A DETERMINANT INVOLVING

GENERALIZED BINOMIAL COEFFICIENTS

D. A. LIND
University of Virginia, Charlottesville, Virginia*
1.

INTRODUCTION

Define the Fibonacci n u m b e r s

(1.1)

by F j = F 2 = 1,

F ^ - F ^ - F
n+2
n+1
n

=0

T h i s difference equation m a y be extended in both d i r e c t i o n s , yielding

= (-Dn+1F

F_
-n

L u c a s [2] h a s shown that the n X n d e t e r m i n a n t

-1

-1

(1.2)

Fn+1

T h i s i s also a consequence of P r o b l e m s B - 1 3 [5] and B-16 [6] in this Q u a r terly.

Note that the r o w s of (1.2) a r e the n e g a t i v e s of the coefficients of the

difference equation (1.1) obeyed by the Fibonacci n u m b e r s .

The s q u a r e s of

the Fibonacci n u m b e r s obey

(1.3)

F2
- 2F2
n+3
n+2

2F

n+1

F2

0 m

If we take the n e g a t i v e s of the coefficients of (1.3) and place them in a d e t e r m i n a n t analogous to (1.2), we find
* Now at Stanford University,,
113

114

A DETERMINANT INVOLVING

(1.4)

-1

-1

-1

-1

[Apr.

n+1 n+2

Equation (1.4), which appears to be new, m a y b e proved by expanding along

the last column and using induction on n. It is our aim to generalize (1.2)
and (1.4), first for the Fibonacci sequence, and then for arbitrary secondorder recurring sequences.
2.

THE FIBONACCI CASE

We define the Fibonacci generalized binomial coefficients

[-]
Note that

F F
,
m m-1
F

m-r+1

1F2

( r > 0)

' ra

by

1 .

is defined for all integers and all non-negative integers r ,

and that
(2.1)

[Y]

= 0

for

m = 0, 1, . . . , r - 1 .

It is convenient to set
(2.2)

[?]

for

r < 0.

Jarden [ l ] showed that the term-by-term product P

each of which obeys (1.1) satisfies

(2.3)

g ( _ 1)J(J+1)/2[ , ]Pn _ j = 0

In particular, if each is the Fibonacci sequence we have

of k - 1 sequences

1971]

GENERALIZED BINOMIAL COEFFICIENTS

|>

(2.4)

j(j+i)/2rk-|Fk-i

115

This becomes (1.1) for k = 2, and (1.3) for k = 3.

Determinants of the form

(2.5)

a3

a4

e|

a-i

a2

a0

a3
a2

..

a0

at

a2

a0

0 J

are known as recurrents. We shall put the coefficients of (2.4) into an n X n

recurrent and show its value is yet another generalized binomial coefficient.
We remark that a general method for evaluating recurrents, from which the
results here would follow, appears to date back to H. Faure (see [3], Vol. 2,
p. 212).

However, our approach seems somewhat more direct, and the

specific results novel enough to warrant separate attention.

Put
D , = det v(a ) ,
n,k
rs
where
a_

rs

= -(-1)

(s-r+l)(s-r+2)/2

k+i

( r , s = 1,2,,n)

[s - r + lj
Recalling (2.1) and (2.2), we see that D ., is simply (1.2), and that
n,x

is (1.4).
For n > k, expansion of det (a
fication gives

n,k

n.)

) along the last column and simplirs

(2.6)

g ( _ 1)W+1 ,/ 2[kr]lw

116

A DETERMINANT INVOLVING

[Apr.

If we define
(2.7)

D Q j k = 1;

D_njk = 0

for

n = 1, 2, , k - 1 ,

then (2.6) remains valid for n > 1. Now for fixed k, i s t n e Product of
k sequences each obeying (1.1), so that using Jarden's result (2.3) we see

??<-y jo+i)/ Tr][v] =

(^
By (2.1) and (2.7),
D

n,k

^ kX]

(n =

and by (2.6) and (2.8) both D n

recurrence relation. Hence,

"k +

ls

"k + 2' " " ' 0) '

and T n ^ k | obey the same (k + l ) s t - o r d e r

fe

.. k - [ tk]

>

Note that this reduces to (1.2) and (1.4) for k = 1, 2,

3.

respectively.

EXTENSION TO SECOND-ORDER RECURRING SEQUENCES

Let the sequence ( u } be defined by U0 = 0, Uj = 1,

(3a)

n+2 " PUn+l

lUn = 0

(q f 0) .

Let a and b be the roots of the auxiliary polynomial x2 - px + q of (3.1).

We deal only with the case in which (3.1) is ordinary in the sense of R. F.
Torretto and J. A. Fuchs [4], i . e . , we assume that either a = b or a f-b

for n > 0. It follows that

1971]

GENERALIZED BINOMIAL COEFFICIENTS

11

- h11
- b

* b
na
if
a = b .
We define the U-generalized binomial coefficients \ \
\

if

117

by

L Ju
LrJn
u

(r>0);

U..U,.-- U

1 2

LoJ

=1

Note that

I1" I

(3.2)

=0

(m = 0, 1, . . . . r - 1) '.

As with the usual binomial coefficients, we define

I"111]

ra -

(3.3)

=0

(r < 0)

L Jarden
Ju
In a generalization of (2.3),
has shown that the term-by-term product

Q n of any k - 1 sequences, each obeying (3.1), satisfies

(-1)V (J ~ 1)/2 [ f l Q n i = 0 .

(3.4)

j=0

LJJ

Equation (3.4) indeed reduces to (2.3) when p = -q = 1.

We shall use the

negatives of the coefficients of (3.4) to form a recurrent as before.

Let

D n>k (U) = d e t ( b r s ) (
where

118

A DETERMINANT INVOLVING
,

_ ( !vS-r+l ( s - r ) ( s - r + l ) |
k + 1 1
q
rs " " ^
L s - r + lJu

[Apr.
.

(r s

'

.
0
2'"^)

We find it convenient to set

D

(3.5)

0,k

(U)

1;

-n,k(U)

(n = 1,2, . . . f k - 1 )

Then expansion of det (b i ) along the last column gives

i s

for all

n^l.

Noticing that

. |

L*JU

is the product of k sequences eacy obeying (3.1),

we see from (3.4) that

(3.7)

s\- J ^- 1)/2 r k * 1 l V v \ =o

Then

-m

D n>k (U) = T

,,

(n = -k + 1, -k + 2, , 0)

and by (3.6) and (3.7), D n k (U) and f"n * k l

recurrence relation.

obey the same (k + l ) s t - o r d e r

Hence,

v k - f ; k]

(3.8)

We conclude by investigating some particular cases of (3.8).

note that it reduces to (2.9) for p = -q = 1. If

P =

s =

s-1

+ F

s+1'

= ("1)S

First

1971]

GENERALIZED BINOMIAL COEFFICIENTS

then U n = F g n ,
L
s
-1

so that for k = 2 ,

<-irT

(3.8) yields

(-DS

L
s

-1

(-1)

s
0

119

-1

s+1

= F

s(n+l)

Putting s = 1 p r o v e s (1.2).
If we let p = 2 ,

q = 1,

then a = b = 1 and U

= n.

In this c a s e ,

["]-(?)
the usual binomial coefficient.

det

L"M)

Equation (3.8) then yields

\s - r + 1/J

\ k j

( ( r , s = 1, 2 ,

n) .

In p a r t i c u l a r , for k = 2 , we find

n + 1 ,
2
0

-1

which f i r s t s e e m s to have been noted by Welstenholme (see [ 3 ] , Vol. 3 , p.

394). Letting k = 3 , we obtain
[Continued on page 162.]

THE

HIDDEN

HEXAGON

SQUARES

V. E. HOGGATT, JR.
San Jose State College, San Jose, California
and
WALTER HANSELL
Mill Valley, California

INTRODUCTION
Pascal's arithmetic triangle has been much studied. Further study continues to produce evidence of the great fertility of this array of numbers.
Here we divulge a very surprising result.
Theorem. Let I

J be such that 0 < n < m ,

of the six binomial coefficients surrounding

m > 2, then the product

J is a perfect integer square.

Proof. The six binomial coefficients are:

m - l\
n - l/'

fm - l \

/m + l\

/m + l\

\n * if

[ n + if

)>

, I
and

\n

if

The product is
(m - 1)1
(n - 1)1 (m - n)!

(m - 1):
n!(m - 1 - n)!

, ml
(n - l)!(m - n + 15"

m!
(n + 1)1 (m - n +

(m + 1)1
X
(m - n)!.

(m + 1)1 ml (m - 1)!
I5
(n - l)!n!(n + 1)1 (m - n - l ) ! ( m - n ) ! ( m - n + l)I J

Since each binomial coefficient is an integer, the product is an integer, and

since the square of a rational number is an integer if and only if the rational
number is an integer, it follows that the product is an integer square.
Corollary.

Each alternate triad of the six binomial coefficients have

equal products.
[Continued on page 133. ]

SOME SPECIAL FIBONACCI

AND LUCAS GENERATING FUNCTIONS
VERNER E. HOGGATT, JR.
San Jose State College, San Jose, California

In [ l ] , Hoggatt and Bicknell derived by matrix methods that

2n+2

i=0
2n+l
i=0

We next list three more similar sums.

k -

^ n

k=0

k=0

E
(I) 4 k
k=0

^ n

Identity (a) is well known, while (b) was in a private communication from D.
Lind, and (c) is a special case of Problem B-88 in the Fibonacci Quarterly,
April, 1966, p. 149.
In [2], various special related results are also derived by matrix
methods. Here, we derive a new class of generating functions by following
the suggestion given in [3], The column generators for Pascal's left -adjusted
triangle are
121

122

SOME SPECIAL FIBONACCI AND

x

[Apr.

(1 - x)
while the generating function for the Fibonacci numbers i s

G(X)

=y

1 - x - x*

Fx

^
n=0

If we now sum

n=0

n=0 k=0

- x Z-/ n i l - x ]
n=0

'

1 -x

'-n^ra
n

1 -

I F2X'
3x + x'
n

n=0
Thus

S(J)
k=0

k = F2n

Now, if we sum

S'^w-LfEia^V-r-rT
n=0

n=0 \k=0

5x2

1 - 5x +

1971]

LUCAS GENERATING FUNCTIONS

123

This is a special case of the general class of identities

L

= Lm x + 5xF2ni
x2 + OAJ
5 L3m
x3 + 5 21?F
x4
0
0
2
m
+
1
.
4m
2
1 - 5Fmx + ( - l r ^ S x
+ 5 2 L C x5 + . . . + 5 k F 0 1
x2k
5m
2km
k
2k+l
+ 5L
+
(2k+l)mX
'"
'
We discuss first a related special case. To see this requires a few identities and a neat trick in algebra.

It is easy to establish that

3 - 2x
1 - 3x + x 2

L
^
2k+2
k=0

OO

X -

X2

1 - 3x + x 2

2k+1 X

k=0

Now,

3x2 - 10x4

- V T
1 - 15x + 25x
f^
k=0
2

2k+2

x2k+2 ck
5

x(l - 5x2)
_ Y * ^F ,, x~ 2 k + 1 5* k
nl
2
4
1 - 15x + 25x
. ft 2 k + 1
k=0
Notice that
+ 3x2 - 5x3 - 10x4 _
1 - 15x2 + 25x4

Next, we need

(x - 2x 2 )(l + 5x + 5x2)
(1 - 5x + 5x 2 )(l + 5x + 5x2)

x - 2x2
1 - 5x + 5x2

124

SOME SPECIAL FIBONACCI AND

[Apr.

OO

x ( l - x)
1 - 2x - 2x

+ x

= **
E Fn2 xn
n=0

Summing a s b e f o r e ,

F^gn(x)
n=0

- E ( S (j)nV
n=0\k=0

oo

oo

=y

F2
n

n=0

1 - X

/ x in
n u~=-iJ

*L_ = _i_ y
(i X ) n + 1 i - x ^

>

F2

n=0

- AT^\ - (T^)' (r^V

x - 2x2
1 - 5x + 5x2
Thus

n=0

U) 4

n _

k=0

x - 2x 2
* _,_ e 2
1 - 5x + 5x^

sk

.
k=0A

( F2k+l
xL01_0)x 2k+l
9^ +
' AAJ 2k+2

and

2n+2

E( 2 \ +2 )

F2

5nL

2n+2

k-0
2n+l

z
k=0

/Fk

2n+1

'

1971]

LUCAS GENERATING FUNCTIONS

which a r e given in the f i r s t p a r a g r a p h of the p a p e r .

k=0
a r e related.

C l e a r l y , t h e n , the

k=0

We r e t u r n now to the s p e c i a l c a s e

= L j x + F 2 5x2 + L 3 5X3 +
1 - 5x + 5x2

To s e e t h i s , we w r i t e

x + x2
1 - 3x + x'.2

l - 3 x

_ T^
s^t
.
k=0A

E~ "F

k+1
2k+lX

nl
2 k

xk

Next,

x ( l + 5x 2 )
V* T
ck

= L> L 2 k + 1 5
1 - 15x 2 + 25x 4
, n
k=0

2k+l

and

5x2
1 - 15x

Thus,

r >F^
+ 25x

E
2
~

2k

rk 2k
5 x

125

126

SOME SPECIAL FIBONACCI AND

[Apr.

oo

x(l + 5x + 5x*) _ y * k,
1-15x^25x4 " f - ; 5 ( F 2 k
k=0

2k
+ xL

2k+l>x

But,
1 - 15x2 + 25x4 = 1 + 10x2 + 25x4 - 25x2 = (1 + 5x 2 ) 2 - (5x)2
= (1 + 5x + 5x 2 )(l - 5x + 5x2) .
Thus,

l - 5 x + 5x

^(Fo,,
+ xL2 01k _,
)x2k
2 k
+ 1

~k=0" "

"^

and
2n

(2kn)

2k =

"

2n

k=0

E ( 2 \ + 1 K = 5%
k=0

We now return to our general class of identities

L x
^_^ .
m
k,
= y^
m+1
2
1 - 5 F m x + (-l)
5x " J ~
^

2k
XL

(2k+l)m

We begin by writing

XL

(1 + ( - l ) m + 1 x 2 )

-7777--2. (2k+l)m
2m

k=0

.+1

)x

1971]

LUCAS GENERATING FUNCTIONS

127

while

xLm(l

(-Dm+15xS)

1-BL,

2m

x2+25x4

"

^ ^

^
k=0

<2k+1'm

Next,

2m5x

1 - 5L0
x + 25x 2
2m

XL

(1 + ( - l ) m + 1 5 x 2 )
m
_

+ xL

(F

-sr^ xr
ck
= 27~1
, F 2 2k kmm 5
k=0

5x)

Tj5k(F0.

2m

xL

2k
'

x L , _ ,

)x2k

k=0

+ (-l)m+15x2)

(1 + 5 x F
m

xL
_

1 - 5 L 0 x 2 + 25x 4
2m

m
x + (-l)m+15x2

1 - 5F
m

since
5L9
= 5 L 2 + 1 0 ( - l ) m " ^ = 5(5 F 2 + 4 ( - l ) m + 2 ( - l ) I i r r )
Zm
m
m
-m-t-i
= 25 F 2 - 1 0 ( - l ) m + 1 .
m
Thus
1 - 5 L 9 x 2 + 25x 4 = 1 + 1 0 ( - l ) m + 1 x 2 + 25x 4 - 2 5 F 2 x 2
zm
ni
or
(1 - 5 L

x 2 + 2 5 x 4 ) = (l + 5 ( - l ) m + 1 x 2 )
= (1 - 5 F

- 2 5 F 2 x2

x + ( - l ) m + 1 5 x 2 )(1 + 5 F m x + 5 ( - l ) m
m

We now r e t u r n to the g e n e r a l p r o b l e m .

x2).

128

[Apr.

SOME SPECIAL FIBONACCI AND

Remember,

x
=

X) 5 <F2km
^ " " 2km
k=0

1 - 5 F x + ( - l ) m + 1 5x2
m
'

+ xL

' o( ,2 ^k +
)x2k
1)m

The generating function for every m th

We start with the general problem.

Fibonacci number is

x
m
1 - L x + v(-1)
m
'

km
2',
.
k=0

x2

k
x

Consider the sum

n=0

(x)

mnSn

_J_f
1 - x u
n=0

/i_\ n
mnll-xl
V
/
"m 1 - x

"

- V ( T ^ )

<-"

(T^)'

x
m
(1 - x)2 - L m x ( l - x) + (-l) m x 2
F

1 - (L m

n=0

k=0

x
m
+ 2)x + ( L m + 1 + (-D m )x 2

km

Now, from
L2 = L 0
+ 2(-l) m
m
2m

1971]

LUCAS GENERATING FUNCTIONS

129

and

L2 - 5 F 2 = 4 ( - l ) L
n
n
one can obtain four useful identities:

L.
+ 2 = LI
4m
2m
L

4xn -

4m+2

=
=

2m "

2m+1

4m+2 "

+ 4

5F

2m

5F

2m+l

2m+1

Thus, for m = 2a (even)

1 - (L2g

2sX

2)x + ( L 2 g

2)x2

n=Q

k=0

'

2skfX

We now d i s c u s s two s p e c i a l c a s e s ,
F,
x
4m

(A)
J

2m

F 0 ( L 0 x)
2m
2m
1 - L 0 (L 0 x) + (L x) 2
2m 2m
2m

2m/

F0
(L
x)n
2mn v 02m

n=0

n
n
E^.LLx
2mn 2 m '

n=0

Thus,

130

SOME SPECIAL FIBONACCI AND

[Apr.

.n
E{J)F._.-Ii.*.
4mk
^ m 2mn
x

k=0
This is Problem H-88, Fibonacci Quarterly, April 1966, p. 149.

F4m+2 X

(B)
1

"

(L

+ 2)x +

4m + 2

(L

4m + 2

2)x

ji2m+l(F2m+lx)

=
1

"

5F

2m+l(F2m+lx)

5 F

2m+1

2m+iy

5 (F

<F2m+lx)2

XT^ k
L~i 5 ( F (4m+2)k
k=0

^
k=0

(4m+2)k

y2

yL

(2k+l)(2m+l))y

2k

2k

2m+l L (2k+l)(2m+l) ) F 2m+l x

2k

Thus,
2n

( 2 k n )' F(4m+2)k
^ - . ^ = 5n n* F(4m+2)n
_ , _ Fr ^2n
2m+l

k=0
2n+l
F

k=0

*(4m+2)k

5nT

F2n+1

' lj (2n+l)(2m+l) *2m+l

Suppose, on the other hand, that we wish to alternate the signs in the above
sums. Consider the sums

1971]

LUCAS GENERATING FUNCTIONS

FUNCTIONS
oo

km

(-l) g

1
1 - x

(x)

1 -

m=0

m 1 -

m 1 - x
F

131

(-

l)m

y.

m
2
(1 - x) + L m x ( l - x) + (-D x 2

-F

1 + (L

again for even m.

