Unit 4

June 12, 2021

1 Cauchy’s General Principle of Convergence for Series

Theorem 1.1: Criterion for Covergence

P
The necessary and sufficient condition that the series an should converge is that, for any given
ϵ > 0, there should exist a positive integer m such that

|an+1 + an+2 + an+3 + ... + an+p | < ϵ for all n≥m

and for all integral values of p.

Proof 1.1

Let

sn = a1 + a2 + a3 + ... + an
sn+p = a1 + a2 + a3 + ... + an + an+1 + an+2 + ... + an+p
sn+p − sn = an+1 + an+2 + ... + an+p
P
an converges if and only if the sequence {sn } converges. By Cauchy’s General Principle of
Convergence of Sequences, {sn } converges if and only if given any ϵ > 0, there exists a positive
integer m such that

|sn+p − sn | < ϵ for all n ≥ m and for all positive integral values of p.
=⇒ |an+1 + an+2 + ... + an+p | < ϵ for all n ≥ m and for all positive integral values of p.

Hence the theorem.

1
Example 1:
an where an = 1 + 12 + 13 + ... + n1 + ...
P
Using cauchy’s criterion of convergence, prove that the series
is divergent.
Solution:
an = 1 + 21 + 13 + ... + n1 + ... Let us assume that the series is convergent. Then, by Cauchy’s
P
Let
general principle of convegrence of series, given ϵ > 0, there exists a positive integer m such that

|an+1 + an+2 + ... + an+p | < ϵ for all n ≥ m and for all positive integral values of p.
1 1 1
|an+1 + an+2 + ... + an+p | = + + ... + < ϵ for all n ≥ m,
n+1 n+2 n+p

and for all positive integral values of p.

In particular, take n − m = p. We have,

1 1 1
|an+1 + an+2 + ... + an+p | = + + ... + (m terms)
m+1 m+2 m+m
1 1 1
> + + ... +
m+m m+m m+m
1 1
>m· >
2m 2

That is |an+1 + an+2 + ... + an+p | ≮ ϵ. Therefore we arrive at a contradiction. Hence the given series
is not convergent.
Example 2:
n
Apply Cauchy’s general principle of convergence to prove that an where an = 1− 21 + 13 −...+ (−1)
P
n
+...
is convergent.
Solution: The solution is left to the reader.

2 Test of Convergence and Divergence of a Series of Positive

Terms
In order to prove that a series is convergent or divergent, we have to apply the definition of conver-
gence and for this we have to find the partial sum. In many cases it will not be possible to evaluate
the partial sum and so the definition of convergence cannot be applied directly. Hence some tests
have been devised to discover whether a given series is convergent or divergent or not, without finding
the partial sum.

2.1 Comparison Test for Convergence and Divergence

Theorem 2.1: Comparison Test for Convergence

P P
If there be two series of positive terms, un and vn such that the terms of the first are less
P P
than or equal to the corresponding terms of the second and if vn is convergent, then un is
also convergent.

2
 Proof 2.1

Let
X
un = u1 + u2 + ... + un + ...
X
vn = v1 + v2 + ... + vn + ...

and An , Bn be respectively the partial sums to n terms of the two series.

It is given that

u1 ≤ v1
u2 ≤ v2
......
......
un ≤ vn

Adding all the inequalities, we get

u1 + u2 + ...... + un ≤ v1 + v2 + ........ + vn
=⇒ An ≤ Bn
∴ lim An ≤ lim Bn
n→∞ n→∞

P
But vn is convergent, and hence

limn→∞ Bn = k, a finite value.

P
Therefore limn→∞ An = l, (a finite value). Hence un is also convergent.

Note: Symbollicaly we write that a series is (C). i.e., convergent similarly we write that a
series is (D) i.e., divergent.
Example 3:
un = 1 + 1!1 + 2!1 + 3!1 + . . . to ∞ is convergent.
P
Prove that the series
Solution:
un = 1 + 1!1 + 2!1 + 3!1 + . . . to ∞.
P
Let

X 1 1 1 1
un = 1 + + + + + . . . to ∞ (1)
1 1.2 1.2.3 1.2.3.4
Consider the second series
X 1 1 1 1
vn = 1 + + + + + . . . to ∞ (2)
1 1.2 1.2.2 1.2.2.2

We find that u1 = v1 , u2 = v3 , u3 = v3 .

