Crux Mathematicorum

VOLUME 42, NO. 5 May / Mai 2016

Editorial Board
Editor-in-Chief Kseniya Garaschuk University of the Fraser Valley
Editorial Assistant Amanda Malloch University of Victoria

Contest Corner Editor John McLoughlin University of New Brunswick

Olympiad Corner Editor Carmen Bruni University of Waterloo
Book Reviews Editor Robert Bilinski Collège Montmorency
Articles Editor Robert Dawson Saint Mary’s University

Problems Editors Edward Barbeau University of Toronto

Chris Fisher University of Regina
Edward Wang Wilfrid Laurier University
Dennis D. A. Epple Berlin, Germany
Magdalena Georgescu University of Toronto
Shaun Fallat University of Regina

Assistant Editors Chip Curtis Missouri Southern State University

Lino Demasi Ottawa, Ontario
Allen O’Hara University of Western Ontario

Guest Editors Alejandro Erickson Durham University

Joseph Horan University of Victoria
Kelly Paton University of British Columbia
Alessandro Ventullo University of Milan
Kyle MacDonald McMaster University

Editor-at-Large Bill Sands University of Calgary

Managing Editor Denise Charron Canadian Mathematical Society

Copyright c Canadian Mathematical Society, 2016

 IN THIS ISSUE / DANS CE NUMÉRO

195 Editorial Kseniya Garaschuk

196 The Contest Corner: No. 45 John McLoughlin
196 Problems: CC221–CC225
199 Solutions: CC171–CC175
202 The Olympiad Corner: No. 343 Carmen Bruni
202 Problems: OC281–OC285
204 Solutions: OC221–OC225
209 Wobbling Bicycle: Solution Luis Goddyn
211 Focus On . . . : No. 22 Michel Bataille
216 The Engel-Titu Inequality Eeshan Banerjee
220 Problems: 4141–4150
225 Solutions: 4041–4050
236 Solvers and proposers index

Crux Mathematicorum
Founding Editors / Rédacteurs-fondateurs: Léopold Sauvé & Frederick G.B. Maskell
Former Editors / Anciens Rédacteurs: G.W. Sands, R.E. Woodrow, Bruce L.R. Shawyer,
Shawn Godin

Crux Mathematicorum
with Mathematical Mayhem
Former Editors / Anciens Rédacteurs: Bruce L.R. Shawyer, James E. Totten, Václav Linek,
Shawn Godin
 195

EDITORIAL
Oral exams – yea or nay? They are a standard practice in many European coun-
tries. In North America, however, this format seems to be reserved for PhD
defences. One explanation might be big class sizes of undergraduate universities,
which make it infeasible to administer individual oral tests; another – potential
lack of consistency and accountability on the part of the examiner as they can
choose to save students with helpful comments or to sink them with killer ques-
tions. This latter impartiality was a point of contention in USSR for years with
circulating claims of discrimination against various groups. The biggest arguments
surrounded admissions of Moskovites versus non-Moscovites as well as Jewish ver-
sus non-Jewish students. An article on this topic that drew my attention was by
Vershik and Shen (The Mathematical Intelligencer (1994), 16: 4) as it includes
examples of killer questions used by examiners entrance exams to Mechanical-
Mathematical Faculty of the Moscow State University in the 1980s. Here are
some of them. Do you think you would have been able to answer these?
1. The faces of a triangular pyramid have the same area. Show that they are
congruent.
2. Draw a straight line that halves the area and the perimeter of a triangle.
3. Show that
1 1 4
2 ≤ 2 +1− 2
sin x x π
π
for 0 < x ≤ 2.

4. Let K be a point on the base AB of a trapezoid ABCD. Find a point M on

the base CD that maximizes the area of the quadrangle formed by the intersection
of triangles AM B and CDK.
5. Compare log3 4 · log3 6 · . . . · log3 80 and 2 log3 3 · log3 5 · . . . · log3 79.
6. Given k segments on the plane, give an upper bound on the number of triangles
all of whose sides belong to the given set of segments.
I think these make for excellent challenge problems. If I find a particularly elegant
solution to any of the above, I will consider using them in my courses.
Kseniya Garaschuk
P. S. My editors recognized some of these problems as appearing in Canadian
sources before. For example, Problem 1 was mentioned in the solution to Crux
problem 478, which was published as an open problem and later settled by Tomasz
Ciesla (Crux 38(2), p. 68–70). Also, Problem 2 is discussed here http://
math.fau.edu/Yiu/GJARCMG2016.pdf and here http://mathcentral.uregina.
ca/mp/previous2011/apr12sol.php. Finally, all the killer problems from Vershik
and Shen’s article have been analyzed by Ilan Vardi in his 1999 essay, which you can
find here http://www.lix.polytechnique.fr/Labo/Ilan.Vardi/mekh-mat.ps

Copyright c Canadian Mathematical Society, 2016

196/ THE CONTEST CORNER

THE CONTEST CORNER

No. 45
John McLoughlin
The problems featured in this section have appeared in, or have been inspired by, a math-
ematics contest question at either the high school or the undergraduate level. Readers
are invited to submit solutions, comments and generalizations to any problem. Please see
submission guidelines inside the back cover or online.
To facilitate their consideration, solutions should be received by February 1, 2017.
The editor thanks André Ladouceur, Ottawa, ON, for translations of the problems.

CC221. What is the smallest positive integer n such that if S is any set
containing n or more integers, then there must be three integers in S whose sum
is divisible by 3?

CC222. What is the value of x in the plane figure shown?

CC223. The letters a, b, c, d, e, f, g, h and i in the figure below represent the

numbers 1, 2, 3, 4, 5, 6, 7, 8 and 9 in a certain order. In each of the nine circles,
we sum the three numbers so that nine sums are obtained. Suppose that all nine
sums are equal. What is the value of a + d + g?

Crux Mathematicorum, Vol. 42(5), May 2016

 THE CONTEST CORNER /197

CC224. What is the smallest positive integer n such that 31 divides 5n + n?

CC225. The three sides of triangle ABC are extended as shown so that
BD = 12 AB, CE = 12 BC and AF = 21 CA. What is the ratio of the area of
triangle DEF to that of triangle ABC?

.................................................................

CC221. Quel est le plus petit entier strictement positif n pour lequel étant
donné n’importe quel ensemble S contenant au moins n entiers, S doit contenir
trois entiers dont la somme est divisible par 3?

CC222. Quelle est la valeur de x dans la figure plane suivante?

CC223. Les lettres a, b, c, d, e, f, g, h et i dans la figure suivante représentent

les nombres 1, 2, 3, 4, 5, 6, 7, 8 et 9 dans un ordre quelconque. Dans chacun des
neuf cercles on additionne les trois nombres contenus dans le cercle. On obtient
ainsi neuf sommes. Sachant que les neuf sommes sont égales, quelle est la valeur
de a + d + g?

Copyright c Canadian Mathematical Society, 2016

198/ THE CONTEST CORNER

CC224. Quel est le plus petit entier positif n pour lequel 5n + n est divisible
par 31?

CC225. Les côtés du triangle ABC sont prolongés comme dans la figure
suivante de manière que BD = 21 AB, CE = 12 BC et AF = 12 CA. Quel est le
rapport de l’aire du triangle DEF à celle du triangle ABC?

Crux Mathematicorum, Vol. 42(5), May 2016

 THE CONTEST CORNER /199

CONTEST CORNER
SOLUTIONS
Statements of the problems in this section originally appear in 2015: 41(5), p. 192–193.

CC171. The zeroes of the polynomial f (x) = x2 − ax + 2a are integers. What

is the sum of all the possible values of the number a?
Originally Problem 21 from the 2015 level 4 phase 2 of the Primavera contest in
Spain.
We received nine submissions of which eight were correct and complete. We present
two solutions.
Solution 1, by Henry Ricardo.
If r1 and r2 are the roots of f , we may write x2 − ax + 2a = (x − r1 )(x − r2 ) =
x2 − (r1 + r2 )x + r1 r2 , so that (1) r1 + r2 = a and (2) r1 r2 = 2a. Substituting (2)
in (1), we get r1 + r2 = (r1 r2 )/2, or r1 (2 − r2 ) + 2r2 = 0. Subtracting 4 from each
side of the last equation gives us r1 (2 − r2 ) + 2r2 − 4 = −4, or (2 − r1 )(2 − r2 ) = 4.
It is easy to verify that the only (unordered) pairs (r1 , r2 ) satisfying this equation
are (0, 0), (1, −2), (3, 6), and (4, 4), yielding a = 0, −1, 9, and 8, respectively, so
that the sum of the values of a is 16.

Solution 2, by Titu Zvonaru.

The discriminant of the quadratic equation x2 − ax + 2a = 0 is a2 − 8a. Since the
roots are integers, let k be an integer such that

a2 − 8a = k 2 ⇐⇒ (a − 4)2 = k 2 + 16 ⇐⇒ (a − 4 − k)(a − 4 + k) = 16.

As (a − 4 − k) + (a − 4 + k) = 2(a − 4), the numbers a − 4 − k and a − 4 + k have

the same parity. We have the possibilities:
(i) a − 4 − k = −8, a − 4 + k = −2 =⇒ a = −1. The equation x2 + x − 2 = 0
has the roots 1 and −2;
(ii) a − 4 − k = −2, a − 4 + k = −8 =⇒ a = −1;
(iii) a − 4 − k = −4, a − 4 + k = −4 =⇒ a = 0. The equation x2 = 0 has the
root 0;
(iv) a − 4 − k = 2, a − 4 + k = 8 =⇒ a = 9. The equation x2 − 9x + 18 = 0 has
the roots 3 and 6;
(v) a − 4 − k = 8, a − 4 + k = 2 =⇒ a = 9;
(vi) a − 4 − k = 4, a − 4 + k = 4 =⇒ a = 8. The equation x2 − 8x + 16 = 0 has
the root 0.

Copyright c Canadian Mathematical Society, 2016

200/ THE CONTEST CORNER

Hence, the sum of all the possible values of the number a is −1 + 0 + 9 + 8 = 16.

CC172. What is the area of regular hexagon ABCDEF with A(0, 0) and
C(7, 1)?
Originally Problem 11 of the 2014 level 4 phase 2 of the Primavera contest in
Spain.
There were ten correct solutions for this problem and one incorrect solution. We
present the solution by Lucia Ma Li and Angel Plaza.
√ √
Let d be the distance between A and C. Then d = 72 + 12 = 50. Notice that d
is the side length of the equilateral triangle ACE, which by Heron’s Formula has
2√
√
area d4 3 = 252 3 .
Let O be the centre of the hexagon. Then OD bisects CE and so triangles ODE
and CDE have equal areas. Repeating this cyclically, we see that
√ the area of ACE
is half the area of ABCDEF, so the area of ABCDEF is 25 3.

