Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
THE
OFFICIAL JOURNAL
OF
VOLUME 6
NUMBER. 6
CONTENTS
PART I Pseudo-Fibonacci Numbers
ADVANCED
.
H. H. Ferns
305
. H. W. Gould
318
322
B a s e s for I n t e r v a l s of R e a l N u m b e r s
335
350
PART H -
ELEMENTARY
T h r e e Diophantine Equations P a r t I .
. Irving Adler
360
H a r m o n i c Design in Minoan A r c h i t e c t u r e .
. Donald A. Preziosi
370
. Joseph S. Madachy
385
390
393
Recreational Mathematics .
L i n e a r R e c u r s i o n Relations 9 L e s s o n Two
E l e m e n t a r y P r o b l e m s and Solutions*
Volume Index . . . . . . .
. . . . , . 408
DECEMBER
. . . . .
. .
Dov Jarden
407
1968
DEVOTED TO THE
WITH SPECIAL
PROPERTIES
EDITORIAL
H. L. A l d e r
M a r j o r i e Bicknell
John L. Brown, J r .
B r o t h e r A. B r o u s s e a u
L. C a r l i t z
H. W. E v e s
H. W. Gould
A. P . Hillman
V. E. Hoggatt, J r .
ASSOCIATION
BOARD
Donald E. Knuth
George Ledin, J r .
D8 A. Lind
C. T. Long
Leo M o s e r
I. D. Ruggles
M. N. S. Swamy
D. E. Thoro
C h a r l e s H. King
L. H. Lange
J a m e s Maxwell
S i s t e r M. DeSales McNabb
C. D. Olds
D W. Robinson
A z r i e l Rosenfeld
John E , Vinson
Lloyd Walker
C h a r l e s R. Wall
PSEUDO-FIBONACCI NUMBERS
H. H. FERNS
Victoria, B.C., Canada
o.+l =
P
(2)
= 0
i+i
0 l
* pt
i+i+XOi
0^=^ = 1
O i + 1 = 2 0 . + i + XO.
<4)
i+2
2P
i+1
O 0 = 0,
*Pi
Ot = 1
P 0 = 1, Pi = 1
Thus the two numbers defined by (1) and (2) satisfy the same recursion
formula but with different initial values.
A Binet-type formula for each of O
and P
(5)
= (1
^JTTxf
(6)
- (1 2 N/1 + X
N/TTX)11
= (1 + ^ / ^ T X ) n + (i _ N/TTX) n
n
2
(Received 1 9 6 3 - - r e v l s e d Feb.
305
1968)
From
306
PSEUDO-FIBONACCI NUMBERS
[Dec.
A + B = 2
A2 = 2A +A
A - B = 2 N/1 + A
B2 = 2B + A
A2 + B 2 = 2(2 + A)
A2 - B2 = 4 N / T T T
AB = -A
Hence
n
u
An - B n
A - B
_ An + B n
n
2
true for i = 2, 3, , k .
O
u
k
and P
that
Assume that it is
Then
= Ak - B k
*k
A - B
+ P
A k -f B k
2
2 A k - - 2 B k + A k + 1 - B k + I - A 1 ^ + AB k
2(A - B)
A k (2 - B) - B k (2 - A) + A k + 1 - B k + 1
2(A + B)
A k + 1 - B k + 1 + Ak+i
2(A - B)
AkHhl - B ~ k + 1
A- B
- Bk+1
ft
U
k+i
This completes the proof of (5). A similar proof holds for (6).
If we let n = -k where k is a positive integer we find from (5) and (6)
that
P
k
O k. = - (~A)k
and
K
k
P " ,k = (-A)
1968]
PSEUDO-FIBONACqi NUMBERS
307
On = 2n"*Fn
n
n
P n *= 3n~*LM
n
n
and
(A = 4)
- i
E:pi
<*
- P * H : - I
i=i
<)
X^Li
*' ^X+i
1}
i=i
308
[Dec.
PSEUDO-FIBONACCI NUMBERS
n
(9 )
1=1
A formula for the sum of the first n O's with even numbered subscripts
is now derived.
n
E* - S
i=i
A 2i - B 2i
A - B
i=l
n
A - B
A2i
E -L' B2i l
i=l
(10)
i=i
L _ |"A2(A2n - 1) _ B 2 (B 2n - 1)
i*]
1
f(A 2 B 2 - A 2 )(A 2n - 1) - (A2B* - B 2 )(B 2n - 1)"
A - B ["
A2B2 - A2 - B 2 + 1
H
1
irr
"S L
A 2 0 2n - 02n+2 + 2
(X + 1)(A - 3)
A2 - 2P2 + 1
J
(A^ 3)
Applying recursion formulas (1) and (3) to (10) takes the form
<
10a)
J^
(A + D(A - 4)0 2n _i + fc2 - A - 4)P 2n _ 1 + 2
I-02i -
( A + i H A + 3>
i=l
,. ,
3)
^
<X
1968]
(10')
PSEUDO-FIBONACCI NUMBERS
309
= | <22nL2n-i + 1)
2>*F2i
i=i
O, - j f s 1 -(-!)']
Hence
"n
(10s)
1
4
Z2i = J E
i=i
9
32
_i=i
si
-I>>
i=l
J
(9n - 1) - f .
Their derivation
(ID
X)P2i
P 2n +2 - 2 0 2 n +2 + 2 - A
A~^"3
<Xt 3)
i=l
(Ha)
A
(A - 4 ) 0 ^ + ! +A(A - 2)0 2 n -i - A + 2
2 ^ ?2i =
"
r
A
- *3
1=1
(Us)
I> i
2
= $
(9Q - 1)
i=i
(11')
2 2 i L 2 i = 4(2 2 n F 2 n - 1)
i=l
(A= 3)
& ^ 3)
310
PSEUDO-FIBONACCI NUMBERS
[Dec.
To find the sum of the first n O's with odd numbered subscripts we use
ii
2n
E2i-i
i==i
"Z^1
i
1=1
i=i
and make use of results already obtained. In this manner we get the following
four identities:
A
< 12 >
fo " D0 2 n +i - 2 O m + 1 - A
L<*-i -
(12a)
& "
(X + i H A - *
4)P
2,02i_!
2n-i
+ X ( X
'
" 2)P2il-2 + 1 - A
(A + lHX - 3)
(A
3)
*3)
i=l
(12s)
Eo2M ' - ^ ^ - D + j
(X=53)
i=i
(120
2*i-F2i_1 - | k ^ L ^ - 3]
2* p tf-
(13)
O m +2 - 3<>2n+i + 1
r~r-5"
<M.a>'
i=l
"
(13a)
^ ^ i "
i=i
(A - 4 ) 0 2 0 . , + X ( A - 2)02n. 2 + 1
1
*" ~~'
A~*^l
"
( A ? l 3 )
1968]
PSEUDO-FIBONACCI NUMBERS
311
S p 2 i - i = Tg (9n -' 1) - J
(13s)
S 2 2 i ~ l L 2l-l
(13')
(A = 3)
^ ^ F ^ + 1)
i=i
We now derive the sum of a series with alternating positive and negative
signs.
From (10) and (12f) we get
=t>2Zi-iF^-E2^i
(-D^VF.
i=i
i=i
i=i
22n+i
= _ _ (I2n-2 -
2Ij
2n-l) -
Hence
2n+i
J(4)
i+
n+i
VF.=2%2M.j;2%2i
i=i
i=i
i=i
22n+2
= _ . (2Lm
- L 2 n -i) - 2
d4)
Y* ( - 1}
V F
lz
r-
(2L
n - W
-2
312
[Dec.
PSEUDO-FIBONACCI NUMBERS
V ( - l ) i + V L . = (-2) n+1 (F
(15)
L~t
n-2
- 2F
n-i
) + 6
i=i
Derivation is given
i r r r [(1 - ("X)I1)]
For brevity we shall denote this by R.
16
( X / 3)
+R
X- 3
i=l
9 n + 1 - 8n + 4(-3) n + 1 + 3
(16s)
( X = 3)
']
f>%> = | [i-rtPa.t + ( - 4 ) 7 - 6]
(16')
i=l
n
E*i-f
(17)
( X - 4)0 2 n +
( X - 2)0 2 n -i
X- 3
- R
(Xj 3)
i=i
(17s)
i=i
(17')
S22iLi
i=l
= 2 2 n + 1 P . i - (-4)
a
n+1
+ 14~
( X = 3)
1968]
PSEUDO-FIBONACCI NUMBERS
i=i
J ] (40*.+ P 2 ) = - ^ [ 9 n + 1 + 8n - 9]
(18s)
i=i
2 ] 2 2 i ( 5 F 2 + Lj) = 8[2 2 n F 2 n _i - 1]
(180
i=i
Ei-
r 1
A - B1
A - B
-i
i=i
1
4(1 + A)
"n
A2i +
_i=i
B2i
- S(AB)i
i=i
A2(A2n_ 1}
1
4(1 +A) I
A2 - 1
i=i
( B2(B2h
._
1}
ABT(AB)n - 1]
B2 - 1
AB - 1
- l
1 + A + 2N/1 + A
- 1 =
1 + A - 2N/1 + A
and
B2
we have
[Dec.
PSEUDO-FIBONACCI NUMBERS
314
Zl =4(14n4+ X)
A2B2-A2-B2 + 1
i=i
1
2(1 +
1
2(1 +
( X * 3)
2R
PJQ-4 - 20 2 n + 2 + 2 - X
+ R
X - 3
x)|
( X / 3)
Ei-
1
2(1 + X)
+ R
X- 3
(X t 3)
i=l
(19)
follows:
AE
2(1 + A)0
O
v
n m
-,11
2(1 + A) A - B
2N/T+T
2 NSTTA
An-nn + B n+m ;_ A n B m
.rH-m _, T-.n+m
A
+ B
:p
Am-Bm
_
AmBn
.m-jm/.n-m
..n-m
A B (A
- B
^ -(-A)mPn w
n+m
n-m.
The proof
1968]
PSEUDO-FIBONACCI NUMBERS
15
i=i
i=l
i=i
2d + A>xVo i+ i = X X + i - L ( - A ) i p *
Using (13a) and the fact that Pj = 1 we have
(20)
\T^
'
2(l + A ) 2 ^ 0 1 Q i + 1
(A - 4 ) 0 2 n + i + (A - 2 ) 0 ^ + 4 - A
" "
^
-+R
(x ^ 3)
i=i
(20S)
2ii+i
125 [32n"^
+ 4( 3)n+1
"
"
8n
~ 15]
i=i
For A = 4 we have
(2 f)
]? 2 l F i F i+i
I D8*** + I [1 " M ) n ] ]
i=i
The proofs of the three following identities are left to the reader,,
(A = 3)
316
(21
PSEUDO-FIBONACCI NUMBERS
2P
>
nP
n+
[Dec.
-^mPn
n m
n+m
n-m
20P
= 0 ,
+ (-X)mO
n m
n+m
n-m
m
2PO
= 0
' - (-A) O
n m
n+m
n-m
(22)
(23)
Following the same procedure that was used above we arrive at the following identities:
(24)
J^
(A - 4)0 2n +i + A(A - 2)O zn + 4 - A
2^P.P.+1 =
- - J
(A^ 3)
i=l
n
(24s)
X] P i P i+i
[32n
& = 3>
i=i
J. *
= 4 J W * - [1 - (-4)]]
^TVL.L.^
(24)
i=i
*
(25)
Z>i+li=
(X + D(X - 3)
^3>
i=i
(25s)
J]Pi+i0i
"A [32n+3 +
8n
{X = 3)
i=i
n
(25)
2^ 2 2 l L i+l F i
[22nL2n
l]
i=i
A,
(26)
22J0. + I L. =
i=i
(X - 4 ) P m + 1 + X(X - 2)P 2n - X2 + X + 4
(x + i)(x-3)
3)
1968
(26s)
PSEUDO-FIBONACCI NUMBERS
Z^i-4 L i
t 32n+3
317
8n + 24(-3)n - 5l]
(X =
3)
i=l
xx
(260
2 ^ 2 2 i F i + 1 L i = i [2 2n L 2n - 3 + ( -4) n ]
1=1
REFERENCE
1. cf.
Vtn)
hn
<1'0)'
n = 1,2,3,--
Paul F. Byrd [l] has shown how to determine the coefficients c, in the
expansion
oo
(i)
tw = i;c k 0 k + 1 (x),
i=o
f (x) = Y, v n >
n=o
and the polynomials <p (x) are defined by the recurrence
(2)
<j@(x) = 0, ^(x)
= 1,
(1 - 2xt - t )"" = ]T V i W i ? 1 *
(3)
n=o
It is our object to point out that the expansion theory involved is a special
case of a general treatment given by the author in [3],
(4)
(C - mxt + y t m ) p = ^ t n P n (m, x, y, p, C) .
n=o
318
"""
DGC
IQAQ
dia
rX
AND A SOLUTION OF PROBLEM P-3
The polynomials P n include the polynomial systems of Louville, Legendre,
* n+1 (x) = P n ( 2 , x , - 1 , - 1 , 1 ) .
