Download as pdf or txt
Download as pdf or txt
You are on page 1of 14

NOTES ON THE BROCARD POINTS AND ANGLES OF A TRIANGLE 49

Notes on the Brocard points and angles of a


triangle
SADI ABU-SAYMEH and MOWAFFAQ HAJJA

1. Introduction – the classical Brocard points and angle


It is known that for every triangle ABC, there exists a unique interior
point Ω1 for which
∠Ω1AB = ∠Ω1BC = ∠Ω1CA, (1)
and a unique interior point Ω2 for which
∠Ω2AC = ∠Ω2BA = ∠Ω2CB; (2)
see Figures 1 and 2, where the common value of the angles in (1), (2) are
denoted by ω1, ω2, respectively.
These points are discussed in a lot of books and papers, and are called
the first and second Brocard points of ABC; see, for example, [1, Chapter
16, pp. 217ff], [2], and [3], and see [4, Chapter XVI, pp. 263ff] for a brief
history. It is also known that ω1 and ω2 turn out to be equal, and that their
common value is called the Brocard angle of ABC and is denoted by ω.
A A

ω1 ω2

Ω1 Ω2
ω1
ω2
ω1 ω2
B C B C
FIGURE 1 FIGURE 2
In Figures 1 and 2, which are not drawn to scale, ω1 and ω2 turn out to
be equal. Their common value is denoted by ω.

One of the properties of the Brocard angle that are proved in the
references above states that
ω ≤ 30°, with equality if, and only if, ABC is equilateral. (3)
It immediately follows that if Ω1 = Ω2, then 6ω = 180°, and hence
ω = 30°, and ABC is equilateral. The converse is also trivially true. Thus
we have
Ω1 = Ω2 ⇔ ABC is equilateral. (4)

Downloaded from https://www.cambridge.org/core. Karolinska Institutet University Library, on 02 Mar 2020 at 18:06:29, subject to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/mag.2020.7
50 THE MATHEMATICAL GAZETTE

In this note, we shall first give, in Theorem 1, a short proof of a rather


stronger form of (3). Then we shall present the Brocard-like points
introduced in [5], in [6], and in [7], and we prove results analogous to (3),
(4), and Theorem 1 for each of these points. The formal similarities among
the results should not result in the false impression that the proofs are
replicas of each other.

2. Preliminaries
The proofs use the concavity of the function sin x on [0, π] , i.e. the
convexity of − sin x. This follows from well-known trigonometric identities,
and it also follows from the fact that
(− sin x) ″ = sin x ≥ 0,
with equality if, and only if, x is 0 or π. Thus we have, by Jensen's
inequality, that
sin x1 + … + sin xn x1 + … + xn
≤ sin
n n
for all xi ∈ [ 0, π] , with equality if, and only if, x1 = … = xn; see [8,
p. 93, Theorems 5.2a and 5.2b]. In particular, if u, v, w ∈ [ 0, π] , then
sin u + sin v + sin w u + v + w
≤ sin ,
3 3
with equality if, and only if, u = v = w. By the AM-GM inequality, we
have that
sin u + sin v + sin w
3
sin u sin v sin w ≤ ,
3
with equality if, and only if, u = v = w. Thus we have proved that if
u, v, w ∈ [ 0, π] , then
sin u + sin v + sin w u+v+w
3
sin u sin v sin w ≤ ≤ sin , (5)
3 3
with one (and hence both) of the equalities if, and only if, u = v = w. This
will be used in the sequel.

We shall also use Ceva's theorem and its trigonometric variant. We


recall that a cevian of triangle ABC is a line segment joining a vertex to a
point on the opposite side (or to a point on the line of the opposite side, but
we will not be concerned with this here). Ceva's theorem states that if X, Y
and Z are points on the sides BC, CA and AB, as in Figure 3, then
BX CY AZ
the cevians AX, BY and CZ are concurrent ⇔ = 1; (6)
CX AY BZ
see [1, Theorem 6.1, p. 69] or [9, (4.2) Theorem, pp. 126-129].

