Endophytic Fungi 263based on only the most common – and perhaps
most generalist – species (Arnold and Lutzoni
2007). It is not surprisin gthat evidence for host
affinity and spatial structure amon gendophytes
has been conflicting, and at times contradictory.
Arnoldet al.(2000, 2003) found stron gevi-
dence for both host affinity and spatial structure
of endophyte communities within and among
forests in Panama, but Cannon and Simmons
(2002) found little structure to endophyte com-
munities in a non-quantitative study in Guyana.
Suryanarayananet al.(2002) found a high degree
of overlap amon gcommunities of endophytes in
many different hosts and forest types in India, but
Arnold and Lutzoni (2007) found little overlap
in endophyte genotypes among different forests
alon ga latitudinal gradient. In the latter study,
some endophyte genotypes were isolated from
numerous hosts at BCI, regardless of the phy-
logenetic placement, leaf defenses, or phenology
of those hosts; however, others were found in
only a single host species. Tropical endophyte
communities appear to contain a mix of gen-
eralist and site- and host-specific species. The
challenge remains to infer these ecological param-
eters for very rare taxa, and to determine whether
apparently rare species are truly rare or simply
compete poorly given a set of culturing conditions
(see Box 15.1).
Distribution data alone do not provide insight
into the mechanisms that underlie apparent host
affinity. Because leaves of tropical trees tend
to be well defended against pathogens (Coley
and Barone 1996), it is plausible that chemi-
cal defenses of leaves may influence host affinity
of endophytes. To test this hypothesis, Arnold
and Herre (2003) incorporated leaf homogenates
as the nutrient source into water agar and
assessed growth rates of endophytesin vitro.
Eighty-six percent of endophytes fromTheobroma
cacao, when tested on media containin gextracts
from each of three host species, grew faster on
extracts fromT. cacaothan on extracts from two
co-occurrin gtree species (Arnoldet al.2003). To
ensure that this result did not reflect a greater
nutritive value inT. cacaoextracts, Arnoldet al.
(2003) grew endophytes from three host species
on leaf-extract media from all hosts, and found
that endophytes grew more rapidly on extracts
of the host in which they were most frequently
recovered in the field. With new methods for
raisin gsterile seedlin gs and inoculatin gthem
with endophytes now available, the stage is set
for much-needed assessments of host affinity
in planta.
Unfortunately, assessin gspatial structure of
endophyte communities – critical for deter-
minin gbeta-diversity (Chave Chapter 2, this
volume) – does not lend itself to such straightfor-
ward laboratory experiments, although reciprocal
transplant experiments could be useful in this
regard. Fungal spores have long been thought
to spread lon gdistances as aerial plankton; clas-
sic examples include wheat leaf rust (Puccinia
triticina), which overwinters in Mexico and blows
north over the Great Plains of the USA in mid-
sprin g(A grios 1997). Whether endophytes move
about at a similar scale is not known. Genotype
data are especially important for comparin gfun-
gal assemblages in different sites – as is recognized
for many pathogenic fungi (Agrios 1997).FUNGA LENDOPHYTES AND
TROPICA LFOREST COMMUNITY
ECOLOGY
Given that plants in the dark forest understory
are carbon limited, why do they host such a large
number of obligate heterotrophs in their leaves?
Given that endophytes are often closely related to
pathogens, and that plants in tropical forests are
well defended against pathogenic fungi (relative to
temperate species; Coley and Barone 1996), why
do plants host such a diversity of fungal species in
their tissues?
These questions have yet to be answered, as
studies assessin gthe ecolo gical roles of tropi-
cal endophytes are yet in their infancy. In gen-
eral, however, there are three main hypotheses
regarding the roles of endophytic fungi: that
they are (1) neutral inhabitants, (2) parasites, or
(3) mutualists of their hosts. Given the tremen-
dous phylogenetic diversity of tropical endophytes
(Figure 15.1), endophytes as a whole likely include
species with the capacity to play each of these
roles, or to change roles over time or under cer-
tain conditions (see Box 15.4). Moreover, it is