The American Mathematical Monthly

ISSN: 0002-9890 (Print) 1930-0972 (Online) Journal homepage: http://www.tandfonline.com/loi/uamm20

Problems and Solutions

Gerald A. Edgar, Daniel H. Ullman & Douglas B. West

To cite this article: Gerald A. Edgar, Daniel H. Ullman & Douglas B. West (2018)
Problems and Solutions, The American Mathematical Monthly, 125:3, 276-283, DOI:
10.1080/00029890.2018.1424478

To link to this article: https://doi.org/10.1080/00029890.2018.1424478

Published online: 22 Feb 2018.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=uamm20
 PROBLEMS AND SOLUTIONS
Edited by Gerald A. Edgar, Daniel H. Ullman, Douglas B. West
with the collaboration of Paul Bracken, Ezra A. Brown, Zachary Franco, Christian Friesen,
László Lipták, Rick Luttmann, Frank B. Miles, Lenhard Ng, Kenneth Stolarsky, Richard
Stong, Walter Stromquist, Daniel Velleman, Stan Wagon, Elizabeth Wilmer, Paul Zeitz,
and Fuzhen Zhang.

Proposed problems should be submitted online at

americanmathematicalmonthly.submittable.com/submit
Proposed solutions to the problems below should be submitted by July 31, 2018 via
the same link. More detailed instructions are available online. Proposed problems
must not be under consideration concurrently at any other journal nor be posted to
the internet before the deadline date for solutions. An asterisk (*) after the number
of a problem or a part of a problem indicates that no solution is currently available.

PROBLEMS
12027. Proposed by Abdul Hannan, Chennai Mathematical Institute, Chennai, India. Let
ABC be a triangle with circumradius R and inradius r. Let D, E, and F be the points where
the incircle of ABC touches BC, CA, and AB, respectively, and let X, Y , and Z be the second
points of intersection between the incircle of ABC and AD, BE, and CF, respectively. Prove
|AX| |BY | |CZ| R 1
+ + = − .
|XD| |Y E| |ZF| r 2

12028. Proposed by Michael Elgersma, Minneapolis, MN, Ramin Naimi, Occidental Col-
lege, Los Angeles, CA, and Stan Wagon, Macalester College, St. Paul, MN. We have n coins,
where n = d + p + q for positive integers d, p, and q. Suppose that whenever any d of the
coins are removed, the rest can be split into sets of size p and q that balance when placed
on a balance with arm lengths q and p, respectively. That is, q times the weight of the p
coins equals p times the weight of the q coins. Must all n coins have the same weight?

12029. Proposed by Hideyuki Ohtsuka, Saitama, Japan. For a > 0, evaluate

n  
k
lim a+ .
n→∞
k=1
n

12030. Proposed by Jonathan Sondow, New York, NY. Let S be the set of positive integers
d such that, for some multiple m of d,
 
m+d
≡ 1 (mod m).
d

(a) Does S contain a prime number?

(b) Does S contain a number with distinct prime factors?
(c)* Does S contain a nontrivial prime power?

doi.org/10.1080/00029890.2018.1424478

276 
C THE MATHEMATICAL ASSOCIATION OF AMERICA [Monthly 125
12031. Proposed by Ovidiu Furdui, Technical University of Cluj-Napoca, Cluj-Napoca,
Romania.
(a) Prove
 1 1 
x
dx dy = 1 − γ ,
0 0 1 − xy
where {a} denotes the fractional part of a, and γ is Euler’s constant.
(b) Let k be a nonnegative integer. Prove
 1 1 k  1  k
x 1
dx dy = dx.
0 0 1 − xy 0 x

12032. Proposed by David Galante (student) and Ángel Plaza, University of Las Palmas
de Gran Canaria, Las Palmas, Spain. For a positive integer n, compute
 n  n   
k−p k n n−k
(−1) 2 .
p=0 k=p
2p k

12033. Proposed by Dao Thanh Oai, Thai Binh, Vietnam, and Leonard Giugiuc, Drobeta
Turnu Severin, Romania. Let ABCD be a convex quadrilateral with area S. Prove
AB2 + AC2 + AD2 + BC2 + BD2 + CD2 ≥ 8S + AB · CD + BC · AD − AC · BD.

