Content uploaded by Celso Feitosa Martins
Author content
All content in this area was uploaded by Celso Feitosa Martins
Content may be subject to copyright.
Pollination of Campanula rapunculus L. (Campanulaceae):
How much pollen flows into pollination and into reproduction
of oligolectic pollinators?
C. Schlindwein
1
, D. Wittmann
2
, C. F. Martins
3
, A. Hamm
2
, J. A. Siqueira
1
,
D. Schiffler
2
,and I. C. Machado
1
1
Departamento de Botaˆ nica, Universidade Federal de Pernambuco, Recife, Brazil
2
Institut fu
¨r Landwirtschaftliche Zoologie und Bienenkunde, Universita
¨t Bonn, Bonn, Germany
3
Departamento de Sistema
´tica e Ecologia-CCEN, Universidade Federal da Paraı
´ba, Joa
˜o Pessoa, PB,
Brazil
Received October 2, 2003; accepted September 8, 2004
Published online: January 31, 2005
Springer-Verlag 2005
Abstract. We studied an isolated population of
Campanula rapunculus and two oligolectic bee
species of Chelostoma (Megachilidae), their main
pollinators. The population of C. rapunculus
consisted of 2808 plants. Measurements of pollen
flow showed that 3.7% of the pollen produced by
a flower contribute to pollination, 95.5% was
collected by bees for their offspring and 0.8%
remained on the styles. Pollen analyses of brood
cells of Chelostoma rapunculi revealed that
females collected on average 4.9 million Campan-
ula pollen to rear one bee. We calculated that
approximately 1588 bees of this species could
have been reared at the study site during the
studied season. The amount of potentially viable
pollen deposited on stigmas was 3.6 to 10.7 times
higher than the number of ovules. We discuss
morphological features of the flowers which may
lower the pollen removal rate per bee visit and
consequently cause a high visitation and pollina-
tion rate.
Key words: Campanula rapunculus L., pollen
partitioning, Chelostoma rapunculi,Chelostoma
campanularum, Megachilidae, effective pollinators.
Pollen has two main functions in ecosystems:
it is essential for the reproduction of plants
and serves as food for flower visiting and
pollinating insects. In the case of pollen col-
lecting female bees, huge amounts of pollen are
withdrawn from the flowers and serve as food
for bee larvae. This pollen is lost for immediate
pollination but indirectly benefits pollination as
it serves to feed future pollinators.
The allocation of pollen between the pollen
producing flower and the pollinating bee can
best be studied in cases in which flowers are
visited almost exclusively by oligolectic polli-
nating bee species. Bees of these species are
specialized to collect pollen only in flowers of
the same genus or family of plant. Such a case is
given in most species of Campanula which are
frequently visited by bees of the genus Chelo-
stoma (Mu
¨ller 1873, Westrich 1989). In several
cases, this close relationship between plant and
bee species is the result of a coevolutionary
process (Schlindwein and Wittmann 1997,
Alves-dos-Santos and Wittmann 2000).
Plant Syst. Evol. 250: 147–156 (2005)
DOI 10.1007/s00606-004-0246-8
Campanula comprises about 400 species,
distributed mainly in temperate Europe, Asia
and North Africa and some species in North
America. Campanula rapunculus is a biannual,
ruderal herb, 30–100 cm high, with funnel to
bell shaped violet flowers. The species occurs
from Northern Germany and the Netherlands
to Northwest Africa and Syria in the South,
Spain in the West and the Caucasus in the East
(Rosenbauer 1996).
Flowers of Campanula have a peculiar
mechanism of secondary pollen presentation:
the anthers with introrse dehiscence open before
anthesis and shed their pollen on the pollen
collecting-hairs of the style (Sprengel 1793,
Mu
¨ller 1873, Kirchner 1897, Knuth 1899, Jost
1918, Shetler 1979, Yeo 1993). At the beginning
of anthesis, all pollen grains adhere to these
pollen-collecting hairs. During this functional
male phase the hairs are retracted into the style
(Mu
¨ller 1873; Leins and Erbar 1990; Nyman
1992, 1993a, b). Leins (2000) suggests that the
retraction, which is stimulated by mechanical
contacts of the flower visitors starts at the apex
of the style and continues in direction of the
base, has the function to gradually liberate the
pollen grains. One to several days later, depend-
ing on the Campanula species, the stigma lobes
spread and the functional female phase starts.
