1

SKOLIAD

No. 114

Va lav

Linek

Please send your solutions to problems in this Skoliad by August 1, 2009.

Solutions should be sent to Lily Yen and Mogens Hansen at the address inside
the ba k over. The Skoliad se tion is in transition and, unfortunately, we
have lost several of the submitted solutions to past ontests. If you have
opies of solutions that you sent to past ontests, please send them again
so that we an mention any orre t solutions we re eive. (This in ludes any
ontest in Skoliad appearing in or after the Mar h 2008 issue of CRUX).
Our rst problem set of the year is the Math Kangaroo Contest Pra ti e
Set. The Kangaroo Contest is international in s ope and supported in Canada
by the Canadian Mathemati al So iety and the Institute of Ele tri al and
Ele troni s Engineers (Northern Se tion).
Our thanks go to Valeria Pandelieva, the Canadian representative of the
Kangaroo Contest, for bringing this ontest to our attention, and for making
us aware of the need for ontests and math-parti ipation in the lower years
in Canada. For that reason, and also sin e this ontest is straightforward
to administer (see www.mathkangaroocanada.com), we are featuring its entire
range of questions over all grades.
Finally, while it is a multiple hoi e test, we ask our readers to send in
omplete solutions showing all the steps and details so that we an evaluate
the solutions and give full redit to the solvers.
Math Kangaroo Contest
Pra ti e Set

Part A (3 points per question)

. (Grades 3-4) In the addition example, ea h letter

represents a digit. Equal digits are represented by the
same letter. Dierent digits are represented by dierent letters. Whi h digit does the letter K represent?
(A) 0
(B) 1
(C) 2
(D) 8
1

+
W

O
K
O

K
O
W

(E) 9

. (Grades 5-6) Ten aterpillars, arranged in a row one behind another,

walked in the park. The length of ea h aterpillar was equal to 8 m, and
the distan e any two adja ent aterpillars kept for safety reasons was 2 m.
What is the total length of their row?
(A) 100 m (B) 98 m
(C) 82 m
(D) 102 m (E) 96 m
2

2
. (Grades 7-8) An ant is running along a ruler ..................................sq...q.r................................................
...
of length 10 m with a onstant speed of 1 m ...
.
1
2
3
4
5
6
7
8
9
.
per se ond (see the gure). Any time when .......................................................................................
the ant rea hes one of the ends of the ruler, it
turns ba k and runs in the opposite dire tion. It takes the ant exa tly 1 se ond to make a turn. The ant starts from the left end of the ruler. Nearest
whi h number will it be after 2009 se onds?
(A) 1 m
(B) 2 m
(C) 3 m
(D) 4 m
(E) 5 m
3

.
....
.

.
....
.

.
....
.

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

.
....
.

.
....
.

.
....
.

.
....
.

. (Grades 9-10) Whi h of the numbers 26 , 35 , 44 , 53 , 62 is the greatest?

(A) 26
(B) 35
(C) 44
(D) 53
(E) 62

. (Grades 11-12) A de orator has prepared a mixed paint, in whi h the

volumes of red and yellow olours were in the ratio 2 : 3. The resulting
olour seemed too light to him, so he added 2 L of red paint. This way, the
ratio of the volumes of the red and yellow olours hanged to 3 : 2. How
many litres of paint did the de orator use?
(B) 6 L
(C) 7 L
(D) 8 L
(E) 9 L
(A) 5 L
5

Part B (4 points per question)

6. (Grades 3-4) Two boys are playing tennis until one of them wins four
times. A tennis mat h annot end in a draw. What is the greatest number of
games they an play?
(B) 7
(C) 6
(D) 5
(E) 9
(A) 8
7. (Grades 5-6) In two years, my son will be twi e as old as he was two years
ago. In three years, my daughter will be three times as old as she was three
years ago. Whi h of the following best des ribes the ages of the daughter and
the son?
(A) The son is older;
(B) The daughter is older;
(C) They are twins;
(D) The son is twi e as old as the daughter;
(E) The daughter is twi e as old as the son.

. (Grades 7-8) Some points are marked on a straight line so that all distan es

1 m, 2 m, 3 m, 4 m, 5 m, 6 m, 7 m, and 9 m are among the distan es

between these points. At least how many points are marked on the line?
(A) 4
(B) 5
(C) 6
(D) 7
(E) 8

. (Grades 9-10) Eva, Betty, Linda, and Cathy went to the inema. Sin e it
was not possible to buy four seats next to ea h other, they bought ti kets for
seats number 7 and 8 in the 10th row and ti kets for seats number 3 and 4
in the 12th row. How many seating arrangements an they hoose from, if
Cathy does not want to sit next to Betty?
(A) 24
(B) 20
(C) 16
(D) 12
(E) 8
9

3
10.
(Grades 11-12) Triangle ABC is
isos eles with BC = AC . The segments
DE , F G, HI , KL, M N , OP , and XY
divide the sides AC and CB into equal
parts. Find XY , if AB = 40 m.
(A) 38 m
(B) 35 m
(C) 33 m
(D) 30 m
(E) 27 m

C
...
... ....

D ..................................... E

F .......................................................... G
H ............................................................................. I
K .................................................................................................. L
M ....................................................................................................................... N
O ............................................................................................................................................ P

X ................................................................................................................................................................ Y
A ........................................................................................................................................................................... B

Part C (5 points per question)

. (Grades 3-4) Matt and Ni k onstru ted

two buildings, shown in the gures, using identi al ubes. Matt's building weighs 200 g, and
Ni k's building weighs 600 g. How many ubes
from Ni k's building are hidden and annot be
seen in the gure?
(B) 2
(C) 3
(A) 1
(D) 4
(E) 5
11

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

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

Ni k's
building

Matt's
building

12. (Grades 5-6) Consider all four-digit numbers divisible by 6 whose digits
are in in reasing order, from left to right. What is the hundreds digit of the
largest su h number?
(A) 7
(B) 6
(C) 5
(D) 4
(E) 3

. (Grades 7-8) A square of side length 3 is divided

by several segments into polygons as shown in the gure. What per ent of the area of the original square is
the area of the shaded gure?
(A) 30%
(B) 33 13 %
(C) 35%
(E) 50%
(D) 40%
13

.........1..........................2.................
.. ...................................................................................................................... ..... 1
. ........................................................................................................ .
2 .. ............................................................................................................................................................................................................................................................................................................................................................................................................................................................................................
.. ................................................................................................................................................................................................................................................................................................................................. ...
......................................................................................................................................................................................................................................................................................................................................................................................................... .. 2
. . .................................................................................................................................................................................................................... ....
.
1 ....
........................................................
2

14. (Grades 9-10) A boy always tells the truth on Thursdays and Fridays,
always tells lies on Tuesdays, and tells either truth or lies on the rest of the
days of the week. Every day he was asked what his name was and six times in
a row he gave the following answers: John, Bob, John, Bob, Pit, Bob. What
did he answer on the seventh day?
(A) John
(B) Bob
(C) Pit
(D) Kate
(E) Not enough information to de ide

. (Grades 11-12) An equilateral triangle and a ir le

M are ins ribed in a ir le K , as shown in the gure.
What is the ratio of the area of K to the area of M ?
(A) 8 : 1
(B) 10 : 1
(C) 12 : 1
(D) 14 : 1
(E) 16 : 1

15

....
.................. .........................
...... K
..........
..... ... ........
....
.... ..... ...............
....
.
.
.
........
...
...
........
...
..
...
........
....M ....
.
........
........ ....
..................... ...
........ ...
......
......
.........
..
....
.........
.
.
.
.....
.
.....
.
.
.
.
..
.........................
........ ....
...
.
........
..
.....
...
.......
..
........
...
...
........
.
.
.... ....
.
...
.
.
.
...
.... ... .............
.
.
.
.
....... ........
...........
......
................. ........................
......

4
Con ours Math Kangaroo
Feuille d'entra^
nement

Partie A (3 points par question)

. (Classes 3-4) Dans l'exemple d'addition idessus, haque lettre dierente


represente

un hire
dierent.

Quel hire la lettre K represente-t-elle

?

O
K
O

(A) 0

(B) 1

(C) 2

+
W

(D) 8

K
O
W

(E) 9

. (Classes 5-6) Dix henilles se promenaient a la le indienne dans un par .

Chaque henille mesurait 8 m et, pour des raisons de se urit

e,
 elles gardaient une distan e de 2 m entre ha une d'elles. Quelle etait

la longueur
totale de leur ortege
 ?

(A) 100 m

(B) 98 m

(C) 82 m

(D) 102 m

(E) 96 m

. (Classes 7-8) Une fourmi ourt le long ..................................sq...q.r................................................

...
d'une regle
 de 10 m de longueur, a la vitesse ...
.
1
2
3
4
5
6
7
8
9
.
onstante de 1 m a la se onde (voir la gure). .......................................................................................
Chaque fois qu'elle atteint une extremit
 e,
 elle
ourt dans la dire tion opposee
 et elle met exa tement 1 se onde pour hanger de dire tion. La fourmi part de l'extremit
 e gau he de la regle.

Pres
 de
quel hire sera-telle apres
 2009 se ondes ?
3

.....
.

(A) 1 m
4

(B) 2 m

(C) 3 m

.....
.

(D) 4 m

.....
.

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

.....
.

.....
.

.....
.

.....
.

(E) 5 m

. (Classes 9-10) Lequel des nombres 26 , 35 , 44 , 53 , 62 est-il le plus grand ?

(A) 26

(B) 35

(C) 44

(D) 53

(E) 62

. (Classes 11-12) Un de orateur


a prepar
 e un melange

de peinture ou les
volumes des ouleurs rouge et jaune etaient

dans un rapport de 2 : 3. Trouvant le melange

trop lair, il ajouta 2 L de peinture rouge. Le rapport des
volumes des ouleurs rouge et jaune devint alors de 3 : 2. Combien de litres
de peinture le de orateur

a-t-il utilise ?
5

(A) 5 L

(B) 6 L

(C) 7 L

(D) 8 L

(E) 9 L

Partie B (4 points par question)

. (Classes 3-4) Deux gar ons jouent au tennis jusqu'a e que l'un d'eux gagne
quatre fois. Un mat h de tennis ne peut nir en un pointage nul. Quel est le
plus grand nombre de jeux qu'ils peuvent jouer ?
6

(A) 8

(B) 7

(C) 6

(D) 5

(E) 9

5
. (Classes 5-6) Dans deux ans, mon ls aura deux fois l'^age qu'il avait il y a
deux ans. Dans trois ans, ma lle aura trois fois l'^age qu'elle avait il y a trois
ans. Quelle reponse

de rit-elle

le mieux l'^age de la lle et du ls ?
(A) Le ls est plus a^ ge ;
(B) La lle est plus a^ gee
 ;
(C) Ils sont des jumeaux ; (D) Le ls est deux fois plus a^ ge que la lle ;
(E) La lle est deux fois plus a^ gee
 que le ls.
7

. (Classes 7-8) Sur une droite on marque des points de sorte que toutes
les distan es de 1 m, 2 m, 3 m, 4 m, 5 m, 6 m, 7 m et 9 m gurent
parmi les distan es entre es points. Combien y a-t-il au minimum de points
marques
 sur ette droite ?
(A) 4
(B) 5
(C) 6
(D) 7
(E) 8
8

9. (Classes 9-10) Liliane, Ni ole, Katia et Charlotte sont allees

 au inema.

Comme il n'etait

pas possible d'a heter quatre pla es ensemble, elles ont
 et d'autres
a hete des billets pour les sieges
 numero
 7 et 8 dans la 10e -rangee
pour les sieges

numero
 3 et 4 dans la 12e -rangee.
 De ombien de manieres

peuvent-elles hoisir de s'asseoir, si Charlotte ne veut pas e^ tre assise a ot
^ e
de Ni ole ?
(B) 20
(C) 16
(D) 12
(E) 8
(A) 24

. (Classes 11-12) Soit ABC un triangle

iso ele
 ave BC = AC . Les segments
DE , F G, HI , KL, M N , OP et XY
divisent les ot
^ es
 AC et CB en parties
egales.

Trouver XY si AB = 40 m.
(B) 35 m
(A) 38 m
(C) 33 m
(D) 30 m
(E) 27 m
10

C
..
.... ....

D ..................................... E

F ......................................................... G
H .............................................................................. I
K .................................................................................................. L
M ...................................................................................................................... N
O ............................................................................................................................................ P

X ................................................................................................................................................................. Y
A .......................................................................................................................................................................... B

Partie C (5 points par question)

. (Classes 3-4) En utilisant des ubes identiques, Mathieu et Ni olas ont onstruit deux
b^atiments, omme illustres
 dans les gures. Le
b^atiment de Mathieu pese
 200 g et elui de Ni olas 600 g. Combien de ubes du b^atiment de
Ni olas sont-ils a hes
 et ne peuvent e^ tre vus
dans la gure ?
(B) 2
(C) 3
(A) 1
(D) 4
(E) 5
11

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

b^atiment
de Ni olas

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

b^atiment
de Mathieu

. (Classes 5-6) On onsidere

 tous les nombres de quatre hires, divisibles
par 6 et dont les hires, lus de gau he a droite, vont en ordre roissant. Quel
est le hire des entaines dans le plus grand de es nombres ?
(A) 7
(B) 6
(C) 5
(D) 4
(E) 3
12

6
. (Classes 7-8) On divise un arre de ot
^ e 3 en polygones ave plusieurs segments omme indique dans
la gure. Quel est le per entage de l'aire de la gure
ombree
 par rapport a elle du arre ?
(A) 30%
(B) 33 13 %
(C) 35%
(E) 50%
(D) 40%
13

.........1..........................2................
.. ..................................................................................................... ..... 1
. ........................................................................................................................ .
2 .. ..............................................................................................................................................................................................................................................................................................................................................................................................................................................................................................
.. .............................................................................................................................................................................................................................................................. ..
................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................... .... 2
................................................................................................................................. ...
1 .....
......
................................................
2

. (Classes 9-10) Un gar on dit toujours la verit

 e les jeudis et vendredis,
ment toujours les mardis et, les autres jours de la semaine, soit il dit la verit
 e
soit il ment. On lui demanda son nom haque jour de la semaine et les six
premieres
 fois, il donna les reponses

suivantes : Jean, Bernard, Jean, Bernard,
Paul, Bernard. Quelle fut sa reponse

le septieme
 jour ?
(A) Jean
(B) Bernard (C) Paul
(D) Lu
(E) Pas possible de de ider

14

. (Classes 11-12) On ins rit un triangle equilat


eral

et un er le M dans un er le K , omme indique dans
la gure. Quel est le rapport de l'aire de K a elle de
M?
(B) 10 : 1
(C) 12 : 1
(A) 8 : 1
(D) 14 : 1
(E) 16 : 1
15

.....................................
..........
........
.................
.....K
..... ... ...............
....
........
.... .....
...
.
.
........
...
...
..
.
.
.
.
.
........
..
....
........
.... .M
........ ....
.............................
........ ..
..........
...
...
...
...
.
.
.
.
........
.......
....
........ ....
.........................
.
........
...
..
..
.......
.
.
.
.
.
...
...
.
.
.
.
.
.
...
..
........
.... .....
...
.......
.
.
.
.
.
.
.
.
.
.
.... ... .........
...............
.....
.........
......
...........................................

Next we shall give solutions to the Mathemati s Asso iation of Quebe

Contest (Se ondary level) February 9, 2006 [2008 : 67-68. We apologize to
any readers who sent in solutions to this ontest but whose solutions we have
lost.
1. A parti ular magi square. It is well known that a magi square is obtained
by putting numbers in a square su h that the sum of ea h row, olumn, and
diagonal is the same, as for example,
8 1
3 5
4 9

6
7
2

Imagine now that we de ide to invent a new form of su h squares by repla ing
the sum by a produ t. We ask you to nd su h a square by repla ing the
asterisks, , by natural numbers, not ne essarily distin t or onse utive, in
the following square:
1
4

7
Solution by the editor.
Suppose that the square A below is a magi square. Then the square
B is a magi square for produ ts. For example, by the Law of Exponents, the
produ t along the rst row of B is xa xb xc = xa+b+c and the produ t along
the rst olumn of B is xa xd xg = xa+d+g and these are the same be ause
a + b + c = a + d + g . The same is true for the other rows, olumns, and
diagonals of B .
A =

a b
d e
g h

c
f
i

xa
xd
xg

B =

xb
xe
xh

xc
xf
xi

Now, if we subtra t 1 from every entry of the rst square given in the question
and if we take x = 2, then the square B below is a solution to the problem.
A =

7 0
2 4
3 8

5
6
1

B =

27
22
23

20
24
28

25
26
21

128
4
8

1
16
256

32
64
2

2. Clovis' outing. Clovis likes to take an outing in the natural numbers.

Ea h day, he starts with a natural number of his hoi e, the biggest possible.
Then, during his day, he passes from number to number using the following
rules. Suppose that the sequen e of numbers is urrently at n.

(1) If n is divisible by 3 without remainder, then the next number is n/3.

(2) If the remainder after dividing
2n + 1.

by

is 1, then the next number is

(3) If the remainder after dividing

2n 1.

by

is 2, then the next number is

(4) If n = 1, then the sequen e stops.

Over the years that he has played this game, he noti ed that, whatever the
starting number, the sequen e always ended up with the number 1. However, he wonders if there is a sequen e that in reases indenitely, with larger
and larger numbers on average, or su h that it ends up in a loop of numbers
that does not ontain 1. Determine if su h a sequen e is possible and give
an example, or show that su h a sequen e does not exist by showing that all
sequen es using the above rules inevitably end up at the number 1.
Here is an example of su h a sequen e: Starting with 55, we get 111,
37, 75, 25, 51, 17, 33, 11, 21, 7, 15, 5, 9, 3 and 1, whi h ends the sequen e.
Solution by the editor.
Note that Clovis' sequen e starting with 55 has a de reasing subsequen e that goes to 1, given by the underlined numbers: 55, 111, 37, 75, 25,

8
51, 17, 33, 11, 21, 7, 15, 5, 9, 3, 1.

