The evolution and phylogenetic placement of invasive Australian Acacia species

Authors

  • Joseph T. Miller,

    Corresponding author
    1. Centre for Australian National Biodiversity Research, CSIRO Division of Plant Industry, Canberra, ACT 2601, GPO Box 1600, Australia
      Joseph T. Miller, Centre for Australian National Biodiversity Research, CSIRO Division of Plant Industry, Canberra, ACT 2601, GPO Box 1600, Australia.
      E-mail: joe.miller@csiro.au
    Search for more papers by this author
  • Daniel J. Murphy,

    1. National Herbarium of Victoria, Royal Botanic Gardens Melbourne, Vic. 3141, Australia
    Search for more papers by this author
  • Gillian K. Brown,

    1. School of Botany, The University of Melbourne, Vic. 3010, Australia
    Search for more papers by this author
  • David M. Richardson,

    1. Centre for Invasion Biology, Department of Botany and Zoology, Stellenbosch University, Matieland 7602, South Africa
    Search for more papers by this author
  • Carlos E. González-Orozco

    1. Centre for Australian National Biodiversity Research, CSIRO Division of Plant Industry, Canberra, ACT 2601, GPO Box 1600, Australia
    Search for more papers by this author

Joseph T. Miller, Centre for Australian National Biodiversity Research, CSIRO Division of Plant Industry, Canberra, ACT 2601, GPO Box 1600, Australia.
E-mail: joe.miller@csiro.au

Abstract

Aim Acacia is the largest genus of plants in Australia with over 1000 species. A subset of these species is invasive in many parts of the world including Africa, the Americas, Europe, the Middle East, Asia and the Pacific region. We investigate the phylogenetic relationships of the invasive species in relation to the genus as a whole. This will provide a framework for studying the evolution of traits that make Acacia species such successful invaders and could assist in screening other species for invasive potential.

Location  Australia and global.

Methods  We sequenced four plastid and two nuclear DNA regions for 110 Australian Acacia species, including 16 species that have large invasive ranges outside Australia. A Bayesian phylogenetic tree was generated to define the major lineages of Acacia and to determine the phylogenetic placement of the invasive species.

Results  Invasive Acacia species do not form a monophyletic group but do form small clusters throughout the phylogeny. There are no taxonomic characters that uniquely describe the invasive Acacia species.

Main conclusions  The legume subfamily Mimosoideae has a high percentage of invasive species and the Australian Acacia species have the highest rate of all the legumes. There is some evidence of phylogenetic clumping of invasive species of Acacia in the limited sampling presented here. This phylogeny provides a framework for further testing of the evolution of traits associated with invasiveness in Acacia.

Introduction

Acacia Mill. was first described by Miller (1754), and the genus has gone through much iteration with Acacia s.l. as we currently understand it largely framed by Bentham (1875). The distributional range of Acacia s.l. is broad, covering much of the world’s subtropical arid areas (Lewis et al., 2005) including the Americas, Africa, Asia and Australia. Through extensive morphological and molecular systematic studies (Chappill & Maslin, 1995; Brain & Maslin, 1996; Miller & Bayer, 2000, 2001; Luckow et al., 2003; Brown et al., 2008), it is now agreed that Acacia s.l. comprises three large lineages, roughly equivalent to subgenera Acacia, Aculiferum Vassal and Phyllodineae (DC.) Seringe, and three newly segregated smaller genera (Box 1). See Maslin et al. (2003) for a detailed overview of the generic history of Acacia s.l.

In 2005, a retypfication of the genus from the African species Acacia nilotica (L.) Willd. Ex Delile, now Vachellia nilotica (L.) P.J.H. Hurter & Mabb. (Mabberley, 2008), to the Australian species Acacia penninervis Sieber ex DC. was approved at the XVII International Botanical Congress in Vienna (McNeill et al., 2005). Under this change, most Australian species belong to the genus Acacia. Acceptance of this retypification still remains controversial (Luckow et al., 2005; Rijckevorsel, 2006; Smith et al., 2006; Moore, 2007, 2008).

There are 1028 species of Acacia s.l. in Australia, of which 1012 (Council of Heads of Australasian Herbaria, 2010); Table 1) belong to a single clade, formerly known as Acacia subg. Phyllodineae (synonymous with Racosperma Mart.), now Acacia s.s. in the new taxonomy. The remaining 16 species found in Australia, which include some naturalized taxa, comprise taxa of Acaciella Britton & Rose (3), Senegalia Rafinesque (2) and Vachellia (11) (Council of Heads of Australasian Herbaria, 2010). Eighteen species of Acacia s.s. occur naturally outside continental Australia, ten of which are not found in Australia. This paper and the others in this special issue of Diversity and Distributions focus on the 1012 species, a group which are often referred to as the Australian acacias (Richardson et al., 2011).

Table 1.   Invasive legume species based on taxonomic category.
Taxonomic groupNumber of invasive speciesTotal number of speciesPercentage invasive
  1. Species totals are from Lewis et al. (2005) and Council of Heads of Australasian Herbaria (2010). The number of woody trees and shrubs that are invasive species is from Richardson & Rejmánek (2011). Many taxa of subfamily Papilionoideae are not woody and therefore not included in the included in the Richardson & Rejmánek (2011) list or in this table.

