First of all I would like to thank Prof. B. Nievergelt for his friendship and for sharing his love for nature with me. The field trip to the Simen Mountains National Park in Ethiopia (Jan.-Feb. 1996) allowed me to get to know the people and projects of the department and to quickly feel at home here. My warmest thanks goes to Karin Hindenlang,who helped me organise my diploma and who supported me through all stages of it. In particular, I would like to express my gratitude to Stefan Imfeld, who generously helped me not only with all questions considering the GIS analysis, but with all other problems that arose during the statistical analysis of my data. I was extremely grateful for his critical remarks and moral support. Much thanks goes to Susanne Reimann for her flair for computers as well as for her talent as art director. Furthermore, I would like to thank Lorenz Gygax and Wolf Blankenhorn for the statistical advice and Gérald Achermann and Stefan Imfeld for helpful comments on the draft. Special thanks to André Utzinger for English corrections and the nice time at the hut. I would especially like to thank Mr. John Potts for proof-reading the manuscript.

I am grateful to many people who helped me find numerous setts: Gérald Achermann, Iris Bättig, Nina Debrunner, Oliver Fischer, Michi Griesser, Stefan Imfeld, Guido Keller, Isabelle Minder, Marco Muhmenthaler, Thomas Nabulon, Josef Niederberger, Dani Wirz and Matthias Wüst.

Exploring the relationships between animal behaviour and ecology is very promising when a wide variability exists in a species’ social and spatial organisation: badgers (Meles meles) adapt their social and spatial organisation to different environments: in areas with a predictable food availability, badgers live in groups, defend territories and occupy a main sett; in areas with a wide seasonal variation in food availability, badgers are solitary, have large overlapping home ranges and use several setts within a range (Cheeseman et al., 1988; Cresswell & Harris, 1988; Skinner & Skinner, 1988). This extraordinary adaptability has undoubtedly contributed largely to their success as a species. Within the Swiss National Foundation grant 3100-40846.91 "Adjustment of spatial behaviour of badger (Meles meles L.) in response to a heterogeneous environment", a study on the European badger concentrates on how variation in resource availability between individual home ranges in the study area is reflected by variation in individual resource utilisation. In order to understand a part of the mechanisms of adjustment of the social and spatial organisation of badgers in different environments, this study focussed on the distribution and utilisation of setts and latrines in Sihlwald.

Badgers are found all over Switzerland up to 2000 m a.s.l.. [ISS: figure omitted]. The badger has adapted to various habitats, but prefers the lowlands and hills. Its distribution in higher regions is limited by the absence of forests, fields and pastures (Graf, 1995). In contrast to the widespread and heterogeneous home ranges, the habitat requirements of the sett sites are rather specific.

As shown in the literature, the distribution and density of badger setts vary considerably between different countries and also between different locations within one country. There are many different factors affecting the distribution and density of badger setts, e.g. the type of soil, the amount of cover and the hilliness of the terrain (Neal, 1986). However, local conditions, particularly available food such as earthworms, also have an important modifying effect on sett choice (Skinner et al., 1991). Thus, the distribution of badgers is strongly determined by how well a certain area fulfils the following ecological requirements:

Setts vary greatly in size, the extent of the tunnels and the number of entrances. A number of them have been excavated and their structure compared (Roper, 1992). The size of a sett is not necessarily determined by the number of entrances, nor does sett size indicate the number of badgers living in it (Neal, 1977). Bock (1986) thinks that the volume of the spoil heaps is probably the best criterion to characterise and quantify a badger's sett.

Various attempts have been made to classify setts according to size and function (Kruuk, 1978; Thornton, 1988). In the studies carried out in Britain, the most important sett in a territory is termed "main sett" which is usually large, is used throughout the year and acts as the main breeding sett. In contrast, there are small setts with one or two holes and a simple, often unbranched tunnel system. They are used only occasionally as temporary refuges and are called "outliers". There may be several of these within a territory. Intermediate levels of size and complexity are found between these two extremes. Setts permanently in use may indicate territoriality; setts used only at certain times of the year (as winter refuge or in areas with a big seasonal variation in food availability) may point to a spatial organisation of extensive, overlapping home ranges.

