EVOLUTION OF PREMATING REPRODUCTIVE ISOLATION AMONG ... · 2E-mail: [email protected] 3 Department of Environmental Biology, University of Rome “La Sapienza,” 00185,

BRIEF COMMUNICATION

doi:10.1111/j.1558-5646.2011.01535.x

EVOLUTION OF PREMATING REPRODUCTIVEISOLATION AMONG CONSPECIFICPOPULATIONS OF THE SEA ROCK-POOLBEETLE OCHTHEBIUS URBANELLIAE DRIVENBY REINFORCING NATURAL SELECTIONDaniele Porretta1,2 and Sandra Urbanelli3

1Department of Ecological and Biological Sciences, University of Tuscia, Largo dell’Universita snc, 01100, Viterbo, Italy2E-mail: [email protected]

3Department of Environmental Biology, University of Rome “La Sapienza,” 00185, Rome, Italy

Received February 21, 2011

Accepted November 10, 2011

Data Archived: Dryad doi:10.5061/dryad.9fs476vt

How natural selection might be involved in speciation remains a fundamental question in evolutionary biology. When two or more

species co-occur in the same areas, natural selection may favor divergence in mating traits. By acting in sympatric but not allopatric

populations, natural selection can also affect mate choice within species and ultimately initiate speciation among conspecific

populations. Here, we address this potential effect in the sea rock-pool beetles Ochthebius quadricollis and O. urbanelliae. The

two species, which inhabit the Mediterranean coasts, co-occurr syntopically in an area along the Italian Tyrrhenian coast and

completed reproductive isolation by reinforcement. In this article, through mating trials under laboratory conditions between

conspecific populations, we found in O. quadricollis no deviations from random mating. Conversely, in O. urbanelliae, we found a

clear pattern of premating isolation between the reinforced populations sympatric with O. quadricollis and those nonreinforced

allopatric. This pattern is consistent with the view that natural selection, which completed the reproductive isolation between

the two species in sympatry, led incidentally also to partial premating reproductive isolation (IPSI estimator from 0.683 to 0.792)

between conspecific populations of O. urbanelliae. This case study supports an until recently underappreciated role of natural

selection resulting from species interactions in initiating speciation.

KEY WORDS: Assortative mating, natural selection, reinforcement, reproductive isolation, speciation.

The way in which natural selection might be involved in speciation

remains one of the most hotly debated questions in evolutionary

biology, with the questions of how natural selection can be in-

volved in initiating speciation of particular interest (Coyne and

Orr 2004; Weissing et al. 2011 for a review). Because speciation

involves the evolution of reproductive isolation between diverg-

ing populations (Coyne and Orr 2004), investigating how natural

selection initiates reproductive isolation is central to understand-

ing its role in the early stages of speciation. One way in which

natural selection acts directly in speciation is through the pro-

cess known as reinforcement. In this process, natural selection

leads to divergence in mating traits by acting against maladap-

tive hybridization between two divergent and potentially inter-

breeding taxa that co-occur in sympatry (Coyne and Orr 2004).

1 2 8 4C© 2012 The Author(s). Evolution C© 2012 The Society for the Study of Evolution.Evolution 66-4: 1284–1295

BRIEF COMMUNICATION

During reinforcement, natural selection and sexual selection may

interact, given that natural selection acts on female preference,

which drives the evolution of divergence in male reproductive

signals (Servedio and Noor 2003; Coyne and Orr 2004; Ritchie

2007). The interplay between the two selective forces, in sym-

patric but not allopatric populations, may therefore lead to a

pattern of geographic divergence in mating traits between con-

specific populations (Ortiz-Barrientos et al. 2009 and references

therein).

Howard in 1993 suggested that during reinforcement, di-

vergence in mate recognition systems in sympatric populations

not only may complete the reproductive isolation between het-

erospecifics, but may also lead incidentally to premating repro-

ductive isolation between conspecific sympatric and allopatric

populations (Howard 1993). This idea was developed further and

empirically demonstrated in the following years (Lemmon 2009;

Ortiz-Barrientos et al. 2009; Pfennig and Pfennig 2009; Hoskin

and Higgie 2010; Rice and Pfennig 2010). The study of Hoskin

et al. (2005) on the green-eyed treefrog Litoria genimaculata first

gave strong empirical evidence of it and is one of the seminal and

most appealing works on this issue. In this system, reinforcement

not only completed reproductive isolation between two differen-

tiated lineages of L. genimaculata, but has also resulted in pre-

mating isolation between populations within the same lineage. In

female-choice trials under laboratory conditions, both allopatric

and sympatric females were indeed observed to mate with males

of their own populations (Hoskin et al. 2005; see also Smadja

and Butlin 2006 for a comment). Following the study of Hoskin

et al. (2005), Pfennig and Ryan (2006), in a simulation study us-

ing an artificial neuronal network, also showed that reproductive

character displacement (defined broadly as “the selective process

by which reproductive traits diverge to minimize costly reproduc-

tive interactions with heterospecifics,” Pfennig and Pfennig 2009)

can lead to reproductive isolation between conspecific popula-

tions, and addressed in which circumstances this is more likely

to occur (see also McPeek and Gavrilets 2006 and Pfennig and

Ryan 2007 for other simulation studies). In addition to the green-

eyed treefrog (Hoskin et al. 2005), some other species systems

furnished empirical support to this prediction, although no full

premating isolation was observed in these systems (for a review,

see Ortiz-Barrientos et al. 2009; Hoskin and Higgie 2010). In most

cases, only sympatric females were observed to reject allopatric

males. The opposite behavior was observed in the Drosophila

serrata/D. birchii system, where allopatric females of D. serrata

rejected conspecific males from the populations sympatric with

D. birchii (Higgie and Blows 2007, 2008). More recently, Ortiz-

Barrientos et al. (2009) have dubbed the mechanism in which

reinforcing selection affects sexual selection within species and,

under some circumstances, leads to speciation events between

sympatric and allopatric populations of a single diverging species,

the Cascade Reinforcement Hypothesis. Finally, Hoskin and Hig-

gie (2010) broadened beyond reinforcement the idea that repro-

ductive character displacement can lead to speciation between

displaced and nondisplaced populations. They included species

interactions other than hybridization that cause selection on mat-

ing traits, such as parasitism and predation, and termed this pro-

cess Reproductive Character Displacement Speciation (Hoskin

and Higgie 2010). Pfennig and Pfennig (2009, 2010) unified eco-

logical character displacement (displacement of traits associated

with resource use in areas where similar species or diverging

taxa co-occur) and reproductive character displacement under the

same conceptual framework and included reinforcement as a par-

ticular case of “character displacement” which occurs by selec-

tion against maladaptive hybridization. These authors addressed

how ecological and reproductive character displacement may pro-

mote each other and how both these processes can be involved

in the evolution of postmating and premating isolation between

incipient species and between conspecific populations within

different selective environments (Pfennig and Pfennig 2009,

2010).

