Driven by the West Wind Drift? A synthesis of southern temperate marine biogeography, with new directions for dispersalism

Authors

  • Jonathan M. Waters

    Corresponding author
      *Jonathan M. Waters, Department of Zoology, University of Otago, PO Box 56, Dunedin, New Zealand.
      E-mail: jonathan.waters@stonebow.otago.ac.nz
    Search for more papers by this author

*Jonathan M. Waters, Department of Zoology, University of Otago, PO Box 56, Dunedin, New Zealand.
E-mail: jonathan.waters@stonebow.otago.ac.nz

Abstract

Aim  Twentieth century biogeographers developed intriguing hypotheses involving West Wind Drift dispersal of Southern Hemisphere biota, but such models were largely abandoned in favour of vicariance following the development of plate tectonic theory. Here I present a synthesis of southern temperate marine biogeography, and suggest some new directions for phylogeographic research.

Location  The southern continents, formerly contiguous components of Gondwana, are now linked only by ocean currents driven by the West Wind Drift.

Methods  While vicariance theory certainly facilitates the development of testable hypotheses, it does not necessarily follow that vicariance explains much of contemporary southern marine biogeography. To overcome the limitations of narratives that simply assume vicariance or dispersal, it is essential for analyses to test biogeographic hypotheses by incorporating genetic, ecological and geological data.

Results  Recent molecular studies have provided strong evidence for dispersal, but relatively little evidence for the biogeographic role of plate tectonics in distributing southern marine taxa. Despite confident panbiogeographic claims to the contrary, molecular and ecological studies of buoyant macroalgae, such as Macrocystis, indicate that dispersal predominates. Ironically, some of the better supported evidence for marine vicariance in southern waters has little or nothing to do with plate tectonics. Rather, it involves far more localized and recent vicariant models, such as the isolating effect of the Bassian Isthmus during Pleistocene low sea-level stands (Nerita).

Main conclusions  Recent phylogeographic studies of southern marine taxa (e.g. Diloma and Parvulastra) imply that passive rafting cannot be ignored as an important mechanism of long-distance dispersal. I outline a new direction for southern hemisphere phylogeography, involving genetic analyses of bull-kelp (Durvillaea) and its associated holdfast invertebrate communities.

The old school: dispersal in southern waters

Biologists have been speculating about biogeographic mechanisms for southern temperate taxa for over 150 years. Charles Darwin (1845) was one of the first naturalists to discuss Southern Hemisphere marine biogeography, noting, for instance, the broad distribution of giant kelp (Macrocystis; from p. 239):

Its geographical range is remarkably wide; it is found from the extreme southern islets near Cape Horn… Cook, who must have been well acquainted with the species, found it at Kerguelen Land, no less than 140° in longitude.

Later, in the sixth edition of his Origin of species, Darwin (1872) noted that ‘as far as regards the occurrence of identical species at points so enormously remote as Kerguelen Land, New Zealand, and Fuegia, I believe that towards the close of the Glacial period, icebergs, as suggested by Lyell, have been largely concerned in their dispersal’. In contrast, Hooker (1853) drew attention to a widespread austral plant biota that ‘has been broken up by geological and climatic causes’. Subsequently, most biogeographers of the early–mid 20th century (e.g. Matthew, 1915; Simpson, 1940; Darlington, 1957) – understandably limited by the geological paradigms of their era – relied on ad hoc dispersalist scenarios (landbridges and repeated waves of north–south migration) concocted under the assumption of continental stability.

In the late 20th century, biogeographers began to develop more explicit hypotheses involving West Wind Drift (WWD)-mediated transport of marine larvae (echinoderms: Fell, 1962, 1967; galaxiid fishes: McDowall, 1970, 1978). These explanations typically involved ‘chains of dispersal’ (McDowall, 1970), with ‘a diminishing trail of species eastwards (down-wind) from the points of entry’ (Fleming, 1979) (Fig. 1). In the case of galaxiid fishes, several of which have an extended marine larval phase, Australian origins were inferred on the basis of taxonomic diversity (McDowall, 1970), with progressive declines in species richness further east (New Zealand, South America and South Africa). McDowall (1970, p. 413) wrote:

Figure 1.

