2.

DIFFERENTIAL CALCULUS OF FUNCTIONS OF ONE VARIABLE

Differential calculus, being a subfield of calculus, is concerned with the study of the rates
at which quantities change, primarily the derivatives of functions. Recall that if the function
y  f (x) has a finite derivative for all x  (a, b)  , then
dy dy
 f '( x) or y   f '( x)
dx dx
is called the gradient function of the curve at every point x  ( a, b) . If y is also a

differentiable function of x in the same interval, then its derivative

d d  dy  d 2 y
y  y     2
dx dx  dx  dx

is called the second (second order) derivative of y with respect to x. Similarly, the n th 
order derivative of y, if it exists , is given by

y (n) 
d
dx
 y ( n1)  ,

for any positive integer n. You have already learnt about derivatives of elementary real-
valued functions of single. In this chapter, we will focus mainly on the application of
derivatives.

2.1 Tangents and normals to general plane curves

Definition 2.1.1

1. The gradient of the curve y  f ( x) at a point x  x0 is given by

dy
y( x0 )  f ( x0 )  ,
dx x  x0

and the tangent line to this curve at x  x0 is the line through  x0 , f ( x0 )  with the

gradient f ( x0 ), i.e. y  f ( x0 )  y( x0 )( x  x0 ) is the tangent line to this curve at x  x0 .

1
 2. A line is normal to a curve at a point if it is perpendicular to the curve’s tangent line
there. The line is called the normal line to the curve at that point.

Recall that if M 1 is the gradient of the tangent line at a point and M 2 is the gradient of
the normal line at the same point, then

M 1. M 2  1.

Example 2.1.1

1. Find the tangent and normal lines to the curve x 2  xy  y 2  7 at the point (1, 2).

2. Find the equations of the tangents with slope  92 to the ellipse 4 x 2  9 y 2  40 .

Solution

1. Using implicit differentiation, we get

dy dx dy
2x  xy  2y 0
dx dx dx
dy y  2 x
 
dx 2 y  x
dy 22 4
  
dx ( x , y ) ( 1,2) 2(2)  1 5

4 4 x  14
 equation of tangent line at (1, 2) is y  2   x  1 , i.e. y  .
5 5

The gradient of the normal is M 2  1   5

4 4
5

5 3  5x
 equation of the normal at (1, 2) is y  2   ( x  1), i.e. y  .
4 4

2. Let ( x0 , y 0 ) be the point of tangency of the required tangent. Then the

point ( x0 , y 0 ) satisfies the equation

4 x0  9 y 0  40 .
2 2
(*)

2
 Also,
dy 4x 4 x0
  .
dx ( x0 , y0 ) 9y ( x0 , y0 )
9 y0

4 x0 2
Thus,    y0  2 x0 . Substituting this in (*) yields
9 y0 9

4x02  36x02  40  x0  1 .

When x0  1, y 0  2 and when x0  1, y 0  2 . Thus, the points of tangency are

(1,2) and (1,2) . Therefore, the required equations of the tangents are
2 2
y  2   ( x  1) and y  2   ( x  1) .
9 9

2.2 Mean Value Theorem

In this section, we state one of the most important and applicable theorems of
mathematics, the Mean-Value theorem. The theorem can be used to prove some
important facts about differentiation. The theorem can also be used to prove
results that are useful in graphing a wide variety of functions. It has many other
applications as well. Before stating the theorem we need one preliminary result.

Theorem 2.2.1 (Rolle’s theorem)

Let f be continuous on the closed interval [ a , b ] and differentiable on the open
interval ( a, b) . If f (a)  f (b)  0 , then there exists at least one number c  (a, b) at which
f ' (c )  0 .

The situation described in Rolle’s theorem is depicted in the diagram below:

y
f ' (c )  0

f (a)  0 f (b)  0

a 0 c b

3
Example 2.2.1
Verify the Rolle’s Theorem for the polynomial function

x3
f ( x)   3x
3

in the interval [ 3, 3].

Solution

Clearly, f is continuous on and differential hence continuous on every point of [3,3]

and differentiable at every interior point of (3,3). Also

(3)3 33
f (3)   3(3)  0  f (3)   3(3).
3 3

f ( x)  x2  3. f (c)  0  c   3.

Therefore, f ( x) is zero at two points c1   3 and c2  3 and that c1 , c2  (3,3).

Hence verified!

We now state the Mean-Value theorem.

