Int. J. Plant Sci. 166(3):475–480. 2005.
Ó 2005 by The University of Chicago. All rights reserved.
1058-5893/2005/16603-0009$15.00
PRIOR SELFING AND GYNOMONOECY IN SILENE NOCTIFLORA L.
(CARYOPHYLLACEAE): OPPORTUNITIES FOR ENHANCED
OUTCROSSING AND REPRODUCTIVE ASSURANCE
Sandra L. Davis1 and Lynda F. Delph
Department of Biology, University of Indianapolis, Indianapolis, Indiana 46227, U.S.A.; and
Department of Biology, Indiana University, Bloomington, Indiana 47405, U.S.A.
The production of seed by both selfing and outcrossing, or mixed mating, may be selected for under
conditions of variable pollinator availability. Seed production was examined following experimental
manipulations of perfect and pistillate flowers of the gynomonoecious plant Silene noctiflora. Style excision
experiments in a greenhouse demonstrated that autonomous selfing occurs in the perfect flowers prior to their
opening and continues as the flower ages; 45% of ovules were fertilized prior to flower opening and 95% by
the third night of opening. Field experiments conducted in four populations showed that autonomous selfing of
perfect flowers also occurs in the field. Perfect flowers excluded from pollinators had similar seed set to perfect
flowers that were open to pollinators, and emasculated flowers had very low levels of seed set. These results are
concordant with the hypothesis that selfing may be selected to confer reproductive assurance when pollinator
visitation limits seed set. Pistillate flowers also showed significant pollinator limitation because hand-pollinated
pistillate flowers had significantly higher fruit set than those open to pollinators. Pollinator visitation, while
low, occurs during both day and night, even though flowers are only fully open at night. The production of
perfect flowers capable of autonomous selfing, together with the plastic production of pistillate flowers, confers
reproductive assurance and enhances the likelihood of producing outcrossed seed.
Keywords: gynomonoecy, mixed mating, outcrossing, pollinator limitation, prior selfing, reproductive
assurance, seed discounting.
Introduction
plants within populations (Takebayashi 2000). The selfing
that does occur as a consequence of such variation in floral
traits may occur at different times within an individual flower’s life span. Hence, mixed mating can occur because of
three different modes of autonomous selfing that occur
within flowers and that refer to the timing of selfing relative
to outcrossing: prior, competing, and delayed. Theory has
shown that the ecological conditions under which each mode
may be selected differ, with the conditions for prior selfing to
evolve being the most stringent (Lloyd 1992). Selection for
reproductive assurance will favor prior selfing (selfing that
occurs before any opportunity for outcrossing), only if the
probability of being outcrossed later is low (chronic pollinator limitation) and/or inbreeding depression is weak (Lloyd
1992; Elle and Hare 2002). At the other extreme is delayed
selfing (selfing that occurs after all opportunities for outcrossing are past), which is almost always favored by selection,
because it provides reproductive assurance without limiting
opportunities for outcrossing via either seeds or pollen (Lloyd
1992).
Mixed mating can also occur in species that produce two
types of flowers on the same plant. For example, some plants
produce both cleistogamous and chasmogamous flowers, and
while obligate selfing occurs in the first, the latter can be outcrossed (e.g., Impatiens capensis; Schemske 1978; Lu 2000).
Gynomonoecy is another system involving two flower types
that allows for mixed mating because plants produce both
perfect (hermaphroditic) and pistillate (male-sterile) flowers.
Many models of the evolution of plant mating systems predict that plants should evolve toward either complete selffertilization or complete outcrossing (Lande and Schemske
1985; Charlesworth et al. 1990; Uyenoyama et al. 1993;
Lande et al. 1994). However, studies investigating selfing
rates in natural systems have indicated that mixed-mating
systems—those that combine both selfing and outcrossing—
may be more common than models have predicted (Brown
and Clegg 1984; Barrett et al. 1996; Vogler and Kalisz
2001). Models and empirical studies attempting to explain
the maintenance of mixed mating within populations have
focused on the effects of inbreeding depression (Uyenoyama
and Waller 1991; Rausher and Chang 1999; Cheptou and
Schoen 2003), seed and pollen discounting (Holsinger 1988;
Herlihy and Eckert 2002), and reproductive assurance (Lloyd
1992; Kephart et al. 1999; Kalisz and Vogler 2003; Tsitrone
et al. 2003).
