PLANT ADAPTIVE STRATEGIES
IN RELATION TO VARIABLE
RESOURCE AVAILABILITY, SOIL
MICROBIAL PROCESSES AND
ECOSYSTEM DEVELOPMENT
SAMI
A IKI O
Department of Biology
OULU 2000
SAMI AIKIO
PLANT ADAPTIVE STRATEGIES IN
RELATION TO VARIABLE RESOURCE
AVAILABILITY, SOIL MICROBIAL
PROCESSES AND ECOSYSTEM
DEVELOPMENT
Academic Dissertation to be presented with the assent
of the Faculty of Science, University of Oulu, for public
discussion in Kuusamonsali (Auditorium YB 210),
Linnanmaa, on June 16th, 2000, at 12 noon.
O U L U N Y L I O P I S TO , O U L U 2 0 0 0
Copyright © 2000
Oulu University Library, 2000
Manuscript received 30 May 2000
Accepted 5 June 2000
Communicated by
Professor Per Lundberg
Docent Timo Vuorisalo
ISBN 951-42-5682-4
ALSO AVAILABLE IN PRINTED FORMAT
ISBN 951-42-5681-6
ISSN 0355-3191
(URL: http://herkules.oulu.fi/issn03553191/)
OULU UNIVERSITY LIBRARY
OULU 2000
Aikio, Sami, Plant adaptive strategies in relation to variable resource availability,
soil microbial processes and ecosystem development
Department of Biology, University of Oulu, FIN-90014 Oulu, Finland
2000
Oulu, Finland
(Manuscript received 30 May 2000)
Abstract
Plants have evolved various adaptive strategies for balancing the benefits and costs of having a
high affinity for resources, plasticity of growth allocation and mycorrhizal symbiosis. The relative
growth rates of mycorrhizal and non-mycorrhizal plants were modelled for stable and variable
nutrient availability. Mycorrhizal plants had higher growth rates at low and non-mycorrhizal plants
at high nutrient availability. Variation in nutrient availability reduced the growth rate of
mycorrhizal plants due to a high affinity for nutrients. However, mycorrhizal plants may be able to
buffer against external fluctuations and therefore experience less environmental variation than nonmycorrhizal plants. Non-mycorrhizal plants may even benefit from variation.
The optimal allocation of growth between shoot and roots depends on the availability of energy
and nutrients. The optimisation model predicted that the requirement for phenotypic plasticity of
shoot/root allocation is greatest in environments with low resource availability. Plants with a high
affinity for resources required more plasticity in order to tolerate variation than plants with a low
affinity. The model predicted a trade-off between the ability to deplete resources and the ability to
tolerate resource fluctuations.
Changes in the availability and ratio of resources lead to changes in the structure and
composition of vegetation during primary succession. The field study of the forested phases of the
land uplift island Hailuoto showed a successional change in the vegetation from the dominance of
bryophytes and deciduous dwarf shrubs to dominance by lichens and evergreen dwarf shrubs. The
humus layer became thinner and the availability of nutrients declined, while the C/N ratio of soil
organic matter increased during succession indicating a decline in the quality of organic matter.
The increased soil respiration rate indicates a successional increase in the energetic costs of
decomposing organic matter.
Nutrients mediate both direct and indirect trophic interactions. Indirect interactions of nutrient
cycling are not explicit in continuous time models. A transformation to a discrete time model was
shown to make the indirect interactions explicit as transition probabilities and allowed their
dynamic contribution to be evaluated with an elasticity analysis. The importance of indirect
interactions was greater in tundra than temperate forest and increased with the rate of nutrient
cycling.
Keywords: Phenotypic plasticity, mycorrhiza, primary succession, nutrient cycling
Acknowledgements
This work was conducted at the Department of Biology, University of Oulu. I wish to
express my warmest gratitude to my supervisors professor Rauni Strömmer and professor
Juha Tuomi for guidance and support during the work. Rauni introduced me to soil
ecology and encouraged me to take a theoretical approach to the study of soil organisms
in ecological processes. Juha has belived in my ideas, helped me forward with them and
seen a sparkle of hope even in my failures.
Although a great proportion of my work has been solitary staring at the computer
screen (doing nothing as an occasional by-passer might have thought), I have had the
pleasure to interact with a number of people. All the collagues and students who have
dropped in for a chat are warmly thanked, since I have never really done anything that is
too important to be interrupted. I have learned more from these informal consultations
than I can ever give in return. The coffee-breaks and discussions with the people of the
Department and the Evolutionary Plant Ecology Group, expecially Marko Hyvärinen, Esa
Härmä, Annamari Markkola, Annu Ruotsalainen, have been both enjoyable and
stimulating. Esa and Annamari have also been great friends during these years and I
would be a different person without them.
The formal and informal comments of the referees, professor Per Lundberg of the
University of Lund and docent Timo Vuorisalo of the University of Turku, are gratefully
acknowledged. Many thank's to Mirkka Jones and Bohdan Sklepkovych for language
revision. I also wish to thank professor Esa Ranta and the Integrated Ecology Unit of the
University of Helsinki for motivating me to do my thesis earlier than I first inteded. This
thesis was financed by the Academy of Finland (project #40951) and the Graduate School
in Evolutionary Ecology.
My cats Tiuhti and Viuhti have made it nearly impossible for me to work at home. This
has lead to long working days at the University but allowed me to save at least a small
fraction of my time and thoughts for things other than science. I am grateful to my parents
for staying safely in the background and knowing little of what I was working with, thus
reminding me of life outside the University. Thank you.
Helsinki, May 2000
Sami Aikio
List of original papers
This thesis is based on the following papers, which are referred to in the text by their
Roman numerals:
I
Aikio S & Ruotsalainen AL (2000) The growth of mycorrhizal and nonmycorrhizal plants under variable nutrient availability. Manuscript.
II
Aikio S & Markkola AM (2000) Resource interaction increases the phenotypic
plasticity of shoot/root allocation. Manuscript.
III
Ohtonen R, Aikio S & Väre H (1997) Ecological theories in soil biology. Soil
Biol Biochem 29: 1613–1619.
IV
Aikio S, Väre H & Strömmer R (2000) Soil microbial activity and biomass in the
primary succession of a dry heath forest. Soil Biol Biochem, in press.
V
Aikio S (2000) Evaluating the direct and indirect interactions of nutrient cycling.
Manuscript.
Contents
Abstract
Acknowledgements
List of original papers
1. Introduction .................................................................................................................11
1.1. Resources and plant strategies..............................................................................11
1.2. Ecosystem development and functioning .............................................................13
1.3. Aims of the study .................................................................................................14
2. Methods .......................................................................................................................15
2.1. Modelling.............................................................................................................15
2.1.1. Plant growth with variable nutrient availability.........................................15
2.1.2. Optimality and plasticity of shoot/root allocation......................................16
2.1.3. Indirect interactions in nutrient cycling.....................................................16
2.2. An empirical study of primary succession ...........................................................17
3. Results .........................................................................................................................18
3.1. Adaptive strategies in response to resource variation...........................................18
3.2. Nutrients in ecosystem processes .........................................................................19
4. Discussion ...................................................................................................................21
4.1. Plant strategies in response to variable resource availability................................21
4.1.1. Trade-off 1: maximum growth rate vs. affinity for resources ...................22
4.1.2. Trade-off 2: growth in response to stable vs. variable resources...............23
4.1.2.1. Mycorrhizal vs. non-mycorrhizal plants.......................................24
4.1.2.2. Phenotypically plastic vs. fixed plants..........................................26
4.2. Nutrients in ecosystem function and development ...............................................28
4.2.1. Resources and competitive interactions.....................................................28
4.2.2. Trophic interactions...................................................................................31
5. Conclusions .................................................................................................................33
6. References ...................................................................................................................35
Original papers
1. Introduction
1.1. Resources and plant strategies
The availability of soil nutrients sets limits to plant growth rate, and nitrogen and
phosphorus are the two most frequently limiting nutrients (DeAngelis 1992, Larcher
1995, Marschner 1995). Trade-offs have been suggested to exist between the abilities to
compete for resources in environments of high and low resource availability. Plants seem
to be unable to both deplete soil nutrients and grow fast when nutrients are abundant, or
tolerate shade and photosynthesise fast under conditions of high amounts of radiation
(Tilman 1982, Grace & Tilman 1990). The depletion of nutrients in soil is a slow process
but in the long run it is suggested to lead to the exclusion of nutrient demanding plant
species or strategies from a nutrient deficient site (Tilman 1982, Grover 1997). When
nutrients are abundant, fast growing plants are able to outgrow the slow growing ones. A
change in the nutrient availability of a site may thus lead to a change in the species
composition of a site, and also to a change within a species from one nutrient uptake
strategy to another.
Mycorrhizal colonisation is variable in many plant species and mycorrhizal and nonmycorrhizal strategies respond indifferent ways to the availability of soil nutrients
(Pairunan et al. 1980, Thomson et al. 1986, Bougher et al. 1990, Smith & Read 1997,
Titus & del Moral 1998). The winning strategy in competition for depleted nutrients often
includes symbiotic association with mycorrhizal fungi (Smith & Read 1997). Due to their
high surface-to-volume ratio fungi have a greater nutrient uptake capacity than plant roots
but they impose an additional energetic cost to a plant, since they may consume up to 10–
30 % of a plant's photosynthates (Fogel & Hunt 1983, Finlay & Söderström 1992, Smith
& Read 1997). This will significantly reduce plant growth rate when resource limitation is
shifted from nutrients to light and photosynthetic carbon. The cost of mycorrhizae may
then exceed the benefits to a plant. Many plants of nutrient rich habitats are therefore nonmycorrhizal or have a lower mycorrhizal colonisation than the plants of nutrient poor
habitats (Alexander & Fairley 1983, Smith & Read 1997).
Much theoretical and empirical research has been carried out to study the relationship
between plant growth and the availability of soil nutrients. Temporal (e.g. Chapin et al.
