Evolution of the New Zealand Mountain Flora

ARTICLE IN PRESS

Organisms, Diversity & Evolution 5 (2005) 237–247 www.elsevier.de/ode

Evolution of the New Zealand mountain ﬂora: Origins, diversiﬁcation and dispersal RichardC. Winkworth a,Ã, Steven J. Wagstaffb, DavidGlenny b, Peter J. Lockhartc aDepartment of Ecology and Evolutionary Biology, Yale University, P.O. Box 208105, New Haven, CT 06520-8105, USA bManaaki Whenua Landcare Research, P.O. Box 69, Lincoln 8152, New Zealand cThe Allan Wilson Centre for Molecular Ecology and Evolution, Massey University, Private Bag 11222, Palmerston North, New Zealand

Received11 May 2004; accepted16 December 2004

Abstract

The New Zealand mountains provide a unique system in which to study the evolution of alpine plants. The relationship between the recent uplift of mountain habitats (5–2 million year ago (mya)) andfloristic diversity has polarizedhypotheses on the evolution of the alpine flora; suggestions have rangedfrom an ancient history in New Zealand to recent arrival by long distance dispersal from the Northern Hemisphere. Molecular phylogenies are now available for numerous New Zealandalpine plant groups andthese provideinsights into the evolution of this unique flora. Taken together with the fossil record, these studies suggest that many alpine lineages first arrived in New Zealand during the late Tertiary and subsequent radiations accompanied environmental upheaval in the Pliocene and Pleistocene. Ongoing studies are investigating the processes that contribute to morphological and ecological diversity in the New Zealandalpine flora. r 2005 Gesellschaft fu¨ r Biologische Systematik. Publishedby Elsevier GmbH. All rights reserved.

Keywords: Biogeography; Dispersal; Ecological diversiﬁcation; Morphological radiation; New Zealand alpine ﬂora

Introduction The New Zealandbiota has developedonan ancient continental landmass that was progressively isolated ‘‘With regardto general problems of biogeography, during the late Cretaceous and early Tertiary. Originally the biota of New Zealandhas been, perhaps, the most linkedto the Gondwanan supercontinent, New Zealand important of any in the world. It has ﬁgured promi- has been separatedfrom its closest continental neighbor, nently in all discussions of austral biogeography, and all Australia, by 2000 km for the last 60 million years (my). notable authorities have felt obligedto explain its The long isolation of New Zealandis reﬂectedin the history: explain New Zealandandthe worldfalls into generally oceanic character of its biota, andin this place aroundit’’ ( Nelson 1975). respect it is similar to Hawaii andthe Gala ´ pagos. However, New Zealandhas a much olderheritage than these strictly oceanic archipelagos, with a long pre- ÃCorresponding author. Tel.: +1 (203) 436 4992; history of dramatic environmental change that has fax: +1 (203) 432 3854. helpedshape its biota. More complex than most oceanic E-mail address: [email protected] (R.C. Winkworth). islands yet more tractable than many continental

1439-6092/$ - see front matter r 2005 Gesellschaft fu¨ r Biologische Systematik. Publishedby Elsevier GmbH. All rights reserved. doi:10.1016/j.ode.2004.12.001 ARTICLE IN PRESS 238 R.C. Winkworth et al. / Organisms, Diversity & Evolution 5 (2005) 237–247 systems, New Zealandprovidesunique opportunities to study evolution. The biogeographic history of New Zealandhas, as Nelson (1975) points out, attractedthe attention of many biogeographers. In this regard, the flora of its mountains (referredto here as the alpine flora) has been one of the most extensively studied elements. However, despite numerous detailed studies and several syntheses, hypotheses about the origins anddiversification of the New Zealandmountain flora have remainedpolarized– largely due to contrasting views on the relative importance of dispersal, geological change, and climatic fluctuations (McGlone et al. 2001). Recent advances in our knowledge of New Zealand’s geological and climatic history, andphylogenetic studies for a growing number of lineages have provided important insights into the origins andevolution of the alpine flora. In this review we have two goals: (i) to discuss insights from recent molecular phylogenetic studies, and (ii) to highlight Fig. 1. Extent of alpine areas in New Zealandandalpine plant distribution patterns. Areas above alpine tree line are indicated avenues for future research. As a framework for this as black silhouettes; montane habitats are more extensive, discussion, we first briefly summarize the alpine setting especially in the North Island. Gray lines indicate generalized in New Zealandandbiogeographic hypotheses concern- boundaries for two endemic-rich centers and an intervening ing the alpine flora. endemic-poor gap. Several individual mountains or mountain ranges described in the text are marked.

