lable at ScienceDirect

Estuarine, Coastal and Shelf Science 86 (2010) 179–188

Contents lists avai

Estuarine, Coastal and Shelf Science

journal homepage: www.elsevier .com/locate/ecss

Gene flow and demographic history of the mangrove crab Neosarmatium meinerti:A case study from the western Indian Ocean

Lapo Ragionieri a,*, Stefano Cannicci a, Christoph D. Schubart b, Sara Fratini a

a Department of Evolutionary Biology ‘‘Leo Pardi’’, via Romana 17, 50125 Firenze, Italyb Biologie I, University of Regensburg, 93040 Regensburg, Germany

a r t i c l e i n f o

Article history:Received 12 January 2009Accepted 9 November 2009Available online 12 November 2009

Keywords:East African coastSeychellespopulation genetic structurelarval dispersalevolutionary significant units

* Corresponding author.E-mail addresses: lapo.ragionieri@unifi.it, lapo.ragion

0272-7714/$ – see front matter � 2009 Elsevier Ltd.doi:10.1016/j.ecss.2009.11.002

a b s t r a c t

Most marine organisms are characterized by at least one planktonic phase during their life history,potentially allowing interconnection of populations separated by several hundred kilometers. For manyyears, the idea that marine species are genetically homogenous throughout their range of distribution,due to passive larval transport, has been a paradigm. Nowadays, a growing number of studies underlinethe existence of boundaries in the marine realm and highlight how larval dispersal is a complex processdepending on biotic as well as abiotic factors. Marine fragmented habitats, such as atolls, mangroves andestuaries, are optimal systems for investigating the marine dispersion process under a metapopulationapproach, since populations can be geographically defined a priori as opposed to those occupying openmarine environments. Within this frame, the present paper investigates the population genetic structureand the demographic history of the mangrove crab Neosarmatium meinerti within the western IndianOcean by partial sequences of the mitochondrial DNA cytochrome oxidase subunit I. A total of 167specimens were sampled from six mangrove sites distributed along the East African coast, from Kenya toSouth Africa, also including a mangrove forest located on Mahe Island, Seychelles. A sharp genetic breakbetween the mainland and the Seychelles is recorded, revealing the existence of two historically distinctgroups that can be defined as independent evolutionary units. Gene flow along the East African coastappears to be high enough to form a single metapopulation, probably by means of stepping stonepopulations. Otherwise, this mainland metapopulation is currently under expansion through a gradualmoving front from the subtropical toward the equatorial populations.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

The majority of marine invertebrate species are characterized byat least one widely potentially dispersive phase during their lifehistory, such as gametes, eggs, larvae or spores, generically calledpropagules (Fairweather, 1991). These planktonic stages arereleased into the open sea where, after a developmental period thatcan vary from hours to months or even years (Levin, 2006), theysettle in the proper habitat where they can grow to adulthood.Thus, the dispersal stages have the ability to potentially disperseacross hundreds of kilometers by means of ocean currents. Larvaeof coastal species, for instance, have been observed in mid-oceanplankton, suggesting to have crossed long distances (Palumbi, 2001,2003). For many years these findings, together with the lack ofevident geographical barriers in the marine realm, supported theidea that marine species are demographically open, with high level

ieri@gmail.com (L. Ragionieri).

All rights reserved.

of population exchange (Becker et al., 2007) and geneticallyhomogenous throughout their geographical range (Cowen et al.,2000). On the other hand, in recent years several studies employinggenetic techniques, chemical tags and/or biological and physicalmodels, revealed many examples of marine species with reducedconnectivity among populations (Knowlton, 2000; DiBacco et al.,2006) and highlighted how larval dispersal is a complex processdepending on several biotic and abiotic factors (Bilton et al., 2002).

Many marine habitats, such as atolls, coral reefs and lagoons, arenaturally fragmented and patchy, consequently allowing theirpopulations to be defined a priori as opposed to those occupyingopen marine environments. This allows us to approach the study ofconnectivity among populations using a metapopulation-basedapproach (DiBacco et al., 2006). With this approach, ‘‘landscapes areviewed as a network of habitat patches or fragments in which speciesoccur as discrete local populations connected by the passive and activemigration of individuals’’ (DiBacco et al., 2006, p. 184). Thus, theexchange of larvae among populations may be described by twoalternative models: the Island Model (Wright, 1940), in whichdefined populations of the same size exchange individuals with

L. Ragionieri et al. / Estuarine, Coastal and Shelf Science 86 (2010) 179–188180

equal probability, and the Stepping Stone Model (Kimura andWeiss, 1964), in which the genetic exchange is supposed to happenexclusively between adjacent demes, and consequently theconnection among populations would be proportional to theirgeographic distance (Palumbi, 2003).

Mangrove forests are another example of a fragmented habitat.Along the East African coast mangroves are on average separated byabout 150 km of stretches of sandy or rocky shores and are popu-lated by several crab and mollusc macrobenthonic species strictlyassociated with this environment (Cannicci et al., 2008). For thesevarious intertidal and semi-terrestrial species a double problemarises: how to optimize the possibilities of dispersal along the coastand how to identify and colonize the most suitable habitat. Mostmangrove crabs release larvae in synchrony with spring tides (Skovet al., 2005), probably to be transported by oceanic waters and thusimproving their dispersal capabilities (Wooldridge, 1991; Papado-poulos et al., 2002). In the open oceanic waters larvae pass througha determined number of stages, all of them characterized bya different pattern of vertical migration (called ontogeneticmigration: Shanks, 1995; Queiroga and Blanton, 2005) in the watercolumn. The vertical migration, depending on the larval stage,should guide both the export of larvae to the open oceanic watersas well as the recolonization of the adequate habitat for thesettlement (Shanks, 1995; Queiroga and Blanton, 2005). The lastlarval stage, the megalopa, returns to the mangroves probablyguided by chemical and physical cues (the ‘‘return migration’’:Shanks, 1995; Queiroga and Blanton, 2005).

The aim of the present study was to describe the populationgenetic structure of the mangrove crab Neosarmatium meinerti(Decapoda: Brachyura: Sesarmidae) and to depict its demographichistory within the western Indian Ocean (WIO). We also aimed toshed light on the historical and ecological factors affecting the geneflow, both in terms of direction and entity. N. meinerti occupies therearward belt of mangrove forests and is widespread throughoutthe Indo-Pacific Ocean (Davie, 1994), but genetically highly struc-tured (Ragionieri et al., 2009). Mangrove crabs such as N. meinertithat occupy the middle-upper belt have a seasonally definedbreeding season, conversely to those species living in the lowerintertidal belt which are continuous breeders (Emmerson, 1994).Moreover, the N. meinerti breeding period is also affected by localconditions: in Kenya and Tanzania ovigerous females release larvaeat spring tide every month during the Northwestern Monsoon(between February and March) (Skov et al., 2005), while in SouthAfrica the spawning happens during the summer, thus a littleearlier and for a shorter period than in Kenya and Tanzania, whenthe temperature is warmer (Emmerson, 1994). After the hatching,larvae have a supposed lasting period in the plankton of about fourweeks (Pereyra Lago, 1989), thus being able to potentially inter-connect populations separated by hundreds of kilometers.

