Dynamical Systems in Population Biology Xiao-

Qiang Zhao pdf download

https://textbookfull.com/product/dynamical-systems-in-population-
biology-xiao-qiang-zhao/

Download more ebook from https://textbookfull.com

We believe these products will be a great fit for you. Click
the link to download now, or visit textbookfull.com
to discover even more!

Dynamical systems in population biology Second Edition

Zhao

https://textbookfull.com/product/dynamical-systems-in-population-
biology-second-edition-zhao/

Synthetic Biology – Metabolic Engineering 1st Edition

Huimin Zhao

https://textbookfull.com/product/synthetic-biology-metabolic-
engineering-1st-edition-huimin-zhao/

Analysis of Chaotic Behavior in Non-linear Dynamical

Systems Micha■ Piórek

https://textbookfull.com/product/analysis-of-chaotic-behavior-in-
non-linear-dynamical-systems-michal-piorek/

Spaces of dynamical systems 2nd Edition Pilyugin S.Yu

https://textbookfull.com/product/spaces-of-dynamical-systems-2nd-
edition-pilyugin-s-yu/
Spaces of Dynamical Systems 2nd Edition Sergei Yu.
Pilyugin

https://textbookfull.com/product/spaces-of-dynamical-systems-2nd-
edition-sergei-yu-pilyugin/

Systems Biology Methods in Molecular Biology 2745

Mariano Bizzarri

https://textbookfull.com/product/systems-biology-methods-in-
molecular-biology-2745-mariano-bizzarri/

Complex Dynamical Systems in Education Concepts Methods

and Applications 1st Edition Matthijs Koopmans

https://textbookfull.com/product/complex-dynamical-systems-in-
education-concepts-methods-and-applications-1st-edition-matthijs-
koopmans/

Dynamical Systems in Theoretical Perspective ■ód■

Poland December 11 14 2017 Jan Awrejcewicz

https://textbookfull.com/product/dynamical-systems-in-
theoretical-perspective-lodz-poland-december-11-14-2017-jan-
awrejcewicz/

Dynamical Systems and Geometric Mechanics An

Introduction Jared Michael Maruskin

https://textbookfull.com/product/dynamical-systems-and-geometric-
mechanics-an-introduction-jared-michael-maruskin/
CMS Books in Mathematics

Canadian
Mathematical Society
Société mathématique

Xiao-Qiang Zhao du Canada

Dynamical Systems
in Population
Biology
Second Edition
 Canadian Mathematical Society
Société mathématique du Canada

Editors-in-Chief
Rédacteurs-en-chef
K. Dilcher
K. Taylor

Advisory Board
Comité consultatif
M. Barlow
H. Bauschke
L. Edelstein-Keshet
N. Kamran
M. Kotchetov

More information about this series at http://www.springer.com/series/4318

Xiao-Qiang Zhao

Dynamical Systems
in Population Biology
Second Edition

123
Xiao-Qiang Zhao
Department of Mathematics and Statistics
Memorial University of Newfoundland
St. John’s, NL, Canada

ISSN 1613-5237 ISSN 2197-4152 (electronic)

CMS Books in Mathematics
ISBN 978-3-319-56432-6 ISBN 978-3-319-56433-3 (eBook)
DOI 10.1007/978-3-319-56433-3

Library of Congress Control Number: 2017938615

Mathematics Subject Classification (2010): 34Cxx, 34Kxx, 35Bxx, 35R10, 37Bxx, 37Cxx, 37N25,
39A05, 92D25, 92D30, 92D40

© Springer International Publishing AG 2003, 2017

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology
now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the
editors give a warranty, express or implied, with respect to the material contained herein or for any errors
or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims
in published maps and institutional affiliations.

Printed on acid-free paper

This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
To Hong and Bob
Preface

Population dynamics is an important subject in mathematical biology.

A central problem is to study the long-term behavior of modeling systems.
Most of these systems are governed by various evolutionary equations such
as difference, ordinary, functional, and partial differential equations (see, e.g.,
[253, 206, 334, 167, 77]). As we know, interactive populations often live in a
fluctuating environment. For example, physical environmental conditions such
as temperature and humidity and the availability of food, water, and other
resources usually vary in time with seasonal or daily variations. Therefore,
more realistic models should be nonautonomous systems. In particular, if the
data in a model are periodic functions of time with commensurate period,
a periodic system arises; if these periodic functions have different (minimal)
periods, we get an almost periodic system. The existing reference books, from
the dynamical systems point of view, mainly focus on autonomous biological
systems. The book of Hess [152] is an excellent reference for periodic parabolic
boundary value problems with applications to population dynamics. Since the
publication of this book, there have been extensive investigations on periodic,
asymptotically periodic, almost periodic, and even general nonautonomous
biological systems, which in turn have motivated further development of the
theory of dynamical systems.
In order to explain the dynamical systems approach to periodic population
problems, let us consider, as an illustration, two species periodic competitive
systems

du1
= f1 (t, u1 , u2 ),
dt (0.1)
du2
= f2 (t, u1 , u2 ),
dt
where f1 and f2 are continuously differentiable and ω-periodic in t, and
∂fi /∂uj ≤ 0, i = j. We assume that, for each v ∈ R2 , the unique solution
u(t, v) of system (0.1) satisfying u(0) = v exists globally on [0, ∞).

vii
viii Preface

Let X = R2 , and define a family of mappings T (t) : X → X, t ≥ 0, by

T (t)x = u(t, x), ∀x ∈ X. It is easy to see that T (t) satisfies the following
properties:
(1) T (0) = I, where I is the identity map on X.
(2) T (t + ω) = T (t) ◦ T (ω), ∀t ≥ 0.
(3) T (t)x is continuous in (t, x) ∈ [0, ∞) × X.
T (t) is called the periodic semiflow generated by periodic system (0.1), and
P := T (ω) is called its associated Poincaré map (or period map). Clearly,
P n v = u(nω, v), ∀n ≥ 1, v ∈ R2 . It then follows that the study of the dy-
namics of (0.1) reduces to that of the discrete dynamical system {P n } on
R2 .
If u = (u1 , u2 ), v = (v1 , v2 ) ∈ R2 , then we write u ≤ v whenever ui ≤ vi
holds for i = 1, 2. We write u ≤K v whenever u1 ≤ v1 and u2 ≥ v2 . By
the well-known Kamke comparison theorem, it follows that the following key
properties hold for competitive system (0.1) (see, e.g., [334, Lemma 7.4.1]):
(P1) If u ≤K v, then P u ≤K P v.
(P2) If P u ≤ P v, then u ≤ v.
Then the Poincaré map P , and hence the discrete dynamical system {P n },
is monotone with respect to the order ≤K on R2 . Consequently, system (0.1)
admits convergent dynamics (see [334, Theorem 7.4.2]).

Theorem Every bounded solution of a competitive planar periodic system

asymptotically approaches a periodic solution.

We use the proof provided in [334, Theorem 7.4.2]. Indeed, it suffices to prove
that every bounded orbit of {P n } converges to a fixed point of P . Given two
points u, v ∈ R2 , one or more of the four relations u ≤ v, v ≤ u, u ≤K v,
v ≤K u must hold. Now, if P n0 u0 ≤K P n0 +1 u0 (or the reverse inequality)
holds for some n0 ≥ 0, then (P1) implies that P n u0 ≤K P n+1 u0 (or the
reverse inequality) holds for all n ≥ n0 . Therefore, {P n u0 } converges to some
fixed point ū, since the sequence is bounded and eventually monotone. The
proof is complete in this case, so we assume that there does not exist such an
n0 as just described. In particular, it follows that u0 is not a fixed point of
P . Then it follows that for each n we must have either P n+1 u0 ≤ P n u0 or
the reverse inequality. Suppose for definiteness that u0 ≤ P u0 , the other case
being similar. We claim that P n u0 ≤ P n+1 u0 for all n. If not, there exists n0
such that
u0 ≤ P u0 ≤ P 2 u0 ≤ · · · ≤ P n0 −1 u0 ≤ P n0 u0
but P n0 u0 ≥ P n0 +1 u0 . Clearly, n0 ≥ 1 since u0 ≤ P u0 . Applying (P2) to the
displayed inequality yields P n0 −1 u0 ≥ P n0 u0 and therefore P n0 −1 u0 = P n0 u0 .
Since P is one to one, u0 must be a fixed point, in contradiction to our
assumption. This proves the claim and implies that the sequence {P n u0 }
converges to some fixed point ū.
 Preface ix

