26
ORIGINAL PAPER Outbreeding causes developmental instability in Drosophila subobscura Zorana Kurbalija Marina Stamenkovic-Radak Cino Pertoldi Marko Andjelkovic Received: 25 June 2009 / Accepted: 29 November 2009 Ó Springer Science+Business Media B.V. 2010 Abstract A possible effect of interpopulation hybridization is either outbreeding depression, as a consequence of breakdown of coadapted gene complexes which can increase developmental instability (DI) of the traits, or increased heterozygosity, which can reduce DI. One of the principal methods commonly used to estimate DI is the variability of fluctuating asymmetry (FA). We analysed the effect of interpopulation hybridization in Drosophila subobscura through the variability in the wing size and the FA of wing length and width for both sexes in parental, F1 and F2 generations. The results of the wing size per se in intra- and interpopulation hybrids of D. subobscura do not explicitly reveal the significance of either of the two hypotheses. However, the results of the FA of the wing traits give a different insight. The FA of wing length and width generally increases in interpopulation crosses in F1 with respect to the FA in the parental generation, which suggests the possibility that outbreeding depression occurred in the first generation after the hybridization event. We generally observed that the FA values for the wing length and width of interpopulation hybrids were higher in F1 and F2 generations, compared to intrapopulation hybrids in same generations. These results suggest that the association between coadaptive genes with the same evolutionary history are the most probable mechanism that maintains the developmental homeostasis in Drosophila subobscura populations. Keywords Coadapted genome Á Fluctuating asymmetry Á Outbreeding depression Á Wing size Z. Kurbalija (&) Á M. Stamenkovic-Radak Á M. Andjelkovic Institute of Biological Research, University of Belgrade, Despot Stefan Blvd. 142, 11000 Belgrade, Serbia e-mail: [email protected] M. Stamenkovic-Radak Á M. Andjelkovic Faculty of Biology, University of Belgrade, Studentski trg 3, 11000 Belgrade, Serbia C. Pertoldi Department of Ecology and Genetics, Institute of Biological Science, University of Aarhus, Ny Munkegade, Building 540, 8000 A ˚ rhus C, Denmark 123 Evol Ecol DOI 10.1007/s10682-009-9342-0

Outbreeding causes developmental instability in Drosophila subobscura

Embed Size (px)

Citation preview

ORI GIN AL PA PER

Outbreeding causes developmental instabilityin Drosophila subobscura

Zorana Kurbalija • Marina Stamenkovic-Radak •

Cino Pertoldi • Marko Andjelkovic

Received: 25 June 2009 / Accepted: 29 November 2009� Springer Science+Business Media B.V. 2010

Abstract A possible effect of interpopulation hybridization is either outbreeding

depression, as a consequence of breakdown of coadapted gene complexes which can

increase developmental instability (DI) of the traits, or increased heterozygosity, which can

reduce DI. One of the principal methods commonly used to estimate DI is the variability of

fluctuating asymmetry (FA). We analysed the effect of interpopulation hybridization in

Drosophila subobscura through the variability in the wing size and the FA of wing length

and width for both sexes in parental, F1 and F2 generations. The results of the wing size

per se in intra- and interpopulation hybrids of D. subobscura do not explicitly reveal the

significance of either of the two hypotheses. However, the results of the FA of the wing

traits give a different insight. The FA of wing length and width generally increases in

interpopulation crosses in F1 with respect to the FA in the parental generation, which

suggests the possibility that outbreeding depression occurred in the first generation after

the hybridization event. We generally observed that the FA values for the wing length and

width of interpopulation hybrids were higher in F1 and F2 generations, compared to

intrapopulation hybrids in same generations. These results suggest that the association

between coadaptive genes with the same evolutionary history are the most probable

mechanism that maintains the developmental homeostasis in Drosophila subobscurapopulations.

Keywords Coadapted genome � Fluctuating asymmetry � Outbreeding depression �Wing size

Z. Kurbalija (&) � M. Stamenkovic-Radak � M. AndjelkovicInstitute of Biological Research, University of Belgrade, Despot Stefan Blvd. 142,11000 Belgrade, Serbiae-mail: [email protected]

M. Stamenkovic-Radak � M. AndjelkovicFaculty of Biology, University of Belgrade, Studentski trg 3, 11000 Belgrade, Serbia

C. PertoldiDepartment of Ecology and Genetics, Institute of Biological Science, University of Aarhus,Ny Munkegade, Building 540, 8000 Arhus C, Denmark

123

Evol EcolDOI 10.1007/s10682-009-9342-0

Introduction

The anthropogenic activities on natural ecosystems increase the risk of stochastic fluctu-

ations in population size and cause changes in the population genetic structure, which

potentially result in inbreeding or outbreeding depressions (Edmands and Timmerman

2003; Frankham 2005). In disturbed habitats, previously isolated populations may come in

contact and, if individuals from two such populations mate, hybridization of the two

different gene pools will occur (Ross and Robertson 1990; Edmands 1999). Mating

between individuals from genetically different populations which are not taxonomically

distinguishable is called intraspecific or interpopulation hybridization (Barton and Hewitt

1985). Hybridization between different populations can lead to heterosis in the first gen-

eration, followed by outbreeding depression in the consecutive generation (Dobzhansky

1950; Andersen et al. 2002; Edmands 2007).

Hybridization can cause outbreeding depression within the affected population due to

breakdown of coadaptive gene complexes (Dobzhansky 1950). A breakdown of coadap-

tation might be displayed by an individual as a decreased ability to develop an optimal

phenotype due to an increased DI (Leary and Allendorf 1989). Developmental instability is

the product of developmental noise or stress, which affects on individuals capacity to

buffer the processes that otherwise result in the development of the specific phenotype

(Zakharov 1981; Palmer 1996). This can be reflected in decreasing fitness components and

an increase in phenotypic variability (Barton and Hewitt 1985). This effect might be

displayed after hybridization in the F1 generation if the hybridization event occurs between

very distinct genomes (Markow and Ricker 1991). In other cases, disruption of coadaptive

gene complexes might not be observed before the F2 generation. The disruption of coa-

dapted gene complexes in F2 is a result of a recombination of the F1 genomes (Graham

1992; Goldberg et al. 2004). Therefore, F2 genomes consist of genes which have evolved

under different selection pressures (Felley 1980).

There is growing evidence that environmental and genomic stress can induce sig-

nificant levels of developmental instability (DI) (Palmer and Strobeck 1986; Palmer

1994, 1996; Møller and Swaddle 1997; Pertoldi et al. 2006a). Two principal methods are

commonly used to estimate DI. Some studies used phenotypic variance of different

morphological traits, where estimate can be blurred by the presence of genetic and/or

environmental variability (Andersen et al. 2002; Pertoldi et al. 2006a, b). Other studies

used fluctuating asymmetry (FA), defined as small deviations from the perfect bilateral

symmetry in morphological traits. This dissimilarity in expression of a given character

on the left and right side cannot be explained by either genotypic or environmental

differences, since the development of bilateral characters in an individual is ensured by

the same genotype under identical environmental conditions (Palmer and Strobeck

1986).

The increase or decrease of DI as a consequence of the genomic stress has been

explained by two hypotheses: heterozigosity (Lerner 1954) and the genomic coadaptation

(Dobzhansky 1950). The heterozygosity theory predicts that levels of heterozygosity will

be inversely correlated with the level of DI (Lerner 1954; Livshits and Kobyliansky 1985;

Pertoldi et al. 2006a). It has been suggested that heterozygosity has a buffering role

through increased biochemical diversity, which enables a dynamic and stable develop-

mental pathway in changing conditions (Livshits and Smouse 1993). Lerner (1954) sug-

gested that heterozygosity in complex multi-genetics systems provides a mechanism for

maintaining potential plasticity and promoting canalization.

Evol Ecol

123

The genomic coadaptation hypothesis predicts that more balanced coadapted gene

complexes, established over the evolutionary history of the populations via natural

selection, will show higher stability in development over time (Markow 1995).

Whether genomic coadaptation or heterozygosity have influence on DI is still unclear.

Although no clear patterns have been found, several trends emerged. Available data

indicate a tendency of FA to increase with inbreeding and population hybridization (Pal-

mer and Strobeck 1986; Waldmann 1999; Lens et al. 2000; Garnier et al. 2006; Andersen

et al. 2008). There is evidence which suggests a positive correlation between FA and

genomic stress (Leary and Allendorf 1989). However, there are several studies that report

exceptions to these patterns (Clarke et al. 1992; Sheridan and Pomiankowski 1997; Pel-

abon et al. 2005). On the other hand, the relationships of DI (measured by FA) with a

breakdown of coadapted gene complexes and heterozygosity are still unclear (Vøllestad

et al. 1999; Alibert and Auffray 2003).

Inversion polymorphism of Drosophila was used as a model system for studying pro-

cesses involved in adaptation and genetic diversity. As crossing-over is suppressed within

the inversion loops of heterokaryotypes, all genes within the inverted segments segregate

as a linked group, representing one physical and functional unit, called the ‘supergene’, so

the different arrangements can be regarded as ‘allelic’ complexes (Krimbas 1993).

Assuming a relatively long-time of selection on the linked genes within inverted regions,

Dobzhansky (1948) developed the coadaptation hypothesis, which proposed that the

selective value of inversions depends on the combinations of alleles, genes and their

interaction. The important aspect of this hypothesis is the effects of heterosis and fitness

epistasis, causing the evolution of the genes evolve after their origin (Hoffmann et al.

2004). The coadaptation hypothesis presumes that different alleles of genes will be pre-

sented in different gene arrangements, and that interpopulation differences exist for the

allelic combinations of the same arrangement (Hoffmann et al. 2004).

Drosophila subobscura is a Palearctic species which displays rich inversion polymor-

phism on all five acrocentric chromosomes of the set (Krimbas and Loukas 1980; Krimbas

1993) which makes that species a good candidate for studying the above mentioned

hypotheses.

In the present paper we focused on coadaptive aspect of inversion polymorphism in

Drosophila subobscura populations from three ecologically and topologically distinct hab-

itats, knowing that they possess a certain degree of genetic differentiation due to their

different evolutionary histories. The aim of this study is to detect variability of the wing size

and FA of wing length and wing width between inter-population and intra-population hybrids

of D. subobscura. The analysis performed over two generations after hybridization was

aimed at comparing the level of fluctuating asymmetry as measure of DI between intra-

population and interpopulation hybrids through generations. However, the most important

aim of the study was to discover if there was association of coadaptive gene complexes and/or

higher heterozygosity maintaining developmental homeostasis in populations.

Materials and methods

For the present study, D. subobscura flies were sampled in Serbia simultaneously at the end

of June 2006 using fermented fruit traps. The flies were collected from three localities

(beech-B, oak-O and Botanical Garden-BG).

The beech (B) and the oak (O) woods are situated in different expositions on mountain

Goc in central Serbia. These two woods have distinct microclimates. Beech wood features

Evol Ecol

123

higher humidity with dense vegetation coverage, whereas the oak has sparser trees and is

slightly warmer. The third locality is the Botanical Garden (BG) situated in the central,

urban part of Belgrade, with a specific microclimate and surrounded by high anthropogenic

activity.

The flies collected in these three localities were used to obtain isofemale lines (IF) and

they were reared on the common cornmeal-sugar-yeast-agar medium for Drosophila. All

cultures were maintained and all experiments performed under constant laboratory condi-

tions, at 19�C, approx. 60% relative humidity, light of 300 lux and 12/12 h light/dark cycles.

We used 63 IF lines from oak, 38 IF lines from beech and 64 IF lines from the Botanical

Garden population. The progeny of these IF lines formed from the field samples were used

as the parental (P) generation in the experiment. Virgin males and females were separated

within each IF line upon emerging and intra- and interpopulation crosses were made 4 days

after eclosion.

The intra- (B 9 B, O 9 O and BG 9 BG) and interpopulation crosses (B 9 O,

BG 9 O and BG 9 B) were made among IF lines of the three D. subobscura populations.

Both direct and reciprocal crosses were made in order to take into account the potential

maternal effect (i.e., direct cross: male from IF line No1 with female from IF No2, and

reciprocal cross: male from IF No2 crossed with female from IF No1 etc.) (Table 1). The

progeny (6 males and 6 females) from each cross was transferred to fresh vials to obtain F1

and F2 generation, respectively.