Downloaded from https://www.cambridge.org/core. Karolinska Institutet University Library, on 02 Mar 2020 at 18:06:29, subject to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/mag.2020.7
NOTES ON THE BROCARD POINTS AND ANGLES OF A TRIANGLE 51

U u

Z
Y

G
W
v
V w
B X C
FIGURE 3

Assuming that the cevians AX, BY and CZ are concurrent, say at G as


shown in Figure 3, and comparing the areas [ABX] and [ACX] of triangles
ABX and ACX, we obtain
BX [ABX] (AB) (AX) sin U c sin U
= = = .
CX [ACX] (AC) (AX) sin u b sin u
Similar formulas are obtained by permuting A, B and C. Thus we have
BX c sin U CY a sin V AZ b sin W
= , = , = .
CX b sin u AY c sin v BZ a sin w
Multiplying out, we obtain
BX CY AZ sin U sin V sin W
= . (7)
CX AY BZ sin u sin v sin w
It follows from this and (6) that if the cevians AX, BY and CZ are
concurrent, then the right-hand side of (7) is equal to 1. The converse is also
easy to prove, and we have that
sin U sin V sin W
the cevians AX, BY and CZ are concurrent ⇔ = 1.
sin u sin v sin w
This form is usually called the trigonometric form of Ceva's theorem; see [1,
p. 71] or [9, (4.7) Theorem, pp. 136-137].

3. A stronger form of (3)


Now we consider the cevians through an arbitrary interior point G of
ABC, and we prove a general result that implies (3). This is given in
Theorem 1 below. This theorem appeared as Problem 5 in the 1991
International Mathematical Olympiad, and is reproduced on page 7 of [10]
with two proofs on pages 66–68. It is also reproduced as Delta 1.2 on page
292 with another proof, and with a generalisation to any convex n-gon in
Delta 21.4. We provide yet another proof that is considerably simpler, and,
more importantly, that can be imitated to prove (13) below.

Downloaded from https://www.cambridge.org/core. Karolinska Institutet University Library, on 02 Mar 2020 at 18:06:29, subject to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/mag.2020.7
52 THE MATHEMATICAL GAZETTE

Theorem 1: Let G be an interior point of triangle ABC, and let AX, BY , CZ


be the cevians through G; see Figure 3 above. Let U , u, V , v, W , w be as in
the Figure. Then at least one of the angles U , V , W is less than, or equal to,
30°. The same holds for the angles u, v, w.
Also, if ω is the Brocard angle of ABC, then ω ≤ 30°, with equality if,
and only if, ABC is equilateral.

Proof: Suppose, by way of contradiction, that min {U, V, W } > 30°. Then
180° > U + V + W > U + 60°,
and therefore U < 120°. The same holds for V and W , and thus we have
120° > U , V, W > 30°. (8)
It follows that U + V + W > 90°, and therefore
u + v + w < 90°. (9)
Now it follows from the trigonometric version of Ceva's theorem that
sin u sin v sin w = sin U sin V sin W. (10)
But it follows from (8) that
1 1 1 1
sin U sin V sin W > · · = ,
2 2 2 8
and it follows from (5) and (9) that

sin u sin v sin w ≤ sin (


u + v + w3
3 )
1
≤ (sin 30°) = ,
3
8
contradicting (10). This proves the first statement.

It immediately follows that ω ≤ 30°. Thus it remains to prove that


ω = 30° if, and only if, ABC is equilateral.
Referring to Figure 3, and taking u = v = w = ω in (10), we obtain
sin 3 ω = sin (A − ω) sin (B − ω) sin (C − ω) .
But

( (A − ω) + (B − ω)(C − ω)
)
3
sin (A − ω) sin (B − ω) sin (C − ω) ≤ sin
3
= sin 3 (60° − ω) ,
with equality if, and only if, A − ω = B − ω = C − ω, i.e. A = B = C.
Therefore
sin 3 ω = sin 3 (60° − ω) , ⇔ A = B = C.
Since 0 ≤ ω ≤ 30°, it follows that
sin 3 ω = sin 3 (60° − ω) ⇔ ω = 60° − ω ⇔ ω = 30°,
as desired.