It is shown in [3] that P (m,x, y,p,C)
(6)
satisfies
CnP
- m(n - 1 - p)xP
+ (n - m - mp)yP
=0,
n
n~* 1
nm
of which (2) is a corresponding special case*
n > m > 1
k=o
with the corresponding inversion
/o\
#p\ /
\&
[n/m]
y
v
\^
1 i\k ip - n + ki p + mk - n ^n-k-p k^
(8) Pl(-mx) = Z, W) f
x ;
) yTT^
yW
k=o
k=o
and
(10,
,)
k=o
MJ^i^-WW-
>
mw C)
'
'
320
[Dec.
This, incidentally, solves his problem P - 3 [ l , page 29] and in a simpler manner than the complicated induction solution given in [4].
Actually, relations
(7) and (8), and hence also (9) and (10), are special cases of the general inversion relations [2, (6.3), (64)] found by the writer:
(ID
FW=
[n/m] /
\
(e-nk+mkjf(n-mk)
k=o
if and only if
(12)
[n/mj
(-Dk\P'k+kj
f(n) =
\
P
p++kk-"nnF(n-mk)-
k=o
Proof of the reciprocal nature of (11) and (12) in turn depends upon general addition theorems for the binomial coefficients, typified by the relation
n
(13)
^qan
A^(a
+ Cih)
k=Q
where
A
/ ,v
A (a b)
k '
/ a + bkl
iT+bk U
'
known, these general addition theorems enter into something on the order of
several hundred papers in the literature.
1968]
321
We wish to note that relations (11) and (12) were used by the writer [2] to
establish certain results about quasi-orthogonal number sets.
The relations
1. INTRODUCTION
Let a f 0, /3 = 0, J a | > |j8j, be any two complex numbers, such that
a + p and a/3 are two relatively prime integers* Then the numbers
n
- j3
n-i ,
n-2^ ,
n o f - j S
, 0n-i
n , 0n
'
a r e integers, since they are expressed as rational integral symmetric functions of the roots a,p
of an algebraic equation
z 2 - - (a + jS)z + o0 = 0
with integral coefficients with leading coefficient unity., One may readily verify that {D } and {S } are
ar second-order recurring sequences satisfying the
L
*
nJ
nJ
common recursion relation
X = (a + j8)X
- a$X
.
n
n-i
n-2
(Since
Do = 0f
D | = 1; S0 = 29
S4 = a + 0 ,
a r e integers*)
that D 2 n = DnSn
Adivisor > 1 of D , n > 1, is said to be primitive (ors characteristic)
if it is relatively prime to any D. with 1 < i < n* The greatest primitive
divisor of D is denoted by
Df A divisor > 1 of S , n > 1, is said to be
J
n
n
n
primitive (or: characteristic) if it is relatively prime to any S. with 0 < i <
n.
D S
(1)
is denoted by S?
196?)
From D ^ =
ueCm
23
For any prime p dividing a certain D. with i > l s a (p) denotes the
smallestpositive subscript nf such that plD . Thus p is a primitive divisor
of D M 9
a(p)
By F we denote the product
(2)
F
n
= n *
d
dn
(3)
x
'
(4)
n
where <f> * s Euler 1 s totient function, and m{n) is the number of distinct prime
factors of n.
The main result achieved by Carmichael is the following
Theorem XXHL If a and p are real and n f- 1,2,6,
then D n con-
tains at least one characteristic factor, except when n = 12, a +'j8 = 1, ofi
= -1.
In the present paper the above Carmichael 1 s results are generalized for
any two complex numbers a f 05 p fi 0$ J a\ > | p\,
a r e two relatively prime integers*
with which any D
o9,
for F
Finally,
the method of proof used here is slightly simpler than the one used by
Carmichael,
By (2)
f C3> or (a - 0) =
and of com-
and op ^ ~ Q
djn
d|n
d|n
d|n
Noting that
2 M(S) d = *<n) ,
d|n
and
E*(S1 -
d|n
for any n > 1, we get
<5)
djn
1968]
325
Let us evaluate
d|n
Note that for any 0 < q < 1 there exists a positive constant A, such that, for
any complex z, for which | z| =^q, there is
| l o g ( l + z)| < A|z|- f
where by log (1 + z) the principal value of log is understood.
log (1 + z)
z
z
" 2
Indeed,
z
3
2 WIN MS)'
"^
-M
!-(!)'
d|n
d|n
^ q. Hence
Allil'Hiirfn-^t
d=i
'Mai
(7)
>-()'
d=i
Ifl
B,
q?fi
4
rn
Luec*
= lots') 2*1 ,
hence
\a\ > 1,
or
>
(9)
C n
log D ; > (log or| *(D) - log p - B > log| a\ ^ g n "
log n
primitive factor*
Remark.
prime, then
I(SW{-()] --*{i-}**.{i-()};V-i>-!^
1968]
327
Theorem 3.
J2w
t^ n
converges,
Proof, From (9) It follows that there is a positive constant D such that,
for all n ^ 1,
C*n
s iloglogn
B
e *n
Hence
n=i
n=i
36 MAIN RESULTS
Lemma 1* Be N the sequence of natural numbers, S a subsequence of
N, and A a reduced arithmetic progression* Then, an infinitude of D* is
composite for
I)
nE S
or
II) n E N ~ S
any
or
II)
according as
1)
no
o28
*- eC*
= 00
pEA
nGN
whence
nGS
pGA
P>q
Then, noting that any prime p ][ 2(a - p)2ap is a factor of a certain Df ([1], p.
45, Theorem XII), any prime member of A not a factor of 2(a-f})2a($ ia a factor of a certain D f , n E N - S, and n) follows as above.
Theorem 4. There is an infinitude of composite D*.
Proof. The theorem is an immediate consequence of Lemma 1,
2
noting
where
r ^ 1 and Pi, , p
are distinct
primes, then there exists an integer u = 1 (mod 4), where (u, 4pi,-- , p r ) =
1, such that, for any prime p = u (mod 4p l9 * , p r ) , it is
Q>
Proof. [ 2 ] , p. 75.
1968]
329
z,
are
distinct primes, then there exists an integer u = 1 (mod 4), where (u,4pj,
> Pr)
sucn
&&U ^ o r
an
y prime p = u (mod 4p lf , p r ) s it is p f x2
(D , . D ) = 1 ( f l l , p. 38, Corollary).
n+i
L J
Hence, putting
in Lemma 4:
b = -o-jS,
r
>
x = D . ,
n+i
= D
n
(a - P)2 f D
or
(a - /3)2 = D
n)
and a? JJ
(9)
I) Suppose that
where
u = 1
(mod 4) ,
(11)
(u, 4 p 1 , * - , p r )
1 ,
(12)
implies
~{a-$)2
,=
_x
x = u (mod 4p l9 , p r )
!
(mod
4QJS)
1968]
SECOND-ORDER RECURRING.SEQUENCES
331
= (a - 0 ) 2 ,
flj8)
it follows
1 =
fe^,plf...,pr)
^ (o0, m 2 P i , - - , p r ) = (oj8,(or-|8.)2)
1 ..
Hence
(4p l9 -,p r9 4^iS) = 4(p l9 ee ',p r ,<*j3) = 4 .
But, by (10), 4|u - 1, hence (13) has a solution u ! , i. e . ,
(14)
Let p be a prime satisfying p = 1 (mod 4c*j8). If aft is odd, then, according to the properties of the Jacobi symbol
(?)-(#)-&)-'
If af$ is even, then p = 1 (mod 8), and aft = 2 t, where k ^ l
Then
(f)-(if(tWf)-(*)-'
in both cases
and 2 [ t.
66Z>
(15)
(16)
<u', 4 P l , . . . , P r ) = 1 ,
If p = uf (mod 4p 4 , , p r ) ,
then
( -fr-P> 2 \ = /-<*(<*-ff)2\
_x
(17)
(mod 4p 4 , , p r ) ,
y = 2,
a = (a -j3) 2 ,
b =
we have
/-(qp)n(g-p)\
If n is even, then
If n is odd, then
/-Wn(g-j8)8\
/-(<*-ff)2\
(apf,
1968]
33a
(18)
p|D n
1 ,
for any odd prime p. Consider now the three following cases.
Case 1; ap = n2 2 ,
where k ^ 0.
m - G)(!f - -
and hence, by (19), p j[ DJn .
Case 2: a/3 = n2 2 q ls ,q r where k ^ 0, r ^ 1, q4, - , q r a r e
distinct odd primes, and t = qj, e , q^ = 1 (mod 4).
Consider the pair of congruences
( x = - 1 (mod 8)
(20)
I
x =
1 (mod t)
(u, 8t) = 1 .
334
Dec. 1968
(22)
(=?=) mm -- - ->
-1 .
x =
c (mod qi)
x =
1 (mod q2>
x =
1 (mod q r )
If r ^ 2, or the system
(x = - 1 (mod 8)
(24)
(x =
c (mod qj)
if r = 1. Since q1? , q r are distinct odd primes, (23) and (24) have a solution v. v satisfies:
(25)
(v, 8t) = 1 ,
(26)
W-
,kn / A \n
(-1X+1)
-'f
= 1
[Cont on p. 406.]
a n d The U n i v e r s i t y , R e a d i n g
i . INTRODUCTION.
In this paper we discuss the problem of representing uniquely each real
number in the interval (0, c ] , where c is any positive real number, as an
infinite series of terms selected from a sequence (b ) of real numbers.
We
choose an integer k ^ 1 and require that any two terms of (b ) whose suffices differ by less than k shall not both be used in the representation of any
given real number. The precise definitions arid results a r e given in the next
section*
In an earlier paper [2] we discussed an analogous problem of representing the integers in arbitrary infinite intervals*
2* STATEMENT OF RESULTS
Throughout this paper k ^ 1 is an integer.
initial term of any sequence is the number 1; e. g*,
Let (c ) be a
sequence of positive real numbers which obey the linear recurrence relation
(2 1)
a^n+k
+ a c
2 n+k-i
+ 9# + a
kcn+i~cn = 0
335
336
[Dec.
smaller than the modulus of any other root of g(z), then p < 1 and c
= Ap
r = b. + b. +
ii
12
where
for
r H *
+ k
real root 0, that 0 is a simple root, and that 0 < 1 . Let R be a real number. We now enunciate our main result.
Theorem 2. Let (b ) be a sequence of real numbers such
that b ^
R
n
b . > 0 for n ^ 1. Then (b ) is a k-series base for (0,0 J1 if and only
n+i
n
if
= eR+n
n
for n ^ 1.
It is not true that all k-series bases are decreasing.
when k = 2,
For instance,
(0,2+0].
0 for n N then b
stant.
>
= A#
for n
>
decreasing.
It follows from Theorem 2 that the sequence
1968]
(2.3)
, e- \.-.,e-\
337
e\ e,...)
^ r and 6"
0-representation
For shortness,
of r .
Then, if an
6-representation1
and the
binary representation of the same real number a r e the same (for when k = 1,
e = J).
A further observation is that any sum T of a finite number of terms of
*0_j.j
the sequence (6
D i n
,8
9**)9
eh
where i
+1
iv
+ eh + . . . + 0*0-1. + Av-i
k for 1 ^ v < a,
where i^ + 1 ^ ip + k for
(2.4)
fl^-1
v ^ 1, simply by putting
= ^Ji<^vk
00
yv
338
[Dee*
and
for n ^ 0? both of which are very easily proved*) This fact is analogous to
e
J 1
] 6
< e for
1 ^ s ^ p9
parts, where
is an integer.
m + 1 points
(NM3?IS
One part of M must contain two of these points; let the corresponding indices
be vi and ^2* Then clearly
1968]
339
for 1 ^ s ^ p, and
I n - i^l > i .
We put
n = N\vt -
P2|
can be expressed
in the form
u
* -x: 2>vk
(3.1)
c = > I > n X s J n
s=i \ t=o
for n ^ 1 ,
where the numbers (a are the distinct roots of g(z), the number (v_ + 1) is
s
s
the multiplicity of the root f for 1 ^ s ^ u, and the numbers B , are suitable complex constants.
Case 1. B
Let
(3.2)
Since c i 5 p
>
for
Then by (3.1),
n ^1 .
a positive constant,
s*o
~ J
340
1 ^s ^p,
[Dec.
V
(ii) for p < s ^ u, if |fg| = | | then B gf . = 0 for q < t < v g ,
if |fg|?> jfij then B g t = 0 for 0 ^ t ^ v g .
and
Then by (3.1)
P
(3.3)
+ R
%=Jjsq%
s=i
y n
where R is the sum of a finite number of non-zero terms of the form Cn f g ,
where C is a complex constant and either If J
Ifl j
<
(3.4)
jfi) => p
or
q > 0 .
(3.5)
For
1 < s ^ p,
^0
as
Hence
n^oo.
let f
= r exp (i a ), where r
and or are the
s
s
s
s
s
modulus and argument of f , respectively. Then by (3.5) and (3.4) respectively,
R / r ? n q ->0
(3.6)
as
n->oo
and either
(3.7)
r! > p
or
q > 0 .
for
n ^ 1 .
1968]
341
From (3.3)
sq
and (3.6)
SQ
(3.8)
+ 0(1)
as
->
w
It is clear from (3.8) that E
m =
m iHq * \
r
""i
m
\
/
p
rj m
Hence,
%
(3.9) c i . > c i ^ > 0
7
m-k+i
m-k+2
8 e
(Ci\
m - f c H N / c * \ m-k+2^ J C A m
>c
> I - I p
> I Ip
> >!J p
m
\p/
\P/
\P/
Therefore,
(3.10)
cm
fc
= a ^
m-k+i v
)
342
mk-t>
[Dec.