We will show that for any number a > 1

in one of Clovis' sequen es, there is always a number b oming after a in the
sequen e su h that a > b. Thus, if Clovis starts with n > 1, then there will
be a subsequen e n, m, p, . . . with n > m > p > and this subsequen e
must eventually hit the number 1 (be ause all of the terms in it are positive,
it annot de rease forever).
If a > 1 and a = 3k, k > 1, then by rule (1) the number b = k omes
right after a and a > b.
If a > 1 and a = 3k + 1, k > 0, then by rule (2) the number 2a + 1 =
2(3k+1)+1 = 6k+3 omes right after the number a, and then by rule (1) the
number (6k + 3)/3 = 2k + 1 omes after 2a + 1. Sin e a = 3k + 1 > 2k + 1,
we see that the number b = 2k + 1 omes after a and a > b.
If a > 1 and a = 3k + 2, k 0, then by rule (3) the number 2a 1 =
2(3k+2)+1 = 6k+3 omes right after the number a, and then by rule (1) the
number (6k + 3)/3 = 2k + 1 omes after 2a 1. Sin e a = 3k + 2 > 2k + 1,
we see that the number b = 2k + 1 omes after a and a > b.
Thus, in all ases where a > 1, there is a number b oming after a in
the sequen e su h that a > b, and we are done.
3. Eight balls in two urns. We give you two similar urns, four white balls,
and four bla k balls. You must separate the balls amongst the two urns
(not ne essarily the same number in ea h urn), after whi h both urns will
be made indistinguishable. How should the balls be distributed to maximize
the han es that, if you draw a ball randomly from a randomly hosen urn,
you will obtain a white ball?

Solution by the editor.

Put 1 white ball in one urn and all the other balls in the other urn. The
probability of hoosing the urn with 1 white ball and then drawing that white
ball from it is 12 1 = 12 and the probability of hoosing the other urn and then
3
3
= 14
drawing a white ball from it is 12 3+4
. Thus, with this distribution, the
overall probability of ultimately obtaining a white ball is p = 12 1+ 12 37 = 57 .
Now let p1 and p2 be the probabilities of drawing white balls from the
two urns (p1 = 1 and p2 = 37 above). The overall probability of ultimately
p2
obtaining a white ball is then p = 12 p1 + 12 p2 = p1 +
, whi h is the average
2
5
of the probabilities p1 and p2 . Therefore, p > 7 implies that p1 > 57 or
p2 > 57 . To make an urn with p1 > 57 (say) we must have (w, b) = (1, 0),
(2, 0), (3, 0), (3, 1), (4, 0), or (4, 1), where the urn ontains w white balls
and b bla k balls. These are the only distributions that ould yield p > 75 .
In the ase of (w, b) = (2, 0), (3, 0), or (4, 0) we have p < 57 , as moving all
but one white ball to the other urn in reases p2 but leaves p1 = 1. Finally,
(w, b) = (3, 1) or (4, 1) yields p = 12 or p = 25 , ea h less than 57 .
Our rst distribution maximizes our han e of obtaining a white ball.

9
........................................
4. The three atta hed barrels. Three big ylindri al
........
.....
....
......
...
....
.
.
...
..
barrels, lying parallel to the earth, are atta hed by
..
..
.......................
.
.
.
.
.
.
.
.
..
.
.
...
.
.
........
...
..... ....
.... ..........
.... ..
a steel able at their onta t points, A and B , su h .................................. ........
... ...........................
....
.....
...
....
....
...
that they stay xed in pla e. Knowing that the two ........... ..............
..
....
.
............................ ....
.... ...............................
.
.
.
smaller ones ea h have a radius of 4 metres and the ............ ........q....A
.
.
.
.
B
.......
........................................................................q
.
..
biggest one has a radius of 9 metres, what is the ............ .............................. ................................. ............
.....................
..............................
.....................
length of the steel able?
Solution by the editor.
Join the entres of the smaller barrels
........................................
..........
.......
.......
.....
.....
....
and drop a perpendi ular to this segment from
....
....
.
.
...
...
.
...
.
.
.
..
.
the entre of the larger barrel, as in the dia..
...
..
...
...
C
.
..
.
...
.
.
gram at right. Sin e the barrels rest on the
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
...
...... .... ...............
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.......... B........................................
........A ..........
...
......
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
earth, the length of CZ is the dieren e of ...... ...........................q............................................................................................................................................q.................. .........
...................................................................................
...
..
their radii, that is, |CZ| = 9 4 = 5. Also ........... X ..............................................................................................Z
... ...
..... .....
Y .........
........
......
.....
.....
.
.................. .........................
.
...............................
.
.
.
.
.....................
.......
the length of CY is the sum of the radii, that
is, |CY | = 9 + 4
= 13. By the Pythagorean
Theorem |ZY | = 132 52 = 12. Thus, |XY | = 24. Finally, sin e triangle
9
216
CAB is similar to triangle CXY , we have |AB| =
|XY | =
.
13
13

5. The magi words. An illusionist is sear hing for

magi words to a ompany his many magi tri ks.
R
He de ides to onstru t his magi words starting D.........................
..
............
......... ....
.........
...
........
...
.........
.
.
.
.
.
.
.
.
.
.
.........
...
...
with the diagram on the right. He takes a path
.....
.
.
.
.
.
.
.........
.
...
...
.
......... ..............
...
...
............
.
...
...
.
.
.
.
.
.
.
.........
through the diagram and jots downs the letters he
...
...
......
.
.
.
.
.
.
.
.
.
.
.
...
...
.........
.....
.
.
.
.
.
.
.
...
...
.
.
.
.
.
.
.........
A
... ..............
.
.
.
.
......... .....
nds on it. Ea h magi word must have exa tly 11
... ..........
.
.
.
.
.
.
....
...
letters and must start and end with the letter A. C
B
Two onse utive letters must never be identi al.
How many magi words are there?
Note: Here are two possible magi words: ABRACADABRA and
ARADCABARBA.

Solution by the editor.

Let Mk be the number of k-letter words that start with A and end with
A and that an be formed by travelling through the bowtie. Let Nk be the
number of k-letter words starting with A but not ending with the letter A
that an be similarly formed.
If k 2, then we see that Mk = Nk1 , be ause removing the letter A
from a word ending with A leaves a word not ending in A (but still starting
with A) and the pro ess an be reversed. Similarly, by deleting the last letter
of a word of length k that starts with A but does not end in A, we see that
Nk = Nk1 + 4Mk1 , be ause any of the four letters dierent from A an
be added to a word not ending in A or else there is only way to extend a
(k 1)-letter word not ending in A to one that still does not end in A. Sin e
Mk1 = Nk2 the last equation be omes Nk = Nk1 + 4Nk2 , where
k 1.

10
We now have N1 = 0, N2 = 4, N3 = N2 + 4N1 = 4 + 4 0 = 4, and so
forth. The results of al ulating the Ni are summarized in the table below:
N1
0

N2
4

N3
4

N4
20

N5
36

N6
116

N7
260

N8
724

N9
1764

N10
4660

Finally M11 = N10 , so there are 4660 magi words altogether.

6. All ten digits. Find the smallest positive natural number N su h that, in
the de imal notation, N and 2N together use all ten digits: 0, 1, 2, . . . , 9.

Solution by the editor.

We have 2(13485) = 26970, and we will prove that if N1 and 2N1
together use all ten digits and N1 N = 13485, then N1 = N .
As N1 has ve digits and N1 N , then N1 = 1 . . . and 2N1 = 2 . . ..
Digits 1 and 2 are now used and 2N1 uses 0 (otherwise N1 uses 0 and 2N1
then uses 0 or 1, a ontradi tion). Thus, N1 = 13 . . .. The smallest available
digit for N1 is now a 4 and N1 N , hen e N1 = 134 . . . and 2N1 = 26 . . ..
The number N1 uses 5, be ause 2N1 uses 0. If N1 = 1345x, then x is a digit
greater than 5 and 2N1 = 2691y, a ontradi tion. Thus, N1 = 134x5 N .
Finally, x 6= 7, hen e x = 8.
Therefore, N1 = N and N is the smallest positive integer with the
given property.
[Ed.: Rolland Gaudet oers the solution N = 6792 if initial zeroes are
allowed, for then 2(6792) = 013584.
7. The pizza toppings. At the Julio pizzeria, all the pizzas have heese and
tomato sau e on them. The hoi e of toppings is limited to bla k olives,
an hovies, and sausage. Of the 200 lients Julio had yesterday, 40 took an hovies, 80 took bla k olives, 120 took sausage, 60 took at the same time
bla k olives and sausage, but none took at the same time an hovies and
bla k olives or an hovies and sausage. How many lients took none of the
three toppings?

Solution by the editor.

Let t be the number of ustomers who took at least one topping. Any
ustomer who took an hovies took no other topping, so t = 40 + x where x
is the number of ustomers who took bla k olives or sausage (or both). There
were 60 ustomers who took both bla k olives and sausage, so 20 = 80 60
took just bla k olives and nothing else. Similarly, 60 = 120 60 ustomers
took sausage and nothing else. Thus, t = 40+x = 40+(20+60+60) = 180,
and the number of ustomers who took no toppings is 200 t = 20.

11

MATHEMATICAL MAYHEM
Mathemati al Mayhem began in 1988 as a Mathemati al Journal for and by
High S hool and University Students. It ontinues, with the same emphasis,
as an integral part of Crux Mathemati orum with Mathemati al Mayhem.
The Mayhem Editor is Ian VanderBurgh (University of Waterloo). The
other sta members are Monika Khbeis (As ension of Our Lord Se ondary
S hool, Mississauga) and Eri Robert (Leo Hayes High S hool, Frederi ton).

Mayhem Problems

Please send your solutions to the problems in this edition by 1 May 2009. Solutions re eived after this date will only be onsidered if there is time before publi ation
of the solutions.
Ea h problem is given in English and Fren h, the o ial languages of Canada.
In issues 1, 3, 5, and 7, English will pre ede Fren h, and in issues 2, 4, 6, and 8,
Fren h will pre ede English.
The editor thanks Jean-Mar Terrier of the University of Montreal for translations of the problems.
M376. Proposed by the Mayhem Sta.

Determine the value of x if

2

2
102009 + 25 102009 25 = 10x .

M377. Proposed by the Mayhem Sta.

An arithmeti sequen e onsists of 9 positive integers. The sum of the
terms in the sequen e is greater than 200 and less than 220. If the se ond
term in the sequen e is 12, determine the sequen e.
M378. Proposed by the Mayhem Sta.
Points C and D are hosen on the semi- ir le with diameter AB so
that C is loser to A. Segments CB and DA interse t at P ; segments AC
and BD extended interse t at Q. Prove that QP extended is perpendi ular
to AB .
M379. Proposed by John Grant M Loughlin, University of New Brunswi k,
Frederi ton, NB.
The integers 27 + C , 555 + C , and 1371 + C are all perfe t squares, the
square roots of whi h form an arithmeti sequen e. Determine all possible
values of C .

12
M380. Proposed by Bru e Shawyer, Memorial University of Newfoundland, St. John's, NL.
Triangle ABC is right-angled at C and has BC = a and CA = b, with
a b. Squares ABDE , BCF G, and CAHI are drawn externally to triangle
ABC . The lines through F I and EH interse t at P , the lines through F I
and DG interse t at Q, and the lines through DG and EH interse t at R. If
triangle P QR is right-angled, determine the value of ab .
M381. Proposed by Mihaly Ben ze, Brasov, Romania.
Determine all solutions to the equation
1
1
1
1
+
+
+
= x2 4x 4 .
x1
x2
x6
x7

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

M376. Propose par l'Equipe
de Mayhem.
Determiner

la valeur de x si

2

2
102009 + 25 102009 25 = 10x .


M377. Propose par l'Equipe
de Mayhem.
Determiner

la suite arithmetique

formee
 de 9 entiers positifs dont la
somme se situe entre 200 et 220 et dont le se ond terme vaut 12.

M378. Propose par l'Equipe
de Mayhem.
A partir du point A, on hoisit deux points C et D sur un demi- er le
de diametre
 AB . Soit P l'interse tion des droites CB et DA, et Q elle des
droites AC et BD. Montrer que la droite P Q est perpendi ulaire a AB .
M379. Propose par John Grant M Loughlin, Universite du NouveauBrunswi k, Frederi ton, NB.
Les entiers 27 + C , 555 + C et 1371 + C sont tous des arres
 parfaits
dont les ra ines arrees
 forment une suite arithmetique.

Trouver toutes les
valeurs possibles de C .
M380. Propose par Bru e Shawyer, Universite Memorial de Terre-Neuve,
St. John's, NL.
Dans un triangle ABC d'angle droit en C , soit BC = a et CA = b,
ave a b. Exterieurement

au triangle ABC , on onstruit les arres
 ABDE ,
BCF G et CAHI . Soit respe tivement P , Q et R les interse tions des droites
F I et EH , F I et DG, DG et EH . Determiner

la valeur de ab pour que P QR
soit un triangle re tangle.

13
M381. Propose par Mihaly Ben ze, Brasov, Roumanie.
Determiner

toutes les solutions de l'equation

1
x1

1
x2

1
x6

1
x7

= x2 4x 4 .

Mayhem Solutions

M338. Proposed by the Mayhem Sta.

Two students mis opy the quadrati equation x2 + bx + c = 0 that
their tea her writes on the board. Jim opies b orre tly but mis opies c;
his equation has roots 5 and 4. Vazz opies c orre tly, but mis opies b; his
equation has roots 2 and 4. What are the roots of the original equation?

Solution by Taylor Thetford, student, Lakeview High S hool, San Angelo, TX,
USA.
The roots of the quadrati equation that Jim writes down are 5 and 4.
His quadrati equation is thus (x 5)(x 4) = x2 9x + 20 = 0. Sin e Jim
opied b orre tly, we an on lude that in the original quadrati equation,
b = 9.
Similarly, sin e Vazz's roots are 2 and 4, his quadrati equation has the
form (x 2)(x 4) = x2 6x + 8 = 0. Sin e Vazz opied c orre tly, then
c = 8.
Thus, the original equation was x2 9x + 8 = 0. Fa toring, we obtain
(x 1)(x 8) = 0. Therefore, the roots of the original equation are 1 and 8.
Also solved by CAO MINH QUANG, Nguyen Binh Khiem High S hool, Vinh Long,
Vietnam; JACLYN CHANG, student, Western Canada High S hool, Calgary, AB; PETER CHIEN,
student, Central Elgin Collegiate, St. Thomas, ON; IAN JUNE L. GARCES, Ateneo de Manila
University, Quezon City, The Philippines; JOHAN GUNARDI, student, SMPK 4 BPK PENABUR,
 IES
Jakarta, Indonesia; RICHARD I. HESS, Ran ho Palos Verdes, CA, USA; RICARD PEIRO,

\Abastos", Valen ia, Spain; JOSE HERNANDEZ
SANTIAGO, student, Universidad Te nologi a

de la Mixte a, Oaxa a, Mexi o; KUNAL SINGH, student, Kendriya Vidyalaya S hool, Shillong,
India; BILLY SUANDITO, Palembang, Indonesia; LUYAN ZHONG-QIAO, Columbia International College, Hamilton, ON; and TITU ZVONARU, Comane
 sti, Romania.

M339. Proposed by the Mayhem Sta.

(a) Determine the number of integers between

whi h ontain exa tly two equal digits.

100

and

199,

in lusive,

(b) An integer between 1 and 999 is hosen at random, with ea h integer

being equally likely to be hosen. What is the probability that the integer has exa tly two equal digits?

14
Solutions by Taylor Thetford, student, Lakeview High S hool, San Angelo,
TX, USA.
1xy .

(a) Ea h of the integers in the given range an be written in the form

There are three ases to onsider.

Case 1. The rst and last digits are the same. Here, we are looking for
integers 1x1 where the middle digit an take any value ex ept 1. This yields
9 possibilities.
Case 2. The rst and se ond digits are the same. Here, we are looking for

integers 11y where the last digit an take any value ex ept 1. This again
yields 9 possibilities.

Case 3. The se ond and last digits are the same. Here, we are looking for

integers 1xx where x is not 1. Again, there are 9 possibilities.

Adding the results from our three ases, we nd that there are 27 numbers between 100 and 199, in lusive, that ontain exa tly two equal digits.
(b) We ount the number of integers in the range 1 to 999, in lusive,
that have exa tly two equal digits.
First, between 1 and 99, there are 9 of these, namely, 11, 22, . . . , 99.
Next, between 100 and 199, we have ounted 27 in part (a). Using the same
argument as in (a), we an show that there are 27 numbers between 200 and
299, in lusive, and for every other interval of one hundred numbers up to
the range of 900 to 999.
There are therefore 9 + 9 27 = 252 numbers between 1 and 999 whi h
ontain exa tly two equal digits.
The probability that a randomly sele ted integer between
28
has exa tly two equal digits is thus 252
=
.
999
111

and

999

Also solved by PETER CHIEN, student, Central Elgin Collegiate, St. Thomas, ON; IAN
JUNE L. GARCES, Ateneo de Manila University, Quezon City, The Philippines; RICHARD
I. HESS, Ran ho Palos Verdes, CA, USA; and LUYAN ZHONG-QIAO, Columbia International
College, Hamilton, ON. There were 3 in orre t solutions and 1 partial solution submitted.

M340. Proposed by the Mayhem Sta.

Let ABC be an isos eles triangle with AB = AC , and let M be the

mid-point of BC . Let P be any point on BM . A perpendi ular is drawn to
BC at P , meeting BA at K and CA extended at T . Prove that P K + P T
is independent of the position of P (that is, the value of P K + P T is always
the same, no matter where P is pla ed).

15
Solution by Cao Minh Quang, Nguyen Binh Khiem High S hool, Vinh Long,
Vietnam.
T
Sin e ABC is isos eles with sides AB
and AC of equal length, we have M A BC .
Also, sin e P T BC , then M A||P T .
A
Sin e M A||P K , then M BA is similar
to P BK sin e ea h is right-angled and they
K
share the angle at B . From this, we obtain
P B MA
PK
MA
, hen e P K = M B .
=
PB
MB
Similarly, sin e M A||P T , then CP T
T
MA
is similar to CM A, when e PP C
=
MC
......
... ....
.... .....
... ....
.
...
...
.....
...
...
...
...
....
...
..........
....
... ... ...
... ..... ......
.
....
.
... ..... .... ......
...
.... ....
.....
...
... ....
...
....
...
.....
....
...
......
...
.
.
.
.
...
.
.. ...
.
.
.
.
...
..
... ....
...
.
.
.
.
...
..
.. ....
.
.
...
.
.
...
..
...
..
.
.
.
.
...
..
...
.....
.
...
.
.
...
.
..
..
.
.
.
.
.
...
..
....
..
...
.
.
.
.
...
...
....
....
.
...
.
.
.
..
...
...
.
.
.
.
..
.
.
..
.
.
.
.
.
.
...............................................................................................................................................

MA
and so P T = P CM C
.
Sin e M B = M C =

PK + PT

1
BC
2

P M

, we an on lude that

(P B + P C) M A
P B MA
P C MA
+
=
MB
MC
MB

BC M A
= 2M A .
MB

Thus, P K + P T is independent of the position of P , sin e it depends

only on the length of M A.
Also solved by EDIN AJANOVIC, student, First Bosniak High S hool, Sarajevo, Bosnia
and Herzegovina; IAN JUNE L. GARCES, Ateneo de Manila University, Quezon City, The
 IES \Abastos",
Philippines; RICHARD I. HESS, Ran ho Palos Verdes, CA, USA; RICARD PEIRO,
Valen ia, Spain; KUNAL SINGH, student, Kendriya Vidyalaya S hool, Shillong, India; BILLY
SUANDITO, Palembang, Indonesia; and TITU ZVONARU, Comane
 sti, Romania. There was 1
in orre t solution submitted.