Beside badgers, the most common sett tenants in Switzerland are foxes (Vulpes vulpes L.). Foxes occasionally dig their own den, but, being lazy diggers, they prefer badger setts when available (Weber, 1990). The fox is a surface dweller that requires secure cover below ground for rearing cubs. After the weaning of the cubs, setts are only frequented by foxes if the weather is harsh or very wet, or where surface cover is scarce (Weber, 1985). Typically, they use unoccupied setts, but in large setts share the spacious accommodation, usually using the perimeter holes (Neal & Cheeseman, 1996). For this reason, setts in Sihlwald were divided into the categories "used by badger", "used by a fox" and "used by badger and/or fox" (see Appendix, Tab. 13).

A latrine consists of one or more dung pits about 15 cm deep. When a pit is full of dung, it is left uncovered and another dug nearby. One can often find several dung pits covering an area of 2-4 m2. Occasionally, instead of a latrine with many pits, a single considerably larger one may be found containing much dung. Sometimes the dung is deposited on the surface, not in pits. Underground latrines found in excavated setts appear to be used when the badgers stay below in periods of very low temperature (Roper, 1992).

Kruuk (1978a) defined two types of latrines: those found near a sett and those that occur in other parts of the territory. The latrines near the border of ranges have territorial significance and are usually larger than those near setts. With regard to utilisation, Roper et al. (1986) distinguished two different types of defecation sites: latrines (several dung pits dug within an area of a few square metres and used more or less all year-round) and temporary defecation sites (one single dung pit in any one month; faeces found only during one or two months of the year). The former are usually located around territory boundaries, the latter elsewhere in the range. Some are near obvious landmarks such as hedgerows, fences or woodedges (Kruuk, 1978a; 1989; Roper et al., 1986).

Given the above-mentioned variability of the badger’s spatial and social organisation, knowledge about the distribution of setts and latrines may help to determine whether badgers in Sihlwald live in groups and occupy a main sett, or whether they are solitary and frequent different setts within the range. With this in mind, the following questions were examined for the habitat analysis of badger setts and latrines:

The field work (March-August 1996) consisted in searching the study area all over for setts and latrines, using paths, brooks and forest divisions as borderlines. It is expected that most of, if not all, the existing setts were found. Owing to the often dense vegetation cover, it is assumed that only a small fraction of the latrines was actually discovered.

The definition of a badger sett in the Protection of Badgers Act (1992) in Britain is "any structure or place which displays signs indicating current use by a badger". Since this definition is rather vague, I defined a sett as at least one entrance leading more than two metres underground. Usually a large mound of excavated soil can be found outside each entrance. If two entrances were further than 25 m apart, they were noted as two separate setts. This definition is independent of the species using a particular sett, since in most areas no distinction could be made between setts used by badgers and those used by foxes. To analyse the distribution of the setts in Sihlwald, Kruuk’s (1978a) main sett definition (>2 entrances) was used to calculate the nearest neighbour distance between "main setts". Furthermore, five different sett types were distinguished: setts dug directly into the ground (Fig. 3), under a boulder/rock (Fig. 4), under a spruce (Picea abies) (Fig. 5), under a deciduous tree and under a tree stump.

Latrines are aggregations of one to several defecations, deposited in one or in several small pits dug by the animals themselves. These dung pits are 5-10 cm deep, sometimes as much as 30 cm, and can contain several different defecations (Fig. 6). However, a few latrines have no pits, and some have all droppings in one large pit (Kruuk, 1978a). Since no distinction could be made in Sihlwald between empty dung pits and pits dug in search of food (beetles), a latrine had to contain dung in at least one pit to be defined as such.

