Professional Documents
Culture Documents
Spatial Dynamics and Pattern Formation in Biological Populations 1St Edition Ranjit Kumar Upadhyay Online Ebook Texxtbook Full Chapter PDF
Spatial Dynamics and Pattern Formation in Biological Populations 1St Edition Ranjit Kumar Upadhyay Online Ebook Texxtbook Full Chapter PDF
https://ebookmeta.com/product/kinematics-and-dynamics-of-
galactic-stellar-populations-1st-edition-rafael-cubarsi/
https://ebookmeta.com/product/computational-intelligence-in-
pattern-recognition-asit-kumar-das/
https://ebookmeta.com/product/preference-based-spatial-co-
location-pattern-mining-big-data-management-lizhen-wang/
https://ebookmeta.com/product/research-methodology-a-step-by-
step-guide-for-beginners-5th-edition-ranjit-kumar/
Nanosensors: Physical, Chemical, and Biological (Series
in Sensors) 2nd Edition Vinod Kumar Khanna
https://ebookmeta.com/product/nanosensors-physical-chemical-and-
biological-series-in-sensors-2nd-edition-vinod-kumar-khanna/
https://ebookmeta.com/product/biological-and-bio-inspired-fluid-
dynamics-theory-and-application-david-e-rival/
https://ebookmeta.com/product/spatial-analysis-using-big-data-
methods-and-urban-applications-spatial-econometrics-and-spatial-
statistics-1st-edition-yoshiki-yamagata/
https://ebookmeta.com/product/transitions-in-post-soviet-eurasia-
identity-polity-and-strategic-choices-1st-edition-archana-
upadhyay-ed/
https://ebookmeta.com/product/spatial-minds-conceptual-
correlations-of-spatial-prepositions-in-hungarian-croatian-and-
english-1st-edition-irena-zovko-dinkovic-linda-gros/
Spatial Dynamics and Pattern
Formation in Biological
Populations
Spatial Dynamics and Pattern
Formation in Biological
Populations
Te right of Ranjit Kumar Upadhyay and Satteluri R. K. Iyengar to be identifed as authors of this work has been
asserted by him in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988.
Reasonable eforts have been made to publish reliable data and information, but the author and publisher cannot
assume responsibility for the validity of all materials or the consequences of their use. Te authors and publishers
have attempted to trace the copyright holders of all material reproduced in this publication and apologize to
copyright holders if permission to publish in this form has not been obtained. If any copyright material has not been
acknowledged please write and let us know so we may rectify in any future reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or
utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including
photocopying, microflming, and recording, or in any information storage or retrieval system, without written
permission from the publishers.
For permission to photocopy or use material electronically from this work, access www.copyright.com or contact the
Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. For works that are
not available on CCC please contact mpkbookspermissions@tandf.co.uk
Trademark notice: Product or corporate names may be trademarks or registered trademarks and are used only for
identifcation and explanation without intent to infringe.
Typeset in Palatino
by codeMantra
Contents
Foreword .........................................................................................................................................ix
Preface..............................................................................................................................................xi
About Authors ............................................................................................................................. xiii
2 Reaction–Diffusion Modeling........................................................................................... 41
2.1 Introduction ................................................................................................................. 41
2.2 Reaction–Diffusion Equations ..................................................................................43
2.2.1 Derivation of Reaction-Diffusion Equation ...............................................43
2.3 Hyperbolic Reaction–Diffusion Equations..............................................................44
2.4 Single-Species Reaction–Diffusion Models............................................................. 46
2.4.1 Model 1: Linear Model of Kierstead and Slobodkin................................. 47
2.4.1.1 KISS Model in Two Dimensions .................................................. 49
2.4.2 Model 2: Nonlinear Fisher Equation........................................................... 50
2.4.2.1 Spatial Steady-State Solution ........................................................ 51
2.4.2.2 Some Analytical Solutions ............................................................ 53
2.4.3 Model 3: Nagumo Equation ......................................................................... 55
2.4.3.1 Numerical Solutions ...................................................................... 58
2.5 Two-Species Reaction–Diffusion Models ................................................................ 60
2.5.1 Turing Instabilities of Two-Species Reaction–Diffusion Systems .......... 60
2.5.1.1 Predator–Prey Reaction–Diffusion Systems...............................64
v
vi Contents
ix
Preface
During the past few years, the world has experienced the emergence of major devastating
epidemic outbreaks and the spread of viruses like Ebola and Zika. As this manuscript
was getting ready to be sent to the publishers, Coronavirus has emerged and has been
threatening the whole world. Understanding the modeling of virus dynamics of infectious
diseases holds the key for designing control strategies from the public health perspec-
tive. Therefore, we felt that collecting all the available literature at one place may be use-
ful for students and research workers in these areas. Emphasis is made on mathematical
modeling using reaction–diffusion systems in biological populations with applications to
ecology, epidemiology, and neural systems. The temporal and spatial dynamics can be
essentially different, and while its results may be misleading in some situations, temporal
analysis provides an important direction for understanding the spatial/spatiotemporal
dynamics. The analysis of patterns enables us to study the dynamics of macroscopic and
microscopic behavior of underlying systems. The travelling wave-type patterns can be
especially observed in dispersive systems.
Chapter 1 introduces the basic concepts of spatial dynamics and spatial/diffusive pro-
cesses of biological populations, Hopf bifurcation analysis, multiple-scale analysis, and
neural systems that play a key role in understanding the functions of the brain and its
dynamics. In Chapter 2, reaction–diffusion modeling is presented to describe the diffu-
sive dispersal of the population, developmental processes, etc. Three single-species mod-
els, a two-species model, six models in applications in biochemistry, and three models
of multispecies reaction–diffusion are analyzed. In Chapter 3, transmission dynamics
of infectious diseases are studied. Five different types of models are analyzed: (i) sus-
ceptible–infected (SI), (infuenza epidemic); (ii) susceptible–infected–susceptible (SIS);
(iii) susceptible–infected–removed (SIR); (iv) susceptible–infected–removed–susceptible
(SIRS); and (v) susceptible–exposed–infected–recovered (SEIR). Chapter 4 deals with fve
models of epidemic spread and outbreak dynamics of the Ebola virus. Chapter 5 deals
with seven models of transmission and outbreak dynamics of the Zika virus. The frst
four models discuss the temporal dynamics and the remaining three models discuss the
temporal and spatial dynamics. In Chapter 6, four biophysical neuron models are exam-
ined for the evolution of the functional mechanism of the brain. Using conductance-based
mathematical models, patterns of spiking activity and qualitative behavior of temporal
activity such as periodic fring, bursting, chattering, mixed-mode oscillations, and chaotic
fring are studied.
Most of the models discussed in the book are solved using the software MATLAB or
MATHEMATICA.
We are extremely grateful to Prof. Dr. Philip K. Maini, FRS, FMedSci, FNA, Wolfson
Centre for Mathematical Biology, Mathematical Institute, Oxford University, Oxford, for
writing the foreword for the book.
We thank all the Professors who reviewed the book and provided constructive sugges-
tions, which gave us the proper direction and impetus for writing the book.
We express our gratitude to all the following authors and their co-authors for giving
approval and providing an opportunity to include their works in the book:
Dr. Abid, W.; Dr. Agusto, F. B.; Dr. Al-Darabsah, I.; Dr. Allen, L. J. S.; Dr. Ambrosio, B.;
Dr. Aziz Alaoui, Dr. Bonyah, E.; Dr. Brauer, F.; Dr. Cai, Y.; Dr. Camacho, A.; Dr. Capone, F.;
xi
xii Preface
Dr. Carrero, G.; Dr. Charles, W. M.; Dr. Chinviriyasit, S.; Dr. Cui, R.; Dr. Deng, K.;
Dr. Djiomba Njankou; Dr. Do, T. S.; Dr. Dubey, B.; Dr. Fitzgibbon, W.; Dr. Funk, S.; Dr. Han,
W.; Dr. Harko, T.; Dr. Hille, B.; Dr. Imran, M.; Dr. Izhikevich, E. M.; Dr. Jin, Z.; Dr. Kao,
Y.; Dr. Kim, K. Ik.; Dr. King, J. R.; Dr. Kucharski, A. J.; Dr. Kumari, N.; Dr. Legrand, J.;
Dr. Li, T.; Dr. Li, X.; Dr. Liu, P.; Dr. Liu, Q. X.; Dr. Ma, M.; Dr. Madzvamuse, A.; Dr. Mao,
X.; Dr. Mariana Ruiz Villarreal.; Dr. Mark Kot.; Dr. Mazin, W.; Dr. Mondal, A.; Dr. Morris,
C.; Dr. Pankaj Seth.; Dr. Parshad, R.D.; Dr. Purves, D.; Dr. Röst, G.; Dr. Roy, P.; Dr. Ruan,
S.; Dr. Samsuzzoha, M.; Dr. Schwartz, Ira. B.; Dr. Schwiening, C. J.; Dr. Shen, J.; Dr. Shi,
J.; Dr. Sun, G. Q.; Dr. Tang, B.; Dr. Tang, Q. L.; Dr. Tsaneva-Atanasova, K.; Dr. Wang, W.;
Dr. Wang, X.; Dr. Wu, C.; Dr. Yang, F.; Dr. Yang, J.; Dr. Yi, F.; Dr. Zhang, J.; Dr. Zheng, Q.;
Dr. Zhou, J.
Our grateful thanks to the following publishers who have given copyright permis-
sion to reproduce fgures from their reputed International journals and books: American
Institute of Physics, American Physical Society, Cambridge University Press, Elsevier,
Hindawi Publishing Corporation, IOP Publishing, John Wiley and Sons, MIT Press,
Oxford Publishing House, Society for Industrial and Applied Mathematics, Springer
Nature, Springer Open Journal, the Physiological Society, and World Scientifc Publishing
Company.
We profusely thank Taylor and Francis for accepting to publish the book and bringing
it in a nice form.
