http://orcid.org/0000-0002-6698-0123
Figures
Abstract
New Caledonia is a biodiversity hotspot, with an extremely high number of endemic species with narrow distribution ranges that are at high risk of extinction due to open-cast nickel mining, invasive species and seasonal man-induced fires. Mentions of micro-endemism permeate the literature on the biota of this archipelago. However, so far there has been no research comparing distribution range in different animal groups. The aim of this study is to examine the implication of different sampling effort variables in order to distinguish micro-endemicity from data deficiency, and evaluate the distribution range, frequency, and extent to which micro-endemism is common to several groups of organisms. We compiled a dataset derived from publications in Zoologia Neocaledonica, comprising 1,149 species, of which 86% are endemic to New Caledonia. We found that the sampling effort variables that were best correlated with distribution range were the number of sampling dates and the number of collectors per species. The median value of sampling dates was used to establish a cut-off point for defining adequately sampled species. We showed that, although only 52% of species were sampled adequately enough to determine their distribution range, the number of species with a very narrow distribution range was still high. Among endemics from New Caledonia, 12% (116 species) have ranges ≤5.2km2 and 3.9% (38 species) have ranges between 23 and 100 km2. Surprisingly, a similar trend was observed in non-endemic species: 22% occurred in areas ≤ 5.2 km2, and 8% in areas 23–100 km2, suggesting that environmental dissimilarity may play an important role in the distribution of these species. Micro-endemic species were predominant in 18 out of 20 orders. These results will contribute to a re-assessment of the IUCN red list of species in this archipelago, indicating that at least 116 species are probably critically endangered.
Citation: Caesar M, Grandcolas P, Pellens R (2017) Outstanding micro-endemism in New Caledonia: More than one out of ten animal species have a very restricted distribution range. PLoS ONE 12(7): e0181437. https://doi.org/10.1371/journal.pone.0181437
Editor: Wolfgang Arthofer, University of Innsbruck, AUSTRIA
Received: February 16, 2017; Accepted: June 30, 2017; Published: July 20, 2017
Copyright: © 2017 Caesar et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: MC has a Ph.D. grant from the Syrian government (grant from the Syrian government-BGF), http://www.syrie.campusfrance.org; http://www.campusfrance.org/fr; http://labex-bcdiv.mnhn.fr. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Micro-endemism, short-range endemism, narrow endemism, restricted distribution range are terms commonly and interchangeably used to describe very small geographical distribution ranges. This is a common form of rarity [1, 2], and the literature has plenty of examples of micro-endemic species and of sites that harbor exceptional rates of micro-endemism [3–5]. Despite its commonness, the quantification of micro-endemism is quite challenging, particularly because the very characterization of a distribution range can be obscured by scaling effects [6] and by differential sampling efforts in space and time, which might lead to a confusion between micro-endemicity and data deficiency.
The massive destruction and transformation of natural habitats [7] is leading to a major crisis in the history of life on Earth [8]. Therefore, characterizing species that are micro-endemic and sites that harbor important rates of micro-endemism is fundamental for establishing adequate conservation strategies for reducing major biodiversity losses and protecting several important and vulnerable functions in the ecosystem [9]. From a conservation perspective, the principal particularity of micro-endemicity is that the population size is often very small (even when density is locally high). As a consequence, a micro-endemic species’ survival is highly dependent on the long-term permanence and integrity of very localized habitats.
Although micro-endemic species can be found virtually anywhere, they are particularly abundant in some regions known for their rich and original biota derived from local evolution and diversification (e.g., the Cape flora, the mammals and frogs of Madagascar) [3–5, 10], or in some environments characterized by strong discontinuity such as, for example, limestone karsts [11,12]. These local diversification bursts have led to biotas that are rich in species with very restricted geographical ranges or with marked differences in the way they exploit local resources.
Among these remarkable regions of the world, New Caledonia is not only famous for its exceptional regional endemism [13] but also for its extremely high local micro-endemism [13–16]. New Caledonia is the oldest oceanic island [17,18]. It is considered a biodiversity hotspot due to its magnificent flora and fauna, which is characterized by a high number of locally endemic species that are under threat due to nickel mining, repeated fires, and invasive species [10, 19–23]. Mentions of micro-endemicity permeate the literature on the New Caledonian fauna and flora. For instance, studies of local insect diversification reveal that most endemic species are allopatric or secondarily sympatric on different small mountains [24–27], or in single creeks [28]. A new harvestman species was found at each newly explored locality [29]. Forty eight percent of land snails studied [30] were found in a single locality; and gecko species from the genus Dierogekko were each found to be restricted to a single site, sometimes very close to one harboring a different species, or isolated without any visible geographical barrier [31]. The same was reported for the flora, with twenty one percent of species defined as micro-endemic [32].
