Download intraspecific variation in sexual isolation in the

Evolution, 54(2), 2000, pp. 567–573
INTRASPECIFIC VARIATION IN SEXUAL ISOLATION IN THE JEWEL WASP NASONIA
SETH R. BORDENSTEIN,1,2 MARK D. DRAPEAU,1,3
1 Department
AND
JOHN H. WERREN1
of Biology, The University of Rochester, Rochester, New York 14627
2 E-mail: [email protected]
Abstract. Divergence in mate recognition systems can lead to reproductive isolation. In this study, we investigate
patterns of intraspecific variation that contribute to premating isolation within and between two haplodiploid species,
Nasonia vitripennis and N. longicornis. In a broad-scale survey of 17 North American isofemale lines encompassing
the two species, we report strong asymmetric sexual isolation between species and a dramatic level of intraspecific
variation for mate discrimination between species. A general lack of incipient speciation was found, with the exception
of low levels of interpopulational sexual isolation within N. vitripennis. Regression analysis shows that the degree of
intraspecific variation for within-species mating frequency is not associated with the degree for between-species mating
frequency. Reinforcement or reproductive character displacement may be involved in some of the variation in interspecies premating isolation.
Key words.
Behavior, premating isolation, reinforcement, sexual selection, speciation.
Received January 20, 1999.
Intraspecific variation in sexual isolation may signify the
early stages of speciation. Such variation can take two forms:
(1) intraspecific variation in reproductive isolation between
populations within a species (incipient speciation); (2) intraspecific variation in reproductive isolation to closely related
species (variable interspecific isolation). Studies of sexual
isolation have characterized both patterns of intraspecific variation (Patterson and Stone 1952; Littlejohn and Loftus-Hills
1968; Zouros and D’Entremont 1980; Krebs 1990; Markow
1991; Wu et al. 1995; Miller et al. 1998). The latter pattern
is often studied in the context of Dobzhansky’s reinforcement
hypothesis, which proposes that natural selection will increase levels of prezygotic reproductive isolation between
populations that produce unfit hybrids (Dobzhansky 1940).
A by-product of reinforcement is that sympatric populations
will show stronger levels of mate discrimination than allopatric populations of the same taxa (Noor 1995; Saetre et al.
1997; Rundle and Schluter 1998). Alternatively, sexual selection can generate locally adapted mating signals within
populations. Divergence between populations could then lead
to intraspecific variation for within and/or between species
sexual isolation.
Sexual isolation (e.g., divergence in mate recognition systems) can arise due to sexual selection (West-Eberhard 1983;
Kirkpatrick 1987; Holland and Rice 1998), genetic drift
(Mayr 1963; Lande 1981), and natural selection (Fisher 1930;
Dobzhansky 1937). In this study, we conduct a broad-scale
survey of sexual isolation within and between two sibling
species of the genus Nasonia. Our objectives are to determine
the levels of variation in intra- and interspecific sexual isolation. The variation uncovered can then be used in genetic
and population genetic experiments to investigate the evolution of premating isolation.
Nasonia is a complex of three closely related sibling species of haplodiploid parasitic wasps. Nasonia vitripennis
(NV) exists worldwide and is a parasite of cycolorrhaphous
flies that are found at carcasses and bird nests. Nasonia gi1 Present address: Department of Ecology and Evolutionary Biology, 321 Steinhaus Hall, University of California, Irvine, California 92697.
Accepted August 18, 1999.
raulti (NG) and the third species N. longicornis (NL) are
commonly found in bird nests across eastern and western
North America, respectively (Darling and Werren 1990). NL
and NG are allopatric to each other, and they occur microsympatrically with NV over much of their ranges.
Previous experiments have uncovered a number of genetic,
microbial, and behavioral isolating factors between NV and
NG (Breeuwer and Werren 1995; Bordenstein and Werren
1998; Drapeau and Werren 1999). The study here characterizes variation in reproductive isolation within and between
geographic populations of NV and its sister species NL. The
two species are estimated to have diverged from each other
approximately 500,000 years ago (Campbell et al. 1993).
We may expect variation in sexual isolation in NV and
NL. First, NV occurs microsympatrically with NL and NG
over parts of their geographic ranges (Darling and Werren
1990), and postzygotic incompatibilities arise due to the cytoplasmic bacterium Wolbachia (Breeuwer and Werren 1990;
Bordenstein and Werren 1998) and F2 hybrid breakdown
(Breeuwer and Werren 1995). Therefore, reinforcement of
premating isolation could generate varying degrees of interspecific mate discrimination among different geographic populations of the species, due to varying potential for interspecific mating. Second, geographic differentiation in mate
recognition (via selection or drift) could be occurring between geographic populations, given that gene flow between
populations may be restricted by geographic habitat barriers,
particularly in western North America. For these reasons, it
is worthwhile to determine whether early stages of divergence
in mate recognition are arising within these species.