M341. Proposed by the Mayhem Sta.

Let ABC be a right triangle with right angle at B . Sides BA and BC
are in the ratio 3 : 2. Altitude BD divides CA into two parts that dier in
length by 10. What is the length of CA?

Solution by Taylor Thetford, student, Lakeview High S hool, San Angelo, TX,
USA.
Let 2x and 3x be the lengths of
B
CB and AB , respe tively. Let y and
y + 10 be the lengths of CD and DA,
3x
respe tively. Let z be the length of
2x
z
BD . We wish to nd 2y + 10, whi h
is the length of CA.
By applying the Pythagorean
A
y
Theorem in ABC , we nd that
y + 10
C
D
(2x)2 + (3x)2 = (2y + 10)2 and so
13x2 = 4y 2 + 40y + 100.
....
..............
.... ... .............
.......
.... ...
.......
.... .....
.
.
.......
.
.......
....
....
.......
...
.
.
.
.
.......
..
..
.
.
.......
.
.
.
..
..
.......
.
.
.
.
.
.......
..
..
.
.
.
.......
.
.
..
..
.......
.
.
.
.
.
.......
..
.
.
.
.
.
.......
.
..
..
.
.......
.
.
.
.
.......
..
..
.
.
.
.
.
.......
..
........
.
.
.......
.
.
.
..
..
. ...
.
.
.
.
.
......................................................................................................................................................................................................

16
Applying the Pythagorean Theorem to BDC and BDA, we nd
that y2 + z 2 = 4x2 and z 2 + (y + 10)2 = 9x2 .
Eliminating z in the last two equations gives 4x2 y2 = 9x2 (y+10)2 .
Therefore, 5x2 = (y + 10)2 y2 = 20y + 100 or x2 = 4y + 20, and so
13x2 = 52y + 260.
Combining this result with 13x2 = 4y2 + 40y + 100, we nd that
52y + 260
4y 12y 160
y 2 3y 40
(y + 5)(y 8)
2

=
=
=
=

4y 2 + 40y + 100 ;
0;
0;
0.

Sin e y > 0, then y = 8, and so CA = 2y + 10 = 26.

Also solved by EDIN AJANOVIC, student, First Bosniak High S hool, Sarajevo, Bosnia
and Herzegovina; JACLYN CHANG, student, Western Canada High S hool, Calgary, AB; IAN
JUNE L. GARCES, Ateneo de Manila University, Quezon City, The Philippines; RICHARD
 IES \Abastos", Valen ia, Spain;
I. HESS, Ran ho Palos Verdes, CA, USA; RICARD PEIRO,
KUNAL SINGH, student, Kendriya Vidyalaya S hool, Shillong, India; BILLY SUANDITO,
Palembang, Indonesia; LUYAN ZHONG-QIAO, Columbia International College, Hamilton, ON;
and TITU ZVONARU, Comane
 sti, Romania. There was 1 in orre t solution submitted.

M342. Proposed by the Mayhem Sta.

Quin y and Celine have to move 10 small boxes and 10 large boxes. The
hart below indi ates the time that ea h person takes to move ea h type of
box.
Celine Quin y
small box 1 min. 3 min.
large box 6 min. 5 min.

They start moving the boxes at 9:00 am. What is the earliest time at whi h
they an be nished moving all of the boxes?
Solution by Mayhem Sta.
Let x represent the number of small boxes and y represent the number
of large boxes that Celine moves. Sin e there are 10 small boxes and 10 large
boxes, then Quin y moves 10 x small boxes and 10 y large boxes.
Given the lengths of time that ea h takes, it takes Celine x+6y minutes
and it takes Quin y 3(10 x) + 5(10 y) = 80 3x 5y minutes. If x = 9
and y = 4, then Celine takes 33 minutes and Quin y takes 33 minutes. We
show that it annot be done faster than this.
If Quin y and Celine nish in fewer than 33 minutes, then ea h takes
at most 32 minutes, so the total working time is at most 64 minutes, so
x + 6y + (80 3x 5y) = 80 2x + y 64 or 2x y 16.
Sin e x and y are nonnegative integers and ea h is less than 10, then
the possible pairs (x, y) that satisfy this inequality are (8, 0), (9, 0), (9, 1),
(9, 2), (10, 0), (10, 1), (10, 2), (10, 3), and (10, 4).

17
Sin e we want ea h of Celine's time and Quin y's time to be at most
minutes, then we need x + 6y 32 and 80 3x 5y 32. The
rst inequality eliminates the pair (10, 4) from the list of possible pairs. The
se ond inequality simplies to 3x + 5y 48; none of the remaining pairs
satisfy this inequality.
Thus, none of these possibilities take any less time than 33 minutes.
Therefore, the earliest possible nishing time is 9:33 a.m.
32

There were 4 in orre t and 3 in omplete solutions submitted.

An expanded treatment of a similar problem appeared in the Problem of the Month
olumn in CRUX with MAYHEM, volume 34, number 2.

M343. Proposed by the Mayhem Sta.

The Fibona i numbers are dened by f1 = f2 = 1 and, for n 2,
by fn+1 = fn + fn1 . The rst few Fibona i numbers are 1, 1, 2, 3, 5, 8,
13, 21, 34, 55, 89, 144, . . . . Find the sum of the rst 100 even Fibona i
numbers.

Solution by Edward T.H. Wang, Wilfrid Laurier University, Waterloo, ON.

Sin e f1 = f2 = 1, f3 = 2, and fn = fn1 + fn2 , then fm is even
if and only if m is a multiple of 3. (This is be ause the parities of the terms
will form the pattern Odd, Odd, Even, Odd, Odd, Even, and so on.)
n
P
f3k , then
If Sn =
k=1

Sn =

n
1 X

(f3k +f3k ) =

k=1

n
1 X

(f3k2 +f3k1 )+f3k

k=1

3n
1 X

fk .

(1)

k=1

Next, we have f1 = f3 f2 , and f2 = f4 f3 , and also f3 = f5 f4 ,

and so on until fr1 = fr+1 fr and fr = fr+2 fr+1 .
Sin e the right side of the sum of the n equations above \teles opes",
it follows that
r
X
fk = fr+2 f2 = fr+2 1 .
(2)
k=1

From (1) and (2), we nd that Sn = 12 (f3n+2 1). In our parti ular
ase, S100 = 12 (f302 1). Maple omputes the value of S100 to be exa tly
290905784918002003245752779317049533129517076702883498623284700.
For the re ord, by Binet's formula for Fibona i numbers we have that

1
1
1
fm = (m m ), where = (1 + 5) and = (1 + 5). Hen e
2
2
5

the required sum is also given by S100 =

1
1
302 302 .
2
2 5

Also solved by EDIN AJANOVIC, student, First Bosniak High S hool, Sarajevo, Bosnia
and Herzegovina; CAO MINH QUANG, Nguyen Binh Khiem High S hool, Vinh Long, Vietnam;
IAN JUNE L. GARCES, Ateneo de Manila University, Quezon City, The Philippines; RICHARD
 IES \Abastos", Valen ia, Spain;
I. HESS, Ran ho Palos Verdes, CA, USA; RICARD PEIRO,

DIVYANSHU RANJAN, Delhi, India; JOSE HERNANDEZ
SANTIAGO, student, Universidad
Te nologi a

de la Mixte a, Oaxa a, Mexi o; and TITU ZVONARU, Comane
 sti, Romania.

18
Problem of the Month

Ian VanderBurgh

Approximation is one of the most important on epts in mathemati s.

Problem (2006 Canadian Open Mathemati s Challenge) Determine, with jus-

ti ation, the fra tion pq , where p and q are positive integers and q

< 100,

that is losest to, but not equal to, 37 .

While it is tempting to get out your al ulator, it an initially only help

so mu h. If you al ulate 37 , you'll obtain 0.428571 . . .. This doesn't help in
any obvious way to answer the question.
A rst approa h after the al ulator might be to go for the fra tion with
the largest possible denominator. This makes a lot of sense in some ways,
as the fra tions with the largest denominators will be losest together and so
would seem to have the best han e of being losest to 37 . In our ase, the
largest possible denominator is q = 99. The given fra tion, 37 , is between
42
43
= 0.424242 . . . and
= 0.434343 . . .. After a qui k look, we an tell
99
99
3
42
dier by
that 7 is loser to 99 . From the de imal approximations, 37 and 42
99
about 0.004. Is this the losest of all possible fra tions?
Another idea is to try to onvert 37 into the equivalent fra tion with
the largest possible denominator and then adjust from there. Multiplying
numerator and denominator by 14, we obtain 42
. We ould then add 1 or
98
41
43
1
1 to the numerator to obtain
or 98 , whi h dier from 37 by 98
. But this
98
means that the dieren e is bigger than 0.01, whi h is worse than before, so
this approa h doesn't give a loser fra tion.
Can we do better than 42
? It is possible that, even though fra tions
99
with smaller denominators are further apart, they an be between some of
the other fra tions that we've looked at, for example between 42
and 37 or
99
between 43
and 37 .
99
to approximate 37 . Let's al ulate
their dieren e, whi h is what we want to minimize:
Solution We want to use the fra tion

p
q





p



 3  =  7p 3q  = |7p 3q|
q


7
7q 
7q

19
What an we do to make this as small as possible? Two approa hes
would be to make the numerator of the dieren e as small as possible or to
make the denominator of the dieren e as large as possible.
Let's fo us initially on the numerator. The numerator annot equal 0
be ause the fra tions pq and 37 are not equal. Thus, the smallest possible
value for the numerator is 1, be ause p and q are integers. So let's try to
nd values of p and q for whi h the numerator equals 1. In this ase, the
1
dieren e equals 7q
whi h is minimized when q is largest.
For the numerator to equal 1, we need 7p 3q = 1. Sin e we also
want to maximize q, we rewrite this as 7p = 3q 1 and work from the largest
possible integer values of q to see when we also get an integer value for p.
If q = 99, the equation be omes 7p = 3(99) 1 = 297 1. Neither
possibility is a multiple of 7.
If q = 98, the equation be omes 7p = 3(98) 1 = 294 1. Neither
possibility is a multiple of 7.
If q = 97, the equation be omes 7p = 3(97) 1 = 291 1. Neither
possibility is a multiple of 7.
If q = 96, the equation be omes 7p = 3(96) 1 = 288 1. Sin e
287 is a multiple of 7, then taking q = 96 and p = 41 gives a dieren e with
numerator 1.


3
1
1
So we have  41
=
=
and this is the smallest possible
96
7
7 96
672
dieren e with the numerator equal to 1.
If the numerator equalled 2 or something larger, then the smallest possible dieren e o urs when the numerator is as small as possible and the
2
denominator is as large as possible, so is 7 299 = 693
. This is the smallest
possible dieren e with numerator at least 2.
1
Combining the ases, the smallest possible dieren e is indeed 672
,
3
and so the losest fra tion to 7 of all of the fra tions under onsideration is
p
41
=
.
q
96
The approximation of fun tions with polynomials is often seen in rstyear university al ulus ourses. As part of these investigations, we learn
how to estimate the amount of error when approximating, for example, sin x
1 5
with x 16 x3 + 120
x . The te hniques used to estimate this type of error
are not dissimilar to what we have seen above, and are very useful in many
types of al ulations.

20

THE OLYMPIAD CORNER

No. 275
R.E. Woodrow

The year has own by, and it has brought many hanges to Crux and
to the Corner. Of ourse it has been overshadowed by the sudden and untimely loss of a great friend and a devoted olleague, Jim Totten, mid-way
through the transition to a new Editor-in-Chief. I think Vazz Linek has done
a wonderful job of stepping in and keeping the journal on tra k with only an
understandable slowing of the produ tion pa e in the interim.
Readers will have noti ed the announ ement at the end of the De ember Corner that Joanne Canape, who has transformed my s ribbles into a
high quality tex le for many years, has de ided that two de ades is enough.
As I ustomarily begin the year by thanking all those who ontributed to the
Corner in the last year, I would be very remiss not to lead o with sin ere
thanks to Joanne.
It is also appropriate to thank those who submitted problem sets for our
use as well as a spe ial thanks to the dedi ated readers who furnish their ni e
solutions whi h we use. Hoping, as always, that I've not missed someone,
here is the list for the 2008 members of the Corner.
Robert Morewood
Arkady Alt
Andrea Munaro
Miguel Amengual Covas
Vedula N. Murty
Jean-Claude Andrieux
Felix Re io
Houda Anoun
Xavier Ros
Ri ardo Barroso Campos
D.J. Smeenk
Mi hel Bataille
Babis Stergiou
Jose Luis Daz-Barrero
Daniel Tsai
J. Chris Fisher
Panos E. Tsaoussoglou
Kipp Johnson
Georey A. Kandall
George Tsapakidis
Jan Verster
Ioannis Katsikis
Edward T.H. Wang
R. Laumen
Salem Maliki
Luyan Zhong-Qiao
Li Zhou
Pavlos Maragoudakis
Titu Zvonaru
Our apologies to Svetoslav Sav hev for the misspelling of his name in
the De ember 2008 Olympiad.
For your problem solving pleasure in the new year we start o with the
problems of the German Mathemati al Olympiad, Final Round, 2006. My
thanks go to Robert Morewood, Canadian Team Leader to the 47th IMO in
Slovenia 2006, for olle ting them for our use.

21
German Mathemati al Olympiad
Final Round, Grades 12{13
Muni h, April 29 { May 2, 2006

First Day

1. Determine all positive integers n for whi h the number

zn = 101
101}
| {z
2n+1 digits

is a prime.

2. Five points are on the surfa e of a sphere of radius 1. Let amin denote
the smallest distan e (measured along a straight line in spa e) between any
two of these points. What is the maximum value for amin , taken over all
arrangements of the ve points?
3. Find all positive integers n for whi h the numbers 1, 2, 3, . . . , 2n an be
oloured with n olours in su h a way that every olour appears twi e and
every number 1, 2, 3, . . . , n appears exa tly on e as the dieren e of two
numbers with the same olor.
Se ond Day

4. Let D be a point inside the triangle ABC su h that AC AD 1 and

BC BD 1. Prove that EC ED 1 for any point E on the side AB .
5. Let x be a nonzero real number satisfying the equation ax2 + bx + c = 0.
Furthermore, let a, b, and c be integers satisfying |a| + |b| + |c| > 1. Prove
that
1
|x|
.
|a| + |b| + |c| 1

6. Let a ir le through B and C of a triangle ABC interse t AB and AC in

Y and Z , respe tively. Let P be the interse tion of BZ and CY , and let X
be the interse tion of AP and BC . Let M be the point that is distin t from
X and on the interse tion of the ir um ir le of the triangle XY Z with BC .
Prove that M is the midpoint of BC

Our se ond problem set for this number is a set of sele ted problems
from the Thai Mathemati al Olympiad Examinations 2005. Again, thanks
go to Robert Morewood, team leader to the 47th IMO in Slovenia 2006, for
olle ting them for the Corner.

22
Thai Mathemati al Olympiad Examinations 2005
Sele ted Problems

1. Let P (x), Q(x), and R(x) be polynomials satisfying

2xP x3

+ Q x x2

Show that x 1 is a fa tor of P (x) Q(x).

1 + x + x2 R(x) .

2. Find all fun tions f : R R su h that

f x + y + f (xy) = f f (x + y) + xy

for all x, y R.
3. Let a, b, and c be positive real numbers. Prove that
1 +

3
6

ab + bc + ca
a+b+c

4. Let n be a positive integer. Prove that n(n + 1)(n + 2) is not a perfe t

square.
5. Find the least positive integer n su h that 2549| n2545 2541 .
6. Do there exist positive integers x, y, and z su h that

2548x + (2005)y = (543)z ?

7. Show that there exist positive integers m and n su h that gcd(m, n) = 1

and 2549| ((25 49)m + 25n 2 49n ).
8. The median AM of a triangle ABC interse ts its in ir le at K and L.
The lines through K and L parallel to BC interse t again at X and Y ,
respe tively. The lines AX and AY interse t BC at P and Q. Prove that
BC = CQ. (Shortlist 2005)
9. Let ABC be an a ute-angled triangle with AB 6= AC , let H be its
ortho entre and M the midpoint of BC . Points D on AB and E on AC
are su h that AE = AD and D, H , and E are ollinear. Prove that HM is
orthogonal to the ommon hord of the ir um ir les of triangles ABC and
ADE . (Shortlist 2005)
10. Assume ABC is an isos eles triangle with AB = AC . Suppose that
P is a point on the extension of side BC . X and Y are points on lines AB
and AC su h that P XkAC and P Y kAB . Let T be the midpoint of ar BC .
Prove that P T XY . (Iran 2004)

23
As a third set of problems we give the 46th Ukrainian Mathemati al
Olympiad Final Round 2006 - 11th form problems. Again, thanks go to Robert
Morewood, team leader to the 47th IMO in Slovenia 2006, for olle ting them
for our use.
46th Ukrainian Mathemati al Olympiad 2006
Final Round

th

11

Form

1. (V.V. Plakhotnyk) Prove that for any rational numbers a and b the graph
of the fun tion
f (x) = x3 6abx 2a3 4b3 , x R
has exa tly one point in ommon with the x-axis.
2. (O.A. Sarana) A ir le is divided into 2006 equal ar s by 2006 points.
Baron Mun hausen laims that he an onstru t a losed polygonal urve
with the set of verti es onsisting of these 2006 points su h that amongst its
2006 edges there are no two whi h are parallel to ea h other. Is his laim
true or false?
3. (T.M. Mitelman)
(a) Prove that for any rational number (0, 1)
exists
an innite
 there


set of real numbers that satisfy the equation x x{x} = and any
two of them have the same fra tional part. (The fra tional part of a real
number a is given by {a} = a a, where a is its integer part, that
is, the greatest integer that does not ex eed a.)
(b) Prove that for any rational number (0, 1)
exists
an innite
 there


set of real numbers that satisfy the equation x x{x} = and any
two of them have dierent fra tional parts.
4. (V.A. Yasinskiy) Two ir les 1 and 2 interse t ea h other at two distin t
points A and B . The tangent line of the ir le 1 at the point A and the
tangent line of the ir le 2 at the point B meet at point C . The rst of
these two lines interse ts the ir le 2 for the se ond time at point T 6= A.
The point X (distin t from A and B ) is on the ir le 1 , and the line XA
interse ts the ir le 2 for the se ond time at point Y (distin t from A). The
lines Y B and XC meet at point Z . Prove that T Z is parallel to XY .
5. (O.O. Kur henko) Prove that for any real numbers x and y
| cos x| + | cos y| + | cos(x + y)| 1 .
6. (T.M. Mitelman) Find all fun tions f : R R su h that
f x3 + y 3

for all real numbers x and y.