Repeated documentation of micro-endemism in several components of the New Caledonia biota highlights a clear need for carrying out a synthesis aimed at understanding not only its frequency, distribution area and main drivers, but also the extent to which it actually occurs in different groups of organisms. Such a study was carried out by Wulff et al. [32], providing data for understanding micro-endemicity within the New Caledonian flora. However, despite the diversity and richness of the local fauna, which has been documented in many taxonomic studies e.g. [16, 33–34], nothing has yet been done to advance our knowledge of the occurrence of micro-endemism in animals. In order to contribute to this task we designed a study for assessing the extent to which micro-endemicity is common within the New Caledonian fauna, and distinguish micro-endemism from sampling deficiency. More specifically, we defined different distribution range classes for New Caledonia and explored (1) the frequency of species in these classes, (2) a set of sampling effort variables in order to find the ones that were more likely to have implications in determining a species distribution range; and (3) we investigated whether micro-endemism significantly occurred in a set of well-sampled species from diverse taxonomic groups. In addition to that, we also searched for the status of these species in IUCN red list. As our dataset also comprised non-endemic species, we compared them to endemics in order to determine the extent to which these distributional patterns might result from local evolution.
Material and methods
Study site
New Caledonia is an oceanic archipelago in the Pacific Ocean located between 18°–23° S and 163°–169° E. It is approximately 1,500 km away from Australia, the nearest continent. It comprises a main island (Grande Terre), considered to be the oldest oceanic island (37Ma) [17–18], and several smaller and much more recent islands, making a total land surface of about 18,760 km2. Grande Terre is a long island oriented northwest-to-southeast with a surface area of 16,372 km2 (350 km in length and 50–70 km in width). Its landscape is characterized by a mountain range running along the length of the island, with five peaks over 1,500 m and a complex series of more or less connected very steeply sloped small mountains.
The dataset
We based this study on species studied in Zoologia Neocaledonica, an ongoing series of 8 volumes on the taxonomy of the New Caledonian fauna, published by the Muséum National d’Histoire naturelle in Paris. We used data published between 1988 and 2014 [33–40], the period with the highest number of publications on the systematics of New Caledonia [27]. Zoologia Neocaledonica is dedicated to the description of species, the revision of genera, and to reporting the natural history of terrestrial and freshwater animals occurring in New Caledonia. The volumes comprise 105 chapters, and represent 25% of publications on this subject over this period. In terms of number of publications, Zoologia Neocaledonica is followed by Zootaxa with 12% (50 publications), and Zoosystema with 6% (26 publications). The remaining publications over the same period of time are spread out between 117 other journals.
We built a database with 9,818 entries, in which we recorded all available information on the taxonomy (order, family, genus, species and any change in name or classification), occurrence (geographical coordinates, locality, altitude, type of ecosystem, soil, and other environmental characteristics) and collection (sampling dates, name of each collector) of each specimen. This dataset was complemented by an exhaustive search using the species name as query (using all names and combinations when there were taxonomic changes) in Web of Science, Zoological Records and Google Scholar using the following Query terms: Topic: ‘new species’ AND Topic: ‘new caledonia’ OR Topic: (new genus) AND Publication Name: (new caledonia) NOT Topic: ‘plant*’ NOT Title: ‘bryophyte’ NOT Topic: ‘moss’ NOT Title: ‘fungi’ NOT Topic: ‘bacteria’ NOT Publication Name: ‘Memoires du Museum National d'Histoire Naturelle’ NOT Topic: ‘tree*’ NOT Title: ‘annonaceae’ NOT Title: ‘lagoons’ NOT Title: deep water NOT Title: marine NOT Publication Name: ‘Systematic Botany’ NOT Publication Name: ‘New bryophyte’ Timespan: 1988–2014. This was followed by a search in Google Scholar and Google using the name of the species and New Caledonia.
