Reproductive alternatives to insect pollination in four species of Thalictrum (Ranunculaceae)

Abstract

Although insect pollination is ancestral in the angiosperms, selection has favored wind pollination and self-fertilization in many lineages. The evolution of clonal growth may have also decreased dependence on insect pollinators. We investigate transitions away from insect pollination in the genus Thalictrum (Ranunculaceae) among four species that vary in pollination syndrome. Thalictrum sparsiflorum is insect pollinated and self-fertilization may also provide reproductive assurance. Thalictrum alpinum is clonal, has a floral morphology associated with wind pollination and is self-compatible. Thalictrum fendleri is a wind-pollinated and dioecious species that sets few fruits naturally and invests little in sexual reproduction, possibly due to a trade-off with clonal growth. Thalictrum dioicum is also wind-pollinated and dioecious, but does not grow clonally at our sites and has a higher investment in sexual reproduction than T. fendleri. The pollen : ovule ratio is largest in wind-pollinated species, but varies considerably among them, possibly reflecting differences in pollination efficiency and/or incidental insect pollination. None of the species appear pollen limited in the study populations, suggesting that factors other than pollen receipt limit female fertility at these sites.

Full text

Plant Species Biology (2004) 19, 73– 80
© 2004 The Society for the Study of Species Biology
Blackwell Science, LtdOxford, UKPSBPlant Species Biology0913-557XThe Society for the Study of Species Biology, 2004August 20041927380Original Article
REPRODUC-
TION IN THALICTRUMJ. C. STEVEN and D. M. WALLER
Correspondence: Janet C. Steven
Email: jancstev@indiana.edu
1
Present address: Department of Biology, 1001 E. Third St., Jordan
Hall, Room. 142, Indiana University, Bloomington IN 47405 USA.
Reproductive alternatives to insect pollination in four
species of
Thalictrum
(Ranunculaceae)
JANET C. STEVEN
1
and DONALD M. WALLER
Department of Botany, University of Wisconsin Madison, Madison, WI 53706 USA.
Abstract
Although insect pollination is ancestral in the angiosperms, selection has favored wind
pollination and self-fertilization in many lineages. The evolution of clonal growth may
have also decreased dependence on insect pollinators. We investigate transitions away
from insect pollination in the genus
Thalictrum
(Ranunculaceae) among four species that
vary in pollination syndrome.
Thalictrum sparsiflorum
is insect pollinated and self-fer-
tilization may also provide reproductive assurance.
Thalictrum alpinum
is clonal, has a
floral morphology associated with wind pollination and is self-compatible.
Thalictrum
fendleri
is a wind-pollinated and dioecious species that sets few fruits naturally and
invests little in sexual reproduction, possibly due to a trade-off with clonal growth.
Thalictrum dioicum
is also wind-pollinated and dioecious, but does not grow clonally at
our sites and has a higher investment in sexual reproduction than
T. fendleri
. The
pollen : ovule ratio is largest in wind-pollinated species, but varies considerably among
them, possibly reflecting differences in pollination efficiency and/or incidental insect
pollination. None of the species appear pollen limited in the study populations, suggest-
ing that factors other than pollen receipt limit female fertility at these sites.
Keywords:
clonal growth, insect pollination, pollen : ovule ratios, self-fertilization,
Thalictrum
,
wind pollination.
Received 19 May 2003; revision received 19 December 2003; accepted 9 January 2004
Introduction
The first flowering plants were insect-pollinated (Thien
et al
. 2000), but many plant species utilize other animal
taxa or even wind and water to carry pollen between
flowers (Cox 1991; Proctor
et al
. 1996). Wind pollination,
the most common form of abiotic cross-pollination,
releases dioecious and self-incompatible plants from
relying on animals for pollen transfer (Regal 1982; Niklas
1985; Berry & Calvo 1989; Goodwillie 1999; Culley
et al
.
2002). Self-fertilization can also provide reproductive
assurance in lineages with poor pollinator service (Steb-
bins 1957; Levin 1972; Solbrig & Rollins 1977; Lloyd 1979;
Uyenoyama
et al
. 1993; Holsinger 1996). The extinction of
pollinators or range expansion in a plant lineage can
favor shifts to abiotic modes of pollination, including
wind pollination and autonomous self-fertilization
(Baker 1955; Stebbins 1957; Regal 1982; Cox 1991; Weller
et al
. 1998). Abiotic pollination may also increase pheno-
logical and developmental flexibility, which may be par-
ticularly beneficial in climates with short growing
seasons (Cox 1991).
Changes in pollination syndrome and levels of self-
fertilization are frequently accompanied by changes in
gamete production. These changes may be adaptive and
are reflected in the ratio of pollen to ovules produced by
a plant (Cruden 1977; Ackerman 2000; Cruden 2000).
