Response distances of wild forest reindeer ( Rangifer tarandus fennicus Lönnb . ) and semi-domestic reindeer ( R . t . tarandus L . ) to direct provocation by a human on foot / snowshoes

#e objective of the study was to examine response distances of wild forest reindeer (Rangifer tarandus fennicus Lönnb.) and semi-domestic reindeer (R. t. tarandus L.) in Finland and Norway to direct provocation by a human on foot/snowshoes in 5 areas and in 15 reindeer herding cooperatives during di$erent seasons in 2010-12. #ere were no signi"cant di$erences in mean herd size or in sight, alert, %ight and closest response distances of wild forest reindeer in the Kuhmo and Suomenselkä areas. #e encounter distance in wild forest reindeer was signi"cantly (P< 0.005) longer than in semi-domestic reindeer in Finland and in Finnmark, Norway, and it increased with the group size. #e sight and the alert distances in wild forest reindeer were signi"cantly (P< 0.001) longer than in semi-domestic reindeer. In addition, the %ight distance for wild forest reindeer (mean 192 m) was signi"cantly (P< 0.001) and almost three times longer than in semi-domestic reindeer in Finland (mean 68 m). #e closest mean distance was in wild forest reindeer 191m (range 100-320 m) but only 44 m (range 2-110 m) in semi-domestic reindeer (P< 0.001). #e sight, alert, %ight and closest response distances were slightly longer in Norwegian than in Finnish semi-domestic reindeer. However, these distances were signi"cantly (P<0.005) longer in Pohjois-Salla (no supplementary feeding) than in other Finnish reindeer herding cooperatives and at the Kaamanen experimental station. #e mean %ight distance of reindeer in Pohjois-Salla was 115 m but only 65 m in other cooperatives (P< 0.001). #e closest distance of semi-domestic reindeer in PohjoisSalla (mean 105 m) was more than 2.5 times longer than in other reindeer herding cooperatives (mean 40 m). #e mean sight, alert and %ight distances in wild forest reindeer in autumn and winter were signi"cantly longer (P<0.005) than in semi-domestic reindeer in Finland. However, during summer these distances in wild forest reindeer herds with young calves were signi"cantly longer (P<0.005). #e mean herd size of Finnish semi-domestic reindeer was almost the same in di$erent seasons, but in wild forest reindeer it was slightly bigger during winter and spring and smaller during summer and autumn, only 7-23 reindeer. #e mean encounter and sight distances in semi-domestic reindeer were signi"cantly longer (P<0.005) in winter, but the mean alert and %ight distances were almost the same in winter and summer and slightly longer than during other seasons. #e results suggest that the supplementary feeding practice during winter may likely cause a reduction in %ight distances in semi-domestic reindeer.