12
1978, Mullen & Schmidt 1993, Grover 1997, Smith & Read 1997, Mullen et al. 1998)
and spatial (Smith & Read 1997) variation in nutrient availability have received little
attention in experimental designs, which rather attempt to keep nutrients at a constant
level (Ingestad & Ågren 1988). The implicit assumption seems to be that growth with a
stable availability of nutrients equals the average growth rate with variable nutrients. Plant
species and strategies that differ in their response to a stable availability of nutrients are
also likely to differ in their response to the variation of nutrients. This can be deduced
from Jensen's inequality (Ruel & Ayers 1999), which predicts the difference between the
average of a function and the function of an average. This mathematical property can be
used for predicting the response of nutrient use strategies to variation in nutrients.
Plants require several resources, which are taken up with different organs and in
variable ratios. Photosynthetic energy is harvested via foliage while nutrients are taken up
by roots or mycorrhizal fungi. Plants therefore have to optimise their above and below
ground growth allocation according to the ratio of resource availability. The optimal
resource use theory (Rapport 1971, Bloom et al. 1985, Tilman 1988) suggests that plants
take up resources according to their proportional limitation by each of the resources. For
example, if growth is limited by the availability of nutrients rather than by photosynthesis,
plants may increase their proportional growth allocation to roots and decrease allocation
to shoots (Bloom et al. 1985). At the optimum, all resources will be equally limiting to
growth. The ability to optimise requires phenotypic plasticity of allocation (Bradshaw
1965, Grime et al. 1986, Sultan 1987, Scheiner 1993). Plasticity increases the range of
resource availabilities a plant can adapt to. According to Chapin (1980), plants with
higher maximum growth rates are phenotypically more flexible in their root vs. shoot
allocation than slowly growing plants. Phenotypic plasticity increases the tolerance of
plants to environmental variation and may therefore extend their geographic distribution.
Ecological interactions among plants often involve competition for nutrients. The
nutritional quality of dead organic matter is one of the major factors determining the
structure and composition of the soil decomposer community (Visser & Parkinson 1992,
Janssen 1996, Mary et al. 1996, Currie 1999) and to a large extent it also determines the
rate of nutrient mineralisation (Ågren & Bosatta 1996a,b). Nutrients become available to
plants by decomposition and therefore have the potential to determine the rate of primary
production of an ecosystem and are perhaps of greater theoretical and practical
importance than generally believed (Ågren & Bosatta 1996b). Proper theoretical
approaches to soil ecology have been sought and a long lasting debate in plant ecology,
the one between J. P. Grime and D. Tilman on plant strategies (Grace 1991, Grover
1997), spread belowground when Grime's (1977, 1979) theory was suggested to apply to
soil systems (Wardle & Giller 1997). Grime (1979) considers plants to be exposed to
stress (resource limitation) and disturbance (biomass removal) to a variable extent, which
leads to corresponding strategies of stress-tolerators (S) and ruderals (R). The
competitive-strategy (C) characterises plants that grow in resource rich (no stress) and
undisturbed habitats. Besides these C, S, and R strategies there are combinations and
intermediates including the indifferent CSR-strategy. Tilman (1982, 1988), on the other
hand, models plant growth as a function of one to several resources and considers plant
strategies to be adaptations to different combinations of the availability of resources.
Tilman (1982) derives the minimum resource availability R*, that corresponds the needs
of a plant, from the interaction of resource uptake by plants and the renewal of resources.
13
From an array of competing plant species, the most competitive is proposed to be the one
with the lowest R*. Since nutrient renewal is dependent on soil microbes, the inclusion of
nutrient mineralisation and other soil processes seems to be a natural extension of
Tilman's theory.
Grime's (1979) theory has been criticised for numerous reasons, including the
definition of stress, which Grime suggests to be a property of a habitat. The problem is
that a low nutrient availability may be experienced as stress by some plant species while it
may be optimal for others. Tilman's (1982) competition is defined as the ability to deplete
the availability of resources to a level that is not sufficient for other species to persist.
Tilman's resource competition theory thus predicts the long term competitive outcome
and shares the common criticism for all equilibrium theories that nature is not at
equilibrium. The properties that make a plant a good competitor under Grime's definition
make it a bad competitor according to Tilman's criteria, since different traits are beneficial
in competition for high relative to low resources. Grime and Tilman also differ in their
view of the forces that lead to a successional change. Grime (1979) sees the successional
increase in soil nutrient availability as declining stress and increasing competition. Tilman
(1985, 1988) argues that the increase in nutrients and decrease in light during a
succession leads to a shift from below-ground competition to above ground competition
according to the ratio of available resources.
1.2. Ecosystem development and functioning
Under fortunate conditions primary succession can be studied as a time series (Rydin &
Borgegård 1991) but empirical studies on succession are often based on the assumption
that a successional time series corresponds to a spatial gradient in a landscape (Drury &
Nisbet 1973, Glen-Lewin et al. 1992, Chapin et al. 1994). A major shortcoming with this
approach is that the correspondence between time and space may be distorted by nonsuccessional spatial gradients while the benefit is that both ecosystem development and
species response to an environmental gradient can be studied along transects representing
primary succession. The successional gradient in terrestrial ecosystems usually involves a
change in the thickness of the humus layer and the availability of soil nutrients (Odum
1969). Changes may also occur in soil parent material due to leaching, nutrient uptake
and podsolisation (Peet 1992). The amount of nutrients increases in the soil from the nonvegetated early successional stages as dead organic matter accumulates on the soil surface
(Chapin et al. 1994), but the trend may be reversed when an increasing proportion of
nutrients becomes bound to biomass (Tilman 1985, Wardle & Ghani 1995). The long
term dynamics of soil nutrient availability are largely determined by the rate of nutrient
input from the atmosphere and leaching of nutrients with percolating soil waters
(DeAngelis 1992). The short term nutrient dynamics and their availability to plants are
largely determined by the activity of decomposers. Soil organisms, especially bacteria and
fungi, have a key role in determining the rate of organic matter decomposition and
thereby nutrient mineralisation. These processes determine the rate of nutrient supply to
primary producers, largely determining the rate of biomass production and other
14
fundamental ecosystem processes in the temporal scale of plant generations (DeAngelis
1992, Ågren & Bosatta 1996a,b).
There are detailed models on the interaction of detritus, decomposers and plants
(Ågren & Bosatta 1996b), and complex and realistic models of nutrient cycling have also
been constructed (e.g. Rastetter et al. 1997). On many occasions, however, the dynamics
of nutrient cycling have to be interpreted from a linear, constant coefficient transition rate
matrix of a differential equation model (Webster et al. 1975, Harrison & Fekete 1980,
Carpenter et al. 1992, Polis & Winemiller 1996, Winemiller 1996). Direct nutrient flows
are often considered to be central to nutrient cycling and threfore the indirect interactions
are often neglected, although their importance has also been acknowledged (Abrams et al.
1996). Indirect interactions in nutrient cycling can emerge e.g. from the possibility that
nutrients move between several compartments during a finite time interval (Aikio 2000).
The sensitivity and elasticity analyses of the dominant eigenvalue of a transition matrix
indicate the contribution of individual transitions to the asymptotic dynamics of a system
(Caswell 1989, Carpenter et al. 1992, Yodzis 1996) but indirect interactions are ignored
when these analyses are based on a continuous time model (Aikio 2000). The inclusion of
indirect interactions in an elasticity analysis would provide information regarding their
dynamic role in nutrient cycling.
1.3. Aims of the study
In this thesis I examine the role of nutrients at several organisational levels from nutrient
acquisition strategies of individual plants to nutrient cycling at the ecosystem level. I
show the existence of a trade-off in resource competition under stable and variable
conditions (I, II) which may provide new insight into the costs and benefits of
mycorrhizal symbiosis (I) and the phenotypic plasticity of above and below ground
growth allocation (II). The strong interaction between plants and soil microbes motivates
the unification of the theories concerning the function of organisms in and above ground,
which will promote the development of both (III). This functional approach guided the
study of primary succession at the island of Hailuoto, the Gulf of Bothnia, where
ecosystem development and the associated changes in plant and soil microbial
communities are a consequence of post-glacial land uplift (IV). The largest scale in
nutrient-mediated interactions is the nutrient cycle of a whole ecosystem. Many studies
have discussed direct nutrient flows while ignoring indirect trophic interactions. I will
propose a measure for the strength of indirect interactions and their contribution to the
dynamics of nutrient cycling (V).
2. Methods
The studies within this thesis employ both theoretical modelling methods and empirical
field research. Only a general outline of the methods is given here, while more explicit
descriptions are given in the original papers.
2.1. Modelling
2.1.1. Plant growth with variable nutrient availability
A model was formulated for the rate of change in plant biomass as a function of soil
nutrient availability (I). Soil nutrients were assumed to be a log-normally distributed
random variable (Lindgren 1976). The growth rate function was fitted by the non-linear
least squares optimisation method using Matlab's Optimisation toolbox (Coleman et al.
1999) to published data on the relative growth rate of mycorrhizal and non-mycorrhizal
plants of several levels of soil nutrient availability. The relative growth rate distributions
were calculated from the distribution of soil nutrients, using the growth rate models as a
non-linear transformation. The average growth rate with variable nutrients was compared
to the growth rate with average availability of nutrients. The model was used for
comparing the growth rate distributions of mycorrhizal and non-mycorrhizal plants and to
detect how different traits of nutrient use, expressed as the parameters of the growth
model, affect the shape and expectation of the growth rate distribution. Maple V (Char et
al. 1991) computer algebra software was used for deriving the growth rate distribution
from the nutrient distribution. The expectation of the growth rate distribution was
computed numerically with Matlab (The MathWorks Inc, 24 Prime Park Way, Natick,
MA 01760, USA).
16
2.1.2. Optimality and plasticity of shoot/root allocation
The relative growth rate of plant biomass was assumed to be a saturating function of the
availability of nutrients and energy (expressed as photosynthetic carbon), which were
assumed to be in a multiplicative interaction with each other resulting in a two-resource
Baule function (O'Neill et al. 1989, II). The component vectors of the Baule function's
gradient were normalised to sum up to unity, which resulted in a measure of the
proportional contribution of the two resources to the length and direction of the gradient.
The contribution of nutrients and energy was used to determine the optimum proportion
of growth allocation between roots and shoots. Optimal allocation proportions were
represented as isoclines on a phase plane of nutrient and carbon availability.
