Pre-zygotic factors best explain reproductive isolation between the hybridizing species of brittle-stars Acrocnida brachiata and A. spatulispina (Echinodermata: Ophiuroidea)

This article was published in the above mentioned Springer issue.The material, including all portions thereof, is protected by copyright;all rights are held exclusively by Springer Science + Business Media.

The material is for personal use only;commercial use is not permitted.

Unauthorized reproduction, transfer and/or usemay be a violation of criminal as well as civil law.

ISSN 0016-6707, Volume 138, Number 6

Pre-zygotic factors best explain reproductive isolation betweenthe hybridizing species of brittle-stars Acrocnida brachiataand A. spatulispina (Echinodermata: Ophiuroidea)

D. Muths • D. Davoult • M. T. Jolly •

F. Gentil • D. Jollivet

Received: 23 April 2009 / Accepted: 8 February 2010 / Published online: 7 March 2010

� Springer Science+Business Media B.V. 2010

Abstract The two brittle-stars Acrocnida brachiata

(Montagu 1804) and A. spatulispina (Stohr and Muths in J

Mar Biol Assoc, 2009) exhibit strong spatial segregation

along the coast of Brittany (France), the first being subti-

dally distributed relative to the other intertidal species.

Despite a very high degree of mitochondrial DNA diver-

gence, previous preliminary results hinted at the potential

for hybridization to occur. Therefore, we specifically aim

to determine local levels of hybridization between these

two species and to investigate the relative roles of pre- and

post- zygotic isolation processes acting to decrease local

hybridization patterns. Mitochondrial DNA, allozymes and

the Internal Transcribed Spacer 2 region of the ribosomal

DNA were all used on 529 brittle-stars sampled locally in

June and September 2005, among six stations in Dou-

arnenez Bay, a site situated at the tip of Brittany. Only

2.6% of all samples analyzed were identified as potential

hybrids. However, these were twice more frequent in June,

just after the reproductive period, than in September after

selective mortality acted to reduce the proportions of

hybrids. In addition to the abrupt bathymetric segregation

between the two species, spawning asynchrony also clearly

restricts hybridization to low levels, which shows the

importance of pre-zygotic mechanisms in maintaining

reproductive isolation. Moreover, both limited hybridiza-

tion events and adult mortalities following reproduction

tend to generate local genetic differentiation at the intra-

species level. On the contrary, the genetic structure is

homogenized by migration of juveniles or adults and

hybrids mortalities over the summer period.

Keywords Hybridization � Migration � Mortality �Spatio-temporal isolation

Introduction

Genetic studies using echinoderms have highlighted the

influence of larval dispersal on patterns of population

structure in the marine environment (e.g. Palumbi et al.

1997; Arndt and Smith 1998), led to calibrate gene-specific

rates of molecular evolution (Bermingham and Lessios

1993) and to delineate species boundaries (Knowlton

1993). In the latter case, such studies have either enabled

the recognition of synonymous taxa (the sea cucumber

Cucumaria sp. Arndt et al. 1996 and the sea star Patiriella

sp. O’Loughlin et al. 2002) or the identification of species

complexes within otherwise morphologically described

species (the sea star Parvulastra exigua Hart et al. 2006,

the sea urchin Echinocardium cordatum, Chenuil and Feral

2003 and the brittle-star Amphipholis squamata, Boissin

et al. 2008). Similarly, a phylogeographic study of the

brittle-star Acrocnida brachiata (Muths et al. 2006) pre-

sented evidence for the existence of two potentially cryptic

D. Muths � D. Davoult � F. Gentil � D. Jollivet

UPMC Univ Paris 6, UMR 7144, BP 74, 29680 Roscoff cedex,

France

D. Muths � D. Davoult � F. Gentil � D. Jollivet

CNRS UMR 7144, Station Biologique de Roscoff, BP 74, Place

Georges Teissier, 29682 Roscoff cedex, France

M. T. Jolly

The Marine Biological Association of the United Kingdom,

The Laboratory, Citadel Hill, Plymouth, Devon PL1 2PB, UK

Present Address:D. Muths (&)

IFREMER Delegation Reunion, Rue Jean Bertho,

97822 Le Port cedex, France

e-mail: [email protected]

123

Genetica (2010) 138:667–679

DOI 10.1007/s10709-010-9441-4 Author's personal copy

species with a clear bathymetric ‘‘intertidal versus subtid-

al’’ segregation at sympatric locations along the western

entrance of the English Channel. A morphological inves-

tigation (Stohr and Muths 2009) recently assessed the

status of A. brachiata as previously described by Montagu

(1804) and actually diagnosed the original type as the

subtidal species. The intertidal lineage was characterized as

a new species named Acrocnida spatulispina in reference

to its spatula-shaped arm spines which is one of the iden-

tification criteria.

Both Acrocnida species are common dwellers of soft

sediments habitats along the North Western European

coastlines. A. brachiata is widely distributed in the

Northeast Atlantic, from the Mediterranean to the shelf

seas around the British Isles, but is never encountered

above 56�N (Koehler 1921). It is often observed as

densely populated aggregates among subtidal habitats

(1,000 ind.m-2 in Ireland, Keegan and Mercer 1986,

900 ind.m-2 along the coast of Normandy, Gentil and

Zakardjian 1989). On the contrary, the presence of

A. spatulispina is seemingly restricted to the intertidal zone

of the Irish sea and the Brittany coastline, where popula-

tions are more sparsely distributed (fewer than

100 ind.m-2 in Brittany, Bourgoin 1987; D. Muths, per-

sonal observation). Despite strong differences in habitat

preferences, it is unlikely that diversifying selective pro-

cesses were directly implicated in the initial separation of

the two lineages. Their emergence is reportedly linked to

large scale vicariance that affected cladogenetic events

similarly among a variety of boreal species belonging to

muddy-fine sediment communities (Jolly et al. 2006;

Muths et al. 2006). This vicariance hypothesis as opposed

to ecotypic speciation is supported by two observations

favouring the secondary contact hypothesis. Firstly,

bathymetric segregation of the two species may be differ-

ent since sampling undertaken in the intertidal zone in

south Brittany (Forest Bay) clearly identified the species

occurring subtidally at the western entrance of the English

Channel (i.e. Acrocnida brachiata). Secondly, neither

species exhibit total reproductive isolation as few individ-

uals displayed a mitochondrial haplotype of one species

and the nuclear background of the other.

How and in which frequency such evolutionary lineages

are able to hybridize despite deep mitochondrial diver-

gence (19.6% based on the COI gene, Muths et al. 2006)

have yet to be determined. In this context, it is important to

note that both temporal and spatial differences in spawning

patterns are often cited as one of the primary mechanisms

contributing to pre-zygotic isolation between closely-

related, broadcast-spawning marine species (e.g. Palumbi

1994). This therefore makes it essential to understand the

extent to which spatial restriction of Acrocnida sp. adults

and temporal isolation may act to limit cross-mating but

still maintain incomplete reproductive isolation between

the two species. Generally speaking, recruitment processes

among Acrocnida sp. populations near the western entrance

of the English Channel (Douarnenez Bay, see Fig. 1) have

been relatively well-studied mainly because of their geo-

graphic location and high densities (Bourgoin et al. 1991).

