Contrasting levels of connectivity and localised persistence characterise the latitudinal distribution of a wind-dispersed rainforest canopy tree

Contrasting levels of connectivity and localised persistencecharacterise the latitudinal distribution of a wind-dispersedrainforest canopy tree

Margaret M. Heslewood • Andrew J. Lowe •

Darren M. Crayn • Maurizio Rossetto

Received: 25 September 2013 / Accepted: 31 May 2014

� Springer International Publishing Switzerland 2014

Abstract Contrasting signals of genetic divergence due

to historic and contemporary gene flow were inferred for

Coachwood, Ceratopetalum apetalum (Cunoniaceae), a

wind-dispersed canopy tree endemic to eastern Australian

warm temperate rainforest. Analysis of nine nuclear

microsatellites across 22 localities revealed two clusters

between northern and southern regions and with vicariance

centred on the wide Hunter River Valley. Within popula-

tions diversity was high indicating a relatively high level of

pollen dispersal among populations. Genetic variation was

correlated to differences in regional biogeography and

ecology corresponding to IBRA regions, primary factors

being soil type and rainfall. Eleven haplotypes were iden-

tified by chloroplast microsatellite analysis from the same

22 localities. A lack of chloroplast diversity within sites

demonstrates limited gene flow via seed dispersal. Network

representation indicated regional sharing of haplotypes

indicative of multiple Pleistocene refugia as well as deep

divergences between regional elements of present popula-

tions. Chloroplast differentiation between sites in the upper

and lower sections of the northern population is reflective

of historic vicariance at the Clarence River Corridor. There

was no simple vicariance explanation for the distribution of

the divergent southern chlorotype, but its distribution may

be explained by the effects of drift from a larger initial

gene pool. Both the Hunter and Clarence River Valleys

represent significant dry breaks within the species range,

consistent with this species being rainfall dependent rather

than cold-adapted.

Keywords Cunoniaceae � Microsatellites � Rainforest

refugia � Vicariance � Wind-dispersed tree

Introduction

Over the last 35MYA, since the separation of the Austra-

lian continent from Antarctica (McLoughlin 2001), the

distribution of dominant mixed broadleaf/conifer forests

akin to modern rainforests has declined (Webb and Tracey

1981). Though many of the families typical of Gondwanan

plant assemblages (e.g. Casuarinaceae, Myrtaceae, Prote-

aceae) have adapted to dryer conditions, fossil evidence

(Hill et al. 1999; Kershaw et al. 2007) shows that broadleaf

lineages (e.g. Cunoniaceae, Elaeocarpaceae) became

restricted to wetter areas, primarily in the tropics and

subtropics. Despite the steady continental aridification and

their current archipelago-like distribution along the east

coast, present day Australian rainforests remain diverse

(Adam 1992; Baur 1957; Webb and Tracey 1981).

In contrast to the northern hemisphere, where glaciation

removed vegetation from large areas which were subse-

quently recolonised from refugia (Heuertz et al. 2006;

Hewitt 2004; Magri et al. 2006), current floristic assem-

blages in Australia are the result of successional changes

driven by rainfall and fire regimes within a continuously

M. M. Heslewood (&) � M. Rossetto

National Herbarium of New South Wales, Royal Botanic

Gardens and Domain Trust, Mrs Macquaries Rd, Sydney,

NSW 2000, Australia

e-mail: [email protected]

M. M. Heslewood � A. J. Lowe

Australian Centre for Evolutionary Biology and Biodiversity,

School of Earth and Environmental Sciences, University of

Adelaide, Adelaide, SA 5005, Australia

D. M. Crayn

Australian Tropical Herbarium, Sir Robert Norman Building

(E2), James Cook University Cairns Campus,

PO Box 6811, Cairns, QLD 4870, Australia

123

Genetica

DOI 10.1007/s10709-014-9771-8

vegetated landscape (Bowman 2000). The detection of

Eucalyptus charcoal originating from the last glaciation

within areas currently covered by tropical rainforests is

evidence of the cyclical contraction/expansion of vegeta-

tion assemblages that has characterised the Quaternary

(Hopkins et al. 1993). Similar cyclical changes in floristic

assemblages and fire frequency have also been described in

southern Australia (Black et al. 2006; Williams et al.

2006). An increasing number of studies are investigating

the impact of these cyclical changes on population and

phylogeographic patterns across the eastern Australian

mesic flora (e.g. Hilbert et al. 2007; Worth et al. 2009;

Mellick et al. 2011; Milner et al. 2012).

Phylogeographic studies are useful for highlighting the

genetic signatures of historical vegetation changes, identi-

fiable as repeated patterns of differentiation and diver-

gence. Interestingly, it is becoming increasingly apparent

that specific life history traits impact on between-popula-

tion dynamics, with certain functional characteristics hav-

ing potential to predict structure and identifiable landscape

genetic patterns. For instance, a range of studies within

fleshy-fruited Elaeocarpus species showed associative

patterns between genetic connectivity, fruit type and the

availability of suitable dispersal vectors (Rossetto et al.

2004b, 2007, 2008, 2009). In order to survive cyclical

disturbance patterns, rainforest species with limited dis-

persal potential often rely on the ability to resprout (Taylor

et al. 2005; Rossetto et al. 2004a; Rossetto and Kooyman

2005). Vegetative growth provides a mechanism for sur-

viving a range of disturbance factors, as well as a localized

competitive edge over seedlings (Bond and Midgley 2001).

Ecological traits that ensure local persistence are common

defence mechanisms against the disturbance events that

characterize most Australian ecosystems (Adams 1992;

Bowman 2000).

Ceratopetalum apetalum D.Don (Cunoniaceae) belongs

to a functional group of widely distributed rainforest trees

that rely on wind dispersal of seed (expected to be less

efficient than dispersal by small fleshy fruits) and can

resprout in response to disturbance events. Within Aus-

tralia the genus Ceratopetalum is restricted to eight species.

Two have broad distributions in warm temperate rainforest

and wetter sclerophyll forests to the east of the Great

Dividing Range in New South Wales (NSW) and southeast

Queensland, and the remaining six are narrow range en-

demics in tropical rainforests of far north Queensland. As

with the majority of the Australian flora, little to nothing is

known of the mating system of this species. Ceratopetalum

flowers are small, cream, sometimes perfumed and pre-

sumed till now to be insect-pollinated (Hoogland 1960,

1981; Rozefelds and Barnes 2002). Fruit are indehiscent

and moderate in size and although the sepals develop

coloration of red through purple, they are woody and

presumably unattractive to frugivores. However their

aerodynamic shape with 4–6 radiating sepals aids wind

dispersal. Fossilised fruit show that the genus had a much

wider distribution in the Eocene-Miocene, extending as far

as South Australia (Barnes and Hill 1999).

We expect that lineages with a long continental history,

but with different functional attributes, such as the closely

related families Cunoniaceae and Elaeocarpaceae, would

have been differentially impacted by the contraction/

expansion cycles of the Quaternary. As a dominant, wind-

dispersed rainforest species, C. apetalum represents dif-

ferent floristic elements from those previously studied, and

one which is potentially resilient to a range of short and

long term landscape-level dynamics in response to climate

change.

