Geography Reference
In-Depth Information
Figure 7.8
Biogeographic analysis of
Brachyderes rugatus
in La Palma by Emerson et
al. (2006; reproduced with permission of
Molecular Ecology
). (a-c) Three equally probable
arrangements of the haplotype network; (d) nesting design of the haplotype network, where
numbered circles refer to specific haplotypes.
Sequences analyzed yielded sixty-nine different haplotypes, with eight haplo-
types reported for more than one locality. Network estimation with TCS resulted
in a single network, which has a loop with three possible arrangements of three
phylogroups (
figs. 7.8a-c
). These three possible arrangements are equally prob-
able when one considers cladistic information alone but differ when the geography
is also considered. Network A (
fig. 7.8a
) juxtaposes Phylogroups 1 and 3 with
Phylogroup 2, a biologically realistic arrangement that does not imply long-dis-
tance dispersal of the species through, but not including, an already inhabited
area. Networks B and C (
figs. 7.8b
and
7.8c
) juxtapose Phylogroups 1 and 3 only
by two mutational differences, implying a biotic connection between the geograph-
ically disparate areas of 1 and 3 over a short time period, at the exclusion of the
geographically intermediate area 2.
Figure 7.8d
shows the haplotype network A
and its nesting design. Based on the statistics provided by the contingency ana-
lysis with GeoDis, Emerson et al. (2006) suggested that a history of contiguous
range expansion was the best explanation for the distribution of
B. rugatus
as a
whole, but other processes were also present in the three major clades. Clade 4.1
is best explained by past fragmentation followed by range expansion, and its in-
ternal Clades 1.1 and 3.2 conform to a contiguous range expansion. Clade 4.2 is
also explained by contiguous range expansion, with none of its subclades having
any significant geographic structure. Clade 4.3 conforms to a history of past frag-
mentation followed by range expansion, but internally Subclade 3.6 is consistent
with allopatric speciation, and Subclade 3.7 is best explained by restricted gene
flow with isolation by distance. The phylogeographic history for
B. rugatus
in La
Palma appears complex. The prediction that ancestral haplotypes should occur
predominantly in the northern shield, with derived haplotypes featuring more on El
Bejenado and Cumbre Vieja terrains, was falsified. The prediction that the newer
Bejenado and Cumbre Vieja terrains would have more derived haplotypes, located
terminally in a haplotype network, than the northern shield terrain was also falsi-
fied. The presence of three distinct phylogroups is clearly at odds with a scenario
of a long-term residency in the northern shield, followed by a more recent expan-
sion into the southern terrains. The authors concluded that rather than a single
origin in the northern shield followed by a more recent expansion into the Cumbre
Vieja region, a series of range expansions has occurred, one into the northern
shield and two into the Cumbre Vieja. The authors speculated that these three
range expansions may have tracked the past range expansions of the host spe-
cies
Pinus canariensis
.
