Junior problems

J559. Let
2n2
an = 1 − √ , n = 1, 2, 3, . . . .
1 + 1 + 4n4
√ √ √
Prove that a1 + 2 a2 + ⋯ + 20 a20 is an integer.

Proposed by Titu Andreescu, University of Texas at Dallas, USA

Solution by Polyahedra, Polk State College, USA

By rationalizing the denominator and Germain’s factoring, we see that
¿ √
√ Á
Á
À 2 1 1 1
n an = n + − (n2 + + n) (n2 + − n)
2 2 2
√ √
1 2 1 1 2 1
= (n + + n) − (n + − n)
2 2 2 2
Therefore, by telescoping,
√ √
√ √ √ 1 1 1 2 1 29 1
a1 + 2 a2 + ⋯ + 20 a20 = (202 + + 20) − (1 + − 1) = − = 14.
2 2 2 2 2 2

Also solved by Brian Bradie, Christopher Newport University, Newport News, VA, USA; Alina Craciun,
Theoretical High School Miron Costin, Pascani, Romania; G. C. Greubel, Newport News, VA, USA; Prodro-
mos Fotiadis, Nikiforos High School, Drama, Greece; Ioan Viorel Codreanu, Satulung, Maramures, Romania;
Corneliu Mănescu-Avram, Ploieşti, Romania; Daniel Văcaru, Pites, ti, Romania; Mohammad Imran, In-
dia; Nicuşor Zlota, Traian Vuia Technical College, Focşani, Romania; Kanav Talwar, Delhi Public School,
Faridabad, India; Arkady Alt, San Jose, CA, USA.

Mathematical Reflections 1 (2021) 1

J560. Let a, b, c be positive real numbers. Prove that
2 5 45 16 24 48
+ 2+ 2 > + + .
a2 b c (a + b) 2 (b + c) 2 (c + a)2

Proposed by Kartik Vedula, James S. Rickards High School, Tallahassee, USA

Solution by the author

We claim that the expression 4
ab + bc
6
+ 12
ca is between both expressions, but equality cannot occurs at the same
time:

1. Note that (a + b)2 ≥ 4ab and similar, so

16 24 48 16 24 48 4 6 12
+ + ≤ + + = + +
(a + b) 2 (b + c) 2 (c + a) 2 4ab 4bc 4ac ab bc ca

2 2 2
2. Note that ( a1 − 2b ) , ( 1b − 3c ) , ( a1 − 6c ) ≥ 0, so adding these together gives

2 5 45 4 6 12
2
+ 2+ 2 ≥ + +
a b c ab bc ca

This means that we have shown 2

a2
+ 5
b2
+ 45
c2
≥ 16
(a+b)2
+ 24
(b+c)2
+ 48
(c+a)2
. However, equality for the first part
occurs when a = b = c, but equality for the second part occurs when a = 2b = 6c . These cannot coincide, so we
1

are done.

Also solved by Polyahedra, Polk State College, USA; Daniel Văcaru, Pites, ti, Romania; Jiang Lianjun,Gui
Lin,China; Arkady Alt, San Jose, CA, USA.

Mathematical Reflections 1 (2021) 2

J561. Solve in nonzero real numbers the system
1 2 1 2 1 2
x− + = y − + = z − + = 0.
x y y z z x

Proposed by Adrian Andreescu, University of Texas at Dallas, USA

Solution by Brian Bradie, Christopher Newport University, Newport News, VA, USA
Solving the equations
1 2 1 2 1 2
x − + = 0, y − + = 0, z − + = 0
x y y z z x
for y, z, and x, respectively, yields
2x 2y 2z
y= , z= , and x = .
1 − x2 1 − y2 1 − z2
Now, let x = tan θ. Then
2 tan θ 2 tan 2θ
y= = tan 2θ, z= = tan 4θ,
1 − tan2 θ 1 − tan2 2θ
and
2 tan 4θ
x= = tan 8θ.
1 − tan2 4θ
Thus,
tan θ + tan 7θ
tan θ = tan 8θ = ,
1 − tan θ tan 7θ
which reduces to
(1 + tan2 θ) tan 7θ = 0.
Because x = tan θ must be nonzero, there are six solutions for θ:
π 2π 3π 4π 5π 6π
, , , , , .
7 7 7 7 7 7
This yields six solutions to the original system:
π 2π 4π
(x, y, z) = (tan , tan , tan )
7 7 7
2π 4π π
= (tan , tan , tan )
7 7 7
3π 6π 5π
= (tan , tan , tan )
7 7 7
4π π 2π
= (tan , tan , tan )
7 7 7
5π 3π 6π
= (tan , tan , tan )
7 7 7
6π 5π 3π
= (tan , tan , tan )
7 7 7

Also solved by Polyahedra, Polk State College, USA; Corneliu Mănescu-Avram, Ploieşti, Romania; Daniel
Văcaru, Pites, ti, Romania; Prodromos Fotiadis, Nikiforos High School, Drama, Greece; Nicuşor Zlota, Traian
Vuia Technical College, Focşani, Romania; Arkady Alt, San Jose, CA, USA.

Mathematical Reflections 1 (2021) 3

J562. Let ABC be a triangle and let D, E, F be points on sides BC, CA, AB, respectively, such that
AD, BE, CF are concurrent in X. Assume that the ratios BD CE AF
[1
DC , EA , F B are in the interval 5 , 5] and
DC + EA + F B = 5 . Evaluate
that BD CE AF 31

AX BX CX
+ + .
XD XE XF
Proposed by Mohammad Imran, India

Solution by Polyahedra, Polk State College, USA

Let x = BD/DC, y = CE/EA, and z = AF /F B. By Ceva’s theorem, xyz = 1. Notice that

AX [ABXC] [ABX] + [CAX] AE AF 1

= = = + = + z.
XD [BCX] [BCX] EC F B y

Therefore,
AX BX CX 1 1 1 31
+ + = + z + + x + + y = zx + xy + yz + .
XD XE XF y z x 5
Let a = xy + yz + zx and
31 1 31
f (t) = (t − x)(t − y)(t − z) = t3 − t + at − 1 = (t − ) (t − 1)(t − 5) + (a − ) t.
5 5 5
If a > 31/5, then f (0) = −1 and f (1/5) > 0, so f has a zero in (0, 1/5), a contradiction. If a < 31/5,
then f (5) < 0, so f has a zero greater than 5, again a contradiction. Therefore, a = 31/5, attained by
{x, y, z} = {1/5, 1, 5}. In conclusion, the value of the expression is 62/5.