The main objective of this study was to evaluate the

structuring of genetic diversity in order to investigate the

response to climatic cycles in C. apetalum, and contrast it

with those of previously studied taxa. In particular we

addressed the following hypotheses:

1. Is seed-mediated gene flow lower in C. apetalum than

typically exhibited by fleshy fruited species?

2. Does genetic structure across the range of the species

correspond to major geographic barriers?

3. If C. apetalum responds to disturbance events through

resprouting rather than rapid recolonisation from local

sources, can we find evidence of multiple refugia,

particularly across areas that have historically endured

greater disturbance?

4. If we find evidence of multiple refugia, can we detect

the genetic signal of local re-expansion events follow-

ing the increase of habitat availability (or a reduction

in disturbance frequency) that potentially characterised

the current postglacial period?

Materials and methods

Study species and sampling strategy

Ceratopetalum apetalum D.Don, Coachwood, is a canopy

tree primarily found in warm temperate rainforest (also

known as simple notophyll vine forest, to simple notophyll

microphyll vine forest), in New South Wales and southeast

Queensland. At present C. apetalum is under no environ-

mental threats and needs no explicit conservation strategy.

The majority of local populations are contained within the

protection of the National Parks reserve system which is

preserving the geographic range and broad genetic diver-

sity of this taxon. At times a locally dominant species it

occupies a naturally fragmented habitat within broader

sclerophyllous vegetation, limited largely by competition

Genetica

123

to gullies on poor soil and in positions close to water

sources such as creeks and waterfalls where it is protected

from intense bushfires (Fig. 1). The indehiscent fruit of C.

apetalum contain single small seeds that readily germinate

after masting events, and persist as suppressed seedlings

ready to exploit opportunities provided by canopy gaps

(Adam 1992; Floyd 1990).

As its common name suggests, the timber of C. apeta-

lum is a prized commodity once widely used for building

coaches and furniture and many sites currently protected in

the National Park system were once logging areas or state

forests. Sampling was designed to represent localities

throughout most of the present distribution of C. apetalum

covering different vegetation types (warm temperate

rainforest through riparian areas in drier sclerophyll), and a

range of elevations both within and between sites, and

included a number of areas that were previously subjected

to logging. 332 individuals from 22 localities were sam-

pled, with individuals (primarily adult trees) collected at

least 50 m apart where possible in order to sample a broad

area of each site (Fig. 1; Table 1). Although coppicing

from the base of trees was observed at many southern sites,

there is no evidence for clonality in this species which

might have impacted on sampling design and diversity

estimates. There are no marked phenotypic differences

across the species range not directly related to environ-

mental conditions at sites that would have affected

sampling.

Fig. 1 Distribution of

populations for C. apetalum

based on nuclear microsatellite

data. STRUCTURE analysis

identified a two population

structure north and south of the

Hunter River. Pies represent

secondary population

assignments identified by

STRUCTURE at K = 7.

Population numbers correspond

to those listed in Table 2

Genetica

123

PCR amplification

All samples were analysed using nuclear microsatellite loci

to derive genetic estimates of current population structure

and connectivity. When examining longer term effects in

populations, such as historical gene flow, the confounding

effect of recombination can be addressed by the use of

haploid non-recombining genomes (Heuertz et al. 2004).

As a consequence, in order to obtain phylogeographic data

and in order to generate pollen/seed migration rate esti-

mates, we also analysed chloroplast microsatellites

(cpSSR). Being maternally inherited in most plants cpDNA

represents the seed dispersal history of a population as well

as revealing longer-term phylogeographic structure. We

sampled fewer individuals from all localities (Petit et al.

2005) using a total sample of 134 individuals (Table 1).

Genomic DNA was extracted using the Qiagen Tissue

Lyser and Qiagen DNeasy 96 Plant kit as per manufac-

turer’s protocols. DNA yields for this taxon were low

regardless of age of the leaf sampled or the length of time

and treatment since collection. After preliminary tests for

suitability, multilocus genotypes were generated from nine

nuclear microsatellites (Heslewood et al. 2009) and four

universal chloroplast microsatellites (Weising and Gardner

1999). Nuclear microsatellite regions were amplified using

multiplex PCRs with up to three non-overlapping loci co-

amplified following the method described in Heslewood

et al. (2009), and cpSSR markers were amplified using the

methods described by Weising and Gardner (1999) with

minor modifications to enable multiplexing during their

analysis. Microsatellites were detected on an ABI 3730

Capillary Sequencer (Applied Biosystems) at the Auto-

mated DNA Sequencing Facility, University of New South

Wales. Microsatellite profiles were examined in GeneM-

apper Version 3.7 (Applied Biosystems).

Null alleles were assessed with MicroChecker 2.2.3 (van

Oosterhout et al. 2006) which uses a Monte Carlo simu-

lation to generate expected homozygote and heterozygote

allele size frequency differences under HWE theory to

detect null alleles, stuttering and large allele dropout in the

Table 1 Population codes for Ceratopetalum apetalum collections used in the population genetic analyses arranged geographically S to N

Code N Population locality Latitude (�S) Longitude (�E) Altitude (m) Substrate

BC 9 Upper Brogers Creek, Budderoo NP, NSW 34.65806 150.70111 590 Sandstone

BF 10 Belmore Falls, Morton NP, NSW 34.64075 150.56042 560 Sandstone

Ro 10 Robertson NR, NSW 34.59175 150.59700 760 Basalt

KF 5 Kelly’s Falls, Royal NP, NSW 34.21364 150.98006 180 Sandstone

R 18 Royal NP, NSW 34.09439 151.07808 0-80 Sandstone

HT 10 Cave Creek, Hilltop, NSW 34.33547 150.49644 560 Sandstone

WF 19 Valley of the Waters Track, Wentworth Falls,

Blue Mountains NP, NSW

33.72667 150.36225 700–900 Sandstone

NG 15 Nellies Glen, Blue Mountains NP, NSW 33.70972 150.29083 800–900 Sandstone

Ev 15 Evans Lookout, Blue Mountains NP, NSW 33.65000 150.32500 900 Sandstone

LC 11 Lane Cove River NP, NSW 33.75253 151.09536 70 Sandstone

EB 10 Elouera Bushland Reserve, NSW 33.72892 151.05014 100–160 Sandstone

DC 39 Deep Creek NR, Narrabeen, NSW 33.70667 151.25889 17 Sandstone

Wo 10 Wollemi site 1, Wollemi NP, NSWa 33.12950 150.45776 660 Sandstone

BT 15 Lower Barrington Tops, NSW 32.21575 151.76161 420 Basalt

We 10 Carrai State Forest, NSW 31.03667 152.33690 1,010

D 24 Never Picnic Area, Dorrigo NP, NSW 30.36242 152.79758 750 Yellow clay

NB 16 Nymboi-Binderay NP, NSW 30.20292 152.74961 500–750 Granite-yellow soils

Wa 22 Coombadjha Rest Area, Washpool NP, NSW 29.47244 152.31733 850–1,000 Granite

