Historic cycles of fragmentation and expansion in Parnassius smintheus (papilionidae) inferred using mitochondrial DNA

Climate oscillations of the Quaternary drove the repeated expansion and contraction of ecosystems. Alpine organisms were probably isolated in sky island refugia during warm interglacials, such as now, and expanded their range by migrating down-slope during glacial periods. We used population genetic and phylogenetic approaches to infer how paleoclimatic events influenced the distribution of genetic variation in the predominantly alpine butterfly Parnassius smintheus. We sequenced a 789 bp region of cytochrome oxidase I for 385 individuals from 20 locations throughout the Rocky Mountains, ranging from southern Colorado to northern Montana. Analyses revealed at lease two centers of diversity in the northern and southern Rocky Mountains and strong population structure. Nested clade analysis suggested that the species experienced repeated cycles of population expansion and fragmentation. The estimated ages of these events, assuming a molecular clock, corresponded with paleoclimatic data on habitat expansion and contraction over the past 400,000 years. We propose that alpine butterflies persisted in an archipelago of isolated sky islands during interglacials and that populations expanded and became more connected during cold glacial periods. An archipelago model implies that the effects of genetic drift and selection varied among populations, depending on their latitude, area, and local environment. Alpine organisms are sensitive indicators of climate change and their history can be used to predict how high-elevation ecosystems might respond to further climate warming.     

Pleistocene origin and population history of a neoendemic alpine butterfly

Alpine environments underwent dramatic transformation during glacial–interglacial cycles, with the consequence that geographical, ecological and demographic changes of alpine populations provided the opportunity for formation of neoendemic species. Several biogeographical models have been proposed to account for the unique history of alpine populations, with different expectations of genetic divergence and speciation. The expanding alpine archipelago model proposes that alpine populations expand spatially and demographically during glacial events, dispersing between mountain ranges. Under this model, alpine populations are unlikely to diverge in isolation due to substantial interpopulation gene flow. In contrast, the alpine archipelago refuge model proposes that gene flow during glacial phases is limited and populations expand demographically during interglacial phases, increasing genetic isolation and the likelihood of speciation. We assess these models by reconstructing the evolutionary history of Colias behrii, a morphologically and ecologically distinct alpine butterfly restricted to the California Sierra Nevada. C. behrii exhibits very low genetic diversity at mitochondrial and nuclear loci, limited population structure and evidence of population expansion. C. behrii and Rocky Mountain C. meadii share identical mitochondrial haplotypes, while in contrast, nuclear data indicate common ancestry between C. behrii and Cascades Range Colias pelidne. The conflict in gene genealogies may be a result of recent expansion in North American Colias, but an isolation with migration analysis indicates that genetic patterns in C. behrii might result from differential introgression following hybridization. Based on the timing of population expansion and gene flow between mountain ranges, the expanding alpine archipelago model is supported in C. behrii.     

Using coalescent simulations to test the impact of quaternary climate cycles on divergence in an alpine plant-insect association

The Quaternary climate cycles forced species to repeatedly migrate across a continually changing landscape. How these shifts in distribution impacted the evolution of unrelated but ecologically associated taxa has remained elusive due to the stochastic nature of the evolutionary process and variation in species-specific biological characteristics and environmental constraints. To account for the uncertainty in genealogical estimates, we adopted a coalescent approach for testing hypotheses of population divergence in coevolving taxa. We compared genealogies of a specialized herbivorous insect, Parnassius smintheus (Papilionidae), and its host plant, Sedum lanceolatum (Crassulaceae), from the alpine tundra of the Rocky Mountains to null distributions from coalescent simulations to test whether tightly associated taxa shared a common response to the paleoclimatic cycles. Explicit phylogeographic models were generated from geologic and biogeographic data and evaluated over a wide range of divergence times given calibrated mutation rates for both species. Our analyses suggest that the insect and its host plant responded similarly but independently to the climate cycles. By promoting habitat expansion and mixing among alpine populations, glacial periods repeatedly reset the distributions of genetic variation in each species and inhibited continual codivergence among pairs of interacting species.     

