Normal Forms and Bifurcation of Planar Vector Fields 1st Edition Shui-Nee Chow 2024 scribd download

Download the full version of the ebook at
https://ebookultra.com
Normal Forms and Bifurcation of Planar Vector

Fields 1st Edition Shui-Nee Chow
https://ebookultra.com/download/normal-forms-and-
bifurcation-of-planar-vector-fields-1st-edition-
shui-nee-chow/
Explore and download more ebook at https://ebookultra.com

Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.
The Method of Normal Forms Second Edition Ali Hasan Nayfeh
https://ebookultra.com/download/the-method-of-normal-forms-second-
edition-ali-hasan-nayfeh/
ebookultra.com
Current Status of Carotid Bifurcation Angioplasty and

Stenting 1st Edition Veith Amor
https://ebookultra.com/download/current-status-of-carotid-bifurcation-
angioplasty-and-stenting-1st-edition-veith-amor/
ebookultra.com
Vector Calculus 1st Edition R. Gupta
https://ebookultra.com/download/vector-calculus-1st-edition-r-gupta/
ebookultra.com
Routledge Handbook of the Chinese Economy 1st Edition

Gregory C. Chow
https://ebookultra.com/download/routledge-handbook-of-the-chinese-
economy-1st-edition-gregory-c-chow/
ebookultra.com
Therapeutic Delivery Solutions 1st Edition Chung Chow Chan
https://ebookultra.com/download/therapeutic-delivery-solutions-1st-
edition-chung-chow-chan/
ebookultra.com
Nonlinear solid mechanics bifurcation theory and material

instability 1st Edition Davide Bigoni
https://ebookultra.com/download/nonlinear-solid-mechanics-bifurcation-
theory-and-material-instability-1st-edition-davide-bigoni/
ebookultra.com
Stochastic Partial Differential Equations 2nd Edition Chow
https://ebookultra.com/download/stochastic-partial-differential-
equations-2nd-edition-chow/
ebookultra.com
Representing Wars from 1860 to the Present Fields of

Action Fields of Vision 1st Edition Claire Bowen
https://ebookultra.com/download/representing-wars-from-1860-to-the-
present-fields-of-action-fields-of-vision-1st-edition-claire-bowen/
ebookultra.com
Bifurcation and Chaos in Complex Systems 1st Edition J.Q.

Sun And A.C.J. Luo (Eds.)
https://ebookultra.com/download/bifurcation-and-chaos-in-complex-
systems-1st-edition-j-q-sun-and-a-c-j-luo-eds/
ebookultra.com
Normal Forms and Bifurcation of Planar Vector Fields
1st Edition Shui-Nee Chow Digital Instant Download
Author(s): Shui-Nee Chow, Chengzhi Li, Duo Wang
ISBN(s): 9780521102230, 0521102235
Edition: 1
File Details: PDF, 3.15 MB
Year: 2009
Language: english
NORMAL FORMS AND BIFURCATION
OF PLANAR VECTOR FIELDS
NORMAL FORMS AND
BIFURCATION OF
PLANAR VECTOR
FIELDS
SHUI-NEE CHOW
Georgia Institute of Technology
CHENGZHI LI
Peking University
DUO WANG
Tsinghua University
AMBRIDGE
UNIVERSITY PRESS
CAMBRIDGE UNIVERSITY PRESS
Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, Sao Paulo
Cambridge University Press

The Edinburgh Building, Cambridge CB2 8RU, UK
Published in the United States of America by Cambridge University Press, New York
www.cambridge.org
Information on this title: www.cambridge.org/9780521372268
© Cambridge University Press 1994
This publication is in copyright. Subject to statutory exception

and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published 1994
A catalogue record for this publication is available from the British Library
Library of Congress Cataloguing in Publication data
Chow, Shui-Nee.
Normal forms and bifurcation of planar vector fields / Shui-Nee
Chow, Chengzhi Li, Duo Wang.
p. cm.
Includes bibliographical references and index.
ISBN 0-521-37226-7
1. Bifurcation theory. 2. Vector fields. 3. Normal forms
(Mathematics) I. Li, Chengzhi. II. Wang, Duo. III. Title.
QA372.C548 1994
515'.35-dc20 93-26510 CIP
ISBN 978-0-521-37226-8 hardback
Transferred to digital printing 2007

Contents
Preface page vii

Chapter 1 Center Manifolds 1
1.1 Existence and Uniqueness of Global Center
Manifolds 3
1.2 Smoothness of the Global Center Manifolds 12
1.3 Local Center Manifolds 27
1.4 Asymptotic Behavior and Invariant
Foliations 35
1.5 Bibliographical Notes 47
Chapter 2 Normal Forms 49
2.1 Normal Forms for Differential Equations
near a Critical Point 49
2.2 Poincare's Theorem and Siegel's Theorem 70
2.3 Normal Forms of Equations with Periodic
Coefficients 89
2.4 Normal Forms of Maps near a Fixed Point 98
2.5 Normal Forms of Equations with Symmetry 105
2.6 Normal Forms of Linear Hamiltonian
Systems 113
2.7 Normal Forms of Nonlinear Hamiltonian
Systems 122
2.8 Takens's Theorem 142
2.9 Versal Deformations of Matrices 148
2.10 Versal Deformations of Infinitesimally
Symplectic Matrices 162
2.11 Normal Forms with Codimension One
or Two 177
v
vi Contents
Chapter 3 Codimension One Bifurcations 191
3.1 Definitions and Jet Transversality Theorem 192
3.2 Bifurcation of Equilibria 197
3.3 Bifurcation of Homoclinic Orbits 213
Chapter 4 Codimension Two Bifurcations 228
4.1 Double Zero Eigenvalue 229
4.2 Double Zero Eigenvalue with Symmetry
of Order 2 259
of Order 3 278
of Order 4 292
of Order >_ 5 328
4.6 A Purely Imaginary Pair of Eigenvalues
and a Simple Zero Eigenvalue 335
4.7 Two Purely Imaginary Pairs of Eigenvalues 353
Chapter 5 Bifurcations with Codimension Higher
than Two 383
5.1 Hopf Bifurcation of Higher Order 383
5.2 Homoclinic Bifurcation of Higher Order 393
5.3 A Codimension 3 Bifurcation:
Cusp of Order 3 409
5.4 A Codimension 4 Bifurcation:
Cusp of Order 4 430
Bibliography 452
Index 469
Preface
The theory of bifurcation of vector fields is the study of a family of

equations that are close to a given equation. For example, the family of
equations could be a system of vector fields depending on several
parameters. An important problem is to understand how the topologi-
cal structure of the flow generated by the family of vector fields changes
qualitatively as parameters are varied. The main purpose of this book is
to present some methods and results of the theory of bifurcations of
planar vector fields.
Since simplifying equations is often a necessary first step in many
bifurcation problems, we introduce the theory of center manifolds and
the theory of normal forms. Center manifold theory is important for the
reduction of equations to ones of lower dimension, and normal-form
theory gives a tool for simplifying the forms of equations to the ones
with the simplest possible higher-order terms near their equilibria. We
introduce vernal deformations of vector fields and define the codimen-
sion of a bifurcation of vector fields. This is illustrated by saddle-node
and Hopf bifurcations. We discuss in detail all known codimension-
two bifurcations of planar vector fields. Some special cases of higher-
codimension bifurcations are also considered.
In Chapter 1, we introduce briefly the basic concepts of center
manifolds. We show the existence, uniqueness, and smoothness of
global center manifolds. The existence, asymptotic behavior, and folia-
tion of local center manifolds are also discussed.
In Chapter 2, we present the theory of normal forms. We first discuss
in detail normal forms of vector fields near their equilibria. We intro-
duce two methods for computing normal forms: the matrix representa-
tion method and the method of adjoints. We also introduce normal
forms of equations with periodic coefficients or with symmetries. Nor-
mal forms of diffeomorphisms and Hamiltonian systems are discussed.
vii
viii Preface
Complete proofs of Poincare and Siegel linearization theorems are
presented. Takens's Theorem gives a relation between diffeomorphisms
near fixed points and the time-one maps of flows of vector fields near
equilibria. We introduce also versal deformations of matrices and of
infinitesimally symplectic matrices and normal forms of vector fields of
codimension one and two.
In Chapters 3, 4, and 5, we discuss bifurcation problems of vector
fields with some degeneracies. We assume that the problems to be
considered are restricted to local center manifolds and are in their
normal forms up to some order. In Chapter 3, we introduce the
concepts of versal deformations and the codimension of a bifurcation of
vector fields. Bifurcations of codimension one near singularities and
homoclinic orbits are considered. In Chapter 4, we deal with bifurca-
tions of codimension two. For vector fields whose linear parts have
double zero eigenvalues, we consider a nonsymmetrical case and the
cases with 1: q symmetrices (q = 2, 3, 4 and q >_ 5). The case of 1: 4
symmetry is the most difficult and is far from being solved completely.
For the cases in which the linear parts have one zero and one pair of
purely imaginary eigenvalues, or two pairs of purely imaginary eigenval-
ues, we reduce them to planar systems and then give complete bifurca-
tion diagrams. In Chapter 5, we discuss higher-codimension bifurcation
problems, including Hopf and homoclinic bifurcations with any codi-
mension and cusp bifurcations with codimension three and four.
In the last section of each chapter we give briefly the history and
literature of material covered in the chapter. We have tried to make
our references as complete as possible. However, we are sure that many
are missing.
We would like to express our special acknowledgment to Max
Ashkenazi, Freddy Dumortier, Jibin Li, Kening Lu, Robert Roussarie,
Christiane Rousseau, Lan Wen, and Henryk Zoladek. They read all or
part of the original manuscript and made many helpful suggestions
which enabled us to correct some mistakes and make improvements.
The second and the third authors would also like to thank Professor
Zhifen Zhang and Professor Tongren Ding for many helpful discus-
sions. They would also like to thank the Department of Mathematics at
Michigan State University and the Center for Dynamical Systems and
Nonlinear Studies and the School of Mathematics at the Georgia
Institute of Technology for their kind hospitality, since most of the book
was written while they were visiting there.
This work was partially supported by grants from DARPA and NSF
(USA) and from the National Natural Science Foundation of China.
1
Center Manifolds
The main goal of this book is to study some bifurcation phenomena of

vector fields. This is, in general, a complicated problem. As a prelimi-
nary step, it is necessary to simplify the problem as much as possible
without changing the dynamic behavior of the original vector field.
There are two steps for this purpose: to reduce the dimension of the
bifurcation problem by using the center-manifold theory, which will be
introduced in this chapter, and to make the equation as simple as
possible by using normal-form theory which will be discussed in the next
chapter.
We first give some rough ideas about center manifolds. Consider a
differential equation
x = Ax + f(x), (A)f
where x E R", A E Y(W , EW ), f E Ck(W, R") for some k >_ 1, f(0) _
0, and Df (0) = 0.
We write the spectrum Q(A) of A as
o(A) =c Uc U r,
where
v,= (AEo(A)IRe A<0),

Q,= (AEo(A)IReA=0),
Qu=(AEo(A)IRe A>0).
Let E, E, and E,, be the generalized eigenspaces corresponding to v
1
2 Center Manifolds
v,, and o , respectively. Then we have
l8"=E,,®E,®E",
with corresponding projections
ar,: R" -> E Tr,: III" -p E, rru: R" - E.

