See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/342093815

Visualize Geometric Series

Preprint · June 2020

CITATIONS READS
0 207

1 author:

Hùng Việt Chu

University of Illinois, Urbana-Champaign
56 PUBLICATIONS   25 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Hùng Việt Chu on 29 September 2020.

The user has requested enhancement of the downloaded file.

 VISUALIZE GEOMETRIC SERIES

HÙNG VIÊ.T CHU

arXiv:2006.05062v1 [math.HO] 9 Jun 2020

A BSTRACT. We review some “proofs without words” for the formula for geometric
series and find a common theme lurking behind them. We also answer negatively the
question raised by Edgar on the existence of other proofs similar to Mabry’s and his.

1. I NTRODUCTION
The well known formula for geometric series
1
1 + x + x2 + x3 · · · = , (1.1)
1−x
for any 0 < x < 1, is very useful due to its ability to convert an infinite sum into a
fraction. A common proof of this formula is as follows:
(1 − x)(1 + x + x2 + x3 + · · · + xn )
= (1 + x + x2 + x3 + · · · + xn ) − (x + x2 + x3 + x4 + · · · + xn+1 ) = 1 − xn+1 .
Hence, we have (1 − x)(1 + x + x2 + x3 + · · · + xn ) = 1 − xn+1 . Dividing both sides
by 1 − x, we arrive at
1 − xn+1
1 + x + x2 + x3 + · · · + xn = .
1−x
Letting n → ∞ and noting that limn→∞ xn+1 = 0 because 0 < x < 1, we have the
desired formula. Though the proof is elegant, it does not give a clear intuition behind
the formula. For this reason, a number of “proofs without words” have been proposed.
The goal of this article is to review proofs without words by various authors, find the
common theme in these proofs, and answer an open question by one of these authors
about the existence of similar proofs.
In 2001, the Viewpoints 2000 Group published Figure 1 to prove (1.1) without words.
In 1999, Mabry [2] published Figure 2 as a visual proof of 41 + ( 14 )2 + ( 41 )3 + · · · = 13 .
Though Mabry’s proof is only for a case of (1.1), the proof provides nice intuition
behind the formula. For convenience, we assume that the area of the outmost triangle
is 1. The idea is to find the colored area in two ways. The first way is to find the ratio
of the colored area in each layer, which is 1/3. The second way is to add up the areas
of all colored triangles, which gives 1/4 + (1/4)2 + (1/4)3 + · · · . The two ways must
give the same answer, so we have the desired identity. Continuing the work, in 2016,
Edgar [1] proved that 94 + ( 94 )2 + ( 94 )3 + · · · = 45 using Mabry’s technique (see Figure
3). At the end of his proof, Edgar asked: “Is it possible to determine which other series
allow ananalogous proof without words?”. We will show that the answer is negative in
Section 2.
1
2 HÙNG VIÊ.T CHU

F IGURE 1. Viewpoints 2000 Group’s proof

··· layer 2

layer 1

F IGURE 2. Mabry’s proof.

F IGURE 3. Edgar’s proof

2. D EMYSTIFY P REVIOUS V ISUAL P ROOFS

In this section, we analyze the technique used by Mabry and Edgar thoroughly. Un-
derstanding the technique is the key to answering the question posed by Edgar and
allows us to connect it with the proof by the Viewpoints 2000 Group to find a common
theme.
2.1. Backbones of the beautiful proofs by Mabry and Edgar.

2010 Mathematics Subject Classification. 40A05.

 VISUALIZE GEOMETRIC SERIES 3

layer 3
layer 2
E D
layer 1
C B

F IGURE 4. Mabry’s and Edgar’s proof in a nutshell

Let n be the total number of triangles in each layer and a be the number of colored
triangles in each layer. Note that 1 ≤ a < n. Also, let r be the ratio of the height of
each layer over the distance from A to the base of that layer. The below analysis uses
Figure 4, which is a sketch of Figures 2 and 3. We assume that △ABC has area 1.
We will compute the colored area, denoted by T , in two ways. Because in each layer,
there are n triangles, a of which are colored, we know that T = a/n. Another way to
compute T is to find the total areas of all colored triangles as follows.
The area of layer 1 is clearly 1 − (1 − r)2 . Hence, each triangle in layer 1 has area
1 − (1 − r)2
,
n
which is also the ratio of the area of each triangle in layer 1 over the area of △ABC
(because the area of △ABC is 1). Due to similarity, the area of each triangle in layer 2
over the area of △ADE is also
1 − (1 − r)2
.
n
Since the area of △ADE is (1 − r)2 , each triangle in layer 2 has area
1 − (1 − r)2
· (1 − r)2 .
n
In general, each triangle in layer n has area
1 − (1 − r)2
· (1 − r)2(n−1) .
n
Therefore, the total colored area is (recall that in each layer, a triangles are colored)
1 − (1 − r)2 1 − (1 − r)2 1 − (1 − r)2
a· +a· (1 − r)2 + a · (1 − r)4 + · · · . (2.1)
n n n
As the two ways of computing T must produce the same result, we have
1 − (1 − r)2 1 − (1 − r)2 2 1 − (1 − r)2 a
a· +a· (1 − r) + a · (1 − r)4 + · · · = .
n n n n
(2.2)
4 HÙNG VIÊ.T CHU