Introduction
ere are many activities in which people may negatively in uence the behaviour of wild Rangifer tarandus, and also a ect their movement and subsequent range use (Wolfe et al., 2000).Human activities and infrastructure contribute with noise for example from power lines, generators, windmills, and also from moving objects like humans on foot, snowshoes and skis, snowmobiles, four-wheelers, cars, aircrafts and helicopters.While roads alone are not likely perceived as a threat to reindeer, increasing roads and tra c are.During the last 30-40 years outdoor ecotourism, hiking, skiing and hunting have been expanding and increasing activities in more remote areas, including mountain habitats of wild and semi-domestic reindeer (Helle & Särkelä, 1993;Colman et al., 2001;Reimers et al., 2006).ere have also been many changes in reindeer herding and husbandry practices over the last years, especially in Finland (Nieminen, 2006).
According to Baskin & Skogland (2001) reindeer are at an early phase of domestication, but semi-domestic reindeer generally exhibit more relaxed fright and ight behaviour compared to wild reindeer (Reimers et al., 2000;2006).erefore, when comparing behaviour between di erent reindeer herds, it is important to know the origin and history of the herds in question.While the caribou subspecies are wild, the Fennoscandian tundra reindeer population includes many domesticated herds (Reimers & Colman, 2006).
e semi-domestic reindeer herds in Northern Finland, Sweden and Norway, as well as the wild reindeer herds (with a mix of wild and domesticated origin) in Southern Norway, are originally Eurasian wild tundra reindeer (Rangifer tarandus tarandus L.).DNA-analyses by Røed et al. (2008) support independent origin of semi-domestic reindeer in Fennoscandia and Russia.e domestic gene pools seem to meet only in eastern Finland, mainly in Halla reindeer herding cooperative.
However, the wild forest reindeer (R. t. fennicus Lönnb.)population in eastern Finland and the wild reindeer populations in central Norway have contributed little or nothing to the domestic gene pool (Røed et al., 2008).Domestication is the rst step of selection, and it is argued that domestication has mostly resulted in quantitative rather than qualitative changes (Mignon-Grasteau et al., 2005).
Human activities a ect reindeer/caribou through the senses of hearing, sight and smell.According to Flydal et al. (2001) the hearing capacity of reindeer ranges from 70 Hz to 38 kHz at a sound pressure of 60 dB.It means that almost all noises and vocalizations are readily perceived by reindeer.Reindeer as other ungulates has apparently also very good day and night vision.Reindeer most likely perceive colours, but no particular colour appears to be dominant, and reindeer probably are unable to distinguish between red and green colour.It is mainly contrasts and movements that betray human presence.Because the eyes of reindeer are laterally positioned, the combined visual elds of both eyes cover virtually 360 o (Nieminen, 1994).It means that reindeer can also spot predators and humans sneaking up from behind.Reindeer´s laterally positioned eyes limit, however, the binocular visual eld.Although reindeer´s sense of smell is not well documented, the capacity to capture scents even under unfavourable wind conditions is well known by reindeer/caribou hunters, hikers and also reindeer herders.Sometimes smell alone can trigger ight of reindeer without input from other senses.Reimers & Colman (2006) predicted that reindeer would also respond at greater distances to the directly approaching person when the wind carried the human scent to the reindeer than when the reindeer could not smell the human intruder.
e strongly elevated nasals of the wild forest reindeer living only in forested regions indicate the very keen sense of smell due to increased olfactory mucous membranes (Nieminen, 1980).
It has long been recognized that learning plays an important role in the manner and degree to which ungulates respond to humans (Geist, 1971), and there are usually three major learned responses which also are valid for reindeer: habituation, attraction and avoidance.Domestication, habituation and sensitisation are essential in shaping adaptability of reindeer and caribou (Reimers & Colman, 2006).Many impact studies focus on the behavioural responses of wildlife to humans, because these attributes are generally more amenable to study than other forms of response (Bejder et al., 2009).e cases of presumed habituation or sensitisation may actually represent di erences in the tolerance level of wildlife to anthropogenic activity.For example reindeer show decreased ight responses in areas with relative high amounts of human activities (Colman et al., 2001;Reimers et al., 2009), indicating the ability to habituate to human activities.Habituation to humans would occur more readily in caribou populations that were not hunted and lived in areas lacking natural predators (Klein, 1980;Aastrup, 2000).Habituation and also former experiences with predators signi cantly in uence an ungulate´s perception of threat.
e populations with few predators ushed at greater distance than those where predators are common.All predator studies reviewed by Stankowich & Blumstein (2005) classi ed humans as the predator and measured di erences in ight initiation distance between ungulate populations that di ered with regard to human exposure.Wild ungulate populations like Svalbard reindeer (R. t. platyrhynchus Vrolik) exposed to a relatively high level of human activities have become habituated to humans in a non-threatening context (Tyler, 1991).ey are likely to perceive less risk when approached by humans than would animals in populations where encounter with humans are rare (Colman et al., 2001;Lund, 2008).Even though semi-domestic reindeer are disturbed by hu-man activities, they can increase their tolerance towards humans if insect harassment is severe during summer (Skarin et al., 2004).Rutting activities during autumn obviously also a ect reindeer behaviour more than the directly approaching human observer (Reimers et al., 2006).
Flight initiation distance is the distance at which an animal begins to ee from an approaching predator or human.It is usually used in studies, because it is easy to measure and correlates with other key aspects of escape behaviour e.g.alert distance (Blumstein et al., 2005).According to Vistnes & Nellemann (2008) accurate assessment of impacts from human activity requires, however, regional-scale and usually long-term studies.e objective of the present study was to examine response distances of wild forest reindeer and semi-domestic reindeer in Finland and Norway to direct provocation by humans on foot or on snowshoes.e observations were collected during di erent seasons in areas subjected to combinations of high or low human activity, supplementary winter feeding or freely-grazing in the forests or on the elds.
e objective was also to determine if ight distances have changed over the last years in response to increased human activity and supplementary feeding of semi-domestic reindeer on natural pastures or in corrals.

