MOLECULAR PHYLOGENETICS AND EVOLUTION
Vol. 8, No. 1, August, pp. 104–113, 1997
ARTICLE NO. FY960392
Phylogeny of the Side-Blotched Lizards (Phrynosomatidae: Uta)
Based on mtDNA Sequences: Support for a Midpeninsular
Seaway in Baja California
Darlene E. Upton and Robert W. Murphy
Center For Biodiversity and Conservation Biology, Royal Ontario Museum, 100 Queen’s Park, Toronto, Ontario, Canada M5S 2C6
Received March 14, 1996; revised September 16, 1996
We investigated the phylogenetic relationships of
western North American side-blotched lizards, genus
Uta, using mitochondrial DNA partial gene sequences
from cytochrome b and ATPase 6. Uta stejnegeri appeared basal in our tree followed by U. palmeri. Uta
stansburiana from the islands of Angel de la Guarda,
Mejia, and Raza formed the next clade, followed by U.
antiqua and other populations of U. stansburiana.
These relationships suggest that the populations on
the Angel de la Guarda island block should be recognized as a distinct species. Remaining populations of
U. stansburiana formed two clades, corresponding to
the northern and southern Baja California peninsula.
Uta stellata and U. squamata occurred within the
northern and southern clades, respectively. The discontinuity requires a long-lasting isolation, the only reasonable explanation being a former midpeninsular
seaway. Correlations between our cladogram and magnetic anomalies in the Gulf of California date formation of the seaway at 1 million years ago. r 1997 Academic
Press
INTRODUCTION
The geographic distribution of organisms has long
intrigued biologists. Among other variables, species’
distributions result from the complex interactions of
dispersal abilities, paleoecology, and geologic events
such as plate tectonics. Simultaneous with the development of plate tectonics theory, the field of phylogenetic
systematics has advanced testable biogeographic hypotheses (Brooks and McLennan, 1991). And now, we
may be in the midst of another methodological progression, the application of DNA sequence data to assessing
paleostratigraphy and the resulting effects on biogeography and speciation.
Uta, side-blotched lizards, occur in much of western
North America, from central Washington to the tip of
Baja California, eastward into western Colorado and
west Texas, and southward into Zacatecas and Sinaloa,
Mexico. They are found in various biomes from below
1055-7903/97 $25.00
Copyright r 1997 by Academic Press
All rights of reproduction in any form reserved.
sea level (Death Valley, CA) to around 2700 m (Stebbins, 1985). The study of these lizards is particularly
important because of their relevance to understanding
insular colonization and extinction, genetic bottlenecks, ecology, and other key evolutionary issues (Soulé
and Sloan, 1966; Ballinger and Tinkle, 1972; Hews,
1990). Yet, with all their relevance, a defensible phylogeny for the genus is wanting, rendering evolutionary
explanations speculative. Consequently, using mitochondrial DNA gene sequences, we investigated the
genealogy of Uta.
MATERIALS AND METHODS
Tissue samples were accumulated through our own
field work and donations. Collecting localities are given
and depicted in Fig. 2. Most tissue samples from
insular populations of Uta were donated by Michael
Soulé and represent remaining material of Soulé and
Yang (1973). The remains are maintained in the frozen
tissue collection of the Royal Ontario Museum. Specific
frozen tissue numbers are available on request, as are
the catalog numbers and locations of preserved vouchers of other tissue samples.
We sequenced 555 base pairs from cytochrome b and
335 bp from ATPase 6. DNA sequencing protocols follow
Murphy et al. (1995) with one exception: rather than
ethanol precipitation, in some samples the DNA was
left in the lysis buffer after removal of PCI with a CI
wash. The DNA was then amplified directly with no
further dilutions. Amplification and sequencing of a
560-bp region of cytochrome b was performed using
four primers flanking the region. Universal cytochrome
b primers were used to sequence the first 307-bp region.
The primers used follow: L14841 58 cca tcc aac atc tca
gca tga tga aa 38, H15149 58 gcc cct cag aat gat att tgt
cct ca 38. The primers are named by their occurrence on
the heavy or light strand and their 38 position in the
Xenopus mitochondrial DNA sequence (Roe et al., 1985).
Both of these primers have been found to work well
with a variety of vertebrates (Kocher et al., 1989). The
remaining bases were amplified and sequenced with
104
105
PHYLOGENY OF SIDE-BLOTCHED LIZARDS
primers originally designed by Birt et al. (1992) to work
with bird species. The sequences L15063 58 gga cga ggc
ttt tac tac ggc tc 38 and H15506 58 ttg ctg ggg tga agt ttt
ctg gtc 38 amplify a 442-bp piece that overlaps the first
region by 80 bp at its 58 end. Amplification and sequencing of a 335-bp region of ATPase 6 mtDNA was performed using two primers. The primers designed by
Haddrath (personal communication) follow: L10031 58
atg aac cta agc ttc ttc gac caa tt 38, and H10437 58 ata
aaa agg cta att gtt tcg at 38.
The sequences were exported from ESEE200 (Cabot
and Beckenbach, 1989) as ASCII text and initially
evaluated using MEGA (Kumar et al., 1993), and
MacClade (Ver. 3.0.4; Maddison and Maddison, 1992).
Phylogenetically informative sites were maintained for
analyses with PAUP (Ver. 3.1.1; Swofford, 1993), Hennig86 (Farris, 1988), and Random Cladistics Ver. 2.1
(Siddall, 1994).
Hillis et al. (1993) suggested that prior to in-depth
analyses, the presence of phylogenetic signal (character
correlation) should be tested. One method to test for
character correlation is the permutation tail probability test (PTP; Faith and Cranston, 1991). Each character state was randomly reallocated to different populations using Random Cladistics Ver. 2.1 (Siddall, 1994).
Nine-hundred ninety-nine randomizations were performed without randomization of the outgroup states. A
second method for analyzing character correlation in
the data set is the g1 statistic (Hillis, 1991; Huelsenbeck, 1991; Hillis and Huelsenbeck, 1992). One thousand randomly chosen trees were used in PAUP to
generate the g1 statistic tested for significance against
the critical value.
