doi: 10.1111/j.1420-9101.2009.01759.x
Time-shifted reproductive behaviours among fall armyworm
(Noctuidae: Spodoptera frugiperda) host strains: evidence
for differing modes of inheritance
G. SCHÖFL, D. G. HECKEL & A. T. GROOT
Department of Entomology, Max Planck Institute for Chemical Ecology, Jena, Germany
Keywords:
Abstract
Lepidoptera;
mating interactions;
mode of inheritance;
prezygotic barriers;
reproductive isolation;
temporal isolation.
The noctuid moth Spodoptera frugiperda consists of two strains associated with
different larval host plants (most notably corn and rice). These strains exhibit
differential temporal patterns of female calling and copulation during
scotophase, with the corn strain more active earlier in the night. We
investigated strain-specific constraints in reproductive timing, mating interactions between the two strains, and the mode of inheritance of timing of female
calling, male calling, copulation and oviposition. We observed an allochronic
shift of all reproductive behaviours by approximately 3 h and a parallel shift of
nonreproductive locomotor activity, suggesting involvement of the circadian
clock. The corn strain was more variable in the timing of calling and
copulation than the rice strain. Rice strain females were more restricted in the
timing of copulation than rice strain males, while such differences between the
sexes were not apparent in the corn strain. There were significant interactions
between the strains affecting onset times of copulation and male calling. The
four investigated reproductive traits differed in their modes of inheritance:
timing of female and male calling exhibited strong maternal effects, timing of
copulation was controlled by a combination of maternal effects and corn strain
dominant autosomal factors, and timing of oviposition was inherited in a corn
strain dominant fashion. We conclude that the allochronic separation of
reproduction between fall armyworm strains is asymmetric, less pronounced
than previously thought, and under complex genetic control.
Introduction
For new species to evolve, reproductive barriers must
build up between divergent populations (Dobzhansky,
1937; Mayr, 1963). Recent studies of reproductive
isolation between well-established pairs of model species
have identified and molecularly characterized genes
causing post-zygotic reproductive isolation (Orr et al.,
2004). Although useful in isolating such ‘speciation
genes’, comparisons between extant species suffer from
the limitation that the isolating barriers that are currently
important in restricting gene flow, are not necessarily the
Correspondence: Gerhard Schöfl, Max Planck Institute for Chemical
Ecology, Department of Entomology, Hans-Knöll Straße 8, D-07745 Jena,
Germany.
Tel.: +49 (0)3641 571559; fax: +49 (0)3641 571502;
e-mail: [email protected]
ones that were important during the early phases of
speciation (Coyne & Orr, 1998, 2004). For species that
have evolved in the absence of physical barriers, e.g. as a
result of ecologically based divergent selection (Schluter,
2001; Via, 2001; Rundle & Nosil, 2005), the early stages
of speciation were likely characterized by appreciable
rates of gene flow (Drès & Mallet, 2002; Mallet, 2005;
Magalhaes et al., 2009). The evolution of isolating barriers under such conditions is still poorly understood
(Coyne & Orr, 2004). Studies of ecologically divergent
populations of the same species are then appropriate to
clarify the early causes of reproductive isolation, even if
they might not always evolve into ‘good’ species (Via &
West, 2008).
A promising system for the study of an ongoing
speciation event is the fall armyworm (Spodoptera
frugiperda J. E. Smith). Spodoptera frugiperda is a highly
ª 2009 THE AUTHORS. J. EVOL. BIOL. 22 (2009) 1447–1459
JOURNAL COMPILATION ª 2009 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
1447
1448
G . S C H Ö F L E T A L .
polyphagous noctuid moth native to the New World,
extending from Argentina to the US. Within the fall
armyworm, two major genetic groups have been recognized that exhibit host-plant associated genetic differentiation at a number of mitochondrial and nuclear loci
(Pashley et al., 1985; Pashley, 1986, 1989; Lu et al., 1992;
Lu & Adang, 1996; McMichael & Prowell, 1999; Levy
et al., 2002; Nagoshi & Meagher, 2003b; Nagoshi et al.,
2006). Larvae of one group have been collected predominantly from maize, sorghum and cotton (referred to as
the corn strain), whereas the other group is found mostly
on rice and various pasture grasses (referred to as the rice
strain) (Pashley, 1986, 1988). No diagnostic morphological features have been described to distinguish these two
strains, but they differ consistently in a number of
physiological, developmental and behavioural features
(Pashley, 1988; Pashley et al., 1992, 1995; Groot et al.,
2008). Using multilocus genotypes to examine the degree
and directionality of hybridization between the
fall armyworm strains in nature, Prowell et al. (2004)
concluded that up to 16% of the individuals sampled
were potential hybrids with many of those not being F1
in origin.
The identity and relative importance of the specific
components of reproductive isolation between these two
strains have not been identified unequivocally. No
intrinsic hybrid inviability or sterility has yet been
reported. Earlier reports of unidirectional incompatibilities in interstrain matings in the laboratory (Pashley &
Martin, 1987) were not confirmed by later studies
(Whitford et al., 1988; Quisenberry, 1991). The role of
habitat choice and ⁄ or host fidelity as a potential isolation
mechanism is contentious. The only study that specifically examined oviposition choice between the corn and
the rice strain (Whitford et al., 1988) found some
evidence for strain-specific preferences. Cross-attraction
experiments in the field using live females as lures
showed a significant strain-biased attraction, although
corn strain females attracted large numbers of rice strain
males as well (Pashley et al., 1992).
The mechanism that has been suggested as the most
important contributor to reproductive isolation between
the two strains is a strong temporal divide in mating
activities throughout the night (Pashley et al., 1992). In
S. frugiperda, corn strain individuals have been found to
be reproductively active during the first part of the night,
while rice strain individuals have been found to be active
during the latter part of the night (Pashley et al., 1992).
In Lepidoptera, other examples of diel differences in
reproductive activity between closely related species or
host races have become known in the past decades
(Konno et al., 1996; Monti et al., 1997; Ueno et al., 2006).
Their potential as a barrier to gene flow has only recently
started to attract wider attention (Devries et al., 2008),
and less is known about the genetics of temporal isolation
than about any other isolating barrier (Coyne & Orr,
2004; p. 210).
To further investigate the dynamics of this isolation
mechanism, we have evaluated the allochronic separation of the two host strains of S. frugiperda by assessing (i)
strain-specific constraints vs. plasticity in reproductive
timing, (ii) the consequences of mating interactions
between the two strains on reproductive timing and
(iii) the mode of inheritance of the strain-specific
differences in reproductive timing, under laboratory
conditions. Reproduction in moths will usually follow a
behavioural sequence roughly reflected by four traits: (1)
long-range mate attraction (female calling), (2) closerange courtship (male calling), (3) copulation and (4)
oviposition (e.g. Alexander et al., 1997). While female
calling is the emission of volatile long-range mate finding
signals, males of many moth species display their
abdominal hair pencils during courtship to emit odours
that are important in mate acceptance and mate choice
(e.g. Birch et al., 1990; Hillier & Vickers, 2004). We
quantified the temporal partitioning of these four reproductive traits in within-strain matings, between-strain
matings and backcross matings using hybrid individuals.
We found that the two strains differed not only in the
timing of the reproductive traits, but also in their
temporal patterns of locomotor activity. The allochronic
separation between the two strains was less pronounced
than previously suggested. The corn strain was more
flexible in the timing of calling and copulation than the
rice strain, and significant interactions between the
strains affected onset times of copulation and male
calling. Surprisingly, we found evidence for differing
modes of inheritance for the four investigated reproductive traits, suggesting that the timing of the reproductive
behaviours is under complex genetic control.
