Risk of Local Extinction of Odonata Freshwater

Contributed Paper

Risk of Local Extinction of Odonata Freshwater Habitat Generalists and Specialists ¨ KI,∗ ‡ JUKKA SALMELA,§ †† AND MARKKU KUITUNEN∗ JUKKA SUHONEN,∗ † ∗∗ ESA KORKEAMA ∗

Department of Biological and Environmental Science, University of Jyväskylä, P.O. Box 35, FIN-40014, Jyväskylä, Finland †Section of Ecology, Department of Biology, University of Turku, FI-20014, Turku, Finland ‡Water and Environment Association of the River Kymi, Tapiontie 2 C, FI- 45160, Kouvola, Finland §Zoological Museum, Department of Biology, University of Turku, FI-20014, Turku, Finland

Abstract: Understanding the risk of a local extinction in a single population relative to the habitat requirements of a species is important in both theoretical and applied ecology. Local extinction risk depends on several factors, such as habitat requirements, range size of species, and habitat quality. We studied the local extinctions among 31 dragonfly and damselfly species from 1930 to 1975 and from 1995 to 2003 in Central Finland. We tested whether habitat specialists had a higher local extinction rate than generalist species. Approximately 30% of the local dragonfly and damselfly populations were extirpated during the 2 study periods. The size of the geographical range of the species was negatively related to extinction rate of the local populations. In contrast to our prediction, the specialist species had lower local extinction rates than the generalist species, probably because generalist species occurred in both low- and high-quality habitat. Our results are consistent with source–sink theory.

Keywords: damselfly, dragonfly, extinction, freshwater, generalist species, geographical range size, specialist species, source–sink Riesgo de Extinci´ on Local de Odonatos de Agua Dulce Generalistas y Especialistas de Hábitat

Resumen: Entender el riesgo de una extinción local en una población u on a los requerim´ nica con relaci´ ientos del h´ abitat de una especie es importante tanto en la ecolog´ıa te´ orica como en la aplicada. El riesgo de extinci´ on local depende de varios factores, como los requerimientos del h´ abitat, el tama˜ no del rango de las especies y la calidad del h´ abitat. Estudiamos las extinciones locales de 31 especies de caballitos del diablo y libélulas entre 1930 y 1975 y entre 1993 a 2000 en el centro de Finlandia. Probamos si las especies especialistas de h´ abitat ten´ıan una tasa de extinci´ on m´ as alta que las generalistas. Aproximadamente el 30% de las poblaciones locales de caballitos del diablo y libélulas fueron extirpadas durante ambos periodos de estudio. El tama˜ no del rango geogr´ afico de las especies estuvo relacionado negativamente con la tasa de extinci´ on de las poblaciones locales. A diferencia de nuestra predicci´ on, las especies especialistas tuvieron una tasa de extinci´ on m´ as baja que las especies generalistas, probablemente porque las especies generalistas est´ an presentes tanto en h´ abitats de baja calidad como en aquellos de alta calidad. Nuestros resultados son consistentes con la teor´ıa de fuente-sumidero. Palabras Clave: agua dulce, caballitos del diablo, dinámicas source-sink, especies especialistas, especies generalistas, extinci´ on, libélulas, tama˜ no del rango geográfico

∗∗ Address

for correspondence: Section of Ecology, Department of Biology, University of Turku, FI-20014, Turku, Finland, email [email protected] ††Current address: Mets¨ ahallitus (Natural Heritage Services), P.O. Box 8016, FI-96101, Rovaniemi, Finland. Paper submitted November 2, 2012; revised manuscript accepted September 3, 2013.