CC173. In the following figure, an isosceles triangle with AB = 12 is divided

into 4 polygons of equal area using segments perpendicular to AB. Find x.

Originally Problem 15 of the 2014 level 4 phase 2 of the Primavera contest in

Spain.
We received eleven correct solutions. We present the solution of Kathleen Lewis.
First, let us label our triangle as follows:
P

A x C D B

Triangle AQC is similar to triangle AP D with sides in the ratio x : 6 and therefore
areas in the ratio x2 : 36. But AQC has the same area as the trapezoid P QCD,√so
2
it has half the area of triangle AP D, which says that x36 = 12 . Therefore x = 3 2.

Crux Mathematicorum, Vol. 42(5), May 2016

 THE CONTEST CORNER /201

√ √
CC174. Evaluate 1111 − 22 and 111111 − 222. Conjecture the result for
√
111111111111111111111111 − 222222222222 and prove it.
Originally Problem 11 of the Swiss “Mathématiques sans frontières” Épreuve de
découverte 2015-16.
There were nine correct solutions for this problem. We present the solution by
Titu Zvonaru.
√ √
1111 − 22 = 33, 111111 − 222 = 333.
. . . 1} = (10n − 1)/9.
Let A = |11 {z
n

. . . 2} = A · 10n + A − 2A = A(10n − 1) = 9A2 .

. . . 1} − |22 {z
We have |11 {z
2n n
…
. . . 1} − |22 {z
The result is that |11 {z . . . 2} = |33 {z
. . . 3} .
2n n n

CC175. Twenty-two mathematics contests were held with five prizes given
out for each one. The organizers notice that for each pair of contests, there is
exactly one participant who has won a prize in both contests. Show that one of
the participants has won a prize in each of the contests.
Originally Problem 4 of the Swiss preliminary contest for the Swiss Math Olympiad
2016.
We received no solutions to this problem.

Copyright c Canadian Mathematical Society, 2016

202/ THE OLYMPIAD CORNER

THE OLYMPIAD CORNER

No. 343
Carmen Bruni
The problems featured in this section have appeared in a regional or national mathematical
Olympiad. Readers are invited to submit solutions, comments and generalizations to any
problem. Please see submission guidelines inside the back cover or online.
To facilitate their consideration, solutions should be received by February 1, 2017.
The editor thanks André Ladouceur, Ottawa, ON, for translations of the problems.

OC281. Find all polynomials P (x) with real coefficients such that

P (P (x)) = (x2 + x + 1) · P (x)

where x ∈ R.

OC282. Let x, y, z be three nonzero real numbers satisfying x + y + z = xyz.

Prove that
X x2 − 1
( )2 ≥ 4.
x

OC283. In isosceles 4ABC, AB = AC, I is its incenter, D is a point inside

4ABC such that I, B, C, D are concyclic. The line through C parallel to BD
meets AD at E. Prove that CD2 = BD · CE.

OC284. A positive integer n is given. If there exist sets F1 , F2 , . . . , Fm satis-

fying the following conditions, prove that m ≤ n.
1. For all 1 ≤ i ≤ m, Fi ⊆ {1, 2, · · · , n}
2. For all 1 ≤ i < j ≤ m, min(|Fi − Fj |, |Fj − Fi |) = 1

OC285. Show that from a set of 11 square integers one can select six numbers
a2 , b2 , c2 , d2 , e2 , f 2 such that a2 + b2 + c2 ≡ d2 + e2 + f 2 (mod 12).

.................................................................

OC281. Déterminer tous les polynômes P (x) à coefficients réels tels que

P (P (x)) = (x2 + x + 1) · P (x)

pour tous réels x.

Crux Mathematicorum, Vol. 42(5), May 2016

 THE OLYMPIAD CORNER /203

OC282. Soit x, y, z trois réels non nuls tels que x + y + z = xyz. Démontrer
que
X x2 − 1
( )2 ≥ 4.
x

OC283. Soit un triangle isocèle ABC où AB = AC et soit I le centre du cercle

inscrit dans le triangle. Soit D un point à l’intérieur du triangle tel que I, B, C
et D soient cocycliques. La droite qui passe au point C et qui est parallèle à BD
coupe AD en E. Démontrer que CD2 = BD · CE.

OC284. Soit n un entier strictement positif. Sachant qu’il existe des ensembles
F1 , F2 , · · · , Fm qui satisfont aux deux conditions suivantes, démontrer que m ≤ n.
1. Pour tous i (1 ≤ i ≤ m), on a Fi ⊆ {1, 2, · · · , n}
2. Pour tous i et j (1 ≤ i < j ≤ m), on a min(|Fi − Fj |, |Fj − Fi |) = 1

OC285.
Étant donné un ensemble des carrés de 11 entiers, démontrer qu’il est possible de
choisir six de ces carrés, a2 , b2 , c2 , d2 , e2 et f 2 , tels que

a2 + b2 + c2 ≡ d2 + e2 + f 2 (mod 12) .

Copyright c Canadian Mathematical Society, 2016

204/ THE OLYMPIAD CORNER

OLYMPIAD SOLUTIONS
Statements of the problems in this section originally appear in 2015: 41(3), p. 101–102.

OC221. From the point P outside a circle ω with center O draw the tangents
P A and P B where A and B belong to ω. In a random point M in the chord
AB we draw the perpendicular to OM , which intersects P A and P B in C and D.
Prove that M is the midpoint of CD.

Originally problem 3 of the 2014 Balkan Mathematical Olympiad Team Selection

Test.

We received 10 correct submissions, consisting of a variety of solutions from many

new readers which is fantastic!

We present the solution by Somasundaram Muralidharan, who actually gave 4

different solutions to this problem. The editor chose the shortest of the solutions
to present.

Since ∠OAC = ∠OM C = 90◦ , O, A, C, M are concylic. Similarly, since ∠OM D =

∠OBD = 90◦ , the points O, M, B, D are concylic. Now,

∠M OC = ∠M AC
= ∠M BP (P A, P B are tangents from an external point)
= ∠M OD

Since OM is perpendicular to CD, it follows that M is the midpoint of CD.

Crux Mathematicorum, Vol. 42(5), May 2016

 THE OLYMPIAD CORNER /205

OC222. Let a, b be natural numbers with ab > 2. Suppose that the sum of
their greatest common divisor and least common multiple is divisible by a + b.
Prove that the quotient is at most a+b
4 . When is this quotient exactly equal to
a+b
4 ?

Originally problem 3 of the 2014 India National Olympiad.

We present the solution by Šefket Arslanagić. There were no other submissions.
First, if a = b then lcm(a, b) = gcd(a, b) = a and thus the given condition is

lcm(a, b) + gcd(a, b) a+a a+b a

= =1≤ =
a+b a+a 4 2
which holds whenever a ≥ 2 which must hold since ab > 2. Equality holds here
when a = b = 2. Now, suppose without loss of generality that a < b. Now, suppose
that gcd(a, b) = d and write a = a1 d and b = b1 d where a1 and b1 are coprime.
Then lcm(a, b) = a1 b1 d and thus,

lcm(a, b) + gcd(a, b) a1 b1 d + d a1 b1 + 1
= =
a+b a1 d + b1 d a1 + b1

If b1 = a1 + 1, then the above becomes

a1 b1 + 1 a2 + a1 + 1
= 1
a1 + b1 2a1 + 1

which is a natural number. Hence, this value times 2 must also be a natural
number. However,
2a21 + 2a1 + 2 a1 + 2
= a1 +
2a1 + 1 2a1 + 1
and thus, a1 + 2 ≥ 2a1 + 1. This implies that a1 = 1 and so a = d and b = 2d and
d > 1 since ab > 2. Hence, in this case,

lcm(a, b) + gcd(a, b) d + 2d a+b 3d

= =1≤ =
a+b d + 2d 4 4
holding since d > 1. Now, suppose that b1 ≥ a1 + 2. Then 2 ≤ b1 − a1 and so
2
4 ≤ b21 − 2a1 b1 + a21 . Rearranging shows that a1 b1 + 1 ≤ (a1 +b
4
1)
and hence

a1 b1 + 1 a1 + b1
≤ .
a 1 + b1 4

Hence, the given inequality holds. Equality holds in the cases

(a, b) ∈ {(2, 2), (a1 d, (a1 + 2)d), ((a1 + 2)d, a1 d)}

where d is a natural number and a1 is an arbitrary odd number (If it were even,
then a1 and a1 + 2 are not coprime and so we could factor out another 2).

Copyright c Canadian Mathematical Society, 2016

206/ THE OLYMPIAD CORNER

OC223. Let Z be the set of integers. Find all functions f : Z → Z such that

f (x)2
xf (2f (y) − x) + y 2 f (2x − f (y)) = + f (yf (y))
x
for all x, y ∈ Z with x 6= 0.
Originally problem 3 from day 1 of the 2014 USAJMO.
We received 2 correct submissions. We present the solution by Oliver Geupel.
It is straightforward to check that the two functions

f : x 7→ 0 and f : x 7→ x2

are solutions. We prove that there are no other ones.

Suppose that f is any solution. For integers x and y, let P (x, y) denote the
assertion that x and y satisfy the proposed functional equation.
For every x 6= 0, the number x divides f (x)2 by P (x, y). Assume towards a
6 0. Then from P (2f (0), 0), we see that
contradiction that f (0) =

f (2f (0))2
2f (0)f (2f (0) − 2f (0)) + 02 f (4f (0) − f (0)) = + f (0f (0))
2f (0)

which simplifies to
Å ã2
f (2f (0))
4f (0) − 2 = .
f (0)
This is a contradiction since the left hand side is divisible by exactly one copy
of 2 whereas the right hand side must be divisible by 4. Hence f (0) = 0. From
f (x)2
the assertions P (x, 0) and P (−x, 0) for x 6= 0, we obtain f (−x) = and
x2
2
f (−x)
f (x) = . Therefore, f (x)4 = x6 f (x), so that f (x) is either 0 or x2 . Also,
x2
f (x) = f (−x). Let us assume that a and b are non-zero integers such that f (a) = 0
and f (b) = b2 . All that remains to be done is to show that this is impossible.
f (x)2
By P (x, a) for x 6= 0, we have xf (−x) + a2 f (2x) = . Thus, f (2x) = 0.
x
Hence b is odd. For every integer x 6= 0, we obtain b f (4x − b2 ) = f (b3 ) applying
2

f (b3 )
P (2x, b); whence f (4x − b2 ) = . Since this holds for all nonzero x, we deduce
b2
that
f (4x − b2 ) = 0.
b2 + b
If b ≡ −1 (mod 4) where b 6= −1, then putting x = leads to
4

0 = f (4x − b2 ) = f (b) = b2 6= 0, (1)

Crux Mathematicorum, Vol. 42(5), May 2016

 THE OLYMPIAD CORNER /207

2
a contradiction. If b ≡ 1 (mod 4), b 6= 1, then choosing x = b 4−b gives (1), which
is impossible. As a consequence, f (−1) = f (1) = 1 and f (x) = 0 for x 6= ±1. By
P (2, 1), we have 0 = 2f (0) + f (3) = f (1) = 1, which is the desired contradiction.