Treeline proximity alters an alpine plant�Cherbivore interaction

Rising treeline threatens the size and contiguity of alpine meadows worldwide. As trees encroach into previously open habitat, the movement and population dynamics of above-treeline alpine species may be disrupted. This process is well documented in studies of the Rocky Mountain apollo butterfly (Parnassius smintheus). However, subtler consequences of treeline rise remain poorly understood. In this study, we examine whether treeline proximity affects feeding behaviour of P. smintheus larvae, due to altered habitat affecting the distribution and availability of their host plant, lance-leaved stonecrop (Sedum lanceolatum). Understanding differential larval exploitation of food resources in relation to the treeline is an important step in predicting the consequences of continued treeline rise. Parnassius smintheus larvae feed more intensively on S. lanceolatum away from the treeline despite the relative paucity of hosts in these areas, and despite higher fitness penalties associated with the plant's herbivory-induced chemical defenses. Sedum lanceolatum growing near the treeline is less attractive, and therefore represents a less significant resource for P. smintheus larvae than its abundance might imply. If treeline rise continues, we suggest that this pattern of altered resource exploitation may represent a mechanism by which larvae are adversely affected even while adult movement among and within meadows appears sufficient for maintaining population health, and total host availability seems ample.     

Influence of landscape on the population genetic structure of the alpine butterfly parnassius smintheus (Papilionidae)

Four microsatellite DNA markers were developed which were used to examine the relationship between landscape and population genetic structure among a set of populations of the butterfly Parnassius smintheus located in the foothills of the Canadian Rockies. Detailed information on the dispersal of adult butterflies among this same set of populations was available. Simple and partial Mantel tests were used to examine the relationships between genetic distances, predicted rates of dispersal, and a number of landscape variables, all measured pairwise for 17 sample sites. Nei's standard genetic distance was negatively correlated with predicted dispersal. We observed a significant pattern of isolation by distance at a very small spatial scale. The distance between sites that was through forest was a stronger predictor of genetic distance than the distance through open meadow, indicating a significant effect of landscape on population genetic structure beyond that of simple isolation by distance. Our results suggest that rises in the tree-line in alpine areas, caused by global warming, will lead to reduced gene flow among populations of P. smintheus.     

Molecular phylogeny of butterflies Parnassius glacialis and P. stubbendorfii at various localities in East Asia.

The phylogeny of butterflies, Parnassius stubbendorfii and P. glacialis, collected at various localities in the Japan archipelago and the eastern part of the Asian continent was analyzed using mitochondrial DNA sequences coding for NADH dehydrogenase subunit 5 (805 bp). The molecular phylogenetic trees revealed that P. glacialis and P. stubbendorfii diverged from a common ancestor, and then the populations inhabiting the Japan archipelago and the Asian continent diverged in each species. The reliability of these divergences was supported by high bootstrap values. The divergences within the Japan archipelago and within the Asian continent in each species were unclear because of low bootstrap values. The genetic distance and a rough time-estimation in the UPGMA tree suggest that the both populations of P. glacialis and P. stubbendorfii may have been isolated in the Japan archipelago at the early time (about 1.7-2.0 Mya) of the glacial period in the Pleistocene. The genetic distance between the Japanese and the continental subspecies may be large enough that they can be classified as different species, in comparison with the genetic distances among some other parnassian species.     

An experimental examination of the effects of habitat quality on the dispersal and local abundance of the butterfly Parnassius smintheus

Abstract 1. Nectar flower abundance was manipulated through flower removal, and sex ratio was manipulated by moving individual butterflies within a series of nine alpine meadows. The movement and abundance of the butterfly Parnassius smintheus in the meadows were monitored using mark–release–recapture methods.
2. A total of 937 butterflies, 698 males and 239 females, was captured. There were 223 observed between-meadow movements. Fifty-two per cent of males and 35% of females moved among meadows.
3. The immigration of male butterflies was related positively to nectar flowers, host plant abundance, and female butterflies. Male emigration was not affected by any of the treatments. The number of males captured was related positively to nectar flowers and host plants but not affected by sex ratio. The number of resident male butterflies was greater in meadows containing flowers and was related positively to host plant abundance, but unaffected by sex ratio.
4. Flower removal, sex ratio, and abundance of Sedum had no significant effect on the abundance, movement, or residence time for female butterflies, in part due to small sample size.
5. The fact that males immigrate to higher quality meadows suggests that male butterflies are assessing meadow quality, either by sampling meadows or potentially from a distance using olfactory cues.     