1 Conservation Biology, Volume 00, No. 0, 1–7 C 2014 Society for Conservation Biology DOI: 10.1111/cobi.12231

Local Extinction of Habitat Generalists

2

Introduction Several studies haveestablished that extinction is likely to be taxonomically nonrandom. In particular, greater numbers and higher proportions of recent species extinctions have occurred in freshwater habitats (Richter et al. 1997; Ricciardi & Rasmussen 1999; Abell 2002). Understanding the patterns of species extinction and threat is a major goal of conservation biology. There are 2 nonmutually exclusive types of variables that may account for variation in risk among species and populations: intrinsic biological factors and extrinsic abiotic or geographical factors. For example, most threatened species have small geographical ranges and narrow habitat requirements and are highly sensitive to environmental changes (Owens & Bennett 2000; Purvis et al. 2000a, 2000b). Rapid environmental change probably leads to a high local extinction risk, which may ultimately lead to the loss of a species in certain areas. In addition, meta-population theory predicts that population size, genetic variation, and stochastic events are also important factors in local extinction of a population (Hanski 1999). Conservation biologists have focused much attention on the significance of several ecological mechanisms, such as geographical range (e.g., Hanski 1982; Korkeamäki & Suhonen 2002), local population size, habitat quality (Pulliam 1988; Watkinson & Sutherland 1995; Suhonen et al. 2010), and specialization of species (e.g., Purvis et al. 2000a, 2000b; Harcourt et al. 2002). The logic behind specialization is clear. When the environment changes, species with broad habitat requirements are more likely to persist than those with narrow habitat requirements (e.g., Purvis et al. 2000a, 2000b). Alternatively, if a generalist species is breeding in both low- (sink) and high-quality (source) habitat, then the local extinction risk between specialist and generalist species may be equal or even higher in generalist species because local extinction risk is higher in the low-quality sink habitats than in the high-quality source habitats and thus mortality exceeds birth rate in true sink populations (Pulliam 1988; Watkinson & Sutherland 1995; Suhonen et al. 2010). The source–sink theory assumes a constant net flow of migrants from source populations to sink populations. Immigration from a nearby source population may also reduce the extinction risk in a true sink population (Pulliam 1988; Watkinson & Sutherland 1995). If migration ceases, another alternative for the local extinction of a sink population is a decline to a stable state, where births and deaths are in balance. This is called a pseudosink (Watkinson & Sutherland 1995; Thomas et al. 1996). More generally, both true sink and pseudosink populations can be defined as constant net receivers of organisms, but true sink populations are outside the fundamental niche (Pulliam 1988) and thus become extinct without constant immigration from a source population.

Conservation Biology Volume 00, No. 0, 2014

We examined local extinction risk relative to the geographical range size of the species and the degree of breeding habitat specialization. We estimated the local extinction risk in breeding habitat generalists and breeding habitat specialists of odonate species in 2 freshwater habitats, running and standing waters, in Central Finland. Each of the dragonfly and damselfly species formed a meta-population, where the local populations were not truly isolated; rather, they were linked via dispersal (Conrad et al. 1999; Hanski 1999). Dragonflies and damselflies are good fliers and can disperse over long distances (e.g., Conrad et al. 1999; but see exception Watts et al. 2007).Furthermore, in most dragonfly and damselfly species the immatures fly away from their emergence site and not to return to the breeding habitat before reaching their reproductive maturity, but some damselfly species fly only a few meters away from the water and sit at the edge of the habitat for a few days waiting to mature (Corbet 1999). Most odonate species can be categorized into 2 breeding groups, generalist and specialist species. Generalist species can breed in both running and standing waters, but specialist species can only breed in either running or standing water. That is, generalist species are breeding in both high-quality source habitats and low-quality sink habitats. In contrast to the generalist species, specialist species breed only in high-quality habitats, either in standing water or running water. We wanted to determine whether the extinction risk was higher in specialist species than in generalist species; thus, we compared the local extinction risk of populations of the 31 most abundant dragonfly and damselfly species (Odonata) living in streams, ponds, and lakes in Finland. This study is based on data collected from 1930 to 1975 (details in Korkeamäki & Suhonen 2002; Suhonen et al. 2010). We conducted field research to reexamine the occurrence of dragonfly and damselfly populations in previously studied sites. First, we investigated whether the local extinction risk of a dragonfly species was associated with the species’ geographical range size. Second, we determined whether the habitat generalist species that occurred commonly in both highand low-quality breeding habitats faced a higher risk of extinction than the specialist species that occurred in only the high-quality source habitats.

Methods Study Sites Local surveys of populations of Finnish Odonata were conducted from 1930 to 1975 (Korkeamäki & Suhonen 2002; Suhonen et al. 2010). We reexamined these populations from 1995 through 2003. The streams, ponds, and lakes examined were all in Central Finland, north of the city of Jyväskylä (26°E, 63°N), at elevations between 75 and 150 m asl. We surveyed 34 small creeks and brooks

Suhonen et al.