It is hoped that the reader will appreciate the elegance and simplicity of
the arguments supporting the above theorem, which are motivated by a now
classical paper of deMottoni and Schiaffino [92] for the special case of periodic
Lotka–Volterra systems. This example also illustrates the roles that Poincaré
maps and monotone discrete dynamical systems may play in the study of pe-
riodic systems. For certain nonautonomous perturbations of a periodic system
(e.g., an asymptotically periodic system), one may expect that the Poincaré
map associated with the unperturbed periodic system (e.g., the limiting pe-
riodic system) should be very helpful in understanding the dynamics of the
original system. For a nonperiodic nonautonomous system (e.g., almost peri-
odic system), we are not able to define a continuous or discrete-time dynamical
system on its state space. The skew-product semiflow approach has proved to
be very powerful in obtaining dynamics for certain types of nonautonomous
systems (see, e.g., [303, 300, 311]).
The main purpose of this book is to provide an introduction to the the-
ory of periodic semiflows on metric spaces and its applications to population
dynamics. Naturally, the selection of the material is highly subjective and
largely influenced by my personal interests. In fact, the contents of this book
are predominantly from my own and my collaborators’ recent works. Also, the
list of references is by no means exhaustive, and I apologize for the exclusion
of many other related works.
Chapter 1 is devoted to abstract discrete dynamical systems on metric
spaces. We study global attractors, chain transitivity, strong repellers, and
perturbations. In particular, we will show that a dissipative, uniformly persis-
tent, and asymptotically compact system must admit a coexistence state. This
result is very useful in proving the existence of (all or partial componentwise)
positive periodic solutions of periodic evolutionary systems.
The focus of Chapter 2 is on global dynamics in certain types of monotone
discrete dynamical systems on ordered Banach spaces. Here we are interested
in the abstract results on attracting order intervals, global attractivity, and
global convergence, which may be easily applied to various population models.
In Chapter 3, we introduce the concept of periodic semiflows and prove
a theorem on the reduction of uniform persistence to that of the associated
Poincaré map. The asymptotically periodic semiflows, nonautonomous semi-
flows, skew-product semiflows, and continuous processes are also discussed.
In Chapter 4, as a first application of the previous abstract results, we
analyze in detail a discrete-time, size-structured chemostat model that is de-
scribed by a system of difference equations, although in this book our main
concern is with global dynamics in periodic and almost periodic systems. The
reason for this choice is that we want to show how the theory of discrete dy-
namical systems can be applied to discrete-time models governed by difference
equations (or maps).
In the rest of the book, we apply the results in Chapters 1–3 to continuous-
time periodic population models: in Chapter 5 to the N -species competition
in a periodic chemostat, in Chapter 6 to almost periodic competitive systems,
x Preface

in Chapter 7 to competitor–competitor–mutualist parabolic systems, and in

Chapter 8 to a periodically pulsed bioreactor model. Of course, for each chap-
ter, we need to use different qualitative methods and even to develop certain
ad hoc techniques.
Chapter 9 is devoted to the global dynamics in an autonomous, nonlocal,
and delayed predator–prey model. Clearly, the continuous-time analogues of
the results in Chapters 1 and 2 can find applications in autonomous models.
Note that an autonomous semiflow can be viewed as a periodic one with the
period being any fixed positive real number, and hence it is possible to get
some global results by using the theory of periodic semiflows. However, we
should point out that there do exist some special theory and methods that
are applicable only to autonomous systems. The fluctuation method in this
chapter provides such an example.
The existence, attractivity, uniqueness, and exponential stability of pe-
riodic traveling waves in periodic reaction–diffusion equations with bistable
nonlinearities are discussed in Chapter 10, which is essentially independent of
the previous chapters. We appeal only to a convergence theorem from Chap-
ter 2 to prove the attractivity and uniqueness of periodic waves. Here the
Poincaré-type map associated with the system plays an important role once
again.
Over the years, I have benefited greatly from the communications, dis-
cussions, and collaborations with many colleagues and friends in the fields of
differential equations, dynamical systems, and mathematical biology, and I
would like to take this opportunity to express my gratitude to all of them. I
am particularly indebted to Herb Freedman, Morris Hirsch, Hal Smith, Horst
Thieme, Gail Wolkowicz, and Jianhong Wu, with whom I wrote research ar-
ticles that are incorporated in the present book.
Finally, I gratefully appreciate, financial support for my research from the
National Science Foundation of China, the Royal Society of London, and the
Natural Sciences and Engineering Research Council of Canada.
Preface to the Second Edition

For this edition, I have corrected some typos, revised Sections 1.1 and 1.3.4 by
using the concepts of global attractors and strong global attractors, deleted
the original subsection on order persistence from Section 1.3, and added three
new sections about persistence and attractors (Section 1.3.3), saddle point
behavior for monotone semiflows (Section 2.5), and solution maps of abstract
nonautonomous functional differential equations (Section 3.5), respectively. I
have also mentioned more related references in the notes sections of Chap-
ters 1–10.
In addition, I have added four new chapters. Chapter 11 is devoted to the
general theory of basic reproduction ratios R0 for compartmental models of
periodic functional differential equations and autonomous reaction–diffusion
systems. Chapter 12 deals with the threshold dynamics in terms of R0 for a
new class of population models with time periodic delays. In Chapter 13, we
study a periodic reaction–diffusion SIS system and investigate the effect of
spatial and temporal heterogeneities on the extinction and persistence of the
infectious disease. The final chapter, Chapter 14, provides a complete analysis
of the disease-free dynamics and global dynamics for a nonlocal spatial model
of Lyme disease. It is expected that Chapters 12–14 may serve as templates for
future investigations on other population models with spatial and temporal
heterogeneities.
My sincere thanks goes to Jifa Jiang, Xing Liang, Yijun Lou, Pierre Magal,
Rui Peng, Wendi Wang, and Xiao Yu, whose joint research articles with me
have been incorporated in the second edition. I am also very grateful to all
collaborators and friends for their encouragements, suggestions, and assistance
with this revision.

St. John’s, NL, Canada Xiao-Qiang Zhao

xi
Contents

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii

1 Dissipative Dynamical Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.1 Limit Sets and Global Attractors . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Chain Transitivity and Attractivity . . . . . . . . . . . . . . . . . . . . . . . . 9
1.2.1 Chain Transitive Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.2.2 Attractivity and Morse Decompositions . . . . . . . . . . . . . . 16
1.3 Strong Repellers and Uniform Persistence . . . . . . . . . . . . . . . . . . 19
1.3.1 Strong Repellers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.3.2 Uniform Persistence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.3.3 Persistence and Attractors . . . . . . . . . . . . . . . . . . . . . . . . . . 26
1.3.4 Coexistence States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
1.4 Persistence Under Perturbations . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
1.4.1 Perturbation of a Globally Stable Steady State . . . . . . . . 33
1.4.2 Persistence Uniform in Parameters . . . . . . . . . . . . . . . . . . 34
1.4.3 Robust Permanence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
1.5 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

2 Monotone Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.1 Attracting Order Intervals and Connecting Orbits . . . . . . . . . . . 44
2.2 Global Attractivity and Convergence . . . . . . . . . . . . . . . . . . . . . . . 48
2.3 Subhomogeneous Maps and Skew-Product Semiflows . . . . . . . . . 52
2.4 Competitive Systems on Ordered Banach Spaces . . . . . . . . . . . . 59
2.5 Saddle Point Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
2.6 Exponential Ordering Induced Monotonicity . . . . . . . . . . . . . . . . 69
2.7 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

3 Nonautonomous Semiflows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.1 Periodic Semiflows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
3.1.1 Reduction to Poincaré Maps . . . . . . . . . . . . . . . . . . . . . . . . 78
3.1.2 Monotone Periodic Systems . . . . . . . . . . . . . . . . . . . . . . . . . 80
xiii
xiv Contents

3.2 Asymptotically Periodic Semiflows . . . . . . . . . . . . . . . . . . . . . . . . . 87

3.2.1 Reduction to Asymptotically Autonomous Processes . . . 88
3.2.2 Asymptotically Periodic Systems . . . . . . . . . . . . . . . . . . . . 90
3.3 Monotone and Subhomogeneous Almost Periodic Systems . . . . 98
3.4 Continuous Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
3.5 Abstract Nonautonomous FDEs . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
3.6 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116

4 A Discrete-Time Chemostat Model . . . . . . . . . . . . . . . . . . . . . . . . 119

4.1 The Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
4.2 The Limiting System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
4.3 Global Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
4.4 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

5 N -Species Competition in a Periodic Chemostat . . . . . . . . . . . 131

5.1 Weak Periodic Repellers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
5.2 Single Population Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
5.3 N -Species Competition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
5.4 3-Species Competition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
5.5 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152

6 Almost Periodic Competitive Systems . . . . . . . . . . . . . . . . . . . . . 155

6.1 Almost Periodic Attractors in Scalar Equations . . . . . . . . . . . . . 156
6.2 Competitive Coexistence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
6.3 An Almost Periodic Chemostat Model . . . . . . . . . . . . . . . . . . . . . 169
6.4 Nonautonomous 2-Species Competitive Systems . . . . . . . . . . . . . 174
6.5 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180

7 Competitor–Competitor–Mutualist Systems . . . . . . . . . . . . . . . 181

7.1 Weak Periodic Repellers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
7.2 Competitive Coexistence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
7.3 Competitive Exclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
7.4 Bifurcations of Periodic Solutions: A Case Study . . . . . . . . . . . . 197
7.5 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210

8 A Periodically Pulsed Bioreactor Model . . . . . . . . . . . . . . . . . . . 213

8.1 The Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
8.2 Unperturbed Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
8.2.1 Conservation Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
8.2.2 Single Species Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
8.2.3 Two-Species Competition . . . . . . . . . . . . . . . . . . . . . . . . . . 224
8.3 Perturbed Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
8.3.1 Periodic Systems with Parameters . . . . . . . . . . . . . . . . . . . 229
8.3.2 Single Species Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
8.3.3 Two-Species Competition . . . . . . . . . . . . . . . . . . . . . . . . . . 236
8.4 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
 Contents xv

