296 Tad C. Theimer and Catherine A. Gehring
implications that difference may have for how
we view the effects of vertebrates on commu-
nity attributes. We begin with a brief overview
of how Australian tropical rainforests may differ
from those in other parts of the world, followed
by a summary of our results on the effect of
vertebrate exclusion on seedling and AMF spore
abundance and diversity. We then propose a sim-
ple conceptual model that argues AMF spores and
seedlings differed in their response to vertebrates
because vertebrates acted primarily to promote
colonization of experimental plots by mycorrhizal
fungal spores while they acted primarily as agents
of local extinction for seedlings without promot-
ing seedling colonization. We end by reiterating
that the effects of a terrestrial vertebrate com-
munity on a seedling community depend on the
relative importance in that plant community of
(1) vertebrate seed dispersal, (2) vertebrate mycor-
rhizal fungal spore dispersal, and (3) the strength
of vertebrate-induced density-dependent seed and
seedling mortality. We suggest that the interplay
of these factors will determine the net effect on
community structure, and that this effect may
differ from that observed or predicted based on
studies of a subset of species, if that subset is not
representative of the community as a whole.
COMPARATIVE ATTRIBUTES OF
AUSTRALIAN RAINFORESTS
The native vertebrate fauna of Australian rain-
forests is depauperate compared with that of other
tropical rainforests (Eisenberg 1981). Especially
lacking are large (>10 kg) herbivorous mammals
like deer, peccaries, and tapirs, though many areas
are increasingly impacted by feral pigs (Susscrofa).
The largest herbivorous terrestrial mammal is the
red-legged pademelon, a small (3.5–6.8 kg) kan-
garoo. The larger tree kangaroos and diverse pos-
sums are arboreal folivores that rarely feed on the
ground. The largest frugivore (1.5–2 m tall) is the
southern cassowary, a flightless bird, with no par-
allel in other tropical faunas except New Guinea.
Severalspeciesof rodentsactasseedpredatorsand
potential seed dispersers (Harringtonet al. 1997,
Theimer 2001), while a small, frugivorous kan-
garoo roughly approximates neotropical agoutis
and acouchies in size and behavior (Dennis 2003).
As in many other forests, these terrestrial ver-
tebrates operate at differing spatial scales, with
cassowaries potentially moving seeds and spores
over hundreds of meters (Westcottet al.2006),
while rodents and smaller marsupials move most
propaguleslessthan50m(Harringtonetal.1997,
Theimer 2001, Dennis 2003). Given the rela-
tively depauperate fauna of Australian rainforests,
the effects of terrestrial vertebrates on seedling
dynamics may be relatively weak compared with
other, more species-rich forests. The mycorrhizal
fungal communities of Australian rainforests also
have similarities and differences when compared
with other rainforests. Arbuscular mycorrhizae
dominate (e.g., Hopkins et al. 1996, Gehring
2003, Gehring and Connell 2006), as is typical
of many rainforests (e.g., St John 1980, Bereau
and Garbaye 1994). However, seedlings of more
than one third of the dominant species on our
study plot are rarely or never colonized by AMF
(Gehring and Connell 2006). This high propor-
tion of non-mycorrhizal species contrasts with
data from many rainforests (St John 1980, Bereau
and Garbaye 1994), but is similar to that in trop-
ical forests in Brazil and China (Zangaroet al.
2000, Zhaoet al. 2001). AMF spore abundance
and diversity are comparable to other rainforests,
with an average of 54 spores per gram of soil
and dominance of the community by members of
the generaAcaulosporaandGlomus(Gehringet al.
2002, Lovelocket al. 2003, Manganet al. 2004).
Several species of terrestrial vertebrates common
in Australian rainforests carry AMF spores in their
feces, including the native rodents, the musky rat
kangaroo, two species of bandicoots, and several
species of ground-frequenting birds (Reddellet al.
1997).RESULTS OF TERRESTRIAL
VERTEBRATE EXCLUSION FROM
AN AUSTRALIAN RAINFOREST
In autumn 1996, we erected fourteen 6.5 m×
7.0 m vertebrate exclosures in the area sur-
rounding Connellet al.’s (1984) long-term study
plot. For the following 4.5 years, we monitored
seedling survival and recruitment on exclosures