Problems and Solutions

Source: The American Mathematical Monthly, Vol. 118, No. 8 (October 2011), pp. 747-754
Published by: Mathematical Association of America
Stable URL: http://www.jstor.org/stable/10.4169/amer.math.monthly.118.08.747 .
Accessed: 31/03/2013 11:49

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 150.108.161.71 on Sun, 31 Mar 2013 11:49:21 AM

All use subject to JSTOR Terms and Conditions
 PROBLEMS AND SOLUTIONS
Edited by Gerald A. Edgar, Doug Hensley, Douglas B. West
with the collaboration of Mike Bennett, Itshak Borosh, Paul Bracken, Ezra A. Brown,
Randall Dougherty, Tamás Erdélyi, Zachary Franco, Christian Friesen, Ira M. Ges-
sel, László Lipták, Frederick W. Luttmann, Vania Mascioni, Frank B. Miles, Bog-
dan Petrenko, Richard Pfiefer, Cecil C. Rousseau, Leonard Smiley, Kenneth Stolarsky,
Richard Stong, Walter Stromquist, Daniel Ullman, Charles Vanden Eynden, Sam Van-
dervelde, and Fuzhen Zhang.

Proposed problems and solutions should be sent in duplicate to the MONTHLY

problems address on the back of the title page. Proposed problems should never
be under submission concurrently to more than one journal. Submitted solutions
should arrive before February 29, 2012. Additional information, such as gen-
eralizations and references, is welcome. The problem number and the solver’s
name and address should appear on each solution. An asterisk (*) after the num-
ber of a problem or a part of a problem indicates that no solution is currently
available.

PROBLEMS
11593. Proposed by Peter McGrath, Brown University, Providence, RI. For positive
integers k and n, let T (n, k) be the n × n matrix with (i, j)-entry ((i − 1)n + j)k .
Prove that for n > k + 1, det(T (n, k)) = 0.
11594. Proposed by Harm Derksen and Jeffrey Lagarias, University of Michigan, Ann
Arbor, MI. Let
 
n k−1
Y Y j ,
Gn = 
k=1 j=1
k

and let G n = 1/G n .

(a) Show that if n is an integer greater than 1, then G n is an integer.
(b) Show that for each prime p, there are infinitely many n greater than 1 such that
p does not divide G n .

11595. Proposed by Victor K. Ohanyan, Yerevan, Armenia. Let P1 , . . . , Pn be the ver-

tices of a convex n-gon in the plane. Let Q be a point in the interior of the n-gon, and
let v be a vector in the plane. Let ri denote the vector Q Pi , with length ri . Let Q i be
the (radian) measure of the angle between v and ri , and let Fi and Yi be respectively
the clockwise and counterclockwise angles into which the interior angle at Pi of the
polygon is divided by Q Pi . Show that
n
X 1
sin(Q i )(cot Fi + cot Yi ) = 0.
i=1
r i

http://dx.doi.org/10.4169/amer.math.monthly.118.08.747

October 2011] PROBLEMS AND SOLUTIONS 747

This content downloaded from 150.108.161.71 on Sun, 31 Mar 2013 11:49:21 AM

All use subject to JSTOR Terms and Conditions
11596. Proposed by Mehmet Sahin (student) Ankara University, Ankara, Turkey. Let
a, b, c be the side lengths of a triangle, and let ra , rb , rc be the corresponding exradii.
Prove that
a2 b2 c2 ra + rb + rc 2
 
+ 2 + 2 =8 − 2.
ra2 rb rc a+b+c

11597. Proposed by Michel Bataille, Rouen, France. Let f (x) = x/ log(1 − x). Prove
that for 0 < x < 1,
∞
X x n (1 − x)n (n) 1
f (x) = − x f (x).
n=1
n! 2

11598. Proposed by Mowaffaq Hajja, Yarmouk University, Irbid, Jordan. Let S be an

additive semigroup of positive integers. Show that there is a finite subset T of S that
generates S and that is contained in every generating set of S.
11599. Proposed by Fred Galvin, University of Kansas, Lawrence, KS, and Péter
Komjáth, Eötvös Loránd University, Budapest, Hungary. Prove that the following
statement is equivalent to the axiom of choice: for any finite family A1 , . . . , An of
sets, there is a finite set F such that |Ai ∩ F| < |A j ∩ F| whenever |Ai | < |A j |.
Here, equivalence is to be judged in the context of Zermelo-Fraenkel set theory,
not assuming the axiom of choice, and to say that |C| < |D| is to say that there is an
injection from C to D, but none from D to C.