The flies from P, F1, and F2 generations, from intra- and interpopulation crosses

(B 9 B, O 9 O, BG 9 BG, B 9 O, BG 9 O, BG 9 B) were frozen (-20�C) and used

for further wing measurements.

Wing length and width analyses

The left and right wings from each fly were cut and mounted on a microscope slide using

double sided scotch (12.7–22.8 mm) and cover slip was placed over them. Each wing was

photographed with a Canon Power Shot camera attached to a Leica stereomicroscope

under 409 magnification. The measurements were performed on photographs, with Image

Tool Software 3.0 (Wilcox et al. 2002). (http://ddsdx.uthscsa.edu/dig/download.html). The

wing length (L) was taken as the distance from the intersection of the third longitudinal

vein (L3) with the anterior cross vein (A1) to the wing tip where the third vein ends.

Table 1 Number and type ofcrosses for direct and reciprocalcrosses

O Oak population, B Beechpopulation, BG botanical gardenpopulation

Type of cross Direct cross Reciprocalcross

No. ofcrosses

Intrapopulation Inter line crosses

O 9 O (63 IF lines) O$ 9 O# O# 9 O$ 63

B 9 B (38 IF lines) B$ 9 B# B# 9 B$ 71

BG 9 BG (64 IF lines) BG$ 9 BG# BG# 9 BG$ 89

Interpopulation

B 9 O B$ 9 O# B# 9 O$ 82

BG 9 O BG$ 9 O# BG# 9 O$ 81

BG 9 B BG$ 9 B# BG# 9 B$ 73

Evol Ecol

123

The wing width (W) was taken as the distance between the ends of the second (L2) and the

fifth longitudinal vein (L5) (Fig. 1).

Statistical analyses

Before interpreting FA estimates, several statistical procedures were completed. The

measurement error was estimated for all samples by the two-way ANOVA on a sample of

30 individuals measured twice (Palmer 1994). There were significant interactions between

wing size and individual FA for both the length (MS = 76.013, p \ 0.01) and width

(MS = 45.918, p \ 0.05), which means that FA has a grater value than the measurement

error. The non-parametric tests, Shapiro–Wilk (W) and Chi-squared (v2), were used to test

(R–L) for departures from normality. There are several avaliable tests for normal distri-

bution, and Shapiro–Wilk is high power test which is optimized for small sample sizes

(N \ 50), and for large sample size we used Chi-squared test. The one-sample t-test was

done to test a departure of the mean of (R–L) from the expected mean of zero. Test for

presence of directional asymmetry (DA) should be conducted in FA studies because, the

presence of DA artificialy inflates the values of certan FA indices (Palmer 1994). To test

size dependence on the absolute FA, linear regression analyses of ((R ? L)/2) on |R–L|

were done for all samples. The FA1 index (Palmer 1994) of each trait was measured as the

absolute (unsigned) |R–L| difference between sides in all samples of intrapopulation and

interpopulation hybrids (both, direct and reciprocal crosses), separately for males and

females through P, F1 and F2 generations. The FA1 index is the one of the most frequently

used index to describe a level of FA in sample. It is also an unbiased estimator of the

sample standard deviation, and recommended for testing FA differences between 3 or more

samples (Palmer and Strobeck 1992).

The F-test and t-test the are very commonly used tests which assume normal distri-

butions. The F-test is for equal variance, while the t-test is for the equality of the means.

These tests were conducted in order to test significant differences in the mean and vari-

ances of the wing length and width between sexes, populations, generations, and type of

cross. All these tests were done using sex, population, generation, and types of cross as

Fig. 1 Wings landmarks used for measuring the two wing traits: wing length distance from the intersectionof the third longitudinal vein (L3) with the anterior cross vein (A1) to the wing tip where the third vein ends;wing width as the distance between the ends of the second (L2) and the fifth longitudinal vein (L5)

Evol Ecol

123

separate variables. The conservative F-test and t-test were used to reduce the possibility of

a Type 1 error. Type 1 error is typically associated with investigations dealing with a large

amount of data, as in the present study. All the statistical analyses were performed using

PAST software (Hammer et al. 2001). Corrections for multiple comparisons were per-

formed using overall Bonferroni correction (Rice 1989).

Results

None of the samples manifest significant deviations from normality (Tables 2, 3) and the

signed right-left (R–L) size analysis show that directional asymmetry (DA) is absent in all

samples. In less than 1% of the samples a positive correlation between |R–L| and the

(R ? L)/2 is found. After sequential Bonferroni correction none of the regressions were

significant, indicating that FA is not correlated with the trait size.

Intrapopulation hybridization

Changes of the mean and variance across generations in males

Analysis of the difference in the mean and variance of the wing length in males is given in

Table 4a. Generally, a significant decrease of the mean is observed in males from P to F2

in all direct and reciprocal crosses. The variance, in general, significantly decreases from P

to F1, and increases in F2, both in direct and reciprocal crosses.

Analysis of the difference in the mean and variance of the wing width is given in

Table 4b. Significant decrease of the mean is obtained in all crosses (both direct and

reciprocal), except in the BG 9 BG direct cross, where different trends through genera-

tions (P [ F1 \ F2) were found. The variance significantly decreases from P to F1, and

towards F2, except in the B 9 B reciprocal cross.

Changes of the mean and variance across generations in females

Analysis of the difference in the mean and variance of the wing length in females is given

in Table 5a. A significant decrease of the mean is observed from P toward F2 in all direct

and reciprocal crosses. The variance generally increases through generations

(P \ F1 \ F2) in both direct and reciprocal crosses.

Analyses of the difference in the mean and variance of the wing width are shown in

Table 5b. A significant decrease of the mean toward F2 is observed both in direct and

reciprocal crosses. The variance change shows no general trend, with a significant dif-

ference between generations, found only in direct crosses.

Changes of the FA across generations in males

Analysis of the FA1 index between generations for wing length in males shows no sig-

nificant differences between generations either in the direct and reciprocal crosses

(Table 6a).

Analysis of FA1 between generations for wing width, shows no significant difference

between generations, except in the hybrids from BG 9 BG reciprocal cross (tP,F1 = -2.17,

p \ 0.05; tP,F2 = -2.30, p \ 0.05) (Table 6b).

Evol Ecol

123

Ta

ble

2R

esult

so

fS

hap

iro

–W

ilk

(W)

and

Ch

i-sq

uar

e(v

2)

test

sfo

rn

orm

ald

istr

ibu

tio

no

fw

ing

len

gth

insa

mple

so

fin

tra-

and

inte

rpo

pu

lati

on

cro

sses

inP

,F

1an

dF

2g

ener

atio

ns

inm

ales

and

fem

ales

(dir

ect

and

reci

pro

cal

cro

sses

)

Cro

ssG

en.

Mal

esF

emal

es

Dir

ect

Rec

ipro

cal

Dir

ect

Rec

ipro

cal

NS

hap

iro

–W

ilk

Ch

i-sq

uar

eN

Sh

apir

o–W

ilk

Ch

i-sq

uar

eN

Sh

apir

o–W

ilk

Ch

i-sq

uar

eN

Sh

apir

o–W

ilk

Chi-

squar

e

Intr

apo

pu

lati

on

O9

OP

16

W=

0.9

6,

p=

0.6

5v2

=2

.00,

p=

0.6

51

6W

=0

.96,

p=

0.7

1v2

=0

.50,

p=

0.7

01

6W

=0

.88,

p=

0.0

5v2

=1

.50,

p=

0.0

51

6W

=0

.92,

p=

0.2

4v2

=8

.5,

p=

0.2

4

F1

62

W=

0.9

8,

p=

0.6

1v2

=1

.35,

p=

0.6

15

0W

=0

.98,

p=

0.7

8v2

=0

.88,

p=

0.7

84

7W

=0

.97,

p=

0.2

9v2

=0

.23,

p=

0.2

84

6W

=0

.97,

p=

0.4

5v2

=1

.82,

p=

0.4

5

F2

11

7W

=0

.96,

p=

0.0

6v2

=1

.13,

p=

0.6

11

10

W=

0.9

9,

p=

0.9

0v2

=1

.78,

p=

0.9

01

46

W=

0.9

9,

p=

0.4

9v2

=1

.78,

p=

0.4

91

05

W=

0.9

7,

p=

0.0

6v2

=1

.62,

p=

0.6

0

B9

BP

13

W=

0.9

3,

p=

0.4

4v2

=0

.84,

p=

0.4

41

3W

=0

.90,

p=

0.1

6v2

=2

.69,

p=

0.1

61

3W

=0

.91,

p=

0.2

0v2

=0

.84,

p=

0.2

01

3W

=0

.82,

p=

0.0

5v2

=2

.69,

p=

0.0

5

F1

45

W=

0.9

7,

p=

0.5

8v2

=2

.34,

p=

0.5

84

5W

=0

.97,

p=

0.4

2v2

=2

.55,

p=

0.4

24

6W

=0

.97,

p=

0.2

8v2

=3

.04,

p=

0.2

94

4W

=0

.96,

p=

0.1

8v2

=4

.54,

p=

0.1

8

F2

97

W=

0.9

8,

p=

0.5

9v2

=0

.36,

p=

0.6

09

8W

=0

.96,

p=

0.0

9v2

=2

.00,

p=

0.0

91

20

W=

0.9

7,

p=

0.0

5v2

=2

.20,

p=

0.0

61

20

W=

2.6

0,

p=

0.0

6v2

=2

.60,

p=

0.0

6

BG

9B

GP

48

W=

0.9

7,

p=

0.1

9v2

=1

.16,

p=

0.1

94

8W

=0

.99,

p=

0.9

6v2

=0

.17,

p=

0.9

64

8W

=0

.98,

p=

0.8

3v2

=1

.16,

p=

0.8

34

8W

=0

.98,

p=

0.8

6v2

=0

.83,

p=

0.8

6

F1

18

8W

=0

.99,

p=

0.6

7v2

=0

.85,

p=

0.6

71

48

W=

0.9

8,

p=

0.0

8v2

=5

.13,

p=

0.0

81

58

W=

0.9

9,

p=

0.9

7v2

=0

.28,

p=

0.9

71

84

W=

0.9

9,

p=

0.8

8v2

=2

.04,

p=

0.8

8

F2

46

4W

=0

.98,

p=

0.0

8v2

=1

.27,

p=

0.0

84

91

W=

0.9

8,

p=

0.6

0v2

=1

.26,

p=

0.0

65

93

W=

0.9

8,

p=

0.1

7v2

=4

.17,

p=

0.1

75

72

W=

0.9

9,

p=

0.0

6v2

=1

.46,

p=

0.0

7

Evol Ecol

123

Ta

ble

2co

nti

nu

ed

Cro

ssG

en.

