A food-web based unified model of "macro"- and "micro-" evolution
Extracted source text ยท arXiv cond-mat/0305322
\documentclass[twocolumn,pre,superscriptaddress]{revtex4}
\usepackage{latexsym}
\usepackage{graphicx}
\usepackage{epsfig,subfigure}
\begin{document}
\title{A food-web based unified model of ``macro''- and ``micro-'' evolution}
\author{Debashish Chowdhury{\footnote{E-mail: debch@iitk.ac.in}}}
\affiliation{Department of Physics, Indian Institute of Technology, Kanpur 208016, India.}
\author{Dietrich Stauffer{\footnote{E-mail: stauffer@thp.uni-koeln.de}}}
\affiliation{Institute for Theoretical Physics, Cologne University, D-50923 K\"oln, Euroland.}
\begin{abstract}
We incorporate the generic hierarchical architecture of foodwebs into
a ``{\it unified}'' model that describes both ``micro'' and ``macro''
evolutions within a single theoretical framework. This model describes
the ``micro''-evolution in detail by accounting for the birth, ageing
and natural death of individual organisms as well as prey-predator
interactions on a hierarchical dynamic food web. It also provides a
natural description of random mutations and speciation/orgination of
species as well as their extinctions. The distribution of lifetimes of
species follows an approximate power law only over a limited regime.
\end{abstract}
\maketitle
%Evolution | Monte Carlo simulation | Extinction | Lifetime distribution
\bigskip
\noindent {\bf PACS Nos. 87.23-n; 87.10.+e}
\section{Introduction}
The questions of ``origin'' and ``evolution'' have always fascinated
scientists in all disciplines. Physicists have focussed attention
mostly on cosmological evolution and origin of universe. On the other
hand, chemists and biologists have studied chemical evolution (i.e.,
formation of elements and compounds) as well as pre-biotic evolution
and ``origin'' of life. Similarly, paleontologists try to understand
the origin of species and evolution of eco-systems by reading ``history
of life written on stone'' in the form of fossil records. In a recent
Letter \cite{csk} we developed a dynamic network model for studying
some generic features of the biological evolution of eco-systems. In
this paper we extend that model incorporating the generic trophic-level
architecture of food webs and show how it can account for evolution at
both ecological as well as geological time scales.
\section{Earlier models and their limitations}
Because of the close similarity between the evolution of interacting
species and that of conventional systems of interacting agents
studied in statistical physics, several models of ``macro''-evolution
of eco-systems have been reported over the last decade in the physics
literature (see \cite{drossel,newman,solerev} for recent reviews).
Some of these describe macro-evolution as random walks on fitness
landscape \cite{kauffman,bak} (see also \cite{peliti,wilke} for reviews),
while some others have been formulated in terms of a matrix of
inter-species interactions \cite{solerev,sole}.
However, most of these models of ``macro''-evolution do not account for
the dynamics of populations of species even in a collective manner.
In other words, such models ignore biological details that are
certainly important on {\it ecological} time scales and, therefore,
cannot provide a natural description of origin, evolution and
extinctions in terms of population dynamics.
On the other hand, the Lotka-Volterra equation \cite{lotka} has been
used extensively in the mathematical modelling of population dynamics
of prey-predator systems. However, for the study of population
dynamics of entire ecological communities one needs a model of the
food web \cite{cohenprs}. A food web \cite{foodweb,foodweb1,foodweb2}
corresponding to an eco-system is a graphic description of prey-predator
relations. More precisely, a food web is a directed graph where each
node is labelled by a species' name and each directed link indicates
the direction of flow of nutrient (i.e., {\it from} a prey {\it to} one
of its predators). However, most often, these models assume static
food webs, where inter-species interactions are assumed to be independent
of time. But, in real eco-systems, species are known to change their
food habits with time \cite{thompson}. These changes in diets may be
caused by scarcity of the normal food and abundance of alternative food
resources. This may also arise from the adaptations of the prey species
that tend to avoid being eaten by predators through camouflage or other
mechanisms. Therefore, Lotka-Volterra type models with time-independent
food webs cannot be expected to account for ``macro''-evolution of the
eco-system over {\it geological} time scales.
Limitations of both these approches are well known \cite{mckane},
and attempts have been made to merge population dynamics and
``macro''-evolution within a single mathematical framework \cite{per}.
