Crux

Mathematicorum
Published by the Canadian Mathematical Society.

http://crux.math.ca/

The Back Files

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

Journal title history:

➢ The first 32 issues, from Vol. 1, No. 1 (March 1975) to
Vol. 4, No.2 (February 1978) were published under the
name EUREKA.
➢ Issues from Vol. 4, No. 3 (March 1978) to Vol. 22, No.
8 (December 1996) were published under the name
Crux Mathematicorum.
➢ Issues from Vol 23., No. 1 (February 1997) to Vol. 37,
No. 8 (December 2011) were published under the
name Crux Mathematicorum with Mathematical
Mayhem.
➢ Issues since Vol. 38, No. 1 (January 2012) are
published under the name Crux Mathematicorum.
:': & -k i': tfc }': ;'c & ;V & j': & & ft s'« j; & & s'c J'J J'J & & }'; J'J s'j s'c s': * s'c s'c *'« A j'; ;'j
ft JV & :V s'« sV s"? & sV s'« sV & JV SV SV JV S V :V S V JV J* sV sV sV s'; JV :V SV S'C :V :?C SV S'C ft
ISSN 0705 - 0348

CRUX MATHFHATICORUH

Vol. 8 S No. 10
December 1982

Sponsored by
Carleton-Ottawa Mathematics Association Hathematique d'Ottawa-Carlecon
Publie par le College Alaonquin, Ottawa
Printed at Carieton University
The assistance of the publisher and the support of the Canadian Mathematical Olympiad
Committee,, the Carieton University Department of Mathematics p.nd Statistics, the University
of Ottawa Department of Mathematics, and the endorsement of the Ottawa Valley Education
Liaison Council are gratefully acknowledged.
ftftftftftftftftftft* ft ft ft ft *ftftftftftftftftftftftftftftftftftftftftftftft
ftftftftftftftftftft1t ft ft ft Mftftftftftftftftftftftftftftftftftftftftftftft

CRUX MATHEMATICORUM is a problem-solving journal at the senior secondary and university

undergraduate levels for those who practise or teach mathematics. Its purpose is primarily
educational, but it serves also those who read it for professional, cultural, or recreational
reasons.
It is published monthly (except July and August). The yearly subscription rate for ten
issues is $22 in Canada, US$20 elsewhere. Back issues: $2 each. 3ojnd volumes with index:
Vols* 1&2 (combined) and each of Vols. 3-7, $17 in Canada and US$15 elsewhere. Cheques and
money orders, payable to CRUX MATHEMATICORUM, should be sent to the managing editor.
All communications about the content (articles, problems, solutions, etc.) should be
sent to the editor. All changes of address and inquiries about subscriptions and back issues
should be sent to the managing editor.
Editor: Leo Sauv§, Algonquin College, 281 Echo Drive, Ottawa, Ontario, Canada K1S IN J.
Managing Editor: F.G.B. Maskell, Algonquin College, 200 Lees Ave., Ottawa, Ontario,
Canada K1S 0C5.
Typist-compositor: Nghi Chung.
Second Class Mail Registration No. 5432. Return Postage Guaranteed.
ft ft ft

CONTENTS
A New Inequality for the Anqles of a Triangle . . . . . . . . . . . 0. Bottema 296
Some Majorization Ineaualities M.S. Klamkin and George Tsintsifas 298
Mathematical Clerihews . . . . . . Alan Wayne 299
The Olympiad Corner: 40 . . . . • . M.S. Klamkin 300
Problems - Probl^mes . . . . . . . . . . . . . . . . . 302
The House Numbers of Europe . . . . . . . . . . Hayo Ah!burg 304
Solutions . . . . . . . . . . . 305
One Two,, Button Your Shoe 327
The Puzzle Corner . . . . . . . . . . . . . . . . . . . . . . . Hans Havermann 328
The Most Dedekindest Cut of All . . . . . . . . . . . . . . . . Leroy F. Meyers 328

- ?°5 -
 - 296 -
A NEW INEQUALITY FOR THE ANGLES OF A TRIANGLE
0. BOTTEMA

Let ABC be a trianqle with sides a9b,e and angles a,3,y in the usual order,
and circumradius R. We will assume that ABC is not right-angled, so that its
orthic trianqle AiBxCi is nondegenerate, and the orthocenters H and Hi of ABC
and A1B1C1 are well-defined points. We will first find an expression for d = HHi
in terms of elements of triangle ABC. We will then use this expression (which is
interesting in its own riqht) to deduce a new inequality for the angles of any
triangle.
We will assume that triangle ABC has been labeled so that y is the larqest
anqle. (The figure shows a triangle with
Y < ir/2, but it can be verified that every- C
thing that follows remains valid if y > TT/2.)
We introduce a Cartesian frame OXY with the
origin 0 at Cx and X-axis along AB. We easily
find that

H = (o, 2i?cos acos 3). (1)

Furthermore, A X C = b cos y and BA! = c c o s 3 ;

hence
Aj = {b cos 3 cosy, c cos 3 sin 3)

and similarly
$1 = (-aCOS a COS y, cCOSaSl'na).

Now BxCj is antiparallel to BC; hence /B1C1A = y and the equation of B ^ i is

a; Sin y + y COS y = 0.

From this it follows that the line through hi perpendicular to B ^ (an altitude
of A ^ j C x ) is
a: COSy - z / S i n y + / c 1 = 0 , (2)

where
ki = c cos 3 sin 3 sin y - b cos 3 cos2y
= -i?sin 23 cos 2y. (3)

In the same way, we find that the line through Bx perpendicular to AiCi has the
equation
 - 2P7 -
x COS y + z/Sl'ny + k2 = 0, (4)

where

fc2 = # S i n 2a COS 2 y . (5)

Solving (2) and (4) simultaneously, we find that the coordinates (a?i »2/i) of ti\
are given by

^ 2 cosy9 ^ 2sinY'
In view of (3) and (5) 9 we therefore have

W\ = (i?COS 2y S i n ( a - 8 ) , -i?COS 2y COS ( a - B ) ) . (6)

Now, from (1) and (6), the value of d = H ^ is given by

d2 = i? 2 COS 2 2y + 4/?2C0S a COS 8 (COS 2y COS ( a - 3 ) + COS a COS 8> . (1)

2
Obviously, it must be possible to write d as a symmetrical function of a,8,y.
We reduce (7) as follows:

d2 = i ? 2 c o s 2 2 y + 4i? 2 cos a cos g { c o s a cos 8 ( l + c o s 2 y ) + s i n a s i n 8 cos 2 y }

= i ? 2 c o s 2 2 y + 4i? 2 cos a cos 8 ( 2 cos a cos 8 c o s 2 y + s i n a s i n 8 cos 2y}
= 8i? 2 cos 2 a c o s 2 8 c o s 2 y + # 2 c o s 2 y ( s i n 2a s i n 28 + cos 2y)
= 8i? 2 C0S 2 a C0S 2 8 C0S 2 y + i?2COS 2 y { s i n 2a Sl'n 28 + COS ( 2 a + 2 8 ) }
= i? 2 (8COS 2 aCOS 2 8 COS2y + COS 2a COS 28 COS 2 y ) , (8)

and this yields the desired expression for d = HHi.

Since d2 > o, we conclude from (8) that
8 COS 2 a COS 2 8 COS2y + COS 2a COS 28 COS 2y > 0 (9)
holds for every triangle in which y * TT/2. But it holds also if y = TT/2 9 for
then (9) reduces to
COS22a (= COS 2 28) > 0 . (10)
So (9) holds for every triangle.
If triangle ABC is equilateral, it is easy to see that H coincides with Hj,
so d = o and equality holds in (9); and it follows from (10) that equality also
holds in (9) when a = 8 = TT/M^ and y = TT/2.
Finally, we note that (9) can be expressed more symmetrically in the form
(l + COS 2a)(l + COS 28)(l + COS 2y) + COS 2a COS 28 COS 2y > 0.

Charlotte de Bourbonstraat 2, 2628 BN Delft, The Netherlands.

 - 298 -

SOME flAJORIZATION INEQUALITIES

M.S. KLAMKIN and GEORGE TSINTSIFAS

In a recent note in this journal [1982: 162], the second author came up with
still another proof of the A.M.-G.M. inequality. He shov/ed that, if x. > o for
i = 1,2,...,n, then (sums and products throughout are for i - l,2,...,n)
n(a? + x.)
{l(x + x.)}

increases monotonically for x > o, and thus

n(ar + x .) Jix.
(1)
{l(x + x.)} {ixA

with equality if and only if x\ = x2 = ... - x . The A.M.-G.M. inequality is then

obtained by letting x -y » in (l). We will relate (l) to other inequalities.
If we set x/x. = a. for x > o, then (1) is equivalent to
t i
11(1 + a^) > (1 + H)n, (2)
where H is the harmonic mean n/z(l/a.). If we now let

A = -~ and G = (na^) 1 ^,

then (2) can be extended to

(1 + A)n > 11(1 + a.) > (1 + G)n > (1 + Hf. (3)
t
The first inequality in (3) follows from the A.M.-G.M. inequality applied to the
n numbers l+o^; for a proof of the second, see Til; and the third follows from
G > H.
Now we prove a stronger result in which the > signs in (3) are replaced by
majorization signs^-.
First, let T , r - o 9 l s ... 5 n, denote the elementary symmetric functions of
the a., which are defined by
11(2 + a.)
^
E TuQzn + Tiz 11
l ' + ... + Tn-1 Az + T n .

We say that (l+A)n majorizes nd+a^) and write (1+A)n > n(l+a.) if, termwise,

(n)Ar >T , v - 0,1,...,n.

p v
To show that
 - 29H -
(i + A)n y n(i + a^ > (i + of >- (i + nf9
al' wri need to prove is that, for v = o,i9.„.9n9

For v = 0, equality clearly holds throughout in (4). For r = 1,2,...,??, we must

show equivalently that
4 > p 1 / ^ > G > H, (5)

where p = T / ( n ) . The last inequality in (5) goes without sayinq, and the first
two follow from the Maclaurin inequalities T2], according to which we have

A =Pl > p 2 1 / 2 > p 3 1 / 3 ^ ... ^ pl/» = G.