Materials and methods
Experimental populations
The corn strain (C) colony used in these experiments
was established from > 100 larvae collected from corn
plants near Homestead in Miami-Dade Co., Florida,
between October and November 2004. The rice strain
(R) colony originated from > 200 larvae collected from
pasture grasses from the Range Cattle Research and
Education Centre, Ona, Hardee Co., Florida, between
May and October 2003. Both colonies were reared in
mass culture for 10 and 21 generations, respectively, on
a pinto bean-based artificial diet at USDA-ARS in
Gainesville, Florida. Since July 2006, colonies have been
maintained in our laboratory in Jena on pinto beanbased artificial diet under a single pair breeding protocol,
designed to minimize inbreeding. Upon receiving the
colonies from Gainesville, 48 putative corn strain and 56
putative rice strain individuals were screened for the
strain-specific cytochrome oxidase I marker. All but
three C individuals were confirmed to be corn strain and
all R individuals were confirmed to be rice strain.
ª 2009 THE AUTHORS. J. EVOL. BIOL. 22 (2009) 1447–1459
JOURNAL COMPILATION ª 2009 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
Temporal isolation of S. frugiperda host strains
Offspring of the three ambiguous C individuals were not
included in the subsequent rearing. Populations are
cultured in environmental chambers at 26 C under a
photoperiod of 14 : 10 light : dark.
Observation of nocturnal reproductive behaviours
All nocturnal behaviours were observed in a walk-in
environmental chamber at 26 C. To facilitate observations, the environmental chamber’s photoperiod was
time-shifted, so that scotophase started at 12:00 h and
ended at 22:00 h. Observations were conducted during
the 10 h scotophase and continued for 1 h into photophase (from 12:00 to 23:00). In each experiment, one
virgin female and one virgin male, each 1–4 days old,
were placed together in clear 16-oz plastic cups and
provided with 10% honey solution. Cups were capped
with cheese cloth. Observations were performed in
three different generations on (i) within-strain matings,
(ii) between-strain matings and (iii) backcross matings
between hybrid and pure-strain individuals. The number of mating pairs observed at each generation ranged
from 320 to 363. All mating pairs observed in one
generation were set up on the same day and simultaneously placed in the environmental chamber 2 h
before the onset of scotophase. Observations were
conducted for three consecutive nights on each set of
mating pairs.
Within-strain matings
To assess strain-specific differences in the timing of
reproduction, we observed 158 corn strain pairs and
162 rice strain pairs (Trial 1). We scored copulation,
female calling (extrusion of pheromone glands), male
calling (extrusion of male abdominal hair pencils) and
oviposition. During the first 2 days, we also scored any
additional actions (e.g. feeding, flight or walking) jointly
as general locomotor activity. Observations were done at
30-min intervals using an electric torch fitted with redlight diodes. Each pair was observed for about 3–5 s. A
single round of observation generally lasted 20 min. The
order of the mating cups in the environmental chamber
was randomized and strain identity was unknown to the
observers.
Between-strain and within-strain matings
To assess the consequences of mating interactions
between the two strains, and to determine the extent
of male and female influence on the timing of reproductive behaviour, we observed 100 C-female · R-male
crosses and 100 R-female · C-male crosses (Trial 2). In
this trial, we also observed 80 corn strain pairs and 80 rice
strain pairs, which served as a control. The betweenstrain matings generated the F1 hybrids, which were used
in the third trial (see below). We used the same mating
procedure as described for the first trial, with the
exception of monitoring general activity patterns.
1449
Backcross matings with hybrid individuals
To assess the mode of inheritance of the strain-specific
differences in the timing of reproduction and to estimate
the relative maternal and paternal contributions to the
timing of reproductive behaviours, we observed F1
hybrids in all possible backcross combinations (i.e.
CR$ · C#, CR$ · R#, RC$ · C#, RC$ · R#, C$ · CR#,
R$ · CR#, C$ · RC# and R$ · RC#) (Trial 3). When
referring to the origin of hybrids, the female parent is
given first. For instance, CR refers to hybrid progeny of a
cross between a corn strain female and a rice strain male.
For each cross combination, we observed 40–57 pairs,
following the same observational protocol as above, with
the exception of monitoring general activity patterns. A
detailed breakdown of the sample sizes for all types of
crosses is presented in Supporting Information Table S1.
Inference of mode of inheritance
In Lepidopterans, females are the heterogametic (ZW)
sex, the Z-chromosome of the females is inherited from
the father. Z-linkage can thus be demonstrated when
female hybrids show trait values similar to the paternal
source population in both reciprocal crosses. W-linkage,
cytoplasmic, or maternal effects are demonstrated when
female hybrids show trait values that are similar to the
maternal source population. Reciprocal crosses that show
trait values intermediate between the parental source
populations indicate additive autosomal gene effects,
while trait values similar to one of the parental source
populations indicate nonadditive (dominance) effects.
Statistical analyses
Assumptions of normality and homogeneity of variances
were not met for onset times of reproductive behaviours
in the within-strain matings of Trial 1 (Shapiro-Wilk
normality tests, P < 0.01 in all cases; Levene’s test for
homogeneity of variances, P = 0.09 to P < 0.001). Therefore, differences between the two strains in the average
onset times of reproductive behaviours were tested using
nonparametric Wilcoxon signed-rank tests corrected for
multiple testing (Holm, 1979). Observing the same
mating pairs for three consecutive nights allowed testing
for differences in onset times between the first time that a
behaviour occurred and subsequent occurrences of the
same behaviour. These differences were tested using
nonparametric Wilcoxon signed-rank tests for each strain
for each trial. To create an overall test for significance, we
used Fisher’s method of combining probabilities (Sokal &
Rohlf, 1995).
For the within- and between-strain matings of Trial 2,
onset times of female calling, copulation and oviposition
departed from normal distributions. Box-Cox transformations (package M A S S 7Æ2–42) were used to improve the
fit to normality. We analysed the transformed onset
times of female calling, copulation and oviposition
ª 2009 THE AUTHORS. J. EVOL. BIOL. 22 (2009) 1447–1459
JOURNAL COMPILATION ª 2009 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
1450
G . S C H Ö F L E T A L .
(k = 1.3, 1.5 and 0.5 respectively), and untransformed
onset times of male calling by fitting linear mixed-effects
models (package N L M E 3.1–89) (Pinheiro & Bates, 2000).
To determine the effects of different types of crosses on
the timing of reproduction, type of cross and the order of
occurrence of a behaviour (i.e. whether a behaviour was
observed for the first time in a pair vs. any subsequent
time) were modelled as fixed effects. To determine the
effects of the female and the male mating partners on
timing of reproduction, strain identity (C or R) of the
female and the male, their interaction, and the order of
occurrence of a behaviour were modelled as fixed effects.
Repeated observations within pairs were modelled as a
random effect. Because we found significant heteroscedasticity between types of crosses in the timing of female
calling and copulation (Levene’s test for homogeneity of
variances: female calling P < 0.001, copulation P <<
0.001), which could not be resolved by transformations,
we fitted heteroscedastic models when dealing with these
two traits. These models equalize residual variances
across groups by weighing them using the VarIdent
variance function implemented in the N L M E package
(Pinheiro & Bates, 2000). The significances of specific
differences between cross types were tested post hoc by
Tukey’s honest significant difference (HSD) tests using
the M U L T C O M P package (1.0–2).
For matings involving hybrids (Trial 3), Box-Cox
transformations removed departures from normality,
except for onset time of oviposition. Therefore, we
analysed untransformed onset times of oviposition by
nonparametric Kruskal–Wallis A N O V A s and performed
post hoc tests for specific differences between types of
crosses via nonparametric multiple comparison based on
relative contrast effects (package N P A R C O M P 1.0–0). We
analysed Box-Cox transformed onset times of female
calling and copulation (k = 1.3 and 1.5, respectively),
and untransformed onset times of male calling by
fitting (heteroscedastic) linear mixed-effects models as
described above. Specific differences between types of
crosses were tested by Tukey’s HSD tests. Models fitting
the strain identity (C, R, CR or RC) of the female and the
male as a fixed effect could not be estimated directly
because of linear dependencies in the model matrix.
