Natural variation in the appendicular skeleton of Triturus carnifex (Amphibia: Salamandridae)

Intraspecific variation in the appendicular skeleton of two geographically isolated populations of Triturus carnifex, one from northern Italy (Rosate, Milano) and one from central Italy (Bagnaia, Perugia), has been studied. A total of 1,746 forelimbs and 830 hindlimbs were examined. Forelimb skeletal variability was much greater in the Rosate than the Bagnaia population. Skeletal variants were present in 36.3% and 13.5% of the forelimbs, respectively, or in 54.7% and 22.7% of the netws (P < 0.0001). There were no predominant skeletal variants in Bagnaia, while in the Rosate population, the majority of the variants consisted of fusion of radiale and prepollicis and of phalangeal formula 1-2-3-2. Hindlimb skeletal variability was similar in the two populations and appeared to be much lower than that of the forelimb, with highly significant differences in the frequency of basipodium variants within the Rosate population and in the frequency of acropodium variants in both populations. Skeletal variants were present in about 9% of the hindlimbs, or in about 12% of the newts from either population. At present, no conclusion can be drawn about the mechanisms, genetic and/or epigenetic, underlying the skeletal variability observed in the Triturus carnifex from northern and central Italy. © 1996 Wiley-Liss, Inc.     

Patterns of genetic differentiation in Thamnophis and Taricha from the Pacific Northwest

Aim The co‐evolutionary interaction between the common garter snake, Thamnophis sirtalis, and the rough‐skinned newt, Taricha granulosa, takes place throughout much of the Pacific Northwest (North America). The biogeography of the Pacific Northwest has been heavily influenced by the last Pleistocene glaciation, which reached a maximum as late as 14,000 yr bp . We researched: (1) what type of population structure is present for garter snakes and newts, (2) whether the population structure of these species is consistent with a Pleistocene glaciation hypothesis, and (3) how population structure and migration possibly affect co‐evolution between these species. Location The Pacific Northwest of North America, specifically northern California, Oregon and Washington in the USA. Methods We sampled approximately 20 populations for each species from three different transects. Using microsatellite markers and tissue samples from both species, we quantified the population structure for both species. Individual‐based assignment tests were used to estimate contemporary migration rates. Results Both Th. sirtalis and Ta. granulosa exhibited little genetic differentiation among our study sites, even among those separated by large distances. Significant population structure was detected on multiple geographic scales. Differences in population structure were observed among transects and between garter snake and newt transects. Contemporary migration rate estimates indicate high levels of genetic exchange between populations. Main conclusions Prior to this study, little was known about the fine‐scale population structure of either species in this region. Patterns of population structure for garter snakes and newts reflect a shared biogeographical history affected by the Pleistocene glaciation in the Pacific Northwest. Both species apparently migrate frequently between populations, thus potentially retarding the process of adaptive co‐evolution. We find that populations from a northern coastal transect (Washington) are most likely to be locally adapted.     

Evidence for strong genetic structure in European populations of the little owl Athene noctua

The little owl Athene noctua is a widespread species in Europe. This mainly sedentary owl experienced reduction in population sizes in some areas due to habitat loss and modification of the landscape. To assess the genetic structure of the populations of western and central Europe, we analysed 333 specimens from 15 geographical areas at 13 microsatellite loci. Statistical analyses and Bayesian clustering procedures detected two major genetically distinct clusters, the first distributed from Portugal to the Czech Republic and the second from the Balkans to Italy. The second cluster was further split into three groups, located in Italy, Sardinia and the Balkans. These groups match four previously‐described mtDNA haplogroups, and probably originated from the isolation of little owl populations in Sardinia and in three glacial refugia (Iberia, south Italy and Balkans) during the ice ages. High genetic admixture was recorded in central and northern Europe, probably as a consequence of the expansion from the refugia during interglacial. The main colonization route originated from the Iberian Peninsula towards central and northern Europe. Contact zones with colonization events from Italy and the Balkans were detected respectively in northern Italy and central Europe. Genetic indices show the existence of moderate levels of genetic variability throughout Europe, although evidence of recent evolutionary bottlenecks was found in some populations. Estimation of migration rates and approximate Bayesian computations highlighted the most likely phylogeographical scenario for the current distribution of little owl populations.     