SOLUTIONS

A Generic Lower Bound for a 2 + b2 + c2 in a Triangle

11460 [2009, 844]. Proposed by Cosmin Pohoaţă, Tudor Vianu National College of
Informatics, Bucharest, Romania. Given a triangle of area S with sides of lengths a, b,
and c, and positive numbers x, y, and z, show that
√ x 2 − yz y 2 − zx z2 − x y
 
2
a 2 + b2 + c2 ≥ 4 3S + a2 + b2 + c2 .
x+y+z x y z
Solution by Marian Dinca, Romania. Since the proposed inequality is homogeneous in
x, y, z, we may assume without loss of generality that x + y + z = 1. The inequality
may be written as
√
ma 2 + nb2 + pc2 ≥ 4 3 S,
where m = 1 − 2(x 2 − yz)/x, n = 1 − 2(y 2 − zx)/y, and p = 1 − 2(z 2 − x y)/z. A
corollary of the Neuberg–Pedoe inequality (see comment below) tells us that
√
ma 2 + nb2 + pc2 ≥ 4S mn + np + pm.
It now suffices to show that mn + np + pm = 3, which may be done as follows:
Let t = 2(x y + yz + zx), so that m = 1 − 2(1 − y − z) + 2yz/x = (2x y + 2yz +
2zx − x)/x = (t − x)/x, n = (t − y)/y, and p = (t − z)/z. Then mn = (t − x)(t −
(x /(x etc., so mn + np + pm = (x + y +


y)/(x y) = zt 2 − z + yz)t + x yz yz),
z)t − t + 3x yz /(x yz) = 3.
2 2

Editorial comment. Several solvers made note of the connection between this inequal-
ity and others already in the literature:

748 c THE MATHEMATICAL ASSOCIATION OF AMERICA [Monthly 118

This content downloaded from 150.108.161.71 on Sun, 31 Mar 2013 11:49:21 AM

All use subject to JSTOR Terms and Conditions
 √
• The Weitzenbock √ inequality [8]: a 2√+ b2 + c2 ≥ 4 3 S; in√fact a 2 + b2 + c2 ≥
√
ab + bc + ca ≥ a bc + b ca + c ab ≥ 3(a 2 b2 c2 )1/3 ≥ 4 3 S.
√
• The Hadwiger–Finsler inequality [4]: a 2 + b2 + c2 ≥ 4 3 S + (a − b)2 + (b −
c)2 + (c − a)2 .
• The Neuberg–Pedoe inequality [5, 6]: for a second triangle of area T with sides of
length x, y, z, we have

a 2 (y 2 + z 2 − x 2 ) + b2 (z 2 + x 2 − y 2 ) + c2 (x 2 + y 2 − z 2 ) ≥ 16ST,

with equality if and only if the triangles are similar.

• The following corollary of the Neuberg–Pedoe inequality: Let √ m, n, p be √
any three
positive
√ numbers. A triangle exists with side lengths x = n + p, y = p + m,
z = √m + n, since the triangle inequality holds for x, y, z. This triangle has area
T = mn + np + pm /2. Then noting 2 2 2
√ y + z − x = 2m, etc., by Neuberg–
2 2 2
Pedoe we have ma + nb + pc ≥ 4S mn + np + pm.
The proposed inequality becomes the Hadwiger–Finsler inequality when x = a,
y = b, z = c. It becomes the Weitzenbock inequality when x = y = z.