In our analyses, all multistate characters were evaluated as nonadditive because there is no a priori reason
to assume order of evolutionary change between nucleotide bases adenine (a), cytosine (c), guanine (g), or
thymine (t) (Swofford et al., 1996). The phylogenetic
analyses with PAUP used a heuristic search, with
random addition sequence, 10 replicates, retaining
minimal trees only, using tree bisection–reconnection
branch swapping with steepest descent and collapsing
zero length branches. Similar analyses were performed
with Hennig86 using the branch swapping command
‘‘mh* bb*,’’ which retains all most parsimonious trees
(MPTs). Strict consensus trees from PAUP were imported into MacClade for further analysis.
Nodal support was assessed for each gene individually and combined. Sankoff matrices (Sankoff and
Rousseau, 1975) were used for differentially weighting
transitions and transversions. Successive approximations (Farris, 1969) was used as an alternative weighting approach. Functional ingroup–outgroup analysis
(FIG/FOG; Watrous and Wheeler, 1981; Fu and Murphy, 1997) was applied. Resulting topologies were examined for congruence with the unweighted analysis. Two
final methods were used on the combined data set only.
These included decay analysis (Bremer, 1988) and
bootstrap (BS; Felsenstein, 1985). Decay analysis
(Bremer, 1988) was performed in PAUP by saving trees
up to two steps longer than the most parsimonious
solution. BS proportions (BSP) using 1000 replicates
were calculated using Random Cladistics Ver. 2.1 (Siddall, 1994) which interfaces with Hennig86 (Farris,
1988). We used the ‘‘mh bb*’’ command in Hennig86
that generates multiple trees and subsequently BS
them to find MPTs. Resulting BSPs were compared to
the preferred tree to assess nodal support.
A preferred tree was chosen and used for subsequent
discussion. The criteria used for tree selection included
a topology based on the combined gene segments,
parsimony, and a high degree of nodal stability. One
goal of this study was to assure that the phylogeny of
the species of Uta was directly reflected in the taxonomy (Farris, 1980), at least at the species level; we
were willing to accept a non-monophyletic taxonomy
below the species level for ease of information retrieval.
It was also highly desirable to have a taxonomy consistent with other reptilian populations at a given hierarchical level. Both Grismer et al. (1994) and Murphy et
al. (1995) used taxonomic hierarchical level as an
objective guideline for not applying species recognition
to Cedros Island red diamond rattlesnakes, Crotalus
ruber—a population on a Pleistocene landbridge island.
Such guidelines remove arbitrary decisions about levels of morphological divergence justifying species designation and provide a uniformity of taxonomic level for
biogeographic inference.
RESULTS
Results of the two single gene analyses showed great
consistency in the formation of three major clades: an
Isla Angel de la Guarda clade, a clade associated with
the northern Baja California peninsula, and one allied
with the southern peninsula.
Our combined gene segments resulted in the analysis
of 890 nucleotides. Of these, 315 were variable, with
191 (21.5%) potentially phylogenetically informative.
The majority of transformations occurred in the third
position (156, or 82%). First and second position transformations were 33 (17%) and 2 (1%), respectively. The
transition to transversion ratio was calculated across
all MPTs and found to be 2.9:1.
PTP analysis with 999 randomizations resulted in
trees ranging in length from 834 steps to 806 steps
(mode 5 819, median 5 817, P value 5 0.001). These
trees were 219 or more steps longer than the MPTs.
Analysis of skewness with 1000 random trees resulted
in g1 5 20.68. This value reflects the strong left skew of
the data set and is less than the critical g1 of 20.09 at
P , 0.01, suggesting the data have significant character correlation.
Analysis of the unweighted data resulted in 7 MPTs,
106
UPTON AND MURPHY
FIG. 1. Strict consensus tree of cytochrome b and ATPase 6
combined data analysis (7 trees, CI 5 0.447, RI 5 0.572). Bootstrap
values are present on those branches supported greater than 85% of
the time.
with lengths of 584 steps (CI 5 0.447, RI 5 0.572). A
strict consensus tree (Fig. 1) depicts Uta stejnegeri
(population 9; Fig. 2) as basal, followed by U. palmeri
(8), then a clade consisting of populations of U. stansburiana from Islas Mejia (3), Angel de la Guarda (4),
and Raza (5). Next, U. antiqua (6) formed a trichotomy
with two clades consisting of the remaining sample
sites, a northern and a southern clade. Five of the seven
cladograms placed U. antiqua basal to the peninsular
clades, whereas in two topologies it was basal to the
northern clade. Relationships among the southern populations were resolved in all trees with U. squamata (13)
located basal in the clade. The population of U. stansburiana from Isla San Esteban (7) was basal in the
northern clade. The remaining northern populations
were essentially unresolved with the exception of the
U. stellata (16)–Isla Cedros (17) clade and the Arroyo de
San Regis (15)– Bajia de los Angeles (14) sister group.
Examination of the MPTs consistently shows paraphyly in Uta stansburiana as currently defined: U. squamata, U. stellata, and U. antiqua branch off from
within U. stansburiana. Because of this finding, a
constraint tree was designed to force monophyly of
U. stansburiana. This analysis produced three trees
with lengths of 646 steps. For U. stansburiana to be
monophyletic an additional 62 steps were required.
Our analysis weighting transversions from 2 to 4
times greater than transitions resulted in three MPTs
with configurations identical to three trees in the
unweighted analysis and depicting the placement of
Uta antiqua (6) basal to the northern and southern
clades. Further, our successive approximations analyses produced one tree identical to one of the weighted
analyses trees. FIG/FOG was completed using four
separate FOGs. In all cases, relationships were congruent to the unweighted analysis and no further resolution was achieved by this method.
Decay analysis was used to examine trees up to two
steps longer than the global MPTs with a strict consensus cladogram produced for all suites of trees found.
Fifty-three trees were found with a length of 585 steps
or less, one step longer than the MPTs. The strict
consensus collapsed the three major clades and Uta
antiqua (6), U. palmeri (8), and U. stejnegeri (9) into an
unresolved polytomy. Relationships within the other
clades were identical to the unweighted analysis. An
analysis saving all trees two steps longer than the
MPTs and less resulted in 202 trees. The strict consensus tree was identical to the previous except that U.
squamata (13) and the Danzante Island population (11)
collapsed into a polytomy with the remaining southern
clade members.