Mal

esF

emal

es

Dir

ect

Rec

ipro

cal

Dir

ect

Rec

ipro

cal

NS

hap

iro

–W

ilk

Chi-

squar

eN

Sh

apir

o–

Wil

kC

hi-

squar

eN

Sh

apir

o–W

ilk

Ch

i-sq

uar

eN

Shap

iro–W

ilk

Chi-

squar

e

Inte

rpo

pu

lati

on

B9

OP

38

W=

0.9

9,

p=

0.9

7v2

=0

.10,

p=

0.9

73

8W

=0

.98,

p=

0.7

9v2

=1

.78,

p=

0.7

93

8W

=0

.98,

p=

0.9

1v2

=0

.52,

p=

0.9

13

8W

=0

.97,

p=

0.4

9v2

=0

.52

,p

=0

.49

F1

15

5W

=0

.95

,p

=0

.05

v2=

1.4

2,

p=

0.0

91

45

W=

0.9

8,

p=

0.1

1v2

=3

.05,

p=

0.1

11

49

W=

0.9

4,

p=

0.0

5v2

=1

.69,

p=

0.0

81

32

W=

0.9

7,

p=

0.0

5v2

=1

.51

,p

=0

.05

F2

26

3W

=0

.98

,p

=0

.06

v2=

1.7

8,

p=

0.4

92

85

W=

0.9

3,

p=

0.0

5v2

=1

.15,

p=

0.5

13

65

W=

0.9

6,

p=

0.0

6v2

=2

.30,

p=

0.0

52

86

W=

0.9

4,

p=

0.0

5v2

=1

.16

,p

=0

.60

BG

9O

P4

5W

=0

.96

,p

=0

.17

v2=

4.3

3,

p=

0.1

74

5W

=0

.98,

p=

0.7

4v2

=4

.33,

p=

0.7

54

5W

=0

.97,

p=

0.3

5v2

=1

.67,

p=

0.3

54

5W

=0

.97,

p=

0.3

7v2

=0

.95

,p

=0

.37

F1

16

9W

=0

.97

,p

=0

.05

v2=

2.9

1,

p=

0.0

71

35

W=

0.9

8,

p=

0.1

0v2

=3

.47,

p=

0.1

01

69

W=

0.9

8,

p=

0.0

5v2

=1

.62,

p=

0.0

51

35

W=

0.9

8,

p=

0.1

6v2

=1

.14

,p

=0

.16

F2

30

6W

=0

.99

,p

=0

.42

v2=

2.0

5,

p=

0.4

22

65

W=

0.9

8,

p=

0.0

5v2

=6

.09,

p=

0.0

63

64

W=

0.9

9,

p=

0.0

5v2

=6

.34,

p=

0.0

53

37

W=

0.9

8,

p=

0.0

5v2

=8

.64

,p

=0

.05

BG

9B

P3

2W

=0

.96

,p

=0

.06

v2=

2.7

5,

p=

0.0

63

2W

=0

.98,

p=

0.9

6v2

=0

.25,

p=

0.7

63

2W

=0

.97,

p=

0.5

1v2

=0

.25,

p=

0.5

13

2W

=0

.95,

p=

0.1

9v2

=3

.00

,p

=0

.19

F1

16

9W

=0

.98

,p

=0

.05

v2=

1.6

2,

p=

0.0

51

54

W=

0.9

7,

p=

0.0

5v2

=4

.83,

p=

0.7

51

50

W=

0.9

9,

p=

0.4

9v2

=1

.62,

p=

0.4

91

43

W=

0.9

8,

p=

0.0

9v2

=4

.15

,p

=0

.06

F2

42

6W

=0

.98

,p

=0

.05

v2=

4.6

9,

p=

0.0

84

40

W=

0.9

7,

p=

0.0

7v2

=2

.41,

p=

0.1

05

48

W=

0.9

8,

p=

0.0

6v2

=5

.78,

p=

0.0

73

65

W=

0.9

8,

p=

0.0

6v2

=1

.70

,p

=0

.06

OO

akp

op

ula

tio

n,

BB

eech

po

pu

lati

on

,B

Gb

ota

nic

alg

ard

enp

op

ula

tio

n,

gen

gen

erat

ion

,R

rig

ht

sid

e,L

left

side.

No

te:

the

pix

elw

asu

sed

asm

easu

rem

ent

un

it

*p\

0.0

5,

**

p\

0.0

1,

**

*p\

0.0

01

Evol Ecol

123

Tab

le3

Res

ult

so

fS

hap

iro

–W

ilk

(W)

and

Ch

i-sq

uar

e(v

2)

test

sfo

rn

orm

ald

istr

ibu

tio

no

fw

ing

wid

thin

sam

ple

so

fin

tra-

and

inte

rpo

pu

lati

on

cross

esin

P,

F1

and

F2

gen

erat

ion

sin

mal

esan

dfe

mal

es(d

irec

tan

dre

cip

roca

lcr

oss

es)

Cro

ssG

en.

Mal

esF

emal

es

Dir

ect

Rec

ipro

cal

Dir

ect

Rec

ipro

cal

NS

hap

iro

–W

ilk

Ch

i-sq

uar

eN

Sh

apir

o–

Wil

kC

hi-

squar

eN

Sh

apir

o–

Wil

kC

hi-

squar

eN

Sh

apir

o–

Wil

kC

hi-

squar

e

Intr

apop

ula

tio

n

O9

OP

16

W=

0.9

3,

p=

0.3

4v2

=2

.00,

p=

0.3

41

6W

=0

.94,

p=

0.4

3v2

=2

.50,

p=

0.4

31

6W

=0

.96,

p=

0.7

5v2

=0

.5,

p=

0.7

51

6W

=0

.87,

p=

0.0

5v2

=2

.50,

p=

0.0

5

F1

62

W=

0.9

8,

p=

0.7

4v2

=0

.58,

p=

0.7

45

0W

=0

.97,

p=

0.3

9v2

=2

.32,

p=

0.3

94

7W

=0

.97,

p=

0.3

5v2

=1

.77,

p=

0.3

64

6W

=0

.97,

p=

0.3

8v2

=3

.56,

p=

0.3

8

F2

16

0W

=0

.99

,p

=0

.64

v2=

2.1

4,

p=

0.6

41

10

W=

0.9

9,

p=

0.6

8v2

=0

.98,

p=

0.6

81

46

W=

0.9

9,

p=

0.9

4v2

=0

.03,

p=

0.9

41

39

W=

0.9

9,

p=

0.8

3v2

=1

.80,

p=

0.8

3

B9

BP

13

W=

0.9

1,

p=

0.2

0v2

=8

.23,

p=

0.2

01

3W

=0

.93,

p=

0.3

4v2

=4

.53,

p=

0.3

41

3W

=0

.96,

p=

0.8

7v2

=0

.23,

p=

0.8

71

3W

=0

.87,

p=

0.0

6v2

=0

.84,

p=

0.0

6

F1

46

W=

0.9

4,

p=

0.0

6v2

=1

.13,

p=

0.0

64

5W

=0

.98,

p=

0.7

8v2

=0

.24,

p=

0.7

84

6W

=0

.95,

p=

0.0

8v2

=0

.95,

p=

0.0

84

4W

=0

.97,

p=

0.3

2v2

=0

.73,

p=

0.3

2

F2

97

W=

0.9

9,

p=

0.9

6v2

=0

.36,

p=

0.9

69

7W

=0

.99,

p=

0.9

0v2

=1

.10,

p=

0.9

01

20

W=

0.9

8,

p=

0.4

6v2

=4

.00,

p=

0.4

51

20

W=

0.9

9,

p=

0.3

2v2

=8

.46,

p=

0.3

2

BG

9B

GP

48

W=

0.9

6,

p=

0.1

4v2

=2

.83,

p=

0.1

44

8W

=0

.96,

p=

0.2

0v2

=0

.83,

p=

0.2

04

8W

=0

.97,

p=

0.1

5v2

=1

.67,

p=

0.1

54

8W

=0

.98,

p=

0.7

5v2

=2

,50,

p=

0.7

5

F1

15

0W

=0

.99

,p

=0

.71

v2=

0.5

1,

p=

0.7

11

49

W=

0.9

8,

p=

0.1

2v2

=3

.15,

p=

0.1

31

58

W=

0.9

8,

p=

0.0

7v2

=3

.87,

p=

0.4

71

74

W=

0.9

8,

p=

0.1

2v2

=1

.17,

p=

0.1

2

F2

44

7W

=0

.99

,p

=0

.81

v2=

0.8

8,

p=

0.8

14

85

W=

0.9

9,

p=

0.6

4v2

=1

.01,

p=

0.6

35

64

W=

0.9

9,

p=

0.3

9v2

=0

.42,

p=

0.3

95

57

W=

0.9

9,

p=

0.3

8v2

=0

.43,

p=

0.3

8

Evol Ecol

123

Tab

le3

con

tin

ued

Cro

ssG

en.

Mal

esF

emal

es

Dir

ect

Rec

ipro

cal

Dir

ect

Rec

ipro

cal

NS

hap

iro

–W

ilk

Ch

i-sq

uar

eN

Sh

apir

o–

Wil

kC

hi-

squ

are

NS

hap

iro

–W

ilk

Ch

i-sq

uar

eN

Sh

apir

o–

Wil

kC

hi-

squ

are

Inte

rpop

ula

tio

n

B9

OP

38

W=

0.9

8,

p=

0.6

6v2

=0

.52,

p=

0.6

53

8W

=0

.98

,p

=0

.82

v2=

0.5

2,

p=

0.8

33

8W

=0

.97,

p=

0.4

1v2

=2

.00,

p=

0.4

13

8W

=0

.98

,p

=0

.61

v2=

1.1

6,

p=

0.6

1

F1

15

5W

=0

.98,

p=

0.0

6v2

=5

.23,

p=

0.0

71

55

W=

0.9

8,

p=

0.6

6v2

=5

.23

,p

=0

.06

14

9W

=0

.98,

p=

0.1

4v2

=1

.34,

p=

0.1

41

49

W=

0.9

8,

p=

0.1

4v2

=1

.74

,p

=0

.14

F2

25

0W

=0

.99,

p=

0.6

0v2

=2

.48,

p=

0.6

02

84

W=

0.9

9,

p=

0.3

7v2

=0

.31

,p

=0

.37

34

4W

=0

.99,

p=

0.4

4v2

=2

.77,

p=

0.4

43

21

W=

0.9

9,

p=

0.7

7v2

=0

.27

,p

=0

.77

BG

9O

P4

5W

=0

.96,

p=

0.8

7v2

=0

.23,

p=

0.8

74

5W

=0

.97

,p

=0

.54

v2=

1.4

9,

p=

0.5

44

5W

=0

.97,

p=

0.3

1v2

=0

.24,

p=

0.8

64

5W

=0

.98

,p

=0

.65

v2=

2.0

2,

p=

0.6

5

F1

16

5W

=0

.98,

p=

0.1

3v2

=1

.71,

p=

0.1

31

42

W=

0.9

8,

p=

0.0

6v2

=6

.22

,p

=0

.05

14

9W

=0

.99,

p=

0.1

3v2

=1

.04,

p=

0.1

41

35

W=

0.9

8,

p=

0.3

4v2

=1

.94

,p

=0

.08

F2

29

8W

=0

.99,

p=

0.8

1v2

=3

.61,

p=

0.8

02

61

W=

0.9

9,

p=

0.4

8v2

=0

.99

,p

=0

.08

34

6W

=0

.99,

p=

0.0

6v2

=6

.34,

p=

0.0

53

32

W=

0.9

9,

p=

0.9

0v2

=1

.80

,p

=0

.91

BG

9B

P3

2W

=0

.96,

p=

0.2

9v2

=0

.05,

p=

0.2

94

5W

=0

.96

,p

=0

.51

v2=

2.9

3,

p=

0.5

13

2W

=0

.96,

p=

0.2

8v2

=0

.50,

p=

0.2

83

1W

=0

.98

,p

=0

.84

v2=

0.3

8,

p=

0.8

4

F1

16

0W

=0

.98,

p=

0.2

5v2

=4

.85,

p=

0.2

51

42

W=

0.9

8,

p=

0.0

8v2

=1

.94

,p

=0

.08

15

0W

=0

.98,

p=

0.2

2v2

=1

.94,

p=

0.3

21

42

W=

0.9

9,

p=

0.7

9v2

=1

.15

,p

=0

.79

F2

42

6W

=0

.99,

p=

0.0

6v2

=6

.34,

p=

0.0

64

31

W=

0.9

9,

p=

0.5

3v2

=2

.05

,p

=0

.53

54

8W

=0

.99,

p=

0.1

0v2

=5

.45,

p=

0.1

15

11

W=

0.9

9,

p=

0.1

8v2

=2

.73

,p

=0

.18

OO

akp

op

ula

tio

n,

BB

eech

po

pula

tio

n,

BG

Bo

tan

ical

gar

den

po

pula

tio

n,

gen

gen

erat

ion

,R

rig

ht

sid

e,L

left

side.

No

te:

the

pix

elw

asu

sed

asm

easu

rem

ent

un

it

*p\

0.0

5,

**

p\

0.0

1,

**

*p\

0.0

01

Evol Ecol

123

Ta

ble

4T

he

mea

nan

dv

aria

nce

for

(a)

win

gle

ngth

and

(b)

win

gw

idth

inm

ales

of

intr

a-an

din

terp

op

ula

tio

ncr

oss

esac

ross

gen

erat

ion

s(d

irec

tan

dre

cip

roca

lcr

oss

es)

Mea

n(R

?L

)/2

Cro

ssG

en.