High pollen production can increase the success of wind
pollination and wind-pollinated species typically have
very high pollen : ovule ratios (Ackerman 2000). In
contrast, self-fertilizing species usually have low
pollen : ovule ratios because a low number of pollen
grains are sufficient for effective pollination (Cruden 1977,
2000).
Plants may also adapt to poor pollinator service by
modifying growth form. Clonal growth results in an
increase in genet size and longevity, which could amelio-

74
J. C. STEVEN AND D. M. WALLER
© 2004 The Society for the Study of Species Biology
Plant Species Biology
19, 73– 80
rate the deleterious effects of ineffective pollination in any
one season (Eriksson 1997). Clonality also allows plants
to forage for spatially distributed resources and poten-
tially gain competitive superiority due to their larger size.
In addition, the multiplication of ramets capable of an
independent existence can reduce the risk of extinction of
the entire genet (Cook 1985). However, clonal growth may
occur at the expense of investment in reproduction
(Waller 1988; Eriksson 1997).
Thus, clonal growth, wind pollination and self-fertili-
zation may all evolve in response to poor pollinator ser-
vice. These major evolutionary shifts may trigger an
increase (in the case of wind pollination) or decrease (in
the case of self-fertilization) in pollen production, or a
decrease in overall allocation to reproduction (in the case
of clonality). Here, we examine pollination success and
patterns of allocation to reproduction among four species
in the genus
Thalictrum
that vary in pollination
syndrome.
Many species in the genus
Thalictrum
have evolved
away from insect pollination. Phylogenetic results sug-
gest that wind pollination has evolved in multiple lin-
eages (Brunet & Liston 1999), potentially in response to
poor reproductive success in insect-pollinated ancestors
(Melampy & Hayworth 1980). Self-compatibility is also
reported for some
Thalictrum
species (Melampy & Hay-
worth 1980; Lubbers & Christensen 1986) and clonal
growth occurs in some species (Boivin 1944; Park & Fes-
terling 1997). Moreover,
Thalictrum
flowers possess vari-
able numbers of floral organs, uniovulate, free carpels
(Park & Festerling 1997), a labile floral phyllotaxy
(Endress 1987) and a potential for homeotic changes in
floral organ identity during development (Lehmann &
Sattler 1994). These characteristics result in a develop-
mental flexibility that can promote rapid changes in allo-
cation to male and female function in response to
selection on pollination syndrome.
In this study, we use reports from the literature,
pollination experiments and observations of plant
growth to determine the mode of pollination, ability to
self-fertilize and presence of clonal growth in our
study species. To determine the consequences of these
conditions on allocation patterns, we also measured
pollen : ovule ratios and biomass allocation to repro-
duction. We expect that all of the study species exhibit
one or more modes of reproduction that reduce or
eliminate dependence on pollinator service. We also
expect that species that grow clonally allocate less to
reproductive structures than non-clonal species due to
a greater allocation of biomass to vegetative growth.
Finally, we expect high pollen : ovule ratios to be asso-
ciated with wind pollination and lower pollen : ovule
ratios to be associated with insect-pollinated and self-
fertilizing species.
Materials and methods
Study species
Based on morphological characters, Kaplan & Mulcahy
(1971) predicted the pollination syndrome of 90 species
of
Thalictrum
and, from this list, we identified species of
Thalictrum
that vary in pollination syndrome and breed-
ing system. We selected two hermaphroditic species:
T.
alpinum
L., scored as anemophilous, and
T. sparsiflorum
Turcz. ex Fisch. & Mey., scored as either wind or insect
pollinated (or possibly both). In addition, we selected
two dioecious species with flowers categorized as ane-
mophilous (
T. dioicum
L. and
T. fendleri
Englem. ex. A.
Gray).
All four of these species are herbaceous perennials.
Thalictrum sparsiflorum
occurs in damp thickets and
woods in northern temperate regions around the globe,
typically growing 30–100 cm tall.
Thalictrum alpinum
grows only 5–30 cm tall and occurs in wet meadows, alka-
line bogs and damp ledges and slopes. The two dioecious
species,
T. dioicum
and
T. fendleri
, commonly grow 30–
80 cm tall.
Thalictrum dioicum
occurs in eastern deciduous
forests in North America and favors mesic woods and
ravines, while
T. fendleri
occurs in western North America
and northern Mexico and is typically found in open forest
or shrub habitats. All four species have no petals, no nec-
tar glands and free uniovulate carpels.
Thalictrum sparsi-
florum
has white, persistent sepals while the other species
have green or greenish-purple sepals that fall off early
(Park & Festerling 1997).
