Introduction
A species introduced outside of its original distribution range is referred to as an alien (or adventive, exotic, foreign, introduced, non-indigenous, non-native; [1]) species. Although no general consensus exist, an alien species may also be classified as invasive, based on impacts on the new environment [2] or transposed ecological and biogeographical barriers [3][4], depending on the school of thought. In either case, when aiming at protecting natural landscapes and the whole ecosystem thereof, however, it is inadvisable to wait until negative impacts caused by known alien species in a new environment are perceived to act. In fact, when the negative effects of a biological invasion are perceived, eradication of invasive species is considerably more expensive and harder to achieve [5]. To understand to what extent an alien species may threaten a native environment or become invasive, it is first necessary to assess its population structure and dynamics, such as survival rate and abundance [6]. Such population parameters also allow for simulation of management strategies to reduce the risk of eradication failures.
However, in some cases, logistical challenges and high data collection costs may impede effective data gathering. This is especially true for island environments in developing countries due to limited personnel and financial resources. Species that are elusive or cryptic may further halt data acquisition. Under such circumstances, estimating wildlife population parameters can become challenging and the probability of generating a sparse dataset – dataset with limited data points – is increased [7]. Still, when dealing with alien species the importance of generating information for management outweighs the uncertainty that comes with analyzing sparse datasets [8].
The rock cavy (Kerodon rupestris; Rodentia, Caviidae) exemplifies the most recent known introduction on the island of Fernando de Noronha (Brazil), intentionally done by the military to be used as eventual food resource [9]. According to Oren [10] and Alves and Leite [11], between 1967 and 1969, four rock cavy specimens (two males and two females) were captured on the continent and released on Fernando de Noronha. Since then, its population grew on the island, benefited by the abundant presence of rocky outcrops, the species’ natural habitat in the Caatinga, the Brazilian dryland [12]. The island of Fernando de Noronha, located in the state of Pernambuco, belongs to an archipelago with the same name. Since its discovery during the 16th century, it has been colonized by several new species. Despite its potential relevance for this island’s conservation, the impacts, colonization stage, and rock cavy population parameters are still largely unknown.
Cost and logistics play an important role on rock cavy data collection on Fernando de Noronha. However, most of the absence of information on this insular rodent population is likely due to considerably low capture rates for capture-recapture studies. In the continent, trapping success has been cited as far less than 1% [13]. Lacher [13] described that the only successful method for capturing wild rock cavies in the continent was by cornering individuals with the help of dogs, and actively constraining them by hand. Zappes, Portella, and Lessa [14] described that using a combination of wires and nets the authors managed to capture one female of Kerodon acrobata, which was only recently described as a separate species from the K. rupestris [15]. The difficulties of capturing rock cavies reported in the past are still largely corroborated by recent studies [12][16][17][18][19] and inevitably lead to the generation of sparse datasets.
The present study aimed at generating unprecedented information about this insular rock cavy population such as its survival and capture probabilities, colony density and, ultimately, total population size. Such information is highly valuable for alien species management planning to protect this UNESCO World Heritage Natural site [20], even if coming from sparse datasets.
Methods
Study area
The oceanic archipelago of Fernando de Noronha is located on the Northeast of Brazil, approximately 360 km ashore. It is constituted by 21 islands and islets with the biggest one being inhabited (Ilha de Fernando de Noronha), measuring 18.4 km2. The climate is tropical with dry (August to March) and rainy (March to August) seasons [21]. The original island vegetation was removed when it served as a prison during the 19th century, but recovered to form a xeromorphic, seasonal, deciduous vegetation, with herbaceous, shrub and forest physiognomies [22]. Of the historically registered vascular plant species, 4,2% are endemic (n=14), but it is likely that some of these species have already been extinct while invasive species such as the evergreen leucaena (Leucaena leucocephala) prosper in the archipelago [22]. Not only the climate and the rocky geology of the island, but also the relatively abundant vegetation in comparison to its original habitat, are a favorable environment for the rock cavy. We chose a rock cavy colony to calculate species density, as well as survival and capture rates. We define a colony as a rocky outcrop formation with defined size, where a group of rock cavies – a gregarious species – can be found. All rock cavy colonies on Fernando de Noronha compose its insular population. Even on the continent, rock cavies are rarely seen outside of their colony and direct surroundings [12], and have been shown to be displaced by tourist presence [23]. The chosen study colony is locally known as Boldró (03°50’45.23’’ S, 32°25’44.80’ W, datum WGS84), where individuals are less skittish due to the constant tourism presence.
