Problems and Solutions

Reviewed work(s):
Source: The American Mathematical Monthly, Vol. 118, No. 2 (February 2011), pp. 178-185
Published by: Mathematical Association of America
Stable URL: http://www.jstor.org/stable/10.4169/amer.math.monthly.118.02.178 .
Accessed: 22/09/2012 23:58

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 [email protected].

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

http://www.jstor.org
 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. Submitted solutions should arrive
at that address before June 30, 2011. Additional information, such as general-
izations and references, is welcome. The problem number and the solver’s name
and address should appear on each solution. An asterisk (*) after the number of
a problem or a part of a problem indicates that no solution is currently available.

PROBLEMS
11551. Proposed by Gregory Galperin, Eastern Illinois University, Charleston, IL, and
Yury Ionin, Central Michigan University, Mount Pleasant, MI. Given a finite set S of
closed bounded convex sets in Rn having positive volume, prove that there exists a
finite set X of points in Rn such that each A ∈ S contains at least one element of X
and any A, B ∈ S with the same volume contain the same number of elements of X .
11552. Proposed by Weidong Jiang, Weihai Vocational College, Weihai, China. In tri-
angle ABC, let A1 , B1 , C1 be the points opposite A, B, C at which the angle bisectors
of the triangle meet the opposite sides. Let R and r be the circumradius and inradius
of ABC. Let a, b, c be the lengths of the sides opposite A, B, C, and let a1 , b1 , c1 be
the lengths of the line segments B1 C1 , C1 A1 , A1 B1 . Prove that
a1 b1 c1 r
+ + ≥1+ .
a b c R
11553. Proposed by Mihály Bencze, Brasov, Romania. For a positive integer k, let α(k)
be the largest odd divisor of k. Prove that for each positive integer n,
n
n(n + 1) X n − k + 1 n(n + 3)
≤ α(k) ≤ .
3 k=1
k 3

11554. Proposed by Zhang Yun, Xi’an Jiao Tong University Sunshine High School,
Xi’an, China. In triangle ABC, let I be the incenter, and let A0 , B 0 , C 0 be the reflections
of I through sides BC, CA, AB, respectively. Prove that the lines AA0 , BB0 , and CC0 are
concurrent.
11555. Proposed by Duong Viet Thong, National Economics University, R 1 Hanoi, Viet-
nam. Let f be a continuous real-valued function on [0, 1] such that R c 0 f (x) d x = 0.
Prove that there exists c in the interval (0, 1) such that c2 f (c) = 0 (x + x 2 ) f (x) d x.
doi:10.4169/amer.math.monthly.118.02.178

178 c THE MATHEMATICAL ASSOCIATION OF AMERICA [Monthly 118

11556. Proposed by Pál Péter Dályay, Deák Ferenc High School, Szeged, Hungary.
For positive real numbers a, b, c, d, show that

9 9 9 9
+ + +
a(b + c + d) b(c + d + a) c(d + a + b) d(a + b + c)
16 16 16
≥ + + .
(a + b)(c + d) (a + c)(b + d) (a + d)(b + c)

11557. Proposed by Marius Cavachi, “Ovidius” University of Constanta, Constanta,

Romania. Let S be a finite set of circles in the Cartesian plane having the property that
any two circles in S intersect in exactly two points, each circle encloses the origin, but
no three circles share a common point. Construct a graph G by taking as the vertices
the set of all intersection points of circles in S, with edges corresponding to arcs of a
circle in S connecting vertices without passing through any intermediate vertex. (Thus,
with four circles, there are 12 vertices and 24 edges.) Show that the resulting graph
contains a Hamiltonian path.

SOLUTIONS

An Arctan Series
11438 [2009, 464]. Proposed by David H. Bailey, Lawrence Berkeley National Labo-
ratory, Berkeley, CA, Jonathan M. Borwein, University of Newcastle, Newcastle, Aus-
tralia and Dalhousie University, Halifax, Canada, and Jörg Waldvogel, Swiss Federal
Institute of Technology ETH, Zurich, Switzerland. Let
∞  
X x −1
P(x) = arctan √ √ .
k=1
(k + x + 1) k + 1 + (k + 2) k + x

(a) Find a closed-form expression for P(n) when n is a nonnegative integer.