REFERENCES

1. O. Bottema and M. S. Klamkin, Joint triangle inequalities, Simon Steven 48 (1974) 3–8.
2. L. Carlitz, An inequality involving the area of two triangles, Amer. Math. Monthly 78 (1971) 772. http:
//dx.doi.org/10.2307/2318014
3. L. Carlitz, Some inequalities for two triangles, Math. Mag. 45 (1972) 43–44. http://dx.doi.org/10.
2307/2688380
4. P. von Finsler and H. Hadwiger, Einige Relationen im Dreieck, Comentarii Mathematici Helvetici 10
(1937) 316–326. http://dx.doi.org/10.1007/BF01214300
5. J. Neuberg, Sur les projections et contre-projections d’un triangle fixe, Academie Royale de Belgique 44
(1891) 31–33.
6. D. Pedoe, An inequality for two triangles, Proceedings of the Cambridge Philosophical Society 38 (1943)
397–398. http://dx.doi.org/10.1017/S030500410002209X
7. D. Pedoe, Thinking geometrically, Amer. Math. Monthly 77 (1970) 711–721. http://dx.doi.org/10.
2307/2316201
8. R. Weitzenbock, Über eine Ungleichung in der Dreiecksgeometrie, Mathematische Zeitschrift 5 (1919)
137–146. http://dx.doi.org/10.1007/BF01203160

Also solved by G. Apostolopoulos (Greece), R. Chapman (U. K.), P. P. Dályay (Hungary), Á. Plaza & S. Falcón
(Spain), M. A. Prasad (India), R. Stong, M. Tetiva (Romania), GCHQ Problem Solving Group (U. K.), and the
proposer.

Eigenvalues, Trace, and Determinant

11463 [2009, 844]. Proposed by Xiang Qian Chang, Massachusetts College of Phar-
macy and Health Sciences, Boston, MA. Let A be a positive-definite n × n Hermitian
matrix with minimum eigenvalue λ and maximum eigenvalue 3. Show that
 n  n
n n
− λ ≤ det(A) ≤ −3 .
tr((A + λI )−1 ) tr((A + 3I )−1 )

Solution by BSI Problems Group, Bonn, Germany. An n-by-n positive-definite Hermi-

tian matrix A has only positive eigenvalues, and it has eigenvectors forming a basis.
Since tr((A + x I )−1 ) and det(A) are invariant under change of basis, we may assume

October 2011] PROBLEMS AND SOLUTIONS 749

This content downloaded from 150.108.161.71 on Sun, 31 Mar 2013 11:49:21 AM

All use subject to JSTOR Terms and Conditions
that A is in diagonal form. We then must show, for positive p1 , . . . , pn with minimum
λ and maximum 3, that
!n n
!n
1 Y 1
1
Pn 1
−λ ≤ pi ≤ 1 Pn 1
−3 .
n i=1 pi +λ i=1 n i=1 pi +3

Dividing the first inequality by λ and the second by 3n , it suffices to show

n

!n n n
!n
1 Y Y 1
1
Pn 1
−1 ≤ xi and z i ≤ 1 Pn 1
−1 ,
n i=1 xi +1 i=1 i=1 n i=1 z i +1

where x1 , . . . , xn ≥ 1 and z 1 , . . . , z n ∈ (0, 1].

The function that maps y to log(1/y − 1) is convex on (0, 1/2] and concave on
[1/2, 1). By Jensen’s inequality,
! n  
1 X 1
n log 1 Pn −1 ≤ log −1 for y1 , . . . , yn ∈ (0, 1/2],
n i=1 yi i=1
yi

and
! n  
1 X 1
n log 1
Pn −1 ≥ log −1 for y1 , . . . , yn ∈ [1/2, 1).
n i=1 yi i=1
yi

Exponentiating these inequalities and setting yi = 1/(1 + xi ) in the first and yi =

1/(1 + z i ) in the second yields the desired results.
Also solved by R. Chapman (U. K.), W. J. Cowieson, P. P. Dályay (Hungary), E. A. Herman, O. Kouba (Syria),
J. H. Lindsey II, O. P. Lossers (Netherlands), N. C. Singer, R. Stong, M. Tetiva (Romania), L. Zhou, GCHQ
Problem Solving Group (U. K.), Microsoft Research Problems Group, and the proposer.