Moreover, the western entrance of the English Channel

might represent a tension zone between the two cryptic

species, with the potential for localised hybridization

(Muths et al. 2006). In fact, not only does the Brittany

coastline lie within a transition zone between boreal

Fig. 1 Geographic distribution

of the two species along

European coasts (according to

Muths et al. 2006 and Stohr and

Muths 2009). Acrocnidaspatulispina and A. brachiataare respectively represented by

black and white items. Squaresand circles represent

respectively intertidal and

subtidal sampling sites. Rightmap: the location of the six

sampling stations in

Douarnenez Bay, along two

transects A and B. Each transect

included one station in the

intertidal zone (zero meter

depth: 0A and 0B) and two in

the subtidal zone, at ten and

thirty meters depth (10A, 10B,

30A, 30B)

668 Genetica (2010) 138:667–679

123

Author's personal copy

(cold-temperate) and lusitanian (warm-temperate) assem-

blages (Cabioch 1968) but the occurrence of tension zones

with mosaically distributed hybrid zones (Mytilus sp.:

Bierne et al. 2003) sharp genetic breaks associated with

contemporary hydrodynamic features (Pectinaria koreni:

Jolly et al. 2005) and habitat suitability (Hydrobia sp.:

Wilke and Pfenninger 2002) have also all been documented.

In addition to the evidence of highly divergent lineages,

Muths et al. (2006) demonstrated that the very short larval

phase (no more than 4 days) largely restricted gene flow

within each species, even between neighbouring popula-

tions. No clear pattern of isolation-by-distance was detec-

ted at macro-geographic scales, which may be an

indication of the predominant role of local scale processes

in masking larger scale genetic patterns (Hellberg et al.

2002). As such, this study also aimed to identify different

migration scenarios.

We employed a targeted approach based on large sample

sizes for well-defined spatial units of Acrocnida sp. col-

lected along a locally restricted bathymetric gradient (less

than 20 km) in Douarnenez Bay. Using transects, genetic

variation was assessed to determine the extent of hybrid-

ization between the two species and to identify potential

patterns of segregation in the spatio-temporal distribution

of parental and hybrid adults. Hybrid diagnosis involved

using the Internal Transcribed Spacer 2 region of the

ribosomal DNA (ITS2) restriction profiles together with

analysis of nucleo-cytoplasmic disequilibrium. The role of

spatial distribution and spawning asynchrony as pre-

zygotic barriers to cross-breeding were also tested. More-

over, using temporal samples collected throughout the

reproductive period should facilitate the detection of

potential counter-selection of hybrids during the course of

the summer period and its influence as a further post-

zygotic barrier preventing genotypes from becoming fixed

in the local population. Finally, this micro-spatial sampling

also served to better determine the mechanisms that influ-

ence the genetic structure within each species.

Materials and methods

Sampling

Samples were collected in 2005 and 2006 along two

transects (Fig. 1) spanning a bathymetric gradient of

30 meters. Each included three stations: one situated in the

intertidal at the lowest tidal limit on the shore (zero meter

depth) and two in subtidal areas at depths of 10 and 30 m,

respectively. Subtidal samples were obtained by sampling

the top 10 cm of sediment with a 0.25 m2 Hamon grab

(‘‘Ophirois’’ cruise) while intertidal samples were collected

by digging at low tide. Despite using a 2 mm-mesh sieve,

all individuals sampled had a disc diameter greater than

7 mm (measured using a binocular and a micrometer).

According to Bourgoin et al. (1991), such individuals

cannot be categorised as newly recruited individuals but as

two-years old adults.

To allow comparative genetic analyses, samples were

collected twice in both habitats, at similar dates in 2005, in

June (during or just after reproduction) and September

(well after reproduction). These dates were found appro-

priate as most of the recruitment process in Douarnenez

Bay occurs in late spring (Bourgoin et al. 1991; present

data). A time series of samples were collected in spring

2006, at one intertidal and one subtidal locality for com-

parison of gonadic development between the two species.

While all specimens from 2005 were whole frozen in

liquid nitrogen for genetic analysis, those collected in 2006

were dissected to quantify the extent of gonad maturity.

Genetic data acquisition

A total of 529 individuals from the two species of

Acrocnida sp. were fingerprinted using sequences of the

mitochondrial Cytochrome Oxidase I gene (COI), restric-

tion profiles of the ribosomal ITS2 internal spacer and

genotypes of five enzyme systems. For each individual, the

disc-shaped body was used for enzymatic starch gel elec-

trophoresis, and the arms for DNA extraction and

sequencing.

Mitochondrial COI sequencing

Specific primers Ab-COIf and Ab-COIr (Muths et al. 2006)

were used to amplify a 598-bp fragment of the COI gene.

Details of the COI amplification and sequencing proce-

dures are described in Muths et al. (2006). Sequences were

run along both directions, then edited using Chromas ver-

sion 1.6 (McCarthy 1997) and aligned using ClustalW

(Thompson et al. 1994) implemented in BioEdit Sequence

Alignment Editor (Hall 1999).

Amplification of the internal transcribed spacer 2 (ITS2)

and restriction fragment length polymorphisms (RFLPs)

Amplification of ITS2 was performed using primers

developed by P.W. H. Holland and collaborators (see

Jollivet et al. 1998). Reactions were performed in 27 ll

containing PCR buffer (19), 1 mM MgCl2, 12.5 lM of

each dNTP, 0.3 lM of each primer, 0.5 U of Thermoprime

Plus Taq polymerase (Abgene), 25 ng CTAB-extracted

genomic DNA. Cycling parameters were 94�C for 5 min,

followed by 40 cycles of 94�C for 45 s, 50�C for 60 s, and

72�C for 60 s and a final elongation at 72�C for 7 min.

PCR products of ten individuals of each lineage were

Genetica (2010) 138:667–679 669

123

Author's personal copy

cloned into a pGEM-T vector, using a Promega pGEM161

T cloning kit (Promega, Madison, WI, USA). Sequencing

of plasmid inserts was performed using universal M13

primers flanking the inserted fragment, and sequences were

submitted to GenBank (Accession nos EU431193–

EU431216). The amplified 535-bp sequences of the ITS2

rDNA displayed seven fixed mutations between Acrocnida

spatulispina and A. brachiata. To discriminate individuals

of each lineage and their potential hybrids, restriction

profiles were identified using a set of three restriction

enzymes (SmlI, Eco88 and XhoI). Enzyme digestions were

carried out according to the manufacturer’s instructions

(Biolabs), using 5 ll of amplified template from each of

the 529 individuals collected. RFLP profiles were then

visualised on a 4% agarose gel made up of TBE (19).

Starch gel electrophoresis

Electrophoretic procedures were performed according to

Pasteur et al. (1987), and the following five enzymatic

systems were used: Glucose Phosphate Isomerase (PGI,

E.C. 5.3.1.9), Mannose Phosphate Isomerase (MPI, E.C.

5.3.1.8), Malic Enzyme (ME, E.C. 1.1.1.40), Phosphoglu-

comutase (PGM, E.C. 5.4.2.2) and Hexokinase (HK, E.C.

2.7.1.1). Frozen samples of the body were homogenized in

250 ll grinding buffer (10 mM Tris–HCl, 2 mM EDTA,

0.05% b-mercaptoethanol, 0.1 mM PMSF, 0.25 mM

sucrose; pH 6.8), centrifuged at 6,500g for 12 min, and the

supernatant stored at -80�C until electrophoresis. Hori-

zontal enzyme electrophoresis was conducted on 12%

starch gels using the following buffers: (1) Tris–Borate-

EDTA (pH 8.6) for PGI, MPI and ME and (2) Tris–HCl

(pH 8.5) for the PGM and HK. Buffer systems were run at

constant amperage, during 5 h (160 V, 35 mA for system

(1) and 150 V, 40 mA, for system (2)). The most frequent

allele was labelled ‘100’ and all others were labelled

according to their relative mobility from allele 100.

Population genetic analyses

All analyses were made in parallel for the two species. For

each station, haplotype (HD) and nucleotide (p) diversities

were estimated with DNAsp 4.0 (Rozas et al. 2003).

Pairwise values of haplotypic frequency-based differenti-

ation (FST) were calculated using ARLEQUIN 2.0

(Schneider et al. 2001).