Therefore, we used likelihood ratio tests (LRT) to determine the respective effects of the mating partners on
timing of female calling, male calling and copulation as
follows. We compared full models fitting types of crosses
(i.e. all eight female · male combinations) as a fixed
effect, to reduced models fitting only the strain identity of
the male or of the female mating partner as a fixed effect.
For onset time of oviposition, the effects of the mating
partners were tested using the Scheirer-Ray-Hare extension of the Kruskal–Wallis test as a nonparametric
alternative (Sokal & Rohlf, 1995;. pp. 445–447). All
analyses were performed using the R 2.8.0 statistical
software (R Development Core Team, 2008; http://
www.r-project.org/).
Results
Within-strain matings (Trial 1)
We observed significant differences between the corn
and rice strains in the onset times of all nocturnal
reproductive activities studied (i.e. female and male
calling, copulation and oviposition), even though onset
times of female calling and male calling were highly
variable in the corn strain (Fig. 1a,b and Table 1). On
average, rice strain females and males started to call 3.0
and 2.6 h later than corn strain females and males,
respectively. In both strains, onset time of female calling
was significantly correlated with onset time of male
calling (in rice strain, only after the removal of two
outliers; Fig. 2).
In the corn strain, onset of copulations peaked
between 3 and 6 h into scotophase. In the rice strain,
copulations were largely restricted to the last 4 h of the
scotophase (Fig. 1c). Only 5% of all rice strain pairs
started copulating earlier. On average, rice strain pairs
started to copulate 3.2 h after corn strain pairs. Despite
the significantly earlier average onset time of copulation
in corn strain (Table 1), there was a considerable overlap
between the two strains: 37% of all corn strain copulations were started during the last 4 h of scotophase.
Considering only pairs that copulated at least twice in
different nights, the average within-individual range in
onset times of copulation was markedly higher in corn
strain (corn strain: 3.4 ± 2.3 h, N = 103; rice strain:
1.6 ± 1.4 h, N = 102). 53% of the corn strain pairs
initiated copulation at least once during the first 6 h
and at least once during the last 4 h of the night; 35%
started copulating exclusively during the first 6 h of the
night; 12% started copulating exclusively during the last
4 h of the night. In contrast, in the rice strain 90% of the
multiply copulating pairs initiated copulation exclusively
during the last 4 h of the night, while none started
copulations exclusively during the first 6 h of the night.
Ovipositions were only observed on the second or third
night after at least one copulation had taken place. Rice
strain females started ovipositing throughout the night,
while a majority of corn strain females initiated egglaying during the first few hours of the scotophase
(Fig. 1d). On average, rice strain females initiated oviposition 2.4 h later than corn strain females.
The onset time of female calling was earlier on
subsequent nights (‘non-first callings’) than on the night
of first occurrence (‘first callings’) (Table 1). Comparable
trends were found in Trials 2 (within- and betweenstrain crosses) and 3 (backcrosses) (data not shown).
Combining data from all three trials showed that this
difference was significant only for the rice strain (Fisher’s
combined probability for corn strain, v2 = 7.4, d.f. = 6,
P = 0.28; for rice strain v2 = 18.5, d.f. = 6, P = 0.005).
The opposite pattern was observed for timing of copulation: both corn and rice strain pairs copulated later the
ª 2009 THE AUTHORS. J. EVOL. BIOL. 22 (2009) 1447–1459
JOURNAL COMPILATION ª 2009 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
Temporal isolation of S. frugiperda host strains
Fig. 1 Temporal distribution of onset
times of (a) female calling, (b) male calling,
(c) copulation and (d) oviposition during
scotophase for corn and rice strain individuals of Spodoptera frugiperda. Histograms and
box plots indicating median, interquartile
range, and range, present data on initiation
times from observations on 158 corn strain
pairs and 162 rice strain pairs pooled over
three consecutive nights. For histograms,
observations were binned in 1-h intervals.
***P < 0.001; Wilcoxon rank sum tests
with continuity correction; P-values adjusted
for multiple testing.
Table 1 Timing of nocturnal behaviours of
Spodoptera frugiperda host strains, averaged
over 3 consecutive nights of observation.
Proportion of individuals that displayed the onset of a trait
(a)
0.30
0.20
1451
(b)
Female calling
0.30
Corn strain
Rice strain
0.20
Male calling
Corn strain
Rice strain
0.10
0.10
0.00
0.00
***
***
2
4
6
8
10
2
(c)
0.30
0.20
4
6
8
10
(d)
Copulation
0.4
Corn strain
Rice strain
Oviposition
Corn strain
Rice strain
0.3
0.2
0.10
0.1
0.00
0.0
***
2
4
6
8
***
10
2
Hours into scotophase
4
6
8
10
Trait
Strain
N
Mean ± SD (h)
Dt (h)*
P-value
Onset time of female calling (first callings)
C
R
C
R
C
R
C
R
C
R
C
R
C
R
C
R
76
88
44
25
72
45
123
119
87
81
117
113
85
77
115
97
4.8
8.0
4.5
7.1
4.5
7.1
4.9
8.1
5.4
8.6
2.7
1.5
2.4
1.3
2.9
5.2
2.5
< 0.001
2.6
< 0.001
2.6
< 0.001
3.2
< 0.001
3.2
< 0.001
1.1
< 0.001
1.1
< 0.001
2.4
< 0.001
Onset time of female calling (nonfirst callings)
Onset time of male calling
Onset time of copulation (first copulations)
Onset time of copulation (nonfirst copulations)
Duration of copulation (first copulations)
Duration of copulation (nonfirst copulations)
Onset time of oviposition
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
3.0
1.6
2.4
2.2
2.7
2.5
2.6
1.7
2.1
1.3
1.9
1.0
1.6
0.8
2.5
2.7
*Average difference in onset times between the two host strains in hours.
Wilcoxon rank sum test with continuity correction; P-values adjusted for multiple testing
(Holm, 1979).
second and third time than the first time they copulated
(Table 1). Combining data from all three trials showed
that this difference was significant in the corn strain
(Fisher’s combined probability, v2 = 20.9, d.f. = 6,
P = 0.002) and marginally significant in the rice strain
(Fisher’s combined probability, v2 = 12.2, d.f. = 6,
P = 0.06). Therefore, order effects were included as fixed
effects in subsequent linear mixed-effects models analyses for female calling and copulation. Too few observations of multiple occurrences of male calling and
oviposition precluded meaningful comparisons for these
behaviours (data not shown).
The temporal profile of general activity patterns (feeding and locomotion) for the two strains paralleled the
ª 2009 THE AUTHORS. J. EVOL. BIOL. 22 (2009) 1447–1459
JOURNAL COMPILATION ª 2009 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
Onset time of male calling
1452
G . S C H Ö F L E T A L .
Corn strain
10
8
8
6
6
4
4
2
2
0
Between-strain matings (Trial 2)
Rice strain
10
0
0
2
4
6 8 10
0 2 4
Onset time of female calling
6
8
10
Fig. 2 Relationship between the onset time of female calling and the
onset time of male calling within strains. Left: corn strain. Pearson
product moment correlation r = 0.65, P < 0.001, n = 41. Right: rice
strain. Pearson product moment correlation (significant only after
the removal of two outliers) r = 0.71, P < 0.001, n = 18. In the
figure, the data are presented including outliers. The dashed lines
have a slope of unity and indicate concurrent observations of female
and male calling. Overlaying data points are denoted by enlarged
symbols.