REFERENCES

1. D.O. Shklarsky, N.N. Chentzov, and I.M. Yaqlom, The USSR Olympiad Problem
Book, Freeman, San Francisco, 1962, Problem 305, solution on pp. 410-411.
2. G.H. Hardy, J.E. Littlewood, and G. Polya, Inequalities, Cambridge Univer-
sity Press, 1952, p. 52.
Department of Mathematics, University of Alberta, Edmonton, Alberta, Canada
T6G 2G1.
Platonos 23, Thessaloniki, Greece.
* ft ft

MATHEMATICAL CLERIHEWS
David Hilbert, Karl Weierstrass2
W.S. Gilbert 1 — Enjoyed a qlass
Each "problemist" Of Rhine wine for drinkinq.
Compiled "a list". He hated fuzzy thinking.
ALAN WAYNE, Holiday, Florida
1
W.S. Gilbert, in The Mikado, Act I:
As some day it may happen that a victim must be found,
I've qot a little list—I've got a little list
Of society offenders who might well be underground,
And who never would be missed — who never would be missed!
Solomon Bochner in The Role of Mathematics in the Rise of Science, Princeton
University Press, 1966, p. 369: "He [Weierstrass1 liked his Rhine wine rich? but
this somehow did not prevent him from becoming a leading enemy of fuzziness in
mathematical thinking in the 19th century."
 - 300 -
TIT OLYMPIAD CORNER: 40
M.S. KLAMKIN

A few months aqo in this column [1982: 70], I qave the problems set in the
first round of the 1982 West German Olympiad. Now, through the courtesy of Bernhard
Leeb, I present the problems set in the second round, for which I solicit elegant
solutions.

WEST GERMAN MATHEMATICAL OLYMPIAD 1982

Second Round

1, Max divides the positive integer p by the positive integer q, where q < loo.
In the decimal expansion of the quotient p/q9 Max finds somewhere after
the decimal point the digit-block 1982. Prove that Max's division is wrong.
2, Decide whether e\/ery triangle ABC can be transformed by orthogonal projec-
tion on a certain plane into an equilateral triangle.

3, The nonneqative real numbers a\, a2 a satisfy ai+a2+...+a = 1.

Show that the sum

SI + £2 + ... + ^
l+a2+a^+..,+a l+ai+a 3 +...+a l+ai+a 2 +...+a
has a minimum and compute it.
£|, If 4 n + 2n + l is a prime number for the positive integer n, then n is a
power of 3.

I now present a set of problems proposed by students attending the 1982 U.S.A.
Mathematical Olympiad Practice Session held last summer at the U.S. Naval Academy
in Annapolis, Maryland. I may later publish the best solutions received from readers.
(Starred problems were submitted without a solution.) For earlier sets of such
problems, see [1980: 210:1982: 991.

1, Proposed by Jeremy Kahn and Steven Newman,

For a positive inteqer n > l, let s(n) denote the sum of all the divisors
of rc, excluding n itself. The integernis said to be deficient, perfect* or abun-
dant accordinq as sin) < n, s(n) = n, or sin) > n.
(a) Determine the maximal length of a block of consecutive integers each of
which is deficient.
(b) Determine the maximal length of a block of consecutive integers each of
which is abundant.
 - 301 -
2 V Proposed by Steven des Javdins.
Given is a sphere of integer radius r. Find the maximum number of points
whid' can be placed on the surface of the sphere in such a way that all pairs of
points are separated by an integer distance
(a) if chorda! distances are used;
(b) if distances are measured along the minor arcs of great circles of the
sphere.

3', Proposed by Andy Tikofsky.

For a positive integer n > 3, does there exist an integer m(n) such that
x + m(n) has 2(n-l) nonconstant polynomial factors with integral coefficients?
L\t Proposed by Noam Elkies.
Find all solutions (x9y9z) of the Diophantine equation

x3 + y3 + s 3 + 6xyz = 0.

5, Proposed by Noam Elkies.

Three disjoint spheres whose centers are not collinear are such that there
exist eight planes each tangent to all three spheres. The points of tangency of
each of these planes are the vertices of a triangle. Prove that the circumcenters
of these eight triangles are collinear.

R, Proposed by Noam Elkies.

Let p be an odd prime, and let n be a factor of
1
IT P "+

Prove that n = 1 (mod 8p).

7, Proposed by Brian Hunt.

In if let X = (# l 5 x2> ..., xn)> Y = (yl9 y2i .... yn)9 and, for p e (0,1),
define

y*,Y) -= (i^'ifti 1 "*. i f f i ^ i 1 * - . ifI P IA "p>-

Prove that

l|X|| + Vll * \\F ( X , Y ) | | ,

II 11^ II \lm II p \ I llm»

where

nxnm = ( k r + i « 2 r + . . . + knn1/m.
8B Proposed by John Steinke.
Determine all six-digit integers n such that n is a perfect square and the
 - 302 -
number formed by the last three diqits of n exceeds the number formed by the first
three digits of n by 1. (n might look like 123124.)

Editor's note. All communications about this column should be sent to Professor
M.S. Klamkin, Department of Mathematics, University of Alberta, Edmonton, Alberta,
Canada T6G 2G1.
5f: it &

PR0BLEMS--PROBLEMES

Problem proposals and solutions should be sent to the editor, whose address
appears on the front page of this issue. Proposals should, whenever possible, be
accompanied by a solution, references, and other insights which are likely to be
of help to the editor. An asterisk (*) after a number indicates a problem submit-
ted without a solution.
Original problems are particularly sought. But other interesting problems
may also be acceptable provided they are not too well known and references are
given as to their provenance. Ordinarily, if the originator of a problem can be
located, it should not be submitted by somebody else without his permission.
To facilitate their consideration, your solutions, typewritten or neatly hand-
written on signed, separate sheets, should preferably be mailed to the editor before
May 1, 1983, although solutions received after that date will also be considered
until the time when a solution is published.

791, Proposed by Alan Wayne, Holiday3 Florida.

At whom might this be shouted?

SO
SCAT
SCAT
5032

The sum of this base ten cryptarithm will answer the question.

792, Proposed by E.J. Barbeau, University of Toronto.

The number 144 has the remarkable property that, with only the three
exceptions

2448 = 17-144, 2736 = 19*144, 3312 = 23*144,

each of its first 50 multiples differs from the first perfect square not less than
it by a perfect square.
Show that, for any positive integer n, there is a numberm such thatn consecutive
positive multiples

km, (k+l)m9 ..., (k+n-l)m

have the property that each differs from the smallest perfect square not less than
it by a perfect square.
 - 303 -
7^31 Proposed by V.N. Murty3 Pennsylvania State University, Capitol Campus.
Consider the followinq double inequality for the Riemann Zeta function:
for n - t92,3,...,

*"
(e-l)(n+l)(n+2)...(n+8-i) T+ C0/(a)<C(e)
V ^* ; < VC^O OT+ (s-l)n(n+l)...
, 0 _ „ W „ . A (n+s-2)
T^r-TT* (1)

where
oo 72
t,(s) = I — and c (e) = 7 — .
s n
k=l k k=l ks

Go as far as you can in determining for which of the inteqers s = 2,3,4,... the
inequalities (l) hold.
(N.D. Kazarinoff asks for a proof that (l> holds for s = 2 in his Analytic
Inequalities, Holt, Rinehart & Winston, 1964, paqe 79; and Norman Schaumberger
asks for a proof or disproof that (1) holds for s = 3 in The Two-Year College Mathe-
matics Journal* 12 (1981) 336.)
79^1, Proposed by J.T. Groenman3 Amhem3 The Netherlands.
Determine the positive inteqers n for which

(a) lim exp(tan\) - exp(sinna;)

x+0 n+2
x
(b) lim exp(a:n) - exp(sin\) _
x+0 n+2
x
(c) lim exp(tannaQ - exp(xn) _±
x+O n+2
x
795, Proposed by Jack Garfunkel3 Flushing, N.Y.
Given a trianqle ABC, let ta>tb,tc be the lenqths of its internal angle
bisectors, and let T 9T,9T be the lenqths of these bisectors extended to the cir-
a b c
cumcircle of the trianqle. Prove that
T + T, + T >% {t + U + t ) .
a b c 3 a b c

796, Proposed by Michael W. Ecker3 Pennsylvania State University 3 Worthington

Scranton Campus.
For an inteqer n > 1, let fin) be the product of all the positive divisors of
n other than n itself. Characterize the fixed points of /, that is, characterize
the inteqers n > l such that f(n) = n. (For example, f(S) - l-2-M- = 8.)
 - 304 -

7Q7. Proposed by H. Kestelman, University College, London, England.

Show that the trace of a real n*n matrix A is eaual to xTAx for some
real column vector x with xTx = n.

798t Proposed by Stanley Rabinowitz, Digital Equipment Corp., Merrimack, New

Hampshire.
For a nonneqative inteqer n, evaluate

7991 Proposed by Allan Wm. Johnson Jr., Washington, B.C.

Can the 25 consecutive primes 1327, ..., 1523 be rearranged into a fifth-
order maqic square?

800, Proposed by Charles W. Trigg, San Diego, California.

