Introduction

Biodiversity conservation of many species is hindered by the fast habitat loss and transformation due to human activity, being it for economic interest or direct occupation and destruction of the habitats (Ceballos and Ehrlich 2002; Arora et al. 2018). The accelerated expansion of industry and agriculture is invading habitats of numerous species in both rural and urban areas (Antrop 2004; McDonald et al. 2008). During this process, some mammal species were forced to move to new habitats or to adapt to the habitat fragmentation which limited their movement capacity, reducing their distribution in areas with high human presence (Tucker et al. 2018). Between 1900 and 2015, 177 mammal species have gone extinct. The rate of population loss in terrestrial vertebrates is extremely high even in “species of low concern” (Ceballos and Ehrlich 2017). In this context, natural patches embedded in urban environments, like parks at road borders or urban gardens, can maintain important wildlife populations (Dickman 1987; Bland et al. 2004; Angold et al. 2006). In Santini et al. (2019), the authors indicate that urban environments act as filters on mammal communities by selecting a variety of “winning traits” like high reproduction rates or behavioural and ecological flexibility which allow certain species to persist within these environments.

Although there are numerous studies on the impact of habitat changes in many species, only scarce information is available in the Iberian Peninsula about the response of two native and sympatric hedgehog species: the European hedgehog (Erinaceus europaeus Linnaeus, 1758) and the Algerian hedgehog (Atelerix algirus Lereboullet 1842). Indeed, there are only a few studies on the European hedgehog, mainly in the UK (Anouschka and Paul 2009; Rondini and Doncaster 2002; Driezen et al. 2007) and Italy (Boitani and Reggiani 1984; Mori et al. 2015).

The European hedgehog is an insectivorous mammal occupying a large portion of middle and west Europe (Burgin et al. 2020). In the Iberian Peninsula, it occupies nearly all provinces being present in different territories with a wide thermal range. In Atlantic, Iberia occupies a very wide range of habitats, both open and wooded, and is probably most abundant in the Atlantic countryside. Its abundance decreases with altitude, although its presence is common below 1000 m. In the Mediterranean area, it is scarcer and prefers the more humid areas, such as forests. It also prefers mountainous areas and semi-urban environments, including gardens (Nores 2007). The Algerian hedgehog, originally from North Africa, is distributed from Mauritania to Libya, including the Djerba island (Burgin et al. 2020). In the Iberian Peninsula, it presents a distribution restricted to the Mediterranean Coast and the inland of Tarragona, areas with a dry and warm climate (Alcover 2007; Morales and Rofes 2008). In the Valencian Community, the distribution of both species overlaps (Crespo 2012). Habitat preferences of the Algerian hedgehog differ from those of the European hedgehog, the former preferring more open areas like cultivated croplands, almond and olive plantations, pine forests and scrublands (Ruiz-Romero 1995).

Due to the anthropization of their distribution ranges, periurban areas represent for both species a diverse mosaic of resourceful habitats. Although these habitats seem favourable on the first place, they also mean new threats when structures easing hedgehog mobility between habitats are not available (Puig et al. 2012). Some of the threats described for the European hedgehog in periurban areas are: predator attacks (dogs, foxes), rodenticide and plaguicide intoxications, loss of forest ecotones or green surfaces in urban areas and road mortality, the latter being the principal threat (Dowding et al. 2010; Huijser 1999; Spinozzi et al. 2012). In 2016, in the Valencian Community, 69 accidents involving hedgehogs were reported to authorities, situating hedgehogs as the third most overrun mammal (Dirección General de Medi Natural i d´Avaluació Ambiental 2017). In the Spanish inland, up to 1.71 dead hedgehogs per road kilometre were recorded in 6 months (Garnica and Robles 1986). Furthermore, it was described the near total disappearance of Algerian hedgehog populations in south France and a remarkable decrease in Iberian populations, with the only non-natural cause of death described being road kills (Alcover 2007).

