EUROPEAN JOURNAL OF ENTOMOLOGY EUROPEAN JOURNAL OF ENTOMOLOGY west-Habitat of Carabus zawadzkii (Coleoptera: Carabidae) in the Eastern Carpathians

. This paper provides an analysis of the rare and important Natura 2000 species Carabus zawadzkii in terms of its ecological niche, conservation status and relationship with other Carabus species in carabid assemblages. Published sources are inconsistent in de ﬁ ning the habitats in which this species occurs. Therefore, a large part of this paper is dedicated to identify-ing the relationship of Carabus zawadzkii with particular non-forest habitats based on ﬁ eld research conducted in 2011–2013. The results revealed an unexpected af ﬁ nity of this species for non-forest habitats in addition to the know relationship with forest habitats. Further analysis, using additional ﬁ eld data collected over the period 2013–2021, also focused on the assessment of the current conservation status of this rare species. Overall, the population of this species seems stable; however, to improve its conservation status it requires particular management measures. Important conservation measures include periodic extensive mowing of mountain hay meadows. This species’ seasonal dynamics provides important information for nature conservationists. It is important to carry out research on this species in May, when it is most active. Presented results provide the most comprehensive overview of the ecological preferences and occurrence of this important and very rare Carpathian species. Carlo permutation test (Ter 2018; & Lepš, 2014) of the full model, with 1999 iterations. Software CANOCO5 was used for this purpose. A signi ﬁ cance level of α = 0.05 was used to test the parameters of the variables. Results for all of the 20 localities studied were tested. Species richness and Shannon’s index were calculated using the statistical package in MS Excel.To evaluate the signi ﬁ cance of the carabid assemblages at the localities we used sample averages (number of individuals at a locality / total number of species), counts (number of species at the localities), recount (number of species at a locality / total number of species), variance (sample variance), total (sum of speci-mens), N2 (Simpsons index = 1 / Σ (pi)2), H´ (Shannon entropy = – ∑ pi * lnpi), H´max (log number of species in samples) and evenness (sample H´ / log (N) ratio). These data were exported from the Canoco5 program using the menu Statistics of composi-tional table (Šmilauer & Lepš, 2014). We chose ordination methods based on the greatest lengths of the environmental gradient (lengths of gradient = SD units) as described by Šmilauer & Lepš (2014).We used 11 environmental variables to de ﬁ ne the characteristics of a site: cover (E1%), abandoned meadows, building mowed meadows and type of (coded as 6 dummy variables: Nardo strictae-Agrostion Calamagrostion Calthion Arrhenatherion , , Caricion lasiocarpae 0.05)