Leguminosae (Fabaceae)12119,3200.63
Subfamily Caesalpinioideae2222500.98
Subfamily Mimosoideae5632701.71
Tribe Acacieae3114502.14
Australian Acacia2310202.25
Tribe Ingeae119511.16
Tribe Mimoseae148691.61

Recent phylogenetic analysis of Acacia s.s. has identified well-supported clades that do not closely resemble the traditional classification of taxa by Pedley (1978) or Bentham (1875). These groups were informally named by Murphy et al. (2010) as comprising five clades. Two of these clades were resolved at basal nodes and consisted mostly of uninerved phyllodinous taxa. One clade was named the ‘Acacia victoriae and Acacia pyrifolia clade’, and the second the ‘Acacia murrayana clade’, with these two groups occurring predominantly in semi-arid and arid regions of Australia. A ‘Pulchelloidea clade’ was named to comprise members of the sections Pulchellae, Alatae, Lycopodiifoliae and some members of sect. Phyllodineae of Pedley (1978). The fourth named clade, the ‘p.u.b. clade’, was a large assemblage of plurinerved and uninerved phyllodinous taxa and also bi-pinnate taxa from section Botrycephalae (members of which were placed in a fifth named informal ‘Botrycephalae subclade’).

Dispersal of Australian acacias has happened several times with presumed long-distance dispersal events to Madagascar, Hawaii and South East Asia (Pedley, 1975, 1986). These dispersal events are hypothesized to predate human intervention in species distribution boundaries, although occurring relatively recently in geological time. However, such hypotheses remain to be tested. Owing to the consequences of human intervention (Griffin et al., 2011; Le Roux et al., 2011; Richardson et al., 2011), the ranges of some Australian acacias have changed dramatically. Many legumes and species of Acacia in particular have been moved around the world by humans for various purposes ranging from sand dune stabilization and forestry as well as for the tannin and perfume industries (Maslin, 2001; Marchante et al., 2008; Kull et al., 2011; Richardson et al., 2011). In many cases, an unforeseen consequence of the cultivation of Acacia species has allowed their establishment and invasion in many parts of the world. Twenty-three species of Australian acacias are now considered invasive outside Australia. Some species such as Acacia mearnsii, Acacia longifolia and Acacia cyclops are invasive in many areas of the world from Africa, Europe, the Middle East, the Americas to Asia. Others such as A. victoriae were only recently recognized as invasive (Richardson & Rejmánek, 2011). The high level of Acacia use worldwide is expected to provide an opportunity for other species to become invasive in the future (Wilson et al., 2011).

The goal of this paper is to explore the molecular phylogenetic relationships of selected invasive species of Acacia within a broader phylogenetic framework of the Acacia s.s. clade using plastid and nuclear ribosomal DNA sequence data. We then use the resulting phylogeny to answer questions regarding invasive species of Acacia such as: (1) Do the invasive species form a monophyletic group(s)? (2) What are the sister taxon relationships to invasive species? (3) Are there key morphological or spatial traits that correlate with the invasive species?

Table Box 1.   Mimosoideae: phylogenetic overview and weedy taxa
inline image
The Leguminosae, is one of the largest plant families with over 700 genera and 19,000 species (Lewis et al., 2005). The family has three subfamilies two of which are monophyletic (Papilionoideae and the Mimosoideae) and each is derived from a paraphyletic Caesalpinoideae.
The legume subfamily, Mimosoideae, is predominantly comprised of three large tribes: Mimoseae, Acacieae and the Ingeae. Molecular phylogenetic research over the last 10 years has tested these tribal boundaries and in turn the relationships of individual genera within the tribes. These results clearly indicate that the three large tribes are not natural lineages (Bukhari et al., 1999; Miller & Bayer, 2000, 2001; Luckow et al., 2003, Brown et al., 2008). The Mimoseae is basal with the smaller tribe Parkieae and part of the Acacieae embedded within it. The tribe Ingeae is paraphyletic in relation to the Acacieae, which is polyphyletic. It appears that a meaningful monophyletic tribal classification for the subfamily will be difficult to circumscribe.
Also these results clearly demonstrate that Acacia s.l. is not monophyletic and that the three main lineages are not closely related. Acacia subg. Acacia is a well-supported clade placed within in the Tribe Mimoseae. It is closely related to the basal Mimoseae such as Prosopis, Neptunia, Desmanthus and Leucaena and to the more derived Mimoseae clade which contains the genera Parkia, Piptadenia, Anadenanthera, Mimosa and Microlobius. Many of these genera contain invasive species.
The largest subgenus, Acacia subg. Phyllodineae, is more closely related to the species of the Tribe Ingeae than to other Acacia s.l. species. (See Box 2 for more phylogenetic information on the Ingeae.)
Acacia subg. Acueiferum was determined to be non-monophyletic and three genera have been named or proposed for the segregate lineages (Rico Arce, 2006; Seigler et al., 2006). The core of Acacia subg. Aculeiferum is monophyletic and consists of over 200 species with a range similar to that of Acacia subg Acacia: ranging from the America, Africa, Asia and into northern Australia.
Invasive species are common in the family with 122 species known to be invasive. The phylogenetic distribution of the invasive species in skewed both due to historical and biological reasons (Richardson et al., 2011).