9 A Nonlocal and Delayed Predator–Prey Model . . . . . . . . . . . . 241

9.1 The Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
9.2 Global Coexistence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
9.3 Global Extinction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
9.4 Global Attractivity: A Fluctuation Method . . . . . . . . . . . . . . . . . 252
9.5 Threshold Dynamics: A Single Species Model . . . . . . . . . . . . . . . 255
9.6 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263

10 Traveling Waves in Bistable Nonlinearities . . . . . . . . . . . . . . . . . 265

10.1 Existence of Periodic Traveling Waves . . . . . . . . . . . . . . . . . . . . . . 266
10.2 Attractivity and Uniqueness of Traveling Waves . . . . . . . . . . . . . 272
10.3 Exponential Stability of Traveling Waves . . . . . . . . . . . . . . . . . . . 277
10.4 Autonomous Case: A Spruce Budworm Model . . . . . . . . . . . . . . . 281
10.5 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283

11 The Theory of Basic Reproduction Ratios . . . . . . . . . . . . . . . . . 285

11.1 Periodic Systems with Time Delay . . . . . . . . . . . . . . . . . . . . . . . . . 286
11.2 A Periodic SEIR Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
11.3 Reaction–Diffusion Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 302
11.4 A Spatial Model of Rabies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
11.5 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315

12 A Population Model with Periodic Delay . . . . . . . . . . . . . . . . . . 317

12.1 Model Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
12.2 Threshold Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
12.3 Numerical Computation of R0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
12.4 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336

13 A Periodic Reaction–Diffusion SIS Model . . . . . . . . . . . . . . . . . . 337

13.1 Basic Reproduction Ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
13.2 Threshold Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
13.3 Global Attractivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 354
13.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 358
13.5 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359

14 A Nonlocal Spatial Model for Lyme Disease . . . . . . . . . . . . . . . 361

14.1 The Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362
14.2 Disease-Free Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
14.3 Global Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373
14.4 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 384

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 385

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 411
1
Dissipative Dynamical Systems

There are many discrete-time population models governed by difference equa-

tions (or maps), and as we mentioned in the Preface, the dynamics of a peri-
odic differential system can be investigated via its associated Poincaré map.
The aim of this chapter is to introduce basic definitions and develop main
tools in the theory of discrete dynamical systems. In Section 1.1 we present
concepts of limit sets and attractors and some fundamental theorems such as
the LaSalle invariance principle, the asymptotic fixed point theorem, and the
global attractor theorems.
Chain transitivity has remarkable connections to the structure of attrac-
tors. In Section 1.2 we first give typical examples and characteristics of chain
transitive sets. Then we show that the Butler–McGehee properties of omega
limit sets are shared by chain transitive sets for a dynamical system, which
enable us to obtain further important properties of chain transitive sets such
as strong attractivity and convergence, and to prove the equivalence between
acyclic coverings and Morse decompositions.
Uniform persistence is an important concept in population dynamics, since
it characterizes the long-term survival of some or all interacting species in an
ecosystem. Looked at abstractly, it is the notion that a closed subset of the
state space is repelling for the dynamics on the complementary set, and then
it gives a uniform estimate for omega limit sets, which sometimes is essen-
tial to obtain a more detailed global dynamics. In Section 1.3 we prove a
strong repeller theorem in terms of chain transitive sets, which unifies earlier
results on uniform persistence, and implies robustness of uniform persistence.
Then we show that a dissipative, uniformly persistent, and asymptotically
compact system must have at least one coexistence steady state, which pro-
vides a dynamic approach to some static problems (e.g., existence of positive
steady states and periodic solutions). We also introduce the concept of gener-
alized distance functions in abstract persistence theory so that the practical
persistence can be easily obtained for certain infinite-dimensional biological
systems.

© Springer International Publishing AG 2017 1

X.-Q. Zhao, Dynamical Systems in Population Biology, CMS Books
in Mathematics, DOI 10.1007/978-3-319-56433-3 1
2 1 Dissipative Dynamical Systems

In Section 1.4 we discuss persistence under perturbations. We prove a gen-

eral result on the perturbation of a globally stable steady state. Then we prove
uniform persistence uniform in parameters, which is very useful in establish-
ing the robustness of global asymptotic stability of an equilibrium solution. A
dissipative and uniformly persistent system is often said to be permanent. For
a class of autonomous Kolmogorov systems of ordinary differential equations
we also obtain a robust permanence theorem.

1.1 Limit Sets and Global Attractors

Let N be the set of integers and N+ the set of nonnegative integers. Let X be
a complete metric space with metric d and f : X → X a continuous map. For
a nonempty invariant set M (i.e., f (M ) = M ), the set W s (M ) := {x ∈ X :
limn→∞ d(f n (x), M ) = 0} is called the stable set of M . The omega limit set of
x is defined in the usual way as ω(x) = {y ∈ X : f nk (x) → y, for some nk →
∞}. A negative orbit through x = x0 is a sequence γ − (x) = {xk }0k=−∞ such
that f (xk−1 ) = xk for integers k ≤ 0. There may be no negative orbit through
x, and even if there is one, it may not be unique. Of course, a point of an
invariant set always has at least one negative orbit contained in the invariant
set. For a given negative orbit γ − (x) we define its alpha limit set as α(γ − ) =
{y ∈ X : xnk → y for some nk → −∞}. If γ + (x) = {f n (x) : n ≥ 0} (γ − (x))
is precompact (i.e., it is contained in a compact set), then ω(x) (α(γ − )) is
nonempty, compact, and invariant(see, e.g., [141, Lemma 2.1.2]).
Let e ∈ X be a fixed point of f (i.e., f (e) = e). Recall that e is said to be
stable for f : X → X if for each  > 0 there exists δ > 0 such that for any
x ∈ X with d(x, e) < δ, we have d(f n (x), e) < , ∀n ≥ 0. The following simple
observation is useful in proving the convergence of a precompact positive orbit
to a fixed point.
Lemma 1.1.1. (Convergence) Let e be a stable fixed point and γ + (x) a
precompact positive orbit for f : X → X. If e ∈ ω(x), then ω(x) = {e}.
Proof. Let  > 0 be given. By stability of e for f : X → X, there exists δ > 0
such that for any y ∈ X with d(y, e) < δ, we have d(f m (y), e) < , ∀m ≥ 0.
Since e ∈ ω(x), there is a subsequence nk → ∞ with f nk (x) → e, and hence
an index k0 such that d(f nk0 (x), e) < δ. Thus d(f nk0 +m (x), e) < , ∀m ≥ 0,
which implies that ω(x) ⊂ {z ∈ X : d(z, e) ≤ }, ∀ > 0. Letting  → 0, we
get ω(x) = {e}.
Definition 1.1.1. Let G be a closed subset of X. A continuous function V :
G → R is said to be a Liapunov function on G of the map f : G → G
(or the discrete system xn+1 = f (xn ), n ≥ 0) if V̇ (x) := V (f (x)) − V (x) ≤ 0
for all x ∈ G.
Theorem 1.1.1. (LaSalle invariance principle) Assume that V is a Li-
apunov function on G of f , and that γ + (x) is a precompact orbit of f and
 1.1 Limit Sets and Global Attractors 3

γ + (x) ⊂ G. Then ω(x) ⊂ M ∩ V −1 (c) for some c = c(x), where M is the

largest invariant set in E := {x ∈ G : V̇ (x) = 0}, and V −1 (c) := {x ∈ G :
V (x) = c}.
Proof. Clearly, the continuous function V is bounded on the compact set
γ + (x) ⊂ G. Let xn = f n (x), n ≥ 0. Then
V (xn+1 ) − V (xn ) = V (f (xn )) − V (xn ) = V̇ (xn ) ≤ 0,
and hence V (xn ) is nonincreasing with respect to n and is bounded from
below. Therefore, there is a real number c = c(x) such that limn→∞ V (xn ) = c.
For any y ∈ ω(x) ⊂ G, there is a sequence nk → ∞ such that limk→∞ xnk = y.
Since V is continuous, limk→∞ V (xnk ) = V (y) = c, and ω(x) ⊂ V −1 (c). Since
ω(x) is invariant, V (f (y)) = c and V̇ (y) = 0. Therefore, ω(x) ⊂ E, and hence
ω(x) ⊂ M .
Recall that a set U in X is said to be a neighborhood of another set V
provided that V is contained in the interior int(U ) of U . For any subsets
A, B ⊂ X and any  > 0, we define
d(x, A) := inf d(x, y), δ(B, A) := sup d(x, A),
y∈A x∈B

N (A, ) := {x ∈ X : d(x, A) < } and N (A, ) := {x ∈ X : d(x, A) ≤ } .

