Jørund Rolstad, Per Wegge, Andrey V. Sivkov, Olav Hjeljord, and Ken Olaf Storaunet

Transcription

1 1032 Size and spacing of grouse leks: comparing capercaillie (Tetrao urogallus) and black grouse (Tetrao tetrix) in two contrasting Eurasian boreal forest landscapes Jørund Rolstad, Per Wegge, Andrey V. Sivkov, Olav Hjeljord, and Ken Olaf Storaunet Abstract: Capercaillie (Tetrao urogallus L., 1758) and black grouse (Tetrao tetrix L., 1758 (= Lyrurus tetrix (L., 1758))) are two sympatric Eurasian lekking grouse species that differ markedly in habitat affinities and social organization. We examined how size and spacing of leks in pristine (Russia) and managed (Norway) forests were related to habitat and social behavior. Leks of both species were larger and spaced farther apart in the pristine landscape. Capercaillie leks were regularly spaced at 2 3 km distance, increasing with lek size, which in turn was positively related to the amount of middleaged and older forests in the surrounding area. Black grouse leks were irregularly distributed at shorter distances of 1 2 km, with lek size explained by the size of the open bog arena and the amount of open habitat in the surroundings. At the landscape scale, spatial distribution of open bogs and social attraction among male black grouse caused leks to be more aggregated, whereas mutual avoidance in male capercaillie caused leks to be spaced out. In the pristine landscape, large-scale and long-term changes in forest dynamics owing to wildfires, combined with an aggregated pattern of huge bog complexes, presumably provide both grouse species with enough time and space to build up bigger lek populations than in the managed landscape. Résumé : Le grand tétras (Tetrao urogallus L., 1758) et le tétras lyre (Tetrao tetrix L., 1758 (= Lyrurus tetrix (L., 1758))) sont deux espèces eurasiennes sympatriques de tétras qui forment des leks et qui différent considérablement dans leurs affinités avec l habitat et leur organisation sociale. Nous examinons de quelle manière la taille et l espacement des leks dans des forêts vierges (en Russie) et aménagées (en Norvège) sont reliésàl habitat et au comportement social. Les leks des deux espèces sont plus grands et plus éloignés les uns des autres dans le paysage vierge. Les leks des grands tétras sont répartis régulièrement à des distances de 2 3 km, mais qui augmentent en fonction de la taille des leks, qui est elle-même en relation positive avec l importance des forêts d âge moyen ou plus avancé dans les environs. Les leks des tétras lyres sont répartis de façon irrégulière à des distances plus courtes de 1 2 km les uns des autres; la taille des leks s explique par l importance de l arène de tourbière ouverte et par la quantité d habitats dégagésàproximité. Àl échelle du paysage, la répartition spatiale des tourbières ouvertes et l attraction sociale entre les mâles de tétras lyres rend la répartition spatiale des leks plus contagieuse, alors que l évitement mutuel des mâles de tétras A résulte en des leks plus dispersés. Dans le paysage non modifié, les changements à grande échelle et à long terme de la dynamique forestière à cause des feux de forêt, combinés àun patron contagieux de complexes immenses de tourbières, fournissent vraisemblablement aux deux espèces de tétras suffisamment de temps et d espace pour l élaboration de plus grandes populations à leks que dans les paysages ménagés. [Traduit par la Rédaction] Introduction Grouse species (Tetraoninae) display a wide range of social organizations, from monogamy, via solitary displaying polygynous males, to classical leks (Hjorth 1970; Kruijt et al. 1972; Wiley 1974; Wittenberger 1978). In general, it has been noted that open steppe and prairie species are lekking and that forest-dwelling species tend towards solitary display (Hamerstrom and Hamerstrom 1961), a dichotomy often ascribed to antipredation behavior (Berger et al. 1963; Koivisto 1965; Boag and Sumanik 1969; Hjorth 1970; Wiley 1974; Bergerud 1988). In North America, this seems to hold true, as all lowland open-dwelling species are highly social and classical lekking, whereas the forest-dwelling species display solitarily in dispersed, separate territories (Johnsgard 1983; Bergerud 1988). Received 2 April Accepted 6 August Published on the NRC Research Press Web site at cjz.nrc.ca on 27 October J. Rolstad 1 and K.O. Storaunet. Norwegian Forest and Landscape Institute, P.O. Box 115, N-1432 Ås, Norway. P. Wegge and O. Hjeljord. Norwegian University of Life Sciences, Department of Ecology and Natural Resource Management, P.O. Box 5003, N-1432 Ås, Norway. A.V. Sivkov. State Natural Reserve Pinezhskiy, 123-a. Pervomayskaya, Archangelsk Region, Russia. 1 Corresponding author ( jorund.rolstad@skogoglandskap.no). Can. J. Zool. 87: (2009) doi: /z09-093

2 Rolstad et al Fig. 1. Distribution of leks of capercaillie (Tetrao urogallus; a, b, c) and black grouse (Tetrao tetrix (= Lyrurus tetrix); d, e, f) at Pinega, northwestern Russia, from 2000 to 2005 (a, d); Varaldskogen, southeastern Norway, from 1980 to 1983 (b, e); and Varaldskogen from 2006 to 2008 (c, f). Thick black circles indicate size of leks scaled to the square-root number of attending males (cf. Fig. 3). Black triangles (e) are leks of unknown size. Thin black circles are scaled to a radius of 1 km (a, b, c, d) and 500 m (e, f). Letters g and h denote subsamples of leks of black grouse with radio-marked males, as shown in detail in Fig. 6. Habitats are denoted by the following colors: open bogs (yellow); clearcuts, burns, and young forest <40 years (light green); middle-aged and old forest combined >40 years (dark green); and open water (light blue). In boreal Eurasian forests three grouse species coexist, of which the monogamous hazel grouse (Bonasa bonasia (L., 