We shall be extremely happy to receive suggestions and comments to improve the
quality of the book.
xiii
1
Introduction to Diffusive Processes
1.1 Introduction
Diffusive/spatial processes like evolution of new species, dynamics of invading species,
maintenance of biodiversity, and movements of animals and plants play central roles in
ecology. A key factor in how ecological communities are shaped is the spatial components
of ecological interactions. Modern tools give us information about rates of dispersal among
subpopulations. The study of the data of ecological species reveals that species spread as
traveling waves and also provide information about the interplay between ecological and
evolutionary dynamics in the spread of the population.
The spatial dynamics of pathogens and immune system cells within individual organ-
isms is important for understanding infectious disease outbreaks [89]. A common feature
of earlier ecological population models is that the interactions were based on the mass action
law, an approach that has its conceptual foundation in modeling chemical reactions [79].
When the reactants are well mixed and have to collide in order to react, the mass action
law states that “the collision rate (hence the reaction rate) is proportional to the product
of the concentrations of the reacting molecules”. The law assumes that the population of
interacting species is large enough to guarantee conditions of well mixing. Ecological situ-
ations in which well-mixed condition does not hold can lead to incorrect predictions, and
a spatial model with local interactions is more appropriate in such cases. The following are
two examples of spatial movements: Andrewartha and Birch [2] observed that some insect
populations become frequently extinct but persisted globally due to recolonization from
local populations. The second example is Huffaker’s [43] laboratory experiment with two
mites, one of which feeds on oranges (Eotetranychus sexmaculatus) and the other feeds on
the predatory mite (Typhlodromus occidentalis) that attacks E. sexmaculatus. Huffaker set up
an array of oranges and rubber balls with different spatial complexity that controlled dis-
persal. He demonstrated that a complex spatially heterogeneous array promoted species
coexistence and later confrmed that spatial subdivision is important for the persistence
of populations.
In modeling the population dynamics, space can be included either implicitly or explic-
itly. The Levins model [55] describes the dynamics of a population in a spatially subdivided
habitat. Population may go to extinction in patches and may subsequently be recolonized
from other occupied patches. Space is implicit in the sense that recolonization is equally
likely from all occupied patches regardless of their locations. The Father of the studies of
ecological diffusion, John Gorden Skellam [96], employed models that include space explic-
itly to describe the invasion of species. These models are similar to Fisher’s model [51]
for the spread of novel allele. Fisher and the famous troika of researchers – Kolmogorov,
Petrovskii, and Piscunov – studied the diffusion models in population dynamics as early
1
2 Spatial Dynamics and Pattern Formation in Biological Populations
as 1936. These models included space, but they did not allow for spatial correlations, since
local populations are effectively infnite [79]. A key aspect of the spatial processes is the
degree to which the subunits of a system or network are connected to one another, and
this degree governs the population persistence, patterns of biodiversity, and ecosystem
function [7]. Complex ecosystems exhibit patterns that are bound to each other and are
observed over different spatial and time scales [36]. Ecosystem theory is full of macroeco-
logical patterns for which causal relationships are still debated [34,68]. Species–area and
diversity–stability relationships of ecological patterns are some of the important areas of
study.
As time progresses, the gradient will grow increasingly shallow until the concentrations
are equalized. It is possible that nonreversible processes such as diffusion can be gen-
erated by microscopic deterministic chaos. Diffusion increases entropy, decreases Gibbs
free energy, and therefore is thermodynamically favorable. Diffusion operates under the
second law of thermodynamics. The second law of thermodynamics gives Gibbs equation
which determines the direction of an irreversible process, relating entropy to a change
in internal energy, volume, and partial masses. Diffusion is important in systems experi-
encing an applied force. In a conducting material, the net motion of electrons in an elec-
trical feld quickly reaches a terminal velocity (resulting in a steady current described
by Ohm’s law) because of the thermal (diffusive) motions of atoms. Einstein’s relation,
D = kT × mobility (discovered by Nernst 1884), relates the diffusion coeffcient to the mobil-
ity of particles [22]. Pais in his book titled, Subtle is the Lord: The Science and the Life of Albert
Einstein [82], gave an interesting review of Einstein’s work on diffusion. Einstein employed
a random walk model for his analysis. Einstein’s work on diffusion can be applied to dairy
technology (colloidal properties of micelle suspension in milk) and construction industry
(link with granular matter) [17,82]. In cell biology, diffusion is the main form of transport
within cells and across cell membranes.
Convection: Convection is the collective movement of groups or aggregates of molecules
within fuids, through advection or through diffusion or as a combination of both of them.
Convection of mass cannot take place in solids, since neither bulk current fows nor sig-
nifcant diffusion can take place in solids. Diffusion of heat can take place in solids, but
that is called heat conduction. Convection can be demonstrated by placing a heat source
at the side of a glass full of a liquid and observing the changes in temperature in the glass
caused by the warmer fuid moving into cooler areas. Convective heat transfer is one of
the major types of heat transfer, and convection is also a major mode of mass transfer in
fuids. Convective heat transfer and mass transfer both take place by diffusion due to the
random Brownian motion of individual particles in the fuid. In advection, matter or heat
is transported by the larger-scale motion of currents in the fuid. In the context of heat
transfer and mass transfer, the term convection is used to refer to the sum of advective and
diffusive transfers.
Advection: Advection is a transport mechanism of a fuid due to the fuid’s bulk motion.
A simple example of advection is the transport of pollutants or silt in a river by bulk water
fow downstream. Another example of advection is energy or enthalpy. The fuid’s motion
is described mathematically as a vector feld, and the transported material is described
by a scalar feld showing its distribution over space. Since advection requires currents
in a fuid, it cannot happen in rigid solids. It does not include transport of substances by
molecular diffusion. In meteorology and physical oceanography, advection often refers to
the transport of some property of the atmosphere or ocean such as heat, humidity (mois-
ture), or salinity. Advection is important for the formation of orographic clouds and the
precipitation of water from clouds, as part of the hydrological cycle.
Dispersion/dispersal: Biological dispersal refers to those processes by which a species
maintains or expands the distribution of its population. Dispersal is necessary in popula-
tions because members of the species compete for the same limited resources within an
ecosystem. Dispersal relieves pressure on resources in an ecosystem. Dispersal mecha-
nisms depend on the competition for these resources. Dispersal may involve replacement
of a parent generation by a new generation with only minor changes in the geographic
area occupied. Dispersal enables the species population to occupy much of the avail-
able habitat and maximize its resources in its favor. Some organisms (plants and espe-
cially sedentary animals) have evolved adaptations for dispersal by taking advantage of
4 Spatial Dynamics and Pattern Formation in Biological Populations
various forms of kinetic energy occurring naturally in the environment like water fow
and wind. Often, dispersal may be purely random, density-dependent, or random plus
density-dependent. Dispersal over long distances can be approximated mathematically
by deterministic partial differential equations or integro-differential equations. Dispersal
over very short distances often results in large spatial correlations. In case of interact-
ing particle systems, local dispersal can result in spatial correlations. The global dispersal
results in a Poisson distribution which allows one to study spatially homogeneous models
that are at the onset of exhibiting spatial correlation, such as when offspring are dispersed
over intermediate distances [79].
Difference between diffusion and dispersion: Dispersive mass transfer in fuid
dynamics is the spreading of mass from highly concentrated areas to less concentrated
areas. Dispersive mass fux is analogous to diffusion. It can be described using Fick’s frst
law, J = −E ( dc/dx ), where c is the mass concentration of the species being dispersed, E is
the dispersion coeffcient, and x is the position in the direction of the concentration gradi-
ent. The following physical model of dispersion is explained well in the book Analysis of
Transport Phenomena by William M. Deen [18]. Consider a convective fow superposed over
the diffusion. If the fow velocity is uniform everywhere (plug fow), then molecules at dif-
ferent places in the fow will move with the same convective velocity, and it is only the dif-
fusion rate that will differentiate between them. Now, consider a fow in which gradients
exist, for example, a fully developed pipe fow. The fuid at the center of the pipe is moving
much faster than the fuid close to the walls. If there is diffusion in the radial direction,
then molecules can hop over from one streamline to the next and thereby will be trans-
ported over different distances due to the difference in velocities. This is caused indirectly
by diffusion in the radial direction. The amount of dispersion reduces with increasing
diffusion coeffcient. This is because the molecules will just be hopping from one stream-
line to another constantly and will not get suffcient time to be transported far from each
other. Depending on the situation, a diffusive process is classifed as self-diffusion, cross-
diffusion, mutual-diffusion, etc. A detailed review of the interpretations of these forms of
diffusion was given by Vanag and Epstein [111]. Wolfenbarger [113] introduced the follow-
ing defnition: When transportation takes place due to energy from within the organism,
the dispersion is termed as active, and when it takes place due to energy from outside the
body, the dispersion is termed as passive. The smaller the organism, the more it is subject
to the effect of environmental turbulence. Thus, the diffusion of small animals should
be considered as partly passive and partly active. For instance, bacteria and pollen in the
air and phytoplankton in the water diffuse almost passively, while many insects in fight
undergo varying proportions of passive diffusion and active diffusion, according to the
degree of movement of the environmental fuid.
Multicomponent diffusion: In gas mixtures or concentrated solutions where more than
one chemical species is present, diffusion coeffcient is not a constant or composition-inde-
pendent. In these cases, diffusion depends on intermolecular dependencies. The diffusion
equation is to take into account relations between the mass fux of one chemical species to
the concentration gradients of all chemical species present. The mathematical equations
are formulated from the Maxwell–Stefan description of diffusion [16,69]. This formulation
is often applied to describe gas mixtures, such as syngas in a reactor or the mix of oxygen,
nitrogen, and water in a fuel cell cathode. In Maxwell–Stefan diffusion, the choices of
dependent variables are not the species concentrations, but rather the species mole or mass
fractions. The diffusive mass fux of each species is, in turn, expressed in terms of the gra-
dients of the mole or mass fractions, using multicomponent diffusion coeffcients ese are
symmetric, so that an n-component system requires n(n − 1)/2 independent coeffcients to
Introduction to Diffusive Processes 5
details of the species distribution, there can be two different patterns – regular or chaotic
corresponding to two different regimes of the system dynamics [86]. Reaction–diffusion
models in ecology were studied by many authors [1,21,54,77,78,95]. Skellam [96,97,98] sug-
gested that the process of biological diffusion cannot be purely random. Animals often
concentrate together to form groups. In such cases, an effect that opposes diffusion occurs
due to behavioral patterns and interaction between individuals. One of the important fea-
tures that distinguish the movement of animals from the random motion of the inorganic
material is this delicate balance between “spreading” and “concentrating”. Dispersion in
the physical environment of species cannot be ignored while considering organism dis-
persal. Horizontal density variation of plankton in the form of plankton patchiness has
been of interest to biological oceanographers [11,101]. While a tendency for the formation
of patches exists, the effect of oceanic diffusion is to supply a mechanism to intra- and
interspecies relationships which causes instability of the ecosystem [56,93]. When move-
ment is diffusive, the traveling wave moves from the stronger population to the weaker.