Study sites
We conducted research on
T. sparsiflorum, T. fendleri
and
T. alpinum
populations located in the Bridger-Teton
National Forest in the Wind River Range of the Rocky
Mountains in Sublette County, Wyoming. We conducted
observations and experiments on a field population of
T.
dioicum
in the Nicolet National Forest, Forest County, Wis-
consin. Voucher specimens are deposited at the Wisconsin
State Herbarium (WIS).
Clonality
We excavated the root systems of
T. fendleri, T. dioicum
and
T. sparsiflorum
plants to investigate clonal growth, and
were careful to preserve connections between plants. We
excavated a total of five female and four male
T. fendleri
,
five male and five female
T. dioicum
and seven
T. sparsiflo-
rum
plants. Due to their small size and the nature of the
soil, we were unable to fully excavate
T. alpinum
plants
.
However, those plants we did excavate were connected to
other shoots via rhizomes and the species is described as
strongly clonal (Boivin 1944; Park & Festerling 1997).

REPRODUCTION IN
THALICTRUM
75
© 2004 The Society for the Study of Species Biology
Plant Species Biology
19, 73– 80
Biomass, flower production and fruit set
We collected above-ground structures of 20–
50 reproductive plants at each site during flowering and
again at fruiting to measure dry-weight biomass alloca-
tion. For clonal species, we considered each ramet to be
an individual. For each plant or ramet, we recorded
flower number and separated the leaves and stems from
the inflorescence. Because carpels are free and uniovulate,
we estimated fruit and seed set by counting the number
of maturing and carpels that failed to mature on
10 flowers per plant and dividing the number of maturing
carpels by the total number of carpels counted. We dried
the structures at 60
∞
C to constant weight and weighed
them to 0.001 g. For the dioecious species, we collected an
equal number of males and females (populations did not
deviate from a 1 : 1 sex ratio). We calculated the mean
proportion of biomass invested in reproduction at flow-
ering and fruiting by averaging males and females.
Pollen and ovule counts
We collected mature flowers with undehisced anthers at
each site and preserved them in ethanol. We also collected
female flowers for ovule counts in the dioecious species
(
n
=
14–19 flowers per species). We lightly crushed the
anthers of each flower to remove the pollen, then used a
Coulter particle counter (Beckham Coulter: Fullerton, CA)
to estimate the number of pollen grains in each flower
(Kearns & Inouye 1993). We also counted the number of
carpels (and therefore ovules) produced by each flower
under a dissecting microscope at 20
¥
magnification.
Pollinator observations
To investigate the effectiveness of insect pollination in
T.
sparsiflorum
, we observed the number of insect visits to
plants in July 2000. We observed two patches of four plants
each for 1.5 h in the morning and a separate pair of patches
of four plants each for 1.5 h in the late afternoon.
Pollination experiments
To test the role of wind, insects and self-fertilization in
reproductive success, we used a subset of five possible
treatments on 10–20 plants per treatment in each popula-
tion. The treatments were: (i) exclosure – inflorescences
were enclosed in self-supporting bags of bridal veil to
exclude insect pollinators, but allow wind pollination. (ii)
Bag effect – inflorescences were enclosed in a bridal veil,
but we removed the bags and hand-pollinated the flowers
when they were receptive with a mixture of pollen from
three or more other plants in the population. This treat-
ment was used to determine whether or not the bags were
shading the carpels or stressing the plant, both of which
could result in reduced seed set. (iii) Pollen addition –
flowers were pollinated by hand when receptive, with the
same mixture of pollen to determine seed set when polli-
nation success is not limiting. (iv) Control plants were
marked, but unmanipulated. (v) Autonomy (cf. Lloyd &
Schoen 1992) – in the hermaphroditic species, flowers
were enclosed in bags to test for the presence of self-
fertilization in the absence of pollinators.
Thalictrum sparsiflorum
received the exclosure, pollen
addition, control and autonomy treatments in early July
2000. We used self-supporting bags of muslin in the auton-
omy treatment. We applied the pollen addition, control
and autonomy treatments in the
T. alpinum
population in
mid-June 2000. Because the flowers were small, we used
aluminum foil bags in the
T. alpinum
autonomy treatments
and we were unable to include the exclosure or bag effect
treatments.
Thalictrum dioicum
received the exclosure, bag
effect, pollen addition and control treatments in mid-May
2000.
Thalictrum fendleri
received the pollen addition and
control treatments in mid-June 2001, but we were unable
to apply the bagging treatments as flowers and inflores-
cences were very small when the stigmas became exposed.
