Content uploaded by Eva Hernandez Plaza
Author content
All content in this area was uploaded by Eva Hernandez Plaza on Oct 16, 2014
Content may be subject to copyright.
Content uploaded by Eva Hernandez Plaza
Author content
All content in this area was uploaded by Eva Hernandez Plaza on Oct 16, 2014
Content may be subject to copyright.
Is reproductive allocation in Senecio vulgaris
plastic?
Jacob Weiner, Lars Rosenmeier, Emma Soy Massoni, Josep Nogue
´s Vera,
Eva Herna
´ndez Plaza, and Maria-Teresa Sebastia
`
Abstract: Several purported cases of plasticity in plant allocation patterns appear to be the effects of size and allometric
growth (‘‘apparent plasticity’’). To ask whether there is true plasticity (i.e., a change in the allometric trajectory) in repro-
ductive allocation in Senecio vulgaris L., we grew S. vulgaris plants at high and low levels of water, nutrients, and compe-
tition, and analyzed the relationship between vegetative and seed biomass. Plant size was the major determinant of
reproductive output, accounting for 83% of the variation in log (seed mass). There were also significant effects of the treat-
ments that were not due to size, accounting for an additional 9% of the variation. The treatments affected the allometric
coefficient (intercept), not the allometric exponent (slope) of the relationship, reflecting a small but significant shift in the
efficiency of conversion of total plant biomass into reproductive biomass. In a second experiment, we grew S. vulgaris
plants at three nutrient levels and allowed all individuals to complete their life cycles. Again, nutrient level had a small
but significant effect on the allometric coefficient. Plasticity in reproductive allocation exists, but is very limited. The pri-
mary effects of the environment on the reproductive output of S. vulgaris occur via plant size.
Key words: allometry, common groundsel, plasticity, size effects.
Re
´sume
´:Plusieurs cas mis de l’avant de plasticite
´dans les patrons d’allocation semblent re
´sulter de la dimension et de la
croissance allome
´trique (« plasicite
´apparente »). Afin de ve
´rifier s’il existe s’il existe une vraie plasticite
´(i.e. un change-
ment de la trajectoire allome
´trique) dans l’allocation reproductive chez le Seneciao vulgaris L. les auteurs ont cultive
´cette
plante a
`des degre
´se
´leve
´s et faibles en eau, nutriments et compe
´tition et ils ont analyse
´la relation entre la biomasse ve
´ge
´-
tative et se
´minale. La dimension des plants constitue le de
´terminant majeur du produit de la reproduction, expliquant 83 %
de la variation en log (masse se
´minale). On observe e
´galement des effets significatifs des traitements qui ne de
´pendent pas
de la dimension expliquant un 9 % additionnel de la variation. Les traitements affectent le coefficient allome
´trique (inter-
ception), et non l’exposant allome
´trique (pente) de la relation, ce qui refle
`te un de
´placement faible, mais significatif de
l’efficacite
´de la conversion de la biomasse totale de la plante en biomasse reproductive. Dans une deuxie
`me expe
´rience,
les auteurs ont cultive
´des plants de S. vulgaris en pre
´sence de trois teneurs en nutriments et ont permis a
`tous les indivi-
dus de comple
´ter leurs cycles vitaux. Encore une fois la teneur en nutriments exerce un effet faible, mais significatif sur le
coefficient allome
´trique. Il existe une plasticite
´dans l’allocation reproductive, mais elle est tre
`s limite
´e. Les effets primai-
res de l’environnement sur la productivite
´reproductive du S. vulgaris s’effectuent via la dimension de la plante.
Mots-cle
´s:allome
´trie, senecio vulgaire, plasticite
´, effets des dimensions.
Introduction
Growth and reproduction are two of the most fundamental
processes for plants. Reproduction is the currency of natural
selection, but plants must grow to build the machinery to re-
produce. After a plant produces biomass, it allocates this bi-
omass to different functions and structures, including
reproductive structures (Bazzaz and Reekie 1985). Because
resources allocated to one function or organ are not avail-
able to other functions or organs, allocation implies trade-
offs. Allocation patterns reflect evolved strategies that are
the results of different selection pressures and constraints
(Bonser and Aarssen 2001).