Those species systems in which reproductive isolation oc-

curred as a consequence of divergence in mating traits by natu-

ral selection provide potentially good systems to investigate the

incidental effects on reproductive isolation between conspecific

populations and to assess how important is the role for natural

selection in initiating speciation and in contributing to the origin

of biodiversity (Ortiz-Barrientos et al. 2009; Pfennig and Pfennig

2009, 2010; Hoskin and Higgie 2010).

The sea rock-pool beetles Ochthebius quadricollis and O.

urbanelliae (=Ochthebius sp. A) (Urbanelli et al. 1996; Audisio

et al. 2010) are one of these systems. They are two morpholog-

ically indistinguishable species of hydraenid beetles inhabiting

marine rock pools in the Mediterranean basin, whose ranges are

largely allopatric. A sympatric area has been detected in Italy,

along the southern Tyrrenean coast, where the two species co-

occur syntopically in the same rock pools (Urbanelli et al. 1996;

Urbanelli 2002). By genetic analysis of sympatric populations,

full reproductive isolation between the two species was evidenced

(no F1 hybrids were found) (Urbanelli et al. 1996; Urbanelli

2002). The finding of introgressed individuals at some diagnos-

tic loci exclusively in the area where the two species co-occur

showed that gene flow occurred in the past but then ceased (Ur-

banelli 2002). Premating isolation between the two species in na-

ture was also shown in sympatric populations by analyzing species

composition of mating couples (Urbanelli and Porretta 2008). Fi-

nally, through mating trials under laboratory conditions, greater

premating isolation among sympatric populations than ones al-

lopatric was found, showing that reproductive character displace-

ment occurred in sympatric populations (Urbanelli and Porretta

2008). Reinforcement was suggested to explain the evolution of

EVOLUTION APRIL 2012 1 2 8 5

BRIEF COMMUNICATION

discriminative mate recognition systems occurring among O.

quadricollis and O. urbanelliae under sympatric populations

(Urbanelli and Porretta 2008).

Here, we aim to test whether the evolution of discriminative

mate recognition systems in sympatric populations affected also

the intraspecific premating isolation within O. quadricollis and/or

O. urbanelliae. The mating behavior of these species makes them

especially suitable for investigation into patterns of premating

isolation between individuals from different populations. Indeed,

in both O. quadricollis and O. urbanelliae, the first step of mat-

ing behavior is the formation of couples between females and

males (with the male on the back of the female) that last for about

1 h (Beier 1956; Urbanelli and Porretta 2008). For both species,

we performed two types of multiple-choice mating trials under

laboratory conditions (Coyne et al. 2005). In the first, we used

individuals from different conspecific allopatric populations. In

the second, we used individuals from conspecific allopatric and

sympatric populations. If divergence in mating traits in sympatric

populations affected intraspecific premating isolation within

O. quadricollis and/or O. urbanelliae, a pattern of assortative mat-

ing between sympatric and allopatric pairs of populations would

be expected. In contrast, the null expectation (i.e., that divergence

in mating traits in sympatric populations did not affect premating

isolation between conspecific sympatric and allopatric popula-

tions) is fully random mating between populations regardless of

their geographic origin.

Materials and MethodsSAMPLING

Adult individuals of O. quadricollis and O. urbanelliae were

collected from Italian allopatric and sympatric sites during the

summers of 2007–2008 for O. quadricollis and in the years

from 2007 to 2010 for O. urbanelliae (Tables 1 and 2; Figs. 1B

and 2B).

The geographic ranges of O. quadricollis and O. urbanelliae

are mainly allopatric along the Italian Tyrrhenian coast, but an area

of sympatry was found along the southern Tyrrhenian coast, where

the two species co-occur syntopically in the same rock pools

(Urbanelli et al. 1996; Urbanelli 2002; Urbanelli and Porretta

2008). Therefore, those sites where the species occur alone are

allopatric and those sites where the species co-occur syntopically,

sympatric (Urbanelli et al. 1996; Urbanelli 2002; Urbanelli and

Porretta 2008). Sampling campaigns performed since 1990 have

shown the stability of the sympatric area in time and space (i.e.,

through the years the co-occurrence of both species was confirmed

in all sites, despite some differences in their proportion: 50–80%)

(Urbanelli et al. 1996; Urbanelli 2002; Urbanelli and Porretta

2008; S. Urbanelli, pers. obs.).

Octhebius quadricollis and O. urbanelliae are morphologi-

cally indistinguishable species; so in sympatric populations, they

may be recognized only a posteriori by genetic analyses (see be-

low). To avoid putting heterospecific individuals in the mating

chamber, we used the sympatric populations where one species

occurs at a higher relative percentage (Nerano and Maratea for

O. urbanelliae and Diamante for O. quadricollis). These popu-

lations, by laboratory mating trials and analyses of couples col-

lected from the field during our previous studies, showed as high

an ability to discriminate between homo- and heterospecific as the

sympatric population of Cirella, where the two species occur at a

very similar percentage (Urbanelli and Porretta 2008). Therefore,

it may be assumed that the differences in the relative abundance of

the two species do not result in differences in the discrimination

ability in mating, as observed in other systems (Nosil et al. 2003;

Peterson et al. 2005).

The females and males collected were manually sorted by

body size, placed in different plastic bottles filled with water

from the rock pool of origin, and brought to the laboratory. Here,

they were kept in separate aquaria until crossing.

MARKING

Because the individuals of conspecific populations of both species

are morphologically and genetically indistinguishable (Urbanelli

et al. 1996), we marked the populations used in mating trials

with different colored paints. Before performing the mating trials,

we tested the paints (1) for their toxicity and their ability to

persist on the cuticle of the insect and (2) for their effects on

mating behavior. The tests were performed in three replicates

using individuals of O. quadricollis and O. urbanelliae collected

in May 2007 from Circeo and Sperlonga, respectively (Urbanelli

2002).

First, we marked two groups of 200 individuals (100 males

and 100 females) with yellow and white Polycolor paint (Maimeri

Spa, Milano, Italy), respectively. Each individual was marked by

using a 000 grade camel hairbrush to apply one yellow or white

dot of paint on the dorsal part of abdomen. We used no anesthetic

during marking, and were careful to prevent paint from getting

onto the articulation between the thorax and the abdomen. A

control group of 200 individuals was treated in similar fashion,

but not marked. Each group was placed in aquaria half-filled with

seawater and covered with a net to prevent the beetles from flying

out. Small rocks covered with algae collected from the sampling

sites were put in the aquaria as both physical support and a food

source for the adults during the breeding experiments (Coyne et al.

2005; Urbanelli and Porretta 2008). The number of beetles that

had lost their paint mark and the number of dead beetles in both

marked and unmarked groups was recorded at 12 h after marking.