Study species
The medium sized rock cavy, Kerodon rupestris (Caviidae), averages 350 mm in body length and weights up to 1,000 g [24]. Although at times reported as crepuscular [25], the rock cavy on the archipelago presents diurnal behavior. The species has a hierarchical familiar organization and relies on running and hiding on rock cracks to avoid predation, with females typically produce 1-2 offspring per litter and breeding potentially multiple times per year, especially during the rainy season [12][24][26][27]. This species has a relatively restricted geographic distribution, occurring from Piauí to northern Minas Gerais state and shows remarkable adaptability to semi-arid conditions, utilizing both terrestrial and arboreal habitats within its range [28]. These areas in northeastern Brazil receive between 400 to 1,200 mm of annual rainfall, with less than 800 mm being most common [29]. Secondary data on the species are scarce. Even regarding the continental populations, there are only 13 behavioral and ecological studies that generated some sort of information about the rock cavy [12][13][17][23][24][26][30][31][32][33][34][35]. Most of the more recent studies citing this species are focused on its morphology, parasites or genetics [15][18][25][36][37][38][39][40][41][42][43][44][45][46][47][48][49][50]1820. There are no scientific studies on this insular population.
Data Collection
Two field trips that lasted 20 days in total were performed in April and October 2014. To estimate survival and capture probabilities, as well as number of individuals in the Boldró colony (i.e., to calculate colony density) we generated two datasets, a capture-mark-recapture by observation (i.e., capture-recapture, detailed in section Capture and marking) and a population survey (i.e., counting, detailed in section Population survey and recapture). This colony was used due to the easier access both for capturing the individuals and for observations. Moreover, the individuals in the colony are more used to humans, which increases the probability of observation, and thus, our dataset for density calculation. We believe, however, that this colony is representative of the other ones observed on the island. The model used to estimate the population parameters from these datasets is described in the section zero-truncated poisson log-normal mixed effects model (ZPNE). We then searched for rock cavy colonies on the island by prospecting transects across the island (Figure 1) and calculated colony area of all those that were detected using satellite imagery. We confirmed the presence of rock cavies on colonies by actively searching for individuals, presence of latrines, and typical cavy herbivory (i.e., leaf and eventual bark consumption). These methods are described in detail in section Identification and measurement of colonies. Using the total area of colonies and the calculated density of rock cavies, we then calculated the total population size (described in the section Calculation of total population size).
Capture and marking
Capture and marking were performed daily, totaling five occasions on the first field trip, and 15 occasions on the second. Animals were trapped with six Tomahawk (two size 900 x 210 x 210 mm, four size 450 x 210 x 210 mm) and two Sherman (430 x 125 x 145 mm) live traps. The eight traps were baited with local and exotic tree leaves and flowers, and placed on the floor around the colony under tree shadows (to avoid death by insolation) at least 3 m apart from each other. The traps were placed strategically to maximize captures (i.e., near often visited trees). In total, eight individuals were trapped and marked. Their complete capture history can be found in Table 1. Traps were opened in the early morning after population survey observations, and closed in the end of the afternoon, comprehending a total of 1.795 hours of trap effort. Only adult individuals were captured.
Captured individuals were submitted to an anesthetic protocol of ketamine (20 mg/kg), xylazine (2 mg/kg), meperidine (15 mg/kg), midazolam (2 mg/kg) and atropine (0.04 mg/kg) IM in the pelvic member (Dr. Paulo R Mangini, DVM; licenses no. 38804-4 and 38804-5). Under anesthesia, two numbered earrings (one metal and one bright colored) were placed on the right (females) or left ear (males). As ear samples were also collected for genetic analysis (not included in this work), each individual presented a unique set of earmarks, allowing for individual identification. Both sides of marked individuals were photographed to aid posterior individual identification by observation. All marked specimens were individually identifiable throughout the study. The animals were released where captured after recovering from anesthesia. Observation of marked individuals for composing the capture-recapture datasets was performed while counting the population, as detailed below.
Population survey and recapture
A modified fixed-radius point count method was used as an abundance assessment method [51]any of which can be used to test for differences in community composition among sites, or for differences in the abundance of a given bird species among sites. These indices are (1, where the observed extent was determined by the colony area, which is clearly delimited. Observations were performed once daily, at approximately the same time (between 6:00 h am and 8:30 h am) using photographic equipment (Sony Cyber-Shot DSC-HX100V 30x Zoom) and yielded a total of 7.18 hours of observation (average of 20 minutes per day of activity). For the population survey, all visible animals were counted while individual differentiation was possible to avoid double counting. At the end of each survey day, the maximum number of different non-marked individuals seen was noted as the daily count. Notes and photographs on all observed marked individuals were taken as individual recapture information.
Zero-truncated poisson log-normal mixed effects model (ZPNE)
The chosen model for this work was the zero-truncated poisson log-normal mixed effects model [52], which we used with a sin link function due to data type and structure. The ZPNE allows different types of input data – both capture-resight and population counting datasets – and is a full-likelihood robust design, allowing for a capture ‘occasion’ within each field trip. This means that each day is considered a different sampling occasion with a closed population within a field trip, and open populations (i.e., death and birth) between field trips. Due to these characteristics, the ZPNE reduces uncertainty in parameters generated by sparse datasets, which is the case of the present study. Outputting survival, capture probability and population size at each sampling occasion, the ZPNE is available off-the-shelf, in the software MARK [53]. Different formulations of the ZPNE were tested and the most appropriate one ZPNE (i.e., smallest AICc) presented constant survival (φ) probability and varying number of individuals in the colony (N) across field trips. Considering the colonies seem to be around carrying capacity (please see discussion regarding carrying capacity on section The rock cavy distribution, population structure, and status on the island) and there are no changes to survival pressures (i.e., stable poaching and predation rates) between the trips, such parameterization choice is easily explained.
