Figure 9.4. Topographic map of
the study watershed on Cerros
Centinelas. The Penas Blancas
headwall cliff is immediately east
of the watershed. Gaps from three
years are presented; fine hatching
denotes those of July 1987 to June
1988, coarse hatching those of
July 1988 to June 1989, solid
those of July 1985 to June 1990.
Contours are from an arbitrary
reference point at an approxi-
mate elevation of 1560 m.Disturbance regimes elsewhere in the Monteverde
forests are similar. Gaps covered 1.5% of five 0.5-km
line transects through windward and leeward cloud
forest each year (Murray 1986, 1988). These Monte-
verde forest values are similar to values from other
neotropical lowland forests (Brokaw 1985, Denslow
1987, Hartshorn 1990). Due to the variation among
years, in the Brillante sample, however, turnover times
calculated from 3-year periods ranged from 51 to 147
years, which emphasizes the importance of long-term
studies.
Variation among gaps Not all gaps in the canopy are
created equal. This observation is central to the sug-
gestion that tropical tree species diversity (and the
animal diversity dependent upon it) is in part the
result of specialization in the "regeneration niche"
(Ricklefs 1977, Denslow 1980, Hartshorn 1980, Orians
1982, Brokaw 1985). Gaps vary dramatically in size,
ranging from hectares (in areas denuded by landslides
or catastrophic storms) to square centimeters (in the
spaces between leaves). Within the Cerro de los Cen-
tinelas study area in Monteverde, gaps larger than
4 m^2 in planar projection, the apparent lower limit
for shade-intolerant species to become established in
elfin forests, were included in gap censuses.
Most gaps are small; on the windward slope of the
study area approximately 55% are <15 m^2 while <5%
are >105 m^2 (Lawton and Putz 1988). The largest 5%,
however, contribute more than one-third of the total
area opened by disturbance. These patterns applies
elsewhere in Monteverde cloud forests. In the five
0.5-km line transects of Murray (1986, 1988), gaps
<5 m^2 accounted for 57% of the number of gaps formed,
but 5% of the gaps were >200 m^2. These few large gaps
accounted for the majority of the area opened.
In a study of 88 gaps in the dwarfed forests of the
windward slope of the study area, the following at-
tributes were measured: gap area, gap aperture (the
angular opening from the gap center to the sky), the
area covered by nurse logs >20 cm diameter, the area
of mineral soil disturbed in uprooted pits and
mounds, the size and identity of the gap-maker tree,
whether the gap-maker snapped or was uprooted, the
height above the ground at which the failure occurred,
and, for a subset of gaps, the photosynthetically ac-
tive radiation (PAR, 400-700 nm) 1.5 m above the
ground at gap center (Lawton and Putz 1988). Of these
gaps, 41% were formed by uprooted trees, 39% by
snapped trees, and the remainder by limbfall, the col-
lapse of epiphyte masses, or dead standing trees killed
by lightning.
A unique characteristic of the elfin forests is that
some trees fall slowly. Snapped trees come down
quickly, but uprooting can take months or even years.
Many elfin forest trees have strong crowns, so when
they topple they do not shatter into a pile of debris309 Ecosystem Ecology and Forest Dynamics