'

4m

- 1 - (-l) m )x 2

- 2)x - (L

Thus,
L0 (-F0
x)
2m
2m

1 + 5 F | m x + 5F|mx2

1 -

5F

2m(-F2mx)

2 m

5(

-F2mx)2

1 - 5F2my

5y^

oo
Kk,_
5 (F

=E

_,_

2mk

v 2k
(2k+l)2m)y

T
yL

k=0
OO

-E

(F

2mk "

\^2k
2mL(2k+l)2m)]

2k

k=0

Thus,
2n

E <-2+k (?)4mk

2n
= F 2m
F2mn

k=0

2n+l

k=0

,2n+l+k f2n + l \

( 1}

r4mk

,JL
TT""
T
2n+l

"

*2m

L,

(2k+1) (2n+l)

'

132

SOME SPECIAL FIBONACCI AND

[Apr.

Proceeding similarly with

F
1

2m+l("L2m+lx)

~ L2m+l(-L2m+lx) " (-L2m+lx)2

2m+ly

2 m + l y " **

S*<(2m+l)k k
r

k=0

^ m + l " (2m+l)k

k=0

Thus

( k | F (4m+2)k

2^ ( D

2m+1 F (2m+l)n

k=0

There remains unsolved

k=0

k=0

for m odd and greater than 3.

Corresponding formulas are given also in

[4] as follows:
2n

T
F
E ( t ) 1(4m+2)k
- - - - -- 5 **----*
(2m+l)2n 2 m + l
n

2n

1J

k=0

V
Z-f

k=0

/2n +
I
k

\T
= 5nF
V2n+1
ij
/ (4m+2)k
* (2m+l) (2n+l) * 2

1971]

LUCAS GENERATING FUNCTIONS

133

2n

2 <-

1}n+kT
;

Tn

= T

^nn^k

^m+Dn

2m+l

k=0
2n

(?)
2

I V

(mml)

4mk "

+ L

f"^

4mk "

5nL

4mnF2m

5llF

2m(2n+l) F 2 n +

k=0

REFERENCES
19 V. E. Hoggatt, J r . , and Marjorie Bicknell, "Some New Fibonacci Ident i t i e s , " Fibonacci Quarterly, Vol. 2, No. 1, Feb. 1964, pp. 29-32.
2. V. E. Hoggatt, J r . , and Marjorie Bicknell, "Matrix Generation of Fibonacci Identities for F 2 uVTto appear, Fibonacci Quarterly.
3. V. E. Hoggatt, J r . ,

A New Angle on Pascal* s Triangle," Fibonacci

Quarterly, Vol. 6, No. 4, Oct. 1968, pp. 221-234.

4.

John Wessner, "Binomial Sums of Fibonacci P o w e r s , " Fibonacci Quarterly, Vol. 4, No. 4, Dec. 1966, pp. 355-358.

5.

H. L. Leonard, J r . , Fibonacci and Lucas Identities and Generating Functions, San Jose State College Master 1 s Thesis, 1969.

[Continued from page 120. ]

SOME FURTHER RESULTS
There are several other configurations which yield products of binomial
coefficients which are squares.

For instance, if two hexagons Hj and H2

have a common entry, then the ten terms obtained by omitting the common
entry have a product which is an integral square.

Thus, one can build up a

long serpentine configuration, or in fact build up snowflake curves.

Secondly, it should be noted in passing that all results above hold for
generalized binomial coefficient a r r a y s , in particular for the FIBONOMIAL
COEFFICIENTS.

ADVANCED PROBLEMS AND SOLUTIONS

Edited By
RAYMOND E.WHITNEY
Lock Haven State College, Lock Haven, Pennsylvania

Send all communications concerning Advanced Problems and Solutions

to Raymond E. Whitney, Mathematics Department, Lock Haven State College,
Lock Haven, Pennsylvania 17745.

This department especially welcomes

problems believed to be new or extending old results. Proposers should submit solutions or other information that will assist the editor.

To facilitate

their consideration, solutions should be submitted on separate signed sheets

within two months after publication of the problems.
H-181

Proposed by L Carlitz, Duke University, Durham, North Carolina.

Prove the identity

(am + en)

(bm + dn)

m,n=0

m n
mini

(1 - ax)(l - dy) - bcxy

where
^-(ax+by)^

H-182

y e -(cx4dy)

Proposed by S. Krishnar, Berthampur, India.

Prove o r disprove

(i)

m
V* A = o
k2
k=lK

(mod 2m + 1) ,

and.
134

Apr. 1971

ADVANCED PROBLEMS AND SOLUTIONS

135

m
(ii)

Y^
= 0
~ (2k - 1)'

(mod 2m + 1)

when 2m + 1 is prime and larger than 3.

[See Special Problem on page 216. ]
SOLUTIONS
GONE BUT NOT FORGOTTEN
H-102

Proposed by J. Arkin, Suffern, New York. (For convenience, the problem is restated, using
B =A

n m->

Find a closed expression for B

(H)

[I]

+ X

in the following recurrence relation.

= Bn " Bn-3 " Bn-4 " Bn-5

+ B

n-7

+ B

n-8

+ B

n-9 " B n-12

where n = 0, 1, 2 , - - and the first thirteen values of B0 through B12 a r e

1, 1, 2, 3, 5, 7, 10, 13, 18, 23, 30, 37, and 47, and [x] is the greatest
integer contained in x.
Solution by the Proposer.

In a recent paper* this author introduced a new notation, and because of

the new method in the paper, we a r e , for the first time, able to find explicit
formulas in such recurrence relations as H-102.
We denote by p (n) the number of partitions of n into parts not exceeding m,

(1)

where

F m (x) = 1/(1 - x)(l - x 2 ) .- (1 - x m ) = J2 Pm ( n ) x I 1

n=0

and p (0) = 1.
^m
The new notation we mentioned above is defined as follows:
*Joseph Arkin, "Researches on Partitions,'' Duke Mathematical Journal, Vol.
38, No. 3 (1970), pp. 304-409.

136

ADVANCED PROBLEMS AND SOLUTIONS

(2)

A(m,n) = 1 if
A(m,n) = 0

[Apr.

m divides n

if m does not divide n ,

whe re
m = 1, 2, 3, ,

n = 0, 1, 2, ,

and
A(m,0) = 1 .
Now, in (1), it is plain that
oo
F 2 (x)/(1 - x )<l - x )(l - x5) = J2 PsWx 11 ,
n=6
3

and we have

(3)

F2(x) = (1 - x 3 )(l - x 4 )(l -

X 5)

oo
J2 PsWx 11
n=0

Then, combining the coefficients in (3) leads to

(4)

p2(n) = p5(n) - p5(n - 3) - p5(n - 4) - p5(n - 5) + p5(n - 7)

+ P5 (n - 8) + p5(n - 9) - p5(n - 12) ,

and it is evident that the right side of (4) is identical to the right side of (H).
Now* it was shown that

^Joseph Arkin, "Researches on Partitions," Duke Mathematical Journal,

Vol. 38, No. 3 (1970), Eq. (6), p. 404.

1971]

ADVANCED PROBLEMS AND SOLUTIONS

137

p2(2u) = u + 1
and
p2(2u + 1) = u + 1

(u = 0, 1, 2, ) ,

so that
(5)

ft

(a) = [n/2] ,

where n = 0 , l , 2 , e " ,

and [x] is the greatest integer contained in x,

Then, combining (5) with the left side of (4) and since
B n = p5(n)

(n = 0, 1, 2, . ) ,

it remains to find an explicit formula for the p5(n).

To this end*, we see that
6n4 + 180n3 + 1860n2 + 7650n + 7719 1
(270n + 2025) ( - l ) n
PsW - j ^ 8 0

1920A(3,n)
2160(A(49n) + A(4,n + 3))

3456A(5,n)

A LARGE ORDER
H-161

Proposed by David Klarner, University of Alberta, Edmonton, Alberta, Canada.

Let

* Joseph Arkin, "Researches on Partitions," Duke Mathematical Journal,

Vol. 38, No, 3 (1970), Eq, (19), p. 406.

138

[Apr.

ADVANCED PROBLEMS AND SOLUTIONS

where the sum is extended over all compositions of n and the contribution to
the sum is 1 when there is only one part in the composition. Find an asymptotic estimate for b (n).
Solution by L Carlitz, Duke University, Durham, North Carolina.
Put

(ai + a A (*2 + a3 V I**-1 + % I

b, (n) = X
k

a1+.*H-ak=nV *2 I \ *3 )

*k

00

ra = E \(n) xn
n=0
It is known (see "A Binomial Identity Arising from a Sorting Problem,"
SIAM Review, Vol. 6 (1964), pp. 20-30), that fk(x) is equal to the following
determinant of order k + 1:
1

1
1

x
1

x
1

1
1

x
1

It follows that

W x > = ***> " xfn-l<x>

Since f0(x) = 1, fj(x) = 1 - x, we find that

F(z) = JT fk(x) z k = k=0

1 - xz
z + xz*

1971]

ADVANCED PROBLEMS AND SOLUTIONS

In the next place,

1 - xz

1 - z + xz

a2

j32

__

a - jS \1 - az

1 - j3z /

where
a + ft = 1,

aft = x

It follows that
ak+2
f

(x)

_ ^+2

a-

'

so that
00

n=0

- 0

Now, if k = 2r + 1,

4 ^ # = TT * ^ "

^ )

S=l

-ii- ,
-27Jls/k\
0 2iris/k\/
= I I (a - j3e
)(a - /3e
)
s=l

x cos
f / l1 -- 44x
= TTI'
sIf we put

^ ]

139

140

ADVANCED PROBLEMS AND SOLUTIONS

r
1
As

(3)

irfi-^co^f)" ^S

g = .,i .X

JL

_ 4x cos*

[Apr.

we find that

2(r-l) TTS
COS

TT f cos 2

0r-l

' -rr
T

- - cos 2

C0S

2(r-l) 7Ts
"F

TT I cos - ^ - cos

t^s

^-J

#s
^ 2 ( r - l ) 7TS

cos

7T(t + S)
T J sin B
Ll^i

-rk

sin

7T(t -

t=l
#s
But
2r
TT sin &

TT

t=l
#s

sin 2 *

+ 2> sin I L Z J O
k

.I,B-1 ^ 1

* _

sin sin 2 "

k

(-D^k
0

o 7TS

TTS

sin*2 - ^ c o s -

so that

(4)

k
sin* c o s 2 ^ 1
q8 1. 2
k
A = (( -- Il )) 8 "" 1
*,-

Then, by (2), and (3) and (4),

^
*

'

S)

1971]

ADVANCED PROBLEMS AND SOLUTIONS

141

,k+2l(k+1)
, s i n 2 , ^ *
^
a - P
_ 2 1 *'" V
, -vS-1 k + 2 w o k + 2
~k+2 __ A+2
k + 2 -2-/
2
A"
1 _ A
a
4x cos 4 k + 2
s=l
8

.k+2 ^ ( k + 1 )

00

S-l

7TS

V * /vi \ n

fS

s=l

IA\

2 n

n=0

oo
4(k+D
n
=
V (4x) V
(-1) 8 " 1 sin2 ^ s Cc0 oS s k + 2 n - ^
k + 2 2L/ * ' 21* l ;
k + 2
k + 2
n=0
s=l

^"S

k + 2

k+2

Therefore, by (1)9

/r\
(5)

k+2n+2 ^ ( k + 1 )
u / \
2
x-^ / .jxS-1 . o 7TS
k+2n TTS
b k (n) = k + 2
2J
(-D
sin2 j ^ - ^ cos
,g

/,
,,v
(k odd)

s=l
This implies the asymptotic formula

9 k+2n+2

b (n

<>

k> ~i m r

sin2

rfs

cos

Next, if k = 2r,

<* " P

02-/32

8=1

r-1 *
s=l *
If we put

1 - 4x cos 2

kTi

(k odd)

142

ADVANCED PROBL1

[Api

we get

cos2(r~2)
k
s

r-1 (

27TS

cos

27Tt \

cos

n \k - J
t=i

2r^2

f r

2(r-2) 7TS
k

27TS

cos2(r"2)
k

27Tt \

T . 7T(t +il>

j7^cos_.cos_j

pr sin-4.

sin_____^__

Since

77

ff

t=1

sin 2ft^j) sin 2__J--i = (-i)3"1

k

#s
(-1)

sin #Jt

k
t=l
sin ^ s i An _ s . inl x ^(r + s)
k~ " " ~TT
k

s-1

k
k

. o 7TS ^ o A"s

sin 4 - r - cos 4 -rk

k

it follows that

8 1

= (-I) "

2 k sin 4 c o s k " 2

ITS

k 7TS

Then

, .

fk

s=i

Q 4
_ 1 sin 42 1-r c o s i , o
Q
0 k+2
a - p
_ _2
<c-^ , vS-1
k +2
k + 2
k+2
J_+2 k + 2 _L, {"1}
.
2 7TS

"*

" 4 x c o s FTT

= FT* E ^ T ^
n=0

s=l

s 2

* FT* cosk+2n
k +

1971]

ADVANCED PROBLEMS AND SOLUTIONS

143

so that

m\

(7)

u /v,\ -

b (n)

9 k+2n+2
2

"

2K
,
V^ / i\s_l

k + 2 2^

^
2

(_1)

sm

fins

FT2 cos

. .
k+2n

TTS

,,

(k e v e n )

FT2

s=l
T h i s i m p l i e s the asymptotic r e s u l t
Q k+2n+2

/o\

i / \

(8)

b (n)

. 9

~ k +2

sm

7TS

FTT

cos

. ,0
k+2n

7T

/,

(k

FT^

even)

We m a y combine (5) and (7) in the single formula

k+2n+2

/^\

/ \

0)

bk(n) = -gy.

[Jfc+i>]
v^

-sS-1 . o

<->

sm2

TTS

FTT

cos

k+2n

s=l
and (6) and (8) in

/IA\
(10)

2js+2n+2
i / \
?
7T
b k (n) ~ - g - ^ sin* ^

LUCA-NACCI
//-73

Proposed by hi. H. Ferns, Victoria, B. C, Canada.

P r o v e the following i d e n t i t i e s :

k=l

k=l

k+2n
cos

77
,g

TTS

irn*

144

ADVANCED PROBLEMS AND SOLUTIONS

where F

and L

a r e the n

Fibonacci and n

[Apr.
Lucas numbers,

respectively.
Solution by A. G. Shannon, Mathematics Department, University of Papua and New Guinea,
Boroko, T.P.N.G.

E 22k " 2L k F k+3

i. s^i;

= L F

i 4 =3 -

k=l
and
2nF^+1 - 1 = 2 * F | - 1 = 3

Assume identity true for n. Then,

2 n

F U

n+1

k=l

k=l

= 22n(F^1

2n
+

L n + i F

( F n+ F n + 2 ) ( F n + 3

Fn+2))-l

= 2 (F 2 + 2F 2
+2FF
+F F
+F
F
) - l
4
u
n+l
n+2
n n+2
n n+1
n+1 n+2 ;
= -2 2 n (2F*
2 F n + 2n+1
Fx1))-l
N
J 0 (N
n+2 + Fn+2
= 22n+2F2
1
2
Fn+2 - 1
which proves the result.
2. It can be readily shown that
<3>

kFk+3 = FkLk+3

+ 4

<-1)k

1971]

ADVANCED PROBLEMS AND SOLUTIONS

145

by using
L k = ak +
and
F k = {ak - pk)(a - iS)"1 .
From (1) above, it follows that
n

<

4)

E *k~2 L k F k + s =

a2n

^ + 1 - pQ+1>1 - 5

k=l
With (3), the left-hand side of (4) becomes

k=l

k=l
n
_\ ^

= 5

1 /

0 2k-2

F L

k k+3

/Q2n+2

+ (2

ixn

,x

<- " 4 ) '

k=l
The right-hand side of (4) reduces to
2 2n ( a 2n + 2 +

^n+2

= (22nLn+1-l)

+ 2(_1)n)

(22n+2(-l)n-4)

and result (2) follows.

/1/M solved by M. Yoder, C. B. A Peck, J. Milsom, M. Hatch ford, D. V. Jaiswal, and the Proposer.

REGULAR POLYHEDRONS AND PASCAL'S TRIANGLE

J. WLGDARSKI
Porz-Westhoven, Federal Republic of Germany

It is known that any convex polyhedron has three parameters.

Numer-

ical values of parameters of all regular polyhedrons are shown below.

Polyhedron

Tetrahedron

2.

Hexahedron

12

3.

Octahedron

12

4.

Dodecahedron

12

20

30

5.

Icosahedron

20

12

30

1.

where F represents the number of faces, V the number of vertices, and E

the number of edges.
Numerical values of these parameters form a sequence:
4, 6, 8, 12, 20, 30 .
It is remarkable that the half-values of all members of this sequence
form two apexes of Pascal's triangle.
The first apex is situated just below the edge-series of ones and the
second one below the first apex.
Both apexes look like this:

146

SEQUENCES WITH A CHARACTERISTIC NUMBER

IRVING ADLER
North Bennington, Vermont

1. A Fibonacci sequence a09 a l9 a 2 , * , a , is called a Fibonacci

sequence if it satisfies the recursion relation

CO

(0

= a ,., + a

n+2

n+1

A well-known property of such a sequence is that there exists a number a

such that
a a - a 2 1 = {-if a
N
n n+2
n+1
'

(2)

for all n = 0, 1, 2 9 e 8 The number a is called the characteristic number

of the sequence [1]. The purpose of this paper is to explore the significance
of the characteristic number [2] and to identify all sequences that have a
characteristic number. We shall consider only sequences of rational numbers*
2 We call a sequence geometric if there exist numbers a and r such
that
(3)

= arn9

n = 09 1, 2,

089

..

If a sequence is geometric, then

(4)
v
'

a a ^ 0 - a^
= 09
J
n n+2
n+1

n = 09 1, 2 9
9

Conversely, suppose Eq. (4) holds 8 If a

/K\
(5)

n+2 _ a n + l
- ,
a
a
n+l
n

/ 0 for all n9 then

_ n - 9
n - 09 1, 2 9
147

148

SEQUENCES WITH A CHARACTERISTIC NUMBER

[Apr.

Then the sequence satisfies (3) with a = ao, and

r = .
a0
If a

= 0 for some n, then by Eq. (4), a

and a

- =0.

2, 3, .

Hence, if a

- = 0. If n > 2 , then by Eq. (4),

a - a2
= 0,
n-20 n
n-1
= 0 for some n,

then a

= 0r for all n = 1,

That i s , either every term of the sequence is 0, or only a0 is not

0. In the first case, the sequence satisfies Eq. (3) with a = 0, and r arbitrary.

In the second case, it satisfies Eq. (3) with a = a 0 ,

and r = 0.

Therefore, a sequence is geometric if and only if it satisfies Eq. (4).

tion (4) is a special case of Eq. (2) with d = 0.

Equa-

Since Eq. (2), with d f 0

represents a minor deviation from the typical behavior of a geometric sequence,

we shall call any sequence satisfying Eq. (2) with d f 0 a parageometric
sequence.
3. We shall call a sequence almost geometric if it is not geometric, but
there exist numbers r

such that

for n = 0, 1, 2, ,

the sequence (r ) approaches a limit as n becomes infinite.

For example,

in the Fibonacci sequence defined by

(6)

F 0 = 1,

Fi = 1,

Fn+2 = Fn+1 + Fn,

n = 0, 1, 2, . ,

the terms of the sequence are given by the Binet formula

(7)
Then

\=^-^-> ^^r^*
n

^ = H^
2

and

1971]

SEQUENCES WITH A CHARACTERISTIC NUMBER

rn

an+1 - f+1
B

0n
- (3

149

~ P (If
x

'(i)(-ir

,n+l

But
o?