More and More Balls in Urns

11464 [2009, 845]. Proposed by David Beckwith, Sag Harbor, NY. Let a(n) be the
number of ways to place n identical balls into a sequence of urns U1 , U2 , . . . in such a
way that U1 receives at least one ball, and while any balls remain, each successive urn
receives at least as many balls as in all the previous urns combined. Let b(n) denote
the number of partitions of n into powers of 2, with repeated powers allowed. (Thus,
a(6) = 6 because the placements are 114, 123, 15, 24, 33, and 6, while b(6) = 6
because the partitions are 111111, 11112, 1122, 114, 222, and 24.) Prove that a(n) =
b(n) for all n ∈ N.
Solution by Jerrold W. Grossman, Oakland University, Rochester, MI. Because a(1) =
b(1) = 1, it suffices to show that both sequences satisfy the same recurrence.
When n is odd, the final urn in an acceptable distribution contains more than half of
the balls, and removing one ball from it gives an acceptable distribution of n − 1 balls.
Thus a(2m + 1) = a(2m) for m ≥ 1. When n is even, then either the final urn contains
more than half the balls, and removing one from it gives an acceptable distribution of
n − 1 balls, or the final urn contains exactly half the balls, and the others contain an
acceptable distribution of n/2 balls. Thus a(2m) = a(2m − 1) + a(m) for m ≥ 1.
When n is odd, a 2-power partition of n contains a 1, and the rest is a 2-power
partition of n − 1. Thus b(2m + 1) = b(2m) for m ≥ 1. When n is even, a 2-power
partition of n contains a 1 plus a 2-power partition of n − 1, or all the parts are even
and the partition is a doubling of a 2-power partition of n/2. Thus b(2m) = b(2m −
1) + b(m) for m ≥ 1.

750 c THE MATHEMATICAL ASSOCIATION OF AMERICA [Monthly 118

This content downloaded from 150.108.161.71 on Sun, 31 Mar 2013 11:49:21 AM

All use subject to JSTOR Terms and Conditions
Editorial comment. The sequence (1, 2, 2, 4, 4, 6, 6, 10, 10, 14, 14, . . .) and its recur-
rence are well known. See Sequence A018819 in The On-Line Encyclopedia of In-
teger Sequences (http://www2.research.att.com/~njas/sequences/) and its
references.
Also solved by M. Andreoli, R. Bagby, C. K. Bailey & M. D. Meyerson, C. Bernardi (Italy), S. Binski,
E. H. M. Brietzke (Brazil), D. Brown, S. M. Bryan, R. Chapman (U. K.), E. Cheng, W. J. Cowieson,
P. P. Dályay (Hungary), C. Delorme (France), S. Eichhorn, D. Finley, J. P. Grivaux (France), J. Hook,
Y. J. Ionin, D. E. Knuth, M. Kochanski, J. Lee (Canada), P. Levande, J. H. Lindsey II, O. P. Lossers (Nether-
lands), J. McKenna, D. Mitchell, J. H. Nieto (Venezuela), Á. Plaza (Spain), B. Popp, S. Post, M. A. Prasad
(India), R. Pratt, B. Schmuland (Canada), J. Simons (U. K.), J. H. Smith, R. Staum, R. Stong, R. Tauraso
(Italy), M. Tetiva (Romania), Z. Vörös (Hungary), S. Xiao (Canada), L. Zhou, Armstrong Problem Solvers,
athenahealth Problem Solving Group, GCHQ Problem Solving Group (U. K.), Missouri State University
Problem Solving Group, and the proposer.

