THE GEOGRAPHY OF MITOCHONDRIAL DNA VARIATION IN TWO SYMPATRIC SPARROWS

Geographic variation in mitochondrial DNA (mtDNA) restriction sites was studied in samples of two sympatric passerine birds, fox (Passerella iliaca) and song (Melospiza melodia) sparrows, collected at the same sites in the western United States. Different levels of variation and differentiation were observed in each species. In 46 fox sparrows taken at nine sites, five clones were observed, partitioned into two distinct east-west groups that meet at the Great Basin-Sierra Nevada interface; percent nucleotide divergence was 0.86 between groups and 0.08 within groups. An additional 43 individuals were examined using at least one of seven diagnostic endonucleases, and all supported the east-west groupings. Considering common mtDNA haplotypes as alleles, an F_ST of 0.50 was computed, which is an order of magnitude greater than that computed from allozyme comparisons (0.019); mtDNA analyses suggest little intergroup gene exchange. Compared to allozymic variation, analysis of mtDNA revealed a greater degree of population structuring and greater consistency with broad patterns of morphological variation. Fifteen clones were observed in 27 song sparrows taken at seven of the same sites at which fox sparrows were sampled; the percent nucleotide divergence averaged 0.27. There was no detectable geographic pattern to the variation, and no evidence of an east-west division as in the fox sparrow. However, the mosaic nature of mtDNA variation in song sparrows suggests limited gene exchange. Considering the 15 clones as alleles yielded an F_st of 0.24, which is reduced to 0.039 when corrected for sampling error. In spite of occupying the same geographic area, mtDNA analyses showed that the two species (or at least their mtDNA gene genealogies) have had different evolutionary histories.     

MITOCHONDRIAL DNA HOMOGENEITY IN THE PHENOTYPICALLY DIVERSE REDPOLL FINCH COMPLEX (AVES: CARDUELINAE: CARDUELIS FLAMMEA-HORNEMANNI)

Breeding redpoll finches (Aves: Carduelinae) show extensive plumage and size variability and, in many cases, a plumage polymorphism that is not related to age or sex. This has been ascribed to extreme phenotypic variation within a single taxon or to moderate variability within distinct taxa coupled with hybridization. The predominant view favors the recognition of two largely sympatric species: Carduelis flammea, comprised of four well-marked subspecies—flammea, cabaret, islandica, and rostrata; and C. hornemanni, comprised of two subspecies—hornemanni and exilipes. We studied representative samples of these putative subspecies (except islandica) for variation in mitochondrial DNA (mtDNA). Using 20 informative restriction enzymes that recognized 124 sites (642 base pairs [bp] of sequence or ≈ 3.7% of the molecule), we identified 17 RFLP haplotypes in the 31 individuals surveyed. The haplotypes formed a simple phylogenetic network with most clones diverging by a single site difference from a common haplotype found in almost half of the individuals. Within populations and taxa, levels of mtDNA diversity were similar to those observed in other avian species. The pattern of mtDNA divergence among populations was statistically unrelated to their geographic or traditional taxonomic relationships, and the estimated distance between the two traditionally recognized species was very small relative to those typically observed among avian sister species.     

HISTORICAL BIOGEOGRAPHY AND CONTEMPORARY PATTERNS OF MITOCHONDRIAL DNA VARIATION IN WHITE-TAILED DEER FROM THE SOUTHEASTERN UNITED STATES

Mitochondrial DNA (mtDNA) was used to characterize patterns of geographic variation among white-tailed deer (Odocoileus virginianus) populations in the southeastern United States. Fifteen restriction enzymes were employed to survey and map 99 restriction sites in 142 deer from 18 localities in five southeastern states. Phylogenetic analysis revealed three primary groups of haplotypes: (1) southern Florida and the Florida Keys, (2) the remainder of peninsular Florida northward to South Carolina, and (3) the Florida panhandle westward to Mississippi. Geographical heterogeneity in haplotype frequencies suggests that stochastic lineage sorting or isolation by distance are not important determinates of mtDNA differentiation among deer populations. The pattern of mtDNA variation in white-tailed deer is concordant spatially with those observed in unrelated taxa suggesting the common influence of historical biogeographic events. The data (1) support previous hypotheses that relate contemporary patterns of intraspecific phylogeography in northern Florida to the physiogeographic history of the region; and (2) suggest that genetic differentiation in southern Florida may be attributable to episodes of Pleistocene dispersal. Despite potentially high vagility and human intervention, ecological and demographic characteristics of deer have effectively preserved the historical pattern of intraspecific mtDNA differentiation.     