There is an emerging consensus among plant ecologists
that allocation patterns, which we originally conceptualized
and analyzed as ratios (e.g., root:shoot ratio) or percent allo-
cation (e.g., ‘‘reproductive effort’’, defined as the percent of
total biomass in reproductive structures; Tuomi et al. 1983)
are better understood and analyzed allometrically (Jasienski
and Bazzaz 1999; Weiner 2004). Plant allocation is usually
allometric in the broad sense: it changes with size as plants
grow. Therefore any factor that affects growth rate and
therefore size will inevitably affect allocation, even if plant
allocation at a given size is totally fixed (Coleman et al.
1994). Effects on allocation solely due to size and allometric
growth have been referred to as ‘‘apparent plasticity’’
(Fig. 1a; McConnaughay and Coleman 1999; Weiner 2004),
Received 3 December 2008. Published on the NRC Research
Press Web site at botany.nrc.ca on 6 May 2009.
J. Weiner1and L. Rosenmeier. Department of Agriculture and
Ecology, University of Copenhagen, DK-1958 Frederiksberg,
Denmark.
E.S. Massoni, J.N. Vera, and M. Sebastia
`.University of
Lleida, Plac¸a Vı
´ctor Siurana 1, 25003 Lleida, Spain.
E.H. Plaza. UdL-IRTA Centre Foundation (Rural Development
and AgriFood Research Institute), Finca ‘‘El Encı
´n.’’ Autovı
´a
A2, Km. 38,2. Alcala
´de Henares, Madrid, 28800 Spain.
1Corresponding author (e-mail: jw@life.ku.dk).
475
Botany 87: 475–481 (2009) doi:10.1139/B09-012 Published by NRC Research Press
since true plasticity implies a change in an allometric
growth trajectory (Fig. 1b), not only the speed at which a
single trajectory is followed. Several patterns that were pre-
viously considered evidence for plasticity in allocation have
been shown to be due to apparent plasticity (Coleman and
McConnaughay 1995; Wright and McConnaughay 2002),
whereas other plant traits show extensive true plasticity.
Understanding which plant behaviors are plastic and which
are fixed would be a major advance for plant ecology and
evolution (Pigliucci and Preston 2004; Weiner 2004).
Here we address the simple and fundamental question: is
reproductive allocation plastic, i.e., does the allometric rela-
tionship between reproductive and vegetative biomass
change in different environments, or can the effects of the
environment on allocation be understood in terms of size
and a fixed allometric growth pattern? We grew the com-
mon cosmopolitan weed Senecio vulgaris L. under different
levels of nutrients, water, and competition, and asked
whether the relationship between seed mass and above-
ground vegetative mass is affected by these treatments.
Materials and methods
Senecio vulgaris (Asteraceae) is a cosmopolitan annual
herbaceous weed that grows up to 40 cm tall, has a thick
taproot, and possesses an ephemeral strategy typical of
many weedy species. A rosette phase of vegetative growth
is followed by a more or less continuous period of seed pro-
duction until the plant dies. There is, however, considerable
variation in the pattern and timing of development. Senecio
vulgaris was the subject of some of the earliest research on
biomass allocation (Harper and Ogden 1970). This species
shows very high rates of self-fertilization, and the degree of
outcrossing is highly variable (Hull 1974).
We grew S. vulgaris plants in pots in the greenhouse of the
Faculty of Life Sciences, University of Copenhagen, Freder-
iksberg, Denmark. Field-collected seed was purchased from
Herbiseed (New Farm, Mire Lane, Twyford, Berkshire, UK).
