American Journal of Botany 84(12): 16851692. 1997.

MATING

STRUCTURE AND INBREEDING AND

OUTBREEDING DEPRESSION IN THE RARE PLANT

GENTIANELLA GERMANICA (GENTIANACEAE)1

MARKUS FISCHER2

AND

DIETHART MATTHIES

Institut fur Umweltwissenschaften, Universitat Zurich, Winterthurerstrasse 190, CH-8057 Zurich, Switzerland
Isolation and small size of populations as a result of habitat destruction and fragmentation may negatively affect plant
fitness through pollinator limitation and increased levels of inbreeding. To increase genetic variation in small populations
of rare plants artificial gene flow has been suggested as a management tool. We investigated whether pollinator limitation
and inbreeding depression could reduce fitness in Gentianella germanica, an endangered biennial of increasingly fragmented
calcareous grasslands in Central Europe. We experimentally excluded pollinators and generated progenies by hand-pollinating
flowers with pollen from different distances. G. germanica was highly selfing. Pollinator exclusion strongly reduced seed
set, indicating that pollinator limitation could potentially reduce plant fitness. Germination rate as well as number of leaves
and rosette size of progeny from 10-m crosses was higher than that of progeny from open pollinations, self-, 1-m, and
interpopulation crosses. After 6 mo of growth differences in the number of surviving plants persisted, whereas differences
in plant size did not. The results suggest that inbreeding depression may reduce plant performance in G. germanica.
Outbreeding depression in the performance of progeny from interpopulation crosses indicates that caution is necessary in
using artificial interpopulation gene flow as a management tool.
Key words: Gentianaceae; Gentianella germanica; hand pollination; inbreeding depression; mating system; outbreeding
depression; pollen quality; reproduction.

As a result of habitat destruction and fragmentation

many formerly common plant species are now restricted
to small and isolated populations (Saunders, Hobbs, and
Margules, 1991; Jennersten et al., 1992; Aizen and Feinsinger, 1994). These populations face an increased risk of
extinction due to random fluctuations of environmental
conditions (Menges, 1991a; 1992; Schemske et al.,
1994). Moreover, other factors including pollinator limitation of reproduction, increased inbreeding depression,
and the erosion of genetic variability may reduce the fitness of plants in small populations and substantially increase the probability of extinction (Menges, 1991b; Widen, 1993; Heschel and Paige, 1995).
Pollinator limitation of reproduction has been demonstrated for a number of species (Bierzychudek, 1981; Calvo, 1990; Primack and Hall, 1990; Johnston, 1991; Burd,
1994). Especially in small and isolated populations plant
pollinator mutualisms may be disrupted, resulting in reduced reproduction (Olesen and Jain, 1994; Bond, 1995).
In plant populations limited dispersal of both seeds and
pollen often leads to a population substructure that makes
selfing or inbreeding with close relatives likely (e.g.,
Wright, 1943; Campbell and Waser, 1987; Dudash, 1990;
Waser, 1993). In small and isolated populations the limited number of mating partners and reduced levels of genetic variation as a result of genetic drift may further
increase the likelihood of inbreeding (Lacy, 1987; Ellstrand and Elam, 1993). This can lead to a reduction of
1 Manuscript received 26 November 1996; revision accepted 23 April
1997.
The authors thank the Swiss National Science Foundation for financial support by the Priority Programme Environment (Module Biodiversity, grants No. 5001-35231 and 500144626), Veronika Stockli for
help in the field, and Mary Price, Bernhard Schmid, Jurg Stocklin, and
Nick Waser for fruitful comments on earlier versions of this manuscript.
2 Author for correspondence.

average plant fitness, in the short term because of inbreeding depression, which may be enhanced further by
the accumulation of mildly detrimental mutations (mutational meltdown, Lande, 1994; Lynch, Conery, and
Burger, 1995), and in the long term because the decrease
in genetic variation reduces the potential to adapt to
changing environmental conditions.
Pollen transfer as well as seed migration between populations seems to be very rare in many species (Slatkin,
1985). Nevertheless, for the maintenance of a sufficient
level of genetic diversity and the prevention of increasing
inbreeding levels occasional gene transfer between isolated populations is crucial (Ellstrand and Elam, 1993).
Increased fitness of offspring from interpopulation crosses has been found in a number of studies and the artificial
increase of gene flow among rare plant populations has
been suggested as a management tool (Van Treuren,
1993; Oostermeijer, Altenburg, and den Nijs, 1995).
However, large-distance gene transfer is not always beneficial. If genes in a local population have adapted to the
genetic environment defined by other genes (intrinsic
coadaptation) or if different populations have become
locally adapted, gene flow might result in outbreeding
depression (Price and Waser, 1979; Templeton, 1986;
Waser and Price, 1989; Lynch, 1991; Waser, 1993). However, there are hardly any studies that have actually investigated the possible effects of increased gene flow between populations of rare plants.
Genetic considerations are probably most important for
declining species with short generation times, i.e., annuals and biennials. For these species pollinator limitation
is also of particular importance, because their persistence
largely depends on seed production. In a previous study
on the demography of the formerly common biennial
Gentianella germanica (Gentianaceae), which is now rare

