Peter W Post 
A Greensmith, Ardea London 
Several Andean bird species are 
widely distributed: the bar-winged 
cinclodes, above, breeds from 
Colombia to Patagonia. The 
distribution of others is more 
restricted: the ochre-naped ground 
tyrant, right, breeds in the 
Patagonian Andes and migrates 
north to winter in the puna of Peru. 
a situation likely to result from in- 
terspecific competition. But in many 
other species, the primary causal 
agent of their Andean distribution 
may be associated with the glacial 
cycles. Competitive interactions may 
occur at all times, but in most cases, 
the ever changing nature of the ex- 
ternal environment would probably 
prevent interspecific competition from 
lasting long enough to settle distri- 
bution patterns as it has with the can- 
asteros. 
So far I have examined some mech- 
anisms that help explain species dy- 
namics in the high Andes once an- 
cestral species occupied the newly 
opened habitats that formed after the 
mountains reached great heights. But 
how old is the fauna, and where did 
the ancestral stocks come from? A 
classic way «of answering these sorts 
of questions has been first, to look 
at levels of endemism — which can pro- 
vide an indication of the age of the 
fauna and the relative importance of 
immigration and local speciation — 
and then to compare the fauna with 
others nearby that might have re- 
ceived or contributed species during 
the time being considered. What do 
these two aspects of faunal analysis 
tell us about high Andean birds? 
In the high Andes, endemic taxa 
would be those restricted to above tim- 
berline. The problem with this defi- 
nition is that while some places have 
a sharp timberline, which effectively 
delimits two drastically different habi- 
tats (for example, cloud forest below, 
paramo grassland above, as in Colom- 
bia), in other areas (including the 
western slopes of the Peruvian Andes) 
there is no timberline. In the analysis 
that Simberloff and I carried out, we 
defined as endemic a taxon that lived 
only in the paramo and puna types 
of vegetation. We studied 147 species 
at forty sites (the total avifauna in- 
cludes about 180) and found 79 to 
be endemic. In other words, nearly 
half the high Andean species are en- 
demic, a rather high percentage. 
In contrast, very few genera — 4 out 
of 85 — are endemic according to the 
same definition: a wader ( Phegornis ), 
two hummingbirds ( Oreotrochilus and 
Oxypogon), and a finch ( Idiopsar ). 
Finally, not a single avian family is 
endemic. 
The low level of endemism at the 
genus level suggests a relatively young 
fauna. The high percentage of endem- 
ism at the species level, however, could 
indicate a rather old fauna. This ap- 
parent contradiction might be resolved 
if the species endemism is the result 
of a high rate of recent speciation. 
This speciation could be either local — 
that is, occurring only within the high 
Andes — or the result of immigration 
followed by differentiation. 
In birds, speciation is generally 
thought to be allopatric, involving the 
geographical division of a single spe- 
cies into two or more daughter species. 
Two good examples of local speciation 
in the high Andes include the white- 
throated and red-backed sierra 
finches, Phrygilus erythronotus and 
P. dorsalis, which look a little like 
oversized juncos, and the plain-capped 
and cinereous ground tyrants ( Muscis - 
axicola alpina and M. cinerea), slen- 
der, long-legged, terrestrial flycatch- 
ers. In both cases, the two species 
replace each other geographically ex- 
cept in a very narrow contact zone. 
A later stage in speciation is sympatry, 
when two formerly isolated species live 
together in the same habitat or share 
the same distributional range. I con- 
sider the plumbeous (P. unicolor) and 
red-backed sierra finches to be such 
a sympatric pair, as are the cinereous 
and puna (M. juninensis) ground ty- 
rants. 
To determine how much speciation 
has occurred in situ, we attempted 
to arrive at an estimate of the number 
of allopatric and sympatric species 
among the 147 birds in our study. 
Allopatric species representing re- 
cently completed speciation number 
about 12 to 20; sympatric species that 
54 
