JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.. British Ecological Society is collaborating with JSTOR to digitize, preserve and extend access to Journal of Ecology.Summary 1 Changes in demography and studies on physical condition of the Riviere George caribou Rangifer tarandus herd have suggested that its size may be primarily regulated by the amount of forage available on the summer range. 2 We therefore document the impact of grazing and trampling on composition and productivity of two plant communities, the shrub tundra and stands of dwarf birch, within this range. Ungrazed sites were rare, but four previously located small areas were used as control sites. 3 For the shrub tundra, the lichen mat was absent in grazed sites and ground previously occupied by lichens was either bare, covered by fragments of dead lichens and mosses or recolonized by early succession lichen species. Ground cover of shrubs not eaten by caribou was lower in grazed sites than in ungrazed sites, and coverage of graminoids, forage shrubs and forbs did not differ significantly between grazed and ungrazed sites. 4 In stands of dwarf birch grazed by caribou, ground cover and leaf biomass of Betula glandulosa was significantly lower than in ungrazed sites. 5 Productivity of forage plant species over the summer range was estimated at 22.5 g m-2 year-l in an ungrazed condition compared to 10.3 g m-2 year' when grazed. 6 At the landscape level, caribou have fragmented the distribution of their food resource by reducing biomass of shrub tundra and stands of dwarf birch to a very low level. 7 The serious negative impact of migratory ungulates on plant productivity of their summer range may be explained by characteristics of the vegetation and the high carrying capacity of winter compared to summer ranges. Significant factors related to the vegetation are its low resilience and productivity and the absence of a response of vascular plants following removal of lichens.

The range used for calving and for the first month of lactation by the Rivière George Caribou Herd (RGH), which peaked at over 600 000 individuals in the mid-eighties, showed signs of overgrazing, in contrast to that used by the adjacent Rivière aux Feuilles Herd. Density of females in the tundra habitat below 600 m asl averaged 11.2 animals/km2 on the overgrazed range in 1988, in comparison with ≈0.5/km2 on the other range. Inadequate foraging conditions during the first month of lactation caused complete exhaustion of fat reserves in females on the poor range; milk production was seemingly insufficient, as calves were lighter and grew at a slower rate than calves born on the better summer range or born in captivity from dams fed ad libitum. Moreover, females on the poor range were shorter and lighter than their counterparts on the good range, illustrating that priority goes to reproduction rather than to growth under suboptimal foraging conditions. Accretion of body fat and protein in RGH females was rapid in autumn, but individuals with a calf at yield accumulated smaller fat reserves than nonreproductive adult females. Condition did not affect ovulation. However, interruption of gestation was suspected in females that were unable to accumulate enough body reserves in autumn or early winter, judging from the progressive decline in the pregnancy rate and in the autumn cow:calf ratio observed for the RGH since 1984, concomitant with the levelling off of the herd. Summer nutrition seemingly regulated the RGH through a combination of decreased fecundity and survival.

In southwestern Québec, non-harvested moose populations stabilize at a density of ≃0.40 animal·km. In an attempt to test whether or not moose were regulated by predators, we investigated wolf predation near this equilibrium density (0.37) and at 2 lower densities (≃0.23, 0.17). Scat analysis in summer and feeding observations in winter indicated a greater use of alternative food resources by wolves at lower moose densities. Each wolf pack killed on average 5.3, 1.8, 1.1 moose·100 days in the area of 0.37, 0.23, and 0.17 moose·km, respectively. Consumption of moose per wolf was 2.8, 1.7, and 1.6 kg/day, respectively. January wolf densities were estimated at 1.38, 0.82, and 0.36 animals·100 km, respectively. Year-long predation rates proved to be density-dependent, increasing with moose density from 6.1 to 19.3% of the postnatal populations. We conclude that moose populations in southwestern Québec are regulated largely by predators (wolves and maybe black bears) at a density where competition for forage produces no detrimental effect. We support the concept that wolf predation can have an important regulatory effect at low moose densities but also a depensatory (inversely density-dependent) effect at higher densities.

