34 Paul V.A. Fineetal.
interact and modulate each other’s effects in a
complex way. Thus, it is difficult to make robust
inferences about process from current distribu-
tions of range size. However, ecological modeling
of lineage and population dynamics over spatial
and temporal gradients is a topic that is ripe for
progress. A good theoretical starting point might
be Chown’s (1997) modified “fission” model in
which the probability of speciation peaks at rel-
atively low values for range size, dispersal rate,
and abundance, but declines slower than the prob-
ability of extinction as those variables increase.
Th eimportanc eof ar ea in driving this patt ern
might become clearer if population genetic mod-
els incorporating those variables were used to
test the GAH. Furthermore, combining theoret-
ical models with population genetic data is also
likely to be productive. In an empirical study,
Martin and McKay (2004) surveyed geographi-
cal patterns of genetic diversity over a wide range
of animal species and found that at low lat-
itudes, genetic divergence between populations
(and hence potential for speciation) is greater
than at high latitudes. Similar studies that focus
on genetic patterns within and between tropi-
cal and extra-tropical biomes are needed to more
adequately test the speciation models underlying
th eGAH.
TESTING THE GAH
In theory, testing the GAH involves simply
calculating a biome’s surface area and count-
ing the number of species found within it, and
then comparing those figures across biomes. This
is difficult to do in practice for several rea-
sons. Biome boundaries may be different depend-
ing on which organisms are considered. Biomes
are traditionally defined by climatic variables
(temperature and rainfall), but these vary in how
they affect the ranges of different kinds of organ-
isms, for example plants versus animals. Birds,
for instance, have an array of behavioral and
physiological adaptations to cold and drought
that ar enot analogous to th eadaptations of
plants. As a consequence, dozens of bird species
hav erang es that span th eNorth Am erican conti-
nent, including boreal, temperate, and subtropical
biomes – in contrast to North American tree
species, which rarely cross even one biome border
(Fin e2001).
A second problem in testing the GAH is the
pervasive lack of data on species distributions.
Once a biome is carefully defined for a group
of organisms, it must be inventoried. For no
group of organisms, not even birds, are com-
prehensive species distribution maps available for
the entire globe. Moreover, even where species
lists exist, caution should be exercised in com-
paring species richness among continents. For
example, North American birds have been inten-
sively studied for centuries, and in some cases
very closely related taxa are considered distinct
species (Zink 2004). In contrast, some Amazo-
nian bird species are morphologically uniform,
but analyses of genetic data reveal substantial
geographic structure, suggesting a plethora of
undiscovered cryptic species in the Amazon (Bates
et al. 1999).
A third challenge in testing the GAH is that it
is an equilibrium hypothesis – that is, the areas
of biomes should correlate with species totals
only if the dynamics of species turnover are at
equilibrium. However, there is little reason to pre-
suppose equilibrium at any given time for all
biomes because paleoclimates have not been stable
(Ricklefs 2004). Current diversities may reflect
th eclimatic history of biom ear eas as much as
they reflect current climatic conditions. Indeed,
sinc eclimat echang eand glaciations dispropor-
tionately affect biomes that are closer to the poles
compared with those closer to the equator, it is
possibl ethat th ebiota of north ern biom es ar e
depauperate and below their equilibrium values,
especially for long-lived and slowly dispersing taxa
like trees (Svenning and Skov 2004). Similarly,
tropical biomes once covered a much larger area
than they do today, and species totals of tropical
rainforests may be “above” equilibrium diversity.
Thus, an adequate test of the GAH must incorpo-
rat eth esiz eof biom ear eas ov er tim e. In th er est
of this chapter, we estimate the geographic extent
of boreal, temperate, and tropical biomes for each
continent over the last 55 million years and corre-
late these areas integrated over time with current
tree species richness to test the geographic area
and ag ehypoth esis (GAAH).