W'

I i I (.yn+1

< 1

6 + 2N/5

Therefore,
= 0 ,

and

lim r

= ae So the Fibonacci sequence, defined by (6), which is para-

geometric with d = 1, is also almost geometric.

4. We shall call a sequence alternating if a , = a, a. - = b, a ^ b,
for all n = 0, 1, 2, e . An alternating sequence satisfies Eq. (2) with d =
a2 - b 2 . Then d = 0 if and only if b = -a.

So, an alternating sequence is

geometric if and only if b = - a , and it is parageometric in all other cases,

However, a parageometric alternating sequence is not almost geometric.
fact, if a = 0 and b f 0, then r
and b = 0, then r

cannot be defined for even n.

cannot be defined for odd n.

zero, then

b
=
n
a

for even n, and

*.-5
for odd n, and

In

If a / 0

If neither a nor b is

150

SEQUENCES WITH A CHARACTERISTIC NUMBER

a
b

so, while r

/ b
a

[Apr.

'

is defined for all n, it does not approach a limit as n becomes

infinite. Hence, every alternating sequence is not almost geometric.

5. We shall call a sequence eventually almost geometric if it is not a l most geometric, but the sequence obtained by deleting the first k t e r m s , for
some positive integer k, is almost geometric.
0, 1, 0, 1, 0, a 5 , a 6 , a?,

8#

-,

where a

For example, the sequence

= F

for m = 0, 1, 2, ,

is parageometric and is not almost geometric, but it is eventually almost geometric.

Similarly, the sequence 8, 5, 3, 2, 1, 1, 0, 1, 0, 1, 0, a^, a^,

, where a

-- = F

for m = 0, 1, 2, ,

is parageometric, is not

almost geometric, but it is eventually almost geometric.

We shall call a sequence eventually alternating if it is not alternating,
but the sequence obtained by deleting the first k t e r m s , for some positive
integer k, is alternating.
a&, , where

afi

For example, the sequence

8, 5, 3, 2, 1, 1, a 7 ,

is 0 for odd n, and is 1 for even n, is parageo-

metric, is not alternating, but is eventually alternating.

6. We can now state our principal result.
Theorem.

If a sequence is not geometric, and no term of the sequence

is 0, it is parageometric if and only if it satisfies the recursion relation

(8)

l0

n+2

= ka

+ a

n+1

for some rational number k. If k = 0, the sequence is alternating, and if

k f 0, the sequence is almost geometric.
A zero term may occur in the sequence only if the absolute value of its
characteristic number is a perfect square.

If there is a zero term in the

sequence, then either the sequence is alternating, or the sequence is eventually

alternating, or the sequence is eventually almost geometric. In the first case,
the sequence satisfies the recursion relation (8) with k = 0.

In the second

case, for some index i > 0, a0;, a l9 , aj is a fragment of an almost geometric sequence satisfying the recursion relation (8) for some k ^ 0,

and

1971

SEQUENCES WITH A CHARACTERISTIC NUMBER

151

i + l 5 a i+2' " " ' a i+n 5 " " i s a n alternating sequence satisfying (8) with k =
0. In the third case, there are two possibilities: (1) For some index j > 0,
ao, a l 5 , , c , a. is a fragment of an alternating sequence satisfying the r e c u r sion relation (8) with k = 0, and a. ,-, a. i n , , a., , is an almost
l+l i+2
j+n
geometric sequence satisfying (8) for some k f 0. (2) For some non-negative
index i, a 0 , aj, , a. is a fragment of an almost geometric sequence satisfying the recursion relation (8) for some k f 0; for some positive index
j > i , a., a. - , , a. is a fragment of an alternating sequence satisfying
the recursion relation (8) with k = 0; and a . , - , a . , 0 , 8 , a., > is an
j+i

j+z

j+n

almost geometric sequence satisfying (8) for some k ^ 0.

Consequently, a

parageometric sequence consists of at most three consecutive segments each

of which satisfies the recursion relation (8) for some value of k.
Proof.

(1) Let (a ) be a sequence that is not geometric and with a

0 for all n = 0, 1, 2, . If it is parageometric, we have

a a l 0 - a2 ,- = (-1) d .
n n+2
n+1
Then

n+l a n+3 " a n+2

Therefore,

V n + 2 " a n+l

V l a n + 3 " <+2

n)

Hence
a

Then, since a

n+l(an+3 - V l '

n+2 ( a n + 2 "

^ 0 for all n,
a

n+3 " a n+l

a

n+2

n+2
a

n+l

152

SEQUENCES WITH A CHARACTERISTIC NUMBER

[Apr.

Thus
a

l0
n+2
a

- a

n _ .

n+l

for some rational constant k and all values of n = 0, 1, 2, .

'
'
satisfies Eq. (8).
Conversely, suppose the sequence satisfies (8). Then
a

n+3 " V l
_.
a
n+2

n+2 " a n

.
= ^

Then (a )
n

n+l

Consequently,
a

n-H a n+3 " a n+2

for all n = 0, 1, 2, ' .

" ( a n a n+2 "

n+l)

If we let d = aoa2 - a|, then we have

V n + 2 ~ an+l = (-1)Dd *

Since the sequence is not geometric, d f- 0, and the sequence is parageometriCo

If k = 0,

-. Hence it is alternating.

f a

then a

= a .

Since the sequence is not geometric,

The characteristic equation associated

with (8) is

(9)

x2 - kx - 1 = 0 ,

whose roots are

/inv

(10)

2
k +^k
r =
5 + 4

s = k -

Then, by the theory of linear recurrence relations [3],

*# 7 1

1971]

153

SEQUENCES WITH A CHARACTERISTIC NUMBER

(11)

= ar

+ bs

w h e r e a and b have the v a l u e s

(12)

a =

b =

^/k2 + 4

a0r - ai

\P"+ 4

d = a 0 a 2 - ai = (a + b)(ar 2 + b s 2 ) - (ar + bs) 2 = a b ( r - s) 2 = ab(k 2 + 4)

Since d f 0,

it follows t h a t a / 0 and b / 0.

If k > 0,

< 1

If k <

0,

< 1

n+1
a

n+1 , , n+1
ar
+ bs
ar

+ bs

r +

b J s\n

1 +

a\rj

|r\
b\s/

+ s
+ 1

If k > 0,

< 1 ,

and

lim r = r.
n -*oo n

If k < 0 9

< 1 ,

and

l i m r = s. Consequently, if k f 0, the sequence i s a l m o s t g e o m e t r i c .

n * oo n

154

SEQUENCES WITH A CHARACTERISTIC NUMBER

[Apr.

(2) If some term a, = 0, then

a

kV2

4 + 1 = (-1)Ild >

and hence
- a | + 1 . = (-l) k d .
If k is odd, d is a perfect square. If k is even, -d is a perfect square.
Since d f 0, a - f 0. If k f 0, we have
a

k-iak+i - 4

-1)n_ld

or

\-iVi =

( 1)k_ld

Then
a

k-lak+l

4+1 '

and
a

k + l (a k-l " V l *

Then, since a, - ^ 0 , a, - = a, y

That i s , every zero term iin the

sequence is flanked by a pair of equal non-zero terms. Consequently, if a =

= 0, with k < m , then m - k > 1. If a, = 0, it is possible that

a
a

k+2

'

anc

-2

** ^ exists.

Then a,

belongs to a sequence of

\+2m

alternate zero terms

k-2 " a k-2+2

"

k-2

where I > 0, 2 < k, and m > 0.

k+2 " ' "

'

1971]
(

SEQUENCES WITH A CHARACTERISTIC NUMBER

\ - -2--1>

k-2+l

\-l

"

\+l

155

k+2m+l

where the parentheses around the term \_2n_1

indicate that it is included
a
J
only if it exists. (That i s , if k - 2 ^ 0.) Then \_2i_^
"5
k_2
a
k+2m+l ? w n i c n i s a segment of the sequence (a ), is an alternating s e quence with zero terms alternating with non-zero terms. Let us extend this
alternating sequence as far as we can to both lower and higher indices by including a, lQ l 0 and a. ,_ i 0 if a, _ i n = 0, and by including a. nn 0
&
J
&
k+2m+2
k+2m+3
k+2m+2
k-2-2
and \_90o
if they exist and a, 9
= 0. Then the following four possibilities a r i s e , depending on whether or not the alternating sequence begins with
a0 on the left and whether or not it terminates on the right:
I. The alternating sequence begins with a 0 , and does not terminate.
II. The alternating sequence begins with a.,

i > 0,

and does not

terminate.
III. The alternating sequence begins with a 0 ,

and terminates with a.,

j > 0.
IV. The alternating sequence begins with a., i > 0, and terminates with
a., i > L
J
In case I, the sequence (a ) is an alternating sequence, with either the
odd-numbered terms or the even-numbered terms equal to zero. That i s , it
has the form 0, a, 0, a, 0, a, or a, 0, a, 0, a, 0, , where a f 0.
Such a sequence satisfies the recursion relation (8) with k = 0.
In case H, a. f 0, a. - = 0, and a. - ^ 0. The infinite sequence
a., a. - ,

is an alternating sequence of the form

a, 0, a, 0, .

We

shall show that for every n < i, a ^ 0.

I n c a s e HI, a. f 0, a. 1- = 0, and a.,- 1 f 0. The finite sequence a 0 ,
J
JJ+
aA, , a. has the form 0, a, 0, a, , 05 a or a, 0, a, 0, , 0, a.
We shall show that for every n > j , a n f. 0.
In case IV, a. ^ 0, a. + 1 = 0, a . ^ 7^ 0, a ^ 0, a
= 0, a
^ 0.
The finite sequence a., , a^ has the form
I

a, 0, a, 0, , 0, a.

We

shall show that for every n < i and every n > j , a n f 0.

Suppose a. ^ 0, a. - = 0, a., - ^ 0 (cases IE and IV).

We shall call

these assumptions Assumptions A. We show that for every n > j , a n ^ 0.

From

156

SEQUENCES WITH A CHARACTERISTIC NUMBER

a. - a.

t1

[Apr.

- a? = (-1) J ~ d

we get a? = (-l)-'d .

(13)

a . a . + 2 - a j + 1 = (-1)1 d = a j .

Therefore,
a

. la . lin , a . )
J J+2
J

We c o n s i d e r f i r s t the c a s e w h e r e -a. > 0.

3
0, and a.^ Q2 > a. > 0.
3+
3
(14)

aj+1aj+3 - a ^

a.,..
3+1
T h e n , s i n c e a . , - ^ 0, a. l 0 - a. >
J+l
3+2
j

= (-l)^d

= -aj

Therefore,
a

Then a . + 3

a
--ii
j + l -_LQ
j+3

a ?j+2
, 0 - a?j > 0 .

i s not z e r o , and h a s the s a m e sign a s a. - .

F r o m (13) and (14),

a . a . i 0 - a?,- + a . , - a . l 0 - a ? l 0 = 0 .
j j+2
j+l
j + l j+3
j+2
Therefore,

V2V2-V

ViV'V^-

Hence a . l 0 - a . , - h a s the s a m e sign

a s a . - and a . l 0 .
5
j+3
j+l
j+l
j+3

T h i s , if a . ( 1 > 0,
j+l

' h+sMVil

j+3

0.

> a

j+r

andif

j+i

<

' aj+3 < a j+r

Now we p r o c e e d by induction.

a. > 0,
that

that a . + 2 n + 1 , a . + 2 f c + 1 ,

1+2

fc

either case

Assume that
k-3' " '

i+l ^ ^

. > a - . o ^ o
the

s a m e

V2k.ll> IV2k.ll> > | V l | > 0

>

sign

>

'''
'

>
and

1971]

SEQUENCES WITH A CHARACTERISTIC NUMBER

(15)

<16)

a
j+2k j+2k+2
a

j+2k-l

" aj+2k+l

j+2k+l

j+2k a j+2k+2 " a j + 2 k + l

j+2k

(a

j+2k = (-Dj+2k"1d

j+2k+2 " a j + 2 k )

Then, since a j + 2 k + 1 > a j + 2 k _ 1 ,

= < ~ 1 ) j + 2 k d = <- X > Jd =

+ a

j+2k-laj+2j+l

j+2k+l

<17>

j+2k+l

>

j+2k+2

j+2k+3 -

(a

= - ]
" a j+2k

j+2k+l " a j + 2 k - l )

and a . + 2 k + 1 = a . + 2 k _ 1

and a . + 2 k > 0, a . + 2 k + 2 - a j + 2 k >

157

'

have the s a m e s i g n ,

0 , and

>

j+2k

"

>

3 > -

]+2k+2 = (-1>j+2k+ld

= <-1>J+ld = " a l

Therefore,

Therefore,

j+2k+laj+2k+3

a. . - and a.

?,

j+2k+2 "

>

have the s a m e sign.

F r o m (15) and (17),

we g e t

j+2k a j+2k+2 " a j + 2 k + l

+ a

j + 2 k + l a j + 2 k + 3 " a j+2k+2 "

'

Then

Therefore,

j+2k+l

(a

j+2k+3 " a j + 2 k + l )

j+2k+2 ( a j+2k+2

a . + 2 n + 3 - a 1 + 2 n + 1 h a s the s a m e sign a s

j+2k+3 9

have the s a m e s i g n , and

j+2k+l'

""

j+l

j+2k)

4+2k+le

Hence

>

158

SEQUENCES WITH A CHARACTERISTIC NUMBER

\^+2k+B\>

>

|V2k+l|

>

"-

|aJ+l|

>

[Apr.

If a. < 0, a similar argument shows that

a

j+2k

<

<

j+2k-2

'"

<

<

'

j+2k+l'

j+2k-l'

""'

j+l

have the same sign, and

|W + i| > I V2k-i| > - > | V i | > Hence, for every n > j , a

Suppose i > 0,

^ 0.

a. f 0,

a. - = 0,

a. - ^ 0 (cases II and IV).

We

shall call these assumptions Assumptions B. Because of the symmetry with

respect to i of the indices in the equation
a

i-lai+l " ai

(- 1 > i " 1(i

(-D i + 1 d ,

and because Assumptions A are symmetrical to Assumptions B with respect to

i if we write i instead of j in Assumptions A, the argument above proceeds
just as well in the direction of decreasing indices.
a

Hence, for every n < i,

f 0. Then by (1), in cases II and IV, the sequence a 0 , , aj_2. satisfies

Eq. (8) for some k f 0, and is a finite segment of an almost geometric s e quence; and in cases III and IV, the sequence a. -, a.

2,

, a

, sat-

isfies Eq. (8) for some k f 0, and is an almost geometric sequence.

This

completes the proof of the theorem.

An example of case IV is given in Section 5.

Another example is the

sequence
58, 24, 10, 4, 2, 0, 2, 0, 2, 0, 2, 8, 34, 144,

In this sequence, the characteristic number a = 4. The sequence is made up

of three consecutive segments:
I.
58, 24, 10, 4;
H.
2, 0, 2, 0, 2, 0, 2;
nr

8, 34, 144, ;

1971]

SEQUENCES WITH A CHARACTERISTIC NUMBER

159

where Segment I is a fragment of an almost geometric sequence satisfying the

recurrence relation a

= -2a

- + a . Segment His a fragment of an alter-

nating sequence satisfying the recurrence relation a

Q = a ; Segment III is
n~r^
n
an almost geometric sequence satisfying the recurrence relation a
= 4a
+a .
n

7. Consider the set of all almost geometric sequences satisfying the

recurrence relation (8) with given k f 0. The associated characteristic equation is (9), where roots are r and s given in (10). If r and s are irrational, the theory of these sequences is analogous to that of rational Fibonacci
sequences.

For example, just as the set of all rational Fibonacci sequences

can be given a field structure isomorphic to the field extension R(a) (see [4]),
the set of all rational sequences satisfying the recurrence relation (8) with
given k / 0 such that r is irrational can be given a field structure isomorphic to the field extension R(r). In fact, we may represent each such sequence
a 0 , a l s by the ordered pair (a0, a j ) ,

since the sequence is fully deter-

mined by its first two terms and the recurrence relation (8).

Then (a0, aj)

a0 + a*r is an isomorphism if we define addition and multiplication of sequences

by
(a0, aA) + (b0, bj) = (a0 + b 0 , aj + bt) .
(a0, ai)(b 0 , bi) = (a0b0 + a ^ , a ^ + a^o + k a ^ ) .

3. If

a1+

X
+

1_
a3

is a continued fraction, the convergents

p
n
=

160

SEQUENCES WITH A CHARACTERISTIC NUMBER

for n = 1, 2, 3, are given by p 0 = 1,

q0 = 0,

pA = a l9

[Apr.
qt = 1,

and

+ p
n = Wl
n - 2 ' q n = a n V l + q n - 2 for n > 1 [5]. If we let H = a2 =
= k / 0, where k is rational, then the equations take the form p 0 = 1,
q0 = 0, P l = k, q4 = 1, and p n = kp jQ _ 1 + p n _ 2 ? q n = k q ^ + q n _ 2 for
n > 1. Moreover, qt = p 0 , and q2 = kqA + a0 = k = p 4 . Hence, for all
n > 0, q =
Then
n
Pn-r

n _

Pn-1

lim C = r , where r is a root of x2 - kx - 1

noo n
0. Moreover, the relation

for n > 0. In this case,

i+2Vl "Pi+lV2

in this case takes the form

PiP i + 2 " Pi + 1 = <"1>i = (" 1 ) i d

where d = 1. Hence the sequence p l 9 p 2 , , p n , is a parageometric
sequence with characteristic number 1,

and is also an almost geometric

sequence satisfying the recursion relation p

= kp

- + p . If k is a pos-

itive integer, the sequence is related to the golden-type rectangle [6],

9. Every sequence that has a characteristic number d is either geometric (with d = 0) or parageometric (with d f 0). If it is parageometric,
it consists of at most three consecutive segments, each of which satisfies the
recursion relation (8) for some value of k .

If it is a geometric sequence

(ar ), and r f 0, it satisfies the recursion relation (8) with k = r = 1/r.

If r = 0, the sequence is a, 0, 0, , and is composed of two consecutive
segments a and 0, 0, , each of which trivially satisfies Eq. (8). Hence,
every sequence that satisfies Eq. (2) and therefore has a characteristic number

1971]

SEQUENCES WITH A CHARACTERISTIC NUMBER

161

a consists of at most three consecutive segments each of which satisfies Eq.

(8) for some value of k
Let us now consider any sequence satisfying Eq. (8), to see if it also
satisfies Eq. (2) and hence has a characteristic number.

If the sequence is

geometric, it satisfies Eq. (2) with d = 0. If the sequence is not geometric,

and no term of the sequence is 0, we have already shown in Section 6 that it
satisfies Eq. (2) with d f 0. Suppose now that the sequence is not geometric
and contains a term a. = 0. Then the method of proof used in Section 6 breaks
down. However, this case can be covered by a general proof that does not r e quire that all terms of the sequence be different from 0.
Let aQ, a l9
value of K.

8,a

, a n _ i , be a sequence satisfying Eq. (8) for some

Let d = a0a2 - aj.

