More than one fifth of the world's mammals are listed in an extinction risk category. However, extinction risk should not be the only criterion used in conservation prioritization schemes. Here, we used a species and spatial prioritization scheme for the conservation of Brazilian mammals that incorporates other parameters in addition to extinction risk. Our results suggest that 3.4% of all Brazilian mammal species are high priorities and 29.1% of all Brazilian ecoregions are also high priorities. Since 62.5% of the species identified as high-priority do not have any national conservation plan, we hope our results will help guide mammal conservation actions.
Mammals are one of the few taxonomic groups whose conservation status has been globally evaluated . Brazil is the world's leading country in mammal richness, with a total of 701 species occurring within the country , representing almost 12% of global mammal diversity . However, the status of Brazilian mammals is deteriorating , with 81 species globally listed as threatened , and 69 as nationally threatened (these two threat levels are not mutually exclusive) . Brazil also ranks high in the discovery of new mammal species [2, 7], indicating a greater mammal diversity than currently recorded for the country.
Extinction risk assessments are used as tools to guide the development of conservation strategies . Due to its transparent and objective criteria, the Red List of the International Union for Conservation of Nature and Natural Resources (IUCN) is the most common extinction risk tool used by governments to guide national policies . However, extinction risk should not be the only criterion for prioritizing species for conservation [10, 11]. For example, using extinction risk as the single parameter to define conservation priorities, we would ignore the fact that species differ in the amount of evolutionary history they represent [12, 13], even if they are classified in the same extinction risk category. Also, the implementation of conservation strategies involves political, cultural and socioeconomic issues, and parameters incorporating these conservation angles should be incorporated into prioritization schemes . Our objective is to develop a species prioritization scheme for the conservation of Brazilian mammals according to a set of parameters to complement extinction risk . We also made a spatial rank for Brazilian mammals, indicating the ecoregions that conservation agents should pay more attention to.
A recent revision updated the mammal species list for Brazil, with a total of 701 native species recorded in the country . We used the conservation priority scheme proposed elsewhere  to rank each species according to four attributes: extinction risk, degree of endemicity, taxonomic uniqueness, and public appeal.
Risk of extinction represents the temporal flexibility needed to address extinction. Species listed as Vulnerable, Endangered and Critically Endangered were assigned scores of one, two and three, respectively [5, 15]. Since national and global red lists do not always match , we adopted both lists [5, 6]. Where a species has different categories between red lists, we adopted the conservative position to list it at its worst category . The degree of endemicity represents both the spatial flexibility to conserve the species and the importance of the national population to the preservation of the species as a whole. Species endemic to Brazil received a score of three; those also occurring in other neotropic countries were assigned a score of two, and species occurring in other biogeographic realms were assigned a score of one .
Taxonomic uniqueness reflects the evolutionary history of a species. Species belonging to a monospecific genus were assigned a score of three; species in a genus with two to 10 species were assigned a score of two; and species belonging to a genus with more than 10 species were assigned a score of one . Public appeal refers to the fact that conservation actions are easier if society supports them. A score of three was assigned to charismatic species (e.g. used as pets or as cultural symbols according to ); a score of two was assigned to species that are used by human populations (e.g. hunted), and a score of one to species that do not attract human interest . According to , a combined priority score was calculated by multiplying the value assigned to each attribute to avoid redundant results. We divided the species' scores in high (>20 points), medium (>10 points =20) or low priority (≤10 points; ).
In order to indicate spatial priorities for mammal conservation in Brazil, we listed the occurrence of mammal species in each Brazilian ecoregion . We overlapped the extension of occurrence of all mammal species  with each Brazilian ecoregion , which generated a presence/absence matrix. Then, we added the scores for the species that occurred in a given ecoregion and divided by the quantity of species that occurred in that ecoregion, which generated the mean priority score per ecoregion. Then, we added the mean priority scores of all ecoregions and divided them into three equal groups, which corresponded to a priority category (high, medium or low).
From all Brazilian mammals (701 species), 14.2% (100 species) were listed in at least one red list (global or national), and therefore were assigned to a priority category. From those, 23% were high priority for conservation (41% were medium and 36% were low priority; appendix 1 and 2). Within high priority mammals, Primates comprised 74% (17 species) of the species, followed by Rodentia (17% - four species), Cetartiodactyla and Carnivora (one species each; appendix 1 and 2).
From all Brazilian terrestrial ecoregions (48 ecoregions), 27% (13 ecoregions) were high priority for mammal conservation (56% were medium and 17% were low priority; Table 1). Within the high priority ecoregions, seven were in the Amazon Forest, five in the Atlantic Forest, and the remnant in the Cerrado biome (Figure 1; Appendix 3).
Of the top priority Brazilian mammals for conservation, four were primates and two were rodents. These two orders are among the best-studied by Brazilian mammalogists . The northern muriqui and the lion tamarins are already conservation flagships for the Atlantic Forest and have action plans dedicated to them [20, 21]. There is a national action plan under preparation for the black-bearded saki . Unfortunately we do not observe this level of attention for the rodents listed as top priorities for mammal conservation in Brazil. Cavia intermedia is the mammal with the smallest range distribution in the world and one of the smallest population sizes [23, 24], highlighting this species not only as a national, but as a global conservation priority. However, there is no formal action plan or long-term research dedicated to this rare mammal. The candango mouse should be one of the country's most celebrated species, but its existence and history are almost forgotten. The candango mouse was discovered in 1965 when Brazil's new capital, Brasília, was being built. It has not been recorded since then, despite exhaustive surveys conducted within its small known historical range. Although listed as Extinct by IUCN , Brazil's national red list classifies the candango mouse as Critically Endangered . This disparity in conservation status makes it a top priority species. If extant, it is under urgent need of actions. If it is really extinct, surveys focusing on this species should put an end to doubts regarding its status.