Finally, BSPs were calculated for the unweighted
strict consensus tree. Nodes having a value of 85% or
higher are shown in Fig. 1. Low BSPs may result from a
small number of unambiguous synapomorphies and
therefore do not indicate a lack of confidence in the node
(Felsenstein, 1985). The strongest support was found
for the Angel de la Guarda clade (BSP 5 1.00), the
sister group relationships between Uta stellata (16) and
Isla Cedros (17; BSP 5 1.00), and the following clade:
(San Pedro la Presa [21] (Isla Monserrate [12] (Isla San
Marcos [10] (San Lucas [20])))) BSP 5 1.00. The
southern clade had a BSP 5 0.96, while the northern
population clade had BSP 5 0.85.
DISCUSSION
Our phylogenetic analyses resulted in the resolution
of multiple most parsimonious trees for side-blotched
lizards. Two assumptions are inherent in this discussion. First, we assume that the character correlation in
our dataset reflects genealogical association. Our PTP
analysis (Faith and Cranston, 1991) and the g1 statistic (Hillis, 1991; Huelsenbeck, 1991; Hillis and Huelsenbeck, 1992) found the patterns in our data to be
significantly different from random. Second, we assume
that the gene trees reflect species phylogeny. We acknowledge that below the species level gene flow may
prevent historical relationships from being identified.
Although the phylogenetic analysis of mtDNA yields a
genealogy of the maternal lineage, this genealogy should
reflect the species phylogeny (Moore, 1995), especially if
few sex-specific differences in dispersal occur (Villablanca,
1994). The life history of these lizards suggests little
sex-specific difference in dispersal (Tinkle, 1967). However, a conclusion about the congruence of the maternal
PHYLOGENY OF SIDE-BLOTCHED LIZARDS
107
FIG. 2. The preferred cladogram. Numbers on the terminal branches of the cladogram correspond to map locations. Vertical bars indicate
the membership in the northern and southern clades. The inferred location of the hypothesized Pleistocene midpeninsular seaway is indicated
by a bar across the peninsula.
genealogy and phylogeny is best made after a nuclear
gene evaluation is completed.
The preferred tree topology (Fig. 2) reflects the
weighted and FIG/FOG analysis. This tree, based upon
analysis of both gene segments, reflects the maternally
inherited mitochondrial genome using the homologous
base pairs as characters. Although our decay and BS
analyses strongly support the formation of the three
major clades, their relationships relative to each other
remain ambiguous.
Taxonomic Implications
If our preferred gene tree (Fig. 2) represents the
phylogenetic relationships of Uta, then a number of
nomenclatural changes will be required in order to
obtain a taxonomy that reflects historical relationships.
Although gene flow among the peninsular populations
may be occurring without interruption, it seems highly
unlikely that any significant gene flow will be occurring
among allopatric populations, including those on islands. Therefore, nomenclatural problems relating to
allopatric populations likely reflect required changes,
whereas those pertinent to continuously distributed
populations may not.
Ballinger and Tinkle (1972), noting the lack of osteological and myological characters, proposed a phylog-
eny for Uta based primarily on morphometrics and
color patterns. Many of our samples represent taxa and
localities considered by Ballinger and Tinkle (1972).
For shared localities, when our data were mapped onto
their phylogeny an additional 56 steps were required,
raising the minimal tree length from 584 to 640 steps.
The relatively basal position of U. palmeri (8) was the
only similarity shared by both analyses.
As in Ballinger and Tinkle’s (1972) phylogeny, paraphyly of U. stansburiana is evident in our analysis. A
constraint tree designed to maintain monophyly of
U. stansburiana required 62 extra steps. Thus, U.
stansburiana is a paraphyletic species either arising
independently in several clades or having new species
arising from within clades. Uta stansburiana in the
Angel de la Guarda clade (3–5) has an origin between
U. antiqua (6) and U. palmeri (8). Uta stellata (16)
branches off as a terminal taxon within the northern
peninsular clade, and U. squamata (13) originated in
the southern clade of U. stansburiana. Although two
distinct clades represent peninsular U. stansburiana,
this distinctiveness may not be seen in a phylogeny
based on nuclear genes due to continued gene flow
along the peninsula.
One of the primary goals of phylogenetic systematics
is to produce taxonomies that directly reflect genealogi-
108
UPTON AND MURPHY
cal relationships. The populations of Uta occurring on
the Angel de la Guarda Island group comprise a distinct
clade (Fig. 2), one long recognized as a phenetically
identifiable group (Soulé, 1967). Consequently, either
the Angel de la Guarda group (3–5) must be recognized
as a separate species to maintain monophyly or U.
antiqua (6), U. squamata (13), and U. stellata (16) must
all be synonomyzed with U. stansburiana. Considering
the location of the Angel de la Guarda Uta on the
cladogram, their morphological distinctness (Soulé,
1966, 1967), and allopatric isolation, we believe they
represent an undescribed species. They were likely
isolated on the islands about 1 million years ago
(Murphy, 1983b) when plate tectonic movements broke
this island group from the peninsula (Moore, 1973).
Recognition of these insular lizards as a species is
further supported by levels of taxonomic recognition of
other reptilian species on the Isla Angel de la Guarda
block. Soulé and Sloan (1967), Murphy (1983c), and
Murphy and Ottley (1984) looked at the distribution of
species in the Gulf of California and found high levels of
endeminism at both the subspecific and specific level on
the Angel de la Guarda Island group. Although recognition of this island group as a distinct species solves one
problem, it still leaves U. stansburiana as a paraphyletic taxon owing to U. stellata and U. squamata.