MATERIALS
METHODS
Strains
All strains are derived from single inseminated females
(isofemale lines) collected from natural populations between
1989 and 1991. Each field strain was maintained in larval
diapause (a prolonged larval state), except for approximately
one to three generations every 1.0–1.5 years, when they were
briefly maintained as breeding strains and reinduced into diapause. Therefore, strains experienced relatively few gener-
567
q 2000 The Society for the Study of Evolution. All rights reserved.
AND
568
SETH R. BORDENSTEIN ET AL.
ations of laboratory rearing prior to experiments. The strains
were removed from diapause and antibiotically cured of their
Wolbachia infection with rifampicin treatment for three sequential generations. Curing of the Wolbachia infection was
confirmed using a polymerase chain reaction (PCR) assay
with primers previously described (Perrot-Minnot et al.
1996). Strains were then maintained in the laboratory for
approximately 12 additional generations prior to the experiment. To test for mate discrimination, strains were divided
into two groups: group I consisted of five NV strains and
four NL strains; group II consisted of five NV and five NL
strains. Two strains (NV Utah and NL Utah 1) were used in
both groups.
The following strains were used (an abbreviated name for
each strain is indicated in parentheses): group I: N. vitripennis: RNVMN 204K (NV Minnesota 1), RNVIN 217K (NV
Indiana), RNVXIDB 431E (NV Idaho), RNVPA 2330 (NV
Pennsylvania), RNVXUTC 406X (NV Utah); N. longicornis:
RNLCA 9304 (NL California), RNLIDB 418E (NL Idaho),
RNLNV 202H (NL Nevada 1), RNLUTN 308V (NL Utah
1); group II: N. vitripennis: RNVMN 206Bi (NV Minnesota
2), RNVXNVB 403D (NV Nevada), RNVWYC 400G (NV
Wyoming), RNVOH 204 (NV Ohio), RNVXUTC 406X (NV
Utah); N. longicornis: RNLUTB 303–7L (NL Utah 2),
NLUTB 305–3D (NL Utah 3), NLUTN 310 (NL Utah 4),
RNLUTN 308V (NL Utah 1), RNLNV 206H (NL Nevada
2).
Strains were chosen from a broad geographic distribution,
ranging from Pennsylvania to California for NV and NL. NV
is cosmopolitan, being found throughout North American and
the world, whereas NL is localized in western North America
(Darling and Werren 1990).
Mating Tests
To assay variation in premating isolation, we tested male
and female mating success in no-choice mating tests. Mating
trials were designed as follows. Males and females of each
strain were collected as virgin pupae approximately one day
before eclosion. Pupae were sorted by gender into glass vials
and kept at 258C until eclosion. Mating frequency was assayed according to the methods below, using one- to twoday-old adults.
In each group, all possible crosses were performed in a
diallel or grid format. Per replicate, two males and five females were placed in 12 3 75-mm glass vials at room temperature (258C). After two hours, males were discarded and
females were placed in individual vials. They were each immediately given one Sarcophaga bullata fleshfly host. Female
wasps were allowed to parasitize these until their death. Progeny were scored for presence or absence of females. Due to
haplodiploidy, production of diploid females indicates that
mating and fertilization occurred in Nasonia. Unmated females produce all-male families. Thus, the assay is designed
to detect differences in mating behavior, rather than mechanical isolation or sperm-egg incompatibilities.
Validity of the Mating Assay
Tests were performed to determine the validity of using
female (daughter) production as an assay for mating. Virgin
FIG. 1. Asymmetrical premating isolation pattern between Nasonia
vitripennis (NV) and N. longicornis (NL). Group I data correspond
to gray filled bars and Group II data correspond to black filled bars.
Sample sizes are denoted above the bars.
males and females were singly paired in a glass vial and
observed for 30 min. Of those replicates in which copulations
were observed, females were subsequently hosted and the
presence/absence of daughters was scored for each family.
A total of 142 intraspecific matings were observed (involving
two NV and two NL strains). Of these, all produced female
offspring. Of 87 interspecific matings observed (involving
the same strains), 86 produced female progeny, whereas of
32 cases where no mating occurred, none produced female
progeny (i.e., all males were produced). These findings establish the validity of using daughter production as an assay
for mating.
Data Analysis
Percent mating was scored based upon the production of
female progeny in a cross. In haplodiploids, mated females
produce female (diploid) and male (haploid) progeny, whereas unmated females produce only male progeny. Mating was
therefore scored (indirectly) by analyzing the percentage of
parental females that produced daughters among their progeny. All statistics were performed with MINITAB 11. Analyses of variance were performed on mating percentages that
were arcsine transformed according to Sokal and Rohlf
(1981). All pairwise comparisons were conducted with proportion tests. Regression analyses were conducted and associations were plotted linearly.