There is little doubt that mountains are a geologically Discussion recent feature in New Zealand. For much of the last 85 my New Zealandhas been an isolatedlandmass of The New Zealand mountains limitedtopographic relief. Having broken away from Gondwana in the late Mesozoic, New Zealand con- Mountains dominate much of the New Zealand tinuedto separate from Australia until the early landscape. With the exception of the central North Paleocene. Throughout the late Cretaceous andearly Islandvolcanoes, the alpine zone is restrictedto a single, Tertiary the ancestral New Zealandlandmass suffered ruggedmountain chain that extendsfrom Stewart Island steady erosion. By the latest Oligocene only a scattered in the south, along the Southern Alps of the South archipelago of low-lying islands remained (Cooper and Island, and through the North Island to East Cape Cooper 1995). The general form of contemporary New (Fig. 1). Although largely contiguous, the New Zealand Zealandis the result of tectonic uplift that began axial mountains vary considerably in geomorphology, approximately 25 million year ago (mya) with the with five regions recognizedbasedon the underlying activation of the modern Pacific-Australian plate parent material. For example, the eastern mountains of boundary (Cooper et al. 1987; Kamp 1992). Miocene the Southern Alps (e.g., the Craigieburn Range) are tectonism ledto rapidchanges in distribution and composedof greywacke andargillite rocks. These topography, but the general uplift trendresultedin a formations underwent complex faulting during their narrow, elongate landmass of rugged relief. It was not uplift andare highly prone to frost shattering that until the Pliocene (5–2 mya) that the formation of the results in the extensive talus andscree slopes Alpine Fault system andthe increasing tempo of characteristic of this region. In contrast, folding and tectonic activity ledto the rapidelevation of the axial tilting of crustal sections formedthe Otago mountains in New Zealand( Ollier 1986; Cox and mountains (e.g., the OldMan Range). These Findlay 1995; Batt et al. 2000). Although the overall more rounded, undulating mountains are composed of form of these mountains reflects the original pattern of harder schist rocks and were not subjected to tectonism, many details of the contemporary relief are extensive glacial erosion (Holloway 1982; Whitehouse the result of extensive modification following uplift. In andPearce 1992 ). Such geological differences, as particular, the alpine landscape of the South Island was well as striking differences in past and present heavy modified by glaciation during the latest Pliocene climates, have resultedin a diversearray of habitat andPleistocene. From about 2.5 mya large areas of the types in the New Zealandmountains ( Whitehouse and Southern Alps were repeatedly glaciated, and many Pearce 1992). contemporary landforms relate directly to glacial ARTICLE IN PRESS R.C. Winkworth et al. / Organisms, Diversity & Evolution 5 (2005) 237–247 239 modeling or modification of glacial constructions. 1978). In contrast, Wardle (1963) and Fleming (1963) Although glaciation was more limitedin the North speculatedthat Tertiary New Zealandwouldnot have Island, cold climate processes, such as periglacial mass provided suitable environments for cool temperate movement, have strongly influencedlower North Island plants. Insteadthey suggestedthat ancestors of modern landscapes (Beu and Edwards 1984; McGlone 1985; alpine lineages may have spent this periodin cool Suggate 1990; Pillans et al. 1992; Markgraf et al. 1995). temperate areas to the south of New Zealand, re- Post-glacial erosion has also playedan important role in establishing in New Zealandonly after suitable habitats shaping the modern alpine landscape (Suggate 1990; became available in the late Pliocene andPleistocene. Pillans et al. 1992). Wardle (1963) hypothesizedthat a southern extension to The most distinctive North Island mountains are the New Zealand(or perhaps islandsto the south; see four volcanic peaks, three on the Central Plateau (Mt. Dawson 1988) wouldhave providedcool climates, Ruapehu, Mt. Ngauruhoe, andMt. Tongariro) andone to whereas Fleming (1963) favoredAntarctica as a the west (Mt. Taranaki). These mountains are younger potential refuge for cool-adapted lineages. than those of the axial ranges, having reachedtheir present A link between the New Zealandalpine flora and elevations during the last Pleistocene glacial or more floras of the northern temperate zone has long been recently. Their contemporary relief has been shapedby recognized. For example, in an analysis by Dawson both volcanic activity anderosion ( Holloway 1982). (1963) genera with north temperate affinities formedthe largest component of the alpine flora. However, Raven Biogeographic themes in the New Zealand mountain (1973) was the first to explicitly suggest that much of the flora New Zealandalpine flora was recently derivedfrom Northern Hemisphere ancestors. Whilst recognizing the The origins of the New Zealandmountain flora have presence of a potentially ancient element, Raven (1973) been a longstanding biogeographic problem. Of the proposedthat most lineages haddispersedto New approximately 600 vascular plant species that occur in Zealandduringthe Pliocene andPleistocene. More the New Zealandmountains more than 90% are specifically, he suggestedthat the mountains of south- endemic. In contrast, no plant families and less than east Asia andNew Guinea – also upliftedduring the late 10% of the genera are endemic to the New Zealand Tertiary andPleistocene – servedas stepping stones for mountains (Mark andAdams 1995 ). Consistent with the dispersal of pre-adapted alpine lineages to Australia, this observation, numerous authors (e.g., Hooker 1853; with dispersal of these groups to the New Zealand Dawson 1963; Fleming 1963; Wardle 1963, 1968; Raven mediated by West Wind Drift (Raven 1973; Raven and 1973; McGlone 1985; McGlone et al. 2001) have noted Raven 1976). Recently, Pole (1994) and Macphail (1997) the diverse affinities of the lineages represented in the have also arguedthat Australia was an important source alpine flora. Two dramatically different hypotheses have for the New Zealandflora. They suggest that New been suggestedto explain this diversity: (i) that these Zealandwas completely or nearly completely inundated lineages have an ancient history in the Southern for at least part of the Oligocene, andconsequently that Hemisphere, or (ii) that they have arrivedmuch more the extant flora must have arrivedfrom Australia since recently from the Northern Hemisphere. reneweduplift began in the Miocene. Several authors (e.g., Cockayne 1928; Fleming 1962, A secondfocus for biogeographical research in New 1963; Wardle 1963, 1968, 1978) have favoredthe idea Zealandhave been patterns of distribution within the that New Zealandalpine lineages have long histories in alpine flora, especially with respect to regions of the Southern Hemisphere. Whilst these hypotheses all endemism and range disjunctions. Strikingly, approxi- involve an ancient presence they suggest quite different mately 6% of alpine species are restrictedto the scenarios for the origin andsurvival of these plants. The northern thirdof the South Island,whereas approxi- earliest hypothesis is that of Cockayne (1928) who mately 8% are confinedto the southern third( Mark and suggested the modern flora was descended from groups Adams 1995). There are also numerous examples of taxa present in Cretaceous age mountains. Following initial that are disjunct between these two endemic centers uplift in the late Jurassic andearly Cretaceous, the (Fig. 1). Three distinct explanations have been pro- ancestral New Zealandlandmass was mountainous, and posed. The first hypothesis favors vicariance (McGlone Cockayne (1928) suggestedthat lineages occurring in 1985; Heads 1998). Over the last 25 my the Pacific- these mountains might have survivedthe Tertiary in Australian plate boundary has experienced 480 km of temperate, low-lying habitats, radiating into contem- lateral displacement along the alpine fault, with as much porary alpine habitats as they arose. Specifically, it has as 100 km of lateral displacement northeast relative to been suggestedthat areas of cool, wet, andinfertile soils the Pacific Plate occurring as recently as the mid- may have provided conditions suitable for the develop- Pliocene (Sutherland1994 ). This model suggests that ment of a cool temperate assemblage prior to renewed historically continuous distributions may have begun to mountain building in the late Pliocene (Wardle 1968, split apart during the Miocene (McGlone 1985; Heads ARTICLE IN PRESS 240 R.C. Winkworth et al. / Organisms, Diversity & Evolution 5 (2005) 237–247