OC224. Let n > 1 be an integer. An n × n-square is divided into n2 unit

squares. Of these unit squares, n are coloured green and n are coloured blue, and
all remaining ones are coloured white. Are there more such colourings for which
there is exactly one green square in each row and exactly one blue square in each
column; or colourings for which there is exactly one green square and exactly one
blue square in each row?

Originally problem 5 of the 2014 South Africa National Olympiad.

We received 2 correct submissions. We present the solution by Kathleen Lewis.

There are more colourings with one green and one blue in each row. To see this,
think of first placing one green square in each row; for both methods there are nn
ways to do that. If we want to place a blue square in each row, there would be
(n − 1)n to accomplish this, since each row has one square already coloured green.
But if we wish to put a blue square in each column, the number of possibilities
depends on the arrangement already made of the green squares. Suppose that there
are ai blank squares
Qn in column i. Then the number of possible arrangements of the
2
blue squares is i=1 ai . The total number of available squares is n −n = n(n−1),
so ni=1 ai = n(n − 1). But for variables with a fixed sum, the
P
Q product is greatest
when all the factors are equal. So, the maximum value of ni=1 ai occurs when
a1 = a2 = · · · = an = n − 1 and ni=1 ai = (n − 1)n . In other cases, the product
Q
would be smaller, even as small as zero if the green squares were all placed in the
same column. So the number of ways of placing a blue square in each column is
always less than or equal to the number of ways to place the blue squares with
one in each row.

OC225. Find the maximum value of real number k such that

a b c 1
+ + ≥
1 + 9bc + k(b − c)2 1 + 9ca + k(c − a)2 1 + 9ab + k(a − b)2 2

holds for all non-negative real numbers a, b, c satisfying a + b + c = 1.

Originally problem 5 of the 2014 Japan Mathematical Olympiad.

We received 3 correct submissions. We present the solution by Arkady Alt.

Let k be such that the original inequality holds for any non-negative real numbers
a, b, c satisfying a + b + c = 1. Then, in particular, if a = 0 and b = c = 1/2, we
get

1/2 1/2 1 4 1
2
+ 2
≥ ⇐⇒ ≥ ⇐⇒ k ≤ 4.
1 + k(1/2) 1 + k(1/2) 2 k+4 2

Copyright c Canadian Mathematical Society, 2016

208/ THE OLYMPIAD CORNER

Let k ≤ 4. By Cauchy’s Inequality

X a X a2
2
=
cyc
1 + 9bc + k(b − c) cyc
a (1 + 9bc + k(b − c)2 )
2
(a + b + c)
≥ P
a (1 + 9bc + k(b − c)2 )
cyc
1
= P
a (1 + 9bc + k(b − c)2 )
cyc
1
=
1 + 9abc (3 − k) + k(ab + bc + ca)
1
= ,
1 + 9q (3 − k) + kp

where p := ab + bc + ca and q := abc. We have

2
(a + b + c)
p = ab + bc + ca ≤ = 1/3,
3
9q = 9abc ≤ (ab + bc + ca) (a + b + c) = p,
9q ≥ 4p − 1.

a (a − b) (a − c) ≥ 0 in p, q notation with normalization by

P
(Schur’s Inequality
cyc
a + b + c = 1).
If k ≤ 3, then
1
9q (3 − k) + kp ≤ p (3 − k) + kp = 3p ≤ 3 · = 1.
3
If 3 < k ≤ 4, then
1
9q (3 − k)+kp ≤ (4p − 1) (3 − k)+kp = k+3p (4 − k)−3 ≤ k+3· (4 − k)−3 = 1.
3
Thus,
X a 1 1 1
2
≥ ≥ =
cyc
1 + 9bc + k(b − c) 1 + 9q (3 − k) + kp 1+1 2

for any k ≤ 4 and, therefore, the maximum value of k is 4.

Crux Mathematicorum, Vol. 42(5), May 2016

 WOBBLING BICYCLE: SOLUTION /209

Wobbling Bicycle: Solution

Proposed by Luis Goddyn, Simon Fraser University, Burnaby, BC. The problem
originally appeared in Crux 42(1), p. 10.

Problem. A wobbling bicycle passes through a mud patch. One of its wheels
traces a part of the curve y = sin x. The other wheel makes a curve with a vertical
inflection point. How long is the bicycle?

In order to eliminate effects due to bicycle geometry, tilting and wheel size, assume
the bicycle has vanishingly thin tires with its front axle always positioned directly
below a vertical headset. Assume also that both wheels were in the mud patch
when the inflection point is traversed. Determine the distance ` between its axles.

Solution. One readily sees that the rear wheel traced the sine curve. Suppose
the bicycle traveled from left to right on the coordinate plane, and that the front
wheel is at (X, Y ) when the rear wheel is at (x, sin x), for x ∈ [0, 2π]. Then
(X, Y ) = (x+`x , sin x+`y ), where (`x , `y ) is a translation vector of length `. Since
the rear wheel of a bicycle is aligned with its frame, the line through (x, sin x) and
(X, Y ) is tangent to the curve y = sin x. This gives two equations,

`y d
`2 = `2x + `2y , = sin x = cos x.
`x dx
√ √
Noting that `x > 0, these imply `x = `/ 1 + cos2 x and X = x + `/( 1 + cos2 x ).
At the vertical inflection, the derivative function

dX ` cos x sin x
=1+
dx (1 + cos2 x)3/2

attains a local minimum value of 0. Since 1 + cos2 x is positive, the function

(1 + cos2 x)3 − (` cos x sin x)2

also attains a local minimum value of 0 at the vertical inflection. We change the
variable to C = cos2 x (so sin2 x = 1 − C). At the local minimum, C is both a

Copyright c Canadian Mathematical Society, 2016

210/ WOBBLING BICYCLE: SOLUTION

zero and a critical point of the resulting function f (C).

f (C) = (1 + C)3 − `2 C(1 − C) = 0 (2)

0 2 2
f (C) = 3(1 + C) − ` (1 − 2C) = 0. (3)

To solve for `2 , we subtract 1 + C times equation (3) from 3 times equation (2) to
get
`2 (1 − 4C + C 2 ) = 0.
Since ` > 0, the second factor equals zero. We “complete the square” in two ways,

(1 + C)2 = (1 − 4C + C 2 ) + 6C = 6C,
(1 − 2C)2 = (1 − 4C + C 2 ) + 3C 2 = 3C 2 .

Using these with (3), we find

3(1 + C)2 3 · 6C √ √
`2 = = √ = ±6 3 = ± 108.
1 − 2C ± 3C 2

√
Thus the bicycle’s “length” is ` = 4 108 ≈ 3.223709. 
Note: With a bit more work, one finds that
√ 1 Ä√ √ ä2
C =2− 3= 6 − 2 ≈ 0.268,
4
and that at the vertical inflection we have the following vallues.
Å » √ ã
√ √ »√
2− 6
x = arccos − 2 − 3 = arccos = π − arcsin 3 − 1 ≈ 2.115,
2
»√
y= 3 − 1 ≈ 0.856,
» √
lx = 3 3 + 3 ≈ 2.863,
» √
ly = − 3 3 − 3 ≈ −1.482,
√ √ » √
2− 6
X = arccos + 3 3 + 3 ≈ 4.978,
»√ 2»
√ Ä√ ä3/2
Y = 3−1− 3 3−3=− 3−1 ≈ −0.626.

It is notable that, at the vertical inflection, the vertical positions of the tires satisfy

Y + y 3 = 0.

Crux Mathematicorum, Vol. 42(5), May 2016

 MICHEL BATAILLE /211

FOCUS ON...
No. 22
Michel Bataille
Constructions on the Sides
Introduction
On the sides of an arbitrary triangle, equilateral triangles are erected externally.
Then the centres of these equilateral triangles are the vertices of a new equilateral
triangle. This result, often called Napoleon’s theorem, if well-known, likely comes
as a surprise to students seeing it for the first time! We will not give a proof here
(see [1] or [2] if necessary), but present a selection of examples in the same vein.
We limit ourselves to configurations involving triangles and/or quadrilaterals and
favour proofs using transformations or complex numbers.

Napoleon’s configuration, slightly modified

We begin with an exercise adapted from problem 2815 [2003 : 111; 2004 : 113]:
Let ABC be a triangle with ∠ACB 6= 60◦ and D, E, F such that
BCD, CAE, ABF are equilateral triangles with D, E external to ∆ABC
and F on the same side of AB as C. Prove that DCEF is a parallel-
ogram.
Without loss of generality, we suppose that ABC is anti-clockwise oriented. Let
rM denote the anti-clockwise rotation with centre M and angle 60◦ . With this
notation, we may write A = rC (E) = rB (F ) and C = rB (D), from which we
deduce
D = r−1 −1
B (C) = rB ◦ rC (C)

and
F = r−1 −1
B (A) = rB ◦ rC (E).

−−→ −−→
Since r−1
B ◦ rC is a translation, we must have DF = CE and the result follows.

Copyright c Canadian Mathematical Society, 2016

212/ FOCUS ON... CONSTRUCTIONS ON THE SIDES

Another long time known problem

The following problem, which generalizes problem 2658 [2001 : 337 ; 2002 : 347],
seems to date back to the eighteenth century:

Squares ACDE and CBGF are drawn externally to an arbitrary tri-

angle ABC. Suppose that AG and BE intersect at M . Show that M
lies on the altitude CN .

The leading idea is to bring out a triangle whose altitudes are the lines CN, AG, BE
(triangle AKB on the figure).

