Table Of ContentMadrono, Vol. 58, No. 2, pp. 78-85, 2011
POLLEN SIRING SUCCESS IN THE CALIFORNIA WILDFLOWER
CLARKIA UNGUICULATA (ONAGRACEAE)
Nancy L. Smith-Huerta
Department of Botany, Miami University, Oxford, OH 45056
[email protected]
Frank C. Vasek12
Department of Botany and Plant Sciences, University of California, Riverside, CA 92521
Abstract
Cryptic self-incompatibility (CSI) is a type of non-random mating observed in self-compatible
plantsinwhichoutcrosspollensiresproportionallymoreseedsthanselfpollenwhenbothpollentypes
areavailableonthe stigma. LevelsofCSI areknownto varyamongindividuals and populations. We
conducted competitive pollinations consisting ofmixtures ofselfand outcross pollen to investigate
reportsofCSIincertainpopulationsofClarkiaunguiculataLindl. Wealsoinvestigatedhowtheorder
of self and outcross pollen deposition on the stigma influences the degree of nonrandom mating.
Finally,welooked atwhetherthe sourcepopulationofoutcrosspollen affectedtheabilityofoutcross
pollentooutcompeteselfpollen.Weutilizedrecessive(white-petaled)maternal(andselfpollendonor)
plantsfromalocalitynearMorro Bay, California, anddominant(pink-petaled)outcrosspollendonor
plants from 17 localities widespread through the species range in California. Progenies from
pollinations made with equal mixtures ofselfand cross pollen included significantly more outcross-
pollinated than self-pollinated offspring in 10 ofthe 17 cross-pollen donor populations. However six
populations showedno significant differencebetween self- and outcross-pollinated offspring, and one
population yielded a majority ofprogeny sired by selfpollen. Progenies from sequential selfpollen
followed by outcross pollinations included significantly more self offspring in 12 of the donor
populations, no significant difference between outcross and selfoffspring in four donor populations
andsignificantlymoreoutcrossoffspringinonedonorpopulation. Progeniesfromsequentialoutcross
pollen followed by self pollinations included significantly more outcross offspring in 15 donor
populations and no significant difference between outcross and self offspring in two donor
populations. Our results confirm the occurrence of non-random mating in C. unguiculata, and
demonstratethatthedegreeofnon-randommatingcandependontheorderofselfvs. outcrosspollen
deposition and the source population ofoutcross pollen. This non-random mating can influence the
proportion ofselfand outcross progeny in sequential pollinations.
Key Words: Clarkia unguiculata, cryptic self-incompatibility, geitonogamy, non-random mating,
sexual conflict.
Cryptic self-incompatibility is essentially a case (Marshall 1998). This fact suggests that the
of non-random mating in which normal seed set source of outcross pollen is likely to affect the
occurs upon self pollination in the absence of degree ofnon-random mating due to CSI.
competitive outcross pollen. However, with Self-incompatibility may evolve in outcrossing
mixtures of self and outcross pollen, weak self populations in response to inbreeding depression.