Genetic differentiation and biogeography in southern European populations of the genus Scardinius (Pisces, Cyprinidae) based on allozyme data

Levels of genetic differentiation and variability were assessed in several populations and species belonging to the primary freshwater fish genus Scardinius by means of electrophoretic analyses of 28 enzymatic loci. We studied one population of S. acarnanicus and one of S. graecus , both endemic to Greece, seven populations of S. erythrophthalmus from Greece, Slovenia, northern and central Italy, and the single extant population of S. scardafa , a species endemic to central Italy. Patterns of genetic relationships placed the Greek species basal in the neighbour-joining (NJ) and maximum parsimony (MP) trees, S. scardafa and S. erythrophthalmus are sister species. Based on the molecular clock, the split between S. acarnanicus and S. graecus occurred around 4.3 Myr ago and the split between ( S. acarnanicus , S. graecus ) and ( S. scardafa , S. erythrophthalmus ) around 3.5 Myr ago, both after the Messinian lacustrine 'Lago Mare' phase of the Mediterranean. The separation between S. scardafa and S. erythrophthalmus was more recent (0.5 Myr ago) and was probably the result of confluences between adjacent river basins during the Pleistocene. Within S. erythrophthalmus , the Greek population showed a certain degree of genetic differentiation, whereas northern and central Italian populations were nearly indistinguishable. These findings demonstrate the impact of the artificial introduction of this species into central Italy, leading to the near extinction of the endemic S. scardafa .     

Phylogeography of an Italian endemic salamander (genus Salamandrina): glacial refugia,postglacial expansions,and secondary contact

The Italian endemic genus Salamandrina has been historically regarded as monotypic but, recently, studies based on both mitochondrial and nuclear markers have indicated the existence of two distinct species of spectacled salamanders: Salamandrina perspicillata, in central and northern Italy, and Salamandrina terdigitata, in southern Italy. We analyzed nucleotide variation at mitochondrial and nuclear genes [cytochrome b, 12S and 16S rRNA, recombination activating gene (RAG 1)] in 223 individuals from 56 locations, aiming to investigate their genetic structure and recent evolutionary histories. Phylogenetic and phylogeographical analyses revealed the existence of three and two genetically distinct groups of populations in northern and southern salamander, respectively. Historical demographic analyses led to the inference of range expansion for both species in the late Pleistocene. During the last glacial stage, each salamander survived in a single refugium, namely the southern in Calabria and the northern in central Italy. At the end of this period, both lineages expanded northward and established secondary contact. Spatial distribution of RAG 1 haplotype variation revealed two differentiated population groups corresponding to the major mitochondrial (mt)DNA clades. Nuclear pattern of introgressive hybridization was more extensive than the highly limited introgression of mtDNA markers. From a conservation standpoint, southern Latium and Calabria proved to be the major genetic diversity reservoirs, thus deserving particular conservation efforts. © 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 104 , 903–922.     

Proposal of a data collection form to record dento-alveolar features--application to two Roman skeletal samples from Italy

Bones and teeth are an important source of information about the lifestyle, health status and dietary habits of past populations. Modifications and lesions of the periodontal tissues also provide information about the health status and behavior of an individual or of an ancient population. In this paper a method to record some dento-alveolar features of both deciduous and permanent teeth is described. The forms that we propose make it easy to record the data in a Microsoft Excel sheet. The form for adults has been applied to two Roman Imperial age skeletal samples from northern Italy - Casalecchio di Reno (II-V c. AD, Bologna) and Emilia Romagna, and central Italy - Quadrella (I-IV c. AD, Isernia, Molise).     

Small edge populations at risk: genetic diversity of the green lizard (<Emphasis Type="Italic">Lacerta viridis viridis</Emphasis>) in Germany and implications for conservation management

Edge and central populations can show great differences regarding their genetic variation and thereby also in their probability of extinction. This fact might be of great importance for the conservation strategies of endangered species. In this study we examine the level of microsatellite variability within three threatened edge populations of the green lizard subspecies Lacerta viridis viridis (Laur.) in Brandenburg (Germany) and compare the observed variation to other edge and central populations within the northern species range. We demonstrate that the northernmost edge populations contain less genetic variation in comparison to the central population. However, there were no observable significant differences to the other edge population included in this study. Surprisingly, we observed a high genetic differentiation in a small geographical range between the three endangered populations in Brandenburg, which can be explained by processes like fragmentation, isolation, genetic drift and small individual numbers within these populations. We also detected unique genetic variants (alleles), which only occurred in these populations, despite a low overall genetic variation. This study demonstrates the potential of fast evolving markers assessing the genetic status of endangered populations with a high resolution. It also illustrates the need for a comparative analysis of different regions within the species range, achieving a more exact interpretation of the genetic variation in endangered populations. This will aid future management decisions in the conservation of genetic diversity in threatened species.     