(?)
and
Dl-k-2
i^.m-k-2
'
Ci>
a contradiction.
(?)
PROOF OF THEOREM 2
] if and only
if
U)
is a k-series base for (0,1]. Hence without loss of generality we assume
that R = 0, so that we shall be discussing k-series bases for
(0,1].
Lemma 2.
00
for
nZ 0 .
k
Proof. Since 6 is a root of f (z) = z + s - 1 and 0 < 6 < 1, we see
that
1968]
343
-V)
for m ^ 0e Since 0 < 1 it follows that
00
1
^=0
and hence
00
~n
"\n+H-i>k
for
n ^ 0 ,
V=0
as required*
Proof of sufficiency.
Let 0 < x ^ 1* F i r s t we construct inductively a sequence (i ) of positive integers such that i , ^ i,, + k for v ^ 1, and
(4,1)
1+k
^ x -S^B
"
> 0 ,
l*=i
= 1 we see that
0 ii-i+k
6Ii-i
0m
344
[Dec.
Let t ^ 1 be an integer and suppose that i*, i 2 , , it are chosen so that (4.1)
holds for m = t,
and i ., ^ i + k for
Vwr\
1 ^v
< t.
Then we choose i, , 4
VT\
such that
eWl
(4.2)
.X-VA>>
l^t
Hence
t+i
9Vf *
-3?
From (4.2) and the assumption that (4.1) holds for m = t it follows that
6*
Hence
i, + ^ i, + k.
> 0 t+1
(i )
follows by
induction.
Since 6 < 1 it follows from (4.1) that there exists a representation of
x in the form
00
oo
(4.3)
where ii 1 and i
^ i " + k for v ^ 1.
This representation of x is unique. For otherwise we may assume
without loss of generality that
, 00
00
2> v -L* v
v=i
v=l
1968]
oo
oo
j t ^ 1 and j
345
- 5^ + k for
by Lemma 2. Hence ii
>
V+* = i
and so 1 is the largest number which has a representation in the form (4.3).
Hence (6n) is a k-series base for (0,1]. This completes the proof of the
sufficiency.
Proof of necessity. We show that if the sequence (b ) is a k-series
and if b n + 1 s> b n > 0 for n s> 1, then b n = 6n for n s> l.
creasing, for if b. = b. for i f- j then clearly some numbers have more than
one k-series representation*
r|r
= Y \
I
1
We denote by B
For n ^ 1 we define
JmmwJ
P=l
; ii = n,
^i+k
Since
oo
= X>"k fr
.
I
>
(4,4>
for v * 1
QS1
(b ) is a positive
346
[Dec*
and
(4.5)
B ^
n
B ^ ^
n+i
0 for
n ^ 1.
such
/m
oo
so that ^ = 0. Hence
(4.6)
B >0
n
as
n-5> oo
B ,,
n+i
= b
n
>
Bm+i
,.,) Suppose that bm < Bm+i
,.. Then we can construct inductively
J a
+
sequence (j^) of positive integers, where ji = m and j + j
k for
v ^ l,
satisfying
B _ + ! e B.
m+i
n
j
if i/ = l ,
or
(4.8)
B ^ - b . - b. - - b.
+ - G B.
Bi+i
Ji
32
J - i- n
J
By (4.5) and the induction hypothesis,
if v ^ 2 .
1968]
xa-i
347
m+i
+I e B .
m+i
n
m
Let S1^ 1 be an integer and suppose that the first 5 - 1
B
terms of
(jp)
if 5 = 2
6-i
B ^ - b . - b . ^ - b.
+~EB.
, if 6 ^ 2 .
m+i
Ji
J2
J62 n
Jfl-1
Hence
00
B ,, - b. - b. - - b .
m+i
+e
ji
Ji
B.
Therefore
B j . - b. - b . - - b .
m+i
l J
i=jJ J. +k
6-i
J2
^ 0
J64
+6 1 /
n
vy
B.
i
348
[Dec.
__, - b. - b. - - b .
B.
00
L-d j 7 .
n+i
00
^LA+i+^k
n+i + J^ b n+kH-i+i'k
Z^=0
b]tl+1 + B
n+k+l
V=Q
= b , + b ,,
n+i
n+k
for
n ^ 0 .
The number 0 is the positive real root of the auxiliary polynomial f (z)
k
= z + z - 1 of this recurrence relation. The modulus of any other root of
f(z) is greater than 0. For if | z |
<:
0 , then since 6 ^ 1 ,
1968]
349
then
| f ( z ) - z + l| = l - | z | - | z | k = 0 ,
1 -|z| - | l - z | = l - | z | so that
|l - z |
1 -|z| ,
= A0 n for
(0,1],
and so it follows that A = 1. This completes the proof of the necessity and of
Theorem 2.
REFERENCES
1. D. E Daykin, "Representation of Natural Numbers as Sums of Generalized
Fibonacci Numbers, f? Journal London Math. Soc. , 35 (I960), pp. 143-160.
2o D. E. Daykin and A, J. W* Hilton,
The Editor acknowledges with gratitude the receipt of $350 from the
Academic Vice President of San Jose State College toward our publication
expenses.
*
JU
.4.
-A-
-W
-A.
Solutions
To facilitate their
L2
+1
zL* Hl2j+i)k
k=o
n
L
Gi)
2j+1 2 J H(2j+i)k
<2j+i)(Q+i) - H_(2j+i) ,
k=
n
(iii)
2J ( - D ^ I j k = (-l)nH2j(n+i)H2jn - H0H.2j
k=o
iv)
<
E <-*=*
k=o
350
Dec, 1968
351
Put
00
xm n
m$n y
msn=o
Show that
n
L^J
B.
nsn
n,n
1 - abx2
(1 - x)(l - ax)(l - bx)(l - abx)
Put
(1 - x H ( l - y)-i(l - axy)^X=
V n ^ ^
m?n=o
Show that
Y"B
x11 = (1 - x T ^ l - ax)"" A.
n=0
ei e<r
e-r
n = Pi1P24*-* P r
is the n
Fibonacci number*
352
[Dec.
Let P
= 2P
be the n
= 169.
H-147 Proposed by George Ledin, J r . , University of San Francisco, San Francisco,
California.
is the k
Xi =
n+i
n->oo
*F n
F
lim
x2
nm
m
n
F
x 3 = nlim
>o
nm
"""n
F
XA
|
nm
limA
x,
limA
n->o
iL - 2 1
n
n
SOLUTIONS
SUM DAY
H-103 Proposed by David Z e i t l i n , Minneapolis, Minnesota.
Show that
m = 1,
1968]
353
n
8
2^
k=o
2x+i =
+ Fx+1
and that
F F _,_ = (-1) X + 1 + F*
X X+2
X+l
Then
2x-3 =
x-2 +
x - i = ( F x - F x _ 2 ) 2 + (Fx_i - F x _ 2 ) 2
= F2 + F2
X
=
F2
+ F2
X-l
+
F2
X-i
= F2 - F2
X
+ F2
X-2
F2
X-2
- F2
X-i
- 2F
X-3
- 2F
X-2
X-3
X-3
F
X-i
_ 2((-l)X"1 + F2
F2
+ F2
X-2
X-3
+ (-1)X""2 + F 2
X-i
= F2 - F
X
)
X-2
+ F2
2X-3
X-3
whence
2F
2X-3 -
F x + Fx__3 .
3n+6 -
3n+3
The r i g h t - h a n d i n c r e a s e i s
=
3n+3M F 3n+6
The f i r s t f a c t o r i s
F 3 n + 5 + F 3 n +4 - F 3 n + 3 = 2 F 3 n + 4
3n+ 3 )
354
[Dec.
The second is
^sn+g + F 3n + 4 + F 3n + 3 = 2F3n+5
Thus the total is
4F3n+4 Fsn+5 (Fsn+e + F 3n +3) .
The left-hand side increase is
8F
3n+4F3n+5FQl+9 *
3n+6
3n+3 5
GENERATOR TROUBLE
H - 1 0 4 Proposed by Verner E. H o g g a t t , J r , San Jose State C o l l e g e , San Jose,
California.
Show
L
x
S
1 - 5F x + (-l) m + 1 5x 2
m
where L
~
k=o
Lucas and Fibonacci numbers, respectively.
Solution by D a v i d Z e i t l i n , M i n n e a p o l i s , M i n n e s o t a .
f?
and Lucas Numbers," this Quarterly, VoL 2, No. 2, 1964s pp. 105-107, we
obtain, respectively
1968]
(1
"d - L 2 m y
y2) Yl
355
2mky k = F 2 m y ,
k=o
00
(2)
= L m + (-Dln+ILmy
since
L
m " ("^
(L3m - L m ^ m )
- k(F
/T.
J] 5
k=o
2mk
+,
, 2k
xL(2k+i) m )x 2
T
L x(l + 5F x + 5 ( - l ) m + i x 2 )
m
m
1 - 5L 2 m x 2 + 25x4
L x
m
1 - 5F x + 5(-l) m + 1 x 2
since
L2m = 2(-l)m + 5 F ^
and so
1 - S L ^ x 2 + 25x4 (1 + 5 F m x + 5 ( l ) m + 1 x 2 ) ( l - 5 F ^ x + 5(-l) m + 1 x 2 K
m
Also solved by Anthony G, Shannon (Australia).
356
pec.
OF PRIME INTEREST
H - 1 0 5 Proposed by Edgar K a r s t , N o r m a n , O k l a h o m a , and S O . Rorem, D a v e n port, Iowa.
Generalize.
Solution by E. W . Bowen, U n i v e r s i t y of N e w E n g l a n d , A u s t r a l i a .
Let b be a prime* m a positive integer, and fi the least positive r e s idue of m modulo b - 1, L e . , for some integer k, m = k(b - 1) + /x where
0 < fx < b - 1.
Clearly n
= 0 = n^(mod b) if n is a multiple of b.
If n is not a
= 1 (mod b) ,
k(b-i)+/x
= n
ik.M
= 1 n
= n
A u\
(mod b) .
= n
(mod b) .
= pel (mod b) .
1968]
357
This is a generalization of the required result since., with 2 and 3 as the primes,
b| f
6)9 i. e . ,
(n + 1)
- n
An m = 1 (mod
- n
= r (mod bj . b s )
if
m = 1 (mod (bi - 1) (b s - 1)) .
Also solved by J . A . H . H u n t e r , Brother A l f r e d Brousseau, D a v i d Singmaster,
Steven W e i n t r a u b , and Anthony Shannon.
BUY MY NOMIAL?
H-106
Proposed by L C a r l i t z , Duke U n i v e r s i t y , D u r h a m , N o . C a r o l i n a .
Show t h a t
(Ni >~
k=o
'
k=o
358
[Dec.
P(x) ==
k=o
and
-E(k)(n:>-n-k
then P(x) s Q(x) is a known identity (see Elementary Problem E 799, American
Math. Monthly, 1948, p. 30). If a and j3 are roots of x 2 - x - 1 = 0, then
Ln = a n - M n ,
F n = (<*n-i3n)A/5,
and thus
P(a2) + P $ 2 ) = Q(a2) + Q<2)
(a)
P(g2) - P<j82)
(b)
Q(a2) - Q(fi2)
%/5
V5
3BE DETERMINANT!
H-107 Proposed by Vladimir Ivanoff, San Carlos, California,
Show that
F
F
p+2n
TP
q+2n
F .
r+2n
for all integers p, q, r and n.
F
p+n
"C"
q+n
F .
r+n
"C1
p
q
F
r
1968]
359
S o l u t i o n by Co C . Y a l a v i g i , Government C o l l e g e , M e r c a r a , I n d i a .
Let
(1)
D =
Pi + n
" pi+rn
F +
" P2 rn
pi
F
p2
F
p2+n
F
"P , + r n
, F F
D = F
rn
F
"Pi+l
F
p2+i
+ 1
pi+(r-i)n
F
/ v
p2+(r~i)n
+i
+ F
"rn-i
P2
F
p2+n
P r + 1 ^(rDn
"Vi^ 1
F
pi+(r-l)n
F
/ v
p 2 +(r-i)n
F
Pr+i
pt
F
p2
(2)
Pi
F
pi+h
Pi+n
F
p2+n
^
.
P r +i + ( r l ) n
Pr+1+n
Vi
F
pi
F
p2
F
pr+1
* * .* *
lo INTRODUCTION
This article deals with the three Diophantine equations
(1)
x2 + (x + l) 2 = z 2
(2)
u 2 + u = 2v2
(3)
s 2 + 21? = 1 .
These equations have been studied by various methods for hundreds of
(See the h i s -
torical note at the end of Part II, Feb.) However, as often happens with old
problems, people not aware of the long history of these equations keep rediscovering them and their solutions.
onacci Quarterly dealt with Eq. (1), and solved it by transforming it into
Eq. (3). Elementary Problem B-102 in the December 1966 issue of the Quarterly (page 373) also links Eq. (1) and the solution to Eq. (3). Another article
recently submitted to the Quarterly deals with Eq. (2),
The three equations are essentially equivalent because, as we shall see
in Section 9, each can be transformed into each of the other two by a linear
transformation.
2. WHY THE EQUATIONS KEEP COMING UP
The equations come up over and over again because they arise in a natural way from some basic problems of number theory.
