International Journal of Mathematical Education in

Science and Technology

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/tmes20

Some variants of Cauchy's mean value theorem

German Lozada-Cruz

To cite this article: German Lozada-Cruz (2020) Some variants of Cauchy's mean value theorem,
International Journal of Mathematical Education in Science and Technology, 51:7, 1155-1163, DOI:
10.1080/0020739X.2019.1703150

To link to this article: https://doi.org/10.1080/0020739X.2019.1703150

Published online: 23 Dec 2019.

Submit your article to this journal

Article views: 432

View related articles

View Crossmark data

Citing articles: 1 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=tmes20
INTERNATIONAL JOURNAL OF MATHEMATICAL EDUCATION IN SCIENCE AND TECHNOLOGY
2020, VOL. 51, NO. 7, 1155–1163
https://doi.org/10.1080/0020739X.2019.1703150

CLASSROOM NOTE

Some variants of Cauchy’s mean value theorem

German Lozada-Cruz
Departamento de Matemática, Instituto de Biociências, Letras e Ciências Exatas (IBILCE) – Universidade
Estadual Paulista (UNESP), São José do Rio Preto, Brazil

ABSTRACT ARTICLE HISTORY

In this note, some variants of Cauchy’s mean value theorem are Received 1 October 2019
proved. The main tools to prove these results are some elementary KEYWORDS
auxiliary functions. Wachnicki’s theorem; Myers’
theorem; Sahoo–Riedel’s
theorem
2010 MATHEMATICS
SUBJECT
CLASSIFICATIONS
26A06; 26A24

1. Introduction
The first contact of our mathematics or engineering undergraduate students with the mean
value theory is in a Differential and Integral Calculus course or in a first Real Analysis
course. The first mean value theorem is the famous Lagrange’s Mean Value Theorem, which
relates the average rate of change of a function at the end of an interval with the value of the
derivative of the function at a point in the same interval. The second mean value theorem
is Cauchy’s Mean Value Theorem, which is a generalization of the Lagrange’s Mean Value
Theorem, it establishes the relationship between the derivatives of two functions and the
variation of these functions on a finite interval.
There are a whole range of results that relate the values of the function and its deriva-
tive or integral. These are called mean value type theorems, and some of these also have
a nice geometric interpretation. In textbooks (Lang, 2005, Theorem 2.3; Protter & Mor-
rey, 1991, Theorem 4.12), some auxiliary functions are used in the proofs of these theorems
and then directly apply Rolle’s theorem, some authors call this ‘happy idea’ (Martínez de
la Rosa, 2008, p. 24). Thinking about these issues and in a way to develop students’ skills
and competencies in manipulating these functions and, consequently, understanding the
proofs in this work, this is the main goal of this note.

Theorem 1.1 (Cauchy’s Mean Value Theorem): Let f , g : [a, b] → R be continuous func-
tions on [a, b], differentiable on (a, b) and g  (x) = 0 for all x ∈ (a, b), then there exists

CONTACT German Lozada-Cruz german.lozada@unesp.br Departamento de Matemática, Instituto de

Biociências, Letras e Ciências Exatas (IBILCE) – Universidade Estadual Paulista (UNESP), São José do Rio Preto, São Paulo
15054-000, Brazil

© 2019 Informa UK Limited, trading as Taylor & Francis Group

1156 G. LOZADA-CRUZ

η ∈ (a, b) such that

f (b) − f (a) f  (η)
=  . (1)
g(b) − g(a) g (η)

In 1958, T.M. Flett (1958) proved the following variant of Lagrange’s mean value
theorem.

Theorem 1.2 (Flett’s Theorem): Let f : [a, b] → R be differentiable on [a, b] and f  (a) =
f  (b), then there exists η ∈ (a, b) such that

f (η) − f (a) = f  (η)(η − a). (2)

In 1977, R.E. Myers (Myers, 1977, Theorem 1 ) proved a slight modification of Flett’s
theorem.

Theorem 1.3 (Myers’ Theorem): Let f : [a, b] → R be differentiable on [a, b] and f  (a) =
f  (b), then there exists η ∈ (a, b) such that

f (b) − f (η) = f  (η)(b − η). (3)

In 1998, P.K. Sahoo and T. Riedel (Sahoo & Riedel, 1998, Theorem 5.2) removed the
boundary condition on f  and obtained a generalization of Flett’s theorem.