3
Since 1 · 2 · 3 ≥ 1 · 2 · 2,

1 1
≤ (i.e.,) u4 ≤ v4
1·2·3 1·2·2
P P
Similarly u5 ≤ v5 and so on. Hence in the two series un and vn ,the first three terms are equal
P P
and afterwards each term of un , is less than the corresponding terms of vn . Therefore by the
P P
comparison test, un is (C) if vn is (C).
But,
X 1 1  1 2  1 3
vn = 1 + + + + + ...
1 2 2 2
1
= 1 + a geometric series whose first term is 1 and common ratio is
2
= convergent, since the geometic series is (C), as its common ratio is < 1.
P
Therefore un is also (C).

Theorem 2.2: Comparison Test for Divergence

P P
If there be two series of positive terms, un and vn such that the terms of the first are
P
greater than or equal to the corresponding terms of the second and if vn is divergent, then
P
un is also divergent.

Proof 2.2

Let
X
un = u1 + u2 + ... + un + ...
X
vn = v1 + v2 + ... + vn + ...

and An , Bn be respectively the partial sums to n terms of the two series. It is given that

u1 ≥ v1
u2 ≥ v2
......
......
un ≥ vn

Adding all the inequalities, we get

u1 + u2 + ...... + un ≥ v1 + v2 + ........ + vn
=⇒ An ≥ Bn
∴ lim An ≥ lim Bn
n→∞ n→∞

4
 P
But vn is divergent, and hence

lim Bn = +∞,
n→∞

P
Therefore limn→∞ An = +∞. Hence un is also divergent.

Example 4:
Prove that the harmonic series 1 + 12 + 13 + . . . to ∞ (i.e.,) ∞ 1
P
n=1 n is divergent.
Solution:
Let ! !
X 1 1 1 1 1 1 1
un =1 + + + + + + +
2 3 4 5 6 7 8
! ! (1)
1 1 1 1 1 1
+ + + ... + + + + ... + + . . . to ∞
9 10 16 17 18 32

Consider the second series

! !
X 1 1 1 1 1 1 1
vn =1 + + + + + + +
2 4 4 8 8 8 8
! ! (2)
1 1 1 1 1 1
+ + + ... + + + + ... + + . . . to ∞
16 16 16 32 32 32

1 1
We find that u1 = v1 , u2 = v3 , u4 = v4 , u8 = v8 , u16 = v16 etc. Since > , u3 > v3 ; Like wise,
3 4
1 1 P P
> , u5 > v5 ; u6 > v6 ; u7 > v7 etc. Hence in the two series un and vn some terms are equal.
5 8 P P
But most of terms of un are greater than the corresponding terms of vn . Therefore by the
P P
comparison test un is (D) if vn is (D).
But on simplification , we find that
X 1 1 1
vn = 1 + + + + . . . to ∞
2 2 2
1
= 1 + sum of geometric series whose first term is and common ratio is 1
2
= divergent, since the geometic series is (D) as its common ratio is 1.
P
Therefore un is also (D).

5
Theorem 2.3: Limit Form

P P
If the two series un and vn ,
un
limn→∞ vn
= l,

a finite quantity ̸= 0, then the two series converge or diverge together.

Proof 2.3

Given: limn→∞ uvnn = l, a finite number ̸= 0. Then by definition of limit, corresponding to a

given ϵ however small, there exists a value m ∈ N such that
 
 un
 vn − l < ϵ for all n ≥ m.

un
=⇒ −ϵ < < ϵ for all n ≥ m
vn
un
=⇒ l − ϵ < < l + ϵ for all n ≥ m
vn
un
Therefore omitting the first m terms of both the series ,we have l − ϵ < < l + ϵ for all n.
vn
Case I:
P un
Let vn be (C). Therefore limn→∞ (v1 + v2 + v3 + . . . + vn ) = k, a finite number. Now < l+ϵ
vn
for all n. (i.e.,) un < (l + ϵ)vn for all n.
Therefore,
n
X n
X
un <(l + ϵ) vn for all n
1 1
n
X n
X
lim un <(l + ϵ) lim vn for all n
n→∞ n→∞
1 1