The availabilities of the resources were assumed to have a bivariate normal distribution
(Lindgren 1976). Four combinations of low and high availability of the two resources
were presented in a phase plane of nutrient and carbon availabilities as circles or ellipses
that correspond to ± 1 SD units of variation in resource availability. The resources were
assumed to be either uncorrelated, or to have positive or negative correlation with each
other. The amount of plasticity in shoot/root allocation, that was needed for adapting to
the variation, was determined from the number of allocation isoclines covered by the
ellipses. The more allocation isoclines an ellipse covers, the more the plant must be able
to vary its shoot/root allocation. Traits associated with resource use were expressed as the
parameters of the Baule function and studied for their effect on a plant's requirement for
plasticity. The Baule function, its gradient and the contribution of the resources, as well as
the confidence ellipses of the resource distribution functions, were calculated analytically
with Maple V (Char et al. 1991). The graphical presentation of the model was computed
numerically with Matlab.
2.1.3. Indirect interactions in nutrient cycling
A linear, constant coefficient system of differential equations was solved for a single time
step with Maple V (Char et al. 1991). A discrete time model was constructed to be
structurally similar to the continuous time model (V). The equality of the two models was
used to derive a transformation from a transition rate matrix of continuous models to a
discrete time transition probability matrix. The transformation was exemplified with the
data of Webster et al. (1975) on nutrient cycling in tundra and temperate forest. The data
were reanalysed for the return time of the ecosystems from perturbations to nutrient
content, the stable nutrient content of the systems, the proportion of indirect interactions,
and the dynamic contribution of the indirect interactions. The latter aspect was carried out
using an elasticity analysis (Caswell 1989, Heppell et al. 2000) of the transition
probability matrices that resulted from the transformation from continuous time models to
discrete time models. The nutrient flow rates of tundra and temperate forest ecosystems
were multiplied with values between 0.1 and 10 to investigate the effect of nutrient
throughflow on the dynamics of the system. The throughflow rate was calculated as the
ratio of nutrient flow to the stable nutrient content of the system. The return time from
17
perturbations, stable nutrient content of the system, the proportion of indirect interactions
and the sum of indirect interaction elasticities were plotted against the nutrient
throughflow in tundra and temperate forest ecosystems. The return times of the systems
were calculated as the inverse of the absolute dominant eigenvalue of the transition rate
matrix. The stable nutrient content of the systems was a product of the negative inverse of
the transition rate matrix and the nutrient input rate vector. The analyses were performed
numerically with Matlab.
2.2. An empirical study of primary succession
Primary succession was studied on three transects in Hailuoto, an island off the coast of
the northern Baltic Sea (IV). The island is uplifting at a relatively constant 9 mm annual
rate (Alestalo 1971). Three transects ran from the low elevations near the shore towards
the highest parts of the island, and 5 study sites representing soils of different ages
(Alestalo 1971, Alestalo 1979) were established on each transect. At each study site, five
10 × 10 m quadrats were established along a line orthogonal to the transect and 5 m apart
from each other.
In 1995, vegetation coverage was estimated visually using a percentage scale from ten
systematically chosen 1 × 1 m squares within each quadrat, and soil samples were taken
from the top 3 cm of the humus layer for the measurements of soil microbial properties.
Another set of soil samples were taken for nutrient analysis from the whole thickness of
the humus layer and the top 15 cm of the mineral soil in august 1997. The soil samples
were homogenised and pooled for each quadrat giving 75 sampling units.
Basal and substrate induced respiration (SIR, Anderson & Domsch 1978) were
determined as the average of two replicates using the Respicond II respirometer apparatus
(Nordgren 1988). SIR was transformed to microbial biomass-C (Anderson & Domsch
1978, with modifications). Plant available cations (Fe, K, Ca, Mg) were determined
spectrophotometrically after ammonium acetate extraction (Halonen et al. 1983) and soil
ammonium (NH4+) colorimetrically from the KCl extract using the method of John
(1970).
Data were analysed using non-metric multidimensional scaling (Kruskal & Wish 1978)
of the vegetation treating microbial and soil nutrient measurements as environmental
variables (Minchin 1988). Soil nutrients were presented as the three principal
components. The vegetation data was subjected to a two-way indicator species analysis
(TWINSPAN, Hill 1979) to form vegetation clusters, which were then compared with
respect to soil nutrient concentrations and microbial biomass and activity. Microbial
variables and soil nutrient concentrations were also compared between altitude classes.
Comparisons were made with a non-parametric Kruskall-Wallis test. The successional
trend in C/N ratio of soil organic matter was assessed with Spearman's rank correlation.
3. Results
3.1. Adaptive strategies in response to resource variation
The maximum growth rate and the half-saturation constant were higher in nonmycorrhizal than in mycorrhizal treatments in all of the reanalysed experiments. A nonzero threshold for growth was found in one case, where non-mycorrhizal plants had a
threshold while mycorrhizal ones did not. The plant growth rate model fitted well to the
relative growth rates of mycorrhizal and non-mycorrhizal plants at various levels of
phosphorus availability (I). The highest nutrient concentrations were an exception, since a
low growth rate at high nutrient level was not accounted for in the model.
The expectation for the plant's relative growth rate distribution in response to variable
nutrient availability was different to the growth rate expectation at stable nutrient
availability. When the growth rate function was decelerating for the whole range of
nutrient availabilities, as was the case with the mycorrhizal plant, the growth rate
expectation was lower in the variable than in the stable availability. When the growth rate
function had an accelerating part at low nutrient availabilities due to a threshold for
growth, as was the case with the non-mycorrhizal plant, growth was expected to be higher
under conditions of variable nutrient availability than under stable nutrient availability,
provided most of the variation in nutrient availability coincides with the accelerating part
of the growth rate function. The difference between the expected growth rate at stable and
variable nutrient availability increased with decreasing half saturation constant but was
not related to the maximum growth rate. Mycorrhizal plants were more adversely affected
by variation in nutrient availability than non-mycorrhizal plants, which were found to
benefit from variation when expected availability was below their threshold of growth.
The shoot/root allocation isoclines were curves radiating from the origin of the carbon
and nutrient availability phase plane (II). The isoclines were close to each other near the
origin and diverged with distance from it. The semiaxes of the confidence ellipses of the
resource availability distributions were parallel to the axes of the phase plane when
nutrients and carbon varied independently of each other. The major semiaxes of the
ellipses had a positive slope when the resources were positively correlated and a negative
slope when the resource correlation was negative. The size (area) of the ellipses increased
19
with increasing variance of resource availability. The coefficient of variation was kept
constant and the ellipses increased in size with increasing resource availability.
When the ellipses were superimposed on the shoot/root allocation isoclines at the
phase plane of carbon and nutrient availability they covered more isoclines when there
was low resource availability than when there was high availability. Thus a plant species
growing under conditions of low resource availability requires more phenotypic plasticity
in shoot/root allocation than the same plant growing under conditions of high resource
availability, in order to adapt to variation. The ellipses cover more isoclines when
resources have a negative correlation with each other than when the correlation is
positive. The need for plasticity was therefore increased by the negative correlation of
resource availability and reduced by positive correlation. Constraints of plasticity restrict
plant growth in environments where the ratio of resource availability is very different
from the optimum for the plant.
The effect of resource utilisation on the plasticity of shoot/root allocation were studied
by changing the parameter value of either maximum growth rate or half saturation
concentration and plotting the resulting allocation isoclines onto a resource availability
phase plane. The phenotypic plasticity of shoot/root allocation was sensitive to variation
in resource availability in a similar way to relative growth rate in Paper I. The allocation
isoclines did not change with changes in maximum growth rate, which indicates that
maximum growth rate of plant biomass is not related to the plasticity of shoot/root
allocation. A low half-saturation constant made the isoclines diverge from each other than
a high half-saturation constant.
3.2. Nutrients in ecosystem processes
Along a successional gradient in Scots pine (Pinus sylvestris L.) forests on the island of
Hailuoto the field layer vegetation changed from a dominance of deciduous dwarf shrubs,
mosses and herbs to a dominance by lichens and evergreen dwarf shrubs (IV). The
nutrient concentration per unit area and the thickness of the humus layer declined with
successional age. Contrary to the expectation, the nutrient concentration of soil organic
matter did not change significantly during the succession. There was an increasing trend
in the carbon-to-nitrogen ratio in the humus, however, suggesting a successional decline
in the nutritional quality of organic matter. Microbial respiration rate per unit weight of
organic matter increased along the successional gradient being highest in the lichen
dominated sites. When soil nutrients, microbial biomass and activity were expressed per
unit area, they declined along the succession. Soil processes were more closely related to
the community structure of vegetation than to the successional age. The three early
successional stages were rather variable in vegetation, nutrient concentration and
microbial properties whereas the two late-successional stages were similar to each other.
Webster et al. (1975) modelled nutrient cycling in tundra and temperate forest
ecosystems with a linear, constant coefficient differential equation model x'(t) = Ax(t) + z,
where vector x(t) is the nutrient content of six ecosystem compartments (plants,
consumers, detritus, saprophytes, plant available nutrients and nutrient reserve) at time t,
20
A is a matrix of nutrient transition between the compartments, and z is a vector of external
nutrient input rate (V). The model was shown to have the same dynamic properties as a
difference equation model x(t + 1) = Px(t) + n, where P is a transition probability matrix
and n is a vector of annual nutrient input to the system. The models had a dynamic
correspondence when the transition matrices had the relationship P = eA. This
relationship was used as a transformation from the specific direct flow rates of matrix A
to the probabilities of direct nutrient flows and indirect interactions in matrix P. Indirect
probabilities came from the possibility of transition through several compartments during
a finite time interval. The probability of an indirect interaction was sometimes greater
than a direct nutrient flow from the same compartment. This occurred when the turn-over
rate of the recipient compartment and the probability of nutrient movement were high.
The indirect interactions were proportionally greater in tundra than in the forest. They had
a weaker contribution to the asymptotic dynamics of the system than direct nutrient flows
in this data set. The return time was shorter in tundra than in forest. An elasticity analysis
of the dominant eigenvalue of tundra and forest transition probability matrices showed
that the contribution of indirect interactions to the asymptotic dynamics of the nutrient
cycle was greater in tundra than in forest. An increase in the speed of internal nutrient
cycling, as measured by the throughflow of nutrients (i.e. the ratio of nutrient flow to the
nutrient content of the system), decreased the return time in both tundra and forest and
increased the resilience of the systems. Throughflow did not affect the stable nutrient
content of the system, however. The proportion of indirect interactions increased with
increasing nutrient throughflow. The contribution of indirect interactions, as measured by
the sum of their elasticities, also increased with increasing nutrient throughflow.