Experiment 1
Plants were grown in 7.5 L pots filled with a mixture (v/v)
of 15% vermiculite, 10% perlite, and 75% Pindstrup ‘‘Fær-
digblandining No. 1’’ (sphagnum with nutrients and lime,
pH 6.8; Pindstrup Mosebrug A/S, DK-8550 Ryomgaard,
Denmark). Several seeds were sown in each pot on 23 Feb-
ruary 2006, and after germination they were randomly
thinned to one or two individuals per pot, depending on the
treatment. There were three factors in a full factorial design:
(i) with and without competition; (ii) high vs. low nutrient
(fertilizer); and (iii) high vs. low water level, giving eight
treatment combinations. There were 12 replicate pots for all
treatments without competition, and 4 replicate pots for all
treatments with competition (see description of the competi-
tion treatment below). Thus, there were a total of 80
S. vulgaris individuals in 64 containers. Pots were arranged
in 12 blocks. Each block included all treatments without
competition, but only one or two of the four treatments
with competition. Supplementary lighting was used to keep
the day length at 18 h.
In the high water level treatment we kept the soil in the
pots moist with osmotically demineralized water. In the low
water treatment, pots were allowed to dry out. When the soil
felt completely dry and leaves began to wilt, we added a
small amount of water. One hundred millilitres was added
at the beginning of the experiment, but as plants grew and
transpiration demands increased, we increased the amount
of water added. But water was only added after some wilt-
ing occurred.
For the high nutrient treatment we added 5 mL of liquid
fertilizer [Hornum Pioner Næring, 5-1-4 + micronutrients
(Brøste A/S, DK-2800 Lyngby, Denmark) in 500 mL water]
once a week. In the low nutrient treatment, no fertilizer was
added to the water throughout the course of the experiment.
To avoid confounding of nutrient and water treatments, we
added 500 mL of water to the low water – low nutrient
treatment whenever we added fertilizer to the low water –
high nutrient treatment.
Fig. 1. Hypothetical example of (a) apparent vs. (b) true plasticity in reproductive allocation in response to nutrient level. In apparent plas-
ticity (H0), changes in allocation are due to the rate of growth along a fixed allometric trajectory. In true plasticity (broken lines), the slope
(H1) or the intercept (H2) of the trajectory is altered by nutrient level.
476 Botany Vol. 87, 2009
Published by NRC Research Press
In the competition treatment, two Centaurea cyanus L.
plants grew in the pots together with two S. vulgaris plants.
This was part of a larger design to look at both species, but
few C. cyanus individuals flowered and very few seeds were
produced, so results are presented only for S. vulgaris.
We collected all fruits produced by S. vulgaris plants as
they matured. We also collected dead plant parts as they
dried. All plants were harvested at the soil level on 17 May
2006, dried at 70 8C, and weighed. Fruits were dried at
25 8C for one day and weighed.
Experiment 2
Because plants were still growing and flowering when
harvested, we performed a second experiment in which
S. vulgaris plants were grown individually in small pots
with only one factor (nutrient level); this allowed for the
completion of the life cycle. Individual S. vulgaris plants
were grown in 0.5 L pots at three different nutrient levels.
The high nutrient level consisted of 100% Pindstrup ‘‘Fær-
digblandining No. 2’’ (Pindstrup Mosebrug A/S, DK-8550
Ryomgaard, Denmark), a commercial greenhouse sphagnum
medium similar to that used in the first experiment but with
higher nutrient levels. The middle nutrient level was 25%
perlite, 25% vermiculite, and 50% of the commercial me-
dium (v/v), and the low nutrient mixture 40% perlite, 40%
vermiculite, and 20% of the commercial medium (v/v).
There were 15 replicates of each, giving at total of 45 pots.
Pots were organized on the greenhouse bench in groups of
three, one of each treatment. Several S. vulgaris fruits were
sown in each pot on 7 February 2007. Seedlings were
thinned to three individuals one week later, and then one in-
dividual 2 weeks later. Plants were watered regularly and
the soil kept moist. No additional nutrients were added
throughout the course of the experiment, with one excep-
tion: owing to an error, the pots were watered once on 31
March with a very low concentration of the fertilizer solu-
tion used for regular watering in the greenhouse. All pots
were heavily watered with deionized water immediately
afterwards to minimize any effects. Mature fruits and dead
parts were collected throughout the course of the experi-
ments. All plants were dead by 5 June and the remaining
plants were harvested on that day. Plant material and seeds
were dried as in the first experiment. We attempted to har-
vest roots by separating them from the rooting medium, but
we were not satisfied with our ability to do so.