Important habitat requirements as mentioned in chapter 1.1, as well as habitat parameters indirectly influencing sett and latrine sites, were recorded for each sett and latrine (see also Appendix , Tab. 12-14). To determine the availability of the recorded details, 86 random points were generated and recorded the same way as the setts and latrines. The coordinates of all setts, latrines and random points are given in the Appendix (Tab. 15-17). [ISS: This Appendix can only be viewed with the permission of the author (Password needed)]

• Sett / latrine location: When a sett or latrine was found, its location was marked on the 1:10'000 map of Sihlwald. The coordinates were fed into a geographical information system, GIS, (Arc/Info V. 7.01) at the Institute for Geography at the University of Zurich for further analysis.

• Altitude: The GIS contained a digital terrain model of Sihlwald with which the altitude was calculated.
• Inclination: The inclination was recorded right at the sett or latrine site taking 10 m above and below the sett or latrine as measurement and put into the following categories for further analysis: 0-10°, 11-20°, 21-30°, 31-40°, >40°.

• Aspect: The aspect (compass direction of the slope) of the sett and latrine sites was calculated by the GIS and put into the following categories for further analysis: N (<= 45°), NE (<= 90°), E (<= 135°), SE (<= 180°), S (<= 225°), SW (<= 270°), W (<= 315°) and NW (<= 360°).

• Topography: The topography was taken within a radius of approximately 25 m of the centre of the sett or latrine. Six different types of topography were defined:

• Vegetation coverage: The degree of vegetation coverage (tree-, shrub-, herb- and moss coverage) around the sett was determined after Braun-blanquet (1964): 0=0%; 1=1-5%; 2=6-25%; 3=26-50%; 4=51-75%; 5= >75%.

• Forest parameters: The forest superintendent's office did a survey on upper-, middle- and lower layer coverage; upper-, middle- and lower coniferous layer coverage; vegetation unit and development stage of the forest stands (Stadtforstamt Zürich, 1990). This data can also be found in the GIS.

• Distance to the nearest path & Distance to the nearest water: These distances were calculated with the GIS and put into the following categories for further analysis: ¾ 50 m, ¾100 m, ¾150 m, ¾ 200 m, >200 m for distances to the nearest path; ¾ 50 m, ¾100 m, ¾ 150 m, >150 m for distances to the nearest water.

• Distance to forest boundary with adjoining agricultural zones (i.e. the forest boundary along the "Sihltalstrasse" was not included): These distances were calculated with the GIS and put into the following categories for further analysis: ¾ 100 m; ¾ 200 m; ....; ¾ 900 m; >900 m.

• Geological situation: Data on the geology of the region was taken from the digitised part of the 1:50'000 geology map of the Canton Zurich (R. Hantke, 1967).

• Distribution: In order to test the distribution of the setts against a random distribution, the nearest neighbour distance between the setts was compared with the nearest neighbour distance for 1000 samples of 123 random points in the area.

• Density: To calculate areas with higher and lower sett density than expected from a random distribution, a Bivariate Normal Kernel Density Estimation was used in conjunction with a Monte Carlo simulation (Imfeld, in prep.).

• Habitat analysis: All parameters (GIS and field data) were plotted against each other to see if a linear correlation exists between them. Owing to the weak correlations, no parameter had to be omitted from further analysis.

• Distribution with regard to availability of habitat parameters: To test the random distribution of setts in the GIS habitat parameter categories (Appendix, Tab. 12), a goodness-of-fit test compared the number of observed setts with the number of expected setts for each category:

Among the recorded 123 setts, the number of entrances per sett varied from one to 11 (Tab. 1). More than half the setts had only one entrance; a third of the setts had two, three or four entrances, respectively. The average number of entrances per sett was 2.26. The setts were classified into small (1 or 2 entrances), middle-sized (3-4 entrances) and big (>4 entrances) setts. Small setts were most common (88), followed by middle-sized (24) and big (11) setts.

The distribution of the setts is shown in Fig. 8 [ISS: This figure can only be viewed with the permission of the author (Password needed)]. The average nearest neighbour distance between two setts is 111 metres, compared to the 158 metres for the average nearest neighbour distance between two generated random points. Thus, the observed distribution of the setts is significantly different from random (p<0.001). If the nearest neighbour distance is calculated for the "main setts" (setts with >2 entrances, Kruuk, 1978a) only, then the average nearest neighbour distance is 311 metres, compared to the 314 metres for the average nearest neighbour distance between two generated random points. This distribution of "main setts" does not differ from random.

