International Journal of Computer Discovered Mathematics (IJCDM)

ISSN 2367-7775 IJCDM

c
November 2015, Volume 0, No.0, pp.32-48.
Received 12 July 2015. Published on-line 15 September 2015
web: http://www.journal-1.eu/
The
c Author(s) This article is published with open access1.

Barycentric Coordinates

Francisco Javier García Capitán

I.E.S. Alvarez Cubero, Priego de Córdoba, Spain
e-mail: garciacapitan@gmail.com
web: http://garciacapitan.99on.com/

Abstract. We give a short introduction to barycentric coordinates.

Keywords. barycentric coordinates, triangle geometry.

Mathematics Subject Classification (2010). 51-04.

1. Barycentric coordinates with respect to a triangle.

1.1. Homogeneous barycentric coordinates. Let ABC be a triangle and

u, v, w ∈ R such that u + v + w 6= 0. For any point O, let P be the point
−→ −→ −−→ −→
on the plane such that (u + v + w)OP = uOA + v OB + wOC. We can prove that
−−→ −−→ −−→ −−→
P does not depend on O. For, if (u + v + w)O0 P 0 = uO0 A + v O0 B + wO0 C, then
−−→ −−→ −−→ −→ −−→ −→ −−→ −→
(u + v + w)(O0 P 0 − OP ) =u(O0 A − OA) + v(O0 A − OA) + w(O0 A − OA) =
−−→
=(u + v + w)O0 O,
−−→ −−→ −→ −−→
therefore O0 P 0 = O0 O + OP = O0 P and P 0 = P .

Hence we can define P as the center of mass of the system formed by the points
A, B, C with weights u, v, w.

The barycentric coordinates of a point P with respect to the triangle ABC is a

list (x : y : z) of three numbers such that
x : y : z = (P BC) : (P CA) : (P AB).

Now we prove that P is the center of mass of the system A, B, C with weights
x, y, z.
1This article is distributed under the terms of the Creative Commons Attribution License
which permits any use, distribution, and reproduction in any medium, provided the original
author(s) and the source are credited.
32
 Francisco Javier Garcı́a Capitán 33

A
−−→ −→ −−→ −→ m − −→
AD =AB + BD = AB + m+n BC,
−→ s −−→ s −→ sm −−→ s
AP = s+t AD = s+t AB + (s+t)(m+n) BC,
−→ −→  −−→ −→  −→ −−→ P
s sm
OP − OA = s+t OB − OA + (s+t)(m+n) OC − OB ,
t
−→ t −→ sn −−→ sm −→ m n
OP = s+t OA + (s+t)(m+n) OB + (s+t)(m+n) OC.
B D C
Then P is the center of mass of the system formed by A, B, C with weights
t(m + n), sn and sm. But,
sn DC (ADC) (P DC) (ADC) − (P DC) 4(P CA)
= = = = = .
sm BD (ABD) (P BD) (ABD) − (P BD) (P AB)

This proves that y : z = (P CA) : (P AB). In the same way, we can prove that
x : y : z = (P BC) : (P CA) : (P AB).

Examples.

A A A

O
G I

B C
B C B C
(1) The centroid G has homogeneous barycentric coordinates (1 : 1 : 1), be-
cause the areas GBC, GCA and GAB are equal to each other.
(2) The incenter I has coordinates a : b : c, because the areas of the trian-
gles IBC, ICA e IAB are respectively 12 ar, 21 br and 12 cr, where r is the
inradius.
(3) The circuncenter O. If R is the circumradius, the coordinates of O are
(OBC) : (OCA) : (OAB) =
= 12 R2 sin 2A : 12 R2 sin 2B : 21 R2 sin 2C =
= sin A cos A : sin B cos B : sin C cos C =
b 2 + c 2 − a2 c 2 + a2 − b 2 a2 + b 2 − c 2
=a · =b· =c· =
2bc 2ac 2ab
=a2 (b2 + c2 − a2 ) : b2 (c2 + a2 − b2 ) : c2 (a2 + b2 − c2 ).
(4) The points on line BC have coordinates of the form (0 : y : z). In the
same way, points on lines CA and AB have coordinates (x : 0 : z) and
(x : y : 0), respectively.