Data and statistical analyses
We avoid the issue of defining structures with multiple
functions as vegetative or reproductive (Bazzaz and Reekie
1985) by taking a very strict view of reproductive alloca-
tion: We consider total biomass of achenes produced by an
individual as reproductive output (R) and all other biomass
as vegetative (V). The values for both of these were log
transformed. Total plant mass was also log transformed to
homogenize variances for the analysis of total plant biomass
in experiment 2.
The effect of treatments on total plant biomass was ana-
lyzed using ANOVA. Regression analysis was used to assess
the relationship between log Rand log Vin both experiments
and to ask (i) whether the treatments had significant effect
on the log R– log Vrelationship, and (ii) if so, how much
additional variation in log Rcan their inclusion account for.
In addition to analyzing each experiment independently, we
also analyzed data pooled from both experiments to the de-
gree possible given the different designs. In the combined
analysis, the eight treatment combinations in experiment 1
and the three levels of nutrient in experiment 2 were nested
within ‘‘experiment.’’
There is no general agreement among researchers con-
cerning the best regression methods to analyze allometric re-
lationships (Riska 1991). While reduced major axis and
other orthogonal methods have the advantage of assuming
error in both variables, they also present a series of disad-
vantages, especially in an experimental context with treat-
ments and potential interactions. Like most researchers, we
have used standard least-squares general linear models, in
which log Ris the dependent variable, log Vis an independ-
ent variable, and treatments are nominal independent varia-
bles (factors). When analyzing only log Rvs. log V, without
treatment effects, we also used reduced major axis regres-
sion. The differences between this and standard regression
were very small. All analyses were performed with JMP
statistical software (SAS Institute Inc. Cary, N.C.).
Results
Experiment 1
All three factors (nutrient level, water level, and competi-
tion) and all their interactions had significant effects on the
biomass of S. vulgaris individuals in experiment 1 (Table 1).
Low water level, low nutrient level, and competition all re-
duced plant size (Fig. 2).
Two extreme outliers were removed from the regression
analyses of Rvs. V, leaving 78 plants. Most of the variation
in log Rcould be accounted for by variation in log V
(Fig. 3). Simple least squares regression fit of log Ron
log Vwithout inclusion of any factors was log R= –0.37 +
0.906 log V,r2= 0.827. The least squares slope was margin-
ally significantly different from 1 (P< 0.05), but an orthog-
onal regression assuming variance in both log Rand log V
gave a slope of 0.995. When all three factors (nutrient level,
water level, and competition), as well as log Vwere included
in the analysis, they were all significant, but there were no
significant interactions among them. Water limitation and
competition had highly significant negative effects on log R
(water: SS = 0.343; df = 1; F= 38.9; P< 0.001. Competi-
tion: SS = 0.358; df = 1; F= 40.6; P< 0.001), whereas low
nutrient level had a marginally significant positive effect
(SS = 0.044; df = 1; F= 4.90; P= 0.03). The complete
model including log Vand all three factors had an adjusted
r2= 0.915. Thus, the inclusion of the three treatment factors
in addition to log Vincreased r2by 0.088.
Experiment 2
Fertility level had a highly significant effect on log total
plant biomass in a one-way ANOVA (SS = 4.09; df = 2;
F= 288.7; P< 0.0001). While the difference between log
biomass (in mg) at low and medium fertility levels (2.58
and 2.69, respectively) was quite small, plants grown at the
highest fertility level were much larger (log biomass =
3.27). In least square means post-hoc tests on log total plant
Weiner et al. 477
Published by NRC Research Press
biomass, all fertility levels were highly significantly differ-
ent from each other (P< 0.0001).