For the allozyme dataset, the allele frequencies per

population, together with the observed (HO) and expected

(HNB) heterozygosities were all estimated using GENEPOP

3.4 software (Raymond and Rousset 1995). Allelic richness

(RS) was further estimated using FSTAT 2.9.3.2 (Goudet

1995) based on a diploid number of 22 individuals.

Deviations from Hardy–Weinberg equilibrium expecta-

tions were tested for each population at each locus. This

was achieved by calculating Wright’s fixation index FIS as

estimated by Weir and Cockerham (1984) and departures

from Hardy–Weinberg equilibrium tested using Fisher’s

exact tests implemented in GENEPOP 3.4. The null

hypothesis of independence between loci was tested from

genotypic disequilibrium analyses using the same software.

Genetic differentiation between pairwise combinations of

samples from each station and from each sampling date

was estimated by calculating Weir and Cockerham’s

(1984) h for each locus and between stations and sampling

dates. The Waples’ (1989) test was also used to estimate

the effect of drift on temporal changes in allele frequencies.

As sampling was undertaken after reproduction, analyses

were conducted according to sampling plan I (individuals

taken after reproduction; see Waples 1989). Using

TEMPTEST 2.2 (Waples 1989), calculations were made

locus by locus, using frequencies of the two most common

alleles. A third ‘‘combined’’ category included all other less

frequent alleles. As the effective population size (Ne) was

unknown, each test was performed using five Ne values

ranging from 50 to 1,000, which seems a reasonable range

given published census data. The null hypothesis of no

change was rejected if changes in allele frequencies

between sampling dates were significantly greater than

expected under conditions of random genetic drift alone.

To estimate the robustness of allozyme divergence

between mitochondrial lineages, exact tests of cytonuclear

disequilibrium were conducted using the CNDM program

(Asmussen and Basten 1994; Basten and Asmussen 1997).

Allozymes significantly correlated to each species were

then used to reconstruct a hybrid index (HdI) in order to

detect putative hybrids in the collected sample. For each

diagnostic allele identified with CNDM, a score of -1 or

?1 was respectively assigned to the characteristic forms of

A. spatulispina and A. brachiata, respectively. With only

three diagnostic loci per species, cumulative HdI values

ranged from -5 to ?6. With nine diagnostic loci for A.

brachiata versus five for A. spatulispina, the central value

was not 0 but closer to 2. Thus, individuals of A. spatu-

lispina displaying an HdI greater or equal to 2 can be

considered as potential hybrids. Conversely, the same

possibility may be expected for individuals of A. brachiata

yielding a HdI strictly lower than 2. The software

STRUCTURE 2.2 (Pritchard et al. 2000) was also used to

test the robustness of genetic isolation between Acrocnida

species and identify putative hybrids. The number of

populations was set from K = 2 to 6 (for the two species to

the 6 sampling sites). The optimal value of K was selected

based on Evanno et al. (2005). Simulations involved

100,000 Markov Chain Monte Carlo (MCMC) iterations

followed by a burn-in period of 100,000.

670 Genetica (2010) 138:667–679

123

Author's personal copy

Spatio-temporal survey of gonad maturity

Sexual maturity was determined for each species based on

the Gonadic Index (GI). The survey was undertaken for A.

spatulispina at station 0A and for A. brachiata at station

10A during the reproductive season, with samples collected

every 15 days, from mid-March 2006 to mid-July 2006.

For each sampling date and each species, ten males and ten

females were fixed in a ‘‘Bouin’’ solution (75% picric acid,

20% formaldehyde, 5% acetic acid; Bourgoin 1987). A

total of 270 individuals were dissected under a binocular

microscope. Gonads were separated from somatic tissue

and both were dried at 60�C for 24 h before being weighed

as dry weight (DW). GI was then estimated as the ratio of

DW of gonads to DW of whole body, and times 100 to give

percentage. Significant differences between sampling dates

were then tested using a Mann & Whitney U-test.

Results

As in Muths et al. (2006), a deep level of mitochondrial

COI divergence (19.6% according to K2-P distance;

Kimura 1980) was revealed between subtidal and intertidal

areas which correspond respectively to A. brachiata and

A. spatulispina. The analysis made with STRUCTURE

strongly suggested that the highest likelihood of obtaining

such data was to consider that two strongly separated

genetic entities existed (K = 2). Moreover, highly signifi-

cant cytonuclear disequilibria (P \ 0.001) were found

from four of the five allozyme loci, all except for the Me

locus (Table 1). However, 2.6% of all individuals analyzed

had a mitochondrial and nuclear genetic signature respec-

tively typical from opposite species. These elements con-

firmed that reproductive isolation is incomplete among

those evolutionary and morphologically divergent species,

at least in Douarnenez Bay. We therefore first searched

among the possible evolutionary mechanisms responsible

for maintaining low levels of hybridisation in the face of

strong spatial segregation, and vice versa. We then ana-

lyzed the genetic structure within each species.

Segregation between species

Bathymetric separation

The bathymetric segregation between the two species is

clear and abrupt, with intertidal or subtidal individuals

displaying completely different sets of haplotypes. In

addition, allozymes also exhibited strong differences

between species. The intertidal A. spatulispina samples had

two diagnostic alleles at the Hk locus: Hk-90 (only found

in A. spatulispina) and Hk-95 (present in only four

A. brachiata samples) (see Fig. 2a). Such exclusive alleles

were however not found to be significantly associated with

the intertidal lineage in the cytonuclear disequilibria

(Table 1). Likewise, in A. brachiata, despite one exception,

all the individuals sampled from subtidal areas (10 and

30 m depths) displayed either allele Hk-100 or Hk-105

(Fig. 2b). This spatial segregation was observed both in

June, just after reproduction, and in September, well after

the reproductive period. This emphasises the strong rela-

tionship between bathymetry and species distribution.

Spawning asynchrony

For all species and sampling dates, the Gonadic Index (GI)

was globally higher for males than for females, but dif-

ferences were never significant (Mann & Whitney U test,

P [ 0.05). Gonad maturation data were therefore pooled

within each site to perform comparisons between species at

each sampling date (Fig. 3). Over the 2006 summer period,

for Acrocnida brachiata, GI increased significantly in April

whereas for A. spatulispina, this increase began in early

May (Mann & Whitney test, U = 2.72 and 3.75, respec-

tively, with P \ 0.001). This increase in gonad maturation

was followed 1 month later, for both species, by an abrupt

drop associated with the release of gametes in the water

column. This induced significant differences in GI between

species during the following months (P \ 0.001). Gamete

release was detected in May for A. brachiata (U = 4.63;

P \ 0.001) and in early June for A. spatulispina

(U = 3.26; P \ 0.001). Overall, the time series of both

species show a clear spawning asynchrony of about

15 days between both species.

Hybridization between species and distribution of hybrids

Out of a total of 529 individuals, the proportion of hybrids

detected using ITS2 restriction profiles was only 0.95%, i.e.

five individuals (results summarized in Table 2). This excess

Table 1 Cytonuclear linkage disequilibrium for the two Acrocnidaspecies: between mitochondrial sequences and enzyme alleles, esti-

mated using CNDM software (Asmussen and Basten 1994; Basten

and Asmussen 1997)

Associated alleles mtDNA Type P

A. spatulispina A. brachiata

HK 80 100/105/110 < 0.001

PGM 90 104/110/116 < 0.001

PGI – 90/104/108 < 0.001

ME – – 0.93

MPI 70 – < 0.001

Values in bold show which alleles of each locus is significantly

associated with each species

Genetica (2010) 138:667–679 671

123

Author's personal copy

of parental types and deficiency of recombinants is illus-

trated by significant cytonuclear disequilibrium (Table 1).

