300 Tad C. Theimer and Catherine A. Gehring
open plots would be greater than that on exclo-
sures, and the difference in rates would be propor-
tional to the difference in mortality rates between
open and protected plots. If seedling mortality
due to vertebrates was more strongly density
dependent than that experienced by seedlings in
their absence, then terrestrial vertebrates could
potentially decrease the rate of species extinc-
tion on open plots relative to protected plots by
allowing rare species to survive by differentially
removing more abundant, superior competitors
(Figure 17.3). The net change in tree seedling
or AMF spore richness over time would depend
on the difference in colonization and extinction
rates.
We argue that this simple model offers an
explanation for the differing responses of AMF
spores and tree seedlings to terrestrial vertebrate
exclosure and has important implications for
how terrestrial vertebrates could alter community
dynamics. In the case of tree seedlings on our
site, we found little evidence that terrestrial ver-
tebrates increased the local species pool available
to colonize plots. The total number of species
recorded on exclosure plots was actually higher
than that on open plots. Although most seedling
species were found in similar relative abundances
on the two plots (Figure 17.1), there was a small
suite of species that were found on exclosure
plots and never recorded on open plots and vice
versa. However, the relative rarity, as well as
the fruit and seed characteristics, of these non-
overlapping species suggested their absence was
most likely due to sampling error or sensitivity
to seed or seedling mortality in the presence of
vertebrates. Only eight species (total of 19 indi-
viduals out of roughly 4000 seedlings recorded)
occurred only on open plots; one species produced
papery, winged seeds that we considered primar-
ily wind dispersed (Cardwellia sublimis), six had
seeds we considered primarily dispersed by arbo-
real birds, and only one (Castanospora alphandi)
produced fruit and seeds that would potentially be
moved primarily by terrestrial vertebrates. In con-
trast, 12 species (47 individuals) occurred only on
exclosure plots, and again included species with a
variety of putative seed dispersal modes, including
wind,arborealbird,andterrestrialvertebrates.We
interpret these results as an indication that there
was no significant increase in the species pool
available to colonize open plots due to increased
seed dispersal by terrestrial vertebrates.
We hypothesize that this lack of effect was
due to characteristics of the vertebrate com-
munity and the tree community at our site.
First, two important terrestrial vertebrate seed
dispersers, southern cassowaries and musky rat
kangaroos, were relatively uncommon on our
plots. We recorded only five cassowary drop-
pings on the 5 ha we frequently traveled over
5 years of study, and none of these fell on a
study plot (interestingly, the only one that yielded
seedlings produced a clump of the most common
canopy tree on the site,Chrysophyllumsp. nov.).
White-tailed rats were common on our plot, but
seedlings of one species for which white-tailed
rats have been argued to be important seed dis-
persers,Beilschmiedia bancroftii(Harringtonet al.
1997, Theimer 2001), were never recorded on
an open plot but did occur in three exclosure
plots,mostlikelybecauseseedpredationandcache
recovery by white-tailed rats resulted in extremely
low densities of surviving seedlings outside of
exclosures. Perhaps more importantly, the major-
ity of tree species on our plot did not depend
on these animals as dispersal agents, and were
instead effectively dispersed by wind (e.g., many
Proteaceae) or canopy birds and bats (e.g., many
Lauraceae).
In contrast, both the number of terrestrial
vertebrate species that contributed to seed and
seedling mortality, and therefore local extinction
rates, and the number of plant species with seeds
and seedlings susceptible to terrestrial vertebrate
mortality were relatively large.The two most com-
mon rodents, the bush rat andMelomys, have not
been reported to disperse seeds, but both have
been documented to eat a variety of rainforest
tree seeds offered to them in captive feeding tri-
als (Harringtonet al.1997). In addition, several
relatively common terrestrial vertebrates on our
study plot (e.g., birds like chowchillas and brush
turkeys, and the marsupial bandicoots) acted as
important agents of seedling mortality, primarily
by uprooting and burying seedlings while foraging
in leaf litter, for some seedling species accounting
for 70–90% of seedling cohort mortality (Theimer
and Gehring 1999).