1685

1686

AMERICAN JOURNAL

and restricted to isolated populations, plants from small

populations were smaller, had less fruits, and produced
less seeds per fruit than plants from large populations
(Fischer, 1996). Average plant fitness per population was
also positively correlated with the amount of genetic variation in a population. These correlations were independent of habitat quality, suggesting that inbreeding depression and, perhaps, pollinator limitation may contribute to the observed pattern (Fischer, 1996).
In this paper we investigate whether pollinator limitation and inbreeding depression could potentially reduce
the fitness of plants of G. germanica, and study the effect
of interpopulation gene flow. We caged flowers and performed hand-pollinations in a population with pollen
from different distances (including pollen from a second
population) and recorded seed set, seed germination, and
growth and survival of offspring. We address the following questions: (1) Are pollinators important for seed set?
(2) What is the natural outcrossing rate? (3) Is there inbreeding depression? (4) Does interpopulation gene flow
affect offspring fitness?
MATERIALS AND METHODS
Study speciesGentianella germanica (Gentianaceae) is a biennial
species that germinates in spring, forms a rosette during the first year,
dies back above ground in winter, re-emerges in spring, and flowers in
the autumn of its second year (Hegi, 1926; Verkaar and Schenkeveld,
1984). Flowering plants typically reach a height of 530 cm and produce 330 flowers (Hegi, 1926). The flowers are mainly pollinated by
diptera and small bees (Fischer, 1996). Each fruit contains up to 100
seeds.
G. germanica is restricted to Central Europe. Typical habitats are
south-facing, nutrient-poor, calcareous grasslands (Zoller and Wagner,
1986; Oberdorfer, 1994). Because of changes in agriculture these types
of habitat have strongly declined, and G. germanica has consequently
become endangered in many parts of Central Europe (e.g., Landolt,
1991; Korneck, Schnittler, and Vollmer, 1996). The remaining populations are isolated from each other and often small. The number of individuals in populations of G. germanica is known to fluctuate strongly
from year to year (Runge, 1963).

OF

BOTANY

[Vol. 84

flower caged to exclude pollinators (to study spontaneous selfing). For

the following treatments we first emasculated and caged the selected
flowers; when stigmas had opened after 2 d we hand-pollinated them
by moving a ripe anther over the stigma: (3a) pollination with pollen
from a different flower of the same plant (to study seed set of selfed
flowers), (3b) pollination with pollen from a plant at a distance of 1 m
(near cross), (3c) pollination with pollen from a plant at a distance of
10 m (far cross), (3d) pollination with pollen from a plant from a second
population (population size ;300 flowering plants) 25 km away. We
performed all cross-pollinations with different paternal plants. For the
paternal plants we took the same size measures as for the maternal
plants. For the last treatment, we transported one flower with ripe pollen
from each of 15 randomly selected plants to the experimental site. We
completed all pollinations within 90 min after collecting the flowers.
After 4 wk we collected the ripe fruits and dried them at room temperature for a week. We determined number and mass of the seeds in
each fruit. Then we stored the seeds at 48C.
On 5 March 1994, we randomly selected 30 seeds in each of the
fruits (seed families) and soaked them in a solution of gibberellic acid
(2 mg GA3/mL water) for 5 d to break dormancy. On 10 March 1993
(subsequently referred to as day 0), we placed five seeds from each seed
family into each of six replicate pots containing a 1:1 mixture of soil
from a field site and sand. We kept the pots in a greenhouse at 258C
with 16 h per day of additional lighting. We weekly recorded the number of germinating seeds. On day 79, when no more seeds had germinated for ;2 wk, we measured the number of leaves and the widest
diameter for each rosette. Then we transferred the pots to the experimental garden next to the Botanical Institute of the University of Basel.
We made the same measurements on day 175 and in the next spring on
day 420. Contrary to our expectations based on demographic data from
the field, the winter mortality from day 175 to day 420 was extremely
high.
For each plant we calculated the cumulative leaf length as the product
of the number of leaves and half the rosette diameter. We calculated
two multiplicative estimates of fitness for days 175 and 420: w1, the
number of surviving offspring resulting from one fruit, was calculated
as number of seeds per fruit times germination rate times survival rate
until day 175 and day 420, respectively. w2 also takes plant size into
account and was calculated as the sum of cumulative leaf lengths of
surviving offspring per fruit.

Mating structureUnder the assumption that biparental inbreeding

can be neglected, estimates of fitness (w) for offspring resulting from
open pollination, selfing, and outcrossing can be used to calculate the
proportion of selfed offspring (s) in a natural population (Charlesworth,
1988).

Test for apomixis and seed set of emasculated flowersIn September 1994, we selected 22 plants of G. germanica in a population of 660
flowering plants near Kleinlutzel, 25 km southwest of Basel, Switzerland. On each plant we marked two flowers that were about to open.
To test the importance of own pollen for seed set we esmaculated one
flower by removing the anthers with a pair of fine forceps. Between
treatments we sterilized the forceps in a flame. We additionally caged
half of the emasculated flowers. Unless there is pseudogamy, this allows
to test for ability to set apomictic seeds. Around these flowers we placed
small bags made from pollen-proof fabric and carefully closed them
with gardening wire. To stabilize the manipulated flowers we fixed the
bags to small poles. After 4 wk we collected the fruits and counted the
number of ovules and seeds in each fruit.

For the estimation of offspring fitness we used two independent performance measures: the germination rate and the rosette diameter on
day 79. We took the fitness estimate for the offspring from 10-m crosses
as an estimate of woutcrossed. We calculated the standard error of s from
the standard errors of the two performance parameters using standard
error propagation formulas (Bronstein and Semendjajew, 1981).