Mixed mating within populations can be accomplished in
a variety of ways and occurs at different levels. For example,
among-plant variation in herkogamy (physical separation of
male and female parts within a flower) may lead to variation
in the degree of selfing compared with outcrossing among
1
Author for correspondence; e-mail [email protected].
Manuscript received August 2004; revised manuscript received December
2004.
475
476
INTERNATIONAL JOURNAL OF PLANT SCIENCES
In this study, we examined the potential for pollinator limitation to influence the mating system of Silene noctiflora, a gynomonoecious annual with plastic pistillate flower production
(Folke and Delph 1997). The potential exists for prior selfing
to occur within the perfect flowers in this species because
some of the anthers dehisce and come into contact with the
styles prior to flower opening. Prior selfing could impact the
potential for later outcrossing in that already selfed ovules
would be preempted from being outcrossed (seed discounting
sensu Lloyd 1992). Therefore, we report results of an experiment to determine the timing of autonomous selfing within
perfect flowers. Second, we report results of emasculation,
bagging, and hand-pollination experiments performed in natural populations to determine the potential for outcrossing via
pollinator visitation in both perfect and pistillate flowers. This
allowed us to determine whether reproductive assurance was
provided by autonomous selfing (Schoen and Lloyd 1992). Finally, we use pollinator exclusion experiments to quantify the
rate and timing of pollinator visitation.
Material and Methods
Study Species
Silene noctiflora L. (Caryophyllaceae) is a gynomonoecious
annual that grows in dense clumps in disturbed soil, e.g.,
along roadsides. As indicated by its name, S. noctiflora is
a night-flowering plant; flowers open in the early evening,
and the petals curl into a ‘‘closed’’ position by midmorning.
Individual flowers open and close in this way for 3 d before
senescing. Flowers are thought to be pollinated principally by
nocturnal moths (Bittrich 1993).
Individual plants produce two types of flowers: perfect
(hermaphroditic) flowers that contain functional ovules and
pollen, and pistillate (male-sterile) flowers that lack pollen
but contain functional ovules. In both flower types, stigmatic
papillae are distributed along the entire adaxial surface of
three styles. Hence, pollen deposited anywhere along the adaxial surface can germinate. Perfect flowers contain two
whorls of five anthers with varying filament lengths. In most
populations, the first whorl of anthers dehisces in the afternoon prior to a flower’s first opening such that pollen comes
into direct contact with the upper third of the styles, which
are located within the floral tube (S. L. Davis and L. F.
Delph, unpublished data). The second whorl of anthers dehisces 1 d later. Pistillate flowers are approximately the same
size as perfect flowers, but they differ in other ways. In addition to producing vestigial anthers on greatly reduced filaments, they also differ in their style length. While the styles
of most perfect flowers do not extend beyond the floral tube,
styles of pistillate flowers extend outward from the opening
ca. 3–4 mm (Folke and Delph 1997).
Timing of Autonomous Self-Fertilization
We quantified the proportion of seeds autonomously selfed
at various stages in the life of the flower, using plants grown
from seed in a greenhouse at Indiana University. We define
autonomous self-fertilization as that which occurs without
the aid of a pollinator, whereas autogamy refers to all forms
of selfing except geitonogamy (sensu Lloyd and Schoen
1992). These seeds were originally collected from a population growing near Mountain Lake Biological Station. Perfect
flowers on each of 28 plants received one of three styleremoval treatments: (1) styles removed in the late afternoon
just prior to the flower’s first opening, (2) styles removed
prior to the second night of flower opening, and (3) styles
removed prior to the third and last night of a flower’s life.
