Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Latitudinal, Longitudinal, Radial and Altitudinal Gradients in Endemism

The overall taxonomic diversity of vascular plants is low towards high latitudes and high towards the tropics. This latitudinal gradient is, in general, also true for endemic vascular plant taxa. A large number of processes which might be responsible for latitudinal gradients of species diversity have been discussed and tested. These include processes such as productivity, competition or mutualism, as well as dispersal, genetical isolation, geographical separation, speciation, survival and extinction processes (Jansson 2003; Gaston 1996; Pianka 1966). Thus, it is clear that environmental heterogeneity and contemporary and historical climate parameters are important factors explaining regional to continental patterns of diversity, independent of the fact that the underlying processes are not yet fully understood.

In contrast to latitudinal or altitudinal gradients, longitudinal and radial gradients of species diversity (cf. Huston 1994) are often not recognised or non-existent. However, hardly any biodiversity hotspots and endemic-rich regions are located far from the sea. Exceptions to this rule are found in Asia (Caucasus, Himalaya and SW China; cf. Huang et al. 2011a, b; Mittermeier et al. 2005; Davis et al. 1994, 1995, 1997).

The inverse richness gradients of endemism from the coasts to the central parts of the continents often reflect a climatic gradient of decreasing humidity and increasing amplitude between summer and winter temperature.

Latitudinal and altitudinal patterns and gradients have often been compared with regard to ecology and biodiversity. The richness of all vascular plant taxa declines with latitude and altitude as well, and biogeographers have long recognised that the altitudinal gradient of biodiversity and climate mirrors the latitudinal gradient (Körner 2000). Stevens (1992) has shown that total maximum species diversity along an altitudinal gradient is often found towards the lower end of the elevation. This finding is compatible with the view that tropical climates harbour the highest biodiversity.

However, the mirror does not work perfectly (see Körner 2000 also for differences of latitudinal and altitudinal richness patterns). The alpine zone in mountain areas of the tropics is climatically different from arctic regions in the North or South because, for example, the amplitudes of temperature, light, vegetation period and also precipitation can differ remarkably if one compares average winter and summer or night and day values in these regions. One of the biologically most important differences between the climate of tropical alpine zones and arctic regions is that the tropics have warm days, cold nights and long vegetation periods whereas the winter of arctic regions is cold and long and the vegetation period is short.

In many temperate, subtropical and tropical regions endemism peaks at mid-altitudes; this does not reflect the altitudinal distribution of vascular plants in total which, in general, shows highest diversity at the lower end of the altitudinal gradient.

The interval of maximum endemic species richness – in absolute numbers – in Nepal is between 3,800 and 4,200 m (Vetaas and Grytnes 2002), in SW China 1,000–2,000 m (Huang et al. 2011a, b), in Taiwan 0–1,500 m (Hsieh 2002), in Peru 2,500–3,000 m (Van der Werff and Consiglio 2004), and in Ecuador 1,300–2,500 m (assessment of the list given in Valencia et al. 2000). Endemics of Corsica, in the Mediterranean, are also concentrated at intermediate levels of elevation between 1,000 and 1,800 m (see figure 48 in Hobohm 2000a). In Europe, endemics of rocks and screes are concentrated between 500 and 1,800 m, grassland endemics peak between 550 and 2,100 m, endemics of scrub and heath formations between 200 and 1,550 m, forest endemics between 300 and 1,500 m, endemics of coastal and saline habitats, arable land, ruderal habitats and settlements at altitudes between 0 and 1,200 m, endemics connected to freshwater habitats, mires, swamps, fens and bogs between 100 and 1,750 m (Hobohm and Bruchmann 2009).

We assume that the higher altitude of maximum endemism in contrast to total taxonomic diversity is related to both the stricter separation/isolation of the high mountain zones and the higher total species diversity at lower altitudes which is the basis for evolution and dispersal (cf. Hobohm 2008b; Vetaas and Grytnes 2002). Interestingly, in continental SW China and in Taiwan (Huang et al. 2011a, b; Hsieh 2002) the number of endemics and the total number of vascular plant taxa seem to peak at the same altitudinal zone in the lowlands, whereas in most other regions endemism peaks at higher elevations than overall diversity.

The number of endemics found in an altitudinal zone is also a function of the area and climate of each zone and depends on maximum elevation. These factors might explain the differences in the maxima in the two neighbouring countries Ecuador and Peru. The coastal zone of Ecuador is relatively large relative to the whole size of the country. This zone is part of the Tumbes-Chocó-Magdalena Hotspot (Mittermeier et al. 2005) and very rich in endemics. The altitude of maximum endemism in Peru is more elevated because the coastal zone is small, extremely dry and relatively poor in endemics (Van der Werff and Consiglio 2004).

The rate of endemism (as a percentage of all taxa) in a mountain range normally increases continuously with altitude. This finding is not in conflict with the patterns discussed above showing that absolute numbers of endemics peak at intermediate or lower altitudes in most cases. The increasing rates in high mountain areas are comparable with values for oceanic islands which are highest far distant from the mainland (Fig. 5.1). However, the altitudinal gradient of the proportion of endemics (in %) is normally not mirrored by a latitudinal gradient (Hsieh 2002; Dhar 2002; Vetaas and Grytnes 2002; Hobohm 2000a; Talbot et al. 1999).

Fig. 5.1
figure 00051

High rates of endemism and total species richness. In the case of oceanic islands far distant from the mainland, a high ratio (E/S) is often combined with a relatively low number of species in total (S)

Full size image

Stebbins and Major (1965: 16) analysed the endemic flora of California and concluded that

“…a high degree of endemism is found in those regions having a great variety of plant habitats.”

Hendrych (1982) examined the diversity of endemic plant species in Europe. He concluded that endemics are concentrated in high mountain areas such as the Alps, Sierra Nevada and Greater Caucasus, which are characterised by high habitat diversity.

