Crux

Mathematicorum
Published by the Canadian Mathematical Society.

[Link]

The Back Files

The CMS is pleased to offer free access to its back file of all
issues of Crux as a service for the greater mathematical
community in Canada and beyond.

Journal title history:

➢ The first 32 issues, from Vol. 1, No. 1 (March 1975) to
Vol. 4, No.2 (February 1978) were published under the
name EUREKA.
➢ Issues from Vol. 4, No. 3 (March 1978) to Vol. 22, No.
8 (December 1996) were published under the name
Crux Mathematicorum.
➢ Issues from Vol 23., No. 1 (February 1997) to Vol. 37,
No. 8 (December 2011) were published under the
name Crux Mathematicorum with Mathematical
Mayhem.
➢ Issues since Vol. 38, No. 1 (January 2012) are
published under the name Crux Mathematicorum.
 CRUX
Mathematicorum
VOLUME 15 # 1 JANUARY / JANVIER 1989

CONTENTS

An Extension of Oppenheim's Area

Inequality for Triangles Ji Chen 1

The Olympiad Corner: No. 101 R.E. Woodrow 3

Problems: 1401-1410 12

Solutions: 1199, 1292-1298, 1300, 1302-1305 14

Q>
Canadian Mathematical Society
Societe Mathematique du Canada
ISSN 0705-0348 Seeond Class Mail Registration Number 5432

Founding Editors: Leopold Sauve, Frederick G.B« Maskell

Editor: G*W, Sands
Managing Editor: G.P. Wright

GENERAL INFORMATION

Crux Mathematicorum is a problem-solving journal at the senior secondary and

university undergraduate levels for those who practise or teach mathematics.
Its purpose is primarily educational, hut it serves also those who read it for
professional* cultural or recreational reasons*

Problem proposals, solutions and short notes intended for publication should
be sent to the Editor:

G*W, Sands
Department of Mathematics & Statistics
University of Calgary
Calgary, Alberta
Canada, T2N 1N4

SUBSCRIPTION INFORMATION

Crux is published monthly (except July and August). The 1989 subscription rate
for ten issues is $17.50 for members of the Canadian Mathematical Society and
$35.00 for non-members. Back issues^ $3.50 each. Bound volumes with index:
volumes 1 & 2 (combined) and each of volumes 3, 4, 7, 8, 9 and 10: $10.00.
(Volumes 5 & 6 are out-of-print). All prices quoted are in Canadian dollars.
Cheques and money orders, payable to the CANADIAN MATHEMATICAL SOCIETY,
should be sent to the Managing Editor:

Graham P. Wright
Canadian Mathematical Society
577 King Edward
Ottawa, Ontario
Canada KIN 6N5

ACKNOWLEDGEMENT

The support of the Departments of Mathematics and Statistics of the University

of Calgary and Carleton University, and of the Department of Mathematics of the
University of Ottawa, is gratefully acknowledged.

© Canadian Mathematical Society, 1989

Published by the Canadian Mathematical Society

Printed at Carleton University
 - 1 -

AN EXTENSION OF OPPENHEIM'S AREA INEQUALITY

FOR TRIANGLES
Ji Chen

The area A of a triangle is a well known function of the lengths a, 6, c of its sides:
A= [s(s-a)(s-b)(s-c)]U2
(where s = (a + b + c)/2), or equivalently
16A2 = 2 6 V + 2c2a2 + 2a262 - a4 - 64 - c4. (1)
In [3], A. Oppenheim proved that, if 0 < p < 1, then
i|A2(a,&,c)lP<i|A2(ap,&p,^) (2)
(here A(x,y,z) denotes the area of the triangle with sides x} y, z), which from (1) can be
written
3 I - P ( 1 6 A 2 ) P < 262pc2P + 2 c 2P a 2p + 2a2H2p - a 4 p - bip - c 4 p , (3)

0 < p < 1. Equality holds (for 0 < p < 1) if and only if a = b = c.

In this paper, we will prove that the reverse of inequality (3) holds for all other
values of p.
Theorem. For p < 0 or p > 1,
Zl-*(16b2y > 2b2pc2p + 2c2pa2p + 2a 2p 6 2p - a 4p - 64p - c 4p ; (4)
equivalently, if ap, 6P, c p are the sides of a triangle,

i|A 2 (a,6,c)] P >i|A 2 (a p ,6 p ,c p ). (5)

Proof. If a p , 6P, cp are not the edges of a triangle, we have
(a p + bp + cp)(bp + cp - a p )(c p + ap - bp){ap + bp - cp) < 0,
and (4) is true obviously. Thus we suppose ap, 6P, cp are the edges of a triangle, and show
(5). Also, equality holds in (5) when a - b - c, so we suppose that a, by c are not all equal.
When p > 1, then 0 < 1/p < 1. Using (2) on the triangle with edges ap, 6P, cp, we
have
p | A 2 ( a p , 6 p , c p ) l l / V < ^A 2 (a,6,c),

which is the same as (5).

Assume p < 0. We define a function / on (-OD,+OD) by
r 2x 2x
b c + c2xa2z + a2xb2x]1/X \aix + b*x + c 4x + 3(16A 2 /3) Z ] l/X
/(*) = x#0,

"16A5 1/2
/(0) = (a&c) 4 ' 3 -(a&c)* 2 / 3
 - 2 -

Then / i^s continuous at x = 0 (take logarithms and apply L'Hopital's rule).

Since we are assuming that a, b, c are not all equal, from the well known inequality
A< ^(abc)2'*

(see item 4.14 of [2]) we know /(0) > 0. Thus we can find q < 0, q > p such that f(q) > 0,
i.e.
^q^ql l q
•b2qc2q + c2qa2q +
' \a4q + biq + ciq + 3(16A2/3)<i] *'«

or
2b2qc2q + 2c2qa2q + 2a2qb2q < aiq + biq + ciq + 3(16A 2 /3)«,
or, from (1),
16A2(aq,bq,cq) < ™*\a,b,c) (6)
For p < q < 0, 0 < q/p < 1. Using (2) on the triangle with edges ap, bp, cp we obtain

^A2(ap,6p,cp)| ^ P < ™A2{aq,bq,cq). (7)

From (6) and (7) we have

16A 2 (a p ,6 p ,c p ) I'V
^A2(a,b,c) >

so that, since q/p > 0,

^A2(a,6,C)lP>i|A2(ap,6p,cp),
which is (5). a

As a special case, take p = —1/2 in (4); then we have the inequality

3^3 N 2 , 2 , 2 1 1 1
T7T- T — -a22 —"5
4 7 T >" -Tl^ + — "t Ta5 X2 T2-
ca
This could also be written

1 1 1 .3^3 1_1 1 1_1' 1T 2

a2 + F +
P - 47T" +
c a + a 7> +
For another application of the theorem, we extend the following inequality of S.
Beatty [1]:
^>_{K-H)^K-5H)t (8)

where
//= (a2 + h2+ c 2 )/2, K^bc+ca+ab
(see also item 4.18 of [2]). Suppose that p > 1 or p < 0 and that the triples (a,h,c) and
(a p ,6 p ,c p ) are each the sides of a triangle. Then from (8) applied to the triangle with sides
ap, P , c p ,
 _ 3 -

A\ap}b^cp) > (Kp - HP){3KP - 5F P ),

where
c
Hp = 5 ± 0 + 9 RP = 6*c* + c ^ + ap&*.
By using (5) we thus get
A2P y 2^nv'\Kp - £Tp)(3irp - 5ffp).

[1] S. Beatty, Upper and lower estimates for the area of a triangle, Trans. Roy. Soc.
Canada III (3), 48 (1954), 1-5.
[2] 0. Bottema et al, Geometric Inequalities, Wolters—Noordhoff, Groningen, 1969.
[3] A. Oppenheim, Inequalities involving elements of triangles, quadrilaterals or
tetrahedra, Univ. Beograd Puhl Elektrotehn. Fak. Ser. Mat. Fiz.9 496 (1974), 257-263

Department of Mathematics
University of Ningbo
Ningbo, Zhejiang
The People's Republic of China
* * *

THE OLYMPIAD CORNER

No. 101
R.E. WOODROW

All communications about this column should be sent to Professor R.E. Woodrow,
Department of Mathematics and Statistics, The University of Calgary, Calgary, Alberta,
Canada, T2N IN4.

This column marks the second anniversary of my taking over the Corner. In spite of
the odd error, the occasional misquoted problem, and misunderstanding of solutions, I hope
that the resulting column provides a forum for discussion of the Olympiads, and is a source
of good problems for those who coach others (or who are preparing themselves) for
mathematics contests. For the rest of us the problems provide the pleasure of a good teaser.
Pm grateful to all those who have sent in problem sets, comments, and solutions. Those
whose contributions were used in Volume 14 include Beno Arbel, Francisco Bellot, Curtis
Cooper, Ed Doolittle, George Evagelopoulos, Richard Gibbs, Douglass Grant, J.T.
Groenman, Branko Grunbaum, R.K. Guy, F.D. Hammer, Walther Janous, Murray Klamkin,
A.H. Lachlan, Andy Liu, Alan Mekler, John Morvay, V.N. Murty, Richard Nowakowski,
 _ 4 -

Peter O'Halloran, Bob Prielipp, Josef Rifa i Coma, Daniel Ropp, Leo Schneider, M. Selby,
Robert E. Shafer, Zun Shan, Bruce Shawyer, D.J. Smeenk, Dan Sokolowsky, D. Vathis,
David Vaughan, O.R. Veldkarnp, and Edward [Link]. Wang. Keep up the contributions —
without your help this column could not exist! (Looking over this list I find surnames
beginning with each letter of the alphabet except F, I, Q, T, U, X, Y, Z. Perhaps this year
we'll hit every letter.)

This month's Olympiad problems begin with the final rounds of the Flanders
Mathematics Olympiad, 1985—86 and 1986—87. Thanks go to Bruce Shawyer who forwarded
them to me. In each exam, students were allowed three hours to solve the problems. Some
of these problems will be familiar to readers, and there are obvious possibilities for
generalization. Please send in your "nice" solutions.