1758) (= Tetrastes bonasia (L., 1758))) inhabits dense forest. The two others, the black grouse (Tetrao tetrix L., 1758 (= Lyrurus tetrix (L., 1758))) and capercaillie (Tetrao urogallus L., 1758), have proven more difficult to classify with respect to the above-mentioned dichotomy (Wiley 1974; de Vos 1979). Beyond doubt, the black grouse is a classical lekking species, with leks located in open habitat and males being highly social throughout the mating season (Koivisto 1965; Kruijt and Hogan 1967; Hjorth 1970; Kruijt et al. 1972; Alatalo et al. 1991). However, in contrast to North American lekking species, male and female black grouse, throughout most of their boreal distribution range, spend most of their time in forested habitat when off the lek (Seiskari 1962; Hjorth 1970; Johnsgard 1983; Klaus et al. 1990; Lindstro m et al. 1998). The capercaillie is an all-year forest-dwelling species (Seiskari 1962; Rolstad et al. 1988; Klaus et al. 1989; Storch 2001), and therefore should subscribe to a dispersed polygynous mating system (de Vos 1979; Bergerud 1988). Yet, all field studies document a lek-type organization, with males clearly aggregated at traditional display sites during the display season (Lumsden 1961; Klaus et al. 1968; Hjorth 1970; Pirkola and Koivisto 1970; Roth and Nievergelt 1975; Rolstad and Wegge 1987a; Wegge and Larsen 1987; Storch 1997; Storch 2001; Mossoll-Torres and Menoni 2006). However, capercaillie leks do differ from classical lekking species in that leks are situated within closed-canopy forest and displaying males are spaced farther apart from each other, although mostly within visual distance. Moreover, when not

3 Table 1. Spatial characteristics of lek populations of capercaillie (Tetrao urogallus) and black grouse (Tetrao tetrix (= Lyrurus tetrix)) at Pinega, northwestern Russia, and Varaldskogen, southeastern Norway, as measured by the R distance-to-nearest neighbor (DNN) and G DNN statistics. DNN CV* (%) R { z P G { P DNN (km) No. of males/ No. of leks/ No. of km km 2 males/lek Reference area (km 2 ) N Capercaillie Pinega < <0.001 Varaldskogen Varaldskogen < Can. J. Zool. Vol. 87, 2009 Black grouse Pinega Varaldskogen Varaldskogen Note: Significant P values are in boldface type. *DNN coefficient of variation (CV). { R = 0: perfectly clumped; R = 1: random; R = 2.15: uniform global spacing (Clark and Evans 1954). { G = (N = 17 10): random; G = 1: uniform local spacing (Brown and Rothery 1978). Excluding 40 km 2 bog area southwest in study area. Based on 10 leks within 30 km 2. attending the clustered display sites, males live solitarily and evenly spaced (Wegge and Larsen 1987; Storch 1997; Eliassen and Wegge 2007), and leks appear regularly distributed at distances of ~2 km (Wegge and Rolstad 1986). Over the years, studies of grouse mating systems have contributed significantly to our current understanding of the driving forces behind the evolution of leks (e.g., Hjorth 1970; Kruijt et al. 1972; Wiley 1978; de Vos 1983; Gibson and Bradbury 1987; Alatalo et al. 1992). The display behavior of the males involves a remarkable variation in sound and visual effects (Hjorth 1970). Understandably, most studies therefore have focused on the behavior of the birds when on the leks, such as dominance hierarchies among competing males and female choice of male ornaments and behavioral characteristics. Less efforts have been devoted to examine landscape-scale interlek patterns, and to the behavior of males and females when outside the lek. Many hypotheses have been put forward to explain the evolution of leks, most commonly seeking to explain the ultimate mechanisms for male aggregation, but also more proximate causes of lek formation and interlek spacing (e.g., Bradbury 1981; Beehler and Foster 1988; Höglund and Alatalo 1995; Gjerde et al. 2000). In the present study, we compare the spatial distribution of capercaillie and black grouse leks in two contrasting boreal forest landscapes to get further insight into the proximate causes explaining size and spacing of leks. In particular, we ask whether capercaillie leks in natural forests are larger and spaced farther apart than in managed forests. This might be hypothesized from previous studies showing that the number of lekking males increases with increasing amount of old forest in the surroundings (Wegge and Rolstad 1986), and males mutually avoiding each other in radially extending daytime home ranges (Wegge et al. 2003, 2005b). Less is known about interlek spacing in black grouse. Being a more social species, with males usually living in flocks when off the lek, we hypothesize that their leks are less regularly spaced and more influenced by the presence of suitable open habitat for lekking. We pursue these hypotheses by comparing the patterns found in a large northwestern Russian Nature Reserve (Pinega, Archangelsk) with a highly managed forest in southeastern Norway (Varaldskogen). Materials and methods Study areas The Varaldskogen study area ( N, E) is located in southeastern Norway on the border to Sweden. It belongs to the middle boreal subzone (Ahti et al. 1968) and has a moderate continental climate with average ( ) July and January temperatures of 15.8 and 5.1 8C, respectively. Annual precipitation averages 700 mm, and snow variably covers the ground during December through April. Topography is gently rolling between 200 and 500 m above sea level., and the forest is characterized by a mixture of Scots pine (Pinus sylvestris L.; 70%), Norway spruce (Picea abies (L.) Karst.); 25%), birches (European white birch (Betula pendula Roth), and downy birch (Betula pubescens Ehrh.); 4%), as well as other northern deciduous species (1%). Within a 50 km 2 area, capercaillie and black grouse have been monitored since 1979 (Rolstad et al.