Theorem 2.2.2 (Mean Value Theorem)

Suppose that y  f ( x) is continuous on a closed interval [ a , b] and differentiable on the

interval’s interior (a, b). Then, there is at least one point c in (a, b) at which
f (b)  f (a )
 f (c) (2.1)
ba

The situation described in Mean Value theorem can be illustrated diagrammatically as

follows:

4
 y

f (c)  slope

0 a c b x

Theorem 2.2.3 (Cauchy’s Generalised Mean Value Theorem)

Suppose that f ( x ) and g ( x) are continuous functions in [ a , b] and differentiable in

(a, b). Then, there is at least one point c in (a, b) at which

f (c) f (b)  f (a)

 , (2.2)
g (c) g (b)  g (a)

provided g (b)  g (a) and f ( x), g ( x) are not simultaneously zero.

NOTE: (2.1) is a special case of (2.2) with g ( x)  x.

Example 2.2.2

Given that f ( x)  2 x 2  7 x  10, satisfies the Mean Value Theorem in the interval [2,5],
find c.

Solution

f (5)  f (2) 25  4
  7,  f ( x)  4 x  7
52 3
f (5)  f (2) 7
 f (c)   4c  7  7  c   (2,5).
52 2

5
The Mean-Value theorem may be put in several useful forms. For example, multiplying
through by (b  a ) , yields
f (b)  f (a)  (b  a) f ' (c), c  ( a, b) .
A simple replacement of b by x yields
f ( x)  f (a)  ( x  a) f ' (c), c  ( a, x ) . (2.3)

Example 2.2.3

Use the Mean-Value theorem to approximate 6 65 .

Solution:

Let f ( x)  6 x , a  64, b  65 .
5

Now, f ' ( x)  x . Thus applying (2.3), we have

1
6
6

f (65)  f (64)  (65  64) f ' (c), c  (64,65) .

Since c is not known, take c  64 . Therefore,

 
65  f (65)  64  (65  64)    2  1  2.00521
6 6 1
5  .
 
 6 6 64  

192

Next, we use the Mean-Value theorem to prove that a function with a zero derivative is
constant.

Theorem 2.2.4
If f is differentiable on ( ,  ) and if f ' ( x)  0 for every x, then
f ( x)  k , a constant function on ( ,  ) .

Proof:
Take any two numbers a and b in ( ,  ) , with a  b . Then by Mean-value theorem there
is a number c  (a, b) such that
f (b)  f (a )
 f ' (c)  0 , by hypothesis.
ba
Since b  a , this can only occur if f (b)  f (a)  0 ,  f (b)  f (a) .

6
Since this result is true for any numbers a and b, f is a constant function in ( ,  ) . Hence
proved!

Corollary 2.2.1
If f ( x)  g ( x) at each point of an interval ( ,  ), then there exists a constant k such
that
f ( x)  g ( x)  k , for all x  ( ,  ).

Proof:
Since f ( x)  g ( x) , then f ( x)  g ( x)  0. Applying Theorem 2.2.4, we have that
f ( x)  g ( x)  k . Hence proved!

Example 2.2.4
Find the function f ( x ) whose derivative is sin x and whose graph passes through
(0, 2).
Solution
Suppose there is a function g ( x) such that f ( x)  g ( x). Then, g ( x)  sin x implying
g ( x)   cos x so that by Corollary 2.2.1
f ( x)   cos x  k .

Since f ( x ) passes through (0, 2), we have that

2   cos 0  k  k  3
 f ( x)   cos x  3.

Theorem 2.2.5 (Extended Mean-Value theorem)

If f (x ) and its n  1 derivatives are continuous on the closed interval [ a , b ] , and if

f ( n ) ( x) exists everywhere on the interval except possibly at the endpoints, then there is at

least one point c  (a, b) such that

f '(a) f ''(a) f ( n ) (c )
f (b)  f (a)  (b  a)  (b  a) 2   (b  a) n . (2.4)
1! 2! n!
When b is replaced with the variable x, (2.4) becomes

7
 f '(a) f ''(a) f ( n ) (c )
f ( x)  f (a )  ( x  a)  ( x  a) 2   ( x  a) n (2.5)
1! 2! n!
for some c between a and x.

When a is replaced with 0, (2.5) becomes

f ' (0) f ' ' (0) 2 f ( n1) (0) n1 f ( n ) (c) n
f ( x)  f (0)  x x  x  x (2.6)
1! 2! (n  1)! n!
for some c between 0 and x.

Theorem 2.2.5 leads us to what is known as the Taylor’s theorem.