Simple regression of log Ron log Vhad an r2= 0.915. In
an analysis with log Vand nutrient level as independent var-
iables, both were significant (log V, SS = 0.443; df = 1; P<
0.0001. Nutrient level: SS = 0.061; df = 2; P= 0.01), with
r2= 0.956 (Fig. 4).
Many plants had stopped reproducing and died by the be-
ginning of May, 3 months after sowing, but some individu-
als produced new branches with flowers, resulting in a
second wave of reproduction (Fig. 5).
Both experiments combined
In an analysis of the data from both experiments together,
least squares regression of log Ron log Vwas log R= –0.57 +
1.026 log V;r2= 0.971 (Fig. 6). In a more complete analysis
with log V, experiment and treatment (factors) nested in ex-
periment, all three variables were highly significant (log V:
SS = 0.801; df = 1; P< 0.0001. Experiment: SS = 0.430;
df = 1; P< 0.0001. Treatment: SS = 0.808; df = 9; P<
0.0001), and r2= 0.981. There was no significant interaction
between log Vand experiment on log R.
Discussion
Within and across both experiments, most of the variation
in log Rcould be accounted for by variation in log V. In ad-
dition, there were small but significant effects of the treat-
ments on log Rin both experiments, but no evidence of
interactions among the treatments, or between the treatments
and log V. In other words, the treatments shifted the location
(i.e., intercept, also called the allometric coefficient) of the
allometric relationship between Rand Vslightly, but not the
slope or the form of the relationship. The absence of signifi-
cant interactions between log Vand the treatments is consis-
tent with the null hypothesis that the treatments do not affect
the allometric slope of the R–Vrelationship. Inclusion of
factors in addition to log Vin the analyses improved the ex-
planatory power of the statistical models by 1% to 9%. This
means that there were small but significant additive effects
of the treatments on reproductive output in addition to those
mediated by size, but size-mediated effects of resources on
reproductive output are much larger than non-size-mediated
effects.
The slope of 1.0 for regression of log Ron log Vin the
combined data set suggests that S. vulgaris has a very sim-
Fig. 2. Mean total biomass of Senecio vulgaris individuals for the
different treatments in experiment 1.
Fig. 3. Relationship between mass of seeds produced by Senecio
vulgaris individuals and their vegetative biomass. Filled symbols, no
competition; open symbols, competition; squares, water limitation;
circles, no water limitation; black, low nutrients; gray, high nutrients.
Single least-squares regression line (shown): log R= –0.37 + 0.906
log V,r2= 0.83.
Table 1. ANOVA of total plant mass of Senecio vulgaris indivi-
duals on the three factors in experiment 1.
Source df SS FP
Water 1 1781.4 154.3 <0.0001
Nutrients 1 395.8 34.3 <0.0001
Competition 1 4523.9 391.7 <0.0001
Waternutrients 1 135.9 11.8 0.0010
Watercompetition 1 746.5 64.6 <0.0001
Nutrientscompetition 1 93.6 8.1 0.0058
Waternutrientscompetition 1 83.2 7.2 0.0091
478 Botany Vol. 87, 2009
Published by NRC Research Press
ple reproductive allocation strategy with relatively constant
‘‘reproductive effort’’, defined as R/(V+R) (Tuomi et al.
1983; Bazzaz and Reekie 1985), over a large range of ma-
ture sizes. There was no evidence of a minimum size for re-
production. This is consistent with the opportunistic, weedy
‘‘r-selected’’ strategy of S. vulgaris, which emphasizes early
reproduction. There were small but significant changes in
this overall reproductive effort owing to resource levels. For
example, almost all the points in the high fertility treatment
in experiment 2 lie above the overall line from both experi-
ments (Fig. 6, solid circles).