Let U1 and U2 denote the centres of the squares ACDE and CBGF , respectively
and let r1 , r2 be the rotations with respective centres U1 , U2 such that r1 (C) = D
and r2 (F ) = C. Then r1 ◦ r2 (F ) = D and r1 ◦ r2 is a rotation with angle 180◦ so
that r1 ◦ r2 = h, the half-turn around the midpoint I of DF . Let K = h(C). In
addition to r1 (A) = C, we have r1 (B) = h ◦ r2 −1 (B) = K, hence CK = AB and
CK ⊥ AB. It follows that the lines KN and CN coincide and KN is an altitude
of ∆AKB.
−−→ −−→
Now, we introduce A0 such that BA0 = KC and the rotation r with centre B
such that r(G) = C. From r(A) = A0 , we deduce AG ⊥ CA0 . However, CKBA0
being a parallelogram, BK is parallel to CA0 and so AG is perpendicular to BK
as well. Thus, AG is an altitude of ∆AKB. Similarly, BE is the third altitude
and therefore intersects AG on the line KN = CN .

Crux Mathematicorum, Vol. 42(5), May 2016

 MICHEL BATAILLE /213

An example with triangles only

In the following example, unlike in Napoleon’s theorem, quite irregular triangles
are constructed on the sides of an arbitrary triangle. However, the angles of the
new triangles are so carefully chosen that some nice results hold. This interesting
problem, proposed in The Mathematical Gazette in November 2014, is slightly
reformulated.
In the plane of a triangle ABC, ∆BCD is described anti-clockwise
with ∠CBD = 63◦ , ∠BCD = 31◦ , ∆CAE is described clockwise with
∠ACE = 60◦ , ∠CAE = 63◦ and ∆ABF is described clockwise with
∠BAF = 31◦ , ∠ABF = 120◦ . Prove that D, E, F are collinear and
that the ratio DE : EF is independent of the triangle ABC.
We work with complex numbers, the small letter k denoting the affix of the point
K. The notation ∠(·, ·) is used for directed angles. The solution rests upon the
following interesting, general result: Let M, N, P, Q be points of the circle with
centre O and radius 1, with M 6= N . If ∠(M N, M P ) = α and ∠(N M, N Q) = β,
the affix of R, the point of intersection of the lines M P and N Q, is given by

neiα sin β − meiβ sin α

r= . (1)
sin(β − α)

−−→ −−→
The proof is straightforward: Since ∠(ON , OP ) = 2α, the affix of P is ne2iα .
Thus, the equation of M P is z + mne2iα z = m + ne2iα and similarly, the equation
of N Q is z + mne2iβ z = n + me2iβ . The affix r of R is the solution of the system
formed by the two equations: we are first led to

r(e2iβ − e2iα ) = me2iβ (1 − e2iα ) + ne2iα (e2iβ − 1)

Copyright c Canadian Mathematical Society, 2016

214/ FOCUS ON... CONSTRUCTIONS ON THE SIDES

and then to (1) after multiplication by e−i(α+β) .

Back to the problem, without loss of generality we suppose that |a| = |b| = |c| = 1.
We set σ = 63◦ , τ = 31◦ and ω = e2πi/3 . Applying the result above three times
yields
√ √
ceiσ sin τ + be−iτ sin σ c( 3/2)eiσ − aω sin σ aω sin τ + b( 3/2)e−iτ
d= , e= , f= .
sin(σ + τ ) sin(σ + 60◦ ) sin(60◦ − τ )

We calculate
sin σ
e−d = ·(ceiσ sin(60◦ −τ )−aω sin(σ+τ )−be−iτ sin(σ+60◦ ))
sin(σ + 60◦ ) sin(σ + τ )
√
(using 2
3
sin(σ + τ ) − sin τ sin(σ + 60◦ ) = sin σ sin(60◦ − τ )) and
√
3/2
f −e = ·(aω sin(σ+τ )+be−iτ sin(σ+60◦ )−ceiσ sin(60◦ −τ ))
sin(60◦ − τ ) sin(σ + 60◦ )
◦
σ sin(60 −τ )
from which we deduce that e−f e−d
= λ where λ = (sin
√
3/2) sin(σ+τ )
.
The result follows since λ is a positive real number independent of ∆ABC.

A general result
Another recent problem of The Mathematical Gazette proposed to construct on
the sides of a parallelogram ABCD equilateral triangles ABP, BCQ, CDR, DAS
with P, Q, R, S external to ABCD and to show that P QRS is a parallelogram.
The problem can be generalized as follows:
Let ABCD be an arbitrary quadrilateral and let α be a nonzero com-
plex number with modulus ρ and argument θ. Denote by SM the
spiral similarity with centre M , factor ρ and angle θ and let SA (B) =
P, SB (C) = Q, SC (D) = R, SD (A) = S. Show that P QRS is a paral-
lelogram if and only if α = 12 or ABCD is a parallelogram.
(This problem was set in a French high school final exam in the 1980s.)
Note that triangles ABP, BCQ, CDR, DAS are directly similar, not necessarily
equilateral.
Once again, complex numbers work wonders! We denote by m the complex affix
of M . From the way P is constructed, we have p − a = α(b − a), that is p =
(1 − α)a + αb. Similarly, q = (1 − α)b + αc, r = (1 − α)c + αd, s = (1 − α)d + αa.
Now, P QRS is a parallelogram if and only if q + s = p + r. A short calculation
shows that the latter is equivalent to

(1 − 2α) ((a + c) − (b + d)) = 0.

1
The conclusion immediately follows. In the case α = 2, P QRS is the Varignon
parallelogram of the quadrilateral ABCD.

Crux Mathematicorum, Vol. 42(5), May 2016

 MICHEL BATAILLE /215

Exercises
Our first exercise is based on problem 858 of The College Mathematics Journal.
1. Let ABCD be a convex quadrilateral that is not a parallelogram. On the sides
AB, BC, CD, DA, construct isosceles triangles KAB, M BC, LCD, N DA external
to ABCD such that the angles at K, L, M, N are right angles. Show that if O is
the midpoint of BD, then one of the triangles M ON or LOK is a 90◦ rotation of
the other around O.
2. Let ABCD be a square and O, P be such that DOC and BCP are equilateral
triangles with O inside ABCD and P external to ABCD. Show that A, O, P are
collinear. [A possible solution follows from the value of ∠AOB found in problem
3458 [2009 : 326 ; 2010 : 347]; preferably solve the problem with the help of a
well-chosen rotation.]

References
[1] H.S.M. Coxeter, S.L. Greitzer, Geometry Revisited, MAA, 1967, p. 61.
[2] M. Bataille, Géométrie plane, avec des nombres, ATOM Vol. XV, p. 36.

Copyright c Canadian Mathematical Society, 2016

216/ THE ENGEL-TITU INEQUALITY

The Engel-Titu Inequality

Eeshan Banerjee
In this article, we present a proof of the ‘Engel-Titu Inequality’ (It’s a long name,
let’s call it E-T!), and some applications of it.
The Engel-Titu Inequality was discovered independently by Arthur Engel in 1998
and Titu Andreescu in 2001. In their book Problems from the Book, Titu An-
dreescu and Gabriel Dopinescu wrote the following note about the inequality:
“[The Engel-Titu Inequality] is clearly a direct application of the Cauchy-
Schwarz inequality. Some will say that it is actually the Cauchy-
Schwarz inequality and they are not wrong. Anyway, this particular
lemma has become very popular among the American students who
attended the training of the USA IMO team. This happened after a
lecture delivered by the first author at the Mathematical Olympiad
Summer Program (MOSP) held at Georgetown University in June,
2001.”
Theorem 1 (E-T) For positive real numbers ak , bk , k ∈ [1, 2, . . . , n], the following
inequality holds.

a21 a2 a2 (a1 + a2 + · · · + an )2
+ 2 + ··· + n ≥ .
b1 b2 bn b1 + b2 + · · · + bn
ai
Equality occurs if and only if all the are equal.
bi
Proof. Let us induct on k.
a2 (a)2
For k = 1, the statement becomes b ≥ b , which is obviously true.
a2 b2 (a+b)2
For k = 2, the statement becomes c + d ≥ c+d . Cross-multiplication yields
Å ã
a2 b2
+ (c + d) ≥ (a + b)2 =⇒
c d
a2 d b2 c
a2 + b2 + + ≥ a2 + 2ab + b2 =⇒
c d
a2 d b2 c
+ ≥ 2ab. (1)
c d
…
a2 d b2 c a2 d b2 c
which is obviously true by AM-GM, as + ≥2 · = 2ab.
c d c d
Now, let us assume that for some positive integer k, the statement is true. That
is, the following inequality is true.

a21 a2 a2 (a1 + a2 + · · · + ak )2
+ 2 + ··· + k ≥ . (2)
b1 b2 bk b1 + b2 + · · · + bk

Crux Mathematicorum, Vol. 42(5), May 2016

 EESHAN BANERJEE /217

Now by (1) and (2) we have

a21 a22 a2 (a1 + a2 + · · · + ak )2 a2k+1 (a1 + a2 + · · · + ak+1 )2

+ + · · · + k+1 ≥ + ≥ .
b1 b2 bk+1 b1 + b2 + · · · + bk bk+1 b1 + b2 + · · · + bk+1
The E-T Inequality follows by induction; the case of equality is easily verified. 2
There’s also a very straightforward proof of the E-T inequality using the Cauchy-
Schwarz inequality. We leave that proof to the readers.

To illustrate the use of the E-T Inequality, let’s work on some problems using this
result.
Example 1 (IMO 1995) Let a, b, c be positive real numbers with product 1. Then
prove that
1 1 1 3
3
+ 3 + 3 ≥ .
a (b + c) b (c + a) c (a + b) 2

1 1 1
Solution. Let us substitute a = , b = , c = . As abc = 1, xyz = 1. We get
x y z
1 1 1 X x2
Å ã Å ã
X X X
= = = .
a3 (b + c) 1 1 1 y+z y+z
cyc cyc · + cyc cyc
x3 y z x3 yz
By E-T and AM-GM inequality, we get
X x2 2 √
X 1 (x + y + z) x+y+z 3 3 xyz 3
3
= ≥ = ≥ = .
cyc
a (b + c) cyc
y+z 2(x + y + z) 2 2 2

Hence we get
1 1 1 3
+ + ≥ .
a3 (b + c) b3 (c + a) c3 (a + b) 2
This completes the proof. 2

The next example is believed to be new.