WaN 13 Washpool NP (north), NSW 29.18669 152.42839 750

Ni 18 Nightcap Track, Nightcap NP, NSW 28.55731 153.34259 760 Rhyolite

NiB 23 Nightcap NP, NSW 28.56497 153.33636 800 Rhyolite

Sp 10 Springbrook NP, SEQ 28.22255 153.28842 680–800

Vouchers for each population lodged with the National Herbarium of New South Wales (NSW)

NP National Park, NR Nature Reserve, NSW New South Wales, SEQ Southeast Queenslanda Locality protected due to co-occurrence of highly endangered species so given LL is for midpoint of NP

Genetica

123

dataset. Data from several loci suggested the presence of

null alleles at some sites, and the software invoked stut-

tering as a possible explanation for alleles at one locus

primarily differing by single repeats. Several individuals

were repeated multiple times (including three internal

controls repeated in each run) and in different multiplex

combinations to ensure that scoring was accurate and

alleles were not an artefact of cross-locus amplification. Up

to 48 % of samples were repeated at least once per locus

from initial trials and optimisation of primers and DNA

quantity through to final runs. Once conditions were opti-

mised the error rate was 1.6 %. No individuals failed to

amplify and no errors were recorded for the cpSSR

genotyping.

Population genetic diversity statistics

Average number of alleles (Na), expected heterozygosity

(He), observed heterozygosity (Ho), number of private

alleles were calculated for each sampled site and averaged

over all sites in GenAlEx 6.4 (Peakall and Smouse 2006).

Fstat 2.9.3.2 (Goudet 1995) was used to calculate fixation

indices (Fis and Fst), and allelic richness, Rs, which

estimates average number of alleles per locus corrected

for different sized sample populations. As the high levels

of heterozygosity measured at microsatellite loci can

reputedly lead to Fst greatly underestimating differentia-

tion amongst populations, we also used Jost’s (2008)

distance-based measure D to estimate true population

differentiation. Calculations of Dest were performed using

the online SMOGD portal (Crawford 2010). Hardy–

Weinberg equilibrium (HWE) and linkage disequilibrium

(LD) values were calculated in GENEPOP 4.0.7 (Rousset

2008) using the exact test with 500 batches and 5,000

iterations per batch. The Markov chain parameters applied

for all global Hardy–Weinberg tests were dememorization

10,000 with 500 batches and 5,000 iterations per batch.

Bonferroni corrections were applied as per Rice (1989).

The only significant pairwise locus combination summed

across sites significant for linkage disequilibrium was

Ca050BGT-Ca110BGT (p \ 0.001). However this was

the result of detected LD in only five of 22 populations

for this locus pair. All sites showing LD were smaller

populations from the Sydney region and could be a

consequence of drift. The data was analysed with and

without either of these two loci, and as no appreciable

differences were noted all loci were included in the

analyses presented.

Population structure

Several estimates of the partitioning of genetic variation

were obtained. The model-based clustering programme

STRUCTURE 2.3.3 detects genetic structure across a

sampled species range and infers the number of popula-

tions, K, present (Pritchard et al. 2000) by probabilisti-

cally assigning individuals to populations using

simulations based on allele frequencies at each locus,

assuming HWE and LE within populations. We used a

parameter set with 50,000 burn-in reps followed by 106

MCMC reps, used all standard parameters under the

ancestry model assuming admixture, with inferred alpha

from an initial value of 1.0 uniform for all populations.

The frequency model used assumed variable Fst for

subpopulations, with allele frequencies among populations

correlated. Values of K from 1 to 22 (the number of

sampled sites) were each simulated under five iterations

from a random starting seed. Posterior probabilities of

different K values were tested to find the value with the

highest likelihood of explaining population structure.

LnPr(X/K) likelihood scores were used to infer DeltaK

values (Evanno et al. 2005). Subsequent partitioned

analyses of the two identified populations separately

revealed no additional structure. Partitioning of genetic

variation as suggested by STRUCTURE was validated

using analyses of molecular variance (AMOVA; GenAlEx

6.4). Principal Coordinate Analysis (PCoA) multivariate

tests were used to visualise major patterns in the dataset.

PCoA plots based on both individual and population

averaged genetic distances were generated in GenAlEx

6.4 to reveal clustering between samples.

To visualise phylogenetic relatedness of haplotypes

generated from the chloroplast microsatellites median-

joining networks were drawn in Network 4.6.0.0 (Fluxus

Engineering). All networks were rooted on a common

haplotype from C. gummiferum (N = 4), which morpho-

logical and molecular evidence have identified as the oldest

extant lineage in the genus ((Rozefelds and Barnes 2002;

Heslewood et al. in prep.).

Seed and pollen flow rates

Ust estimates of differentiation for both biparental and

maternal microsatellites were taken from AMOVA analyses

of nuclear and chloroplast microsatellite data, respectively.

Values for chloroplast microsatellite data were also esti-

mated as Gst, Nst and Rst using PermuteCpSSR 2.0 (Pons

and Petit 1996), the latter two measures have been suggested

to give more accurate estimates of differentiation for hap-

lotypic loci, taking into consideration differences in muta-

tion models and repeat number amongst alleles. To estimate

relative pollen and seed gene flow rates these differentiation

values were inserted into the migration equation of

Angelone et al. (2007), derived from (Ennos 1994):

mp/ms = [(1/FSTb - 1) - 2(1/FSTm - 1)]/(1/FSTm - 1);

where p = pollen, s = seed, b = biparental, m = maternal.

Genetica

123

Results

Genetic diversity: nSSRs

Diversity statistics for C. apetalum based on nine nuclear

microsatellites are summarised in Table 2. In total 179

alleles were detected, ranging from 10 to 29 per locus

(average 19.9). All loci were polymorphic in C. apetalum,

though locus Ca184BGT was monomorphic at two south-

ern sites (BC and Wo). Not all samples amplified for the

nine nuclear loci. 1, 2, 6, 8, 9 or 28 of the 332 individuals

failed to amplify at six loci, 1-3(-8) of these failures

occurring within single sites. Only 3 individuals failed to

amplify at two loci, suggestive of problematic DNA for

these individuals. Allelic richness, R7, ranged from 4.5 to

6.4 across subpopulations, excepting BC, the southernmost

site sampled, with a value of only 3.05, and the average R7

(excluding site KF, N = 5) was 5.3. Forty alleles were

unique to a site, with 0–5 private alleles per site. Average

allelic richness, heterozygosity and proportion of private

alleles were all marginally higher at northern sites.

Thirteen of 22 sites departed significantly from HWE

(uncorrected p \ 0.05), eight being highly significant

(uncorrected p \ 0.001), all 13 showing significant hetero-

zygote deficit. Average He was 0.681, ranging from 0.461

(BC) to 0.773 (R). Average Ho (Table 2) ranged from 0.448

(BC) to 0.722 (We) with an average of 0.62. Wright’s fixa-

tion index (Fis) ranged from 0.018 (Ro and We) to 0.274

(R) suggesting the occurrence of inbreeding at some sites.