Methods

Taxon sampling

The sampling consisted of 121 OTUs representing 110 Acacia species (Appendix 1). The species were selected based on the main lineages of Acacia (Murphy et al., 2010) and the list of invasive and non-invasive taxa present in South Africa (Richardson et al., 2011). Sixteen sampled species are invasive. Three separate data alignments and phylogenetic analyses were conducted. First, a dataset consisting of 60 species of Acacia was developed which represents all the main lineages of Acacia so far identified (Murphy et al., 2010). This will be referred to as the ‘overall’ dataset and analysis. Owing to high levels of variation, portions of the DNA sequences were not able to be aligned; therefore, two subset analyses were also performed. The subset analyses allowed better sequence alignment and homology assessment, and resulted in more sites included in the phylogenetic analyses. The first subset consisted of 43 OTUs and 40 species and will be referred to as the ‘mearnsii’ dataset because of the presence of this representative invasive species. The second comprised 46 OTUs and 44 species and will be referred to as the ‘melanoxylon’ dataset because of the presence of this representative invasive species.

Invasive species of Acacia are represented in all three datasets. Placeholder species from the mearnsii and melanoxylon datasets were included in the overall dataset. Outgroup taxa, Parachidendron pruionsum and Paraserianthes lophantha subsp. lophantha, were chosen based on results of previous studies (Miller & Bayer, 2000, 2001; Luckow et al., 2003; Brown et al., 2008). Fresh leaf samples were collected either in the field or from cultivated plants of known provenance, and where no other material was available, from herbarium specimens.

DNA isolation, amplification and sequencing

Genomic DNA was extracted from 10 to 100 mg of fresh or silica gel–dried leaf tissue, or from herbarium material, using the DNeasy Plant Mini kit (Qiagen, Valencia, CA) either individually or in the 96-well plate format. Six regions were amplified and sequenced of which four were plastid loci and two nuclear. The four plastid loci sequenced were psbA-trnH intergenic spacer, trnL-F intron and intergenic spacer, rpl32-trnL intergenic spacer and a portion of the matK region. All amplifications were performed using the PCR profile outlined by Shaw et al. (2005). The primers used were as follows: psbA-trnH [Sang et al. (1997)], trnL-F [Taberlet et al. (1991)], rpl32-trnL [Shaw et al. (2007)] and matK 59R/6 [Johnson & Soltis (1994)]. The complete sequences of nuclear ribosomal DNA internal (ITS) and external (ETS) transcribed spacers were amplified and sequenced using the primers and protocols described by Murphy et al. (2010). All sequences are lodged in Genbank (JF419907JF420546).

Phylogenetic analyses

Contiguous sequences were edited using Sequencher™ v.3.0 (Gene Codes Corporation, Ann Arbor, MI) and manually aligned in BioEdit sequence alignment editor v.4.8.6 (Hall, 1999). Sequence alignments and PAUP/Nexus formatted files are available from the authors upon request, and all sequences are lodged in Genbank (see Appendix 1).

Any uncertain base positions, generally located close to priming sites, and highly variable regions with uncertain sequence homology, were excluded from phylogenetic analysis. Individual base positions were coded as unordered multistates, and potentially informative insertions/deletions (indels) were coded as additional binary characters.

Bayesian analyses were performed using MrBayes version 3.1.2 (Ronquist & Huelsenbeck, 2003). Posada & Crandall (1998) (Modeltest v.1.1) determined that the GTR + I + gamma model was the best-fit model for both the plastid and nuclear partitions, and it was applied to each DNA sequence partition. Indel characters were included as a separate partition, and a standard (morphology) discrete state model with a gamma shape parameter was applied to this partition (Lewis, 2001). The Markov chain Monte Carlo search was run for 5 million generations with trees sampled every 1000 generations. MrBayes performed two simultaneous analyses starting from different random trees (Nruns = 2), each with four Markov chains (Nchains = 4). The first 2000 trees were discarded from each run. A Bayesian consensus phylogram with posterior probability values plotted was calculated in MrBayes. Maximum parsimony analyses were performed with the heuristic search option (excluding uninformative characters) in PAUP* 4.02 (Swofford, 1999). A four-step search method for multiple islands was performed with 10,000 random replicates (Olmstead & Palmer, 1994). Support for internal branches was evaluated by the fast bootstrap method with 10,000 replicates (Felsenstein, 1985). A partition homogeneity test was conducted in PAUP* 4.02 (Swofford, 1999) using 100 random replications to test whether the plastid and nuclear data partitions are congruent.

Results

Phylogenetic analyses

In the overall dataset that contained the broadest sampling of species of Acacia, the four concatenated plastid sequences aligned to 3337 nucleotides, while the nuclear ribosomal DNA aligned to 1263 nucleotides, and for this dataset, 29 indel characters were scored. The two subset analyses had shorter alignments, because of fewer indels than found in the overall dataset.

