F:\ALCES\Vol_39\p65\3924.PDF ALCES VOL. 39, 2003 PERSSON – VISIBILITY OF MOOSE PELLETS 233 SEASONAL AND HABITAT DIFFERENCES IN VISIBILITY OF MOOSE PELLETS Inga-Lill Persson Department of Animal Ecology, Swedish University of Agricultural Sciences, SE-90183 Umeå, Sweden ABSTRACT: Counts of faecal pellet groups have been widely used to estimate population densities and trends of large ungulates like moose (Alces alces). The visibility of pellet groups affects the accuracy of estimates, decreases with time, and varies among habitat types. I investigated the impact of season and habitat type on how, over time, visibility of moose pellets decreased along a forest productivity gradient in boreal forests of northeastern Sweden. Visibility decreased at the fastest rate during the transition from spring to summer due to concealment by new vegetation. Visibility also varied significantly among habitat types and was correlated with vegetative litter production. After one winter of exposure, more than 95% of all pellet groups were visible independent of habitat type, but thereafter visibility decreased fast in more productive habitats. The results demonstrated that if study plots are cleared in late autumn after the vegetation period and then visited as soon as possible after snowmelt, pellet counts can be used to estimate population trends and habitat use of moose in winter without bias caused by differences in visibility within different habitat types. Also, the correlation with litter production suggests that if a sightability correction factor is developed, pellet counts could be used to estimate habitat use and population distribution during the vegetation period and with longer periods between plot visits. ALCES VOL. 39: 233-241 (2003) Key words: Alces alces, habitat, management, moose, pellet counts, productivity, season, visibil- ity Counts of faecal pellet groups have been widely used to estimate population densities and trends of large ungulates ( W a l l m o e t a l . 1 9 6 2 , N e f f 1 9 6 8 , Timmermann 1974, Harestad and Bunnell 1987). It is a less expensive technique than p o p u l a t i o n e s t i m a t e s f r o m a i r c r a f t (Härkönen and Heikkilä 1999) and might be more precise in forests with dense overstory vegetation (Jordan et al. 1993), although there are methodological problems (Neff 1968). The visibility of pellet groups is an im- portant factor affecting the estimates from pellet counts (Wallmo et al. 1962, Lehmkuhl et al. 1994), but is rarely attributed much importance (Harestad and Bunnell 1987, Aulak and Babinska-Werka 1990). Visibil- ity decreases with time as a result of con- cealment by vegetation and decay proc- esses. The commonly used “clearance- plot” method is based on faecal accumula- tion in previously cleared plots. Correcting for pellet groups which have become invis- ible between consecutive plot visits is criti- cal for the reliability of the method (Massei et al. 1998). How fast visibility decreases depends to a great extent on habitat type (Smith 1968, Lavsund 1975, Harestad and Bunnell 1987, Lehmkuhl et al. 1994, Massei et al. 1998). Visibility decreases faster in moister habitats with more vegetation than in dry habitats with scarce vegetation (Lavsund 1975, Harestad and Bunnell 1987, Lehmkuhl et al. 1994). Faster concealment in more VISIBILITY OF MOOSE PELLETS – PERSSON ALCES VOL. 39, 2003 234 vegetated habitats is suggested to be a primary factor contributing to observed dif- ferences in visibility of pellet groups (Harestad and Bunnell 1987), and vegeta- tive litter is especially important (Lavsund 1975). The decrease of visibility also varies by season and decreases faster during the growing season and in autumn than in win- ter (Aulak and Babinska-Werka 1990, Massei et al. 1998). There is a general lack of information on the fate of moose pellets, especially in