Evolution of life-histories in stochastic environments:
Cole’s paradox revisited
David Tesar
Department of Ecology and Systematics Division of Population Biology
University of Helsinki Finland
Academic dissertation
To be presented, with the permission of the Faculty of Science of the University of Helsinki, for public criticism in the Lecture Room of the Department of
Ecology and Systematics, P. Rautatiekatu 13, Helsinki, on December 1, 2000, at 12 o’clock noon.
Helsinki 2000
Author’s address:
Integrative Ecology Unit
Department of Ecology and Systematics Division of Population Biology
P.O. Box 17 (Arkadiankatu 7) FIN-00014 University of Helsinki e-mail: david.tesar@helsinki.fi
ISBN 952-91-2582-8 (nid) ISBN 952-91-2583-5 (PDF)
Helsinki 2000
Oy Edita Ab
Life history evolution in stochastic environments iii
Evolution of life-histories in stochastic environments:
Cole’s paradox revisited
David Tesar
Integrative Ecology Unit
Department of Ecology and Systematics Division of Population Biology
P.O. Box 17 (Arkadiankatu 7) F
IN-00014 University of Helsinki, Finland
This thesis is based on the following articles which are referred to in the text by their Roman numerals:
I Tesar, D. 2000: Lamont Cole’s study on the population consequences of life history phenomena: An ongoing challenge. — Manuscript.
II Ranta, E., V. Kaitala, S. Alaja, and D. Tesar. 2000. Nonlinear dynamics and the evolution of semelparous and iteroparous reproductive strategies. — The American Naturalist 155, 294-300
III Ranta, E., D. Tesar, S. Alaja, and V. Kaitala. 2000. Does evolution of iteroparous and semelparous reproduction call for spatially structured systems? — Evolution, 54, 145-150.
IV Tesar, D., V. Kaitala, and E. Ranta. 2000. Stochasticity and spatial coexistence of semelparity and iteroparity as life histories. — Manuscript.
V Ranta, E., D. Tesar, and V. Kaitala. 2000. Environmental variability and semelparity vs. iteroparity as life histories. — Manuscript.
VI Ranta, E., D. Tesar, and V. Kaitala. 2000. Local extinctions promote coexistence of semelparous and iteroparous life histories. — Manuscript.
VII Kaitala, V., D. Tesar, and E. Ranta. 2000. Semelparity vs. iteroparity and the number of age classes. — Manuscript.
CONTRIBUTIONS
I II III IV V VI VII
Original
idea IKP IKP IKP IKP IKP IKP DT,VK
Study design and methods
DT IKP IKP IKP IKP ER IKP
Manuscript
preparation DT DT,ER
VK,SA
DT,ER SA,VK
DT,VK ER
DT,ER VK
DT,ER VK
VK,DT ER IKP = Integrative Ecology Unit, here DT, ER and VK.; (SA = Susanna Alaja)
Supervised by: Prof. ESA RANTA
University of Helsinki Finland
Prof. VEIJO KAITALA
University of Helsinki/ University of Jyväskylä Finland
Reviewed by: Prof. PER LUNDBERG
University of Lund Sweden
Doc. JUHA MERILÄ
University of Uppsala Sweden
Examined by: Doc. KARI LEHTILÄ
University of Stockholm Sweden
EVOLUTION OF LIFE HISTORIES IN STOCHASTIC ENVIRONMENTS: COLE’S PARADOX REVISITED
1. Introduction
The large diversity of life histories asks for classification into smaller groups of common characteristics. Classifying life histories based on the number of reproductive events per lifetime of an organism and hence creating semelparity (type of life history with only one reproductive event) and iteroparity (type of life history with more than one reproductive event) was introduced by Cole (1954).
Cole (1954) was interested in finding the evolutionary cause why some organisms reproduce only once in their entire life while others reproduce repeatedly. Cole took a demographic approach and built a simple model which lead him to conclude that semelparity is a reproductive strategy favoured before iteroparity, since a semelparous reproductive type, so he reasoned, would be easily capable of reproducing one more offspring to achieve the same growth rate as an iteroparous reproductive type. This finding does not correspond to the actual spread of semelparity in natural populations. Later this finding was coined with the name “Cole’s paradox”.