Only 2–4% of the marked beetles lost their marks, showing that

the paint marks used were durable. Furthermore, there were no

1 2 8 6 EVOLUTION APRIL 2012

BRIEF COMMUNICATION

Ta

ble

1.

Mu

ltip

le-c

ho

ice

mat

ing

tria

lsb

etw

een

Och

theb

ius

qu

adri

colli

sp

op

ula

tio

ns.

The

Dia

man

tep

op

ula

tio

nis

sym

pat

ric

wit

hO

.urb

anel

liae.

All

the

oth

erp

op

ula

tio

ns

stu

die

d

are

allo

pat

ric

wit

hO

.urb

anel

liae

(Fig

.1B

).

Cro

sses

Hom

o-pa

irs

Het

ero-

pair

sO

chth

ebiu

squ

adri

coll

is( ♀×

♂)(♀×

♂)IA

PSI

AB

/B

A

Pop.

-1Po

p.-2

Dat

eof

colle

ctio

nA

AB

BA

BB

AN

I PSI

(SD

)∗ P

(SD

)∗∗

P

allo

patr

ic×

allo

patr

icC

irce

o×

Cas

tiglio

ncel

loJu

ly20

07a

2218

1917

760.

049

(0.1

17)

0.68

90.

994

(0.0

53)

0.83

9Se

ptem

ber

2008

b16

1413

1255

0.09

4(0

.138

)0.

503

0.99

4(0

.085

)0.

837

Cir

ceo

×Pi

zzo

July

2007

a21

2419

2286

0.05

0(0

.111

)0.

666

1.01

0(0

.050

)0.

938

Sept

embe

r20

08b

1817

1615

660.

062

(0.1

26)

0.63

90.

996

(0.0

59)

0.86

4Pi

zzo

×C

astig

lionc

ello

July

2007

a23

2723

2093

0.07

6(0

.106

)0.

478

0.98

8(0

.048

)0.

753

Sept

embe

r20

08b

2118

1917

750.

042

(0.1

18)

0.71

50.

995

(0.0

49)

0.84

6sy

mpa

tric

×al

lopa

tric

Dia

man

te×

Cir

ceo

July

2007

a22

2120

1881

0.06

3(0

.113

)0.

580

0.99

3(0

.501

)0.

838

Sept

embe

r20

08b

2326

1719

850.

155

(0.1

10)

0.15

71.

020

(0.0

81)

0.86

1D

iam

ante

×C

astig

lionc

ello

July

2007

a21

2415

1878

0.15

5(0

.114

)0.

182

1.03

6(0

.092

)0.

737

Sept

embe

r20

08b

2223

2118

840.

075

(0.1

10)

0.50

30.

988

(0.0

52)

0.74

6D

iam

ante

×Pi

zzo

July

2007

a25

2219

2086

0.09

3(0

.110

)0.

403

1.00

7(0

.057

)0.

995

Sept

embe

r20

08b

2419

1719

860.

088

(0.1

15)

0.44

41.

014

(0.0

63)

0.91

4

AA

and

BB

=n

um

ber

of

cou

ple

sco

mp

ose

do

fin

div

idu

als

bel

on

gin

gto

the

sam

ep

op

ula

tio

n;A

Ban

dB

A=

nu

mb

ero

fco

up

les

com

po

sed

of

ind

ivid

ual

sb

elo

ng

ing

tod

iffe

ren

tp

op

ula

tio

ns

(AB

=fe

mal

eo

f

po

p.-

1×

mal

eo

fp

op

.-2;

BA

=fe

mal

eo

fp

op

.-2

×m

ale

of

po

p.-

1);N

=to

taln

um

ber

of

cou

ple

sco

llect

edin

each

cro

ss;I

PSI=

esti

mat

or

of

sexu

alis

ola

tio

n(s

tan

dar

dd

evia

tio

n);

IAPS

IA

B/B

A=

asym

met

ryin

dex

for

het

ero

typ

icm

atin

gs

(sta

nd

ard

dev

iati

on

);∗ P

and

∗∗P

=tw

o-t

ailp

rob

abili

tyo

fre

ject

ing

the

nu

llh

ypo

thes

isin

the

bo

ots

trap

resa

mp

ling

dis

trib

uti

on

wit

h10

0,00

0re

plic

ates

for

the

I PSI

esti

mat

or

and

the

IAPS

IA

B/B

Ain

dex

,res

pec

tive

ly.

Dat

eo

fco

llect

ion

issh

ow

nfo

rea

chcr

oss

;sam

esu

per

scri

pt

lett

er(a

or

b)

mea

ns

that

ind

ivid

ual

su

sed

inth

ecr

oss

esb

elo

ng

toth

esa

me

sam

plin

gca

mp

aig

n.


BRIEF COMMUNICATION

Ta

ble

2.

Mu

ltip

le-c

ho

ice

mat

ing

tria

lsb

etw

een

Och

theb

ius

urb

anel

liae

po

pu

lati

on

s.Th

eM

arat

eaan

dN

eran

op

op

ula

tio

ns

are

sym

pat

ric

wit

hO

chth

ebiu

sq

uad

rico

llis.

All

the

oth

erp

op

ula

tio

ns

stu

die

dar

eal

lop

atri

cw

ith

O.q

uad

rico

llis

(Fig

.2B

).

Cro

sses

Hom

o-pa

irs

Het

ero-

pair

sO

chth

ebiu

sur

bane

llia

e(♀×

♂)(♀×

♂)IA

PSI

AB

/B

A

Pop.

-1Po

p.-2

Dat

eof

colle

ctio

nA

AB

BA

BB

AN

I PSI

(SD

)∗ P

(SD

)∗∗

P

allo

patr

ic×

allo

patr

icE

gnaz

ia×

Tala

mon

eSe

ptem

ber

2007

a15

1814

1360

0.09

9(0.

132)

0.46

10.

995

(0.0

82)

0.83

2Ju

ly20

08b

2522

1917

830.

135(

0.11

1)0.

223

0.98

6(0

.068

)0.

741

July

2009

c22

2420

1884

0.09

6(0.

110)

0.39

50.

989

(0.0

55)

0.77

3Se

ptem

ber

2009

d24

2116

1778

0.15

5(0.

114)

0.18

01.

013

(0.0

85)

0.97

1E

gnaz

ia×

Sper

long

aSe

ptem

ber

2007

a25

2119

2388

0.04

6(0.

110)

0.68

31.

014

(0.0

53)

0.89

7Ju

ly20

08b

2317

1316

690.

159(

0.12

1)0.

187

1.04

4(0

.109

)0.

712

July

2009

c27

1814

1776

0.17

7(0.

117)

0.14

31.

045

(0.1

12)

0.72

0Se

ptem

ber

2009

d21

2415

1878

0.15

6(0.