1998). In contrast, the two other hypotheses invoke suggesteda late Tertiary, or more recent, arrival of these more recent events. One suggests that the origin of the plant groups in Australasia (Table 1). For example, endemic centers relates directly to mountain building uncertainty about the phylogeny andtaxonomy of (McGlone 1985). Geological evidence indicates that the Myosotis had made it difficult to differentiate between endemic centers have been relatively stable over then last hypotheses suggesting either Northern Hemisphere or 5 my, whereas the central Southern Alps have been Australasian origins (e.g., Grau andLeins 1968 ; Raven tectonically much more active. This explanation implies 1973). However, molecular phylogenetic analyses of the that tectonically stable areas have maintainedspecies nuclear ribosomal ITS (nrITS) andthree chloroplast loci diversity or provided opportunities for the evolution of have clarifiedrelationships within Myosotis (Winkworth local endemics, and that habitat instability has limited et al. 1999, 2002a). These analyses strongly suggest a such opportunities in the intervening region (McGlone Northern Hemisphere origin; furthermore, the level of 1985). The final hypothesis suggests that contemporary genetic divergence between austral taxa and Northern distribution patterns are the result of glaciation and Hemisphere relatives implies that long distance dispersal glacial climates over the last 2.5 my. During the late has playedan important role in establishing the current Pliocene andPleistocene repeatedclimatic changes and geographic range of Myosotis. Indeed molecular age glaciation strongly influencedNew Zealandenviron- estimates basedon nrITS sequences suggest that the ments. In particular, geological evidence suggests that austral lineage was derived from an ancestor that most the central portion of the Southern Alps were more likely arrivedfrom the Northern Hemisphere within the heavily glaciatedthan areas to the north andsouth. last 15 my (Winkworth et al. 2002a). Several authors (e.g., Cockayne 1928; Wardle 1963, Molecular phylogenetic studies have supported Ra- 1988; Burrows 1965) have arguedthat alpine taxa may ven’s (1973) assertion that many of New Zealand’s have been restrictedto these relatively ice-free regions alpine plant lineages are recent arrivals from the during glacial maxima, and that contemporary range Northern Hemisphere. In addition to Myosotis, analyses disjunctions are explained by failure to recolonize the of Asteraceae (Breitwieser et al. 1999), Brassicaceae central Southern Alps (Fig. 1). (Heenan et al. 2002), the Hebe complex (Wagstaff et al. There are also distinct differences between the alpine 2002), Ranunculus (Lockhart et al. 2001), and Scler- floras of the North andSouth Islands.Specifically, the anthus (Smissen et al. 2003b) suggest Northern Hemi- alpine flora of North Islandis smaller in size andhas sphere origins. However, these phylogenetic studies are lower diversity. Several factors may have limited the not consistent with the route of dispersal suggested by evolution of alpine plant diversity in the North Island. Raven (1973). That is, these analyses do not provide These include: (a) the younger age and more limited evidence that the mountains of southeast Asia and extent of mountain habitats; (b) a post-glacial climatic Australia actedas stepping-stones. An expectation of regime that probably further reduced the extent of stepping-stone models is that taxa from southeast Asia North Islandalpine habitats; (c) the relative isolation andAustralia wouldbe basal to the New Zealand andinstability of habitats on the volcanic peaks radiation in gene trees. However, to date species from (McGlone 1985; Wardle 1988). stepping-stone locations have been nestedwithin predominantly New Zealand clades, suggesting that they are derived rather than ancestral. In Myosotis, for Insights from molecular data example, Winkworth et al. (2002a) suggest that lineages currently present in the mountains of Australia andNew Pre-molecular studies of the New Zealand alpine flora Guinea arose following dispersal from New Zealand. have provided important clues to the evolutionary Although stepping-stone hypotheses are not currently history of these plants. However, these studies have supportedby molecular phylogenies, we shouldremain been unable to distinguish between competing hypo- cautious about these inferences since phylogenetic theses of origin. Recently, attention has focusedon the analyses may be limitedin this respect. Specifically, if use of molecular phylogenetic approaches as a means to Pleistocene climatic fluctuations resultedin the extinc- help test hypotheses and provide additional insights into tion of stepping-stone populations, then it wouldbe the history of New Zealandalpine plants. Here we misleading to draw conclusions based on extant review recent phylogenetic studies that are contributing lineages. One possible approach to this problem would to our understanding of the origins and evolution of the be to use molecular age estimates; if contemporary mountain flora. species were shown to have persistedin the stepping- stone regions throughout the Pleistocene, then it would The importance of dispersal be difficult to explain survival of these species and not Molecular phylogenetic analyses have provided im- the stepping-stone forms (Winkworth et al. 2002b). portant insights into the origins of New Zealandalpine Although recent analyses have suggestedthe impor- plant lineages. To date, these studies have consistently tance of the Northern Hemisphere as a source for ARTICLE IN PRESS R.C. Winkworth et al. / Organisms, Diversity & Evolution 5 (2005) 237–247 241

Table 1. New Zealandalpine plant taxa for which molecular phylogenetic analyses suggest recent origins anddiversiﬁcation

Taxon Approx. number of Inferredorigins Molecular age Reference New Zealandspecies a estimate for lineage

Apiaceae/apioidgenera b 60 Northern n/ac Mitchell et al. (1998), Hemisphere Radford et al. (2001) Asteraceae/Gnaphalieae 75 Equivocald n/a Breitwieser et al. (1999), Wagstaff andBreitwieser (2002), Smissen et al. (2004) Asteraceae/Brachyglottis 30 Equivocale n/a Wagstaff andBreitwieser (2002, 2004) Boraginaceae/Myosotis 34 Northern 2.0–14.7 mya Winkworth et al. (2002a) Hemisphere Brassicaceae/Pachycladon and 8 Northern 1.0–3.5 mya Heenan et al. (2002) relatives Hemisphere Caryophyllaceae/Scleranthus 3 Northern 1.2–7.7 mya Smissen et al. (2003b) Hemisphere Gentianaceae/Gentianella 30 South America 1.4–2.7 mya von Hagen andKadereit (2001), Glenny (2004) Ranunculaceae/Caltha 2 South America n/a Schuettpelz andHoot (2004) Ranunculaceae/Ranunculus 41 Northern n/a Lockhart et al. (2001) Hemisphere Scrophulariaceae/Ourisia 10 South America n/a Meudt and Simpson (in press) Scrophulariaceae/Hebe complexf 100 Northern 3.9 mya Wagstaff et al. (2002) Hemisphere Stylidiaceae/Forstera and 7 Equivocalg 6.0 mya Wagstaff andWege (2002) Phyllache Stylidiaceae/Oreostylidium 1 Australia 2.0 mya Wagstaff andWege (2002)

aSpecies numbers compiled from individual references and additional sources (Mark andAdams1995 ; Wilton andBreitwieser 2000 ; McGlone et al. 2001). bIncludes Aciphylla, Anisotome, Gingidium, and Lignocarpa. A fifth member of this clade, Scandia, has no alpine representatives. cn/a ¼ a divergence time estimate is not available, although relative branch lengths are consistent with recent events. dMolecular analyses suggest close affinities to Australian taxa, with multiple independent dispersals between New Zealand, Australia, and New Guinea. However, the root position remains equivocal, andtherefore it is difficultto make strong inferences. eOrigins of the group in New Zealandare equivocal. This morphologically extremely diversegenus is closely relatedto several Australian species, but exact relationships remain uncertain. fIncludes Chionohebe, Hebe, Heliohebe, and Parahebe. gThe two genera are not monophyletic, Phyllache being nestedwithin Forstera. The New Zealandradiationis most closely relatedto a South American species, andin turn this cladeis sister to the Tasmanian Forstera. Therefore, South American andTasmanian origins are equally parsimonious.

