The Population Dynamics of Two Rodents in Two Coastal Marshes in Virginia

Abstract

The communities of small mammals were evaluated for 13 months with capture-mark-recapture methods in two Spartina-Juncus marshes of the Atlantic coast in Northampton County, Virginia. Small mammals were trapped for three days each month using live traps placed on floats on two study grids. Two rodents were numerically dominant (~90% of small mammals) there: marsh rice rat, Oryzomys palustris, and meadow vole, Microtus pennsylvanicus. Monthly estimates of population density were greater for rice rats (peak: 45/ha) than for those of meadow voles (peak: 30/ha). Survival rates were generally low, especially for rice rats, indicating highly vagile populations. Both species had greatest breeding activity in spring and autumn, with lower rates in summer and winter. Sex ratios favored males in rice rats but were unity in meadow voles. Although marsh rice rats, being semi-aquatic and capable swimmers, are more highly adapted to living in flooded marsh environments, meadow voles can thrive there too. INTRODUCTION Two species of rodent, marsh rice rat, Oryzomys palustris, and meadow vole, Microtus pennsylvanicus, are dominant in the marshes of the coast and the nearby barrier islands in eastern Virginia (Bloch and Rose 2005, Cranford and Maly 1990, Dueser et al. 1979). These species have been frequently studied elsewhere, but rarely together because the meadow vole is a boreal species near its southern limit in eastern Virginia and Oryzomys, a tropical genus, is widespread only from coastal Delaware southward. Early studies reporting the presence of the marsh rice rat in Virginia tidal marshes include Goldman (1918) on Wreck and Smith islands and Bailey (1946) on Wallops Island. Later, Paradiso and Handley (1965) found rice rats and meadow voles as the dominant marsh rodents in their survey of the small mammal fauna of the northernmost barrier island, Assateague Island. But the most complete survey of tidal marsh and island mammals was conducted in the mid-1970s by Dueser et al. (1979), who trapped on 11 islands; nine had marsh rice rats but only three islands also had meadow voles. Corresponding author: brose@odu.edu 1 Current address: 520 East Main Street, Suite 608, Richmond Virginia 23219 2 Virginia Journal of Science, Vol. 64, No. 1, 2013 http://digitalcommons.odu.edu/vjs/vol64/iss1 18 VIRGINIA JOURNAL OF SCIENCE All studies confirm the numerical dominance of marsh rice rats and, when present, meadow voles in the grassand sedge-dominated tidal marshes. Two studies used regular trapping on study grids to obtain density estimates of small mammals on barrier islands: Adkins (1980) in an Assateague Island marsh and Cranford and Maly (1990), on Wallops Island. Adkins (1980) found modest densities (10-15/ha) of both species in late autumn of two years with few or none of either species during the summer months. Cranford and Maly (1990) reported densities of 25 and 30/ha in two Novembers for rice rats but higher peaks in late winter (45 and 50/ha in March) for meadow voles. Later, Bloch and Rose (2005), also using capture-markrecapture (CMR) methods on two study grids, report density estimates for both species in mainland tidal marshes in Northampton County, Virginia, with populations of both species fluctuating around a mean of 10/ha on one grid but on the second grid meadow voles had August-September peaks of 65 and 75/ha and rice rats had comparable densities but with peaks 3-4 months later. This paper describes the dynamics of populations of both species on the sites used by Bloch and Rose (2005) immediately after our study ended. Meadow voles and marsh rice rats differ in two important ways: their diets and their activity periods. The meadow vole is considered a strict herbivore and the marsh rice rat an omnivore. Wolfe (1982) and others believe the marsh rice rat to be the secondmost carnivorous North American rodent, behind the grasshopper mice (genus Onychomys) of the western US. However, a recent study of the diet of marsh rice rats in mainland marshes of Northampton County, Virginia indicates that local rice rat populations rely heavily on plant material throughout the year (Rose and McGurk, 2006), suggesting that Virginia populations may be more herbivorous than those living farther south (Negus et al., 1961; Wolfe, 1982). Regarding activity patterns, the marsh rice rat is strictly nocturnal, in part confirmed by being common in the diets of owls (e.g., Blem and Pagels 1973, Harris 1953, Jemison and Chabreck 1962), whereas meadow voles are intermittently active both day and night (e.g., Webster and Brooks 1981). Both species are similar in size, with fully adult meadow voles weighing 40-60 g and marsh rice rat adults slightly larger, up to 80 g. The objectives of our study were to evaluate the population dynamics of the two species, including such features as density, survival rates, and reproduction, and to compare these patterns to those of other geographic