Taxonomy and conservation: a tale of two tinamou species groups (Tinamidae,Crypturellus)

Species delimitation has important consequences for the management of endangered species. Species‐level taxonomy in the genus Crypturellus (Tinamidae) has been based largely on plumage characters and species limits in several groups have been difficult to establish. Because some of the forms of uncertain taxonomic status are currently threatened with extinction, a basic understanding of species limits is crucial not only for taxonomists but also for conservation biologists and managers. We analysed vocal variation to assess species limits in two Crypturellus species‐groups, the red‐legged complex (Crypturellus erythropus and allied forms) and the brown tinamou Crypturellus obsoletus. In the red‐legged complex, where several species‐level taxa have been recognized by some authors, there is no obvious geographic variation in vocalizations and populations appear mostly continuously distributed, with plumage variation largely explicable in terms of environmental conditions. In the brown group, a single species is recognized, but we found marked geographic variation in vocalizations and populations have disjunct distributions; we propose that at least one of the populations in this group likely merits recognition as a separate species. We conclude that incomplete knowledge of patterns of variation in relevant traits in addition to the momentum carried by traditional taxonomy may potentially mislead conservation actions.     

Morphological and molecular evolution and their consequences for conservation and taxonomy in the Le Conte's thrasher Toxostoma lecontei

We evaluated geographic variation and subspecific taxonomy in the Le Conte's thrasher Toxostoma lecontei by analyzing DNA sequences from 16 nuclear loci, one mitochondrial DNA locus, and four study skin characters, and compared these data sets with previously published data on plumage coloration and different mtDNA genes. Morphological support for the southernmost taxon, T. l. arenicola, is relatively weak: multivariate analyses of morphometrics or back coloration do not provide diagnostic support, although one color character differs statistically. However, combined DNA analyses indicate that T. l. arenicola is diagnosable and reciprocally monophyletic, diverging from T. l. lecontei at least 140 000 yr ago. Little to no past introgression across a very short geographic distance despite the long period of isolation is strong evidence of independently evolving taxa. We suggest that the lack of morphological divergence in traits related to niche use has prevented the two taxa from invading each other's range. Despite relatively weak morphological differences we suggest that these two deeply divergent lineages merit species status, and we suggest Vizcaíno thrasher for the common name corresponding to T. l. arenicola. The population size of T. l. arenicola is small and the taxon is in need of preservation attention.     

MITOCHONDRIAL DNA VARIATION IN THE FUNGUS ATKINSONELLA HYPOXYLON INFECTING SYMPATRIC DANTHONIA GRASSES

The purpose of this study was to determine how host identity and geographic proximity were related to genetic variation in the fungus Atkinsonella hypoxylon infecting four sympatric Danthonia grasses. We analyzed 192 A. hypoxylon isolates from 48 geographic sites for mitochondrial DNA (mtDNA) variation by restriction digestion of total DNA with EcoRI, BamHI, and PstI, and subsequent hybridization with purified A. hypoxylon mtDNA. Thirtynine haplotypes were identified. UPGMA cluster analysis showed that mtDNA type is highly correlated with host-species group; isolates from Danthonia compressa and D. spicata formed one branch of the phenogram, and isolates from D. epilis and D. sericea formed the other. Restriction maps of the most common mtDNA haplotypes infecting each host-species pair revealed a 30-kb size difference and a minimum of eight length changes and one restrictionsite change between them. Mapping of the mutation differences among all haplotypes occurring at North Carolina sites resulted in two distinct gene trees corresponding to the two Danthonia species groups, corroborating the phenetic analysis. The results indicate a high degree of host-dependent isolation and establish the existence of host races in A. hypoxylon. Possible mechanisms responsible for this isolation are discussed. Little differentiation existed between isolates from within a Danthonia species pair, and some variation was explained by geographic origin. Analysis of progeny from a natural sexual cross revealed that mitochondria are maternally inherited in A. hypoxylon.     

