Retrospective and current trend of wild‐cat trade in Peru

Several species of wild cats are threatened with extinction due to habitat loss, persecution or retaliatory killing by humans as a result of real or perceived livestock depredation, and illegal trade. The trade of individuals or their parts has been a recurring threat over the years, especially prior to the establishment of CITES (Convention on International Trade in Endangered Species of Wild Fauna and Flora) in 1975. We assessed the extent of trade of wild cats in Peru (eight species) using export records available from 1946 to 1973, before the establishment of CITES and after the implementation of CITES using confiscation data available for 2001–2020. The ocelot and the jaguar were the most exploited species for their skins in the pre‐CITES period, with 228,376 and 17,301 individuals, respectively; as well as post‐CITES, although at significantly lower levels, with 67 individuals and 107 body parts, and 27 individuals and 99 body parts, respectively. Post‐CITES trade, however, shows an increasing trend for the jaguar and all wild cat species. Currently in Peru, the illegal wildlife trade is considered opportunistic, but its impact on wild populations has not been properly documented. We recommend improving knowledge regarding population status of wild cats to inform conservation status, and to increase efforts to reduce illegal trade at both national and regional level.

Wild feline populations, especially those of large-bodied species (≥15 kg), have been decreasing in recent years, in some cases with up to 95% of range contraction (e.g., tigers or lions; Ripple et al., 2014;Wolf & Ripple, 2017). This decrease is associated with threats such as the loss and degradation of habitat, persecution or retaliatory killing by humans as a result of real or perceived livestock depredation, human use (e.g., traditional medicine, skins, and trophies), and population declines of their natural prey species (Ripple et al., 2014).
A recurring threat to wild cats is the illegal trade in body parts. This is a primary cause of population decline for tigers (Panthera tigris) due to the demand from China, South Korea, and other Asian countries for body parts and derivatives for traditional medicine and decorative purposes, especially high during the 1980s and 1990s (Dinerstein et al., 2007;Jackson, 2010;Loveridge, Wang, Frank, & Seidensticker, 2010). Currently, this threat has been recognized as a major concern for lions and other large cats (Everatt, Kokes, & Lopez Pereira, 2019;Loveridge et al., 2010;Williams, Loveridge, Newton, & Macdonald, 2017). The trafficking of body parts has also been recognized as a conservation problem for the jaguar, the largest feline in the Americas (Morcatty et al., 2020).
Several spotted-cat species populations in Latin America, including the ocelot (Leopardus pardalis), margay (Leopardus wiedii), and jaguar (Panthera onca) have been impacted due to the international trade of skins, especially for the fashion industry in Europe and North America in the 1960s and 1970s (Loveridge et al., 2010). This trade resulted in an annual removal of~80,000 ocelots and~15,000 jaguars in the Brazilian Amazon during 1960s (Smith, 1976), and up to~800,000 ocelots and margays, and~180,000 jaguars between 1904 and 1969 (Antunes et al., 2016). That level of overexploitation of Neotropical wild cats motivated their listing in Appendix I of the Convention on International Trade in Endangered Species (CITES) in 1975, prohibiting international commercial trade. In 1986 the European Union banned all imports of body parts of Latin American cats, while several smaller cat species were listed as Appendix I between 1989 and 1992 (Loveridge et al., 2010). Due to trade controls at both national and international levels and a resultant decreasing demand of spotted cat skins, trade in cat products decreased (Broad, 1987b). Furthermore, Graham (2017) reviewed the CITES trade database  and found that the ocelot trade decreased after the legal fur-trade period. Confiscations during that time included live wild born and captive-bred individuals as well as body parts. Nevertheless, illegal trade in ocelot skins and other body parts still persists (Graham, 2017). The same situation applies to jaguars (Morcatty et al., 2020). In fact, both species show decreasing population trends, mainly due to habitat loss, persecution, and illegal trade of body parts (de la Torre, Gonz alez-Maya, Zarza, Ceballos, & Medellín, 2017;Paviolo et al., 2015).