areas where moose densities are high and where pellet counts could be an alternative to helicopter surveys and other methods of population estimation. My study objective was to estimate seasonal and habitat differ- ences in visibility of moose pellets (i.e., the percentage of the original surface area of each pellet group still visible). More spe- cifically, I tested how fast visibility de- creased in spring, summer, and autumn re- spectively, if visibility was correlated to habitat productivity, and if some plants and substrates (i.e., the ground cover the pellet groups were lying on) contributed more than others to a decrease in visibility. STUDY AREA The study was done in the middle boreal zone (Ahti et al. 1968) of coastal northern Sweden (Fig. 1), at 8 sites (Table 1) situated 50 - 90 km north and north-west of Umeå (63°50' N, 20°18' E). The length of the vegetation period (average day tempera- ture > 5° C) was 120 - 150 days with onset between 10 and 20 May. Yearly precipita- tion was 600 - 700 mm (Raab and Vedin 1996), and precipitation during the vegeta- tion period was 300 - 350 mm (Nilsson 1996). Snow covered the ground approxi- mately from 20 - 25 October to 5 - 15 May (Raab and Vedin 1996). All sites used in this investigation were young forest stands of Scots pine (Pinus sylvestris), inter- spersed with various deciduous trees; mainly b i r c h e s ( B e t u l a p u b e s c e n s a n d B . p e n d u l a ) , b u t a l s o r o w a n ( S o r b u s aucuparia), aspen (Populus tremula), and willows (Salix spp.). Raspberry (Rubus idaeus), wavy hair grass (Deschampsia f l e x u o s a ) , b l u e b e r r y ( V a c c i n i u m myrtillus), lingonberry (V. vitis-idaea), heather (Calluna vulgaris), and fireweed (Epilobium angustifolium) were common. To prevent bias caused by differences in soil moisture among sites, the sites were selected to have as little slope as possible. METHODS The study sites (Table 1) were selected to represent a forest productivity gradient, covering the range of forest types in the region (Hägglund and Lundmark 1987, Fridman et al. 2001). Site productivity was estimated as Site Index (estimated top height at 100 years) for Scots pine using methods developed for young forest stands (Lindgren et al. 1994, Elfving and Kiviste 1997). Site index is a common measurement of habitat productivity in forestry in Sweden (Persson Fig. 1. Map of Sweden showing the study area. # Umeå # Stockholm N EW S Study area ALCES VOL. 39, 2003 PERSSON – VISIBILITY OF MOOSE PELLETS 235 Table 1. Study sites ranked by increasing Site Index for Scots pine (mean top height in meters at 100 years). Litter production (g dry mass per m2 and year; all treatment plots per exclosure pooled) estimated 2001-02 (Persson 2003), mean age of trees (years), geographic location (WGS84), altitude (m above sea level), and major tree species present. Site Site Litter Mean Geographic Altitude Tree Index Age Location Species1 Lögdåberget 12.9 18.08 16 64o 00' N, 18 o 45' E 300 B, P, Po, (S, Sa) Skatan 14.7 15.85 9 64 o 13' N, 19 o 09' E 265 B, P, (Po, Sa) Djupsjöbrännan 24.3 15.83 9 64 o 06' N, 19 o 12' E 250 B, P, Sa Åtmyrberget 24.8 55.02 9 64 o 12' N, 19 o 17' E 305 B, P, S, Sa, (Po) Selsberget 26.3 11.40 7 64 o 15' N, 19 o 16' E 175 B, P, S, (Po, Sa) Mörtsjöstavaren 26.4 56.38 7 64 o 22' N, 20 o 07' E 280 B, P, S, Sa Ralberget 27.3 27.15 9 64 o 13' N, 20 o 42' E 250 B, P, S, Sa, (Po) Rönnäs 27.9 12.01 9 64 o 02' N, 20 o 40' E 62 B, P, (Po) 1 B = Betula spp., P = Pinus sylvestris, Po = Populus tremula, S = Sorbus aucuparia, and Sa = Salix spp. Tree species occurring sparsely are in brackets. 