Genetic differentiation among morphological variants of Acacia saligna (Mimosaceae)

This study investigated the pattern of genetic variation in Acacia saligna (Labill.) H.L. Wendl. to facilitate its development as a crop species for dryland salinity management. A. saligna is a morphologically variable species and four main variants are recognized. The genetic structure within A. saligna was investigated in populations across the geographic range of the species using nuclear restriction fragment length polymorphism loci. The analysis identified considerable genetic variation within A. saligna that was genetically structured into three groups. Two of the three groups corresponded to variants recognized in the field study; the third group encompassed the other two variants which, though morphologically different, were not genetically differentiated. The level of genetic differentiation between the groups suggests they may represent different taxa and a taxonomic revision of the species may be required. Identification of different taxa within A. saligna will have implications for the utilization and domestication of the species, as the taxa will need to be evaluated separately to determine their suitability for agroforestry. The high genetic variation between and within groups suggests there is a large genetic base available for breeding improved cultivars.     

GENETIC SUBDIVISIONS AMONG SMALL CANIDS: MITOCHONDRIAL DNA DIFFERENTIATION OF SWIFT,KIT, AND ARCTIC FOXES

Gene flow can effectively suppress genetic divergence among widely separated populations in highly mobile species. However, the same may not be true of species that typically disperse over shorter distances. Using mtDNA restriction-site and sequence analyses, we evaluate the extent of divergence among populations of two small relatively sedentary North American canids, the kit and swift foxes (genus Vulpes). We determine the significance of genetic differentiation among populations separated by distance and those separated by discrete topographic barriers. Our results show the among-population component of genetic variation in kit and swift foxes is large and similar to that of small rodents with limited dispersal ability. In addition, we found two distinct groupings of genotypes, separated by the Rocky Mountains, corresponding to the traditional division between kit and swift fox populations. Previous workers have characterized these morphologically similar populations either as separate species or subspecies. Our mtDNA data also suggest that kit and swift fox populations hybridize over a limited geographic area. However, the sequence divergence between kit and swift foxes is similar to that between these taxa and the arctic fox (Alopex lagopus), a morphologically distinct species commonly placed in a separate genus. This result presents a dilemma for species concepts, and we conclude that kit and swift foxes should be recognized as separate species.     

GENE FLOW,REFUGIA, AND EVOLUTION OF GEOGRAPHIC VARIATION IN THE SONG SPARROW (MELOSPIZA MELODIA)

We surveyed mtDNA restriction-site variation in song sparrows taken from across their continental range. Despite marked geographic variation in size and plumage color, mtDNA variation was not geographically structured. Subspecies were not identifiable by mtDNA analysis. We suggest that postglaciation dispersal scattered mtDNA haplotypes across the continent, explaining the lack of mtDNA geographic patterns. Evolution of size and plumage coloration has probably proceeded faster than mtDNA evolution, leading to the well-structured continental pattern of morphological variation. We suggest that the nonordered geographic distribution of haplotypes reflects the recency of population establishment following completion of range expansion. Dispersal distance was estimated from the mtDNA data at 6.1 km per generation, an order of magnitude greater than that (0.3 km) estimated from demographic data. Island samples were not especially different from continental ones. Rooting the haplotype cladogram with a putative primitive haplotype identified Newfoundland and the Queen Charlotte Islands as potential sites of recent refugia. We question whether study of geographic variation in song sparrows leads to insights concerning speciation.     

HISTORICAL BIOGEOGRAPHY OF THE BANANAQUIT (COEREBA FLAVEOLA) IN THE CARIBBEAN REGION: A MITOCHONDRIAL DNA ASSESSMENT

We analyzed mitochondrial DNA (mtDNA) restriction-site variation in bananaquit (Coereba flaveola; Aves, Coerebinae) populations sampled on 12 Caribbean islands and at 5 continental localities in Central America and northern South America. Multiple fixed restriction-site differences genetically defined several regional bananaquit populations. An mtDNA clade representing all Jamaican bananaquits was the most divergent; the estimated average sequence divergence (d_xy) between Jamaican and all other mtDNA haplotypes surveyed was 0.027. Three groups of populations, representing Central America, northern South America, and the eastern Antilles (Puerto Rico to Grenada) were nearly equally differentiated among themselves (average d_xy = 0.014), and may represent a single, recent range expansion. Within the eastern Antilles, three geographically restricted haplotype groups were identified: Puerto Rico, north-central Lesser Antilles (U.S. Virgin Islands to St. Lucia), and Grenada–St. Vincent. The evolutionary relationships of these groups were not clear. Genetic homogeneity of the island populations from the U.S. Virgin Islands to St. Lucia suggested a recent spread of a specific north-central Lesser Antillean haplotype through most of those islands. Haplotype variation across this region indicated that this spread may have occurred in two waves, first through the southernmost islands of St. Lucia, Martinique, and Dominica, and more recently from Guadeloupe to the north. The geographic distribution of mtDNA haplotypes, and of bananaquit populations, suggests periods of invasiveness followed by relative geographic quiescence. Although most genetic studies of bird populations have revealed homogeneity over large geographic areas, our findings provide a remarkable counterexample of strong geographic structuring of mtDNA variation over relatively small distances. Furthermore, although the mtDNA data were consistent with several subspecific distinctions, it was clear that named subspecies do not define equally differentiated evolutionary entities.     