Proportion of individuals
that were active
1.0
Corn strain
Corn strain
Rice strain
Rice strain
0.8
0.6
0.4
0.2
0.0
0
2
4
6
Hours into scotophase
8
10
Fig. 3 Temporal distribution of locomotor activity of the two host
strains of Spodoptera frugiperda during scotophase. Observations on
158 corn strain pairs and (light line) and 162 rice strain pairs (dark
line) were pooled over two consecutive nights.
temporal patterns of reproductive behaviours (Fig. 3).
After an initial peak in activity after lights-off, corn strain
individuals remained active at relatively high levels until
shortly before the end of scotophase. Rice strain individuals were less active than corn strain individuals during
the first half of the night, but increased activity during
the latter part of the night until shortly before lights-on.
The difference in activity between the two strains was
highly significant for both females (Kolmogorov–Smirnov two-sample test, D2063,1823 = 0.222, P < 0.001) and
males (D2311,2100 = 0.217, P < 0.001). We detected no
significant differences between the sexes within strains
(corn strain: D2063,2311 = 0.038, P = 0.08; rice strain:
D1823,2100 = 0.034, P = 0.22).
Comparison of within-strain matings and between-strain
matings allowed estimation of the influence of each
strain on the timing of reproductive behaviours of the
other. For onset time of female calling, the linear mixedeffects model revealed a significant main effect of female
strain identity (t106 = 4.98, P < 0.001), but not of male
strain identity (t106 = )0.17, P = 0.86). There was no
significant interaction effect between male and female
strain identity (t106 = 0.24, P = 0.81). Thus, males did not
influence the onset time of female calling (Fig. 4a).
With onset times of male calling, the linear mixedeffects model revealed significant main effects of female
strain identity (t68 = 2.18, P = 0.03) and male strain
identity (t68 = 2.86, P = 0.006), but no significant interaction effect between male and female strain identity
(t68 = )0.57, P = 0.57). This shows that both males and
females influenced onset time of male calling (Fig. 4b).
With onset time of copulation, the linear mixed-effects
model revealed significant main effects of female strain
identity (t304 = 12.7, P < 0.001) and male strain identity
(t304 = 3.86, P < 0.001). In this case, there was also a
significant interaction effect between male and female
strain identity (t304 = )3.57, P < 0.001). Rice strain
females paired with different male partners did not differ
in their onset time of copulation, while corn strain
females copulated significantly later when paired with
rice strain males than when paired with corn strain males
(Tukey’s HSD, z = 3.86, P < 0.001; Fig. 4c). The magnitude of the effect predicted from the linear mixed-effects
model was considerably larger for female strain identity
(b^female = 8.53; 95% CI = 7.21–9.85), than for male
strain identity (b^male = 3.22; 95% CI = 1.58–4.86) or
the interaction between female and male strain identity
(b^femalemale = )3.51; 95% CI = )5.44 to )1.57).
For onset time of oviposition, we found a significant
main effect of female strain identity (t184 = 4.88,
P < 0.001), but not of male strain identity (t184 = 1.79,
P = 0.08). There was no significant interaction effect
between male and female strain identity (t184 = )1.31,
P = 0.19). Thus, males did not influence onset time of
oviposition (Fig. 4d).
Backcross matings with hybrid individuals (Trial 3)
In hybrid · parental strain pairings (CR$ · C#,
CR$ · R#, RC$ · C#, RC$ · R#, C$ · CR#, R$ · CR#,
C$ · RC# and R$ · RC#), the strain identity of the
mating partners affected reproductive timing in a similar
way as found in crosses between parental strains (previous paragraph). Only females significantly influenced
timing of female calling (LRT for the female effect,
LR = 53.5, P < 0.001; for the male effect, LR = 8.5,
P = 0.08). Both males and females significantly influenced timing of male calling (LRT for the female effect,
LR = 15.1, P = 0.005; for the male effect, LR = 14.6,
ª 2009 THE AUTHORS. J. EVOL. BIOL. 22 (2009) 1447–1459
JOURNAL COMPILATION ª 2009 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
Temporal isolation of S. frugiperda host strains
Hours into scotophase
(a)
10
Fig. 4 Average onset times of (a) female
calling, (b) male calling, (c) copulation and
(d) oviposition in crosses within and
between host strains of Spodoptera frugiperda.
Figures show mean, SE (boxes) and SD
(whiskers). Different letters indicate significant differences [Tukey’s honest significant
difference (HSD) post hoc tests, a = 0.05,
see Table S2].
(b)
Female calling
10
8
8
6
6
4
4
2
2
(c)
10
1453
(d)
10
Copulation
Male calling
Oviposition
8
8
6
6
4
4
2
2
P = 0.006) and copulation (LRT for the female effect,
LR = 61.9, P < 0.001; for the male effect, LR = 16.7,
P = 0.002). Only females significantly influenced timing
of oviposition (Scheirer-Ray-Hare extension of the Kruskal–Wallis A N O V A , H = 36.1, Pfemale << 0.001, H = 1.64,
Pmale = 0.44).
To infer modes of inheritance, female hybrids mated to C
or R males in Trial 3 were compared with the pure strain
pairs generated in the previous generation in Trial 2
(likewise male hybrids mated to C or R females in the case
of male calling) (Fig. 5). Combining data across trials here
is valid because no consistent generational effects were
observed (Supporting Information Fig. S1). For subsequent analyses, we pooled hybrids of the same class when
no significant influence of the strain identity of their
mating partner was detected (Table S3). However, in
Fig. 5, we present all data for the hybrids separated with
respect to their mating partners, irrespective of whether
they showed a significant influence on onset times or not.
deeper level by directly modelling the effects of the
parental strain identity of each female on onset time of
female calling (cf. Nygren et al., 2006). Taking into
account all possible hybrid · parental strain pairings of
Trial 3 as well as the within-strain pairings of Trial 2
simultaneously, this approach allows estimating the
relative maternal and paternal contributions to the
timing of a female calling. Employing a linear mixedeffects model, the effect of the female’s maternal strain
identity was highly significant (P < 0.001) and showed
the largest predicted effect size, while the effect of the
female’s paternal strain identity was marginally significant (P = 0.06) and the predicted effect size considerably smaller (Table 2). There was no significant
interaction between the female’s maternal and the
female’s paternal strain identity. This corroborates that
the strain-specific differences in onset time of female
calling are controlled by a predominantly maternal
pattern of inheritance, which may be modulated by
some additive effects.
Inheritance of onset time of female calling
Onset times of female calling were not statistically
different between pure corn strain females and CR-hybrid females, or between pure rice strain females and
RC-hybrid females (Fig. 5a). CR- and RC-hybrid
females differed significantly from each other (Tukey’s
HSD, z = 2.97, P = 0.02). This suggests that the strainspecific differences in onset time of female calling are
controlled by a maternal pattern of inheritance. The
analysis of patterns of inheritance can be taken to a
Inheritance of onset time of male calling
The analysis of onset time of male calling was complicated by the fact that both sexes significantly influenced
male calling times. Disregarding the biasing effect of
female mating partners of different strain identity, Fig. 5b
suggests an overall pattern of maternal inheritance
similar to what we observed for female calling. Statistical
support for maternal inheritance provides the observation that CR-hybrid males called earlier than RC-hybrid
ª 2009 THE AUTHORS. J. EVOL. BIOL. 22 (2009) 1447–1459
JOURNAL COMPILATION ª 2009 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
1454
G . S C H Ö F L E T A L .
(a)
(b)
Hours into scotophase
Female calling
10
10
8
8
6
6
4
4
2
2
(c)
10
Male calling
(d)
10
Copulation
Oviposition
8
8
6
6
4
4
2
2
Table 2 Effects of parental strain identity of females (and males) on
the onset times of reproductive behaviours.