The phylogeographical clade trade: tracing the impact of human-mediated dispersal on the colonization of northern Europe by the oak gallwasp Andricus kollari

Human dispersal of organisms is an important process modifying natural patterns of biodiversity. Such dispersal generates new patterns of genetic diversity that overlie natural phylogeographical signatures, allowing discrimination between alternative dispersal mechanisms. Here we use allele frequency and DNA sequence data to distinguish between alternative scenarios (unassisted range expansion and long range introduction) for the colonization of northern Europe by an oak-feeding gallwasp, Andricus kollari. Native to Mediterranean latitudes from Portugal to Iran, this species became established in northern Europe following human introduction of a host plant, the Turkey oak Quercus cerris. Colonization of northern Europe is possible through three alternative routes: (i) unassisted range expansion from natural populations in the Iberian Peninsula; (ii) unassisted range expansion from natural populations in Italy and Hungary; or (iii) descent from populations imported to the UK as trade goods from the eastern Mediterranean in the 1830s. We show that while populations in France were colonized from sources in Italy and Hungary, populations in the UK and neighbouring parts of coastal northern Europe encompass allozyme and sequence variation absent from the known native range. Further, these populations show demographic signatures expected for large stable populations, rather than signatures of rapid population growth from small numbers of founders. The extent and spatial distribution of genetic diversity in the UK suggests that these A. kollari populations are derived from introductions of large numbers of individuals from each of two genetically divergent centres of diversity in the eastern Mediterranean. The strong spatial patterning in genetic diversity observed between different regions of northern Europe, and between sites in the UK, is compatible with leptokurtic models of population establishment.     

Inhibition and axial deviation of limb regeneration in the newt by means of a digit implanted into the amputated limb

This research was designed to follow up the observation of Thornton and Kraemer ('51) that regressed, denervated limbs of Ambystoma larvae will not regenerate upon reinnervation if all digits on the limbs were not completely resorbed. The object of this experiment was to determine whether the presence of an apical structure, protruding past the amputation surface, would affect the regenerative process. Both forearms of adult newts were amputated midway between the elbow and the wrist. One limb served as a normal regeneration control, and in the other limb the third digit from the removed hand was implanted in place of the removed radius, so that the three distal phalangeal segments protruded past the plane of amputation. Blastema formation in the experimental limbs was delayed by several weeks as compared with control limbs. Approximately one third of the experimental limbs did not regenerate. The regenerates that did form were strongly deviated (45–90°) radially from the longitudinal axis of the limb. Experimental analysis showed that the delay in regeneration is due largely to the projecting part of the digit. The radial deviation of the regenerates is not due to the digital implant, but rather to the removal of the radius. Trauma alone does not account for this phenomenon.     

Adaptive significance of amylase polymorphism in Drosophila I. The geographical pattern of allozyme polymorphism at the amylase locus in Drosophila subobscura

Allelic variation at the Amy locus was studied in eight natural populations from the central and northern range of D. subobscura, and the geographical pattern of Amy polymorphism over the range of this species was described. Even though regional and local differences in gene frequencies were found, in general the same alleles occur at high, intermediate and low frequencies, in nearly all populations. There are no significant differences in allele frequencies, but there is significant difference in the degree of heterozygosity among groups of populations from the northern, central and southern range. An analysis of population subdivision indicates that heterogeneity within populations is higher than between populations. Genetic distance values indicate that there is a variable degree of geographical differentiation between local populations. Variability within and between continental and insular populations is also discussed.     