Recognition of the Islas San Benito side-blotched
lizards as a species, U. stellata (16), also results in a
paraphyletic taxonomy for U. stansburiana. Uta stellata is unambiguously resolved as the sister taxon to
the Isla Cedros (17) population of U. stansburiana, the
geographically nearest region. Isla Cedros, a landbridge island, has no endemic species. The populations
of Uta on the Cedros and San Benito islands have
nearly identical sequences, suggesting that the colonization of the oceanic San Benito Islands likely occurred
after all other landbridge islands were formed. Consequently, if the San Benito population is recognized as a
species, U. stellata, then all other insular populations
having an earlier origin must also be named new
species in order to maintain a taxonomy that reflects
history. Given this situation, and the relatively small
morphological divergence, it is necessary to synonymze
U. stellata with U. stansburiana, hereafter referred to
as U. stansburiana stellata.
Uta squamata from Isla Santa Catalina (13) is resolved as the sister taxon of the southern peninsular
clade. Its position within the cladogram also leaves U.
stansburiana as a paraphyletic taxon. However, in this
case, because of the likelihood of continued gene flow
among peninsular populations, which would effectively
place U. squamata as the next taxon branching off after
U. antiqua, and before the northern and southern
peninsular clades, we recommend the continued recognition of this species pending an evaluation based on
nuclear genes. The alternative option is to consider the
southern clade of U. stansburiana as a species (U.
elegans).
Murphy (1983a, 1983b) considered the population of
Uta on Isla Monserrate (12) as a distinct, unnamed
species based on allozyme divergence. The analysis of
our mtDNA sequence data does not support his conclusion. Either we have observed a case of extremely rapid
allozyme differentiation which obscures phylogenetic
relatedness or we detected a relatively recent overwater colonization of Monserrate from another source and
have not found the mitochondrial genome of the founding population. Until additional data and samples
become available, we cannot differentiate between these
two scenarios and recommend that the population not
be formally named pending further study.
Three of the four remaining species of Uta are
restricted to landbridge islands. The desirability of
naming landbridge island populations as species can be
addressed by looking at both the taxonomic levels
accorded to other reptilian species on landbridge islands and the genealogical relationships of populations
in our study. In the former case, Murphy and Ottley
(1984) noted only a single case of a Gulf of California
reptile on a landbridge island being accorded species
status, and this species, Sauromalus slevini, also occurrs on an oceanic Gulf island. Recently, Grismer
(1988), Grismer et al. (1994), Grismer and Mellink
(1994), and Murphy et al. (1995) sank the only three
Pacific Baja California landbridge island species, all of
which occur on Isla Cedros (17). Plainly, history has
borne out the undesirability of naming recently isolated landbridge populations of reptiles as species. (The
fourth remaining species, U. nolascensis, occurs on an
oceanic island, Isla San Pedro Nolasco, off the coast of
Sonora, Mexico; tissue samples were not available for
evaluation.)
In our cladogram, peninsular populations frequently
appear to be more closely related to nearby insular
populations than they are to contiguous peninsular
localities: note the associations of peninsular San Lucas (20) with Isla San Marcos (10) and the northern
peninsular population (e.g, 19) with Isla Cedros (17;
Fig. 2). In addition, although not shown, our cytochrome b sequence data associated San Pedro la Presa
(21) with Isla Espiritu Santo. These genealogical relationships were predicted by Soulé (1967) based on a
phenetic analysis of 36 morphological characters.
Clearly, some peninsular populations of U. stansburiana appear to be more closely related to island population than to one another. Thus, if the island populations
are given species status, a paraphyletic taxonomy
results. Similar results have been reported by Patton
and Smith (1994), but in their case, the paraphyly is
attributable to hybridization. Further, Theriot (1991)
discussed the use of autapomorphies to define lineages,
even if this results in a paraphyletic taxonomy. This
PHYLOGENY OF SIDE-BLOTCHED LIZARDS
FIG. 3. Illustration of problems of paraphyly. (a) Four peripheral
populations. (b) Four peripherals where source populations are
becoming homogenous over time. (c) Phylogeny representing the
sequence of peripheral isolation. (d) Phylogeny representing recoverable relationships in a, b.
issue has been discussed in detail by Graybeal (1995).
Nevertheless, in the case of Uta, we can also infer the
successiveness of isolation of insular populations by
Gulf channel depth, the most shallow being the most
recently isolated. Consequently, we have independent
evidence of genealogical history, even if nearby populations are resolved as sisters of each other.
Figure 3 further illustrates the problem of paraphyly.
Figure 3a represents peripheral isolate populations
ranging from (A) recently completely separated from
the mainland (M), to (D) remaining connected via a
large landbridge connection. As time passes and sea
levels rise, each island population will sequentially
become completely isolated, first A, then B, C, and
finally D (Figs. 3b and 3c). If D is diagnosable morphologically, and A–C are not, then D cannot be named
without creating a paraphyletic taxonomy because the
taxonomy will not reflect genealogical history. This is
the critical issue. Further, because some peninsular
populations are most closely related to nearby insular
populations, M becomes the sister population to all four
insular isolates, initially resulting in a paraphyletic
taxonomy (Fig. 3d). In time, the source areas for the
islands will become more closely related to one another
than to the island isolates, as evidenced by Uta on the
Angel de la Guarda block (4). Once this occurs, the
phylogeny for landbridge islands will appear as an
109
unresolved polytomy (Murphy and Ottley, 1984). At
such a time, species status may be given unquestionably to the peripheral isolates. However, isolates A–C
must be named to produce a monophyletic taxonomy. If
all populations are not named, then the resulting
taxonomy will be paraphyletic.
The naming of these landbridge populations as species could be viewed as being equivalent to naming
every introduced allopatric population as a species
when some diagnosable attribute is observed, whether
or not the attribute has a heritable or environmental
basis. For example, any Eurasian species introduced to
the Americas could be named as a new species if, due to
a different temperature of egg incubation, pH, or salinity, they were morphologically distinctive because of a
developmental response to a different environment.
Nothing would be gained by naming these populations,
and much would be lost, namely historical relationships. Regarding insular populations associated with
the Baja California peninsula, the varying ocean depths
in the Gulf of California are not trivial characteristics
of biogeography but rather reliable indicators of age of
isolation. They very well may explain the seemingly
super-saturated reptile fauna on post-Pleistocene landbridge islands (Wilcox, 1978) and hence the sequence of
isolation events.