Effects of pollinator exclusion and interparental distanceIn September 1993, we randomly selected and marked 90 plants in a population of 1400 flowering plants near Langenbruck, 35 km south of Basel,
Switzerland. We took five measures of plant size: numbers of branches,
leaves, and flowers, plant height, and widest diameter.
We randomly selected one flower on each plant that was about to
open, but with both stigma and anthers still closed. We subjected the
selected flowers to one of six different treatments (15 replicates per
treatment): (1) no experimental manipulation (open pollination), (2)

Statistical analysisWe tested the effect of emasculation with a

paired t test. We analyzed the other data with analyses of variance
(ANOVA). Sources of variation were pollination mode, hand-pollination mode (interparental distance) nested within pollination mode, maternal plant (seed family), and pot. We tested both the effects of pollination mode and of hand-pollination mode against variation among maternal plants (seed families). The effect of seed families was tested
against the variation among pots (Table 1).

s 5 (wopen 2 woutcrossed)/(wselfed 2 woutcrossed).

December 1997]

FISCHER

AND

MATTHIESEXPERIMENTAL

1687

POLLINATION OF A RARE PLANT

TABLE 1. Skeleton ANOVA for the effect of different pollination treatments. Pollination modes were open pollination, exclusion of pollinators
and hand-pollination. The hand-pollination treatment consisted of hand-pollinations with four different interparental distances.
Source of variation

df

Mean squares

Pollination mode
Hand-pollination mode nested within pollination mode
Maternal plant (seed family) nested within two upper levels
Residual variation among pots

2
3
84
394

MSP
MSH
MSF
MSI

MSP/MSF
MSH/MSF
MSF/MSI

RESULTS
Test for apomixis and seed set of emasculated flowersEmasculation and caging had no effect on the mean
(61SE) number of ovules per fruit (70.7 6 4.16, N 5
38), but significantly affected the number of seeds per
fruit. The 12 flowers that had been both emasculated and
caged produced no seeds at all. Unless the species is
pseudogamous, this indicates that there is no apomixis in
the study population. Emasculation without caging reduced seed set by 66% compared with open pollination
(P , 0.001; Fig. 1), indicating that pollination within the
same flower is an important part of natural pollination.

parameters between hand-pollinated and open-pollinated

plants.
Hand-pollination with pollen from a distance of 10 m
resulted in the highest number of seeds per fruit, but this
trend was not significant (Fig. 2). Because there was no
significant effect of crossing distance on seed set, differ-

Effects of pollinator exclusion and interparental distancePlant size did not affect seed production per fruit
in G. germanica. There was a weak negative relationship
between the number of flowers of the maternal plant and
the number (N 5 72, r 5 20.16, P , 0.18) and mass of
seeds per fruit (N 5 72, r 5 20.16, P , 0.17). The
position of a fruit along the inflorescence did not affect
its number of seeds. The size of the paternal plant also
had no influence.
The exclusion of pollinators by cages reduced the number of seeds per fruit by 55% (Fig. 2) and seed mass per
fruit by 65% compared with open pollination (Tables 2,
3). This indicates that pollinators contribute considerably
to natural pollination, and that pollinator limitation could
potentially reduce fitness of G. germanica. In contrast,
there were no significant differences in these reproductive

Fig. 1. The effect of emasculation on the mean number of seeds

per fruit in Gentianella germanica. Vertical bar denotes 1 SE of difference.

Figs. 23. The effect of six pollination treatments on (2) the mean
number of seeds per fruit and (3) the germination rate in Gentianella
germanica. Flowers were either caged to exclude pollinators, open-pollinated, or hand-pollinated after emasculation. Hand-pollination was
carried out with self pollen and with pollen from 1 m, 10 m (same
population), and 25 km (different population) distances. Error bars indicate 1 SE.

1688

AMERICAN JOURNAL

TABLE 2. Summary of analyses of variance (F values) of the effects

of pollination mode, interparental distance, and seed family (maternal plant) on plant reproduction and offspring fitness in Gentianella germanica. The fitness estimate w1 is the number of surviving offspring per fruit, w2 is the product of w1 and the mean
cumulative leaf length. Data on germination, survival, and multiplicative fitness w1 were angular transformed prior to analysis, data
on leaf length and w2 were log transformed. P , 0.05; P ,
0.01; P , 0.001.
Parameter

Seeds per fruit

Seed mass per fruit (mg)
Mean seed mass (mg)
Germination rate

Pollination mode

Interparental
distance

Seed family

14.8
10.5
3.50
0.0389

0.193
1.15
1.97
3.80

3.20

OF

BOTANY

[Vol. 84

TABLE 3. The effect of (A) different pollination modes and (B) different interparental distances in the hand-pollination treatment on
reproduction and offspring fitness in Gentianella germanica. Flowers were either caged to exclude pollinators, open-pollinated, or
hand-pollinated after emasculation. Hand-pollination was carried
out with self pollen and with pollen from 1 m, 10 m (same population), and 25 km (different population) distances. The fitness estimate w1 is the number of surviving offspring per fruit, w2 is the
product of w1 and the mean cumulative leaf length. Data on germination, survival, and multiplicative fitness w1 were angular-transformed prior to analysis; data on leaf length and w2 were log transformed. Back-transformed means are given in the table. Values in
a line with different superscripts are different at the 5% level (Tukeys HSD procedure).
A)
Pollination mode

Day 79
Number of leaves
Rosette diameter (mm)
Cumulative leaf length (mm)

2.57
1.32
2.14

7.15
4.14
5.95

1.25
2.12
1.99

Day 175
Surviving plants per seed
Number of leaves
Rosette diameter (mm)
Cumulative leaf length (mm)
w1
w2