The density of the setts in Sihlwald (10 km2) is 12.3/100 ha. In Fig. 9 [ISS: This figure can only be viewed with the permission of the author (Password needed)], the light coloured patches indicate areas with a higher density, dark coloured patches areas with a lower density than expected from a random distribution. Compared to the dark coloured patches, the light coloured patches "Sihlzopf", "Wüesttobel", "Schüeppenloch", "Eschentobel" and "Silrain" are very close to the forest boundary and to the adjoining agricultural areas. It is interesting to see that the setts are especially aggregated from "Häuliboden" to "Sihlzopf".

Five different entrance types were distinguished: entrances dug directly into the ground, under a boulder/rock, under spruce trees, under deciduous trees and under tree trunks. As setts may have entrances underground as well as under spruces or under boulders, not the setts themselves, but each entrance was classified as entrance type (Tab. 2). Most entrances were dug directly into the ground ( Fig.3). The next common type was under a spruce tree ( Fig.5). It is remarkable that 31 entrances were under spruces, whereas only 7 were under deciduous trees. 22 entrances led under a boulder ( Fig.4), 12 were under tree stumps.

Since small, middle-sized and big setts did not differ significantly from each other for any habitat parameter, all setts were taken for further analysis. The backward stepwise regression gave the best equation for the following 6 parameters: inclination, topography, distance to nearest path, middle layer-, herb- and moss coverage and explained 79.81% of the variation. Positive "B" values indicate a positive effect, negative "B" values a negative effect on sett distribution. The results show that inclination, distance to the nearest path and convex slopes have a positive effect, but flat areas, gentle and concave slopes, as well as middle layer-, herb- and moss coverage have a negative effect on sett distribution (Tab. 3).

Tab. 3: Habitat parameters that best explain the distribution of the setts. The coefficient labeled "B" is the partial regression coefficient. Postive "B" values indicate a positive effect, negative "B" values a negative effect on sett distribution. The significant p-values (p<0.05) are printed in bold numbers. Multiple logistic regression.

In Fig.9 [ISS: This figure can only be viewed with the permission of the author (Password needed)], areas with a higher sett density than expected from a random distribution are shown as light coloured patches and probably represent ideal sett sites. Therefore, a comparison was made between the setts in those areas and all other setts. However, because of the high sett concentration in the light coloured patches (Fig.9) and because of the fact that these patches are especially aggregated from "Häuliboden" to "Sihlzopf", there is a high degree of dependency between the clumped setts and the general characteristics of the southern range of the study area. Consequently, the following results should be considered with care. As the mean rank shows (Tab. 4), the setts in areas with a higher sett density than expected from a random distribution differ significantly from the other setts in the following parameters: they were found at a lower altitude (p<0.0001), had in general a lower vegetation coverage (p<0.02) and were nearer to paths than the other setts (p<0.0001).

For a more detailed analysis, this chapter will compare the individual categories of the GIS habitat parameters with the availability of those categories within the study area (goodness-of-fit test). The number of setts will be compared with the number of random points for each habitat parameter measured in the field (field data) as well (Chi ²-test for two independent samples). The habitat parameters for which the distribution of the setts is non-random are then analysed with the Chi ²-test in conjunction with a Bonferroni z statistic to find out which parameter categories explain the distribution of the setts in Sihlwald best. Trends are then tested with a Mann-Whitney U-test. It should be kept in mind that, with this number of parameters, the probability of a random significance is high. Fig. 10-21 help to clarify the ecological significance of the parameters. Therefore, the significance of the individual classes is less relevant than the pattern of each parameter. Likewise, the individual parameters are only parts of a more complex system of habitat factors influencing the choice of sett sites in Sihlwald.

The distribution of the setts is decidedly non-random for the following GIS parameters (Tab. 5): altitude (p<0.001) and distance to the nearest path (p<0.001) seem to have the biggest effect on the distribution of the setts. The vegetation coverage in general seems to play a key role. The distribution of the setts also depends on aspect (p<0.02) and distance to nearest water (p<0.05). The parameter categories affecting the distribution of the settts are shown below (Fig. 10-16). No significant difference was attained for the following parameters: upper- and middle layer coverage, upper coniferous layer coverage, vegetation unit and development stage of the forest stands.