Sample sizes were n ¼ 27, 27, and 28, respectively. The percentage of ovules fertilized was determined for each treatment by collecting the nearly fully developed fruit at a later
date and recording the number of swollen, developing ovules
and the number of shrunken, unfertilized ovules. There was
no indication of any abortion of ovules after this period, so
these numbers are an accurate indication of final seed production. These data allowed us to determine the timing of
self-fertilization, with the first treatment estimating the proportion of ovules that would be autonomously self-fertilized
prior to any potential pollinator visitation (i.e., prior selfing)
and the final treatment indicating the total percentage of
ovules autonomously fertilized. Therefore, seed production in
these treatments indicates the potential for reproductive assurance in the absence of pollinators.
Data were analyzed using a one-way ANOVA using the
statistical program Systat, with the proportion of ovules fertilized as the dependent variable and treatment as the independent variable. A Tukey’s a posteriori test was done to
determine statistical differences between specific pairs of
treatments. Variances among treatments were unequal and
an arcsine (square-root) transformation still did not provide
equal variances. A nonparametric Kruskal-Wallis test, which
completes the analysis on ranked values, was therefore performed to corroborate the results of the one-way ANOVA.
Field Studies
Field studies were conducted in natural populations growing in the vicinity of the University of Virginia’s Mountain
Lake Biological Station in Giles County, Virginia. Experiments were conducted at four major sites located at: (1)
37°17.259N, 80°34.969W (elevation 609 m); (2) 37°19.809N,
80°29.499W (elevation 619 m); (3) 37°20.189N, 80°33.519W
(elevation 959 m); and (4) 37°17.979N, 80°36.119W (elevation 519 m). The populations were observed from June 1,
2002 through August 1, 2002.
Selfing vs. outcrossing in perfect flowers. Perfect flowers
of S. noctiflora appear to be constructed in such a way that
allows for prior selfing. However, if some ovules remain unfertilized after the flower opens, there is still the potential for
perfect flowers to produce seeds through outcrossing. This
possibility was examined by manipulating perfect flowers on
plants growing in the four natural populations described
above. Populations varied in size, and consequently 15–25
plants from each population were chosen. Three flowers on
each plant that were to open for the first time in the coming
evening were randomly assigned to one of three different
treatments: bagged, emasculated, or control. Because all
three treatments were applied to every plant, this design controlled for maternal effects. Nylon mesh bags were placed
around bagged flowers to prevent pollinator visitation.
DAVIS & DELPH—PRIOR SELFING IN GYNOMONOECIOUS SILENE NOCTIFLORA
Bagged flowers could only produce seed autonomously. In
emasculated flowers, anthers were removed while the flower
was still in bud to prevent them from being able to produce
seed via autogamous selfing, leaving only geitonogamy and
outcrossing as options. While emasculation was unlikely to
alter signals to floral visitors because the anthers are contained within the floral tube, this placement did complicate
anther removal. Emasculation was performed by making
a small slit in the side of the calyx tube with a pair of sharp
forceps, which were then used to pull the anthers off the filaments. The filaments were then gently pushed back into the
tube. To control for this manipulation, the flowers in the
other two treatments were also slit, but the anthers were not
removed. The control flower on each plant was otherwise
not manipulated, thereby potentially producing autogamously selfed, geitonogamously selfed, and outcrossed seed.
Fruits were collected after seeds had developed but prior to
fruit opening (ca. 16 d after treatment) in order to prevent
seed loss. The number of seeds and unfertilized ovules from
each of these fruits were counted to determine the proportion
of developing seed (seed set) under each pollination treatment.
Differences in seed set were analyzed with the general linear
model ANOVA procedure of SPSS 11.0 for Macintosh, with
treatment as a fixed effect, population as a random effect, and
Type III sum of squares. Variances among groups were heterogeneous because of 0 seed set by some flowers in the emasculated treatment. Therefore, P values should be considered
approximate. However, given the low P values we obtained,
this is not likely to change our results or interpretation.
