Orchid biogeography and factors associated with rarity in ... - CiteSeerX [PDF]

Journal of Biogeography (J. Biogeogr.) (2011) 38, 487–501

ORIGINAL ARTICLE

Orchid biogeography and factors associated with rarity in a biodiversity hotspot, the Southwest Australian Floristic Region Ryan D. Phillips1,2*, Andrew P. Brown3, Kingsley W. Dixon1,2 and Stephen D. Hopper2,4

1

Kings Park and Botanic Garden, Botanic Gardens and Parks Authority, West Perth, WA 6005, Australia, 2School of Plant Biology, The University of Western Australia, Nedlands, WA 6009, Australia, 3Department of Environment and Conservation, Species and Communities Branch, Locked Bag 104, Bentley Delivery Centre, WA 6983, Australia, 4Royal Botanic Gardens, Kew, Richmond, Surrey TW9 3AB, UK

ABSTRACT

Aim The causes of orchid diversification and intrinsic rarity are poorly resolved. The Orchidaceae of the Southwest Australian Floristic Region use a diversity of pollination strategies and sites of mycorrhizal infection, and occupy a diversity of habitats. We combined a biogeographic analysis with analysis of factors associated with rarity to establish: (1) the landscape features correlated with taxon turnover and speciation, and (2) the possible role in taxon rarity of geographic region, pollination strategy, edaphic habitat and site of mycorrhizal infection. Location Southwest Australian Floristic Region. Methods The distributions of 407 orchid taxa (species and subspecies) were mapped at the quarter-degree scale using 13,267 collections in the Western Australian Herbarium. This database was used to map taxon richness, for a biogeographic analysis and to quantify rarity of taxa. Using herbarium records, rarity was expressed as mean abundance, mean distribution and incidence of rarity based on abundance and distribution for each genus. We tested for differences in rarity of species between pollination strategies, edaphic habitats and sites of mycorrhizal infection. Results Taxon richness was highest in the High Rainfall Province. Biogeographic provincial boundaries for orchids were aligned with rainfall, while district boundaries tended to follow geological formations. When rarity was defined as either low abundance or small distribution, the greatest number of rare taxa occurred in areas of high taxon richness and naturally fragmented edaphic environments. For both abundance and distributional extent, sexual deception had a significantly higher incidence of rarity than food-rewarding taxa. There was no significant difference in rarity with site of mycorrhizal infection. Main conclusions While large-scale edaphic and climatic variation are correlated with orchid taxon turnover and speciation in a similar fashion to the flora in general, the processes responsible for patterns of diversity may differ. Fragmented edaphic environments appear to be associated with a higher incidence of rare species due to limited dispersal/colonization opportunities or radiations of taxa in allopatry. The high incidence of rarity in sexually deceptive taxa could be due to either low fruit set or the risk of specializing on a single pollinator species.

*Correspondence: Ryan Phillips, Kings Park and Botanic Garden, Botanic Gardens and Parks Authority, West Perth, WA 6005, Australia. E-mail: [email protected]

ª 2010 Blackwell Publishing Ltd

Keywords Biogeographic provinces, conservation, endemism, mycorrhiza, Orchidaceae, pollination, rarity, Western Australia.

http://wileyonlinelibrary.com/journal/jbi doi:10.1111/j.1365-2699.2010.02413.x

487

R. D. Phillips et al. INTRODUCTION The Orchidaceae are an exceptionally diverse family of c. 26,000 species (Royal Botanic Gardens, Kew, 2009). The family is characterized by the presence of mycorrhizal endophytes (Rasmussen, 1995) and a diversity of pollination strategies (Adams & Lawson, 1993; Tremblay et al., 2005). The prevalence of pollination by deceit (Cozzolino & Widmer, 2005), the rapid effects of genetic drift in small populations with highly skewed reproductive success (Tremblay et al., 2005), mycorrhizal specificity (Otero & Flanagan, 2006), and habitat specialization (Gravendeel et al., 2004) have all been implicated in the diversification of the family. The relative role of these factors in orchid rarity and diversification is unknown. Research combining the disciplines of pollination and mycorrhizal biology could serve to elucidate the drivers of orchid diversification and rarity and prove crucial to their conservation. Differences in pollination strategy may have profound effects on the ecology and evolution of orchid species. Among orchids, food-rewarding species have the highest fruit set on average, and sexually deceptive species the lowest (Neiland & Wilcock, 1998; Tremblay et al., 2005; Phillips et al., 2009a). Specificity also varies between pollination strategies, with foodrewarding species generally having the lowest pollinator specificity and sexually deceptive species the highest (Cozzolino & Widmer, 2005; Phillips et al., 2009a). Specialized systems with low levels of fruit set may be more predisposed to rarity. Further, more specific pollination systems may lead to more rapid speciation through pollinator-mediated isolation between populations. In the extreme case of the highly specific sexual deception pollination system, speciation may arise rapidly through switching between pollinators. The ecology of mycorrhizal fungi, the specificity of their relationship with orchids, and the phenology of infection may play a role in orchid speciation and rarity (Swarts & Dixon, 2009). For example, rare or highly specific symbionts could afford opportunities for rapid genetic divergence in local allopatry, facilitating the origin of daughter species. The influence of the site of fungal infection on other aspects of orchid biology is unknown. Most orchids are root- or stem (rhizome)-infected (Rasmussen, 1995). However, Australasia is unique in also possessing orchids where the infection occurs primarily in a specialized subsoil stem-collar at the base of the leaf (Ramsay et al., 1986). In collar-infected species, there is a single point of infection at leaf emergence with no spreading roots to increase the probability of intersection with fungal inoculum in the soil. Thus, patchiness in fungal distribution in soil may act to limit recruitment and plant survival compared with root-infected species. The influence of infection site on rarity has not been investigated. Delineation of biogeographic provinces and centres of rarity gives an indication of the broad-scale features responsible for species turnover, restricted distributions and, historically, speciation events (e.g. Stebbins & Major, 1965; Hopper & Gioia, 2004). Analysis of the factors associated with rarity 488

could reveal whether any strategy has a predisposition to rarity and is limiting distribution at a more local scale. Coupling these two approaches has the potential to provide initial clues to the features influencing orchid speciation and rarity. The Orchidaceae of the Southwest Australian Floristic Region (SWAFR, sensu Hopper & Gioia, 2004) are an ideal study group with which to adopt this approach because of the diverse array of pollination strategies (Brown et al., 2008), mycorrhizal infection sites (Ramsay et al., 1986), and intrinsically rare species (Brown et al., 1998). All orchid taxa in the SWAFR are terrestrial herbaceous geophytic perennials, with Cryptostylis ovata R.Br. the sole evergreen species (Brown et al., 2008). Rarity can be defined in terms of abundance, distributional extent and habitat specificity. Rabinowitz (1981) combined these three variables to establish seven possible forms of rarity. However, in the Western Australian flora, rare species tend to occur on predictable soil types, although their local presence within soil types is far less predictable and is linked to complex biological and environmental interactions over millions of years (e.g. Byrne & Hopper, 2008). To make a start in formulating testable hypotheses, we followed the definition of rarity by Gaston (1994), where rarity is defined simply in terms of low abundance or small range size. A general model for the patterns of richness and rarity for the flora of the SWAFR has been presented by Hopper (1979) and expanded upon more recently by Hopper & Gioia (2004; Fig. 1). Based on species composition, four broad provinces have been delineated. In order of decreasing species richness, the provinces are the Transitional Rainfall Province (TRP, 300–600 mm rainfall per annum); the South-east Coastal Province (SCP, 300–600 mm); the High Rainfall Province (HRP, 600–1500 mm) and the adjoining Arid Zone (AZ, < 300 mm). The TRP and SCP form the Transitional Rainfall Zone (TRZ) with rainfall decreasing progressively inland. Nodes of particularly high species richness and endemism occur in all the more mesic provinces, but particularly in the TRZ (Hopper & Gioia, 2004). The high diversity in the TRZ arose from the more diverse topography and erosional dynamism and climatic fluctuation of the Neogene and Quaternary creating the opportunity for repeated bouts of speciation (Hopper & Gioia, 2004). Given the unique relationships orchids have with their pollinator and mycorrhizal endophytes, the process may differ from those that have generated the remarkable floristic diversity and endemism of this region, which is dominated by woody perennials. Patterns of biogeography and rarity may help to resolve the influences of pollinators, mycorrhiza, edaphic conditions and historical events in determining speciation and rarity in the orchids of the SWAFR. We tested the following hypotheses: (1) the patterns of orchid species richness and endemism match those of the flora in general; (2) biogeographic provinces for orchids correspond to climatic and edaphic variation; (3) the proportion of rare species varies with site of fungal infection and pollination strategy; and (4) naturally Journal of Biogeography 38, 487–501 ª 2010 Blackwell Publishing Ltd

