Fire-adaptive Trait Evolution - Fire in Mediterranean Ecosystems - page 270

resprouters (e.g. see Table 3.1 ). Thus, fire-adaptive traits, climate and geology (see
Fig. 1.4 ) all seem to have played a role in accounting for different patterns of
diversification in resprouters and seeders.
This interaction between fire and geology can be invoked in explaining patterns
of fire traits in many MTC genera. The South African grass Ehrharta comprises
both resprouters and obligate seeders and much of its radiation appears to be tied
to substrate heterogeneity (Verboom et al. 2004 ). In California large outcrops of
serpentine substrate also have had selective effects on seeders vs. resprouters
(Safford & Harrison 2004 ).
Orogenic changes such as Quaternary mountain building also have been con-
sidered a factor in the diversification of the MTC taxa such as the Californian
genus Arctostaphylos , which is dominated by obligate seeding species (Raven &
Axelrod 1978 ). Although intuitively appealing it needs to be recognized that there
is little evidence that radiation was this recent. The genus comprises two lineages,
both of which are almost entirely composed of postfire seeders, and the majority
of species are obligate seeders. Many of the species are localized endemics and
some are strictly confined to unusual substrates, e.g. A. myrtifolia found only on
outcroppings of Eocene lateritic soils, and others are serpentine endemics (Parker
et al. 2009 ). Many of the endemic species are in coastal southern and central
California and some of these ranges were up to 2000 m in the Miocene (Stadum &
Weigand 1999 ).
Pleistocene mountain building has also been used to explain the origin of many
narrow endemics in the very large woody Californian genus Ceanothus , most of
which are obligate seeders (Raven & Axelrod 1978 ). However, this ignores the
substantial evidence that there was significant elevational variation throughout
the Miocene (Wolfe et al. 1998 ) that could have provided habitat differentiation.
In addition, latitudinal replacement of localized endemics within the state appears
to be more important than altitudinal variation and rain shadows (Cody 1999 ).
One could argue that only about 10% of the extant species are reported from the
middle to late Miocene fossil record, and thus most are more recent. However,
assigning these fossil impressions to contemporary species is problematical
(Edwards 2004 ), which is not surprising since no specialist today could identify
most species solely from leaf impressions. The difficulty in answering this question
of timing of diversification with fossils is that much of the present diversity in
Ceanothus comprises localized endemics. Even if they were extant during the
Tertiary, macrofossil samples comprise only a tiny fraction of that landscape
and thus sampling theory alone would predict that relatively little of the diversity
would have been recovered from fossils. In support of recent speciation in Ceano-
thus , one could cite the fact that there is very little molecular separation evident
between species within subgenera; however, rather than indicative of recent sep-
aration, this also could be explained by introgression, which is well known in the
genus (Hardig et al. 2000 ).
Some diverse genera have radiated relatively recently, for example the Mediter-
ranean Basin Cistus (Guzma´ n et al. 2009 ), possibly in response to intensification of