Theorem 1.4 (Sahoo–Riedel’s Theorem): Let f : [a, b] → R be differentiable on [a, b],

then there exists η ∈ (a, b) such that

1 f  (b) − f  (a)
f (η) − f (a) = f  (η)(η − a) − (η − a)2 . (4)
2 b−a

In 2012, D.Çakmak and A.Tiryaki (Çakmak & Tiryaki, 2012, Theorem 2.1) proved a
slight modification of Sahoo–Riedel’s theorem which it reduces to Myers’ Theorem when
f  (a) = f  (b).

Theorem 1.5 (Çakmak–Tiryaki’s Theorem): Let f : [a, b] → R be differentiable on [a, b],

then there exists η ∈ (a, b) such that

1 f  (b) − f  (a)
f (b) − f (η) = f  (η)(b − η) + (b − η)2 . (5)
2 b−a

2. Variants of Cauchy’s mean value theorem

In this section, we prove some variants and generalizations of Cauchy’s mean value
theorem.
In 2000, E. Wachnicki (Wachnicki, 2000, Teorema 1.3) proved the following variant of
Cauchy’s mean value theorem.
 INTERNATIONAL JOURNAL OF MATHEMATICAL EDUCATION IN SCIENCE AND TECHNOLOGY 1157

Theorem 2.1 (Wachnicki’s Theorem): Let f , g : [a, b] → R be differentiable on [a, b].

Suppose that g  (x) = 0 for all x ∈ [a, b] and
f  (a) f  (b)
= . (6)
g  (a) g  (b)
Then, there exists η ∈ (a, b) such that
f (η) − f (a) f  (η)
=  . (7)
g(η) − g(a) g (η)

Also, he removed the condition (6) and obtained the following result which is an
analogous result to Sahoo–Riedel’s theorem.

Theorem 2.2: Let f , g : [a, b] → R be differentiable on [a, b]. Suppose that g  (x) = 0 for all
x ∈ [a, b] Then, there exists η ∈ (a, b) such that
 
f (η) − f (a) f  (η) 1 f  (b) f  (a) g(η) − g(a)
=  − − . (8)
g(η) − g(a) g (η) 2 g  (b) g  (a) g(b) − g(a)

Remark 2.1: Assuming that g is an increasing function, M. Ivan (2002) proved that
Wachnick’s Theorem is nothing else than Flett’s theorem applied to the particular func-
tion F = f ◦ g −1 and Theorem 2.2 is in fact Sahoo–Riedel’s theorem applied to the same
function F.

Now we prove our first result, that is, a variant of the Theorem 2.2.

Theorem 2.3: Let f , g : [a, b] → R be differentiable on [a, b]. Suppose that g  (x) = 0 for all
x ∈ [a, b]. Then, there exists η ∈ (a, b) such that
  
f (η) − f (a) f  (η) n − 1 f  (b) f  (a) g(η) − g(a) n−1
=  − − , n ∈ N. (9)
g(η) − g(a) g (η) n g  (b) g  (a) g(b) − g(a)

Proof: Let n ∈ N and consider the auxiliary function ψ : [a, b] → R given by

ψ(x) = f (x) + λ(g(x) − g(a))n , λ ∈ R.

We choose λ in such a way to satisfy the condition

ψ  (a) ψ  (b)
= . (10)
g  (a) g  (b)
The function ψ is differentiable on [a, b] and

ψ  (x) = f  (x) + λn(g(x) − g(a))n−1 g  (x).

Thus

ψ  (a) = f  (a),
ψ  (b) = f  (b) + λn[g(b) − g(a)]n−1 g  (b).
1158 G. LOZADA-CRUZ

Now, from (10) we get

ψ  (a) ψ  (b) f  (a) f  (b)

= ⇔ = + λn [g(b) − g(a)]n−1
g  (a) g  (b) g  (a) g  (b)
 
1 f  (b) f  (a) 1
⇔λ=− 
− 
n g (b) g (a) (g(b) − g(a))n−1

and the function ψ takes the form

 
1 f  (b) f  (a) (g(x) − g(a))n
ψ(x) = f (x) − − .
n g  (b) g  (a) (g(b) − g(a))n−1

Now applying Wachnicki’s theorem (Theorem 2.1), there exists η ∈ (a, b) such that

ψ(η) − ψ(a) ψ  (η)

=  .
g(η) − g(a) g (η)

Where we immediately get equality (9). 