Overall average Fis was 0.130, but the average among

southern sites was higher than in the north, 0.164 versus

0.105. Highly significant levels of inbreeding were detected

Table 2 Nuclear and plastid diversity statistics for the Ceratopetalum apetalum populations sampled

Pop Nuclear (9 loci) Chloroplast (4 loci)

N NA Mean A

per locus

Ap R7 Ho He Fis N Chlorotype He(s.e.) R3

1 BC 9 29 3.222 0 3.053 0.448 0.461 0.088 3 1 0.000 1.000

2 BF 10 53 5.889 0 5.171 0.584 0.669 0.179* 4 1 0.000 1.000

3 HT 10 50 5.556 1 4.928 0.631 0.651 0.086 5 1 0.000 1.000

4 Ro 10 51 5.667 1 5.024 0.715 0.690 0.018 5 1,2 0.064 (0.064) 1.867

5 KF 5 37 4.111 0 n.a. (R4 3.734) 0.589 0.606 0.141 5 3 0.000 1.000

6 R 18 80 8.889 1 6.277 0.583 0.773 0.274*** 14 1,3 0.245 (0.099) 1.897

7 WF 19 69 7.667 3 5.383 0.537 0.693 0.252*** 9 3 0.000 1.000

8 NG 15 57 6.333 1 4.934 0.643 0.678 0.089* 15 3 0.000 1.000

9 Ev 15 73 8.111 2 5.910 0.688 0.725 0.086*** 4 2,4 0.150 (0.091) 1.964

10 LC 11 67 7.444 2 6.396 0.637 0.783 0.237*** 4 3 0.000 1.000

11 EB 10 57 6.333 3 5.672 0.640 0.712 0.159*** 4 3 0.000 1.000

12 DC 37 87 9.667 4 5.554 0.595 0.723 0.191*** 11 5,6 0.059 (0.059) 1.751

13 Wo 10 46 5.111 2 4.559 0.591 0.624 0.105 4 6 0.000 1.000

Ave South 13.8 6.462 1.5 (20) 5.238 0.606 0.676 0.164 6.7 (1–6) 1.268

14 BT 15 59 6.556 0 4.858 0.513 0.633 0.222*** 4 6 0.000 1.000

15 We 10 61 6.778 3 5.751 0.722 0.698 0.018 4 7 0.000 1.000

16 D 24 82 9.111 5 5.593 0.689 0.729 0.076* 8 8 0.000 1.000

17 NB 16 66 7.333 2 5.603 0.627 0.706 0.141* 6 8, 9, 10 0.111 (0.068) 2.545

18 Wa 22 64 7.111 1 4.478 0.545 0.564 0.056 7 11 0.000 1.000

19 WaN 13 55 6.111 2 5.020 0.607 0.676 0.141** 4 11 0.000 1.000

20 Ni 18 76 8.444 5 5.846 0.641 0.734 0.154*** 6 11 0.000 1.000

21 NiB 23 87 9.667 2 5.974 0.694 0.740 0.085 4 11 0.000 1.000

22 Sp 10 54 6.000 1 5.305 0.704 0.702 0.051 4 11 0.000 1.000

Ave North 16.8 7.457 2.2 (20) 5.381 0.638 0.687 0.105 5.2 (6–11) 1.172

ALL 14.9 61.8 6.851 1.8 (40) 5.297 0.620 0.681 0.130 6.1 11 0.029 (0.017) 1.230

N number of populations, NA number of alleles, AP number of private alleles, R allelic richness, Ho He observed and expected heterozygosities,

Fis Wright’s fixation index

* p \ 0.05; ** p \ 0.01; *** p \ 0.001

Values in bold represent statistics averaged across all 22 sampled population, or across the southern (Pops 1–13) and northern (Pops 14–22)

populations identified by the structure analysis

Genetica

123

in all but one of the lowest altitude sites sampled from sea

level up to 420 m and a few at higher altitudes (Tables 1, 2).

Genetic diversity: cpSSRs

Four variable loci each produced 2–4 alleles across the

sampled C. apetalum resulting in 11 chloroplast haplotypes

(chlorotypes) (Tables 2, 3). All but five populations were

fixed for a chlorotype (Table 2). Chlorotype diversity was

distributed across the species range but was lower in the

north. The five northernmost sites spread across 168 km

shared a single haplotype while over a similar geographic

range (*176 km), six chlorotypes were found amongst 13

southern sites. Average heterozygosity for the plastid data

was low at only 0.034, ranging from 0 (at 17 of 22 sites) to

0.245, with corrected allelic diversity (R3) ranging from 1

to 2.545 per site, an average of 1.228 (Table 2).

Genetic structure: nSSRs

Overall pairwise Fst values ranged from 0.019 to 0.171

(mean Fst = 0.077) indicating high gene flow between

sites. Fst was lower in the north (mean Fst = 0.047,

p \ 0.01) than amongst southern sites (mean Fst = 0.074,

p \ 0.01). Jost’s estimator produced an average overall

estimate of D = 0.211. Dest = 0.148 among the 13

southern sites, and Dest = 0.110 among the nine northern

sites. Overall both measures are higher in the south, and D

values were higher than Fst values, which corroborate our

measurements of high levels of within-site heterozygosity

found in the nuclear microsatellite data.

For the STRUCTURE analysis of the nuclear microsatel-

lite data, K = 2 (DK = 203.01) gave the highest assignment

values to clusters (average per population of 0.743–0.983;

Fig. 1). These two clusters corresponded to a north–south split

amongst populations located on either side of the Hunter River

Valley (Fig. 1), with sampled sites in the northern and

southern population separated by a minimum of 159 km. The

second highest DeltaK peak identified at K = 7

(DK = 77.85) grouped geographically meaningful clusters

corresponding to three and four subclusters north and south of

the Hunter River Valley. The three northern clusters reflected

localised geographic proximity (with moderate to high

assignment values). The northernmost sites clustered (Sp,

NiB, Ni); the two sites from the Washpool area clustered and a

third cluster included sites NB, D and We. Individuals from

the southernmost ‘northern’ site BT had moderate assignment

to both of the latter two clusters (0.386 and 0.457). Four

clusters were identified amongst the southern population but

there was little obvious geographic structuring (Fig. 2).

Three-way AMOVA of the data in regions identified by

STRUCTURE (K = 2) revealed nuclear variation was

partitioned primarily within sites (82 %; p B 0.001), and

equally among sites (10 %; p B 0.001) and among regions

(9 %; p B 0.001). Two-way AMOVA of data segregated

into sites in only the northern or the southern populations

revealed significant but moderate PhiPT of 0.116 and 0.106

(p = 0.01), with 88–89 % of variation contained within

sites. Principal coordinate analysis of genetic distances

between all samples revealed a large aggregation of data

points with general separation of southern and of northern

samples but with a small zone of overlap (Fig. 3a). PCoA

of the genetic distances averaged by site showed no clear

clustering among the southern site (Fig. 3b). PCoA of the

northern sites revealed geographic clustering (Fig. 3c)

corresponding to five of the eight regions which comprise

the Gondwanan Rainforests of Australia.

