BULL. AUSTRAL. MATH. SOC.

30A32, 30A34
VOL. 16 ( 1 9 7 7 ) , 415-425.

Coefficient estimates for

starlike functions

M.L. Mogra and O.P. Juneja

Let 5?(a, $) denote the class of functions

ftz) = z + I az11
n
n=k+l

analytic in the unit disc A = {s : | s | < 1} and satisfying

for some a, g (0 5 a < 1, 0 < S 5 l) and for all z £ A . In

the present paper, sharp coefficient estimates for functions in
SjHot, 3) have teen obtained. The results thus obtained not only

generalize the corresponding results of Thomas H. MacGregor

{Michigan Math. J. 10 (1963), 277-281), A.V. Boyd (Proa. Amer.
Math. Soa. 17 (1966), 1016-1018) and others, but also give rise
to analogous results for various other subclasses of starlike
functions.

1. Introduction
Let f(z) be regular in the unit disc A E {3 : |s| < 1} and
normalized by the conditions /(0) = 0 , /'(0) = 1 . The power series
representation for such a function is

(1.1) ftz) = 3 + I a zn .
n=2

Received 26 January 1977.

415

Downloaded from https://www.cambridge.org/core. IP address: 18.206.13.133, on 13 Jun 2020 at 11:33:26, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0004972700023492
 416 M . L . Mogra and O . P . J u n e j a

If f(z) satisfies the condition

(1.2) Re[z ^ 1 } > 0

for all z i A , then it is well known [ 9 , p. 221] that (1.2) is both
necessary and sufficient for / to be univalent and starlike with respect
to the origin in A . For starlike functions the Bieberbach conjecture
\a | 5 n holds for all n , and equality occurs only for the functions

f(z) = js/Cl+es) , where |e| = 1 . MacGregor [6] obtained upper bounds

for the moduli of the coefficients of a starlike function whose power
series representation in A is of the form

GO

(1.3) f(z) = z + £ a zzn

n=k+l

Boyd [7] extended MacGregor ' s result to the class S* of starlike

functions of order a (0 £ a < 1) ; that is, / ( S ' if it is univalent

and Re{zf (z )//(s)) > a for z € A . Various other subclasses of star-

like functions have been considered and corresponding coefficient estimates
obtained for them by Singh [73, 7 4 ] , Padmanabhan [70], Eenigenburg [ 2 ] ,
McCarty [ 7 ] , Mogra [S] , and others.

Recently, the authors [4] introduced the class of starlike functions

of order a (0 £ a < 1) and type $ (0 < f$ 5 1) and made a preliminary
study of its properties. Later, Gupta and Jain [3] obtained certain
results for functions of this class when all a are negative. These

results were the analogues of the corresponding results obtained by

Silverman [72] for starlike functions of order a . However, to broaden
the scope of applicability of the results obtained, the authors [5] have
very recently modified the definition of starlike functions of order a
(0 5 a < 1) and type & (0 < 3 £ 1) , which is as follows.

CO

DEFINITION. Let f{z) = z + £ a zn be analytic in the unit disc

n
n=2
A . Then f is said to be starlike of order a and type 3 , if it
satisfies the condition

Downloaded from https://www.cambridge.org/core. IP address: 18.206.13.133, on 13 Jun 2020 at 11:33:26, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0004972700023492
 Coefficient estimates 417

for some a, B (0 < a < 1, 0 < B < 1) and z dA .

The class of starlike functions of order a (0 5 a < l) and type B

(0 < g < 1) we shall denote by S*{a, B) .

It can be easily seen that S*(a, B) includes the subclasses of

starlike functions for different values of the parameters a and B •
Hence a study of its various properties leads to a unified study of these
subclasses. In the present note, we determine sharp coefficient estimates
for the class Sj[(a> B) of starlike functions of order a and type B

whose power series representation is of the form (1.3). The results thus
obtained not only generalize the corresponding results of MacGregor [ 6 ] ,
Boyd [/], and others, but also give rise to analogous results for the
functions of the form (1.3) belonging to the classes introduced and studied
by Singh [73, 74], Padmanabhan [10], Wright [15], McCarty [7], and
Eenigenburg [2].