4. Discussion
4.1. Plant strategies in response to variable resource availability
The relative growth rate of plants typically increases with resource availability and
becomes saturated at some maximum (Grover 1997). In a simple case, which can be
described by a Michaelis-Menten function (Marschner 1995), the growth rate increases
rapidly with an increase from zero to low nutrient availability but is smaller for conditions
of high resource availability. In other words, the relative growth rate is a decelerating or
concave function of resource availability (I). The response to resource availability is
expressed with two parameters in Michaelis-Menten function: one for the maximum
growth rate and the other for the steepness of the function. In some cases it is necessary to
include a third parameter to account for the threshold nutrient concentration for growth.
The threshold indicates the difference between the measurable concentration of a resource
and the availability of the resource for a plant. A threshold parameter should be included
in a model if there are reasons to assume that some of the traits or species under
comparison are able to utilize a fraction of the resource which is not accessible to other
plant types. A growth rate function with a threshold is more complicated than a
Michaelis-Menten function, since it is accelerating (convex) at low and decelerating
(concave) at high resource availabilities. The predictions that are based on a growth rate
function with a threshold may therefore be qualitatively different from those of a function
without a threshold (I).
Plants are usually limited by multiple resources, which often have an interactive effect
on relative growth rate (Tilman 1982, 1988, O'Neill et al. 1989, Grover 1997, II). The
concept of modelling plant growth rate as a function of resource availability is easy to
extend to multiple resources by assuming a maximum growth rate and resource-specific
half saturation parameters to account for the limitation by each of the resources. In the
simplest case this yields the Baule function (O'Neill et al. 1989, II), which is essentially a
product of two Michaelis-Menten functions. A single-resource model makes it possible to
study the sensitivity of growth rate to variation in resource availability (I). A similar
analysis could be carried out with multiple-resource models, which also enable the study
of optimisation and growth allocation between resource uptake organs (II). The variation
22
in the allocation optimum makes it further possible to study the phenotypic plasticity of
allocation.
The assumption of non-interacting resources has been found to be adequate in the
experiments regarding resource competition by aquatic algae, which due to their
unicellularity are limited in their allocation patterns (Tilman 1982, Grover 1997).
Vascular plants in terrestrial conditions experience more variation in resource
availabilities and environmental conditions than algae and their structural specialisation
enables considerable phenotypic plasticity in resource allocation (Hutchings & de Kroon
1994, Ågren & Bosatta 1996b, Grover 1997). Plants may also have some capacity to
compensate for the low availability of a resource for instance by increasing the uptake of
another, more abundant resource, which in turn increases the uptake of the depleted
resource. This may occur e.g. in nutrient uptake by mycorrhizal plants, which can increase
their carbon allocation to mycorrhizal fungi in nutrient-poor conditions and thereby
increase their nutrient gain (IV).
4.1.1. Trade-off 1: maximum growth rate vs. affinity for resources
The trade-off between maximum growth rate and affinity for resources is one of the
reasons why no single best strategy for resource competition has evolved (Chapin 1980,
Grace & Tilman 1990). Both mycorrhizal and non-mycorrhizal plants exist since the
benefit of being mycorrhizal or non-mycorrhizal varies with resource availability.
Mycorrhizal plants have a lower maximum growth rate than non-mycorrhizal plants and
they saturate at lower nutrient availabilities. This supports the hypothesis that being
mycorrhizal is the best strategy for growth under conditions of low nutrient availability
(I).
Good tolerance of low resource availability may lead to low maximum growth rate
(Smith & Read 1997, I). Mycorrhizae are a net benefit to a plant when nutrients are in low
supply but may become a net cost when nutrients are abundant. As the availability of
nutrients varies seasonally, the benefit of mycorrhizal symbiosis to plant is also seasonally
variable. Plants have some capacity to regulate their degree of mycorrhizal colonisation
and seasonal changes in mycorrhizal colonisation have been reported (e.g. Brundrett &
Kendrick 1988, Mullen & Schmidt 1993, Mullen et al. 1998). Plants may refuse to form
mycorrhiza or be able to reduce the colonisation percentage if nutrient concentration does
not favor being mycorrhizal (Bougher et al. 1990, Setälä et al. 1997), but fungi may
benefit from plants even under such conditions. Especially the vesicular-arbuscular
mycorrhiza forming fungi may resist the attempts of a plant to prevent infection, since
they are unable to survive without a host plant (Smith & Read 1997). The costs of
maintaining a physiological machinery that restricts fungal infection are probably
comparable to the costs of avoiding fungal parasites (Harborne 1988).
Some of the reanalysed data (I) showed a reduction in plant growth rate under high
nutrient concentrations (e.g. Howeler et al. 1982, Amijee et al. 1989). This phenomenon
may be a toxic effect of unnaturally high nutrient concentrations, which the saturating
growth model is not capable of describing. It is one of the simplest functions to capture
23
the essential features of plant growth as a function of nutrients (e.g. Marschner 1995).
Alhough experiments have frequently demonstrated the toxicity of very high nutrient
concentrations, it is questionable if this is a common phenomenon in nature. Plant species
that frequently experience high nutrient concentrations are likely to be adapted to them,
e.g. ecotypes have evolved to tolerate the extreme nutrient concentrations of serpentine
bedrocks (Rune 1957).
4.1.2. Trade-off 2: growth in response to stable vs. variable resources
It is often assumed, perhaps implicitly, that the average growth rate when resource
availability varies is the same as the growth rate with a mean resource availability. A
more realistic situation can presumably be predicted from Jensen's inequality, which is a
mathematical property of non-linear functions (Ruel & Ayers 1999). This inequality
predicts that the expectation of a non-linear function of a random variable differs from the
value of the same function which has the expected value of the random variable as an
argument, i.e. E[f(x)] ≠ f[E(x)]. Plant growth rate may be considerably affected by
variation in resource availability and the difference in growth rate under conditions of
variable versus stable resource availabilitity depends on the shape of the function. If f(x)
is accelerating, average growth rate will increase with increasing variance of x, but if f(x)
is decelerating the average growth rate will decrease with increasing variance of x.
The decelerating Michaelis-Menten (I) and Baule (II) growth rate functions are more
sensitive to fluctuations in resources of low availability than resources of high
availability. This has important consequences for plant resource utilization strategies,
since traits that increase the efficiency of resource uptake and long term competitive
success under conditions of stable resource availablity increase the concavity of the
growth rate functions and make plants increasingly sensitive to fluctuations in resources
(I, II). There is thus a trade-off between the ability to grow at a low but stable availability
of resources and the ability to grow when resource availability fluctuates. The relationship
between resource variation and average plant growth rate becomes complicated when the
growth rate function has both a concave and a convex part due to the existence of a
threshold. Resource variation can then increase growth resulting in a positive relationship
rather than a trade-off between relative growth rates at stable and variable resources.
However, such a positive relationship only applies to a narrow range of low resource
availabilities above which there is again a trade-off between resource variation and
average growth rate.
The trade-off between the ability to deplete resources at constant availability and the
ability to tolerate fluctuations in resources may allow the coexistence of species, that
might otherwise outcompete one another in a stable environment (Levins 1979). Since
good resource competitors, those that are able to deplete resources, are more affected by
variation than inferior resource competitors, variation in resource availability may inhibit
competitive exclusion. Resource fluctuations are predicted to be particularly detrimental
to plants that are good competitors for stable resources. These plants would also gain
greater benefit from adaptations that stabilize the rate of resource uptake. In nutrient
24
uptake such strategies could include increased nutrient storage and mycorrhizal
symbiosis, which might stabilise the flow of nutrients to plants. The number of
environmental dimensions doubles when the variation in each resource or environmental
condition is considered to be a separate variable. This can increase the number of
ecological niches (Hutchinson 1957). The increasing number of environmental variables
increases the potential number of species that an environment can support and thus
variation in resource availability has potential to increase diversity (Levins 1979, Tilman
1982). The trade-off between the ability to tolerate resource fluctuations and grow at low
resource availabilities may therefore be a factor that increases diversity.
4.1.2.1. Mycorrhizal vs. non-mycorrhizal plants
Mycorrhizal and non-mycorrhizal plants respond in different ways to the availability of
nutrients as well as to the temporal and spatial variation in nutrient availability (I). When
nutrient availability is stable, growth rate is predicted to be constant but variation in
nutrient availability also causes growth rates to vary. Therefore, when nutrient availability
varies, relative growth rates form a distribution with some expectation. The distributions
of mycorrhizal and non-mycorrhizal plants differ in shape and consequently mycorrhizal
and non-mycorrhizal plants differ in their response to variation in nutrient availabilities
(I). Variation reduces plant grow rate if the growth rate function is decelerating, as it is
for mycorrhizal plants. The expectation for growth rate in mycorrhizal plants declines
more due to variation in nutrients than the expectation for growth rate in non-mycorrhizal
plants. This is a consequence of the greater concavity of the growth rate function of
mycorrhizal plants. The growth rate of non-mycorrhizal plants can be increased under
certain conditions by variation in nutrient availability (I). This increase in growth rate
requires that the growth rate function has a convex part coincident with most of the
variation in nutrient availability. Convexity can be introduced to a growth rate function by
including a threshold parameter.
The variation in nutrient availabilities was assumed to be the same for both
mycorrhizal and non-mycorrhizal plants such that the more adverse effects on mycorrhizal
plants were due to the greater concavity of the growth rate function of mycorrhizal alone.
The prediction that mycorrhizal plants are more sensitive to variation in nutrient
availability than non-mycorrhizal plants seems to contradict empirical observations.
