Combining demography and genetic analysis to assess the population structure of an amphibian in a human-dominated landscape

In this article, we applied demographic and genetic approaches to assess how landscape features influence dispersal patterns and genetic structure of the common frog Rana temporaria in a landscape where anthropogenic perturbations are pervasive (urbanization and roads). We used a combination of GIS methods that integrate radiotracking and landscape configuration data, and simulation techniques in order to estimate the potential dispersal area around breeding patches. Additionally, genetic data provided indirect measures of dispersal and allowed to characterise the spatial genetic structure of ponds and the patterns of gene flow across the landscape. Although demographic simulations predicted six distinct groups of habitat patches within which movement can occur, genetic analyses suggested a different configuration. More precisely, BAPS5 spatial clustering method with ponds as the analysis unit detected five spatial clusters. Individual-based analyses were not able to detect significant genetic structure. We argue that (1) taking into account that each individual breeds in specific breeding patch allowed for better explanation of population functioning, (2) the discrepancy between direct (radiotracking) and indirect (genetic) estimates of subpopulations (breeding patches) is due to a recent landscape fragmentation (e.g. traffic increase). We discuss the future of this population in the face of increasing landscape fragmentation, focusing on the need for combining demographic and genetic approaches when evaluating the conservation status of population subjected to rapid landscape changes.     

Invasion and persistence of plant parasites in a spatially structured host population

Spatial structure is of central importance in the dynamics of plant-parasite interactions and is imposed by the growth habit and distribution of host plants and by parasite dispersal which is frequently restricted. To investigate the effects of spatial heterogeneity on the dynamics of plant parasites we introduce a simple model for epidemic development within a spatially structured host population. Here the host population is subdivided into a number of patches which are linked to allow for transmission from one patch to another with the connections defining the spatial structure of the host population. Three key parameters are identified that play a critical role in the ability of the parasite to invade and persist within the host population: the within-patch parasite basic reproductive number which characterises the infection dynamics at the local spatial scale; and the neighbourhood of interaction which describes which patches interact with which and the strength of coupling between patches within the neighbourhood which together characterise the spread of the parasite over larger spatial scales. Using both deterministic and stochastic formulations of the model, we investigate how the thresholds and probabilities of invasion and persistence are affected by these parameters, by demographic stochasticity and by differences in the initial level of infection.     

Rapid increase in dispersal during range expansion in the invasive ladybird Harmonia axyridis

The evolutionary trajectories associated with demographic, genetic and spatial disequilibrium have become an issue of growing interest in population biology. Invasive species provide unique opportunities to explore the impact of recent range expansion on life‐history traits, making it possible to test for a spatial arrangement of dispersal abilities along the expanding range, in particular. We carried out controlled experiments in laboratory conditions to test the hypothesis of an increase in dispersal capacity with range expansion in Harmonia axyridis, a ladybird that has been invading Europe since 2001. We found a marked increase in the flight speed of the insects from the core to the front of the invasion range in two independent sampling transects. By contrast, we found that two other traits associated with dispersal (endurance and motivation to fly off) did not follow the same spatial gradient. Our results provide a striking illustration of the way in which predictable directional genetic changes may occur rapidly for some traits associated with dispersal during biological invasions. We discuss the consequences of our results for invasion dynamics and the evolutionary outcomes of spatially expanding populations.     

Impacts of gene flow and logging history on the local genetic structure of a scattered tree species, Sorbus torminalis L. Crantz

Sorbus torminalis L. Crantz is a colonizing tree species usually found at low density in managed European forests. Using six microsatellite markers, we investigated spatial and temporal patterns of genetic structure within a 472-ha population of 185 individuals to infer processes shaping the distribution of genetic diversity. Only eight young stems were found to be the result of vegetative reproduction. Despite high levels of gene flow (standard deviation of gene dispersal = 360 m), marked patterns of isolation by distance were detected, associated with an aggregated distribution of individuals in approximately 100-m patches. This spatial structure of both genes and individuals is likely to result from patterns of seedling recruitment combined with low tree density. Our results suggest that landscape factors and logging cycles markedly shape the distribution of favourable sites for seedling establishment, which are then colonized by sibling cohorts as a result of joint seed transportation by frugivores. These combined genetic and demographic processes result in similar genetic structure both within and among logging units. However, conversion to high forest may enhance genetic structuring.     