Downloaded from https://www.cambridge.org/core. Karolinska Institutet University Library, on 02 Mar 2020 at 18:06:29, subject to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/mag.2020.7
NOTES ON THE BROCARD POINTS AND ANGLES OF A TRIANGLE 53

4. Cevian Brocard-like points


In analogy with the treatment of the (classical) Brocard points presented
in Section 1, it is proved in [5] that for every triangle ABC, there exists a
unique interior point Ψ1 through which the cevians AX, BY and CZ have the
property that
∠AZY = ∠BXZ = ∠CYX, (11)
and a unique interior point Ψ2 through which the cevians AX, BY and CZ
have the property that
∠AY′Z′ = ∠BZ′X′ = ∠CX′Y′; (12)
see Figures 4 and 5, where the common value of the angles in (11), (12) are
denoted by ψ1, ψ2, respectively. These points first appeared in [5], where
they are called the first and second cevian Brocard points of ABC, and
where it is proved that for these points, the angles ψ1 and ψ2 are equal. Their
common value is called the cevian Brocard angle of ABC and will be
denoted by ψ. We should mention here that ψ is called Ω in [5].
Among the properties of the cevian Brocard angle is the analogue of (3)
given by
ψ ≤ 60°, with equality if, and only if, ABC is equilateral. (13)
Also, it immediately follows that if Ψ1 = Ψ2, then
(2ψ + A) + (2ψ + B) + (2ψ + C) = 3 (180°) ,
and hence ψ = 60°, and A = B = C = 60°. The converse is also trivially
true. Thus we have
Ψ1 = Ψ2 ⇔ ABC is equilateral, (14)
A A

Z ψ1 Z′
ψ2
Y ψ2 Y′
ψ1
Ψ1 Ψ2

ψ1 ψ2
B X C B X′ C

FIGURE 4 FIGURE 5
In Figures 4 and 5, which are not drawn to scale, ψ1 and ψ2 turn out to be
equal. Their common value is denoted by ψ.

5. A stronger form of (13)


The next theorem is stronger than (13). It is the analogue of
Theorem 1 for the cevian Brocard angle.

Downloaded from https://www.cambridge.org/core. Karolinska Institutet University Library, on 02 Mar 2020 at 18:06:29, subject to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/mag.2020.7
54 THE MATHEMATICAL GAZETTE

Theorem 2: Let G be an interior point of triangle ABC, and let AX, BY , CZ


be the cevians through G; see Figure 6. Let X, x, Y , y, Z, z be as in the
Figure. Then at least one of the angles X, Y , Z is less than, or equal to, 60°.
The same holds for the angles x, y, z.
Also, if ψ is the cevian Brocard angle of ABC, then ψ ≤ 60°, with
equality if, and only if, ABC is equilateral.

Proof: Referring to Figure 6, we let


X = ∠BXZ, Y = ∠CYX, Z = ∠AZY, x = ∠YXC, y = ∠ZYA, z = ∠XZB,
and we suppose, by way of contradiction, that min {X, Y, Z } > 60°.
A

Y′
Z Z
y
z Y
Y
G

X x
B X C
FIGURE 6

Let Y′ be the point on the line segment AY for which ∠AZY′ = 60°.
Then
AZ AZ sin ∠ZY′A 2 sin (120° − A)
< = = . (15)
AY AY′ sin 60° 3
Letting
α = 120° − A, β = 120° − B, γ = 120° − C, (16)
we rewrite (16) and its permutes as
3 AZ 3 BX 3 CY
sin α > , sin β > , sin γ > .
2 AY 2 BZ 2 CX
Multiplying and using Ceva's theorem, we obtain
3 3
sin α sin β sin γ > . (17)
8
But it also follows from (5) and (16) that

( )
3
α + β + γ 3 3
= (sin 60°) =
3
sin α sin β sin γ ≤ sin ,
3 8
contradicting (17). This proves the first statement.