Also solved by Alina Craciun, Theoretical High School Miron Costin, Pascani, Romania; Daniel Văcaru,
Pites, ti, Romania.

Mathematical Reflections 1 (2021) 4

J563. Let a, b, c ≥ 0 be real numbers such that ab + bc + ca = a + b + c > 0. Prove that
1 1 1 7
1≤ + + ≤ .
1 + 2a 1 + 2b 1 + 2c 5

Proposed by An Zhenping, Xianyang Normal University, China

Solution by Polyahedra, Polk State College, USA √

We may assume a ≥ b ≥ c ≥ 0. If c = 0, then 0 ≠ ab = a + b ≥ 2 ab, so ab ≥ 4. Therefore,

1 1 2(1 + ab) 2
0< + = ≤ .
1 + 2a 1 + 2b 1 + 6ab 5
Next, suppose c > 0. Let x = 1/a, y = 1/b, and z = 1/c. Then

bc + ca + ab c + a + b
x+y+z = = = xy + yz + zx,
abc abc
thus
xy + yz + zx xy + yz + zx xy + yz + zx
a= , b= , c= .
x(x + y + z) y(x + y + z) z(x + y + z)
By the Cauchy-Schwarz inequality,

1 x(x + y + z) x2
∑ =∑ = (x + y + z) ∑ 3
cyc 1 + 2a cyc x(x + y + z) + 2(xy + yz + zx) cyc x + 3x y + 3x z + 2xyz
2 2

(x + y + z)3
≥ = 1.
∑cyc (x3 + 3x2 y + 3x2 z + 2xyz)

On the other hand, since

1 1 1 3 + 8(a + b + c)
+ + = ,
1 + 2a 1 + 2b 1 + 2c 1 + 6(a + b + c) + 8abc

the right inequality is equivalent to a + b + c + 28abc ≥ 4, that is,

(x + y + z)2 (xy + yz + zx)2 + 28(xy + yz + zx)3 ≥ 4xyz(x + y + z)3 .

Notice that (xy + yz + zx)2 ≥ 3xyz(x + y + z) and

(x + y + z)(xy + yz + zx)2 + 84xyz(xy + yz + zx) − 4xyz(x + y + z)2

≥(x + y + z)(xy + yz + zx)2 − 4xyz (x2 + y 2 + z 2 ) ≥ x3 (y − z)2 + y 3 (z − x)2 + z 3 (x − y)2 ≥ 0.

Also solved by Christopher Newport University, Newport News, VA, USA; Daniel Văcaru, Pites, ti, Roma-
nia; Jiang Lianjun, Southwest University, Chong Qing, China; Arkady Alt, San Jose, CA, USA.

Mathematical Reflections 1 (2021) 5

J564. Find all complex numbers z such that for each real number a and each positive integer n

(cos a + z sin a)n = cos na + z sin na.

Proposed by Adrian Andreescu, University of Texas at Dallas, USA

Solution by Polyahedra, Polk State College, USA

Letting a = π/2 and n = 2, we get z 2 = −1, so z ∈ {±i}.
By DeMoivre’s formula, for all real a and natural n,

(cos a + i sin a)n = cos na + i sin na

and
(cos a − i sin a)n = [cos(−a) + i sin(−a)]n = cos na − i sin na.
Therefore, z = ±i.

Also solved by Brian Bradie, Christopher Newport University, Newport News, VA, USA; G. C. Greubel,
Newport News, VA, USA; Prodromos Fotiadis, Nikiforos High School, Drama, Greece; Corneliu Mănescu-
Avram, Ploieşti, Romania; Daniel Văcaru, Pites, ti, Romania.

Mathematical Reflections 1 (2021) 6

 Senior problems

S559. Let a1 , a2 , . . . , an be positive real numbers such that a1 + a2 + ⋯ + an ≤ n. Find the minimum value of
1 1 1
+ 2 +⋯+ n.
a1 2a2 nan

Proposed by Nguyen Viet Hung, Hanoi University of Science, Vietnam

Solution by the author

Using the AM-GM inequality we obtain
1
+ a1 ≥ 2,
a1
1 a2 a2 3
2
+ + ≥ ,
2a2 2 2 2
... ... ...,
1 an an
n
+ +⋯+ ≥ n + 1.
nan n n
´¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¸¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¶
n

Summing up these inequalities we obtain

1 1 1 n
k+1
+ 2 + ⋯ + n + a1 + a2 + ⋯ + an ≥ ∑ .
a1 2a2 nan k=1 k

From here and using the given condition we get

n
1 1 1 1
+ 2 + ⋯ + n ≥ ∑ (1 + ) − n
a1 2a2 nan k=1 k
n
1
=∑ .
k=1 k

The equality occurs if and only if a1 = a2 = ⋯ = an . Hence

n
1 1 1 1
min ( + 2 +⋯+ n) = ∑ .
a1 2a2 nan k=1 k

Also solved by Brian Bradie, Christopher Newport University, Newport News, VA, USA; Prodromos
Fotiadis, Nikiforos High School, Drama, Greece; Nicuşor Zlota, Traian Vuia Technical College, Focşani,
Romania; Arkady Alt, San Jose, CA, USA.

Mathematical Reflections 1 (2021) 7

S560. Let a, b, c be nonnegative real numbers such that a + b + c = 2. Prove that
a b c 10
+ + +8≥ .
b2 + bc + c2 c2 + ca + a2 a2 + ab + b2 ab + bc + ca

Proposed by Marius Stănean, Zalău, România

Solution by the author

After homogenizing, the inequality becomes as follows

a b c 16 5(a + b + c)
+ + + ≥ ,
b2 + bc + c2 c2 + ca + a2 a2 + ab + b2 a + b + c ab + bc + ca
or multiplying by a + b + c
a(a + b + c) 5(a + b + c)2
∑ + 16 ≥ ,
cyc b + bc + c ab + bc + ca
2 2

or
a2 + b2 + c2 + ab + bc + ca − b2 − bc − c2 5(a2 + b2 + c2 + ab + bc + ca)
∑ + 16 ≥ + 5,
cyc b2 + bc + c2 ab + bc + ca
or f (a, b, c) ≥ 5, where
ab + bc + ca 8(ab + bc + ca)
f (a, b, c) = ∑ + 2 2 2 .
cyc a + ab + b a + b + c + ab + bc + ca
2 2