Kraatz, 1854, in the Carpathian foothills. In older literature on its occurrence in the Ukrainian Carpathians, this species is listed as 2 subspecies, C. zawadskyi zawadskyi Kraatz, 1854and C. zawadskyi ronayi Csiki, 1905(Kryzhanovskij, 1983Kryzhanovskij et al., 1985). Also, 3 specimens of Carabus zawadzkii were collected in the 19 th Century in Galicia, currently on the border between Poland and Ukraine, which are now deposited in the Natural History Museum of Sibiu Entomological Collections (Muntean et al., 2020). ern Ukraine, north-eastern Hungary and northern Romania (Figs 1, 2). In Poland it occurs in the Bieszczady mountains and their foothills. The north-western border of its range is in Poland. There are 7 sites known in Poland where it is found in small numbers (Pawlaczyk et al., 2004). The fi rst information from Poland in the year 1991 related to its occurrence was based on one specimen caught on 11th of July 1974. Specimens from several sites are in the collections of the Museum and Institute of Zoology of the Polish Academy of Sciences in Łomno near Warsaw. In Poland this species, is found mainly in meadows and rotten stumps and lying logs. Due to the lack of specifi c data, it was previously not classifi ed as a native beetle (Burakowski et al., 2000). Over the last two decades there was only one confi rmed record for the year 2011 (BioMap diversity, 2022). It is also found very rarely in Slovakia and only in north-eastern Slovakia (especially Vihorlat mountains, Ondavská and Laborecká highlands, Bukovské and Čergov mountains) (Olšovský, 2015). In Romania, Barloy & Prunar (2012) document its occurrence in 15 localities in the Maramureş and Bistriţa Năsăud districts. There are two confi rmed localities in Ukraine, namely Dilove and Bogdan in the Maramureș mountains and Podpula Mts respectively (Barloy et al., 2014). Putchkov (2011Putchkov ( , 2012 documents the occurrence of two subspecies in the Ukrainian Carpathian range: C. zawadskyi seriatissimus Reiter, 1896, in the Carpathian moutains, and C. zawadskyi zawadskyi   Fig. 1 (Rizun, 2003).
There is relatively little published information on its habitat preferences and distribution. Art. 17 reporting according to HD defi nes this species as follows: "It lives predominantly in forests and forest margins of hills and mountains, but it also occurs in river valleys, both in riparian forests and in wet meadows". In Romania it is reported by Barloy & Prunar (2012), particularly the subspecies Cara-bus (Morphocarabus) scheidleri zawadzkii seriatissimus Reiter, 1896. Its habitats are classifi ed as forests or forest edges, or bushes at the boundaries of meadows in the vicinity of semi-permanent brooks. Based on its occurrence in the Poloniny region of Slovakia it is classifi ed as a forest species (Jászay, 2001). Only one expert considers this species to be one that inhabits foothills, lower mountain  (Niedl, 1957).  meadows, well-lit forests with glades and edges of forests near meadows with extensive management and with plenty of natural hiding places (Olšovský, 2015). Its occurrence in the fi rst half of the 20th century in the former Czechoslovakia and in surrounding countries are summarized by Niedl (1957), where this species is referred to 2 subspecies of Carabus scheidleri, namely C.s. zawadskyi and C.s. ronayi. This species is reported in eastern Slovakia and also in Hungary and Ukraine (Figs 3, 4).

Taxonomy
There are very few studies on the taxonomy of this species and most of them indicate it is problematic in terms of its various subspecies (Barloy et al., 2014). Zawadzkii's ground beetle was a species with unclear taxonomic status in the past. For specialists on Carabidae, it is only a subspecies of Scheidler's ground beetle (Carabus scheidleri Panz.) in the Western Carpathians, which is widespread in Central Europe. Also, in the last Catalogue of Palearctic Coleoptera (Löbl & Löbl, 2017) it is a subspecies of Scheidler's ground beetle as Carabus scheidleri zawadzkii (according to Annex II of HD it is Carabus zawadzkii). Carabus scheidleri zawadzkii seriatissimus Reiter, 1896 recorded from Romania and Ukraine is on the other hand in the Catalogue of Palearctic Coleoptera listed as separate subspecies Carabus scheidleri seriatissimus (Löbl & Löbl, 2017) and zawadzkii is not mentioned. However, there are experts who recognize it as a completely separate species. From the point of view of European legislation, the process of declaration of Natura 2000 sites is based on the Habitats Directive, where under the listed name Carabus zawadzkii there are currently Carabus scheidleri zawadzkii zawadzkii, Carabus scheidleri zawadzkii ronayii and Carabus scheidleri seriatissimus (listed also by some experts as Carabus scheidleri zawadzkii seriatissimus, Carabus zawadzkii seriatissimus) and this understanding is accepted in the present paper.

Aim of the research
The aim of this research is to use fi eld data to clarify the species' niche in terms of its association with mainly nonforest habitats and identify what management measures will better maintain or improve the conservation status of Carabus zawadzkii. Forest habitats are already confi rmed habitats of this species and not the main focus of this research although it also includes some fi ndings related to forest ecosystems. A signifi cant part of the research is also dedicated to determining the most appropriate monitoring methods for assessing the seasonal dynamics of this species. An additional aim is to present a comprehensive overview of the distribution of this important and very rare Carpathian species for which there is currently very little information.

Study area
Carabus zawadzkii and carabid assemblages were studied in the Slovakian part of the Eastern Carpathians. This study is mainly an inventory of the non-forest habitats of this species and an assessment of its conservation status.