RESULTS
Mating in Intra- and Interspecific Crosses
For the first group of strains (group I), all possible pairwise
crosses were performed between five isofemale NV lines and
four isofemale NL lines spanning nine different North American populations. Summary data of overall specieswide mating frequencies are shown in Figure 1. As expected, intraspecific matings occur readily in comparison to interspecific
matings (93.4% vs. 39.1%). In interspecific crosses, sexual
569
INTRASPECIFIC VARIATION IN SEXUAL ISOLATION
TABLE 1. Nested ANOVA on interspecific mating frequency with
female and male strain nested within the cross direction (Nasonia vitripennis [NV] male 3 N. longicornis [NL] female vs. NL male 3 NV
female).
Source
df
SS
MS
Group I
Cross
Female strain
Male strain
Error
Total
1
7
7
24
39
10,298.3
7332.3
996.3
2977.0
21,603.9
10,298.3
1047.5
142.3
124.0
Group II
Cross
Female strain
Male strain
Error
Total
1
8
8
32
49
17,300.5
2000.3
1174.9
3705.8
24,181.6
17,300.5
250.0
146.9
115.8
F
83.0
8.4
1.2
149.4
2.2
1.3
P
0.000
0.000
0.368
0.000
0.059
0.294
isolation is strong, but not complete. There is a highly significant asymmetrical premating isolation pattern: NL male
3 NV female matings occur more readily (59.2%) than NV
male 3 NL female matings (18.6%; Fig. 1, Table 1, F 5
83.0, P , 0.001). Analysis of variance on the premating
asymmetry (and thus interspecies mating frequency) shows
that female strain contributes significantly to the observed
variation in interspecies mating frequency, whereas male
strain does not (Table 1). Female strain accounts for 33.9%
of the variation in interspecific crosses, whereas male strain
only accounts for 4.6% of the variation. These results support
the view that differences in female preference govern the
TABLE 2. Group I male and female mating frequency. NV and NL
denote Nasonia vitripennis and N. longicornis, respectively. A number
following a strain name designates a specific isofemale line from that
state; n denotes the total number of females in a cross.
Cross
Strain
Male mating
frequency (n)
Female mating
frequency (n)
Intraspecific
NV
NV
NV
NV
NV
NV
Minnesota 1
Indiana
Idaho
Pennsylvania
Utah
93.4
95.9
97.3
93.1
97.6
(122)
(122)
(110)
(102)
(125)
96.3
96.7
98.2
93.8
91.5
(136)
(122)
(113)
(128)
(82)
NL
NL
NL
NL
California
Idaho
Nevada 1
Utah 1
86.4
90.6
89.7
92.7
(88)
(85)
(78)
(82)
83.7
95.4
96.7
84.7
(104)
(65)
(92)
(72)
NV
NV
NV
NV
NV
Minnesota 1
Indiana
Idaho
Pennsylvania
Utah
14.6
17.4
18.8
19.4
24.1
(96)
(86)
(85)
(93)
(83)
87.91 (91)
52.3 (111)
41.6 (89)
55.4 (92)
60.3 (68)
NL
Interspecific
NV
NL
NL
NL
NL
NL
California
Idaho
Nevada 1
Utah 1
55.6
53.8
58.2
67.7
(108)
(106)
(110)
(127)
0.9
52.2
16.8
12.0
(116)
(90)
(137)
(100)
TABLE 3. Proportion of variation attributed to female and male strain
obtained from the analysis of variance on mating success. NV, Nasonia
vitripennis; NL, N. longicornis.
Cross
(male 3
female)
Trait
Group I %
of the variance
Group II %
of the variance
NV 3 NV
female strain
male strain
13.6
17.8
30.7
16.9
NV 3 NL
female strain
male strain
85.6**
3.5
42.2*
10.6
NL 3 NV
female strain
male strain
41.9
14.7
27.3
16.9
NL 3 NL
female strain
male strain
58.7*
4.9
3.4
5.4
* P , 0.05; ** P , 0.0001.
asymmetry. Several points discussed below corroborate this
view.
Table 2 shows patterns of male and female mating success
across all possible crosses. To determine levels of intraspecific variation in mating success, an analysis of variance was
conducted on mating successes within each of the four specieswide cross directions (Table 3). Male and female strains
were included as factors. Although there is a moderate range
of variation in male mating success in heterospecific matings,
male strain tends not to have a significant effect in any of
the four cross directions (Table 3). NL female strain does
have a highly significant effect on the observed mating success, whereas NV female strain does not (Table 3). For instance, NL female strain explains 58.7% and 85.6% of the
variation in intra- and interspecific crosses, respectively.