The first serious attempt to solve Cole’s paradox was undertaken by Murphy (1968). He concluded from computer simulations that environmental fluctuations favour the iteroparous reproductive strategy. His finding was supported by the intuitively obvious idea that an organism’s reproductive success is increased the more reproductive events fall into “good” seasons for reproduction. Murphy (1968) further discussed relevant life history components and their implications in the evolution of semelparity and iteroparity. Many of his ideas and the suggestion to differentiate between stability of juvenile survival and stability of the adult survival, were taken and advanced further in more recent works (Charnov and Schaffer 1973, Schaffer 1974).
The biggest deficiency of Cole’s (1954) theoretical model was the lack of mortality for the iteroparous type. With the introduction of juvenile and adult survival rates, Cole’s paradox was to a large extent solved by Charnov and Schaffer (1973). They proposed that evolution to semelparity would occur when the ratio between adult and juvenile survival rate would fall below a certain value, depending on the fecundities of semelparity and iteroparity.
Schaffer (1974) added more complexity into this problem by suggesting that semelparity was favoured by natural selection when adult survival was uncertain, while iteroparity was favoured when fecundity was uncertain. His analysis of the
impact of environmental fluctuations on life histories was based on static version of Levins model (1967).
A further milestone was reached when density dependence was introduced into the analysis of life histories (Bulmer 1985). Based on the assumption of stable population dynamics and placed into an invasion scenario with an abundant resident reproductive strategy and a rare mutant reproductive strategy that tries to invade the resident population, Bulmer (1985) proposed two inequalities, which determine when semelparity can invade iteroparity and vice versa (see Fig. 1 on page 9).
It has increasingly become clear that life histories are not just the result of some constant extrinsic or intrinsic impacts on vital rates and other life history traits, but in most cases, consequences of uncertainties in the environment, the demography or dispersal. Hence, the attention to the semelparity – iteroparity problem has moved towards analysing the effect of fluctuations in the vital rates on the adaptive success of reproductive strategies.
Orzack and Tuljapurkar (1989) investigated the effect of environmental fluctuations on the growth rate of a range of iteroparous life histories. They found that uncertainty in fecundity does not necessarily lead to an iteroparous reproductive strategy as Schaffer (1974) predicted. Based on the fitness measure of equal life time weighted reproduction and on the condition of equal environmental variation, Orzack and Tuljapurkar (1989) concluded that a life history with dispersed reproduction is equally fit as a slightly longer living life history, with reproduction concentrated in the last few age-classes, mitigating Murphy’s (1968) findings.
In a thorough investigation of reproductive strategies in stochastic and density-dependent environments Benton and Grant (1999a) found that the optimal reproductive strategy in a given system depends on a number of factors, such as deterministic population dynamics, what vital rates are affected by stochastic processes, how those vital rates are correlated with each other, what their distribution is, from which the variation of the stochastic process is drawn, and what the ESS in the deterministic case is. Benton and Grant (1999a) differentiate between the effect of density-dependence and variation stemming from catastrophic events (density-independent effects) that occur periodically and reduce vital rates to a value of 0.01. They extend Schaffer’s (1974) result by suggesting that when fecundity is submitted predominantly to density- independent fluctuations, and survival is affected by density-dependent effects, then iteroparity as an optimal reproductive strategy is to be expected. On the other hand, when survival is submitted to predominantly density-independent fluctuations, and fecundity is affected by density-dependent effects, then semelparity as an optimal reproductive strategy is to be expected. Benton and Grant (1999a) further realize that if density dependence acts on survival, fecundity will vary less strongly and semelparity will be favoured. And if
Life history evolution in stochastic environments 3 density dependence acts on fecundity, survival will vary comparatively less strongly and iteroparity will be favoured. The question, why do some organisms reproduce only once in their life while others reproduce several times, is — despite of tremendous recent advancements — still very much on the agenda of modern evolutionary research. The study of semelparity and iteroparity has often made use of simple models. The challenge in the further study of this topic lies in the use of more realistic dynamic models.