Exercises.
(1) Show that the sum of coordinates of the circumcenter equals to 4S 2 , where
S is twice the area of ABC.
34 Barycentric coordinates

We consider the Heron formula for the area of the triangle and do some
algebraic manipulation:
a2 (b2 + c2 − a2 ) + b2 (c2 + a2 − b2 ) + c2 (a2 + b2 − c2 ) =
=2a2 b2 + 2a2 c2 + 2b2 c2 − a4 − b4 − c4 =
=(a + b + c)(−a + b + c)(a − b + c)(a + b − c) =
a + b + c −a + b + c a − b + c a + b − c
=4 · 4 · · · · =
2 2 2 2
=4S 2 .
(2) Find the coordinates of the excenters.
We consider the following figure that shows the triangle ABC and the
excenter Ib .

A rb

Ib

B C
The barycentric coordinates of Ib are
(Ib BC) : (Ib CA) : (Ib AB) = arb : −brb : crb = a : −b : c.
We observe that the orientation of triangle Ib CA is opposite the orienta-
tion of the other two, giving the negative second coordinate. In the same
way, we have Ia = (−a : b : c) and Ic = a : b : −c.

1.2. Absolute barycentric coordinates. Let P be a point with (homogeneous

barycentric) coordinates (x : y : z). If x+y+z 6= 0, we get the absolute coordinates
of P by normalizing the coefficients so that their sum becomes 1:
x·A+y·B+z·C
P = .
x+y+z
If we have absolute coordinates of P and Q, the point that divides the segment
P Q in the ratio P X : XQ = p : q has absolute coordinates qPp+q +pQ
. However,
since it is more convenient avoiding denominators in our calculations, we adapt
the previous formula in the following way:
If P = (u : v : w) and Q = (u0 : v 0 : w0 ) are homogeneous barycentric coordinates
satisfying u + v + w = u0 + v 0 + w0 , then the point X that divides P Q in the ratio
P X : XQ = p : q has homogeneous coordinates (qu + pu0 : qv + pv 0 : qw + pw0 ).

Exercises.
(1) The orthocenter lies on Euler lines and divides OG in the ratio 3 : −2.
Prove that their barycentric coordinates can be written
H = (tan A : tan B : tan C),
 Francisco Javier Garcı́a Capitán 35

or equivalently,
 
1 1 1
H= : 2 : 2 .
b + c − a c + a − b a + b2 − c 2
2 2 2 2 2

We have seen that

O = a2 (b2 + c2 − a2 ) : b2 (c2 + a2 − b2 ) : c2 (a2 + b2 − c2 ) ,

G = (1 : 1 : 1),
with sum 4S 2 and 3 respectively. We first consider these coordinates with
equal sum 12S 2 :
O = 3a2 (b2 + c2 − a2 ) : 3b2 (c2 + a2 − b2 ) : 3c2 (a2 + b2 − c2 ) ,

G = (4S 2 : 4S 2 : 4S 2 ).
Now the first coordinate of H is
(−2)3a2 (b2 + c2 − a2 ) + 3 · 4S 2 =
= − 6(a2 b2 + a2 c2 − a4 ) + 3(2a2 b2 + 2a2 c2 + 2b2 c2 − a4 − b4 − c4 ) =
=3a4 − 3b4 + 6b2 c2 − 3c4 =
=3(a2 − b2 + c2 )(a2 + b2 − c2 ).
In the same way, the second and third coordinates are 3(−a2 + b2 +
c )(a2 + b2 − c2 ) and 3(−a2 + b2 + c2 )(a2 − b2 + c2 ). If we divide them by
2

(−a2 + b2 + c2 )(a2 − b2 + c2 )(a2 + b2 − c2 )

we get
 
1 1 1
H= : 2 : 2 .
b + c − a c + a − b a + b2 − c 2
2 2 2 2 2

Now,
1 sin A
1 S = tan A ,
= 2 2bc
2 2 =
2 2
b +c −a 2 b +c −a cos A S
2bc
threfore
H = (tan A : tan B : tan C).
(2) Use that the nine point center N divides OG in the ratio 3 : −1 to show
that their barycentric coordinates can be written as
N = (a cos(B − C) : b cos(C − A) : c cos(A − B)).
Starting from
O = 3a2 (b2 + c2 − a2 ) : 3b2 (c2 + a2 − b2 ) : 3c2 (a2 + b2 − c2 ) ,