Without loss of generality, assume that c = min{a, b, c}. Notice that we have equality for a = b, c = 0 which
leads us to prove a mixing of form f (a, b, c) ≥ f (a + c, b + c, 0). Indeed, we have

ab + bc + ca a + c c(ab − c2 )
− = ≥ 0,
b2 + bc + c2 b + c (b2 + bc + c2 )(b + c)

ab + bc + ca b + c c(ab − c2 )
− = ≥ 0,
c2 + ca + a2 a + c (c2 + ca + a2 )(a + c)

ab + bc + ca (a + c)(b + c)
−
a + ab + b (a + c) + (a + c)(b + c) + (b + c)2
2 2 2

ab + bc + ca (a + c)(b + c)
≥ 2 − 2
a + ab + b 2 a + ab + b2 + c(a + b)
abc(a + b + c)
= 2 ≥ 0,
(a + ab + b )(a2 + ab + b2 + c(a + b))
2

ab + bc + ca (a + c)(b + c)
−
a2 + b2
+ c + ab + bc + ca (a + c) + (a + c)(b + c) + (b + c)2
2 2

ab + bc + ca ab + bc + ca + c2
≥ 2 2 2 − 2 2 2
a + b + c + ab + bc + ca a + b + c + ab + bc + ca + c(a + b)
c(ab − c2 )(a + b + c)
= 2 2 2 ≥ 0.
(a + b + c + ab + bc + ca)(a2 + b2 + c2 + ab + bc + ca + c(a + b))

Mathematical Reflections 1 (2021) 8

 It remains to show that f (a + c, b + c, 0) ≥ 5, that is

a+c b+c 9(a + c)(b + c)

+ + ≥ 5,
b + c a + c (a + c) + (a + c)(b + c) + (b + c)2
2

a+c b+c
or if we denote t = + + 1 ≥ 3,
b+c a+c
9 (t − 3)2
+ t − 1 ≥ 5 ⇐⇒ ≥ 0,
t t
clearly true.

Also solved by Nicuşor Zlota, Traian Vuia Technical College, Focşani, Romania; Arkady Alt, San Jose,
CA, USA.

Mathematical Reflections 1 (2021) 9

S561. Let p be a prime number. Solve in positive integers the equation

(x2 − yz)3 + (y 2 − zx)3 + (z 2 − xy)3 − 3(x2 − yz)(y 2 − zx)(z 2 − xy) = p2 .

Proposed by Alessandro Ventullo, Milan, Italy

Solution by the author

Let P (x, y, z) = (x2 − yz)3 + (y 2 − zx)3 + (z 2 − xy)3 − 3(x2 − yz)(y 2 − zx)(z 2 − xy). Using the identity

a3 + b3 + c3 − 3abc = (a + b + c)(a2 + b2 + c2 − ab − bc − ca)

with a = x2 − yz, b = y 2 − zx and c = z 2 − xy, we have

P (x, y, z) = (x2 + y 2 + z 2 − xy − yz − zx)(x + y + z)2 (x2 + y 2 + z 2 − xy − yz − zx)

= [(x + y + z)(x2 + y 2 + z 2 − xy − yz − zx)]2
= (x3 + y 3 + z 3 − 3xyz)2 .

So, the given equation becomes

x3 + y 3 + z 3 − 3xyz = p. (1)
(i) If p = 2, then the equation becomes x3 + y 3 + z 3 − 3xyz = 2, i.e.

(x + y + z)(x2 + y 2 + z 2 − xy − yz − zx) = 2.

Since x, y, z are positive integers, then x + y + z ≥ 3, contradiction.

(ii) If p = 3, then the equation becomes x3 + y 3 + z 3 − 3xyz = 3, i.e.

(x + y + z)(x2 + y 2 + z 2 − xy − yz − zx) = 3.

Since x, y, z are positive integers, then x + y + z ≥ 3, which gives x + y + z = 3, i.e. x = y = z = 1. But then
x2 + y 2 + z 2 − xy − yz − zx = 0, contradiction.

(iii) If p > 3, equation (1) has been solved in T. Andreescu, D. Andrica, I. Cucurezeanu - An Introduction
p+1 p+1 p−2
to Diophantine Equations, pag. 9. We get the solutions (x, y, z) ∈ ( , , ) and the corresponding
3 3 3
p+2 p−1 p−1
permutations or (x, y, z) = ( , , ) and the corresponding permutations.
3 3 3
Note: The problem can be formulated for the particular case p = 2017.

Also solved by Kanav Talwar, Delhi Public School, Faridabad, India.

Mathematical Reflections 1 (2021) 10

S562. Let a, b, c, d be nonnegative real numbers. Prove that

(a + b + c + d)3 + 9 (abc + abd + acd + bcd) ≥ 4 (a + b + c + d) (ab + ac + ad + bc + bd + cd) .

Proposed by An Zhenping, Xianyang Normal University, China

Solution by Daniel Văcaru, Pites, ti, Romania

By calculus, we obtain

(a + b + c + d)3 + 9 (abc + abd + acd + bcd) ≥ 4 (a + b + c + d) (ab + ac + ad + bc + bd + cd) ⇔

a3 + b3 + c3 + d3 + 3 (abc + abd + acd + bcd) ≥ a2 b + a2 c + a2 d + ab2 + b2 c + b2 d + ac2 + bc2 + c2 d + ad2 + bd2 + cd2

Again, by calculus, this is equivalent to 4 (a3 + b3 + c3 + d3 )+15 (abc + abd + acd + bcd) ≥ (a + b + c + d)3 (∗).
Let E (a, b, c, d) = 4 (a3 + b3 + c3 + d3 ) + 15 (abc + abd + acd + bcd) − (a + b + c + d)3 . Without loss of gene-
rality, assume a ≤ b ≤ c ≤ d. We prove that

E (a, b, c, d) ≥ E (0, a + b, c, d) ≥ 0

We have E (a, b, c, d) − E (0, a + b, c, d) = 4 [a3 + b3 − (a + b)3 ] + 15ab (c + d) = 3ab [5 (c + d) − 4 (a + b) ≥ 0]

not
Set x = a+b; we need to show that E (0, x, c, d) ≥ 0,and E (0, x, c, d) = 4 (x3 + c3 + d3 )+15xcd−(x + c + d)3 .
But that is equivalent to Schur’s inequality

x3 + c3 + d3 + 3xcd ≥ xc (x + c) + cd (c + d) + dx (d + x) .