Inventory
This part of the research was carried out mainly at the end of the last century and fi rst years of the present century with the aim of collecting data on the distribution of the epigeic fauna especially the species of community interest (Natura 2000 species) including Carabus zawadzkii. Only a few records are available for the period since 1955 and 1970s. An overview of the localities where Carabus zawadzkii occurs is presented in Table 1 and Fig.  4. Sampling was carried out mainly by using formalin pitfall traps without bait, which is identical with that described in subchapter "Sampling". This was supplemented by collecting specimens. Most of the specimens of Carabus zawadzkii collected were deposited in the Šariš Museum of Entomological Collections (Bardejov, Slovakia). Our chosen model area was the Poloniny National Park, which extends over a large part of the Bukovské mountains and includes a small area in Slovakia, located in the eastern part, on the borders with Poland and Ukraine. The topography of the area is mainly mountaineous, with a more or less rugged relief of the fl ysch zone and dominance of deciduous forests only in the eastern part, with mixed forests here and there.
The landscape in the area is typically Carpathian and the localities studied were chosen for their similarity with other places in the known range of Carabus zawadzkii. According to Bezák (2010), meadow ecosystems currently cover about 10% of the area. These are mainly typical mountain meadows, which are located on the main ridges of the Bukovské mountains. In addition to mountain meadows, various non-forest ecosystems are also present: pastures; moist, mesophilic, ruderal, xerothermic and permanent grasslands; and other types of non-forest habitats. The research was carried out at 20 localities, with 7 types of nonforest habitats characteristic of the Eastern Carpathians (Table 2 and Fig. 5).
Localities studied: "Mountain meadows" The mountain meadows studied are Ruské 1 (R1), Ruské 2 (R2) and Ruské 3 (R3). R1 is a typical mowed mountain meadow belonging to the Nardo strictae-Agrostion tenuis alliance. Depending on the time since last mowing, they were at different successional stages. Tree species coverage is rather low, but species composition changes over time (Ružičková & Halada, 2002;Halada et al., 2004). There was a fi re at this site in the spring of 2012. R2 and R3 are meadows belonging to the Calamagrostion arundinacae alliance. The communities of this association primarily occur at high altitudes, but also occur secondarily at low altitudes in areas free of forest due to grazing. In the Bukovské mountains, the occurrence of communities of this association are secondary and occur on the main ridge and sides of high ridges at altitudes above 800 m (Ružičková & Halada, 2002;Halada et al., 2004). The R2 site was not mown during the study period and that at R3 was mown once.
"Wet meadows and fens" Wet meadows and bogs occur at Ruské 4 (R4), Ruské 8 (R8), National Nature Reserve (NNR) Pod Ruským 10 (R10), Kolbasov 17 (Ko17) and Nová Sedlica 18 (Se18). These are moist meadow habitats with a relatively stable groundwater level, where the soil surface is rarely completely dry. All were created by human activ-ity. They belong to the Calthion alliance. All areas are mown once a year (Ružičková & Halada, 2002;Halada et al., 2004). The Se18 locality is a peat meadow belonging to the Caricion lasiocarpae alliance. At Pod Ruským (R10) there are natural communities of moist waterlogged meadows and wetlands on the fl oodplain of the Cirocha River. The overall character of the vegetation is that of tall wetland plants typical of swamp and wetland communities.
"Intensively used mesophilic meadows" The localities Príslop 12 (Pr12) and Topoľa 14 (To14) are nonforest habitats. They are intensively-used mesophilic meadows at middle mountain locations and lowlands on nutrient rich or fertilized soils. They belong to the alliance Arrhenatherion elatioris. The areas surveyed were mown twice a year.
"Mesophilic meadows with specifi c management" The localities Príslop 13 (Pr13) and Topoľa 15 (To15) are nonforest habitats of this type. These are mesophilic meadows at mid mountain locations and in lowlands belonging to the Arrhenatherion elatioris alliance, which are similar to the previous type, but importantly unlike them in terms of biodiversity and are managed in a specifi c way (e.g. use of only organic fertilizers, limits on the dosage thereof, etc.) as specifi ed in the document "Rural Development Program SR 2007-2013" (Ministry of Agriculture of SR, 2010). The funds for their protection and management are provided by higher subsidies than for intensively used meadows. The areas surveyed were mown once a year. "Nitrophilous ruderal plant communities" Two localities, namely Ruské 6 (R6) and Ruské 9 (R9), are of this type. These are nitrophilic ruderal fringe communities outside settlements, belonging to the Galio-Alliarion alliance (R6) and the Aegopodion podagrariae alliance (R9), association Cherophylletum aromatici. There was a fi re at locality R9 in 2013.
"Abandoned meadows" The localities Starina 11 (St11) and Nová Sedlica 19 (Se19) include abandoned meadows. St11 is an unmown acid meadow with a fl uctuating water regime infl uenced by groundwater from the Starina reservoir. It belongs to the alliance Caricion lasiocarpae. It is at succession stage of moist thistle and Filipendula meadows of the Polygalo-Cynosurenion alliance. These meadows are more species-rich with a unique species composition (Ružičková & Halada, 2002;Halada et al., 2004). Se19 is an unmown degraded meadow belonging to the alliance Calthion palustris.