(b) Show that limx→−1+ P(x) exists, and find a closed-form expression for it.
Solution by Hongwei Chen, Christopher Newport University, Newport News, VA.
(a) Notice that

√1 √1
√ √
k+1
− k+n k+n− k+1
= √ √ .
1+ √1 · √1 1+ k+1· k+n
k+1 k+n

Rationalizing the numerator gives

√1 − √1
k+1 k+n n−1
= √ √ .
1+ √ 1
k+1
1
· √k+n (k + n + 1) k + 1 + (k + 2) k + n

From the identity

α−β
 
arctan α − arctan β = arctan ,
1 + αβ

February 2011] PROBLEMS AND SOLUTIONS 179

we obtain
∞  
X n−1
P(n) = arctan √ √
k=1
(k + n + 1) k + 1 + (k + 2) k + n
∞  
X 1 1
= arctan √ − arctan √ .
k=1
k+1 k+n

Clearly, P(1) = 0. The series for P(0) telescopes to give

1 π
P(0) = − arctan 1 + lim arctan √ =− .
k→∞ k+1 4
In general, for n ≥ 2, the series telescopes into the form
n
X 1
P(n) = arctan √ .
k=2
k
(b) Now use the inequality arctan t < t for t > 0. If k ≥ 2 and x ≥ −1, then
 
|x − 1| |x| + 1
arctan √ √ ≤ √ √ .
(k + x + 1) k + 1 + (k + 2) k + x k k + 1 + (k + 2) k − 1
By the Weierstrass M-test, the series P(x) converges uniformly, and therefore it is
continuous for x > −1. As in (a), we have
∞  
X 1 1
P(x) = arctan √ − arctan √ ,
k=1
k+1 k+x
so
1
P(x + 1) = P(x) + arctan √ .
1+x
Thus,
π
 
1 3π
lim P(x) = lim P(x + 1) − arctan √ = P(0) − =− .
x→−1+ x→−1+ 1+x 2 4

Editorial comment. The proposers report that they discovered the value −3π/4 ex-
perimentally. They ask whether there are more general closed forms for P, say at
half-integers.
Also solved by R. Bagby, N. Bagis (Greece) & M. L. Glasser, D. Beckwith, M. Benito, Ó. Ciaurri, E. Fernández
& L. Roncal (Spain), M. Chamberland, R. Chapman (U.K.), Y. Dumont (France), M. Goldenberg &
M. Kaplan, O. Kouba (Syria), G. Lamb, O. P. Lossers (Netherlands), R. Nandan, M. Omarjee (France),
A. H. Sabuwala, R. Stong, M. Tetiva (Romania), M. Vowe (Switzerland), GCHQ Problem Solving Group
(U.K.), Microsoft Research Problems Group, and the proposers.

A Vector Differential Equation

11440 [2009, 547]. Proposed by Stefano Siboni, University of Trento, Trento, Italy.
Consider the vector differential equation
 