The Kuratowski measure of noncompactness, α, is defined by
α(B) = inf{r : B has a finite open cover of diameter ≤ r},
for any bounded set B of X. We set α(B) = +∞, whenever B is unbounded.
For various properties of Kuratowski’s measure of noncompactness, we
refer to [242, 91] and [304, Lemma 22.2]. The proof of the following lemma is
straightforward.
Lemma 1.1.2. The following statements are valid:
(a) Let I ⊂ [0, +∞) be unbounded, and {At }t∈I be a nonincreasing family
of nonempty closed subsets (i.e., t ≤ s 
implies As ⊂ At ). Assume that
α(At ) → 0, as t → +∞. Then A∞ = At is nonempty and compact,
t≥0
and δ(At , A∞ ) → 0, as t → +∞.
(b) For each A ⊂ X and B ⊂ X, we have α(B) ≤ α(A) + δ(B, A).
 m
For a subset B ⊂ X, let γ + (B) := f (B) be the positive orbit of B
m≥0
for f , and
 
ω(B) := f m (B)
n≥0 m≥n

the omega limit set of B. A subset A ⊂ X is positively invariant for f if

f (A) ⊂ A. We say that a subset A ⊂ X attracts a subset B ⊂ X for f if
limn→∞ δ(f n (B), A) = 0.
4 1 Dissipative Dynamical Systems

It is easy to see that B is precompact (i.e., B is compact) if and only if

α(B) = 0. A continuous mapping f : X → X is said to be compact (com-
pletely continuous) if f maps any bounded set to a precompact set in X.
The theory of attractors is based on the following fundamental result,
which is related to [141, Lemmas 2.1.1 and 2.1.2].
Lemma 1.1.3. Let B be a subset of X and assume that there exists a compact
subset C of X which attracts B for f . Then ω(B) is nonempty, compact,
invariant for f and attracts B.

Proof. Let I = N+ , the set of all nonnegative integers, and


An = f m (B), ∀n ≥ 0.
m≥n

Since C attracts B, from Lemma 1.1.2 (b) we deduce that

α(An ) ≤ α(C) + δ(An , C) = δ(An , C) → 0, as n → +∞.

So the family {An }n≥0 satisfies the conditions of assertion (a) in Lemma 1.1.2,
and we deduce that ω(B) is nonempty, compact, and δ(An , ω(B)) → 0, as
n → +∞. So ω(B) attracts B for f . Moreover, we have
⎛ ⎞
 
f⎝ f m (B)⎠ = f m (B), ∀n ≥ 0,
m≥n m≥n+1

and since f is continuous, we obtain

f (An ) ⊂ An+1 , and An+1 ⊂ f (An ), ∀n ≥ 0.

Finally, since δ(An , ω(B)) → 0, as n → +∞, we have f (ω(B)) = ω(B).

Definition 1.1.2. A continuous mapping f : X → X is said to be point

(compact, bounded) dissipative if there is a bounded set B0 in X such that
B0 attracts each point (compact set, bounded set) in X; α-condensing (α-
contraction of order k, 0 ≤ k < 1) if f takes bounded sets to bounded sets
and α(f (B)) < α(B) (α(f (B)) ≤ kα(B)) for any nonempty closed bounded
set B ⊂ X with α(B) > 0; α-contracting if limn→∞ α (f n (B)) = 0 for any
bounded subset B ⊂ X; asymptotically smooth if for any nonempty closed
bounded set B ⊂ X for which f (B) ⊂ B, there is a compact set J ⊂ B such
that J attracts B.

Clearly, a compact map is an α-contraction of order 0, and an α-contraction

of order k is α-condensing. It is well known that α-condensing maps are asymp-
totically smooth (see, e.g., [141, Lemma 2.3.5]). By Lemma 1.1.2, it follows
that f : X → X is asymptotically smooth if and only if limn→∞ α (f n (B)) = 0
 1.1 Limit Sets and Global Attractors 5

for any nonempty closed bounded subset B ⊂ X for which f (B) ⊂ B. This
implies that any α-contracting map is asymptotically smooth.
A positively invariant subset B ⊂ X for f is said to be stable if for any
neighborhood V of B, there exists a neighborhood U ⊂ V of B such that
f n (U ) ⊂ V, ∀n ≥ 0. We say that A is globally asymptotically stable for f if,
in addition, A attracts points of X for f .
By the proof that (i) implies (ii) in [141, Theorem 2.2.5], we have the
following result.
Lemma 1.1.4. Let B ⊂ X be compact, and positively invariant for f . If B
attracts compact subsets of some neighborhood of itself, then B is stable.
Definition 1.1.3. A nonempty, compact, and invariant set A ⊂ X is said to
be an attractor for f if A attracts some open neighborhood of itself; a global
attractor for f if A is an attractor that attracts every point in X; and a strong
global attractor for f if A attracts every bounded subset of X.
Remark 1.1.1. The notion of attractor and global attractor was used in
[164, 304]. The strong global attractor was defined as global attractor in
[141, 358, 286]. In the case where the dimension of X is finite, it is easy to see
that both global attractor and strong global attractor are equivalent. In the
infinite-dimensional case of X, however, there exist discrete- and continuous-
time dynamical systems that admit global attractors, but no strong global
attractors, see Example 1.3.3 and [241, Sections 5.1–5.3].
The following result is essentially the same as [142, Theorem 3.2]. Note
that the proof of this result was not provided in [142]. For completeness, we
state it in terms of global attractors and give an elementary proof below.
Theorem 1.1.2. (Global Attractors) Let f : X → X be a continuous
map. Assume that
(a) f is point dissipative and asymptotically smooth;
(b) Positive orbits of compact subsets of X for f are bounded.
Then f has a global attractor A ⊂ X. Moreover, if a subset B of X admits
the property that γ + (f k (B)) is bounded for some k ≥ 0, then A attracts B
for f .
Proof. Assume that (a) is satisfied. Since f is point dissipative, we can find
a closed and bounded subset B0 in X such that for each x ∈ X, there exists
k = k(x) ≥ 0, f n (x) ∈ B0 , ∀n ≥ k. Define

J(B0 ) := {y ∈ B0 : f n (y) ∈ B0 , ∀n ≥ 0} .

Thus, f (J(B0 )) ⊂ J(B0 ), and for every x ∈ X, there exists k = k(x) ≥ 0 such
that f k (x) ∈ J(B0 ). Since J(B0 ) is closed and bounded, and f is asymptot-
ically smooth, Lemma 1.1.3 implies that ω(J(B0 )) is compact invariant, and
attracts points of X.
6 1 Dissipative Dynamical Systems

Assume, in addition, that (b) is satisfied. We claim that there exists an ε >
0 such that γ + (N (ω(J(B0 )), ε)) is bounded. Assume, by contradiction, that
1
γ + N ω(J(B0 )), n+1 is unbounded for each n > 0. Let z ∈ X be fixed.
1
Then we can find a sequence xn ∈ N ω(J(B0 )), n+1 , and a sequence of
integers mn ≥ 0 such that d(z, f mn
(xn )) ≥ n. Since ω(J(B0 )) is compact,
we can always assume that xn → x ∈ ω(J(B)), as n → +∞. Since H :=
{xn : n ≥ 0}∪{x} is compact, assumption (b) implies that γ + (H) is bounded,
a contradiction. Let D = γ + (N (ω(J(B0 )), ε)). Then D is closed, bounded,
and positively invariant for f . Since ω(J(B0 )) attracts points of X for f , and
ω(J(B0 )) ⊂ N (ω(J(B0 )), ε) ⊂ int(D), we deduce that for each x ∈ X, there
exists k = k(x) ≥ 0 such that f k (x) ∈ int(D). It then follows that for each
compact subset C of X, there exists an integer k ≥ 0 such that f k (C) ⊂ D.
Thus, the set A := ω(D) attracts every compact subset of X. Fix a bounded
neighborhood V of A. By Lemma 1.1.4, it follows that A is stable, and hence,
there is a neighborhood W of A such that f n (W ) ⊂ V, ∀n ≥ 0. Clearly, the
set U := ∪n≥0 f n (W ) is a bounded neighborhood of A, and f (U ) ⊂ U . Since
f is asymptotically smooth, there is a compact set J ⊂ U such that J attracts
U . By Lemma 1.1.3, ω(U ) is nonempty, compact, invariant for f , and attracts
U . Since A attracts ω(U ), we have ω(U ) ⊂ A. Thus, A is a global attractor
for f .
To prove the last part of the theorem, without loss of generality we assume
that B is a bounded subset of X and γ + (B) is bounded. We set K = γ + (B).
Then f (K) ⊂ K. Since K is bounded and f is asymptotically smooth, there
exists a compact C which attracts K for f . Note that f k (B) ⊂ f k (γ + (B)) ⊂
f k (K) , ∀k ≥ 0. Thus, C attracts B for f . By Lemma 1.1.3, we deduce that
ω(B) is nonempty, compact, invariant for f and attracts B. Since A is a
global attractor for f , it follows that A attracts compact subsets of X. By the
invariance of ω(B) for f , we deduce that ω(B) ⊂ A, and hence, A attracts B
for f .

Remark 1.1.2. From the first part of the proof of Theorem 1.1.2, it is easy to
see that if f is point dissipative and asymptotically smooth, then there exists
a nonempty, compact, and invariant subset C of X for f such that C attracts
every point in X for f .

The following lemma provides sufficient conditions for the positive orbit
of a compact set to be bounded.

Lemma 1.1.5. Assume that f is point dissipative. If C is a compact subset

of X with the property that for every bounded sequence {xn }n≥0 in γ + (C),
{xn }n≥0 or {f (xn )}n≥0 has a convergent subsequence, then γ + (C) is bounded
in X.