Analysis of the distance of the localities from forest habitats
Since the preference of Carabus zawadzkii for forest habitats is mentioned by several authors, it is necessary to determine the distance of the areas studied from the nearest forest when determining and confi rming a possible relationship with other habitats (in our case non-forest habitats). Distance is especially important in terms of determining the extent to which nearby forest ecosystems can affect the occurrence of this species. For this purpose, a GIS analysis of the distance of the localities studied from the nearest forest habitats was carried out. This analysis was performed in Arc GIS 10.1 using the NEAR function, which measures the shortest distance to specifi ed elements (in this case, polygons of forest habitats). Based on the recorded distances of the 20 localities the average value of the distance of all areas from forest habitats was calculated.

Sampling
Field research was conducted over a two-year period: June 2011 -June 2013. Pitfall traps were used to catch beetles of the genus Carabus (Coleoptera) especially Carabus zawadzkii. At every locality a line of 5 pitfall traps (plastic cups of 10 cm diameter and 0.5 l capacity) were set at 5 m intervals and monitored on a regular basis.
Diversity of Carabus assemblages recorded in the habitats at the localities studied Diversity was evaluated using three parameters: number of taxa, α-diversity, Shannon index of diversity (H) and Pielou index of equitability (J).

Evaluation of ecological factors and gradients
For determining the statistical signifi cance of the relationships of LEV (landscape environment variables) with the distribution of Carabus beetles at the localities studied, we used a Monte- Carlo permutation test (Ter Braak & Šmilauer, 2018;Šmilauer & Lepš, 2014) of the full model, with 1999 iterations. Software CANOCO5 was used for this purpose. A signifi cance level of α = 0.05 was used to test the parameters of the variables. Results for all of the 20 localities studied were tested. Species richness and Shannon's index were calculated using the statistical package in MS Excel.
To evaluate the signifi cance of the carabid assemblages at the localities we used sample averages (number of individuals at a locality / total number of species), counts (number of species at the localities), recount (number of species at a locality / total number of species), variance (sample variance), total (sum of specimens), N2 (Simpsons index = 1 / Σ(pi)2), H´ (Shannon entropy = -∑ pi * lnpi), H´max (log number of species in samples) and evenness (sample H´ / log (N) ratio). These data were exported from the Canoco5 program using the menu Statistics of compositional table (Šmilauer & Lepš, 2014). We chose ordination methods based on the greatest lengths of the environmental gradient (lengths of gradient = SD units) as described by Šmilauer & Lepš (2014).
We used 11 environmental variables to defi ne the ecological characteristics of a site: altitude, vegetation cover (E1%), abandoned meadows, building rubbish, mowed meadows and type of vegetation (coded as 6 dummy variables: Nardo strictae-Agrostion tenuis, Calamagrostion arundinaceae, Calthion, Arrhenatherion elatioris, Galio-Alliarion, Caricion lasiocarpae). We tested the statistical signifi cance (pα = 0.05) of the null hypothesis (environmental factors do not affect the species groupings) using a Monte Carlo permutation test with 499 permutations. The success of the analysis was measured by the value of the coeffi cient of determination (R 2 ) and its adjusted value R 2 adj , which considers sample size and number of variables. We test-ed the simple effects, which summarize the independent effects of all the explanatory variables, and conditional effects, which summarize the conditional (partial) effect of each predictor. We verifi ed the tightness of the mutual relationship (correlation) with environmental factors using the nonparametric Spearman correlation coeffi cient in the Statistika.cz program (StatSoft, 2004). For this verifi cation we used the correlation matrix in Canoco5 (log fi le, p α = 0.05). We evaluated The strength of the linear bond was evaluated using the r coeffi cient and the categories proposed by Conelly (2012): strength of correlation r < 0.2 slight, r = 0.2-04 low, r = 0.4-0.7 moderate, r = 0.7-0.9 high, r > 0.9 very high.