There are studies of post-release hedgehog monitoring in other countries (Morris 1997; Bunnell 2002; Molony et al. 2006); in Spain, there is only one study evaluating home ranges of released European hedgehogs (Cahill et al. 2011). Until now, no study compared the spatial behaviour of the European and the Algerian hedgehogs in the same habitat. The coexistence of both species may give rise to additional problems such as possible competitive relationships between the two species that could further limit the use of available habitats. Possible sources of competition could be feeding on the same prey pool (mainly insects), using the same refuges for hibernation or breeding, possible territorial behaviours, etc. Additionally, an increase in mobility, for example, motivated by territorial behaviours, could increase the risk of death by car accidents as roads limit hedgehog dispersal.

The aim of this study is to analyse the differences in spatial behaviour and habitat use between sympatric European and Algerian hedgehogs in a periurban environment. To do so, we analyse their home ranges and travelled distances, obtained by radio tracking of individuals. Furthermore, we evaluate the effect of translocation on home ranges and habitat use, comparing translocated and indigenous individuals of both species. We hypothesized that the two species show habitat and home range segregation, in order to avoid competition. Additionally, we expect translocated individuals to have larger home ranges and higher activity levels than indigenous ones, due to the “release effect” in an unknown environment (e.g. Cahill et al. 2011).

Material and methods

Study area

The study area is located in the suburban area of Valencia City (W 0°22′38.6″ N 39°28′11.1″) in the eastern Iberian Peninsula (Fig. 1). Mean annual rainfall in this area is 445 mm, mainly in April and October, with a mean annual temperature of 17.4 °C (Pérez-Cuevas 1994). The study area is 447 Ha and is located in a periurban area of the municipality Godella. We distinguished six habitat types: high-density Urban (HDU), characterized by actively used human settlements; low-density urban (LDU), areas like abandoned constructions, generally with higher abundance of bush and scrub vegetation; abandoned crops (AC), former croplands with a dense bush vegetation; crops being used (CU), terrains with an active production of oranges with a continuous human presence; pinewood (PW), forests dominated by Pinus halepensis and scrub; and ravine (RV) an abrupt lowering of the terrain level with a dense bush vegetation, where the water from the seasonal rainfall runs off.

Fig. 1
figure 1

Location of the study area (1) and releasing point of translocated individuals (2)

Full size image

Fieldwork

Between March 2017 and May 2018, we captured a total of 18 indigenous male hedgehogs (10 Algerian and 8 European) and released 13 males from the wildlife centre “La Granja” in El Saler (4 Algerian and 9 European). The latter were taken to the wildlife centre for multiple reasons and were originally from the metropolitan area of Valencia. Rehabilitated animals were always released in optimal health conditions after a recovery period. Females were not included because their home ranges are significantly reduced during the breeding season, which would add a bias to the results. We only selected individuals from urban or periurban areas for this study. For the capture of indigenous individuals, we carried out nocturnal transects and used traps. The transects consisted in walking systemically over the whole area with bright flash lights (Nitecore, 1000 lumens), capturing by hand every hedgehog detected with the flash light. We began the transects 1 h after sunset and ended after a maximum duration of 3 h. Additionally, we used 12 Havahart rabbit traps with 81 × 25 × 30 cm, baited with peanut butter (following Riber (2006) and Hof and Bright (2009)) which were placed every sampling night and distributed equidistantly on fixed points, covering the whole study area. Traps were installed at dusk and removed at dawn (Haigh 2011).

All captured or recovered individuals were sexed and weighed using a digital scale (PATISZON, with a precision of 0.01 g). Then, we marked them in order to avoid repeated sampling, combining the application of an individual code of differently coloured solvent-free paint (Titanlux) applied to the tips of the spines and the attachment of two coloured thermoretractile tubes (small plastic tubes that deform with heat) on the spines, which allowed the identification of the hedgehogs from afar, without the need of capturing them (Campbell 1973). We also attached a VHF radio transmitter (Pip3, Biotrack), gently cutting the spines in the dorso-lateral area and applying an adherent (masilla Ceys) between the spines and the emitter. The emitter weighted always less than 4% of the individual’s body weight, following Young et al. (2006). The weighting, sexing and marking of the individuals were done when the stress derived from the capturing allowed the manipulation of the hedgehogs and they were in good conditions (not rolled up and/or emitting hissing sounds), following the criteria of Bunnell (2002). After the attachment of the emitter, we released the individuals. Indigenous individuals were released at the corresponding capture point. Translocated individuals were released all at the same point (2 in Fig. 1) in the study area, near sites with food and water supply for local free ranging cats, maintained by a local animal protection organization (Fig. 1). This release area was chosen as it has a high connectivity to other habitats and there are already established populations indicating that it is a suitable habitat for their survival.