x(t)
x (t) = p(t, x(t), x (t))x (t) ×
00 0 0
(1)
kx(t)k

180 c THE MATHEMATICAL ASSOCIATION OF AMERICA [Monthly 118

where x(t) = (x1 (t), x2 (t), x3 (t)), kuk denotes the usual Euclidean norm of a vector
u, × is the standard cross-product, and p and its first partial derivatives are real-valued
and continuous.
(a) Show that all solutions to (1) are defined on all of R.
(b) Show that any nonconstant solution tends to infinity as t → +∞.
x(t)
(c) Show that for any nonzero solution x(t), limt→+∞ kx(t)k exists.
Solution by Robin Chapman, University of Exeter, Exeter, U.K.
(a) Consider a nonconstant solution of (1) on an open interval I . From (1),
x(t) · x00 (t) = x0 (t) · x00 (t) = 0
on I . Therefore,
d 0
(x (t) · x0 (t)) = 2x0 (t) · x00 (t) = 0,
dt
which implies x0 (t) · x0 (t) = A, where A is a constant. Certainly A ≥ 0. If A = 0, then
x0 (t) = 0 on I , so x(t) would have to be constant. Hence A > 0. Next,
d
(x(t) · x0 (t)) = x0 (t) · x0 (t) + x(t) · x00 (t) = A.
dt
This implies that x(t) · x0 (t) = At + B for some constant B. Also, d
dt
(x(t) · x(t)) =
2x(t) · x0 (t) = 2At + 2B. This in turn implies that
x(t) · x(t) = At 2 + 2Bt + C,
where C is a constant.
By the Cauchy–Schwarz inequality,
(x(t) · x0 (t))2 ≤ (x(t) · x(t))(x0 (t) · x0 (t)).
Upon substituting the results above, this becomes
(At + B)2 ≤ A(At 2 + 2Bt + C).
Thus B 2 ≤ AC. If B 2 = AC, then x(t) and x0 (t) are linearly dependent, so x0 (t) ×
(x(t)/||x(t)||) = 0. Thus by (1), x00 (t) = 0, so x0 (t) is constant. In this case the solution
has the form x(t) = (t + k)u for a fixed k ∈ R and vector u; this extends to all of R.
Moreover, x(t)/||x(t)|| = u/||u|| for t 6 = k.
Suppose now that B 2 < AC. We then have
(At + B)2 + (AC − B 2 ) AC − B 2
x(t) · x(t) = ≥ > 0.
A A
Therefore, a problem with the differential equation in (1) being ill-defined when x(t) =
0 does not arise. From the theory of ordinary differential equations, a solution on an
open interval I with an endpoint a extends to a larger open interval J containing a
provided neither x(t) nor x0 (t) tends to infinity as t → a. Our formulas for x(t) · x(t)
and x0 (t) · x0 (t) prevent this. Hence x(t) extends to a solution on all of R.
(b), (c) Since x(t) · x(t) → ∞, it follows that x(t) → ∞ as t → ∞. Let y(t) =
0 (t)
x(t)/||x(t)||. Now dtd ||x(t)|| = x(t)·x
||x(t)||
. Hence,

x0 (t) (x(t) · x0 (t)) x(t) (x(t) · x(t)) x0 (t) − (x(t) · x0 (t))x(t)

y0 (t) = − = .
||x(t)|| ||x(t)||3 ||x(t)||3

February 2011] PROBLEMS AND SOLUTIONS 181

Consequently,
||x(t)||2 ||x0 (t)||2 − (x(t) · x0 (t))2
||y0 (t)||2 =
||x(t)||4
A(At 2 + 2Bt + C) − (At + B)2 AC − B 2
= = .
(At 2 + 2Bt + C)2 (At 2 + 2Bt + C)2
R∞
Hence ||y0 (t)|| = O(1/t 2 ) as t → ∞. The integral 0 y0 (s) ds then converges abso-
lutely, and so
Z t Z ∞
y(t) = y(0) + y0 (s) ds → y(0) + y0 (s) ds,
0 0

a finite limit as t → ∞.
Also solved by R. Bagby, W. J. Cowieson, H. Guggenheimer, O. Kouba (Syria), J. H. Lindsey II, O. P. Lossers
(Netherlands), J. Simons (U.K.), R. Stong, N. Thornber, and the proposer.

Triangles in a Subdivided Polygon

11441 [2009, 547]. Proposed by Y. N. Aliyev, Qafqaz University, Khyrdalan, Azerbai-
jan. Let n ≥ 4, let A0 , . . . , An−1 be the vertices of a convex polygon, and for each i
let Bi be a point in the interior of the segment Ai Ai+1 . (Here, and throughout, indices
of points are taken modulo n.) Let Ci denote the intersection of diagonals Bi−2 Bi and
Bi−1 Bi+1 . Let a( p, q, r ) denote the area of the triangle with vertices p, q, r . Show that
n−1 n−1
X 1 X 1
≥ .
i=0
a(Ai , Bi , Bi−1 ) i=0
a(Ci , Bi , Bi−1 )