0.376
0.491
2.03
1.94
8.60
8.56

3.51
0.604
2.03
0.373
3.34
4.17

2.78
2.78
1.46
1.99

Caged

Open

Seeds per fruit

Seed mass per fruit (mg)
Average seed mass (mg)

23.5a
3.28a
0.137a

52.4b
9.28b
0.192b

60.8b
9.88b
0.161ab

Day 79
Germination rate
Number of leaves
Rosette diameter (mm)
Cumulative leaf length (mm)

0.507
6.22
21.8
69.5

0.531
6.30
23.5
78.2

0.542
6.65
24.9
86.7

0.424
11.55
36.41
214
5.12a
92.1a

0.439
11.15
39.85
227
16.3b
299b

0.491
11.08
38.77
219
25.1b
446b

Parameter

ences in pollen viability between different hand-pollination treatments are unlikely. In contrast, the germination
rate of seeds resulting from far crosses within a population was much higher than that of seeds resulting from
the other treatments. The germination rate of seeds resulting from pollinations with pollen from a distance of
10 m was . 40% higher than that of seeds from selfed
flowers, from flowers pollinated with pollen from a distance of 1 m, or from flowers pollinated with pollen from
a different population (Fig. 3). This indicates inbreeding
depression after short-distance crosses and outbreeding
depression after interpopulation crosses.
Similarly, after 79 d of growth the number of leaves,
the rosette diameter, and the cumulative leaf length were
highest for the progeny of the 10-m treatment. The differences in plant size decreased with time. On day 175
survivorship of progeny of the 10-m treatment and that
of other treatments was still significantly different,
whereas the size of the surviving plants did not significantly differ between treatments any more. The survival
rate per pot (after angular transformation) from day 79
until day 175 was positively correlated with the mean
number of leaves (r 5 1 0.169, F 5 10.01; P , 0.002)
and the mean diameter of the plants in a pot (r 5 1
0.139, F 5 6.75; P , 0.01) on day 79. This indicates
that mainly small plants died during this period, which
diminished the differences in plant size between treatments.
Multiplicative fitness was lowest in the caged treatment
and highest in the hand-pollination treatment with the
intermediate outcrossing distance of 10 m (Table 3).
These differences were significant for day 175. This indicates consistent inbreeding depression after short-distance crosses and outbreeding depression after interpopulation crosses until day 175.
Overwinter mortality from day 175 to day 420 was

Day 175
Surviving plants per planted
seed
Number of leaves
Rosette diameter (mm)
Cumulative leaf length (mm)
w1
w2

Hand

B)
Interparental distance
Parameter

Selfed

1m

Day 79
Germination rate
Number of leaves
Rosette diameter (mm)
Cumulative leaf length (mm)

0.490a
0.513a
0.714b 0.453a
6.08a
6.43ab
7.10c
6.98bc
22.9a
23.0a
28.5b
25.4ab
72.4a
78.4a
104.3b
91.5ab

0.413a
11.2
36.4
208
18.9a
330a

0.473a
11.0
39.2
220
25.4ab
455ab

64.4
11.5
0.18

25 km

59.8
10.6
0.176

Day 175
Surviving plants per planted
seed
Number of leaves
Rosette diameter (mm)
Cumulative leaf length (mm)
w1
w2

60.9
8.93
0.142

10 m

Seeds per fruit

Seed mass per fruit (mg)
Average seed mass (mg)

0.700b
10.7
40.9
220
40.9b
792b

57.9
8.45
0.145

0.373a
11.4
38.6
224
17.2a
277a

extremely high (95.6%), drastically reducing statistical

power. Therefore, although differences between means
remained large, on day 420 none of the treatment effects
was significant any more.
Mating systemThe percentage of selfed seeds in the
studied population of G. germanica was 82.3% (SE
19.7%) if estimated from the germination rate (angular
transformed) and 89.3% (SE 23.0%) if estimated from
the rosette diameter on day 79, respectively. Both estimates are very similar and indicate a mixed mating system of G. germanica.

December 1997]