Fell (1962) hypothesis of West Wind Drift-mediated dispersal of southern temperate echinoderms. Numbers indicate described taxa per region, and ‘D’ represents the inferred point of origin for each lineage; reproduced from Fleming 1979.

More than 20 galaxiids are recorded from Australia, 14 from New Zealand, 4–5 from South America, but only one from South Africa. These details suggest that South Africa is a most unlikely source of origin and dispersal, and that it is, rather, the end of a chain of dispersal areas, under the influence of the west wind drift…G. maculatus is present in Australia, New Zealand and South America, but not South Africa, and this suggests that it originated in Australia and spread eastwards.

As with vicariance models (see below), such sweeping biogeographic hypotheses tend to be simplistic, but nevertheless represent useful ‘false models’ from which ‘truer theories’ can be developed (Wimsatt, 1987). With regard to Fell (1962) echinoderm hypothesis (Fig. 1), for instance, subsequent molecular phylogenetic analyses indicated that the genus Patiriella sensu lato was in fact polyphyletic (Hart et al., 1997; Waters et al., 2004a). Nevertheless, Patiriella exigua, a species now allocated to Parvulastra (O'Loughlin & Waters, 2004), does indeed show evidence for recent WWD-mediated dispersal from Africa. An African origin for the clade is suggested both by the diversity of lineages present in South Africa, and by oceanographic data (Fig. 2; Waters & Roy, 2004a). This finding conflicts with Fell (1962) hypothesis that South Africa represents the biogeographic ‘tail-end’ of Patiriella sensu lato (Fig. 1). Recently, Hart et al. (2006) speculated that divergent genetic lineages of the P. exigua clade might correspond to reproductively isolated species, but future study is required to assess this possibility.

Figure 2.

 Distribution of the asterinid sea-star Parvulastra exigua (grey lines) lines, with sample sites indicated by black dots. The phylogenetic relationships of P. exigua haplotypes are inferred from ML analysis of combined mtDNA control region and cytochrome oxidase I (after Waters & Roy, 2004a).

Plate tectonics and vicariant isolation of marine taxa

Following the development of plate tectonic theory in the mid- to late 1900s, biogeographic interpretations of southern taxa shifted dramatically in favour of vicariance (Croizat, 1964; Nelson, 1974; Ball, 1975) and marine biota were by no means exempt from this paradigm shift (e.g. Knox, 1979, 1980; Zinsmeister & Camacho, 1982; Garbary, 1987; Heads, 2005). Some southern marine taxa do indeed show some substantial intercontinental genetic divergences (e.g. Afrolittorina and Austrolittorina; Williams et al., 2003; Waters et al., 2007), and may thus provide support for vicariance. For the most part, however, molecular studies have provided strong evidence for dispersal, in contrast to relatively little evidence for the biogeographic role of plate tectonics in distributing southern taxa (e.g. Ovenden et al., 1992; O'Foighil et al., 1999; Booth & Ovenden, 2000; Waters et al., 2000, 2005; Burridge & Smolenski, 2003, 2004; Casey et al., 2004; Donald et al., 2005; Burridge et al., 2006b).

Regardless, marine biogeographers have eagerly adopted plate tectonic models to explain wide southern distributions – even those involving single species (or morphologically similar congeners) of buoyant macroalgae such as Durvillaea (Cheshire et al., 1995) and Macrocystis (Chin et al., 1991). Chin et al. (1991), for example, proposed (on p. 9) that:

Lessoniaceae and Macrocystis evolved by more or less in situ differentiation out of a widespread ancestral complex distributed: circum-Antarctic, Tierra del Fuego–Peru–California–Washington, and Australia–New Zealand–northwest America. A process of vicariance at family, genus and species levels also explains why, for example, Lessonia itself is found only in the south and not in the northeast Pacific…There is no need to assume any long-distance migration to or from anywhere – the groups have simply evolved where they are…the distribution will be found to be of Mesozoic age.

