Tropical regions around the world are known for their high biodiversity and endemism of flora and fauna (Hamer & Hill 2004; Lodh & Agarwala, 2016; Meseguer et al. 2022). However, they are under threat of species extinction due to deforestation and habitat loss caused by increasing human demands (Lodh & Agarwala, 2016; Moreira et al. 2023). Among these regions, the Andean tropical montane forest (ATMF) is considered the principal hotspot of global biodiversity, and it has extremely high species richness in most taxonomic groups of plants and animals and contains sites with high speciation rates (Pennigton & Lavin, 2010; Toledo-Aceves et al. 2011; Ledo et al. 2012; Horwath et al. 2019; Bax et al. 2021; Meseguer et al. 2022), and it is also home to many endemic species with restricted geographic ranges (Luebert & Weigend, 2014; Hutter et al. 2017; Pérez-Escobar et al. 2017; Meseguer et al. 2022). However, the ATMF is among the most threatened ecosystems on the planet, with a speedy rate of loss of its natural vegetation cover due to human land-use change, such as agricultural expansion, urbanization, and deforestation (Armenteras et al. 2011; Toledo-Aceves et al. 2011; Quitián et al. 2018; Aide et al. 2019; Horwath et al. 2019; Moreira et al. 2023). This has put many endemic species at risk of extinction due to the bottleneck process in their populations (ToledoAceves et al. 2011; Mahecha-Jiménez et al. 2011; Soh et al. 2019; Bax et al. 2021). Moreover, due to their restricted elevation range, many ATMF plant and animal species have naturally limited distribution in the cloud belt of tropical mountains, ATMF distribution is often considered analogous to an archipelago or islands, with deforestation enhancing the isolation of remaining ATMF fragments (Toledo-Aceves et al. 2011; Mahecha et al. 2019).

Over the last two decades, there has been a gradual loss of tropical forests in Latin America, specifically the ATMF, that have been cleared and replaced with pastures and crops, leading to a permanent loss of forest cover, and due to their high population density, the ATMF has been subjected to intense agricultural activity, which has led to their reduction (Castellanos-Mora & Agudelo-Hz, 2020; Moreira et al. 2023). Extensive areas of cloud forests are being cut down for timber, cleared for agriculture at mid-elevations, and burnt for grazing at high elevations and the act of burning is likely the primary reason behind the widespread and dramatic disappearance of vast portions of lowland and montane tropical forests in South America for centuries (Arroyo-Kalin, 2012; Aide et al. 2019). It is increasingly clear that the impact of human activity in the tropical Andes has been intense and longlasting, with evidence dating back to the pre-Columbian era, and this activity has greatly contributed to the deforestation of the region, particularly at high elevations near timberline, and on a larger scale than what we see today (Loughlin et al. 2018). For the above reasons, ATMF is widely regarded as a conservation priority worldwide because provide important ecosystem services to millions of people (Ivory & Misrachi, 2014; Moreira et al. 2023) due to its critical role in the maintenance of hydrological and nutrient cycles, carbon sequestration and storage, and prevention of erosion and its high biodiversity (Toledo-Aceves et al. 2011; Soh et al. 2019; Bax et al. 2021). For that reason, there is an immediate need for conservation in highly diverse tropical ecosystems, especially in remote areas like mountain habitats. However, the lack of data has been a barrier to effective planning and implementation of conservation efforts (Lodh & Agarwala, 2016).

Conservation International introduced the Rapid Assessment Program (RAP) in 1990 for speciesrich areas, to create baseline data which includes species checklists, records of habitat structure, and of disturbance and anthropogenic pressures in a short time, and vegetation type, etc. (Alonso et al. 2011; Lodh & Agarwala, 2016). To address this, Rapid Biodiversity Assessment (RBA) is a cost- and timeeffective method for exploring the biodiversity of tropical habitats and it is considered a good method for predicting species richness using indicator taxa (Pearson, 1994; Lodh & Agarwala, 2016; Oliveira et al., 2020; Tettey et al. 2020). In addition, the RBAs only use a taxonomic group to indicate the site’s status (Tettey et al. 2020). Butterflies are commonly used to assess the health of both pristine and anthropogenic ecosystems. They have also been utilized to study various aspects of tropical forest ecology in natural, managed, and degraded ecosystems (Tettey et al. 2020). Moreover, butterflies are a promising group of insects for rapid assessment because they are widespread, conspicuous, and easily recognizable, taxonomically well-known, and effective indicators of forest health (Lodh & Agarwala, 2016; Henao-Bañol & Gantiva-Q., 2020; Tettey et al. 2020; Sharma & Sharma, 2021). Also, Lepidoptera diversity is an indicator of biodiversity because they depend on plant diversity and different habitats according to ecosystem ecology (Ferrer-Paris et al. 2013; Lodh & Agarwala, 2016; Tettey et al. 2020; Sharma & Sharma, 2021).