Then, for n = 0,

Eq. (2). We now proceed by induction.

the sequence satisfies

Assume

a a l 0 - a2 ,- = N(-1) d
n n+2
n+1
for some fixed n.
a

n+lan+3

n+2

= a , n (ka ^
n+1 n+2

n+1

- aa2
n+2

2
+ aa 2
n+lan+2 - a
n+l
n+2
= a n + 2 ( k a n + l - an+2> + a n + l

ka

n+2 ( " a n )

n+l

-(anan+2-an+l>

= (-l)n+1d.
Hence , every sequence satisfying Eq. (8) also satisfies Eq. (2), and therefore
has_a characteristic number.
REFERENCES
Brother U. Alfred, "On the Ordering of the Fibonacci Sequence," Fibonacci
Quarterly, Vol. 1, No. 4, Dec. 196 3, pp. 43-46.
Donna B. May, "On a Characterization of the Fibonacci Sequence," Fibonacci Quarterly, Vol. 6, No. 5, Nov. 1968, pp. 11-14.

SEQUENCES WITH A CHARACTERISTIC NUMBER

162

A p r . 1971

3.

J a m e s A. J e s k e , " L i n e a r R e c u r r e n c e R e l a t i o n s , P a r t 1,M Fibonacci Q u a r -

4.

Eugene L e v i n e , " F i b o n a c c i Sequences with Identical C h a r a c t e r i s t i c V a l u e s , "

t e r l y , Vol. 1, No. 2 , April 1963, pp. 69=74.

Fibonacci Q u a r t e r l y , Vol. 6, No. 5, Nov. 1968, pp. 75-80.
5.

C. D. O l d s , Continued F r a c t i o n s , 1963, Random H o u s e , pp. 19-28.

6.

J o s e p h A. R a a b , "A G e n e r a l i z a t i o n of the Connection between the Fibonacci

Sequence and Pa seal 1 s T r i a n g l e , " Fibonacci Q u a r t e r l y , Vol. 1, No. 3 , Oct.
1963, pp. 2 1 - 3 1 .

7.

C h a r l e s H. K i n g , "Conjugate S e q u e n c e s , " Fibonacci Q u a r t e r l y ,

Vol.

6,

No. 1, pp. 4 6 - 4 9 .
[Continued from page 119. ]
-3

-1
= ^ ( n + l)(n + 2)
0 - 1
0

3 - 3
0

REFERENCES
1.

Dov J a r d e n , R e c u r r i n g S e q u e n c e s , Published by Riveon L e m a t i m a t i k a ,

J e r u s a l e m ( I s r a e l ) , 1966.

2.

E. L u c a s , " T h e o r i e d e s F o n c t i o n s N u m e r i q u e s S i m p l e m e n t P e r i o d i q u e s , "
A m e r . J. of Math. , 1 (1878), pp. 184-240 and 2 8 9 - 3 2 1 .

3.

T h o m a s M u i r , The T h e o r y of D e t e r m i n a n t s (4 V o l s . ) , D o v e r , New Y o r k ,
1960.

4.

R. F . T o r r e t t o and J . A. F u c h s , " G e n e r a l i z e d Binomial

Coefficients,"

Fibonacci Q u a r t e r l y , 2 (1964), pp. 296-302.

5.

P r o b l e m B - 1 3 , Fibonacci Q u a r t e r l y ,

1 (1963), No. 2 , p. 86; Solution 1

(1963), No. 4 , p. 79.

6.

P r o b l e m B - 4 6 , Fibonacci Q u a r t e r l y 2 (1964), p. )2; Solution 3 (1965),

pp. 76-77.

A NEW ANTHESIS
JOSEPH P. MUtMZENRIDER
Warner Robbins, Georgia

Louis Pasteur pursued and assessed his studies in the light of his belief that:
" . . . t h e r e is a cosmic dissymmetric influence which presides
constantly and naturally over the molecular organization of principles
[sic] immediately essential to life; and that, in consequence of this,
the species of the three kingdoms, by their structure, by their form,
by the disposition of their tissues, have a definite relation to the
movements of the universe. M [1]
Fathoming the occurrence of a specific angle of dynamic orientation
throughout a range of pehnomena involving gravitation andelectromagnetism:
stellar and atomic systems and living molecules, may fulfill such belief that
there is unitary, fundamental interdependency (relativity) of each such system one upon the others, mutually generated of cosmic necessity.
In 1783, Herschel found in the constellation Hercules the point among
Earth's neighborhood of "fixed" stars toward which the Solar System moves,
the Solar Apex. The pole of the ecliptic, known from ancient times, is in
the constellation Draco. Though the individual members of the Solar System
a r e variously oriented, the system as a whole spirals toward Hercules along
a trajectory inclined about 37 from the pole of the ecliptic [2].
Galactocentrically, that is of great interest.

The plane of the ecliptic

intersects the plane of the Milky Way, our Galaxy, at points in Sagittarius
and Gemini. Beyond Sagittarius lie the mass center and dynamic foci of the
Galaxy. Thus, the Solar System plane is about perpendicular to the galectic
center, while the axis of the Solar System is inclined to the plane of the
Galaxy. As the Solar System spirals along its galactic orbit, towardCephaus,
it would seem that we would rise out of the Galaxy, but, the fact is that the
Galaxy itself moves overall at an inclination to its plane, at an angle that

may be determinable with respect to our neighbors in the local cluster of

163

164

ANEWANTHESIS

[Apr.

galaxies. It may also be determinable how galactic clusters in general are

oriented: their members among themselves; one cluster to others [3].
The axis of the overall motion of the Solar System is of additional interest with respect to the "abandoned" theory of the JEther.

Adolph Griinbaum

[4] argues that acceptance or rejection of the AEther depends substantially

upon oneTs philosophical and historical comprehension of the issue, as well as
upon scientific criteria. Arthur Moestler [5] reviews the history of the r e jection of the JEther and concludes that contemporary scientists have glossed
the issue. On theoretical grounds, P. A. M. Dirac [6] has reconsidered and
found a reconceived JEther necessary.
Somewhat like Laplace not needing the "hypothesis" of God, one may
say, Einstein [7] early argued that Relativity Theory eliminated the need for
JEther postulates, experiments, and interpretations.

Firm in his belief that

God does not cast dice, however, by 1952, Einstein wrote:

" . . . the foundation of electromagnetic theory taught that a particular inertial system must be given preference, namely, that of the
luminiferous aether at r e s t . . .
"Since the special theory of relativity revealed the physical equivalence of all inertial systems, it proved the untenabilityof the hypothesis
of an aether at r e s t . . .
"It appears therefore more natural to think of physical reality as
a four-dimensional existence, instead of, as hitherto, the evolution of
a three-dimensional existence.
"This rigid four-dimensional space of the special theory of relativity is to some extent a four-dimensional analogue of H. A. Lorentz's
rigid three-dimensional aether. " [8]
Inquiry as to the results of the Michaelson-Morely experimental program would, then, seem as legitimate as it is interesting.

As summarized

by Robert W. Wood:
"The most exhaustive series of observations extending over a
period of thirty years have been made by D. C. M i l l e r . . . " He "com-

1971]

A NEW ANTHESIS

165

puted the velocity and direction of the earth 1 s absolute motion in space,
on the assumption that the observed effects were real.

Astronomical

observations indicate that the solar system is moving with a velocity

of 19 kms/sec with respect to the brighter stars toward the c o n s t e l lation of Hercules* Miller 1 s results showed an absolute motion in the
opposite direction of 208 k m s / s e c . . . .

A very full and convincing

account of these laborious observations and calculations will be found

in Review of Modern Physics, Vol. 5, No. 3, July, 1933." [9]
Herbert Dingle, a member of the British solar eclipse expeditions of
1927, 1932, and 1940, to test Einstein 1 s prediction of the bending of light
rays passing through the gravitational field of the Sun, has long questioned
the Special Theory of Relativity and all its consequents, arguing that there
is a serious e r r o r at the root of Einstein* s mathematical reasoning.

The

history of science is largely a human story of accurate results obtained in

terms of inadequate theory. Dingle insists that, some time, however inconvenient it may be, the inadequacies of Special Relativity must be faced [10].
The e i t h e r is still considerable.
fruits.

But a concept is best judged by its

Beginning in 1925, in terms of his revamping of the ^Ether, Carl F.

Krafft [11] discovered what is fully described by the title of his first, p r i vately published monograph of 1927, Spiral Molecular Structures the Basis
of Life, which is replete with diagrams.

In that and subsequent writings,

which Mendelianly remain unrecognized, Krafft fully developed a theory of

helical molecular structure for proteins, with full understanding of the genetic import of his discovery, and much more.
The years 19251927 were those when Schroedinger,

Meisenberg,

Born, Jordan, Wigner, Pauli, F e r m i , Dirac, de Broglie, Base, Einstein,

et al, were developing the fundaments of quantum mechanics.
In 1948, Linus Pauling [12] discovered the base of contemporary
knowledge of helical molecular structure of proteins, in terms of which
F. H. C. Crick and J. D. Watson discovered the helical molecular structure
of Deoxyribose Nucleic Acid (DNA). Presenting their theory and structure,
Crick and Watson wrote:

166

A NEW ANTHESIS

[Apr.

"We have assumed an angle of 36 between adjacent residues in

the same c h a i n . . . .

The structure is an open one, and its water con-

tent is rather high. At lower water contents we would expect the bases
to tilt so that the structure could become more compact. " [13]
That assumed 36 molecule of the DNA double helix, articulated with
respect to the molecular axis, to which the purine and pyrimidine bases of
the genetic code are perpendicular and planar.
In 1927, Krafft theorized that the spiral structure of proteins provided
an explanation of optical activity. This is as yet unsettled.

Discussing the

"Origin and Role of Optical Isomery in Life," A. S. Garay states:

"Living organisms possess only one of two possible optical isom e r s . There is no generally accepted theory for the origin of this
asymmetry. " [14]
It i s , however, generally thought that polarized light is necessary to
the origin, development, and maintenance of life.

Garey asks: "What is the

source of circularly polarized light in nature?" A. Dauvillier notes:

"Rectilinear polarized light exists in the solar light diffused by
the sky and is produced in nature by reflexion, at an incidence of 37,
from the surface of water or on flat crystalline facets.
the sky is not polarized elliptically.

The light from

Circularly polarized light, which

is obtained by causing rectilinear polarized light to fall on a quarterwave plate or a Fresnel parallelepiped, may also be produced naturally by polarized light falling on a birefringent quartz or apar crystal."
[15]
In the molecule of water, the hydrogens are bonded at an angle of 104.6
across the pole of the oxygen. Regarded in an upsidedown, Alice sort of way,
one may say that the hydrogens a r e oriented to oxygen at 180 2 x 37.7
[16].
Investigating "Rotary Brownian Movement. The Shape of Protein Molecules as Determined from Viscosity and Double Refraction of Flow," John T.
EdsaLl wrote:

1971]

ANEWANTHESIS

167

"All these measurements involve the rotation and partial orientation of protein molecules in an external field of f o r c e . . . the orientation achieved is only partial, since it is opposed by the disorienting
action of the Brownian movement of the molecules... involving rotation of the molecules about their axes, arising from thermal agitation.
Its effect is to produce a purely random distribution of molecular o r ientations, in the absence of external orienting forces. In the presence
of such orienting forces, a steady state is gradually achieved, a state
intermediate between the two limiting conditions of complete orientation and of complete disorder. The exact character of this intermediate
state depends on the magnitude of the orienting forces relative to that
of the rotary Brownian movement

[17]

In terms of the rationale then presented, the experiment performed on

various proteins and other substances involved having a fixed core within a
concentric tube rotatable to impart motion to a solution contained in the tube,
through which polarized light is passed, its behavior being measured.

For

at least myosin, the protein of muscle, the angular parameters emerged as

53 and/or 37.
(Edsall, in another discussion, makes the only reference to the work
of Carl F. Krafft which I have yet found [18].)
Gunther S. Stant, both historian and practitioner of molecular biology,
discussing the seemingly startling permanence of the genetic code over geologic time, ventures the possible explanation that:

. . . there exists some as yet unfathomed geometrical or stereo-

chemical relation between the anticodon nucleotids triplet and the amino
acid which it represents.

Indeed, if such a relation exists, it would be

bound to hold one of the keys to understanding the origin of life. M [19]

In "Toward aDefinition of Mind," citing the work of D. L. Reiser [20],

Harold Kelman wrote this beautiful passage:

168

ANEWANTHESIS
n

[Apr.

The forming process is metaphorically a spiral, constituted of

an intimately connected sequence of levels or a continuum of transformations with movement possible from depth to surface and vice versa.
The helix is of crucial import in Indian cosmology.

Nature moves in

a helical pattern in time, so that spiral forms get ingrained at many

levels....

This is all part of a galactic rotation in which a Cosmic

field plays an important part in transmitting spin (angular momentum)

to m a t t e r . ' This is one facet of Reiser f s concept of cosmic imagination which moves in similar directions to my ideas on cosmic minding.
"X-ray crystellography reveals DNA as a double stranded alpha
helix [ s i c ] . . . .

It directs protein synthesis and heredity.

What was

intuited thousands of years ago regarding life and living is being confirmed by science or science confronts us with ancient truth." [21]
In terms of the evidence marshaled for the universal occurrence of an
angle of about 37?^-38, it seems that now and henceforth, it shall be less
metaphorical to make such assertions.
A. N. Whitehead wrote:
?T

.. .the search for a reason is always the search for an actual

fact which is the vehicle of reason. The antological principle... , constitutes the first step in the description of the universe as a solidarity
of many actual entities. M [22]
and:
"The task of reason is to fathom the deeper depths of the manysidedness of things. We must not expect simple answers to far-reaching
questions. fT

[23]

A fact is not an answer, but should be of service in approaching an a n s wer, however Zenoic the process of approach may ultimately be.
Musing in Autumn on "The Secret of Life," Loran Eiseley says:
" . . . I have come to suspect that the mystery may just as well be
solved in a carved and intricate seed case out of which the life has
flown, as in the seed itself. " [24]

1971]

ANEWANTHESIS

169

Natural Philosophy, as contemporary as it is ancient and honorable, is

especially pertinent in this two-cultured era, when I am "credentially" a
"litterateur.

As such, however, I am particularly a student of the

Phil-

osophy of Owen Barfield, whose preface compelled my attention to E. GrantWatson7 s exposition [25] of many beautifully amazing aspects of structure
and behavior of living creatures.

This was my introduction to the botanical

principle of Phyllotaxis; the aesthetic principle of the Golden Section; and to

the significance of the Fibonacci Series.
1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144, ,
or
1/1, 1/2, 2 / 3 , 3/5, 5/8, 8/13, 13/21, 21/34, 34/55, 55/89, 89/144,* .
These fractions,

1/2 = 0.5; 2/3 = 0.66; 3/5 = 0.6; successively vacillate

until a/b = 0.6180 .

This Fibonacci Number, 0.6180- , is the value of the Sine of 38.166.
"Elected Silence, sing to me
And beat upon my whorled e a r . . . .
Be shelled, eyes, with double dark
And find the uncreated light... . f !
G. M. Hopkins, S. J. , "The Habit of Perfection. "
Having discovered the equiangular, logarithmic spiral, which is the
shape of the shell of the Chambered Nautilus; whose equation is satisfied by
the Fibonacci Number, Jacob Bernoulli (1654-1705) had this figure inscribed
on his tombstone, with the inscription:
"Eadem mutata resurgo (Though changed I shall rise the same)."
[26]
The French poet, Paul Vallery, having caught Einstein's allusion to a
"mollusc" of reference [27], essaying on "Man and the Sea Shell," wrote:

170

ANEWANTHESIS
"Without the slightest effort life creates a very
relativity....

[Apr.
T

generalizedT

"It does not separate its geometry from its physics but endows
each species with all the axioms and more or less 'differential 1 invariants it needs to maintain a satisfactory harmony between the individual
and the world around i t . . . .
"The pattern of the colored furrows or bands that curve around the
shell, and of the bands that intersect them, reminds us of T geodesic
lines' and suggests the existence of some sort of f field of forceT which
we are unable to discern, but whose action would give the growth of the
shell the irresistible torsion and rhythmic progress we observe in the
finished product. " [28]
A line segment is a Golden Section if it may be split A.. .B
such that AB/BC = BC/AC.

Turning the segment AB perpendicular to BC and completing the

rectangle produces the Golden Rectangle.
A Golden Rectangle may be constructed geometrically by taking any
line as the base of a square; drawing the square; bisecting the square; drawing a diagonal in one of the created interior rectangles; using that diagonal as
a radius; swinging that radius until it intersects a line extended from the
original base; from which point, completing a new rectangle containing the
original square.
The Golden Section, and forms based thereon, occur throughout a r t i s tic endeavors from the Greeks through the Renaissance through Bartok T s
music [29]. A beautiful and extraordinarily meaningful example of a Golden
Rectangle is this graph by Linus Pauling:
100
80
f

60

40
20

20

40

60

80

100 120 140 160

"Fig. 8. A curve of proton number Z as a function of neutron

number N, calculated as described in the text. The horizontal lines
show the ranges of stable isotopes for alternate Z-even elements (for
large Z the four most stable isotopes). " [30]

1971]

ANEWANTHESIS
The value of the Tangent of 38.1+ is 0.7854,

171
which is the value of

(pi)/4, which is generated by the Gregory/Leibniz Series: [31]

1 - 1/3 + 1/5 - 1/7 + 1/9 .
Assessing the mass ratios of electron, positron, proton, neutron, the
mesons and hyperons, etc., John J. Grebe found "A Periodic Table for Fundamental Particles, M a s :
"The existence of unique relations among the fundamental particles based on exponentials of (pi)/4 has been discovered in the search
for symmetry, unity, and simple structure. n

[32]

Grebe has also employed the logarithmic spiral ingeniously in plotting

a three-dimensional Chambered Nautilus of a graph ordering all things in
terms of frequencies of cycles per nanosecond through cycles per

eons

[33].
Hermann Weyl, whose discussion [34] of the significance of the Fibonacci Series eases one's mind in daring such sweeping claims as are being
presented herein, has stated:
?t

Tl

Perhaps the philosophically most relevant feature of modern

science is the emergence of abstract symbolic structures as the hard

core of objectivity behind as Eddington put it the colorful tale of
the subjective storyteller m i n d . . . .

In the progress of science such

elementary structures as roughly correspond to obvious facts are often

later recognized as founded on structures of a deeper level, and in this
reduction the limits of their validity are revealed. M [35]
The objection that the Fibonacci Number and its associated Golden Angle
at best only approximate to established parameters in astronomy, from which
inference about gravitation haaybe made; in the interaction of electromagnetic radiation with matter, fr^m which inference about electromagnetism and
about chemical structure may be made; in microphysics; and in biology; may

172

A NEWANTHESIS

[Apr.

be met by stating that appropriate investigation of the deviations from equality may be profoundly significant.
Space Time may be involved.