Population dynamics is monitored in Abramson's ``macro''-evolutionary
model \cite{abramson} in a simplified manner. However, Abramson
postulated an oversimplified model of dynamically evolving food web
that, essentially, consists of a single food chain.
Amaral and Meyer \cite{amaral} developed a ``macro''-evolutionary model
with a dynamically evolving food web where niches are arranged in a
hierarchical trophic level architecture. However, population dynamics
of the species does not enter explicitly in this model. The strength
of this model is its simplicity as some of its properties, e.g., its
self-organized criticality, can be studied analytically
\cite{drossel98,camacho00}. However, we feel, more details need to be
included to address a wider range of biologically relevant questions.
\section{The ``unified'' eco-system model}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{figure}[tb]
\begin{center}
\includegraphics[angle=-90,width=0.9\columnwidth]{foodweb0.eps}
\end{center}
\caption{A schematic representation of the network model, with {\it
random} foodweb architecture, considered in \cite{csk}. The circles
represent the niches in the eco-system. The arrows indicate the
directions of nutrient flows {\it to} the species at an arbitrary
stage during the evolution of the eco-system.}
\label{fig-0}
\end{figure}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
To our knowledge, our recent ``unified'' model \cite{csk} is one of the
first few \cite{dudek,lasz} that describes not only ``macro''-evolution
of origin/speciation and extinction of species on geological time scales
but also ``micro''- evolutionary processes like, for example, the birth,
growth (ageing) and natural death of individual organisms as well as the
effects of prey-predator interactions on their populations. Our "unified"
model, reported in \cite{csk}, can be schematically represented by the
random network shown in figure \ref{fig-0}. Each node of this
network, denoted by the circles, represents a niche that can be occupied
by at most one species at a time. In that Letter \cite{csk} we postulated
a simple random, but dynamic, food web ignoring the hierarchical
organization of species in food webs. In this paper we postulate a generic
hierarchical food web, where niches are arranged in different trophic
levels, with biologically realistic inter-species interactions.
\subsection{Architecture of the network}
As in our earlier work \cite{csk}, we model the eco-system as a dynamic
{\it network} each node of which represents a niche that can be occupied
by at most one species at a time. We assume a generic {\it hierarchical
architecture} of this network (see fig.\ref{fig-1}) in order to capture
the organization of species in different trophic levels of foodwebs
\cite{foodweb}. If the $i$-th species occupies the $\nu$-th node at the
${\ell}$-th trophic level of the food web, we denote its position by the
ordered pair ${\ell}, \nu$. We assume only one single species at the
highest level $\ell = 1$. Each node at level ${\ell}$ leads to $m$
branches at the level ${\ell} + 1$; therefore, the maximum allowed number
of nodes in level ${\ell}$ is $m^{{\ell}-1}$ and the allowed range of
${\ell}$ is $1 \leq {\ell} \leq {\ell}_{max}$. The hierarchical
architecture helps us in capturing a well known fact that in the normal
ecosystems the higher is the trophic level the fewer are the number of
species.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{figure}[tb]
\begin{center}
\includegraphics[angle=-90,width=0.9\columnwidth]{foodweb1.eps}
\end{center}
\caption{A schematic representation of the network model, with {\it
hierarchical} foodweb architecture. The circles represent the niches
in the eco-system. Each arrow represents direction of nutrient flow.
All possible nutrient flows {\it to} the species occupying the second
node at the second level and that occupying the highest level are
shown explicitly. }
\label{fig-1}
\end{figure}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\subsection{The network is dynamic}
The faster dynamics within each node captures "micro"-evolution, i.e.,
the birth, growth (ageing) and natural death of the individual organisms.
Moreover, the network itself evolves slowly over sufficiently long time
scales. For example, the adaptive evolution of the species takes place
through alterations in some of their crucial characteristics by random
mutations.
Furthermore, as the eco-system evolves with time, the populations of
some species would drop to zero, indicating their extinction, and the
corresponding nodes would be slowly re-occupied by new species through the
process of speciation.
At any arbitrary instant of time $t$ the model consists of $N(t)$
{\it species} each of which occupies one of the nodes of the dynamic
network. The total number of species cannot exceed
$N_{max} = (m^{{\ell}_{max}}-1)/(m-1)$, the total number of nodes.