Example 2 Again letting a, b, c be positive real numbers with product 1, prove that
a3
≥ 32 .
P
cyc
b+c
P a3 P a4
Solution. We may write cyc = cyc . Now by E-T, we get
b+c ab + ac

2
X a4 a2 + b2 + c2
≥ =⇒
cyc
ab + ac 2(ab + bc + ca)

2
2
a2 + b2 + c2 a2 + b2 + c2 a2 + b2 + c2


X a3
≥ ≥ = .
cyc
b+c 2(ab + bc + ca) 2 (a2 + b2 + c2 ) 2

Copyright c Canadian Mathematical Society, 2016

218/ THE ENGEL-TITU INEQUALITY

Therefore, by Power Mean Inequality and AM-GM Inequality with abc = 1, we

have

Å ã Ç √ å2 Å ã2
X a3 a2 + b2 + c2 a+b+c 2 3 3 3 abc 3 3 3 3
= · ≥ · = · = ,
cyc
b + c ≥ 3 2 3 2 3 2 2

which completes the proof. 2

Example 3 (Turkey 1997) Let n ≥ 2. Find the minimal value of

x51 x52 x5n

+ + ··· + ,
x2 + x3 + · · · + xn x1 + x3 + · · · + xn x1 + x2 + · · · + xn−1

where x1 , x2 , · · · xn are positive real numbers satisfying x21 + x22 + · · · + x2n = 1.

Solution. Let s = x1 + x2 + · · · + xn . By E-T inequality we have
Å n
ã2
x3i
P
n n
X x5i X x6i (x31 + x32 + · · · + xn )3 i=1
= ≥ n = .
s − xi sxi − x2i s2 − 1
(sxi − x2i )
P
i=1 i=1
i=1

Now, by the Power-Mean Inequality we have

n
! 13 n
! 21 n
X x3 i
X x2 i
X xi
≥ ≥
i=1
n i=1
n i=1
n

and hence
n
! 23
X x3 i 1 s2
≥ ≥ 2.
i=1
n n n

So we have
1 s2 1 1
≥ 2 =⇒ s2 ≤ n =⇒ s2 − 1 ≤ n − 1 =⇒ 2 ≤
n n s −1 n−1
and
n
!2 n
!2
X x3 i 1 X 1
≥ 3 =⇒ x3i ≥ .
i=1
n n i=1
n

So Å ã2
n
x3i
P
i=1 1
≥ .
s2 − 1 n(n − 1)
Thus, the minimal value of the given expression is 1
n(n−1) . 2

Crux Mathematicorum, Vol. 42(5), May 2016

 EESHAN BANERJEE /219

Problems for the Reader

Problem 1 For positive real numbers a, b, c, prove that
a b c
+ + ≤ 1.
2a + b 2b + c 2c + a
Problem 2 (IMOSL 1993) For arbitrary positive real numbers a, b, c, d prove
the inequality
a b c d 2
+ + + ≥ .
b + 2c + 3d c + 2d + 3a d + 2a + 3b a + 2b + 3c 3
Problem 3 (Japan 1997) Prove that for any positive real numbers a, b, c, the
following inequality holds

(b + c − 1)2 (c + a − 1)2 (a + b − 1)2 3

2 2
+ 2 2
+ 2 ≥
a + (b + c) b + (c + a) c + (a + b)2 5

Problem 4 (Gabriel Dopinescu) Prove that if a, b, c, d > 0 satisfy abc + bcd +

cda + dab = a + b + c + d, then
… … … …
a2 + 1 b2 + 1 c2 + 1 d2 + 1
+ + + ≤ a + b + c + d.
2 2 2 2
Problem 5 (MOSP 2001) Prove that for any positive real numbers a1 , a2 , a3 , a4 , a5 ,
a1 a2 a3 a4 a5 5
+ + + + ≥ .
a2 + a3 a3 + a4 a4 + a5 a5 + a1 a1 + a2 2

References
[1] AoPS Community - www.artofproblemsolving.com/community
[2] Samin Riasat, Basics of Olympiad Inequalities, Unpublished.
[3] Arthur Engel, Problem Solving Strategies, Problem Book in Mathematics,
Springer, 1995.
[4] Titu Andreescu, Gabriel Dopinescu, Problems from the Book, XYZ Press, 2008

Copyright c Canadian Mathematical Society, 2016

220/ PROBLEMS

PROBLEMS
Readers are invited to submit solutions, comments and generalizations to any problem in
this section. Moreover, readers are encouraged to submit problem proposals. Please see
submission guidelines inside the back cover or online.

To facilitate their consideration, solutions should be received by February 1, 2017.

The editor thanks André Ladouceur, Ottawa, ON, for translations of the problems.

4141. Proposed by Leonard Giugiuc, Daniel Sitaru and Oai thanh Dao; modi-
fied by the editor.
a) Let A0 A1 . . . An−1 be a convex n-gon for which there exists an interior point
T such that ∠Ai−1 T Ai = 2π n , i = 1, 2, . . . n (with An ≡ A0 ). Construct
regular n-gons Πi externally on the sides Ai−1 Ai . Prove that
n
1X
[A0 A1 . . . An−1 ] ≤ [Πi ]
n i=1

(where square brackets denote area).

b) ? Does the inequality continue to hold if the given convex polygon is arbi-
trary?

4142. Proposed by Daniel Sitaru.

Prove that if a, b, c ∈ (0, ∞) then:
(a+b+c)2
 a2 + b2 + c2  a2 +b2 +c2  a  b  c
1+ ≤ 1+ 1+ 1+ .
ab + bc + ca b c a

4143. Proposed by Roy Barbara.

For any real number x ≥ 1, let y = x1/2 + x−1/2 .
a) Express x in terms of y by a radical formula and check that no rational
fraction F (t) can exist such that x = F (y). (A rational fraction is an expres-
sion of the form f (t)/g(t), where f (t) and g(t) are polynomials with rational
coefficients.)
b) Find a closed form formula x = F (y) containing no radicals.
c) ? Is there a complex fraction such that x = F (y)? (A complex fraction is
a function of the form f (z)/g(z), where f (t) and g(t) are polynomials with
complex coefficients.)

Crux Mathematicorum, Vol. 42(5), May 2016

 PROBLEMS /221

4144. Proposed by George Apostolopoulos.

Let a, b and c be positive real numbers such that a + b + c = 1. Find the maximum
value of the expression
Å ã Å ã Å ã
1 3 1 3 1 3
a− + b− + c− .
2 2 2

4145. Proposed by Leonard Giugiuc.

Prove that the system
  

 1 2 3
 3 2 2
A + A B + AB + ABA =  0 1 2 ,

  


0 0 1




  
−1 0 3




3 2 2

B + B A + BA + BAB = 0 −1 0 

  

 

0 0 −1
has no solutions in the set of 3 × 3 matrices over complex numbers.

4146. Proposed by Mehmet Berke Işler.

Let a,
√ b, c √
be non-negative
√ real numbers such that a2 + b2 + c2 = 2(ab + bc + ca)
and a + b + c = 2. Prove that at least one of the numbers a, b, c is equal to 1.

4147. Proposed by Mehtaab Sawhney.

Let {ai } be a sequence of real numbers. Suppose that |ai − aj | ≥ 2i−j if i > j,
then find the minimal value of
X
(aj − ai )2 .
1≤i<j≤n

4148. Proposed by Lorian Saceanu.

For positive real numbers x, y and z, show that
» » »
xy(x + y) + yz(y + z) + xz(x + z)
» 2xyz
≥ (x + y)(y + z)(z + x) + (x + y + z) .
3(xy + yz + xz)

Copyright c Canadian Mathematical Society, 2016

222/ PROBLEMS

4149. Proposed by Daniel Sitaru.

h i
Prove that if [a, b] ⊂ 0, π4 then:

3(a tan b + b tan a) ≥ ab(6 + a tan a + b tan b).

4150. Proposed by Leonard Giugiuc.

Let (xn )n≥1 be a sequence of positive real numbers such that
Å ã
2 32
lim xn + 2xn + 3 = 12.
n→∞ xn
Show that limn→∞ xn exists and find its value.

.................................................................

4141. Proposé par Leonard Giugiuc, Daniel Sitaru et Oai thanh Dao; modifié
par l’éditeur.
a) Soit A0 A1 . . . An−1 un n-gone convexe avec un point intérieur T tel que
∠Ai−1 T Ai = 2π n , i = 1, 2, . . . , n (avec An ≡ A0 ). On construit des n-gones
réguliers externes Πi sur les côtés Ai−1 Ai . Démontrer que
n
1X
[A0 A1 . . . An−1 ] ≤ [Πi ]
n i=1

(les crochets représentent l’aire).

b) ? L’inégalité est-elle toujours vérifiée si le polygone convexe donné est plutôt
un polygone arbitraire?

4142. Proposé par Daniel Sitaru.

Sachant que a, b, c ∈ (0, ∞), démontrer que
(a+b+c)2
 a2 + b2 + c2  a2 +b2 +c2  a  b  c
1+ ≤ 1+ 1+ 1+ .
ab + bc + ca b c a

4143. Proposé par Roy Barbara.

Étant donné un réel x (x ≥ 1), soit y = x1/2 + x−1/2 .
a) Exprimer x en fonction de y au moyen d’une expression contenant des rad-
icaux et montrer qu’il n’existe aucune fraction rationnelle F (t) telle que
x = F (y). (Une fraction rationnelle est une expression de la forme f (t)/g(t),
f (t) et g(t) étant des polynômes avec coefficients rationnels.)

Crux Mathematicorum, Vol. 42(5), May 2016

 PROBLEMS /223

b) Déterminer une formule explicite x = F (y) qui ne contient aucun radical.

c) ? Est-il possible d’exprimer x = F (y) au moyen d’une fraction complexe?
(Une fraction complexe est une expression de la forme f (z)/g(z), f (t) et
g(t) étant des polynômes avec coefficients complexes.)

4144. Proposé par George Apostopoulos.

Soit a, b et c des réels strictement positifs tels que a + b + c = 1. Déterminer la
valeur maximale de l’expression
Å ã Å ã Å ã
1 3 1 3 1 3
a− + b− + c− .
2 2 2

4145. Proposé par Leonard Giugiuc.

Démontrer que le système
  

 1 2 3
 3 2 2
A + A B + AB + ABA =  0 1 2 

  


0 0 1




  
−1 0 3




3 2 2

B + B A + BA + BAB =  0 −1 0 

  



0 0 −1
n’admet aucune solution, A et B étant des matrices 3 × 3 de nombres complexes.