3

Table 1. Number of local populations and their extinction rates (%) between 1930–1975 and 1995–2003 of the 31 most abundant odonate species in running and standing waters in Central Finland. Running water Species Aeshna caerulea Aeshna grandis Aeshna juncea Aeshna subartica elisabethae Calopteryx splendens Calopteryx virgo Coenagrion armatum Coenagrion hastulatum Coenagrion johanssoni Coenagrion lunulatum Coenagrion pulchellum Cordulegaster boltoni Cordulia aenea Enallagma cyathigerum Erythromma najas Gomphus vulgatissimus Lestes sponsa Leucorrhinia dubia Leucorrhinia albifrons Leucorrhinia caudalis Leucorrhinia rubicunda Libellula quadrimaculata Onychogompus forcipatus Ophiogomphus cecilia Platycnemis pennipes Pyrrhosoma nymphula Somatochlora arctica Somatochlora flavomaculata Somatochlora metallica Sympetrum danae Sympetrum flaveolum ∗ Main

Standing water

Habitat∗

n

%

n

%

S S S S R R S S S S S R S S S R S S S S S S R R R R S S R S S

7 10 11

100 0 64

4 16 18 9

100 0 11 33

8 26

50 23

11 5

18 100

7 25 18 4 6

0 4 22 100 17

19 8 8 8 3 8 6

37 75 38 38 67 63 50

21 13 15

5 31 7

7 7 15 7 4 21

86 57 67 43 25 71

15 17 10 10 19 22

13 41 40 20 0 5

20 7

0 29

3 8 7 12 14 11

67 75 29 8 0 18

breeding habitat of odonates in Finland based on Valle’s (1952) classification. Abbreviations: S, standing water; R, running water.

and 23 ponds and lakes, all located within 150 km of each other (see fig. 1 in Suhonen et al. 2010). Most of the previously recorded data were collected from 1930 until 1950 and consisted of 300 records from lakes and ponds (standing waters) and 230 records from streams, creeks, and rivers (running waters) (530 old records all together from 31 species; Table 1). If 2 old records for the same species were observed within the same catchment and if there was a connecting stream between a lake and a pond, we considered these separate standing-water populations. For running-water populations, 2 records were assigned as being from different populations if each observation was made in a different lake or stream. We examined on average 17 (SD 10; range 3–36) populations for each species (Table 1). Our data included 12 species of damselflies (Zygoptera) and 19 species of dragonflies (Anisoptera) (Askew 1988). Current Persistence of Local Populations In previous studies of dragonfly populations (see references in Korkeamäki & Suhonen 2002; Suhonen et al.

2010), the presence of a species was determined by confirming the presence of adults. We used the same method to determine the presence of current populations. We determined the persistence of each odonate population by surveying all streams during the summers of 1995 and 1996 and the lakes and ponds during the summers of 2002 and 2003. We recorded and identified each adult dragonfly and damselfly that was present at each of the selected water bodies. To locate the adults, we visited each site during sunny weather at least 4 times during the flight period (from early June to early September). We identified different species in different parts of the water bodies. In areas of the water bodies where populations were difficult to observe, the search was continuous. Those species that could not be identified while in flight were caught with a butterfly net and identified. We included all odonate species observed and recorded within 50 m of each stream, pond, or lake. We assumed we would find all species as adults. In each study area, however, all exuviae encountered on the shoreline were also collected. We assumed a species had been extirpated if we did not find any larvae, exuviae, or adults around