For each dataset, the nuclear and plastid sequence partitions were analyzed separately. The partition homogeneity test indicated that the two partitions were congruent and the resulting phylogenies (not shown) were broadly concordant. The minor discrepancies between the nuclear and plastid phylogenies are only at the branch tips where the posterior probabilities and bootstrap values are low.

The main clades resolved are broadly congruent in the three trees presented here (Figs 1–3: overall, melanoxylon and mearnsii). Two invasive taxa, namely Acacia saligna and A. victoriae, were only included in the overall analysis and not in the subset analyses because these taxa are not members of the two subset clades.

Figure 1.

 Bayesian phylogenetic tree of the ‘overall’ dataset. Number above node is the Bayesian posterior probability (PP) followed by the bootstrap value. Bold branches indicate PP > 0.90. Taxa in bold italic are invasive with the following codes indicating areas of invasiveness. Af, Africa; Am, North and South America and Caribbean Islands; EME, Europe, Middle East and Atlantic Islands; AsPac, Asia, Australia, New Zealand, Indian and Pacific Islands.

Figure 2.

 Bayesian phylogenetic tree of the ‘melanoxylon’ dataset. Notes as in Fig. 1.

Figure 3.

 Bayesian phylogenetic tree of the ‘mearnsii’ dataset. Notes as in Fig. 1.

In the overall tree (Fig. 1), the major lineages shown, similar to Murphy et al. (2010), are supported with maximal posterior probability values (PP = 1.00). These are (A) The A. victoriae and A. pyrifolia clade, (B) The Pulchelloidea clade, (C) The A. murrayana clade, (D) The melanoxylon clade and (E) The mearnsii clade. The latter two are equivalent to the p.u.b. clade of Murphy et al. (2010). As previously found and based on the current sampling, none of the sections of Pedley (1978) are resolved as monophyletic. However, additional resolution was found for taxa within the melanoxylon and mearnsii clades than previously discovered. It is notable that the current phylogenetic analysis has identified new taxa that should be placed within the lineages found by Murphy et al. (2010).

Clades that include invasive taxa (Figs 1–3) have been identified as follows:

  • A. Resolved at the earliest diverging node of the overall tree (Fig. 1, clade A) is A. victoriae, placed in a clade with A. pyrifolia and Acacia dempsteri (the A. victoriae and A. pyrifolia clade, PP = 1.00).
  • B. In the Pulchelloidea clade (Fig. 1, clade B, PP = 1.00), A. saligna, with A. alata, is the sister group (PP = 0.98) to the remaining eleven sampled Pulchelloidea taxa. Acacia saligna is the sole invasive taxon so far identified in the Pulchelloidea clade.
  • C. There are no invasive species within the A. murrayana clade (Fig. 1, clade C).
  • The melanoxylon (Fig. 1, clade D) and mearnsii (Fig. 1, clade E) clades together (Fig. 1, PP: 1.00) are equivalent to the largely unresolved p.u.b. clade of Murphy et al. (2010). Additional taxa were included in the present subset analyses for the melanoxylon clade (Fig. 2) and the mearnsii clade (Fig. 3), and most of the invasive taxa are found in these clades.

  • D. The melanoxylon clade contains six recognized invasive species in four broad groups:
    • (i)In the Acacia cognata subclade (Fig. 2, top), a notable group of invasive species occurs. These are A. implexa and A. melanoxylon as sister taxa (PP = 1.00), related to a clade, with very low PP support (PP = 0.51), that include A. verticillata, A. genistifolia, A. baeuerlenii, and A. elongata.
    • (ii)Acacia cyclops and A. ixiophylla (PP = 0.61) are the sister clades to the remaining taxa in the A. cognata clade of which A. cyclops is invasive.
    • (iii)An A. longifolia clade, as identified by Brown et al. (2010), is resolved (with PP = 0.99) to include Acacia longissima, Acacia mucronata and the invasive A. longifolia.
    • (iv)In the Acacia aneura subclade (Fig. 2, bottom), only a single invasive lineage is identified; A. crassicarpa and A. holosericea are sister clades (PP: 0.65) to A. aulacocarpa, an apparently non-invasive species.
  • E. The large mearnsii clade (Fig. 3), which includes uninerved phyllodinous and bipinnate taxa also includes A. penninervis, the newly designated type species of Acacia (Orchard & Maslin, 2003). This clade includes seven invasive species and therefore has the largest number of invasive taxa within it.

Some grouping of invasive species is noted within the mearnsii clade. Five invasive species (Acacia baileyana, A. dealbata, A. decurrens, A. mearnsii and A. podalyriifolia) occur in this clade along with 10 non-invasive species (PP = 1.00). This clade represents taxa with both bipinnate and phyllodinous mature vegetative leaves. The other invasive taxa in the mearnsii clade, A. pycnantha and A. elata do not group closely together.