Bees of numerous species are cited as flower
visitors of Campanula (Blionis and Vokou
2001). Besides the two Chelostoma species
recorded in our study (Ch. rapunculi and Ch.
campanularum), Westrich (1989) lists further
eight species out of four families as Campanula
oligoleges: Andrena curvungula,A. pandellei,
A. rufizona (Andrenidae), Dufourea dentiventris,
D. inermis (Halictidae), Chelostoma distinctum,
Osmia mitis (Megachilidae) and Melitta
haemorrhoidalis (Melittidae).
Where does the pollen go? In the case of
the oligolectic bee Ptilothrix plumata (Antho-
phoridae) Schlindwein and Martins (2000)
calculated how many pollen grains and flowers
of Pavonia cancellata (Malvaceae) are neces-
sary to feed one bee larva. However, in
general, there are no quantitative data about
1) the amount of pollen of a plant population
collected by female bees to rear offspring-these
data permit to calculate the potential popula-
tion size of an oligolectic bee species at a given
site and 2) the amount of pollen that reached
the stigmas-a measure for the efficiency of the
pollination mechanism and the functioning of
the plant-pollinator system.
Thus, in this study we asked: How much
pollen is available in the population of C. ra-
punculus at the study site? How much of this
pollen is transferred to the stigmas of the
flowers of C. rapunculus and how much is
needed to fertilize all ovules? How much pollen
flows into (offspring of) Chelostoma rapunculi?
Material and methods
Study site and population size of Campanula rapun-
culus. The field study was performed during June
and July 2001 and 2002 in an abandoned and
isolated 50 ha gravel-pit area which has been under
protection as ‘‘Nature Reserve Du
¨nstekoven’’ since
1988. It is located W of Bonn, Germany (E
656¢0900,N5042¢03¢). The vegetation is charac-
terized by pioneer plants. This reserve borders on
one side a forest area and is surrounded by
plantations of different cereal species, Brassica
napus,Sinapis arvensis and Beta vulgaris (Fig. 1).
To estimate the availability of Campanula pollen
we mapped the study site and its surroundings and
counted all flowering plants of C. rapunculus. The
average number of flowers produced in the lifetime
of a plant was determined from counts on 45
individuals near the end of the flowering period.
The total number of flowers was calculated by
multiplying the number of flowering plant
individuals with the average number of flowers per
plant.
Flower morphology, anthesis, floral longevity
and breeding system. Floral diameter, length of the
flower and length of the pollen presenting area on
the style were measured in ten flowers. Buds of 79
flowers were marked individually and anthesis of
these flowers, which were accessible to flower
visitors, was monitored until senescence paying
particular regard on duration of functional male
and female phases. We considered the beginning of
anthesis and the functional male phase when the
petals opened wide enough to permit flower visitors
to enter. Functional female phase was defined to
148 C. Schlindwein et al.: How much Campanula pollen flows into pollination and reproduction of bees?
begin with separation of the stigma lobes (Evanhoe
and Galloway 2002). The duration of anthesis of
the Campanula flowers is given in daylight hours at
the study site (14 hours in June/July).
We determined the breeding system by con-
trolled pollination tests. 1. Spontaneous selfing:
flower buds were bagged before anthesis and
maintained enclosed until flower senescence. 2.
Hand self-pollination: flower buds were bagged
before anthesis and pollinated with pollen of the
same flower during the female phase and 3. Hand
cross-pollination: bagged flowers were opened and
pollinated during the female phase with pollen
from two other individuals of C. rapunculus.4.
Open, free pollination: for control marked flowers
were kept accessible to pollinators.
Pollen counts. Pollen grain numbers were
determined with a particle counter (CASY I;
Scha
¨rfe, Germany) which measures the exact
number, size and volume of pollen grains in a
sample. For the measurements, the grains were
dispersed in 10 ml Casyton, an isotonic liquid
provided by Scha
¨rfe, Germany.
To evaluate the number of pollen grains
presented at the pollen collecting hairs, ten flowers
were collected before anthesis. Pollen grains which
adhered to the pollen collecting hairs were washed
with 10% KOH solution, centrifuged, transferred
to the isotonic liquid and counted with the particle
counter. From a sub-sample of each, a microscope
slide with basic fuchsine stained glycerin was
prepared (Louveaux et al. 1978) to determine the
percentage of pollen grains without protoplasm
(empty, not developed grains) by counting 600–
800 grains. Grains which remained at the style and
other floral parts were collected with glycerin
gelatin and embedded on microscope slides.
