Introduction

In the south-west botanical region of Western Australia, isolation, climatic fluctuations and poor soils have contributed to a rich and diverse flora composed of many endemic and restricted species (Hopper, 1979). Approximately 8000 species have been identified to date, 75% of which are endemic (Hopper, 1992). Associated with this relatively high proportion of rare and geographically restricted species is a large number of taxa with historically fragmented or disjunct distributions (Hopper et al, 1996). Such range fragmentation would be expected to lead to strong genetic structure within species (Moran and Hopper, 1983). This is supported by studies on taxa from several different genera which suggest that unusually high levels of population differentiation are a feature of this flora (Coates, 2000).

Many genetic studies of Western Australian flora have concentrated on species which are rare or have highly restricted distributions. These studies have provided valuable insights into localised stochastic processes and life history characteristics and the importance these factors may have in the persistence of small populations in more recently fragmented landscapes. Genetic processes associated with rarity in plants have received significant attention over the last decade (see Karron et al, 1988; Hamrick and Godt, 1989; Barrett and Kohn, 1991; Ellstrand and Elam, 1993; Gitzendanner and Soltis, 2000). Theoretically, rare plants might be expected to show low levels of genetic variation both at the species and population levels due to selection under a narrow range of environmental conditions, and genetic drift and inbreeding in small isolated populations (Barrett and Kohn, 1991; Ellstrand and Elam, 1993). Low genetic variation within rare species may also be due to founder events associated with recent speciation (see Young and Brown, 1996). Interestingly, while some rare taxa in south-west Australia conform with theoretically expected low levels of genetic variation (Moran and Hopper, 1983, 1987) many others do not (Moran and Hopper, 1987; Coates, 1988, 1992; Tischler, 1998), maintaining levels which are consistently higher than those exhibited by rare plants on a worldwide basis (Hamrick and Godt, 1989; Hamrick et al, 1992; Young and Brown, 1996).

Similar findings have been documented elsewhere (see Young and Brown, 1996) and are not necessarily unexpected given the broad range of life histories and evolutionary patterns found in the rare plants (Karron, 1987). An alternative approach to a better understanding of rarity, small population size and restricted geographic range in plants has been proposed by Karron et al (1988) and others (see Gitzendanner and Soltis, 2000). Comparisons of the patterns of genetic variation in closely related groups or between pairs of rare and widespread or common species should at least reduce the confounding effects of life history and evolutionary factors (Young and Brown, 1996).

In contrast to rare or restricted taxa, widespread species are predicted to exhibit higher levels of genetic diversity although both classes distribute this diversity in a similar manner (Hamrick and Godt, 1989). The large continuous populations of widespread taxa are theoretically less susceptible to genetic drift and hence loss of diversity, in contrast to the small and ecologically limited populations characteristic of many endemic species (Hamrick and Godt, 1989). As more studies have been completed, however, it has become evident that a range of values exist within each geographical class (Gitzendanner and Soltis, 2000).

While the evolutionary patterns of genetic variation observed within rare and geographically restricted south-west Australian taxa often reflect that expected of extremely fragmented and disjunct populations, patterns associated with widespread species remain largely unexplored. Widespread species tend to have more continuous population systems in a broader range of environments. One possibility is that these species might be expected to show less genetic structure due to increased influences of gene flow even though they cover a much greater geographic range. Alternatively, as pointed out by Godt and Hamrick (1999), widespread species might be expected to exchange fewer genes across their geographic range and have overall higher levels of population divergence than geographically restricted species.

The Acacia acuminata Benth. (Mimosaceae) complex is a group of woody shrubs and small trees comprised of three formally described taxa (A. oldfieldii F.Muell., A. acuminata Benth. ssp. acuminata and A. acuminata ssp. burkittii F.Muell. ex Benth. Kodela and Tindale), two informal taxa (‘small seed’ and ‘narrow phyllode’), and three variants of uncertain status (‘variant 1’, ‘variant 2’ and ‘variant 1/2’) (Maslin et al, 1999; Maslin, 2001). In this analysis the as yet undescribed ‘small seed’ and ‘narrow phyllode’ are treated as distinct taxa rather than subgroups within the A. acuminata complex as has been done elsewhere (Maslin et al, 1999; M Byrne, unpublished). The complex is distributed across much of the south-west land division of Western Australia and extending eastwards through the arid zone into central Australia (Figure 1). Both ssp. acuminata and ssp. burkittii have relatively large geographical ranges (Figure 1). The morphologically distinct ‘small seed’ taxon is localised, while the ‘narrow phyllode’ taxon is widely distributed between ssp. acuminata and ssp. burkittii. The three variants which exhibit morphological affinities with both ssp. acuminata and ssp. burkittii are restricted to the Geraldton region (Maslin et al, 1999) while the rare and extremely geographically restricted A. oldfieldii is the fifth taxon within the complex.