Orchid biogeography and rarity 120°E

116°E

124°E

26°S

26°S Sampled species richness

Floristic Region Floristic Province Floristic District

375

112°E

Kalbarri 0

Wongan 30°S

Mt Lesueur

Transitional Rainfall Province Narrogin

Lesueur

High Rainfall Province

Fitzgerald Boxwood Hills

Stirling Muir Leeuwin Walpole Albany Naturaliste Ridge Stirling Range 112°E

Southeast Coastal Province

Hyden

Perth Greater Perth Darkan 34°S

30°S

116°E

120°E

0

34°S

Esperance

100 km

124°E

Figure 1 Species richness and biogeographic regions for the Southwest Australian Floristic Region for all plant taxa. Locations mentioned in the text are also included. Figure modified from Hopper & Gioia (2004).

fragmented habitats have a higher incidence of rare species than continuous habitats. MATERIALS AND METHODS The distribution of 407 currently recognized native orchid taxa from the SWAFR was mapped as presence/absence data on a grid of quarter-degree cells using the 13,267 independent records from the Western Australian Herbarium (Perth) as of June 2006. The Western Australian Herbarium has the most accurate distributional data available for the Orchidaceae in south-western Australia (cf. Hopper, 1983; Hoffman & Brown, 1998; Jones, 2006; Brown et al., 2008). Surveys underpinning major taxonomic revisions of most Western Australian orchid genera have greatly improved our knowledge of conservation status and distribution (see references in Appendix S1 in Supporting Information). The species that are poorly represented in the Western Australian Herbarium are generally those recently recognized taxa where extensive collections are yet to be made. For taxa that are poorly collected, but have well known distributions, additional sites were included (see Appendix S2). Taxa were included if: (1) formally described, (2) not yet formally described but listed as ‘phrase name’ taxa on the Western Australian Herbarium database (these taxa are recognized as distinct but are awaiting formal taxonomic description) (Western Australian Herbarium 2009), or (3) while included in the herbarium database only as an ‘aff.’ of another species, they are known to be morphologically consistent and are awaiting taxonomic description. The Journal of Biogeography 38, 487–501 ª 2010 Blackwell Publishing Ltd

undescribed species included are detailed in Appendix S3 and are illustrated in Brown et al. (2008) with voucher herbarium specimens cited on pp. 21–23. Diuris corymbosa Lindley was discounted from all analyses on the basis that it is believed to represent a species complex with very poorly known species boundaries (A.P. Brown, unpublished data). Working in the SWAFR, Hopper & Gioia (2004) found little difference between raw collection data and data standardized for collection effort, so no standardization for collector effort was undertaken. The larger size of the cells in the north of the region means there is a progressive increase northwards in a bias towards high species richness. However, the trend evident in the present study is sufficiently strong for this bias to be inconsequential (R.D. Phillips, unpublished data). To enable comparison with the patterns of richness exhibited by the flora in general, taxon richness for all quarterdegree grid squares was plotted on a map of southern Western Australia (Fig. 2). A similar map encompassing all the flora of the SWAFR was presented in Hopper & Gioia (2004) and has been reproduced here (Fig. 1). A map of the number of rare taxa per cell was also produced (for definition of rarity, see below). The unweighted pair-group method using arithmetic averages (UPGMA) in Primer 5.0 (Belbin, 1994) was used to delineate orchid biogeographic provinces. Euclidean distance was used as the measure for the distance matrix. Ordinations generated by non-metric multidimensional scaling (100 randomizations) in Primer 5.0 (Belbin, 1994) were used to confirm the discreteness of clusters. For comparison with the 489

R. D. Phillips et al.

28 S

30 S

1 – 10 species 11 – 25 species 26 – 50 species 51 – 75 species 76 + species

32 S 126 E

124 E

34 S 120 E 116 E

118 E

122 E 0

100 km

Figure 2 Taxon richness of orchids in the Southwest Australian Floristic Region. Cells represent quarter-degree squares.

provinces and districts of the flora in general, see Fig. 1 (modified from Hopper & Gioia, 2004). Following the nomenclature of Hopper & Gioia (2004), major biogeographic divisions are referred to as orchid provinces, while minor divisions are referred to as orchid districts. Orchid provinces were established using presence/absence of taxa in one-degree grid squares. A finer resolution for areas of higher taxon richness was achieved by repeating the UPGMA analysis at the half-degree grid-square scale. For both analyses, grid squares were included only if they contained more than 15 taxa and more than a quarter of the grid square was terrestrial. The number of taxa endemic to each orchid district, the total number of taxa, and the area of the district were tabulated to enable calculation of percentage endemism and endemism per unit area. Rarity was defined in terms of abundance and distributional extent. We conducted separate analyses using mean abundance and mean distribution, and the incidence of rarity based on abundance and distribution. Incidence of rarity was included because, if means are used, the large number of herbarium records for a small subset of species could mask associations present in the remainder of the taxa. The number of independent specimen records in the Western Australian Herbarium was used as a surrogate measure of abundance (e.g. Holmgren & Poorter, 2007). A taxon was classified as rare if there were fewer than 10 independent herbarium records. Cases where the species had been collected on 10 or fewer occasions, but is known to be more widely distributed, were not included as rare (A.P. Brown, unpublished data). Distributional extent was quantified by using the number of quarterdegree grid squares from which a taxon was recorded in the Western Australian Herbarium (e.g. Peat et al., 2007). Taxa were classed as rare if they were recorded from six or fewer quarter-degree cells. This classification of rarity follows from Gaston (1994), where rare species were the least abundant 25% 490

of species. In the present study, this value was modified slightly to align the cut-off with a point where there was a natural disjunction in the species abundance distribution. Distributional extent was not calculated using minimum convex polygons formed from the locations of herbarium records, because numerous taxa have naturally fragmented distributions or are known from outlying populations or individuals well beyond the regular distribution (for examples see Hopper & Brown, 2001, 2004, 2007). The number of rare taxa collected per quarter-degree cell in the Western Australian Herbarium was plotted for the SWAFR. For each orchid biogeographic district, we calculated the average number of rare taxa per quarter-degree cell (in terms of both abundance and distribution) and the average proportion of taxa classed as rare per cell. For each genus, the mean number of herbarium records, occupied grid squares, and the proportion (incidence) of rare taxa (in terms of both distribution and abundance) were calculated. For the analysis of pollination strategy, Caladenia was treated at the subgeneric level (Hopper & Brown, 2001) because the genus is unique in containing multiple origins of the sexual deception pollination strategy (Kores et al., 2001; Phillips et al., 2009a). Both food deception and sexual deception have been recorded in Caladenia subg. Calonema and Caladenia subg. Phlebochilus (Stoutamire, 1983; Phillips et al., 2009a). In Caladenia, means were calculated separately for each pollination strategy. Using genera as replicates, we used Kruskal–Wallis tests to test for differences between pollination strategies and sites of mycorrhizal infection in (1) the mean number of herbarium records, (2) the mean number of occupied grid squares, (3) the proportion of rare taxa in terms of abundance, and (4) the proportion of rare taxa in terms of distribution. For the analysis of habitat, species were used as replicates and Kruskal–Wallis tests undertaken on abundance and distributional extent. In all analyses, Kruskal– Wallis tests were used rather than ANOVA because the Journal of Biogeography 38, 487–501 ª 2010 Blackwell Publishing Ltd