Data analysis
We compared the proportion of total plant biomass
invested in reproduction at flowering and at fruiting
among species using Analysis of Variance (
ANOVA
). We
also used
ANOVA
to compare the number of flowers per
plant and overall plant biomass among species, applying
a log transformation to both variables to increase the
equality of variance. We compared the fraction of carpels
matured under natural conditions (arcsine square root
transformed) among species using
ANOVA
. We compared
the number of pollen grains per flower and the number
of ovules per flower among species using
ANOVA
, com-
paring means a posteriori using Tukey’s method. We used
an independent samples
t
-test with pooled variance to
compare the pollen : ovule ratio of the two hermaphro-
ditic species. We compared the observed distribution of
pollinator visits per plant in
T. sparsiflorum
to a Poisson
distribution with
l=
0.8 (the observed mean) using a
c
2
goodness-of-fit test to assess whether plants had equal
probabilities of receiving insect visits. Finally, we assessed
the effects of pollination treatments within species by
comparing fruit set (arcsine square root transformed)
using
ANOVA
.
Results
Clonality
Thalictrum dioicum
and
T. sparsiflorum
showed no evidence
of clonal growth. Root systems produced multiple shoot
buds in some cases (in one instance two stems of
T.

76
J. C. STEVEN AND D. M. WALLER
© 2004 The Society for the Study of Species Biology
Plant Species Biology
19, 73– 80
dioicum
grew from the same root system), but none of the
plants had rhizomes. In contrast,
T. fendleri
was strongly
clonal and averaged 1.6 ramets per genet. All the plants
we excavated had horizontal rhizomes with root-bearing
nodes and four had rhizomes that branched at one or
more nodes. The distance between nodes averaged
13.5 cm. The longest intact rhizome we excavated mea-
sured 114 cm, with no more than 30 cm between nodes.
The longest distance between above-ground shoots was
72 cm. We found a few dead rhizomes, suggesting that
connections between ramets may decay over time.
Biomass, flower production and fruit set
Based on dry-weight biomass at flowering,
T. alpinum
pro-
duces the smallest plants, while
T. fendleri
and
T. sparsiflo-
rum
are similarly sized and
T. dioicum
plants are the
largest (
F
3,146
=
544.71,
P
<
0.0001; Table 1). The number of
flowers per plant also differed (
F
3,257
=
76.59,
P
<
0.0001),
with
T. dioicum
producing significantly more flowers per
plant than the other species (Table 1). At flowering,
T.
alpinum
allocated over half of its above-ground biomass
to reproduction, while the other three species allocated
significantly less (
F
3.146
=
609.83,
P
<
0.0001; Table 1). This
difference was also present at fruiting (
F
3,107
=
300.92,
P
<
0.0001; Table 1). These four species also varied widely
in the fraction of carpels that matured into fruits under
natural conditions (
F
3,213
=
42.061,
P
<
0.0001; Table 1).
Thalictrum sparsiflorum
matured more than 90% of its
carpels, compared to about 50% in
T. dioicum
and less than
30% in
T. alpinum
and
T. fendleri
.
Pollen and ovule counts
Both pollen and ovule production per flower differed
significantly among species (
F
3,61
=
103.76,
P
<
0.0001 and
F
3,61
=
50.493,
P
<
0.0001; Fig. 1). Based on Tukey’s
method and a = 0.05, both the dioecious, wind-pollinated
species produced more pollen grains than either her-
maphroditic species. Although the pollen count for T.
sparsiflorum did not differ from that of T. alpinum, T. alpi-
num had far fewer ovules per flower and thus a higher
pollen : ovule ratio. Thalictrum sparsiflorum also produced
more ovules per flower than the dioecious species (all
differences significant by Tukey’s method at a = 0.05).
The pollen-ovule ratio was greatest in T. dioicum, inter-
mediate in T. alpinum and T. fendleri, and least in T. spar-
siflorum, with strong differences between the two
hermaphroditic species (T. alpinum and T. sparsiflorum,
t = 6.06, d.f. = 33, P < 0.0001; Table 1).
Table 1 Summary of growth and allocation patterns and pollination characteristics of the four study species.
Thalictrum sparsiflorum Thalictrum alpinum Thalictrum fendleri Thalictrum dioicum
Breeding system Hermaphroditic Hermaphroditic Dioecious Dioecious
Clonal growth? No Yes Yes No
Total plant biomass at flowering (mg) 848 ± 377a 44 ± 11b 704 ± 266a 2043 ± 952c
Number of flowers produced per plant 15.9 ± 13.6a8.4 ± 2.4a8.9 ± 5.6a69.7 ± 40.2b
Fraction of biomass in inflorescence at
flowering
0.018 ± 0.010a0.53 ± 0.11b0.037 ± 0.032ac 0.053 ± 0.036c
Fraction of biomass in inflorescence at
fruiting
0.057 ± 0.045a0.54 ± 0.14b0.024 ± 0.041a0.052 ± 0.040a
Carpels per flower 12.1 ± 3.12a3.6 ± 0.87b9.5 ± 2.11c7.3 ± 1.64d
Pollen-ovule ratio 4240 : 1 29 980 : 1 16 880 : 1 54 380 : 1
Fraction of carpels matured under
natural conditions
0.914 ± 0.072a0.189 ± 0.266b0.230 ± 0.241b0.487 ± 0.250c
Probable pollination syndrome Insect pollination,
autonomy
Wind pollination,
autonomy
Wind pollination Wind pollination
Within each row, means sharing the same superscript letter are not significantly diffent at P = 0.05 based on ANOVA and an a posteriori
comparison using Tukey’s method.