When information on sampling efforts was absent, we assumed the existence of one collector and one collection date for each sample. For specimens collected with Malaise traps or other long-stay capture methods, we assigned a single sampling date independently of the number of days the trap was active.
Estimating distribution range, sampling effort and accessing conservation status.
The distribution range of each species was calculated using the convex hull method in ArcGis 10.2. The convex hull calculates the minimum convex polygon that encompasses all available records for a given species. It was used here for all species with at least three records at different points. For species with only one point of occurrence we assigned a value of 1/3 of the smallest area estimated with the Convex Hull. For those with two points, the area was calculated as follows:
where mA is the minimum area estimated with a polygon, and D is the distance between the two points of occurrence.
For 130 species that had at least 15 distinct records of occurrence, we also estimated the area with MaxEnt 3.3.3 [41], using a polygon of the New Caledonia archipelago as a mask. Environmental data were obtained from the WORLDCLIM Version 1.4 database (http://www.worldclim.org; [42]), at the highest resolution (30arc-seconds (~1 km)). Analyses were carried out with 19 Bioclim variables plus altitude. Comparison with the convex hull showed that the areas estimated with MaxEnt were much smaller (t = 16.5, df = 131.5, p-value < 2.2e-16). For this reason, and considering that the minimum convex polygon method is the one used by the IUCN for defining the Area of Occupancy (AOO) and the Extent of Occurrence (EOO) [43], we based our analysis on results obtained by the convex hull method.
Six sampling effort variables were examined in order to assess their implications on distribution range estimates. The first three described the sampling effort actually employed for sampling each species: the number of sampling dates, the number of collectors that sampled the species, and the number of specimens per species. The other three represented inferences on the possibility of being collected: the number of times a species’ collector was in the field, the number of points sampled by a species’ collector, and the mean area of the polygons sampled by a species’ collector.
Threat status was accessed from the IUCN red list at http://www.iucnredlist.org in December 27th 2016 [44].
Results
a) Characterization of the dataset
Our dataset comprised 1,149 species belonging to 352 genera, 91 families, and 20 orders (11 insect orders and 9 orders from very distantly related organisms such as Squamata, Araneae, and Mollusca). Forty four percent of genera and 59% of species were newly described. 78% of all species and 90% of newly described species were only known from Zoologia Neocaledonica without any other mention in the literature. Almost all new species (98%) were known only from New Caledonia. Sampling involved 358 collectors, 41,289 specimens, and 1,803 different sampling points (Table 1). Species distribution area ranged from a single point to 11,500 km2.
Despite differences in total numbers of endemic and non-endemic, species with narrow distribution ranges were the most frequent in the dataset (Fig 1).
b) Main trends in species that are endemic and non-endemic to New Caledonia
982 (86%) species were known only from the New Caledonian archipelago, and therefore considered regionally endemic, whereas 167 (14%) species were also recorded elsewhere. On average, the distribution area was significantly lower for species endemic to New Caledonia (less than half that of non-endemics), although the sampling effort was similar in the two sets (Table 2). For both sets of species, the median was consistently much lower than the mean, indicating that distribution area and sampling effort are highly skewed to the lowest values.
The comparison was made with a Student t test and significant values are indicated in bold.
c) Relationship between distribution range and sampling effort
The relationship between area and sampling effort was very similar in New Caledonian endemic and non-endemic species. In all cases, the correlation coefficient r between the area and sampling effort was low (Table 3). Nevertheless, there was a statistical difference between the sampling variables dealing with actual sampling and those dealing with potential sampling. On the one hand, two of the three variables describing the way the species were sampled (the number of collectors and the number of sampling dates per species) were significantly correlated with area. The number of collected specimens was the exception. On the other hand, the variables associated with the possibility of being collected were not found to be correlated with area (Table 3).
Significant correlations are indicated in bold.
d) Choosing a variable to characterize the sampling effort in the dataset
The two variables that were significantly correlated with distribution range (the number of sampling dates and the number of collectors) were only slightly correlated when considering all species together (r = 0.058, p = 0.05), and not significantly correlated when considering endemics and non-endemics separately (r = 0.031, p = 0.32 for endemics) and (r = 0.131, p = 0.09 for non-endemics). This confirms, as expected, their independence.