Setts are found significantly closer to the forest boundary than random points (Mann-Whitney U-test, two tailed; U=4003, p<0.005). Comparison of the habitat parameters of the setts measured in the field with those of the random points (Tab. 6) shows that setts differ significantly from random points in the following parameters: inclination, topography (both p<0.001), herb- (p<0.05) and moss coverage (p<0.01). No significant difference was attained for shrub coverage. The parameter categories affecting the distribution of the setts are shown below (Fig. 17-21). Categories containing less than 3 data points were pooled.

Tab. 6: Comparison of the habitat parameters of the setts with those of the random points. Only the parameters for which the distribution of the setts is significantly non-random are listed here. chi2-test for two independent samples; df=degrees of freedom.

As shown in Fig. 10, the number of observed setts differs significantly from the number of expected setts (p<0.05) for each altitude category. There are fewer setts than expected below 600 m a.s.l., but more than expected above 600 m a.s.l. (Mann-Whitney U-test, two tailed; U=0, p<0.001).

The observed and expected number of setts do not differ for any aspect category other than the category North, for which fewer setts than expected were found (p<0.05; Fig. 11).

No trend can be seen for lower layer coverage (Fig. 12). Setts are preferably dug in areas lacking a lower- and middle coniferous layer coverage (Fig. 13, Fig. 14), rather than in areas containing a lower- and middle layer coverage (both: Mann-Whitney U-test, two tailed; U=0, p<0.001).

As shown in Fig. 15, there were significantly fewer setts than expected in category £50 m (p<0.05), but more than expected in the category 51-100 m to the nearest path (p<0.05). In general, fewer setts are found £50 m to nearest path than >50 m to the nearest path (Mann-Whitney U-test, two tailed; U=0, p<0.001). No trend can be seen for any category of distance to nearest water (Fig. 16). Significantly more setts than expected were found in the distance category 51-100 m (p<0.05) from the forest boundary (Fig. 17), fewer than expected in the distance categories 301-400 m and >900 m from the forest boundary (p<0.05).The difference between setts found <= 300 m and those found >300 m from the forest boundary is significant (Mann-Whitney U-test, two tailed; U=0, p<0.001).

Setts are rarely dug in flat areas (p<0.05; Fig. 18). They are preferably dug on slopes with an inclination <= 20°, rather than on slopes with an inclination >20° (Mann-Whitney U-test, two tailed; U=38.5, p<0.001). Topography plays a key role in the choice of sett site (Fig. 19): convex slopes are the preferred sett locations (p<0.05) whereas flat areas, gentle and concave slopes are avoided (p<0.05).

As shown in Fig. 20, setts are preferably dug in areas with a sparse herb coverage, rather than in areas with a dense herb coverage (Mann Whitney U-test, two tailed; U=0, p<0.001). The same trend can be seen in Fig. 21: areas without moss are significantly preferred to areas with moss (Mann-Whitney U-test, two tailed; U=0, p<0.001). See Appendix , (Tab.18-29) for exact data on Chi ²-test in conjunction with a Bonferroni z statistic.

The recorded setts vary greatly in size, extent of tunnels and the number of entrances. Possible explanations for these observations are:

• setts of different sizes are used for different purposes (Thornton, 1988);

• setts are larger and have more entrance in regions where the soil is more easily dug (Neal, 1977);

• the size of a sett, especially of a main sett, reflects its age, since badgers tend to go on extending their setts for as long as the occupation continues (Neal & Roper, 1991).