However, by incorporating behaviorally induced directed movement toward the stronger
population, the weaker one can slow the traveling wave down, even reversing its direc-
tion. Hence, movement responses can switch the predictions of traditional mechanistic
models [87].
Why diffusion in neural systems? The electrical activities in the neurons are governed
via movement of ionic currents through neuron membranes. Izhikevich [45] has discussed
the mechanism of how ions diffuse down the concentration gradient through the mem-
brane and produce outward current. The positive and negative charges accumulate on the
opposite sides of the membrane surface creating an electric potential gradient across the
membrane called as the transmembrane potential or membrane voltage. This potential
slows the diffusion of K+ ions which are attracted to the negatively charged interior and
repelled from the positively charged exterior of the membrane. Most of these membrane
currents involve one of the four ionic species: sodium (Na+), potassium (K+), calcium (Ca2+),
and chloride (Cl−). The concentrations of these ions are different on the inside and the
outside of a cell, which creates electrochemical gradients – the major driving forces of
neural activity. The extracellular medium has high concentration of Na+, Cl−, and Ca2+ ions,
and the intracellular medium has high concentration of K+ ions. According to Fick’s frst
law, the ions will move from a region of high concentration to a region of low concentra-
tion gradient. In cell membrane, Na+ and Ca2+ ions move from outside the cell membrane
to inside the cell membrane, whereas the K+ ions move from inside to outside the cell
membrane. The movement of these ions is described mathematically by the diffusion pro-
cesses. For example, in models of nerve conduction, only the membrane potential spatially
interacts; the recovery and other variables interact only through the membrane potential
[24,109]. Reaction–diffusion equations describe a variety of phenomena in neuroscience.
The Hodgkin–Huxley equations describe the propagation of nerve pulses and the potas-
sium and calcium ion concentrations in cortical structures. Reaction–diffusion systems
have been shown to be able to propagate structured wave patterns without attenuation
to form stable patterns [39,67,73] and to select the maximum stimulus using only local
interactions [60,71,72,94,108]. A reaction–diffusion system requires the presence of a pair of
antagonistic neurotransmitters. Note that a neurotransmitter may interact with more than
one other neurotransmitter. The same system may also model diffusion neurotransmis-
sion of a single neurotransmitter without the antagonistic inhibitor. This can be treated
as a special case of a reaction–diffusion (RD) system [60]. The solutions of RD equations
display a wide range of behaviors including the formation of traveling waves in nerve
conduction, for example, traveling wave solutions of the full Hodgkin–Huxley equations
Introduction to Diffusive Processes 7
[6]. Some of the processes that can be modeled by the RD equations are the following: (i)
propagation of an action potential along an electric cable (HH equations), (ii) concentra-
tion changes of ionic species under the infuence of local reactions, and diffusion of ions
down their concentration gradients without expense of energy in an aqueous medium
(Kolmogorov–Petrovsky–Piskunov equations), and (iii) electrodiffusion (diffusion of
charged particles under the infuence of an electric feld), which is a nonlinear transport
process whose essence is diffusion of ions combined with their migration in an electric feld
(Nernst–Planck equations). However, electrodiffusion is the principal means of migration
only in the presence of large number of ions. Electrodiffusion has not yet been reconciled
with action potential propagation, and alternative ways (based on Maxwell’s equations
instead of RD systems) are appearing in theoretical models of information handling in the
brain [6].
˙ N ( x + ˝x , t) − N ( x , t) ˘ ( ˝x)2 ˙ N ( x + ˝x , t) − N ( x , t) ˘
F = −ˇ = − .
ˆ 2 a˝t 2 ˝t ˇˆ a( ˝x)2
The concentration per unit volume C and the diffusion coeffcient D are respectively,
defned as
N ( x , t) ( °x)2
C( x , t) = , and D = .
a°x 2°t
Hence,
D
F=−
˙x
[C(x + ˙x, t) − C(x, t)].
In the limit, we obtain the equation F = −DCx, which is Fick’s frst law of diffusion. Here,
x is the space coordinate measured normal to the section. The modern mathematical form
of the law is written as N i = − Di˛ci , where N i is the molar fux, Di is the diffusion coef-
fcient, and ci is the concentration of the species i. Crank [15] in his book The Mathematics
of Diffusion gave an excellent account of methods for the solution of diffusion equations.
8 Spatial Dynamics and Pattern Formation in Biological Populations
Fick’s second law: Consider the case when there is diffusion at the front and rear sur-
faces of an incremental planar volume. Fick’s second law states that “the rate of accumula-
tion (or depletion) of concentration within the volume is proportional to the local curvature
of the concentration gradient”. The second law of diffusion is a linear equation with the
dependent variable being the concentration of the chemical species under consideration.
Diffusion of each chemical species occurs independently. The mass transport systems
described by Fick’s second law are easy to simulate numerically. Accumulation is positive
when the curvature is positive (that is, when the concentration gradient is more negative
on the front end of the planar volume and less negative on the rear end so that more fux
is driven into the volume at the front end than is driven out of the volume at the rear end)
[102]. The accumulation (˜ C/˜ t) is proportional to the diffusivity D and the second deriva-
( )
tive (or curvature) of the concentration ˜ 2 C/˜ x 2 . Fick’s second law is given by
Ct = DCxx . (1.1)
˜C ˝ ˜ 2C ˜ 2C ˇ ˜C ˝ ˜ 2C ˜ 2C ˜ 2C ˇ
= Dˆ 2 + , = Dˆ 2 + + . (1.2)
˜t ˙ ˜x ˜ y 2 ˘ ˜t ˙ ˜x ˜ y 2 ˜ z 2 ˘
˜C ˜ ˛ ˜C ˆ ˜ ˛ ˜C ˆ ˜C ˜ ˛ ˜C ˆ ˜ ˛ ˜C ˆ ˜ ˛ ˜C ˆ
= ˙D ˘+ D , = ˙D ˘+ D + ˙D ˘. (1.3)
˜ t ˜ x ˝ ˜ x ˇ ˜ y ˙˝ ˜ y ˘ˇ ˜ t ˜ x ˝ ˜ x ˇ ˜ y ˙˝ ˜ y ˘ˇ ˜ z ˝ ˜ z ˇ
In the cylindrical and spherical polar coordinates, we have the following equations in
three dimensions:
˜C 1 ˜ ˝ ˜C ˇ ˜ ˝ D ˜C ˇ ˜ ˝ ˜C ˇ
Diffusion in cylinder: = ˆ rD + ˆ + ˆ rD (1.4)
˜t r ˜ r ˙ ˜ r ˘ ˜° ˙ r ˜° ˘ ˜ z ˙ ˜z ˘
˜C 1 ˜ ˙ 2 ˜C ˘ 1 ˜ ˙ ˜C˘ D ˜ 2C
Diffusion in sphere: = 2 ˇr D + ˇ Dsin° + (1.5)
˜t r ˜ r ˆ ˜ r sin° ˜° ˆ ˜° sin 2° ˜˛ 2
˜C
or as = div(D grad C).
˜t
Each term of the equation represents a potential energy per unit mass of water at a given
point. Here, hg = ° is the total energy per unit mass and also the potential ˜ of the water,
gz is the gravitational energy, and p/˜ is the pressure energy. Since hg = °, we get
K
q = −K grad h = − grad ˛. (1.7)
g
The force per unit mass acting upon the water at a given point is given by
where ˜ = k/(c° ) is the thermal diffusivity of the material. This is the simplest diffu-
2
sion equation. The domain may be semi-infnite, 0 < x < ° ,t > 0, or infnite −° < x < °, t > 0.
Suitable initial and boundary conditions are provided to solve the problem completely.
The diffusion equation (1.10) is also written as ut = Duxx , where D is the diffusion coef-
fcient. If density ρ or specifc heat c depends on the location x, then α is a function of x. In
this case, we may write the equation in the self-adjoint form as ut = ˙˝ K ( x ) ux ˆˇ x . However, it
is not always possible to write the equation in self-adjoint form. If the material properties
depend on t also, then the diffusion equation may be written as ut = [ a( x , t)ux ]x .
Consider the case of diffusion in a cylinder of infnite length with unit cross section. Let
u denote the concentration. Observe that u = Pt −1/2 e
− x2
is a solution of the above equa-
( 4° 2 t)
tion (1.10). If M is the amount of substance deposited in the plane x = 0 at time t = 0, then
the amount of substance diffusing remains constant [15]. The total amount of substance M
diffusing in the cylinder is given by (Crank [15]):
ˇ ˇ
( )dx.
˜ ˜
− x 2 / 4˛ 2 t
M= udx = Pt −1/2 e
−ˇ −ˇ
˜
2
M = 2P° e − ˛ d ˛ = 2P° (˙)1/2 ,
−ˇ
u = Pt −1/2 e
(
− x 2 / 4˜ 2 t )= M − x 2 /( 4˜ 2 t )
.
1/2 e
2˜ ( ˙t )
u( x , t) =
4K
˙ ˜ 2n1+ 1 e
n= 0
− ˇ 2 dn2 t
sin ( dn x ) , (1.11)
Introduction to Diffusive Processes 11
x
u( x ,t) = K erfc , (1.12)
2α t
∞
2
∫
2
where erfc(z) = 1 − erf(z) = e −t dt.