Pollen limitation in pistillate flowers. Because they are unisexual, pistillate flowers are dependent on pollen received
through pollinator visitation for seed set. To determine
whether fruit or seed production by pistillate flowers is limited
by the amount of pollen received, we performed the following
experiment. When two pistillate flowers opened simultaneously on the same plant, each was randomly assigned either
to be hand-pollinated with pollen from another plant (growing
at least 5 m away) or left to be open-pollinated. Fruit was collected ca. 16 d after pollination. Whether the flower developed
into a fruit was recorded (fruit set). Fruit set was recorded as
1.0 if a fruit developed and 0.0 if a fruit failed to develop. A
nonparametric sign test was performed on 33 pairs of flowers
using the statistical program SPSS 11.0 for Macintosh to test
for differences in fruit set between hand- and open-pollinated
flowers. Population was not incorporated as a variable because there were too few plants in any given population with
two pistillate flowers open simultaneously to adequately test
for an effect of population. Because most flowers did not produce fruit, seed counts were not made.
Diurnal vs. nocturnal pollination. The floral shape, color,
and nocturnal anthesis of S. noctiflora indicate moths as the
likely pollinator. However, during the course of our studies, insects were observed visiting flowers of S. noctiflora during
daylight hours, including flies, small bees, and bumble bees. Although the petals of S. noctiflora flowers curl inward during
the day, they do not obstruct the opening of the corolla tube,
so pollination is still feasible. To determine the possible contribution of diurnal pollinators, triplets of unopened perfect
flowers were chosen on 30 plants at site number 1. Each
flower was emasculated to remove anthers as described above,
477
so any seed set would be the result solely of pollinator visitation. One flower on each plant was bagged at dusk each evening and unbagged at dawn. Another flower was bagged at
dawn and unbagged at dusk. The final flower was left open to
allow for visitation during both the night and the day. Bagging
was performed using mesh string-tie sock style mantles for
Coleman lanterns. Bagging treatments were maintained through
initial flower opening until the flower senesced from June 23
to June 28, 2002. Fruits were collected ca. 16 d later and
scored for fruit set. Fruit set was recorded as 1.0 if a fruit developed and 0.0 if a fruit failed to develop. Road crews
mowed part of the field site during the course of the experiment, so the final sample size receiving all three treatments
was reduced to 18 plants. Because the majority of flowers
failed to set fruit, most values were 0.0. Differences among
the treatments in fruit set were analyzed by a x2 analysis using
the Crosstabs procedure of SPSS 11.0 for Macintosh. Because
most flowers did not produce fruit, seeds were not counted.
Results
The Timing of Autonomous Self-Fertilization
The timing of style removal had a significant effect on the
number of ovules fertilized (one-way ANOVA, F2; 79 ¼ 33:6,
P < 0:001; fig. 1). The results of a Kruskal-Wallis analysis
yielded similar results (Kruskal-Wallis test statistic ¼ 38:74,
df ¼ 2, P < 0:001). When styles were removed just prior to
flower opening, the number of ovules fertilized was significantly lower than if styles were removed after one or two
nights of being open, indicating that prior selfing, while high,
was <100%. However, after two nights of being open, autonomous selfing resulted in all of the ovules being fertilized
in many of the flowers, with an overall average of 95%. Because not all ovules were fertilized at the time of flower
opening, in addition to the high levels of prior selfing, at least
some autonomous seed set in Silene noctiflora may be caused
by the other two forms of autonomous selfing (competing
and delayed selfing).
Fig. 1 Timing of pollination in perfect flowers of Silene noctiflora
as determined by style removal at different times throughout the life of
the flower. Bars indicate means (61 SE). Treatments designated with
different letters are significantly different according to Tukey’s a
posteriori test (P < 0:05).
INTERNATIONAL JOURNAL OF PLANT SCIENCES
478
Selfing vs. Outcrossing in Perfect Flowers
Treatment had a significant effect on seed set in perfect
flowers in the field (table 1; fig. 2). Emasculated flowers had
significantly reduced seed set as compared to control and
bagged flowers (P < 0:001 for both comparisons), but there
was no significant difference between control and bagged
flowers (P ¼ 0:95). Although the effect of population was
not significant, the treatment by population interaction was,
indicating that the effect of treatment varied among populations. Control flowers set slightly more seeds in one population than bagged flowers, whereas it was the other way
around in three populations (fig. 2). However, emasculated
flowers had lower seed set than the other two treatments in
all populations, including two populations in which emasculated flowers produced no seeds, with an overall mean of
only 6%. Overall, these results indicate that most fertilization in perfect flowers in the field occurred via autonomous
selfing without the aid of pollinators.