We have equality for a = 0 and b = c = d (or any cyclic permutation), and also for a = b = 0 and c = d (or
any permutation).

Mathematical Reflections 1 (2021) 11

S563. Let a, b, c be distinct real numbers. Prove that at least one of the numbers

1 2 1 2 1 2
(a + ) (1 − b4 ) , (b + ) (1 − c4 ) , (c + ) (1 − a4 )
a b c
is not equal to 4.

Proposed by Titu Andreescu, University of Texas at Dallas, USA

First solution by the author

Assume for the sake of contradiction that all three numbers are equal to 4. Then a2 , b2 , c2 are less than 1.
The first equality can be rewritten as
1 4
a2 + 2 = − 2.
a 1 − b4
π
Let a2 = tan u, b2 = tan v, c2 = tan w, where u, v, w ∈ (0, ). We have
4

tan2 u + 1 tan2 v + 1
=2 ,
tan u 1 − tan2 v
implying
2 tan u 1 − tan2 v
= .
1 + tan2 u 1 + tan2 v
π π
It follows that sin 2u = cos 2v and, since 2u, 2v ∈ (0, ), we have 2u + 2v = . Similarly, from the second
2 2
π
equality, 2v + 2w = , implying u = w, a contradiction. Hence the conclusion.
2

Second solution by the author

Assume by contradiction that all three numbers are equal to 4. The first equality can be rewritten as

2a2 1 − b4
= .
1 + a4 1 + b4
Then,
2 2
2a2 1 − b4
1−( ) = 1 − ( ) ,
1 + a4 1 + b4
i.e.
1 − a4 2b2
= .
1 + a4 1 + b4
However, since
1 − a4 2c2
= ,
1 + a4 1 + c4
we conclude that b = c, contradiction.

Also solved by Nicuşor Zlota, Traian Vuia Technical College, Focşani, Romania; Kanav Talwar, Delhi
Public School, Faridabad, India.

Mathematical Reflections 1 (2021) 12

S564. Let x, y, z be nonegative real numbers. Prove that
x3 + y 3 + z 3 + 3xyz 5 1 1 1
+ ≥ (xy + yz + zx)[ + + ].
∑cyc xy(x + y) 4 (x + y) 2 (y + z) 2 (z + x)2

Proposed by Marius Stănean, Zalău, Romania

Solution by the author

If one of x, y, z are equal with 0, say z, then the inequality becomes
(x − y)2
≥ 0,
4(x + y)2
clearly true.
If x, y, z are positive numbers, then a = y + z, b = z + x, c = x + y are side lengths of a triangle with the
circumradius R, the inradius r, and the semiperimeter s. It follows that x = s − a, y = s − b, z = s − c, and
since ab + bc + ca = s2 + r2 + 4Rr the inequality becomes
5 (r + 4Rr) [(ab + bc + ca) − 4abcs]
2 2
s3 − 3r(4R − r)s
+ ≥ ,
s(r2 + 4Rr) − 3sr2 4 16R2 r2 s2
or
s2 + 3r2 − 12Rr 5 (4R + r) [s − 2r(4R − r)s + r (4R + r) ]
4 2 2 2
+ ≥ ,
4Rr − 2r2 4 16R2 rs2
or
⎡ 2 ⎤
R2 s2 3r2 12r 5 4R + r ⎢⎢ s2 r2 4r R2 2r(4R−r) ⎥⎥
( + − )+ ≥ +( 2 + ) 2 −
2Rr−r2 R2 R2 R 2 8r ⎢⎢ R2 R R s R2 ⎥.
⎥
⎣ ⎦
Consider the functions f , defined as
⎡ 2 ⎤
s2 R2 s2 4R + r ⎢⎢ s2 r2 4r R2 ⎥⎥
f ( 2) = ( )− +( 2 + ) 2 ⎥ .
R 2Rr−r2 R2 8r ⎢⎢ R2 R R s ⎥
⎣ ⎦
Since
2
s2 R2 4R + r 4R + r r2 4r R4
f′ ( ) = − + ( ) ( + )
R2 2Rr−r2 8r 8r R 2 R s4
2
2R + r 4R + r r2 4r R4
= +( ) ( 2 + ) 4 ≥ 0,
8(2R − r) 8r R R s
we deduce that f is a increasing function.
If we denote t2 = 1 − 2r
R ∈ [0, 1), then by Blundon’s Inequality

s2 (1 − t2 )2 (1 − t)(t + 3)3
≥ 2 + 5(1 − t2
) − − 2t3
= .
R2 4 4
Hence, it suffices to prove that
4 (1 − t)(t + 3)3 3(1 − t2 )2 5
[ + − 6(1 − t2 )] +
(1 − t )(3 + t )
2 2 4 4 2
9 − t2 (1 − t)(t + 3)3 (1 − t2 )2 (9 − t2 )2 (1 − t2 )(7 + t2 )
≥ [ + − ],
8(1 − t2 ) 4 4(1 − t)(t + 3)3 2
or after some computations,
t2 (t − 1)2
≥ 0,
t2 + 3
clearly true. The equality holds when t = 0, so when the triangle is equilateral which means x = y = z.

Also solved by Nicuşor Zlota, Traian Vuia Technical College, Focşani, Romania.