The observed changes in the log R– log Vrelationship can
perhaps be better interpreted in terms of constraints than
adaptations (Weiner 1988). For example, water stress and
competition in experiment 1 and lower nutrient levels in ex-
periment 2 not only reduced plant size, but also slightly re-
duced reproductive output at a given size compared with
individuals growing in more favorable environments. Simi-
larly, competition resulted in smaller plants, and these plants
produced slightly less reproductive biomass at a given size
than did plants growing with less competition, as observed in
a study on Plantago major growing at three densities (Weiner
2004). Plants grow as much as their environment permits, and
they allocate resources to reproduction according to a rela-
tively fixed allometric trajectory. In S. vulgaris, resource lev-
els can alter the overall efficiency of this conversion (i.e., the
allometric coefficient) slightly, but they do not seem to affect
the form or slope of the relationship. Size and the allometric
relationship between vegetative mass and reproductive output
determine a plant’s potential reproductive output. The rate of
development determines whether or not this potential is
achieved, and variation in the rate or stage of development
has been misinterpreted as variation in allometry (Clauss and
Aarssen 1994; J. Weiner, L.G. Campbell, J. Pino, and
L. Echarte, unpublished data).
The observed effects of resource levels on the intercept
of the R–Vrelationship would result if the concentration of
different resources is more constrained in reproductive tis-
sues than in vegetative tissues. For example, if a resource
(e.g., nitrogen) is limiting plant growth, that resource may
be in lower concentration in the plant’s tissues than if it
were more available. If the concentration of this resource
in the seeds is physiologically constrained and therefore
cannot be lowered to the same extent in the seeds produced
as in vegetative parts, then the conversion of vegetative bi-
omass into seeds will be at a lower overall rate when the
resource is limiting than if the resource were more abun-
dant. This would be reflected as change in the allometric
coefficient, as observed here not the exponent. Our results
for reproductive allocation are similar to those on root vs.
shoot allocation in a study with 27 herbaceous species
(Mu
¨ller et al. 2000), in which nutrient level generally had
no effect on root–shoot allometry or altered the allometric
coefficient.
The one result that does not seem to fit this explanation is
the slight and marginally significant increase in reproductive
allocation (i.e., the allometric coefficient) at low fertility in
experiment 1 (solid vs. gray symbols in Fig. 3). We attribute
this to a ‘‘postponement effect’’ with increased nutrient lev-
els. The higher fertility level resulted in more growth and
more potential reproductive output, but it also postponed re-
production, and the experiment was concluded before this
potential could be realized. This interpretation is supported
by the results of the second experiment, in which all plants
were allowed to complete their life cycles. Here there was a
positive effect of increased nutrient levels on the R–Vallo-
metric coefficient. The allometric approach focuses on size
rather than time, but most studies collect data at given
points in time (Coleman et al. 1994) so the available data
are often far from ideal for allometric analyses. Effects of
treatments on the timing of reproduction have been a major
problem in analyzing and interpreting allometric data in
studies that do not include the whole life cycle, and that is
almost all studies.
Our analyses suggest that the factors determining a plant’s
size are much more complex than those determining its po-
tential reproductive output, given the size achieved. Plant
size in the first experiment was affected by all factors and
all interactions among the factors were significant (Table 1).
We do not yet understand how different resources and fac-
tors interact to determine plant growth. The allometric rela-
tionship between Rand V, however, was only slightly
affected by the factors and there were no significant interac-
tions. Thus, the primary effects of the environment on a
plant’s reproductive output occur via plant size. Other ef-
fects are relatively small, although they could still be impor-
tant under some circumstances.
Allocation to reproduction may be one of the least plastic
of allocation patterns. Patterns of root vs. shoot allocation
are functional, and plants can grow more by changing this
allocation pattern (Gedroc et al. 1996). Reproduction in an
annual plant involves the conversion of the maximum
Fig. 4. Relationship between the mass of seeds produced by Sene-
cio vulgaris individuals and their vegetative biomass in the experi-
ment 2. Circles, low nutrient level; triangles, middle nutrient level;
squares, high nutrient level. There was no difference between the
low and medium nutrient levels in the least squares regression, but
the estimate of the intercept was significantly different at the high-
est fertility level (lines shown).