Our model allows $N(t)$ to fluctuate with time over the range
${\ell} \leq N(t) \leq N_{max}$. The population (i.e., the total number
of organisms) of a given species, say, $i$, at any arbitrary instant of
time $t$ is given by $n_i(t)$. The {\it intra}-species interactions
among the organisms of the same species for limited availability of
resources, other than food, imposes an upper limit $n_{max}$ of the
allowed population of each species. Thus, the total number of organisms
$n(t)$ at time $t$ is given by $n(t) = \sum_{i=1}^{N(t)} n_i(t)$. Both
$N_{max}$ and $n_{max}$ are time-independent parameters in the model.
\subsection{Interactions in the food web}
Between any two species $i,k$ that occupy two adjacent trophic levels
there is either a link ($J_{ik}=\pm1$) or no link ($J_{ik}=0$). The
sign of $J_{ik}$ gives the direction of trophic flow, i.e. it is $+1$
if $i$ eats $k$ and it is $-1$ if $k$ eats $i$. Thus, $J_{ik} = 0$
means that there is no prey-predator relation between the two species
$i$ and $k$.
If we neglect parasites and herbivorous insects on trees, then, in
general, predators are rarer and bigger than their prey \cite{cohen03}.
This is very naturally incorporated in the hierarchical food web
structure of our model by assuming that each predator needs $m$ prey
animals to survive (see factor $m$ below). The maximum number of
individuals on each level $\ell$ is $m$ times bigger than on its
predator level $\ell-1$ in the model, and when we imagine the predator
mass to be $m$ times the prey mass, then the maximum (and initial) amount
of biomass on each level is the same. In this way, the body size and
abundance of a species are strongly correlated to the food web and its
interactions with other species \cite{cohen93,cohen03}.
The $J$ account not only for the {\it inter}-species interactions but
also {\it intra}-species interactions. Let $S_i^+$ be the number of all
prey individuals for species $i$ on the lower trophic level, and $S_i^-$
be $m$ times the number of all predator individuals on the higher
trophic level. Since a predator eats $m$ prey per time interval, $S_i^+$
gives the available food for species $i$, and $S_i^-$ is the contribution
of species $i$ to all predators on the higher level. If the available
food $S_i^+$ is less than the requirement, then some organisms of the
species $i$ will die of {\it starvation}, even if none of them is killed
by any predator. This way the model can account not only
for the inter-species prey-predator interactions but also for the
intra-species interactions arising from the competition of individual
organisms during shortage of food supply.
Note that the food resources of a given species are not restricted to
only the lower branches emanating from that node but it can also
exploit the species at the lower-level nodes emanating from other
nodes at its own trophic level. Moreover, note that although there is no
direct interaction between species at the same trophic level in our model,
they can compete, albeit indirectly, with each other for the same food
resources available in the form of prey at the next lower trophic level.
\subsection{The collective characteristics of species}
An arbitrary species $i$, occupying the $\nu$-th node at the ${\ell}$-th
level is {\it collectively} characterized by \cite{csk}:\\
(i) the {\it minimum reproduction age} $X_{rep}(i)$,\\
(ii) the {\it birth rate} $M(i)$,\\
(iii) the {\it maximum possible age} $X_{max}(i)$. \\
An individual of the species $i$ can reproduce only
after attaining the age $X_{rep}(i)$. Whenever an organism of
this species gives birth to offsprings, $M(i)$ of
these are born simultaneously. None of the individuals of this
species can live longer than $X_{max}(i)$,
even if an individual manages to escape its predators.
Note that, in several earlier works the reproductive success was modelled
mathematically by assigning a ``fitness'' to a species or to an individual
organism. The use of the term ``fitness'' has an interesting history
\cite{brookfield}. In contrast to these earlier works, in our models, we
assign a minimum reproductive age, a maximum possible age and the birth
rate to model the reproductive success (or failure). It has been felt
\cite{brookfield} that fitness merely summarizes, instead of explainig,
the ability to survive and reproduce. On the other hand, the interplay
of the $M, X_{rep}$ and $X_{max}$, we hope, will be able to explain why
some species survive while others become extinct.
