MA4101 — Real Analysis: Series of Functions

and Calculus of Several Variables

• Differentiation of functions

Dr. Hiranmoy Pal1

September 10, 2025
1 Department of Mathematics
National Institute of Technology Rourkela
Directional derivative

Let φ : D ⊆ R → R. Now φ is differentiable at X0 if

φ(X0 + t) − φ(X0 )
lim exists.
t→0 t

Definition. Let f : D ⊆ Rm → Rn and X0 be an interior point of

D. For a unit vector u ∈ Rm , if
f (X0 + tu) − f (X0 )
lim exists,
t→0 t
then the limiting value, denoted by Du f (X0 ), is called the
directional derivative of f at X0 along the direction u.

1 / 100
Example. Let f : R2 → R be defined by

f (x, y) = x2 + y 2

Find the directional derivaive of f at (x0 , y0) ∈ R2 along

 the
1 1
direction u1 = (1, 0), u2 = (0, 1) and u3 = √2 , √2 .

2 / 100
Solution. Observe that
f ((x0 , y0 ) + t(1, 0)) − f (x0 , y0 )
lim
t→0 t
f (x0 + t, y0 ) − f (x0 , y0 )
= lim
t→0 t
((x0 + t)2 + y02 ) − (x20 + y02 )
= lim = 2x0 .
t→0 t
The limit exists, and hence

Du1 f (x0 , y0 ) = 2x0 .

Similarly,
√
Du2 f (x0 , y0 ) = 2y0 , Du3 f (x0 , y0 ) = 2(x0 + y0 ).

3 / 100
Problem. [State true or false]
Let f : D ⊆ Rn → R. Suppose D contains a nbd of X0 .

• If the directional derivatives of f at X0 along the principal

directions (that is, along u1 = (1, 0) and u2 = (0, 1)) exists,
then f is continuous at X0 .
• Is the converse true?

4 / 100
Example. Let f : R2 → R be defined by
 xy

2 2
, if (x, y) 6= (0, 0)
f (x, y) = x + y
0, if (x, y) = (0, 0)

Here f is not continuous at (0, 0). However

f ((0, 0) + t(1, 0)) − f (0, 0) f (t, 0) − f (0, 0)

Du1 f (0, 0) = lim = lim
t→0 t t→0 t
0−0
= lim = 0.
t→0 t

Similarly, Du2 f (0, 0) = 0 Hence, both the directional derivative

along the principal direction exists at (0, 0).

5 / 100
Other directions?

Consider a unit vector v = (v1 , v2 ). Then

f ((0, 0) + t(v1 , v2 )) − f (0, 0) f (tv1 , tv2 ) − f (0, 0)

lim = lim
t→0 t t→0 t
(tv1 )(tv2 )
−0
(tv1 )2 + (tv2 )2
= lim
t→0 t
v1 v2
= lim
t→0 t

The limit does not exist whenever v1 v2 6= 0.

6 / 100
Example. f (x, y) = |x| + |y| is continuous at origin. Now,

f ((0, 0) + t(1, 0)) − f (0, 0)

lim
t→0 t
f (t, 0) − f (0, 0)
= lim
t→0 t
|t|
= lim does not exist.
t→0 t

The function f is continuous at (0, 0), however, both the

directional derivatives do not exist at (0, 0) along the principal
directions.

7 / 100
Problem. [State true or false]
Let f : D ⊆ Rn → R. Suppose D contains a nbd of X0 .

• If the directional derivative at X0 along every direction exist,

then f is continuous.
• Is the converse true?

8 / 100
Example. Suppose f : R2 → R be defined by

2
 xy , if (x, y) =

6 (0, 0)
f (x, y) = x2 + y 4
0,

if (x, y) = (0, 0)

Let u = (u1 , u2 ) be a unit vector. Then

f (tu1 , tu2 ) − f (0, 0) u1 u2 2

Du f (0, 0) = lim = lim 2
t→0 t t→0 u1 + t2 u2 4


0,
 if u1 = 0
= u2 2

 , if u1 6= 0
u1

The directional derivatives exist at (0, 0) along all the directions,

however f is not continuous at (0, 0).
9 / 100
Observation

Let φ : E ⊆ R → R and E contain a nbd of x0 .

φ(x0 +t)−φ(x0 )
• φ is differentiable at x0 ⇔ lim t exists.
t→0

• Equivalently, there is λ ∈ R s.t.

φ(x0 + t) − φ(x0 ) − λt
lim =0
t→0 t
• Equivalently, ∃ a linear transformation T : R → R s.t.

φ(x0 + t) − φ(x0 ) − T (t)

lim = 0.
t→0 t

10 / 100
Definition. Let f : D ⊆ Rn → Rm and X0 ∈ D. Suppose D
contains a nbd of X0 .

The function f is said to be differentiable at X0 if there exists a

linear transformation T : Rn → Rm s.t.
f (X0 + h) − f (X0 ) − T (h)
lim = 0.
h→0 khk

• Equivalently, f is differentiable at X0 , if there exists a matrix

of B order m × n s.t.
f (X0 + h) − f (X0 ) − Bh
lim = 0.
h→0 khk

• The matrix B is called the derivative of f at X0 , which is

denoted by Df (X0 ).

11 / 100
Theorem. Let f : D ⊆ Rn → Rm be differentiable at X0 ∈ D,
then f is continuous at X0 .

Proof. Since f is differentiable at X0 , So

f (X0 + h) − f (X0 ) − (Df (X0 ))h

lim = 0.
h→0 khk

Note that

lim [f (X0 + h) − f (X0 )]

h→0
  
f (X0 + h) − f (X0 ) − (Df (X0 ))h
= lim k h k +Df (X0 )h
h→0 khk
=0 + 0 = 0.