Figure 1
figure 1

Distribution of taxa and study sites within the A. acuminata complex. Refer to Figure 1 for site numbers.

Full size image

The taxa within the A. acuminata complex exhibit varying degrees of geographic distribution providing a unique opportunity to compare population genetic patterns within and among the populations and taxa. Levels of genetic diversity within populations were also compared between the geographically restricted and more widespread taxa to assess the genetic implications of rarity in this complex, and whether the geographically restricted taxa were genetically depauperate. In addition, we sought to determine relationships between the taxa and how they may contribute to a better understanding of the complex evolutionary processes characteristic of the flora in this region.

Materials and methods

Sampling strategy

Seed was collected from populations of each taxon across the known distributions (Figure 1). Within each population several legumes were randomly collected from individual plants (10) and kept separate. Seed was later extracted and maintained as individual collections from which 10 seeds were randomly selected. Seed was surface sterilised with 1% sodium hypochlorite, rinsed three times, covered with boiling water and left to cool to room temperature. The following day any unimbibed seed were nicked to stimulate germination. Seedlings were frozen at −72°C until processed.

Isozyme electrophoresis

Seedlings (1–5) from individuals within each population were ground in extraction buffer (Broadhurst et al, 1999) and centrifuged for 5 mins. Allozymes were run using the cellulose acetate system (Helena Laboratories, Beaumont, TX, USA) and stained according to Richardson et al (1986) with modifications as described by Coates (1988) with the exception that 4 ml agar were used with all stain ingredients proportionally increased. Running times varied according to each enzyme system. Twelve enzyme systems were assayed: aspartate aminotransferase (Aat, EC2.6.1.1), esterase (Est, EC3.1.1.-), glucose-6-phosphate dehydrogenase (G-6-pdh, EC1.1.1.49), glutamate dehydrogenase (Gdh, EC1.4.1.2), glucose-6-phosphate isomerase (Gpi, EC5.3.1.9), isocitrate dehydrogenase (Idh, EC1.1.141), malate dehydrogenase (Mdh, EC1.1.1.37), malate dehydrogenase (oxaloacetate-decarboxylating) (Me, EC1.1.1.40), menandione reductase (Mnr, EC1.6.99.-), phosphoglucomutase (Pgm, EC5.4.2.2), phosphogluconate dehydrogenase (Pgdh, EC1.1.1.44), shikimate de- hydrogenase (Skd, EC 1.1.1.25). Sixteen zones of activity were scored and each zone was assumed to represent a single locus.

Genetic analyses

Single locus diversity measures including the average number of alleles per polymorphic locus (AP), percentage of polymorphic loci per population (P, criterion 99%), observed heterozygosity (Ho) and gene diversity (expected panmictic heterozygosity) (He) were estimated using POPGENE (Yeh and Boyle, 1997). Differences in genetic variation between species at the population level were assessed using nested analysis of variance for allelic richness and gene diversity, t-test for polymorphism and split-plot analysis of variance for heterozygosity with loci treated as split-plot treatments (Weir, 1990) using MINITAB Release 12. Fixation indices (FIS) (Wright, 1978) were estimated to examine population deviation from random mating using POPGENE (Yeh and Boyle, 1997). Total genetic diversity (HT) and its distribution within (HS) and between (DST) populations and the proportion of interpopulational differentiation (GST) were calculated using GENESTAT (Whitkus, 1985).

An UPGMA phenogram based on populations representing the five recognised taxa, A. oldfieldii, ssp. acuminata, ssp. burkittii, ‘small seed’ and ‘narrow phyllode’, was constructed using Nei's genetic distance (D) and standard error bars calculated with Ritland's Genetic Distance and Clustering (GD) program (Ritland, 1989). A second UPGMA was constructed using populations of the five taxa and the Geraldton variants to investigate whether any relationship(s) between these variants and the taxa that was not evident from morphological studies were present.

