Scribd
Upload
22 views

+ 4 + 2 + I I 1 + I I 2 + 3: University of Santa Clara, Santa Clara CA 95053

Maths.

Uploaded by

Charlie Pinedo
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
22 views

+ 4 + 2 + I I 1 + I I 2 + 3: University of Santa Clara, Santa Clara CA 95053

Maths.

Uploaded by

Charlie Pinedo
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

[Aug.

276

AN IMPLICIT TRIANGLE OF NUMBERS


DAVE LOGOTHETTI

University

of Santa Clara,

Sketch reprinted from California

Santa Clara CA 95053

Mathematics

5 ( 2 ) , October, 1980.

To Voxn Hoggattt whole common AenAe, plain language,


energetic
enthiutaAm brought real mathematics into
Lives ol div&tte people throughout the would.

and
the

This elementary note introduces a new triangle of numbers that is implicitly


defined in PascalTs Triangle. It shares many properties with Pascal's Triangle,
including the generation of Fibonacci numbers. It differs from Pascalfs Triangle in that it is not symmetrical (and therefore is not a special case of the
Fontene-Ward Triangle [I]). When I asked Vern Hoggattwho seemed to know everything there is to know about Pascalfs Triangleabout the Implicit Triangle,
he surprised me by replying that he did not know of either the triangle or any
of its properties. Therefore, the following may add to our readers1 list of
"Neat Little Facts about Integers."
The question that led to the discovery of the Implicit Triangle is: "How
do we get the squares out of Pascalfs Triangle?" One fairly well-known way is
to note that
0 + 1 = 1, 1 + 3 = 4, 3 + 6 = 9, ..., (" 2 X ) + (2)

"*'

This can be generalized using Eulerian numbers, so that


0 + 4(0) + 1 = 1, 0 + 4(1) + 4 = 8 , 1 + 4(4) + 10 = 27, ...,

()+4(^)+("32)--)

+ ii

("t1)+iiri2)+(":3)-'-

etc. See [2], for example. But there is another way to get the squares out of
Pascal's Triangle, and this is not so well known:

( ? ) ' ( S M I ) ( ; ) - ( ! ) (;)) *u)-

(j)+a)*(?)*a)-'.-.(":,)+("2V(T)+(")--'-

1981]

AN IMPLICIT TRIANGLE OF NUMBERS

277

These squares are generated by adding rhombuses of entries from Pascal1s


Triangle. By adding other rhombuses, we generate our new triangle:
0 1 - -11
/ 2 / 1 / 0 /
0100
0
/ 2 / 3 / 1
/ 0 /
0-1
1
0
00
/ 2 / 5 / 4 / 1 / 0 /
0 1-2
1
000
/ 2 / 7 / 9 / 5 / 1 / 0 /
0
1
3 _ 3 _ i
o0 0
/ 2 / 9 /16/14/ 6 / 1/ 0 /
0
-1 4 6 4
10 0 0
/ 2 /ll /25/30/20/ 7/ 1 / 0 /
0
I_5_IO105
10
0
0
/ 1 / 1 3 / 3 6 / 5 5 / 5 5 / 27 / 8 / 1 / 0 /
0
1- 615 2015
6 1 0 00
/ 2 /15/49/91 /105/77/35/ 9 / 1 / 0 /
0-1
721353521
7
1 0-00
/ 2 /17 /64/l40/l96/l82/ll2/44/l0/ 1 / 0 /
0r8-2856705628
8 1 0 00
/ 2 / 19/81/204/336/378/294/156/54/11 / 1 / 0 /
0 r 9 3684-126-12684 36 9 1 0 0
0
Suppressing the entries from Pascal's Triangle, we get the (almost) triangular
array:
2
1
2

2
2
2
2
2

11
13

15
17

19

64
81

5
14

30
55

91

140

25

49

16

36

6
20

50

105

196

1
1
7
27

77

182

8
35

112

204 336 378 294

1
1
9
44

156

1
10

54

1
11

This Implicit Triangle has the generating formula

where J(n, k) is the Implicit Triangle entry in the nth row, kth diagonal,
n = 1, 2, 3, ..., Zc = 0, 1, 2, 3, ... .
(The zeroth row is missing from this new triangle.)

278

AN IMPLICIT TRIANGLE OF NUMBERS

[Aug.

Although it lacks the symmetry of Pascalfs Triangle, the Implicit Triangle


shares many of its properties.
Tkzonem 1:

I(n

- 1, k - 1) + I(n

- 1, k) = I(n9

k).

Psioofi:- This version of Pascal s Identity follows from that identity in


Pascal's Triangle.

-i.>'-+*<-i.-{'l:l)*ti:l)*(i:l)

+ (r-3

+ e: ?M\-VC:: D+ (V)
-(Z:a*'(Z:iHi;:;MV)
-(:?) ( V M S :;)("*-')
= T(n, Zc).
This is not really surprising, since the Implicit entries are linear combinations of Pascal entries, and these linear combinations carry along the properties of Pascal's Triangle.
Tko.on.2m 2 {"ChAAJi>m(U Stocking Tke.on.em")'*
k+r

Y, ifa, k) = i(k + v + 1, k + i).


n=k

Tke.onem 3 ["Hockey Stick

Tke.on.em")
k=n

I(n9

P) =

( - l ) k " 2 , " 1 J ( n + 1,

k).

k=r+l

Tkconejn 4 ["Ftboncicct UumbeJi. Tke.on.ejm"):

I(n

- k, k)

=Fn+2.

k=0

( exploits the fact that proceeding up a diagonal we eventually get all O's.)
Tkconem 5 (" Attcnncuting

Row Sam Tkconem"):

n+1

(-l)*J(n, k) = 0, n = 2, 3, 4, ... .
k=o

Vnoo^k*- All of these theorems follow from the fact that the Implicit entries are linear combinations of the Pascal entries.
And then there are properties different from, but analogous to, properties
of Pascal's Triangle. For examples,
Tkconem 6 {"Luc<u Hambcn Tkconem");
]T l(n
k=0

- ks n + 1 - Ik) =

Lk+1.