The radius of curvature in non-Euclidian

The variation of gravitation with time may be

involved.
Considering such things in "Gravitation An Enigma5?T E. H. Dicke
says:
"The chief conclusion... is that it is a serious lack of observational data that keeps one from drawing a clear portrait or gravitation.
Each tiny fragment of information appears as a star shining through a
murky haze. Conclusions regarding the most fundamental of physical
concepts are based on numbers which may be off by a factor of 100.M
[36]
Rectification may be achieved by recognition and pursuit of the unification that seems possible in terms of the Golden Angle.
Before finding the details presented here, in an unpublished paper written in January 1968, on the significance of Carl F. Krafft f s work and neglect,
I perhaps gave myself as litterateur too free a rein when I wrote:
"Krafft unambiguously formulated, and, over several y e a r s :
amplified and exploited his idea of spiral molecular structure as the
basis of life.

In terms of his development of Descartes 1 vortices and

his unique f panpsychism, T ... one can visualize Life as an inherent

function of energized matter:

From electrons spiraling about nuclei,

which consist of spiraling mason clouds, yielding atoms which aggregate into stars and planets which spiral about each other and about a
focus of a spiral galaxy; which galaxies perhaps ultimately spiral about
each other as their overall form of motion in the universe, which may
not be expanding if Krafft's ((and my independent)) interpretation of the
red shift in the spectra of galaxies prove more accurate than the
Doppler effect interpretation of that red shift.

As an inherent product

of universal spiral motion, under suitable conditions, matter is energized to spiral into molecular structures which live, evolve, and finally
are energized to such density of redundant interaction as to resonate

1971]

ANEWANTHESIS

173

"self-consciousness, yieldingTeilhard deChardin ! s Noosphere in a way

he doubtless would have rejoiced in understanding.
"It maybe flamboyants but I think that the model of DNA ? sculpted ?
by Crick and Watson, and the models of the alpha-helix by Pauling and
Corey; of hemoglobin by Perutz; of myoglobin by Kendrew; and of insulin by Frederick Sanger; should be enshrined along with if not above
the works of Praxiteles and Michelangelo, as the greatest testimony to
the truest humanistic value of science (man's cosmically emplacedimpulse to know) in fathoming the secret of the ultimate, or at least penultimate, formative principle of life.. . . M
Perceiving the European philosophical tradition as a series of footnotes
to Plato, A. N. Whitehead wrote:
M

. . . i f we had to render Plato T s general point of view with the least

changes made necessary by the intervening two thousand years of human

experience in social organization, in aesthetic attainments, in science,
and in religion, we should have to set about the construction of a philosophy of organism. ??

[37]

Ludwig von Bertalanffy, independent cofounder with Whitehead and others

of contemporary Organismic Philosophy, recently concluded an excellentoritique of psychology in the modern world:
M

.. . science is more than an accumulation of facts and technological ex-

ploitation of knowledge in the service of the Establishment; it may still

be able to present a grand view and to become deeply humanistic in its
endeavor.

If we achieve as much as contributing a bit toward human-

ization of science, we have done our share in the service of society and
civilization. TT

[38]

G. D. Birkhoff, whose gravitational theory needs review of itself if not

in the context of this anthesis; who perceptively explored Aesthetics mathematically [39], wrote:
"The prophetic conjecture that Nature is mathematical is one

which which goes back to Pythagoras and the ancient Greeks. The scientific

174

A NEW ANTHESIS

[Apr.

"development of the intervening 25 centuries has only served to establish this conjecture to a remarkable degree. The complementary fact
that mathematics is natural i s , however, just beginning to be grasped...
"The essential genetic foundation here is obvious.

The mental

codification of the facts of Nature in logical and mathematical terms

has its origin in the uniformity of Nature and of Mind. " [40]
Many thinkers, addressing themselves to problems engendered by the
growth of detailed knowledge, advocate radically new departures.

This

Anthesis, however, is rooted in Tradition embracing Goethe, Leonardo de

Vinci, Leonard of Pisa/Fibonacci, Plato, Pythagoras, and so many more.
This Anthesis affirms Man and his capacity to know, and, what though, not
thinking, at the foot of the cross men cast dice for the seamless raiment,
this Anthesis affirms Einstein's belief that God does not cast dice.

REFERENCES
1. Vallery-Radot, Rene, The Life of Pasteur.
Devonshire.

Translated by Mrs. R. L.

Garden City, New York. Not dated, p. 198.

2. Flammarion, Camille, The Flammerion Book of Astronomy.

Translated by Annabel and Bernard Page.

1955-1960 Edition,

(1880).
revised

and updated 1964. Simon and Schuster, New York, pp. 52-53.
3. de Vaucouleurs, G., "The Case for a Hierarchical Cosmology," Science,
Vol. 167, February 1970, pp. 1203-1213.
4. Grunbaum, A. , "The Bearing of Philosophy on the History of Science,"
Science, Vol. 143, March 1964, pp. 1406-1412.
5. Koestler, Arthur, The Act of Creation, Macmillan, New York, 1964,
pp. 243-244.
6. Dirac, P. A. M. , "Is There an iEther?" Nature, Vol. 168, November
1951, pp. 906-907.
7. Einstein, A. , Relativity, The Special and the General Theory. A Popular
Exposition.

Translated by Robert W. Lawson, Crown Pubs. , New York,

1961, p. 53.
8. Ibid. , pp. 148-151.
9. Wood, Robert W. , Physical Optics, Macmillan, New York, 3rd Edition,
1934 (13th Printing 1959), pp. 822-823.

1971]
10.

ANEWANTHESIS

175

Dingle, H e r b e r t , "Relativity and E l e c t r o m a g n e t i s m : An E p i s t e m o l o g i c a l

A p p r a i s a l , " P h i l o s o p h y of Science, Vol. 2 7 , July 1960, 223 ff.,

through

" T h e C a s e Against Special R e l a t i v i t y , " N a t u r e , Vol. 216, O c t o b e r 1967,

pp. 119-122.
11.

Krafft, C a r l F . , v a r i o u s w o r k s , e s p e c i a l l y : "Life A V o r t e x P h e n o m e n o n , " B i o d y n a m i c s , Vol. 1, No. 18, D e c e m b e r 1936.

12.

P a u l i n g , L i n u s , " T h e Stochastic Method of the S t r u c t u r e of P r o t e i n s , "

A m e r i c a n S c i e n t i s t , Vol. 4 3 , A p r i l 1955, p . 293,

13.

C r i c k , F . H. C. , and Watson, J. D. , "A S t r u c t u r e for

Deoxyribose

Nucleic A c i d , " N a t u r e , Vol. 1 7 1 , A p r i l 1953, pp. 737-738.

14.

G a r a y , A. S. , "Origin and Role of Optical I s o m e r y in L i f e , " N a t u r e ,

Vol. 219, J u l y 1968, p . 338.

15.

D a u v i l l i e r , A . , The P h o t o c h e m i c a l Origin of Life. T r a n s l a t e d by Scripta

T e c h n i c s , Inc. , A c a d e m i c P r e s s , New York and London, 1965, p . 111.

16.

P r i v a t e c o m m u n i c a t i o n , Linus Pauling.

17.

Cohn, Edwin J . , and E d s e l l , John T, , E d i t o r s : P r o t e i n s , Amino A c i d s ,

and P e p t i d e s a s Ions and D i p o l a r Ions. A m e r i c a n C h e m i c a l Society
Monograph S e r i e s , Rheinhold P u b s . , New York, 1943; r e p r i n t e d 1965,
pp. 506-507, and p . 536.

18.

Ibid. , p. 334.

19.

Stent, Gunther S. , The Coming of the (Golden Age: A View of the End of
P r o g r e s s , The N a t u r a l H i s t o r y P r e s s , Garden C i t y , New Y o r k , 1969,
p. 64.
20. R e i s e r , O. L. , The Integration of Human Knowledge, P o r t e r S a r g e n t ,
Boston, 1958, p. 140.
2 1 . K e l m a n , H a r o l d , in T h e o r i e s of the Mind, J o r d a n M. S c h e r , E d i t o r .
F r e e P r e s s of G l e n c o e , New York; M a c m i l l a n , London, 1962, p. 264.
22. Whitehead, A. N. , P r o c e s s and Reality: An E s s a y in Cosmology, M a c m i l l a n , New Y o r k , 1929 and 1957 (5th P r i n t i n g 1960), p . 65.
23. Ibid. , p. 519.
24. E i s e l e y , L o r e n , The I m m e n s e J o u r n e y , Random H o u s e , New Y o r k ,
1957 (7th P r i n t i n g ) , p. 196.
25.
26.
27.

G r a n t - W a t s o n , E. , The M y s t e r y of P h y s i c a l Life, A b e l a r d - S c h u m a n ,
London, New Y o r k , T o r o n t o , 1964.
B e l l , E. T. , Man of M a t h e m a t i c s , E s s a n d e s s P a p e r b a c k , Simon and
S c h u s t e r , New Y o r k , 1937 (5th P r i n t i n g , 1966), p . 135.
E i n s t e i n , A. , op. cit. , p . 100.

176
28.

ANEWANTHESIS
Valery, Paul, Aesthetics.

A p r . 1971

T r a n s l a t e d by Ralph M a n h e i m ,

S e r i e s , Vol. 13 of h i s W o r k s .

Bollingen

Pantheon B o o k s , New Y o r k , 1964, pp.

25-27.
29.

Read, Herbert,

Education Through A r t , Pantheon B o o k s , New Y o r k ,

3 r d Edition, 1956, pp. 14-22.

L e n d v a i , E r n o , "Duality and Synthesis in the Music of Bele B a r t o k , n
in Module, P r o p o r t i o n ,

Symmetry,

Rhythm,

Gyorgy K e p e s ,

Editor,

George B r a z i l l e r , New Y o r k , 1966, pp. 174-193.

K e p e s , Gyorgy, The New L a n d s c a p e in A r t and Science, P a u l Theobald
and Co. , C h i c a g o , 1956, p. 329, and p. 294.
30.

Pauling, Linus,

The C l o s e - P a c k e d - S p h e r o n T h e o r y and N u c l e a r F i s -

s i o n , " Science_, Vol. 150, O c t o b e r 1965, p . 301.

31.

Turnbull, Herbert Westren,

Mathematics,

The G r e a t M a t h e m a t i c i a n s , " The World of

J a m e s R. Newman, E d i t o r ,

Simon and S c h u s t e r , New

Y o r k , 1956, Vol. 1, p. 138.

32.

G r e b e , John J . , "A P e r i o d i c Table for Fundamental P a r t i c l e s , " A n n a l s ,

New York A c a d e m y of S c i e n c e s , Vol. 76, A r t . 1, S e p t e m b e r 1958, p . 14.

33.

G r e b e , John J . , " T i m e : Its B r e a d t h and Depth in Biological R h y t h m s , "

A n n a l s , New Y o r k A c a d e m y of S c i e n c e s , Vol. 9 8 , A r t . 4 , O c t o b e r 1962,
pp. 1206-1210.

34.

Weyl, H e r m a n n , " S y m m e t r y , " in: Newman a s in Note 3 1 , Vol. 1, pp.

67.1-724.

35.

Weyl, H e r m a n n , Philosophy of M a t h e m a t i c s and N a t u r a l Science, A t h e n e u m p a p e r b a c k , New York, 1963, p. 237.

36.

D i c k e , R. H. , " G r a v i t a t i o n An E n i g m a , " A m e r i c a n

Scientist, Vol.

4 7 , M a r c h 1959, p. 40.
37.

Whitehead, A. N. , op. cit. , p . 63.

38.

von Bertalanffy, Ludwig, R o b o t s , Men and M i n d s , B r a z i l l e r , New Y o r k ,

1967, pp. 114-115.

39.

Birkhoff, G. D. , " M a t h e m a t i c s of A e s t h e t i c s , " in: Newman a s in Note

3 1 , Vol. 4 , pp. 2185-2195.

40.

Birkhoff, G. C. , " T h e M a t h e m a t i c a l N a t u r e of P h y s i c a l
A m e r i c a n Scientist, Vol. 3 1 , O c t o b e r 1943, pp. 281-282.

Theories,"

FIBONACCI, LUCAS, AND THE EGYPTIANS

SAL LA BARBERA
San Jose State College, San Jose, California

1.

INTRODUCTION

One of the obvious distinctions between Egyptian m a t h e m a t i c s and the

m a t h e m a t i c s of o t h e r c u l t u r e s i s i t s additive c h a r a c t e r of the dependent a r i t h m e t i c . A typical e x a m p l e i s r e c o g n i z e d when we e x a m i n e the a l g o r i t h m used
by the Egyptians in doing multiplication in c o m p a r i s o n to o t h e r a l g o r i t h m s .
Multiplication (Egyptian Style) i s done by a d o u b l i n g - s u m m i n g p r o c e s s
s i m i l a r to the one shown in the following e x a m p l e .
ing p r o b l e m :

19 x 65.

L e t u s solve the follow-

The E g y p t i a n s noted that the n u m b e r

19 w a s equal

to 1 + 2 + 16 (the s u m of p o w e r s of two), h e n c e , by the addition of a p p r o p r i a t e m u l t i p l e s of 65 the Egyptians a r r i v e d at the d e s i r e d r e s u l t .

We m a y

a r r a n g e the p r o b l e m in the following way:

<
doubling

1*
.2*
\ .

doubling \
doubling

*65
*130
260

f t

^ ^
19

520
*1040

doubling
doubling
doubling
doubling

1235

Upon careful examination of the p r o c e s s e s u s e d in this a l g o r i t h m , we

find t h a t t h e r e a r e two b a s i c c o n c e p t s that contribute to i t s efficiency.

Name-

l y , they a r e the concepts of d i s t r i b u t i v i t y and c o m p l e t e n e s s . The l a t t e r c o n ceived by P r o f e s s o r V e r n e r E . Hoggatt, J r .

[1].

We can e a s i l y identify the r o l e which i s played by the d i s t r i b u t i v e law

in the a l g o r i t h m , for e x a m p l e , in the p r e c e d i n g p r o b l e m
2 + 16).

65 x 19 = 65 (1 +

H o w e v e r , the contribution m a d e by the concept of c o m p l e t e n e s s i s

not self-evident.

Let u s t u r n to the definition of c o m p l e t e n e s s before we e x -

a m i n e i t s r o l e in the Egyptian a l g o r i t h m .

177

178

FIBONACCI, LUCAS, AND THE EGYPTIANS

Definition.

A sequence

[Apr.

S of positive i n t e g e r s i s said to be c o m p l e t e

if and only if e v e r y e l e m e n t n ,

w h e r e n is an e l e m e n t of the positive i n t e -

g e r s can be r e p r e s e n t e d a s a s u m of d i s t i n c t e l e m e n t s of S.
The sequence u s e d in the Egyptian method of multiplication the a u t h o r
shall d e s c r i b e a s T ,

where T

=2

(n > 0 ) .

In o r d e r to show that T

is

c o m p l e t e , we m u s t f i r s t p r o v e the following l e m m a .
L e m m a 1.

Proof.

T 0 + T-, + T? + TQ + + T - = T - 1.
u
L
6
i
n-1
n

We shall p r o v e this l e m m a by m a t h e m a t i c a l induction.

Here,

we have
P(n) : T 0 + T j + T 2 + T 3 + + T ^
Then

TR

- 1 .

P ( l ) : T 0 = Tt - 1 i s e a s i l y seen to be t r u e since

1 = 2-1.

T h u s , we have a c c o m p l i s h e d o u r inductive b a s i s .
Now, suppose t h a t
P(K) : T 0 + T i + T 2 + T 3 + - . + T ^

= T fe - 1

i s t r u e (the inductive a s s u m p t i o n ) , and we m u s t then p r o v e :

P(K + 1) : T 0 + T i + T 2 + T 3 + + T k = T k + 1
By o u r inductive a s s u m p t i o n , we know that
T0 + T i + T2 + T 3 + - - - + T k _ x = T k - 1
H e n c e , by substitution into P(k + 1), we have that
T, - 1 + T. = T. - - 1.
k
k
k+1
It follows that
2T

hence,

2T,. = T . + 1 .

k -

k +

i -

Since T k = 2 k , we have that 2T fc = T k + 1 -

Therefore,

1971]

FIBONACCI, LUCAS 5 AND THE EGYPTIANS

we have shown that if

P(K)

i s t r u e , then

P(K + 1)

179

i s t r u e , and we have

completed the inductive t r a n s i t i o n .

Employing L e m m a 1, we m a y p r o v e the following t h e o r e m .
T h e o r e m 1.

= 2n

The sequence T , w h e r e T

(n > 0) is a complete

sequence.
Proof.

As an inductive b a s i s , we know that

1
2
3
4
5
6
7

=
=
=
=
=
=
=

1
2
1 +
4
1 +
2 +
1 +

2
4
4
2+4,

etc.

H e n c e , we m u s t a s s u m e that t h e r e a r e r e p r e s e n t a t i o n s for all the positive

i n t e g e r s N:
1 < N < 2 n + 1 - 1.
T h e r e f o r e , we m u s t show that t h e r e a r e r e p r e s e n t a t i o n s for all positive i n t e gers

M:
2n+1 - 1 < M < 2n+2 - 1 .

By s u b t r a c t i n g 2

from the above inequality, we have that

-1 < M - 2 n + 1 < 2n+2 - 2 n + 1 - 1 .

n-t-1

Let Q = M - 2

n+1

; hence,

-1 < Q < 2

- 1 . T h i s l e a d s u s to the c o n -

clusion that Q is r e p r e s e n t a b l e a s a s u m of p o w e r s of 2 by o u r inductive

assumption.

And, from t h i s , we can conclude that M i s r e p r e s e n t a b l e as a

s u m of p o w e r s of 2 since M = Q + 2
n+1
1 = l + 2+2
2

and
2

+2

+ ---+2

H e n c e , we have completed o u r inductive t r a n s i t i o n .

180

FIBONACCI, LUCAS, AND THE EGYPTIANS

2.

[Apr.

FIBONACCI-EGYPTIAN METHOD

A s we noted in the introduction, the n e c e s s a r y and sufficient conditions

for the Egyptian a l g o r i t h m to " w o r k " a r e c o m p l e t e n e s s and d i s t r i b u t i v i t y .
The a u t h o r , upon r e a c h i n g this conclusion, went in s e a r c h of

other

s e q u e n c e s that would satisfy the above conditions. The f i r s t sequence e x a m ined p r o v e d to be fruitful.

It w a s the F i b o n a c c i sequence.

the d i s t r i b u t i v e law i s s a t i s f i e d ,

It i s obvious that

since we a r e working solely with positive

i n t e g e r s ; h o w e v e r , it i s not so obvious that the Fibonacci sequence i s c o m plete.

L e t u s then p r o v e this fact.

A s b e f o r e , we m u s t p r o v e a l e m m a before proving the m a i n t h e o r e m .

It i s the following:
L e m m a 2.
F

Proof.

n + 2 - 1 = F i + F2 + F 3 + F 4 + + F n .

We shall p r o v e the l e m m a by m a t h e m a t i c a l induction.

P(n) : F n + 2 - 1 = F A + F 2 + F 3 + F 4 + + F n .

Then P(J.) : F 3 - 1 = F j which i s t r u e , s i n c e

a c c o m p l i s h e d o u r inductive b a s i s .

2 - 1 = 1.