Note that:
If n = 1 in (9), we get the Wachnicki’s Theorem (Theorem 2.1).
If n = 2 in (9), we get the Theorem 2.2. In this case we have the auxiliary function
 
1 f  (b) f  (a) (g(x) − g(a))2
ψ(x) = f (x) − −
2 g  (b) g  (a) g(b) − g(a)

(Wachnicki, 2000, Theorem 4).

If n = 3 in (9), we get a slight modification of Theorem 2.2
  
f (η) − f (a) f  (η) 2 f  (b) f  (a) g(η) − g(a) 2
=  − − . (11)
g(η) − g(a) g (η) 3 g  (b) g  (a) g(b) − g(a)

In this case, we have the auxiliary function

 
1 f  (b) f  (a) (g(x) − g(a))3
ψ(x) = f (x) − − .
3 g  (b) g  (a) (g(b) − g(a))2

Remark 2.2: If g(x) = x in (7) and (8), then Theorems 2.1 and 2.2 become in Flett’s
Theorem (Theorem 1.2) and Sahoo–Riedel’s Theorem (Theorem 1.4), respectively.

Remark 2.3: If g(x) = x in (9), then Theorem 2.3 becomes in a variant of Sahoo–Riedel’s
Theorem (see Lozada-Cruz, 2019), that is,

n − 1 f  (b) − f  (a)
f (η) − f (a) = f  (η)(η − a) − (η − a)n , n ∈ N. (12)
n (b − a)n−1

Now we prove our second result, that is, a variant of Wachnicki’s Theorem.
 INTERNATIONAL JOURNAL OF MATHEMATICAL EDUCATION IN SCIENCE AND TECHNOLOGY 1159

Theorem 2.4: Let f , g : [a, b] → R be differentiable on [a, b]. Suppose that g  (x) = 0 for all
x ∈ [a, b] and
f  (a) f  (b)
= . (13)
g  (a) g  (b)
Then, there exists η ∈ (a, b) such that
f (b) − f (η) f  (η)
=  . (14)
g(b) − g(η) g (η)
Proof: Consider the function G : [a, b] → R given by
⎧
⎪
⎪ f (b) − f (x)
⎨ , if x ∈ [a, b)
g(b) − g(x)
G(x) = f  (b)
⎪
⎪
⎩  , if x = b.
g (b)
Note that G is continuous on [a, b], differentiable on [a, b) and
−(g(b) − g(x))f  (x) + g  (x)(f (b) − f (x))
G (x) = ,
(g(b) − g(x))2
or equivalently
−f  (x) G(x)g  (x)
G (x) = + , x ∈ [a, b). (15)
g(b) − g(x) g(b) − g(x)
To get (14) is sufficient to proof that there exists η ∈ (a, b) such that G (η) = 0. Assume
that G (x) = 0, ∀x ∈ (a, b). If G (x) > 0, ∀x ∈ (a, b), then G is strictly increasing in [a, b].
Thus
G(a) < G(b). (16)
Now,
f  (b)
G(b) = .
g  (b)
From (15) and using (14), we have
f  (a) g(b) − g(a)  f  (b) g(b) − g(a) 
G(a) = 
+ 
G (a) =  + G (a).
g (a) g (a) g (b) g  (a)
From (16), we have
g(b) − g(a) 
G (a) < 0. (17)
g  (a)
Assume that g  (x) > 0, ∀x ∈ [a, b]. Then, there exists ξ ∈ (a, b) such that g(b) − g(x) =
g  (ξ )(b − x), ∀x ∈ [a, b). Thus g(b) > g(x), ∀x ∈ [a, b). Hence, (g(b) − g(x))/g  (x) > 0,
∀x ∈ [a, b). In particular,
g(b) − g(a)
> 0. (18)
g  (a)
Finally, using (18) in (17) we have G (a) < 0. But this is a contradiction because G (x) > 0,
for all x ∈ [a, b). The proof is completed because in the case G (x) < 0 for all x ∈ (a, b),
the reasoning is analogous. 
1160 G. LOZADA-CRUZ

Removing the condition (13), we get a generalization of Theorem 2.4, which is an

analogous result to the Ç akmak–Tiryaki’s Theorem (Theorem 1.5).
Our third result is the following result.