Immigration was assumed to be irrelevant due to both the short time between field trips and as studies in captive colonies show that the rock cavy is extremely territorial and do not allow immigration to a colony which is at its carrying capacity [26]. The only study available on K. rupestris’ social system [54] observed that emigration, and consequently immigration, involve juvenile individuals exclusively. Therefore, the parameter that accounts for ‘unavailability’ (γ’; or temporary emigration) was fixed to zero. Permanent emigration from the colony was effectively equivalent to mortality, as studies showed that increasing the density by even a small amount (0.09 animals/m2) in a captive experiment resulted in severe territorial disputes, with immigrant individuals typically being violently rejected and killed when attempting to join the established colony [13]. Moreover, the survival of an individual lacking the protection of the colony is also unlikely due to high predation rates of rock cavies by feral cats [55], as well as dogs and poachers. Capture by observation (p²) and individual heterogeneity (σ) were considered to be constant through time as the landscape did not present significant changes between field trips. All colonies were also assumed to be homogenous with regards to their structure, density, and behaviour. Such assumption is necessary to allow us to calculate the insular population size. Although only available through extrapolation, such information is extremely valuable for management purposes. Considering the complete lack of information on this population, and due to the small island size and the very narrow genetic lineage, we believe it is an acceptable assumption for the time being.
Identification and measurement of colonies
Rock cavy colonies are found in open rocky formations or rocky outcrops (i.e., large stones piled up). We identified 68 colonies through 13 transects walked across all different habitats on the island, totalling ~30 km (Figure 1). We marked the location of the colonies using a GPS device (Garmin eTrex 30® and Garmin Colorado 300®) and overlayed the points on a imagery captured by the ©DigitalGlobe Quickbird satellite (65 cm pan-sharpened resolution) using the QGIS software [56]. All colonies identified on the ground were seen in the satellite imagery except for one small colony (approximately 2 m2) observed during the transects, where the rocks were completely covered by vegetation. Additionally, we identified 32 ‘new’ colonies in the digital imagery which were not assessable on the ground and could, therefore, only be confirmed in the imagery. All colonies were then manually measured in the imagery. To ensure imagery measurements were accurate for calculating the total colony area, the Boldró Colony was manually measured and compared to its digital measurements.
Calculation of total population size
We calculated the rock cavy density estimates for the Boldró colony by dividing the average number of individuals calculated by the ZPNE by its total area (651 m2). These calculated densities were then multiplied by the total extent of each identified colony. The sum of all population sizes resulted in the total number of rock cavies on the archipelago. The CI95% for the entire population was calculated using a similar approach, replacing mean estimated population size by the lower and the upper confidence bounds.
Results
Faeces, herbivory and/or rock cavy individuals were observed in 100% of the surveyed colonies. Although this might not be common for the cave-dwelling rodents, all indirect signs of activity observed were recent. Population parameters’ estimates generated by the ZPNE for the Boldró rock cavy colony are presented in Table 2. Across both field trips, the population remained relatively stable, while the model estimated survival probability confidence interval at 95% as between 0 and 1. Mean capture probability by observation was estimated at 0.858 (CI95% 0.597 – 0.961). The mean density estimate for the Boldró colony was 0.0316 individuals/m2. Minimum and maximum densities found on the island were 0.018 individuals/m2 and 0.0555 individuals/m2, respectively. The total population size was calculated as 5,473 individuals (CI95% 3,114 – 9,622). Finally, the digital measurement of the Boldró colony presented less than 6% difference in area size (i.e., smaller) in comparison to the manual measurement. Although we only managed to measure one colony, which could be improved in the future to increase reliability, this result suggests that the digital measure might be considered accurate enough to be used in this study.
Discussion
The present study assessed the distribution and estimated population parameters of the exotic rock cavy on the island of Fernando de Noronha, Brazil, for the first time. We calculated rock cavy colony density per m2, and the total abundance of individuals for the main island of Fernando de Noronha. We also identified strengths and weaknesses of applying the ZPNE model to our rock cavy sparse dataset. Despite the challenges involved in the calculation of total population size from one to all colonies on the island, our study provides unprecedented information on the insular rock cavy population. Moreover, our results provide a warning case on the establishment of (native) species in new habitats and could support the planning of management actions for this and other exotic species.
The rock cavy distribution and population structure
According to Alves and Leite [11], within 23 years the population of rock cavies was perceived as spread across the whole main island of Fernando de Noronha. This conference abstract is, however, anecdotal; it only presents descriptive data in regards to the species’ abundance on different colonies on the island. The present study confirmed rock cavy presence for all of the 68 surveyed colonies, which scientifically supports this anecdote. Neither rock cavies, its latrines, or its presence signs were observed outside a colony’s surroundings. This corroborates ecological observations on the species in the continent, where the rock cavy is believed to be very restricted to this type of habitat [12][24][26][27].