4

water bodies, where the species had previously been recorded, in either of the study years. At each stream, we sampled larvae with hand nets (Korkeamäki & Suhonen 2002; Suhonen et al. 2010 for details). In our data set, we determined the local extinction of a species on the basis of its actual presence or absence, which may bias the true extinction frequency because an odonate species classified as locally extinct may later reoccur at a site due to subsequent immigration from another site where it is present. Breeding Habitats of Adults Odanate females may lay eggs on aquatic macrophytes and mosses in lakes, ponds, pools, creeks, rivers, or streams (Corbet 1999). The data on the breeding habitats of odonates are based on Valle’s (1952) work from the same area in Central Finland. In Central Finland, the 15 species we examined occur frequently in both standing and running waters (Valle 1952; Korkeamäki & Suhonen 2002; Suhonen et al. 2010). We categorized each of these 31 species as preferring 1 of the 3 main breeding habitat types (Table 1): running or still water or both. Seven species occurred only in running water, 9 species only on still water, and 15 species occurred frequently in both habitats. This method was based on Valle’s (1952) classification. Thus, we designated the main breeding habitat of each species as being of high quality and the less preferred breeding habitat as being of low quality (Valle 1952). Range Size of Study Species in Finland Although the studied water bodies were situated in Central Finland, we found it important to get an estimate of the geographical range size of each species in Finland (Valtonen 1980). The geographical range size for each of the 31 studied species in Finland was measured according to the previously published distribution maps. The distribution maps provided by Valtonen (1980) were considered the most accurate because they are based upon an extensive atlas on dragonfly distribution in Finland. The geographical range size of the species is presented as the frequency, that is, the number of occupied standardized coordinate system squares (10 × 10 km) reported by Valtonen (1980). Each occupied square was considered a separate unit, and occupied squares for each of the species were counted. Statistical Analyses The extinction probability for each of the species was calculated by dividing the number of vanished populations by the number of old populations in water bodies. Unfortunately, we could not analyze the recolonization rates from the old data set because we were


not sure whether all the species were recorded in the old data set. We tested the effect of species distribution on the extinction probability with generalized linear models with type III errors. The link function was logit when we tested the relationship between the species’ geographical range size in Finland and probability of extinction (probability distribution binomial). In the model, we used events per trial option, where the number of extinct populations varied by event and the number of old populations varied by trial. In this model, each species’ geographical range size in Finland was used as a covariate and breeding habitat type was used as factor. The interaction between these 2 variables was not statistically significant (all habitats; Wald = 2.49, df = 2, p = 0.288 and only main breeding habitat; Wald = 4.45, df = 2, p = 0.108); therefore, we removed all interactions from the final models for simplification. All the data analyses were performed in the IBM SPSS statistical package, version 20.

Results On average 35% (SD 26; range 0–100%) of odonate populations had become extinct (Fig. 1). None of the previously known populations of the brown hawker (Aeshna grandis) and dark bluet (Coenagrion armatum) became extinct. In contrast, all previously known population of azure hawker (Aeshna caerulea) and crescent bluet (Coenagrion lunulatum) were extirpated. Both of the studied variables, geographical range size of the species and habitat quality, affected local extinction risk. The geographical range size of the species in Finland was negatively related to the extinction rate of local populations (Figs. 1, 2a, and b). There were also consistent differences between the extinction risk of the breeding habitat generalist species (breeding in both running and standing water) and the breeding habitat specialist species (breeding only in running or standing water) (generalized linear models, Wald = 12.41, df = 2, p = 0.002; Figs. 1 & 2). Probability of local extinction risk was lower in the specialist species than it was in the generalist species, for which the effects of the distribution area on local extinction were controlled (covariate) (generalized linear models, Wald = 56.28, df = 1, p < 0.001; Figs. 1 & 2a). There were no differences in local extinction risk between standing water and running water species (Wald = 2.50, df = 1, p = 0.114), but both species groups had lower extinction risk than generalist species, which were breeding in both habitat types (Wald = 12.18, df = 1, p < 0.001 and Wald = 4.15, df = 1, p = 0.042, respectively). However, when only main breeding habitat was selected from each generalist species, there were no differences between types of species in local extinction risk (Wald = 5.71, df = 2, p = 0.057; Fig. 2b).

Suhonen et al.

5

Figure 1. Relationship between local extinction rate (as measured in 34 streams and 23 ponds and lakes in Central Finland) of 31 of the most frequently recorded odonate species and geographical range size of each species in Finland (open circles, standing water species; filled circles, running water species; filled triangles, species that occur in both running and standing water).