We now apply (2.2) to have the two identities proven by Edgar [1] and Mabry [2]. In
Mabry’s proof, we have (n, a, r) = (3, 1, 1/2). Plugging these numbers into (2.2) gives
 2  3
1 1 1 1
+ + +··· = .
4 4 4 3
In Edgar’s proof, we have (n, a, r) = (5, 4, 1/3). Plugging these numbers into (2.2)
gives
 2  3
4 4 4 4
+ + +··· = .
9 9 9 5
2.2. More beautiful pictures? A natural question is whether we can produce more
beautiful pictures like Mabry’s and Edgar’s for other geometric series of the forms
v + v2 + v3 + · · · . (2.3)
The answer is, unfortunately, negative. To show this, we prove two restrictions on r,
which are simultaneously satisfied only when r is 1/2 or 1/3.
Comparing (2.1) and (2.3) shows that the second term of the geometric series is the
square of the first term; hence, we have
2
1 − (1 − r)2 1 − (1 − r)2

2
a· (1 − r) = a · ,
n n
which is equivalent to
(1 − r)2 a
2
= . (2.4)
1 − (1 − r) n
Because a/n < 1, we know that
1
0.29 < 1 − √ < r < 1. (2.5)
2
This is the first restriction on r.
Furthermore, each triangle in layer 1 has area r 2 , while the area of layer 1 is 1 − (1 −
r) . Therefore, to have pictures like Mabry’s and Edgar’s, we need
2

1 − (1 − r)2
∈ N,
r2
which implies that
2
∈ N. (2.6)
r
Then the number of triangles in each layer is 2/r − 1, or n = 2/r − 1. Plugging this
back to (2.4), we have
(1 − r)2 a
2
= ,
1 − (1 − r) 2/r − 1
which is simplifed to
1 2
a = 2
+1− .
r r
 VISUALIZE GEOMETRIC SERIES 5

Due to (2.6) and a ∈ N, it follows that r = 1/m for some m ∈ N. This is the
second restriction on r. Combining this restriction with (2.5), we know that m ∈ {2, 3}.
Therefore, we answer negatively the question raised by Edgar [1] on whether there are
more similar “proofs without words” for other series.

3. T HE COMMON THEME

At the first glance, the proof by the Viewpoints 2000 Group and the proof by Mabry
are little related. However, there is a common theme lurking behind them. Figure 5 is
obtained by repositioning √ the Viewpoints 2000 Group’s. For convenience, we substitute
a = 1 and replace r by r.

1√
1− r A

layer 3
r
layer 2

1
r2
1 layer 1

B
C 1 1√
1− r

F IGURE 5. Repositioning the Viewpoints 2000 Group’s picture

We can now apply the Mabry’s technique to prove the formula for geometric series. The
colored area is clearly 21 + 12 r + 12 r 2 + 21 r 3 + · · · . On the other hand, in each layer, the
colored area is 1 1 the area of the layer. Hence, the colored area in total is 1 1 the
1+r 2 1+r 2
1
area of △ABC, which is 1 . Therefore, we have
2(1−r 2 )

1 1 1 1 1 1
+ r + r2 + r3 + · · · = 1 · 1 ,
2 2 2 2 1 + r 2 2(1 − r 2 )
which is simplified to
1
1 + r + r2 + r3 + · · · = .
1−r
 6 HÙNG VIÊ.T CHU

R EFERENCES
[1] T. Edgar, Proof without words: sum of powers of 49 , Math. Mag. 89 no.3 (2016) 191.
[2] R. Mabry, Proof without words: 14 + ( 14 )2 + ( 14 )3 + · · · = 13 , Math. Mag. 72 no.1 (1999) 63.
[3] The Viewpoints 2000 Group, Proof without words: geometric series, Math. Mag. 74 no.4 (2001) 320.
E-mail address: [email protected]

D EPARTMENT OF M ATHEMATICS , U NIVERSITY OF I LLINOIS AT U RBANA -C HAMPAIGN , U RBANA ,

IL 61820, USA

View publication stats