SOLUTIONS
A Radical Bound
11906 [2016, 400]. Proposed by Robert Bosch, Archimedean Academy, FL. Let x, y, and z
be positive numbers such that xyz = 1. Prove


x+1 y+1 z+1 √
+ + ≤ 3 2.
x −x+1
2 y −y+1
2 z −z+1
2

Solution by Ramya Dutta, Chennai Mathematical Institute, Chennai, India. Since xyz = 1,
we can choose a, b, c ∈ R+ such that x = a/b, y = b/c, and z = c/a. The inequality then
becomes

 b(a + b) √
≤ 3 2,
cyc
a − ab + b
2 2


where cyc τ (a, b, c) denotes the cyclic sum τ (a, b, c) + τ (b, c, a) + τ (c, a, b). Since


 b(a + b)  b(a + b)  b
= ≤ 2
cyc
a − ab + b
2 2
cyc
1
4
(a + b) + 4 (a − b)
2 3 2
cyc
a+b

and since

 b √ b
= b+c·
cyc
a + b cyc
(a + b)(b + c)

March 2018] PROBLEMS AND SOLUTIONS 277

  1/2  1/2  1/2
  b √  b(a + b + c)
≤ (b + c) · = 2· ,
cyc cyc
(a + b)(b + c) cyc
(a + b)(b + c)

by the Cauchy–Schwarz inequality, it suffices to prove

 b(a + b + c) 9
≤ .
cyc
(a + b)(b + c) 4

This is equivalent to
 b(a + b + c)

3  ac 3
1− ≥ , or ≥ ,
cyc
(a + b)(b + c) 4 cyc
(a + b)(b + c) 4

which is in turn equivalent to


 3 3 
ac(a + c) ≥ (a + b)(b + c)(a + c) = 2abc + ac(a + c) .
cyc
4 4 cyc

Solving this for the cyclic sum, we obtain


ac(a + c) ≥ 6abc.
cyc

This follows from the AM–GM inequality applied to the six terms a2 b, ab2 , a2 c, ac2 , b2 c,
and bc2 . Therefore,

 b(a + b)  b √ 9 √
≤2 ≤2 2· = 3 2.
cyc
a − ab + b
2 2
cyc
a+b 4

Also solved by R. A. Agnew, T. Amdeberhan & V. H. Moll, R. Boukharfane (France), P. Bracken,

M. V. Channakeshava (India), H. Chen, P. P. Dályay (Hungary), M. Dincă (Romania), H. Evans, D. Fleischman,
S. Gayen (India), J.-P. Grivaux (France), N. Grivaux (France), O. Kouba (Syria), M. E. Kuczma (Poland),
K.-W. Lau (China), J. H. Lindsey II, S. Malikić (Canada), V. Mikayelyan (Armenia), M. Omarjee (France),
Á. Plaza (Spain), J. C. Smith, A. Stenger, R. Stong, R. Tauraso (Italy), T. Wiandt, M. R. Yegan (Iran),
Con Amore Problem Group (Denmark), FAU Problem Solving Group, GCHQ Problem Solving Group (U. K.),
GWstat Problem Solving Group, NSA Problems Group, and the proposer.

An Inequality Applied To Eigenvalues

11907 [2016, 400]. Proposed by Xiang-Qian Chang, Massachusetts College of Pharmacy
and Health Sciences, Boston, MA. Let A be an n × n positive-definite Hermitian matrix,
with minimum and maximum eigvenvalues λ and ω, respectively. Prove
   
1 Tr(A) ωn n 1 −1
+ ≤ det A + ωA ,
ω n Tr(A) ω
  n  
1 n Tr(A−1 ) 1
−1
+ λ ≤ det A + λA−1 .
λ Tr(A ) n λ
Solution by Richard Stong, Center for Communications Research, San Diego, CA. The func-
tion f (t ) = log(t + 1/t ) = log(t 2 + 1) − log t has

1 2(1 − t 2 )
f

(t ) = + ≥0
t2 (1 + t 2 )2

278 
C THE MATHEMATICAL ASSOCIATION OF AMERICA [Monthly 125
for 0 < t ≤ 1. Hence by Jensen’s inequality, for any x1 , . . . , xn ∈ (0, 1], we have
 n  n
x1 + · · · + xn n
+ ≤ (xk + xk−1 ).
n x1 + · · · + xn k=1

