Figure 5.6. Aftermath of the 1987 crash of anuran
populations in the Monteverde area in relation to local
patterns of distribution and diversity. During 1990-94,
40% of the species were missing (left column) from the
study area of Pounds et al. (1997). The remaining
columns depict hypothetical percentages of missing
species under different assumptions concerning the
spatial pattern of the crash. Each percentage was
calculated by tallying the number of species missing in
the corresponding distributional category (Table 5.2) and
dividing by the total number of species (N = 50). Because
many of the species restricted to either the Pacific or the
Caribbean slope extend to the highest elevations (Zones 3
and 4), these mountaintop areas are affected in each
scenario. If the highest elevations alone had been
affected, only three species would have disappeared.
Since completion of the 1990-94 study, one species has
reappeared (Forrer's Leopard Frog on the Pacific Slope in
1997).
The Middle American herpetofauna appears to
have evolved into a distinctive assemblage during a
long period of isolation (Savage 1982). According to
this view, an oceanic barrier isolated the fauna from
that of South America, and mountain building and
climate change isolated it from the North American
fauna. With the uplift of the Middle American high-
lands and the resultant closure of this oceanic barrier
late in the Tertiary, the Middle American and South
American assemblages intermixed. About 25% of the
genera represented at Monteverde appear to be of
South American origin (Savage 1982, J. Savage, un-
publ. data).
This uplift of the central axis of volcanic moun-
tains also played a key role in the diversification ofamphibians and reptiles within Middle America.
The once relatively uniform assemblage of wide-
ranging lowland species became fragmented into
eastern and western elements. Upland faunas, derived
from lineages that rode the uplift or invaded the
mountains afterward, became isolated in three prin-
cipal highland areas—the Sierra Madres of Mexico,
the highlands of northern Central America, and the
Cordilleras of Costa Rica and Western Panama. These
upland assemblages evolved to be strikingly different
from one another. As a result, Monteverde's herpeto-
fauna has relatively little in common with the high-
land faunas of northern Central America. A mix of
upland specialists and principally lowland forms, the
fauna shares more species with La Selva in the nearby
Caribbean lowlands than it does with Sierra de las
Minas in the mountains of Guatemala (Fisher exact
test comparing the calculated proportions, P= .00001;
Table 5.3).
The difference between the highland faunas is es-
pecially striking for amphibians. The proportion of
species in common between Monteverde and Sierra
de las Minas is much lower for frogs, toads, and sala-
manders than it is for lizards and snakes (P= .00001).
Compared to reptiles, amphibians have undergone
greater differentiation into distinctive upland assem-
blages in the different highland areas. Consistent with
this pattern, the proportion of species shared between
Monteverde and La Selva is lower for amphibians
than it is for reptiles (P = .001; Table 5.3). A majority
of Monteverde's reptile species are primarily lowland
forms that have colonized the mountain slopes; a rela-
tively large fraction of the amphibian species are
upland specialists that do not inhabit the lowlands.
The difference is also reflected in patterns of ende-
mism. About 40% of Monteverde's amphibian species
are restricted to the uplands of Costa Rica and west-
ern Panama, compared to about 14% of the reptile
species (P= .0002; Table 5.4; Appendix 8). The results
are similar when anurans are compared to lizards and
snakes (36% vs. 14%; P= .002). A large core element
of Monteverde's amphibians belong to a distinctive
upland assemblage rich in endemic species.
Amphibians have apparently experienced stronger
geographic isolation than reptiles. During Pleistocene
glaciations, climatic and vegetation zones on tropi-
cal mountains extended to lower elevations than they
do today (Colinvaux et al. 1996). Although these epi-
sodes probably facilitated the movement of highland
species between massifs (Duellman 1979, Savage
1982), amphibians might have been less likely than
reptiles to accomplish this movement, as they are
generally poorer dispersers. Also, lowland environ-
ments might be stronger barriers to amphibians than
to reptiles if the former are more montane in distri-156 Amphibians and Reptiles