Of the 23 known invasive Acacia species, seven are not sampled in this phylogeny. Based on previous knowledge, we can estimate that A. mangium and A. auriculiformis would group in the aneura subclade of the melanoxylon clade, possibly near the invasive species A. crassicarpa and A. holosericea. No confident prediction can be undertaken as yet for the phylogenetic placement of the other invasive species A. iteaphylla, A. paradoxa, A. retinodes, A. salicinia and A. stricta. These species await future molecular phylogenetic research.

Discussion

With over 19,000 species, the legumes are one of the largest families of flowering plants (Lewis et al., 2005). The family comprises a paraphyletic subfamily, the Caesalpinioideae and two monophyletic subfamilies, the Papilionoideae and the Mimosoideae. The largest subfamily, Papilionoideae, contains over 13,800 species, many of which are important as food crops such as Glycine, Pisum and Vigna.

Richardson & Rejmánek (2011) identified 121 woody legume species that are clearly invasive (sensuPyšek et al., 2004) somewhere in the world. However, weed species are apparently not evenly spread in a phylogenetic sense across subfamilies or within them. For example, 56 of the 3270 recognized species in subfamily Mimosoideae are considered invasive, whereas only 22 of 2250 Caesalpinioideae species are invasive, a rate is 5.5 times higher (Table 1). The Australian acacias have the highest rate of invasiveness of any large lineage of the legume family with 2.16% (22 of 1020) of the species known to be invasive. This percentage of invasive species is higher than other large woody plant families such as the Diptocarpaceae (0.3%) and the Fagaceae (0.7%). The percentage of invasive species in Acacia is higher than in the Myrtaceae, but much lower than in Pinus (12%), both of which have been planted worldwide (Richardson & Rejmánek, 2004).

There does not appear to be a high correlation between a legume species being used as a food and its invasiveness. The subfamily Papilinoideae contains most of the human cultivated food species including soybean, pea, cowpea and dry beans but has fewer invasive species than the less species-rich subfamily Mimosoideae. The Mimosoideae has more woody perennial species than the Papilinoideae, and perhaps it is this life history that drives the higher rate of invasiveness. These data suggest that there are particular affinities towards invasiveness in some plant lineages but fewer in others.

Other than the Australian Acacia species that are the main subject of this special issue, there are several other weedy mimosoid legumes with large invasive ranges. Several genera with invasive species cluster together near the base of the phylogenetic tree depicted in Box 1. These include Prosopis spp. (mesquite), Acacia (Vachellia) nilotica (prickly acacia), Leucaena leucocephala and Mimosa pigra. All these species are highly invasive and can be found in Africa, the Americas and the Asian Pacific region (Richardson & Rejmánek, 2011).

In Australia, the Commonwealth Government has identified the 20 worst weeds: the Weeds of National Significance (WONS) list (http://www.weeds.org.au/natsig.htm). These were determined based on invasiveness, impact, potential for spread and other social and environmental impacts. Of these 20 species, five are legumes, including three mimosoid legumes mesquite, prickly acacia and M. pigra.

The level of invasiveness of different Australian acacias seems to have more to do with human-mediated events than with biological features of the species (Carruthers et al., 2011; Gibson et al., 2011; Griffin et al., 2011). However the woody, arborescent habit of Mimosoid legumes appears to makes it more receptive to invasiveness. The widespread use of Mimosoid legumes as forage species and in various types of forestry and agroforestry programmes worldwide has radically enhanced their invasiveness potential (Griffin et al., 2011; Richardson et al., 2011).

Australian Acacias and invasive relationships

This study has built on previous phylogenetic analyses of Acacia s.s. by increasing taxon sampling and particularly by increasing the amount of DNA sequence data sampled to provide a more comprehensive phylogeny with greater phylogenetic resolution than previously available. The overall aim and focus of this study was to place in a broad phylogenetic context some of the known invasive species of Acacia and to provide some general insights into the evolution of invasiveness in the Acacia s.s. clade. The current study also provides some insights that should be useful for predicting future invasions in the group.

Overall, an important discovery from the phylogenetic analysis in the current study is that invasive taxa do not form a single clade. Rather, invasiveness is spread across the phylogeny of Acacia. However, given this, there are some clades in which several invasive taxa occur; probably, the most notable of these is the subclade that contains A. melanoxylon (Fig. 2) which contains six species: A. melanoxylon, A. implexa, A. verticillata, A. genistifolia, A. baeuerlenii and A. elongata. The first three of these are known to be invasive (Richardson & Rejmánek, 2011). A. genistifolia, currently not known as invasive in its introduced range, may well become invasive and should be carefully monitored. The range of the species in this grouping is in southeastern Australia (Fig. 4). In all cases, the distribution of the invasive species is much larger than the non-invasive sister species (see Hui et al., 2011).

Figure 4.

 Distribution map of selected species of the melanoxylon clade (see Fig. 2). Data derived from an edited version of the Australian Virtual Herbarium Council of Heads of Australasian Herbaria (2010).