Generally mycorrhizae are assumed to stabilize nutrient uptake in heterogenous
environments (Cui & Caldwell 1996a,b, Smith & Read 1997). This apparent contradiction
comes from the assumption of the growth rate model that the availability of resources
fluctuates with an equal magnitude for both mycorrhizal and non-mycorrhizal plants. This
will not be the case if mycorrhizal plants have more nutrients per unit biomass than nonmycorrhizal ones, due to luxury accumulation (Stribley et al. 1980, Bolan 1991). Such
accumulation would enable mycorrhizal plants to buffer the periods of low nutrient
availability. The storage of nutrients in mycorrhizal plants would reduce the effect of
fluctuations in external nutrient availability and stabilize the flow of nutrients reguired for
physiological processes. The fungal hyphae could have similar stabilizing role, both by
25
effectively providing plants with nutrients, even when they are at low concentration
(Smith & Read 1997) and by increasing the volume of nutrient storage. Mycorrhizal
plants therefore appear to have a stronger buffer of internal nutrient concentration against
variation in external nutrients than non-mycorrhizal plants. Therefore mycorrhizal plants
experience smaller environmental fluctuations than non-mycorrhizal plants. Since a good
ability to compete for nutrients involves a cost in making a plant more sensitive to
variation in nutrient availability, mycorrhiza may have a dual role in nutrient competition
by both increasing affinity for nutrients and stabilizing the flow of nutrient to a plant. In
fluctuating and patchy environments the latter may be an important factor in the
competition for nutrients and a reason for plants to maintain mycorrhiza.
The trade-off between the ability to deplete resources and the ability to tolerate
resource fluctuation depends on the colonization by mycorrhizae. Variation in nutrient
availability may also affect the composition of mycorrhizal fungal types. Herbs and
grasses are relatively plastic in their shoot/root allocation and their vesicular-arbuscular
(VA) -mycorrhizal colonization may be variable during a season (Smith & Read 1997).
VA-mycorrhizae increase the nutrient uptake maximum (I) and may thereby facilitate the
utilization of nutrient pulses increasing plant growth rate when resources are abundant.
Ectomycorrhizal plants (most trees) may be less plastic than herbs and the degree of
colonization by ectomycorrhizal fungi is often less variable than the colonization by VAmycorrhiza (Smith & Read 1997). Ectomycorrhizal plants generally grow in nutrient
deficient conditions where the stability of nutrient uptake is presumably of greater
importance than fast growth during nutrient pulses (Smith & Read 1997, I).
In the present analysis (I), variation in nutrient availability increased average plant
growth rate when nutrient availability only exceeded the threshold during peak
availabilities. Deserts provide an example where the availability of water is variable and
only exceeds the threshold of growth during rainy periods. Some ephemeral plants may
tolerate dry periods as seeds, and only grow and reproduce during the rainy periods.
Among the analysed experiments, a threshold was only found in the experiment on
Eucalyptus diversicolor (Bougher et al. 1990), where the nutrient level treatments were
properly chosen and their number was high enough. However, experimental set-up may
not be the only explanation. E. diversicolor is an ectomycorrhizal plant while the rest of
the experiments were carried out with VA-mycorrhizal plant species. Ectomycorrhizal
plants are able to utilize otherwise inaccessible sources of nutrients, such as organic
nitrogen (Northup et al. 1995, Näsholm et al. 1998, Perez-Moreno & Read 2000) and
mineral nutrients of the soil parent material (Jongmans et al. 1997). A threshold in growth
response to soil nutrient availability may be demonstrated when naturally ectomycorrhizal
plant species are grown with and without mycorrhizal inoculation. The VA-mycorrhizal
plants treated with mycorrhiza or without it probably do not differ this much in amount of
nutrients that they can potentially exploit.
26
4.1.2.2. Phenotypically plastic vs. fixed plants
Variation associated with the low availability of resources sets greater requirements for
phenotypic plasticity of a plant's shoot/root allocation than the variation associated with
high resource availability (II). The prediction was made with the assumption that
resources of high and low availabilties have the same coefficient of variation. If the
variance is equal under conditions of low and high availability or if the low resource
availability has a higher variance than the high availability, the difference between
conditions of low vs. high resource availability is even greater. When resource availability
is low, even slight variation can bias the resource ratio from the optimum for a plant. A
plant growing under resource-poor conditions will therefore benefit more from plasticity
of shoot/root allocation than the same plant growing under resource-rich conditions. The
interactive-essential resources have little ability to complement each other when resource
availability is low. Essential resources can not be replaced, but their complementarity
increases with increasing resource availability due to the increased interaction between
resources. The ability to vary shoot/root allocation is limited under conditions of low
resource availability, since the interaction of resources is low and plants are sensitive to
variation in both energy and nutrients.
The requirement of a plant for plasticity is greatly affected by the correlation of the
availability of resources, and the effect of correlation is greatest when resources are at a
low availability. A negative correlation between resources increases the variation in the
resource ratio while a positive correlation reduces the variation in the resource ratio.
Plants therefore need to express a wider range of allocation alternatives in an environment
where resources are negatively correlated than in an environment where resources are
positively correlated. A strict negative correlation of resources would require complete
plasticity and a strict positive correlation would remove all needs for plasticity. At a large
scale, environments have a negative correlation between the availability of nutrients and
the amount of photosynthetic energy (Grover 1997). In dense vegetation there is little
light at the field layer but the accumulation of organic matter provides higher soil nutrient
concentrations than open habitats with plenty of light. Primary succession starts in
habitats where nutrients are scarce but light may even be superabundant and suitable for
quite different plant species than closed vegetation with a greater availability of nutrients
but less light. At the landscape level, negative correlation between the availability of
nutrients and light promotes diversity by reducing the probability of adaptation to the
variation by means of phenotypic plasticity.
The stabilisation of resource inflow may be of competitive importance (Grover 1997, I,
II). Some plants, e.g. coniferous trees, grow in habitats where they experience low
nutrient availabilities but high amounts of photosynthetically active radiation (Larcher
1995). Other plants, e.g. understorey herbs of deciduous forests, grow on more nutrient
rich soils but under a dense vegetation and receive low amounts of radiation (Tilman
1988). Conifers have a relatively high affinity for nutrients while herbs have a high
affinity for light, which suggests that conifers are sensitive to variation in nutrient
availability while herbs are affected by variation in the radiation environment.
Ectomycorrhizae of coniferous trees may effectively stabilise the nutrient flow to a plant.
Deciduous plants that live in shade only experiencing temporary light flecks may be more
27
affected by variation in the radiation environment than in nutrients, since there are few
possibilities for stabilizing photosynthesis with respect to the variation in amount of
incoming light (Ruel & Ayers 1999). Understorey herbs with a high maximum growth
rate may have costs as declined growth rate decline due to variation in light availability.
Plants that are adapted to low, or alternatively, to high resource availability respond
differentially to changes in resource availabilities. Plants with low resource requirements
usually have a low maximum growth rate and already saturate at low resource
availabilities. If resources are moderately available they only respond weakly to a surplus
of resources. Plants with high resource requirements respond strongly to additional
resources (Chapin 1991). This suggests that plants from resource-rich environments are
more plastic than those from resource-poor environments. The analysis shows that the
difference is not caused by a high maximum growth rate of nutrient demanding plants but
by their higher half saturation concentration. If there is a lot of variation in the availability
of resources that have a high mean availability, the ability to efficiently utilise the
resource peaks is favoured. In such a case, species from highly productive habitats are
predicted to have a greater flexibility in resource foraging.
Resources have a patchy distribution, especially when they are at low availability
(Kleb & Wilson 1997). The plasticity of resource uptake organs enables foraging in
resource-rich patches and the avoidance of resource-poor patches. Since fungal hyphae
can explore soil with less costs than plant roots, mycorrhizal symbiosis probably makes
the exploitation of soil resources more efficient in this respect, too. The ability to control
mycorrhizal colonzation can also be a source of plasticity. A good availability of light or
high concentration of atmospheric CO2 (Woodward et al. 1991, Rogers et al. 1994)
enhance photosynthesis thus enabling the plant to increase aboveground growth, and also
to allocate more resources below ground to facilitate nutrient uptake in nutrient poor
environments (Zak et al. 1993, IV). This complementarity of resources is limited, as can
is suggested by the acclimation of photosynthesis to elevated CO2 levels (Stitt & Krapp
1999). In this experiment, plant growth rate increased with increasing CO2, but declined
to the original growth rate faster in nitrogen-limited plants than in plants with a high
supply of nitrogen.
The plasticity requirements of a plant were determined from the properties of the
gradient of the Baule function (O'Neill et al. 1989). The curvature of the relative growth
rate isoclines increases with increasing resource availability. This is a general property of
the functions that assume the growth rate to be proportional to the product of the
availability of the two resources (II). The multiplicative interaction of interactive essential
resources was assumed in several models, including the one that fitted best to multiple
data sets (O'Neill et al. 1989), and seems to be a rather safe assumption. The prediction of
increasing complementarity and a reduced requirement for plasticity with increasing
resource availability should be both general and robust. The actual shape of the growth
rate function and its isoclines depend on the parameter values of the function. In the
Baule-function (II) these were the maximum growth rate and the half-saturation constants
for carbon and nutrients. Two species, or two genets of the same species, will probably
differ from each other at least in one of the parameter values. If maximum growth rate
alone differs, the two species are predicted to have equal requirements for the plasticity of
shoot/root allocation. If they differ in their half-saturation constant, the plant which is
more effective at exploiting low resources of constant availability (the one with the lower
28
half-saturation constant) is predicted to be more affected by resource fluctuations than the
plant which is inferior at exploiting depleted resources.
4.2. Nutrients in ecosystem function and development
Much of the theory on plant growth and resource allocation, including the contribution of
this thesis (I, II), is based on the response of plants to the availability of nutrients. For
example, the well-known allocation principle is based on assumption of resource
limitation (Cody 1966). The effect of nutrients extends from the physiological needs of
individual plants to competitive interactions that modify plant community structure during
succession (IV) and to ecosystem level trophic interactions that are mediated by nutrients
(V). The availability of nutrients is largely determined by decomposers and nutrient
mineralisation which therefore control the primary productivity of a system (III).