<(l + ϵ)k
(i.e.,) < a finite number.

n
X P
Therefore limn→∞ un is finite. Hence un is (C).
1
Case II:
P un
Let vn be (D). Therefore limn→∞ (v1 + v2 + v3 + . . . + vn ) = ∞. Now > l − ϵ for all n.
vn
(i.e.,) un > (l − ϵ)vn for all n.
Hence,
n
X n
X
un > (l − ϵ) · vn
1 1
n
X n
X
lim un > (l − ϵ) · lim vn → ∞
n→∞ n→∞
1 1

P
Hence un is also (D).

6
2.2 The k-series
The series,
1 1 1 1
k
+ k + k + ... + k + ...
1 2 3 n
is (C) if k > 1 and (D) when k ≤ 1.
Let
X 1 1 1 1
un = k + k + k + . . . + k + . . . (1)
1 2 3 n
Case I: Let k > 1.
P
We write un as ! !
X 1 1 1 1 1 1 1
un = k + + + + k+ k+ k+ k
1 2k 3k 4 5 6 7
! (2)
1 1 1
k
+ k + ... + k + ...
8 9 15

such that the nth bracket of (2) contains 2n−1 terms of the series (1).
Consider the second series,
! !
X 1 1 1 1 1 1 1
vn = k + k
+ k + + k+ k+ k+ k
1 2 2 4 4 4 4
! (3)
1 1 1
k
+ k + ... + k + ...
8 8 8

We find that u1 = v1 , u2 = v2 , u4 = v4 , u8 = v8 , u16 = v16 etc. Since 3k > 2k , k being >1.

1 1
=⇒ < (i.e.,) u3 < v3 .
3k 2k

Similarly

1 1
< and so u5 <v5 ,
5k 4k
P P
and u6 <v6 , u7 <v7 and so on. Hence in the two series un and vn , some terms are equal. But
P P
most of the terms of un are less than the corresponding terms of vn .
P P
Therefore the comparison test , un is (C) if vn is (C). On simplification,
X 1 2 4 8
vn = k
+ k + k + k + ...
1 2 4 8
1 1 1 1
= k + k−1 + k−1 + k−1 + . . .
1 2 4 8
!2 !3
1 1 1 1
= k + k−1 + + + ...
1 2 2k−1 2k−1

1
This is a geometric series with common ratio .
2k−1
1
As k>1, k − 1 is +ve; i.e., k − 1>0. Therefore 2k−1 >20 ; i.e.,>1. and k−1 < 1. Therefore
P
vn is
P 2
(C). Hence un is also (C).

7
Case II: If k = 1,then (1) becomes the harmonic series.

1 1 1 1
+ + + ... + + ...
1 2 3 n

which has been shown to be divergent earlier.

Case III: Let k < 1.
Now
X 1 1 1 1
un = k + k + k + . . . + k + . . . (4)
1 2 3 n
Consider the second series,
X 1 1 1 1
vn = + + + . . . + + . . . (5)
1 2 3 n
1 1
As k<1, nk <n. Therefore k > . i.e., un >vn for all values of n. Hence each term of
P
un after the
n n P P
first is greater than the corresponding term of the harmonic series vn . But vn is (D).
P
Therefore un is also (D).

2.3 Application of the Comparison Test

With the help of the comparison test enunciated in theorem 2.3, we can determine the convergency
P P
or divergency of some given series un . For this ,we have to construct a second series vn (known
as the auxiliary series) whose convergence and divergence is known before hand. We then compare
P P P
un with vn . In most cases, the auxiliary series vn is chosen by giving particular values to k
in the series of 2.2, (i.e.,) in the series

1 1 1 1
k
+ k + k + ... + k + ...
1 2 3 n

The method is illustrated in the following worked examples.

