Result.Math.

48 (200S) 310-32S
1422-6383/0S/040310-16
DorIO.1007/s0002S-00S-0203-z Results in Mathematics
© Birkhiiuser Verlag, Basel, 200S

POSITIVE PERIODIC SOLUTIONS OF SYSTEMS OF FIRST ORDER ORDINARY

DIFFERENTIAL EQUATIONS

Donal O'Regan and Haiyan Wang*

Abstract

Consider the n-dimensional nonautonomous system x(t) = A(t)G(x(t)) - B(t)F(x(t - r(t))) Let

U = (Ul, ... ,Un ), fJ = limllull_o~, f~ = limllull_oo~, i = 1, ... ,n, F = (f', ... ,r), Fo =
max.=l, ... ,nUJ} and Foo = max'=l, ... ,nU~}. Under some quite general conditions, we prove that ei-

ther Fo = 0 and F = 00, or Fo = 00 and F = 0, guarantee the existence of positive periodic solutions
00 00

for the system for all A > O. Furthermore, we show t.hat Fo = F = 0, or Fo = F = 00 guarantee the
00 00

multiplicity of positive periodic solutions for the system for sufficiently large, or small A, respectively.

We also establish the nonexistence of the system when either Fo and F 00 > 0, or Fo and F 00 < 00 for
sufficiently large, or small A, respectively. We shall use fixed point theorems in a cone.

Keywords: positive periodic solutions, existence, fixed point theorem.

1 Introduction

The existence of periodic solutions of the equation of the form

x'(t) = a(t)g(x(t» - >"b(t)f(x(t - ret»~). (Ll)

and its generalizations have attracted much attention in the literature. See, e.g., Chow [1], Hadeler and

Tomiuk [5], Kuang [8, 9J, Kuang and Smith [lOJ, Tang and Kuang [12J. The equation of the form (1.1) has
2000 Mathematics Subject Classification. Primary 34K13
• Corresponding author
O'Regan and Wang 311

been proposed as models for a variety of population dynamics and physiological processes such as production

of blood cells, respiration, and cardiac arrhythmias, See, for example, the above references, and [4, 11, 16].

Motivated by multiple-species ecological models, it is natural to explore nonautonomous n-dimensional

systems. Nonautonomous systems are more realistic since real-world models often require us to incorporate

temporal inhomogeneity in the models. One of the methods of incorporating temporal nonuniformity of the

environments in models is to assume that the parameters are periodic with the same period of the time

variable. In this paper, we shall study the existence of positive w-periodic solutions for the nonautonomous

n-dimensional system

x(t) = A(t)G(x(t)) - >.B(t)F(x(t - T(t))), (1.2)

where A(t) = diag[al (t), a2(t), ... ,an(t)], B(t) = diag[bl (t), b2(t), . .. ,bn(t)], F(x) = [fl (x), F(x), ... ,r(xW,
G(x) = [gl(X),g2(X), ... ,gn(x)]T and>' > °is a positive parameter.
In(1.2), we assume that

(HI) ai, bi E C(IR, [0,00)) are w-periodic functions such that J;;' ai(t)dt > 0, J;;' bi(t)dt > 0, i = 1, ... , n.

T E C(IR, JR.) is an w-periodic function.

(H2) fi: JR.+ --> [0,00) is continuous, gi : JR.+ --> °

[I, L], < 1 < L < 00 is continuous, i = 1, ... ,n.

(H3) fi(U) °
> for Ilull > 0, i = 1, ... , n.

In order to state our theorems, we introduce some notation. Let

fi = lim fi(u) I 00= l' Ji(u) 1

o Iiuli-+o Ilull '
i
1m
Iiuli-+oo
-II-II
U
,uE,,+,
nn'
l = , ... ,n
(1.3)
Fo =. max
t.=l, ... ,n
un, Foo = . max u~}·
t.=l, ... ,n

A solution u(t) = (UI(t), ... , un(t)) is positive if, for each i = 1, ... , n, Ui(t) ~ °for all t E IR and there is

at least one component which is positive on lR..

Our main results are:

312 O'Regan and Wang

Theorem 1.1 Assume (H1)-(H2) hold.

(a). If Fo = 0 and F = 00, then for all >. > 0 (1.2) has a positive w-periodicsolution.
00

(b). If Fo = 00 and F = 0, then for all >. > 0 (1.2) has a positive w-periodic solution.
00

Theorem 1.2 Assume (H1)-(H3) hold.

