264 A. Elizabeth Arnold
plausible that the most important ecological roles
of endophytes are manifested with regard not to
the plants they inhabit, but instead to insects or
pathogens that attack those plants. Here, I present
evidence for and against these general models of
endophyte–host interactions as a foundation for
emergent questions regarding the ecological roles
of tropical endophytes.
ENDOPHYTES AS NEUTRAL
INHABITANTS OF THEIR HOSTS
Over the past two decades, some authors have sug-
gested that endophytes simply inhabit their hosts
without interactin gdirectly (e. g., Carroll 1988).
Under this scenario, endophytism is an incidental
part of the life cycle of fungi whose primary eco-
logical role lies elsewhere. However, endophytes
selectively colonize particular hosts (Arnold and
Herre 2003), implyin gan interaction between
endophytes and host defenses and/or other traits
(see also Arnoldet al.2003). Further, endophytes
actively penetrate leaf cuticles durin gcoloniza-
tion, and only rarely enter leaves in a passive man-
ner (i.e., through open stomata; Mejiaet al.2003,
Herreet al.2005b). Endophytes remain metabol-
ically active durin gthe intercellular colonization
phase, and grow slowly but actively within host
foliage following infection (Deckertet al.2001,
Arnoldet al.2003). In each of these stages, fungi
exude the organic molecules needed for cuticular
penetration and absorptive nutrition (Van Schöll
et al.2006). Given the close phylogenetic relation-
ship between endophytes and pathogens, it is likely
that tropical plants are sensitive to such exudates.
Can endophytes avoid inducin ghost defenses dur-
in gcolonization? If so, then how do plants in
the forest understory tolerate carbon use by these
heterotrophic colonists?
ENDOPHYTES AS PARASITES
Based in part on the observation that endo-
phytes subsist on carbon from the host (Clay
2001), the potential role of endophytes as plant
parasites has lon gbeen reco gnized. The evo-
lutionary transiency between endophytism and
pathogenicity also underscores the possibly neg-
ative roles of endophytes (Arnold 2007). Yet
leaf area, plant growth rates, and total biomass
do not differ given the presence or absence of
endophytes in seedlings of tropical angiosperms
such asTheobroma cacao,Gustavia superba, and
Faramea occidentalis(Arnold unpublished data).
Similarly, Arnold and Engelbrecht (2007) found
that endophyte infection did not influence leaf
fresh weight, dry weight, or water content under
well-watered conditions. While apparently sup-
portin gthe neutralism hypothesis, these studies
raise two questions: (1) to what degree are such
outcomes sensitive to the makeup of endophyte
communities in particular leaves; and (2) how are
endophyte–host interactions shaped by pressure
from natural enemies or abiotic stressors?
Few data are available to address the first ques-
tion, but several studies have started to explore
the second. For example, Arnold and Engelbrecht
(2007) found that endophyte-infected leaves of
Theobroma cacaolose water two times faster than
endophyte-free leaves under severe drought condi-
tions. In addition to the immediate consequences
of desiccation, plants sufferin gfrom increased
water stress may be more susceptible to fungal and
bacterial pathogens (e.g.,Botryosphaeria dothidea,
Maet al.2001;Xylella fastidiosa, McElroneet al.
2001, McElrone and Forseth 2004) and abi-
otic stresses (Thaler and Bostock 2004). Effects
on host water relations are likely most impor-
tant in strongly seasonal tropical forests. In
turn, beneficial effects such as anti-pathogen
defense (see below) may be more important
durin gwet seasons or in everwet forests. Simi-
larly, apparently symptomless infections also can
influence photosynthetic activity. For example,
Pinto et al. (2000) found that infections by
two endophytic Ascomycota (Colletotrichum musae
and Fusarium moniliforme) reduced photosyn-
thetic capacity in maize and banana. Because
photosynthetic capacity is associated with plants’
tolerance of herbivory (Agrawal 2000), endo-
phyte infections may restrict the ability of plants
to cope with damage. In the carbon-limited
environment of the forest understory, the com-
bined cost of reduced photosynthesis and a
decrease in damage tolerance may be especially
problematic.