Example 5:
Test for convergence of the series

1 3 5
+ + + ...
1.2.3 2.3.4 3.4.5

Solution:
P
Let the given series be un . Then,

2n − 1
un =
n(n + 1)(n + 2)

Consider the second series,

X 1 1 1
vn = 2
+ 2 + ... + 2 + ...
1 2 n
un 2n − 1 1
= ÷ 2
vn n(n + 1)(n + 2) n
n(2n − 1)
=
n + 1)(n + 2)

8
Therefore,

un n(2n − 1)
lim = lim
n→∞ vn n→∞ n + 1)(n + 2)
 
1 2 − n1
= 1lim   
n
→0 1 + 1 1 + 2
n n

=2 = (a finite quantity) ̸= 0.
P P P
Therefore by the comparison test, un and vn (C) or (D) together, But vn is (C) being a
particular case of the k series
1 1 1
k
+ k + ... + k + ...
1 2 n
which is known to be (C) when k>1. Here k = 2.
P
Therefore un is also (C).
P
Note: The choice of the comparison series vn is due to the following consideration. In un , the
degree n in the numerator is 1 and its degree in the denominator is 3. For large values of n, un is
2n 2 1
approximately equal to (i.e.,) 2 . So un is of the order of 2 and hence we compare it with
n.n.n n n
P 1
the series .
n2
Example 6:
Prove that the general harmonic series

1 1 1
+ + + ...
a a + b a + 2b

where a and b are positive numbers diverge to ∞.

Solution:
P
Let the given series be un . Then

1
un =
a + (n − 1)b

P 1
Consider the second series vn where vn = .
n
un 1 1
= ÷
vn a + (n − 1)b n
n
=
a + (n − 1)b
un n
lim = lim
n→∞ vn n→∞ a + (n − 1)b
1
=  
a 1
n
+ 1 − n
b
1
= = a finite quantity̸= 0 (Since b is positive)
b

9
 P P P
∴ By the comparison test, un and vn (C) or (D) together. But vn is (D) being a particular
case of the k-series
1 1 1
k
+ k + ... + k + ...
1 2 n
which is known to be (D) when k<1. Here k = 1.
P
Therefore un is also (D).
Example 7:
∞ 
X √ 
Test for convergence of the series n2 + 1 − n .
n=1
Solution:
P
Let the given series be un . Then
√ 
u n = n2 + 1 − n
√ √ 
n2 + 1 − n n2 + 1 + n
= √ 
2
n +1+n
n2 + 1 − n2
= √ 
n2 + 1 + n
1
= √ 
n2 + 1 + n

Consider the second series,

X 1 1 1
vn = + + ... + + ...
1 2 n
un 1 1
= √ ÷
vn n2 + 1 + n n
un n
lim = lim √ 
n→∞ vn n→∞
n2 + 1 + n
1
= 1lim r
→0

n 1
1+ n2
+1
1
= = (a finite quantity) ̸= 0
2
P P P
∴ By the comparison test, un and vn (C) or (D) together. But vn is (D) being a particular
case of the k-series,
1 1 1
k
+ k + ... + k + ...
1 2 n
which is known to be (D) when k<1. Here k = 1.
P
Therefore un is also (D).
Example 8:
Discuss the convergence of the series
∞ √ √
X n+1− n
1
np

10
Solution:
P
Let the given series be un . Then,
√ √
n+1− n
un = p
√ n √ √ √ 
n+1− n n+1+ n
= √ √ 
n p n+1+ n
n+1−n
= √ √ 
n p n+1+ n
1
= √ √ 
n p n+1+ n

P 1
Consider the second series vn , where vn =
1
p+
n 2

un 1 1
= √ √ ÷ 1
vn np n + 1 + n
p+
n 2
1 √
n2 n
= √ √  = √ √ 
n+1+ n n+1+ n
√
un n
lim = lim √ √ 
n→∞ vn n→∞
n+1+ n
1
= 1lim !
n
→0 q
1
1+ n
+1

1 1
= = = (a finite quantity) ̸= 0.
1+1 2
P P P
∴ By the comparison test, un and vn (C) or (D) together. But vn is a particular case of the
k-series,
1 1 1
k
+ k + ... + k + ...
1 2 n
which is known to be (C) when k>1 and (D) if k ≤ 1. Here k = p + 21 . Therefore
P
vn is (C) if
p + 12 >1 (i.e.,) if p> 21 and it is (D) if p ≤ 21 . Hence un is also (C) if p> 12 and (D) if p ≤ 21 .
P