Priority categories for Brazilian's terrestrial ecoregions. From the 48 terrestrial ecoregions, 13 were high priority for conservation.
The results of the ecoregion priorities for mammal conservation show that from the top 13 sites, seven are in the Amazon, five are in the Atlantic Forest, and one in the Cerrado. The top priority sites identified here at the national scale are also sites of global importance in other prioritization schemes, such as WWF Global 200 [26, 27], Biodiversity Hotspots [28, 29, 30], Crisis Ecoregions , Wilderness Areas [29, 32], the Last Frontier Forests  and Last of the Wild .
Two important considerations need to be made in our spatial analysis. First, the aim of our analysis was only to indicate the most important regions for conservation of mammal species and to guide resource allocation by decision-makers and conservationists. However, on-the-ground conservation actions must be planned on a finer spatial scale. Second, the geographic data used for each species (extension of occurrence) tend to overestimate the actual geographic distribution of a given species, generating commission errors. This means that the assumption that a species occurs in every single locality within its distribution is wrong. It is important to be careful when interpreting priority scores of species with large ranges, but which may have small areas of occupancy. However, as discussed above, these issues would be taken into account when on-the-ground conservation actions are implemented, a natural next step from our current analytical scheme.
Although the majority of on-the-ground conservation actions taken by governments and non-governmental organizations are based solely on extinction risk schemes, it is already known that other parameters must be taken into account [10, 11]. Setting conservation priorities is not only a scientific process, but also a socio-political one , and criteria-based conservation approaches should take this into account. One example is the thin-spinned porcupine Chaetomys subspinosus, which in a threat-only prioritization would be considered medium priority, but in this multiple approach is high priority, because despite being listed as Vulnerable, it is endemic to Brazil, it is the only species in its genus, and it is a mammal that attracts human attention.
Implications for conservation
The methodology used in this study is simple, comprehensible, and can be used for other taxonomic groups and/or regions. Since 62.5% (15 species) of all high-priority Brazilian mammals do not have any national action plan, we hope our results are taken into account when national environmental agencies (e.g. ICMBio) decide which species will be the targets for national action plans in the future .
We thank Lucas Gontijo and Pedro H. P. Braga for suggestions and analytical help in an earlier version of the manuscript. Davi Alves's research was supported by a masters scholarship from Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) and Daniel Brito's research is supported by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) (project #305631/2009-8).
Risk categories, attributes and priority scores of Brazilian's mammals (N - National; G - Global; ER - Extinction risk; DE - Degree of endemicity; TU - Taxonomic uniqueness; PU - Public appeal; PS - Priority score). We used the list with the higher extinction risk category due to the precautionary principle.
Priority categories for Brazilian's mammals. From 100 priority species, 23 were high priority for conservation.
The mean priority scores of each Brazilian's terrestrial ecoregions. The mean score is explained in the text. Ecoregions: Acr – Atlantic Coast restingas; Adf – Atlantic dry forests; Amf – Araucária moist forests; Aos – Amazon-Orinoco-Southern Caribbean mangroves; Apa – Alto Paraná Atlantic forests; Bcf – Bahia coastal forests; Bif – Bahia interior forests; Bs – Beni savanna; C – Cerrado; Ca – Caatinga; Cdf – Chiquitano dry forests; Cem – Caatinga enclaves moist forests; Cmf – Caqueta moist forests; Crm – Campos Rupestres montane savanna; Dc – Dry chaco; Ghm – Guianan highlands moist forests; Gmf – Guianan moist forests; Gpl – Guianan piedmont and lowland moist forests; Gs - Guianan savanna; Gv - Gurupa várzea; Hc – Humid chaco; Iv – Iquitos várzea; Jpm – Juruá-Purus moist forests; Jpn – Japurá-Solimões-Negro moist forests; Mav – Monte Alegre várzea; Mbf – Maranhão babaçu forests; Mgs – Mato Grosso seasonal forests; Mtm – Madeiras-Tapajós moist forests; Mv – Marajó várzea; Nbm – Negro-Branco moist forests; Neb – Northeastern Brazil restingas; P – Pantanal; Pcf – Pernambuco coastal forests; Pif – Pernambuco interior forests; Pm – Purus-Madeira moist forests; Pt – Pantepui; Pv – Purus várzea; Rnc – Rio Negro campinarana; Sam – Southern Atlantic mangroves; Scm – Southern Cone Mesopotamian savanna; Sjm – Solimões-Japurá moist forests; Smc – Serra do Mar coastal forests; Swa – Southwest Amazon moist forests; Tpm – Tocantins/Pindare moist forests; Txm – Tapajós-Xingu moist forests; Us – Uruguayan savanna; Utm – Uatuma-Trombetas moist forests; Xta – Xingu-Tocantins-Araguaia moist forests.