Dendritic network structure and dispersal affect temporal dynamics of diversity and species persistence

Landscape connectivity structure, specifically the dendritic network structure of rivers, is expected to influence community diversity dynamics by altering dispersal patterns, and subsequently the unfolding of species interactions. However, previous comparative and experimental work on dendritic metacommunities has studied diversity mostly from an equilibrium perspective. Here we investigated the effect of dendritic versus linear network structure on local (α‐diversity), among (β‐diversity) and total (γ‐diversity) temporal species community diversity dynamics. Using a combination of microcosm experiments, which allowed for active dispersal of 14 protists and a rotifer species, and numerical analyses, we demonstrate the general importance of spatial network configuration and basic life history tradeoffs as driving factors of different diversity patterns in linear and dendritic systems. We experimentally found that community diversity patterns were shaped by the interaction of dispersal within the networks and local species interactions. Specifically, α‐diversity remained higher in dendritic networks over time, especially at highly connected sites. β‐diversity was initially greater in linear networks, due to increased dispersal limitation, but became more similar to β‐diversity in dendritic networks over time. Comparing the experimental results with a neutral metacommunity model we found that dispersal and network connectivity alone may, to a large extent, explain α‐ and β‐diversity dynamics. However, additional mechanisms, such as variation in carrying capacity and competition–colonization tradeoffs, were needed in the model to capture the detailed temporal diversity dynamics of the experiments, such as a general decline in γ‐diversity and long‐term dynamics in α‐diversity.     

Population Dynamics Modelling in an Hierarchical Arborescent River Network: An Attempt with Salmo trutta

The balance between births and deaths in an age-structured population is strongly influenced by the spatial distribution of sub-populations. Our aim was to describe the demographic process of a fish population in an hierarchical dendritic river network, by taking into account the possible movements of individuals. We tried also to quantify the effect of river network changes (damming or channelling) on the global fish population dynamics. The Salmo trutta life pattern was taken as an example for.We proposed a model which includes the demographic and the migration processes, considering migration fast compared to demography. The population was divided into three age-classes and subdivided into fifteen spatial patches, thus having 45 state variables. Both processes were described by means of constant transfer coefficients, so we were dealing with a linear system of difference equations. The discrete case of the variable aggregation method allowed the study of the system through the dominant elements of a much simpler linear system with only three global variables: the total number of individuals in each age-class.From biological hypothesis on demographic and migratory parameters, we showed that the global population dynamics of fishes is well characterized in the reference river network, and that dams could have stronger effects on the global dynamics than channelling.     

Effects of landscape structure on the invasive spread of black cherry Prunus serotina in an agricultural landscape in Flanders, Belgium

Analysing invasive spread from a landscape ecological perspective forms an important challenge in plant invasion ecology. The present study examines the effects of landscape structure on the spatial and temporal dynamics of an expanding black cherry Prunus serotina population within a rural landscape in Flanders, Belgium, carrying a dense network of interconnected hedgerows. The study area, 251 ha in size, harboured a total of 2962 P. serotina individuals. The population was characterised by a negative exponential age distribution, a high growth rate and an early and continuous reproduction throughout the species' life cycle. The historical rate of spread of the species through the hedgerow network progressively increased with time, especially during the last decade. Spatial point pattern analysis revealed that the individuals had a significantly clustered distribution pattern and were spatially aggregated around seed sources, hedgerow intersections and roosting trees. Logistic regression analysis confirmed the effect of landscape structure on P. serotina occurrence, suggesting directional long distance dispersal by avian dispersal vectors, resulting in a differential seed pressure throughout the hedgerow network due to the preference of dispersing birds for roosting in structurally rich hedgerow with large trees near hedgerow intersections. Hence, the distribution of P. serotina in agricultural landscapes was strongly mediated by dispersal processes. Furthermore, decreasing spatial aggregation along the species life cycle, with especially seedlings and saplings being significantly aggregated while adult individuals were mostly distributed at random, and a relative outward shift of seedling recruitment curves with time indicate density dependent mortality, probably caused by intraspecific competition.     