2. Objective of the thesis
A number of investigations in the evolution of life histories assume constant life history parameters (I). As pointed out above, the investigation of the evolution of life histories has recently extended onto the dynamical aspect of evolution.
For a full understanding of the evolution of life histories, it is not enough to analytically determine under what constant conditions semelparity or iteroparity is favoured by natural selection under stable conditions. This should be especially clear as we all know that natural populations are subject to stochastic processes of various kind. Until about a decade ago powerful tools to investigate dynamical aspects of evolution were missing or rare. Today, modern desktop computers allow simulations of increasingly complex population systems, studies on the impact of stochasticity on long-term persistence of different reproductive strategies can be launched.
Why is it important to investigate the conditions for coexistence between semelparity and iteroparity? Foremost, it is of fundamental evolutionary interest to know what the conditions for coexistence between different reproductive strategies are. Fitness of an individual is measured as its contribution in units of offspring, that survive at least to maturity, relative to the total population in the next generation. In assuming a particular fitness measure (e.g., intrinsic rate of increase), different life history types of equal fitness can be compared with each other and it can be determined under what conditions they can potentially coexist. The condition for coexistence between different reproductive strategies in constant environments are usually easily determined when the fitnesses of the considered reproductive strategies are equated. Typically only one coexistence point exists in such simple environments.
Temporal variability in vital rates, demographic stochasticity or stochastic dispersal tend to distort this simple coexistence condition. A stochastic process affecting given life histories results in an unpredictability of which reproductive type will outcompete the other ones. Eventually several life history types remain in the system on the long term, despite differences in reproductive strategy and fitness level.
The lifetime reproductive rate, R0, exclusive and the intrinsic rate of increase, r, partially refer to stationary populations and cannot be used as a fitness
measures in the context of variable environments (Pásztor et al. 1996). In highly variable environments or in stochastic environments the best fitness measure that exists (Metz et al. 1992) is the dominant Lyapunov exponent. It is the summation of logarithmic changes in the population size over a given period of time. The dominant Lyapunov exponent is directly applicable to an invasion scenario where a rare mutant strategy tries to invade an abundant resident strategy for a limited period of time. A positive dominant Lyapunov exponent or invasion exponent, as it is also referred to in ecological context (Heino 1998), indicates that the mutants can invade.
It is of paramount evolutionary interest to know what factors promote coexistence between life history types (e.g., iteropary, semelpary), and what type of variation enables long-term persistence of a life-history strategy in its ambient environment with another life history. In contrast to constant environments, the conditions for coexistence in fluctuating environments include a range of parameter combinations mainly depending on the presence of a spatial structure, nonlinear population dynamics, demographic stochasticity and stochastic dispersal.
The main goal of this thesis work was to investigate the potential for coexistence of semelparity with iteroparity. A departure for answering that question was to investigate, what impact nonlinear dynamics (or complex, or chaotic), spatial linking of populations, local extinctions, environmental stochasticity and stochastic dispersal have on the potential of semelparity and iteroparity to coexist.
3. Summary
The study of the evolution of life histories has a central role in modern ecological research (Roff 1992, Stearns 1992). In (I) I review the large body of literature that exists on the evolution of semelparity and iteroparity concepts which were introduced by Cole (1954). The review presents the ideas which lead Cole (1954) to construct his simple model and shows how the model was further elaborated. Alternative approaches to Cole’s original approach and recent advancements in the study of semelparity and iteroparity are presented. Finally, a number of components are identified which add complexity to the study of reproductive strategies. I conclude that these components need to be studied further in order to better understand why some organisms reproduce only once in their life while others reproduce repeatedly.
Initially, the question about the fitness trade-off between semelparity and iteroparity was approached in an analytical way (Cole 1954, Gadgil and Bossert 1970, Bryant 1971, Charnov and Schaffer 1973, Bell 1976, Bulmer 1985), keeping environmental parameters constant. More recently, interest has been
Life history evolution in stochastic environments 5 devoted to the effect of environmental variability on the evolution of life histories (Orzack and Tuljapurkar 1989, Benton and Grant 1999a).