Example 9:
Examine the convergence of ∞ 1
P
1 sin n .
Solution:
P
Let the given series be un . Then,
1
un = sin
n

11
 1
P
Consider the second series vn , where vn = n

un sin n1
= 
vn 1
n

un sin 1
lim = lim  n
n→∞ vn n→∞ 1
n

sin 1
= 1lim  n
→0 1
n
n
 sin θ 
=1 ∵ lim=1
θ→0 θ

=(a finite quantity) ̸= 0.

P P P
Hence by the comparison test, un and vn (C) or (D) together. But we know that vn is (D),
being a particular case of the k-series

1 1 1
k
+ k + ... + k + ...
1 2 n
P
which is known to be (D) if k ≥ 1. Here k = 1. So un is also (D).

Homework Sums
P∞  √
3 3

1. Test for convergence the series n=1 n +3−n .

2. Test for convergence the series

2 3 4 5
p
+ p + p + p + ...
1 2 3 4

3. Discuss the convergence of the series

∞
X 1
(a, b, p, q being all positives).
n=1
(a + n)p (b + n)q

4. Test for covergency or divergency the series

∞ 
X √ √ 
n4 + 1 − n4 − 1 .
1

12
3 Binomial Theorem
Theorem 3.1: Binomial Theorem for Any Rational Index

If n is any number, positive or negative, integral functional, and −1 < x < 1 (i.e.,) |x|<1, the
sum of the series

n(n − 1) 2 n(n − 1) . . . (n − r + 1) r
1 + nx + x + ... + x + . . . to ∞ is (1 + x)n .
2! r!

Proof 3.1

n(n − 1) 2
For all values of n, the series 1 + nx + x + . . . to ∞ known to be absolutely convergent
2!
when |x|<1. Let its sum be determined.
Consider the two series

n1 n2 nr
f (n) = 1 + x + x2 + . . . + xr + . . .
1! 2! r!

and

m1 m2 2 mr r
f (m) = 1 + x+ x + ... + x + ...
1! 2! r!

Since these two series are absolutely convergent when |x| < 1. They can be multiplied and the
resulting product series is also an absolutely convergent series.
Now h n1 n2 2 nr r i
f (m).f (n) = 1 + x + x + . . . + x + . . . ×
1! 2! r!
h m1 m2 2 mr r i
1+ x+ x + ... + x + ...
1! 2! r!
m n1   m2 m1 n1 n2  2
1
=1+ + x+ + + + x + ...
1! 1! 2! 1! 1! 2!
In the R.H.S ., the (r + 1)th term (which is the coefficient of xr )

mr mr−1 n1 mr−2 n2 nr
= + + + ... +
r! (r − 1)! 1! (r − 2)! 2! r!
1 h r! r! i
= mr + .mr−1 .n1 + .mr−2 .n2 + . . . + nr
r! (r − 1)!1! (r − 2)!2!
1
= [m1 +r C1 .mr−1 .n1 +r C2 .mr−2 .n2 + . . . + nr ]
r!
1
= (m + n)r (by Vandermonde’s theorem)
r!

Therefore

(m + n)1 (m + n)2 2 (m + n)r r

f (m).f (n) = 1 + x+ x + ... + x + ...
1! 2! r!
= f (m + n) in our notation,

13
Similarly

f (m).f (n).f (p) =f (m + n).f (p)

=f (m + n + p) and so on
Alsof (1) = 1 + x and
f (0) =1 by actual substitution in f (n).

We consider the following three cases.

Case I: Let n be a positive integer.

f (n) =f (1 + 1 + . . . to n terms )
=f (1)f (1)f (1) . . . to n factors
=[f (1)]n
f (n) =(1 + x)n

and this is simply the binomial theorem for a positive integral index.
Case II: Let n be a positive fraction.
r
Put n = where r and s are positive integers.
s
r r r r 
f f . . . to s factors = f + + . . . to s terms
s s  rs s
=f ·s
s
= f (r)
= (1 + x)r (by case I as r is a positive integer)
h  r is
f .s = (1 + x)r
s
r r
f = (1 + x) s
s
(i.e., ) f (n) = (1 + x)n .

