Abstract:
For my doctoral thesis, I have used a combination of experiments and simulations to
investigate the various factors that affect the dynamics and stability of spatiallystructured
as well as spatially unstructured populations. A brief description of my work is
as follows:
Metapopulation stability
Although classical population ecology theory treats individuals as being homogeneously
distributed over space, most natural populations exhibit some degree of spatial structuring
into metapopulations: ensembles of local populations (henceforth, subpopulations) that
are connected by migration. Using Ricker-based coupled map lattice simulations, I show
that the precise spatial arrangement of the subpopulations does not interact with
migration in determining metapopulation stability. This indicates that the fine-scale
details of the spatial arrangement of subpopulations can often be safely ignored while
modeling metapopulation dynamics. In a continuation of this work, I show that, at least
for systems in which the subpopulations follow Ricker dynamics, maximum
metapopulation stability is attained at intermediate migration rates, regardless of whether
the migration rate is density-dependent, density-independent or stochastic. However,
migration rate can stabilize the dynamics of a metapopulation only when the migration
events take place very frequently. These results were found to be robust to different
spatial arrangements of patches.
4
The above studies indicated that a metapopulation would be most stable at intermediate
rates of migration - a prediction that I tested using laboratory metapopulations of
Drosophila melanogaster. I show that a low migration rate (10%) stabilizes D.
melanogaster metapopulations by inducing asynchrony between neighboring
subpopulations. On the other hand, higher migration rate (30%) synchronizes the
neighboring subpopulations, thus leading to metapopulation instability. Simulations
based on a simple non-species specific population growth model (Ricker map) captured
most features of the data, suggesting that the results are generalizable. A subsequent
simulation study indicated that, contrary to the concern raised by some other workers,
asynchrony at intermediate migration rates is a very likely outcome in real
metapopulations.
I have also empirically investigated the effects of constant localized perturbations on the
stability of metapopulations. The experimental data suggests that constant addition of
individuals to a particular subpopulation in every generation stabilizes that population
locally, but does not have an effect on the dynamics of the metapopulation in any way.
Simulations of the experimental system, based on the Ricker map, were able to recover
the empirical findings, indicating the generality of the results. I also simulated the
possible consequences of perturbing more subpopulations, increasing the strength of
perturbations and different rates of migration, but found that none of these conditions
were expected to alter the outcomes of our experiments. Finally, I show that the main
results of this study are robust to the presence of local extinctions in the metapopulation.
5
Stability of spatially unstructured populations
Prior studies have indicated that the dynamics of D. melanogaster single populations are
affected by three major density-dependent feedback loops: larval density acting on 1)
larval survivorship and 2) adult fecundity, and 3) the effects of adult density on adult
fecundity. In an experimental study on replicate D. melanogaster single populations, I
altered the relative strengths of these loops by manipulating the quantity and quality of
nutrition available to the larvae and the adults. This study led to several insights into how
the three density-dependent loops interact to shape the dynamics of D. melanogaster
populations in the laboratory.
In an experimental study, I examined the effects of four different rates of adult mortality
(control, 20%, 40% and 60%) on the stability of replicate D. melanogaster single
populations under two different nutritional regimes. When the intrinsic growth rate was
low, there was no significant effect of different mortality rates on stability. However,
under high rates of intrinsic growth, the effects of mortality rates on stability varied based
on the index chosen to quantify stability. Specifically, under high growth rates, the
variation in population size (as measured by coefficient of variation, CV) across
generations, decreased monotonically with increasing rates of mortality. However, the
average one-step fluctuation in population size (as measured by fluctuation index, FI)
was significantly larger at lower mortality rate (20%). The extinction probabilities of the
low mortality treatment were also found to be different from the controls.
6
I also investigated the issue of evolution of population stability as a result of selection
acting on the life history of organisms. Although there were several hypotheses about the
mechanism of evolution of population stability, none of them had any empirical support.
A previous study had provided the first experimental demonstration that population
stability can evolve as a correlated (and not direct) response to selection on life-history
traits. In a subsequent study, which extends the above work, I show that the evolution of
one type of stability property (constancy) does not necessarily guarantee that other
stability properties would also evolve simultaneously. Moreover, manifestation of
stability properties was found to depend critically on the fine details of the environment
under which the populations are maintained.
Finally, in another experimental study, I demonstrate that minor variations in pre-assay
rearing conditions can lead to systematic bias in life-history traits like fecundity. This
underlines the importance of an often-neglected source of stochastic variations that can
potentially affect the dynamics of populations, even under controlled laboratory
conditions.