At regional to continental scales, the diversities of higher plant species and endemic species increase with increasing habitat diversity. Areas with high habitat diversity generally also have high species diversity, although this is not always the case (cf. Cowling and Lombard 2002; Trinder-Smith et al. 1996; Huston 1994).

The main reasons are positive effects such as the number of speciation processes caused by a greater species pool and the avoidance of negative influences, including anthropogenic impacts, which are more effective in habitat-poor than habitat-rich environments. It is more likely that suitable habitats will remain in habitat-rich as compared to habitat-poor regions.

Habitat diversity is not an obligatory precondition for the coexistence of (endemic) species under stable ecological conditions. This finding does not question the positive effects of spatial heterogeneity for endemism. Spatial heterogeneity might be primarily important for the survival of species under changing climate conditions.

On the other hand, it is very easy to find high-elevation and habitat-rich ecoregions with relatively poor endemism. For example, mountain areas in northern North America, South Chile and Argentina, Norway, Central Australia, and the Cape Verde Islands are such regions. All these areas have been influenced by strong climate change, the northern and southern territories by Pleistocene glaciation cycles, Central Australia and the Cape Verde Islands by several dramatic cycles of wet and dry periods during the Quaternary (Crisp et al. 2001; Brochmann et al. 1997; Hendrych 1982).

Size of area, which is a factor often strongly correlated with the diversity of (endemic) species (see e.g. discussion in Körner 2000; Rosenzweig 1995a, b), has similar effects. According to Rosenzweig’s area hypothesis the size of geographic area is responsible for the latitudinal gradient of species diversity. This hypothesis is based on the fact that the tropics, as the largest zone and having a continuous cover stretching into the northern and the southern hemisphere, can host species with larger range sizes, resulting in higher speciation rates and lower extinction rates (cf. Rohde 1992; Rapoport 1982). But, as the discussion of Chown, Gaston and Storch (Chown and Gaston 2000; Storch 2000) has shown, too little is currently known about range sizes of species in a given latitudinal gradient or in a humidity gradient. And actually, such gradients may divide the tropics into a larger number of different wet to dry ecozones than higher latitude areas (e.g. Bailey 1998). We are not aware of any studies that have tried to examine range size differences of vascular plant taxa in relation to area size of different climate zones or ecoregions.

However, area as a driver of biological processes means very little (cf. Ricklefs and Lovette 1999; Buckley 1985; MacArthur and Wilson 1967; Abbott 1977). But, in the context of negative influences on biodiversity in combination with a variety of environmental conditions area becomes important for the probability of the survival of individuals and populations. Survival of the biota in large regions under changing conditions depends on having a sufficient number of patches so that changes in one patch will not affect other patches, and consequently there is only little or no net change over a large area. For example, almost all the mid- and high altitudes of the Alps were covered by Pleistocene glaciers. But, the Alps harbour many endemic vascular plants which survived in the valleys or in the foothills of the Alps close to the Mediterranean Basin. Stability at this scale is a stochastic phenomenon where local and regional processes such as catastrophes, gap dynamics and successions are involved, and is different from the deterministic mathematical equilibrium of the competitive exclusion principle.

2 Oceanic Islands, Continental Islands and Mainland Regions

Oceanic islands are defined by their origin and development within the marine environment. When an oceanic island rises above sea level, colonisation from other islands or mainland regions can begin. All plants which are endemic to an oceanic island must have been developed and evolved in three steps: (i) dispersal of ancestors, which in the case of a great distance is an improbable event, (ii) establishment of a founder-population which normally involves finding adequate ecological conditions and overcoming a genetic bottleneck, and (iii) speciation processes on the basis of a reduced genepool and genetic drift in an environment which is more or less different from the original habitat. Because of reduced dispersal across the sea, oceanic islands are often relatively poor in total species numbers. On the other hand, isolation favours speciation. Thus, the percentage of endemics can be high.

In contrast, continental islands were by definition part of the mainland before they became separated. When they first became islands they were normally already covered by vegetation. This is the reason why continental islands do not normally reach high rates of endemism. Because of the larger species pool the conditions for increasing endemism are favourable compared to oceanic islands, and continental islands, in general, are not poorer in endemic vascular plant taxa than oceanic islands.

Unlike islands, mainland regions do not have clear natural borders, such as a coastline, everywhere. Biogeographical information about countries such as Austria or Switzerland is often much better than information related to mountain ranges which belong to two or several countries, such as the Alps. This is the reason why terrestrial ecoregions are often analysed with a strong focus on countries.

3 Habitats

Endemic vascular plants are found in diverse habitat types in almost all tropical, subtropical, temperate, boreal and arctic climate zones of the world. Tables 5.1, 5.2, 5.3, 5.4, and 5.5 present examples of regions with the respective numbers of endemics and information about the main physiognomic habitat types to be found there. Many of the examples show a relatively high total number of endemic species. This might be a result of the major scientific effort, especially in species-rich areas, and overestimations due to taxonomic double, or even triple, identity or to the revised taxonomic status of species (Kier et al. 2005; Yena 2007). We assume that the scientific effort into biogeography in general is greater where biodiversity is high or unique.

Table 5.1 Examples of biodiversity hotspots, ecoregions and landscapes with numbers of endemic vascular plants and the main habitats in various parts of the Americas (partly nested or overlapping), from North to South or West to East, respectively
Full size table

Most regions and landscapes with endemic vascular plants represent more than a single group of physiognomic habitats. Furthermore, many endemics occur in more than a single vegetation unit. Thus, it is very often difficult to determine how many endemics are related to a particular habitat type. Even those scientists who have expert knowledge of a certain region, such as a biodiversity hotspot or an ecoregion, might find it difficult to estimate where endemics are concentrated or which habitat type harbours most endemics.