Discussion Extinction risk of local populations was high, especially for generalist species with a narrow geographical range (Figs. 1 & 2). Species with a narrow and fragmented geographical range (Hanski 1982; Korkeamäki & Suhonen 2002; Segura et al. 2007) and occurring in low-quality habitats (e.g., Thomas et al. 1996; Johnson 2004; Suhonen et al. 2010) have been reported to be sensitive to environmental changes.Several previous studies show that species with limited ecological and geographical range size are vulnerable to extinction (e.g., Johnson 1998; Purvis et al. 2000a; Collen et al. 2011). Locally, common species become widely distributed because of their low-extinction and high-colonization rates (e.g., Hanski & Gyllenberg 1997). Persistence of extinction-prone populations may be enhanced through migration from the other local populations. There are at least 2 not mutually exclusive explanations for the hypotheses that widely distributed species have a low risk of extinction. First, range size and local abundance are often correlated (e.g., Brown 1984). If extinction risk declines with increasing local population size, the range–size effect could be an effect of local abundance. The second mechanism is based on the idea that with a large range size, there are more occupied and suitable sites from which more colonists can migrate and thus prevent extinction. Unfortunately, we could not analyze recolonization rates from the old data set because we were not sure weather all species were recorded. Therefore, we had no data on species recolonizations. Thus, our estimated extinction rates are likely to be underestimates because there were long gaps between repeated surveys.

Many dragonfly and damselfly species with narrow geographical range size seemed sensitive to habitat quality, too. This may be due to a change in the water quality of streams between the 2 study periods, agriculture, forestry, and construction have caused some extensive disturbance and degradation in areas otherwise suitable for the dragonflies (e.g., Korkeamäki & Suhonen 2002). Water chemistry does not seem to affect the odonate larvae directly; rather, changes in water chemistry change species composition and available food items (McPeek 1990; Buchwald 1992). That species with narrow geographical range sizes have a high-extinction risk may stem from their lower population sizes and greater environmental sensitivity relative to widespread species (Angermeier 1994). Small populations are more prone to extinction than larger populations (e.g., Schoener & Spiller 1992; Hanski et al. 1995; Eisto et al. 2000). An elevated risk of the local extinction is connected to the fact that demographic, genetic, and environmental stochasticities affect small populations more severely than large populations (Lande 1988; Sacceri et al. 1998). When we controlled species’ geographical range size, extinction risk was lower in the specialist species than in the generalist species. This phenomenon occurred in the generalist species only when both the high- and the low-quality habitats were pooled. This result is the opposite of earlier observations and is not consistent with ecological theory (e.g., Hughes et al. 2000; Harcourt et al. 2002). However, this result may be explained by source– sink theory (Pulliam 1988; Watkinson & Sutherland 1995; Thomas et al. 1996). There are 2 not mutually exclusive and general mechanisms for this. There is a standing crop of individuals in the sink population, sustained by


6

Figure 2. Relationship between the predicted local extinction probability of 31 odonate species in (a) all habitat types and (b) only the main breeding habitats and the geographical range size of each species in Finland (open circles, standing water species; filled circles, running water species; filled triangles, species that occur in both running and standing waters).

immigration from the source population. That standing crop can be pushed toward zero (and then count as an extinction) if either the sink population itself becomes unfit or fewer immigrants arrive from the source population, maybe because the quality of the source populations’ habitat has declined or the population has been extirpated. Our results support the latter case, especially in relation to the peatland-associated species. Ex-