lineages in the New Zealandalpine flora, several studies distance dispersal. However, fossil evidence suggests have highlightedlineages with origins in southern South that Antarctica may have actedas a corridorfor America. Molecular phylogenetic analyses of Caltha movements between these two regions until relatively (Schuettpelz andHoot 2004 ), Gentianella (von Hagen recently (e.g., Webb andHarwood1991 ; Ashworth and andKadereit 2001 ), and Ourisia (Meudt and Simpson in Cantrill 2004), a possibility recognizedby many authors press) all indicate South American origins for the New (e.g., Darwin 1859; Skottsberg 1960; Swenson and Zealandrepresentatives. Phylogenetic analyses of mor- Bremer 1997a; Wardle et al. 2001). In this case phological characters suggest similar origins for Abro- distinguishing between alternative dispersal routes will tanella (Swenson andBremer 1997a, b ). As for lineages require a much more detailed knowledge of the fossil with Northern Hemisphere origins, the route of record. dispersal between South America and New Zealand Taken together, molecular phylogenetic andpalyno- remains uncertain. Phylogenies basedon contemporary logical studies (e.g., Fleming 1979; Mildenhall 1980) species couldbe interpretedas favoring directlong- provide convincing evidence that many alpine plant ARTICLE IN PRESS 242 R.C. Winkworth et al. / Organisms, Diversity & Evolution 5 (2005) 237–247 lineages are recent additions to the New Zealand flora. of recent diversification in the New Zealand mountains Indeed dispersal appears to have played a very include the Australasian Apioideae (Mitchell et al. 1998; important role in the assembly of this element, lineages Radford et al. 2001), Brachyglottis (Wagstaff and having arrived at different times and from different Breitwieser 2004), the Gnaphalieae (Breitwieser et al. sources. The apparent influx of alpine lineages into the 1999; Smissen et al. 2003a, 2004), the Hebe complex New Zealandmountains since the late Tertiary may be (Wagstaff andGarnock-Jones 1998, 2000 ; Wagstaff et al. directly linked to the uplift of montane environments. 2002), Myosotis (Winkworth et al. 1999, 2002a), Ourisia However, this correlation might also reflect an increase (Meudt and Simpson in press), the Pachycladon complex in the availability of open niches, both in montane and (Heenan et al. 2002; Heenan andMitchell 2003 ), and lowlandenvironments. That is, comparedto the Ranunculus (Lockhart et al. 2001). environmental stability of the early Tertiary, the rapidly Rapiddiversification of plant lineages in the New changing environments of the late Pliocene andPleisto- Zealandmountains appears to be correlatedwith a cene may have createda range of new, open habitats periodof environmental instability duringthe late that offeredopportunities for the establishment of Pliocene andPleistocene. Molecular clock analyses of dispersing lineages. Although long-distance dispersal nrITS sequence data from Myosotis (Winkworth et al. clearly has been an important influence on the composi- 2002a), Pachycladon (Heenan et al. 2002), and Ranun- tion of the New Zealandalpine flora, the overall culus (Lockhart et al. 2001) suggest that the contem- contribution of dispersed lineages to this assemblage porary diversity arose within the last 5 my, a period of remains uncertain. Our current knowledge is biased rapidgeological uplift andfrequent climatic fluctuations towards larger, morphologically and ecologically diverse in New Zealand. Certainly the expansion of montane groups, and understanding the assembly of the alpine habitats during this period may have provided oppor- element requires the inclusion of lineages that poten- tunities for the diversification of contemporary moun- tially represent alternative evolutionary histories. For tain plant lineages (Raven 1972). That is, we might example, several predominantly lowland forest groups expect that the topography of the newly formed also have representatives in the mountain flora (e.g., mountains wouldhave introducedbarriers to gene flow, Coprosma, Myrsine, Pittosporum,andPodocarpus). It thereby promoting the evolution of novel, local variants has been suggestedthat such lineages might have long (Wardle 1963). However, it also seems likely that the histories in New Zealand, with the alpine forms arising repeatedclimatic fluctuations duringthe late Pliocene only after the onset of mountain building and climate andPleistocene have playedan important role in alpine change (Smith 1986; Dawson 1988). Indeed molecular plant diversification. Several authors have previously phylogenetic analyses appear to support such an implicatedcyclical environmental change in modelsof explanation for the origin of Phyllocladus alpinus plant speciation (e.g., Ehrendorfer 1958; Rattenbury (Podocarpaceae). In this case, despite a long fossil 1962; Morton 1972; Stebbins 1984; Hewitt 1996; Ferris record, molecular evidence indicates that the modern et al. 1999), andit seems possible that the intense species diversity arose very recently; indeed the moun- selective regimes imposedby Pliocene andPleistocene tain ecotype of P. alpinus appears to have arisen from a glacial cycles may have influencedthe radiationof New lowlandform within the last 6.3 my ( Wagstaff 2004). Zealandmountain plant lineages. A potential cyclical model might involve a two phases: (i) climate change Rapid morphological differentiation produces new habitats that allow range expansion and The New Zealandmountain flora is characterizedby provide opportunities for the differentiation of local numerous large, morphologically andecologically di- morphological or ecological variants; (ii) ongoing verse genera. For example, Aciphylla, Celmisia, Epilo- climate change leads to range contraction, which may bium, Hebe, and Ranunculus each contain more than 35 result in the formation of additional local variants species andhave many montane species ( Mark and either by local differentiation or perhaps hybridization Adams 1995; Wilton andBreitwieser 2000 ). Despite (Fig. 2). Through multiple climatic cycles such a model uncertainty about the origins of this flora, it has might be expectedto producemany variants, especially generally been assumedthat speciation during the if additional habitats become available or changing Pliocene andPleistocene explains the presence of these migration paths result in interactions (e.g., hybridization large, morphologically andecologically diverse lineages or introgression) between previously isolatedforms. (Wardle 1968; Raven 1973). Recent molecular phyloge- Given the large environmental changes associated netic analyses are also consistent with this suggestion. with glacial cycles, as well as smaller fluctuations within Specifically, species with striking morphological and each cycle, we might expect that such a model of species ecological differences exhibit little or no genetic differ- diversification would result in considerable admixture, entiation at commonly assayedmarker loci (e.g., nrITS andtherefore complex patterns of phylogenetic relation- andchloroplast loci), suggesting that differentiation has ship. Indeed this expectation is consistent with several occurredover a relatively short periodof time. Examples recent molecular analyses. For example, Lockhart et al. ARTICLE IN PRESS R.C. Winkworth et al. / Organisms, Diversity & Evolution 5 (2005) 237–247 243

Fig. 2. Potential model for the rapid diversification of New Zealand alpine plant lineages during the Pliocene and Pleistocene: (A) Climate change leads to range expansion. An initial population (light gray) forms several isolated daughter populations (dark gray) that differentiate from one another. (B) Further climate changes result in range contraction. Due to altered migration pathways or novel ecological tolerances these new populations (dark gray) may occupy different habitats relative to the initial population. Furthermore, such range contraction may result in different local forms occurring in one area, providing opportunities for the origin of further novel variation through hybridization. (C) Repeated cycles of range expansion and contraction associated with later climate fluctuations may promote further diversification, especially if new habitats become available or interactions (e.g., hybridization or introgression) between previously isolated forms are possible.

