Population dynamics of a long-lived alpine plant with sexual and clonal reproduction

This thesis focuses on the demography,
reproduction, and pre-dispersal seed
predation of the alpine clonal plant Geum
reptans. In chapter 2, I studied the
population dynamics of G. reptans, thereby
I particularly addressed the relative
importance of sexual vs. clonal reproduction
for population growth in this longlived
species. In chapter 3, the balance
between sexual and clonal reproduction in
plants of different sizes and the frequency
of life-cycle stages in G. reptans were
studied in respect to the environmental
gradients altitude and succession. In
chapter 4, I demonstrated the impact of a
host-specific seed predator on reproducetion
and population growth of G. reptans.
G. reptans is a typical pioneer species of
recently deglaciated areas showing also
high persistence in late successional communities.
Since G. reptans occurs along
broad ranges of altitude and succession, its
environment is mainly characterised by
high natural fragmentation, frequent disturbances,
and sharp gradients of climatic
conditions which may constrain growth
and reproduction. Many plants adapted to
alpine habitats are slow-growing, thereby
showing long life-spans and pronounced
clonal reproduction (Hartmann 1957;
Billings & Mooney 1968; Bliss 1971).
Longevity and particularly the ability for
clonal growth can result in high individual
persistence for centuries or even for
hundreds of years (Steinger et al. 1996;
Molau 1997; Morris & Doak 1998).
Generally, the importance of clonal growth
tends to increase with altitude (Bliss 1971;
Klimeš et al. 1997). Contrary, there is large
evidence that recruitment from seeds
declines with altitude (e.g. Jolls & Bock
1983) supporting the assumption that in
high alpine habitats, local population
dynamics of many species are strongly
dominated by clonal growth (cp. Crawley
1990; Eriksson 1992; Silvertown et al.
1993). However, alpine pioneer species are
characterised by high seed production
(Stöcklin & Bäumler 1996). Thus, sexual
reproduction providing dispersal and
colonisation of new sites must be of
particular importance for a pioneer species
situated in the patchy alpine landscape.
Nevertheless, knowledge about population
dynamics of alpine plants is scarce and the
importance of sexual vs. clonal reproduction
for population growth remains largely
unclear. In a demographic study in two
field populations, I focused on the relevance
of life-cycle stages as well as on the
significance of sexual and clonal reproductive
transitions for population dynamics in
long-lived G. reptans (chapter 2).
Implying a trade-off between sexual
and clonal reproduction in G. reptans
depending on meristem availability, the
balance between both reproductive modes
is likely to vary depending on different
selection pressures. Environmental heterogeneity
is considered to promote
genetic variation in life-history traits
(Stratton 1994; Prati & Schmid 2000) and
large-scale environmental gradients are
well known to result in locally adapted
genotypes (Clausen et al. 1947; Linhart &
Grant 1996). The question, if the
relationship between sexual and clonal
reproduction varies in contrasting
altitudes and successional stages is one of
the main topics in this thesis (chapter 3).
Opposing selection pressures for reproductive
traits is predicted by
metapopulation theory (e.g. Olivieri et al.
1995; Husband & Barrett 1996): In
recently established populations, traits
that enhance dispersal should be favoured
since they are more likely founded by
genotypes with high dispersal ability. With
increasing population age, within-population
selection should act against dispersal
resulting in a decrease in sexual
reproduction whereas clonal reproduction
increases. If the importance of clonal
growth increases with altitude, in high
altitude plants more allocation to clonal
reproduction at the expense of sexual
reproduction is expected. In 20 populations
of G. reptans, I compared the
proportion of clonal reproduction in low
vs. high altitude habitats and in early vs.
late successional habitats and studied the
influence of plant size on reproduction of
this species.
Pre-dispersal seed predation can
substantially influence the relative
reproductive success of individuals by
limiting seed production (e.g. Louda &
Potvin 1995; Briese 2000). Moreover,
evidence is given, that seed consumption
prior to dispersal can limit a species
population growth and therefore may have
severe consequences for its persistence
(e.g. Louda 1982; Kelly & Dyer 2002).
However, effects of pre-dispersal seed
predation on population dynamics are still
poorly understood. A reduction in seed
output may be of particular importance in
high alpine habitats, where recruitment
from seeds can be strongly limited by
harsh environmental conditions (Scherff et
al. 1994; Forbis 2003). Many populations
of G. reptans show heavy predation by
host-specific gall midge larvae feeding on
developing seeds. In an exclusion experiment,
I used insecticide to study the consequences
of pre-dispersal seed predation on
reproduction in a population of G. reptans
(chapter 4). Since this species does not
form a persistent seed bank and immigration
of seeds from other sites is negligible
(Tackenberg & Stöcklin, unpublished),
impacts of seed loss are expected to
directly affect recruitment. I used a greenhouse
germination experiment to test for
seed viability due to predation and matrix
modelling allowed predicting consequences
for population growth based on
the observed seed loss.