Another group of taxa found in the aneura subclade of the melanoxylon clade contains A. holosericea and A. crassicarpa. This clade is also the likely place for two other invasive species that were not sampled: A. mangium and A. auriculiformis. All are northern Australian species and have been used, with the exception of A. holosericea, in South East Asian forestry projects (Griffin et al., 2011). Interestingly, as in the melanoxylon clade, the ecological tolerances and distribution of A. crassicarpa are broader than its closest relatives, A. peregrina and A. midgleyi (McDonald & Maslin, 2000).

Furthermore, the mearnsii clade (Fig. 3), which while being the target of increased taxon sampling for the current analysis, also has a large number of invasive species within it: A. dealbata, A. baileyana and A. decurrens. These species group with A. cardiopylla, A. silvestris, A. pubescens and A. spectablis. With the exception of A. spectablis, the invasive species have larger natural ranges.

The species range distribution should be interpreted with caution as the Australian ranges shown may include range expansion because of the species’ invasiveness (Hui et al., 2011). In general, the native distribution of invasive species is smaller than the current distribution in Australia. For example, the native distribution A. baileyana is restricted to a small area of NSW, but it is naturalized in much of SE Australia (Orchard et al., 2001). However, for most species, the circumscription of native and naturalized boundaries is unclear. Therefore, when compared to their sister species, it appears that the characteristics that allow a species to be invasive may also have an effect in the native range distribution.

The phylogenetic clustering of invasive species may have less to do with them possessing traits associated with invasiveness per se than with them having traits that make them more important in forestry and other industries that gave the species a foothold in many areas of the world.

Additionally, the sampling in this study is not random. With over 1000 species in the genus, it is unlikely that we have sampled all the sister species of the known invasive species. The ‘clustering’ of invasive species may therefore be an artefact of the limited species sampled. This can only be overcome by highly intensive phylogenetic sampling.

Gallagher et al. (2011) investigated difference in several functional traits between invasive and non-invasive Acacia species. They found invasive species to be taller, more prone to seed dispersal by vertebrates, have a larger native range including adapted to a broader range of annual precipitation than non-invasive Acacia species. No differences were found in seed mass, specific leaf area, relative growth rate and genome size (Gallagher et al., 2011). Data are needed for other functional traits, and full testing of these hypotheses will require a fully sampled and more resolved phylogeny.

Plant morphological characters, especially leaf and inflorescences traits, have been used to classify Acacia species into sections (Pedley, 1978). These groupings have allowed convenient discussion of the variation within Acacia but are not considered to be natural groups (Maslin et al., 2003). The most important taxonomic character is leaf type. Two sections contain only taxa with bipinnate leaves. The other sections are phyllodinous and were divided based on the number of prominent nerves in the phyllode. There is no correlation of leaf type with invasiveness. Some invasive species have bipinnate leaves, while others have phyllodes with either single or multiple nerves.

The major taxonomically important characters of the inflorescence are their shape and arrangement. The inflorescences are either globose or spicate and can be arranged in racemes or along the stem. Again there is no correlation of invasiveness and inflorescence form.

In conclusion, the legume subfamily Mimosoideae and in particular the Australian species of Acacia have a high percentage of invasive species compared to other legumes. It is becoming clear that invasiveness is closely associated with human-mediated introduction and dissemination, so it is very likely that the percentage of invasive Acacia species will rise in the future. There is some evidence of phylogenetic clumping of invasive species of Acacia in the limited sampling presented here with invasive species tending to have a larger native distribution than their non-invasive sister species. While no major taxonomic character is shared among the invasive species, this phylogenetic framework provides a structure for further testing the evolution of traits associated with invasiveness.

Table Box 2.   Tribe Ingeae: phylogenetic overview and weedy taxa
Tribe Ingeae is pantropical with 36 genera and c. 1000 species. It is differentiated from tribe Acacieae by one character: fusion of stamens into a tube in Ingeae and free stamens in Acacieae (Elias, 1981; Nielsen, 1981). Tribe Ingeae has repeatedly been shown to be paraphyletic, with Acacia sensu stricto nested within it (Brown et al., 2008; Dayanandan et al., 1997; Lavin et al., 2005; Luckow et al., 2003, 2000; Miller et al., 2003a, 2003b; Miller & Bayer, 2000, 2001, 2003; Wojciechowski et al., 2004). An intra-tribal classification was proposed by Barneby & Grimes (1996), which described five alliances based on developmental and macro-morphological characteristics. Some modifications have since been suggested (see Brown, 2008) and several alliances have been disputed by a molecular phylogenetic analysis (Brown et al., 2008). Phylogenetic relationships within tribe Ingeae are not well understood with many terminal clades resolved but the deeper relationships within the tribe not yet determined (Figure a).
Like many acacias, numerous Ingeae taxa are significant weeds world-wide, including species of Albizia Durazz., Lysiloma Benth., Paraserianthes sensu lato I.C. Nielsen, Pithecellobium Martius and Samanea Merr. The placement of these weedy taxa, if known, are scattered across the phylogeny (see Figure a). However, several weedy taxa —Albizia lebbeck (L.) Benth, Albizia saponaria Blume ex. Miq. and Samanea saman (Jacq.) Merr. — are united in the Samanea group of Brown et al. (2008); Figure a).
Relationships of some taxa are well supported, for example Paraserianthes and Pithecellobium. Pithecellobium dulce is related to Ebenopsis, Havardia, Sphinga and Painteria, in the Pithecellobium-alliance (Barneby & Grimes, 1996; Brown et al., 2008). Paraserianthes sensu lato is the closest relative to Acacia sensu stricto. It includes four species, two of which are widely planted and invasive taxa: Paraserianthes lophantha and Paraserianthes falcataria (=Falcataria moluccana).
Relationships of other weedy Ingeae taxa, however, are not understood. For example, Lysiloma, which includes several weed species (L. acapulcense, L. bahamensis and L. latisiliqua). Lysiloma was placed in the Chloroleucon-alliance of Barneby & Grimes (1996), however, Lewis & Rico (2005) did not think it belonged there and left it unplaced within the alliances. Molecular phylogenies suggest that Lysiloma is monophyletic and related to Hesperalbizia, of the Samanea-alliance (Barneby & Grimes, 1996), but these studies include less than a quarter of known species of Lysiloma (Luckow et al., 2003; Miller et al., 2003b; Brown et al., 2008).
Limited data is available on the intraspecific variation of weedy Ingeae, e.g. A. lebbeck (Aparajita & Rout, 2009). However, work is underway on some taxa (e.g. P. lophantha).
inline image
(a) Summary molecular phylogeny of Tribe Ingeae based on nrDNA sequences of the ITS and ETS (Brown et al., 2008).