To achieve our goals, we investigated the pattern of geneticvariation in the mitochondrial cytochrome oxidase subunit 1(mtDNA CoxI) in six populations of the East African coast and in onepopulation from Seychelles Islands. Two main scenarios can besupposed: (1) Neosarmatium meinerti populations of the WIO area single metapopulation with high level of genetic exchange, oralternatively (2) some barriers to dispersal occur, inducing aninterruption of gene flow between the mainland and island pop-ulations and/or along the continental margin. Different separationscenarios are possible. On the one hand, the Seychelles Islands areseparated from the East African coast by more than 1500 km ofopen sea, a geographic gap that could exceed the distance thelarvae may cover (see Palumbi, 2003). On the other hand, a sepa-ration among tropical (Kenya) and subtropical populations (SouthAfrica) may be induced by local scale currents and ebbs as well asby larval and adult local adaptations to sharp temperature

gradients, as recorded along the South African coast for the estu-arine mud prawn Upogebia Africana by Teske et al. (2008). More-over, in a metapopulation which is latitudinally widespread overa wide range, connectivity should be interrupted due to repro-ductive isolation among populations under different seasonalconditions.

2. Materials and methods

2.1. Main circulation of western Indian Ocean

The main WIO circulation is characterized by the South Equa-torial Current (SEC) which flow from the eastern Indian Oceantoward the western Indian Ocean at around 10� of latitude southbefore splitting into two main currents when reaching the easterncoast of Africa (Fig. 1). One of these currents flows southward andincludes the Mozambique Currents and Madagascar Currents,while the other one flows northward comprising the East AfricaCoastal Current (EACC). It should separate populations of Tanzaniaand Kenya from the populations of Mozambique and South Africa.Even the monsoon plays an important role with its seasonalchanging of the Somali Current direction. During the SoutheastMonsoon, the EACC extends further becoming the Somali Currentand later joining the Indian Monsoon Currents, while during theNortheast Monsoon the northerly flow of the EACC is reduced andturns eastward becoming the Equatorial Counter Current, togetherwith the Somali Current which flows southward (Fig. 1).

2.2. Sample collections and DNA extraction

A total of 167 specimens of Neosarmatium meinerti werecollected along the East African coast from Kenya (Lamu n ¼ 18,Mida Creek n ¼ 26, Gazi Bay n ¼ 27), Tanzania (Dar es Salaamn ¼ 26), Mozambique (Inhaca Island n ¼ 28) and South Africa(Durban n ¼ 22) (Fig. 1). In addition, we sampled a population fromthe Seychelles (Mahe Island n ¼ 20) (Fig. 1). From each specimen,muscle tissue was removed from a walking leg and immediatelyplaced in a vial with absolute ethanol. Subsequently, the specimenswere released into their natural environment. The genomic DNAwas extracted using the Puregene Kit (Gentra System) and thenresuspended in TE or distilled water and stored at �20 �C.

2.3. DNA amplification

A fragment of 658 base pairs (bp) of the cytochrome oxidasesubunit 1 gene (CoxI) was amplified by polymerase chain reaction(PCR) using the following primers: COL6b 50-aca aat cat aaa gat atygg-30 (Schubart and Huber, 2006) and HCO2198 50-taa act tca gggtga cca aaa aat ca-30 (Folmer et al., 1994). The PCR amplification wasperformed in a Perkin Elmer 9600 thermal cycler with thefollowing PCR conditions: 40 cycles with 45 s at 94 �C for dena-turation, 1 min at 48 �C for annealing, 1 min at 72 �C for extension,preceded by 10 min at 94 �C for initial denaturation and 10 min at72 �C for final extension. Subsequently, PCR products were visual-ized on agarose gels, purified by precipitation with Sure Clean(Bioline) and then resuspended in water. The sequence reactionswere performed with the Big Dye terminator mix (Big Dye Termi-nator� V 1.1 Cycle Sequencing kit; Applied Biosystems) followed byelectrophoreses in an ABI Prism automated sequencer (ABI Prism�310 Genetic Analyzer; Applied Biosystems). The sequences werecorrected manually with Flinch TV 1.4.0 (Geospiza�) and 620 bpaligned manually with Bioedit (Hall, 1999).

Fig. 1. Map of collection localities along the East African Coast and western Indian Ocean (from north toward south along the coast: Lamu Archipelago, Mida Creek, Gazi Bay, Dar esSalaam, Inhaca Island and Durban, with Seychelles in front of Kenya) and main oceanic currents of the western Indian Ocean. Dashed lines and circles represent circulation duringthe North East Monsoon of the Somali Current. SC, Somali Current; ECC, Equatorial Counter Current; EACC, East Africa Costal Current; SEC, South Equatorial Current; MaC,Madagascar Current; MC, Mozambique Current; AC, Agulhas Current. Normal line: main oceanic circulation; dashed lines: during Southwest Monsoon; dotted lines: duringNortheast Monsoon.

L. Ragionieri et al. / Estuarine, Coastal and Shelf Science 86 (2010) 179–188 181

2.4. Genetic diversity indexes, population genetic structure, anddemographic history

The nucleotide composition was calculated with MEGA 3.1(Kumar et al., 2004). Genetic heterogeneity within each populationwas estimated as haplotype diversity (h, indicating the probabilitythat two randomly chosen haplotypes are different in a population)and nucleotide diversity (p, indicating the percentage meannumber of differences between all pairs of haplotypes in a pop-ulation) (Nei, 1987), as implemented in ARLEQUIN 3.01 (Excoffieret al., 2005).

Differentiation among populations was estimated by the anal-ysis of molecular variance (AMOVA) (Excoffier et al., 1992) asimplemented in ARLEQUIN (Excoffier et al., 2005). We calculatedthe Fst indices (Excoffier et al., 1992) using haplotypic frequenciesand genetic distances. The distance method we used was the Tajimaand Nei model (Tajima and Nei, 1984), as suggested for unequalnucleotide frequencies. Significance of the fixation indices, underthe null hypothesis of no differentiation among populations, wastested using a non-parametric permutation approach (10,000

permutations of haplotypes among populations). In this case, theP value of the test is the proportion of permutations with Fst valueslarger or equal to the observed one (Excoffier et al., 1992).

The spatial distributions of haplotypes among populations werealso calculated by contingency c2 tables using a Monte Carlosimulation without pooling rare haplotypes (Roff and Bentzen,1989) as implemented in CHIRXC (Zaykin and Pudovkin, 1993).Probability of heterogeneity in haplotypic distribution was esti-mated by comparison between observed c2 and simulated onceobtained from 20,000 random permutations of original data.

A Mantel test was carried out as implemented in ARLEQUIN totest whether genetic divergences (pairwise Fst values, calculatedfrom genetic distance values) were correlated with geographicdistances (calculated for each site pairwise) (Mantel, 1967).Significance was assessed by means of 10,000 permutations of rowand columns of one matrix, keeping the other constant.

The historical demographic history of our populations wasreconstructed applying the mismatch distribution analysis (i.e. thedistribution of the observed number of differences between pairs ofhaplotypes) as implemented in ARLEQUIN and DnaSP 4.10 (Rozas

L. Ragionieri et al. / Estuarine, Coastal and Shelf Science 86 (2010) 179–188182

et al., 2003). Populations at the equilibrium are expected to showa multimodal distribution of haplotype frequencies, while pop-ulations having recently passed through a demographic expansionor a range expansion, are predicted to have an unimodal distribu-tion (Rogers and Harpending, 1992).