To determine the amount of pollen which was
gradually withdrawn from flowers the same proce-
dure was performed with 16 flowers which were
accessible to bee visitors. Six of them were collected
3 hours after beginning of anthesis and ten flowers
at the end of anthesis (wilted flower). In these cases
pollen grains were counted under a microscope,
because the CASY counter is not designed to count
very low particle numbers.
At the end of anthesis we removed the stigmas
from ten flowers and prepared microscope slides
each containing the three stigma lobes. Fuchsine
stained glycerin was used for mounting to improve
pollen counting. The pollen grains which adhered
to the stigma lobes were identified and counted.
Frequency and behavior of flower visitors. Fre-
quency of flower visitors was determined by
counting female and male bees of the different
species at the flowers from 11:00h (first flower
visits) to 18:15h for 15 minutes per hour. Nine
flowers in male and nine flowers in female phase
Fig. 1. Study site ‘‘Natural Reserve Du
¨nstekoven’’ with 27 mapped patches of Campanula rapunculus (circles
with numbers of individual plants)
C. Schlindwein et al.: How much Campanula pollen flows into pollination and reproduction of bees? 149
were monitored simultaneously. We noted whether
the visitors collected pollen or nectar.
In order to determine the origin and number of
pollen grains necessary to rear one bee larva, we
identified and counted the pollen grains in nine
brood cells from two freshly provisioned nests of
Chelostoma rapunculi. These cells were obtained
from trap nests, small pieces of wood with burrows
of 4–10 mm diameter, which were placed at the
study site to attract female bees of solitary species
looking for nesting sites.
Specimens are deposited in the Entomological
Collection of the Institute for Agricultural Zoology
and Apiculture of the University of Bonn, Ger-
many and in the Entomological Collection of the
Laboratory of Plant Ecology of the Federal
University of Pernambuco, Recife, Brazil.
Results
Pollen production and flower characteristics. At
the study site, C. rapunculus flowered from
beginning of June until mid of July 2001. The
total population of C. rapunculus consisted of
2808 plants which grew in 27 patches (Fig. 1).
Besides C. rapunculus, no other Campanula
species occurred at the nature reserve. No
Campanula plants were found in the surround-
ing forest and plantations. As each plant had
on average 35 flowers (sd ¼26, N ¼45),
the local population presented approximately
98280 flowers. On average a flower of C. rapun-
culus produced 82935 (sd ¼15674, N ¼10,
range 66080 – 107330) pollen grains. On
average 17.4% (range 8.2 – 45%) of them were
empty; this means that on average 68504 grains
were potentially viable.
Opening time of the flowers was not synchro-
nized. Flowers opened or changed from male to
female phase at any hour between 10 – 18:00 h.
Flowers which were visited by bees had an average
longevity of 21.2 hours of daylight (Fig. 2).
Duration of functional male and female phases
were similar (Mann-Whitney U test ¼611.0,
p > 0.05, Nfemale ¼52, Nmale ¼30).
Floral diameter was 1.8 cm (sd ¼0.24),
length of the flower 1.3 cm (sd ¼0.13), length
of the flower tube 0.8 cm (sd ¼0.09) and
length of the pollen presenting area 0.9 cm
(sd ¼0.1). The bases of the filaments are
triangular, dilated and right above the base
they adhere to the style in the middle of the
flower, forming a nectar chamber. Between the
basal parts of the filaments there are narrow
slits which give access to the nectar. Each petal
has a row with five to nine white setae (max.
length 2.5 mm) which insert on the upper 2/3
of its midrib. These setae touch the style in the
center of the flower (Fig. 3a). During the
female phase the hairs wilted and adhered to
the corolla. Thus, in the beginning of the male
phase when the flower tubes are still narrow
and the setae rigid, the flowers are divided into
five compartments. Each of them is limited by
two rows of setae and the style in the middle.