FISCHER

AND

MATTHIESEXPERIMENTAL

Genetic variation in offspring fitnessSeeds from

different maternal plants differed in germination rate, and
the resulting seedlings differed strongly in growth and
survival (Table 2). Mean germination rate per seed family
ranged from 0 to 0.90, mean cumulative leaf length at
day 175 from 72 to 595 mm, and mean survival rate from
day 79 to day 175 from 0.25 to 1. Size of the maternal
plant did not affect offspring demography. Treatment effects remained unchanged, when any of the measures of
maternal plant size (number of branches, leaves and flowers; plant height and diameter) were included as a covariate in any of the analyses. This indicates that differences among seed families in offspring fitness were not
due to maternal effects, but represent genetic variation.
DISCUSSION
Our results indicate that pollinator limitation could potentially reduce fitness of G. germanica. Natural seed set
of G. germanica strongly depended on pollinators (Table
3A, Fig. 2) and on pollen transfer within the flower (Fig.
1). It is likely that within-flower pollination is partly due
to pollinators (facilitated self-pollination; Lloyd and
Schoen, 1992). A reduced seed set after the exclusion of
pollinators has also been found for the related perennial
species Gentiana pneumonanthe and Gentiana cruciata
(Petanidou, den Nijs, and Ellis-Adam, 1991; Petanidou
et al., 1995; Petanidou, den Nijs, and Oostermeijer,
1995). In the monocarpic Gentianella amarella, however,
spontaneous selfing was as effective as open pollination
(T. Petanidou, Aristotle University, Thessaloniki, Greece,
personal communication).
We did not investigate the effect of supplementary
hand-pollination, but in our experiment seed set of handpollinated flowers was not higher than that of open-pollinated flowers. Therefore it appears that pollinator limitation does not play an important role in the study population of G. germanica. However, if there is a trade-off
between resources invested into natural pollen attraction
and seed provisioning, pollen limitation would only become manifest as reduced seed set after an experimental
reduction in pollen load (Haig and Westoby, 1988).
Moreover, the addition of pollen may not result in higher
seed set for a variety of reasons other than pollen limitation (Young and Young, 1992). Therefore our results
do not allow a final conclusion on the presence of pollen
limitation in the study population.
Moreover, pollen limitation has been demonstrated to
become more important with decreasing population size or
density (Sih and Baltus, 1987; Kunin, 1992, 1993; Jennersten and Nilsson, 1993; Byers, 1995), and may therefore affect only very small populations of G. germanica.
In order to explore the importance of pollinator limitation
in G. germanica in populations of different size we suggest direct observations of pollinator visits, counts of stigma pollen loads, and the comparison of natural seed set
with seed set following supplementary hand-pollination.
In the related species Gentiana cruciata and Gentianella
amarella supplementary hand-pollination increased seed
set compared to open pollination, indicating pollen limitation, but in Gentiana pneumonanthe it did not (Petanidou, den Nijs, and Ellis-Adam, 1991; Petanidou, den Nijs,
and Oostermeijer, 1995; Petanidou et al., 1995).

POLLINATION OF A RARE PLANT

1689

Offspring resulting from selfing and near-neighbor (1

m) crosses showed inbreeding depression in germination,
early growth, and survival compared with offspring of
the 10-m pollination treatment (Table 3B, Fig. 3). The
similarly low performance of offspring from the selfing
and 1-m treatments suggests that near neighbors are close
relatives (Price and Waser, 1979). A likely explanation is
the very restricted dispersal of seeds of G. germanica
(Verkaar, Schenkeveld, and van der Klashorst, 1983). The
significant variation among seed families in germination,
survival, and early plant size (Table 2) also indicates a
population structure shaped by limited seed and pollen
dispersal (e.g., Wright, 1943; Campbell and Waser, 1987;
Dudash, 1990; Waser, 1993). Inbreeding depression after
pollination with near neighbors was demonstrated for
several species (Levin, 1984; Dudash, 1990).
Because levels of inbreeding increase in declining populations (Ellstrand and Elam, 1993), a combination of
inbreeding depression and reduction in population size
could further reduce average plant performance in G. germanica. However, because the species is highly selfing,
the expected reduction in fitness due to increased inbreeding would be relatively small. Nevertheless, even a small
fitness reduction could crucially affect the persistence of
small populations. Moreover, the magnitude of inbreeding depression in the field could be considerably greater
than in the experimental garden (Dudash 1990).
A reduction in fitness due to inbreeding depression
may not have negative consequences for population survival if there is substantial self-thinning, because increased mortality due to inbreeding could be compensated by reduced self-thinning. However, in demographic
studies of seven natural populations no evidence for density-dependent mortality of seedlings in G. germanica
was found (Fischer, 1996). It is therefore likely that inbreeding depression reduces average fitness in natural
populations of G. germanica. However, in declining populations deleterious alleles might be purged because of
increased levels of inbreeding, counteracting a fitness decrease, and eventually even leading to a net increase of
average fitness (Barrett and Charlesworth, 1991). Therefore, it may appear unlikely that inbreeding depression
could pose a threat to population persistence. However,
inbreeding depression in populations that have been small
and isolated for a number of generations may rather represent the effect of the accumulation of mildly detrimental mutations than of the genetic load present before the
isolation (Lande, 1994; Lynch, Conery, and Burger,
1995). Negative effects of the accumulation of mildly
detrimental mutations are expected to increase with decreasing population size and increasing duration of isolation (Lande, 1995). We conclude that inbreeding depression may contribute to the observed lower fitness of
plants in small populations of G. germanica (Fischer,
1996).
Based on the reduced performance of selfed offspring
in comparison to that of offspring from 10-m crosses
(outcrossed) the selfing rate was calculated as ;85%.
However, the observed reduction in the number of seeds
produced by emasculated flowers would suggest a selfing
rate of 66%. The difference between these estimates
could be due to facilitated selfing, geitonogamy, inbreeding with close relatives, or natural variability, because the