Proof. Since f is point dissipative, we can choose a bounded and open subset
V of X such that for each x ∈ X there exists n0 = n0 (x) ≥ 0 such that
 1.1 Limit Sets and Global Attractors 7

f n (x) ∈ V, ∀n ≥ n0 . By the continuity of f and the compactness of C, it

follows that there exists a positive integer r = r(C) such that for any x ∈ C,
there exists an integer k = k(x) ≤ r such that f k (x) ∈ V . Let z ∈ X be
fixed. Assume, by contradiction, that γ + (C) is unbounded. Then there exists
a sequence {xp } in γ + (C) such that

xp = f mp (zp ), zp ∈ C, and lim d(z, xp ) = ∞.

p→∞

Since f is continuous and C is compact, without loss of generality we can

assume that

lim mp = ∞, and mp > r, xp ∈

/ V, ∀p ≥ 1.
p→∞

For each zp ∈ C, there exists an integer kp ≤ r such that f kp (zp ) ∈ V . Since

xp = f mp (zp ) ∈
/ V , there exists an integer np ∈ [kp , mp ) such that

yp = f np (zp ) ∈ V, and f l (yp ) ∈

/ V, ∀1 ≤ l ≤ lp = mp − np .

Clearly, xp = f lp (yp ), ∀p ≥ 1, and {yp } is a bounded sequence in γ + (C).

We only consider the case where {yp } has a convergent subsequence since
the proof for the case where {f (yp )} has a convergent subsequence is similar.
Thus, without loss of generality we can assume that limp→∞ yp = y ∈ V .
In the case where the sequence {lp } is bounded, there exist an integer l̂ and
sequence pk → ∞ such that lpk = l̂, ∀k ≥ 1, and hence,

d(z, f l̂ (y)) = lim d(z, f l̂ (ypk )) = lim d(z, xpk ) = ∞,

k→∞ k→∞

which is impossible. In the case where the sequence {lp } is unbounded, there
exists a subsequence lpk → ∞ as k → ∞. Then for each fixed m ≥ 1, there
exists an integer km such that m ≤ lpk , ∀k ≥ km , and hence,

f m (ypk ) ∈ X \ V, ∀k ≥ km .

Letting k → ∞, we obtain

f m (y) ∈ X \ V, ∀m ≥ 1,

which contradicts the definition of V .

The following result on the existence of strong global attractors is implied

by [142, Theorems 3.1 and 3.4]. Since the proof of this result was not provided
in [142], we include a simple proof of it.

Theorem 1.1.3. (Strong Global Attractors) Let f : X → X be a

continuous map. Assume that f is point dissipative on X, and one of the
following conditions holds:
8 1 Dissipative Dynamical Systems

(a) f n0 is compact for some integer n0 ≥ 1, or

(b) f is asymptotically smooth, and for each bounded set B ⊂ X, there exists
k = k(B) ≥ 0 such that γ + (f k (B)) is bounded.
Then there is a strong global attractor A for f .

Proof. The conclusion in case (b) is an immediate consequence of Theo-

 ninteger n0 ≥ 1, it
rem 1.1.2. In the case of (a), since f n0 is compact for some
suffices to show that for each compact subset C ⊂ X, f (C) is bounded.
n≥0
By applying
 Lemma 1.1.5 to f = f n0 , we deduce that for each compact subset
C ⊂ X, n
f (C) is bounded. So Theorem 1.1.2 implies that f has a global
n≥0

attractor A ⊂ X. We set B = f k A . By the continuity of f , it
0≤k≤n0 −1
then follows that B is compact and attracts every compact subset of X for f ,
and hence, the result follows from Theorem 1.1.2.

Remark 1.1.3. It is easy to see that a metric space (X, d) is complete if and
only if for any subset B of X, α(B) = 0 implies that B is compact. How-
ever, we can prove that Lemmas 1.1.3 and 1.1.4 also hold for non-complete
metric spaces by employing the equivalence between the compactness and the
sequential compactness for metric spaces. It then follows that Theorems 1.1.2
and 1.1.3 are still valid for any metric space. We refer to [64, 286] for the
existence of strong global attractors of continuous-time semiflows on a metric
space.

Clearly, if the global attractor is a singleton {e}, then e is a globally

attractive fixed point. Let A be the global attractor claimed in Theorem 1.1.2
with X being a Banach space and with “asymptotically smooth” replaced by
“α-condensing.” The following asymptotic fixed point theorem implies that
there is at least one fixed point in A. For a proof of it, we refer to [257, 143]
or [141, Section 2.6].
Theorem 1.1.4. (Asymptotic fixed point theorem) Suppose E is a Ba-
nach space. If f : E → E is α-condensing and compact dissipative, then f has
a fixed point.
Let Λ be a metric space. The family of continuous mappings fλ : X → X, λ ∈
Λ, is said to be collectively asymptotically smooth if for any nonempty closed
bounded set B ⊂ X for which fλ (B) ⊂ B, λ ∈ Λ, there is a compact set
Jλ = J(λ, B) ⊂ B such that Jλ attracts B under fλ and ∪λ∈Λ Jλ is precompact
in X. We then have the following result on the upper semicontinuity of global
attractors. For a proof, we refer to [141, Theorem 2.5.3].
Theorem 1.1.5. Let f : Λ×X → X be continuous, fλ =: f (λ, ·), and suppose
there is a bounded set B that attracts points of X under fλ for each λ ∈ Λ, and
for any bounded set U , the set V = ∪λ∈Λ ∪n≥0 fλn (U ) is bounded. If the family
 1.2 Chain Transitivity and Attractivity 9

{fλ : λ ∈ Λ} is collectively asymptotically smooth, then the global attractor Aλ

of fλ is upper semicontinuous in the sense that limλ→λ0 supx∈Aλ d(x, Aλ0 ) = 0
for each λ0 ∈ Λ.

1.2 Chain Transitivity and Attractivity

In this section we continue to assume that X is a metric space with metric d,
and that f : X → X is a continuous map.

1.2.1 Chain Transitive Sets

Definition 1.2.1. A point x ∈ X is said to be chain recurrent if for any

 > 0, there is a finite sequence of points x1 , . . . , xm in X (m > 1) with
x1 = x = xm such that d(f (xi ), xi+1 ) <  for all 1 ≤ i ≤ m − 1. The set of all
chain recurrent points for f : X → X is denoted by R(X, f ). Let A ⊂ X be a
nonempty invariant set. We call A internally chain recurrent if R(A, f ) = A,
and internally chain transitive if the following stronger condition holds: For
any a, b ∈ A and any  > 0, there is a finite sequence x1 , . . . , xm in A with
x1 = a, xm = b such that d(f (xi ), xi+1 ) < , 1 ≤ i ≤ m − 1. The sequence
{x1 , . . . , xm } is called an -chain in A connecting a and b.

Following LaSalle [212], we call a compact invariant set A invariantly con-

nected if it cannot be decomposed into two disjoint closed nonempty invariant
sets. An internally chain recurrent set need not have this property, e.g., a pair
of fixed points. However, it is easy to see that every internally chain transitive
set is invariantly connected.
We give some examples of internally chain transitive sets.
Lemma 1.2.1. Let f : X → X be a continuous map. Then the omega (al-
pha) limit set of any precompact positive (negative) orbit is internally chain
transitive.

Proof. Let x ∈ X and set xn = f n (x). Assume that x has a precompact

orbit γ = {xn }, and denote its omega limit set by ω. Then ω is nonempty,
compact, and invariant, and limn→∞ d(xn , ω) = 0. Let  > 0 be given. By
the continuity of f and compactness of ω, there exists δ ∈ (0, 3 ) with the
following property: If u, v are points in the open δ-neighborhood U of ω with
d(u, v) < δ, then d(f (u), f (v)) < 3 . Since xn approaches ω as n → ∞, there
exists N > 0 such that xn ∈ U for all n ≥ N .
Let a, b ∈ ω be arbitrary. There exist k > m ≥ N such that d(xm , f (a)) <

3 and d(xk , b) < 3 . The sequence

{y0 = a, y1 = xm , . . . , yk−m = xk−1 , yk−m+1 = b}

10 1 Dissipative Dynamical Systems

is an 3 -chain in X connecting a and b. Since for each yi ∈ U , 1 ≤ i ≤ k − m,

we can choose zi ∈ ω such that d(zi , yi ) < δ. Let z0 = a and zk−m+1 = b.
Then for i = 0, 1, . . . , k − m we have
d(f (zi ), zi+1 ) ≤ d(f (zi ), f (yi )) + d(f (yi ), yi+1 ) + d(yi+1 , zi+1 )
< /3 + /3 + /3.

Thus the sequence z0 , z1 , . . . , zk−m , zk−m+1 is an -chain in ω connecting a

and b. Therefore, ω is internally chain transitive. By a similar argument, we
can prove the internal chain transitivity of alpha limit sets of precompact
negative orbits.
Let {Sn : X → X}n≥0 be a sequence of continuous maps. The discrete
dynamical process (or process for short) generated by {Sn } is the sequence
{Tn : X → X}n≥0 defined by T0 = I = the identity map of X and

Tn = Sn−1 ◦ Sn−2 ◦ · · · ◦ S1 ◦ S0 , n ≥ 1.