Dir

ect

Rec

ipro

cal

NM

ean

±S

Et-

test

pV

aria

nce

F-t

est

pN

Mea

SE

t-te

stp

Var

iance

F-t

est

p

(a)

Win

gle

ngth

Intr

apopula

tion

O9

OP

16

603.9

10.8

8P

[F

1,

t=

2.2

7*

1,8

94.4

6P

[F

1,

F=

2.7

1**

16

590.3

9.0

7P

[F

1,

t=

0.3

21,3

15.6

0P

[F

1,

F=

1.2

6

F1

59

584.2

3.4

4P

[F

2,

t=

4.1

6***

697.6

0P

[F

2,

F=

1.3

150

587.1

4.5

8P

[F

2,

t=

4.7

8***

1,0

47.4

5P

\F

2,

F=

1.0

3

F2

117

561.0

3.5

2F

1[

F2,

t=

4.2

0***

1,4

50.0

1F

1\

F2,

F=

2.0

8**

109

543.1

3.5

3F

1[

F2,

t=

7.2

5***

1,3

57.1

9F

1\

F2,

F=

1.2

9

B9

BP

13

598.2

7.7

6P

[F

1,

t=

2.1

4*

783.6

2P

[F

1,

F=

1.5

713

595.9

7.9

4P

[F

1,

t=

0.1

2820.5

0P

\F

1,

F=

1.2

6

F1

46

582.3

3.3

0P

[F

2,

t=

6.4

2***

499.9

9P

\F

1,

F=

1.1

445

594.8

4.7

8P

[F

2,

t=

4.6

4***

1,0

30.4

5P

\F

2,

F=

1.4

2

F2

97

541.9

3.0

3F

1[

F2,

t=

8.1

5***

892.3

1F

1\

F2,

F=

1.7

8*

97

549.9

3.4

7F

1[

F2,

t=

7.4

2***

1,1

68.3

9F

1\

F2,

F=

1.1

4

BG

9B

GP

48

598.2

3.3

7P

[F

1,

t=

2.7

4**

544.1

9P

[F

1,

F=

1.0

048

602.4

3.5

2P

[F

1,

t=

4.9

2***

596.5

6P

[F

1,

F=

1.3

4

F1

188

587.8

1.7

0P

[F

2,

t=

5.6

0***

543.5

3P

\F

2,

F=

2.1

6**

194

585.2

1.5

1P

[F

2,

t=

6.6

2***

443.3

5P

\F

2,

F=

1.7

2*

F2

463

569.8

1.5

9F

1[

F2,

t=

6.6

2***

1,1

76.2

0P

\F

2,

F=

2.1

7***

490

571.0

1.4

5F

1[

F2,

t=

5.7

1***

1,0

27.0

9F

1\

F2,

F=

2.3

2***

Inte

rpopula

tion

B9

OP

38

586.3

5.2

9P

[F

1,

t=

1.0

71,0

64.1

4P

[F

1,

F=

1.3

638

590.3

4.9

9P

[F

1,

t=

2.1

1*

944.9

0P

[F

1,

F=

1.4

1

F1

155

580.7

2.2

4P

[F

2,

t=

4.4

4***

781.4

1P

[F

2,

F=

1.1

2145

579.9

2.1

5P

[F

2,

t=

4.9

7***

669.4

0P

\F

2,

F=

1.2

6

F2

263

562.3

1.9

0F

1[

F2,

t=

6.0

8***

949.0

7F

1\

F2,

F=

1.2

1284

561.0

2.0

5F

1[

F2,

t=

5.9

2***

1,1

91.2

8F

1\

F2,

F=

1.7

8***

BG

9O

P45

562.1

5.6

2P

\F

1,

t=

-5.5

2***

1,4

22.0

6P

[F

1,

F=

3.2

8***

45

578.0

3.6

0P

\F

1,

t=

-1.6

9581.8

6P

[F

1,

F=

1.0

1

F1

165

585.6

1.6

2P

\F

2,

t=

-0.9

9432.7

8P

[F

2,

F=

1.4

4153

584.8

1.9

4P

[F

2,

t=

0.3

7574.4

7P

\F

2,

F=

2.6

5***

F2

306

567.2

1.8

0F

1[

F2,

t=

6.7

6***

987.9

7F

1\

F2,

F=

2.2

8***

265

575.7

2.4

1F

1[

F2,

t=

0.3

71,5

40.9

5F

1\

F2,

F=

2.6

5***

BG

9B

P32

580.9

4.6

0P

[F

1,

t=

0.0

4677.1

3P

[F

1,

F=

1.4

745

578.0

3.6

0P

\F

1,

t=

-2.6

6***

587.4

4P

\F

1,

F=

2.3

6**

F1

169

580.7

1.6

5P

[F

2,

t=

1.8

6458.8

5P

\F

2,

F=

1.3

6141

604.2

3.1

2P

[F

2,

t=

1.8

8***

1,3

72.2

6P

\F

2,

F=

1.6

3

F2

426

570.6

1.4

7F

1[

F2,

t=

3.9

5***

922.8

6F

1\

F2,

F=

2.0

1***

440

575.4

1.4

7F

1[

F2,

t=

9.1

6***

959.6

7F

1[

F2,

F=

1.4

3**

Evol Ecol

123

Ta

ble

4co

nti

nu

ed

Mea

n(R

?L

)/2

Cro

ssG

en.

Dir

ect

Rec

ipro

cal

NM

ean

±S

Et-

test

pV

aria

nce

F-t

est

pN

Mea

SE

t-te

stp

Var

iance

F-t

est

p

(b)

Win

gw

idth

Inta

rpopula

tion

O9

OP

16

383.1

6.0

9P

[F

1,

t=

2.1

2*

594.2

1P

[F

1,

F=

1.5

616

373.9

5.6

0P

[F

1,

t=

-0.4

9501.8

2P

[F

1,

F=

1.0

2

F1

62

370.9

2.4

7P

[F

2,

t=

3.3

2**

379.6

2P

[F

2,

F=

1.0

250

377.1

3.1

4P

[F

2,

t=

3.9

7***

492.3

4P

\F

2,

F=

1.0

1

F2

117

361.7

2.2

3F

1[

F2,

t=

2.5

8*

583.3

0F

1\

F2,

F=

1.5

4109

350.1

2.1

5F

1\

F2,

t=

-0.4

9502.3

6F

1\

F2,

F=

1.0

2

B9

BP

13

379.4

5.7

7P

[F

1,

t=

0.8

9432.5

2P

[F

1,

F=

1.3

313

377.3

5.9

0P

\F

1,

t=

-1.0

2452.6

7P

\F

1,

F=

3.0

5*

F1

46

374.2

2.6

6P

[F

2,

t=

4.9

9***

324.5

8P

[F

2,

F=

1.2

245

388.3

5.5

4P

[F

2,

t=

3.9

8***

1,3

83.2

7P

[F

2,

F=

1.0

8

F2

97

351.3

1.9

1F

1[

F2,

t=

6.8

9***

353.3

7F

1\

F2,

F=

1.0

997

353.1

2.0

8F

1[

F2,

t=

7.2

7***

418.0

5F

1[

F2,

F=

3.3

1***

BG

9B

GP

48

379.0

2.4

4P

[F

1,

t=

1.7

4285.7

6P

[F

1,

F=

1.1

648

382.1

2.6

8P

[F

1,

t=

3.3

6***

345.6

5P

[F

1,

F=

1.5

2

F1

188

374.5

1.4

4P

\F

2,

t=

-1.2

1246.1

8P

\F

2,

F=

2.4

7***

194

373.5

1.0

8P

[F

2,

t=

4.3

5***

227.8

8P

\F

2,

F=

1.4

8

F2

521

383.7

1.1

6F

1\

F2,

t=

-4.4

8***

704.8

3F

1\

F2,

F=

2.8

6***

491

367.4

1.0

2F

1[

F2,

t=

3.4

4***

511.9

5F

1\

F2,

F=

2.2

5***

Inte

rpopula

tion

B9

OP

38

368.1

4.1

0P

\F

1,

t=

-0.7

5635.6

1P

[F

1,

F=

1.6

7*

38

373.9

3.5

8P

[F

1,

t=

1.7

6488.2

8P

[F

1,

F=

1.9

3**

F1

155

370.9

1.5

7P

[F

2,

t=

2.4

1*

380.2

0P

[F

2,

F=

1.3

6145

368.3

1.3

2P

[F

2,

t=

3.8

8***

252.5

7P

\F

2,

F=

1.0

8

F2

263

358.8

1.3

3F

1[

F2,

t=

5.7

1***

465.3

9F

1\

F2,

F=

1.2

2284

358.5

1.3

7F

1[

F2,

t=

4.5

9***

529.8

6F

1\

F2,

F=

2.0

9***

BG

9O

P45

355.5

3.5

8P

\F

1,

t=

-5.3

1***

576.2

6P

[F

1,

F=

2.5

0***

45

368.6

2.4

0P

\F

1,

t=

-0.5

5259.4

5P

[F

1,

F=

1.1

3

F1

165

371.0

1.1

8P

\F

2,

t=

-1.8

0230.1

2P

[F

2,

F=

1.1

9153

370.0

1.2

2P

[F

2,

t=

0.5

8229.4

5P

\F

2,

F=

2.4

9***

F2

306

361.9

1.2

6F

1[

F2,

t=

4.7

6***

485.7

5F

1\

F2,

F=

2.1

1***

265

366.3

1.5

6F

1[

F2,

t=

1.6

4645.4

0F

1\

F2,

F=

2.8

1***

BG

9B

P32

367.2

3.5

3P

\F

1,

t=

-1.7

1399.9

5P

[F

1,

F=

1.6

8*

32

374.3

3.1

0P

\F

1,

t=

-0.0

6307.3

3P

[F

1,

F=

1.0

4

F1

169

372.6

1.1

9P

[F

2,

t=

-0.2

1238.4

6P

[F

2,

F=

1.0

8154

374.5

1.3

9P

[F

2,

t=

1.1

296.1

6P

\F

2,

F=

1.3

7

F2

426

367.9

0.9

3F

1[

F2,

t=

2.7

9*

368.9

1F

1\

F2,F

=1.5

5**

440

370.2

0.9

8F

1[

F2,

t=

2.3

4*

422.1

2F

1\

F2,

F=

1.4

2

OO

akpopula

tion,

BB

eech

popula

tion,

BG

bota

nic

algar

den

popula

tion,

gen

gen

erat

ion,

Rri

ght

side,

Lle

ftsi

de.

Note

:th

epix

elw

asuse

das

mea

sure

men

tunit

*p

\0.0

5,

**

p\

0.0

1,

***

p\

0.0

01

Evol Ecol

123

Ta

ble

5T

he

mea

nan

dv

aria

nce

for

(a)

win

gle

ngth

and

(b)

win

gw

idth

infe

mal

eso

fin

tra-

and

inte

rpo

pu

lati

on

cro

sses

acro

ssg

ener

atio

ns

(dir

ect

and

reci

pro

cal

cross

es)

Mea

n(R

?L

)/2

Cro

ssG

en.