Peru has a large diversity of felines, with eight species (Felidae) , one of them considered threatened (Leopardus jacobita), two are near threatened (Panthera onca and Puma concolor) and three are data deficient (Leopardus garleppi, Leopardus tigrinus, and Leopardus wiedii) (SERFOR, 2018). In 1970 the Peruvian government banned trade and exportation of jaguar skins (Ministerial Resolution N 5056-70-AG) and in 1973 declared an indefinite prohibition on Amazonian wildlife trade (Veda de Caza, 934-73-AG), including spotted-cats (Broad, 1987a(Broad, , 1987b. In 2017 the Peruvian government published the national strategy to reduce illegal wildlife trafficking (SERFOR, 2017), whose implementation is hoped to reduce illegal trade. Peru also has the secondlargest population of jaguars (~22,200 individuals) after Brazil (Jędrzejewski et al., 2018). Unfortunately, wild felids have been poorly studied in Peru and large knowledge gaps exist regarding population ecology, distribution, and conservation issues, among others (Cossíos et al., 2012), making an accurate assessment of threats to these species difficult.
In this paper, we assessed trends in trade of spotted cats in Peru prior to the implementation of CITES. We reviewed records on legal exports from the city of Iquitos from 1946 to 1973 available in Grimwood (1969) and Broad (1987a). These records provided the number of skins for spotted cat species exported per year during that period. Export records from Iquitos were utilized because it was the main export location for wildlife skins during that time (Broad, 1987a;Grimwood, 1969). In addition to reviewing these legal exports, we assessed trends for all wild cat body-parts and living individuals (hereafter referred as specimens) seized between 2001 and 2020, using data provided by national authorities such as the National Forest and Wildlife Service (SERFOR) and Regional Governments of Loreto, Ucayali, and Madre de Dios (GOREs). Following Heinrich et al. (2020), we considered specimens seized as examples of illegal wildlife trade. Specimen confiscations were generally planned, and intelligence driven, but also sometimes complaintdriven and occurred in markets, airports, bus station terminals, and custom posts. Using both past legal export data and more recent seizures of illegally traded specimens allow for an updated assessment on the current illegal trade of wild cats in Peru within the context of past population impacts due to historical trade. Thus, we believe that this approach allows us to better assess the effect of current threats (e.g., habitat loss, illegal trade, and human-wildlife conflicts) on previously impacted populations.
We used generalized additive models (GAM) to estimate trends over time for the number of exported spotted-cat skins from 1946 to 1973 (pre-CITES period). GAM is a kind of general linear model (GLM) where the linear predictor depends linearly on smooth functions of predictor variables and also allows the potential temporal autocorrelation to be incorporated into the structure of the models (Wood, 2017). The number of exported spotted-cat skins by each species was used as the response variable: spotted-cat skins = f Year ð Þ, where year was the predictor variable, modeled using the GAM framework to fit a smooth function (f ). We treated time (Year) as a continuous variable, which allows annual patterns to be visualized more easily (Heinrich et al., 2020). The model was fitted using a log link function under a Poisson distribution (Model 1) and a negative binomial variance function (Model 2), which are recommended when there are over-dispersion problems in the data (Zuur, Ieno, & Elphick, 2010). We fitted GAM models for the species with the highest number of traded skins (ocelots and jaguars).
In addition, we fitted GAM models to estimate the temporal trend in the total number of wild cats seized from 2001 to 2020. We also ran GAM models to analyze trends for the two species with the highest number of specimens seized (ocelot and jaguar). Skins, skulls, teeth, legs, and claws were considered body parts (specimens) because records did not identify these to individuals. We considered the number of specimens by year as the variable response "number of wild cat specimens seized". As the total number of confiscation incidents ("confiscation effort") was not evenly distributed across the years, it was included as an offset in modeling (Zuur, Hilbe, & Ieno, 2013). The average number of wild cat specimens seized by each confiscation incident was 1.5 ± 0.58. The comparison of the models was based on the Akaike information criterion corrected for small samples (AICc), and subsequent inferences were made from the model with the lowest AICc value when the difference with the next best model was ΔAICc <2 (Burnham & Anderson, 2002). All data were analyzed in R (R Core Team, 2020). The models were run using the "mgcv" package (Wood, 2011) and the selection of the models was made using the "MuMIn" package (Barton, 2020).