(a). If Fo = 0 or F = 0, then there exists a >'0 > 0 such that (1.2) has a positive w-periodic solution for
00

>. > >'0'

(b). If Fo = 00 or F = 00, then there exists a >'0 > (} such that (1.2) has a positive w-periodic solution for
00

o < >. < >'0.

(c). If Fo = F 00 = 0, then there exists a >'0 > 0 such that (1.2) has two positive w-periodic solutions for

>. > >'0'

(d). If Fo = F 00 = 00, then there exists a >'0 > 0 such that (1.2) has two positive w-periodic solutions for

o < A < AO.

(e). If Fo < 00 and F 00 < 00, then there exists a >'0 > 0 such that for all 0 < >. < >'0 (1.2) has no positive

w-periodic solution.

(f). If Fo > 0 and F 00 > 0, then there exists a >'0 > 0 such that for all >. > >'0 (1.2) has no positive w-periodic

solution.

For n = 1, the existence, multiplicity and nonexistence of positive w-periodic solution of (1.2) with a

parameter>' was discussed in Wang [14]. Jiang, Wei and Zhang [6] obtained some existence results for the

case when when gi == 1, i = 1, ... , n. In a recent paper, Wang, Kuang and Fen [15] proved multiplicity and

nonexistence results for a similar equation when gi == 1, i = 1, ... , n.

This paper is organized in the following ways. In Section 2, we transform (1.2) into a system of integral

equations, and then to a fixed point problem of an equivalent operator in a cone. FUrther, we establish two

inequalities which allow us to estimate the operator. In Section 3, we apply the fixed point index to show the

existence, multiplicity and nonexistence of positive w-periodic solutions of (1.2) based on the inequalities.
O'Regan and Wang 313

2 Preliminaries

In this section, we recall some concepts and conclusions on the fixed point index in a cone in [2, 3, 7J. Let E

be a Banach space and K be a closed, nonempty subset of E. K is said to be a cone if (i) au + {3v E K for

all u, v E K and all a, {3 ;:: 0 and (ii) u, -u E K imply u = O. Assume 0 is a bounded open subset in E with

the boundary 80, and let T : K n n -> K is completely continuous such that Tx #x for x E 00 n K, then

the fixed point index itT, K n 0, K) is defined. If itT, K n 0, K) # 0, then T has a fixed point in K n O.

The following well-known result on the fixed point index is crucial in our arguments.

Lemma 2.1 ([2,3,7]). Let E be a Banach space and K a cone in E. Forr > 0, define Kr = {u E K : IIxll <

r}. Assume that T : K r -> K is completely continuous such that Tx ~ x for x E oKr = {u E K : Ilxll = r}.

(i) If IITxl1 ;:: Ilxll for x E 8Kro then itT, K" K) = O.

(ii) If IITxl1 5: Ilxll for x E 8Kro then itT, K" K) = 1.

In order to apply Lemma 2.1 to (1.2), let X be the Banach space defined by

X = {u(t) E C(lR, lRn) : u(t + w) = u(t), t E lR}

n
with a norm lIuli = L
;=1 tE[O,wj
sup lu;(t)l, for u = (Ul' ... , un) EX. For u E X or lRf., Ilull denotes the norm of u
in X or lRf., respectively.

Define

K = {u = (Ul, ... ,un) EX: Ui(t) ;:: O'f(1 - ~D sup IUi(t)l, i = 1, ... , n, t E [O,wJ,}
1 - O'i tE[O,wj

where O'i = e- J; a,(t)dt, i = 1, ... , n. It is clear K is cone in X.

For r > 0, define Or = {u E K: Ilull < r}. It is clear that 80r = {u E K: Ilull = r}. Let T,\ : K -> X

be a map with components (Tl, ... , 11'):

T,{u(t) = r
Ait
Hw
Gi(t, s)bi(s)fi(u(s - r(s)))ds, i = 1, .. . ,n, (2.1)

where
e- It' a,(8)g'(u(8))d8
Gi(t, s) = 1 _ e- I; a,(8)g'(u(8))d8
314 O'Regan and Wang

Note that
aL 1
-1• L ~Gi(t,s)~-I--I' t~s~t+w,i=I, ... ,n.
-ui -O"i

Lemma 2.2 Assume (Hl)-(H2) hold. Then TA(K) c K and TA : K ---> K is continuous and completely

continuous.