π
z
The solution of the diffusion equation in composite mediums in which the diffu-
sion coefficients are different, for example, for x > 0, the diffusion coefficient is D1 ,
and for x < 0, the diffusion coefficient is D2 can be obtained in terms of the basic
( )
solutions erf x 2 D1t and erf x 2 D2t .
( )
c. Consider the initial and boundary conditions as (generalization of (a)) u( x , 0) = f ( x),
∀x ∈[ 0, L ] , u(0, t) = u(L,t) = 0, ∀t > 0. The general solution is given by
∞ L
nπ
2
u( x ,t) = ∑ ∫ f (ξ )sin nπL ξ sin nπL x exp −α
2
L
n= 1
2
t dξ .
L
(1.13)
0
u( x ,t) = T1 + (T0 − T1 ) ∑(−1) erfc L −2xα+√t2nL + erfc L +2xα+√t2nL .
n= 0
n
(1.14)
If f (t) is taken as the unit step temperature, f (t) = uτ (t) = 0, for t < τ , and = 1, for
t ≥ τ . Then, u( x ,t) reduces to
t
x 1 ( )dτ = erfc x = 1 − erf x .
∫τ
− x 2 / 4α 2 τ
u( x , t) = e 2α t (1.16)
2α π 3/2
2α t
0
12 Spatial Dynamics and Pattern Formation in Biological Populations
2
ˆ 2 2
u( x , t) =
˙ ˜˜
0
˘ f (° )sin(˛° )d° e − ˛ t sin(˛ x)d˛ .
˘
ˇ0
(1.17)
g. Diffusion in a long circular cylinder [15]: In the case of a long solid circular cyl-
inder, diffusion is in the radial direction. The governing differential equation
becomes
˜u 1 ˜ ˛ ˜uˆ
= ˙ rD ˘ˇ , (1.19)
˜t r ˜r ˝ ˜r
˜u ˝ ˜ 2u 2 ˜uˇ
= Dˆ 2 + . (1.20)
˜t ˙ ˜r r ˜ r ˘
Under the transformation U = ur, the equation reduces to the linear fow equation:
˜U ˜ 2U
=D 2 .
˜t ˜r
Therefore, in many problems, it may be possible to express the solution in terms of
the solutions of the corresponding linear problems.
Consider the initial value or Cauchy problem for the heat equation on the real line:
with u( x , 0) = g( x), where f and g are given smooth functions. Let uˆ (˜ , t) denote the Fourier
transformation of u in the space variable defned by
1
uˆ (° , t) =
2˝ ˜e − i° x
u( x , t)dx ,
Introduction to Diffusive Processes 13
where ˜ ° is a parameter. Applying the Fourier transform to the heat equation (1.21), we
obtain
dû
= − ˜ 2 uˆ + fˆ (˜ , t), û(˜ , 0) = gˆ (˜ ).
dt
˜ fˆ (° , s)ds.
2 2
û(° , t) = ĝ(° )e − ° t + e − ° (t − s)
1
u( x , t) =
2˝ ˜e i° x
û(° , t)d°
=
˜ K(x − y, t)g(y)dy + ˜˜ K(x − y, t − s) f (y, s)dyds,
0
(1.22)
˜u ˜2u 1 ˜u ˜2u
= D(t) 2 , or = . (1.23)
˜t ˜x D(t) ˜ t ˜ x 2
t
Under the transformation T =
˜ D(° ) d° , or dT = D(t)dt, this equation becomes
0
˜u ˜2u
= .
˜ T ˜ x2
Therefore, the solution of the equation with constant D can be used to fnd u as a function
of T. The solution can then be written in terms of t.
When diffusion is dependent on concentration, the diffusion equation becomes
˜u ˜ ˛ ˜uˆ
= ˙D ˘. (1.24)
˜t ˜ x ˝ ˜ x ˇ
Under the Boltzmann variable transformation ˜ = x/(2 ˛ t), the diffusion equation reduces
to an ordinary differential equation as
du d ˝ du ˇ
−2˜ = D .
d˜ d˜ ˆ˙ d˜ ˘
14 Spatial Dynamics and Pattern Formation in Biological Populations
The above transformation is usually used when diffusion takes place in an infnite or
semi-infnite media when the concentration is initially constant. Further, the initial and
boundary conditions must be expressible in terms of ˜ alone. For example, the typical
initial conditions in an infnite medium u = u1 , x < 0, t = 0; u = u2 , x > 0, t = 0; transform to
u = u1 , ˜ = −˝ ; u = u2 , ˜ = +˝.
˜u ˝ ˜2u ˜2uˇ
= Dˆ 2 + 2 , (1.25)
˜t ˙ ˜x ˜y ˘
P ˆ x2 + y 2 P 2
ˆ r2
M=
˜ ˜− − t
exp ˇ−
˘ 4 Dt
dx dy =
t ˜ ˜
0 0
exp ˇ− r dr d° = 4DP.
˘ 4Dt
M ˝ r2 ˇ
Thus, concentration is given by u = exp ˙− ˘.
4°Dt ˆ 4Dt
The heat distribution in a rectangular plate is governed by the linear constant coeffcient
parabolic diffusion equation:
(where the diffusion coeffcient is taken as 1) under suitable initial condition and boundary
conditions. This is the simplest diffusion equation in two dimensions. Assume the condi-
tions as the following: Initial condition: u( x , y , 0) = f ( x , y ). Boundary condition: u( x , y , t) = 0,
on the boundary. The separable solution of (1.26) is given by
˝ n˛y ˇ − mnt
u( x , y , t) = ˜˜ A
m = 1 n= 1
mn
˝ m˛x ˇ
sin ˆ
˙ x0 ˘ sin ˆ
˙ y 0 ˘
e , (1.27)
4 x0 y0
˙ m˝x ˘ ˙ n˝y ˘ ˙ m 2 n2 ˘
where Amn =
x0 y 0 ˜ ˜ 0 0
f ( x , y )sin ˇ
ˆ x0 sin ˇ
ˆ y0 dy dx , and °mn = ˝ 2 ˇ 2 + 2 .
ˆ x0 y 0
˜2u 1 ˜u ˜2u
+ + = 0. (1.28)
˜ r 2 r ˜ r ˜ z2
˜u
u = 0, z = 0, r ˛ a; = 0, z = 0, r > a;
˜z
u = K , r ° 0, z = ˛; u = K , z ° 0, r = ˛ ;
where K is the concentration in the bulk of the solution. The solution is obtained in terms
of Bessel’s function of order zero as
2K sin(ma)
K−u=
˝ ˜
0
m
J 0 (mr )e −mz dm
(1.29)
2K ˇ a˙2
= tan −1 2 2 2
, where R = r + z − a .
˝ 2 2
˘ R + R + 4z a
˙ °u˘ 2K 2K
ˇˆ
° z z= 0
=−
˜ sin ( ma) J (mr ) dm =
0
0
a2 − r 2
. (1.30)
˜S ˝ ˜ 2S ˜ 2S ˜ 2Sˇ
= Dˆ 2 + 2 + 2 , (1.31)
˜t ˙ ˜x ˜y ˜z ˘
M
S( x , y , z, t) = , r 2 = x2 + y 2 + z2 .
(4˝Dt) 3/2
( 2
exp −r (4Dt) )
The ground is assumed to be a refecting plane at z = 0, implying the boundary condition:
(˜ S/˜ z ) = 0, at z = 0. If the point source is located at the ground, the solution of (1.31) on a
16 Spatial Dynamics and Pattern Formation in Biological Populations
semi-infinite domain (z > 0) with this boundary condition is twice the solution in infinite
space. Thus, for a ground source,
2M
S( x , y , z,t) = .
(4πDt)3/2 exp − r 2 (4Dt) ( )
The amount of release during an infinitesimal time dt is denoted by Qdt. Replacing M by
Qdt in the above solution and integrating with respect to time, the solution is obtained as
t
2Q r2 Q
S ( r,t ) =
∫ 4πDt′
exp −
4Dt′
dt′ =
r 1/2
,
0 2 πDr 1 − ϕ
( 4Dt )
If the diffusion process involves r density functions ui = ui (t , x), i = 1, 2,…, r , and admits
convection, then the system of coupled reaction–diffusion–convection equations can be
written as
m m
∂ ui ∂ i ∂ ui ∂ ui
− ∑
∂ t j , k = 1 ∂ x j
d jk (t , x)
∂ xk
+ ∑
j =1
b ij (t, x)
∂ xj
( )
= f i t , x , u1 , u2 ,…, ur ,
i = 1, 2,…, r ,
Introduction to Diffusive Processes 17
where d ijk are coeffcients of diffusion and b ij = b ij (t , x), i = 1,2, …., r, is a drift vector.
A reaction–diffusion (RD) model for biochemical cell polarization was proposed by Mori
et al. [75]. They found a wave-based phenomenon whereby a traveling wave is initiated at
one end of a fnite, homogeneous 1D domain, moves across the domain, but stalls before
arriving at the opposite end. They refer to this behavior as wave-pinning and observed
that this phenomenon was obtained from a two-component RD system obeying the fol-
lowing assumptions: (i) Mass is conserved and limited (there is no production or removal,
only exchange between one species and the other), (ii) one species is far more mobile than
the other (due to binding to immobile structures, or embedding in a lipid membrane), and
(iii) there is feedback (autocatalysis) from one form to further conversion to that form. Mori
et al. [76] also analyzed a bistable reaction–diffusion (RD) model for two interconverting
chemical species that exhibits a phenomenon of a form of wave-pinning (a wave of acti-
vation of one of the species is initiated at one end of the domain, moves into the domain,
decelerates, and eventually stops inside the domain, forming a stationary front).
Köhnke and Malchow [50] studied the emergence of stationary fronts in two-species
competition–diffusion models with particular emphasis on the stability against envi-
ronmental perturbations. Wave-pinning in the considered one-dimensional models is
not stable against environmental noise. They have demonstrated the wave-pinning in
competition–diffusion models in variable environments.