Based on current density data for jaguars in Peru, 3.00 (2.41 ± 3.62) individuals/100 km 2 as suggested by Jędrzejewski et al. (2018), the annual loss of~618 jaguars prior to the establishment of CITES implied the annual local extirpation of jaguar individuals in an approximate area of 20,167 (15,900-24,433) km 2 . This area is the equivalent in size to protected areas such as the Pacaya Samiria National Reserve or the Sierra del Divisor National Park. Grimwood (1969) pointed out that during the 60s, jaguars disappeared from areas surrounding human settlements and were considered rare throughout the Peruvian amazon. Due to the jaguar's biological characteristics, including slow reproduction rate and long generation time van de Kerk, de Kroon, Conde, & Jongejans, 2013), it would seem that this level of exploitation was unsustainable. Nevertheless, Antunes et al. (2016) considered that a refuge-harvestable area model (i.e., source-sink dynamics) was the only reasonable way to explain increasing or continued commercial-scale harvest for Amazonian wildlife species in Brazil, including jaguars and ocelots. It is possible that F I G U R E 1 The fitted smoother curves obtained with GAM depicting the effect of time (years) on commercial harvest (skins) by both ocelot and jaguar (solid line) and 95% confidence interval (shaded) based on data from Grimwood (1969) and (Broad, 1987a). The grey dotted line indicates the year of the hunting band (Veda de Caza, 934-73-AG) jaguar dispersal patterns may have allowed them to recolonize previously harvested territories. Studies have shown that jaguar dispersal is mostly male-biased, while females are more philopatric, remaining close to each other Lorenzana et al., 2020). In addition, jaguar home-ranges are estimated at 35-2915 km 2 for males and 13-1155 km 2 for females, and are related to habitat productivity, forest cover and road density (Morato et al., 2016;Thompson et al., 2021). These biological characteristics, in combination with the availability of unexploited and undisturbed habitats, possibly allowed for the existence of source-sink dynamics and prevented a population collapse during pre-CITES levels of overexploitation and explains the current persistence of jaguars in previously hunted areas. Likewise, the high dispersal potential of jaguars has also probably resulted in high levels of gene flow among populations. In that sense, Lorenzana et al. (2020) showed no evidence of recent genetic bottlenecks for Amazonian populations; instead, jaguars show high levels of genetic diversity. As Antunes et al. (2016) pointed out, if hunting of wild cat populations in Brazil during the trade period included adequate source populations, felines may be resilient since high-levels of commercial harvest failed to extirpate them during that period. This likely also occurred for jaguars and ocelots in Peru during the pre-CITES period, and also the 2001-2020 period. Broad (1987b) pointed out that CITES annual reports showed that the number of cats skins traded in Peru was drastically reduced (zero in 1985). However, illegal trade in wild cat body parts, although not at pre-CITES levels, seems to be widespread in several locations throughout the Peruvian Amazon (Graham, 2017;Morcatty et al., 2020;Shanee, 2012). In Peru, the wild cat trade for medicinal and mystical utilization is considered opportunistic and, although illegal, it is not recognized as a criminal act due to cultural legacy reasons (Arias et al., 2021;Braczkowski et al., 2019;Leberatto, 2016;Leberatto, 2017). A recent study showed evidence of illegal trade of jaguar body parts in the main cities of the Peruvian Amazon, where skins and teeth are sold as handicrafts in markets (SERFOR and WCS, 2020) (Figure 3), with Iquitos (Loreto) being the most important city for this type of trade. These findings are in agreement with the data regarding specimen seizures, which showed the Amazonian departments of Loreto and Ucayali as having the highest number of seizures after Lima, which is the main departure location for international destinations.