Model terms
Onset time of female calling
Female’s mother
Female’s father
Female’s mother · female’s
father
Onset time of male calling
Female’s mother
Female’s father
Male’s mother
Male’s father
Onset time of copulation
Female’s mother
Female’s father
Male’s mother
Male’s father
Onset time of oviposition
Female’s mother
Female’s father
Female’s mother · female’s
father
Effect
SE
d.f.
t
P-value
3.49
1.09
0.60
0.85
0.59
1.07
167
167
167
4.08
1.86
0.56
< 0.001
0.06
0.58
0.63
1.08
1.21
)0.01
0.41
0.36
0.35
0.33
158
158
158
158
1.56
2.97
3.45
)0.4
0.12
0.003
< 0.001
0.97
2.96
1.25
1.61
0.91
0.39
0.43
0.34
0.31
398
398
398
398
7.68
2.92
4.72
2.96
< 0.001
0.004
< 0.001
0.003
262
262
262
)0.37
)0.24
2.42
)14.7
)4.09
104.9
39.6
17.4
43.3
0.71
0.81
0.02
Effect and SE indicate the predicted effect sizes of each term on
(transformed ⁄ ranked) onset times of reproductive behaviours and
their standard errors obtained from linear mixed-effects models fit by
R E M L . For onset time of male calling and copulation interaction
terms could not be estimated jointly because of linear dependencies
in the model matrix.
Fig. 5 Average onset times of (a) female
calling, (b) male calling, (c) copulation and
(d) oviposition of Spodoptera frugiperda host
strains in within-strain crosses (outer plots)
and crosses between F1 hybrid and parental
strain individuals (inner plots). The figures
show mean, SE (boxes) and SD (whiskers).
Data for hybrid females (males in the case of
male calling) is presented separate with
respect to the strain identity of the mating
partner (inner plots). Lines below letters join
types of crosses that were pooled for statistical comparisons of classes of females
(males) when no significant difference with
regard to the male (female) mating partner
was detected (see Supporting Information
Table S3). Different letters indicate significant differences (Tukey’s HSD post hoc tests,
a = 0.05, see Table S4).
males in the presence of rice strain females (Tukey’s HSD,
z = )4.88, P < 0.001). This notion was further supported
when testing for the effect of the parental strain identity
of each female and of each male using all possible
backcross combinations of Trial 3 and the within-strain
pairings of Trial 2. We found significant effects for the
male’s maternal strain identity but not for the male’s
paternal strain identity (Table 2). Onset time of male
calling was also significantly affected by the female’s
paternal strain identity (Table 2).
Inheritance of onset time of copulation
Analyses of between-strain matings suggested that
females exerted more control over onset times of copulation than males (see above). Therefore, we considered
this trait from the females’ perspective (Fig. 5c). With
onset time of copulation, the two reciprocal hybrid
categories did not differ significantly from each other.
However, while RC-hybrid females were significantly
different from both parental strains, CR-hybrid females
(when paired with C-males) did not differ significantly
from pure corn strain females (Fig. 5c), suggesting some
degree of corn strain dominance. Estimating the maternal and paternal contributions in both sexes to the timing
of a copulation revealed significant effects of both the
female’s and the male’s maternal strain identity, as well
as the female’s and the male’s paternal strain identity
(Table 2). The predicted effect size of the female’s
ª 2009 THE AUTHORS. J. EVOL. BIOL. 22 (2009) 1447–1459
JOURNAL COMPILATION ª 2009 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
Temporal isolation of S. frugiperda host strains
maternal strain identity (b^ = 2.96; 95% CI = 2.20–3.71)
was significantly greater than the predicted effect sizes of
the female’s paternal strain identity (b^ = 1.25; 95% CI =
0.41–2.10) and both the male’s maternal (b^ = 1.61; 95%
CI = 1.05–2.17) and paternal strain identity (b^ =0.92;
95% CI = 0.40–1.43). This supports the notion that onset
time of copulation is controlled to a larger degree by
females than by males, and that onset time of mating is
influenced by a combination of maternal and corn strain
dominant effects.
Inheritance of onset time of oviposition
Neither CR-hybrid-females nor RC-hybrid females differed significantly from pure corn strain females in onset
time of oviposition, whereas both F1 hybrid classes
differed significantly from pure rice strain females
(Fig. 4d). In a model testing for the effect of the parental
strain identity of each female, only the interaction
between the females’ maternal and paternal strain
identity was significant (Table 2). This suggests an autosomal corn strain dominant mode of inheritance for
onset time of oviposition.
Discussion
Our observations of a total of 480 pure strain pairs, 200
between-strain pairs and 320 pairs involving hybrids for
three consecutive nights have given us a detailed picture
of timing differences of the four reproductive traits
(female calling, male calling, copulation and oviposition)
between the two strains in S. frugiperda. As we observed
pairs in all possible cross combinations, we could determine the effects of the mating partner, as well as of their
parents, on these traits. We first summarize each of the
traits separately, then we discuss the interaction effects
between the mating partners and the mode of inheritance of the traits.
Individual traits
The average initiation time of female calling differed
significantly between the two strains, even though corn
strain females showed a very broad distribution of onset
time of calling throughout the night (Fig. 1a). When
first-time calling was compared with subsequent nights,
rice strain females started calling earlier the second or
third time, which further reduced the average gap
between the two strains. We found no indication that
the timing of female calling was affected by the presence
of the male mating partner. The onset time of female
calling was predominantly maternally inherited:
C$ · R#-hybrids behaved similar to the pure corn strain
females, while R$ · C#-hybrids behaved similar to pure
rice strain females (Fig. 5a).
The average differences in timing of male calling
between the two host strains, as well as their distribu-
1455
tions, matched those of female calling (Fig. 1b), and the
initiation times of male and female calling were closely
correlated within strains (Fig. 2). Onset time of male
calling was controlled not only by the males but also by
the females (further discussed below). Surprisingly, the
onset time of male calling was also maternally inherited
(Fig. 5b and Table 2).
Copulations in the corn strain peaked between 3 and
6 h into scotophase (Fig. 1c), although roughly 37% of
all corn strain copulations coincided with the later rice
strain copulations. The rice strain copulated mostly in the
last part of the night, only 5% of all rice strain matings
occurred earlier than during the last 4 h of scotophase.
Considering only multiple copulations of the same pairs
at different nights, the corn strain showed large withinindividual variability, i.e. there were no distinct groups of
‘late copulating pairs’ and ‘early copulating pairs’. Onset
time of copulation seems to be controlled by a combination of corn-strain dominant and maternal effects:
CR-hybrid females and RC-hybrid females copulated
significantly earlier than pure rice strain females, but
CR-hybrid females (mated to C males) were not significantly different from pure corn strain females (Fig. 5c),
indicating some degree of corn strain dominance. The
RC-hybrids were significantly different from both parental strains in their timing of copulations (Fig. 5c),
suggesting a maternal contribution to onset times of
copulation. Both the female’s and male’s maternally
inherited genotype exhibited a greater influence over
onset time of copulation than the two sexes’ paternally
inherited genotypes, confirming the suggestion of a
maternal contribution to onset time of copulation
(Table 2).
Onset times of oviposition were biased towards the first
4 h into scotophase in the corn strain, while they were
roughly uniformly distributed throughout the scotophase
in the rice strain (Fig. 1d). Onset time of oviposition
showed no indication of maternal effects. Instead, this
trait was inherited in a purely corn strain dominant
fashion: CR-hybrid females and RC-hybrid females
oviposited significantly earlier than pure rice strain
females, but were not different from pure corn strain
females (Fig. 5d). When testing for the effects of the
parental strain identity of the females on the onset times
of oviposition, only the interaction between the female’s
maternally and paternally derived genotypes was significant. This indicates nonadditive effects.