Life histories are shaped by the properties of both physical and biological environment, including conspecifics and individuals from other species. Such environments are commonly affected by stochastic processes and show a spatial structure. However, the role of stochastic processes and spatial structure on the evolution of life histories is still poorly understood (I). One possible approach is to investigate how stochastic processes affect the parameter space of coexistence between different life history types, and how the assumption of spatial structure, i.e., dispersal or level of isolation betweeen sub-populations, affects the parameter range under which coexistence is possible.
Fig. 1. Graphical representation of Bulmer’s (1985) model delimiting the parameter space where semelparity is ESS (hatched triangle) and the parameter space where iteroparity is ESS (above hatched triangle). Bulmer’s model allows coexistence between semelparity and iteroparity only exactly on the borderline between these parameter spaces (hypothenuse of hatched triangle; notation after III).
Bulmer’s (1985) discrete-time population model is a starting point for most of my research on the evolution of life histories. The model determines for which survival and fecundity combinations semelparity wins over iteroparity and vice versa. It also shows a linear parameter space in where coexistence is possible (see Fig. 1). In this model the geometric growth rate of a semelparous and an iteroparous population are represented in the following way:
S =PJbS, and I =PJbI +PA (1)
0 0.25 0.5 0.75 1
0 0.25 0.5 0.75 1
Adult survival, P A In this area
an iteroparous life history cannot invade a semelparous population
In this area a semelparous life history cannot invade an iteroparous population
P is the survival rate and b the offspring number or fecundity. The subscripts S and I stand for ‘semelparous’ and ‘iteroparous’, and J and A for ‘juvenile’ and
‘adult’. Bulmer has elaborated his initial density dependence function of his population model in 1985, and published an elegant solution a decade later (Bulmer 1994) accounting for density dependence affecting juvenile survival:
PJ = pjexp{− [bSNS(k)+bINI(k)]} (2) where pj is the maximum juvenile survival that is achieved without density effects, N is the population density, and k represents the number of reproductive seasons.
Equation (2) allows, in contrast to Bulmer’s (1985) initial model, to investigate, also for fluctuating populations, under which conditions a semelparous reproductive type can invade an iteroparous reproductive type and under which conditions an iteroparous type can invade a semelparous reproductive type and finally what are the conditions that allow coexistence between two reproductive strategies?
In (II) we address these questions in a density-dependent population model with nonlinear population dynamics by means of computer simulations. We chose the geometric population growth rate of the semelparous population so that nonlinear population dynamics would result. The parameter combinations of the vital rates, as in (IV, V and VI), were chosen so that semelparity would be favoured according to Bulmer’s (1985) inequality. To our surprise, coexistence between iteroparity and semelparity was nevertheless possible. In the entire range of the geometric growth rate causing nonlinear (regularly oscillating, chaotic) dynamics coexistence between semelparity and iteroparity was possible. Despite the fact that the invader (iteroparous type) population when seperated from the resident (semelparous type) population performed stable population dynamics, at the same geometric growth rate, combined with the second population via a density dependence function, the iteroparous population showed nonlinear dynamics as well. We made an extended search for parameters allowing invasion of one breeding strategy into the other. Generally, invasion by a semelparous breeder appeared to be easier than invasion by an iteroparous breeder. Parameter combinations allowing invasions were usually highly sensitive to vital rates and to the scaling factor between semelparous and iteroparous number of offspring.
From (II) it has become clear that nonlinear dynamics would enable coexistence between iteroparity and semelparity. We therefore had to choose the geometric growth rate of the semelparous population to be in the range resulting in stable population dynamics. We further extended the analysis of invasion possibilities by using spatially structured populations (III). Separate populations of pure reproductive strategies were linked with dispersal. First, we tracked invasion attempts in a spatial system in which dispersal was not affected by distance (spatially implicit model). Next, we followed invasion attempts in a
Life history evolution in stochastic environments 7 spatial system in which dispersal was affected by distance (spatially explicit model). A calculation of dominant Lyapunov exponents allowed us a more exact investigation of when invasions by a rare breeding type were possible. The results from the spatial exploration of invasion possibilities stand in contrast to results from spatially unstructured models, where coexistence for the selected parameters is not possible. We summarized the invasion outcomes of the implicit vs. explicit spatial model graphically and noted that both reproductive strategies as rare mutants can invade a resident population with a different breeding strategy. Successful semelparous invasions have shown to depend only marginally on whether space was assumed to be implicit or explicit. No such difference was found in the case of iteroparous invasions. An analysis of the population dynamics during an invasion revealed that the invader after exceeding a threshold density causes nonlinear dynamics which also affect the resident population. Such nonlinear dynamics appear after a transient period possibly lasting several hundred time steps. This transient phase is variable, depending on the vital parameter values but also on the spatial configuration values. We have compared local population time-series and have found that the semelparous population often fluctuates in opposite phase to the iteroparous population.