Acknowledgements

We acknowledge financial support from the Oppenheimer Memorial Trust and Stellenbosch University towards the attendance of the October 2010 Acacia workshop in Stellenbosch by J.T.M. and D.J.M. D.M.R. acknowledges support from the Working for Water Programme and the DST-NRF Centre of Excellence for Invasion Biology through their collaborative research project on ‘Research for Integrated Management of Invasive Alien Species’. J.T.M. acknowledges the Hermon Slade Foundation and the Taxonomy Research and Information Network (TRIN) which is funded by the Australian Commonwealth Environment Research Facilities program for funding the phylogenetic research.

Biosketches

Joseph T. Miller is a molecular systematist and focuses his research on the plant genus Acacia. Daniel J. Murphy and Gillian K. Brown are also molecular systematists with major interests in the evolution and classification of Acacia and other Mimosoid legumes. Carlos E. González-Orozco is a GIS scientist with broad interests in the Australian flora. David M. Richardson is an invasion ecologist with an interest in legumes as invasive species.

Author contributions: J.T.M., D.J.M. and D.M.R. conceived the ideas, J.T.M. and C.E.G.-O. collected the data, J.T.M., D.J.M. and G.K.B. analysed the data and J.T.M. led the writing.

Editor: Andrew Lowe

Appendix

Appendix 1 Material used in this study

TaxonHerbarium voucher specimen or collector number
Acacia abbreviata MaslinCANB 793276
Acacia acuminata Benth.Mt Annan BG 866885
Acacia adoxa PedleyANBG 8212874
Acacia adunca A.Cunn. ex Don.ANBG 8502778
Acacia alata R.Br.CANB 00579597
Acacia anceps DC.CANB 793283
Acacia aneura F.Muell. ex Benth.CANB 635377
Acacia aneura F.Muell. ex Benth.Clarke12a
Acacia argyrophylla Hook.CANB 793288
Acacia aspera Lindl.CANB 793290
Acacia aulacocarpa A.Cunn. ex Benth.ClarkeB
Acacia auriculiformis A.Cunn. ex Benth.ATSC 15688
Acacia baeuerlenii Maiden & R.T.BakerClarke22b
Acacia baileyana F.Muell.CANB 00693196
Acacia beckleri TindaleANBG 9707897
Acacia binervata DC.ATSC 16245
Acacia brachystachya Benth.CANB 793304
Acacia calamifolia Lodd.CANB 793310
Acacia calcicola Forde & IsingCANB 793311
Acacia cardiophylla A.Cunn. ex Benth.CANB 492118
Acacia cognata DominANBG 9101965
Acacia cognata DominCANB 615708
Acacia confluens Maiden & BlakelyCANB 793329
Acacia crassicarpa A.Cunn. ex Benth.ATSC 15698
Acacia cultriformis A.Cunn. ex G.DonCANB 793341
Acacia cupularis DominCANB 633912
Acacia cyclops A.Cunn. ex G.DonCANB 793345
Acacia dealbata LinkANBG 9101960
Acacia dealbata LinkCANB 738126.1
Acacia dealbata LinkClarke3d
Acacia deanei (R.T.Baker) M.B.Welch, Coombs & McGlynnClarke20d
Acacia decurrens Willd.CANB 793354
Acacia dempsteri F.Muell.ANBG 680141
Acacia doratoxylon A.Cunn.Clarke33d
Acacia dorothea MaidenClarke40d
Acacia elata A.Cunn. ex Benth.ANBG 632927
Acacia elongata Sieber ex DC.Clarke27e
Acacia epacantha (Maslin) MaslinCANB 623291
Acacia euthycarpa (J.M.Black) J.M.BlackCANB 793378
Acacia extensa Lindl.CANB 793382
Acacia falcata Willd.Clarke4f
Acacia fimbriata A.Cunn. ex G.DonClarke26f
Acacia flexifolia Benth.CANB 793390
Acacia flexifolia Benth.Clarke6f
