Global Ecology and Biogeography, (Global Ecol. Biogeogr.) (2009) 18, 596–606 Blackwell Publishing Ltd

RESEARCH PAPER

Dispersal traits linked to range size through range location, not dispersal ability, in Western Australian angiosperms Aaron D. Gove1,2*, Matthew C. Fitzpatrick3,4, Jonathan D. Majer1 and Robert R. Dunn1,2

1

Centre for Ecosystem Diversity and Dynamics in the Department of Environmental and Aquatic Sciences, Curtin University of Technology, Perth, Western Australia, Australia, 2 Department of Biology, North Carolina State University, Raleigh, NC, USA, 3Harvard University, Harvard Forest, 324 North Main Street, Petersham, MA 01366, USA, 4 Department of Biological Sciences, University of Rhode Island, 100 Flagg Road, Kingston, RI 02881-0816, USA

ABSTRACT

Aim We examine the relative importance of seed dispersal mode in determining the range size and range placement in 524 species from six focal plant families (Agavaceae, Euphorbiaceae, Malvacaeae, Sapindaceae, Proteaceae and Fabaceae (Acacia)). Location Western Australia. Methods Taxa were categorized by dispersal mode and life-form and their distributions modelled using maxent. Geographical range size was compared amongst dispersal mode, life-form and biome using phylogenetically independent contrasts. Geographical range placement was considered in a similar manner. Results Range size did not vary with dispersal mode (ant versus wind and vertebrate dispersal) or life-form, and instead varied primarily as a function of the biogeographical region in which a species was found. Range placement, however, did vary among dispersal modes, with the consequence that diversity of wind- and ant-dispersed plants increased with latitude while the diversity of vertebratedispersed plants was more evenly distributed.

*Corresponding author: Department of Environmental and Aquatic Science, Curtin University of Technology, PO Box U1987 Perth, Western Australia 6845, Australia. E-mail: [email protected]

Main conclusions For the taxa studied, range sizes were a function of the biogeographical region in which species were found. Although differences in range size may exist among species differing in dispersal modes, they are likely to be far smaller than differences among species from different biogeographical regions. The trait most likely to affect species geographical range size, and hence rarity and risks associated with other threats, may simply be the geographical region in which that species has evolved. Keywords Angiosperm diversity, biogeography, dispersal, latitude, myrmecochory, Western Australia.

At the regional scale, patterns of species diversity are a consequence of both species range sizes and their positions in space (Colwell & Lees, 2000; Arita & Rodríguez, 2002; Colwell et al., 2004). Both range size and range position can in turn be influenced by species traits such as dispersal ability. However, the role of dispersal in determining range size is poorly understood, while its influence on range positions is considered better understood but is rarely considered empirically.

Species vary in geographical distribution from globally distributed species, such as the osprey, to species with geographical distributions that were, even prior to anthropogenic disturbance, no more than a few dozen hectares (e.g. Dunn & Romdal, 2005). Theory suggests that one of the traits likely to influence range size is dispersal ability (Hanski, 1999; Dynesius & Jansson, 2000; Hubbell, 2001). Poorly dispersing species are predicted to have smaller ranges because they are less able to establish or maintain gene flow to distant sites, or more likely to become locally adapted and more likely to speciate (Kunin & Gaston, 1997;

596

DOI: 10.1111/j.1466-8238.2009.00470.x © 2009 Blackwell Publishing Ltd www.blackwellpublishing.com/geb

INTRODUCTION

Ecological constraints on range size Latimer et al., 2005; Urban et al., 2008). However, empirical results for plants, as an example, have been mixed (Edwards & Westoby, 1996; Kelly & Woodward, 1996; Murray, 2002a,b; Pocock et al., 2006) and proposed mechanisms linking dispersal ability with range size are not generally well supported (Lester et al., 2007). Just why dispersal mode and range size are only sometimes correlated remains unresolved but may depend upon taxa, scale and particular context (Lester et al., 2007). Here, we gain insight into these factors by examining the interplay between range size and range placement and focusing upon a terrestrial environment with pronounced biome boundaries. The effects of dispersal mode on range size may be contingent on several other species-level traits. One trait particularly likely to influence seed dispersal distance is life-form (Murray et al., 2002b). Larger plants may achieve longer dispersal distances than do smaller plants, both because of the height from which seeds drop, and thus their access to more powerful dispersers, and because of their ability to produce more seeds per plant (Westoby et al., 1990; Ruokolainen & Vormisto, 2000; Ruokolainen et al., 2002). Lineages with dispersal modes with larger average dispersal distances should have larger ranges (as above), and for a given dispersal mode larger plant life-forms (e.g. trees versus herbs) should achieve larger ranges. An alternative reason why findings among studies differ with regard to the relationships between dispersal mode, or other traits, and range size is that the trait may vary along environmental gradients (e.g. Bond et al., 1991, Márquez et al., 2004). This possibility has rarely been considered, but given that dispersal traits can vary dramatically in space (Willson et al., 1989; Moles et al., 2007), it is a realistic one. Under this scenario, range sizes of taxa would be governed primarily by aspects of the biogeographical region in which they are found, rather than dispersal-related traits. Species with a particular trait may have larger ranges, for example, simply as a consequence of being more common in biomes or biogeographical regions where habitat is at its largest extent. If range sizes are primarily a result of attributes of biogeographical regions rather than species traits (or some interaction of the two), relationships between such traits and range size are expected to vary among regions and to be difficult to generalize. Amongst terrestrial ecosystems, those that include seed dispersal by ants (myrmecochory), wind and vertebrates are excellent candidates for the study of dispersal mode, dispersal distance and their consequences for geographical range size. Ant-dispersed species possess some of the shortest dispersal distances documented (Gómez & Espadaler, 1998) and thus should be the most likely to show the effects of dispersal limitation on range size. If dispersal ability is associated with range size, plant species dispersed by ants should have smaller geographical ranges than those dispersed by birds or wind. To date, empirical results provide only modest support for a relationship between dispersal mode and range size in ecosystems with seed dispersal by ants and birds or wind. In Australia, Oakwood et al. (1993) found that larger range sizes in vertebratedispersed plant species depended upon the region considered. Similarly McDonald et al. (1995), found that ant-dispersed