115)

0.18

41.

037

(0.0

96)

0.73

7E

gnaz

ia×

Gal

lipol

iSe

ptem

ber

2007

a24

1921

2286

−0.0

04(0

.111

)0.

955

1.00

1(0

.038

)0.

928

July

2008

b21

2023

1781

0.01

8(0.

114)

0.89

30.

991

(0.0

54)

0.88

8Ju

ly20

09c

1824

2125

88−0

.051

(0.1

09)

0.62

70.

994

(0.0

39)

0.79

0Se

ptem

ber

2009

d16

1414

1357

0.05

2(0.

138)

0.70

70.

997

(0.0

71)

0.88

7Ta

lam

one

×Sp

erlo

nga

Sept

embe

r20

07a

1417

1315

590.

051(

0.13

3)0.

700

1.01

4(0

.077

)0.

991

July

2008

b27

2421

1890

0.13

4(0.

106)

0.21

80.

980

(0.0

63)

0.64

0Ju

ly20

09c

2429

1722

920.

155(

0.10

5)0.

145

1.05

0(0

.090

)0.

554

Sept

embe

r20

09d

2831

2224

105

0.12

6(0.

097)

0.20

31.

013

(0.0

58)

0.89

0Ta

lam

one

×G

allip

oli

Sept

embe

r20

07a

2021

2024

85−0

.057

(0.1

08)

0.59

20.

996

(0.0

36)

0.89

3Ju

ly20

08b

1618

1718

69−0

.015

(0.1

24)

0.90

31.

002

(0.0

47)

0.92

3Ju

ly20

09c

2220

2519

86−0

.184

(0.1

10)

0.85

41.

002

(0.0

46)

0.96

9Se

ptem

ber

2009

d27

2124

2092

0.04

1(0.

107)

0.69

80.

042

(0.0

42)

0.79

4Sp

erlo

nga

×G

allip

oli

Sept

embe

r20

07a

1615

1215

580.

073(

0.13

4)0.

610

1.02

4(0

.096

)0.

920

July

2008

b19

2620

2186

0.04

1(0.

109)

0.72

51.

004

(0.0

44)

0.95

9Ju

ly20

09c

1418

1915

86−0

.031

(0.1

27)

0.79

31.

005

(0.0

57)

0.97

7Se

ptem

ber

2009

d23

2018

1778

0.10

4(0.

115)

0.37

40.

995

(0.0

61)

0.84

2sy

mpa

tric

×al

lopa

tric

Mar

atea

×Ta

lam

one

Sept

embe

r20

07a

2426

35

580.

734(

0.08

9)<

0.00

11.

697

(1.0

68)

0.60

8Ju

ly20

08b

3633

54

780.

773(

0.07

1)<

0.00

10.

991

(0.6

98)

0.62

2M

arat

ea×

Sper

long

aSe

ptem

ber

2007

a24

353

668

0.74

4(0.

081)

<0.

001

1.97

4(1

.215

)0.

400

July

2008

b18

214

144

0.79

2(0.

087)

<0.

001

0.38

3(0

.427

)0.

103

Mar

atea

×G

allip

oli

Sept

embe

r20

07a

3035

53

730.

788(

0.07

2)<

0.00

10.

782

(0.6

09)

0.37

1Ju

ly20

08b

3631

44

720.

790(

0.07

0)<

0.00

11.

197

(0.8

64)

0.85

5M

arat

ea×

Egn

azia

Sept

embe

r20

07a

3026

25

630.

788(

0.07

5)<

0.00

12.

231

(1.2

86)

0.33

3

Co

nti

nu

ed.


BRIEF COMMUNICATION

Ta

ble

2.

Co

nti

nu

ed.

Cro

sses

Hom

o-pa

irs

Het

ero-

pair

sO

chth

ebiu

sur

bane

llia

e( ♀×

♂)(♀×

♂)IA

PSI

AB

/B

A

Pop.

-1Po

p.-2

Dat

eof

colle

ctio

nA

AB

BA

BB

AN

I PSI

(SD

)∗ P

(SD

)∗∗

P

July

2008

b33

275

469

0.74

5(0.

081)

<0.

001

0.97

1(0

.659

)0.

606

Ner

ano

×Ta

lam

one

July

2009

c27

302

665

0.77

0(0.

076)

<0.

001

2.52

8(1

.390

)0.

203

Sept

embe

r20

09d

3934

45

820.

785(

0.06

7)<

0.00

11.

435

(0.9

98)

0.88

3N

eran

o×

Sper

long

aJu

ly20

09c

1821

34

460.

704(

0.10

5)<

0.00

11.

405

(0.8

71)

0.85

6Se

ptem

ber

2009

d27

324

365

0.79

3(0.

074)

<0.

001

0.95

0(0

.737

)0.

576

Ner

ano

×G

allip

oli

July

2009

c25

337

574

0.66

3(0.

089)

<0.

001

0.86

5(0

.450

)0.

469

Sept

embe

r20

09d

3531

83

770.

732(

0.07

6)<

0.00

10.

519

(0.3

30)

0.08

5N

eran

o×

Egn

azia

July

2009

c31

266

569

0.68

3(0.

090)

<0.

001

0.98

7(0

.577

)0.

647

Sept

embe

r20

09d

3735

47

830.

744(

0.07

2)<

0.00

11.

817

(1.1

76)

0.46

3sy

mpa

tric

×sy

mpa

tric

Mar

atea

×N

eran

oJu

ly20

1025

2418

2188

0.11

7(0.

109)

0.28

71.

022

(0.0

69)

0.80

4Se

ptem

ber

2010

1721

1519

720.

055(

0.12

1)0.

672

1.01

9(0

.072

)0.

901

AA

and

BB

=n

um

ber

of

cou

ple

sco

mp

ose

do

fin

div

idu

als

bel

on

gin

gto

the

sam

ep

op

ula

tio

n;

AB

and

BA

=n

um

ber

of

cou

ple

sco

mp

ose

do

fin

div

idu

als

bel

on

gin

gto

dif

fere

nt

po

pu

lati

on

s(A

B:

fem

ale

of

po

p.-

1×

mal

eo

fp

op

.-2;

BA

:fem

ale

of

po

p.-

2×

mal

eo

fp

op

.-1)

;N=

tota

lnu

mb

ero

fco

up

les

colle

cted

;IPS

I=

esti

mat

or

of

sexu

alis

ola

tio

n(s

tan

dar

dd

evia

tio

n);

IAPS

IA

B/B

A=

asym

met

ryin

dex

for

het

ero

typ

ic

mat

ing

s(s

tan

dar

dd

evia

tio

n);

∗ Pan

d∗∗

P=

two

-tai

lpro

bab

ility

of

reje

ctin

gth

en

ull

hyp

oth

esis

inth

eb

oo

tstr

apre

sam

plin

gd

istr

ibu

tio

nw

ith

100,

000

rep

licat

esfo

rth

eI P

SIes

tim

ato

ran

dth

eIA

PSI

AB/B

Ain

dex

,

resp

ecti

vely

.