Population dynamics of Geum
reptans:
Population dynamics of Geum reptans are
mainly characterised by slow yearly increase
and longevity which are typical features
of many alpine plants as well as of
pioneer species of primary successions.
Despite considerable variation in growth
rates (�) among sites and years, dynamics
in both populations were very similar.
However, a slightly negative � in one year
out of four mainly resulting from higher
adult mortality indicates that population
dynamics of slow-growing G. reptans are
likely to be sensitive to environmental stochasticity.
Sexual and clonal reproduction
occurred regularly in both populations
during the study period but the frequency
of reproducing adults was always low. This
was demonstrated for populations of G.
reptans situated in different environments
(chapter 3), suggesting that yearly variable
availability of resources for reproduction
and the cost of reproduction may cause
annually variation in reproductive output
similar to mast seeding in trees (Webb &
Kelly 1993). Interestingly, in G. reptans
sexual reproduction plays an equally
important role for population growth as
the reproduction by stolons. Thus, my
results clearly contradict the general
assumption, that the role of sexual
reproduction in clonal plants from high
altitudes is mainly restricted to the
maintenance of genetic variation and longdistance
dispersal. Nevertheless, the two
reproductive modes may have different
significance for population dynamics: The
constantly higher establishment of stolons
compared to seeds suggests that clonal
propagation serves as a low-risk alternative
compared to more unpredictable and
unsecured establishment from seeds. Reproduction
by stolons alone may guarantee
continuously slow population growth
leading to a positive � and thereby assuring
persistence of the population even in
unfavourable periods. Although germination
occurred regularly in both studied
populations, percent germination varied
by nearly 200% among years implying that
seed recruitment in G. reptans may be
particularly susceptible to environmental
variation. On the other hand, in favourable
years, reproduction by seeds may be an
important contributor to population
growth.
In G. reptans, population growth rate
was much more sensitive to changes in the
survivorship of adults than to growth or to
reproductive parameters confirming the
importance of persistence. Long-lived species
living in stressful environments and
exhibiting low recruitment rates may be
particularly vulnerable to adult mortality.
Since G. reptans as a species of glacier
forelands is highly subject to changes in
habitat conditions, it is a reasonable
strategy for this species to privilege
investments to maintenance of adults at
the expense of growth and fecundity. Thus,
adult longevity is expected to serve as a
buffer against temporal variation which
may be of particular importance in
frequently disturbed alpine habitats. My
results indicate that even in harsh alpine
environments, population growth of a
long-lived species does not rely on clonal
reproduction alone.
Variation of sexual and clonal
reproduction:
In Geum reptans, the relationship between
sexual and clonal reproduction largely
varied among populations. However, the
variation in reproductive behaviour could
only be explained to a small degree by
contrasting altitudes or successional
stages. There may be two possible explanations
for this result: Firstly, reproductive
differences between populations may
reflect mainly small-scale variation in
environmental conditions not related to
altitude or succession overriding the
effects of gradients. Secondly, high phenotypic
plasticity may restrict or even
prevent adaptation to contrasting environmental
conditions. Highly plastic genotypes
have been demonstrated for species
dispersed over broad geographic and environmental
ranges showing little genetic or
morphological differentiation (Williams et
al. 1995; Hermanutz & Weaver 1996).
Accordingly, G. reptans as a colonising
species may have the capacity of broad
plastic responses allowing rapid spread
into new habitats without experiencing
adaptation through selection. Additionally,
high individual variation in stolon and
flower head production could add to
population variability which was also
detected in chapter 2.
Nevertheless, in comparison with intermediate
populations, clonal reproduction
increased in populations at high altitude as
well as at low altitude which was partly in
accordance with my expectations. The
higher frequency of stolons at high
altitudes demonstrates that abiotic environmental
conditions may restrict seedling
establishment leading to a higher fraction
of clonal reproduction. At low altitudes,
habitats are often characterised by increased
inter-specific competition and G.
reptans is very susceptible to competition
(Pluess & Stöcklin 2005). Additionally,
seedlings lacking the support of adults in
contrast to vegetative off-spring may be
particularly vulnerable for crowding what
might explain the higher proportion of
stolons also found in low altitude habitats.
However, sensitivity analysis of matrix
projection models can reveal the direction
and intensity of selection on life-history
characteristics of species (van Groenendael
et al. 1988). Since sensitivity of � was
highest to stasis of small adults and very
low to reproductive transitions (chapter 2),
it is expected that in G. reptans, traits
promoting persistence may be stronger
under selection compared to reproductive
behaviour. In a common garden experiment
studying the effects of population
origin and environment in G. reptans, in
spite of high variation, differences in
reproductive behaviour were not explained
by population origin from successional
gradients supporting the assumption that
adaptation in reproductive behaviour to
contrasting habitats is limited in this
species (Pluess & Stöcklin 2005). My
results suggest that the observed variation
among populations in the frequency of
clonal reproduction largely results either
from differences in local environmental
conditions and probably also from random
drift among populations.
