysede la loi asymptotique
de la distribution
des nombrespremiers
gneralisks,Acta Math., vol. 68, pp. 255-291(1937).
3 These Annalsthisissue,pp. 237-304.

This content downloaded from 188.44.67.235 on Sat, 27 Sep 2014 21:17:42 PM

All use subject to JSTOR Terms and Conditions
308 ATLE SELBERG

Now considerthe expression

On
=
d XdXd F1=x E
I'd
+
d
OZ [ ]
Xd)
(2.4) zn! n<z din d<z d dx d_

j X
=X E A(d) og2 Z + o u2 )=xE(d)
d O(X).
d<z d d \dx d d oz
lo2
d
+

This on the otherhand is equal to, by (2.3),

2

ORn-log2x + E log p log -

n<x Pa<X P

+ 2 log p logq= log2p

VcaqB?x p<:
v<q

(2.5) + E logplogq+O (1log p log)

pq?z \p~z P1

+0( Z og2x) +O( E logplogq)

jpa < paq< z
x>l a>1

+ log2x = log2p++ logp logq + O(x).

<X Vpq~z

The remainderterm being obtained by use of (1.4) and (1.9). Hence from
(2.4) and (2.5),

(2.6) E log2p + E logp logq = x 2L?(dQ)log2x + O(x).

(2.6) pq<_ dz d d

It remainsnow to estimatethe sum on the right-hand-side.To this purpose

we need the formulas

(2.7) = log z + c1 + 0(Z-*),

and

(2.7) = og2 Z + c2log z + C3+ O(Z )

where the c's are absolute constants,(2.7) is well known,and (2.7') may be
easily derivedby partial summationfromthe well-knownresult

>T(v) = zlogz+c4z+0(A/).

From (2.7) and (2.7') we get

log2z = 2 2
p:z X
(G) +
V
c5
w'VXV
I+ C6+ O(z1).

This content downloaded from 188.44.67.235 on Sat, 27 Sep 2014 21:17:42 PM

All use subject to JSTOR Terms and Conditions
 THE PRIME-NUMBER THEOREM 309

By takingherez = x/d,we get

1,,ud 2X = 2 E jAxd), r(v) +C5 iAg(d) E

d?xd d d<z d a xld V d~z d v~xld V

+ c6 Z ,8(d) + O(x Z dt4) = 2Z /(d)T(v)

d~x d d~z dv-z< dv

+ c6 E 1d) + C6 E ) + 0(1)
dax dv dz_ d

=2 n d r + c5 Z (d)

+ 0(1) = 2 j - + c5+ 0(1) = 2 logx + 0(1).

Weused herethatEd/n ,(d) r(n/d) = 1,and thewell-known d <z (,u(d))/d= 0(1).

Now (2.6) yields
(2.8) E log2p + E logp logq = 2x logx + O(x).
pz pqz

in theformgivenin theintroduction
mayalso be written
Thisformula

(2.9) t6(x)logx + Z logp 6 - = 2x logx + O(x),

that
by noticing
Z
vZ
log2p = t(x) log x + O(x).

we getfrom(2.8)
By partialsummation

(2.10) E log p+ E logp logq 2x + 0 (x)

V:5.x VQ_: logpq logx
Thisgives

E
Pq_$
logp logq = Pz
I logp E
q~zIp
logq = 2x P-z logP z P

-E logp E q gllogrg 0 E log \

pE~z: qr~z/p log qr V dp (1+log-))

= 2x logx - E log qlogr (f) + O(x loglogx).

qr~z logqr \qr/
thisforthesecondtermin (2.8) we get
Inserting

(2.11) t,(x)logx = E log p log q 6 ( - ) + O(x log logx).

V.< log pq \pq/

This content downloaded from 188.44.67.235 on Sat, 27 Sep 2014 21:17:42 PM

All use subject to JSTOR Terms and Conditions
310 ATLE SELBERG

now
Writing
6(x) = x + R(x) , (2.9) easily gives

(2.12) R(x) logx = -Z logp R ( + O(x),

posx p
and (2.11) yieldsin thesamemanner
lo p
(2.13) R(x) logx log qR + O(x log log x),
,~log pq \pq/
since
lo1glo1gq =~logx + O(logO
logx),
x pq logpq
from
by partialsummation
whichfollows
E logP logq = x + O(logx),
log2
vq < pq

whichagainfollowseasilyfrom(1.4).
The (2.12) and (2.13) yield
2j R(x)IlogzxS E logp RI-)

E logp logq |R (X)

+ + O(x log logx).
logpq \pq vq<
Fromthis,by partialsummation,

I
2j R(x)logx < pE n logp+ E log Pq }

{ R - R) + }+O(x log logx),

orby (2.10)
2 R(x)I
logz 2 2n{R(n) - (n l)|

+0 1 logn ) +- O(x loglogx)

=2 IR ) + (Xloglog x),

This content downloaded from 188.44.67.235 on Sat, 27 Sep 2014 21:17:42 PM

All use subject to JSTOR Terms and Conditions
 THE PRIME-NUMBER THEOREM 311

or

(2.14) IR(x)I < log1X n?Ei| R(?)