A. When the general solution of the equation x2 + y2 = z2 is studied, it
is natural to consider the special case in which x and y are consecutive integers. This leads to Eq. (1).
B. When people play with figurate numbers, and, in particular, with the
triangular numbers
360
Dec. 1968
361
T(u) = | u ( u + 1) ,
and the square numbers
S(v) = v2 ,
they soon observe that
36 = S(6) = T(8) .
This observation naturally suggests the problem of finding all the triangular
numbers that are also square numbers* This problem leads to Eqe (2).
C. There is no rational number s/t equal to the square root of 2. That
is 9 there are no positive integers s and t such that
S2 - 2t2 = 0 .
(4)
approximations with a small e r r o r naturally leads to consideration of the equation obtained from Eq. (4) by requiring the right-hand member to be 1 instead
of 0. This leads to Eqe (3).
3. SOLUTIONS BY TRIAL AND ERROR
One way of finding some positive integers that satisfy Eq0 (1) is to substitute first 1, then 2S e t c , , for x in the expression x2 + (x + I) 2 to identify
values of x which make the expression a perfect square,, Similarly, solutions
of Eq. (2) can be found by identifying by trial and e r r o r some positive integral
values of u that make
fu(u + 1)
a perfect square. And solutions of Eq, (3) can be found by identifying some
positive integral values of t that make 1 + 2t2 a perfect square. Anyone with
362
[Dec*
patience and a table of squares, or who has access to a computer can discover
in this way at least a few of the solutions of each of the three equations.
It will be useful to us to identify not only positive solutions, but nonnegative solutions. The first five non-negative solutions of Eqs. (1), (2), and
(3) are shown in the table below:
Solutions of
Equation (1)
Solutions of
Equation (2)
z
0
u
0
20
29
119
169
696
985
Solutions of
Equation (3)
s
t
0
1
3
17
12
49
35
99
70
288
204
577
408
x0
0,
z 0 = 1,
xt = 3,
x2 = 20,
zi
z2 = 29,
(5)
(6)
dx
Vt
n-i
+ b
Vi
+ ez
n-i
+ c
+ f
1968]
363
Then we have to use only the first four values of x and z to determine what
the values of a, b, c, d, e and f must be. Taking n equal to 1, 2,
and 3
20 = a * 3 + b5 + c
5 = d0 + e - l +f
<29 = d 3 + e*5 + f
119 = a 20 + b - 29 + c ,
/169 = d 20 + e 29 + f
b = 2,
c = 1,
d = 4,
e = 3,
and
f = 2 .
Equations (5) and (6) a r e merely guesses. However, the fact that the values of
a, b, c, d, e and f that we calculated on the basis of these guesses turns out
to be integers, and small ones, at that, is presumptive evidence in favor of
these guesses. Let us continue operating with these guesses. If Eqs* (5) and
(6) a r e true, then they must take this form:
(7)
(8)7
x = 3x
+-2z
+ 1 ,
n
n-d
n-i
z = 4x
+ 3z
+ 2 .
n
n-i
n-i
We can obtain more evidence for or against our guesses by using Eqs. (7) and
(8) to calculate x 4 and z 4 :
x 4 = 3(119) + 2(169) + 1 = 696 ;
z 4 = 4(119) + 3(169) + 2 = 985 .
Since these values of x 4 and z 4 calculated by means of Eqs e (7) and (8) agree
with the values of x 4 and z 4 in the table, the evidence tends to support the
correctness of Eqs (7) and (8). We now know that Eqs. (7) and (8) are true
when n = 1, 2, 3, or 4. This gives us the confidence to seek a proof that
they a r e true for all positive integral values of n. The proof is given in the
next section*
364
[Dec.
EXERCISES
fVi
1. Let (u , v ) be the n
assume that
u = alu
+ bv u + c ,
n
n-i
n-i
'
and
v =' du
+ ev , + f ,
n
n-i
n-i
*
then what values must a, b, c, d, e and f have in these formulas?
2. Let (I(S , t ) be the n
we assume that
s = as
+ bt
+ c ,
n
n-i
n-i
'
and
t
= ds
n-i
+ et
n-i
+ f ,
,z
< 3 Vi
+ 2
Vi +1 -
Vi
+ 3
V.
+ 2)
1968]
365
X2 + (X + I ) 2 = Z 2 ,
(f)
f zf = 4x + 3z + 2 .
feffzf)->fe,z)
366
[Dec.
x = 3xf - 2z f + 1 ,
?
! z = -4x + 3z - 2 .
Then
iV = i
for every positive integer n. That i s ,
f n g n (x,z) = (x,z) .
We shall show first that if (x, z) is a solution of Eq. (1), with x > 0, z > 0,
then
fccijZj)
= g(x,z)
then
x\ + (x| + I)2 - 2xf+ 2xi + 1 = 2(3x - 2z + I)2 + 2(3x - 2z +' 1) + 1
= 18x2 + 8z2 - 24xz + 18x - 12z + 5
= 16x2 + 8z2 - 24xz + I6x - 12z + 4 + (2x2 + 2x + 1)
= 16x2 + 9z 2 - 24xz + 16x - 12z + 4 ,
since
1968]
367
then
xf + (xj + I ) 2 = M x + 3z - 2) 2 = zf .
Therefore (x^Zj) is a solution of Eq* (1). Now we aim to show that xj > 0,
zj > 0, and z\ < z. The condition xj > 0 is equivalent to 3x - 2z + 1 > 0
o r 2z < 3x + l.
or
2z < 3x + 1 ,
and
3z > 4x + 2 .
z 2 = 2x2 + 2x + 1 = 4x2 +' 4x +. 1 - 2X2 - 2x
= (2x + I) 2 - 2x(x + 1)< (2x + I) 2 ,
since x > 0, and hence 2x(x + 1) > 0. Therefore z < 2x + 1. Since
z2 = 2x2 + 2x + 1 ,
and x > 0,
then
9z2 = 18x2 + 18x + 9 > 16x2 + 16x + 4 ~ (4x + 2) 2
Therefore 3z > 4x + 2
4z2
8x 2
+ 8z + 4 = 9x2 + 6x + 1 ~ X2 + 2x + 3
Since x > 0, we see from the table of solutions of Eqe (1) that x > 3* Then
368
[Dec
x2 > 3x = 2x + x > 2x + 3 .
Then
2x + 3 - x2 < 0 .
Consequently
4z 2 < 9x2 + 6x + 1 = (3x + I) 2 ,
and
2z < 3x + 1 .
We have shown that if (x, z) is a solution of Eq. (1) for which x > 0 and
z > 0, then
(xi,Zi) = g(x,z)
is a solution for which xj > 0, zj > 0, and z^ > z. If xA > 0 we can repeat
the process to obtain a solution
(X 2 ,Z 2 )
= g(Xi,Zi)
= g 2 (x, Z)
with x2 > 0, z2 > 09 and z 2 < zj. Continuing in this way as long as x. >
0, i = 1, 2j B * , we get a descending sequence of positive integers z > zj >
z 2 > . Since this sequence must terminate, there exists a positive integer
n for which
(x , z ) = g (x, z)
N
& N
n n
'
is a solution of Eq
M (1)/ with xn = 0. Then z n = 1,' and
(0,1) = ( x n , z n ) = g n (x,z) .
1968]
369
Then
^ ( 0 , 1 ) = ^ g n ( x , z ) = (x,z) .
This completes the proof of P a r t B.
If we return now to the notation of Eqs. (7) and (8), we can say that all
solutions of Eq. (1) are given by the formula
(
(11)
Vzn)
fQ
<1>
n =
1.2.3,---
= 3u
n
(13)
+ 4v
n-i
n-i
+ 1 ,
'
v = 2u
+ 3v
+ 1 .
n
n-i
n-i
(14)
4.
VVn)
^(O^O),
n = 1,2,3,
= 3s
n
(16)
(Continued on p. 317.)
t
n
+ 4t
n-i '
= 2s
+ 3t
n-i
n-i
n-i
H A R M O N I C DESIGN IN M I N O A N ARCHITECTURE
DONALD-A. PREZIOSI
Instructor in Classical Airehaoelogy, Yale University, Nev> Haven, Conn.
During 1964-66 a study was undertaken of the remains of Bronze Age construction in the Aegean Basin in an attempt to establish certain underlying principles of Minoan, Mycenaean, and other Aegean architectural design. Some 330
structures at 90 sites were examined and measured; evidence for planning and
layout-procedures, andfor modular and proportional design canons were sought
The results of the project, presented (June 1968) in partial fulfillment of the
requirements for a doctoral degree at Harvard* are outlined below, i
Between 2000 BG and 1500/1400 BC, Minoan Crete generated the first
large-scale, complex townscapes in Europe and a sophisticated architecture
comparable to the contemporary architectures of Egypt and the Lewant* During
this period a number of large structures, known conventionally as "palaces,."
were constructed at Knosses, Phaistos and Mallia, and (slightly later) at Kato
Zakro, Gournia and Plati (Fig. 1). The largest of these, atKnossos, maybeenclosed within a square roughly 150 meters on a side* The palaces are generally similar in groundplan;
also have extensive paved courtyards bordering their western facades; in all
cases building material is stone, frequently in the form of finely squared
masonry blocks, particularly on the outer facades*
The remains have come to light only since the turn of the century; 2 since
that time Crete has become one of the most thoroughly explored areas of Greece;
!
Begin,ning with the excavations at Knossos by Sir Arthur Evans in 1899; final
publication: A* Evans 9 The Palace of Minos at Knoss ., 4 vols*, 1921-1936;
3h 1900 the palace at Phaisto'sbegantobe uncovered* L. Pernier and L. Banti,
11 Palazzo Mincico di Festos, 2 vols,, 1935 and 1951.
370
Dec.
N -*
KATO ZAKRO
501
oJ MILES
CRETE
Qp
371
372
[Dec.
scores of settlements, large and small, have come to light on this island not
much larger than Long Islands
palaces mentioned above (Kato Zakro)was discovered in 1963, and last Summer
(1967) a new town belonging to the Early Bronze Age (2700 2000 BC) was
uncovered on the South c o a s t
Despite the archaeological familiarity of Minoan remains, a thorough
study of the architecture has yet to be published ? which fact is partly responsible for a good deal of misinformation about its nature* Another contributing
cause is the great contrast Minoan architecture makes with Greek temple
design of a millenium later; the great complexity and seeming irregularity of
the former have provided more than one Classical scholar, trained to appreciate the apparent clarity and simplicity of the Greek temple, with nearly insuperable obstacles to understanding, 1
There are other factors contributing to the general misunderstanding, the
most relevant here being that the palaces (and most notably Knossos) underwent periodic rebuilding and remodelling during the centuries of their use.
This has tended to obscure the fact that each large complex was designed
initially as a coherent whole* Archaeological research has shown that in some
cases (e. g., Phaistos) 2 different sections of a palace were constructed at different times: necessarily, construction of such enormous structures would
have been phased for varying reasons.
That
one of the keys to the solution involves the ratios of a Fibonacci Series will
2
As late as 1957, the author of one of the major textbooks on Greek architecture could write "It appears that the Minoans did not object to disorderly
planning as such; they obviously saw no advantage in symmetry and may have
been lovers of the picturesque at all costs; in fact their architecture r e s e m bles their other arts in showing no sense of form. M A. W. Lawrence,, Greek
Architecture (Penguin, 1957), 34.
The earlier palace at Phaistos was built in at least four separate phases, b e ginning on the south and working north, E. Fiandra, "I periodi struttivi del
primo palazzo di Festos, " Kritika Khronika 15/16 (1961-62) 112 ff. Nevertheless, the entire plan is" a unity, as demonstrated by a study of the
measurements.
1968]
373
become apparent below. Within the last decade the researches of Prof. Graham
of Toronto have provided a number of initial insights into the nature of Minoan
metrology* 1 He postulated the existence of a builders' modules (which he called
the "Minoan Foot") used in the layout of thepalacess the value was set at .3036. 2
That Graham 1 s conclusions were premature was shown by the results of
the aforementioned project, in which some 12*000 measurements were made on
structures both on Crete and elsewhere in the Aegean and Greece: evidence of
four modules was found* Of the 330 structures examined, 217 revealed clear
evidence of modular usage (or at least were sufficiently well-preserved to admit of careful measurement):
MODULE
values
times found:
A_
.2704
.3380
.4330
.3036
104
67
17
other 3
20
An immediate curiosity was that the distinction in usage revealed no consistent geographical or chronological pattern; L e., A was not found in area X
to the exclusion of B, C, and D9 nor was its use limited to one chronological
period within the Bronze Age, e t c
builder was more or less free to choose any of the four in laying out his
structure. 4
A comparison was then made of the scales based on each module (Fig, 2);
it was noted that there were certain consistent points of contact among the
scales of a single unit of measurement.
geometrically (Fig. 4); if a rectangle is constructed with the short side equal
to .5408, and the long side .676 (= 2 x .338), then the diagonal = 2 x .434. The
relationship is 10:16 or 5:8. It is of interest also that the diagonal bisecting
i j . W. Graham, The Palaces of Crete (Princeton, 1962) ch* XDI, w. refs.
2
Graham ? s study involved a sample of measurements of palace sections similar in design at various places; a by-product of a study of window-recesses,
the author did not have as his purpose a comprehensive metrological examination of architectural remains of the Bronze Age in Greece.