Monitoring -conservation status
In order to assess the conservation status, specifi c fi eld visits were carried out in the period 2013-2021 (Fig. 6) and additional data collected in 2011-2013. The plots monitored were visited repeatedly in order to record quantitative and qualitative data using a standardized method (Saxa et al., 2015). This method was based primarily on monitoring abundance using animal traps distributed along a transect (5 traps at 5 m intervals) and recording parameters related to the quality of the species' habitat, future prospects, threats and pressures. At 4 plots selected for monitoring Carabus zawadzkii, 20 fi eld visits were made. All the plots monitored were in non-forest habitats, mostly habitats in hay meadows.
The conservation status based on individual parameters and overall assessment were divided into 3 categories: favourable (FV), unfavourable-inadequate (U1) and unfavourable-bad (U2).
Evaluation of the species' conservation status using this categorization is based on the evaluation of these parameters: (a) Quality of the population at a locality (b) Quality of the habitat at a locality (c) Future prospects of the species' habitat at a locality 17.9 10.5 9.2 2.3 6.4 8.5 3.1 5.6 5.9 3.4 9.1 0.9 2.4 1.1 2 0.9 0.5 2.6 6.1 1.6 100 No. of species 6 9 9 7 9 11 9 8 8 6 6 3 3 4 4 5 2 5 7 8 These three parameters were evaluated. For each parameter and category of status, the percentage values were estimated on the basis of an expert assessment during short fi eld visits (two days visits per year). The process by which this data was used to assess the overall status of the individual parameters was as follows.
The overall status for a particular parameter, a, b or c, is favourable when the following values are achieved: -favourable ≥ 85% of assessments are favourable, or favourable when ≥ 70% of assessments are favourable and no assessements are unfavourable-bad.
The overall status for a particular parameter is unfavourablebad when the following values are achieved: -unfavourable-bad assessments are ≥ 50% of the total assessments.
All other combinations of percentages result in an unfavourable-inadequate status of the given parameter.
This assessment of parameters was followed by a joint evaluation of all assessed parameters, which combines the results of the evaluations of all the parameters. The parameter which scored the worst determines the total assessment at the locality level. So, where all three parameters are favourable (FV), the overall conservation status at the locality is also assessed as favourable (FV). If one or more parameters are assessed as unfavourable-bad (U2), the overall conservation status at the locality is assessed as unfavourable-bad (U2). All other combinations result in an unfavourable-inadequate conservation status (U1). This method of assessment was used separetly for each plot monitored. The overall approach and assessment were based on the standardized conservation status assessment reported for Slovakia (Janák et al., 2015). After the calculations at the locality level, the fi nal conservation status was defi ned on the basis of the number of FV, U1 and U2 assessments. The parameter which occurred most often was recorded as the conservation status. In addition, the average % for the quality of the population, quality of the habitat and future prospect was calculated as a mean value for the assessments of all 20 fi eld visits.

Inventory of the samples of Carabus zawadzkii
There are 73 records of Carabus zawadzkii recorded by experts in the period from 1976-2020 (Table 1). Based on the prevalence of records (41 vs. 29) and abundance (118 vs. 96 specimens) this species mainly occurs in non-forest habitats (Table 4). It is evident that this species also occurs in forest habitats. In forest habitats Carabus zawadzkii is recorded occuring in beech forests, fi r-beech forests, alder forests, thermophilous oak and oak-hornbeam forests (Table 4). In the case of forest habitats, it is possible that the edges of the forests are also suitable as at Babličke in oak-hornbeam forest, during the same period of time, this species was more numerous in traps at the edge of the forest than in those located in the forest (9 vs. 4 specimens).