\subsection{The dynamics of the eco-system}
The state of the system is updated in discrete time steps as follows:
\noindent {\it Step I- Birth}: Assuming, for the sake of simplicity, the
reproduction to be {\it asexual}, each individual organism $\alpha$
($\alpha = 1,...,n_i(t)$) of the species $i$ ($i=1,2,...N(t)$) is
allowed to give birth to $M(i;t)$ offsprings at every time step $t$
with probability (per unit time) $p_b(i,\alpha;t)$ which is non-zero
only when the individual organism's age $X(i,\alpha;t) \ge X_{rep}(i;t)$.
\noindent {\it Step II- Natural death}: At any arbitrary time step $t$ the
probability (per unit time) of ``natural'' death (due to ageing) of
an individual organism $\alpha$ of species $i$ is $p_d(i,\alpha;t)$.
\noindent {\it Step III- Mutation}: With probability $p_{mut}$ per unit time,
each of the species simultaneously increases or decreases, with equal
probability, their $X_{rep}$, $X_{max}$ and $M$ by unity. (The ages are
restricted to the interval from 1 and 100, and $M > 0$.) Moreover,
with the same probability $p_{mut}$ per unit time, they also re-adjust one of
the links $J$ from prey and one of the links $J$ to predators \cite{sole};
if the link $J$ was zero, it is assigned a new value of $\pm 1$
whereas if the link was non-zero it is assigned a new
value of zero. These re-adjustments of the incoming and outgoing
(in the sense of nutrient flow)
interactions are intended to capture the facts that each species
tries to minimize predators but look for new food resources.
\noindent {\it Step IV- Starvation death and killing by prey}:
If $n_i-S_i^+$ is larger than $S_i^-$
then food
shortage will be the dominant cause of premature death of a
fraction of the existing population of the species $i$. On the
other hand, if $S_i^- > n_i-S_i^+$, then a fraction
of the existing population will be wiped out primarily by the
predators. In order to capture these phenomena, at every time
step $t$, in addition to the natural death due to ageing, a
further reduction of the population by
\begin{equation}
C ~~\max(S_i^-,~n_i- S_i^{+})
\label{eq-kill}
\end{equation}
is implemented where $n_i(t)$ is the population of the species $i$
that survives after the natural death step above. $C$ is a constant
of proportionality. If implementation of these steps makes $n_i \leq 0$,
species $i$ becomes extinct.
\noindent {\it Step V- Speciation}: After the extinction of, typically,
half of the species in a trophic level, the niches (nodes) left empty
are re-filled by new species, with probability $p_{sp}$. All the
simultaneously re-filled nodes in a trophic level of the network
originate from {\it one common ancestor} which is picked up randomly
from among the surviving species at the same trophic level. All the
interactions $J$ of the new species are identical to those of their
common ancestor. The characteristic parameters $X_{max}$, $X_{rep}$,
$M$ of each of the new species differ randomly by $\pm 1$ from the
corresponding parameters for their ancestor.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{figure}[tb]
\begin{center}
\includegraphics[angle=-90,width=0.9\columnwidth]{foodweb2a.eps}
\includegraphics[angle=-90,width=0.9\columnwidth]{foodweb2b.eps}
\end{center}
\caption{Log-log plots of the distributions of the lifetimes of the
species in an eco-system with $n_{max} = 10^2$ to $10^4$ and 600 to
60000 iterations.
The line with slope $-2$ corresponds to a power law distribution
that has been predicted by many theories.
The common parameters for both plots are
$m=2, {\ell} = 5$ (i.e. $N_{max} = 31), p_{sp} = 0.1, p_{mut} = 0.0001,
C = 0.05$). In the upper plot, the symbols $ +, \times$ and $\ast$ correspond to
$n_{max} = 10^2, 10^3, 10^4$ averaged over 6400, 640 and 64 systems
respectively. In the lower plot, $n_{max} = 1000$ (except for the line where
$n_{max} = 100$) and the maximum simulation
time is 600 (+) and 60000 ($\times$ and line) iterations; $\ast$
corresponds to $m=12, \ell = 3$ after 6000 iterations; 640 systems were
averaged over for short and intermediate times, and 64 for the longest time.
