The concept of ecological exchangeability, together with genetic exchangeability, is central to both the Cohesion Species Concept as well as to some definitions of Evolutionarily Significant Units. While there are well-established criteria for measuring genetic exchangeability, the concept of ecological exchangeability has generated considerable confusion. We describe a procedure that uses the complementary strengths, while recognising the limitations, of both molecular genetic data and ecological experiments to determine the ecological exchangeability of local populations within a species. This is the first synthesis of a combined approach (experiments and genetics) and the first explicit discussion of testing ecological exchangeability. Although it would be ideal to find functional genes that interact to influence quantitative traits resulting in ecological differences (e.g. growth, size, fecundity), we suggest that our current knowledge of functional markers is too limited for most species to use them to differentiate adaptively different local populations. Thus, we argue that ecological experiments using whole organisms combined with neutral markers that indicate evolutionary divergence, provide the strongest case for detecting adaptive differences among local populations. Both genetic divergence and ecological experiments provide the best information for infering ecological exchangeability. This procedure can be used to decide which local populations should be preserved to maintain intraspecific variation and to determine which populations would enhance captive-breeding programs, augment endangered local populations and could best be used to re-introduce native species into historically occupied areas.

The elephant population in Kruger National Park, Republic of South Africa, is growing rapidly. To prevent damage to the Park's ecosystems, the management has culled about 7% of the population annually. Such culls are very controversial. At first glance, contraceptives seem an attractive alternative means of control. We examine contraception as a management option, review the relevant aspects of elephant reproduction, physiology and demography and conclude that this optimism is probably misplaced. First, contraceptives have a wide range of physiological and behavioural side-effects that may prove to be damaging to the individual female and those around her. Second, the elephants in the Park have near-maximal growth rates with inter-calving intervals of less than four years. To achieve zero population growth, about three-quarters of the adult female elephants would need to be on contraceptives. There are no simple alternatives. The smallest numerical target for controlling population numbers is to kill or sterilize females about to become pregnant for the first time. Such a solution is unlikely to appease those who consider killing elephants to be unethical. It may, however, be the one closest to the natural patterns of elephant mortality.

We carried out extensive field surveys in the dry forest portions of Madagascar to document the species of bats occurring in these regions. These data combined with information in the literature and museum specimen records indicate that 28 species of Chiroptera occur in this region of the island, of which we documented 27 during our inventories. The community composition at sites occurring in areas of water-eroded sedimentary rock is notably different from sites on alluvial substrates. In contrast to the majority of native land mammal species on Madagascar, much of the microchiropteran fauna is not dependent on large tracts of intact forest and anthropogenic perturbations of forests may have less direct impact on their long-term survival. Conservation strategies for Chiroptera in the dry regions of the island should focus on reducing various types of human disturbance of cave environments.

Vocal individuality is widely suggested as a method for identifying individuals within a population. But few studies have explored its performance in real or simulated conservation situations. Here we simulated the use of vocal individuality to monitor the calling corncrake (Crex crex), a secretive and endangered land rail. Our data set contained 600 calls from 30 individuals and was used to simulate a population of corncrakes being counted and monitored. We tested three different neural network models for their ability to discriminate between and to identify individuals. Neural networks are non-linear classification tools widely applied to both biological and non-biological identification tasks. Backpropagation and probabilistic neural networks were used to simulate the reidentification of members of a known population (monitoring) and a Kohonen network was used to simulate the counting of a population of unknown size (census). We found that both backpropagation and probabilistic networks identified all individuals correctly all the time, irrespective of sample size. Kohonen networks were more variable in performance but estimated population size to within one individual of the actual size. Our results indicate that neural networks can be used effectively together with recordings of vocalizations in census and monitoring tasks.