4 Rolstad et al Fig. 2. Distance between neighboring leks (DN) of capercaillie (Tetrao urogallus; left panel) and black grouse (Tetrao tetrix (= Lyrurus tetrix); right panel) in relation to lek size (mean number of lekking males per pairs of leks). Symbols denote leks at Pinega, northwestern Russia, from 2000 to 2005 (solid circles); Varaldskogen, southeastern Norway, from 1980 to 1983 (open circles); and Varaldskogen, from 2006 to 2008 (open downward-facing triangles). Lek size is scaled to square-root values. Thick solid line is the pooled OLS (ordinary least square) regression of leks of capercaillie. Broken lines are the expected relationships if leks are randomly distributed in a close-packed lattice with overall male densities of males per km 2. For details see the Materials and methods. 1988; Wegge et al. 1992). Some years, more extensive surveys were conducted in a 200 km 2 surrounding area. Varaldskogen has been intensively managed for several hundred years. During the first half of the 20th century, timber was harvested by means of high-grading selective cuttings, giving rise to a rather homogenous landscape dominated by sparsely stocked medium to old-aged forest stands. Stand-replacing clearcutting methods were launched in the 1950s, gradually transforming the landscape into a checkerboard mosaic of clearcuts and even-aged, young plantations. By the time of 1979, half of the older, semi-natural forest had been replaced, and in 2008, less than 25% remained (Fig. 1). Harvested blocks range between 5 and 50 ha in size, but most clearcuts are in the lower end of this range. Densities of capercaillie and black grouse at Varaldskogen are low to medium-high, with spring densities of lekking males varying between 0.5 and 1 capercaillie and 1 and 2 black grouse per km 2. The Russian study area is centrally located ( N, E) within the Pinega State Forest Reserve (515 km 2 ) in the Archangelsk Region of northwestern Russia (Wegge et al. 2005a). It belongs to the northern boreal subzone, being more continental than Varaldskogen, with July and January temperatures averaging 14.3 and C. Annual precipitation averages 600 mm, and the ground is snow-covered from October to mid-may. The terrain is very flat, mostly varying between 70 and 130 m above sea level. The Pinega reserve was established in 1974, partly because of its unique gypsum karst formations. Underground and surface karst formations are mostly confined to the northwestern and southeastern parts of the Reserve (Goryachkin 2006). Grouse leks have been surveyed since Here we report results from the central 240 km 2 section of the reserve, which was thoroughly surveyed between 2000 and 2005 (Fig. 1). This part of the landscape contains less karst, and is dominated by large bog complexes and extensive poorly drained stands of spruce (Picea obovata Ledeb. Picea abies) on strongly leached loamy soils. Scots pine occurs in bog areas and on scattered moraine soils. Pinega Reserve is mostly a pristine boreal landscape comprising large blocks ( ha) of old forest (50%), fireregenerated younger forest (25%), and open bogs (25%). The landscape dynamics is governed by large forest fires, with return intervals >100 years. Old forests (>100 years) are mainly spruce dominated (80%), admixed with smaller stands of pine (20%) scattered with older birch (Betula spp.) and Siberian larch (Larix sibirica Ledeb.) trees. Secondary successional forests (<40 years) are dominated by birch, admixed with patches of old pine and larch that survived the fires. A few logging operations were conducted in the s, confined to two restricted patches in the eastern and central part covering 6.5% of the study area. These cuttings differed from recent clearcuts at Varaldskogen in leaving more residual green trees and snags within the harvested areas, thereby resembling the forest structure following natural fires. Densities of capercaillie and black grouse are low to medium-high, varying between 0.5 and 1 lekking males per km 2. Fieldwork The study areas were surveyed for display tracks with snowmobile and by cross-country skiing in late winter early spring. When display tracks were found, the sites were repeatedly visited to confirm that birds were using the sites on a regular basis, and to estimate how many birds were present at each site. Well-trained field personnel counted number of displaying males during early morning hours during peak lekking season in late April early May. In most cases, leks were easily identified by site-specific, regularly attending males. In capercaillie, this also applied to a few solitary displaying males (five cases at Varaldskogen), which therefore were included in the data set. Capercaillie males born the previous summer do not display their first spring. This segment of the population was therefore not included in the survey, although they were often ob-

5 1036 Can. J. Zool. Vol. 87, 2009 Table 2. Model parameters of OLS regression of distance between neighboring leks (DN) of capercaillie (Tetrao urogallus) against lek size at Pinega, northwestern Russia, and Varaldskogen, southeastern Norway. Slope Intercept N * Coefficient SE Odds (>0) { interval (CI) 99% CI Coefficient SE Odds (>0) { 95% CI 99% CI 95% confidence Pinega :1 0.04, , :1 0.56, , 3.99 Varaldskogen : , , :1 0.98, , 3.43 Varaldskogen :1 0.45, , :1 0.34, , 4.01 Varaldskogen combined :1 0.15, , :1 0.20, , 2.65 All leks combined >1000:1 0.52, , :1 0.22, , 1.84 Note: For an explanation of the model see the Materials and methods. *The number of DN readings averaged 1.7 readings per lek. Lek sizes averaged between pairs of compared leks. { Odds = [1 (P/2)]/(P/2). served at the lek during peak season. In black grouse, single males and small groups of 2 3 males were sometimes observed displaying on an irregular basis, especially prior to the peak in matings. Such vagrant singletons and small groups were not included in the data set unless it was confirmed that they were faithful to the display site. This segment of the population presumably consisted mostly of yearlings, which do display in their first spring. Almost without exception, aggregations of displaying males (leks) were clearly separated from neighboring aggregations. Between-male distances were and m in black grouse and capercaillie, respectively, whereas interlek distances were 800 and 1000 m, respectively. Exceptions were two capercaillie males displaying 200 m from each other on each side of an open bog, and two groups of black grouse displaying 200 m