Known from a handful of specimens : analyzing the worldwide patterns of occurrence and conservation of rodents and shrews recorded only from the type locality

Author Details: Giovanni Amori is a professor of zoology and works at the National Council of Research and is a leading expert of small mammal ecology, conservation and evolution. Guido Alari Esposito is a master student in biology at the Sapienza University of Rome and is interested in the mammal conservation. Luca Luiselli is a professor of ecology and works at the IDECC and at the Rivers State University of Science and Technology, Port Harcourt (Nigeria) and is a tropical ecologist and conservation biologist.


INTRODUCTION
Narrow distribution ranges and small/restricted populations are key elements of extinction risk (e.g., Gaston 1994), with species having restricted distributions and low population sizes being at greatest immediate risk of extinction.Some taxa are known only from a handful of individuals and may be functionally analogous to 'singletons' in ecological communities (Gaston 1994).As with singletons these species are still largely unknown, and it is open to debate whether face a high risk of extinction or their rarity is a collection artefact.
The orders Rodentia and Soricomorpha include about 50% of mammal species, with 2,705 species out of a total of 5,416 (Wilson & Reeder 2005).They inhabit almost every habitat and terrestrial and freshwater niche available, including flying species (Churchfield 1990;Hafner et al. 1998;Nowak 1999), and also play a key role in natural ecosystems (Golley et al. 1975;Carpaneto et al. 2011).Traditionally, conservation research and applications have been focused on large mammals (Amori & Gippoliti 2000;Amori et al. 2011aAmori et al. , 2011b)), despite the fact that many small mammal species are known from a handful of specimens.Some have been suspected to be extinct, with novel specimens not having been found for several decades (e.g., the Togo Mouse Leimacomys buettneri, see Schlitter 1989;Wilson & Reeder 2005), although some have been recently rediscovered (e.g., Řeháková et al. 2015).This study highlights the distribution patterns, main habitats and conservation implications of the rarest/most neglected species of Rodentia and Soricomorpha, known only from their type locality and less than 10 individuals (hereby FETP species).

MATERIALS AND METHODS
In this paper we considered species of Rodentia and Soricomorpha fulfilling two concurrent criteria: (i) being known just from the type locality; (ii) being known from <10 individuals.Species fulfilling these criteria are defined as FETP species.The list of FETP species was obtained using Wilson & Reeder (2005).Distribution, habitat and conservation status were obtained from the IUCN database available at www.iucnredlist.org(accessed on 15 August 2015).For each FETP species we also recorded the corresponding biogeographical region.We excluded from analysis: (i) species known from few individuals that were recorded from distinct areas.For example instance the rodent Sigmodontomys aphrastus, is known from only seven specimens originating from widely separated locations in Costa Rica, Panama and the western slope of Ecuador (McCain et al. 2007) (ii) species described after 2005 (e.g., Fukomys ilariae Gippoliti & Amori 2011), on the assumption that additional individuals of recently-described species may be found in the future field surveys.
Differences in the frequencies of species belonging to the various IUCN Red List categories between Rodentia and Soricomorpha were assessed by χ 2 test.The correlation between number of rodent species per country and number of FETP species was run by Pearson's correlation coefficient, and the same analysis was also performed to explore the correlation between number of species of each genus and number of FETP species per genus.Data were log-transformed to achieve normality when necessary, prior to apply any parametric tests.All analyses were performed by PAST 3.0 statistical software.

RESULTS
The list of FETP species for both Rodentia and Soricomorpha is given in Table 1.Overall, 91 species of Rodentia and 19 species of Soricomorpha fulfilled the inclusion criteria of our study.The number of species per genus was positively correlated with the number of FETP species in each genus (r = 0.436, r 2 = 0.190, P < 0.001; Fig. 1).
Concerning rodents, the majority of selected species occur in the Neotropical, Afrotropical and Oriental regions, whereas for soricomorphs in the Afrotropical and Oriental regions (Fig. 2).Looking at the country of origin, we observed that a high number of FETP species occurred in Argentina and Indonesia for rodents, whereas no country emerged as for the soricomorphs (Table 2).Overall, there was a significantly positive relationship between species richness of rodents per country and number of FETP species (r = 0.540, r 2 = 0.292, P < 0.01; Fig. 3).Thus, the high number of FETP species in Argentina and Indonesia merely depended on a overall high number of rodent species in these countries.

Biogeographic region (%) frequency of species
Known from a handful of specimens: rodents and shrews Amori et al.
In terms of habitat type, the FETP species of rodents and soricomorphs showed similar patterns, with most species being found in rainforest (Fig. 4).
In terms of IUCN Red List Status, the great majority of FETP species were Data Deficient (DD) in both Rodentia (71.4%, total n=91) and Soricomorpha (84.2%, total n=19).Critically Endangered (CR) species accounted for 16.5% of Rodentia and 5.3% of Soricomorpha, Endangered (EN) for, respectively, 6.6% and 10.5%, Vulnerable (VU) for 3.3% and 0%.Surprisingly, there was also a case of a Least Concern (LC) species in Rodentia (Archboldomys musseri).Based on the frequencies of species belonging to the various IUCN Red List categories, there was a statistically significant difference between Rodentia and Soricomorpha (χ 2 = 14.25, df = 5, P = 0.014), with a statistically higher frequency of CR species in rodents.In Rodentia, the majority of CR species (n = 15) came from the Neotropical region (60%), whereas there was only a single CR species in Soricomorpha, coming from India.Overall, pooling CR, EN, and VU, it resulted that most of the recorded rodents came from the Neotropical region

DISCUSSION
Our study showed that a relatively high number of Rodentia (4%) and Soricomorpha (4.4%) species are FETP.The Neotropical region housed few soricomorphs and many rodents, most likely because the former are not found in South America and only a few live in central America (Churchfield 1990).Other important regions for both rodents and soricomorphs were the Oriental and Afrotropical regions, which also represent the regions with most speciose groups (Churchfield 1990;Ghazoul & Sheil 2010).Rainforests emerged as the most important habitat type for both groups in terms of FETP species, which may be due to (i) the relative inaccessibility of wide rainforest zones in the tropics (Lewis & Berry 2012), and also (ii) overall species richness is higher in this habitat type compared to other habitats (e.g., Ghazoul & Sheil 2010).However, the pattern of occurrence of FETP species may be substantially influenced by possible biases in research efforts in individual geographic areas as well as by current attractiveness of individual taxonomic groups (e.g., dormice and squirrels vs. mice and rats).
Surprisingly, there is a lack of homogeneity in terms of IUCN Red List category (ranging from LC to CR) for FETP species, with the majority being listed as DD.We think that this results from the heterogeneity of assessors and their assumption that when a rodent species is extremely rare it is due to suboptimal research efforts.For instance, the Togo mouse Leimacomys buettneri, albeit (i) is known from just two specimens dating back to 1890, and (ii) obtained from a relatively well explored area (Bismarckburg, Adéle area in southwestern Togo), is listed as DD despite it is presumed to be extinct by several authorities (e.g., Schlitter 1999).Therefore, we would urge the IUCN authorities to revise all FETP species and give them precautionary CR status, at least when a reasonable time-span (for instance, we may tentatively suggest > 25 years) has passed since the last records.Currently, 15.8% of rodent species worldwide are DD, and application of machine-learning models to DD species predicted that several DD species should be instead considered to be threatened (Bland et al. 2015).Similarly, Howard & Bickford (2014) showed that DD amphibian species are likely to be more threatened with extinction than their fully assessed counterparts, with some regions (in the Neotropical, Afrotropical and Oriental biogeographic regions) being particularly at risk due to lack of species knowledge and higher extinction risk than currently recognized.The field effort devoted to study a given FETP species is certainly also important in determining its real status, but this variable is difficult to quantify.
In our view, a precautionary re-assessment as CR is especially due to monotypic genera, given that their loss would mean the extinction of a full evolutionary lineage instead of a single species within a lineage.Our plea is urgently needed because, till a FETP species is listed as DD, there will be virtually no way to get funds for exploring their true status, as nearly all the conservation grant sources available tend to focus on presently threatened species, i.e. those that are already listed as CR, EN or VU.Therefore, we suggest to change the paragraph 8.1 of the IUCN Red List guidelines, where it is recommended that '…if a taxon is only known from its type locality and there is no information on its current status or possible threats, the taxon should be listed as DD.If there are no plausible threats, and the area is relatively well known, Least Concern is appropriate, unless criterion A, B or C is met.If people have searched for the taxon, both at the type locality and at a reasonable number of other potential localities, and no more than 50 mature individuals are estimated, then the taxon would be listed as 'Critically Endangered'.

Figure 1 .
Figure 1.Correlation between number of species per genus and number of FETP species in each genus.For the statistical details, see the text.

Figure 2 .
Figure 2. Percent of occurrence of FETP species of Rodentia and Soricomorpha, by biogeographical region.Percentages are calculated on the total number of species included in our analysis, and not on the total number of species of the two orders worldwide.

Figure 3 .
Figure 3. Relationships between rodent species richness per country and number of FETP species.For the statistical details, see the text.

Figure 4 .
Figure 4. Percent of occurrence of the species of Rodentia and Soricomorpha known from just the type locality, with few individuals, by habitat type.