We monitored the released individuals on a daily routine (once per day), using a Yagi directional antenna, during 1 month after release or until emitter failure or death of the individual. In all localizations (tracking events), we annotated the UTM coordinates, using a GPS (Garmin etrex 20). Part of the localizations (9%) was obtained during the day, in order to detect burrows and to have a starting point for the nocturnal monitoring. The rest of the localizations (91%) were obtained during the night, when the hedgehogs were active. In both cases, the approximation to the radio tracked individuals was done by the homing technique (White and Garrott 1990), as the terrain and the behaviour of the individuals allowed to follow them until visual contact was possible, granting an exact localization.

Data analysis

We estimated home ranges by the traditionally used minimum convex polygon method, using all points (MCP 100) (Mohr 1947; Abu Baker et al. 2017) and also by the Kernel method, using 90% of the points (Kernel 90) and 50% for the core area (Kernel 50) (Worton 1989), using the default (quartic) smoothing factor. We estimated the home range overlap as the percentage of the shared area, in relation to the total area used by each of the individuals (Millspaugh and Marzluff 2001). Additionally, we calculated the mean distance travelled per hour, standardizing the distance between two consecutive localizations by the time in hours passed, as a measure of individual activity. The differences in home range size and travelled distances were analysed fitting a generalized linear model (GLM) (Zuur et al. 2009). Because the Shapiro-Wilks normality tests showed that home range sizes and travelled distances are not normally distributed (Shapiro-Wilks test, Kernel 50: W = 0.56 p < 0.01; Kernel 90: W = 0.63 p < 0.01; MCP 95: W = 0.63 p < 0.01; MCP 100: W = 0.64 p < 0.01; distances: W: 0.81 p < 0.01), we fitted a GLM with Poisson error distribution and a log linking function, considering the following predictors: season (spring (21 March–21 June), summer (22 June–23 September), autumn (24 September–21 December) and winter (22 December–20 March)), origin (indigenous and translocated), species, body weight and the interactions species*season, species*origin and species*weight. Furthermore, we evaluated the correlation between the individual weight and extension of home range performing a Spearman correlation test.

To study how habitat use differs between species and origins, we estimated the used habitat by calculating the proportion of the surface each habitat type occupied in each Kernel 90 and Kernel 50 home ranges in relation to the total surface of each home range. We considered the home ranges estimated by the Kernel method rather than the MCP, as the latter is more affected by outliers. Then, for each habitat type and kernel home range (90 and 50), we fitted a GLM with normal error distributions and identity linking function, with the proportions of the used habitat as dependent variable, and considering the following predictors: species, origin and the interaction of both. To do so, we previously normalized the proportions by the arcsine transformation (Zar 1999). We tested in all GLMs the parametric assumptions of residual normality and homoscedasticity with a Shapiro-Wilks and a Breusch-Pagan test, respectively. We also fitted a resource selection function (RSF) following Manly et al. (2002), randomly generating a number of points in each Kernel 90 home range area equal to the number of locations (used) to estimate the home range and assigned the habitat type each point or location was in. These random points served as a measure of the habitat availability. Then, we fitted a binomial GLM with the used/random as dependent variable and the habitat type as predictor. Home range estimations and habitat proportion calculations were carried out using the software QGIS (2021) v2.18 with the plugin density analysis. All the statistical analyses were ran using the statistics software R v4.0.3 (R Core Team 2020). The packages used were “lme4” for the GLM fitting and “lmtest” for the testing of parametric assumptions.