Nevertheless, when considering the frequency of species in different classes of sampling effort, a strong correlation between the number of sampling dates and the number of collectors was observed (r = 0.964, r = 0.897, r = 0.966, for endemic, non-endemic, and all species, respectively, with p< 0.00001 in all cases). Species sampled in very few dates or by very few collectors were the most frequent, followed by a marked decrease in frequency in subsequent classes (Fig 2).
Based on the strong correlation between these two variables, and considering that dates had stronger correlation coefficients with area, we used the number of sampling dates as a measure of the sampling effort in subsequent analyses. In addition, we used the median number of sampling dates from the entire dataset (= 3) as a cut-off point for establishing the adequacy of sampling for the inference of distribution range. That is, a species was considered adequately sampled if it was collected on at least 3 different sampling dates (Fig 3).
The black line indicates the median sampling date (n = 3), used as a cut-off point between species that are considered to be correctly sampled and those that are considered data deficient.
The relationship between distribution range and the number of sampling dates roughly described a triangular shape. This is due to the fact that a) the number of species collected with very low sampling effort was very high in all classes of distribution range; b) the number of species with a narrow distribution range was very high in all classes of sampling effort; and c) the number of species between these two extremes was much smaller (Fig 3).
e) Distribution range of adequately sampled species
51% (502) of endemic and 57% (95) of non-endemic species were sampled a number of times equal to or higher than the median (i.e. ≥ 3 sampling dates) (Table 4). Despite the difference in total number, species with small distribution ranges were the most frequent in the two datasets (Fig 4A and 4B). 116 endemic species (12%) had distribution area ≤ 5.2 km2, and 38 (3.9%) occurred in areas between 23–100 km2. The percentage of non-endemics with a small distribution range in New Caledonia was much higher (22%–36 species—had a distribution area ≤ 5.2 km2, and 8% (13 species) had a distribution range between 23–100 km2) (S1 Table and S2 Table).
Number of endemic (A) and non-endemic (B) species sampled on at least 3 sampling dates in different classes of distribution range.
Percentage values not in parentheses are for the total of each order. Percentage values in parentheses are based on the total number of species of the corresponding subset considered (total or sampled at least 3 times).
It is to be noted that although the distribution range of the set of species examined varied widely (see Table 2), there were major gaps in this continuum, with some classes of area without any species (see Fig 3). Hence, these gaps helped us define cut-off points for the different distribution range classes. This is how we arrived at the value of ≤5.2km2 for micro-endemic species. Similarly, the next group of species had distribution ranges varying from 23 km2 to 89 km2, which we synthetized as 23–100 km2.
f) Contribution of organisms from different orders
Among the 20 orders represented in this dataset, Diptera, with 419 species, was by far the most diverse, followed by Coleoptera, Collembola and Squamata. All orders had at least one, often several, species endemic to New Caledonia, whereas the non-endemic species included here belonged to only 10 different orders. Diptera was also the richest in terms of number of species endemic to New Caledonia, followed by Coleoptera, Squamata and Hemiptera. Non-endemics were present in relatively high numbers in Diptera, Collembola and Crustacea, whereas they were poorly represented or absent in the other orders. The proportion of species endemic to New Caledonia was very high for all orders, except for Crustacea (43%) and Collembola (65%).
Despite the marked difference in the total number of endemic and non-endemic species, the proportion of species sampled on 3 or more sampling dates was quite high for several orders, for both endemic (51% of the total) and non-endemic species (57% of the total) (Table 4).
The frequency of species from the most abundant orders across different classes of distribution range is presented in Fig 5. We found that high frequency in very small distribution ranges was common to almost all orders (except for Myriapoda and Orthoptera). In addition, only Diptera had species with the maximum distribution range estimated in this dataset (11,519km2). Squamata and Coleoptera also had species with large distribution ranges (up to 7,020km2). In the other orders, species distribution range never exceeded 5,500km2.
g) Conservation status in the IUCN red list
Investigation of the IUCN red list shows that the number of species from our database that have been already assessed is very low (only 76 out of the 1,149) and that only a few groups of organisms are represented (Crustacea, Teleostei and Squamata). In addition, only Squamata species have an assessment with detailed information. Crustacea and Teleostei were only classified as LC (least concern) or DD (data deficient) (Table 5). Concerning the micro-endemic species only 15 are included in the IUCN red list, and only 12 occur in protected areas (S1 Table and S2 Table).
Only three orders from the present dataset have been assessed by the IUCN. DD = data deficient; LC = least concern; NT = near threatened; VU = vulnerable; EN = endangered; CR = critically endangered.