Theorem 2.5.6 (Taylor’s theorem)

( n 1)
Suppose a function f has n continuous derivatives, that f exists on the interval

 a, b and x0   a, b . For every x0   a, b , there exists a number  between x0 and x

with

f ( x)  Pn ( x)  Rn ( x) (2.7)

where
f '( x0 ) f ''( x0 ) f ( n ) ( x0 )
Pn ( x)  f ( x0 )  ( x  x0 )  ( x  x0 ) 2   ( x  x0 ) n
1! 2! n!
n
f (i ) ( x0 )
 ( x  x0 )i
i 0 i!
and
f ( n1) ( )
Rn ( x)  ( x  x0 )n1.
(n  1)!

Here Pn ( x ) is called the n th Taylor polynomial for f about x0 and Rn ( x) is called

the remainder term associated with Pn ( x). When x0  0, the Taylor polynomial is often
called the Maclaurin polynomial.

8
NOTE: 1. The value of  cannot be determined explicitly, we just know that it lies

between x0 and x. This means that f ( n1) ( ) cannot be determined explicitly;

we can just state its bounds when x is in some specified interval.
2. When n  0, we get the Mean Value theorem.

Example 2.2.5
1. Find the nth Maclaurin polynomial for
(a) f ( x)  e x (b) f ( x)  cos x

2. Find the Taylor polynomial for f ( x)  ln x at x0  1.

Solutions:
1. For the Maclaurin polynomial at x0  0 ,

(a) f ( x)  e x  f ( x)  e x , f ( x)  e x ,..., f ( n ) ( x)  e x , f ( n1) ( x)  e x

 f (0)  e0  1, f (0)  e0  1,..., f ( n ) (0)  e0  1

f '(0) f ''(0) 2 f ( n ) (0) n
Pn ( x)  f (0)  x x   x
1! 2! n!
x2 xn
 1  x   ... 
2 n!
(b) f ( x)  cos x  f ( x)   sin x, f ( x)   cos x,...

sin x, n  3, 7,11,15,...
 sin x, n  1,5,9,13,... 0, n is odd
 
f ( n ) ( x)    f (0)  1 and f ( n ) (0)  1, n  4,8,12,16,...
cos x, n  4,8,12,16,... 1, n  2, 6,10,14,...
 
 cos x, n  2, 6,10,14,...
f '(0) f ''(0) 2 f ( n ) (0) n
Pn ( x)  f (0)  x x   x
1! 2! n!
x2 x4 x2n
 1   ...  (1) n
2 4! (2n)!
2. f ( x)  ln x  f (1)  0

1
f ( x)   f (1)  1
x

9
 1
f ( x)    f (0)  1
x2
2
f (3) ( x)  3  f (3) (1)  2
x
2.3 3!
f (4) ( x)   4   4  f (4) (1)  6
x x

(1) n 1 (n  1)!
f (n)
( x)  n
 f ( n ) (1)  (1) n 1 (n  1)!
x
f '(1) f ''(1) f ( n ) (0)
Pn ( x)  f (1)  ( x  1)  ( x  1) 2   ( x  1) n
1! 2! n!
( x  1) 2( x  1)
2 3
(n  1)!( x  1) n
 0  ( x  1)    ...  (1) n 1
2 3! n!
( x  1) ( x  1)
2 3
( x  1) n
 ( x  1)    ...  (1) n 1
2 3 n

2.3 Differentials
Definition 2.3.1
Let y  f ( x) be a differentiable function, where x is the independent variable. The

quantity dx is called the differential of x . Then dy , the differential of y is defined by

dy  f ( x)dx,

which can also be written as

df  f ( x)dx.

y  f ( x) Q  x

P ( x, y ) y

x  dx

0 x0

10
Differentials are frequently useful in approximations. If x  dx is an increment of x ,
then the true change of f as x changes from x0 to x0  dx is

f  f ( x0  dx)  f ( x0 )

and the differential estimate is

df  f ( x0 )dx.

When x is small, the approximation is usually quite close so that the approximate error
f  df   x

is quite small and f  df .

Example 2.3.1
1. Find dy if (a) y  sin 3 x (b) y  2 x ln x.
1
2. The function f ( x ) 
x

changes value when x changes from x0 to x0  dx, where x0  0.5 and dx  0.1.
Find
(a) the true change of f
(b) the differential estimate of f
(c) the approximation error.

3
3. Use differentials to approximate 26.5.

Solution
 x1 x 
1. (a) y  sin 3x  dy  (3cos 3x)dx (b) y  2 ln x  dy   2 .  2 ln 2 ln x  dx.
x

 x 
1
2. (a) f ( x)   f  f ( x0  dx)  f ( x0 )
x

 f (0.6)  f (0.5)

11
 5
 2
3
1
 .
3

1
(b) f ( x)  
x2
1 2
 df  f ( x0 )x   2
.(0.1)  (4)(0.1)   .
(0.5) 5
1  2 1
(c) Approximation error  f  df        .
3  5  15

Let f ( x)  x . We observe that 26.5 is close to 27, so we take x0  27.