Overlap in reproductive timing between the strains
The behaviour of the corn strain in our study differed
from the behaviour of the corn strain in the first
report on differential timing of nocturnal activities in
S. frugiperda (Pashley et al., 1992). In the previous study,
corn strain copulations were clearly restricted to the first
6 h of a 10-h scotophase and did not overlap with rice
strain copulations (see Fig. 1 in Pashley et al., 1992).
ª 2009 THE AUTHORS. J. EVOL. BIOL. 22 (2009) 1447–1459
JOURNAL COMPILATION ª 2009 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
1456
G . S C H Ö F L E T A L .
Three explanations are possible for this discrepancy. First,
the strains used here have been reared in the laboratory
for much longer (15 and 26 generations for the corn and
the rice strain, respectively) than the strains used in
Pashley et al.’s (1992) study (field collection, 3 and 10
generations for two corn strain samples and a rice strain
sample, respectively). However, this does not explain
why constraints should relax because of prolonged
rearing in the laboratory in the corn strain alone, and
not in the even older rice strain colony. Second, the
sample sizes employed in the previous study may not
have been large enough to observe the full range of
variation in these timing traits in the corn strain of the
fall armyworm. Pashley et al.’s (1992) data for the corn
strain were based on observations on 16 mating pairs,
while our data is based on observations on 123 corn
strain pairs. Finally, there may be geographic variation in
these behavioural traits. In our study, the corn strain
derived from South Florida while in the previous study
eight corn strain pairs derived from Louisiana and eight
pairs came from Florida (the precise place of origin
within Florida was not given).
Effects of an interaction between the sexes on the
timing of reproductive behaviours
In the behavioural sequence of reproduction, onset times
of female calling and oviposition are expected to be
determined by the female alone, because under field
conditions neither behaviour requires (or will usually be
performed in) the presence of a male. As males of many
moth species have been found to extrude their hair
pencils in the close proximity of the female (Birch et al.,
1990), the timing of male calling is likely to be the result
of an interaction between the two sexes. The onset time
of copulation clearly depends on an interaction between
the female and the male. While it is difficult to directly
determine the extent of male and female influence over a
behaviour such as onset time of copulation, cross-strain
mating experiments can be used to address this issue,
because the average activity peaks of the two mating
partners are different (see, e.g. Holwell, 2008).
As predicted, the timing of male calling was not only
dependent on the male but also strongly affected by the
strain identity of the female mating partner. In interstrain
crosses, C-males paired with R-females, and R-males
paired with C-females, called on average at a time
intermediate between pure corn strain and rice strain
pairs (Fig. 4b). The correlation between the onset times
of female and male calling (Fig. 2) suggests that the
interaction with the female may be elicited by the female
sex pheromone. Even though this pheromone may not
be confined to the mating cups but could also emanate
into the climate chamber, this ‘chemical background
noise’ did not affect the timing of male calling in purestrain crosses: R-males paired with C-females called
earlier than R-males paired with R-females. Hence, the
concentration of the background pheromone is most
likely too low in comparison to that inside the mating
cups to elicit a response from males. In addition, other
time-shifted cues of female receptiveness may affect the
temporal control of male calling. This may explain why
males sometimes started calling before females (cf. Fig 2).
While both sexes exerted a similar influence on the
timing of male calling, the pattern was more complicated
for the timing of copulation. C-females paired with Rmales copulated on average 1.5 h later than when paired
with C-males, whereas copulation times of R-females
were not affected as much by the strain of the males they
were paired with (see Fig. 4c). Interestingly, model
predictions of effect sizes of female and male strain
identity on onset times of copulation implied that females
exert a greater influence over the onset of copulation
than males. The asymmetric interaction between the
strains suggests that (i) the corn strain is more flexible in
the timing of copulation relative to the rice strain and (ii)
the mating phase is more restricted for R-females than for
R-males. Therefore, rice strain females rather than rice
strain males seem to cause the restriction of rice strain
matings to the last four hours of the scotophase.
Differing modes of inheritance and the genetic basis
of differential timing of reproduction
All four reproductive behaviours measured here appear
to be shifted co-ordinately by approximately 3 h between
the two strains (Fig. 1). Moreover, their time shift is
paralleled by a temporal shift in general locomotor
activity (Fig. 3). This suggests that a single common
genetic factor controls the timing of these activities,
mostly likely a gene involved in the circadian system (e.g.
Sakai & Ishida, 2001; Miyatake et al., 2002; Tauber et al.,
2003; An et al., 2004; Rymer et al., 2007).
Surprisingly, the results of our crosses indicate differing
modes of inheritance for the differences in timing of the
four investigated reproductive behaviours. Two traits
(timing of copulation and oviposition) showed evidence
for autosomal corn strain dominant control. Three traits
(timing of female calling, male calling and copulation)
showed varying degrees of maternal inheritance. As
female lepidopterans are the heterogametic sex (ZW) and
males are the homogametic sex (ZZ), a maternal mode of
inheritance can be caused either by the (maternally
transmitted) W-chromosome, by maternal effects (cytoplasmatic factors) or both. In traits limited to the female
sex (onset time of female calling), these factors are
confounded. In a trait expressed in both sexes (onset time
of copulation) or limited to the male sex (onset time of
male calling), these factors can be separated. For male
calling and copulation, our results imply that cytoplasmatic effects rather than W-linked factors cause the
maternal inheritance.
The idea of a single genetic factor controlling the timing
of reproduction and other activities in the S. frugiperda host
ª 2009 THE AUTHORS. J. EVOL. BIOL. 22 (2009) 1447–1459
JOURNAL COMPILATION ª 2009 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
Temporal isolation of S. frugiperda host strains
strains seems to be at odds with the finding of differing
modes of inheritance for these behaviours. It is conceivable, however, that the timing all four behaviours is under
the control of one or a few corn strain dominant autosomal
factor(s), which is overridden – to varying degrees – by
maternal cytoplasmic effects.
Our study showed no evidence for a Z-chromosomal
influence on the timing of any of the four traits. This is
striking, because a Z-chromosomal control for timing of
reproduction in S. frugiperda has been suggested before
(J. McNeil, cited as unpublished data in Prowell, 1998;
Prowell et al., 2004). In other moths, temporal partitioning of copulation (Chilo suppressalis, Samudra et al., 2002)
and of female calling (Spodoptera latifascia and Spodoptera
descoinsi, Monti et al., 1997) have been shown to be
under polygenic additive autosomal control. None of
these studies attempted to infer the mode of inheritance
of more than one component of reproductive behaviour.
Ours is thus the first to suggest that a time shift that
equally affects different reproductive behaviours may
show differing modes of inheritance.
Differential timing of reproduction as a prezygotic
barrier to gene flow
As isolating barriers act sequentially in the life cycle of
hybridizing taxa, early-acting barriers may contribute
more to total isolation than later-acting barriers, even if
they are of equal strength (Schemske, 2000; Ramsey
et al., 2003). The temporal partitioning of reproduction in
S. frugiperda has been suggested to be the major prezygotic isolation barrier between the two strains (Pashley
et al., 1992). However, our study shows that, although
being significantly differentiated, there is a large overlap
in the timing of both female calling and copulation
between the two strains. Our data indicate that rice strain
females are more restricted in the timing of copulation
than rice strain males. Therefore, if the differences in
onset time of copulation were a major contributor to
prezygotic isolation, this should translate into R$ · C#hybrids being rarer than the reciprocals. Molecular
evidence for directionality of hybridization in the field
does not support this notion, but suggests a roughly equal
representation of CR- and RC-hybrids (Prowell et al.,
2004) or even an under-representation of CR-hybrids
(Nagoshi & Meagher, 2003a).