species were more likely to be endemic to the Cape Province of South Africa than wind-dispersed species, though they did not explicitly consider range size. Here, we examine the relationship between dispersal mode, life-form and range placement in Western Australia, a region in which myrmecochory – a globally rare dispersal mode – is very common (Berg, 1975) and has evolved often (Dunn et al., 2007, 2008; Lengyel et al., 2009). First, we test whether range size varies with dispersal mode and life-form, and more specifically whether ant-dispersed species have smaller geographical ranges than species with other dispersal vectors. We also consider whether once we account for biogeographical gradients in dispersal mode by focusing on Western Australia’s primary biogeographical regions, the relationship between dispersal mode and range size persists. Secondly, we examine whether dispersal mode varies with latitude and longitude and consider the consequences of such variations in range placement. Finally, we use our distribution models to identify the important environmental variables associated with the distribution of species within each class of dispersal mode. METHODS Dispersal mode We chose six plant families for consideration (Euphorbiaceae, Agavaceae, Sapindaceae, Malvaceae, Proteaceae and Fabaceae (Acacia)). We focussed on these families because morphological adaptations of diaspores for dispersal are known to vary within families and dispersal of seeds by ants is well represented, while many of the phylogenetic relationships are well defined. We use the term seed throughout for convenience, but technically the relevant dispersal unit (diaspore) varies from seeds to multiseeded fruits. In assigning seeds to dispersal groups, we focused on morphological adaptations for dispersal. We assigned dispersal morphology at the level of plant genera (79 in total). Exceptions were Grevillea (Proteaceae) and Acacia, which were split at the species level due to a large degree of within-genus variation of dispersal mode. We assigned dispersal mode based on a combination of observations of diaspores and the consultation of systematic revisions, regional floras, interactive keys and publications explicitly focussing on diaspore dispersal. When diaspores possessed white, or more generally pale, fleshy appendages smaller in size than the ‘body’ of the diaspore, these appendages were considered to be elaiosomes and the seeds were considered to be ant dispersed. When diaspores possessed larger, more brightly coloured appendages, or were covered in flesh, they were coded as vertebrate dispersed. When diaspores possessed wings or wing-like appendages, they were coded as wind dispersed. All other species were coded as passively dispersed. In order to examine the role of life-form as a correlate of biogeographical attributes, we designated species to life-forms (trees, shrubs or herbs) based upon descriptions provided by FloraBase (Department of Environment and Conservation, 2005). Taxa that fell within two categories were assigned to the larger of the two life-forms.

Global Ecology and Biogeography, 18, 596–606, © 2009 Blackwell Publishing Ltd

597

A. D. Gove et al. Geographical range data Geographical range data were compiled from Western Australian Herbarium records (PERTH; accessed May 2005). Western Australia and its plants are ideal for these tests because a variety of dispersal modes are represented within the region and plant diversity is exceptionally high (e.g. Hopper, 1979, 1992; Hopper & Gioia, 2004). We used the distribution of the six plant families, which together include 524 species in Western Australia. The state of Western Australia covers an area of approximately 2,770,000 km2 and 22° of the 29° (75%) of latitude covered by mainland Australia. We estimated the size of species ranges by modelling their distributions using maxent 2.3.3 (Phillips et al., 2006). maxent is a recent implementation of a statistical approach called maximum entropy that characterizes probability distributions from incomplete information (Phillips et al., 2006). In the context of modelling distributions of species using maximum entropy, the assumptions are: (1) that occurrence data represent an incomplete sample of an empirical probability distribution; (2) that this unknown distribution can be most appropriately estimated as the distribution with maximum entropy (i.e. the probability distribution that is most uniform) subject to constraints imposed by environmental variables; and (3) that this distribution of maximum entropy approximates the potential geographical distribution of the species (see Phillips et al., 2006, for more details). maxent has several characteristics that make it particularly suitable for our study, including a deterministic algorithm, the ability to use presence-only distribution data and the option to automatically batch process using command line scripts. Further, maxent has been found to be a promising and robust approach for modelling species distributions (Elith et al., 2006; Hernandez et al., 2006; Pearson et al., 2007; Sérgio et al., 2007). To avoid potential problems relating to small sample sizes (Stockwell & Peterson, 2002), we developed models only for species that had at least 20 unique distribution records, thereby leaving 573 – 49 = 524 species for analysis. We used Phillips et al.’s (2006) suggested default values for the convergence threshold (10–5) and maximum number of iterations of 1000. Setting of regularization values, which address problems of over-fitting and selection of ‘features’ (environmental variables and/or functions derived from combinations of such variables), were performed automatically by the program per the default rules dependent on the number of occurrence records. Environmental variables (from which ‘features’ are based) included those considered important to plant distribution in general (Woodward, 1987) and Western Australia in particular (Hopper & Maslin, 1978; Hnatiuk & Maslin, 1988; Beard, 1990; Groom & Lamont, 1996; Keighery, 1996; Lamont & Connell, 1996; Cowling & Lamont, 1998; B. Lamont, pers. comm.). These included major soil types (Digital Atlas of Australian Soils, Department of Agriculture, Fisheries and Forestry, 2005), an index of humidity (the ratio of mean actual annual evapotranspiration to mean potential annual evapotranspiration, AET/PET), mean annual sum of precipitation, mean annual temperature, maximum temperature