3
3.
Then,

x0  dx  26.5  dx  0.5.

1 2  1 1 1
dy  f ( x0 )dx   x0  3  dx  2 .( 0.5)   , i.e. f  df   .
3  3.(27) 3 54 54

 f  f ( x0  dx)  f ( x0 )  
1
 f (26.5)  f (27)
54

1
 f (26.5)  3 26.5  3   2.9815.
54

2.4 Limits

Definition 2.4.1
Let f :  be a given function and c and L be real numbers. Then f (x ) is said to tend
to L as x approaches c if for any given positive number  there is a positive number 

such that f ( x)  L   whenever 0  x  c   .

We write f ( x)  L as x  c or lim f ( x)  L
x c

12
NOTE: 1. The definition implies that there can be at most one limit L.

2. The inequality x  c   implies that c    x  c   . The part of the inequality

which states that 0  x  c merely means that x is not allowed to be equal to c

itself.

3. The inequality f ( x)  L    L    f ( x)  L  

Example 2.4.1
Use the definition of the limit to show that
2
(a) lim(5x  2)  13 (b) lim  x 2  x   0 (c) lim  x xx3 2   
2

x 3 x 1 x 0 3
Solution
In each case, we wish to show that, given any   0 , we can find a   0 such that

| f ( x)  L |  whenever 0  x  c   .

(a) Here c  3, L  13 and 0  x  c    0  x  3   .

Let   0 be given. Then,

5x  2  13  5( x  3)  5 x  3 .
1
Choose    . Then,
5

| f ( x)  L | 5 x  3  5  5  15     whenever 0  x  3  
as required.

(b) Let   0 be given. Then, whenever 0  x  1  

| f ( x)  L | x2  x  0 | x( x  1) || x || x  1| .

We can take x  1  1 so that 1  x 1  1  0  x  2 . This means that

| f ( x)  L || x || x  1| 2 | x  1| .
1
Choose    . Then,
2

| f ( x)  L | 2 x  1  2  2  12     whenever 0  x  1   .

13
Hence shown!

(c) Let   0 be given. Then, whenever 0  x  0    0 | x | 

x2  x  2  2  x2  x  2 2 3x 2  3x  6  2 x  6 3 x 2  5 x | x || 3 x  5 |
f ( x)  L        
x3  3 x3 3 3( x  3) 3( x  3) 3| x 3|

Using | x | 1, we have that

1  x  1  3  3x  3  2  3x  5  8
and
1 1 1
1  x  1  2  x  3  4    .
4 x3 2
Thus,
| x || 3x  5 | 8 4
f ( x)  L   | x | | x | .
3 | x  3 | 3.2 3
3
Now, choose    so that
4

4 4 43 
f ( x)  L  | x |        whenever 0  x   .
3 3 34 

Hence shown!

2.4.1 Properties of limits

1. Uniqueness of limits: Suppose that f ( x)  L1 as x  c and f ( x)  L2 as

x  c . Then L1  L2 .
2. Limit of a constant: If k is a constant and f ( x)  k for all values of x then
for any number c lim f ( x)  k .
x c

3. If c is real number and f ( x)  x for all x, then lim f ( x)  c .

x c

4. Limits of equal functions: Suppose that there is a number h  0 such that

f ( x )  g ( x) for all x for which x  c  h . Suppose also that lim f ( x)  L ,

x c

then lim g ( x)  L
x c

5. Limit of a sum of functions: If f and g are two functions, with lim f ( x)  L1

x c

14
 and lim g ( x)  L2 , then lim[ f ( x)  g ( x)]  L1  L2 .
x c x c

6. Limit of a product of functions: If f and g are two functions, with

lim f ( x)  L1 and lim g ( x)  L2 , then lim[ f ( x)  g ( x)]  L1  L2 .
x c x c x c

7. Limit of a Quotient of functions: If f and g are two functions, with

lim f ( x)  L1 and lim g ( x)  L2 , then lim f ( x) / g ( x)  L1 / L2 .
x c x c x c

8. Limit of a composite of a function: Suppose that f and g are functions, b and

c are numbers, f (b) is defined, and lim f ( x)  f (b) , lim g ( x)  b . Then
x b x c

lim f [ g ( x)]  f (b) .

x c

9. If n is a positive integer and c  0 , then lim n x  n c .

x c

10. If n is positive integer, L  0 and lim f ( x)  L , then lim n f ( x)  n L .

x c x c

You used most of these properties in MAT 1100 to evaluate limits.