2003). However, Site Index is developed for conifers, and studies have shown that conifers and hardwoods have fundamen- tally different soil-plant interactions (Ollinger et al. 2002). Litter production might be a better biological measurement of habitat productivity (Persson 2003). Thus, I also tested for relationships between the visibil- ity of pellet groups and vegetation litter production in each plot. The study of pellet visibility was done in connection with an experimental study, where browsing, defecation and urination of different levels of moose population den- sity were simulated in 8 exclosures (i.e., the study sites) along a forest productivity gra- dient (Persson 2003). To simulate defeca- tion, 39 pellet groups of 0.8 litres were laid out at each study site on 4 occasions (May 1999, October 1999, June 2000, and August 2000). The number of pellet groups was estimated based on defecation rates of moose (Persson et al. 2000). The pellet groups were deposited within circles of 35 cm in diameter and marked with a plastic stick in the centre. The moose pellets were col- lected from a nearby moose farm. Pellet groups showing clear evidence of decom- position were not collected. The animals were using mainly natural habitats and had free access to natural food (Nyberg and Persson 2002). It was thus an experimental study with pellet groups of equal size and origin. Twelve randomly selected pellet groups from each of the 4 age classes of pellets (i.e., pellet groups laid out May 1999, Octo- ber 1999, June 2000, and August 2000) were investigated at each site; 48 pellet groups per site and 384 for the whole study (one pellet group was not found in spring and some others were not found in summer and autumn; the total n was therefore 383 in spring, 376 in summer, and 371 in autumn). The reason for the missing pellet groups was that some markers were accidentally kicked down by fieldworkers during the experimental treatment. The same pellet groups were investi- gated in spring (21 - 31 May), in summer (16 VISIBILITY OF MOOSE PELLETS – PERSSON ALCES VOL. 39, 2003 236 - 20 July), and after the vegetation period (22 - 31 October) in 2001. At each visit, visibility (from standing position) was estimated to the nearest 5%. All pellet groups which became invisible disappeared due to concealment by vegeta- tion, not as a result of decay. No signs of other wildlife having disturbed the pellet groups by scratching in search for insects or nematodes were found. The relationships between visibility and Site Index, as well as litter production, were checked. Percentage of the different plants which covered the pellet groups either by growing over them or by litterfall was visu- ally estimated to the nearest 5%, and clas- sified as lichens, mosses, grasses, forbs, dwarf shrubs, raspberry, ferns/horsetails, and litter. Because there were few values for some plant groups, site differences were only tested for the most common groups. Ground cover was recorded as lichens, mosses, grass, or bare ground. Data were analysed statistically using the SAS System for Windows (version 8.2, 2001). The Kruskal Wallis test was used to test for differences in visibility among study sites, plants concealing the pellet groups, and substrate types (Siegel and Castellan 1988). Correlations between visibility, Site Index, and litter production were checked using the Pearson correlation coefficient (Fry 1999) or (if the assumptions for a parametric test failed) the Spearman corre- lation coefficient (Siegel and Castellan 1988). When significant correlations were found, linear regression models were devel- oped. The significance level was set at α = 0.05. RESULTS The mean visibility (the percentage of the original surface area of each pellet group visible) decreased with time (Fig. 2). The largest decrease in visibility was 10% per month, and occurred during the transi- tion from spring to summer after one winter (9 months of exposure). Thereafter, visibil- ity decreased at a slower rate. The mean visibility differed significantly among study sites for all seasons: (spring: Kruskal