It is well known that if A is hyperbolic, that is, v, = 0, then the flow
of (A) f in a small neighborhood dZ of the equilibrium point x = 0 is
topologically equivalent to the flow of the linearized equation at x = 0
x = Ax. (A)0
Since x(t) = e`9tx(0) is the solution of (A)0 and a, = 0, any nonzero
solution in E, (or E") tends to the equilibrium x = 0 exponentially as
t -* + oo (or t -' - co). Therefore, the structure of flow in fl is simple;
it is also stable with respect to any small perturbation on the right-hand
side of equation (A)0. See Hartman [1], for example.
However, if o,, 0, then the situation will be different from the
above in two aspects. First, the topological structure for (A) f is not, in
general, the same as for (A)0 any more; this will be shown in a lot of
examples in Chapters 3-5. Second, more complicated structure of the
flow for (A) f may exist on an invariant manifold W `(f ), and the
dimension of W '(f) is equal to the dimension of E,
In fact, if f = 0, then all bounded solutions of (A)0, including all
equilibria and periodic orbits, are contained in the subspace E, which
is invariant under (A)o. So we take W'(0) = E, We will prove that the
aforementioned W `(f) exists for f * 0, it is tangent to E, at x = 0,
and W'(f) contains all solutions of (A)f that stay in fl for all t E If81.
In particular, WC(f) contains all sufficiently small equilibria, periodic
orbits, and homoclinic and heteroclinic orbits. And if a" = 0, then all
solutions of (A)f (in ft) will converge exponentially to some solutions
on WV ) as t -' + c. Therefore, instead of the n-dimensional equa-
tion (A)f, we can consider a lower-dimensional equation on W '(f) for a
bifurcation problem, and WC(f) is called a center manifold. The precise
definition will be given subsequently in Section 1.1.
We will prove the existence, uniqueness, and smoothness of global
center manifolds in Sections 1.1-1.2 under a quite strong condition
which says the Lipschitz constant of f is globally small. In Section 1.3
the cut-off technique is used to get the local center manifolds from the
Existence, Uniqueness of Global Center Manifolds 3
global theory, and the above Lipschitz condition will be satisfied auto-
matically since f (O) = 0 and Df (0) = 0. But a new problem arises: The
local center manifold is not unique. In fact, different cut-off functions
can give different local center manifolds. Hence, it is needed to show
the equivalence (in some sense) between different local center mani-
folds concerning the bifurcation problems. Finally, in Section 1.4 we
discuss the center-stable and center-unstable manifolds, give the asymp-
totic behavior of any solution of (1.1) in R", and describe the invariant
foliation structure.
1.1 Existence and Uniqueness of Global Center Manifolds
Consider the equation

x =Ax + f(x), (1.1)
where x E 1R", A E 5°(IP",18"), f E Ck(R ,18") for some k > 1, f(0) _

0, and Df (0) = 0.
We keep the notations Ee, E, E and are, are, aru throughout this
chapter, and let
Eh = ES ® Eu , arh = are + art .
As usual, we denote by I y I the norm of y in some Banach space. Let

X, Y be Banach spaces and Ck(X, Y) be the set of all Ck mappings
from X into Y. We define the Banach space
Cb(X,Y) _ {w E Ck(X,Y) IIWIIck := max suplDjw(x)I < oo

0<j<k XEX
If X = Y, we write Cbk(X, X) as Cb(X ). We let

IIDwII = supXExIDw(x)I.
Similarly, we define
IIDkw(x) - Dkw(y)II
Cb'1(X,Y) = r w E Cb(X,Y) sup 00,
IIx - yllx <
x,y EX,x 0 y
1
4 Center Manifolds
with norm
II Dkw(x) - Dkw(x)ll
11W11-k,1 = IIwII ck + sun x,yEX,x#y.
IIx - yllx
Finally, we denote by i(t, x) the solution of (1.1) with the initial

condition i(O, x) = x.
Now we state the main result of this section, and will prove it by
using several lemmas.
Theorem 1.1. (i) There is a positive number So which depends only on A

in (1.1) such that if f E Cb.1(l8") and Lip(f) < So, then the set
W`:= (xE08" suplkrhz(t, x)I < 0} (1.2)

teR
is invariant under (1.1) and is a Lipschitz submanifold of 08"; more

precisely, there exists a unique Lipschitz function !i E Cb(EC, Eh) such
that
W`={xc+ii(x,)IxcEEj. (1.3)
(ii) If ¢ E Cb(Ec, Eh), and the set
MO := {xc+4)(xc)IxcEEj (1.4)
is invariant under (1.1), then M. = W` and 0 = 41.
Definition 1.2. W` is called the global center manifold of (1.1).
Remark 1.3. If f E Cb(08"), then we will usually replace the condition

Lip(f) < So by IIDf 11 < So.
Remark 1.4. The uniqueness conclusion (ii) should be understood in

the following sense: If M,, is invariant under (1.1), then ¢ E Cb(EC, Eh)
is determined uniquely. This is not true if we replace the condition
¢ E Cb(EC, Eh) by 0 E C°(E, Eh) unless AI E, is semisimple and f

has compact support (see Sijbrand [1] and Vanderbauwhede [3] for
more details).
Lemma 1.5. For any integer k > 0, there are constants K > 1, a > 0,
and /3 > 0 such that ka < /3, and
leAt,rrcl <Ke"IÌ, t E 18,
leAtirsl <- Ke-Rt, t > 0, (1.5)
leAt7ruI < Kept, t < 0.
Proof. Let
/3 =min{IReAllA E=- E au Uv51} -e,
0<e<a<1
where a and a are sufficiently small. Thus, the existence of K is
obvious by the properties of eAt. 0
Let y satisfy
a<y</3. (1.6)
Define a Banach space by
Cy :_ {x E C°(ll ,QBn) llxll7 := sup e-YItIlx(t)I <

t(=-R
The following lemma gives a different criterion for W`.
Lemma 1.6. Suppose f E Cb.1(Rn) and (1.6) is satisfied. Then

(i)
W` _ (x E Rn 1 1(., x) E CY). (1.7)

6 Center Manifolds
(ii) Consider the integral equation
r
eA(t-T)lTCf(y(T))dT
y(t) = eAt?rcx + f0
+ f teA(t-T)Truf(y(T))dT + f t eA(t-T)7r f(y(r))dr. (1.8)

m - o0
We have
W = {y(0) E IIB" I E C,, and satisfies (1.8) forsomex E V8"}.
(1.9)
Proof. By the variation of constants formula, for to, t E R we have
x(t, x) = eA(t-to)x(to, x) + foeA(t-T)f(x(T, x))dT. (1.10)
Denote by W` the right-hand side of (1.7), and by W` the right-hand

side of (1.9). We will show that W c C W ` c W c c WC.
(a) Suppose x E W c; then by (1.2)
sup e-y't'I'rrhx(t, x)I < supITrhx(t, x)I < oo. (1.11)

teR teR
Taking to = 0 in (1.10) we obtain
ir,x(t, x) = eAtircx + foteA(t-7)7rcf(x(T, x))dT. (1.12)
Using Lemma 1.5 and (1.6), we have from (1.12) that
Ilf llcO I,
I1rTx(t, x)I < Key1tllxl + KII f IIco f tey(t-T)dr s Keyltl(IxI +
0 y
whence
sup e-y"t1Iirrx(t, x)I < 00. (1.13)

tER
It follows from (1.11) and (1.13) that x E W`, and this implies W` c W c.
(b) Suppose now x E If/', that is, x) E C,. From (1.10) we have
irux(t, x) = eA(t-to),rrux(to, x) + feA(t-T)Truf(x(T, x))dr. (1.14)
Fixing t E C$ and to >_ max(t, 0), we obtain from (1.5)
IeA(t-to)rrux(to, x)I < KeO(t-to)I x(to, x)I
< KePt-(P
x) 117 -* 0 as to - +00,
since x) E C. and y < P. Therefore, taking the limit in (1.14) as

to -* +-, we have
'Jrux(t, x) = f x)) dT. (1.15)
Similarly, we can obtain
7rsx(t, x) = f t eA(t-T)Tr5f(x(r, x))dT. (1.16)
Combining (1.12), (1.15), and (1.16), we see that

x x) E W'. Therefore W` c W'.
(c) Suppose yo E W`, that is, there is a function E CY, which
satisfies (1.8) for some x E fly" and y(O) = yo. Then from (1.8)
y(t) = eAtSTrcx + f o e-A7,r f(y(T))dT + f °e-ATrruf(y(T))dr}
+ f teA(t-T)f(y(T))dT
0
= eAtyo + f teA(t-T)f(y(T))dr.
0
Hence y(t) is the solution of (1.1) with initial value y(0) = y0. Using
(1.5) and (1.8), it follows that
Iirruy(t)I <_ KlIf Ilco f e,(t-T)d-r If Ilc& < 00,

t
8 Center Manifolds
and
K
1ir5Y(t)I - a Ilfiico < 00,
since f E Cb(R'). Hence ITrhy(t)I < oo. Thus yo E W`'. This implies
W`c W`.
Now we consider the integral equation defined by (1.8). Let F:

Ec - Cy be defined by
F(i;)(t) = E Ec, (1.17)
and G: Cy - C., be defined by
G(Y('))(t) = f'eA(r_ 1rcY(T)dT + f leA(`-r)TrUY(T)dT
+ f l eA(t-T)TrSY(T)dT. (1.18)
We denote the previous three integrals by Gc(y( ))(t), and

respectively. We will use these notations repeatedly in this
chapter.
Define J: E, x CY -p Cy by
J(6, y) = F(e) + G(f(y('))) (1.19)
Obviously, if 6 E Ec then y = is a fixed point of J(e, ) if and

only if y*(t) is a solution of (1.8) with x = 6.
Lemma 1.7. There is a number So > 0, which depends only on A, such

that, if Lip(f) < So, then for any f E E, J(e, y), defined by (1.19), has
a unique fixed point y = x * ( , 6).
Proof. Note that
j( YI) -J( ,Y2) = G(.f(Y1('))) - G(f(Y2(')))

= G(f(yi(')) -.f(Y2('))), (1.20)
and by (1.5) we have
IGG(f(Y1(.)) - f(y2(.)))(t)l
fteA(t-T)Trc(f(Y1(T))
-f(y2(T)))dr
t
<KLip(f)I foealt TIIYl(T) -Y2(T)IdT
<KLip(f)I f
0
tealt-TIeYIT1(sup
TEQ$
e-vITIIy1(T) -Y2(T)I)dT
erltl
<
y-a KLip(f)IIY1 -Y21I,. (1.21)
Similarly, we have
erltl
IGu(f(y1(.)))(t) - Gu(f(Y2(.)))(t)I < - yK Lip(f )11Y1 - Y211-1,
R
(1.22)
eyltl
IGS(f(y1(.)))(t) - GS(f(y2(.)))(t)I - R- yK Lip(f)11Y1 - Y2II,..
(1.23)
These estimates give
sup e-"I11IG(f(Y1(.)))(t) - G(f(y2(.)))(t)I

tE=R
2
< KI 1 + )Lip(f )IIY1 - Y2111-
-y -a /3-y
We choose
1
y)1-
So
K1 a +
I( llII
10 Center Manifolds
If Lip(f) < So then
Ky 1
a
+ y)Lip(f) < 1, (1.24)
a?
and
1
IIG(f(Y1(-))) - G(f(y2(.)))Ill s 3IIY1 -Y21I,.. (1.25)
Thus, for any 6 E E,, by (1.20), we have
1
Y1) - J(4, Y2)117 s 31IY1 - Y2IIY1 (1.26)
as long as Lip(f) < So.

By the Uniform Contraction Mapping Theorem, has a unique
fixed point y = x*(t, 6) for each 6 E E,.
Lemma 1.8. If Lip(f) < So, then there exists a unique Lipschitz function
41 E Cb(EC, Eh) such that
W`= {xc+4i(xj I xcEEj.
Proof. By Lemmas 1.7 and 1.6, (1.8) has a unique solution x*(t,
x(t, x*(O, a)), for any E E,. By Lemma 1.6,
W` = {x*(0, 6) I lj E Ej.
Note that
x*(0, 6) = J(6, e))(0) _ 6 + e E Ec,
where
440 = fè-ATTruf(x*(T, f o
e-aT,r5f(x*(T, 6))dT.
(1.27)
We need to prove the boundedness and Lipschitz continuity of fi.