Dir

ect

Rec

ipro

cal

NM

ean

±S

Et-

test

pV

aria

nce

F-t

est

pN

Mea

SE

t-te

stp

Var

iance

F-t

est

p

(a)

Win

gle

ngth

Intr

apopula

tion

O9

OP

16

642.4

7.9

8P

[F

1,

t=

1.8

21,0

19.7

4P

\F

1,

F=

1.5

316

662.3

7.3

8P

[F

1,

t=

4.7

0***

872.5

0P

\F

1,

F=

1.1

4

F1

47

622.4

5.7

5P

[F

2,

t=

4.9

6***

1,5

56.4

7P

\F

2,

F=

1.2

446

619.9

4.6

5P

[F

2,

t=

7.6

5***

996.4

4P

\F

2,

F=

1.4

8

F2

146

596.3

2.9

4F

1[

F2,

t=

4.2

6***

1,2

66.5

6F

1[

F2,

F=

1.2

3139

590.9

3.0

5F

1[

F2,

t=

4.8

8***

1,2

90.1

0F

1\

F2,

F=

1.2

9

B9

BP

13

638.9

5.4

8P

[F

1,

t=

1.8

5390.7

8P

\F

1,

F=

1.7

913

642.1

6.2

3P

[F

1,

t=

1.9

9505.0

0P

\F

1,

F=

2.0

3

F1

46

624.3

3.9

1P

[F

2,

t=

4.0

2***

701.8

0P

\F

2,

F=

3.7

0*

45

623.1

4.7

7P

[F

2,

t=

4.4

0***

1,0

25.9

9P

\F

2,

F=

2.4

5

F2

120

595.8

3.4

7F

1[

F2,

t=

4.6

6***

1,4

44.4

2F

1\

F2,

F=

2.0

6**

120

598.0

3.2

1F

1[

F2,

t=

4.1

7***

1,2

40.5

8F

1\

F2,

F=

1.2

1***

BG

9B

GP

48

652.8

3.7

7P

[F

1,

t=

4.2

1***

688.4

7P

\F

1,

F=

1.0

348

642.2

3.4

9P

[F

1,

t=

3.2

0**

586.4

6P

[F

1,

F=

1.0

3

F1

158

634.3

2.1

2P

[F

2,

t=

8.0

2***

710.8

4P

\F

2,

F=

1.4

4174

629.8

1.8

1P

[F

2,

t=

5.9

0***

569.9

9P

\F

2,

F=

1.3

1

F2

592

615.3

1.2

9F

1[

F2,

t=

6.9

6***

989.7

7F

1\

F2,

F=

1.3

9*

571

617.9

1.1

6F

1[

F2,

t=

5.0

9***

765.9

9F

1\

F2,

F=

1.3

4*

Inte

rpopula

tion

B9

OP

38

628.1

4.4

0P

\F

1,

t=

-0.4

5734.7

5P

[F

1,

F=

1.3

338

628.2

4.4

2P

[F

1,

t=

0.6

1741.8

3P

\F

1,

F=

1.2

0

F1

149

630.1

1.9

2P

[F

2,

t=

3.6

7***

550.6

8P

\F

2,

F=

1.2

5130

624.9

2.6

1P

[F

2,

t=

5.4

0***

887.3

1P

\F

2,

F=

1.4

1

F2

365

609.3

1.5

9F

1[

F2,

t=

7.4

8***

920.8

3F

1\

F2,

F=

1.6

7***

335

598.7

1.7

7F

1[

F2,

t=

6.0

8***

1,0

47.2

7F

1\

F2,

F=

1.0

5

BG

9O

P45

634.2

5.0

5P

\F

1,

t=

-0.2

41,1

49.1

3P

[F

1,

F=

1.9

1**

39

598.0

5.9

5P

\F

1,

t=

-6.0

0***

1,3

81.8

9P

[F

1,

F=

1.5

6

F1

149

635.3

2.0

1P

[F

2,

t=

4.0

2***

602.3

2P

[F

2,

F=

1.0

1135

632.4

2.5

6P

\F

2,

t=

-1.6

8883.0

6P

\F

2,

F=

1.0

4

F2

346

612.6

1.8

2F

1[

F2,

t=

7.3

8***

1,1

48.1

3F

1\

F2,

F=

1.9

1***

337

608.8

2.0

7F

1[

F2,

t=

6.4

7***

1,4

44.0

2F

1\

F2,

F=

1.6

3**

BG

9B

P32

635.4

6.0

9P

[F

1,

t=

0.5

71,1

88.0

3P

[F

1,

F=

1.7

9*

32

623.8

5.6

9P

\F

1,

t=

-2.0

7*

1,0

37.5

1P

[F

1,

F=

1.2

7

F1

150

632.4

2.1

0P

[F

2,

t=

3.6

3***

663.6

2P

[F

2,

F=

1.3

8143

635.7

2.3

9P

[F

2,

t=

0.9

3814.7

9P

[F

2,

F=

1.0

9

F2

548

615.9

1.2

5F

1[

F2,

t=

6.2

6***

858.8

7F

1\

F2,

F=

1.2

9520

618.6

1.3

5F

1[

F2,

t=

5.9

5***

953.9

6F

1\

F2,

F=

1.1

7

Evol Ecol

123

Ta

ble

5co

nti

nu

ed

Mea

n(R

?L

)/2

Cro

ssG

en.

Dir

ect

Rec

ipro

cal

NM

ean

±S

Et-

test

pV

aria

nce

F-t

est

pN

Mea

SE

t-te

stp

Var

iance

F-t

est

p

(b)

Win

gw

idth

Intr

apopula

tion

O9

OP

16

412.6

5.9

4P

[F

1,

t=

1.7

5565.3

4P

\F

1,

F=

1.3

816

421.9

4.6

1P

[F

1,

t=

4.2

5***

340.4

6P

\F

1,

F=

1.2

4

F1

47

398.9

4.0

8P

[F

2,

t=

4.7

1***

781.2

4P

[F

2,

F=

1.2

646

397.2

3.0

3P

[F

2,

t=

7.3

2***

421.9

0P

\F

2,

F=

1.3

7

F2

146

386.0

1.7

5F

1[

F2,

t=

3.3

4***

447.9

3F

1[

F2,

F=

1.7

4*

139

380.6

1.8

3F

1[

F2,

t=

4.5

5***

468.3

1F

1\

F2,

F=

1.1

1

B9

BP

13

406.5

4.6

1P

[F

1,

t=

1.0

8276.1

8P

\F

1,

F=

1.2

913

403.6

7.0

5P

[F

1,

t=

0.3

8647.0

2P

[F

1,

F=

1.6

4

F1

46

400.3

2.7

8P

[F

2,

t=

2.3

8*

356.2

3P

\F

2,

F=

2.8

6*

44

401.0

2.9

9P

[F

2,

t=

2.5

2*

394.4

5P

[F

2,

F=

1.1

7

F2

120

387.6

2.5

6F

1[

F2,

t=

2.8

2**

789.6

3F

1\

F2,

F=

2.2

2**

120

386.2

2.1

4F

1[

F2,

t=

3.7

4***

552.2

6F

1\

F2,

F=

1.4

0

BG

9B

GP

48

414.3

3.1

8P

[F

1,

t=

2.5

7*

490.8

1P

[F

1,

F=

1.3

348

409.8

2.3

4P

[F

1,

t=

2.5

7*

262.2

6P

\F

1,

F=

1.2

7

F1

158

405.8

1.5

3P

[F

2,

t=

4.9

7***

369.6

1P

\F

2,

F=

1.0

0174

402.3

1.3

8P

[F

2,

t=

3.5

2***

333.0

6P

\F

2,

F=

1.5

0

F2

592

397.7

0.9

1F

1[

F2,

t=

4.1

9***

492.0

0F

1\

F2,

F=

1.3

3*

572

399.4

0.8

3F

1[

F2,

t=

1.7

2393.8

8F

1\

F2,

F=

1.1

8

Inte

rpopula

tion

B9

OP

38

396.5

4.5

3P

\F

1,

t=

-2.6

2**

780.5

6P

[F

1,

F=

2.5

3***

38

401.1

2.7

8P

[F

1,

t=

0.9

2293.5

1P

\F

1,

F=

1.4

5

F1

149

406.0

1.4

4P

[F

2,

t=

1.4

8308.6

2P

[F

2,

F=

1.8

5**

130

397.7

1.8

1P

[F

2,

t=

4.6

5***

425.8

6P

\F

2,

F=

1.5

3

F2

346

391.1

1.1

0F

1[

F2,

t=

7.7

5***

421.7

4F

1\

F2,

F=

1.3

7*

335

384.5

1.1

6F

1[

F2,

t=

6.0

8***

449.5

9F

1\

F2,

F=

1.0

5

BG

9O

P45

399.9

3.7

2P

\F

1,

t=

-0.7

5622.5

0P

[F

1,

F=

1.9

2**

45

379.0

3.6

6P

\F

1,

t=

-5.0

8***

602.0

9P

[F

1,

F=

1.3

6

F1

149

402.4

1.4

8P

[F

2,

t=

2.0

0*

324.6

7P

[F

2,

F=

1.2

1135

398.2

1.8

1P

\F

2,

t=

-2.8

5**

441.9

5P

\F

2,

F=

1.0

9

F2

346

392.6

1.2

2F

1[

F2,

t=

4.6

8***

515.1

4F

1\

F2,

F=

1.5

9**

337

390.5

1.4

0F

1[

F2,

t=

3.0

7**

657.8

8F

1\

F2,

F=

1.4

9**

BG

9B

P32

407.2

3.6

8P

[F

1,

t=

4.6

1***

434.2

7P

\F

1,

F=

1.1

332

397.7

3.2

9P

\F

1,

t=

-2.2

2*

346.4

5P

\F

1,

F=

1.1

3

OO

akpopula

tion,

BB

eech

popula

tion,

BG

bota

nic

algar

den

popula

tion,

gen

gen

erat

ion,

Rri

ght

side,

Lle

ftsi

de.

Note

:th

epix

elw

asuse

das

mea

sure

men

tunit

*p

\0.0

5,

**

p\

0.0

1,

***

p\

0.0

01

Evol Ecol

123

Ta

ble

6T

he

FA

1in

dex

dif

fere

nce

sb

etw

een

gen

erat

ion

san

dty

pe

of

cross

esfo

r(a

)w

ing

len

gth

and

(b)

win

gw

idth

inm

ales

and

fem

ales

ind

irec

tan

dre

cip

roca

lcr

oss

es

FA

1=

|R-

L|

win

gle

ng

th

Cro

ssG

en.

Mal

esF

emal

es

Dir

ect

Rec

ipro

cal

Dir

ect

Rec

ipro

cal

NM

ean

±S

Et-

test

pN

Mea

SE

t-te

stp

NM

ean

±S

Et-

test

pN

Mea

SE

t-te

stp

(a)

Win

gle

ng

th

Intr

apop

ula

tio

n

O9

OP

16

3.5

0.5

9P

[F

1,

t=

0.0

51

63

.90

±0

.75

P[

F1

,t

=1

.35

16

3.4

0.6

3P

[F

1,

t=

1.0

31

62

.86

±0

.40

P\

F1

,t

=-

0.1

4

F1

59

3.4

0.3

9P

[F

2,

t=

0.5

75

02

.99

±0

.30

P[

F2

,t

=2

.17

*4

72

.73

±0

.33

P[

F2

,t

=0

.91

46

2.9

0.3

8P

[F

2,

t=

0.6

1

F2

11

73

.12

±0

.24

F1[

F2

,t

=0

.80

10

92

.71

±0

.18

F1[

F2

,t

=0

.84

14

62

.89

±0

.18

F1\

F2

,t

=-

0.4

41

39

2.5

0.1

6F

1[

F2

,t

=1

.13

B9

BP

13

2.2

0.3

3P

\F

1,

t=

-1

.46

13

2.7

0.4

8P

\F

1,

t=

-0

.77

13

3.1

0.5

5P

[F

1,

t=

0.3

01

33

.11

±0

.36

P\

F1

,t

=-

0.2

4

F1

46

3.3

0.3

9P

\F

2,

t=

-1

.78

45

3.2

0.3

1P

\F

2,

t=

-1

.00

46

2.9

0.3

1P

\F

2,

t=

-0

.19

45

3.2

0.3

8P

\F

2,

t=

-0

.56

F2

97

3.5

0.2

5F

1\

F2

,t

=-

0.3

19

73

.50

±0

.27

F1\

F2

,t

=-

0.6

21

20

3.3

0.2

6F

1\

F2

,t

=-

0.7

51

20

3.4

0,2

1F

1\

F2

,t

=-

0.4

6

BG

9B

GP

48

4.4

0.4

2P

[F

1,

t=

0.1

64

84

.73

±0

.48

P[

F1

,t

=0

.36

48

3.5

0.3

8P

\F

1,

t=

-1

.20

*4

83

.68

±0

.42

P\

F1

,t

=-

1.1

3

F1

18

84

.34

±0

.23

P\

F2

,t

=-

0.0

91

94

4.5

0.2

5P

[F

2,

t=

0.2

01

58

4.5

0.2

6P

\F

2,

t=

-2

.38

**

17

44

.22

±0

.22

P\

F2

,t

=-

1.7

8

F2

46

34

.47

±0

.15

F1\

F2

,t

=-

0.4

54

90

4.6

0.1

6F

1\

F2

,t

=-

0.3

15

92

4.7

0.1

4F

1\

F2

,t

=-

0.5

75

71

4.6

0.1

6F

1\

F2

,t

=-

1.4

4

Evol Ecol

123

Ta

ble

6co

nti

nu

ed

FA

1=

|R-

L|

win

gle

ng

th

Cro

ssG

en.