1981]

AN IMPLICIT TRIANGLE OF NUMBERS

279

Thtohm 7 ["Row Sam Tkzotim"):


k) = 2 n " 1 (3).

J^I(n9
k=o

Vtioofa.
Both of t h e s e theorems may be proved j u s t a s t h e i r a n a l o g u e s a r e
proved f o r P a s c a l ' s T r i a n g l e . Theorem 7 may be proved v e r y e a s i l y w i t h t h e a i d
of Theorem 8.
ThdOKOm S ["Co&&&sL(U.2.nt Tkzotiejfn")'
e x p a n s i o n of (2x + 1) (as + l ) n _ 1 e

I(n9

k)

i s the coefficient

of xn~k

in the

VK.OO^i From the identity

we can see that the Implicit Triangle is formed from the binomial coefficients
of two overlapping Pascal Triangles:

- = U)+(V)-

The theorem then follows from the fact that


(2a? + l)(x

+ l ) n ~ 1 = x(x + l ) n _ 1 + (x + l ) n .

We are now in a position to look at a Generalized Implicit Triangle:


1

(a + 1)

a
a
a
a
a

(3a + 1)

(4a + 1)

(5a + 1)

(2a + 1)

(a + 2)

(3a + 3)

(6a + 4)

(a + 3)

(4a + 6)

(a + 4)

(10a + 5) (10a + 10) (5a + 10)

(a + 5)

Here the generating identity is


G(n9

k) = G(n - 1, k - 1) + G(n - 1, k)

G(n9 0) = a9 G(rc, n) = 1;

for a = 1, this is just Pascal s Identity.

IkdoK^m 9 (" Generalized dodijldloiit

xn~k

Thejotiom")'- G(n9 k) is the coefficient of

in the expansion of (ax + 1) (x + I ) " " 1 .

P/LOO^: The Generalized Implicit Triangle is again just the overlap of Pascal flT7frTangle and Pascal's Triangle with every entry multiplied by a - 1. The
theorem follows from the identity
(ax + l)(x

+ I)"1-1

= (x + l) n + (a - l)x(x

+ l)n-1.

Since each entry of the Generalized Implicit Triangle is a linear combination of entries from Pascal's Triangle9 those foregoing theorems whose proofs
were based on linear combinations will hold in the general cases with appropriate modifications; for example5 the row sums will be of the form 2n~ (a + 1).

280

FRACTIONAL PARTS (nr - s), ALMOST ARITHMETIC SEQUENCES,


AND FIBONACCI NUMBERS

[Aug.

Had Vern Hoggatt been able to coauthor this article he would no doubt have
found many more results. Perhaps our readers will celebrate his memory by
looking for further results themselves.
REFERENCES
1.
2.

H. W. Gould. "The Bracket Function and Fontene-Ward Generalized Binomial


Coefficients with Application to Fibonomial Coefficients." The
Fibonacci
Quarterly
1 (1969):23-40.
Dave Logothetti. "Rediscovering the Eulerian Triangle." California
Mathematics 4 (1979):27-33.

FRACTIONAL PARTS (nr

s), ALMOST ARITHMETIC SEQUENCES,

AND FIBONACCI NUMBERS


CLARK KIMBERLING

University

of Evansville,

Evansville

IN 47702

To tkd mmosiy o{ \Zznn HoQQeutt, wiAk gsicutvtude, and adbruAxutxon.


Except where noted otherwise, sequences ian}, {bn}, and {cn}
to satisfy the following requirements, as stated for {an}i

are understood

(i) the indexing set {n} is the set of all integers;


(ii) an is an integer for every n;
(iii) {an} is a strictly increasing sequence;
(iv) the least positive term of {an} is a.
We call {an} almost arithmetic
if there exist real numbers u and B such that
(1)

\an - un\ < B

for all n, and we write an ^ un if (1) holds for some B and all n,
Suppose r is any irrational number and s is any real number. Put
cm = [mr - s] = the greatest integer less than or equal to mr -

s,

and let b be any nonzero integer. It is easy to check that cm + b - cm = [br],


if (mr - s) < (-br), and = [br] + 1, otherwise.
Let an be the nth term of the sequence of all m satisfying cm+b~ cm= [br].
In the following examples, r = (1 + /~5)/2, the golden mean, and s - 1/2.
Selected values of m and cm arei
-5, -4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16
-9, - 7 , -6, - 4 , -3,-1, 1, 2, 4, 5, 7, 9,10,12,14, 15, 17, 18, 20, 22, 23, 25.
When b = 1 we have [br] = 1, and selected values of n and an are:
-1, 0, 1, 2, 3, 4, 5, 6
-4, -2, 1, 3, 6, 9, 11, 14.
When b = 2 we have [2?r] = 3, and selected values of n and a n are:
-4, -3, - 2 , -1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11
-5, -4, -3, -2, 0, _1_, 2., _3, 5_9 6, 81, 9, 10, 11, 13_, 14.
Note here the presence of Fibonacci numbers among the an.
Methods given in
this note can be used to confirm that the Fibonacci sequence is a subsequence
of {an} in the present case.

You might also like