G = (4S 2 : 4S 2 : 4S 2 ),
the first coordinate of N is
(−1)3a2 (b2 + c2 − a2 ) + 3 · 4S 2 =
= −3a2 b2 − 3a2 c2 + 3a4 + 3(2a2 b2 + 2a2 c2 + 2b2 c2 − a4 − b4 − c4 )
= 3(a2 b2 + a2 c2 + 2b2 c2 − b4 − c4 ).
36 Barycentric coordinates

To arrive the desired result, we use the formulas

a2 + c 2 − b 2
cos B = , S = ac sin B
2ac
(and the corresponding ones for angle C). Then,
cos(B − C) = cos B cos C + sin B sin C =
a2 + c 2 − b 2 a2 + b 2 − c 2 S S
= · + =
2ac 2ab ac ab
(a2 + c2 − b2 )(a2 + b2 − c2 ) + 4S 2
= =
4a2 bc
2a2 b2 + 2a2 c2 + 4b2 c2 − 2b4 − 2c4
= =
4a2 bc
1 a2 b2 + a2 c2 + 2b2 c2 − b4 − c4
= · .
a 2abc
Thus we have
N = (a cos(B − C) : b cos(C − A) : c cos(A − B)).

2. Conway formula

2.1. Notation. If θ is any angle and S is twice the area of ABC, we define
Sθ = S cot θ. In particular,
b 2 + c 2 − a2 c 2 + a2 − b 2 a2 + b 2 − c 2
SA = , SB = , SC = .
2 2 2
Given two angles θ and φ, we define Sθφ = Sθ · Sφ .
In this notation, we have the formulas :
(1) SB + SC = a2 , SC + SA = b2 , SA + SB = c2 .
(2) SAB + SBC + SCA = S 2 .
The first one is easy. The second one comes from the identity
cot A cot B + cot B cot C + cot C cot A = 1.
To prove this,
cot A cot B + cot B cot C + cot C cot A =
= cot A (cot B + cot C) + cot B cot C =
 
cos A cos B cos C cos B cos C
= + + · =
sin A sin B sin C sin B sin C
cos A sin C cos B + sin B cos C cos B cos C
= · + · =
sin A sin B sin C sin B sin C
cos A sin(B + C) cos B cos C
= · + · =
sin A sin B sin C sin B sin C
cos A cos B cos C
= + =
sin B sin C sin B sin C
cos B cos C − cos(B + C)
= =
sin B sin C
sin B sin C
= = 1.
sin B sin C
 Francisco Javier Garcı́a Capitán 37

Examples.
(1) The orthocenter has coordinates
 
1 1 1
: : = (SBC : SCA : SAB ) .
SA SB SC
(2) The circumcenter has coordinates

a2 SA : b2 SB : c2 SC = (SA (SB + SC ) : SB (SC + SA ) : SC (SA + SB )) .

In this way, the sum of coordinates is 2(SAB + SBC + SCA ) = 2S 2 .

2.2. Conway formula.

Given a triangle ABC and a point P , we consider the A

swing angles of P with respect to BC as θ = ∠CBP and
ϕ = ∠BCP , considered in the range − π2 6 θ, ϕ 6 π2 . The
angle θ is positive or negative according as angles ∠CBP
and ∠CBA have different or the same orientation. For
example, θ is taken positive if ∠CBP and ∠CBP have
different or the same orientation.
The Conway formula gives the barycentric coordinates of θ ϕ
B C
a point from its swing angles θ and ϕ:
P = (−a2 : SC + Sϕ : SB + Sθ ). P
To prove this, we use the sine theorem to triangle P CB,
BP CP a a sin ϕ a sin θ
= = ⇒ BP = , CP = .
sin ϕ sin θ sin(θ + ϕ) sin(θ + ϕ) sin(θ + ϕ)
The area of the triangle P BC is
1 a2 sin θ sin ϕ
(P BC) = · BC · BP · sin θ = .
2 2 sin(θ + ϕ)
We can calculate the areas (P CA) and (P AB) in a similar way, giving:
(P BC) : (P CA) : (P AB) =
a2 sin θ sin ϕ ba sin θ sin(ϕ + C) ca sin ϕ sin(θ + B)
=− : : =
2 sin(θ + ϕ) 2 sin(θ + ϕ) 2 sin(θ + ϕ)
ab sin(ϕ + C) ac sin(θ + B)
= − a2 : : =
sin ϕ sin θ
= − a2 : ab cos C + ab sin C cot ϕ : ca cos B + ac sin B cot θ =
= − a2 : SC + Sϕ : SB + Sθ .