Mal

esF

emal

es

Dir

ect

Rec

ipro

cal

Dir

ect

Rec

ipro

cal

NM

ean

±S

Et-

test

pN

Mea

SE

t-te

stp

NM

ean

±S

Et-

test

pN

Mea

SE

t-te

stp

Inte

rpop

ula

tio

n

B9

OP

38

4.4

0.8

1P

\F

1,

t=

-1

.71

38

4.1

0.4

7P

\F

1,

t=

-0

.99

38

4.3

0.5

5P

\F

1,

t=

-0

.85

38

5.0

0.6

3P

[F

1,

t=

1.3

3

F1

15

55

.94

±0

.39

P[

F2

,t

=0

.12

14

54

.66

±0

.26

P\

F2

,t

=-

0.2

41

49

4.8

0.2

9P

\F

2,

t=

-0

.54

13

04

.32

±0

.42

P\

F2

,t

=-

0.1

2

F2

26

34

.35

±0

.18

F1[

F2

,t

=4

.17

**

*2

84

4.2

0.1

5F

1[

F2

,t

=1

.56

36

54

.60

±0

.15

F1[

F2

,t

=0

.91

33

55

.17

±0

.26

F1\

F2

,t

=-

1.9

3

BG

9O

P4

54

.10

±0

.37

P\

F1

,t

=-

1.2

94

53

.78

±0

.39

P\

F1

,t

=-

2.1

*4

54

.00

±0

.39

P\

F1

,t

=-

1.4

23

93

.74

±0

.42

P\

F1

,t

=-

1.7

5

F1

16

54

.79

±0

.26

P\

F2

,t

=-

1.2

51

53

5.2

0.3

6P

\F

2,

t=

-1

.03

14

94

.84

±0

.32

P\

F2

,t

=-

2.0

7*

13

54

.89

±0

.33

P\

F2

,t

=-

1.9

7*

F2

30

64

.78

±0

.20

F1[

F2

,t

=0

.03

26

54

.31

±0

.18

F1[

F2

,t

=2

.49

**

34

65

.22

±0

.21

F1\

F2

,t

=-

1.0

23

37

4.9

0.1

9F

1\

F2

,t

=-

0.0

2

BG

9B

P3

23

.73

±0

.33

P\

F1

,t

=-

0.9

53

24

.29

±0

.49

P\

F1

,t

=-

1.2

33

13

.69

±0

.48

P\

F1

,t

=-

1.8

44

83

.68

±0

.43

P\

F1

,t

=-

1.8

2

F1

16

94

.33

±0

.29

P\

F1

,t

=-

2.1

2*

14

15

.17

±0

.32

P\

F2

,t

=-

1.5

01

50

5.1

0.3

3P

\F

2,

t=

-2

.45

**

14

35

.23

±0

.47

P\

F2

,t

=-

2.8

8*

*

F2

42

65

.25

±0

.19

F1\

F2

,t

=-

2.4

5*

*4

40

5.4

0.2

0F

1\

F2

,t

=-

0.6

05

48

5.7

0.1

9F

1\

F2

,t

=-

1.4

55

20

5.4

0.1

8F

1\

F2

,t

=-

0.4

3

Evol Ecol

123

Ta

ble

6co

nti

nu

ed

FA

1=

|R-

L|

win

gle

ng

th

Cro

ssG

en.

Mal

esF

emal

es

Dir

ect

Rec

ipro

cal

Dir

ect

Rec

ipro

cal

NM

ean

±S

Et-

test

pN

Mea

SE

t-te

stp

NM

ean

±S

Et-

test

pN

Mea

SE

t-te

stp

(b)

Win

gw

idth

Intr

apop

ula

tio

n

O9

OP

16

1.8

0.4

7P

\F

1,

t=

-1

.98

16

2.9

0.5

9P

[F

1,

t=

0.6

41

62

.46

±0

.51

P\

F1

,t

=-

1.4

71

62

.15

±0

.63

P\

F1

,t

=-

0.4

5

F1

62

3.1

0.3

1P

\F

2,

t=

-1

.92

50

2.5

0.2

7P

[F

2,

t=

1.2

34

73

.53

±0

.39

P\

F2

,t

=-

0.3

84

62

.41

±0

.27

P\

F2

,t

=-

1.2

1

F2

11

73

.03

±0

.21

F1[

F2

,t

=0

.38

10

92

.27

±0

.17

F1[

F2

,t

=0

.80

14

62

.65

±0

.16

F1[

F2

,t

=2

.46

*1

39

2.8

0.1

7F

1\

F2

,t

=-

1.1

9

B9

BP

13

2.0

0.3

3P

\F

1,

t=

-1

.44

13

2.3

0.2

9P

[F

1,

t=

-0

.18

13

2.4

0.4

7P

[F

1,

t=

0.1

51

32

.94

±0

.67

P[

F1

,t

=0

.33

F1

46

2.7

0.2

6P

[F

2,

t=

0.8

44

52

.47

±0

.29

P[

F2

,t

=-

0.8

24

62

.34

±0

.32

P\

F2

,t

=-

0.3

04

42

.73

±0

.30

P\

F2

,t

=-

0.3

6

F2

97

3.1

0.2

5F

1[

F2

,t

=-

0.8

49

72

.91

±0

.24

F1\

F2

,t

=-

1.0

91

20

2.6

0.2

0F

1\

F2

,t

=-

0.7

71

20

3.1

0.2

0F

1\

F2

,t

=-

1.1

8

BG

9B

GP

48

3.1

0.4

0P

\F

1,

t=

-2

.17

*4

84

.22

±0

.44

P\

F1

,t

=-

0.1

74

84

.35

±0

.46

P\

F1

,t

=-

0.8

94

83

.43

±0

.35

P\

F1

,t

=-

2.4

6*

F1

18

84

.15

±0

.22

P\

F2

,t

=-

2.3

0*

19

44

.32

±0

.25

P\

F2

,t

=-

0.5

91

58

4.8

0.2

6P

[F

2,

t=

0.0

51

74

4.6

0.2

4P

\F

2,

t=

-2

.00

*

F2

52

14

.37

±0

.16

F1\

F2

,t

=-

0.7

14

91

4.5

0.1

6F

1\

F2

,t

=-

0.7

35

92

4.3

0.1

5F

1[

F2

,t

=1

.54

57

24

.64

±0

.17

F1[

F2

,t

=0

.04

Evol Ecol

123

Ta

ble

6co

nti

nu

ed

FA

1=

|R-

L|

win

gle

ng

th

Cro

ssG

en.

Mal

esF

emal

es

Dir

ect

Rec

ipro

cal

Dir

ect

Rec

ipro

cal

NM

ean

±S

Et-

test

pN

Mea

SE

t-te

stp

NM

ean

±S

Et-

test

pN

Mea

SE

t-te

stp

Inte

rpop

ula

tio

n

B9

OP

38

2.8

0.4

2P

\F

1,

t=

-3

.22

**

38

3.6

0.4

3P

\F

1,

t=

-1

.32

38

4.0

0.4

0P

\F

1,

t=

-1

.69

38

3.8

0.4

7P

\F

1,

t=

-1

.58

F1

15

55

.28

±0

.37

P\

F2

,t

=-

2.6

1*

*1

45

4.4

0.3

0P

\F

2,

t=

-2

.14

*1

49

5.2

0.3

5P

\F

2,

t=

-1

.13

13

05

.19

±0

.43

P\

F2

,t

=-

2.0

5*

F2

26

34

.67

±0

.26

F1[

F2

,t

=1

.37

28

45

.08

±0

.24

F1\

F2

,t

=-

1.5

03

46

4.8

0.2

1F

1[

F2

,t

=1

.17

33

55

.54

±0

.27

F1\

F2

,t

=-

0.7

0

BG

9O

P4

53

.32

±0

.38

P\

F1

,t

=-

2.5

3*

45

4.2

0.4

7P

\F

1,

t=

-1

.06

45

3.8

0.5

1P

\F

1,

t=

-1

.61

45

3.8

0.4

0P

\F

1,

t=

-2

.71

**

F1

16

54

.67

±0

.26

P\

F2

,t

=-

2.1

9*

15

34

.86

±0

.30

P\

F2

,t

=-

0.3

01

49

5.1

0.4

2P

\F

2,

t=

-1

.57

13

55

.46

±0

.31

P\

F2

,t

=-

0.8

7

F2

30

64

.55

±0

.21

F1[

F2

,t

=0

.34

26

54

.38

±0

.22

F1[

F2

,t

=1

.29

34

64

.72

±0

.18

F1[

F2

,t

=1

.18

33

74

.35

±0

.19

F1[

F2

,t

=3

.09

**

BG

9B

P3

23

.40

±0

.66

P\

F1

,t

=-

1.1

43

23

.94

±0

.59

P\

F1

,t

=-

1.0

53

24

.20

±0

.42

P\

F1

,t

=-

0.6

73

25

.98

±1

.51

P[

F1

,t

=0

.75

F1

16

95

.07

±0

.39

P\

F2

,t

=-

0.5

31

54

4.7

0.3

4P

\F

2,

t=

-1

.21

15

74

.73

±0

.35

P\

F2

,t

=-

0.5

61

43

5.2

0.3

7P

[F

2,

t=

1.5

4

F2

42

64

.34

±0

.17

F1[

F2

,t

=2

.00

*4

40

4.8

0.2

0F

1\

F2

,t

=-

0.2

15

48

4.5

0.1

6F

1[

F2

,t

=0

.41

52

04

.79

±0

.18

F1[

F2

,t

=1

.22

OO

akp

op

ula

tio

n,

BB

eech

po

pula

tio

n,

BG

bo

tan

ical

gar

den

po

pula

tio

n,

gen

gen

erat

ion

,R

rig

ht

sid

e,L

left

side.

No

te:

the

pix

elw

asu

sed

asm

easu

rem

ent

un

it

*p\

0.0

5,

**p\

0.0

1,

**

*p\

0.0

01

Evol Ecol

123

Changes of the FA across generations in females

Analysis of FA1 between generations for wing length in females shows a significant

increase of FA through generations (P \ F1 \ F2) only in direct cross BG 9 BG

(tP,F1 = -1.20, p \ 0.05; tP,F2 = -2.38, p \ 0.001) (Table 6a).

The FA1 analysis between generations for wing width shows no significant differences

between generations (Table 6b), except in the hybrids from the direct cross O 9 O

(tF1,F2 = 2.46, p \ 0.05). In the hybrids of the reciprocal crosses within the BG population,

a significant difference is obtained between generations (tP,F1 = -2.46, p \ 0.05;

tP,F2 = -2.00, p \ 0.05).

Interpopulation hybridization

Changes of the mean and variance across generations in males

The analysis of the mean and variance of the wing length in males is given in Table 4a. A

significant decrease of the mean (as the one observed for intrapopulation hybrids) is found

in the hybrids from B 9 O, both in direct and reciprocal crosses. But, in the hybrids from

BG 9 O and BG 9 B, a different general pattern is obtained. In direct crosses, the mean

significantly increases in F1 and significantly decreases in the F2 generation

(P \ F1 [ F2). The same tendency is found in hybrids of the BG 9 B reciprocal crosses.

The variance, in general, significantly decreases in F1 and increases in F2, except in the

hybrids of BG 9 B, where the opposite was obtained in the reciprocal crosses.

Analysis of the mean and variance of the wing width in males is given in Table 4b. The

same general pattern as for the wing length is obtained, with a significant decrease of the

mean through generations, in B 9 O, both in direct and reciprocal crosses. In hybrids from

direct crosses of BG 9 O and BG 9 B a significant increase of the mean in F1 is followed

by a significant decrease in F2.

The variance, in general, significantly decreases in F1 and an increase in F2, both in

direct and reciprocal crosses (Table 4b.).

Changes of the mean and variance across generations in females

Analysis of the mean and variance of the wing length in females is shown in Table 5a. The

mean significantly decreases in hybrids from both direct and reciprocal B 9 O crosses. In

the hybrids from the BG 9 O cross, the mean increases in F1 and significantly decreases in

the F2 generation. A similar response is found in the reciprocal crosses of BG 9 O.

However, in the hybrids from BG 9 B, the mean significantly increases in consecutive

generations of direct crosses, and a different trend is obtained in hybrids from the reci-

procal crosses (P \ F1 [ F2). The variance, in general, significantly decreases in F1 and

increases in F2, in both direct and reciprocal crosses.

Analysis of the mean and variance of the wing width in females shows a significant

increase of the mean in F1 and a significant decrease in the F2 generation in B 9 O hybrids

from direct crosses (Table 5b). The mean significantly decreases through generations in

reciprocal crosses (P [ F1 [ F2). In BG 9 O hybrids the mean increases in the F1 gen-

eration and significantly decreases in F2, in both direct and reciprocal crosses. However, in

the hybrids from the BG 9 B direct cross, a significant decrease of the mean in F1 and a

significant increase in F2 is found. In the reciprocal crosses, a different result is obtained

(P \ F1 [ F2).