A subsequent molecular study of Macrocystis shows that the above panbiogeographic scenario is, simply, wrong. Specifically, ITS1 and ITS2 sequences (fast-evolving nuclear loci) show little genetic differentiation across a wide southern geographic range spanning Chile, South Africa, Marion Island, Tasmania, mainland Australia, and New Zealand (Coyer et al., 2001); this finding can only be explained by much more recent dispersal. Indeed, in heavily glaciated regions such as South Georgia it is difficult to envisage any credible alternatives to dispersal when explaining the subsequent recolonization of littoral and sublittoral biotas of that island (Davenport & Stevenson, 1998). In rejecting dispersal, Chin et al. (1991) argued that ‘although detached fronds may drift over long distances there is no evidence that these produce reproductive blades’. But more recent studies by Macaya et al. (2005) and Hernández-Carmona et al. (2006) have produced strong evidence to the contrary:

Here, we describe the presence of sporophylls (with sporogenous tissues) on floating kelp rafts of Macrocystis spp. along the coast of Chile. Fifteen (28.6%) of the 56 samples had sporophylls, indicating the maintenance of sporophylls after detachment…Some of the kelp sporophytes with reproductive blades showed signs of having been afloat for long periods (indicated by the large size of attached stalked barnacles). Additionally, experiments showed that floating kelps released viable zoospores…suggesting zoospores may be dispersed by floating rafts…an alternative mechanism of long-distance dispersal. (Macaya et al., 2005, p. 913)

Chin et al. (1991) imply that this hypothesis of Mesozoic differentiation of marine algae can be extended more broadly, even to divergences within sea lions, fur seals and a single genus of pelagic marine fishes (Sardinops). But this ancient timeframe for otarid diversification is clearly at odds with recently published timescales of mammalian evolution (e.g. Cenozoic diversification of Carnivora: Penny et al., 1999; Springer et al., 2003; Penny & Phillips, 2004). Furthermore, comprehensive genetic analyses of Sardinops (Bowen & Grant, 1997, 1998) suggest that regional populations of Sardinops‘trace back to a common ancestor within the Pleistocene’. South African and Australian populations are particularly close relatives, with a high-frequency shared mtDNA haplotype. Bowen & Grant (1997) suggest that this shallow genetic history represents repeated extinctions and recolonizations, noting that ‘coastal upwelling zones are probably not stable sardine habitats’.

Following the acceptance of plate tectonic theory, dispersalist biogeography has received ongoing criticism (Croizat et al., 1974; Rosen, 1978; Craw, 1979; Humphries, 2000; Nelson & Ladiges, 2001; Ebach & Humphries, 2003). In particular, some researchers have questioned the value of research on individual taxa: they prefer a ‘top down’ approach involving the study of ‘area relationships’ (e.g. Humphries, 2000; Ebach et al., 2003; Ebach & Humphries, 2003). Craw (1982), for example, implied that any ‘discerning student of biogeography’ would avoid ‘looking…through the dirty spectacles of the technical problems surrounding their own narrow specialities’. But it could equally be argued that the ‘broad-brush’ vicariant explanations favoured by some biogeographers frequently fail to withstand detailed scrutiny. Indeed, concerned by the panbiogeographic tendency to overlook phylogenetic data, Mayden (1991) wondered how many associated ‘ad hoc biogeographic hypotheses and tracks…could be eliminated if at least the relationships of species were known?’. His reservations have proven to be well founded to date, as subsequent molecular phylogenetic analyses have enabled researchers to reject southern panbiogeographic hypotheses (e.g. Trewick, 2000; Trewick & Wallis, 2001; Paterson et al., 2006).

In summary, the use of vicariance theory certainly facilitates the development of testable phylogenetic hypotheses (e.g. see Mayden, 1988; Waters et al., 2001; Burridge et al., 2006a). It does not necessarily follow, however, that vicariance explains much of contemporary southern marine biogeography. For instance, if a rifted fragment of Gondwana was subsequently drowned (e.g. Chatham Islands: Campbell, 1998; Campbell et al., 2006) – and its terrestrial and littoral biotas thus completely lost – then associated panbiogeographic scenarios involving vicariant origins for extant lineages (e.g. Craw, 1988, 1989; Heads, 1989) must be false.

Evidence for recent vicariance in southern marine taxa?