Another factor altering the relationship
of sexual and clonal reproduction was
plant size. In G. reptans, the number of
reproductive meristems as well as the
probability of producing flower heads and
stolons simultaneously increased with the
number of rosettes. Accordingly, small
plants being rather limited in resources
invested preferentially in low-cost stolon
production. Larger plants having more
resources available invested into both
reproductive modes thereby tending to
prefer sexual reproduction confirming
once more the importance of sexual reproduction
for a clonal alpine plant (chapter
2).
Pre-dispersal seed predation:
Pre-dispersal seed predation by gall midge
larvae heavily decreased seed mass in
Geum reptans causing seed viability to be
substantially reduced (-98%). However,
matrix modelling revealed that the population
growth rate (�) was only slightly
affected by the observed seed loss
suggesting that persistence of G. reptans
does not depend on seed production alone.
In natural habitats, sexual reproduction
and seedling establishment are very likely
to show high inter-annual variation
depending on environmental conditions
and individual variation in reproduction
(chapter 2). While high seed production
can rapidly increase population growth,
low seed availability does not necessarily
lead to a negative population growth since
continuous clonal reproduction can balance
potential seed losses. Accordingly,
population growth rate would only
decrease as a result of a general lack of
reproduction. However, since elasticity
analysis revealed that � is most sensitive to
survival of adults, limitation in sexual
reproduction by seed predation is not
expected to have more than only slight
effects on population growth. Thus, the
local abundance of G. reptans does not
critically depend on seed supply suggesting
that population dynamics is probably
not limited by pre-dispersal seed predation
which has also been shown for other
perennial species (Andersen 1989; Fröborg
& Eriksson 2003). Nevertheless, predation
could be an important factor generating
differences in the reproduction of individuals,
thereby decreasing effective
population size, and thus facilitating
genetic change, particularly genetic drift
(Crawford 1984; Heywood 1986).
Surprisingly, seed predation resulted in
an increase in stolon dry-weight in predated
plants suggesting a change in
resource allocation due to less resources
used for seed ripening after pre-dispersal
seed predation favouring clonal reproduction.
However, as leafy stolons of G.
reptans are expected to mainly support
themselves, it seems unlikely that a higher
stolon mass may increase clonal establishment
and therefore directly compensate
for reduced sexual establishment due to
seed loss.
Although pre-dispersal seed predation
by host-specific gall midge larvae did not
limit population growth of G. reptans, the
production of viable seeds was substantially
reduced. Therefore, infested
populations may have reduced amounts of
seeds available for dispersal and colonisation
of new sites which may be of
particular relevance in the naturally
fragmented alpine landscape.
Conclusions
In summary, my results demonstrate that
sexual and clonal reproduction both have
an equally important role for population
dynamics of the high alpine plant Geum
reptans. Reproduction by seeds acts as
mechanism promoting not only genetic
diversity and dispersal but also rapid
population growth in favourable years.
Clonal reproduction serves to ensure
population increase by slow but
continuous growth even under distinct
environmental variability and unfavourable
weather conditions. The relative
importance of sexual and clonal reproduction
is highly variable among
populations of G. reptans. However, the
observed variation could only be explained
to a small degree by contrasting habitats. A
higher frequency of clonal reproduction
occurred in populations from high and low
altitudes in comparison with intermediate
populations. Taken together, my results
suggest that variation in reproduction
could either be attributed to individual
plasticity in response to different habitat
conditions or resulted from genetic
differentiation among populations, probably
partly random in nature. Plant size is
suggested to be an additional factor
influencing the relationship between
sexual and clonal reproduction demonstrating
a higher threshold size for
investment in sexual reproduction.
The outstanding importance of adult
survival for the life-history of G. reptans
highly exceeding the impact of reproduction
demonstrates the significance of
longevity and persistence for this pioneer
species of glacier-forelands. In long-lived
G. reptans, adult survivorship is expected
to serve as a buffer against temporal
variation in reproduction that may be
particularly important in frequently disturbed
alpine habitats.
The importance of persistence and the
ability for clonal growth balancing reduced
seedling recruitment explain why, despite
substantial seed losses, population growth
of G. reptans might not be limited by predispersal
seed predation. In G. reptans as
a long-lived perennial clonal plant being
most vulnerable to adult mortality, seed
supply is of minor importance for population
dynamics. In unpredictable habitats
where successful recruitment by seeds may
not occur every year, this might be an
important adaptation. However, seed
predation by reducing the proportion of
viable seeds restricts successful colonisation
of new sites, thereby contributing to
the not very high frequency of this species
in the alpine landscape and reducing geneflow
among populations.