rejection reactions promote preferential fertiliza- If a normally outcrossing population carries a
tion by outcross pollen (Bateman 1956; Weller genetic load, self-incompatibility could confer a
and Ornduff 1977; Eckert and Allen 1997; fitness advantage to maternal plants if they
Kruszewski and Galloway 2006). produce more outcross progeny than selfed
Nonrandom mating in plants is not limited to progeny in mixed pollinations. In contrast to
examples of CSI but also occurs in mixtures of “complete” self-incompatibility, plants exhibiting
outcross pollen from several donors, as in CSI maintain the ability to produce progeny
Raphanus sativus L. (Marshall and Ellstrand through self-fertilization. This ability should
1986; Marshall 1991, 1998). Ofparticular interest confer the fitness advantage of reproductive
in these cases is the observation that pollen assurance to annual plants that grow where
donors may differ significantly in their ability to access to pollinators or mates is limited. During
sire seed, and the rank order oftheir siring ability growing seasons when pollinators and other
can be consistent across several maternal plants resources are plentiful, non-random mating by
favoring outcross pollen allows maternal plants
to increase fitness by increasing the proportion of
1Retired. outcross progeny they produce. During less
2Present address: 3418 Mono PI., Davis, CA 95618. optimal growing seasons, when resources and
2011] SMITH-HUERTA AND VASEK: POLLEN SIRING SUCCESS IN CLARKIA 79
pollinators are limiting, maternal fitness is by self vs. cross pollen. Test plants were
enhanced through self-pollination. genetically marked for petal color. Flower color
Another important consideration is the role in C. unguiculata varies from white to red or
that non-random mating by favoring outcross purple but most populations are characterized by
pollen might play in promoting outcross pollina- pink or lavender-pink flowers. White flowers are
tion in plants with varied patterns of pollen conditioned by a single recessive allele and pink
deposition. The number ofpollen grains deposit- or red flowers are conditioned by a dominant
edonthestigmainnaturalpopulationsofClarkia allele at the same locus (Rasmuson 1920; Vasek
unguiculata Lindl. increases with time (Nemeth 1965, plus extensive unpublished data; Bowman
and Smith-Huerta 2003), indicating that pollina- 1984). Plants with pink flowers contain malvidin
tors make multiple visits to flowers, depositing 3, 5-diglucoside in theirflowers, leaves, stems and
loadsofpollensequentiallyovertime. Theseloads seedlings (Bowman 1987) whereas plants with
of pollen are likely to vary in composition from white flowers lack this and other anthocyanins.
mixtures ofoutcross pollen toloads ofselfpollen. Consequently, flower color can be determined in
Geitonogamous self pollination is possible since the seedling stage merely by scoring seedlings as
C. unguiculata is strongly protandrous (Vasek green (non-anthocyanous) or red (anthocyanous)
A
1968, 1977; Vasek et al. 1987). pollinatorcould becausegreen seedlings growup to producewhite
visit a flower in the male phase and then visit flowers and red seedlings grow up to produce
another flower on the same plant in the female pink flowers (Bowman 1987; Vasek 1965).
phase resulting in self pollination. It is possible
that CSI could promote outcross pollination in
Organization ofExperiments
cases where the stigma ofa flower first receives a
loadofselfpollenandsoonafterreceivesaloadof Ten white-flowered families were developed
outcross pollen. In fact, although geitonogamy is from available stocks from a natural population
possible (and probable) in C. unguiculata, out- near Morro Bay, California because they were
crossing rates in a natural population have been
measuredatgreaterthan90% (Vasek 1965). Non- available and we knew they were fully self-fertile
randommatingbyfavoring outcross pollen (CSI) (Vasek 1986). Each family consisted of approx-
maywell have contributed to these high outcross- imately6—8 siblings (homozygous forwhitepetal)
ingrates. In fact, CSIisknownto occurinatleast andwas used asa line ofmaternaltestplants, and
one population ofC. unguiculata (Bowman 1987) simultaneously as self-pollen donors.
andinonepopulationofC. gracilisA. Nelsonand Seventeen pink-flowered families were devel-
J. F.Macbr.(Jones 1994). Interestingly, CSI was oped from available stocks from Morro Bay and
reported to be completely absent from another 16 widespread localities in California (Table 1).
population of C. unguiculata (Travers and Mazer Each family consisted of about 6-8 siblings and
2000). This suggests that the strength and nature was used as a line ofcross pollen donors. These
and of non-random mating could vary from plantswere homozygous forpink flowers because
populationtopopulationandalsowiththesource they were grown from selected, available stocks
ofoutcross pollen. known to have produced only pink-flowered
The purpose of our investigation was to progeny.
examine a single population of C. unguiculata Seeds from the source localities were sown on
maternal plants to determine whether the order vermiculite in December, 1987 in a University of
that self and outcross pollen is deposited on the California greenhouse in Riverside, California.
stigma affects non-random mating. Furthermore, Subsequently, seedlings were transplanted to 6
UC
we examined ifthe source population ofoutcross inch pots of standard soil mix, irrigated as
pollen affects levels ofCSI. This closer examina- needed, fertilized weekly with half strength
tion of non-random mating in C. unguiculata is Hoagland’s solution and grown to maturity in
important given the possibility of sequential and the same greenhouse. Experimental pollinations
geitonogamous pollination in C. unguiculata were made during a period ofabout two months
(Vasek 1968, 1977; Vasek et al. 1987; Nemeth in the spring of 1988.
and Smith-Huerta 2003), the variation in levels of
CSI documented previously in two different Mixed-Donor Pollinations
populations of C. unguiculata (Bowman 1987;
Travers and Mazer 2000), and that pollen donors All experimental pollinations utilized pollen
maydiffer significantly in their ability to sire seed from a pink-flowered cross-pollen donor and
(Marshall 1998). pollen from a white-flowered self pollen donor.