Intriguingly, some of the better supported evidence for marine vicariance in southern temperate seas has little or nothing to do with Gondwana or plate tectonics. Rather, it involves far more localized and recent vicariant models, such as the isolating effect of the Bassian Isthmus that connected Tasmania to mainland Australia during Pleistocene low sea-level stands (Davies, 1974). Specifically, several marine biologists have hypothesized that this barrier promoted allopatric divergence between eastern and western populations of temperate taxa in southern Australia (e.g. Dartnall, 1974; Wilson & Allen, 1987).

Burridge (2000) provides three independent cases in which temperate Australian cirrhitoid fishes show east–west mtDNA divergence more or less consistent with this hypothesis. Similar phylogeographic evidence for east–west disjunction was also detected in temperate shallow sublittoral sea-stars (Waters et al., 2004b). Most recently, Waters et al. (2005) showed strong genetic and distributional evidence for east–west disjunction in intertidal gastropod genus Nerita (Fig. 3), with biogeographic turnover across Wilsons Promontory in southern Victoria, the northern ‘residue’ of the former Bassian Isthmus. The sharp biogeographic turnover detected for Nerita within southern Australia contrasts with (1) this gastropod's lengthy planktotrophic larval phase (5–6 months); (2) evidence for gene flow across the Tasman Sea (homogenization of eastern Australian and New Zealand populations); and (3) evidence of recent dispersal between Australasia and Easter Island (see Waters et al., 2005).

Figure 3.

 East–west disjunction of the intertidal gastropod Nerita (Gastropoda) in southern Australia (after Waters et al., 2005). The phylogeographic break at Wilsons Promontary (between Walkerville and Waterloo Bay) is consistent with Dartnall's 1974 hypothesis of vicariance associated with the Bassian landbridge linking Tasmania and Victoria during Pleistocene low sea-level stands.

Another possible example of relatively recent vicariance in southern marine taxa involves the development of the Benguela current, a cold upwelling system thought to isolate eastern vs. western populations of marine taxa in southern Africa (Bowen et al., 2001). East–west phylogeographic disjunctions along the South African coast have been recently reported for sea-stars (Parvulastra; Waters & Roy, 2004a), brittle-stars (Amphipholis; Sponer, 2002) and estuarine crustaceans (Teske et al., 2006), although the sites of disjunction vary between taxa. In addition, Waters & Roy (2003) and Lessios et al. (1999, 2001, 2003) speculated that genetic divergences detected between Atlantic vs. Indo-Pacific echinoderms might be attributable to the formation or strengthening of the Benguela cold-water barrier.

The fascinating observation that some southern marine taxa (e.g. Nerita, Parvulastra) show biogeographic evidence of both long-distance dispersal and relatively localized vicariance (see above) emphasizes the continuing need for hypotheses ‘framed within broad conceptual approaches to biogeography’ (McDowall, 2004).

The revival of long-distance dispersal: under the influence of molecular genetics

As discussed above, the view that species distributions are driven chiefly by geological (rather than biological) processes has held sway among many biogeographers for several decades (Croizat, 1964; Rosen, 1978; Nelson & Platnick, 1981; Craw et al., 1999; Humphries & Parenti, 1999). Under this vicariant vision, we apparently need only to examine earth history in order to reveal biogeographic relationships (e.g. Humphries, 2000), or to analyse biotic distributions to elucidate geological history (e.g. Craw et al., 1999); views that might suggest life forms are merely excrescences associated with underlying rock strata. Recently, however, dispersal – previously rejected as inappropriate for scientific endeavour (Croizat et al., 1974; Ball, 1975; Craw, 1979, 1982; Nelson & Ladiges, 2001) – has re-emerged as a credible scientific topic (de Queiroz, 2005), and particularly so in the Southern Hemisphere (Jordan, 2001; Winkworth et al., 2002; Chiswell et al., 2003; Sanmartin & Ronquist, 2004; Waters & Roy, 2004a; Donald et al., 2005). This transformation is largely due to the advent of new DNA sequencing methods. Amongst others, Penny & Phillips (2006) noted that ‘in the past, some aspects of evolution were outside the current boundaries of falsifiable science, but increasingly new techniques and ideas are expanding the boundaries of science’.