The five ITS2 profiles that were identified as potential

hybrids all displayed an A. spatulispina mitochondrial

haplotype but, based on the hybrid index, only four indi-

viduals displayed an A. brachiata nuclear background. All

four were sampled in the intertidal habitat (station 0A) in

June. On the contrary, the remaining hybrid detected with

ITS2 showed a clear mitochondrial and nuclear genetic

signature of A. spatulispina and was sampled in the subtidal

station (station 10B) in September. This shows that a strong

spatial segregation between hybrid types might also occur, at

least just after reproduction in June.

Strictly based on the hybrid index (see Fig. 4 and details in

Table 2), nine additional individuals could also represent

introgressed individuals. When the search for putative hybrids

was performed with STRUCTURE (Pritchard et al. 2000)

using five allozyme loci together with the ITS2 data, the same

nine individuals were identified as intermediate types.

On these 14 putative hybrids (2.6% of all individuals

analyzed), most were sampled in June from only two sta-

tions: one intertidal (0A, n = 4) and one subtidal (30B,

n = 5). The geographical distribution of hybrids was less

homogeneous in September, with five putative hybrids

found from four different stations (0A, 10A, 10B and 30B).

Temporal changes in the localisation of putative hybrids

could also be identified by looking at the temporal variation

Fig. 2 Haplotype and allozyme frequencies in A. spatulispina(Fig. 2a, n = 162) and A. brachiata (Fig. 2b, n = 367) Symbols ontop of histograms indicate when samples differ significantly between

June and September on the basis of a permutation test on the estimator

h of Weir and Cockerham (1984) (* P \ 0.05 and *** P \ 0.001) and

whether these differences are explained by drift alone on the basis of

the Waples’ (1989) test with ¤ (¤ \ 0.05 and ¤¤¤ \ 0.001).

Concerning haplotype frequencies (in top left corner), HIi and HSi

are the more frequently found haplotypes in A. spatulispina and

A. brachiata respectively. Sh. corresponds to shared haplotypes

(found in more than one population of a given species but at low

frequency). Priv. corresponds to private haplotypes (found in only one

population). Concerning allozyme frequencies, the allele labeled 100

refers to the most frequent one

0

20

40

60

03/17

/06

04/13

/06

04/26

/06

05/15

/06

05/29

/06

06/16

/06

07/13

/06

Gon

daic

inde

x (G

I, %

)

* **

*

Fig. 3 Time series showing the evolution of the Gonadic Index (GI)

in the two Acrocnida species sampled in Douarnenez Bay (n = 270).

Bold and standard lines represent A. brachiata and A. spatulispina,

respectively. * indicates a significant difference in GI, at a given date

672 Genetica (2010) 138:667–679

123

Author's personal copy

in the frequencies of ‘diagnostic’ alleles between sampling

dates (e.g. Hk-100 and Hk-105 in station 0A and Hk-95 in

station 30B; see Fig. 2).

Spatial and temporal genetic variation within Acrocnida

spatulispina

While mitochondrial sequences displayed high levels of

haplotype (HD) and nucleotide (p) diversities at both

intertidal stations surveyed (see values in Table 3), both

mean values were lower in June than in September. This

decrease was associated with the appearance of a new

haplotype (HI5) and an increase in the frequency of hap-

lotype HI4 at both intertidal stations in September

(Fig. 2a). Nevertheless, the proportion of shared and

‘‘private’’ haplotypes did not vary significantly between

sampling dates, resulting in non-significant temporal

genetic differences in pairwise FST estimates obtained from

mtCOI data (Table 4a).

The number of alleles at the five polymorphic enzyme loci

screened varied from 3 to 6, with an average of 4.8. Whereas

allelic richness (RS) was similar stations and sampling dates,

allelic frequencies (f, Fig. 2a) differed significantly between

stations and sampling dates. For example, at both intertidal

stations, the frequency of Pgi-104 increased significantly

from June to September, and inversely for Pgi-100. The

frequency of Me-100 at station 0B also increased signifi-

cantly from 50 to 84% between June and September, with a

concomitant decrease in Me-110. For the Me, frequencies

only remained constant at station 0A. Taken together, allele

frequencies between the two intertidal stations were far more

similar in September (h = 0.002, P [ 0.05, Table 4c) than

they were in June (h = 0.099, P \ 0.001, Table 4c), just

after reproduction. Values of temporal genetic differentia-

tion were also highly significant for both stations

(P \ 0.001) and were mostly associated with significant

monolocus P-values in the Waples’ test (1989; Fig. 2a). The

observed heterozygosity (HO) was higher than the expected

value (HNB) in September but not in June. Generally, there

were no significant deviations from Hardy–Weinberg equi-

librium expectations. Heterozygote deficiencies were only

observed in June for station 0B (Table 5).

Spatial and temporal genetic variation within Acrocnida

brachiata

Both haplotype (HD) and nucleotide (p) diversities were

higher in A. brachiata than in A. spatulispina (Table 3).

Pairwise FST estimates (Table 4b) indicated that most

values of genetic differentiation were low and non-signif-

icant for subtidal stations, except between stations 10B and

Table 2 Hybrid diagnosis

between Acrocnida species

The following details are given

for the 14 putative hybrids: the

sampling dates and stations; the

mtDNA signature (As for A.spatulispina, Ab for A.brachiata); the respective

genotypes at the 5 allozyme loci

and the corresponding

hybridization index (HdI: values

strictly lower than 2 or higher

than 2 respectively indicates A.brachiata or A. spatulispinanuclear background); the ITS2

restriction profiles (As for A.spatulispina, Ab for A.brachiata, H for Hybrid)

Month Station mtDNA HK PGM PGI ME MPI Hdl ITS2

0A As 090095 100104 100100 100110 090100 -1 H

June 0A As 100100 100100 100104 100100 090100 3 H

0A As 100100 100104 100104 100100 090100 4 H

0A As 100105 100104 100100 100100 090100 3 H

Sept. 0A As 090100 100104 100104 100110 090090 2 As

30B Ab 100105 100100 100100 100100 070100 1 Ab

30B Ab 100105 100100 100100 100110 070100 1 Ab

June 30B Ab 095105 096104 100100 110110 100100 1 Ab

30B Ab 095105 100100 100100 100100 100100 0 Ab

30B Ab 095105 100104 100100 100100 100100 1 Ab

10B Ab 100105 110110 100108 100100 090100 5 H

Sept. 10A Ab 100105 090100 100100 100100 100100 1 Ab

10B Ab 105105 100100 100100 110110 070090 1 Ab

30B Ab 100105 100104 100100 100100 070070 1 Ab

0

50

100

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6

Hybrid index (HDi)

Nb

of

ind

ivid

ual

s

A.spatulispina

A. brachiata

Fig. 4 Frequency distribution of the hybridization index HdI in the

whole sampling period of 2005, based on the result of the CNDM

software (Asmussen and Basten 1994; Basten and Asmussen 1997).