Evol Ecol

123

The variance, in general, significantly decreases in F1 and increases in F2 in the hybrids

from both direct and reciprocal crosses, with the exception of BG 9 B (P \ F1 [ F2) in

direct crosses (Table 5b).

Changes of FA across generations in males

Analysis of the FA1 between generations for wing length in males is shown in Table 6a. A

significant increase of FA in F1 and its significant decrease in F2 is observed for B 9 O

hybrids both from direct and reciprocal crosses. In the hybrids from the BG 9 B direct

cross, FA1 increases across generations (P \ F1 \ F2).

The FA1 index analysis between generations for wing width in males is shown in

Table 6b. In the hybrids from B 9 O (direct crosses) FA significantly increases in F1 and

decreases in F2. However, a different trend (P \ F1 \ F2) is found in the reciprocal

crosses. In direct crosses, BG 9 O and BG 9 B, a trend (P \ F1 [ F2) is obtained,

without significant differences in reciprocal crosses.

Changes of the FA across generations in females

The results of the FA1 analysis between generations for wing length in females are shown

in Table 6a. The FA1 significantly increases across generations in the hybrids both from

direct and reciprocal BG 9 O crosses. Also, in the hybrids from BG 9 B direct crosses,

FA1 increases through generations (P \ F1 \ F2) and the same significant trend is found

in the hybrids from the reciprocal crosses.

Analysis of FA1 between generations for wing width in females is shown in Table 6b.

No significant difference between generations in the hybrids from the B 9 O direct crosses

is obtained, but in the reciprocal crosses, a significant increase of FA between generations

(P \ F1 \ F2) is observed. In the hybrids from the BG 9 O reciprocal crosses, FA sig-

nificantly increases in F1 and decreases in F2. No significant difference of FA is obtained

between generations of BG 9 B both in direct and reciprocal crosses.

Discussion

In the present paper we focused on the coadaptive aspect of genetic variability at popu-

lation level, and its relation to genomic stress, such as interpopulation hybridization.

Populations of D. subobscura from three ecologically and topologically distinct habitats

were analyzed, presuming that they possess a certain degree of genetic differences due to

their different evolutionary histories. Previous analyses of the inversion polymorphism

show that these populations differ in the frequencies of some gene arrangements (Andj-

elkovic et al. 2003; Stamenkovic-Radak et al. 2008; Jelic et al. 2009). The different gene

arrangements are carriers of various alleles that are differently favoured in diverse envi-

ronmental conditions and prove in most cases to be the major factor determining the gene

arrangement frequencies in natural populations of D. subobscura (Andjelkovic et al. 2003).

Assuming that all genes within inversion segregate as a linked group, and that they have

existed together for relatively long-time under selection, we could considered each

inversion as a coadapted gene complex (Krimbas 1993). The coadaptation hypothesis

presumes that different alleles of genes will be presented in different gene arrangements,

and that interpopulation differences exist for the allelic combinations of the same

Evol Ecol

123

arrangement (Hoffmann et al. 2004), which makes these three populations suitable for

testing the coadaptation versus heterozygosity hypothesis.

In most studies with different Drosophila species and with various traits (Anderson

1968; David 1979), heterosis is found in F1 hybrids, and lost in F2, as a consequence of

disruption of the balanced polygene complexes. This supports the coadaptation hypothesis.

Our results show that the means of the wing length and width for intrapopulation hybrids

significantly decrease in F1 and F2 generations with respect to parental values in both

sexes, in direct and reciprocal crosses. As opposed to that, hybridization between eco-

logically different and distant populations causes an increase of the wing size in F1 and a

decrease in F2. This confirms that distance is important in hybridization between the

closely situated beech and oak populations. Their hybrids show a similar decrease of the

mean wing size through generations as hybrids from intrapopulation crosses of each

population, suggesting that a higher gene flow is probably involved between these two

populations. Some studies showed evidence of coadaptation in this species when different

experimental design, populations, and traits were considered (Orengo and Prevosti 1996;

Banerjee and Singh 1998).

The results concerning wing size and variance per se obtained in intra- and interpop-

ulation hybrids of D. subobscura in our paper cannot explicitly reveal the significance of

either of the two hypotheses. However, the observed results of the FA of the wing traits,

give a different insight.

It is generally believed that the degree of DI in hybrids is related to the genetic distance

between hybridizing populations, as a result of the balance between the stabilizing effect

due to the increased heterozygosity and the disruption caused by the breakdown of the

coadapted gene complexes (Markow and Ricker 1991). Investigations of the genetic basis

of the DI, using several approaches, gave inconclusive results, for several reasons. Both

genomic coadaptation theory and the heterozygosity theory lead to similar predictions of

the DI decrease in hybrids with respect to their parental strains (Andersen et al. 2002). A

problem associated with the hybrid approach is that parental strains are not completely

homozygous, especially in the lines from the field. Furthermore, the negative correlation

between heterozygosity and DI may only become apparent in certain ecological and

population contexts, possibly because DI is related to various exogenic and endogenic

factors (Pertoldi et al. 2006b). Also, to obtain the highest fitness level, sexual reproducing

individuals have to aim for optimal outbreeding. Optimal outbreeding is where the crosses

between two individuals can raise offspring with highest fitness (fitness optimum) without

expression of any detrimental effects. The optimal outbreeding can be described as a

continuous variable that is under influence by other variables like population size and is

flanked by the two extremes of inbreeding and outbreeding (Sagvik et al. 2005).

In the interpopulation hybrids we generally observed an increase of the FA in F1, which

suggests that outbreeding depression occurred in the first generation after hybridization.

Theoretically, outbreeding depression occurs as a consequence of disruption of the coa-

dapted genome phenomenon which can occur after the second generation upon hybrid-

ization (Graham 1992). The Dobzhansky-Muller model (Orr and Turelli 2001) suggested

that outbreeding depression is considered to be more likely expressed in the F2, because

different populations can accumulate genetic incompatibilities more easily (i.e., alleles

with negative epistatic interactions). It has become widely accepted that the evolution of

epistatic incompatibility is explained by the observation that isolated populations gradually

accumulate neutral or advantageous mutations over time. Furthermore, selection for

positive epistasis may result in development of unique coadapted gene complexes within

each isolated population (Whitlock et al. 1995; Fenster et al. 1997).

Evol Ecol

123

The occurrence of significant outbreeding depression in F1 generation of populations

which are not too distant from each other is more surprising. Outbreeding depression is

usually considered an important issue when divergent populations, often recognized as

distinct subspecies, are brought into contact, rather than two demes in a local metapopu-

lation. But, there is some evidence that outbreeding depression may also occur in the F1,

due to factors such as underdominance, epistatic interactions (heterozygote-heterozygote

interactions or interaction involving sex chromosomes) or disruption of local adaptations

(i.e., extrinsic isolation) (Edmands 2007; Escobar et al. 2008). Furthermore, if an indi-

vidual have parents originating from different populations its fitness depends on the type of

selection acting on the parental populations. For example, if recessive deleterious alleles

are present in the parental populations in high frequency, hybridization will mask such

recessive variation and heterosis is expected. In contrast, under some scenarios of over-

dominant selection, outbreeding depression might be expected, particularly if different

pairs of alleles are most fit in different populations or if only one of the populations is

subject to overdominant selection at the locus (Charlesworth and Hughes 1999; Garnier

et al. 2006; Andersen et al. 2008).

We generally observed that the FA value for wing length and width, of interpopulation

hybrids is higher in F2 generation compared to the FA value in intrapopulation hybrids, for

both sexes. This was expected if we take into account that the intrapopulation crosses, in

fact represents the simulation of random mating in parental natural populations. In each

parental population, new mutations may arise and in increase of FA but, if these mutations

are either neutral or advantageous for the populations they are predicted to accumulate and

become fixed in the genome. Here they may be incorporate in specific gene complexes on

which selection for local adaptation may invoke. In the other words, they become coa-

dapted and increase the fitness of populations which suggest that the specific association of

coadaptive gene complexes are the most probable mechanism that maintains the devel-

opmental homeostasis in populations. In the case of hybridisation between two different

populations (interpopulation crosses), coadaptive gene complexes were disrupted because

the new recombination due to divergence in selection for local adaptation between the two

populations So, the new gene complexes will be expected to decrease overall fitness and

individuals and populations will experience outbreeding depression (Dobzhansky 1936;

Orr 1995; Burke et al. 1998; Andersen et al. 2002, 2008; Edmands 2007). The impact of

outbreeding depression on populations differs and depends on the amount of genes

involved in the coadapted gene complexes. If only few genes are involved a coadapted

gene complex, outbreeding depression may be temporary, because selection may effi-

ciently re-establish fitness. But, if the most genes within genome are involved in coad-

aptation the chances of re-establish fitness are very low.

We expected the breakdown of coadapted gene complexes to be equal for males and

females in interpopulation crosses. However, our results show that effects of the break-

down differ between sexes which are in accordance to previous studies (see Andersen et al.

2002 and references therein). The results of the FA of wing length and width in inter-

population hybrids, suggest a different response in males and a general increase of FA in

F1, followed by a reduction of FA in the F2 generation. We found a different response in

females, which was a constant increase of FA through the generations. The results show a

greater trend of FA change in males than in females both for wing length and width.

The observed difference in FA between direct and reciprocal crosses suggests the

presence of the maternal effect. A large part of FA variation is non-additive and is

influenced by cytoplasimic maternal effect, reflecting unpredictable interactions between

genetic backgrounds that have diverged to stochastic processes. Its importance is not

Evol Ecol

123

questionable in literature, and there are experiments indicating that a maternal effect can

account for the variation of the individual size in offspring, survival and behavioural

differences and may have a huge impact on the fitness (Cheverud and Moore 1994;

McAdam et al. 2002; Andersen et al. 2005). Maternal effects arise due to a lot of different

causes, both intrinsic and extrinsic. There are investigations showing that extrinsic changes

in parental environment, such as temperature (Gilchrist and Huey 2001), pollution,

nutrition (Gliwich and Guisande 1992; Rossiter 1996) and oviposition site (Mousseau and

Fox 1998) can influence the offspring. Similarly, various more direct intrinsic effects as

maternal size (Emlet and Hoegh-Guldberg 1997; Hunt and Simmons 2000), parental care

(Rauter and Moore 2002) and maternal age (Kern et al. 2001) influence offspring across

one or more generations. There is also growing evidence that some maternal effects are not

simply the accidental transmission of environmental information from one generation to

the next. Rather, type and function of maternal effects appear often to have been shaped by

natural selection (Hunter 2002).

The results of our study suggest that D. subobscura populations are genetically diverse

even at small geographic scale with frequently strong and unpredictable consequences of

outbreeding. That pattern is likely to become increasingly pronounced as a result of

ongoing habitat fragmentation and destruction by the increased anthropogenic activity. In

such a situation, it seems that one preventive action to increase survival may be to increase

the gene flow by translocation of individuals between populations, and in that way hinder

inbreeding and loss of genetic variation. Our results indicate that care should be taken with

such translocations because that kind of activity could have strong consequences on the

population fitness, and could potentially increase the risk of outbreeding depression. Cli-

mate change is therefore also altering the immigration-emigration dynamics which in turn

will have profound consequences for the population genetic structure as increased gene-

flow typically increase the genetic variability within of populations simultaneously

reducing their local adaptation (Pertoldi and Topping 2004). In a population, the actual

degree of adaptation is the residual effect of the dynamic interaction between the selective

pressure and gene flow. Hence, the high levels of gene flow can reduce or impede the

capacity of adaptation to a stressor (Roush and McKenzie 1987) or may introduce essential

new genes for future adaptation or increase in tolerance (Slatkin 1995). The importance of

gene flow as a force for the maintenance of genetic diversity and avoidance of inbreeding

depression is therefore quite evident. However, as the results of this study are showing high

levels of gene flow also have the potential to introduce poorly adapted genes (outbreeding

depression) that can reduce viability of the population, even if the consequences of out-

breeding have been shown to not being predictable. Furthermore, different outcomes could

be expected depending on the origin of the populations, microhabitat adaptation and

evolutionary history of each population. Clearly, we know too little about long-term

consequences of outbreeding. Overall, among-population heterogeneity in genetic archi-

tecture makes it difficult to assess short/long time results of different population crossings.