Fig. 1 Mean pollen and ovule number per flower for four species
of Thalictrum. Bars represent ±1 SD from the mean. Data points
for T. sparsiflorum and T. alpinum represent counts collected from
within the same flower for pollen and ovules. Data points for T.
fendleri and T. dioicum represent counts from separate plants in
the same population.

REPRODUCTION IN THALICTRUM 77
© 2004 The Society for the Study of Species Biology Plant Species Biology 19, 73–80
Pollinator observations
We observed visitors from three insect orders (Hemiptera,
Hymenoptera and Diptera) to T. sparsiflorum. Twenty vis-
its were distributed over the 16 plants observed, with the
observed number of visits resembling a Poisson (random)
distribution (c2 = 1.94, d.f. = 4, P = 0.747).
Pollination experiments
Despite wide variation in the fraction of carpels matured
under natural conditions, the pollen addition treatment
did not significantly increase fruit set in any of the four
species (Table 2, Fig. 2). The plants that were hand polli-
nated and not bagged in T. dioicum had greater fruit set
than the plants that were hand pollinated and bagged with
a bridal veil, suggesting that bags affect fruit set in T.
dioicum. The two hermaphroditic species, T. alpinum and
T. sparsiflorum, both produced fruit in the autonomy treat-
ment, but the fraction of mature ovules was somewhat
reduced relative to the control in T. sparsiflorum (Fig. 2).
Discussion
None of these four species of Thalictrum depend entirely
on insect pollinators for reproductive success. Although
T. sparsiflorum receives insect visits, it is also capable of
self-fertilizing in the absence of pollinators. Thalictrum
alpinum has the floral morphology and high pollen : ovule
ratios characteristic of wind pollination, is autonomously
self-fertilizing and grows clonally. The dioecious species
T. fendleri and T. dioicum have wind-pollinated floral mor-
phology and pollen : ovule ratios, with T. fendleri also
capable of clonal growth (Table 1). Thus, all these species
are employing reproductive and/or growth strategies
that ensure fitness even when insect pollinators are scarce
or unreliable.
The two species with clonal growth (T. alpinum and T.
fendleri) also produce the fewest flowers and mature the
smallest fraction of fruits, suggesting that they are allocat-
ing resources to clonal growth at the expense of reproduc-
tion, as demonstrated in other clonal species (Waller 1988;
Eriksson 1997; Piquot et al. 1998). Clonal growth can
increase plant longevity and thus increase lifetime fitness
when sexual reproduction is likely to be poor in some
years (Eriksson 1997). Therefore, if T. alpinum and T. fen-
dleri have a history of poor pollination success, selection
could favor reduced allocation to reproduction and
increased allocation to growth, reducing fruit set even
when pollen is not limiting. Thalictrum fendleri also allo-
cated a low fraction of its biomass to reproduction in the
absence of pollen limitation. In contrast, flowering ramets
of T. alpinum mature a small fraction of carpels, but allo-
cate a large proportion of their above-ground biomass to
inflorescences at both flowering and fruiting. However, if
nearby non-flowering ramets connected to the flowering
ramet contribute resources to reproduction, our measure
overestimates reproductive allocation in this species.
A variety of generalist pollinators were observed to
visit hermaphroditic T. sparsiflorum. Pollination by a vari-
ety of insect taxa is reported for other Thalictrum species
as well (Robertson 1928; Kaplan & Mulcahy 1971;
Melampy & Hayworth 1980; Lubbers & Christensen 1986;
Davis 1997) and is also typical of small, white, dish-
shaped flowers (Faegri & van der Pijl 1979). However,
when we excluded insects from flowers with bags of
bridal veil, seed set was not significantly reduced, sug-
gesting that wind and/or autonomous self-pollination
also contribute to fruit set. While pollen : ovule ratios in
T. sparsiflorum are much lower than in the wind-pollinated
study species (Table 1), wind might play a role in self-
fertilization within a flower or occasionally transfer pol-
len between flowers. Flowers not exposed to either wind
or insects matured 74% of their ovules, which was signif-
icantly less than the 91% matured in the control treatment
(Fig. 2). This reduction is possibly due to decreased pollen
receipt, or early acting inbreeding depression and/or
weak self-incompatibility due to self-fertilization (Keller
Table 2 Analysis of variance of fraction of maturing carpels per plant with pollination treatment as the factor for each species of
Thalictrum.