Since recent bottlenecks can mimic the effect of a populationexpansion (Ramırez-Soriano et al., 2008), we also applied threeneutrality tests, Tajima’s D test, Fu’s Fs test, and the R2 test in orderto better depict the demographic history of our populations.Ramos-Onsins and Rozas (2002) classified the neutrality tests intothree major classes: Class I statistics use information on themutation frequency and are appropriate for distinguishing pop-ulation growth from constant size population. The Tajima D test(Tajima, 1989a,b) and the R2 test (Ramos-Onsins and Rozas, 2002)belong to this class. The Class II includes tests that use informationfrom the haplotype distribution such as the Fs test (Fu, 1997). TheClass III tests use information from the mismatch distribution suchas the raggedness index rg (Harpending et al., 1993; Harpending,1994) which measures the smoothness of the mismatch distribu-tion. The significance level of the D test and Fs test was estimatedby generating random samples under the null hypothesis ofselective neutrality and population equilibrium using a coalescentsimulation (Hudson, 1990) as implemented in ARLEQUIN. The R2test significance level was estimated using DnaSP 4.10 based on1000 simulated re-sampling replicates, while the rg significancewas tested by a parametric bootstrap approach (10,000 replicates)under the null hypothesis of population expansion as implementedin ARLEQUIN.

We also estimated the expansion parameters Tau (s), Theta0 (q0) and Theta 1 (q1) under a demographic or spatial expansionhypothesis by a generalized non-linear least-square approach(Schneider and Excoffier, 1999). The time (t) at which the demo-graphic or spatial expansion began was calculated applying the Liformula (Li, 1977): t ¼ s/2m, where m is the mutation rate per site peryear (in our case assumed to be 1.66% per million years as in marinesesarmids from Jamaica, see Schubart et al., 1998). Approximateconfidence intervals for the demographic parameters wereobtained by 1000 parametric bootstrap replicates.

2.5. Coalescent based analysis

In addition, we employed two analysis methods based on coa-lescent theory (Kingman, 1982), for investigating the relationshipbetween population genetic structure and geography (i.e. theNested Clade Phylogeographic Analysis, NCPA: Templeton et al.,1995; Templeton, 1998) and for calculating the ongoing gene flowamong predefined groups of populations (i.e. Isolation withMigration, IM: Hey and Nielsen, 2004).

The NCPA is based on a minimum spanning network using TCS1.21 (Clement et al., 2000) with 95% of plausible sets for linkagesof haplotypes. The NCPA nesting design was constructed by handon the resulting statistical parsimony (SP) cladogram, solvingambiguities (such as loops) following the rules derived fromcoalescent theory (Templeton et al., 1992; Crandall and Temple-ton, 1993; Pfenninger and Posada, 2002). The program Geodis 2.6(Posada et al., 2000) was used for calculating the categorical test ofsignificant associations between nested clades and geography,using 10,000 random permutations. For a single-locus nestedclade analysis, the significance of the test is corrected with theDunn–Sidak correction incorporated in Geodis 2.6. Finally thehistorical and current processes responsible for the observedpattern of genetic variation were inferred using the latest versionof the inference key given in Templeton (2004) (updated onDecember 15, 2008).

For investigating the ongoing migration rates along the EastAfrican coast and between the mainland and Seychelles we usedthe program IM (Hey and Nielsen, 2004). Migration rates along thecoast were estimated grouping equatorial populations (Kenya andTanzania) and subtropical populations (Mozambique and SouthAfrica), while the migration rates between inland and mainlandpopulations were estimated considering Seychelles and consid-ering equatorial mainland populations. Whenever the effectivesample size (ESS) values were low and posterior probability diffi-cult to interpret, we forced the effective population sizes to be equalfor both groups of populations, as suggested by Teske et al. (2007).Then we used a HKY model (Hasegawa et al., 1985) and an inheri-tance scalar of 0.25 as suggested for mitochondrial markers (Hey,2007). Finally, the following search strategy was adopted:�b 500,000 �l 0.5 �n20 �k20 �fg �g1 0.95 �g2 0.99 �j 4 (or �j 1whenever possible). For each analysis the IM program was run asfar as the ESS reached a value bigger than 50 for at least three times.For each comparison, migration rate parameters were converted tothe effective number of females migrating per generation, halvingthe product of the population mutation rate (4Nu) by the migrationrate from a population to the other one (m/u), where N is theeffective size of the population, u is the neutral mutation rate forthe gene per generation and m is the migration rate per gene pergeneration from one population into the other, in the coalescent.

3. Results

3.1. Genetic diversity indices, population genetic structure, anddemographic history

Our fragment has an A-T rich nucleotide composition(C ¼ 16.22%; A ¼ 37.17%; T ¼ 29.10%; G ¼ 17.51%), as reported forother arthropod mitochondrial DNA (Simon et al. 1994). From thealigned 620 base pairs, we found 44 variable and 23 parsimony-informative sites, resulting in 36 haplotypes (Appendix A). All buttwo mutations are silent. The overall average number of differencesbetween pairwise haplotypes is k ¼ 5.97, but most haplotypeswithin geographic regions differ from one another by singlemutations (Appendix A). Haplotypes 1–26 and 31–36 are onlypresent within the East African populations, while haplotypes 27–30 are exclusive to the Seychelles population (Table 1); manyhaplotypes are rare (Table 1). Haplotype diversity (h) and nucleo-tide diversity (p) values are reported in Table 1. All populationsexcept for the Seychelles are characterized by high haplotypicdiversity values (>0.65), and low nucleotide diversity indexes(<0.5). In contrast, the Seychelles population has a very reducedgenetic heterogeneity, both whether expressed as haplotypic ornucleotide diversity (Table 1).

The AMOVA tests, both based on haplotypic frequencies andgenetic divergence data, reveal the existence of a strong populationgenetic structure within the study area (haplotypic frequencies:Fst ¼ 0.130, P < 0.001; genetic distance data: Fst ¼ 0.630,P < 0.001). The Fst pairwise values indicate that this structure isdue to the genetic separation of the Seychelles population from allthe others (Table 2). Otherwise, all the pairwise Fst values betweenEast African populations are relatively low and non-significantwhen testing for lack of gene flow (Table 2).

The contingency c2 tables of haplotype frequencies detecta strong genetic differentiation among all the populations(c2 ¼ 398.81, df ¼ 210, P < 0.001). All the pairwise comparisonsregarding the Seychelles population are highly significant (Lamu–Seychelles: c2 ¼ 38, df ¼ 12, P < 0.001; Mida–Seychelles: c2 ¼ 46,df ¼ 14, P < 0.001; Gazi–Seychelles: c2 ¼ 47, df ¼ 15, P < 0.001; Dares Salaam–Seychelles: c2 ¼ 46, df ¼ 10, P < 0.001; Inhaca–Seychelles: c2 ¼ 48, df ¼ 12, P < 0.001; Durban–Seychelles:

Tab

le1

Hap

loty

pe

dis

trib

uti

onw

ith

inth

ese

ven

pop

ula

tion

sof

Neo

sarm

atiu

mm

ein

erti

alon

gth

eEa

stA

fric

aco

ast.

N,n

um

ber

ofsp

ecim

ens

anal

yzed

;N

hap

,nu

mb

erof

hap

loty

pes

for

each

pop

ula

tion

;h

and

p,h

aplo

typ

ed

iver

sity

and

nu

cleo

tid

ed

iver

sity

for

each

pop

ula

tion

;re

spec

tive

ly.