FLANDERS MATHEMATICS OLYMPIAD 1985-86 (final round)

L A circle with radius R is divided

into twelve equal parts. The
twelve dividing points are connected with the
centre of the circle, producing twelve rays.
Starting from one of the dividing points a segment
is drawn perpendicular to the next ray in the
clockwise sense; from the foot of this
perpendicular another perpendicular segment is
drawn to the next ray, and the process is
continued ad infinitum. What is the limit of the
sum of these segments (in terms of R)l

2. Prove that, for every natural number n, we have

3. A sequence of numbers {a/J is defined as follows:

a0 ~ 0,
aj c+1 = 3ajj + 1, k > 0.
Show that ai55 is divisible by 11.

4. To put a marble with radius 1 cm in a cube it is obvious that the cube must
have an edge with at least a length of 2 cm. What is the minimum length of
the edge of a cube which can contain two marbles of radius 1 cm? (Prove your answer.)
 - 5 -

FLANDERS MATHEMATICS OLYMPIAD 1986-87 (final round)

1. A rectangle ABCD is given. On ^ L 2 n B

the side AB} n different points are 1
chosen strictly between A and B. Similarly, m
2
different points are chosen on the side AD. Lines
are drawn from the points parallel to the sides.
W/////////A 7////7A
How many rectangles are formed in this way? m
(One possibility is shown in the figure.) D

2. Two parallel lines a and 6 meet two other lines c and d. Let A and A' be the
points of intersection of a with c and d9 respectively. Let B and J3' be the
points of intersection of h with c and d, respectively. If Xis the midpoint of the line segment
AA' and Fis the midpoint of the segment BB', prove that
]XY]i\AB\ +\A<B>

(\XY\ represents the length of the line segment XY.)

3. Determine all continuous functions / : R -* IR such

(/(z)) 3 = ffi + fxf(x) + 16(/(x)) 2 ]
4. Prove that, for every r € IR with r > 1
l r + 2 r + - ••+(n-l)r+nr+(n-l)r+- • .+2r+lr|
lim + 0D.
I n2 /
What is the value of the limit when r- 11

Next we give four problems from the Arany Daniel Competition 1987, Junior Level
(age 15), from Hungary. These were collected by Gy. Karolyi and J. Pataki, and forwarded
to me by Bruce Shawyer.
t. The real numbers x, y, z satisfy the following equation:
y2 + z2 - x2 + z2 + x2 - y2 , x2 + y2 - z2
2yz 2zx = 1.
2xy
Prove that two of the three fractions have the value 1

2. The median lines of a convex quadrilateral divide it into four smaller ones.
Prove that the sum of the areas of two of the quadrilaterals with no common
side equals the sum of the areas of the other pair of such quadrilaterals.
3. Choose n points on a circle and label them with the numbers 1,2,...,n. Say
that two non-neighbouring points A and B are connectable if the points on at
least one of the two arcs containing A and B are all labelled with numbers that are less than
 _ 6 -

those for A and B. Prove that the number of connectable pairs of points is n — 3.
4. Let four distinct points be given in space. Determine all the planes having
the same distance from each of the points.

Now we return to problems from the 1987 I.M.O. (Havana) given in the October 1987
number of the Corner.

France 2. [1987: 246]

Let ABC be a triangle. For each point M of the segment BC denote by B' and C'
the orthogonal projections of M on the lines AC and AB, respectively. Determine those
points Mfor which the length of B' C' is minimum.

Solutions independently by George Evagelopoulos, law student, Athens, Greece, and by

Zun Shan and Edward T.H. Wang, Wilfrid Laurier University, Waterloo, Ontario.
The four points A, C, M
and B' lie on a circle with AM as a
diameter. Using the law of sines it
follows from this that
| B ' C " | - |;4M|sin A
[Let I be the center of the circle. In
triangle IC' B' we have
zC'/fl' =24,
UC'B' = UB'C =90° -A
and so
\IB'\ \AM\
\B'C sin 2 A - - j - 2 sin A cos A.]
sin(90 - A) cos
But this means that the length of B' C' is minimized where ^4Mis as short as possible. This
occurs when AM is an altitude, if angles B and C are both acute. If B (or C) is obtuse the
point Mshould be chosen to be B (C, respectively).

Great Britain 1. [1987: 246]

Prove that if the equation
x4 + ax2 + bx + c - 0
has all its roots real then ab < 0.
Solution I [without the calculus] by Zun Shan and Edward T.H. Wang, Wilfrid Laurier
University, Waterloo, Ontario.
Let p(x) = xA + ax2 + bx + c - 0 and let V(p(x)) denote the number of sign changes
(variations) among the coefficients of p(x). We use Descartes' rule of signs to prove the
 - 7 -

assertion by contradiction. We denote by r*(p), rQ(p)} and r'(p) the number of positive,
zero, and negative real roots of p(x), respectively. Suppose ah > 0. Then there are two
cases:
Case (i). a > 0 and b > 0. There are three subcases.
If c> 0, then V(p(x)) = 0 and V(p(-x)) = 2. Thus
r*(p) = 05 r°(p) = 0 and r~(p) = 2 or 0.
If c = 0, then V{p(x)) = 0, and V(p(-^)) = 1. Thus
r
*(p) = 0? r°(p) = 1 and r"(p) = 1.
If c < 0, then F(p(x)) - 1 and V[p(-x)) = 1. Thus
r+(p) = 1, r°(p) = 0 and r"(p) = 1.
Case (ii). a < 0 and 6 < 0. We simply apply the conclusion of case (i) to
g(a?) = p(-x) = xA + Ax% + Bx + C
where ^4 = —a, B = — 6, C = c.
In both cases p(x) can have at most two real roots, a contradiction.
Solution II [using Rollers Theorem] by Murray S. Klamkin, University of Alberta.
More generally, assume that the equation
rf1 + ai^" 1 + asaf1"3 + a4xn"4 + • • • + an = 0 (1)
has all of its roots real. We show that a\az < 0.
It follows from the above assumption that all the derivatives of the polynomial (1)
have all real zeros. In particular, the equation
x3 + a\x? + a'z = 0,
where a\ and a'z are positive (real) multiples of a\ and a^ respectively, has real roots, rh r2,
r3, say. Then
—a\ = r\ + r2 + r3
0 = r2r3 + ran + rir2
-a' 3 = rxr2n.
If the three roots are 0, then trivially a\a'z - 0. If not, there is some pair of roots whose sum
is not zero. Suppose it is r2 + r$. Then

Hence
a\a\ = (n +r2+ r3) W 3 = (r2 + r3 - ^_pL_) ( - & £ * £ ) .

Hence ai<23 < 0.

 - 8 -

[Editofs note: This problem was also solved by George Evagelopoulos, law student, Athens,
Greece, and by M.A. Selby, Department of Mathematics, University of Windsor. They used
methods similar to the above but did not generalize.]

Great Britain 2. [1987: 246]

Numbers d(n,m), where n, m are integers and 0 < m < n, are defined by
d(n,0) - d(n,n) = 1 for all n > 0
and
m*d(n,m) - m- d(n — l,m) + (2n — m)-d{n — l,ra — 1)
for 0 < m < n. Prove that all the d(n,m) are integers.
Solutions by George Evagelopoulos, law student, Athens, Greece, R.K. Guy,
Department of Mathematics and Statistics, The University of Calgary, and by Zun Shan and
Edward T.H. Wang, Wilfrid Laurier University.

m)(
77 I 77'
- —IT—-- n- are integers, since they
m\{n - m)\ ° ' J are the number

of ways of choosing m things from n. Thus is an integer. We show by induction that

d(n,m) = (*) .

We assume 0 < m < n, and recall that 0! = 1. Now

d(i,o) = i = (J) a = (J) 3 = d(i,i).

So assume 0 < k, and

d(k-l,m)=(k-1)2
for 0 < m < k - 1. Then for 0 < m < k,

m-d(k,m) = m\k-1]2 + (2k-m)[^Z1]]2

»ii)-
Thus
d(k,m) = (£) .
As

d(M) = l = ( J ) 2 = ( * ) a = d ( M )
this completes the induction step and we conclude d(n,m) - \n\ for 0 < m < n.

Editor's note. R.K. Guy asks "Is there a nicer combinatorial proof?"
 - 9 -

Great Britain 3. [1987:247]

Find, with proof, the smallest real number c with the following property: for every
sequence {X{} of positive real numbers such that
Xt+ X2 + ••• + Xn <Xnn for n= 1,2,3,...
we have
4TX + JXi + • . . + ^ < cJXx + X2 + . . . + Xn
for n = 1,2,3,... . [c is to be independent of the X\ and independent of n\

Solutions by George Evagelopoulos, law student, Athens, Greece, by Murray S.

Klamkin, Department of Mathematics, The University of Alberta, and by Zun Shan and
Edward T.H. Wang, Wilfrid Laurier University, Waterloo, Ontario.
The sequence Xi = 2 1 " 1 satisfies
Xx + X2 + ••• + Xn<Xnn (1)
since
1 + 2 + ••• +2 f l " 1 = 2 n - l .
Thus we must have
1 + 2 i ' 2 + 2 2 / 2 + ••- + 2 ( n " 1 ? / 2 = 2%'2 ~ l < c ( 2 n ~ l ) 1 / 2 .
21'2 - 1
This gives
N 1 2* / 2 - 1
c>
V^ - 1 (2 n - l ) 1 ' 2
for all n. Since
,. 2n'2 - 1 ,. 1 - 2"n/2 ,
lim = lim - =1
THOD (2n - l) 1
' 2
THQD I x _ 2 - n

we therefore require
c> — = v ^ + 1.

We show c = \/2 + 1 does have the required property by induction on n.