Find triads of consecutive triangular numbers whose products are square
numbers.
;': s'c s':

THE HOUSE NUMBERS OF EUROPE

Georq Freiherr von Vega was born on 23 March 1754 in Zagorica (Krain), in the
Austrian Empire, and murdered on 17 September 1802 near Vienna. In this computer
age, it may be permitted to recall that he published his Thesaurus Logarithmorum
Completus in Leipzig in 1791 (to ten decimal places). Its 94th edition (seven
decimal places, no misprints, no errois) was published in Berlin in 1935. It had
become ever more reliable and was still widely used after a century and a half.
Numerous checks and calculations to many decimal places (e.g.,
log cos 24°55'30" = -1 + 0.95754 03499 9866
shows that, rounded to seven decimal places, the last digit is 3, not 4) had made
of Vega's Tables a computer's "Bible".
The German philosopher Kuno Fischer (1824-1907) tells from his school days
(the late 1830s) that two of his classmates had a rather simple-minded uncle. He
sometimes looked over their homework, although he rarely grasped what it was all
about. Once he found them doing mathematics, and for the first time in his life
he saw Vega's Log Tables. The big book containing numbers, nothing but numbers,
fascinated him and aroused his curiosity. He asked what it was. One of the boys,
his face suitably despondent, in keepina with the enormous task, answered: "It is
the house numbers of Europe." The uncle said nothing, but did not doubt his word.
After all, why shouldn't numbers be house numbers? That evening, he told friends
that in his own youth school had been difficult enough, but that was nothing compared
to the present. "Why, my poor nephews are sitting at home, and what do they learn?
The house numbers of Europe!" Well, it might come in handy, he mused, if the Germans
ever conquer Paris again. Wouldn't it be helpful to know all the house numbers
right away?
HAYO AHLBURG
Benidorm, Alicante, Spain
 - 305 -
SnU'TIONS

No problem is ever permanently closed. The editor will always be pleased to

consider for publication new solutions or new insights on past problem.

687, C1981: 276] Proposed jointly by J. Chris Fisher., University of Reginaj

and Roger Servranckx, University of Saskatchewan at Saskatoon.
(a) Show that there exists a number y such that the equation
x
c
C - X

has three solutions whenever o < o < y.

(b) How many solutions does the equation

x
Q

o'
O - X

have when there are 2n e's in the ladder and o < c < y?
Solution by the proposers.
(a) For a fixed c e (0,1), let
x
fix) := o° :
then f(o) = c > o, /(») = i, and fix) is strictly increasinq. Thus, if we find a
solution to the equation fix) - x for which flix) > l, then there will be at least
three solutions. Every c in (0,1) can be written uniquely as c = a~a for some a > l.
Then o1/a = i/a and fil/a) = l/a. Furthermore,

fix) = (lnc)2cc cx,

and so f il/a) = (In a ) 2 > 1 for a > e. Note that a > e if and only if o < e~e,
so we let y = e~e . To see that there are at most three solutions, note that

f"ix) = C\r\c)3c° oX(l + ox]nc)

has exactly one zero, x = {-ln(-ln c)}/lnc, and therefore fix) has only one inflec-
tion point.
(b) With f{x) as defined above, let

fix) = fix), f-+±ix) = fifix)), n = 1,2,3,... .

We show that f'ix) = x has three solutions for all n when c e (o,y). Let x - s\
and x - s% be the first and third solutions, respectively, of fix) = x. (The middle
 - 30R -
solution is always x = l/a according to part (a).) We have o < f' (sj) < 1, so si
can be approximated by the iteration process: /"(x) -> s\ a s m - ^ » for any x < l/a.
Even more, o < fix) < s\ for x < sj, so f * ( x ) increases to s\\ consequently,
/"ix) = x for no x < s\. Similarly, f * i x ) decreases to si a s m - ^ » for a?€(si, l/a),
so that f " i x ) = x for x\o x e is\, l/a). The analogous arguments for s 2 imply that,
for any n, f ' ( x ) - x can have no solution except for s l 5 l/a, and s 2 . In fact, as
n -* °°, /*(#) approaches the step function

/ si, for a? < l/a,

^*} =
(s2s for a: > l/a.
A comment was received from LEROY F. MEYERS, The Ohio State University.

Editorrs comment.
Meyers noted that the results of this problem, and a great deal of related
material, can be found in a recent article by Knoebel [l]. In that reference,
note particularly Figure 4, which shows that fix) = x has three solutions if
o < a < e~e, one solution if e~e < c < 1, and two solutions if 1 < c < e .

REFERENCE

1. R. Arthur Knoebel, "Exponentials Reiterated", American Mathematical

Monthly3 88 (1981) 235-252.
* J'C *

688, C1981: 276] Proposed by Robert A. Stumps Hopewell, Virginia.

Let o denote a binary operation on the set of all real numbers such that,
for all real numbers a9b9c,
(i) 0°a = -a; (ii) a°(fc°c) = <3°(&°a).
Show that a°(&°<?) = (a°£)°(-c).
Solution by Stanley Rabinowitz, Digital Equipment Corp.3 Merrimack, New Hamp-
shire.
For all real numbers a9b9c9
a°(b°c) - (o(-a))°(&°<?) = (o°(0°a))°(£°e) = (a°(0°0))°(£°c)
= (a°(-0))o(b°c) - (a°0)°(b<>c) = c°(b°(a°0))
= co(oo(aab)) - (aob)o(Ooc) = (a<>fc)°(-c). •

When doing a problem like this, I always like to know if there can be such
an operator, just to make sure that I am not playing the null game. In this case,
at least one such operator exists, namely, that defined by a°b - a-b.
 - 307 -
Also solved by JAMES J. BOWE, Erskine College, Due West, South Carolina; E.C.
BUISSANT DES AMORIE, Amstelveen, The Netherlands; CURTIS COOPER, Central Missouri
State University; CLAYTON W. DODGE, University of Maine at Orono; JOEL ERICKSON,
student, Fort Lewis College, Durango, Colorado; G.C. GIRI, Midnapore College, West
Bengal, India; J.T. GROENMAN, Arnhem, The Netherlands; J.D. HISCOCKS, University of
Lethbridge, Alberta; PETER A. LINDSTROM, Genesee Community College, Batavia, N.Y.;
ANDY LIU, University of Alberta; LEROY F. MEYERS, The Ohio State University; DAN
SOKOLOWSKY, California State University at Los Angeles; DAVID R. STONE, Georgia
Southern College; ANDREAS STRIEBICH, student, Goethe Gymnasium, Gaggenau, West
Germany; RAM REKHA TIWARI, Radhaur, Bihar, India; ROBERT TRANQUILLE, College de
Maisonneuve, Montreal, Quebec; and the proposer.

Editor's comment.
Nearly all solutions received for this easy problem were essentially equivalent,
so presentation was the deciding factor in selecting one for publication. In the
editor's opinion, the one we have featured won by a whisker.
Meyers noted that the restriction to real numbers is unnecessary, since the
desired conclusion holds for the elements of any group, or, more generally, for
the elements of any set in which a unary operation "-" (such that -0=0 for some
element 0 and -(-a) = a for all elements a) and a binary operation "o" are defined
which satisfy (i) and (ii).
S'I & &

689, [1981: 276] Proposed by Jack Garfunkel, Flushing, N.Y.

Let 777 ,777-,777 denote the lenqths of the medians to sides a,b9c9 respective-
a b c
ly, of triangle ABC, and let M ,AL ,M denote the lengths of these medians extended
to the circumcircle of the trianqle. Prove that
M fm + Mn /X + M /m > 4.
a a b b 0 0

I. Solution by Stanley Rabinowitz, Digital Equipment Corp., Merrimacks New

Hampshire„
Let G be the c e n t r o i d of the t r i a n g l e , and l e t AG, BG, CG meet the c i r c u m c i r c l e
again i n A ' , B ' , C , r e s p e c t i v e l y . I t i s known (see |~1] or [ 2 ] ) that

i i . J G CG_ (1)
GA* + G F +
GC " d
* K J

It is easily shown (or see [3]) that, for positive p9q,r,

1 1 1 9
- + - + ->
p q r p+q+r
with equality just when p = q = r. Applying this to (1), we obtain
 - 308
whence, since M /AG = l + GA'/AG, e t c . ,
M /AG + AL/BG + M /CG > 6.
a £ c

Finally, since AG = (2/3)m , etc., we have

M /m + M^/m, + M /m > 4.
a a b b o o

II. Solution by W.J. Blundon, Memorial University of Newfoundland.

2 2 2 2
From m = (2b +2c -a )/^ and two similar results, we get
8
2 2 .8 2 4
2
9 b 9 e 9 a
and two similar results. From m (M -m ) = (a/2) 2 , we now qet
a a a
2 o . 2
a 4 9a 9Z? 9 e

and so

Va = I + lH/mP +
¥m»
With this and two similar results, we obtain, with sums cyclic over a9b9c9

W lm
a a - I + ¥HK + ml/mP * *>
since x1 + l/x2 > 2, with equality just when m = mh = m , that is, just when the
triangle is equilateral.
III. Solution by M.S. Klamkini University of Alberta.
Since m (M -m ) = a 2 A , etc., the proposed inequality is equivalent to
a a a
2 2
= (2)
ITW? W+z^-a* - *•
a
where the sums, here and later, are cyclic over a9b9c. It is a known result
(see [H] or my comment following my solution of Crux 589 [1981: 308]) that if
Ka9b,c9m m^ m ) > 0
a 9 b9 Q

is a valid inequality, then so is the dual inequality

Km m. ,m ,%aAbAo) > 0,
a9 b o

and conversely. The dual of (2) is

4/77
1
a
L
9a* 9a*
 - 309 -
and this is equivalent to
2
(b c2s ,c2 a2, ,a2 b2, e

which follows immediately from x2 + 1/x2 > 2. There is equality if and only if the
triangle is equilateral. •
It is known [5] that, if a,b,c are the sides of a triangle, then so are an,
n n
b ,c for anynsuch that o < n < l. With n = J, it therefore follows from (2) that

L
2b+2e-a ~ '
which is not so easy to prove directly.
More generally, I show that, for k > l,

K6)
l<k(b+c)-a ~ 2k-V
with equality if and only if a = b = c.
Proof. Let x = s-a9 y = s-b9 z = s-os where s is the semi perimeter; then
a = i/+2, b = z+x, a = x+y,
and (3) is equivalent to

2(fc-l)Ti > (bk-7)TiT2 + 9(fe+l)T3,

where Tx = £#, T2 = Zj/s, T3 = rci/s, with sums cyclic over x,y9z. Since
r 3 < T\/27 and ^ ^ < Ti/39

we have
3 3
(Sk-DTM + 9<fc+l)T3 < r5^"37)^ + 9 ( ^ ) T l = 2(fc-l)Ti.