T h u s , we

Now we m u s t suppose that

P(K) : F k + 2 - 1 = F t + F 2 + F 3 + F 4 + + F k
i s t r u e (the inductive a s s u m p t i o n ) , and we m u s t then p r o v e :
P(K + 1) : F k + 3 - 1 = F i + F 2 + F 8 + F 4 + + F k + 1 .
By the addition of F,
F

- to both s i d e s of the equation P(K), we have

k+2 + F k + 1 - 1 = Ft

F2

F3

which H e a d s us to
F

k+3 " 1 = F1

F2

F3

...

k + 1

Fk+1,

have

1971]

FIBONACCI, LUCAS, AND THE EGYPTIANS

181

by the r e c u r s i o n r e l a t i o n for Fibonacci n u m b e r s , n a m e l y

F

= F
n+3

+ F
n+2

n+1

Using t h i s l e m m a , we m a y p r o v e the following t h e o r e m .

T h e o r e m 2.

T h e Fibonacci n u m b e r s f o r m a complete s e q u e n c e .

Proof. The inductive proof will be c o n s i d e r e d in the following way. We

o b s e r v e that
1 = F i = F2
2 = F3 = F 2 + Fj[
3 = F4 = F3 + F 2
4 = F4 + F2 = F3 + F2 + F l f
We shall u s e t h i s a s o u r inductive b a s i s .

etc.

Next, we m u s t a s s u m e t h a t t h e r e

a r e r e p r e s e n t a t i o n s for all positive i n t e g e r s N ,

such that

n+2
is true.

We m u s t t h e r e f o r e show that t h e r e a r e r e p r e s e n t a t i o n s for all p o s -

itive i n t e g e r s M,

such t h a t
F

By s u b t r a c t i n g an F

n+2

- 1 < M < F

from the above inequality, we have that

-1 < M - F
1

Let Q = M - F

+2

- 1.
n+3

n+2

< F
- F
- 1 .
*n+3
n+2

; hence,
-1 < Q < F n + 1 - 1 .

This l e a d s u s to the conclusion that Q i s r e p r e s e n t a b l e a s a s u m of F i b o n a c c i n u m b e r s by o u r inductive a s s u m p t i o n .

And from t h i s , we m a y conclude

182

FIBONACCI, LUCAS, AND THE EGYPTIANS

[Apr.

t h a t M i s r e p r e s e n t a b l e a s a s u m of Fibonacci n u m b e r s , s i n c e
M = Q

Fn+2

and
F _, 0 - 1 = Fi + Fo + F 8 + + F
i

n+2

&

H e n c e , we have completed o u r inductive a r g u m e n t .

L e t us examine the F i b o n a c c i - E g y p t i a n method for multiplication.
e x a m p l e , c o n s i d e r the p r o b l e m

19 x 65.

For

We note that

19 = 1 + 5 + 13 ,
a l l of which a r e Fibonacci n u m b e r s .

T o g e t h e r with the F i b o n a c c i r e c u r s i o n

r e l a t i o n , and the following s e t - u p , we m a y a p p r o a c h the p r o b l e m in the following way:

1*

* 65

+
2

130

+
+

+
3
5*

195
*325
520

13"

*845

19

1235

+
+

One m a y o b s e r v e that in the p r e c e d i n g e x a m p l e , the e n t i r e Fibonacci

sequence w a s not used.

Upon e x a m i n a t i o n , one will find that the f i r s t n u m -

b e r of the sequence h a s been t r u n c a t e d .

T h i s does not, h o w e v e r ,

e i t h e r the c o m p l e t e n e s s of the sequence n o r the distributivity.

effect

The author

shall r e f e r to the Fibonacci sequence with one e l e m e n t omitted a s the Deleted

F Sequence.

H e n c e , let u s p r o v e the following t h e o r e m .

T h e o r e m 3.
arbitrary

The deleted F s e q u e n c e , w h e r e f

not u s e d , i s c o m p l e t e .

= F

(n > 1) with

1971]

FIBONACCI, LUCAS, AND THE EGYPTIANS

Proof.

183

From the previously proven theorem, it was noted that we may

represent any positive integer n,

the Fibonacci numbers

Ft

where

l < n < F

through F _-,

+ 1

- l

by using only

without using F .

Hence, we

shall consider F n as the arbitrary Fibonacci number to be omitted. We may

observe that F n + 1 can represent itself.

Since this is true, it is noted that

we now have representations for 1 < n < 2F

- - 1. Since we have increased

our upper bound from what it was formerly, we may use this particular technique so that we may have representations for any positive integer without
using F . For example, if F

= 1, which is proposed to be the deleted

number, then the sequence would remain complete.

Therefore, we have another method for multiplication which may be
employed by those who have not mastered the traditional algorithm.
3. LUCAS-EGYPTIAN METHOD
Another sequence which proves fruitful in using our algorithm is the
Lucas sequence. The Lucas sequence is composed of the numbers
(1, 3, 4, 7, 11, 18, 29, 47, )

and can be used effectively for the base sequence in an Egyptian multiplication problem.

However, there is one acute difficulty in the consideration of

this sequence for our algorithm; it does not have any representation for the
positive integer 2. Therefore, something must be done to the sequence before we can apply it to our algorithm, since without a representation for the
number 2 it is not complete.
The author chose to augment the sequence in the following way and define his Augmejite^lJ-^^

A = L

- , where A t = 2, A2 = 1 5

A3 = 3, and so on.
The reader will observe that this augmented sequence has a representation for 2 and also observe the recursion relation for the Lucas Sequence,
namely A

= A +A

-. Hence, we may use it for our base sequence in

the Egyptian algorithm.

The problem 18 x 54 may be set up in the follow-

ing fashion.

184

FIBONACCI, LUCAS, AND THE EGYPTIANS

+
+
+
+

2
1
3
4

7
+ 11
+ 18
18

[Apr.

108
54 +
+
162
216 +
378 +
594 +
972 +
972

The augmented L u c a s sequence i s c o m p l e t e and m a y be p r o v e d to be in a

s i m i l a r fashion to T h e o r e m 2 , by u s e of L e m m a 3 , which s t a t e s
L e m m a 3.
L 0 + L i + L 2 + L 3 + - + L n = L n + 2 - 1.
Proof.

Using an inductive proof, we have as o u r b a s i s

P ( l ) : L0 + L j = L3 - 1

which i s t r u e , since 2 + 1 = 4 - 1 . O u r inductive a s s u m p t i o n is

P(K) : L 0 + L j + L 2 + L 3 + ' ' ' + L k = L k + 2 - 1 .
We m u s t then p r o v e that
P(K + 1) : L 0 + L i + L 2 + L 3 + + L k + 1 = L k + 3 - 1
is true.

T h i s m a y be a c c o m p l i s h e d by adding a L k + 1 to both s i d e s of P(K).

H e n c e , we have that
L 0 + L i + L2 + + L k + L k + 1 -

L k + 2 + L k + 1 - 1,

which l e a d s u s to the fact that

L0 + L i + L2 + L3 + L4 + + L k + 1 = L k + 3 - 1

1971]

FIBONACCI, LUCAS, AND THE EGYPTIANS

185

Hence, our induction transition is complete.

Invoking this lemma, we may prove the following theorem.
Theorem 4. The augmented Lucas sequence is complete.
Proof. As our inductive basis, we have that
1 = Lt
2 = L0
3 = L2
4 = L3
5 = L3 + Lj, etc.
As our inductive assumption, we assume that for N,

a positive integer,

there are representations for N in terms of Lucas numbers so that

1

<

<

n+2 "

Hence, we must prove that for M, a positive integer, M is representable

a s a sum of Lucas numbers between the intervals of

I 0

- 1 < M < L

n+2

_ Q - 1 .
n+3

Using the same idea as described in the previously proven theorems, we

shall subtract an L

from the above inequality. Hence, we have that

- K M Let Q = M - L n + 2 -

Ln+2 < Ln+3 - Ln+2 - 1 .

Therefore,
- 1 < Q < L n + 3 - L n + 2 - 1.

This leads us to the conclusion that

-1 < Q < L n + 1 - 1 .

186

FIBONACCI, LUCAS, AND THE EGYPTIANS

[Apr.

We may conclude that Q is representable as a sum of augmented Lucas numbers. And from this, we can conclude that M is representable as a sum of
augmented Lucas numbers, since M = Q + L

?.

Other sequences may be investigated and tested for completeness; however, no others with starting values other than (1,1), (1,2), and (2,1) will
be found which satisfy the generalized Fibonacci recursion relation.

In gen-

eral, other sequences that are complete will follow the following generalized
recursion relation
n-1
G

J = (2,3,4,...),

q=n-j
and where the starting values for the above sequences are taken from either
the augmented Lucas sequence or the deleted F sequence.

For example,

let us examine the Tribonacci sequence, where three numbers are added.
The generalized recursion relation would look like the following:
n-1

= y

Zi

q=n-3
Hence, the sequence would be
(1, 2, 3, 6, 11, 20, ) .
In general j determines the number of terms to be added together and also
the number of starting values to betaken from either the deleted F sequence
o r the augmented Lucas sequence.
The author at this point feels that it would be valuable for the reader to
have a simple method for determining whether a sequence is or is not complete. It was observed and proven by John L. Brown, J r . [2] that the necessary and sufficient conditions for a sequence to be complete is that the
sequence satisfy the following general summation formula

1971]

FIBONACCI, LUCAS, AND THE EGYPTIANS

187

n
A

n+1 *

1 +

i=l

where Aj = 1. Hence, we now have a convenient way in which to determine

a sequence complete.
The material submitted in this paper is not completely theoretical and
does have very definite practical application.

The author used both the de-

leted F sequence and the augmented Lucas sequence in conjunction with the
Egyptian method in a class of "slow learners. M The results were phenomenal.
Those students who could not multiply by traditional means were then given
a method even they could handle. You see, all one needs to be proficient in
the methods given above is an adequate understanding of simple addition.
The author found that most slow learners could add correctly, however, they
could not multiply.

Therefore, this algorithm best fit the needs of those

students.
The concepts mentioned throughout the paper may also be used in advanced mathematics classes.

Hence, as one can see, the utility of these

topics and their applications is boundless.

It is the author's intent that the reader search for other complete s e quences and establish those concepts revealed in this paper, so that he may
transfer the concepts to others and hence, give many an algorithm for multiplication which they may not already have.
The author would also like the reader to be aware of the fact that it is
sometimes advantageous to use one complete sequence over another.

For

example, it is better to use the Lucas sequence when multiplying the numbers
18 x 432, than it is to use the Fibonacci sequence or the powers of two s e quence, since 18 is an element of the Lucas sequence. Therefore, this was
the primary reason the author went in search of other complete sequences.
The author hopes that the methods for multiplication developed in this
paper will be tried, and hopes that the success of those using them will be as
rich as his own.

[Continued on page 194. J

THE GOLDEN RATIO IN AN ELECTRICAL NETWORK

J. WLODARSK!
Proz-Westhoven, Federal Republic of Germany

At the end of June 1967, Poland called together an international physics

olympiad for grammar-school students in Warsaw.

Five countries partici-

pated: Bulgaria, Poland., Romania, Czechoslovakia, and Hungary.

During this competition, the following problem was presented, among
others:
An infinite network consists of the resistors r. Calculate the r e s i s t ance between points A and B.
T

A*

VVV

1-j

v W

; r .-jD!
D
The solution of this problem can be presented in different ways.
B

One

quite brief version is possible as follows:

Suppose the resistance of the infinite network on the right-hand side of
points C and D is equal to r .
If we go one step to the left from points C and D to C

and D', the

resistance of the network would be

r r
r

CTDf

n
r + r
n

in accordance with relation:

_J_ = I
r

CTDT

I.
r
n

The next resistor r is added on the left behind the resistance r r T

therefore, the resistance between A and B is
[Continued on page 194. ]
188

KAPREKAR'S ROUTINE WITH TWO-DIGIT INTEGERS

CHARLES W.TRIGG
San Diego, California

Kaprekar T s routine consists of rearranging the digits (not all alike) of

an integer, N 0 , to form the largest and the smallest possible integers, finding their difference,

Nj, and repeating the operation on N4 and on the sub-

sequent differences until a terminal situation is reached.

He found p.] that

when the routine is applied to any four-digit (not all alike) integer in the decimal system, the self-producing 6174 is eventually reached.

The routine has

been expanded to other number bases [ 2 ] , [3], and to three-digit [4] and
five-digit [5] integers.
When applied to two-digit integers, an integer and its reverse are involved, the smaller being subtracted from the l a r g e r in each step of the routine.

In the system with base r, all differences are multiples of r - 1.

Each step may be called a reversal-subtraction-operation (RSQ).

There are three possible terminal situations which may result when the
routine is applied to a two-digit integer, namely:
A. If at any step of the repetitious routine, an integer with two like
digits is produced, its Nj will be 00.
B. A self-producing integer is formed.

That i s , the integer is r e p r o -

duced when subjected to an RSO. For example: 37 in the scale of

eleven, where 73 - 37 = 37.
C. A regenerative loop is formed, in which an RSO on one member
produces the next member.

For example: in the scale of nine,

53 - 35 = 17, 71 - 17 = 53, and so on.

In each of these categories, if all two-digit integers in a particular system lead to the same result, it is said to be unanimous.
The two-digit ordered integers ab and a + k b + k have the same NA,
Hence, to investigate the entire field in the scale of r ,

it will be sufficient

to examine only the r - 1 integers with the form r - 1 b, where b < r - 1.

Each of these is the representative (rep) of all two-digit integers from which
an RSO produces the same NA. Any two-digit integer can be converted into
its rep by addition of an appropriate multiple of 11, ThuSj
189

52 + 44 = 96 in

190

[Apr.

KAPREKAR'S ROUTINE WITH TWO-DIGIT INTEGERS

the decimal system, so 52 - 2 5 = 27 = 9 6 - 69. All integers with the same

rep have the same value of a - b.

THE EXAMINATION PROCEDURE

An RSO is performed on each of the reps.
converted into its own rep.

Each difference,

Nj,

is

These results are assembled into flow charts

such as those given below. All NjTs with the same rep are placed below a
common subtraction line. Their common rep is placed below them on the
left.

The category letter of each terminal situation is placed below it.

Thus, only r - 1 RSO!s are necessary to examine the entire field in

the base r.

The number of steps necessary to convert any integer into the

terminal result can be read directly from the chart after locating its rep. In
the scale of five, three RSOTs convert 12 (which has the rep 43) into the selfproducing 13.
In the charts for bases eleven and twelve, the symbols X and E stand
for the digits ten and eleven, respectively.

OPERATIONAL FLOW CHARTS

Base Two

Base Three

Base Four

Base Five

10

21

30

31

43

01

15

03

13

34

41
14

01

02

-21

12

04

22
A

B
32

40

02

23

11
A

-03

04
31

20

C
42
24
13
B

1971]

KAPREKAR T S ROUTINE WITH TWO-DIGIT INTEGERS

B a s e Six

B a s e Seven

191

B a s e Eight

50

53

65

63

61

70

75

71

05

35

_56

36

16

07

57

17

r41

14

06

24

42

-61

16

52

52

51

60

64

72

73

74

25

15

06

46

27

37

47

23

32

51

15

43

34

25

B
76

54

62

45

26

67

05

33

07

Base Ten

B a s e Nine

85

81

83

58

18

38_

09

79

26

62

44

81

18

90

97

A
84

87

92

48

78

29

95
59

35

08

63

36

86

80

96

91

68

08

69

19

17

71

27

72

82

94

93

^8

49

39

L53

45

54

98

_89
L09

1971]

KAPREKAR'S ROUTINE WITH TWO-DIGIT INTEGERS

192

X9

Base Eleven
X7
X5
X3

9X
OX

5X
46

XO

X8

X4

E2

E7

OX

8X

4X

2E

7E

91

19

55

83

38

3X
64

7X
28

Base Twelve
XI

EO

E9

IX

OE
,rXl

9E

82

IX

A
X2

E6

E3

2X

6E

3E

73

47

74

X6

E8

El

6X

8E

IE

37

29

92

B
E4

E5

4E

5E

65

56

EX
XE
LOE
C
SUMMARY AND GENERALIZATIONS
1. Every system with an odd base has a sequence leading to 00, since
r - 1 (r - 3)/2 - (r - 3)/2 r - 1 = (r - l)/2 (r - l)/2
In bases three and seven, 00 is unanimous.
2. If a self-producing integer,
with base r , then

kx with k < x,

exists in a system

1971]

KAPREKAR'S ROUTINE WITH TWO-DIGIT INTEGERS

193

(rx + k) - (rk + x) = rk + x
whereupon
r = <2x - k)/(x - 2k) = 2 + 3k/(x - 2k) .
Then, since x < r , self-producing integers, which will have the form
k 2k + 1, exist in and only in systems with bases of the form 3k + 2.
Such bases are two (in which 01 is unanimous), five, eight, and eleven.
3. Both r - 1 c and r - l r - 3 - c
of each other, since
r - l c - c r - 1

have N ^ s which are the reverse

= r - 2 - c c

+ l

and
r - l r - 3 - c - r - 3 - c r - l

= c + l r - 2 - c

Hence, the NjT s have the same rep.

4. If r is even and not of the form 3k + 2, the result of application
of RSO!s to the reps in that system is a unanimous regenerative loop of r / 2
elements.
REFERENCES
1. D. R. Kaprekar, "An Interesting Property of the Number 6174," Scripta
Mathematica, 21 (December, 1965), p. 304.
2. J. H. Jordan, "Self-Producing Sequences of Digits," American Mathematical Monthly, 71 (January, 1964), pp. 61-64.
3* Charles W. Trigg, "Predictive Indices for Kaprekar's Routine," Journal
of Recreational Mathematics, scheduled for publication in October, 1970.
4. Charles W. Trigg, "All Integers Lead to ," submitted to The Arithmetic Teacher.
5 Charles W. Trigg, "Kaprekar* s Routine with Five-Digit Integers," submitted to Mathematics Magazine.

194
THE GOLDEN RATIO IN AN ELECTRICAL NETWORK
[Continued from page 188. ]
, r rn
*AB
r + r

Apr. 1971

Since the network is infinite, we can disregard the addition of one s e c tion of each sequence. This allows to determine the resistance between points
A and B as equal to the resistance between C and D.
Consequently,
r r
= r +

r + r
n

After solving this equation, we have *

rn = r ^ -

= r 0

where </) is the Golden Ratio.

See also, S. L. Basin, "The Fibonacci Sequence as it Appears in N a t u r e , " Fibojna^cJ^u^rterij;, Vol. 1, No. 1, p. 53.
[Continued from page 187. ]
Editorial Note: The question remains how the students are to find the Fibonacci
or Lucas representation for the first factor. To find the Fibonacci representation for 28, we subtract the largest Fibonacci number not exceeding 28,
namely 21. This leaves 28 - 21 = 7y so our next choice is 5' 28 - 2 1 5
= 2, a Fibonacci number.

Thus, 28 = 21 + 5 + 2.

This will always yield

the representation with the least number of summands.

REFERENCES
1. V. E. Hoggatt, J r . , Fibonacci and Lucas Numbers,

Houghton-Mifflin

Company, Boston (1969), pp. 69-72.

2. John L. Brown, J r . , "Note on Complete Sequence of Integers," American
Mathematical Monthly, Vol. 67 (1960), pp. 557-560.

A NUMBER PROBLEM
J. WLODARSKI
Porz-Westhoven, Federal Republic of Germany

The last digit of one number is 6.

front of a given number.