A range-wide genetic bottleneck overwhelms contemporary landscape factors and local abundance in shaping genetic patterns of an alpine butterfly (Lepidoptera: Pieridae: Colias behrii)

Spatial and environmental heterogeneity are major factors in structuring species distributions in alpine landscapes. These landscapes have also been affected by glacial advances and retreats, causing alpine taxa to undergo range shifts and demographic changes. These nonequilibrium population dynamics have the potential to obscure the effects of environmental factors on the distribution of genetic variation. Here, we investigate how demographic change and environmental factors influence genetic variation in the alpine butterfly Colias behrii. Data from 14 microsatellite loci provide evidence of bottlenecks in all population samples. We test several alternative models of demography using approximate Bayesian computation (ABC), with the results favouring a model in which a recent bottleneck precedes rapid population growth. Applying independent calibrations to microsatellite loci and a nuclear gene, we estimate that this bottleneck affected both northern and southern populations 531–281 years ago, coinciding with a period of global cooling. Using regression approaches, we attempt to separate the effects of population structure, geographical distance and landscape on patterns of population genetic differentiation. Only 40% of the variation in F_ST is explained by these models, with geographical distance and least‐cost distance among meadow patches selected as the best predictors. Various measures of genetic diversity within populations are also decoupled from estimates of local abundance and habitat patch characteristics. Our results demonstrate that demographic change can have a disproportionate influence on genetic diversity in alpine species, contrasting with other studies that suggest landscape features control contemporary demographic processes in high‐elevation environments.     

Late Pleistocene origin of the entire circumarctic range of the arctic‐alpine plant Kalmia procumbens

The circumarctic ranges of arctic‐alpine plants are thought to have been established in the late Pliocene/early Pleistocene, when the modern arctic tundra was formed in response to climate cooling. Previous findings of range‐wide genetic structure in arctic‐alpine plants have been thought to support this hypothesis, but few studies have explicitly addressed the temporal framework of the genetic structure. Here, we estimated the demographic history of the genetic structure in the circumarctic Kalmia procumbens using sequences of multiple nuclear loci and examined whether its genetic structure reflects prolonged isolation throughout the Pleistocene. Both Bayesian clustering and phylogenetic analyses revealed genetic distinction between alpine and arctic regions, whereas detailed groupings were somewhat discordant between the analyses. By assuming a population grouping based on the phylogenetic analyses, which likely reflects a deeper intraspecific divergence, we conducted model‐based analyses and demonstrated that the intraspecific genetic divergence in K. procumbens likely originated during the last glacial period. Thus, there is no need to postulate range separation throughout the Pleistocene to explain the current genetic structure in this species. This study demonstrates that range‐wide genetic structure in arctic‐alpine plants does not necessarily result from the late Pliocene/early Pleistocene origin of their circumarctic ranges and emphasizes the importance of a temporal framework of the current genetic structure for understanding the biogeographic history of the arctic flora.     

Pleistocene speciation is not refuge speciation

A number of researchers working on the origin of extant Neotropical biodiversity implicitly and without appropriate proofs assume that Pleistocene speciation should necessarily follow the rules of the refuge hypothesis. A recent example is provided by a study of Neotropical butterflies. Although the analysis showed that these groups experienced their main diversification burst during the last 2.6 million years, coinciding with the Pleistocene glacial cycles (Garzón‐Orduña et al., 2014, Journal of Biogeography, 41 , 1631–1638), a causal link between the speciation chronology and the evolutionary mechanisms proposed by the refuge hypothesis is not provided. Without more detailed studies on the environmental drivers, geographical patterns and speciation modes, establishing a causal link between speciation chronology and a particular speciation model – of which the refuge hypothesis is only one among many possibilities – is too speculative. Here I provide a six‐step conceptual framework for linking the speciation chronology with the environmental drivers and the ecological and evolutionary mechanisms potentially involved.     

Phylogeography of the prehensile-tailed skink Corucia zebrata on the Solomon Archipelago