For the Fickian equation, any constant state is a steady state. For the Wereide and Chapman
equations, steady states are non-constant if the diffusivity is not constant. The Chapman
equation is a Fokker-Planck-type diffusion equation. These two equations are (i) satisfed
by the probability density functions of a stochastic process when the Stratonovich and
Itô integrals are considered, respectively, and (ii) considered as mathematical diffusion
models in a heterogeneous environment [4,110]. All the three diffusion laws show different
behaviors. Choi and Kim [14] have calculated and compared their steady states as follows:
Consider the one-dimensional case with the Neumann boundary conditions.
18 Spatial Dynamics and Pattern Formation in Biological Populations
˝
ut = ˙ˇ a( x) ( b( x)u)˝ ˘ , for 0 < x < 1, t > 0,
ˆ
(1.38)
˛
The steady state of the problem exists only if c0 = c1 and satisfes ˝ˆ a( x) ( b( x)u)˛ ˇ = 0.
˙ ˘
The steady state is given by
x
1 c1 c
u( x) =
b( x) ˜ a(y) dy + b(x) ,
0
0 < x < 1.
The steady states corresponding to the three diffusion laws (1.35)–(1.37) are
x
c1
u( x) =
˜ D(y) dy + c,
0
(1.39)
x
1 c1 c
u( x) =
D( x) ˜0
D( y )
dy +
D( x)
, (1.40)
c1 x + c
u( x) = . (1.41)
D( x)
If the zero fux boundary condition is given, c1 = 0, then the steady states are given by
u( x) = c, u( x) = c/ D( x) , and u( x) = c/D( x), respectively. Therefore, the steady state of Fick’s
law is constant, while the other steady states are functions of x, even if there is no fux
across the boundary.
corresponding Itô formula was not available. The authors [64] extended the Lyapunov
direct method to the Itô stochastic reaction–diffusion systems. They formulated the corre-
sponding Lyapunov stability theory and discussed stability in probability, asymptotic sta-
bility in probability, and exponential stability in mean square. They applied their theory
to study the stability of the Hopfeld neural network. Their results generalized the results
of Holden et al. [40] and Liao [61]. In the following, we briefy report the work of Luo et al.
[64]. The authors [47,64] consider the stochastic reaction–diffusion equation:
( )
dv(t , x) = ˜ x2 ( v(t , x)) + f ( t , x , v(t , x)) dt + g ( t , x , v(t , x)) dW (t , x) (1.42)
( t, x ) ˛t+ × G, with the initial condition v(t0 , x) = ˜ (x), x ˛G, and the boundary condition
0
(˜ v (t, x )/˜ N ) = 0, (t, x ) ˙t+ × ˜ G, where G = {x, x < l < +ˆ} ˇ r ; and
0
f ˜˛˝ + × G × n , n ˙ˆ , g ˜ ˛˝ + × G × n , n × m ˆ˙
(
defned on complete probability space ˛, , ( t )t ˝I , with natural fltration {t }t ˛ 0 , N )
is the outside unit normal vector of ˜G, and ˜ (x ) is a suitably smooth known function.
The authors assume that g ( t , x , v(t , x)) satisfes the integral linear growth condition, and in
addition, f and g satisfy the Lipschitz condition, that is, there exists a constant C > 0 such
that
(
g (t , x , v ) ˛ C 1 + v G , )
f ( t , x , v1 ) − f ( t, x , v2 ) G ˝ C v1 − v2 G ,
g ( t , x , v1 ) − g ( t, x , v2 ) G ˝ C v1 − v2 G ,
where v ( x ,.) G
˜v ( x,.) dx .
G
Without loss of generality, it was assumed that
Class K function If µ (.) ˆ C ˇ˘[ 0, r ] , is a strictly increasing function and µ(0) = 0, then
the function µ is said to be a Class K function. Suppose that µ °K. If µ (.) ˝C ˙ˆ + , + ˇ˘ and
µ °K, lim µ(r ) = +˝, then µ °K.
r˙+˝
As the authors in their works [58,115,117,120] have pointed out, a continuous function
V (t , ˜ ) is said to be positive-defnite if V (t , 0) = 0, and for some µ °K, V (t , ˜ ) ˙ µ ˜ . Write ( )
( )
C1,2 + × n ; + for the family of all nonnegative functions V (t , ˜ ) on + × n that are con-
tinuously twice differentiable in ˜ and once in t. If V ( t, ˜ ) ˝C1,2 + × n ; + , then defne ( )
an operator V ( t, ˜ ) from + × n to with respect to (1.42) by
1
V ( t , ˜ ) = Vt ( t, ˜ ) + V˜T ( t, ˜ ) f ( t , x , ˜ ) + trace ˆˇ g T ( t , x , ˜ ) V˜˜ ( x , t ) g ( t , x , ˜ ) ˘ , (1.43)
2
° V (t , ˜ ) ˙ ° V (t , ˜ ) ° V (t , ˜ ) ˘ ˙ ° 2 V (t , ˜ ) ˘
where Vt (t , ˜ ) = , V˜T (t , ˜ ) = ˇ ,…, , and V ( x , t ) = ˇ .
°˜ n
˜˜
°t ˆ °˜1 ˆ °˜ i °˜ j n × n
t ˜ t0 , G
˜
Applying the Itô formula to V ( t , v(t , x )) dx along system (1.42), one obtains that for all
˜
d V ( t , v(t , x)) dx
G
=
˜ ˘( LV (t, v(t, x)) + V (t, v)°ˇv(t, x)) dt +V (t, v)g(t, x, v(t, x))dW(t) dx.
G
T
v
T
v
(1)
The existence of the function V (t , v ) ˜C1,2 and other conditions in the classical Lyapunov
theorem on the stability of (1.42) are needed [66]. The following are some defnitions:
(
Lyapunov-A and Lyapunov-B functions: V ˛ C1,2 + × n ; + is called as a Lyapunov-A )
˜
function for (1.42), if L V (t , v)dx ° 0, and is called as a Lyapunov-B function for (1.42), if
G
˜ ˜
L V (t , v) dx ° −b V (t , v) dx , in which b > 0.
G G
The fundamental quantities in reaction–diffusion models are individual entities such as
atoms, molecules, bacteria, cells, or animals, which move and/or react in a stochastic man-
ner. If the number of individuals is large, then accounting for each individual is ineffcient.
In this case, often PDE models describing the average or mean behavior of the system are
used. If the number of individuals is large in certain regions and small in others, then a
stochastic model may be ineffcient in one region, and a PDE model is inaccurate in the
other. When a small number of individuals are involved, stochastic effects can play an
Introduction to Diffusive Processes 21
important role in the survival and spatiotemporal distribution of individuals. For example,
when chemical reactions occur in discrete steps at the molecular level, then the processes
are invariantly stochastic, which have been demonstrated experimentally for single-cell
gene expression events [56,81].
Bates et al. [3] discussed the random attractors for stochastic reaction–diffusion equa-
tions on unbounded domains. Ferm et al. [26] derived an adaptive hybrid method for sim-
ulation of stochastic reaction–diffusion equations. Kelkel and Surulescu [48] studied the
stochastic reaction–diffusion system modeling the pattern formation on seashells. Within
the same species of seashells, they exhibit a huge variety of beautiful and highly complex
patterns. The importance of the work was to study whether the diversity is due to a single
or more mechanisms. Following the work of Gierer-Meinhardt [35], the authors [48] pro-
posed a model and proved the existence of a positive solution for the resulting system. The
authors have also performed numerical simulations and compared them with the solu-
tions obtained using a deterministic approach.
Li and Kao [59] investigated the mean square stability of a class of SRD equations with
the Markovian switching and impulsive perturbations by means of the Lyapunov function
and stochastic analysis. Jifeng Hu et al. [41] discussed the stochastic analysis of RD pro-
cesses. The authors derived an estimator for the appropriate compartment size for simulat-
ing the RD systems and introduced a measure of fuctuations in a discretized system. They
have also given a computational algorithm for implementing the method.
Stochastic RD simulations were successfully used in many biological applications. Some
examples are the following: (i) models of signal transduction in E. Coli Chemotaxis [62], (ii)
oscillation of Min proteins in cell division [25], (iii) MAPK pathway [106], (iv) intracellular
calcium dynamics [29], (v) models of Hes1 gene regulatory network [103], and (vi) Actin
dynamics in flopodia [23].
Fabian Spill et al. [100] derived “Hybrid approaches for multiple-species stochastic
reaction-diffusion models”. They considered the case where the fundamental quantities
in such models are individual entities that move and/or react in a stochastic manner. The
authors developed a scheme that couples a stochastic reaction–diffusion system in one
part of the domain with its mean-feld analog (a discretized PDE model) in the other part
of the domain. The interface between the two domains occupies exactly one lattice site
and is chosen such that the mean-feld description is still accurate there and the errors due
to the fux between the domains are small. Each species comprises individuals that can
migrate to neighboring lattice sites, or react locally with entities of the same or other spe-
cies. The algorithm preserves mass at the interface between the stochastic and determin-
istic domains, and these domains need to be neither static nor connected. The scheme can
account for multiple dynamic interfaces separating multiple stochastic and deterministic
domains, and the coupling between the domains conserves the total number of particles.
The authors showed that the method is signifcantly faster to simulate on a computer than
the pure stochastic model. The authors applied a hybrid algorithm to solve the stochastic
Fisher-Kolmogorov equation:
˜n ˆ n ˜2n
= °n˘ 1 − + D 2 , (1.44)
˜t ˇ ˙ ˜x
where D is the diffusion coeffcient, ˜ is the growth rate, and ˜ is the carrying capacity.
The algorithm was used to simulate traveling wave solutions as the model parameters are
varied. The authors have also applied the hybrid algorithm for solving the spatial stochas-
tic Lotka-Volterra system. They considered the interaction reaction such that each time a
22 Spatial Dynamics and Pattern Formation in Biological Populations
prey (N) is eaten by a predator (M), a single new predator is born. The mean feld limit and
continuum limit correspond to the classical spatial Lotka-Volterra equations:
˜n ˜2n
= an − bnm + DN 2 , (1.45a)
˜t ˜x
˜m ˜2m
= − cm + bnm + DM , (1.45b)
˜t ˜ x2
where n = n( x , t) and m = m( x , t) are prey and predator densities related to N ( k ) and M( k ),
respectively. In the stochastic model, each species can jump to neighboring lattice sites,
the prey reproduce at rate a, predators die at rate c, and consume prey and produce at rate
b. They considered the following four cases depending on whether each of the prey and
predator evolves deterministically or stochastically: (i) deterministic prey and determin-
istic predator, (ii) deterministic predator and stochastic prey, (iii) stochastic predator and
deterministic prey, and (iv) stochastic prey and stochastic predator.
