Here's a piece of his paper reproduced WITHOUT permission (hope he doesn't mind)
Phylogeography
All morphological and molecular analyses performed in this study support a western
origin of L. getula in the eastern U.S. The evolutionary history of eastern L. getula
appears to be linked to the geological history of the Coastal Plains derived from ancient
coastlines. Fossil remains indicate that ancestral western L. getula utilized a Gulf Coast
corridor to disperse east into Florida during low sea levels in the Pliocene between 2-5
million years ago (Mya) (Auffenburg, 1963). Estimates of mtDNA sequence divergence
for reptile species range from 0.47 to 1.32% per million years (Zamudio and Greene,
1997). The smallest percentage sequence divergence between the eastern and western
clades of L. getula is 5.56%, which translates into 4.2-11.8 Mya, a time frame that
overlaps the sequence of events hypothesized by Auffenburg (1963). Although some
midwestern samples were much closer geographically to eastern samples, they are much
less closely related relative to the range of divergences within the eastern clade. A rapid
range expansion from southern refugia into northern areas as glaciers retreated could
account for the high haplotype diversity and low genetic distances (maximum d = 0.91%)
in the east in comparison with other geographic areas. Interestingly, additional snake
taxa with similar distributions in the midwestern and eastern U.S. have been reported to
exhibit similar phylogeographical patterns (Rodriguez- Robles and Jesus- Escobar, 1999b,
1999a; Burbrink et al., 2000; Burbrink, 2001). Thus, the evolution of the three identified
123
subclades within the eastern clade is probably related to more recent glacial events during
the Pleistocene (10,000-2 Mya).
The lowest sequence divergence between the eastern and midwestern clades is
2.18%, which translates into 1.65-4.6 Mya (early Pleistocene to early Pliocene). There
were four major glacial events during the Pleistocene, with the first being the greatest
followed by less intense episodes (Webb, 1990; Brown and Lomolino, 1998). During
these glacial events sea level was >100 meters (m) lower than present-day (Lidz and
Shinn, 1991), Florida was nearly twice its present size and land was continuous from the
present Florida mainland to the Dry Tortugas (MacNeil, 1950; Hoffmeister and Multer,
1968). In order to escape from cooler climates and ice masses, extreme northeastern L.
getula populations must have been extirpated or pushed southward, whereas peninsular
and panhandle populations experienced an expansion following suitable habitat. These
conditions of displacement and expansion would have promoted mixing among SE U.S.
populations. This is evident because of their close genetic relationships as well as certain
shared morphological characters such as ontogenetic interband lightening found in
Florida and as far north as the Outer Banks of North Carolina (Barbour and Engles, 1942;
Lazell and Musick, 1973; Blaney, 1977). However, a less likely but alternative
hypothesis is that the character of ontogenetic interband lightening evolved multiple
times along the Coastal Plains. During major interglacial events much of present-day
Florida was inundated except for higher elevations in the peninsula and panhandle.
These areas of higher elevations have been identified as ancient shorelines in the
peninsula (Jackson, 1973; Webb, 1990; Clark et al., 1999), as well as ancient barrier
islands in the eastern Apalachicola Lowlands (Brenneman, 1957; Brenneman and Tanner,
124
1958). During these times, Atlantic populations were capable of expanding their range
northward following warmer climate, but peninsular and panhandle populations were
geographically isolated. Peninsular populations were probably found in mesic habitats
extending into the southern peninsula (Watts and Hansen, 1988). Panhandle populations
were probably isolated on barrier islands in the eastern Apalachicola Lowlands
(Brenneman, 1957), which correlate well with a large number of endemic and relict
plants and animals there (Table 3-10; James, 1961; Clewell, 1977; Yerger, 1977; Ward,
1979; Judd, 1982; Gilbert, 1987; Coile, 1996; Chafin, 2000; Chaplin et al., 2000). Thus,
the distributions of the Apalachicola Lowlands endemic plants and animals possibly were
a result of the same geological event. Since the last glacial maximum around 18,000
years before present (BP), North American glaciers have melted and retreated northward,
sea level has risen steadily (Randazzo and Halley, 1997). Again, populations of eastern
L. getula will eventually become isolated in the peninsula and panhandle, while others
expand their range northward.