Grismer (1994a) recently described three new species, Uta lowei, Uta encantadae, and Uta tumidarostra,
from the Islas Las Encantadas Archipelago, the northern-most islands associated with the peninsula of Baja
California. This archipelago had a landbridge connection with the peninsula, as did many other islands
(Gastil et al., 1983). If Grismer’s taxonomy is accepted,
then all of the other landbridge island populations
separated by an equal or greater ocean depth (including
Islas Danzante (11), Carmen, San Jose, San Francisco,
and Cedros (17); see Wilcox, 1978) must also be named
as species in order to maintain a monophyletic classification. Grismer’s species are anatomically diagnosed
by a single, highly adaptive functional complex (enlarged salt glands and associated osteological and
squamosal modifications), modification of the postzygapophyseal process of the atlas, and color pattern variation. Following the evolutionary species concept (Simpson, 1961; Wiley, 1978) as elucidated by Frost and Hillis
(1990), Grismer (1994a) recognized these species because they were allopatric in distribution and possessed discrete character states. However, Grismer
(1994a) failed to adhere to the criterion of maintaining
monophyly in the taxonomy (Frost and Hillis, 1990).
Rather, regarding paraphyly, he stated ‘‘until character
evidence (synapomorphies) is presented indicating
which of the various populations of U. stansburiana are
more closely related to particular insular species (i.e.
demonstrable paraphyly . . .), changes in the nomenclature . . . should not be considered.’’ Our data clearly
indicate paraphyly and the close association of land-
110
UPTON AND MURPHY
bridge island populations with nearby peninsular populations. Although we have no tissue samples for evaluation, given the cladogenic patterns observed among our
samples, the very recent age of the landbridge islands
(#12,000 years), that these islands have been formed
multiple times during the Pleistocene, that no other
reptiles occurring exclusively on landbridge islands
have been accorded species status, and the evident
paraphyletic taxonomy resulting from recognition of
these three species, we consider Grismer’s (1994a)
species as subspecies: U. stansburiana lowei, U. stansburiana encantadae, and U. stansburiana tumidarostra. In doing so, we recognize paraphyletic relationships at the subspecies level and accept this for the
efficiency of information retrieval about the highly
adaptive nasal modifications in a Linnean taxonomic
system.
Paleobiogeography
We found a notable midpeninsular break in gene
sequence continuity within U. stansburiana (Fig. 2).
The break was resolved in separate analyses of each
gene and in the combined gene segments. Among the
893 base pairs sequenced, the absolute differences (AD)
between sequences of northern landbridge islands and
northern peninsular sites yielded a mean AD of 44.5
(68.2; N 5 35). Similarly, southern landbridge islands
compared to southern peninsular sites yielded an average AD of 43.0 (619, N 5 15). In contrast, ADs of
northern versus southern peninsular sites were almost
double these values (mean AD 5 84.0). As supported by
parsimony, distance, and BS analyses, this break was
not a spurious observation, but rather highly significant.
The midpeninsular mtDNA discontinuity requires a
long-lasting isolation of the northern and southern
lizard populations. Two possibilities may explain this
midpeninsular break. First, an ecological change, such
as one facilitated by climate, may have prevented lizard
contact in the midpeninsular region (Savage, 1960).
Climatic changes associated with Pleistocene glaciation have unquestionably resulted in far more mesic
transpeninsular habitats than occur today (Betancourt
et al., 1990; Grismer and McGuire, 1993). However,
because U. stansburiana occurs in a variety of habitats,
a climatic scenario seems improbable.
A persistent physiographic barrier is the only alternative explanation for a midpeninsular disruption of
distribution. Further, a temporary seaway is the only
practicable explanation for this disruption. While inconclusive, limited stratigraphic data from marine bed
sediments found inland at Santa Rosalia (18 km north
of San Lucas, 20; Fig. 2) support the concept of Pleistocene midpeninsular flooding (Anderson, 1950;
Durham and Allison, 1960). Such a seaway has long
been hypothesized, but based solely on low midpeninsular elevations and without the existence of corroborative data (Nelsen, 1921; Johnson, 1924; Beal, 1948).
Strong support for the midpeninsular seaway is also
derived from the distribution of marine organisms.
Forty-two species of marine fishes and several invertebrates common to the Gulf of California and the Pacific
Ocean do not have continuous distributions around the
southern tip of the peninsula (Present, 1987). Most of
these occur in temperate waters and not in tropical
domains. Such disjunct distributions have been explained by Pleistocene changes in ocean isotherms.
However, significant differences in allozyme frequencies among fishes (Present, 1987; Crabtree, 1983; Orton
and Buth, 1984), and presumably other marine organisms, likely would not have accrued if the once contiguous populations were disrupted only 12,000 years ago
when marine isotherms shifted southward.
When did the seaway exist? Our phylogenetic analysis of the combined genes resolved seven most parsimonious trees. The multiple hypotheses showed alternative explanations in two regions of the tree, near
terminal branches in the northern clade, and in the
position of U. antiqua. This species occurs on the
midriff islands of Salsipuedes, San Lorenzo Norte, and
San Lorenzo Sur (6), which were linked together during
the Pleistocene. Uta antiqua occurs on the cladogram
either at the base of the northern clade or at the base of
the northern and southern clades together, (Fig. 2). The
branch lengths separating the major clades in this
region of the cladogram are very short (Fig. 2), suggesting the occurrence of almost simultaneous speciation
events. Magnetic anomalies associated with the Delfin
basin suggest that the Gulf midriff islands broke away
from the peninsula about 1 million years ago (Ma) as a
consequence of sea floor spreading (Moore, 1973). Assuming that U. antiqua was isolated on the midriff
islands as they were formed (Murphy, 1983b), and
given its position on the cladogram relative to the
peninsular north–south break, we date the formation of
the midpeninsular seaway at about 1 Ma.