Results

From the 31 marked individuals, two lost the emitter and one Algerian hedgehog disappeared without a trace; five individuals died (four translocated and one indigenous). After the necropsies, we determined that all translocated individuals died from lung damage and the indigenous individual was found crushed by a rock (Table 1).

Table 1 Mean ± SD of number of locations per individual (L), weight, home range size and mean activity for each species and origin. Also, the number of monitored individuals (N), the range of the number of locations per individual, the number of individuals with lost emitters and individuals which died during the study
Full size table

Home ranges

The GLM for the MCP 100 showed a significant effect for the predictors: season and origin and the interaction species*weight. The MCP 100 were largest in spring and smallest in winter (Table 2). Translocated individuals presented greater MCP 100 (10.96 ± 4.88 Ha, N = 13) than indigenous ones (7.20 ± 1.92 Ha, N = 18), independent of the species (Table 2). The correlation between the extension of the MCP 100 and the weight differed between species, being positive and significant for the European hedgehog (rho = 0.48, p = 0.04, N = 14) and not significant for the Algerian hedgehog (rho = 0.53, p = 0.53, N = 17).

Table 2 Results of the GLMs for each home range. Significant results (p < 0.05) are highlighted in bold. Below also the mean ± SE of home range size for each species in each season, together with the number of monitored individuals (N) and the mean ± SD of number of locations per individual (L)
Full size table

In the case of the Kernel 90, the significant predictors were the same as for the MCP 100. The Kernel 90 was largest in spring and smallest in winter, independent of the species (Table 2). Translocated individuals presented larger Kernel 90 (25.60 ± 11.79 Ha) than indigenous ones (14.92 ± 3.77 Ha) (Table 2). Furthermore, the correlation of the Kernel 90 and the weight was positive for both hedgehog species, especially in the European hedgehog (rho = 0.56, p = 0.03 for Algerian hedgehogs and rho = 0.68, p < 0.01, for European hedgehogs) (Fig. 2). Finally, regarding the core area (Kernel 50), the GLM showed a significant effect only for the predictor season, being largest in spring and smallest in winter, independent of the species (Table 2).

Fig. 2
figure 2

Correlations between the Kernel 90 (upper graphic) and activity (lower graphic) with weight for the European hedgehog (triangles and continuous line) (N = 14) and the Algerian hedgehog (circles and dashed line) (N = 17)

Full size image

The home range overlap in the case of the Kernel 90 was higher for the Algerian hedgehog (82.9%) than for the European hedgehog (31.2%) (Table 3, Figure S1) and higher for indigenous individuals (90.2%) than for translocated ones (36.2%) (Table 3, Figure S1). The overlap of the Kernel 50 followed the same pattern, being higher for Algerian (40.9%) than for European hedgehogs (14.9%) and higher for indigenous individuals (83.3%) than for translocated ones (28.7%) (Table 3, Figure S1).

Table 3 Home range overlaps for each species and origin
Full size table

Distances travelled

The GLM for the mean distance travelled showed a significant effect of all the interactions of the predictors season, origin and weight with species. For the seasons, the European hedgehog presented maximum activity in spring, while the Algerian hedgehog presented the maximum in summer (Figure S2). Regarding the origin, in the case of the European hedgehog, translocated individuals presented a greater activity level (22.36 ± 6.79 m/h, N = 9) than indigenous ones (14.30 ± 4.35 m/h, N = 8), while in the Algerian hedgehog, it was the contrary, translocated individuals presenting a lower activity level (13.72 ± 4.65 m/h, N = 4) than indigenous ones (32.15 ± 7.60 m/h, N = 10). Finally, the correlation of the mean weight was positive for the European hedgehog (rho = 0.68; p < 0.01) and negative for the Algerian hedgehog (rho =  − 0.56; p = 0.06) (Fig. 3).