We were only able to obtain information for a few regions on both the relationship between endemism and habitat type and the importance (size) of different habitat types within the region. These regions are e.g. Europe, Cape Floristic Region, Madagascar, New Caledonia, and New Zealand (cf. Tables 5.1, 5.2, 5.3, 5.4, and 5.5).

In Europe, most endemics are found in habitats with rocks and screes or in grasslands and scrub which together cover much smaller areas than cropland, forest or urban habitats. Most endemics in Europe are basiphytes or indifferent to soil-pH, but acidic substrates are dominant throughout the continent; acidic substrates harbour fewer endemics (Ewald 2003; Hobohm 2008a).

The predominant habitat in the Cape Floristic Region is fynbos. This habitat type harbours the largest number of endemics, almost all of which grow on nutrient-poor and acidic soils.

In New Caledonia, the majority of endemics inhabit wet evergreen forest which nowadays has been reduced to a fifth of its former range. A further large proportion of the endemics is found in the maquis vegetation which covers a third of the archipelago’s area. Endemism on this archipelago is much higher on ultrabasic substrates than elsewhere.

On Madagascar, most endemics are woody plants that inhabit forest and woodland. The majority inhabit tropical and subtropical humid forest. Many others inhabit more or less dry woodlands and thickets. Today, secondary grassland and wooded grassland cover most of the island. These open vegetation types are extremely poor in endemics.

Table 5.2 Examples of biodiversity hotspots, ecoregions and landscapes with numbers of endemic vascular plants and the main habitats in various parts of the Atlantic Ocean and Caribbean (partly nested or overlapping), from North to South or West to East, respectively
Full size table
Table 5.3 Examples of biodiversity hotspots, ecoregions and landscapes with numbers of endemic vascular plants and habitats in various parts of Europe and Africa (partly nested or overlapping), from North to South or West to East, respectively
Full size table
Table 5.4 Examples of biodiversity hotspots, ecoregions and landscapes with numbers of endemic vascular plants and the main habitats in various parts of the Arctic, Asia and Indian Ocean Islands (partly nested or overlapping), from North to South or West to East, respectively
Full size table
Table 5.5 Examples of biodiversity hotspots, ecoregions and landscapes with numbers of endemic vascular plants and the main habitats in various parts of Australasia and Pacific Ocean Islands (partly nested or overlapping), from North to South or West to East, respectively
Full size table

In New Zealand, the majority of the endemics are found in alpine habitats although low and mid-altitude habitats are dominant in the landscape. Forest, woody scrub, wet and aquatic habitats also house many endemics in New Zealand.

In the following we group habitats according to Davies et al. (2004) or Song and Xu (2003), respectively.

3.1 Coastal and Saline Habitats

Many coastal and saline habitats that extend in narrow strips along coasts or shorelines represent very old arrays of environmental conditions. The abiotic factors of certain regions in this transition zone might be as old as the ocean. However, marine currents, winds and migrating birds support dispersal processes which lead to a low likelihood of genetic isolation.

At a regional scale, the database for endemism in these habitats is largely fragmentary.

Analyses of the European endemic flora show that at least 450 endemic species and subspecies are restricted to coastal habitats (Hobohm 2008a; Hobohm and Bruchmann 2009, Hobohm, database EvaplantE in progress).

Van der Maarel and Van der Maarel-Versluys (1996) found that many of the coastal endemics are not necessarily restricted to coastal habitats but that the latter represent their optimal habitat. They found that about 30 % of the endemic species are dune and beach species and another 30 % are species of maritime rocks. It should be noted that this distribution pattern varies greatly, depending on the different coastal regions of Europe, as the coastlines extend from Arctic and Subarctic to Mediterranean regions (further analyses in Van der Maarel and Van der Maarel-Versluys 1996).

The Eastern mangroves of the Indian, Indo-Pacific and Pacific Ocean on the coasts of Indonesia, Madagascar, Southeast Asia and North Australia are more species-rich than western mangroves on the Atlantic Ocean coasts of Africa and America. This might also be the case for endemic species (Schroeder 1998; Wikramanayake et al. 2002). However, at present we can only speculate on this, and further research and data are indispensable.

3.2 Inland Waters, Mires, Bogs and Fens

3.2.1 Overview

Only few comprehensive analyses on endemic vascular plant taxa deal with lakes, rivers, bogs, mires, swamps, or other aquatic habitats. These are, for example, associated with the vernal pools of California, wetlands, mires, bogs, fens and swamps of Europe, aquatic and semiaquatic vegetation on Madagascar and New Zealand, and swamps and other freshwater habitats of the Nile Delta in Egypt (Bruchmann 2011; Hobohm and Bruchmann 2009; McGinley 2008h; Hobohm 2008b; Lazar 2004; Gautier and Goodman 2003; Ranarijaona in Goodman and Benstead 2003; McGlone et al. 2001; Ferry et al. 1999; Davis et al. 1994, 1995, 1997). Information about single endemic plants and their (aquatic or wet) habitats can be found on a number of websites (e.g. IUCN Red List data, www.iucnredlist.org) and many regional floras.

Only a few regions show moderate or high endemism associated with water bodies or wetlands. The vernal pools of California harbour c. 140 endemics in ephemeral freshwater communities and related vegetation types (Lazar 2004; Keeler-Wolf et al. 1998; Zedler 1990; Holland and Jain 1988).

In Baja California only few taxa are associated with water habitats or wetlands, e.g. the palm Brachea armata which is found in permanent oases (Vanderplank in lit.).

In the swamp habitats and other freshwater wetlands of the Nile Delta in Egypt McGinley (2008f) identified 8 endemic species (of a total of 553 plant species).