and so (3) is established. •

Finally, I leave as open problems to find all fixed k > 2 such that

y z _ az 2, z > 3
Lk(b +c )-a ~ 2k-l>
and, even more generally, to determine all fixed k and n such that
n Q
y s > 3.
7 /tW ?^v w 2k-i
k(b +o )-a
Also solved by J.T. GROENMAN, Arnhem, The Netherlands; GALI SALVATORE, Perkins,
Quebec; GEORGE TSINTSIFAS, Thessaloniki, Greece; and the proposer. A comment was
sent by S.C. CHAN, Singapore ; and one incorrect solution was received, wherein it
is "proved" that the inequality of the problem must be reversed!
 - 310 -

REFERENCES

1. Thu Pham, solution to Problem 1119 (proposed by K.R.S. Sastry), Mathematics

Magazine, 55 (1982) 180-182.
2. Georqe Tsintsifas, Problem 723 (b) s this journal, 8 (1982) 77.
3. S. Barnard and J.M. Child, Higher Algebra* Macmillan, London, 1936 (re-
printed many times), p. 224, Ex. 4.
4. Aufqabe 677, solution by M.S. Klamkin, Elemente der Mathematik .
5. Solutions to Problem E 1366 (proposed by V.E. Hoqqatt, Jr.), American Mathe-
matical Monthly, 67 (1960) 82-84.
ft ft ft

5^)0, [1981: 276"! Propose par Hippolyte Charles3 Waterloos Quebec,

n £tant un entier positif donn£, soit
n+k in\, . . n+k ,n%, _. n+k . . M-1/ n %
S = n
k " (i' ( n " 1 ) +
( 2'(n " e,°+ (_1)
Vl'.
Calculers, pour k = osi,2,3.
I. Solution by Gali Salvatore3 Perkins^ Quebec.
Let
fix) = (ex - l)n , (1)

that is,
fix) = em - (*fe<«-l>* + (nyn-2)x _ _ + ^n-1 (n_y + ( _1}^

D i f f e r e n t i a t i n g n+k times, we qet

Jn+k), . _ n+k nx i n\, , v +k in-l)x . (K\, n,n+k in-2)x w ,^n-lf n sx
f ix) = n e \ 1 )in-l) e + \ 2Kn-2) e - . . , + (-1) (n_j)e »

from which it is clear that s* - f (n+k) (o). Therefore s, = (n+fc)! x the coeffi-
cient o f / + in the Maclaurin expansion of (l). This expansion, as far as we need,
is
4ft \ J1 . n n+1 , «(3«+l) n+2 n2in+l) n+3
fix) = a* + ^c + ———x + —^ - g — ^ + ... ,

and therefore (' .

So = n\ 9
a - in+l)ln
2
(2)
„ (n+2)!n(3n+l)
9
^2 = JH
o _ (n+3)ln2in+l)
S3 =
48 "
 - 311 -
II. Solution by V.N. Murty_, Pennsylvania State Universitys The Capitol Campus.
Let n distinguishable boxes and n+k distinguishable balls be given, and let E,
be the event that no box is empty after the n+k balls have been distributed ran-
domly among the n boxes. The evaluation of Pr (E. ) is a classical occupancy problem.
Feller has shown Ci] that

nn+kPr(Ek) = Sr
We easily (but tediously) find that

Pr(E0)=^, Pr(El)~.L-i^L, Pr (%) = <K±22M^±il, Pr( ^) = (2±32l2£i5iii,

n 2n 24rc 48n

from which the results (2) follow.

III. Comment by M.S. Klamkin, University of Alberta.
These results are known and can be obtained by means of Stirling Numbers of the
second kind. The values reguired here, and more, are given in a table in Gould [2].
Also solved by G D P . HENDERSON, Campbellcroft, Ontario; KESIRAJU SATYANARAYANA,
Gagan Mahal Colony, Hyderabad, India? and the proposer.

REFERENCES

1. William Feller, An Introduction to Probability Theory and its Applications*

Vol. l, Third Edition, Wiley, New York, 1968, p. 60 (Eguation 11.6) and p. 65
(Problems 16 and 17).
2. H.W. Gould, Combinatorial Identities, Morgantown, West Virginia, 1972, p. 3,
where the reference is to "A remarkable combinatorial formula of Heselden", Proc
W, Va. Acad, of Sci.

691, [1981: 301] Proposed by J.A. McCallum, Medicine Eats Alberta.

Here is an alphametic about the man who coined the word "alphametic",
the well-known author of the syndicated column Fun With Figures, J.A.H. Hunter:

H E«S
THAT
FUN
MATHS,
HUNTER
The apostrophe has no mathematical significance and the answer, l i k e the man himself,
is unigue.
Solution by Clayton W. Dodge3 University of Maine at Orono.
It is immediately apparent that H = 1 5 M = 9S and U = o. Now
 - 312 -
2S + T + N = R + 10a 9 a = 1 or 2,
a + A + 1 = 10,
3 + F = 10,

and
1+T+A=N+10.

Thus F = 7, A = 8, and a = l. Now T = N + l; hence 2S + 2N = R + 9 and R is odd,

that is, R = 3 or 5 and S + H = 6 or 7, respectively. But R = 5 must be rejected,
since then {S,N} = {3,4} and there is no value for T = N + l. Thus R = 3, S = 2 S
N = 4, T = 5, which leaves E = 6. The uniaue solution is
162
5185
704
98512
10M-563

Also solved by E.G. BUISSANT DES AMORIE, Amstelveen, The Netherlands; RICHARD I.
HESS, Rancho Palos Verdes, California; J.A.H. HUNTER, Toronto, Ontario; W.C. IGIPS,
Danbury, Connecticut; ALLAN WM. JOHNSON JR., Washington, D.C.; STANLEY RABINOWITZ,
Digital Equipment Corp.,Merrimack, New Hampshire; DONVAL R. SIMPSON, Fairbanks, Alaska;
ROBERT TRANQUILLE, College de Maisonneuve, Montreal, Quebec; CHARLES W. TRIGG, San
Diego, California; KENNETH M. WILKE, Topeka, Kansas; DAVID ZAGORSKI, student, Massa-
chusetts Institute of Technology; and the proposer.

692.ri981: 301] Proposed by Dan Sokolowsky3 California State University at Los

Angeles.
S is a set of n distinct objects. For a fixed k > l, 2k subsets of S are
n n
denoted by A.,B.9 i - i,...,fc. Find the larqest possible value of n for which the
following conditions (a)-(d) can hold simultaneously for i - i,...,fc.
(a) A. u B. = s .
% ^ n
(b) A . n B. = 0.
i i
(c) For each pair of distinct elements of s , there exists an i such that
the two elements are either both in A . or both in 5..
(d) For each pair of distinct elements of s , there exists an i such that one
n
of the two elements is in A . and the other is in £..
^ ^
Solution by Curtis Cooper3 Central Missouri State University, Warrensburg,
Missouri.
Let k > 1 be a fixed inteqer and let S = {1,2,...,w>. We will show that the
answer to our problem is n = 2 ~ . But first we must set up some machinery.
 - 313 -
Let E be the set of vertices of the ^-dimensional unit "cube" (more preciselys
E is the Cartesian product {0,1} ) . For all x,y e E, we define

x ~y <=> x - y or x + y = (1,1,... ,1),

where the addition in E is, as usual, coordinatewise. Then ~ is an equivalence
k-l
relation on E which partitions E into 2 equivalence classes. Number these classes
from 1 to 2 and pick a representative P. from the tth equivalence class.
The desired result is a consequence of the following theorem:
k-l
(i) If n < 2 3 then there exist 2k subsets of S which satisfy (a)-(d),
n
k-l
(ii) If n > 2 3 then no collection of 2k subsets of S satisfy all of (a)-(d).
The theorem is clearly true if k = 1, so we assume that k > 1.
k~l
"Proof of (i). Suppose n < 2 and, for i = 1,2,...,£, let
A . = {j I 1 < j < n and the ith coordinate of P. is 0}

and
B. = {j I 1 £ j < n and the ith coordinate of P . is 1}.
Clearly, (a) and (b) hold for i = 1,2,...,&. To show that (c) and (d) hold, let
1 < i9j < n with i * j . Since P. and P. are in different equivalence classes,
there exist integers t 0 and j 0 , 1 < t 0st 7 0 ^ &» s u c n that
i 0 th coordinate of P. = t 0 th coordinate of P.

and
j 0 th coordinate of P. * jo^h coordinate of P..
t 0
Thus € and j are either both in A. or both in B. , and one of i and J is in 4. and
to *o Jo
the other is in B . .
JO VA

Proof of (ii). Suppose n > 2 and let >!.,£., j = l,2,...,fc, be 2fc subsets of
V V

S . If the subsets i4.,B. satisfy (a) and (b), then there is a function f;Sn + E
defined by
(0, if i e A.,
d
jth coordinate of f{i) = I
1, if i e B. 9
J
fc-l
for t = l,2,...,n and j = l,2,...,k. Since n > 2 , there exist distinct integers
t0 and Jo, 1 < to Jo * «» such that /(to) ~ f(jo)9 and either (c) or (d) is fa1se0
Also solved by J.D. HISCOCKS, University of Lethbridge, Alberta; and the proposer.
Comments were received from PAUL R. BEESACK, Carleton University, Ottawa; and DAVID
R„ STONE, Georgia Southern College, Statesboro, Georgia.
 - 311 -
6^3, [1981: 301] Proposed by Fervell Wheeler, student, Texas A & M University.
On a ux^ tick-tack-toe board, a winninq path consists of four squares in
a row, column, or diaqonal. In how many ways can three X's be placed on the board,
not all on the same winninq path, so that if a qame is played on this partly-filled
board, X goinq first, then X can absolutely force a win?
Editor's comment.
No solution was received for this problem, which therefore remains open.
it »'• &

694, ("1981: 3021 Proposed by Jack Garfunkel, Flushing, N.Y.