Many authors focused on constructing SRD models to solve specifc biological questions
or to analyze SRD algorithms or to develop SRD software. There are a number of recent
SRD approaches like (i) combining together in a multiscale framework [28,38] or (ii) com-
bining stochastic models with mean-feld descriptions [30]. There are many other works on
SRD models that we have not quoted or discussed.
ut = f1 ( ˜ , u, v ) + d1uxx ,
u ( x , 0 ) = u0 ( x ) , v ( x, 0 ) = v0 ( x ) ,
Defne F ( ˜ ,U ) by
˝ f1 (˜ , u, v) − A(˜ )u − B(˜ )v ˇ
F(˜ , U) = ˆ , where U = (u, v)T X.
ˆ˙ f2 (˜ , u, v) − C(˜ )u − D(˜ )v ˘
Introduction to Diffusive Processes 23
Then, system (1.46) can be rewritten into the following abstract form:
dU
= L(˜ )U + F(˜ , U). (1.47)
dt
dU
= L ( ˜0 )U + F0 (U ), where F0 (U ) = F ( ˜0 ,U ) . (1.48)
dt
The authors [116] considered the real and complex-valued Sobolev spaces.
Let A(˜ ) = f1u (˜ , 0, 0), B(˜ ) = f1v (˜ , 0, 0), C(˜ ) = f2 u (˜ , 0, 0), and D(˜ ) = f2 v (˜ , 0, 0). Defne the
linear operator L(˜ ) with the domain DL ( ° ) = X as
˙ °2 ˘
ˇ d1 2 + A(˜ ) B(˜ )
°x
L(˜ ) = ˇ . (1.49)
ˇ °2
ˇ C(˜ ) d2 + D(˜ )
ˆ °x 2
ˆ ˆ d n2
˘ A(˜ ) − ˘ 1 2 B(˜ )
˘ ˇ
Ln ( ˜ ) = ˘ , n 0 . (1.50)
˘ ˆ d2 n
2
C(˜ ) D(˜ ) − ˘ 2
˘ ˇ
ˇ
˜ 2 − ˜ Tn ( ° ) + Dn ( ° ) = 0, n ˘ 0 (1.51)
d1d2 n4 n2
Dn (˜ ) = 4
− [ d1D(˜ ) + d2 A(˜ )] 2 + A(˜ )D(˜ ) − B(˜ )C(˜ ) (1.52)
d˜
2 ( °0 ) = ˘Aˆ ( °0 ) + Dˆ ( °0 ) . (1.53)
d°
Condition 1 has the following equivalent form:
d˜
Therefore, ( °0 ) ˙ 0. Now, ˜ 0 = Dn ( °0 ) , and B ( ˜0 ) ,C ( ˜0 ) cannot be equal to zero simul-
d°
taneously. Therefore, the transversality condition holds and the system (1.46) undergoes
Hopf bifurcation. This establishes the existence of Hopf bifurcation. The authors used the
central manifold theorem and normal form as defned in Hassard et al. [37] to obtain the
conditions under which a family of periodic solutions bifurcates from the positive steady-
state solution of (1.46), when the control parameter crosses through the critical value. The
authors derived the formulas to decide the direction of the Hopf bifurcation, stability, and
period of bifurcating periodic solutions arising through the Hopf bifurcation. The authors
transformed the system (1.48) into the following system in (z, w) coordinates:
dz dw
= i˜ 0 z + q* , F0 , = L ( °0 ) w + H ( z, z , w), (1.54)
dt dt
where H ( z, z , w ) = F0 − q* , F0 q − q * , F0 q , F0 = F0 ( zq + z q + w ) .
The frst equation of (1.54) gives the equation of reaction–diffusion system (1.46) restricted
on the center manifold at ( ˜0 , 0, 0 ) as defned by Hassard et al. [37]. The application of the
theory gives the equation:
dz
dt
= i° 0 z + ˜ k
g kj k j
! j !
( )
4
z z +˛ z , (1.55)
2˘ k + j ˘ 3
The dynamics of (1.54) can be determined by the dynamics of (1.55). In addition, it can
be observed from the work of Hassard et al. [37] that when ˜ approaches ˜0 suffciently
closely, the Poincare normal form of (1.47) has the form
M
z = (° (˛ ) + i˝ (˛ )) z + z ˜c (˛ )(zz ) ,
j=1
j
j
(1.56)
1 * 2 ° 1 * 1 *
− q , Qq q *
˝˛ + q , Qw 11q + 2 q , Qw 20 q + 2 q , Qqqq , (1.57)
3
1
µ2 = ° ˘˘(0) = −
˛ ˘ ( °0 )
( ) (
Re C1 ( °0 ) , ˝ 2 = 2Re C1 ( °0 ) . ) (1.58)
The direction of Hopf bifurcation and stability of the bifurcating periodic solutions of
(1.46) at ( ˜0 , 0, 0 ) can be determined by the sign of Re ˝˙ c1 ( ˜0 ) ˆˇ . For ˜ = ˜ ( s), where s is suf-
fciently small, there exists a family of T ( s) periodic, continuously differentiable solutions
( u(s), v(s))(x, t) of system (1.46) such that u(0) = 0 = v(0), and
T2 = Tˇˇ(0) = −
4
( )
Im c1 ( °0 ) −
( )
Re c1 ( °0 ) ˜ ˇ ( °0 )
2 . (1.59)
˜ 0 ˛ ˇ ( °0 )
Introduction to Diffusive Processes 25
The following Hopf bifurcation theorem for the general RD equations (1.46) summarizes
the above results.
Assume that Condition 1 holds. Then, the model system (1.46) undergoes
i. a supercritical (or subcritical) Hopf bifurcation at (0,0) when ˜ = ˜0 , if µ2 > 0 (or < 0),
˜ ˇ ( °0 ) ]Re[ c1 ( °0 ) < 0 (or > 0).
i.e., ˘1/
ii. In addition, if all the other eigenvalues of L ( ˜0 ) have negative real parts, then
the bifurcating periodic solutions are stable (or unstable) when ˜ 2 < 0 (or > 0), i.e.,
Re ˝˙ c1 ( ˜0 ) ˆˇ < 0 (or > 0).
iii. T2 determines the period of the bifurcating periodic solutions. The period increases
if T2 > 0 and decreases if T2 < 0.
˜u
= f (u, v) + d1˝ 2 u + d12 ˝ 2 v = a11u + a12 v + d1˝ 2 u + d12 ˝ 2 v , (1.60a)
˜t
˜v
= g(u, v) + d21˝ 2 u + d2 ˝ 2 v = a21u + a22 v + d21˝ 2 u + d2 ˝ 2 v , (1.60b)
˜t
where ˜ 2 is the two-dimensional Laplace operator and ( x , y ) ˜ R 2 . The initial and boundary
conditions are
˜ u( x , y , t) ˜ v( x , y , t)
= = 0, t > 0, ( x , y ) ˙˜ ˆ, ˆ ˙ R 2 . (1.61b)
˜° ˜°
Amplitude of the modes in the solution of (1.60) cannot be directly determined. One can
obtain an approximation to the amplitude using the Taylor series expansion. Expand the
right-hand sides in (1.60) about the equilibrium point ( u0 , v0 ), and truncate the expansion
at third order. We can use multiple-scale analysis to derive the amplitude equations, when
k = kc . When the controlled parameter d12 = d12
cr
is larger than the critical value of the Turing
point, the solution of the system (1.60) can be expanded as (Zheng and Shen [122]),
N
c = c0 + ˜(Z e ) ,
i=1
i
iki ˆr
26 Spatial Dynamics and Pattern Formation in Biological Populations
with k = kc , Z j and the conjugate Z j are the amplitudes associated with the modes k j and
−k j. Close to onset of d12 = d12
cr
, one obtains
˜ Zi
= si Zi + Fi ( Zi , Z j ,) ,
˜t
where si is the coeffcient of the linear term of the variable Zi . Applying the center manifold
theory near the Turing bifurcation point, it can be concluded that amplitude Z j satisfes the
equation:
˜ Zi
˜t
(
= Fi Zi , Zi , Z j , Z j , . )
From the standard multiple-scale analysis, the spatiotemporal evolution of the amplitudes
up to the third order in the perturbations can be written as
˛ Zi
°0
˛t
= µZi + ˜h lm
lm Zl Zm + ˜g
lm
lmn Zl Zm Zn . (1.62)
˛ Zi iki r
°0
˛t
e = µZi e iki r + ˜h
lm
lm Zl Zm e i( kl + km )r + ˜g
lm
lmn Zl ZmZne i( kl + km + kn )r . (1.63)
Comparing (1.62) with (1.63), it can be concluded that the two equations hold only if
ki = kl + + km. Let the system (1.60) be rewritten as
˜c
= Lc + N(c), (1.64)
˜t
° u ˙ ° a + d 2 a12 + d12 2 ˙
11 1
where c = ˝ ˇ , L = ˝ ˇ,
˛ v ˆ ˝˛ a21 + d21 2 a22 + d2 2 ˇˆ
˛ f u2 + 2 f uv + f v 2 ˆ ˛ f u3 + 3 f u2 v + 3 f uv 2 + f v 3 ˆ
1 uu uv vv 1 uuu uuv uvv vvv
N= ˙ ˘+ ˙ ˘.