Strong genetic impoverishment from the centre of distribution in southern Europe to peripheral Baltic and isolated Scandinavian populations of the pearly heath butterfly

Aim Climatic changes and fluctuations in the past have strongly influenced the distribution of animal and plant species. Such fluctuations are also reflected in the patterns of genetic diversity on both local and global scales. The genetic pattern of the pearly heath butterfly, Coenonympha arcania, was used to evaluate the genetic differentiation of isolated (in north‐western Europe), peripheral (in north‐eastern Europe) and central (in southern Europe) populations in the context of post‐glacial distributional changes of the species. Location Europe (Sweden, Germany, the Baltic states, Italy, Slovenia, Hungary, Romania, Bulgaria). Thus, samples were collected from large parts of the species’ distribution representing the three categories mentioned above. Methods We analysed 18 loci of 569 individuals from 28 populations by allozyme electrophoresis. We used both individual‐based and population‐based analyses, including F‐statistics, various clustering methods and Markov chain Monte Carlo simulations. Results All loci, except Fum, were polymorphic. The mean F_ST for all samples was 0.18. The mean genetic distance among populations was 0.046. Two major genetic lineages were distinguished. Populations from the centre of the distributional range in southern Europe and the northern periphery of the distributional range differed significantly in their level of genetic variability. The central populations of south‐eastern Europe showed high levels of genetic diversity and no differentiation among populations. Main conclusions Most probably the two major genetic lineages evolved during glacial isolation in two disjunct Mediterranean refugia. The lack of genetic differentiation across south‐eastern Europe implies a continuous Würm ice age distribution in this area, thus supporting the functional existence of steppe forests throughout this region. The peripheral‐isolated populations in Sweden seem to have suffered from one or more severe bottlenecks, resulting in substantial genetic impoverishment. The peripheral‐connected eastern Baltic populations, on the other hand, are affected by post‐glacial and possibly recurrent gene flow from more central parts of the distribution.     

North‐facing slopes and elevation shape asymmetric genetic structure in the range‐restricted salamander Plethodon shenandoah

Species with narrow environmental tolerances are often distributed within fragmented patches of suitable habitat, and dispersal among these subpopulations can be difficult to directly observe. Genetic data can help quantify gene flow between localities, which is especially important for vulnerable species with a disjunct range. The Shenandoah salamander (Plethodon shenandoah) is a federally endangered species known only from three mountaintops in Virginia, USA. To reconstruct the evolutionary history and population connectivity of this species, we generated both mitochondrial and nuclear data using sequence capture from individuals collected across all three mountaintops. Applying population and landscape genetic methods, we found strong population structure that was independent of geographic distance. Both the nuclear markers and mitochondrial genomes indicated a deep split between the most southern population and the genetically similar central and northern populations. Although there was some mitochondrial haplotype‐splitting between the central and northern populations, there was admixture in nuclear markers. This is indicative of either a recent split or current male‐biased dispersal among mountain isolates. Models of landscape resistance found that dispersal across north‐facing slopes at mid‐elevation levels best explain the observed genetic structure among populations. These unexpected results highlight the importance of incorporating landscape features in understanding and predicting the movement and fragmentation of this range‐restricted salamander species across space.     

The effect of biogeographic and phylogeographic barriers on gene flow in the brown smoothhound shark,Mustelus henlei,in the northeastern Pacific

We assessed the effects of the prominent biogeographic (Point Conception and the Peninsula of Baja California) and phylogeographic barriers (Los Angeles Region) of the northeastern Pacific on the population connectivity of the brown smoothhound shark, Mustelus henlei (Triakidae). Data from the mitochondrial control region and six nuclear microsatellite loci revealed significant population structure among three populations: northern (San Francisco), central (Santa Barbara, Santa Catalina, Punta Lobos, and San Felipe), and southern (Costa Rica). Patterns of long‐term and contemporary migration were incongruent, with long‐term migration being asymmetric and occurring in a north to south direction and a lack of significant contemporary migration observed between localities with the exception of Punta Lobos that contributed migrants to all localities within the central population. Our findings indicate that Point Conception may be restricting gene flow between the northern and central populations whereas barriers to gene flow within the central population would seem to be ineffective; additionally, a contemporary expansion of tropical M. henlei into subtropical and temperate waters may have been observed.     