Three conaruent circles with radical center R lie inside a given triangle
with incenter I and circumcenter 0. Each circle touches a pair of sides of the
trianqle. Prove that 0, R, and I are collinear.
(This generalizes Problem 5 of the 1981 International Mathematical Olympiad
[1981: 2231, where it was specified that the three circles had a common point.)
Solution by Roland H. Eddy, Memorial University of Newfoundland,
This proposal is essentially the same as the oriqinal International Olympiad
problem. The radical center R, irrespective of the position of the three circles,
is simply the circumcenter of the trianqle formed by joininq the centers of the
three circles. This trianqle and the oriqinal one are nomothetic with nomothetic
center I. Hence the circumcenters 0 and R, which are corresponding points in this
homothecy, are collinear with the nomothetic center I.

Also solved by E.C. BUISSANT DES AMORIE, Amstelveen, The Netherlands; JORDI DOU,
Barcelona, Spain; ROLAND H. EDDY, Memorial University of Newfoundland (second solu-
tion); J.T. GROENMAN, Arnhem, The Netherlands; F.G.B. MASKELL, Algonquin College,
Ottawa; STANLEY RABINOWITZ, Digital Equipment Corp., Merrimack, New Hampshire;
KESIRAJU SATYANARAYANA, Gagan Mahal Colony, Hyderabad, India; and GEORGE TSINTSIFAS,
Thessaloniki, Greece.
s'c ft &

695, [ 1 9 8 1 : 302 ; 1982: 301 ( C o r r e c t e d ) Proposed by J.T. Groenman, Arnhem, The

Netherlands.
For i = 1,2,3, A. are the vertices of a triangle with sides a. and excircles
with centers I. touching a. in B.. For j 9k * i9 M. are the midpoints of B.B.; and
m. are the lines through M. perpendicular to a.. Prove that the m. are concurrent.
Solution by Kesiraju Satyanarayana, Gagan Mahal Colony, Hyderabad, India.
We will for typographical convenience rename the trianqle ABC and its sides a9b9c
in the usual order (with r, i?, and s having their usual meaninqs), but leave the
rest of the notation unchanged. Suppose mi9m29m3 meet a9b9c in D,E,F, respectively.
 - 315 -
I t i s well known T l ] t h a t the m. arei concurrent
concurrei i f and only if

(BD2 - DC2) + (CE2 - EA2) + (AF2 - FB2) = 0.

Referring now to the f i g u r e , we have

BD2 - DC2 = (BD + DC)(BD - DC)

= a(BB£ - B^C)
= a ( s - a ) ( c o s B - cos C)

= ( 4 i ? s i n ^ c o s ~ ) ( r c o t ^ ) ( c o s B - cos C)

= 4i?2*C0S2r-(C0S B - COS C)

= 2i?p(l + cos A)(cos B - c o s C ) .

With t h i s and two s i m i l a r r e s u l t s , we have, w i t h sums c y c l i c over A,B,C,

l(Bp2 DC2)
2i?r = l ( c o s B - cosC) + FcosA(cosB - cosC) = o,

and the desired r e s u l t f o l l o w s .

Also solved by DAN SOKOLOWSKY, California State University at Los Angeles; and
the proposer. Comments (pointing out the error in the original formulation of the
problem) were received from JORDI DOU, Barcelona, Spain? ROLAND H. EDDY, Memorial
University of Newfoundland; and STANLEY RABINOWITZ, Digital Equipment Corp., Merrimack,
New Hampshire.

REFERENCE

1. H.S.M. Coxeter, Introduction to Geometry* Second Edition, Wiley, 1969, p. 16,

Ex. 6.
 - 316 -
F^F,T1981: 302] Proposed by George Tsintsifas, Thessaloniki, Greece,
Let ABC be a t r i a n g l e ; a9b,c i t s s i d e s ; and s9r,R i t s semi perimeter,
inradius and circumradius. Prove t h a t , with sums cyclic over A,B,C9
(a) S + 4£cosJ(P-C) > E C O S A ;
(b) ZaCOSi(B-C) > s(l + 2r/R).

Solution by V.N, Murty, Pennsylvania State University, Capitol Campus.

(a) From
B-C = c o sB+C
cos — . B . C|
— + n2 sin-sin

= sin^ + (2nsin^)csc^

and two similar results (sums and products throuqhout are cyclic over A,B,C), we
net
z c o s — ^ = E s i n ^ t (2nsin~)zcsc^; (1)

and from
A R-f A
cos B + cos C = 2 s i n ^ c o s ^ - 2 s 1 n ^ <2)

and two similar results, we net

A
E sin - > E cos A. (3)

We will also use the well-known inequality [l, p. 31]

A
E esc - > 6. (4)

Now, from ( l ) , (3), and ( 4 ) , we obtain

z c o s — ^ > E cos A + I 2 n s i n ^ . (5)

Finally, we apply to (5) the classical result

E cos A = i + m sin -

and obtain
R-P
E C O S - — > ECOSA + 3 ( E C O S A - l ) = 4E COS A - 3 , (6)

which is equivalent to the inequality proposed in part (a). Equality holds in (3)
and (4), and hence in (6), iust when the trianqle is equilateral,
(b) It is easy to show (or see [2]) that
Eacos — - = rCZH<5)s1n| ,
 - 317 -
and them, from (2), we have
A
l(b+o)sir)^ > ii(b+o)(cosB + cosC).

Now, from a = b cos C + ccosB, etc., we have

JE(£+C)(COS B + cosC) = s + la cos A,

where s is the semi perimeter. Thus

R-P
la cos -^- > s + EacosA.

Now, with K and i? the area and circumradius of the trianale, respectively, we have

la cos A = Rl sin 2A = 4 M sin A = ~ = ^ ,

XT XI

and therefore
(7)
racos^,a(l + f).
In all the inequalities we have used, and hence in (7), equality holds just when
the triangle is equilateral. •
If we compare
E C O S ^ > 4ECOS A - 3 (6)

and Garfunkel's inequality (Crux 613 [1982: 138])

ECOS^£ > --LssinA, (8)

i t is natural to ask which inequality is sharper. With

x = | = E cos A - 1 and y = | = E sin A,

XT XT

(6) and (8) can be expressed as

I COS 5 l £ > 1 + i\X (9)

and
ECOS^ > 2|. (10)

With the inequalities in this form, our question is easily answered.

It is known [1, p. 22] that o < x < J. For the notation and other results I
will now use, see my article recently published in this journal [1982: 62-68].
For all trianqles of type I, we have y > /3(l+#), so
 - 318 -

% > 2(1 +x) > 1 + hx,

and (10) is sharper than (9). The line

y = &{1 + ija?)

divides the type II region (shaded with horizontal lines in the figure on page 65)
into two parts, in one of which we have

2 (l+4ff) < y < /3(l-w?) (11)

and in the other

0 < y < /3
~(1 + to). (12)

Inequality (10) is sharper than (9) for the type II triangles which satisfy (11) 9
and (9) is sharper than d o ) for the type II triangles which satisfy (12).
Also solved by KESIRAJU SATYANARAYANA, Gagan Mahal Colony, Hyderabad, India;
and by the proposer.

REFERENCES

1. 0. Bottema et al., Geometric Inequalities, Wolters-Noordhoff, Groningen,

1969.
2. C.V. Durell and A. Robson, Advanced Trigonometry, G. Bell and Sons, London,
1959, p. 3, Ex. 36.

P97, [1981: 302] Proposed by G.C. Giri3 Midnapore College^ West Bengal^ India.
Let

a - tan e + tan $,
b = sec e + sec $,
c = esc e + esc <j>.
If the angles e and <j> are such that the requisite functions are defined and be * o 9
show that 2a/bc < 1.
Solution by W.J, Blundon^ Memorial University of Newfoundland.
Let u - tan (0/2) and v - tan (<f>/2). It follows from the hypothesis that neither
u nor v can eaual any of -l, o, l. Now
2u 2v 2(
.. _ u+v)(l-uv)
a
" TU? +
T^v7 (±-uz)(i-v2)'
 - 319 -
1+u2 1+t
h . + ^ 2(1+UV)(1-U.V)
+
TIF T^v7 ~ '(i-uz)(i-vz) 9
lm2 lw2 _ (u+v)(l-ntv)
° 2u 2v 2uv
Therefore;

be (l+uv)z x
(l+uv)* ^ 1
'

The inequality is strict, since uv = 1 implies £ = o, contrary to the hypothesis.

Also solved by E.C. BUISSANT DES AMORIE, Amstelveen, The Netherlands; JACK
GARFUNKELf Flushing, N.Y.; J.T. GROENMAN , Arnhem, The Netherlands; W.C. IGIPS,
Danbury, Connecticut? V,N. MURTY, Pennsylvania State University, Capitol Campus;
KESIRAJU SATYANARAYANA, Gagan Mahal Colony, Hyderabad, India; and the proposer.
A comment was received form STANLEY RABINOWITZ, Digital Equipment Corp., Merrimack,
New Hampshire.

Editor's comment.
Several solvers proved instead that be > 2a, which they claimed was equivalent
to 2a/bc < l- This would have been true if the proposal had read: "Prove that
be > 2a if b andc are of the same sign." (This, in fact, is exactly the wording
the proposer had used, and Rabinowitz found the problem, in the exact notation and
wording of the proposer, in Hobson [13.) But the proposal as modified by the
editor is slightly more general and allows the possibility be < o, so be > 2a is
not equivalent to 2a/be < 1.

REFERENCE

1. E.W. Hobson, A Treatise on Plane and Advanced Trigonometry, Dover, New York,
1957, p. 101, Ex. 6M-.
ft & &

698, C1981: 302] Propose" par Eippolyte Charles, Waterloo, QuSbee.