From (1.5) it follows that
l'0e-ATVuf(x*(T,6))dTl =I
f E Cb (III"). Similarly,
f° e-AT <00. (1.29)
Hence Vi is bounded.
In (1.22) and (1.23), we take yl(t) = x*(t, ) and y2(t) = x*(t, ),
i E E, and then using the condition (1.24) and letting t = 0, we
obtain
-+G( )I 3 Ilx*(', ) -x*(., )Ily. (1.30)
On the other hand, for , 2 E E,
Ix*(t,) -x*(t,)I
_IJ(,x*(',))(t) -J(,x*(',))(t)I
-I F(6 - )(t)I +I G(f(x*(',e)))(t) - G(f(x*(',)))(t)I
Using (1.5) and (1.25), respectively, we have
11x*(.,) -x*(.,4)I1,
',) -x*(',)IIY,
whence,
IIx*(', ) -x*(., )IIy I

- I
(1.31)
2
12 Center Manifolds
From (1.30) and (1.31) we have finally that
I+G() -
K
2
I -I for 6, 4 E E,. (1.32)
Thus the lemma is proved.
Remark 1.9. By the definition of C.,, we can rewrite (1.31) as
1x*(t, ) -x*(t, )I < (1.33)
for any y E (a, /3), , E E, and all t E R.
Proof of Theorem 1.1. Since x(t1, i(t2, x)) = z(t1 + t2, x), the set W`
defined by (1.2) is invariant under (1.1). The remaining conclusions in
(i) are proved in Lemma 1.8.
Now we prove the uniqueness of ¢ in (ii). Suppose 0 E Cb(EC, Eh)
and M. defined by (1.4) is invariant under (1.1). Then z(t, xc +
4(xc)) E M. for all t E P and any xc E Ec. By the definition of M., it
follows that
7rhx(t, xc + `V(xc)) = 4('1Tcx(t, xc + `r(xc))).
Since 0 E ME, Eh), the boundedness of 0 implies the bounded-

ness of lrhx(t, xc + cb(xc)), and hence, by the definition (1.2), xc +
O(x,) E W` for any x, E E, In Lemma 1.8 we have proved that such a
0 is unique. Hence 4) = i/i, and M. = W.
1.2 Smoothness of the Global Center Manifolds
We have proved the existence and uniqueness of the global center

manifolds W c under the conditions f E Cb , 1(OB") and Lip(f) is suffi-
ciently small. If, in addition, f E Cb (W') for some k >_ 1, then we will
show that W' is smooth. The main result of this section is the following
theorem.
Smoothness of Global Center Manifolds 13
Theorem 2.1. Suppose f E Cb (R") for some k >- 1, f (O) = 0 and Df (0)
= 0. Then there is a number Sk > 0 such that if II Df II < Sk, the unique
global center manifold W` is of class Ck, that is, 4 E Cb(EC, Eh), where
a, is given by (1.27) and is related to W` by (1.3). Moreover, Lip(r(i) <
1, qr(0) = 0 and D+/(0) = 0. Furthermore, if x e W` and ic(t)
arrx(t, x), then zc(t) satisfies the following equation
.zc =Axc +7rcf(xc + O(xc)), x, r= Ec. (2.1)
We will prove this theorem by induction on k, and consider first the

case k = 1.
We remark here that if a > 0, then C. c C,y+o and IIxII,+Q < IIxHI,
Hence, there exists a continuous inclusion from C,, into C,,+,. The
choices of spaces (Cfor different 77 in the following discussion are
very important.
To prove .r E C', by (1.27), we need to prove first that x*(t, ) E C'
with respect to E E, Since x*(t, ) is the unique solution of (1.8), we
have
eAt6 + G(f(x (2.2)

x*(t, 6) = fi)))(t),
where G is defined in (1.18).
Lemma 2.2. Suppose that f E Cb(18"), II Df II < So. Then there exists a
number o > 0 such that the map from Ec to C,,+o, is
differentiable.
Proof. Let
u(t, , ) = x*(t, ) - x*(t, ), f, E E, (2.3)
and
f*(t, , ) =f(x*(t, )) -f(x*(t, )) -Df(x*(t,

(2.4)
14 Center Manifolds
Define
L(u(., , )) =G(Df(x*(., 0)u(',f,4)), (2.5)

and
N(u(., (2.6)
Then we obtain from (2.2) that

(I (2.7)
where I is the identity operator and F is defined in (1.17). Obviously, F
is a bounded linear operator. If we replace (f(y1) - f(y2)) by
Df(x*(t, ))u(t, 4, 4) and replace Lip(f) by II Df II in (1.21)-(1.24), then
instead of (1.25) we can obtain
3IIu(-,6,4) Ill.
This implies the norm of L, as an operator from C,, to C.Y satisfies
1
IILII <- 3
Hence (I - L)-' exists and is bounded, and (2.7) can be written as

(I - L) -'F( - 4) + (I - (2.9)
We will prove that there exists a o, > 0 such that
r+1 =o(I6-4I) as 6-->4. (2.10)
Hence, by the definition of a derivative, the map 6 H 6): E, -

Cy+o is differentiable.
It is obvious that if o > 0 is sufficiently small, and we replace y by
y + v (in some cases later, we need to replace y by y + kff for some
integer k > 0), then (1.6) and (1.24) still hold. We fix such a number o,.
We will prove that for every small e > 0, there exists a µ > 0 such
that if 16 - 41 < µ, , 4 E E, then
sup e-cv+o>I"II N(u(., , ))(t) <E16-41. (2.11)
ten
From (2.6) and (2.4) we have
NS, (2.12)
where
N, = ,))(t),
N. =Gulf*( ,))(t), (2.13)
NS = Gs(f*(',))(t).
We will find an estimate only for N, since it is similar for Na and N.
Choose T > 0 so large that
3K2 E
IIDf Ile aT < - (2.14)
y-a 6
where the constants a and K are the same as in (1.5).

We consider two cases:
(i)Itl<-T.
Without loss of generality, we assume 0 < t < T. By (2.13), (2.4) and
(1.31), we have
INJ = foteA(t-T) 1 c[ f(x*(T, )) -f(x*(T, ))

``
-Df(x*(r,))(x*(r,S) -x*(T,S))Jdr
f teA(t-`)Tr,1 f 1[Df(Ax*(T, ) + (1 -
0 0
-Df(x*(T,4))]dA)(x*(T,f) -x*(r, ))dT
3K 2 e«It'I4
2
- o
IfItle(y-«)T
x(f'IDf(Ax*(T,6) + (1 - A)x*(z,4)) IdA dT

0
3K 2 foTfo'
< eyltll
- I
(1 - A)x*(T,
-Df(x*(r,2))IdAd-r. (2.15)
16 Center Manifolds
Since f E C'(R) and the last integral is taken over a compact region
[0, TI x [0, 1] c R2, there exists a /.tj > 0 such that if
I4 - 41 < i then
sup e-(Y+o)ItllNl 3 16 - I. (2.16)

ItI-T
(ii) Itl > T.
Without loss of generality, we assume t > T, and let Ne = N(1) + N(2)
where
N(i) =
f eA(t-T).c{'*(T,4, )dT, N(2) = fTeA(t-T)Tfcf*(Te )dT.
Similarly to (2.15) and (2.16), there exists a µ2 > 0 such that if (4 -I

< µ2, then
sup e-(Y+O)ItIlN(1)I < I4 - 41. (2.17)

Itl>T
6
Using (2.4) and (1.5) we have
teA(t-')?rc f*(T,
IN(2)1=
f
T
6, )dr
s K f te"(t-7)(21I Df III x* (T, 4) - ,4)1)dT.

T
From (1.33) it follows that
3K
s 2 eYTI -I (2.18)
Hence
IN(2)I - 3K2IIDf 1114 - If

T
teate(Y-«)TdT < 3K21IDf1114 - I y - a evt.
From the above estimate and condition (2.14) we obtain
3K 2 e
sup e-(y+Q)ItIIN(2)) <
y-a
IIDf II e-aTl t -I <
6
I -I. (2.19)
ItI>T
We choose µ, = min(/1, µ2). If 16 - EI < i- then (2.16), (2.17) and

(2.19) give
I - El.
Similarly, we can find µu and µs such that
IINull1+Q <
3
I - 4I when I: - EI < µu,
and
IINSIIy+Q< 31 - I when
Let µ = min(,,, µu, t5). If l - EI < µ and E E, then (2.11)

holds, and hence (2.10) holds. o
The following lemma gives a more general result which will be used
repeatedly in the rest of this section.
Lemma 2.3. Suppose that E is a Euclidean space with norm II ' HE and,
for each y E E, the map y H y) from E to CP for some p E (a,,6)
satisfies
(i) g(t, y) is continuous in (t, y) E Il8 X E;
(ii) Iig(', y)IIP < M for some constant M > 0, where M is independent
of Y.
Then for any E (p,,6) and yo E E, we have
lim 11G(g(', y)) - G(g(', yo)) 11; = 0,

II YoIIE0
where G: CP -> C, is defined in (1.18).
Proof. This lemma can be proved by using the same arguments as in

the proof of Lemma 2.2. For a given small e > 0, we find a T > 0 such
that e/6. Then divide the integrals in G into two parts
G(') and G(2). For the noncompact part I tl >- T, we use condition (ii)
and the continuous inclusion from CP to Q to get y) -
18 Center Manifolds
yo))IIC s E/2; for the compact part I tl < T, we use the uniform
continuity of g(t, y) (condition (i)) to find a µ > 0 such that if
ly - yol < µ then IIG(2)(g(., y) - yo))II s E/2. o
We have proved that under the hypothesis of Lemma 2.2, x*(-, 6), as
a mapping from E, to C'Y+Q, is differentiable.
It is known that
x*(r, 6 + Arl) - x* (t, )

DEx*(t,)rl = lim (2.20)
A->o A
Note that for each E E, 4) is a linear mapping from TAE, =

E, to TX*(.,C)C7± = CY+Q, where TAE, and TX*(. )CY+0. are the tangent
spaces of E, and CY+Q at 6 and 6), respectively. For all E E,
we consider 6) as a mapping from E, to .°(E,, CY+o). We will
show in the next two lemmas that as a mapping from E, to
Y (E,, Cy+3Q), is continuous in 6 E E,
Lemma 2.4. Suppose that f 'E=- C'(l ) and II Df II < So. Then the map
ti E, -+2'(E,, CY+2Q) satisfies the following integral equa-
tion
c(t)rl = eA`rl + Vi EEC, t E ft,

(2.21)
where G: CY+o - CY+2o is defined in (1.18).
Proof. Let 6, rl E E, be fixed, A # 0, and
x*(t,6+Arl)
g(r A) = (2.22)
A
Since limx.o(g(t, A)),+P = Dex*(t, 6)77 exists (Lemma 2.2) and

6) is Lipschitz continuous in l: in Cy norm (see (1.31)), g(t, A) is
continuous in (t, A) E R X R. From (2.2) we have for A # 0,
g(t, A) = eA`rl + A))(t), (2.23)

where
h(t, A)
f(x*(t,6+Ai7)) -f(x
_ (f1Df(ox*(t, 6 + A77) + (1 - 9)x*(t, f))dO)g(t, A).