In recent decades, the abundance of otter, Lutra lutra, decreased in many European countries. This study aimed at determining if varying occurrence of otters is related to land features and if this occurs at many spatial scales. Values of landscape features were compared among three categories of otter occurrence, i.e. absent, sporadic and common, for 25 European countries, for 19 regions and 22-40 departments in France. At all scales, significantly higher human and road densities were detected in areas where otters are absent, which illustrates the pervasive impact of human presence on otter population fragmentation. Moreover, particular natural features (‘percentage other use’ defined as non-agricultural and non-forested) may also limit otter populations at the scale of Europe. Our study failed to detect significant differences in landscapes occupied by common or sporadic otter populations, probably owing to a limited sample size, or to the resilience of otters. The conservation of the European otter in Europe would benefit from an improved list of common large-scale land descriptors, the recognition of large gaps of naturally unsuitable habitats, and the development of international large-scale conservation studies to support existing regional otter conservation efforts.

Invasive alien species pose a significant threat to biodiversity worldwide and many eradication programmes are now underway in an effort to reduce the impact they may have on native species and ecosystems. The spatial structure of such invasive species populations is likely to have important implications for designing effective control strategies. Here, a simple source–sink population model is used to address the following question: if a population of an invasive alien species is source–sink in nature, what is the best way of dividing limited resources for its control? Results from this model indicate that allocation of resources solely to the source population does not always result in the most effective control strategy. The most efficient control measure is determined by the relative strengths (net gains and losses) of the source and sink populations and, crucially, the nature of dispersal between them. We present a case study for the control of an invasive species illustrating the use of these types of model.

The European mink Mustela lutreola is one of Europe's most endangered carnivores, with few vulnerable populations remaining. Surprisingly, a recent phylogeny placed a single mink specimen within the polecat (M. putorius, M. eversmannii) group, suggesting a recent speciation and/or the effects of hybridization. The analysis has now been extended to a further 51 mink and polecats. As before, phylogenetic methods failed to resolve the relationships between the species. One haplotype (C11) was found in both species, and predominated in European mink from Spain and eastern Europe. The known M. lutreola fossils are of very young date, so either mink arose recently, or else the situation is confused by hybridization and a biased fossil recovery. The study highlights the dangers of using a single genetic marker in defining Evolutionarily Significant Units (ESUs). Polecats and European mink are clearly distinct in their morphology and ecology, and should still be considered as separate ESUs, but without further data it is difficult to define Management Units. Following the precautionary principle, we recommend that for the moment European mink in eastern Europe (Belarus, Estonia and Russia) and Spain should be managed separately.

Between the years 1994 and 2000, we assessed the mortality rate of adult-sized, migrant Asian houbara bustards (Chlamydotis [undulata] macqueenii) via banding and satellite tracking programmes. The birds were captured either on their breeding grounds in Kazakhstan (Taukum desert and Buzachi peninsula) and in China (Jungar basin) or on their wintering grounds in Pakistan (Cholistan desert) and in the United Arab Emirates (Baynunah area). For the banding programme, 104 houbara bustards were marked and statistical analysis was performed on the number of transmitters and rings returned by falconers. This allowed us to calculate a 0.045 (SD 0.061) annual mortality rate attributable to hunting activities. The tracking programme analysed the fate of 41 houbara bustards equipped with satellite transmitters (PTTs). Overall, the satellite-tracked birds were followed for 10,965 days and had travelled all over the distribution range of the species in Asia. The causes of mortality were assessed from the interpretation of the ARGOS data relayed by the sensors embarked on the PTTs. The annual mortality rate was derived from the calculation of the average finite daily survival rate using the Trent & Rongstad (1974) method. With this method, the overall annual mortality rate in adult-sized houbara bustards migrating through Asia was 0.283 (95%lcl 0.186-95%ucl 0.404). The hunting and poaching pressure could explain 73.5% of the observed mortality. The mortality rate during winter, when the houbara is under severe hunting pressure, was 11.2 times higher than during the breeding season. Houbara appear to be especially at risk in Pakistan, Afghanistan and Iran. A VORTEX simulation showed that under current hunting and poaching pressures, the probability of extinction of the houbara population modelled can be anticipated at 50 years with a 94% probability. The maximum sustainable yield was estimated at 7.2% of the adult-sized population whereas as much as 20.8% is currently taken. This work stresses the need for conservationists and Arab falconers to take appropriate concerted measures as soon as possible.