Orchid biogeography and rarity variances were heterogeneous. For both analyses, when significant variation was detected, Mann–Whitney U-tests were used to establish the source of the variation. All statistical tests were undertaken in spss 11.0. Phylogenetic independent contrasts were not used because the layout of the phylogeny resulted in the use of the same clades in multiple contrasts, inflating the degrees of freedom. The classifications of genera into sites of fungal infection and mechanisms of pollination attraction are given in Appendix S4. Categories of fungal infection sites follow those of Ramsay et al. (1986): (1) stem tuber infection, (2) underground stem infection, (3) stem-collar infection, (4) root infection, and (5) root–stem infection. The sole departure from this classification is for the genus Drakaea, which was originally classed in category (5) but has since been found to have infection pattern (3) (K.W. Dixon, unpublished data). Species were categorized by pollination strategy based on the published literature (Appendix S4) and field observations (A.P. Brown and R.D. Phillips, unpublished data). We recognized three broad pollination strategies based on the mechanism of attraction: food reward, food deception, and sexual deception. Species that self-pollinate but also utilize one of these attraction mechanisms were included within the relevant attraction category. The only cleistogamous taxon within the study region, Caladenia bicalliata subsp. cleistogama Hopper & A.P. Brown (Hopper & Brown, 2001), was discounted from the analyses of pollination strategy. In cases where the mechanism of pollinator attraction has not been recorded, species were classified on the basis of closely related congeners. While the groups responsible for pollination of Corybas, Pterostylis and Rhizanthella are known (Jones, 2006; Brown et al., 2008), the mechanisms of attraction have not been established (Adams & Lawson, 1993), and these genera have thus been omitted from analyses involving pollination strategy. For the analysis of habitat type, all species were classified according to habitats using Hopper & Brown (2001, 2004, 2006, 2007), Brown et al. (2008), and the habitat descriptions provided with collections in the Western Australian Herbarium (2009). We followed the definition of habitat provided by Hall et al. (1997): ‘habitat is the resources and conditions present in an area that produce occupancy – including survival and reproduction – by a given organism’. Due to the rapid turnover of species in the SWAFR (Hopper & Gioia, 2004), habitat classifications were based primarily on edaphic conditions and to a minor extent on rainfall. Taxa were classified as occurring in the following habitat types (sensu Beard, 1981): (1) coastal: confined to coastal dune or limestone formations, (2) granite: confined to shallow soils on granite outcrops and inselbergs, (3) salt lake: confined to the margins of inland salt lakes, (4) swamp: confined to either permanent or seasonally inundated swamps, creeklines and moist flats, (5) woodland: confined to forests, woodlands and heathlands, and (6) variable: occur in more than one of the above habitats. Species that typically occur in only one habitat but are rarely recorded on contrasting habitats were Journal of Biogeography 38, 487–501 ª 2010 Blackwell Publishing Ltd

classified under their typical habitat, rather than as being of variable habitat preference. Forests, woodlands and heathlands represent continuous, generalized habitats in comparison with granite, salt lake and swamp habitats (Beard, 1981), and were all classed together under the woodland category. Many species are shared between forest, woodland and heathland habitats and, when analysed separately, show no difference in the degree of rarity (R.D. Phillips, unpublished data). The fragmented and continuous environments have a similar geographic extent, but fragmented environments have a smaller total area. We calculated the incidence of rarity (in terms of both abundance and distribution) for each habitat type, and calculated the mean abundance and distribution based on herbarium records. Due to heterogeneity of variances we used a Kruskall–Wallis test to test for differences between habitats, and Mann–Whitney U-tests to establish the source of the variation. RESULTS Orchid biogeographic divisions and endemism Orchid species richness was highest in the HRP, followed by the SCP, the TRP and the AZ (nomenclature of regions follows Hopper & Gioia, 2004) (Appendix S5; Fig. 1). Coastal areas of the HRP had high richness, with nodes occurring at the Swan Coastal Plain, Leeuwin–Naturaliste Ridge, and the south coast between Walpole and Albany. All these regions contain relatively high rainfall and a diversity of edaphic environments, particularly forests (on varying soils), swamps and coastal dunes. Using degree blocks, broad-scale biogeographic orchid provinces corresponded closely to the HRP, SCP and TRP presented in Hopper & Gioia (2004) (Figs 3 & 4). Within the TRP, this analysis recognized the Kalbarri and Northern Wheatbelt regions as orchid districts (Fig. 3). These areas of comparatively low diversity were omitted from analysis at the half-degree scale. At the half-degree scale, the 12 orchid districts were recognized (Fig. 4). The Moore region was a discrete cluster that, in the ordination, was intermediate between the HRP and the TRP (Fig. 5). This region lies on the margin between the Swan and Northern Sandplain orchid districts, and has no species unique to it or having its centre of distribution within it. The Moore cluster was included in the Northern Sandplain orchid district due to closer proximity to the TRP sites in the ordination, and to having a relatively low species richness, equivalent to those observed in the Northern Sandplain. The Leeuwin–Naturaliste, Southern Forests and Swan orchid districts had the highest number of endemics for their areas (Appendix S5). The Esperance (SCP), Kalbarri and Northern Sandplain (TRP) orchid districts had a moderately high level of orchid endemism. Kalbarri had a particularly high number of endemics relative to species richness. While the Northern Wheatbelt had a number of endemics similar to the districts of the HRP, it is over five times larger than the next 491

R. D. Phillips et al.

Southern Wheatbelt Kalbarri

Central and Northern Wheatbelt

Northern Sandplain Brookton High Rainfall Province

15

34 / 119 33 / 122 33 / 123 33 / 119 33 / 120 33 / 121 33 / 118 33 / 117 32 / 117 28 / 114 27 / 114 28 / 115 32 / 121 32 / 120 32 / 123 29 / 116 30 / 119 31 / 119 31 / 120 29 / 117 30 / 118 30 / 117 31 / 117 31 / 118 32 / 118 32 / 119 30 / 116 29 / 115 30 / 115 31 / 116 32 / 116 35 / 116 34 / 115 35 / 117 33 / 115 31 / 115 32 / 115 34 / 118 33 / 116 34 / 116 34 / 117

Esperance

Trainfall Rainfall Province

10

South-east Coastal Province

5

0

Distance

South-west Coastal Southern Forests

Figure 3 UPGMA cluster analysis of orchid taxon composition in one-degree squares in the Southwest Australian Floristic Region. Only squares from which > 15 taxa were recorded were included in the analysis. Clusters were used to delineate orchid botanical provinces. Regions given are approximate, and were further refined by cluster analysis at the half-degree scale before being classed as districts. Numbers adjacent to the dendrogram refer to latitudes and longitudes of degree cells.

