Feynman
Feynman
Feynman
C'
U
A'
V
W
A
p -1
B'
Figure 1
As pointed out by De Villiers (2005) and others in feedback in a later issue of the
Mathematical Gazette, this delightful result can be generalised as follows.
This is an extension of De Villiers, M. (2005). Feedback: Feynman's Triangle. The Mathematical Gazette, 89
(514), March 2005, p. 107. All rights reserved by the Mathematical Association, http://www.m-a.org.uk
Theorem 1
For a triangle in the plane, if each vertex is joined to the point
1
(p > 2) along the opposite
p
side (measured say anti-clockwise), then the area of the inner triangle formed by these lines
is exactly
( p 2 )2
p2 p + 1
4
3
, for p = 5, it is , etc. Probably the easiest way to derive
13
7
this formula is to generalise the affine proof given in Cook & Wood (1994) as follows.
Proof
Consider an equilateral triangle ABC as shown in Figure 1. Let AB = BC = CA = p, thus C'B
= 1. Then by the cosine rule CC 2 = p2 + 1 2 pcos60 = p 2 p 1. Therefore,
CC =
p p +1
2
p
=
p p +1
2
1
, BU (= CV) =
p p +1
2
p
, and so VU =
p p +1
2
p(p 2)
. By symmetry, UVW is also an
p2 p + 1
p( p 2)
1
=
2
p p +1 p
p 2
of the sides of ABC.
p p +1
2
( p 2)2
of the area ABC.
p2 p + 1
Since the theorem only involves affine properties such as ratios of lines and areas, and
every triangle is affinely equivalent to an equilateral triangle the general result follows.
Feynman's triangle can also be generalized to a parallelogram as pointed out be De Villiers
(2005) by observing as follows that the side opposite a vertex in a triangle is the same as
the alternate side from that vertex:
Theorem 2
For a parallelogram in the plane, if each vertex is joined to the point
1
(p 2) along the
p
alternate side (measured say anti-clockwise), then the area of the inner parallelogram
This is an extension of De Villiers, M. (2005). Feedback: Feynman's Triangle. The Mathematical Gazette, 89
(514), March 2005, p. 107. All rights reserved by the Mathematical Association, http://www.m-a.org.uk
p2 2 p + 1
of the area of the initial parallelogram
p2 + 1
D'
E
A'
C'
G
F
B
B'
Figure 2
So for p = 2, the fraction is
1
2
9
, for p = 3, it is , for p = 4, it is
, etc. Since a square is
5
5
17
affinely equivalent to a parallelogram, the easiest way to derive and prove this formula is to
consider the special case of a square.
Exercise
1.
1
(p 2)
p
(measured say anti-clockwise), then find (and prove) a formula relating the area of
ABC with that of A'B'C' (triangle formed by these points). See Figure 3.
B
C'
A'
C
B'
Figure 3
This is an extension of De Villiers, M. (2005). Feedback: Feynman's Triangle. The Mathematical Gazette, 89
(514), March 2005, p. 107. All rights reserved by the Mathematical Association, http://www.m-a.org.uk
2.
1
(p 2)
p
(measured say anti-clockwise), then find (and prove) a formula relating the area of
ABCD with that of A'B'C'D' (parallelogram formed by these points). See Figure 4.
D'
A'
C'
B'
Figure 4
Note
Sketchpad sketches for exploring and demonstrating all four results and problems can be
downloaded in zipped format from:
http://mysite.mweb.co.za/residents/profmd/feynman.zip
These dynamic geometry figures can also be viewed and manipulated with the Free Demo
of Sketchpad which can be downloaded from:
http://www.keypress.com/sketchpad/sketchdemo.html
References
Cook, R.J. & Wood, G.V. (2004). Note 88.46: Feynman's Triangle. The Mathematical
Gazette, July, Vol. 88, No. 512, pp. 299-302.
De Villiers, M. (2005). Feedback: Feynman's Triangle. The Mathematical Gazette, 89
(514), March 2005, p. 107.