Number of species of Carabus recorded in non-forest habitats
During the period 2011-2013, a total of 1399 individuals of the genus Carabus belonging to 16 species, were recorded in seven non-forest habitats ( Table 3). The most numerous species in non-forest habitats in the area studied were the 3 eudominant species, Carabus cancelatus (26.02%), Carabus violaceus (18.44%) and Carabus arcensis (14.72%). Five Carabus species were dominant in the area studied. Of the Annex II HD species, two were

Biodiversity
The carabid assemblages in the samples were evaluated in terms of species richness, diversity and equitability (Fig.  7). The highest number of species was recorded in nitrophilous ruderal plant communities (R6), with 11 species. High values of the diversity index were recorded in habitats at localities R7 and R8, and the highest value of equitability at localities Pr12, To14, To15 and Bz16 (above 0.9). The carabid assemblages recorded at locality Pr13 had the lowest equitability value (0.24) due to the low number of identifi ed species and of the eudominant Carabus zawadzkii. The lowest number of species (2 to 3 species) was recorded in the meadow communities at Príslop (Pr12 and 13) and wet meadow at Kolbasov (Ko17).
Detailed statistics and calculated indexes of Average, Count, Relative Count, Variance, N2, H´, H´max and Evenness for individual localities are in Table S1.

Carabus zawadzkii in non-forest habitats
This species was recorded more frequently in non-forest habitats. The species was recorded at most localities (13 of the 20 localities studied) and in all types of non-forest habitats. The results indicate that in addition to other habitats this species regularly occurs abundantly in mountain meadows and was recorded in all the mountain meadows studied (Table 3).
It was most abundant at Pr13, which is one of the 2 managed sites (Table 3). Thirty-two individuals were recorded here over 2 years. A relatively high abundance of this species (7 individuals) was also recorded at Pr12, but it is 4.5 times less than at Pr13. It is interesting that these localities had low α-diversities and therefore Carabus zawadzkii was subject to little competition. In contrast, the lowest frequency of occurrence of Carabus zawadzkii was recorded in wet habitats and was recorded at only one of the 5 wetland sites studied. The presence in this type of habitat is probably accidental, because it was recorded only once and at the edge of wetland habitat R4, possibly due to a marginal effect. The wetland at R4 is relatively small and the composition of the invertebrates caught by the lateral trap and pitfall traps was possibly affected by extensivelyused meadows located close by. Based on these results we assume that this species avoids wetland habitats.
The unimodal analysis of the response (species) data provided a gradient value of 2.7 SD units, which indicates a linear method (Redundancy analysis) should be used. The analysis of the relationship (Fig. 8) between non-forest habitats and occurrence of the studied species suggests a connection between Carabus zawadzkii and extensivelymanaged mountain meadows that are mown once a year, or even less frequently, with the presence of some degree of secondary succession. Occurrence of this species was also recorded in uncultivated meadows that had not been mown for a long time, although in these areas the abundance of this species was very low. Based on these fi ndings, we conclude that this species prefers extensivelyfarmed mountain meadows (apart from forest habitats) and that the population declines with time since they were last mowed. On the other hand, there was a low abundance of this species in ruderal and forest areas. The preferred plant communities in non-forest habitats included the phytocenological alliances Nardo strictae-Agrostion tenuis and Calamagrostion arundinacae. We recorded similar habitat preferences (only occasionally mown or unmown) for the species Carabus auronitens, Carabus intricatus, Carabus obsoletus and Carabus cancelatus.  The vegetation cover factor and ruderal communities (building rubbish) are highly positively correlated (Spearman's r = 0.76) and cover is a statistically signifi cant factor. The vegetation types Calamagrostion arundinaceae (dominant tall grass in meadows at Poloniny) and Nardo strictae-Agrostion tenuis (r = 0.92) also correlate positively. Calamagrostion arundinaceae correlates very strongly and statistically signifi cantly with the mowed meadow factor (r = -0.92).
The unimodal analysis of the response (species) data provided a gradient value of 2.7 SD units, which indicates a linear method (Redundancy analysis) should be used. The species table includes 16 species of Carabus recorded at over 20 localities and 59.7% zero values, i.e a particular species was not present at a particular location. Of the selected factors, the vegetation types Galio-Alliarion and Caricion lasiocarpae, were not included in the analysis as they are correlated with other factors. The remaining factors constrained by the RDA model explain 69.29% of the variability of the species data (explanatory variables account for 60.41%). The adjusted value after Benferroni correction is relatively high at 41.65% (adjusted explained variation is 24.78%). The values of the simple and conditional effects of the Monte Carlo permutation test of the null hypothesis are presented in Table 5.
Based on these results it seems that this species, as previously suggested, has a wider ecological niche (Table 6) and is also able to some extent to survive in highly anthropically-disturbed habitats (nitrophilous ruderal communities).
The ordination analysis shows that Carabus zawadzkii prefers abandoned meadows at high altitudes with tall vegetation (Calamagrostis arundinacea). Mowed meadows, bogs and moist herbaceous communities (Calthion) are less suitable habitats, as these factors negatively correlate with the 1st ordination axis, p = 0.002. The vegetation cover factor positively correlates with the 2nd ordination axis, p = 0.024. NEAR analysis provided information on the distance of research plots in non-forest habitats from the nearest forest habitats. The results indicate that the average distance of the research plots from forest was 225 m. Based on these fi ndings the occurrence of specimens in non-forest habitats is not occasional or random. Although the dispersal ability of this species can play a role in its distribution in nonforest areas, its confi rmed presence at a high number of non-forest localities is signifi cant and that this species has broad ecological niche, which includes both non-forest and forest habitats. Detailed calculation of the distances is presented in Table S2.