Managers trying to preserve populations of endangered carnivores are often forced to choose between restoring habitat to allow larger breeding populations or reduce risks of mortality to increase survival rates. We modelled the viability of a metapopulation of the Iberian lynx (Lynx pardinus) under several scenarios (habitat restoration, anti-poaching, reduction in road kills) in a real landscape to evaluate their relative effects. Increasing carrying capacity was highly effective when performed on the local populations that acted as sources but had no effect when carried out in the sinks. Realistic scenarios consisting of partial removal of the human-related mortality (assuming additive effects of causes) predicted high risk of extinction. When combined, the effects of both management options are highly dependent on where they are carried out. If the sinks are the only targets of carrying capacity enlargement, a complete removal of human-caused mortality is required, whereas increases in the carrying capacity of sources are always effective. The metapopulation risk of extinction decreases dramatically (from 45.5% to 2.1% in 100) if connectivity among source populations can be improved. According to our work, only a detailed knowledge of the spatial and demographic structure of the populations, combined with simulations of realistic situations, can help managers to select the a priori optimal strategy, which probably combines different management options.

An outbreak of rabies in a pack of endangered wild dogs (Lycaon pictus) in the Madikwe Game Reserve, South Africa, is described. The outbreak, which occurred in early 2000, was the second outbreak of rabies in wild dogs followed their re-introduction to the Reserve in 1995. The outbreak resulted in the death or disappearance of 10 out of 12 eight-month-old pups. Gene sequence analysis indicates that the virus is distinct from that of the previous, 1997, outbreak, although it falls within the same group as other canid isolates from northern South Africa. Since jackals (Canis mesomelas) are the principal host species of rabies in the area and domestic dogs do not occur in the Reserve, it is assumed that the jackal cycle was the source of the infection. Whereas the pups had not been vaccinated, the five adults, all of which survived, had each been vaccinated by the parenteral route at least twice and each had significant rabies serum neutralising antibodies. This indicates that multiple vaccination against rabies is effective in protecting wild dogs against challenge. Our findings extend previous observations that indicated failure of protection against rabies with single parenteral vaccinations in African wild dogs.

We document the recolonization of an indigenous large herbivore into its historic range. Eighteen wood bison (Bison bison athabascae) were reintroduced into the Mackenzie Bison Sanctuary of the Northwest Territories, Canada, in 1963. The population subsequently increased in number and range, peaking at about 2400 in 1989; numbers were estimated at about 1900 in 1998. Recolonization occurred through a series of increases in local areas followed by pulses of dispersal and range expansion. This pattern was originally described for exotic species' introductions. Differences in diet and overwinter survival of calves over the bison's range suggest that intraspecific competition for food provided the stimulus for range expansion. For a conservation strategy, the reintroduction of animals into several independent sites in their historic range would facilitate recolonization and achieve a faster spread than a reintroduction into one site followed by waiting for the population to spread as a result of its own density dependent responses.

This study examined the effects of alteration of tropical rainforest vegetation structure and composition on bird community structure and the influence of life-history traits on species persistence in the Kalakad-Mundanthurai Tiger Reserve, Western Ghats. Systematic sampling for vegetation and point count surveys for birds were carried out in cardamom plantations abandoned for 5 and 15 years, plantation-rainforest edges, a selectively logged forest patch and adjoining undisturbed rainforest sites. Principal components analysis of vegetation variables revealed clear differences between undisturbed and altered sites in woody plant and cane densities, canopy cover and vertical stratification. Bird species richness was lowest in cardamom plantations abandoned for 5 years and highest in logged and undisturbed forest. Bird species richness and similarity with undisturbed forest were significantly positively related to the vegetation component representing woody plant and cane (Calamus spp.) densities. Sites that were more similar in tree species composition had more similar bird communities whereas similarity in foliage profile between sites did not influence bird community similarity. Birds that were rare, were large-bodied and belonged to the carnivore, omnivore, bark-surface feeder and terrestrial insectivore guilds were adversely affected by habitat alteration. Restoring woody plant and cane densities and rainforest floristic composition in disturbed habitats may be required for management and conservation of bird communities typical to the region.