Fig. 3
figure 3

Mean ± SE of the proportion (in %) of each habitat type in the Kernel 90 and Kernel 50 home ranges for each hedgehog species, considering translocated (black) and indigenous (grey) individuals separately

Full size image

Habitat use

In the Kernel 90, the GLM for the HDU habitat showed a significant effect of the interaction species*origin. In both species, translocated individuals used HDU less than indigenous ones; however, this difference was more important in the Algerian hedgehog than in the European hedgehog (Table 4, Fig. 3). For the PW habitat, the significant predictors were origin and species. In both origins, Algerian hedgehogs used PW more than European hedgehogs, and in both species, translocated individuals used PW more than indigenous ones (Table 4, Fig. 3). For the RV habitat, the GLM showed a significant effect of the interaction species*origin. In both species, translocated individuals used RV less than indigenous ones; however, this difference was more important in the Algerian hedgehog than in the European hedgehog (Table 4, Fig. 3). In LDU, CU and AC, neither of the predictors showed a significant effect on habitat use.

Table 4 Results of the GLMs for the proportions of the different habitat types in the Kernel 50 and Kernel 90 home ranges. Significant results (p < 0.05) are highlighted in bold
Full size table

Regarding the core area (Kernel 50), the GLM for the HDU habitat showed a significant effect of the interaction species*origin. In the case of the European hedgehog, translocated individuals presented a higher use than indigenous ones, while in the Algerian hedgehog, translocated individuals presented a lower use than indigenous ones (Table 4, Fig. 3). The use of AC differed between species, being more used by the Algerian hedgehog than by the European hedgehog, independent of the origin (Table 4, Fig. 3). In the other habitats, none of the predictors showed to have a significant effect.

The analysis of habitat preference by the RSF showed a clear tendency to avoid HDU (Table 5).

Table 5 Results of the resource selection function for each habitat type with the total number of locations (used) and randomly generated points (random), together with the percentage respect the total number of points in brackets. The results are shown joining both species and for each species separately. Significant results (p < 0.05) are highlighted in bold
Full size table

Discussion

Home ranges

We did not find differences in home range sizes between species. This agrees with other studies carried out separately for each species (García et al. 2009; Cahill et al. 2011; Marco and López 2015). Some studies recorded different home range sizes of the European hedgehog in similar periurban areas in other countries. For example, Rondini and Doncaster (2002) described in the UK smaller mean home ranges (5.00 ± 0.7 Ha) than in our study. However, it is known that the home range size of the European hedgehog is highly variable, ranging from 5.5 to 102.5 Ha (Boitani and Reggiani 1984). Thus, the results have to be taken with caution, as they depend on the local environment, number of localizations and monitored time period.

Supporting our hypothesis, we found a clear tendency of translocated hedgehogs to present larger home ranges than indigenous ones. This can be explained by an increased activity and mobility of the individuals when introduced to an unknown environment. This tendency was also observed in Cahill et al. (2011), where translocated individuals had a mean home range of 55.3 ± 117.0 Ha, with values ranging up to 465 Ha. In other studies, translocated individuals ended up using similar home ranges as indigenous ones, adapting to the new environment without problems (Morris et al. 1992). As we tracked the hedgehogs for a period of 30 days, we could not statistically evaluate the adaptation; however, we did observe that translocated individuals tended to use the same feeding grounds and burrows as indigenous individuals, without apparent intraspecific competition. There is no evident territoriality described for these species. However, there are models like the ideal non-territorial despotic distribution (Fretwell and Lucas 1970) which postulates that competitors are free to move to feeding grounds, if they maintain a certain distance from other individuals. In this constellation, bad competitors could not access the best feeding grounds if they are saturated with good competitors. In the study of Cassini (1996), the author empirically proved this model in free-ranging European hedgehogs in meadow environments, concluding that, under certain circumstances, the density of hedgehogs increases proportionally to food availability. The results of this interaction are dependent on the ability of the bad competitors to move to other territories or shift their diet. Furthermore, in Haigh et al. (2012), the authors showed that hedgehogs did not select their habitats in relation to the availability but presented a philopatric tendency, showing a repetitive annual pattern of habitat selection.