4146. Proposé par Mehmet Berke Işler.

c des réels non négatifs tels que a2 + b2 + c2 = 2(ab + bc + ca) et
Soit a,√b et √
√
a + b + c = 2. Démontrer qu’au moins un des nombres a, b et c est égal à 1.

4147. Proposé par Mehtaab Sawhney.

Soit {ai } une suite de réels. Sachant que |ai − aj | ≥ 2i−j lorsque i > j, déterminer
la valeur minimale de X
(aj − ai )2 .
1≤i<j≤n

4148. Proposé par Lorian Saceanu.

Soit x, y et z des réels strictement positifs. Démontrer que
» » »
xy(x + y) + yz(y + z) + xz(x + z)
» 2xyz
≥ (x + y)(y + z)(z + x) + (x + y + z) .
3(xy + yz + xz)

Copyright c Canadian Mathematical Society, 2016

224/ PROBLEMS

4149. Proposé par Daniel Sitaru.

h i
Soit [a, b] ⊂ 0, π4 . Démontrer que

3(a tan b + b tan a) ≥ ab(6 + a tan a + b tan b).

4150. Proposé par Leonard Giugiuc.

Soit (xn )n≥1 une suite de réels strictement positifs tels que
Å ã
32
lim x2n + 2xn + 3 = 12.
n→∞ xn

Démontrer que limn→∞ xn existe et déterminer sa valeur.

Crux Mathematicorum, Vol. 42(5), May 2016

 SOLUTIONS /225

SOLUTIONS
No problem is ever permanently closed. The editor is always pleased to consider for
publication new solutions or new insights on past problems.
Statements of the problems in this section originally appear in 2015: 41(5), p. 212–215.

4041. Proposed by Arkady Alt.

Let a, b and c be the side lengths of a triangle ABC. Let AA0 , BB 0 and CC 0 be
the heights of the triangle and let ap = B 0 C 0 , bp = C 0 A0 and cp = A0 B 0 be the
sides of the orthic triangle. Prove that:
a) a2 (bp + cp ) + b2 (cp + ap ) + c2 (ap + bp ) = 3abc;
b) ap + bp + cp ≤ s, where s is the semiperimeter of ABC.
We received 15 correct solutions and present the solution by Michel Bataille.
We show (a) and (b) in the case when ∆ABC has no obtuse angle and provide a
counter-example in the opposite case.
First, suppose that angles A, B, C are acute. Since ∆AB 0 B is right-angled with
∠AB 0 B = 90◦ , we have AB 0 = c · cos A. Similarly, AC 0 = b · cos A, and it follows
that

B 0 C 02 = c2 cos2 A + b2 cos2 A − 2bc cos3 A

= (c2 + b2 − 2bc cos A) cos2 A = a2 cos2 A

and so ap = B 0 C 0 = a cos A. In a similar way, we obtain bp = A0 C 0 = b cos B and

cp = A0 B 0 = c cos C.
Now we calculate X = a2 (bp + cp ) + b2 (cp + ap ) + c2 (ap + bp ) as follows:

X = a2 b cos B + a2 c cos C + b2 c cos C + b2 a cos A + bc2 cos B + c2 a cos A

= ab(a cos B + b cos A) + bc(b cos C + c cos B) + ca(c cos A + a cos C)
= abc + bca + cab = 3abc,

as desired. Denoting by r and R the inradius and the circumradius of ∆ABC and
using the Law of Sines, we get

ap + bp + cp = a cos A + b cos B + c cos C

= R(sin 2A + sin 2B + sin 2C)
= 4R sin A sin B sin C
abc 4rRs 2r
= 4R · 3
= 2
=s·
8R 2R R
and the result ap + bp + cp ≤ s follows from Euler’s inequality 2r ≤ R.

Copyright c Canadian Mathematical Society, 2016

226/ SOLUTIONS

If ∆ABC is right-angled, say ∠BAC = 90◦ , results (a) and (b) continue to hold
if we take, as is natural, ap = 0, bp = cp = h, where h = AA0 . Indeed, we have
3abc = 3a · ah = 3a2 h and

a2 (bp + cp ) + b2 (cp + ap ) + c2 (ap + bp ) = a2 · 2h + b2 · h + c2 · h

= h(b2 + c2 + 2a2 ) = 3a2 h.
2
√ ap +b2p +cp2 ≤ s 2rewrites as 4h ≤ a+b+c or 4bc ≤ a +a(b+c).
Also, the inequality
Since b + c ≥ 2 bc and a = b + c ≥ 2bc, we have
√ √ √
a2 + a(b + c) ≥ 2bc + 2 2bc · 2 bc = (2 + 2 2)bc ≥ 4bc.

None of these results is correct, however, if an angle of ∆ABC is obtuse, as the

following example shows. Consider a triangle ABC with ∠BAC = 120◦ and
√ a
√
c 3
AB = AC. Then b = c, a = c 3, and ap = bp = cp = 2 = 2 . One easily finds
√
that 3abc = 3c3 3, while
√
a2 (bp + cp ) + b2 (cp + ap ) + c2 (ap + bp ) = 5c3 3.

Also, √ √
3c 3 (2 + 3)c
ap + bp + cp = > = s.
2 2

4042. Proposed by Leonard Giugiuc and Diana Trailescu.

Let a, b and c be real numbers in [0, π/2] such that a + b + c = π. Prove the
inequality
√ a b c √
2 2 sin sin sin ≥ cos a cos b cos c.
2 2 2

We received 14 correct solutions. We present the solution by Scott Brown. Similar

solutions came from Arslanagić S̆efket, Michel Bataille, Andrea Fanchini, and
John Heuvel.
In [1] and [2] respectively, we find the identities

a b c r
sin sin sin = (1)
2 2 2 4R
and
s2 − 4R2 − 4Rr − r2
cos a cos b cos c = , (2)
4R2
where R, r, and s are the circumradius, inradius, and semiperimeter of the triangle.
We square both sides of the original inequality to obtain the equivalent statement

a b c
8 sin2 sin2 sin2 ≤ cos a cos b cos c,
2 2 2

Crux Mathematicorum, Vol. 42(5), May 2016

 SOLUTIONS /227

into which we substitute the identities (1) and (2). The resulting inequality is
equivalent to one due to Gerretsen [3]:

s2 ≤ 4R2 + 4Rr + 3r2 .

References
[1] Anders Bager. “A family of goniometric inequalities.” Univ. Beograd. Publ.
Elektrotehn. Fak. Ser. Mat. Fiz., No. 338 352 (1971), p. 10.
[2] Anders Bager. “Another family of goniometric inequalities.” Univ. Beograd.
Publ. Elektrotehn. Fak. Ser. Mat. Fiz., No. 412 460 (1973), p. 209.
[3] D. S. Mitrinovic et al. Recent Advances in Geometric Inequalities. Kluwer,
Dordrecht, 1989.
Editor’s Comment. Digby Smith pointed out that the inequality is equivalent to
Crux Problem 974, proposed by Jack Garfunkel in Volume 10, (8), October 1984,
and solved by Murray Klamkin in Volume 11 (10), December 1985. The solution
to 974 is based on Crux Problem 836, proposed by Vedula N. Murty in Volume
9 (4), April 1983, and solved, again by Klamkin, in Volume 10 (7), August 1984.

4043. Proposed by Michel Bataille.

Suppose that the lines m and n intersect at A and are not perpendicular. Let
B be a point on n, with B 6= A. If F is a point of m, distinct from A, show
that there exists a unique conic CF with focus F and focal axis BF , intersecting
n orthogonally at A. Given  > 0, how many of the conics CF have eccentricity ?
We recieved two correct solutions and present the solution submitted by the pro-
poser.
Since m 6= n, the perpendicular to m through F and the perpendicular t to n at
A intersect, say at K. Note that K is distinct from both F and A (since F 6= A).
Define p to be the perpendicular to BF through K. Then A ∈ / p (otherwise we
would have t ⊥ BF , implying n k BF , a contradiction). We also have F ∈ / p
(otherwise KF ⊥ BF , implying BF = m and B ∈ m, contradicting B 6= A).
We first show uniqueness: Suppose that CF exists. Note that t is the tangent
to CF at A. Since ∠KF A = 90◦ and K ∈ t, K must be on the directrix of CF
associated with F (see [2], Theorem 1 p. 14). Thus, CF must be the unique conic
AF
with focus F , directrix p and eccentricity d(A,p) . Conversely, because the line p
misses the distinct points A and F , we can consider the unique conic with focus F ,
AF
directrix p and eccentricity d(A,p) . This conic passes through A (by the definition
of eccentricity) and is tangent to AK at A (since K ∈ p and ∠KF A = 90◦ ); it
therefore intersects n orthogonally at A. Also, its focal axis is BF (since BF ⊥ p).
Thus, this conic satisfies the required conditions for CF .
Note that the eccentricity of CF is also equal to FF B A (see [2], Theorem 4, p. 18)
and that if F, F 0 are two distinct points on m (F, F 0 6= A), then the conics CF and

Copyright c Canadian Mathematical Society, 2016

228/ SOLUTIONS

CF 0 are distinct (their focal axes are distinct). From these remarks, we see that
there are as many conics CF with eccentricity  as points of M ∈ m that belong to
the locus E of points for which M B
M A = . If  = 1, E is the perpendicular bisector
of AB; it intersects m (since m and n are not perpendicular), so that exactly one
conic CF is a parabola. If  6= 1, then E is a circle—the circle of Apollonius—which
can intersect m in at most two points. The collection of all these circles (as  varies
over the positive real numbers except 1) forms a nonintersecting pencil of circles
with limiting points A and B, one through each point of the plane not on the
perpendicular bisector of AB (see [1], Section 6.6). It follows that there are at
most two conics CF corresponding to a given value of .

To be more specific, E has diameter JJ 0 where J, J 0 are the points of n defined

−→ −−→ −−→ −−→
by (1 + )AJ = AB and (1 − )AJ 0 = AB. The centre U of E is such that
−→ −−
→
(1 − 2 )AU = AB; its radius is ρ = JU = AU = |1− AB
2| . If H, H 0 are the
0
orthogonal projections of B, U onto m, respectively, then UBH H
= ABAU
= |1−1
2| ;

hence U H 0 = |1−2 | (where BH is the distance from B to m which, of course, is

BH

always less than AB). We conclude that no, one, or two conics CF have eccentricity
 according as U H 0 is greater than, equal to, or less than ρ, which is equivalent to
 less than, equal to, or greater than BH
AB . So, for example, CF could never be a
circle (for which  = 0).

References

[1] H.S.M. Coxeter, Introduction to Geometry, Wiley, 1961.