Relocation of endangered species can be an effective conservation tool if it does not mix populations that represent significant intraspecific variation. The threatened Italian agile frog, Rana latastei, has small populations with low genetic diversity: translocation has been proposed to improve the likelihood of survival of populations. Using a common environment experiment and field surveys, we investigated whether there were differences in larval growth and developmental rate between foothill and lowland R. latastei populations, to evaluate if they are evolutionarily significant units. In nature, the colder climate of the foothills causes delayed metamorphosis. Conversely, in a common environment, larvae from foothill populations show faster growth and development. We did not find a significant egg-size related maternal effect or any differences in size at metamorphosis. We hypothesise that counter-gradient selection promoted fast growing phenotypes in a cold environment, where low temperatures slow down larval development. Foothill populations, despite being only a small geographical distance away from lowland populations, seem to be adapted to a colder climate and represent an evolutionarily significant unit. Different populations should, therefore, be managed independently, avoiding translocation. We suggest that evolutionary divergence between populations should be verified prior to planning relocation programmes, to prevent the risk of genetic homogenisation.

During the twentieth century otter (Lutra lutra) populations in central and western Europe declined and became fragmented because of habitat alterations, chemical pollution and direct persecution. In this study we used microsatellites to describe spatial patterns of genetic diversity and subdivision in otters from eight populations in Europe. Genetic diversity was moderately high within populations (He = 0.45-0.77), and significantly partitioned among locations (FST = 0.17; RST = 0.16; P < 0.001). A Bayesian cluster analysis of multilocus genotypes assigned individuals to seven genetically distinct groups, which were partly concordant with the geographical origin of the samples. An assignment test of the individuals to the populations showed that the Danish and, to a lesser extent, the Spanish populations were unique and distinct, whereas the other populations were partially admixed. Inference of past demographic fluctuations from coalescent analysis suggested that otter populations probably declined several thousand years ago, with the exception of the Irish population for which no such decline could be detected. No genetic evidence for recent bottlenecks was found. The historical decline could stem from post-glacial founding events and recolonization of northern Europe after the last glacial maximum, or from more recent dry stages in the early Holocene in central Europe. On the basis of these results, we recommend that recovery plans should promote the expansion of existing natural populations through improvements of river and wetland habitats.

Introduced predators are a leading threat to seabird populations world-wide and cats (Felis catus) have probably had the most universally damaging effect (Moors & Atkinson, 1984). Eradication of feral cat populations from seabird colonies is a conservation priority (Tershy et al., 2002) and there are many studies that demonstrate the benefits of these actions for seabirds (e.g. Forsell, 1982; Cooper et al., 1995). However, detailed estimates of the effects of cat predation on seabird population viability are lacking in spite of the fact that such data could provide important support for land managers attempting to promote eradication programmes for seabird restoration.

We present a case study of the use of simulation modelling to develop and test strategies for managing populations under uncertainty. Strategies that meet a stock conservation criterion under a base case scenario are subjected to a set of robustness trials, including biased and highly variable abundance estimates and poaching. Strategy performance is assessed with respect to a conservation criterion, the revenues achieved and their variability. Strategies that harvest heavily, even when the population is apparently very large, perform badly in the robustness trials. Setting a threshold below which harvesting does not take place, and above which all individuals are harvested, does not provide effective protection against over-harvesting. Strategies that rely on population growth rates rather than estimates of population size are more robust to biased estimates. The strategies that are most robust to uncertainty are simple, involving harvesting a relatively small proportion of the population each year. The simulation modelling approach to exploring harvesting strategies is suggested as a useful tool for the assessment of the performance of competing strategies under uncertainty.