The Determinant of a Very Square Matrix

11467 [2009, 940]. Proposed by Xiang Qian Chang, Massachusetts College of Phar-
macy and Health Sciences, Boston, MA. Find in closed form the determinant of the
n × n matrix with entries ai, j given by
(P
i−1
( j − k)2 if i ≤ j;
ai, j = Pk=0 j
(n if i > j.
2
Pi− j−1 2
k=1 k + k=0 − k)

Solution by Jaime Vinuesa Tejedor, University of Cantabria, Santander, Spain. The

answer is
n n−2 (n + 1)(2n + 1)[(n + 2)n − n n ]
(−1)n−1 .
12
For i from n − 1 to 1, subtracting row i from row i + 1 does not change the deter-
minant but transforms the matrix to a cyclic matrix with constant diagonals 12 , . . . , n 2 .
The determinant of a cyclic matrix with elements a1 , . . . , an is n−1 j=1 a j ω
Q Pn k( j−1)
k=0 ,
where ω = e 2πi/n
(see, for example, A. C. Aitken, Determinants and Matrices,
U. M. T. Oliver and Boyd (1946)).
In our case, since a j = j 2 , the inner sum is n(n + 1)(2n + 1)/6 when k = 0. To
evaluate the other factors, it follows when ζ n = 1 that
n
X
(1 − ζ ) 2
j 2 ζ j−1 = n 2 (ζ − 1) − 2n
j=1

and hence
n
X n 2 (ωk − n+2 )
j 2 ωk( j−1) = n
. (1)
j=1
(ω − 1)
k 2

x n −1
k=1 (x − ωk ); setting x =
Qn−1 n+2
For x 6= 1, we have x−1
= n
yields
n−1    n 
Y n+2 n n+2
− ωk = −1 .
k=1
n 2 n
x n −1
k=1 (x − ω ) =
To multiply the denominators in (1), note that n−1 k
Pn
x k−1 .
Q
x−1
= k=1
Letting x = 1 in each polynomial yields k=1 (1 − ωk ) = n.
Qn−1

October 2011] PROBLEMS AND SOLUTIONS 751

This content downloaded from 150.108.161.71 on Sun, 31 Mar 2013 11:49:21 AM

All use subject to JSTOR Terms and Conditions
 Now we can compute the determinant:
n−1 X
n n−1 n
Y n(n + 1)(2n + 1) Y X 2 k( j−1)
a j ωk( j−1) = j ω
k=0 j=1
6 k=1 j=1

n(n + 1)(2n + 1) Y n 2 (ωk − n+2 )

n−1
n
=
6 k=1
(ωk − 1)2

n(n + 1)(2n + 1) 2 n−1 (−1)n−1 n (n + 2)n − n n

= (n ) .
12 n2 2 nn

Editorial comment. Many of the solvers provided solution formulas not in closed form,
but we have listed them anyway.
Also solved by D. Beckwith, R. Chapman (U. K.), C. Curtis, P. P. Dályay (Hungary), M. Goldenberg & M.
Kaplan, E. A. Herman, Y. J. Ionin, O. Kouba (Syria), O. P. Lossers (Netherlands), K. McInturff, M. Omarjee
(France), Á. Plaza (Spain), J. Simons (U. K.), R. Stong, M. Tetiva (Romania), Z. Vörös (Hungary), P. Zwier,
GCHQ Problem Solving Group (U. K.), Microsoft Research Problems Group, Missouri State University Prob-
lem Solving Group, and the proposer.

A Nonobtuse Altitude Inequality

11480 [2010, 87]. Proposed by Omran Kouba, Higher Institute for Applied Sciences
and Technology, Damascus, Syria. Let a, b, and c be the lengths of the sides opposite
vertices A, B, and C, respectively, in a nonobtuse triangle. Let h a , h b , and h c be the
corresponding lengths of the altitudes. Show that
 2  2  2
ha hb hc 9
+ + ≥ ,
a b c 4

and determine the cases of equality.