Flower visitors, their frequency and
behavior. Nine species were recorded as flower
visitors of C. rapunculus (Table 1). Only Ch. ra-
punculi (body length 9–10 mm) and the tiny Ch.
campanularum (body length 4.5–6 mm) visited
the flowers frequently. Females and males of
both species visited the flowers of Campanula
rapunculus in the same manner with their sterna
oriented towards the styles.
The difference in abundance of females of
these two species was not significant (Fig. 4).
Chelostoma rapunculi females were most abun-
dant at 13:00 h in male as well as in female
phase flowers, while Ch. campanularum did not
show a distinct peak of activity.
Chelostoma rapunculi spent significantly
more time (12.4 sec, sd ¼9.3, N ¼33) collect-
Fig. 2. Duration of functional male and female
phase, and total time of anthesis of Campanula
rapunculus.Malephase:mean¼12.0 ± 5.33 h,
range ¼2–21 h, N ¼52; female phase: mean ¼9.6
±5.31h,range¼5–21 h, N ¼30; total duration of
anthesis: mean ¼21.2 hours ± 4.19 h, range ¼11–
28 h, N ¼27). One day equals 14 daylight hours
150 C. Schlindwein et al.: How much Campanula pollen flows into pollination and reproduction of bees?
ing pollen than collecting nectar (4.6 sec,
sd ¼1.4, N ¼15; Mann-Whitney U test ¼69.0,
p < 0.001). Females of Ch. campanularum
needed also significantly more time for pollen
collection (62.3 sec, sd ¼46.8, N ¼18) than for
nectar collection (5.9 sec, sd ¼3.7, N ¼22;
Mann-Whitney U test ¼2.50, p<0.001).
Pollen collecting visits of Ch. campanularum
were significantly longer than those of Ch.
rapunculi (U ¼36.5 p<0.001 NCh. campanula-
rum ¼18, NCh. rapunculi ¼33). Chelostoma
campanularum females were also observed to
collect pollen grains that had become attached
to the style in female phase flowers. This
gleaning behavior was rare for females of the
larger Ch. rapunculi.
Males of both species were significantly less
abundant as flower visitors than females
(U test ¼0.0 p<0.001 for Ch. rapunculi and
U test ¼8.0 p<0.05 for Ch. campanularum,
N¼156). Patrolling flights of males which
inspected the flowers of Campanula in search
of females were common. Males sporadically
visited the flowers to take up nectar during
patrolling flights. At night males of Ch.
rapunculi were observed to sleep in the flowers
of C. rapunculus.
When females entered a flower at the
beginning of the male phase, they generally
restricted their visit and their pollen collecting
activity to one of the five compartments
limited by the setae (Fig. 3b). Later in the
Fig. 3. Flowers of Campanula rapunculus during the male phase. aBeginning of the male phase showing long setae
on corolla midribs. bFemale of Chelostoma rapunculi collecting pollen in a compartment of a flower of
C. rapunculus
Table 1. Flower visiting insects of Campanula rapunculus at the Natural Reserve Du
¨nstekoven (27.6.2001/
28.6.2001, 11:00 - 18:15h, 18 flowers, 15 min counts/hour)
Species Family Females Males** Sum
Hymenoptera
Andrena flavipes (Panzer, 1799) Andrenidae 4 – 4
Andrena bicolor (Fabricius, 1775) Andrenidae 1 – 1
Andrena chrysosceles (Kirby, 1802) Andrenidae 2 – 1
Bombus pascuorum (Scopoi, 1763) Apidae 1 – 1
Lasioglossum cfr. pygmaeum (Schenck, 1868) Halictidae 3*
Chelostoma rapunculi (Lepeletier, 1841) Megachilidae 86 8 (147) 241
Chelostoma campanularum (Kirby, 1802) Megachilidae 38 8 (106) 152
Diptera
Episyrphus balteatus Syrphidae 10*
Sphaerophoria sp. Syrphidae 9*
* sex of the flower visitors not identified
** Numbers in parentheses refer to patrolling males that hovered in front of the flowers without contact to
floral parts (include multiple counting of individuals)
C. Schlindwein et al.: How much Campanula pollen flows into pollination and reproduction of bees? 151
male phase, when all setae had wilted, females
could rotate in the flowers during pollen
collection without any restriction to compart-
ments. In general, females and males touched
the trilobed stigma with the legs and ventral
part of meso- and metasoma when entering
flowers in the female phase.