Extreme negative scores correspond to a nuclear background of

‘pure’ A. spatulispina whereas extreme positive scores to ‘pure’

A. brachiata. The Y axis represents the number of individuals of each

species

Genetica (2010) 138:667–679 673

123

Author's personal copy

Table 4 Spatial and temporal genetic differentiation within each Acrocnida species

(4b) 10 A 10 B 30 A 30 B(4a) 0A 0B 10 A -0.005 0.001 0.001 -0.0050A 0.001 -0.034 10 B -0.003 0.006 0.023*** 0.0060B -0.019 -0.034 30 A -0.005 -0.003 -0.005 -0.002

30 B 0.005 -0.000 -0.005 -0.002

(4d)(4c) 0.023* 0.019* 0.032*** 0.0030A 0.060*** 0.002 0.035*** 0.027* 0.075*** 0.038***0B 0.099*** 0.051*** 0.033*** 0.052*** 0.010 0.014*

0.036* 0.018* 0.063*** 0.093***

A. spatulispina A. brachiata

0A 0B 10 A10 B30 A30 B

10 A 10 B 30 A 30 B

Allo

zym

esm

tDN

A

The two-first tables (Tables 4a, b) are based on mtDNA dataset (FST) and the two-second ones (Tables 4c, d) on allozymes (h) dataset. Tables 4a

and c are for A. spatulispina and Tables 4c and d for A. brachiata. White cells represent spatial estimates of differentiation obtained between

stations sampled in June 2005, and light grey cells represent estimates obtained in September 2005. Dark grey cells represent the temporal

genetic differentiation observed between June and September, for any given station. Tests of significance were performed for mtDNA data and

for allozymes with ARLEQUIN 2.0 (Schneider et al. 2001) and with GENEPOP 3.4 (Raymond and Rousset 1995) respectively: *P \ 0.05;

***P \ 0.001; after a Bonferroni correction, only values significant at P \ 0.001 remain significant

Table 3 Temporal variation in genetic diversity in Douarnenez bay, for both Acrocnida species, and for all sampling stations and dates (n is the

sample size)

Species Station June September

n HD P RS HNB HOBS n HD P RS HNB HOBS

A. spatulispina 0A 53 0.779 0.0038 3.1 0.343 0.328 45 0.833 0.0042 3.5 0.457 0.469

0B 42 0.672 0.0039 3.1 0.458 0.381 22 0.833 0.0047 2.8 0.415 0.464

A. brachiata 10A 33 0.917 0.0056 3.5 0.480 0.454 48 0.950 0.0047 3.0 0.488 0.492

10B 57 0.911 0.0040 3.3 0.545 0.428 31 0.890 0.0037 3.6 0.538 0.485

30A 36 0.942 0.0049 3.0 0.492 0.375 52 0.988 0.0070 3.4 0.527 0.403

30B 51 0.892 0.0069 2.8 0.432 0.349 59 0.940 0.0048 3.2 0.483 0.381

Mitochondrial DNA diversity values are represented in italics (HD, haplotype diversity; p, nucleotide diversity). Indices from allozyme data are

represented by RS (allelic richness based on a diploid number of 22 individuals), HNB, (unbiased expected heterozygosity, Nei 1987) and HO

(observed heterozygosity)

Table 5 Estimates of the fixation index FIS for each Acrocnida species, within each station (J refers to June; S refers to September)

A. spatulispina A. brachiata

OAJ OAS 0BJ 0BS 10AJ 10AS 10BJ 10BS 30AJ 30AS 30BJ 30BS

HK -0.01 -0.070 -0.159 -0.066 -0.019 -0.35 -0.165 0.079 0.011 -0.144 0.186 0.009

PGM -0.049 0.104 -0.091 -0.027 0.128 -0.016 0.282*** 0.484*** 0.305* 0.004 0.221* 0.267*

PGI -0.097 -0.313 0.052 -0.400 0.314* 0.076 0.092 -0.161 0.124 -0.079 -0.120 0.06

ME 0.633*** 0.024 0.814*** -0.167 0.364*** 0.192*** 0.514*** 0.166*** 0.341*** 0.279*** 0.664*** 0.542***

MPI -0.021 0.113 0.142* 0.031* 0.426*** 0.200*** 0.308*** 0.45*** 0.309*** 0.542*** 0.198*** 0.467***

Multilocusaverage

0.043 -0.027 0.171*** -0.119 0.055 -0.008 0.239*** 0.238*** 0.217*** 0.101*** 0.194*** 0.213***

Linkagedisequilibrium

PGM-MEME-

PGIME-MPI HK-MPI HK-ME

Tests of significance were performed with GENEPOP 3.4 (Raymond and Rousset 1995); *P \ 0.05; ***P \ 0.001. Significant linkage dis-

equilibria estimated between pairs of loci at each sampling station are also indicated

674 Genetica (2010) 138:667–679

123

Author's personal copy

30A in September. This was associated with the common

haplotype HS1 comprising the biggest proportion of hap-

lotypes in 30A when compared to 10B.

In terms of allozyme diversity, both expected (HNB) and

observed (HO) heterozygosities were higher in A. brachiata

than in A. spatulispina (Table 3) although there were no

differences in allelic richness (RS). In contrast to the mito-

chondrial dataset, clear differences in allozyme frequencies

between sampling dates were observed at nearly all stations.

For example, at four subtidal stations, an increase in Pgi-108

between June and September occurred concomitantly with

an increase in Me-100 and a decrease in Mpi-100. These

temporal variations in allele frequencies corresponded to

significant monolocus P-values in the Waples’ (1989) test

(Fig. 2b) and with significant temporal genetic differentia-

tion observed within the two stations situated at 10 m depth

(10A and 10B). Highly significant temporal changes also

occurred at station 30B but not at station 30A (Table 4d),

suggesting a high degree of self-recruitment and a limited

degree of hybridization at this site. Globally, genetic dif-

ferentiation was more pronounced in June between the two

stations of a same transect (10A vs. 30A or 10B vs. 30B),

than in September between stations of the same bathymetric

level (10A vs. 10B or 30A vs. 30B). Nevertheless, spatial

differentiation between subtidal stations in June and in

September was always highly significant. Overall, depar-

tures from Hardy–Weinberg proportions were highly

significant in both June and September (see FIS values in

Table 5), with heterozygote deficiencies observed at all

stations except for station 10A. In relation to the observed

departures from Hardy–Weinberg equilibrium at the Me and

Mpi loci, significant genotypic disequilibrium between the

two loci was also apparent.

Discussion

Spatial isolation and spawning asynchrony as

pre-zygotic mechanisms limiting local hybridization

On the basis of hybridization laboratory experiments and in

situ observations, Mallet (2005) showed a linear negative

relationship between species divergence (and so the time

since divergence) and cross-fertilization. Similarly, hybrid

viability also decreases as a function of divergence (Bolnick

and Near 2005). The occurrence of hybridization at a fre-

quency of 2.6% between the two spatially segregated and

highly divergent Acrocnida species (DAVG-COI = 19.6%;

hALLOZYMES = 0.22) clearly indicates incomplete repro-

ductive isolation among natural populations. The distribu-

tion of the hybridization index is clearly bimodal with rare

hybrids and predominating parental forms. While bimo-

dality is supposed to be associated with assortative mating

or preferential fertilization (Jiggins and Mallet 2000)

experimental trials established to test for preferential

assortative mating and hybrid viability of Acrocnida failed

thus leaving a number of unanswered questions.

The two Acrocnida species largely segregate with depth,

with an abrupt transition at low tidal levels. The temporal

stability of this bathymetric segregation may suggest that

strong selective mechanisms associated with depth operate.

While selection against migrants across the low tide

boundary may occur through habitat-specific differential

mortality, the short dispersal phase of pelagic larvae

(around 4 days) and the gregarious behaviour of the species

both act to reduce larval settlement or juvenile/adult dis-

persal in opposite habitats. Studies have shown the direct

(Schmidt and Rand 2001) and indirect (Cruz et al. 2004)

effects of selection by depth on the spatial restriction of

genotypes. In addition, the fact that individuals of both

species are strongly restricted to different bathymetric

levels drastically reduces further the probability of cross-

fertilization. Even in the extreme case of long-distance

fertilization of the sea-star Acanthaster planci, success

rates drops abruptly 40 meters from the source (Babcock

et al. 1994). Acrocnida sp. is not a long distance fertilizer

and the rapid dilution of gametes may be significantly

increased in the intertidal zone, compared to the subtidal

(Denny and Shibata 1989). Even given full gametic com-

patibility and long-distance dispersal, taxa that spawn in

allopatry are obviously less likely to cross-fertilise (Lessios

1984; Byrne et al. 1994). Together with bathymetric seg-

regation and the short larval dispersal phase, marine cur-

rents—e.g. strong outward currents at the entrance of the

study area (station 30A; Blanchet et al. 2004)—may also

act to limit the meeting of gametes between neighbouring

demes and favour self recruitment processes.