Solution by Jim Simons, Cheltenham, U.K. Fix all the B j (and therefore all the C j ),
and all the A j except Ai and Ai+1 for some particular i, and consider varying the
line Ai Ai+1 through Bi . Let |Bi−1 Bi+1 | = l, α = ∠Bi Bi−1 Bi+1 ,β = ∠Bi Bi−1 Ai ,
γ = ∠Bi Bi+1 Bi−1 , δ = ∠Bi Bi+1 Ai+1 and θ = ∠Ai Bi Bi−1 , so that α + γ − θ =
∠Ai+1 Bi Bi+1 . Now
l sin γ l sin α
|Bi−1 Bi | = and |Bi Bi+1 | = .
sin(α + γ ) sin(α + γ )
If h i is the distance of Ai from the line Bi−1 Bi , then
|Bi−1 Bi | |Bi Bi+1 |
hi = and h i+1 = .
cot β + cot θ cot δ + cot(α + γ − θ )
Writing 1 for 1/a(Ai Bi Bi−1 ) + 1/a(Ai+1 Bi+1 Bi ), we conclude that
2 sin2 (α + γ ) 2 sin2 (α + γ )
1= (cot β θ) cot δ + cot(α + γ − θ ) .


+ cot +
l sin γ
2 2
l sin α
2 2

Differentiating with respect to θ here gives

2 sin2 (α + γ )
 
d1 −1 1
= + .
dθ l2 sin2 γ sin2 θ sin2 α sin2 (α + γ − θ )
Thus d1/dθ = 0 when sin γ sin θ = ± sin α sin(α + γ − θ ). Since the sines are all
positive, the only valid case is when sin γ sin θ = sin α sin(α + γ − θ ), and this gives a

182 c THE MATHEMATICAL ASSOCIATION OF AMERICA [Monthly 118

minimum of 1 since 1 → ∞ as θ → 0 and as θ → α + β. Therefore 1 is minimized
when sin γ sin θ = sin α sin(α + γ ) cos θ − sin θ cos(α + γ ) , or equivalently, when

sin α sin(α + γ ) sin α sin(α + γ )

tan θ = =
sin γ + sin α sin(α + γ ) sin γ + sin α cos α cos γ − sin2 α sin γ
sin α sin(α + γ )
= = tan α.
sin γ cos2 α + sin α cos α cos γ
This occurs when Ai Ai+1 is parallel to Bi−1 Bi+1 . Thus in a configuration that min-
1/a(Ai , Bi , Bi−1 ) for a given value of i=0 1/a(Ci , Bi , Bi−1 ), every
Pn−1 Pn−1
imizes i=0
Ai Ai+1 is parallel to the corresponding Bi−1 Bi+1 . In that case every Ai Bi−1 Ci Bi is
a parallelogram, so that every a(Ai , Bi , Bi−1 ) = a(Ci , Bi , Bi−1 ), and therefore
n−1 n−1
X 1 X 1
= .
i=0
a(Ai , Bi , Bi−1 ) i=0
a(Ci , Bi , Bi−1 )

Also solved by R. Chapman (U.K.), P. P. Dályay (Hungary), J. H. Lindsey II, Á. Plaza & S. Falcón (Spain), R.
Stong, GCHQ Problem Solving Group (U.K.), and the proposer.

That’s Sum Inequality

11442 [2009, 547]. Proposed by José Luis Dı́az-Barrero and José Gibergans-Báguena,
Universidad Politécnica de Cataluña, Barcelona, Spain. Let hak i be a sequence of
positive numbers defined by an = 12 (an−1
2
+ 1) for n > 1, with a1 = 3. Show that
" n ! n !#1/2  
X ak X 1 1 a1 + an
≤ √ .
k=1
1 + ak k=1 k
a (1 + ak ) 4 a1 an

Solution by Jim Simons. Cheltenham, U.K. This is an extraordinarily weak inequal-

ity! The left side exceeds 1 for the first time when n = 9, at which point the right
side exceeds 1025 . To see that it is true, we first note that ak > 2k . To prove this by
induction, note a1 = 3 > 2 and a2 = 5 > 4; beyond that, ak > 4 and ak+1 > 2ak .
k−1
(A stronger bound of ak ≥ 2(3/2)2 is also easy.) Since ak /(1 + ak ) < 1, we have
k=1 ak /(1P
+ ak ) < n. Since 1 + ak ≥ 4 and ak > 2k , we have 1/(ak (1 + ak )) <
P n