Despite the availability of few captive studies [13][17][24][30][31][32][33], only three have generated some sort of in situ information on this specie’s ecology [12][26][35], all in the continent. There is a complete gap of knowledge regarding this alien species on the on Fernando de Noronha. Captive studies, although rarely comparable to in situ ones, might improve reliability of estimates by setting maximum thresholds to these. Considering the rock cavy, captive populations to which nests, water and food were provided ad libitum, did not endure more than 0.09 individuals per m2 [13]. In a recent longer-term in situ study (i.e., in the continent) mean rock cavy density was found to be 0.0012 (CI95% 0.0007 – 0.0022) individuals/m2 [12].
The density calculated using the the ZPNE was 35% smaller than the estimated maximum density observed by the captive study and 26 times larger than the mean estimate calculated by the in situ study. This demonstrates that the rock cavy density on the island is considerably higher in comparison to the continent, and close, although still lower, than the maximum density tolerated in captivity. The last fact suggests further that the island population might be at carrying capacity. Higher density than in situ studies can be explained, by the marked wet season – which results in abundance of resources for the rock cavy – and decreased poaching and predation pressures on the island than in the continent. On the other hand, lower density than ex situ ones is easily explained by the lack of direct subsidies and existing poaching and predation of individuals in situ. Moreover, this demonstrated that our estimates are ecologically sound. It is important, however, to acknowledge that the current effort only used one point of validation for the measurement of the colonies (i.e., in situ versus digital imagery), which could introduce uncertainties in the calculated densities and should be revised by future works.
Extrapolation of results from one colony to the entire island
Extrapolation of estimates and predictions is a considerable challenge to scientists, but it has become a sine qua non in ecological studies in the last decades [60]. Although there are important advantages on extrapolating data from a colony to the whole population, there are also disadvantages. Both are well described by Miller et al. [60]. As a general rule, the more severe the extrapolation, the lower the reliability of the predictions. In a context where precision is of high importance, for example when planning management of highly invasive species in favorable environments, blindly trusting information that presents low uncertainty can lead to failures in accomplishing eradications [61]. Invasive rodent eradication failures, for example, happen at least two times more often on tropical than on temperate islands, in part due to lower certainty of population parameters and their interactions with the tropical island environment [62].
On the other hand, information with high uncertainty is still better than no information and can be used as an abundance index to be monitored by further studies. Even if highly uncertain, an estimate of abundance might provide local managers and decision makers with a comprehensive context of the target species’ situation, allowing for initial management feasibility discussions. Invasive species’ management, for example, can be planned targeting the upper populations’ abundance threshold, instead of its average. This increases the financial burden of a management plan, but is a possible trade-off to improve the chances of success. In the present case, direct scaling up the rock cavy density from a colony to the whole island is believed to be sufficient at this time due to (1) the small sized island and the (2) urgent need of information on population abundance. The information generated by the present work can be considered ecologically sound and, with caution, may be used to simulate management interventions for the rock cavy. Moreover, it may provide managers with enough information about parameters that need improved precision and those that may be sufficient for management decisions.
The ZPNE model used and its application to a sparse dataset
The zero-truncated poisson log-normal mixed effects model (ZPNE) has been extensively used and proven to provide good population estimates. Still, when it comes to using it with sparse datasets for developing management plans, generated estimates should be used cautiously. Our model, although the best option for a sparse dataset available off-the-shelf, presented two main weaknesses: (1) an uninformative confidence interval for the survival probability, and (2) a lower estimate for population size than the minimum number of individuals known to be present. Regarding the first weakness, although it is not possible directly assessing bias for real data, extreme values on standard errors of parameter’s estimates might provide information regarding its accuracy. The ZPNE presented an uninformative confidence interval for survival (Table 2), due to its considerably large standard deviation. Still, even though uninformative confidence interval might indicate that some model assumptions were violated [63][64] a large standard error not necessarily mean that estimates are biased. Regarding the second weakness, the ZPNE estimated a lower confidence bound than the minimum number of individuals known to be alive, which is the known number of marked individuals summed by the total non-marked individuals counted. On the other hand, recapture probability’s uncertainty was calculated to be within a reasonable interval, especially for a sparse dataset. This parameter suggested that recapture by observation is highly efficient for this species. Moreover, as previously noted, our estimates seem to be ecologically sound. However, even though we were able to estimate urgently needed population parameters of the Fernando de Noronha’s rock cavy, we advise caution when using these parameter estimates for planning rock cavy management due to its high level of uncertainty.
Rock cavy management considerations
While the precautionary principle is valuable for managing known invasive species, especially during early establishment, addressing well-established populations such as the rock cavy require substantial resources and comprehensive management plans [57] prevention of its spread to non-affected areas (e.g., sites, regions, and cross-continent. While the rock cavy population has established itself beyond geographical barriers [3][4], it is critical to note that there have not been any formal impact assessments for the species.