Results

No monomorphic loci were observed within the complex although the alternate allele(s) for loci Gdh-1, Mnr-1 and Pgdh-1 was either rare or at very low frequencies. While no fixed differences between species were observed, the most frequent allele for Aat-2, Aat-3, Gdh-1, Idh-1, Mdh-2 and Pgdh-1 in A. oldfieldii was alternate to that in the other taxa.

Variation within and among the taxa

Genetic variation within populations and population means for allelic richness, polymorphism, gene diversity and heterozygosity for all taxa and Geraldton variants are shown in Table 1. Within populations of the most widespread taxon, ssp. burkittii, AP (average number of alleles per polymorphic locus) ranged from 2.4 to 2.6, P (percentage polymorphic loci) from 81.3% to 93.8%, He (gene diversity) from 0.275 to 0.313 and Ho (hetero- zygosity) from 0.136 to 0.194. The FIS ranged from 0.263 to 0.414. In the other widespread taxon, ssp. acuminata, AP varied from 1.9 to 2.6, P from 56.3% to 87.5%, He from 0.187 to 0.287 and Ho from 0.099 to 0.194. Significant differences between these two taxa were noted for both AP (p = 0.001) and He (p = 0.008).

Table 1 Single locus diversity measures based on 16 loci for populations in the A. acuminata complex. See for site locations
Full size table

Diversity measures in populations of ‘narrow phyl- lode’, which is distributed between ssp. acuminata and ssp. burkittii, were similar to those of ssp. burkittii (AP, range 2.4–2.6; P, 75–87.5%; He, 0.263–0.330; Ho, 0.133–0.154). This taxon also had high FIS values ranging from 0.397–0.499. While no significant differences were observed for any measure between ‘narrow phyllode’ and ssp. burkittii, differences were noted between ‘narrow phyllode’ and ssp. acuminata for both AP (p = 0.007) and He (p = 0.011).

In the highly restricted A. oldfieldii populations, mean values for all measures of variation were clearly lower than population means for all other taxa and Geraldton variants. The mean AP was 2, P 59.4%, He 0.168, Ho 0.078, while FIS was moderate at 0.332. All measures were significantly different between this taxon and ssp. burkittii (AP, p = 0.000; He, p = 0.000; Ho, p = 0.000), ssp. acuminata (AP, p = 0.020; He, p = 0.002; Ho, p = 0.007) and ‘narrow phyllode’ (AP, p = 0.001; He, p = 0.000; Ho, p = 0.001). The localised ‘small seed’ exhibited similar mean values to those of ssp. acuminata (AP, 2.3; P, 71.9%; He, 0.245; Ho, 0.155). The FIS was the closest to zero of all the taxa but still exhibited homozygote excesses (0.220). This taxon did not differ significantly for any value with ssp. acuminata or ‘narrow phyllode’ but was different to ssp. burkittii for AP (p = 0.020) and to A. oldfieldii for Ho (p = 0.000) and He (p = 0.005).

Differences at the species level

Pooled single locus diversity measures also varied between the taxa. The measures of AP and P were highest in the three widespread taxa, ssp. acuminata, ssp. burkittii and ‘narrow phyllode’ (AP 2.9, 3.3 and 3.1; P 87.5%, 100% and 87.5%, respectively) and lowest in the restricted A. oldfieldii and localised ‘small seed’ (AP, 2.3 and 2.4; P, 75% and 81.3%, respectively). Values for He were lowest in A. oldfieldii (0.173) and in the other taxa ranged from 0.252 in ‘small seed’ to 0.319 in ‘narrow phyllode’. A similar pattern was observed for Ho, while fixation indices were extremely high in ‘narrow phyllode’ (0.495) and at similar levels in all other taxa (0.360–0.388).

The taxa also exhibited differences in genetic diversity indices (Table 2). Total genetic diversity was lowest in A. oldfieldii (HT, 0.178), highest in ‘narrow phyllode’ (0.332) and ssp. burkittii (0.318) with ssp. acuminata and ‘small seed’ being mid-range (0.266 and 0.259, respectively).