Hybridization is also physically limited by temporal iso-

lation linked to reproductive asynchrony, with a 2-weeks

delay in spawning activity between species. Spawning

asynchrony is a characteristic of many marine sibling species

living in sympatry (Knowlton 1993). This is an evolutionary

mechanism which, when gametic, habitat and/or behavioural

isolation are incomplete, may lead to parapatric and sym-

patric forms of speciation. Differences in gamete maturation

between subtidal and intertidal samples of what was at time

thought to represent only A. brachiata were noticed 20 years

ago by Bourgoin et al. (1991) but were only viewed as the

result of differences in the thermal regime affecting indi-

viduals of the same population. It is now clear that this

spawning shift actually corresponds to the existence of two

morphologically distinct species. The earlier spawning per-

iod demonstrated for the subtidal species may reflect the

more stable conditions found at this bathymetric level,

compared to the intertidal zone where stress linked to

emersion is high. Spawning in temperate waters is highly

Genetica (2010) 138:667–679 675

123

Author's personal copy

dependent on seasonal variations of temperature and nutri-

ents. Such parameters greatly vary with latitude but also with

depth where environmental conditions are buffered com-

pared to the intertidal zone. This hypothesis was emitted by

Nichols and Barker (1984) to explain a similar spawning

asynchrony between intertidal and subtidal populations of

Asterias rubens in the English Channel. A further explana-

tion is based on the gregariousness of externally fertilising

species: the more abundant the individuals are, the more they

will interact and the earlier spawning may take place relative

to the case where individuals are less abundant. This could

help explain the earlier spawning event observed in the

subtidal area where density is at least ten times more

important (Bourgoin 1987; D. Muths, personal observation).

The timing of spawning activity may also be a heritable trait

since demic populations are often composed of a mixture of

individuals that reproduce at different times, and this can be

essential to the stability of the whole meta-population.

Genetic exchanges within a given population (or deme) may,

however, decrease with increasing differences in reproduc-

tive timing, thus generating a process of Isolation-By-Time

(IBT) and adaptation (Hendry and Day 2005). Environ-

mental determinism and heritability in reproductive timing

are, however, not mutually exclusive explanations. Levitan

et al. (2004) showed that a shift of only 2 h in the spawning of

two coral species was sufficient to greatly reduce fertilization

success. The spawning asynchrony that characterizes

Acrocnida species is surely a common feature among other

sister species with similar bathymetric restrictions, for

example among intertidal and subtidal lineages of the

polychaete tubeworm Owenia fusiformis (Jolly et al. 2006).

Our results indicate that the combined effect of spatial

ecological constraints and temporal spawning asynchrony

act as a strong mechanism of pre-zygotic isolation greatly

reducing the likelihood of hybridization between lineages.

Nevertheless, such pre-zygotic barriers are incomplete to

offset hybridization.

Limited role of post-zygotic mechanisms in preventing

hybridization

The presence of even low levels of hybridization among the

two species of Acrocnida sp. reinforces the idea that hybrid

zones can persist over long periods of time through a balance

between dispersal of parental types and selection against

their hybrids (Harrison 1993). In our case, results indicate

that the average age between lineages (over 3.5 million years

ago if we assume an average constant mutation rate; Muths

et al. 2006) is still too short relative to the accumulation of

significant post-zygotic incompatibilities. Indeed, this is

demonstrated by the existence of different levels of intro-

gression which indicates that hybrids might not be sterile or

poorly adapted to local reproduction. On the contrary, the

fact that fewer hybrids were sampled in September compared

to June 2005, as well as the decrease at station 0A between

June and September of one diagnostic allele of A. brachiata

(Hk-100), might also indicate selective mortality operating

against Acrocnida sp. hybrids. Thus, some of our data indi-

cates that post-zygotic isolating mechanisms are acting to

reinforce pre-zygotic ones. Due to our restricted sampling

scheme, we cannot rule out the alternative explanation of

migration by juvenile hybrids. These results emphasise the

idea that multiple pre-zygotic and post-zygotic factors are

needed to create a definite barrier to genome re-homogeni-

zation, especially among closely related species living in

sympatry (Bierne et al. 2002).

There are a number of ways by which post-zygotic

processes may have affected small scale patterns of genetic

differentiation observed within each species, and which

could explain the transient nature of genetic structure

between June and September. Whereas hybridization

events tend to homogenize genetic signature between

species, it also creates some spatio-temporal structure at

the within-species level due to different levels of intro-

gression and heterogeneity of hybrid distribution. Conse-

quently, post-zygotic processes such as hybrid mortalities

which should homogenize hybrid repartition will decrease

genetic differentiation between two stations of a same

species but will also generate differentiation between two

temporal samples of the same station. But, keeping in mind

the low proportion of hybrids, this counter-selection, alone,

can definitively not explain all the within-species spatial

genetic structure observed in Douarnenez Bay.

Migration dynamics is shaping the within-species

genetic structure differently according to habitat

In June, a pronounced genetic structure was observed

between populations of both species, with all pairwise values

of genetic differentiation being significant. This may indi-

cate that gene flow is reduced even between neighbouring

demes, which fit well with expectations based on a very

short-larval phase (Muths et al. 2006). This could also be

caused by strong selective mortalities associated with

increased intra/inter-specific competition during the first few

months of the benthic juvenile life (Gosselin and Qian 1997).

In September, the extent and significance levels of

spatial differentiation in both species was far from the

structure observed in June. This is especially true for the

two intertidal stations where A. spatulispina was sampled

and for which we observed highly significant temporal

changes, followed by spatial genetic homogeneity. Tem-

poral genetic changes implicating a reduction in spatial

genetic structure was also observed for A. brachiata,

although generally to a much lesser extent, with temporal

differentiation only marginally significant (10A and 10B) or

676 Genetica (2010) 138:667–679

123

Author's personal copy

not significant at all (30A). Significant values of the

Waples’ (1989) test suggest that temporal variations in

allele frequencies are greater than expected under the

influence of drift alone (sampling or genetic drift). The

differentiation observed between June and September is

likely due to allele frequency changes at one or more loci, in

relation to migration and selective mortality. While the

increase of Pgi-108 among A. brachiata samples at subtidal

stations could reflect selective effects, the appearance of

Pgi-90 at station 30A or Hk-110 at three of the four subtidal

stations in September clearly implicate migration processes.

Among A. spatulispina, these migration processes can be

illustrated in September by the appearance of high propor-

tions of new haplotypes (HI5 & HI4), spatial genetic

homogenization at intertidal stations and an increase in both

mitochondrial and nuclear diversities at station 0A.

Considering migrations in Acrocnida species, one

peculiar feature of our data is that all the sampled brittle-

stars were characterised as juveniles or adults (all more than

7 mm), and no newly recruited individuals were found.

Therefore, migration events of both juvenile and adult

stages may be considered as a feasible alternative explain-

ing the spatio-temporal genetic structure in Dournenez Bay.

The migratory behaviour of adult brittle-stars, including

A. brachiata, has already been suggested to occur during

summer periods in the Bristol Channel (Tyler and Banner

1977). In this case, migration could either be due to rafting

(Hendler and Miller 1991; Hendler et al. 1999) or active

swimming with the tide, a behaviour also noticed for other

benthic invertebrates (Farke and Berghuis 1979; Guillou

1980; Goodacre 2006). Migration could also be brought

about by passive displacements caused, for example, by

high rates of sediment re-suspension as a result of storm

activity. This has frequently been observed for soft sedi-

ment marine invertebrates (Tamaki 1987; Zuhlke and Reise

1994; Grant et al. 1997) and in Dournenez Bay itself

(Pectinaria koreni; Jolly and Gentil, personal observation).

All the above affect population stability, and which might

explain, not only the strong temporal changes in allele

frequencies and the spatial genetic homogenization of

intertidal A. spatulispina samples, but also the pronounced

and more stable spatial genetic structure observed among

subtidal A. brachiata samples.

For both species, the homogenizing effect on allele

frequencies observed within and between stations over the

summer period is best explained by active migratory events

(juvenile, adults) in conjunction with hybrid mortalities.

On the contrary, even limited hybridization events during

the spawning season coupled to strong adult mortalities

following reproduction (reported by Bourgoin et al. 1991)

tend to generate local genetic structure within a species as

seen in June. Both selection and migration therefore play

an important role in the genetic population structure of

each lineage, at least at this small spatial scale and over this

short time frame. As these two evolutionary forces may act

synergistically but differently at each locality within the

embayment (whether subtidal or intertidal), larger scale

patterns such as historical isolation by distance may easily

be masked.

This study illustrates the fact that incomplete repro-

ductive isolation may occur between highly divergent

species and that local hybridization together with selective

pressures may affect population structure. It also shows

how reproductive isolation is maintained during the for-

mation of species, or after secondary contact, through a

combination of asynchronous population dynamics and

local selective constraints. Pre-zygotic processes such as

ecological preferences, which may be caused by phyloge-

netic niche conservatism (Wiens and Donoghue 2004), and

spawning asynchrony are the most likely pre-zygotic bar-

riers to interspecific matings contributing to reproductive

isolation in A. brachiata. Hybrid inviability does not seem

to be a strong contributing post-zygotic factor in sympatric/

parapatric speciation via reproductive isolation. Infertility

of the F1 & F2 generation should however be tested

experimentally. In addressing the particular environmental

conditions favouring the spatial segregation of ecotypes

and genotypes, new in situ post-larval data collection and

eco-physiological experiments (such as in Edmands and

Deimler 2004; Gardner and Thompson 2001) would truly

be an interesting complement to the present study.

Acknowledgments We are very grateful to people who helped us in

the collection of samples, particularly to the crews of the N/O ‘‘Cotes

de la Manche’’ and ‘‘Cotes d’Aquitaine’’. Many thanks also to the

divers of the ‘‘Station Biologique de Roscoff’’, Caroline Broudin

and to Yvan Lebras. This work was funded by the ‘‘PNEC-AT’’

(Programme National sur l’Environnement Cotier), the European

network of excellence ‘‘Marbef’’ (Marine Biodiversity and Ecosystem

Functioning) and supported by a PhD grant from the French Ministry

of Research.

References

Arndt A, Smith MJ (1998) Genetic diversity and population structure

in two species of sea cucumber: differing patterns according to

mode of development. Mol Ecol 7:1053–1064

Arndt A, Marquez C, Lambert P, Smith MJ (1996) Molecular

phylogeny of eastern Pacific sea cucumbers (Echinodermata:

Holothuroidea) based on mitochondrial DNA sequence. Mol

Phyl Evol 6:425–437

Asmussen MA, Basten CJ (1994) Sampling theory for cytonuclear

disequilibria. Genetics 138:1351–1363

Babcock RC, Mundy CN, Whitehead D (1994) Sperm diffusion

models and in situ confirmation of long-distance fertilization in

the free-spawning asteroid Acanthaster planci. Biol Bull

186:17–28

Basten CJ, Asmussen MA (1997) The exact test for cytonuclear

disequilibria. Genetics 146:1165–1171

Bermingham E, Lessios HA (1993) Rate variation of protein and

mitochondrial DNA evolution as revealed by sea urchins

Genetica (2010) 138:667–679 677

123

Author's personal copy

separated by the Isthmus of Panama. Proc Nat Acad Sci USA

90:2734–2738

Bierne N, David P, Boudry P, Bonhomme F (2002) Assortative

fertilization and selection at larval stage in the mussels Mytilusedulis and M. galloprovencialis. Evolution 56:292–298

Bierne N, Borsa P, Daguin C et al (2003) Introgression patterns in the

mosaic hybrid zone between Mytilus edulis and M. galloprovin-cialis. Mol Ecol 12:447–461

Blanchet A, Chevalier C, Gaffet JD, Hamon D (2004) Bionomie

benthique subtidale en Baie de Douarnenez, Vol. Ifremer, Brest,

p. 51

Boissin E, Feral J, Chenuil A (2008) Defining reproductively isolated

units in a cryptic and syntopic species complex using mitochon-

drial and nuclear markers: the brooding brittle star, Amphipholissquamata (Ophiuroidea). Mol Ecol 17:1732–1744

Bolnick DI, Near TJ (2005) Tempo of hybrid inviability in

Centrarchid fishes (Teleostei: Centrarchidae). Evolution

59:1754–1767

Bourgoin A (1987) Ecologie et demographie d’Acrocnida brachiata(Montagu) (Echinodermata : Ophiuroidea) en baie de Douarne-

nez (Bretagne). These 3eme cycle, Oceanographie, Universite de

Bretagne occidentale, p. 146

Bourgoin A, Guillou M, Glemarec M (1991) Environmental instability

and demographic variability in Acrocnida brachiata (Echinoder-

mata : Ophiuropidea) in Douarnenez Bay (Brittany : France). Mar

Ecol 12:89–104

Byrne M, Mundy C, Keesing J, Oliver J (1994) Hybridization of

sympatric Patiriella species (Echinodermata, Asteroidea) in

New South Wales. Evolution 48:564–576

Cabioch L (1968) Contribution a la connaissance des peuplements

benthiques de la Manche Occidentale. Cah Biol Mar 9:493–720

Chenuil A, Feral JP (2003) Sequences of mitochondrial DNA suggest

that Echinocardium cordatum is a complex of several sympatric

or hybridizing species : a pilot study. In: David F (ed)

Echinoderm research 2001. Swets & Zeitlinger, Lisse, pp 15–21

Cruz R, Carballo M, Conde-Padin P, Rolan-Alvarez E (2004) Testing

alternative models for sexual isolation in natural populations of

Littorina saxatilis: indirect support for by-product ecological

speciation? J Evol Biol 17:288–293

Denny MW, Shibata M (1989) Consequences of surf-zone turbulence

for settlement and external fertilization. Am Nat 134:859–889

Edmands S, Deimler JK (2004) Local adaptation, intrinsic coadap-

tation and the effects of environmental stress on interpopulation

hybrids in the copepod Tigriopus californicus. J Exp Mar Biol

Ecol 303:183–196

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

clusters of individuals using the software structure : a simulation

study. Mol Ecol 14:2611–2620

Farke H, Berghuis EM (1979) Spawning, larval development and

migration behaviour of Arenicola marina under field conditions

in the western Wadden Sea. Neth J Sea Res 13:529–535

Gardner JPA, Thompson RJ (2001) The effects of coastal and

estuarine conditions on the physiology and survivorship of the

mussels Mytilus edulis, M. trossulus and their hybrids. J Exp Mar

Biol Ecol 265:119–140

Gentil F, Zakardjian B (1989) Premieres donnees sur la dynamique de

population d’Acrocnida brachiata (Montagu) en Baie de Seine

orientale. Vie marine HS 10:148

Goodacre SL (2006) Marathon mollusc migrations. Heredity 97:1

Gosselin L, Qian P (1997) Juvenile mortality in benthic marine

invertebrates. Mar Ecol Prog Ser 146:265–282

Goudet J (1995) FSTAT (vers. 1.2): a computer program to calculate

F-statistics. J Hered 86:485–486

Grant J, Turner S, Legendre P, Hume T, Bell R (1997) Patterns of

sediment reworking and transport over small spatial scales on an

intertidal sandflat, Manujau Harbour, New Zealand. J Exp Mar

Biol Ecol 216:33–50

Guillou J (1980) Les peuplements de sables fins du littoral Nord-

Gascogne. Universite de Bretagne Occidentale, p 209

Hall TA (1999) BioEdit : a user-friendly biological sequence

alignment editor and analysis program for Windows 95/98/NT.

Nucleic Acid Symp Ser 41:95–98

Harrison R (1993) Hybrids and hybrid zones: historical perspective.

In: Hybrids zones and the evolutionary process. Oxford Univer-

sity Press, Oxford, pp 3–12

Hellberg ME, Burton RS, Neigel JE, Palumbi SR (2002) Genetic

assessment of connectivity among marine populations. Bull Mar

Sci 70:273–290

Hendler G, Miller JE (1991) Swimming ophiuroids—Real and

imagined. In: Yanagisawa Y, Oguro, Suzuki & Motokawa (ed)

Proceedings of the Seventh International Echinoderm Confer-

ence, Biology of Echinodermata. Balkema, Rotterdam, Atami,

9–14 September, 1990

Hendler G, Baldwin CC, Smith DG, Thacker CE (1999) Planktonic

dispersal of juvenile brittle stars (Echinodermata: Ophiuroidea)

on a caribbean reef. Bull Mar Sci 65:283–288

Hendry AP, Day T (2005) Population structure attributable to

reproductive time: isolation by time and adaptation by time.

Mol Ecol 14:901–916

Jiggins CD, Mallet J (2000) Bimodal hybrid zones and speciation.

Trends Ecol Evol 15:250–255

Jollivet D, Comtet T, Chevaldonne P, Hourdez S, Jouin-Toulmond C,

Desbruyeres D, Dixon DR (1998) Unexpected relationship

between dispersal strategies and speciation within the association

Bathymodiolus (Bivalvia) - Branchipolinoe (Polychate) using the

rDNA neutral ITS2 marker. Cah Biol Mar 39:359–362

Jolly MT, Jollivet D, Gentil F, Thiebaut E, Viard F (2005) Sharp

genetic break between Atlantic and English Channel populations

of the polychaete Pectinaria koreni, along the north coast of

Brittany, France. Heredity 94:23–32

Jolly MT, Viard F, Gentil F, Thiebaut E, Jollivet D (2006)

Comparative phylogeography of two coastal polychaete tube-

worms in the North East Atlantic supports shared history and

vicariant events. Mol Ecol 15:1841–1855

Keegan BF, Mercer JP (1986) An oceanographical survey of Killary

Harbour on the West coast of Ireland. Proc R Ir Academy 86:91–

99

Kimura M (1980) A simple method for estimating evolutionary rates

of base substitutions through comparative studies of nucleotide

sequences. J Mol Evol 16:111–120

Knowlton N (1993) Sibling species in the sea. Annu Rev Ecol

Systemat 24:189–216

Koehler R (1921) Faune de France—Echinodermes, Vol 1, Paris, p 210

Lessios HA (1984) Possible prezygotic reproductive isolation in sea

urchins separated by the Isthmus of Panama. Evolution 38:1144–

1148

Levitan DR, Fukami H, Jara J, Kline D, McGovern TM, McGhee KE,

Swanson CA, Knowlton N (2004) Mechanisms of reproductive

isolation among sympatric broadcast-spawning corals of the

Montastraea annularis species complex. Evolution 58:308–323

Mallet J (2005) Hybridization as an invasion of the genome. Trends

Ecol Evol 20:231–237

McCarthy C (1997) Chromas, Version 1.41. Griffith University,

Brisbane

Montagu G (1804) Description of several marine animals found on the

south coast of Devonshire. Trans Linn Soc 7:61–85. pl VI–VII

Muths D, Davoult D, Gentil F, Jollivet D (2006) Incomplete cryptic

speciation between intertidal and subtidal morphs of Acrocnidabrachiata (Echinoderm : Ophiuroidea) in the North East

Atlantic. Mol Ecol 15:3303–3318

678 Genetica (2010) 138:667–679

123

Author's personal copy

Nichols D, Barker MF (1984) A comparative study of reproductive

and nutritional periodicities in two populations of Asteriasrubens (Echinodermata : Asteroidea) from the English Channel.

J Mar Biol Assoc UK 64:471–484

O’Loughlin PM, Waters JM, Roy SM (2002) Description of a new

species of Patiriella from New Zealand, and review of Patiriellaregularis (Echinodermata, Asteroidea) based on morphological

and molecular data. J Roy Soc New Zeal 32:697–711

Palumbi SR (1994) Genetic divergence, reproductive isolation and

marine speciation. Annu Rev Ecol Sys 25:547–572

Palumbi SR, Grabowsky G, Duda T, Greyer L, Tachino N (1997)

Speciation and population genetic structure in tropical pacific sea

urchins. Evolution 51:1506–1517

Pasteur N, Pasteur G, Bonhomme F, Catalan J, Britton-Davidan J

(1987) Manuel technique de genetique par electrophorese des

proteines, Vol, Paris, p 459

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

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

Raymond M, Rousset F (1995) GENEPOP (version 1.2): a population

genetics software for exact tests and ecumenicism. J Hered

86:248–249

Rozas J, Sanchez-DelBarrio JC, Messeguer X, Rozas R (2003)

DnaSP, DNA polymorphism analyses by the coalescent and

other methods. Bioinformatics 19:2496–2497

Schmidt PS, Rand DM (2001) Adaptive maintenance of genetic

polymorphism in an intertidal barnacle: habitat- and life-stage-

specific survivorship of Mpi genotypes. Evolution 55:1336–1344

Schneider S, Roessli D, Excoffier L (2001) Arlequin (v 2.0): software

for population Genetics Data Analysis. Genetics and Biometry

Laboratory. Department of Anthropology, University of Geneva,

Geneva

Stohr S, Muths D (2009) Morphological diagnosis of the two genetic

lineages of Acrocnida brachiata (Echinodermata: Ophiuroidea),

with description of a new species. J Mar Biol Assoc UK, 1–13.

doi:10.1017/S0025315409990749

Tamaki A (1987) Comparison of resistivity to transport by wave

action in several polychaete species on an intertidal sand flat.

Mar Ecol Prog Ser 37:181–189

Thompson JD, Higgins DG, Gibson TJ (1994) CLUSTAL W:

improving the sensitivity of progressive multiple sequence

alignment through sequence weighting, positions-specific gap

penalties and weight matrix choice. Nucleic Acids Res 22:4673–

4680

Tyler PA, Banner FT (1977) The effect of coastal hydrodynamics on

the Echinoderm distribution in the sublittoral of Oxwich Bay,

Bristol Channel. Estuar Coast Shelf Sci 5:293–308

Waples RS (1989) Temporal variation in allele frequencies: testing

the right hypothesis. Evolution 43:1236–1251

Weir BS, Cockerham CC (1984) Estimating F-statistics for the

analysis of population structure. Evolution 38:1358–1370

Wiens JJ, Donoghue MJ (2004) Historical biogeography, ecology and

species richness. Trends Ecol Evol 19:639–644

Wilke T, Pfenninger M (2002) Separating historic events from

recurrent processes in cryptic species: phylogeography of mud

snails. Mol Ecol 11:1439–1451

Zuhlke R, Reise K (1994) Response of macrofauna to drifting tidal

sediments. Helgol Mar Res 48:277–328

Genetica (2010) 138:667–679 679

123

Author's personal copy