Taxonomy
Under the Apomorphic Species Concept (ASC), species are considered to be wellsupported
minimal monophyletic groups. In all morphological and molecular analyses in
this study, the eastern populations of L. getula represent the smallest well-supported
monophyletic group, which suggests recognition as a separate species. Recently, users of
the various Phylogenetic and Evolutionary Species Concepts (PSC and ESC respectively)
have objected to the subspecies concept because they claim it is not operational within
their definitions of reciprocal monophyly (Mayr and Ashlock, 1991; Baum, 1992;
Burbrink et al., 2000). Additionally, beginning with Wilson and Brown (1953) the
125
subspecies concept has been criticized because subspecies were commonly identified on
too few and arbitrary delimited characters. And in many cases where several characters
were used, each character varied independently because of differing locally adaptive
pressures, which resulted in different subspecies distributions and arbitrary cline slices
(Frost and Hillis, 1990). Nonetheless, under the ASC one may recognize subspecies as
subclades within species, where subclades have weaker support because of occasional
gene flow or recent evolutionary divergence (Mishler and Theriot, 2000). Mayr (1969)
and Smith et al. (1997) also recognize subspecies as geographic races with relatively
homogenous phenotypic patterns, which have the ability to exchange genes with adjacent
races. Areas where subspecies come into genetic contact are identified as intergradation
zones (Mayr, 1969; Smith et al., 1997). Because gene flow will obscure boundaries of
subspecies and prevent them from attaining reciprocal monophyly at the mtDNA level,
this criterion should not invalidate subspecies recognition (Patton and Smith, 1994;
Rodriguez-Robles et al., 1999c).
The traditional subspecies recognized in L. getula in the eastern U.S. correspond
closely to the phylogenetic patterns that I have uncovered. After examination of
hundreds more specimens than any previous researcher for a study of geographic
variation using morphology (Chapter 3), I identified three subspecies in Florida: 1) L. g.
floridana from central and southern peninsular Florida, 2) L. g. getula from NW
peninsular Florida north to southern New Jersey, and 3) unnamed populations in the
eastern Apalachicola Lowlands. In the morphological and molecular analyses, the three
consistently identified subclades correspond exceptionally well to the three identified
geographic races (Chapter 3). Because each of these subspecies can be diagnosed by at
126
least one synapomorphy, this evidence suggests that they represent evolutionary
processes rather than ecotypic variants.
Individuals from the western Apalachicola Lowlands were once believed to
represent a distinct subspecies, L. g. goini (Neill and Allen, 1949). However, Blaney
(1977) invalidated L. g. goini by speculating that it represented a Pleistocene intergrade
between panhandle L. g. getula and now disjunct peninsular L. g. floridana.
Additionally, Means (1977) believed that L. g. goini indeed represented an intergrade
population, but L. g. goini was intermediate between unnamed populations in the eastern
Apalachicola Lowlands and L. g. getula that surrounds the region. The most consistent
subclade with statistical support in morphological and molecular analyses consisted
primarily of eastern Apalachicola Lowlands and panhandle individuals. Additionally,
there are five synapomorphies supporting the monophyly of the eastern Apalachicola
Lowlands populations. Morphological data examined on a much finer geographic scale
(Chapter 3) revealed that the eastern Apalachicola Lowlands populations possess
relatively homogeneous color patterns that grade into populations in surrounding areas
(i.e., gene flow) consisting of L. g. getula. Because L. g. goini possesses intermediate
characters between eastern Apalachicola Lowlands snakes and L. g. getula, I relegate L.
g. goini to intergrade status. Therefore, because L. g. getula and L. g. floridana are
already recognized subspecies, the eastern Apalachicola Lowlands populations deserve
equivalent taxonomic recognition as well. As systematists it is our responsibility to
document biological diversity, and it is essential to recognize both species and subspecies
otherwise everyone worldwide will fail to notice biological diversity without names
127
(Dobzhansky, 1970; Smith et al., 1997). A taxonomic amendment regarding the eastern
Apalachicola Lowlands populations is in progress and will be treated in a separate paper.
I was unable to find evidence supporting the recognition of both L. g. brooksi and L.
g. sticticeps using morphology and mtDNA. Individuals from the extreme southern
Florida peninsula were once believed to represent a distinct subspecies, L. g. brooksi
(Barbour, 1919), yet its validity quickly came into question (Blanchard, 1920, 1921;
Wright, 1935). Becasue only 15 individuals from the Florida peninsula were sequenced
for mtDNA, this small sample size might have been insuffiecient to find genetic
partitioning within the peninsula. Nonetheless, Duellman and Schwartz (1958), Blaney
(1977), Krysko (1995), and Krysko (this study) could not find morphological characters
for diagnosing L. g. brooksi, and the mtDNA data indicate that the Florida peninsula is a
panmictic population (Tables 5-4, 5-5). Barbour and Engles (1942) described L. g.
sticticeps from the Outer Banks of North Carolina based on head morphology and
ontogenetic interband lightening of the dorsal pattern (Lazell and Musick, 1973).
However, Hillestad et al. (1975), Blaney (1977), Gibbons and Coker (1978) and Palmer
and Braswell (1995) did not accept that a distinct geographic race existed on the Outer
Banks. My morphological and molecular data further corroborate these authors, as I
found snakes from the Outer Banks with mtDNA haplotypes identical to adjacent
mainland snakes as well as those as far away as SW Georgia.