1690

AMERICAN JOURNAL

estimates are based on data from different years and different populations. Outcrossing rates may vary among
populations (Schemske and Lande, 1985), among years,
and among seasons (Molau et al., 1989). Nevertheless,
both independently derived estimates suggest high selfing
rates. Self compatibility might allow G. germanica to establish a new population from a single seed after longdistance dispersal (Baker, 1955, 1967). A high selfing rate
is also of advantage for species whose populations show
large annual fluctuations in the number of individuals
(Stebbins, 1957), as reported for G. germanica (Runge,
1963). Inbred offspring of G. germanica suffered on the
other hand from reduced fitness, and selfing should therefore not be favored. The findings are in line with the
interpretation of mixed mating systems as a result of the
conflict between moderate levels of inbreeding depression that would favor selfing, and the genetic association
of outcrossing and high fitness, which favors some outcrossing (Uyenoyama, Holsinger, and Waller, 1993).
In a previous study we found significant genetic variation among 25 populations of G. germanica and reduced
levels of genetic variation in smaller populations (Fischer,
1996). Significant genetic variation among isolated populations also has been found in several other rare species
(Waller, OMalley, and Gawler, 1987; Brauner, Crawford,
and Stuessy, 1992; Raijmann et al., 1994). This suggests
that the average degree of relatedness of parent plants
used in this study was lower for parents of interpopulation crosses than for within-population crosses.
The most unusual aspect of our study is the detection
of significant outbreeding depression after interpopulation crosses. There are several mechanisms by which
gene transfer into a population could lead to reductions
in fitness: through disruption of coadapted gene complexes (Price and Waser, 1979; Templeton, 1986; Dudash,
1990), or by the disruption of local adaptations (Price and
Waser, 1979), a phenomenon that in plants has even been
observed within distances of 30 m (Waser and Price,
1994). Reproductive isolation in plants could also evolve
as a direct result of adaptive differentiation in sympatry
(MacNair and Christie, 1983).
Botanists have been reluctant to accept the phenomenon of outbreeding depression (Waser, 1993). In contrast,
zoologists have been aware already a decade ago of the
problem of simultaneously minimizing detrimental effects of inbreeding and outbreeding depression in breeding programs or management of natural populations
(Templeton 1986). In principle, enhanced gene flow
among isolated populations could be beneficial for average plant fitness in the short term through heterosis
effects (van Treuren et al., 1993; Hauser and Loeschke,
1994) and in the long term because of enhanced genetic
variation in the population (Ouborg and van Treuren,
1994; van Treuren et al., 1994). For example, in the rare
gentian G. pneumonanthe seed set and several measures
of offspring fitness, e.g., adult mass, increased with increasing outcrossing distance (Oostermeijer, Altenburg,
and Den Nijs, 1995). This was mainly the result of low
fitness of selfed offspring and high fitness of offspring
resulting from interpopulation crosses. Similarly, crosses
between individuals of different populations substantially
increased measures of offspring fitness such as biomass
and adult survivorship in Scabiosa columbaria (Van

OF

BOTANY

[Vol. 84

Treuren et al., 1994). Accordingly, it has been suggested

that management of rare plant species should try to improve the genetic exchange among populations by artificial gene flow (Ellstrand and Elam, 1993; Van Treuren,
1993). However, there are still very few studies that have
analyzed the consequences of in- and outbreeding over
the whole life cycle of a plant (Waser and Price, 1994).
In addition, the final consequences of outcrossing among
populations may only become apparent in the F2 and
subsequent generations (Lynch, 1991). It is therefore still
difficult to provide general suggestions on how to include
genetic aspects into the management of natural populations. However, to keep inbreeding coefficients low and
to reduce the problems of the accumulation of deleterious
mutations management must seek to maximize population sizes. This has the additional advantage of buffering
threats by environmental and demographic stochasticity
(Menges 1991a, 1992). To avoid the detrimental effects
of genetic drift and inbreeding and outbreeding depression, management could aim at maintaining or re-establishing gene flow in natural populations at the level prevailing prior to fragmentation or isolation. However, the
significant outbreeding depression after interpopulation
crosses detected in our study suggests that artificial interpopulation gene flow as a management tool to counter
genetic erosion should be used with caution.
LITERATURE CITED
AIZEN, M. A., AND P. FEINSINGER. 1994. Habitat fragmentation, native
insect pollinators, and feral honey bees in Argentine Chaco Serrano. Ecological Applications 4: 378392.
BAKER, H. G. 1955. Selfcompatibility and establishment after
longdistance dispersal. Evolution 9: 347348.
. 1967. Support for Bakers lawas a rule. Evolution 21: 853
856.
BARRETT, S. C. H., AND D. CHARLESWORTH. 1991. Effects of a change
in the level of inbreeding on the genetic load. Nature 352: 522
524.
BIERZYCHUDEK, P. 1981. Pollinator limitation of plant reproductive effort. American Naturalist 117: 838840.
BOND, W. J. 1995. Assessing the risk of plant extinction due to pollinator and disperser failure. In J. H. Lawton and R. M. May [eds.],
Extinction rates, 131146. Oxford University Press, Oxford.
BRAUNER, S., D. J. CRAWFORD, AND T. F. STUESSY. 1992. Ribosomal
DNA and RAPD variation in the rare plant family Lactoridaceae.
American Journal of Botany 79: 14361439.
BRONSTEIN, I. N., AND K. A. SEMENDJAJEW. 1981. Taschenbuch der
Mathematik. 20th edition. Harri Deutsch, Thun.
BURD, M. 1994. Batemans principle and plant reproduction: the role
of pollen limitation in fruit and seed set. Botanical Reviews 60:
83139.
BYERS, D. L. 1995. Pollen quantity and quality as explanations for low
seed set in small populations exemplified by Eupatorium (Asteraceae). American Journal of Botany 82: 10001006.
CALVO, R. N. 1990. Four-year growth and reproduction of Cyclopogon
cranichoides (Orchidaceae) in South Florida. American Journal of
Botany 77: 736741.
CAMPBELL, D. R., AND N. M. WASER. 1987. The evolution of plant
mating systems: multilocus simulations of pollen dispersal. American Naturalist 129: 593609.
CHARLESWORTH, D. 1988. A method of estimating outcrossing rates in
natural populations of plants. Heredity 61: 469471.
DUDASH, M. R. 1990. Relative fitness of selfed and outcrossed progeny
in a self-compatible, proterandrous species, Sabatia angularis L.
(Gentianaceae): a comparison in three environments. Evolution 44:
11291139.
ELLSTRAND, N. C., AND D. R. ELAM. 1993. Population genetic conse-