Traditionally, the potential limitation by nutrients is treated implicitly as a constant
carrying capacity of the system. This implies that resource dynamics are fast compared to
the dynamics of consumers (Strauss 1991, Abrams et al. 1996). The dynamics of nutrients
and soil organisms and their nutrient-mediated interaction with plants have received less
attention. Theoretical approaches to link the dynamics of the soil subsystem to plant
ecology have however been presented (DeAngelis 1992, Ågren & Bosatta 1996b, III). A
theory that builds on the decomposition and mineralisation processes as determinants of
the productivity of ecosystems seems to be a promising approach to soil ecology (III).
The vast diversity of soil organisms and the problems involved in studying them have
probably restricted rather than promoted the emergence of theory and generalisations,
when the dynamics of soil processes have been interpreted from the properties of species.
The availability of various resources and their mutual relationships vary during
ecosystem development, which leads to a change in resource-mediated competitive
interactions during ecosystem succession (Tilman 1985, 1988, Grover 1997, IV) and
alters the competitive hierarchy of plant species. Resource competition is an indirect
interaction (Strauss 1991) but trophic interactions include both direct nutrient flows and
indirect interactions between trophospecies (Yodzis 1996) that do not have a direct
consumer–resource interaction with each other. The dynamic importance of indirect
trophic interactions increases with the rate of internal nutrient circulation and may
occasionally exceed the importance of direct feeding interactions. It is therefore important
to evaluate their dynamic contribution as well as the contribution of direct flows (V).
4.2.1. Resources and competitive interactions
Ecologists have classified species as having various strategies, which help to epitomise
the traits of resource utilisation and habitat requirements (Grover 1997). Classification
must, however, be understood to serve only descriptive purposes. Grime's (1979)
classification of plant species as being either competitors (C), stress tolerators (S) or
29
ruderals (R), or their intermediates, is based on the comparison of one species to another.
This means, that the strategy label of a species depends on the sample of species which is
considered in the classification, and the addition of a new species to the scheme may
change the position of the previously classified species. The vast number of species living
in soil and the practical difficulties that are associated with their identification guarantee
that there are new species to be found and new descriptive information to be gathered for
a long time in the future. The limited knowledge of their in situ functioning implies that a
classification based on the current state of our knowledge would require major changes in
future classifications. The classification of soil organisms to Grime's strategies, as
suggested by Wardle and Giller (1997), must therefore be considered an unstable building
block in theoretical soil ecology (III). Alternatives do exist, however. Quantities like
biomass, nutrient concentration and respiration are based on species-independent physical
measures. One of the goals in theoretical ecology is to unite fields that are otherwise
treated as separate. This requires that studies are focused on the function of the
subsystems rather than on patterns of species composition, and that subsystems share an
independent unit of measurement, such as those presented above. The study of processes
like energy flow and nutrient cycling promote this agenda, and decomposition, nutrient
mineralisation and symbiotic interactions of soil organisms with plants should be studied
in the soil subsystem.
The long term competitive success of a plant requires the ability to reduce the
availability of a soil nutrient to a level that is sufficient for its own persistence but too low
for its competitors (Tilman 1982). The ability to tolerate nutrient deficiency is associated
with a slow growth rate and low biomass turnover rate, which requires in turn that the loss
rate due to senescence is low. This further leads to a low input rate of litter to soil and a
slow increase in the soil organic matter reservoir. A low rate of organic matter
accumulation results in a thin humus layer, which will also dry easily (increasing the
potential risk of fire). This was studied on a primary succession gradient (IV). Vegetation
shifted from domination by bryophytes and deciduous dwarf shrubs to domination by
evergreen dwarf shrubs and lichens. The microbial biomass and activity, as well as soil
nutrient concentrations, decreased towards the late-successional stages. This supports the
view that the availability of soil nutrients is a key factor determining both the vegetation
composition and microbial properties (III).
Since the nutrient concentrations varied little when expressed per unit organic matter,
the change in nutrient availability was connected with the considerable reduction in the
thickness of the humus layer and the amount of organic matter in the soil. The rate of
nutrient leaching probably increases with the thinning of the humus layer, since the
variation in soil moisture is greater when there is little organic matter to retain rain water
and thus buffer against the drying of the soil. The decrease in soil nutrients towards the
later stages of succession is not a new observation (Crocker & Major 1955, Odum 1969,
Tilman 1988, Peet 1992) but the age of the soil when nutrient concentrations began to
decrease, was very young in comparison with that in some other chronosequences (e.g.
Wardle & Ghani 1995), apparently due to the coarse texture of soil.
The change in vegetation composition suggests a decreasing quality of detritus during
succession, which was supported by the increasing trend in the C/N ratio. This suggests a
decline in the decomposability of organic matter (Kaye & Hart 1997). When microbial
activity and biomass were expressed per unit mass of organic matter, the highest basal
30
respiration rate was in the lichen-dominated sites at the highest altitudes. The increase in
soil respiration from bryophyte-dominated towards lichen-dominated vegetation has also
been noted by Ohtonen and Väre (1998). The high soil respiration rate in this kind of
situation becomes understandable, if energetic costs are involved, as recalcitrant, lignified
compounds have a low concentration of nutrients (Wareing & Patrick 1975). Thus, if the
C/N ratio of the detritus is high, the decomposers may not need to maximize carbon
fixation in their biomass, and a greater proportion of the carbon can be used to supply
energy for the processes of nutrient uptake. The C/N ratio of organic matter has been
associated with high levels of recalcitrant compounds (Northup et al. 1995), which are
abundant in nutrient-deficient conditions. In pine forests, organic nitrogen is the main
nitrogen pool and is bound to compounds that are inaccessible to plants that are not
engaged in symbiosis with ericoid or ectomycorrhizal fungi (Northup et al. 1995,
Näsholm et al. 1998). The majority of plant detritus has a C/N ratio that is higher than the
critical value for microbes, suggesting that there is also nitrogen limitation among
decomposers. Plants and microbes are therefore potential competitors for nitrogen (Kaye
& Hart 1997).
Since plants also become more nutrient-limited during succession, they may produce
photosynthetically fixed carbon in excess to their physiological needs and have the option
of allocating the carbon to mycorrhizal fungi (Smith & Read 1997, IV). The increased
carbon allocation to mycorrhizae would increase their performance and contribute to the
increased basal respiration. The ratio of sporocarp yields of mycorrhizal to saprophytic
fungi is higher in dry, nutrient-poor soils than in moist, nutrient-rich ones (Väre et al.
1996). A similar trend can also be expected in fungal communities below-ground, which
means that the fungal community is less carbon-limited in nutrient poor than in nutrient
rich ecosystems.
Different units for expressing soil properties are appropriate when the soil is examined
as a medium where plants take up nutrients rather than as the habitat of decomposers (IV).
Vegetation studies and ecosystem level questions require that nutrient concentration, soil
respiration, microbial biomass and other measures of a similar nature are expressed per
unit area of the habitat or alternatively per unit volume of soil. Area is an ecosystemindependent measure that facilitates cross-ecosystem comparisons. The competitive
interactions of plants also take place in space and nutrients are taken from certain area (or
volume) of soil. The areal properties of soil are thus highly related to the performance of
a plant in a given habitat. When soil properties are expressed per unit mass of organic
matter, they describe the prevailing conditions in microbial subsystems, e.g. indicate the
decomposability of dead organic matter. Although decomposers live on the surfaces of
organic matter, the surface area of this complex, porous substrate is hard to analyse and
nutrients concentrations can not be expressed per unit area of organic matter. The
problems of averaging (Ruel & Ayers 1999, I) should also be considered when soil
processes are studied. When measurements are made of soil samples that are
homogenised for some volume, the measurements are averages for that volume.
Decomposers experience soil at a much more precise scale than is practical in analysis
methods. If the measurements are made from litres of soil volume, they can not indicate
the small scale variation of soil. Decomposers may choose favourable microhabitats
within that volume and experience greater nutrient concentrations that can be revealed by
the larger scale measurements.
31
4.2.2. Trophic interactions
The dynamics of nutrient cycling reflect the functioning of a whole ecosystem (DeAngelis
1992). Area and volume-based measurements enable comparisons between different
ecosystems and the different components of a given ecosystem. A common unit of
measurement is a requirement for the construction of models of nutrient cycling. The
input of nutrients to the system may be rather constant but the output of nutrients is
proportional to the nutrient content of a system compartment. If, say, the proportion of
nutrients that leach from a compartment is constant (as assumed in V), more nutrients
leach from a high concentration than from a low one, and the system may remain at a
stable state of nutrient content where the input and output of nutrients balance each other.
The nutrient cycling model that was considered in Paper V will reach an asymptotic
state where the nutrient content and distribution are stable. The rate of approach to the
stable state, i.e. return time from perturbations, can be interpreted from the dominant
eigenvalue of the matrix of nutrient transition rate between ecosystem compartments
(Neubert & Caswell 1997). The dominant eigenvalue of the matrix depends on all the
elements of the matrix and their contribution to the dynamics can be evaluated from an
elasticity or sensitivity analysis of the dominant eigenvalue (Caswell 1989, Heppel et al.
2000). The elasticity of a zero valued, "non-existent" flow is zero but the sensitivity of
such a transition is also zero if calculated according to the equation of Mesterton-Gibbons
(2000). This approach neglects the dynamic contribution of indirect interactions. The
transformation of a continuous time model to discrete time revealed the indirect
interactions (V) and enabled a more thorough investigation of the interactions of nutrient
cycling than continuous time models.
When a nutrient cycling model is constructed in continuous time the indirect
interactions are implicit but included in the dynamics, and can be revealed with a
transformation to discrete time. When a nutrient cycling model is constructed in discrete
time in the first place, the data may be insufficient for giving non-zero transition
probability estimates for indirect interactions. This is likely when the rate of nutrient
throughflow is small and the measurement of a nutrient flow is based on a technique
where the movement of a quantity of labelled nutrient is followed (Peterson & Fry 1989).
The greatest proportion of the labelled nutrient is likely to be found in compartments that
receive a direct flow from the compartment into which the labelled form was introduced.
Only small concentrations can accumulate in compartments that are remote from the
compartment into which the nutrient was originally introduced. It can therefore be
recommended that a nutrient cycling model is first constructed in continuous time and
then transformed to discrete time, if indirect interactions are to be studied (V). This is
especially important in systems with rapid dynamics, where the contribution of indirect
interactions is considerable. On the other hand, it is fortunate that the probability of
detecting indirect interactions is greatest in systems where they are of greatest dynamic
importance.