The invasion possibilities of semelparity and iteroparity in a spatial context are further explored in (IV). In contrast to (III) where the complexity in the system arose from combined nonlinear population dynamics, in (IV) the complexity stems from the stochasticity of dispersal. For this second spatial investigation of the parameter range of coexistence, we used a new substantially simplified model. Space was represented by a row of 1002 habitable patches.
Initially only the first patch was occupied by a semelparous breeding type and the last patch by an iteroparous breeding type. Each time step and for every patch a random proportion of the population in the patch dispersed equally to the neighbouring patches, with the exception at the habitat ends, where the dispersers were reflected. This represents a bi-directional process, which should lead to a dispered global population over the entire one-dimensional habitat.
Here again, in order to exclude that only nonlinear population dynamics cause coexistence between the two breeding strategies, we chose the population growth rate to be low.
The outcome indicated that coexistence of semelparity with iteroparity was possible, provided the variation in the dispersal fractions was high enough.
Local dispersal has resulted in coexistence at ranges for geometric growth rates which would not have permitted coexistence in a spatially unstructured model with stable population dynamics. Further, numerical simulations have shown that different ranges of dispersal variability can have an immense effect on population density of particularly the semelparous population (IV). Also the autocorrelation structure of the dispersal variability was shown to affect population densities and coexistence between semelparity and iteroparity.
Negatively autocorrelated dispersal variability enabled the largest parameter range
for coexistence between the two reproductive strategies. In other words, in an environment in which a low dispersal rate is followed in the subsequent year by a high dispersal rate and vice versa offered better conditions for different life history types to coexist. It is not surprising that the populations in the patches fluctuate more asynchronously if dispersal variability occur on a local, as opposed to, a global scale.
In (V) we investigated how environmental variability affects the range of coexistence between semelparity and iteroparity in a density-dependent population model. This problem was approached in the following way: A density-dependent system with a semelparous and an iteroparous population was submitted to variation in fecundity, juvenile or adult survival. The variation which created by a noise function capable of producing autocorrelated environmental time-series, can act either only on one or several vital rates. We believe that environmental variation in natural populations, though not with the same sensitivity, affects all vital rates simultanously. Nevertheless we analyse also the situation where only part of the vital rates is affected by environmental variation.
In simulations where only one vital rate is varying, we found no potential for coexistence. With at least two of the three vital rates varying, a wide range of parameter combinations allowed coexistence between semelparity and iteroparity.
With uncertain juvenile survival and fecundity coexistence was not possible and with the chosen Bulmer (1985) inequality, semelparity was an ESS (Maynard Smith 1982). A more complex picture emerged when adult and juvenile survival rates were varying. Three broad parameter regions emerged, where either semelparity or iteroparity was an ESS or both strategies coexisted. In contrast, Bulmer’s (1985) simple model predicts rather a linear than a broad parameter region. Testing for coexistence resulting from autocorrelated environmental time- series, we conclude that autocorrelation structure in environmental variability had only marginal effect on coexistence. On the other hand, when evaluating to what extent the range of variability was responsible for coexistence, we found that a smaller variability range reduced the potential for coexistence.
In our third spatial investigation of the invasion possibilities between semelparity and iteroparity (VI), we focused on the function of local extinctions promoting coexistence between the two reproductive strategies. We used an implicit spatially structured model with 100 populations, in between which dispersal occurred unaffected by distance and added demographic stochasticity to dynamics of the local populations. Dispersal was implemented as a constant proportion of the local population leaving the patch. Demographic stochasticity was realized by drawing random numbers from a normal distribution and adding them to the population density in each patch at the end of each time step.