Our results do not invalidate the idea that temporal
partitioning of reproduction is an important prezygotic
isolation barrier, but they do call for a re-evaluation of
other probable prezygotic isolating barriers in the fall
armyworm, and their potential interactions. Studies
testing for habitat isolation of the S. frugiperda host
strains, either through differences in oviposition preferences or differential larval survival on different host
plants, have been inconclusive to date (Pashley, 1988;
Whitford et al., 1988; Pashley et al., 1995; Veenstra et al.,
1995; Prowell et al., 2004) and need to be further
1457
investigated. The other most obvious behavioural prezygotic barrier is differential mating preferences because
of sex pheromone differences (Pashley et al., 1992; Groot
et al., 2008). Recently, we (Groot et al., 2008) found
significant differences in the female sex pheromone
composition between the two host strains. Mate attraction tests in the field using live females of both strains as
a lure showed a significant strain-biased preference,
although corn strain females still attracted large numbers
of rice strain males (Pashley et al., 1992). These data
considered only total trap catches throughout the night
and not the potential interaction of pheromone chemistry with differences in the timing of female calling. The
total reproductive isolation achieved in nature is potentially much higher than the individual evaluations of
allochronic separation and differential attraction to sex
pheromones suggests. Both may act simultaneously and
could therefore contribute multiplicatively rather than
cumulatively to reproductive isolation. To test this, the
next step is to assess to what extent these different
pheromone blends are differentially attractive to corn
and rice strain males, and to take into account the timing
of mate attraction during the night.
In summary, the two strains of the fall armyworm
differ in the nocturnal timing of four reproductive traits
and also in their locomotor activity. The allochronic
separation of reproduction between the two strains is less
pronounced than previously thought. The corn strain is
more flexible in the timing of calling and copulation
relative to the rice strain. Interactions between the sexes
suggested that rice strain females were more restricted to
the timing of copulation than rice strain males. In the
corn strain, differences between the sexes were not
apparent. The co-ordinated time shift between reproductive traits and locomotor activity suggests an involvement of the circadian clock. Surprisingly, however, the
four investigated reproductive traits showed evidence for
differing modes of inheritance: the timing of female and
male calling was controlled mainly by maternal effects,
the timing of copulation was controlled by a combination
of maternal effects and corn strain dominant autosomal
factors, and the timing of oviposition was inherited in a
purely corn strain dominant fashion. This indicates that
the different reproductive behaviours are under complex
genetic control. The effectiveness of allochronic isolation
needs to be tested in interaction with other isolation
mechanisms in the laboratory and the field.
Acknowledgments
The authors are most grateful to Rob Meagher for
providing us with the fall armyworm colonies used in
this study. The authors thank Sabine Hänniger, Steffen
Reifarth, and Johannes Fleischmann for assistance
with the rearing, the setting-up of the crosses, and
the observations. This work was supported by the
Max-Planck-Gesellschaft.
ª 2009 THE AUTHORS. J. EVOL. BIOL. 22 (2009) 1447–1459
JOURNAL COMPILATION ª 2009 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
1458
G . S C H Ö F L E T A L .
References
Alexander, R.D., Marshall, D.C. & Cooley, J.R. 1997. Evolutionary perspectives on insect mating. In: The Evolution of Mating
Systems in Insects and Arachnids (B.J. Crespi & J.C. Choe, eds),
pp. 4–31. Cambridge University Press, Cambridge.
An, X., Tebo, M., Song, S., Frommer, M. & Raphael, K.A. 2004.
The cryptochrome (cry) gene and a mating isolation mechanism in Tephritid fruit flies. Genetics 168: 2025–2036.
Birch, M.C., Poppy, G.M. & Baker, T.C. 1990. Scents and eversible
scent structures of male moths. Annu. Rev. Entomol. 35: 25–58.
Coyne, J.A. & Orr, H.A. 1998. The evolutionary genetics of
speciation. Philos. Trans. R. Soc. Lond. B Biol. Sci. 353: 287–305.
Coyne, J.A. & Orr, H.A. 2004. Speciation. Sinauer Associates, Inc.,
Sunderland, MA.
Devries, P.J., Austin, G.T. & Martin, N.H. 2008. Diel activity and
reproductive isolation in a diverse assemblage of Neotropical
skippers (Lepidoptera: Hesperiidae). Biol. J. Linn. Soc. Lond. 94:
723–736.
Dobzhansky, T. 1937. Genetics and the Origins of Species. Columbia
University Press, New York.
Drès, M. & Mallet, J. 2002. Host races in plant-feeding insects
and their importance in sympatric speciation. Philos. Trans. R.
Soc. Lond. B Biol. Sci. 357: 471–492.
Groot, A.T., Marr, M., Schöfl, G., Lorenz, S., Svatoš, A. & Heckel,
D.G. 2008. Host strain specific sex pheromone variation in
Spodoptera frugiperda. Front. Zool. 5: 20.
Hillier, N.K. & Vickers, N.J. 2004. The role of heliothine
hairpencil compounds in female Heliothis virescens (Lepidoptera: Noctuidae) behavior and mate acceptance. Chem. Senses
29: 499–511.
Holm, S. 1979. A simple sequentially rejective multiple test
procedure. Scand. J. Stat. 6: 65–70.
Holwell, G.I. 2008. Male control of copulation duration in
Ciulfina praying mantids (Insecta Mantodea). Ethol. Ecol. Evol.
20: 61–70.
Konno, Y., Tsukuba, K. & Tanaka, F. 1996. Mating time of the
rice-feeding and water-oat-feeding strains of the rice stem
borer, Chilo suppressalis (Walker) (Lepidoptera: Pyralidae). Jpn.
J. Appl. Entomol. Z. 40: 245–247.
Levy, H.C., Garcia-Maruniak, A. & Maruniak, J.E. 2002. Strain
identification of Spodoptera frugiperda (Lepidoptera: Noctuidae)
insects and cell line: PCR-RFLP of cytochrome oxidase C
subunit I gene. Fla. Entomol. 85: 186–190.
Lu, Y.J. & Adang, M.J. 1996. Distinguishing fall armyworm
(Lepidoptera: Noctuidae) strains using a diagnostic mitochondrial DNA marker. Fla. Entomol. 79: 48–55.
Lu, Y.J., Adang, M.J., Isenhour, D.J. & Kochert, G.D. 1992.
Restriction fragment length polymorphism analysis of genetic
variation in North American populations of the fall armyworm
Spodoptera frugiperda (Lepidoptera: Noctuidae). Mol. Ecol.
1: 199–208.
Magalhaes, I.S., Mwaiko, S., Schneider, M.V. & Seehausen, O.
2009. Divergent selection and phenotypic plasticity during
incipient speciation in Lake Victoria cichlid fish. J. Evol. Biol.
22: 260–274.
Mallet, J. 2005. Hybridization as an invasion of the genome.
Trends Ecol. Evol. 20: 229–237.
Mayr, E. 1963. Animal Species and Evolution. Belknap Press of
Harvard University Press, Cambridge, Mass.
McMichael, M. & Prowell, D.P. 1999. Differences in amplified
fragment length polymorphisms in fall armyworm (Lepidop-
tera: Noctuidae) host strains. Ann. Entomol. Soc. Am. 92: 175–
181.
Miyatake, T., Matsumoto, A., Matsuyama, T., Ueda, T., Toyasato,
T. & Tanimura, T. 2002. The period gene and allochronic
reproductive isolation in Bactrocera cucurbitae. Proc. R. Soc. Lond.
(Biol.) 269: 2467–2472.
Monti, L., Genermont, J., Malosse, C. & Lalanne-Cassou, B.
1997. A genetic analysis of some components of reproductive
isolation between two closely related species, Spodoptera
latifascia (Walker) and S. descoinsi (Lalanne-Cassou and Silvain) (Lepidoptera: Noctuidae). J. Evol. Biol. 10: 121–134.
Nagoshi, R.N. & Meagher, R. 2003a. Fall armyworm FR
sequences map to sex chromosomes and their distribution in
the wild indicate limitations in interstrain mating. Insect Mol.