from each other on each side of a narrow strip of pine trees within a large bog complex. These two cases were both treated as one lek. Male capercaillie were captured on leks by luring the birds into ground nets using playback of the display song. Male black grouse were captured with drop-traps at leks and in snow burrows during winter. Radio transmitters (142 MHz), weighing <3% of the body mass, were fitted using necklace and backpack harness. Males were located daily during display seasons, and weekly during falls and winters, either by flushing the birds or triangulating their position within a radius <150 m using hand-held antennas (Wegge and Larsen 1987; Gjerde and Wegge 1989; Wegge et al. 2003, 2005b). Data analyses and statistical inference Leks were surveyed and monitored at Varaldskogen from 1979 to 2008 and at Pinega from 1986 to Because of fluctuations in population sizes and because not all leks were counted every year, for the purpose of this study we used mean bird counts from 1980 to 1983 (4 years) and from 2006 to 2008 (3 years) at Varaldskogen, and from 2000 to 2005 (6 years) at Pinega. During these periods, search effort was similar and males of both species were radio-monitored at 2 3 leks in both study areas. We chose two periods at Varaldskogen because the study area was much smaller than at Pinega. Because of intensive logging and subsequent forest regrowth, the landscape mosaic at Varaldskogen changed so much between the two periods (23 years) that the two data sets could be treated as independent samples. In some analyses, we therefore combined them to increase the power of the statistical tests. Digitized maps based on aerial photos, satellite images, and detailed forest survey plans were used to measure interlek distances and habitat composition, using ArcView GIS version 3.3 (Environmental Systems Research Institute (ESRI) Inc., Redlands, California, USA). In this study, we classified the landscape into four main habitats: (1) open bogs; (2) clearcuts, burns, and young forest <40 years; (3) middle-aged forest years, and (4) old forest >70 years. Based on previous knowledge about habitat selection in the two species, we defined open habitats (bogs, burned areas, and clearcuts) and young forest as suitable habitat for black grouse (Hjeljord and Fry 1995; Lindström et al. 1998; Angelstam 2004), and middle-aged and old forest as suitable habitat for capercaillie (Miettinen et al. 2005;

6 Rolstad et al Fig. 3. Distribution of lek sizes of capercaillie (Tetrao urogallus) and black grouse (Tetrao tetrix (= Lyrurus tetrix)) compared with a random Poisson distribution (broken lines) at Pinega, northwestern Russia, from 2000 to 2005; Varaldskogen, southeastern Norway, from 1980 to 1983; and Varaldskogen from 2006 to **, P < 0.01; ***, P < 0.001; 1), probability of 4 leks with 2 males is < Fig. 4. Lek sizes of capercaillie (Tetrao urogallus; left panel) and black grouse (Tetrao tetrix (= Lyrurus tetrix); right panel) in relation to the proportion of suitable habitat surrounding the leks. Symbols denote leks at Pinega, northwestern Russia, from 2000 to 2005 (solid circles); Varaldskogen, southeastern Norway, from 1980 to 1983 (open circles); and Varaldskogen from 2006 to 2008 (open triangles). Surrounding area was within a radius of 1 km in capercaillie (both study areas) and black grouse (Pinega), and within 500 m in black grouse at Varaldskogen. Suitable habitats are middle-aged and old forest in capercaillie (>40 years) and open bogs, clearcuts, and young forest (<40 years) in black grouse. Rolstad et al. 2007). In black grouse, we also assessed the size of open bogs harboring leks by measuring the radius (m) from the center of the lek to the nearest forest edge. Habitat was recorded as the proportion of suitable habitat within a 1 km radius around capercaillie leks. In black grouse, we used a 1 km radius at Pinega and a 0.5 km radius

7 1038 Can. J. Zool. Vol. 87, 2009 Fig. 5. Lek size of black grouse (Tetrao tetrix (= Lyrurus tetrix)) in relation to the size (radius) of open bogs at the lek sites. Symbols denote leks at Pinega, northwestern Russia (circles, N = 14) and at Varaldskogen (triangles, N = 16, mean values are given for leks at the same site in and ). Broken line is secondpower function (y = x 2 ) fitted to the solid symbols (N = 24), tentatively indicating maximum lek sizes in the study areas. at Varaldskogen to comply with the marked differences in interlek distances. Spatial patterns of leks were inferred using two test procedures based on distance-to-nearest-neighbor (DNN). To check for global spacing patterns across the whole study areas, we applied the R statistic of Clark and Evans (1954), as modified by Donnelly (1978). R values >1 indicate overdispersion and R values <1 indicate aggregation, with values of 2.15 corresponding to a perfect regular spacing as in a simple hexagonal close-packed (HCP) lattice and 0 denoting complete overlap. However, the R statistic is not well suited for recognizing small-scale spacing patterns, and may overlook local regularity despite global aggregation or randomness (e.g., Blomqvist and Elander 1988). To examine spacing patterns on a smaller scale, we used the G statistic of local regularity (Brown and Rothery 1978), which is the ratio of the geometric to the arithmetic means of the squared DNNs. Deviations from randomness was tested using random permutations, with G values of indicating increasing local regularity. (Random G values varied from 0.49 to 0.53, depending on sample size.) The Clark and Evans R statistic and Brown and Rothery G statistic (based on DNN) do not take into account the size of the leks. To assess whether the number of lekking males influenced the spacing pattern, we measured the distance from each lek to its nearest neighbor in the four cardinal directions up to a distance of twice the mean DNN. We refer to this parameter as the distance between neighboring leks (DN), as opposed to the distance-to-nearest-neighbor lek (DNN) that we used in the R and G statistical tests. Avoiding DNs being measured twice between pairs of leks, we retrieved 1 4 readings per lek (mean 1.7 readings per lek). This still introduced a certain amount of pseudoreplication (see below). The number of males was calculated as the mean of each pairwise lek comparison, linearized by taking the square-root values. If we assume that males are randomly distributed across the landscape and that leks are initiated (for whatever reason) with a random sample of these