Since neither size of the setts nor extent of the tunnels were examined in this study, the setts in Sihlwald were classified in small, middle-sized and big setts according to the number of entrances. The number of entrances per sett varied from one to 11 with an average of 2.26 entrances per sett. This is well below the average found in the literature: Kruuk (1978a) found main setts with three to 21 entrances and outliers with one or two entrances resulting in an average of 10.5 entrances per sett. In Belgium, the number of setts varied from one to 38 entrances with an average of 13.3 entrances per sett (Anrys & Libois, 1988). This difference can be explained by the fact that this study did not differentiate between fox dens and badger setts. Fox dens usually have fewer entrances than badger setts (Stubbe, 1980). This could be the reason why the average numbers of sett entrances in Sihlwald is lower than the average mentioned by other authors. Nevertheless, badger hairs were found in several setts consisting of one entrance only, indicating the presence of badgers in single entrance setts as well. As mentioned before, it was seldom possible to clearly distinguish between setts used by badgers and those used by foxes. The differentiation was especially difficult for setts with 1 or 2 entrances. Badger hairs indicated that a badger had visited the sett, but gave no indication for how long or how often the particular sett was used. Tracks were only found on freshly dug spoil heaps or around the entrance after it had rained. So, owing to this uncertainty concerning most of the setts, all setts (unless mentioned elsewhere) were used for further analysis.

The setts in Sihlwald were non-randomly spaced. The average nearest neighbour distance was 111 metres. Since too little is known about the utilisation of the different setts, they could not be classifed into main setts and outliers. Nevertheless, in order to compare the nearest neighbour distance between the main setts as defined in Kruuk (1978a), all setts with more than 2 entrances were taken for further analysis. The distribution of the main setts did not differ from random and the average nearest neighbour distance between two main setts was 311 metres. The average nearest neighbour distance corresponds fairly well with that of the main setts in southern England (300 metres, Kruuk, 1978a), although the distribution of those setts was non-random.

The density of setts in Sihlwald is 12.3/100 ha. Clements et al. (1988) think that 10 setts/100 ha is an average sett density. Compared to the low densities in former Czechoslovakia (0.18/100 ha, Pelikan & Vackar, 1978) and in Switzerland (Canton of Neuchatel: 0-0.22/100 ha, Monnier, 1993; Canton of Berne: 0.42/100ha, Graf, 1988) and the relatively high sett densities all over Great Britain (0-26/100 ha, Cresswell et al., 1990), Sihlwald seems to have an average sett density for European standards. However, for Swiss standards Sihlwald has a high sett density.

The areas "Sihlzopf", "Wuesttobel", "Eschentobel", "Schüeppenloch" and "Silrain" have a higher sett density than expected from a random distribution. These areas are also very close to the forest boundary and to the adjoining pastures and agricultural zones (see also 4.6). In the Canton of Zurich, badger setts are mainly found in the richly structured forest- and agricultural zones (Hindenlang, 1992). This mixture of woodland and pastures, woodland and arable land is among the habitat types preferred by the badger (Neal, 1977; Zejda & Nesvadbova, 1983). Blab et al. (1989) mention two possible reasons: 1. the topography in such regions is extremely varied, which is advantageous for building setts on slopes; 2. important food resources (earthworms, insects) are found in higher densities in this diverse landscape. Neal (1986) also thinks that the presence of a plentiful and varied food supply is one of the most important factors influencing the choice of a sett site. If this is true, the areas in Sihlwald with a higher sett density than expected from a random distribution are ideal sett sites. The setts are situated close to pastures and arable land, meaning that badgers can adapt their foraging strategies to seasonal changes in food availability in the different habitats.

The area "Häuliboden" is an exception. It is not close to any forest boundary, yet there are several setts in this area. It seems that the biggest sett there has been used for generations and is an ideal sett site regarding the heterogeneous topography and sparse vegetation cover.

Areas with a lower sett density than expected from a random distribution (dark coloured patches, Fig. 9) are relatively flat, are close to the village "Langnau a. A." and are therefore near frequently used paths. There is also a higher hunting pressure in this region than toward Sihlzopf. Thus, these areas are less attractive sett sites.

This study was centered on finding habitat parameters likely to be relevant to the distribution of sett and latrine sites. Yet, it is often impossible to separate the relative importance of various ecological parameters,as they are obviously inter-related and differ in priority from one area to another. This study probably does not contain all relevant parameters for a sett site for two reasons: 1. it is difficult to know beforehand which parameters are likely to be relevant; and 2. the more parameters are measured, the more difficult it is to find strong relationships between parameters and the more difficult it gets to interpret the results correctly.