As if (to use a New Haven example) Brockett f s street grid of 1643 employed
rods and each of the Yale colleges was laid out on a different system (meters,
feet, fathoms, etc.). Examples may be found in World Weights and Measures,
UN Handbook M/2l/rev 1. (1966), for simultaneous usage of several systems
in a given country.
374
A
.2704
.5408
.8112
1.0816
1.3520
1.6224
1.8928
2.1632
2.4336
2.7040
2.9744
3.2448
3.5152
3.7856
4.0560
4.3264
4.5968
4.8678
5.1376
5.4080
6.7600
8.1120
9.4640
10.8160
12.1680
13.5200
14.8720 .
16.2240
17.5760
18.9280
20.2800
21.6320
22.9840
24.3360
256880
27.0400
5.6784
9.1936
14.8720
24.0656
38.9376
63.0032
101.9408
164.9440
[Dec.
.3380
.6760
1.0140
1.3520
1.6900
2*0280
2.3660
2.7040
3.0420
3.3800
3.7180
4.0560
4.3940
4.7320
5.0700
5.4080
5.7460
6.0840
6.4220
6.7600
8.4500
10.1400
11.8300
13.5200
15.2100
16.9000
18.5900
20.2800
21.9700
23.6600
25,3500
27.0400
28.7300
30.4200
32.1100
33.8000
.4330
.8660
1.2990
1,7320
2.1650
2.5980
3.0310
3.4640
3.8970
4.3300
4.7630
5.1960
5.6290
6.0620
6.4950
6.9280
7.3610
7.7940
8.2270
8.6600
10.8250
12.9900
15.1550
17.3200
19.4850
21.6500
23.8150
25.9800
28.1450
30.3100
32.4750
34.6400
36.8050
38,9700
41.1350
43.3000
.3036
.6072
.9108
1.2144
1.5180
1.8216
2.1252
2.4288
2.7324
3.0360
3.3396
3.6432
3.9468
4.2504
4.5540
4.8576
5.1612
5.4648
5.7684
6.0720
7.5900
9.1080
10.6260
12.1440
13.6620
15.1800
16.6980
18.2160
19.7340
21.2520
22.7700
24.2880
25.8060
27.3240
288420
30.3600
7.0980
11.4920
18.5900
30.0820
48.6720
78.7540
127.4260
206.1800
9.0930
14.7220
23.8150
38.5370
62.3520
100.8890
143.2410
244.1300
6.3756
10.3224
16.6980
27.0204
43.7184
70.7388
114.4572
185*1960
1968]
METRIC
,2704
e678
1.08
1*352
11.52
2 8 02
2*16
2.704
3.04
3.38
4.056
4.33
5 9 408
6.07
6.76
8.112
9*10
9,464
. 9,72
10*82
11.24
12,16
13.52
14.872
15.21
16.224
18,20
20.28
21e22
21.64
27,04
30.36
33.80
4Q e 56
47.32
54*08
60.72
2.5
4
5
67.60
7580
.3380
.4330
25
4
5
9
10
12
13
16
18
20
24
27
28
15
18
20
25
30
32
60
36
40
44
45
48
75
60
80
100
125
150
175
200
250
4*5
5
7.5
8
10
45
50
55
.3036
u 0
35
38
40
375
80
90
100
120
140
160
180
200
(approximate)
9
10
11
10
14
18
20
21
30
25
26
28
32
. 36
37
40
45
35
50
42
60
49
50
70
70
90
100
125
140
\180
200
175
250
376
[Dec.
1968]
377
the central axis is .3025 (.3036 = D>. Using .3380 as base integer, a F i b o n acci Series may be generated in which all four modules appear:
.3380 / . 6 7 6 0 / 1.014/ 1.690 / 2.704 / 4 . 3 9 4 / 7.098/11.49/18.60/30.1.
1
13
2.704
4.33
21
34
55
89
30,36)
The value of the first integer 9 *3380? was tentatively taken as the base
unit of measurement on which the three variations depend.
This quadripartite system forms the basis of the harmonic system of
Minoan architectural design., and brings into focus the complicated system of
relative proportions of various subsections of a structure. An excellent example is the western-facade section of the palace at Mallia (Fig. 5).
The section with which we are concerned consists of three subsections
further articulated into three wall-planes, two projections and one recess per
subsection* The designer gave the wall-planes the proportions shown in Fig* 6S
A = 8, B = 5 + 5 + 5, C = 8. The Fibonacci integers (base = 2 x .338 =
.676) are also indicated;
SECTIONS
SUBSECTION:
(actual)
1. 2. 3
3*31/3*13/3*31
4f 5, 6
6*03/6*06/6.09
79 8, 9
3*72/2*82/3*65
(ideal)
3*38/ r /3*38
6*02/6*02/6*02
3*72/ r
10 / 10 / 10
INTEGER NO*:
/ r /4(= 10)
/3 e 72
9/ r / 9 ,
where r = remainder (i* e*, Al, A3$ C79 C9 were staked out from outer edges
inward; r having a metrological value of null)*
C approximate in toto the 11th Fibonacci Integer of this series (9*75 vs* 9*75
(A) and 10*19 .(C)); the latter is in e r r o r by *445 or one unit of Module G (*433)s
the module generally used in the layout of the palace*
The system of proportions employed by the Minoan architect in the detailed articulation of the perimetral walls extends also to the underlying grid
of a palace 1 s groundplan*
^Factors such as solar orientation of buildings., as well as alignment of c e r tain building axes on prominent landscape features, play a role in design also 9
as yet not fully understood* Cf* V* Scully? The Earth g the Temple and the
Gods (Yale 1962 and 1968) ch* 2; and below, n* 11*
37.8
PORTE NORD
ESPLAND:
QUES'
PORTE SUD
1968]
~o 3C~h'
71 0)
x>
T3 (Q
O
CD
3
"7
C*m
o
3
Fig, 6 Analysis of Mallia West Facade
379
380
[Dec.
grid based on a rectangle with the proportion 5s8 (160B x 200 B).
The basic
rectangle was presumably laid out with pegs and ropes of fixed length,,
The
center point was found (by means of diagonals or rope, which would have to be
200 C in length 1 ), and four quadrants were further indicated.
quadrants will delimit the central court, the two we stern the west-central block
of the palace* At the center of the overall rectangle (or along an EW axis p a s sing through that point) was constructed a room of (presumably) some ritual
significance, the so-called Pillar Crypt
The entire palace may be generated by subdivisions or additions of fixed
modular size to the central grid-rectangle; the procedures vary in detail among
the palaces. It is noteworthy that the subdivisions of a grid coincide generally
with the functional subdivisions of a palace*
A Palace maybe described as a grid of squares of varying size, the sizes
determined by a sequence of interrelationships based on proportions such as
3:5, 5:8, 8:13, e t c . , as well as 1:2,
the diagram; the sequence of six integers with .067 as base is found in various
subsections of the board; overall proportion is 8:13 (Fig. 9).
1968]
381
382
1.375
.547
.80
.81
,28
.28
2
3
[Dec,
1968]
383
384
[Dec.
This brief consideration of various aspects of Minoan design and its r e lation to the Fibonacci system 1 would best be concluded with the following observation* We are not now in a position to understand the full significance of
the harmonic system of Minoan design* The principles of the Fibonacci system
were certainly understood, as evidenced by the monuments themselves. Whether
these principles were a trade secret of a small class of artisans or more widely understood is not known. 2 Of several scripts used by the Minoans during
their history, only one* the last one to be employed, has been deciphered; it
is a primitive form of Greek, and it was employed for bureaucratic purposes
(listing of commodities, produce, etc.) only; no literature survives, and we
certainly have no mathematical t r e a t i s e s . 3
It is now known that contemporary Egyptian architecture reveals design
principles based in part on the Fibonacci system; 4 it has been known for some
time that early in the palatial period Minoan craftsmen were employed in the
construction of at least one major Egyptian monument. 5 It would seem reasonable to assume that such a situation would provide an opportunity for the diffusion to Crete of the principles of this system. If this was the case, 6 it should
be borne in mind that it was only the principles which were diffused, for the
Minoan system is grounded in a Minoan metrological system, and the Egyptian
is based on a native cubit-measure system.Whatever the case, the essential point remains: inlaying the foundations
of architecture in Europe,
d i s c u s s e d in detail as it applies to some 50 structures in the author's d i s s e r tation, "Minoan Palace Planning and its Origins" (Harvard 1968) Chapter HI
(unpublished).
2
One of the primary ritual S3/mbols of Minoan Crete, the n double-axe n sign,
incised on walls within the palaces, may be an ideogram of the 5:8 triangle of
Fig. 4; for a similar situation, cf. A, Badawy, Ancient Egyptian Architec
tural Design (UCLA 1965) 40-46.
3
Cf. J. Chadwick, The Decipherment of Linear B (New York 1958); M. Ventris
and J. Chadwick, Documents in Mycenaean Greek (Cambridge 1959) 117.
4
Ibid., 14. A measuring rod of .676 was found at Kahun (| = .338) which .(not
being an Egyptian measure) might be connected with the Minoan workmen
employed in the construction of the pyramid of Sesostris n. On the other
handj Levantine workmen were also employed there; there is no firm basis
for deciding to whom the rod should be attributed.
[Continued on p* 317* ]
RECREATIONAL MATHEMATICS
Joseph S, Madachy
4761 Bigger R d . , K e t t e r i n g , O h i o
A FIBONACCI CONSTANT
I would like to introduce a new constant, if it hasn ? t been done before.
If you evaluate the continued fraction
1 + 1
2 + 3
5 + 8
13 + 21
etc.
386
RECREATIONAL MATHEMATICS
[Dec.
The
Idea of instant arithmetic is not new, having appeared in the Fibonacci Quarterly [2, 3] and elsewhere [4, 5J.
I wondered If there were other integers, N, such that when any one or
more of the initial digits were transposed intact to the right, the result would
be N/k, where k is any integer. In other words, as an example, Is there an
integer which can be divided by 7 by moving its initial digits, 317, to the right?
The answer is yes. Although not all integers possess the desired property,
there are an infinite number of integers that do.
Trigg [ l ] shows that, for single-digit transposition
where d is the initial digit to be transposed to the right and F is the proper
fraction which, when written as a decimal for one period, or cycle, r e p r e sents the integer sought
F = 16/39 = .410256410256 .
Therefore, the smallest integer which can be divided by 4 by transposing the
initial digit to the right is 410256.
Now, I will show how to find integers such that the transposition is not
restricted to single digits, nor need N be divisible by the transposed digits.
Following Trigg's format, let D represent the initial digit or digits to be
transposed from left to right,
Then
1968]
RECREATIONAL MATHEMATICS
387
N/k = 0 . - - - D - - - D - - - .
Therefore
10 n N - N / k = D,
or
(1)
Dk
10 n k -
N = (with no d e c i m a l point)
12/19
= 631,578,947,368,421,052
63/69
= 9,130,434,782,608,695,652,173
73
37
317
317
2701/3699 = 730,197,350,635,306,839,686,401
388
RECREATIONAL MATHEMATICS
[Dec.
Instant division in other bases can be done also. We have, for any base
N =
Dk
b \ -
N =
Dk
10 n k - 1
Base
Three
12
101/122
= 1202122110201001
Four
23
201/233
= 23032332220312131331101
Five
13
31
1003/3044 = 1310022231202000303
(2)
N =
-^a
10 q - p
obtaining
RECREATIONAL MATHEMATICS
1968]
389
P
3
JL
2
_ _ _ _
1,176,470,588,235,294
3
7
2
11
2,352,941,176,470,588
3,203,883,495,145,631,067,961,165,048,543,689
2
7
3
3
1
2
11
?L
4,705,882,352,941,176
428,571
11,262,798,634,812,286,689,419,795,221,843,003,412,969,283,276,450,511,945,392,491,467,576,791,808,873,720,136,518,771,331,058,020,477,815,699,658,703,071,672,354,948,805,460,750,853,242,320,819 (146 digits)
p = 3,
and q = 2,
we obtain
N = 7,058,823,529,411,764
which is not a solution for two reasons:
7, is not equal to
D, nor is the integer produced by transposing the 7 to the right in the ratio
3:2 to the calculated N.
Tentatively, I have found that, for proper solutions Dq must be less
than (q/p)(10 n q - p). Perhaps readers can provide further insight, or provide definite criteria.
NOTE: In [ 3 , problem 2] it is proven that there is no integer which is
doubled when the initial digit is transposed to the r i g h t
However, I found
and
and
and
125000 000
250000 000
375000 000
By including as many 9 f s or 0 f s as necessary, you can get as close to d o u b ling as you desire,.
It is possible* however,
to*double
by moving two or
390
RECREATIONAL MATHEMATICS
[Dec.
TRIANGLE DISSECTIONS
Mel Stover first asked Til if it is possible to cut an obtuse triangle in
smaller triangles, all of them acute. It was proven that it can be done and that
no more than seven acute triangles are necessary [ 2 ] . Martin Gardner [ i j
showed that a square can be dissected into no less than eight acute triangles,
and then asked if a square could be dissected into less than eleven acute i s o s celes triangles.
and Free
Thus, since t r i -
angle A may be dissected into seven acute isosceles triangles, it follows that
a square may be dissected into 10 acute isosceles triangles.