Table 2 Gene diversity statistics over all loci unbiased for sample size and population number for Acacia acuminata taxa
Full size table

Differences within and among the Geraldton variants

The Geraldton variant populations exhibited a range of single locus diversity measures both within and among the different variants. The mean AP ranged from 2.3 in the ‘variant 1/2’ population to 2.6 in ‘variant 1’. The mean P was 81.3% in ‘variant 1’ and ‘variant 1/2’ and 71.9% in ‘variant 2’ while He was similar in ‘variant 1’ and ‘variant 1/2’ (0.265 and 0.263, respectively) and marginally lower in ‘variant 2’ (0.257). The mean Ho was highest in ‘variant 1’ (0.164) and lowest in ‘variant 1/2’ (0.111). Fixation indices were highest in ‘variant 1/2’ (FIS, 0.465), lower in ‘variant 2’ (0.406) and considerably lower again in ‘variant 1’ (0.296).

Genetic divergence among populations and phylogenetic relationships

All taxa strongly partitioned total genetic diversity within populations (HS) rather than among (DST). Consequently, the levels of interpopulational differentiation in the taxa (GST) were relatively low (Table 2). The proportion of interpopulation differentiation is highest in the three most widespread taxa, ssp. acuminata (GST, 10.8%), ssp. burkittii (GST, 9.8%) and ‘narrow phyllode’ (GST, 12.5%), and lowest in the two geographically restricted taxa, A. oldfieldii (GST, 6.9%) and ‘small seed’ (GST, 5.3%).

Excluding the Geraldton variants, populations clustered according to systematic boundaries (Figure 2). Acacia oldfieldii was clearly a discrete taxon, significantly distant from the remaining taxa (D, 0.85). The ‘small seed’ populations also grouped together as a significant cluster. While the remaining populations were not strongly differentiated, these did group within respective taxon boundaries. All ssp. burkittii populations were clustered together but as two discrete groups. Populations of ssp. acuminata were clustered together with one population of ‘narrow phyllode’ (10) while the remaining populations of this latter taxon (9 and 11) clustered separately.

Figure 2
figure 2

Phylogenetic relationships between populations within the A. acuminata complex excluding Geraldton variants. Branches are significant if error bar is less than half the branch length (Ritland, 1989).

Full size image

Including the Geraldton variants in the UPGMA analysis confused some of the taxonomic alliances (Figure 3). Acacia oldfieldii was again significantly distant to all other taxa while ‘small seed’ and the ‘narrow phyllode’ populations 9 and 11 again remained as discrete groups. However, the ssp. burkittii populations 24 and 25 were separated from the remaining three populations and allied with a ‘narrow phyllode’ population (10) previously included within ssp. acuminata, and two ‘variant 1’ populations (15 and 17). The other ssp. burkittii populations were associated with another ‘variant 1’ population (16). The ssp. acuminata population 3 was removed from all other ssp. acuminata populations which were associated with the remaining Geraldton variant populations.

Figure 3
figure 3

Phylogenetic relationships between populations within the A. acuminata complex including Geraldton variants. Branches are significant if error bar is less than half the branch length (Ritland, 1989).

Full size image

Discussion

Apart from A. oldfieldii, levels of genetic diversity observed within both the taxa and Geraldton variants of the A. acuminata complex were considerably higher than that expected of long-lived woody angiosperms (He 0.183, s.e. 0.011; Hamrick et al, 1992). Most taxa and variants exhibited levels of gene diversity consistent with that expected of widespread species (He 0.257, s.e. 0.039; Hamrick et al, 1992) with two exceptions. Gene diversity in ssp. burkittii and ‘narrow phyllode’ were at considerably higher levels than that observed in other similarly distributed species (He 0.311 and 0.319, respectively).

As expected of a rare and highly restricted species, gene diversity in A. oldfieldii was considerably lower (He 0.173) than that observed in any of its closely related relatives. These levels of diversity were still more than twice that exhibited by other endemic taxa on a worldwide basis (He 0.078, s.e. 0.016) and were more similar to narrowly distributed or regional species (Hamrick et al, 1992). While A. oldfieldii does not conform with the average expected for endemic taxa, recent studies have shown that some rare and geographically restricted species not only have higher than expected levels of genetic variation, but also have levels comparable to closely related widespread common congeners (Young and Brown, 1996; Gitzendanner and Soltis, 2000).