The mtDNA data also reveal much about overwater
island colonization by the lizards, and thus they are
also critical to studies of island biogeography (Murphy,
1983b; Case, 1975, 1983). Isla Raza (5), a tiny (0.68
km2 ), Holocene volcanic island, is located 9.6 km north
of Isla Salsipuedes, the home of U. antiqua (6). It is also
8.3 km south of small Isla Partida and 21.5 km south of
large Isla Angel de la Guarda (4). Our mtDNA data
unquestionably demonstrate that the colonization of
Isla Raza occurred from the north, rather than the
south, the peninsula (20.8 km from Isla Raza), or
mainland Mexico (.40 km). This finding of a northern
Gulf association is surprising given that oceanic currents in the region consist primarily of upwelling tidal
mixings as opposed to strong directionally dominating
summer currents (Maluf, 1983). Moreover, because Isla
Mejia (3) is a land bridge island to Isla Angel de la
Guarda (4; Gastil et al., 1983), our cladogram reveals
that the Isla Raza (5) population had its origin about
111
PHYLOGENY OF SIDE-BLOTCHED LIZARDS
the same time as did Isla Mejia, between 6,000 and
12,000 years ago. Owing to prehistoric and historic
human activity on Isla Raza (Bahre, 1983), and the
absence of significant genetic differentiation, we cannot
rule out the possibility of a human-facilitated introduction of Uta.
Our data are equally important to understanding
species associations on other islands. Regarding other
insular colonizations, U. stansburiana stellata in the
Pacific Ocean must have colonized the Islas San Benito
(16) by overwater colonization, for these islands have
never had a connection with the peninsula. Not unexpectedly, U. stansburiana stellata shares a sister group
relationship with populations from the nearest land
mass, Isla Cedros (17; Fig. 2). The almost identical
similarity in mtDNA sequences suggests that the colonization of Islas San Benitos must have occurred very
recently, likely during the last 106 years, but undoubtedly during the Pleistocene.
Finally, it has been suggested that, based on color
patterns, some lizards and snakes on Isla Santa
Catalina (13) had an origin from the west coast of
mainland Mexico (Grismer, 1994b), as opposed to the
adjacent peninsula. Our data convincingly demonstrate an association of U. squamata from Isla Santa
Catalina with the southern peninsular clade. Its membership in the southern clade is accredited by a BSP of
96%, by noting that removal from the southern clade
requires a minimum of 11 additional steps (homoplastic
additions in 11 characters), and by statistical comparison using Templeton’s (1983) test (Ts 5 276, n 5 191,
p ,0.001). A similar conclusion was reached by Soulé
(1967) based on a phenetic evaluation of morphological
data. If the continental island’s faunal assemblage
reflects its former peninsular connection, as suggested
by Murphy (1983b), and not an aggregate of overwater
colonizations, then we may expect that all insular
species have a closer phylogenetic affinity with congeners in Baja California Sur. Hypothesized affinities
with western Mexico (Grismer, 1994b) more likely
reflect homoplastic changes in permutable color patterns.
Given the significance of the midpeninsular break in
mtDNA, it is possible that two species of Uta occur on
the peninsula. However, we believe that future investigations will not verify this to be true. Our data are
taken from the maternally inherited mitochondrial
genome. As such, they reflect female dispersion, and
not ongoing interbreeding and gene flow. Such expectations of a single peninsular species do not diminish the
power of mtDNA data in forming biogeographic hypotheses. Because of the maternal inheritance, we have
discovered the first compelling evidence for the existence of a midpeninsular seaway.
The significance of this find extends beyond sideblotched lizards. A midpeninsular seaway could explain
the occurrence of dual ‘‘peninsular effects’’ (Murphy,
1991) and the distributions of other terrestrial and
marine organisms and refocus studies on other taxa in
the region. If this pattern is replicated in additional
organisms, then the evidence for the seaway would
seem to be irrefutable.
We hope that these mtDNA data will help to inspire
further investigations on the stratigraphic development of the Baja California peninsula and associated
islands. Molecular analyses based on nuclear genes
alone may have overlooked this significant geologic and
biogeographic event. It is clear that future collaborative
interdisciplinary studies synthesizing morphological,
molecular genetic, and abiotic data should provide a
better understanding of our biodiversity and the associations between plate tectonics, paleostratigraphy, and
speciation.
Sequence data. Sequences have been submitted to
GenBank and are available. ATPase 6 sequences are
found for 23 populations, Accession Nos. U46695–
U46717. Cytochrome b sequences are found for 26
populations, Accession Nos. U46718–U46743.
ACKNOWLEDGMENTS
Tissue samples were generously donated by M. Soulé. Specimen
collection was carried out with permission of Direccion General de la
Fauna Silvestre, Mexico. Import permits for frozen tissues and
preserved specimens were issued by Agriculture Canada, and all
collecting and specimen euthanasia followed approved procedures.
This research was supported by the Natural Sciences and Engineering Research Council (NSERC) of Canada Grant A3148 and the Ft.
Irwin NTC to R.W.M.; special thanks to S. Ahmann, W. Johnson, and
M. Quilman of Ft. Irwin. M. Bobyn, D. Currie, A. Lathrop, R.
MacCulloch, M. Soulé, R. Winterbottom, and two anonymous reviewers provided valuable comments on the manuscript. This is contribution 56 from the Centre for Biodiversity and Conservation Biology of
the Royal Ontario Museum.
REFERENCES
Anderson, C. A. (1950). The 1940 E. W. Scripps cruise to the Gulf of
California. Part I: Geology of the islands and neighboring land
areas. Geol. Soc. Am. Mem. 43: 1–53.
Bahre, C. J. (1983). Human impact: The midriff islands. In ‘‘Island
Biogeography in the Sea of Cortez’’ (T. J. Case and M. L. Cody,
Eds.), pp. 290–306, Univ. California Press, Berkeley.
Ballinger, R. E., and Tinkle, D. W. (1972). Systematics and evolution
of the genus Uta (Sauria: Iguanidae). Misc. Pub. Mus. Zool. Univ.
Michigan 145: 1–83.
Beal, C. H. (1948). Reconnaissance of the geology and oil possibilities
of Baja California, Mexico. Geol. Soc. Am. Mem. 31: 1–138.
Betancourt, J. L., Van Devender, T. R., and Martin, P. S. (1990).