Dat

eo

fco

llect

ion

issh

ow

nfo

rea

chcr

oss

;sam

esu

per

scri

pt

lett

er(a

,b,c

,or

d)

mea

ns

that

ind

ivid

ual

su

sed

inth

ecr

oss

esb

elo

ng

toth

esa

me

sam

plin

gca

mp

aig

n.


BRIEF COMMUNICATION

Circeo x Castiglioncello

0

10

20

30

CiCi CaCa CiCa CaCi

N°

ofco

uple

sc o

llect

ed

Couple composition ( )

Circeo x Pizzo

0

10

20

30

N°

ofco

uple

sco

llect

ed

CC PP CP PCCouple composition ( )

Allopatric x Allopatric

Diamante x Circeo

0

10

20

30

40

DD CC DC CD

DD CC DC CD


N°

ofco

uple

sco

l lec t

edDiamante x Castiglioncello

0

10

20

30

40N

°of

coup

les

colle

cted

Couple composition( )

Sympatric x AllopatricA

Pizzo x Castiglioncello

0

10

20

30

40

N°

ofco

uple

sco

llect

ed

PP CC PC CP


Diamante x Pizzo

0

10

20

30

40

DD PP DP PD

N°

ofco

uple

sco

llect

e d


40

40

0 50 100km

Circeo

Castiglioncello

Diamante

Tyrrhenian

Sea

Ad ria t ic

S e a

Pizzo

ItalianPeninsula

Ochthebius quadricollisBallopatric populations

sympatric populations

with O. urbanelliae

Figure 1. (A) Multiple-choice mating trials between conspecific populations of Ochthebius quadricollis. Histograms show the mean

number of couples collected in different replicate trials ± standard deviation (Table 1). The couple composition is shown by using the first

letters of the name of population which the female and male belong to. The population of Diamante is sympatric with O. urbanelliae.

The IPSI value was not significant (∗P > 0.05) either in the allopatric × allopatric or sympatric × allopatric mating trials (Table 1). (B) Map

showing the populations of O. quadricollis studied. The sympatric area between O. urbanelliae and O. quadricollis along the coast of the

Tyrrhenian Sea is shown by the gray rectangle.

significant differences in the mortality rate between marked and

unmarked groups (all χ2 tests P > 0.05), showing that the paint

marks were not toxic for the individuals.

Second, we tested whether the paints affected the number

of couples and the male and/or female choice. We marked, as

described above, 100 males and 100 females with white or yel-

low paints in all possible color combinations (i.e., males white

and females yellow; both males and females white, both males

and females yellow; males yellow and females white). A control

group of 200 individuals (100 males and 100 females) was treated

in similar fashion, without being marked. Each group of beetles

was placed in aquaria prepared as described above. The observed

mating couples were thus removed from the aquaria at intervals of

2 h during the first 12 h of the trial. We did not observe significant

differences in the number of mating couples collected between

the marked and unmarked groups (all χ2 tests P > 0.05). Further-

more, no biases in couple composition caused by different colors

were observed (all χ2 tests P > 0.05) (data available in Dryad).

On the whole, all tests showed that the paint marks used were

durable, nontoxic, and that they did not affect mating behavior of

O. quadricollis or O. urbanelliae.

MATING TRIALS

Within two days of the sampling, breeding experiments were per-

formed in aquaria under the environmental conditions described in

Urbanelli and Porretta (2008). Mating trials consisted of multiple-

choice mating without replacement, in which males and females

of both mating types are placed in a mating chamber (Coyne et al.

2005). In the beetles O. quadricollis and O. urbanelliae, this is the

experimental design that most realistically mimics mate choice in

nature (Urbanelli and Porretta 2008).

For each mating trial, 100 females and 100 males from each

population were used, for a total of 400 individuals for each

mating chamber. The observed couples were removed from the

aquaria at intervals of 2 h during the first 12 h of the trial. We

marked each population with a specific color (yellow or white)

and we alternated the colors between trials.

In a multiple-choice design, the element of choice may di-

minish as individuals are removed from the aquaria and the trial

is lengthened to obtain the maximum number of mates (Casares

et al. 1998). To avoid this issue, we stopped the trials below the

limit of 50% of the potential matings, following the suggestions

of Casares et al. (1998), which empirically demonstrated this to be


BRIEF COMMUNICATION

Egnazia x Talamone

0

10

20

30

EE TT ET TE

N°

ofco

uple

sco

llect

ed


Egnazia x Sperlonga

0

10

20

30

N°

ofco

uple

sc o

llect

ed

EE SS ES SECouple composition ( )

Talamone x Sperlonga

0

10

20

30

N°

ofco

uple

sco

l lect

ed

TT SS TS STCouple composition ( )

Allopatric x Allopatric

Maratea x Egnazia

N°

ofco

uple

sco

llec t

ed

0

10

20

30

40


MM EE ME EM

Nerano x Egnazia

NN EE NE EN0

10

20

30

40

N°

ofco

uple

sco

ll ect

ed

0

10

20

30

40Maratea x Nerano

MM NN MN NMCouple composition ( )

N°

ofco

uple

sco

llect

ed

Sympatric x Sympatric

Maratea x Talamone

0

10

20

30

40

MM TT MT TMCouple composition ( )

N°

ofco

uple

sco

llec t

ed

Nerano x Talamone

NN TT NT TN


0

10

20

30

40

N°

ofco

uple

sc o

llect

ed

Maratea x Sperlonga

0

10

20

30

40

N°

ofco

uple

sc o

llect

ed

MM SS MS SMCouple composition( )

Nerano x Sperlonga

NN SS NS SNCouple composition ( )

0

10

20

30

40

N°

ofco

uple

sco

l lect

ed

Sympatric x AllopatricA

0 50 100km

EgnaziaSperlonga

Maratea

Talamone

Nerano

Ochthebius urbanelliae

Tyrrhenian

Sea

Ad r ia t ic

S e a

Gallipoli

Ita lianP enin su la

Egnazia x Gallipoli

0

10

20

30

40

N°

ofco

uple

sco

ll ect

ed

EE GG EG GE


Sperlonga x Gallipoli

0

10

20

30

40

N°

ofco

uple

sco

llect

ed

SS GG SG GS


Talamone x Gallipoli

0

10

20

30

N°

ofco

uple

sco

llect

ed

TT GG TG GT


Maratea x Gallipoli

0

10

20

30

40

MM GG MG GM

N°

ofco

uple

sco

l lec t

ed


Nerano x Gallipoli

NN GG NG GN


0

10

20

30

40

N°

ofco

uple

sc o

ll ect

ed

40

40

40

40

B


sympatric populations

allopatric populations

with O. quadricollis

Figure 2. (A) Multiple-choice mating trials between conspecific populations of Ochthebius urbanelliae. Histograms show the mean

number of couples collected in different replicate trials ± standard deviation (Table 2). The couple composition is shown by using the

first letter of the name of population which the female and male belong to. The populations of Maratea and Nerano are sympatric with


BRIEF COMMUNICATION

a safe, although conservative, value. Each component (male and

female) of a collected couple belonging to the sympatric samples

of Nerano, Maratea, and Diamante was analyzed by genetic anal-

yses to verify its taxonomic status using the allozymic diagnostic

loci described in Urbanelli (2002), and we discarded the couples

in which heterospecific individuals were found (data available in

Dryad).

