Why do islands have endemic species

Genetic diversity and population sizes tend to be limited, and species often become concentrated in small confined areas. The legacy of a unique evolutionary history, many island species are endemic—found nowhere else on Earth. Islands harbour higher concentrations of endemic species than do continents, and the number and proportion of endemics rises with increasing isolation, island size and topographic variety. The island of Cuba is home to 18 endemic mammals, while mainland Guatemala and Honduras, both nearby, have only three each.

Conservation is important in order to prevent global loss of biodiversity. When one species becomes threatened or goes extinct, the effect is widespread. Sometimes, it is only one endemic species that fills a need within an ecosystem; when that need is left unfilled, a chain reaction of events occurs leading to continued biodiversity loss. Estimates suggest that when one endemic plant species goes extinct, between 10 and 30 additional animal species become extinct.

Therefore, the conservation of endemic species, in particular, is important. Biodiversity is the foundation of healthy global ecosystems, in turn, healthy ecosystems sustain life, including human life. Without biodiversity, and specifically endemic species, the earth would no longer be able to produce sufficient amounts of our most basic necessities: food, water, and air.

This is why biodiversity conservation efforts should focus principally on endemic plant and animal species. Unless drastic measures are taken quickly, these species and their unique habitats will continue to decline and disappear. Red Ruffed Lemur, an endemic species found in the island country of Madagascar. Carly Dodd June 28 in Environment.

Puma, Cougar, Or Mountain Lion? Bhopal Gas Tragedy. All rights reserved. What is an Endemic? An endemic is a distinct, unique organism found within a restricted area or range. A restricted range may be an island, or a group of islands, and in the case of some endemic mammals within the Alexander Archipelago, a restricted region such as the North Pacific Coast.

The North Pacific Coast is a hot spot for endemism because of its historical isolation, ecological complexity, and narrow distribution between the Pacific Ocean and coastal mountain ranges. The long- term viability of these endemic populations is unknown, but of increasing concern.

Another possibility is that taxa belonging to these relatively young genera went extinct on the continent and only those on islands survived, which would account for their apparent range-restriction 13 , Although more research is needed, this would go against the traditional way of viewing neo-endemics, which are often associated with oceanic islands and considered to be peripheral isolates that have evolved by in-situ speciation 14 , Areas of insular paleo-endemism were found all over the world, and, contrary to theoretical expectations, on islands with contrasting histories.

Paleo-endemics are expected to be found on islands produced via fragmentation of the continental crust and are thought to have arisen through relictualization However, extinctions on the continent may also have caused the evolutionary isolation of taxa found on oceanic islands, giving birth to species considered as paleo-endemics.

Madeira, an island of oceanic origin, provided climatic refugia for tree species that went extinct on the mainland because of extreme climate changes 4.

Although many tree species currently present on Madeira are different from those in the fossil record, they are the last remnants of ancient lineages and are considered to be paleo-endemics 4. Despite the fact that paleo-endemic lineages are ancient and range-restricted, they are found on islands that may be either ancient South Island in New Zealand or recent Sri Lanka in geological terms.

In addition, as mentioned above, some young islands are also areas of paleo-endemism. One possible explanation is that these paleo-endemics are relicts of a clade that colonized different islands of an archipelago, moving from one island to another as the one they lived on degraded and subsided 23 , This hypothesis could be tested with geological data showing whether this scenario could have occurred in regions with high paleo-endemism. Other types of dispersal events cannot be discounted.

For example, ancient lineages may have dispersed from continents where they could have become extinct. Fossil records documenting the presence of similar or closely related clades on continents could confirm this kind of event. Tasmania is a nice example of how geological history and island age estimate may influence phylogenetic endemism. Consequently, Tasmania was classified as an area of paleo-endemism despite its relatively recent isolation.

This is a well-known example, however, for many islands, information on the geological age is lacking. This highlights the need to synthesize the different parameters that are used to estimate island age and facilitate their integration in a database. A change in the endemism category may reflect the difference in relative age between genera restricted to islands and those also found on continents.

For example, a change from mixed-endemism to paleo-endemism e. Tasmania may suggest that island endemics could be relatively older than genera that are also found on the mainland, favoring the relictualization hypothesis. Moreover, a relatively high proportion of islands did not change category between expanded and restricted phylogenetic endemism.

This is the case for most of islands identified as significant areas of restricted endemism, in particular those situated in the Indian Ocean e. Madagascar, Sri Lanka, Socotra. An island may be identified as belonging to the same category with either index because even when continental genera are included, the global age of the community is determined by the age of genera that are only found on islands.

Another explanation may be that, on those islands, endemic genera that are restricted to islands and those that are also found on the mainland generally have a similar age. Latitude was the localization variable that contributed the most to nearly all categories of endemism, for both restricted and expanded phylogenetic endemism. This is congruent with global patterns of biodiversity distribution, where higher richness and endemism are observed at lower latitudes for most taxonomic groups 5 , The fact that latitude is a significant factor for all categories of endemism suggests that different and non-exclusive processes may underlie the higher diversity observed in the tropics as already indicated by Gaston 37 , and that, for monocots at least, the tropics may be both a cradle, i.

