•VESICULAR-ARBUSCUlAR MYCORRHIZAE OF EPIPHYTIC AND
TERRESTRIAL PIPERACEAE
Blase Maffia, Nalini ML Nadkarni,'& David P Janosn order to survive the oligotrophic (nutrient-poor)
conditions in forest canopies (Benzing 1981,
1990), epiphytes have developed morphological
and physiological adaptations for garnering and re-
taining nutrients. Epiphytes derive nearly all of their
nutrients from atmospheric sources (Nadkarni and
Matelson 1991), which are generally dilute, especially
in phosphorus (Vitousek and Sanford 1986). Vesicular-
arbuscular mycorrhizal fungi (VAM) are one means
by which epiphytes might increase nutrient uptake.
The ability of VAM to reach the canopy is uncertain,
however, because the large spores of these fungi are
not normally dispersed by wind (Janos 1993). Several
studies have documented VAM in the Piperaceae (St.
John 1980, Garcia Yazquez-Yanes 1985, Bermudes
and Benzing 1987, Mohankumar and Mahadevan
1987), a common tropical plant family with epiphytic
and terrestrial species. However, nine species of this
family collected in montane forests of Costa Rica did
notons/have VAM (Lesica and Antibus 1990).
We surveyed the occurrence of VAM among 27 spe-
cies of Piperaceae in the field in Monteverde (Maffia
et al. 1993). We focused on this family because it is
species-rich (Burger 1971) and includes many epi-
phytes. Some of these are facultative (found in both
terrestrial and arboreal habitats), which allows for com-
parison of VAM between the canopy and the ground.
Preliminary work documented low levels of VAM of
terrestrial Pothomorphe umbellatum L. (Piperaceae)
and revealed VAM structures (e.g., internal and exter-
nal hyphae, vesicles, and spores) in root sections col-
lected from random locations in canopy root mats
(Maffia 1990).
We collected roots of epiphytic and terrestrial
Piperaceae in and adjacent to the leeward cloud for-
est study site and secondary vegetation (pasture trees,
secondary forest, and along roadside), including 17
species of Peperomia, 8 species of Piper, and 1 spe-
cies each of Pothomorphe and Sarcorachis (Maffia
et al. 1993). Epiphytic samples were taken from bare
branches, tree trunks, and interwoven "mats" of live
roots and humus on horizontal branches. We sepa-
rated apparently living roots following St. John and
Uhl (1983), preserving them in formalin-propanol-
acetic acid within 3 hr of collection, and cleared and
stained samples following Phillips and Hayman
(1970). The amount of mycorrhizal colonization was
determined by recording the presence or absence of
VAM in 1-cm sections of root. We also noted extra-
matrical spores and auxiliary cells associated with
root segments.
Of the 27 species examined, two terrestrial Pepero-
mia species, two terrestrial Piper species, and Potho-
morphe umbellatum contained internal vesicles and/
or arbuscules. Overall, colonization was extremely
low, with only 21 cm of 4867 cm of fine roots exam-
ined having internal colonization (<1%). The terres-
trial Po. umbellatum accounted for half of all inter-
nal colonization. In contrast, we found moderate
amounts of typical coarse VAM external hyphae as-
sociated with both epiphytic and terrestrial root seg-
ments, which included 15% of all epiphytic and 13%
of all terrestrial Piperaceae root segments. Glomus and
Acaulospora spores and Gigaspora auxiliary cells oc-
curred in both terrestrial and epiphytic habitats, al-
though fewer than 1% of root segments had associ-
ated spores (Maffia et al. 1993).
The low colonization of roots of Piperaceae sug-
gests no consistent pattern of usage of VAM by ter-
restrial or epiphytic plants. A greenhouse study
carried out on Peperomia costaricensis produced sig-
nificantly higher amounts of internal colonization (up
to 28%) with native and commercial inoculum when
the plants were raised on an infertile substrate (Maf-
fia 1990, Maffia et al. 1993). Two non-exclusive hy-
potheses may explain the paucity of VAM on Pipera-
ceae in the field. First, low colonization levels could
be a consequence of relatively high site fertility. High
atmospheric inputs of dissolved inorganic nutrients,
characteristic of tropical cloud forests (Kellman et al.
1982, Clark 1994), may make VAM unnecessary to
meet nutrient requirements. Second, VAM formation
in the canopy might be limited by a lack of inoculum
(Benzing 1983, Lesica and Antibus 1990).
We hypothesize that the canopy comprises a "habi-
tat mosaic" (Janos 1993). Developing canopy mats
accruing vegetation might progressively accumulate
VAM inocula dispersed by any of several potential
vectors (e.g., birds, ants, small mammals). The great-
est inoculum potentials in the canopy should occur
in old-growth forest with low tree and branch turn-
over. Chance deposition of large numbers of VAM
propagules, however, would cause a "young" canopy
mat to have a high inoculum potential, but we pre-
dict that this occurs less often than gradual accretion
of inoculum.
Although spatially patchy, canopy mats with abun-
dant VAM might serve as inoculum sources through338 Ecosystem Ecology and Forest DynamicsI