We performed two types of mating trials for each species.

In the first, we used individuals from different conspecific al-

lopatric populations. In the second, we used individuals from

conspecific allopatric and sympatric populations. If divergence

in mating traits in sympatric populations affected intraspecific

premating isolation within O. quadricollis and/or O. urbanelliae,

a pattern of assortative mating between sympatric and allopatric

pairs of populations would be expected. In contrast, the null ex-

pectation is that fully random mating occurs between populations

regardless of their geographic origin (Ortiz-Barrientos et al. 2009

and references therein).

Because our results indicated a clear pattern of assorta-

tive mating between sympatric and allopatric populations of O.

urbanelliae (see Results and Discussion section), we performed,

for this species only, a third type of mating trials that involved

crossing individuals from the sympatric populations of Nerano

and Maratea (Fig. 2B). In this case, we would expect to find full

random mating between them.

We carried out at least two replicates of each type of mat-

ing trial (Tables 1 and 2). Briefly, we sampled and used as de-

scribed above: allopatric versus allopatric and sympatric ver-

sus allopatric populations of O. quadricollis, twice (in July

2007 and September 2008); allopatric versus allopatric and

sympatric (Maratea, Fig. 2B) versus allopatric populations of

O. urbanelliae, twice (in September 2007 and July 2008); al-

lopatric versus allopatric and sympatric (Nerano, Fig. 2B) ver-

sus allopatric populations of O. urbanelliae, twice (in July 2009

and September 2009); sympatric versus sympatric populations

(Maratea and Nerano) of O. urbanelliae, twice (in July 2010 and

September 2010).

The degree of sexual isolation among conspecific popula-

tions in both O. quadricollis and O. urbanelliae was assessed by

using the IPSI estimator (Rolan-Alvarez and Caballero 2000) as

implemented in the software JMATING (Carvajal-Rodriguez and

Rolan-Alvarez 2006).

Following Rolan-Alvarez and Caballero (2000), for each of

the four mating combinations (two homotypic aa and bb and two

heterotypic combinations ab and ba), an estimator of pair sexual

isolation (PSI) is computed as the number of observed mating

pairs for a particular combination over the number of matings for

this combination that would be expected if mating among individ-

uals that actually mated was random. The IPSI estimator is com-

puted from the four PSI values (PSIaa, PSIbb, PSIab, and PSIba) and

quantifies the sexual isolation component based mainly on mate

choice by removing the effects of different mating propensities

of distinct populations (Carvajal-Rodriguez and Rolan-Alvarez

2006). It can take values between −1 and 1, where zero indi-

cates random mating, and 1 and −1 indicate that only homo- or

heterotypic matings were observed. Bootstrapping (100,000 re-

samplings) was used to calculate mean bootstrap values, standard

deviations, and the two-tail probabilities for rejecting the null

hypothesis of random mating (IPSI = 0). A comparative study

showed IPSI to be the most reliable estimator of sexual isolation

caused by mating preferences, and also when there is uncertainty

in the frequency of the two morphs or in the mating propensity of

individuals, as in our case (Perez-Figuero et al. 2005; Urbanelli

and Porretta 2008).

Sexual isolation may be asymmetric because of differences in

the ability of females and/or males to discriminate between homo-

and heterotype partners (Svensson et al. 2007). To assess the pos-

sible occurrence of asymmetric isolation, we therefore used the

indices of asymmetry IAPSI as implemented in the software JMAT-

ING (Carvajal-Rodriguez and Rolan-Alvarez 2006). As with IPSI ,

the indices of asymmetry (IAPSI) can be calculated by taking the

ratio of PSI values for the two homo- and heterotypic combi-

nations, respectively, (PSIab and PSIba) (Carvajal-Rodriguez and

Rolan-Alvarez 2006). An IAPSI of 1 indicates symmetric mating

frequencies between the two mating combinations that were com-

pared. Bootstrapping (100,000 resamplings) was used to calculate

mean bootstrap values, standard deviations, and the two-tail prob-

abilities for rejecting the null hypothesis of symmetry in mating

(IAPSI = 1).

Results and DiscussionDifferent patterns of intraspecific matings were observed in the

two species. For O. quadricollis, no deviations from random mat-

ing were observed either in crosses between allopatric populations

or in crosses between allopatric and sympatric populations (for all

IPSI statistics ∗P > 0.05), regardless of the geographic origin of

Figure 2. Continued.

O. quadricollis. In allopatric × allopatric mating trials, for each replicate IPSI value was not significant, ∗P > 0.05. In sympatric × allopatric

mating trials, for each replicate IPSI value was significant, P < 0.001. In sympatric × sympatric mating trials, for each replicate IPSI value

was not significant, ∗P > 0.05 (Table 2). (B) Map showing the populations of O. urbanelliae studied. The sympatric area between O.

urbanelliae and O. quadricollis along the coast of the Tyrrhenian Sea is shown by the gray rectangle.


BRIEF COMMUNICATION

the samples (Fig. 1, Table 1). No deviations from symmetry were

observed in any trials (for all IAPSI, statistics ∗∗P > 0.05). Con-

versely, in O. urbanelliae, although no deviations from random

mating were observed in crosses between allopatric populations

(for all IPSI statistics ∗P > 0.05) and between the sympatric pop-

ulations of Nerano and Maratea (for all IPSI statistics ∗P > 0.05)

(Fig. 2A, Table 2), in the crosses between allopatric and sympatric

populations, highly significant departures from random mating

were observed (for all IPSI statistics ∗P < 0.001) (Fig. 2A, Ta-

ble 2). For O. urbanelliae only, therefore, mating trials showed

(1) a clear pattern of assortative mating between the sympatric

populations with O. quadricollis and those allopatric; (2) random

mating between sympatric populations and between allopatric

populations. This is the mating pattern expected if divergence

in mating traits in sympatric populations affected intraspecific

premating isolation as well.

Sympatric–allopatric populations are separated by unsuit-

able habitats for the beetles such as sand stretches, and rocky

shores with steep slope, whereas sympatric populations are not.