Figure 1
Figure 2
1968]
RECREATIONAL MATHEMATICS
391
Step 2,
If, In Figure 3 (which is Triangle A of Figure 1), we cut off an isosceles
triangle of vertex angle 15 , the remaining triangle is obtuse with A_ = 15,
B = 97.5, and = 67.5. In [5] it was proven that any obtuse triangle can
be dissected into eight acute isosceles triangles.
However, if an obtuse t r i -
angle is such that B * 90s B - A < 90, and B - < 90, then only seven
a r e needed.,
triangles*
Figure 3
Step 3.
Let the triangle with angles 15, 97.5, and 67*5 (which can be cut into
seven already) have an isosceles triangle with vertex angle 15 removed,
leaving a triangle with angles 15, 67,5, and 97.5 which can be cut into seven
acute isosceles triangles.
Isosceles triangles.
needed to get any n ^ 10 (recall we already have 10, 11, and 12). However,
at the point where you had the 10-piece dissection, you can draw lines joining
the midpoints of, say, the equilateral triangle (in Fig. 1) to go from 10 to 13,
Then Steps 2 and 3 can go from 13 to 14 to 15.
desired, or, for that matter, one of any shape as in the 10-piece dissection.
REFERENCES
1. Martin Gardner, "Mathematical Games* " Scientific American; Vol. 202?
No. 2 (February 1960), p. 150 (Solution in the same Journal, Vol. 202, No*
3 (March 1960), pp. 176-178).
392
RECREATIONAL MATHEMATICS
2. Wallace Manheimer,
Dec. 1968
REFERENCES
1. Charles W. Trigg,
3, No. 3
(October 1965), page 201 (Solution by Robert H. Anglin in the same J o u r nal, Vol. 5, No. 1 (February 1967), pp. 75-76). An instant multiplication
problem.
3. J. Wlodarski, "A Number Problem," Fibonacci Quarterly, Vol. 6, No. 2
(April 1968), pp. 161 175. Another instant multiplication problem.
4. D. O. Shklarsky, N. N. Chentzov, I. N. Yaglom, The USSR Olympiad
Problem Book (Edited by Irving Sussman, translated by John Maykovich),
W. H. Freeman and Co., San Francisco, 1962. Specifically Chapter 2.
5. Sidney Penner, "Problem E1350," American Math. Monthly, Vol. 69, No.
7 (Aug.-Sept. 1962), p. 667 (Solution by J. W. Ellis in the same Journal,
Vol. 70, No. 4 (April 1963), pp. 441-442).
* * * *
ticular type known as the linear recursion relation which by its simplicity,
range of application, and interesting mathematical properties deserves special
consideration.
and the method of expressing its terms by means of the roots of an auxiliary
equation analyzed.
tions in more variables which characterize the plane in three-space, the hyperplane in four-space, etc. By further analogy, one speaks of linear differential
equations in which the dependent variable and its derivatives are not found in a
degree higher than one. In this context it is natural to call a recursion r e l a tion of the form:
(1)
T n + i = a i T n + a2Tn1 + a 8 T n _ 2 + ' + a r T n - r + i
If a
is the last
non-zero coefficient, then this would be spoken of as a linear recursion r e l a tion of order r.
Note that there is no allowance for a constant term.
no restriction.
then since
= 3T
n-i
-2T
393
n-2
+4T
n-3
+8
This, however, is
394
[Dec*
(2)
T , = r T
n+i
n
Evidently, this
nique that Is being developed for relating the terms of the sequence with the
roots of an auxiliary equation, we set up the equation corresponding to this
recursion relation, namely:
(3)
x - r
= 0
which has the one root r* The term of the sequence can be written as a multh
tiple of the n
power of this root* thus:
Tn = ( a/ / r ) r n
That this term satisfies the recursion relation (2) follows from (3), since-:on
substituting r for x, we have*
r = r
and on multiplying both sides by r '
n
r
n-i
= r - r
1968]
395
Note that the powers of the root have the same recursion relation as the terms
(2)! So if
= (a/r)rn+1
and
n+i
r
n
= r r
T ^ = r ( a / r ) r n '= r T .
n+i
' '
n
Perhaps due to the simplicity of this case, the considerations are confusing!
But let us pass on to a second-order linear relation where the operations are
not so obvious.
n+i
= 5T
- 6T
n
n-i
If all terms are brought to one side and equated to zero, the result is:
(5)
N
'
T ,. - 5T + 6 T
= 0 .
n+i
n
n-i
If now the successive terms are replaced bypowers of x one obtains the auxiliary equation
(6)
x 2 - 5x + 6 = 0
whose roots are r = 3, s - 2. Since they satisfy the equation (5), it follows
that
396
pec.
r 2 = 5r - 6
s 2 = 5s - 6
Since we may multiply by any power of r or s,
r I1+1 = 5 r n - Gr11"1
(7)
s n + 1 = 5 s n - es 1 1 " 1
Note that the powers of r and s satisfy the same recursion relation (4) as the
terms of the sequence T . Hence if we express these terms as linear c o m binations of powers of r and s, we should obtain expressions that satisfy the
recursion relation (4). Set
= a r ^ + bfl 11 - 1
T
T
= ar 1 1 + b s n
n+i
= 5T
- 6T
n-i
Then
= a(5r n - 6 ^ )
+ b(5s n - 6s11""1)
or
T , = ar
n+i
+ bs
so that the form of the term persists for all values of n once it is established
for two initial values.
What this implies is that given any two starting values Tj = p, T 2 = q
it is possible to find a sequence
= a3n + b 2 n
1968]
397
2 = a . 2 + b3
7 = a 22 + b - 3 2
Solving for a and b we obtain a = - 1 / 2 , b = 1, so that in general,
= (-l/2)2 n + 3 n
If the roots r and s are real and distinct with r s f 0, it will always be
possible to solve the above set of equations for the determinant of the coefficients of the equations:
p = a r + bs
a = ar 2 + bs 2
is
= rs(s - r)
For
10x2 - 31x + 30
398
[Dec.
so that
'T
n+i
= 10 T - 31T
+ 30T
n
n-i
n-2
Thus
= (-5/6)2 n + 2 3 n + (~4/l5)5 n .
Evidently, there are many questions that require further study; the case
of equal roots of the auxiliary equation; what happens if the roots are irrational;
the situation in which the roots are complex; and various combinations of these
cases. Such matters will receive attention in a number of subsequent lessons.
PROBLEMS
1. Find the recursion relation for the sequence beginning 3, 10 with
terms in the form
T
= a + 2% ,
T n+i
__, = 8T n - 15T n-i4 ,
1968]
express T
399
equation.
3. The sequence
5, 13, 61, 349, 2077, 12445, 74653, 447901,
obeys a linear recursion relation of the second order.
express T
equation,
4. A sequence with initial terms 3, 7, 13 has an auxiliary equation
x 3 - 6x2 + l l x - 6 = 0 .
Express the term T
equation,
5. A third-order recursion relation governs the terms of the sequence:
1, 6, 14, 45, 131, 396, 1184, 3555, 10661, 31986, 95954, 287865, 863589 .
Determine the coefficients in this recursion relation and express the term T
n
as a linear combination of powers of the roots of the auxiliary equation.,
LESSON TWO SOLUTIONS
1.
= -4 + (7/2) 2 n
T
n
F i r s t ten terms^ 3, 10, 24, 52, 108, 220, 444, 892, 1788, 3580.
2*
Tn
(13/6) 3 a + (-3/10) 5 n
= 17/5 + (4/15) 6 n
T __, = 7T - 6T ,
n+i
n
n-i
T
-2 + 3*2 a + (-1/3) 3 n
T , = 3T + T
- 3T 0
n+i
n
n-i
n-2
T
Each problem or
solution should be submitted In legible form, preferably typed in double spacing, on a separate sheet or sheets, in the format used below. Solutions should
be received within three months of the publication date,
B-148
Let F
and L
n
11
F, t v = F L L0 L L /o f-i \
(2%)
n n 2n 4n
(21 TI)
B-149
Show that
L
B-150
n+1
= 5F F i .
M L L + 4(-l)
n+i
n+3
n n+4
Show that
L2 - F 2 = 4F F ^ .
n
n
n-i n+i
B - 1 5 1 Proposed by Hal Leonard, San Jose State C o l l e g e , San J o s e , C a l i f .
Pn(x) = n
(l+x1)
= a 0 + ajx + -
+ a m xm .
i=i
Let q be the number of integers k such that both 0 < k < m and a, = 0.
Find a recurrence relation for the q .
n
400
Dec. 1968
401
m+n
B-153
m+i n+i
m-i n-i
Proposed by K I a u s - G u n t h e r Recke, G o t t i n g e n , G e r m a n y .
P r o v e that
With the centers of the smaller circles placed at the vertices of a regular
pentagon, the smaller circles cover the larger one with a ratio of diameters
equal to the golden ratio (1 + \ / 5 ) / 2 . There may exist another arrangement of
the five circles which results in a smaller ratio.
EVEN AND ODD SEQUENCES
B-131a Proposed by Charles R. Wall, University of Tennessee, Knoxville, Tenn.
Let JH I be a generalized Fibonacci sequence, i. e . ,
H
n+2
= H
+1
+ H
H0 = q* Hi = p,
negative subscripts,,
402
[Dec.
and since F
IH
= (-1)
F , we have
I = (-l) n (qF
H
J -n J
'
n-i
- (p
q)F )j = IqF
F - H/
n I
p n-i4
- F(p - q)Fn I.
|
Let U
and V be solutions of
m
n
W , = aW , + bWn
n+2
n+i
n
= |Wn
EXPONENT PROBLEM
B-132
belongs modulo
1968]
403
- F 2 = (-l) n
A F
n+i
n-i4
n
we have
F^ E ( - l ) n + 1 (mod F n + 1 ) ,
Now F
1 (mod F
') as 1 / F < F , for n > 3. If n is odd then
n
n+i
n
n+i
Fn = 1 (mod Fn~r"i
, ). If n is even then Fn2 = - 1 (mod Fn"+i
, ). Now
2
F3 = - F = F
(mod F , )
n
n
n-i
n+i
and F
AN OLD P R O B L E M I N FIBONACCI C L O T H E S
B-133 Proposed by Douglas Lind / University of Virginia, Charlottesville, V a .
Let r =
is the larger?
FJOOO
anc
s = F
r
and s , which
404
pec.
vi'v>
Hence r S > s r .
by
aA = a2 = 1,
a2k+i = a2k
a 2 k-i
and
a
for
k > 1.
2k
Show that
n
2_s\
k=i
n
=
2n+i - h
2^ a 2 k ~ 1
k=i
a4n+1
" a2n+i
]T}a k = ^ a
k=i
2 k
k=i
= a 2 n + i - at = a 2n +t - 1 .
Then,
n
k=i
a2k
-i
h
+
n
a
2n
k=i
=
k=-i
k=i
(a4n+i ~ !) - (a 2 n + j - 1)
k=i
=
a^+j - a 2 n + i .
1968]
405
GENERALIZED SUMS
B-135 Proposed by L. Carlitz, Duke University, Durham, No. Carolina.
Put
n-1
n-1
1,
n
F,' = 2 - F
n
n+2 '
L ' = 32
n
- L + ' :.
n 2
Knoxville; Te,nnessee.
H'
k=o
H'n
for all n
1.
Identity (A_ can be verified for small n; assume that (A) holds for n.
Then 'since
(H
we have
-H
n+3
n+1
)+H
n+2
-(H
n+2
-H)
n+1
406
n
H
n+i = E
k=o
- Hn
2H; = 2 n + i H 2 - 2H n+2
H n 2+\
Dec. 1968
- H ^
Thus (A) holds for all n > 1. To obtain the identities given by Carlitz, we
note that F 2 = 1, L2 = 3.
Also solved by Herta T. Freitag, D 9 V . Jaiswal (India), Bruce W . K i n g , C . B A .
Peck, A . C. Shannon (Australia), David Z e i t l i n , and the proposer.
* *
ERRATA
Please make the following correction in the October Elementary Problems and
Solutions: In the third equation from the bottom, on p. 292, delete
F
2k
2k+2
F2k
^2kH
F2k*i
F2k
<
F 2 k-i
"T^~
F2k+2
Fifcf?
F 2 k+!
F2k+2
F 2 k-i
"FiF
* * * * *
[Continued from p. 334. ]
Hence, by (13), p \ DJn
In each case we have found a reduced arithmetic progression no prime
member of which is a factor of a certain D| n . Hence, by Lemma 1, II), there
is an infinitude of composite DJ n +j.
REFERENCES
1. R. D. Carmiehael, "On the Numerical Factors of the Arithmetic Forms
a n / 3 n , " Annals of Mathematics, 15 (1913-1914), pp. 30-70,
2. W. J. LeVeque, Topics in Number Theory, I (1958).
*
* *
=A B ,
where
A = 5F 2 - 5F + 1, B = 5F 2 + 5F + 1,? A < B , (A , B ) = 1 ,
n
n
n
n
n
n
n
n'
n' n
where F
= 151,
= 11L , and
(1)
< 3F
33 < 5(F
(2)
n
- 1)
(n> 2) ,
(n > 5)
Equation (1) is easily verified for n = 3,4. If (1) is valid for n, n + 1, its
validity for n + 2 follows by addition of the corresponding inequalities sidewise. Similarly (2) is shown* Hence
HLn<
1 1 . 3 P n = 3 3 F n < 5(F n - l ) F n = 5 F ^ - F n
< 5Fn2 - F n + 1 = An .-
407
VOLUME INDEX
ABLER, Irving,
360-369.