MITOCHONDRIAL-DNA VARIATION IN THE CRESTED NEWT SUPERSPECIES: LIMITED CYTOPLASMIC GENE FLOW AMONG SPECIES

The crested newt has a widespread European distribution and encompasses four taxa recently elevated to full species: Triturus cristatus, T. carnifex, T. dobrogicus, and T. karelini. These are distinct on morphological, chromosomal, and isozymic grounds and have fairly sharp transition zones. A widespread survey (12 countries, 49 geographic sites, 210 individuals) of mtDNA variation (20–27 restriction enzyme sites mapped per individual) was made in order to 1) correlate mtDNA variation with morphological features defining the species, 2) determine the degree of differentiation within and among species, and 3) detect any introgression among species. The mtDNAs of these species were clearly differentiated (d = 3.9–7.1%). Additionally, geographic structuring was observed within T. carnifex and T. karelini, each displaying two divergent mitochondrial genome types (d = 3.5% and 4.7%, respectively). The other two (more northerly distributed) species were genetically homogeneous over most (T. cristatus) or all (T. dobrogicus) of their ranges. In the case of T. cristatus, one may infer bottlenecking as a result of Pleistocene glaciation events. This may also apply in part to T. dobrogicus, but high population connectedness and gene flow in this lowland river species may alone be sufficient for homogenization of mtDNA. Patterns of mtDNA variation were largely concordant with morphology; some interspecific mitochondrial gene flow was observed, but only close to or in the transition zones. Analyses of mapped restriction-site data by UPGMA and parsimony methods (using the closely related T. marmoratus as an outgroup) produce very similar dendrograms. The levels of divergence found concur with the systematics of the group, but the differentiation within T. carnifex and T. karelini is notable.     

EVIDENCE FOR A MOSAIC HYBRID ZONE IN THE GRASS SHRIMP PALAEMONETES KADIAKENSIS (PALAEMONIDAE) AS REVEALED BY MULTIPLE GENETIC MARKERS

Molecular techniques provide powerful tools for studying the geographic structure of hybrid zones and the dynamics of gene exchange between incipient species. We examined allozyme variation at five loci (PGM, GPI, MDH-1, MDH-2, and LDH) for 27 populations of Palaemonetes kadiakensis from the central, coastal, and eastern regions of Texas. Central Texas populations of P. kadiakensis exhibited highly significant linkage disequilibrium and departures from Hardy-Weinberg genotype proportions. In populations with linkage disequilibrium, allelic differences at GPI defined two types of P. kadiakensis, designated A and B. Both types existed in central Texas with little or no evidence of interbreeding, whereas the populations from all other localities showed complete introgression of type B alleles into the type A gene pool. We also examined ribosomal DNA (rDNA) and mitochondrial DNA (mtDNA) variation in a subset of populations, chosen to cover a range of geographic locations and levels of linkage disequilibrium. Two groups of mtDNA haplotypes and two restriction fragment patterns for the rDNA corresponded to allozyme type A and B individuals in populations exhibiting linkage disequilibrium. In populations with ongoing hybridization, all hybrid animals (N= 15) exhibited type A mtDNA. Exhibition of type A mtDNA indicated that type A females had mated successfully with type B males, but type B females had not mated successfully with type A males. Genotype distributions suggest reduced reproduction by hybrid offspring in central Texas populations. These patterns are consistent with a mosaic model of hybrid zone dynamics.     