Wallis test, χ2 = 100.92, P < 0.001; summer: χ2 = 174.25, P < 0.001; and au- tumn: χ2 = 149.60, P < 0.001; df = 7 in all tests, Fig. 3). No correlations were found between the mean visibility and Site Index for all pellet groups pooled or for spring, summer, and autumn respectively for pellet groups laid out in 1999 and 2000 combined (Spearman r, P > 0.16 for all tests) or pellet groups laid out in 2000 (P > 0.07 for all tests). However, there were significant negative correlations between mean visibil- ity and litter production for pellet groups laid out in 1999 and 2000 (Fig. 4a), which could be described by linear regressions (Table 2). For pellet groups laid out in 2000 there was a trend, although not significant, that the mean visibility of all pellet groups pooled decreased with litter production (Fig. 4b). Significant negative correlations were also found between litter production and visibil- ity of pellet groups investigated in spring, summer, and autumn, respectively (Table 2). The relative proportions of plant groups concealing pellet groups differed signifi- 10 15 20 25 30 0 20 40 60 80 100 % v is ib ili ty Months of exposure Fig. 2. The change in percent visibility (mean and standard deviation) with time of moose faecal pellets, all study sites pooled. ALCES VOL. 39, 2003 PERSSON – VISIBILITY OF MOOSE PELLETS 237 cantly among study sites for all seasons (Table 3, Fig. 5). Grasses contributed most to the concealment in all seasons, and at all study sites except Lögdåberget and Skatan. Mosses were also important in all seasons and at all study sites, whereas lichens only were important at Lögdåberget and Skatan. Åtmyrberget and Mörtsjöstavaren had the richest field vegetation, and in addition to grass, raspberry, forbs, ferns, and horsetails covered the pellet groups at those sites in summer, whereas litter was important in autumn. The visibility of pellet groups differed significantly among substrate types for all seasons: (spring: Kruskal Wallis test, χ2 = 73.61, P < 0.001; summer: χ2 = 128.44, P < 0.001; and autumn: χ2 = 114.44, P < 0.001; df = 2 in all tests, Fig. 6). Only 4 pellet groups were on barren ground, and because low sample size makes statistical tests un- reliable, these groups were omitted from the analysis. Visibility was highest on li- chen substrate and lowest on grass. The difference among substrate types was Fig. 3. The mean visibility of moose faecal pel- lets at different study sites ranked according to Site Index, investigated in spring (a), sum- mer (b), and autumn (c). a) b) c) Fig. 4. The negative correlation between mean visibility (%) and habitat productivity esti- mated as litter production (g dry mass per m2 and year): (a) all faecal pellet groups and inves- tigations pooled for pellet groups laid out in 1999 and 2000, exposed for an average of 18 months; and (b) all pellet groups and investi- gations pooled for pellet groups laid out in 2000, exposed for an average of 12 months. a) b) 8 10 12 14 16 18 20 22 24 26 0 20 40 60 80 100 3a ÅtmyrbegetMörtsjöstavaren Skatan Lögdåberget Lögdåberget Skatan Djupsjöbrännan Åtmyrberget Selsberget Mörtsjöstavaren Ralberget Rönnäs Spring % v is ib ili ty Months of exposure 10 12 14 16 18 20 22 24 26 28 0 20 40 60 80 100 Months of exposure 3b Mörtsjöstavaren Åtmyrberget Skatan Lögdåberget Summer % v is ib ili yy 12 14 16 18 20 22 24 26 28 30 0 20 40 60 80 100 3c Months of exposureÅtmyrberget and Mörtsjöstavaren Skatan Lögdåberget Autumn % v is ib ili ty 1.5 2.0 2.5 3.0 3.5 4.0 2.0 2.5 3.0 3.5 4.0 4.5 Ln (% visibility) = 4.27 - 0.90 (Ln Litter) R square = 0.60, p = 0.025 L n M ea n v is ib ili ty ( % ) LnLitter (g per m2 and year) 1.5 2.0 2.5 3.0 3.5 4.0 2.5 3.0 3.5 4.0 4.5 Ln (% visibility) = 4.29 - 0.63 (Ln Litter) R square = 0.49, p = 0.054 L n M ea n v is ib ili ty ( % ) LnLitter (g per m2 and year) VISIBILITY OF MOOSE PELLETS – PERSSON ALCES VOL. 39, 2003 238 higher in summer and autumn than in spring: In summer and autumn, the mean visibility was 13 - 16 times higher for pellet groups lying on lichen than on grass, whereas vis- ibility was 3 times higher in spring. Visibility was 4.5 - 5 times higher for pellet groups lying on moss than on grass in summer and autumn, and only 1.7 times higher in spring. DISCUSSION My study demonstrated that visibility of moose pellets decreased considerably with time. As a result of concealment by new vegetation, the largest decrease occurred in the transition from spring to summer after one winter of exposure. There were large habitat differences in how fast visibility decreased. I found no correlations between visibility and Site Index, whereas visibility was negatively correlated with vegetative litter production. Litter production could therefore be used to predict how fast the visibility of pellet groups decreases. Com- position of the ground cover vegetation was an important factor affecting the visibility of pellet groups. Visibility was highest on lichen-rich sites without grass (i.e., Lögdåberget and Skatan) and lowest on sites with rich field vegetation of grass, forbs, ferns, horsetails, and raspberry (i.e., Åtmyrberget and Mörtsjöstavaren). Pellet groups disappeared as a result of Table 2. Statistics from the regression models (F, P, and R2) describing correlations between mean visibility (%) of faecal pellet groups and vegetative litter production (g dry mass per m2 and year) for pellet groups laid out in 1999 and 2000 combined, as well as in 2000. The regressions are estimated as ln visibility = ln litter for all pellet groups and investigations pooled for each site (“Total”), as well as for all pellet groups at each site of the investigations in spring, summer, and autumn 2001, respectively. 1999 and 2000 2000 Total F = 8.84, P = 0.025, R2 = 0.60 F = 5.69, P = 0.054, R2 = 0.49 Spring F = 9.34, P = 0.022, R2 = 0.61 F = 8.01, P = 0.030, R2 = 0.57 Summer F = 12.08, P = 0.013, R2 = 0.67 F = 8.88, P = 0.023, R2 = 0.60 Autumn F = 13.41, P = 0.011, R2 = 0.69 F = 8.00, P = 0.030, R2 = 0.57 Table 3. Results of Kruskal Wallis test for statistical differences in the relative contribution of various plant groups concealing moose pellet groups among study sites. χ2 and P-values are presented in the table, df = 7 for all tests. Forbs, shrubs, raspberry, ferns, horsetails, and litter (in summer) contributed less to the concealment and were omitted in the tests. Vegetation Spring (χ2, P) Summer (χ2, P) Autumn (χ2, P) Lichens 140.13, < 0.001 165.31, < 0.001 192.11, < 0.001 Mosses 14.13, = 0.049 26.51, < 0.001 37.52, = 0.005 Grass 139.80, < 0.001 139.80, < 0.001 192.11, < 0.001 Litter 161.97, < 0.001 - 207.59, < 0.001 ALCES VOL. 39, 2003 PERSSON – VISIBILITY OF MOOSE PELLETS 239 concealment by vegetation rather than de- cay processes. Studies from Vancouver Island, British Columbia (Harestad and Bunnell 1987), and the Olympic Peninsula, 1 2 3 4 5 6 7 8 0 20 40 60 80 100 5a Rönnäs Ralberget Mörtsjöstavaren Selsberget Åtmyrberget Djupsjöbrännan Skatan Lögdåberget Spring R el at iv e % c o ve r Lichens Mosses Grass Litter Fig. 5. The relative contribution (mean and stand- ard deviation) of the most important vegeta- tion types concealing moose faecal pellet groups at different study sites ranked after Site Index; investigated in spring (a), summer (b), and autumn (c). Forbs, shrubs, raspberry, ferns, horsetails, and litter (in summer) con- tributed less to the concealment and were omitted in the figures. Lichens Mosses Grass 0 20 40 60 80 100 % v is ib ili ty Substrate type Spring Summer Autumn Fig. 6. The visibility (mean and standard devia- tion) of moose faecal pellets lying on different substrate types, all study sites pooled. a) c) 1 2 3 4 5 6 7 8 0 20 40 60 80 100 5b Rönnäs Ralberget Mörtsjöstavaren