Hence, f is continuous at X0 .

12 / 100
Uniqueness of Derivative:

Let C be another m × n matrix s.t.

f (X + h) − f (X) − Ch (C − B)h
lim =0 ⇒ lim = 0.
h→0 khk h→0 khk

Let h = tej , where t ∈ R and ej = (0, . . . , 0, 1, 0, . . .).

In the case t → 0, it follows that h → 0, and

(C − B)ej t
lim = 0.
t→0 |t|

Hence, the vector (C − B)ej , the j-th column of C − B, is

identical to 0. That is B = C.

13 / 100
Example. Let f : R2 → R3 be defined by,
 
1 2 " #
 x
f (x, y) =  3 4  .

y
5 6

Is f differentiable at (0, 0) ∈ R2 ?

14 / 100
Solution. Let X0 = (0, 0) and consider h = (h1 , h2 ) ∈ R2 .
 
1 2 " #
 h1
f ((0, 0) + (h1 , h2 )) − f (0, 0) =  3 4 

h2
5 6
 
1 2
If B =  3 4  then
 

5 6

f (X0 + h) − f (X0 ) − Bh
lim = 0.
h→0 khk
 
1 2
Hence f is differentiable at (0, 0) with Df (0, 0) =  3 4 .
 

5 6

15 / 100
Problem. Let f : Rn → Rm be defined by

f (X) = BX + b, X ∈ Rm

where B is an m × n matrix and b ∈ Rm . Is f differentiable on Rn .

(For each X ∈ Rn , Df (X) = B).

16 / 100
Problem. State true or false.

• If f : D ⊆ Rn → Rm is differentiable at X0 ∈ D, then all

directional derivative of f at X0 exist. (True)
• Is the converse holds? (False)

Example
Let f : R2 → R be defined by

x2 y

x4 +y 2
, if (x, y) 6= (0, 0)
f (x, y) =
0, if (x, y) = (0, 0)

17 / 100
Directional Derivatives from Differentiability

Theorem
If f : D ⊆ Rn → Rm is differentiable at X0 ∈ D, then all
directional derivatives of f at X0 exist, and

Du f (X0 ) = Df (X0 ) · u,

for every unit vector u.

18 / 100
Proof (Part 1)

If f is differentiable at X0 ∈ D, then

f (X0 + h) − f (X0 ) − Df (X0 )h

lim = 0.
h→0 khk

Let u be a unit vector and set h = tu with 0 < |t| <  for some
 > 0 sufficiently small:

f (X0 + tu) − f (X0 ) − Df (X0 )tu

lim = 0.
t→0 |t|

This implies

f (X0 + tu) − f (X0 )

lim = Df (X0 )u.
t→0+ t

19 / 100
Proof (Part 2)

Similarly, for t → 0− :

f (X0 + tu) − f (X0 ) − Df (X0 )tu

lim = 0,
t→0− −t
so
f (X0 + tu) − f (X0 )
lim = Df (X0 )u.
t→0− t

Combining both limits:

f (X0 + tu) − f (X0 )

lim = Df (X0 )u.
t→0 t

Du f (X0 ) = Df (X0 ) · u.

20 / 100
Problem. Let f : R2 → R3 be defined by,
 
" #! 1 2 " #
x  x
f = 3 4 

y y
5 6

Find Du f (0, 0) for u = (1, 0).

Solution.
   
1 2 " # 1
 1
Du f (0, 0) = Df (0, 0) · u =  3 4  =  3 .
  
0
5 6 5

21 / 100
Definition
Let A ⊆ Rm → R. The j − th partial derivative of f at a is the
directional derivative of f at a along the direction ej .

Problem. Let f : R2 → R defined by

f (x, y) = xy 2 + y 3

Find D1 f (0, 0).

Solution. Here D1 f (x, y) = y 2 . [differentiating w.r.t x keeping y

as constant]. Then D1 f (0, 0) = 0.

22 / 100
Theorem
Let f : D ⊆ Rn → R. If f is differentiable at X0 ∈ D, then

Df (X0 ) = [D1 f (X0 ), D2 f (X0 ), . . . , Dn f (X0 )]1×n

Proof.
If Df (X0 ) exists, then

Df (X0 ) = [λ1 , λ2 , . . . , λn ]1×n

Since f is differential at X0 , for all j = 1, 2, . . . , n:

Dj f (X0 ) = Df (X0 ) · ej = λj .

23 / 100
Theorem
Let f : D ⊆ Rn → Rm and D contains a nbd atX0 ∈ D. Let
fj : D ⊆ Rn → R for j = 1, 2, . . . , m so that
 
f1 (x)
f2 (x)
 
 
f (X) =  .. , then
.
 
 
fm (x)

(i) f is differentiable at X0 iff fj is differentiableat X0 . 

Df1 (x)
 Df2 (x)
 

(ii) If f is differentiable at X0 , then Df (X0 ) =  .. 
.
 
 
Dfm (x) m×n

• The ij − th entry of Df (X0 ) is Di fj (X0 ). The matrix

Df (X0 ) is called the Jacobian matrix of f at X0 .
24 / 100
Proof.
Since f is differentiable at X0 , we have

f (X0 + h) − f (X0 ) − Bh
lim = 0.
h→0 khk

if and only if

fj (X0 + h) − fj (X0 ) − (j th row of B)h

lim = 0.
h→0 khk

25 / 100
Differentiability at the Origin

Problem. Check the differentiability of f at (0, 0):

1. f (x, y) = (x2 + y 2 , xy)

2.  
 xy sin 1 , x2 y 2 ,

if x 6= 0,
f (x, y) = x

0, if x = 0
3.
xy

 , if (x, y) 6= (0, 0),
x2 + y2

f (x, y) =

0, if (x, y) = (0, 0)

26 / 100