According to Ranarijaona (in Goodman and Benstead 2003: 251) c. 128 aquatic and semiaquatic vascular plant taxa are endemic to Madagascar. This number is relatively low compared to the total number of endemics on the island (8,000–10,000; see Gautier and Goodman 2003). Ferry et al. (1999) suggest that the low number of endemic aquatic plants on Madagascar may be a result of Quaternary climate fluctuations; during dry periods, freshwater and semi-aquatic habitats were essentially dry, eliminating local floras which depended on wet conditions. The small total area of water bodies and wetlands and the fact that lakes and running waters are normally relatively young and discrete units are not the best conditions for promoting endemism. Compared to other regions in the world the percentage of wetland inhabitants that are endemic to Madagascar is high. Of the 338 Malagasy vascular plants that belong to the flora of aquatic and semiaquatic vegetation, 128 (or 38 %) are endemics. The relatively high ratio can be explained by the fact that Madagascar is most likely the oldest island in the world and separated from the mainland of Africa about 165 million years ago (Goodman and Benstead 2003).

For New Zealand a figure of 264 endemic species is given for wetlands (McGlone et al. 2001). The relatively high number for the much smaller areas of New Zealand (269,000 km2) compared to Europe (275 endemic taxa, 10,500,000 km2) or Madagascar (128 endemic taxa, 587,000 km2) might be explained by less marked effects of climate fluctuations, higher precipitation rates, and the higher proportion of wetland areas in New Zealand in general (De Lange et al. 2006, 2009; Connor 2002; Davis et al. 1994, 1995).

We assume that the global proportion of endemics in wetlands and water bodies is indeed very low (Photos 5.1, 5.2, and 5.3).

Photo 5.1
figure 00052

Colchicum figlalii, extremely rare and restricted to temporarily inundated depressions in serpentinite areas of Turkey (Photographed by Gerhard Pils)

Full size image
Photo 5.2
figure 00053

Nutrient-poor mire with carnivorous Nepenthes madagascariensis (S Madagascar near Fort Dauphin; photographed by Carsten Hobohm)

Full size image
Photo 5.3
figure 00054

Freshwater lagoon with Pandanus peyrierasii, Pandanus rollotii, Pandanus longistylus, Typhonodorum lindleyanum (large leaves), Barringtonia racemosa (to the left). All but one Malagasy Pandanus spp. are endemic to the island (Mandena, S Madagascar; photographed by Carsten Hobohm)

Full size image

In the following we give an overview of the endemics which are associated with the wetlands of Europe.

3.2.2 Wet Habitats of Europe

We here present an analysis based on an earlier publication (Hobohm and Bruchmann2011). The improvement of the list of endemics and discussions with colleagues resulted in a few changes. However, the main result of the earlier publication remain valid.

3.2.2.1 Analysis of Endemism Associated with Wet Habitats in Europe

The area covered here is Europe as defined in Fontaine et al. (2007). We divided Europe into 42 geographical units representing islands or groups of islands, nations, groups of small nations or, in the case of the former Soviet Union, parts of a nation (cf. Bruchmann 2011; Hobohm and Bruchmann 2009; Tutin et al. 1968–1993, see Fig. 5.2). The details are given in Bruchmann (2011).

Fig. 5.2
figure 00055

Distribution of 275 endemic vascular plant taxa associated with wet or regularly inundated habitats in Europe

Full size image

The data base EvaplantE (cf. Bruchmann 2011; Bruchmann and Hobohm 2010; Hobohm and Bruchmann 2009; Hobohm 2008a) contains information about most endemic vascular plant species or subspecies of Europe. The version EvaplantE 11/2012 shows 6,244 vascular plant taxa as endemics of Europe.

The taxonomic status and our species concept is primarily based on Flora Europaea (Tutin et al. 1968–1993) plus floras and lists of the Canary Islands, the Madeira archipelago and Cyprus (Borges et al. 2008; Izquierdo et al. 2004; Press and Short 1994; Meikle 1977, 1985). Many other regional and national floras were also used to gather information about habitats, ecological conditions, altitudes, and so on (cf. Bruchmann 2011; Hobohm 2008a, b; Kliment 1999).

In general, biogeographical analyses based on field data of plant compositions show some degree of bias. An imbalance in the biogeographical data arises as a result of geographically different perceptions and activities in the field and from the use of different taxonomies (different floras). We tried to minimise this problem by using a broad species concept and international floras and checklists, such as Flora Europaea (Tutin et al. 1968–1993, reprints 1996) and Euro + Med plantbase (cf. e.g. Greuter and Raab-Straube 2012) as primary sources.

We filtered and analysed the recent version of the data base (EvaplantE; version 11/2012) with a focus on European endemic vascular plants occurring under wet conditions or in the succession stages following inundation. We excluded taxa which primarily inhabit coastal habitats such as saltmarshes or rocky habitats near the sea, taxa that only exceptionally occur in wet or inundated habitats, and apomictic microspecies (Alchemilla, Taraxacum, Ranunculus auricomus). Recently, we excluded Marsilea azorica which is conspecific with M. hirsuta from Australia. Thus, this species was reclassified from endemic to neophytic (Schaefer et al. 2011).

Additionally, we verified or corrected the nomenclature and geographical distribution for most taxa on our list using the Euro+Med plantbase on the internet (Euro+Med 2006–2011). In a few cases, this procedure had the effect that taxa changed their designation from endemic to subendemic because they have since been found in North African countries, such as Morocco, Algeria, Tunesia, or in parts of western Asia. These taxa were also eliminated from the list.

The analysis is founded on absolute numbers of taxa per region and descriptive statistics. As a first step, we summed up numbers of species or subspecies for the whole of Europe, and numbers of species and subspecies per region. The regions are of different sizes. At the moment it is impossible to reliably determine numbers of European endemic vascular plant taxa in relation to habitat for an artificially defined grid cell. This is the reason why certain statistical methods cannot be applied (see also the discussions of species-area and endemics-area relationships, e.g. in He and Hubbell 2011; Dengler 2009; Werner and Buszko 2005; Green and Ostling 2003). Preliminary investigations, for example, showed a negative correlation between area and endemism of vascular plants in Europe (log-log-space, all endemics) because many small regions in the South of Europe show high endemism, whereas very large regions in Scandinavia and Russia are poor in endemics. Further problems and restrictions related to the statistics are discussed in Bruchmann (2011).