‘‘Packrat Middens: The Last 40,000 Years of Biotic Change,’’ Univ.
Arizona Press, Tucson.
Birt, T. P., Vriesen, V. L., Green, J. M., Montevecchi, W. A., and
Davidson, W. S. (1992). Cytochrome b sequence variation among
parrots. Heredity 117: 67–72.
Bremer, K. (1988). The limits of amino acid sequence data in
angiosperm phylogenetic reconstruction. Evolution 42: 795–803.
Brooks, D. R., and McLennan, D. H.(1991). ‘‘Phylogeny, Ecology, and
Behavior,’’ Univ. Chicago Press, Chicago.
112
UPTON AND MURPHY
Cabot, E. L., and Beckenbach, A. T.(1989). Simultaneous editing of
multiple nucleic acid and protein sequences with ESEE. Comp.
Appl. Biosci. 5: 233–234.
Case, T. J. (1975). Species numbers, density compensation, and the
colonizing ability of lizards on islands in the Gulf of California.
Ecology 56: 3–18.
Case, T. J. (1983). The reptiles: Ecology. In ‘‘Island Biogeography in
the Sea of Cortez’’ (T. J. Case and M. L. Cody, Eds.), pp. 159–209,
Univ. California Press, Berkeley.
Crabtree, C. B. (1983). Assessment of genetic differentiation within
the atherinid genus Leuresthes. Isozyme Bull. 16: 77.
Durham, J. W., and Allison, E. C. (1960). Geological history of Baja
California and its marine environments. Syst. Zool. 9: 47–91.
Eernisse, D. J., and Kluge, A. G. (1993). Taxonomic congruence verses
total evidence, and amniote phylogeny inferred from fossiles,
molecules, and morphology. Mol. Biol. Evol. 10: 1170–1195.
Faith, D. P., and Cranston, P. S. (1991). Could a cladogram this short
have arisen by chance alone?: On permutation tests for cladistic
structure. Cladistics 7: 1–28.
Farris, J. S. (1969). A successive approximations approach to character weighting. Syst. Zool. 18: 374–385.
Farris, J. S. (1988). ‘‘Hennig86, Version 1.5: Computer program and
documentation,’’ Port Jefferson Station, New York.
Felsenstein, J. (1985). Confidence limits of phylogenies: an approach
using the bootstrap. Evolution 39: 783–791.
Frost, D. R., and Hillis, D. M. (1990). Species in concept and practice:
Herpetological applications. Herpetologica 46: 87–104.
Fu, J., and Murphy, R. W. (1997). Phylogeny of Chinese Oreolalax and
the use of functional outgroups to select among multiple equally
parsimonious trees. Asiatic Herpetol. Res. 7: 38–43.
Gastil, G., Minch, J., and Phillips, R. P. (1983). The geology and ages
of islands. In ‘‘Island Biogeography in the Sea of Cortez’’ (T. J. Case
and M. L. Cody, Eds.), pp. 13–25, Univ. California Press, Berkeley.
Graybeal, A. (1995). Naming of species. Syst. Biol. 44: 237–250.
Grismer, L. L. (1994a). Three new species of intertidal side-blotched
lizards (Genus Uta) from the Gulf of California, Mexico. Herpetologica 50: 451–474.
Grismer, L. L. (1994b). Geographic origins for reptiles on islands in
the Gulf of California. Herpetol. Nat. Hist. 2: 17–40.
Grismer, L. L. (1988). Geographic distribution, taxonomy, and biogeography of the anguid genus Elgaria (Reptilia: Squamata) in Baja
California, Mexico. Herpetologica 44: 139–154.
Grismer, L. L., and McGuire, J. A. (1993). The oasis of central Baja
California, Mexico. Part I. A preliminary account of the relict
mesophillic herpetofauna and the status of the oases. Bull. Southern Calif. Acad. Sci. 92: 2–24.
Grismer, L. L., McGuire, J. A. and Hollingsworth. B. D. (1994). A
report on the herpetofauna of the Vizcaino Peninsula, Baja California with a discussion of its biogeographic and taxonomic implications. Bull. Southern Calif. Acad. Sci. 93: 45–80.
Grismer, L. L., and Mellink, E. (1994). The addition of Sceloporus
occidentalis to the herpetofauna of Isla de Cedros, Baja California,
Mexico and its historical and taxonomic implications. J. Herpetol.
28: 120–126.
Hews, D. (1990). Examining hypotheses generated by field measures
of sexual selection on male lizards, Uta palmeri. Evolution, 44:
1956–1966.
Hillis, D. M. (1991). Discriminating between phylogenetic signal and
random noise in DNA sequences. In ‘‘Phylogenetic Analysis of DNA
Sequences’’ (M. M. Miyamoto, and J. Cracraft, Eds.), pp. 278–294,
Oxford Univ. Press, Oxford.
Hillis, D. M., Allard, M. W., and Miyamoto, M. M. (1993). Analysis of
DNA sequence data: Phylogenetic inference. Methods Enzymol.
242: 456–487.
Hillis, D. M., and Huelsenbeck, J. P. (1992). Signal, noise, and
reliability in molecular phylogenetic analyses. J. Hered. 83: 189–
195.
Huelsenbeck, J. P. (1991). Tree-length distribution skewness: an
indicator of phylogenetic information. Syst. Zool. 40: 257–270.
Johnson, I. (1924). Expedition of the California Academy of Sciences
to the Gulf of California in 1921. The botany (vascular plants).
Proc. Calif. Acad. Sci. 12: 951–1218.
Kocher, T. D., Thomas, W. K., Meyer, A., Edwards, S. V., Paabo, S.,
Villablanca, F. X., and Wilson, A. S. (1989). Dynamics of mitochondrial DNA evolution in animals: Amplification and sequencing with
conserved primers. Proc. Natl. Acad. Sci. USA 86: 6196–6200.
Kumar, S., Tamura, K., and Nei, M. (1993). ‘‘MEGA: Molecular
Evolutionary Genetics Analysis, Version 1.0,’’ The Pennsylvania
State University, University Park, PA.