Let the eigenvalues of A (with multiplicities) be λ1 , . . . , λn > 0. Applying this inequality

to xk = λk /ω, we obtain the first inequality requested. Applying it to xk = λ/λk ≤ 1, we
obtain the second inequality.
Editorial comment. The problem statement contained a typographical error: The exponent
n was missing from the second inequality.
Also solved by H. Chen, P. P. Dályay (Hungary), R. Dutta (India), D. Fleischman, N. Grivaux (France),
E. A. Herman, O. Kouba (Syria), J. H. Lindsey II, O. P. Lossers (Netherlands), V. Mikayelyan (Armenia),
M. Omarjee (France) & R. Tauraso (Italy), J. C. Smith, FAU Problem Solving Group, GCHQ Problem Solving
Group (U. K.), and the proposer.

A Generalized Bijection for Partitions

11908 [2016, 504]. Proposed by George E. Andrews, The Pennsylvania State University,
University Park, PA, and Emeric Deutsch, Polytechnic Institute of New York University,
Brooklyn, NY. Let n and k be nonnegative integers. Show that the number of partitions of n
having k even parts is the same as the number of partitions of n in which the largest repeated
part is k (defined to be 0 if the parts are all distinct). For example, 7 has three partitions with
two even parts (4 + 2 + 1 = 3 + 2 + 2 = 2 + 2 + 1 + 1 + 1) and also three partitions in
which the largest repeated part is 2 (3 + 2 + 2 = 2 + 2 + 2 + 1 = 2 + 2 + 1 + 1 + 1).
Solution I by Meghana Madhyastha, International Institute of Information Technology,
Bangalore, India. Fixing k, we find the generating functions of the two quantities, indexed
by n. In a partition where k is the largest repeated part, each part smaller than k can appear
any number of times, k appears at least twice, and parts larger than k appear at most once.
Hence, the generating function is
k−1 ∞
 1 x2k 
(1 + xi ).
i=1
1 − x i 1 − x k
i=k+1

For a partition with exactly k even parts, consider the even and odd parts separately.
In the conjugate of the partition using the even parts, each part occurs an even number of
times, and the largest part is k (occurring at least twice). There is no restriction on the use
of odd parts. Hence, the generating function is
k−1 ∞
 1 x2k  1
.
i=1
1 − x 2i 1 − x 2k
i=1
1 − x2i−1

Straightforward manipulation shows that both generating functions equal

∞
 1 − x2(k+i)
x2k .
i=1
1 − xi

Solution II by Nicolas Allen Smoot, Georgia Southern University, Statesboro, GA. We prove
the following generalization: Given nonnegative integers n and k and a positive integer d,
the number of partitions of n having exactly k parts divisible by d is the same as the number
of partitions of n in which k is the largest part that occurs at least d times.
When n = 0, the claim is trivial, so assume n > 0. We construct a bijection. Let λ be a
partition of n having exactly k parts divisible by d. Let A consist of all the parts in λ that

March 2018] PROBLEMS AND SOLUTIONS 279

are divisible by d, and let B consist of the other parts (A is empty when k = 0). Map A to
its conjugate partition A∗ , in which the largest part is k and every part occurs a multiple of
d times.
We map B, which has no part divisible by d, to a partition B
, in which no part occurs at
least d times, bijectively. For this we use Glaisher’s bijection (J. W. L. Glaisher, A theorem
in partitions, Messenger of Math. 12 (1883) 158–170). This turns B into B
by iteratively
combining d equal parts into one part until no instance of d identical parts remains. The
proof that this is a bijection relies on the fact that every positive integer is expressible as a
power of d times a number not divisible by d in a unique way.
Note that in the union of A∗ and B
, the largest part occurring at least d times is k. To
invert the map, we separate a partition μ in which k is the largest part occurring at least d
times into the contributions A∗ and B
, where A∗ will have each part occurring a multiple
of d times (k being the largest part) and B
will have no part occurring at least d times.
For each part i, occurring mi times in μ, put d mi /d of the copies of i into A∗ . Put the
remaining copies into B
; no part occurs at least d times among these. This is the only way
that μ can be separated into two partitions in the specified families. We can now invert the
two maps separately and recombine the outcomes to retrieve the only partition λ that maps
to μ under the given function.
Editorial comment. Mingjia Yang also proved the generalization in Solution II.
Also solved by D. Beckwith, K. David, Y. J. Ionin, P. Lalonde (Canada), P. W. Lindstrom, G. Lord, O. P. Lossers
(Netherlands), R. Nandan, M. Sawhney, J. H. Smith, R. Stong, R. Tauraso (Italy), V. Walavalkar (India),
E. T. White, M. Wildon (U. K.), M. Yang, GCHQ Problem Solving Group (U. K.), and the proposers.