Loca

lity

Hap

loty

pes

NN

hap

hp

12

34

56

78

91

01

11

21

31

41

51

61

71

81

92

02

122

23

24

25

26

27

28

29

30

31

32

33

34

35

36

Lam

u(K

enya

)4

70

00

01

00

10

00

01

00

01

11

10

00

00

00

00

00

00

01

89

0.8

2�

0.0

70

.43�

0.2

7M

ida

Cre

ek(K

enya

)6

70

00

11

00

20

01

14

11

10

00

00

00

00

00

00

00

00

02

61

10

.87�

0.0

40

.47�

0.2

8

Gaz

iB

ay(K

enya

)9

61

11

13

11

11

10

00

00

00

00

00

00

00

00

00

00

00

02

71

20

.85�

0.0

50

.43�

0.2

6D

ares

Sala

am(T

anza

nia

)1

11

00

00

01

00

10

00

01

00

00

00

00

00

00

00

01

10

00

02

67

0.6

9�

0.0

60

.38�

0.2

4

Inh

aca

Isla

nd

(Moz

amb

iqu

e)6

70

00

01

00

70

00

03

00

00

00

01

11

10

00

00

00

00

02

89

0.8

4�

0.0

30

.46�

0.2

7

Du

rban

(Sou

thA

fric

a)4

50

20

00

00

40

00

00

00

00

03

00

00

00

00

00

01

11

12

29

0.8

9�

0.0

40

.47�

0.2

8

Mah

eIs

lan

d(S

eych

elle

s)0

00

00

00

00

00

00

00

00

00

00

00

00

01

71

11

00

00

00

20

40

.28�

0.1

30

.08�

0.0

8

Tot

min

us

Seyc

hel

les

40

42

13

12

71

11

61

11

19

11

11

14

11

11

10

00

01

11

11

11

47

32

0.8

2�

0.0

20

.46�

0.2

6

Tot

40

42

13

12

71

11

61

11

19

11

11

14

11

11

11

71

11

11

11

11

16

73

60

.83�

0.0

20

.45�

0.2

6

L. Ragionieri et al. / Estuarine, Coastal and Shelf Science 86 (2010) 179–188 183

c2 ¼ 42, df ¼ 12, P < 0.001); while the only significant pairwisecomparison between two East African populations is between Dares Salaam and Durban (c2 ¼ 20.74, df ¼ 12, P ¼ 0.006). Moreover,the c2 value after the exclusion of the Seychelles population wasnot significant (c2 ¼ 171.80, df ¼ 155; P ¼ 0.510).

Applying the Mantel test, we find a significant correlationbetween pairwise population Fst values and geographic distances(Z ¼ 12.61, P ¼ 0.02) (Fig. 2). This suggests that a model of isolationby distance (IBD) may explain the recorded genetic structure. Onthe other hand, Fig. 2 gives evidence for the existence of twoseparated clusters of data points, distributed in parallel to the x-axis, invalidating the significance of the test. The points with Fstvalues near to 1 (empty circles) are those corresponding to thepairwise population comparisons between the Seychelles and theEast African populations, while the other points are those corre-sponding to the East African population pairwise comparisons.Applying the Mantel test only to the East African populations, anIBD model (Z ¼ 0.18, P ¼ not significant) is not further supported.

The population demographic history was reconstructed usingthe mismatch distribution analysis and three neutrality tests foreach population separately, for the entire sample, and groupingmainland populations. All the separate populations (data notshown) as well as the mainland group (Fig. 3a) have an unimodalmismatch distribution, as expected for populations that haverecently experienced a demographic or range expansion (Rogersand Harpending, 1992; Ray et al., 2003; Excoffier, 2004), while themismatch distribution calculated for the whole sample hasa bimodal distribution (Fig. 3b). All the applied neutrality testsrevealed significant deviations from mutation-drift equilibrium forthe pooled East African populations (Table 3), while only the Fu’stest resulted negative and significant, if the Seychelles populationwas also included. Considering separately each population, onlytwo (Gazi Bay by the Fu’s test and Seychelles by the Tajima D test)seem to have experienced a recent event of population expansion(Table 3). Moreover, based on the raggedness index, it is notpossible to reject the null hypothesis of populations being out ofequilibrium (Table 3).

Under the assumption of a demographic expansion, the value ofs for the pooled East African populations is 4.58 (95% confidenceinterval: 1.2–7.46), corresponding to a population expansion whichbegan at around 3.8 Mya, while under the spatial expansionhypothesis the value of s decreases to 2.53, corresponding to anexpansion time of approximately 2.09 Mya. The Seychelles pop-ulation is characterized by lower values of the demographicindices: under the assumption of demographic expansion, thevalue of s is 3.0, corresponding to a population expansion starting2.49 Mya, while in case of a spatial expansion the value of s is 1.59,corresponding to a t value of 1.31 Mya.

3.2. Coalescent based analysis

The SP analysis identified two non-connected haplotypenetworks, one including all the haplotypes from the East Africanmainland populations (Mida, Gazi, Dar es Salaam and Durban) andclose islands (Lamu and Inhaca), and the second one, separatedfrom the former by up to 15 mutational steps, including all thehaplotypes from Mahe Island (Fig. 4). Haplotype 2 had the highestroot probability (P ¼ 0.15) and for this it is supposed to be theancestral haplotype (Castelloe and Templeton, 1994). The NCPArecorded significant association between haplotype clades andgeographic distribution in three clades (Table 4). The oldest infer-red event was a past fragmentation or a long distance colonizationbetween all the populations of the East African coast (Clades 3-1and 3-2) and the population of Mahe Island (Clade 3-3). WithinClade 2-1 there was indication for Contiguous Range Expansion.

Table 2Pairwise analysis of molecular variance (AMOVA) calculated among the seven populations of Neosarmatium meinerti along the East Africa coast. The pairwise Fst values werecalculated based on haplotypic frequencies, below the diagonal, and on genetic divergence, above the diagonal. Values with P < 0.05 are reported in bold.

Lamu(Kenya)

Mida Creek(Kenya)

Gazi Bay(Kenya)

Dar es Salaam(Tanzania)

Inhaca Island(Mozambique)

Durban(South Africa)

Mahe Island(Seychelles)

Lamu (Kenya) * �0.02034 �0.00454 �0.03286 �0.00798 �0.01516 0.91746Mida Creek (Kenya) �0.01191 * �0.00258 0.01329 �0.01510 0.00696 0.89900Gazi Bay (Kenya) �0.01903 0.02983 * �0.02797 0.03717 0.04550 0.90661Dar es Salaam (Tanzania) �0.00917 0.01073 �0.00200 * 0.03881 0.02865 0.49671Inhaca Island (Mozambique) 0.00090 �0.00980 0.01870 0.02089 * �0.01306 0.90036Durban (South Africa) �0.00267 0.00609 0.01388 0.04811 0.03881 * 0.40720Mahe Island (Seychelles) 0.45377 0.40586 0.41452 0.91611 0.41479 0.90680 *

L. Ragionieri et al. / Estuarine, Coastal and Shelf Science 86 (2010) 179–188184

While in clade 2-3 the sampling design was inadequate todiscriminate between Isolation by Distance (short distance move-ment) and Long Distance Colonization.

The analysis of the ongoing gene flow along the East Africancoast (using a migration model, mi > 0) supports asymmetricmigration rates, with immigration into equatorial populations(2Neqmeq ¼ 136.5) greatly exceeding immigration into subtropicalpopulations (2Nstmst ¼ 0.2) (Fig. 5a). Otherwise, following theresults of the AMOVA test, we used an Isolation model (mi ¼ 0) fordating back the fragmentation event separating the Seychelles fromthe mainland populations. Unfortunately, we were not able toestimate the qa because the posterior probability densities were flat(Fig. 5b). The time of fragmentation is set during Pleistocene(t ¼ 1.6 Mya), while the time of most recent common ancestorcomes before (TMRCA ¼ 2.09 Mya) as logically supposed by anIsolation model (Wakeley and Hey, 1997; Hey and Nielsen, 2004).