2.3.2 Limits at infinity and infinity limits

1
If f is the function defined by f ( x)  , then
x
1 1
(i) lim f ( x)  0 (ii) lim f ( x)  0 (iii) lim   and (iv) lim   .
x  x  x 0 
x x 0 x

1
y  f ( x) 
x

0 x

15
Note that the notation x   is used if x increases without bound through positive values
and x   is used if x decreases without bound. Note also that as x approaches 0 from
the right (or from the left), f (x ) increases (or reduces) without bound.

We now consider some important rules of operations for limits:

R1 . If lim f ( x)   (or  ) and lim g ( x)  b , then lim[ f ( x)  g ( x)]   (or  ).

x c x c x c

On the other hand, if lim g ( x)   , nothing in general can be said about

x c

lim[ f ( x)  g ( x)] .
x c

Further investigation of the particular limit will be necessary.

R2 . If lim f ( x)   (or  ) and lim g ( x)  b , then

x c x c

(i) lim[ f ( x)  g ( x)]   (or  ) if b  0 .

x c

(ii) lim[ f ( x)  g ( x)]   (or  ) if b  0 .

x c

If b  0 , further investigation is needed.

R3 . If lim f ( x)   (or  ) and lim g ( x)  b , then lim g ( x) / f ( x)  0 . If

x c x c x c

lim g ( x)   (or  ) , other methods for evaluating the limit may be necessary.
x c

2.4.3 Indeterminate Forms

Definition 2.4.2
If lim f ( x)  A and lim g ( x)  B, where A and B are either both zero or infinite, then
x  x0 x  x0

 f ( x) 
x  x0 g ( x ) 
lim 
 

16
 0 
is called an indeterminate of the form or , respectively.
0 

L’Hospital’s rule uses derivatives to evaluate such limits.

Theorem 2.4.1 (L’Hospital’s rule)

1. If f ( x ) and g ( x) are differentiable in the interval ( a, b) except possibly at a

point x0 in this interval, and if g ( x)  0, for x  x0 , then

 f ( x)   f ( x) 
lim    lim  , (2.8)

x  x0 g ( x )
 x x0  g ( x) 

whenever the limit on the right can be found. In the case where f ( x) and g ( x)
satisfy the same conditions as f ( x ) and g ( x) given above, the process can be
repeated.

2. If lim f ( x)   and lim g ( x)  , the result (2.8) is also valid.

x  x0 x  x0

The forms (1) and (2) can be extended to cases where x   or  , and to cases where

x0  a or x0  b in which only one sided limit, such as x  a  or x  b , are involved.

Other so-called indeterminate forms are 0., 0 , 00 , 1 and   , and can be evaluated
on replacing them by equivalent limits for which the above rules are applicable.


Example 2.4.2  Indeterminate form
0
and 
0 

Evaluate the following limits:

 1  cos  x  ex cot x
(a) lim  2  (b) lim (c) lim .
x 1 x  2 x  1
  x  x x 0 cot 2 x

17
Solutions:

0
a) The limit has indeterminate form , so applying L’Hospital’s rule, we get
0

 1  cos  x    sin  x    2 cos  x   2

lim  2 
 x 1 
lim   lim   .
x 1 x  2 x  1
   2 x  2  x 1  2  2

ex 
b) Note that e   as x   . Thus lim
x
is of the form .
x  x 

ex ex
 lim  lim   .
x  x x  1

cot x 
c) Note that lim is of the form . Thus,
x 0 cot 2 x 

cot x csc2 x csc2 x cot x

lim  lim  lim
x 0 cot 2 x x 0 2 csc2 2 x x 0 4 csc2 2 x cot 2 x
.
Note that each application of the rule results in the form of type /  . Instead, we
try trigonometric substitution.
cot x tan 2 x 2 sec2 2 x 2
lim  lim  lim  2
x 0 cot 2 x x 0 tan x x 0 sec2 x 1

Example 2.4.3  Indeterminate form 0. and    

Evaluate the following limits:

(a) lim ( x 2 ln x) (b) lim(csc x  1x ) .
x 0 x 0

Solutions:
0 
These may be evaluated by first transforming them in form of type or .
0 
(a) The limit has the indeterminate form 0., so rewriting it and applying
L’Hospital’s rule, we get

18
  ln x   1x   x2 
lim( x ln x)  lim 
2
  lim  2   lim     0.
x 0 x 0
 x 2  x 0   x 3  x 0  2 
1

(b) lim(csc x  1x ) is of the type    .