Mathematical Reflections 1 (2021) 13

 Undergraduate problems

U559. Evaluate √
1
∫ 1 + dx
x

Proposed by Nguyen Viet Hung, Hanoi University of Science, Vietnam

Solution by the author

Letting √
1 1 −2t
1+ =t⇒x= 2 ⇒ dx = 2 dt.
x t −1 (t − 1)2
Hence
√
1 −2t2
∫ 1 + dx = ∫ dt
x (t2 − 1)2
1 1
= −2 ∫ ( 2 + 2 ) dt
t − 1 (t − 1)2
2
1 1 1 1
= −2 ∫ dt − 2 ∫ [ ( − )] dt
t2 − 1 2 t−1 t+1
dt 1 1 1 2
= −2 ∫ 2 − ∫ ( + − 2 ) dt
t −1 2 (t − 1) 2 (t + 1) 2 t −1
dt 1 1 1
= −∫ − ∫ ( + ) dt
t2 − 1 2 (t − 1)2 (t + 1)2
1 1 1 1 1 1
= ∫ ( − )− ∫ ( dt + ) dt
2 t+1 t−1 2 (t − 1)2 (t + 1)2
1 t+1 1 1 1
= ln ∣ ∣+ ( + )+C
2 t−1 2 t+1 t−1
1 (t + 1)2 t
= ln 2 + 2 +C
2 ∣t − 1∣ t − 1
⎡ √ 2⎤ √
1 ⎢⎢ ⎛ 1 ⎞ ⎥⎥ 1
= ln ⎢∣x∣ 1 + + 1 ⎥ + x 1 + + C.
2 ⎢ ⎝ x ⎠ ⎥ x
⎣ ⎦

Also solved by Brian Bradie, Christopher Newport University, Newport News, VA, USA; Alina Craciun,
Theoretical High School Miron Costin, Pascani, Romania; G. C. Greubel, Newport News, VA, USA; Henry
Ricardo, Westchester Area Math Circle, USA; Nicuşor Zlota, Traian Vuia Technical College, Focşani, Ro-
mania; Taes Padhihary, Disha Delphi Public School, Kota, Rajasthan, India; Arkady Alt, San Jose, CA,
USA.

Mathematical Reflections 1 (2021) 14

U560. Let 1, 1, 2, 3, 5, 8, ⋯ be the Fibonacci sequence. Evaluate

1 n cot ( Fj+1 )
1
lim ∑ .
n→∞ n j=0 cot ( 1 )
Fj+2

Proposed by Toyesh Prakash Sharma, St. C. F. Andrews School, Agra, India

Solution by G. C. Greubel, Newport News, VA, USA

The sum
n cot ( 1 )
Fj+1
Sn = ∑
j=0 cot ( F )
1
j+2

can also be seen in the form

n
1 1
Sn = ∑ cot ( ) cot ( ).
j=0 Fj+1 Fj+2
Using
1 1
cot ( ) = x − +⋯
x 3x
1 1 1
tan ( ) = + +⋯
x x 3 x2
then
n ⎛ Fj+1 Fj+1 1 ⎞
Sn = ∑ + − +⋯
j=0 ⎝ Fj+2 ⎠
3
3 Fj+2 3 Fj+1 Fj+2
n Fj+1
≈∑ + O(n0 )
j=0 Fj+2
n
1
≈∑ + O(n0 )
j=0 α
n+1
≈ + O(n0 ),
α
√
where Fn ≈ αn , 2α = 1 + 5, was used. Now,

1 n cot ( Fj+1 )
1
1 n+1
lim ∑ = lim ( + O(n0 ))
n→∞ n n→∞ n α
j=0 cot ( F )
1
j+2

1 1
= lim (1 + )
α n→∞ n
1
= .
α

Also solved by Brian Bradie, Christopher Newport University, Newport News, VA, USA; Toyesh Prakash
Sharma, St. C. F. Andrews School, Agra, India; Alina Craciun, Theoretical High School Miron Costin,
Pascani, Romania; Arkady Alt, San Jose, CA, USA.

Mathematical Reflections 1 (2021) 15

U561. The Fibonacci numbers Fn an the Lucas nubers Ln satisfy

Fn+2 = Fn+1 + Fn , F0 = 0, F1 = 1;

Ln+2 = Ln+1 + Ln , L0 = 2, L1 = 1.
Prove that
5
2Fn Fn+1 − 2Fn5 Fn+1 = Fn6 + Fn+1
6
− Fn+1
2
− Fn2 .

Proposed by Ángel Plaza, Universidad de Las Palmas de Gran Canaria, Spain

Solution by the author

The proposed identity may be written as

2Fn Fn+1 (Fn+1

4
− Fn4 ) = Fn2 (Fn4 − 1) + Fn+1
2
(Fn+1
4
− 1) .

Now, since Fn4 = Fn−2 Fn−1 Fn+1 Fn+2 − 1, we have

2Fn Fn+1 (Fn+1

4
− Fn4 ) = Fn2 (Fn4 − 1) + Fn+1
2
(Fn+1
4
− 1)
2Fn−1 Fn Fn+1 Fn+2 (Fn Fn+3 − Fn−2 Fn+1 ) = Fn2 (Fn−2 Fn−1 Fn+1 Fn+2 ) + Fn+1
2
(Fn−1 Fn Fn+2 Fn+3 )
2Fn−1 Fn Fn+1 Fn+2 (Fn Fn+3 − Fn−2 Fn+1 ) = Fn−1 Fn Fn+1 Fn+2 (Fn−2 Fn + Fn+1 Fn+3 ) .

So the problem is reduced to prove that

2 (Fn Fn+3 − Fn−2 Fn+1 ) = Fn−2 Fn + Fn+1 Fn+3 .

Since 5Fn Fm = Ln+m − (−1)m Ln−m , last identity is equivalent to

2 (L2n+3 − (−1)n L3 − L2n−1 + (−1)n L3 ) = L2n−2 − (−1)n L2 + L2n+4 − (−1)n+1 L2

2 (L2n+3 − L2n−1 ) = L2n+4 + L2n−2

which is true since both terms are equal to 10F2n+1 .

Also solved by Brian Bradie, Christopher Newport University, Newport News, VA, USA; G. C. Greubel,
Newport News, VA, USA; Nicuşor Zlota, Traian Vuia Technical College, Focşani, Romania; Arkady Alt, San
Jose, CA, USA.

Mathematical Reflections 1 (2021) 16

U562. Let ABCD be a square. The variable points P and Q are taken on the side AB and BC, respectively
such that ∠P DQ = 45○ . Find the locus of the orthocenter of the triangle P DQ.