The high availability of energy in the tropics, as reflected by the contribution of vapor pressure and temperature to endemism rates, may allow the co-occurrence of range-restricted species resulting in high phylogenetic endemism 17 , At low latitudes, temperatures are high nearly all year round, with much greater diurnal than seasonal variations. These diurnal variations are known to increase rock degradation, soil creation, organic matter decomposition and nutrient cycling, all of which are factors that are crucial for the establishment of plants.

In addition, high mean temperatures imply lower environmental selection on organismal life cycles, so that annual, biannual, and perennial organisms may co-occur. Therefore, the high contribution of mean annual temperature to islands classified as areas of mixed and super-endemism with the PE R index probably highlights the influence of this environmental factor on the establishment and diversification of plants on islands Wind speed was found to have a significant negative effect on the occurrence of areas of mixed and super-endemism, which may reflect the substantial contribution of wind on dispersal filtering rather than on environmental filtering.

Island floras that are exposed to strong winds, frequent storms and cyclones would be expected to be mainly composed of organisms capable of withstanding these harsh conditions. However, our results suggest that strong winds may on the contrary be a force selecting for organisms that tend to be more ubiquitous and able to colonize islands multiple times. Low wind speed, on the other hand reduces the role of wind for dispersal, thereby contributing to endemism. Similarly, island remoteness may act as a dispersal filter, and offer conditions for species survival and diversification, promoting phylogenetic endemism and in particular paleo-endemism Low wind speed is also a factor that decreases the stress on island floras and may be related to climatic stability.

Climatic stability was shown to be very important for the persistence of old clades as well as for promoting diversification, as found for continental faunas Notably, we observed that slow velocity of past climate change correlates with significant phylogenetic endemism.

These results highlight the importance of integrating wind speed, direction and variation in studies of island biogeography. It also shows the need to integrate phylogenetic information with trait diversity, which could be done for monocots, in order to assess the way wind dispersal and island isolation may select plant species and lead to island diversity. Finally, habitat availability had a key importance to explain phylogenetic endemism.

Area contributed highly to all categories of endemism e. Indeed, large islands not only offer the possibility of large population sizes, they also have diverse environments forming geographical barriers that could lead to allopatric speciation in situ 2 which, over time, could ultimately lead to single island endemics see the model from Whittaker et al.

Corroborating this assumption, the number of ecoregions also contributed highly to age categories, suggesting the importance of habitat availability for the co-occurrence of rare short and long branches Finally, elevation was also found to be an important contributor for nearly all categories of endemism.

This result highlights the importance of mountains, which support diverse habitats resulting from altitudinal gradients, different climatic conditions between windward and leeward sites, and high precipitation regimes in certain altitudinal belts.

In addition, elevation has played an important role in providing refuges of suitable habitats during the climatic fluctuations of the Pliocene and Pleistocene. For islands near the tropics, particularly in the Southern Hemisphere, elevation has contributed to the maintenance of humid forests during glacial periods when places at low altitudes became much drier In particular, the high contribution of elevation to paleo-endemism, for both expanded and restricted phylogenetic endemism, is consistent with the idea that these refuges played a key role in the survival of ancient lineages It is conceivable that on some islands that became periodically dry, mountains offered more suitable conditions for these lineages.

We cannot exclude that the data used here contained certain biases that are inherent to natural history collections records These include the fact that places that are rich in biodiversity, easily accessible, close to academic structures, or of economic importance will tend to be better sampled e.

Therefore, it is likely that remote islands with low economic or biodiversity attractiveness and with no links to academically developed countries currently or in the past will be less sampled. Another bias is the human tendency to pay more attention to rare organisms Although this tendency would increase the diversity estimate of a region, by encouraging the sampling of rare and endemic taxa, it could impact our analysis by inflating estimates of endemism.

Nevertheless, as most of the largest herbaria of the world are nearly completely digitized 47 the number of available recorded occurrences is huge we used more than 2 million. We performed several procedures to reduce potential biases i. Therefore, we predict that future field surveys and further investigation of island floras are unlikely to change the categories identified here.

They may, however, lead to the identification of additional islands with significant phylogenetic endemism. To conclude, this study allowed to identify islands across the world with significant phylogenetic endemism for monocots. As phylogenetic endemism captures how much of the tree of life is restricted to a single or to a few places worldwide, these results highlight the irreplaceable quality of these islands especially those detected with PE R index , and the need for conservation in order to avoid the loss of deep branches in the tree of life 17 , 20 , In general, the islands identified here are already areas of top priority for conservation on the basis of their species richness, level of endemism or because they are under threat e.

By conducting our study on a large number of islands worldwide and by including phylogenetic information, we were able to identify potentially undetected key areas for conservation e. The rationale for using phylogenetic diversity as a central measure for biodiversity conservation is well known 49 , 50 , 51 , Based on the assumption that shared traits are due to a shared evolutionary history, phylogenetic diversity measures the variety of features produced by life. It is also considered the best way to assess the variety of options that could allow adaptation to a changing environment and provide benefits for future human needs 49 , Our approach based on phylogenetic endemism may therefore identify islands that may be irreplaceable for, as stated above, the preservation of deep branches of the tree of life but also of option-values.