tremely high extinction rates of the peatland-associated species also lend support to the idea that sink populations are really vulnerable. In Central Finland, the proportion of unditched peatland area was reduced from 80% to 20% within 50 years due to forestry (Karjalainen 1991). Probably most of the previously known populations of peatland-associated species (Coenagrion johanssoni, Pyrrhosoma nymphula, and A. caerulea) disappeared because their source habitat was severely degraded between the survey periods (Table 1). However, without extensive demographic study of each species, the relationship between habitat quality and local extinction cannot be proven. However, there is also an alternatively potential explanation for our results: local extinction rates were higher in the low-quality habitats. For instance, dragonflies and damselflies may have ideal-free distributions (Fretwell & Lucas 1970);thus, no habitat is a demographic source or sink. It could still be the case that rarely used habitats have a low carrying capacity. That all by itself makes extinction due to either demographic or environmental stochasticity more likely. This might also account for our observed pattern of extinction risk in species that bred in both standing and running water habitats. Our results highlight the necessity for conservation biologists to study directly the quality of freshwater habitats because this is likely to be an important factor in the likelihood of extinction for small populations of aquatic insects. Some researchers assume species’ geographical range sizes are explained solely by habitat characteristics (i.e., species is assumed to be present if there is habitat and absent if there is no habitat). However, our results show that habitat use, at least in the Finnish odonates, is unlikely to be easy to determine unambiguously. In devising an effective conservation plan for a threatened or endangered population scattered over a heterogeneous area, it is never easy to determine which one of the patches is the true sink population for that species (e.g., Watkinson & Sutherland 1995; Thomas et al. 1996; Johnson 2004). Although it is not easy to assess a true sink population (Watkinson & Sutherland 1995), it is important to identify source population for conservation purposes. Protection of only the sink population will probably produce extinction of a species if the area or the habitat quality of the source population is reduced (e.g., Pulliam 1988; Watkinson & Sutherland 1995; Suhonen et al. 2010). Only a few previous studies have included both the habitat quality and the meta-population factors within their attempts to understand the local extinction rates (Armstrong 2005; Franzén & Nilsson 2010). Future empirical and theoretical studies thus need to incorporate more detailed information regarding habitat requirements, habitat quality and size, and geographical range size in order to elucidate the risk of local population extinction.

Suhonen et al.

Acknowledgments We thank J. P. Rodriguez and two anonymous referees for their valuable comments that greatly improved the manuscript. We are also grateful to M. Kuitunen for searching literature and to K. Juvonen for reading our manuscript before submitting. This study was supported by the Finnish Biodiversity Research Programme (FIBRE), Academy of Finland (J.S., M.K.), Jenny and Antti Wihuri Foundation (M.K.), EU IMPERIA-project (LIFE11 ENV/FI/905; MK), University of Jyväskylä (M.K.), Ministry of Environment (J.S.), Maj and Tor Nessling Foundation (J.S.), and Academy of Finland: Finnish Centre of Excellence Programme (2000–2005) (J.S.). Literature Cited Abell, R. 2002. Conservation biology for the biodiversity crisis: a freshwater follow-up. Conservation Biology 16:1435–1437. Angermeier, P. L. 1994. Ecological attributes of extinction-prone species; loss of freshwater fishes of Virginia. Conservation Biology 9:143–158. Armstrong, D. P. 2005. Integrating the metapopulation and habitat paradigms for understanding broad-scale declines of species. Conservation Biology 19:1402–1410. Askew, R. R. 1988. The dragonflies of Europe. Harley Books, London. Brown, J. H. 1984. On the relationship between abundance and distribution of species. American Naturalist 124:255–279. Buchwald, R. 1992. Vegetation and dragonfly fauna—characteristics and examples of biocenological field studies. Vegetatio 101:99– 107. Collen, B., L. McRae, S. Deinet, A. De Palma, T. Carranza, N. Cooper, J. Loh, and J. E. M. Baillie. 2011. Predicting how populations decline to extinction. Philosophical Transactions of Royal Society B 366:2577– 2586. Conrad, K. F., K. H. Willson, I. F. Harvey, C. J. Thomas, and T. N. Sherratt. 1999. Dispersal characteristics of seven odonate species in an agricultural landscape. Ecography 22:524–531. Corbet, P. S. 1999. Dragonflies: behaviour and ecology of odonata. Harley Books, Essex. Eisto, A.-K., M. T. Kuitunen, A. Lammi, V. Saari, J. Suhonen, S. Syrjäsuo, and P. Tikka. 2000. Persistence and offspring fitness of the rare bellflower Campanula cervicaria in relation to population size. Conservation Biology 14:1413–1421. Franzén, M., and S. G. Nilsson. 2010. Both population size and patch quality affect local extinctions and colonizations. Proceedings of the Royal Society London B 277:79–85. Fretwell, S. D., and H. L. Lucas. 1970. On territorial behavior and other factors influencing habitat distribution in birds. Acta Biotheoretica 19:16–36. Hanski, I. 1982. Dynamics of regional distribution: the core and satellite species hypothesis. Oikos 38:210–221. Hanski, I. 1999. Metapopulation ecology. Oxford University Press, Oxford. Hanski, I., and M. Gyllenberg. 1997. Uniting two general patters in the distribution of species. Science 275:397–399. Hanski, I., J. P¨ oyry, T. Pakkala, and M. Kuussaari. 1995. Multible equilibrium in metapopulation dynamics. Nature 377:618– 621.

