-Relationships of area with numbers of species and individuals of spring migrants were examined for 69 shelterbelts (forest islands) in eastern South Dakota. Total abundance and number of species were as highly correlated with area during spring migration as during the breeding season. The relationships of area with total abundance and number of species were highly similar between two years of study. Area was more important in determining abundance and number of species than either diversity of plant species or isolation of the islands. The dispersion of migrants among islands, indicated by the relationships of area with total abundance and species numbers, could have been a result of passive dispersal, selection for larger area, or behavioral interactions. Passive dispersal was unlikely because the isolation of an island had no significant influence on abundance or diversity. Also, the diversity and abundance of migrants were modified by habitat conditions (i.e., vegetation diversity), indicating that migrants select the forest islands they inhabit. Migrants may select large areas but they should then have increased with area at an increasing rate, but this did not occur. Dispersion among islands may be the best way for migrants to replenish their energy reserves when food is scarce. Two facts suggest that migrants may interact to disperse themselves relative to food. First, application of a model for interacting species provided increasingly better fits to ecological groups that increasingly confined their foraging within shelterbelts. Second, the density of birds was greater in smaller islands than in larger islands owing to species that did not feed solely within the islands. The density of birds that did feed primarily within the islands remained more or less constant with changing area. Shelterbelts consist of rows of trees and shrubs forming long and narrow island configurations. Shelterbelt plantings began largely in the mid-1930's under the auspices of the Shelterbelt Project and the Prairie State Forestry Project directed by the U.S. Forest Service, Department of Agriculture (Read 1958), in order to reduce wind erosion and protect homes and livestock on the Great Plains. These islands of planted trees supplied potential habitat for many species of woodland birds. The vegetation structure of a particular shelterbelt modifies the composition, diversity, and abundance of bird species using that shelterbelt during spring migration and the summer breeding season (Martin and Vohs 1978). A strong relationship between numbers of breeding bird species and island area has been found for many real and habitat island systems (MacArthur and Wilson 1963, 1967; Abbot 1974, Johnson 1975, Diamond and Mayr 1976, Galli et al. 1976). With the exception of a recent study by Willson and Carothers (1979) in Arizona, the influence of area on the number of transient bird species and individuals using habitat islands during migration has not been reported. I examine here the influence of area of shelterbelts in eastern South Dakota on the diversity, abundance, and ecological structure of transient bird communities during the spring migrations of 1976 and 1977. STUDY AREA AND METHODS The 69 shelterbelts examined in this study were scattered throughout South Dakota between its eastern border and the Missouri River. There is approximately one shelterbelt per 115 ha in this area of the state, representing 1.1% of the land area (Walker and Suedkamp 1977). These small forests represented true habitat islands, being isolated from one another by fields of natural and planted grasses. The floral diversity was low and included only 15 tree and 14 shrub species. The most common tree species were Siberian elm (Ulmus pumila), American elm (U. americana), green ash (Fraxinus pennsylvanica), and hackberry (Celtis occidentalis), while the most common shrubs were honeysuckle (Lonicera tatarica), lilac (Syringa vulgaris), American plum (Prunus americana), and chokecherry (Prunus virginiana). I quantified the structural complexity of the vegetation for each shelterbelt. Ground and canopy coverages were measured by noting the presence or absence of vegetation through an ocular tube (James and Shugart 1970). Density of understory vegetation was measured using a density board (Giles 1971). Readings were taken between each planted row at three locations in each belt. Measurement of percent slope, from a fixed distance, estimated vegetation height.
Read full abstract