The most common accompanying species of Carabus
According to Olšovský (2015), the most common accompanying carabid species are Carabus violaceus violaceus, Carabus glabratus glabratus, Carabus obsoletus and Carabus nemoralis nemoralis. Based on the results of our fi eld study, the most common accompanying species of Carabus in non-forest habitats are Carabus obsoletus, Carabus cancellatus and Carabus arcensis. Carabus zawadzkii has a negative, and statistically signifi cant, correlation with Carabus ulrichii. Carabus violaceus, one of the most common species, seems to prefer different types of habitat.

Spatial distribution
The spatial distribution of this species was uneven (Table  7) as it was often only caught by one or two of the fi ve pitfall traps set at 9 localities. A more even spatial distribution was recorded at St12, where similar numbers of this spe-  cies (1 or 2 individuals) were caught by all 5 traps. This is a locality where the grass was mown twice a year. In contrast, the neighbouring locality was extensively used (occasionally mowed) and this species was caught by all but one trap, but the number captured varied signifi cantly, from 1 to 23. The differences in the evaluated periods 2011-2012 and 2012-2013 are interesting. While in the period 2011-2012 this species was mainly recorded in the Cirocha basin, at high altitudes, in the following year it was signifi cantly more often caught by traps located in the Ulička valley (low altitude). It is likely that the population density varies from year to year and the factors that cause this need to be investigated. Overall, the total number of individusls caught in both periods was almost the same, which indicates that the population in the area is stable, though not large. A study of the dispersal of Carabus hungaricus (Elek et al., 2014) indicates that its spatial distribution is infl uenced by its movement, which is locality dependent. The average distances moved by males and females at each of the localities ranges between 47 and 207 m (Elek et al., 2014). Dispersal of Carabus zawadzkii needs further study, however the average distance moved is likely to be similar to that of Carabus hungaricus.

Sesonal dynamics
Seasonal activity of Carabus zawadzkii in both years was concentrated in May, when almost 73% of the individuals were recorded (Fig. 9). From the end of September to the end of October and during the winter period (November to end of April), no individuals were caught. In the other months (June to the end of September) very few individuals were caught. The period of activity coincides with their reproductive period. The beetles had one marked period of activity that extended from the end of April to end of May and very few were caught from end of May to end of September.

Conservation status
The conservation status was assessed based on the results of 20 fi eld visits to 4 plots (Fig. 10). The average quality of their habitat and future prospects were assessed as mostly unfavourable (39% U1 and 16% U2). The average quality of the population was assessed as 69% unfavourable (29% U1 and 40% U2), with only 31% assessed as FV. Overall, 14 fi eld visits resulted in an assessment of unfavourable -inadequate (U1) and 6 unfavourable -bad (U2). The overall conservation status is therefore assessed as unfavourable -inadequate (U1).
The most frequently recorded threat to non-forest habitats is that of insuffi cient mowing and abandonment of traditional management practices as this species is most abundant in those non-forest habitats that are frequently mown.