males as focal points, each focal male will be surrounded by a drawing area extending halfway to its neighboring focal males. If the other males gather freely from this drawing area to their nearest focal male (which becomes the lek), DN will increase with increasing lek size and decrease with increasing overall density of males in the landscape. In a closed-packed lattice, DN follows: ½1Š DN ¼ 2 0:5 p 0:5 males per lek males per km 2 which could be seen as a null model of ideal free settlement. In an OLS (ordinary least square) regression of DN against lek size, this null model would be supported by a positive slope and zero intercept, meaning that DN increases with lek size at a constant density of males in the drawing areas. Increased overall density of males would steepen the slope. A positive intercept would suggest that bigger leks are situated closer together (and smaller leks farther apart) than predicted from the null model (i.e., higher density of males within the drawing areas of big leks), whereas a negative intercept would indicate the opposite. A zero or negative slope means that DN is constant or actually decreasing with increasing lek size, a pattern that is possible only if male density is very much higher around big leks. Deviations from the null model (i.e., positive slope and zero intercept) were assessed by means of odds ratios and confidence intervals of the likely range of the true value (Oakes 1986; Hopkins 2002; Lombardi and Hurlbert 2009). To comply with the certain degree of pseudoreplication in reading DNs, we present both 95% and 99% confidence intervals (CI) of slope and intercept. We examined if the distribution of lek sizes departed from random, i.e., if there was an excess of smaller or larger leks, by ranking the leks from the biggest to the smallest and testing it against a random Poisson distribution. Relationships between variables were inferred from Pearson product moment correlation coefficients (r). Non-normal variables were transformed with square root or logarithmic functions, depending of the degree of right skewness. Differences among means of different landscapes (Varaldskogen , , and Pinega ) were tested using ANOVA followed by Bonferroni Dunn post hoc tests. Results Size and spacing of leks Capercaillie Capercaillie leks were regularly spaced in both study areas at mean DNNs of approximately 2 and 3 km at Varaldskogen and Pinega, respectively (Fig. 1, Table 1). The regular spacing was consistent at both regional (whole study areas) and local scales. Whereas overall density of males was similar in both landscapes, mean lek size was twice as big in Pinega compared with Varaldskogen (F [2,36] = 8.24, P = 0.001, P Pinega vs. Varald 1980 = 0.003, P Pinega vs. Varald 2006 = 0.001). Mean lek size was similar at Varaldskogen during the two periods (P Varald 1980 vs. Varald 2006 = 0.96) (Table 1).

8 Rolstad et al Fig. 6. Spatial distribution of radio-marked male black grouse (Tetrao tetrix (= Lyrurus tetrix)) during seasons of peak lek attendance (spring and fall), shown for two neighboring leks at Pinega, northwestern Russia, in 2002 (left panel) and for two neighboring leks at Varaldskogen, southeastern Norway, from 2006 to 2008 (right panel). These leks are marked with (g) and (h) in Fig. 1. Total number of attending males, number of radio-marked males, and number of individual locations are as follows: Pinega (lower lek, open circles): 42 males, 15 marked, 86 fixes. Pinega (upper lek, solid triangles): 18 males, 7 marked, 56 fixes. Varaldskogen (left lek, open circles): 10 males, 10 marked, 70 fixes. Varaldskogen (right lek, solid triangles): 8 males, 8 marked, 47 fixes. Circles are drawn at 0.5 and 1 km radii. In both study areas, the distance between neighboring leks (DN) increased with increasing number of males attending the leks (Fig. 2). The odds in favor of a positive slope were 52:1 and 124:1 at Pinega and Varaldskogen (both periods combined), respectively (Table 2). Except for Varaldskogen ( ), all individual 95% CIs were >0. However, among the 99% CIs, only the interval for all leks combined was >0. Alltogether, we interpret the statistics fairly strongly in favor of a positive slope, i.e., supporting the null model (eq. 1). When interlek distances were compared with the null model (eq. 1), bigger leks tended to be closer together and smaller leks farther apart than predicted, being statistically highly likely only when all leks were pooled (odds in favor of positive intercept 180:1; Table 2). The pattern was consistent between the two periods at Varaldskogen, and between study areas, thereby justifying pooling of the data. This suggests that density of males was slightly higher within the drawing areas of the bigger leks compared with the smaller leks, thereby presumably deviating from the null model (Fig. 2). The size distributions of leks at Varaldskogen did not differ from a random Poisson distribution, implying that there was no excess of neither big nor small leks. This also was the case at Pinega, except for one lek (24 males), which was bigger than expected (Fig. 3). Black grouse At the scale of the total study areas (using the R statistic), leks of black grouse were more irregularly distributed; they were not different from a random model at Varaldskogen and tended towards clumping at Pinega. However, when examined at a smaller scale (using the G statistic), leks were regularly spaced in both study areas, at mean DNNs of approximately 1 and 2 km at Varaldskogen and Pinega, respectively (Fig. 1, Table 1). Thus, leks of black grouse were distributed less regularly and at closer neighbor distances than leks of capercaillie. Notably, whereas overall density of lekking black grouse was 4 times lower at Pinega than at Varaldskogen, mean lek size was similar (F [2,35] = 0.02, P = 0.98, P Pinega vs. Varald 1980 = 0.85, P Pinega vs. Varald 2006 = 0.99, P Varald 1980 vs. Varald 2006 = 0.85). In contrast to capercaillie, there was no relationship between lek size and interlek distance in black grouse (odds in favor of slope >0 ranging 1.0:1 to 1.4:1, Fig. 2). At Varaldskogen, the size distribution of leks followed a random Poisson distribution, but there was an excess of big leks and a deficiency of small leks at Pinega (Fig. 3). Habitat Capercaillie At Pinega, capercaillie leks were distributed across the whole study area, except for a large open bog complex in the southwest corner (Fig. 1). Although most leks were situated in old forest, three leks were located within a large fireregenerated area in the eastern part. Although the fire occurred 60 years ago, much of the forest was younger than