Each system started from a new random initial state.}
\label{fig-2}
\end{figure}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\subsection{Probability of birth}
We assume the {\it time-dependent} probability $p_b(i,\alpha)$ (of
individual $\alpha$ in species $i$) of giving birth per unit time to
decrease linearly with age, from its maximum value, attainable at the
minimum reproduction age, down to zero at the maximum lifespan. It is
multiplied with a Verhulst factor $ 1 - n_i/n_{max}$ and equals this
factor at $X = X_{rep}$. Thus in the limit of vanishingly small
population, i.e., $n_i \rightarrow 0$, we have $p_b(i,\alpha) \rightarrow 1$
if $X(i,\alpha) = X_{rep}(i)$ and, thereafter, $p_b$ decreases linearly
\cite{austad} as the organism grows older. However, since the eco-system
can support only a maximum of $n_{max}$ individual organisms of each
species, $p_b(i,\alpha;t) \rightarrow 0$ as $n_i(t) \rightarrow n_{max}$,
irrespective of the age of the individual organism $\alpha$ \cite{cebrat}.
\subsection{Probability of natural death}
Similarly, we assume the probability $p_d$ of ``natural'' death (due to
ageing) to increase linearly with age \cite{carey} and to reach unity at
the maximum lifespan $X_{max}$ of the species:
$p_d = (X M - X_{rep})/(X_{max}M -X_{rep})$.
(For $X < X_{rep}$ the death probability, instead, has the constant
value that $p_d$ attains at $X = X_{rep}$; if the above denominator is
negative, $p_d=1$.) Note that, for a given $X_{max}$ and $X_{rep}$,
the larger is the $M$ the higher is the $p_d$ for any age $X$.
Therefore, each species has a tendency to increase $M$ for giving
birth to larger number of offsprings whereas the higher mortality
for higher $M$ opposes this tendency \cite{tradeoff}.
\section{Results}
In our simulations. initially, $M = 10, \; X_{max}$ is distributed
randomly between $2$ and $99$ independently for each species, $X_{rep}$
randomly between $1$ and $X_{max}$, the population randomly between $1$
and $n_{max}/2$. The ages of the individuals vary randomly between $1$
and the $X_{max}$ of their species.
The longest runs in our computer simulations were continued upto a
a million time steps. If each time step in our
model is assumed to correspond to a real time of the order of one
year, then the time scale of a million years, over which we have
monitored our model eco-system, is comparable to real speciation time
scales.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{figure}[tb]
\begin{center}
\includegraphics[angle=-90,width=0.9\columnwidth]{foodweb3.eps}
\end{center}
\caption{Semi-log plot of the distributions of $X_{rep}$ ~($\times$) and $X_{max}$, taken from the simulations symbolized by the curved line in the lower
part of fig.\ref{fig-1}: $m=2, \ell=5, n_{max}=100, t=60000$, 640 systems.
}
\label{fig-3}
\end{figure}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\subsection{Lifetime distributions}
The average distributions of the lifetimes of the species are plotted
in fig.\ref{fig-2} for various sets of values of the parameters. Only very
approximately,
the data are consistent with a power-law; the effective exponent, which
is, ~~approximately, $2$, is also consistent with the corresponding estimate
quoted in the literature \cite{drossel,newman}. However, in
fig.\ref{fig-2} the power law holds only over a limited range \cite{chu} of
times; for longer times a plateau seems to develop.
Since real eco-systems are much more complex than our model eco-system
and the available fossil data are quite sparse, it is questionable
whether real extinctions follow power laws and, if so, over how many
orders of magnitude.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{figure}[tb]
\begin{center}
\includegraphics[angle=-90,width=0.9\columnwidth]{foodweb4.eps}
\end{center}
\caption{Semi-log plot of the distribution of $M$. The parameter
values are same as those in fig.\ref{fig-3}; shorter and longer simulations
are added to show further broadening of the distribution.
The symbols $+$, $\times$ and $\ast$ correspond to $600$, $60000$ and $600000$
iterations, respectively, using 6400, 640 and 1 systems. The lower lines, using
64 lattices with $n_{max}=100, \; t=6000$, show the broadening with increasing
mutation rate $p_{mut} = 0.00001, \; 0.001$ and 0.01.
