Department of Mathematics

Indian Institute of Technology Delhi

MTL 100: Calculus
Lecture Notes
August 26, 2016
You
may
want
to
read about
the life &
works
of
Newton.

Limits & Continuity

Isaac Newton
(1643 1727)

Limits

Definition 1. Let f : D R, where D R, and let x0 be such that D contains an interval of the
form (x0 , x0 + ), except possibly the point x0 . We say that f (x) tends to ` R as x tends to
x0 , if to every  > 0, there corresponds a > 0 (depending on both the function f and the point
x0 ) such that




f (x) ` <  whenever 0 < x x0  < .
(1)
Theorem 1. (Uniqueness of limit)
The limit of a function, if it exists, is unique.
Proof. Suppose f (x) tends to both `1 and `2 as x tends to x0 . Let  > 0. By (1) there exist
positive real numbers 1 and 2 such that




f (x) `1  <  whenever 0 < x x0  < 1 ,
2





f (x) `2  <  whenever 0 < x x0  < 2 .
2

For = min(1 , 2 ), we therefore have both





f (x) `1  <  and f (x) `2  < 
2
2

whenever 0 < |x x0 | < . But then



|`1 `2 | =  f (x) `1 f (x) `2  f (x) `1  + f (x) `2  <

Since |`1 `2 | <  holds for every  > 0, `1 = `2 .


2


2

= .


Remark 1. Theorem 1 justifies the use of a notation for the definition of the limit. We write
limxx0 f (x) = ` to denote f (x) tends to ` R as x tends to x0 .
Example 1. Consider the function
(

f (x) =

1 if x > 0;
1 if x < 0.

We show that limx0 f (x) does not exist. Suppose, on the contrary, limx0 f (x) = `. With  = 1,
we choose > 0 such that 0 < |x 0| < implies |f (x) `| < 1. Since both x = /2 and x = /2
satisfy 0 < |x 0| < , we must have |1 `| < 1 and | 1 `| < 1. From |1 `| < 1 we get ` > 0,
and from | 1 `| < 1 we get ` < 0. Therefore no such ` exists.
Example 2. Consider the function
(

f (x) =

1 if x Q;
0 if x
/ Q.

We show that limxx0 f (x) does not exist for each x0 R. Fix x0 R. Suppose, on the contrary,
limx0 f (x) = `. With  = 1/2, we choose > 0 such that 0 < |x x0 | < implies |f (x) `| < 1/2.
Choose x1 Q and x2
/ Q satisfying 0 < |x x0 | < . Then we must have |1 `| < 1/2 and
|0 `| < 1/2. From |1 `| < 1/2 we get ` > 1/2, and from |0 `| < 1/2 we get ` < 1/2. Therefore
no such ` exists.
Example 3. Consider the function
(

f (x) =

x if x Q;
0 if x
/ Q.

We show that limx0 f (x) = 0. Let  > 0 be arbitrarily chosen. Since f (x) is one of 0, x, we have
|f (x)| |x|. Therefore 0 < |x 0| <  implies |f (x) 0| < . This proves our claim.
Theorem 2. (Properties of limits)
If limxx0 f (x) = `1 and limxx0 g(x) = `2 , then

(i) limxx0 (f (x) g(x) = `1 `2 ;

(ii) limxx0 (f (x)g(x) = `1 `2 ;

(iii) limxx0 (f (x)/g(x) = `1 /`2 provided `2 6= 0.

In particular, (ii) implies limxx0 lim(cf (x) = c limxx0 f (x) = c`1 .

Proof. Let  > 0 be arbitrarily chosen. Then there exist 1 and 2 such that |f (x) `1 | < 2
whenever 0 < |x x0 | < 1 and |g(x) `2 | < 2 whenever 0 < |x x0 | < 2 . So for = min(1 , 2 ),
both |f (x) `1 | < 2 and |g(x) `2 | < 2 hold whenever 0 < |x x0 | < .
(i) Hence |(f + g)(x) (`1 + `2 )| = |(f (x) `1 ) + (g(x) `2
)| |f (x) `1 | + |g(x) `2 | < 
whenever 0 < |x x0 | < . Therefore limxx0 (f (x) + g(x) = `1 + `2 .
The same argument gives |(f g)(x) (`1 `2 )| = |(f (x) `1 ) (g(x) `
2 )| |f (x) `1 | +
|g(x) `2 | <  whenever 0 < |x x0 | < . Therefore limxx0 (f (x) g(x) = `1 `2 .
(ii)

|(f g)(x) `1 `2 | = |f (x) g(x) `2 + `2 f (x) `1 |

|f (x)||g(x) `2 | + `2 |f (x) `1 |
<
=


2 (`1

2 `1

+ 2 ) + 12 `2 
+ `2 +


2

whenever 0 < |x x0 | < . Therefore limxx0 (f (x)g(x) = `1 `2 .

2

(iii) Exercise.


Proposition 1. If p(x) is a polynomial, then for every x0 R, limxx0 p(x) = p x0 .

Proof. Fix x0 R. Since limxx0 x = x0 , we have limxx0 xk = xk0 by Theorem 2, (ii) and so
P
also limxx0 ck xk = ck xk0 . So with p(x) = nk=0 ck xk , we have by Theorem 2, (i) and the particular
case of (ii)
!
lim p(x) = lim

xx0

xx0

n
X

ck xk

k=0

n
X

k=0

lim ck xk =

xx0

n
X

ck xk0 = p x0 .

k=0


Proposition 2. (The Squeeze Principle)
Suppose that there exists > 0 such that
g1 (x) f (x) g2 (x) for all x satisfying 0 < |x x0 | < .
If limxx0 g1 (x) = ` = limxx0 g2 (x), then limxx0 f (x) = `.
Proof. Let  > 0 be arbitrarily chosen. Choose positive real numbers 1 and 2 such that




g1 (x) ` <  whenever 0 < x x0  < 1 ,





g2 (x) ` <  whenever 0 < x x0  < 2 .