The self pollen donor was also the seed parent.
Materialsand Methods Pollen was always applied to stigmas which were
1-2 days old as judged by the degree of stigma
Mixed-donor pollinations were employed to expansion not greaterthan 180degrees (seefigure
determine the relative frequency of fertilization 2 in Smith-Huerta and Vasek 1984).
MADRONO
80 [Vol. 58
Tdiarbeclteion1.frSoemeMdorCroollBeacyt:iSon=LCoocaasltiRtainegse(sAtloltihensCoaulthi;foNrn=iaC)oafsotrRaCnrgoesss-tPootlhelennortDho;nEor= SPilearnrtasN.evDaIdRat=o
the East; DIS = approximate linear distance in km from Morro Bay.
Population Locality County DIR DIS
—
15 Morro Bay - Atascadero Rd San Luis Obispo 00
4 Santa Maria Santa Barbara S 58
5 Pinnacles Natl. Monument San Benito N 126
11 Rancheria Rd Kem E 185
6 Kern River Kern E 188
13 Caliente Hills, Low Canyon, Kern E 189
17 Caliente Hills, High Canyon Kern E 190
8 Santa Paula Ventura S 193
7 Sequoia Tulare E 220
9 Bouquet Canyon Los Angeles S 238
14 Laurel Canyon Los Angeles S 260
2 Jackson Amador E 330
3 Riverside -Fairmount Park Riverside S 350
10 San Luis Rey San Diego S 398
16 Old Castle Rd San Diego s 406
N
1 Clear Lake Lake 430
12 Feather River, North Fork Butte E 485
Oneprotocol (MIX)followed themethodology they were grown to maturity for direct scoring of
used by Bowman (1987). Cross pollen and self flower color.
pollen were mixed togetherin equal amounts and
then applied to the stigma of the maternal test Analysis
plant, whichwasalso the self-pollen donor. In the
second protocol (S/X), self pollen was applied We expect the progeny from competitive
first followed immediately by cross pollen. In the (MIX) pollinations to include halfpink-flowered
third protocol (X/S), cross pollen was applied outcrosses if mating is non-random. However,
first, followed immediately by selfpollen. preferential functioning of either cross or self
We pollinated 246 flowers in 82 competitive pollen will yield ratios of white to pink progeny
pollinations. Each pollination included one pol- significantly different from 1:1. The frequency of
lination by each of the three protocols described outcrosses in each resulting progeny was multi-
above. Several siblings ofeach cross pollen donor plied by 100 and expressed as a percent. The
population were used, for a total of 47 pollen progeny of (S/X) pollinations should include
donors from the 17 cross-pollen donor lines. significantly more selfpollinated white than pink
Thus, the actual competitive comparisons in- progeny ifmating is random. We expect the first
volved closely related siblings rather than identi- pollen grains on a receptive stigma should
cal individuals. interfere with the normal germination of pollen
Following competitive pollinations, the result- arriving later. The progeny of (X/S) pollinations
ingcapsules were harvested at maturity, the seeds should include significantly more pink than white
werethen separated, counted, weighed and stored progeny.
over the summer. Each ofthe 246 progenies from MIX, S/X and
X/S pollinations was tested for significant depar-
Data Harvest ture from an expected 1:1 rAatio of outcrosses to
seifs byaG-test (Zar 1984). BonferroniP-value
Beginningin October 1988, seedswere sown on correction was applied to avoid inflated type-I
vermiculite in small plastic pots and put in a error (avoid high risk of false positive results)
temperature controlled chamber. Temperatures across the multiple tests (Zar 1984). Progenies
were approximately 20 degrees C with a 12 h day from all ofthecompetitivepollinationswithin the
12 h night lighting schedule. Two weeks after 17 populations were also tested for significant
germination, seedlings were moved to a green- departure from an expected 1:1 ratio ofoutcross
house for continued development because space to seifs by a G-test (Zar 1984).
