Introduction:
Patterns in the composition and diversity of species in a community are
the result of many interacting processes. Borrowing concepts from
population genetics, Vellend (2010, 2016) distilled these down to four
fundamental processes in a conceptual synthesis of community ecology:
selection, ecological drift, speciation, and dispersal. Since speciation
and dispersal are responsible for the introduction of new variation into
the community, the dynamics of a closed community is essentially
governed by selection (also referred to as species sorting) and
ecological drift alone. Species sorting is natural selection at the
level of species, which will produce distinct assemblages of species in
different habitats, each local community consisting of those species
best adapted to their local conditions of growth. This classical
“niche-based” view asserts that species coexistence is due to
functional differences between species and predicts deterministic
dynamics. Ecological drift is genetic drift at the level of species,
which will produce a distinctive assemblage of species in any given
place whose composition is unrelated to local conditions. This
“neutral-based” view assumes the functional equivalency of species and
predicts stochastic community dynamics. The relative importance of these
two processes in structuring communities has been vigorously debated in
the last two decades and many attempts have been made to show that one
of these processes is much more important than the other (Wright 2002,
Hubbell 2006, Rosindell et al. 2011, Wennekes et al. 2012). It would be
difficult to support either extreme view, however, because both
processes will be active at all times everywhere, and the main goal of
community ecology should be to understand how the balance between them
depends on the underlying physical and biotic characteristics of sites.
Many recent developments have been proposed to resolve the niche-neutral
controversy (Leibold and McPeek 2006, Adler et al. 2007, Haegeman and
Loreau 2011, Chase 2014, Fisher and Mehta 2014, Matthews and Whittaker
2014, Shoemaker et al. 2020, Siqueira et al. 2020), and there is a
growing body of empirical experimental work aimed at disentangling
stochastic from deterministic processes in community assembly (Chase
2010, Gilbert and Levine 2017, Ron et al. 2018).
Species sorting and ecological drift will have directly opposed effects
on the species composition of communities under a given set of
environmental conditions. Under species sorting, communities that
initially differ in composition will converge on the same composition,
which represents the stable equilibrium community for that set of
conditions. Under ecological drift, communities that are initially
identical in composition will diverge over time. The relative
contributions of these two processes to community dynamics (changes over
time in species composition) can therefore be estimated by setting up
replicated communities with different initial composition. Ecological
drift will cause divergence of replicate communities of any given
initial composition. Species sorting will cause convergence of
communities that initially differ in composition.
A third factor which might influence how a community changes over time
is its initial state, as both drift and sorting may be historically
contingent (Chase 2003, Fukami 2015). A species’ initial frequency may
influence its sensitivity to ecological drift due to the tendency of
stochasticity to increase in importance in smaller effective
populations. Likewise, species sorting might depend on a mechanism that
favoured abundant species, such as priority effects or growth inhibition
by exudates. The contributions of sorting, drift and initial state will
sum to the overall change in composition observed over a given period of
time.
Such experiments have been done for single-species populations to
estimate the contributions of natural selection, genetic drift and
ancestry to the evolution of fitness and of phenotypes such as cell size
in bacteria and the evolution of heterotrophy in Chlamydomonas(Travisano et al. 1995, Bell 2013). Despite the clear analogy of these
processes in population genetics to community ecology (Vellend 2010), no
similar work has yet been done in multi-species communities. Here we
extend experimental evolution into ecology to estimate the relative
contributions of species sorting (the ecological equivalent of natural
selection), ecological drift (genetic drift), and initial state
(ancestry) to community species dynamics.
We assembled experimental communities of floating aquatic macrophytes
from the family Lemnaceae that frequently coexist in the field.
These are highly reduced angiosperms that consist of a single leaf-like
frond which may or may not bear a submerged unbranched root, depending
on the species. Reproduction is nearly always asexual and vegetative,
which results in extremely short generation times of less than a week in
eutrophic conditions. Many species are widespread and abundant in lentic
ecosystems and often coexist in multi-species communities consisting of
hundreds of thousands to millions of individuals. Because of their small
size and short generation time, they are being increasingly used as a
model system in ecology and evolution (Laird and Barks 2018, Hart et al.
2019, Vu et al. 2019) and enable us to run highly-replicated experiments
lasting more than a dozen generations in a single season. Here we report
the results of a basic community dynamics experiment using semi-natural
communities consisting of four such species of Lemnaceae . By
manipulating initial relative abundances of species and following
changes in species composition over time, we can estimate the relative
contributions of these opposing processes to community species dynamics.