276
A.M. Zink and J. Klicka
These events in earth history share the common feature of providing geographic
isolating barriers that allow for allopatric speciation. Typically barriers such as the
uplift of a mountain range are invoked as the force driving speciation for specific
taxonomic groups. Other barriers likely had more pervasive influences on the biota.
The periodic glaciations of the Pleistocene are thought to have affected many organisms simultaneously. For example, in North America, taxonomists such as
Mengel (1964, 1970) concluded that numerous bird species evolved as a result of
geographic isolation induced by the last one or two cycles of glaciation. The models
are often very specific as illustrated, for instance, by Hubbard's (1973) prediction
that the speciation event leading to modern Calcarius lapponicus and C. ornatus
was precipitated by the advance ca. 100,000 years before present (ybp) of the
Laurentide Ice Sheet, the most recent major southward glacial advance. During
this glacial period, Hubbard postulated that the common ancestral species was split
into two refugia, allowing speciation to occur. Upon retreat of the glacier beginning
18,000 ybp, the new daughter species enlarged their ranges, eventually to their
current sizes. Many other species pairs are thought to have arisen during this period or the previous glacial cycle, the Illinoian (beginning ca. 250,000 ybp). We
(Klicka and Zink 1997) termed this hypothesis the Late Pleistocene Origins (LPO)
model.
Morphological data do not allow testing such a hypothesis beyond the general
phylogenetic pattern. That is, phenotypic patterns might be independent responses
to environmental conditions over long periods rather than historical reflections of
speciation at one or two points in the recent past. Molecular data offer tests of such
historical theories by corroborating that species pairs are in fact sister species, and.
by documenting the degree of genetic differentiation between them. The degree to
which DNA sequence evolution is clock-like is controversial (Avise 1994; Hillis et
al. 1996); however, rough approximations can be made.
Klicka and Zink (1997) measured the mitochondrial DNA distance between 35
pairs of sister species of North American songbirds. If, as previously hypothesized,
these species had originated as a result of the last one or two glacial cycles, then the
mtDNA differences between sister species should be low. How "low" depends on
the calibration of the molecular clock. Klicka and Zink (1997) used a widely accepted rate of 2% sequence divergence per million years, based on estimates for
several avian orders (reviewed by Klicka and Zink 1997). More recently Fleischer
et al. (1998) estimated a rate of 1.6% for cytochrome b sequences in Hawaiian
songbirds. Given either of these rates, the 35 songbird species pairs should have
had a mean mtDNA distance of less than 0.5%, if indeed speciation occurred during the Illinoian or Wisconsinian. However, the average divergence was 5.1 %, a
tenfold greater value than expected. Therefore we concluded that most of the 35
speciation events occurred much earlier than the Illinoian, some even in the Late
Pliocene. Two other lines of evidence supported our rejection of the LPO model.
The distribution of mtDNA distances between 13 additional songbird sister-species
pairs, not previously hypothesized to fit the LPO model, did not differ significantly
from that of the 35 pairs, indicating that what had been considered extremely re-
A.M. Zink and J. Klicka
These events in earth history share the common feature of providing geographic
isolating barriers that allow for allopatric speciation. Typically barriers such as the
uplift of a mountain range are invoked as the force driving speciation for specific
taxonomic groups. Other barriers likely had more pervasive influences on the biota.
The periodic glaciations of the Pleistocene are thought to have affected many organisms simultaneously. For example, in North America, taxonomists such as
Mengel (1964, 1970) concluded that numerous bird species evolved as a result of
geographic isolation induced by the last one or two cycles of glaciation. The models
are often very specific as illustrated, for instance, by Hubbard's (1973) prediction
that the speciation event leading to modern Calcarius lapponicus and C. ornatus
was precipitated by the advance ca. 100,000 years before present (ybp) of the
Laurentide Ice Sheet, the most recent major southward glacial advance. During
this glacial period, Hubbard postulated that the common ancestral species was split
into two refugia, allowing speciation to occur. Upon retreat of the glacier beginning
18,000 ybp, the new daughter species enlarged their ranges, eventually to their
current sizes. Many other species pairs are thought to have arisen during this period or the previous glacial cycle, the Illinoian (beginning ca. 250,000 ybp). We
(Klicka and Zink 1997) termed this hypothesis the Late Pleistocene Origins (LPO)
model.
Morphological data do not allow testing such a hypothesis beyond the general
phylogenetic pattern. That is, phenotypic patterns might be independent responses
to environmental conditions over long periods rather than historical reflections of
speciation at one or two points in the recent past. Molecular data offer tests of such
historical theories by corroborating that species pairs are in fact sister species, and.
by documenting the degree of genetic differentiation between them. The degree to
which DNA sequence evolution is clock-like is controversial (Avise 1994; Hillis et
al. 1996); however, rough approximations can be made.
Klicka and Zink (1997) measured the mitochondrial DNA distance between 35
pairs of sister species of North American songbirds. If, as previously hypothesized,
these species had originated as a result of the last one or two glacial cycles, then the
mtDNA differences between sister species should be low. How "low" depends on
the calibration of the molecular clock. Klicka and Zink (1997) used a widely accepted rate of 2% sequence divergence per million years, based on estimates for
several avian orders (reviewed by Klicka and Zink 1997). More recently Fleischer
et al. (1998) estimated a rate of 1.6% for cytochrome b sequences in Hawaiian
songbirds. Given either of these rates, the 35 songbird species pairs should have
had a mean mtDNA distance of less than 0.5%, if indeed speciation occurred during the Illinoian or Wisconsinian. However, the average divergence was 5.1 %, a
tenfold greater value than expected. Therefore we concluded that most of the 35
speciation events occurred much earlier than the Illinoian, some even in the Late
Pliocene. Two other lines of evidence supported our rejection of the LPO model.
The distribution of mtDNA distances between 13 additional songbird sister-species
pairs, not previously hypothesized to fit the LPO model, did not differ significantly
from that of the 35 pairs, indicating that what had been considered extremely re-