Population Trends and Habitat Associations of Rodents in Southern Texas
Windberg, Lamar A., The American Midland Naturalist
ABSTRACT.-Population trends for seven genera of rodents were monitored by snap-trap capture rates on a study area (700 km2) in Webb County, Texas, during winter from 1976 to 1986. Population irruptions occurred in 1982 (46-fold) and 1986 (13-fold) for Sigmodon, and in 1982 (17-fold) for Reithrodontomys. Annual variability in abundance was positively correlated between the following pairs of genera: Sigmodon-Reithrodontomys, Peromyscus-Onychomys, Onychomys-Reithrodontomys, and Dipodomys-Chaetodips. Annual abundance of Sigmodon in winter was positively correlated with rainfall in the prior growing season. Analysis of the distribution of the composite rodent community among five vegetative habitats revealed greater abundance of rodents in the most dense and diverse shrub community (high-density shrub/mixed-grass habitat).
Population fluctuations of desert rodents are associated with variations in precipitation which influence plant growth and seed production (Whitford, 1976; Brown and Zeng, 1989; Brown and Heske, 1990). Coexisting species with different life histories may exhibit contrasting patterns of fluctuation in the same environment (Brown and Heske, 1990).
The population dynamics (Whitford, 1976; Brown and Zeng, 1989) and community structure (Brown, 1989; Brown and Heske, 1990) of rodents in the Chihuahuan Desert of North America have been investigated extensively. However, rodent populations in the semiarid Tamaulipan Biotic Province of southern Texas and northeastern Mexico (Blair, 1952;Jahrsdoerfer and Leslie, 1988) have received limited attention. Brown and Heske (1990) stressed the importance of long-term studies in providing insights about the community dynamics of small mammals. Here I present data for trends in population abundance of seven genera of rodents in the Tamaulipan brushland of southern Texas for 11 y (1976-1986). examine the effects of environmental factors and interspecific interactions on rodent population dynamics by analyses for relationships between variations in annual rodent abundance and rainfall and for covariation among taxa.
Rodent habitat use, and their competitive interactions for resources, has also been wellstudied in the Chihuahuan Desert (Hallett, 1982; Brown, 1989; Rogovin et al., 1991). In this study, I examine habitat use by the rodent community in southern Texas by analysis of rodent abundance in five vegetative habitats during a 4-gamma period (1981-1984), and assess intraspecific competition for habitats by analyses of rodent body mass.
The study area of approximately 700 km^sup 2^ was located 5-40 km NE of Laredo, Webb County, Texas. The climate of the region is semiarid; annual rainfall at Laredo averages 51 cm (Windberg et al., 1985) but is erratic among and within years (Norwine, 1978; Jahrsdoerfer and Leslie, 1988). Topography is level to rolling; drainages flow into the Rio Grande River (Windberg et al., 1985). Soils are generally shallow and vary from fine sandy loam to clay.
Vegetation on the study area was typical of the Rio Grande Plains of southern Texas (Gould, 1975; Archer et al., 1988). The original vegetative community was a grasslandsavannah climax, but present communities are dominated by dense stands of shrubs (Archer et al., 1988). Extensive brush control by landowners results in varying stages of secondary plant succession (Jahrsdoerfer and Leslie, 1988). The study area was privately owned rangeland used for grazing by cattle.
In conjunction with other studies (Windberg and Mitchell, 1990; Windberg, 1995), four 24-km routes along unimproved roads were used for sampling rodent abundance in winter. The relative abundances of seven genera were estimated by capture rates in snap-trap surveys conducted from mid-January to early March, 1976-1986. Trap-transects (n = 100) were spaced 1 km apart along each of the four routes. Ten M-4 Victor rat traps with expanded treadles (5 by 5 cm), baited with a mixture of peanut butter and rolled oats, were spaced 10 m apart on each transect. …