are so scarce that they are virtually 
useless. Biogeographers are, however, 
beginning to learn about the history 
of some Andean environments through 
the work of palynologists, who study 
pollen, and about the history of the 
mountains themselves through the ef- 
forts of geologists. 
In the Andes today, permanent gla- 
ciers are found only at the highest 
altitudes, but in the past they extended 
much farther down the mountainsides. 
In several areas of Venezuela, Colom- 
bia, and Peru, moraines and fluvio- 
glacial accumulations of rocks and 
other materials attest to between one 
and three major glacial advances. But 
if the Andes, like parts of Europe and 
North America, witnessed four major 
advances, why are they not all re- 
corded? One view is that at the time 
of the earliest glacial advances in the 
Northern Hemisphere, the Andes 
were not sufficiently high to have had 
glaciers. The high, alpinelike environ- 
ments of the Andes (the wet paramos 
of Venezuela, Colombia, and Ecuador 
and the drier puna of Peru, Bolivia, 
northern Argentina, and northern 
Chile) probably only date from the 
mid- to late-Pleistocene era, one to 
two million years ago. 
While not definitive, pollen data 
also suggest that the Andes have been 
high for a geologically short time. 
Palynologists have carried out several 
detailed studies of cores taken from 
ancient bogs in the Andes of Ven- 
ezuela, Colombia, and Bolivia. They 
have dated the peat from different 
core depths and counted and identified 
by species the pollen grains found in 
the peat. The resultant “pollen pro- 
files” indicate how much pollen was 
contributed by different species of 
trees or grasses at different times and 
different places. For example, at one 
locality in the Colombian Andes, cur- 
rently at an altitude of 13,500 feet 
and situated fully within the paramo 
zone, the vegetation apparently 
changed repeatedly during the last 
15,000 to 20,000 years. At times, 
woodlands occurred at the site, while 
in other periods the vegetation resem- 
bled the wet grassland that now domi- 
nates the area. The overall picture 
of the Andes is one of vegetation zones 
that have been alternately lower and 
higher over about the past 100,000 
years. During cold, glacial episodes, 
alpinelike vegetation reached much 
lower altitudes than it does now. By 
contrast, at the peak of warmer, inter- 
glacial episodes, the zones moved up, 
so that the paramo and puna region 
was even more restricted to moun- 
taintops than today. 
If a given bird species now found 
exclusively in alpinelike grassland had 
the same habitat requirements in the 
past, its distribution would have 
changed over time according to the 
periodic contractions and expansions 
of its environment. Theoretically, this 
“cycle” of habitat disruption could 
have led to the formation of new spe- 
cies, and not surprisingly, different 
subspecies or, indeed, species of birds 
do sometimes occur on mountaintops 
that are now geographically isolated 
from one another. In other cases, dif- 
ferentiated taxa are found in close geo- 
graphical contact; careful study of 
these contact zones has revealed hy- 
bridization in some instances and 
slight range overlaps without hybrid- 
ization in others. 
We have here a model for speci- 
ation, and hence faunal buildup, from 
within the system. The bearded hel- 
metcrest ( Oxypogon guerinii), for ex- 
ample, is a small hummingbird inhab- 
iting the high altitudes of four isolated 
Andean regions: three in Colombia (in 
the Santa Marta Mountains, the east- 
ern Andes, and the central Andes) 
and one in Venezuela. Because the 
birds in each of the four allopatric, 
or geographically isolated, populations 
are morphologically differentiated, or- 
nithologists once treated them as spe- 
cies; they are now considered sub- 
species. A possible sequence of events 
leading to this pattern is not hard to 
imagine. During a glacial period, a 
formerly widespread ancestor may 
have lived all along the central and 
eastern Andes and extended north into 
Venezuela and the Santa Marta 
Mountains. At this time, the paramo 
zone, in which the ancestral helmet- 
crest lived, was lower and more con- 
tinuous than today. As glaciers 
melted, the paramo zone moved up- 
ward and became restricted to small 
pockets on mountaintops. Today, only 
the tallest summits, possessing the ap- 
propriate type of scrub and woodland, 
harbor the helmetcrest, whose isola- 
tion into four patches might have be- 
gun as recently as 20,000 years ago. 
Should the climate become colder 
and glaciers push down the paramo 
habitats once again, I predict that the 
four isolates would come in secondary 
contact as their preferred habitat be- 
came continuous or nearly so. If com- 
plete speciation has taken place by 
the time of this secondary contact, 
geographical overlap might occur, and 
a region formerly inhabited by one 
ancestral species could be occupied 
by two daughter species. Such spe- 
ciation may be a corollary of fluc- 
tuations in the environment that, in 
turn, resulted from climatic changes 
during the glacial-interglacial cycles. 
Much of the previous reasoning is 
based on the assumption that a spe- 
cies’ habitat requirements are largely 
invariant in time. This is not, however, 
necessarily the case. The bar-winged 
cinclodes ( Cinclodes fuscus), a brown, 
dipperlike bird that runs along the 
ground in search of insects and other 
small animals, is one of several wide- 
ranging high Andean birds that have 
geographically variable requirements. 
In some parts of the high Andes, this 
species of cinclodes is found in rel- 
atively dense and moist grassland, 
whereas elsewhere it inhabits much 
drier and more open scrub. If its habi- 
tat requirements vary geographically, 
there is no reason to believe they did 
not also vary temporally. 
Another complicating factor is that 
species do not live alone; they share 
their preferred habitat with other spe- 
cies. To what extent are the habitat 
preferences of one species modified 
by the presence (or the absence) of 
another species, especially an ecologi- 
cally similar relative? Does inter- 
specific competition somehow rule 
how many species can live together 
in the same patch of habitat? While 
competition between species is an im- 
portant ecological — and hence also 
evolutionary — factor, the precise role 
of competition in determining species 
ranges and influencing speciation (or 
extinction) is currently under great 
debate. 
After an extensive analysis involv- 
ing seventy-nine species of land and 
freshwater birds endemic, or re- 
stricted, to the high Andes (represent- 
ing about 44 percent of the total 
avifauna), Daniel Simberloff of Flor- 
ida State University and I found good 
evidence that competition had been 
a causal factor in the distribution pat- 
tern for no more than six of the sev- 
enty-nine species. For example, in the 
case of canasteros (small, drab birds 
of the genus Asthenes that look like 
long-tailed wrens and build dome- 
shaped, covered nests), the distribu- 
tion pattern is a mosaic in which sev- 
eral species do seem to replace each 
other at given study sites, rather than 
coexist. We concluded that the species 
of canasteros are mutually exclusive. 
53 