Study areas and reindeer herds
In total, 55 di erent reindeer herds were included in the present study: 17 wild forest reindeer herds (Rangifer tarandus fennicus Lönnb.) in the Suomenselkä area (9 in Perho, 1 in Kyyjärvi and 7 in Alajärvi municipalities) and 3 in Kuhmo municipality in the Kainuu area, 32 semi-domestic reindeer herds (R. t. tarandus L.) in 15 reindeer herding cooperatives in Finland including a herd at Kaamanen experimental station (150 reindeer, fenced area, 44 km 2 ), and 3 herds in Northern Norway (2 in Kautokeino, West-Finnmark and 1 in Pykeija, East-Finnmark) (Fig. 1).Approximate herd size was 40 animals in wild forest reindeer and 115 animals in semi-domestic reindeer in Finland and Norway.e herds of wild forest reindeer were found with the help of GPS-collared females.In the studied reindeer herding cooperatives in Finland total numbers of counted reindeer varied between 2 200-8 300 animals.Reindeer herds were studied during di erent seasons in the central areas of the cooperatives.According to the reindeer herders these reindeer herds represent a mixture of animals owned by many herders (the animals have di erent earmarks), i.e. constitute of animals belonging to the main herd of the respective cooperative.
e wild forest reindeer disappeared from Finland for decades in the beginning of the ing the last years the number of wild reindeer has decreased due to predation.In Kuhmo municipality, near the Russian border, the population of wild forest reindeer has also decreased, and the total number of reindeer is today about 900 (Kojola et al., 2009).Northern Finland di ers from the south mainly by the type of forest, but also by the elevated watershed areas in central and eastern Finland.Suomenselkä and Kainuu belong to the middle boreal vegetation zone (Ahti et al., 1968).e landscapes are dominated by mires, and they are located on watersheds.Both areas are diverse in altitudes, mountains, hills and valleys, barren areas with forests, bogs and lakes.e landscapes are dominated by Norwegian spruce (Picea abies) and Scots pine (Pinus sylvestris) forests with ericaceous heather, lichen (Cladonia, Cladina spp.) and boggy areas.e northern reindeer herding cooperatives are situated in the north boreal vegetation zone (Ahti et al., 1968).e terrain is dominated by rolling hills with different aged forest stands of mainly Scots pine.Mountain birch (Betula pubescens czerepanowii) grows in the slopes of the highest hills, and only the tops of the highest fells are barren.e cooperatives in the middle and southern parts of reindeer herding area belong, like Kainuu, to the middle boreal vegetation zone (Ahti et al., 1968).Agriculture is common and there are many elds in the reindeer herding cooperatives and also in wild forest reindeer areas.Because the numbers of wild forest reindeer have decreased as described above, no reindeer was hunted during the study period in the Kuhmo and Suomenselkä areas.
In the reindeer husbandry area in Finland, and also in Finnmark, Norway, lichens pastures are generally strongly or very strongly worn (lichen biomass < 100-300 kg dry weight/ ha) (Kumpula et al., 2009;Mattila, 2006).As results of reindeer grazing, lichen ranges in mountain areas, large national parks, nature reserves and other wilderness areas are worn in Finland (Nieminen, 2010).In the northern cooperatives also the reindeer summer pastures are worn, and grazing has been the main reason to change vegetation and cause erosion in some places.
e current condition of winter pastures in combination with a continuous state of change for the worse show that maintaining the current number of reindeer on their natural winter pasture is no longer possible.However, the body condition of reindeer in Finland has usually been good even during hard winters, due to the intensive supplementary feeding practised in the cooperatives.Totally over 40 million kg feed (calculated as dried hay) are used yearly during winters for feeding of semidomestic reindeer, mainly on natural pastures or in corrals in the middle or southern reindeer herding cooperatives (Nieminen, 2006).Supplementary winter feeding of reindeer in Finnmark area, Norway, is not common, but in Finland only Pohjois-Salla reindeer herding cooperative has herded reindeer on the natural pastures without supplementary feeding during winters.e wild forest reindeer are freelygrazing in Suomenselkä and Kuhmo, and both winter and summer pastures are in rather good condition.In Kuhmo the amount of lichen biomass (dry weight) has over the last years been seven times higher than in the nearby Halla reindeer herding cooperative (Mattila, 2004).e road network in wild forest reindeer areas and also in reindeer husbandry areas are well developed, and every year some wild forest reindeer and over 4 000 semi-domestic reindeer die by tra c.Also, some adult wild forest reindeer and more than 4 000 semi-domestic reindeer are killed yearly by big predators in Finland (Nieminen, 2012).