9 1040 Can. J. Zool. Vol. 87, years owing to poor regeneration. Two of these leks were rather big, with 14 and 13 lekking males, respectively. When the number of lekking males was regressed against the amount of old forest within a circle of 1 km radius, no relationship was discernable (r = 0.04, P > 0.20, N = 17) (Fig. 4). This also was the case when we included middleaged forest and removed the three leks in the fire area (r = 0.30, P > 0.20, N = 14). At Varaldskogen, capercaillie leks were also distributed throughout the whole study area. Between 1980 and 1983, there was a significant positive relationship between lek size and amount of old forest within a 1 km radius (r = 0.59, P = 0.005, N = 21) (Fig. 4). Here, 11 leks from a larger surrounding area were included, but the pattern was upheld when only leks within the core area were retained in the model (r = 0.63, P = 0.05, N = 10). Between 2006 and 2008, no significant relationship was observed, although a positive trend was noted (r = 0.43, P = 0.16, N = 12). Combining the two data sets from Varaldskogen confirmed the positive relationship between number of lekking males and amount of old forest, which became highly significant because of the larger sample size (r = 0.56, P < 0.001, N = 33). Justified by the same density of males in the two study areas, we included the Pinega data, which resulted in a poorer fit of the model but still highly significant because of the increased statistical power (r = 0.51, P < 0.001, N = 50). Including middle-aged forest stands as an explanatory variable in the model did not change the results notably. This forest age class was almost nonexistent at Pinega and at Varaldskogen between 1980 and 1983, but it made up a significant proportion (30%) at Varaldskogen between 2006 and Black grouse At Pinega, leks of black grouse were located at open bogs, largely confined next to the younger fire-regenerated forest in the east and to the huge bog complex in the southwest (Fig. 1). Only two leks were located in the central area dominated by old forest. Neither the proportion of young forest (<40 years), the proportion of open bogs, nor a combination of the two habitats within a 1 km surrounding radius explained the number of lekking males at Pinega (r < 0.30, P > 0.20, N = 14 for all three comparisons). The size distributions of the leks were, however, highly skewed to the right, towards larger leks (Fig. 3). Notably, two small leks (with two males each) were located among several larger ones within the burnt-over young forest in the east surrounded by >90% suitable habitat (Fig. 4). Excluding these two leks from the data set revealed a significant positive relationship between the amount of young forest and open bog, combined, and the number of lekking males (r = 0.58, P = 0.05, N = 12; Fig. 4). At Varaldskogen, leks of black grouse invariably were located at open bogs, which were scattered throughout the study area (Fig. 1). Similar to Pinega, neither the amount of young forest nor the amount of open bogs (within a 0.5 km radius) explained the variation in lek size. However, when these habitats were combined, lek size increased significantly with increasing proportion of suitable habitat, both between 1980 and 1983 (r = 0.68, P = 0.03, N = 10) and between 2006 and 2008 (r = 0.63, P = 0.02, N = 14). Because of the marked differences in densities of males between the study areas, we did not combine the data sets. In both study areas, leks of black grouse increased with increasing size of the open lek arena itself (Pinega: r = 0.56, P = 0.05, N = 13, excluding one lek at bog radius of 500 m; Varaldskogen: r = 0.72, P = 0.002, N = 16, using the mean number of males at leks located at the same bog in both periods). To tentatively indicate maximum lek sizes in relation to the size of the open bogs in our study areas, we least-square-fitted a second-power function to the pooled data set, without six leks at Pinega that apparently had room for more males (lek size = bog radius 2 ; Fig. 5). Spacing of males outside the lek During the display season, male capercaillie in both study areas lived solitarily in daytime ranges within ~1 km radius of the lek center (Varaldskogen: Wegge and Larsen 1987; Wegge et al. 2005b; Pinega: Wegge et al. 2003). This happened even though interlek distances were longer and male daytime ranges overlapped more at Pinega. Males were not strictly territorial outside the lek, but site-related or temporal dominance behavior still produced a spaced-out distribution of individuals in both study areas. For the most part, males also lived solitarily during winter, within overlapping home ranges that extended farther out from lek centers than during spring, but still closest to their home lek (Varaldskogen: Gjerde and Wegge 1989; Pinega: Lande 2002). Male black grouse, on the other hand, were mostly social and stayed within groups of varying sizes throughout fall, winter, and spring seasons. Although socializing with members of the same lek group during peak seasons of lekking activity (fall and spring), radio-tracking data indicated that males were distributed in separate, nonoverlapping, lek populations (Fig. 6). Daily cruising radius was, however, larger at Pinega than at Varaldskogen, with males from different Pinega leks overlapping during winter. Discussion Leks of capercaillie were regularly distributed in both landscapes at interlek distances of approximately 2 3 km, increasing with increasing number of attending males. This result is in accordance with earlier studies invoking a malebased component of lek spacing (Wegge and Rolstad 1986; Beshkarev et al. 1995; Storch 1997). Habitat contributed to the spacing pattern of leks by influencing the number of attending lekking males. Large blocks of old and middle-aged forest provide daytime ranges for more males than smaller, scattered forest patches. As hypothesized by Rolstad and Wegge (1987b), this resulted in bigger leks being spaced farther apart in the landscape dominated by older continuous forest at Pinega compared with the fragmented landscape at Varaldskogen. However, bigger leks tended to be slightly closer together than predicted from the null model of ideal free settlement, i.e., density of males was higher within the drawing area of the biggest leks. This presumably is the outcome of male home ranges decreasing in size with increasing proportion of suitable habitat, as shown by Wegge and Rolstad (1986). Thus, spacing of leks in capercaillie seems to be a product of male spacing behavior and habitat compo-