Case III: Let n be any negative quantity (integral or fractional).

Put n = −m, so that m is positive.
Now

f (n).f (m) =f (n + m)
=f (−m + m) = f (0) = 1.

14
 Therefore

1 1
f (n) = = (by case I and II as m is a positive)
f (m) (1 + x)m
=(1 + x)−m
=(1 + x)n .

Therefore, for all values of n

f (n) = (1 + x)n

n(n − 1) 2 n(n − 1)(n − 2) . . . (n − r − 1) r

(i.e., ) (1 + x)n = 1 + nx + x + ... + x + ...
2! r!

Note:

1. Throughout we have assumed that the series f (n) is absolutely convergent for all values n
provided |x| < 1.

2. The series

n(n − 1) 2 n(n − 1)(n − 2) . . . (n − r − 1) r

1 + nx + x + ... + x + . . . to ∞
2! r!

is called the Binomial Series. It is the binomial expansion for the function (1 + x)n .

3. The binomial series can be written as

∞
X n(n − 1)(n − 2) . . . (n − r + 1)
xr
r=0
r!

Then (r + 1)th term, namely

n(n − 1)(n − 2) . . . (n − r + 1) r
x
r!

is called the general term.

4. The coefficient of the general term will be zero only when one of the factors of its numerator
is zero. When n − r + 1 = 0 i.e., When r = n + 1, the series will stop at the rth term. But r is
a positive integer. So when r = n + 1, it follows that n must be a positive integer. Hence the
expansion by the binomial theorem extends to n + 1 terms when n is a positive integer and to
an infinite number of terms in all the other cases.

5. Let the (r + 1)th term in the expansion of(1 + x)n be denoted by tr−1 and the rth t erm by tr .
Then
n(n − 1)(n − 2) . . . (n − r + 1) r
tr−1 = x (1)
r!
n(n − 1)(n − 2) . . . (n − r + 2) r−1
tr = x (2)
(r − 1)!

15
 Dividing (1) by (2),

tr+1 n−r+1
= x
tr r
This simplification will be very useful in working problems.

4 Exponential Theorem
Theorem 4.1: Exponential Theorem

For all values of x, positive or negative, integral or fractional, the sum of series

x x2 x3 xr
1+ + + + ... + + . . . to ∞ is ex .
1! 2! 3! r!

Proof 4.1

For all values of x, the series

x x2 x 3 xr
1+ + + + ... + + . . . to ∞.
1! 2! 3! r!

is known to be absolutely convergent. Let its sum be denoted by f (x) . Consider the two series

x x2 x3 xr
f (x) =1 + + + + ... + + . . . to ∞
1! 2! 3! r!
y y2 y3 yr
and f (y) =1 + + + + ... + + . . . to ∞
1! 2! 3! r!

Since these two series are absolutey convergent for all values of x and y, they can be multiplied
and the resulting product series is also an absolutely convergent series.
Now
" #
2 3 r
x x x x
f (x).f (y) = 1 + + + + ... + + . . . to ∞
1! 2! 3! r!
" #
2 3 r
y y y y
× 1+ + + + ... + + . . . to ∞
1! 2! 3! r!
!
x2 y2
 
x y x y
=1+ + + + . + + ...
1! 1! 2! 1! 1! 2!

16
In the R.H.S., the r + 1th term

xr xr−1 y xr−2 y 2 yr
= + + + ... +
r! (r − 1)! 1! (r − 2)! 2! r!
 
1 r r! r−1 1 r! r−2 2 r
= x + x y + x y + ... + y
r! (r − 1)!.1! (r − 2)!.2!
1 r
x +r C1 xr−1 y 1 +r C2 xr−2 y 2 . . . + y r

=
r!
1
= [x + y]r (by the binomial theorem)
r!