in the temperature controlled chamber was
limited. Red pigments in stems and along leaf Results
veins develop well with cool temperatures and
bright daylight. Our greenhouse conditions were Both the order ofpollen deposition and source
not always optimum for red pigment develop- of outcross pollen had variable affects on the
ment (e.g., warm weather). Consequently, if any degree ofnon-random mating due to CSI in our
seedlings were not clearly red or not clearly green competitive pollinations (Table 2, Fig. 1). In the
2011] SMITH-HUERTA AND VASEK: POLLEN SIRING SUCCESS IN CLARKIA 81
Table 2. G-Test Values for all Progeny within single populations of plants (Jones 1994;
Produced in Competitive Pollinations. The Travers and Mazer 2000) or used a commercial
symbols “*,” “***,” and “ns” indicate seed source (Bowman 1987). Our study included
significant difference from a 1:1 ratio of self to maternal plants derived from one population and
outcross progeny at the 5%, 1%, and 0.1% levels, and outcross pollen plants derived from 17 popula-
not significantly different, respectively.
tions distributed throughout California. In the
Population MIX S/X X/S present study, competitivepollinationsmadewith
pollen from 10 of the 17 populations yielded
68.219*** 48.753*** 61.926***
21 16.934*** 0.134 ns 9.150** results similar to those found by Bowman (1987)
3 34.464*** 1.590 ns 51.451*** and Jones (1994) with a majority of pollinations
4 17.357*** 0.0055 ns 71.653*** favoring outcross pollen. Competitive pollina-
5 32.142*** 19.534*** 55.477*** tions with outcross pollen derived from six ofthe
6 12.659*** 29.500*** 46.019*** populations yielded results similar to those found
7 9.419** 11.027*** 16.933*** by Travers and Mazer (2000), with outcross
8 14.710*** 37.016*** 42.659*** pollen favored in fewer than halfthe pollinations.
9 7.905** 80.327*** 2.987 ns In contrast to all of the previous studies,
1110 71..269676*n*s 2192.315642****** 3104..307029****** competitive pollinations made with outcross
12 2.337 ns 7.703** 24.856*** pollen derived from 1 of our study populations
13 0.032 ns 26.640*** 23.774*** yielded a majority of progenies in which self
14 0.643 ns 9.550** 25.641*** pollen was favored significantly over outcross
15 0.0303 ns 144.658*** 3.412 ns pollen.
16 0.024 ns 0.468 ns 13.832*** Non-random mating when outcross pollen is
17 11.512*** 106.878*** 9.662*** favored over self pollen has the potential to
reduce the negative effects of inbreeding depres-
MIX competitive pollinations, 10 ofthe 17 cross sion in populations through the increased pro-
pollen donor populations exhibited CSI, yielding duction of outcross progeny, at the same time
progenies that significantly favored outcross preserving the ability of individuals to produce
pollen (Table 2, Fig. 1). However six of the 17 offspring by selfing. Inbreeding depression has
pollen donor populations did not exhibit CSI, been documented in several populations of C.
and onepopulationyielded amajority ofprogeny tembloriensis Vasek (Holtsford and Ellstrand
sired by selfpollen (Table 2, Fig. 1). 1990) and the magnitude of this inbreeding
In the S/X competitive pollinations, one ofthe depression was shown to vary between popula-
17 cross pollen donor populations yielded prog- tions. It is possible that populations of C.
enies that significantly favored outcross pollen, unguiculata could vary with respect to genetic
12pollen donor populations significantly favored load, and selective pressure promoting non-
selfpollen, and four ofthe donor populations did random mating favoring self pollen might vary
not differ significantly from a 1:1 ratio (Table 2, between populations. This could account for the
Fig. 1). variation observed in non-random mating in
Fifteen ofthe pollen donor populations in X/S previous studies (Bowman 1987; Jones 1994;
competitive pollinations significantly favored Travers and Mazer 2000).