PROOF In view of the definition of K, for u E K, we have, i = 1, ... , n,

t+
= AI
2w
(T{u)(t + w) Gi(t + W, s)bi(s)Ji(u(S - r(s)))ds
t+w
=A I
t
t+w
Gi(t + w, 0 + w)bi(O + w)fi(u(O + w - r(O + w)))dO

= AI t+
t
w
Gi(t, s)bi(s)fi(u(S - r(s)))ds

= (l1 u )(t).

It is easy to see that ftt+w bi(s)fi(u(s-r(s)))ds is a constant because ofthe periodicity of b;(t)J'(u(t-r(t))).

Notice that, for u E K and t E [D,w], i = 1, ... , n,

l1u(t) 2: ~A
aL It+w bi(s)J'(u(s - r(s)))ds
1- a; t

= ar LA r bi(s)fi(U(S - r(s)))ds
10
l
1 - a;

' - L'
=aI,(1 a!) 1 w .
IA b;(s)f'(u(s - r(s)))ds
1 - ai 1 - ai 0

2: ar(1 - ~D sup IT{u(t)l.

1 - ai tE[O,w]

Thus TA(K) c K and it is easy to show that TA : K -+ K is continuous and completely continuous. 0

Lemma 2.3 Assume that (Hl)-(H2) hold. Then u E K is a positive periodic solution of (1.2) if and only if

it is a fixed point of TA in K.
O'Regan and Wang 315

PROOF If u = (UJ,"" Un) E K and T>.u = u, then, for i = 1, ... ,n,

d It+w
u;(t) = diY' Gi(t, s)bi(s)t(u(s - T(s»)ds)
t

= AGi(t, t + w)bi(t + w)fi(u(t + w - T(t + w» - AGi(t, t)bi(t)fi(U(t - T(t»)

+ ai(t)gi(u(t»T{u(t)
= A[Gi(t, t + w) - Gi(t, t)]bi(t)t(u(t - T(t») + ai(t)gi(u(t»T{u(t)
= ai(t)l(u(t»Ui(t) - Abi(t)fi(u(t - T(t»).

Thus u is a positive w-periodic solution of (1.2). On the other hand, if u = (UJ,"" un) is a positive

w-periodic function of (1.2), then Abi(t)fi(u(t - T(t») = a,(t)gi(u(t»Ui(t) - ui(t) and

T{u(t) = AI t+ t
W

Gi(t,s)bi(s)fi(U(S - T(s»)ds

= I t+t
w
Gi(t, s)(ai(s)gi(u(s»Ui(S) - u;(s»ds

= I t+w Gi(t, s)ai(s)gi(u(s»Ui(S)ds - It+w Gi(t, s)u:(s)ds

t t

t+w
= I Gi(t, s)ai(s)gi(u(s»u,(s)ds - Gi(t, s)u,(s)11+
t w

I w Gi(t, s)ai(s)gi(u(s»Ui(S)ds
-
t+
t

= Ui(t).

Thus, T>.u = u, Furthermore, in view of the proof of Lemma 2.2, we also have Ui(t) 2': C7tJ~~fl) SUPtE[O,w) Ui(t)
for t E [D,w]. That is, u is a fixed point of T>. in K. 0

Define r = mini=J, ... ,n{l~tf Jaw b,(s)ds}mini=l, ... ,n{C7tJ~~fl)} > Dand we have the following lemma.

Lemma 2.4 Assume that (H1)-(H2) hold. For any TJ > D and u = (uJ,"" un) E K, if there exists a

component fi of F such that fi(u(t» 2': ~j=J Uj(t)TJ for t E [D,w], then
316 O'Regan and Wang

PROOF Since u E K and ji(U(t)) 2: E7=1 Uj(t)T} for t E [O,wj, we have

For each i = 1, .. , n, let ji(t) : IR+ -> IR+ be the function given by

p(t) = max{fi(u) : u E IR~ and Ilull :::; t}.

Lemma 2.5 ([13]) Assume (H2) holds. Then jJ = fJ and jfx, = ffx" i = 1, ... , n.

Lemma 2.6 Assume (H1)-(H2) hold and let r > o. If there exits an c > 0 such that

p(r):::; cr, i = 1, ... ,n,

then

where C- = Ei=1
n
R1 J.w0 bi(s)ds.

