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