2. Some coefficient estimates

00

THEOREM. Let f(z) = z + £ a zn be in S*(a, B) .

n=k+l

B(l-a)" .
(a) If B(l-a) > fc(l-S) , let M = Then

for mk + 1 2 n £ (m+l)k , m = 1 , 2 , . . . , M+l and

(b) If B(l-a) < fe(l-B) , then

(2.3)

The estimates in (2.1) are sharp for n = mk + 1 , m = 1, 2, . . . , while

the estimates in (2.3) are sharp for all n .
Proof. We employ the technique used by MacGregor [ 6 ] . Thus, l e t
/ € Stta, B) ; then we have

Downloaded from https://www.cambridge.org/core. IP address: 18.206.13.133, on 13 Jun 2020 at 11:33:26, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0004972700023492
 418 M . L . Mogra andO . P . Juneja

(2 k) MB) - zf'(z)-f(z)
( K
' ' 2&{zf (z)-*f(z))-{zf' (z)-f(z)}

where h is regular in A and satisfies |7z(s)| < 1 in A • Also the

power series for h{z) begins with c,z + e, z + ... . Equating

C9efficients of the same powers on both sides of the equation

zf'(z) - f(z) = Mz){2&(zf'{z)-af{z))-[zf'(z)-f(z))}

or
n
(2.5) ££ (n-l)az
az n
n
n=k+l
k+1
k+1z + ...J|28(l-c)a

we o b t a i n

(2.6) (n-l)an = 2&{l-a)cn_1 for n = k+1, k+2, . . . , 2k

Since \h(z)\ < 1 , it follows that Y |e | « 1 and so

n
n=k

(2-7) I \a \2 5 1 .
n=k

From ( 2 . 6 ) a n d ( 2 . 7 ) , we f i n d t h a t

2k
(2.8) £ (n-l)2|a |2< l * 2 2
n

(2.5) can be rewritten in the form

(2.9) | (n-l)«/ + £ d zn
n n
n=k+l n=p+l

{ n=fe+l
p-k
p-K. \

2B(l-a)s + £ ((2B-l)n+l-2a8)a zn\

n
Since (2.9) has the form n=k+l >
F(z) = h{z)G{z) , where |?i(s) | < 1 , i t follows
that

(2.10) i- 0 |F(rei(f) |2dj) 5 i

Downloaded from https://www.cambridge.org/core. IP address: 18.206.13.133, on 13 Jun 2020 at 11:33:26, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0004972700023492
 Coefficient estimates 419

for each r (0 < r < l) . Expressing (2.10) in terms of the coefficients

in (2.9), we get

(2.11) | (n-l)2|a

((26-l)n+l-2ctB)2|a
n
n=k+l

In particular, (2.11) implies

(2.12) I (n-l)2\a\2r2n
n
n=k+l
p-k
((2B-l)n+l-2a6)2|a |
n
n=k+l

Letting v -»• 1 in (2.12), we conclude that

(2.13) (n-D2\an | 2 |a n | 2 .
n=k+l

This inequality is equivalent to

(2.H0 f; (n-l)2|aj2
n=p-k+l
p-k
2 2
i ( ) X {((28-l) n +l-2a6) 2 -(«-l) 2 Ha | 2
n=k+l n

Now two cases arise.

(a) If g(l-a) > fe(l-3) , then by an inductive argument we will

establish the inequalities

(m+l)k
(2.15b)

m-

T
g(l-a)"
for m = 1, 2, .... M+l ; M = where [p] denotes the greatest

Downloaded from https://www.cambridge.org/core. IP address: 18.206.13.133, on 13 Jun 2020 at 11:33:26, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0004972700023492
 420 M.L. Mogra and O.P. Juneja

integer not greater than p .