Discussion
Although micro-endemicity has frequently been reported in researches for different groups of the New Caledonian fauna [10, 14, 16, 23, 26–27, 31], the present study is the first to be specifically designed to quantify its distribution range, estimate its frequency, and determine the extent to which these findings are common to several taxonomic groups. It is also the first time that the distribution range of this fauna is analyzed taking sampling effort into account.
The dataset we used was certainly incomplete in terms of the species covered. However, it had the advantage of including a large number of orders, comprising organisms with different life habits, sizes and natural histories. In addition, it included a high number of species that were new to science. Our dataset was compiled from species descriptions that were done within an explicit taxonomic framework, which means that scientific names and specimen occurrence data were based on the work of experts. This is undoubtedly preferable to using second hand faunistic compilations or natural history collection databases that often include uncertain taxonomic treatments and doubtful identifications.
Sampling effort variables and their implications on estimated distribution ranges
Among the sampling effort variables examined, only the number of times the species was sampled and the number of collectors were correlated with area. This highlights the importance of diversity in the way of sampling in order to increase the probability finding a wider distribution range for a given species. This result is in accordance with the topography and distribution of remnants of different forests in New Caledonia, which are marked by important discontinuities. The relief formed by the central chain, which is divided into a series of mountains with steep slopes, certainly constrains the mobility of fieldworkers. This makes it difficult to increase the area surveyed without moving from one slope to the other, or from one mountain to the next, which requires long walks or driving to access sampling sites from another flank. Thus, researchers often focus on nearby spots on a same day, and start surveying more distant sites at another time. The fact that some forest remnants are distributed across several scattered patches also contributes to that surveys are spread out over time. As expected from previous studies on the effect of sampling conditions, accessibility constraints have an impact on specific richness estimates [45, 46].
The fact that the number of specimens for a given species was not correlated with area was not a surprise, as the number of specimens is often a choice of the collectors. However, the low figures observed here suggest that in general densities are low, which is in accordance with our (PG, RP) field observations and the repeated remarks in the literature [16]. Nevertheless, more studies on population ecology are required to confirm the extent to which low population densities are prevalent across taxa in New Caledonia.
Unlike the actual sampling effort variables, we found a lack of relationship between the variables that relate to the possibility of being collected and area, which was unexpected. Intuitively, one would expect that the more the collectors sampled the territory, the higher the chances of finding a species in the entire range of its occurrence. Therefore, our interpretation for the lack of relationship between area and these variables (i.e. number of dates, number of points, area covered by all the collectors that sampled the species) is that, for many species, the estimated distribution range might be very close to reality and not a mere sampling artifact.
Another important question in this analysis was the definition of a cut-off point above which the sampling would be considered adequate. We chose to use the median sampling effort. Although arbitrary, this allows for comparisons between different datasets.
Extent of micro-endemicity
Our results show that the rate of micro-endemicity is indeed extremely high. If we had based our inferences on the entire dataset, 22% of the species endemic to New Caledonia could be considered to be very narrow endemics (with a distribution range no larger than 5.2km2). This figure is quite similar to the estimates of Wulff et al. [32], who showed that 21.7% of plants in New Caledonia are found in at least three different localities and over less than 10km2 [32].
By considering the impact of sampling effort on distribution range estimates, we refined these estimates, which led to considerations about the nature of the distribution of this fauna. The set of species that were sampled well enough to allow a sound inference of their distribution range was much smaller than the total number of species assessed. This brings the rates of confirmed micro-endemism down to 12%, (116 species with areas ≤ 5.2 km2). In addition, 3.9% of these well-studied species also occur in small areas (ranging from 23–100 km2).
Our study suggests that micro-endemism represents a general characteristic of the New Caledonian fauna, which is consistent with what has been found for the New Caledonian flora [32]. The extreme scale of this micro-endemism is remarkable as it probably involves extremely small populations, given the prevalence of small areas and low densities. Previous case studies of various groups showed that these species are well-differentiated from a morphological, molecular or behavioral point of view [22–25]. Therefore, micro-endemism arose from normal evolutionary differentiation equivalent to other highly diverse but less micro-endemic faunas.
What about non-endemic species?