Home ranges were generally larger in spring, coinciding with the end of hibernation and the mating season (Nores 2007). Between March and June, hedgehogs show a great activity related to the search of food and mates. The fact that we exclusively monitored males induces a bias in the results, as males travel larger distances seeking females (Rautio et al. 2013), presenting increased mobility and home range sizes in comparison to females or pups (García and Puig-Montserrat 2014). This is a pattern commonly observed in mammals, including the European hedgehog (Riber 2006; Haigh 2011). According to the known phenology, it was to be expected that the winter is the period with the lowest activity, as hedgehogs reduce their activity from November to December, starting hibernation. In very anthropized environments, like our study area, human activity can alter the phenology. In previous studies, it was demonstrated that the hibernation of hedgehogs can be interrupted when disturbed (Blanco 1998; Nores 2007) or lead to a facultative hibernation (de Miguel et al. 2018).

Interestingly, independent of the method used to estimate home ranges, we observed a clear positive correlation between home range size and weight, being more important in the European than in the Algerian hedgehog. This correlation was not observed for the core areas. A possible explanation of this correlation could be that the size of the home range is directly linked to the energy requirements of the individual, larger individuals presenting increased energy requirements. Thus, large individuals need to range over large areas to satisfy their higher requirements, but maintaining the size of the core areas similar to small individuals (Lindstedt et al. 1986). Our study did not include young individuals and all translocated individuals weighted over 600 g; however, our personal observations indicate that translocated individuals are reproducing in the study area. This represents a good opportunity to describe the movement pattern of young individuals of a translocated population, recording information about important biological aspects like the dispersal and spatial behaviour of these individuals. This was done in other species, with very interesting results (Rasmussen et al. 2019). Thus, we strongly suggest to carry out a similar study for both hedgehog species.

Regarding home range overlaps, our results evidence a greater overlap in the Algerian than in the European hedgehog, in both their total home range and core areas. Currently, there is no study focusing on the simultaneous distribution of both sympatric species in a small spatial scale, where a possible interspecific competition could be observed. However, there are references for other mammal species in the Iberian Peninsula, like Martes martes or Genetta genetta, where in sympatric areas, each species is distributed separately, showing differences in habitat use and a trophic resource partition (Mangas et al. 2007). Both studied hedgehog species are described as opportunistic insectivores, feeding on arthropods, worms and even small reptiles, mammals or bird eggs (Nores 2007; Sayah et al. 2009). However, several studies analysing the diet of the European hedgehog indicate a possible facultative herbivory (Hernández 2006; Rautio et al. 2016). This can be indicative of the species’ generalism, which is also reflected in the home range overlaps. The fact that the Algerian hedgehog greatly overlapped its home range between individuals can be related to the remarked habitat selectivity, while the generalism of the European hedgehog lead to a use of a much broader range of habitats and, thus, lower overlap values between individuals. Finally, translocated hedgehogs showed less overlap than indigenous ones. In accordance with the results of home range size, the overlap can also be explained by the increased exploratory spatial behaviour of the individuals in unknown environments.

Travelled distances

We found phenologic differences between species regarding mean travelled distances, the European hedgehog was more active in spring and the Algerian hedgehog in summer. Such differences were not described in previous studies, but they coincide with the moment when home ranges are maximum. The obtained mean travelled distances are similar to those obtained in studies where localizations were taken each hour (Morris 1988). Furthermore, our results indicate that translocated European hedgehogs moved larger distances than indigenous ones, contrary to the Algerian hedgehog. Some studies show that the trajectories of translocated hedgehogs are highly variable in both length and direction, supporting that they explore the unknown environment instead of following a fixed set of spatial behaviour rules (Doncaster et al. 2001). Comparative studies of translocated and indigenous animals evidenced that non-indigenous individuals dispersed over greater distances than indigenous ones, in accordance with our results for the European hedgehog. The fact that we did not observe this pattern (and even observed the opposite) in the Algerian hedgehog can be related to the aforementioned habitat selectivity of the species, tending to disperse less or being more cautious when exploring the territory. Finally, we also found a positive correlation of activity level and weight in both species. This agrees with our home range results. Increased energetic requirements of heavier/larger hedgehogs would also lead to longer travelled distances.