32
The proportion and dynamic contribution of indirect interactions has the potential to
increase with the number of compartments that take part in nutrient cycling (V). Indirect
interactions are of greatest importance in systems where each compartment only has one
direct flow to another compartment and the flow of nutrients is high compared to the
nutrient content of compartments. The number of species increases at least during the
early phases of primary succession, which suggests an increase of indirect interactions in
nutrient cycling. The unforested stages that are vegetated by herbaceous plants and
deciduous bushes attract herbivores, while forested stages probably experience less
herbivory. The early and late successional stages may therefore differ in the dynamics of
nutrient cycling in similar way to tundra and forest ecosystems (V). Odum (1969)
suggested a successional decrease in the rate of nutrient exchange (comparable to nutrient
throughflow rate). The probability matrix and elasticity analysis approach would reveal
this trend as a decrease in the probability of indirect interactions and their elasticities.
5. Conclusions
Nutrients play an important role at all organisational levels of an ecosystem. The growth
of an individual plant is crucially dependent or limited by the availability of nutrients and
other resources, and several strategies exist for resource competition. Different traits
facilitate resource competition under conditions of high vs. low resource availability. Low
nutrient availability requires strategies of efficient uptake of nutrients with mycorrhizal
symbionts and internal nutrient circulation. These traits have trade-offs in reducing the
maximal growth rate and the capacity for fast uptake of abundant nutrients. The symbiosis
with mycorrhizae is an energetic cost, which is only worth paying when nutrients, rather
than energy, limit plant growth. Environments where both nutrients and light are abundant
favour species that grow fast, and reproduce and colonise other temporarily non-vegetated
patches. Such ruderal species characteristically do not form mycorrhizae.
Variation in nutrient availability complicates the interactions involved in nutrient
competition. The traits that increase the affinity for plants to nutrients reduce the
tolerance of plants to variation in nutrient availability. For example, the growth of
mycorrhizal plants is adversely affected by variation in nutrient availability. The literature
suggests that mycorrhizal plants are able tolerate variation in external nutrient availability
to a greater extent than non-mycorrhizal plants. This may be a consequence of high
nutrient storage by mycorrhizal plants, which buffers physiological processes against
fluctuations in nutrient availability. The plants that benefit from variation in nutrients are
able to exploit peak nutrient availabilities efficiently. If the average nutrient availability of
a habitat is insufficient for a plant it may persist by effective utilisation of peak nutrient
availabilities.
Further trade-offs in plant strategies include the allocation of growth between the shoot
and roots. A limited potential for growth requires optimal allocation, which is greatly
dependent on the availability and ratio of nutrients and energy. Optimally, plant growth is
equally limited by all resources. The availability of nutrients and energy is variable and
consequently optimal allocation is also variable. A plant can not choose its habitat and its
persistence therefore depends on its ability to accommodate to variation. Phenotypic
plasticity of shoot/root allocation increases the range of environmental conditions that can
be tolerated. Environmental variation may be greater than the range tolerated by a plant.
34
The plasticity of allocation may be insufficient, especially when a plant grows under
conditions of low availability of negatively correlated resources.
Soil processes have an important contribution to the structure and species composition
of vegetation and the primary productivity of an ecosystem. A successional increase in
biomass and productivity may reverse to a decrease when a significant proportion of
nutrients is bound to the biomass and soil and a great proportion of the remaining soil
nutrients are leached. This leads to a plant community which produces litter of low
nutritional quality, causing the soil decomposer community to become dominated by
species that are able to utilise the recalcitrant organic compounds, which increase in
proportion in organic matter. Plants may increase carbon allocation to mycorrhizae when
nutrients become more limiting and decomposers may use more carbon per unit of
assimilated nutrient. This increases soil respiration rate in the low nutrient conditions
which prevail at the late successional stages.
Plant adaptive strategies, ecosystem development and trophic interactions are too
complex to be thoroughly examined in a single study. This thesis illuminates selected
aspects of the role of nutrients at the different levels of ecosystems and shows that the
functioning of a complex entity such as an ecosystem becomes tractable when general
processes such as resource acquisition and nutrient cycling are emphasised. Ecosystem
modelling does not have to mimic the whole complexity of a natural system but isolated
questions can be studied with simple mathematical tools.
6. References
Abrams P, Menge BA, Mittelbach GG, Spiller D & Yodzis P (1996) The role of indirect effects in
food webs. pp. 371–395 In: Polis GA & Winemiller KO (eds) Food Webs: Integration of
Patterns & Dynamics. Chapman & Hall, New York.
Ågren GI & Bosatta E (1996a) Quality: a bridge between theory and experiment in soil organic
matter studies. Oikos 76: 522–528.
Ågren GI & Bosatta E (1996b). Theoretical ecosystem ecology: understanding element cycles.
Cambridge University Press, Cambridge.
Aikio S (2000) Response to Mesterton-Gibbons. Trends Ecol Evol, in press.
Alestalo R (1971) Dendrochonological interpretation of geomorphic processes. Fennia 105.
Alestalo R (1979) Land uplift and development of the littoral and aeolian morphology on Hailuoto,
Finland. Acta Univ Oul A 82: 109–120.
Alexander I & Fairley RI (1983) Effects of N fertilization on populations of fine roots and
mycorrhizas in spruce humus. Plant Soil 74: 49–53.
Amijee F, Tinker PB & Stribley DP (1989) The development of endomycorrhizal root systems.
VII. A detailed study of effects of soil phosphorus on colonization. New Phytol 111: 435–446.
Anderson JPE & Domsch KH (1978) A physiological method for the quantitative measurement of
microbial biomass in soils. Soil Biol Biochem 10: 215–221.
Bloom AJ, Chapin FS & Mooney HA (1985) Resource limitation in plants - an economic analogy.
Annu Rev Ecol Syst 16: 363–392.
Bolan NS (1991) A critical review of the role of mycorrhizal fungi in the uptake of phosphorus by
plants. Plant Soil 134: 189–207.
Bougher NL, Grove TS & Malajczuk N (1990) Growth and phosphorous acquisition of karri
(Eucalyptus diversicolor F. Muell.) seedlings inoculated with ectomycorrhizal fungi in relation
to phosphorous supply. New Phytol 114: 77–85.
Bradshaw AD (1965) Evolutionary significance of phenotypic plasticity in plants. Adv Genet 13:
115–155.
Brundrett MC & Kendrick B (1988) The mycorrhizal status, root anatomy, and phenology of plants
in a sugar maple forest. Can J Bot 66: 1153–1173.
Carpenter SR, Kraft CE, Wright R, He X, Solanno PA & Hodgson JR (1992) Resilience and
resistance of a lake phosphorus cycle before and after food web manipulation. Am Nat 140:
781–798.
Caswell H (1989) Matrices in Population Biology. Sinauer, Sunderland, MA, USA.
Chapin III FS (1980) The mineral nutrition of wild plants. Annu Rev Ecol Syst 11: 233–260.
Chapin III FS (1991) Effects of multiple environmental stresses on nutrient availability. In:
Mooney MA, Winner WE & Pell EJ (eds) Response of Plants to Multiple Stresses. Academic
Press, London. pp. 67–88.
Chapin III FS, Barsdate RJ & Barèl D (1978) Phosphorus cycling in Alaskan coastal tundra: a
hypothesis for the regulation of nutrient cycling. Oikos 31: 189–199.
36
Chapin III FS, Walker LR, Fastie CL & Sharman LC (1994) Mechanisms of primary succession
following deglaciation at Glacier Bay, Alaska. Ecol Monogr 64: 149–175.
Char BW, Geddes KO, Gonnet GH, Leong BL, Monagan MB & Watt SM (1991) Maple V
Language Reference Manual. Springer-Verlag, New York.
Cody ML (1966) A general theory of clutch size. Evolution 20: 174–184.
Coleman T, Branch MA & Grace A (1999) Optimization Toolbox for use with Matlab. The
MathWorks Inc., Natick, MA, USA.
Crocker RL & Major J (1955) Soil development in relation to vegetation and surface age at Glacier
Bay, Alaska. J Ecol 43: 427–448.
Cui M & Caldwell MM (1996a) Facilitation of plant phosphate acquisition by arbuscular
mycorrhizas from enriched soil patches I. Roots and hyphae exploiting the same soil volume.
New Phytol 133: 453–460.
Cui M & Caldwell MM (1996b) Facilitation of plant phosphate acquisition by arbuscular
mycorrhizas from enriched soil patches II. Hyphae exploiting root-free soil. New Phytol 133:
461–467.
Currie WS (1999) The responsive C and N biogeochemistry of the temperate forest floor. Trends
Ecol Evol 14: 316–320.
DeAngelis D (1992) Dynamics of Nutrient Cycling and Food Webs. Chapman & Hall, USA.
Drury WH & Nisbet ICT (1973) Succession. J Arnold Arbor 54: 331–368.
Finlay R & Söderström B (1992) Mycorrhiza and carbon flow to the soil. pp. 134–160. In: Allen
MJ (ed) Mycorrhizal Functioning: an Integral Plant-Fungal Process. Chapman & Hall, London.
Fogel R & Hunt G (1983) Contribution of mycorrhizae and soil fungi to nutrient cycling in a
Douglas-fir ecosystem. Can J For Res 13: 219–232.
Glenn-Lewin DC, Peet RK & Veblen TT (eds 1992) Plant Succession. Chapman & Hall, London.
Grace JB & Tilman D (1990) Perspectives in Plant Competition. Blackwell, Oxford.
Grace JB (1991) A clarification of the debate between Grime and Tilman. Funct Ecol 5: 583–587.
Grime JP (1977) Evidence for the existence of three primary strategies in plants and its relevance to
ecological and evolutionary theory. Am Nat 111: 1169–1194.
Grime JP (1979) Plant Strategies and Vegetation Processes. John Wiley & Sons, New York.
Grime JP, Crick JC & Rincon JE (1986) The ecological significance of plasticity. pp. 5–29 In:
Jennings DH & Trewars AJ (eds) Plasticity in Plants. Symposio of the Society of Experimental
Biology, Cambridge.
Grover JP (1997) Resource Competition. Chapman & Hall, London.
