Eric E. Jorgensen

Rodent Use of Microhabitat Patches in Desert Arroyos

-We studied microhabitat use and movements of rodents in a desert arroyo. In arroyos, habitat patches can be identified in which microhabitat is distributed linearly. We recognized five types of patch: (1) wash zone, (2) N shoulder, (3) S shoulder, (4) N terrace and (5) S terrace. Rodents were rarely captured in wash zones. They were captured on shoulders half as much as terraces. Dipodomys merriami preferentially moved parallel to arroyos. We confirmed past observations that cricetines tend to move farther than heteromyids. Exposed/open patches were avoided by all taxa in this rodent community; this may be a general pattern when patches are distributed continuously and consistently in space.

Emission of Volatile Compounds by Seeds Under Different Environmental Conditions

Abstract Small mammals locate buried wet seeds more efficiently than buried dry seeds. This may be attributable to emission of volatile compounds by the seeds. To test this hypothesis I measured emission of volatile compounds from seeds of three plant species (Pinus contorta, Purshia tridentata and Achnatherum hymenoides) under wet and dry conditions using solid phase micro-extraction, gas chromatography/mass spectrometry headspace analysis. Seeds responded in two ways: (1) wet seeds released different, generally greater, amounts of volatile gas than dry seeds and (2) wet seeds and dry seeds released different compounds. Pinus contorta seeds release greater amounts of three compounds when wet; Purshia tridentata seeds release two compounds when dry that are not released when wet, and release increased amounts of two compounds when wet, and increased amounts of two compounds when dry; Achnatherum hymenoides releases at least 22 compounds, one of which is released in large concentrations when wet. These data suggested two mechanisms by which small mammals locate buried seeds. First, small mammals may be sensitive to release of differing concentrations of volatile compounds by seeds. Second, small mammals may be sensitive to compounds released by wet seeds that are not released by dry seeds. Ability of seeds to survive depredation by granivores may be an adaptive trait influenced by natural selection.

Modeling habitat suitability for small mammals in Chihuahuan desert foothills of New Mexico

Desert arroyos make up only 2-4% of the desert landscape but may be unique habitat for some wildlife species. Habitat suitability models for small mammals could indicate habitat condition in these uncommon areas and, through monitoring, provide a tool for arid lands wildlife management. During 1993 and 1994, we modeled habitat associations for 18 species of small mammals in 6 habitats associated with desert foothills in the Sacramento Mountains, New Mexico. Habitat associations of small mammals remained relatively consistent through time, although community abundance decreased 34% from spring 1993 to fall 1994. These habitat associations allowed us to use discriminant analysis to model habitat associations based upon designation of habitat as capable of supporting high, moderate, or low relative abundance. The procedure successfully classified 80-93% of habitats as supporting high, moderate, or low abundance of 13 species of heteromyid and murid rodents. Murid rodents (Peromyscus, Reithrodontomys, Neotoma, Sigmodon) were found primarily within arroyos where shrub diversity, canopy cover, and height were greatest compared to surrounding habitats. The modeling approach described herein should be useful to managers because stochastic temporal change in the relative abundance of small mammals is effectively removed from consideration.

CATASTROPHIC DECLINE OF A HIGH-DENSITY POPULATION OF HISPID COTTON RATS (SIGMODON HISPIDUS) IN CENTRAL OKLAHOMA

Abstract Along the northern periphery of their range, populations of the hispid cotton rat, Sigmodon hispidus, are vulnerable to major reductions in density and occasional local extinctions as a result of severe winter weather. Between our sampling periods on 3 December 2000 and 14 January 2001, 3 independent winter weather events, in conjunction with the coldest month in the state since 1983, affected central and eastern Oklahoma. We recorded a drastic decline in the population of hispid cotton rats at the Center for Subsurface and Ecological Assessment Research in central Oklahoma following these winter weather events. Densities dropped from 58.6 cotton rats/ha on 3 December 2000 to 1.2 cotton rats/ha on 14 January 2001. Although hispid cotton rat densities were declining before these winter weather events occurred, we attributed the dramatic decrease to severe winter weather and below-normal temperatures. As of 19 November 2001, the population of hispid cotton rats at the site had not recovered. Abundances between January and November 2001 ranged from 0.6 to 2.6 cotton rats/ha compared with a range of 30.1 to 112.5 during the same period in 2000. Additionally, we present evidence of populations of hispid cotton rats being affected at a statewide scale as a result of the weather in December 2000. We suspect that severe winters, such as the events described, might slow the northward advance of hispid cotton rats and serve to indirectly regulate populations along the intermediate and northern fringes of its range.

Population Dynamics of Harvest Mice (Reithrodontomys fulvescens and R. montanus) Across a Nitrogen-amended Old Field

Abstract We conducted a mark-recapture experiment to examine population dynamics of the fulvous harvest mouse (Reithrodontomys fulvescens) and plains harvest mouse (R. montanus) in response to low-level nitrogen amendments (16.4 kg N/ha/y) in an old-field grassland. The experimental design consisted of 16, 0.16-ha plots with four replicates of each treatment combination (fenced, nitrogen amendment; unfenced, nitrogen amendment; fenced, control; unfenced, control). We predicted that densities, survival, and transition probabilities would be greater for both species on nitrogen-amended plots because of greater aboveground biomass (i.e., enhanced concealment from predators). We observed no distinct patterns in survival or transition probabilities of R. fulvescens or R. montanus with regard to treatments. Although population densities of R. fulvescens did not exhibit any distinct patterns with regard to treatments, densities of R. montanus tended to be highest on nitrogen plots, but lowest on nitrogen-fenced plots during winter 1999–2000. As low-level nitrogen amendments continue to be applied, we predict survival and densities of R. montanus and R. fulvescens on control plots, especially fenced plots with no nitrogen amendment, will eventually exceed those on nitrogen-amended plots as a result of higher plant species diversity, food availability and better quality cover.

A comparison of modelling techniques for small mammal diversity

Abstract Development pressures frequently dictate that managers’ need to make decisions about which local sites will be developed and which will be protected. When management for diversity is the goal, it would be helpful if models could aid these decisions. We compared three methods for modelling site-specific small mammal diversity at 48 0.58-ha study sites distributed within six habitats in foothills of the Sacramento Mountains, south-central New Mexico, spring and fall, 1993–1994. Methods included; 1) direct richness prediction with discriminant analysis (classification success rate of 15.1%, mean error=1.6 species), 2) prediction of richness based upon expected species-specific habitat suitability with discriminant analysis (classification success rate 20.3%, mean error=1.6 species), and 3) prediction of relative richness (high vs. low) (classification success rate=91.1%). The mean error of methods 1 and 2 (1.6 species) exceeds the difference known to distinguish high richness habitats from low (1.3 species) in this ecosystem. Therefore, we conclude that the appropriate conceptual technique for modelling diversity is to proceed by distinguishing high and low diversity habitats. We found this technique preferable when compared to pursuit of error-prone models for actual richness that have mean errors larger than those known to characterize the system.