Phylogeography of Scots pine in Europe and Asia based on mtDNA polymorphisms

We analyzed mitochondrial DNA polymorphisms to search for evidence of the genetic structure and patterns of admixture in 124 populations (N = 1407 trees) across the distribution of Scots pine in Europe and Asia. The markers revealed only a weak population structure in Central and Eastern Europe and suggested postglacial expansion to middle and northern latitudes from multiple sources. Major mitotype variants include the remnants of Scots pine at the north-western extreme of the distribution in the Scottish Highlands; two main variants (western and central European) that contributed to the contemporary populations in Norway and Sweden; the central-eastern European variant present in the Balkan region, Finland, and Russian Karelia; and a separate one common to most eastern European parts of Russia and western Siberia. We also observe signatures of a distinct refugium located in the northern parts of the Black Sea basin that contributed to the patterns of genetic variation observed in several populations in the Balkans, Ukraine, and western Russia. Some common haplotypes of putative ancient origin were shared among distant populations from Europe and Asia, including the most southern refugial stands that did not participate in postglacial recolonization of northern latitudes. The study indicates different genetic lineages of the species in Europe and provides a set of genetic markers for its finer-scale population history and divergence inference.     

Frequency distribution of mitochondrial DNA haplogroups in Corsica and Sardinia

Mitochondrial DNA (mtDNA) polymorphisms were analyzed by polymerase chain reaction amplification and haplogroup-specific restriction screening in populations from Corsica and Sardinia. These included 56 individuals from the area of Corte, central Corsica (France), 51 individuals from Gallura, northern Sardinia (Italy), and 45 individuals from Barbagia, central Sardinia. The screening revealed that about 95% of mtDNAs could be grouped in 8 of the 9 European haplogroups, including H-K, T-V, and X. Our results confirmed that these haplogroups encompass virtually all the mitochondrial lineages present in Europe and can be detected in both northern and southern European populations. We also discovered 2 restriction sites (-73 Alw441 and +75 SphI) that allow the detection of informative nucleotide changes in the second hypervariable segment of the control region, which help to detect the haplogroup identity of mtDNAs without requiring further DNA sequencing. Haplogroup H was the most common mtDNA lineage in this sample, reaching frequencies from about 40% in Corsican and Gallurese populations, to about 65% in the Barbagian population. Haplogroup V, possibly originating in the Iberian peninsula, was found only in the central Sardinian sample. Of the 5 Corsican mtDNAs belonging to the haplogroup T, 4 had a restriction fragment length polymorphism found only in this population. It seems that this mutation originated in Corsica and has had time to spread in the area, since the maternal grandmothers of the subjects came from different villages of the island. The sample from central Sardinia shows a remarkable discontinuity with those from the northern part of the island and from Corsica. Gallura and Corsica seem to have undergone a more recent peopling event, possibly related to the arrival of new mitochondrial variability from continental Italy, while Barbagia has apparently maintained more archaic haplotypes.     

Phylogeography, genetic diversity and conservation of the large copper butterfly Lycaena dispar in Europe

We argue that insect species conservation at large scales should take account of the distribution of genetic diversity among populations. Maintenance of genetic diversity may be vital in retaining a species' adaptive capacity and evolutionary potential. We illustrate the concept using the example of the large copper butterfly Lycaena dispar in Europe. This species has become extinct in parts of its range and is declining rapidly in others, whilst conversely, increasing in many areas. The latter has recently reduced its conservation status. Mitochondrial DNA analysis is used to construct a phylogeography from a preliminary sample set obtained from across Europe. A cytochrome b fragment of 402 base pairs was sequenced and 10 haplotypes were found. Relatedness among populations suggest that those from northern and central Europe are closely related and probably form one evolutionary significant unit (ESU) reflecting post-glacial colonization from southeast Europe. In contrast, the sample from Italy is divergent and should be considered a separate ESU. Our results, combined with ecological data, suggest that conservation action for this species should be targeted on specific regions and populations.     

Italian Peninsula preserves an evolutionary lineage of the fat dormouse Glis glis L. (Rodentia: Gliridae)

The present study examines the population genetic structure of fifty‐nine specimens of Glis glis (Linneaus, 1766) from thirteen localities in central Europe, sequencing a 400‐bp segment of the mitochondrial cytochrome b (cyt b) gene and a 673‐bp segment of the cytochrome c oxidase subunit I (COI) gene. The consensus tree obtained from Bayesian analysis revealed a robust dichotomy, showing two sister groups: one clade includes samples from a wide geographical area, extending from north‐central Europe to northern Italy (major branch sensu Bilton), and the other comprises samples collected in central and southern Italy and in Sicily (Italian branch). According to the Tajima–Nei model, the two phylogroups were separated by a sequence divergence of 0.8% (cyt b) – 2.6% (COI), showing the COI gene to be more informative than cyt b. On a smaller geographical scale, the Italian clade was further substructured, displaying geographical differentiation along the Peninsula. The gene pool in this area was patchy; whereas populations from Sicily Island demonstrated fixed cyt b and COI haplotypes, assuming processes of isolation and selection. © 2010 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 102 , 11–21.     