If we compare numbers of regions of the same area then we can use these values as direct measurements for the density of endemics (Bruchmann 2011). Also in other cases we can compare density values directly (see Chap. 2). In our division of Europe this allows a direct comparison of density values of many pairs of regions.

3.2.2.2 Endemic Vascular Plants in Wetlands of Europe

The data base EvaplantE comprises at least 275 endemic plant taxa – species and subspecies – which occur more or less regularly in wetland communities (Table 5.6). All of these are restricted to the boundaries of Europe as defined in Fontaine et al. (2007). This number is a minimum value because c. 20 % of the listed taxa in EvaplantE have not yet, or have only inadequately, been characterized in relation to ecological conditions or habitat.

Table 5.6 Comparison of the density (E/A) of wetland endemics in different parts of Europe (Left column: increasing area; abbreviations refer to the geographic areas mapped in Fig. 5.2)
Full size table

However, compared with endemics of other habitat types the number of 275 is low (Bruchmann 2011).

Only few of the taxa listed are hydrophytes living in standing or running waters. Many more taxa are not strongly associated with wet habitats and occur in both wet and other habitats (Hobohm and Bruchmann 2011, Table 5.6).

Figure 5.2 shows distribution patterns of endemic vascular plants related to wet habitats. The numbers in the South-west of Europe are higher than in the North-east. The numbers for a triangle-shaped region between Spain, former Yugoslavia and Germany are much higher than for the rest of Europe. The highest absolute numbers of taxa were found in France and Spain. Austria, Italy and Germany also have high numbers.

We assume that this fact reflects a combination of processes and conditions which favour endemism in general and endemism of wet habitats in particular (e.g. oceanic climate, high precipitation rate, humidity). Southern France is located within and between two high-mountain ranges with high environmental heterogeneity, the Pyrenees and the Alps. The country is also located between two marine environments which influence and stabilise the climate. Therefore, three major climate regimes occur in France: Mediterranean, temperate Atlantic and high-mountain climate. Mainland France is connected with other species-rich regions – e.g. the Iberian Peninsula, the Alps, Italy. Some of the Pleistocene refugia are located in the country or not far away. Thus, migration distances after glacial cycles might have been relatively short (Medail and Diadema 2009; Krebs et al. 2004).

The East of Europe shares many taxa with the West of Asia. Many landscapes, habitat types and ecological conditions are quite similar West and East of the border between Europe and Asia. The southern and south-eastern part of the continental border, in particular, is artificially defined (along the narrow Bosporus and along the Ural River). Thus, the marginal regions bordering Asia necessarily have fewer endemics than they would have if these regions were located in another part of Europe. This applies particularly to the European part of Turkey, which is rather small.

These factors together might explain the relatively high numbers of European endemic vascular plant taxa related to wet habitats in France and neighbouring countries (cf. Rull 2004; Rosenzweig 1995a, b) and relatively low numbers to the East. However, this cannot easily been verified at present.

Table 5.6 shows the result of all (250) pairwise comparisons of the density of endemism (E/A) that can be calculated directly. Some regions, such as the Madeira archipelago, Corsica, Switzerland, Austria, Italy, Germany, mainland Spain and France are only to be found in the left column. These countries show relatively high density values. Others occur only in the right column: Regions such as Svalbard and the eastern, northern and central divisions of the former Soviet Union have relatively low density values. All other regions occur in both columns.

The comparison shows that the density of endemics increases roughly from the continental regions in eastern Europe to the more oceanic regions towards the west. However, Switzerland has a higher density value than Ireland. Austria, the Czech Republic and Slovakia have higher values than Great Britain.

The highest density values on a N-S gradient are in general not represented by the Mediterranean regions, e.g. Austria has a higher density than Portugal, Greece or former Yugoslavia, and even the Faroe Islands towards the North-west have a higher density than the Canary Islands, Crete, or Cyprus, to the South. In general, temperate regions do not have lower values than Mediterranean regions. However, the boreal-arctic regions towards the north, such as Svalbard, Iceland, Finland and northern Russia, have extremely low values.

We obtained altitudinal range data for 156 taxa on the list. Wetland endemics occur at all altitudes between sea level and alpine zones. The average of the minima (median) is 300 m above sea level, the average of the maxima 1,800 m. This means that most endemics occur in the montane and subalpine zones. Gentiana bavarica occurs in damp places, e.g. moors with spring water, wet alpine meadows, and snow patches. This species represents the absolute maximum of 3,600 m above sea level in our list.

Forty-one endemics are basiphytes, 31 acidophytes. Many species are indifferent to pH or have not yet been characterized (Table 5.7).

Table 5.7 List of European endemics that are connected with wet habitats (Excerpt from EvaplantE, version 11/2012)
Full size table

3.3 Forest and Woodland

Many publications cited in Tables 5.1, 5.2, 5.3, 5.4, and 5.5 name forest or woodland as the only or as one of the main habitat types – often in the warm-temperate and tropical zones of the world. Data for the boreal-arctic and cold temperate zones is scarce and seldom provides quantified information on patterns of endemism.

In the coastal forest of the Eastern Africa Biodiversity Hotspot many endemics have been recorded from forest habitats. However, it is difficult to give quantified data because numbers for the areas studied diverge considerably between sources. While Burgess et al. (1998, 2003, 2004) state that of the 1,750 strictly endemic species 554 species are confined to the coastal forest habitat and 812 species inhabit the surrounding non-forested habitats, Mittermeier et al. (2005) claim that some 1,225 (about 70 % of all endemic species in the hotspot region) have been recorded from forest.

Approximately one half of the 1,605–1,957 endemic species within the Chilean-Winter-Rainfall-Valdivian-Forest Hotspot belong to shrub and forest ecosystems, which constitute 50 % of the whole area (Mittermeier et al. 2005).