Estimating range expansion of wildlife in heterogeneous landscapes: A spatially explicit state‐space matrix model coupled with an improved numerical integration technique

Dispersal as well as population growth is a key demographic process that determines population dynamics. However, determining the effects of environmental covariates on dispersal from spatial‐temporal abundance proxy data is challenging owing to the complexity of model specification for directional dispersal permeability and the extremely high computational loads for numerical integration. In this paper, we present a case study estimating how environmental covariates affect the dispersal of Japanese sika deer by developing a spatially explicit state‐space matrix model coupled with an improved numerical integration technique (Markov chain Monte Carlo with particle filters). In particular, we explored the environmental drivers of inhomogeneous range expansion, characteristic of animals with short dispersal. Our model framework successfully reproduced the complex population dynamics of sika deer, including rapid changes in densely populated areas and distribution fronts within a decade. Furthermore, our results revealed that the inhomogeneous range expansion of sika deer seemed to be primarily caused by the dispersal process (i.e., movement barriers in fragmented forests) rather than population growth. Our state‐space matrix model enables the inference of population dynamics for a broad range of organisms, even those with low dispersal ability, in heterogeneous landscapes, and could address many pressing issues in conservation biology and ecosystem management.     

Comparing short and long‐distance dispersal: modelling and field case studies

Dispersal is a factor of great importance in determining a species spatial distribution. Short distance dispersal (SDD) and long distance dispersal (LDD) strategies yield very different spatial distributions. In this paper we compare spatial spread patterns from SDD and LDD simulations, contrast them with patterns from field data, and assess the significance of biological and population traits. Simulated SDD spread using an exponential function generates a single circular patch with a well‐defined invasion front showing a travelling‐wave structure. The invasive spread is relatively slow as it is restricted to reproductive individuals occupying the outer zone of the circular patch. As a consequence of this dispersal dynamics, spread is slower than spread generated by LDD. In contrast, the early and fast invasion of the entire habitat mediated by power law LDD not only involves a significantly greater invasion velocity, but also an entirely different habitat occupation. As newly dispersed individuals soon reach very distant portions of the habitat as well as the vicinity of the original dispersal focus, new growing patches are generated while the main patch increases its own growth absorbing the closest patches. As a consequence of both dispersal and lower density dependence, growth of the occupied area is much faster than with SDD. SDD and LDD also differ regarding pattern generation. With SDD, fractal patterns appear only in the border of the invasion front in SDD when competitive interaction with residents is included. In contrast, LDD patterns show fractality both in the spatial arrangements of patches as well as in patch borders. Moreover, values of border fractal dimension inform on the dispersal process in relation with habitat heterogeneity. The distribution of patch size is also scale‐free, showing two power laws characteristic of small and large patch sizes directly arising from the dispersal and reproductive dynamics. Ecological factors like habitat heterogeneity are relevant for dispersal, although its importance is greater for SDD, lowering the invasion velocity. Among the life history traits considered, adult mortality, the juvenile bank and mean dispersal distance are the most relevant for SDD. For LDD, habitat heterogeneity and changes in life history traits are not so relevant, causing minor changes in the values of the scale‐free parameters. Our work on short and long distance dispersal shows novel theoretical differences between SDD and LDD in invasive systems (mechanisms of pattern formation, fractal and scaling properties, relevance of different life history traits and habitat variables) that correspond closely with field examples and were not analyzed, at least in this degree of detail, by the previously existing models.     

Host–parasitoid spatial models: the interplay of demographic stochasticity and dynamics

Host–parasitoid metapopulation models have typically been deterministic models formulated with population numbers as a continuous variable. Spatial heterogeneity in local population abundance is a typical (and often essential) feature of these models and means that, even when average population density is high, some patches have small population sizes. In addition, large temporal population fluctuations are characteristic of many of these models, and this also results in periodically small local population sizes. Whenever population abundances are small, demographic stochasticity can become important in several ways. To investigate this problem, we have reformulated a deterministic, host–parasitoid metapopulation as an integer-based model in which encounters between hosts and parasitoids, and the fecundity of individuals are modelled as stochastic processes. This has a number of important consequences: (1) stochastic fluctuations at small population sizes tend to be amplified by the dynamics to cause massive population variability, i.e. the demographic stochasticity has a destabilizing effect; (2) the spatial patterns of local abundance observed in the deterministic counterpart are largely maintained (although the area of ''spatial chaos'' is extended); (3) at small population sizes, dispersal by discrete individuals leads to a smaller fraction of new patches being colonized, so that parasitoids with small dispersal rates have a greater tendency for extinction and higher dispersal rates have a larger competitive advantage; and (4) competing parasitoids that could coexist in the deterministic model due to spatial segregation cannot now coexist for any combination of parameters.     

