10. Terrestrial Vertebrate Diversity in New World Temperate Rainforests 221
longipilis , and O. longicaudatus). Captures of the remaining four species
constituted only 3.8 % , as compared to about 39% for five remaining
species in a similar sampling effort conducted in La Picada in 1983
(Patterson, Meserve, & Lang , 1989). Therefore , small forest islands
appear to support a more depauperate small-mammal fauna than contiguous, intact forests.
Meserve, Kelt, & Martinez (1991a) sampled 15 primary-growth temperate.rainforest sites between approximately 41°S and 45°45' S latitude,
using a standarized trapping protocol during 1985-1987. The transect
followed the Carretera Austral (Southern Highway) as it was being constructed through formerly undisturbed Valdivian and north Patagonian
rainforest. Small-mammal species number declined progressively, from a
maximum of seven species in La Picada to a minimum of three near the
southern end of the transect in Chile Chico; latitude was the single most
predictive variable explaining species number in stepwise regressions.
Thus, point diversity appears to decline progressively to the south, and
widespread species such as A. olivaceus and O. longicaudatus predominate
at more southerly sites (73% of the captures for the five southern sites;
Meserve et al., 1991a). While it may be tempting to conclude that this is
due to a peninsular effect as rainforest becomes more narrowly restricted
in the southern terminous of South America, examination of vegetation
maps (e .g. , Hartwig, 1991) shows that the most conspicuous narrowing of
hygrophytic forests occurs near the Patagonian icefields (ca. 47°S), well
to the south of the transect. Furthermore, virtually all small-mammal
species occurring at the northern end of the transect have been historically
documented at the southern end with exception of the marsupials
(Rhyncholestes raphanurus and Dromciops gliroides) , which do not occur
south of 42°S on the mainland (Kelt & Martinez, 1989; Meserve et al. ,
1991a). Clearly, other factors must be operating here. We suggest that
widespread extirpation of forests in the Puerto Aisen region, which was
settled much earlier than the intervening forest sampled by the transect,
has already resulted in local depauperization. Although the central valley
region also has suffered massive deforestation between 37°Sand 42°S,
coastal forests still retain most species, perhaps because they are relatively
continuous and/or secondary-growth forests such as at San Martin that
have recovered sufficiently (ca. 100 yrs). Even in the latter case , however,
most small mammals are comprised of two species (A. olivaceus and O.
longicaudatus) , unlike the precordilleran forests in Antillanca and La
Picada but similar to north Patagonian forests sampled by Meserve et al.
(1991a) .
Finall y, Johnson, Franklin , Iriarte (1990) describe an apparent
" dilemma" of high small-mammal species numbers at the interface of the
southern forests and steppes in Chilean Patagonia (ca. 50°38' S). Here a
total of 38 species of mammals are reported, of which 16 species are
nati ve rodents. This is in contrast to Valdivian forests , where rodents
longipilis , and O. longicaudatus). Captures of the remaining four species
constituted only 3.8 % , as compared to about 39% for five remaining
species in a similar sampling effort conducted in La Picada in 1983
(Patterson, Meserve, & Lang , 1989). Therefore , small forest islands
appear to support a more depauperate small-mammal fauna than contiguous, intact forests.
Meserve, Kelt, & Martinez (1991a) sampled 15 primary-growth temperate.rainforest sites between approximately 41°S and 45°45' S latitude,
using a standarized trapping protocol during 1985-1987. The transect
followed the Carretera Austral (Southern Highway) as it was being constructed through formerly undisturbed Valdivian and north Patagonian
rainforest. Small-mammal species number declined progressively, from a
maximum of seven species in La Picada to a minimum of three near the
southern end of the transect in Chile Chico; latitude was the single most
predictive variable explaining species number in stepwise regressions.
Thus, point diversity appears to decline progressively to the south, and
widespread species such as A. olivaceus and O. longicaudatus predominate
at more southerly sites (73% of the captures for the five southern sites;
Meserve et al., 1991a). While it may be tempting to conclude that this is
due to a peninsular effect as rainforest becomes more narrowly restricted
in the southern terminous of South America, examination of vegetation
maps (e .g. , Hartwig, 1991) shows that the most conspicuous narrowing of
hygrophytic forests occurs near the Patagonian icefields (ca. 47°S), well
to the south of the transect. Furthermore, virtually all small-mammal
species occurring at the northern end of the transect have been historically
documented at the southern end with exception of the marsupials
(Rhyncholestes raphanurus and Dromciops gliroides) , which do not occur
south of 42°S on the mainland (Kelt & Martinez, 1989; Meserve et al. ,
1991a). Clearly, other factors must be operating here. We suggest that
widespread extirpation of forests in the Puerto Aisen region, which was
settled much earlier than the intervening forest sampled by the transect,
has already resulted in local depauperization. Although the central valley
region also has suffered massive deforestation between 37°Sand 42°S,
coastal forests still retain most species, perhaps because they are relatively
continuous and/or secondary-growth forests such as at San Martin that
have recovered sufficiently (ca. 100 yrs). Even in the latter case , however,
most small mammals are comprised of two species (A. olivaceus and O.
longicaudatus) , unlike the precordilleran forests in Antillanca and La
Picada but similar to north Patagonian forests sampled by Meserve et al.
(1991a) .
Finall y, Johnson, Franklin , Iriarte (1990) describe an apparent
" dilemma" of high small-mammal species numbers at the interface of the
southern forests and steppes in Chilean Patagonia (ca. 50°38' S). Here a
total of 38 species of mammals are reported, of which 16 species are
nati ve rodents. This is in contrast to Valdivian forests , where rodents