10 Rolstad et al sition, acting in concert to produce a highly regular spatial pattern. Lek size of capercaillie increased with increasing amount of suitable habitat within a 1 km surrounding radius. Still, lek size did not exceed males. What prevents leks from growing bigger? Since males mutually avoid each other when off the lek (Wegge et al. 2005b), an increasing number of males would require a successively larger surrounding area. Commuting between daytime ranges and lek territories may become increasingly costly at longer distances. This could be resolved by male ranges overlapping more on bigger leks, as has been documented at Pinega (Wegge et al. 2003) and in the Bavarian Alps (Storch 1997). However, a threshold level of density might exist at which younger males, which recruit from the outside (Wegge and Larsen 1987), choose smaller leks as their destination, thereby increasing their lifetime probability of achieving mating success. Although male capercaillie are not territorial in the common sense outside the central lek area, at big leks they appear to be temporarily spaced farther apart from each other than expected from a random distribution (Wegge et al. 2003). It is therefore conceivable during population peaks that high numbers of attending males contribute to space out leks maximally. These leks may remain spaced out at the same interlek distances also during subsequent population lows, especially in landscapes with low frequency of disturbance from logging or fire. Thus, although intralek spacing of males may not limit lek size up to a certain threshold level, temporal mutual avoidance behavior still may be the ultimate reason for leks to be spaced out. Whereas leks of black grouse for the most part remained at the same locations throughout the study period, leks of capercaille were spatially more variable, especially at Varaldskogen. Here, the shifting locations of leks conforms to the hypothesis of a female-based (late-winter hotspot) mechanism governing the formation of new leks in the interlek vacant zones (Gjerde et al. 2000; Eliassen and Wegge 2007). The highly shifting landscape mosaic caused by logging at Varaldskogen presumably speeds up the dynamic process of lek formation in capercaillie, thereby preventing leks from becoming as large as those at Pinega. In black grouse, habitat appeared more influential than male spacing behavior in determining the distribution and size of leks, at least at a regional scale. Leks were considerably less regularly distributed and were at shorter interlek distances than in capercaillie. First, the size of open bogs used for lekking was an important prerequisite, as has been shown in other studies (Hjeljord and Fry 1995). Second, provided open bogs were large enough, the amount of surrounding open habitat and young forest appeared important in determining number of attending males, which also conforms to earlier accounts (Angelstam 2004). At the scale of the total study areas, the importance of habitat appeared even more pronounced. In the burnt-over area of eastern Pinega, overall density of lekking black grouse was 10 times higher (1.16 males/km 2 ), and density of leks 5 times higher (11 leks/100 km 2 ), than in the central old-growth-dominated area (0.10 males/km 2 and 2 leks/100 km 2, respectively) (Fig. 1). Social behavior may also explain why there were vacant spaces devoid of leks of black grouse, most pronounced at Pinega, but also evident at Varaldskogen (Fig. 1). Possibly, suitable lek sites in these areas were not occupied because male black grouse were socially attracted to already established leks. Thus, at the landscape scale, social attraction among male black grouse may cause leks to be more aggregated, whereas mutual avoidance among male capercaillie space leks out. At local scales, leks of black grouse were regularly spaced. Thus, also in black grouse it appears to be behavioral mechanisms that prevent leks from getting closer than a certain minimum distance of ~1 km. In contrast to capercaillie, male black grouse are highly social during the breeding season. Hence, mutual avoidance in more or less solitary daytime ranges, as seen in capercaillie, cannot explain interlek spacing. However, in both study areas we found little overlap between males from neighboring leks during seasons of lek activity (Fig. 6). At Pinega, males overlapped during winter, although they were never observed in multilek flocks. At Varaldskogen, winter flocks of males almost exclusively ranged within 500 m of the leks (P. Wegge and J. Rolstad, unpublished data). A certain degree of local interlek spacing may be the net result of local lek groups using more or less exclusive spring ranges. This was the case in a lek population of Scottish black grouse (Johnstone 1969; Robel 1969), and has been suggested to apply to lekking sharptailed grouse (Tympanuchus phasianellus (L., 1758)) in Alberta, Canada (Rippin and Boag 1974). Where does this leave us with respect to the dichotomy of forest grouse that are displaying solitarily in dispersed territories versus open habitat and steppe grouse that are congregating at communal leks? A closer look at the lekking species reveals that the sharp-tailed grouse might be the North American counterpart of black grouse. In the northern parts of its distribution range (Tympanuchus phasianellus caurus (Friedmann, 1943), Tympanuchus phasianellus kennicotti (Suckley, 1862), and Tympanuchus phasianellus phasianellus (L., 1758)), sharp-tails spend a large part of their time arboreal, described as being intermediate between forest grouse and open lekking prairie chickens (Tympanuchus cupido (L., 1758) and Tympanuchus pallidicinctus (Ridgway, 1873)) in terms of habitat selection (Hamerstrom and Hamerstrom 1951). Leks of sharp-tails seem to be smaller, in terms of attending males, in these northern grass woodland mosaic landscapes (Bergerud 1988). In forested areas short of suitable open habitat for lekking, black grouse has been shown to display mainly solitary and dispersed (Bocca 1995; Höglund and Stöhr 1997). Thus, lekking in grouse appears unanimously associated with group living in open habitat. Presumably, there are multiple reasons for male grouse to aggregate, such as female preference (Bradbury 1981; Alatalo et al. 1992), hotshot attractiveness (Beehler and Foster 1988), and female hotspots (Bradbury and Gibson 1983; Gjerde et al. 2000). Although these mechanisms may account for the ultimate reason of lekking, none of them convincingly explains why lekking grouse almost exclusively are social and confined to open habitat. The capercaillie might hold a key to this issue. Lekking in dense habitat actually may be more dangerous in terms of predator attacks than lekking in the open (Koivisto 1965; Hjorth 1970; Wiley