The biogeography of islands is often strongly influenced by prior geological events. Corucia zebrata (Squamata: Scincidae) is endemic to the geologically complex Solomon Archipelago in Northern Melanesia. We examined the level of divergence for different island populations of C. zebrata and discussed these patterns in light of Pleistocene land bridges, island isolation, and island age. Corucia zebrata was sampled from 14 locations across the Solomon Archipelago and sequenced at two mitochondrial genes (ND2 and ND4; 1697 bp in total) and four nuclear loci (rhodopsin, an unknown intron, AKAP9, and PTPN12). Measures of genetic divergence, analyses of genetic variation, and Bayesian phylogenetic inference were used and the data assessed in light of geological information. Populations of C. zebrata on separate islands were found to be genetically different from each other, with reciprocal monophyly on mitochondrial DNA. Populations on islands previously connected by Pleistocene land bridges were marginally less divergent from each other than from populations on other nearby but isolated islands. There are indications that C. zebrata has radiated across the eastern islands of the archipelago within the last 1-4 million years. Nuclear loci were not sufficiently informative to yield further information about the phylogeography of C. zebrata on the Solomon Archipelago. Analyses of the mitochondrial data suggest that dispersal between islands has been very limited and that there are barriers to gene flow within the major islands. Islands that have been isolated during the Pleistocene glacial cycles are somewhat divergent in their mitochondrial genotypes, however, isolation by distance (IBD) and recent colonization of isolated but geologically younger islands appear to have had stronger effects on the phylogeography of C. zebrata than the Pleistocene glacial cycles. This contrasts with patterns reported for avian taxa, and highlights the fact that biogeographic regions for island species cannot be directly extrapolated among taxa of differing dispersal ability.     

Winter is coming: Miocene and Quaternary climatic shifts shaped the diversification of Western‐Mediterranean Harpactocrates (Araneae,Dysderidae) spiders

Past climatic shifts have played a major role in generating and shaping biodiversity. Quaternary glacial cycles are the better known examples of dramatic climatic changes endured by ecosystems in temperate regions. Although still a matter of debate, some authors suggest that glaciations promoted speciation. Here we investigate the effect of past climatic changes on the diversification of the ground‐dwelling spider genus Harpactocrates, distributed across the major mountain ranges of the western Mediterranean. Concatenated and species‐tree analyses of multiple mitochondrial and nuclear loci, combined with the use of fossil and biogeographic calibration points, reveal a Miocene origin of most nominal species, but also unravel several cryptic lineages tracing back to the Pleistocene. We hypothesize that the Miocene Climatic Transition triggered major extinction events in the genus but also promoted its subsequent diversification. Under this scenario, the Iberian mountains acted as an island‐like system, providing shelter to Harpactocrates lineages during the climate shifts and favouring isolation between mountain ranges. Quaternary glacial cycles contributed further to the diversification of the group by isolating lineages in peripheral refugia within mountain ranges. In addition, we recovered some unique biogeographic patterns, such as the colonization of the Alps and the Apennines from the Iberian Peninsula.     

Tracking the ice: Subterranean harvestmen distribution matches ancient glacier margins

Biogeographic studies often underline the role of glacial dynamism during Pleistocene (1.806–0.011 Mya) in shaping the distribution of subterranean species. Accordingly, it is presumed that present‐day distribution of most specialized cold‐adapted (cryophilic) cave‐dwelling species should bear the signatures of past climatic events. To test this idea, we modelled the distribution of specialized cold‐adapted subterranean alpine harvestmen (Arachnida: Opiliones: Ischyropsalididae: Ischyropsalis). We found that the distance from the glacier margins during Last Glacial Maximum (LGM; about 22,000 years ago) was the most important predictor of their present‐day distribution. In particular, the peak in the probability of occurrence of alpine subterranean Ischyropsalis was found to be in close proximity to the LGM glacier, with a sharp drop at a distance of 30 km from the ice margin. In light of the role played by past climatic events in determining the species current range, we briefly discuss their biogeographic history and the role played by glacial refugia dynamics in determining the current distribution of these species. We argue that low dispersal harvestmen such as our model species can be used as biological indicators for tracking past glaciations and other similar biogeographic events.     

Testing hypotheses of Pleistocene population history using coalescent simulations: phylogeography of the pygmy nuthatch (Sitta pygmaea)

In this paper, we use mitochondrial NADH dehydrogenase subunit 2 sequences to test Pleistocene refugial hypotheses for the pygmy nuthatch (Sitta pygmaea). Pygmy nuthatches are a common resident of long-needle pine forests in western North America and demonstrate a particular affinity with ponderosa pine (Pinus ponderosa). Palaeoecological and genetic data indicate that ponderosa pine was isolated in two Pleistocene refugia corresponding to areas in the southern Sierra Nevada in the west and southern Arizona and New Mexico in the east. We use coalescent simulations to test the hypothesis that pygmy nuthatches tracked the Pleistocene history of their preferred habitat and persisted in two refugia during the periods of glacial maxima. Coalescent simulation of population history does not support the hypothesis of two Pleistocene refugia for the pygmy nuthatch. Instead, our data are consistent with a single refuge model. Nucleotide diversity is greatest in the western populations of southern and coastal California. We suggest that the pygmy nuthatch expanded from a far western glacial refuge into its current distribution since the most recent glacial maximum.     