The level of interspecific mating frequency differed significantly between strains. Results were most dramatic in
analyses of interspecific mating frequency of NL females. A
50-fold difference in NL female interspecific mating frequency exists between Idaho females (52.2%) and California
females (0.9%; Table 2). To confirm this extraordinary difference in NL female heterospecific mating frequency, two
additional experiments were conducted. Following the previous design, NL Idaho and NL California females were
crossed to NV Idaho males in both one- and two-hour assays.
The results, which are shown in Table 4, reconfirm the variation initially described in Table 2. In the one-hour assay,
NL Idaho females mated with NV Idaho males in 20.0% of
the replicates, whereas NL California females did not mate
TABLE 4. Reconfirmation of intraspecific variation for interspecific
mate discrimination. Values are the percentage of those females that
mated. n, the total number of females in that cross; NV, Nasonia vitripennis; NL, N. longicornis.
Cross (male 3 female)
Interspecific
NV Idaho 3 NL Idaho
NV Idaho 3 NL California
Intraspecific
NL Idaho 3 NL Idaho
NL California 3 NL California
One-hour
assay (n)
Two-hour
assay (n)
20.0% (15)
0.0% (31)
57.0% (86)
2.6% (193)
100.0% (14)
83.3% (18)
87.9% (23)
90.9% (55)
570
SETH R. BORDENSTEIN ET AL.
TABLE 5. Group II male and female mating frequency. NV and NL
denote Nasonia vitripennis and N. longicornis, respectively. A number
following a strain name designates a specific isofemale line used from
that state. n denotes the total number of females in a cross.
Cross
Strain
Male mating
frequency (n)
Female mating
frequency (n)
Intraspecific
NV
NV
NV
NV
NV
NV
Minnesota 2
Nevada
Wyoming
Ohio
Utah
99.1
94.3
95.8
94.0
95.0
(114)
(105)
(72)
(100)
(80)
99.0
97.3
96.4
92.6
87.1
(98)
(111)
(137)
(94)
(31)
NL
NL
NL
NL
NL
Utah 2
Utah 3
Utah 4
Utah 1
Nevada 2
97.1
94.2
93.1
94.0
95.0
(102)
(104)
(72)
(83)
(60)
94.1
94.7
96.4
92.3
94.3
(102)
(95)
(111)
(26)
(87)
NV
NV
NV
NV
NV
Minnesota 2
Nevada
Wyoming
Ohio
Utah
23.7
22.6
18.9
23.8
33.8
(97)
(102)
(74)
(80)
(74)
90.3
85.3
66.7
65.1
76.0
(103)
(102)
(123)
(83)
(25)
NL
NL
NL
NL
NL
Utah 2
Utah 3
Utah 4
Utah 1
Nevada 2
93.1
67.2
75.3
74.1
88.3
(72)
(125)
(81)
(81)
(77)
16.4
39.3
15.7
33.3
23.5
(104)
(107)
(108)
(27)
(81)
NL
FIG. 2. Intraspecific variation in within species mating frequency.
NV, Nasomia vitripennis; NL, N. longicornis. Self and nonself refer
to within-strain and between-strain within-species crosses, respectively. Group I data correspond to gray filled bars and group II data
correspond to black filled bars. Sample sizes are denoted above the
bars.
with NV Idaho males at all (0.0%; Table 4). These percentages differ from the original two-hour experiment, but show
the same trend in mate acceptance. A replicate experiment
of the two-hour assay reconfirms the large difference in interspecific sexual isolation between these strains. Mating frequencies when NV Idaho males are crossed to NL Idaho and
NL California females are 57.0% and 2.6%, respectively (Table 4).
We also analyzed the data to look for evidence of incipient
sexual isolation by comparing the mating frequencies of within strain (self) and between strain within species (nonself)
crosses. Figure 2 shows that for both species, nonself crosses
yield lower mating percentages than self crosses, but the
finding is only significant for NV (P , 0.01). For instance,
crosses with either NV Minnesota 1 and NV Indiana males
to NV Utah females show significantly reduced mating frequencies, 81.8% (n 5 22) and 87.5% (n 5 16), in comparison
to the self-crosses, 100.00% (n . 16, P , 0.05).