Breeding system. Bagged flowers without
treatment (spontaneous self-pollination) pro-
duced only one fruit with two seeds (Table 2).
Fruit set in hand self-pollinated flowers was
50% and in hand cross-pollinated flowers 90%
of the manipulated flowers. Fruits resulting
from hand cross-pollination produced five
times more seeds than those resulting from
hand self-pollination. The bee pollinated flow-
ers had 100% fruit set and produced on average
372 (sd ¼70.8, N ¼22) seeds per fruit.
Pollen withdrawal and deposition on
stigmas. Three hours after the beginning of
anthesis bees had already removed 61.6% of
the mean amount of pollen per flower and by
the end of anthesis only 0,8% remained at the
style (Table 3).
Analysis of the pollen grains deposited on
the stigma lobes of C. rapunculus at the end of
anthesis showed that on average 99.1% were
conspecific. On average bees had deposited
3075 (range 1570 - 4693) pollen grains of
C. rapunculus on the stigma (Table 4). This is
3.7% of the average pollen produced by a
C. rapunculus flower. Pollen of other plants
came mainly from Asteraceae (3 types) and
Convolvulus (Convolvulaceae).
Flowers of C. rapunculus contained a
mean of 361.4 ovules (sd ¼44.4; range 287
– 409, N ¼10, pollen/ovule ratio is 229.7).
Considering that on average 17.4% of the
deposited grains were empty, 2540 potentially
viable pollen grains were deposited on the
stigmas. This is 7 times (range 3.6 – 10.7) the
number of ovules. However, this number
diminishes by an unknown percentage
when pollen comes from one of the few
open flowers in the male phase of the same
plant.
Brood cells of Chelostoma rapunculi.In
two nests of Ch. rapunculi which contained 4
and 5 brood cells, pollen analysis of the
larval provisions revealed that the bees had
exclusively collected pollen from C. rapun-
culus. On average the brood cells contained
4.90 million (range 2,98 – 6,54) pollen grains
(Table 5). This is the mean pollen content of
59.1 (range 35,9 – 78,8) flowers of C.
rapunculus.
Fig. 4. Visits of females of Chelostoma spp. during
one day to flowers of Campanula rapunculus in male
and female phase. C.rap.M ¼Chelostoma rapunculi
visiting flowers in male phase; C.rap.F ¼C. rapunculi
visiting flowers in female phase; C.cam.M ¼C.
campanularum visiting flowers in male phase; C.cam.
F¼C. campanularum visiting flowers in female phase
Table 2. Breeding system of Campanula rapunculus. Fruit set and average seed set of bagged flowers
without treatment (spontaneous self-pollination), hand self-pollinated flowers, hand cross-pollinated
flowers and flowers pollinated by flower visitors (controls)
Treatment N Produced
fruits (N)
Fruit
set (%)
Seeds per
fruit (mean)
Spontaneous self-pollination 22 1 4 2
Hand self-pollination 12 6 50 11
Hand cross-pollination 10 9 90 53
Controls 22 22 100 372
152 C. Schlindwein et al.: How much Campanula pollen flows into pollination and reproduction of bees?
Discussion
At the end of anthesis only a few pollen grains
– less than 1% - remained in the Campanula-
flowers. They were widely distributed over the
style so that these grains could not be gleaned
even by the tiny Ch. campanularum. This
shows that to the two pollen collecting oligo-
lectic species of Chelostoma,Campanula pollen
is scarce in the ‘‘Natural Reserve Du
¨nsteko-
ven’’ and that the females of these species
should compete for this resource with each
other and with the sporadic flower visitors.