Selsberget Åtmyrberget Djupsjöbrännan Skatan Lögdåberget Summer R el at iv e % c o ve r Lichens Mosses Grass 1 2 3 4 5 6 7 8 0 20 40 60 80 100 5c Rönnäs Ralberget Mörtsjöstavaren Selsberget Åtmyrberget Djupsjöbrännan Skatan Lögdåberget Autumn R el at iv e % c o ve r Lichens Mosses Grass Litter b) Washington (Lehmkuhl et al. 1994), have reported the opposite. However, their stud- ies were done in areas with milder climate and higher annual precipitation. The rela- tively cold and dry climate in my study area was probably the main explanation for lower decay rates. Vegetation concealment likely results in higher moisture which increases decomposition rate and decreases visibility in more moist and vegetated habitats (Lehmkuhl et al. 1994). It is likely that the sites with the richest field vegetation also offered the moistest conditions at ground level. Reliable methods to estimate population density, habitat use, and distribution of moose are important for wildlife and forest man- agement (Härkönen and Heikkilä 1999). Unless pellet counts are done immediately after snow melt before green up, one needs to know how fast visibility decreases, and habitat differences must be regarded. Af- ter one winter of exposure, more than 95% of the deposited pellet groups were visible (i.e., visibility > 0 from standing position) independent of habitat type in the study area. Population trends and habitat use by moose in winter can thus be estimated with- out biases caused by visibility differences in boreal forest habitats with a relatively dry and cold climate. Because of the fast VISIBILITY OF MOOSE PELLETS – PERSSON ALCES VOL. 39, 2003 240 decline in visibility after the first winter, it is important to clear the study plots in late autumn after the vegetation period and then visit them as soon as possible after snowmelt. However, moose have different food pref- erences in summer and winter (Cederlund et al. 1980, Bergström and Hjeljord 1987), and habitat use and distribution of moose in the vegetation period is also interesting to reveal for moose managers. The correla- tions between visibility and litter production suggest that visibility can be estimated as a function of habitat productivity. More stud- ies of the relationship between visibility and habitat productivity should be done to estab- lish a sightability correction factor for vari- ous habitat types. The pellet count method could then be used to estimate habitat use and population distribution in the landscape during the vegetation period and with longer periods between plot visits. ACKNOWLEDGEMENTS Thanks to Christer Johansson for al- lowing me to collect moose pellets at his moose farm in Bjurholm, and the forest companies Assi Domän and Holmen Skog where the study sites were located. I am also grateful to Professor Kjell Danell and Professor Roger Bergström for help with the planning of the study as well as valuable discussions and comments concerning the manuscript. The study was financed by the Swedish Council for Forestry and Agricul- tural Research and the Swedish Environ- mental Protection Agency (grants to Kjell Danell and Roger Bergström). REFERENCES AHTI, T., L. HÄMET-AHTI, and J. JALAS. 1968. Vegetation zones and their sec- tions in northwestern Europe. Annales Botanici Fennici 5:169-211. AULAK, W., and J. BABINSKA-WERKA. 1990. Estimation of roe deer density based on the abundance and rate of disappear- ance of their faeces from the forest. Acta Theriologica 35:111-120. BERGSTRÖM, R., and O. HJELJORD. 1987. Moose and vegetation interaction in northwestern Europe and Poland. Swedish Wildlife Research Supplement 1:213-228. C E D E R L U N D , G . , H . L J U N G Q U I S T , G . MARKGREN, and F. STÅLFELT. 1980. Foods of moose and roe deer at Grimsö in central Sweden - results of rumen content analyses. Swedish Wildlife Research Supplement 11:169-247. ELFVING, B., and A. KIVISTE. 