Halonen O, Tulkki H & Derome J (1983) Nutrient analysis methods. Metsäntutkimuslaitoksen
tiedonantoja 121: 1–28.
Harborne JB (1988) Introduction to Ecological Biochemistry, 3. ed. Academic Press, London.
Harrison GW & Fekete S (1980) Resistance of nutrient cycling systems to perturbations of the flow
rates. Ecol Modell 10: 227–241.
Heppell S, Pfister C & de Kroon H (2000) Elasticity analysis in population analysis: methods and
applications. Ecology 81: 605–606.
Hill MO (1979) TWINSPAN - a Fortran Program for Analyzing Multicariate Data in an Ordered
Two-Way Table by Classification of Individual Attributes. Ithaca, New York.
Howeler RH, Asher CJ & Edwards DG (1982) Establishment of an effective endomycorrhizal
association on cassava in flowing solution culture and its effect on phosphorus nutrition. New
Phytol 90: 229–238.
Hutchings MJ & de Kroon H (1994) Foraging in plants: the role of morphological plasticity in
resource acquisition. Adv Ecol Res 25: 159–238.
Hutchinson GE (1957) Concluding remarks. Cold Spring Harbor Symposium on Quantitative
Biology 22: 415–427.
Ingestad T & Ågren GI (1988) Nutrient uptake and allocation at steady state-nutrition. Physiol
Plant 72: 450–459.
Janssen BH (1996) Nitrogen mineralization in relation to C:N ratio and decomposability of organic
materials. Plant Soil 181: 39–45
John MK (1970) Colorimetric determination of phosphorus in soil and plant materials with
ascorbic acid. Soil Science 100: 214–220.
37
Jongmans AG, van Breemen N, Lundström U, van Hees PAW, Finlay RD, Srinivasan M, Unestam
T, Giesler R, Melkerud P-A & Olsson M (1997) Rock-eating fungi. Nature 389: 682–683.
Kaye JP & Hart SC (1997) Competition for nitrogen between plants and soil microorganisms.
Trends Ecol Evol 12: 139–143.
Kleb HR & Wilson SD (1997) Vegetation effects on soil resource heterogeneity. Am Nat 150:
283–298.
Kruskal JB & Wish M (1978) Multidimensional Scaling. Sage, Beverly Hills, CA, USA.
Larcher W (1995) Physiological Plant Ecology, 3. ed. Springer Verlag, Berlin.
Levins R (1979) Coexistence in a variable environment. Am Nat 114: 765–783.
Lindgren BW (1976) Statistical Theory, 3. ed. Macmillan Publishing Company, New York.
Marschner H (1995) The Mineral Nutrition of Higher Plants. Academic Press, London.
Mary B, Recous S, Darwis D & Robin D (1996) Interactions between decomposition of plant
residues and niteogen cycling in soil. Plant Soil 181: 71–82.
Mesterton-Gibbons M (2000) A consistent equation for ecological sensitivity in matrix population
analysis. Trends Ecol Evol 15: 115.
Minchin PR (1988) DECODA: Database for Ecological Community Data. Anutech Pty Ltd.,
Canberra.
Mullen RB & Schmidt SK (1993) Mycorrhizal infection, phosphorus uptake, and phenology in
Ranunculus adoneus: implications for the functioning of mycorrhizae in alpine systems.
Oecologia 94: 229–234.
Mullen RB, Schmidt SK & Jaeger CH (1998) Nitrogen uptake during snow melt by the snow
buttercup, Ranunculus adoneus. Arct Alp Res 30: 121–125.
Näsholm T, Ekblad A, Nordin A, Giesler R, Högberg M & Högberg P (1998) Boreal forest plants
take up organic nitrogen. Nature 392: 914–916.
Neubert MG & Caswell H (1997) Alternatives to resilience for measuring the responses of
ecological systems to perturbations. Ecology 78: 653–665.
Nordgren A (1988) Apparatus for the continuous, long-term monitoring of soil respiration rate in
large numbers of samples. Soil Biol Biochem 20: 955–957.
Northup RR, Yu Z, Dahlgren RA & Vogt KA (1995) Polyphenol control of nitrogen release from
pine litter. Nature 377: 227–229.
Odum EP (1969) The strategy of ecosystem development. Science 164: 262–270.
Ohtonen R & Väre H (1998) Vegetation composition determines microbial activities in a boreal
forest soil. Microb Ecol 36: 328–335.
O'Neill RV, DeAngelis DL, Pastor JJ, Jackson BJ, & Post WM (1989) Multiple nutrient limitation
in ecological models. Ecol Modell 46: 147–163.
Pairunan A, Robson AD & Abbott LK (1980) The effectiveness of vesicular-arbuscular
mycorrhizas of subterranean clover in uptake of from phosphorus sources of different
solubilities. New Phytol 84: 327–338.
Peet RK (1992) Community structure and ecosystem function. pp. 103–151 In: Glenn-Lewin DC,
Peet RK & Veblen TT (eds) Plant Succession. Chapman & Hall, London.
Perez-Moreno J & Read DJ (2000) Mobilization and transfer of nutrients from litter to tree
seedlings via the vegetative mycelium of ectomycorrhizal plants. New Phytol 145: 301–309.
Peterson BJ & Fry B (1989) Stable isotopes in ecosystem studies. Ann Rev Ecol Syst 18: 293–320.
Polis GA & Winemiller WO (eds, 1996) Food Webs: Integration of Patterns & Dynamics.
Chapman & Hall, New York.
Rapport DJ (1971) An optimization model of food selection. Am Nat 105: 575–578.
Rastetter EB, Ågren GI & Shaver GR (1997) Responses of N-limited ecosystems to increased CO2:
a balanced-nutrition, coupled-element-cycles model. Ecol Appl 7: 444–460.
Rogers HH, Runion GB & Krupa SV (1994) Plant responses to atmospheric CO2 enrichment with
emphasis on roots and the rhizosphere. Pedobiologica 35: 313–325.
Ruel JJ & Ayers MP (1999) Jensen's inequality predicts effects of environmental variation. Trends
Ecol Evol 14: 361–366.
Rune O (1957) Serpentinicola element i Fennoscandiens flora. Sv Bot Tidskr 51: 43–105.
Rydin H & Borgegård S-O (1991) Plant characteristics over a century of primary succession on
islands: Lake Hjälmaren. Ecology 72: 1089–1101.
38
Scheiner SM (1993) Genetics and evolution of phenotypic plasticity. Annu Rev Ecol Syst 24: 35–
68.
Setälä H, Rissanen J & Markkola AM (1997) Conditional outcomes in the relationship between
pine and ectomycorrhizal fungi in relation to the biotic and abiotic environment. Oikos 80: 112–
122.
Smith SE & Read DJ (1997) Mycorrhizal Symbiosis, 2nd ed. Academic Press, London.
Stitt M & Krapp A (1999) The interaction between elevated carbon dioxide and nitrogen nutrition:
the physiological and molecular background. Plant Cell Env 22: 583–621.
Strauss SY (1991) Indirect effects in community ecology: their definition, study and importance.
Trends Ecol Evol 6: 206–209.
Stribley DP, Tinker PB & Rayner JH (1980) Relation of internal phosphorus concentration and
plant weight in plants infected by vesicular-arbuscular mycorrhiza. New Phytol 86: 261–266.
Sultan SE (1987) Evolutionary implications of phenotypic plasticity in plants Evol Biol 21: 127–
178.
Thomson BD, Robson AD & Abbot LK (1986) Effects of phosphorus on the formation of
mycorrhizas by Gigaspora calospora and Glomus fasciculatum in relation to root
carbohydrates. New Phytol 103: 751–765.
Tilman D (1982) Resource Competition and Community Structure. Princeton University Press,
Princeton.
Tilman D (1985) The resource ratio hypothesis of plant succession. Am Nat 125: 827–852.
Tilman D (1988) Plant Strategies and the Dynamics and Structure of Plant Communities. Princeton
University Press, Princeton.
Titus JH & del Moral R (1998) Vesicular-arbuscular mycorrhizal influence on Mount St. Helens
pioneer species in greenhouse experiments. Oikos 81: 495–510.
Väre H, Ohenoja E & Ohtonen R (1996) Macrofungi of oligotrophic Scots pine forests in northern
Finland. Karstenia 36: 1–18.
Visser S & Parkinson D (1992) Soil biological criteria as indicators of soil quality: Soil
microorganisms. Am J Alt Agr 7: 33–37.
Wardle DA & Giller KE (1997) The quest for a contemporary ecological dimension to soil ecology.
Soil Biol Biochem 28: 1549–1554.
Wardle DA & Ghani A (1995) A critique of the microbial metabolic quotient (qCO2) as a
bioindicator of disturbance and ecosystem development. Soil Biol Biochem 27: 1601–1610.
Wareing PF & Patrick J (1975) Source-sink relations and partition of assimilates in the plant. pp.
481–500 In: Cooper JP (ed) Photosynthesis and Productivity in Different Environments.
Cambridge University Press, Cambridge.
Webster JR, Waide JB & Patten BC (1975) Nutrient recycling and stability of ecosystems. pp. 1–27
In: Howell F, Gentry JB & Smith MH (eds) Mineral Cycling in Southeastern Ecosystems.
ERDA Conf-750513, Math. Tech. Inf. Serv., Dept. Commer., Springfield, Virginia, USA.
Winemiller KO (1996) Factors driving temporal and spatial variation in aquatic floodplanin food
webs. pp. 298–312 In: Polis GA & Winemiller KO (eds) Food Webs: Integration of Patterns &
Dynamics. Chapman & Hall, New York.
Woodward FI, Thompson GB & McKee IF (1991) The effects of elevated concentrations of carbon
dioxide in individual plants, populations, communities and ecosystems. Ann Bot 67: 23–38.
Yodzis P (1996) Food webs and perturbation experiments: theory and practice. pp. 192–200 In:
Polis GA & Winemiller KO (eds) Food Webs: Integration of Patterns & Dynamics. Chapman &
Hall, New York.
Zak DR, Pregitzer KS, Curtis PS, Teeri JA, Fogel R & Randlett DL (1993) Elevated atmospheric
CO2 and feedback between carbon and nitrogen cycles. Plant Soil 151 105–117.