The Pyrenean population of Spanish wild goat (Capra pyrenaica pyrenaica) is nearly extinct. To find the most appropriate source of individuals for supplementing the Pyrenean population, we identified Evolutionary Significant Units (ESUs) among populations of the Capra pyrenaica species. We have examined sequence variability of portions of the mitochondrial DNA (mtDNA) control region and cytochrome b (cyt b) gene. Samples were from seven populations of Spanish wild goat distributed over the species' geographic range. The level of divergence between the Pyrenean and other Spanish haplotypes is almost as high as the divergence between the Alpine (C. ibex ibex) and the Spanish wild goats. In addition, the Pyrenean goat is morphologically distinct from other Spanish wild goats. Therefore the Pyrenean population should be considered as an ESU. For the reinforcement, we suggest either using individuals from the most polymorphic Spanish population, or mixing individuals from diverse Spanish origins, since all the other Spanish populations are equally genetically distant from the Pyrenean population.

Individual identification was assessed for 30 brown bears (Ursus arctos) from the endangered Apennine population in central Italy, mostly through non-invasive sampling. Shed hairs were used as a source of DNA to obtain single genotype profiling at 12 microsatellite loci. Average expected heterozygosity (0.438) and mean number of alleles per locus (2.2) showed that the level of genetic diversity was one of the lowest recorded for brown bear populations across their European and North American ranges. A large portion of the original genetic variation may have been lost through random genetic drift during the recent period of isolation. The level of genetic variability proved sufficient to identify the bears individually by using nine microsatellite loci. The probability that two animals shared by chance the same multilocus genotype was estimated to be 1 in 100. Despite the recent history of small population size, the average relatedness indicated that the majority of individuals are not first-order relatives. Simulated paternity tests showed that a 12-loci genotype may be necessary to assign paternity with an 80–95% confidence level. Lack of genetic diversity, as well as the maintenance of an adequately effective number, may seriously jeopardise the long-term survival of this population. Furthermore, high mortality rates, poaching and encroachment by human activities represent immediate, urgent concerns. A population increase through enforced protection and habitat restoration is recommended. Presently, restocking with bears from other source populations should be discouraged to avoid genetic extinction. Non-invasive sampling provided reliable population and individual genetic data. Microsatellite genotyping proved a valuable genetic tagging method and a feasible alternative to conventional field counts for the brown bear in central Italy.

Recent studies have suggested that contemporary losses of species are unevenly distributed over phylogenies. Here, we show that species of Australian marsupials are at higher risk if they belong to genera that are species-poor, old or phylogenetically distinct. Extinction risk in this group is also related to habitat and body size, being higher for species from non-forested habitats and of intermediate body size. We tested the extent to which the phylogenetic selectivity of extinction was explained by this ecological pattern. We found that while genus size and distinctiveness explain no variation in extinction risk that is not accounted for by habitat and body size, there is a significant residual association of genus age with extinction risk. This suggests that while species in small and distinctive genera are at high risk because they are overrepresented in non-forested habitats and intermediate body size classes, species in old genera are at higher risk over the range of body sizes and habitats and may be intrinsically vulnerable to extinction.

We studied variables affecting spatial and inter-annual variations in presence/absence and abundance of little bustard (Tetrax tetrax) males in extensive cereal farmland in central Spain. Both variables were positively influenced by the percentage of arable surface and also by substrate heterogeneity. In addition, male abundance varied significantly between years and among sectors of the study area, even when taking into account variation in landscape variables. A significant interaction between year and sector showed that abundance variations did not follow a fixed pattern, meaning that interannual variation in male abundance among sectors cannot be predicted. We discuss potential alternatives to explain such variation, as well as the implications of our results for the species' population estimates, status assessment and conservation.