Monocots are a particularly important group regarding option-values, when we consider all the benefits they already provide for humanity.

A final important point is that we are currently unable to predict the way biodiversity will evolve, and which lineages ancient or young will be favored If we take, for example, the flowering plant Amborella or the lizard-like reptile Tuatara , we realize that their rates of molecular evolution are quite high, despite being found on long deep branches 16 , 18 , thus contradicting the a priori assumptions that ancient lineages are evolutionary dead-ends and that young lineages have a greater capacity to diversify and adapt.

All that we know is that increased phylogenetic diversity will enhance the probability of having the right feature at the right time 52 , which leads us to expect that phylogenetic diversity might represent a good marker of evolutionary potential By preserving areas of significantly high PE we may increase the probability of having species that can contribute to the evolutionary potential of a clade.

In this study, we used the monocot phylogeny of Tang et al. We extracted all records, except fossil specimens, from the GBIF portal and selected monocot taxa by using all names found in the eMonocot database as queries.

The eMonocot database provides a comprehensive list of synonyms for the great majority of species, distribution verified by experts within polygons at the TDWG level 4 scale 54 , and native or non-native species status. We used the eMonocot database to i check for synonymy ii verify the native and non-native status of GBIF occurrences, and iii check for discrepancies in distribution records, and we consequently excluded data from GBIF that deviated inexplicably from the eMonocot data.

Regarding ii , we aimed to keep only native genera Supplementary Fig. S2 and proceeded as follows. First, we compared native genera found in the eMonocot database with the occurrence data extracted from GBIF.

As different spatial scales for islands and polygons were used in eMonocot and GBIF, this procedure was only valid for a small set of islands. For these islands, we kept only genera that were present in the eMonocot database and identified as native Supplementary Fig.

All other islands were associated with a higher scale polygon defined in the eMonocot database, i. TDWG4 polygons. When a genus was identified as non-native to this polygon, the occurrence was discarded.

However, as information about native and non-native status was sometimes missing in the eMonocot database, we employed a second test to identify genera occurrences found outside their native range referring to points ii and iii.

This was done by building polygons delimited by the maximal and minimal latitude and longitude of their known native ranges and discarding all occurrences outside these limits. We used the Global Island Database provided by the United Nations Environment Programme 55 , which comprises information on , islands worldwide, as the basis of our island dataset. We considered islands to be isolated areas surrounded by water and smaller than Australia; we included only continental and oceanic islands, and excluded those found within continents e.

Only islands where at least one monocot genus was present were included, regardless of the number of species found there. Although the study was designed at the genus level we were able to estimate sampling effort using a species richness approach Supplementary Method S3.

We used modeling to predict species richness and excluded islands where the observed species richness was 5 times lower than the predicted richness, which would be an indication of under-sampling.

We chose this threshold because it allowed us to i keep all monocot genera that occur on islands we found that all genera from the islands that fell below the threshold were present on at least another island , ii remove only islands with few genera the maximum number of genera present on such islands was 8 and iii avoid excluding complete clusters of spatially close islands.

In the end, we were able to include 2,, occurrences representing 15, species from 1, genera found on 4, islands. Crossing data from three large databases - the GBIF, eMonocot and UNEP-WCMC databases , allowed us to compile, to our knowledge, the most comprehensive dataset comprising information on the largest number of islands used in island biogeography research. We used the measure of phylogenetic endemism PE of Rosauer et al. PE departs from the traditional definition of endemism as it does not represent the confinement of a species to a discrete geographic unit but measures the geographical concentration of evolutionary history compared to a broad landscape where diversity is distributed to varying degrees 17 , 19 , The PE index weights each branch of a phylogenetic tree by the inverse of the combined spatial range of all species it supports.

Here, we replaced the spatial range by the number of islands on which a genus occurs. Thus, each branch length was weighted by the inverse of the number of islands harboring the genera supported by that branch. We then summed the weighted branch lengths joining those genera.

This may be interpreted as PE on the scale of islands worldwide, which we called expanded insular phylogenetic endemism, PE E.

At this stage, we did not take into consideration that a genus can occur both on islands and continents. In a second step, we defined a measure of PE that gives more weight to genera that are only present on islands. We devised this measure to meet the objective of estimating the relative age of lineages having arisen on islands through diversification or relictualization but taking into account the specificities of the PE index.

Branches supporting continental genera were weighted by the inverse of the maximum number of islands where a genus could be found i. We called this measure restricted phylogenetic endemism, PE R , which is approximately the PE of the set of genera only found on islands and absent from continents.

With this weighting method, genera occurring on continents had very little influence on the value of insular phylogenetic endemism. This strategy avoids the exclusion of continental genera, which would have been misleading because i some deep branches supporting genera that are island endemics also support genera occurring on continents, and ii the differentiation between neo-, paleo-, and mixed endemism requires calculations based on the entire set of genera found on islands and not only on endemic ones see the following section.

We used these two indices because they capture slightly different aspects of the distribution of phylogenetic endemism.