}
\label{fig-4}
\end{figure}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{figure}[tb]
\begin{center}
\includegraphics[angle=-90,width=0.9\columnwidth]{foodweb5.eps}
\end{center}
\caption{Log-log plot of the distribution of lifetimes for speciation
probabilities $p_{sp}$ = 0.02 (+) and 0.5 ($\times$), and (squares, with
$p_{sp}$ = 0.1) for
Gompertz mortality assumption: $p_d = \exp[(\max(X,X_{rep})-X_{max})/M]$,
using 640 systems for $n_{max}=100$ and $t=6000$.}
\label{fig-5}
\end{figure}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{figure}[tb]
\begin{center}
\includegraphics[angle=-90,width=0.9\columnwidth]{foodweb6.eps}
\end{center}
\caption{Log-log plot of the distribution of lifetimes for the whole
ecosystem of $\ell$ trophic layers, with $\ell=5,m=2$ and $\ell=3,m=6$,
from 1 and 10 systems only; $n_{max}=100$.
}
\label{fig-6}
\end{figure}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\subsection{Distributions of Species Characteristics}
Figs.\ref{fig-3},\ref{fig-4} show the time-averaged distributions of
$X_{max}$, $X_{rep}$ and $M$. We see that the minimum age of reproduction
$X_{rep}$ is quite small, as usual in a similar ageing model
\cite{stauffer}. The age distribution (not shown) decays stronger than a simple
exponential, indicating a mortality increasing with age as it should
be \cite{carey}. The genetic death ages $5 < X_{max} < 100$ reach ages far
above the upper end $\simeq 50$ of the age distribution (for the species on top
of the food web), as is
appropriate for animals in the wild \cite{austad}. Finally,
fig.\ref{fig-4} shows the distribution of $M(i)$ which is still broadening
even after 60000 iterations.
We have also observed (not shown)
that the higher is the mutation probability $p_{mut}$ the lower
is the lifetime of the eco-system;
this is consistent with the intuitive expectation that
a higher rate of mutation leads to higher levels of biological activity
in the eco-system thereby leading to the extinction of larger number of species.
% new remark added below; also figure capotion 4 corrected and expanded
Fig.\ref{fig-4} from these data shows that the broadening of the histogram
for $M$, i.e. the equilibration process, is determined by the product $p_{mut}t$
giving the average number of mutations per species.
But, $p_{sp}$ had weaker effect on the same data as shown in
Fig. \ref{fig-5}. The same figure also shows a somewhat better power law
at short times if the above linear increase of the mortality with age
is replaced by an exponential increase (Gompertz law \cite{austad}).
\subsection{Collapse of fragile ecosystems}
We model an eco-system with a {\it fixed} number $\ell$ of trophic levels;
thus as soon as we find one level to be extinct completely, we regard the
eco-system as destroyed and try to build a new one for the same parameters,
changing only the random numbers. Hundreds of such attempts are needed for
a successful system lasting the prescribed number (like 6000) of iterations,
see fig.\ref{fig-6}.
This method simulates the billions of years which natural evolution
needed to build the present life on earth.
\section{Summary and conclusion}
In summary, we have presented a unified model which describes not only
the birth, ageing and death of individuals as well as population dynamics
on short time scales but also the long-time evolution of species, their
origination/speciation and extinction. The total number of species, the
inter-species interactions and the collective characteristics, namely,
$X_{rep}, X_{max}$ and $M$, of each species vary following a stochastic
dynamics with Darwinian selection. Thus, our model is capable of
{\it self-organization}.
\noindent {\bf Acknowledgements}
We thank J.E. Cohen for emphasizing to us the importance of food webs and the
Supercomputer Center J\"ulich for computer time on their CRAY-T3E.
This work is supported by Deutsche Forschungsgemeinschaft through a
Indo-German joint research project.
\bibliographystyle{plain}
\begin{thebibliography}{1}
\bibitem{csk} D. Chowdhury, D. Stauffer and A. Kunwar, Phys. Rev. Lett.
{\bf 90}, 068101 (2003).
\bibitem{drossel} B. Drossel, Adv. Phys. {\bf 50}, 209 (2001).
\bibitem{newman} M. E. J. Newman and R. G. Palmer, adap-org/9908002;
{\it Modeling Extinction}, (Oxford University Press, New York, 2002);
D. M. Raup, {\it Bad Genes or Bad Luck}, (W. W. Norton, New York, 1991) p. 81.
\bibitem{solerev} R. V. Sole, in: {\it Statistical Mechanics of Biocomplexity}, eds. D. Reguera, J. M. G. Vilar and J. M. Rubi, Lec. Notes in Phys. vol.527, 217 (Springer, 1999)
\bibitem{kauffman} S. Kauffman, {\it The Origins of Order: Self-organization
and selection in Evolution} (Oxford University Press, New York, 1993).