Reciprocal Fibonacci Arctangents

11910 [2016, 504]. Proposed by Cornel Ioan Vălean, Teremia Mare, Romania. Let Gk be
the reciprocal of the kth Fibonacci number, for example, G4 = 1/3 and G5 = 1/5. Find
∞

(arctan G4n−3 + arctan G4n−2 + arctan G4n−1 − arctan G4n ) .
n=1

Solution by Ramya
√ Dutta, Chennai Mathematical Institute, Chennai, India. The sum is
π /2 + arctan(( 5 − 1)/2). To see this, we write Fn for the nth Fibonacci number, and we
make use of Catalan’s identity Fn2 − Fn−1 Fn+1 = (−1)n+1 and d’Ocagne’s identity Fn Fn+1 +
Fn Fn−1 = F2n . Since
1 1 F2n+2 − F2n F2n+1 1
arctan − arctan = arctan = arctan 2 = arctan ,
F2n F2n+2 F2n F2n+2 + 1 F2n+1 F2n+1
we have
1 1 1 1
arctan + arctan = arctan − arctan
F4n−3 F4n−1 F4n−4 F4n−2
1 1 1 1
+ arctan − arctan = arctan − arctan . (1)
F4n−2 F4n F4n−4 F4n
Equation (1) holds for all positive integers n, including n = 1, provided that we interpret
arctan(1/0) to be π /2. We also have
Fn−1 Fn Fn+1 Fn−1 − Fn2
arctan − arctan = arctan
Fn Fn+1 Fn Fn−1 + Fn Fn+1
(−1)n 1
= arctan = (−1)n arctan . (2)
F2n F2n

280 
C THE MATHEMATICAL ASSOCIATION OF AMERICA [Monthly 125
Thus,
∞ 
 
1 1 1 1
arctan + arctan + arctan − arctan
n=1
F4n−3 F4n−2 F4n−1 F4n
∞ 
  ∞ 
 
1 1 1 1
= arctan − arctan + arctan − arctan by (1)
n=1
F4n−4 F4n n=1
F4n−2 F4n
∞ ∞  
π  1 π  Fn−1 Fn
= + (−1)n−1 arctan = − arctan − arctan by (2)
2 n=1
F2n 2 n=1
Fn Fn+1
π Fn π 1
= + lim arctan = + arctan ,
2 n→∞ Fn+1 2 ϕ
√
where ϕ = (1 + 5 )/2. This gives the claimed result.
Also solved by K. Adegoke (Nigeria) & Á. Plaza (Spain), B. Bradie, M. V. Channakeshava (India),
P. P. Dályay (Hungary), D. Fleischman, D. Fritze (Germany), M. Goldenberg & M. Kaplan, S. Hitotu-
matu (Japan), O. Kouba (Syria), M. E. Kuczma (Poland), P. Lalonde (Canada), O. P. Lossers (Netherlands),
R. Nandan, M. Omarjee (France), A. Rajkumar & F. Mawyer, M. Sawhney, A. Stenger, R. Stong, R. Tauraso
(Italy), D. Terr, D. B. Tyler, M. Wildon (U. K.), J. Zacharias, L. Zhou, GCHQ Problem Solving Group (U. K.),
NSA Problems Group, and the proposer.

A Symmetric Inequality for Real Triples

11911 [2016, 504]. Proposed by Marian Cucoanes, Focşani, Romania, and Leonard
Giugiuc, Drobeta-Turnu Severin, Romania. Let a, b, and c be positive real numbers such
that 1 + ab + bc + ca = a + b + c + 2abc. Prove a3 + b3 + c3 + 5abc ≥ 1, and determine
when equality holds.
Solution by Marcin E. Kuczma, University of Warsaw, Poland. The claim holds trivially
when max{a, b, c} ≥ 1. Hence we assume a, b, c ∈ (0, 1), which yields abc < 1. In terms
of the elementary symmetric polynomials A = a + b + c, B = ab + bc + ca, and C = abc,
the constraint says 1 + B = A + 2C. Let
X = a3 + b3 + c3 + 5abc − 1 = A3 − 3AB + 8C − 1 = A3 + (4 − 3A)B + (3 − 4A).
We must show that X is nonnegative.
The AM–GM inequality implies B/3 ≥ C2/3 . Combining this with C < 1 yields B ≥
3C > 3C > 2C, which implies A = 1 + B − 2C > 1. If 4 − 3A > 0, then B = A + 2C −
2/3