Proposed by Mircea Becheanu, Canada

Solution by the author

We shall use coordinates. Take D ≡ O the origin of the coordinate axes, A(1, 0) ∈ (OX) and C(0, 1) ∈ (OY ).
Then, the points P and Q have coordinates P (1, s) and Q(t, 1) where 0 < s, t < 1 are variables related by
s+t
1= (2)
1 − st
Let H be the orthocenter of the triangle OP Q. In order to find faster the coordinates of H we remind the
problem J551: if AC intersects OP and OQ in E and F , respectively then QE ⊥ OP and P F ⊥ OQ. Hence,
we can find H to be the intersection point of QE and P F. The coordinates of the point E are obtained by
solving the system of linear equations:
x + y = 1 and y = sx.
We obtain the solution:
1 s
xE = ; yE = . (3)
1+s 1+s
From the system of equations
1
x + y = 1 and y = x
t
we obtain the coordinates
t 1
xF = ; yF = (4)
1+t 1+t
Using formulae (2) and (3), the equation of the line QE is x + sy = s + t and the equation of the line P F is
x + sy = s + t. Solving the system given by these equations one obtains:

xH = 1 − s; yH = 1 − t. (5)

Using the relation (1) we can eliminate the parameters s and t and we obtain that xH and yH satisfy the
equation
xy − 2x − 2y + 2 = 0.
We can write it in the form
(x − 2)(y − 2) = 2,
which shows, after a translation given by the vector v = (2, 2), that the locus is an equilateral hyperbola. It
passes through the points A and C.

Also solved by Miguel Amengual Covas, Cala Figuera, Mallorca, Spain; Brian Bradie, Christopher
Newport University, Newport News, VA, USA; Sarah B. Seales, Prescott, AZ, USA.

Mathematical Reflections 1 (2021) 17

U563. Find all polynomials P (x) with real coefficients such that, for all real numbers x,

(P (x2 ) + x) (P (x3 ) − x2 ) = P (x5 ) + x.

Proposed by Titu Andreescu, University of Texas at Dallas, USA

Solution by G. C. Greubel, Newport News, VA, USA

Let P (x) be a polynomial with the expansion P (x) = a0 + a1 x + a2 x2 + ⋯ for which

P (x2 ) + x = a0 + x + a1 x2 + a2 x4 + a3 x6 + a4 x8 + a5 x10 + a6 x12 + a7 x14 + ⋯

P (x3 ) − x2 = a0 − x + a1 x3 + a2 x6 + a3 x9 + a4 x12 + a5 x15 + ⋯
P (x5 ) + x = a0 + x + a1 x5 + a2 x10 + a3 x15 + ⋯

and leads to

a0 + x + a1 x5 + a2 x10 + a3 x15 + ⋯ =
a20 + a0 x + a0 (a1 − 1) x2 + (a0 a1 − 1) x3 + a0 a2 x4 + a21 x5 + ⋯.

Comparing the coefficients of the powers of x leads to the values of an being an ∈ {1, 1, 0, 0, 0, ⋯}n≥0 . This
yields P (x) = 1 + x.

Mathematical Reflections 1 (2021) 18

 4n
−3k 2 + 1 − k 4
U564. Evaluate lim (−1)n ⋅ sin ( ∑ arctan )
n→+∞
k=1 −k 2 + 7 + k 4

Proposed by Paolo Perfetti, Università degli studi di Tor Vergata, Roma, Italy

Solution by the author

−3k 2 + 1 − k 4 −1 + qk 4 − 2k 2
= ⇐⇒ q k =
−k 2 + 7 + k 4 1 + qk k4 + k2 + 3
thus
−3k 2 + 1 − k 4 −1 + qk
arctan = arctan = arctan(−1) + arctan qk =
−k + 7 + k
2 4 1 + qk
−π 4 − 2k 2
= + arctan 4
4 k + k2 + 3
Moreover

k+1 k−1
k+1
1+(k+1)2
− 1+(k−1)
k−1
2
arctan − arctan = arctan =
1 + (k + 1) 2 1 + (k − 1) 2
1 + 1+(k+1)2 1+(k−1)2
k+1 k−1

4 − 2k 2
= arctan
k4 + k2 + 3

and

n
−3k 2 + 1 − k 4 n
−π k+1 k−1
arctan ∑ = ∑ [ + arctan − arctan ]=
k=1 −k + 7 + k 1 + (k + 1) 1 + (k − 1)2
2 4 2
k=1 4
n ⎡
⎢
−nπ k+1 k
= + ∑ ⎢⎢arctan − arctan +
4 k=1⎢ 1 + (k + 1) 2 1 + k2
⎣
⎤
k k−1 ⎥
+ − arctan ⎥=
1 + (k − 1) ⎥⎥
arctan
1+k 2 2
⎦
−nπ n+1 1 n
= + arctan − arctan + arctan =
4 1 + (n + 1)2 2 1 + n2
−nπ
n+1
1+(n+1)2
+ 1+n
n
2 1
= + arctan − arctan =
4 1 − 1+(n+1)2 1+n2
n+1 n 2
3 2
−nπ 3n + 2n3 + 1 + 3n2 1
3n+2n +1+3n
2+2n2 +n4 +n+2n3
− 12
= + arctan − arctan = arctan 3 +1+3n2 1 =
4 2 + 2n2 + n4 + n + 2n3 2 1 + 3n+2n2 4 3
2+2n +n +n+2n 2
−nπ −n + 2n + 4n − 5n
4 3 2
= + arctan 4
4 2n + 6n3 + 7n2 + 5n + 5

Mathematical Reflections 1 (2021) 19

 4n
−3k 2 + 1 − k 4 −(4n)4 + 2(4n)3 + 4(4n)2 − 20n
∑ arctan = −nπ + arctan
k=1 −k 2 + 7 + k 4 2(4n)4 + 6(4n)3 + 7(4n)2 + 20n + 5
and finally
4n
−3k 2 + 1 − k 4
lim (−1)n sin [ ∑ arctan ]=
n→+∞
k=1 −k 2 + 7 + k 4
−(4n)4 + 2(4n)3 + 4(4n)2 − 20n
= lim (−1)n (−1)n sin [arctan ]=
n→+∞ 2(4n)4 + 6(4n)3 + 7(4n)2 + 20n + 5
−(4n)4 + 2(4n)3 + 4(4n)2 − 20n
= sin [arctan lim ]=
n→+∞ 2(4n)4 + 6(4n)3 + 7(4n)2 + 20n + 5
−1 −1 1 −1
= sin arctan = √ =√
2 2 1+ 1 5
4

Mathematical Reflections 1 (2021) 20

 Olympiad problems

O559. Let x, y, z be real numbers such that none of them lies in the open interval (−1, 1) and
1 1 1
+ + + x + y + z = 0.
x y z
z
Find the minimum value of .
x+y

