(Swanston 1974), vegetation structure (Swanston 1969,
1970,1974), and human activities can also influence
their likelihood. These relationships raise numerous
questions: Are spatial and temporal variation in land-
scape dynamics dictated by landslide formation? Is
the relatively wetter Caribbean slope more prone to
sliding than the Pacific slope? Do deforestation and
reforestation influence the likelihood of landslides?
Does vegetation mass progressively burden the under-
lying substrate as succession proceeds on old land-
slides? Does the recurrence interval of landslides in-
fluence vegetation stature by limiting site stability
(Raup 1957)?
A landslide chronosequence in the Luquillo Moun-
tains of Puerto Rico suggested that forest basal area
requires at least 50 years to recover following land-
slides (Guariguata 1990). The rate and species com-
position of revegetation on landslides depend strongly
on the nature of the exposed substrates (Garwood
1985, Guariguata 1990, Myster and Fernandez 1995).
Revegetation is typically more vigorous in the lower
zone of deposition (Garwood 1985, Guariguata 1990),
which reflects both greater colonization and more
favorable growing conditions. Seed input on land-
slides may be very patchy in both lower and upper
zones (Myster and Fernandez 1995). Although soil
nutrient status is also patchy, a gradient of increas-
ing availability may exist from the top to the bottom,
and from the center to the periphery in the upper
zone of a landslide (Guariguata 1990, Myster and
Fernandez 1995). Water availability is typically
greater in the lower zone, due to both downslope flow
and increased soil volume (Carson and Kirkby 1972,
Crozier 1986).
To evaluate the contribution of landslides to land-
scape patterns, we need information on the size class
distribution of slides, their temporal frequency, and
their spatial distribution. The distribution of land-
slides influences the physical structure of vegetation
and soil, energy flow (due to the impact on vegeta-
tion structure and organic matter export), and bio-
geochemical cycling (particularly sediment export
and rejuvenation of soil profiles). Because landslides
are conspicuous on aerial photographs, photogram-
metric analysis of aerial photographs of the Cordillera
de Tilaran could document the spatiotemporal den-
sity of slides (e.g., Gupta and Joshi 1990). Multivari-
ate techniques, including discriminant analysis (e.g.,
Jibson and Keeper 1989) and logistic multiple regres-
sion, can quantify landslide risk.
Vegetation dynamics can also be used in the study
of geomorphological processes in Monteverde (Hack
and Goodlett 1960, Sigafoos 1964, LaMarche 1968,
Veblen et al. 1980). Stand composition can serve as a
marker and delimiter of past events of mass wasting
and floodplain deposition (Hack and Goodlett 1960,
Veblen et al. 1980). Vegetation "markers" (e.g., bent
or damaged trees) can determine relative and abso-
lute dating of events of scour and deposition (Sigafoos
1964), and rates of surface creep and erosion (Hack
and Goodlett 1960, LaMarche 1968).2.4. Paleoecology of the
Cordillera de TilaranPaleoecology is the study of past plant and animal
communities and their relationships to environmen-
tal conditions. Paleoecological interpretation of pol-
len in sediments depends on the accuracy of informa-
tion on species distributions. Oaks, for instance, are
characteristic elements of montane forests in Central
America and, if present in pollen records, indicate
montane conditions (Leyden 1984, Bush et al. 1992).
However, the recent discoveries of oaks at 500 m el-
evation on the ridges of the Osa Peninsula (Soto 1992)
and at 800 m on ridges in the Penas Blancas valley
(W. Haber, pers. comm.) illustrate the difficulties of
inferring paleoconditions from the presence of pollen.
There have been no formal paleoecological stud-
ies in the Cordillera de Tilaran. Lake sediments from
other Central American and northern Andean sites
record lake geochemistry, sediment character, and
pollen composition, allowing paleoecological recon-
struction of glacial and postglacial conditions in the
area (van der Hammen 1974, Binford 1982, Leyden
1984, Bush et al. 1990,1992, Islebe et al. 1995). Tem-
peratures in both lowland and montane neotropical
regions were about 5°C lower at the last glacial maxi-
mum (18,000-20,000 years ago). Palynological evi-
dence (i.e., fossil pollen record) suggests that charac-
teristic elements of the vegetation ranged as much as
1500 m lower in the Andes at the last glacial maximum
(van der Hammen 1974, Hooghiemstra 1989, Bush
et al. 1990) and 600-900 m lower in Central America
(Leyden 1984, Bush et al. 1992, Islebe et al. 1995).
Montane species limited by warmer temperatures
at the lower limits of their range would have had
much larger ranges in the Cordillera de Tilaran if tem-
peratures were 5°C cooler than at present. The cloud
forests of the Cordillera, however, seem influenced
as much by exposure to the mechanical stresses and
clouds of the trade winds as by temperature (Lawton
and Dryer 1980). Knowledge of the paleoecology in
the range requires more detail on the paleoclimate,
particularly of trade winds, the ITCZ, the impact of cold
fronts from continental North America, and empirical
studies of lake sediments within the Cordillera.
A remarkably well-distributed set of small lakes
exists in the Cordillera. The lake on Cerro Chato26 The Physical Environment