December 1997]

FISCHER

AND

MATTHIESEXPERIMENTAL

quences of small population size: implications for plant conservation. Annual Review of Ecology and Systematics 24: 217242.
FISCHER, M. 1996. Experimental population biology of the rare Gentianella germanica. PhD. dissertation, University of Basel, Basel.
HAIG, D., AND M. WESTOBY. 1988. On limits to seed production. American Naturalist 131: 757759
HAUSER T. P., AND V. LOESCHKE. 1994. Inbreeding depression and mating-distance dependent offspring fitness in large and small populations of Lychnis floscuculi (Caryophyllaceae). Journal of Evolutionary Biology 7: 609622.
HEGI, G. 1926. Illustrierte Flora von Mitteleuropa, vol. 5. Lehmanns,
Munchen.
HESCHEL, M. S., AND K. N. PAIGE. 1995. Inbreeding depression, environmental stress, and population size variation in scarlet gilia
(Ipomopsis aggregata). Conservation Biology 9: 126133.
JENNERSTEN, O., J. LOMAN, A. P. MLLER, J. ROBERTSON, AND B. WIDEN.
1992. Conservation biology in agricultural habitat islands. In L.
Hansson [ed.], Conservation biology by ecological principles, 396
425. Elsevier, London.
, AND S. G. NILSSON. 1993. Insect flower visitation frequency
and seed production in relation to patch size of Viscaria vulgaris
(Caryophyllaceae). Oikos 68: 283292.
JOHNSTON, M. O. 1991. Pollen limitation of female reproduction in
Lobelia cardinalis and L. siphilitica. Ecology 72: 15001503 .
KORNECK, D., M. SCHNITTLER, AND I. VOLLMER. 1996. Rote Liste der
Farn- und Blutenpflanzen (Pteridophyta et Spermatophyta)
Deutschlands. Schriftenreihe fur Vegetationskunde 28: 21187.
KUNIN, W. E. 1992. Density and reproductive success in wild populations of Diplotaxis erucoides (Brassicaceae). Oecologia 91: 129
133.
. 1993. Sex and the single mustard: population density and pollinator behavior effects on seed set. Ecology 74: 21452160.
LACY, R. C. 1987. Loss of genetic diversity from managed populations:
interacting effects of drift, mutation, immigration, selection, and
population subdivision. Conservation Biology 1: 143158.
LANDE, R. 1994. Risk of population extinction from fixation of new
deleterious mutations. Evolution 48: 14601469.
. 1995. Mutation and conservation. Conservation Biology 9:
782791
LANDOLT, E. 1991. Gefahrdung der Farn- und Blutenpflanzen in der
Schweiz mit gesamtschweizerischen und regionalen roten Listen.
Bundesamt fur Umwelt, Wald und Landschaft (BUWAL), Bern.
LEVIN, D. A. 1984. Inbreeding depression and proximitydependent
crossing success in Phlox drummondii. Evolution 38: 116127.
LLOYD, D. G., AND D. J. SCHOEN. 1992. Self- and cross fertilization in
plants. I. Functional dimensions. International Journal of Plant Sciences 153: 358369
LYNCH, M. 1991. The genetic interpretation of inbreeding depression
and outbreeding depression. Evolution 45: 622629.
, J. CONERY, AND R. BURGER. 1995. Mutation accumulation and
the extinction of small populations. American Naturalist 146: 489
518
MACNAIR, M. R., AND P. CHRISTIE. 1983. Reproductive isolation as a
pleiotropic effect of copper tolerance in Mimulus guttatus? Heredity 50: 295302.
MENGES, E. S. 1991a. The application of minimum viable population
theory to plants. In D. A. Falk and K.E. Holsinger [eds.], Genetics
and conservation of rare plants, 4761. Oxford University Press,
New York, NY.
. 1991b. Seed germination percentage increases with population
size in a fragmented prairie species. Conservation Biology 5: 185
164.
. 1992. Stochastic modeling of extinction in plant populations.
In P. L. Fiedler and S. K. Jain [eds.], Conservation biology: the
theory and practice of nature conservation, preservation and management, 253276. Chapman and Hall, New York, NY.
. HILL. 1989. Mating
MOLAU, U., M. CARLSSON, A. DAHLBERG, AND O
system and pollenmediated gene flow in Bartsia alpina. Oikos
55: 409419.
OBERDORFER, E. 1994. Pflanzensoziologische Exkursionsflora. 7. Auflage. Ulmer, Stuttgart.
OLESEN, J. M. AND S. K. JAIN. 1994. Fragmented plant populations and