0
To prove Lemma 2.4, we need to take limits on both sides of (2.23) in

Cy+o, and show that as A -> 0,
G(h(-, 0)) in C,+2.,'
In fact, the continuity of h(t, A) comes from the continuity of g(t, A)

and f E Cb(R"). By (2.22) and (1.31), we have
3K
A)Ily+v <- A) 11, <- 50 2 1771.
Thus the hypotheses of Lemma 2.3 are satisfied.
Lemma 2.5. Suppose f E C'(IP"). Then there exists a number S1 < S0

such that if II Df II < 81, then the map H 6) from Ec to
2'(E, Cy+3Q) is continuous in E Ec.
Proof. We consider as a solution of (2.21), and G as a

mapping from Cy+2o to CY+3v. Then for any q G Ec we have
(Dfx*(.'6) -DDx*(."))7711,+3,,,
I G[(Df(x*(-,s)) - Df(x*(., )))D x*(' )" Iy+3Q
(2.24)
Similarly to obtaining the estimate (1.25) (it comes from (1.21), (1.22),
20 Center Manifolds
and (1.23)), we can find 31 > 0, 31 < 8°, such that if II Df II < 51, then
11
G[Df(x*( 6))(Dx*(., ) - DDx*(., jjy+3o
3 II
(Dgx*(', 6)
- DIx*(', ))'1, Iy+3,' (2.25)
Noting Df(x*(t, 6)) is continuous in (t, 6) E R X E., and

3K
IIDf11 <31, Dgx*(', )nI y+2v < 2 I ?1I.
we can use the same argument as in Lemmas 2.2 and 2.3 to find µ > 0
such that if 16 - I < µ then
-Df(x*(',2))D6x*(',s)'nh y+3o<e1711. (2.26)
From (2.24), (2.25), and (2.26) we obtain finally
i (D,x*(., ) - D x*(', 0)"711 y+3Q 2 E1i11,
which implies
3
-D,x*( ,2)IY(Ec,cy+3o)<
2e
if l -I < µ.
Lemma 2.6. Suppose f E Cb(II8"), f(0) = 0, and Df(0) = 0. Suppose

II Df I I < 31, where 51 is given by Lemma 2.4. Then 4 E Cb (Ec, Eh),
I0) = 0, D4i(0) = 0, Lip(4i) < 1, where 0 is given by (1.27). Further-
more, if 2 E W` and zc(t) rrcz(t, x), then zc(t) satisfies equation
(2.1).
Proof From Lemma 2.5 we have E C°(Ec, 5f(Ec, Cy+3u)),

and from (1.33) we have
K e(y+3Q)171I
32
I DDx*(T, S )771 < 2 eylTll.nl <
nI, d , 'q E Ec, T E R.
This inequality and (1.5) imply that for each r] E E, and all E E,
f°e-Ar7r.D f(x*
(T, f)) Dfx* (T, f) ldT
3K23
2
1 I"]I
-° e[P-(y+3-)]Tdr
3K2S11'iI
(2.27)
2(13 - (y + 3o,))
Similarly, for each 77 E E, and all f E E, we have
3K2511771
f° e-AT,r5Df(x*(T,6))Dfx*(T,6)?7dT
2(a-(y+3a))
(2.28)
Hence, we can take the derivative under the integral sign with respect
to 6 E E, on the right-hand side of (1.27). This gives for each 77 E E,
D*(f)rl = f
f° e-ArTSDf(x*(T, (2.29)
Moreover, the uniform convergence of the above integrals with respect

to E E, (see (2.27) and (2.28)), and the continuity of Df, x*(-, ), and
imply the continuity of D4i(e). From equations (2.27) and
(2.28), we obtain the boundedness of D+/. Therefore, 4 E Cb(E,, Eh).
Since f (O) = 0, Df (0) = 0, by using the uniqueness of solutions of
(1.1) with initial conditions, we have
x*(t,0) =i(t,x*(0,0)) =z(t,0) = 0,

and by (1.27) and (2.29)
4,(0) = 0 and D4i(0) = 0.

22 Center Manifolds
It is obvious from (2.27) and (2.28) that if we choose S1 small enough,
then
Lip(tfi) < 1.
Finally for x E W`, if we take i(t, x) = e(t) + p(i(t)), where

6(t) E E, then
(I + Dql) =Ae +A+/i(6) + f(e +

Projecting both sides of the above equality onto E, and noting that 41:
E, - Eh, Drs: TE, = E, -p TEh = Eh, we have
e E E,,
and this is equation (2.1).
Proof of Theorem 2.1. The conclusions for the case k = 1 have been
proved in Lemma 2.6. The case k >- 2 is slightly different from the case
k = 1, although the basic arguments are eventually the same. We will
prove the case k = 2; the general case can be obtained by an induction
on k. In the following we assume a < y < ky < 0 and o, > 0 is
sufficiently small.
Suppose that f e Cb(D'). Then by Lemma 2.5, Dex*(t, ) E
C°(E,,.(E,, Cy+3o)). We prove first that H as a map-
ping from E, to 2°(E,, Cuy+3a>+,-), is differentiable. We will use the
same idea as in the proof of Lemma 2.2 and consider equation (2.21)
instead of (2.2). Let
v = v(t, , 2, n) = Dx*(t, 6)71 - Dx*(t,
e, , 71 E EC, t E R,
and
L(v) = 2,,q)).
Then from (2.21) we obtain
(I - L)v = ,i)), (2.30)

where
S(t, 71)
=Df(x*(t,6))D
Df(x*(t, ))Dx*(t, ),1 - Df(x*(t, )) v

_ (Df(x*(t, 6)) - Df(x*(t, 4)))Dx*(t, 6) 71
= [f'D2f(ox*(t,) + (1 - 0)x*(t, ))dOI
X ((x* (t, ) - x*(t, )), Dex*(t, )r7)
= [f'D2f(ox*(t,) + (1 -
0
X Dx*(t, (1 - 0)4)d01( - ))
0
and ( , ) denotes the action of the bilinear map D 2f. We define the
following bilinear form:
[j1D2f(oX*(t , 6) + (1 - 0)x*(t, 4))d01
X (D x*(t, a, f 1D6x*(t, 0 + (1
o
- b ).
/
Thus, (2.30) becomes
(I - L)(v) =
Hence, in order to prove the differentiability of 6) (as a

24 Center Manifolds
mapping from E, to '(E,, C2(y+3(,)+a)), we only need to verify the
following two facts:
(i) 4)( , )) is a bounded bilinear operator from E, to
C2(y+3a)+a, Where 2E, = E, x E,;
(ii) 6, j) - j,))( 0 as 16 -I -
0.
Note that
B(t, S, 4)i17l2 = D2f (x*(t, S ))DDx*(t, S )771 - DDx*(t, )j2
For any 7711 712 E E,
11
4)1712) 02(y+3a)+a
1 + 1
< sup e-(2(y+3a)+a)ltlIl f IIc2
t(=R 2(y + 3v) - a 6 - (2y + 6o,
-e
(3K)2
2
I'0I7721
< M17711117211 for some M> 0.
Thus, (i) holds. On the other hand, for any '171,772 E E, we consider
4) - 4, 4N02) as a mapping from C2(-,+3,) to
C,,+3,)+,, and can prove that
11G(h(-, as16-4I -+0
by completely the same way as in the proof of Lemma 2.3. Hence (ii)
holds.
Similarly to Lemma 2.4, we can obtain an equation satisfied by
E, -Y(Ec,C2(y+3a)+2a)
In fact, if we let
Dcx*(t, + Ag12)7)1 - DDx*(t, )711

8(t, A) fl1'?12EEc,
A
limg(t,A) =D£x*(t, )771712,

A-0
9(t, A) = (2.32)
1
h(t, A) = A [Df(x*(t, + A772))D x*(t, + AT12)r11
-Df(x*(t, f))Dx*(t, )'n1]

1
A
[Df(x*(t, 6 + A712))
x (DDx*(t, + A'y72)771 - Dx*(t, 6) 711)

+(Df(x*(t
6 + A712)) - Df(x*(t, )))Dx*(t, 0711]
= Df(x*(t, + A712))g(t, A)
+ I f 1D2f(Ox*(t, + A772) + (1 - 9)x*(t, ))de)

0 /
.D,x*(t, 0 e( + A712) + (1 - 0)6)do '172-

0
Using Lemma 2.3 and taking limits on both sides of (2.32), we obtain
x*(t,
D2 )(711712) = G(Df(x*(., ))D2x*(., )711712
+D2f(x*(., ))DDx*(., )711D6x*(., 0712)
G(Df(x*(., ))D2x*(.,S)711712 +H1(', )).

(2.33)
As in the proof of Lemma 2.5, we obtain that there exists a number

< S1 such that if IIDf II < S2, then e) as a mapping from E,
322
26 Center Manifolds
to .(Ec,C2(y+3o)+30) is continuous in E E. In fact, the term
in (2.21) has no influence in the proof. The only difference is the
existence of the additional term H1(t, t; ) in (2.33). But H1(t, ) is
continuous in (t, ) E 118 X E, Hence, by Lemma 2.3,
IIG(H1(', ) - H1(', ))II C2(Y+3o)+30 - 0 as 16 - el - 0.
Here we consider G as a mapping from C2(y+3o)+2o to C2(y+3o)+30

Finally, by the same reasoning as in Lemma 2.6, we can take the
derivative under the integral sign with respect to i; in (2.29), and obtain
.\ l
D2+f( 0771772
f 0 e-Armu(Df(x*(T,
))D2x*(T, 6) 771772 + H1(T, 6))dT
+ f0 e 0711712 + H1(T, ))dr,
(2.34)
where H1(T, ) is defined in (2.33). Besides, the continuity and bound-

edness of D2f, D x*(t, ), and H1(t, 6) implaV E Cb(EC, Eh).
We have just proved Theorem 2.1 for k = 2. Suppose now the
conclusions are true for k = j > 2, that is:
(1) 6) exists as a mapping from Ec to
Cj(y+3o)+(j-2)3a+o), where Eci = Ec X E, X X E, (j times),
and satisfies the equation (as a mapping from Ec to
9'(E,, Ci(Y+3o)+(j-2)3o+20))
Djx*(t, 6) (771 ... 77i)
= G(Df(x*(', ))Dix*(', 6)(711 ...77i) + Hj-1(', ))(t),
77i E Ec, (2.35)
where Hj_1(t, ) is a finitesum of terms involving

Df(x*(t, i)), ... , Dif(x*(t, i)); Dex*(t, (i = 1, ... , j -
1), ..., Dt-1x*(t, 4x711 ... 77j-1).
Local Center Manifolds 27
(2) There exists a number Sj < Sj_ such that if II Df II < Sj, then
1
as a mapping from E, to .(Ej,Cj(,y+3o)+(j-1)3o) is

continuous in E E, and hence
(3) 4 E Cb(EC, Eh).
If f E Cb+1(l"), we need to prove the above conclusions are true for
Butt
k = j + 1. the procedure is completely the same as that done for

k = 2. Therefore, Theorem 2.1 is proved. 0
1.3 Local Center Manifolds
In the previous two sections we established the existence, uniqueness,

and smoothness of the global center manifolds for equation (1.1). The
condition f E Cb (W?) is natural. But the hypothesis Lip(f) < So (or
II Df II < So) for a small So > 0 is quite strong. If we consider a bifurca-
tion problem only near an equilibrium point of (1.1), then we need a
local center manifold. This can be obtained from the global center
manifold of a modified equation by using the cut-off technique, and the
hypothesis II Df II < So will be satisfied automatically since f(0) = 0 and
Df (0) = 0. Let us discuss this in detail.
We consider a cut-off function x: W' - R with the following proper-
ties:
(i) X(x) E C";
(ii) 0<X(x)<1,VxEl8";
(iii) X(x) = 1 if IxI < l and X(x) = 0 if IxI >_ 2.
Related to f(x) in (1.1) and for a given p > 0, we define
.fP(x) =f(x)X(p l
Vx E l ". (3.1)
,
Thus, as a modification of equation (1.1), we consider
.z=Ax+ff(x). (3.2)
Obviously, if we restrict x to the domain IxI < p, then equations (1.1)

and (3.2) are the same. The following lemma shows that if we choose p
sufficiently small, then II DfP II can be very small. Hence we can apply the
global center manifold theory to (3.2), and get some local results
for (1.1).
28 Center Manifolds
Lemma 3.1. If f E Ck(') for some k >_ 1, and f(0) = 0 and Df(0) = 0,
then fp(x) E Cb(l") for a given p > 0, and
limIlDfp11=0.
P-0
Proof. Since f E Ck and X E C°°, fp E Ck. For a given p > 0, fp(x) = 0

if I xl >_ 2p, whence fp c- Cb (l "). From (3.1) we have that
Hence,
1
IIDfpll s sup IDf(x)I + - IIDxII sup If(x)I. (3.4)
IxI52p P IxI52p
The condition f(0) = 0 implies f(x) = foDf((1 - A)x)xdA. This gives
sup l f(x)I < sup IxI( sup IDf(x)l) < 2p sup lDf(x)l.
IxI52p IxI52p IxI52p IxI52p
Substituting the above inequality into (3.4), we obtain that
IlDfpll <- (1 + 2IIDxII) sup I Df(x)l.