Biological studies should, for example, confirm the specie’s potential impact on native flora and facilitation of invasive plant species. Concerns about landscape modification through rock erosion [11][58] also warrant scrutiny. Health impact assessments, including investigation of potential Chagas disease transmission [59], would also provide valuable data for informed management decisions. The priority moving forward should be a comprehensive, scientific assessment of the rock cavy’s ecological impact on the main island of the archipelago. This evaluation would enable evidence-based allocation of conservation resources, ensuring that management efforts target the most significant threats to the archipelago’s ecosystem.
Conclusion
Our observations show that the rock cavy is distributed across the entire island, though confined to rocky outcrops and their surroundings, similar to its habitat on the mainland. We estimated survival probability and density within one colony and used this information to calculate the total abundance of individuals on Fernando de Noronha. Future studies could enhance abundance and survival estimates, as well as recruitment rates, by developing tailored models for this species – models that can better accommodate sparse data. As demonstrated here, observational recapture is highly effective and could serve as a cornerstone for such customized models, especially for species that are difficult to trap. Although the information on population abundance could help managers understand the potential implications of this species on the archipelago, the ecological impacts of the rock cavy on the island require scientific assessment, as the population seems well established and has an important colonization potential. Importantly, we generated unprecedented information on the rock cavy population on Fernando de Noronha. While this data should be interpreted with caution due to the high uncertainty, it may nonetheless provide local managers with a foundation for initial management feasibility discussions.
Acknowledgements
CPNq (Brazilian National Council for Scientific and Technological Development) for funding the first author (246171/2012-0), Rufford Foundation (Project 15896-1), PVE/CAPES/USP (Coordination of Improvement of Higher-Level Personnel, 88881.0304552013-01), Graduate Academy of TU Dresden and the Friends of TU Dresden for the funding of field trips and presentations. TM was supported by NWT Cumulative Impact Monitoring Program (NWT CIMP, Project No. CIMP220) and by Deutsche Forschungsgemeinschaft (DFG, Project No. 529743012) during the development of this manuscript.
References
1. Richardson DM, Pyšek P, Carlton JT. A Compendium of essential concepts and terminology in invasion ecology. In: Richardson DM (ed.). Fifty years of invasion ecology: The legacy of Charles Elton. 1st ed., Hoboken, NJ, USA: Blackwell Publishing Ltd; 2011, p. 409-420.
2. Davis MA, Chew MK, Hobbs RJ, Lugo AE, Ewel JJ, Vermeij GJ, et al. Don’t judge species on their origins. Nature 2011; 474: 153-154. doi: 10.1038/474153a.
3. Richardson DM, Pysek P, Rejmanek M, Barbour MG, Panetta FD, West CJ. Naturalization and invasion of alien plants: concepts and definitions. Diversity and Distributions 2000; 6: 93-107. doi: 10.1046/j.1472-4642.2000.00083.x.
4. Blackburn TM, Pyšek P, Bacher S, Carlton JT, Duncan RP, Jarošík V, et al. A proposed unified framework for biological invasions. Trends in Ecology & Evolution 2011; 26: 333-339. doi: 10.1016/j.tree.2011.03.023.
5. Simberloff D, Martin J-L, Genovesi P, Maris V, Wardle DA, Aronson J, et al. Impacts of biological invasions: what’s what and the way forward. Trends in Ecology & Evolution 2013; 28: 58-66. doi: 10.1016/j.tree.2012.07.013.
6. Parker IM, Simberloff D, Lonsdale WM, Goodell K, Wonham M, Kareiva PM, et al. Impact: Toward a framework for understanding the ecological effects of invaders. Biological Invasions, 1999; 1: 3-19. doi: 10.1023/A:1010034312781.
7. Stephenson PJ, Ntiamoa-Baidu Y, Simaika JP. The use of traditional and modern tools for monitoring wetlands biodiversity in Africa: Challenges and opportunities. Front Environ Sci, 2020; 8: 61. doi: 10.3389/fenvs.2020.00061.
8. Baker CM, Bode M. Recent advances of quantitative modeling to support invasive species eradication on islands. Conservat Sci and Prac, 2021; 3. doi: 10.1111/csp2.246.
9. Schulz-Neto A. Observando aves no Parque Nacional Marinho de Fernando de Noronha: guía de campo. Brasília/DF: Instituto Brasileiro do Meio Ambiente e dos Recursos Naturais Renováveis; 1995.
10. Oren D. Resultados de uma nova expedição zoológica a Fernando de Noronha. Boletim do Museu Paraense Emílio Goeldi, 1984; Zoologia: 19-44.
11. Alves ML, Leite LMRM. Distribuição geográfica do Kerodon rupestris (Mammalia, Rodentia) em Fernando de Noronha/PE, Brasil. In: Overal W (ed.). Resumos XII Congresso Latino-Americano de Zoologia / XIX Congresso Brasileiro de Zoologia, Belem: Sociedade Brasileira de Zoologia; 1992, p. 156.