A second set of crosses (group II) yield generally similar
results. Crosses were performed with a different set of strains,
except NV Utah and NL Utah 1 are common to the two
groups. All possible pairwise crosses were set up with five
NV and five NL isofemale lines. Summary data of overall
specieswide mating success are shown in Figure 1. Data show
that intraspecific matings occur readily in comparison to interspecific matings. A highly significant premating isolation
asymmetry was again found in the same direction as that of
group 1: NL male 3 NV female matings occur more readily
than NV male 3 NL female matings (Table 1, F 5 149.39,
P , 0.001). Female strain had a marginally significant effect
on the mating asymmetry (8.3% of the variation), whereas
male strain did not (4.9% of the variation). The majority of
the variation (71.5%) is due to the species (i.e., cross direction). The finding differs from group I data, in which female
strain explained a larger proportion of the variation. This
difference may be attributed to reduced genetic variation
among the group II strains, because four of the five NL lines
are from the same state (Utah).
Interspecific
NV
NL
Table 5 shows patterns of male and female mating success
across all possible group II crosses. To determine levels of
intraspecific variance in mating frequencies, an analysis of
variance was conducted within each of the four species-wide
cross directions. Male and female strain were included as
factors. While there is a moderate range of variation in male
mating success in heterospecific matings, male strain tends
not to have a significant effect in any of the four cross directions (Table 3, Two-way ANOVA, P k 0.05). On average,
the range of male mating success variation in heterospecific
crosses is greater for NL males (67.2% to 93.1%) than for
NV males (18.9% to 33.3%). NL female strain does have a
significant effect on the observed mating frequency variation,
while NV female strain does not (Table 3).
Intraspecific variance in interspecies mating frequency was
again revealed in our analysis. In contrast to group I, the
greatest range of female mating frequency occurs within NV.
For example, the frequency of NV Ohio females that mated
with NL males is 65.1% versus 90.3% for NV Minnesota 2
females (Table 5). Interestingly, NV Minnesota 2 females
mate with NL males almost as readily as they do with NV
males (99.0%). Such female strain variation in interspecific
mating frequency suggests that female preference may be
actively evolving within local populations of NV. The Minnesota population is different from others in that NV was not
found microsympatrically with either NL or NG, in contrast
to all other populations tested (unpubl. data).
To detect patterns of incipient sexual isolation, mating fre-
571
INTRASPECIFIC VARIATION IN SEXUAL ISOLATION
FIG. 3. Linear plots of interspecific female mating frequency regressed against intraspecific female mating frequency. Each circle/
diamond denotes a geographic strain. Solid black circles denote data for Nasonia vitripennis and unfilled diamonds for N. longicornis.
(A) Intraspecific mating frequency is calculated as the nonself mating frequency. (B) One outlier was excluded from the analysis and
relative intraspecific mating frequency is calculated as the ratio of nonself mating frequency to self mating frequency.
quencies of self and nonself crosses in both species were
compared. No significant differences were detected (NV:
94.4% vs. 96.1%, NL: 94.9% vs. 94.2%). Thus, there was an
overall lack of intraspecies variation in within-species mating
success in the group II data.
the data here, we conclude that the variation in interspecific
mating success is not due to divergence in intraspecific mating success between populations. Therefore, there is no evidence that the inter- and intraspecific behavioral patterns
share a common cause.
Variance in Interspecific Mating Frequency:
The Role of Sexual Selection
DISCUSSION
A regression analysis was conducted to make statistical
inferences on the role of sexual selection in the observed
intraspecific variance in interspecific mating frequency. The
reasoning is as follows. If the variation in interspecific premating isolation is due to incipient divergence between populations (e.g., sexual selection), then the level of premating
isolation within species may be associated with the level of
premating isolation between species. Assuming the targets
of sexual selection (e.g., pheromones, courtship displays) are
the same in both closely related species, subtle differences
in mate discrimination between populations within a species
could be expressed as much larger effects on interspecific
isolation. One prediction follows: the strains showing a higher degree of intraspecific discrimination are most likely to
show higher levels of interspecific discrimination. A regression analysis on female mating success was conducted for
both NV and NL. In Figure 3A, interspecific mating frequency is regressed against intraspecific (nonself) mating frequency. No association is found for NV (R2 5 0.004, P 5
0.858), and a positive but nonsignificant association is found
for NL (R2 5 0.329, P 5 0.106). Similarly, in Figure 3B,
interspecific mating frequency is regressed against relative
intraspecific mating frequency (calculated as the ratio of nonself mating frequency to self mating frequency), a measure
that eliminates the confounding effects of mating propensity.
No significant association was detected (NV: R2 5 0.133, P
5 0.334 and NL: R2 5 0.028, P 5 0.692). Furthermore, NV
Utah, an NV strain with the highest level of self versus nonself intraspecific mating, had typical levels of female mating
frequency with NL males (60.3%). Similarly, the NL California strain, which rarely mated with NV males (0.9%)
showed no evidence of intraspecific discrimination. Based on
The biological species concept defines species as populations that do not typically interbreed and are thus reproductively isolated (Mayr 1940). The widespread acceptance
of Mayr’s concept had a major impact on speciation studies.