The orbit of x ∈ X under this process is the set γ + (x) = {Tn (x) : n ≥ 0},
and its omega limit set is

ω(x) = y ∈ X : ∃nk → ∞ such that lim Tnk (x) = y .
k→∞

If there is a continuous map S on X such that Sn = S, ∀n ≥ 0, so that Tn

is the nth iterate S n , then {Tn } is a special kind of process called the discrete
semiflow generated by S. By an abuse of language we may refer to the map
S as a discrete semiflow.
Definition 1.2.2. The process {Tn : X → X} is asymptotically autonomous
if there exists a continuous map S : X → X such that

nj → ∞, xj → x ⇒ lim Snj (xj ) = S(x).

j→∞

We also say that {Tn } is asymptotic to S.

It is easy to see from the triangle inequality that if limn→∞ Sn = S uni-
formly on compact sets, then the process generated by {Sn } is asymptotic
to S.
Lemma 1.2.2. Let Tn : X → X, n ≥ 0, be an asymptotically autonomous
discrete process with limit S : X → X. Then the omega limit set of any
precompact orbit of {Tn } is internally chain transitive for S.
Proof. Let N+ = N+ ∪ {∞}. For any given strictly increasing continuous
function φ : [0, ∞) → [0, 1) with φ(0) = 0 and φ(∞) = 1 (e.g., φ(s) = 1+s
s
),
we can define a metric ρ on N+ as ρ(m1 , m2 ) = |φ(m1 ) − φ(m2 )|, for any
m1 , m2 ∈ N+ , and then N+ is compactified. Let X := N+ × X. Define a
mapping S : X → X by
 1.2 Chain Transitivity and Attractivity 11

S(m, x) = (1 + m, Sm (x)), S(∞, x) = (∞, S(x)), ∀m ∈ N+ , x ∈ X.

By Definition 1.2.2, S : X → X is continuous. Let γ + (x) be a precompact

orbit of Tn . Since

S n ((0, x)) = (n, Sn−1 ◦ Sn−2 ◦ · · · ◦ S1 ◦ S0 (x)) = (n, Tn (x)), ∀n ≥ 0,

and N+ is compact, it follows that the orbit γ + ((0, x)) of S n is precompact

and {∞} × ω(x) = ω(0, x), where ω(0, x) is the omega limit set of (0, x) for
S n . By Lemma 1.2.1, ω(0, x) is invariant and internally chain transitive for
S, which, together with the definition of S, implies that ω(x) is invariant and
internally chain transitive for S.

Definition 1.2.3. Let S : X → X be a continuous map. A sequence {xn } in

X is an asymptotic pseudo-orbit of S if

lim d(S(xn ), xn+1 ) = 0.

n→∞

The omega limit set of {xn } is the set of limits of subsequences.

Let {Tn } be a discrete process in X generated by a sequence of continuous

maps Sn that converges to a continuous map S : X → X uniformly on
compact subsets of X. It is easy to see that every precompact orbit of Tn :
X → X, n ≥ 0, is an asymptotic pseudo-orbit of S.
Example 1.2.1. Consider the nonautonomous difference equation

xn+1 = f (n, xn ), n ≥ 0,

on the metric space X. If we define Sn = f (n, ·) : X → X, n ≥ 0, and let

T0 = I, Tn = Sn−1 ◦ · · · ◦ S1 ◦ S0 : X → X, n ≥ 1,

then xn = Tn (x0 ), and {xn : n ≥ 0} is an orbit of the discrete process

Tn . If f (n, ·) → f¯ : X → X uniformly on compact subsets of X, then Tn
is asymptotically autonomous with limit f¯. Furthermore, in this case any
precompact orbit of the difference equation is an asymptotic pseudo-orbit of
f¯, since d(f¯(xn ), xn+1 ) = d(f¯(xn ), f (n, xn )) → 0.

Lemma 1.2.3. The omega limit set of any precompact asymptotic pseudo-
orbit of a continuous map S : X → X is nonempty, compact, invariant, and
internally chain transitive.

Proof. Let (N+ , ρ) be the compact metric space defined in the proof of
Lemma 1.2.2. Let {xn : n ≥ 0} be a precompact asymptotic pseudo-orbit
of S : X → X, and denote its compact omega limit set by ω. Define a metric
space
Y = ({∞} × X) ∪ {(n, xn ) : n ≥ 0}
12 1 Dissipative Dynamical Systems

and

g : Y → Y, g(n, xn ) = (n + 1, xn+1 ), g(∞, x) = (∞, S(x)).

By Definition 1.2.3 and the fact that d(xn+1 , S(x)) ≤ d(xn+1 , S(xn )) +
d(S(xn ), S(x)) for x ∈ X, n ≥ 0, it easily follows that g : Y → Y is con-
tinuous. Let γ + (0, x0 ) = {(n, xn ) : n ≥ 0} be the positive orbit of (0, x0 ) for
the discrete semiflow g n : Y → Y, n ≥ 0. Then γ + (0, x0 ) is precompact in Y ,
and its omega limit ω(0, x0 ) is {∞} × ω, which by Lemma 1.2.1 is invariant
and internally chain transitive for g. Applying the definition of g, we see that
ω is invariant and internally chain transitive for S.
Let A and B be two nonempty compact subsets of X. Recall that the
Hausdorff distance between A and B is defined by

dH (A, B) := max (sup{d(x, B) : x ∈ A}, sup{d(x, A) : x ∈ B}) .

We then have the following result.

Lemma 1.2.4. Let S, Sn : X → X, ∀n ≥ 1, be continuous. Let {Dn } be a
sequence of nonempty compact subsets of X with limn→∞ dH (Dn , D) = 0 for
some compact subset D of X. Assume that for each n ≥ 1, Dn is invariant
and internally chain transitive for Sn . If Sn → S uniformly on D ∪(∪n≥1 Dn ),
then D is invariant and internally chain transitive for S.

Proof. Observe that the set K = D (∪n≥1 Dn ) is compact. Indeed, since
an open cover of K also covers D, a finite subcover provides a neighborhood
of D that must also contain Dn for all large n. If x ∈ D, then there exist
xn ∈ Dn such that xn → x. Since Sn (xn ) ∈ Dn and Sn (xn ) → S(x), we see
that S(x) ∈ D. Thus S(D) ⊂ D. On the other hand, there exist yn ∈ Dn such
that Sn (yn ) = xn . Since dH (Dni , D) → 0, we can assume that yni → y ∈ D
for some subsequence yni . Then xni = Sni (yni ) → S(y) = x, showing that
S(D) = D.
By uniform continuity and uniform convergence, for any  > 0 there exist
δ ∈ (0, /3) and a natural number N such that for n ≥ N and u, v ∈ K with
d(u, v) < δ, we have

d(Sn (u), S(v)) ≤ d(Sn (u), S(u)) + d(S(u), S(v)) < /3.

Fix n > N such that dH (Dn , D) < δ. For any a, b ∈ D, there are points
x, y ∈ Dn such that d(x, a) < δ and d(y, b) < δ. Since Dn is internally chain
transitive for Sn , there is a δ-chain {z1 = x, z2 , . . . , zm+1 = y} in Dn for Sn
connecting x to y. For each i = 2, . . . , m we can find wi ∈ D with d(wi , zi ) < δ,
since Dn is contained in the δ-neighborhood of D. Let w1 = a, wm+1 = b. We
then have

d(S(wi ), wi+1 ) ≤ d(S(wi ), Sn (zi )) + d(Sn (zi ), zi+1 ) + d(zi+1 , wi+1 )

< /3 + δ + δ < 
 1.2 Chain Transitivity and Attractivity 13

for i = 1, . . . , m. Thus {w1 = a, w2 , . . . , wm+1 = b} is an -chain for S in D

connecting a to b.

Let Φ(t) : X → X, t ≥ 0, be a continuous-time semiflow. That is, (x, t) →

Φ(t)x is continuous, Φ(0) = I and Φ(t) ◦ Φ(s) = Φ(t + s) for t, s ≥ 0. A
nonempty invariant set A ⊂ X for Φ(t) (i.e., Φ(t)A = A, ∀t ≥ 0) is said to be
internally chain transitive if for any a, b ∈ A and any  > 0, t0 > 0, there is a
finite sequence {x1 = a, x2 , . . . , xm−1 , xm = b; t1 , . . . , tm−1 } with xi ∈ A and
ti ≥ t0 , 1 ≤ i ≤ m−1, such that d(Φ(ti , xi ), xi+1 ) <  for all 1 ≤ i ≤ m−1. The
sequence {x1 , . . . , xm ; t1 , . . . , tm−1 } is called an (, t0 )-chain in A connecting
a and b. We then have the following result.

Lemma 1.2.1 Let Φ(t) : X → X, t ≥ 0, be a continuous-time semiflow.

Then the omega (alpha) limit set of any precompact positive (negative) orbit
is internally chain transitive.