598

of the warmest month, and winter (JJA) and summer (DJF) averages of precipitation, temperature and potential evapotranspiration (Bureau of Meteorology, 2005). All layers were at a resolution of 0.025° (approximately 2.5 km × 2.5 km in Australia) and were projected to Alber’s equal-area conic to preserve the area characteristics of resulting range predictions. We converted the relative suitability values (0–100) from maxent to presence/absence (1/0) using the minimum threshold value that predicted all known occurrences as present. This approach is a conservative estimate that determines the smallest predicted range possible, but which contains all known occurrences. In ecological terms, the predicted range thus contains all environmental conditions that are at least as suitable as those where the species has been observed. Using arcmap 9.1 (ESRI, 2006), we determined the area of the predicted ranges and their latitude–longitude centre point. For all diversity analyses, we derived maps of species richness by calculating the sum of overlapping presence/absence ranges in each cell. The area under the curve (AUC) of the receiver operator characteristic (ROC) plot of sensitivity versus (1 – specificity) and pseudo-absences rather than observed absences (Phillips et al., 2006) was used to test the agreement between observed species presence and the modelled potential distribution. Models for all 524 species exhibited acceptable performance (minimum AUC = 0.86). We examined the relative importance of individual environmental variables as predictors of the distributions of each species in each dispersal mode using maxent’s built-in jackknife procedure. To quantify the importance of a variable for each dispersal mode, we calculated the mean training gains when the variable of interest was used in isolation in the maxent runs for all species within that mode. Analyses, range size and midpoint We compared the geographical range sizes of species among dispersal modes and life-forms and variation in range midpoint location among dispersal modes. We also tested whether occupied biome (botanical provinces based on Beard, 1980) had an association with range size. Species were assigned to one of three biomes based on the location of the species’ range midpoint. We focus here on three biomes – the South West, Eremaean and Northern provinces. We chose these provinces because, in contrast to the more finely divided, for example, World Wildlife Fund (WWF) ecoregions (Olson et al., 2001) these provinces have very separate histories (Beard, 1980) and represent areas of strong transition in species composition (M. C. Fitzpatrick, unpublished data), which is not necessarily the case with WWF ecoregions where turnover amongst region may be as much as that within regions (McDonald et al., 2005). Areas and midpoints of ranges were analysed using phylogenetic comparative analyses (Purvis & Rambaut, 1995). We tested whether dispersal mode, life-form and biome were related to range size and whether the location of each species (latitude and longitude) was related to dispersal mode in terms of the independent occasions on which evolutionary transitions in the given trait had occurred. Our phylogenetic hypotheses

Global Ecology and Biogeography, 18, 596–606, © 2009 Blackwell Publishing Ltd

Ecological constraints on range size Table 1 Results of phylogenetically independent contrasts on sister clades for the relationship between plant life-history traits and biome on geographical range size and position. Variable

Contrasts Positive contrasts F

Geographical range size Dispersal mode 13 Life form 28 Biome 36

7 15 32

P

1.04 0.327 0.01 0.929 61.37 < 0.001

Latitude Dispersal mode

11

9

15.45

0.003

Longitude Dispersal mode

11

8

10.69

0.008

phylogenetic comparisons for dispersal mode (ant versus vertebrate and wind), 28 comparisons for life-form and 36 contrasts in the two bioregion types. Area was log-transformed before analysis. Because dispersal modes were non-randomly distributed among these biomes, comparisons of range size among dispersal modes at the scale of Western Australia may confound the effects of dispersal mode per se and the effects of biome size. As such, we repeated the comparisons of range size among dispersal modes within biomes. RESULTS Figure 1 Phylogenetically independent contrasts between range size and (a) dispersal mode, (b) life-form and (c) biogeographical region.

were based on the Angiosperm Phylogeny Group (2003) for the families, Whitlock et al. (2001) for the Malvaceae (Sterculiaceae), Hoot and Douglas (1998) for the Proteaceae and Wurdack et al. (2004, 2005) and Kathriarachchi et al. (2006) for the Euphorbiaceae. Agavaceae was separated into several clades recognized by the Angiosperm Phylogeny Group (2003). The phylogenies for Sapindaceae, Acacia and Grevillea were resolved as polytomies. As the comparison of independent contrasts relies on a dichotomy of independent variables (Purvis & Rambaut, 1995), or a clear gradient in more than two categorical variables, we compare ant-dispersed with wind- and vertebrate-dispersed species (assumed to be ‘long dispersal’) combined. Likewise, we chose to examine the effect of biomes in terms of their size: the smaller biomes of the North and Southwest provinces, are of similar area (3.3 × 107 ha vs. 3.2 × 107 ha, respectively), and we combined data from the two to form a ‘small biomes’ treatment to compare with species occupying the larger Eremaean Province (‘large biome’). Existing phylogenetic data allowed for 13

Does range size vary with dispersal mode, life-form or biome? When we considered the region as a whole (disregarding biomes), range sizes did not vary between ant-dispersed and other dispersal modes (F1,12 = 1.04, P = 0.327; Fig. 1a, Table 1). No individual test comparing pairs of dispersal modes (ant-, wind- or vertebrate-) was significant (P = 0.37–0.88); but power was particularly low at this level (three to nine contrasts). Range size was not associated with difference in life-form (F1,27 = 0.01, P = 0.929; Fig. 1b, Table 1). In contrast, the size of the biogeographical region considered was strongly associated with range size among lineages. Geographical ranges were almost seven times larger in the Eremaean biome as compared to the smaller South West and Northern biomes (F1,35 = 61.37, P < 0.001, Fig. 1c). Within biomes, there was a marginal relationship between dispersal mode and range size in the smaller biomes, but not in the larger biome (Table 2). We found no relationship between life-form and range size. Thus, range size is likely to be primarily associated with the biome in which it is situated. With greater statistical power, we may detect a significant statistical signal of dispersal mode on range size within each biome, but given the larger differences between biogeographical regions in range size, such a difference is likely to remain small.

Global Ecology and Biogeography, 18, 596–606, © 2009 Blackwell Publishing Ltd

599

A. D. Gove et al. Table 2 Results of phylogenetically independent contrasts on sister clades for the relationship between plant species traits and geographical range size within biomes. Contrasts