2.3. Examples.

2.3.1. Square constructed on a side of the triangle.

38 Barycentric coordinates

A
We consider the square BCCA BA constructed externally
on the side BC of the triangle ABC. The point BA
has swing angles θ = 90◦ and ϕ = 45◦ , therefore since
cot 90◦ = 0 and cot 45◦ = 1, we have
BA =(−a2 : SC + Sϕ : SB + Sθ ) =
(−a2 : SC + S : SB ). B 45º C
Similary,
CA = (−a2 : SC : SB + S).
We caln calculate the midpoint of the square as the mid- MA
point of BA and C = (0 : 0 : S), giving
MA = (−a2 : SC + S : SB + S).
BA CA

2.3.2. Equilateral triangles constructed on a side.

√ triangle BCX externally on

If we erect the equilateral
◦
BC, since cot 60 = 1/ 3, we have
 
2 S S
X = −a : SC + √ : SB + √ .
3 3 A
Similarly, if we erect equilateral triangles CY A and AZB
on CA and AB,
 
S 2 S
Y = SC + √ : −b : SA + √ ,
3 3
 
S S 2
Z = SB + √ : SA + √ : −c .
3 3 B 60º 60º C
If we erect equilateral triangles BX C, CY 0 A, AZ 0 B in-
0

ternally, we get the points

 
0 2 S S
X = −a : SC − √ : SB − √ ,
3 3
 
0 S 2 S X
Y = SC − √ : −b : SA − √ ,
3 3
 
0 S S 2
Z = SB − √ : SA − √ : −c .
3 3

3. Cevians and traces

We call cevians of a point P = (x : y : z) the lines joining A
P with the vertices of the triangle.
The intersections AP , BP , CP of these cevians and the
sides of the triangle are called the traces of P . CP
BP
P
The coordinates of the traces are easily deduced from
their geometrical meaning:
AP = (0 : y : z), BP = (x : 0 : z), CP = (x : y : 0). B AP C
 Francisco Javier Garcı́a Capitán 39

3.1. Ceva theorem. Three points X, Y, Z on BC, CA and AB respectively are

the traces of a point if and only if they have coordinates of the form X = (0 : y : z),
Y = (x : 0 : z) and Z = (x : y : 0) for some x, y, z.

3.2. Examples.

3.2.1. Gergonne point. The points of tangency of the incircle with the sides are
X = (0 : s − c : s − b), Y = (s − c : 0 : s − a) and Z = (s − b : s − a : 0) that can
1 1 1 1 1 1
be rewritten as X = (0 : s−b : s−c ), Y = ( s−a : 0 : s−c ) and Z = ( s−a : s−b : 0).
1 1
Therefore AX, BY and CZ intersect at the point with coordinates ( s−a : s−b :
1
s−c
). This point is known as the Gergonne point Ge of triangle ABC.

3.2.2. Nagel point. The points of tan-

gency of the excircles and the correspon-
ding sides have coordinates A

X 0 = (0 : s − b : s − c),
Y 0 = (s − a : 0 : s − c), Z' Y'
N
0
Z = (s − a : s − b : 0).
These are the traces of the point with B X' C
coordinates (s − a : s − b : s − c), known
as Nagel point Na of triangle ABC.

3.2.3. Exercise. Show that the Nagel point Na lies on line IG and Na divides IG
in the ratio 3 : −2.
We consider I = (3a : 3b : 3c) y G = (2s : 2s : 2s) both with sum 6s. Then,
−2I + 3G = (6s − 6a : 6s − 6b : 6s − 6c) = (s − a : s − b : s − c) = Na .

3.2.4. Fermat Points. We have calculated the coordinates of the points X, Y, Z

such that BXC, CY A and AZB are equilateral triangles constructed externally
on the sides BC, CA, AB of the triangle ABC.
These coordinates can be written as
!
1 1
X= ∗∗∗: S
: S
,
SB + √
3
SC + √
3
!
1 1
Y = S
:∗∗∗: S
,
SA + √
3
SC + √
3
!
1 1
Z= S
: S
:∗∗∗ ,
SA + √
3
SB + √
3

where the asterisks substitute quantities whose value is not needed.

The lines AX, BY , CZ concur at the point
!
1 1 1
F+ = S
: S
: S
,
SA + √
3
SB + √
3
SC + √
3
40 Barycentric coordinates

known as first Fermat point.

