Most studies of habitat selection by large herbivores focus on the resource availability and interactions with other species, but neglect the importance of an animal being familiar with an area due to past use. Yet, studies of the establishment and retention of territories, home ranges, birth sites, and feeding site choices in experimental settings have shown the importance of spatial familiarity at these scales. We used GPS locations of translocated wapiti Cervus elaphus, resource selection functions (RSF), and time-to-return to examine whether previous site use was important for selection of sites by wapiti in west-central Alberta. To construct RSFs, we used logistic regression that included spatial familiarity (presence of a previous GPS location within a 50-m radius) as well as estimates of herbaceous and shrub biomass, elevation, aspect, slope, and predation risk to wapiti from wolf predation, as dependent variables. We found that previous use had a strong positive relationship with subsequent site use, indicating that wapiti were not avoiding previously visited locations, as would be expected if memory of forage depletion (which we did not measure) determined response to familiar locations. Revisited sites were of higher quality, i.e. had more moderate terrain, higher forage, and lower predation risk, than sites that were not revisited, indicating that the selection of familiar locations was likely not the result of avoidance of unfamiliar locations. Finally, animals demonstrated preference for familiar locations that it had visited most recently, indicating that memory (which would decline with time) of higher site quality, rather than high quality alone, influenced selection for familiar locations. We conclude that spatial familiarity is important not only for large scale processes such as selection of home range and territory, but for smaller scale habitat selection and foraging as well.

No abstract

Based on evidence of negative relationships between forest fragment area and mouse density, earlier studies suggested that forest edge might be high quality habitat for white‐footed mice (Peromyscus leucopus). However, we have found lower densities of this species and higher prevalence of parasitism near woodland edges. To examine the overall effect of edge habitat, we measured food availability (tree seeds and litter invertebrates), reproductive effort, and risk of predation along edge‐to‐interior gradients in eastern deciduous forests. We also compared woods adjacent to restored prairies to woods adjacent to agricultural fields to determine if the type of habitat adjacent to the edge affected these factors. Results showed slightly higher production of tree seeds in forest interior but no edge‐to‐interior pattern in abundance of invertebrates. Reproductive effort did not differ between edge and interior either. Risk of predation, however, showed a significant increase with proximity to the woodland edge. Type of adjacent habitat did not consistently affect factors determining edge‐to‐interior patterns. Mass of tree seeds differed little between woodland edges adjacent to different habitats in spring but was greater adjacent to agriculture in fall. Invertebrate abundance did not differ between woodland edges with different adjacent habitats. Risk of predation was higher at edges adjacent to restored prairies than at edges adjacent to agricultural fields. We conclude that forest edges are generally lower quality habitat than forest interior for white‐footed mice because of increased predation, and that relatively natural edges are likely to have higher risk of predation than anthropogenic edges.

Geographical variation in morphometric characters in heteromyid rodents has often correlated with climate gradients. Here, we used the long-term database of rodents trapped in the Sevilleta National Wildlife Refuge in New Mexico, USA to test whether significant annual changes in external morphometric characters are observed in a region with large variations in temperature and precipitation. We looked at the relationships between multiple temperature and precipitation variables and a number of morphological traits (body mass, body, tail, hind leg, and ear length) for two heteromyid rodents, Dipodomys merriami and Perognathus flavescens. Because these rodents can live multiple years in the wild, the climate variables for the year of the capture and the previous 2 years were included in the analyses. Using multiple linear regressions, we found that all of our morphometric traits, with the exception of tail length in D. merriami, had a significant relationship with one or more of the climate variables used. Our results demonstrate that effects of climate change on morphological traits occur over short periods, even in noninsular mammal populations. It is unclear, though, whether these changes are the result of morphological plasticity or natural selection.

Several studies have reported higher densities of white‐footed mice in small fragments than in large fragments of eastern deciduous forests. The edge hypothesis states that higher densities in smaller fragments reflect an increase in relative amount of edge habitat, which supports higher densities of mice because of its higher quality. To test this hypothesis we live trapped white‐footed mice along edge‐to‐interior gradients in forest fragments of east‐central Illinois. Our results indicated a greater abundance of mice in the forest interior than near the edge, which did not support the edge hypothesis. This pattern could occur because dominant adults hold larger territories of higher quality habitat, thereby reducing density and increasing fitness near the edge (an ideal despotic distribution). We found some evidence of increased reproductive success (juveniles per female) at the edge, but this could reflect density‐dependent demographic processes rather than habitat quality. Furthermore, other indicators of dominance (body weight, and reproductive activity) did not show an increase at the edge, and other studies indicate higher prevalence of natural enemies at edges, which could account for lower densities there. Reduced competition from larger rodents and reduced predation could cause higher densities in small fragments but the distributions of competitors and predators do not strongly support these hypotheses. We suggest two additional hypotheses that could account for greater densities in smaller fragments: 1) estimates of high densities could be artifacts of the large effect that a few captures can have on density estimates for very small fragments, and 2) densities in smaller fragments are overestimated because mice use a relatively larger area of surrounding habitat as fragment size decreases.