The subtribe Pronophilina Reuter, 1896 is one of the groups of Lepidoptera almost entirely restricted to ATMF region (Casner & Pyrcz, 2010; Mahecha et al. 2019; Pyrcz et al. 2019). Thanks to numerous contributions, over the last two decades, in all aspects of their taxonomy, phylogeny and natural history it turned from one of the least known into one of the better known, if not the best-known group of Andean Lepidoptera, and they present a number of characteristics which make them a group particularly suitable for researching biodiversity (Pyrcz et al. 2009; Casner & Pyrcz, 2010; Pyrcz et al. 2019; Díaz-Suárez et al. 2022). In particular, Pronophilina are one of the most abundant and arguably the most species-rich group among AMTF with well over 600 known species (Lamas et al. 2004; Pyrcz, 2010). They are highly habitat-specific, restricted to well-defined ranges of altitude, an issue which has been studied with considerable focus, and are extremely sensitive to habitat degradation which affects their elevational ranges (Mahecha-Jiménez et al. 2011; Pyrcz & Garlacz, 2012; Díaz-Suárez et al. 2022). Species are often geographically limited, confined to single ranges or valleys, with only a few having wider distributions (Pyrcz et al. 2020). Even though their highly complicated alphataxonomy still require some research, it can be considered as fair, and fast species identification only occasionally presents any serious obstacles (Marín et al. 2017; Padrón et al. 2021). In addition, the spatial distribution patterns of Pronophilina are well-known and fine-tuned both geographically and altitudinally, and this is fundamental from the perspectives of biogeography and population ecology. (Padrón et al. 2021). The presence of Pronophilina is generally high in bamboo thickets, where the adults remain close to their host plants, which are mainly Chusquea Kunth bamboos (Poaceae) (Pyrcz et al. 2009; Mahecha et al. 2019; Pyrcz et al. 2020; Padrón et al. 2021). Finally, and importantly, the use of van Someren-Rydon traps baited with decomposing animal matter guarantees the obtaining of a numerically large, statistically viable sample, frequently representative of local Pronophilina communities within a limited time frame (Pyrcz et al. 2009; Díaz-Suárez et al. 2022).

On a global scale, the number of species in most animal groups, including Lepidoptera, either decreases steadily with increasing elevation or has a peak at mid-elevation (Brehm et al. 2007; Grytnes & Maccain, 2007; Despland et al. 2018). However, recent studies conducted throughout the Andes yield a somewhat different pattern, with several groups of Lepidoptera whose diversity increases with altitude, sometimes reaching the highest values close to timberline at some 2600-3000 m, for example some taxa of the Larentiinae subfamily of Geometridae (Brehm et al. 2003), or to that matter the Pronophilina (Pyrcz, 2010). In this study, we did an RBA of Pronophilina butterfly community structure and diversity along two transects in the uppermost cloud forests near timberline in Manu National Park, Cusco department in SE Peru with the aim to confirm that Pronophilina are good habitat preservation indicators in the Andes.

This is the first RBA study for Pronophilina butterflies diversity in a sanctuary protected by law in Manu National Park, Cusco department in SE Peru, and our study proves once again that sampling with van Someren-Rydon baited traps is an extremely efficient method in carrying out quick surveys of Lepidoptera diversity and community structure in both lowland and montane rainforest habitats for local or regional studies (De Vries & Walla, 2001; Maicher et al. 2020), as well as in ecological assessments of habitat preservation versus anthropogenic activity within a limited time frame (Aduse-Poku et al. 2012; Whitworth et al. 2016). Such a method is efficient not only by attracting large numbers of individuals which enables to produce viable samples for statistical analysis, but it is also relatively specifically unbiased being equally attractive for all the species of a butterfly community, at least true for cloud forest Satyrinae (Pyrcz et al. 2009; Pyrcz & Garlacz, 2012). Biased sampling method is frequently a problem in similar studies, thus the use a wide array of baits is necessary (Maicher et al. 2020).

This study is, in our opinion, also valuable because it is one of the few to offer a comprehensive comparison of two ecologically similar and geographically marginally set apart localities, allowing a fine-tuned insight into intricate zoogeography of one of topographically most complex areas of the Andes. In fact, the two sampling transects are only 30 km away, both situated on the eastern slopes of the Vilcanota range and are not separated by any noticeable topographical event. And yet, their Sørensen similarity index is surprisingly low, and several interesting differences can be pointed out. For example, the absence of Pedaliodes uniformis the fifth most abundant species overall, and Lymanopoda inaudita along the Trocha Unión transect most likely reflects their south-eastern distribution limit, which is relevant from a zoogeographical perspective as it uncovers a discreet geographical pattern. On the other hand, the absence of Panyapedaliodes phila, a widespread Andean species common at 2600-3000 m along the Qurqurpampa transect is difficult to explain. For comparison, the faunas of Pronophilina of two sites separated by 50 km in the Venezuela Cordillera de Merida were also studied (Pyrcz & Wojtusiak, 2002; Pyrcz & Garlacz, 2012). In that case, the Sørensen similarity index was significantly higher, reaching 0.96. The two samples in Venezuela differed by one species only. Our results indicate much higher beta diversity, understood as species turn-over on a geographical scale, in the Peruvian Andes. Also, the species richness in Venezuela, whereas the altitudinal span covered was much wider, 700 m compared to 400 m in this study, was considerably lower with 22 and 23 species respectively. Higher diversity of two Peruvian localities was also reflected by Shannon and Fisher alpha indices.