= x2 f (x) + yf (y 2 )

24
7. (V.A. Yasinskiy) A point M lies inside a ube ABCDA1 B1 C1 D1 . Points
A , B , C , D , A1 , B1 , C1 , and D1 belong to the rays M A, M B , M C ,
M D , M A1 , M B1 , BC1 , and M D1 respe tively. Prove that if the polyhedron A B C D A1 B1 C1 D1 is a parallelepiped (that is, all of its fa es are
parallelograms), then it is a ube.
8. (V.A. Yasinskiy) There are n 3 soldiers in aptain Petrenko's squad,
no two of the same height. The aptain orders them to stand single-le (not
ne essarily sorted by height). A \wave" is any subsequen e of (not ne essarily next to ea h other) soldiers in this line su h that the rst (leftmost) soldier
in the wave is taller than the se ond soldier in it, but the se ond soldier in it
is shorter than the third one, who is in turn taller than the fourth one, and
so on. (For example, if n = 9, the soldiers are enumerated by height, and
the aptain lines them up as 9, 3, 5, 7, 1, 2, 6, 4, 8 then a longest wave for
this line-up is 9, 3, 7, 1, 6, 4, 8. However, if the aptain lines them up as
1, 2, 3, 4, 5, 6, 7, 8, 9, then a longest wave onsists of (any) one soldier.)
For every n, onsider the number of possible lines with the longest waves of
even lengths and the number of possible lines with the longest waves of odd
lengths. Whi h of these numbers is bigger?

Continuing with problems for readers to solve we give the Cze h-PolishSlovak Mathemati s Competition written on June 26-28, 2006 at Z ilina,
Slovakia. Thanks again go to Robert Morewood, Canadian team leader to
the 47th IMO in Slovenia 2006, for olle ting them for our use.
Cze h-Polish-Slovak Mathemati s Competition 2006
1. Five distin t points A, B , C , D, and E lie in this order on a ir le of
radius r and satisfy AC = BD = CE = r. Prove that the ortho entres
of the triangles ACD, BCD, and BCE are the verti es of a right-angled
triangle.
2. There are n hildren sitting at a round table. Erika is the oldest among
them and she has n andies. No other hild has any andy. Erika distributes
the andies as follows. In every round, all the hildren with at least two
andies show their hands. Erika hooses one of them and he/she gives one
andy to ea h of the hildren sitting next to him/her. (So in the rst round
Erika must hoose herself to begin the distribution.) For whi h n 3 is it
possible to redistribute the andies so that ea h hild has exa tly one andy?
3. The sum of four real numbers is 9 and the sum of their squares is 21.
Prove that these four numbers an be labelled as a, b, c, and d so that the
inequality ab cd 2 holds.

25
4. Prove that for every positive integer k there is a positive integer n su h
that the de imal representation of 2n has a blo k of exa tly k onse utive
zeros, that is, 2n = a00 0b , where a and b are nonzero digits with
k zeros between them.
5. Find the number of integer sequen es (an )
n=1 su h that an 6= 1 and
an+2 =

an + 2006
an+1 + 1

for every positive integer n.

6. Is there a onvex pentagon A1 A2 . . . A5 su h that for ea h i the lines
and Ai+1 Ai+2 interse t in Bi and the points B1 , B2 , . . . , B5 are
ollinear? (By onvention A6 = A1 , A7 = A2 , and A8 = A3 .)

Ai Ai+3

Our nal problem set for this issue is the XXI Olimpiadi Italiano della
Matemati a, Cesenati o, written 5 May 2006. Thanks again go to Robert
Morewood, Canadian team leader to the 47th IMO in Slovenia, for olle ting
them for our use.
XXI Olimpiadi Italiano della Matemati a
Cesenati o
May 5, 2006

1. Rose and Savino play a game with a de k of traditional Neapolitan playing

ards whi h onsists of 40 ards of four dierent suits, numbered 1 to 10. At
the start ea h player has 20 ards. Taking turns, one shows a ard on the
table. Whenever some ards on the table add to exa tly 15, these are then
removed from the game (if the sum 15 an be obtained in more than one way,
the player who last moved de ides whi h ards adding to 15 to remove). At
the end of the game only one ard, a 9, is left on the table. Savino holds two
ards numbered 3 and 5, and Rose holds one ard. What is the number of
Rose's ard?
2. Find all values of m, n, and p su h that
pn + 144 = m2 ,

where m and n are positive integers and p is a prime number.

3. Let A and B be two points on a ir le su h that AB is not a diameter.
Let P be a point on dierent from A and B , and let H be the ortho entre
of the triangle ABP . Find the lo us of H as P varies over all points of
dierent from A and B .

26
4. On an innite hessboard all the
positive integers are written in as ending order along a spiral, starting
from 1 and pro eeding anti lo kwise;
a portion of the hessboard is shown
in the gure.
A "right half-line" of the hessboard is the set of squares given by
a square C and by all squares in the
same row as C and to the right of C .

.......................................................................................................................................................................................
... 17 ...... 16 ...... 15 ...... 14 ...... 13 ...
......................................................................................................................................................................
...... ..... ..... ..... .... ....
..... 18 ..... 5 .... 4 .... 3 ..... 12 ....
................................................................................................................................
...... ...... ..... ...... ..... .....
..........19
... 6 ... 1 .. 2 ... 11 .
... ..............................................................................................................................................
...... 20 ....... 7 ....... 8 ....... 9 ....... 10 ......
.................................................................................................................................................................
...... 21 ........ 22 ........ 23 ........ 24 ........ 25 ........
...................................................................................................................................................................................

(a) Prove that there exists a right half-line none of whose squares ontains
a multiple of 3.
(b) Determine if there exist innitely many pairwise disjoint right half-lines
none of whose squares ontains a multiple of 3.
5. Consider the inequality
(x1 + + xn )2 4(x1 x2 + x2 x3 + + xn x1 ) .

(a) Determine for whi h n 3 the inequality holds true for all possible
hoi es of positive real numbers x1 , x2 , . . . , xn .
(b) Determine for whi h n 3 the inequality holds true for all possible
hoi es of any real numbers x1 , x2 , . . . , xn .
6. Albert and Barbara play a game. At the start there are some piles of oins
on the table, not all ne essarily with the same number of oins. The players
move in turn and Albert starts. At ea h turn a player may either take a oin
from a pile or divide a pile into two piles with ea h pile ontaining at least
one oin (a player may exer ise only one of these options).
The one who takes the last oin wins the game. In terms of the number
of piles and the number of oins in ea h pile at the start, determine whi h of
the players has a winning strategy.

Now we turn to our le of solutions from the readers to problems from

the Mar h 2008 number of the Corner and the Estonian IMO Sele tion Contest 2004-2005, given at [2008 : 79-80.
3. Find all pairs (x, y) of positive integers satisfying (x + y)x = xy .

Solution by Konstantine Zelator, University of Toledo, Toledo, OH, USA.

We show that there are exa tly two su h pairs, (x, y) = (2, 6), (3, 6).
We will make use of two basi fa ts from elementary number theory.

27
(a) If a, b Z and gcd(a, b) = 1 , then gcd(am , bn ) = 1 for any positive
integers m and n.
(b) If a|b, a is a positive integer, b Z, and gcd(a, b) = 1, then a = 1.

x and y be positive integers that satisfy (x + y)x = xy . Write

d = gcd(x, y) and let x = dx1 , y = dy1 , where x1 and y1 are positive
integers that are relatively prime. Substituting for x and y yields

Let

dx1 d (x1 + y1 )x1 d = x1y1 d dy1 d .

(1)

dd(x1 y1 ) (x1 + y1 )x1 d = xy11 d .

(2)

We make ases by omparing the sizes of x1 and y1 .

Case 1. Suppose that x1 = y1 . Sin e gcd(x1 , y1 ) = 1, we have x1 = y1 = 1.
Thus, equation (1) be omes dd 2d = dd , whi h is impossible sin e d 1.
Case 2. Suppose that y1 < x1 . Then x1 y1 is a positive integer and from
equation (1) we obtain
Sin e gcd(x1 , y1 ) = 1 it follows
that gcd(x1 + y1 , x1 ) = 1. By (a) above, we


have gcd (x1 + y1 )x d , xy1 d = 1. However, by equation (2), the positive
integer (x1 +y1 )x d is a divisor of xy1 d . Sin e these two integers are relatively
prime, it follows by (b) that (x1 + y1 )x d = 1, whi h is impossible sin e
x1 + y1 2 and x1 d 1.
Case 3. Suppose that x1 < y1 . From (1) we obtain
1

(x1 + y1 )x1 d = x1y1 d dd(y1 x1 ) .

(3)


Sin e gcd(x1 , y1 ) = 1 we have gcd (x1 + y1 )x1 d , xy11 d = 1 and from equation (3) we see that xy11 d is a divisor of (x1 + y1 )x1 d , whi h implies that
xy11 d = 1. Sin e y1 d is a positive integer this means that x1 = 1. Going ba k
to equation (3) we see that (1 + y1 )d = dd(y1 1) , hen e
1 + y1 = dy1 1 .

(4)

Note that d 6= 1; otherwise equation (4) be omes 1 + y1 = 1, ontrary to the

fa t that y1 is a positive integer. Thus, d 2. Sin e y1 = 1 does not satisfy
equation (4), we also have y1 2. Setting k = y1 1 equation (4) then
be omes k + 2 = dk . By Indu tion (or the Binomial Theorem) we obtain
2k > k + 2 for all integers k 3. Sin e dk 2k , it follows from k + 2 = dk
that k = 1 or k = 2.
For k = 2 we have 4 = d2 , hen e d = 2. From 2 = k = y1 1 we then
have y1 = 3. Re all that x1 = 1. Going ba k, we have x = x1 d = 1 2 = 2
and y = y1 d = 3 2 = 6. This is the solution (x, y) = (2, 6).
Similarly, for k = 1 we have d = 3. Then y1 = k + 1 = 2 and sin e
x1 = 1 we obtain x = dx1 = 3 1 = 3 and y = dy1 = 3 2 = 6. This is the
other solution (x, y) = (3, 6).

28
4. Find all pairs (a, b) of real numbers su h that all roots of the polynomials
6x2 24x 4a and x3 + ax2 + bx 8 are non-negative real numbers.

Solution by Titu Zvonaru, Comane

 sti, Romania.
Let 1 , 2 , and 3 be the roots of the polynomial x3 + ax2 + bx 8, so
that x3 + ax2 + bx 8 = (x 1 )(x 2 )(x 3 ). Comparing oe ients
yields 1 + 2 + 3 = a and 1 2 3 = 8. Sin e 1 , 2 , and 3 are
nonnegative real numbers, by the AM-GM Inequality we have
1 + 2 + 3 3

p
3

1 2 3 ,

hen e a 6 or a 6. The equation 6x2 24x 4a = 0 has real roots

if and only if 242 4 (4a) 6 0, whi h implies 242 + 24 4a 0 and
hen e a 6. Therefore, a = 6.

Now we have 1 + 2 + 3 = 6 = 3 1 2 3 , from whi h it follows

that 1 = 2 = 3 = 2 and b = 12. Thus, the only pair satisfying the
ondition is (a, b) = (6, 12).
3

Next we turn to a solution to a problem of the Trentieme


Olympiad
Mathematique

Belge Maxi Finale, Mer redi 20 avril 2005 given at [2008 : 80.
3. Dans le triangle ABC , les droites AE et CD sont les bisse tri es
interieures

des angles BAC et ACB respe tivement ; E appartient a BC
et D appartient a AB . Pour quelles amplitudes de l'angle ABC a-t-on
ertainement
(b) |AD| + |EC| > |AC|?
(a) |AD| + |EC| = |AC|?
( ) |AD| + |EC| < |AC|?

Solution by Titu Zvonaru, Comane

 sti, Romania
As usual let a = BC , b = CA,
and c = AB . The rst equation below
is the Angle Bise tor Theorem; the following equations are equivalent to it:
BE
EC
BE
EC
BE + EC
EC

EC

Similarly, AD =

=
=
=
=
bc
.
a+b

AB
;
AC
c
;
b
b+c
b
ab
b+c

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

;
.

29
We have
|AD| + |EC| |AC| =
= b

bc
a+b

ab
b+c

c(b + c) + a(a + b) (a + b)(b + c)

(a + b)(b + c)

By the Law of Cosines a2 + c2 b2 = 2ac cos B , so the equation above an

be rewritten as
|AD| + |EC| |AC| =

2abc cos(B)

1
2

(a + b)(b + c)

Hen e,
|AD| + |EC| = |AC|
|AD| + |EC| > |AC|
|AD| + |EC| < |AC|

ABC = 60 ,
ABC < 60 ,
ABC > 60 .

Next we turn to solutions from our readers to problems of the 2005

Vietnam Mathemati al Olympiad given at [2008 : 81.
1. Find the smallest and largest values ofthe expression
P = x + y , where
x and y are real numbers satisfying x 3 x + 1 = 3 y + 2 y .

Solved by Arkady Alt, San Jose, CA, USA; and Mi hel Bataille, Rouen, Fran e.
We give the solution of Bataille.

and
Pmax = 9 + 3 15.
We show that Pmin = 21 9 + 3 21


First, x = 1 and y = 12 11 + 3 21 satisfy the onstraint equation

2

x3 x + 1 = 3 y + 2y (easily he ked using 10+2 21 =

3+ 7 )

and P = 12 9+3 21. Similarly, for x = 12 10+3 15 , y = 21 8+3 15 ,
we have P = 9 + 3 15 and the onstraintis satised. Now, let x and y
satisfy the onstraint equation. Then P = 3 x + 1 + 3 y + 2, so that


p
P 2 = 9 P + 3 + 2 (x + 1)(y + 2)

(1)

It follows that P 0 and P 2 9P 27 0. Thus, P is not less than

2
the positive solution

of the quadrati x 9x 27 = 0 and we dedu e that
1
P 2 9 + 3 21 . From the AM-GM Inequality and (1), we obtain
P 2 9(P + 3 + x + 1 + y + 2) = 9(2P + 6) = 18P + 54 ,

P 2 18P 54 0, whi h implies that P 9 + 3 15. The proof

or
omplete.

is

30
4. Find all
real-valued fun tions f dened on R that satisfy the identity
f f (x y) = f (x)f (y) f (x) + f (y) xy .

Solved by George Apostolopoulos, Messolonghi, Gree e; Mi hel Bataille,

Rouen, Fran e; and Daniel Tsai, student, Taipei Ameri an S hool, Taipei,
Taiwan. We give the write up of Bataille.
It is readily he ked that the fun tion f (x) = x for all x is a solution.
We show that there are no other solutions. Let f satisfy


f f (x y) = f (x)f (y) f (x) + f (y) xy

for all x, y R and let a = f (0). Taking x = y

and then taking y = x shows that

=0

(1)

in (1) gives f (a) = a2

f (x)2 = x2 + a2
2
2
4
2
for all real
x. In parti ular f (a)
= 2a , that is, a = 2a , hen e
numbers
a {0, 2, 2}. Assume
that a = 2. Then for any given x we have
f (x) = x2 + 2 or f (x) = x2 + 2 and taking y = 0 in (1), we obtain

2 1 f (x) +
2.
(2)

Now, f 2 = f (a) = a2 = 2 so that f f 2 = f (2) = 6, but then

2
taking x = 2 in (2) yields
a ontradi tion sin e 6 6= 3 2 2 . Similarly,

the assumption a = 2 leads to a ontradi tion. It follows that a = 0,

hen e f (x) = x or f (x) = x for any given
x. However, if f (x0 ) = x0 for
some nonzero real
number x0 , then
f f (x0 ) = f (x0 ) = x0 , while from (1)
we have f f (x0 ) = f f (x0 0) = f (x0 ) = x0 . This is impossible
sin e x0 6= 0, hen e f (x) = x for all real x.


f f (x) =

5. Find all triples of non-negative integers (x, y, n) su h that

(with the onvention 0! = 1).

x! + y!
= 3n
n!

Solved by Daniel Tsai, student, Taipei Ameri an S hool, Taipei, Taiwan; and
Konstantine Zelator, University of Toledo, Toledo, OH, USA. We give the
solution of Tsai.
Let S be the set of all ordered triples (x, y, n) of nonnegative integers
+ y!
su h that x! n!
= 3n , or equivalently x! + y! = 3n n!. For n = 0, it is
seen at on e that there is no orresponding (x, y, n) S . For n = 1, let
(x, y, n) S ; then if x 3 or y 3 we have x! + y! 7 > 3 = 31 1!,
thus x, y < 3 and simple he king yields that {(0, 2, 1), (1, 2, 1), (2, 0, 1),
(2, 1, 1)} S .
Lemma Let (x, y, n) S with n 2. Then x, y n and x > n or y > n.

Proof: If x, y n, then x! + y! 2n! < 3n n!, so x > n or y > n. If

x! + y!
x!
y!
x!
x > n and y < n, then
=
+
= 3n , a ontradi tion sin e
is
n!
n!
n!
n!

31
y!
an integer but n!
is not. Thus, if x > n, then y
y > n we have x n by symmetry.

n.

Similarly, in the ase

We shall prove that for n 2 there is no orresponding (x, y, n)

by onsidering ases on n modulo 3.

Case 1. n 0 (mod 3). Let (x, y, n) S and assume without loss of

generality that x y. By the Lemma, n x y and one of these two
+ y!
inequalities is stri t. If x > n, then from x! n!
= 3n it follows that
(n+1)|3n . However, n+1 has a prime divisor other than 3, a ontradi tion.
Therefore, n = x < y, and onsequently
x! + y!
n!

x!
n!

y!
n!

= 1 + (n + 1)(n + 2) y = 3n .

Thus, 3 divides 1+(n+1)(n+2) y and (n+1)(n+2) y 2 (mod 3),

whi h implies that y = n + 2. However, 1 + (n + 1)(n + 2) < 3n for n 3
(by indu tion) and 1 + (2 + 1)(2 + 2) 6= 32 , ontradi ting the fa t that
(x, y, n) S .
Case 2. n 1 (mod 3). Let (x, y, n) S and assume without loss of
generality that x y. By reasoning similar to that of Case 1 it follows that
x = n and y = n + 1. However, 1 + (n + 1) < 3n for ea h integer n 2,
ontradi ting the fa t that (x, y, n) S .
Case 3. n 2 (mod 3). Let (x, y, n) S and assume without loss of
generality that x y. By the Lemma, n x y and one of these two
+ y!
inequalities is stri t. If x n + 2, then from x! n!
= 3n it follows that
(n + 2)|3n . However, n + 2 has a prime divisor other than 3, a ontradi tion,
hen e n x < n + 2.
If x = n, then n = x < y and
x! + y!
n!

x!
n!

ontradi ting n + 1 0
If x = n + 1, then

y!
n!