(2001) identified a clade (their Group II) containing address the role of hybridization within alpine plant three morphologically distinct Ranunculus species en- lineages are clearly needed. demic to the South Island. Analyses of nrITS sequences Recent molecular systematic studies on New Zealand indicate complex patterns of relationship between these alpine plant lineages show striking similarities to those taxa perhaps linkedto reticulate evolution ( Lockhart et from both oceanic islandandcontinental floras. In al. 2001; Winkworth et al. 2005). Similarly, incongru- particular, the patterns of relationship exhibitedby ence between gene trees for chloroplast andnuclear loci many New Zealandalpine plant groups parallel plant suggest reticulate patterns of evolution in Raoulia species radiations on other island archipelagos. Specifi- (Smissen et al. 2004) andthe Australasian Apioideae cally, radiations on the Hawaiian Islands (e.g., Carr (Winkworth andLockhart, unpublished). These obser- 1987; Robichaux et al. 1990; Baldwin 1992), the Juan vations are certainly consistent with earlier suggestions Fernandez Islands (e.g., Sang et al. 1994, 1995), andthe that hybridization may have been of considerable Macaronesian Islands (e.g., Bo¨ hle et al. 1996; Kim et al. evolutionary importance in the New Zealandmountain 1996) all exhibit extensive morphological andecological flora (Raven 1972, 1973) andwith the reported differentiation but limited genetic variation. For such frequency of hybridization – natural hybrids are known groups analyses of morphology, biochemistry, cytology for Ranunculus (Fisher 1965), Leptinella (Lloyd1972 ), andgenetics tendto favor hypotheses of recent arrival Epilobium (Raven 1972; Raven andRaven 1976 ), and by transoceanic dispersal, followed by rapid and the Australasian Apioideae (Webb andDruce 1984 ). potentially adaptive radiation (Givnish 1997). In addi- However, it remains unclear whether the apparent tion, studies on New Zealand alpine plants are also frequency of hybrids in the contemporary alpine flora consistent with analyses of continental lineages that is correlatedwith the evolutionary significance of this suggest the importance of environmental change during phenomenon. Phylogenetic studies that specifically the late Tertiary andQuaternary. Such investigations ARTICLE IN PRESS 244 R.C. Winkworth et al. / Organisms, Diversity & Evolution 5 (2005) 237–247 suggest that rapidandrepeatedclimatic fluctuations Alps (Lockhart et al. 2001; Lockhart, unpublished). associatedwith Pleistocene glaciation have hadpro- This pattern was interpretedas consistent with restricted foundeffects on the floras of Europe andNorth gene flow due to range contraction into separate refugia America (Comes andKadereit 1998 ). Clearly, environ- during Pleistocene glaciations (Lockhart et al. 2001; mental change during the last 5 my has been an McGlone et al. 2001). Lockhart et al. (2001) also suggest important stimulus for the evolution of modern floristic that lower sea levels during the glacial may have diversity at both local and global scales. facilitatedthe movement of Ranunculus into the mountains of the North IslandandStewart Island.In Species distributions other work, analyses of ISSR markers andnrITS Previous hypotheses have invokedancient tectonism sequences have been usedto investigate phylogenetic (e.g., Heads 1998), mountain building (e.g., McGlone relationships within Raoulia (Asteraceae). Phylogenetic 1985), or glaciation (e.g., Wardle 1988) to explain analyses recover two well-supportedclades:one contain- distribution patterns in the New Zealand alpine flora. ing exclusively North Islandpopulations, andthe other Although molecular phylogenetic analyses provide in- only representatives of South Islandpopulations ( Smis- sights into the origin anddivergence of these lineages, sen et al. 2003a). These results suggest that, although they have made little impact on our understanding of Pleistocene climate change may have been an important distribution patterns. In general this reflects the limited influence on contemporary species diversity and dis- resolution provided by commonly used molecular tribution, it is insufficient to fully explain patterns of markers such as the nrITS andchloroplast loci in what endemism in Raoulia. Specifically, the level of genetic appear to be recently differentiated groups. In most differentiation between North Island and South Island studies relationships between the austral species are populations suggests that this divergence occurred prior poorly resolved, either represented as polytomies or, if to glaciation, although glacial cycling might have clades are resolved, relationships are weakly supported influencedpatterns of distribution in the South Island (e.g., Wagstaff et al. 2002; Winkworth et al. 2002a). (Smissen et al. 2003a). Whilst our understanding of distribution patterns To date, molecular phylogenetic studies provide only remains limited, several recent studies have provided preliminary insights into the origin of distribution some preliminary insights (e.g., Lockhart et al. 2001; patterns in the New Zealandmountains. However, these Heenan andMitchell 2003 ; Smissen et al. 2003a). studies do suggest that the alpine flora is a more recent Combinedanalyses of morphological andnrITS data assemblage than previously thought (e.g., Heads 1998). produced a largely well-resolved phylogeny for Pachy- In particular, molecular-clock analyses suggest that cladon, a group of about 10 species distributed primarily environmental change during the latest Tertiary and in the mountains of the South Island, but with a single Quaternary has been most important for determining representative in Australia (Heenan andMitchell 2003 ). contemporary patterns of diversity and distribution. In this tree the New Zealandspecies fall into three Furthermore, these studies indicate that no single factor clades, two of which are restricted to mountain habitats. can fully explain patterns of alpine plant diversity; for a The two alpine lineages are differentiated both mor- given plant group, patterns of diversification and phologically andecologically – one is restrictedto schist distribution appear to have been determined by a substrates in the southern South Island, whereas the complex, andpotentially unique, set of influences other occurs on greywacke substrates to the north. relatedto climate, geology andecological tolerances Heenan andMitchell (2003) hypothesize that adaptation (McGlone et al. 2001). to geological substrates may have stimulatedthe initial diversification of Pachycladon. However, they suggest that the disjunct distribution of P. fastigiata and restricteddistributions of several species (e.g., P. wallii Conclusions and outlook and P. stellata) reflects range contraction into isolated glacial refugia during the Pleistocene. Molecular systematic studies are contributing greatly Studies on Raoulia and Ranunculus also suggest that to our understanding of the New Zealand alpine flora. glaciation has been important for shaping contemporary Phylogenetic analyses, in conjunction with palynological patterns of alpine plant biodiversity. Lockhart et al. and distributional data, suggest that contemporary (2001) described molecular phylogenetic analyses of patterns of diversity and distribution in this assemblage nuclear andchloroplast loci for a broadsample of New have arisen only recently. More specifically, these Zealandalpine Ranunculus. These analyses identified studies indicate two general patterns: (i) that ancestors four major lineages andprovidedsome resolution of of many New Zealandalpine plant lineages arrivedby relationships within these groupings. In two of these long-distance dispersal during the late Tertiary or major lineages, accessions from southwestern popula- Quaternary, and(ii) that alpine lineages have diversi- tions are genetically distinct from those in the central fiedrecently, perhaps in response to environmental ARTICLE IN PRESS R.C. Winkworth et al. / Organisms, Diversity & Evolution 5 (2005) 237–247 245 instability during the Pliocene and Pleistocene. Further- Breitwieser, I., Glenny, D.S., Thorne, A., Wagstaff, S.J., 1999. more, molecular data suggest that alpine plant distribu- Phylogenetic relationships in Australasian Gnaphalieae tion patterns are unlikely to have been determined by a (Compositae) inferredfrom ITS sequences. N. Z. J. Bot. single factor, but insteadappear to be influencedby 37, 399–412. both climate andgeology. These findings mirror results Burrows, C.J., 1965. Some discontinuous distributions of from molecular systematic studies on island plant plants within New Zealandandtheir ecological signifi- radiations and continental floras. cance. Part II: disjunctions between Otago-Southland and Nelson-Marlborough andrelateddistributionpatterns. Despite these advances we are still far from a detailed Tuatara 13, 9–29. understanding of the evolutionary patterns and pro- Carr, G.D., 1987. Beggars ticks andtarweeds:masters of cesses in the New Zealandmountains. Many of the adaptive radiation. TREE 2, 192–195. questions that remain also have implications for under- Cockayne, L., 1928. The Vegetation of New Zealand. standing wider problems in plant evolution. Addressing Engelmann, Leipzig. these issues will require synergistic approaches drawing Comes, H.P., Kadereit, J.W., 1998. The effects of Quaternary on numerous disciplines. We expect future studies in climatic changes on plant distribution and evolution. New Zealandwill provideimportant insights into (a) the Trends Plant Sci. 3, 432–438. evolutionary significance of hybridization in plant Cooper, A., Cooper, R.A., 1995. The Oligocene bottleneck diversification, (b) the role of plant breeding systems andthe New Zealandbiota: genetic recordof a past in adaptive radiation, (c) the concept and evolutionary environmental crisis. Proc. R. Soc. London, Ser. B, Biol. significance of glacial refugia, (d) the development of Sci. 261, 293–302. analytical methods and ecological models for studying Cooper, A.F., Barreiro, A.B., Kimbrough, D.L., Mattinson, species radiation, and (e) the genetics underlying J.M., 1987. Lamprophyre dike intrusion and age of the alpine fault, New Zealand. Geology 15, 941–944. morphological andphysiological adaptation. Cox, S.C., Findlay, R.H., 1995. The main divide fault zone andits role in formation of the southern alps, New Zealand. N. Z. J. Geol. Geophys. 38, 489–499. Acknowledgements Darwin, C., 1859. On the Origin of Species by Means of Natural Selection (reprinted1950). Watts & Co., London. The authors thank Reto Nyffeler for the opportunity Dawson, J.W., 1963. Origins of the New Zealandalpine flora. to present this work during the ASPT colloquium Proc. N. Z. Ecol. Soc. 10, 12–15. Dawson, J.W., 1988. From Forest Vine to Snow Tussock. entitled‘‘Alpine plant biogeography on the large scale: Victoria University Press, Wellington. current stages andprospects’’. The authors also Ehrendorfer, F.E.R., 1958. Differentiation–hybridization cy- acknowledge financial support from the New Zealand cles andpolyploidy.ColdSpring Harbor Symp. Quant. Marsden Fund, the New Zealand Foundation for Biol. 24, 141–152. Research, Science andTechnology, andthe Alexander Ferris, C., King, R.A., Hewitt, G.M., 1999. Isolation within von Humboldt Foundation. species andthe history of glacial refugia. In: Hollingsworth, P.M., Bateman, R.M., Gornall, R. (Eds.), Molecular Systematics andPlant Evolution. Taylor & Francis, References London, pp. 20–34. Fisher, F.J.F., 1965. The Alpine Ranunculi of New Zealand. Ashworth, A.C., Cantrill, D.J., 2004. Neogene vegetation of Botany Division, Department of Scientific andIndustrial the Meyer Desert Formation (Sirius Group) Transantarctic Research, Wellington. Mountains, Antarctica. Paleogeogr. Paleoclimatol. Paleoe- Fleming, C.A., 1962. New Zealandbiogeography – a col. 213, 65–82. paleontologist’s approach. Tuatara 10, 53–104. Baldwin, B.G., 1992. Phylogenetic utility of the internal Fleming, C.A., 1963. Age of the alpine biota. Proc. N. Z. Ecol. transcribedspacers of nuclear ribosomal DNA in plants: Soc. 10, 15–18. an example from the Compositae. Mol. Phylogenet. Evol. Fleming, C.A., 1979. The Geological History of New Zealand 1, 3–16. andits Life. AucklandUniversity Press. Batt, G.E., Braun, J., Kohn, B.P., McDougall, I., 2000. Givnish, T.J., 1997. Adaptive radiation and molecular systema- Thermochronological analysis of the dynamics of the tics: issues andapproaches. In: Givnish, T.J., Sytsma, K.J. southern alps, New Zealand. Bull. Geol. Soc. Am. 112, (Eds.), Molecular Evolution and Adaptive Radiation. Cam- 250–266. bridge University Press, Cambridge, pp. 1–54. Beu, A.G., Edwards, A.R., 1984. New Zealand Pleistocene and Glenny, D., 2004. A revision of the genus Gentianella in New late Pliocene glacio-eustatic cycles. Paleogeogr. Paleoclima- Zealand. N. Z. J. Bot. 42, 361–530. tol. Paleoecol. 46, 119–142. Grau, J., Leins, P., 1968. Pollenkorntypen undSektionsglie- Bo¨ hle, U.-R., Hilger, H.H., Martin, W.F., 1996. Island derung der Gattung Myosotis L. Ber. Dt. Bot. Ges. 81, colonisation andevolution of the insular woodyhabit in 107–117. Echium L. (Boraginaceae). Proc. Natl. Acad. Sci. 93, Hagen, K.B.von., Kadereit, J.W., 2001. The phylogeny of 11740–11745. Gentianella (Gentianaceae) andits colonization of the ARTICLE IN PRESS 246 R.C. Winkworth et al. / Organisms, Diversity & Evolution 5 (2005) 237–247