Effects of Culling on Mesopredator Population Dynamics

Anthropogenic changes in land use and the extirpation of apex predators have facilitated explosive growth of mesopredator populations. Consequently, many species have been subjected to extensive control throughout portions of their range due to their integral role as generalist predators and reservoirs of zoonotic disease. Yet, few studies have monitored the effects of landscape composition or configuration on the demographic or behavioral response of mesopredators to population manipulation. During 2007 we removed 382 raccoons (Procyon lotor) from 30 forest patches throughout a fragmented agricultural ecosystem to test hypotheses regarding the effects of habitat isolation on population recovery and role of range expansion and dispersal in patch colonization of mesopredators in heterogeneous landscapes. Patches were allowed to recolonize naturally and demographic restructuring of patches was monitored from 2008–2010 using mark-recapture. An additional 25 control patches were monitored as a baseline measure of demography. After 3 years only 40% of experimental patches had returned to pre-removal densities. This stagnant recovery was driven by low colonization rates of females, resulting in little to no within-patch recruitment. Colonizing raccoons were predominantly young males, suggesting that dispersal, rather than range expansion, was the primary mechanism driving population recovery. Contrary to our prediction, neither landscape connectivity nor measured local habitat attributes influenced colonization rates, likely due to the high dispersal capability of raccoons and limited role of range expansion in patch colonization. Although culling is commonly used to control local populations of many mesopredators, we demonstrate that such practices create severe disruptions in population demography that may be counterproductive to disease management in fragmented landscapes due to an influx of dispersing males into depopulated areas. However, given the slow repopulation rates observed in our study, localized depopulation may be effective at reducing negative ecological impacts of mesopredators in fragmented landscapes at limited spatial and temporal scales.     

Life history determines genetic structure and evolutionary potential of host-parasite interactions

Measures of population genetic structure and diversity of disease-causing organisms are commonly used to draw inferences regarding their evolutionary history and potential to generate new variation in traits that determine interactions with their hosts. Parasite species exhibit a range of population structures and life-history strategies, including different transmission modes, life-cycle complexity, off-host survival mechanisms and dispersal ability. These are important determinants of the frequency and predictability of interactions with host species. Yet the complex causal relationships between spatial structure, life history and the evolutionary dynamics of parasite populations are not well understood. We demonstrate that a clear picture of the evolutionary potential of parasitic organisms and their demographic and evolutionary histories can only come from understanding the role of life history and spatial structure in influencing population dynamics and epidemiological patterns.     

Classical metapopulation dynamics and eco‐evolutionary feedbacks in dendritic networks

Eco‐evolutionary dynamics are now recognized to be highly relevant for population and community dynamics. However, the impact of evolutionary dynamics on spatial patterns, such as the occurrence of classical metapopulation dynamics, is less well appreciated. Here, we analyse the evolutionary consequences of spatial network connectivity and topology for dispersal strategies and quantify the eco‐evolutionary feedback in terms of altered classical metapopulation dynamics. We find that network properties, such as topology and connectivity, lead to predictable spatio‐temporal correlations in fitness expectations. These spatio‐temporally stable fitness patterns heavily impact evolutionarily stable dispersal strategies and lead to eco‐evolutionary feedbacks on landscape level metrics, such as the number of occupied patches, the number of extinctions and recolonizations as well as metapopulation extinction risk and genetic structure. Our model predicts that classical metapopulation dynamics are more likely to occur in dendritic networks, and especially in riverine systems, compared to other types of landscape configurations. As it remains debated whether classical metapopulation dynamics are likely to occur in nature at all, our work provides an important conceptual advance for understanding the occurrence of classical metapopulation dynamics which has implications for conservation and management of spatially structured populations.     

Coexistence of competitors in metacommunities due to spatial variation in resource growth rates; does R * predict the outcome of competition?