Positive contrasts

F

P

Within large biome Dispersal mode Life-form

7 14

3 4

0.35 1.84

0.577 0.198

Within small biomes Dispersal mode Life-form

11 22

9 12

4.10 0.26

0.070 0.615

Are there biogeographical gradients in dispersal mode? In phylogenetic comparisons, locations of the mid-points of species geographical ranges were significantly associated with dispersal mode (Table 1). Ant-dispersed species were positioned further to the south (F1,10 = 15.45, P = 0.003) and to the west (F1,10 = 10.69, P = 0.008) than species dispersed by other modes (Fig. 2). In all plant clades, ant-dispersed species were concentrated in the South West, and vertebrate-dispersed species were richest in three separate regions (Northern, Central West and South West). Like ant-dispersed plant species, winddispersed species were also concentrated in the South West (Fig. 2c). As a consequence of these patterns in range placement, ant- and wind-dispersed species showed a strong latitudinal gradient in diversity. Notably, the gradient was in the opposite direction of the ‘canonical gradient’ in diversity, such that ant- and wind-dispersed plants were both most diverse furthest from the equator. Environmental variables and species distributions For both wind- and ant-dispersed species, the environmental variables best associated with species distributions were winter precipitation, followed by summer temperature and actual/ potential evapotranspiration (Fig. 3a,b). Wind- and ant-dispersed species had very similar profiles of importance of environmental variables, which coincides closely with their similar distribution patterns in south-western Australia, a region of mediterranean climate (Fig. 2c,d). The two most important variables for vertebrate-dispersed species were winter precipitation and winter potential evapotranspiration followed by mean annual temperature and summer precipitation (Fig. 3c). DISCUSSION Does range size vary with dispersal mode, life-form and biome? Theory suggests that species with poor dispersal ability may move more slowly across the landscape, maintain less gene flow

600

and consequently have smaller geographical ranges (Carter & Prince, 1988; Rosenzweig, 1995; Cody & Overton, 1996). However, like several other authors (e.g. Kelly & Woodward, 1996; Thompson et al, 1999; Murray et al., 2002a; Lester et al., 2007), we found little empirical support for the relationship between range size and dispersal mode over such a large scale. Despite having average dispersal distances of the order of metres, ant-dispersed plant species did not consistently have smaller ranges than plant species dispersed by, for example, birds, for which dispersal distances could easily be hundreds of metres (Westcott et al., 2005; Weir & Corlett, 2006). Similarly, plant size, which may influence range size through its effects on initial seed dispersal distances during falling (Ruokolainen & Vormisto, 2000), was not correlated with range size within Western Australia. Why doesn’t dispersal mode correlate more strongly with range size? One reason is that despite the short average distances associated with ant-dispersal of seeds (Gómez & Espadaler, 1998), rare long-distance dispersal events may occur often enough to maintain large-scale gene flow and hence large geographical ranges (Clark, 1998; Higgins & Richardson, 1999; Vellend et al., 2003; Calviño-Cancela et al., 2006; He et al., 2009). While the short average dispersal distance of ant-dispersed species undoubtedly affects local gene flow patterns (e.g. Kalisz et al., 1999), at larger scales, rare long-distance dispersal events may actually be independent of dispersal morphology (Higgins et al., 2003; He et al., 2009). Emus (Dromaius novaehollandiae) are known to occasionally disperse ant-dispersed plant species in Western Australia (Calviño-Cancela et al., 2006), as are white-tailed deer in North America (Vellend et al., 2003). Any of a variety of non-standard means of dispersal may be sufficient to move propagules across species potential ranges, particularly in regions such as Western Australia where climate has been relatively stable for millions of years due to weak Milankovitch cycles (Dodson, 1998). If dispersal mode does not mediate range size then what does? The most obvious alternative is that range sizes of species are primarily a function of the size of their occupied biogeographical regions. Western Australia can be divided into three biogeographical regions that differ both in their current climate and in the history of their biotas. These biogeographical regions have relatively hard boundaries, such that few species have ranges that cross them (Beard, 1980; M. C. Fitzpatrick et al., unpublished data). We found that the two biogeographical regions that are much smaller – the South West and Northern biogeographical regions – had species ranges that were on average, almost seven times smaller than those in the Eremaean region (Fig. 1c). The most parsimonious explanation for the larger range sizes in larger biomes is that, in those biomes, species have more geographical space in which environmental requirements are met (Oakwood et al., 1993; Smith et al., 1994; Hughes et al., 1996). For instance, in our study, despite their ostensibly greater dispersal ability, wind-dispersed species are confined to precisely the same environmental (and hence geographical) space as ant-dispersed species (Fig. 3). The data suggest that it is the association of dispersal mode and environmental conditions,

Global Ecology and Biogeography, 18, 596–606, © 2009 Blackwell Publishing Ltd

Ecological constraints on range size

Figure 2 Geographical patterns of (a) mean annual precipitation, and species diversity separated by dispersal mode for focal angiosperm groups in Western Australia: (b) vertebrate-dispersed species, (c) wind-dispersed species, (d) ant-dispersed species. Outlined areas are the Northern, Eremaean and South West biomes.

Global Ecology and Biogeography, 18, 596–606, © 2009 Blackwell Publishing Ltd

601

A. D. Gove et al.

Figure 3 Results of jackknife tests of variable importance for (a) ant-, (b) wind- and (c) vertebrate-dispersed species using all occurrence points. Bars represent the mean training gain, a measure of the average likelihood of the point localities under the maxent probability distribution, across all species of each dispersal mode when each variable is used in isolation. Variables with the highest gains contain the most useful information for predicting the distribution of species of that dispersal mode.

rather than dispersal limitation, which determines the extent of species range boundaries. Alternative explanations are also possible, such as the environmental gradients within the two smaller regions being more pronounced, leading to smaller ranges. Nonetheless, our result makes it clear that there are strong differences among regions in range size that are largely independent of species traits. These differences make it very likely that linking species traits and range size, and hence rarity and extinction risk, will remain a difficult endeavour. To date, most studies of the relationship between range size and dispersal mode have tended to use numbers of occupied regions/quadrats or a simple rectangle based on latitudinal and longitudinal extent. Here, we chose instead to use species distribution models to estimate range size. Species distribution models offer much more nuance in modelling the occurrence of a species. However, species distribution models may overestimate the area actually inhabited by a given species by including potential rather than occupied sites. This is only a problem for our analysis if our models disproportionately overestimate the sizes of species ranges from a particular dispersal mode, which is unlikely. In fact, we found that range sizes derived from maxent