= 1 + (n + 1)(n + 2) y = 3n ,

(mod 3).

x! + y!
x!
y!
=
+
= (n + 1) + (n + 1)(n + 2) y = 3n .
n!
n!
n!

If furthermore y = n + 1, then
2(n + 1) = (n + 1) + (n + 1)(n + 2) y = 3n

is even, a ontradi tion. Thus, y > n+1 and (n+1) 1+(n+2) . . . y = 3n .

It follows that 1 + (n + 2) y 0 (mod 3) and (n + 2) y 2 (mod

3),
whi h implies that y = n + 3. However, (n + 1) 1 + (n + 2)(n + 3) 6= 3n

32
for 1 n 5 and by indu tion (n + 1) 1 + (n + 2)(n + 3)
ontradi ting the fa t that (x, y, n) S .

< 3n

for n 6,

6. Let the sequen e x1 , x2 , x3 , . . . , be dened by x1 = a, where a is a real

number, and the re ursion xn+1 = 3x3n 7x2n + 5xn for n 1.
Find all values of a for whi h the sequen e has a nite limit as n tends
to innity, and nd this limit.

Solved by Arkady Alt, San Jose, CA, USA; Mi hel Bataille, Rouen, Fran e;
and Daniel Tsai, student, Taipei Ameri an S hool, Taipei, Taiwan. We give
Bataille's write-up.
Let f (x) = 3x3 7x2 + 5x and g(x) = f (x) x = x(x 1)(3x 4).
The sequen e {xn }, whi h satises xn+1 = f (xn ) for all positive integers n,
an only onverge to a root of g(x) = 0. Thus, the only possible nite limits
of {xn } are 0, 1, and 43 . We show that the sequen e is onvergent if and only
if 0 a 43 , in whi h ase the limit is 1 ex ept if a = 0 and n
lim xn = 0 or
4
4
if a = 3 and n
lim xn = 3 .
Suppose rst a < 0. Sin e g(x) < 0 when x < 0, it follows that
xn < x1 = a < 0 for all positive integers n. If {xn } had a nite limit, ,
we would have a, ontradi ting the fa t that {0, 1, 43 }. Thus, {xn }
is divergent when a < 0. Using the fa t that g(x) > 0 for x > 43 , similar
reasoning shows that {xn } is divergent when a > 43 .
If a {0, 1, 43 }, then the sequen e {xn } is onstant.
If a (1, 34 ), then using f (x) 1 = (x 1)2 (3x 1) an easy indu tion shows that 1 < xn+1 < xn for all positive integers n. Thus, {xn } is
de reasing and bounded, hen e onvergent. Its limit satises 1 and
{0, 1, 43 }, that is, = 1.
If a
[ 13 , 1) then x2 = f (a) 1 and x2 < 43 , as the maximum of f on
< 43 . From the previous ase, we see that lim xn = 1.
[0, 1] is f 59 = 275
243
n
1
1
It remains to study the ase a 0, 13 . Then, 3m+1
a < m for
3
some unique positive integer m. If any of the numbers x2 , x3 , . . . , xm is not
less than 13 , let xk be the one with the smallest index. Then 13 xk < 43 and
by the previous ases {xn }nk onverges to 1 and n
lim xn = 1. Otherwise,


noting that f (x) 3x = x(x 2)(3x 1) is positive for x 0, 13 , we have

x2

f (x1 ) > 3x1 = 3a

x3

f (x2 ) > 3x2

xm

1
3m1

f (xm1 ) > 3xm1

3m

,
1
32

and nally xm+1 > 13 . So {xn }nm+1 onverges to 1 and again n

lim xn = 1.

33
To nish this number of the Corner we give solutions from the readers
to problems of the 2005 German Mathemati al Olympiad, given at [2008 : 82.
1. Determine all pairs (x, y) of reals, whi h satisfy the system of equations
x3 + 1 xy 2 y 2
y 3 + 1 x2 y x2

=
=

0,
0.

Solved by George Apostolopoulos, Messolonghi, Gree e; Konstantine Zelator, University of Toledo, Toledo, OH, USA; and Titu Zvonaru, Comane
 sti,
Romania. We give the write-up of Apostolopoulos.
We subtra t the two equations of the system to obtain
x3 y 3

+ xy(x y) + x2 y 2 = 0 ,

whi h upon fa toring be omes

(x y)(x + y)(x + y + 1) = 0 .

Thus, y = x or y = x or y = x 1.
If y = x, then x = 1, hen e (x, y) = (1, 1) or (x, y) = (1, 1).
If y = x, then again x = 1, hen e (x, y) = (1, 1) or (x, y) = (1, 1).
If y = x 1 we substitute into the rst equation to obtain
x3 + 1 x(x + 1)2 (x + 1)2 = 3x(x + 1) = 0 ,

hen e x = 0 or x = 1 and (x, y) = (0, 1) or (x, y) = (1, 0).

2. Let A, B , and C be three distin t points on the ir le k. Let the lines h
and g ea h be perpendi ular to BC with h passing through B and g passing
through C . The perpendi ular bise tor of AB meets h in F and the perpendi ular bise tor of AC meets g in G. Prove that the produ t |BF | |CG| is
independent of the hoi e of A, whenever B and C are xed.

Solved by Miguel Amengual Covas, Cala Figuera, Mallor a, Spain; Mi hel

Bataille, Rouen, Fran e; Konstantine Zelator, University of Toledo, Toledo,
OH, USA; and Titu Zvonaru, Comane
 sti, Romania. We give the solution of
Amengual Covas.
Let A be the point diametri ally opposite to A. Let M and N be
the midpoints of the segments AB and AC , respe tively. Let B and C
be the se ond points of interse tion of the lines g and h with the ir le k,
respe tively.

34
.
.
....
....
...
....
....
....
....
....
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
...
.
.
.
.
.
.
.
.
.
.
...
.
.
.
..........
...................................................
.
.
.
.
.
.
.
.
.
.
.
.
.
.
....
.
.
.
.
.
.
...
.......
...... ... .... .... ............. ...............
.
.
.
.
.
.
.
.
.
.
.
.
....
.
.
...
.
.
........
........
. ... ...
......
......
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
....
....
.
.
.
.
.
.......
........
. .... ...
......
.....
.
.
.
.
.
.
.
.
.
.
.
...
.
.
.
.
.
.
.
.
....
.
.
.
.
........
......
. ... ...
.....
....
.
.
.
.
.
.
.
.
.
.
.
.
.
.
....
.
.
.
.
.
.
..
........ ....
...
.
...
..
........ ..... ....
...
.... ........ ..................
.
.
.
...
........ ... ..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
........ ... ...
... .... ...........

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

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

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

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

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

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

Sin e C BC and BCB are right angles, the segments BB and CC

are diameters of the ir le k. Thus, the two quadrilaterals AC A C and
ABA B are re tangles, and hen e parallelograms, so we have
|C A | = |CA|

and

|A B | = |AB| .

Sin e the right triangles BM F and A C A are similar, as are the right triangles CN G and A B A, we have
BF
BM
=
=

AA
C A

and
from whi h we obtain

CG
CN
=
=

AA
A B
BF CG =

1
AB
2

CA
1
CA
2

AB

1
|AA |2
4

as the square of the radius of k, whi h is independent of the hoi e of A.

3. A lamp is pla ed at ea h latti e point (x, y) in the plane (that is, x and
y are both integers). At time t = 0 exa tly one lamp is swit hed on. At
any integer time t 1, exa tly those lamps are swit hed on whi h are at a
distan e of 2005 from some lamp whi h is already swit hed on. Prove that
every lamp will be swit hed on at some time.

Solution by Titu Zvonaru, Comane

 sti, Romania.
Assume
that at time t = 0 the lamp at O(0, 0) is swit hed on. Sin e

2005 = 13572 + 14762 then at some time the lamps at the following latti e

35
points will be swit hed on:
A1 (1357, 1476) ,
A2 (3 1357, 1476) ,

..
.

Ak (2k 1) 1357, 1476 ,

O1 (2 1357, 0) ,
O2 (4 1357, 0) ,

..
.

Ok (2k 1357, 0) ;

and then the lamps at these latti e points will be swit hed on:
B1 (2k 1357 2005, 0) ,
B2 (2k 1357 2 2005, 0) ,
B3 (2k 1357 3 2005, 0) ,
..
.
Bt (2k 1357 t 2005, 0) .
The equation 2k 1357 2005t = 1 is the same as 2714k 2005t = 1,
whi h has a solution in positive integers k, t be ause gcd(2714, 2005) = 1,
for example, 2714 1134 2005 1535 = 1. Thus the lamp at (1, 0) will be
swit hed on at some time. It follows (by symmetry) that every lamp will be
swit hed on at some time.
4. LetQ(n) denote thesum of the digits of the positive integer n. Prove


that Q Q Q(20052005 ) = 7.

Solution by Titu Zvonaru, Comane

 sti, Romania.
It is well known that Q(n) n (mod
then k 20052005 (mod 9) and we have
20052005

9).




Q Q(20052005 ) ,

Let k = Q

(2)2005 = 2 22004 = 2 23

2(1)668 = 2 7 (mod 9)

668

so that k 7 (mod 9).

The number
20052005 has at most 4 2005 = 8020 digits. Hen e,


2005
Q 2005
is at most 9 8020 = 72180. This implies that




Q Q 20052005
5 9 = 45




and hen e k = Q Q Q(20052005 ) is at most Q(39) = 12.
Altogether, k is an integer satisfying 0 k 12 and k 7 (mod 9),
hen e k = 7, as desired.

That ompletes the Corner for this issue. Send me your ni e solutions,
generalizations, and Olympiad problem sets.

36

BOOK REVIEWS
Amar Sodhi

Putnam and Beyond

By Razvan Gel a and Titu Andrees u, Springer S ien e + Business Media LLC,
New York, 2007
ISBN-13: 978-0-387-25765-5, soft over, 798+xvi pages, US$69.95
e-ISBN-13: 978-0-387-68445-1
Reviewed by Jeff Hooper, A adia University, Wolfville, NS
One of my favourite problem books is the one by Andrees u and Gel a,
Mathemati al Olympiad Challenges (Birkhauser, 2000), in whi h the authors
olle t numerous problems entred mainly around past Olympiads, and group
them together by similar topi . In Putnam and Beyond, the authors again
ombine to deliver a similar problem book, based this time around on the
Putnam Competition.
Stru turally, the book onsists of six major hapters: Methods of Proof,
Algebra, Real Analysis, Geometry and Trigonometry, Number Theory, and
Combinatori s and Probability. Ea h hapter is divided into numerous se tions and subse tions, ea h of whi h fo uses on an important problem idea.
For instan e, the initial se tion of the Algebra hapter is Identities and
Inequalities, and here we nd the topi divided into a number of important ideas: Algebrai Identities, the equation x2 0, the Cau hy-S hwartz
Inequality, the Triangle Inequality, the AM{GM Inequality, and so on. Ea h
se tion ontains explanations of the key ideas and several worked problems,
along with a number of problems to solve. Full solutions are provided at
the ba k of the book, along with sour es and/or helpful referen es where
ne essary. The initial hapter is a parti ularly ni e introdu tion for students.
There is some overlap, of ourse, with the topi s overed in the authors'
previous book, and I was expe ting there to be mu h in ommon. To their
redit, the authors have avoided that sort of mild heating: for the topi s in
ommon with their rst book, the examples and problems they oer are new.
This new book is also a mu h larger and far more extensive eort. In addition
to all of the examples provided, the book ontains more than 900 problems.
And these are Putnam-level problems, so they are mainly a level up from
the earlier book. In fa t the topi s in the book go very deep, and over most
of the major ideas found in the undergraduate mathemati s urri ulum. So
there's lots here.
The book is not error-free, however, and a tea her or oa h who uses
this book should be a little areful. I ertainly haven't worked through every
problem in the book, but did dip into them in a number of se tions. Just
to give an example, Problem 33 asks the solver, "Given 50 distin t positive
integers stri tly less than 100, prove that some two of them sum to 99." A
little thought shows that one an take the numbers 50, 51, . . . , 99 and the
statement fails.

37
But these problems are of the minor variety. This wonderful book is an
ex ellent problems resour e and should be ome a part of any serious library
for problem solving. By olle ting together problems by topi , the authors
provide readers the han e to study ea h of these important problem-solving
te hniques and ideas in isolation, and help them begin to see the inherent
patterns. This should be one of the rst books onsidered as a resour e by
anyone oa hing groups of problem solvers.
Mathemati al Conne tions: A Companion for Tea hers and Others
By Al Cuo o, Mathemati al Asso iation of Ameri a, 2005
ISBN 978-0-88385-739-7, hard over, 239+xix pages, US$54.95
Reviewed by Peter S. Brouwer, State University of New York, Potsdam,
NY, USA
Al Cuo o is the Dire tor of the Center for Mathemati s Edu ation at
the Edu ation Development Center in Newton, MA, where he works in the
areas of urri ulum development and professional development of tea hers.
This book joins a number of re ent others in addressing se ondary s hool
mathemati s ontent topi s from an advan ed (or deeper) perspe tive. The
primary audien e is high s hool mathemati s tea hers, and the book is based
on the assumption that providing a more advan ed treatment of some of the
mathemati al topi s taught at that level is a valuable form of professional development. The author's emphasis is on making onne tions between topi s
and developing mathemati al habits of mind.
The hoi e of topi s is somewhat idiosyn rati , and re e ts the interrelated topi s that Cuo o is interested in exploring. The hapter titles are:
1. Dieren e Tables and Polynomial Fits, 2. Form and Fun tion: The Algebra of Polynomials, 3. Complex Numbers, Complex Maps, and Trigonometry,
4. Combinations and Lo ks, and 5. Sums of Powers.
The strength of this book is that it is essentially a problems book (on
the above topi s). There are many problems, in luding 90 in the rst hapter alone, and the reader is asked to work them sequentially while reading
through the text. These are grouped by themes, whi h aids oheren e. In
addition, there are many problems given as exer ises. The author in ludes
helpful notes on sele ted problems at the end of ea h hapter.
Many of the problems in this book are quite hallenging, but its in remental and themati approa h helps. As polynomial algebra (and patterns in
polynomial oe ients) appear in every hapter, the reader must be omfortable manipulating rather ompli ated algebrai expressions. I would re ommend this book for serious, mathemati s-based professional development
programs as well as for experien ed independent readers who would enjoy
pursuing a fruitful intelle tual journey through sele ted advan ed se ondary
mathemati s topi s.

38
Velo ity Analysis: an Approa h to Solving
Geometry Problems

Peng YuChen

Introduction

We introdu e velo ity analysis for solving otherwise omplex geometry problems, then we give two examples of the use of this method: (1) to give a very
brief proof of the opti al property of the ellipse, and (2) to nd the length of
a logarithmi spiral. At the end of this note we leave some problems for the
reader; the solutions will be very brief if velo ity analysis is used.

Velocity Analysis

In analyti geometry, a urve is a set of points whose oordinates (x, y) satisfy a ertain formula. For example, the set of points {(x, y) : x2 + y2 = R2 }
is a ir le or radius R.
As we all know, the tra e of a moving point is a urve. In the analyti
notion of a urve above, we mainly pay attention to the position of a moving
point, rather than its velo ity. In fa t, either the position or the velo ity of
a moving point an des ribe the ourse of its motion, and in some ases it
is more onvenient to study the velo ity rather than the position. If instead
of studying the oordinates of a moving point we study its velo ity, then
it is very simple for us to dedu e ertain geometri properties of the tra ed
urve without engaging in omplex mathemati s, espe ially when seeking the
tangent or ar length. (As we know, the velo ity ve tor of a moving point
gives a tangent to the urve tra ed by the point.)
Now we introdu e the basi idea of velo ity analysis in solving geometry
problems.

A point P is moving under a ertain restri tion on its velo ity V . For

example, if V OP 0 (here OP is the position ve tor of P ), then obviously

the point P tra es out a ir le whose entre is O. That is to say, we nd an
equivalent way of dening the ir le by studying the velo ity rather than the
oordinates of a moving point.
When studying velo ity, usually it is more onvenient to study omponents. In our example of the ir le, we break the velo ity of P into two

dire tions: along OP and orthogonal to OP . Name the two omponents V1

Copyright

c 2009

Canadian Mathemati al So iety

Crux Mathemati orum with Mathemati al Mayhem, Volume 35, Issue 1

39

and V2 respe tively; if the omponent V1 along OP satises V1 0, then

the point P tra es out a ir le.
This is the basi idea of velo ity analysis: if we de ompose the velo ity
of a moving point into two appropriate dire tions, giving ertain restri tions
to the omponents of the velo ity, the tra e of the moving point will be ome
a ertain urve. In this setting it is very easy to analyse the tangent to the
urve, for we already know the restri tion on the velo ity of the point.
Example 1 The opti al property of the ellipse. To show the power of the ve tor analysis approa h, we use it to solve this lassi al geometry problem. In
an ient times people found that oni se tions have very spe ial and beautiful opti al properties. One example is this: if a ray of light leaves one fo us
of an ellipse and strikes the ellipse, it will be re e ted to the other fo us of
the ellipse (see Figure 1).
L1

...

.
..........

..
................................................................................
.
.
.
.
.
.
.
.
.
.
..
... ....
....
..... ...............
......
..
...
.. ......
......
....
.........
...
.............
.
.
.
...
.
.
.
......
.
.
...
.....
.
.
.
.
.
.
.
.
.
.
.
..
.....
..
.
.
.
.
.
.
.
..
.
.
.
.
...
.
...
........
.....
...
........
.
.
.
.
.
.
.
.
.
.
.
....................
................................r..........................................................................................................................................................r.................................................
.
... F1
F2 ..
.....
...
....
.
....
.
.
......
....
.
.....
.
.
.
........
.
...
............... ......... .....................
.............................

.
.........

....
.

Figure 1

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

...........................................
.............
........
.
.
.
.
.
.
.
..
....
.
.
.......
.
.
...
....
..
... r
r ....
...
... F1
F2 ...
....
.
......
....
.
.
.
.
........
.
...............
.......
........................................

Figure 2

Using al ulus to prove this property would be ompli ated, so we use

the approa h of analysing the velo ity of a moving point to give a mu h
shorter proof. As in Figure 2, to prove the opti al property of the ellipse,
we need to prove that = , whi h is enough to satisfy the Law of Re e tion. Here L1 is tangent to the ellipse.
We usually dene an ellipse like
this: two points F1 and F2 in the plane

V2
are xed. A point P moves so that

.........
V
..
|P F1 |+|P F2 | is always a onstant. We

all the tra e of P an ellipse. In a or................................r..P

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

dan e with the idea of velo ity analy........