The first parameter to be discussed here is the type of soil. This parameter seems to be important in the choice of a sett site (Neal, 1986; Dunwell & Killingley, 1969; Zejda & Nesvadbova, 1983). No exact data was collected on different soil types at each sett. Information about soil came from the corresponding map (Bodeneignungskarte des Kantons Zürich, 1: 50'000). The dominant soil types in Sihlwald are sandy to silty clay and argillaceous sand. Sandy soil is normally preferred to one of clay (Neal, 1986). The only disadvantage of sandy soils is that if they are very light there is a risk of the roof of the sett collapsing. Neal (1986) mentions, however, that these soils are often associated with woodland and that badgers choose sett sites where tree roots provide the necessary strength. In Sihlwald, setts built under tree roots were seen on several occasions (Fig. 4) (see also 5.3). All in all, Sihlwald seems to have an ideal soil for digging setts (see also topography, p. 22). This factor could possibly explain the relatively high sett density in this forest compared to other densities in Switzerland (see above).

Altitude: fewer setts than expected were found below 600 metres a.s.l., more setts than expected were found above 600 metres. A possible reason for this observation could be the proximity of the "Sihltalstrasse" in the lower ranges, the main road connecting Zurich with Central Switzerland. This road is permanently busy and represents along with the railroad and the river Sihl, a major barrier for the animals. It is unclear whether it is the noise of the passing cars and trucks or the level of traffic mortality that keeps the badgers from digging their setts in proximity of the road (Skinner et al., 1991). However, only one or two badgers are killed on the "Sihltalstrasse" yearly (Leutwyler, pers. comm.) As already mentioned above, it is likely that food distribution (see also distance to forest boundary, p. 23) plays a bigger role in the distribution of the setts than the noise of passing vehicles or traffic mortality.

Aspect: significantly fewer setts than expected faced North. However, no preference was attained for any other category. In general, slopes facing north seem to be unattractive sett sites (Monnier, 1993; Skinner et al., 1991; Anrys & Libois, 1983). Owing to the orientation of the forest (SSE to NNW) and thus to lack of slopes facing south, a preference for this aspect could not be confirmed during the study.

Inclination: most setts in Sihlwald were dug in a slope with an inclination between 20°-40°. Fewer than the number of expected setts were dug in flat areas (£10°). Setts are usually dug in slopes of some sort (Neal, 1986; Skinner et al, 1991). Thornton (1988) showed that main sett density increased with the hilliness of the district. This is advantageous to the badger in various ways (Neal, 1986):

Topography: this parameter is closely connected with inclination. Setts in Sihlwald were preferably dug in convex slopes. Such setts have the following advantages: the badgers can pick up scents from several directions without having to leave the security of the sett and thus have several directions from which they can leave the sett. It is also possible that setts on convex slopes are easier to enlarge, since the rounded shape of the slope gives the badgers more opportunities for digging entrances and connecting tunnels than the shape of a gentle slope. Topography is a parameter that has never been mentioned in the literature before and seems to play a key role in the choice of a sett site. It appears to be important to pay attention to the small-scale topography around the setts.

Vegetation cover in general: the preferred sett sites are tree stands with sparse ground cover (i.e. a low herb- and moss coverage). As mentioned in Neal (1986) and Zejda & Nesvadbova (1983), coniferous stands were avoided. In Sihlwald, badgers clearly preferred areas with a sparse lower- and middle coniferous layer coverage. However, as shown above (4.5), single spruces were popular sett sites. According to Neal (1986), cover seems to play an important role in the choice of a sett site. Cover allows the badgers to emerge inconspicuously, and it allows the young cubs to play near the entrance without being visible to people or potential predators. However, the parameter cover is never clearly defined. Cover does not only have to be the vegetation cover; the small-scaled topography around a sett can be a type of cover, too. The distance to the nearest path (just "being out of view") can indirectly be a type of cover as well. It should be mentioned that no comparisons were made between the field data tree coverage and the lower-, middle-, and upper layer coverage, since the data of the latter are averages of the individual forest stands and thus too vague to be compared with the exact data of tree coverage at each site.

