An Elementary Proof of the Formula $\Sum^\infty_{k = 1} 1/k^2 = \pi^2/6$

Author(s): Yoshio Matsuoka

Source: The American Mathematical Monthly, Vol. 68, No. 5 (May, 1961), pp. 485-487
Published by: Mathematical Association of America
Stable URL: http://www.jstor.org/stable/2311110 .
Accessed: 31/01/2015 06:10

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.

Mathematical Association of America is collaborating with JSTOR to digitize, preserve and extend access to
The American Mathematical Monthly.

http://www.jstor.org

This content downloaded from 128.235.251.160 on Sat, 31 Jan 2015 06:10:11 AM

All use subject to JSTOR Terms and Conditions
1961] MATHEMATICAL NOTES 485

I k(t)+agi(t) I = ||kjj+ |gi(t) . From (3) we conclude that I gi(t) <e. Since gi can
be any of the functionsoccurringin (2), it followsthat tG V. That is, k only as-
sumes its maximumon V, and hence not on B, whichcontradictsthe assumption
that B has property(i). Therefore,if B has property(i), FCB as claimed.
References
1. RichardArensand I. M. Singer,Functionvalues as boundaryintegrals,Proc. Amer.
Math. Soc., vol. 5, 1954,pp. 735-745.
2. Heinz Bauer,Un problhme de Dirichletpour la frontiere
de Silov d'un espace compact,
C. R. Acad. Sci. Paris,vol. 247, 1958,pp. 564-567.
3. ErrettBishopand Karel de Leeuw,The representation of linearfunctionalsby measures
on setsofextremepoints(to appear).
4. L. H. Loomis,An Introduction to AbstractHarmonicAnalysis,New York,1953.

AN ELEMENTARY PROOF OF THE FORMULA F,'* 1/k2

=ir2/6

YOSHIO MATSUOKA,Kagoshima University,Japan

The formulain the title has been well known,but its various known proofs
are less elementary,(see for examnple,[1], p. 219, 360, [3], p. 237, 267, 324,
[4], Problem 99, p. 196, [5], p. 379). The followingproofis quite elementaryin
character.
For any positive integern, we consider
fr //2 2n
cos tdt.

Applying integrationby parts twice, we obtain

J/2 nCos tdt= [l s2 t]12 + 2n 2n-1 tsin tdt

7/2 (w12 2 n1 CS-21S 2

t2
n[co
2
C02n-1 t
sin ] - n t [-(2n-1) cos tsin t + cos t]dt

- -2n2 J t cos
7r/2 2 2n
tdt+ n(2n-1) Jt w/2 2
cos
2n-2
Idt

- 2n212 + n(2n- 1)12n_2,

whereI2n = fo/2t2 CoS2n tdt.Hence

+ n(2n -
r 7r/2 2n
cos tdt.
-2nt2I2n 1)I2n-2 =

As is known, (see, forexample, [2], p. 226)