outcross pollen, and two donor populations did As stated above, non-random mating by
not differ significantly from a 1:1 ratio (Table 2, favoring outcross pollen not only promotes the
Fig. 1). production of outcross progeny, but also pre-
serves the ability ofa plant to produce offspring
Discussion through selfpollination. This provides reproduc-
tive assurance to annual plants that grow where
Overall, the results of our MIX pollinations access to mates may be limiting. Numbers of
confirm the occurrence of CSI and non-random individuals in populations of C. unguiculata can
mating in Clarkia unguiculata. More than halfof vary from only a few plants to thousands of
the competitive MIX pollinations yielded proge- individuals (Lewis and Lewis 1955; Vasek 1964)
nies that significantly favored outcross pollen, and those growing in more marginal areas ofthe
and approximately 20% of competitive pollina- species range may experience large seasonal
tions yielded progenies with significantly more fluctuations in population size (Lewis and Lewis
offspring produced by self pollen. Only about 1955). During seasons when plant populations
23% ofcompetitivepollinationsyieldedprogenies are small, pollinators are rare, and access to
that did not differ from a 1:1 ratio. Our mates limited, fitness might be enhanced by the
investigation differs in detail and scope from production of selfprogeny.
previous studies ofCSI and non-random mating Our sequential pollinations address the ques-
in Clarkia (Bowman 1987; Jones 1994; Travers tion ofwhether non-random mating by favoring
and Mazer 2000), and may help to explain their outcross pollen can somehow mitigate the effects
variable results. Previous studies were conducted of geitonogamy when self pollen arrives first on
MADRONO
82 [Vol. 58
Outcross Progenysig. <50%
Outcross Progenynotsig. diff. from 50%
Outcross Progenysig. >50%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Outcross Pollen Source Population
Fig. 1. Percent oftotal competitive pollinations within each of 17 cross pollen donor populations that yielded
outcross progenies either significantly more or significantly less than expected by chance, or yielded outcross
progeny not significantly different from chance expectations. For the MIX treatment, cross and selfpollen were
mixedtogetherinequalamountsandthenappliedtothe stigmaofthematernaltestplant, whichwas also the self-
pollendonor. ForS/X, selfpollenwasappliedfirstfollowedimmediatelybycrosspollen. ForX/S,crosspollenwas
applied first, followed immediately by selfpollen.
the stigma followed by outcross pollen. Rates of nation is likely. Geitonogamy does not provide
geitonogamy are influenced by the number of reproductive assurance and is considered to be
flowers that are open simultaneously on a plant “an unavoidable by-product of selection for
(Karron et al. 2004). Clarkia unguiculata is outcrossing success” in plants that have numer-
strongly protandrous (anthers shed pollen before ous flowers that are open simultaneously in an
the stigma becomes receptive), and may have as inflorescence (Goodwillie et al. 2005). It is
many as 12 open flowers per spike (Vasek 1968, possible that non-random mating by favoring
1977; Vasek et al. 1987). Anthers of individual outcross pollen could mitigate the unavoidable
flowers shed pollen for up to 11 days before the selfing that occurs when pollinators visit more
stigma becomes receptive (Vasek 1968, 1977; that one flower per plant in C. unguiculata, and
Vasek et al. 1987). Ifpollinators visit more than our S/X sequential pollinations were designed to
one flower per plant, geitonogamous self polli- explore this idea. If mating is random, S/X
2011] SMITH-HUERTA AND VASEK: POLLEN SIRING SUCCESS IN CLARKIA 83
pollinations (outcross pollen grains deposited outcross pollen in single donor pollinations in
immediately after selfpollen grains) should yield these two Clarkia species (Smith-Huerta 1996;
progenies with significantly more self than Kerwin and Smith-Huerta 2000; Nemeth and
outcross offspring. Twelve of the donor popula- Smith-Huerta 2002). It is possible that self and
tions did yield progenies that significantly fa- outcross pollen must be present together on the
vored self pollen, even though six of these stigma and in the style for the non-random
populations displayed non-random mating in mating of CSI to occur. This appears to be the
MIXpollinations. In thesecases, the documented case in Clarkia. Pollen germination was signifi-
presence of non-random mating did not appear cantly reduced in two donor pollinations ofself+
to mitigate the effects of geitonogamy. In outcross pollen and in outcross + outcross pollen
contrast, S/X sequential pollinations in five of (from two different donors) in C. unguiculata
the donorpopulations did not favor selfprogeny. (Nemeth and Smith-Huerta 2002). Further,
Of these, four of the populations produced germination of pollen decreased with increasing
progeny that did not differ from a 1:1 ratio of contact between pollen grains (Nemeth and
selfto outcross, and one population significantly Smith-Huerta 2002). Pollen-pollen interactions,
favored outcross pollen. With one exception mediated by the stigma, might provide a possible
(population 16), these were the donor popula- mechanism to explain the differences in the
tions that displayed the strongest non-random relative success of self and outcross pollen
mating by favoring outcross pollen in MIX observed in reported non-random mating and
pollinations. It appears that strong non-random CSI in Clarkia.