PROOF From the definition of T A, for u E anT! we have

:::;
n 1 r
~ 1 - 0"1 A Jo b;(s)ji(r)ds

1 l
: :; L ---=---r
n
1
i=1 0"; 0
w
b;(s)dsAcllull = ACcllull·
o

The following two lemmas are weak forms of Lemmas 2.4 and 2.6.
O'Regan and Wang 317

Lemma 2.7 Assume (HI )-(H3) hold. If u E ann r > 0, then

IIT>.ull ~ Amr i=~~,)I-'uf

uL r
Jo bi(s)ds}

where mr = min{fi(u): u E lR+ and ur::; Ilull ::; r, i = 1, ... ,n.} > 0, and u = rnini=1, ... ,n{0'\~;;I)}.

PROOF Note r = Ilull = L:~=1 sup[O,w] IUi(t)1 ~ L:~=1 inf[o,w] IUi(t)1 ~ u L:~=1 sup [O,w] IUi(t)1 = ur. Thus

fi(U(t)) ~ mr for t'E [O,w], i = 1, ... ,n. A slight modification of the proof in Lemma 2.4 yields the result.

Lemma 2.8 Assume (Hl)-(H3) hold. If u E ann r > 0, then

where Mr = max{fi(u) : u E lR+ and Ilull ::; r, i = 1, ... , n} > 0 and C is the positive constant defined in
Lemma 2.6

PROOF Since Ji(u(t)) ::; Mr for t E [O,w], i = 1, ... , n, a slight modification of the proof in Lemma 2.6

guarantees the result. 0

3 Proof of Theorem 1.1

PROOF Part (a). F o = 0 implies that fj = 0, i = 1, ... , n. It follows from Lemma 2.5 that jj = 0,
i = 1, ... , n. Therefore, we can choose r1 > 0 so that ji(rd ::; Cr1, i = 1, ... , n, where the constant c > 0
satisfies

AcC < 1,

and C is the positive constant defined in Lemma 2.6. We have by Lemma 2.6 that

IIT>.ull ::; AcC11u11 < Ilull for u E anr,.

Now, since F 00 = 00, there exists a component fi of F such that f;'" = 00. Therefore, there is an if > 0
such that
318 O'Regan and Wang

for u= (Ul' ... , un) E lR+ and Ilull ~ iI , where 1/ > 0 is chosen so that

Let r2 = max{2rl' ~iI}, where a = mini=I •...• n{'Ttl~~f:)}. If u = (Ul' ... ,un) E anr" then
n
min '" Ui(t) ~ allull =
a<t<w~
ar2 >
-
iI,
- - i=l

which implies that

n
fi(U(t)) ~ 1/ ~ Ui(t) for t E [O,w].
i=l

It follows from Lemma 2.4 that

By Lemma 2.1,

It follows from the additivity of the fixed point index that

Thus, i(T>.,n r , \ nr"K) i- 0, which implies T>. has a fixed point u E n r , \ n

r, by the existence property

of the fixed point index. The fixed point u E nr , \ nr1 is the desired positive solution of (1.2).

Part (b). If Fa = 00, there exists a component Ji such that fj = 00. Therefore, there is an rl > 0 such

that

for u = (Ul' ... , Un) E lR+ and lIull ~ rl, where 1/ > 0 is chosen so that

n
fi(u(t)) ~ 1/ ~ Ui(t), for t E [0, w].
i=1
O'Regan and Wang 319

Lemma 2.4 implies that

We now determine nr,. Foe = 0 implies that f':x, = 0, i = 1, ... ,n. It follows from Lemma 2.5 that j':x, = 0,
i = 1, ... , n. Therefore there is an r2 > 2rl such that

where the constant c > 0 satisfies

>..cC < 1,

and C is the positive constant defined in Lemma 2.6. Thus, we have by Lemma 2.6 that

By Lemma 2.1,

i(T)"nr"K) = 0 and i(T)"nr"K) = 1.

It follows from the additivity of the fixed point index that i(T)" nr, \ Or" K) = 1. Thus, T), has a fixed

point in nr2 \ Or" which is the desired positive solution of (1.2). 0

4 Proof of Theorem 1.2

PROOF Part (a). Fix a number rl > O. Lemma 2.7 implies that there exists a >"0 > 0 such that

IIT),ull> Ilull, for .u E an,,,>,, > >"0.

If Fa = 0, then f~ = 0, i = 1, ... , n. It follows from Lemma 2.5 that

jJ = 0, i = 1, ... , n.