Weiner et al. 479
Published by NRC Research Press
amount of accumulated resources, which is reflected in its
vegetative biomass, into reproduction. Given a plant’s size
and reproductive system, the R–Vrelationship, if allowed to
fully unfold, represents the limiting condition of how much
the plant can reproduce, given its resources.
Although seed quality can be as important as seed number
for plant reproduction, it is not clear how seed quality can
be quantified and thus included in quantitative models of re-
productive allocation. Most studies of plant fitness use seed
production as the estimate of fitness, as we do here.
Most taxonomic sources describe S. vulgaris as mono-
carpic because it is an annual. This is a misconception. For
the length of its life cycle, S. vulgaris starts reproducing
quite early. In the second experiment most plants germi-
nated around 12 February and most were dead by 24 May,
giving an average life span of around 102 d. Most plants
were flowering by 9 April, 56 d after emergence, and con-
tinued to do so until they died. Thus, plants were reproduc-
ing for around 45% of their lifespan, which can hardly be
considered a monocarpic strategy. The opportunistic, annual,
weedy strategy of S. vulgaris (Ollerton and Lack 1992;
Leiss and Muller-Scharer 2001), growing as long and as
much as the environment permits, means that seeds pro-
duced when an individual is young/small can germinate in
the same growing season, thus fitting the demographic defi-
nition of iteroparous. Annuals should be considered mono-
carpic if they reproduce only towards the end of their lives.
The allometric approach focuses on size rather than time,
but a more complete understanding of plant behavior must
encompass size, time, and development. There was one ma-
jor wave of reproduction in which all plants participated in
experiment 2 (Fig. 5). Those individuals that continued to
produce flowers and fruits after this period did not do so by
extending this wave, but rather by producing a subsequent
Fig. 5. Temporal distribution of mature fruiting heads over time in experiment 2. Black bars, high fertility; dark gray, medium fertility; light
gray, low fertility.
Fig. 6. Relationship between mass of seeds produced by Senecio
vulgaris individuals and their vegetative biomass for both experi-
ments combined. Circles indicate data from experiment 2, all other
data are from experiment 1. Single regression line (shown):
log R= –0.57 + 1.026 log V;r2= 0.971.
480 Botany Vol. 87, 2009
Published by NRC Research Press
smaller wave of new shoots with flowers. There are several
possible explanations for this pattern, which is often ob-
served when plants are resource limited. One possibility is
that the number of flowers (or in the case of S. vulgaris, in-
florescences) to be produced by a shoot is determined before
flowering starts, and the plant cannot adjust this upwards
later (e.g., Inouye 1986). Individuals that have resources re-
maining after the first wave of reproduction is complete
must make new shoots and flowers if they are to reproduce
more. Alternatively, this behavior may be part of a bet-
hedging strategy by relatively robust individuals to invest in
a later reproductive episode rather than simply extending the
current one. Such behavior can be seen as a step in the di-
rection of perenniation, in which a generally senescent indi-
vidual produces new ‘‘youthful’’ shoots. The size a plant
achieves determines in large part its potential reproductive
output, but the allometric relationship between reproductive
output and size does not tell us the course of this reproduc-
tion over time.
Acknowledgements
Funding was provided by the Danish Natural Science Re-
search Council (grant No. 21-04-0421) and the European
Region Action Scheme for the Mobility of University Stu-
dents (ERASMUS) Program. We thank Elze Astrup and
Mai-Britt Sauer for help with the experiments and two anon-
ymous reviewers for helpful comments on the manuscript.
References
Bazzaz, F.A., and Reekie, E.G. 1985. The meaning and measure-
ment of reproductive effort in plants. In Studies on plant demo-
graphy. Edited by J. White. Academic Press, London, UK.
pp. 373–387.
Bonser, S.P., and Aarssen, L.W. 2001. Allometry and plasticity of
meristem allocation throughout development in Arabidopsis
thaliana. J. Ecol. 89: 72–79. doi:10.1046/j.1365-2745.2001.
00516.x.