12. Conceição AM, Bocchiglieri A. Population density and use of space by Kerodon rupestris: An endemic and threatened rodent in the semiarid areas of Brazil. Journal of Arid Environments, 2021; 186: 104425. doi: 10.1016/j.jaridenv.2020.104425.
13. Lacher T. Rates of growth in Kerodon rupestris and an assessment of its potential as a domesticated food source. Papéis Avulsos de Zoologia (São Paulo), 1979; 33: 67-76.
14. Zappes I, Portella A, Lessa G. Description of Karyotype of Kerodon acrobata, an endemic rodent in Brazilian Cerrado. Braz J Biol. 2014; 74: 251-256. doi: 10.1590/1519-6984.23512.
15. Lessa G, Gonçalves PR, Pessôa LM. Variação geográfica em caracteres cranianos quantitativos de Kerodon rupestris (WIED, 1820) (RODENTIA, CAVIIDAE). Arquivos do Museu Nacional, 2005; 63: 75-88.
16. Freitas RR, Rocha PLB da, Simões-Lopes PC. Habitat structure and small mammals abundances in one semiarid landscape in the Brazilian Caatinga. Rev Bras Zool. 2005; 22: 119-129. doi: 10.1590/S0101-81752005000100015.
17. Sousa RA, Menezes AAL. Circadian rhythm of motor activity of the Brazilian rock cavy (Kerodon rupestris) under artificial photoperiod. Biological Rhythm Research, 2006; 37: 443-450. doi: 10.1080/09291010600869836.
18. Xavier SCC, Vaz VC, D’Andrea PS, Herrera L, Emperaire L, Alves JR, et al. Mapping of the distribution of Trypanosoma cruzi infection among small wild mammals in a conservation unit and its surroundings (Northeast-Brazil). Parasitology International, 2007; 56: 119-128. doi: 10.1016/j.parint.2007.01.003.
19. Delciellos AC. Mammals of four Caatinga areas in northeastern Brazil: inventory, species biology, and community structure. Cl, 2016; 12: 1916. doi: 10.15560/12.3.1916.
20. UNESCO. Brazilian Atlantic Islands: Fernando de Noronha and Atol das Rocas Reserves 2001.
21. Serafini TZ, França GB de, Andriguetto-Filho JM. Ilhas oceânicas brasileiras: biodiversidade conhecida e sua relação com o histórico de uso e ocupação humana. RGCI, 2010; 10: 281-301. doi: 10.5894/rgci178.
22. Mello TJ, Oliveira AAD. Making a bad situation worse: An invasive species altering the balance of interactions between local species. PLoS ONE, 2016; 11: e0152070. doi: 10.1371/journal.pone.0152070.
23. Barcelos D, Vieira EM, Pinheiro MS, Ferreira GB. A before-after assessment of the response of mammals to tourism in a Brazilian national park. Oryx, 2022; 56: 854-863. doi: 10.1017/S0030605321001472.
24. Mares M, Lacher TE. Ecological, morphological and behavioral convergence in rock-dwelling mammals. Current Mammalogy, 1987; 1: 307-348.
25. Oliveira FG, Nascimento-Júnior ESD, Cavalcante JC, Guzen FP, Cavalcante JDS, Soares JG, et al. Topographic specializations of catecholaminergic cells and ganglion cells and distribution of calcium binding proteins in the crepuscular rock cavy (Kerodon rupestris) retina. Journal of Chemical Neuroanatomy, 2018; 90: 57-69. doi: 10.1016/j.jchemneu.2017.12.007.
26. Lacher T. The comparative social behavior of Kerodon rupestris and Galea spixii and the evolution of behavior in the Caviidae. vol. 17. Pittsburgh: Carnegie Museum of Natural History; 1981.
27. Nutt K. Socioecology of rock-dwelling rodents. In: Wolff J, Sherman PW (ed.). Rodent societies: an ecological and evolutionary perspective, Chicago and London: University of Chicago Press; 2007, p. 416-426.
28. IUCN. The IUCN Red List of Threatened Species 2016.
29. Oliveira J, Bonvicino C. Ordem rodentia. Mamíferos do Brasil, Londrina, Brazil: Universidade Estadual de Londrina; 2006, p. 347-406.
30. Lacher TE, Willig MR, Mares MA. Food preference as a function of resource abundance with multiple prey types: An experimental analysis of optimal foraging theory. The American Naturalist, 1982; 120: 297-316.
31. Roberts M, Maliniak E, Deal M. The reproductive biology of the rock cavy, Kerodon rupestris, in captivity: A study of reproductive adaptation in a trophic specialist. Mammalia, 1984; 48. doi: 10.1515/mamm.1984.48.2.253.
32. Tasse J. Maternal and paternal care in the rock cavy, Kerodon rupestris, a South American hystricomorph rodent. Zoo Biol. 1986; 5: 27-43. doi: 10.1002/zoo.1430050105.
33. Willig MR, Lacher TE. Food selection of a tropical mammalian folivore in relation to leaf-nutrient content. Journal of Mammalogy, 1991; 72: 314-321.