If we consider equilateral triangles constructed internally on the sides, we get the

second point of Fermat:
!
1 1 1
F− = S
: S
: S
.
SA − √
3
SB − √
3
SC − √
3

From

S

cos A 1

S √ 
SA − √ =S ±√ =√ 3 cos A ± sin A =
3 sin A 3 3 sin A
√ !
2S 3 1
=√ cos A ± sin A =
3 sin A 2 2
2S
=√ (sin 60◦ cos A ± cos 60◦ sin A) =
3 sin A
2S sin(60◦ ± A)
=√ =
3 sin A
4RS sin(60◦ ± A)
=√ ,
3 a

the coordinates of Fermat points can be written as follows:

 
a b c
F+ = : : ,
sin(60◦ + A) sin(60◦ + B) sin(60◦ + C)
 
a b c
F− = : : .
sin(60◦ − A) sin(60◦ − B) sin(60◦ − C)

4. Area of a triangle

If P = (x1 , y1 ), Q = (x2 , y2 ) and R = (x3 , y3 ) are three points on the plane, we

know that the area (P QR) of the triangle P QR is given by
 
 1 1 1 
1 
(P QR) =  x1 x2 x3  .
2 y y y 
1 2 3

If the barycentric coordinates of P , Q and R with respect to the triangle ABC

are P = (u1 : v1 : w1 ), Q = (u2 : v2 : w2 ) and R = (u3 : v3 : w3 ), then

(u1 + v1 + w1 )P = u1 A + v1 B + w1 C,
(u2 + v2 + w2 )Q = u2 A + v2 B + w2 C,
(u3 + v3 + w3 )R = u3 A + v3 B + w3 C,
 Francisco Javier Garcı́a Capitán 41

If we put A = (r1 , s1 ), B = (r2 , s2 ) and C = (r3 , s3 ), these equations can be

written as
(u1 + v1 + w1 )x1 = u1 r1 + v1 r2 + w1 r3 ,
(u1 + v1 + w1 )y1 = u1 s1 + v1 s2 + w1 s3 ,
(u2 + v2 + w2 )x2 = u2 r1 + v2 r2 + w2 r3 ,
(u2 + v2 + w2 )y2 = u2 s1 + v2 s2 + w2 s3 ,
(u3 + v3 + w3 )x3 = u3 r1 + v3 r2 + w3 r3 ,
(u3 + v3 + w3 )y3 = u3 s1 + v3 s2 + w3 s3 .
and therefore
(u1 + v1 + w1 )(u2 + v2 + w2 )(u3 + v3 + w3 )(P QR) =
 
 u1 + v1 + w1 u 2 + v 2 + w 2 u 3 + v 3 + w3

1  
=  (u1 + v1 + w1 )x1 (u2 + v2 + w2 )x2 (u3 + v3 + w3 )x3  =
2  (u + v + w )x (u + v + w )y (u + v + w )y 
1 1 1 1 2 2 2 2 3 3 3 3
 
 u1 + v1 + w1 u2 + v2 + w2 u3 + v3 + w3 
1  
=  u1 r1 + v1 r2 + w1 r3 u2 r1 + v2 r2 + w2 r3 u3 r1 + v3 r2 + w3 r3 =
2 u s +v s +w s u s +v s +w s u s +v s +w s 
1 1 1 2 1 3 2 1 2 2 2 3 3 1 3 2 3 3

    
 u v w1   1 r1 s1   u1 v1 w1 
1  1 1    
=  u2 v2 w2   1 r2 s2  =  u2 v2 w2  (ABC).
2  u v w  1 r s   u v w 
3 3 3 3 3 3 3 3

When the homogeneous coordinates of P , Q, R are normalized, we have

 
 u1 v1 w1 

(P QR) =  u2 v2 w2  (ABC).
 u3 v3 w3 

5. Lines

5.1. Equation of the line through two points. We can use the formula given
in the previous section for the area of the triangle to establish the equation of the
line through two points (u1 : v1 : w1 ) and (u2 : v2 : w2 ):
 
 u1 v1 w1 
 
 u2 v2 w2  = 0.
 
 x y z 

5.1.1. Examples.
(1) The equations of the sides BC, CA and AB are, respectively, x = 0, y = 0
and z = 0. For example, since B = (0 : 1 : 0) and C = (0 : 0 : 1), the line
BC has equation
 