2 −k−2
and√ nk=1 1/ak (1 + ak ) < 1/4. Combining these, we see that the left side is
less than n/2. For n ≥ 8, the right side satisfies
 √ r
2n/2

1 a1 + an an n
√ > √ > √ > .
4 a1 an 4 3 4 3 2
Direct calculation shows that the inequality holds for smaller n, the closest call being
at n = 3.
Also solved by R. Chapman (U.K.), L. Csete (Hungary), P. P. Dályay (Hungary), J.-P. Grivaux (France), E. Hys-
nelaj (Australia) & E. Bojaxhiu (Germany), O. Kouba (Syria), J. H. Lindsey II, O. P. Lossers (Netherlands),
K. McInturff, P. Perfetti (Italy), C. R. & S. Selvaraj, R. Stong, M. Tetiva (Romania), Z. Vörös (Hungary),
GCHQ Problem Solving Group (U.K.), Microsoft Research Problems Group, and the proposer.

An Integral with Fractional Parts

11447 [2009, 647]. Proposed by Ovidiu Furdui, Campia Turzii, Cluj, Romania. Let a
be a positive number, and let g be a continuous, positive, increasing function on [0, 1].

February 2011] PROBLEMS AND SOLUTIONS 183

Prove that
Z 1 Z 1
n n oa 1
lim g(x) d x = g(x) d x,
n→∞ 0 x a+1 0

where a > 0 and {x} denotes the fractional part of x.

Solution by Ralph Howard, University of South Carolina, Columbia, SC. The result
holds in greater generality; we claim that:
If β : R → R be a bounded measurable function that is periodic with period 1, so that
β satisfies β(z + 1) = β(z), and if g ∈ L 1 ([0, 1]), then
Z 1   Z 1 Z 1
n
lim β g(x) d x = β(z) dz g(x) d x. (1)
n→∞ 0 x 0 0

Assuming (1) and taking β(z) = {z}a , which has period one and is bounded for a > 0,
we have
Z 1 n oa Z 1 Z 1 Z 1
n 1
lim g(x) d x = z a dz g(x) d x = g(x) d x.
n→∞ x=0 x 0 x=0 a + 1 x=0
For the proof of (1), we extend the Riemann–Lebesgue lemma:

Lemma. If f ∈ L 1 (R) and β : R → R is a bounded measurable function such that

β(z + 1) = β(z), then
Z ∞ Z 1 Z ∞
lim β(ny) f (y) dy = β(z) dz f (y) dy. (2)
n→∞ −∞ 0 −∞

The proof of this lemma proceeds just as in one of the standard proofs of the Riemann-
Lebesgue lemma: It is easy to check that it holds for f = χ[a,b] , the characteristic
function of an interval. By linearity, it then holds for finite linear combinations of
characteristic functions of intervals, that is, for step functions. However, step functions
are dense in L 1 (R), so the result holds for all f ∈ L 1 (R) by approximation.
To obtain (1) from the lemma, let g ∈ L 1 ([0, 1]), and define f : R → R by f (y) =
y g(1/y) for y ≥ 1 and f (y) = 0 otherwise. Letting y = 1/x in the change of vari-
−2

able formula yields

Z ∞ Z ∞ Z 1
f (y) dy = −2
y g(1/y) dy = g(x) d x.
−∞ 1 0

This equation holds as well with absolute value bars on the integrands, and therefore
f ∈ L 1 (R). The same change of variable yields
Z ∞ Z ∞ Z 1  
g(1/y) n
β(ny) f (y) dy = β(ny) 2
dy = β g(x) d x.
−∞ 1 y 0 x
The required result is now an application of the lemma.
Also solved by K. F. Andersen (Canada), M. R. Avidon, R. Bagby, D. Borwein (Canada), R. Chapman (U.K.),
W. J. Cowieson, M. Eyvasi (Iran), P. J. Fitzsimmons, J.-P. Grivaux (France), E. A. Herman, M. Kochanski
(U.K.), O. Kouba (Syria), J. H. Lindsey II, O. P. Lossers (Netherlands), V. S. Miller, M. Omarjee (France),
E. Omey (Belgium), P. Perfetti (Italy), Á. Plaza & S. Falcón (Spain), K. Schilling, J. Simons (U.K.), A. Stadler
(Switzerland), A. Stenger, R. Stong, J. V. Tejedor (Spain), M. Tetiva (Romania), E. I. Verriest, L. Zhou, GCHQ
Problem Solving Group (U.K.), NSA Problems Group, and the proposer.