Species Source Sum of Squares d.f. F-ratio Significance
Thalictrum sparsiflorum treatment 0.686 3 5.601 0.002
error 2.778 68
Thalictrum alpinum treatment 0.293 2 1.202 0.307
error 7.922 65
Thalictrum fendleri treatment 0.026 1 0.266 0.614
error 1.470 15
Thalictrum dioicum treatment 0.368 3 2.919 0.042
error 2.227 53
Fraction of maturing carpels was arcsine square root transformed prior to analysis.

78 J. C. STEVEN AND D. M. WALLER
© 2004 The Society for the Study of Species Biology Plant Species Biology 19, 73–80
Fig. 2 Results of pollination experiments in four species of Thalictrum. Bars represent mean fraction of mature carpels within a treatment.
Error bars are ±1 SE from the mean. Hermaphroditic species are indicated with the letter H and dioecious species are indicated with
the letter D. Within each graph, means that share letters are not significantly different from one another based on Tukey’s method at
a = 0.05. Results of analyses of variance for each species are presented in Table 2.
& Waller 2002). Alternatively, the bags used in the auton-
omy treatment may also have damaged the plant stems
or shaded the carpels, reducing local photosynthesis and
carbon gain. Photosynthesis in carpels contributes to the
carbon balance of morphologically similar flowers in
Ranunculus adoneus (Ranunculaceae) (Galen et al. 1993).
Our treatments in T. dioicum also suggest that the bag
itself may affect fruit set, as flowers hand pollinated
within bags produced fewer fruits than flowers hand pol-
linated and not in bags (Fig. 2). Thus, our results suggest
that both insect pollination and self-fertilization poten-
tially contribute to reproductive success in T. sparsiflorum.
While T. sparsiflorum has pollen : ovule ratios more similar
to insect-pollinated outcrossing species than species with
high levels of self-fertilization (Cruden 2000), self-
fertilization could still provide reproductive assurance
when pollinators are scarce or ineffective.
Although self-compatible, T. alpinum also has a
pollen : ovule ratio that is not similar to self-fertilizing
species, but is more typical of wind-pollinated species

REPRODUCTION IN THALICTRUM 79
© 2004 The Society for the Study of Species Biology Plant Species Biology 19, 73–80
(Cruden 2000). The non-directed dispersal of pollen in
wind-pollinated systems favors higher investment in pol-
len production (Faegri & van der Pijl 1979; Cruden 2000)
and the high ratio in T. alpinum may indicate that both
cross-fertilization via wind pollination and self-fertiliza-
tion occur. While the flowers appear to be strongly
adapted for wind pollination, insect visits could also con-
tribute to fruit set in this species. The low number of
ovules per flower in T. alpinum as compared to T. sparsi-
florum may also represent an adaptive reduction in ovule
production in favor of increased pollen production. Alter-
natively, the evolution of small flower and plant size in T.
alpinum may have resulted in reductions in carpel number
per flower due to genetic constraints.
The two dioecious obligately outcrossing species, T.
dioicum and T. fendleri, both have high pollen : ovule ratios
similar to other wind-pollinated species (Cruden 2000).
However, T. dioicum has a larger pollen : ovule ratio than
T. fendleri, perhaps reflecting differences in pollination
efficiency or levels of incidental insect pollination. The
open, windy habitat found in the Wind River Range could
assure effective wind pollination at lower pollen : ovule
ratios in T. fendleri. While insect pollination could poten-
tially contribute to reproductive success and also select
for reduced pollen : ovule ratios in this species, receptive
female T. fendleri flowers lack showy sepals, nectar and
pollen, and are therefore not likely to receive reliable
insect visits. Clonality in T. fendleri could also favor real-
location of resources away from pollen production and
towards vegetative growth. The closed temperate forest
habitat of T. dioicum may result in low pollination effi-
ciency, selecting for high pollen : ovule ratios. When we
excluded insect pollinators from female T. dioicum flow-
ers, fruit set was not reduced below natural levels, sug-
gesting that wind is the primary and possibly exclusive
pollen vector in this species.
Adding pollen by hand failed to increase fruit set in any
of the species studied suggesting that current pollen vec-
tors and levels of self-fertilization delivered adequate pol-
len loads in these populations during our study. However,
if pollen limitation varies among plants and years, hand
pollinations in one year may not be sufficient to detect it.
The high fraction of carpels matured in T. sparsiflorum
may also decrease our ability to detect low levels of pollen
limitation in this species. In the other three species, factors
other than pollen limitation appear to limit the proportion
of carpels matured.