Acacia floribunda (Vent.) Willd.ANBG 9611057
Acacia floribunda (Vent.) Willd.Clarke7fl
Acacia genistifolia LinkCANB 793395
Acacia gonoclada F.Muell.ATSC 14721
Acacia guinetii MaslinCANB 793406
Acacia hakeoides A.Cunn. ex Benth.CANB 793281
Acacia hakeoides A.Cunn. ex Benth.Clarke19h
Acacia hakeoides A.Cunn. ex Benth.Mt Annan BG 842668
Acacia hakeoides A.Cunn. ex Benth.CANB 793407
Acacia hammondii MaidenCANB 793410
Acacia hemiteles Benth.CANB 633963
Acacia heteroclita Meisn.CANB 793415
Acacia holosericea A.Cunn. ex G.DonATSC 15669
Acacia howittii F.Muell.CANB 793419
Acacia implexa Benth.Clarke11i
Acacia irrorata Sieber ex Spreng.CANB 793423
Acacia ixiophylla Benth.CANB 793426
Acacia jonesii F.Muell. & MaidenMt Annan BG 20051433
Acacia kempeana F.Muell.CANB 793435
Acacia lasiocalyx C.R.P.AndrewsCANB 793438
Acacia leioderma MaslinCANB 793443
Acacia ligulata A.Cunn. ex Benth.ANBG 8210071
Acacia lineata A.Cunn. ex G.DonMt Annan BG 842542
Acacia linifolia (Vent.) Willd.ANBG 9409682
Acacia longifolia (Andrews) Willd.JN782
Acacia longissima Hort. ex H.L.Wendl.CANB 793457
Acacia mearnsii De Wild.ANBG 12
Acacia mearnsii De Wild.CANB 793467
Acacia mearnsii De Wild.Clarke37m
Acacia meisneri Lehm. ex Meisn.CANB 793468
Acacia melanoxylon R.Br.Mt Annan BG 860538
Acacia montana Benth.Clarke32m
Acacia mucronata Willd. ex H.L.Wendl.CANB 615743
Acacia muelleriana Maiden & R.T.BakerCANB 634004
Acacia multispicata Benth.ANBG 9710019
Acacia murrayana F.Muell. ex Benth.CANB 793477
Acacia neriifolia A.Cunn. ex Benth.Clarke8n
Acacia oswaldii F.Muell.CANB 793495
Acacia pachyacra Maiden & BlakelyMELU- SRA 239
Acacia parvipinnula TindaleMELU- SRA 31
Acacia penninervis Sieber ex DC.CANB 793506
Acacia pentadenia Lindl.CANB 793507
Acacia perryi PedleyCANB 793511
Acacia podalyriifolia A.Cunn. ex G.DonANBG 9406554
Acacia pravifolia F.Muell.Clarke39p
Acacia pravissima F.Muell. ex Benth.CANB 793515
Acacia prominens A.Cunn. ex G.DonMt Annan BG 981404
Acacia pruinosa A.Cunn. ex Benth.CANB 793518
Acacia pubescens (Vent.) R.Br.MEL 2111926
Acacia pycnantha Benth.CANB 793526
Acacia pyrifolia DC.CANB 793527
Acacia ramulosa W.Fitzg.CANB 793528
Acacia retinodes Schltdl.CANB 587946
Acacia rigens A.Cunn. ex G.DonCANB 634045
Acacia rostellifera Benth.CANB 793537
Acacia saliciformis TindaleMt Annan BG 884112
Acacia saligna (Labill.) H.L.Wendl.CANB 634053
Acacia saligna (Labill.) H.L.Wendl.CANB 793541
Acacia schinoides Benth.CANB 793542
Acacia silvestris TindaleCANB 793549
Acacia spectabilis A.Cunn. ex Benth.CANB 793553
Acacia stenophylla A.Cunn. ex Benth.CANB 793555
Acacia stigmatophylla A.Cunn. ex Benth.CANB 793556
Acacia suaveolens (Sm.) Willd.ANBG 643849
Acacia subrigida MaslinMELU- SRA 106
Acacia subulata Bonpl.ANBG 657697
Acacia triptera Benth.Clarke18t
Acacia triquetra Benth.CANB 793573
Acacia venulosa Benth.ANBG 9705373
Acacia verniciflua A.Cunn.Mt Annan BG 13007
Acacia verticillata (L’Her.) Willd.CANB 793581
Acacia vestita Ker Gawl.CANB 793583
Acacia victoriae Benth.AD 99835210 s51
Acacia viscidula Benth.Clarke1v
Acacia wattsiana F.Muell. ex Benth.CANB 793588
Pararchidendron pruinosum (Benth.) I.C.NielsenANBG 820099
Paraserianthes lophantha (Willd.) I.C.NielsenMEL 2057862

Ancillary