2 ˙ g uuu2 + 2g uv uv + g vv v 2 ˘ˇ 6 ˙˝ g uuuu3 + 3g uuv u2 v + 3g uvv uv 2 + g vvv v 3
˝ ˇ˘
cr
Here, L is the linear term and N is the nonlinear term. When d12 is close to d12 , expand d12 as
d12 − d12
cr
= ˜ d12
(1)
+ ˜ 2 d12
(2)
+ ˜ 3 d12
(3)
+ o ˜3 , ( )
where ε is a small parameter. Expand c and N in terms of ˜ as
Introduction to Diffusive Processes 27
ˆ u ˆ u1 ˆ u2 2
c=˘ =˘ ˜ +˘ ˜ +…
ˇ v ˇ v1 ˇ v2
ˆ f u2 + 2 f u ° + f ° 2 ˆ f u u + f ( u ° + u ° ) + f ° °
1 uu 1 uv 1 1 uv 1 uu 1 2 uv 1 2 2 1 uv 1 2
N= ˘ ˜ 2 + ˘
2 ˘ g uu u1 + 2g uv u1°1 + g uv° 12
2
˘ g uu u1u2 + g uv ( u1° 2 + u2° 1 ) + g vv° 1° 2
ˇ ˇ
ˆ f u3 + 3 f u2 v + 3 f u v 2 + f v 3
1 uuu 1 uuv 1 1 uvv 1 1 vvv 1
+ ˘ ˜ 3.
6 ˘ g uuu u1 + 3g uuv u1 v1 + 3g uvv u1 v1 + g vvv v13
3 2 2
ˇ
L = Lc + ˜ Md12
(1)
+ ˜ 2 Md12
(2)
+ o ˜2 ,( ) (1.65)
˝ a + d ˛2 cr 2 ˇ ˝
where Lc = ˆ
11 1 a12 + d12 ˛
M=ˆ 0 ˛2 ˇ .
,
ˆ˙ a21 + d21˛ 2 a22 + d2˛ 2 ˘ ˙ 0 0 ˘
˜ ˜ ˜ ˜
= +° + °2 +
˜ t ˜ T0 ˜ T1 ˜ T2
˛ u ˆ
3
˛ xi ˆ
c=˙ ˘=
˝ v ˇ
˜ ˙˝
i=1
yi ˇ
ik r
˘ e i + c c.
This expression implies that the bases of the solutions have nothing to do with time and
the amplitude Z is a variable that changes slowly. As a result, one can write
˜Z ˜Z ˜Z
=° + °2 + (1.66)
˜t ˜ T1 ˜ T2
Comparing different orders of ˜ in (1.65), one obtains
˜ u1 ˝
Lc ˛ ˆ = 0, (1.67)
° v1 ˙
˝ u3 ˇ ˜ ˝ u2 ˇ ˜ ˝ u1 ˇ (1)
˝ u2 ˇ (2)
˝ u1 ˇ
Lc ˆ = ˆ + ˆ − d12 M ˆ − d12 M ˆ
˙ v3 ˘ ˜ T1 ˙ v2 ˘ ˜ T2 ˙ v1 ˘ ˙ v2 ˘ ˙ v1 ˘
˝ f u3 + 3 f u 2 v + 3 f u v 2 + f v 3 ˇ ˝ G ˇ
1 uuu 1 uuv 1 1 uvv 1 1 vvv 1 u
− ˆ =ˆ . (1.69)
6 ˆ guuu u13 + 3guuv u12 v1 + 3guvv u1 v12 + g vvv v13 ˘ ˙ Gv ˘
˙
Consider the case of the frst order in ˜ . Since Lc is the linear operator of the system close
to the onset, ( u1 , v1 ) is the linear combination of the eigenvectors that correspond to the
T
° u ˙
3
° xi ˙
˝ v ˇ=
˛ ˆ
˜ ˝˛
i=1
yi ˆ
ik r
ˇ e i + c c,
a22 − d2 kcr2
we obtain that xi = By i , where B = . If y i = 1 is assumed, then xi = B, and
d21kcr2 − a21
° u1 ˙ ° B ˙ ° 3
˙
˝ ˇ =˝ ˇ˝
˛ v1 ˆ ˛ 1 ˆ ˛
˜W e
i=1
i
iki r
+ c c ˇ , i = 1, 2, 3
ˆ
˛ u2 ˆ
(1 A ) e −ik .r Lc ˙
i
˘ = 0.
˝ v2 ˇ
Putting the value of ( u1 , v1 ) in the above equation and equating the coeffcient of e ik1r, one
obtains the equations
˜ W1
(A + B) = − kcr2 d12
(1)
W1 + ( f2 + Af3 ) W2W3 .
˜ T1
˝W2
( A + B) = −kcr2 d12
(1)
W2 + ( f2 + Af3 ) W1W3 . (1.70)
˝T1
˝W3
(A + B) = −kcr2 d12
(1)
W3 + ( f2 + Af3 ) W1W2 ,
˝T1
Introduction to Diffusive Processes 29
˛ u2 ˆ ˛ U 0 ˆ 3 ˛ Uj ˆ 3 ˛ U jj ˆ ˛ U 12 ˆ i( k − k )r
Write ˙ ˘ =˙ ˘+
˝ ° 2 ˇ ˝ V0 ˇ
˜ ˙
˙ Vj
j=1 ˝
˘ˇ
ik r
˘e j + ˜ ˙
˙ Vjj
j =1 ˝
˘ˇ
2ik r
˘e j +˙
˝ V12
˘e
ˇ
1 2
˛ U 23 ˆ i( k − k )r ˛ U 31 ˆ i( k − k )r
+˙ ˘e
2 3
+˙ ˘e
3 1
+ c.c. (1.71)
˝ V23 ˇ ˝ V31 ˇ
Substituting (1.71) into (1.68), collecting the coeffcients and comparing, one gets the
equations:
−1
˛ U 0 ˆ ˛ a11 a12 ˆ ˛ − f2 ˆ
˙ ˘ =˙
˝ V0 ˇ ˝ a21
˘ ˙
a22 ˇ ˙˝ − f3
(
2 2
˘ W1 + W2 + W3
˘ˇ
2
)
(1.72)
˛ Zu0 ˆ
=˙
2 2
˘ W1 + W2 + W3
˝ Zv0 ˇ
2
( )
U j = BVj , j = 1, 2, 3 (1.73)
−1
˜ f2 ˝
˜ U 11 ˝ ˜ a11 − 4 kcr2 d1 a12 − 4 kcr2 d12
cr ˝ ˛ − ˆ ˜ Zu1 ˝ 2
2
˛ ˆ =˛ ˆ ˛ ˆ W12 = ˛ ˆ W1 , (1.74)
° V11 ˙ ˛° a21 − 4 kcr d21 a22 − 4 kcr2 d2 ˆ˙ ˛ − f3 ˆ ° Zv 1 ˙
2
˛° 2 ˆ˙
−1
˜ U 12 ˝ ˜ a11 − 3 kcr2 d1 a12 − 3 kcr2 d12
cr ˝ ˜ − f2 ˝ ˜ Zu2 ˝
˛ ˆ =˛ ˆ ˛ ˆ W1 W2 = ˛ ˆ W1 W2 . (1.75)
° V12 ˙ ˛° a21 − 3 kcr d21
2
a22 − 3 kcr d2 ˆ˙
2 ˛° − f3 ˆ˙ ° Zv 2 ˙
By permuting the suffxes, one can get the coeffcients of other terms of (1.71).
( )
T
For the third order of ˜ as in (1.69), the coeffcient of e ik1 .r . , denoted by Gu1 Gv1 , is given by
˛ ˛ W1 V1 ˆ ˆ
˛ G1 ˆ ˙ B ˙˝ T + T ˘ˇ ˘ ˛ (1) ˆ ˛ BV1 ˆ ˛ (2) ˆ ˛ BW1 ˆ
˘ = ˙˙ ˘ + k2 0 d12 0 d12
u 2 1
˘ + k cr ˙
2
˙ ˘ cr ˙ ˘˙ ˘˙ ˘
˙˝ Gv1 ˘ˇ W1 V1 ˙˝ 0 0 ˘ˇ ˝ V1 ˇ ˙˝ 0 0 ˘ˇ ˝ W1 ˇ
˙ + ˘
˙˝ T2 T1 ˘ˇ
˙
˛
((( f uu (( f B + f ) ( Z + Z )
B + f uv ) ( Zu0 + Zu1 ) + ( f uv B + f vv ) ( Zv0 + Zv1 ) W1 + ) 2
uu uv u0 u2
ˆ
˘
˙ ˘
+ ( f B + f ) ( Z + Z )) ( W + W )) W + f (W V + W V )
2 2
˙ uv vv v0 v2 2 3 1 2 2 3 3 2 ˘
−˙ ˘
˙
˙ ((( g B + g ) ( Z + Z ) + ( g B + g ) ( Z + Z ))) W + (( g B + g ) ( Z + Z )
uu uv u0 u1 uv vv v0 v1 1
2
uu uv u0 u2
˘
˘
˙ ˘
˙ + ( g B + g ) ( Z + Z )) ( W + W )) W + f (W V + W V ) ˘
2 2
˙˝ uv vv v0 v2 2 3 1 3 2 3 3 2 ˘ˇ
˛
˙
−˙
(W 1
2 2
+ W2 + W3
2
)( f uuu B3 + 3 f uuv B2 + 3 f uvv B + f vvv ) ˆ
˘ (1.76)
˘ W1 .
˙
˝ (W 1
2 2
+ W2 + W3
2
)( g uuu B + 3g uuv B + 3g uvv B + g vvv
3 2
) ˘
ˇ
30 Spatial Dynamics and Pattern Formation in Biological Populations
( ) ( )
T T
Similarly, Gu2 Gv2 and Gu3 Gv3 can be obtained by permuting the subscripts.