Take away this digit and put it in

Thus a new number can be formed which will be 6 times bigger than the
original one.
What smallest of all possible numbers satisfies this condition?
Simple Solution
6
36
216
1296
7776
46656

1016949

6779664

L-a period with 58 digits*

The number to be found is:

1016949152542372881355932203389830508474576271186440677966 .

Find this number in another way using the Fibonacci terms.

Remark. 2, 8, 34, 144, , etc. are third, sixth, ninth, twelfth,
(general: F . , where n = 1, 2, 3, 4, ' ' ) ' , etc. terms of Fibonacci sequence 1, 1, 2, 3, 5, .
The solution by using the Fibonacci numbers is given on the following page.
[Continued on page 198. J
195

NUMBERS THAT ARE BOTH TRIANGULAR AND SQUARE

THEIR TRIANGULAR ROOTS AND SQUARE ROOTS
R. L. BAUER
St. Louis, Missouri

There is an infinite series of numbers, N, which for integral T and

S:
| T ( T + 1) = N = S2 .

(1)

The first nine members of the series are tabulated below, together with their
triangular roots, square roots, and index numbers, n.
N
0

1
2

1
8

36

3
4

49
288

1225
41616

35
204

1681

1413721

1189

6
7
8

9800

48024900

6930

57121

1631432881

40391

332928

55420693056

235416

By inspection of the tabulation, we note the recursive formula for N:

(2)

N = 34 N - - N 0 + 2 ,
n
n-1
n-2
'

from which we can develop a generalized formula for N:

(3)

n = A [(17

+ 12

^)n

+ (17

Similarly,
196

" 12 N^") n - 2 ] .

Apr
p

(4)

1971

NUMBERS THAT ARE

BOTH TRIANGULAR AND SQUARE
T = 7T - - 7T 0 + T 0 9
n
n-1
n-2
n-3 '

iy

'

and

(5)

T n = | [(3 + 2 *J2 f

+ (3 - 2 ^ )

- 2 ].

Also:
<6)

n =

6S

n-l "

n-2

and
N/2 [(3 + 2 ^ 2 ) n - (3 - 2 ^ 2 ) n ]

Sn = |

Other recursive formulas and relations were found by inspection of the

tabulation:
(7)

S 0 = S (S _,_- - S - )
2n
n n+1
n-1

(8)

T 0 - = (T - T , )2
2n-l
n
n-1

(9

>

(10)

n "

n-1

T 2 n = 8N n

(11)
'

- T
n

<12>

<13>
<14>

2n-1 =

, = S + S n-1
n
n-1

2n-1 =

2 n - 1 = <Sn '

n -

n-1 =

(S

(S

+ S

n-l)2

n-l)(Tn " Tn-1>

n "

n - l ) ( T n " Tn-1>

*This simplification of the author f s more complicated formula was furnished

by Hoggatt

198

(15)
(16)
(17)

NUMBERS THAT ARE

BOTH TRIANGULAR AND SQUARE
T
= 8S 2
2n
n

Apr 1971

S0 - = (S - S - )T% 2n-l
n
n-1 2n-l
N - N - = ( S - S - ) ( S + S - ) .
n
n-1
n
n-1 n
n-1

By the use of the recursive formulas, the tabulation was extrapolated for
negative index numbers. It was found to be perfectly reflexive about 0 except that the values of S became negative for negative index numbers, while
the values of N and T remained positive.

All generalized formulas and

recursive formulas and relations held for the reflected series.

[Continued from page 195. ]

Solution by Using the Fibonacci Terms

2
8
34
144
610
2584
10946
46368
196418
832040

33893 X 3389 = 1016949 .

RECREATIONAL MATHEMATICS
Edited by
JOSEPH S. MADACHY
4761 Bigger Road, Kettering, Ohio

ASYMPTOTIC EUCLIDEAN TYPE CONSTRUCTIONS

WITHOUT EUCLIDEAN TOOLS1
JEAIM J. PEDERSEN
University of Santa Clara, Santa Clara, California

INTRODUCTION
" . . . , Gauss made the remarkable discovery that those, and only
those, regular polygons having a prime number of sides p can be constructed with straight edge and compasses if and only if p is of the
form 22 + 1. Now the ancient Greeks had found how to construct with
straight edge and compasses regular polygons of 3, 4, 5, 6, 8, 10 and
15 sides. If in the formula p = 22 + 1 we set n = 0 and 1, we obtain the primes 3 and 5 respectively cases already known to the
Greeks.
Therefore

For n = 2,

we find p = 17, which is a prime number.

Gauss showed that a regular polygon 6f 17 sides is con-

structive with straight edge and compasses, which was unknown to the
Greeks. Gauss was vastly proud of this discovery, and he said that it
induced him to choose mathematics instead of philology as his life workT.T 2
This quote from Howard W. Eves ! recent two-volume set, In Mathematical Circles, suggests that the construction of regular polygons having a
prime number of sides is not easy, even when possible, with a straight edge
and compass. Note that Gauss showed it is impossible to construct with a
ruler and compass the regular seven-sided polygon.

Furthermore, one

method for showing that a general angle 6 cannot be trisected with Euclidean tools involves showing that it is impossible to trisect the angle whose

*Text and illustrations copyright 1971 by Jean J. Pedersen.

2
HowardW. Eves, In Mathematical Circles, Quadrants UI and IV, Prindle,
Weber and Schmidt, Inc., Boston, 1969, p. 113.

199

200

RECREATIONAL MA THE MA TICS

[Apr.

measure is 7r/3 hence, the nine-sided regular polygon is not constructible

with a ruler and compass

either. *

The first part of this article deals with a way to approximate, by folding a paper strip, any regular polygon whose number of sides is of the form
2 1 , for some natural number n. Note that when n = 3, the expression
2 n 1 yields 7 and 9.
A modification of the iterative folding sequences used on paper strips
is presented.
ure 6/(2

It suggests a method for approximating an angle having m e a s -

+ 1), where n is any natural number and 0 is any given angle

whose measure is between 0 and TV Particularly interesting is the case

when n = 1, which produces a trisection approximation process.
Finally, as an illustration, instructions are given describing how paper
strips may be used to construct models of regular convex dodecahedra.

The

constructions suggest, as will be seen, that a "parallel strip" classification

of certain polyhedra might provide an interesting point of view from which to
study their properties.
FOLDING SEQUENCES INVOLVING ONE ITERATIVE EQUATION
As an elementary example, take a roll of ordinary adding machine tape
and make a fold on any straight line, t 0 , near the end of the tape so that
t0 crosses one of the parallel edges of the tape at a point, A0.

Fold again

through A0 to bisect one of the angles formed by t0 and an edge of the tape.
Do this so that the newly created transversal, tA, goes towards the roll of
paper. One endpoint of tj is A 0 , the other endpoint is named

Aj. M Now

fold the tape through A1;> bisecting the obtuse angle created by tj and the
edge of the tape. This fold yields yet another transversal, 1^, whose endpoints are Aj, A2. To continue this folding process
ing through A , the obtuse angle, having sides t
thereby obtaining a new transversal, t

always bisect, by foldand an edge of the tape;

- , having endpoints A , A

n = 1, 2, 3, ) . The acute angle formed by t

- (for

and an edge of the tape is

denoted x n-1...
*Howard W. Eves, An Introduction to the History of Mathematics, Rinehart
and Company, Inc. , New York, 1953, pp. 96-98, p. 107.

1971]

RECREATIONAL MATHEMATICS

201

For the most accurate results, both in this case and all other examples
which follow, fold the tape so that whenever transversals are formed, the tape
remains folded on these creases and the next fold always occurs on the portion of the paper strip which comes from the top of the existing configuration.
Thus, the triangles which are formed will either stack up or form a zig-zag
type pattern in the folding plane, but the configuration formed will never need
to be turned over during the folding process.

One quickly discovers, how-

ever, that certain rotations of the configuration in the folding plane facilitates
the folding process.

Figure 1 illustrates one case of how the unfolded tape

appears after the folding process has taken place.

When the above folding process is accurately carried out, an accordianlike stack of triangles results. And, it soon becomes visually apparent that
successive triangles are getting more and more alike consequently, the
measure of x

must approach TT/S as n gets large.

For skeptics, the proof can be ascertained.

F i r s t , note that since the

edges of the tape are parallel, the measures of successive acute angles always satisfy the equation
2x

where n = 1, 2, 3, .

V l

= '

Successive computations of x 1? x 2 , x 3 , e t c . ,

yields
xn = |

[ 1 + (-1/2)1

(_l/2)2 +

...

(-i^)11"1] + (-l/2)nx0

which can be verified by mathematical induction.

Then, using the formula

for the sum of a geometric sequence, it follows that

xft = I [1 - (-l/2) n ] + ( - l / 2 ) n x0 .

Consequently,
lim x = 7T/3 .
n* n

202

RECREATIONAL MATHEMATICS

[Apr.

/
/
(Si
CD

CM
CD
PH

bJD

rH

PH

\
P

\ ,w
*
*-

iSM

CN/

*\X I
00 \ | CO

/
/

,H

*9CXI

CD
?H

>^

rH

Hi

K"

\x
\

\
4J \

Ki!

3CM
CD

x 2H

rH

X
CM

\
\

\ O

bJD

\ J>

, 0

1971]

RECREATIONAL MATHEMATICS
Notice that the difference, in radians, between x
nH

by the (n + 2)

203
n

(which is formed

fold of the tape) and ir/3 is

( - l / 2 ) n [ x 0 - 7T/3] .

This means each new accurate fold on the tape produces an angle whose
measure is twice as close to 7r/3 as its predecessor. In fact, the maximum
value for the actual e r r o r (which occurs when x0 approaches zero) indicates
that one can always expect an approximation of tr/3 with accuracy better than
one minute after 14 folds.

But, as one is not likely to choose x0 close to

zero, this degree of accuracy will occur, in most cases, when n < 14.
It turns out to be practical, in the paper tape construction of models,
to have the following:
Visual Criterion.

When the consecutive longest transversals formed

on a tape by an iterative folding process, appear to be of the same length,

then the tape is called usable.
Suppose the length of successive transversals obtained from some iterative folding process approaches some fixed value, in the limit sense.

Then

there must exist some number k f 0 such that consecutive acute angles
formed by those transversals and an edge of the tape converge to an angle
having measure 7r/k.
Definition.

On a usable tape, whose successive smallest acute angles

converge to 77/k; when the portion not satisfying the Visual Criterion is cut
off, the remaining tape is denoted

T(7r/k). n

Accordingly, the usable tape produced in the above example is denoted

?r

T(7r/3)Tf and called a Mpi thirds tape. n

The method of obtaining T(ir/k) implies that there will always be some

natural number, p, such that the transversals t , where n < p, will not
appear on that tape. But, it is not necessary to identify p. Thus, in describing constructions, reference to a transversal t n on T(7r/k) will mean any
transversal on T(7f/k).

However, once t

has been identified for use in a

particular construction, then t + (where q is any natural number) will

th
mean the q
transversal following t .

204

RECREATIONAL MATHEMATICS

[Apr.

Since T(7r/3) contains approximations of equilateral triangles, it may

be used to construct models of hexagons and deltahedra. As an example, cut
T(7r/3) on t and t + 1 A> then fold the ten triangle strip on t and t fi.
Now, because straight lines are easier to fold than to cut, the t 1 0 end of
the tape is wrapped around t when the tape is folded on t q to complete
the model of a hexagon. Note that the definitive edges do not include either
of the cut edges
t n , t n+10
&
lirk.
The above folding process generalizes in the following way.
Theorem 1. If
(1) n is some fixed natural number.
(2) A paper tape of width w is folded on some transversal, t 0 , which
crosses one of the parallel edges of the tape at A 0 .
(3) One angle formed by t0 and an edge of the tape is then divided into
2

parts, by folding through A 0 ; creating, in order, the new set

of transversals, t l s t 2 , t 3 , , t , where tj < t2 < t3 < < t .

The measure of the acute angle formed by t
tape is denoted x 0 .

The endpoint of t

and the edge of the

which lies on the opposite

edge of the tape from A0 is called AA.

(4) In general, folds are made so as to divide into 2

parts the obtuse

angle having vertex A, and an interior with no transversals. The

new transversals,
< W

<

tkn+1> t ^ ,

, t k n + n > are such that

The endpoint of t ^ ,

on the opposite edge of the tape from A,

t ^

called A f c + 1 , lies

(for k = 1, 2, 3, ) .

The measure of the acute angle formed between t

and an edge

of the tape is denoted x, .

Then , lim x, = 7r/(2 n + 1) and consequently, this folding process produces T(7r/(2n + 1)).
Proof.

From the description of the folding process, it follows that the

measures of successive acute angles satisfy the equation

(1)

2nxk + x ^

where k = 1, 2, 3, " .

Then, using mathematical induction, it can be

shown that

= 7T ,

1971]

RECREATIONAL MATHEMATICS
x,k

2n + 1

for k = 1, 2, 3, .

[1 - ( - l / 2 n ) k ] + ( - l / 2 n ) k x 0

205
,

But, since | - l / 2 n | < 1, it follows immediately that

, lim x = 7r/(2 n + 1) .
K - CO

The theorem is surprisingly fruitful.

For example, Figure 2(a) illus-

trates how the folded tape appears just after the folding process has taken
place with n = 2. Figure 2(b) shows how this same tape appears when it is
unfolded.

This folding process produces the usable tape, T(7r/5). If T(7r/5)

is cut on t 2 n and t 2 n + 6 , and folded on t ^ ,

t 2 n + 3 , t 2 n + 5 , a model of the

regular pentagon shown in Figure 2(c) is formed.

The sides of this pentagon

approximate w/sin (27r/5). But, a regular pentagon whose sides approximate

w/sin

(TT/5)

t 0 ,- and t 0

may also be formed from T(7r/5).

TO

see this, cut T(7f/5) on

1Q ; then fold in a winding fashion on the transversals

t0 , 0 ,

2n+4' W e ' Sn+S' W o * Sn+12 T h e r e s u l t ' a m o d e l o f a r e u l a r P e n "

tagon with a pentagonal hole in the center, is shown in Figure 2(d).
As another example, consider the results of the theorem when n = 3.
Figure 3(a) shows how the beginning of the tape which produces T(7T/9) might
appear. Once T(7r/9) has been obtained, it may be used to construct models
of regular 9-gons whose sides approximate either w/sin (ir/9), w/sin (27T/9),
or w/sin (4ir/9). This is done by folding T(7T/9) on consecutive transversals

whose labels are equal to 0 (mod 3), 2 (mod 3), and 1 (mod 3), respectively.
Figure 3(b) illustrates the regular 9-gon which is formed by folding on t
t

4,

-,

to ,7> j to joo'* and whose sides approximate w/sin (47T/9).

In general,

T(7r/(2n + 1)) will produce models of n non-congruent

regular (2 + l)-gons whose sides approximate w/sin (2 7r/(2 +1)), k = 1,

2, , (n - 1). The actual construction involves folding T(7r/(2 + 1))

on

successive transversals whose labels are equal to 0 (mod n), (n- 1) (mod n),
(n - 2) (mod n), , 1 (mod n),

respectively.
A BONUS

Suppose the folding process described in the theorem takes place on a

piece of paper whose straight edges are not parallel* Thus, suppose angle

RECREATIONAL MATHEMATICS

206

|\

\
$\

00 \

\ CJ

I
I
I

l *>
'-P

rip/"
1/

\l'*f -

V
\
\
\

\
\

\
K

\0

[Apr.

1971]

RECREATIONAL MATHEMATICS

207

ABC, having measure 0 (between 0 and IT), and supplementary angle A BD,
occur so that DBC lies on the edge of a pieaa of paper. Then the paper is
cut along the line AB (see Figure 4). A point, A 0 , is selected between D
and B and the paper is folded, through A0 on some line, t 0 , which is not
parallel to AB. The transversals t

(where k = 1, 2, ) are formed by

folding so that tA bisects the angle formed by to and A 0 B, determining a

point,

Aj,

angle A.

on the line containing AB.

2 A.

And, in general, t,

bisects the

- B , determining a point A, on the line containing A, ~B

(when k >. 2). The measure of the angle A ^ Q B is denoted x0 and half the
measure of angle
A, n A
B is denoted x. - for k > 2.
&
k-2 k_i
k-1
Then, since the sum of the measures of the interior angles in any t r i angle is always equal to 1T it follows that
2 x k + X k _ 1 + (7T - 0)

when k = 1, 2, 3, .

= 7T

Thus,

(2)

2x k + x k _ x = 6 ,

when k = 1, 2, 3, .
But this is similar to Equation 1, where n = 1* In fact, a review of
the proof for Theorem 1 reveals that it would not have been any more difficult
if

Tt Tf

77

were replaced with

M M

and that Equation 2 would lead to the result

. lim x.

= e/3 .

Thus the method illustrated in Figure 4 really represents a trisection

approximation method for angles whose measure is between 0 and 77. A s a
practical matter it is not, in this case, possible to fold accurately indefinitely, as was the case with parallel lines. Nevertheless, the method is effective
especially when judicious choices of A0 and to are made i. e. , choose
A0 as far away from B as the paper will allow and make a visual guess
when folding to so that when x0 is formed, it will be as close to 0/3 as
possible.

208

RECREATIONAL MATHEMATICS

[Apr.

Figure 4
OTHER FOLDING SEQUENCES
The folding sequences considered thus far have involved just one iterative equation. But, as the next theorem shows, other folding sequences do
exist.
Theorem 2. If
(1) n is some fixed natural number greater than 1.
(2) A paper tape, of width w, is folded on a transversal, t 0 , which
crosses an edge of the tape at some point, A 0 .
(3) The angle formed by t0 and the edge of the tape having vertex A0
is divided, by folding, into 2

parts producing transversals t l9

1971]

RECREATIONAL MATHEMATICS

209

k > ^3 9 ''' t so that ti < t2 < t 3 < - < t ; and t has endpoints A0, A 4 . The measure of the acute angle which t n makes
with the edge of the tape is denoted x 0 .
(4) The obtuse angle at A j is bisected, creating a new transversal
t

-. It has endpoints A l5 A 2 and forms an acute angle with an

edge of the tape, denoted

xt.

(5) In general, either (i) the obtuse angle at A, is divided into 2

parts, when k is even, so that each new transversal is longer than
its predecessor and the last transversal folded creates the point
A, - on the opposite edge of the tape from A, ; or (ii) the obtuse
angle at A,

is bisected, when k is odd.

In either case, the

measure of the acute angle between the transversal joining A, ,

A, - and an edge of the tape is denoted x, .
Then , lim x 91 = 7r/(2
sequence produces T(7r/(2

- 1) and consequently, the folding

- 1)).

Proof. By the description of the folding process, it follows that the

measures of consecutive acute angles satisfy

,v

2x

2 k - l + X 2 k - 2 = \
n
J for k = 1, 2, 3,
A
2 x
2k + x 2 k - l = * )

Solving for x 2 . - in the first iterative equation, then for x^.

In the second

yields
X

2k

(7r+X

2k-2)/2n+1 '

It can then be shown, by mathematical induction, that

x,2k

[1 - ( l / 2 u " r r ] + U / 2 n ^ ) x0 ,

n n+l

for k = 1, 2, 3, - .