The trade of tiger body parts, mainly associated with medicinal attributes, has had repercussions for other big cats as tiger populations have decreased and regulation of tiger trade has increased. Species such as leopards, lions, and jaguars may be considered as tiger substitutes (Morcatty et al., 2020;Villalva & Moracho, 2019).
The fitted smoother curves obtained with GAM depicting the effect of time (years) on all wild cats, ocelot, and jaguar seized (solid line) and 95% confidence interval (shaded) The grey dotted line indicates the year of publication of the national strategy to reduce illegal wildlife trafficking (2017) Fortunately, in Peru, there does not yet seem to be a relationship of illegal trade with the demand for wild cat parts for the Asian market, as it is for other Latin American countries (Morcatty et al., 2020). However, data on the relationship between the Asian markets and illegal wildlife trade in Peru is lacking. The recent increase in jaguars and other species of wild cats seized is a wake-up call to increase measures to determine the drivers of the illegal wildlife trade and reduce wildlife trafficking. As 't Sas-Rolfes, Challender, Hinsley, Veríssimo, and Milner-Gulland (2019) has pointed out, seizure data at the national level are subject to the enforcement or reporting capacity, so the extent of the illegal wildlife trade is probably underestimated.
Currently, jaguar populations are being strongly threatened by land-use change, which leads to deforestation, forest fragmentation and the replacement of original forests by crops, (Oliveira et al., 2007;S anchez-Cuervo et al., 2020). Between 2001 and 2020 Peru lost 33,900 km 2 of primary forest equivalent to 4.3% of its forest cover, and 3.1% of its primary forest (Global Forest Watch, 2021), which equals approximately 1000 individual jaguars lost based on Jędrzejewski et al. (2018) density estimates. In addition to land use change, persecution and population decline of their natural prey base are considered important threats in Peru. However, these threats have not been fully assessed. Therefore, even if biological and ecological factors favored the survival of spotted cat populations during past commercial harvests when trade was legal, the current illegal trade which is likely underestimated, in combination with the aforementioned main threats, pose a significant challenge to wild cat conservation.
The reduction of big cat populations has repercussions not only for prey dynamics but also for ecosystem functioning, due to the role of large cats as top predators (Hoeks et al., 2020;Ripple et al., 2014). Unfortunately, information on the population ecology of wild cats is limited in Peru, with a few exceptions for ocelots (Kolowski & Alonso, 2010) and jaguars (Mena et al., 2020;Tobler et al., 2018;Tobler, Carrillo-Percastegui, Zúñiga Hartley, & Powell, 2013). We recommend improving knowledge regarding population status of wild cats to inform conservation status and efforts to reduce illegal trade in alignment with Peru's national strategies. We also recommend improving public education campaigns for potential suppliers and purchasers of wildlife products as well as government authorities regarding the illegality of trade, its criminal repercussions, and the importance of conserving wild cats.

ACKNOWLEDGMENTS
We thank the National Forest and Wildlife Service (SERFOR) and the GOREs (Regional Governments of F I G U R E 3 During seizures, live specimens of wild cats are recovered, as well as parts such as whole skins or transformed into handicrafts to be more tempting to the consumer. WCS© Loreto, Ucayali, and Madre de Dios) for providing data about wildlife seized. We are grateful to Catherine Sahley for reviewing previous versions of the manuscript.

CONFLICT OF INTEREST
The authors declare no conflict of interest.

AUTHOR CONTRIBUTIONS
José Luis Mena: designed the study, collated, analyzed, and interpreted the data and wrote the manuscript. Rosa Vento, Jorge Luis Martínez and Ana Gallegos: supervised the development of the manuscript, and assisted with the study design, analysis, and editing.

DATA AVAILABILITY STATEMENT
All data are freely available from the referenced sources.