In some biodiversity hotspots with high endemism, such as Wallacea or the Guinean forests of West Africa (Tables 5.3 or 5.5, respectively), tropical rain forest is also predominant (Mittermeier et al. 2005). In the wet tropics of Mesoamerica many endemic species inhabit rain forest, cloud forest and montane forest (e.g. Mittermeier et al. 2005) but it was not possible to find specific numbers at smaller scales for this region. However, according to Gentry (1986) in montane regions of Central America and especially the Andes, local endemism seems to result mostly from a veritable explosion of speciation in relatively few taxa, mostly of shrubs, herbs and epiphytes. These groups constitute almost half of the neotropical flora. Many of them are not associated with forest but can also be found there. And they account for most of the excess floristic diversity of the Neotropics compared with the Paleotropics. In lowland Amazonia endemism is prevalent to habitat islands. Most of the taxa involved are canopy trees and lianas, with derivative species in specialised habitats such as white sands or seasonally inundated swamp forests. In some cases, it is clear that local endemic habitat specialists are derived from wide-ranging ancestors of the terra firme forest. Habitat specialisation has obviously been the prevalent evolutionary pathway of giving rise to local endemics in Amazonia (Gentry 1986).

Photo 5.4
figure 00056

Arbutus canariensis (Los Tilos, La Palma) is endemic to the western Canary Islands (Photographed by Carsten Hobohm)

Full size image
Photo 5.5
figure 00057

Sonchus canariensis (in a Pinus canariensis forest near Villaflor, Tenerife; photographed by Carsten Hobohm) is endemic to Tenerife and Gran Canaria, Canary Islands

Full size image
Photo 5.6
figure 00058

The genus Micronychia (Micronychia macrophylla Ranomafana, Madagascar; photographed by Carsten Hobohm), including three species, is endemic to Madagascar (Schatz 2001)

Full size image
Photo 5.7
figure 00059

Ravenala madagascariensis is a pioneer tree in humid forest regions of Madagascar. The endemic monotypic genus (Schatz 2001) is planted in many regions of the humid tropics today (Photographed by Carsten Hobohm; near Brickaville, Madagascar)

Full size image

On New Caledonia (Table 5.5) endemism peaks at an extraordinarily high level, with the highest percentage numbers reached in humid evergreen and sclerophyllous forest (Lowry 1998; Morat 1993; Schneckenburger 1991).

Some authors claim that the wet tropics or the tropical rain forest harbour most (endemic) vascular plant species (e.g. Mittermeier et al. 2005; Parmentier et al. 2007). Very often, tropical rainforest is named as the centre of endemism, with extraordinarily high endemism; however, this assumption should be checked carefully, as results from several ecoregions show a contrary trend. For the Northern Territory of Australia, for example, Woinarski et al. (2006) showed that rain forest-associated species have less propensity for endemism (602 species with 41 endemics) than species associated with other habitats (3,611 species with 526 endemics). Knowledge of the occurrence, distribution patterns, and diversity of endemic vascular plants in forest ecosystems (all layers) is quite limited and sometimes fragmentary. The analyses of Kreft et al. (2004) indicate that knowledge of the structure and diversity of neotropical forests was for a long time based on incomplete floristic inventories which misjudged large groups of vascular plants.

3.4 Scrub, Heath and Succulent Shrubland

Shrub-dominated communities may either be a mature vegetation type (e.g. in subalpine zones or some types of drylands) or may occur as temporary communities (successional stages) that develop following intense disturbances or degradations of woodland or forest. If disturbances such as fire, clearing, or heavy grazing regimes are regular phenomena then the adapted shrub vegetation may remain stable over a long period of time, even if forest could grow there.

Several sources state that Mediterranean climate and semi-desert ecosystems in Southern Africa, Southwest Australia, California, the Mediterranean Basin, and Mexico harbour a high number of endemic vascular plants. Many endemic-rich areas of the world have been found to be areas of shrubland vegetation.

In the Cape Floristic Region the highest numbers of endemic taxa are mostly found in the dominant fynbos vegetation (Mittermeier et al. 2005; Davis et al. 1994; Cowling 1992).

On New Caledonia the second largest group of endemics is associated with maquis vegetation (Lowry 1998; Morat 1993; Schneckenburger 1991).

We assume that the ecological conditions in many shrub-dominated landscapes promote endemism, especially under soil conditions and precipitation rates that are relatively stable over long evolutionary time periods (Desmet 2007; Ojeda et al. 2001; Wisheu et al. 2000; Cowling and Hilton-Taylor 1999; Cowling et al. 1992, 1994, 1996). Strong influences of the major ocean surface currents in the neighbourhood might favour high climatic constancy in these regions; thus relatively low precipitation totals (including mist) in some shrub- and dwarf-shrub dominated semi-desert and Mediterranean-climate areas are highly predictable (Hobbs et al. 1995) (Photos 5.8 and 5.9).

Photo 5.8
figure 000510

Dry thicket (above) with Adansonia rubrostipa (with treads for gathering fruits), Didierea madagascariensis (background, comose) and Euphorbia laro (right) near Ifaty, Madagascar (Photographed by Carsten Hobohm)

Full size image
Photo 5.9
figure 000511

Euphorbia canariensis (cactaceous habit), forming succulent scrub communities in semi-desert zones, is endemic to the Canary Islands (Barranco Seco, Tenerife, photographed by Carsten Hobohm)

Full size image

3.5 Grassland and Habitats Dominated by Forbs, Mosses or Lichens

As Mucina and Rutherford stated, “the term grassland is one of the most used, misused and abused terms of vegetation ecology” … (Mucina and Rutherford 2006: 350, see also Gibson 2009). In fact, the term grassland is often used as a generic term irrespective of the huge variety of environmental parameters, the different patterns of floristic or plant functional compositions, habitat extension and interface structures. Here we define grassland as a group of physiognomic habitats dominated by herbs and grasses which are not used as cropland. Sometimes shrubs or single trees occur but these cover less than 10 % of the area. This definition includes steppes, grassy savannas and paramo, and is consistent with other physiognomic classification systems (cf. EUNIS classification; Davies et al. 2004; White 1983).