Let {Xi} satisfy (1) for all n. Certainly JX{ < ( # + 1 ) ^ 7 holds, so the required
inequality holds with n = 1. Now suppose
VC^ + - - - + sPQ < (V^ + l)V^i + ••• + A n .
Then
i/37 + • • • + A ^ < (V5 + 1 ) / ^ + 1 / X ^ I ,
where
F n = Xx + • • • + Xn.
We shall show
(VS + 1 ) / ^ + VX^I < ( ^ + l V F n + * n * i ,
by arguing that, more generally, if 0 < A < B we have
 -10-

(y/2 + l)y[A + JB<(J2 + 1)JA + B.

Squaring, we get the equivalent inequality
(3 + 2JT)A + 2{yj2 + l)JA4B + B < (3 + 2j2)(A + B),
that is,
2(fi + \)JAJB < 2(^5 + 1)5,
which is clear for 0 < A < B (and strict unless A - B). This completes the induction step
and the proof.

Greece 1. [1987: 247]

Consider the regular 1987—gon AiA2...Aig^ with center O. Show that the sum of
vectors belonging to any proper subset of M- {OAy. j - 1,2,...,1987} is nonzero.

Solutions by George Evagelopoulos, law student, Athens, Greece, by Murray S.

Klamkin, Mathematics Department, The University of Alberta, Edmonton, and by Zun Shan
and Edward T.H. Wang, Wilfrid Laurier University, Waterloo, Ontario.
The number 1987 can be replaced by any prime p with the same result. We can
represent the given set of vectors by the p complex numbers
l,U,UJ2,...,U^ (1)
2l l v
where UJ = e * ' satisfies the equation
1 + u+ u2 + ••• + up'x^0. (2)
It is also a known result (which follows by using Eisenstein's criterion) that the cyclotomic
polynomial J[x) - 1 + x + • • • + xp'1 is irreducible over the reals, i.e. there is no real
polynomial of degree less than p — 1 which is satisfied by u.
Our proof is indirect. Assume that there is a proper subset of the ujjs whose sum is
zero. It follows from (2) that the sum of the remaining uj's is also zero. This gives two real
polynomial equations in u (after dividing out the lowest power of u in one of them)
1 + o/*l + J* + . . . + u*r = o
1 + Jl + J* + . . . + Js =0
where the a's and /?'s are natural numbers less than p. One of these equations must be of
degree less than p — 1 which contradicts the fact that / i s irreducible. Thus there is no
proper subset of the uj's whose sum is zero.
Editor's note. Murray Klamkin adds the following comment.
There is an extension of this result to n dimensions but it is less sophisticated.
Consider the n + 1 vectors from the centroid of a regular n-dimensional simplex to its
vertices Ao,...,An. The sum of these n + 1 vectors is zero. However, the sum of any subset
of these vectors
VJ*0 + TJX\+ ••• +~UTr^=rUUt+0 (r<n+l)
where G and Gr are the centroids of the given n-dimensional simplex and the (r — 1)-
 -11-

dimensional simplex with vertices A^A\[Link]^ respectively.

Greece 2. [1987: 247]

Solve the equation
28 x = 19» + 87 2
where x, y, z are integers.

Solution by Dave McDonald, Crimson Elk, Alberta.

Suppose (x,y,z) is a solution.
(i) None of x, y, z are negative, since otherwise there would be a prime
(respectively 2, 19, 3) in the denominator of one term which could not be matched by
another.
(ii) From (i), y > 0, z > 0 imply x > 0 and it is easy to see that x $ 1, so x > 2.
(iii) From (ii), 0 = 3y + V mod 16. Now 7Z = 7 or 1 mod 16 according as zis odd
or even, respectively, while 3 y = 1, 3, 9 or 11 mod 16, according as y = 0, 1, 2 or 3 mod 4. So
2 is odd and y = 2 mod 4.
(iv) Working modulo 9, 1 = 1 + (~3) 2 , so z > 2, and since zis odd, z > 3.
(v) By (iv), 1 = (-8) y mod 27, so 3 divides y, and by (iii), y = 6 mod 12.
(vi) Working modulo 7, 0 E (-2) y + 3*. By (v), y is a multiple of 6, so 3 2 = - 1
mod 7. This gives z = 3 mod 6.
(vii) Now, modulo 13, 2X = 6^ + (-4) 2 . By (v), y is a multiple of 6, so 6y = - 1
mod 13. By (vi), z is a multiple of 3, so (-4)* = 1 mod 13. This gives 2Z = 0 mod 13, a
contradiction!
Thus there are no integer solutions of the equation.

Editor's note. A second correct, but somewhat more involved, solution was
submitted by George Evagelopoulos, law student, Athens, Greece.

Hungary 1. [1987: 247]

Does there exist a set M in the usual Euclidean space such that for any plane a, the
intersection Affl a is finite and non-empty?
Solution by George Evagelopoulos, law student, Athens, Greece, and also by Zun Shan
and Edward T.H. Wang, Wilfrid Laurier University, Waterloo, Ontario.
The answer is yes. Consider the curve
M= {(t,t\tb): i6K}.
If the plane a has equation Ax + By + Cz + D = 0 (not all of A, B, C zero) then the points
of intersection are given by the solutions of At + Btz + Ctb + D = 0 which is a polynomial of
odd degree. Thus there is at least one and there are only finitely many solutions so M n a is
 -12-

fmite and nonempty.

* * *

Next month we continue with solutions to these I.M.O. problems. The Olympiad
season is fast approaching. Please remember to collect the Olympiads available to you and
send them in to me for use in the Corner.
* * *

PROBLEMS
Problem proposals and solutions should be sent to the editor , whose address appears
on the inside front cover of this issue. Proposals should, whenever possible, be accompanied
by a solution, references, and other insights which are likely to be of help to the editor. An
asterisk (*) after a number indicates a problem submitted without a solution.

Original problems are particularly sought. But other interesting problems may also be
acceptable provided they are not too well known and references are given as to their
provenance. Ordinarily, if the originator of a problem can be located, it should not be
submitted by somebody else without his or her permission.

To facilitate their consideration, your solutions, typewritten or neatly handwritten on

signed, separate sheets, should preferably be mailed to the editor before August 1, 1989,
although solutions received after that date will also be considered until the time when a
solution is published.

1401. Proposed by P. Penning, Delft, The Netherlands.

Given are a circle C and two straight lines / and m in the plane of C that
intersect in a point S inside C. Find the tangent(s) to Cintersecting I and m in points P and
Q so that the perimeter of ASPQ is a minimum.

1402. Proposed by George Tsintsifas, Thessaloniki, Greece.

Let M be an interior point of the triangle A\A2Az and Bh B 2 , £3 the feet of
the perpendiculars from Af to sides A2A^ A%Ah AXA2 respectively. Put ri - BiM, i- 1,2,3.
R' is the circumradius of AB{B2B^ and R, r the circumradius and inradius of AA{A2A^.
Prove that
R' Rr> 2rir 2 r 3 .
1403. Proposed by Walther Janous, Ursulinengymnasium, Innsbruck, Austria.
For n > 2, prove or disprove that

n
for all natural numbers £1,2:2,...,zn satisfying
X\ + £2 + • • • + £n = XrX2-...-Xn .
 -13-

1404 Proposed by J.T. Groenman, Arnhem, The Netherlands, and D.J. Smeenk,
Zaltbommel, The Netherlands.
Let ABC be a triangle with circumradius R and inradius p. A theorem of
Poncelet states that there are an infinity of triangles having the same circtimcircle and the
same incircle as AABC.
(a) Show that the orthocenters of these triangles lie on a circle.
(b) If R = 4p, what can be said about the locus of the centers of the nine-point
circles of these triangles?

1 4 0 5 . Proposed by Murray S. Klamkin and Andy Liu, University of Alberta.

Two distinct congruent n—gons P and P ' are inscribed in a noncircular
ellipse E. Prove or disprove that if n > 4, P' must be obtainable from P by a reflection
across the axes or center of E. (For the cases n = 3 and 4 see [1988: 131, 139].)

1 4 0 8 . Proposed by R.S. Luthar, University of Wisconsin Center, Janesville.

If 0 < 0 < 7T, prove without calculus that
cot 0/4 - cot 0 > 2.

1407. Proposed by G.R. Veldkamp, De Bilt, The Netherlands.

Given a rectangle ABCD with AB = CD > AD = BC, construct points X, Y
on CD between C and D such that AX - XY = YB.
1408. Proposed by Jordi Dou, Barcelona, Spain.
Given the equilateral triangle ABC, find all positive real numbers rfor which
there is a point P(r) such that
EA = ER = IP
1 r r2 '
and describe the locus of P(r).
1409. Proposed by Shailesh Shirali, Rishi Valley School, India.
Show that
n , 2n(n - 1) , 3n(n - l ) ( n - 2) _n
F T T + (n + l){n + 2) + (n + lj(n + 2)(n + 3) ^ " °' " 2"
for all positive integers n. What if n > 0 is not an integer?
1410. Proposed by Svetoslav Bilchev and Emilia Velikova, Technical University,
Russe, Bulgaria.
Given is a triangle with circumcentre O and circumradius R. Interior points
P, P' are isogonal conjugates, and rh r2, r3 are the distances from P to the sides of the
triangle. Prove that
(R2 - TJP2)2(R2 - 7JP'2) = 8r{r2r,R\
* * *
 -14-

SOLUTIONS
No problem is ever permanently closed. The editor will always he pleased to consider
for publication new solutions or new insights on past problems.

1 1 9 9 . [1986: 283; 1988: 87] Proposed by D.S. Mitrinovic and J.E. Pecaric,
University of Belgrade, Belgrade, Yugoslavia. (Dedicated to Leo Sauve.)
Prove that for acute triangles,
27i?2
52 < (2jR + T)\
2 2
27R - 8r
where s, r, R are the semiperimeter, inradius, and circumradius, respectively.
II. Comment by Walther Janous, Ursulinengymnasium, Innsbruck, Austria.
Let h\, /12, h$ be the altitudes of an acute triangle and d\, di, d% the distances
from the circumcenter to the sides. Then since
3 3

jQ hi = ^f- and Y[ di = j[s2 ~ (2R + r) 2 ],

1=1 1= 1

the above inequality has a remarkable interpretation, namely that

3 3

i.e.
n^n""
i=i i=i

geometric mean |^>o^WH > geometric mean {^1,^2,^3}.