Constancy and change of basipodial variation patterns: a comparative study of crested and marbled newts —Triturus cristatus,Triturus marmoratus — and their natural hybrids

To test the concept of stable and lineage-specific developmental constraints the stability of biased limb skeletal variation patterns was examined by the comparative approach: 384 limbs of crested and marbled newts as well as their natural hybrids were cleared and stained for cartilage and bone. The variation in autopodium was recorded and compared between samples. The frequencies of variations were the same for crested and marbled newt samples. In the hybrids digital variation was significantly elevated, but the basipodium was as conservative as in the parental crested and marbled newts. The tarsus had to be excluded from further statistical analysis, since variation in tarsus was too rare to allow rigorous tests. The carpal variation was highly biased; out of twelve geometrically possible fusions of two neighbouring elements only three were found in the 219 forelimbs. The frequency of the fusions was the same in the parental as well as in the hybrid populations. In one sample of the marbled newt from a marginal population, the frequency of one fusion type was elevated, leading to an even more pronounced bias in phenotypic variation. Hence, neither in the hybrids nor in the marginal population could a breakdown of the constraint be observed. Also the pattern of variation was the same in all Triturus-samples examined here. The comparison of our results with within-population variation from other genera showed qualitative differences. The most common variant from our samples was found as a natural variant and even as the typical pattern in some species, but has not been observed in species of other genera. On, the other hand, fusions commonly observed in other genera were not observed in our sample. Even if all populations exhibit a highly bound variation pattern which may restrict the possible directions of morphological evolution, the direction of variation changes from genus to genus. It is suggested that shifts in constrained variation patterns have to be considered as key-events in the evolution of a clade. The cause of these shifts are unknown.     

Cranial trauma in iron age Samnite agriculturists, Alfedena, Italy: implications for biocultural and economic stress

The Samnites are an Iron Age protohistoric people from the central region of Italy. The skeletal remains are from the Alfedena necropolis, 6th through 5th centuries B.C. Macchiarelli et al. (Antropologia Contemporanea 4 (1981) 239-243) were the first to report on cranial trauma for this population, presenting four cases with extreme injuries. We re-examined this well documented skeletal population for additional examples of trauma. Previously unexamined remains from Alfedena, excavated at the turn of the 20th century, are also included in our analysis (Mariani. 1901. "Aufidena", ricerche archeologiche e storiche del Sannio settentrionale. Roma: Acc Naz Dei Lincei). Of the 209 adult crania examined, 12.9% of them exhibited trauma. Analysis of location and frequency of cranial trauma revealed that cranial injuries to the head appear to originate from all directions. The high rate of cranial trauma underscores the violent circumstances experienced during the Iron Age protohistoric period of central Italy. Males are much more likely to exhibit cranial injury than females (P = 0.009). We conclude that the injuries received by Samnite male farmer-warriors occurred while defending pastoral-agricultural resources. Trauma rates are similar for some Iron Age populations and not for others. Behavior associated with violence during the Iron Age period can not be generalized for all populations found in Italy.     

Genetic diversity and historical population structure in the New Zealand mayfly Acanthophlebia cruentata

1. Nucleotide sequences of a 280 base pair region of the cytochrome b gene were used to assess genetic diversity and to infer population histories in the New Zealand mayfly Acanthophlebia cruentata. 2. A hierarchial examination of populations from 19 streams at different spatial scales in the central and northern North Island of New Zealand found 34 haplotypes. A common haplotype was found in all central region streams and unique haplotypes in northern streams. Several central streams had region specific haplotypes with genetically differentiated populations at the 70–100 km scale. 3. Haplotype diversity was high (0.53–0.8) at most sites, but low (0–0.22) in some central sites. amova analyses found significant genetic diversity among regions (69%) and among catchments (58%). Most population pairwise F_ST tests were significant, with non‐significant pairwise tests among sites in the central region and pairs of sites between neighbouring streams. 4. The levels of sequence divergence are interpreted as the result of Pleistocene divergence in multiple refugia, leading to the evolution of regionally unique haplotypes. The low diversity in some central region populations may result from recent colonisation following local extinctions, associated with volcanic events.