Maddison, W., Maddison, D. (1992). ‘‘MacClade: Analysis of Phylogeny and Character Evolution, Version 3.04,’’ Sinauer, Sunderland,
MA.
Maluf, L. Y. (1983). The physical oceanography. In ‘‘Island Biogeography in the Sea of Cortez’’ (T. J. Case and M. L. Cody, Eds.), pp.
26–48, Univ. California Press, Berkeley.
Moore, D. G. (1973). Plate-edge deformation and crustal growth, Gulf
of California structural province. Geol. Soc. Am. Bull. 84: 1833–
1906.
Murphy, R. W. (1983a). Paleobiogeography and genetic differentiation of the Baja California Herpetofauna. Occ. Pap. Calif. Acad.
Sci. 137: 1–48.
Murphy, R. W. (1983b). The reptiles: origins and evolution. In ‘‘Island
Biogeography in the Sea of Cortez’’ (T. J. Case and M. L. Cody,
Eds.), pp. 130–157, Univ. California Press, Berkeley.
Murphy, R. W. (1983c). A distributional checklist of the reptiles and
amphibians on the islands in the Sea of Cortéz. In ‘‘Island
Biogeography in the Sea of Cortez’’ (T. J. Case and M. L. Cody,
Eds.), pp. 429–437, Univ. California Press, Berkeley.
Murphy, R. W. (1991). Plate tectonics, peninsular effects, and the
borderlands herpetofauna of western North America. In: ‘‘Environmental Hazards and Bioresource Issues of the United StatesMexico Borderlands’’ (P. Ganster and H. Walter, Eds.), pp. 433–457,
Latin American Center, Univ. California (UCLA), Los Angeles.
Murphy, R. W., and Ottley, J. R. (1984). Distribution of amphibians
and reptiles on islands in the Gulf of California. Ann. Carnegie
Mus. 53: 207–230.
Murphy, R. W., Kovac, V., Haddrath, O., Allen, G. S., Fishbein, A., and
Mandrak, N. (1995). MTDNA gene sequence, allozyme, and morphological uniformity among red diamond rattlesnakes, Crotalus ruber
and Crotalus exsul. Can. J. Zool. 73: 270–281.
Nelsen, E. E. (1921). Lower California and its natural resources.
Natl. Acad. Sci. 16: 1–194.
Orton, R. D., and Buth, D. G. (1984). Minimal genetic differentiation
between Girella nigricans and Girella simplicidens (Perciformes:
Kyphosidae). Isozyme Bull. 17: 66.
Patton, J. L., and Smith, M. F. (1994). Paraphyly, polyphyly, and the
nature of species boundaries in pocket gophers (genus Thomomys).
Syst. Biol. 43: 11–26.
Present, T. M. C. (1987). Genetic differentiation of disjunct Gulf of
California and Pacific outer coast populations of Hypsoblennius
jenkinsi. Copeia 1987: 1010–1024.
Roe, B. A., Ma, D. P., Wilson, R. K., and Wong, J. F. H. (1985). The
complete nucleotide sequence of the Xenopus levis mitochondrial
genome. J. Biol. Chem. 260: 9759–9774.
Sankoff, D., and Rousseau, P. (1975). Locating the vertices of Steiner
tree in arbitrary space. Math. Progr. 9: 240–246.
Savage, J. M. (1960). Evolution of a peninsular herpetofauna. Syst.
Zool. 9: 184–212.
PHYLOGENY OF SIDE-BLOTCHED LIZARDS
Siddall, M. E. (1994). ‘‘Random Cladistics, Version 2.1,’’ Department
of Zoology, Univ. Toronto, Toronto.
Simpson, G. G. (1961). ‘‘Principles of Animal Taxonomy,’’ Columbia
Univ. Press, New York.
Soulé, M. (1966). Trends in the insular radiation of a lizard. Am. Nat.
100: 47–64.
Soulé, M. (1967). Phenetics of natural populations. I. Phenetic
relationships of insular populations of the side-blotched lizard.
Evolution 21: 584–591.
Soulé, M., and Sloan, A. J. (1966). Biogeography and distribution of
reptiles and amphibians on the islands in the Gulf of California,
Mexico. Trans. San Diego Soc. Nat. Hist. 14: 137–156.
Soulé, M., and Yang, S. Y. (1973). Genetic variation in side-blotched
lizards on islands in the Gulf of California. Evolution 27: 593–600.
Stebbins, R. C. (1985). ‘‘A Field Guide to Western Reptiles and
Amphibians,’’ Houghton Mifflin, Boston.
Swofford, D. L. (1993). ‘‘PAUP: Phylogenetic Analysis Using Parsimony, Version 3.1.1: Computer Program and Documentation,’’
Illinois Natur. Hist. Survey, Urbana.
Swofford, D. L., Olsen, G. J., Waddell, P. J., and Hillis, D. M. (1996).
113
Phylogenetic inference. In ‘‘Molecular Systematics’’ (D. M. Hillis,
C. Moritz, and B. Mable, Eds.), pp. 407–514, Sinauer, Sunderland, M.A.
Templeton, A. R. (1983). Phylogenetic inference from restriction endonuclease cleavage site maps with particular reference to the
evolution of humans and apes. Evolution 37: 221–244.
Theriot, E. (1991). Clusters, species concepts, and morphological
evolution of diatoms. Syst. Biol. 41: 141–157.
Tinkle, D. W. (1967). The life and demography of the side-blotched
lizard, Uta stansburiana. Misc. Pub. Mus. Zool. Univ. Michigan
132: 1–182.
Villablanca, F. X. (1994). Spatial and temporal aspects of populations
revealed by mitochondrial DNA. In ‘‘Ancient DNA’’ (B. Hermann,
and S. Hummel, Eds.), Springer-Verlag, New York.
Watrous, L. E., and Wheeler, Q. D. (1981). The outgroup comparison
method of character analysis. Syst. Zool. 30: 1–11.
Wilcox, B. A. (1978). Supersaturated island faunas: A species-age
relationship for lizards on post-Pleistocene land-bridge islands.
Science 199: 996–998.
Wiley, E. O. (1978). The evolutionary species concept reconsidered.
Syst. Zool. 27: 17–26.