Suttie et al. (2005) describe grassland ecosystems as one of the originally largest ecosystems in the world; estimates of the extent of this biome range from 31 to 40.5 % of the world’s terrestrial area, depending on the definition of the term ‘grassland’. Nowadays, large areas of grassland habitats are declining in quality and quantity, and some ecosystems are even dramatically threatened by extinction (Cremene et al. 2005; Mackay 2002; Dömpke and Succow 1998; Breymeyer 1990). Therefore, it is necessary that the great variety, and thus diversity, of grassland ecosystems should receive more attention than in the past (Photos 5.10, 5.11, 5.12, and 5.13).

Photo 5.10
figure 000512

The composite Cirsium dissectum is endemic to W Europe and threatened by the decline in traditional management practices (Photographed by Carsten Hobohm; wet meadow in NW Germany)

Full size image
Photo 5.11
figure 000513

Orchis champagneuxii is endemic to the Iberian Peninsula, S France and the Balearic Islands (Plus Morocco?; photographed by Carsten Hobohm in Serra da Estrela, Portugal)

Full size image
Photo 5.12
figure 000514

Phelypaea boissieri is a parasite, endemic to Macedonia, Greece and Turkey and rare throughout almost all of its range (Photographed by Carsten Hobohm; subalpine grassland near Lake Prespa, Greece)

Full size image
Photo 5.13
figure 000515

Espeletia timotensis is a character species of the paramos of the northern Andes (Photographed by Marinus Werger; Venezuela, Mucuchie, Piedras Negras, 4,100 m)

Full size image

The characteristic aspects of grassland-dominated ecosystems are widely influenced by their different developmental history. In the past, grassland was stabilised naturally either by specific climatic conditions such as low summer precipitation and winter drought, by periodical fire events, or by strong grazing regimes. Such habitats might be very old and thus characterised by long ecological continuity. On the other hand, there is the young semi-natural grassland that could potentially become forest if anthropogenic influences (cutting, burning, grazing, mowing) were less intensive (e.g. many grasslands at low- and mid altitudes in Europe).

The recent studies on occurrence of habitat specialists in grassland ecosystems show that, along with habitat persistence at the Holocene scale (Hájek et al. 2011), nutrient availability is also an important factor underlying specialist occurrence, while habitat-specialized grassland species tend to occur in nutrient-poor habitats (Fajmonová et al. 2012).

Mucina and Rutherford (2006) showed that some of the endemic-rich habitats of South Africa are linked to high altitude regions such as Drakensberg or Wolkensberg. The Drakensberg Alpine region, which is dominated by grassland, harbours about 334 endemic taxa (Mucina and Rutherford 2006) and for this reason it has been named a Centre of Plant Endemism (Van Wyk and Smith 2001). For many high altitude regions of the world, endemic-rich grassland such as alpine meadow, is mentioned in the literature, most often in connection with stony or rocky habitats (e.g. European Alps, Carpathians, Mountains of Central Asia and others). In the European flora, the group of grassland endemics is the second largest and comprises many more taxa than e.g. forest ecosystems; this is not an area-effect because the area size of forest is much larger than that of grassland (Hobohm 2008a; Hobohm and Bruchmann 2009).

Beard et al. (2000) listed 241 endemic species for the savanna landscape of the Northern Province of Western Australia.

In summary, grasslands around the world are inhabited by a huge amount of endemics. Because this type of habitat is globally decreasing in area and quality many of the endemics are threatened with extinction (cf. Smolenice Grassland Declaration; EDGG 2010, and Hohhot Declaration 2008; on the internet).

3.6 Rock and Scree Habitats

3.6.1 Overview

Rock and scree habitats are found all over the world. Rocky habitats are dominant in many high mountain zones, in landscapes with steep slopes, and in coastal and desert regions. In these habitats differences in ecological conditions (light, water, wind speed, etc.) are much higher within small distances – a few meters – than in other habitat categories, e.g. grassland or forest. This also means that the impact of climate change might be small compared to habitats which are less heterogeneous (cf. Leuzinger et al. 2011).

The environmental conditions of rock and scree vegetation seem to have been relatively constant through time. Erosion and sedimentation patches are narrowly meshed. To find adequate ecological conditions under the pressure of climate change (glaciation periods and global warming) adapted vegetation types and endemic species would simply have to move a few hundred metres up or down the mountains (vertical displacement, Rull 2004). Rock and scree habitats as stepping stones that support the survival probability of the (endemic) species can be found at almost all altitudes of mountain ranges.

In the datasets analysed (Tables 5.1, 5.2, 5.3, 5.4, and 5.5) these habitats were not named very often as host areas for endemic species. On the other hand, in Europe, and possibly in many boreal-arctic and temperate regions worldwide, these are the habitats with very high or even the highest endemism. Studies on the endemic flora of Europe show that rocky habitats and screes represent the majority (>2,772 of c. 6,250–6,500 taxa) that are known to be endemic to Europe (Hobohm and Bruchmann 2009, data base EvaplantE in progress). In the Alps, 35–60 % of the 350–400 endemics species are found in habitats with rocks and screes, while in the Carpathians, which are not as high, only 10–15 % of 100–120 endemic species inhabit rocky habitats (Hobohm 2008a; Casazza et al. 2005, 2008; Dullinger et al. 2000; Pawlowski 1969). However, in Europe the highest endemism is confined to a habitat type which only represents a small percentage of the whole area (no exact data available yet). There is also data available on the endemic flora of the alpine regions of Iran (Noroozi et al. 2008) that reports a relationship between endemics and scree habitats. Talbot et al. (1999) list some endemic species of arctic regions that live in rock- and scree-dominated tundra vegetation.