Therefore, we can choose 0 < r2 < rl so that

320 O'Regan and Wang

where the constant c > 0 satisfies

AcC < 1,

and C is the positive constant defined in Lemma 2.6. We have by Lemma 2.6 that

If F 00 = 0, then 1/:0 = 0, i = 1, ... , n. It follows from Lemma 2.5 that j/:o = 0, i = 1, ... , n. Therefore there is
an T3 > 2Tl such that

where the constant c > 0 satisfies

AcC < 1,

and C is the positive constant defined in Lemma 2.6. Thus, we have by Lemma 2.6 that

IIT),ull :s; AcCliul1 < Ilull for u E ao r3 ·

It follows from Lemma 2.1 that

i(T)"Or"K) = 0, i(T)"Or.,K) =1 and i(T)"Or3,K) = 1.

Thus i(T)" Or, \ nr" K) = -1 and i(T)"Or3 \ nr" K) = 1. Hence, T), has a fixed point in Or, \ nr• or
OT3 \ nT, according to Fo = 0 or F 00 = 0, respectively. Consequently, (1.2) has a positive solution for A > AO'

Part (b). Fix a number Tl > O. Lemma 2.8 implies that there exists a >'0> 0 such that

IIT),ull < Ilull, for u E ao T" 0 < A < Ao.

If Fo = 00, there exists a component Ii of F such that IJ = 00. Therefore, there is a positive number T2 < Tl
such that

for u = (Ul, ... , un) E lR'i- and Ilull :s; T2, where T} > 0 is chosen so that

ArT} > 1.
O'Regan and Wang 321

Then
n
fi(U(t)) ~ TJ L Ui(t),
i=l

for u = (U1' ... , Un) E anr>! t E [0, w]. Lemma 2.4 implies that

If F 00 = 00, there exists a component fi of F such that f!x, = 00. Therefore, there is an iI > 0 such that

for u= (U1' ... , un) E lR+. and Ilull ~ iI , where TJ > 0 is chosen so that

Let 1'3 = max{2T1'~}, where a = mini=1, ... ,n{ut1~;!:)}' If u = (U1' ... ,Un) E an r., then

n
min "Ui(t) ~ allull
o<t<w~ .
= aT3 ~ iI,
- - i=l

which implies that

n
fi(u(t)) ~ TJ L Ui(t) for t E [O,w].
i=l

It follows from Lemma 2.4 that

It follows from Lemma 2.1 that

or nr • \ nr , according to Fo = 00 or F 00 = 00, respectively. Consequently, (1.2) has a positive solution for

o < A < AO.

Part (c). Fix two numbers 0 < 1'3 < 1'4. Lemma 2.7 implies that there exists a AO > 0 such that we have,
for A > AO,

IIT>.ull > Ilull, for u E an r " (i = 3,4).

322 O'Regan and Wang

Since Fo = 0 and F 00 = 0, it follows from the proof of Theorem 1.2 (a) that we can choose 0 < rl < r3/2
and rz > 2r4 such that

IITAul1 < Ilull, for u E anr" (i = 1,2).

It follows from Lemma 2.1 that

and

i(TA,nr.,K) =0, i(TA,nr.,K) =0

and hence, i(TA,nr • \nr"K) = -1 and i(TA,nr , \nr.,K) = 1. Thus, TA has two fixed points Ul(t) and

U2(t) such that Ul(t) E nr • \ nr} and uz(t) E nr , \ nr. , which are the desired distinct positive periodic
solutions of (1.2) for A > Ao satisfying

Part (d). Fix two numbers 0 < r3 < r4. Lemma 2.8 implies that there exists a Ao > 0 such that we have,
for 0 < A < AO,

IITAul1 < lIull, for u E anr ;! (i = 3,4).

Since Fo = 00 and F = 00, it follows from the proof of Theorem

00 1.2 (b) that we can choose 0 < rl < r3/2

and r2 > 2r4 such that

It follows from Lemma 2.1 that

and

i(TA,nr.,K) = 1, i(TA,nr.,K) = 1

and hence, i(T A, nr • \ nr " K) = 1 and i(TA, n r , \ nr • , K) = -1. Thus, T A has two fixed points Ul (t) and

uz(t) such that Ul(t) E nr • \ nr } and U2(t) E nr , \ nr • , which are the desired distinct positive periodic
O'Regan and Wang 323

solutions of (1.2) for 0 < A < AO satisfying

Part (e). Since Fo < 00 and F 00 < 00, then f~ < 00 and f~ < 00, i = 1, ... , n. It is easy to show (see

[13]) that there exists an e > 0 such that

fi(U) :s:; ellull for u E R+, i = 1, ... , n.