Proposed by Marius Stănean, Zalău, România

Solution by the author

Notice that if {a, b, c} = {x, y, z}, then

1 c2 + 1 c
(a + b) (1 + )=− Ô⇒ < 0.
ab c a+b
Suppose that a ≥ b ≥ c, then obviously a ≥ 1, c ≤ −1 and
a+b+c ab + bc + ca 1 1 b
=− =− − > 0 Ô⇒ > −1.
a+c abc(a + c) ac b(a + c) a+c

We have 2 more cases to evaluate:

a
Case 1: b ≥ 1, then as above > −1.
b+c
1 1
Let us denote A = a + b + + ≥ 4, so c2 + Ac + 1 = 0. Solving the quadratic equation it follows that
√ a b
−A − A2 − 4
c= ≤ −1, and hence
2
√ ¿
2c −A − A − 4
2 1 Á À(1 + 1 ) −
2
4
= =−1− −Á
a+b a+b ab ab (a + b)2
¿
1 Á 2
À(1 + 1 ) − 1
≥−1− −Á
ab ab (ab)2
√
1 2 √
=−1− − 1+ ≥ −2 − 3,
ab ab
where above we used that
4 1
≥ ⇐⇒ (ab)2 − 1 + a2 (b2 − 1) + b2 (a2 − 1) + (ab − 1)2 ≥ 0,
(a + b)2 (ab)2
so √
c 3
≥ −1 − .
a+b 2
√
The equality holds when a = b = 1, c = −2 − 3.

Mathematical Reflections 1 (2021) 21

 c
Case 2: b ≤ −1, then as above > −1.
a+b
1 1
Let us denote B = −b − c − − ≥ 4, so a2 − Ba + 1 = 0. Solving the quadratic equation it follows that
√ b c
B + B2 − 4
a= ≥ 1, and hence
2
√ ¿
2a B + B2 − 4 1 Á 2
À(1 + 1 ) − 4
= =−1− −Á
b+c b+c bc bc (b + c)2
√
≥ − 2 − 3,

as in Case 1, so √
a 3
≥ −1 − .
b+c 2
√
The equality holds when a = 2 + 3, b = c = −1. √
We conclude that the required minimum is −1 − 3
2 .

Mathematical Reflections 1 (2021) 22

O560. Prove that there are infinitely many triples (a, b, c) of integers for which ab + bc + ca = 1 an that for
each such triple (a2 + 1) (b2 + 1) (c2 + 1) is a perfect square.

Proposed by Titu Andreescu, University of Texas at Dallas, USA

Solution by Daniel Văcaru, Pites, ti, Romania

⎧
⎪ ⎧
⎪
⎪a + b = 1 ⎪b = 1 − a
We have a2 + 1 = (a + b) (a + c) (1)We obtain, for example ⎨ ⇒ ⎨ .It is clear
⎪ ⎪c = a2 − a + 1
⎩a + c = a + 1
2
⎪ ⎪
⎩
that (a, 1 − a, a2 − a + 1) , a ∈ Zis (an) triple which fulfilled ab + bc + ca = 1. Observe relationships b2 + 1 =
(b + a) (b + c) (2) and c2 + 1 = (c + a) (c + b) .We obtain

(a2 + 1) (b2 + 1) (c2 + 1) = [(a + b) (b + c) (c + a)]2 ,

Also solved by Kanav Talwar, Delhi Public School, Faridabad, India; Prodromos Fotiadis, Nikiforos High
School, Drama, Greece; Ioan Viorel Codreanu, Satulung, Maramures, Romania; Mohammad Imran, India;
Prithwijit De, HBCSE, Mumbai, India; Arkady Alt, San Jose, CA, USA; Taes Padhihary, Disha Delphi
Public School, Rajasthan, India .

Mathematical Reflections 1 (2021) 23

O561. Let ABC be a triangle. Prove that

a2 b2 c2 2
+ + ≤ (a2 + b2 + c2 )
1 + cos B + cos C 1 + cos C + cos A 1 + cos A + cos B 3
2 2 2 2 2 2

Proposed by An Zhenping, Xianyang Normal University, China

Solution by Taes Padhihary, Disha Delphi Public School, Rajasthan, India

Observe that
a2 a2
=
1 + cos2 B + cos2 C 1 + ( c +a −b )2 + ( a2 +b2 −c2 )2
2 2 2
2ca 2ab
a2
≤ (c2 +a2 −b2 +a2 +b2 −c2 )2
1+ 4c2 a2 +4a2 b2
2
a
= a2
1+ b2 +c2
(ab) + (ac)2
2
= .
a2 + b2 + c2
Summing up the three relations, we obtain

2((ab)2 + (bc)2 + (ca)2 ) 2

LHS ≤ = 2 2 2 ((ab)2 + (bc)2 + (ca)2 )
a +b +c
2 2 2 a +b +c
2 (a2 + b2 + c2 )2
≤ 2 2 2⋅
a +b +c 3
2 2 2 2
≤ (a + b + c ).
3

Also solved by Ioan Viorel Codreanu, Satulung, Maramures, Romania; Arkady Alt, San Jose, CA, USA.

Mathematical Reflections 1 (2021) 24

O562. Find all functions f ∶ R ← R such that

x2 f (y) + f (y 2 f (x)) = xyf (x + y),

for all real numbers x, y.

Proposed by Prodromos Fotiadis, Nikiforos High School, Drama, Greece

Solution by the author

It’s clear that f (x) = x, ∀x ∈ R and f (x) = 0, ∀x ∈ R satisfy the given condition.
Let P (x, y) ∶ x2 f (y) + f (y 2 f (x)) = xyf (x + y) P (0, 0) ∶ f (0) = 0. Suppose that there exist t ∈ R∗ such that
f (t) = 0. P (t, t) ∶ f (2t) = 0.
t≠0
P (t, x) ∶ t2 f (x) = txf (x + t) ⇐⇒ tf (x) = xf (x + t) (1)

Since f (2t) = 0 we have similarly

2tf (x) = xf (x + 2t) (2)
Comparing (1) and (2) we have that 2f (x+t) = f (x+2t), ∀x ∈ R∗ . Also 2f (0+t) = 0 = f (2t) = f (2t) = f (0+2t)
so 2f (x) = f (x + t), ∀x ∈ R. Using (1) we have that tf (x) = 2xf (x) which means that f (x) = 0 or x t/2. If
t
f (t/2) = 0 then f (x) = 0, ∀x ∈ R. If f (t/2) ≠ 0 then putting x = −t/2 in (1) we get tf (−t/2) = f (t/2) ≠ 0,
2
a contradiction since −t/2 ≠ t/2 ⇒ f (−t/2) = 0. Suppose now that there is no such t, i.e. f (u) = 0 ⇔ u = 0.
P (x, −y) ∶ x2 f (−y) + f (y 2 f (x)) = −xyf (x − y). For the following suppose that x, y ≠ 0. Let Q(x, y) ≡
P (x, y) − P (x, −y) i.e. Q(x, y) ∶ x(f (y) − f (−y)) = yf (x + y) + yf (x − y).