populations. Our study lasted for a calendar year, from May 1994 through May 1995. MATERIALS AND METHODS The Study Sites The two study sites were seaside marshes on Nature Conservancy property in Northampton County, Virginia. At Grid 1, located 4.4 km east of US Highway 13 at the southern edge of the village of Oyster, the vegetation was representative of the salt grass community. The dominant plants were Spartina alterniflora, S. patens, Iva frutescens, Juncus roemarianus, and on slightly higher ground, Phragmites australis (communis). Low-lying areas were subject to more frequent flooding, but most of the grid was on higher ground and remained relatively dry except during the high tides associated with the full moon. A thick ground cover of S. patens blanketed most of Grid 1. Grid 2 was located in a marsh locally known as Steelman’s Landing, east of Townsend. A larger grid could be placed in this marsh, but the three rows closest to the Virginia Journal of Science, Vol. 64, No. 1, 2013 http://digitalcommons.odu.edu/vjs/vol64/iss1 POPULATION DYNAMICS OF TWO RODENTS 19 mud flats had sparse vegetation and deeper water during the periods of daily flooding. Grid 2 flooded less often than did Grid 1 because it was farther away from the shoreline. But the surface was flatter and so it flooded more uniformly, and was usually wetter than Grid 1 when flooding did occur. Trapping Procedures After preliminary trapping in March 1994 confirmed the presence of both species, grids were established at both sites, and monthly trapping began in May 1994. Grid 1 was irregular in shape to conform to the area of herbaceous vegetation; its 75 trapping stations were placed at 10-m intervals (maximum of 7 rows, 14 columns) for an effective trapping area of 0.75 ha. Grid 2 had 130 traps on a 13 X 10 grid, for a trapping area of 1.3 ha. One Fitch live trap (Rose, 1994) was placed at each grid coordinate (= station). However, because of the daily flooding or danger of flooding, traps were strapped onto floats made of 31 cm by 21 cm rectangles of 1.6 cm thick insulation Styrofoam, using large rubber bands. Wire ties and monofilament line secured each float to the wooden stake marking each coordinate on the grid. Thus, during periods of flooding the float raised the trap, enabling a rodent to swim to the trap or preventing trapped animals from drowning. Each grid was trapped for three days per month, for a total of 1872 trap-nights on Grid 1 and 3584 trap-nights on Grid 2 (1 trap set for 1 night = 1 trap-night). Traps baited with mixed bird seed and sunflower seeds were set in the late afternoon and then checked early in the next three mornings. In summer, traps were locked open in the morning and set again in late afternoon to prevent death from confinement to a trap during the heat of the day. In winter, polyfill was added to each trap for insulation. At first capture, each rodent was given a uniquely numbered ear tag, examined for its reproductive condition, weighed with a Pesola scale, and released at the point of ® capture. Reproductive information included position of testes (abdominal or descended/scrotal) for males and for females we evaluated condition of the vagina (perforate or not), size of nipples (small, medium, large) and the condition of the pubic symphysis (closed, slightly open, open). Heavily pregnant females were recorded as such. We defined the age classes of meadow voles using the criteria of Krebs et al. (1969), with juveniles (<22g), sub-adults (22-29g), and adults ($ 30 g). For rice rats, we used the criteria of Wolfe (1985), with juveniles (<30g), sub-adults (31-50g), and adults (> 50 g). The trapping methods followed the animal handling guidelines of the American Society of Mammalogists (most recent: Sikes, Gannon et al. 2011). This study was conducted before Old Dominion University had established an IACUC protocol for the field study of wild mammals. Statistical analysis We used the Minimum Number Alive (MNA) method to estimate density of the populations (Krebs, 1966) and the statistical package JOLLYAGE to calculate timespecific survival rates of both adults and young and recruitment (Pollock et al., 1990). These parameters were analyzed using a Model-II 3-factor analysis of variance (ANOVA), with the factors being grid, season, and species. Where necessary, a RyanEinot-Gabriel-Welsch (REGWF) test was used to investigate the role of significant factors. Seasons were defined as summer (June-August), autumn (SeptemberNovember), winter (December-February), and spring (March-May). Chi-Square tests Virginia Journal of Science, Vol. 64, No. 1, 2013 http://digitalcommons.odu.edu/vjs/vol64/iss1 20 VIRGINIA JOURNAL OF SCIENCE were used to determine whether sex ratios were unity. The p <0.05 level of significance was used for all statistical tests. RESULTS The Small Mammal Community During the 13-month study period, 185 small mammals of five species were identified on Grid 1 (Table 1), of which 65 were rice rats and 62 were meadow voles; together these species con