28 S Kalbarri

Northern Wheatbelt

30 S Northern Sandplain Brookton Swan

32 S

Southwest Wheatbelt Esperance Mallee

Darkan

Southern Wheatbelt

34 S

Fitzgerald

Stirlings

Esperance

Southern Forests

Leeuwin Naturaliste 116 S

118 S

120 S

122 S

largest district. As such, the mesic south-western districts and the Kalbarri district are the most significant regions in terms of orchid endemism. Rarity A total of 85 taxa were classified as rare based on abundance, and 107 based on distributional extent. Seventy-nine species were classed as rare under both definitions. The number of 492

124 S

126 S

Figure 4 Orchid biogeographic provinces and districts of orchids of the Southwest Australian Floristic Region. Dark grey = High Rainfall Province; light grey = Southeast Coastal Province; white = Transitional Rainfall Province.

herbarium records and distributional extent was strongly correlated (R = 0.96). Accordingly, the geographic pattern of rarity was very similar when species were classed as rare based on abundance or distribution (Fig. 6). At the provincial level, the HRP generally had the highest numbers of rare species per cell. Within this province, the Leeuwin–Naturaliste Ridge and the Swan Coastal Plain had exceptional numbers of rare species (Fig. 6). In the TRP, the Kalbarri district had a similar number of rare species to that exhibited by the HRP (Table 1). When Journal of Biogeography 38, 487–501 ª 2010 Blackwell Publishing Ltd

Orchid biogeography and rarity 1

1

(a)

0

0 0

-1

-2

-2

1

0

-1

1

2

Dimension 2

-3

-1

-1

-2

-2

Dimension 1

Dimension 2

(b)

Dimension 3

1

1

0.5

0.5

0

0 -1

0

1

0.5

0.5

0.5

Dimension 2

0

Figure 5 A non-metric multidimensional scaling ordination of the orchid taxa composition of the orchid botanical districts of the Southwest Australian Floristic Region. Each point represents a quarter-degree square. While one ordination was produced, close-ups of the different provinces are presented to allow districts to be discerned. (a) High Rainfall Province, (b) South-east Coastal Province, (c) Transitional Rainfall Province.

1

0

-1

-0.5

1

-0.5

-1.5

-1.5

Dimension 3

Dimension 1

Southern Forests Swan Leeuwin - Naturaliste Darkan - Stirlings

rarity was considered as the proportion of rare species, differences between provinces were minimal regardless of whether rarity was considered in terms of abundance or distribution. However, in both cases the Kalbarri district stood out as a region with an exceptional proportion of rare species (Table 1). The site of mycorrhizal infection showed no significant relationship with incidence of rarity, abundance or distributional extent (Table 2). Pollination strategy showed no significant relationship with mean abundance or mean distributional extent (Table 3). There was significant variation Journal of Biogeography 38, 487–501 ª 2010 Blackwell Publishing Ltd

1

Dimension 3

Dimension 1

(c)

0

-1

Northern Wheatbelt Northern Sandplain Esperance Mallee Southwest Wheatbelt Southern Wheatbelt

Fitzgerald Esperance Brookton Moore

in the incidence of rarity between pollination strategies when rarity was expressed in terms of abundance and distributional extent (abundance: P = 0.028, distribution: P = 0.037; Table 3). Sexual deception had the highest incidence of rarity, food deception showed an intermediate level, and almost no rare taxa provide a food reward (Table 3). When rarity was classified based on abundance or distribution, species with variable habitat requirements had the lowest incidence of rarity (Table 4). Woodland and coastal areas were intermediate, granite and swamp had high incidence, and salt lakes had an extremely high incidence of rarity (Table 4). 493

R. D. Phillips et al. (a)

28 S

30 S

1 species 2 species 3 species 4-5 species 6-8 species

32 S 126 E

34 S

124 E 120 E

122 E 0

116 E

100 km

118 E

(b)

28 S

30 S

1 species 2 species 3 species 4-5 species 6-8 species

32 S 126 E

34 S

124 E 120 E

122 E 0

116 E

100 km

118 E

Figure 6 Incidence of rarity of orchid taxa in the Southwest Australian Floristic Region. Cells represent quarter-degree squares. Taxa were classed as rare if they are known from (a) 10 or fewer herbarium records (abundance), or (b) six or fewer quarter-degree cells (distribution).

There was significant variation in the abundance and distributional extent of taxa between habitats (P < 0.05, Kruskal– Wallis test; Table 4). Abundance and distributional extent were significantly higher for species occurring in woodlands or with variable habitat preferences than in the remaining habitats (Table 4, P < 0.05, Mann–Whitney U-test). Of the remaining habitats, the sole significant difference was between the greater abundance of swamp-dwelling versus granite-dwelling taxa. DISCUSSION Biogeography and orchid species richness The Orchidaceae of the Southwest Australian Floristic Region (SWAFR) show a pattern of species richness markedly different 494

from the remainder of the flora. For the majority of plant genera for which there are data available, highest species richness occurs in the Transitional Rainfall Zone (TRZ), though there are exceptions (Hopper, 1979; Hopper et al., 1992; : Lyons et al., 2000; Phillips et al., 2009b). For some genera of annuals and perennial herbs, the south coast appears to have been a major centre of speciation (Hopper et al., 1992; Phillips et al., 2009b). For the vascular flora in general, nodes of high species richness occur at Mt Lesueur in the Transitional Rainfall Province (TRP), Stirling Range, Boxwood Hills (east of the Stirling Range) and Fitzgerald in the South-east Coastal Province (SCP), and the Swan Coastal Plain centred on Perth (High Rainfall Province, HRP), with several less prominent nodes in the TRP (Hopper & Gioia, 2004). Alternatively, orchids have their highest diversity in coastal areas of the HRP Journal of Biogeography 38, 487–501 ª 2010 Blackwell Publishing Ltd

Orchid biogeography and rarity Table 1 Geographic variation in orchid rarity in the Southwest Australian Floristic Region. Rare species per cell Orchid province

Regions

Hopper & Gioia (2004)

Abundance

Distribution

High Rainfall Province (HRP)

DST LN SF SW BRO FITZ ESP ESPM

Darkan, Stirling W. Muir N., S., E. Muir Greater Perth N. Darkan Fitzgerald Esperance N. Esperance, N. Fitzgerald Kalbarri Lesueur Wongan, Hyden S. Hyden Narrogin

0.73 1.44 1.06 1.82 0.38 0.62 0.50 0.12

± ± ± ± ± ± ± ±

0.18 0.49 0.20 0.28 0.18 0.12 0.17 0.08

1.08 1.44 1.51 2.23 1.12 0.5 0.60 0.35

± ± ± ± ± ± ± ±

0.25 0.45 0.26 0.38 0.40 0.15 0.16 0.12

2.55 2.80 2.66 4.19 1.32 4.21 2.34 1.29

± ± ± ± ± ± ± ±

0.64 0.91 0.54 0.63 0.78 0.93 0.82 0.93

3.77 2.87 3.49 5.53 2.81 2.75 2.74 2.69

± ± ± ± ± ± ± ±

0.89 0.78 0.59 1.18 0.95 0.80 0.79 1.07

1.95 0.56 0.14 0.15 0.32

± ± ± ± ±

0.34 0.12 0.04 0.08 0.13

1.82 0.80 0.15 0.2 0.58

± ± ± ± ±

0.36 0.15 0.04 0.09 0.14

14.64 4.57 1.30 1.31 2.41

± ± ± ± ±

2.57 1.02 0.42 0.83 1.25

13.89 6.13 1.37 1.59 3.82

± ± ± ± ±

2.76 1.06 0.42 0.85 1.07

South-east Coastal Province (SCP)

Transitional Rainfall Province (TRP)

KAL NSAND NWB SWB SWWB

Abundance (%)

Distribution (%)

Cells were a quarter-degree in size. Taxa were classed as rare if known from 10 or fewer herbarium records (abundance), or six or fewer quarter-degree cells (distribution). Regions: DST = Darkan–Stirling Range, LN = Leeuwin–Naturaliste, SF = Southern Forests, SW = Swan (High Rainfall Province), BRO = Brookton (margin of High Rainfall Province), FITZ = Fitzgerald, ESP = Esperance (South-east Coastal Province), ESPM = Esperance Mallee, KAL = Kalbarri, NSAND = Northern Sandplain, NWB = Northern Wheatbelt, SWB = Southern Wheatbelt, SWWB = Southwest Wheatbelt. For location of regions, see Figs 1 and 4. Regions defined by Hopper & Gioia (2004) are for the entire vascular flora.