Data collection
In 2010-12, during the four sampling periods of September-November (autumn and rutting period), February-March (winter), April (spring) and July (summer), a single and same person (the observer) on foot or on snowshoes (used in deep snow), dressed in dark clothing, disturbed wild forest and semi-domestic reindeer during daylight hours by directly approaching them.e observer used binoculars and camera (Nikon D80) to document behaviour and places and measured later (using 1 m steps) response distances between the reindeer and the observer and the resultant distances by the reindeer after taking ight.Upon location of a group (≥ 4 reindeer), 10 parameters were recorded: 1) sample month, 2) group size (small: < 20 animals, medium: 20-49 animals, and large: >50 animals), 3) group composition (females and calves, males), 4) dominant activity of the group when rst sighted (lying or grazing, rutting, moving), 5) wind direction relative to the observer (no wind, upwind or downwind), 6) vegetation type (open eld, marsh, forest), 7) topography of the surrounding area (level, mountain, lake ice and feeding place), 8) visibility/weather (sunny, cloudy, raining/ snowing), 9) snow depth and 10) temperature.
When a group of reindeer was rst sighted, the observer took pictures, and used the so called direct approach method: advancing directly towards the centre of the group at a constant speed (about 4 km/hour) with < 10 second stops, to take pictures and later measured the four additional response distances de ned below.When reindeer are rst disturbed they show signs of awareness and fright by raising their heads and tails, urinating and sometimes jumping (deVos, 1960;Horejsi, 1981).e initial ight is often followed by curiosity behaviour: the disturbed reindeer circles around the intruder to catch the scent.All measurements were made from the position of the directly approaching observer to the nearest reindeer.
e wildlife response distance terminology and methodology recommended by Taylor & Knight (2003) with the modi cations following Reimers et al. (2003;2009) were used in this study: 1) Encounter distance (END): the distance used as the starting point of the disturbance.
e reindeer rst discovers the provoker by sight or scent, indicated by looking, standing, turning their head or pausing from eating in a manner visible to the observer.2) Sight distance (SD): the distance between the observer and closest reindeer when reindeer in the group displayed an alerted behaviour directed at the observer.
3) Alert distance (AD): the distance at which the reindeer group displayed an increased alert response by grouping together or by individuals urinating with one hind leg extended outward at an exaggerated angle, while staring at the directly approaching observer (Fig. 2).4) Flight distance (FD): the distance from the directly approaching observer to the group when the reindeer initially took ight.5) Closest distance (CD): the distance from the directly approaching observer to the nearest animal if a group approached the observer immediately before nal withdrawal.6) Escape distance (ED): the shortest straightline distance from where the reindeer took ight in response to the observer to where the reindeer resumed grazing or bedded down.7) Assessment time: the time elapsed from alert to ight initiation estimated from measured distances and assuming a constant observer speed of about 4 km/hour.

Statistical analyses
Initially, data of all areas and both wild forest reindeer and semi-domestic were pooled to analyse the e ect of each factor on di erent distances separately.Correlation among dependent variables was tested with original (untransformed) data using Spearman rank correlation.e response variables (END, SD, AD, FD and CD) were rst transformed into their natural logarithms prior to analysis.Sight, alert, ight and closest distances were analysed with Mixed Models Analysis.Marginal F-tests were used for the full models contained area, season, group size, wind direction relative to the observer, vegetation type and dominant activity of reindeer.To test for di erences for provocation methods and to relate independent variables to reindeer responses, a mixed, stepwise analysis of ANOVA (analysis of variance) was used. 2 -test was used to assess the relative seasonal frequency in group-size classes.e statistical differences response distances in di erent groups were tested using t-test.Statistical tests were carried out by use of SPSS ver.7.0 for Windows.
e data were examined for statistical signi cance at P<0.05.