4. Discussion

Neosarmatium meinerti is a widespread species in the Indo-Pacific Ocean and considered to have potentially good dispersalability, having a larval phase lasting about a month and a larvalrelease during the high spring tides. Notwithstanding, within theWIO it shows a sharp genetic break between the East Africanpopulations (including islands close to the mainland) and thepopulation of Mahe Island, Seychelles. These two geographicgroups are highly genetically distinct and appear to be two inde-pendent evolutionary units (Moritz, 1994), historically separatedfrom each other and not sharing any haplotype (complete lineagesorting). This conclusion is strongly supported by the Fst statistic,the c2 analysis based on rare haplotypes, and the NCPA. Also thebimodal shape of the mismatch distribution recorded for thepooled sample points to the presence of two distinct phylogroupsas suggested by Alvarado Bremer et al. (2005) and discussed by

Mantel Test

-0,2

0

0,2

0,4

0,6

0,8

1

0 500 1000 1500 2000 2500 3000 3500 4000

Distances

Φst

Fig. 2. Pairwise population differentiation (Fst) based on nucleotide diversity plottedas a function of the geographic distance. The full colored rumbles are the pairwise Fstvalues corresponding to population along the East Africa Coast, while the emptyrumbles correspond to pairwise Fst between Mahe Island and the rest of populations.

Patarnello et al. (2007). Conversely, along the East African coast,a metapopulation with high levels of gene flow among its sub-populations is present.

The isolation of Seychelles from the mainland may be explainedby different hypotheses. First, we should consider the presence ofsome geographic barriers or of some local gyres preventing larvaldispersal, but we exclude this hypothesis based on a studyregarding the population genetic structure of the portunid crabScylla serrata sampled in the same area, which revealed that theSeychelles population is genetically connected to the East Africancoast (Fratini et al., in press). Secondly, we should hypothesize thatphysiological constraints, acting on Neosarmatium meinerti larvalsurvival as well as active horizontal migrations in the plankton,prevent larvae to cross thousands of kilometers of open ocean,inducing retention near the parental habitat. Finally, we maysuppose that in this crab species connection among WIO islandsand mainland is interrupted for the lack of stepping stone

Fig. 3. Mismatch distribution for the whole population (a), including Seychelles, withobserved and expected distribution under a constant size population model, andmismatch distribution for the East African metapopulation (b) with observed andexpected distribution under a sudden demographic expansion.

Table 3Neutrality tests, Tajima D test (D), R2 test and Fu Fs-test (Fs) for each population, forthe whole population minus Seychelles and for the overall populations. Raggednessindex (rg) for the mismatch distribution. The significant P values are reported inbold.

Locality D Fs R2 rg

Lamu (Kenya) �0.27 �2.50 0.11 0.09Mida Creek (Kenya) 0.06 �2.87 0.12 0.05Gazi Bay (Kenya) �0.50 L4.32 0.09 0.07Dar es Salaam (Tanzania) 0.42 �0.16 0.14 0.19Inhaca Island (Mozambique) 0.04 �1.01 0.12 0.07Durban (South Africa) �0.42 �1.59 0.11 0.04Mahe Island (Seychelles) L1.64 �1.61 0.12 0.42Tot minus Seychelles L1.48 L19.82 0.04 0.05Tot �0.73 L8.9 0.06 0.15

Table 4Results of nested clade analysis with Chi-square test of geographical association ofclades. Results of non-significant tests have been omitted. P, probability of obtaininga c2-statistic larger than or equal to the observed statistic by randomly permutingthe original contingency table 10,000 (�1) times. Inferences were obtainedfollowing the latest key given in Templeton (2004). Abbreviations for the inferenceare: CRE, continuous range expansion; IbD, isolation by distance; LDC, long distancecolonization; LDD, long distance dispersal.

Clades nestedwith

Permutationalc2 statistic

P Chain ofinference

Inference

Clade 2-1 23.66 0.005 2-11-12-NO CREClade 2-3 26.97 0.019 2-3-5-6-7-8-NO IbD or LDDEntire cladogram 170.66 0.000 2-3-5-15-NO Past fragmentation

or LDC

L. Ragionieri et al. / Estuarine, Coastal and Shelf Science 86 (2010) 179–188 185

populations, since the distance between the two farthest pop-ulations placed along the coast (i.e. Lamu Archipelago and Durban)is greater than that of Seychelles Islands from Africa mainland(about 4500 km versus 2000 km, respectively). This is also inagreement with what was reported by Palumbi (2003) and Shankset al. (2003) calculating the mean dispersal distance in marineinvertebrates ranging between 25 km and 150 km, a distanceconsiderably lower than that which separates Seychelles from theEast Africa. In our opinion, this latter hypothesis is the most plau-sible, since it is also allow us to explain the presence of a meta-population widespread from Northern Kenya to South Africa: EastAfrican mangroves, in fact, are patchily distributed and on averageseparated by no more than 150 km.

Considering in detail the population genetic structure along theEast African coast, gene flow, as estimated by Fst values, does notseem to be obstructed either by the recorded temporal shift in thespawning season between tropical and subtropical latitudes(Emmerson, 1994; Skov et al., 2005), nor by the splitting of theSouth Equatorial Current off Madagascar into a northward branch(the East African Coastal Current) and a southward one (theMozambique Current). Otherwise, the analysis of the ongoing geneflow by a migration model underlines a strong asymmetric gene

Fig. 4. Statistical parsimony cladogram with complete nesting design. The square shape hapthe frequencies in the populations.

exchange into equatorial populations from subtropical ones. Thus,notwithstanding the presence of an East African metapopulation,this result suggests the presence of shallow biogeographic barrierspreventing southward larval dispersal. Such a barrier may berelated to local adaptations of larvae to specific environmentalfactors such as coastal water temperaturedas also suggested byTeske et al., 2008 for some Uca spp.dand/or some physical factorssuch as local upwelling systems, near shore currents and salinitygradient.

The demographic history of the East African populationssupports a sudden expansion from a small initial population(Hudson, 1990; Rogers and Harpending, 1992). Populations thathave recently experienced a contraction or bottleneck have largepositive values of the Tajima D parameter (Tajima, 1989a,b), whilea recent population expansion leads to low values for R2 and largeand negative Fs values (Fu, 1997). Moreover, Ramırez-Soriano et al.(2008) recently highlighted that strong bottlenecks tend to behaveas expansions, while weak bottleneck as contractions. Thus for theEast African metapopulation the results of the neutrality testssupport the hypothesis of a recent population expansion. Aconfirmation of a recent event of population expansion of thismetapopulation also comes from the high values of haplotypic andnucleotidic diversity indices (Grant and Bowen, 1998), as well as

lotype is the supposed ancestral haplotype. The size of haplotypes is proportionate to

0

0.001

0.002

0.003

0.004

0.005

0.006

0.007

0.008

0.009

1 61 121 181 241 301 361 421 481 541 601 661 721 781 841 901 961

mt

θ

0

0.005

0.01

0.015

0.02

0.025

0.05

4.358.65 1317.3

21.6

25.9

30.2

34.5

38.8

43.1

47.4

51.7 56

60.3

64.6

68.9

73.2

77.5

81.8

86.1

90.4

94.7 99

Post

erio

r pr

obab

ility

0

0.005

0.01

0.015

0.02

0.025

0.03

0.035

0.47 4181.6122

163

203

244

284

325

366

406

447

487

528

568

609

650

690

731

771

812

852

893

934

θeq

θst

mst

meq

θSey

θeq

θa

0

0.0005

0.001

0.0015

0.002

0.0025

0.003

0.0035

0.004

0.0045

0.021.31 2.6

3.895.186.477.769.0510.3

11.6

12.9

14.2

15.5

16.8

18.1

19.4

20.7

21.9

23.2

24.5

25.8

27.1

28.4

29.7

θ

a b

Fig. 5. Posterior probability plots of the parameters q (population mutation rate), m (migration rate) and t (time since splitting) under a migration model between the equatorial andsubtropical group of populations (a) and under an isolation model between equatorial and Seychelles populations (b).