....
......
.
.
sis, we des ribe an ellipse as follows:
......
.
....
...
.
.........
...
V1
Denition. Two points F1 , F2 in the
...
...
r
r
.
...
plane are xed (see Figure 3). A mov.
.
... F1

F2 ..
ing point, P , has velo ity ve tor V .
....
.

......
....
.
.
.
.
Let V1 and V2 be the omponents of V
........
.
.......
................
towards F1 and away from F2 , respe ......................................

tively. If |V1 | |V2 |, then the tra e of

Figure 3
P is an ellipse.
..
..........

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

40
It is obvious that the two denitions are equivalent. For when P moves
towards F1 a short distan e, it must then move away from F2 by that same

distan e to ensure that |P F1 | + |P F2 | is always a onstant.

Now V is a tangent ve tor of the ellipse. Sin e |V1 | |V2 |, then = .

Sin e = , we also have = . That is all.
Example 2 The length of the logarithmi spiral. As another example, we nd

the length of the logarithmi spiral = 0 ea between = 1 and = 2 .

Usually we annot gure out this problem without using a lot of al ulus, so
we introdu e a physi al model.
As in Figure 4, three
points A, B , and C are ea h at a vertex of an equilateral triangle of side 30 . Point A moves towards B , B moves towards
C , and C moves towards A. The speed of ea h point is s. Where will the
three points meet and how far will they travel before meeting?

V2........ A

....

Figure 4

....

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

O
r

....

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

....

..
.
.
.
...

O
r

...
...
...
.

...........................................
..
.. .. ..
.. ... ..
.
.. ..... .... .....
. . .. ..
... .... ..... .....
...
.. ..
...
...
....
.....
...
.....................
...
..
.
...
.
.
.
1 .......
...
.
...
..
.
...
..
...
...
...
...
.
...
.
..
...
.
.
...
..
.
.
...
.
.
.
...
...
...
.
...
..
.
...
..
.
...
.
.
.
...
...
...
.
...
..
.
...
.
.
.
...
..
.
.
...
...
...
.
.
...
.
.
...
..
.
.
.
...........................................................................................................................................................................................................

...
...
...
.

....

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

....

..
.
.
.
...

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

Figure 5

The solution of this problem is easy: obviously the three points will
meet at the point O, whi h is the entre of the triangle. We de ompose the

velo ity V of A at ea h instant into two omponents: one towards the point

O and the other perpendi ular to AO , and we all these two

omponents V1

and V2 , respe tively. We have that |V1 | = | V | cos 30 = 23 s (see Figure 5),
0
0
so the time until they meet is t =
and the distan e traveled
=
is

2
d = st = 0 .
3

|V 1 |

( 3/2)s

Now we onsider the urve tra ed by A. We use a polar oordinate

system with the origin at O and with the ve tor V2 pointing in the dire tion

of in reasing . At ea h instant, |V2 | = | V | sin 30 = |V1 | tan 30 . Sin e

d
|V 2 | =
and |V1 | = d
, we have d = ad, the dierential equation
dt
dt

of the spiral = 0 ea with a = cot 150 = 3. Thus, the length of

2 |
2
the urve from = 1 to = 2 is |1
s = |1 2 |. We leave

|V 1 |

41

........

the problem of nding a suitable physi al model for other values of a to the
reader.
We leave three more problems for the reader to
...
Figure 6
solve. They ould be solved by using analyti ge...
ometry and al ulus, but are more onveniently
.
solved using the approa h of velo ity analysis.
..... ..... A
1. The pursuant traje tory problem. As in Figure
.. ..r 1
.
.
6, suppose that an obje t A starts from the point
..
.
(0, 1), and moves with a onstant speed s in the
.
B .
dire tion of the positive y{axis. At the same time
..
.
.
another obje t B starts from the point (1, 0),
.r.....
and moves with speed 2s and always in the dire 0
1
tion of the obje t A. When will the obje ts meet?
2. The opti al property of a hyperbola. If a light ray leaves one fo us of a
hyperbola and strikes the hyperbola, then the (reverse) extended line of its
re e tion will pass through the other fo us of the hyperbola (see Figure 7).
3. The opti al property of a parabola. If a light ray leaving the fo us is re e ted in the parabola, then its re e tion is parallel to the axis of symmetry
(see Figure 8).

.......

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

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

F2

Figure 7

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

F1...

..
...............
.
.
.
.
.
.
.
.
.
.
.
.
.
.........
............
.
.
.
.
.
.
.
.
.
......
........
.
.
.
.
.
..
.....
...
... r
... F
.....
.......
........
.........
...........
............
.............. Figure 8
................
......

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

.........
.

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

...... ...
.......... ..
.......... ...
. ....... ..
........
......
........ ....
..... ....
... .. .... ....
... .. .... ..
... .. . ...
...
... ..
...
... ..
.
..........................................................................r.................................................
................................................................r
.

Referen es

[1 Cai Suilin, Ordinary Dierential Equations, ZheJiang University Press,

1988, pp. 46-48.
[2 R.T. Coman and C.S. Ogilvy, The \Re e tion Property" of the Coni s,
Mathemati s Magazine, Vol. 36, No. 1 (Jan., 1963), pp. 11-12.
Peng YuChen
Lantian-1 4083
Zi Jingang Campus, ZheJiang University
388 Yu Hangtang Road, Hangzhou, China
007-sunboy@163.com

42

PROBLEMS
Solutions to problems in this issue should arrive no later than 1 August 2009.
An asterisk () after a number indi ates that a problem was proposed without a
solution.
Ea h problem is given in English and Fren h, the o ial languages of Canada.
In issues 1, 3, 5, and 7, English will pre ede Fren h, and in issues 2, 4, 6, and 8,
Fren h will pre ede English. In the solutions' se tion, the problem will be stated in
the language of the primary featured solution.
The editor thanks Rolland Gaudet of the University College of Saint Bonifa e
and Jean-Mar Terrier of the University of Montreal for translations of the problems.
3401. Proposed by Tigran Sloyan, Basi Gymnasium of SEUA, Yerevan,
Armenia.
Let ABCDE be a onvex pentagon su h that BAC = EAD and
BCA = EDA, and let the lines CB and DE interse t in the point F .
Prove that the midpoints of CD, BE , and AF are ollinear.
3402. Proposed by Mihaly Ben ze, Brasov, Romania.
Let D and E be the midpoints of the sides AB and AC in triangle
ABC , respe tively. Prove that CD is perpendi ular to BE if and only if
5BC 2 = AC 2 + AB 2 .

3403. Proposed by D.J. Smeenk, Zaltbommel, the Netherlands.

The ir les 1 and 2 interse t at P and Q. A line through P interse ts
1 and 2 for the se ond time at A and B , respe tively. The tangents to
1 and 2 at A and B interse t at C . If O is the ir um entre of ABC
determine the lo us of O when rotates about P .
3404. Proposed by Mi hel Bataille, Rouen, Fran e.
Let Q be a y li quadrilateral. The perpendi ulars to ea h diagonal
through its endpoints form a parallelogram, P . Chara terize the entre of P
and show that opposite sides of Q interse t on a diagonal of P .
3405. Proposed by Mi hel Bataille, Rouen, Fran e.
Find the minimum value of
| cos | + | cos | + | cos | + | cos( )| + | cos( )| + | cos( )| ,

where , , and are real numbers.

43
3406. Proposed by Jose Luis Daz-Barrero and Miquel Grau-San hez,

Universitat Polite ni a

de Catalunya, Bar elona, Spain.
Find
"

#
n
1 Y

lim ln

2n

2+

k=1

n2

3407. Proposed by Roy Barbara, Lebanese University, Fanar, Lebanon.

Let S be a set of positive integers ontaining the integer 2007 and su h
that

(a) If x, y S and x 6= y, then |x y| S , and

(b) If x S , then

x3 1007x + 3007 S .

Prove that S is the set of all positive integers.

3408. Proposed by Slavko Simi , Mathemati al Institute SANU, Belgrade,

Serbia.
Let {ci }
i=1 be a sequen e of distin t positive integers, and let |q| < 1.
Prove that the inequality

ci q ci

i=1

1+

q ci

q
1q

i=1

holds for all su h sequen es {ci }

i=1 if and only if q

 1
0, 2 .

3409. Proposed by Jose Luis Daz-Barrero, Universitat Polite ni a


de
Catalunya, Bar elona, Spain.
Let a, b, c, and d be positive real numbers. Prove that
ab + bc + ca
ab + bd + da
ac + cd + da
bc + cd + db
+ 3
+ 3
+ 3
a 3 + b3 + c3
a + b3 + d3
a + c3 + d3
b + c3 + d3

min

a 2 + b2
c2 + d2 a 2 + c2
b2 + d2 a 2 + d2
b2 + c2
+
,
+
,
+
3/2
3/2
3/2
3/2
3/2
(ab)
(cd)
(ac)
(bd)
(ad)
(bc)3/2

3410. Proposed by Joe Howard, Portales, NM, USA.

Let a, b, and c be the sides of triangle ABC , let R be its ir umradius,
and let F be its area. Prove that
X bc sin2 A/2

y li

b+c

F
2R

44
3411. Proposed by Mihaly Ben ze, Brasov, Romania.
Let a, b, and c be positive real numbers su h that
a6 + b6 + c6 <

32
33

a3 + b3 + c3

Prove that at least one of the quadrati s

cx2 + ax + b has no real roots.

2

ax2 + bx + c, bx2 + cx + a,

or

3412. Proposed by Cao Minh Quang, Nguyen Binh Khiem High S hool,
Vinh Long, Vietnam.
Let a, b, and c be positive real numbers su h that abc = 1. Prove that
X

y li

1
1.

3
a + 2b3 + 6

3413. Proposed by Vo Quo Ba Can, Can Tho University of Medi ine and
Pharma y, Can Tho, Vietnam.
Let a, b, c, and d be real numbers in the interval [1, 2]. Prove that
a+b
c+d

c+d
a+b

a+c
b+d

3
2

.................................................................
3401. Propose par Tigran Sloyan, Ly ee
 de la SEUA, Erevan, Armenie.

Soit ABCDE un pentagone onvexe tel que BAC = EAD et
BCA = EDA. Soit F le point d'interse tion des droites CB et DE .
Montrer que les milieux des CD, BE et AF sont olineaires.

3402. Propose par Mihaly Ben ze, Brasov, Roumanie.
Soit le triangle ABC , ou D et E sont les milieux des ot
^ es
 AB et AC ,
respe tivement. Montrer que CD est perpendi ulaire a BE si et seulement
si
5BC 2 = AC 2 + AB 2 .
3403. Propose par D.J. Smeenk, Zaltbommel, Pays-Bas.
Les er les 1 et 2 interse tent a P et Q. Une ligne passant par P
interse te une se onde fois 1 et 2 a A et B , respe tivement. Les tangentes
de 1 et 2 a A et B interse tent a C . Si O est le entre du er le ir ons rit
du ABC determiner

le lieu de O lorsque tourne autour de P .

45
3404. Propose par Mi hel Bataille, Rouen, Fran e.
Soit Q un quadrilatere
 y lique. Les perpendi ulaires a haque diagonale issues de ses sommets forment un parallelogramme

P . Cara teriser

le
^ es
 opposes
 de Q se oupent sur une dia entre de P et montrer que les ot
gonale de P .
3405. Propose par Mi hel Bataille, Rouen, Fran e.
Determiner

la valeur minimum de
| cos | + | cos | + | cos | + | cos( )| + | cos( )| + | cos( )| ,

ou , et sont des nombres reels.



3406. Propose par Jose Luis Daz-Barrero et Miquel Grau-San hez,


Universite Polyte hnique de Catalogne, Bar elone, Espagne.
Cal uler
"
#
n 
1 Y

lim ln

2n

2+

k=1

n2

3407. Propose par Roy Barbara, Universite Libanaise, Fanar, Liban.

Soit S un ensemble d'entiers ontenant l'entier 2007 et tel que
(a) Si x, y S et x 6= y , alors |x y| S , et


(b) Si x S , alors x3 1007x + 3007 S .
Montrer que S est l'ensemble de tous les entiers positifs.
3408. Propose par Slavko Simi , Institut de Mathematiques

SANU,
Belgrade, Serbie.
Soit {ci }
i=1 une suite d'entiers positifs distin ts, et soit |q| < 1. Montrer que l'inegalit

e

ci q ci

i=1

1+

q ci

q
1q

i=1

vaut pour toutes les suites de e type si et seulement si q

 1
0, 2 .

3409. Propose par Jose Luis Daz-Barrero, Universite Polyte hnique de

Catalogne, Bar elone, Espagne.
Soit a, b, c et d des nombres reels
 positifs. Montrer que
ab + bc + ca
ab + bd + da
ac + cd + da
bc + cd + db
+ 3
+ 3
+ 3
a 3 + b3 + c3
a + b3 + d3
a + c3 + d3
b + c3 + d3

min

a 2 + b2
c2 + d2 a 2 + c2
b2 + d2 a 2 + d2
b2 + c2
+
,
+
,
+
(ab)3/2
(cd)3/2 (ac)3/2
(bd)3/2 (ad)3/2
(bc)3/2

46

3410. Propose par Joe Howard, Portales, NM, E-U.
Soit a, b et c les ot
^ es
 du triangle ABC , soit R le rayon de son er le
ir ons rit et F son aire. Montrer que
X bc sin2 A/2
F

b+c
2R
y lique

3411. Propose par Mihaly Ben ze, Brasov, Roumanie.

Soit a, b et c des nombres reels
 positifs tels que
a6 + b6 + c6 <

2
32 3
a + b3 + c3
33

Montrer que au moins une des quadratiques ax2 + bx + c, bx2 + cx + a et

cx2 + ax + b n'a au une ra ine reelle.

3412. Propose par Cao Minh Quang, College
 Nguyen Binh Khiem, Vinh
Long, Vietnam.
 positifs tels que abc = 1. Montrer
Soit a, b et c trois nombres reels
que
X
1
1.
3
3
y lique

a + 2b + 6

3413. Propose par Vo Quo Ba Can, Universite de Mede ine


et Pharma ie
de Can Tho, Can Tho, Vietnam.
Soit a, b, c et d quatre nombres reels
 dans l'intervalle [1, 2]. Montrer
que
a+b
c+d
a+c
3
+

.
c+d

a+b

b+d

47

SOLUTIONS
No problem is ever permanently losed. The editor is always pleased
to onsider for publi ation new solutions or new insights on past problems.
Last year we re eived a bat h of orre t solutions from Steven Karp,
student, University of Waterloo, Waterloo, ON, to problems 3289, 3292, 3294,
3296, 3297, 3298, and 3300, whi h did not make it into the De ember issue
due to being misled. Our apologies for this oversight.
3301. [2008 : 44, 46 Proposed by Ovidiu Furdui, University of Toledo,
Toledo, OH, USA.
Prove that


1
1
1
1
1

ln 2
+
+ +
1 + + +
X
X
n+1
n+2
2n
2
n
=
n
(2n + 1)(2n + 2)
n=1
n=1

What is this ommon value ?


Solution by Manuel Benito, Os ar
Ciaurri, and Emilio Fernandez, Logrono,
~
Spain, expanded by the editor.
Let A and B denote the summations on the left side and the right side
of the proposed equality, respe tively. Also, let Hn = 1 + 21 + + n1 . Then
1
n+1
=

1
1
+ +
= H2n Hn
n+2
2n


2n
X
1
1
1
(1)j1
H2n 2
+ + +
=
2
4
2n
j
j=1
+

Sin e it is well known that

P
(1)j 1

of the double summation we have

A

X
1

n=1

k=1

j1

(1)

H j1 =
2

Hk
(2k + 1)(2k + 2)

j=2n+1

X
(1)j1

j=3

= ln 2,

j=1

Hk

k=1

1
2k + 1

X
1+

n=1

by hanging the order

X
(1)j1

j=3

1
2

j1
2

X 1

n=1

2k + 2

+ +

1
n

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

= B.

48
To nd the ommon value of the two absolutely onvergent series, let

X
(1)j1

f (x) =

j=3

H j1 xj ,
2

where the power series for f onverges for all x (1, 1). Then
f (x)

=
=

(1)j1 H j1 xj1 =

j=3

(1)j H j xj
2

j=2

(Hn x2n Hn x2n+1 ) = (1 x)

n=1

Hn x2n .

(1)

n=1

Now, it is well known that

1
1x

[Ed : Multiply 1 1 x

ln

1
1x

= 1+x+x2 +

Hn xn .

(2)

n=1

with ln(1x) = x+ 12 x2 + 13 x3 +

and observe that the oe ient of xn is 1 + 12 + + n1 .

From (2) we have (1 x)
Thus, (1 x2 )

n=1

(1 x)

n=1

Hn xn = ln(1 x).

Hn x2n = ln(1 x2 ),

n=1

Hn x2n =

1
1+x

and it follows that

ln(1 x2 ) .

(3)

From (1) and (3) we obtain

f (x) =

1
ln(1 x2 ) .
1+x

Sin e f (0) = 0 and the last improper integral below is onvergent, by

applying Abel's Continuity Theorem for power series we have
A =

lim f (x) =

x1

f (x) dx =

1
0

ln 1 x2
dx .
1+x

(4)

It remains to evaluate the last integral in (4).

With the hange of variable x = 2u 1, we have
Z

1
0

ln 1 x2
1+x

dx =

1/2

ln 4u(1 u)
u

du = I1 + I2 ,

(5)

49
where
I1

1



ln(4u)
1 2
1
du =
ln (4u)
=
(ln 4)2 (ln 2)2
u
2
2
1/2
1/2
1
3
1
(ln 4 + ln 2)(ln 4 ln 2) = (ln 8)(ln 2) = (ln 2)2 .
2
2
2

=
=

(6)

On the other hand, using integration by parts and then making the
hange of variable u = 1 t, we have
I2

1
Z 1


ln(1 u)
ln u
=
du = (ln u) ln(1 u) 
+
du

u
1/2
1/2 1 u
1/2
Z 1/2
Z 1
ln(1 t)
ln(1 t)
= (ln 2)2 +
dt = (ln 2)2 +
dt I2 ,
t
t
0
0
Z

from whi h we obtain

2I2 = (ln 2)2 +

1
0

ln(1 t)
dt .
t

(7)

[Ed : Sin e the integrals in the omputations above

are improper, are must
be taken ; e.g., the evaluation of (ln
u)
ln(1

u)
at u = 1 must be done by


omputing lim (ln u) ln(1 u) using L'Hospital's
^
Rule.
u1
Finally,



Z 1
1
1
ln(1 t)
dt =
ln
dt
t
1t
0 t
0
!
Z 1

Z 1 n1
X
1 X tn
t
=
dt =
dt
t
n
n
0
n=1
n=1 0
Z

n=1

X
tn 1
1
2
=


n2 0
n2
6
n=1

From (4) { (8), we on lude that A = B =

(8)

2
(ln 2)2 .
12

Also solved by PAOLO PERFETTI, Dipartimento di Matemati a, Universita degli studi di

Tor Vergata Roma, Rome, Italy ; and the proposer. There were also three in omplete solutions,
all of whi h only demonstrated that the two given summations are equal.