Results
During the four sampling periods in 2010-12 totally 55 independent reindeer groups were encountered and used in statistical analysis.A total of 2 216 reindeer were observed.Of these, 739 were wild forest reindeer and 1 477 semidomestic reindeer.
Although 27 reindeer groups were encountered during autumn, these groups represented only 22 % of the groups in the large size class (> 50 animals) and 9.3 % of all of the reindeer sampled.During autumn, rutting season, most reindeer were in few large, mixed-sex groups.Totally 14 groups were encountered during winter, 50 % of the groups was in the large size class and represented 20 % of all studied reindeer.During spring and summer, 8 and 6 groups, respectively, were encountered, and only 1-2 groups were in the large size class.During summer, reindeer were distributed usually in small-and medium-sized female-calf e reindeer group displayed an increased alert response by grouping together or by individuals urinating with one hind leg extended outward at an exaggerated angle, while staring at the directly approaching observer.Photo Mauri Nieminen.
groups.Mean herd size of wild forest reindeer was 37, which was almost the same as for semidomestic reindeer in Finland (mean 39).e groups observed in Norway (winter only) were bigger (mean 178 animals).
e encounter (END), sight (SD), alert (AD), ight (FD) and closest distance (CD) were positively correlated between areas, with a general decrease in the correlation coe cients from SD to CD (Table 1).FD increased with increasing encounter distance (END) (Table 2), as did also SD and AD, indicating that when the observer approached reindeer from farther away they responded at longer distances.No signi cant di erences, depending on whether the approach was made on foot or on snowshoes in winter, were seen in any of the response distances.ere were also no signi cant di erences depending on area (Kuhmo and Suomenselkä) in mean herd size, END, SD, AD, FD or CD of wild forest reindeer.e Encounter distance (mean ± standard deviation) of wild forest reindeer (332 ± 24 m) was, however, signi cantly (P< 0.005) longer than that of semi-domestic reindeer in Finland and Norway (226 ± 13 and 250 ± 6 m, respectively) (Table 3).END increased with group size in both wild forest reindeer and semi-domestic reindeer.SD (253 ± 16 m) and AD (216 ± 13 m) in wild forest reindeer were signi cantly (P< 0.001) longer than in Finnish semi-domestic reindeer (144 ± 7 and 94 ± 6 m, respectively).e FD in wild forest reindeer (192 ± 14 m) was almost three times longer than in semi-domestic reindeer (68 ± 5 m) in Finland (P<0.001) (Fig. 3).e escape distance (ED) of reindeer was possible to measure only in the mountain areas, and the mean ED of semi-domestic reindeer in Muotkatunturi reindeer herding cooperative in Finland and also in Kautokeino, Norway was 360 m.
e mean distance between alert and ight was 24 m in wild forest reindeer and almost the same, 26 m in semi-domestic reindeer in Finland.With an encounter speed about 4 km/ hour, these distances suggest that there was a separation of 22-24 seconds, on average, from when the reindeer groups became alert until they took ight.
e mean closest distance was 191 m (range 100-320 m) in wild forest reindeer but only 44 m (range 2-110 m) in semi-domestic reindeer in Finland (P<0.001).
e mean SD, AD, FD and CD in Norwegian semi-domestic reindeer herds were slightly, but not signi cantly longer than in Finnish semidomestic reindeer.However, the mean SD, AD, FD and CD of the semi-domestic reindeer in Pohjois-Salla (no supplementary feeding) were signi cantly (P<0.005)longer than in the other Finnish semi-domestic herds (Table 4).
e mean FD of the semi-domestic reindeer in Pohjois-Salla was 115 m, but only 65 m in the other herds.e mean CD of semi-domestic  reindeer in Pohjois-Salla was 105 m, which was more than 2.5 times longer than in the other herds (mean 40 m) (Table 4).e mean END, SD, AD and FD of wild forest reindeer were long in autumn and winter, and signi cantly longer (P<0.005)than in semi-domestic reindeer in Finland (Table 5, Fig. 4).However, during summer these distances of wild forest reindeer herds with young calves were signi cantly longer (P<0.005).
e mean herd size of Finnish semi-domestic reindeer was almost the same during the four seasons (37-43 reindeer) (Table 6).e mean herd size of wild forest reindeer was slightly larger during winter and spring (52-63 reindeer), and smaller during summer and autumn, only 7-23 reindeer.e mean END and SD of semi-domestic reindeer (285 m and 183 m, respectively) were signi cantly longer in winter (P<0.005), while mean AD and FD did not differ signi cantly depending on season.