408
1968]
VOLUME INDEX
409
CARLITZ, L.
" B e r n o u l l i N u m b e r s , " Vol. 6, No. 3 , pp e 7 1 - 8 5 . " F i b o n a c c i
R e p r e s e n t a t i o n s , " VoL 6, No. 4, pp. 193-220.
Problems Proposed:
H-126, Vol. 6, No. 1, p . 50; B - 1 1 6 , VoL 6, No. 1, p . 93; B - 1 3 5 ,
Vol. 6, No. 1, p . 90; B - 1 1 7 , VoL 6, No. 1, p. 94; H-134, VoL 6.,
No. 2, p. 143; H-139, VoL 6, No. 4, p. 251; H-97 5 VoL 6, No. 4,
p . 256; B - 1 3 5 , VoL 6, No. 6, p. 405; H-106, Vol. 6, No. 6, p . 357;
H-144, VoL 6, No. 6, p. 351. P r o b l e m s Solved: H - 8 1 , VoL 6, No. 1,
p. 52; H - 8 5 , VoL 6, No. 1, p . 56; B - 1 1 6 , VoL *6, No. 1, p. 94; B - 1 1 7 ,
VoL 6, No. 1, p. 94; H - 9 2 , VoL 6, No. 2, p . 145; H - 9 3 , VoL 6,
No. 2, p . 148; B - 1 2 0 , VoL 6, No. 2, p. 188; H - 8 8 , Vol. 6, No. 4,
p. 254; B - 1 3 4 , VoL 6, No. 6, p. 405; B - 1 3 5 , VoL 6, No. 6, p . 406.
CAVIOR, Stephen R.
161.
CLARK, E . M.
P r o b l e m s Solved:
VoL 6, No. 2, p . 188.
B-120,
CURL, J a m e s C.
" F i b o n a c c i N u m b e r s and the Slow L e a r n e r , " VoL 6, No.
4, pp. 266-274.
CUSICK, T. W.
"On a C e r t a i n I n t e g e r A s s o c i a t e d with a G e n e r a l i z e d F i b o n a c c i Sequence, " Vol. 6, No. 2, pp. 117-126.
DAIEV, V a s s i l i .
DAYKIN, D. E.
" B a s e s for I n t e r v a l s of R e a l N u m b e r s , " VoL
p . 335 (Co-author, A. J . W. Hilton).
6,
No.
6,
DENCE, T h o m a s .
Problem Proposed:
B - 1 2 4 , VoL 6, No. 4, p. 296,
P r o b l e m s Solved: H - 8 1 , VoL 6, No. 1, p . 52; B - 1 1 2 , VoL 6, Hfo.. 1,
p . 91; B - 1 2 9 , VoL 6, No. 4, p . 296.
DESMOND, J a m e s E.
P r o b l e m P r o p o s e d : H - 1 3 5 , VoL 6, No. 2, p. 143.
P r o b l e m Solved: H - 9 2 , VoL 6, No. 2, p . 145.
DUNCAN, Dewey C.
P r o b l e m s Solved: B - 1 2 4 , VoL 6, No. 4, p . 290; B - 1 2 5 ,
VoL 6, No. 4 , p . 2 9 1 .
EDGAR, G e r a l d , P r o b l e m P r o p o s e d : B - 1 1 2 , VoL 6, No. 1, p. 90.
Solved: B-112, Vol. 6, No. 1, p . 90.
Problem
ENGLUND, David.
ESTRIN, David L.
EVES, Howard.
p . 88.
FERNS, H. H.
" P s e u d o - F i b o n a c c i N u m b e r s , " VoL 6, No. 6, pp. 305-317.
P r o b l e m P r o p o s e d : B - 1 1 5 , VoL 6, No. 1, p. 92. P r o b l e m Solved:
B - 1 1 5 , VoL 6, No. 1, p . 93.
410
VOLUME INDEX
[Dec.
FIELDER, Daniel C, "Special Integer Sequences Controlled by Three P a r ameters, M Vol. 6, No. 3, pp. 64-70. "Generation of Stirling Numbers
by Means of Special Partitions of Numbers," Vol. 6, No. 4, pp. 1-10.
FORD, Gary G. Errata for "Recurrence Relations for Sequences Like {Fp } , "
Vol. 6, No. 1, p. 21.
FORGET, T. W. "Pythagorean Triads of the Form X, X + 1, Z Described
by Recurrence Sequences," Vol. 6, No. 3, pp. 94-104 (Co-author, T. A.
Larkin).
FRIETAG, Herta T. "A Magic Square Involving Fibonacci Numbers, " Vol. 6,
No. 1, pp. 77-80. Problems Solved: B-119, Vol. 6, No. 2, p. 137;;
B-127, Vol. 6, No. 4, p. 295; B-131, Vol. 6, No. 4, p. 304; B-131a,
Vol. 6, No. 6, p. 402;, B-134, Vol. 6, No. 6, p. 405; B-135, Vol. 6,
No. 6, p. 406,
GALE, Gene B. "Factorization of 2 x 2 Integral Matrices with Determinant
1 , " Vol. 6, No. 1, pp. 3-21. E r r a t a for above, Vol. 6, No. 5, p. 10.
GOMMEL, A.
Operator,"
1968]
VOLUME INDEX
411
IVANOFF, Vladimir.
IVIE, John. Problems Solved: B-118, Vol. 6, No. 2, p. 187; B-119, VoL 6,
No. 2, p. 187; B-131a, VoL 6, No. 6, p. 402; B-134, VoL 6, No. 6,
p. 405,
JACKSON, William D. Problem Proposed: B-142, Vol. 6, No. 4, p. 288.
Problem Solved: B-133, VoL 6, No. 6, p. 404.
JAISWAL, D. V. Problems Solved: B-112, VoL 6, No. 1, p. 91; B-116,
Vol. 6, No. 1, p. 95; B-131a, VoL 6, No. 6, p. 402; B-132, VoL 6,
No. 6, p. 403; B-134, VoL 6, No. 6, p. 405; B-135, VoL 6, No. 6,
p. 406.
JAMISON, Free. "Dissection of a Square into n Acute Isosceles Triangles,"
VoL 6, No. 6, pp. 390-392 (Co-author, V. E. Hoggatt, J r . )
JARDEN, Dov. "Formulas for Decomposing F 3 n / F n t F ^ / F n and L 5 n / L n
into a Sum or Difference of Two Squares," VoL 6, No. 1, p. 96. "Any
Lucas Numbers L 5 p, for any Prime p > 5, F^s at J ^ s t Two Distinct
Primitive Prime Divisors," VoL 6, No. 6, p. 107. "On.-the Existence of an Infinitude of Composite Primitive Divisors of Second-Order
Recurring Sequences," VoL 6, No, 6, pp9 322-334 (Co-author; Moshe
Jarden).
412
VOLUME INDEX
[Dec,
JARDEN, M o s h e .
"On the E x i s t e n c e of an Infinitude of C o m p o s i t e P r i m i t i v e
D i v i s o r s of S e c o n d - O r d e r R e c u r r i n g S e q u e n c e s , " Vol. 6, No. 6, pp. 322334 (Co-author, Dov J a r d e n ) .
JORDAN, J . H. "The Twin P r i m e P r o b l e m and Goldbach's Conjecture in the
G a u s s i a n I n t e g e r s , " Vol. 6, No. 5, pp. 81-85 (Co-author, C. A. Halben).
J U H L K E , Clifford.
KARST, E d g a r .
P r o b l e m s P r o p o s e d : B-147, Vol. 6, No. 4, p . 289; H - 1 0 5 ,
Vol. 6, No. 6, p . 356.
KING, B r u d e W. P r o b l e m s Solved: B - 1 1 8 , Vol. 6, No. 2,
Vol. 6, No. 2 , p . 187; B-124, Vol. 6, No. 4, p . 290;
No. 4, p . 296; B-129, VoL 6, No. 4, p . 296; B-131a,
p . 402; B-134, VoL 6, No. 6, p. 405; B - 1 3 5 , Vol. 6,
p . 187; B - 1 1 9 ,
B - 1 2 8 , VoL 6,
Vol. 6, No. 6,
No. 6, p . 406.
KING, C h a r l e s H.
"Conjugate G e n e r a l i z e d Fibonacci S e q u e n c e s , " VoL
No. 1, pp. 4 6 - 4 9 .
6,
KRAVITZ, Sidney.
P r o b l e m s P r o p o s e d : B-130, VoL 6, No. 1, p . 89;
B-130a, Vol. 6, No. 6, p . 4 0 1 . P r o b l e m Solved: B-130a, Vol. 6, No.
6, p . 4 0 1 .
LARKIN, T. A. " P y t h a g o r e a n T r i a d s of the F o r m X, X + 1, Z D e s c r i b e d
by R e c u r r e n c e Sequences, " Vol. 6, No. 3, pp. 94-104 (Co-author, T. W.
Forget).
LEDIN, G e o r g e , J r . "Is E r a t o s t h e n e s O u t ? " VoL 6, No. 4 ,
P r o b l e m P r o p o s e d : H-147, VoL 6, No. 6, p . 352.
pp. 261-265.
L E E , Geoffrey.
P r o b l e m s Solved: B - 1 1 8 , Vol. 6, No. 2, p . 187;
VoL 6, No. 2, p . 187.
B-119,
LEONARD, H. T . , J r .
P r o b l e m s P r o p o s e d : H - 1 4 1 , Vol. 6, No. 4, p . 252;
B - 1 5 1 , Vol. 6, No, 6, p. 400.
LEVINE, Eugene.
"On the G e n e r a l i z e d Langford P r o b l e m , " VoL 6, No. 2,
pp. 135-138. " F i b o n a c c i Sequences with I d e n t i c a l C h a r a c t e r i s t i c
V a l u e s , " VoL 6, No. 5, pp. 75-80.
"The E x i s t e n c e of P e r f e c t
3 - S e q u e n c e s , " VoL 6, No. 5, pp. 108-112.
1968]
VOLUME INDEX
413
LIND, D. A. " T h e Q u a d r a t i c F i e l d Q ( N / 5 ) and a C e r t a i n D i o p h a n t i n e E q u a tion, tf Vol. 6, No. 3, pp. 8 6 - 9 3 . "Symbolic Substitutions into Fibonacci
P o l y n o m i a l s , " Vol. 6, No. 5, pp. 55-74 (Co-author, V. E. Hoggatt, J r . ) .
E r r a t a for "A P r i m e r for the F i b o n a c c i N u m b e r s , P a r t VI, f f Vol. 6,
N 0 . 1, p . 49. P r o b l e m s P r o p o s e d : H-128, Vol. 6, No. 1, p . 51;
B - 1 3 3 , Vol. 6, No. 1, p. 89; B - 1 3 4 , VoL 6, No. 1, p . 90; B - 1 1 3 , VoL
6, No. 1, p . 91; H - 9 3 , Vol. 6, No. 3 , p . 145; B - 1 3 8 , VoL 6, No. 2,
p . 185; B - 1 3 9 , VoL 6, No. 2, p. 185; H-136, VoL 6, No. 4, p. 250;
H-140, VoL 6, No. 4, p . 251; B - 1 4 5 , VoL 6, No. 4, p . 289; B - 1 2 5 ,
Vol. 6, No. 4, p . 291; B-130, VoL 6, No. 4 , p . 297; B - 1 3 3 , VoL 6,
No. 6, p . 403; B - 1 3 4 , Vol. 6, No. 6, p . 404; H - 1 4 5 , Vol. 6, No. 6,
p . 351; P r o b l e m s Solved: B - 1 1 2 , Vol. 6, No. 1, p . 91; ffi-113, VoL
6, No. 1, p . 91; B - 1 1 6 , Vol. 6, No. 1, p. 94; B - 1 1 7 , VoL 6, No. 1,
p . 95 B - 1 1 8 , Vol. 6, No. 2, p . 186; B - 1 1 9 , VoL 6, No. 2, p . 187;
B - 1 2 0 , Vol. 6, No. 2, p. 188; B - 1 2 1 , Vol. 6, No. 2 , p . 188; H - 9 6 ,
Vol. 6, No. 4, p . 254; H - 9 9 , VoL 6, No. 4, p . 257; B - 1 2 5 , Vol. 6,
No. 4 , p . 293; B - 1 3 0 , VoL 6, No. 4 , p . 298; B - 1 3 3 , Vol. 6, No. 6, p.
404; B-134, VoL 6, No. 6, p . 405; H-107, Vol. 6, No. 6, p . 359.
LOMBARD, W i l l i a m .
LOWERRE, G e o r g e F . P r o b l e m s Solved:
B - 1 3 4 , Vol. 6, No. 6, p . 405.
B - 1 3 3 , VoL 6,
No. 6,
p . 404;
MADACHY, J o s e p h S. " R e c r e a t i o n a l M a t h e m a t i c s , " VoL 6, No. 1, pp. 6 0 68; No. 2, p p . 162-166; No. 4, pp. 299-304' No. 6, pp. 385-392.