On the other hand it is known that

arithmetic mean jo-^o^WM i arithmetic mean {^1,^2,^3}

(see e.g. item 12.3 of Bottema et al, Geometric Inequalities). Thus one could pose the
problem: find all exponents t ± 0 such that

holds for all acute triangles, where

Mt(u,v,W) = (»' + f + " ' ) " ' .

*

1 2 9 2 . [1987:320] Proposed by Jack Garfunkel, Flushing, N.Y.

It has been shown (see Crux 1083 [1987: 96]) that if A, B, Care the angles of
a triangle,
 -15-

sin A < B - C\ COS

cos
V2 v3
where the sums are cyclic. Prove that
^ cosf—^\ < — [ Y sin A + T cos ~

which if true would imply the right hand inequality above.

Solution by Walther Janous, Ursulinengymnasium, Innsbruck, Austria.

Proceeding as in the solution of Crux 1083 [1987: 96] we put
7T ~ A
a- , etc.
so that a, /?, 7 are the angles of an acute triangle. Then, using the relations given in Crux
[1982: 67-68],
B - C V
^ _cos(/?-7)
/* .A
= x2 + y2 + 2x - 2
_ cos 2 " ZJ VP / ;
" 2
} cos 4y = N sin a - y,

\ sin A = 2 , s*n 2^ = 2in/,

where x = r/i?, t/ = s/R (r, R, s being the inradius, circumradius, and semiperimeter of the
a — P — 7 triangle). The claimed inequality now reads
x2 + y2 + 2£ - 2 , y + 2xy
2
73
i.e.
j/ 2 ~ — y(l + 2x) + x2 + 2x-2< 0, (1)
V2
where 0 < x < 1/2 and 0 < y < 3VJ/2. In order to show (1) we have to prove that y\ < y < y?,
where
J/J = -1(1 + 2x - V? - 2x + 7)

and
2/2 - ( 1 + 2x + V? - 2x + 7).
VJ
We first show y{ < 0. Indeed,
(1 + 2z)2 = 4x2 + Ax + 1 < z2 - 2x + 7
is equivalent to
a;2 + 2z < 2,
which is true for 0 < x < 1/2.
Hence we have to show y < y% i.e.
yfi<l + 2x+ Jx2 - 2x + 7. (2)
From item 5.4 of Bottema et a], Geometric Inequalities we take the inequality
 -16-

y<2 + ( 3 # - 4 ) x ,
and thus inequality (2) will follow if we prove
2 ft + (9 - 4 ^ ) x < 1 + 2x + V? - 2x + 7,
i.e.
<s) := 2fi - 1 + (7 - 4fi)x < ^ i 2 - 2x + 7 =: r(a;). (4)
Now ^(x) increases, whereas r(x) decreases for 0 < x < 1/2. As furthermore
£(1/2) = 5/2 = r(l/2),
inequality (4) follows.

1 2 9 3 . [1987: 320] Proposed by Steve Maurer, Swarthmore College, Swarthmore,

Pennsylvania and Stanley Rabinowitz, Alliant Computer Systems Corp.,
Littleton, Massachusetts.
Solve the following "twin'1 problems (in both problems, 0 is the center of the
circle and OA±AB).

Figure (a) Figure (b)

(a) In Figure (a), AB = BCand z 4 B C - 60°. Prove CD = OAfi.

(b) In Figure (b), OA = SCand zABC= 30°. Prove CD = ABfi.

Solution by Jordi Dou, Barcelona, Spain.

(a) Let E be on line AB such that AEBD is equilateral. Then CD = EA and EO
(extended) is the perpendicular bisector of BD. Thus [Link] - 30°, so CD = EA- OAJ3.
(b) Let E be the other intersection of AB with the circle, and let F b e on BD so
that EFLBD. Then EF = EB/2 = 4 5 . Since LEBD = 30°, we have lEOD - 60° so that
ED^EO^ OB. Therefore DF = 0 4 - CB, and thus CD = BF = EFfi = 4 5 ^ .

4/50 so/ved 6y S4M BAETHGE, Science Academy, Austin, Texas; S.C CHAN,
Singapore; HANS ENGELHAUPT, Gundelsheim, Federal Republic of Germany; C
 -17-

FESTRAETS-HAMOIR, Brussels, Belgium; RICHARD A. GIBBS, Fort Lewis College,

Durango, Colorado; J.T GROENMAN, Arnhem, The Netherlands; JORG HARTERICH,
Winnenden, Federal Republic of Germany; RICHARD I. HESS, Rancho Palos Verdes,
California; WALTHER JANOUS, Ursulinengymnasium, Innsbruck, Austria; KEE-WAI
LAU, Hong Kong; Z.F. LI, University of Regina; P. PENNING, Delft, The Netherlands; D.J.
SMEENK, Zaltbommel, The Netherlands; DAN SOKOLOWSKY, Williamsburg, Virginia;
GEORGE TSINTSIFAS, Thessaloniki, Greece; C WILDHAGEN, Tilburg University,
Tilburg, The Netherlands; and the proposers.
* * *

1294. [1987: 320] Proposed by P. Penning, Delft, The Netherlands.

Find a necessary and sufficient condition on a convex quadrangle ABCD in
order that there exist a point P (in the same plane as ABCD) such that the areas of the
triangles PAB, PBC, PCD, PDA are equal. C

Solution by Jordi Dou, Barcelona, Spain.

The simplest condition I have found is that
the midpoint of one diagonal lies on the other. The
point P will then be the midpoint of the other B
diagonal. Convexity is not necessary.
If M and N are the midpoints of AC and
BD respectively, the locus of points P such that
area APAB = area APBC is the line BM. This
and analogous results show that if the areas of all four triangles PAB, PBC, PCD, PDA are
to be equal, P must lie on lines BM and DM, and also on lines CN and AN. Thus either M
lies on BD or JVlies on AC.

Also solved by J.T GROENMAN, Arnhem, The Netherlands; GEORGE

TSINTSIFAS, Thessaloniki, Greece; and the proposer. The conditions found appear to be
more or less the same.
* * *

1295. [1987:321] Proposed by J.T Groenman, Arnhem, The Netherlands.

Let A \A2A3 be a triangle with Ih I2, h the excenters and Bh B2, £3 the feet
of the altitudes. Show that the lines I\Bi, I2B2, hB$ concur at a point collinear with the
incenter and circumcenter of the triangle.
 -18-

I. Solution by Shiko Iwata, Gifu, Japan.

{I,r), (0,R), and
(0' ,R') are the centers and radii of
the incircle and circumcircle of
A A \A 2A 3 and the circumcircle of
Alihhj respectively. Then, since /
is the orthocenter of Ahhh and the
circumcircle {0,R) of AAiA2A$ is
the nine—point circle of Ahhh ([1],
page 197), 0 ' , O and / a r e collinear
and R' = 2R. Also,
zO / / 1 i4 3 = 9 0 o - 5 Z / 1 O ' / 2

= 90°-z/3
= £A2I\I= ^A2AzI
= 90°-Zi42i43/i,
/
so i4 2 i4 3 ±0 /i, i.e. ^iBiHO'/i. Let
P be the meeting point of JiBi and l1
10, and let P be on PI\ such that
IP || 0 ' /i. Then we have
PhPO' = IP:0'Ix. (1)
On the other hand,

i4iBi ~ i4i/i T\ T\~ s ~ s

where 5 is the semiperimeter of AA x A 2 Az, so that
n , _ avA1B1^2& _2r
s s '
where A is the area of AAiA2A^ Thus from (1)
PI:PO' = 2 r # ' = r.R.
It follows that P is independent of 7i, so lies on I2B2 and 73B3 too.

Reference:
[1] R.A. Johnson, Advanced Euclidean Geometry, Dover, New York, 1960.

II. Generalization by the proposer.

We prove more generally that if P is any point in the plane of AA\A2Az, Q is
its isogonal conjugate, and P\ is the intersection of A\P and A2A^ [with analogous definitions
for P2, -P3), then the lines P\I\, P2I2, P3/3 are concurrent at a point collinear with Q and the
incenter I of the triangle. The given result follows by letting P be the orthocenter so that Q
is the circumcenter.
We use normal homogeneous triangular coordinates, with
 -19-

At = (1,0,0), A2 = (0,1,0), A3 = (0,0,1),

^= (1,1,1), / s = (1,1,-1), etc.
Let P have coordinates (pi,p2,pi). Then P 3 = (pi,p2,0), so P 3 / 3 has the equation
P2X-piy+ (p2-pi)z= 0.
Analogous equations hold for PJX and P 2 / 2 - Thus to show these lines intersect we have to
prove
Pi ~P\ Pi - P\ j
Pi - Pi Pi -Pi = 0.
-Pi Pi - Pi Pi I
This is easy as the sum of the three rows is zero. The coordinates of the intersection point 5
we get from the equations
P2X-PM+ (P2-Pi)z= 0
(PI-P2)X+ PlV~P2Z= 0
from which comes
S = (P1P2 + PlPl ~ PlPl , P\P2 ~ PlPl + PlPl , ~P\P2 + PlP\ + PlPl)-
Next we can put
Q = (PiPhPiPhPiPi)
so to show that S, Q, and / are collinear we have to prove
P2P3 P1P1 P1P2
P1P2 + P1P1-P2P1 P1P2 - P1P1 + P2P1 -V1P2 + P1P1 + P2P1 I = 0,
1 1 1
or
P2P1 P1P1 P1P2
P1P2 + P1P1 + P1P1 P1P1 + P1P1 + P2P1 P1P2 + P1P1 + P1P1= 0,
1 ,1 1
which is indeed true.
Also solved by C. FESTRAETS-HAMOIR, Brussels, Belgium; HIDETOSI
FUKAGAWA, Yokosuka High School, Aichi, Japan; CLARK KIMBERLING, University of
Evansville, Evansville, Indiana; P. PENNING, Delft, The Netherlands; D.J. SMEENK,
Zaltbommel, The Netherlands; GEORGE TSINTSIFAS, Thessaloniki, Greece; and JOSE
YUSTY PITA, Madrid, Spain.
Kimberling notes that the proposal is a known result (John Casey, Analytic Geometry,
2nd ed., Hodges &: Figgis, Dublin, 1893, p.85).