For = min(1 , 2 ), we therefore have both





g1 (x) ` <  and g2 (x) ` < 

whenever 0 < |x x0 | < . So whenever 0 < |x x0 | < , we have

`  < g1 (x) f (x) g2 (x) < ` + .
Thus |f (x) `| <  whenever 0 < |x x0 | < , so that limxx0 f (x) = `.

Continuity

Definition 2. Let f : D R, where D R and contains an interval of the form (a , a + ).

We say that f (x) is continuous at a if limxa f (x) = f (a). We say that f is continuous on a set
S if f is continuous at a, for each a S.
Proposition 3. Suppose f is continuous at a. If f (a) > 0, there is a > 0 such that f (x) > 0 for
x (a , a + ). If f (a) < 0, there is a > 0 such that f (x) < 0 for x (a , a + ).
Proof.
Suppose f is continuous at a, and f (a) > 0. Corresponding to  = f (a)/2, there is
a > 0 such that |f (x) f (a)| <  whenever |x a| < . So for a < x < a + we have
f (x) > f (a) = f (a)/2 > 0. The same argument with  = f (a)/2 supplies the proof for the
case f (a) < 0.

Theorem 3. Let f : D R, and let S D. Then f is continuous
on S if and only if whenever

{sn } is sequence in S that converges to s S, then the sequence f (sn ) converges to f (s).

Proof. Suppose f is continuous on S. Let {sn } be a sequence in S that converges to s S. Let

 > 0 be arbitrarily chosen. Since f is continuous at s, there exists > 0 such that |x s| <
implies |f (x) f (s)| < . Since {sn } s, there
exists
N > 0 such that |sn s| < whenever


n > N . So for n > N , |f (sn ) f (s)| < . Thus f (sn ) f (s).
Conversely suppose f is discontinuous at some point s S. Then there exists  > 0, and for each
> 0, x S, such that |f (x ) f (s)|  and |x s| < . Let = n1 . For all sufficiently large
integer n, this gives rise to a sequence {sn } in S satisfying |sn s|< n1 and
|f (sn ) f (s)| .

Thus the sequence {sn } in S converges to s S, but the sequence f (sn ) does not converge to
f (s). Hence
there is at least one sequence {sn } in S that converges to s S for which the sequence


f (sn ) does not converges to
f (s). This prove that if whenever {sn } is sequence in S that con
verges to s S, the sequence f (sn ) converges to f (s), then f is continuous at s, for each s S. 
Theorem 4. (Intermediate Value Theorem, Bolzano)
Let f be a real-valued function, continuous on [a, b]. For any x1 , x2 [a, b], x1 < x2 , and any c
lying between f (x1 ) and f (x2 ), there exists x0 [x1 , x2 ] such that f (x0 ) = c.
Proof. If f (x1 ) = f (x2 ), choose x0 = x1 . Otherwise suppose f (x1 ) < c < f (x2 ), without loss of
generality. Let
S = {x [x1 , x2 ] : f (x) < c}.
Then S 6= ; let sup S = x0 . We shall show that f (x0 ) = c.
For each sufficiently large integer
n,
choose sn S such that sn > x0 n1 . Then {sn } x0 , and


since f is continuous on S, f (sn ) f (x0 ). Now since f (sn ) < c, f (x0 ) c.
If x0 = x2 , then f (x0 ) > c, which is impossible. Therefore x0 6= x2 .
Hence x0 < x2 ; let tn = x0 + n1 for n 1. The sequence {tn } lies in the interval
[x1 , x2 ] for all

sufficiently large integer n, and converges to x0 . Since f is continuous on S, f (tn ) f (x0 ).
Since tn
/ S, f (tn ) c. Therefore f (x0 ) c.
Thus f (x0 ) = c.

Theorem 5. (Extreme Value Theorem, Weierstrass)
Let f be a real-valued function, continuous on [a, b]. Then there exists x1 , x2 [a, b] such that
f (x1 ) f (x) f (x2 ) for all x [a, b].
Proof. If f is not bounded, for each n N, there exists xn [a, b] for which |f (xn )| > n. By
Bolzano-Weierstrass theorem, {xn } has a subsequence {xnk } that converges to some x0 [a, b].
Since f is continuous
at x0 , the sequence {f (xnk )} must converge to f (x0 ) and hence be bounded.


But f xnk  > nk and {nk } . This contradiction proves that f is bounded on [a, b].
Suppose m f (x) M for x [a, b]. We claim that both sup f and inf f belongs to the range of
f on [a, b]. For each
n N, choose sn [a, b] such that M n1 < f (sn ) M . Observe that the

sequence f (sn ) converges to M . By Bolzano-Weierstrass theorem, {sn } has a subsequence


{snk } that converges to some
s0 [a, b]. Since f is continuous at s0 , the sequence
{f
s
} must
n
k


converge to f (s0 ). But {f snk } being a subsequence of the convergent sequence f (sn ) must also
converge to M . Thus f (s0 ) = M for some s0 [a, b].


References
Richard R. Goldberg, Methods of Real Analysis, Second Edition, John Wiley & Sons, 1976.
Kenneth A. Ross, Elementary Analysis: The Theory of Calculus, Third Edition, McGraw Hill
International Editions, 1976.