Clauss, M.J., and Aarssen, L.W. 1994. Patterns of reproductive ef-
fort in Arabidopsis thaliana: confounding effects of size and de-
velopmental stage. Ecoscience, 1: 153–159.
Coleman, J.S., and McConnaughay, K.D.M. 1995. A non-func-
tional interpretation of a classical optimal-partitioning example.
Funct. Ecol. 9: 951–954.
Coleman, J.S., McConnaughay, K.D.M., and Ackerly, D.D. 1994.
Interpreting phenotypic variation in plants. Trends Ecol. Evol.
9: 187–191. doi:10.1016/0169-5347(94)90087-6.
Gedroc, J.J., McConnaughay, K.D.M., and Coleman, J.S. 1996.
Plasticity in root/shoot partitioning: optimal, ontogenetic, or
both? Funct. Ecol. 10: 44–50. doi:10.2307/2390260.
Harper, J.L., and Ogden, J. 1970. The reproductive strategy of
higher plants. I. The concept of strategy with special reference
to Senecio vulgaris. J. Ecol. 58: 681–698. doi:10.2307/2258529.
Hull, P. 1974. Self-fertilization and the distribution of the radiate
form of Senecio vulgaris L. in Central Scotland. Watsonia, 10:
69–75.
Inouye, D.W. 1986. Long-term preformation of leaves and inflores-
cences by a long-lived perennial monocarp, Frasera speciosa
(Gentianaceae). Am. J. Bot. 73: 1535–1540. doi:10.2307/
2443919.
Jasienski, M., and Bazzaz, F.A. 1999. The fallacy of ratios and the
testability of models in biology. Oikos, 84: 321–326. doi:10.
2307/3546729.
Leiss, K.A., and Muller-Scharer, H. 2001. Adaptation of Senecio
vulgaris (Asteraceae) to ruderal and agricultural habitats. Am. J.
Bot. 88(9): 1593–1599. doi:10.2307/3558403.
McConnaughay, K.D.M., and Coleman, J.S. 1999. Biomass alloca-
tion in plants: ontogeny or optimality? A test along three re-
source gradients. Oecologia (Berl.), 113: 447–455.
Mu
¨ller, I., Schmid, B., and Weiner, J. 2000. The effect of nutrient
availability on biomass allocation patterns in 27 species of her-
baceous plants. Perspect. Plant Ecol. Evol. Syst. 3: 115–127.
doi:10.1078/1433-8319-00007.
Ollerton, J., and Lack, A.J. 1992. Flowering phenology: an exam-
ple of relaxation of natural selection? Trends Ecol. Evol. 7(8):
274–276. doi:10.1016/0169-5347(92)90175-B.
Pigliucci, M., and Preston, K. (Editors). 2004. Phenotypic integra-
tion: studying the ecology and evolution of complex phenotypes.
Oxford University Press, New York, N.Y.
Riska, B. 1991. Regression models in evolutionary allometry. Am.
Nat. 138(2): 283–299. doi:10.1086/285218.
Tuomi, J., Hakala, T., and Haukioja, E. 1983. Alternative concepts
of reproductive effort, costs of reproduction, and selection in
life-history evolution. Am. Zool. 23(1): 25–34.
Weiner, J. 1988. The influence of competition on plant reproduc-
tion. In Plant reproductive ecology: patterns and strategies. Edi-
ted by J. Lovett Doust and L. Lovett Doust. Oxford University
Press, New York, N.Y. pp. 228–245.
Weiner, J. 2004. Allocation, plasticity and allometry in plants. Per-
spect. Plant Ecol. Evol. Syst. 6(4): 207–215. doi:10.1078/1433-
8319-00083.
Wright, S.D., and McConnaughay, K.D.M. 2002. Interpreting phe-
notypic plasticity: the importance of ontogeny. Plant Species
Biol. 17: 119–131. doi:10.1046/j.1442-1984.2002.00082.x.
Weiner et al. 481
Published by NRC Research Press
A preview of this full-text is provided by Canadian Science Publishing.
Content available from Botany
This content is subject to copyright. Terms and conditions apply.