 0 1 0 
 
 0 0 1  = 0 ⇔ x = 0.
 
 x y z 
(2) The equation of the perpendicular bisector of BC is (b2 − c2 )x + a2 (y −
z) = 0. For, this line goes through the midpoint of BC, with coordinates
(0 : 1 : 1) and the circumcenter O of ABC with coordinates
a2 (b2 + c2 − a2 ) : b2 (c2 + a2 − b2 ) : c2 (a2 + b2 − c2 ).
42 Barycentric coordinates

Therefore the equation of the perpendicular bisector of BC is

 
 2 2 02 1 1
 

 a (b + c − a2 ) b2 (a2 + c2 − b2 ) c2 (a2 + b2 − c2 )  = 0 ⇔
 
 x y z 
⇔ (a2 c2 − c4 − a2 b2 + b4 )x − a2 (b2 + c2 − a2 )(y − z) = 0 ⇔
⇔ (a2 c2 − c4 − a2 b2 + b4 )x + a2 (b2 + c2 − a2 )(y − z) = 0 ⇔
⇔ (b2 + c2 − a2 )((b2 − c2 )x + a2 (y − z)) = 0 ⇔
⇔ (b2 − c2 )x + a2 (y − z) = 0.
(3) The internal bisector of angle A joins the vertex A = (1 : 0 : 0) and the
incenter I = (a : b : c). The equation is
 
 1 0 0 
 
 a b c  = 0 ⇒ cy − bz = 0.
 
 x y z 

5.2. Parallel lines.

5.2.1. Infinite points. In order to get the equation of a parallel line we consider
the infinite points. We know that each line has an infinite point and all infinite
points lie on a line called line at infinity.
The line at infinity has equation x + y + z = 0, since x + y + z 6= 0 always return
an ordinary point.
The infinite point of the line px + qy + rz = 0 is (q − r : r − p : p − q), since their
coordinates have sum zero and it lies on the line px + qy + rz = 0.
On the other side if P = (u1 : v1 : w1 ) and Q = (u2 : v2 : w2 ) with u1 + v1 + w1 =
u2 + v2 + w2 , we can prove that the infinite point of line P Q has coordinates
(u1 − v1 , u2 − v2 , u3 − v3 ).
For example, since the orthocenter is H = (SBC : SCA : SAB ), the foot of the
altitude from A is AH = (0 : SCA : SAB ) = (0 : SC : SB ), with SC + SB = a2 and
the infinite point of the altitude through A = (a2 : 0 : 0) is −a2 : SC : SB ).

5.2.2. Parallel through a point. The line that goes through P = (u : v : w) parallel
to px + qy + rz = 0 has equation
 
 q−r r−p p−q 
 
 u v w  = 0.
 
 x y z 

5.2.3. Exercises.
(1) Find the equations of the lines through P = (u : v : w) parallel to the
sides of the triangle.
The infinite point of BC is, subtracting coordinates of C from B, (0, 1, −1),
then the line through P parallel to BC has equation
 
 0 1 −1 
 
 u v w  = 0 ⇔ (v + w)x − u(y + z) = 0.
 
 x y z 
 Francisco Javier Garcı́a Capitán 43

The parallels to CA and AB are (w + u)y − v(x + z) = 0 and (u + v)z −

w(x + y) = 0.
(2) Let DEF be the medial triangle of ABC. Given a point P , let XY Z be
the cevian triangle with respect to ABC and U V W the medial triangle of
XY Z. Find the point P such that lines DU , EV and F W are parallel to
the internal angle bisectors of angles A, B and C respectively.
We have :
A = (1 : 0 : 0), B = (0 : 1 : 0), C = (0 : 0 : 1),
D = (0 : 1 : 1), E = (1 : 0 : 1), F = (1 : 1 : 0),
X = (0 : v : w), Y = (u : 0 : w), Z = (u : v : 0).
Since
Y = (u : 0 : w) = ((u + v)u : 0 : (u + v)w),
Z = (u : v : 0) = ((u + w)u : (u + w)v : 0),
we have
U = ((2u + v + w)u : (u + w)v : (u + v)w).
If DU is parallel to AI, both lines have the same point at infinity. Now
since 2u2 + uv + uw + vw is the sum of coordinates of U , we consider
D = (0 : u2 + uv + uw + vw : u2 + uv + uw + vw),
and by subtracting coordinates of D and U we get that the infinity point
of line DU is
(2u2 + uv + uw : −u2 − uw : −u2 − uv) = (2u + v + w : −u − w : −u − v).
The angle bisector AI goes through A = (a+b+c : 0 : 0) and I = (a : b : c),
hence it infinite point is (b + c : −b : −c).
The two infinite points are the same when u + w = kb, u + v = kc for
some k. The same calculations for EV and F W , give
v + u = hc w + u = tb
 