IxI52p
The desired result (3.3) follows from the above estimate and the
condition Df (0) = 0.
Theorem 3.2. Suppose that f E Ck(I!") for some k >_ 1, and f(0) = 0,
Df(0) = 0. Then there exists Vi E Cb(EC, Eh) and an open neighborhood
fl of x = 0 in F" such that
(i) the manifold
My {xC + 4r(x,) I xc E Ej (3.5)
is locally invariant under (1.1). More precisely,
X(t,x) EMO, VxEM.nff, VtEJn(x),

where At, x) is the flow of (1.1) with x(0, x) = x, and JJ(x) is the
maximal interval of existence of the solution x) with respect to fl;
(ii) 0(0) = 0 and Da/r(0) = 0;
(iii) if x E fl and J0(x) = R, then x E Mk.
Proof. Let Sk be given as in Theorem 2.1. By Lemma 3.1, we can find a

p > 0 such that fp E Cb (I!") and II Df f ll < Sk, where f,, is defined in
(3.1). By Theorem 1.1 and Theorem 2.1, we can obtain the global center
manifold M,4 of equation (3.2), M. is defined by (3.5), 41 E Cb (R" and
p(0) = 0, Dii(0) = 0.
On the other hand, from the properties of the cut-off function X,
we know that the equations (3.2) and (1.1) are the same if x E fl
(x E E" I I I x II < p). Thus, the conclusions (i) and (ii) are proved.
We suppose now that x E fl and J0(x) = R. Then x(t, x)
1p(t, x) c fl, V t E R, whence sup1 ERlTrhx(t, x)l < -. By (1.2), x E My,
and conclusion (iii) follows.
Definition 3.3. If 4) E Ck(E,, Eh), k >_ 1, 4)(0) = 0, D¢(0) = 0, and

MM (x, + ¢(x,) I x, E E,) is locally invariant for the flow of (1.1),
then M,, is called a Ck local center manifold of (1.1).
Lemma 3.4. Suppose that f E C 1(W"), f (O) = 0, and Df (0) = 0; and

-0 E C1(E,, Eh), 0(0) = 0, and D4)(0) = 0. Then M. {x, +
¢(x,) I x, E E,} is a local center manifold of (1.1) if and only if there is a
neighborhood ftc of the origin in Ec such that for all xc E ft,
Do(xc)7rw(Ax, +f(xc + 4)(xc))) = 7rh(A4)(xC) +f(xc + O(xc)))

(3.6)
Proof. Suppose that such an f2c exists and (3.6) holds. For each
xc (=- flc, let zi(t) be the solution of the following initial value problem
zC _ -rrc(Aic + f(-'c + 4)(xc))), 1 JO) = x,
and let z(t) = xc(t) + 4)(zJt)). It is obvious that x,(t) E f.l, if Iti is
30 Center Manifolds
sufficiently small. Then by (3.6) we have
z(t) = (I + D4(i,(t)))7rr(Ai,(t) + f(f (t) + 4(i,(t))))
= Or, + -rrh)(Ai(t) + f(i(t))), Itl is sufficiently small.
This means i(t) = i,(t) + 4(i,(t)) is a solution of (1.1) if Iti is suffi-

ciently small. Hence, M. is locally invariant under (1.1), and, by
Definition 3.3, it is a local center manifold of (1.1).
Suppose that M. is locally invariant under (1.1) in a neighborhood fl
of the origin in R". Let fl, be an open neighborhood of the origin in E,
such that x, + 4(x,) E fl if x, E Sty. For any x, e ( let i(t) =
i(t, x, + 4(x,)), which is the solution of (1.1) with the initial condition
i(0) = x, + 4(x,). Then the local invariance of M. under (1.1) implies
that for t sufficiently small we have
Trhx(t) = 4 rcx(t)).
Differentiating the above equality with respect to t, and using (1.1), we

obtain
7rh(`4-'(t) +f(x(t))) = Dcb(ircx(t)) rr(Ai(t) +f(i(t))).

Taking t = 0 and noting 0: E, -+ Eh, we get (3.6).
Theorem 3.2 gives the existence of a local center manifold. But, in

general, it is not unique.
Example 3.5. Consider the planar system
x = x2, Y = -y.
It is easy to see that E, = {(x, 0) I x E R) and Eh = ((0, y) I Y E RI.

Suppose 0 E C'(E, Eh) gives the local center manifold M. = (x +
4(x) I x E R). Then by Lemma 3.4, we have
0,(x)x2
= -4(x), 0(0) = 0'(0) = 0.
Hence
O(X) __ ae','X for x < 0;

0 for x >_ 0,
where a E O is a constant. Each different a gives a different M.. This

means that local center manifolds are not unique, even in a sufficiently
small neighborhood of the origin.
Fortunately, the nonuniqueness of local center manifolds is not a
serious problem when we consider bifurcation phenomena of vector
fields. In fact, every local center manifold of (1.1) contains all bounded
solutions of (1.1), for example, equilibrium points, periodic orbits, or
homoclinic or heteroclinic orbits, provided they stay in a sufficiently
small neighborhood of the origin. To show this, we need the following
result which says that each local center manifold of (1.1) can be
obtained from the global center manifold of a related vector field.
Theorem 3.6. Suppose f r C"(R") for some k >_ 1, f(0) = 0 and

Df(0) = 0, and 0 E Ck+'(E,, Eh) defines a local center manifold MO of
(1.1). Let S E (0, So], where So is defined in Theorem 1.1. Then there
exists a neighborhood fl of the origin in Fl" and mappings f E Cb (I8")
and 4 E Cb "(Ec, Eh) such that
(i) f(x)=f(x),VxEE fZ;
(ii) II Df II < S;
(iii) M4, n SZ = M. n fl,
where M. is the unique global center manifold of the following equation
x =Ax +AX). (3.7)
Proof. Part (I): a special case. We suppose that M,, = Ec is a local

center manifold of (1.1), that is, there exists some neighborhood fl, of
the origin in Fl", such that
¢(xc) = 0, Vxc E=- Ec n fft.
By Lemma 3.4, there exists a d > 0 such that
Trh f (xc) = 0, V xc E Ec and IIxcII < d. (3.8)

32 Center Manifolds
Let 0 < p < d/2, and X E C(P", fl) be a cut-off function. We define
f(x) =f(x)X(p)' Vx E R";
then
T1hf(xC) = (Trhf(xc))X(p) = 0, Vx, E E, (3.9)
In fact, if Ixj < 2p < d, then 71h f(X,) = 0 by (3.8); if I xj > 2p, then
X(x,/p) = 0. Lemma 3.4 and (3.9) give the invariance of E, under
(3.7). By Lemma 3.1, we can choose p so small that I I Df I I < S. On the
other hand, since f E=- C6 (R") and II Df II < S < So, it follows from
Theorem 1.1 that (3.7) has a unique global center manifold Mo. Hence,
M o = E, We define r/i(x,) ° 0 and fl R' Hence,
41 E Cb +1(E,, Eh) and the conclusions (i)-(iii) are satisfied.
Part (II): the general case. Let .0 E Ck+'(E,, Eh), and M. is a local
center manifold of (1.1). By Lemma 3.4, we can find some do > 0 such
that (3.6) holds for I x,I < do and x, E E, Let d E (0, do) and define
E CGk+ l(Ec Eh) by
gf(xc) = -O(xc)X(d V xc E Ec. (3.10)
We make a transformation
y = x - +Ii('r x) P(x), V x E R". (3.11)
Then it is easy to verify that

(a) ''P(M,,) = Ec;
(b) 41-1(y) = y + i(7rcy), V y E U8".
Under the transformation (3.11), (1.1) becomes
y =Ay +g(y), (3.12)
where
g(y) =A./i('rrcy) +f(y + -A(7rcy))
- D4(7rcy)7rc(Ay +f(y + 41 (7rcy))). (3.13)

Since E, is a local center manifold of (3.12) (see the property (a)), the
results in Part (I) imply that there exist some g E Cb (P.') and some
neighborhood fl of the origin in V1" such that
g(Y) = g(y), V Y E fl, (3.14)
and E, is the unique global center manifold of the equation
y =Ay +g(y). (3.15)
Furthermore, IIDgII can be smaller than any given positive number. Now
if we take the inverse transformation x = 'I'1(y) = y + a f (7r y ), (3.15)
becomes an equation of the form (3.7), where
f(x) =Dfi(7rx)A7rx -A4i(irx) +g(x - 4r(7rx))

+ Dr/i(-rrx)-rrrg(x - +/(-rrx)). (3.16)
Since g E Cb (R"), / E Cb + 1(E,, Eh), and 4 has bounded support, we

have f E Cb(W'). We claim that if we take Cl = and let Cl be
sufficiently small, then the conclusions (i)-(iii) are satisfied.
In fact, `d x E Cl = 'I-1(SZ), y = x - /r(7r,x) E (I, and hence (3.14)
holds. Noting 41, DIIi: E, - Eh, and substituting (3.13) into (3.16), we
obtain f(x) = f(x) for x E Cl.
Next, by using the following equalities
4r(7r,x) = (f'D((1 - 0)7rx)dB)(-7r,x),
Ox - 41(7rx)) = (f1Dg((1 - 0)(x -',(Trx)))d6)O/(Tr'x) - x),
we can obtain from (3.16) that
II Df II < (2IAI117r11)IIDiiII + (1 + IIir IIIIDiPII)211Dg11.
By (3.10) and Lemma 3.1, I I DIr I I - 0 as d - 0. Hence we can choose d

so small that (2IAIIkr II)IIDrII < 8/2. Fix such a d > 0; then 41 and g
are well defined. By Part (I), we can choose Cl properly such that
(1 + 5/2. Thus, we have IIDf II < 5.
34 Center Manifolds
Finally, from (3.10) it is obvious that ¢(x,) = ir(x,) if IxJ 5 d. Let SZ
be sufficiently small so that I wry x 15 d if x E fl. Hence, M. n SZ =
M4, n fl. Since E, is the unique global center manifold of (3.15) and
T-'(E,) = Md,, M. is invariant under (3.7). But II Df II < S 5 So, so by
Theorem 1.1, M1, is the unique global center manifold of (3.7).
Theorem 3.7. Under the assumptions of Theorem 3.6, there exists a

bounded neighborhood fZ of the origin in uR" such that if x e ft and
J0(x) = 1k, then x E M1,. (For the definition of Ju(x), see Theorem 3.2).
Proof. We use Theorem 3.6 with S = So, and assume that f found by
Theorem 3.6 is bounded (otherwise, just shrink it to a bounded one).
Suppose that x e H and Jn(x) = fk, which means Z(t, x) E fZ for all
t E 1k, where i(t, x) is the solution of (1.1) with 1(0, x) = x. By the
conclusion (i) of Theorem 3.6, X(t, x) is a solution of (3.7), and it is
globally bounded. By the conclusion (ii), we can use Theorem 1.1, and
hence x E M,,. By the conclusion (iii), x e Mo.
Remark 3.8. Theorem 3.7 says that if M. is a Ck (k >_ 1) local center

manifold of (1.1), then it must contain all small bounded solutions of
(1.1). In particular, M. must contain all sufficiently small equilibria,
periodic orbits, and homoclinic and heteroclinic orbits.
Theorem 3.9. Suppose that f E Ck(1") for some k >_ 1, f(0) = 0 and
Df(0) = 0, and M., and Mo2 are two Ck+1 local center manifolds of
(1.1). Then we have
D'41(0) = D'42(0), 1 5 j 5 k. (3.17)
Proof. We use Theorem 3.6 for 01 and c62 with S = Sk 5 So, where So
and Sk are defined in Theorems 1.1 and 2.1, respectively, and they
depend only on A E £9(OR",1k") in (1.1). Then there exist corresponding
t., f, E Cb(OR"), and i/r, E Cbk+'(Ec, Eh) satisfying the conclusions
(i)-(iii) for i = 1 and 2, respectively. Let ft = H1 n 512; then we have
Asymptotic Behavior and Invariant Foliations 35
the following conclusions:

(a) f(x) = f1(x) = f2(x) for x E fl;
(b) IIDf,II<5=Sk<S.for i=1,2;
(c) MM.nfl=M,, nf1fori=1,2,
where M,,,, is the unique global center manifold of the equation
z=Ax+f;(x).
From (1.27) and the proof of Theorem 2.1 we know that D'ii1(0)
(1 < j < k) is completely determined by A and f(x) for x in a
sufficiently small neighborhood of the origin. Hence, by (a) and (b),
Di4i1(0) = DNi2(0), 1 < j < k, and then by (c), D'¢1(0) = D42(0), 1 <
j<k.
Remark 3.10. The conclusions in Theorem 3.7 and Theorem 3.9 give
partial uniqueness of local center manifolds. At the end of the next
section we will introduce a new result by Burchard, Deng, and Lu [1]
which says that the flows on any two Ck+1 local center manifolds of
(1.1) are locally Ck conjugate. Hence, we can choose any local center
manifold to study the bifurcation phenomena.
1.4 Asymptotic Behavior and Invariant Foliations
In this section we will generalize the results on global center manifolds

in Sections 1.1-1.2 to the cases of global center-stable and center-
unstable manifolds. Then we will discuss asymptotic behavior of solu-
tions outside these invariant manifolds. This is related to foliations of
OB". We will go back to the local situation by using the cut-off technique,
as in Section 1.3, and study stability properties of local center mani-
folds.
Denote
7TC" = 7Tc + 7r", Tres = 7T, + 7T5,
E,"=E,®E", E,,.=E,®E5.
We now introduce the center-unstable manifold for which the proof of
results is completely similar to that in Sections 1.1 and 1.2. Suppose
36 Center Manifolds
that the positive numbers a, /3 are determined in Lemma 1.5, and
a<y</3.
Theorem 4.1. (i) There is a positive number S,u such that if f E Cb' 1(Rn)
and Lip(f) < Scu, then the set
W'u = l(x E ll" I suplirsx(t,x) I <00}

t50 /
_ {xEWIsup e7tIi(t,x)I <oo} (4.1)

ts0
is invariant under (1.1), and is a Lipschitz submanifold of Di", that is,

there exists a unique Lipschitz function 41 E Cb(ECu, Es) such that
Wù = {xCu + i/i(xcu) I xCu E Ec"}. (4.2)
(ii) If ¢ E Cb(Ecu, Es) and the manifold
MO {xCu + la(xcu) I xcu E Ecu} (4.3)
is invariant under (1.1), then M. = W CU, ¢ = 41.
We say that W`" is the unique global center-unstable manifold of

(1.1). Using the same approach of Section 1.2, for any k >_ 1, we have
the following:
Theorem 4.2. Suppose f E Cb (l r) for some k >_ 1. Then there is a

number S' > 0 such that when II Df II < S u, the conclusion of Theorem
4.1 holds with t/t E Cb(Ecu, E).
We next consider the existence of global invariant foliations. For

y E R, a Banach space is defined by
Cy = Z C C°(l ,R") I IIzIIy = sup eytIz(t)I <

tE=R+
We fix y such that a < y < k y < /3, where k is the positive integer in
Theorem 4.2. Suppose that x(t,1), i(t, x) are the two solutions of (1.1)
satisfying the initial conditions 1(0,1) =.i and 1(0, x) = x, x, x E n".
Asymptotic Behavior and Invariant Foliations 37
We define the set
M,(x) = {x E Rn I z(t) =.z(t, z) - z(t, x) E CY }. (4.4)
We call M,(x) the stable leaf of x E R".
Theorem 4.3. Suppose f E C°' 1(Rn) and Lip(f) < S, Then there exists
a uniformly continuous mapping J,: l X E, - Ec" such that for each
x E C$" the following conclusions hold:
(i) M,(x) = (x, + J,(x, x) I xs E Es) is a Lipschitz manifold;
(ii) Ms(x) has a unique intersection point with W`";
(iii) there is a stable foliation of l":
R " = U M,(x). (4.5)

xeW"'
(iv) The foliation is invariant under (1.1), that is,
MS(i(t, X)) = X(t, MM(x)), `d t > 0, `d x E R'.
The proof of Theorem 4.3 needs several lemmas.
Lemma 4.4. Suppose f E Cb.1(Rn), Lip(f) < 8c", and z(t) E C,. For
x E R", x) + z is a solution of (1.1) if and only if there exists some
xs E Es such that
z(t) = eA`x, + foè'`-T)7rf(T; x, z(T))dT
+ f teA(t-T)1rcuf( r; x, z(T))dr,
where!: F X Ii" X V8" - R" is defined by
f(t; x, z(t)) = f(z(t, x) + z(t)) - f(z(t, x)). (4.7)
Proof. If x) + z is a solution of (1.1), then z satisfies the equation
i=Az+f(t;x,z). (4.8)
Exploring the Variety of Random
Documents with Different Content
CHAPTER X
TIT FOR TAT
After that it was a comparatively easy matter to get Henry Stowell

to tell the details of what had been done. Several times Dock Wesley
tried to stop him, but finally he also capitulated and became almost
as humble as the sneak.
“It was only a bit of fun,” said Wesley. “Can’t a fellow do something
on the last night at school?”
“Sure!” answered Fred.
“But you’ve got to take your dose in return,” was Fatty Hendry’s
comment.
Thereupon Codfish and Wesley admitted that they and four other
cadets had entered the rooms occupied by the Rovers and their
chums and taken away all their clothing and their bed things.
“Everything is locked up safe and sound in Room Forty-two,” said
Codfish. “You know, that room hasn’t been occupied this term.”
“How did you get the key?” asked Andy.
“We got it one day from the janitor when he was cleaning up. He
thought he had lost it, and so locked up with a duplicate.”
“Where is the key now?” asked Jack.
“I—I let Dock keep it,” faltered Codfish.
“Say, you needn’t put off everything on me,” growled Wesley. “You
had as much to do with this as anybody. The key is on a hook in that
closet,” and Wesley nodded toward a closet in a corner.
“Now we want to know who the other fellows were,” declared Fred,
after the key had been secured.
“Oh, you had better not ask that,” pleaded the sneak. “If we give
them away they may hammer the daylights out of us.”
“You talk up, Codfish, or you may get the hammering right now,”
put in Gif.
Thereupon Codfish mentioned the names of four cadets who had
been more or less chummy with him since the term had started. Two
were new boys, and all were fellows with whom the Rovers and their
chums had had little to do.
“Now put on your slippers and come along with us,” ordered Jack.
“What do you want of us?” questioned Wesley.
“First of all, you’re going to bring all that stuff back,” declared the
young major. “After that we’ll see what we’ll do.”
“Why don’t you make the other fellows join us?” asked Codfish. He
thought there might be safety in numbers.
“We’ll take care of them later on,” put in Gif grimly.
Finding themselves cornered, Codfish and Wesley accompanied
the others to Room 42, and there on the bed, on the chairs, and on
the floor they found all the things taken from the Rovers and Gif and
Spouter.
“I call this something of a mess,” declared Fatty, who had come
along. “Here, give me some of that clothing! I’ll help carry it.”
Even with the assistance of those who had suffered from the joke,
it was necessary to make several trips back and forth to get all the
things where they belonged. During the last trip Fred and Andy
noticed some other cadets hiding in the shadows at the end of the
corridor and laughing softly among themselves.
“They think they’ve got the joke on us,” whispered Fred. “Come
on, let us make a break for them.”
“Not yet. I’ve got a better plan,” came from Randy.
After everything had been restored to the rooms, the Rovers and
their chums marched Codfish and Wesley back to their own
quarters.
“Now then, I think we’ll give you a dose of your own medicine,”
said the young major. “Boys, pick up all that extra clothing and all
those quilts and bedsheets and put them in the closet over there.”
“Say, what does this mean?” demanded Wesley.
“You’ll see in a minute.”
The others were quick to catch the idea, and all the bed coverings,
as well as the wearing apparel in the room, were quickly transferred
to the closet.
“We’ll leave you your pajamas, for you might catch cold,” said
Randy. Then the closet door was locked and the key taken away.
“Now, don’t try to raise a row, or you’ll be sure to get the worst of
it,” said Jack, as the crowd prepared to leave the room.
“We can’t stay here with nothing on the beds!” cried Codfish.
“You thought we could do it, didn’t you?” asked Andy. “It’s simply tit
for tat. Go on and lie down and enjoy yourselves.” And thereupon the
Rovers and their chums withdrew, locking the door after them.
“I guess that will hold them for a while,” remarked Spouter. “They
can’t get their things unless they break open the door, and I don’t
think they’ll go that far. And they can’t get out unless they go on the
fire escape, and the door from there to the corridor is locked on the
inside—they’d have to go through some of the other fellows’ rooms.”
“Now then, how are we going to square up with those other
fellows?” asked Gif.
“I was thinking I might sneak down and get old Huxley’s garden
syringe—the one he uses to spray the bushes and flowers with,” said
Andy. “We might give ’em all a dose of ice-water, or something like
that.”
“Old stuff,” declared Fred. “Can’t we think of something new?”
“We might blow some smoke through the keyholes or under the
doors,” suggested Randy. “Then we could bang on the door and let
them think there was a fire.”
“Gosh! that isn’t half bad,” said Fred. “But how shall we make the
smoke? We can’t build a fire, or anything of that sort.”
“Some wet paper will do the trick.”
“I don’t think you ought to try that, boys,” declared Jack. “It might
bring on a panic, and we don’t want any one to be hurt on this, the
last night at the Hall. Come on and see if we can’t get hold of those
fellows.”
They passed around a corner of the corridor, and as they did so
Gif suddenly clutched the youngest Rover by the arm.
“There go some fellows now!” he whispered. “See them crawling
along over there? I wonder who they are and what they’re up to?”
The lights in the hallway had been turned low, and the Rovers and
their chums could just make out the forms of four cadets slinking
along silently. Then they disappeared from view around one of the
numerous corners.
Curious to know what new fun might be in the air, the Rovers and
the others followed the crowd like so many shadows. They saw the
four cadets who were ahead stop in front of the room which they had
left but a few moments before.
“Gee, I know that crowd!” exclaimed Andy, in a low voice. “Those
are the very fellows Codfish and Wesley mentioned—the fellows
who helped them take our things.”
“They must be wanting to know what we were doing here,”
suggested Gif. “Say, why can’t we pounce on ’em and make ’em
prisoners? We are seven to four.”
“I’m game if you fellows are,” answered Randy readily.
A plan was hastily formed, and just as the four cadets had begun
their talk with Codfish and Wesley, out of the semi-darkness pounced
the Rovers and their chums.
“Give in! Give in!” was the whispered command. “Give in or you’ll
get the licking of your lives!”
“Hi! Stop that!” roared one of the cadets, a lad named Morris. “Let
up!”
“Do-do-don’t ch-choke me to death!” spluttered a cadet named
Shamberg. “Let up, I tell you!”
“It’s the Rovers!” came from a third of the lads.
“They’ve found us out!” wailed the fourth, a fellow who was just as
much of a sneak and coward as Codfish had ever been.
Surrounded and taken completely off their guard, the four cadets
were speedily made prisoners. Then, almost before they knew what
was happening, they were taken to the two adjoining rooms which
they chanced to occupy. One of the rooms had a rather large closet
which at one time had been a storeroom. It had a small window
about five feet from the ground.
“I’ve got an idea,” said Jack. “Throw a mattress in here on the
floor.”
The others quickly caught on and in a trice a mattress from one of
the beds was flung on the floor of the storeroom. Then the four
cadets who had been captured were forced into the place.
“Now you fellows can stay here until morning,” declared Jack. “You
didn’t want us to have a decent night’s sleep, so now you can get
along in any old way you please. Don’t dare to make a rumpus, or
we’ll be after you in a way you least expect.”
“Gee, we’ll smother to death in here this warm night!” declared
Morris.
“No, you won’t,” said Spouter. “You can take turns at looking out of
the window. But I’d advise you not to crawl out, because it’s about
twenty-five feet to the ground.”
“We’ll report this to-morrow, you see if we don’t,” grumbled
Shamberg.
“Report and be hanged,” retorted Gif. “If you say a word to Colonel
Colby we’ll tell him what you did.” And thereupon the Rovers and
their chums withdrew, locking the storeroom door and then locking
the door to the corridor.
It was a good quarter of an hour after Gif, Spouter and Fatty had
left them that the Rovers were able to rearrange their beds so that
they could lie down. All were now thoroughly tired out and Andy
could scarcely keep his eyes open. But there was to be little sleep
for any of the cadets during that last night at Colby Hall. Half a dozen
parties were wandering around, making all the fun possible, and
presently Professor Snopper Duke came after some of the boys,
trying to quiet them.
“This is disgraceful!” stormed the irate teacher. “I want you boys to
keep quiet.”
Then came an alarm from Codfish and Wesley, as several other
cadets broke into their room, bent upon bringing the sneak and his
chum to terms for something done in the classroom the week before.
Into this row Snopper Duke precipitated himself, and as a
consequence was struck in the nose by a baseball which one of the
lads threw at Codfish.
“Oh, oh, my nose! Who threw that baseball?” roared the teacher.
Then, as the blood began to flow from the injured organ, he
hastened off to the nearest bathroom where he might bathe it.
It was all of three o’clock before the Rovers got any sleep at all. By
half past six they were again awake and busy packing their things,
ready to depart. Then Randy and Andy sneaked away and liberated
Morris, Shamberg and the other two with them.
“Hope you slept well,” said Andy, grinning.
“You let me get my hands on you, and I’ll show you how I slept,”
stormed Morris. But then Andy ran off laughing and his twin followed
him. The other boys were very sore, but did not dare to do anything.
“And now to get the girls and start for home!” said Jack, a short
while after breakfast.
“And then for our vacation!” added Fred. “If only we knew where it
was going to be!”
“You’ll know very soon,” declared Andy. “Randy and I have made
up our minds to tell you as soon as we are ready to leave Haven
Point.”
CHAPTER XI
A MYSTERIOUS PLOT
“Good-by, boys. I wish all of you the best of luck.”