602

models and simple convex hull polygons were strongly correlated (R2 = 0.79, P < 0.001), suggesting that modelled potential ranges would not affect comparisons of range sizes amongst dispersal modes. Are there biogeographical gradients in dispersal mode? In our study, dispersal mode is strongly related to region: ant- and wind-dispersed species are concentrated in the South; vertebrate-dispersed species are concentrated in three areas (Northern, Central West and South West). One explanation for a preponderance of large-seeded, and hence vertebrate-dispersed, species in the tropics has been the suggestion that high rainfall (Lord et al., 1997) or denser shade (Bolmgren & Eriksson, 2005) leads to higher resource requirements in order to become established. This may help to explain the occurrence of vertebratedispersed species in the tropical north of Western Australia. An alternative explanation for spatial patterns in the diversity of different dispersal modes is that those patterns reflect differences in the availability of dispersal agents (e.g. Beattie, 1983; but see

Global Ecology and Biogeography, 18, 596–606, © 2009 Blackwell Publishing Ltd

Ecological constraints on range size Herrera, 1987, Jordano, 1995, and Márquez et al., 2004). In Western Australia, bird and mammal diversity peak in the north, and commensurate with that pattern the diversity of vertebrate dispersed plants is, relative to other plants, biased toward the north. The preponderance of ant-dispersed plants, on the other hand, is unlikely to be due solely to the availability of ants as dispersers. Gove et al. (2007) have shown that species richness and ant activity were not positively associated with seed removal rate along a 1650-km latitudinal transect in Western Australia. At the habitat scale, Mossop (1989) found no difference in ant activity, despite habitats differing in the frequency of ant-dispersed plant species (see Bond et al., 1991, for similar observations in South Africa). Furthermore, previous estimates of the frequency of myrmecochory predicted a gradient, ‘in which species richness and abundance of myrmecochores and diaspore-dispersing ants increase with decreasing latitude’ (Beattie, 1983, p. 249) – a prediction contrary to our observations. The preponderance of ant-dispersed plant species is often explained by low-nutrient soils (Beattie, 1985; Westoby et al., 1991), but soil type was the least important factor when modelling species distributions (Fig. 3a) in our study. High fire frequency and intensity (Majer, 1982; Bond & Slingsby, 1984; Hughes & Westoby, 1992) are considered other important predictors. CONCLUSIONS Overall, we find that range size in Western Australia is primarily a function of biome size (at least for the variables considered here). The most parsimonious general explanation for the association between biogeographical region and range size is that many species have, because of the relative climatic stability of Western Australia over the last 5 Myr and occasional longdistance dispersal, realized their potential ranges. Species potential ranges are in turn relatively constrained to those conditions within the biogeographical region in which they have evolved. As the Eremaean Province is large and expanding (e.g. Fitzpatrick et al., 2008), so are the geographical ranges of the species with their ranges centred in the Eremaean Province. In contrast, the environmental conditions characteristic of the Northern and South West provinces have contracted, leaving ranges of most species within those botanical provinces small. Under these conditions, it is unlikely that any life-history traits correlate well with plant species range size, but instead that history and the history of climate in Western Australia is far more important. This hypothesis deserves further attention in other regions and continents, but where it has been considered it finds support. For instance, several bird studies have found a relationship between species range size and extent of habitat (Duncan et al., 1999; Hawkins & Diniz-Filho, 2006). A dispersal mode, such as myrmecochory, should not necessarily be associated with increased risk of extinction due to reduced range size, as factors other than dispersal ability can be drivers of geographical range size. Despite this suggestion, ant-dispersed species, for example, are still predisposed to extinction due to their relatively high frequency within particular biomes. In our

case, ant-dispersed species are most common in the south-west of Australia, a relatively small region that is highly likely to be further reduced in habitat size as a consequence of climate change (Fitzpatrick et al., 2008). Biological trait-based approaches to predicting rarity or proneness to extinction (e.g. Foden et al., 2008; Mattila et al., 2008) may be ineffectual if not considered in terms of their relationship with geographical range placement. In the face of large-scale environmental changes, which are not expected to affect the entire globe homogenously, geographical placement per se will often be the largest factor in determining a species’ fate. ACKNOWLEDGEMENTS This research was funded by an Australian Research Council Discovery Grant to J.D.M. and R.R.D. and a Fulbright Fellowship to R.R.D. We thank Tim Barraclough, Kevin Burns, Brad Murray, Pajaro Morales and Monica Sanchez for suggested improvements to the manuscript. This paper is contribution CEDD41–2009 of the Centre for Ecosystem Diversity and Dynamics, Curtin University of Technology. REFERENCES Angiosperm Phylogeny Group. (2003) An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG II. Botanical Journal of the Linnean Society, 141, 399–436. Arita, H.T. & Rodríguez, P. (2002) Geographic range, turnover rate and the scaling of species diversity. Ecography, 25, 541– 550. Beard, J.S. (1980) A new phytogeographic map of Western Australia. Research Notes of the Western Australian Herbarium, 3, 37–58. Beard, J.S. (1990) Plant life of Western Australia. Kangaroo Press, Kenthurst. Beattie, A.J. (1983) Distribution of ant-dispersed plants. Sonderbaende de Naturwissenschaftlichen Vereins in Hamburg, 7, 249–270. Beattie, A.J. (1985) The evolutionary ecology of ant–plant mutualisms. Cambridge University Press, Cambridge. Berg, R.Y. (1975) Myrmecochorous plants in Australia and their dispersal by ants. Australian Journal of Botany, 23, 475–508. Bolmgren, K. & Eriksson, O. (2005) Fleshy fruits – origins, niche shifts, and diversification. Oikos, 109, 255–272. Bond, W.J. & Slingsby, P. (1984) Collapse of an ant-plant mutualism: the Argentine ant, Iridomyrmex humilis and myrmecochorous Proteaceae. Ecology, 65, 1031–1037. Bond, W.J., Yeaton, R. & Stock, W.D. (1991) Myrmecochory in Cape fynbos. Ant–plant interactions (ed. by C.R. Huxley & D.F. Cutler), pp. 448–462. Oxford University Press, Oxford. Bureau of Meteorology (2005) Climate data online. http:// www.bom.gov.au/climate/averages/ (accessed October 2005). Calviño-Cancela, M., Dunn, R.R., van Etten, E.J.B. & Lamont, B.B. (2006) Emus as non-standard seed dispersers and their potential for long-distance dispersal. Ecography, 29, 632–640.