Molecular and plumage analyses indicate the incomplete separation of two woodpeckers (Aves,Picidae)

We evaluated the relationship between Celeus undatus and Celeus grammicus, with the objective of clarifying their evolutionary history. We analysed fragments of the mitochondrial and nuclear genes of 57 specimens. For comparative purposes, we inspected the plumage patterns of 77 skins. Our findings highlight the absence of reciprocal monophyly between the two taxa, given their reduced genetic divergence, and the lack of any clear separation of the two forms in the haplotype networks. A similar situation was found in the STRUCTURE analysis, with reciprocal contributions from the two taxa to the respective clusters, indicating that C. grammicus and C. undatus cannot be differentiated using the molecular markers. Corroborating the genetic data, our plumage analyses also failed to find any clear diagnostic characters between the polytypic C. undatus and C. grammicus, as they are defined at present. The genetic profile is consistent with either extensive historical gene flow between the species or, alternatively, incomplete lineage sorting, rather than recent secondary contact. The lack of monophyly between the two taxa impeded subspecies‐level phylogeographic inferences, with the subspecific variation being interpreted as a probable artefact of the phenotypic plasticity of the two forms. These findings indicate clearly that the two taxa form a single evolutionary unit, in which the morphological differentiation used to diagnose the species, combined with their geographic distribution, is at odds with the incomplete separation of the taxa. This may reflect disparities in the rates of differentiation between molecular and phenotypic markers, which is possibly due to the variation in selection pressures along a humidity gradient in Amazonia.     

Extensive phenotypic diversification coexists with little genetic divergence and a lack of population structure in the White Wagtail subspecies complex (Motacilla alba)

Geographically clustered phenotypes often demonstrate consistent patterns in molecular markers, particularly mitochondrial DNA (mtDNA) traditionally used in phylogeographic studies. However, distinct evolutionary trajectories among traits and markers can lead to their discordance. First, geographic structure in phenotypic traits and nuclear molecular markers can be co‐aligned but inconsistent with mtDNA (mito‐nuclear discordance). Alternatively, phenotypic variation can have little to do with patterns in neither mtDNA nor nuclear markers. Disentangling between these distinct patterns can provide insight into the role of selection, demography and gene flow in population divergence. Here, we examined a previously reported case of strong inconsistency between geographic structure in mtDNA and plumage traits in a widespread polytypic bird species, the White Wagtail (Motacilla alba). We tested whether this pattern is due to mito‐nuclear discordance or discrepancy between morphological evolution and both nuclear and mtDNA markers. We analysed population differentiation and structure across six out of nine commonly recognized subspecies using 17 microsatellite loci and a combination of microsatellites and plumage indices in a comprehensively sampled region of a contact between two subspecies. We did not find support for the mito‐nuclear discordance hypothesis: nuclear markers indicated a subtle signal of genetic clustering only partially consistent with plumage groups, similar to previous findings that relied on mtDNA. We discuss evolutionary factors that could have shaped the intricate patterns of phenotypic diversification in the White wagtail and the role that repeated selection on plumage ‘hotspots’ and hybridization may have played.     

GEOGRAPHIC DISTRIBUTION OF MITOCHONDRIAL DNA VARIANTS AND THE HISTORICAL BIOGEOGRAPHY OF THE SPOTTED SALAMANDER,AMBYSTOMA MACULATUM

I analyzed geographic partitioning of mitochondrial DNA (mtDNA) restriction-site variants in the spotted salamander, Ambystoma maculatum. Two highly divergent and geographically separate genetic lineages were identified that differed by a minimum of 19 restriction sites (6% sequence divergence). One of the lineages has a disjunct distribution with very closely related haplotypes occurring in Missouri, Arkansas, North Carolina, and Virginia. The other lineage is found in Michigan, Illinois, and Alabama. The geographic separation of highly divergent mtDNA haplotypes, a pattern that was predicted based on the sedentary nature of these salamanders, is evidence for long-term barriers to gene flow. In contrast, the large-scale disjunction of very similar haplotypes suggests recent, long-distance gene flow and does not match the phylogeographic expectation for a small terrestrial vertebrate. I explain this potential contradiction in the level of importance assigned to gene flow by a scenario in which historical barriers to gene flow account for the two divergent mtDNA assemblages, but stochastic sorting of ancestral polymorphism is responsible for the large-scale geographic disjunction. Ten of 16 populations collected in the Ozark Highlands were fixed for the same haplotype. I attribute this lack of detectable variation to recent colonization of this area, a hypothesis that is supported by paleoecological data and demonstrates the potential benefits of combining data from paleobotany, geology, and other disciplines to reconstruct the historical biogeography of a species.     