Table 2 Differences in the abundance, distribution and incidence of rarity in orchid genera in relation to site of mycorrhizal infection in the Southwest Australian Floristic Region. Parameter

Infection

n

Mean ± SE

Rank average

Chi-squared

P

Abundance

1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

2 2 9 12 2 2 2 9 12 2 2 2 9 12 2 2 2 9 12 2

16 32.04 82.25 54.06 45.29 9.5 18.67 50.29 32.99 24.46 50 6.82 15.62 14.75 0 50 17.42 18.91 16.82 12.5

4.50 10.50 17.67 13.75 12.00 4.50 10.50 17.11 14.25 11.50 17.25 17 14.11 13.95 7.5 17 16.75 13.88 13.375 12.5

5.313

0.257

5.239

0.263

2.282

0.684

0.758

0.944

Distribution

Rarity – abundance

Rarity – distribution

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

6.00 11.87 18.14 10.76 25.04 4.50 4.25 13.29 5.89 13.21 50 2.27 6.79 6.98 0 50 9.84 8.02 7.08 12.5

Mean = parametric mean; statistical test = Kruskall–Wallis test. Classification of infection sites follows Ramsay et al. (1986). Taxa were classed as rare if they are known from 10 or fewer herbarium records (abundance), or six or fewer quarter-degree cells (distribution).

(Fig. 2), though some areas of high species richness are in common (e.g. Swan Coastal Plain). The high diversity in highrainfall regions might be attributed partially to greater habitat diversity and niche conservatism, with orchids being a predominantly mesic family (Cribb et al., 2003). High species Journal of Biogeography 38, 487–501 ª 2010 Blackwell Publishing Ltd

richness in the HRP indicates that a high proportion of species evolved here, necessitating an alternative model of speciation to climatic fluctuation acting in concert with geological variability in the TRP to drive diversification. While there is some evidence that the effects of climatic fluctuation may have 495

R. D. Phillips et al. Table 3 Differences in abundance, distribution and incidence of rarity in orchid genera in relation to pollination strategy in the Southwest Australian Floristic Region. Parameter

Pollination

n

Mean (SE)

Mean rank

Chi-squared

P

Abundance

Sex Food Reward Sex Food Reward SexAB FoodaC Rewardbc SexDE Foodd Rewarde

9 13 8 9 13 8 9 13 8 9 13 8

48.39 65.21 63.81 27.96 39.28 38.46 31.90 14.92 1.44 32.43 19.27 2.40

13.11 15.69 17.88 12.89 15.65 18.19 20.28 15.88 9.5 19.72 16.27 9.5

0.609

0.552

0.758

0.479

7.12

0.028

6.55

0.037

Distribution

Rarity – abundance

Rarity – distribution

± ± ± ± ± ± ± ± ± ± ± ±

12.97 14.28 11.04 7.69 8.79 6.68 8.86 4.38 1.01 9.39 4.98 1.62

Mean = parametric mean; statistical test = Kruskall–Wallis test. Pollinator attraction strategy: sex = sexual deception, food = food deception, reward = food reward provided. Taxa were classed as rare if they are known from 10 or fewer herbarium records (abundance), or six or fewer quarter-degree cells (distribution). Letters indicate significant differences at P < 0.05 (upper case, higher value; lower case, lower value).

Table 4 Differences in abundance, distribution and incidence of rarity (in terms of abundance and distribution) in orchid genera in relation to habitat in the Southwest Australian Floristic Region. Habitat

Total taxa

Rare taxa

Percentage rare

Abundance (±SE)

Coastal Granite Salt lake Swamp Variable Woodland

22 25 6 72 68 211

8 9 4 21 1 40

36.4 36.0 67.0 29.2 1.5 19.0

10.9 10.5 6.8 18.9 57.3 35.5

± ± ± ± ± ±

1.7ab 3.0abd 1.1abc 2.3abD 7.3AC 2.8Bc

Rare taxa

Percentage rare

Distribution (±SE)

7 10 5 26 2 55

31.8 40.0 83.3 36.1 2.9 26.1

7.3 7.7 4.5 11.1 34.7 20.5

± ± ± ± ± ±

1.0ab 1.8ab 1.3abc 1.3ab 1.3AC 1.5Bc

Variable: taxa that utilize more than one of the listed habitats. For both mean abundance and mean distribution, there is significant variation between habitats (P < 0.05, Kruskal–Wallis test). The source of variation was established using Mann–Whitney U-tests. Letters indicate significant differences at P < 0.05 (upper case, higher value; lower case, lower value).

influenced speciation and population genetic structure in wetter regions (Wheeler & Byrne, 2006), the unique pollinator and mycorrhizal specialization in orchids may have played a pivotal role in diversification within the HRP. Despite a different pattern of richness, orchids exhibit phytogeographic provinces similar to those of the entire flora (Fig. 1; Hopper & Gioia, 2004). Rainfall appears to determine main boundaries between provinces, with the 600-mm isohyet being closely correlated with the boundary of the TRP for both orchids and the entire flora. At a more local scale, soil type is also critical (Hopper, 1979). For example, the boundaries between the Swan and Darkan–Stirling orchid provinces coincide with a change from sandplain to predominantly lateritic soil. Similarly, the Leeuwin–Naturaliste ridge forms a separate orchid province from the remaining high-rainfall areas. Therefore at regional scales, rainfall is the dominant factor correlated with orchid species turnover and speciation between regions, while at subregional scales, edaphic specialization is a critical correlate. The correlation of climatic and edaphic variables with species turnover and the boundaries between sister taxa 496

suggests that these environmental variables may play a role in the speciation of orchids in the SWAFR. In the SWAFR, support for provincial boundaries being regions favouring genetic divergence and speciation is accumulating through population genetic and phylogeographic research in other families (e.g. Kennington & James, 1998; Wheeler & Byrne, 2006). At a finer scale, Bussell et al. (2006) revealed that taxa from several families show genetic provenances within the Swan Coastal Plain, either from north to south or between soil types. Due to insufficient data for the SWAFR, it is unknown whether local specialization for edaphic environment, turnover of pollinator species, or turnover of potential mycorrhizal endophytes is responsible for the changes in orchid composition across environmental gradients. When considering the flora in general, areas of endemism largely coincide with centres of high species richness, although there is a pronounced relative increase in endemism on the northern sandplain (Hopper & Gioia, 2004). In the Orchidaceae, areas of high endemism also tended to follow areas of high species richness (Appendix S5). This analysis confirms the HRP as a centre of endemism, but also brings attention to the Journal of Biogeography 38, 487–501 ª 2010 Blackwell Publishing Ltd

Orchid biogeography and rarity Kalbarri district, which has a relatively high number of orchid endemics and a very high level of endemism relative to the number of orchid taxa present. This provides further evidence that the processes driving the accumulation of orchid diversity in the SWAFR may differ from those of the flora in general. Rarity The number of rare orchid taxa showed strong geographic variation. Due to high species richness, regions in the HRP tended to have the highest number of rare taxa in terms of both abundance and distributional extent (Table 1). Within the HRP, the Leeuwin–Naturaliste Ridge, the Swan Coastal Plain and parts of the south coast had exceptionally high numbers of rare taxa (Fig. 6). The Leeuwin–Naturaliste Ridge is a unique formation within the HRP of a granitic, faultbounded horst of Precambrian antiquity (Myers, 1990) with diverse surface soils of granite outcropping, limestone, laterite, coastal sands and loams. In the TRP, the Kalbarri region had an exceptionally high number of rare orchid taxa. The Kalbarri region has a diverse geology including sandplain and the granitic Northampton complex, and is the only part of the SWAFR containing part of the Carnarvon Basin (Hocking, 1990; Myers, 1990). There are several instances of taxa from species complexes usually associated with more mesic environments to the south that have persisted in occasional seasonally moist, relictual environments and speciated, resulting in a high level of orchid endemism (Brown et al., 2008; Phillips et al., 2009b). The characteristics of these regions may indicate a role of unique edaphic environment and a relictual state in the rarity of taxa. Orchid taxa that are rare in terms of low abundance and/or restricted distributions were most strongly associated with the naturally fragmented habitats of salt lakes, granites and swamps within the SWAFR. While the semi-arid salt lakes often form continuous systems along ancient palaeorivers (Beard, 1999), the area of available habitat is restricted to a narrow, intermittent strip around the periphery of the lake that remains non-saline but seasonally moist. The prevalence of naturally rare species from these habitats may be the result of the rarity of suitable habitat, low colonization possibilities due to the disjunct nature of suitable habitat, or a radiation of taxa through allopatric speciation. In these habitats, taxa extend their distribution further into drier regions than is typical (Hopper & Brown, 2001, 2004; Brown et al., 2008), providing further evidence that moist environments play a role in supporting often rare, relictual taxa. The combination of geographic patterns of rarity and pronounced variation in rarity between habitats suggest that specialization with edaphic environment may be the primary determinant of intrinsic rarity of orchids in the south-west. This could arise directly from physiological specialization with edaphic environment, or indirectly through an absence of suitable pollinators and mycorrhiza in other edaphic environments. Pollination strategy was shown to have a significant association with the incidence of rarity in terms of both Journal of Biogeography 38, 487–501 ª 2010 Blackwell Publishing Ltd