Thus
lim

x9.

= 7r/(2n+1 - 1)

210

RECREATIONAL MATHEMATICS

[Apr.

In general, if T(ir/(2 n - 1)) is folded on all t .,, t.

n+1
0 (mod (n + 1)), a regular (2
- l)-gon will be formed.

9,

where k =

As an example, suppose n = 2 in Theorem 2. Figure 5(a) illustrates

how the beginning of this tape, which produces T(7r/7), might appear.
If T(7r/7) is folded, in a winding fashion, on all t, - , t,

9,

where

k = 0 (mod 3), the model formed is a regular seven-sided polygon (Figure

5(b)), whose sides approximate w/sin (ir/7).
Likewise, if

T(TT/7)

is folded on all t. where k f 1 (mod 3), the r e -

sult is a seven-sided polygon whose sides approximate w/sin (BTr/7). If this

is done so that the folds on t , when k = 0 (mod 3), wrap the tape around
the polygon being formed; then the result appears as shown in Figure 5(c).
Note, however, that as illustrated in Figure 5(d), if T(7r/7) is folded
on all t

where k f 2 (mod 3), a regular seven-sided star polygon is

formed whose sides approximate w/sin (7r/7). It can be shown that the shortest distance between consecutive vertices approximates w/sin (27r/7).
CONSTRUCTING DODECAHEDRA WITH T(fl/5)
When cash register tape (which is more porous than adding machine
tape) is used with white glue, surprisingly sturdy models of polyhedra may
be made.
To construct a dodecahedron, for example, fold the cash register tape
to obtain T(7r/5) containing at least 90 usable triangles.
t9

and tQ

Cut T(7r/5)

on

_fi, then fold the resulting strip, glueing the overlapping por-

tions in position as shown in Figure 6. Label the edges of the pentagons as

shown. The polyhedron is completed by .first forming a ring-like figure and
glueing one of the shaded parallelograms on top of the other.

Then join the

remaining 18 pairs of edges so that edges labeled with like numbers c o r r e s pond with each other. Tabs for joining the edges may be conveniently obtained by cutting on nineteen successive long transversals of T(7r/5).
If the tabs are labeled so that when they are glued in place it preserves
the numbers shown on each of the edges, it is then possible, upon completion
of the dodecahedron, to observe that
The dodecahedron, formed from

T(TT/5)

of width w, and whose edge

approximates w/sin (27T/5) may be constructed with no fewer than six

1971]

RECREATIONAL MATHEMATICS

211

*P
3\
\
<D
m

K'

4->

\
C0\

\
\

X.

\ <*
CD

4->

\
\
4-)

^N

CN\

P \

iH

212

RECREATIONAL MATHEMATICS

[Apr.

^P

U
rH

1971]

RECREATIONAL MATHEMATICS

213

bands, each of which contains 12 consecutive triangles from T(7r/5).

(In practice, an extra triangle would be required on some bands but,
since it serves only as a tab, it is not counted.)
To see that this is true, take a strip of

T(TT/5)

which contains 12 t r i -

angles and observe that it is possible to position it on the completed dodecahedron so that its short transversals all coincide with edges whose label includes the symbol

l. , T But, it may also be positioned in five other ways so

that its short transversals coincide with the edges each of whose labels include the symbols"2," "3, M M 4, M " 5 , " and

ff

6, n respectively.

Because the

label on every edge contains at least one number, six bands are sufficient for
this particular construction of ihe dodecahedron.

Note that if any one num-

ber were removed from the labels on this dodecahedron, there would be some
edges with no label.

Therefore, at least six bands are necessary for the

construction of this dodecahedron.

This model may be used to show that if a dodecahedron were constructed
from six bands, each containing 12 consecutive triangles from

T(TT/5),

there

would be six edges crossed by exactly two bands and those edges would be
oriented so that (a) their midpoints are the vertices of an inscribed octahedron; (b) the collection of pentagonal diagonals parallel to those six edges
form the edges of an inscribed cube; and, since alternative vertices of a cube
define vertices of a tetrahedron,

(c) the vertices of two distinct inscribed

tetrahedra may be identified on this model.

A second, somewhat different, dodecahedron may be constructed using
T(7r/5). This model is particularly easy to make from gummed tape.

Cash

register tape and white glue produce a better looking model but, having one
side gummed makes the description of the construction easier. Accordingly,
the following instructions are given for gummed tape.
F i r s t , cut from T(fl/5) six strips of 22 triangles each. The first portion of a typical strip is shown, with the gummed side down, in Figure 7.
Label the ungummed side of each of the strips by replacing the letter "X"
shown in Figure 7 with the letters "A,"

B , " ! f C," " D , " " E , " " F , " on the

first, second, third, fourth, fifth, and sixth strips, respectively.

As an

example, the first strip, called "strip A , " will have i* p-leven long transversals labeled " A j , " "A 2 ," 9 "An," consecutively, and all transversals will

214

RECREATIONAL MATHEMATICS

[Apr.

be l a b e l e d with an a r r o w which points to the endpoint of the next long t r a n s versal.

The following notational device i s convenient: If X, Y, Z
m e m b e r s of
then

fA, B , C , D , E , F } and if

TT

X **Y

m, n, k

are natural

represent
numbers,

m e a n s that: the g u m m e d s i d e of s t r i p X i s glued onto the

ungummed side of s t r i p Y so that the t r a n s versal marked

fT

transversal marked

lf

X **~*-Y

" c o i n c i d e s with the

M

Y " and the a r n

r o w s point in the s a m e d i r e c t i o n .
m e a n s that: the g u m m e d side of s t r i p X i s glued onto
the ungummed side of s t r i p
transversal marked

so that the

-V
coincides
m
with the t r a n s v e r s a l m a r k e d n Y - "
and
n
the a r r o w s point in opposite d i r e c t i o n s .
" X m - * . Y n - - > Z k " m e a n s that: X m Y n

and

Yn-+Zk.

Using this notational d e v i c e , the dodecahedron i s a s s e m b l e d a s follows:

E7 - A e
Dr - E 6

III.

II.

A 5 *-* D 8
B 5 <-- E 8

CT - ^ D g

C5w

A8

Br - c 8
A 7 >Bg

D5 ~

B8

E5*-C8

The F s t r i p m a y now be woven in and out so that

D9->F2
F3 ->B4
E9 F4
F

-C

A9 - F 6
F 7 -D4
B9 F8
Fg - E 4
C

*F10

F1I-^F1^A4

19^1]

RECREATIONAL MATHEMATICS
IV.

A3<~+ C10

V. An - > A4 - > E 2

B3 <->D10

B i i - * Bi - * A 2

C3 E10

On Ci - - B 2

D3 <-> A10

Du >Dj C2

E 3 * B10

E1A * Et D2

215

This dodecahedron is also formed from exactly six bands, but each
band contains 20 triangles (not counting the overlapping tabs) from

T(TT/5).

Comparing the two completed polyhedra, one will note many similarities and
differences.

The first and most obvious difference is that the one has some

holes in it and that itappears to be "woven together." A most effective model

of the second dodecahedron may be made if six different colored strips are
used in its construction.

In fact, it is not even necessary to use glue, for

one can hold the various strips together as indicated by the instructions,
with 30 paper clips.

Then, when the dodecahedron is finished, all of the

paper clips, except those six which hold three thicknesses of tape together,
may be removed.
If the places where bands overlap themselves are discounted, all of the
edges of the second dodecahedron are crossed by exactly two bands.

If one

imagines the arrows on this dodecahedron to be roads on which travel is permitted only in the direction of the arrows, it can be seen that, if one leaves
the pentagonal cycle A i i B i i C u D i i E u , all roads lead to the cycle

D 9 F3E 9 F5A9F 7 B 9 F9C 9 F 1 1

and, leaving that cycle, all roads lead to the cycle A7B7C7D7E7, from which
there is no escape.
Many other polyhedra may be constructed with paper strips.

If the

reader wishes to try devising some paper tape constructions for other polyhedra, the following references may be useful.

216

RECREATIONAL MATHEMATICS

April 1971

REFERENCE BOOKS
Ball, W. W. R. , Mathematical Recreations and Essays, rev. by H. S. M.
Coxeter, A paperback, published by The Macmillan Company, New York,
1962.
Beck, A. , and Bleicher, M. N. , and Crowe, D. W. , Excursions into Mathematics, Worth Publishers, New York, 1969.
Coxeter, H. S. M. , Regular Polytopes, Second Ed. , Macmillan Company,
New York, 1963.
Cundy, H. M. , and Rollet, A. P. , Mathematical Models, Second Ed., Oxford
University P r e s s , New York, 1964.
Fejes-Toth, L., Regular Figures, International Series of Monographs in Pure
and Applied Mathematics, Vol. 48, Pergamon P r e s s , New York, 1964.
Jacobs, Harold R. , Mathematics A Human Endeavor, W. H. Freeman and
Company, San Francisco, 1970.
Stover, Donald W. , Mosaics, Houghton Mifflin Mathematics Enrichment
Series, New York, 1966.
Wenninger, Magnus J. , Polyhedron Models for the Classroom,

National

Council of Teachers of Mathematics, Supplementary Publication, Washington, D. C. , 1968.

[Continued from page 135. ]

SPECIAL ADVANCED PROBLEM
H-182S Proposed by Paul Erdos, University of Colorado, Boulder, Colorado.
Prove that there is a sequence of integers nj < n2 <
o-(nk)

satisfying

or(cr(nk))
^00

and

Tr-*-l
a ( )

where

w = dl

(the sum of the integer divisors of n.)

[From Conference on NUMBER THEORY, March 24-27, Washington State
University, Pullman, Washington. ]

ELEMENTARY PROBLEMS AND SOLUTIONS

Edited By
A. P. H1LLMAIM
University of New Mexico, Albuquerque, New Mexico

Send all communications regarding Elementary Problems and Solutions

to Professor A. P. Hillman, Dept. of Mathematics and Statistics, University
of New Mexico, Albuquerque, New Mexico 87106.

Each problem or solu-

tion 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.

Contributors (in the United States) who desire acknowledgement of r e ceipt of their contributions are asked to enclose self-addressed stamped
postcards.
B-208

Proposed by V. Hoggatt, Jr., San Jose State College, San Jose, California.
Let

F0 = 0,

Fl

= 1,

Fn+2 = Fn+1 + Fn,

L 0 = 2,

Ll

= 1, L n + 2 = L n + 1 + L n .

Prove both of the following and generalize:

(a)

F' n + 2 = 3 F ^ + 1 - F*n - 2 ( - l ) n

(b)

I ^ + 2 = 3L* n + 1 - L ^

B-209

10(-l) n .

Proposed by V. Hoggatt, Jr., San Jose State College, San Jose, California.

Do the analogue of B-208 f o r the P e l l sequence defined by

P0 = 0,
B-210

Pl

= 1,

Pn+2 = 2Pn+1

Pn,

and

Qn = P n

P ^ .

Proposed by Guy A. R. Guil/otte, Montreal, Quebec, Canada.

Let Fi = F 2 = 1 and F f t + 2 = F n + 1 + F n where

217

Prove that S > 803/240,

218

ELEMENTARY PROBLEMS AND SOLUTIONS

^i

B-211

F2

[Apr.

F3

Proposed by V. Hoggatt, Jr., San Jose State College, San Jose, California.

th
Let F be the n term in the Fibonacci sequence 1, 1, 2, 3, 5,
Solve the recurrence
D _L1 = 2D + F 0 ^
n+1
n
2n+l
subject to the initial conditions Dj = 1 and D2 = 3.
B-212

Proposed by Tomas Djerverson, Albrook College, Tigertown on the Rio.

Give examples of interesting functions f and g such that

f(m,n) = g(m + n) - g(m) - g(n) .
(One example is f(m ? n) = mn and g(n) = (^ J = n(n - l)/2.)
B-213

Proposed by L Cariitz, Duke University, Durham, North Carolina.

Given n points on a straight line, find the number of subsets (including

the empty set) of the n points in which consecutive points are not allowed.
Also find the corresponding number when the points are on a circle.
SOLUTIONS
A SIXTY-ORDER FIBONACCI-LUCAS IDENTITY
B-190

A repeat of B-186 with a typographical error corrected.

Let L

be the n
L

Lucas number and show that

5n/ L n = fL2n " ^ " ^ V

^2<

'

Solution by Phil Mana, University of New Mexico, Albuquerque, New Mexico.

Since L 5n
and
r

1971]

ELEMENTARY PROBLEMS AND SOLUTIONS

{[L2n-3(-l)n]2

219

(-l)n25F^Ln

satisfy the same sixth-order linear homogeneous recurrence, the result is

proved by verifying it for n = - 2 , - 1 , 0, 1, 2, and 3 (and then relying on
mathematical induction).
Also solved by W. C. Barley, Wray G. Brady, Warren Cheves, Herta T. Freitag, Jo Carol Gordon,
John A. Hitchcock, Edgar Karst, Bob Topley, Andrew Wyatt, Rev. Robert Zuparko, and the Proposer.

THE HUNTER UNVEILED

B-191

Proposed by Guy A. R. Guillotte, Montreal, Quebec, Canada.

In this alphametic, each letter represents a particular but different

digit, all ten digits being represented here.

It must only be that well-known

mathematical teaser from Toronto, J. A. H. Hunter, but what is the value

of HUNTER?
MR
HUNTER
MADE
A_
TEASER
Solution by David Zeitlin, Minneapolis, Minnesota.

The value of HUNTER is 198207, where the unique solution is given by

57
198207
5340
_
3
203607

We note that

220

ELEMENTARY PROBLEMS AND SOLUTIONS

(1)

[Apr.

2R + E + A = 10Cj + R ,

(2)

C1 + M + E + D

= 10C 2 + E ,

(3)

C2 + T + A = 10 C3 + S ,

(4)

C3 + N + M = 10C 4 + A ,

(5)

C4 + U = 10C 5 + E ,

and
(6)

C 5 + H = T,

where C., i = 1, 2, , 5, are carry-overs from the i

column. Since

C 5 = 1, we find from (5) that C4 = 1, with U = 9 and E = 0; and thus,

from (1), that Ct = 1. From (2),

C2 = 1,

All cases for C3 = 1 are non-solutions.

and from (3),

For C3 = 0,

C3 = 0 or 1

the single solution

is obtained when
(A,D f E,H,M,N f R,S,T,U)

= (3,4,0,1,5,8,7,6,2,9) .

Also solved by W. C. Barley, Wray G. Brady, Albert Gommel, Jo Carol Gordon, J. A. H. Hunter,
Edgar Karst, John W. Milsom, C. B. A. Peck, Darla Perry, Azriel Rosenfeld, and the Proposer.

A FOURTH-ORDER F-L IDENTITY

B-192 Proposed by Warren Cheves, Littleton, North Carolina.
Prove that F 3 n = L Q F 2 n - ( - l ) n F n .
Solution by Herta T. Freitag, Hollins, Virginia.

One needs to show that

J I W 2n
-vn,(an - prji,) ,
a 3n - 0p3n = /(an +, ft
)(a
- 0p2n, ) - ,(-1)

where a and p are (1 ' s /5)/2.

1971]

ELEMENTARY PROBLEMS AND SOLUTIONS

221

This, however, is immediately seen by using the relationship a/3 = - 1 ,

and simplifying.
Also solved by W. C. Barley, Wray G. Brady, Mike Franusich, Jo Carol Gordon, John A. Hitchcock,
Stu Hohbs, Edgar Karst, John Kegel, Scott King, John W. Milsom, C. B. A. Peck, Darla Perry, Patricia
Shay, Don C. Stevens, Boh Tepley, Andrew Wyatt, David Zeitlin, Rev. Robert Zuparko, and the
Proposer.

ANOTHER F - L IDENTITY
B-193

Proposed by V. E. Hoggatt, Jr., San Jose State College, San Jose, California.

Show that L , L
is 5F F or L L 9 depending on the choice
&
n+p
n-p
p n
p n
^
of sign on whether p is even or odd.
Solution by John Kegel, Fort Lauderdale, Florida.

It is a well-known fact that

(1)
w

n
un
= 2_lb
n
a - b

and
(2)

LM

= a11 + b n ,

where
a = |(1 +

N/5)

b = | (1 -

N/5)

and

which also gives

222

ELEMENTARY PROBLEMS AND SOLUTIONS

(3)

ab = - 1

and
(a - b)2 = 5 .

(3)
Now

n
L p Ln = (apy + bFp)(a
)(a"n + bb")
)

= an+p+bn+p+apbn

bpan

= Ln+p + (ab) p (a n - p + b n " p )

Thus
L L = L __
, + (-1) P L
p n
n+p
n-p

(4)

Likewise,

5F F
P n

\ a - b / \ a - b /
5 (a n + p + b n + p - a p b n - a n b p )
(a - b) 2
! t L n + p " <a p b p )(a n ~ p + b n " p ) ] ;

Thus
5F F = L
- (-DPL
p n
n+p
n-p

(5)

Hence (4) and (5) give

(6)

L , L
= 5F F
n+p
n-p
p n

or L L
p n

(p odd o r even)
^

[Apr.

1971]

ELEMENTARY PROBLEMS AND SOLUTIONS

223

and the proof is complete,

Also solved by W. C. Barley, Wray G. Brady, Herta T. Freitag, Jo Carol Gordon, David Zeitlin, and
the Proposer.

SECOND ORDER IN n, FIFTH IN k

B-194 Proposed by Phil Mana, University of New Mexico, Albuquerque, New Mexico. (Corrected
Statement).

Show that
L

n +

4k-Ln

5F

k[Fn+3k

-1)kFn+k]-

Solution by C. B. A. Peck, State College, Pennsylvania.

From results of Brother Alfred (Fibonacci Quarterly, VoL 1, No. 4,

p. 55), or H. H. Ferns (Fibonacci Quarterly, VoL VII, No. 1, p. 1), we
have
5F F
U

= IX _, - (-1) V L

1.

U-VJ

L U+V

Replace v by k and u successively by n + 3k and n + k.

second of these identities by (-1)

Multiply the

and add to the first; this gives the (cor-

rected) desired result.

7

Also solved and corrected by Wray G. Brady, Herta T. Freitag, John Kegel, David Zeitlin, and the
Proposer. The error in the statement was also noted by W. C. Barley.

GENERALIZED FIBONOMIALS
B-195 Proposed by David Zeitlin, Minneapolis, Minnesota.

Let JH denote W^'"

n _ r + i/

L L

i 2 '"8

r - Show that

224

ELEMENTARY PROBLEMS AND SOLUTIONS

Apr. 1971

Solution by A. K. Gupta, University of Arizona, Tuscon, Arizona.

From formula (2) (on p. 447 of Fibonacci Quarterly, Vol. 8, No. 4) of
the P r o p o s e r s solution to B-176, we have

(A,
where H

2H

=H1H!H|*2]-2[J1]

[|,

satisfies
H

n+2

n+1

and
n
fLrJ
]

= Hn Hn-1 Hn - r +^ 1 /M 1 HoL Hr .

The desired result is obtained from (A) for H + L

n

and
HtHgHs = L i L 2 L 3 = 2-1-3 = 6

Also solved by W. C. Barley, may G. Brady, Warren Cheves, Herta T. Freitag, John Kegel, John W.
Milsom, and the Proposer.