1 > A − 1 yields
X > A3 + (4 − 3A)(A − 1) + (3 − 4A) = (A − 1)3 > 0.
If 4 − 3A ≤ 0, then the Cauchy–Schwarz inequality yields A2 ≤ 3(a2 + b2 + c2 ) = 3(A2 −
2B), and thus B ≤ A2 /3. Therefore,
X ≥ A3 + (4 − 3A)(A2 /3) + (3 − 4A) = (2A − 3)2 /3 ≥ 0.
Equality requires 2A − 3 = 0 as well as equality in the Cauchy–Schwarz application; the
latter occurs when a = b = c. Thus equality holds if and only if a = b = c = 1/2.
Also solved by A. Alt, P. P. Dályay (Hungary), M. Dincă (Romania), D. Fleischman, N. Grivaux (France),
Y. Ionin, K.-W. Lau (China), J. H. Lindsey II, T. L. McCoy, R. Stong, T. Wiandt, GCHQ Problem Solving
Group (U. K.), NSA Problems Group, and the proposer.

Bounds for the Sum of the Mixtilinear Radii

11912 [2016, 505]. Proposed by Pál Péter Dályay, Szeged, Hungary. Let ω be the circum-
scribed circle of triangle ABC, and let R and r be the radii of its circumcircle and incircle,
March 2018] PROBLEMS AND SOLUTIONS 281
respectively. Let rA , rB , and rC be the radii of the A-, B-, and C-mixtilinear incircles of ABC
and ω, respectively. Prove 4r ≤ rA + rB + rC ≤ (5R + 6r)/4.
Solution by Tamas Wiandt, Rochester Institute of Technology, Rochester, NY.
Lower bound: Let the angles at A, B, and C be α, β, and γ , respectively. We have
rA = r sec2 (α/2), and similarly for rB and rC . (See, for example, L. Bankoff, A mixtilinear
adventure, Crux Mathematicorum 9 (1983) 2–7.) Now we have
 
α/2 + β/2 + γ /2 π 4
sec2 = sec2 = .
3 6 3
For x ∈ (0, π ), let f (x) = sec2 (x/2). Since f

(x) ≥ 0, we may apply Jensen’s inequality to

f and obtain
 
α/2 + β/2 + γ /2
4r = 3r sec2
3
sec (α/2) + sec2 (β/2) + sec2 (γ /2)
2
≤ 3r = rA + rB + rC .
3
Upper bound: Let a, b, c denote the side lengths opposite angles A, B, C, respectively.
Let x, y, z be the distances from A, B, C, respectively, to the points of tangency of the in-
circle. Since a = y + z, b = z + x, and c = z + y, the semiperimeter s of the triangle is
x + y + z. Now sec2 (α/2) = tan2 (α/2) + 1 = r2 /x2 + 1, and similarly for β and γ . So
rA = r(r2 /x2 + 1), rB = r(r2 /y2 + 1), and rC = r(r2 /z2 + 1), and the desired inequality be-
comes
 2 
2 r r2 r2 3r2 5Rr
r 2
+ 2
+ 2
+ 3r 2
≤ + .
x y z 2 4
The area T of the triangle is given by any of the three formulae T = rs, T = abc/4R, or
T 2 = s(s − a)(s − b)(s − c) = sxyz (Heron’s formula). From the first and the third, we ob-
tain r2 = xyz/s. From the first and second, we obtain Rr = abc/4s. Substituting these ex-
pressions into the desired inequality yields
 