It was Colonel Colby who spoke as he shook hands with the Rover
boys and a number of the other cadets.
“Good-by, Colonel. I hope we see you again some time,” returned
Jack.
“You must come and visit us at our home when you can get time,”
put in Fred.
“I will certainly come when I can get away,” was the reply from the
master of the school.
Breakfast was at an end and all was bustle and confusion as the
cadets were hurrying in all directions, suitcases in hand, ready to
leave the Hall. Many were going away in automobiles which lined
one side of the campus drive. Others were to go to the Haven Point
railroad station. A motor truck had already taken two loads of trunks
away and was now back for a third.
“Good-by, fellows!” cried Gif. “Hope you have a good time.” He
and Spouter had arranged to go up on the coast of Maine with Dan
Soppinger and their folks.
“Good-by!” cried the Rovers, and a few minutes later had entered
the touring car which was to take them away.
“Here is something to remember us by!” shouted Spouter gayly,
and threw a bunch of confetti over the Rovers.
“And here is something to remember me by!” yelled Andy, as the
car moved away and he hurled an old shoe he had picked up at
Spouter, catching that cadet in the stomach, causing him to give a
grunt of surprise. Then the touring car rolled out of the grounds, all of
the boys waving their hands as the place faded from their sight.
“Now it is good-by to Colby Hall and hurrah for a vacation!”
exclaimed Fred. Then he added quickly: “Now then, Andy and
Randy, where are we to go? Don’t keep me waiting any longer. I’m
all on fire with suspense,” and the youngest Rover put on a tragic air.
“Wait till we pick up the girls,” pleaded Andy. “No use in going over
the whole thing twice. They’ll want to know about it, anyway.”
It had already been arranged that Martha and Mary, along with
Ruth, were to accompany the lads to New York City. Although the
others did not know it, Jack went armed, having obtained the loan of
a pistol from Colonel Colby, who had been told the particulars
regarding the rascality of Carson Davenport.
“I feel that I am responsible for the safety of my sister and my
cousin,” the young major had told the master of the school. “I want to
be sure that they get home safely.” And thereupon Colonel Colby
had somewhat reluctantly permitted Jack to take his own private
nickel-plated pistol.
When the boys arrived at Clearwater Hall they found the three girls
waiting for them. A few minutes later the whole crowd was off for the
Haven Point railroad station.
“Have you seen or heard anything more of that man Davenport?”
questioned Mary anxiously, as they rode along.
“Not a thing, Mary,” answered her brother. “Have you?”
“Two or three times we saw somebody skulking in the bushes
back of the school,” said the girl. “It was rather dark, and the man
was so far off we couldn’t tell who he was, although Martha thought
he walked like the fellow who tried to push us into the auto.”
It took but a few minutes to reach the railroad station, and during
that time Andy and Randy had no opportunity to speak of the trip the
lads intended to take during their vacation. At the station they fell in
with a number of the cadets, including Phil Franklin.
“I’ve arranged to stay with Mrs. Logan,” said Phil. “And I think
Barry and I are going to have some bang-up times.”
“Don’t forget to look for the silver trophy,” said Jack quickly.
“Oh, I’ve already spoken to Barry about that,” answered the boy
from the oil fields. “We’re going to make a systematic hunt. Of
course, it isn’t going to be very easy to locate the exact spot where
the vase went down.”
“It was opposite that clump of big pines,” declared Randy. “I
noticed the pines just as I went overboard,” he added, with a sickly
grin.
“I’ll remember that—it ought to help us in locating the spot,” said
Phil, and then walked away to bid some of his other friends good-by.
“Now then, Andy and Randy, tell us where we’re going!” cried
Fred, when the Rover boys and girls and Ruth were left for a
moment to themselves.
“You’re going out West,” answered Andy dryly.
“Out West? Where?” came from Jack and Fred.
“You’re going out to the Rolling Thunder gold mine,” said Randy.
“Rolling Thunder! What a name!” exclaimed Ruth, dimpling.
“Where in thunder is Rolling Thunder?” demanded Fred. “I never
heard of such a gold mine.”
“I have,” put in Jack quickly. “It’s the one Uncle Tom invested in a
couple of years ago. I’m right, am I not?” he questioned of the twins.
“That’s it. It’s away out in the Rocky Mountains near a place called
Maporah. It’s on what is known as Sunset Trail.”
“Gee, that sounds good! Sunset Trail!” murmured Fred.
“How are we to go? In an auto?” queried Jack.
“Hardly! We’re to take the train to Chicago and then another train
to Maporah. From there we take horses and ride to a place called
Gold Hill Falls where the mine is located. Dad says we ought to have
a dandy time on Sunset Trail.”
“He says it’s a very wild country, with plenty of good hunting and
fishing, and all that sort of thing,” came from the other twin. “He says
we can go out either with a guide or by ourselves, just as we please.”
“That sounds mighty good to me,” said Fred, his eyes brightening.
“I’d like to spend a few weeks in the saddle, and I’d like to go where
there is some real fishing.”
“Suppose some Indians catch you and scalp you?” put in his sister
mischievously.
“Indians! Humph! If there are any Indians out there more than
likely some of them are from college and on the baseball or football
teams,” was the quick retort. “The old-fashioned Indians exist only in
the story books.”
The boys and girls became greatly interested in the subject of the
outing and talked about it freely until it was time for the train to arrive.
Then they bustled around to say good-by to those who were to leave
in the opposite direction.
“Gee, it makes me feel awfully queer to think I’m never coming
back to Colby Hall!” murmured Fred, as he shook hands with one
and another of the cadets.
“This place has certainly been a second home to us,” answered
Jack. “No matter what happens in the future, I’ll never forget the
days spent here.”
“None of us will!” cried Randy.
“They were great days, the best of days, in spite of such fellows as
Codfish, Gabe Werner, Bill Glutts, and Professor Duke,” declared the
young major.
The girls were likewise in a flutter bidding farewell to their chums
and also several of the teachers who were leaving. In the midst of all
this excitement the train rolled in and a few seconds later boys and
girls climbed aboard and the Rovers rushed down the aisle to get
comfortable seats.
“Good-by to Haven Point!” shouted Andy, out of the window, and
then opening a bag of popcorn he had purchased he scattered the
entire contents over the heads of those left behind.
“Oh, my, look at that!” was the cry. “Popcorn! Did you ever!”
“That was Andy Rover! He’s always cutting up!”
“Here you are, Andy!” yelled Phil Franklin, in excitement, and just
as the train started he sent a rubber ball whizzing through the open
window of the car. The ball struck Andy in the ear, then bounced
away into Ruth’s lap.
“Hi! We don’t want your ball!” called out Andy, and, catching it up,
he threw it through the window, hitting the cadet named Morris in the
chin. Then the train rolled away, and the journey to New York City
was begun.
As the train passed out of sight two men, one about middle age
and the other very much younger, stepped from a corner of a
baggage room which was located close to where the Rover boys and
those with them had been standing.
“I guess you got the right dope that time, Davenport,” said the
younger man, as both walked away unnoticed and entered a
roadster standing on a side road behind some bushes.
“I think I did,” answered Carson Davenport, his manner showing
his satisfaction. “So they are going to Chicago and then to Maporah,
and then out on Sunset Trail, eh? I’ll have to look into that.”
“Do you know anything about the Sunset Trail territory?”
questioned the younger man.
“I do and I don’t,” was Davenport’s reply. “I was never there
myself. But Tate, the fellow I’ve been telling you about, came from
that district and he’s often told me about it. He spoke about this
Rolling Thunder mine, too. He knows some of the fellows working
there.”
“Then what you’ve got in mind ought to be easy, Davenport.”
“I don’t know about its being so easy! Those Rovers are not fools
and since we made a mess of things the other day, more than likely
they’ll be on their guard. I reckon I made something of a mistake
when I called on Dick Rover. I should have waited until I had things
better in hand.”
“What is the next move?”
“I think we had better follow them to New York, and then you had
better find out a few more details of their plans.”
“Why don’t you do that yourself?”
“They know me, and they don’t know you.”
“They saw me out riding with you.”
“True! But I don’t think they’ll remember you. Anyway, you can
easily put on some sort of a disguise. You can bump into the boys
and pretend to get friendly and all that sort of thing,” went on the
man from the oil fields.
“All right, Davenport, I’ll do whatever you want me to,” returned the
younger man. “But understand, I’m not doing this for nothing.”
“I understand that well enough. And I’m not doing it for nothing
either. If we work this thing right there will be a small fortune in it for
all of us.”
CHAPTER XII
HOME ONCE MORE
“Here we are at last!”