abundance and distributional extent. Sexually deceptive genera exhibit a significantly higher incidence of rarity than foodrewarding genera. This correlation was not evident when considering mean abundance and distribution, and is probably due to the large effect that a small number of abundant species have on the means, particularly in monotypic genera. There was no evidence for a difference in incidence of rarity between sexually deceptive and food-deceptive clades. However, a resolved phylogeny of the Caladeniinae would allow inference of the number of evolutions of each pollination strategy and when they occurred, permitting a more powerful analysis using phylogenetic independent contrasts. The difference in incidence of rarity could be driven through greater fruit set from the provision of a reward. Orchids that produce floral nectar have, on average, higher fruit set (Neiland & Wilcock, 1998; Tremblay et al., 2005), and in some cases nectar supplementation can result in a higher visitation rate (Smithson & Gigord, 2001; Jersakova & Johnson, 2006). Furthermore, limited investigations of Ophrys and sexually deceptive Australian species have revealed, on average, comparatively low fruit set (Tremblay et al., 2005; Phillips et al., 2009a). However, some genera of sexually deceptive orchids that occur in the SWAFR show high fruit set that is comparable with that in some rewarding species (Phillips, 2010). Whether low fruit set results in rarity depends on the availability of orchid recruitment sites and the presence of suitable mycorrhiza. This concept must be evaluated by combining studies on pollination ecology with data on the longevity of individual plants and the availability of recruitment sites with suitable mycorrhizal endophytes. An alternative hypothesis is that the specialization for a single pollinator in sexual deception (e.g. Coleman, 1930; Stoutamire, 1983; Phillips et al., 2009a) leaves the orchid vulnerable to changes in pollinator abundance. This is less likely to arise in foodrewarding or food-deceptive species, which attract a suite of foraging insects. This hypothesis would be supported if it is shown that rare species are associated with rare pollinators. Previous studies of rarity in orchids have focused on the role of pollination strategy, with varying results. Neiland & Wilcock (1998) found that, in Britain, rarity is associated with nonrewarding species. Alternatively, in the Netherlands, orchid rarity is related to habitat rather than pollination strategy, with orchids confined to wet grasslands and heathlands suffering greater losses than those confined to forests or calcareous grasslands (Jacquemyn et al., 2004). In Estonia, species associated with calcareous grassland and woodland habitats showed the greatest decline (Kull & Hutchings, 2006). The results of the present study support those of Neiland & Wilcock (1998). However, the variation in conclusions from these four studies suggests that the drivers of rarity in orchids are dependent on regional variation in anthropogenic impacts and biology of the orchids. It was predicted that collar-infected genera would show a greater predisposition to rarity due to a perceived restricted ability to acquire fungal endophytes. However, there was no evidence from this study that site of mycorrhizal infection is 497

R. D. Phillips et al. linked to rarity. Furthermore, there are widespread and common genera in four of the five mycorrhizal infection types. We propose that, for Western Australian terrestrial orchids, the major limiting factor on recruitment is locating a suitable fungus for germination, although lateral growth of roots may increase the likelihood of encountering suitable fungi, thereby enhancing the effectiveness of clonality. If mycorrhizal associations do play a role in rarity, it is more likely to result from specificity of the relationship or coarse differences in the distribution of fungi within the environment in response to microhabitat. An issue with any analysis of rarity in a human-modified landscape is the role of both intrinsic and human-induced rarity. We have focused on the intrinsic features of the biology of species on rarity. However, how anthropogenic influences affect the analysis must be considered. An inherent assumption is that anthropogenic influences will be equal on all pollination strategies, for all sites of mycorrhizal infection, in all geographic regions, and across all habitats. No data are available to evaluate the effect of anthropogenic effects on the breakdown of pollinator and mycorrhizal relations in the SWAFR. The anthropogenic influence varies considerably between regions and habitats, particularly from land clearing (Shepherd et al., 2002), with the highest impact on the woodlands of the TRP and woodlands and swamps of the Swan Coastal Plain. In the TRP, very few rare taxa occur in woodlands, so this will have a negligible impact on the analysis. Due to few early collections and a history of taxonomic confusion (for examples see Hopper & Brown, 2001, 2007), it is difficult to assess whether the rare taxa on the Swan Coastal Plain were intrinsically rare. Given the specialized habitat requirements of most of the rare Swan Coastal Plain species (Hopper & Brown, 2001, 2007; Brown et al., 2008), they were probably always uncommon or localized, but this position has been accentuated by land clearance and habitat alteration. An additional influence was the consumption of some common orchids by Noongar Aboriginal people (Drummond, 1842), present in the region for at least 35,000 years (Allen, 1998). An important area of future research will be to examine the effects of anthropogenic change on species with differing habitat requirements, pollination strategies and mycorrhizal ecology.

cannot be used as a surrogate for assessing the importance of a region for orchid conservation and vice versa. Restricted edaphic environments were the spatially most strongly associated with rarity in orchids of the SWAFR, in particular swamps, inland salt lake margins and granite outcrops. The natural rarity and fragmentation of these habitats mean that the orchids may have evolved a genetic system that can cope with the associated inbreeding, as seen in some other SWAFR taxa (James, 1965; Samson et al., 1988; Byrne & Hopper, 2008). However, the maintenance of suitable habitats to facilitate dispersal events in the circumstance that existing locations become unsuitable needs to be considered. While granite outcrops, although sometimes degraded, are reasonably well protected in conservation reserves, the other habitats remain under threat. Salt-lake margins are a vulnerable habitat due to rising saline water tables, resulting from the removal of up to 95% of the original vegetation in the Western Australian wheatbelt (Anon., 2007). Swamplands are generally well protected in the state forests in southern Western Australia. However, the orchid-rich swamps of the Swan Coastal Plain have mostly been cleared, and ephemeral swamps in the Leeuwin–Naturaliste district are threatened by a current proposal to tap the Yarragadee aquifer to supplement Perth’s declining water supply (Horwitz et al., 2008). The long-term effects of pervasive changes, such as a pronounced reduction in rainfall (Li et al., 2005), changing fire regimes (Brown et al., 1998), and disturbance from introduced pests (Brown et al., 1998), remain to be seen. Future conservation efforts should take into account the propensity towards rarity in sexually deceptive species. Due to the specificity of the plant–pollinator relationship, particular attention should be paid to the biology and requirements of the pollinator. In particular, if there are ample sites for recruitment, an increase in abundance of the pollinator may lead to an increase in orchid recruitment. With the exception of the ant-pollinated Leporella (Peakall, 1989), all sexually deceptive taxa in the SWAFR utilize parasitic wasps (see references in Appendix S4; Ridsdill Smith, 1970). Parasitoids are believed to be particularly sensitive to environmental change (Tscharntke & Brandl, 2004), making the biology of the wasps and the orchids they pollinate of particular concern. In the longer term, changes in the abundance of a pollinator may precede those of the orchid.