1 2 9 6 . [1987: 321] Proposed by Murray S. Klamkin, University of Alberta,

Edmonton, Alberta.
Let r\, r2, r3 be the distances from an interior point of a triangle to its sides
 -20-

Cj, a2, 03, respectively, and let R be the circumradius of the triangle. Prove that
flirl + a2fi + azr\ < {2R)n-2a1a2a3
for all n > 1, and determine when equality holds.
Solution by Hidetosi Fukagawa, Yokosuka High School, Aichi, Japan.
Since ri < 2R,
axr\ + a2r^ + a3fi < {axrx + a2r2 + azrz)(2R)n-1
= 2-Aiea(A1A2A3)-(2R)n-1
= £igjp.(2fl)»-i = o1o2fl3(2i2)n-2.

Equality holds when n - 1.

4teo solved by C. FESTRAETS-HAMOIR, Brussels, Belgium; J.T. GROENMAN,

Arnhem, The Netherlands; WALTHER JANOUS, Ursulinengymnasium, Innsbruck, Austria;
GEORGE TSINTSIFAS, Thessaloniki, Greece; and the proposer.
* * *

1 2 9 7 . [1987: 321] Proposed by Walther Janous, Ursulinengymnasium, Innsbruck,

Austria. (To the memory of Leo.)
(a) Let C > 1 be a real number. The sequence £1,%,... of real numbers
satisfies 1 < zn and z\ + • • • + zn < Czn+\ for n > 1. Prove the existence of a constant a > 1
such that zn > an, n > 1.
*
(b) Let conversely z\ < z2 < ... be a strictly increasing sequence of positive real
numbers satisfying zn > an, n> 1, where a > 1 is a constant. Does there necessarily exist a
constant Csuch that z\ + • • • + zn < Czn*\ for all n > 1?
Solution by C. Wildhagen, Tilburg University, Tilburg, The Netherlands.
(a) Let 0 = C"1 so that 0 < 0 < 1. Then
Zn > 0(Zi + Z2+ ••• + Zn-l)
for each n > 2. This gives
2n > 0[*1 + 22 + • • • + Zn-2 + 0(Zi + Z2 + • • ' + ^-2)]
= 0(1+ ff)(Zi + Z2+ • ' • +2n- 2 ),
and using an inductive argument
Zn > 0(1 + 0)h-\zi + Z2 + • • • + **-*)
for each k, 1 < k < n — 1. In particular when fc = n — 1,
^ > fei(l+ <?) n - 2 -D n .A n
where
M
0*i 1 + 0
^ " (1 + V [ l + 0/2J ' A ~ l + °l2'
Take an AT 6 IN such that Dn > 1 for each n> N and hence zn > An for n> N. Since zn > 1
for each n > 1, there exists some ^ > 1 such that zn > /xu for 1 < n < AT. Letting
 -21-

a = min{A,/i}, we have a > 1 and zn > an for all n > 1.

(b) The answer is "no"! Take ^ = 2 and, for each integer k > 2, let
Zi =
I* + (k - i)ifc(Jb + l)J
for each integer i satisfying
k{k-l) <i<k(k + 1).
l
Clearly zi^> Zi >2 for each i. Let
ft

*i

1= 1
for each n € IN. For n - k(k + 1) we have
n -2

5n.2> £ «i

i = M fc-1)

> (2A-l)« k ( k . 1 )
1 1 ?t( Jt+i)
=(2*-!)* + ^
while
I Tfeffc+l)
Zn-l< k+
F-n:
This implies
+ 1 \ fc( fc+l)
fsc2>(2Jfc-l) r A; + 1
+ 1
r Jfe — 1
fc( k+1)
2
(2Jfc-l) 1 -
(k- l)(k + 1)
* + FH;
-» QD ( a s h-* OD).

Hence the sequence (5 n /z n+1 ) is unbounded.

Part (a) also solved by SEUNG-JIN BANG, Seoul, Korea; KEE-WAI LAU, Hong
Kong; and the proposer.

1298. [1987: 321] Proposed by Len Bos, University of Calgary, Calgary, Alberta.
Let A - (aij) be an n * n matrix of positive integers such that | det A\ = 1,
and suppose that z\,Z2,...,zn are complex numbers such that
Z\ %2 ' ' " Zn - 1

for each i = l,2,...,n. Show that z\ = 1 for each f.

 -22-

I. Solution by Seung—Jin Bang, Seoul, Korea.

Note that
Nflil|*|fl""- \zn\ai-=\zVl---z^\^l.
Taking logs, we have the system of linear equations
fliilog|*i| + ••• + ainlog\zn\ = 0 (i= l,...,n)
with coefficient matrix A. Since det A $ 0, we have
log |^i | = ••• = l o g l ^ l =0,
that is,
z1=e^,...,^ =e^ (1)
for some reals 0\,...,0n. From the original equation
fiiifiii ... £in =1

we see that
Q>i\0\ + - • • + ainOn - 2kiKJ i-l,...,n,
for integers k{. Since |det A\ = 1, by Cramer's rule we have Q{ - 2£i?r for some integers l{,
i = l,...,n. Hence by (1) we know that z\- • • • = zn = 1.

II. Solution by C. Wildhagen, Tilburg University, Tilburg, The Netherlands.

We slightly extend the problem by allowing A to have arbitrary integer
entries (provided still that |det A\ = 1).
By applying a finite number of the following elementary row operations:
(i) multiplying a row by —1;
(ii) interchanging two rows;
(iii) adding an integral multiple of one row to another row,
A can be transformed to an upper triangular integral matrix T - (tij) with det T - 1 and
ti i = 1 for all i. Moreover the property (of A) that
4 U 4 ' 2 - - - *£in = i (*= 1,2,...,T0
is preserved by each of (i), (ii), (iii), and therefore
ti\ti2 B . J1 in _ i

Now putting i = n we have £nl = tU2 = • • • = £n n -i = 0, fnn = 1 so that zn = 1; putting t = n — 1

we have
7? . . . 7° r1 , rn-lft - 1l
*l ' m ' *n-2n Zn-l *• ~
so that zn.\-l\ and so on. Continuing in this way we find Z{ - 1 for all i.
Also solved by CHRIS GODSIL, University of Waterloo, and EDWARD T.H. WANG,
Wilfrid Laurier University] WALTHER JANGUS, Ursulinengymnasium, Innsbruck, Austria]
LEROY F. MEYERS, The Ohio State University; M.A. SELBY, University of Windsor, and
the proposer. Two incorrect solutions were received.
Except for Wildhagen's solution II, all solutions received were similar to solution I,
and go through if the entries of A are arbitrary integers. This extension was only pointed out
 -23-

by Wildhagen, however.
*

1 3 0 0 . [1987: 321] Proposed by D.J. Smeenk, Zaltbommel, The Netherlands.

ABC is a triangle, not right angled, with circiimcentre O and orthocentre H.
The line OH intersects CA in K and CB in L, and OK- EL. Calculate angle C.
I. Solution by J. T. Groenman, Arnhem, The Netherlands.
Let a, /?, 7 be the angles and a, 6,
c the sides of A ABC, and let R be the
circumradius. Label the figure as shown, and also
put 0= zCKO. Since
zifCO = z#CL(= 90° - / ? ) ,
we have

sin 0 si B~D mTYKUD

HL ^ CH
sin zHCL sin z CLif
- 2i? cos 7
"siniChk
and thus
sin /.CLH= 2 sin 0 cos 7
= sin(0 + 7) + sin(0 - 7).
But 0+ 7 + zCLff= 180°, so
sin zCLH= sin(0 + 7),
and thus
s i n ( 0 - 7 ) = 0,
i.e. 0= j . Now
tan 7 = tan 0 = tan z / O F
- HNJZOM= 2i? cos a cos 7 - fi cos /?
" AN - A M " c cos a - 6/2
- 3fi cos a cos 7— i? s i n a sin 7
~ 2R sin 7 cos a - J? sin /?
- 3 cos a cos 7 — sin a sin 7
sin 7 cos a — sin a cos 7
- 3 - tan q tan 7
tan 7 — tan a
2
Thus tan 7 = 3, so that 7 = 60° or 120°.

II. Solution by P. Penning, Delft, The Netherlands.

We use triangular coordinates based on the distances to the sides of AABC
Then
 -24-

0(cos i4,cos B,cos C)

and
F(cos B cos C, cos C cos A , cos A cos J3),
so the straight line OH has equation
x y z
cos A cos S cos C = 0.
cos JB COS C COS C COS 4 cos A cos £

Thus the intersection of OH with AC:y - 0 is

if(cos C(cos2,4 - cos2S) , 0 , cos ,4(cos 2 C- cos2B)). (1)
The actual (signed) distance of 0 from BC is X6 = R cos A, of H from BC is
Xh, = 2i? cos B cos C, and of if from BCis (from (1))
2R si n A s i n B sin C - cos C(cos2,4 - cos 2B)
(2)
sin A-cos C(cos 2 ,4 - cos 2 5) + sin C-cos A(cos2C — cos2£?)
which, with quite some manipulation1, equals
X h = 2i?-cos c(cos2A - cos2B)
2 cos A cos C - cos B (3)
The condition OK - HL implies
X\ — Xv - %h — Xi (4)
which with (3) (and Xz - 0) becomes
2 cos C(cos2A - cos25) _ A 0 _ D _ n
T^-———/ " n „~n r>} — cos A = 2 cos S cos G,
2 cos A cos 6 — cos 5 *
i.e.
2 cos C(COS2J4 - cos 2 5) = (2 cos A cos C - c o s B){2 cos £ cos C + cos ,4).
This leads to
cos 4 COS B(4 c o s 2 C - 1) = 0,
so cos C = ±1/2, and thus C = 60° or 120°. (Note that A = 90° or £ = 90° are possible
solutions, but are excluded in the problem.)