7

Harcourt, A. H., S. A. Coppeto, and S. A. Parks. 2002. Rarity, specialization and extinction in primates. Journal of Biogeography 29:445– 456. Hughes, J. N., G. C. Daily, and P. R. Ehrlich. 2000. Conservation of insect diversity: a habitat approach. Conservation Biology 14:1788–1797. Johnson, C. N. 1998. Species extinction and the relationship between distribution and abundance. Nature 394:272–274. Johnson, D. M. 2004. Source-sink dynamics in a temporally, heterogeneous environment. Ecology 85:2037–2045. Karjalainen, H. 1991. Uhanalainen Luonto ja Metsienhoito. Forssa (In Finnish). Korkeamäki, E., and J. Suhonen. 2002. Distribution and habitat specialization of species affect local extinction in dragonfly Odonata populations. Ecography 25:459–465. Lande, R. 1988. Genetics and demography in biological conservation. Science 241:1455–1460. McPeek, M. A. 1990. Determination of species composition in the Enallagma damselfly assemblages of permanent lakes. Ecology 71:83– 98. Owens, I. P. F., and P. M. Bennett. 2000. Ecological basis of extinction risk in birds: habitat loss versus human persecution and introduced predators. Proceedings of National Academy of Sciences 97:12144– 12148. Pulliam, H. R. 1988. Sources, sinks and population regulation. American Naturalist 132:652–661. Purvis, A., J. L. Gittleman, G. Cowlishaw, and G. M. Mace. 2000a. Predicting extinction risk in declining species. Proceedings of the Royal Society London B 267:1947–1952. Purvis, A., K. E. Jones, and G. M. Mace. 2000b. Extinction. BioEssays 22:1123–1133. Ricciardi, A., and J. B. Rasmussen. 1999. Extinction rates of North American freshwater fauna. Conservation Biology 13:1220–1222. Richter, B. C., D. P. Braun, M. A. Mendelson, and L. L. Master. 1997. Threats to imperilled freshwater fauna. Conservation Biology 11:1081–1093. Sacceri, I, M. Kuussaari, M. Kankare, P. Vikman, W. Fortelius, and I. Hanski. 1998. Inbreeding and extinction in a butterfly metapopulation. Nature 392:491–494. Schoener, T. W., and D. A. Spiller. 1992. Is extinction rate related to temporal variability in population size? An empirical answer for orb spiders. American Naturalist 139:1176–1207. Segura, C., M. Feriche, J. M. Pleguezuelos, and X. Santos. 2007. Specialist and generalist species in habitat use: implications for conservation assessment in snakes. Journal of Natural History 41:2765–2774. Suhonen, J., M. Hilli-Lukkarinen, E. Korkeamäki, M. Kuitunen, J. Kullas, J. Penttinen, and J. Salmela. 2010. Local extinction of dragonfly and damselfly populations in low- and high-quality habitat patches. Conservation Biology 24:1148–1153. Thomas, C. D., M. C. Singer, and D. A. Boughton. 1996. Catastrophic extinction of population sources in a butterfly metapopulation. The American Naturalist 148:957–975. Valle, K. J. 1952. Die Verbteitungsverhältnisse der ostfennoskandischen Odonaten. (Zur Kenntnis der Odonatenfauna Finnlands 6.). Acta Entomologica Fennica 10:1–87. Valtonen, P. 1980. Die Verbteirung der Finnischen Libellen (Odonata). Notulae Entomology 60:199–215. Watkinson, A. R., and W. J. Sutherland. 1995. Sources, sinks and pseudosinks. Journal of Animal Ecology 64:126–130. Watts, P. C., F. Rousset, I. J. Saccheri, R. Leblois, S. J. Kemp, and D. J. Thompson. 2007. Compatible genetic and ecological estimates of dispersal rates in insect (Coenagrion mercuriale: Odonata: Zygoptera) populations: analysis of ‘neighbourhood size’ using a more precise estimator. Molecular Ecology 16:737–751.