southern hemisphere as revealedby nuclear andchloroplast Mildenhall, D.C., 1980. New Zealand late Cretaceous and DNA sequence variation. Org. Divers. Evol. 1, Cenozoic plant biogeography: a contribution. Paleogeogr. 61–79. Paleoclimatol. Paleoecol. 31, 197–233. Heads, M.J., 1998. Biogeographic disjunction along the Alpine Mitchell, A.D., Webb, C.J., Wagstaff, S.J., 1998. Phylo- fault, New Zealand. Biol. J. Linn. Soc. 63, 161–176. genetic relationships of the species of Gingidia andrelated Heenan, P.B., Mitchell, A.D., 2003. Phylogeny, biogeography genera (Apiaceae, subfamily Apioideae). N. Z. J. Bot. 36, and adaptive radiation of Pachycladon (Brassicaceae) in the 417–424. mountains of South Island, New Zealand. J. Biogeogr. 30, Morton, J.K., 1972. Phytogeography of the west African 1737–1749. mountains. In: Valentine, D.H. (Ed.), Taxonomy, Phyto- Heenan, P.B., Mitchell, A.D., Koch, M., 2002. Molecular geography and Evolution. Academic Press, London, systematics of the New Zealand Pachycladon (Brassicaceae) pp. 221–236. complex: generic circumscription andrelationships to Nelson, G., 1975. Reviews: biogeography, the vicariance Arabidopsis s. l. and Arabis s. l. N. Z. J. Bot. 40, 543–562. paradigm, and continental drift. Syst. Zool. 24, 489–504. Hewitt, G.M., 1996. Some genetic consequences of ice ages, Ollier, C.D., 1986. The origins of alpine landforms in andtheir role in divergenceandspeciation. Biol. J. Linn. Australasia. In: Barlow, B.A. (Ed.), Flora and Fauna of Soc. 58, 247–276. Alpine Australasia. CSIRO, Melbourne, pp. 3–26. Holloway, J.T., 1982. The Mountain Lands of New Zealand. Pillans, B., Pullar, W.A., Selby, M.J., Soons, J.M., 1992. The Tussock Grasslands & Mountain Lands Institute, Christch- age and development of the New Zealand landscape. In: urch. Soons, J.M., Selby, M.J. (Eds.), Landforms of New Hooker, J.D., 1853. Botany of the Antarctic Voyage of the Zealand. Longman Paul, Auckland, pp. 31–62. Discovery Ships ‘Erebus’ and‘Terror’ in the Years Pole, M., 1994. The New Zealandflora – entirely long-distance 1831–1843. II. Flora Novae-Zelandiae. Part I. Flowering dispersal? J. Biogeogr. 21, 625–635. plants. Lovell Reeve, London. Radford, E.A., Watson, M.F., Preston, J., 2001. Phylo- Kamp, P.J.J., 1992. Tectonic architecture of New Zealand. In: genetic relationships of species of Aciphylla (Apiaceae, Soons, J.M., Selby, M.J. (Eds.), Landforms of New subfamily Apioideae) and related genera using molecular, Zealand. Longman Paul, Auckland, pp. 1–30. morphological andcombineddatasets. N. Z. J. Bot. 39, Kim, S.-C., Crawford, D.J., Francisco-Ortega, J., Santos- 183–208. Guerra, A., 1996. A common origin for woody Sonchus and Rattenbury, J.A., 1962. Cyclic hybridisation as a survival five relatedgenera in the Macaronesian islands:molecular mechanism in the New Zealandforest flora. Evolution 16, evidence for extensive radiation. Proc. Natl. Acad. Sci. 93, 348–363. 7743–7748. Raven, P.H., 1972. Evolution andendemismin the New Lloyd, D.G., 1972. A revision of the New Zealand, Sub- Zealandspecies of Epilobium. In: Valentine, D.H. (Ed.), antarctic, andSouth American species of Cotula section Taxonomy, Phytogeography andEvolution. Academic Leptinella. N. Z. J. Bot. 10, 277–372. Press, London, pp. 259–274. Lockhart, P.J., Mclenachan, P.A., Havell, D., Glenny, D., Raven, P.H., 1973. Evolution of the subalpine andalpine plant Huson, D., Jensen, U., 2001. Phylogeny, radiation and groups in New Zealand. N. Z. J. Bot. 11, 177–200. transoceanic dispersal of New Zealand alpine buttercups: Raven, P.H., Raven, T.E., 1976. The Genus Epilobium molecular evidence under split decomposition. Ann. Mo. (Onagraceae) in Australasia: a Systematic andEvolution- Bot. Gard. 88, 458–477. ary Study. Botany Division, Department of Scientific and Macphail, M.K., 1997. Comment on M. Pole, 1994. The New Industrial Research, Wellington. Zealand flora – entirely long-distance dispersal? J. Bio- Robichaux, R.H., Carr, G.D., Liebman, M., Pearcy, R.W., geogr. 21, 625–635; 24, 113–114. 1990. Adaptive radiation of the Hawaiian silversword Mark, A.F., Adams, N.M., 1995. New Zealand Alpine Plants. alliance (Compositae–Madiinae): ecological, morphologi- Godwit Publishing Ltd., Auckland. cal, and physiological diversity. Ann. Mo. Bot. Gard. 77, Markgraf, V., McGlone, M.S., Hope, G., 1995. Neogene 64–72. paleoenvironmental andpaleoclimatic change in southern Sang, T., Crawford, D.J., Kim, S.-C., Stuessy, T.F., 1994. temperate ecosystems – a southern perspective. TREE 10, Radiation of the endemic genus Dendroseris (Asteraceae) 143–147. on the Juan Fernandez Islands: evidence from sequences of McGlone, M.S., 1985. Plant biogeography andthe late the ITS region of the nuclear ribosomal DNA. Am. J. Bot. Cenozoic history of New Zealand. N. Z. J. Bot. 23, 81, 1494–1501. 723–749. Sang, T., Crawford, D.J., Kim, S.-C., Silva, O.M., 1995. ITS McGlone, M.S., Duncan, R.P., Heenan, P.B., 2001. Ende- sequences andthe phylogeny of the genus Robinsonia mism, species selection andthe origin anddistributionof (Asteraceae). Syst. Bot. 20, 55–64. the vascular plant flora of New Zealand. J. Biogeogr. 28, Schuettpelz, E., Hoot, S.B., 2004. Phylogeny and 199–216. biogeography of Caltha (Ranunculaceae) basedon chlor- Meudt, H.M., Simpson, B.B., in press. The biogeography of oplast andnuclear DNA sequences. Am. J. Bot. 91, the austral, subalpine genus Ourisia (Plantaginaceae) based 247–253. on molecular phylogenetic evidence: South American origin Skottsberg, C., 1960. Remarks on the plant geography of the anddispersalto New ZealandandTasmania. Biol. J. Linn. southern coldtemperate zone. Proc. R. Soc. London,Ser. Soc. B, Biol. Sci. 152, 447–457. ARTICLE IN PRESS R.C. Winkworth et al. / Organisms, Diversity & Evolution 5 (2005) 237–247 247