/2 2it (2n - 1)!! ir
cos tdt==
Jo II 2
~~~~~~~~~~~(2n)

This content downloaded from 128.235.251.160 on Sat, 31 Jan 2015 06:10:11 AM

All use subject to JSTOR Terms and Conditions
486 MATHEMATICAL NOTES [May

where, as usual,
(2n)!!!1 2.4 ... (2n - 2)(2n), 0! ! = 1;
(2n + 1)!! = 1.3 * (2n - 1)(2n + 1), (-1) !! 1.
Thus we have
(2n - 1)!! 7r
- 2n2I2 + n(2n -1)I2n_2 =
(2n)!! 2

(2n)!! (2n - 2)!! -r 1

(2n - 1)!! (2n - 3)!! I4 n2
This implies that

(2n
(2n)!!
- 1)!
0!!
01)!!
n

k-l
r[

7r
(2k
n
(2k)!!

1
1)!!
(2k - 2)!! I
(2k -3)!! ]
4 k-1 2

and hence that

(2n)!! w3 ' 1 w [I-72

(2n- 1)!! 24 4 k=1 k2 4 L6k6 k2

It is sufficientto prove that

(1)lim (2n)!! = 0
(2n - 1)!!2
Now we have
, 2 7r2 2 2n
rf22 2
12n t Cos td? td) sin tCos
[C2 r2 72
U0
2n 2n+2
-[ cos tdt - cos tdi]
4 oo
73 r(2n - 1)!! (2n + 1)!! 73 (2n -1)!!
w
8 L (2ni)! ! (2n + 2)!!j 8 (2n + 2)!1
Therefore
(2n) !7! 3 1
(2n- 1)!! - 8 2n + 2
Thus we establish (1); hence the formulais proved.
Finally, the author wishes to express his thanks to a refereefor valuable
criticism.

This content downloaded from 128.235.251.160 on Sat, 31 Jan 2015 06:10:11 AM

All use subject to JSTOR Terms and Conditions
1961] CLASSROOM NOTES 487

References
1. T. J. Bromwich, An Introduction
to theTheoryofInfinite Series(2nd ed.), London,1926.
2. J. Edwards,The IntegralCalculus,vol. 1, New York,1954.
3. K. Knopp,Theoryand Applicationof Infinite Series,Londonand Glasgow,1928.
4. A. Ostrowski, Vorlesungen
iuberDifferential-und Integralrechnung,vol. 2, Basel, 1951.
5. , Vorlesungen
uber Differential-
und Integralrechnung,vol. 3, Basel, 1954.

CLASSROOM NOTES
EDITED BY C. 0. OAKLEY, HaverfordCollege
All material
forthisdepartment
shouldbe sentto C. 0. Oakley,Department
ofMathe-
matics,Haverford
College,
Haverford,
Pa.

AN ABSTRACT FORMULATION OF A PROBLEM RELATED TO

GOLDBACH'S CONJECTURE
A. A. MULLIN, University
of Illinois
0. Introduction.Goldbach's conjecture [1] states that every even integer
greater than four can be representedas the sum of two odd prime numbers.
This paper gives an abstract formulationof a problemthat is at least as general
as Goldbach's conjecture. Specificallythe problem deals with a minimalcondi-
tion associated withsubsets of odd integersin the additive semigroupof strictly
positive integers.
1. Definitions.
DEFINITION 1.1. Let (A, *) denotea nonemptyset A together with a closed
binarycompositionlaw "*" definedon A. Call (A, *) an algebraic system.By a
mutantof(A, *) is meanta subsetMIIofA thatsatisfiestheconditionthatM * MC Ml,
whereM * M= {a * b: a G M, bC M} and M is theset of all of theelementsof A
notin Al. ClearlyM mustbe a propersubsetofA. If and onlyif all of theelements
ofA are idempotent withrespectto "*" lettheemptysetbe theonlymutantof (A, *).
Withthisconventioneveryalgebraicsystemhas a mutant.
DEFINITION 1.2. A mutantM of (A, *) is said to be a maximalmutantof (A, *)
if thereis no mutantof (A, *) whichproperlycontainsM. If and onlyif all of the
elementsofA are idempotent withrespectto '*" lettheemptysetbe theonlymaximal
mutantof (A, *). WiththisconventiontheHausdorffextremalprinciple [2] guar-
antees thateveryalgebraicsystemhas a maximal mutant.
DEFINITION 1.3. Let N be a maximal mutantof (A, *). If thereexistsa non-
emptyclass of mutantsMiCN of (A, *) such thatMi * M N, for all iCI, then
nier Mi is said to be a potentialminimaxmutantin N of (A, *). WhenN is poten-
tial minimax mutantin N of (A, *) it is said to be thetrivialpotentialminimax
mutantin N of (A, *).

This content downloaded from 128.235.251.160 on Sat, 31 Jan 2015 06:10:11 AM

All use subject to JSTOR Terms and Conditions