A surprising result in this study concerns species that are not endemic to New Caledonia. Even if on average they had larger distribution areas (Table 2), an important fraction had very restricted ranges (22% in areas ≤ 5.2 km2, and 8% in areas 23–100 km2). Several hypotheses might explain these restricted distributions. The most likely is a) that their niches are clumped together on the island, or 2) that some of them are occasional species, which are typically found in low abundance with different habitat requirements [47]. Considering the present dataset, some of these species would be classified as common-rare-rare by Rabinowitz [1], meaning that, although they are found across distant locations, their total distribution is due to several, but very localized populations. More data on their distribution range could clarify this point. However, it highlights the difficulty of estimating distribution range and, consequently, of defining categories of threat in species with this pattern of distribution (see [43] for a discussion of this topic).
As a perspective for future research, this result should help in the interpretation of the role of environmental factors and in understanding the origin of micro-endemism in local species. The restricted distribution areas of non-endemic species could mean that distribution is primarily controlled by environmental parameters and their small-scale variation in a patchy landscape.
Distribution and conservation status
As already stated by Wulff et al. [32] concerning the New Caledonian flora, our results indicate that the IUCN red list needs an urgent re-assessment concerning the endemic fauna of New Caledonia. Our results provide estimates of the distribution range in a wide range of species, which will facilitate the re-assessment of the status of species from at least 12 different orders (S1 Table and S2 Table). Considering the ongoing environmental pressure and destruction of several habitats in New Caledonia [10, 19–23], the fact that all species with a distribution range <10 km2 occur in a single locality, and that population sizes are frequently very small, and that very few of these species occur in areas legally protected, a great number of these micro-endemic species are very likely to be Critically Endangered.
In conclusion, our study shows that the New Caledonian biota is composed of very narrowly distributed species. Based on existing information, 12% of the species were characterized as micro-endemic, with known distribution ranges no greater than 5.2 km2. Nearly 4% of species were shown to occur in areas smaller than 100 km2. This result is consistent with a previous assessment for the flora. By taking the sampling effort variables into account, we were able to show that these distribution patterns are not a spurious effect of undersampling but are the hallmark of rarity in a constraining landscape. This study sets up the foundation for a global study of micro-endemism, which should now focus on the environmental parameters that possibly determined its evolution. This would represent an invaluable source for future phylogeographic studies and for establishing policies of land management and biodiversity conservation.
Acknowledgments
This study was carried out within the framework of the Ph.D. thesis of Maram Caesar at the ED 227 of the Muséum National d’Histoire Naturelle and ISYEB. MC has a Ph.D. grant from the Syrian government (Grant from the Syrian government-BGF).
We thank the local authorities for permitting the study of the New Caledonian fauna in various circumstances, and the Labex BCDIV for supporting this research on several occasions. We are very grateful to Stéphane Prigent for his careful work on assembling the database; to Jérôme Sueur for helping with the analysis in R and designing some R scripts for analyzing the data; to Hervé Jourdan for fruitful comments and suggestions on the first drafts of our analysis protocol. We also thank Dr. Tony Whitten for constructive comments during the review of an earlier version of the manuscript.
References
- 1.
Rabinowitz D (1981) Seven forms of rarity. In: Synge H, editor. The Biological Aspects of Rare Plant Conservation: John Wiley & Sons Ltd. pp. 205–217.
- 2.
Gaston KJ (1994) Rarity. London: Chapman & Hall. 205 p.
- 3. Richardson JE, Weitz FM, Fay MF, Cronk QCB, Linder HP, Reeves G, et al. (2001) Rapid and recent origin of species richness in the Cape flora of South Africa. Nature 412: 181–183. pmid:11449273
- 4. Wilme L, Goodman SM, Ganzhorn JU. (2006) Biogeographic evolution of Madagascar’s microendemic biota. Science 312: 1063–1065. pmid:16709785
- 5. Vences M, Wollenberg KC, Vieites DR, Lees DC (2009) Madagascar as a model region of species diversification. Trends in Ecology & Evolution 24: 456–465.
- 6. Peterson AT, Watson DM (1998) Problems with areal definitions of endemism: the effects of spatial scaling. Diversity and Distributions 4, 189–194.
- 7.
Mace GM, Masundire H, Baillie JEM (2005) Millenium Ecosystem Assessment: current states and trends Biodiversity. Washington, DC. 77–121 p.
- 8. Barnosky AD, Matzke N, Tomiya S, Wogan GOU, Swartz B, Quental TB, et al. (2011) Has the Earth's sixth mass extinction already arrived? Nature 471: 51–57. pmid:21368823
- 9. Mouillot D, Bellwood DR, Baraloto C, Chave J, Galzin R, Kulbicki M, et al. (2013) Rare species support vulnerable functions in highdiversity ecosystems. Plos Biology 11(5): e1001569. pmid:23723735
- 10.
Pellens R, Ahrends A, Hollingsworth PM, Grandcolas P (2016) Assessing hotspots of evolutionary history with data from multiple phylogenies: An analysis of endemic clades from New Caledonia. In: Pellens R, Grandcolas P, editors. Biodiversity Conservation and Phylogenetic Systematics: Preserving Our Evolutionary Heritage in an Extinction Crisis pp. 237–262.
- 11. Christman MC, Culver DC, Madden MK, White D (2005) Patterns of endemism of eastern North American cave fauna. Journal of Biogeography 32: 1441–1452.
- 12. Clements R, Sodhi NS, Schilthuizen M, Ng PKL (2006) Limestone Karsts of Southeast Asia: Imperiled Arks of Biodiversity. Bioscience 56(9): 773–742.
- 13. Kier G, Kreft H, Lee TM, Jetz W, Ibisch PL, Nowicki C, et al. (2009) A global assessment of endemism and species richness across island and mainland regions. Proceedings of the National Academy of Science USA 106: 9322–9327.
- 14. Chazeau J (1993) Research on New Caledonian terrestrial fauna: achievements and prospects. Biodiversity Letters 1: 123–129.
- 15. Morat P, Jaffré T, Veillon JM (1999) Menaces sur les taxons rares et endémiques de la Nouvelle–Calédonie. Bulletin de la Société Botanique du Centre–Ouest (SBCO), 19: 129–144.
- 16. Grandcolas P, Murienne J, Robillard T, Desutter-Grandcolas L, Jourdan H, Guilbert E, et al. (2008) New Caledonia: a very old Darwinian island? Philosophical Transactions of the Royal Society of London, B 363: 3309–3317. pmid:18765357
- 17. Grandcolas P (2016) Ten false ideas about New Caledonia biogeography. Cladistics, in press.
- 18. Nattier R, Pellens R, Robillard T, Jourdan H, Legendre F, Caesar M, et al. (2016) Old Clades on Oceanic Islands: Updating the Dating of New Caledonia Biodiversity, Scientific Reports, in press.
- 19. Bouchet P, Jaffré T, Veillon JM (1995) Plant extinction in New Caledonia: protection of sclerophyll forest urgently needed. Biodiversity and Conservation 4: 415–428.
- 20. Myers N, Mittermeier RA, Mittermeier CG, Fonseca GAB, Kent J (2000) Biodiversity hotspots for conservation priorities. Nature 403: 853–858. pmid:10706275
- 21.
Beauvais ML, Coléno A, Jourdan H, editors (2006) Les espèces envahissantes dans l'archipel néocalédonien. Un risque environnemental et économique majeur. Paris: IRD editions. 259 p. + CD.
- 22. Pascal M, Richer de Forges B, Le Guyader H, Simberloff D (2008) Mining and other threats to the New Caledonia biodiversity hotspot. Conservation Biology 22: 498–499. pmid:18402591
- 23.
Pellens R, Grandcolas P (2010) Conservation and management of the biodiversity in a hotspot characterized by short-range endemism and rarity: the challenge of New Caledonia. In: Rescigno V, Maletta S, editors. Biodiversity Hotspots. New York: Nova Publishers. pp. 139–151.
- 24. Murienne J, Grandcolas P, Piulachs MD, Bellés X, D'Haese C, Legendre F, et al. (2005) Evolution on a shaky piece of Gondwana: is local endemism recent in New Caledonia? Cladistics 21: 2–7.
- 25. Murienne J, Pellens R, Budinoff RB, Wheeler WC, Grandcolas P (2008) Phylogenetic analysis of the cockroach Lauraesilpha endemic to New Caledonia. Testing competing hypotheses for the diversification of the biota. Cladistics 24: 802–812.