III. Editofs comments.

After much manipulation indeed, the editor came up with the following
argument to show (2) equals (3). We need to show
sin A sin B sin C(2 cos A cos C— cos B) = sin A cos C(cos2,4 — cos 2 5)
+ sin Ccos i4(cos2C— cos 2 5)
which can be written
cos A cos Cfsin .4(sin B sin C— cos A) + sin C(sin A sin B — cos C)}
+ cos25(sin A cos C + sin (7 cos ;4) = sin A sin 5 sin Ccos B. (5)

*See III, Editor's comments.

 -25-

Using
cos A = -cos(J9 + C) = sin B sin ( 7 - cos B cos C, etc.
(5) becomes
cos A cos C(sin 4 cos B cos (7+ sin C cos A cos 5)
+ cos2JB(sin A cos C + sin C cos A) = sin A sin B sin C cos £
or
cos J3(cos A cos C + cos J3)(sin A cos (7 + cos A sin C) = sin 4 sin J5 sin Ccos B.
This last equation now follows from
cos B = sin J4 sin C — cos 4 cos C
and
sin ;4 COS C + cos i4 sin C = sin(<A + C) = sin B.
None of the solutions received for this problem were completely satisfactory, in that
they don't appear to work in all cases.
Groenman's proof, for instance, depends on the
given diagram, while Penning's, which seems to be
the most general, uses (at (4)) that point K is
farther from line BC than the circumcentre 0 is.
In fact, if (as is reasonable) one allows the points
K and L to lie on the extended lines CA and CB
respectively, the conclusion of the problem may
not hold! An interesting counterexample is the
triangle ABC illustrated at the right, where D is
the foot of the perpendicular from B, and DC = 1,
AD = 2, BD = 3. The reader can check that OK =
HL, while of course LC $ 60 .

Also solved by JORDIDOU, Barcelona, Spain; C. FESTRAETS-HAMOIR, Brussels,

Belgium; GEORGE TSINTSIFAS, Thessaloniki, Greece; JOSE YUSTY PITA, Madrid,
Spain; and the proposer.

1 8 0 2 . [1988: 12] Proposed by Mihaly Bencze, Brasov, Romania.

n

Suppose a& > 0 for k = 1,2,...,n and \ t a n h 2 ^ = 1. Prove that

k=l

sin ait
sinh
Y
la
}
sinh ak
> n Y
- Za cosh
^
h= i k=i
 -26-

Solution by Vedula N. Murty, Pennsylvania State University at Harrisburg.

We may without loss of generality assume that 0 < ot\ < • • • < otn. Let
1 ... = tanh
w\ . „ ! 2*aj; , k- 1,2,...,n.
Then it is easily seen that
e
Xi > X2 > • • > Xn

and

i.e. the correlation coefficient between the afs and w's is < 0. Therefore by Chebyshev's
inequality
n n
n 1 1

L X k IfcI"*
1 X""'l

k=l , fc = i =i J)
Since

Z J ^ k ~ \J tanh 2 a^ = 1
k=l fc = 1
is given, and
tanh2^?; _ sinh a^
sinh a* ~ cosh 2 a^
the required inequality follows.
Also solved by SEUNG-JIN BANG, Seoul, Korea] C FESTRAETS-HAMOIR,
Brussels, Belgium] JORG HARTERICH, Winnenden, Federal Republic of Germany]
RICHARD L HESS, Rancho Palos Verdes, California] WALTHER JANOUS,
Ursulinengymnasium, Innsbruck, Austria] MURRA Y S. KLAMKIN, University of Alberta]
KEE-WAI LAU, Hong Kong] COLIN SPRINGER, student, University of Waterloo] C.
WILDHAGEN, Breda, The Netherlands] and the proposer.

As in Crux 1288 [1988: 312], about half the solvers simply applied Chebyshev7s
inequality.

1 3 0 3 . [1988: 12] Proposed by George Tsintsifas, Thessaloniki, Greece.

Let ABC and A\B\C\ be two triangles with sides a, b, c and a\, b\, c\ and
inradii r and r\, and let P be an interior point of A ABC. Set AP = x, BP = y, CP = z. Prove
that
aix2 +
Jiyl l
+ ClZ
' > 4m.
a + b+ c -
I. Solution by Walther Janous, Ursulinengymnasium, Innsbruck, Austria.
First proof Let F and F\ be the areas of triangles ABC and AiBiC\,
 -27-

respectively. The given inequality can then be written as

axx2 + bty2 + ctz2 > 8Frh (1)
In [1] (see also item 12.56 of [3]) is shown the inequality
(atx + hty + az)2 > y + 8FFh (2)

where
M= ] T a](b2 + c2 - a2),
the sum being cyclic. Moreover, in [2] inequality (2) is extended to all points P of the space.
Applying the Cauchy—Schwarz inequality to (2) we get

1 X flH - I5J ai) I X aix2 = 2si

) X aix2' ^
Furthermore, we take from item 10.8 of [3] the f!Neuberg—Pedoe" inequality
M>WFFh (4)
Finally (2), (3) and (4) lead to the better estimation

X ai7? - 3TI + 4Fri - 8Frh

the inequality holding for all points of the space.

Second proof. The polar moment of inertia inequality [4] states: if u}v,w > 0 then
(u + v + w)(ux2 + vy2 + wz2) > a2vw + h2wu + c2uv. (5)
Furthermore, from Crux 1181(a) [1988: 25] the inequality
a2vw + b2wu + c2uv > £FJuvw(u + v + w) (6)
is known. (5) and (6) yield
UVW /?x
ux2 + vy2 + wz2 > 4F
+ v + w'
T^
Putting in (7) u- a^ v- b^ w- C\ and noting ai&iCi = 4Fii?i we get
a^ 2 + iii/2 + c{z2 > 4FIfSS = 4 F ^ 7 ^ . (8)
i 2si
As R\ > 2ri we obtain from (8) the desired inequality (1). Note that (8) gives another
interpolation of (1). Again (as can be seen from [4]) there are no restrictions on the position
of the point P.
References:
[1] 0 . Bottema and M.S. Klamkin, Joint triangle inequalities, Simon Stevin 48, I—II
(1974) 3-8.
[2] D.S. Mitrinovic and J.E. Pecaric, Note on 0 . Bottema's inequality for two triangles,
C.R. Math. Rep. Acad. Sci. Canada 8, No. 2 (1986) 141-144.
[3] 0 . Bottema et al, Geometric Inequalities, Groningen, 1968.
[4] M.S. Klamkin, Geometric inequalities via the polar moment of inertia, Math.
Magazine 48 (1975) 44-46.
 -28-

II. Solution by Murray S. Klamkin, University of Alberta.

First proof. The given result will follow by successive use of some known
stronger inequalities. First we use the polar moment of inertia inequality [1]
(W\ + W2+ Wz)(W\X2 + W2y2 + W$Z2) > W2Wza2 + W$Wib2 + W\V)2C2
where W\, v)2y w$ are arbitrary real numbers. (A simple proof follows by expanding out
(wiX+ w2Y + w3Z)2> 0,
where X, Y, Z are vectors from P to J4, B, C, respectively.) Letting (wuW2,w$) = (tti,6i,ci),
we get
a\X* + bit/2 + C\z2 y bjCia2 + ciQ,\b2 + a\b\e2
a + b+ c ~~- (a + b + c)(ai + bx + c\)
_{b\C\a2 + c\a\b2 + a\b\c2)rr\
=' —WFl '
It thus suffices to show that
biCia2 + cxaib2 + aib{c2 > 16FFh (9)
or equivalently,
l+V+<t>$F. (10)
To prove (10), we use the known stronger inequality ([2], eq. 41)
aH. + l + lyWJKfi (n)
fl262C2
al b\ cV ' ^ '
where fl2, &2, ^2, ^2 are the sides and area of a third triangle. In (11), let

to give
a* 62 + c^ > 8/ygffi
ai 61 ci -
\Ja\b\c
It now remains to show that
2 ^ # , 1

or
±R\F2fi > aifiici = 4Fii2 h
or
RIFT>/S > *V (12)
But this follows from the Finsler-Hadwiger inequality ([3], item 10.3)
4*1 > F1}/5,
and the known inequality ([3], item 4.14)
3R\y/S > 4 F h
which is equivalent to the fact that the largest triangle (in area) that can be inscribed in a
circle is the equilateral one.
 -29-

An inequality similar to (9), and due to the proposer, appears as problem E3154 of
the Amer. Math. Monthly (solution in Vol. 95 (1988), pp.659-660). Here one was to show
that
btcta2 + aaib2 + athtc2 > 4F 2
where the triangle AiBiCi is inscribed in A ABC.
References:
[1] M.S. Klamkin, Geometric inequalities via the polar moment of inertia, Math.
Magazine 48 (1975) 44-46.
[2] M.S. Klamkin, Asymmetric inequalities, Publ. Electrotehn. Fur. Ser. Mat Fiz. Univ.
BeogradNo. 357-380 (1971) 33-44.
[3] O. Bottema et al, Geometric Inequalities, Wolters-Noordhoff, Groningen, 1968.
[Editofs note. Klamkin gave a second solution, somewhat like Janous' first proof
above. References to lines in this proof have been added by the editor.]
Second proof. Let r > 1. We start out with the power mean inequality
axxT + biyr + CjZr > la{x + b\y + CJZV
ax + b\ + c\ ~ I a\ + b\ + ct J
Then using (2) and (4),
aixr + b\yr + c\zr > {a\x + bxy + ctz)r
a b c
+ + "" 25(25l)r"1

N (4jFFj)r _2r(FFx)r'2
'ss^-^r ssV1
The given inequality corresponds to the case r = 2.