[2] C. V. Durell, A Concise Geometrical Conics, MacMillan, 1952.

4044. Proposed by Dragoljub Milošević.

Let x, y, z be positive real numbers such that x + y + z = 1. Prove that

x+1 y+1 z+1 27

+ + ≤ .
x3 + 1 y 3 + 1 z 3 + 1 7

We received 24 submissions, of which 22 were correct and complete. There were

two main approaches: Jensen’s inequality, or comparing each term to a linear
function. We present two solutions, one for each approach.

Solution 1, by Fernando Ballesta Yagüe.

As x + y + z = 1 and x, y, z are positive, we have x, y, z ∈ (0, 1). For x in the

interval (0, 1), consider the rational function

x+1 1
f (x) = = 2 .
x3 + 1 x −x+1

Crux Mathematicorum, Vol. 42(5), May 2016

 SOLUTIONS /229

Let’s take the second derivative to check its convexity:

−2x + 1
f 0 (x) = ,
(x2 − x + 1)2
−2(x2 − x + 1)2 − (−2x + 1) · 2 · (x2 − x + 1) · (2x − 1) 6x(x − 1)
f 00 (x) = = 2 .
(x2 − x + 1)4 (x − x + 1)3
Since x ∈ (0, 1), we have x − 1 < 0, but 6x > 0 and x2 − x + 1 > 0 (note that
−x + 1 > 0 for x ∈ (0, 1)). So in the interval x ∈ (0, 1) we have f 00 (x) < 0 and
hence f is concave. By Jensen’s Inequality for concave functions,
Å ã Å ã
1 1 1 9
(f (x) + f (y) + f (z)) ≤ f (x + y + z) = f = ;
3 3 3 7
in other words,
1 1 1 27
+ 2 + 2 ≤ ,
x2 −x+1 y −y+1 z −z+1 7
which is equivalent to the inequality we wanted to prove. Note that equality holds
when x = y = z = 31 .

Solution 2, by Paul Bracken.

1
As in the previous solution, define f (x) = x2 −x+1 and show that f is concave for
x ∈ (0, 1). Therefore, if t(x) is a tangent line to f (x) at some point x0 ∈ (0, 1)
then the inequality f (x) ≤ t(x) holds for x ∈ (0, 1). Let us calculate the tangent
line to f (x) at x0 = 13 :
Å ã Å ã Å ã
0 1 1 1 27 54
t(x) = f · x− +f = x+ .
3 3 3 49 49
The inequality f (x) ≤ t(x) then gives us xx+1 27 54
3 +1 ≤ 49 x + 49 for x ∈ (0, 1). We

obtain similar inequalities by replacing x by y and z respectively, then add the

three inequalities to get
x+1 y+1 z+1 27 54 27
3
+ 3 + 3 ≤ (x + y + z) + 3 · = ,
x +1 y +1 z +1 49 49 7
where for the last equality we used x + y + z = 1.

4045. Proposed by Galav Kapoor.

Suppose that we have a natural number n such that n ≥ 10. Show that by changing
at most one digit of n, we can compose a number of the form x2 + y 2 + 10z 2 , where
x, y, z are integers.
We received two correct solutions. We present the solution by Roy Barbara.
Recall that Legendre’s three-square theorem states that a natural number is a sum
of three squares if and only if it is not of the form 4m (8k + 7). In particular, any
natural number of the form 4k + 2 is a sum of three squares.

Copyright c Canadian Mathematical Society, 2016

230/ SOLUTIONS

Now let 2k +1 be any odd natural number. Then we can write 4k +2 = a2 +b2 +c2 .
Using a2 + b2 + c2 ≡ 2 (mod 4), it is clear that exactly one of a, b, c is even, say c.
Setting x = 21 (a + b), y = 12 (a − b), c = 2z yields

4k + 2 = 2x2 + 2y 2 + 4z 2 ,

whence
2k + 1 = x2 + y 2 + 2z 2 .

Finally let n ≥ 10. By changing the last digit of n to a 5 (if necessary), we obtain
a number of the form 10k + 5 for which we have

10k + 5 = 5x2 + 5y 2 + 10z 2 = (2x + y)2 + (2y − x)2 + 10z 2 .

4046. Proposed by Michel Bataille.

√ √ √
Let a, b, c be nonnegative real numbers such that a + b + c ≥ 1. Prove that
»
a2 + b2 + c2 + 7(ab + bc + ca) ≥ 8(a + b)(b + c)(c + a).

Two correct solutions were received. A purported counterexample that was submit-
ted had an error. We present both solutions.

Solution 1, by Madhav R. Modak.

»
8(a + b)(b + c)(c + a)
√ √ √ »
≤ ( a + b + c) 8(a + b)(b + c)(c + a)
» »
= (4ab + 4ca)[2(a + b)(c + a)] + (4bc + 4ab)[2(a + b)(b + c)]
»
+ (4ca + 4bc)[2(b + c)(c + a)]
1 1
≤ [(4ab + 4ca) + 2(a + b)(c + a)] + [(4bc + 4ab) + 2(a + b)(b + c)]
2 2
1
+ [(4ca + 4bc) + 2(b + c)(c + a)]
2
= 4(ab + bc + ca) + (a + ab + ca + bc) + (b2 + ab + bc + ca) + (c2 + ca + bc + ab)
2

= a2 + b2 + c2 + 7(ab + bc + ca),

which yields the desired result.

Solution 2, by the proposer.

Since

(a2 + 3ab + 3ca + bc)2 = 8a(a + b)(b + c)(c + a) + (a − b)2 (a − c)2 ,

Crux Mathematicorum, Vol. 42(5), May 2016

 SOLUTIONS /231

it follows that
√ »
a2 + 3ab + 3ca + bc ≥ a( 8(a + b)(b + c)(c + a)).

Similarly
√ »
b2 + 3ab + 3bc + ca ≥ b( 8(a + b)(b + c)(c + a))

and
√ »
c2 + 3ca + 3bc + ab ≥ c( 8(a + b)(b + c)(c + a)).

Adding these three inequalities yields the result.

Editor’s comment. Equality holds if and only if a = b = c = 1/9.

4047?. Proposed by Abdilkadir Altintaş.

Let ABC be a triangle with circumcircle O, orthocenter H and ∠BAC = 60◦ .
Suppose the circle with centre Q is tangent to BH, CH and the circumcircle of
ABC. Show that OH ⊥ HQ.

All 14 submissions we received were correct. We feature two solutions.

Copyright c Canadian Mathematical Society, 2016

232/ SOLUTIONS

Solution 1 is a composite of solutions by Václav Konečný and Edmund Swylan.

The statement of the problem is faulty: Because the plane is partitioned into as
many as eight regions by the circumcircle of triangle ABC and the lines HB and
HC, there could be eight tritangent circles and, consequently, eight choices for
Q, of which some lie on the Euler line OH (in which case the lines OH and HQ
would be coincident, not perpendicular).
[Editor’s comment: Since the centres of all circles tangent to HB and HC would
lie on a bisector of ∠BHC, the requirement that the circle be tangent also to the
circumcircle was perhaps included to limit the choice of tritangent circle to the
incircle of the curvilinear triangle HBC (formed by the line segments HB and
HC and the circular arc BC). Then the problem has been correctly stated for an
acute ∆ABC, but it is still not correct when there is an obtuse angle at B or C.]
The exact location of Q is not relevant to the correct theorem:
For any circle tangent to the lines HB and HC, its centre Q must
belong to one of the two bisectors of ∠BHC, and so must O.
The claim for Q is a familiar theorem, while the claim for O depends on ∠BAC =
60◦ and must be proved.
As in the figure, denote the midpoints of AC and AB by M and N , respectively,
and the feet of the altitudes to these lines by F and G. Then the segments F M
and GN are congruent:

AC 1
F M = |F A − M A| = AB · cos 60◦ − = |AB − AC|
2 2

and
AB 1
GN = |GA − N A| = AC · cos 60◦ − = |AC − AB|.
2 2
Then the lines BF, CG, M O, N O form the sides of a rhombus for which the line
OH is a diagonal. Thus OH bisects one of the angles formed by the lines HF and
HG, as claimed.

Solution 2 is a composite of similar solutions by Šefket Arslanagić, Ricardo Barroso

Campos, Prithwijit De (done independently), and Adnan Ibrić with Salem Malikić.
As in the figure that accompanies the statement of the problem, we assume that
the given triangle is acute, and that F and G are the feet of the altitudes from
B and C, respectively. Observe that ∠BHC = ∠F HG = 120◦ (since ∠A is
60◦ and is opposite ∠F HG in the circle whose diameter is AH). Furthermore,
∠BOC = 120◦ also (because O is the centre of the circumcircle so that the angle
there is twice the angle BAC = 60◦ which is inscribed in that circle). Because
O and H are on the same side of BC, it follows that B, O, H, C are concyclic.
Finally, note that because ∆BOC is isosceles, ∠OCB = 30◦ . Since HQ is the

Crux Mathematicorum, Vol. 42(5), May 2016

 SOLUTIONS /233

bisector of ∠BHC,
1
∠BHQ = ∠BHC = 60◦ .
2
Combine this with
∠OHB = ∠OCB = 30◦ ,
and conclude that
∠OHQ = ∠OHB + ∠BHQ = 90◦ .

Editor’s Comments. Essentially the same problem has appeared before in Crux
[1988: 165; 1990: 103] as Problem M1046, which was taken from the 1987 U.S.S.R
journal Kvant:
If ∠A = 60◦ then one of the bisectors of the angle between the altitudes
from B and C passes through O.
This and related properties were discussed under the heading “Property 3” in
the article “Recurring Crux Configurations 3: Triangles Whose Angles Satisfy
2B = C + A” [2011: 350].

4048. Proposed by Leonard Giugiuc and Daniel Sitaru.

Let n ≥ 2 be an integer and let ak ≥ 1 be real numbers, 1 ≤ k ≤ n. Prove the
inequality
Å ã Å ã Å ã
1 1 1 1
a1 a2 · · · an − ≥ a1 − + a2 − + · · · + an −
a1 a2 · · · an a1 a2 an
and study equality cases.
Thirteen solutions were received, all of which established the inequality. Two of
them did not get all the possible conditions for equality, while three others neglected
to consider when equality occurred. The solutions were all similar to the one
presented below.
Let
1
f (x) = x −
x
and observe that, for x, y ≥ 1,

f (xy) − f (x) − f (y) = (xy)−1 (xy − 1)(x − 1)(y − 1) ≥ 0

with equality if and only if at least one of x and y is equal to 1.