Our results show that coexistence between semelparity and iteroparity is possible in a spatially structured population system with demographic
Life history evolution in stochastic environments 9 stochasticity and dispersal. Demographic stochasticity in this system is crucial for coexistence to occur.
We conclude that coexistence between the two breeding strategies is possible due to spatial structure, demographic stochasticity, stablizing the dynamics on the one hand, and creating empty patches on the other hand. Dispersal enables recolonization of the empty patches.
A topic largely neglected in the literature is long-lived semelparity. Most studies investigating the fitness trade-off between semelparity and iteroparity assumed an annual life history type as the semelparous reproductive strategy. We took the Leslie matrix approach (Leslie 1945, 1948) to model the semelparous and iteroparous population in a combined density-dependent system. We aimed to find under which conditions semelparity is an ESS, under which conditions iteroparity is an ESS and when both reproductive strategies can coexist. First, we explored the parameter space keeping the number of age-classes equal. We found that all present/absent combinations were possible. For certain parameter combinations (VII) both reproductive strategies went extinct. In the second step, we increased the number of age-classes of the semelparous reproductive type but kept the number of age-classes of the iteroparous type below the one of semelparity. Here, for a wide range of parameter combinations the semelparous reproductive strategy was an ESS. When exploring the parameter range where iteroparity has a higher number of age-classes than the semelparous type, we found that either of the strategies was an ESS in part of the parameter range.
When the geometric growth rate of iteroparity exceeded that of semelparity, then iteroparity was an ESS. Which breeding type was an ESS often depended not only on the number of age-classes and geometric growth rates, but also on which strategy was resident and which invading.
4. General discussion Evidence for coexistence
A review of the main contributions to the study of the evolution of semelparity and iteroparity (I) has revealed numerous ideas and hypotheses about which natural conditions should favour iteroparity and which should favour semelparity (Murphy 1968, Charnov and Schaffer 1973, Hart 1977, Roff 1992, Stearns 1992, Charlesworth 1994). Some of these ideas should be viewed with reservations, e.g., the hypothesis that environmental variation favours iteroparity may be too simple (Roff 1992). My investigation has demonstrated that environmental variation can promote coexistence between a semelparous and an iteroparous reproductive strategy, and hence, supports Roff’s doubt about environmental stochasticity clearly favouring iteroparity. A further claim which was mitigated to some extent (II), summarized by Charlesworth (1994, p. 249), states that iteroparity should be favoured when adult survival is high and the
ratio between semelparous to iteroparous number of offspring is low. I found long-term persistence of both reproductive strategies even for extreme juvenile and adult survival values.
In my search for conditions enabling coexistence between semelparity and iteroparity I demonstrate numerically for which parameter combinations the fitness trade-off between semelparity and iteroparity can be eliminated. I have focussed on factors affecting life histories such as nonlinear population, dynamics, spatial structure, environmental stochasticity, demographic stochasticity, stochastic dispersal and the impact of longevity of the various reproductive strategies on the coexistence between them was analysed.
The main conclusions of my thesis are:
• Nonlinear dynamics caused by a high geometric growth rate in a density- dependent population model can promote coexistence between semelparous and iteroparous breeding systems (II).
• A spatial structure separating stable populations of pure reproductive strategy linked together by dispersal can also be counted as a factor promoting coexistence between semelparity and iteroparity (III).
• Stochastic dispersal in a spatially structured model was able to promote coexistence where spatial structure alone did not permit coexistence (IV).
• Negative autocorrelation in the dispersal fraction time-series enabled coexistence better than no or positive autocorrelated time-series (IV).
• Similarly, also demographic stochasticity reinforces the potential for coexistence (VI).
• The widely debated impact of environmental stochasticity on life history evolution seems to be a further factor promoting coexistence between the semelparous and the iteroparous type (V).
• On the other hand, for the rarely debated topic of long-lived semelparous organisms I found that all presence/absence combinations between a semelparous and an iteroparous population could occur (VII).