3302. [2008 : 44, 47 Proposed by Mihaly Ben ze, Brasov, Romania.

Let s, r, and R denote the semiperimeter, the inradius, and the
ir umradius of a triangle ABC , respe tively. Show that
(s2 + r 2 + 4Rr)(s2 + r 2 + 2Rr) 4Rr(5s2 + r 2 + 4Rr) ,

and determine when equality holds.

50
Solution by Mi hel Bataille, Rouen, Fran e.
First, using the two known formulas s2 + r2 + 4Rr = ab + bc + ca and
abc = 4Rrs, where a, b, and c are the sides of the triangle, we dedu e that
(a + b)(b + c)(c + a) =
=
=

(a + b + c)(ab + bc + ca) abc

2s s2 + r 2 + 4Rr 4Rrs


2s s2 + r 2 + 2Rr .

For onvenien e, let e1 = a + b + c, e2 = ab + bc + ca, and e3

It follows that the required inequality is su essively equivalent to
(a + b)(b + c)(c + a)e2
(a + b)(b + c)(c + a)e2


ab + a b + bc2 + b2 c + ca2 + c2 a e2
2

a2 b3 + a3 b2 + b2 c3 + b3 c2 + c2 a3 + c3 a2

= abc.

8Rrs(5s2 + r 2 + 4Rr) ,


2e3 e21 + e2 ,
2e3 e21 ,

2a2 b2 c + 2a2 bc2 + 2ab2 c2 ,

and nally to
a2 (b c)2 (b + c) + b2 (c a)2 (c + a) + c2 (a b)2 (a + b) 0 .

The last inequality is obviously true, whi h ompletes the proof. Equality
holds if and only if a = b = c, that is, if and only if the triangle ABC is
equilateral.
Also solved by ARKADY ALT, San Jose, CA, USA ; GEORGE APOSTOLOPOULOS,

 University of Sarajevo, Sarajevo, Bosnia and
Messolonghi, Gree e ; SEFKET
ARSLANAGIC,
Herzegovina ; ROY BARBARA, Lebanese University, Fanar, Lebanon ; CHIP CURTIS, Missouri
Southern State University, Joplin, MO, USA ; CHARLES R. DIMINNIE, Angelo State
University, San Angelo, TX, USA ; OLIVER GEUPEL, Bruhl,
 NRW, Germany ; JOE HOWARD,
Portales, NM, USA ; WALTHER JANOUS, Ursulinengymnasium, Innsbru k, Austria ; KEE-WAI
LAU, Hong Kong, China ; THANOS MAGKOS, 3rd High S hool of Kozani, Kozani, Gree e ; SALEM
 student, Sarajevo College, Sarajevo, Bosnia and Herzegovina ; ANDREA MUNARO,
MALIKIC,
student, University of Trento, Trento, Italy ; PANOS E. TSAOUSSOGLOU, Athens, Gree e ;
GEORGE TSAPAKIDIS, Agrinio, Gree e ; PETER Y. WOO, Biola University, La Mirada, CA, USA ;
TITU ZVONARU, Comane
 sti, Romania ; and the proposer.

3303. [2008 : 44, 47 Proposed by Mihaly Ben ze, Brasov, Romania.

Let a, b, and c be positive real numbers. Show that

Y

y li

2(a + b)3

y li

(a + s1 )(bc + s2 ) ,

where s1 = a + b + c and s2 = ab + bc + ca.

51

Composite of similar solutions by Manuel Benito, Os ar
Ciaurri, Emilio
Fernandez, and Luz Ron al, Logrono,
~ Spain ; and Chip Curtis, Missouri
Southern State University, Joplin, MO, USA.
By straightforward omputations, we have
2(a + b)(b + c)(c + a) (a + s1 )(bc + s2 )
= 2(a + b)(b + c)(c + a) (2a + b + c)(ab + 2bc + ca)
= 2(a2 b + ab2 + b2 c + bc2 + c2 a + ca2 + 2abc)
(2a2 b + ab2 + 2b2 c + 2bc2 + c2 a + 2ca2 + 6abc)
= ab2 + ac2 2abc = a(b c)2 0 .

Hen e,

2(a + b)(b + c)(c + a) (a + s1 )(bc + s2 ) .

Similarly, we have

2(a + b)(b + c)(c + a) (b + s1 )(ca + s2 ) ;

2(a + b)(b + c)(c + a) (c + s1 )(ab + s2 ) .

The result now follows by multiplying a ross the last three inequalities.
Also solved by ARKADY ALT, San Jose, CA, USA ; GEORGE APOSTOLOPOULOS,

 University of Sarajevo, Sarajevo, Bosnia and HerMessolonghi, Gree e ; SEFKET
ARSLANAGIC,
zegovina ; DIONNE BAILEY, ELSIE CAMPBELL, and CHARLES R. DIMINNIE, Angelo State University, San Angelo, TX, USA ; MICHEL BATAILLE, Rouen, Fran e ; CAO MINH QUANG, Nguyen
Binh Khiem High S hool, Vinh Long, Vietnam ; OLIVER GEUPEL, Bruhl,
 NRW, Germany ;
 student, Sarajevo College, Sarajevo, Bosnia and Herzegovina ; ANDREA
SALEM MALIKIC,
MUNARO, student, University of Trento, Trento, Italy ; NGUYEN MANH DUNG, High s hool
student, HUS, Hanoi, Vietnam ; PAOLO PERFETTI, Dipartimento di Matemati a, Universita
degli studi di Tor Vergata Roma, Rome, Italy ; DANIEL TSAI, student, Taipei Ameri an S hool,
Taipei, Taiwan ; PANOS E. TSAOUSSOGLOU, Athens, Gree e ; GEORGE TSAPAKIDIS, Agrinio,
Gree e ; PETER Y. WOO, Biola University, La Mirada, CA, USA ; TITU ZVONARU, Comane
 sti,
Romania ; and the proposer.

3304. [2008 : 45, 47 Proposed by Mihaly Ben ze, Brasov, Romania.

Let a1 , a2 , . . . , an be positive real numbers and identify an+1 with a1 .
Prove that
n
n
X
X
a3k
ak a2k+1 .
k=1

k=1

Similar solutions by Mi hel Bataille, Rouen, Fran e ; Walther Janous, Ursulinengymnasium, Innsbru k, Austria ; Steven Karp, student, University of
Waterloo, Waterloo, ON ; Xavier Ros, student, Universitat Polite ni a

de Catalunya, Bar elona, Spain ; and Daniel Tsai, student, Taipei Ameri an S hool,
Taipei, Taiwan.
Reorder the numbers a1 , a2 , . . . , an from smallest to largest and rename them x1 , x2 , . . . , xn . Then 0 x1 x2 xn and if yi = x2i

52
for ea h i then we also have 0 y1 y2 yn . The Rearrangen
n
P
P
ment Inequality states that
xi yi
xi y(i) for any permutation of
{1, 2, . . . , n}.

n
P

i=1

i=1

n
P

n
P

Sin e
=
xi yi and
ak a2k+1
k=1
i=1
k=1
appropriate permutation , the result follows.
a3k

n
P

xi y(i)

i=1

for an

Also solved by MOHAMMED AASSILA, Strasbourg, Fran e ; ARKADY ALT, San Jose,

 UniCA, USA ; GEORGE APOSTOLOPOULOS, Messolonghi, Gree e ; SEFKET
ARSLANAGIC,
versity of Sarajevo, Sarajevo, Bosnia and Herzegovina ; ROY BARBARA, Lebanese University,

Fanar, Lebanon ; MANUEL BENITO, OSCAR
CIAURRI, EMILIO FERNANDEZ, and LUZ
RONCAL, Logrono,
~ Spain ; CAO MINH QUANG, Nguyen Binh Khiem High S hool, Vinh Long,
Vietnam ; CHARLES R. DIMINNIE, Angelo State University, San Angelo, TX, USA ; OVIDIU
FURDUI, University of Toledo, Toledo, OH, USA ; OLIVER GEUPEL, Bruhl,
 NRW, Germany ;
 student, Sarajevo College, Sarajevo, BosJOE HOWARD, Portales, NM, USA ; SALEM MALIKIC,
nia and Herzegovina ; DUNG NGUYEN MANH, High S hool of HUS, Hanoi, Vietnam ; PAOLO
PERFETTI, Dipartimento di Matemati a, Universita degli studi di Tor Vergata Roma, Rome,
Italy ; HENRY RICARDO, Medgar Evers College (CUNY ), Brooklyn, NY, USA ; PANOS
E. TSAOUSSOGLOU, Athens, Gree e ; GEORGE TSAPAKIDIS, Agrinio, Gree e ; PETER Y. WOO,
Biola University, La Mirada, CA, USA ; TITU ZVONARU, Comane
 sti, Romania ; and the proposer.
Ri ardo omments that this problem appears as Problem 11.7 on p. 148 of Elementary
Inequalities by D.S. Mitrinovi (P. Nordho, 1964), but that no solution is provided there.

3305. [2008 : 45, 47 Proposed by Stanley Rabinowitz, MathPro Press,

Chelmsford, MA, USA.
Prove that
tan

6
2
+ 4 sin
13
13

4
+ 4 sin
13
5
tan
+ 4 sin
13

= tan
=

13
q

2
=
13 + 2 13 .
13


Solution by Manuel Benito, Os ar
Ciaurri, Emilio Fernandez, and Luz Ron al,
Logrono,
~ Spain, in memory of Jim Totten.
We will prove that
tan

2
6
+ 4 sin
13
13

5
+ 4 sin
13
6
4 sin
= tan
13
= tan

and
tan

4
13

+ 4 sin

13

= tan
= tan

13
3
13

+ 4 sin
+ 4 sin

2
13
q

5
=
13 + 2 13
13

(1)

3
13
4
13

13 2 13 .

(2)

We will make use of two elegant results due to K.F. Gauss and in luded in
the Se tio VII of the Disquisitiones Arithmeti  (DA).

53
Lemma (DA, art. 362, II). Let n > 1 be an odd number and =
k

is any of the numbers 1, 2, . . . , n 1. Then,

2k
,
n

where




tan = 2 sin(2) sin(4) + sin(6) + sin (n 1) .

Theorem (DA, art. 356). Let n > 1 be an odd prime number, R be the set of

the (positive and less than n) quadrati residues modulo n, and N be the set
of the (positive and less than n) quadrati non-residues modulo n. Then,

and


X
2m
2r
n if

cos
=
cos
0 if
n
n
rR
mN
X

rR

sin


X
2r
2m
0

sin
=
n
n
n
mN

For n = 13 with =

2k
n

if
if

n 1 (mod 4)
n 3 (mod 4)

,
,

n 1 (mod 4)
n 3 (mod 4)

,
.

and 1 k 12 the Lemma yields



tan = 2 sin(2) sin(4) + sin(6) sin(8) + sin(10) sin(12) .

We ompute with dierent values of k in this identity as follows.

If k = 1 and = 2
, then the Lemma yields
13

. (3)



6

2
3
4
5
6
tan
= 2 sin
+ sin
+ sin
+ sin
+ sin
+ sin

. (4)

tan

2
4
5

3
6
2
= 2 sin
sin
+ sin
+ sin
sin
+ sin
13
13
13
13
13
13
13

If k = 3 and =
13

then the Lemma yields

13

If k = 4 and =
tan

6
,
13

8
,
13

13

then tan 5
13

13

= tan

13

8
13

13

13

and the Lemma yields

. (5)

. (6)

. (7)

5
3
6
4

2
5
= 2 sin
+ sin
+ sin
+ sin
sin
sin
13
13
13
13
13
13
13

By omparing the equations (3), (4), and (5) we see that the rst three
expressions in equation (1) are equal.
If k = 2 and = 4
, then the Lemma yields
13
tan

4
5
3
2
6

4
= 2 sin
+ sin
sin
sin
sin
+ sin
13
13
13
13
13
13
13

If k = 5 and =
tan

10
,
13

then tan 3
13

= tan

10
13

and the Lemma yields

5
2
4
3
3
= 2 sin
sin
sin
+ sin
+ sin
sin
13
13
13
13
13
13
13

54
If k = 6 and =
tan

12
,
13

then tan 13

= tan

12
13

and the Lemma yields

2
4
6
5
3

= 2 sin
sin
+ sin
sin
+ sin
sin
13
13
13
13
13
13
13

. (8)

By omparing the equations (6), (7), and (8) we see that the rst three
expressions in equation (2) are equal.
6

and B = tan 4
Now we take A = tan 2
+ 4 sin
+ 4 sin .
13
13
13
13
Clearly A and B are positive numbers. From (3) and (6) it follows that



3
4
A + B = 4 sin
+ sin
+ sin
13

and

13

13



5
6
2
sin
+ sin
A B = 4 sin
13

Then,

13

13





3
4
2
5
6
A2 B 2 = 16 sin
+ sin
+ sin
sin
sin
+ sin




2
3
4
5
6
A2 B 2 = 8 cos
+ cos
+ cos
cos
cos
+ cos

13

13

13

13

13

13

Applying the identity 2 sin a sin b = cos(a b) cos(a + b), we have

13

13

13

13

13

13

However, for n = 13 1 (mod 4), the sets R and N in the Theorem are
R = {1,4,9,3,12,10} and N = {2,8,6,11,5,7} ; thus, by the Theorem



2
6
5

3
4
2 cos
+ cos
cos
+ cos
+ cos
cos
=
13 ,
13

13

and therefore

13

13

13

13

A2 B 2 = 4 13 .

Similarly, using the identity 2 sin2 a = 1 cos(2a) we dedu e that

AB

=
=




2
3
4
5
6
4 sin
+ sin
+ sin
+ sin
sin
+ sin
13
13
13
13
13
 13


2
3
4
5
6
sin
sin
+ sin
+ sin
+ sin
sin
13
13
13
13
13
13


2
3
4
5
6

+ cos
+ cos
cos
cos
+ cos
6 cos
13
13

6 13
= 3 13 .
2

13

13

13

13

For positive real numbers

A and B with

A > B , the solutions of the equations A2 B 2 = 4 13 and AB = 3 13 are

A =

13 + 2 13

and

B =

13 2 13 .

55
This ompletes the proof of the identities (1) and (2). The following similar
identities an be dedu ed when n = 11 :
tan

+ 4 sin
11
11

2
5
+ 4 sin
11
11
3
2
4

tan
+ 4 sin
= tan
+ 4 sin
11
11
11
11
5
4
tan
4 sin
11
11

= tan
=
=
=

11 .


 University of Sarajevo, Sarajevo, Bosnia and HerAlso solved by SEFKET
ARSLANAGIC,
zegovina ; MICHEL BATAILLE, Rouen, Fran e ; APOSTOLIS K. DEMIS, Varvakeio High S hool,
Athens, Gree e (2 solutions) ; JOHN HAWKINS and DAVID R. STONE, Georgia Southern University, Statesboro, GA, USA ; and PETER Y. WOO,
Biola University, La Mirada, CA, USA.
q

All solvers noted that tan 13 +4 sin 13 6= 13 + 2 13, as did George Apostolopoulos,
Messolonghi, Gree e ; and Luyan Zhong-Qiao, Columbia International College, Hamilton, ON.
Wagon used Mathemati a to he k that the rst and third identities are orre t and the se ond
is in orre t. The proposer oered a partially orre t solution.

Woo wondered if similar results hold for 5 , 7 , 11

or 17
. For the ase of 11
Benito et al.
answered (above) in the armative. The interested reader may want to investigate the other
ases. Woo also hallenges the readers to nd geometri proofs for the equalities in (1) and (2).

3306. [2008 : 45, 47 Proposed by Stanley Rabinowitz, MathPro Press,

Chelmsford, MA, USA.
Find a real number t and polynomials f (x), g(x), and h(x) with integer
oe ients, su h that
f (t) =

2,

g(t) =

3,

and

h(t) =

7.

Solution by Roy Barbara, Lebanese University, Fanar, Lebanon.

Set = 2 + 3 + 7 and = 2 3 7. Computing 3 , 5 , and 7 ,
we obtain

2 +
3 +
7

16 2 + 15 3 + 11 7 + 3

This is a

1 3
,
2

1 5
281 2 + 241 3 + 161 7 + 60 =
,
4

1 7
4796 2 + 3975 3 + 2611 7 + 1043 =
.
8


linear system for 2, 3, 7, and . Solving for 2, 3,
=

and

56
yields

59
1
1 5
3 +
,
20
2
80
313
297 3
67 5
3 7
=

+

,
80
160
320
640
469
377 3
71 5
3 7
=
+

+
.
80
160
320
640
=

2+ 3+ 7

Finally, setting t = 80
=
we obtain 2 = f (t), 3 = g(t),
80

and 7 = h(t), where the polynomials f (x), g(x), and h(x) have integer
oe ients :

f (x)

= 236x

g(x)

= 313x

h(x)

= 469x +

2
1
2

(80)3 x3 + (80)4 x5 ,
(80)2 297x3 +
1
2

1
4

(80)2 377x3

(80)4 67x5
1
4

3
8

(80)6 x7 ,

(80)4 71x5 +

3
8

(80)6 x7 .

Also solved by MOHAMMED AASSILA, Strasbourg, Fran e ; BRIAN D. BEASLEY,


Presbyterian College, Clinton, SC, USA ; MANUEL BENITO, OSCAR
CIAURRI, EMILIO
FERNANDEZ, and LUZ RONCAL, Logrono,
~ Spain ; CHIP CURTIS, Missouri Southern State
University, Joplin, MO, USA ; WALTHER JANOUS, Ursulinengymnasium, Innsbru k, Austria ;
and the proposer.

3307. [2008 : 45, 47 Proposed by D.E. Prithwijit, University College Cork,

Republi of Ireland.
Eliminate from the system
cos(2)
sin(2)

= cos( + ) ,

= 2 sin( + ) .


Similar solutions by Manuel Benito, Os ar
Ciaurri, Emilio Fernandez, and
Luz Ron al, Logrono,
~ Spain ; Joe Howard, Portales, NM, USA ; and George
Tsapakidis, Agrinio, Gree e.
The given system an be rewritten as a linear system in cos and sin :
cos cos sin sin
sin cos + cos sin

= (cos2 sin2 ) ,
= sin cos .

Its solution is then

cos = cos3 ,

and

sin = sin3 ;

57
when e,

(cos )2/3 + (sin )2/3 = 2/3 .