x 0

 x  sin x 
But lim(csc x  1x )  lim  is of type 0 .
x 0 x 0
 x sin x  0

 x  sin x  1  cos x
 lim(csc x  1x )  lim   lim
x 0 x 0
 x sin x  x  0 sin x  x cos x

 sin x  0
 lim    0.
x 0 2 cos x  x sin x
  2

Example 2.4.4 (Indeterminate types of the form 0 0 ,  0 and 1 )

Evaluate the following limits:
1
(a) lim x kx , k  0 (b) lim x x (c) lim1  1x x .
x 0 x  x 

Solutions:

In these forms we exploit the fact that if lim y is one of these types, then lim (ln y ) is of

type 0. which has been dealt with already.

(a) lim x kx , k  0 is of the form 0 0 . Thus, set y  lim x kx . Then,
x 0 x 0

x 0

ln y  ln lim x kx  limln x kx   limkx ln x 
x 0 x 0

ln x 1/ x
 lim  lim  lim(kx)  0
x 0 1 / kx x 0  1 / kx 2 x 0

But ln y  0  y  1 .

 lim x kx  1
x 0

1 1
(b) lim x x is of the form  0 . Thus, we set y  lim x x . Then,
x  x 

 1
 1
ln x 1/ x
ln y  ln lim x   lim ln x x  lim 1x ln x  lim
x
 lim  0 .  ln y  0 .
x  x  x  x  x x  1
 

19
 1
 lim x  1 x
x 

(c) lim1  1x x is of the form 1 . We set y  lim1  1 x

x
 . Thus,
x  x


ln y  ln lim1 
x 
x

1 x
  lim ln1 
x 

1 x
x  lim x ln1 
x 
1
x


 lim
ln1  1x 
 lim
 1 / x 2 /1  1x 
x  1/ x x   1/ x 2
1
 lim 1
x  (1  1 )
x

 ln y  1  y  e .

 lim1  
1 x
x  e.
x 

2.5 Curvature

We now consider a way of measuring how fast or sharp a curve can turn or bend and
differentiation is very handy in doing this. This measure is called curvature. For example,
the curve below bends faster at P than at Q and so we can say that curvature at P is
greater than at Q .
y Q

0 x

We first define the parameter called arc length parameter which will be used to find
curvature.

20
2.5.1 Length of the arc

Let f a, b  be a differentiable function such that f  is continuous. Such a Function

is said to be smooth and its graph is said to be a smooth curve. The arc of a curve is said to
be simple if it does not intersect itself.
Suppose we wish to find the length of a simple arc of a smooth curve.

yk
( x0 , y0 ) y k sk ( xn , yn )
yk 1 xk

0 a xk 1 xk b x

If we partition the interval  a, b then the length of each line segment sk can be found
by Pythagorean Theorem
sk  (xk )2  (yk ) 2 .

Thus, the length of the arc is obtained by summing up all the sk ' s and finding the limit as
n  . Therefore,
n   b
2
  
2
n
 y   dy 
s  lim  (xk )  (yk )  lim  1  
2 2 k
 .xk   a 1    dx.
k 1  xk   dx 
n  n 
k 1
  

Definition 2.5.1

Let the function given by y  f ( x) represent a smooth curve on the interval [a, b]. The
arc length of f between a and b is

  dx   1   f ( x)  dx.
b dy 2 b
s 1
2
dx (2.9)
a a

For a  t  b,

  dx
t dy 2
s(t )   1  dx (2.10)
a

is called the arc length parameter.

If the curve is given parametrically by x  n(t ), y  m(t ),   t   , then

21
  dxdt    dydt 
 2
s
2
dt (2.11)


and

 dxdt    dydt  dt
 2
s(  )  
2
(2.12)


for some variable  such that      .

NOTE: (a) If the function is given by x  g ( y ) in the interval  c, d  , then (2.9) and (2.10)

can be expressed as

  dy   1   g ( y )  dy and s(t )   1    dy
d 2 d t 2
s 1 dx 2 dx
dy dy respectively.
c c c

(b) (2.10) and (2.12) imply that

ds
  ds
 dxdt   
2 2 2
 1 dy
dx
and   dy
dt
respectively.
dx dt

Example 2.5.1

x3 1
Find the arc length of the graph of y  
6 2x
on the interval  12 , 2.