Smissen, R.D., Breitwieser, I., Ward, J.M., McLenachan, P.A., Wagstaff, S.J., Bayly, M.J., Garnock-Jones, P.J., Albach, Lockhart, P.J., 2003a. Use of ISSR andITS-sequences to D.C., 2002. Classification, origin anddiversificationof the study the biogeography of alpine cushion plants in the New Zealandhebes (Scrophulariaceae). Ann. Mo. Bot. genus Raoulia (Asteraceae). Pl. Syst. Evol. 239, 79–94. Gard. 89, 38–63. Smissen, R.D., Garnock-Jones, P.J., Chambers, G.K., 2003b. Wardle, P., 1963. Evolution and distribution of the New Phylogenetic analysis of ITS sequences suggests a Pliocene Zealandflora, as affectedby Quaternary climates. N. Z. J. origin for the bipolar distribution of Scleranthus (Caryo- Bot. 1, 3–17. phyllaceae). Aust. Syst. Bot. 16, 301–315. Wardle, P., 1968. Evidence for an indigenous pre-quaternary Smissen, R.D., Breitwieser, I., Ward, J.M., 2004. Phylogenetic element in the mountain flora of New Zealand. N. Z. J. Bot. implications of trans-specific chloroplast DNA sequence 6, 120–125. polymorhism in New ZealandGnaphalieae (Asteraceae). Wardle, P., 1978. Origins of the New Zealand mountain flora, Pl. Syst. Evol. 249, 37–53. with special reference to trans–Tasman relationships. N. Z. Smith, J.W., 1986. Origins of the Australasian tropicalpine and J. Bot. 16, 535–550. alpine floras. In: Barlow, B.A. (Ed.), Flora and Fauna of Wardle, P., 1988. Effects of glacial climates on floristic Alpine Australasia. CSIRO, Melbourne, pp. 109–128. distribution in New Zealand 1. A review of the evidence. Stebbins, G.L., 1984. Polyploidy and the distribution of the N. Z. J. Bot. 26, 541–555. arctic-alpine flora: new evidence and a new approach. Bot. Wardle, P., Ezcurra, C., Ramı´ rez, C., Wagstaff, S.J., Hel. 94, 1–13. 2001. Comparison of the flora andvegetation of the Suggate, R.P., 1990. Late pliocene andquaternary glaciations southern Andes and New Zealand. N. Z. J. Bot. 39, of New Zealand. Q. Sci. Rev. 9, 175–197. 69–108. Sutherland, R., 1994. Displacement since the Pliocene along Webb, C.J., Druce, A.P., 1984. A natural intergeneric hybrid, the southern section of the alpine fault, New Zealand. Aciphylla squarrosa Gingidia montana, andthe frequency Geology 22, 327–330. of hybrids among other New Zealand apioid Umbelliferae. Swenson, U., Bremer, K., 1997a. Pacific biogeography of the N. Z. J. Bot. 22, 403–411. Asteraceae genus Abrontanella (Senecioneae, Blennosper- Webb, P.N., Harwood, D.M., 1991. Late Cenozoic glacial matinae). Syst. Biol. 22, 493–508. history of the Ross embayment, Antarctica. Q. Sci. Rev. 10, Swenson, U., Bremer, K., 1997b. Patterns of floral evolution of 215–223. four Asteraceae genera (Senecioneae, Blennospermatinae) Whitehouse, I.E., Pearce, A.J., 1992. Shaping the mountains of andthe origin of white flowers in New Zealand.Syst. Biol. New Zealand. In: Soons, J.M., Selby, M.J. (Eds.), Land- 46, 407–425. forms of New Zealand. Longman Paul, Auckland, Wagstaff, S.J., 2004. Origin anddiversificationof the austral pp. 144–160. genus Phyllocladus (Podocarpaceae). J. Biogeogr. 31, Wilton, A.D., Breitwieser, I., 2000. Composition of the 1569–1577. New Zealandseedplant flora. N. Z. J. Bot. 38, Wagstaff, S.J., Breitwieser, I., 2002. Phylogenetic relationships 537–549. of New ZealandAsteraceae inferredfrom ITS sequenes. Pl. Winkworth, R.C., Robertson, A.W., Ehrendorfer, F., Lock- Syst. Evol. 231, 203–224. hart, P.J., 1999. The importance of dispersal and recent Wagstaff, S.J., Breitwieser, I., 2004. Phylogeny andclassifica- speciation in the flora of New Zealand. J. Biogeogr. 26, tion of Brachyglottis (Senecioneae, Asteraceae): an example 1323–1325. of a rapid species radiation in New Zealand. Syst. Bot. 29, Winkworth, R.C., Grau, J., Robertson, A.W., Lockhart, P.J., 1003–1010. 2002a. The origin andevolution of the genus Myosotis L. Wagstaff, S.J., Garnock-Jones, P.J., 1998. Evolution and (Boraginaceae). Mol. Phylogenet. Evol. 24, 180–193. biogeography of the Hebe complex (Scrophulariaceae) Winkworth, R.C., Wagstaff, S.J., Glenny, D., Lockhart, P.J., inferredfrom ITS sequences. N. Z. J. Bot. 36, 425–437. 2002b. Plant dispersal N.E.W.S. from New Zealand. TREE Wagstaff, S.J., Garnock-Jones, P.J., 2000. Patterns of diversi- 17, 514–520. fication in Chionohebe and Parahebe (Scrophulariaceae) Winkworth, R.C., Bryant, D., Lockhart, P.J., Havell, D., inferredfrom ITS sequences. N. Z. J. Bot. 38, 389–407. Moulton, V., 2005. Biogeographic interpretation of split Wagstaff, S.J., Wege, J., 2002. Patterns of diversification in graphs: least squares optimization of branch lengths. Syst. New ZealandStylidiaceae.Am. J. Bot. 89, 865–874. Biol. 54, 56–65.