If we do not take into account predators
and nest parasites as limiting factors for the
Table 3. Gradually declining numbers of pollen grains per flower during anthesis
Pollen grains N average sd min max %
Total pollen grains per flower 10 82935 15674 66080 107330 100
Three hours after start of anthesis 6 31865 11622 12880 44850 38.4
End of anthesis 12 702 397 194 1403 0.8
Table 4. Pollen grains deposited on the stigma lobes at the end of anthesis. Other pollen came from
Asteraceae and Convolvulaceae
Pollen grains on stigmas %Pollen of
Campanula
Flower Campanula Other plants Total
1 2247 7 2254 99.7
2 2897 103 3000 96.6
3 4153 0 4153 100.0
4 1943 4 1947 99.8
5 3386 2 3388 99.9
6 2780 1 2781 99.9
7 4064 87 4151 97.9
8 3020 40 3060 98.7
9 1570 14 1584 99.1
10 4693 32 4725 99.3
average 3075 29 3104 99.1
sd 959 36 969 1.1
Table 5. Pollen grains of Campanula rapunculus from two nests of Chelostoma rapunculi
Nest Brood cells Number pollen grains % developed % empty grains
1 1 6342501 80.4 19.6
1 2 3116677 79.1 20.9
1 3 6536252 81.9 18.1
1 4 5573332 89.1 10.9
1 5 5011253 91.5 8.5
2 1 2983332 63.2 36.8
2 2 4972920 75.3 24.7
2 3 4399170 83.0 17.0
2 4 5191255 78.6 21.4
average 4902965 80.2 19.8
sd 1174671 8.2 8.2
C. Schlindwein et al.: How much Campanula pollen flows into pollination and reproduction of bees? 153
populations of the two Chelostoma species,
Campanula pollen appears to determine the
carrying capacity of the nature reserve for
these bee species. This was even more so, as the
population of C. rapunculus at the study site
was isolated by surrounding cultivated areas
and a forest where neither plants of C. rapun-
culus nor of other Campanula species grew.
Furthermore, the small size of the Chelostoma
bees makes it impossible for them to include
distant Campanula resources into their forag-
ing range. This specific situation of a closed
system allows us to calculate the potential
maximum size of the Chelostoma population
and to make quantitative estimations about
pollen fate between recipient flowers and the
oligolectic bee species.
Our data show that one larva of Ch. rapun-
culi is reared on 59 flowers of C. rapunculus.
This is the equivalent to the total pollen
production of about 1.7 plants, taking 35
flowers as the average number per plant.
Of the approximate 8.15 billion pollen
grains produced during the season by the
98280 flowers, 65 million might remain as
uncollectable in the flowers and 301 million
might be transferred to the stigmas, leaving
7.78 billion (95.5%) for the bees. If all these
pollen grains would flow into reproduction of
Ch. rapunculi, 1588 brood cells of this species
could be provisioned.
Certainly this number is not reached for
reasons such as competition for pollen between
all kinds of flower visiting insects, due to
predators, nest parasites and diseases and also
by loss of plants following disturbances. From
the view of habitat conservation it is notable
that the loss of 1.7 plants of C. rapunculus could
cause the loss of one bee larva of Ch. rapunculi.
Pollination success and oligolectic
bees. The considerably high seed set (100%)
in control flowers of our breeding experiments
indicates that the two oligolectic bee species
cause a very high pollen flow between conspe-
cific plants. This is either due to a very high
number of bees or, in this case, to a mechanism
which causes bees to visit the flowers in high
frequencies. In the flowers of Campanula, such
a mechanism may be exhibited by special
morphological and physiological traits.
Several authors have considered pollen-
collecting hairs as either a morphological
structure facilitating pollen presentation or as
a mechanism to guide nectar seeking insects to
the nectaries (Mu
¨ller 1873, Kirchner 1897,
Knuth 1899). Other authors have focused on
stimulation of the pollen-collecting hairs which
shortens the duration of the male phase
(Richardson and Stephenson 1989; Nyman
1992, 1993b). Erbar and Leins (1989, 1995)
and Leins (2000) interpreted the retraction of
the pollen collecting hairs at the style as a
mechanism to gradually liberate and portion
the pollen as well as to remove self-pollen from
the flower before the beginning of the female
phase. They assumed that the pollen falls in
succession down and adheres to flower visiting
insects. However, we observed that all females
actively collected pollen directly from the style.
Even so, Erbar and Leins’ (1989) interpreta-
tion is valid as it points to a mechanism which
is apt to reward bees with small amounts of
pollen and keeps them ‘‘harried’’ and ‘‘under-
fed’’ (Feinsinger 1983) and thus forces them to
frequent flower visits.