Solution:

2
 dy   x 1  x4 1 1
2
x3 1 dy x 2 1 2
y              4 .
6 2x dx 2 2 x 2  dx   2 2 x 2  4 2 4x

Thus,

2
x4 1 1  x2 1 
1  
dy 2
dx
x4 1 1
 1   4    4    2 
4 2 4x 4 2 4x  2 2x 

22
 2
 x2 1  2 x 1 
  dx  
2
b dy 2 2
s 1 dx   2  dx  1   2 dx
 2 2x   2 2x 
a 1
2 2

2
 x3 1  33
   
 6 2 x  12 16

Definition 2.5.2

Let s denote the arc length on a curve, and suppose that a definite direction on the curve
has been selected for the direction of increasing s. Let  be the angle that the tangent line
to the curve makes with the x  axis. Then the curvature of the curve, denoted by , is
defined by
d
 .
ds

Note that   0 .

Theorem 2.5.1

1) If the curve is given by y  f ( x), then curvature at the given point ( x, y ) is

| y |
 3 . (2.13)
1  ( y)2 
2

2) If the curve is given parametrically in terms of the parameter t , the curvature is

| xy  yx |
 3 .
( x)  ( y) 
2 2 2

Proof:

1) y y  f ( x)

P ( x, y )

A
 x

23
 From some fixed point A of the curve, the length of the arc AP is s and the angle
made by the tangent to the curve at P with the positive direction of the x  axis is
 . Then, the gradient of the curve is dy
and the gradient of the tangent is tan .
dx
Thus,

tan 
dy
dx
   arctan dydx 
d  dx  dx 2
d dy d2y
y
  dx  
dx 1   dy  1   dy  1  ( y) 2
2 2
dx dx

By definition,
d y
d 1  ( y)2 y
  dx   3 .
ds ds 1  ( y) 2 1  ( y)2  2
dx

The proof follows by the fact that   0.

d
d d
2) Since   , we have that    dt .
ds ds ds
dt
d
We now find an expression for
dt
. Using   arctan   , we get
dy
dx

 dy dx 
  arctan  dy
dx   arctan  
 dt dt 
 dx d 2 y dy d 2 x   dx 
2

 dx d 2 y dy d 2 x
 . .    .  .
d  dt dt dt dt 2   dt 
2
   dt dt 2 dt dt 2
2 2 2
dt  dy dx   dx   dy 
1      
 dt dt   dt   dt 

dx d 2 y dy d 2 x
.  .
dt dt 2 dt dt 2
2 2
d  dx   dy  dx d 2 y dy d 2 x
    .  .
     
dt dt dt dt 2 dt dt 2 .
   dt 3
ds 2 2
 dx  2  dy  2 
 dx   dy 
2

dt           
 dt   dt   dt   dt  

Hence proved!

24
NOTE: If the curve is given by x  g ( y ), then (2.13) becomes
| x |
 3 .
1  ( x) 2 
2

Let y  f ( x) be a curve with curvature  at point P. The circle passing through point
1
P with radius   is called the circle of curvature if the circle lies on the concave

side of the curve and shares a common tangent line with the curve at P.
y

P

0 x
Centre of curvature

 is called the radius of curvature at P , and the centre of the circle is called the centre
of curvature.
Example 2.5.2
1. Find the curvature of the curve given by
x  a(t  sin t ), y  a(1  cos t ), a  0, 0  t  2 .

x2
2. Given the curve y  x  , sketch the circle of curvature at (2,1).
4
Solutions:
1. x  a(t  sin t )  x  a(1  cos t )  x  a sin t
 ( x)2  a 2 (1  cos t )2  a 2  2a 2 cos t  a 2 cos2 t
y  a(1  cos t )  y  a (sin t )  y  a cos t
 ( y) 2  a 2 sin 2 t
| xy  yx | | (a  a cos t )(a cos t )  (a sin t )(a sin t ) | | a 2 cos t  a 2 |
  
a  2a 2 cos t  a 2 cos 2 t  a 2 sin 2 t   2a  2a 2 cos t 
3 3 3
( x)2  ( y)2  2 2
2 2 2

25
 a 2 |1  cos t | 1
 1  1 .
1  cos t 1  cos t  2 .a 1  cos t 
3 3 2 3 2
2
2 .(a ) 2 2 2

Using the identity 2sin 2  2t   1  cos t , we have that

1 1 1
   .
4a sin  2t 
1 1

2 2.a 1  cos t  2 2.a  2sin 2  2t  

3 2 2
3

2. We first find the curvature.

2
x2 x 1  x x2
y  x  y  1   y   and ( y)2  1    1  x 
4 2 2  2 4
1 1

| y | 2 2
  3  3  3 .
1  ( y) 
2
 2
  2

2 2 2
x x
     
4   4 
1 1 x 2 x
 
Thus, at (2,1),
1
2 1
 x 2  3 
 22  2
2

  
4 
2 2


1 1
Therefore, radius of curvature    2
 12

y
P(2,1)
x2
 2 y  x
4
0 x

(2, 1)

Since   2, the centre of curvature must be at (2, 1) .