Therefore

(x + y) (x + y)2 (x + y)r
f (x).f (y) = 1 + + + ... + + . . . to ∞
1! 2! r!
= f (x + y) ( in our notation)

Similarly

f (x).f (y).f (z) = f (x + y) · f (z)

= f (x + y + z) and so on
Also by actual substitution in f (x), we havef (0) = 1

1 1
and f (1) = 1 + + + . . . to ∞ = e
1! 2!

We now consider the following three cases.

Case I: Let x be a positive integer.

f (x) = f (1 + 1 . . . to x terms)
= f (1).f (1) . . . to x factors
= [f (1)]x
= ex

Case II: Let x be a positive fraction.

Put x = (r/s) where r and s are positive integers.

17
    
r r
f .f . . . to s factors
s s
 
r r r
=f + + . . . to s terms
s s s
 
r
=f ·s
s
= f (r)
= er (by Case I, as r is a positive integer)
"  #s
r
i.e., f = er
s
Taking the sth root of the both sides,
 
r r
f = es
s
i.e., f (x) = ex

Case III: Let x be any negative quantity (integral or fractional). Put x = −y, so that y is
positive. Now

f (x).f (y) = f (x + y)
= f (−y + y) = f (0) = 1

Therefore

1
f (x) =
f (y)
1
= y
e
= e−y
= ex .

Therefore, for all values of x,

f (x) = ex
x x2 x3 xr
i.e., ex = 1 + + + ... + + . . . to ∞
1! 2! 3! r!
This is known as the Exponential Theorem. Since the variable occurs as an exponent in the
function ex represented by the above series. The function is called an exponential function and
the corresponding series an Exponential Series.

18
5 Logarithmic Series
The student is already familiar with the logarithmic notation.

If N = aN , (a > 0) (3)

x is called the logarithm of N to the base a and is written as

x = loga N (4)

N
Hence from (3) and (4), N = aloga which is a very useful result.

Common and Natural Logarithms:

In the arithmetic work, logarithms are taken to the base 10 and they are called common logarithms.
In theoretical work logarithms are taken to the Napierian base e and these are called Natural loga-
rithms.

The following elementary theorems concerning logarithms are of frequent occurrence:

1. loga N = logb N × loga b

2. logb a · loga b = 1

From these, we have the following simple relation between the systems of logarithms:
log10 N
loge N = log10 N · loge 10 =
log10 e
Hence log10 N = loge N · log10 e

Hence, if the logarithm of N to the base e is known, the common logarithm is obtained by mul-
tiplying by the constant factor log10 e. This constant factor is called the modulus and its value is
0.43429 nearly.

Hereafter, unless otherwise specified, logarithms are always natural logarithms.

5.1 The Logarithmic Series

If |x| < 1 (i.e., ) − 1 < x < 1, then
x 2 x3 xn
log(1 + x) = x − + − . . . + (−1)n−1 + . . .
2 3 n
If |x| < 1, we have for all values of y, by the binomial theorem

y(y − 1) 2 y(y − 1)(y − 2) 3

(1 + x)y = 1 + yx + x + x + ... (5)
2! 3!

19
 Since this series is absolutely convergent, it can be re-arranged in powers of y. Then the coefficient
of y will be

(−1) 2 (−1)(−2) 3 (−1)(−2)(−3) 4

=x+ x + x + x + ...
2! 3! 4!
1 1 1
= x − x2 + x3 − x4 + . . .
2 3 4

Also by the exponential theorem,

(1 + x)y = eylog(1−x)
ylog(1 + x) (ylog(1 + x))2
=1+ + + ... (6)
1! 2!
(5) and (6) are different expansions of the same expression. So the coefficient of corresponding powers
of y in the two must be equal. Hence equating the coefficient of y, we get
1 1 1
log(1 + x) = x − x2 + x3 − x4 + . . .
2 3 4
Note:
When x = 1, the above series for log(1 + x) is known to be convergent. So that,
1 1 1
log 2 = 1 − + − + ...
2 3 4
Hence we may state the theorem as follows:
For the values of x such that −1 < x < 1,
1 1 (−1)n−1 n
log(1 + x) = x − x2 + x3 − . . . + x + ...
2 3 n

20