mating by favoring outcross pollen has the The present investigation goes beyond an
potential to mitigate the effects of geitonogamy examination of self vs. outcross pollen perfor-
when self pollen arrives first on the stigma mance within a single population, and examines
followed by outcross pollen. the performance ofoutcross pollen derived from
Our X/S pollinations explored further the foreign populations. These outcross pollen source
interactions ofnon-random mating and geitono- populations occur from 58 to 485 km from the
gamy in Clarkia. The X/S pollinations were maternal (self pollen) population. Clearly, the
expected to produce significantly more outcross maternal plants in our study would not normally
than self progeny in all pollinations, since the encounter outcross pollen from these popula-
advantage ofoutcross pollen arriving first on the tions. Interestingly, the Morro Bay population,
stigma should be enhanced by non-random which provided all of our maternal plants, did
mating. This proved to be the case in all but not show CSI in MIX crosses when Morro Bay
two of the donor populations which showed no plants were the source of outcross pollen. In
significant difference between the number of self contrast, levels of CSI could be very high when
and outcross progeny produced. Interestingly, pollen from foreign populations was used. This
these two donor populations were the only great difference in outcross pollen donor success
populations in which selfpollen was significantly suggests that pollen-pistil interaction may evolve
favored in all oftheir sequential S/X pollinations. differently in each population of C. unguiculata,
The physical mechanism responsible for differ- as a result of sexual conflict between male and
ential siring success in non-random mating and female function. In plants, sexual conflict occurs
CSI has been investigated in other studies when optimal reproductive fitness strategies for
(Hessing 1989; Weller and Ornduff 1989; Aizen pollen differ from those of the maternal plant.
et al. 1990; Eckert and Allen 1997; Kruszewski The trait ofCSI has the potential to enhance the
and Galloway 2006; Figueroa-Castro and Holts- fitness ofovules but not ofpollen, thus creating a
ford2009). In all ofthe above investigations, with male-female sexual conflict. Although sexual
one exception, outcross pollen germinated faster conflict has been studied mostly in animals,
on the stigma and grew faster through the style several recent studies have examined the occur-
than selfpollen. The single exception occurred in rence of sexual conflict in plants (Prasad and
Campanulastrum americanum Small in which Bedhomme 2006; Lankinen and Larsson 2009;
pollen tube growth rates did not differ between Madjidian and Lankinen 2009). In one instance,
self and outcross pollen (Kruszewski and Gallo- similar to the present study, male-female interac-
way 2006). Although we did not measure pollen tions were investigated in cross pollinations
germination and tube growth rates in the present between plants derived from 4 different popula-
study, this was the focus of previous studies in tions of Collinsia heterophylla Graham (Madji-
both C. unguiculata (Nemeth and Smith-Huerta dian and Lankinen 2009). In this plant, the onset
2002; Smith-Huerta et al. 2007) and C. temblor- ofstigma receptivity may be affected by both the
iensis (Smith-Huerta 1996; Kerwin and Smith- source and recipient of the pollen, with early
Huerta 2000). Similar to Campanulastrum amer- receptivity and fertilization resulting in the
icanum (Kruszewski and Galloway 2006), no production of fewer seeds than late receptivity.
differenceinpercentgermination orrateofpollen In experimental cross pollinations, it was found
tube growth was observed between self and that pollen donors from foreign populationswere
MADRONO
84 [Vol. 58
more successful at inducing stigma receptivity longiflora and N. plumbaginifolia pollen tube
:
than pollen donors derived from the same growth rate, offspring paternity and hybridization.
population (Madjidian and Lankinen 2009). Sexual Plant Reproduction 22:187-196.