Mitochondrial DNA Variation Follows a Geographic Pattern in Japanese Beech Species

The amount and distribution of mitochondrial (mt) DNA restriction fragment length polymorphism was determined among individual tree samples of two Japanese beech species, Fagus crenata and F.japonica. Individual plants were collected from 16 F. crenata populations throughout the range of the species, and from three F. japonica populations. We detected enough variation to characterize eleven and three chondriome types in F. crenata and F.japonica, respectively. The grouping of beech chondriome types based upon the cladistic analysis of mtDNA polymorphism allowed us to recognize the apparent geographical patterns of mtDIMA diversity: the resulting three main groups occupied distinct geographic areas. This geographic differentiation is likely to reflect the history of the Japanese beech forests after the last glacial period of the Pleistocene. In addition, the mtDNA polymorphism encountered within F. crenata encompassed all the variation observed in F.japonica. Our result suggests the need for re-evaluation of their phylogenetic relationships.     

Phylogeny and species limits in the Palaearctic chiffchaff Phylloscopus collybita complex: mitochondrial genetic differentiation and bioacoustic evidence*

Nucleotide sequences of the mitochondrial cytochrome b gene (1041 bp), analysis of vocalizations and behavioural evidence from zones of contact were used to reassess the species limits and phylogenetic relationships at the species and subspecies levels in the Phylloscopus collybita complex. A new classification is proposed which recognizes four biological species. Phylloscopus brehmii (Iberia) and P. canariensis (Canary Islands) are genetically and bioacoustically highly distinct. There is no mitochondrial gene flow between them or with P. collybita. The Mountain Chiffchaff P. sindianus (with subspecies sindianus and Iorenzii) is equally distinct genetically from southwest Asian subspecies of P. collybita (caucasicus, brevirostris). The status of the Siberian form tristis, which shares potentially synapomorphic characters with the Mountain Chiffchaff (ascending song notes, grey-brown adult plumage) but genetically closely resembles P. c. collybita and P. c. abietinus, remains uncertain. In two zones of secondary contact between taxa with “greenish” (P. collybita) and brownish plumage, hybridization is either unrecorded (caucasicus v lorenzii, Caucasus Mountains) or its extent is insufficiently known (abietinus v tristis, west of Ural Mountains). A phylogeny reconstructed from nucleotide sequences agrees with one based on song and some morphological characters in identifying P. brehmii as the oldest branch within the chiffchaff group. Of the remaining taxa, four fall into a clade with greenish plumage (P. collybita ssp.), two into one with brownish plumage (P. s. sindianus, P. s. lorenzii), while the position of P. canariensis with respect to these two clades is uncertain. Molecular and phenotypic phylogenies contradict each other regarding the affinities of P. (ssp.?) tristis.     

High levels of population differentiation for mitochondrial DNA haplotypes inPinus radiata,muricata, andattenuata

We analyzed mitochondrial (mt) DNA restriction fragment length polymorphisms (RFLPs) associated with cytochrome oxidase, subunit I (coxI)-related gene sequences in 268 trees derived from 19 natural populations of three species of pines from California (USA): Monterey pine (Pinus radiata D. Don), bishop pine (P. Muricata D. Don), and knobcone pine (P. attenuata Lemm.). Total genomic DNA was digested with four restriction endonucleases and probed with a 750-bp fragment of the mitochondrialcoxI gene amplified fromP. attenuata via the polymerase chain reaction (PCR). ThecoxI gene is repeated at least 4 times in some populations, and all variants that we observed resulted from complex rearrangements rather than from point mutations. There was limited intrapopulation variation, but strong differentiation among populations. When applied to haplotype frequencies, Nei's gene diversity within populations (H_s) averaged 7% (±3), and G_st varied from 75% forP. Radiata to 96% forP. muricata. The high degree of population differentiation for mtDNA suggests that it can be a powerful marker of population differences, but its rapid rate of structural evolution appears to result from recombination among a limited number of repetitive elements-giving frequent homoplasious fragment phenotypes. The phylogenetic trees disagreed with results from chloroplast DNA, nuclear gene, and morphological studies.