Tree species composition varies dramatically from place to place in the
Amazon rain forest. Prior analyses almost invariably assumed that
habitat preferences, or niche specialisation, would explain much of this
variation. However, recent PhD work of Pitman showed, to the contrary,
that most, if not all, species in their Peruvian and Ecuadorian plots
were habitat generalists in stead of habitat specialists. Furthermore, a
number of species was found to dominate tree communities in plots both
in Peru and in Ecuador at ca. 1400km distance. In fact, Hubbell’s’
neutral model, that assumes zero habitat specialisation for TRF trees
and is based on explaining beta-diversity not
in terms of habitat specificity but dispersal limitation, predicts just
this. Condit indeed showed that such a neutral model could explain, very
significantly, the differences in composition by distance.
The Guianas beg to disagree with a neutral community structure, and
differ from Western Amazonia in having a lower alpha diversity,
relatively high endemism and habitat speciality. Largest differences in
species composition are found when podzols are compared with ultisols/oxisols.
Still, for any given distance, forests on white sands show more
similarity with each other than comparisons between forests on different
soil types, where the distance effect proves to be of overriding
importance.
A central question to this PhD-work is how the local species pool is
derived from the regional species pool. Is it mainly through a random
process and influenced mainly by dispersal limitation or are
environmental filters important as well? To achieve maximum resolution
as to environmental filters we use plots of the two soil types that show
greatest differentiation in species composition: white sands and brown
sands.
In the Guianas new combinations of plots need to be established in such
a way that a variety of distances between plots is achieved, while
having at least plots on podzols and oxi-/ultisols (and some leptosols)
close by. Life history characteristics of species will be compared with
local abundance to look for patterns and functional groups in relation
to habitat preferences. To assess whether species dominant on the
contrasting soil types asses different nutrient levels in the soils we
propose a ‘natural phytometer’ experiment. A comparison with patterns in
beta-diversity in Western Amazonia and the
related patterns in life-history characteristics in close collaboration
with Pitman will complete this project.
[Home] [Staff] [Research
Projects] [Publications] [Meetings]
[Links]
[Nationaal
Herbarium Nederland]