These results may be interpreted to suggest that Goodwillie, C., S. Kalisz, andC. G. Eckert. 2005.
there is “sexually antagonistic coevolution” The evolutionary enigma ofmixed mating systems
between maternal plants and pollen within in plants: occurrence, theoretical explanations, and
empirical evidence. Annual Review of Ecology,
populations of C. heterophylla (Madjidian and Evolution, and Systematics 36:47-79.
Lankinen 2009). It appears that when maternal Hessing, M. B. 1989. Differential pollen tube success
plants receive pollen from foreign populations in Geranium caespitosum. Botanical Gazzette
theyare somehow “released fromthecost oflocal 150:404^110.
pollen” (Madjidian and Lankinen 2009). In the Holtsford, T. P. and N. C. Ellstrand. 1990.
present study, CSI and non-random mating Inbreeding effects in Clarkia tembloriensis (Ona-
varied extensively depending on the population graceae) populations with different natural out-
source of outcross pollen. It is possible that this crossing rates. Evolution 44:2031-2046.
reflects a similar “release from the cost of local Jones, K. N. 1994. Nonrandom mating in Clarkia
gracilis (Onagraceae): a case ofcryptic self-incom-
pollen” in our experiments.
patibility. AmericanJournal ofBotany 81:195-198.
raInndosmum,mawteindgocaunmdentCStIheinocccurrorsesnesceboeftwneoenn- KarrJ.onM,. BJ.elDl.,, aR.ndJ.BM.iFtucnhke.ll2,00K4.. TGh.eHionlfmluqeunicseto,f
different populations of Clarkia unguiculata, floral display size on selfing rates in Mimulus
and demonstrate that this non-random mating ringens. Heredity 92:242-248.
can also influence the proportion of self and Kerwin, M. A. and N. L. Smith-Huerta. 2000.
outcross progeny in sequential pollinations. The Pollen and pistil effects on pollen germination and
non-random mating observed in CSI may pro- tube growth in selfing and outcrossingpopulations
mote outcrossinginprotandrousplants subject to of Clarkia tembloriensis. International Journal of
geitonogamous pollination and contribute to Plant Science 161:895-902.
reproductive assurance when access to mates is KrusEzxepwlasiknii,ngL.outJc.roasnsidngL.ratFe. iGnalCalmopwaanyu.las2t0r0u6.m
limited. Finally, the fact that levels of CSI vary americanum (Campanulaceae): geitonogamy and
between sources of pollen donors suggests that cryptic self-incompatibility. International Journal
sexual conflict between pollen and maternal ofPlant Sciences 167:455-461.
plants may result in coevolution unique to each Lankinen, A. and M. C. Larsson. 2009. Conflicting
population of C. unguiculata. selection pressures on reproductive functions and
speciation in plants. Evolutionary Ecology 23:
Acknowledgments 147-157.
We are grateful to the University of California LewiUsn,ivHe.rsaintyd Mof. CEa.liLfeowrinsi.a 1P9u5b5l.icTahteiognesnuisn CBloartkainay.
Agricultural Experiment Station for support, to Mi-
20:241-392.
VchianeclenHtuWgehensgffoorrhhieslptewchintihcatlheasssitasttiasntciec,alaanndaltyosiCsa,rltao Madjidian, J. A. and A. Lankinen. 2009. Sexual
tSsumwgiogtehsatfniooornnasyssmfioosrutasinmceprrewovivitehewmseeernestdcfoooflrletchttehieominar.nWuvseecrrayilpstoh.etlhTpahfniuksl Marscaohnnnafullailcl,tpalDna.ndt.sLe.PxuL1ao9l9Sl1y.OanNNtoaEngor4n:aiesn5td4i7oc7m.comeavotliuntgioinniwnialnd
work is dedicated to the memory of our mentor roafdimsuhl:tviaprlieatpiaotneirnniptoylleanmdoonngorosnuec-cestsoasnidx-efdfoencotrs
Professor Harlan Lewis.
pollinations. American Journal of Botany
Literature Cited 7—8:1404-1418.
1998. Pollen donor performance can be
.