Discussion
ere were no signi cant di erences depending on area (Kuhmo or Suomenselkä) in mean herd size, encounter, sight, alert, ight or closest distances of wild forest reindeer in Finland.Wild forest reindeer were transferred from Kuhmo to Suomenselkä more than 30 years ago (Nieminen & Laitinen, 1983).When approached by humans, reindeer are likely to respond as prey encountered by a predator (Frid & Dill, 2002).e behavioural responses of wild forest reindeer to human activity seem to be the same in both of the studied areas, where reindeer are today exposed to humans rather frequently.During autumn and winter wild forest reindeer are also grazing on the cultivated elds, quite close to human houses and other buildings.e encounter distance (END) of wild forest reindeer was, however, rather long (mean 332 m) and signi cantly longer than that of the studied semi-domestic reindeer.Di erences in mean END suggest that the observer would have been in view with wild forest reindeer about two minutes longer than with semi-domestic reindeer.Rangifer with a wild origin appear to have longer response distances than reindeer with domesticated origin (Reimers & Colman, 2006), and females with calves are more easily alarmed and more likely to ee from a potential threat than are adult only groups (Lent, 1966;Bergerud, 1974;Stankowich, 2008).END increased with group size in wild forest reindeer as well as in semi-domestic reindeer.According to Reimers et al. (2006) the END of feral reindeer (wild reindeer with a semi-domestic origin) in mountain areas of south eastern Norway was longer than that of my observation of wild forest reindeer in the present study, and it also increased with group size.Klein (1979), Baskin (1986) and Baskin & Hjälten (2001) et al., 2006).
In reindeer, olfaction is generally considered a more dependable sense than vision for early warnings.Scent would thus result in faster reactions at farther distances to humans.However, feral reindeer did not respond di er-ently during events when they were downwind of the observer compared to when they were upwind (Reimers et al., 2006).us for reindeer in open or mountain areas vision might often be relatively more important than detection by scent or sound compared to the relative importance of these senses for ungulates in forest habitat.Usually both wild and semidomestic reindeer that have visually detected a disturbance source, for example a human, will often approach or circle the source until they are able to con rm by scent the apparent need for ight.In the present study the mean closest distance (CD) was 191 m (range 100-320 m) in wild forest reindeer but only 44 m (range 2-110 m) in semi-domestic reindeer.According to Reimers et al. (2006) also Norwegian feral reindeer approached by the observer usually rst moved towards the observer before they took nal ight.e closest approach by feral reindeer in winter was the same as in Finland with wild forest and semi-domestic reindeer.
e mean sight (SD) and the alert distances (AD) were signi cantly longer in wild forest reindeer than in semi-domestic reindeer, and the mean FD was almost three times longer in wild forest reindeer than in semi-domestic reindeer in Finland.Although reindeer are considered to be at an early phase of domestication (Baskin & Skogland, 2001), semi-domestic reindeer generally exhibit more relaxed fright and ight behaviour compared to wild reindeer (Reimers et al., 2000;2006).According to Tarlow & Blumstein (2007) an important interacting factor is frequency of interaction with humans, and reindeer -like other ungulates -show reduced ight responses in areas with frequent contact with humans compared to those in areas where human encounters are rare (Colman et al., 2001;Stankowich, 2008).
In the present study the mean SD, AD, FD and CD in Norwegian semi-domestic reindeer herds were slightly longer than in Finnish semidomestic reindeer.ese distances were, however, signi cantly longer in reindeer in Pohjois-Salla reindeer-herding cooperatives than in other Finnish cooperatives.
ere was no supplementary feeding of reindeer in Pohjois-Salla.In other herding cooperatives almost all reindeer are fed 3-5 months every winter.e mean FD was 115 m in semi-domestic reindeer of Pohjois-Salla but only 65 m in other cooperatives in Finland.Also the mean CD of reindeer in Pohjois-Salla was 105 m and more than 2.5 times longer than in other cooperatives.
e results suggest a strong taming e ect due to the supplementary feeding practice, which in some areas have been going on for 40 years in Finland.Increased snowmobile use has also expanded the area where humans are daily in contact with reindeer during winter and spring.Several studies have also shown that people and dogs elicit usually greater ight responses than machines, e.g.skiers vs. snowmobiles (Freddy et al., 1986).e mean SD, AD and FD were longer in wild forest reindeer and also in semi-domestic reindeer in Finland during winter and autumn.It is in accordance with earlier results with wild mountain reindeer in southern Norway (Reimers et al., 2009).Earlier studies (Dervo & Muniz, 1994;Kind, 1996;Eftestøl, 1998) on reactions of wild reindeer to humans on foot or skis have also revealed longer reaction distances during winter than in the other seasons, indicating that reindeer are especially vulnerable to disturbance during winter, usually a period of negative energy balance (Reimers et al., 2003).
e distance the feral reindeer in south eastern Norway moved away in response to the approaching human was greatest during summer (Reimers et al., 2006).Farther alert and ightinitiation distance in winter may be explained by the observer being easier to detect against snow (Reimers et al., 2009).Shorter alert and ight distances for larger than smaller groups are also in general agreement with previous studies (deVos, 1960;Baskin & Hjälten, 2001;Reimers et al., 2006).e mean herd size of wild forest reindeer was slightly bigger during winter and spring, and smaller during summer and autumn.e mean END and SD of semidomestic reindeer were also signi cantly longer in winter, but the mean AD and FD were almost the same in winter and summer and only slightly longer than during other seasons.I acknowledge the herders in the studied reindeer herding cooperatives and also like to thank Kaamanen experimental station.I especially want to thank Jukka Siitari at RKTL, Reindeer Research Station in Kaamanen for his assistance.I like to thank Øystein Holand for his valuable comments on my manuscript.I also thank two anonymous referees for their critical evaluation.