Recent ecological, phylogenetic and phylogeographic studies have implied that passive rafting cannot be ignored as an important mechanism of long-distance dispersal in southern waters (O'Foighil et al., 1999; Waters & Roy, 2004a; Donald et al., 2005; Thiel & Haye, 2006). If dispersal biogeography, in general, has been criticized as the study of ‘miracles’ (Craw, 1979), then the phenomenon of long-distance rafting may seem even more outlandish. Nevertheless, some marine taxa that appear intrinsically non-dispersive, with pelagic larval phases that are short-lived (e.g. Diloma nigerrima; Donald et al., 2005), or even non-existent (e.g. Parvulastra exigua; Waters & Roy, 2004a), have morphologically indistinguishable populations on isolated continents and far-flung, geologically recent, oceanic islands. Moreover, these populations exhibit very closely related mtDNA haplotypes, an obvious sign of recent gene flow. In addition, there exists strong, if circumstantial, support for rafting on macroalgae of D. nigerrima (New Zealand, Chile and Juan Fernandez), in particular, as it is the only member of its genus that is regularly found grazing on beach-cast bull-kelp (Durvillaea) (although an undescribed species of Diloma has been found inhabiting the dome-like holdfasts of Durvillaea antarctica in southern New Zealand; J.M. Waters, unpublished data).

It should be noted that rafting is not a new idea for southern taxa: Mortensen (1933), for instance, suggested that P. exigua colonized St Helena by rafting from South Africa on the holdfasts of detached Ecklonia, a buoyant seaweed that regularly drifts north to St Helena via the Benguela Current (see Fig. 2). More recent studies have even noted rafting communities associated with plastic debris (Barnes & Fraser, 2003; Thiel & Gutow, 2005a).

One potential shortcoming of dispersal biogeography, as Nelson & Ladiges (2001) argue, is ‘to render superfluous any geographic comparison across taxa’. But this need not be the case. For instance, many recent biogeographic studies of unrelated taxa with long-lived pelagic larvae show clear evidence of genetic connectivity across the Tasman Sea (Ovenden et al., 1992; Waters et al., 2000, 2005; Burridge & Smolenski, 2003; Chiswell et al., 2003). The facts that about 30% of New Zealand's inshore fish species are also found in Australian waters (Wilson & Allen, 1987), and that roughly 20% of Tasmanian echinoderm species are shared with New Zealand (Rowe & Vail, 1982), attest to the general ecological importance of dispersal in southern waters (although a few of these wide distributions almost certainly reflect anthropogenic translocation: Dartnall, 1969; Clements et al., 2000; Waters & Roy, 2004b; see Barnes & Fraser, 2003, and Castilla et al., 2005, for some other examples of artificial translocation of southern taxa). An extremely long-lived planktotrophic phase would be required for trans-Pacific larval dispersal (Castilla & Guiñez, 2000), thus rafting seems a preferable explanation for most cases of dispersal across the south Pacific. The WWD potentially links all southern continents (Briggs, 1995), providing ample scope for comparative analysis and hypothesis testing. Oceanic islands are likely to share strong biogeographic affinities with continents upstream but not downstream in the WWD. We may predict, therefore, that colonization of the oceanic Juan Fernández and Desventuradas islands, for instance, was facilitated by west–east dispersal events across the South Pacific involving both closely related (Burridge et al., 2006b) and distantly related (Santelices, 1992; Pequeño & Lamilla, 2000; Williams et al., 2003; Donald et al., 2005) taxa.

Returning to the roots: putting biology back into historical biogeography

If one is to attempt to study rare biological events – such as long-distance rafting – one should at least maximize the chance of success by focusing on species that have an ecological propensity for rafting (e.g. Edgar, 1987; Thiel & Gutow, 2005a,b; Thiel & Haye, 2006). In this section, I therefore outline a new direction for phylogeography, integrating both ecological and historical biogeographic perspectives, and involving genetic analysis of kelp and its associated rafting invertebrate communities. To this end, southern kelps of the genus Durvillaea (Fig. 4) represent a potentially powerful system for southern marine biogeography. This species complex comprises a number of solid-bladed (non-buoyant) species that are geographically restricted (Durvillaea willana, southern New Zealand; D. potatorum, southern Australia and Tasmania; D. chathamensis, Chatham Islands; D. sp. ‘Antipodes Is’, Antipodes Islands (see Hay, 1994; Adams, 1997), and a single honeycomb-bladed (buoyant) species (D. antarctica) that has a wide subantarctic distribution (including New Zealand and its subantarctic islands, Macquarie Island, Chatham Island, Kerguelen, Marion Island, Crozet Island, Heard Island, Chile, Falkland Islands, South Georgia) (Fig. 4).

