Arc length of a curve in Cartesian

coordinates

Approximating the length of a curve

The problem under investigation here is to find the length of a curve between two points (say P
and Q). The curve is given in terms of a function y = f(x).

Q y = f (x)
P
s

0
x

We will designate the required distance s. As with differential calculus, we will arrive at an exact
expression for s by using the theory of closer and closer approximations.

y Q
A curve between
s Q P and Q of length s
Q
P
y = f (x)
P approximated by

straight-line segments
between P and Q.
P
0
x
dx dx

We will approximate the length of the curve by adding up the straight-line segments that join
points between P and Q. As we increase the number of segments the approximation gets better
and better.

© blacksacademy.net

1
Suppose the length of one of these straight-line segments is designated ds and the corresponding
small changes in the x and y directions are designated dx and dy respectively.

Q y = f (x)
y + dy = f (x + dx)

P ds dy
y = f (x)
dx

x
0 x x + dx

Then, as the above diagram indicates

ds   dx   dy 

2 2 2

and

dx   dy 
2 2
ds 

Then the approximate length along the curve is the sum of the straight-line segments
s   ds

Hence, the exact curve is

s  lim  ds
 x 0

It is the limit of the sum of the small increments along the curve from P to Q as the length of the
increment gets smaller and smaller. Also the point P is given by, say, y  f a  and the point Q by

y  f b  .

Q y = f (x)
f (b)
f (a) s
P

0
x
a b

© blacksacademy.net

2
 b b
s  lim  dx   dy    dx   dy 
2 2 2 2

dx  0
a a

This formula is not in a very manageable form. However, we are allowed to manipulate the terms
dx and dy as if they were functions, which in fact they are – they are called differentials. This
enables us to write the arc length in the following more convenient form.
b

dx   dy 
2 2
s
a

 dy  
b 2

dx 
2
 1 
 dx  
2
a 
b 2 
 dy  
  1   dx
  dx  
a
 
This computes the arc length in terms of the variable x¸ but we could also write it in terms of y.
f b   2
 1   dx   dy
s 
f a 
 
 dy  
 
We will now proceed to illustrate the application of this formula.

The circumference of the circle

Suppose we wish to find the circumference of a circle radius a.

(x, y)

O
x

A circle of radius a has an equation x 2  y 2  a2 . Since there are two y values corresponding to a
given x value, this is not a function. This makes integrating it difficult. However, the length of the
circle can be treated as four times the length of one of the quadrants.

© blacksacademy.net

3
 y

O
x

On this segment it is possible to make a one-one function. Here we may write y  a 2  x 2 with
the convention that we take only the positive root. Then the arc length of one segment is given by
a  2
 dy   dy
 1 
 dx 
 
dx . This formula requires us to find
dx
. To do so
0
 
1


y  a2  x 2  2

1
dy 1 2 x
 

 a  x2 2  2x 
dx 2 a  x2
2

2 2
 dy  x
   2
 dx  a  x2

Then on substituting into the formula

a 2
 dy  
s   1   dx
  dx  
0
 
a
 x 2

  1  2 2 
 dx
0
  a  x  
a 
a2  x 2  x 2
   dx

0 a2  x 2 

a 
a2
  2  dx

0 a x
2 

a
 1 
 a    dx
2 
0 a x 
2

a
  x 
 a sin 1   
  a  0
 a sin 1 1  0 
 
 a   0
2 
a

2

© blacksacademy.net

4
This gives the arc length of one quarter of the circumference of the circle. Hence the full
circumference is

C  4
a  2 a
2
This is the familiar formula for the circumference of the circle, which we have now proven. You
may not have been aware that C  2 a is actually a theorem – you have been so familiar with it as
a result! Logically, the formula for area comes first

A  r 2
This is not a theorem but actually a definition – it is the definition of  (pi) as the ratio of the
area of the circle (A) to the area of the square of its radius
A
 
r2

That is  (pi) is the ratio of the area of a circle of radius r to the area of a square that has a side
of length r. Having defined  in this way, it is then that we prove the formula for the

a  2 
 1   dy   dx , for the arc-length of a curve in Cartesian
circumference by means the formula, 0   dx  
 
 
coordinates that we have just introduced.