Proof. Let ω = ω(x) be the omega limit set of a precompact orbit γ(x) =
{Φ(t)x : t ≥ 0} in X. Then ω is nonempty, compact, invariant and
limt→∞ d(Φ(t)x, ω) = 0. Let  > 0 and t0 > 0 be given. By the uniform
continuity of Φ(t)x for (t, x) in the compact set [t0 , 2t0 ] × ω, there is a
δ = δ(, t0 ) ∈ (0, 3 ) such that for any t ∈ [t0 , 2t0 ] and u and v in the open
δ-neighborhood U of ω with d(u, v) < δ, we have d(Φ(t)u, Φ(t)v) < 3 . It
then follows that there exists a sufficiently large T0 = T0 (δ) > 0 such that
Φ(t)x ∈ U , for all t ≥ T0 . For any a, b ∈ ω, there exist T1 > T0 and T2 > T0
with T2 > T1 + t0 such that d(Φ(T1 )x, Φ(t0 )a) < 3 and d(Φ(T2 )x, b) < 3 . Let
m be the greatest integer that is not greater than T2t−T 0
1
. Then m ≥ 1. Set

y1 = a, yi = Φ(T1 + (i − 2)t0 )x, i = 2, . . . , m + 1, ym+2 = b,

and
ti = t0 for i = 1, . . . , m; tm+1 = T2 − T1 − (m − 1)t0 .
Then tm+1 ∈ [t0 , 2t0 ). It follows that d(Φ(ti )yi , yi+1 ) < 
3 for all i = 1, . . . , m+
1. Thus the sequence

{y1 = a, y2 , . . . , ym+1 , ym+2 = b; t1 , t2 , . . . , tm+1 }

is an ( 3 , t0 )-chain in X connecting a and b. Since yi ∈ U for i = 2, . . . , m + 1,

we can choose zi ∈ ω such that d(zi , yi ) < δ. Let z1 = a and zm+2 = b. It
then follows that
d(Φ(ti )zi , zi+1 ) ≤ d(Φ(ti )zi , Φ(ti )yi ) + d(Φ(ti )yi , yi+1 ) + d(yi+1 , zi+1 )
< /3 + /3 + /3, i = 1, . . . , m + 1.

This proves that the sequence {z1 = a, z2 , . . . , zm+1 , zm+2 = b; t1 , t2 , . . . , tm+1 }

is an (, t0 )-chain in ω connecting a and b. Therefore, ω is internally chain tran-
sitive. By a similar argument we can prove the internal chain transitivity of
alpha limit sets of precompact negative orbits.
14 1 Dissipative Dynamical Systems

With Lemma 1.2.1 it is easy to see that there are analogues of Lem-
mas 1.2.2 and 1.2.3 for continuous-time semiflows. The following result is an
analogue of Lemma 1.2.4 for continuous-time semiflows.

Lemma 1.2.4 Let Φ and Φn be continuous-time semiflows on X for n ≥ 1.

Let {Dn } be a sequence of nonempty compact subsets of X with
limn→∞ dH (Dn , D) = 0 for some compact subset D of X. Assume that for
each n ≥ 1, Dn is invariant and internally chain transitive for Φn . If for each
T > 0, Φn → Φ uniformly for (x, t) ∈ (D ∪ (∪n≥1 Dn )) × [0, T ], then D is
invariant and internally chain transitive for Φ.

Proof. It is easy to see that K = D (∪n≥1 Dn ) is compact and D is invariant
for Φ. By uniform continuity and uniform convergence, for any  > 0 and
t0 > 0 there exists δ ∈ (0, /3) and a natural number N such that for n ≥
N , t ∈ [0, 2t0 ], and u, v ∈ K with d(u, v) < δ, we have d(Φnt (u), Φt (v)) ≤
d(Φnt (u), Φt (u)) + d(Φt (u), Φt (v)) < /3. Fix n > N such that dH (Dn , D) < δ.
For any a, b ∈ D, there are points x, y ∈ Dn such that d(x, a) < δ and
d(y, b) < δ. Since Dn is chain transitive for Φn , there is a (δ, t0 )-chain {z1 =
x, z2 , . . . , zm+1 = y; t1 , . . . , tm } in Dn for Φn , with t0 ≤ ti < 2t0 connecting x
to y. For each i = 2, . . . , m we can find wi ∈ D with d(wi , zi ) < δ, since Dn is
contained in the δ-neighborhood of D. Let w1 = a, wm+1 = b. We then have

d(Φti (wi ), wi+1 ) ≤ d(Φti (wi ), Φnti (zi )) + d(Φnti (zi ), zi+1 ) + d(zi+1 , wi+1 )
< /3 + δ + δ < 

for i = 1, . . . , m. Thus {w1 = a, w2 , . . . , wm+1 = b; t1 , . . . , tm } is an (, t0 )-

chain for Φ in D connecting a to b.

Example 1.2.2. Note that if in Lemma 1.2.4 Dn is an omega limit set for
Φn (and therefore internally chain transitive by Lemma 1.2.1 ), the set D
need not be an omega limit set for the limit semiflow Φ, although it must be
chain transitive. Easy examples are constructed with Φn = Φ, ∀n ≥ 1. For
example, consider the flow generated by the planar vector field given in polar
coordinates by
r = 0, θ = 1 − r.
The unit circle D = {r = 1}, consisting of equilibria, is chain transitive but is
not an omega limit set of any point, yet D is the Hausdorff limit of the omega
limit sets Dn = {r = 1 + n1 }.

Lemma 1.2.5. A nonempty compact invariant set M is internally chain tran-

sitive if and only if M is the omega limit set of some asymptotic pseudo-orbit
of f in M .

Proof. The sufficiency follows from Lemma 1.2.3. To prove the necessity,
we can choose a point x ∈ M since M is nonempty. For any  > 0,
the compactness of M implies that there is a finite sequence of points
 1.2 Chain Transitivity and Attractivity 15

{x1 = x, x2 , . . . , xm , xm+1 = x} in M such that its -net in X covers M ;

i.e., M ⊂ ∪m i=1 B(xi , ), where B(xi , ) := {y ∈ X : d(y, xi ) < }. For each
1 ≤ i ≤ m, since M is internally chain transitive, there is a finite -chain
{y1i = xi , y2i , . . . , yni i , yni i +1 = xi+1 } in M connecting xi and xi+1 . Then the
sequence {y11 , . . . , yn1 1 , y12 , . . . , yn2 2 , . . . , y1m , . . . , ynmm , ynmm +1 } is a finite -chain
in M connecting x and x, and its -net in X covers M .
For each integer k, letting  = k1 in the above claim, we have a finite k1 -
chain {z1k = x, z2k , . . . , zlkk , zlkk +1 = x} in M whose k1 -net in X covers M . It
then easily follows that the infinite sequence of points

{z11 , . . . , zl11 , z12 , . . . , zl22 , . . . , z1k , . . . , zlkk , . . . }

is an asymptotic pseudo-orbit of f in M and its omega limit set is M .

Block–Franke Lemma ([33], Theorem A) Let K be a compact metric

space and f : K → K a continuous map. Then x ∈ R(K, f ) if and only
if there exists an attractor A ⊂ K such that x ∈ W s (A) \ A.

Lemma 1.2.6. A nonempty compact invariant set M is internally chain tran-

sitive if and only if f |M : M → M has no proper attractor.

Proof. Necessity. Assume that there is a proper attractor A for f |M : M →

M . Then A = ∅ and M \A = ∅. Since A is an attractor, there is an 0 > 0 such
that A attracts the open 0 -neighborhood U of A in M . Choose a ∈ M \ A
and b ∈ A and let {x1 = a, x2 , . . . , xm = b} be an 0 -chain in M connecting
a and b. Let k = min{i : 1 ≤ i ≤ m, xi ∈ A}. Since b ∈ A and a ∈ A, we
have 2 ≤ k ≤ m. Since d(f (xk−1 ), xk ) < 0 , we have f (xk−1 ) ∈ U and hence
xk−1 ∈ W s (A) \ A. By the Block–Franke lemma, xk−1 ∈ R(M, f ), which
proves that M is not internally chain recurrent, and a fortiori not internally
chain transitive.
Sufficiency. For any subset B ⊂ X we define ω(B) to be the set of limits
of sequences of the form {f nk (xk )}, where nk → ∞ and xk ∈ B. Since f |M :
M → M has no proper attractor, the Block–Franke lemma implies that M is
internally chain recurrent. Given a, b ∈ M and  > 0, let V be the set of all
points x in M for which there is an -chain in M connecting a to x; this set
contains a. For any z ∈ V , let {z1 = a, z2 , . . . , zm−1 , zm = z} be an -chain in
M connecting a to z. Since

lim d(f (zm−1 ), x) = d(f (zm−1 ), z) < ,

x→z

there is an open neighborhood U of z in M such that for any x ∈ U ,

d(f (zm−1 ), x) < . Then {z1 = a, z2 , . . . , zm−1 , x} is an -chain in M con-
necting a and x, and hence U ⊂ V . Thus V is an open set in M . We
further claim that f (V ) ⊂ V . Indeed, for any z ∈ V , by the continu-
ity of f at z, we can choose y ∈ V such that d(f (y), f (z)) < . Let
{y1 = a, y2 , . . . , ym−1 , ym = y} be an -chain in M connecting a and y. It
16 1 Dissipative Dynamical Systems

then follows that {y1 = a, y2 , . . . , ym−1 , ym = y, ym+1 = f (z)} is an -chain

in M connecting a and f (z), and hence f (z) ∈ V . By the compactness of M
and [141, Lemma 2.1.2] applied to f : M → M , it then follows that ω(V )
is nonempty, compact, invariant, and ω(V ) attracts V . Since f (V ) ⊂ V , we
have ω(V ) ⊂ V and hence ω(V ) = f (ω(V )) ⊂ V . Then ω(V ) is an attractor
in M . Now the nonexistence of a proper attractor for f : M → M implies
that ω(V ) = M and hence V = M . Clearly, b ∈ M = V , and hence, by the
definition of V , there is an -chain in M connecting a and b. Therefore, M is
internally chain transitive.