L. Ragionieri et al. / Estuarine, Coastal and Shelf Science 86 (2010) 179–188186

from the NCPA that supports a contiguous range expansion froma small size initial population (gradual moving front of rangeexpansion) in this geographic area.

Based on the demographic parameters of the mismatchdistribution analysis, under a pure demographic expansionmodel, the East African populations seem to have expandedabout 3.8 Mya, while under a spatial expansion model this wouldhave occurred around 2.07 Mya. This tentative time frameencompasses a period of great climatic changes along the EastAfrican coast and therefore appears more than plausible. Thecloser of the Indonesia Gateway occurred around 3–4 Mya andproduced great climatic changes in WIO (Cane and Molnar, 2001;Gourlan et al., 2008). Moreover, successively during the Plio-Pleistocene, the Indian Monsoon repeatedly changed in intensityand phase, and it should have repercussion on the oceaniccirculation of the WIO (Clemens et al., 1996).

Conversely, considering the Seychelles population, themismatch distribution and the Tajima D test suggest a recentbottleneck event. In addition, the low values of haplotypic andgenetic diversity are consistent with a recent prolonged or severedemographic bottleneck (Grant and Bowen, 1998), followed byexpansion. Also, the NCPA explained the isolation of Seychellesfrom the East African coast as due to a past fragmentation event(before which we can suppose a unique metapopulation wide-spread throughout the western Indian Ocean) or as the conse-quence of long distance colonization (that should have occurred insynchrony with the Pleistocenic sea-level fluctuations). At the lightof the results of Ragionieri et al. (2009), indicating that other WIOislands (Aldabra, Mauritius and Rodriguez) belong to theSeychelles’ evolutionary unit, the former hypothesis seems to bethe most plausible. Conversely, in case of episodic LDC, we shouldexpect the presence of different evolutionary units in the WIOislands, each of them characterized by a reduced number of sharedhaplotypes. Finally, by means of the mismatch distribution analysis,our investigation for providing a time frame for the last expansionevent of Seychelles suggests a Plio-Pleistocene context. This timeframe corresponds to those calculated by the Isolation Model: the

TMRCA is set during late Pliocene (2.01 Mya) and the splitting timebetween mainland populations and Seychelles results to be1.6 Mya. Such results are in agreement with previous splitting timeestimated for Neosarmatium meinerti populations from other WIOIslands and East African coast (Ragionieri et al., 2009). This timeestimated is highly plausible, since numerous sea level fluctuationsoccurred during this period and are recognized to have causedpopulation size fluctuations in many marine species (Hewitt 2000).Moreover, concerning our study area, traces of recent expansionshave also been recorded in two other decapod species, the tigerprawn Penaeus monodon (Duda and Palumbi, 1999; Forbes et al.,1999; Benzie et al., 2002; You et al., 2008) and the portunid crabScylla serrata (Gopurenko et al., 1999; Fratini et al., in press).

5. Conclusions

According to our own results as well as those of Ragionieri et al.(2009), Neosarmatium meinerti populations of the western IndianOcean comprised a single metapopulation during the late Pliocene.Such metapopulation underwent a fragmentation event during theearly Pleistocene with the consequent formation of two evolu-tionary significant units, the East African mainland metapopulationand the western Indian Ocean island group (including Seychelles,Mauritius, Rodriguez and Aldabra). Nowadays gene flow betweenthese two groups appears to be completely interrupted, probablyfor the lack of stepping stone populations between the Africanmainland and western Indian Ocean islands.

On the contrary, gene flow along the East African coast does notseem to be obstructed neither by the main western Indian Oceancirculations nor by temporal shift in the spawning season betweenequatorial and subtropical latitudes. Otherwise we recorded strongasymmetrical gene flow between equatorial and sub-tropicalpopulations probably due temperature and salinity gradients (thusinducing local adaptation of larvae and adults) as well as localcoastal currents. Our results highlight that even in species withhigh potential dispersal and without population genetic structure,biogeographic boundaries may still result in strong asymmetries in

L. Ragionieri et al. / Estuarine, Coastal and Shelf Science 86 (2010) 179–188 187

gene flow among populations which, in time, could end up in soundgenetic separation. Indeed, the presence of boundaries or barriersin the marine realm, denied until the recent past, has been clearlydemonstrated in several genetic studies (see Knowlton, 2000), withthe consequent growing in the number of cryptic and siblingspecies (Pfenninger and Schwenk, 2007). Within this debate, thediscovery of two evolutionary units within Neosarmatium meinerti,which are genetically completely separate within the WIO, enforcethe new idea in evolutionary marine biology that allopatric speci-ation in the marine realm is a likely phenomenon, more commonthan what was believed for many years.

Acknowledgments

We would like to thank two anonymous reviewers for theiruseful suggestions. We would also like to thank Isabelle Papado-pulos and Brent Newman for the help in collecting specimens inSouth Africa. Many thanks to Silke Reuschel, Peter Koller, TobiasSantl and Sara Mannini for invaluable help with computerprograms and laboratory work. Mariella Baratti, Fabrizio Bartoliniand Giuseppe Messana were invaluable in collecting samples. Wealso express our gratitude to the M. Goti and M. Gori of theC.I.B.I.A.C.I. (University of Florence) for sequencing analysis. Specialthank is due to Shannon Johnson for introducing the first author tothe IM model and especially for always being helpful in explaininghow to run IM simulations. Many thanks to Tessa Marikova forlanguage revision. Part of this work was carried out using theresources of the Computational Biology Service Unit from CornellUniversity which is partially funded by Microsoft Corporation. Thisstudy was financed by the University of Florence research funds, bya FIRB project (Italian MIUR), and from Fondi d’Ateneo to S. Cannicciand M. Vannini (ex 60% University of Florence).

Appendix. A

Supplementary information for this article can be accessedonline at doi:10.1016/j.ecss.2009.11.002

References

Alvarado Bremer, J.R., Vinas, J., Mejuto, J., Ely, B., Pla, C., 2005. Comparative phylo-geography of Atlantic bluefin tuna and swordfish: the combined effects ofvicariance, secondary contact, introgression, and population expansion on theregional phylogenies of two highly migratory pelagic fishes. Molecular Phylo-genetics and Evolution 36, 169–187.

Becker, B.J., Levin, L.A., Fodrie, F.J., McMillan, P.A., 2007. Complex larval connectivitypatterns among marine invertebrate populations. Proceedings of the NationalAcademy of Sciences 104, 3267.

Benzie, J.A.H., Ballment, E., Forbes, A.T., Demetriades, N.T., Sugama, K., Moria, S.,2002. Mitochondrial DNA variation in Indo-Pacific populations of the giant tigerprawn, Penaeus monodon. Molecular Ecology 11, 2553–2569.

Bilton, D.T., Paula, J., Bishop, J.D.D., 2002. Dispersal, genetic differentiation andspeciation in estuarine organisms. Estuarine, Coastal and Shelf Science 55,937–952.

Cane, M.A., Molnar, P., 2001. Closing of the Indonesian seaway as a precursor to eastAfrican aridification around 3–4 million years ago. Nature 411, 157–162.