11 1042 Can. J. Zool. Vol. 87, ; Bergerud 1988; Höglund and Alatalo 1995), although antipredation behavior may not be the ultimate reason for lekking. Being the largest member of the grouse subfamily, the capercaillie clearly is less vulnerable than the smaller black grouse and hazel grouse to predation from northern goshawk (Accipiter gentilis (L., 1758)), the main predator on adult grouse in boreal Eurasia (Widén 1987; Tornberg 2001; Borchtchevski et al. 2003). In particular, this applies to male capercaillie during lekking season, when female goshawk is incubating, and only the smaller male is hunting. Hence, there has been little selective pressure on male capercaillie to lek in open habitat. Acknowledgements This contribution was mainly funded by the Norwegian Directorate for Nature Management and the Norwegian Research Council. Additional support was received from the Forestry Division of the Ministry of Food and Agriculture to J.R. and K.O.S. through the Norwegian Forest and Landscape Institute. In recent years, fieldwork was facilitated with the assistance of Morten Odden, Mats H. Finne, Øyvind Andreassen, Kai Lande, Svein Bjørk, Anders Halland, Trond Wegge, John Gunnar Dokk, and Erlend Rolstad. References Ahti, T.L., Hämet-Ahti, L., and Salas, J Vegetation zones and their sections in north western Europe. Ann. Bot. Fenn. 5: Alatalo, R.V., Höglund, J., and Lundberg, A Lekking in black grouse a test of male viability. Nature (London), 352(6331): doi: /352155a0. Alatalo, R.V., Höglund, J., Lundberg, A., and Sutherland, W.J Evolution of black grouse leks: female preferences benefit males in larger leks. Behav. Ecol. 3(1): doi: / beheco/ Angelstam, P Habitat thresholds and effects of forest landscape change on the distribution and abundance of black grouse and capercaillie. Ecol. Bull. 51: Beehler, B.M., and Foster, M.S Hotshots, hotspots, and female preference in the organization of lek mating systems. Am. Nat. 131(2): doi: / Berger, D.D., Hamerstrom, F., and Hamerstrom, F.N., Jr The effect of raptors on prairie chickens on booming grounds. J. Wildl. Manage. 27(4): doi: / Bergerud, A.T Mating systems in grouse. In Adaptive strategies and population ecology of northern grouse. Edited by A.T. Bergerud and M.W. Gratson. University of Minnesota Press, Minneapolis. pp Beshkarev, A.B., Blagovidov, A., Teplov, V., and Hjeljord, O Spatial distribution and habitat preference of male capercaillie in the Pechora-Illych Nature Reserve. In Proceedings of the 6th International Grouse Symposium, Udine, Italy, September Edited by D. Jenkins. World Pheasant Association, P.O. Box 5, Reading, UK, and Istituto Nazionale per la Fauna Selvatica, Ozzzano dell Emilia, Italy. pp Blomqvist, S., and Elander, M King Eider (Somateria spectabilis) nesting in association with Long-tailed Skua (Stercorarius longicaudus). Arctic, 41: Boag, D.A., and Sumanik, K.M Charcteristics of drumming sites selected by ruffed grouse in Alberta. J. Wildl. Manage. 33(3): doi: / Bocca, M Dispersion and habitat selection of displaying male Black Grouse in the Mont Avic Natural Park, western Italian Alps. In Proceedings of the 6th International Grouse Symposium, Udine, Italy, September Edited by D. Jenkins. World Pheasant Association, P.O. Box 5, Reading, UK, and Istituto Nazionale per la Fauna Selvatica, Ozzzano dell Emilia, Italy. pp Borchtchevski, V.G., Hjeljord, O., Wegge, P., and Sivkov, A.V Does fragmentation by logging reduce grouse reproductive success in boreal forests? Wildl. Biol. 9: Bradbury, J.W The evolution of leks. In Natural selection and social behavior. Edited by R.D. Alexander and D.W. Twinkle. Chiron Press, New York. pp Bradbury, J.W., and Gibson, R.M Leks and mate choice. In Mate choice. Edited by P.P.G. Bateson. Cambridge University Press, Cambridge, UK. pp Brown, D., and Rothery, P Randomness and local regularity of points in a plane. Biometrika, 65(1): doi: / biomet/ Clark, P.J., and Evans, F.C Distance to nearest neighbor as a measure of spatial relationships in populations. Ecology, 35(4): doi: / de Vos, G.J Adaptedness of arena behaviour in black grouse (Tetrao tetrix) and other grouse species (Tetraoninae). Behaviour, 68(3): doi: / x de Vos, G.J Social behaviour of black grouse: an observational and experimental field study. Ardea, 71: Donnelly, K.P Simulations to determine the variance and edge effect of total nearest-neighbour distance. In Simulation methods in archaeology. Edited by I.R. Hodder. Cambridge University Press, Cambridge, UK. pp Eliassen, S., and Wegge, P Ranging behaviour of male capercaillie Tetrao urogallus outside the lekking ground in spring. J. Avian Biol. 38(1): doi: /j x. Gibson, R.M., and Bradbury, J.W Lek organization in sage grouse: variations on a territorial theme. Auk, 104: Gjerde, I., and Wegge, P Spacing pattern, habitat use and survival of Capercaillie in a fragmented winter habitat. Ornis Scand. 20(3): doi: / Gjerde, I., Wegge, P., and Rolstad, J Lost hotspots and passive female preference: the dynamic process of lek formation in capercaillie Tetrao urogallus grouse. Wildl. Biol. 6: Goryachkin, S.V The Fourth International Conference on Cryopedology (Arkhangelsk Pinega, August 1 8, 2005). Eurasian Soil Sci. 39(10): doi: / S Hamerstrom, F.N., Jr., and Hamerstrom, F Mobility of the sharp-tailed grouse in relation to its ecology and distribution. Am. Midl. Nat. 46(1): doi: / Hamerstrom, F.N., Jr., and Hamerstrom, F Status and problems of North American grouse. Wilson Bull. 73: Hjeljord, O., and Fry, G The size of black grouse lek populations in relation to habitat characteristics in southern Norway. In Proceedings of the 6th International Grouse Symposium, Udine, Italy, September Edited by D. Jenkins. World Pheasant Association, P.O. Box 5, Reading, UK, and Istituto Nazionale per la Fauna Selvatica, Ozzzano dell Emilia, Italy. pp Hjorth, I Reproductive behaviour in Tetraonidae, with special reference to males. Viltrevy (Stockh.), 7(4): Höglund, J., and Alatalo, R Leks. Princeton University Press, Princeton, N.J. Höglund, J., and Stöhr, S A non-lekking population of Black