DISCUSSION
As there is very little information on Carabus zawadzkii this study provides new fi ndings on its relationship with forest and non-forest habitats and important information on its seasonal dynamics. This research did not focus on the composition of the population (age structure and repro-ductive characteristics) so in the future it would be interesting to study the whole life-cycle of this species.
The results presented are important for conservation agencies and organizations that are responsible for monitoring this species as it should enable them to optimize the use of funds for the collection of fi eld data. For instance, the monitoring method currently used in Slovakia is based on recording individuals of this species living under stones and by monitoring every locality using baited non-lethal pitfall traps during the period 1st of May to 15th of October. The number of visits per year is set at 3, and the second and third visit should follow no later than 3 days after the previous visit. With such monitoring it is possible that sites are visited at the wrong time of year resulting in wrong conclusions. Therefore, based on our results, we suggest that monitoring should be carried out mainly in May. As sampling over such a short period of time provides little and potentially unreliable information we suggest that inaddition to intensive monitoring during the whole of the period of activity in May the number of fi eld visits should be increased in order to increase the reliability of the data.
In addition this study defi nes the habitat in which this species normally lives. Previous studies indicate it lives mainly in forest habitats, but this study indicates it is also abundant in non-forest habitats including human-modifi ed and disturbed habitats (intensively used meadows, abandoned meadows and pastures, nitrophilous ruderal plant communities, etc.). A similar wide ecological niche is also proposed by Andorkó & Kádár (2009) for a closely related species Carabus scheidleri. It is likely that the mosaic-like landscape in the Eastern Carpathians, with its open and also afforested parts, meets the habitat requirements of this important European species.
This habitat should be the choice for the location of monitoring plots in both forest and non-forest ares in future attempts to map this species' distribution.
According to the offi cial data based on art. 17 HD reporting (EEA, 2021), the population size and trends of this species in Poland are unknown. In Slovakia, the population size is currently estimated at 1000 to 5000 individuals (probably underestimates) and its occurrence is based on 196 1 × 1 km grids, but the trend in population and distribution of this species are unknown. In Hungary, it occurs in 81 1 × 1 km grids, conservation status is listed as favourable and trends are classifi ed as stable, however there is no detailed information on its population size and therefore most probably the results are based only on the judgement of experts rather than on recently collected fi eld data. In Romania, both the trend in and population size of this species are unknown and the population based 1000 1 × 1 km grids is very likely to be an overestimate as there are only a few records for a few localities. Based on this information it is clear that our knowledge of this species is inadequate and the assessment of its conservation status in the 27 EU states is questionable and should be improved/corrected in the future.
Lack of data is a general problem and clearly there is a need to gather new information from across the entire Carpathian range. For example, information on the recent distribution of this species in Ukraine comes only from very limited set of localities (Barloy et al., 2014) and its distribution is based only on old data (Rizun, 2003).

CONCLUSION
There is little and in some cases contrary information on the ecology of the rare species Carabus zawadzkii. This research contributes by presenting information regarding its ecological niche. Carabus zawadzkii is a species of European interest and is listed in Annex 2 of the HD (a species for which a Natura 2000 protected area is declared), which makes the need for research and new knowledge even more pressing. The results presented indicates that it evident that Carabus zawadzkii not only occurs in forest, but also in non-forest habitats. They also indicate a need for EU member states to adopt more appropriate monitoring systems and establish more monitoring plots based on these new fi ndings, especially in Slovakia, Romania, Hungary and Poland. The results also provide important information for improving the management and conservation status of this species. Particular attention should be paid to the implementation of appropriate measures at Natura 2000 sites where this species is protected, as well as in other areas where this species occurs naturally. Regarding suitable management regimes in non-forest habitats, extensive mowing seems to be the most benefi cial and in fact seems to be the most important factor for the survival of this species in addition to the presence of adequate forest ecosystems. This study also provides insights into the seasonal dynamics, showing that it is most active in May, followed by July. Despite the improvement in the knowledge of this protected species provided by this study, it is necessary to do further research on its distribution and ecology.