Examples of higher than expected genetic diversity have been observed in other restricted species of Acacia within Western Australia. Genetic diversity values in the rare A. sciophanes were higher than those for both ssp. acuminata and ssp. burkittii but significantly lower than estimates for its closely related widespread congener A. anfractuosa (Tischler, 1998). The sexually reproducing populations of the A. anomala exhibit similar A and P values (2.0 and 61.1%, respectively) to A. oldfieldii and a marginally higher gene diversity (He 0.209; Coates, 1988) while indices for the single large population of the highly localised Acacia sp. ‘Dandaragan’ are somewhat lower but still higher than expected (A, 1.55; P, 55%; He, 0.109) (Elliott et al, 2002). These studies indicate that rare and geographically restricted Acacia species can exhibit relatively high levels of genetic diversity although they are generally lower than their widespread and common close relatives.

Of the three widespread taxa within the A. acuminata complex, allelic richness and gene diversity were significantly higher in ssp. burkittii and ‘narrow phyllode’ than in ssp. acuminata. Acacia acuminata ssp. burkittii is very widely distributed across a harsh and arid environment of inland Australia. Populations residing in challenging environments can exhibit higher genetic diversity (Brown and Schoen, 1992) and the levels of genetic variability observed in this species may serve to buffer against various environmental stresses experienced across the large geographic range. Lower levels of genetic diversity within ssp. acuminata may reflect its distribution in a more ecologically favourable environment where the climate is generally more mesic.

The ‘narrow phyllode’ taxon is geographically distributed between ssp. acuminata and ssp. burkittii. It is considered to be most closely related to ssp. acuminata and both taxa can difficult to distinguish when their geographic ranges overlap (Maslin et al, 1999). This taxon also has morphological affinities with ssp. burkittii but can be readily differentiated by fruit structure (Maslin et al, 1999). In this study ssp. acuminata and ‘narrow phyl- lode’ were associated together rather than with ssp. burkittii suggesting that these are more closely related (Figure 2). It is possible, given its geographic distribution, that ‘narrow phyllode’ may be the result of extensive and ongoing hybridisation between ssp. acuminata and ssp. burkittii. Hybridisation events can create new species and/or increase genetic variability (Mayr, 1970; Richards, 1997). Although levels of population differentiation within the A. acuminata complex are relatively low, ‘narrow phyllode’ exhibits one of the highest levels of population differentiation (GST, 12.5%) within the complex. Extensive and ongoing hybridisation over large temporal and geographical scales might readily produce hybrid populations with a diverse array of genetic products. Such circumstances could result in higher than usual levels of population differentiation.

The similarity of diversity levels between the ‘small seed’ taxon and ssp. acuminata would suggest that although this taxon appears to have a limited distribution (Figure 1), it was/is as widespread as ssp. acuminata. To date, relatively few collections of ‘small seed’ exist, but it might also be distributed along the south coast of Western Australia (Maslin et al, 1999). Unfortunately much of the land between the study sites and the south coast is now extensively cleared and it may be impossible to determine the true range of this taxon.

Within the Geraldton variants there is no clear relationship between genetic variation and geographical distribution. All these variants have narrow geographic ranges but exhibit mean diversity indices which are higher than the restricted A. oldfieldii and in many instances are similar to or slightly higher than levels observed in the widespread taxa. Given that phylogenetic relationships within the complex are confused when the Geraldton variants are included in the analysis (Figure 3) and their morphological delimitation is unclear, some other explanation may be more appropriate.

Taxonomically, the Geraldton variants exhibit combinations of morphological attributes associated which each of the three widespread taxa (Maslin et al, 1999). As such these may simply represent the extremes of variation found on the edge of the species’ ranges (Maslin et al, 1999). The diversity indices do suggest that the Geraldton variants are not, or at least do not behave as, restricted taxa and are consistent with them being range extensions. An alternative possibility is, however, that like ‘narrow phyllode’, these variants are also of hybrid origin. Indeed, the Geraldton variants exhibit a similar degree of population differentiation (GST, 12.7%) to that of the ‘narrow phyllode’ taxon. The ‘variant 1’ may represent hybridisation between ‘narrow phyllode’ and ssp. burkittii, ‘variant 2’ may represent meetings of ssp. acuminata and ssp. burkittii while the origins of ‘variant 1/2’ remain unclear (Maslin et al, 1999). The apparent breakdown of taxonomic boundaries when these variants are included in the analysis does support the notion that these have hybrid origins.