Theorem 2.5: Let f , g : [a, b] → R be differentiable [a, b]. Suppose that g  (x) = 0 for all
x ∈ [a, b] Then, there exists η ∈ (a, b) such that
 
f (b) − f (η) f  (η) 1 f  (b) f  (a) g(b) − g(η)
=  + − . (19)
g(b) − g(η) g (η) 2 g  (b) g  (a) g(b) − g(a)

Proof: Consider the auxiliary function φ : [a, b] → R given by

 
1 f  (b) f  (a) ((g(x) − g(b))2
φ(x) = f (x) − − .
2 g  (b) g  (a) g(b) − g(a)
The function φ is differentiable on [a, b] and
   
  f (b) f  (a) g(x) − g(b) 
φ (x) = f (x) −  − g (x).
g (b) g  (a) g(b) − g(a)
Then
f  (b) 
φ  (a) = g (a),
g  (b)
φ  (b) = f  (b).

Therefore the condition

φ  (a) φ  (b)
=
g  (a) g  (b)
is satisfied. Since the function φ and g satisfy the hypotheses of Theorem 2.4, there exists
η ∈ (a, b) such that
φ(b) − φ(η) φ  (η)
=  , (20)
g(b) − g(η) g (η)
which implies

1 f  (b) f  (a) (g(η)−g(b))2
f (b) − f (η) − 2 g  (b) − g  (a) g(b)−g(a)
g(b) − g(η)
f  (b) f  (a)
f  (η) − g  (η)
g(η)−g(b)
g  (b) − g  (a) g(b)−g(a)
=
g  (η)
  
f (b) − f (η) 1 f  (b) f  (a) g(b) − g(η)
+ −
g(b) − g(η) 2 g  (b) g  (a) g(b) − g(a)
     
f (η) f (b) f (a) g(b) − g(η)
=  +  −
g (η) g (b) g  (a) g(b) − g(a)
  
f (b) − f (η) f (η) 1 f (b) f  (a) g(b) − g(η)

=  + − . 
g(b) − g(η) g (η) 2 g  (b) g  (a) g(b) − g(a)
 INTERNATIONAL JOURNAL OF MATHEMATICAL EDUCATION IN SCIENCE AND TECHNOLOGY 1161

Our last result in this note is a variant of Theorem 2.5.

Theorem 2.6: Let f , g : [a, b] → R be differentiable on [a, b]. Suppose that g  (x) = 0 for all
x ∈ [a, b]. Then, there exists η ∈ (a, b) such that
  n−1
f (b) − f (η) f  (η) n − 1 f  (b) f  (a) g(b) − g(η)
=  + − , n ∈ N. (21)
g(b) − g(η) g (η) n g  (b) g  (a) g(b) − g(a)

Proof: Let n ∈ N and consider the auxiliary function φ : [a, b] → R given by

φ(x) = f (x) + λ(g(x) − g(b))n , λ ∈ R.

We choose λ in such a way to satisfy the condition

φ  (a) φ  (b)

=  . (22)
g (a) g (b)

The function φ is differentiable on [a, b] and

φ  (x) = f  (x) + λn(g(x) − g(b))n−1 g  (x).

Thus

φ  (a) = f  (a) + nλ[g(a) − g(b)]n−1 g  (a)

φ  (b) = f  (b).

Now, from (22) we get

φ  (a) φ  (b) f  (a) n−1 f  (b)

= ⇔ + nλ[g(a) − g(b)] =
g  (a) g  (b) g  (a) g  (b)
  
1 f (b) f  (a) 1
⇔λ= −
n g  (b) g  (a) (g(a) − g(b))n−1

and the function φ takes the form

 
1 f  (b) f  (a) (g(x) − g(b))n
φ(x) = f (x) + − .
n g  (b) g  (a) (g(a) − g(b))n−1

Since the functions φ and g satisfy the hypotheses of Theorem 2.4, there exists η ∈ (a, b)
such that
φ(b) − φ(η) φ  (η)
=  , (23)
g(b) − g(η) g (η)
1162 G. LOZADA-CRUZ

which implies

1 f  (b) f  (a) (g(η)−g(b))n
f (b) − f (η) + n g  (b) − g  (a) (g(a)−g(b))n−1
g(b) − g(η)
f  (b) f  (a) (g(η)−g(b))n−1
f  (η) + g  (b) − g  (a) (g(a)−g(b))n−1
g  (η)
=
g  (η)
  
f (b) − f (η) 1 f  (b) f  (a) (g(η) − g(b))n
− −
g(b) − g(η) n g  (b) g  (a) (g(b) − g(η))(g(a) − g(b))n−1
   
f  (η) f  (b) f  (a) (g(η) − g(b))n−1
=  +  − . (24)
g (η) g (b) g  (a) (g(a) − g(b))n−1