This effect might be strengthened by another
feature which the flowers exhibit at least at the
beginning of the male phase, when large
amounts of pollen are available at the pollen
collecting hairs. The long setae on the midribs of
the corolla lobes divide the corolla tube around
the style in compartments and thus restrict
females to collect pollen only from limited parts
of the style. This might further lower the pollen
removal rate per visit and consequently cause
high visitation and pollination rates. The num-
ber of pollen grains deposited on the stigma was
very high, several times higher than the number
of ovules. As there are only one or a few open
flowers in the male phase at the same plant by
the same time, this was predominately outcross
pollen. The stigmatic pollen load, however, can
often not be directly related to the number of
ovules. In some species a minimum load of
pollen grains is required on a stigma to trigger
seed set or to stimulate pollen tube growth
154 C. Schlindwein et al.: How much Campanula pollen flows into pollination and reproduction of bees?
(Schemske and Fenster 1983, Cruzan 1986,
Cruden 2000). A minimum of 4–6 pollen grains
per ovule seem to be necessary for maximum
seed set (Cruden 2000).
Nyman (1993b) has pointed out that the
more often a flower is visited by pollinators, the
more frequently the pollen collecting-hairs are
stimulated. This results in a shortened male
phase and accelerates the onset of the female
phase. Furthermore, for C. americana Evanhoe
and Galloway (2002) showed that increasing
pollen deposition shortens the female phase.
In sum, the flower gradually releases pollen
and temporarily restricts the access of polli-
nators to the pollen collecting hairs thus
causing them to frequently visit the flowers.
In return flower visitors frequently stimulate
the pollen-collecting hairs and thus shorten the
male phase. Frequent visits to flowers in the
female phase not only cause a 100% fertiliza-
tion rate but also shorten the phase. Thus, the
frequent flower visits, which are forced by
specific floral traits, cause shortening of male
and female phases. This increases male and
female fitness (Evanhoe and Galloway 2002).
The study was supported by a joint project
CAPES / DAAD (Probral 112/00). We thank
NaBu (Naturschutzbund) Bonn and the Untere
Landschaftsbeho
¨rde Siegburg for the permission to
work at the Natural Reserve Du
¨nstekoven.
References
Alves-dos-Santos I., Wittmann D. (2000) Legiti-
mate pollination of the tristylous flowers of
Eichhornia azurea (Pontederiaceae) by Ancylos-
celis gigas bees (Anthophoridae, Apoidea). Plant
Syst. Evol. 223: 127–137.
Blionis G. J., Vokou D. (2001) Pollination ecology
of Campanula species on Mt Olympos, Greece.
Ecography 24: 287–297.
Cruden R. W. (2000) Pollen grains: why so many?
In: Dafni A., Pacini E., Hesse M. (eds.) Pollen
and pollination. Springer, Berlin, pp. 143–165.
Cruzan M. B. (1986) Pollen tube distributions in
Nicotiana glauca: evidence for density dependent
growth. Amer. J. Bot. 73: 902–907.
Erbar C., Leins P. (1989) On the early floral
development and the mechanisms of secondary
pollen presentation in Campanula, Jasione and
Lobelia. Bot. Jahrb. Syst. 111 (1): 29–55.
Erbar C., Leins P. (1995) Portioned pollen release
and the syndromes of secondary pollen presen-
tation in the Campanulales - Asterales - com-
plex. Flora 190: 323–338.
Evanhoe L., Galloway L. F. (2002) Floral lon-
gevity in Campanula americana (Campanula-
ceae): a comparison of morphological and
functional gender phases. Amer. J. Bot. 89 (4):
587–591.
Feinsinger P. (1983) Coevolution and pollination.
In: Futuyma D. J., Slatkin M. (eds.) Coevolution.
Sinauer Associates, Sunderland, pp. 282–310.
Jost L. (1918) Die Griffelhaare der Campa-
nulablu
¨te. Flora 111: 478–489.
Kirchner O. (1897) Blu
¨teneinrichtungen der Cam-
panulaceen. Jahreshefte des Vereins fu
¨r Vater-
la
¨ndische Naturkunde in Wu
¨rttemberg 53: 193–
228.
Knuth P. (1899) Handbuch der Blu
¨tenbiologie unter
Zugrundelegung von Hermann Mu
¨llers Werk:
‘‘Die Befruchtung der Blumen durch Insekten’’.
II. Band. Wilhelm Engelmann, Leipzig.
Leins P. (2000) Blu
¨te und Frucht. Morphologie,
Entwicklungsgeschichte, Phylogenie, Funktion,
O
¨kologie. E. Schweizerbart’ sche Verlagsbuch-
handlung, Stuttgart.