Conservation implications In terms of conservation of orchid communities, regions important for the flora in general will often not satisfy the needs of orchid conservation in terms of preserving high species richness and local endemics. A similar result has been obtained on other continents at a regional scale when testing for a correlation in species richness between different groups of flora and fauna (Prendergast et al., 1993; Howard et al., 1998), and conforms to the observation of Gaston (2000) that exceptions to patterns of biodiversity become more prevalent at lower taxonomic ranks. The present study demonstrates that this is the case for Orchidaceae. The diversity of the total flora 498

ACKNOWLEDGEMENTS Funding was provided by the School of Plant Biology at the University of Western Australia and grants from the Australian Orchid Foundation and Holsworth Wildlife Research Endowment to R.D.P. R.D.P. was supported by an Australian Postgraduate Award. We are very grateful to the dedicated orchidologists, particularly members of the Western Australian Native Orchid Study and Conservation Group, who have provided numerous collections to the Western Australian Herbarium. Thanks to Paul Gioia for allowing the use of the figure of species richness in the SWAFR. Thanks to Ann Journal of Biogeography 38, 487–501 ª 2010 Blackwell Publishing Ltd

Orchid biogeography and rarity Smithson and two anonymous reviewers for comments that improved the quality of the manuscript. REFERENCES Adams, P.B. & Lawson, S.D. (1993) Pollination of Australian orchids: a critical assessment of the literature 1882–1992. Australian Journal of Botany, 41, 553–575. Allen, J. (1998) When did humans first colonise Australia? Archaeology of aboriginal Australia (ed. by T. Murray), pp. 50–60. Allen & Unwin, Sydney. Anon. (2007) State of the environment Western Australia – land – theme 3. Available at: http://www.soe.wa.gov.au/report/ land.html (accessed 12 April 2009). Beard, J.S. (1981) Vegetation survey of Western Australia – Swan. University of Western Australia Press, Perth. Beard, J.S. (1999) Evolution of the river systems of the southwest drainage division, Western Australia. Journal of the Royal Society of Western Australia, 82, 147–164. Belbin, B. (1994) PATN: pattern analysis package. CSIRO Division of Wildlife and Ecology, Canberra. Brown, A.P., Thomson-Dans, C. & Marchant, N. (1998) Western Australia’s threatened flora. Department of Conservation and Land Management, Perth. Brown, A.P., Dundas, P., Dixon, K.W. & Hopper, S.D. (2008) Orchids of Western Australia. University of Western Australian Press, Perth. Bussell, J.D., Hood, P., Alacs, E.A., Dixon, K.W., Hobbs, R.J. & Krauss, S.L. (2006) Rapid genetic delineation of local provenance seed-collection zones for effective rehabilitation of an urban bushland remnant. Austral Ecology, 31, 164– 175. Byrne, M. & Hopper, S.D. (2008) Granite outcrops as ancient islands in old landscapes: evidence from the phylogeography and population genetics of Eucalyptus caesia (Myrtaceae) in Western Australia. Biological Journal of the Linnean Society, 93, 177–188. Coleman, E. (1930) Pollination of some Western Australian orchids. Victorian Naturalist, 46, 62–66. Cozzolino, S. & Widmer, A. (2005) Orchid diversity: an evolutionary consequence of deception? Trends in Ecology and Evolution, 20, 487–494. Cribb, P.J., Kell, S.P., Dixon, K.W. & Barrett, R.L. (2003) Orchid conservation: a global perspective. Orchid conservation (ed. by K.W. Dixon, S.P. Kell, R.L. Barrett and P.J. Cribb), pp. 1–24. Natural History Publications, Sabah. Drummond, J. (1842) On the botany of Western Australia. The Inquirer, 107, 3. Gaston, K.J. (1994) Rarity. Chapman & Hall, London. Gaston, K.J. (2000) Global patterns in biodiversity. Nature, 405, 220–227. Gravendeel, B., Smithson, A., Slik, F.J.W. & Schuiteman, A. (2004) Epiphytism and pollinator specialisation: drivers for orchid diversity? Philosophical Transactions of the Royal Society B: Biological Sciences, 359, 1523–1535. Journal of Biogeography 38, 487–501 ª 2010 Blackwell Publishing Ltd

Hall, L.S., Krausman, P.R. & Morrison, M.L. (1997) The habitat concept and a plea for standard terminology. Wildlife Society Bulletin, 25, 173–182. Hocking, R.M. (1990) Carnarvon Basin. Geology and mineral resources of Western Australia: Western Australian geological survey, Memoir 3 (ed. by J.S. Meyers & R.M. Hocking), pp. 457–460. State Printing Division, Perth. Hoffman, N. & Brown, A.P. (1998) Orchids of south-western Australia, 2nd edn. University of Western Australia Press, Perth. Holmgren, M. & Poorter, L. (2007) Does a ruderal strategy dominate the endemic flora of the West African forests? Journal of Biogeography, 34, 1100–1111. Hopper, S.D. (1979) Biogeographical aspects of speciation in the southwest Australian flora. Annual Review of Ecology and Systematics, 10, 399–422. Hopper, S.D. (1983) Atlas of the Western Australian flora pilot project – orchids. Western Australian Wildlife Research Centre, Perth. Hopper, S.D. & Brown, A.P. (2001) Contributions to Western Australian orchidology: 2. New taxa and circumscriptions in Caladenia (Spider, Fairy and Dragon Orchids of Western Australia). Nuytsia: Bulletin of the Western Australian Herbarium, 14, 27–307. Hopper, S.D. & Brown, A.P. (2004) Robert Brown’s Caladenia revisited, including a revision of its sister genera Cyanicula, Ericksonella and Pheladenia (Caladeniinae: Orchidaceae). Australian Systematic Botany, 17, 171–240. Hopper, S.D. & Brown, A.P. (2006) Australia’s wasp-pollinated flying duck orchids revised (Paracaleana: Orchidaceae). Australian Systematic Botany, 19, 211–244. Hopper, S.D. & Brown, A.P. (2007) A revision of Australia’s hammer orchids (Drakaea: Orchidaceae), with some field data on species-specific sexually deceived wasp pollinators. Australian Systematic Botany, 20, 252–285. Hopper, S.D. & Gioia, P. (2004) The Southwest Australian Floristic Region: evolution and conservation of a global diversity hotspot. Annual Review of Ecology, Evolution, and Systematics, 35, 623–650. Hopper, S.D., Keighery, G.J. & Wardell-Johnson, G. (1992) Flora of the Karri forest and other communities in the Warren Botanical Subdistrict of Western Australia. CALM Occasional Paper No. 2/92, pp. 1–32. Conservation and Land Management, Perth. Horwitz, P., Bradshaw, D., Hopper, S.D., Davies, P., Froend, R. & Bradshaw, F. (2008) Hydrological change escalates risk of ecosystem stress in Australia’s threatened biodiversity hotspot. Journal of the Royal Society of Western Australia, 91, 1–11. Howard, P.C., Viskanic, P., Davenport, T.R.B., Kigneyi, F.W., Baltzer, M., Dickinson, C.J., Lwanga, J.S., Mathews, R.A. & Balmford, A. (1998) Complementarity and the use of indicator groups for reserve selection in Uganda. Nature, 394, 472–475. Jacquemyn, H., Brys, R., Hermy, M. & Willems, J.H. (2004) Does nectar reward affect rarity and extinction probabilities 499