The study of the origin of species became the study of reproductive isolating factors (Coyne 1992). For practical purposes, most studies have investigated strongly expressing isolating mechanisms that are fixed between divergent groups.
However, such systems are most likely beyond the incipient
stages of speciation.
Studies of intraspecific variation in reproductive isolation
have been conducted in systems such as fruit flies (Patterson
and Stone 1952; Krebs 1990; Noor 1995), mosquitoes (Laven
1959), flour beetles (Wade and Johnson 1994; Wade et al.
1995, 1997), wolf spiders (Miller et al. 1998), and treefrogs
(Littlejohn and Loftus-Hills 1968). The present study concerns intraspecific variation in reproductive isolation in the
haplodiploid jewel wasp Nasonia.
We found significant variation in sexual isolation both
within and between species. However, the variation detected
in between-species mating trials was much greater than the
variation detected in mating trials between geographic populations of the same species. In fact, intraspecific trials with
only one group of strains showed significantly higher mating
frequencies in within-strain to between-strain crosses, and
only in one species (NV). Thus, we found relatively little
evidence for interpopulational premating isolation within the
two species.
In contrast, a large range of within-species variation in
interspecific sexual isolation was uncovered. Both species
showed dramatic differences across geographic populations
in mating frequency from heterospecific crosses. The greatest
range of variation was seen between female strains, rather
572
SETH R. BORDENSTEIN ET AL.
than male strains (Tables 3, 4). In part, this suggests that
female preference may be actively evolving in localized populations. There are a few a priori reasons why this is to be
expected in Nasonia. First, females determine whether copulation occurs in Nasonia; forced copulations are not possible
in these species (Assem and Werren 1994). For any mating
to occur, the female must be stimulated to a state of receptivity by the male. Second, females rarely mate twice, whereas males will mate repeatedly; thus, there may be stronger
selection on females to be ‘‘choosy’’ about their mates. Third,
selection against maladaptive hybridization could drive the
evolution of mate discrimination (e.g., reinforcement) in regions of sympatry.
The view that interspecies mating can lead to reinforcement
of premating isolation is still controversial, although the possibility has gained some recent theoretical (Liou and Price
1994; Kelly and Noor 1996; Kirkpatrick and Servedio 1999)
and empirical (Noor 1995; Saetre et al. 1997; Rundle and
Schluter 1998) support. There are several reasons for why
reinforcement might be occurring in natural populations of
Nasonia. First, there is potential for hybridization between
the species in nature. NL and NV can co-occur microsympatrically, often in the same bird nests where hosts occur
(Darling and Werren 1990; J. Werren, unpubl. data). Thus,
in areas of sympatry and in the absence of mate discrimination, hybridization would be a frequent occurrence. Second, interspecific hybridization is maladaptive in Nasonia.
Wolbachia-induced cytoplasmic incompatibility (CI) acts as
a primary postzygotic isolating barrier between all three Nasonia species; CI can completely prevent hybrid production
between the species (Breeuwer and Werren 1990; Bordenstein and Werren 1998; S. R. Bordenstein, J. A. J. Breewer,
and J.H. Werren, unpubl. data). Even in the absence of Wolbachia-induced cytoplasmic incompatibility, other postmating incompatibilities exist (Breeuwer and Werren 1995) that
could drive the evolution of premating isolation.
Although reinforcement was not the main subject of these
experiments, some of the data are consistent with the theory.
First, the interspecies premating asymmetry is in the expected
direction (Watanabe and Kawanishi 1979): NL is more discriminatory than NV. There is likely to be an asymmetry in
exposure to hybridization because NL is embedded within
the geographic distribution of NV (Darling and Werren
1990). Additionally, NV is a generalist parasitoid found on
a wider range of hosts and apparently at much higher population densities than NL. NL specializes on blowfly pupae
inhabiting bird nests (Darling and Werren 1990; J. Werren,
unpubl. data). Thus, NL likely experiences relatively higher
incidences of microsympatry (i.e., occurring in same bird
nests with NV) than NV.
A second line of evidence for reinforcement comes from
patterns observed in particular crosses. NV female mating
frequency in group I interspecific crosses varies widely across
geographic regions (Table 2). In particular, NV Minnesota
female mating frequency is high in matings to NL males
(;88%), whereas the other NV strains show a significantly
stronger degree of sexual isolation (;40–60%). The finding
is notable, because the NV Minnesota strain is the most ‘‘allopatric-like’’ strain in our analysis. The other NV strains
either occur sympatrically with NL in the western United
States or with NG in the northeastern United States (Darling
and Werren 1990). The same trend is observed in the group
II dataset with a different NV Minnesota strain, although the
differences are not to the same degree (Table 5). Future work
will focus on a more systematic investigation of the possible
role of reinforcement in the evolution of premating isolation.