Simple mathematical models are used to investigate the coexistence of two consumers using a single limiting resource that is distributed over distinct patches, and that has unequal growth rates in the different patches. Relatively low movement rates or high demographic rates of an inefficient resource exploiter allow it to coexist at a stable equilibrium with a more efficient species whose ratio of movement to demographic rates is lower. The range of conditions allowing coexistence depends on the between‐patch heterogeneity in resource growth rates, but this range can be quite broad. The between‐patch movement of the more efficient consumer turns patches with high resource growth rates into sources, while low‐growth‐rate patches effectively become sinks. A less efficient species can coexist with or even exclude the more efficient species from the global environment if it is better able to bias its spatial distribution towards the source patches. This can be accomplished with density independent dispersal if the less efficient species has a lower ratio of per capita between‐patch movement rate to demographic rates. Conditions that maximize the range of efficiencies allowing coexistence of two species are: a relatively high level of heterogeneity in resource growth conditions; high dispersal (or low demographic rates) of the superior competitor; and low dispersal (or high demographic rates) of the inferior competitor. Global exclusion of the more efficient competitor requires that the inferior competitor have sufficient movement to also produce a source‐sink environment.     

Evolution of dispersal in metapopulations with local density dependence and demographic stochasticity

In this paper, we predict the outcome of dispersal evolution in metapopulations based on the following assumptions: (i) population dynamics within patches are density-regulated by realistic growth functions; (ii) demographic stochasticity resulting from finite population sizes within patches is accounted for; and (iii) the transition of individuals between patches is explicitly modelled by a disperser pool. We show, first, that evolutionarily stable dispersal rates do not necessarily increase with rates for the local extinction of populations due to external disturbances in habitable patches. Second, we describe how demographic stochasticity affects the evolution of dispersal rates: evolutionarily stable dispersal rates remain high even when disturbance-related rates of local extinction are low, and a variety of qualitatively different responses of adapted dispersal rates to varied levels of disturbance become possible. This paper shows, for the first time, that evolution of dispersal rates may give rise to monotonically increasing or decreasing responses, as well as to intermediate maxima or minima.     

Comparative host-parasite population structures: disentangling prospecting and dispersal in the black-legged kittiwake Rissa tridactyla

Although much insight is to be gained through the comparison of the population genetic structures of parasites and hosts, there are, at present, few studies that take advantage of the information on vertebrate life histories available through the consideration of their parasites. Here, we examined the genetic structure of a colonial seabird, the black-legged kittiwake (Rissa tridactyla) using seven polymorphic microsatellite markers to make inferences about population functioning and intercolony dispersal. We sampled kittiwakes from 22 colonies across the species' range and, at the same time, collected individuals of one of its common ectoparasites, the tick Ixodes uriae. Parasites were genotyped at eight microsatellite markers and the population genetic structure of host and parasite were compared. Kittiwake populations are only genetically structured at large spatial scales and show weak patterns of isolation by distance. This may be due to long-distance dispersal events that erase local patterns of population subdivision. However, important additional information is gained by comparing results with those of the parasite. In particular, tick populations are strongly structured at regional scales and show a stepping-stone pattern of gene flow. Due to the parasite's life history, its population structure is directly linked to the frequency and spatial extent of within-breeding season movements of kittiwakes. The comparison of host and parasite gene flow therefore helps us to disentangle the intercolony movements of birds from that of true dispersal events (movement followed by reproduction). In addition, such data can provide essential elements for predicting the outcome of local co-evolutionary interactions.     

The effect of habitat fragmentation on dispersal patterns, mating behavior, and genetic variation in a pika (Ochotona princeps) metapopulation

 Habitat fragmentation is becoming increasingly common, yet, the effect of habitat spatial structure on population dynamics remains undetermined for most species. Populations of a single species found in fragmented and nonfragmented habitat present a rare opportunity to examine the effect of habitat spatial structure on population dynamics. This study investigates the impact of highly fragmented habitat on dispersal patterns, mating behavior, and genetic variation in a pika (Ochotona princeps) population with a mainland-island spatial structure. Juvenile dispersal patterns in fragmented habitat revealed that individuals tended to disperse to neighboring habitat patches. However, within-patch band-sharing scores from multilocus DNA fingerprints did not differ from what would be expected if individuals were assorting randomly among habitat patches each year. Multiple, short-distance dispersal targets for juveniles and occasional long-distance dispersal events suggest that habitat fragmentation on this scale has not resulted in restricted dispersal and a genetically subdivided population. Although pikas tended to mate with the closest available partner, DNA fingerprinting band-sharing scores between mated pairs were consistent with a random mating hypothesis. Random mating in this population appears to be an incidental effect of dispersal in a fragmented habitat. This pattern is distinct from that found in nonfragmented habitat (large talus patches) where mating was non-random and consistent with mating between individuals of intermediate relatedness. DNA fingerprinting data revealed within-species variation in the mating habits of the pika directly attributable to habitat spatial structure. Received: 4 November 1996 / Accepted: 30 June 1997     