Habitat use

We found a clear tendency of the hedgehogs to use preferably pine forest areas and to avoid all other habitat types, being the HDU the most avoided habitat. These results contrast those from other studies in the Iberian Peninsula in the case of the Algerian hedgehog, where they prefer open croplands with a certain degree of vegetation (García and Puig-Montserrat 2014). In fact, in our case, the Algerian hedgehog presented an even more marked tendency to prefer pine forests in their Kernel 90 and ravines in their core areas than the European hedgehog. This could be explained by a possibly better balance of food or refuge availability or accessibility and a lower predation risk in these areas in comparison to croplands. Furthermore, the differences in habitat selection between both hedgehog species show that the European hedgehog tends to use a wider range of habitats, from forests to anthropized environments (Reeve 1994). Nonetheless, the preference of open habitats was also described for the European hedgehog in previous studies (Riber 2006; Young et al. 2006; Shanahan et al. 2007). This, again, suggests that the European hedgehog is more generalist, while the Algerian hedgehog is more specialist. This is also supported by the results in Castro et al. (2007), where the authors ran predictive models in order to compare favourable areas for both species in the Iberian Peninsula.

The preference for pine forests can be related to the presence of abundant vegetal cover and natural refuges in logs, stumps or root formations, mainly absent in the other habitats (LDU, HDU, AC, CU), combined with a lower degree of human disturbance. The adaptation of hedgehogs to suburban environments was demonstrated in other countries like the UK (Reeve 1994; Harris 1984; Rondinini and Doncaster 2002), but a key factor in areas with high human presence is permeable residential garden. Some studies demonstrate how the patterns of use of the gardens were affected by the characteristics of the general environment (city, village, rural), decreasing with the increase of the degree of urbanization in most species, including hedgehogs (Baker et al. 2007). This explains the avoidance shown by our individuals of areas with human presence, as it is a very urbanized environment and the hedgehogs are only able to use gardens with permeable structures.

We did not find traces of natural predators in the study area, like badgers or foxes, but we registered the presence of dogs, with a variable density. This agrees with studies revealing that hedgehog densities depend on those of predators; thus, the shown habitat preference could reflect variable predator presence (Rondonini and Doncaster 2002).

We also want to remark the accessibility of the individuals to the different environments. From their core areas, mainly presenting pine forests and low-density urban areas, the individuals always had feeding grounds (cat food for the local free ranging cat populations) to their disposal, which could be attracting them to these areas. In order to travel between urban habitats (LDU and HDU) and between pine forests (PW) and croplands (AC and CU), the hedgehogs needed to use the ravine (RV) habitat, as walls and highly frequented roads act as barriers (Rondinini and Doncaster 2002). The fragmentation of the habitat is an important problem for these species (Berger et al 2020; Gazzard et al. 2021). Some authors emphasize the importance of conserving beneficial agroenvironmental systems, acting as feeding points for hedgehog populations (Hof and Bright 2010).

Regarding the effect of translocation on habitat use, we found a more marked tendency of translocated individuals to use the pine forest habitat and to avoid ravine and high density urban areas, compared to indigenous hedgehogs. Some studies suggest that the presence of torrents favour longer travels, as they act as corridors, given their dense vegetation of bramble and reeds serving as refuge (Cahill et al. 2011). In our case, although the individuals did not show a preference for this habitat, we could see that they use it to travel between different environments. However, several weeks after release, the individuals adapted and used the same feeding grounds and resting points as indigenous hedgehogs.

In conclusion, our results do not support the first hypothesis. In a sympatric environment, both species tend to cohabit without presenting an apparent spatial competition, commonly using habitats offering protection by the vegetation, major refuge availability and lower degree of disturbance. Algerian hedgehogs showed a more specialist character regarding habitat selection than European hedgehogs and a tendency to move less when introduced into a new environment. We could confirm the second hypothesis, translocated individuals presenting greater home ranges and activity levels than indigenous hedgehogs. Our results also highlight the need to analyse in greater detail which habitat structures act as corridors, increasing the connectivity and which ones act as barriers. We also suggest to continue monitoring the population densities of both species in similar environments, allowing to estimate the effect of habitat fragmentation on sympatric populations. Finally, this study also shows the need to monitor young individuals, as we expect interesting results regarding the dispersal behaviour of the species in fragmented habitats.