Les sommes partielles de la seYie harmonique (laquelle, on le sait bien,
est divergente) sont dgfinies par

00

La se>ie £ - est-elle converqente ou divergente?

n=l n
Solution de Leroy F. Meyers, The Ohio State University.
On voit ais£ment que
 - 7 ?0 -
Done l/e > l/n, et la serie proposee diverqe par comparaison avec la seYie harmo-
nique.
Also solved by BERNARD BAUDIFFIER, College de Sherbrooke, Quebec; PAUL R.
BEESACK, Carleton University, Ottawa; CLAYTON W. DODGE, University of Maine at
Orono; JORDI DOU, Barcelona, Spain; MICHAEL W. ECKER, Pennsylvania State University,
Worthington Scranton Campus; MILTON P. EISNER, Mount Vernon College, Washington, D.C.;
YVES GALIPEAU, College de Sherbrooke, Quebec; JACK GARFUNKEL, Flushing, N.Y.; J.T.
GROENMAN, Arnhem, The Netherlands; RICHARD I. HESS, Rancho Palos Verdes, California;
J.D. HISCOCKS, University of Lethbridge, Alberta; FRIEND H. KIERSTEAD, JR., Cuyahoga
Falls, Ohio; F.G.B. MASKELL, Algonquin College, Ottawa; V.N. MURTY, Pennsylvania
State University, Capitol Campus; KESIRAJU SATYANARAYANA, Gagan Mahal Colony, Hy-
derabad, India; DAN SOKOLOWSKY, California State University at Los Angeles; DAVID R.
STONE, Georgia Southern College; KENNETH S. WILLIAMS, Carleton University, Ottawa;
and DAVID ZAGORSKI, student, Massachusetts Institute of Technology.

Editor's commment.
Several (but by no means all) of the solutions to this easy problem were equiv-
alent to the trivial one qiven above. We chose to feature one submitted in French
precisely because the problem was easy, to enable our readers to improve their
French, if not their mathematics.
* ii «

6^9, C1981: 302] Proposed by Charles W. Trigg3 San Diego3 California.

A quadrilateral is inscribed in a circle. One side is a diameter of the
circle and the other sides have lenqths of 3, 4, and 5. What is the length of the
diameter of the circle?
Solution by the proposer.
Generalizina the problem, we let the required diameter be x, the three given
sides a9b9a9 and the diagonals e9f. As indicated in the figures, there are three
possible situations:
I. a convex quadrilateral;
II. a crossed quadrilateral with two vertices on one side of the diameter;
III. a crossed quadrilateral with vertices on opposite sides of the diameter.
Case I. Applyinq the theorems of Ptolemy and Pythaqoras, we have
ef = ae + bx, a2 + f2 = x2, e2 + c2 = x2.
Elimination of e and / leads to (x2-o2)(x2-a2) = (ac+bx)2, which, since x * 0, is
equivalent to

x3 - (a2+b2+c2)x - 2abo = 0 . (1)

(This result has appeared frequently in the literature. See, e.g., Cl], where more
references are aiven.) With the qiven values of a,b,o, equation (1) becomes
 - 321 -

-^f

III a III b

x3 - 50x - 120 = 0. (2)

This equation has one positive and two neqative roots, the positive root beina

x = l ^ l c O S (|ArCCOS ?§) « 8.05581036.

(For information as to how this root was obtained, see Rausen [2].) We note that,
since the central angles subtended by the chords a,b,c are additive on one side
of the diameter, the chords may be taken in any of 3! = 6 orders.
Cases II and III. In case II, we have
ao = ef + bx, e2 + c2 = x2, f2+a2=x29
 - 322 -

and case III qives

bx = ac + e/, a 2 + / 2 = x2, e 2 + c 2 = a:2.

In each of these cases, proceeding as in case I leads to the equation

a:3 - (a2+b2+c2)x + 2abc = 0 S

or, with the qiven values of a9b,c, to

3
fix) E X - 50# + 120 = 0.

This equation has one neqative and two positive roots. Since /(o) - / ( 3 ) < o , one
of the positive roots is less than 3, an impossible diameter with chords greater
than 3. The other positive root, which is acceptable, is

x = ^ C O S ( | - j U r c C O S ^ I ) » 5.18026779.

In a circle with this diameter, there is one crossed quadrilateral (and its reflec-
tion) of type II with sides in the order (4,3,5,x)9 and two crossed quadrilaterals
(and their reflections) of type III with sides in the orders (3,4,5,ic) and (3,5,4,#).

Also solved by ELWYN ADAMS, Gainesville, Florida; PAUL R. BEESACK, Carleton

University, Ottawa; W.J. BLUNDON, Memorial University of Newfoundland; S.C. CHAN,
Sinqapore; CLAYTON W. DODGE, University of Maine at Orono; JORDI DOU, Barcelona,
Spain; MILTON P. EISNER, Mount Vernon College, Washington, D.C.; JACK GARFUNKEL,
Flushing, N.Y.; J.T. GROENMAN, Arnhem, The Netherlands; MARIE JANE HAGUEL, College
de Sherbrooke, Quebec; RICHARD I. HESS, Rancho Palos Verdes, California; W.C. IGIPS,
Danbury, Connecticut; FRIEND H. KIERSTEAD, JR., Cuyahoga Falls, Ohio; LEROY F. MEYERS,
The Ohio State University; STANLEY RABINOWITZ, Digital Equipment Corp., Merrimack,
New Hampshire; KESIRAJU SATYANARAYANA, Gaqan Mahal Colony, Hyderabad, India; DAN
SOKOLOWSKY, California State University at Los Anaeles; ROBERT TRANQUILLE et YVES
LACOMBE, College de Maisonneuve, Montreal, Quebec; DIMITRIS VATHIS, Chalcis, Greece;
KENNETH M. WILKE, Topeka, Kansas; DAVID ZAGORSKI, student, Massachusetts Institute
°f Technology; and one anonymous solver from Georgia Southern College.

Editor's comment.
Did the proposal say that the quadrilateral was convex? It did not. Yet, the
editor reqrets to report, in a rare display of unanimity each of the twenty-three
other solvers tacitly assumed that the quadrilateral was convex and thus was led
to consider only equation (2). Poets may rhapsodize about majority rule (democracy
is for the bards), but here we have a proof that you can fool all of the people some
of the time. Our proposer, an old pro, constitutes an overwhelmina majority of
one. Caveat lectori

REFERENCES
1 , Problem 880 (proposed by Richard Corary) , Mathematics Magazine, 48 (January
1975) 53.
2. John Rausen, "Solving cubic equations on a pocket calculator", this journal,
6 (1980) 5-7.
 - 323 -
701, C1982: 1M-] Proposed by Alan Wayne, Holiday, Florida.
Ever since her X-rated contribution to the solution of Crux 411 [1979: 299]
shocked the old lady from Dubuque, EDITH ORR (long may she swing!) has maintained
a low profile. Yet we all know that she is always THERE at the editor's elbow, like
as not blowing in his ear. So we should all be able to solve the subtraction

EDITH
- ORR.
THERE"
Solution by Sam Baethge3 Southwest High School3 San Antonio3 Texas.
Expressing the problem as an addition,

THERE
+ ORR ,
EDITH
we see immediately that H = 9, D = o 9 and

E = T + 1, E + R = 9, 2R = T + a, (1)

where a = o or 10. Eliminating E and T from (l), we find that 3R = 8 + a, from

which a = 10, so R = 6, E = 3, and T = 2. Now

4 + 0 = I + 10, or 0-1 = 6,

for which the remaining digits give 0 = 7 and I = l as the only possible values.
The given subtraction therefore has the unique solution

30129
- 766
29363 '

Also solved by HAYO AHLBURG, Benidorm, Alicante, Spain; CLAYTON w. DODGE, Uni-
versity of Maine at Orono; MILTON P. EISNER, Mount Vernon College, Washington, D.C.;
MEIR FEDER, Haifa, Israel; J.T. GROENMAN, Arnhem, The Netherlands; J.A.H. HUNTER,
Toronto, Ontario; W.C. IGIPS, Danbury, Connecticut; ALLAN WM„ JOHNSON JR., Washington,
D.C.; ROBERT S. JOHNSON, Montreal, Quebec; FRIEND H. KIERSTEAD, JR., Cuyahoga Falls,
Ohio? F.G.B. MASKELL, Algonquin College, Ottawa; STANLEY RABINOWITZ, Digital Equip-
ment Corp., Merrimack, New Hampshire; RAM REKHA TIWARI, Radhaur, Bihar, India;
ROBERT TRANQUILLE, College de Maisonneuve, Montreal, Quebec; CHARLES W. TRIGG, San
Diego, California; KENNETH M. WILKE, Topeka, Kansas; ANNELIESE ZIMMERMANN, Bonn, West
Germany; and the proposer.

702. [1982: 14] Proposed by George Tsintsifas, Thessaloniki3 Greece.

Given are three noncollinear points Ai,BlsCi and three positive real
numbers l9m,n whose sum is l. Show how to determine a point M inside the circum-
circle of triangle A 1 B 1 C 1 such that, if Ai^Bj M J C J M meet the circumcircle again in
 - ?24 -
A,B,C, respectively, then we have

rBMC] = [CMA] = [AMB] = [ABC1 (1)

I m n '

where the brackets denote the area of a trianqle.

"Partial solution by the proposer (revised by the editor).

Let the qiven points AisBi ,Ci have affixes a\,$i,YIS respectively, in the com-
plex plane, and suppose there exists a point M inside the circumcircle such that
the resulting trianqle ABC satisfies (l). If the affixes of the unknown points
M,A,B,C are z9a9$,y9 respectively, then it is well known that

Z - Id + 7776 + tty. (2)

The power of the interior point 1^ with respect to the circle is

k = (z-a)(z-a\ ) * 09

from which we obtain

Ik 7- 7-
•= Iz - la, /oN
Z-cxi (3)
and similarly
mk - - ,, N
—r- = wz - m$, (4)
z -pi
nk - - /f_x
= nz - ny. (5)
z-yi
Since l+m+n = 1 and k * o, addinq ( 3 ) , ( 4 ) , ( 5 ) , and usinq (2) r e s u l t i n

_1_ + _?L_ + _ J 1 _ = o
z-a\ a-3i S-YI

and clearinq the fractions qives the equivalent equation

z2 - {(m+n)a,i+(n+l)$i+(l+Tn)yi}z + (Z8iYl+^Ylal+^al3i ) = 0. (6)

One root of this equation is the affix of the point M, so M and the resultinq tri-
anqle ABC can be identified.