However, our results confirm that the studies dealing with DI seem to provide valuable

insights into the importance of the genetic factors in the long term survival of isolated

populations.

Acknowledgments We are grateful to anonymous reviewers for helpful and valuable suggestions on themanuscript. This work was supported by the Ministry of Science and Technology of the Republic of Serbia,Grant No. 143014.

Evol Ecol

123

References

Alibert P, Auffray JC (2003) Genomic coadaptation, outbreeding depression, and developmental stability.In: Polak M (ed) Developmental instability: causes and consequences. Oxford University Press,Oxford, pp 116–134

Andersen DH, Pertoldi C, Scali V, Loeschcke V (2002) Intraspecific hybridization, developmental stabilityand fitness in Drosophila mercatorum. Evol Ecol Res 4:603–621

Andersen DH, Pertoldi C, Scali V, Loeschcke V (2005) Heat stress and age induced maternal effects onwing size and shape in partenogenetic Drosophila mercatorum. J Evol Biol 18:884–892

Andersen DH, Pertoldi C, Loeschcke V, Cavicchi S, Scali V (2008) The impact of genetic parental distanceon developmental stability and fitness in Drosophila buzzatii. Genetica 134:223–233

Anderson WW (1968) Further evidence for coadaptation in crosses between geographic populations ofDrosophila pseudoobscura. Genet Res Camb 12:317–330

Andjelkovic M, Savkovic V, Kalajdzic P (2003) Inversion polymorphism in Drosophila subobscura fromtwo different habitats from the mountain of Goc. Hereditas 138:241–243

Banerjee R, Singh BN (1998) Evidence for coadaptation in geographic populations of Drosophilabipectinata. J Zool Syst Evol Res 36:1–6

Barton NH, Hewitt GM (1985) Analysis of hybrid zones. Ann Rev Ecol Syst 16:113–148Burke JM, Voss TJ, Arnold ML (1998) Genetic interactions and natural selection in Louisiana iris hybrids.

Evolution 52:1304–1310Charlesworth B, Hughes KA (1999) The maintenance of genetic variation in life-history traits. In: Singh RS,

Krimbas CB (eds) Evolutionary genetics: from molecules to morphology, vol 1. Cambridge UniversityPress, Cambridge, pp 369–392

Cheverud JM, Moore AJ (1994) Quantitative genetics and the role of environment provided by relatives inthe evolution of behaviour. In: Quantitative genetic studies of behavior evolution. University ofChicago Press, Chicago, pp 67–100

Clarke GM, Oldroyd BP, Hunt P (1992) The genetic basis of developmental stability in Apis mellifera:heterozygosity versus genetic balance. Evolution 46:753–762

David JR (1979) Utilization of morphological traits for the analysis of genetic variation of wild populations.Aquilo Ser Zool 20:49–61

Dobzhansky T (1936) Studies on hybrid sterility. II. Localization of sterility factors in Drosophila pseud-oobscura hybrids. Genetics 21:113–135

Dobzhansky T (1948) Chromosomal variation in populations of Drosophila pseudoobscura which inhabitnorthern Mexico. Amer Nat 82:97–106

Dobzhansky T (1950) Genetics of natural populations. XIX. Origin of heterosis through natural selection inpopulations of Drosophila pseudoobscura. Genetics 35:288–302

Edmands S (1999) Heterosis and outbreeding depression in interpopulation crosses spanning a wide range ofdivergence. Evolution 53:1757–1768

Edmands S (2007) Between a rock and a hard place: evaluating the relative risks of inbreeding andoutbreeding for conservation and management. Mol Ecol 16:463–475

Edmands S, Timmerman CC (2003) Modelling factors affecting the severity of outbreeding depression.Conserv Biol 17:883–892

Emlet RB, Hoegh-Guldberg O (1997) Effects of egg size on postlarval performance: experimental evidenceform a sea urchin. Evolution 51:141–152

Escobar JS, Nicot A, David P (2008) The different sources of variation in inbreeding depression, heterosis,and outbreeding depression in a metapopulation of Phisa acuta. Genetics 180:1593–1608

Felley J (1980) Analysis of morphology and asymmetry in Bluegill Sunfish Lepomis macrochirus in theSoutheastern United States. Copeia 1:18–29

Fenster CB, Galoway LG, Chao L (1997) Epistasis and its consequences for the evolution of naturalpopulations. Trends Eco Evol 12:282–286

Frankham R (2005) Genetics and extinction. Biol Conserv 126:131–140Garnier S, Gidaszewski N, Charlot M, Rasplus JY, Alibert P (2006) Hibridization, developmental stability,

and functionality of morphological traits in the ground beetle Carabus soliery (Coleoptera, Carabidae).Biol J Linn Soc 89:151–158

Gilchrist GW, Huey RB (2001) Parental and developmental temperature effects on the thermal dependenceof fitness in Drosophila melanogaster. Evolution 55:209–214

Gliwich ZM, Guisande C (1992) Family planning in Daphnia: resistance to starvation in offspring born tomothers grown at different food levels. Oecologia 91:463–467

Goldberg TL, Grant EC, Inendino KR, Kassler TW, Claussen JE, David PP (2004) Increased infectiousdisease susceptibility resulting from outbreeding depression. Conserv Biol 19:455–462

Evol Ecol

123

Graham JH (1992) Genomic coadaptation and developmental stability in hybrid zones. Acta Zool Fenn191:121–131

Graham JH, Felley J (1985) Genomic coadaptation and developmental stability within introgresed popu-lations of Enneachantus gloriosus and E. obesus (Pisces, Centrarchidae). Evolution 39:283–289

Hammer Ø, Harper D, Ryan PD (2001) PAST: Paleontological statistics software package for education anddata analysis. Paleontol Electron 4:9

Hoffmann AA, Sgro CM, Weeks A (2004) Chromosomal inversion polymorphism and adaptation. TrendsEcol Evol 9:482–488

Hunt J, Simmons LW (2000) Maternal and paternal effects on offspring phenotype in the dung beetleOnthophagus taurus. Evolution 54:936–941

Hunter MD (2002) Maternal effects and the population dynamics of insects on plants. Agric For Entomol4:1–9

Jelic M, Kenig B, Kurbalija Z, Stamenkovic-Radak M, Andjelkovic M (2009) Intra-species differentiationamong Drosophila subobscura from different habitats in Serbia. Arch Biol Sci 61:513–521

Kern S, Ackermann M, Stearns SC, Kawecki TJ (2001) Decline in offspring viability as a manifestation ofaging in Drosophila melanogaster. Evolution 55:1822–1831

Krimbas CB (1993) Drosophila subobscura: biology, genetics, and inversion polymorphism. VerlagDr. Kovac, Hamburg

Krimbas CB, Loukas M (1980) The inversion polymorphism of Drosophila subobscura. Evol Biol 12:163–234

Leary RF, Allendorf FW (1989) Fluctuating asymmetry as an indicator of stress: implication for conser-vation biology. Trends Ecol Evol 4:214–217

Lens L, Van Dongen S, Galbusera P (2000) Developmental instability and inbreeding natural bird popu-lations exposed to different levels of habitat disturbance. J Evol Biol 13:889–896

Lerner IM (1954) Genetic homeostasis. Wiley, New YorkLivshits G, Kobyliansky E (1985) Lerners concept of developmental homeostasis and the problem of

heterozygosity level in natural populations. Heredity 55:341–353Livshits G, Smouse PE (1993) Relationship between fluctuating asymmetry, morofological modality and

heterozygosity in elderly Israely populations. Genetica 89:155–166Markow TA (1995) Evolutionary ecology and developmental instability. Ann Rev Entomol 40:105–120Markow TA, Ricker P (1991) Developmental stability in hybrids between the sibiling species pair,

Drosophila melanogaster and Drosophila simulans. Genetica 84:115–121McAdam AG, Boutin S, Reale D, Berteaux D (2002) Maternal effects and the potential for evolution in

natural populations of animals. Evolution 56:846–851Møller AP, Swaddle JP (1997) Asymmetry, developmental stability, and evolution. Oxford University Press,

OxfordMousseau TA, Fox CW (1998) Maternal effects as adaptations. Oxford University Press, New YorkOrengo DJ, Prevosti A (1996) Temporal changes in chromosomal polymorphism of Drosophila subobscura

related to climatic changes. Evolution 50:1346–1350Orr HA (1995) The population-genetics of speciation—the evolution of hybrid incompatibilities. Genetics

139:1805–1813Orr HA, Turelli M (2001) The evolution of postzygotic isolation: accumulating Dobzhansky–Muller

incompatibilities. Evolution 55:1085–1094Palmer AR (1994) Fluctuating asymmetry analyses: a primer. In: Developmental instability: its origins and

evolutionary implications, pp 335–364Palmer AR (1996) Waltzing with asymmetry: is fluctuating asymmetry a powerful new tool for biologists or

just an alluring new dance step? Bio Sci 46:518–532Palmer AR, Strobeck C (1986) Fluctuating asymmetry: measurement, analysis, patterns. Ann Rev Ecol Syst

17:391–421Palmer AR, Strobeck C (1992) Fluctuating asymmetry as a measure of developmental stability: implications

of non-normal distributions and power of statistical tests. Acta Zool Fenn 191:55–70Pelabon C, Carlson ML, Hansen TF, Armbruster WS (2005) Effect of crossing distance of offspring fitness

and developmental stability in Dalechampia scadens (Euphorbiaceae). Am J Bot 92:842–851Pertoldi C, Topping C (2004) Impact assessment predicted by means of genetic agent-based modelling. Crit

Rev Toxicol 34:487–498Pertoldi C, Kristensen TN, Andersen DN, Loeschcke V (2006a) Developmental instability as an estimator of

genetic stress. Heredity 96:122–127Pertoldi C, Sørensen JG, David JR, Loeschcke V (2006b) Lerner’s theory on the genetic relationship

between heterozigosity, genomic co-adaptations, and developmental instability revised. Evol Ecol Res8:1487–1498

Evol Ecol

123

Rauter CM, Moore AJ (2002) Evolutionary importance of parental care performance, food resources, anddirect and indirect genetic effects in a burying beetle. J Evol Biol 15:407–417

Rice WR (1989) Analysing tables of statistical tests. Evolution 43:223–225Ross KG, Robertson JL (1990) Developmental stability, heterozigosity, and fitness in two introduced fire

ants (Solenopsis invicta and S. richteri) and their hybrids. Heredity 64:93–103Rossiter MC (1996) Incidence and consequence of inherited environmental effects. Ann Rev Ecol Syst

27:451–476Roush RT, McKenzie JA (1987) Ecological genetics of insecticide and arcaricide resistance. Ann Rev

Entomol 32:361–380Sagvik J, Uller T, Olsson M (2005) Outbreeding depression in the common frog, Rana temporaria. Conserv

Genet 6:205–211Sheridan L, Pomiankowski A (1997) Fluctuating asymmetry, spot asymmetry and inbreeding depression in

the sexual coloration of male guppy fish. Heredity 79:515–523Slatkin M (1995) A measure of population subdivision based on microsatellite allele frequencies. Genetics

139:457–462Stamenkovic-Radak M, Rasic G, Savic T, Kalajdzic P, Kurbalija Z, Kenig B, Andjelkovic M (2008)

Monitoring of the genetic structure of natural populations: change of the effective population size andinversion polymorphism in Drosophila subobscura. Genetica 133:57–63

Vøllestad LA, Hindar K, Møller AP (1999) A meta-analysis of fluctuating asymmetry in relation to het-erozigosity. Heredity 83:138–144

Waldmann P (1999) The effect of inbreeding and population hybridization on developmental instability inpetals and leaves of the rare plant Silene diclinis (Caryophyllaceae). Heredity 83:138–144

Whitlock MC, Phillips PC, Moore FBG, Tonsor S (1995) Multiple fitness peaks and epistasis. Ann Rev EcolSyst 26:601–629

Wilcox DC, Brent SD, McDavid WD, Greer DB (2002). ImageTool 3.0. Department of Dental DiagnosticScience at the University of Texas Health Science Center, San Antonio, (http://ddsdx.uthscsa.edu/dig/download.html)

Zakharov VM (1981) Fluctuating asymmetry as an index of developmental homeostasis. Genetica 13:241–256

Evol Ecol

123