Genetic structure: cpSSRs

There was some evidence for regional sharing of chloro-

types but no clear evidence of explosive diversification

Table 3 Composition and

distribution of chlorotypes in

Ceratopetalum apetalum

Chlorotype

(S to N)

cpSSR Locus Individuals Sites Range

(km)ccmp4 ccmp5 ccmp6 ccmp10

1 139 134 120 118 26 5 76

2 139 135 120 118 2 2 108

3 139 135 121 127 41 6 84

4 138 134 120 118 3 1 –

5 138 134 121 118 9 1 –

6 138 133 121 118 10 3 172

7 139 133 121 127 4 1 –

8 138 133 121 127 12 2 18

9 138 133 122 127 1 1 –

10 138 134 121 127 1 1 –

11 139 136 120 118 25 5 168

Total 2 4 3 2 134 22 762

Genetica

123

from a common internal chlorotype (although within the

northern range three chlorotypes differed from the common

NB/D chlorotype by single mutations at a different micro-

satellite locus). This included two unique individuals from

NB and a fixed difference at We to the south. These four

northern chlorotypes were 1–6 (or 9–14) mutations diver-

gent from the other two chlorotypes found in the north.

Representation of the 11 C. apetalum chlorotypes in a

median-joining network included two four-way reticula-

tions indicating some homoplasy or admixture. When the

network was rooted by a chlorotype of a partially sympatric

sister taxon, Ceratopetalum gummiferum (in black, Fig. 2)

only one reticulation remained amongst those chlorotypes.

The large divergence in the centre of the network is due to

a 9 bp size difference at ccmp10. Evidence from a number

of taxa (Rossetto pers. obs., H. McPherson pers. comm.)

also show this pattern at ccmp10 suggesting that it may be

prone to large mutations as opposed to the typical single

base (step-wise) mutations at other chloroplast loci. If that

is true the 9 bp size difference between the only two alleles

present at this locus may represent a single mutational step.

All other chlorotypes diverged by only single base muta-

tions. The predominant Sydney metropolitan area chloro-

type, 3, was somewhat divergent from remaining southern

chlorotypes, separated by four or 12 mutations. The

remaining five southern chlorotypes successively differed

by single mutations. The northernmost chlorotype, 11, was

separated by five mutations from chlorotype 6, the only

chlorotype common to both populations identified from the

nuclear microsatellite analyses and shared by

Fig. 2 Geographical

distribution of chlorotypes in C.

apetalum based on chloroplast

microsatellite data. A

representation of the 11

chlorotypes in a median-joining

network is also shown (inset)

rooted on a sister taxon, C.

gummiferum (in black). Dashes

indicate the number of changes

between chlorotypes.

Population numbers correspond

to those listed in Table 2

Genetica

123

geographically nearest sites separated by up to 172 km.

The pronounced divergence between the northernmost and

remaining northern chlorotypes corresponds to divergence

over the dry Clarence River Corridor in far northern NSW.

Comparative genetic differentiation: nSSRs

versus cpSSRs

Estimates of Fst, Ust, Gst, Rst and Nst were compared for

the chloroplast microsatellite data to estimate the effect of

mutation model and allele sizes, and to determine their

effect on estimates of pollen/seed migration rates (Table 4).

All estimates of Rst (which takes into account differences in

repeat number) and Nst (which takes into account genetic

distance between alleles) were significantly greater than Gst

(which considers all alleles equally divergent) indicating

there is structure present in the species.

Estimates of differentiation were taken from AMOVA

analyses of matching sample sets of nuclear and chloro-

plast microsatellite data for 134 individuals and from

estimates of Gst, Rst and Nst calculated in PermuteCpSSR

2.0. The overall pollen/seed flow estimate across the

Fig. 3 Principal coordinate

analysis plots of genetic

distances: a between all

samples; b averaged by site;

c averaged by site among

northern localities only

Genetica

123

species was high (range 42.01–137.04) indicating most

gene flow in the species is due to pollen movement. When

partitioned to look at gene flow within the two populations

identified from STRUCTURE analysis, all differentiation

measures showed a substantially higher pollen/seed

migration rate within the northern population (range

86.68–4,643.33) indicating ineffective seed dispersal

(Table 4).

Discussion

Pollen rather than seed mediated gene flow

UST estimates of nuclear and chloroplast differentiation

across 134 C. apetalum individuals, suggest a relatively

high pollen to seed flow rate of 52:1. Similar patterns are

thought to be common, albeit highly variable, in many

other temperate and tropical plant species with wind-pol-

lination and wind-dispersal habits (Jordano 2010). Petit

et al. (2005) examined data from 183 species in 52 families

and found a median pollen to seed ratio of 17:1 but with

wide variance, primarily related to reproductive charac-

teristics. High pollen to seed flow estimates (i.e. [90:1)

have been measured in wind-pollinated outcrossing species

of Quercus (Ennos 1994; Pakkad et al. 2008) while pollen

to seed flow rates of \3:1 were estimated in the insect-

pollinated and fleshy-fruited Sorbus torminalis (Angelone

et al. 2007; Oddou-Muratorio et al. 2001).

Previous studies of local rainforest species suggested

that pollen-mediated gene flow is often limited, as sup-

ported by the observation that species with restricted fruit

dispersal have high levels of genetic structure even across

small geographic areas (Rossetto et al. 2008, 2009). Here

we discover that wind pollination might have a signifi-

cantly greater role than previously expected for a species

thought to be insect-pollinated (Hoogland 1960, 1981;

Rozefelds and Barnes 2002). We also confirm the relatively

limited dispersal potential of the seed-bearing fruit, which

as noted earlier is woody, unpalatable to animal dispersers,

and thus limited to passive or wind dispersal. Interestingly,

the differences in structure and gene flow observed across

the distribution of C. apetalum are likely to be linked to

local conditions and respective population sizes. In the

southern distributional range, most populations are found

in gullies within highly divided landscapes, effectively

preventing large scale dispersal of wind-dispersed fruit and

pollen even in a canopy tree. In contrast, northern sites

often comprise very large mature stands separated by a

range of geographic distances, but potentially more

accessible to wind dispersed pollen than seed (as suggested

by the higher pollen to seed dispersal ratio). It is also likely

that these larger northern populations provide fewer

opportunities for external colonisers.

Compared to animals or herbaceous plants, many trees

maintain genetic diversity even through repeated habitat

expansions and reductions through factors such as long

juvenile stage and adult longevity. Thus through shorter

temporal disturbances, long-lived trees can ‘ride it out’ and

wind-pollinated trees in fragments can maintain gene flow

with pollen dispersal as previously demonstrated in rain-

forest trees (Rossetto et al. 2004a, 2008; Rossetto and

Kooyman 2005). Our study of C. apetalum confirms that

trees with inefficient dispersal combined with persistence

at sites through traits such as coppicing and a long term

seedling bank can maintain high nuclear and chloroplast

diversity over a broad range.