Also solved by SVETOSLAV J. BILCHEV and EMILIA A. VELIKOVA, Technical

University, Russe, Bulgaria; VEDULA N. MURTY, Pennsylvania State University at
Harrisburg; COLIN SPRINGER, student, University of Waterloo; G.R. VELDKAMP, De
Bilt, The Netherlands; and the proposer.
Bilchev and Velikova also obtained the stronger inequality (8) in Janous7 second
solution.
* * *

1 3 0 4 . [1988: 12] Proposed by M.S. Klamkin, University of Alberta, Edmonton,

Alberta.
If p, q, r are the real roots of
xz-6x2 + 3x + 1 = 0,
determine the possible values of
p2q + q2r + r2p
 -30-

and write them in a simple form.

Solution by Sam Baethge, Science Academy, Austin, Texas.
Let
A - p2q + q2r + r^p , B - p2r + q2p + r2g,
the only two possible values of expressions of the given type. We also have
p + q+ r= 6,
pq + qr + rp = 3,
pqr = —1.
In the equations that follow, all summations are symmetric over p, q and r.
(i) 18 = (p + q + r){pq + qr + rp) = V p2g + 3pqr = V p2g - 3

or
4 4- 5 = Y p2g = 21.

(ii) 216 = (p + q + r) 3 = Y p 3 + 3 Y p2g + 6pgr

or
Y p 3 = 216 - 3(21) - 6(-l) - 159.

(iii) 27 = {pq + qr + rp) 3 = Y p V + 3 Y p2q2r + Qp2q2r

or
Y p V = 27 - 6(1) - 3pgr Y p2q = 21 - 3(-l)(21) - 84.

(iv) J4B = Y p4qr + Y p3g3 + 3p2g2r2 = pqr Y p 3 + 84 + 3

= (—1)(159) + 87 = -72.
Using (i) and (iv), A and B are the roots of
y2-21y-72 = 0,
so the possible values are 24 and —3.

Also solved by FRANCISCO BELLOT ROSADO, Emilio Ferrari High School, and
MARIA ASCENSION LOPEZ CHAMORRO, Leopoldo Cano High School, Valladolid, Spain]
HANS ENGELHAUPT, Gundelsheim, Federal Republic of Germany; C FESTRAETS-
HAMOIR, Brussels, Belgium; JACK GARFUNKEL, Flushing, N.Y.; RICHARD L HESS,
Rancho Palos Verdes, California; ERIC HOLLEMAN, student, Memorial University of
Newfoundland; WALTHER JANOUS, Ursulinengymnasium, Innsbruck, Austria; SIDNEY
KRAVITZ, Dover, New Jersey; KEE-WAI LAU, Hong Kong; P. PENNING, Delft, The
Netherlands; COLIN SPRINGER, student, University of Waterloo; G.R. VELDKAMP, De
Bilt, The Netherlands; C WILDHAGEN, Breda, The Netherlands; and the proposer. There
were two partial solutions submitted.
 -31-

Some solvers, including the proposer, answered the problem for an arbitrary cubic.
* * *

1 3 0 5 . [1988: 12] Proposed by J. T. Groenman, Arnhem, The Netherlands.

Let AiA2A3 be an acute triangle with circumcenter O. Let Pu Qx (Q\f At)
denote the intersection of AxO with A2A3 and with the circumcircle, respectively, and define
P2, Q2, P3, <?3 analogously. Prove that

(a) VFl-VFT-VFj
>i;
~PTQi-T7<72~-PW3
TJF1 TJP~. VFx >3;
(b) + r
7T$i T 2~m TTQi
fc\ A\P\- A2P2-A3Pz > 27
TWi-TW*-P!Qi~
Solution by Colin Springer, student, University of Waterloo
We write | OP\\ for OP\ etc. and also
write I T\ for the area of the triangle T.
Let
«i = /LAIOAI,
a2 = LA-),OA\,
az = 1A1OA2,
and, without loss of generality,
\OAt\ = \OA2\ = \OAz\ = 1 .
Then, since AA\A2A3 is acute,
\OPA-- OPx ___AA2AiAi
<AOAlA2 I + I
sinai
sin #3 + sm a2'
with similar expressions for j OP2\ and | OP3\. Thus
11PM
1Vi| ^ l ^ - ll O P l]i l ^ 8 1 1 1 sm
a
l + as i +; asm
? + as i n a
\ ' etc.
2 3

Let
x - —sin ai + sin a 2 + sin as,
y - sin ai — sin a 2 + sin a3,
z = sin ai + sin a 2 — sin a3;
then
 -32

\0P\ QP2l-|OP3;.
TTQi P2Q2\-\PzQz\ \hi
(—sinai+sina2+sino!3)(sina i-sina2+si na3)(sin^!+sina2~sina3)
(y + z)(z + x)(x + y)
Sxyz

-\1L + Z (z_±_x\ (x + y\ >

by the A.M.—G.M. inequality. This is (a).

For (b),
OPx\ OP? I OP- DP^ ~UFl
+ + >3
Ira P 2 Q 2 | - \PzQz\
>3

by the A.M.—G.M. inequality and part (a).

Finally, since
sin a
1\A,P,\
1 1 1 =1+ \OPt\
> 1! = l + [Link] i +
sin ? asin
? + as i n a3
, ' etc., '
3

we have
U1P1 • ^2^2 • ^ 3 P 3 |
\P1Q1 • P2Q2 • ^aQal
^ sinQfi+sina2+sina3 f___T__^
(—sinai+sina2+sina3)(sinai—Sina2+sina3)(sinai+sina2~sina3)
> ^ ^ 27sinarsina2'Sinof3 ^
~ (~sinai+sina2+sina3)(sinai-sina2 + sina3)(sinai+sina2--sina3)
> 27,
again by the A.M.—G.M. inequality and part (a).

Also solved by C. FESTRAETS-HAMOIR, Brussels, Belgium; JACK GARFUNKEL,

Flushing, N.Y.; JORG HARTERICH, Winnenden, Federal Republic of Germany;
WALTHER JANOUS, Ursulinengymnasium, Innsbruck, Austria; MURRAY S. KLAMKIN,
University of Alberta; VEDULA N. MURTY, Pennsylvania State University at Harrisburg;
D.J. SMEENK, Zaltbommel, The Netherlands; G.R. VELDKAMP, De Bilt, The Netherlands;
and the proposer.

Janous' solution to part (c) reduced it to Crux 1199 [1988: 81].

* * *
 CMS SUBSCRIPTION PUBLICATIONS

1989 RATES

CANADIAN JOURNAL OF MATHEMATICS

Editor-in-Chief; D. Dawson and V. Dlab

This internationally renowned journal is the companion publication to the
Canadian Mathematical Bulletin. It publishes the most up-to-date research in
the field of mathematics, normally publishing articles exceeding 15 typed
pages. Bimonthly, 256 pages per issue.
Non-CMS Members $250.00 CMS Members $125.00
Non-CMS Members obtain a 10% discount if they also subscribe to the Canadian
Mathematical Bulletin. Both subscription must be placed together.

CANADIAN MATHEMATICAL BULLETIN

Editors^ J. Fournier and D. Sjerve

This internationally renowned journal is the companion publication to the
Canadian Journal of Mathematics. It publishes the most up-to-date research in
the field of mathematics, normally publishing articles no longer than 15 pages.
Quarterly, 128 pages per issue.
Non-CMS Members $120.00 CMS Members $60.00
Non-CMS Members obtain a 10% discount if they also subscribe to the Canadian
Journal of Mathematics. Both subscriptions must be placed together.

Orders by CMS members and applications for CMS membership

should be submitted using the form of the following page.

Orders by non-CMS Members for the

CANADIAN MATHEMATICAL BULLETIN and the
CANADIAN JOURNAL OF MATHEMATICS should be
submitted using the form below:

Order Form I

Canadian Mathematical Society • Please enter my subscription to both the CJM and CM B I

<j> Societe Mathematique du Canada Q

•
(Regular institutional rate $250 + $120, combined discount rate $333)
Please enter my subscription to the CJM only
Institutional rate $250
Please enter my subscription to the CMB only
I
I
I
I
Institutional rate $120 I
D Please bill me I
D I am using a credit card I
D 1 enclose a cheque made payable to the University of Toronto Press J
D Send me a free sample of [ | CJM Q CMB I

Visa / Bank Americard / Barclaycard j

DDDD DDD DDD DDD

MasterCard / Access / Interbank 4-digit bank no. 1

Inquiries and order:

University of Toronto Press
DDDD DDDD •••• DDDD DDDD
Journals Department, 5201 Dufferin St.
Downsview, Ontario M3H 5T8 Expiry date Signature
 CMS SUBSCRIPTION PUBLICATIONS

1989 RATES

CRUX MATHEMATICORUM

Editor: W. Sands
Problem solving journal at the senior secondary and university undergraduate
levels. Includes "Olympiad Corner" which is particularly applicable to
students preparing for senior contests.
10 issue per year. 36 pages per issues.
Non-CMS Members: $35.00 CMS Members: $17.50

CMS NOTES

Editors: E.R. Williams and P.P. Narayanaswami

Primary organ for the dissemination of information to the members of the C.M.S.
The Problems and Solutions section formerly published in the Canadian
Mathematical Bulletin is now published in the CMS Notes.
8-9 issues per year.
Non-CMS Members: $10.00 CMS Members FREE

Orders by CMS members and applications for CMS Membership

should be submitted using the form on the following page.