Biol. 12: 453–458.
Nagoshi, R.N. & Meagher, R.L. 2003b. FR tandem-repeat
sequence in fall army-worm (Lepidoptera: Noctuidae) host
strains. Ann. Entomol. Soc. Am. 96: 329–335.
Nagoshi, R.N., Meagher, R.L., Adamczyk, J.J., Braman, S.K.,
Brandenburg, R.L. & Nuessly, G. 2006. New restriction
fragment length polymorphisms in the cytochrome oxidase I
gene facilitate host strain identification of Fall armyworm
(Lepidoptera: Noctuidae) populations in the Southeastern
United States. J. Econ. Entomol. 99: 671–677.
Nygren, G.H., Nylin, S. & Stefanescu, C. 2006. Genetics of host
plant use on life history in the comma butterfly across Europe:
varying modes of inheritance as a potential reproductive
barrier. J. Evol. Biol. 19: 1882–1893.
Orr, H.A., Masly, J.P. & Presgraves, D.C. 2004. Speciation genes.
Curr. Opin. Genet. Dev. 14: 675–679.
Pashley, D.P. 1986. Host associated genetic differentiation in fall
armyworm: a sibling species complex? Ann. Entomol. Soc. Am.
79: 898–904.
Pashley, D.P. 1988. Quantitative genetics, development and
physiological adaptation in sympatric host strains of fall
armyworm. Evolution 42: 93–102.
Pashley, D.P. 1989. Host-associated differentiation in armyworms: an allozymic and mtDNA perspective. In: Electrophoretic Studies on Agricultural Pests (H. Loxdale & M.F. Claridge,
eds), pp. 103–114. Oxford University Press, London, UK.
Pashley, D.P. & Martin, J.A. 1987. Reproductive incompatibility
between host strains of the fall armyworm (Lepidoptera:
Noctuidae). Ann. Entomol. Soc. Am. 80: 731–733.
Pashley, D.P., Johnson, S.J. & Sparks, A.N. 1985. Genetic
population structure of migratory moths: the fall armyworm
(Lepidoptera: Noctuidae). Ann. Entomol. Soc. Am. 78: 756–
762.
Pashley, D.P., Hammond, A.M. & Hardy, T.N. 1992. Reproductive isolation mechanisms in fall armyworm host strains
(Lepidoptera: Noctuidae). Ann. Entomol. Soc. Am. 85: 400–405.
Pashley, D.P., Hardy, T.N. & Hammond, A.M. 1995. Host effects
on developmental and reproductive traits in fall armyworm
strains (Lepidoptera: Noctuidae). Ann. Entomol. Soc. Am. 88:
748–755.
Pinheiro, J.C. & Bates, D.M. 2000. Mixed-Effects Models in S and
S-PLUS. Springer, New York.
Prowell, D.P. 1998. Sex linkage and speciation in Lepidoptera.
In: Endless Forms: Species and Speciation (D.J. Howard & S.H.
Berlocher, eds), pp. 309–319. Oxford University Press, Oxford,
UK.
Prowell, D.P., McMichael, M. & Silvain, J.F. 2004. Multilocus
genetic analysis of host use, introgression, and speciation in
ª 2009 THE AUTHORS. J. EVOL. BIOL. 22 (2009) 1447–1459
JOURNAL COMPILATION ª 2009 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY
Temporal isolation of S. frugiperda host strains
host strains of fall armyworm (Lepidoptera: Noctuidae). Ann.
Entomol. Soc. Am. 97: 1034–1044.
Quisenberry, S.S. 1991. Fall armyworm (Lepidoptera: Noctuidae) host strain reproductive compatibility. Fla. Entomol. 72:
194–199.
R Development Core Team 2008. R: A Language and Environment for Statistical Computing. R Foundation for Statistical
Computing, Vienna.
Ramsey, J., Bradshaw, H.D. & Schemske, D.W. 2003. Components of reproductive isolation between the monkeyflowers
Mimulus lewesii and M. Cardinalis (Phrymaceae). Evolution 57:
1520–1534.
Rundle, H.D. & Nosil, P. 2005. Ecological speciation. Ecol. Lett. 8:
336–352.
Rymer, J., Bauernfeind, A.L., Brown, S. & Page, T.L. 2007.
Circadian rhythms in the mating behavior of the cockroach,
Leucophaea maderae. J. Biol. Rhythm. 22: 43–57.
Sakai, T. & Ishida, N. 2001. Circadian rhythms of female mating
activity governed by clock genes in Drosophila. Proc. Natl. Acad.
Sci. USA 98: 9221–9225.
Samudra, I.M., Emura, K., Hoshizaki, S., Ishikawa, Y. & Tatsuki,
S. 2002. Temporal differences in mating behavior between
rice- and water-oats-populations of the striped stem borer,
Chilo suppressalis (Walker) (Lepidoptera: Crambidae). Appl.
Entomol. Zool. 37: 257–262.
Schemske, D.W. 2000. Understanding the origin of species.
Evolution 54: 1069–1073.
Schluter, D. 2001. Ecology and the origin of species. Trends Ecol.
Evol. 16: 372–380.
Sokal, R.R. & Rohlf, F.J. 1995. Biometry. The Principles and Practice
of Statistics in Biological Research, 3rd edn. W. H. Freeman and
Co, New York, NY.
Tauber, E., Roe, H., Costa, R., Hennessy, J.M. & Kyriacou, C.P.
2003. Temporal mating isolation driven a behavioral gene in
Drosophila. Curr. Biol. 13: 140–145.
Ueno, H., Furukawa, S. & Tsuchida, K. 2006. Difference in the
time of mating activity between host-associated populations of
the rice stem borer, Chilo suppressalis (Walker). Entomol. Science
9: 255–259.
Veenstra, K.H., Pashley, D.P. & Ottea, J.A. 1995. Host-plant
adaptation in fall armyworm host strains: comparison of food
1459
consumption, utilization, and detoxification enzyme activities.
Ann. Entomol. Soc. Am. 88: 80–91.
Via, S. 2001. Sympatric speciation in animals: the ugly duckling
grows up. Trends Ecol. Evol. 16: 381–390.
Via, S. & West, J. 2008. The genetic mosaic suggests a new role for
hitchhiking in ecological speciation. Mol. Ecol. 17: 4334–4345.
Whitford, F., Quisenberry, S.S., Riley, T.J. & Lee, J.W. 1988.
Oviposition preference, mating compatibility, and development of two fall armyworm strains. Fla. Entomol. 71: 234–243.
Supporting information
Additional supporting information may be found in the
online version of this article:
Table S1 Sample sizes for all three trials.
Table S2 Tukey’s honest significant difference (HSD) post
hoc tests between crosses in Trial 2 (within- and betweenstrain matings).
Table S3 Tukey’s HSD post hoc tests (nonparametric
comparison of relative contrast effects in the case of
oviposition) between crosses in Trial 3 (backcrosses).
Table S4 Tukey’s HSD post hoc tests (nonparametric
comparison of relative contrast effects in the case of
oviposition) between hybrid classes of females (males, in
the case of male calling) in Trial 3 (backcrosses).
Figure S1 Average onset times of (a) female calling, (b)
male calling, (c) copulation and (d) oviposition in crosses
of females of either strain (males in the case of male
calling) with pure strain and hybrid males (females in the
case of male calling).
Please note: Wiley-Blackwell are not responsible for
the content or functionality of any supporting materials
supplied by the authors. Any queries (other than missing
material) should be directed to the corresponding author
for the article.
Received 24 November 2008; revised 22 March 2009; accepted 27 March
2009
ª 2009 THE AUTHORS. J. EVOL. BIOL. 22 (2009) 1447–1459
JOURNAL COMPILATION ª 2009 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY