Assessment of the rarity and conservation status of the Colombian endemic brown hairy dwarf porcupine Coendou vestitus

Abstract The brown hairy dwarf porcupine Coendou vestitus is a poorly studied Andean species endemic to Colombia. Its current Red List category is contradictory: globally it is categorized as Data Deficient but in Colombia it is categorized as Vulnerable. This contradiction has limited the implementation of conservation programmes. We evaluate the level of rarity of the species and provide consolidated information for a new assessment of its Red List status. We reviewed literature, photographs, and voucher specimens in natural history collections. Using the confirmed records, we estimated the extent of occurrence (EOO) based on the minimum convex polygon and the area of occupancy (AOO) summing the area of grid squares occupied by the species. We found that C. vestitus is very rare, with a small range, low estimated population density, occurrence in only one habitat type and small body size. The species has an EOO of 3,323 km2 and an AOO of 24 km2, based on six confirmed localities, all on the western slopes of the Eastern Cordillera, in the central Andean region of Colombia. Based on the species' rarity, restricted distribution, and threats to its natural habitat, we recommend its categorization as Endangered on the IUCN Red List. This will help delineate research and conservation efforts for this porcupine, which has a highly restricted range and inhabits the threatened Andean forest.


Introduction
T he brown hairy dwarf porcupine Coendou vestitus is one of the rarest of the seven species of porcupines (genus Coendou) that occur in Colombia (Ramírez-Chaves et al., ). It is a small species (head-body length - mm), characterized by having three types of fur: long dorsal fur, bicolored defensive fur, and bristles (Voss, ). Since its description more than a century ago (Thomas, ), it has been recorded from only six localities. This species is considered endemic to both sides of the Eastern Cordillera in the Colombian Andes, which is a complex ecosystem with topographical and biological diversity and high levels of endemism (Olson & Dinerstein, ; Armenteras et al., ; Sánchez-Cuervo et al., ). Andean ecosystems are a global conservation priority as only % of their original extent remains (Armenteras et al., ).
Although C. vestitus is considered rare (Ramírez-Chaves et al., ), this condition has not been evaluated, and it has been suggested based only on the absence of data and the paucity of voucher specimens. Information on the ecology, genetics, natural history and conservation status of the species is also scarce, and in the case of the latter, contradictory. This porcupine is categorized as Data Deficient on the IUCN Red List (Weksler et al., ), with the assessment considering the presence of the species from only two localities, although in the same assessment three localities were mentioned. Nationally, the species has been categorized as Vulnerable based on its reduced geographical range as a result of habitat loss and fragmentation (Amori & Gippoliti, ; Alberico & Moreno, ; MADS, ). Currently, it is the only porcupine species categorized as threatened in Colombia (Alberico & Moreno, ). Here, based on available literature, specimens in natural history museums and collections, and data from recent records, we evaluate the level of rarity of the species and reassess its conservation status.

Assessing rarity
To evaluate the rarity of the species we followed the criteria of Yu  () local population density, () range, () the number of habitat types in which the species occurs and () body size. In addition, we suggest factors that may have determined the rarity category of this species. As the species population density has not been assessed, we documented the number of records per year since the species description. We consider the population density to be low if the number of records evaluated is ,  for each -year interval (there are c. , records of mammals in databases for Cundinamarca, the department in which C. vestitus has been historically recorded). We also estimated the species range and related this to the rarity level using the extent of occurrence (EOO) and area of occupancy (AOO). We calculated the EOO using the minimum convex polygon (linking the known points of occurrence for the species), and the AOO by summing the area of grid squares in which the species is known (using grid squares of  km  as recommended in IUCN, ) in GeoCAT (Bachman et al., ). For this we used data from six confirmed localities. Three of these are voucher specimens housed at Colombian collections and the other three are photographic records ( Table ). The photographs showed characters used to differentiate C. vestitus from other species in this genus: dorsal pelage with long blackish fur that partially or completely conceals defensive quills, and bicoloured bristle-quills (Voss & da Silva, ; Voss, ; Plate ). We excluded the record ICN  from our analyses because the specimen was transported from another locality in the western part of the Eastern Cordillera; the location on the label of the specimen is in a market area (Alberico & Moreno, ).
We also considered the number of habitats occupied by the species. For this, we overlaid the EOO on the ecoregions of Colombia, following the classification of terrestrial ecoregions (Olson et al., ). We evaluated body size based on information from the labels of the reviewed voucher specimens and from the literature (Voss, ; Ramírez-Chaves et al., ), and compared this trait with other species of Coendou.

Conservation status
To reassess the species conservation status, we used information on the level of rarity, EOO and AOO (IUCN, ). We included the level of rarity and the ecoregions the species inhabits to infer the conservation status of the species because of the absence of information on other factors that can influence the AOO, such as biotic interactions (predation, competition) and landscape (connectivity and shelter) (Bernard et al., ). We also examined whether the EOO or AOO of C. vestitus overlaps with protected areas, by using the protected areas layer for Colombia (October ; SINAP, ), and evaluated the per cent of forest area that remained unchanged during - within the polygon of the species range (IDEAM, ). To evaluate overlap of the AOO and EOO with protected areas, we estimated the extent (km  ) of protected areas inside the EOO polygon using as a limit the elevational range of the species, and determined the number of confirmed localities inside protected areas.

Rarity
Our findings indicate that C. vestitus matches the criteria of an extremely rare species (Category H; Yu & Dobson, ), based on all four factors evaluated.
Population density From the description of C. vestitus to the present, we found only  voucher specimens and photographs of three living specimens. Three records had no precise locality information. The records date from the species description in  to photographic records from Cundinamarca in  (Plate , Table ). The scarcity of records suggests a low population density, considering that several biological expeditions have visited the area in which the species occurs (Cundinamarca is the Department in which Bogotá, the capital of Colombia and home of the main Colombian academic institutions, is located).
Range It has been suggested that the elevational range of C. vestitus is ,-, m (Barthelmess, ), -, m (Alberico et al., ) and -, m (Ramírez-Chaves et al., ). However, based on the information from confirmed localities, and discarding a dubious record from Villavicencio, the elevational range appears to be from , m (Cundinamarca, Quipile) to , m (Cundinamarca, Chicaque). The estimated area of the species range is based on only six confirmed localities ( Table , Fig. ), with an estimated EOO of , km  and an AOO of  km  .

Number of habitat types in which the species occurs
Overlaying the EOO on the terrestrial ecoregions map indicated that the species only occurs in tropical moist broadleaf forest (Olson et al., ). This ecoregion corresponds to the highly threatened sub-Andean and Andean forests.
Body size The adult body size of C. vestitus (total length of head and body - mm) is within the range observed for small Coendou species (C. insidiosus - mm; C. nycthemera - mm; C. pruinosus - mm; C. rufescens - mm; C. melanurus - mm; C. quichua - mm; C. speratus - mm; C. spinosus - mm; C. bicolor - mm; C. prehensilis - mm), being the third smallest species of the genus after C. ichillus (- mm) and C. roosmalenorum ( mm). Coendou vestitus has, however, a considerably shorter tail than the latter species.
Conservation status We found only two localities (of the six confirmed) within conservation areas (.% of the occurrences), although the EOO polygon intersected with  protected areas, managed by two institutions (Corporación Autónoma Regional and Parques Naturales Nacionales de Colombia). Specifically, these  areas include one Soil Conservation District, five Regional Protected Forest Reserves, eight Regional Integrated Management Districts, and  Natural Civil Society Reserves (UNEP-WCMC, ). The portions of the protected areas within the EOO polygon have a total area of , km  , with , km  of the EOO not lying within protected areas (Fig. a). When overlapping the forest coverage with the EOO polygon, the forest coverage during - was  km  (.%). Considering the species' range and its rarity, we recommend that C. vestitus is recategorized from Data Deficient to Endangered based on the following criteria: () Bb(iii,iv,v) and Bb(iii,iv,v): with an EOO of , , km  (B) and continuing decline inferred (b) in extent and/or quality of habitat (iii), number of locations or subpopulations (iv) and the number of mature individuals (v), and similarly with an AOO ,  km  (B); () Cb: population size estimated to number , , mature individuals with a continuing decline, observed, projected, or inferred (C), and, continuing decline, observed, projected or inferred, in numbers of mature individuals, and extreme fluctuations in number of mature individuals (b).