184 c THE MATHEMATICAL ASSOCIATION OF AMERICA [Monthly 118

 Asymptotics of a Product
11456 [2009, 747]. Proposed by Raymond Mortini, Université Paul Verlaine, Metz,
France. Find
n  
Y 1 5
lim n 1− + .
n→∞
m=1
m 4m 2

Solution
√ by Oliver Geupel, Brühl, Germany. We use Stirling’s formula, which says that
n+1/2 −n
n! ≈ 2π n e , together with the infinite product
∞ 
4z 2
Y 
cosh z = 1+ .
m=1
(2m − 1)2 π 2

(Abramowitz & Stegun, Handbook of Mathematical Functions with Formulas, Graphs,

and Mathematical Tables, Dover 1972, Formula 4.5.69, p. 85).
We have
n  n
(2m − 1)2 + 4

Y 1 5 Y
lim n 1− + = lim n
n→∞
m=1
m 4m 2 n→∞
m=1
(2m)2
n n
Y (2m − 1)2 Y (2m − 1)2 + 4
= lim n ·
n→∞
m=1
(2m)2 m=1
(2m − 1)2
n 
n · (2n)!2 Y 4π 2

= lim · 1 +
n→∞ 16n · n!4
m=1
(2m − 1)2 π 2

n · 2π · (2n)4n+1 e−4n cosh(π )

= lim · cosh(π ) = .
n→∞ 16 · (2π ) · n
n 2 4n+2 ·e −4n π

Editorial comment. A generalization was provided by Jerry Minkus (San Francisco,

CA): Let b be a positive integer, and c a nonzero constant such that bm=1 (m 2 − bm +
Q
c) 6= 0. Letting 1 = b2 /4 − c, we have
n  
Y b c
lim n(n − 1)(n − 2) · · · (n − b + 1) 1− + 2
n→∞
m=1
m m
b
1Y 2 1
= (m − bm + c) · √ √ .
c m=1 0(b/2 − 1 ) 0(b/2 + 1 )
Several other solvers provided the generalization with b = 1 but general c.
Also solved by R. A. Agnew, K. F. Andersen (Canada), M. S. Ashbaugh (U.S.A.) & F. V. Prado (Chile),
R. Bagby, D. H. Bailey (U.S.A.) & J. M. Borwein (Canada), M. Bataille (France), K. A. Beres, P. Bracken,
B. S. Burdick, M. A. Carlton, R. Chapman (U.K.), H. Chen, P. P. Dályay (Hungary), O. Furdui (Romania),
M. Goldenberg & M. Kaplan, G. C. Greubel, J. Grivaux (France), W.-P. Heidorn (Germany), E. A. Herman,
F. Holland (Ireland), A. Ilić (Serbia), M. E. H. Ismail, T. Konstantopoulos (U.K.), O. Kouba (Syria), O. P.
Lossers (Netherlands), S. de Luxán & Á. Plaza (Spain), R. Martin (Germany), V. S. Miller, J. Minkus, B. Mu-
lansky (Germany), M. Muldoon (Canada), D. K. Nester, M. Omarjee (France), J. Posch, H. Riesel (Sweden),
O. G. Ruehr, B. Schmuland (Canada), J. Simons (U.K.), N. C. Singer, A. Stenger, R. Stong, T. Tam, R. Tauraso
(Italy), J. V. Tejedor (Spain), M. Tetiva (Romania), D. B. Tyler, M. Vowe (Switzerland), S. Wagon, B. Wal-
lace, T. Wiandt, P. Xi (China), S. Xiao, GCHQ Problem Solving Group (U.K.), NSA Problems Group, and the
proposer.

February 2011] PROBLEMS AND SOLUTIONS 185