Faunal structure of Pronophilina evidenced in this study is similar with other sampled Andean localities, such as Monte Zerpa and El Baho in the Venezuelan Cordillera de Merida (Pyrcz & Wojtusiak, 2002; Pyrcz & Garlacz, 2012), Alto del Zuque-Bogotá (Mahecha-Jiménez et al. 2011) and Frailejonal- La Calera (Díaz-Suárez et al. 2022) in the Colombian Eastern Cordillera; Belmira (Pyrcz et al. 2018) in the Colombian Central Cordillera; Tambito (Pyrcz & Wojtusiak, 1999), and Golondrinas (Pyrcz et al. 2009) on the Pacific slopes in Colombia and Ecuador, and Chachapoyas in northern Peru (Pyrcz, 2004). All the communities of Pronophilina are characterized by the dominance, in terms of species-richness and abundance, of Pedaliodes¸ with one to three strong dominant or co-dominant species, and a comparative under-representation of other genera in the sample. Such a faunal structure is not an artefact of the method of sampling nor seems to be related with the geographic setting of sampling locality or altitudinal span covered. The species dominance of Pedaliodes in the sample is not surprising given that it is also the most species-rich genus of Pronophilina accounting for nearly 50% of all known species in this subtribe (Pyrcz, 2004). However, the dominance of Pedaliodes in terms of abundance is less straightforward to interpret. Also, in the less diverse communities the dominance of a single species of Pedaliodes is more accentuated, which is particularly striking in the case of El Baho sample where Pedaliodes minabilisPyrcz, 2008 constitutes as much as 53% of the sample (Pyrcz & Garlacz, 2012). Interestingly, dominant species of Pedaliodes in each locality belong to completely unrelated clades of the genus (Pyrcz et al. in prep.).

Our results suggest good preservation of the patches of uppermost forest where both samplings were carried out. This is shown, first, by the presence of species indicators of well-preserved habitats, in particular Corades sp. nov., Apexacuta astoreth (Thieme, 1907), Lymanopoda inauditaPyrcz, 2010, or Pedaliodes phaeinomorpha Viloria, Lamas & Pyrcz, 2008, the latter one of the co-dominant species in the study area, elsewhere extremely localized and rare, for example along the nearby Acjanaco - Manu Road due to high disturbance of forest cover at 3000-3200 m. At the same time, very few species related with disturbed forest habitats were found in the study area, in particular only one individual of Panyapedaliodes drymaea (Hewitson, 1858) which is one of the most prominent indicators of high anthropogenic disturbance, very few specimens of Steremnia umbracina (Butler, 1873), another species related with secondary habitats, and none of Pedaliodes palaepolis (Hewitson, 1878), again a secondary forest indicator species (Pyrcz, 2004; Mahecha-Jiménez et al. 2011; Díaz-Suárez et al. 2022). Second, the RAD models fit in Trocha Unión (Zipf-Mandelbrot) and Qurqurpampa (log-normal) are also indicative of undisturbed, natural, varied and mature communities (may, 1975; Hill et al. 1995; Hill & Hamer, 1998; Marimon et al. 2015; Harterreiten-Souza et al. 2020; Nallis, 2021). In fact, the Zipf-Mandelbrot model assumes that, in the succession processes, late colonizing species have a specialized niche and thus are more difficult to find, while early colonizing species are more generalist and dominate in abundance regardless of the time of succession (Magurran, 2013; Harterreiten-Souza et al. 2020; Nallis, 2021). This model also assumes that colonization is initially stochastic depending on the regional abundance of generalist species, but that mature communities are determined according to the niche of late species (Nallis, 2021). The log-normal model assumes that ecological communities will have many species with intermediate abundance and few rare or common species (Schluter & Ricklefs, 1993; Harterreiten-Souza et al. 2020; Nallis, 2021). It has been argued that the use of RAD is a much more reliable method in comparing communities than non-parametric indices of diversity (Kim et al. 2013; Sæther et al. 2013). The latter, frequently used in the studies of biodiversity are shown to oversimplify the complex structural nature of communities and are not independent of sampling intensity, thus on sample size as well (Mouillot & Lepretre, 2000; Dornelas et al. 2011). Interestingly, this does not seem to be the case in our study. Most indices show closely similar values despite the fact that the Trocha Unión sample contains two times as many records as Qurqurpampa.

In sum, our study confirms that Pronophilina, thanks to their ecological and also behavioural qualities, such as habitat fidelity coupled with low adult vagility, and high sensibility to human made disturbance, which immediately express in their community structure, can be explored in the assessment of cloud forest state of preservation, and thus in formulating future conservation policies. In addition, to their other features, such as easy high species richness, standard sampling procedures, rapid data gathering protocols, and good knowledge of taxonomy among others, they surpass as potential habitat preservation indicators any other group of Andean Lepidoptera.