Assume v(t) is a positive solution of (1.2) . We will show that this leads to a contradiction for 0 < A < AO,

where
1
AO = n 1 rw .
2:i=1 R Jo bi(s)dse
In fact, for 0 < A < AO, since TAV(t) = v(t) for t E [O,w], we find

IIvll IIT>.vll
n
"" max T{v(t)
LO<t<w
i=l - -

:s:; t;
n 1
1 - O"j 10
{w
bi(s)dSAellvll

< Ilvll,

which is a contradiction.

Part (f). Since Fo > 0 and F 00 > 0, there exist two components fi and fi of F such that f& > 0 and

fix, > O. It is easy to show (see [13]) that there exist positive numbers 1}, TI such that

(4.2)

and

(4.3)

u!'(l-u!)
here 0" = mini=I, ... ,n{~} . Assume v(t) = (VI, ... , v n) is a positive solution of (1.2). We will show that

this leads to a contradiction for A > AO = f,;. In fact, if Ilvll :s:; TI, (4.2) implies that
n
fi(v(t)) ~ 1} L Vi(t) , for t E [O,wj.
i=l
324 O'Regan and Wang

On the other hand, if Ilvll > 1'1, then

n
min ' " Vi(t) ~
o<t<wL...
CTllvll > CTT1,
- - i=l

which, together with (4.3), implies that

n
fj(v(t)) ~ 1] Lv;(t), for t E [O,wl·
i=l

Since T>. v(t) = v(t) for t E [0, wj, it follows from Lemma 2.4 that, for A > AO,

Ilvll IIT>.vll

which is a contradiction. 0

References

[1) S-N. Chow, Existence of periodic solutions of autonomous functional differential equations, J. Differential Equa-

tions 15(1974), 350-378.

[2) K, Deimling, "Nonlinear Functional Analysis," Springer, Berlin, 1985.

[3) D. Guo and V. Lakshmikantham, "Nonlinear Problems in Abstract Cones," Academic Press, Orlando, FL, 1988.

[4) W.S. Gurney, S.P. Blythe and R.N. Nisbet, Nicholson's blowflies revisited, Nature, 287(1980), 17-21.

[5) K. P. Hadeler and J. Tomiuk, Periodic solutions of difference differential equations, Arch. Rat. Mech. Anal. 65

(1977), 87-95.

[6) D. Jiang, J. Wei and B. Zhang, Positive periodic solutions of functional differential equations and population

models. Electron. J. Differential Equations 2002, No. 71, 1-13.

[7) M. Krasnoselskii, Positive solutions of operator equations, Noordhoff, Groningen, 1964.

[8) Y. Kuang, Delay Differential Equations with Applications in Population Dynamics, Academic Press, New York,

1993.
O·Regan and Wang 325

[9J Y. Kuang, Global attractivity and periodic solutions in delay-differential equations related to models in physi-

ology and population biology. Japan J. Indust. App!. Math. 9 (1992) , 205-238.

[lOJ Y. Kuang and H. L. Smith, Periodic solutions of differential delay equations with threshold-type delays, Oscil-

lations and Dynamics in Delay Equations, Contemp. Math. 129 (1992), 153-176.

[l1J M.C. Mackey and L. Glass, Oscillations and chaos in physiological control systems, Science, 197 ( 1997), 287-289.

[12J B. Tang and Y. Kuang, Existence, uniqueness and asymptotic stability of periodic solutions of periodic functional-

differential systems. Tohoku Math. J . (2) 49 (1997), 217-239

[13J H. Wang, On the number of positive solutions of nonlinear systems, J. Math. Ana!. App!. , 281(2003) 287-306.

[14J H. Wang, Positive periodic solutions of functional differential equations, Journal of Differential Equations 202

(2004), 354-366.

[15J H. Wang, Y. Kuang, M. Fen, Periodic Solutions of Systems of Delay Differential Equations.

[16J M. Wazewska-Czyzewska and A. Lasota, Mathematical problems of the dynamics of a system of red blood cells.

(Polish) Mat. Stos. 6 (1976), 23-40 .

Department of Mathematics

National University of Ireland

Galway, Ireland

Email address: donal.oregan@nuigalway.ie

Department of Mathematical Sciences and Applied Computing

Arizona State University

Phoenix, AZ 85069-7100, U.S .A.

E-mail address: wangh@asu.edu

Eingegangen am 4. Juni 2005