xyz xyz xyz xyz 3xyz 5abc
2
+ 2 + 2 + ≤ .
s sx sy sz 2s 16s
Expressing a, b, c in terms of x, y, z and rearranging, we produce the equivalent inequality
3(2x2 y2 + 2y2 z2 + 2z2 x2 ) + 2(2x2 yz + 2y2 zx + 2z2 xy)
≤ 5(x3 y + y3 x + y3 z + z3 y + z3 x + x3 z). (∗)
We now recall Muirhead’s inequality, which asserts the following (in the case of three
variables). Let (a, b, c) and (p, q, r) be two nonnegative triples satisfying the conditions a ≥
b ≥ c, p ≥ q ≥ r, a ≥ p, a + b ≥ p + q, and 
a + b + c = p + q + r. For all nonnegative
real numbers x, y, and z, we have xa yb zc ≥ x p yq zr , where the sums are taken over all
3! = 6 permutations of the three variables x, y, z.
Applying Muirhead’s inequality with (a, b, c) = (3, 1, 0) and (p, q, r) = (2, 2, 0), we
get
x3 y + y3 x + y3 z + z3 y + z3 x + x3 z ≥ 2x2 y2 + 2y2 z2 + 2z2 x2 .
Applying Muirhead’s inequality with (a, b, c) = (3, 1, 0) and (p, q, r) = (2, 1, 1), we get
x3 y + y3 x + y3 z + z3 y + z3 x + x3 z ≥ 2x2 yz + 2y2 zx + 2z2 xy.
Together these give the required inequality (∗).

282 
C THE MATHEMATICAL ASSOCIATION OF AMERICA [Monthly 125
Also solved by O. Geupel (Germany), O. Kouba (Syria), M. E. Kuczma (Poland), R. Stong, J. Zacharias,
GCHQ Problem Solving Group (U. K.), and the proposer.

An Integral Inequality
11918 [2016, 613]. Proposed by Phu Cuong Le Van, College of Education, Hue University,
Hue City, Vietnam. Let f be n times continuously differentiable on [0, 1], with f (1/2) = 0
and f (i) (1/2) = 0 when i is even and at most n. Prove
 1 2  1
1
f (x) dx ≤ ( f (n) (x))2 dx.
0 (2n + 1)2 2n (n!)2
0

Solution by Patrick J. Fitzsimmons, University of California, San Diego, La Jolla, CA. Let
F be an antiderivative of f . Using Taylor’s theorem with remainder in integral form, we
expand F in powers of t − 1/2 to obtain
n−1 (k)    t (n)
f (1/2) 1 k+1 f (x)
F (t ) = F (1/2) + t− + (t − x)n dx
k=0
(k + 1)! 2 1/2 n!

for any t in [0, 1]. In particular, with t = 1,

 1 
n−1 (k)    1 (n)
f (1/2) 1 k+1 f (x)
f (x) dx = + (1 − x)n dx,
1/2 k=0
(k + 1)! 2 1/2 n!

and with t = 0,
 1/2 
n−1 (k)    1/2 (n)
f (1/2) 1 k+1 f (x)
f (x) dx = − − + (−x)n dx.
0 k=1
(k + 1)! 2 0 n!

When we add these, the terms for odd k cancel, while the terms for even k vanish by hy-
pothesis. It follows that
 1  1
f (x) dx = g(x) f (n) (x) dx,
0 0

where

(−x)n /n! when 0 ≤ x ≤ 1/2;
g(x) =
(1 − x)n /n! when 1/2 ≤ x ≤ 1.
Now the desired inequality follows from the Cauchy–Schwarz inequality, because
 1  1/2 2n  1
x (1 − x)2n 1
g(x)2 dx = dx + dx = .
0 0 (n!) 2
1/2 (n!) 2 (2n + 1)22n (n!)2

Also solved by U. Abel (Germany), K. F. Andersen (Germany), P. Bracken, R. Chapman (U. K.), H. Chen,
P. P. Dályay (Hungary), R. Dutta (India), N. Grivaux (France), A. Harnist (France), E. A. Herman, K. Koo
(China), O. Kouba (Syria), M. E. Kuczma (Poland), J. H. Lindsey II, O. P. Lossers (Netherlands), F. Marino
(Italy), V. Mikayelyan (Armenia), R. Nandan, M. Omarjee (France), Á. Plaza & F. Perdomo (Spain),
M. A. Prasad (India), M. Sawhney, A. Stadler (Switzerland), R. Stong, R. Tauraso (Italy), E. I. Verriest,
T. Wiandt, L. Zhou, GCHQ Problem Solving Group (U. K.), Missouri State University Problem Solving Group,
and the proposer.

March 2018] PROBLEMS AND SOLUTIONS 283