34. De Matos Dias D, de Campos CB, Guimarães Rodrigues FH. Behavioural ecology in a predator-prey system. Mammalian Biology, 2018; 92: 30-36. doi: 10.1016/j.mambio.2018.04.005.
35. Conceição AM, Bocchiglieri A. Temporal variation in the diet of the endemic and threatened rodent Kerodon rupestris in the semiarid area of Brazil. Mammalia, 2021; 85: 537-540. doi: 10.1515/mammalia-2021-0043.
36. da Silva Neto EJ. Morphology of the regiones ethmoidalis and orbitotemporalis in Galea musteloides Meyen 1832 and Kerodon rupestris (Wied-Neuwied 1820) (Rodentia: Caviidae) with comments on the phylogenetic systematics of the Caviidae. Journal of Zoological Systematics and Evolutionary Research, 2000; 38: 219-229. doi: 10.1046/j.1439-0469.2000.384135.x.
37. Oliveira M, Carter A, Bonatelli M, Ambrosio CE, Miglino MA. Placentation in the rock cavy, Kerodon rupestris (Wied). Placenta, 2006; 27: 87-97. doi: 10.1016/j.placenta.2004.11.012.
38. Nascimento ES, Souza APM, Duarte RB, Magalhães MAF, Silva SF, Cavalcante JC, et al. The suprachiasmatic nucleus and the intergeniculate leaflet in the rock cavy (Kerodon rupestris): Retinal projections and immunohistochemical characterization. Brain Research, 2010; 1320: 34-46. doi: 10.1016/j.brainres.2010.01.034.
39. Soares JG, Cavalcanti JRLP, Oliveira FG, Pontes ALB, Sousa TB, Freitas LM, et al. Nuclear organization of the serotonergic system in the brain of the rock cavy (Kerodon rupestris). Journal of Chemical Neuroanatomy, 2012; 43: 112-119. doi: 10.1016/j.jchemneu.2012.03.001.
40. Cavalcanti JRLP, Soares JG, Oliveira FG, Guzen FP, Pontes ALB, Sousa TB, et al. A cytoarchitectonic and TH-immunohistochemistry characterization of the dopamine cell groups in the substantia nigra, ventral tegmental area and retrorubral field in the rock cavy (Kerodon rupestris). Journal of Chemical Neuroanatomy, 2014; 55: 58-66. doi: 10.1016/j.jchemneu.2014.01.002.
41. de Medeiros Silva A, de Santana MAD, de Góis Morais PLA, de Sousa TB, Januário Engelberth RCG, de Souza Lucena EE, et al. Serotonergic fibers distribution in the midline and intralaminar thalamic nuclei in the rock cavy (Kerodon rupestris). Brain Research, 2014; 1586: 99-108. doi: 10.1016/j.brainres.2014.08.047.
42. Santos AC dos, Aro MM de, Bertassoli BM, Viana DC, Vasconcelos BG, Rici REG, et al. Morphological characteristics of the tongue of the rock cavy- Kerodon rupestris wied, 1820 (Rodentia, caviidae). Bioscience Journal, 2015; 31: 1174-1182. doi: 10.14393/BJ-v31n4a2015-26061.
43. Santos F, Magalhaes-Junior JT, de Oliveira Carneiro I, Lambert SM, da Silva Souza BMP, de Pauda AD, et al. Wild mammals involved in the transmission of Trypanosoma cruzi and food sources of Triatoma sherlocki in an endemic region of northeastern Brazil. Medical Vet Entomology, 2023; 37: 396-406. doi: 10.1111/mve.12641.
44. Almeida CE, Faucher L, Lavina M, Costa J, Harry M. Molecular individual-based approach on Triatoma brasiliensis: Inferences on Triatomine Foci, Trypanosoma cruzi Natural Infection Prevalence, Parasite Diversity and Feeding Sources. PLoS Negl Trop Dis, 2016; 10: e0004447. doi: 10.1371/journal.pntd.0004447.
45. Labruna MB, Nava S, Marcili A, Barbieri ARM, Nunes PH, Horta MC, et al. A new argasid tick species (Acari: Argasidae) associated with the rock cavy, Kerodon rupestris Wied-Neuwied (Rodentia: Caviidae), in a semiarid region of Brazil. Parasites Vectors, 2016; 9: 511. doi: 10.1186/s13071-016-1796-7.
46. Resende NR, Soares Filho PL, Peixoto PPA, Silva AM, Silva SF, Soares JG, et al. Nuclear organization and morphology of cholinergic neurons in the brain of the rock cavy (Kerodon rupestris) (Wied, 1820). Journal of Chemical Neuroanatomy, 2018; 94: 63-74. doi: 10.1016/j.jchemneu.2018.09.001.
47. Fundação Oswaldo Cruz, Souza MVD, Chaves SADM, Hugot J-P, Iñiguez AM. New parasite records from Kerodon rupestris (Rodentia, Caviidae) an endemic species to northeastern Brazil. Oecol Ast, 2020; 24: 196-203. doi: 10.4257/oeco.2020.2401.18.