Plants commonly produce more ovules than are
matured into seeds, especially when fruit and seed costs
are high and/or hermaphroditic flowers primarily serve
a male function (Bell 1985; Charlesworth 1989). Muta-
tional load can also lead to the abortion of fertilized
ovules (Charlesworth 1989). In addition, the resources
available to mature fertilized ovules may limit fruit set in
our study populations (Bierzychudek 1981; Stephenson
1981; Burd 1994). Lubbers & Christensen (1986) demon-
strated that light availability limited fruit set in T. thalic-
troides at the end of the growing season. Light could limit
fruit set in T. dioicum and T. fendleri as both species grow
in forest understories where light levels can be low.
Manipulating nutrient, water and/or light levels while
tracking male and female components of fitness could
further elucidate limits to fruit set in these species.
In conclusion, the four species of Thalictrum we studied
have multiple strategies to ensure reproductive success in
the absence of pollinators. Clonal growth can provide
increased genet longevity and fitness, and autonomous
self-fertilization can assure seed production when a
flower does not receive outcross pollen. In addition, wind
pollination is common in the genus. Thalictrum lineages
may have been predisposed to the evolution of wind pol-
lination if selection pressure to attract pollinators in the
absence of nectar rewards increased pollen production.
This increase, combined with the lack of tight plant-pol-
linator relationships in Thalictrum and the developmen-
tally plastic floral structure characteristic of the
Ranunculaceae, may have promoted the evolution of
effective wind pollination and further increases in pollen
production.
Acknowledgments
The authors wish to thank Johanne Brunet for helpful
comments on an earlier draft and Tom Givnish, Tom Shar-
key, Linda Graham and Murray Clayton for helpful input
and discussion. We are also grateful to Jeanne Sheahan
and Brian Anacker for help with fieldwork. This study
constitutes part of Janet Steven’s Dissertation research at
the University of Wisconsin-Madison, and is based upon
work supported under a National Science Foundation
Graduate Fellowship to Janet Steven and grants from the
Beta Chapter of SDE-Graduate Women in Science, Sigma
Xi Grants in Aid of Research and the J. J. Davis fund of
the Department of Botany at the University of Wisconsin-
Madison.
References
Ackerman J. D. (2000) Abiotic pollen and pollination: Ecological,
functional, and evolutionary perspectives. Plant Systematics
and Evolution 222: 167–185.
Baker H. G. (1955) Self-compatibility and establishment after
‘long-distance’ dispersal. Evolution 9: 347–349.
Bell G. (1985) On the function of flowers. Proceedings of the Royal
Society, London B224: 223–265.
Berry P. E. & Calvo R. N. (1989) Wind pollination, self-
incompatibility and altitudinal shifts in pollination systems
in the High Andean genus Espeletia (Asteraceae). American
Journal of Botany 76: 1602–1614.

80 J. C. STEVEN AND D. M. WALLER
© 2004 The Society for the Study of Species Biology Plant Species Biology 19, 73–80
Bierzychudek P. (1981) Pollinator limitation of plant reproduc-
tive effort. American Naturalist 117: 838–840.
Boivin B. (1944) American Thalictra and their old world allies.
Rhodora 46: 337–487.
Brunet J. & Liston A. (1999) The evolution of dioecy in the genus
Thalictrum (meadow rue) (Ranunculaceae). International Botan-
ical Congress, St. Louis, MO.
Burd M. (1994) Bateman’s principle and plant reproduction: The
role of pollen limitation in fruit and seed set. Botanical Review
60: 83–139.
Charlesworth D. (1989) Evolution of low female fertility in
plants: Pollen limitation, resource allocation and genetic load.
Trends in Ecology and Evolution 4: 289–292.
Cook R. E. (1985) Growth and development in clonal plant pop-
ulations. In: Population Biology and Evolution of Clonal Organ-
isms (Eds J. B. C. Jackson, L. W. Buss & R. E. Cook), pp. 259–
296. Yale University Press, New Haven.
Cox P. A. (1991) Abiotic pollination: An evolutionary escape for
animal-pollinated angiosperms. Philosophical Transactions of
the Royal Society of London Series B. Biological Sciences 333: 217–
224.
Cruden R. W. (1977) Pollen-ovule ratios: a conservative indica-
tor of breeding systems in flowering plants. Evolution 31: 32–
46.
Cruden R. W. (2000) Pollen grains: why so many? Plant System-
atics and Evolution 222: 143–165.
Culley T. M., Weller S. G. & Sakai A. K. (2002) The evolution of
wind pollination in angiosperms. Trends in Ecology and Evolu-
tion 17: 361–369.
Davis S. L. (1997) Stamens are not essential as an attractant for
pollinators in females of cryptically dioecious Thalictrum
pubescens Pursch. (Ranunculaceae). Sexual Plant Reproduction
10: 293–299.
Endress P. K. (1987) Floral phyllotaxis and floral evolution. Bot-
anische Jahrbuecher Fuer Systematik Pflanzengeschichte und
Pflanzengeographie 108: 417–438.