1.2.2 Attractivity and Morse Decompositions

Recall that a nonempty invariant subset M of X is said to be isolated for

f : X → X if it is the maximal invariant set in some neighborhood of itself.
Lemma 1.2.7. (Butler–McGehee-type lemma) Let M be an isolated in-
variant set and L a compact internally chain transitive set for f : X → X.
Assume that L ∩ M = ∅ and L ⊂ M . Then
(a) there exists u ∈ L \ M such that ω(u) ⊂ M ;
(b) there exist w ∈ L \ M and a negative orbit γ − (w) ⊂ L such that its α-limit
set satisfies α(w) ⊂ M .

Proof. Since M is an isolated invariant set, there exists an  > 0 such that
M is the maximal invariant set in the closed -neighborhood of M . By the
assumption, we can choose a ∈ L ∩ M and b ∈ L with d(b, M ) > . For
any integer k ≥ 1, by the internal chain transitivity of L, there exists a
1 1
k -chain {y1 = a, . . . , ylk +1 = b} in L connecting a and b, and a k -chain
k k

{z1 = b, . . . , zmk +1 = a} in L connecting b and a. Define a sequence of points

k k

by

{xn : n ≥ 0} := {y11 , . . . , yl11 , z11 , . . . , zm

1
1
, . . . , y1k , . . . , ylkk , z1k , . . . , zm
k
k
, . . . }.
k
Then for any k > 0 and for all n ≥ N (k) := j=1 (lj + mj ), we have
1
d(f (xn ), xn+1 ) < k+1 , and hence limn→∞ d(f (xn ), xn+1 ) = 0. Thus {xn }n≥0
⊂ L is a precompact asymptotic pseudo-orbit of f : X → X. Then there are
two subsequences xmj and xrj such that xmj = a and xrj = b for all j ≥ 1.
Note that d(xsj +1 , f (x)) ≤ d(xsj +1 , f (xsj )) + d(f (xsj ), f (x)). By induction, it
then follows that for any convergent subsequence xsj → x ∈ X, j → ∞,
we have limj→∞ xsj +n = f n (x) for any integer n ≥ 0. We can further
choose two sequences lj and nj with lj < mj < nj and limj→∞ lj = ∞
such that d(xlj , M ) > , d(xnj , M ) > , and d(xk , M ) ≤  for any integer
k ∈ (lj , nj ), j ≥ 1. Since {xn : n ≥ 0} is a subset of the compact set L, we can
assume that upon taking a convergent subsequence, xlj → u ∈ L as j → ∞.
Clearly, d(u, M ) ≥  and hence u ∈ L \ M . Since u ∈ L and L is a compact in-
variant set, we have ω(u) ⊂ L. We further claim that ω(u) ⊂ M , which proves
Exploring the Variety of Random
Documents with Different Content
 Pig

KANGAROO
This animal has its home in Australia, where the birds are songless
and the trees give no shade.
It has a very powerful tail which serves with the two long hind legs
as its support and for making enormous bounds. The fore legs,
much like arms, are used with surprising dexterity by this strange
animal. The nose, throat and breast are very light, the rest of the
body is reddish brown.

Kangaroo

CAMEL
It is with significance that this beast is called the ship of the desert.
Like a ship crossing a wide ocean of water, the camel travels across
a great expanse of sand carrying heavy loads of freight.
It has a wonderful endurance and can go for a week without
drinking, subsisting on the coarse grass of that waste region, and
the water stored up in its humps. Its dreamy eyes, sullen nature,
angular figure and neutral grey-brown color—all seem to resemble
the rocky desert itself.
 Camel

GIRAFFE
This creature (Fig. 9) is, unquestionably, the tallest of all the beasts
of the forests of Africa. It does not stoop to obtain its living from the
ground, but browses on the tops of trees.
It is also called the camelopard, suggesting a resemblance in shape
to the camel and in color to the leopard.
It is cream-colored with a shower of dark-brown spots on its back
and sides.

Fig. 9.
 Giraffe

SWAN ROCKER
All the parts may be made from 1/4" stock. The two sides are made
like the pattern. The seat is 2-1/4" wide, the back 2-1/2". Both are
3" long, and serve to unite the two sides.
When the parts are finished and ready for assembling, mark on the
sides the exact location of seat and back; also the position of each
nail on all parts. Hold the two sides together and make small holes
thru them where the nails are to be driven. Nail, in turn, the seat
and back to the first side, and then to the second side.
Color white and decorate appropriately.
 Swan Rocker

BALANCING PEACOCK
Saw out the shape of the fowl as outlined in the drawing.
There is good chance for display of fine color in dark-green and blue.
The breast should be colored brown.
Make a small stick about 2" long for a perch. Drive a small nail into
each end of the stick, and tie the ends of a fine cord to each nail.
Press a brad thru the middle of the stick up into the foot of the
peacock. It may then be hung up by the long loop of the string and
swung freely.
 Peacock

TOY DOG
This dog may be made of 3/16" or 1/4" stock. When put together,
the two pieces for the body are separated by the head, tail and two
circular pieces. The legs are fastened to the outside of the body by
two long nails that reach thru the five thicknesses. The ends of the
nails are bent back.
Smooth the parts and color white, with large black spots on head,
back, tail and legs.
 Toy Dog

TEDDY BEAR
This animal has two pieces for the body. The head and two circular
pieces hold these two pieces apart. The head and legs move on tight
joints so that the bear will stay in the desired position.
 Teddy Bear

PARROT
This may be made of 1/4" stock. The base has two holes bored thru
it for the uprights with fit tight into the holes (Fig. 10). The perch is
2" long and has a brad in each end to swing on. These brads fit
loose thru the uprights near the top ends.
The parrot is sawed out, and a 3/4" brad driven up thru the foot
which also fits into the hole in the middle of the perch.
The parrot may be colored white with black trimmings, yellow beak
and eyes, red crest, tail and foot. The wings are green.
The stand should have a green base, red uprights, and yellow perch.
The bird should balance well and swing freely.
 Fig. 10.

Parrot

DOLL'S HOBBY HORSE

The two body pieces of the horse with rocker are sawed from 1/4"
wood. The seat and back are made alike, as are also the shelf and
foot-rest. The ends of these four pieces should be at right angles to
the sides, and the edges should be slightly rounded.
Mark carefully on both side pieces where the cross-pieces are to be
fastened. Hold the two together and make fine holes thru the two
thicknesses where the nails are to go thru. By driving the brads thru
these holes, fasten all the cross pieces to one side, and watch that
the brads go straight. Then fasten the other side in a similar way.
Color the horses white, rockers red, and seat blue.

Doll's Hobby Horse

STERN WHEELER
This boat is sawed out according to the drawing, and notches cut on
the arms at the stern as a place for the rubber band which serves
both as axle and motive power. The paddles are sawed out to fit
together to form a stern wheel with four paddles.
The elastic is made in a double loop of four thicknesses, one of
which is placed in each of the four angles of the paddle. By twisting
the elastic band, power is stored up sufficient for the boat to attain a
fair rate of speed. A spool serves as smoke-stack.
Paint the boat white and smoke-stack red.

THE WEATHER-COCK
This fowl may be made of 3/8" stock; a shingle will do. Let the grain
run vertically. The perch may be made of similar stock, triple
thickness, with the middle piece short to allow room for the foot of
the rooster. The pieces are securely fastened together with nails,
and a hole bored in the other end of the perch for a metal pivot, on
which it should swing to tell the way the wind blows. Give it two
coats of paint, using brilliant colors.
 Stern Wheeler
Weather Cock

ARM-CHAIR AND ROCKING-CHAIR

These two doll's chairs are similar in construction. Make them from
1/4" wood. Saw out two sides, a back and a seat for each chair.
Have edges finished accurately. The seats are alike for the two
chairs. The backs, also, are alike, except that the one for the rocker
is 1/4" longer than the other. Assemble as shown in dotted lines, and
fasten with 1" brads.
The chairs may be colored white or mahogany. With due care and
skill, this furniture may be made very attractive and valuable.

Rocking Chair
Arm Chair

DOLL'S CRADLE
Every little girl loves to own a cradle for her doll. Here is one that is
strong and pretty, and can easily be made from thin wood about
1/4" thick. Saw out the two ends after tracing the outline from the
full-sized drawing. The two sides and bottom have their dimensions
given. When properly shaped and smooth, the bottom is fastened
with 1" brads between the ends. One-inch brads are driven thru the
sides into the edges of the ends. Be careful when driving the brads
that the sides do not split and that the brads go straight.
Paint the cradle white on the outside and violet or pink inside.
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

textbookfull.com