Cannicci, S., Burrows, D., Fratini, S., Smith, T.J., Offenberg, J., Dahdouh-Guebas, F.,2008. Faunal impact on vegetation structure and ecosystem function inmangrove forests: A review. Aquatic Botany 89, 186–200.

Castelloe, J., Templeton, A.R., 1994. Root probabilities for intraspecific gene treesunder neutral coalescent theory. Molecular Phylogenetics and Evolution 3,102–113.

Clemens, S.C., Murray, D.W., Prell, W.L., 1996. Nonstationary phase of the Plio-Pleistocene Asian Monsoon. Science 274, 943–948.

Clement, M., Posada, D., Crandall, K.A., 2000. TCS: a computer program to estimategene genealogies. Molecular Ecology 9 1667–1659.

Cowen, R.K., Lwiza, K.M.M., Sponaugle, S., Paris, C.B., Olson, D.B., 2000. Connectivityof marine populations: Open or closed? Science 287, 857–859.

Crandall, K.A., Templeton, A.R., 1993. Empirical tests of some predictions fromcoalescent theory with applications to intraspecific phylogeny reconstruction.Genetics 134, 959–969.

Davie, P.J.F., 1994. Revision of Neosarmatium Serene and Soh (Crustacea: Brachyura:Sesarminae) with descriptions of two new species. Memoris of the QueenslandMuseum 35, 35–74.

DiBacco, C., Levin, L.A., Sala, E., 2006. Connectivity in marine ecosystems: theimportance of larval and spore dispersal. In: Crooks, K.R., Sanjayan, M. (Eds.),Connectivity conservation. Cambridge University Press, pp. 184–212.

Duda Jr., T.F., Palumbi, S.R., 1999. Population structure of the black tiger prawn,Penaeus monodon, among western Indian Ocean and western Pacific pop-ulations. Marine Biology 134, 705–710.

Emmerson, W.D., 1994. Seasonal breeding cycles and sex ratios of eight species ofcrabs from Mgazana, a mangrove estuary in Transkei, Southern Africa. Journalof Crustacean Biology 14, 568–578.

Excoffier, L., 2004. Patterns of DNA sequence diversity and genetic structure aftera range expansion: lessons from the infinite-island model. Molecular Ecology13, 853–864.

Excoffier, L., Smouse, P.E., Quattro, J.M., 1992. Analysis of molecular variance infer-red from metric distances among DNA haplotypes: application to humanmitochondrial DNA restriction data. Genetics 131, 479–491.

Excoffier, L., Laval, L.G., Schneider, S., 2005. Arlequin, version 3: An integratedsoftware package for population genetics data analysis. Evolutionary Bio-informatics. Online 1, 47–50.

Fairweather, P.G., 1991. Implications of ‘‘supply-side’’ ecology for environmentalassessment and management. Trends in Ecology & Evolution 6, 60–63.

Folmer, O., Black, M., Hoeh, W., Lutz, R., Vrijenhoek, R., 1994. DNA primers for ampli-fication of mitochondrial cytochrome c oxidase subunit I from diverse metazoaninvertebrates. Molecular Marine Biology and Biotechnology 3 (5), 294–299.

Forbes, A.T., Demetriades, N.T., Benzie, J.A.H., Ballment, E., 1999. Allozymefrequencies indicate little geographic variation among stocks of giant tigerprawn Penaeus monodon in the south-west Indian Ocean. South African Journalof Marine Science 21, 271–277.

Fratini, S., Ragionieri, L., Cannicci, S. Stock structure and demographic history of theIndo-West Pacific mud crab Scylla serrata. Estuarine, Costal and Shelf Science, inpress, doi:10.1016/j.ecss.2009.10.009.

Fu, Y.X., 1997. Statistical tests of neutrality of mutations against population growth,hitchhiking and background selection. Genetics 147, 915–925.

Gopurenko, D., Hughes, J.M., Keenan, C.P., 1999. Mitochondrial DNA evidence forrapid colonisation of the Indo-West Pacific by the mudcrab Scylla serrata.Marine Biology 134, 227–233.

Gourlan, A.T., Meynadier, L., Allcgre, C.J., 2008. Tectonically driven changes in theIndian Ocean circulation over the last 25 Ma: Neodymium isotope evidence.Earth and Planetary Science Letters 267, 353–364.

Grant, W.S., Bowen, B.W., 1998. Shallow population histories in deep evolutionarylineages of marine fishes: insight from sardines and anchovies and lessons forconservation. The Journal of Heredity 89 (5), 415–426.

Hall, T.A., 1999. BioEdit: a user-friendly biological sequence alignment editor and anal-ysis program for Windows 95/98/NT. Nucleic Acids Symposium Series 41, 95–98.

Harpending, H.C., 1994. Signature of ancient population growth in a low-resolutionmitochondrial DNA mismatch distribution. Human Biology 66, 591–600.

Harpending, H.C., Sherry, S.T., Rogers, A.R., Stoneking, M., 1993. The genetic struc-ture of ancient human populations. Current Anthropology 34, 483–496.

Hasegawa, M., Kishino, H., Yano, T., 1985. Dating the human-ape split by a molecularclock of mitochondrial DNA. Journal of Molecular Evolution 22, 160–174.

Hewitt, G., 2000. The genetic legacy of the Quaternary ice ages. Nature 405, 907–913.Hey, J., 2007. Using the IM program. http://lifesci.rutgers.edu/wheylab/

ProgramsandData/Programs/IM/Using_IM_3_5_2007.pdfHey, J., Nielsen, R., 2004. Multilocus methods for estimating population sizes,

migration rates and divergence time, with applications to the divergence ofDrosophila pseudoobscura and D. persimilis. Genetics 167, 747–760.

Hudson, R.R., 1990. Gene genealogies and the coalescent process. In: Futuyama, D.,Antonovics, J.D. (Eds.), Oxford Surveys in Evolutionary Biology. University Press,Oxford, pp. 1–44.

Kimura, M., Weiss, G.H., 1964. The stepping stone model of population structureand the decrease of genetic correlation with distance. Genetics 49, 561–576.

Kingman, J.F.C., 1982. The coalescent. Stochastic Processes and their Applications 13,235–248.

Knowlton, N., 2000. Molecular genetic analyses of species boundaries in the sea.Hydrobiologia 420, 73–90.

Kumar, S., Tamura, K., Nei, M., 2004. MEGA3: Integrated software for molecularevolutionary genetics analysis and sequence alignment. Briefings in Bio-informatics 5, 150–163.

Levin, L.A., 2006. Recent progress in understanding larval dispersal: new directionsand digressions. Integrative and Comparative Biology 46, 282–297.

Li, W.H., 1977. Distribution of nucleotide differences between two randomly chosencistrons in a finite population. Genetics 85, 331–337.

Mantel, N., 1967. The detection of disease clustering and a generalized regressionapproach. Cancer Research 27, 209–220.

Moritz, C., 1994. Defining ‘Evolutionarily Significant Units’ for conservation. Trendsin Ecology & Evolution 9, 373–375.

Nei, M., 1987. Molecular Evolutionary Genetics. Columbia University Press, NewYork, 512 pp.

Palumbi, S.R., 2001. The ecology of marine protected areas. In: Bertness, M.D.,Gaines, M.D., Hay, M.E. (Eds.), Marine Community Ecology. Sinauer Associates,Sunderland, MA, pp. 509–530.

Palumbi, S.R., 2003. Population genetics, demographic connectivity, and the designof marine protected reserves. Ecological Applications 13, 146–158.