Editor's comment.
Still unresolved in the above solution of this interestinq problem are the
followinq questions:
(a) The editor would like to receive a simple proof or a reference for the
"well-known" relation (2), which appears to be correct. It would qreatly ease the
editor's burden if solvers were to assume that he knows far less than they do,
and give proofs or references for results that are "well known" to only a few.
 - 325 -
(b) If the point M is assumed to exist, which root s 0 of (6) gives its affix,
and what is the geometrical significance, if any, of the other root?
(c) Conversely, if z0 is the appropriate root of (6), it must be shown that
the point M whose affix is s 0 , and the resulting triangle ABC, satisfy (l).
(d) Can the problem be extended, perhaps by removing the requirement that
l,>m9n be positive and using signed areas?
Readers are invited to answer these questions, or else to find an entirely
different approach to the problem.
* * *
704, C1982: 14] Proposed by H. Kestelman, University College, London, England.
Suppose A\ , A 2 9 . . . , A . are square matrices of any orders and no two have a
common eigenvalue. If f\9f2 /• are a n Y given polynomials, then there is a poly-
nomial F such that F(A ) = f (A ) for all r.
J
r r r
Solution by the proposer.
Let g be the product of the characteristic polynomials of all the A with
s * r. Since the gr are pairwise co-prime, there exist polynomials $x, <f>2, ...s Q
<J>.
such that

r=l

for all t. Set

Fit) = \ *(t)g(t)f(t).
r~±
By the Cayley-Hamilton theorem, g (A ) = 0 if r * s; hence
F( and J =
V = VV^V^V WW
for all s, that is, F(As ) = f s (As ).
jfe * "

7D5, C1982: 15] Proposed by Andy Liu, University of Alberta.

On page 315 of U.I. Lydna's definitive treatise Medieval Religion (pub-
lished in 1947) is a little piece of irrelevancy entitled "Pagan Island". This
island was of circular shape and had along its coast exactly 26 villages whose names,
in cyclic order, had the initials A, B, ..., Z. At various times in its history, the
island was visited by exactly 26 missionaries whose names all had different initial
letters. Each missionary first went to the village which matched his initial. While
more than one missionary miqht be on the island at the same time, no two ever ap-
 - 32F -
peared (or disappeared) in the same village at the same time. As the name of the
island implies, all the villaqes were initially pagan. When a missionary arrived
at an unconverted village, he converted it and then moved on to the next village
along the coast. When a missionary arrived at a converted village, he was promptly
devoured and the village became unconverted again.
While the missionaries' fate was never in doubt, how many villages remained
converted after their combined effort?
I. Solution by Leroy F. Meyersj The Ohio State University.
In the end, Pagan Island remains entirely pagan, since each village receives
exactly two visits from missionaries.
Each villaqe is converted at least once, either by the missionary who lands there
or by the missionary who first converts the preceding village. (At least one village
is converted by the missionary who lands there, so the process can begin.) But in
either case there is another visit, by the first missionary to convert the preceding
village or by the missionary who lands at the village.
On the other hand, no village is visited more than twice. For suppose that some
village is amonq those tied for first to receive a third visit (for reconversion,
after deconversion, after conversion to — what else? — hierophagy, or the devouring
of the next missionary to arrive at the village). Then at least two of the three
visits must have been by missionaries who had previously (re)converted the preceding
village. But a reconversion implies a preceding deconversion. Hence the preceding
village must have already had three visits by the time the village under consider-
ation has had its third, so that the village cannot have been tied for first to
receive a third missionary. Thus no village can be tied for first, and so no village
can have a third visit.
Hence each village receives exactly two visits from missionaries, and so, after
the missionary activity has ended, is aqain pagan.

II. Incorrect solution by X3 who deserves to take on holy orders and receive
a one-way ticket to "Pagan Island,
Let u denote the number of unconverted villages on the island (at any particular
point in time) and let m denote the number of (live) missionaries on the island at
that time. I claim that urn is always 26.
Proof. Initially, u = 26, 777 = 0, so u-m = 26. There are only two events that
can occur:
Eoent 1. Missionary arrives at an unconverted village. Here m increases bv l
and u decreases by 1.
 - 327 -
Event 2. Missionary arrives at a converted villaqe. Here (alas) 777 decreases
by 1 and u increases by 1.
In all cases, u+m is invariant. D
Thus, when all the missionaries were gone, m - 0, and we must have u = 26.

Also solved by HAYO AHLBURG, Benidorm, Alicante, Spain; SAM BAETHGE, Southwest
High School, San Antonio, Texas; PAUL R. BEESACK, Carleton University, Ottawa;
CLAYTON W. DODGE, University of Maine at Orono; ROBERT S. JOHNSON, Montreal, Quebec;
FRIEND H. KIERSTEAD, JR., Cuyahoga Falls, Ohio; ROBERT TRANQUILLE, College de Mai-
sonneuve, Montreal, Quebec; ANNELIESE ZIMMERMANN, Bonn, West Germany; and the
proposer.

Editor's comment.
All solutions credited above were correct, although a few contained more arm-
waving than strict application of logic. All solutions assumed that all missionaries
travel in the same direction, but Kierstead showed in addition that some villages
may remain converted if the missionaries do not all travel in the same direction in
qoing from one village to the next. Beesack proved more generally that if k mis-
sionaries visit the island, where 1 < k < 26, then 26-fc villages remain converted
at the end. The debacle at Pagan Island inspired one solver to guote:
A man convinced against his will
Is of the same opinion still.

This is one of the famous misguotations of literary history (another is "a rose is
a rose is a rose", about which more next month in the solution of Crux 711). The
correct guotation is
He that complies against his will
Is of his own opinion still.

(Samuel Butler (1612-1680), in Hudibras , Part III, Canto III, line 547.)
Our first solver's phrase "conversion to ... hierophagy" shows that cannibalism
is the religion of the missionaries* whose fleeting effects last only until the
next meal of long pig. Meyers also sent with his solution a copy of a cartoon from
Monster Rally, by Charles Addams, showing a group of cannibals seated around a boil-
ing missionary-size kettle. One of them says to another: "Oh, I like missionary,
all right, but missionary doesn't like me." To which we may reply: "Well, just
leave him there and eat the noodles."
f; it &

ONE TWO, BUTTON YOUR SHOE

Good Queen Bess was more than a wonder, she was a Tudor. (Anonymous)
I ate my tenderloin with a fork, then I nine my elevenderloin with a fivek.
(Victor Borge)
 - 328 -
THF PUZZLE CORNER
Puzzle No. 26: Alphametic

The following quiz could be plenty fun9

Just remember to work in base twenty-one.
NOW
WE
ARE
SIX ... a summation of fact;
That ONE divides SIX requires no tact;
That TWO is prime is perfectly valid,
But SIX must be perfect, and thus ends our ballad.
Puzzle No. 27: Alphametic

Prime STEAK from a square COW:

Haven't had it before? At least you've heard of it now.
Once in a while, just as a treat,
Of this I partake 'cause I
EAT
REST'
Puzzle No. 28: Alphametic

SQUARE is a square but Q is not cubic.

There's a cube number of cubes in the form of RUBIK;
Only a cube number of which the first criteria still meet...
All the other solutions you will have to delete.
If I look like a cube it's because of your eyes;
For I've noticed each iris is a square in disguise.
HANS HAVERMANN
Weston, Ontario

Answer to Puzzle No. 23 [1982: 275]: Rhumb, rum.

Answer to Puzzle No. 24 [1982: 2753: Nonplussed (no ilfplussed).
Answer to Puzzle No. 25 [1982: 275]: Iconic.
s'i *': s':

A Double dactyl
THE MOST DEDEKINDEST CUT OF ALL
Upper class, lower class, "How can I deal with their
Herr Richard Dedekind, Irrationality?
Plagued by real numbers, was Ho! By Eudoxus, I'll
Stuck in a rut. Start with a cut!"
LEROY F. MEYERS
The Ohio State University
 - 329 -
INDEX TO VOLUPE 8 , 1982

AUTHORS OF ARTICLES AND NOTES

AHLBURG, HAYO. Sweet are the Uses of Mathematics, 172.

. Passing Cars, 232.
. The House Numbers of Europe, 304.
ANDREE, RICHARD V. The Puzzle Corner, 35.
. An Interesting Recursive Function, 69.
. Postscript To "An Interesting Recursive Function", 96.
BANKOFF, LEON. An Heronian Oddity, 206.
BOTTEMA, 0. On the Area of a Triangle in Barycentric Coordinates, 228.
. A New Inequality for the Angles of a Triangle, 296.
and GROENMAN, J.T. A Note on Wallace's Theorem, 126.
COOPER, CURTIS. A Knockout Tournament Problem, 94.
DALEZMAN, MICHAEL. The Gaussian "30", 2.
DODGE CLAYTON W. Fractions on a Calculator, 36.
GOLOMB, SOLOMON W. The Numerology of Periodic Decimal Expansions, 32.
GROENMAN, J.T. and BOTTEMA, 0. A Note on Wallace's Theorem, 126.
HAVERMANN, HANS. The Puzzle Corner, 76, 158, 260, 328.
JOHNSON, ALLAN WM., JR. Magic Squares in Minimal Primes, 97.
_______ A Bordered Prime Magic Square, 172.
KENNEDY, ROBERT E. Digital Sums, Niven Numbersf and Natural Density, 129.
KILLGROVE, R.B. Field Planes, 196.
KLAMKIN, M.S. The Olympiad Corner: 31 to 40, 12, 39, 70, 99, 133, 164, 223, 237,
269, 300.
. Mathematical Swifties, 244, 275.
______ and TSINTSIFAS, GEORGE. Some Majorization Inequalities, 298.
KUMARI, KAMALA. A Magic Circle for 1982, 38.
MEYERS, LEROY F. Notes On Notation: V, VI, 10, 207.
. Ducci's Four-Number Problem: A Short Bibliography, 262.
. The Most Dedekindest Cut of All, 328.
MURTY, VEDULA N. A New Inequality for R, r, and s, 62.
ORR, EDITH. Mathematical Swifties, 275.
TRIGG, CHARLES W. A 1982 Potpourri, 6.
_______ Some Notable Palindromic Triangular Numerals, 267.
TSINTSIFAS, GEORGE. Another Proof of the A.M.-G.M. Inequality, 162.
______ and KLAMKIN, M.S. Some Majorization Inequalities, 298.
VAN DE CRAATS, JAN. On Six Concyclic Points, 160.
WAYNE, ALAN. The Puzzle Corner, 5, 76, 96, 163, 208, 275.
. Mathematical Clerihews, 128, 194, 231, 299.
TITLES OF ARTICLES AND NOTES

ANOTHER PROOF OF THE A.M.-G.M. INEQUALITY. George Tsintsifas, 162.