\l
+0( loglogXx)
\logx /
whichis the resultwe will use in the following.4

3. Some propertiesof R(x)

From (1.4) we get by partial summationthat

E 6(n) = log x + 0(1),

n~x fl2

or
z R(n) 0(1).

This means thereexistsan absolute positiveconstantK1, so that forall x > 4

and x' > x,
() < K1.
(3.1) E

Accordinglywe have, if R(n) does not change its sign betweenx and x', that
thereis a y in the intervalx < y ? x', so that

(3.2) < K221.

19xi

This is easily seen to hold trueifR(n) changesthe signalso.'

Thus for an arbitraryfixedpositive a < 1 and x > 4, there will exist a y
in the intervalx y 5_ e K2/6 x, with
(3.3) (R(y) <6y.
From (2.10) we see that fory < y',

0 < E log p ? 2(y' - y) + ? Y

fromwhichfollowsthat

R(y') - R(y) y' - y + o (I'Y,)

4 Apparentlywe have herelost somethingin the orderof the remainder-term
compared
to (2.8). Actuallywe could insteadof (2.14) have used the inequality

I5
IR(x) 2 Lon +
1 1: IR (n ? (lga

whichcan be provedin a similarway.

5 Because therewill then be a I R(y) I < log y.

This content downloaded from 188.44.67.235 on Sat, 27 Sep 2014 21:17:42 PM

All use subject to JSTOR Terms and Conditions
 312 ATLE SELBERG

Hence, if y/2 ? y' < 2y,y > 4,

R(y') -R(y) y' - y + o y

or

IR(y') I _ R(y)1+ I y'- y) + O (l'y,).

Now consider an interval (x, eK2I x), according to (3.3) there exists a y
in this intervalwith
I R(y) I < by.
Thus forany y' in the intervaly/2 ? y' < 2 y, we have

I R(y') I _ by+ IY
R~~~y') - y I + K3l '
g x'
or

R(y') < 26 + 1 I + l3

Hence ifx > e "'/ and e~812) ? y'/yS we get

R(y') 2+(e812-1+_45
I~ < 25 + (e21) + 3 < 43.

Thus forx > e K/8 the interval (x, e K2/ZX) will always contain a sub-interval
such that JR(z) J< 48zif z belongs to this sub-interval.
(yi, e8/2y1),

4. Proofof the prime-number

theorem
We are now goingto prove the
THEOREM.

limx = 1.
Zx-*oo X

Obviouslythis is equivalentto

(4.1) lim =() 0.

_-*oO x

We knowthat forx > 1,

(4.2) jR(x) I < K4x.
Now assume that forsome positivenumbera < 8,
(4.3) j R(x) I < ax,
holds for all x > x0. Taking 5 = a/8, we have accordingto the preceding

This content downloaded from 188.44.67.235 on Sat, 27 Sep 2014 21:17:42 PM

All use subject to JSTOR Terms and Conditions
 THE PRIME-NUMBER THEOREM 313

section (since we may assume that xo > eK3I8),that all intervals of the type
(x, eK2/8X)withx > xo, containan interval(y, e812y) such that
(4.4) | R(z) I < az/2,
fory < z ? e6/2y.
The inequality(2.14) thengives,using (4.2),

iR(x) I <! lo x E R (n + ? (+Ixc

<K4-X '< N 0 oxxz Rx\
log1x (Ixo)<n~zxn log1x n?(z 0o)n x n \/logx)X
writingnow p = using (4.3)
eK2/8,we get further, and (4.4),

R~)I<ax > l ax
i R(x) < log X n?(zlxo) n 2 log x ?<?<(log (zxz0)/logp)

n = ax- 2 log x ifv (log(xzlo)logp)2

y,<n!zy~e(8-2)
a2
(\/log x) __

p-<yr~v

x
VAnalog 4 log p + (

a ~~x+ 0( Jx< a(13of~)

( 256K2) ? V log x) < t( 300K )2
forx > xi. Since the iteration-process
( 2

an+1 = an 11 - _

obviouslyconvergesto zero if we startforinstancewitha, = 4 (one sees easily

that thenan < K5/V/n),thisproves (4.1) and thus our theorem.
FINAL REMARK. As one sees we have actually never used the full force of
(2.8) in the proof,we could just as well have used it with the remainderterm
o(x log x) instead of O(x). It is not necessaryto use the full force of (1.4)
either,ifwe have herethe remainder-term o(log x) but in additionknowingthat
t~(x) > Kx for x > 1 and some positive constantK, we can still prove the
theorem. However,we have thento make some changein the argumentsof ?3.

THE INSTITUTE FOR ADVANCED STUDY

AND
SYRACUSE UNIVERSITY

This content downloaded from 188.44.67.235 on Sat, 27 Sep 2014 21:17:42 PM

All use subject to JSTOR Terms and Conditions