A
So from the definition of  as   we prove that the circumference of a circle is C  2 r .
r2

This leads us into the question of the properties of  , which is a fascinating study. The question
that haunted the Greeks was whether the area of the circle could be constructed from the area of
the square using only a compass and ruler. In the C19th Galois showed that this was not possible,
which meant that  could not be a rational number. Later Lindenbaum showed that  is a
transcendental number, which means that it is not the root of any polynomial function with
rational coefficients

© blacksacademy.net

5
Further examples
Example (1)

Find the arc length of a parabola x 2  4y from the vertex to x  2

Solution

4y = x 2

s
x
2

x2
We have y  which is a parabola. Hence
4
dy x

dx 2
The arc length formula is
b 2 
 dy  
s   1   dx
  dx  
a
 
and on substitution we get
2 2
x 
s 
0
1    dx
2
2
x2
 
0
1
4
dx

2
4  x2
 
0 4
dx

2
 1  4  x 2 dx
2
0

The integral of this expression is a standard result found by a hyperbolic substitution

(with which you should be familiar at this stage) and is given by

1 1  x  
s x x  4  2sinh   
2

2  2 

© blacksacademy.net

6
 1 a2 x 
This is found by substitution into the formula  x 2  a 2dx 
2
x x 2  a2 
2
sinh 1   .
a 
The result also can be given in logarithmic form.
2
1  x  x 2  4 
s  x x 2  4  2ln  
2  2 
  0


1 1
2  2
  
 2 8  2 ln 2  8  ln 2 

 
22 2
 2  ln  
 2 


 2  ln 1  2 
 2.296 3 s.f.

Example (2)
(i) Prove that the length of the arc of the curve y  log2  x  between the points

3
1
where x  2 and x  3 is given by s   1  dx
2 x 2 ln 2  x 

1
(ii) Use the substitution sinh 2 t   to evaluate this integral to 3 s.f.
x 2 ln 2 2

Solution
ln  x 
i y  x   log2  x  
ln 2
dy 1

dx x ln 2
2
 dy  1
   2 2
 dx  x ln 2
b 2
 dy  
On substitution into the formula s    1    dx .
  dx  
a
 
3
1
s 
2
1
x 2 ln 2  x 
dx

Proven.

© blacksacademy.net

7
 1
 ii   sinh 2 t 
x ln 2 2
2

1
x
ln  2 sinh t 
cosh t 
dx   dt
ln 2 sinh 2  x 
 1 
t  sinh 1 
 x ln 2 
 

3
1
s 
2
1
x 2 ln 2  x 
dx

 1 
sinh 1 


 3ln 2 
  cosh t  
  1
 1 
1  sinh 2 t  
 ln 2 sinh 2  x   dt
sinh 


 2ln 2 
  
 1 
sinh 1 


 3ln 2 
  cosh 2 t  
     dt
 ln 2 sinh t  
1
 1  2
sinh  
 2ln 2 
 

 
 3 ln 2   1  sinh t  
1 2
sinh 1  
1
ln 2 sinh 1  2 ln1 2   sinh 2 t  
  
  dt
  
 1 
1
sinh  
1
 cosech t   1 dt
 3 ln 2 

ln 2 
 
 1 
 2
1
sinh  
 2 ln 2 
 

1
 1 
sinh  
1 d 
coth t   t   dt coth t   cosech t  
 3 ln 2 
   2

ln 2  sinh 1
 1 
   
 2 ln 2 
  
1
 1 
sinh  
1   3ln 2 
 1+cosech 2
t   t    coth 2 t   cosech 2 t   1
ln 2   sinh 1  1 

 2ln 2 
  
3
1  1  1   1 
 x ln 2 1  2 2  sinh 1 
 x ln 2    2 2  sinh 2 t   cosech 2 t   x 2 ln 2 2 
ln 2  x ln 2    x ln  
2 
2
3
1   1 
  1  x 2 ln 2 2  sinh 1   
ln 2   x ln 2  2
 1.16 3.s .f .

© blacksacademy.net

8
© blacksacademy.net