,
v + w = ha w + v = ta
for some h and t. This gives k = h = t and u, v, w are the solutions of the
system of equations 
 u + v = kc

u + w = kb ,

 v + w = ka
that is , u = k(b + c − a), v = k(a + c − b), w = k(a + b − c) or P =
(b + c − a : a + c − b : a + b − c) is the Nagel point of triangle ABC.
A

Z
U E
F Y
P
V W

B D X C
44 Barycentric coordinates

5.3. Line intersection. The intersection point of two lines

p1 x + q1 y + r1 z = 0,


p2 x + q 2 y + r2 z = 0
is the point     

 q1 r1 
 : −  p1 r1  :  p1 q1  .
   

 q2 r2   p2 r2   p2 q2 

The infinite point of a line l can be regarded as the intersection of l and the line
at infinity with equation l∞ : x + y + z = 0.

Three lines pi x + qi y + ri z = 0, i = 1, 2, 3 are concurrent if and only if

 
 p1 q 1 r1 
 
 p2 q2 r2  = 0.
 
 p3 q 3 r3 

5.3.1. Examples.
(1) Let DEF be the medial triangle of ABC. Find the equation of the line
DIa , joining D and the A-excenter. Show that the lines DIa , EIb and F Ic
are concurrent and find their common point.
The equation of line DIa is
 
 x y z 
 
 0 1 1  = 0 ⇒ (b − c)x + ay − az = 0.
 
 −a b c 

Similarly we get
EIb : −bx + (c − a)y + bz = 0.
F Ic : cx − cy + (a − b)z = 0.
The three lines concur if and only if the determinant of coefficients is
zero:
   
 b−c a −a   −c c b − a 
   
 −b c − a b  =  −b c − a
  b  = 0.

 c −c a − b   c −c a − b 
To find the common point we solve the system formed by the two first
equations:
     
 a −a   b − c −a   b − c
 : − a 
(x : y : z) =  :  =
c−a b   −b b   −b c − a 
= (a(b + c − a) : b(a + c − b) : c(a + b − c)) =
= (a(s − a) : b(s − b) : c(s − c)) ,
known as Mittenpunkt.
(2) Let DEF be the medial triangle of ABC and X, Y, Z the midpoints of the
altitudes of ABC. Find the equations of lines DX, EY , F Z, show that
they are concurrent and find their common point.
The orthocenter is H = (SBC : SCA : SAB ), hence the foot of the altitude
from A is AH = (0 : SCA : SAB ) = (0 : SC : SB ), with SC + SB = a2 .
 Francisco Javier Garcı́a Capitán 45

Therefore, the midpoint of AH and A = (a2 : 0 : 0) is X = (a2 : SC : SB ).

The equation of line DX is
   
 x y
 z   x y
  z−y 

 0 1 1 = 0 1 0  = (SB − SC )x + a2 y − a2 z = 0.
 2   2 
 a SC SB   a SC SB − SC 

Since SB − SC = c2 − b2 , we get DX : (c2 − b2 )x + a2 y − a2 z = 0. Similarly

we get

EY : −b2 x + (a2 − c2 )y + b2 z = 0,
F Z : c2 x − c2 y + (b2 − a2 )z = 0.

Since
 
 c 2 − b2 a2 −a2 
 
 −b2 a2 − c 2 b2 =0
 2 2

 c −c b − a2
2 

because the first row is the sum of the other two, the three lines are
concurrent.
To find the common point we solve:

(x : y : z) =
  2   
a2 −a2   c − b2 −a2   c2 − b2 a2

 
=  2
 : −   :   =
a − c2 b2   −b2 b2   −b2 a2 − c2 
= a2 (a2 + b2 − c2 ) : b2 (a2 + b2 − c2 ) : c2 (a2 + b2 − c2 ) =

=(a2 : b2 : c2 ),

and the three lines concur at the symmedian point of ABC.

(3) (Vecten points) We have already seen that the midpoint of the square
BCCA BA constructed externally on BC is

MA = (−a2 : SC + S : SB + S).