Solution by Richard Bagby, New Mexico State University, LasCruces, NM. By scale
invariance it suffices to restrict our attention to triangles of unit area. For such triangles
we have
 2  2  2
ha hb hc 4 4 4
H := + + = 4 + 4 + 4.
a b c a b c

Let C be the largest angle, and consider varying a and b with C fixed.
Since the area is 21 ab sin C, this amounts to fixing ab. We first compare H to the
√ √
value H0 obtained for the triangle with sides lengths ( ab, ab, c0 ), where

c02 = ab + ab − 2ab cos C = 2ab(1 − cos C).

We observe that

c2 − c02 = (a 2 + b2 − 2ab cos C) − 2ab(1 − cos C) = (a − b)2

and

c2 + c02 = (a + b)2 − 4ab cos C.

752 c THE MATHEMATICAL ASSOCIATION OF AMERICA [Monthly 118

This content downloaded from 150.108.161.71 on Sun, 31 Mar 2013 11:49:21 AM

All use subject to JSTOR Terms and Conditions
Consequently, we compute
4 4 4 4 4 4
H − H0 = + 4+ 4− 2 2− 2 2− 4
a4 b c a b a b c0
4(a 2 − b2 )2 4(c4 − c04 )
= +
a 4 b4 c4 c04

2 (a + b) (a + b)2
2
 
4ab cos C
= 4(a − b) − + .
a 4 b4 c4 c04 c4 c04
Now π/3 ≤ C ≤ π/2. Since
c4 c04 ≥ c08 = 16a 4 b4 (1 − cos C)4 ≥ a 4 b4
for π/3 ≤ C ≤ π/2, we see that H − H0 ≥ 0 with equality if and only if a = b. Thus
it suffices to look at isosceles triangles.
In this case, unit area gives a 2 = b2 = 2 csc C and c2 = c02 = 4(1 − cos C) csc C so
we compute
9 sin2 C 9
H0 − = 2 sin2 C + 2
−
4 4(1 − cos C) 4
1 + cos C 9
= 2(1 − cos2 C) + −
4(1 − cos C) 4
2(2 cos C − 1)2 cos C
= ≥ 0,
4(1 − cos C)
with equality if and only if C = π/3 or C = π/2. Thus we get equality for an equilat-
eral triangle or an isosceles right triangle.
Editorial comment. Many solvers noted that
ha 1
=
a cot B + cot C
and substituted x = cot A, y = cot B, and z = cot C, to reduce the problem to
1 1 1 9
+ + ≥
(x + y) 2 (y + z) 2 (z + x)2 4
for x y + yz + zx = 1, which was a problem on the 1996 Iranian Mathematical
Olympiad. However the earliest reference submitted was to J. F. Rigby, Sextic in-
equalities for the sides of a triangle, Univ. Beograd. Publ. Elektrotehn. Fak. Ser. Mat.
Fiz. no. 498–541 (1975) 51–58.
Also solved by A. Alt, G. Apostolopoulos (Greece), M. Bataille (France), M. Can, C. Curtis, P. P. Dályay
(Hungary), J. Fabrykowski & T. Smotzer, F. Holland (Ireland), J. H. Lindsey II, C. Pohoata (Romania), C. R.
Pranesachar (India), R. Stong, E. Suppa (Syria), M. Tetiva (Romania), Z. Vörös (Hungary), GCHQ Problem
Solving Group (U. K.), and the proposer.

Countable Dense Partition

11481 [2010, 182]. Proposed by Ron Rietz, Gustavus Adolphus College, St. Peter, MN.
Let X be a countable dense subset of a separable metric space M with no isolated
points. Show that there exists a countable partition (X 1 , X 2 , . . .) of X such that each
X n is dense in M.