Not much literature exists on the coexistence between semelparity with iteroparity under the impact of stochastic processes. Schaffer (1974) studied convex-concave curves of reproductive effort. Such models permit coexistence between semelparity and iteroparity as optimal life history strategies if there is variation in the reproductive effort among individuals. Also Orzack and Tuljapurkar (1989) consider the possibility of two different iteroparous reproductive strategies coexisting in a fluctuating environment. Orzack and Tuljapurkar’s (1989) results are based on an investigation considering stochastic growth rate and correlation between vital rates.
The occurrence of natural species with similar ecological requirements but with completely different reproductive strategies (Schaffer 1974, Schaffer and Schaffer 1979, Young and Augspurger 1991), suggests that if there is a fitness
Life history evolution in stochastic environments 11 trade-off between reproducing semelparously and iteroparously, then it can be mitigated in natural populations possibly by stochastic processes in the environment.
Concluding remarks on semelparity and iteroparity
I took a demographic approach based on numerical simulations to investigate the fitness merits of the two different reproductive strategies, semelparity and iteroparity. As a main analytical tool I used an invasion scenario which allowed me to investigate the adaptive success of a certain breeding type affected by stochastic process(es).
Irrespective of using invasion of a rare breeding strategy into a different resident breeding strategy as a fitness measure, it became increasingly clear that invasibility was a major component of the semelparous life history. Several indices suggest this: First, it’s larger number of offspring facilitates invasions as compared to invasions by the iteroparous reproductive strategy, as noted in (II).
Second, in competition with the iteroparity the recovery potential from low densities of the semelparous type is best met in density-independent environments, as pointed out in (I). Third, smaller minimal viable semelparous than iteroparous populations, as reported in (II), facilitate the long-term persistence of newly established populations. And fourth, descriptions of habitat of semelparous organisms often indicate frequent habitat disturbances (Hart 1977, Pitelka 1977), which make an area newly colonizable. Colonization is associated with a great risk, but the gain that can be achieved in terms of fitness within one breeding season surpasses that of iteroparity. The evolution of semelparity can also be viewed under many other aspects. For example, space or resource limitations may affect the adaptative value of semelparity as well as autocorrelation structure can modify it (IV).
The dominating breeding strategy in nature is iteroparity. The principle of bet-hedging (Philippi and Seger 1989), i.e., spreading reproduction over an extended period of time or over an extended area, may mostly contribute to that fact. Even some semelparous plants use bet-hedging by establishing a seed bank, to evade the uncertain fate of semelparity. Further the phenomenon of parental care in iteroparous organisms, makes sure that the invested reproductive effort bears “fruits” after it has reached maturity. As a matter of fact, iteroparity includes such a large diversity of life histories that semelparity might be considered only as an extreme case of iteroparity. Orzack and Tuljapurkar (1989), as a matter of fact, consider only iteroparous breeding strategies and focus on the question which circumstances cause concentration or spread of reproduction in an organism’s life history.
Not all aspects affecting the evolution of life histories could be covered in this thesis. The role of extinction risk in the evolution of semelparity or
iteroparity, respectively, would have been a further research project. I touched upon that topic only marginally (IV, V). The aspects that I covered in this thesis, i.e., the effect of environmental and demographic stochasticity, spatial structure and stochastic dispersal on the evolution of semelparity and iteroparity, have been dealt with specific methods in special models. One should bear in mind that the approach used here is associated with a number of assumptions, most of which I have delineated above. The interpretation of the results should occur in this light.
5. Acknowledgements
Many thanks to ESA RANTA, who gave me the opportunity to work in his lab and guided me through the work. Also to VEIJO “TA-DAA” KAITALA go many thanks. In the final stages to play ping pong with unfinished manuscripts via email with him, was a pleasure. I thank them especially for the enjoyable collaboration and the supervision of my thesis during the past 26 months it has taken from scattered ideas to completion. I am also thankful to the rest of the IKP team, some of whom have been important in showing me the correct key combination to get a special symbol into this text or in giving me and my wife a crucial advice which lead us to find our future home: ANNA, ANNE, CRAIG, HEIKKI, KATRIINA, MARIANNE, MIKKO H., MIKKO K., NINA, SAMI, SUSANNA A., SUSANNA P. HANNU RITA helped to formulate expression (5) in Chapter (VI).