POLLINATION OF A RARE PLANT

1691

their lost interactions. In V. Loeschke, J. Tomiuk, and S.K. Jain

[eds.], Conservation genetics, 417426. Birkhauser, Basel.
OOSTERMEIJER, J. G. B., R. G. M. ALTENBURG, AND H. C. M. DEN NIJS.
1995. Effects of outcrossing distance and selfing on fitness components in the rare Gentiana pneumonanthe (Gentianaceae). Acta
Botanica Neerlandica 44: 257268.
OUBORG, N. J., AND R. VAN TREUREN. 1994. The significance of genetic
erosion in the process of extinction. IV. Inbreeding load and heterosis in relation to population size in the mint Salvia pratensis.
Evolution 48: 9961008.
PETANIDOU, T., H. C. M. DEN NIJS, AND A. C. ELLISADAM. 1991.
Comparative pollination ecology of two rare dutch Gentiana species, in relation to population size. Acta Horticultura 288: 308
312.
,
, AND J. G. B. OOSTERMEIJER. 1995. Pollination ecology
and constraints on seed set of the rare perennial Gentiana cruciata
L. in The Netherlands. Acta Botanica Neerlandica 44: 5574.
,
,
, AND A. C. ELLIS-ADAM. 1995. Pollination
ecology and patch-dependent reproductive success of the rare perennial Gentiana pneumonanthe L. New Phytologist 129: 155163.
PRICE, M. V., AND N. M. WASER. 1979. Pollen dispersal and optimal
outcrossing in Delphinium nelsonii. Nature 277: 294297 .
PRIMACK, R. B., AND P. HALL. 1990. Cost of reproduction in the pink
ladys slipper orchid: a four year experimental study. American
Naturalist 136: 638656.
RAIJMANN, L. E., N. C. VAN LEUWEN, R. KERSTEN, J. G. B. OOSTERMEIJER, H. C. M. DEN NIJS, AND S. B. J. MENKEN. 1994. Genetic
variation and outcrossing rate in relation to population size in Gentiana pneumonanthe. Conservation Biology 8: 10141026.
RUNGE, F. 1963. Die Artmachtigkeitsschwankungen in einem Nordwestdeutschen EnzianZwenkenrasen. Vegetatio 11: 237240.
SAUNDERS, D. A., R. J. HOBBS, AND C. R. MARGULES. 1991. Biological
consequences of ecosystem fragmentation: a review. Conservation
Biology 5: 832.
SCHEMSKE, D. W., B. C. HUSBAND, M. H. RUCKELSHAUS, C. GOODWILLIE,
I. M. PARKER, AND J. G. BISHOP. 1994. Evaluating approaches to
the conservation of rare and endangered plants. Ecology 75: 584
606.
, AND R. LANDE. 1985. The evolution of self-fertilization and
inbreeding depression in plants. II. Empirical observations. Evolution 39: 4152.
SIH, A., AND M. S. BALTUS. 1987. Patch size, pollinator behavior, and
pollinator limitation in catnip. Ecology 68: 16791690.
SLATKIN, M. 1985. Gene flow in natural populations. Annual Review
of Ecology and Systematics 16: 393430.
STEBBINS, G. L. 1957. Selffertilization and population variability in
higher plants. American Naturalist 91: 337354.
TEMPLETON, A. R. 1986. Coadaptation and outbreeding depression. In
M.E. Soule [ed.], Conservation biology: the science of scarcity and
diversity, 105116. Sinauer, Sunderland, MA.
UYENOYAMA, M. K., K. E. HOLSINGER, AND D. M. WALLER. 1993. Ecological and genetic factors directing the evolution of self fertilization. Oxford Surveys in Evolutionary Biology 9: 327381.
VAN TREUREN, R. 1993. The significance of genetic erosion for the
extinction of locally endangered plant populations. PhD. dissertation, University of Groningen, Groningen.
, R. BIJLSMA, N. J. OUBORG, AND M. M. KWAK. 1994. Relationships between plant density, outcrossing rates and seed set in natural and experimental populations of Scabiosa columbaria. Journal
of Evolutionary Biology 7: 287302.
,
,
, AND W. VAN DELDEN. 1993. The significance
of genetic erosion in the process of extinction: IV. Inbreeding depression and heterosis effects caused by selfing and outcrossing in
Scabiosa columbaria. Evolution 47: 16691680.
VERKAAR, H. J., AND A. J. SCHENKEVELD. 1984. On the ecology of short
lived forbs in chalk grasslands: life history characteristics. New
Phytologist 98: 659672.
,
, AND M. P. VAN DER KLASHORST. 1983. The ecology
of short-lived forbs in chalk grasslands: dispersal of seeds. New
Phytologist 95: 335344.
WALLER, D. M., D. M. OMALLEY, AND S. C. GAWLER. 1987. Genetic
variation in the extreme endemic Pedicularis furbishiae (Scrophulariaceae). Conservation Biology 1: 335340.

1692

AMERICAN JOURNAL

WASER, N. M. 1993. Population structure, optimal outbreeding, and

assortative mating in angiosperms. In N.W. Thornhill [ed.], The
natural history of inbreeding and outbreeding: theoretical and empirical perspectives, 173199. University of Chicago Press, Chicago, IL.
, AND M. V. PRICE. 1989. Optimal outcrossing in Ipomopsis
aggregata: seed set and offspring fitness. Evolution 43: 10971109.
, AND
. 1994. Crossing-distance effects in Delphinium
nelsonii: outbreeding and inbreeding depression in progeny fitness.
Evolution 48: 842852.

OF

BOTANY

[Vol. 84

WIDEN, B. 1993. Demographic and genetic effects on reproduction as

related to population size in a rare, perennial herb, Senecio integrifolius. Biological Journal of the Linnean Society 50: 179195.
WRIGHT, S. 1943. Isolation by distance. Genetics 28: 114138.
YOUNG, H. J., AND T. P. YOUNG. 1992. Alternative outcomes of natural
and experimental high pollen loads. Ecology 73: 639647
ZOLLER, H., AND C. WAGNER. 1986. Ruckgang und Gefahrdung von
Mesobromion-Arten im Schweizer Jura. Veroffentlichungen des
Geobotanischen Instituts der ETH Zurich, Stiftung Rubel 87: 239
259.