48. Lima-Neiva V, Toma HK, Abrantes Aguiar LM, Lopes CM, Dias LP, Monte Gonçalves TC, et al. The connection between Trypanosoma cruzi transmission cycles by Triatoma brasiliensis brasiliensis: A threat to human health in an area susceptible to desertification in the Seridó, Rio Grande do Norte, Brazil. PLoS Negl Trop Dis, 2021; 15: e0009919. doi: 10.1371/journal.pntd.0009919.
49. Souza MV, Chaves SAM, Iñiguez AM. New observations from the intestinal fauna of Kerodon rupestris (Wied, 1820) (Rodentia, Cavidae), Brazil: a checklist spanning 30,000 years of parasitism. Braz J Biol. 2021; 81: 989-998. doi: 10.1590/1519-6984.232838.
50. Saldanha BM, Chame M, Nunes GKM, Sianto L, Leles D. Parasites of the Brazilian Rock Cavy, Kerodon rupestris: Revealing Their History in the Brazilian Semiarid Region. Journal of Parasitology 2022;108:395–402. doi: 10.1645/20-56.
51. Hutto RL, Pletschet SM, Hendricks P. A Fixed-radius Point Count Method for Nonbreeding and Breeding Season Use. The Auk 1986; 103: 593-602. doi: 10.1093/auk/103.3.593.
52. McClintock BT, White GC, Antolin MF, Tripp DW. Estimating abundance using mark-resight when sampling is with replacement or the number of marked individuals is unknown. Biometrics, 2009; 65: 237-246. doi: 10.1111/j.1541-0420.2008.01047.x.
53, McClintock BT, White GC. From NOREMARK to MARK: software for estimating demographic parameters using mark–resight methodology. Journal of Ornithology, 2012; 152: 641-650. doi: 10.1007/s10336-010-0524-x.
54. Adrian O, Sachser N. Diversity of social and mating systems in cavies: a review. Journal of Mammalogy, 2011; 92: 39-53. doi: 10.1644/09-MAMM-S-405.1.
55. Gaiotto JV, Abrahão CR, Dias RA, Bugoni L. Diet of invasive cats, rats and tegu lizards reveals impact over threatened species in a tropical island. Perspectives in Ecology and Conservation, 2020; 18: 294-303. doi: 10.1016/j.pecon.2020.09.005.
56. Quantum GIS Development Team. Development Team, 2012. Quantum GIS Geographic Information System. Open Source Geospatial Foundation Project. Free Software Foundation, India 2015.
57. Poland TM, Juzwik J, Rowley A, Huebner CD, Kilgo JC, Lopez VM, et al. Management of Landscapes for Established Invasive Species. In: Poland TM, Patel-Weynand T, Finch DM, Miniat CF, Hayes DC, Lopez VM (eds.). Invasive species in forests and rangelands of the United States, Cham: Springer International Publishing; 2021, p. 133-184. doi: 10.1007/978-3-030-45367-1_7.
58. da Rosa CA, de Almeida Curi NH, Puertas F, Passamani M. Alien terrestrial mammals in Brazil: current status and management. Biological Invasions, 2017. doi: 10.1007/s10530-017-1423-3.
59. Lilioso M, Reigada C, Pires-Silva D, Fontes FVHM, Limeira C, Monsalve-Lara J, et al. Dynamics of food sources, ecotypic distribution and Trypanosoma cruzi infection in Triatoma brasiliensis from the northeast of Brazil. PLoS Negl Trop Dis 2020; 14: e0008735. doi: 10.1371/journal.pntd.0008735.
60. Miller Jr, Turner M, Smithwick EAH, Dent CL, Stanley EH. Spatial Extrapolation: The Science of Predicting Ecological Patterns and Processes. BioScience. 2004 Apr 1; 54(4): 310-320. https://doi.org/10.1641/0006-3568(2004)054[0310:SETSOP]2.0.CO;2
61. Seymour A, Varnham K, Roy S, Harris S, Bhageerutty L, Church S, et al. Mechanisms underlying the failure of an attempt to eradicate the invasive Asian musk shrew Suncus murinus from an island nature reserve. Biological Conservation. 2005 Sep; 125(1): 23-35. https://doi.org/10.1016/j.biocon.2005.03.005
62. Russell JC, Holmes ND. Tropical island conservation: Rat eradication for species recovery. Biological Conservation. 2015 May; 185: 1-7. https://doi.org/10.1016/j.biocon.2015.01.009
63. Breslow N. Tests of Hypotheses in Overdispersed Poisson Regression and other Quasi-Likelihood Models. Journal of the American Statistical Association. 1990 Jun; 85(410): 565-571. https://doi.org/10.1080/01621459.1990.10476236
64. Schmutz JA, Ward DH, Sedinger JS, Rexstad EA. Survival estimation and the effects of dependency among animals. Journal of Applied Statistics. 1995 Nov; 22(5-6): 673-682. https://doi.org/10.1080/02664769524531