Plant molecular phylogeography in China and adjacent regions: Tracing the genetic imprints of Quaternary climate and environmental change in the world's most diverse temperate flora

The Sino-Japanese Floristic Region (SJFR) of East Asia harbors the most diverse of the world's temperate flora, and was the most important glacial refuge for its Tertiary representatives ('relics') throughout Quaternary ice-age cycles. A steadily increasing number of phylogeographic studies in the SJFR of mainland China and adjacent areas, including the Qinghai-Tibetan-Plateau (QTP) and Sino-Himalayan region, have documented the population histories of temperate plant species in these regions. Here we review this current literature that challenges the oft-stated view of the SJFR as a glacial sanctuary for temperate plants, instead revealing profound effects of Quaternary changes in climate, topography, and/or sea level on the current genetic structure of such organisms. There are three recurrent phylogeographic scenarios identified by different case studies that broadly agree with longstanding biogeographic or palaeo-ecological hypotheses: (i) postglacial re-colonization of the QTP from (south-)eastern glacial refugia; (ii) population isolation and endemic species formation in Southwest China due to tectonic shifts and river course dynamics; and (iii) long-term isolation and species survival in multiple localized refugia of (warm-)temperate deciduous forest habitats in subtropical (Central/East/South) China. However, in four additional instances, phylogeographic findings seem to conflict with a priori predictions raised by palaeo-data, suggesting instead: (iv) glacial in situ survival of some hardy alpine herbs and forest trees on the QTP platform itself; (v) long-term refugial isolation of (warm-)temperate evergreen taxa in subtropical China; (vi) 'cryptic' glacial survival of (cool-)temperate deciduous forest trees in North China; and (vii) unexpectedly deep (Late Tertiary/early-to-mid Pleistocene) allopatric-vicariant differentiation of disjunct lineages in the East China-Japan-Korea region due to past sea transgressions. We discuss these and other consequences of the main phylogeographic findings in light of palaeo-environmental evidence, emphasize notable gaps in our knowledge, and outline future research prospects for disentangling the evolution and biogeographic history of the region's extremely diverse temperate flora.     

A multivariate approach to the determination of faunal structures among European butterfly species (Lepidoptera: Rhopalocera)

Multivariate analyses of 393 butterfly species over 85 geographical areas (R- and Q-data matrices) in Europe and North Africa have produced a consistent pattern of faunal structures (units and regions). Prominent features to emerge are the latitudinal zonation of geographical units and the division of the Mediterranean into western and eastern components; southwards in Europe, endemicity increases whereas faunal structures decrease in spatial dimensions. Central Europe–from the Urals to the British Isles–forms a single large faunal structure (extent unit and region). A model has been constructed to account for Pleistocene evolutionary changes and endemism in European butterflies and for the east-west taxonomic divisions in the extent faunal structure which dominates central Europe. Periodic Pleistocene climatic changes have resulted in cycles of population extinction, isolation, evolution and migration, but the nature and timing of events has depended on the environmental tolerances of species belonging to different faunal units. During Pleistocene glaciations, southern species have been relatively static and more isolated and have evolved independently. By comparison, northern species have been more mobile and have migrated over large distances. Contact and hybrid zones among cosmopolitan species in northern Europe are probably of some antiquity. They result from persistent survival and isolation of refuge populations in the west and east Mediterranean during glacial phases; dispersal from these refuges leads to their regeneration during each interglacial.     

Historical biogeography of two alpine butterflies in the Rocky Mountains: broad-scale concordance and local-scale discordance