Pollen Limitation in Pistillate Flowers
Pistillate flowers can produce fruit only if pollinators visit
them. Hand-pollinated pistillate flowers set significantly
more fruit than pistillate flowers that were not manipulated
and that were left open to pollination (P < 0:006 using a
two-tailed sign test); mean fruit set (61 SE) was 65:52 6 0:09
vs. 17:86 6 0:07, respectively. These results indicate that fruit
set in pistillate flowers was limited by the amount of pollen
received.
Diurnal vs. Nocturnal Pollination
Fruit set in all three treatments (both nocturnal and diurnal pollination, nocturnal pollination alone, and diurnal pollination alone) was very low (table 2). A x2 test showed
there was no statistical difference among the three treatments
(x2 ¼ 0:328, P ¼ 0:849). This is congruent with the data
from the selfing vs. outcrossing experiment, which indicate
that the amount of pollination by insects is low. The data
from this experiment indicate that pollinator visits to flowers
are indeed low and seem to be fairly well divided between diurnal and nocturnal insects.
Discussion
Our results show that the combination of prior selfing and
gynomonoecy results in a mixed-mating system in Silene noc-
Table 1
Summary of Two-Way ANOVA in Seed Set in Perfect Flowers
Following Experimental Pollination Treatments (Control,
Bagged, Emasculated) in Four Natural
Populations of Silene noctiflora
Source
Treatment
Population
Treatment 3 population
Error
df
Mean square
F
P
2
3
6
200
6.232
0.898
0.331
0.008
21.34
2.72
4.01
...
0.001
0.137
0.001
...
Fig. 2 Percentage of ovules (61 SE) maturing into seed in four
different populations of Silene noctiflora in control flowers (possibility
of autogamy, geitonogamy, and outcrossing), bagged flowers (possibility of autonomous autogamy), and emasculated flowers (possibility of
geitonogamy and outcrossing). Emasculation resulted in a significant
reduction in seed set compared with either of the other two treatments,
but control and bagged treatments did not significantly differ from each
other. No emasculated flowers set seed in populations 2 and 4, so the
bars for these populations have a value of 0 with no variance.
tiflora. Hence, the dual advantages of selfing and outcrossing
(see Lloyd 1992) are afforded to the plant by making two
types of flowers. Self-fertilization advantages include the
transmission advantage (Fisher 1941) and reproductive assurance (Darwin 1876), whereas the advantage of outcrossing is
the production of high-quality offspring that do not suffer
from inbreeding depression (Darwin 1876). We discuss our
results relative to these dual advantages.
Reproductive Assurance
Autonomous selfing can be selected for when pollinator
visitation is low because it provides reproductive assurance
(Darwin 1876; Baker 1955). Evidence that selfing evolved to
allow reproductive assurance comes from studies on species
that exhibit among-population variation in the degree of selfing. In these species, self-pollinating populations occur in
areas with lower plant density and/or lower pollinator visitation than outcrossing populations (e.g., Eichornia paniculata,
Barrett et al. 1989; Blandfordia grandiflora, Ramsey and
Vaughton 1996; Clarkia xantiana, Fausto et al. 2001; Collinsia parviflora, Elle and Carney 2003). One can also evaluate
whether pollinator limitation may have led to the evolution
of selfing by examining the extent to which autonomous selfing compared with pollen deposition by pollinators contributes to seed production within natural populations (Schoen
and Lloyd 1992; Herlihy and Eckert 2002; Kalisz and Vogler
2003). This is the approach we took here. Our results are
concordant with the hypothesis that autonomous selfing provides reproductive assurance in S. noctiflora.