Fig. 2
Fig. 2. e alert distance (AD) of a wild forest reindeer group in the Suomenselkä area, in February 2010.e reindeer group displayed an increased alert response by grouping together or by individuals urinating with one hind leg extended outward at an exaggerated angle, while staring at the directly approaching observer.Photo Mauri Nieminen.

Fig. 4 .
Fig. 4. Response distances (mean ± SD) of wild forest reindeer in Suomenselkä (WFR-Suo) and of semi-domestic reindeer in Finland (SDR-Fin) in groups disturbed by a directly approaching human on foot/snowshoes during autumn and winter 2010-12.Sample sizes are: END (encounter), SD (sight) AD (alert), and FD ( ight distance), n = 10 for WFR in autumn and n = 7 in winter, n = 17 for SDR-Fin in autumn and n = 6 in in winter, and CD (closest distance), n = 7 for WFR in autumn and n = 5 in winter, n = 14 for SDR-Fin in autumn and n = 4 in winter.

Table 2 .
Linear mixed-e ects model for predicting ight-initiation distances (In transformed) of groups of wild forest reindeer and semi-domestic reindeer disturbed by anapproaching observer on foot/snowshoes in di erent areas in Finland and Norway in 2010-12.

Table 3 .
Observed response distances of wild forest reindeer and semi-domestic reindeer in Finland and also of semidomestic reindeer in Finnmark, Norway, when provoked by an observer on foot/snowshoes in 2010-12 (data pooled across years).
have also reported behavioural di erences between forest living

Table 4 .
Observed response distances of semi-domestic reindeer in di erent Finnish reindeer-herding cooperatives and areas (winter feeding common) and in Pohjois-Salla (no feeding), when provoked by an observer on foot/ snowshoes in 2010-12 (data pooled across years).

Table 5 .
Observed response distances of wild forest reindeer during di erent seasons when provoked by an observer on foot/snowshoes in 2010-12 (data pooled across years).

Table 6 .
Observed response distances of semi-domestic reindeer in Finland during di erent seasons, when provoked by an observer on foot/snowshoes in 2010-12 (data pooled across years).