MANA, P h i l .
P r o b l e m s Proposed: B-136,
VoL 6, No. 2, p . 185; B - 1 2 0 , Vol. 6,
No. 2, p. 188; B - 1 5 2 , VoL 6, No. 6, p.
Vol. 6, No. 1, p . 90; B-120, Vol. 6,
No. 2, p . 189; B - 1 2 3 , VoL 6, No. 2,
p . 404.
MANDELSON, J o s e p h .
" A m a t e u r I n t e r e s t s in the F i b o n a c c i S e r i e s I l l R e s i d u e s of u with R e s p e c t to any Modulus, " Vol. 6, No. 4, pp. 2 7 5 n
278.
MARSHALL, A r t h u r .
B-126,
414
[Dec.
VOLUME INDEX
MOHANTY, S. G.
"On a P a r t i t i o n of G e n e r a l i z e d Fibonacci N u m b e r s ,
6, No. 1. pp. 2 2 - 3 3 .
M O N T L E A F , A. H.
P r o b l e m P r o p o s e d : B - 1 2 2 , VoL 6,
P r o b l e m Solved: B-122, Vol. 6, No. 2, p. 190.
PADILLA, G l o r i a C.
P r o b l e m P r o p o s e d : B-114,
P r o b l e m Solved: B-114, Vol. 6, No. 1, p. 92.
Vol. 6,
No. 2,
No. 1,
ff
VoL
p . 189.
p . 92.
PARKER, F r a n c i s D.
"A F i b o n a c c i F u n c t i o n , " Vol. 6, No. 1, pp. 1-2.
P r o b l e m s Solved: H - 8 9 , Vol. 6, No. 2, p . 144; B - 1 1 8 , Vol. 6, No. 2,
p . 187; B-119, Vol. 6, No. 2, p . 187; H - 8 8 , Vol. 6, No. 4, p . 254;
H-124, Vol. 6, No. 4, p . 290; H - 1 0 3 , Vol. 6, No. 6, p . 354; H-107,
Vol. 6, No. 6, p . 359.
PECK, C. B. A.
P r o b l e m s Solved: H - 8 1 , Vol. 6, No. 1, p. 52; H - 3 2 ,
H - 4 4 , H - 4 5 , H - 6 7 , Vol. 6, No. 2, p. 150; B - 1 1 8 , Vol. 6, No. 2, p . 187;
B-119, Vol. 6, No. 2, p . 187; B-120, Vol. 6, No. 2, p . 188; B-124,
Vol. 6, No. 4 , p . 290; B - 1 2 5 , Vol. 6, No. 4, p . 293; B-126, VoL 6,
No. 4, p . 294; B-127, VoL 6, No. 4, p . 295; B-128, Vol. 6, No. 1*
p . 295; B-129, VoL 6, No. 4, p. 296; B - 1 3 1 , Vol. 6, No. 4, p. 304;
B-131a, Vol. 6, No. 6, p . 400; B - 1 3 3 , Vol. 6, No. 6, p. 404; B-134,
VoL 6, No. 6, p . 405; B - 1 3 5 , Vol. 6, No. 6, p . 406; H - 1 0 3 , VoL 6,
No. 6, p . 353; H-107, VoL 6, No. 6, p . 359.
POND, J e r e m y C.
pp. 97-108.
6,
No.
2,
RABINOWITZ, Stanley.
P r o b l e m P r o p o s e d : H-129, Vol. 6, No. 1, p . 5 1 . .
P r o b l e m s Solved: B - 1 1 2 , VoL 6, No. 1, p . 91; B - 1 1 5 , VoL 6, No. 1,
p . 93.
RAMANNA, J .
P r o b l e m s Solved: B - 1 1 8 , Vol. 6, No. 2, p . 187;
VoL 6, No. 2, p . 188; H - 1 0 3 , VoL 6, No. 6, p . 353.
RECKE, K l a u s - G u n t h e r .
ROREM, S. O.
B-120,
SCOTT, Allan.
" S c o t t ' s F i b o n a c c i S c r a p b o o k , " VoL 6, No. 2, p . 176.
E r r a t a for above, Vol. 6, No. 3, p. 70.
" F i b o n a c c i a n I l l u s t r a t i o n of
L f H o s p i t a l s R u l e , " VoL 6, No. 2, p . 138. "Continuous E x t e n s i o n s of
Fibonacci I d e n t i t i e s , " VoL 6, No. 4, pp. 245-249.
SHAH, A. P .
1968]
VOLUME INDEX
415
SMITH, P a u l .
P r o b l e m s Solved:
VoL 6, No. 4, p . 296.
SQUIRE, M r s . William.
B - 1 2 7 , Vol. 6, No. 4, p.
Problem Proposed:
295;
B-129,
SWAMY, M. N. S.
" F u r t h e r P r o p e r t i e s of M o r g a n - V o y c e P o l y n o m i a l s , "
Vol. 6, No. 2, pp. 167-175.
P r o b l e m s P r o p o s e d : H-127, VoL 6,
No. 1, p . 51; B - 1 2 8 , VoL 6, No. 4, p. 295. P r o b l e m s Solved: H - 8 2 ,
VoL 6, No. 1, p . 52; H - 8 3 , VoL 6, No. 1, p . 54; B - 1 1 2 , VoL 6,
No. 1, p . 91; B - 1 1 6 , VoL 6, No. 1, p. 94; B - 1 1 7 , VoL 6, No. 1,
p . 95; H - 9 3 , VoL 6, No^ 2, p . 146; H - 9 5 , VoL 6, No. 2, p. 148;
H - 8 8 , VoL 6, No. 4, p. 252; H-99, VoL 6, No. 4, p . 258; H - 1 0 1 ,
VoL 6, No. 4, p . 259; B - 1 2 8 , VoL 6, No. 4 , p . 295; B-134, VoL 6,
No. 6, p . 405.
TALLMAN, M a l c o l m .
Problem Proposed:
TAUBER, Selmo.
"On Q - F i b o n a c c i P o l y n o m i a l s , " VoL 6, No. 2, pp. 127134 t " L a h N u m b e r s for Fibonacci and L u c a s P o l y n o m i a l s , " VoL 6,
No. 5, pp. 93-99. "Lah N u m b e r s for R - P o l y n o m i a l s 9 " VoL 6, No. 5,
pp. 100-107.
TRIGG, C h a r l e s W. E r r a t a to " L e t t e r to the E d i t o r , " VoL 6, No. 1, p . 2 1 .
E r r a t a to "Getting P r i m e d for 1 9 6 7 , " to " C u r i o s a in 1 9 6 7 , " and to "A
Digital B r a c e l e t for 1 9 6 7 , " VoL 6, No. 1, p . 33. P r o b l e m s Solved:
B - 1 4 1 , VoL 6, No. 2, p. 186; B - 1 2 6 , VoL 6, No. 4, p. 293.
UMANSKY, H a r l a n .
WALL, C h a r l e s R.
VoL 6, No. 2,
No. 4, p . 257;
p . 298; B - 1 3 2 ,
416
VOLUME INDEX
D e c e m b e r 1968
WEINTRAUB, Steven.
WESSNER, John.
Problems
Vol. 6, No. 2, p . 187;
No, 4, p. 290; B - 1 2 7 ,
p . 296; B - 1 3 1 , VoL 6,
WHITNEY, Raymond E .
E d i t o r of "Advanced P r o b l e m s and Solutions, " VoL
6, No. 1, pp. 50-57; No. 2, pp. 142-150; No. 4, pp. 250-260; No. 6,
pp. 350-359.
WINTHROP, Henry.
" M a t h e m a t i c a l Models for the Study of the P r o p a g a t i o n
of Novel Social B e h a v i o r , " VoL 6, No. 2, pp. 151-156.
WLODARSKI, J.
" P a s c a l ' s T r i a n g l e and Some F a m o u s N u m b e r S e q u e n c e s , "
Vol. 6, No. 2, p . 192. " C u r i o u s P r o p e r t y of One F r a c t i o n , " VoL 6,
No. 2, p . 156. "A N u m b e r P r o b l e m , " Vol. 6, No. 2, p . 161.
"More
about the ! Golden Ratio' in the World of A t o m s , " Vol. 6, No. 4, p . 244.
WOOLUM, J i m .
P r o b l e m P r o p o s e d : B - 1 1 9 , Vol. 6, No. 2, p . 187. P r o b l e m
Solved: B - 1 1 9 , VoL 6, No. 2, p. 187.
YALAVIGI, C. C.
ZEITLIN, David.
P r o b l e m P r o p o s e d : H - 1 0 3 , VoL 6, No. 6, p. 354.
P r o b l e m s Solved: B - 1 2 2 , Vol. 6, No. 1, p . 90; B - 1 5 5 , Vol. 6, No. 1,
p. 98; B-116, VoL 6, No. 1, p . 94; B - 1 1 8 , VoL 6, No. 2, p. 187;
B - 1 2 0 , Vol. 6, No. 2, p . 188; H - 9 7 , Vol. 6, No. 4 , p . 256; B - 1 2 4 ,
VoL 6, No. 4, p . 290; B - 1 2 6 , Vol. 6, No. 4 , p . 294; B - 1 2 7 , VoL 6,
No. 4, p . 295; B - 1 2 8 , VoL 6, No. 4, p . 296; B-131a, Vol. 6, No. 6,
p . 402; B - 1 3 3 , VoL 6, No. 6, p. 404; B - 1 3 4 , VoL 6, No. 6, p . 405;
B - 1 3 5 , Vol. 6, No. 6, p . 406; H-104, VoL 6, No. 6, p . 354; H-106,
VoL 6, No. 6, p . 358; H-107, Vol. 6, No. 6, p . 359.
* *
SUSTAINING MEMBERS
*EL L@ Alder
EL D Allen
R. Ho Anglin
*Joseph Arkin
CoL R9 S, B e a r d
Be C. B e l l a m y
Murray Berg
H. Jo B e r n s t e i n
Leon B e r n s t e i n
*Marjorie Bicknell
John H* Biggs
F r a n k Boehm
Monte Boisen, J r 0
<T Aa B r e n n a n
Ce A* B r i d g e r
Leonard Bristow
*Maxey Brooke
*Bro A, B r o u s s e a u
^Fe LB Brown, J r .
No So C a m e r o n
Lo C a r l i t z
PB Vo Charland
P e J e Cucuzza
*Calvin C r a b i l l
J e R C r e n s h a w
D* E 8 Daykin
Joseph DeCelis
Fo DeKovan
Jo E. D e s m o n d
J* Ho Desmond
Ao Wo Dickinson
No Ao D r a i m
D, Co Duncan
Mrs* Wo Co Dunckel
M a r g u e r i t e Dunton
Mo H* E a s t m a n
Co F* E l l i s
Ho So E l l s w o r t h
Merritt Elmore
Ho W. E v e s
R8 Ao F a i r b a i r n
Ao J* F a u l c o n b r i d g e
*Ho Ho F e r n s
D* C* F i e l d e r
To Wo F o r g e t
Eo To F r a n k e l
Go R0 Glabe
Eo Lo Godfrey
*Ho Wo Gould
Nicholas G r a n t
Go Bo G r e e n e
Bo Ho Gundlach
*Jo Ho Halton
Vo Co H a r r i s
Lo Bo Hedge
Cletus H e m s t e g e r
*Ao P 8 Hillman
B r u c e Ho Hoelter
Elizabeth Hoelter
*V0 Eo Hoggatt, J r 8
^Ao Fo H o r a d a m
Do Fo Howells
Jo A Ho Hunter
Wo Lo Hutchings
Stephen Hytch
*Dov J a r d e n
*So Ko J e r b i c
Jo Ho J o r d a n
B j a r n e Junge
Eo Jo K a r c h m a r
C h a r l e s King
Kenneth K l o s s
^Donald Knuth
Eugene Kohlbecker
Sidney Kravitz
George Ledin, Jr
Hal Leonard
Eugene Levine
*Do Ao Lind
^Co To Long
Ao Fo Lopez
Fo Wo Ludecke
Jo So Madachy
Vo Ao Maxwell
* S i s t e r M e DeSales McNabb
John M e l l i s h
C a r l To M e r r i m a n
Lo R. Morton
P . Jo O ' B r i e n
Fo Jo O s s l a n d e r
Lo A. Pape
R. Jo P e g i s
Mo Mo Risqueno
*D8 Wo Robinson
*Azriel Rosenfeld
To Jo R o s s
Fo Go Rothwell
*L Do Ruggles
Ho Jo Schafer
Jo Ao Schumaker
Bo Bo Sharpe
Lo Ro Shenton
Go Singh
Ao N. Spitz
Do Eo Stahl
Mo No S. Swamy
Ao V. Sylwester
R. Bo Thompson, J r 0
*Dmitri Thoro
C. Wo T r i g g
H. Lo U m a n s k y
R o b e r t Vogt
M. Waddill
*Co Ro Wall
*L. A. Walker
R. Ao White
V. White
Ro Eo Whitney
Luroff Ho Williams
P* Ao Willis
*To Po W i n a r s k e
Alan Wise
C h a r l e s Ziegenfus
^Charter Members
WESTMINSTER COLLEGE
Fulton, M i s s o u r i
DUKE UNIVERSITY
D u r h a m , No Carolina
THE CALIFORNIA
MATHEMATICS COUNCIL