L. Ragionieri et al. / Estuarine, Coastal and Shelf Science 86 (2010) 179–188188

Papadopoulos, I., Wooldridge, T.H., Newman, B.K., 2002. Larval life history strategiesof sub-tropical southern African estuarine brachyuran crabs and implicationsfor tidal inlet management. Wetlands Ecology and Management 10, 249–256.

Patarnello, T., Volckaert, F., Castilho, R., 2007. Pillars of Hercules: is the Atlantic-Medi-terranean transition a phylogeographical break? Molecular Ecology 16, 4426–4444.

Pereyra Lago, R., 1989. The larval development of the red mangrove crab Sesarmameinerti de Man (Brachyura: Grapsidae) reared in the laboratory. South AfricanJournal of Zoology 24, 199–211.

Pfenninger, M., Posada, D., 2002. Phylogeographic history of the land snail Candi-dula unifasciata (Helicellinae, Stylommatophora): fragmentation, corridormigration, and secondary contact. Evolution 56 (9), 1776–1788.

Pfenninger, M., Schwenk, K., 2007. Cryptic animal species are homogeneouslydistributed among taxa and biogeographical regions. BMC Evolutionary Biology7, 121.

Posada, D., Crandall, K.A., Templeton, A.R., 2000. GeoDis: a program for the cladisticnested analysis of the geographical distribution of genetic haplotypes. Mole-cular Ecology 9, 487–488.

Queiroga, H., Blanton, J., 2005. Interactions between behaviour and physical forcingin the control of horizontal transport of decapod crustacean larvae. Advances inMarine Biology 47, 107–214.

Ragionieri, L., Fratini, S., Vannini, M., Schubart, C.D., 2009. Phylogenetic andmorphometric differentiation reveal geographic radiation and pseudo-crypticspeciation in a mangrove crab from Indo-West Pacific. Molecular Phylogeneticsand Evolution 52, 825–834.

Ramırez-Soriano, A., Ramos-Onsins, S.E., Rozas, J., Calafell, F., Navarro, A., 2008.Statistical Power Analysis of neutrality tests under demographic expansions,contractions and bottlenecks with recombination. Genetics 179, 555–567.

Ramos-Onsins, S.E., Rozas, J., 2002. Statistical properties of new neutrality testsagainst population growth. Molecular Biology and Evolution 19, 2092–2100.

Ray, N., Currat, M., Excoffier, L., 2003. Intra-deme molecular diversity in spatiallyexpanding populations molecular biology and evolution. Molecular Biology andEvolution 20 (1), 76–86.

Roff, D.A., Bentzen, P., 1989. The statistical analysis of mitochondrial DNA poly-morphisms: Chi2 and the problem of small samples. Molecular Biology andEvolution 6 (5), 539–545.

Rogers, A.R., Harpending, H., 1992. Population growth makes waves in the distribu-tion of pairwise genetic distances. Molecular Biology and Evolution 9, 552–569.

Rozas, J., Sanchez-DelBarrio, J.C., Messeguer, X., Rozas, R., 2003. DnaSP, DNA poly-morphism analyses by the coalescent and other methods. Bioinformatics 19,2496–2497.

Schneider, S., Excoffier, L., 1999. Estimation of past demographic parameters fromthe distribution of pairwise differences when the mutation rates vary amongsites application to human mitochondrial DNA. Genetics 152, 1079–1089.

Schubart, C.D., Huber, M.G.J., 2006. Genetic comparisons of German populations ofthe stone crayfish, Austropotamobius torrentium (Crustacea: Astacidae). BulletinFrancais de la Peche et de la Pisciculture 380, 1019–1028.

Schubart, C.D., Diesel, R., Hedges, S.B., 1998. Rapid evolution to terrestrial life inJamaican crabs. Nature 393, 363–365.

Shanks, A.L., 1995. Mechanisms of cross-shelf dispersal of larval invertebrates andfishes. In: McEdward, L. (Ed.), Ecology of Marine Invertebrate Larvae. CRC Press,Boca Raton, FL, pp. 323–365.

Shanks, A., Grantham, B.A., Carr, M.H., 2003. Propagule dispersal distance and thesize and spacing of marine reserves. Ecological Applications 13 (1), 159–169.

Simon, C., Frati, F., Beckenbach, A., Crespi, B., Liu, H., Flook, P., 1994. Evolution,weighting, and phylogenetic utility of mitochondrial gene sequences anda compilation of conserved polymerase chain reaction primers. Annales of theEntomological Society of America 87, 651–701.

Skov, M.W., Hartnoll, R.G., Ruwa, R.K., Shunula, J.P., Vannini, M., Cannicci, S., 2005.Marching to a different drummer: crabs synchronize reproduction to a 14-month lunar-tidal cycle. Ecology 86, 1164–1171.

Tajima, F., 1989a. Statistical method for testing the neutral mutation hypothesis byDNA polymorphism. Genetics 123, 585–595.

Tajima, F., 1989b. The effect of change in population size on DNA polymorphism.Genetics 123, 597–601.

Tajima, F., Nei, M., 1984. Estimation of evolutionary distance between nucleotidesequences Molecular Biology and Evolution 1 (3), 269–285.

Templeton, A.R., 1998. Nested clade analyses of phylogeographic data: testinghypotheses about gene flow and population history. Molecular Ecology 7,381–397.

Templeton, A.R., 2004. Statistical phylogeography: methods of evaluating andminimizing inference errors. Molecular Ecology 13, 789–809.

Templeton, A.R., Crandall, K.A., Sing, C.F., 1992. A cladistic analysis of pheno-typic associations with haplotypes inferred from restriction endonucleasemapping and DNA sequence data III. Cladogram estimation. Genetics 132,619–633.

Templeton, A.R., Routman, E., Phillips, C.A., 1995. Separating population structurefrom population history: a cladistic analysis of the geographical distribution ofmitochondrial DNA haplotypes in the tiger salamander, Ambystoma tigrinurn.Genetics 140, 767–782.

Teske, P.R., Papadopoulos, I., Zardi, G.I., McQuaid, C.D., Edkins, M.T., Griffiths, C.L.,Barker, N.P., 2007. Implications of life history for genetic structure and migra-tion rates of southern African coastal invertebrates: planktonic, abbreviated anddirect development. Marine Biology 152 (3), 697–711.

Teske, P.R., Papadopoulos, I., Newman, B.K., Dworschak, P.C., McQuaid, C.D.,Barker, N.P., 2008. Oceanic dispersal barriers, adaptation and larval retention:an interdisciplinary assessment of potential factors maintaining a phylogeo-graphic break between sister lineages of an African prawn. BMC EvolutionaryBiology 8, 341.

Wakeley, J., Hey, J., 1997. Estimating Ancestral Population Parameters. Genetics 145,847–855.

Wooldridge, T.H., 1991. Exchange of two species of decapod larvae across anestuarine mouth inlet and implications of anthropogenic changes in thefrequency and duration of mouth closure. South African Journal of Science87, 519–525.

Wright, S., 1940. Breeding structure of populations in relation to speciation. TheAmerican Naturalist 74, 232–248.

You, E.M., Chiu, T.S., Liu, K.F., Tassanakajon, A., Klinbunga, S., Triwitayakorn, K., de laPena, L.D., Li, Y., Yu, H.T., 2008. Microsatellite and mitochondrial haplotypediversity reveals population differentiation in the tiger shrimp Penaeus mono-don in the Indo-Pacific region. Animal Genetics 39, 267–277.

Zaykin, D.V., Pudovkin, A.I., 1993. Two programs to estimate significance of Chi2

values using pseudo-probability tests. Journal of Heredity 84, 152.