Carnivores living in areas of deep snow face additional energy expenditures during winter owing to increased locomotory costs. Such costs may vary in function of snow depth and hardness (sinking depth of animal) and travel speed. We estimated energetic costs of locomotion through snow in wild coyotes (Canis latrans) using three coyote-sized domestic dogs (Canis familiaris) to develop regression models predicting heart rate (as surrogate for energy expenditure) in relation to sinking depth and travel speed. In the absence of snow, heart rates of dogs increased linearly with travel speed (R 2 = 0.24), whereas when snow was present, track sinking depth affected heart rate substantially more than did travel speed. To assess whether our results with domestic dogs could help explain the behaviour of wild coyotes, we snow-tracked coyotes in southeastern Quebec, Canada, during two winters. During a normal harsh winter, coyotes relied on artificially packed snow (snowmobile and animal trails) more than during a mild winter. Coyotes typically exerted a fine-scale selection for snow depth and hardness that effectively reduced their sinking depth by~2 cm. We estimated that travelling over snow increased coyote heart rate by 4%-6% in comparison with locomotion on hard surfaces, whereas fine-scale selection saved a similar amount of extra energy. We hypothesize that the use of snow packed by anthropogenic activities, especially snowmobile trails, may not only facilitate coyote movements in deep snow environments but also allow occupation of marginal habitats such as forested areas of northeastern North America.Résumé : Les carnivores qui vivent dans les régions où la neige est épaisse ont des dépenses énergétiques additionnelles durant l'hiver à cause des coûts accrus de la locomotion. Ces coûts peuvent varier en fonction de l'épaisseur et de la consistance de la neige (profondeur d'enfoncement de l'animal) ainsi que de la vitesse de déplacement. Nous avons estimé les coûts énergétiques de la locomotion dans la neige de coyotes (Canis latrans) sauvages en utilisant trois chiens domestiques (Canis familiaris) de la taille de coyotes pour mettre au point des modèles de prédiction du rythme cardiaque (comme indice de la dépense d'énergie) en fonction de la profondeur d'enfoncement et de la vitesse de dé-placement. En l'absence de neige, le taux cardiaque des chiens augmente avec la vitesse de déplacement selon une fonction linéaire (R 2 = 0,24); dans la neige, cependant, la profondeur d'enfoncement des pistes affecte considérable-ment plus le rythme cardiaque que ne le fait la vitesse de déplacement. Pour vérifier si les résultats obtenus chez les chiens pouvaient servir à expliquer le comportement de coyotes sauvages, nous avons suivi durant deux hivers les pistes de coyotes dans le sud-est du Québec, Canada. Durant un hiver rigoureux ordinaire, les coyotes utilisent la neige tassée artificiellement (par les motoneiges ou par le passage des animaux) plus que durant un hiver doux. Les coyotes font normalement une sélection de la ...

Optimality models of food selection by herbivores assume that individuals are capable of assessing forage value, either directly through the currency used in the model or indirectly through other variables correlated with the currency. Although energy and protein are the two currencies most often used, controversy exists regarding their respective influence on food choice. Part of the debate is due to the difficulty of teasing apart these two nutrients, which are closely correlated in most natural forages. Here we offer a test of the assumption that energy and protein contents of the forage are both currencies that large mammalian herbivores can use when selecting their food. We observed feeding behavior of 47 wild white-tailed deer (Odocoileusvirginianus) during winter while individuals were presented with four experimental foods representing two levels of energy and protein (dry matter digestibility: 40-50%; crude protein: 12-16%). Using experimental foods allowed us to separate the influences of energy and protein and clearly distinguish between the roles of these two nutrients. Deer discriminated between foods through partial selection, and selected diets higher in energy but lower in protein. The observed choices appeared consistent with physiological needs of deer wintering at the study site, where digestible energy was in short supply in the natural environment while protein was probably not. Results are in good agreement with recent findings on domesticated ruminants. They support a basic assumption of optimality models of food selection that use energy and/or protein as a currency, although the physiological mechanisms behind the food selection process remain unclear. We urge students of food selection by herbivores to replicate our experiment with other foods and/or in other circumstances before more general conclusions are drawn.