Genetic diversity and geographic structure

The level of nuclear diversity in C. apetalum is comparable

to other local rainforest tree species. Elaeocarpus grandis,

an early successional, fleshy-fruited lowland tree from

subtropical and tropical rainforests, showed slightly lower

diversity (He = 0.57) across its northern NSW range but

very little structure and landscape-level differentiation

(Rossetto et al. 2004b). However, when populations from

the Wet Tropics where compared to northern NSW

Table 4 Pollen/seed flow estimate across all Ceratopetalum apetalum populations sampled, and across northern and southern sites only

Data Pollen/seed migration

UST Gst Rst Nst UST Gst Rst Nst

Overall nuc 0.144

cp 0.904 0.881 0.959* 0.923* 53.98 42.01 137.04 69.26

Southern cluster nuc 0.115

cp 0.859 0.828 0.933** 0.884* 44.9 35.05 105.16 56.65

Northern cluster nuc 0.097

cp 0.998 0.905 0.998*** 0.965* 4,643.3 86.68 4,643.33 254.67

Significance level * p \ 0.05; ** p \ 0.01; *** p \ 0.001

Genetica

123

populations a higher level of between-region divergence

(23 %; Rossetto et al. 2007) was detected in E. grandis

then in C. apetalum. Although this might be unexpected in

view of the more easily dispersed nature of Elaeocarpus

fruits, this could be a consequence of the much larger gap

between the sampled E. grandis populations. A recent

study of nDNA microsatellite variation in Podocarpus el-

atus, a widely distributed dry rainforest conifer with fleshy

fruits, found similar levels of diversity (Hs = 0.53) across

its entire distribution of over 2,000 km and also identified a

major genetic break corresponding to the Clarence River

Corridor (Mellick et al. 2011). Marked latitudinal variation

in genetic diversity has been noted in several southern

Australian woody species, Acacia melanoxylon R.Br.

(Playford et al. 1993), Elaeocarpus angustifolius (Rossetto

et al. 2007), Nothofagus moorei (Taylor et al. 2005), the

genera Telopea (Rossetto et al. 2012) and Lomatia (Milner

et al. 2012).

We identified two important biogeographic barriers for

C. apetalum in eastern Australia: the Hunter River Valley

and the Clarence River Corridor which both represent wide

dry forest breaks (Di Virgilio et al. 2012). The Hunter

River Valley which separates the Sydney Basin and North

Coast bioregions is the driest and one of the largest coastal

valleys in NSW with a wide floodplain and vast estuarine

wetlands. It is a known barrier for a number of species.

Floristically, the Hunter is a known genetic barrier for a

number of species e.g. Acacia melanoxylon R.Br. (Playford

et al. 1993) and Lomatia R.Br. (Milner et al. 2012). The

Hunter region is also designated as one of the major geo-

graphic barriers for birds in mainland Australia (Ford 1987)

and several authors have described phenotypic and distri-

butional (Ford 1987; Nicholls and Austin 2005; Schodde

and Mason 1997) or genetic breaks across the Hunter

(Joseph et al. 2008). Ceratopetalum apetalum to the south

of the Hunter occupies sites almost exclusively on sand-

stone substrates and comprise both warm temperate rain-

forest pockets and wet sclerophyll forests from 900 m

down to sea level. At many southern sites trees attain only

around 10 m height and are often heavily multi-stemmed in

response to drier habitat and greater disturbance from fire

in sites enclosed by dry sclerophyll forests. In comparison

in the north, C. apetalum occurs only at higher altitudes,

from 420 to over 1,000 m, and is often present in larger

continuous stands with mature adults, usually comprising

single primary stems [25 m in height, common. Soils in

the northern sites are all derived from volcanic elements—

rhyolites, granites, basalt.

The Clarence River Corridor is the largest coastal river

catchment in NSW with a broad floodplain. This region

constitutes a wide dry break in northern NSW vegetation

and is the divide between the North Coast and South East

Queensland bioregions. Sea level at the Clarence had a low

of approximately -100 m below present levels but has

been stable for roughly 6,500 years (Tulau 1999). Rainfall

in the area is summer-autumn dominant as it is in the

Hunter Valley, and varies over the floodplain from 1,000

around Grafton to 1,500 mm at the coast. Although either

of these two biogeographic barriers has been shown to

independently impact on the distribution of genetic diver-

sity in plants before, this is the first study reporting the

combined impact of the Hunter River Valley and the

Clarence River Corridor on the genetic structure of a single

species.

Climatic cycles, contractions and expansions,

and refugia

In Australian cool-temperate rainforests, strong phylogeo-

graphic structure among 30 chlorotypes was identified

across Tasmania, Victoria and southern NSW populations

of the Australian species Tasmannia lanceolata, and was

attributed to the existence of multiple isolated refugia at the

LGM and Pleistocene near present day sites (Worth et al.

2010). Even though this species is bird dispersed, dispersal

is infrequent or limited to short distances such that strong

phylogeographic structure has been maintained. Similar

‘patchiness’ of chlorotype distributions was recorded for C.

apetalum. Chloroplast diversity, indicative of seed medi-

ated gene flow, was generally absent within sites and

regional sharing of chlorotypes where it occurred was

limited to sites no more than 172 km apart suggestive of

limits to longer distance seed dispersal. Sites separated by

much shorter distances than that in the south showed

chlorotype differences, sometimes fixed, indicating limits

or barriers to seed dispersal over shorter scales. This spatial

distribution demonstrates that there are sometimes limits to

seed dispersal in this species over even a small scale. The

four Blue Mountains sites also possessed from 1 to 3 pri-

vate alleles amongst the nuclear loci and among site dif-

ferences in genetic diversity are illustrative of possible

barriers to pollen flow as well between sites in this dis-

sected landscape. Overall, the pattern of diversity in C.

apetalum is suggestive of persistence in a number of ref-

ugial areas across the present range with little evidence of

post-Pleistocene expansion.

Lack of concordant geographic structure amongst

nuclear and chloroplast genes suggest differences in pat-

terns of seed and pollen dispersal, and is supported by the

strong differences observed in ratio between these param-

eters. While C. apetalum comprises two clear contempo-

rary populations with limited current gene flow, chloroplast

diversity is locally differentiated but not simply structured

according to north–south boundaries or geographic distri-

butions. The two terminal chlorotypes in the C. apetalum

cpSSR network are isolated from other chlorotypes found

Genetica

123

in their contemporary populations, but otherwise the

median-joining network (Fig. 1) shows a reasonable geo-

graphic arrangement. The absence of a clearly identifiable

ancestral chlorotype suggests that the present distribution

of chlorotypes reflects local losses of diversity from a pre-

Pleistocene gene pool rather than a postglacial radiation of

haplotypes. This may have resulted in the loss of ‘northern’

haplotypes from the south, or intermediate southern hap-

lotypes may have become extinct. The cpSSR data is

suggestive of multiple refugia existing within the species

range, as evidenced by the presence of regional sharing of

at least six out of eleven chlorotypes. In addition, the level

of differentiation shown in the chloroplast indicates a deep

divergence between the upper and lower portions of the

northern population, coincident with a vicariant break at

the Clarence River Corridor. North of the Clarence sub-

populations of C. apetalum sampled over a 168 km range

were fixed for a single unique chlorotype. This suggests

possible expansion from a major refugial area in the north,

though moderately high nuclear differentiation among

northern sites suggests longstanding presence at current

locations or that this area existed as a metapopulation until

recently.