1997. Con- struction of site index equations for Pinus sylvestris L. using permanent plot data in Sweden. Forest Ecology and Management 98:125-134. FRIDMAN, J., G. KEMPE, P. NILSSON, H. TOET, and B. WESTERLUND. 2001. Forestry statistics 2001. Official Statistics of Sweden. Swedish University of Agri- cultural Sciences, Umeå, Sweden. (In Swedish with English summary). FRY, J. C. 1999. Biological data analysis: a practical approach. Oxford University Press, Oxford, UK. HÄGGLUND, B., and J-E. LUNDMARK. 1987. H a n d l e d n i n g i b o n i t e r i n g m e d Skogshögskolans boniteringssystem. D e l 2 , d i a g r a m o c h t a b e l l e r . Skogsstyrelsen, Jönköping, Sweden. (In Swedish). HARESTAD, A. S., and F. L. BUNNELL. 1987. Persistence of black-tailed deer fecal pellets in coastal habitats. Journal of Wildlife Management 51:33-37. HÄRKÖNEN, S., and R. HEIKKILÄ. 1999. Use of pellet group counts in determining density and habitat use of moose Alces alces in Finland. Wildlife Biology 5:233- 239. JORDAN, P. A., R. O. PETERSON, P. CAMPBELL, and B. MCLAREN. 1993. Comparison of pellet counts and aerial counts for esti- mating density of moose at Isle Royale: ALCES VOL. 39, 2003 PERSSON – VISIBILITY OF MOOSE PELLETS 241 a progress report. Alces 29:267-278. LAVSUND, S. 1975. Investigations on pellet groups. Research Notes 23. Depart- ment of Vertebrate Ecology, Royal College of Forestry, Stockholm, Swe- den. (In Swedish with English sum- mary). LEHMKUHL, J. F., C. A. HANSEN, and K. SLOAN. 1994. Elk pellet-group decom- position and detectability in coastal for- ests of Washington. Journal of Wildlife Management 58:664-669. LINDGREN, D., C. C. YING, B. ELFVING, and K. LINDGREN. 1994. Site index variation with latitude and altitude in IUFRO Pinus contorta provenance experi- ments in western Canada and northern Sweden. Scandinavian Journal of For- est Research 9:270-274. MASSEI, G., P. BACON, and P. GENOV. 1998. Fallow deer and wild boar pellet group disappearance in a Mediterranean area. J o u r n a l o f W i l d l i f e M a n a g e m e n t 62:1086-1094. NEFF, D. J. 1968. The pellet-group count technique for big game trend, census, and distribution: a review. Journal of Wildlife Management 32:597-614. NILSSON, N.-E. 1996. The forest. National Atlas of Sweden. Bokförlaget Bra Böcker, Höganäs, Sweden. NYBERG, Å., and I.-L. PERSSON. 2002. Habi- tat differences of coprophilous fungi on moose dung. Mycological Research 106:1360-1366. OLLINGER, S. V., M. L. SMITH, M. E. MARTIN, R. A. HALLETT, C. L. GOODALE, and J. D. ABER. 2002. Regional variation in foliar chemistry and N cycling among forests of diverse history and composition. Ecology 82:339-355. PERSSON, I.-L. 2003. Moose population density and habitat productivity as driv- ers of ecosystem processes in northern boreal forests. Ph.D. thesis, Swedish University of Agricultural Sciences, Umeå, Sweden. , K. DANELL, and R. BERGSTRÖM. 2000. Disturbance by large herbivores in boreal forests with special reference to moose. Annales Zoologici Fennici 37:251-263. RAAB, B., and H. VEDIN. 1996. Climate, lakes and rivers. National Atlas of Sweden. Bokförlaget Bra Böcker, Höganäs, Sweden. SAS INSTITUTE. 2001. SAS System for Windows. Version 8.2. SAS Institute Incorporated, Cary, North Carolina, USA. SIEGEL, S., and J. N. CASTELLAN. 1988. Nonparametric statistics for the behav- ioural sciences. Second edition. McGraw-Hill Book Company, Singa- pore. SMITH, R.H. 1968. A comparison of sev- eral sizes of circular plots for estimating deer pellet-group density. Journal of Wildlife Management 32:585-591. TIMMERMANN, H. R. 1974. Moose inventory methods: a review. Naturaliste Canadien 101:615-629. WALLMO, O. C., A. W. JACKSON, T. L. HAILEY, and R. L. CARLISLE. 1962. Influence of rain on the count of deer pellet groups. Journal of Wildlife Man- agement 26:50-55.