We found that pollinator limitation occurred in all four of
our study populations, even though both diurnal and nocturnal insects contributed to pollination. Emasculated perfect
DAVIS & DELPH—PRIOR SELFING IN GYNOMONOECIOUS SILENE NOCTIFLORA
Table 2
Number of Fruit Produced by Emasculated Perfect Flowers
Open to Diurnal and/or Nocturnal Pollination
Bagging
treatment
Not bagged
Bagged during the night
Bagged during the day
No. flowers
that set fruit
No. flowers
that aborted
3
2
2
15
16
16
flowers and pistillate flowers left open to pollination had
greatly reduced seed set as compared with intact or handpollinated flowers. While pollinator limitation may lead to
the evolution of autonomy because it provides reproductive
assurance, it is not required in all populations or all years for
this evolution to occur (Schoen and Brown 1991). Nevertheless, Jürgens et al. (1996) also observed significant autogamy
and low levels of pollination by insects in populations of
S. noctiflora in Germany. Pollinator limitation may therefore
be a common condition for this species.
In spite of limited opportunities for outcrossing, we found
that most of the ovules of perfect flowers in the field are fertilized because of the plant’s ability to autonomously self.
This was demonstrated by our finding that bagged perfect
flowers had similar seed set to those left open to pollinators
(62% vs. 60%, respectively) and that perfect flowers blooming in a pollinator-free greenhouse had most of their ovules
fertilized. Furthermore, we showed that a significant portion
of this autonomy comes in the form of prior selfing. In our
style-removal experiment, 45% of ovules were fertilized prior
to any possible insect visit, as a result of direct dehiscence of
half of the anthers onto receptive stigmatic papillae when the
flower is in bud. Although this leaves 55% of the ovules still
unfertilized when the flower is open to pollination, insect visitation appeared to add little to fertilization. While delayed
selfing is adaptive whenever inbreeding depression is <1,
prior selfing is predicted to evolve only when pollinators are
chronically scarce because this minimizes or even eliminates
seed and pollen discounting (Lloyd 1992). As mentioned, this
seems to be the case in our study populations, given that
emasculated perfect flowers had extremely low seed set.
Enhancing Opportunities for Outcrossing
Because fertilization in perfect flowers is primarily through
autonomy, the production of pistillate flowers enhances the
opportunity for outcrossing to occur. In this way, S. noctiflora could be classified as having a mixed-mating system.
Impatiens capensis, which produces both cleistogamous and
chasmogamous flowers, has a somewhat analogous way of
achieving mixed mating. One flower type produces seed via
autonomous selfing (cleistogamous flowers) whereas the other
does so primarily via outcrossing (chasmogamous flowers)
(Schemske 1978; Lu 2000). Chasmogamous flowers, while
not male-sterile, cannot autogamously self because of strong
protandry; the fused anthers form a cap over the stigma, and
the stigma is exposed to pollination only after this cap falls
off (Schemske 1978).
479
Gynomonoecy is prevalent in the composite family, where
pistillate flowers are usually ray flowers located on the outer
perimeter of the inflorescence. This location increases the
probability that they will be outcrossed because pollinators
usually land on the edges of inflorescences and deposit pollen
from previously visited plants (Cheptou et al. 2001; Gibson
and Tomlinson 2002). Gynomonoecy is relatively rare in
other plant families, perhaps because geitonogamy reduces
outcrossing and the advantages related to it. For example,
Collin and Shykoff (2003) found that pistillate flowers
produced on hermaphroditic individuals of gynodioecious
Dianthus sylvestris were more selfed than perfect flowers
on the same plants because pollinators tended to visit the
larger perfect flowers before visiting the smaller pistillate
flowers.
In S. noctiflora, the level of geitonogamy is likely to be low
because individual plants have few flowers open at one time
(S. L. Davis, personal observation). Furthermore, the low
level of fruit set in individually emasculated perfect flowers
indicates that pollinator-mediated pollen transfer (both outcrossing and geitonogamy) is very low (Schoen and Lloyd
1992). However, seed production in pistillate flowers is dependent on insect visitation and appears to be strongly pollen
limited; hand pollination of pistillate flowers increased their
fruit set nearly four times over that of unmanipulated flowers. The advantage of producing pistillate flowers therefore
depends on the level of pollinator activity and inbreeding depression. We have found significant family-level variation in
the level of inbreeding depression in S. noctiflora, with some
families exhibiting strong inbreeding depression (L. F. Delph
and S. L. Davis, unpublished data), indicating that the production of outcrossed seed would increase the fitness of the
plants. Pollinator activity is likely to be highly environmentally dependent and may change rapidly over the course of
the season. Such environmental and temporal variation may
favor individuals who are able to gauge their surroundings
and plastically alter their reproductive strategy (Tsitrone
et al. 2003). Folke and Delph (1997) found that S. noctiflora
individuals growing under higher temperatures produced a
significantly higher percentage of pistillate flowers than individuals growing in lower temperatures. This also corresponds
to conditions under which insects are likely to be more active. Hence, producing pistillate flowers under conditions of
higher temperature may also increase the probability that
they will be pollinated. This phenotypic plasticity in pistillate
flower production reduces the consequences to the plant of
any seed discounting within perfect flowers. It also corresponds with the prediction that the advantages of selfing for
reproductive assurance depend on the environment in which
plants live (Lloyd 1992; Kalisz and Vogler 2003).