Global Ecology and Biogeography, 18, 596–606, © 2009 Blackwell Publishing Ltd

603

A. D. Gove et al. Carter, R.N. & Prince, S.D. (1988) Distribution limits from a demographic viewpoint. Plant population ecology (ed. by A.J. Davy, M.J. Hutchings & A.R. Watkinson), pp. 165–184. Blackwell Scientific Publications, Oxford. Clark, J.S. (1998) Why trees migrate so fast: confronting theory with dispersal biology and the paleorecord. The American Naturalist, 152, 204 –224. Cody, M.L. & Overton, J.M. (1996) Short-term evolution of reduced dispersal in island plant populations. Journal of Ecology, 84, 53–61. Colwell, R.K. & Lees, D.C. (2000) The mid-domain effect: geometric constraints on the geography of species richness. Trends in Ecology and Evolution, 15, 70–76. Colwell, R.K., Rahbek, C. & Gotelli, N.J. (2004) The mid-domain effect and species richness patterns: What have we learned so far? The American Naturalist, 163, E1–E23. Cowling, R.M. & Lamont, B.B. (1998) On the nature of Gondwanan species flocks: diversity of Proteaceae in Mediterranean south-western Australia and South Africa. Australian Journal of Botany, 46, 335–355. Department of Agriculture, Fisheries and Forestry (2005) Digital atlas of Australian soils. http://www.daff.gov.au/brs/data-tools/ daas-download (accessed October 2005). Department of Environment and Conservation (2005) FloraBase. http://florabase.calm.wa.gov.au/ (accessed May 2005). Dodson, J.R. (1998) Timing and response of vegetation change to Milankovitch forcing in temperate Australia and New Zealand. Global and Planetary Change, 18, 161–174. Duncan, R.P., Blackburn, T.M. & Veltman, C.J. (1999) Determinants of geographical range sizes: a test using introduced New Zealand birds. Journal of Animal Ecology, 68, 963–975. Dunn, R.R. & Romdal, T.S. (2005) Mean latitudinal range sizes of bird assemblages in six Neotropical forest chronosequences. Global Ecology and Biogeography, 14, 359–366. Dunn, R.R., Gove, A.D., Barraclough, T.G., Givnish, T.J. & Majer, J.D. (2007) Convergent evolution of an ant–plant mutualism across plant families, continents, and time. Evolutionary Ecology Research, 9, 1349–1362. Dunn, R., Gove, A. & Majer, J.D. (2008) Seed dispersal mutualisms with ants and patterns of plant diversity in Western Australia. In: Fronteiras do conhecimento em Insetos Sociais (ed. by E.F. Vilela, I.A. Santos, J.H. Schoereder, L.A.O. Campos & J.E. Serrão), pp. 325–347. Editora Universidade Federal de Viçosa, Viçosa, Brazil. Dynesius, M. & Jansson, R. (2000) Evolutionary consequences of changes in species’ geographical distributions driven by Milankovitch climate oscillations. Proceedings of the National Academy of Science USA, 97, 9115–9120. Edwards, W. & Westoby, M. (1996) Reserve mass and dispersal investment in relation to geographic range of plant species: phylogenetically independent contrasts. Journal of Biogeography, 23, 329–338. Elith, J., Graham, C.H., Anderson, R.P., Dudík, M., Ferrier, S., Guisan, A., Hijmans, R.J., Huettmann, F., Leathwick, J.R., Lehmann, A., Li, J., Lohmann, L.G., Loiselle, B.A., Manion, G., Moritz, C., Nakamura, M., Nakazawa, Y., Overton, J.,

604

Peterson, A.T., Phillips, S.J., Richardson, K.S., ScachettiPereira, R., Schapire, R.E., Soberón, J., Williams, S., Wisz, M.S. & Zimmermann, N.E. (2006) Novel methods improve prediction of species’ distributions from occurrence data. Ecography, 29, 129–151. ESRI (2006) ArcMap, Version 9.1. Environmental Systems Research Institute, Redlands, CA, USA. Fitzpatrick, M.C., Gove, A.D., Sanders, N.J. & Dunn, R.R. (2008) Climate change, plant migration and range collapse in a global biodiversity hotspot: the Banksia (Proteaceae) of Western Australia. Global Change Biology 14, 1337– 1352. Foden, W., Mace, G., Vie, J., Angulo, A., Butchard, S., Devantier, L., Dublin, H., Gutsche, A., Stuart, S. & Turak, E. (2008) Species susceptibility to climate change impact. The 2008 review of the IUCN Red List of threatened species (ed. by J. Vie, C. HiltonTaylor & S. Stuart). IUCN, Gland, Switzerland. Gómez, C. & Espadaler, X. (1998) Myrmecochorous dispersal distances: a world survey. Journal of Biogeography, 25, 573– 580. Gove, A.D., Majer, J.D. & Dunn, R.R. (2007) A keystone ant species promotes seed dispersal in a ‘diffuse’ mutualism. Oecologia 153, 687–697. Groom, P.K. & Lamont, B.B. (1996) Ecogeographical analysis of Hakea (Proteaceae) in south-western Australia, with special reference to leaf morphology and life form. Australian Journal of Botany, 44, 527–542. Hanski, I. (1999) Metapopulation ecology. Oxford University Press, Oxford. Hawkins, B.A. & Diniz-Filho, J.A.F. (2006) Beyond Rapoport’s rule: evaluating range size patterns of New World birds in a two-dimensional framework. Global Ecology and Biogeography, 15, 461–469. He, T., Lamont, B.B., Krauss, S.L., Enright, N.J., Miller, B.P. & Gove, A.D. (2009) Ants cannot account for interpopulation dispersal of the arillate pea Daviesia triflora. New Phytologist, 181, 725–733. Hernandez, P.A., Graham, C.H., Master, L.L. & Albert, D.L. (2006) The effect of sample size and species characteristics on performance of different species distribution modelling methods. Ecography, 29, 773–785. Herrera, C.M. (1987) Vertebrate-dispersed plants of the Iberian peninsula: A study of fruit characteristics. Ecological Monographs, 57, 305–331. Higgins, S.I. & Richardson, D.M. (1999) Predicting plant migration rates in a changing world: the role of long-distance dispersal. The American Naturalist, 153, 464–475. Higgins, S.I., Nathan, R. & Cain, M.L. (2003) Are long-distance dispersal events in plants usually caused by nonstandard means of dispersal? Ecology, 84, 1945–1956. Hnatiuk, R.J. & Maslin, B.R. (1988) Phytogeography of Acacia in Australia in relation to climate and species-richness. Australian Journal of Botany, 36, 361–383. Hoot, S.B. & Douglas, A.W. (1998) Phylogeny of the Proteaceae based on atpB and atpB-rbcL spacer region sequences. Australian Systematic Botany, 11, 301–320.

Global Ecology and Biogeography, 18, 596–606, © 2009 Blackwell Publishing Ltd

Ecological constraints on range size 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. (1992) Patterns of plant diversity at the population and species levels in south-west Australian mediterranean ecosystems. Biodiversity of mediterranean ecosystems in Australia (ed. by R.J. Hobbs), pp. 27– 46. Surrey Beatty & Sons, Chipping Norton, NSW. Hopper, S.D. & Gioia, P. (2004) The southwest Australian floristic region: evolution and conservation of a global hot spot of biodiversity. Annual Review of Ecology and Systematics, 35, 623–650. Hopper, S.D. & Maslin, B.R. (1978) Phytogeography of Acacia in Western Australia. Australian Journal of Botany, 26, 63–78. Hubbell, S.P. (2001) The unified neutral theory of biodiversity and biogeography. Princeton University Press, Princeton. Hughes, L. & Westoby, M. (1992) Fates of seed adapted for dispersal by ants in Australian sclerophyll vegetation. Ecology, 73, 1285–1299. Hughes, L., Cawsey, E.M. & Westoby, M. (1996) Geographic and climatic range sizes of Australian eucalypts and a test of Rapoport’s rule. Global Ecology and Biogeography, 5, 128–142. Jordano, P. (1995) Angiosperm fleshy fruits and seed dispersers: a comparative analysis of adaptation and constraints in plant-animal interactions. The American Naturalist, 145, 163– 191. Kalisz, S., Hanzawa, F.M., Tonsor, S.J., Thiede, D.A. & Voigt, S. (1999) Ant-mediated seed dispersal alters patterns of relatedness in a population of Trillium grandiflorum. Ecology, 80, 2620–2634. Kathriarachchi, H., Samuel, R., Hoffmann, P., Mlinarec, J., Wurdack, K.J., Ralimanana, H., Stuessy, T.F. & Chase, M.W. (2006) Phylogenetics of tribe Phyllantheae (Phyllanthaceae; Euphorbiaceae sensu lato) based on nrITS and plastid matK DNA sequence data. American Journal of Botany, 93, 637–655. Keighery, G.J. (1996) Phytogeography, biology and conservation of Western Australian Epacridaceae. Annals of Botany, 77, 347– 355. Kelly, C.K. & Woodward, F.I. (1996) Ecological correlates of plant range size: taxonomies and phylogenies in the study of plant commonness and rarity in Great Britain. Philosophical Transactions of the Royal Society B: Biological Sciences. 351, 1261–1269. Kunin, W.E. & Gaston, K.J. (1997) The biology of rarity: causes and consequences of rare-common differences. Springer, Berlin. Lamont, B.B. & Connell, S.W. (1996) Biogeography of Banksias in southwestern Australia. Journal of Biogeography, 23, 295–309. Latimer, A.M., Silander, J.A. & Cowling, R.M. (2005) Neutral ecological theory reveals isolation and rapid speciation in a biodiversity hot spot. Science, 309, 1722–1725. Lengyel, S., Gove, A.D., Latimer, A.M., Majer, J.D. & Dunn, R.R. (2009) Ants sow the seeds of global diversification in flowering plants. PLoS ONE, 4, e5480. doi: 10.1371/journal.pone.0005480. Lester, S.E., Ruttenberg, B.I., Gaines, S.D. & Kinlan, B.P. (2007) The relationship between dispersal ability and geographic range size. Ecology Letters, 10, 745–758.

Lord, J.M., Egan, J., Clifford, T., Jurado, E., Leishman, M., Williams, D. & Westoby, M. (1997) Larger seeds in tropical floras: consistent patterns independent of growth form and dispersal mode. Journal of Biogeography, 24, 205–211. McDonald, D.J., Juritz, J.M., Cowling, R.M. & Knottenbelt, W.J. (1995) Modelling the biological aspects of local endemism in South African fynbos. Plant Systematics and Evolution, 195, 137–147. McDonald, R., McKnight, M., Weiss, D., Selig, E., O’Connor, M., Violin, C. & Moody, A. (2005) Species compositional similarity and ecoregions: do ecoregion boundaries represent zones of high species turnover? Biological Conservation, 126, 24– 40. Majer, J.D. (1982) Ant–plant interactions in the Darling Botanical District of Western Australia. Ant–plant interactions in Australia (ed. by Buckley, R.C.), pp. 45–61. Dr W. Junk Publishers, The Hague. Márquez, A.L., Real, R. & Vargas, J.M. (2004) Dependence of broad-scale geographical variation in fleshy-fruited plant species richness on disperser bird species richness. Global Ecology and Biogeography, 13, 295–304. Mattila, N., Kotiaho, J.S., Kaitala, V. & Komonen, A. (2008) The use of ecological traits in extinction risk assessments: a case study on geometrid moths. Biological Conservation, 141, 2322– 2328. Moles, A.T., Ackerly, D.D., Tweddle, J.C., Dickie, J.B., Smith, R., Leishman, M., Mayfield, M.M., Pitman, A., Wood, J.T. & Westoby, M. (2007) Global patterns in seed size. Global Ecology and Biogeography, 16, 109–116. Mossop, M.K. (1989) Comparison of seed removal by ants in vegetation on fertile and infertile soils. Australian Journal of Ecology, 14, 367–374. Murray, B.R., Thrall, P.H., Gill, A.M. & Nicotra, A.B. (2002a) How plant life-history and ecological traits relate to species rarity and commonness at varying spatial scales. Austral Ecology, 27, 291–310. Murray, B.R., Thrall, P.H. & Lepschi, B.J. (2002b) Relating species rarity to life history in plants of eastern Australia. Evolutionary Ecology Research, 4, 937–950. Oakwood, M., Jurado, E., Leishman, M. & Westoby, M. (1993) Geographic ranges of plant species in relation to dispersal morphology, growth form and diaspore weight. Journal of Biogeography, 20, 563–572. Olson, D.M., Dinerstein, E., Wikramanayake, E.D., Burgess, N.D., Powell, G.V.N., Underwood, E.C., D’Amico, J.A., Itoua, I., Strand, H.E., Morrison, J.C., Loucks, J., Allnutt, T.F., Ricketts, T.H., Kura, Y., Lamoreux, J.F., Wettengel, W.W., Hedao, P. & Kassem, K.R. (2001) Terrestrial ecoregions of the world: A new map of life on Earth. BioScience, 51, 933–938. Pearson, R.G., Raxworthy, C.J., Nakamura, M. & Peterson, A.T. (2007) Predicting species’ distributions from small numbers of occurrence records: a test case using cryptic geckos in Madagascar. Journal of Biogeography, 34, 102–117. Phillips, S.J., Anderson, R.P. & Schapire, R.E. (2006) Maximum entropy modeling of species geographic distributions. Ecological Modelling, 190, 231–259.

Global Ecology and Biogeography, 18, 596–606, © 2009 Blackwell Publishing Ltd

605

A. D. Gove et al. Pocock, M., Hartley, S., Telfer, M., Preston, C. & Kunin, W. (2006) Ecological correlates of range structure in rare and scarce British plants. Journal of Ecology, 94, 581–596. Purvis, A. & Rambaut, A. (1995) Comparative analysis by independent contrasts (CAIC): an Apple Macintosh application for analysing comparative data. Computer Applications in the Biosciences, 11, 247–251. Rosenzweig, M.L. (1995) Species diversity in space and time. Cambridge University Press, Cambridge. Ruokolainen, K. & Vormisto, J. (2000) The most widespread Amazonian palms tend to be tall and habitat generalists. Basic and Applied Ecology, 1, 97–108. Ruokolainen, K., Tuomisto, H., Vormisto, J. & Pitman, N. (2002) Two biases in estimating range sizes of Amazonian plant species. Journal of Tropical Ecology, 18, 935–942. Sérgio, C., Figueira, R., Draper, D., Menezes, R. & Sousa, J.A. (2007) Modelling bryophyte distribution based on ecological information for extent of occurrence assessment. Biological Conservation, 135, 341–351. Smith, F.D.M., May, R.M. & Harvey, P.H. (1994) Geographical ranges of Australian mammals. Journal of Animal Ecology, 63, 441–450. Stockwell, D.R.B. & Peterson, A.T. (2002) Effects of sample size on accuracy of species distribution models. Ecological Modelling, 148, 1–13. Thompson, K., Gaston, K.J., & Band, S.R. (1999) Range size, dispersal and niche breadth in the herbaceous flora of central England. Journal of Ecology, 87, 150–155. Urban, M.C., Leibold, M.A., Amarasekare, P., Meester, L.D., Gomulkiewicz, R., Hochberg, M.E., Klausmeier, C.A., Loeuille, N., de Mazancourt, C., Norberg, J., Pantel, J.H., Strauss, S.Y., Vellend, M. and Wade, M.J. (2008) The evolutionary ecology of metacommunities. Trends in Ecology and Evolution, 23, 311–317. Vellend, M., Myers, J.A., Gardescu, S. & Marks, P.L. (2003) Dispersal of Trillium seeds by deer: implications for longdistance migration of forest herbs. Ecology, 84, 1067–1072. Weir, J.E.S. & Corlett, R.T. (2006) How far do birds disperse seeds in the degraded tropical landscape of Hong Kong, China? Landscape Ecology, 22, 131–140. Westcott, D.A., Bentrupperbäumer, J., Bradford, M.G. &

606

McKeown, A. (2005) Incorporating patterns of disperser behaviour into models of seed dispersal and its effects on estimated dispersal curves. Oecologia, 146, 57–67. Westoby, M., Rice, B. & Howell, J. (1990) Seed size and plant growth form as factors in dispersal spectra. Ecology, 71, 1307– 1315. Westoby, M., French, K., Hughes, L., Rice, B. & Rodgerson, L. (1991) Why do more plant species use ants for dispersal on infertile compared with fertile soils? Australian Journal of Ecology, 16, 445–455. Whitlock, B.A., Bayer, C. & Baum, D.A. (2001) Phylogenetic relationships and floral evolution of the Byttnerioideae (‘Sterculiaceae’ or Malvaceae s.l.) based on sequences of the chloroplast gene, ndhF. Systematic Botany, 26, 420–437. Willson, M.F., Irvine, A.K. & Walsh, N.G. (1989) Vertebrate dispersal syndromes in some Australian and New Zealand plant communities, with geographic comparisons. Biotropica, 21, 133–147. Woodward, F.I. (1987) Climate and plant distribution. Cambridge University Press, Cambridge. Wurdack, K.J., Hoffmann, P., Samuel, R., de Bruijn, A., van der Bank, M. & Chase, M.W. (2004) Molecular phylogenetic analysis of Phyllanthaceae (Phyllanthoideae pro parte, Euphorbiaceae sensu lato) using plastid rbcL DNA sequences. American Journal of Botany, 91, 1882–1900. Wurdack, K.J., Hoffmann, P. & Chase, M.W. (2005) Molecular phylogenetic analysis of uniovulate Euphorbiaceae (Euphorbiaceae sensu stricto) using plastid rbcL and trnL-F DNA sequences. American Journal of Botany, 92, 1397–1420. BIOSKETCH Aaron Gove is a Postdoctoral Researcher at Curtin University of Technology. He is interested in plant–animal interactions and the formation and conservation of biological diversity. His recent research has concentrated on the role of myrmecochory in the formation of biogeographical patterns in angiosperms. Editor: Thomas Gillespie

Global Ecology and Biogeography, 18, 596–606, © 2009 Blackwell Publishing Ltd