For m = 1 , (2.15a) gives
2k
I (n-l)2\a n
| 2 5 l+e2(l-a)2
n=k+l

which i s t h e same a s ( 2 . 8 ) . Thus (2.15a) i s v a l i d for m = 1 . We can

prove (2.15t>) for m =1 by using (2.8) as follows:

2k
T {((2g-l)n+l-2a6) 2 -( w -l) 2 }|a n I 2
fcl

w ± 2. ,2
2 L ~' \ann
n
k n=k+l

How suppose t h a t (2.15a) and (2.15b) hold for m = 1 , 2, . . . , <y-l . Using

(2.I**) with p = (q+l)k and the inductive hypothesis concerning (2.15a),
we obtain the inequalities

(q+Dk p ?
I (n-D2\an\2
n=qk+l

+ ^ {((23-l)n+l-2 aB ) 2 -(n-l) 2 Ha | 2
n
n=k+X
q-1 (m+l)k o c o
+ I I {((2B-l)«+l-2aB) 2 -(»-l) 2 }|a | 2
m=l n=mk+l

feTT f(2B-Dy
m
- p =ol

The last equality can be easily obtained by an inductive argument on q .

This last sequence of inequalities implies (2.15a) where m = q .

Continuing our argument, we ure (2.15a) with m = q to deduce (2.15b)

for m - q as follows:

Downloaded from https://www.cambridge.org/core. IP address: 18.206.13.133, on 13 Jun 2020 at 11:33:26, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0004972700023492
 Coefficient estimates 421

[q+l)k o o o
I {((23-l)n + l-2 a B) 2 -(n-l) 2 }|a n | 2
n=qk+l

£ (n_±f\a| | 2

q k n=qk+l

2,2
q k
-1
= {^7 f t
This completes the proof of (2.15a) and (2.15t>). Now (2.1) follows from
(2.15a).

To prove (2.2), suppose n > (M+2)k . Putting p = (q+l)k in

(2. ll*), we have

(q+l)k o o o o qk „ „ „
(n-D 2 |a | 2 < UB2(l-a)2 + I { ((2B-l)n+l-2aB) -(n-1) 2 } \a | 2 .

Hence, for n > (M+2)k , we have

(2.16) (n-D2\an\2

5 U B 2 (l-a) 2 + \ { ( )
{((2B-l)«+l-2}a3)2-(n-l)
| a2}|a n | 2
k
n=k+l

B (1a) + 7 {((2B-l)n+l-2ae)2-(n-l)2}|a | 2
n
n=k+l

{((2e-l)K+l-2aB)2-(n-D2Ha | 2
n

M+l (m+l)k
£ £ {((28-l)«+l-2ctB)2-(n-l)2}|a | 2
m=l n=mk+l
q-X (m+l)k „
E I {((2B-l)n+l-2aB)-(n-l) 2 }|a | 2
2 =mk+l
M+l (m+l)k
s i*B2(l-a)2 + E I {((26-l)«+l-2aB)2-(«-l)2}|a | 2 .
n
m=l n=mk+l

Using (2.15a) in (2.l6), we obtain

Downloaded from https://www.cambridge.org/core. IP address: 18.206.13.133, on 13 Jun 2020 at 11:33:26, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0004972700023492
 422 M.L. Mogra and O.P. Juneja

that is,

l a j ~ (n-lHM+l)l ft ((28-1)11 + 2 g ^- g ) ] , n > (M+2)fc

This proves (2.2).

(b) If B(l-a) £fe(l-e), then (2.13) gives

^ (*-D2|a | 2 £ Jtf"

(n-D2\an\2 < l+62(l-a)2 if n 2 fc+i ;

that is,

| | 2p(l-«) lf
1
n' n-1

which gives (2.3).

The function / , given by

shows that the estimates in (2.1) are sharp for n = mk + 1 ,

m = 1, 2, 3, while the estimates in (2.3) are sharp for the function

f{z) = z exP{
where B(l-a) < &(l-3) and n >
Putting 3 = 1 in the theorem, we get the following result due to
Boyd [I}.
GO

COROLLARY 1. If f(z) = s + V a zn is starlike of order a ,

n
n=k+l
0 s a < 1 , then

Downloaded from https://www.cambridge.org/core. IP address: 18.206.13.133, on 13 Jun 2020 at 11:33:26, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0004972700023492
 Coefficient estimates 423

KJ - (n-D(m-l)I JJQ [y
where mk+1 < n s (m+l)?c , m = 1 , 2 , . The result is sharp for
n = mk + 1 , m = 1 , 2 , /or the function