We establish the result by induction.
The foregoing shows that it is true for n = 2. Suppose that the inequality holds
for n = m ≥ 2 with equality iff all but at most one of a1 , a2 , . . . , am is equal to 1.
Then, by the foregoing property of f and the result for n = m,
m
X m+1
X
f (a1 a2 · · · am am+1 ) ≥ f (a1 a2 · · · am )+f (am+1 ) ≥ f (ak )+f (am+1 ) = f (ak ).
k=1 k=1

Copyright c Canadian Mathematical Society, 2016

234/ SOLUTIONS

Equality holds if and only if either

• a1 a2 · · · am = 1, in which case a1 = a2 = · · · = am = 1, or
• am+1 = 1 and f (a1 a2 · · · am ) = m
P
k=1 f (ak ).

In either case, all but at most one of a1 , a2 , . . . , am+1 is equal to 1.

Editor’s comments. One can also peel off the last two terms in the product so that
the induction step becomes
m−1
X m+1
X
f (a1 a2 . . . am am+1 ) ≥ f (ak ) + f (am am+1 ) ≥ f (ak ).
k=1 k=1

Edmund Swyland observed that if, for any i and j, you replaced the pair (ai , aj )
by (ai aj , 1), the left side f (a1 a2 · · · an ) of the inequality remained unchanged, but
the right side increased. Thus we can reduce the problem to establishing that it
holds when all but two of the ai are equal to 1, and this now involves dealing with
the case n = 2.
Kee-Wai Lau pointed out that an easy induction argument yields
n n
X X (a1 a2 · · · ak−1 − 1)(ak − 1)(a1 a2 · · · ak − 1)
f (a1 a2 · · · an ) − f (ak ) = .
a1 a2 · · · ak
k=1 k=2

4049. Proposed by Mihaela Berindeanu.

Evaluate
sin x − x cos x
Z
dx
(x + sin x)(x + 2 sin x)
for all x ∈ (0, π/2).
We received 16 submissions all of which were correct. We present a composite of
the nearly identical solutions given by Adnan Ali, Michel Bataille, Prithwijit De,
Joseph Ling and Albert Stadler, all done independently
Let I denote the given integral. Since it is readily checked that

(1 + cos x)(x + 2 sin x) − (1 + 2 cos x)(x + sin x) = sin x − x cos x,

we have
Z Å ã
1 + cos x 1 + 2 cos x
I= − dx
x + sin x x + 2 sin x
= ln(x + sin x) − ln(x + 2 sin x) + C
Å ã
x + sin x
= ln + C,
x + 2 sin x

where C is an arbitrary constant.

Crux Mathematicorum, Vol. 42(5), May 2016

 SOLUTIONS /235

4050. Proposed by Mehtaab Sawhney.

Prove that Ç å Ç å2
2n
X 4n 4n
=
k, k, 2n − k, 2n − k 2n
k=0

for all nonnegative integers n.

We received twelve correct solutions which were split between an arithmetic proof
and a proof by double counting, so we present a solution of each type.
Solution 1, by C.R. Pranesachar.
We have
2n Ç å 2n Ç åÇ å2
X 4n X 4n 2n
=
k, k, 2n − k, 2n − k 2n k
k=0 k=0
Ç å 2n Ç åÇ å
4n X 2n 2n
=
2n k 2n − k
k=0
Ç å2
4n
= ,
2n

where the last equality is due to Vandermonde’s identity.

Solution 2, by Joseph DiMuro.

Let’s say we have a classroom with 4n students. The teacher wants to choose
2n of them to work on one project and 2n of them to work on a second project
(independently of each other – students may be assigned to both or neither of the
projects). In how many ways can the teacher assign students to the projects?
On one hand there are 4n


2n ways to choose the students for each of the two projects,

2
thus 4n
2n possibilities altogether.
On the other hand note that if k students are assigned to both projects then 2n−k
will be assigned to just the first project, 2n − k to just the second project, and
k to neither project. So the teacher can proceed as follows: first decide on the
number k of students that will be assigned to both projects, then partition the
class into four groups – of size k (both projects), 2n − k (first project), 2n − k
(second project), and k (neither project). There are
2n Ç å
X 4n
k, k, 2n − k, 2n − k
k=0

ways to do this. Thus the two sides in the problem are equal.

Copyright c Canadian Mathematical Society, 2016

236/ AUTHORS’ INDEX

AUTHORS’ INDEX
Solvers and proposers appearing in this issue
(Bold font indicates featured solution.)

Proposers
George Apostolopoulos, Athens, Greece : 4144
Roy Barbara, Lebanese University, Fanar, Lebanon : 4143
Mehmet Berke Işler, Denizli, Turkey : 4146
Leonard Giugiuc, Drobeta Turnu Severin, Romania : 4145, 4150
Leonard Giugiuc, Daniel Sitaru and Oai thanh Dao, Romania and Vietnam : 4141
Lorian Saceanu, Harstad, Norway : 4148
Mehtaab Sawhney, Commack High School, Commack, NY, USA : 4147
Daniel Sitaru, Drobeta Turnu Severin, Romania : 4142, 4149

Solvers - individuals
Arkady Alt, San Jose, CA, USA : 4041, 4042, 4044, 4049
Adnan Ali, Munbia, India : 4041, 4042, 4044, 4048, 4049, 4050
George Apostolopoulos, Messolonghi, Greece : 4041, 4044, 4049
Šefket Arslanagić, University of Sarajevo, Sarajevo, Bosnia and Herzegovina : CC172,
CC173, CC174, 4041, 4042, 4044, 4047, 4049
Roy Barbara, Lebanese University, Fanar, Lebanon : 4044, 4045
Ricardo Barroso Campos, University of Seville, Seville, Spain : 4041, 4047
Michel Bataille, Rouen, France : 4041, 4042, 4043, 4044, 4046, 4047, 4048, 4049, 4050
Michaela Berindeanu, Bucharest, Romania : 4049
Paul Bracken, University of Texas, Edinburg, TX, USA : 4044
Scott Brown, Auburn University, Montgomery, AL, USA : 4042
Matei Coiculescu, East Lyme High School, CT, USA : CC173
Prithwijit De, Homi Bhabha Centre for Science Education, Mumbai, India : 4041, 4047,
4049, 4050
Paul Deiermann, Southeast Missouri State University, Cape Gurardeau, MO, USA : 4049
Joseph DiMuro, Biola University, La Mirada, CA, USA : 4050
Marian Dincǎ, Bucharest, Romania : 4042
Andrea Fanchini, Canù, Italy : CC172, CC173, 4041, 4042, 4044, 4047
Leonard Giugiuc, Drobeta Turnu Sevein, Romania : 4049
Elnaz Hessami Pilehrood, Marc Garneau Collegiate Institute, Toronto, ON: CC171
John G. Heuver, Grande Prairie, AB : CC173, 4041, 4042, 4047
Jacob Hyder, Auburn University Montgomery, Montgomery, Al, USA: CC174
Galav Kapoor, India : 4045
Vaclav Konecny, Big Rapids, MI, USA : 4047, 4049
Aimee Krug, Northern Kentucky University, Highland Heights, KY, USA : 4050
Kee-Wai Lau, Hong Kong, China : 4044, 4048
Kathleen Lewis, University of the Gambia, Brikama, Republic of the Gambia: CC171,
CC172, CC173
Joseph M. Ling, University of Calgary, Calgary, AB : 4049
Salem Malikić, student, Simon Fraser University, Burnaby, BC : 4042, 4044, 4048
David E. Manes, SUNY at Oneonta, Oneonta, NY, USA : CC171, CC172, CC173, CC174

Crux Mathematicorum, Vol. 42(5), May 2016

 AUTHORS’ INDEX /237

Phil McCartney, Northern Kentucky University, Highland Heights, KY, USA : 4042, 4044
Dragoljub Milošević, Gornji Milanovac, Serbia : 4044
Madhav R. Modak, formerly of Sir Parashurambhau College, Pune, India : 4041, 4043,
4044, 4046, 4047, 4048, 4049, 4050
Paolo Perfetti, Dipartimento di Matematica, Università degli studi di Tor Vergata Roma,
Rome, Italy : 4044
Ricard Peiró i Estruch. IES “Abastos” València, Spain : 4041, 4044, 4047
Angel Plaza, University of Las Palmas de Gran Canaria, Spain : CC173, CC174
C.R. Pranesachar, Indian Institute of Science, Bangalore, India : 4041, 4050
Jordan Price, Auburn University Montgomery, Montgomery, Al, USA : CC172
Henry Ricardo, Tappan, NY, USA : CC171, CC174, 4044, 4049
Mehtaab Sawhney, Commack High School, Commack, NY, USA: 4050
Joel Schlosberg, Bayside, NY, USA : 4041, 4044, 4050
Digby Smith, Mount Royal University, Calgary, AB : CC171, CC172, CC173, CC174,
4042, 4044, 4049, 4050
Albert Stadler, Herrliberg, Switzerland : 4044, 4049, 4050
Edmund Swylan, Riga, Latvia : 4041, 4042, 4047, 4050
Peter Y. Woo, Biola University, La Mirada, CA, USA : 4047
Fernando Ballesta Yagüe, I.E.S. Infante don Juan Manuel, Murcia, Spain, CC171, CC172,
CC173, CC174, 4044
Konstantine Zelator, University of Pittsburgh, Pittsburgh, PA, USA : CC171, CC172,
CC173, CC174
Titu Zvonaru, Cománeşti, Romania : CC171, CC172, CC173, CC174, 4041, 4042,
4044, 4047, 4048

Solvers - collaborations
Dionne Bailey, Elsie Campbell, and Charles R. Diminnie, Angelo State University, San
Angelo, TX, USA : 4044
Dionne Bailey, Elsie Campbell, Charles Diminnie, and Karl Havlak, Angelo State
University, San Angelo, TX, USA : 4049
Prithwijit De and M.A. Prasad, Mumbai, India : 4048
Leonard Giugiuc and Daniel Sitaru, Romania : 4048
Leonard Giugiuc and Diana Trailescu, Romania : 4042
Adnan Ibric, student, University of Tuzla, Bosnia and Herzegovina, and Salem Malikić,
Burnaby, BC : 4047
Angel Plaza, University of Las Palmas de Gran Canaria, Spain and Lucia Ma Li, Isabel
de España High School, Las Palmas de Gran Canaria, Spain : CC172

Copyright c Canadian Mathematical Society, 2016