Using the Fredholm solvability condition as in the second order of ˜ , we obtain
˛ Gj ˆ
(1 A ) ˙˙ u
˘ = 0, j = 1, 2, 3,
u ˘
j
˝ G ˇ
˙ ˝W1 ˝V1 ˘
(A + B) ˇ +
ˆ ˝T2 ˝T1
(
= −kcr2 d12(1)V1 + d12(2)W1 + h1 W2V3 + W3V2 − G1 W1
2
) ( ) (
(1.77)
(
+ G2 W2 + W3
2 2
)) W , 1
where
h1 = f2 + Af3 ,
˘
(
ˆ ( fuuB + fuv ) + A ( g uuB + g uv ) ( Zu0 + Zu1 ))
,
G1 = −
˘
( )
ˇ + ( fuv B + f vv ) + A ( g uv B + g vv ) ( Zv 0 + Zv 1 ) + ( f 4 + Ag 4 )
˘
(
ˆ ( fuuB + fuv ) + A ( g uuB + g uv ) ( Zu0 + Zu2 ))
,
G2 = −
˘
( )
ˇ + ( fuv B + f vv ) + A ( g uv B + g vv ) ( Zv 0 + Zv2 ) + ( f 4 + Ag 4 )
Similarly, the other two equations can be obtained by permuting the subscripts. The ampli-
tudes, Z j , j = 1, 2, 3, are taken as variables that change slowly with respect to time so that
( )
˜ Z j ˜ T0 = 0. Hence,
˜ Zj ˜ Zj ˜ Zj
˜t
=°
˜ T1
+ °2
˜ T2
+ o °2 . ( ) (1.78)
( )
Setting Z j = ˜ Wj + ˜ 2Vj + o ˜ 2 , and using the above results, one can get the amplitude
equations corresponding to Z1 , Z2 , Z3 as follows:
˜0
° Z1
°t
2
(2 2
= µZ1 + hZ2 Z3 − g1 Z1 + g 2 Z2 + g 2 Z3 Z1 , )
˜0
° Z2
°t
2
(
2 2
= µZ2 + hZ1Z3 − g1 Z2 + g 2 Z1 + g 2 Z3 Z2 , )
˜0
° Z3
°t
2
(2 2
= µZ3 + hZ2 Z1 − g1 Z3 + g 2 Z2 + g 2 Z1 Z3 , ) (1.79)
Another random document with
no related content on Scribd:
CHAPTER XII.
CONCLUSION.
The war now being concluded, and the questions which had for more
than a century proved so vexatious, being permanently settled, the people of
the United States had plenty of leisure time to count up the cost of the
"economy" which their rulers had been treating them to, ever since the end
of the Civil War.
As has already been shown, these demagogues, while prating about the
surplus, and the tariff, and the down-trodden laboring man, and the crime of
spending the people's money for anything but river and harbor and public
building jobs, and exorbitant premiums on immature bonds; had permitted
the United States Navy to go to decay, from motives of "economy;" had
utterly refused to offer even decent rates of compensation for the carrying
of foreign mails in American ships, for the sake of "economy;" had declined
to encourage the establishment of an auxiliary naval force, by the payment
of an annual rental for the privilege of employing swift American built
steamships as cruisers in time of war, because by not doing so, the treasury
would save $5,000,000 or $10,000,000 a year, and these statesmen could
continue to pose before the country as champions and apostles of
"economy;" and had neglected year after year to fortify the seaboard cities,
notwithstanding the constant and oft repeated warnings uttered by military
and naval experts, intelligent and thoughtful writers, and farsighted
statesmen of both political parties; because, forsooth, when year after year
the River and Harbor Bill and the Public Buildings Bill had been inflated to
the largest possible figures that would be likely to escape a presidential
veto, they found that they could not spend any money on fortifications
without exceeding the appropriations made by their predecessors, and
would thus render themselves liable to be considered by their constituents
as lacking in the great essential element of "economy."
Well, these economical statesmen had had their way—and their day—
and their constituents and masters—the People of the United States—said to
them, "Gentlemen, render an account of your stewardship. Let us see what
your loud professions of 'economy' for the past twenty-five or thirty years
amount to. Let us have an itemized account, debit and credit, and see how
far your acts have been justified by results."
CREDIT.
DEBIT.
"There are other items, gentlemen, which could easily be made to swell
the above debit balance; but these are sufficient. You may step down and
out. The people of the United States are the most wealthy and liberal people
in the world, but ten years more of such "economy" as yours has been,
would bankrupt us. We wish you no harm, gentlemen, but we have no
further need of your services."
At the first glance, the Battle of the Swash seemed to have been a most
disastrous event for the United States.
England got all the glory and all the money, and the United States got
Canada and—the experience. But the latter proved to be worth infinitely
more than it cost, in that it exploded the absurd system of miscalled
"economy," which only "saved at the spigot to waste at the bung."
Let us rejoice that in this year of grace 1930, we have so profited by the
errors of our ancestors, that we now occupy unchallenged, the foremost
position among the Nations of the earth; and that with our 200,000,000 of
intelligent, prosperous and contented citizens—each one a sovereign in his
own right—we can afford to look with indifference upon the wars and
struggles of our less fortunate contemporaries on the other side of the
Atlantic.
Too late, alas! had the truth and wisdom of these words—written by that
great founder of the Government, Thomas Jefferson—become manifest.
"But on the seaboard they are open to injury, and they have then, too, a
commerce which must be protected.
"If particular Nations grasp at undue shares of our commerce, and more
especially, if they seize on the means of the United States, to convert them
into aliment for their own strength, and withdraw them entirely from the
support of those to whom they belong, defensive and protective measures
become necessary on the part of the Nation whose marine sources are thus
invaded, or it will be disarmed of its defense, its productions will be at the
mercy of the Nation which has possessed itself exclusively of the means of
carrying them, and its politics may be influenced by those who command its
commerce.
The "disgrace and losses" incurred by our ancestors in this brief but
disastrous campaign, had indeed brought "home to their feelings the evils of
having abandoned" the great interest thus earnestly pleaded for by the
greatest statesman of his day; and the absurd folly of the so-called
"economy," which prompted its abandonment, was at length reluctantly
conceded by the noisiest and bitterest advocates of free trade throughout the
land.
THE END.
APPENDIX.
SECTION 1. Be it enacted that from and after the passage of this Act,
the Secretary of the Navy is authorized and directed to execute contracts
with the owners or agents of steam vessels, built and owned in the United
States of America, and sailing under the American flag, upon the terms and
conditions hereinafter set forth; and the faith of the United States is hereby
pledged to carry said contracts into effect, and to make all payments arising
thereunder, in strict and prompt accordance with the terms thereof.
1st. The said vessels must be of at least four thousand tons burden.
2nd. The said vessels must have a maximum speed of at least nineteen
statute miles per hour, the same to be established by actual test.
3rd. The said vessels must be built, both as to their hulls and engines,
on plans to be approved by a Bureau of Naval Construction and
Engineering, consisting of five members, to be appointed for that
purpose by the President and the Secretary of the Navy; and every
such vessel shall have carriages or platforms for two or more guns
of at least eight inches calibre, and also suitable accommodations
for an auxilliary battery of rapid firing guns.
SECTION 3. Upon the completion of any such vessel, and the filing in
the office of the Secretary of the Navy of certified copies of the report of
the Bureau of Naval Construction and Engineering, approving the plans of
the hull and engines of such vessel, and also certifying that the estimate of
the cost of such vessel is not excessive; and after actual tests of the speed of
such vessel, under regulations imposed by the Secretary of the Navy; the
said Secretary shall, if requested to do so by the owner or owners of such
vessel, enter into a contract with such owner or owners, for a period of not
less than ten, nor more than fifteen years, by which, in consideration of the
privilege of taking such vessel by charter or purchase, at any time during
the pendency of the said contract, on ten day's notice, the United States
shall pay as an annual rental, to the owners or agents of such vessel, interest
on her cost, as certified by the Bureau of Construction and Engineering, as
follows:
1st. For vessels having a speed of nineteen statute miles or over, and
less than twenty statute miles per hour, four per cent.
2nd. For vessels having a speed of twenty statute miles or over, and less
than twenty-one statute miles, per hour, four and one-half per cent.
3rd. For vessels having a speed of twenty-one statute miles or over, per
hour, five per cent.
Provided. That such annual rental shall in no case exceed the sum of one
hundred thousand dollars for any single vessel.
Such valuation shall govern until such vessel shall have attained the age
of five years; and thereafter a deduction of six per cent. shall be made
therefrom annually, to cover depreciation and wear and tear.
SECTION 5. In case the Secretary of the Navy shall elect to charter such
vessel or vessels, instead of purchasing them, the rate to be paid therefor
shall be five dollars per month per registered ton; such vessel to be at the
risk of the United States during the pendency of such charter.
Updated editions will replace the previous one—the old editions will
be renamed.
1.D. The copyright laws of the place where you are located also
govern what you can do with this work. Copyright laws in most
countries are in a constant state of change. If you are outside the
United States, check the laws of your country in addition to the terms
of this agreement before downloading, copying, displaying,
performing, distributing or creating derivative works based on this
work or any other Project Gutenberg™ work. The Foundation makes
no representations concerning the copyright status of any work in
any country other than the United States.
• You pay a royalty fee of 20% of the gross profits you derive from
the use of Project Gutenberg™ works calculated using the
method you already use to calculate your applicable taxes. The
fee is owed to the owner of the Project Gutenberg™ trademark,
but he has agreed to donate royalties under this paragraph to
the Project Gutenberg Literary Archive Foundation. Royalty
payments must be paid within 60 days following each date on
which you prepare (or are legally required to prepare) your
periodic tax returns. Royalty payments should be clearly marked
as such and sent to the Project Gutenberg Literary Archive
Foundation at the address specified in Section 4, “Information
about donations to the Project Gutenberg Literary Archive
Foundation.”
• You comply with all other terms of this agreement for free
distribution of Project Gutenberg™ works.
1.F.
1.F.4. Except for the limited right of replacement or refund set forth in
paragraph 1.F.3, this work is provided to you ‘AS-IS’, WITH NO
OTHER WARRANTIES OF ANY KIND, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO WARRANTIES OF
MERCHANTABILITY OR FITNESS FOR ANY PURPOSE.
Please check the Project Gutenberg web pages for current donation
methods and addresses. Donations are accepted in a number of
other ways including checks, online payments and credit card
donations. To donate, please visit: www.gutenberg.org/donate.
Most people start at our website which has the main PG search
facility: www.gutenberg.org.