Loss of animal seed dispersal increases extinction risk in a tropical tree species due to pervasive negative density dependence across life stages

Overhunting in tropical forests reduces populations of vertebrate seed dispersers. If reduced seed dispersal has a negative impact on tree population viability, overhunting could lead to altered forest structure and dynamics, including decreased biodiversity. However, empirical data showing decreased animal-dispersed tree abundance in overhunted forests contradict demographic models which predict minimal sensitivity of tree population growth rate to early life stages. One resolution to this discrepancy is that seed dispersal determines spatial aggregation, which could have demographic consequences for all life stages. We tested the impact of dispersal loss on population viability of a tropical tree species, Miliusa horsfieldii, currently dispersed by an intact community of large mammals in a Thai forest. We evaluated the effect of spatial aggregation for all tree life stages, from seeds to adult trees, and constructed simulation models to compare population viability with and without animal-mediated seed dispersal. In simulated populations, disperser loss increased spatial aggregation by fourfold, leading to increased negative density dependence across the life cycle and a 10-fold increase in the probability of extinction. Given that the majority of tree species in tropical forests are animal-dispersed, overhunting will potentially result in forests that are fundamentally different from those existing now.     

The effect of migration on metapopulation stability is qualitatively unaffected by demographic and spatial heterogeneity

Coupled map lattices (CMLs), using two coupled logistic equations, have been extensively used to model the dynamics of two-patch ecological systems. Such studies have revealed that migration rate plays an important role in determining the dynamics of the system, particularly when the two maps differ in their intrinsic growth rate parameter, r. However, under more realistic assumptions, a metapopulation can be expected to consist of more than two subpopulations, each with its own demographic parameters, which will in part be a function of the environment of that patch. The role of the spatial arrangement of heterogeneous (i.e. with different r values) subpopulations in shaping the dynamics of such a metapopulation has rarely been investigated. Here, we study the effect of demographic and spatial heterogeneity on the stability of one- and two-dimensional systems of 64 coupled Ricker maps with different r values, under periodic and absorbing boundary conditions. We show that the effects of migration rate on metapopulation stability do not depend upon either the precise spatial arrangement of the subpopulations in the lattice, or on the presence of a moderate proportion of vacant (uninhabitable) patches in the lattice. The results, thus, suggest that metapopulation models are robust to variation in spatial arrangement of patch quality and, hence, of demographic parameters. We also show that for any given arrangement of the patches, maximum stability of the metapopulation occurs when the migration levels are intermediate, a result that agrees well with previous studies on two-map CML systems.     

Viability analyses with habitat-based metapopulation models

Population viability analysis (PVA) models incorporate spatial dynamics in different ways. At one extreme are the occupancy models that are based on the number of occupied populations. The simplest occupancy models ignore the location of populations. At the other extreme are individual-based models, which describe the spatial structure with the location of each individual in the population, or the location of territories or home ranges. In between these are spatially structured metapopulation models that describe the dynamics of each population with structured demographic models and incorporate spatial dynamics by modeling dispersal and temporal correlation among populations. Both dispersal and correlation between each pair of populations depend on the location of the populations, making these models spatially structured. In this article, I describe a method that expands spatially structured metapopulation models by incorporating information about habitat relationships of the species and the characteristics of the landscape in which the metapopulation exists. This method uses a habitat suitability map to determine the spatial structure of the metapopulation, including the number, size, and location of habitat patches in which subpopulations of the metapopulation live. The habitat suitability map can be calculated in a number of different ways, including statistical analyses (such as logistic regression) that find the relationship between the occurrence (or, density) of the species and independent variables which describe its habitat requirements. The habitat suitability map is then used to calculate the spatial structure of the metapopulation, based on species-specific characteristics such as the home range size, dispersal distance, and minimum habitat suitability for reproduction. Received: April 1, 1999 / Accepted: October 29, 1999