It was Fred who spoke as the long train rolled into the Grand
Central Terminal, New York City, and came to a stop. The boys had
collected their hand baggage and soon the Rovers and Ruth were in
the midst of the crowd that was pouring through the gateway into the
waiting room of the big station.
“Here you are—and glad to see you!” exclaimed Dick Rover, as he
came up, followed by his wife and Fred’s mother.
There was a general handshaking and many kisses, and then Dick
Rover took possession of the young folks’ checks for their trunks and
led the way to a side street where two of the family touring cars
waited.
The trip to the metropolis had been without special incident save
for the fact that a number of the cadets, including Andy and Randy,
were inclined to indulge in more or less horseplay on the way. They
had had to make one change at the Junction, and on account of the
heavy travel had been compelled to come down in an ordinary day
coach in place of getting seats in a parlor car. They had managed,
however, to get lunch on the train and had had considerable fun
during the meal.
“I am certainly glad to see you young folks home again,” remarked
Dora Rover, as she gazed affectionately at her son and daughter
and then at the others. “And you are more than welcome, Ruth,” she
added, tapping the visitor on the shoulder.
“Maybe we’re not glad to see little old New York again!” cried Fred.
“I don’t think I’d call it ‘little old New York,’” answered Ruth, with a
smile. “To me it’s a wonderfully big and busy city. When I first arrive
here I always feel like shrinking back until I can get my bearings.”
“Oh, New York is just all right. I wouldn’t want it any better,”
answered Randy.
“But you don’t want to stay here even when you come,” put in his
Aunt Grace. “You just stay at home a few days and then away you
go on one of those trips.”
“Well, I’m a Rover by name, so why not be a rover by nature?”
was the sly reply, and this brought on a general laugh.
Soon the young folks were aboard the two automobiles. In the
meantime Dick Rover had turned the checks for the trunks over to an
expressman and in a few minutes more the whole crowd was
headed for Riverside Drive. Here a surprise awaited them. Not only
was Mrs. Tom Rover on hand to greet them, but likewise their
grandfather, Anderson Rover, and their old Aunt Martha and Uncle
Randolph, who had come from Valley Brook Farm on a short visit to
the city.
“My gracious, this is fine!” cried Fred. “A regular family reunion!”
and then came more hugs and kisses all around.
“My, my! how big you boys are getting!” said old Aunt Martha, as
she surveyed them through her spectacles. “The first thing you know,
you won’t be boys any more—you’ll be men.”
“Well, you couldn’t expect them to remain boys all their life, could
you?” queried Uncle Randolph. “Now they have graduated from
Colby Hall, I suppose they’ll either have to go to college or go into
business.”
“No use of shoving them ahead too quickly,” came from
Grandfather Rover, as he sat down and rested his chin on the top of
his cane. “They have been studying pretty hard for years—let ’em
take a rest. They might take a whole year, if it was necessary.”
“Gee, Granddad, you’re a pippin!” exclaimed Randy, going up and
placing his arm around the old man’s shoulder. “A year’s vacation
would be all to the mustard.”
“It might be if you could only get rid of some of your slang in the
meantime,” put in his mother. Yet she had to smile as she spoke.
The boys were glad to get back into their old quarters, and in the
meantime Martha escorted Ruth to the room she was to occupy
during her visit. All the connecting doors of the three houses had
been thrown wide open, making the residences virtually one. While
this was going on Dick Rover hurried back to Wall Street, for
business with The Rover Company was brisk and he was needed at
the offices.
“You must be making a lot of money, Dad,” remarked Jack, as his
parent was leaving.
“Well, we’re holding our own, Jack,” was the reply.
“How are the oil wells making out?”
“Very fine.” Dick Rover stepped closer to his son. “Did you hear
anything from Carson Davenport?” he asked in a low tone so that the
others might not hear.
“Nothing since the girls met him. They said they sent word about
that.”
“You want to be very careful, Jack. We’ll talk the whole thing over
to-night. That rascal is certainly going to put one over on us if he
possibly can.”
“Why did they let him out of prison?”
“I don’t know. He may have got a number of important friends to
appear for him before a board of pardons, or something like that.
Then again, you must remember that what he was tried for was his
trouble with his partners. I did not want to appear against him
because it would have taken too much of my time, which, just then,
was very valuable to our concern. It’s possible that he got the very
people he swindled—or tried to swindle—to sign a petition in his
favor and in favor of his other partners, Tate and Jackson. But I must
hurry now. We can talk the whole thing over later.”
During the afternoon the twins went out to renew their
acquaintance with some of their former boy chums while Jack and
Fred accompanied the girls on a sightseeing and shopping
expedition.
“I’ll be awfully sorry to leave you, Ruth,” said Jack, when he got a
chance to speak to the visitor alone.
“Well, then you’d better stay,” she answered mischievously.
“Oh, you know I couldn’t do that,” he returned hastily. “What would
the other fellows say?”
“I was only joking, Jack. You go ahead and have your outing. I
hope you enjoy every minute of it. Only, please don’t get into any
trouble,” and the girl’s face clouded.
“I think we’ll be able to take care of ourselves, Ruth. And you take
care of yourself, too.”
“Are you going to write?”
“Sure I am! And I’ll expect you to answer, too. You will, won’t you?”
“Why, of course.”
There was a brief silence, neither of them seeming to know what
to say next. Then the former major of the Colby Hall battalion
stepped closer.
“I’m going to take that photograph of you along—you know, the
one you gave me some time ago,” he said in a low tone.
“Never!” she returned quickly. “Oh, Jack, suppose—suppose the
others saw it!”
“I don’t care! I’m going to take it,” he answered firmly.
“Well, if you’re set on it, I suppose I can’t stop you,” answered
Ruth. Her eyes were shining like stars. Then Jack caught her hand
and pressed it warmly just as the others came up and interrupted
what might have proved a very interesting tête-à-tête.
Dinner that evening was a grand affair, and Ruth, who sat next to
Jack, declared she had never enjoyed anything so much in all her
life. The twins and Tom Rover were full of fun, and Tom told several
stories which convulsed everybody with laughter.
“Gee, Dad, you’re a wonder!” breathed Randy, trying to stop
laughing. “I can see where Andy gets his wit from.”
“Yes, and I know where you get your habit of playing tricks from,”
put in his mother, gazing fondly at her husband.
“Now, now! No knocking!” cried Tom gayly. “The boys are just all
right! They may cut up a little now and then, but as they both bear
marks of their mother’s good looks, that will be forgiven them,” and
then Tom dodged back, as his wife made a move as if to pull his hair.
Ruth was quite a pianist and had cultivated that talent carefully
during her days at Clearwater Hall. After dinner Dora Rover insisted
that the girl give them some music. After playing one of her best
compositions Ruth gathered all the boys and girls around her and
they sang one popular song after another.
“A touch of old times, eh?” said Dick Rover to Dora, as, with his
arm around her waist, they surveyed the scene.
“It’s history repeating itself, Dick,” she answered. And then she
looked at her husband questioningly and nodded toward where Jack
was carefully turning the sheets of music for Ruth. “What do you
think of them?” she whispered.
“I think Jack is hit pretty hard,” he returned.
“Well, Ruth seems to be an awfully nice girl, Dick.”
“I agree. I wouldn’t ask for a better girl,” he answered.
“But Jack is so young!”
“He isn’t any younger than I was when I came after you and saved
your mother from old Crabtree.”
“Oh, well, that was different!” murmured Dora.
So far the boys had had no opportunity to speak to Tom Rover
about the proposed trip to the West. But soon the twins broached the
subject, and then the crowd around the piano broke up and Mary
and Martha retired, taking Ruth with them.
“We want to talk to the boys in the library,” said Tom Rover to his
wife and his sisters-in-law, and thereupon the ladies took the hint
and also left them.
“Now, Dad, tell us all about the Rolling Thunder mine and Sunset
Trail!” cried Randy. “Gee, I wish I was out there right now!”
“And on horseback!” put in his twin. “Say, we’ll have the best times
ever!”
“I certainly hope so,” returned Dick Rover. “At the same time, I
want to caution you.”
“Don’t scare the boys into fits, Dick,” said Tom. “You’ll spoil the
whole outing if you do.”
“I’m not going to scare them into fits, Tom,” answered the older
brother. “But I am going to give them some advice that I think they
ought to have.”
“I think so too,” came from Sam Rover. “If any fellow ever got on
my nerves, it’s that rascal, Carson Davenport.”
CHAPTER XIII
A NEW ACQUAINTANCE
The mention of Carson Davenport’s name made all the boys look
serious.
“Has that fellow made another demand?” questioned Jack quickly.
“Not directly,” answered his father. “But I have heard in an indirect
way, through a detective working for one of the local agencies, that
he is watching us very carefully. He has been seen in the vicinity of
our offices several times, and you have seen him twice in the vicinity
of Colby Hall and Clearwater Hall. That’s enough for me to realize
that the scoundrel means business.”
“You forgot to mention one thing, Dick,” came from Fred’s father.
“Another one of the detectives from that agency saw Davenport in
this vicinity less than three weeks ago.”
“What do you mean? Here at the houses?” questioned Randy.
“Yes. He was out on the Drive, skulking up and down looking at all
the doors and windows. And he asked one of the tradesmen who
lived here, evidently to make sure that he had the right place.”
“Why don’t they arrest him?” questioned Andy impatiently.
“That’s what we’re going to do as soon as we can get any real
evidence against him,” answered his Uncle Dick. “I’d like to catch
him red-handed at something.”
“I’ve got a scheme!” exclaimed Randy. “Jack, you’d be the fellow
to put it through because you’re Uncle Dick’s son and it’s Uncle Dick
that Davenport is sore on.”
“What’s the idea?” questioned his cousin.
“Lay a trap for Davenport by placing yourself in such a position
that he can get at you. Then, when he thinks he’s got you, let the

Normal Forms and Bifurcation of Planar Vector Fields 1st Edition Shui-Nee Chow 2024 scribd download

Uploaded by

Copyright:

Available Formats

Normal Forms and Bifurcation of Planar Vector Fields 1st Edition Shui-Nee Chow 2024 scribd download

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Normal Forms and Bifurcation of Planar Vector Fields 1st Edition Shui-Nee Chow 2024 scribd download

Uploaded by

Copyright:

Available Formats

Download the full version of the ebook at

Normal Forms and Bifurcation of Planar Vector

Explore and download more ebook at https://ebookultra.com

The Method of Normal Forms Second Edition Ali Hasan Nayfeh

Current Status of Carotid Bifurcation Angioplasty and

Vector Calculus 1st Edition R. Gupta

Routledge Handbook of the Chinese Economy 1st Edition

Nonlinear solid mechanics bifurcation theory and material

Stochastic Partial Differential Equations 2nd Edition Chow

Representing Wars from 1860 to the Present Fields of

Bifurcation and Chaos in Complex Systems 1st Edition J.Q.

Cambridge University Press

© Cambridge University Press 1994

This publication is in copyright. Subject to statutory exception

First published 1994

Library of Congress Cataloguing in Publication data

ISBN 978-0-521-37226-8 hardback

Transferred to digital printing 2007

Preface page vii

The theory of bifurcation of vector fields is the study of a family of

The main goal of this book is to study some bifurcation phenomena of

v,= (AEo(A)IRe A<0),

ar,: R" -> E Tr,: III" -p E, rru: R" - E.

1.1 Existence and Uniqueness of Global Center Manifolds

Consider the equation

where x E 1R", A E 5°(IP",18"), f E Ck(R ,18") for some k > 1, f(0) _

As usual, we denote by I y I the norm of y in some Banach space. Let

Cb(X,Y) _ {w E Ck(X,Y) IIWIIck := max suplDjw(x)I < oo

If X = Y, we write Cbk(X, X) as Cb(X ). We let

Finally, we denote by i(t, x) the solution of (1.1) with the initial

Theorem 1.1. (i) There is a positive number So which depends only on A

W`:= (xE08" suplkrhz(t, x)I < 0} (1.2)

is invariant under (1.1) and is a Lipschitz submanifold of 08"; more

(ii) If ¢ E Cb(Ec, Eh), and the set

is invariant under (1.1), then M. = W` and 0 = 41.

Definition 1.2. W` is called the global center manifold of (1.1).

Remark 1.3. If f E Cb(08"), then we will usually replace the condition

Remark 1.4. The uniqueness conclusion (ii) should be understood in

¢ E Cb(EC, Eh) by 0 E C°(E, Eh) unless AI E, is semisimple and f

leAtirsl <- Ke-Rt, t > 0, (1.5)

leAt7ruI < Kept, t < 0.

/3 =min{IReAllA E=- E au Uv51} -e,

Define a Banach space by

Cy :_ {x E C°(ll ,QBn) llxll7 := sup e-YItIlx(t)I <

The following lemma gives a different criterion for W`.

Lemma 1.6. Suppose f E Cb.1(Rn) and (1.6) is satisfied. Then

W` _ (x E Rn 1 1(., x) E CY). (1.7)

+ f teA(t-T)Truf(y(T))dT + f t eA(t-T)7r f(y(r))dr. (1.8)

W = {y(0) E IIB" I E C,, and satisfies (1.8) forsomex E V8"}.

Proof. By the variation of constants formula, for to, t E R we have

x(t, x) = eA(t-to)x(to, x) + foeA(t-T)f(x(T, x))dT. (1.10)

Denote by W` the right-hand side of (1.7), and by W` the right-hand

sup e-y't'I'rrhx(t, x)I < supITrhx(t, x)I < oo. (1.11)

Taking to = 0 in (1.10) we obtain

ir,x(t, x) = eAtircx + foteA(t-7)7rcf(x(T, x))dT. (1.12)

Using Lemma 1.5 and (1.6), we have from (1.12) that

sup e-y"t1Iirrx(t, x)I < 00. (1.13)

irux(t, x) = eA(t-to),rrux(to, x) + feA(t-T)Truf(x(T, x))dr. (1.14)

Fixing t E C$ and to >_ max(t, 0), we obtain from (1.5)

IeA(t-to)rrux(to, x)I < KeO(t-to)I x(to, x)I

-+G( )I 3 Ilx(', ) -x(., )Ily. (1.30)

IIx(', ) -x(., )IIy I

1x(t, ) -x(t, )I < (1.33)

u(t, , ) = x(t, ) - x(t, ), f, E E, (2.3)

f(t, , ) =f(x(t, )) -f(x(t, )) -Df(x(t,

INJ = foteA(t-T) 1 c[ f(x(T, )) -f(x(T, ))

-Df(x(T,4))]dA)(x(T,f) -x*(r, ))dT

x(f'IDf(Ax(T,6) + (1 - A)x(z,4)) IdA dT

x(r, 6 + Arl) - x (t, )

_ (f1Df(ox(t, 6 + A77) + (1 - 9)x(t, f))dO)g(t, A).