Aim We inferred the phylogeography of the alpine butterfly Colias meadii Edwards (Pieridae) and compared its genetic structure with that of another high elevation, co‐distributed butterfly, Parnassius smintheus Doubleday (Papilionidae), to test if the two Rocky Mountain butterflies responded similarly to the palaeoclimatic cycles of the Quaternary. Location Specimens were collected from 18 alpine sites in the Rocky Mountains of North America, from southern Colorado to northern Montana. Methods We sequenced 867 and 789 nucleotides of cytochrome oxidase I from an average of 19 and 20 individuals for C. meadii and P. smintheus, respectively, from each of the same 18 localities. From the sequence data, we calculated measures of genetic diversity within each population (H, θ), genetic divergence among populations (F_ST), and tested for geographic structure through an analysis of molecular variance (amova ). Population estimates were compared against latitude and between species using a variety of statistical tests. Furthermore, nested clade analysis was implemented to infer historic events underlying the geographic distribution of genetic variation in each species. Then, we compared the number of inferred population events between species using a nonparametric Spearman's rank correlation test. Finally, we ran coalescent simulations on each species’ genealogy to test whether the two species of Lepidoptera fit the same model of population divergence. Results Our analyses revealed that: (1) measures of within‐population diversity were not correlated with latitude for either species, (2) within‐site diversity was not correlated between species, (3) within a species, nearly all populations were genetically isolated, (4) both species exhibited significant and nearly identical partitioning of genetic variation at all hierarchical levels of the amova , including a strong break between populations across the Wyoming Basin, (5) both species experienced similar cycles of expansion and contraction, although fewer were inferred for C. meadii, and (6) data from both species fit a model of three refugia diverging during the Pleistocene. Main conclusions While our findings supported a shared response of the two butterfly species to historic climate change across coarse spatial scales, a common pattern was not evident at finer spatial and temporal scales. The shared demographic history of the two species is consistent with an expanding–contracting archipelago model, suggesting that populations persisted across the geographic range throughout the climate cycles, experiencing isolation on ‘sky islands’ during interglacial periods and becoming connected as they migrated down‐slope during cool, wet climates.     

High mountains of the Japanese archipelago as refugia for arctic–alpine plants: phylogeography of Loiseleuria procumbens (L.) Desvaux (Ericaceae)

According to previous phylogeographic studies, high mountains at low latitudes are important areas for the study of the evolutionary history of arctic–alpine plants in surviving the Pleistocene climatic oscillations. To evaluate this hypothesis, we elucidated the genetic structure of the arctic–alpine plant, Loiseleuria procumbens , in the Japanese archipelago, which corresponds to one of the southernmost limits of its distribution, using 152 individuals from 17 populations that covered the entire distribution of the Japanese archipelago and Sakhalin, in addition to samples from Sweden. Based on 854 bp of chloroplast DNA, we detected eight haplotypes. Along with haplotype distribution, strong genetic differentiation between populations in central and northern Japan was elucidated by a neighbour-joining tree (100%) and spatial analysis of molecular variance (79%), which is consistent with other alpine plants in Japan, regardless of the species' range. In addition, the southernmost populations from northern Japan showed specific genetic structure, although the remaining areas of northern Japan and Sakhalin harboured an homogenous genetic structure. Our results suggest that the populations in central Japan persisted for a long time during the Pleistocene climatic oscillation and that genetic divergence occurred in situ , supporting our hypothesis in conjunction with a previous study of another arctic–alpine plant, Diapensia lapponica subsp. obovata .  © 2009 The Linnean Society of London, Biological Journal of the Linnean Society , 2009, 97 , 403–412.     

Population structure within an alpine archipelago: strong signature of past climate change in the New Zealand rock wren (Xenicus gilviventris)

Naturally subdivided populations such as those occupying high‐altitude habitat patches of the ‘alpine archipelago’ can provide significant insight into past biogeographical change and serve as useful models for predicting future responses to anthropogenic climate change. Among New Zealand's alpine taxa, phylogenetic studies support two major radiations: the first correlating with geological forces (Pliocene uplift) and the second with climatic processes (Pleistocene glaciations). The rock wren (Xenicus gilviventris) is a threatened alpine passerine belonging to the endemic New Zealand wren family (Acanthisittidae). Rock wren constitute a widespread, naturally fragmented population, occurring in patches of suitable habitat over c. 900 m in altitude throughout the length of the South Island, New Zealand. We investigated the relative role of historical geological versus climatic processes in shaping the genetic structure of rock wren (N = 134) throughout their range. Using microsatellites combined with nuclear and mtDNA sequence data, we identify a deep north–south divergence in rock wren (3.7 ± 0.5% at cytochrome b) consistent with the glacial refugia hypothesis whereby populations were restricted in isolated refugia during the Pleistocene c. 2 Ma. This is the first study of an alpine vertebrate to test and provide strong evidence for the glacial refugia hypothesis as an explanation for the low endemicity central zone known as the biotic ‘gap’ in the South Island of New Zealand.