If n is even, from (24) we get

  
f (b) − f (η) 1 f  (b) f  (a) (g(b) − g(η))n−1
+ −
g(b) − g(η) n g  (b) g  (a) (g(b) − g(a))n−1
   
f  (η) f (b) f  (a) (g(b) − g(η))n−1
=  +  − 
g (η) g (b) g (a) (g(b) − g(a))n−1
  
f (b) − f (η) f  (η) n − 1 f  (b) f  (a) g(b) − g(η) n−1
=  + − .
g(b) − g(η) g (η) n g  (b) g  (a) g(b) − g(a)

If n is odd, from (24) we get

  
f (b) − f (η) 1 f  (b) f  (a) (g(b) − g(η))n−1
+ −
g(b) − g(η) n g  (b) g  (a) (g(b) − g(a))n−1
   
f  (η) f (b) f  (a) (g(b) − g(η)))n−1
=  +  − 
g (η) g (b) g (a) (g(b) − g(a))n−1
  
f (b) − f (η) f  (η) n − 1 f  (b) f  (a) g(b) − g(η) n−1
=  + − .
g(b) − g(η) g (η) n g  (b) g  (a) g(b) − g(a)

Note that:
If n = 1 in (21), we get Theorem 2.4.
If n = 2 in (21), we get Theorem 2.5.
If n = 3 in (21), we get a slight variant of Theorem 2.5
  2
f (b) − f (η) f  (η) 2 f  (b) f  (a) g(b) − g(η)
=  + − .
g(b) − g(η) g (η) 3 g  (b) g  (a) g(b) − g(a)

In this case, we have the auxiliary function

 
1 f  (b) f  (a) (g(x) − g(b))3
φ(x) = f (x) + − .
3 g  (b) g  (a) (g(a) − g(b))2
 INTERNATIONAL JOURNAL OF MATHEMATICAL EDUCATION IN SCIENCE AND TECHNOLOGY 1163

3. Conclusion
The use of elementary auxiliary functions to prove type mean value theorems is an
effective didactic resource. This has already been seen in the proof of Lagrange’s mean
value theorem, see references Mercer (2002), Myers (1977), Sahoo and Riedel (1998) and
Tong (2005), for example.
In this note using some elementary auxiliary functions and Differential Calculus some
variants of Cauchy’s mean value theorem were proved (Theorems 2.3–2.6).

Acknowledgments
The author sincerely thanks the referees for their valuable suggestions improving this note.

Disclosure statement
No potential conflict of interest was reported by the authors.

ORCID
German Lozada-Cruz http://orcid.org/0000-0003-2860-954X

References
Çakmak D., & Tiryaki A. (2012). Mean value theorem for holomorphic functions. Electronic Journal
of Differential Equatiions, 2012(34), 1–6.
Flett T. M. (1958). A mean value problem. Mathematical Gazette, 42, 38–39.
Ivan M. (2002). A note on a Cauchy-type mean value theorem. Demonstratio Mathematica, 35(3),
493–494.
Lang S. (2005). Undergraduate analysis (2nd ed.) Undergraduate texts in mathematics. New York:
Springer-Verlag.
Lozada-Cruz G. (2019). Some variants of Lagrange’s mean value theorem.
Martínez de la Rosa F. (2008). Panorámica de los Teoremas de Valor Medio. Miscelánea Matemática,
47, 23–38.
Mercer P. R. (2002). On a mean value theorem. College Mathematics Journal, 33(1), 46–47.
Myers R. E. (1977). Some elementary results related to the mean value theorem. Two-Year College
Math. J., 8(1), 51–53.
Protter M. H., & Morrey Jr. C. B. (1991). A first course in real analysis (2nd ed.) Undergraduate Texts
in Mathematics, New York: Springer-Verlag, Inc.
Sahoo P. K., & Riedel T (1998). Mean value theorems and functional equations. River Edge (NJ):
World Scientific.
Tong J. (2005). A new auxiliary function for the mean value theorem. Journal of North Carolina
Academy Sciences, 121(4), 174–176.
Wachnicki E. (2000). Une variante du théorème de Cauchy de la valeur moyenne. Demonstratio
Mathematica, 33(4), 737–740.