Most empirical studies of mixed mating and the factors
that contribute to its maintenance have been carried out in
plant species that alter the levels of selfing and outcrossing
through altering levels of either dichogamy or herkogamy
within hermaphroditic flowers. Here we show that S. noctiflora enhances the possibility of mixed mating within plants
by producing two types of flowers. We conclude that by producing autonomously selfing perfect flowers when pollinators
are scarce and pistillate flowers when pollinator activity increases, S. noctiflora can preserve the reproductive assurance
480
INTERNATIONAL JOURNAL OF PLANT SCIENCES
advantage of autonomy and still maintain the potential benefit of producing outcrossed offspring.
Acknowledgments
Fieldwork was funded by a Faculty Development Grant to
S. L. Davis by the Howard Hughes Undergraduate Biologi-
cal Sciences Education Program at the University of Louisiana at Monroe and a fellowship to S. L. Davis from the
Mountain Lake Biological Station. This research was also
supported by a National Science Foundation grant BSR9010556 to L. F. Delph. We thank A. Caraway and J. Snow
for help counting seeds. We also thank C. Herlihy and two
anonymous reviewers for comments on an earlier draft.
Literature Cited
Baker HG 1955 Self-compatibility and establishment after ‘‘longdistance’’ dispersal. Evolution 9:347–348.
Barrett SCH, LD Harder, AC Worley 1996 The comparative biology
of pollination and mating in flowering plants. Philos Trans R Soc
Lond B 351:1271–1280.
Barrett SCH, MT Morgan, BC Husband 1989 The dissolution of
a complex genetic polymorphism: the evolution of self-fertilization
in tristylous Eichornia paniculata (Pontederiaceae). Evolution 43:
1398–1416.
Bittrich V 1993 Caryophyllaceae. Pages 206–236 in K Kubitzki, J
Rohwer, V Bittrich, eds. The families and genera of vascular plants.
Vol II. Springer, Berlin.
Brown BA, MT Clegg 1984 Influence of flower color polymorphism
on genetic transmission in a natural population of the common
morning glory, Ipomoea purpurea. Evolution 38:796–803.
Charlesworth D, MT Morgan, B Charlesworth 1990 Inbreeding
depression, genetic load, and the evolution of outcrossing rates in
a multilocus system with no linkage. Evolution 44:1469–1489.
Cheptou PO, J Lepart, J Escarre 2001 Differential outcrossing rates
in dispersing and non-dispersing achenes in the heterocarpic plant
Crepis sancta (Asteraceae). Evol Ecol 15:1–13.
Cheptou PO, DJ Schoen 2003 Frequency-dependent inbreeding depression in Amsinckia. Am Nat 162:744–753.
Collin CL, JA Shykoff 2003 Outcrossing rates in the gynodioeciousgynomonoecious species Dianthus sylvestris (Caryophyllaceae). Am
J Bot 90:579–585.
Darwin CR 1876 The effects of cross- and self-fertilization in the
vegetable kingdom. J. Murray, London.
Elle E, R Carney 2003 Reproductive assurance varies with flower size
in Collinsia parviflora (Scrophulariaceae). Am J Bot 90:888–896.
Elle E, JD Hare 2002 Environmentally induced variation in floral
traits affects the mating system in Datura wrightii. Funct Ecol 16:
79–88.
