and carbon in montane forests may be bound in or-
ganic forms for long periods of time (Edwards and
Grubb 1977, Tanner 1985) before they are released by
invertebrate, fungal, and microbial decomposers.
Only preliminary work on leaf decomposition in
Monteverde has been published (Nadkarni and Matel-
son 1991). One study involved the incubation of fallen
litter in tethered litterbags in the canopy of the lee-
ward forest site to quantify the amount of input of
intercepted host tree litterfall into epiphyte commu-
nities. Litterbags (N= 140) were tethered to random-
ized locations on the upper surface of large interior
branches of four trees; subsets were retrieved at tri-
monthly intervals for 15 months. A linear regression
was performed using the natural logarithm of the
mean dry mass remaining over time to calculate k, the
fractional loss rate for the entire 15-month study pe-
riod, following the formula
where Xt and X 0 are the mass remaining at time t and
time 0, respectively (Olson 1963).
Biomass loss from the beginning to the end of the
study was approximately 69% of the original biomass.
The kd value was 0.36, with 64% of the variance ex-
plained by an exponential model. The regression of
percentage biomass loss with time was significant (lin-
ear regression, p < .03). Calculated leaf litter turnover
is 2.8 yr, which is considerably slower than decompo-
sition rates in other tropical montane forests (Wiegert
and Murphy 1970, Edwards 1977). Further work is
needed to quantify long-term litter decomposition rates
of other abscised materials (stems, reproductive parts,
branches, whole trunks) and the processes that regu-
late the immobilization and release of energy and nu-
trients from dead organic matter in this forest.
A second study monitored short-term rates of EM
decomposition. We carried out a six-month field study
to measure changes of biomass and nutrient concen-
tration from EM (N. Nadkarni and T. Matelson, un-
publ. data). In April 1988, we collected epiphyte mats
from the upper canopy (17-22 m) of three Ficus tuerck-
heimii (Moraceae) trees in the study area. These were
cut into small pieces (10 x 10 x 3 cm) of humus/dead
organic matter (65%), vascular plant roots (15%), vas-
cular plant shoots (5%), and bryophytes (15%). Each
of a further 20 replicates was placed on top of the leaf
litter layer at random locations on the forest floor, 10
in recent treefall gaps, and 10 randomly located in
the forest interior. All samples were tethered to the
ground and collected after six months.
Decomposition of individual pieces of EM was vis-
ually apparent during the study period. Tethered epi-
phyte mats on the ground had a significantly lower
biomass at the end of the six-month field incubationthan their original biomass (mean percentage loss =
16.4). Nutrient concentration and total amount of nu-
trients for all macronutrients except potassium were
significantly higher, and potassium was significantly
lower, than the original concentration and amounts
at the end of the incubation, suggesting that epiphyte
mats are net immobilizers of all of the measured
macronutrients except potassium in the short term
(N. Nadkarni and T. Matelson, unpubl. data). Extrapo
lation of net immobilization and flux by fallen EM to
longer-term time scales is an important measurement
to make in the future.- Future Research Directions
Assembling the jigsaw puzzle of the ecology of Monte-
verde ecosystems has only just begun. Existing data
are fragmentary and undoubtedly specific to the sites
where research was carried out. Nonetheless, some of
the primary precursors for understanding the Monte-
verde landscape at the ecosystem level are in place.
Several protected permanent study plots have been
established where trees have been identified and
mapped; baseline data on some of the fundamental
descriptors of forest structure, composition, biomass,
and nutrient capital have been described; and we have
the capacity to carry out basic laboratory studies. A
small cadre of trained local field assistants exists. We
lack sophisticated laboratory equipment and highly
trained technical assistance of field stations such as
at La Selva or BCI. Knowledge of basic abiotic factors
that influence the biota (e.g., soil physical properties,
hydrological attributes such as canopy interception,
groundwater fluxes, micrometeorological conditions)
is incomplete, and an interdisciplinary group of sci-
entists with whom to collaborate on the complex eco-
system-level interactions that occur in Monteverde
needs to be assembled.Acknowledgments We thank Rodrigo Solano for his
help with field and managerial aspects of this project.
We also thank Eric Vance, Frank Setaro, William
Haber, Eric Bello, John Campbell, Jim Crisp, John T.
Longino, the Monteverde community, Doug Schaefer,
and Steve and Karen Ingram. Support was provided
from research grants and Research Experience for
Undergraduate Grant Supplements from the National
Science Foundation (BSR 87-14935, BSR 90-18006,
BIR 93-07771, and BSR 96-15341), the Whitehall
Foundation, the National Geographic Society Com-
mittee on Research and Exploration, the University
of California Santa Barbara Academic Senate, The
Marie Selby Botanical Gardens, and The Evergreen
State College.335 Ecosystem Ecology and Forest Dynamics