Q(x, x) ∶ x(f (x) − f (−x)) = xf (2x) + xf (0) ⇔ f (x) − f (−x) = f (2x) (3)

Q(−x, x) ∶ −x(f (x) − f (−x)) = xf (0) + xf (−2x) ⇔ f (x) − f (−x) = −f (−2x) (4)
x
Comparing (3), (4) we get f (−2x) = −f (2x) i.e. f is odd. Now Q(x, y) becomes R(x, y) ∶ 2 f (y) = f (x+y)+
y
f (x−y). R(y, x)+R(x, y) ⇒ x f (y)+y f (x) = xyf (x+y). However, P (x, y) ∶ x f (y)+f (y f (x)) = xyf (x+y)
2 2 2 2

so f (y 2 f (x)) = y 2 f (x). Put x = 1 ∶ f (y 2 f (1)) = y 2 f (1). Also f (−y 2 f (1)) = −f (y 2 f (1)) = −y 2 f (1).
Since f (1) ≠ 0 we have that f (x) = x, ∀x ∈ R.
Therefore f (x) = x, ∀x ∈ R and f (x) = 0, ∀x ∈ R are the only solutions.

Also solved by Kanav Talwar, Delhi Public School, Faridabad, India.

Mathematical Reflections 1 (2021) 25

O563. Prove that in any triangle ABC,
√ √ √
ma mb mc 6(ab + bc + ca)
+ + + ≥ 5.
ha hb hc (a + b + c)2

Proposed by Nguyen Viet Hung, Hanoi University of Science, Vietnam

Solution by the author

b2 + c2
ma ≥ .
4R
This result can be rewritten as
abc b2 + c2 S b2 + c2 b2 + c2
ma ≥ ⋅ = ⋅ = ha ⋅ .
4R abc a bc 2bc
It follows that
ma b2 + c2
≥ .
ha 2bc
Therefore
√ √
ma b2 + c2
≥
ha 2bc
b2 + c2
=√
(2bc)(b2 + c2 )
2(b2 + c2 )
≥ .
(b + c)2

It is enough to show that

2(b2 + c2 ) 6(ab + bc + ca)
∑ + ≥ 5.
cyc (b + c) (a + b + c)2
2

This is equivalent to
2(b2 + c2 ) 6(ab + bc + ca)
∑( − 1) + − 2 ≥ 0,
cyc (b + c) 2 (a + b + c)2
or
(b − c)2 2(ab + bc + ca − a2 − b2 − c2 )
∑ + ≥ 0,
cyc (b + c) (a + b + c)2
2

or
(b − c)2 (b − c)2 + (c − a)2 + (a − b)2
∑ − ≥ 0,
cyc (b + c) (a + b + c)2
2

or
1 1
∑( − ) (b − c)2 ≥ 0
cyc (b + c) 2 (a + b + c)2
which is true because
1 1 a(a + 2b + 2c)
− = ≥ 0.
(b + c) 2 (a + b + c) 2 (b + c)2 (a + b + c)2
The proof is completed. The equality occurs is and only if the triangle ABC is equilateral.

Also solved by Ioan Viorel Codreanu, Satulung, Maramures, Romania.

Mathematical Reflections 1 (2021) 26

O564. Let n ≥ 3 be an integer. For every sequence −1 = x1 < x2 < ⋯ < xn = 1 of real numbers and every
k = 1, 2, . . . , n we define

Dk (x2 , . . . , xn−1 ) = ∣(xk − x1 )∣ . . . (xk − xk−1 (xk − xk+1 . . . (xk − xn )))

and denote
rn = max x2 ,...,xn−1 min k Dk (x2 , . . . , xn−1 ).
Assume that rn is achieved at the sequence −1 = a1 < a2 < ⋯ < an = 1. Prove that

D2 (a2 , . . . , an−1 ) = ⋯ = Dn−1 (a2 , . . . , an−1 ).

Proposed by Navid Safaei, Sharif University of Technology, Tehran, Iran

Solution by the author

At the outset, we shall prove that for each k = 1, . . . , n − 1, we have

min(Dk (a2 , . . . , an−1 ), Dk+1 (a2 , . . . , an−1 )) = rn .

Assume by contradiction that min(Dk (a2 , . . . , an−1 ), Dk+1 (a2 , . . . , an−1 )) > rn . Let’s define

⎧
⎪ak + ε, i=k>1
⎪
⎪
⎪
bi = ⎨ak+1 − ε i = k + 1 < n
⎪
⎪
⎪
⎪
⎩ai otherwise.

ak+1 − ak
For some ε > 0 that satisfies ε < .
2
Then, the fact that min(Dk (b2 , . . . , bn−1 ), Dk+1 (b2 , . . . , bn−1 )) > rn remains unchanged. Now, for i ≠
k, k + 1, 2 ≤ i ≤ n − 1, one can find that Dk (b2 , . . . , bn−1 ) − Di (a2 , . . . , an−1 ) =
RRR RRR
RRR n RRR
= RRR ∏ (ai − aj )RRRRR ∣(ai − ak − ε)(ai − ak+1 + ε) − (ai − ak )(ai − ak+1 )∣
R
R
RRR j=1 RRR
RRRj≠i,k,k+1 RRR
RRR RRR
RRR n RRR
= RRR ∏ (ai − aj )RRRRR ∣ε(ak+1 − ak − ε)∣ > 0.
R
R
RRR j=1 RRR
RRRj≠i,k,k+1 RRR
This means min 1≤i≤n Di (b2 , . . . , bn−1 ) = min 2≤i≤n−1 Di (b2 , . . . , bn−1 ) > rn . This contradicts with the choice
of rn . Thus,
D2 (a2 , . . . , an−1 ) = ⋯ = Dn−1 (a2 , . . . , an−1 ) = rn .

Mathematical Reflections 1 (2021) 27