To the following persons I am grateful because they either helped me out when I needed to have technical problems solved before public presentations, or helped me to get official forms or were so kind to translate documents, because of my ignorance of the Finnish language. SUSANNA ALAJA, ANNE LUOMA, MARIANNE FRED, TEIJA SEPPÄ, ILPO HANSKI and HEIKKI HIRVONEN.
To a next group of people I am grateful because they helped me through the administrative process, which, as a rumour tells, I could not understand even if I would speak Finnish. TIINA AIRAMO, HANNU PIETIÄINEN, and TEIJA SEPPÄ.
I also would like to thank PER LUNDBERG and JUHA MERILÄ who kindly reviewed my thesis. Last but definitely not least I want to thank my wife, HENRIIKKA, for being patient and understanding when my work extended off office hours, as it happened in the final stages of my PhD studies.
This is a contribution of the plant ecology program (68453/47573) and the MaDaMe program (72400/49643) supported by the Finnish Academy.
Life history evolution in stochastic environments 13
6. References
Bell, G. 1976. On breeding more than once. The American Naturalist 110: 57- 77.
Benton, T.G., and A. Grant. 1999a. Optimal reproductive effort in stochastic, density-dependent environments. Evolution 53: 677-688.
———. 1999b. Elasticity analysis as an important tool in evolutionary and population ecology. TREE 14: 467-471.
Bryant, E.H. 1971. Life history consequences of natural selection: Cole's result.
The American Naturalist 104: 75-76.
Bulmer, M.G. 1985. Selection for iteroparity in a variable environment. The American Naturalist 126: 63-71.
———. 1994. Theoretical Evolutionary Ecology. Sinauer Associates, Sunderland, MA, USA.
Charlesworth, B. 1994. Evolution in age-structured populations. Cambridge University Press, Cambridge, UK.
Charnov, E.L., and W.M. Schaffer. 1973. Life-history consequences of natural selection: Cole's result revisited. The American Naturalist 107: 791-793.
Cole, L.C. 1954. The population consequences of life history phenomena.
Quart. Rev. Biol. 29: 103-137.
Gadgil, M. and W.H. Bossert. 1970. Life historical consequences of natural selection. The American Naturalist 104: 1-24.
Hart, R. 1977. Why are biennials so few? The American Naturalist 111: 792- 799.
Heino, M. 1998. Evolution of age schedules of reproduction and environmental feedback. PhD dissertation, University of Helsinki.
Levins, R. 1967. Evolution in changing environments. Princeton University Press, Princeton, NJ, USA.
Maynard Smith, J. 1982. Evolution and the theory of games. Cambridge University Press, Cambridge, UK.
Metz, J.A.J., R.M. Nisbet, and S.A.H. Geritz. 1992. How should we define
‘fitness’ for general ecological scenarios? TREE 7: 198-202.
Murphy, G.I. 1968. Pattern in life history and the environment. The American Naturalist 102:391-403.
Orzack, S.H., and S. Tuljapurkar. 1989. Population dynamics in variable environments. VII. The demography and evolution of iteroparity. The American Naturalist 133: 901-923.
Pásztor, E., G. Meszéna, and E. Kisdi. 1996. R0 or r: A matter of taste? Journal of Evolutionary Biology 9: 511-516.
Pitelka, L.F. 1977. Energy allocation in annual and perennial lupines (Lupinus:
Leguminosae). Ecology 58: 1055-1065.
Roff, D.A. 1992. The evolution of life histories: Theory and analysis. Chapman
& Hall. New York.
Schaffer, W.M. 1974. Optimal reproductive effort in fluctuating environments.
The American Naturalist 108: 783-790.
Schaffer, W. M., and M. V. Schaffer. 1979. The adaptive significance of variations in reproductive habit in the Agavaceae II: Pollinator foraging behavior and selection for increased reproductive expenditure. Ecology.
60: 1051-1069.
Stearns, S.C. 1992. The evolution of life histories. Oxford University Press, Oxford, UK.
Young, T.P., and C.K. Augspurger. 1991. Ecology and Evolution of Long-lived Semelparous Plants. TREE 6: 285-289.
