Lecture 3

1 Sequences ctd..
Theorem 1.0.1 (Sandwich theorem for sequences). Let {an }, {bn } and {cn } be three sequences
such that an ≤ bn ≤ cn for all n ∈ N. If lim an = L and lim cn = L, then lim bn = L.
n→∞ n→∞ n→∞

Proof. Let  > 0 be given. As lim an = L, there exists N1 ∈ N such that

n→∞

n ≥ N1 =⇒ |an − L| < . (1.1)

Similarly as lim cn = L, there exists N2 ∈ N

n→∞

n ≥ N2 =⇒ |cn − L| < . (1.2)

Let N = max{N1 , N2 }. Then, L −  < an (from (1.1)) and cn ≤ L +  ( from (1.2)). Thus

L −  < an ≤ bn ≤ cn ≤ L + .

Thus |bn − L| <  for all n ≥ N . Hence the proof. ///

Examples 1.0.2.
n cos n o∞ −1 cos n 1
(i) Consider the sequence . Then ≤ ≤ . Hence by Sandwich theorem
n n=1 n n n
cos n
lim = 0.
n→∞ n
1 1
(ii) As 0 ≤ 2n ≤and n1 → 0 as n → ∞,
n
1
2n also converges to 0 by Sandwich theorem.
√
(iv) If b > 0, then lim n b = 1.
n→∞
1
First assume that b > 1. Let an = b n − 1. As b > 1, an > 0 for all n ∈ N. Further,

b = (1 + an )n ≥ 1 + nan .

b−1 1
Then 0 ≤ an ≤ . Thus an → 0, i.e., b n → 1 as n → ∞.
n
Now if b < 1, then take c = 1b and it is easy to show the result. ///

Examples 1.0.3.
√
(i) lim n n = 1.
n→∞

nx
(ii) If x > 0 then lim n = 0.
n→∞ (1+x)

log(n)
(iii) If p > 0, then lim p = 0.
n→∞ n

1
 1
Solution. (i) Let an = n n − 1. Then 0 ≤ an ≤ 1 for all n ∈ N. Further,

n(n − 1) 2
n = (1 + an )n ≥ an .
2
q q
2 2
Thus 0 ≤ an ≤ (n−1) (n ≥ 2). As (n−1) → 0 as n → ∞, by Sandwich theorem, an → 0, i.e.,
1
n → 1 as n → ∞.
n

(ii) Let k be an integer such that k > x, k > 0. Then for n > 2k,

n! xk k nk xk
(1 + x)n > n Ck x
k
= xk = Π [n − i + 1] > k .
k!(n − k)! k! i=1 2 k!

Hence,
nx 2k k! x−k
0< < n .
(1 + x)n xk
nx
As x − k < 0, nx−k → 0. Thus → 0 as n → ∞.
(1 + x)n
(iii) By Archimedian property, for any n ∈ N there exists m ∈ N such that

m ≤ np < (m + 1)

or equivalently
1 1
m p ≤ n < (m + 1) p .

Therefore,
1 log n 1 log(m + 1)
p
< p < .
n n p m
Also
1 log(m + 1) 1 log(m + 1) m + 1
= .
p m p m+1 m
So it is enough to show logn n → 0 as n → ∞. for this,
1
let us start with  > 0. From the previous problem we have n n → 1 as n → ∞. This implies
there exists N ∈ N such that
1 log n
n n ∈ (e− , e ), ∀ n ≥ N (equivalently) ∈ (−, ), ∀ n ≥ N.
n
log n
That is, n → 0. ///
Definition 1.0.4. Subsequence: Let {an } be a sequence and {n1 , n2 , ...} be a sequence of
positive integers such that i > j implies ni > nj . Then the sequence {ani }∞i=1 is called a
subsequence of {an }.
Theorem 1.0.5. If the sequence of real numbers {an }∞
1 , is convergent to L, then any subse-
quence of {an } is also convergent to L.
Proof. Let {ni }∞ ∞
i=1 be a sequence of positive integers such that {ani }i=1 forms a subsequence of

2
{an }. Let  > 0 be given. As {an } converges to L, there exists N ∈ N such that

|an − L| < , ∀ n ≥ N.

Choose M ∈ N such that ni ≥ N for i ≥ M . Then

|ani − L| < , ∀ i ≥ M.

Hence the proof. ///

Theorem 1.0.6. 1. If a sequence {an } converges to a. Then {|an |} converges to |a|.

√ √
2. If an ≥ 0 and an → a, then an converges to a.

Proof. Proof of (1) follows from the inequality

||a| − |b|| ≤ |a − b|, ∀a, b ∈ R.

(2) follows from the fact that if a 6= 0, then

√ √ |an − a|
| an − a| ≤ √ √ .
| an + a|

The case a = 0 is easy and is left as an exercise.

n
( )
1X
Theorem 1.0.7. If a sequence {an } converges to a, then ak also converges to a.
n
k=1

Proof. Proof is left as an exercise. ///

Next we study some fine properties of sequences that imply convergence/divergence.
Theorem 1.0.8. For any sequence {an } with an > 0
an+1
lim a1/n = lim
n→∞ n n→∞ an

provided the limit on the right side exists.

Proof. Let  > 0 be arbitrary. Suppose the second limit exists (say l), then there exists N ∈ N
such that
an+1
l−< < l + , ∀n ≥ N.
an
Taking n = N, N + 1, ...., m − 1 and multiplying we get
am
(l − )m−N < < (l + )m−N , ∀m ≥ N + 1
aN

equivalently,
N 1 1 N 1
(l − )1− m aN
m
< (am ) m < (l + )1− m aN
m
, ∀m ≥ N + 1.

3
 1/m
Now the result follows from the fact that lim (l ± )1−N/m aN = l ± . ///
m→∞
an+1
Corollary 1.0.9. (i) If an > 0 and lim = l < 1, then lim an = 0.
n→∞ an n→∞

an+1
(ii) If an > 0 and lim = l > 1, then an → ∞.
n→∞ an

Proof. If L < 1, then choose 0 such that L + 0 < 1. Then for this 0 there exists N0 such that
an+1
n ≥ N0 =⇒ < L + 0
an

Therefore from the hypothesis of (i)

1
n ≥ N0 =⇒ ann < L + 0

Therefore an < (L + 0 )n → 0 as n → ∞. A similar argument will imply the (ii). ///

Examples 1.0.10. (i) lim a1/n = 1, if a > 0 (ii) lim nα xn = 0, if |x| < 1 and α ∈ R.
an+1
Solution: (i) Take an = a, then lim = 1.
n→∞ an
an+1 1
(ii) If x 6= 0, take an = nα xn , then lim = lim (1 + )α |x| = |x|.
n→∞ an n→∞ n