- 26. Nattier R, Grandcolas P, Elias M, Desutter-Grandcolas L, Jourdan H, Couloux A, et al. (2012) Secondary sympatry caused by range expansion informs on the dynamics of microendemism in a biodiversity hotspot. PLoS ONE 7(11): e48047. pmid:23139758
- 27. Nattier R, Grandcolas P, Pellens R, Jourdan H, Couloux A, Poulain S, et al. (2013) Climate and soil type together explain the distribution of microendemic species in a biodiversity hotspot. PLoS ONE 8(12): e80811. pmid:24367480
- 28. Espeland M, Johanson KA (2010) The effect of environmental diversification on species diversification in New Caledonian caddisflies (Insecta: Trichoptera: Hydropsychidae). Journal of Biogeography 37: 879–890.
- 29. Sharma P, Giribet G (2009) A relict in New Caledonia: phylogenetic relationships of the family Troglosironidae (Opiliones: Cyphophthalmi). Cladistics 25: 279–294.
- 30. Haase M, Bouchet P (1998) Radiation of crenobiontic gastropods on an ancient continental island: the Hemistomia–clade in New Caledonia (Gastropoda: Hydrobiidae). Hydrobiologia 367: 43–129.
- 31. Bauer AM, Jackman T, Sadlier RA, Whitaker AH (2006) A Revision of the Bavayia validiclavis group (Squamata: Gekkota: Diplodactylidae), a clade of New Caledonian Geckos Exhibiting Microendemism. Proceedings of the California Academy of Sciences 57: 503–547.
- 32. Wulff AS, Hollingsworth PM, Ahrends A, Jaffré T, Veillon JM, L’Huillier L, et al. (2013) Conservation Priorities in a Biodiversity Hotspot: Analysis of Narrow Endemic Plant Species in New Caledonia. PLoS ONE 8(9): e73371. pmid:24058470
- 33. Najt J, Grandcolas P, editors (2002) Zoologia Neocaledonica 5. Systématique et endémisme en Nouvelle–Calédonie. Mémoires du Muséum National d'Histoire Naturelle, 187: 1–283.
- 34. Guilbert E, Robillard T, Jourdan H, Grandcolas P, editors (2014) Zoologia neocaledonica: 8. Biodiversity studies in New Caledonia. Mémoires du Muséum National d'Histoire Naturelle, 206: 1–315.
- 35. Chazeau J, Tillier S, editors (1991) Zoologia Neocaledonica 2. Mémoires du Muséum National d’Histoire Naturelle, 149: 1–358.
- 36. Grandcolas P, editor (2008) Zoologia Neocaledonica 6: Biodiversity studies in New Caledonia. Mémoires du Muséum National d’Histoire Naturelle, 197: 1–326.
- 37. Grandcolas P, editor (2009) Zoologia Neocaledonica 7: Biodiversity studies in New Caledonia. Mémoires du Muséum National d’Histoire Naturelle, 198: 1–440.
- 38. Najt J, Matile L, editors (1997) Zoologia Neocaledonica 4. Mémoires du Muséum National d’Histoire Naturelle, 171: 1–399.
- 39. Tillier S, editor (1988) Zoologia Neocaledonica 1. Mémoires du Muséum national d’Histoire naturelle, 142: 1–158.
- 40. Tillier S, editor (1993) Zoologia Neocaledonica 3. Mémoires du Muséum National d’Histoire Naturelle, 157: 1–218.
- 41.
Phillips S, Dudík M, Schapire RE (2004) A Maximum Entropy approach to species distribution modeling In Proceedings of the twenty–first international conference on Machine learning, ACM, Banff, Alberta, Canada: 655–662.
- 42. Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A (2005) Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology 25: 1965–1978.
- 43. Hartley S, Kunin WE (2003) Scale dependency of rarity, extinction risk, and conservation priority. Conservation Biology 17: 1559–1570.
- 44.
IUCN (2012) IUCN Red List Categories and Criteria: Version 3.1. Second edition. Gland, Switzerland and Cambridge, UK: IUCN.
- 45. Crisp MD, Laffan HP, Linder HP, Monro A (2001) Endemism in the Australian flora, Journal of Biogeography 28: 183–198.
- 46. Schulman L, Ruokolainen K, Junikka L, Saaksjarvi IE, Salo M, Salo J, et al. (2007) Amazonian biodiversity and protected areas: Do they meet? Biodiversity and Conservation 16: 3011–3051.
- 47. Magurran AE, Henderson PA (2003) Explaining the excess of rare species in natural species abundance distributions. Nature 422: 714–716. pmid:12700760