The following r e s u l t , due to MacGregor [ 6 ] , can be obtained by taking

(a, 8) = ( 0 , 1) in the theorem.
OO
n
COROLLARY 2. If f(z) = z + Y, a
a zz is in S*(0, 1) , then
n
n=k+l
m-1
Uiv+{2/k)]
-DI" u=0
where mk+1 5 n £ (m+l)fe , m = 1, 2, — . The estimates are sharp for
n = mk + 1 (m = 1, 2, ...) , for

REMARKS. (i) The coefficient estimates determined in [5] for the

starlike functions of order a and type 3 can be obtained by putting
k =1 in the theorem.

(ii) Putting

3 =% ,
(a, g) = (0, %) ;
(a, 8) = (0, (26-D/26 where 6 > %) ;

(a, 6) = (U-Y)/Un), (l+Y)/2)

where 0 < y S 1 , and replacing a by 1 - a and B by % in the
theorem, we get respectively the corresponding coefficient estimates for
the functions of the form (1.3) belonging to the classes introduced by
McCarty [ 7 ] , Singh [J3, 74], Padmanabhan [JO], and Eenigenburg [2].

(iii) The results due to SeniId [7J], Singh [73, 74], Eenigenburg
[2]*Mogra [S], McCarty [ 7 ] , can be obtained by taking different values of
the parameters a,8 ( 0 < o < 1, 0 < J < 1) with k = 1 in the theorem.

Downloaded from https://www.cambridge.org/core. IP address: 18.206.13.133, on 13 Jun 2020 at 11:33:26, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0004972700023492
 424 M.L. Mogra and O . P . Jusneja

References

[7] A.V. Boyd , "Coefficient estimates for starlike functions of order

a ",Proe. Amer. Math. Soc. 17 (1966), 1016-1018.

[2] P.J. Eenigenburg, "A class of starlike mappings in the unit disc",
Compositio Math. 24 (1972), 235-238.

[3] V.P. Gupta and P.K. Jain, "Certain classes of univalent functions
with negative coefficients", Bull. Austral. Math. Soo. 14
(1976), kO9-kl6.

[4] O.P. Juneja and M.L. Mogra, "On starlike functions of order a and
type S ". Notices Amer. Math. Soo. 22 (1975), A-38U; Abstract
No. 75T-B80.

[5] O.P. Juneja and M.L. Mogra, "On starlike functions of order a and
type 3 ", Rev. Roumaine Math. Pures Appl. (to appear).

[6] Thomas H. MacGregor, "Coefficient estimates for starlike mappings",

Michigan Math. J. 10 (1963), 277-281.

[7] Carl P. McCarty, "Starlike functions", Proo. Amer. Math. Soc. 43

(19710, 361-366.

[£] M.L. Mogra, "On a class of starlike functions in the unit disc I",
J. Indian Math. Soo. (to appear).

[92 Zeev Nehari , Conformal mapping (McGraw-Hill, New York, Toronto,

London, 1952).

[70] K.S. Padmanabhan, "On certain classes of starlike functions in the

unit disk", J. Indian Math. Soc. (U.S.) 32 (1968), 89-103.

[7 7] Albert Schild, "On starlike functions of order a ", Amer. J. Math.

87 (1965), 65-70.

[72] Herb Silverman, "Univalent functions with negative coefficients",

Proc. Amer. Math. Soc. 51 (1975), 109-116.

[73] Ram Singh, "On a class of star-like functions", Compositio Math. 19

(1968), 78-82.

[74] Ram Singh, "On a class of starlike functions. II", Ganita 19 (1968),
103-110.

Downloaded from https://www.cambridge.org/core. IP address: 18.206.13.133, on 13 Jun 2020 at 11:33:26, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0004972700023492
 Coefficient estimates 425

[7 5] D.J. Wright, "On a class of starlike functions", Compositio Math. 21

(1969), 122-12U.

Department of Mathematics,
Indian Institute of Technology Kanpur,
Kanpur,
India.

Downloaded from https://www.cambridge.org/core. IP address: 18.206.13.133, on 13 Jun 2020 at 11:33:26, subject to the Cambridge Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0004972700023492