Therefore, the assortative mating observed between sympatric

and allopatric populations of O. urbanelliae could alternatively

be driven by casual changes due to geographic isolation between

groups of populations. This latter hypothesis, however, is un-

likely. Indeed, unsuitable habitats occur also between allopatric–

allopatric populations and the pattern of premating isolation ob-

served in our study does not agree either the geographical barri-

ers between populations or the pattern of genetic differentiation

among them (Urbanelli et al. 1996; Urbanelli 2002). For example,

the allopatric populations used in this study of Sperlonga and Tala-

mone are separated by long sand stretches (see, e.g., Google Earth

for detailed maps). These habitats (associated with the low disper-

sal ability of the species, Beier 1956) can reduce the exchange of

individuals between the two populations. Previous genetic stud-

ies on O. urbanelliae populations using allozyme markers support

this. Indeed, the genetic analyses showed that the populations of

Latium (such as Sperlonga, Fig. 2) and of Tuscany (such as Ta-

lamone, Fig. 2) belong to distinct genetic groups of populations

which show significant values of the Weir and Cockerman (1984)

FST differentiation index (FST = 0.452, P < 0.001), and gene flow

values (Nm(FST)) of 0.30 (Urbanelli 2002). The unsuitable habi-

tats between Sperlonga and Talamone, however, do not affect the

premating pattern between them. Indeed, they show no premating

barriers (Table 2; Fig. 2). More generally, allopatric populations,

regardless of their genetic differentiation or the occurrence of un-

suitable habitats, do not show premating barriers (Table 2; Fig. 2).

On the contrary, premating isolation occurs only between the al-

lopatric populations and those sympatric with O. quadricollis.

Despite the significant premating isolation observed between

conspecific allopatric and sympatric populations of O. urbanel-

liae, we observed some heteropopulation couples in the crosses

between them (Fig. 2A, Table 2). However, it is noticeable that: (1)

under laboratory conditions, premating isolation may be weaker

than in nature (Rundle and Schluter 2004; Jennings and Etges

2010); (2) the observed frequency of heteropopulation couples

was within the observed frequency range of heterospecific cou-

ples in laboratory conditions (11–16%) between sympatric pop-

ulations of O. quadricollis and O. urbanelliae, which are fully

isolated in nature (Urbanelli and Porretta 2008); (3) our experi-

mental system underestimates premating isolation, as we analyze

only the coupling behavior but ignore all postcoupling stages of

mating up to egg fertilization. Therefore, it may be reasonably hy-

pothesized that in nature the premating isolation would be greater

than that observed here. It is also to be noted that, as discussed

above, the low dispersal ability of the species and the occurrence

of unsuitable habitats between sympatric and allopatric groups

of populations can restrict migration between sympatric and al-

lopatric groups of populations (Urbanelli 2002). The analyses of

other stages of mating, such as sperm transfer, in the heteropopu-

lation couples could give us insights into the effective premating

isolation between individuals (Arnqvist and Rowe 2005). Like-

wise, genetic studies aimed at investigating the genetic diver-

gence specifically between allopatric and sympatric populations

would give us insights into the actual reproductive isolation be-

tween these groups of populations in nature (Rice and Pfennig

2010). If reduced gene flow occurs between sympatric and al-

lopatric populations, greater genetic distance would be expected

between sympatric versus allopatric populations than between

sympatric versus sympatric and allopatric versus allopatric pop-

ulations. On the other hand, because natural selection may drive

rapid mating trait divergence (Higgie et al. 2000; Hoskin et al.

2005) and likely acts on a limited number of genes (Coyne and

Orr 2004), allopatric and sympatric populations may not differ

at neutral loci (Ortiz-Barrientos et al. 2009; Hoskin and Higgie

2010). No genetic studies on O. urbanelliae populations have

investigated genetic divergence specifically between sympatric

versus allopatric populations. Highly polymorphic genetic mark-

ers such as microsatellites would be useful to address this issue

(Rice and Pfennig 2010).

The possible contribution of both reproductive and ecological

displacement in the evolution of the mating pattern observed in the

Ochthebius system is another issue that could be the focus of fur-

ther research (Pfennig and Pfennig 2009, 2010). When two similar

species or diverging taxa come into contact, they may also com-

pete for resources as well as for reproduction (Pfennig and Pfennig

2010 and references therein). This is the case, for example, with

the spadefoot toads Spea multiplicata and S. bombifrons. In this

species system, ecological character displacement (in tadpoles’

trophic morphology) and reinforcement contributed both to com-

plete the process of speciation and to initiate speciation between

sympatric and allopatric populations of S. multiplicata (Pfennig


BRIEF COMMUNICATION

and Rice 2007; Rice and Pfennig 2010 and references therein).

If some divergence in resource use has occurred, morphological

differences would be expected between the species or between

conspecific populations (Kawano 2002; Dayan and Simberloff

2005). This is not the case with O. quadricollis and O. urbanel-

liae, where no morphological differences have been observed

between species and conspecific populations (for a brief review

see Urbanelli et al. 1996; Urbanelli and Porretta 2008). However,

it cannot be completely excluded that, in addition to reinforce-

ment, ecological character displacement between the species may

have contributed to the patterns of assortative mating observed

(Pfennig and Pfennig 2009, 2010).

In conclusion, in this article, we tested in two species of sea

rock-pool beetles, O. quadricollis and O. urbanelliae, the pre-

diction that the evolution by natural selection of divergence in

mate recognition systems in sympatric but not allopatric popula-

tions can affect mate choice within species as well. Our results

showed significant premating reproductive isolation between the

populations of O. urbanelliae sympatric with O. quadricollis and

their conspecific allopatric populations. The case of O. urbanel-

liae strongly supports the above prediction, as it does the active

role in initiating speciation of natural selection originating from

species interactions.

ACKNOWLEDGMENTSWe thank V. Mastrantonio, G. Batani, I. Lauria, M. Bitetti for samplecollection; A. Spano for technical assistance, D. Canestrelli for comments;M. Eltenton for the linguistic revision; the Editor and three anonymousreviewers for providing many helpful comments on the manuscript.

LITERATURE CITEDArnqvist, G., and L. Rowe. 2005. Sexual conflict. Princeton Univ. Press,

Princeton, NJ.Audisio, P., M. Trizzino, A. De Biase, G. Rossetti, E. Mancini, and G. An-

tonini. 2010. Molecular and morphological evidence of a new siblingspecies of Calobius (Coleoptera: Hydraenidae) of the C. quadricollis

complex from peninsular Italy. It. J. Zool. 77:29–37.Beier, M. 1956. ´Uber Korperbau und Lebensweise von Ochtebius quadricol-

lis steinbuhleri Rtt. Und O. Adriaticus Rtt. (Col. Hydroph. Hydraen.).Thalassia Jugosl. 1:93–242.