The line AMA has equation

 
 x y z 
 
 1 0 0  = 0 ⇒ (SB + S)y − (SC + S)z = 0.
 
 −a2 SC + S SB + S 

Similarly, we have

BMB : (SA + S)x − (SC + S)z = 0

CMC : (SA + S)x − (SB + S)y = 0
46 Barycentric coordinates

AB

AC
A CB
MB

MC VE

BC
B C

MA

BA CA
The three lines are concurrent:
 

 0 SB + S −(SC + S) 

 SA + S
 0 −(SC + S) =

 SA + S −(SB + S) 0 
 

 0 SB + S −(SC + S) 
=  0 SB + S −(SC + S)  = 0.
 SA + S −(SB + S) 0 

The common point is

VE = ((SB + S)(SC + S) : (SC + S)(SA + S) : (SA + S)(SB + S)) ,
known as the Vecten point of ABC.
If we construct squares internally on the sides of the triangles, the we
find the inner Vecten point of the triangle, whose coordinates are
VI = ((SB − S)(SC − S) : (SC − S)(SA − S) : (SA − S)(SB − S)) .

5.4. Perpendicular lines. Given a line L : px + qy + rz = 0, it is interesting to

calculate the point at infinity of all lines perpendicular to L. The line L intersects
CA and AB at the points Y = (−r : 0 : p) and Z = (q : −p : 0). We want
to calculate the perpendicular line from A to L. First, we find the equations of
perpendiculars from Y to AB and from Z to CA. These are
   
 SB SA −c2   SC −b2 SA 
   
 −r 0 p  = 0,  q −p 0  = 0,
   
 x y z   x y z 
or
SA px + (c2 r − SB p)y + SA rz = 0,
SA px + SA qy + (b2 q − SC p)z = 0.
 Francisco Javier Garcı́a Capitán 47

X'

Z Y

B C
The intersection point of these lines is the orthocenter of AY Z,
X 0 =(∗ ∗ ∗ : SA p(SA r − b2 q + SC p) : SA p(SA q + SB p − c2 r)) =
=(∗ ∗ ∗ : SC (p − q) − SA (q − r) : SA (q − r) − SB (r − p)),
since SA + SC = b2 and SA + SB = c2 .

Now the perpendicular from A to L is the line AX 0 , whose equation is

 
 1 0 0 
 
 ∗ ∗ ∗ SC (p − q) − SA (q − r) SA (q − r) − SB (r − p)  = 0,
 
 x y z 
or −(SA (q − r) − SB (r − p))y + (SC (p − q) − SA (q − r))z = 0.

Then, if we call (f : g : h) = (q − r : r − p : p − q) to the infinite point of L, the

perpendicular to L from A has equation
−(SA f − SB g)y + (SC h − SA f )z = 0,
with (f 0 : g 0 : h0 ) = (SB g − SC h : SC h − SA f : SA f − SB g) as infinite point, which
will be also the infinite point of any line perpendicular to L.

5.4.1. Examples.
(1) Show that the perpendicular to the sides through the points of tangency
of the excircles are concurrent.
Let X = (0 : s − b : s − c), Y = (s − a : 0 : s − c) and Z = (s − a : s − b : 0) the
points of tangency of the excircles with the corresponding sides.

The infinite point of BC is (0 : 1 : 0) − (0 : 0 : 1) = (0 : 1 : −1). The infinite point

of any perpendicular to BC is
(SB · 1 − SC (−1) : SC (−1) − SA · 0 : SA · 0 − SB · 1) =
=(SB + SC : −SC : −SB ) = (−a2 , SC , SB ).
and the perpendicular to BC through X has equation
 
 0 s−b s−c 
 
 −a2 SC SB  = 0,

 x y z 
that is equivalent to s(b − c)x + a(s − c)y − a(s − b)z = 0.

If we calculate the perpendiculars to CA, AB through Y , Z, we get :

−b(s − c)x + s(c − a)y + b(s − a)z = 0,
c(s − b)x − c(s − a)y + s(a − b)z = 0.
48 Barycentric coordinates

The point of concurrence of these lines is known as the Bevan point of ABC.

References
[1] P.Yiu, Introduction to the Geometry of the Triangle, 2001, version of 2013, http://math.
fau.edu/Yiu/YIUIntroductionToTriangleGeometry130411.pdf.