Fausto JA, Jr, VM Eckhart, MA Geber 2001 Reproductive assurance
and the evolutionary ecology of self-pollination in Clarkia xantiana
(Onagraceae). Am J Bot 88:1794–1800.
Fisher RA 1941 Average excess and average effect of a gene substitution. Ann Eugen 11:53–63.
Folke SH, LF Delph 1997 Environmental and physiological effects
on pistillate flower production in Silene noctiflora L. (Caryophyllaceae). Int J Plant Sci 158:501–509.
Gibson JP, AD Tomlinson 2002 Genetic diversity and mating system
comparisons between ray and disc achene seed pools of the
heterocarpic species Heterotheca subaxillaris (Asteraceae). Int J
Plant Sci 163:1025–1034.
Herlihy CR, CG Eckert 2002 Genetic cost of reproductive assurance
in a self-fertilizing plant. Nature 416:320–323.
Holsinger KE 1988 Inbreeding depression doesn’t matter: the genetic
basis of mating-system evolution. Evolution 42:1235–1244.
Jürgens A, T Witt, G Gottsberger 1996 Reproduction and pollination
in central European populations of Silene and Saponaria species.
Bot Acta 109:316–324.
Kalisz S, DW Vogler 2003 Benefits of autonomous selfing under
unpredictable pollinator environments. Ecology 84:2928–2942.
Kephart SR, E Brown, J Hall 1999 Inbreeding depression and partial
selfing: evolutionary implications of mixed-mating in a coastal
endemic, Silene douglasii var. oraria (Caryophyllaceae). Heredity
82:543–554.
Lande R, DW Schemske 1985 The evolution of self-fertilization and
inbreeding depression in plants. I. Genetic models. Evolution 39:
24–40.
Lande R, DW Schemske, ST Schultz 1994 High inbreeding depression, selective interference among loci, and the threshold selfing
rate for purging deleterious mutations. Evolution 48:965–978.
Lloyd DG 1992 Self- and cross-fertilization in plants. II. The
selection of self-fertilization. Int J Plant Sci 153:370–380.
Lloyd DG, DJ Schoen 1992 Self- and cross-fertilization in plants. I.
Functional dimensions. Int J Plant Sci 153:358–369.
Lu Y 2000 Why is cleistogamy a selected reproductive strategy in
Impatiens capensis (Balsaminaceae)? Biol J Linn Soc 75:543–553.
Ramsey M, G Vaughton 1996 Inbreeding depression and pollinator
availability in a partially self-fertile perennial herb Blandfordia
grandiflora (Liliaceae). Oikos 76:465–474.
Rausher MD, SM Chang 1999 Stabilization of mixed-mating systems by differences in the magnitude of inbreeding depression for
male and female fitness components. Am Nat 154:242–248.
Schemske DW 1978 Evolution of reproductive characteristics in
Impatiens (Balsaminaceae): the significance of cleistogamy and
chasmogamy. Ecology 59:596–613.
Schoen DJ, AHD Brown 1991 Whole- and part-flower self-pollination
in Glycine clandestina and G. argyrea and the evolution of autogamy.
Evolution 45:1665–1674.
Schoen D, DG Lloyd 1992 Self- and cross-fertilization in plants. III.
Methods for studying modes and functional aspects of selffertilization. Int J Plant Sci 153:381–393.
Takebayashi N 2000 Self-fertilization, genetic associations, and
genetic architecture: understanding the components of matingsystem evolution. PhD diss. Indiana University, Bloomington.
Tsitrone A, S Duperron, P David 2003 Delayed selfing as an optimal
mating strategy in preferentially outcrossing species: theoretical
analysis of the optimal age of first reproduction in relation to mate
availability. Am Nat 162:318–331.
Uyenoyama M, KE Holsinger, DM Waller 1993 Ecological and
genetic factors directing the evolution of self-fertilization. Oxford
Surv Evol Biol 9:327–381.
Uyenoyama M, DM Waller 1991 Coevolution of self-fertilization
and inbreeding depression. II. Symmetric overdominance in
viability. Theor Popul Biol 40:47–77.
Vogler DW, S Kalisz 2001 Sex among the flowers: the distribution of
plant mating systems. Evolution 55:202–204.