Carvajal-Rodriguez, A., and E. Rolan-Alvarez. 2006. JMATING: a softwarefor the analysis of sexual selection and sexual isolation effects frommating frequency data. BMC Evol. Biol. 6:40–44.

Casares, P., M. C. Carracedo, B. D. Rio, R. Pineiro, L. Garcıa-Flores, andA. R. Barros. 1998. Disentangling the effects of mating propensity andmating choice in Drosophila. Evolution 52:126–133.

Coyne, J. A., and H. A. Orr. 2004. Speciation. Sinauer Associates, Sunderland,MA.

Coyne, J. A., S. Elwyn, and E. Rolan-Alvarez. 2005. Impact of experimentaldesign on Drosophila sexual isolation studies. Direct effect and compar-ison to field hybridization data. Evolution 59:2588–2601.

Dayan, T., and D. Simberloff. 2005. Ecological and community-wide characterdisplacement: the next generation. Ecol. Lett. 8:875–894.

Higgie, M., and M. W. Blows. 2007. Are traits that experience reinforcementalso under sexual selection? Am. Nat. 170:409–420.

———. 2008. The evolution of reproductive character displacement conflictswith how sexual selection operates within a species. Evolution 62:1192–1203.

Higgie, M., S. Chenoweth, and M. W. Blows. 2000. Natural selection and thereinforcement of mate recognition. Science 290:519–521.

Hoskin, C. J., and H. Higgie 2010. Speciation via species interactions:the divergence of mating traits within species. Ecol. Lett. 13:409–420.

Hoskin, C. J., H. Higgie, K. R. McDonald, and C. Moritz. 2005. Reinforcementdrives rapid allopatric speciation. Nature 437:1353–1356.

Howard, D. J. 1993. Reinforcement: origin, dynamics and fate of an evolu-tionary hypothesis. Pp. 46–69 in R. G. Harrison, ed. Hybrid zones andthe evolutionary process. Oxford Univ. Press, New York.

Jennings, J. H., and W. J. Etges. 2010. Species hybrids in the laboratory butnot in nature: a reanalysis of premating isolation between Drosophila

arizonae and D. mojavensis. Evolution 64:587–598.Kawano, K. 2002. Character displacement in giant rhinoceros beetles. Am.

Nat. 159:255–271.Lemmon, E. M. 2009. Diversification of conspecific signals in sympatry:

geographic overlap drives multidimensional reproductive character dis-placement in frogs. Evolution 63:1155–1170.

McPeek, M. A., and S. Gavrilets. 2006. The evolution of female matingpreferences: differentiation from species with promiscuous males canpromote speciation. Evolution 60:1967–1980.

Nosil, P., B. J. Crespi, and C. P. Sandoval. 2003. Reproductive isolation drivenby the combined effects of ecological adaptation and reinforcement.Proc. R. Soc. Lond. B 270:1911–1918.

Ortiz-Barrientos, D., A. Grealy, and P. Nosil. 2009. The genetics and ecologyof reinforcement. Ann. NY Acad. Sci. 1168:156–182.

Perez-Figuero, A., A. Caballero, and E. Rolan-Alvarez. 2005. Comparing theestimation properties of different statistics for measuring sexual isolationfrom mating frequencies. Biol. J. Linn. Soc. 85:307–318.

Peterson, M. A., B. M. Honchak, S. E. Locke, T. E. Beeman, J. Mendoza, J.Green, K. J. Buckingham, M. A. White, and K. J. Monsen. 2005. Rel-ative abundance and the species-specific reinforcement of male matingpreference in the Chrysochus (Coleoptera: Chrysomelidae) hybrid zone.Evolution 22:2639–2655.

Pfennig, K. S., and D. W. Pfennig. 2009. Character displacement: ecologicaland reproductive responses to a common evolutionary problem. Quat.Rev. Biol. 84:253–276.

Pfennig, D. W., and K. S. Pfennig. 2010. Character displacement and theorigins of diversity. Am. Nat. 176:26–44.

Pfennig, K. S., and A. M. Rice. 2007. An experimental test of characterdisplacement’s role in promoting postmating isolation between conspe-cific populations in contrasting competitive environments. Evolution61:2433–2443.

Pfennig, K. S., and M. J. Ryan. 2006. Reproductive character displace-ment generates reproductive isolation among conspecific populations:an artificial neural network study. Proc. R. Soc. Lond. B 273:1361–1368.

Rice, A. M., and D. W. Pfennig. 2010. Does character displacementinitiate speciation? Evidence of reduced gene flow between pop-ulations experiencing divergent selection. J. Evol. Biol. 23:854–866.

Ritchie, M. G. 2007. Sexual selection and speciation annu. Rev. Ecol. Evol.Syst. 38:79–102.

Rolan-Alvarez, E., and A. Caballero. 2000. Estimating sexual selectionand sexual isolation effects from mating frequencies. Evolution 54:30–36.

Rundle, H. D., and D. Schluter. 2004. Natural selection and ecological spe-ciation in sticklebacks. Pp. 192–209 in U. Dieckmann, M. Doebeli,


BRIEF COMMUNICATION

J. A. J. Metz, and D. Tautz, eds. Adaptive speciation. Cambridge Univ.Press, UK.

Servedio, M. R., and M. A. F. Noor. 2003. The role of reinforcement inspeciation: theory and data. Annu. Rev. Ecol. Syst. 34:339–364.

Smadja, C., and R. Butlin. 2006. A new role for reinforcement. Heredity96:422–423.

Svensson, E. I., K. Karlsson, M. Friberg, and F. Eroukhmanoff. 2007. Genderdifferences in species recognition and the evolution of asymmetric sexualisolation. Curr. Biol. 17:1943–1947.

Urbanelli, S. 2002. Genetic divergence and reproductive isolation in theOchthebius (Calobius) complex (Coleoptera: Hydraenidae). Heredity88:333–341.

Urbanelli, S., and D. Porretta. 2008. Evidence of reinforcement of prematingisolation between two species of the genus Ochthebius (Coleoptera:Hydraenidae). Evolution 62:1520–1527.

Urbanelli, S., P. Sallicandro, E. De Vito, E. Colonnelli, and L. Bullini. 1996.Molecular reexamination of the taxonomy of Ochthebius (Calobius)(Coleoptera: Hydraenidae) from the Mediterranean and MacaronesianRegions. Ann. Entomol. Soc. Am. 89:623–636.

Weissing, F. J., P. Edelaar, and G. Sander van Doorn. 2011. Adaptivespeciation theory: a conceptual review. Behav. Ecol. Sociobiol. 65:461–480.

Associate Editor: B. Fitzpatrick


Documents

EVOLUTION OF PREMATING REPRODUCTIVE ISOLATION AMONG ... · 2E-mail: [email protected] 3 Department of Environmental Biology, University of Rome “La Sapienza,” 00185,