The nature of some localities e.g. highly dissected gul-

lies along the sandstone escarpments of the Blue Moun-

tains might promote the formation of microrefugia and any

fine scale population substructuring would impact on

equilibrium and diversity measures estimated in sites.

Additionally, a number of populations exhibited significant

levels of inbreeding which could partially account for

population-level differentiation. However the southernmost

sampled site at Brogers Creek (BC), only 14 km from the

next sampled site above Belmore Falls (BF), on all evi-

dence demonstrated lower levels of diversity than any other

site (R7 = 3.05, He = 0.461). BC clustered apart from the

remaining southern sites in PCoA analysis but showed no

sign of inbreeding and contained no private alleles. BC

represents an example of how even small, isolated popu-

lations can survive across mostly unsuitable habitat (along

a creek line protected by sandstone cliffs within sclerophyll

woodland), an occurrence more common in the south than

the north where greater habitat availability may support

longer term connectivity.

Rainforest in the Washpool area, which includes the

largest extant stand of C. apetalum, is distinctive in resulting

from recolonisation by seed dispersed taxa and relictual

rainforest species. Kooyman et al. (2011) showed evidence

of this by demonstrating denser clustering of community

phylogenetic structure in rainforests at Washpool and also

Dorrigo relative to Nightcap or the AWT reflecting large

losses of taxonomic richness in these areas due to habitat

contractions and subsequent patterns of recolonisation. This

is consistent with our findings in Washpool. Lower

heterozygosity at Washpool (Ho = 0.545, R7 = 4.5 Wa;

Ho = 0.607, R7 = 5.0 WaN) than all other northern sites

excepting Lower Barrington, could be explained by recent

recolonisation and founder events from the Nightcap and

McPherson Ranges to the north (as suggested in Kooyman

et al. 2011). Although small coachwood stands may have

persisted at Washpool through glacial cycles, subsequent

pollen and seed migration from other refugia appear to have

homogenised chloroplast and nuclear diversity. In contrast,

C. apetalum is likely to have remained at Dorrigo in a more

substantive refugium resulting in the maintenance of higher

genetic diversity (Ho = 0.689, R7 = 5.6).

Conclusion

We confirmed overall limited seed-mediated gene flow in

this non-fleshy-fruited taxon, coupled with unexpectedly

high pollen-mediated gene flow. Gene flow, either seed or

pollen mediated, was highly impacted by local landscape

characteristics, with marked restriction over even short

distances in the most dissected habitat in the south. Our

findings reinforce the concept that small pockets of

coachwood persisted throughout the landscape during

rainforest contractions and might have preserved diversity

for the species as well as providing habitat for other taxa,

and therefore preserving floristic (rainforest) diversity.

Acknowledgments We thank Steve Clarke, Maria Cotter, Robert

Kooyman, Hannah McPherson, Louisa Murray, Carolyn Connelly,

Marlien van der Merwe, Peter Weston and Michael Whitehead for

assisting with fieldwork or supplying specimens. We thank Chris

Togher, Chris Allen and Rohan Mellick for assistance with the GIS

and the map. The project was funded by the Australian Research

Council (DP0665859).

References

Adam P (1992) Australian rainforests. Oxford University Press,

Oxford

Angelone S, Hilfiker K, Holderegger R, Bergamini A, Hoebee SE

(2007) Regional population dynamics define the local genetic

structure in Sorbus torminalis. Mol Ecol 16:1291–1301

Barnes RW, Hill RS (1999) Ceratopetalum fruits from Australian

Cainozoic sediments and their significance for petal evolution in

the genus. Aust Syst Bot 12:635–645

Baur GN (1957) Nature and distribution of rainforests in New South

Wales. Aust J Bot 5:190–233

Black MP, Mooney SD, Martin HA (2006) A[43000-year vegetation

and fire history from Lake Baraba, New South Wales, Australia.

Quat Sci Rev 25:3003–3016

Bond WJ, Midgley JJ (2001) Ecology of sprouting in woody plants:

the persistence niche. Trends Ecol Evol 16:45–51

Bowman DMJS (2000) Australian rainforests—Islands of green in a

land of fire. Cambridge University Press, Cambridge, UK

Crawford NG (2010) SMOGD: software for the measurement of

genetic diversity. Mol Ecol Resour 10:556–557

Genetica

123

Di Virgilio G, Laffan SW, Ebach MC (2012) Fine-scale quantification

of floral and faunal breaks and their geographic correlates, with an

example from south-eastern Australia. J Biogeogr 39:1862–1876

Ennos RA (1994) Estimating the relative rates of pollen and seed

migration among plant-populations. Hered 72:250–259

Evanno G, Roegnaut S, Goudet J (2005) Detecting the number of

clusters of individuals using the software STRUCTURE: a

simulation study. Mol Ecol 14:2611–2620

Floyd AG (1990) Australian rainforests in New South Wales. Surrey

Beatty, Sydney

Ford J (1987) Minor isolates and minor geographical barriers in avian

speciation in continental Australia. EMU 87:90–102

Goudet J (1995) FSTAT (Version 1.2): a computer program to

calculate F-statistics. J Hered 86:485–486

Heslewood MM, Porter C, Avino M, Rossetto M (2009) Isolation and

characterization of nuclear microsatellite loci from Ceratopeta-

lum apetalum (Cunoniaceae). Mol Ecol Resour 9:566–568

Heuertz M, Fineschi S, Anzidei M et al (2004) Chloroplast DNA

variation and postglacial recolonization of common ash (Frax-

inus excelsior L.) in Europe. Mol Ecol 13:3437–3452

Heuertz M, Carnevale S, Fineschi S et al (2006) Chloroplast DNA

phylogeography of European ashes, Fraxinus sp (Oleaceae): roles

of hybridization and life history traits. Mol Ecol 15:2131–2140

Hewitt GM (2004) Genetic consequences of climatic oscillations in

the Quaternary. Phil Trans R Soc B 359:183–195

Hilbert DW, Graham A, Hopkins MS (2007) Glacial and interglacial

refugia within a long-term rainforest refugium: the wet tropics

bioregion of NE Queensland, Australia. Palaeogeogr Palaeoclim

Palaeoecol 251:104–118

Hill RS, Truswell EM, McLoughlin S, Dettmann ME (1999) The

evolution of the Australian flora: fossil evidence. In: Orchard AE

(ed) Flora of Australia, vol 1., IntroductionCSIRO Publishing,

Melbourne, pp 251–320

Hoogland RD (1960) Studies in the Cunoniaceae. I. The genera

Ceratopetalum, Gillbeea, Aistopetalum, and Calycomis. Aust J

Bot 8:318–341

Hoogland RD (1981) Studies in the Cunoniaceae III. Additional notes

on Ceratopetalum and Acrophyllum. Brunonia 4:213–216

Hopkins MS, Ash J, Graham AW, Head J, Hewett RK (1993)

Charcoal evidence of the spatial extent of the Eucalyptus

woodland expansions and rainforest contractions in north

Queensland during the late Pleistocene. J Biogeogr 20:357–372

Jordano P (2010) Pollen, seeds and genes: the movement ecology of

plants. Hered 105:329–330

Joseph L, Dolman G, Donnellan S et al (2008) Where and when does

a ring start and end? Testing the ring-species hypothesis in a

species complex of Australian parrots. Proc Trans R Soc B

275:2431–2440

Jost LOU (2008) GST and its relatives do not measure differentiation.