Downloaded from https://www.cambridge.org/core. Karolinska Institutet University Library, on 02 Mar 2020 at 18:06:29, subject to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/mag.2020.7
NOTES ON THE BROCARD POINTS AND ANGLES OF A TRIANGLE 55

It follows that ψ ≤ 60°, and thus it remains to prove ψ = 60° if, and
only if, ABC is equilateral.
Referring to Figure 6, we use Ceva's theorem and the facts that
AZ sin y BX sin z CY sin x
= , = , =
AY sin Z BZ sin X CX sin Y
to conclude that
sin x sin y sin z = sin X sin Y sin Z.
Putting x = y = z = ψ, we see that
sin 3 ψ = sin (A + ψ) sin (B + ψ) sin (C + ψ) .
But
sin (A + ψ) sin (B + ψ) sin (C + ψ)

(
≤ sin
(A + ψ) + (B + ψ) + (C + ψ) 3
3 )
= sin 3 (60° + ψ) ,

with equality if, and only if, A = B = C. Therefore


sin 3 ψ = sin 3 (60° + ψ) ⇔ A = B = C.
Also,
sin 3 ψ = sin 3 (60° + ψ) ⇔ 2ψ + 60° = 180° ⇔ ψ = 60°,
because ψ ≤ 60°. Therefore ψ = 60° ⇔ A = B = C, as desired.

6. Yff's Brocard-like points


In another analogy with the treatment in Section 1, it is proved in [6]
that for every triangle ABC, there exists a unique interior point Λ1 through
which the cevians AX, BY and CZ have the property that
BX = CY = AZ, (18)
and a unique interior point Λ2 through which the cevians AX′, BY′ and CZ′
have the property that
CX′ = AY′ = BZ′; (19)
see Figures 7 and 8, where the common value of the lengths in (18), (19) are
denoted by λ1, λ2, respectively. These points first appeared in [6], where it is
proved that for these points, the numbers λ1 and λ2 are equal. Their common
value will be denoted by λ. The points Λ1 and Λ2 may be called the first and
second Yff Brocard-like points of ABC.

In Figures 7 and 8, which are not drawn to scale, λ1 and λ2 turn out to be
equal. Their common value is denoted by λ.

Downloaded from https://www.cambridge.org/core. Karolinska Institutet University Library, on 02 Mar 2020 at 18:06:29, subject to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/mag.2020.7
56 THE MATHEMATICAL GAZETTE
A A

λ1
λ2

Z Z′
Y Y′

λ2
Λ1 λ1 Λ2

B λ1 X C B X′ λ2 C

FIGURE 7 FIGURE 8
It easily follows that if Λ1 = Λ2, then BX = CY = AZ and
CX = AY = BZ and hence ABC is equilateral. The converse is also trivial,
and thus
Λ1 = Λ2 ⇔ ABC is equilateral. (20)
This is the analogue of (4) and (14). However, (3) and (13) do not seem to
have analogues for Λ1 and Λ2. In other words, the questions about bounds on
λ do not seem to have been asked.

Theorem 3 below implies that λ = 16 (a + b + c), with equality if, and


only if, ABC is equilateral. This is a satisfactory analogue of (3) and (13). In
the proof, we shall use material from [11, Section 3, pp. 221–222]. Although
we could have supplied the details needed for making the treatment self-
contained, we chose to make our use of [11] more visible.

Theorem 3: Let ABC be a triangle with side lengths a, b and c in the


standard order. Let G be an interior point of ABC, and let AX, BY and CZ be
the cevians through G. Let
x = BX, y = CY, z = AZ.
Then at least one of the numbers x, y and z is less than, or equal to
6 (a + b + c).
1

Also, if λ is as defined earlier (in Figures 7 and 8), then λ ≤ 16 (a + b + c),


with equality if, and only if, ABC is equilateral.
Similarly for the requirement CX = AY = BZ.

Proof: We suppose that


min {x, y, z} > t, (21)
where
a + b + c
t = ,
6
and we reach a contradiction.

Downloaded from https://www.cambridge.org/core. Karolinska Institutet University Library, on 02 Mar 2020 at 18:06:29, subject to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/mag.2020.7
NOTES ON THE BROCARD POINTS AND ANGLES OF A TRIANGLE 57

By Ceva's theorem, we have xyz = (a − x) (b − y) (c − z). By (21),


we have xyz > t 3 and (a − x) (b − y) (c − z) < (a − t) (b − t) (c − t).
Therefore t 3 < (a − t) (b − t) (c − t), or equivalently, f (t) < 0, where f
is defined by
f (x) = x3 − (a − x) (b − x) (c − x) .
We shall reach a contradiction by proving that f (t) ≥ 0. We then complete
the proof by showing that if f (t) = 0, then a = b = c.