(1)

Comments from the Spanish team. Note that the original system has a solution if and only if and satisfy (1). In parti ular, letting x = cos and
y = sin , we see that for 1 || 2 the solutions an be represented
by the interse tion points of the unit ir le x2 + y2 = 1 with the astroid
x2/3 + y 2/3 = 2/3 . Thus for ea h with absolute value between 1 and 2,
(1) will be satised for eight values of ; for {1, 2}, it will be satised by four values of . There an be no real solutions for other values of
.
Also solved by ARKADY ALT, San Jose, CA, USA ; GEORGE APOSTOLOPOULOS,

 University of Sarajevo, Sarajevo, Bosnia and
Messolonghi, Gree e ; SEFKET
ARSLANAGIC,
Herzegovina ; MICHEL BATAILLE, Rouen, Fran e ; CHIP CURTIS, Missouri Southern State
University, Joplin, MO, USA ; APOSTOLIS K. DEMIS, Varvakeio High S hool, Athens, Gree e ;
JOSE LUIS D I AZ-BARRERO, Universitat Polite ni a

de Catalunya, Bar elona, Spain ; OLIVER
GEUPEL, Bruhl,
 NRW, Germany ; JOHN HAWKINS and DAVID R. STONE, Georgia Southern
University, Statesboro, GA, USA ; WALTHER JANOUS, Ursulinengymnasium, Innsbru k,
Austria ; DAVID E. MANES, SUNY at Oneonta, Oneonta, NY, USA ; ANDREA MUNARO,
student, University of Trento, Trento, Italy ; PAOLO PERFETTI, Dipartimento di Matemati a,
Universita degli studi di Tor Vergata Roma, Rome, Italy ; XAVIER ROS, student, Universitat Polite ni a

de Catalunya, Bar elona, Spain ; BOB SERKEY, Leonia, NJ, USA ; PANOS E. TSAOUSSOGLOU, Athens, Gree e ; and the proposer. There was one in orre t submission.
Our readers produ ed solutions in a variety of formats. Here are a few of the ni est.
Instead of (1), Alt, Bataille, and the proposer independently obtained the equivalent equation
sin2 (2) =

3
4 2 1
272

Geupel found that in terms of a real parameter t, the solutions of the given system satisfy
= arctan t + m ,

= arctan(t ) + n ,
3

and

m+n

= (1)

(1 + t2 )3
1 + t6

for integers m and n. In addition, there were several impli it solutions where
the solver simply
presented an equation for in terms of or ; for example, = arctan 3 tan + k ame
from Arslanagi and from Ros.

3308. [2008 : 45, 48 Proposed by D.E. Prithwijit, University College Cork,

Republi of Ireland.

Given ABC, let AD be the altitude to BC . If AB : AC = 1 : 3,

prove that AD 23 BC . When does equality hold ?

I. Solution by Joe Howard, Portales, NM, USA.

It su es to take AB = 1 and AC = 3 ; onsequently, AD = sin B .

Writing a = BC , we therefore must show that

3
sin B
a.
2

58
We start with the inequality
1
=

sin(60 + A) = sin 60 cos A + sin A cos 60

3
1
cos A +
sin A ,
2
2

whi h is equivalent to

3 cos A sin A .
(1)

By the osine law, a2 = 4 2 3 cos A, so the inequality (1) is equivalent to

2

a2 2 sin A .

By the sine law,

sin A
sin B
= ,
a
3

(2)

or

a =

3 sin A
sin B

when e (2) is equivalent to

a

3 sin A
sin B

2 sin A ,

whi h redu es immediately to what we were to show.Equality o urs when

60 + A = 90 , in whi h ase A = 30 and a2 = 4 2 3 cos 30 = 1. Thus,
equality holds if and only if ABC is isos eles with A = C = 30 .
II. Solution by Mi hel Bataille, Rouen, Fran e.
We are given that vertex A belongs to the lo us of those points P for
B
1
whi h P
= . We re ognize this lo us to be a ir le alled the ir le of
PC
3
Apollonius ; it interse ts symmetri ally the line joining B to C in points
that

divide the segment BC internally and externally in the ratio 1 : 3. As an

easy onsequen e, the entre K of satises BK = 21 BC , and its radius

is

3BC
.
2

Let N N be the diameter of that is perpendi ular to BC . Then

and N are the points of that are farthest from the line BC , hen e

3BC
AD = d(A, BC) d(N, BC) =
2

Equality holds if and only if A is situated at N or N ', in whi h ase ABC

is isos eles with BA = BC and ABC = 120 .
Also solved by MOHAMMED AASSILA, Strasbourg, Fran e ; ARKADY ALT, San


Jose, CA, USA ; GEORGE APOSTOLOPOULOS, Messolonghi, Gree e ; SEFKET
ARSLANAGIC,
University of Sarajevo, Sarajevo, Bosnia and Herzegovina ; ROY BARBARA, Lebanese

University, Fanar, Lebanon ; MANUEL BENITO, OSCAR
CIAURRI, EMILIO FERNANDEZ, and
LUZ RONCAL, Logrono,
~ Spain ; CHIP CURTIS, Missouri Southern State University, Joplin,

59
MO, USA ; CHARLES R. DIMINNIE, Angelo State University, San Angelo, TX, USA ; IAN
JUNE L. GARCES, Ateneo de Manila University, Quezon City, The Philippines ; FRANCISCO
 IES Alvarez

JAVIER GARC I A CAPITAN,
Cubero, Priego de Cordoba,

Spain ; OLIVER GEUPEL,
Bruhl,

NRW, Germany ; STEVEN KARP, student, University of Waterloo, Waterloo, ON ;

 Y,
 Big Rapids, MI, USA ; KEE-WAI LAU, Hong Kong, China ; THANOS
V ACLAV
KONECN
 student, Sarajevo
MAGKOS, 3rd High S hool of Kozani, Kozani, Gree e ; SALEM MALIKIC,
College, Sarajevo, Bosnia and Herzegovina ; ANDREA MUNARO, student, University of
Trento, Trento, Italy ; JOEL SCHLOSBERG, Bayside, NY, USA ; PANOS E. TSAOUSSOGLOU,
Athens, Gree e ; GEORGE TSAPAKIDIS, Agrinio, Gree e ; DANIEL TSAI, student, Taipei
Ameri an S hool, Taipei, Taiwan ; PETER Y. WOO, Biola University, La Mirada, CA, USA ; TITU
ZVONARU, Comane
 sti, Romania ; and the proposer.
q
More generally, Kone ny proved that AD 2
BC if AC : AB = q with q > 1.
q 1
His argument was mu h like that of solution II above.

3309. [2008 : 45, 48 Proposed by Virgil Ni ula, Bu harest, Romania.

Let , , and be xed non-zero real numbers. Show that the system
x + y + z
xy + yz + zx

=
=

1,
1,

has a unique solution for (x, y, z) if and only if

2 + 2 + 2 + 1 = 2( + + ) ,

and, in this ase nd that unique solution.

Solution by George Tsapakidis, Agrinio, Gree e, modied by the editor.
Substituting x = 1 y z into (xy + yz + zx) = we obtain
(y + z)(1 y z) + yz = , whi h upon simplifying yields the following
quadrati equation in y


y 2 1 + ( )z y + z 2 z + = 0 .

(1)

Equation (1) has a unique solution in y if and only if

[1 + ( )z] 4(z 2 z + ) = 0 ,
2

whi h an be written as the following quadrati equation in z



( )2 4 z 2 + 2( + )z + 1 4 = 0 .

(2)

Equation (2) has a unique solution in z if and only if



( + )2 ( )2 4 (1 4) = 0 ,

whi h, by straightforward omputations, redu es to

2 + 2 + 2 + 1 = 2( + + ) .

(3)

60
Now, suppose (3) holds. Then we have
( )2 4 = 2 + 2 + 2 2 2 2 = 1 .

Thus, (2) redu es to z 2 + 2( + )z + 1 4 = 0, the unique

solution of whi h is given by z = + .
Using this and (3), we nd that
y

=
=

1 + ( )( + )

1 + ( )2 2

2
2
2 2 + 2 + 2
1 + 2 2 + 2 2
=
= +.
2
2

Finally, using (3) we have

x =
=

1 y z = 1 ( + ) ( + )
2 + + ,

from whi h it follows that x = + .

Therefore, the unique solution to the given system is
(x, y, z) = ( + , + , + ) .

Also solved by ARKADY ALT, San Jose, CA, USA ; GEORGE APOSTOLOPOULOS, Messolonghi, Gree e ; ROY BARBARA, Lebanese University, Fanar, Lebanon ; MICHEL BATAILLE,

Rouen, Fran e ; MANUEL BENITO, OSCAR
CIAURRI, EMILIO FERNANDEZ, and LUZ RONCAL,
Logrono,
~ Spain ; CHIP CURTIS, Missouri Southern State University, Joplin, MO, USA ; OLIVER
GEUPEL, Bruhl,
 NRW, Germany ; STEVEN KARP, student, University of Waterloo, Waterloo,
ON ; PAOLO PERFETTI, Dipartimento di Matemati a, Universita degli studi di Tor Vergata
Roma, Rome, Italy ; DANIEL TSAI, student, Taipei Ameri an S hool, Taipei, Taiwan ; PETER
Y. WOO, Biola University, La Mirada, CA, USA ; TITU ZVONARU, Comane
 sti, Romania ; and
the proposer. There was also one partially in orre t solution submitted.

3310. [2008 : 46, 48 Proposed by Virgil Ni ula, Bu harest, Romania.

Let a, b, and c denote, as usual, the lengths of the sides BC , CA, and
AB , respe tively, in ABC . Let s be the semiperimeter of ABC , r the
inradius, ha the altitude to side BC , and ra , rb , and rc the exradii to A, B ,
and C , respe tively.

(a) Show that for x > 0, we have ha

or x = rc .

(b) Show that for x > 0, we have ha

x = r or x = ra .

2s(s a)x
x2 + s(s a)

if and only if x = rb

2(s b)(s c)x

|x2 (s b)(s c)|

if and only if

61
Similar solutions by Mi hel Bataille, Rouen, Fran e ; Oliver Geupel, Bruhl,

NRW, Germany ; Titu Zvonaru, Comane
 sti, Romania ; and the proposer.
It is well known that the area F of ABC
p an be variously expressed
as 12 aha , rs, ra (s a), rb (s b), rc (s c), or s(s a)(s b)(s c). We
have
rb + rc

sb

sc

F (2s b c)

(s b)(s c)

a F (s a)
as(s a)
=
s(s a)(s b)(s c)
F

2as(s a)
aha

2s(s a)

(1)

ha

and
rb rc

F2
(s b)(s c)

s(s a)(s b)(s c)

(s b)(s c)

= s(s a) .

(2)

It follows from (1) and (2) that x = rb and x = rc are the solutions of
ha x2 2s(s a)x + ha s(s a) = 0 ,

hen e these are the solutions of the equation in (a).

We also have
ra r

=
=

F
sa

F
s

aF
s(s a)

2(s b)(s c)
a F (s b)(s c)
=
2
F
ha

and
ra r =

F2
= (s b)(s c) .
s(s a)

(3)
(4)

By (3) and (4) it follows that the equation

ha x2 2(s b)(s c)x (s b)(s a)ha = 0

has the solutions x = ra and x = r and that the equation

ha x2 + 2(s b)(s c)x (s b)(s a)ha = 0

has the solutions x = ra and x = r. Thus, the only positive solutions of

the pre eding two equations are r and ra , and it follows that these are the
only positive solutions to the equation in (b).
Also solved by GEORGE APOSTOLOPOULOS, Messolonghi, Gree e ; MANUEL BENITO,

OSCAR
CIAURRI, EMILIO FERNANDEZ, and LUZ RONCAL, Logrono,
~ Spain ; CHIP CURTIS,
Missouri Southern State University, Joplin, MO, USA ; GEORGE TSAPAKIDIS, Agrinio, Gree e ;
and PETER Y. WOO, Biola University, La Mirada, CA, USA.

62
3311. [2008 : 46, 48 Proposed by Mi hel Bataille, Rouen, Fran e.
Let n be an integer with n 2. Suppose that for k = 0, 1, . . . , n 2
we have


n2
(1)k (k + 1) (mod n) .
k

Show that n is a prime.

Solution by Oliver Geupel, Bruhl,

 NRW, Germany.
It is su ient to prove that if n is a omposite integer with n 2, then
there exists an integer k with 0 k n 2 su h that


n2
k

6 (1)k (k + 1) (mod n)

(1)

Toward that end, let p be the least prime fa tor of n. If (1) holds for some
k < p, then we are done. Otherwise the given ongruen e holds in parti ular
for k = p 2, and we have




n2
n2
np
(p 1)
=

(n p 1)
p
p2
p


n
(1)p2 (p 1)
1 (p 1)
p


n
(p 1) (1)p p + 1
(mod n) .
p
Be ause p 1 is oprime to the modulus n, we an divide both sides by it
and on lude that (1) holds with k = p :


n2
p



n
(1)p p + 1
6 (1)p (p + 1) (mod n)
p


Also solved by MANUEL BENITO, OSCAR
CIAURRI, EMILIO FERNANDEZ, and
LUZ RONCAL, Logrono,
~ Spain ; JOHN HAWKINS and DAVID R. STONE, Georgia Southern
University, Statesboro, GA, USA ; STEVEN KARP, student, University of Waterloo, Waterloo,
 student, Sarajevo College, Sarajevo, Bosnia and Herzegovina ; DAVID
ON ; SALEM MALIKIC,
E. MANES, SUNY at Oneonta, Oneonta, NY, USA ; JOEL SCHLOSBERG, Bayside, NY, USA ; and
the proposer.


The onverse is a known result : If p is a prime, then p2
(1)k (k + 1) (mod p)
k
for k = 0, 1, . . . , p 2. Karp provided a simple proof (by indu tion on k) ; Bataille provided
a referen e : E. Lu as, Theorie

des nombres, A. Blan hard (1961), p. 420.

3312. [2008 : 46, 48 Proposed by Mi hel Bataille, Rouen, Fran e.

Let n be a positive integer ongruent to 1 modulo 6. Show that 3/n an
be expressed as
1

a1

a2

+ +

ak

for some distin t positive integers a1 , a2 , . . . , ak , and nd the minimal value

of k.

63
Solution by Steven Karp, student, University of Waterloo, Waterloo, ON.
We solve the problem for all positive integers n. If n 0 or 3 (mod 6),
1
then n3 = n/3
. If n 1 (mod 6), then
3
=
n

If n 2

1
n+1
3

n(n+1)
3

(mod 6), then

3
1
1
= n +
n
n
2

If n 1

(mod 6)

and n has a fa tor c su h that c 1

3
=
n

1
n+c
3

n(n+c)
3c

(mod 6),

then

We see that k is minimal in all of these ases, sin e n3 = a1 for some positive
1
integer a1 only if n 0 (mod 3). Now, if n > 1, n 1 (mod 6) and all
divisors of n are ongruent to 1 modulo 6, then
3
=
n

1
n+1
2

1
+
n

1
n(n+1)
2

and we laim that k = 3 is minimal in this ase. To prove this suppose for
the sake of ontradi tion that
3
n

1
a

1
b

for distin t positive integers a and b. Then

n =

b(3a n)
a

and we have a|b(3a n). Therefore, a = pq where p and q are positive


integers su h that p|b and q|(3a n). Then pb n, when e pb 1 (mod 6).
Sin e q|a and q|(3a n), we also have q|n, so q 1 (mod 6). We now have
1 qn =

 
b
p

(3a n) 3a n 1 or 2 (mod 6)

a ontradi tion.
Finally, for n = 1 we have
3 = 1+

1
1
1
1
1
1
1
1
1
1
1
1
+ + + + + + + +
+
+
+
2
3
4
5
6
7
8
9
10
15
230
57960

A omputer sear h shows that k = 13 is minimal in this ase.

64
Also solved by ROY BARBARA, Lebanese University, Fanar, Lebanon ; MANUEL

BENITO, OSCAR
CIAURRI, EMILIO FERNANDEZ, and LUZ RONCAL, Logrono,
~ Spain ; CHIP
CURTIS, Missouri Southern State University, Joplin, MO, USA ; JOHN HAWKINS and DAVID
R. STONE, Georgia Southern University, Statesboro, GA, USA ; WALTHER JANOUS, Ursulinengymnasium, Innsbru k, Austria ; DAVID E. MANES, SUNY at Oneonta, Oneonta, NY, USA ; JOEL
SCHLOSBERG, Bayside, NY, USA ; PETER Y. WOO, Biola University, La Mirada, CA, USA ; TITU
ZVONARU, Comane
 sti, Romania ; and the proposer.
Benito et al. refer to the note by Thomas R. Hagedorn, A proof of a onje ture on Egyptian
fra tions, Amer. Math. Monthly, 107 (2000) 62-63, where it is proved that for ea h odd integer
n 3 not divisible by 3 there exist distin t odd, positive integers a, b, and c su h that
3
1
1
1
=
+ +
n
a
b
c

The result had been onje tured by R. Hardin and N. Sloane. When n = 6p+ 1 Hagedorn gives
the de omposition
3
3
1
1
1
=
=
+
+
n
6p + 1
2p + 1
(2p + 1)(4p + 1)
(4p + 1)(6p + 1)

Janous refers to a paper by Andrzej S hinzel, Sur quelques propriet

 es
 des nombres 3/n
et 4/n, ou n est un nombre impair, Mathesis 65 (1956) 219-222, for a treatment of our problem.
Sin e the improper fra tion 3 = 3/1 arises, he asks if any CRUX readers know the minimum
number (n) of distin t Egyptian fra tions needed to represent the positive integer n.

A Happy New Year to all CRUX with MAYHEM readers. The Jim Totten
spe ial issue was slated to be ompleted in May of this year, but due to delays
we are now going to release the spe ial issue in September 2009 instead.
This year we plan on improving our database of names and aliations
of you, the readers. If your name does not look quite right, for example, if
the a ents are not quite right or your family name is in orre t, et ., then
please let us know and we will update our les. Many international readers
subs ribe to CRUX with MAYHEM and we want to get these (fas inating !)
details right.
Va lav

Linek.

Crux Mathemati orum

with Mathemati al Mayhem

Former Editors / An iens Reda teurs


: Bru e L.R. Shawyer, James E. Totten
Crux Mathemati orum
Founding Editors / Reda teurs-fondateurs

: Leopold

Sauve & Frederi k G.B. Maskell
Former Editors / An iens Reda teurs

: G.W. Sands, R.E. Woodrow, Bru e L.R. Shawyer
Mathemati al Mayhem
Founding Editors / Reda teurs-fondateurs

: Patri k Surry & Ravi Vakil
Former Editors / An iens Reda teurs

: Philip Jong, Je Higham, J.P. Grossman,
Andre Chang, Naoki Sato, Cyrus Hsia, Shawn Godin, Je Hooper