\bibitem{bak} P. Bak and K. Sneppen, Phys. Rev. Lett. {\bf 71}, 4083 (1993);
M. Paczuski, S. Maslov and P. Bak, Phys. Rev. E {\bf 53}, 414 (1996).
\bibitem{peliti} L. Peliti, cond-mat/9505003; 9712027.
\bibitem{wilke} C. O. Wilke, C. Ronnewinkel and T. Martinez, Phys. Rep.
{\bf 349}, 395 (2001).
\bibitem{sole} R. V. Sol\'e and S. C. Manrubia, Phys. Rev. E
{\bf 54}, R42 (1996); {\bf 55}, 4500 (1997).
\bibitem{lotka} J. Hofbauer and K. Sigmund, {\it The Theory of Evolution and Dynamical Systems} (Cambridge University Press, 1988).
\bibitem{cohenprs} J. E. Cohen, T. Luczak, C. M. Newman and Z. -M Zhou, Proc. Roy. Soc. Lond. B {\bf 240}, 607 (1990).
\bibitem{foodweb} G. A. Polis and K. O. Winemiller (eds.) {\it Food
Webs: Integration of Patterns and Dynamics} (Chapman and Hall, New York, 1996).
\bibitem{foodweb1} S. L. Pimm, {\it Food Webs} (Chapman and Hall, London, 1982).
\bibitem{foodweb2} Drossel B and McKane, A.J., Modelling food webs, p.
218 in: {\it Handbook of Graphs and Networks - From the Genome to the
Internet}, S. Bornholdt and H. G. Schuster (eds.), (Wiley-VCH, Weinheim, 2003).
\bibitem{thompson} J. N. Thompson, Science {\bf 284}, 2116 (1999).
\bibitem{mckane} C. Quince, P. G. Higgs, and A. J. McKane,
in: {\it Biological Evolution and Statistical Physics}, M. L\"assig and
A. Valleriani (eds), (Springer Verlag, Berlin-Heidelberg 2002).
\bibitem{per} P. A. Rikvold and R. K. P. Zia, nlin.AO/0303010.
\bibitem{abramson} G. Abramson, Phys. Rev. E {\bf 55}, 785 (1997).
\bibitem{amaral} L. A. N. Amaral and M. Meyer, Phys. Rev. Lett. {\bf 82}, 652 (1999).
\bibitem{drossel98} B. Drossel, Phys. Rev. Lett. {\bf 81}, 5011 (1998).
\bibitem{camacho00} J. Camacho and R. V. Sole, Phys. Rev. E {\bf 62},
1119 (2000).
\bibitem{dudek} A. Nowicka, A. Duda and M. R. Dudek, cond-mat/0207198.
\bibitem{lasz} A. Laszkiewicz, Sz. Szymczak and S. Cebrat, Int. J. Mod. Phys. C {\bf 14}, No.6 (2003).
\bibitem{cohen03} J. E. Cohen, T. Jonsson and S. R. Carpenter, Proc.
Natl. Acad. Sci. USA {\bf 100}, 1781 (2003).
\bibitem{cohen93} J. E. Cohen, S. L. Pimm, P. Yodzis and J. Saldana, J.
Animal Ecology {\bf 62}, 67 (1993).
\bibitem{brookfield} J. F. Y. Brookfield, Nature {\bf 411}, 999 (2001).
\bibitem{austad} K. W. Wachter, and C. E. Finch, eds.
{\it Between Zeus and the Salmon: The Biodemography of Longevity},
(National Academy Press, Washington DC, 1997), in particular p. 162 by
Austad, S.N.
\bibitem{cebrat} J. S. S\'a Martins, and S. Cebrat, Theory in Biosciences
{\bf 199}, 156 (2000).
\bibitem{carey} J. R. Carey, Exp. Gerontology {\bf 37}, 567 (2002).
\bibitem{tradeoff} C. K. Ghalambor, and T. E. Martin, Science {\bf 292},
494 (2001).
\bibitem{chu} C. Adami and J. H. Chu, Phys. Rev. E {\bf 66}, 011907 (2002).
\bibitem{stauffer} D. Stauffer and J. P. Radomski, Exp. Gerontology
{\bf 37}, 175 (2001).
\end{thebibliography}
\end{document}