Aizen, M. A., K. B. Searcy, and D. L. Mulcahy. consistent across maternal plants in wild radish
1990. Among- and within-flower comparisons of (Raphanus sativus, Brassicaceae): a necessary
pollen tube growth following self- and cross- condition for the action of sexual selection.
pollinations in Dianthus chinensis (Caryophylla- American Journal ofBotany 85:1389-1397.
ceae). American Journal ofBotany 77:671-676. — and N. C. Ellstrand. 1986. Sexual selection
Bateman, A. J. 1956. Cryptic self-incompatibility in in Raphanus sativus: experimental data on the
the wallflower. Heredity 10:257-261. nonrandom fertilization, maternal choice, and
Bowman, R. N. 1984. Experimental non-stigmatic consequences of multiple paternity. American
pollinations in Clarkia unguiculata (Onagraceae). Naturalist 131:736-759.
A—merican Journal ofBotany 71:1338-1346. Nemeth, M. B. and N. L. Smith-Huerta. 2002.
1987. Cryptic self-incompatibility and the Effects ofpollen load composition and deposition
.
breeding system of Clarkia unguiculata (Onagra- pattern on pollen performance in Clarkia unguicu-
ceae). American Journal ofBotany 74:471^-76. lata. International Journal of Plant Science
Eckert, C. G. and M. Allen. 1997. Cryptic self- 163:795-802.
incompatibility in tristylous Decodon verticillatus AND . 2003. Pollen deposition, pollen
(Lythraceae). American Journal of Botany 84: tube growth, seedproduction, and seedling perfor-
1391-1397. mance in natural populations of Clarkia unguicu-
Figueroa-Castro, D. M. and T. P. Holtzford. lata (Onagraceae). International Journal of Plant
2009. Post-pollination mechanisms in Nicotiana Science 164:153-164.
2011] SMITH-HUERTA AND VASEK: POLLEN SIRING SUCCESS IN CLARKIA 85
Prasad, N. G. and S. Bedhomme. 2006. Sexual . 1965. Outcrossing in natural populations.
conflictsin plants. Journal ofGenetics 85:161-164. II. Clarkia unquiculata. Evolution 19:152-
Rasmuson, H. 1920. Die Hauptergebnisse von einigen —156.
genetischen Versuchen mit verschiedenen von 1968. The relationships of two ecologically
.
Tropaeolum, Clarkia und Impatiens. Hereditas marginal, sympatric Clarkia populations. Ameri-
1:270-276. c—an Naturalist 102:25-40.
Smitthu-beHugerrotwat,h iNn.seLl.fin1g99a6n.dPooultlcernosgseirnmgipnoaptuiloantiaonnds Cla.rk1i9a77.sePchteionnotyPphiaceovsatroimatai.onSaysntdemaadtaipctatBiootnaniyn
ofClarkia tembloriensis (Onagraceae). Internation- 2—:251-279.
al JaounrdnalF.ofC.PlVaansteSkc.ie1n9c8e4.15P7o:l2l2e8n-2l3o3n.gevity and Sel.ec1t9i8o6n. Efovroluhtaiirolneasrsnyesmso.difAimceartiiconaninJColuarrnkaila.oIf.
stigma pre-emption in Clarkia. American Journal B—otany 73:517-523.
of,BoS.taRn.y C71a:r1r1i8n3-o1-1K9y1.ker, and A. J. Huerta. 198,7.V.EfWfeecntgs,ofR.miJn.erBaelavneurtr,itaionndoCn.cKo.mpHounzeanrt.s
2007. Theeffects ofmaternal andpaternalnutrient of reproduction in Clarkia unguiculata. Aliso
status on pollen performance in the wildflower 11:599-618.
Clarkia unguiculata (Onagraceae). Journal of the Weller, S. G. and R. Ornduff. 1977. Cryptic self-
Torrey Botanical Society 134:451-457.
Travers, S. E. and S. J. Mazer. 2000. Theabsenceof incompatibilityinAmsinckiagrandiflora. Evolution
31:47-51.
c(rOynpaigcrasceelaf-ei)n.comApamteirbiilciatyn inJoCulranraklia oufnguBiocutlaantya —AND . 1989. IncompatibilityinAmsinckia
grandiflora (Boraginaceae). American Journal of
87:191-196.
Vasek, F. C. 1964. TheevolutionofClarkiaderivatives Botany 78:801-804.
adaptedtorelativelyxericenvironments. Evolution Zar, J. H. 1984. Biostatistical analysis, 2nd ed.
18:26-12. Prentice-Hall, Inc., Englewood Cliffs, NJ.