Discussion
Our recommendation to categorize C. vestitus as Endangered follows IUCN () recommendations to assess poorly known taxa based on information on inferred habitat loss and restricted distribution, to avoid assigning a Data Deficient category. The current IUCN information (Weksler et al., ) is incomplete (with AOO and EOO unknown) because it includes data for only from two localities (Voss, ). The periodic re-evaluation of a species' Red List status is an important tool for the planning, monitoring and management of biodiversity conservation (Hoffmann et al., ; Angulo et al., ).
Besides being an endemic and rare species, our findings confirm that C. vestitus is restricted to an ecosystem in which habitat is being lost and there are continuing threats from anthropogenic transformation (Thiollay, ; Armenteras et al., ; Angulo et al., ). The Eastern Cordillera comprises % of Andean ecosystems, but only % of its original vegetation cover remains. Although this region is characterized by high species richness and endemism it is one of the least known and least protected ecosystems (Armenteras et al., ; Rodríguez-Erazo et al., ). Habitat loss affects the persistence of small mammals, which play important ecological roles, for example as seed dispersers of pioneer species, and in trophic and predator-prey relationships (Decher, ; Brose et al., ; Lima et al., ; Tsai et al., ).
Although C. vestitus has previously been considered a rare species because of the scarcity of records (Ramírez-Chaves et al., ), there are other reasons for it to be considered rare (Cofré & Marquet, ). Although other porcupines in Colombia are known from fewer specimens (e.g. C. ichillus), they are not endemic to the country, having a wider distribution (Voss, ). Our confirmation of rarity is based on a combination of apparent low local population density, a small range, occurrence in only one habitat type, and small body size.
In general, small-sized species of Coendou (e.g. C. ichillus, C. roosmalenorum and C. vestitus) have more restricted distributions in the northern part of South America compared to the larger species that have only one type of quills in adulthood (e.g. C. prehensilis and C. bicolor; Voss, ).
The rarity of C. vestitus is perhaps associated with homoplastic functional traits such as the presence of three types of hairs in adulthood (i.e. with less protection against predators than species with a body fully-covered by quills) (Voss et al., ) and small body size (Gaston & Blackburn, ; Yu & Dobson, ). Several species of African small mammals have been categorized as rare or Vulnerable (Schlitter, ; Decher, ) because their size could influence predator-prey relationships (Brose et al., ; Tsai et al., ). In this context, additional morphological characters may confer an adaptive advantage to large Coendou species: a body mostly covered by quills in adulthood provides a possible advantage against predators (Scharf et al., ; Speed & Ruxton, ), the swollen nasofrontal sinuses protect the brain, and a larger tail facilitates arboreal locomotion, as observed in C. bicolor and C. prehensilis (Voss et al., ; Voss, ). The morphological and/or evolutionary reasons for the restricted range of small-sized Coendou are, however, as yet unclear, and require further research.
Although  genera,  species and  subspecies of rodents are considered threatened and have high endemism (Ceballos &  only .% of the species' range (Fig. a). In addition to the small EOO, the area of the  protected areas within the EOO is small (a mean of  km  per protected area). The area surrounding the EOO is severely affected by extensive commercial plantations and urban settlements, with only c. % of the ecoregion unaffected (Armenteras et al., ; Sánchez-Cuervo et al., ; Angulo et al., ). This limits connectivity (Bright, ), which is important for the persistence of a species (Passos et al., ). Maintenance, extension or connection of protected areas, connecting the relict habitats, could help to protect C. vestitus (Cofré & Marquet, ; Amori & Gippoliti, ; Armenteras et al., ).
Conservation strategies and financial resources need to be established for threatened and endemic species and for species with restricted distributions (Cofré & Marquet, ; Isaac et al., ). However, for mammals most monitoring and conservation efforts are directed at large or charismatic species. Less attention has been directed at rodents, even though the group has a high extinction rate (Amori & Gippoliti, , ; Armenteras et al., ). Prioritizing the conservation of C. vestitus has the potential to contribute to the protection of the ecosystems in which it occurs and of co-occurring species. We recommend that national agencies prioritize this porcupine species, together with other species in urgent need of monitoring.
Porcupines remain a poorly known group, both at national and Neotropical levels (Voss, ). Knowledge of the ecology of C. vestitus is mostly based on inference from other porcupine species (e.g. Alberico & Moreno, ; Weksler et al., ). Our compilation of information on C. vestitus highlights the need for further fieldwork and data collection. Nevertheless, threats to porcupines are evident, in particular loss of habitat, illegal trade, road-kills, and hunting for consumption (de Freitas et al., ; Racero-Casarrubia et al., ). In Colombia, illegal captivity has also been documented (on a voucher specimen label; Table ) as a threat to the species. Our compilation of data and our findings form the basis for further research and for the establishment of conservation strategies and future evaluations of the distribution and conservation status of C. vestitus.