Mol Ecol 17:4015–4026

Kershaw AP, Bretherton SC, van der Kaars S (2007) A complete

pollen record of the last 230 ka from Lynch’s Crater, north-

eastem Australia. Palaeogeogr Palaeoclim Palaeoecol 251:23–45

Kooyman R, Rossetto M, Cornwell W, Westoby M (2011) Phyloge-

netic tests of community assembly across regional to continental

scales in tropical and subtropical rain forests. Global Ecol

Biogeogr 20:707–716

Magri D, Vendramin GG, Comps B et al (2006) A new scenario for

the Quaternary history of European beech populations: palaeo-

botanical evidence and genetic consequences. New Phytol

171:199–221

McLoughlin S (2001) The breakup history of Gondwana and its impact

on pre-Cenozoic floristic provincialism. Aust J Bot 49:271–300

Mellick R, Lowe A, Rossetto M (2011) Consequences of long- and short-

term fragmentation on the genetic diversity and differentiation of a

late successional rainforest conifer. Aust J Bot 59:351–362

Milner ML, Rossetto M, Crisp MD, Weston PH (2012) The impact of

multiple biogeographic barriers and hybridization on species-

level differentiation. Am J Bot 99:2045–2057

Nicholls JA, Austin JJ (2005) Phylogeography of an east Australian

wet-forest bird, the satin bowerbird (Ptilonorhynchus violaceus),

derived from mtDNA, and its relationship to morphology. Mol

Ecol 14:1485–1496

Oddou-Muratorio S, Petit RJ, Le Guerroue B, Guesnet D, Demesure

B (2001) Pollen-versus seed-mediated gene flow in a scattered

forest tree species. Evolution 55:1123–1135

Pakkad G, Ueno S, Yoshimaru H (2008) Genetic diversity and

differentiation of Quercus semiserrata Roxb. in northern Thai-

land revealed by nuclear and chloroplast microsatellite markers.

For Ecol Manag 255:1067–1077

Peakall R, Smouse PE (2006) GENALEX 6: genetic analysis in

Excel. Population genetic software for teaching and research.

Mol Ecol Notes 6:288–295

Petit RJ, Duminil J, Fineschi S et al (2005) Comparative organization

of chloroplast, mitochondrial and nuclear diversity in plant

populations. Mol Ecol 14:689–701

Playford J, Bell JC, Moran GF (1993) A major disjunction in genetic

diversity over the geographic range of Acacia melanoxylon

R.Br. Aust J Bot 41:355–368

Pons O, Petit RJ (1996) Measuring and testing genetic differentiation

with ordered versus unordered alleles. Genetics 144:1237–1245

Pritchard JK, Stephens M, Donnelly P (2000) Inference of population

structure using multilocus genotype data. Genetics 155:945–959

Rice WR (1989) Analyzing tables of statistical tests. Evolution

43:223–225

Rossetto M, Kooyman RM (2005) The tension between dispersal and

persistence regulates the current distribution of rare palaeo-

endemic rain forest flora: a case study. J Ecol 93:906–917

Rossetto M, Gross CL, Jones R, Hunter J (2004a) The impact of

clonality on an endangered tree (Elaeocarpus williamsianus) in a

fragmented rainforest. Biol Conserv 117:33–39

Rossetto M, Jones R, Hunter J (2004b) Genetic effects of rainforest

fragmentation in an early successional tree (Elaeocarpus gran-

dis). Hered 93:610–618

Rossetto M, Crayn D, Ford A, Ridgeway P, Rymer P (2007) The

comparative study of range-wide genetic structure across related,

co-distributed rainforest trees reveals contrasting evolutionary

histories. Aust J Bot 55:416–424

Rossetto M, Kooyman R, Sherwin W, Jones R (2008) Dispersal

limitations, rather than bottlenecks or habitat specificity, can

restrict the distribution of rare and endemic rainforest trees. Am J

Bot 95:321–329

Rossetto M, Crayn D, Ford A, Mellick R, Sommerville K (2009) The

influence of environment and life-history traits on the distribu-

tion of genes and individuals: a comparative study of 11

rainforest trees. Mol Ecol 18:1422–1438

Rossetto M, Allen C, Thurlby K, Weston P, Milner M (2012) Genetic

structure and bio-climatic modeling support allopatric over

parapatric speciation along a latitudinal gradient. BMC Evol

Biol 12:149

Rousset F (2008) GENEPOP ‘ 007: a complete re-implementation of

the GENEPOP software for Windows and Linux. Mol Ecol

Resour 8:103–106

Rozefelds AC, Barnes RW (2002) The systematic and biogeograph-

ical relationships of Ceratopetalum (Cunoniaceae) in Australia

and New Guinea. Int J Plant Sci 163:651–673

Schodde R, Mason IJ (1997) Psittacidae. In: WWK Houston, Wells A

(eds) Aves (Columbidae to Coraciidae). Zoological Catalogue of

Australia Vol. 37.2. CSIRO Publishing, Melbourne, pp 109-218.

Taylor KJ, Lowe AJ, Hunter RJ et al (2005) Genetic diversity

and regional identity in the Australian remnant Nothofagus

moorei. Aust J Bot 53:437–444

Genetica

123

Tulau MJ (1999) Acid sulfate soil management priority areas in the

Lower Clarence floodplain (Report). Department of Land and

Water Conservation, Sydney

van Oosterhout C, Weetman D, Hutchinson WF (2006) Estimation

and adjustment of microsatellite null alleles in nonequilibrium

populations. Mol Ecol Notes 6:255–256

Webb LJ, Tracey JG (1981) Australian rainforests: patterns and

change. In: Keast A (ed) Ecological Biogeography of Australia.

W. Junk, The Hague, pp 605–694

Weising K, Gardner RC (1999) A set of conserved PCR primers for the

analysis of simple sequence repeat polymorphisms in chloroplast

genomes of dicotyledonous angiosperms. Genome 42:9–19

Williams NJ, Harle KJ, Gale SJ, Heijnis H (2006) The vegetation

history of the last glacial–interglacial cycle in eastern New South

Wales. J Quat Sci 21:735–750

Worth JRP, Jordan GJ, McKinnon GE, Vaillancourt RE (2009) The major

Australian cool temperate rainforest tree Nothofagus cunninghamii

withstood Pleistocene glacial aridity within multiple regions:

evidence from the chloroplast. New Phytol 182:519–532

Worth JRP, Jordan GJ, Marthick JR, McKinnon GE, Vaillancourt RE

(2010) Chloroplast evidence for geographic stasis of the

Australian bird-dispersed shrub Tasmannia lanceolata (Winter-

aceae). Mol Ecol 19:2949–2963

Genetica

123