E1478

Author(s): Jonathan Sondow, Erwin Just, Norman Schaumberger, Wallace Manheimer and David
Zeitlin
Source: The American Mathematical Monthly, Vol. 69, No. 3 (Mar., 1962), pp. 234-235
Published by: Mathematical Association of America
Stable URL: http://www.jstor.org/stable/2311065
Accessed: 31-03-2015 00:43 UTC

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 Tue, 31 Mar 2015 00:43:42 UTC
All use subject to JSTOR Terms and Conditions
 234 ELEMENTARY PROBLEMS AND SOLUTIONS [March

Re f(n) =n2- n+41

E 1477 [1961, 667]. Proposed by Sidney Kravitz,Dover,New Jersey
It is known that f (n) = n2-n+41 yields prime numbersforn = 1, * * , 40.
Prove that (a) f(n) is neverdivisible by a positive integer <41, (b) f(n) is never
a perfectsquare except for n =41, (c) for each n there exists an m such that
f (m) =f(n)f(n + 1), (d) f (1 722) is the smallest f with four,not necessarilydis-
tinct, prime factors.
SolutionbyD. C. Stevens,New York University.More generally,suppose that
f (n) = n2-n + p yields primenumbersforn- 1, * , p-1. (a) If q, some prime
less than p, divides f(n) for some n, then q divides f(n') where n'_ n (mod q)
and O <n' ? q. But f(n') is prime by assumption,so f(n) is never divisible by a
number <p. (b) If f(n) =a2, the above implies that a>_p. f(p)= p2. Since f is
strictlyincreasing,and f(a) < a2, f(a + 1) > a2 fora > p, f(n) = a2 is impossible for
a> p. (c) f(n)f(n+ 1) =f(n2 +p). (d) The smallest numberwith fourfactors > p
iS p4. But f(p2) <p4 and f(p2+1) >p4. The next possibilityis p3(p+2). The
assumptions imply p +2 is a prime.f(p(p + 1)) =p3(p +2).
The above is valid forp = 3, 5, 11, 17, 41, and possibly othervalues.

Also solvedby J.W. Baldwin,M. T. L. Bizley,BrotherAlfred, A. L. Buchman,F. G. Carty,

D. I. A. Cohen,D. Drasinand J. Yeager(jointly),UnderwoodDudley,PhilipFung,N. V. Glick,
MichaelGoldberg, S. H. Greene,J.E. Homer,Jr.,ErwinJustand NormanSchaumberger (jointly),
P. B. Manchester, WallaceManheimer, D. C. B. Marsh,M. V. Mielke,P. R. Nolan,HughNoland,
R. J. Oberg,StantonPhilipp,N. A. Robins,I. D. Ruggles,D. L. Silverman,The Cincinnatians,
Guy Torchinelli, J. E. Vinson,W. C. Waterhouse, David Zeitlin,and theproposer.Late solutions
by RonaldAlter,D. A. Breault,Gus DiAntonio,A. J. Kokar,D. B. Lloyd,R. H. C. Newton,and
PhilipRose.
The proposerpointedoutthatthesmallestfwiththreefactorsoccursat n =421 (Problem395,
Mathematics Magazine,Nov. 1959), and the smallestf withsix factorsoccursat n = 139,564
(Recreational Mathematics Magazine,June1961,p. 50).

DigitalRootsofPerfectNumbers
E 1478 [1961, 667]. Proposed by JonathanSondow, Universityof Wisconsin
Prove that the digital root of every even perfectnumberexcept 6 is 1.
I. Solution by Erwin Just and Norman Schaumberger,Bronx Community
College.An even perfectnumber must have the form2n-1(2n-1), where 2n-1
is a prime,which impliesn is odd forn > 2. Now 2n-1(2n-1) -1 = 22n-1- 2-1-I
(2"+ 1) (2-1- 1), which is divisible by 9 since each of the last two factorsis
divisible by 3.
II. Solution by Wallace Manheimer,Franklin K. Lane High School, Brook-
lyn,New York.Let N= 2k-1(2k4- 1). For k> 2 and even, 2k - 1 is composite,and
N is not a perfectnumber. For odd k, the digital roots of 2; -1 formthe recur-
ringsequence1, 7, 4, * . . . rootsof 2k-1 are 1, 4, 7, * .
The corresponding .

and N 1 (mod 9).

This content downloaded from 128.235.251.160 on Tue, 31 Mar 2015 00:43:42 UTC
All use subject to JSTOR Terms and Conditions
1962] ELEMENTARY PROBLEMS AND SOLUTIONS 235

III. Remarks by David Zeitlin, RemingtonRand Univac. Since the digital

root of a number (base 10) is the remainderwhen divided by 9, we have the
followingcomments from L. E. Dickson, History of the Theoryof Numbers,
vol. 1, New York, 1952:
p. 9. "Tartaglia (1506-1559) observed that any perfectnumberexcept 6 yields
the remainder1 when divided by 9."
p. 25. "H. Novarese (1887) proved that every perfectnumber of Euclid's
type ends in 6 or 28, and that each one >6 is of the form9k+1."
Also solvedby R. E. Bowen,D. A. Breault,BrotherAlfred,J. L. Brown,Jr.,F. G. Carty,
D. I. A. Cohen,UnderwoodDudley,N. V. Glick,MichaelGoldberg,R. P. Goldberg, SidneyKrav-
itz,WarrenMcCausland,AndrzejMakowski,D. C. B. Marsh,R. J. Oberg,F. D. Parker,D. L.
Silverman, W. C. Waterhouse,
D. C. Stevens,GuyTorchinelli, and theproposer.Late solutionsby
J. F. Dillonand C. F. Pinzka.
Severalsolverspointedout thatthisproblemappears,withsolution,in Maxey Brooke,"On
thedigitalrootsofperfectnumbers,"Mathematics Magazine,vol.34,no.2 (Nov.-Dec. 1960),p. 100.
RestrictedChoices
E 1479 [1961, 668]. Proposed by MortonAbramson,McGill University
Find the numberof ways of choosing k elementsfromn elementsxl; . , Xn
so that no threeconsecutive elementsappear in any choice.
Solution by the proposer. We wish to find the number of distinguishable
arrays of n-k dashes and k dots along a straight line with the restrictionof
no threeconsecutive dots appearing. To do this we array only the n - k dashes
along a straightline and considerthe n - k+ 1 spaces or slots, those between the
dashes, the space beforethe firstdash, and the space afterthe last dash. What
then is the numberof ways of insertingthe k dots into the spaces with at most
2 dots being insertedinto any one space? There are
(n-k+ 1)(n-k+ 1i) (n k+ 1)(k-

such distinguishableinsertionswith exactly i spaces each containing 2 dots.

It followsthat our required numberis
tk n-
n k + 1A {k-

Also solvedby BrotherT. C. Wesselkamper, VirginiaChristian,S. J. Einhorn,Seymour

Geisser,W. M. Gentleman,S. H. Greene,J. F. Leetch,NathanMantel,D. C. B. Marsh,Michael
Skalsky,D. C. Stevens,WarrenTeitelman,and W. C. Waterhouse.Late solutionby J. D. Nulton.
Many ofthesesolutionsdisagreedwiththesolutionpublishedabove.

ofa Polygon
Probability
E 1480 [1961, 668]. Proposed byLeo Flattoand A. G. Konheim,IBM, York-
townHeights,New York
Let the line segment [0, 1] be divided into n + 1 segments by n points

This content downloaded from 128.235.251.160 on Tue, 31 Mar 2015 00:43:42 UTC
All use subject to JSTOR Terms and Conditions