Similarly high concentrations may exist in mountain areas of the tropics (e.g. the tropical Andes with about 15,000 endemic species) but no quantified data on endemics was found which could prove this assumption (Photo 5.14).

Photo 5.14
figure 000516

Papaver sendtneri is endemic to the Alps and shows preference for scree habitats (Photographed by Carsten Hobohm; Cogne Valley, Italian Alps)

Full size image

3.6.2 Comparison of High Endemism in Rocky Habitats with Low Endemism in Wetlands

In general, the number of endemics in wetlands is low compared to the numbers for rocky habitats and screes, grasslands, or scrub and heath landscapes (e.g. Bruchmann 2011; Bruchmann and Hobohm 2010 for Europe).

This fact should also be discussed in the context of zonal, extrazonal and azonal vegetation types. Walter (1954) defined zonobiomes as zonal vegetation types that are controlled by climate, orobiomes as different belts of mountain ranges, and pedobiomes as soil-dependent vegetation types that are azonal. For example, boreal spruce forest belongs to zonal vegetation, and aquatic vegetation, swamps, fens, bogs and riparian forest belong to azonal vegetation (cf. e.g. Bohn et al. 2000, 2003; Bailey 1998). However, the relationship between zonality and endemism is not clear. Some azonal vegetation types are very rich in endemics. Serpentine soils, for example, often give rise to sparse associations with many endemic plants (Chiarucci and Baker 2007; Alexander et al. 2007; Stevanović et al. 2003; Roberts and Proctor 1992). Rocky habitats which are strongly affected by both climate and substrate also harbour many endemic plants (Bruchmann 2011). In contrast, azonal vegetation types such as aquatic vegetation, reeds or bogs are normally poor in endemics (e.g. Parolly 2004; Meusel and Jäger 1992).

Wetlands cover a small part (a few percent) of the world’s surface (Hobohm and Bruchmann 2009, 2011; Revenga et al. 2000). This is also true for rocks and screes which are inhabited by far more endemic taxa than is the case for wetlands. Thus, the differences in endemism cannot be explained by the size of area because both habitat types represent vegetation types which cover only a very small proportion of the earth.

In contrast to rocky habitats, most freshwater habitats, mires and swamps are very young, often much younger than 10,000 years. These habitats are characterised by low ecological continuity during the late Pleistocene and Holocene. Many lakes, for example, originated as dead ice holes during the late Pleistocene and silted up or developed into fens, bogs or peaty substrates covered by forest during the Holocene (Pott 2010; Hobohm and Bruchmann 2009). Furthermore, sudden changes in the physico-chemical conditions can impact whole water bodies. Perhaps as an adaptation to this, many aquatic and wetland plants, such as Alisma plantago-aquatica, Eleocharis acicularis, Lemna minor, Phalaris arundinacea or Phragmites communis, have long-distance dispersal abilities (wind, migratory birds) and very large ranges of distribution (Meusel and Jäger 1992; Meusel et al. 1978, 1965; Hultén 1971). In Europe, even the endemics of wetlands have, on average, larger ranges than endemics of rocks and screes (Hobohm and Bruchmann 2009). The relationship between the distribution patterns of specialists and refugial history of mires in the West Carpathians and Bulgaria, for example, has been shown by Hájek et al. (2011) and Horsák et al. (2007).

Rocky habitats, in general, are composed of different micro-habitats with very different environmental conditions with respect to light, water, organic material, soil, dynamics, etc. Under changing conditions, inhabitants of rocky habitats can normally find suitable survival conditions very close to the place where they are located. The higher rates of endemism here might be due to the relatively fixed structural complexity of their environment, compared with the dynamics of the lotic and lentic aquatic environment (see Scherrer and Körner 2011). Therefore, we assume that rocky habitats and screes are less affected by changing physico-chemical conditions than standing and running waters, banks, mires and swamps.

3.7 Arable, Horticultural and Artificial Habitats

In Europe a few regional endemics (e.g. Anthemis lithuanica, Bromus secalinus ssp. multiflorus, Bromus interruptus, Carduus litigiosus, Centaurea polymorpha, Erucastrum gallicum, and Urtica atrovirens; see Tutin et al. 1996a, b, c, d, e) might be largely restricted to anthropogenic habitats, such as arable land or ruderal habitats, and so are not expected to occur in natural or seminatural habitats. Discussing the fact that certain plant species only occur in anthropogenic habitats, Gams (1938) and Pignatti (1978, 1979) explored the possibility that plant evolution is influenced by, or dependent on, human activities. Another explanation for the existence of endemic taxa in anthropogenic habitats could be that the original habitats of these taxa were destroyed, while associated endemics survived under similar environmental conditions. A combination of both explanations might also be correct.

413 European endemics occur in this group of relatively young but widespread habitats (Hobohm and Bruchmann 2009). However, we did not find much information about endemics in anthropogenic habitats outside of Europe.

Because of the duration of human influence we expect that this group of endemics might also exist in Africa and perhaps in Asia, but most probably neither in Australia nor in the Americas.

According to the IUCN Red List c. 63 vascular plant species on earth are extinct in the wild and still survive in horticulture. Rauer et al. (2000) have discussed the role of botanic gardens in protecting threatened vascular plant taxa (Photos 5.15 and 5.16).

Photo 5.15
figure 000517

Delonix regia, endemic to W Madagascar, is widely cultivated in tropical regions (Private garden in S Madagascar near Fort Dauphin; photographed by Carsten Hobohm)

Full size image
Photo 5.16
figure 000518

Melanoselinum decipiens, species and genus endemic to Madeira, is widely cultivated for fodder on this island. Original habitats are shady rocks and banks in ravines of northern Madeira (Photographed by Carsten Hobohm near Fontes, Madeira)

Full size image