It is not difficult, or one may use an algebra package, to see that


F(t) = 54f (t)
= − (a3 + b3 + c3) + 6 (a2b + b2a + b2c + c2b + c2a + a2c) − 33abc.
We now use the material in [11, Section 3, pp. 221–222], where it is proved
that if ABC is non-degenerate (as we have always assumed), and if
L = − (a3 + b + c3) + 2 (a2b + b a + b c + c2b + c2a + a2c) − 9abc,
3 2 2

M = (a3 + b3 + c3) − (a2b + b2a + b2c + c2b + c2a + a2c) + 3abc,


then
L ≥ 0, with equality if, and only if, a = b = c,
M ≥ 0, with equality if, and only if, a = b = c.
Since F = 5L + 4M, it follows that f ≥ 0, with equality if, and only if,
a = b = c, and the proof is complete.

7. Goormaghtigh's Brocard-like points


In yet another analogy with the treatment in Section 1, it is proved in [7]
that for every triangle ABC, there exists a unique interior point Γ1 from
which the perpendiculars Γ1D, Γ1E and Γ1F to the sides BC, CA, AB have the
property that
BD = CE = AF,
and a unique interior point Γ2 from which the perpendiculars Γ2D, Γ2E and
Γ2F to the sides BC, CA, AB have the property that
CD′ = AE′ = BF′;
see Figures 9 and 10, where the common value of the lengths in (18), (19) is
denoted by γ1, γ2, respectively. These points first appeared in [7], where it is
proved that for these points, the numbers γ1 and γ2 are equal. Their common
value will be denoted by γ. The points Λ1 and Λ2 may be called the first and
second Goormaghtigh Brocard-like points of ABC.

In Figures 9 and 10, which are not drawn to scale, γ1 and γ2


turn out to be equal. Their common value is denoted by γ.

Downloaded from https://www.cambridge.org/core. Karolinska Institutet University Library, on 02 Mar 2020 at 18:06:29, subject to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/mag.2020.7
58 THE MATHEMATICAL GAZETTE
A A

γ1
γ2

F F
E E

γ2
Γ1 γ1 Γ2

B γ1 D C B D′ γ2 C

FIGURE 9 FIGURE 10
It is very easy to see that
Γ1 = Γ2 ⇔ ABC is equilateral.
This is the analogue of (4), (14) and (20).

The next theorem is an analogue of Theorems 1, 2, and 3, but it differs


in that it gives both upper and lower bounds for γ. The proof uses a Ceva-
like theorem of Carnot stating that if X, Y and Z are points on the sides BC,
CA and AB of triangle ABC such that BX = x, CY = y, AZ = z, as shown
in Figure 11, then the perpendiculars erected from X, Y and Z on the sides
BC, CA and AB are concurrent if, and only if,
x2 + y2 + z2 = (a − x)2 + (b − y)2 + (c − z)2 ; (22)
see, for example, [12, Theorem 6.3.1, p. 96], [13], or [14, Theorem 2.2,
p. 19].
A

Z
Y

B x X C
FIGURE 11

Theorem 4: Let ABC be a triangle with side lengths a, b and c in the standard
order. Let P be an interior point of ABC, and let PX, PY and PZ be the
perpendiculars from P to BC, CA and AB, respectively. Let
x = BX, y = CY, z = AZ.

Downloaded from https://www.cambridge.org/core. Karolinska Institutet University Library, on 02 Mar 2020 at 18:06:29, subject to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/mag.2020.7
NOTES ON THE BROCARD POINTS AND ANGLES OF A TRIANGLE 59

Then at least one of the numbers x, y and z is less than 14 (a + b + c), and at
least one of them is greater than, or equal to, 16 (a + b + c).
Also, if x = y = z (= γ), i.e. if P is the first Goormaghtigh Brocard-
like point, then
a + b + c a + b + c
≤ γ < , (23)
6 4
with equality if, and only if, ABC is equilateral.

Similarly for the requirement x = CD, y = AE, z = BF.