Eriksson O. (1997) Clonal life histories and the evolution of seed
recruitment. In: The Ecology and Evolution of Clonal Plants (Eds
H. de Kroon & J. van Groenendael), pp. 211–226. Backhuys
Publishers, Leiden.
Faegri K. & van der Pijl L. (1979) The Principles of Pollination
Ecology. Pergamon Press, New York.
Galen C., Dawson T. E. & Stanton M. L. (1993) Carpels as leaves:
Meeting the carbon cost of reproduction in an alpine butter-
cup. Oecologia 95: 187–193.
Goodwillie C. (1999) Wind pollination and reproductive assur-
ance in Linanthus parviflorus (Polemoniaceae), a self-incom-
patible annual. American Journal of Botany 86: 948–954.
Holsinger K. E. (1996) Pollination biology and the evolution of
mating systems in flowering plants. Evolutionary Biology 29:
107–149.
Kaplan S. M. & Mulcahy D. L. (1971) Mode of pollination and
floral sexuality in Thalictrum. Evolution 25: 659–668.
Kearns C. A. & Inouye D. W. (1993) Techniques for Pollination
Biologists. The University Press of Colorado, Niwot, CO.
Keller L. F. & Waller D. M. (2002) Inbreeding effects in wild
populations. Tr ends in Ecology and Evolution 17: 230–241.
Lehmann N. L. & Sattler R. (1994) Floral development and home-
osis in Actaea rubra (Ranunculaceae). International Journal of
Plant Sciences 155: 658–671.
Levin D. A. (1972) Competition for pollinator service: a stimulus
for the evolution of autogamy. Evolution 26: 668–669.
Lloyd D. G. (1979) Some reproductive factors affecting the selec-
tion of self-fertilization in plants. American Naturalist 113: 67–
79.
Lloyd D. G. & Schoen D. J. (1992) Self- and cross-fertilization in
plants. I. Functional dimensions. International Journal of Plant
Sciences 153: 358–369.
Lubbers A. E. & Christensen N. L. (1986) Intraseasonal variation
in seed production among flowers and plants of Thalictrum
thalictroides (Ranunculaceae). American Journal of Botany 73:
190–203.
Melampy M. N. & Hayworth A. M. (1980) Seed production and
pollen vectors in several nectarless plants. Evolution 34: 1144–
1154.
Niklas K. J. (1985) The aerodynamics of wind pollination. Botan-
ical Review 51: 328–386.
Park M. M. & Festerling D. Jr (1997) Thalictrum. In: Flora of North
America North of Mexico (Ed. Flora of North America Editorial
Committee), pp. 258–271. Oxford University Press, New York.
Piquot Y., Petit D., Valero M., Cuguen J., deLaguerie P. & Vernet
P. (1998) Variation in sexual and asexual reproduction among
young and old populations of perennial macrophyte Sparga-
nium erectum. Oikos 82: 139–148.
Proctor M., Yeo P. & Lack A. (1996) The Natural History of
Pollination. Timber Press, Portland, OR.
Regal P. J. (1982) Pollination by wind and animals: Ecology of
geographic patterns. Annual Review of Ecology and Systematics
13: 497–524.
Robertson C. (1928) Flowers and Insects: Lists of Visitors of Four
Hundred and Fifty-Three Flowers. Carlinville, IL.
Solbrig O. T. & Rollins R. C. (1977) The evolution of autogamy
in species of the mustard genus Leavenworthia. Evolution 31:
265–281.
Stebbins G. L. (1957) Self fertilization and population variability
in the higher plants. American Naturalist 41: 337–354.
Stephenson A. G. (1981) Flower and fruit abortion: Proximate
causes and ultimate functions. Annual Review of Ecology and
Systematics 12: 253–279.
Thien L. B., Azuma H. & Kawano S. (2000) New perspectives on
the pollination biology of basal angiosperms. International
Journal of Plant Sciences 161: S225–S235.
Uyenoyama M. K., Holsinger K. E. & Waller D. M. (1993) Eco-
logical and genetic factors directing the evolution of self-
fertilization. Oxford Surveys in Evolutionary Biology 9: 327–381.
Waller D. M. (1988) Plant morphology and reproduction. In: Plant
Reproductive Ecology: Patterns and Strategies (Eds J. Lovett
Doust & L. Lovett Doust), pp. 203–227. Oxford University
Press, New York.
Weller S. G., Sakai A. K., Rankin A. E., Golonka A., Kutcher B. &
Ashby K. E. (1998) Dioecy and the evolution of pollination
systems in Schiedea and Alsinidendron (Caryophyllaceae: Als-
inoideae) in the Hawaiian islands. American Journal of Botany
85: 1377–1388.