Leins P., Erbar C. (1990) On the mechanism of
secondary pollen presentation in the Campa-
nulales-Asterales-complex. Botanica Acta 103:
87–92.
Louveaux J., Maurizio A., Vorwohl G. (1978).
Methods of melissopalynology. Bee World 59
(4): 139–157.
Mu
¨ller H. (1873) Die Befruchtung der Blumen
durch Insekten und die gegenseitigen Anpassun-
gen beider. Wilhelm Engelmann, Leipzig.
Nyman Y. (1992) Pollination mechanisms in six
Campanula species (Campanulaceae). Plant Syst.
Evol. 181: 97–108.
Nyman Y. (1993a) The pollen-collecting hairs of
Campanula (Campanulaceae). I. Morphological
variation and the retractive mechanism. Amer. J.
Bot. 80(12): 1437–1443.
Nyman Y. (1993b) The pollen-collecting hairs of
Campanula (Campanulaceae). II. Function and
adaptive significance in relation to pollination.
Amer. J. Bot. 80(12): 1427–1436.
C. Schlindwein et al.: How much Campanula pollen flows into pollination and reproduction of bees? 155
Richardson T. E., Stephenson A. G. (1989) Pollen
removal and pollen deposition affect the dura-
tion of the staminate and pistillate phases in
Campanula rapunculoides. Amer. J. Bot. 76(4):
532–538.
Rosenbauer A. (1996) Campanulaceae. In: Sebald
O., Seybold S., Philippi G., Wo
¨rz A. (eds.) Die
Farn- und Blu
¨tenpflanzen Baden-Wu
¨rttembergs.
Band 5: Spezieller Teil. (Spermatophyta, Unter-
klasse Asteridae) Buddlejaceae bis Caprifolia-
ceae. Verlag Eugen Ulmer, Stuttgart, pp. 417–449.
Schemske D. W., Fenster C. (1983) Pollen-grain
interactions in a neotropical Costus: effects of
clump size and competitors. In: Mulcahy D. L.,
Ottaviano E. (eds.) Pollen: biology and implica-
tions for plant breeding. Elsevier, New York, pp.
405–410.
Schlindwein C., Wittmann D. (1997) Micro foraging
routes of Bicolletes pampeana (Colletidae) and bee
induced pollen presentation in Cajophora arecha-
valetae (Loasaceae). Botanica Acta 110: 177–183.
Schlindwein C., Martins C. F. (2000) Competition
between the oligolectic bee Ptilothrix plumata
(Anthophoridae) and the flower closing beetle
Pristimerus calcaratus (Curculionidae) for floral
resources of Pavonia cancellata (Malvaceae).
Plant Syst. Evol. 224: 183–194.
Shetler S. G. (1979) Pollen-collecting hairs of
Campanula (Campanulaceae). I. Historical
review. Taxon 28: 205–215.
Sprengel C. K. (1793) Das entdeckte Geheimnis der
Natur im Bau und in der Befruchtung der
Blumen. Friedrich Vieweg, Berlin.
Westrich P. (1989) Die Wildbienen Baden-Wu
¨rt-
tembergs. Allgemeiner Teil. (I) and Die
Wildbienen Baden-Wu
¨rttembergs. Spezieller
Teil (II). Verlag Eugen Ulmer, Stuttgart.
Yeo P. F. (1993) Secondary pollen presentation.
Form, function and evolution. Plant Syst. Evol.
Suppl. 6. Springer, Wien New York, 268 p.
Addresses of the authors: Clemens Schlindwein
(e-mail: schlindw@ufpe.br), Jose
´Alves Siqueira,
Isabel Cristina Machado, Departamento de Bota-
nica, Universidade Federal de Pernambuco, Av.
Prof. Moraes Rego, s/n, 50670–901 Recife, Brazil.
Dieter Wittmann, Andre
´Hamm, Dirk Schiffler,
Institut fu
¨r Landwirtschaftliche Zoologie und
Bienenkunde, Universita
¨t Bonn, Melbweg 42,
63127 Bonn, Germany. Celso Feitosa Martins,
Departamento de Sistema
´tica e Ecologia – CCEN,
Universidade Federal da Paraı
´ba, Joao Pessoa, PB,
Brazil.
156 C. Schlindwein et al.: How much Campanula pollen flows into pollination and reproduction of bees?