Proof: We shall prove (23) first. It follows from Carnot's theorem (22) that
a2 + b + c2 = 2 (ax + by + cz) . (24)
2

Putting x = y = z (= γ), we obtain


a2 + b2 + c2
γ = . (25)
2 (a + b + c)
Thus
a + b + c a + b + c
≤ γ <
6 4
a + b + c a2 + b2 + c2 a + b + c
⇔ ≤ <
6 2 (a + b + c) 4
⇔ 2 (a + b + c)2 ≤ 6 (a2 + b + c2) < 3 (a + b + c) . (26)
2 2

But (26) can be proved as follows:


6 (a2 + b2 + c2) − 2(a + b + c)2 = 4 ((a2 + b2 + c2) − (ab + bc + ca))
= 2 ((a − b)2 + (b − c)2 + (c − a)2)
> 0, with equality if, and only if, a = b = c;
3(a + b + c)2 − 6 (a2 + b2 + c2) = 3 (− (a2 + b2 + c2) + 2(ab + bc + ca))
= 3 (a(b + c − a) + b(c + a − b) + c(a + b − c))
≥ 0.
This proves (26), and it also proves that equality holds if, and only if, ABC is
equilateral.
To prove the first statement of the theorem, suppose, by way of
contradiction, that
a + b + c
max {x, y, z} < .
6

Downloaded from https://www.cambridge.org/core. Karolinska Institutet University Library, on 02 Mar 2020 at 18:06:29, subject to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/mag.2020.7
60 THE MATHEMATICAL GAZETTE

Then it follows from (22) that

a2 + b2 + c2 < 2 (a + b + c) ( a + 6b + c ) , i.e.
6 (a2 + b + c2) < 2 (a + b + c)2 ,
2

contradicting (26). Similarly the assumption that


a + b + c
min {x, y, z} ≥
4
leads to the contradiction
6 (a2 + b2 + c2) ≥ 3 (a + b + c)2 .
This completes the proof.

8. Remarks and questions


In this section, we make several remarks that raise issues for future
research.

8.1 Coincidence of Brocard-like points


The eight special points Ω1, Ω2, Ψ1, Ψ2, Λ1, Λ2, Γ1, Γ2 appearing in
Theorems 1-4 are not triangle centres for ABC, since they are not invariant
under permuting the vertices of ABC. But one may still ask whether a given
pair of these points may coincide for a non-equilateral triangle. A priori, this
gives rise to 28 questions. Although some of these questions are already
answered in the existing literature, there remain many others that may turn
out to form a rich source for composing problems that can be used in
training for, and including in, mathematical contests and competitions.
One may also define the points Ω, Ψ, Λ and Γ to be the midpoints of the
segments that join the relevant pairs of points, i.e. by
Ω1 + Ω2
Ω =
, etc.
2
One then asks the question raised in the previous paragraph about these
points.

8.2 Positivity of symmetric forms


In Section 6, we benefited a lot from the material in [11]. In that paper,
it is proved, among other things, that if f (X, Y, Z) is a symmetric cubic
form in the variables X, Y and Z, then
(a) f (X, Y, Z) ≥ 0 for all X, Y, Z ≥ 0
⇔ f (1, 1, 1) , f (1, 1, 0) , f (1, 0, 0) ≥ 0,
(b) f (X, Y, Z) ≥ 0 for all X, Y , Z that can serve as the side-
lengths of a (degenerate or non-degenerate) triangle
⇔ f (1, 1, 1) , f (2, 1, 1) , f (1, 1, 0) ≥ 0.

Downloaded from https://www.cambridge.org/core. Karolinska Institutet University Library, on 02 Mar 2020 at 18:06:29, subject to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/mag.2020.7
NOTES ON THE BROCARD POINTS AND ANGLES OF A TRIANGLE 61

It also gives a method (which we used in Section 7) for finding conditions


for f (X, Y, Z) to be 0. Interested readers can find a generalisation to any
number of variables in [15], and for shorter proofs in [16]. For similar
results on 2 variables, we refer the readers to [17].