The manner in which the Geraldton variants are distributed when all populations are included in the analysis further suggests that not only are these hybrid derivatives, but that these may actually represent a series of hybridisation events. Spatially and temporally separated meetings between the various taxa might be expected to produce hybrid derivatives with a diverse array of both genetic and morphological types and which present affinities with either or both parental taxa. Indeed, repeated crossing between hybrids and one parental type can produce a population resembling the recurrent parent, but varying in the direction of the other (Grant, 1981).

The Geraldton variants, ‘narrow phyllode’ and ‘small seed’ are distributed in a region known as the Transitional Rainfall Zone which is well known for high levels of speciation, endemism and allopatric replacement of congeneric species (Hopper, 1979). Unfortunately few genetic studies of flora within this region are known although a study of two subspecies of Geleznowia verrucosa (Rutaceae) has suggested that a series of ancient hybrid derivatives also exist within the same area as the purported Geraldton hybrids (Broadhurst et al, 1999). In addition, a chloroplast study of the A. acuminata complex has indicated that the region where the Geraldton variants occur is an evolutionary hotspot for this complex (Byrne et al, in review).

Unlike studies on a number of other species in the south-west Australian flora (Coates, 2000) there is no clear geographical population genetic structure in either the geographically restricted or widespread taxa in the A. acuminata complex. However, the widespread taxa, ssp. acuminata, ssp. burkittii and ‘narrow phyllode’ have moderately higher levels of genetic differentiation among populations than the geographically restricted A. oldfieldii and ‘small seed’. The lack of population genetic structure tends to confirm field observations that taxa in the complex generally occur as large relatively continuous population systems throughout their range, and that the moderately higher levels of among population differentiation in the more geographically widespread taxa are the result of reduced gene flow across their broader geographic range.

Although mating system studies were not carried out on the taxa in the A. acuminata complex, studies on other Acacia species indicate very high rates of outcrossing (Moran et al, 1989a; McGranahan et al, 1997). Large population size typical of taxa within this complex combined with high outcrossing rates may provide one explanation for the high levels of genetic diversity observed. However, average fixation index for all populations within taxa are positive and in many cases significantly greater than zero suggesting significant levels of inbreeding within populations. The fixation estimates for this study are based on seed progeny not maternal parents and while high levels of inbreeding may occur within populations, selection against inbred individuals may ensure these do not reach reproductive maturity. As shown in other Acacia species (McGranahan et al, 1997; Tischler, 1998) and in mating system studies on a number of other plant species there can be preferential survival of heterozygotes as plants mature from seedling to adult (Hamrick and Schnabel, 1985).

Apart from A. oldfieldii the relatively low level of phylogenetic separation between taxa in the complex indicates quite recent evolutionary divergence. The high degree of genetic differentiation between A. oldfieldii and the other A. acuminata taxa (D, 0.85) was somewhat unexpected given the clear morphological affinities that this species has with those taxa. Indeed, the major morphological features distinguishing A. oldfieldii are glabrous phyllodes rather than those with marginal hairs, and short pedunculate flower spikes (Maslin et al, 1999). This high degree of genetic differentiation was also reflected in the chloroplast study which placed the divergence of A. oldfieldii at approximately 2.5–3 My ago (M Byrne, unpublished).

The levels of genetic diversity within the A. acuminata complex taxa are generally much higher than expected based on geographical distribution and when compared with other woody plant species with similar life histories. Other south-west Australian Acacia species also show comparable levels of genetic diversity (Tischler, 1998; Elliott et al, 2002). These levels are noticeably higher than that observed in Acacia species from Eastern Australia (Moran et al, 1989a,b; Playford et al, 1993; Searle et al, 2000). A feature of the flora in the south-west region of Australia is its antiquity associated with prolonged geological stability and no obvious extinction episodes caused by recent glaciation. Taxa in the A. acuminata complex have probably persisted in relatively large continuous populations for extended periods of time. This may provide one explanation for the high levels of genetic diversity observed. In addition, hybridisation would appear to be an important component of recent evolution and may also be a contributing factor to the high levels of genetic diversity in some taxa. Hybridisation may also explain the unresolved patterns of differentiation evident in some parts of the geographical range of the complex.