Indirect evidence from NV and the eastern North American
species N. giraulti (NG) is also consistent with reinforcement
(or reproductive character displacement). Most parasitic
wasps mate outside the host following emergence (but see
Suzuki and Hiehata 1985). However, NG females have a high
propensity to mate within the host prior to emergence, and
thus often mate with siblings (Drapeau and Werren 1999).
NG occurs microsympatrically with NV in bird nests over
much of its range, and it has been hypothesized that withinhost mating has evolved as a hybridization escape mechanism
from the more common NV species. Nevertheless, the suggestion of reinforcement of premating isolation in Nasonia
remains largely anecdotal at this time.
An alternative model for the evolution of a mating asymmetry has been posited by Kaneshiro (1983). According to
this model, females from the ancestral lineage discriminate
against males from the derived lineage, due to a loss of male
mating components associated with a founder event in the
derived lineage. Our data are inconsistent with this model,
because the species showing more discrimination (NL) is the
derived lineage (Campbell et al. 1993).
Sexual selection operating within populations could generate variation in interspecific mating, because subtle differences in interpopulational mate discrimination governed by
sexual selection may be amplified in heterospecific crosses.
According to this view, the variation in heterospecific mating
is merely a by-product of sexual selection within species,
rather than a character directly under natural selection (e.g.,
reinforcement). One prediction of the model is that the strains
showing a higher degree of intraspecific discrimination are
most likely to show high levels of interspecific discrimination. This was not the case: We failed to detect an association
between intra- and interspecific female mating frequency for
either NV or NL (Fig. 3). The finding suggests that divergence
in interspecific mating is not due to divergence in intraspecific
mating (e.g., via sexual selection). However, sample sizes
are insufficient to confidently rule out this effect.
The availability of visible and molecular markers in the
Nasonia complex (Gadau et al. 1999) will provide tools for
a more detailed genetic analysis of premating isolation in
Nasonia. Future work will focus on the genetic and ecological
basis of large differences in mate discrimination in Nasonia.
ACKNOWLEDGMENTS
We thank C. Kennedy, P. Mukhopadhyay, and S. Patel for
technical assistance and S. Perlman for setting up pilot studies. We thank N. Johnson, C. Jones, T. Markow, D. Presgraves, and one anonymous reviewer for their critical reading
of the manuscript and J. Coyne, N. Johnson, H. A. Orr, and
M. Noor for pointing us to relevant literature. This research
was supported by a National Science Foundation grant to
JHW.
INTRASPECIFIC VARIATION IN SEXUAL ISOLATION
LITERATURE CITED
Assem, J. van den, and J. H. Werren. 1994. A comparison of the
courtship and mating behavior of three species of Nasonia. J.
Insect Behav. 7:53–66.
Bordenstein, S. R., and J. H. Werren. 1998. Effects of A and B
Wolbachia and host genotype on interspecies cytoplasmic incompatibility. Genetics 148:1833–1844.
Breeuwer, J. A. J., and J. H. Werren. 1990. Microorganisms associated with chromosome destruction and reproductive isolation between two insect species. Nature 346:558–560.
———. 1995. Hybrid breakdown between two haplodiploid species:
the role of nuclear and cytoplasmic genes. Evolution 49:
705–717.
Campbell, B. C., J. D. Steffen-Campbell, and J. H. Werren. 1993.
Phylogeny of theNasonia species complex (Hymenoptera: Pteromalidae) inferred from an rDNA internal transcribed spacer
(ITS2). Insect Molec. Biol. 2:255–237.
Coyne, J. A. 1992. Genetics and speciation. Nature 355:511–515.
Darling, C. D., and J. H. Werren. 1990. Biosystematics of two new
species ofNasonia (Hymenoptera: Pteromalidae) reared from
birds’ nests in North America. Ann. Entomol. Soc. Am. 83:
352–370.
Dobzhansky, T. 1937. Genetics and the origin of species. Columbia
Univ. Press, New York.
———. 1940. Speciation as a stage in evolutionary divergence. Am.
Nat. 74:312–321.
Drapeau, M. D., and J. H. Werren. 1999. Differences in mating
behavior and sex ratio between three sibling species of Nasonia.
Evol. Ecol. Res. 1:223–239.
Fisher, R. A. 1930. The genetical theory of natural selection. Clarendon Press, Oxford.
Gadau, J., R. E. Page, and J. H. Werren. 1999. Mapping of hybrid
incompatibility loci in Nasonia. Genetics 153:173–174.