Figure 4.

 Map of the Southern Ocean, showing geographic distributions of buoyant hollow-bladed (Durvillaea antarctica; indicated by dotted lines) vs. non-buoyant solid-bladed (indicated by arrows and solid lines) Durvillaea kelp taxa.

The correlation between buoyancy and geographic range of Durvillaea species is, in itself, circumstantial evidence for rafting as a dispersal mechanism for D. antarctica. This pattern apparently contradicts the panbiogeographic assertion that ‘taxa with good means of dispersal were no more widely distributed than those with poor means’ (Grehan, 2001). Moreover, as Durvillaea holdfasts support exceptionally high marine invertebrate biodiversity (Fig. 5; Smith & Simpson, 2002), D. antarctica probably presents an important rafting mechanism for southern coastal invertebrates. Indeed, the Southern Ocean has an estimated 70 million bull-kelp rafts at any one time (Smith, 2002) and these rafts may float for vast distances. Holdfasts of D. antarctica (Fig. 5), for example, are known to wash up on the west and south coasts of Tasmania (Dartnall, 1974; G. J. Edgar, personal communication), regions where D. antarctica does not grow. These presumably originate from a western source such as Kerguelen, some 5000 km distant, or perhaps even Chile. Vicariant biogeographers may be interested to note that some D. antarctica plants are responsible for the distribution of their geological substrates (Garden, 2005; Thiel & Gutow, 2005a), rather than vice versa!

Figure 5.

 Beach-cast bull-kelp (Durvillaea antarctica) from Otago, southern New Zealand (above), and associated holdfast-dwelling invertebrates (below). Photos courtesy of Ceridwen Fraser.

Although ecological evidence for the importance of D. antarctica rafting seems compelling (Thiel & Gutow, 2005a,b; Thiel & Haye, 2006), genetic data are lacking. I therefore propose undertaking phylogeographic analyses of both Durvillaea and its associated invertebrate community to test directly for rafting as a dispersal mechanism. This broad focus, incorporating kelp-dwelling species from diverse taxonomic groups, will also circumvent the ‘anecdotal’ nature of dispersal studies of one or two closely related species. By testing explicit WWD-rafting hypotheses for a number of independent taxa, we can finally put these old criticisms of dispersal biogeography to bed.

Acknowledgements

This paper was presented as part of the ‘Maritime connectivity’ symposium (International Biogeography Society, Third Biennial Conference, Tenerife), organized by Michael Dawson, John Wares and Anuschka Faucci. The author gratefully acknowledges the University of Otago's financial support for international travel, and the Spanish Government's support for the IBS Meeting as a whole. Chris Burridge, Dave Craw, Margaret Finney, Ceridwen Fraser, Bob McDowall, Hamish Spencer, Martin Thiel and Graham Wallis provided constructive comments on early versions of the manuscript, and Chris Garden, Cameron Hay and Abigail Smith provided informative discussions on kelp biogeography. Ken Miller assisted with figure production and Ceridwen Fraser provided bull-kelp photos.

Biosketch

Jonathan Waters is a Senior Lecturer in Zoology at the University of Otago. His research programme focuses on the phylogeography and evolution of Southern Hemisphere marine and freshwater biota. Current projects are centred on the use of freshwater vicariant events to calibrate molecular clocks, and the importance of macroalgae as facilitators for oceanic rafting.

Editor: David Bellwood

Special Issue: This article arose from a paper presented at the third biennial meeting of the International Biogeography Society, held in Puerto de la Cruz, Tenerife, Canary Islands, 9–13 January 2007.

Ancillary