AREA OF A TRIANGLE IN BARYCENTRIC COORDINATES, ON THE. 0. Bottema, 228.
BORDERED PRIME MAGIC SQUARE, A. Allan Wm. Johnson Jr., 172.
CARLETON UNVERSITY PUBLICATIONS. 260.
CAVIARE TO THE GENERAL. 60.
CROOKS WHAT?. 11.
DIGITAL SUMS, NIVEN NUMBERS, AND NATURAL DENSITY. Robert E. Kennedy, 129.
DUCCI'S FOUR-NUMBER PROBLEM: A SHORT BIBLIOGRAPHY, Leroy F, Meyers, 262.
 - 330 -

FIELD PLANES. R.B. Killgrove, 196.

FRACTIONS ON A CALCULATOR. Clayton W. Dodge, 36.
GAUSSIAN "30", THE. Michael Dalezman, 2.
HERONIAN ODDITY, AN. Leon Bankoff, 206.
HOMAGE TO HERMAN NYON. 268.
HOUSE NUMBERS OF EUROPE, THE. Hayo Ahlburg, 304.
INTERESTING RECURSIVE FUNCTION, AN. Richard V. Andree, 69.
INTERNATIONAL CAMPAIGN-MASSERA. 34.
KNOCKOUT TOURNAMENT PROBLEM, A. Curtis Cooper, 94.
MAGIC CIRCLE FOR 1982, A. Kamala Kumari, 38.
MAGIC SQUARES IN MINIMAL PRIMES. Allan Wm. Johnson Jr. 97.
MATHEMATICAL CLERIHEWS. Alan Wayne, 128, 194, 231, 299.
MATHEMATICAL SWIFTIES. 244, 275.
MOST DEDEKINDEST CUT OF ALL, THE. Leroy F. Meyers, 328.
NEW INEQUALITY FOR R, r, AND s, A. Vedula N. Murty, 62.
NEW INEQUALITY FOR THE ANGLES OF A TRIANGLE, A. 0. Bottema, 296
1982 POTPOURRI, A. Charles W. Trigg, 6.
NOTE ON WALLACE'S THEOREM, A. O. Bottema and J.T. Groenman, 126
NOTES ON NOTATION: V, VI. Leroy F. Meyers, 10, 207.
NOT EUCLID ALONE HAS LOOKED ON BEAUTY BARE. 124.
Golomb, 32.
NUMEROLOGY OF PERIODIC DECIMAL EXPANSIONS, THE. Solomon W.
70, 99, 133, 164, 223,
OLYMPIAD CORNER, THE: 31 to 40. Murray S. Klamkin, 12, 39,
237, 269, 300.
ONE TWO, BUTTON YOUR SHOE. 327.
PASSING CARS. Hayo Ahlburg, 232.
POSTSCRIPT TO "AN INTERESTING RECURSIVE FUNCTION". Richard V. Andree, 96.
POSTSCRIPT TO "ON SIX CONCYCLIC POINTS". 244.
PUZZLE CORNER THE. 5, 35, 76, 96, 98, 158, 163, 208, 210, 260, 275, 328.
SIX CONCYCLIC POINTS, ON. Jan van de Craats, 160.
SOME MAJORIZATION INEQUALITIES. M.S. Klamkin and George Tsintsifas, 298
SOME NOTABLE PALINDROMIC TRIANGULAR NUMERALS. Charles W. Trigg, 267.
SWEET ARE THE USES OF MATHEMATICS. Hayo Ahlburg, 172.

PROBLEMS AND SOLUTIONS

The numbers refer to the pages in which the corresponding name appears with a
problem proposal, a solution, or a comment.
AEPPLI, ALFRED: 174. COXETER, H.S.M.: 276.
AHLBURG, HAYO: 18, 28, 78, 145, 182, 279. CRAY-1, S/N 22^COMPUTER: 152
AITKEN, DONALD: 47. CSIRMAZ, LASZLO: 174.
ALDINS, JANIS: 120. DELLINGER, JOE: 282.
BAETHGE, SAM: 323. DODGE, CLAYTON W.: 119, 121, 284, 311.
BANKOFF, LEON: 16, 215, 216, 285. DOU, JORDI: 19, 51, 142, 144, 156, 175,
BARBEAU, E.J.: 281, 302. 176, 211, 253, 289, 294.
BEESACK, PAUL R. : 54, 77, 184. ECKER, MICHAEL W.: 107, 177, 248, 303.
BHATTACHARYA, JAYANTA: 143. EDDY, ROLAND H.: 115, 155, 278, 314.
BLUNDON, W.J.: 82, 117, 122, 188, 284, ELKIES, NOAM D.: 49.
308, 318. ERD8S, PAUL: 49.
BOKLAN, KENT: 209. EVES, HOWARD: 256.
BOTTEMA, O.: 57, 246, 251, 253, 257, 277. FEDER, MEIR: 278.
BUISSANT DES AMORIE, E.C.: 218, 221. FISHER, BENJI: 120.
CHAN, S.C.: 107, 115, 117, 144, 154, 190. FISHER, J. CHRIS: 148, 175, 277, 305.
CHARLES, HIPPOLYTE: 27, 113, 310, 319. GARFUNKEL, JACK: 27, 55, 91, 106,
COOPER, CURTIS: 312.
115, 136, 138, 154, 174, 180, 190,
COVILL, RANDALL J.: 193.
210, 221, 280, 303, 307, 314.
 - 331 -
GIRI, G.C. 53, 78, 107, 217, 318. SATYANARAYANA, KESIRAJU: 56, 78, 85, 210,
GROENMAN, J.T. 15, 48, 56, 78, 89, 217, 248, 285,
91, 106, 116, 137, 293, 314.
153, 173, 209, 211, SAUVE, LEO: 245, 279.
212, 222, 246, 292, SERVRANCKX, ROGER: 305.
303, 314. SHUM, K.P.: 210.
HAMMER, F. DAVID: 113. SMEENK, D.J.: 48.
HENDERSON, G.P.: 48, 57, 108. SOKOLOWSKY, DAN: 78, 139, 248, 312.
HUNTER, J.A.H. 16, 82, 106, 175, STUMP, ROBERT A.: 85, 306.
246, 285, 286. TAN, KAIDY: 193, 218, 289.
HUTCHINSON, JOAN P.: 20. TIWARI, RAM REKHA: 137.
IZARD, ROGER: 136. TRANQUILLE, ROBERT: 86,
JINGZHONG, ZHANG: 174. TRIGG, CHARLES W.: 15, 29, 82, 107, 117,
JOHNSON, ALLAN WM., JR.: 22, 135, 150, 135, 152, 186, 213,
186, 209, 219, 217, 245, 247, 249,
245, 252, 304. 286, 294, 304, 320.
KATZ, DICK: 78. TSINTSIFAS, GEORGE: 14, 17, 24, 49, 77,
KESTELMAN, H.: 14, 78, 106, 137, 173, 83, 107, 108, 110,
210, 246, 278, 304, 325. 135, 138, 148, 153,
KIERSTEAD, FRIEND H., JR.: 85, 89, 282, 174, 190, 209, 220,
KLAMKIN, M.S.: 58, 120, 145, 156, 181, 246, 256, 287, 291,
189, 191, 194, 220, 279, 316, 323.
280, 308, 311. TURNER-SMITH, R.F.: 210.
KLINE, JAMES S.: 120. UMBUGIO, EUCLIDE PARACELSO BOMBASTO: 20.
KOMORNICKI, WOJCIECH: 49. VATHIS, DIMITRIS: 175.
KONHAUSER, JOE: 18, 121. WAGON, STANLEY: 20, 120.
KRAVITZ, SIDNEY: 119. WAYNE, ALAN: 14, 29, 59, 77, 173, 217,
LARSON, L O R E N C : 210. 276, 302, 323.
LEEB, BERNHARD: 48. WHEELER, FERRELL: 142, 282, 314.
LINDSTROM, PETER A.: 252. WILKE, KENNETH M.: 17, 22, 89, 122, 283.
LIU, ANDY: 15, 113, 249, 325. WILLIAMS, KENNETH S.: 18.
LONNBERG, ANDERS: 138.
LU, YANG: 174.
LYNCH, J. WALTER: 278.
LYNESS, R.C.: 277, 287.
MAENNEL, HARTMUT: 48.
MASKELL, F.G.B.: 15, 16, 28, 51, 121.
MAVLO, DMITRY P 109, 157.
McCALLUM, J.A.: 47, 145, 311.
MEYERS, LEROY F 25, 53, 59, 112, 119,
178, 279, 319, 326.
MILLER, FRED A. : 86, 285.
MOSER, WILLIAM: 259.
MURTY, V.N.: 15, 48, 254, 303, 311, 316.
NGO, TAN: 27, 212.
NYON, HERMAN: 88.
ORR, EDITH: 188.
PEDOE, DAN: 79, 141, 174.
PRIELIPP, BOB: 118, 191, 213, 245, 282.
RABINOWITZ, STANLEY: 14, 49, 78, 107,
136, 143, 152, 157,
174, 177, 210, 219,
222, 247, 255, 259,
277, 284, 304, 306,
307.
RUDERMAN, HARRY D.: 48, 257.
SALVATORE, GALI: 16, 24, 25, 111, 150, 182,
310.