October 2011] PROBLEMS AND SOLUTIONS 753

This content downloaded from 150.108.161.71 on Sun, 31 Mar 2013 11:49:21 AM

All use subject to JSTOR Terms and Conditions
Solution by Michael Barr, McGill University, Montréal, QC, Canada. We derive the
conclusion under the weaker condition that M is first countable and Hausdorff. (A
space M is first countable if for every x in M there is a sequence U1 , U2 , . . . of neigh-
borhoods of x such that for every neighborhood U of x, at least one of the Ui is a
subset of U . It is Hausdorff if distinct points have disjoint neighborhoods.)
We begin with a few observations. First, since M is Hausdorff and has no isolated
points, if N is a neighborhood of x, then N − {x} is open and nonempty. Second, M
is infinite. Third, any dense subset Y of M is infinite.
The plan now is to first show that we can split (partition into two components) any
countable dense subset Y of M into dense subsets U and V . If we can do that, the
problem is solved, because we can take Y = X , take X 1 = U , split V in the same
fashion to get an X 2 , then split the other part of V to get X 3 , and so on. To split Y
in the desired manner, we begin by showing that for every y in Y there are disjoint
sequences of points in Y − {y} that converge to y. (A sequence (y j ) converges to y if
for every neighborhood N of y there is an index k such that for all j ≥ k, y j ∈ N .)
By hypothesis, there is a sequence (N1 , N2 , . . .) of neighborhoods of y such that
every neighborhood of y contains one of the N j . It will be convenient to require that
N j ⊆ Nk for j ≥ k, and this can be achieved by replacing Nk with ∩kj=1 N j . Since M
has no isolated points, each N j contains a point other than y. Since Y is dense, for
each j, Y ∩ (N j − {y}) is nonempty.
We construct a sequence (s1 , s2 , . . .) by taking y j from Y ∩ (N j − {y}), j =
1, 2, . . .. Every neighborhood of y contains some N j as a subset, and hence all Nk with
k ≥ j since the N j are nested. Thus, (s j ) has the property that for every neighborhood
N of y, there is an index k such that for k ≥ j, sk ∈ N . That is, (s j ) converges to y
without hitting y.
By deleting duplicates and renumbering, we make (s j ) a sequence of distinct ele-
ments. The disjoint sequences (s2 j ) and (s2 j+1 ) then provide our sequences converging
to y from within Y .
Since Y is countable, there is a sequence (y1 , y2 , . . .) of distinct elements of Y that
exhaust Y . Using our construction above, for each k ≥ 1 we build a disjoint pair (u j,k )
and (v j,k ) of sequences, free of repetion, drawn from Y − {yk } and converging to yk .
We put Uk = {u j,k : j ≥ 1} and Vk = {v Sj,k : j ≥0 1}. We put S U10 = U1 , V10 = V1 , and
for k ≥ 2 we put Uk = Uk − ∪ j<k (U j V j ), Vk = ∪ j<k (U j V j ). Since sequences
0

converging to distinct limits share only finitely many entries, each Uk0 and Vk0 is infinite.
Seen as sequences, Uk0 and Vk0 converge within Y − {yk } to yk . Furthermore, any two
distinct sets from the collection of Uk0 and Vk0 are disjoint.
Now take U = ∪∞ 0 ∞ 0
k=1 Uk , V = ∪k=1 Vk . Then Y ⊆ U and Y ⊆ V , so U = V = M.
By construction, U and V are disjoint. Adding any unused elements of Y to U , we
have our claimed partition of Y into disjoint subsets U and V , which completes the
proof.
Also solved by J. Bryant, P. Budney, B. S. Burdick, R. Chapman (U. K.), J. Cobb, W. J. Cowieson, N. Eldredge,
P. J. Fitzsimmons, J. Grivaux (France), J. W. Hagood, E. A. Herman, G. A. Heuer, O. P. Lossers (Netherlands),
M. D. Meyerson, V. Pambuccian, D. Rose, K. A. Ross, B. Schmuland (Canada), J. Simons (U. K.), R. Stong, B.
Tomper, J. Vinuesa (Spain), GCHQ Problem Solving Group (U. K.), Northwestern University Math Problem
Solving Group, NSA Problems Group, and the proposer.

754 c THE MATHEMATICAL ASSOCIATION OF AMERICA [Monthly 118

This content downloaded from 150.108.161.71 on Sun, 31 Mar 2013 11:49:21 AM

All use subject to JSTOR Terms and Conditions