8.3 Lower bounds on ω, ψ and λ


Theorem 4 gives both an upper and a lower bound on γ, while
Theorems 1, 2 and 3 give only upper bounds on ω, ψ and λ. This makes it
natural to wonder whether there are such bounds in the first place. For
example, if λ is as defined in Section 5, then it is not difficult to prove, using
the theorem in Section 7.2 above, to prove that there does not exist any
r > 0 such that λ ≥ r (a + b + c). This also follows by proving that λ can
take values as close to 0 as desired.
It is also true that there does not exist any r > 0 for which ψ ≥ rπ. In
fact, it is proved in [5, Theorem 2(iii)] that ψ can take all values in (0, 60°).
As for ω, one can also prove that it can take all values in (0, 30°). One
can also use the identity cot ω = cot A + cot B + cot C and find the
extrema of cot A + cot B + cot C as A, B, C range over the angles of a
triangle. This has the advantage of giving a new proof that ω ≤ 30°. In this
respect, one may mention that there is a systematic and fairly short
procedure, described in [18, Section 3, p. 16], for finding the extrema of any
symmetric function in cot A, cot B, cot C, as long as it can be expressed in
terms of the elementary symmetric functions
cot A + cot B + cot C, cot A cot B + cot B cot C + cot C cot A, cot A cot B cot C.

References
1. G. Leversha, The geometry of the triangle, The United Kingdom
Mathematics Trust, Pathways, No. 2, University of Leeds (2013).
2. P. Yff, On the Brocard points of a triangle, Amer. Math. Monthly 67
(1960) pp. 520-525.
3. R. Shail, Some properties of Brocard points, Math. Gaz. 80 (November
1996) pp. 485-491.
4. R. A. Johnson, Advanced Euclidean Geometry, Dover Publications,
New York (1929).
5. S. Abu-Saymeh and M. Hajja, Some Brocard-like points of a triangle,
Forum Geom. 5 (2005) pp. 65-74.
6. P. Yff, An analogue of the Brocard points, Amer. Math. Monthly, 70
(1963) pp. 495-501.
7. M. R. Goormaghtigh, Sur deux points du plan d’un triangle et sur une
généralisation des points de Brocard, Ann. de. Mathémat., 4° série, t.
XVIII, (Novembre 1918) pp. 417-424.
8. I. Niven, Maxima and minima without calculus, The Dolciani
Mathematical Expositions, No. 6, MAA, Washington, D. C. (1981).

Downloaded from https://www.cambridge.org/core. Karolinska Institutet University Library, on 02 Mar 2020 at 18:06:29, subject to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/mag.2020.7
62 THE MATHEMATICAL GAZETTE

9. I. M. Isaacs, Geometry for college students, AMS, Providence, RI


(2001).
10. M. E. Kuczma, International Mathematical Olympiads, 1986-1999,
MAA, Washington, D. C. (2003).
11. J. Rigby, A method for obtaining related inequalities, with applications,
Univ. Beograd. Publ. Elektrotehn. Fak. Ser. Mat. Fiz., 412-460 (1973),
pp. 217-226.
12 C. J. Bradley, Challenges in geometry, Oxford University Press, New
York (2005).
13. P. E. Clement, The concurrence of perpendiculars, Amer. Math.
Monthly, 65 (1958) pp. 601-605.
14. T. Andreescu, S. Korsky and C. Pohoata, Lemmas in Olympiad
geometry, XYZ Press, LLC (2016).
15. M. D. Choi, T. Y. Lam and B. Reznick, Even symmetric sextics,
Math. Z. 195 (1987) pp. 559-580.
16. M. Hajja, Copositive symmetric cubic forms, Amer. Math. Monthly,
112 (2005) pp. 462-466.
17. M. Hajja, Radical and rational means of degree 2, Math. Inequal. Appl.
6 (2003) pp. 581-593.
18. J. Habeb and M. Hajja, A method for establishing certain trigonometric
inequalities, J. Inequal. Pure Appl. Math. (JIPAM) 8 (2007) Art. 29.
10.1017/mag.2020.7 SADI ABU-SAYMEH
2271 Barrowcliffe Drive, Concord, NC 28027, USA
P. O. Box 963708, 11196 – Amman – Jordan
MOWAFFAQ HAJJA
P. O. Box 388 – Al-Husun, 21510 – Irbid – Jordan, Jordan
e-mails: ssaymeh@yahoo.com; mowhajja@yahoo.com

Downloaded from https://www.cambridge.org/core. Karolinska Institutet University Library, on 02 Mar 2020 at 18:06:29, subject to the
Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/mag.2020.7

You might also like