26
For a given point P( x1 , y1 ) on the curve y  f ( x) , we can derive formulas for finding the
centre of curvature,  xc , yc  , so that the equation of the circle of curvature is
( x  xc )2  ( y  yc )2   2 ,

where  is the radius of curvature at P. We assume that the arc of the curve lies in the
first quadrant.

y y  f ( x)
C  xc , yc 

P( x1 , y1 )

A
 x

1
The gradient of PC is  dy
implying that
dx

1
yc  y1   dy  xc  x1  (i)
dx

Since P lies on the circle of curvature, we have that

( x1  xc )2  ( y1  yc )2   2 (ii)
Solving for xc and yc from equation (i) and (ii), we get

y 1  ( y) 2 
xc  x1 
y
1  ( y) 2
yc  y1 
y

Example 2.5.3
x2
1. Find the centre of the circle of curvature of the curve given by y  x  at (2,1).
4
2. Determine the centre of curvature at the point (0.5, 1) of the curve y 2  2 x.

27
Solutions:

2
x2 x 1  x x2
1. y  x   y  1   y   and ( y)2  1    1  x 
4 2 2  2 4

 x  x2   x  x2 
y 1  ( y) 2  1   1  1  x   1    2  x  
 2 4  2 4 0
xc  x1   2  2  2 2
y 
1

1

1
2 2 2
2
x
 2 x
1 ( y ) 4  1  2  2  4  1  2  1
2 22
yc  y1   1
y 
1

1
2 2
 centre of circle of curvature is (2, 1).

 dy dy d2y
2
dy dy 1 d2y 1  dy  1
2. y  2 x  2 y
2
2   2  .  y. 2   0  2   .     3
dx dx y  dx dx dx  dx y  dx  y

1   1 2  1   11 2 
y 1  ( y) 2 
1 1
 
 0.5   1   0.5     0.5  2  2.5
y y 1
xc  x1 
y  y3  ( 1)3
1
1

1  y 
2
1  ( y) 2
1
2
yc  y1   1   1   1
y  y31
1
 centre of circle of curvature is (2.5,1).

y2  2x

(2.5,1)

0 x

28
2.6 Intrinsic Coordinates

dy
From the definition of the curvature of a curve at a point, we have seen that  tan
dx
and using the arc length parameter, we have that
2 2
 ds   dy  ds dx 1 we have
   1     1  tan   sec  . Thus, dx  sec and since
2 2
 ,
 dx   dx  ds ds
dx
dx dy
that  cos . Similarly, we can show that  sin . The coordinates ( s, ) are called
ds ds
intrinsic coordinates. The differential relations

dy dx dy
 tan ,  cos and  sin
dx ds ds

make it possible to move from an equation of a curve in Cartesian form y  f ( x) to the

equivalent equation of the curve s  g ( ).

Example 2.6.1
1. A curve has Cartesian equation
y  cosh x, x  .

Find an intrinsic equation of the curve in the form s  g ( ), where s is measured

from the point (0,1) and  is the angle the tangent to the curve makes with the
positive x  axis.
2. A curve has Cartesian equation
2 3
y ( x  1) 2 , x  , x  1.
3

Find an intrinsic equation of the curve in the form s  g ( ), where s is measured

from the point (1, 0) and  is the angle the tangent to the curve makes with the
positive x  axis.

29
Solutions:

2
dy  dy 
1. y  cosh x,   sinh x     sinh 2 x.
dx  dx 
dy
Since  tan , we have that sinh x  tan .
dx

 
x dy 2 x
s 1 d d   1  sinh 2  d .
0 0

Using the identity cosh 2 x  sinh 2 x  1, we get

x x
 1  sinh 2  d   cosh  d  sinh   sinh x.
x
0 0 0

Expressing s in terms of  , we observe that

s  sinh x  tan ,

i.e. s  tan .

2
2 dy 2 3  dy 
 .  x  1 2   x  1 2
3 1 1
2. y  ( x  1) 2      x  1.
3 dx 3 2  dx 
dy
 tan , we have that  x  1 2  tan  x  tan 2   1  sec2  . We
1
Since
dx
now find s.

 
x dy 2 x x
s 1 d d   1    1 d    d
1 1 1

2 3x 2 3
  2  ( x 2  1).
3 1 3

We can now express s in terms of  by substituting the value of x,

2
i.e. s  (sec3   1).
3

THE END!

30