Distance to nearest path: Sihlwald is a popular recreation area for joggers, mountain bikers and hikers, especially on weekends. Fortunately, owing to the topography and the often dense vegetation of this forest, people rarely leave the paths. However, even though badgers are nocturnal, they avoid digging their setts close to the paths. Setts are most common in the distance categories >50 metres to the nearest path.

Distance to nearest water: there is no area in Sihlwald that is very far from water. Thus, it is assumed that this is the reason why no significant difference was attained for any distance category to nearest water. It is not clear why this parameter affects the distribution of the setts, but it probably correlates with other habitat parameters relevant to sett sites.

Distance to forest boundary with adjoining agricultural zones: setts are found significantly closer to the forest boundary than random points. Ideal sett sites are determined by at least two factors: the proximity to the forest boundary and adjoining agricultural zones, guaranteeing the presence of a good food supply on the one hand, and the forest itself, providing slopes, cover and little disturbance on the other hand.

The setts found in areas with a higher sett density than expected from a random distribution (light coloured patches, Fig. 9 were generally situated at a lower altitude, had a lower vegetation coverage and were nearer to paths than the other setts. This can be explained by the fact that the area from "Häuliboden" to "Sihlzopf" is actually at a lower altitude than the rest of the forest. The forest is narrower in this area, so that the setts were closer to the paths and to the forest boundary.

In order to understand the results of the analysis of the setts mentioned above, Fig. 22 shows the different parameters affecting sett sites in Sihlwald.

During the survey 40 latrines were recorded. The number of dung pits per latrine varied from one to 20 (Tab. 7). Nearly half the latrines only had a single pit. Four latrines had three dung pits, two had four, four had five, and two had six dung pits. The biggest latrine consisted of 20 dung pits. The average of dung pits per latrine was 3.5. The number of defecations varied from one to seven, with an average of 1.475 defecations per latrine.

Fig. 23 [ISS: This figure can only be viewed with the permission of the author (Password needed)] shows the distribution of the latrines in Sihlwald. Since only 40 latrines were found, the distribution of the latrines was not tested against a random distribution. All that can be said is that the minimum density of latrines in Sihlwald is 4/100 ha.

There is a significant difference between the latrines near (<= 5 m) setts and those that are not in the immediate vicinity of a sett. As the mean rank shows (Tab. 8), latrines near setts contain less dung pits (p=0.0103) and less used dung pits (p=0.0377) than latrines further away from setts.

Too few latrines were found to test the following details. Nevertheless, certain observations are worth noting: of the 40 latrines found, 23 had a clear runway nearby. More than two thirds (28) of the latrines were found near (<= 5 m) a conspicuous landmark (Tab. 9). The most commonly used landmark was near a sett, followed by near a path. Four latrines were found at the edge of the forest (one of which was near a fence post), two on the border of two different forest stands (mixed=deciduous and coniferous trees/deciduous trees) and two on the border of two development stages (old mixed/young deciduous stand). One latrine was found in an entrance of a sett. A third of the latrines (12) could not be placed in any of the above categories.

The GIS habitat parameters of the latrines for which the distribution of the latrines is decidedly non-random are listed in Tab. 10. Altitude seems to have the biggest effect on the distribution of the latrines (p<0.001), followed by distance to nearest water (p<0.05).

No significance was attained for any habitat parameter in the Chi²-test for two independent samples (latrines versus random points).

Tab. 10: Comparison of the GIS habitat parameter categories of the latrines with the availability of those parameter categories within the study area.

Fig. 24: Distribution of the latrines for altitude categories (Bonferroni z statistic, Neu et al., 1974). Dark coloured bars = number of setts, light coloured bars = number of random points.

Fig. 25: Distribution of the latrines for categories to nearest water. (Bonferroni z statistic, Neu et al., 1974). Dark coloured bars = number of setts, light coloured bars = number of random points.