Orders by non-CMS members for

CRUX MATHEMATICORUM or the CMS NOTES
should be submitted using the form below:

Order Form

Please enter subscriptions:

dtfP^HX^ Canadic
Canadian Mathematical Society • Crux Mathematicorum ($35.00)
Mathematique du Canada D C.M.S. Notes ($10.00)

Lj Please bill me
•~l I am using a credit card
D I enclose a cheque made payable to the Canadian Mathematical Society

Visa

DDDD DDD DDD DDD

MasterCard

Inquiries and order:

Canadian Mathematical Society
DDDD DDDD •••• DDDD DDDL^
577 King Edward
Ottawa, Ontario KIN 6N5 Expiry date Signature
 Canadian Mathematical Society 577 KING EDWARD, OTTAWA. ©NT
CANADA K1N 8N5
Swsete Mathematique dM Canada
< ! >
1989 MEMBERSHIP APPLICATION FORM 1989
(Membership period: January 1 to December 31)

CATEGORY DETAILS
1 students and unemployed members $ 15 per year
2 retired professors, postdoctoral fellows, necondary
I junior college teachers s 25 per year
3 members with salaries under $30,000 per year $ 45 per year
4 members with salaries from $30,000 - $60,000 $ 60 per year
5 members with salaries of $60,000 and more $ 75 per year
10 Lifetime membership for members under age 60 $ 1000 i [ i l i)
15 Lifetime membership for members age 60 or older $ 500

<i) Hembers of the AHS and/or MAA WHO RESIDE OUTSIDE CANADA are eligible for a 15% reduction in the basic
membership fee*

Cii) Members of the Allahabad, Australian, Brazilian, Calcutta, French, German, Hong Kong, Italian, London, Mexican, Polish
of New Zealand mathematical societies, WHO RESIDE OUTSIDE CANADA are eligible for a 50% reduction in basic membership
fee for categories 3,4 and 5.

(iii) Payment may be made in two equal annual installments of $500

APPLIED MATHEMATICS NOTES: Reduced rate for members $6.00 (Regular $12.00)
CANADIAN JOURNAL OF MATHEMATICS: Reduced rate for members $125.00 (Regular $250.00)
CANADIAN MATHEMATICAL BULLETIN: Reduced rate for members $60.00 (Regular $120.00)
CRUX MATHEMATICORUM: Reduced rate for members $17.50 (Regular $35.00)

FAMILY NAME FIRST NAME INITIAL TITLE

MAILING ADDRESS CITY

PROVINCE/STATE COUNTRY POSTAL CODE TELEPHONE ELECTRONIC MAIL

PRESENT EMPLOYER POSITION

HIGHEST DEGREE OBTAINED GRANTING UNIVERSITY YEAR

PRIMARY FIELD OF INTEREST (see list on reverse) MEMBER OF OTHER SOCIETIES (See (i) and (ii))

Membership ne renenaI CATEGORY RECEIPT NO. _________

#
Basic membership fees (as per table above)
* Contribution towards the Work of the CMS
Publications requested
Applied Mathematics Notes C$ 6.00)
Canadian Journal of Mathematics ($125.00)
Canadian Mathematical Bulletin ($ 60.00)
Crux Mathematicorum ($ 17.50)

TOTAL REMITTANCE: %

CHEQUE ENCLOSED (MAKE PAYABLE TO CANADIAN MATHEMATICAL SOCIETY) - CANADIAN CURRENCY PLEASE

PLEASE CHARGE VISA MASTERCARD

ACCOUNT NO. EXPIRY DATE

SIGNATURE
L _2_
BUSINESS TELEPHONE NUMBER

(*> INCOME TAX RECEIPTS ARE ISSUE® TO ALL MEMBERS FOR MEMBERSHIP FEES AND CONTRI BUT IONS ONLY
MEMBERSHIP FEES AND CONTRIBUTIONS MAT BE CLAIMED ON TOUR CANADIAN TAX RETURN AS CHARITABLE DONATIONS
SI
y ^ ' " ^ M<H**nwHc*l ZonetY S77 KING EDWARD . OTTAWA, ONT
£ « , , , « - . j ^. i CANADA KIN 8N5
r^ Societe Mathemattque au i annua
1989 FORMULAIRE P ' A D H E S I O N 1989

CLa cotisation est pour l'am£e civile: 1 Janvier - 31 d^cembre)

CATEGORIES DETAILS COTISATION

1 itudiants et chomeurs 15$ par annee
2 professeurs a la retraite, boursiers postdoctoraux, 25$ par annee
enseignants des £coles secondaires et des colleges
3 revenu annuel brut moins de 30,000$ 45$ par annee
4 revenu annuel brut 30,000$ - 60,000$ 60$ par annee
5 revenu annuel brut plus de 60,000$ 75$ par annee
10 Membre a vie pour membres ages de moins de 60 ans 1000$ (iii)
15 500$
Membre a vie pour membres ag6s de 60 ans et plus
(i) La cotisation des membres de I'AMS et MAA est requite de 15% SI CEUX-CI NE RESIDENT PAS AU CANADA.

Cii) Suivant I'accord de reciprocity, la cotisation des membres des categories 3, 4 et 5 des societes
suivantes: Allahabad, Allemagne, Australie, Br£sil, Calcutta, France, Londre, Mexique, Nouvelle
Zllande, Pologne, Italie, Hong Kong, est nkiuite de 50% SI CEUX-CI NE RESIDENT PAS AU CANADA.

(iii) Les frais peuvent etre r6gl£s en deux versements annuels de 500,00$

NOTES DE MATHEMATIQUES APPLIOUEES: Abonnement des membres 6$ (R6gulier 12$)

JOURNAL CANADIEN DE MATHEMATIQUES: Abonnement des membres 125$ (Regulier 250$)
BULLETIN CANADIEN DE MATHEMATIQUES: Abonnement des membres 60$ (Regulier 120$)
CRUX MATHEMATICORUM: Abonnement des membres 17,50$ (Regulier 35$)

NOM DE FAMILLE PRENOM INITIALE TITRE

ADRESSE DU COURRIER VILLE

PROVINCE/ETAT PAYS CODE POSTAL TELEPHONE

ADRESSE ELECTRONIQUE EMPLOYEUR ACTUEL POSTE

DIPL0ME LE PLUS ELEVE UNIVERSITE ANNEE

DOMAINE D'INTERET PRINCIPAL (svp voir liste au verso) MEMBRE D'AUTRE SOCIETE (Voir (i) et (ii))

Membre nouveau

•
* Cotisation (voir table plus haut)
renouvellement
n CATEGORIE: NO. DE RECU:

* Don pour les activites de la Societe

Abonnements desires:
Notes de mathematiques appliquees (6.00$)
Journal canadien de mathematiques (125.00$)
Bulletin canadien de mathematiques (60.00$)
Crux Mathematicorum (17.50$)

TOTAL DE VOTRE REMISE:

CHEQUE INCLUS (PAYABLE A LA SOCIETE MATHEMATIQUE DU CANADA) - EN DEVISES CANADIENNES S.V.P.

PORTER A MON COMPTE j | VISA f I MASTERCARD \ \

NUMERO DE COMPTE DATE D'EXPIRATION

)
SIGNATURE TELEPHONE D'AFFAIRE

(*) UN RECU POUR FIN D'IMPuT SERA EMIS A TOUS LES MEMBRES POUR LES DONS ET LES COTISATIONS SEULEMENT
LES FRAIS D'AFFILIATION ET LES DONS SONT DEDUCTIBLES DMMPOT A CONDITION TOUTEFOIS D'ETRE INSCRITS DANS
LA RUBRIQUE "DON DE CHARITE" DES FORMULAIRES DMMPOT FEDERAL
 !!!!! BOUND VOLUMES !!!!!
THE FOLLOWING BOUND VOLUMES OF CRUX MATHEMATICORUM
ARE AVAILABLE AT $ 10.00 PER VOLUME

1 & 2 (combined), 3, 4, 7, 8, 9 and 10

PLEASE SEND CHEQUES MADE PAYABLE TO

THE CANADIAN MATHEMATICAL SOCIETY

The Canadian Mathematical Society

577 King Edward
Ottawa, Ontario
Canada KIN 6N5
Volume Numbers Mailing:
Address
volumes X $10.00 = $

!!!!! VOLUMES RELIES !!!!!

CHACUN DES VOLUMES RELIES SUIVANTS A 10$:

1 & 2 (ensemble), 3, 4, 7, 8, 9 et 10

S.V.P. COMPLETER ET RETOURNER, AVEC VOTRE REMISE LIBELLEE

AU NOM DE LA SOCIETE MATHEMATIQUE DU CANADA, A L'ADRESSE SUIVANTE:

Societe mathematique du Canada

577 King Edward
Ottawa, Ontario
Canada KIN 6N5

Volumes: Adresse:

volumes X 10$ = $
 PUBLIC A TIONS

The Canadian Mathematical Society

577 King Edward, Ottawa, Ontario KIN 6N5
is pleased to announce the availability of the following publications:

1001 Problems in High School Mathematics

Collected and edited by E.J. Barbeau, M.S. Klamkin and W.O.J. Moser.

Book I: Problems 1-100 and Solutions 1-50 58 pages ($5.00)

Book II: Problems 51-200 and Solutions 51-150 85 pages ($5.00)
Book III: Problems 151-300 and Solutions 151-350 95 pages ($5.00)
Book IV: Problems 251-400 and Solutions 251-350 115 pages ($5.00)
Book V: Problems 351-500 and Solutions 351-450 86 pages ($5.00)

The Canadian Mathematics Olympiads (1968-1978)

Problems set in the first ten Olympiads (1969-1978) together with suggested
solutions. Edited by E.J. Barbeau and W.O.J. Moser. 89 pages ($5.00)

The Canadian Mathematics Olympiads (1979-1985)

Problems set in the Olympiads (1979-1985) together with suggested solutions.

Edited by C M . Reis and S.Z. Ditor. 84 pages ($5.00)

Prices are in Canadian dollars and include handling charges.

Information on other CMS publications can be obtained by writing
to the Executive Director at the address given above.