R. D. Phillips et al. of orchid species? An assessment using historical records from Belgium and the Netherlands. Biological Conservation, 121, 257–263. James, S.H. (1965) Complex hybridity in Isotoma petraea I. The occurrence of interchange heterozygosity, autogamy and a balanced lethal system. Heredity, 20, 341–353. Jersakova, J. & Johnson, S.D. (2006) Lack of floral nectar reduces self-pollination in a fly-pollinated orchid. Oecologia, 147, 60–68. Jones, D.L. (2006) A complete guide to the native orchids of Australia. Reed New Holland, Sydney. Kennington, W.J. & James, S.H. (1998) Allozyme and morphometric variation in two closely related mallee species from Western Australia, Eucalyptus argutifolia and E. obtusiflora (Myrtaceae). Australian Journal of Botany, 46, 173–186. Kores, P.J., Molvray, M., Weston, P.H., Hopper, S.D., Brown, A.P., Cameron, K.M. & Chase, M.W. (2001) A phylogenetic analysis of Diuridae (Orchidaceae) based on plastid DNA sequence data. American Journal of Botany, 88, 1903– 1914. Kull, T. & Hutchings, M.J. (2006) A comparative analysis of decline in the distribution ranges of orchid species in Estonia and the United Kingdom. Biological Conservation, 129, 31–39. Li, Y., Cai, W.J. & Campbell, E.P. (2005) Statistical modelling of extreme rainfall in south-west Western Australia. Journal of Climate, 18, 852–863. Lyons, M.N., Keighery, G.J., Gibson, N. & Wardell-Johnson, G. (2000) The vascular flora of the Warren bioregion, southwest Western Australia: composition, reservation status and endemism. CALMScience, 3, 181–250. Myers, J.S. (1990) Pinjarra orogen. Geology and mineral resources of Western Australia: Western Australian geological survey, Memoir 3 (ed. by J.S. Meyers & R.M. Hocking), pp. 265–274. State Printing Division, Perth. Neiland, M.R.M. & Wilcock, C.C. (1998) Fruit set, nectar reward, and rarity in the Orchidaceae. American Journal of Botany, 85, 1657–1671. Otero, J.T. & Flanagan, N.S. (2006) Orchid diversity – beyond deception. Trends in Ecology and Evolution, 21, 64–65. Peakall, R. (1989) The unique pollination of Leporella fimbriata (Orchidaceae): pollination by pseudocopulating male ants (Myrmecia urens, Formicidae). Plant Systematics and Evolution, 167, 137–148. Peat, H.J., Clarke, A. & Convey, P. (2007) Diversity and biogeography of the Antarctic flora. Journal of Biogeography, 34, 132–146. Phillips, R.D. (2010) Landscape, pollinator and mycorrhizal specificity and their contribution to rarity in Drakaea (Hammer Orchids). PhD thesis, The University of Western Australia, Perth. Phillips, R.D., Faast, R., Bower, C.C., Brown, G.R. & Peakall, R. (2009a) Implications of pollination by food and sexual deception for pollinator specificity, fruit set, population

500

genetics and conservation of Caladenia (Orchidaceae). Australian Journal of Botany, 57, 287–306. Phillips, R.D., Backhouse, G., Brown, A.P. & Hopper, S.D. (2009b) Biogeography of Caladenia (Orchidaceae), with special reference to the Southwest Australian Floristic Region. Australian Journal of Botany, 57, 259–275. Prendergast, J.R., Quinn, R.M., Lawton, J.H., Eversham, B.C. & Gibbons, D.W. (1993) Rare species, the coincidence of diversity hotspots and conservation strategies. Nature, 365, 335–337. Rabinowitz, D. (1981) Seven forms of rarity. The biological aspects of rare conservation (ed. by H. Synge), pp. 205–217. John Wiley & Sons, Melbourne. Ramsay, R.R., Dixon, K.W. & Sivasithamparam, K. (1986) Patterns of infection and endophytes associated with Western Australian orchids. Lindleyana: the Scientific Journal of the American Orchid Society, 1, 203–214. Rasmussen, H.N. (1995) Terrestrial orchids from seed to mycotrophic plant. Cambridge University Press, Melbourne. Ridsdill Smith, T.J. (1970) The biology of Hemithynnus hyalinatus (Hymenoptera: Tiphiidae), a parasite on scarabaeid larvae. Journal of the Australian Entomological Society, 9, 183–195. Royal Botanic Gardens, Kew (2009) World checklist of selected plant families. Available at: http://apps.kew.org/wcsp/ prepareChecklist.do?checklist=selected_families%40%40259 160920/00929686 (accessed on 12 April 2009). Samson, J.F., Hopper, S.D. & James, S.H. (1988) Genetic diversity and the conservation of Eucalyptus crucis Maiden. Australian Journal of Botany, 36, 447–460. Shepherd, D.P., Hopkins, A.J.M. & Beeston, G. (2002) Native vegetation in Western Australia: extent, type and status. Western Australian Department of Agriculture, Perth. Smithson, A. & Gigord, L.D.B. (2001) Are there fitness advantages to being a rewardless orchid? Reward supplementation experiments with Barlia robertiana. Proceedings of the Royal Society B: Biological Sciences, 268, 1435–1441. Stebbins, G.L. & Major, J. (1965) Endemism and speciation in the Californian flora. Ecological Monographs, 35, 1–35. Stoutamire, W.P. (1983) Wasp-pollinated species of Caladenia (Orchidaceae) in South-western Australia. Australian Journal of Botany, 31, 383–394. Swarts, N.D. & Dixon, K.W. (2009) Terrestrial orchid conservation in the age of extinction. Annals of Botany, 104, 543–556. Tremblay, R.L., Ackerman, J.D., Zimmerman, J.K. & Calvo, R.N. (2005) Variation in sexual reproduction in orchids and its evolutionary consequences: a spasmodic journey to diversification. Biological Journal of the Linnean Society, 84, 1–54. Tscharntke, T. & Brandl, R. (2004) Plant–insect interactions in fragmented landscapes. Annual Review in Entomology, 49, 405–430.

Journal of Biogeography 38, 487–501 ª 2010 Blackwell Publishing Ltd

Orchid biogeography and rarity Western Australian Herbarium (2009) Western Australian Herbarium. Department of Environment and Conservation. Available at: http://florabase.calm.wa.gov.au (accessed on 12 April 2009). Wheeler, M.A. & Byrne, M. (2006) Congruence between phylogeographic patterns in cpDNA variation in Eucalyptus marginata (Myrtaceae) and geomorphology of the Darling Plateau, south-west of Western Australia. Australian Journal of Botany, 54, 17–26.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

BIOSKETCH SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article: Appendix S1 Taxonomic reviews of south-west Australian Orchidaceae from 1980 onwards. Appendix S2 Taxa in which herbarium records were supplemented with known locations for analyses of taxon richness and biogeography. Appendix S3 Taxa included in the analyses, other than currently described taxa. Appendix S4 Incidence of rarity, mean distribution and mean abundance of orchid genera in the Southwest Australian Floristic Region with reference to pollination strategy and site of mycorrhizal infection. Appendix S5 An analysis of the taxon richness and endemism in the biogeographic districts of the Orchidaceae of the Southwest Australian Floristic Region.

Journal of Biogeography 38, 487–501 ª 2010 Blackwell Publishing Ltd

Ryan Phillips is post-doctoral research fellow at The Australian National University. He undertook his PhD at Kings Park and Botanic Garden and The University of Western Australia, investigating the role of mycorrhiza and pollinators in controlling rarity and speciation in Drakaea. His current interests include the causes of orchid diversification, the evolutionary interactions of orchids and their pollinators, and the ecology and evolution of pollination systems in southwestern Australia. Author contributions: R.D.P., A.P.B., K.W.D. and S.D.H. conceived the ideas, R.D.P. collected the data with supplementary information from A.P.B. for poorly known taxa, R.D.P. analysed the data, and R.D.P. led the writing with revision by A.P.B., K.W.D. and S.D.H.

Editor: Pauline Ladiges

501