Holland, B., and W. R. Rice. 1998. Chase-away sexual selection:
antagonistic seduction versus resistance. Evolution 52:1–7.
Kaneshiro, K. Y. 1983. Sexual selection and the direction of evolution in the biosystematics of Hawaiin Drosophiladae. Annu.
Rev. Entomol. 28:161–178.
Kelly, J. K., and M. A. F. Noor. 1996. Speciation by reinforcement:
a model from studies of Drosophila. Genetics 143:1485–1497.
Kirkpatrick, M. 1987. Sexual selection by female choice in polygynous animals. Annu. Rev. Ecol. Syst. 18:43–70.
Kirkpatrick, M., and M. R. Servedio. 1999. The reinforcement of
mating preferences on an island. Genetics 151(2):865–884.
Krebs, R. A. 1990. Courtship behavior and control of reproductive
isolation in Drosophila mojavenis: genetic analysis of population
hybrids. Behav. Genet. 20:535–543.
Lande, R. 1981. Models of speciation by sexual selection and adaptation on polygenic traits. Proc. Natl. Acad. Sci. USA 78:
3721–3725.
Laven, H. 1959. Speciation by cytoplasmic isolation in the Culex
pipiens complex. Cold Spring Harbor Symp. Quant. Biol. 24:
166–173.
Liou, L. W., and T. D. Price. 1994. Speciation by reinforcement of
premating isolation. Evolution 48:1451–1459.
573
Littlejohn, M. J., and J. J. Loftus-Hills. 1968. An experimental
evaluation of premating isolation in theHyla ewingi complex.
Evolution 22:659–663.
Markow, T. A. 1991. Sexual isolation among populations of Drosophila mojavenis. Evolution 45:1525–1529.
Mayr, E. 1940. Speciation phenomena in birds. Am. Nat. 74:
249–278.
———. 1963. Animal species and evolution. Belknap Press, Cambridge, MA.
Miller, G. L., G. E. Stratton, P. R. Miller, and E. Heberts. 1998.
Geographical variation in male courtship behavior and sexual
isolation in wolf spiders of the genus Schizocosa. Anim. Behav.
56:937–951.
Noor, M. A. 1995. Speciation driven by natural selection in Drosophila. Nature 375:674–675.
Patterson, J. T., and W. S. Stone. 1952. Evolution in the genusDrosophila. Macmillan, New York.
Perrot-Minnot, M. J., L. Guo, and J. H. Werren. 1996. Double and
singleWolbahchia infections in Nasonia vitripennis. Genetics
143:961–972.
Rundle, H. D., and D. Schluter. 1998. Reinforcement of stickleback
mate preferences: sympatry breeds contempt. Evolution 52:
200–208.
Saetre, G. P., T. Moum, S. Bures, M. Kral, M. Adamjan, and J.
Moreno. 1997. A sexually selected character displacement in
flycatchers reinforces premating isolation. Nature 387:589–592.
Sokal, R. R., and F. J. Rohlf. 1981. Biometry. Freeman, San Francisco, CA.
Suzuki, Y., and K. Hiehata. 1985. Mating systems and sex ratios
in the egg parasitoids, Trichogramma dendrolimi and T. papilionis (Hymenoptera: Trichogrammatidae). Anim. Behav. 33:
1223–1227.
Wade, M. J., and N. A. Johnson. 1994. Reproductive isolation between two species of flour beetles, Tribolium castaneum and T.
freemani: variation within and among geographical populations
of T. castaneum. Heredity 72:155–162.
Wade, M. J., N. W. Chang, and M. McNaughton. 1995. Incipient
speciation in the flour beetleTribolium confusum: premating isolation between natural populations. Heredity 75:453–459.
Wade, M. J., N. A. Johnson, R. Jones, V. Siguel, and M. McNaughton. 1997. Genetic variation segregating in natural populations ofTribolium castaneum affecting traits observed in hybrids with T. freemani. Genetics 147:1235–1247.
Watanabe, T. K., and M. Kawanishi. 1979. Mating preference and
the direction of evolution in Drosophila. Science 205:906–907.
West-Eberhard, M. J. 1983. Sexual selection, social competition
and speciation. Q. Rev. Biol. 58:155–183.
Wu, C. I., H. Hollocher, D. J. Begun, C. F. Aquadro, Y. Xu, and
M. Wu. 1995. Sexual isolation inDrosophila melanogaster: a
possible case of incipient speciation. Proc. Natl. Acad. Sci USA.
92:2519–2523.
Zouros, E., and C. J. D’Entremont. 1980. Sexual isolation among
populations ofDrosophila mojavenis: response to pressure from
a related species. Evolution 34:421–430.
Corresponding Editor: T. Markow