Butterflies of Glacier National Park, Montana

Materials and Methods

Census Sites

During 1987, 58 sites were sampled with the objective of censusing extensively throughout the park. The sites were selected more on the basis of their position in the topographic and elevational gradients in Glacier National Park than by their habitat type. Temperature and moisture gradients, quite independent of habitat types defined by vegetation, are often of primary importance in determining the distribution and local abundance of many terrestrial plant and animal taxa (Whittaker, 1952; Terborgh, 1970; Brussard, 1985). Several methods of site categorization were considered (e.g., Elton and Miller, 1954; Southwood et al., 1979; Bunce and Shaw, 1973). The technique used consisted of (1) marking the location of each site on a 7.5-min United States Geological Survey (USGS) topographic map, (2) recording the elevation, slope, and exposure of the site, and (3) briefly categorizing the habitat type of the site (e.g., xeric meadow, riparian, lodgepole pine forest). On the basis of these data, each site was plotted on a two-dimensional graph representing the available "ecological space" in Glacier National Park where ecological space is defined graphically with elevation as the ordinate and moisture conditions (ranging from hydric to xeric on the flats and based on aspect, slope, and exposure on mountainsides) as the abscissa. Additional sites then were chosen to sample the full range of combined topographic and elevational conditions in the Park.

The 1987 sites varied in both size and shape. Those that were inaccessible by road, could be visited only once. Temporal replication was necessary because of phenological differences among butterflies; for example, some species emerge early in the summer. whereas others emerge in late summer. To increase the number of temporal replicates, I sampled only 24 sites in 1988 and 1989; this was the maximum number that could be visited at least twice during the summer. Thus, each site was surveyed more intensively, but the Park, as a whole, was sampled less extensively. Nonetheless, the 24 sites selected are representative of the range of geographic and environmental variation within the Park (Fig. 1). Each of the 1988-89 sampling sites was 1 square kilometer as defined by Universal Transverse Mercator coordinates on USGS topographic maps. Field experience in 1987 indicated that sampling sites must be at least this size to represent the range of small-scale patchiness and microhabitats adequately.

Glacier National Park biodiversity sites

Fig. 1. Glacier National Park biodiversity sites for butterflies in 1988 and 1989. Open circles represent sites surveyed during only one year, whereas closed circles represent sites surveyed during both 1988 and 1989. See Debinski (1991) for detailed descriptions and names of sites.

Although standard community-sampling procedures call for use of sampling sites that are homogeneous in structure and composition (Gauch, 1982), I chose to maximize sampling of habitat diversity in the 1988-89. Nonetheless, the 24 sites selected are representative of the range of geographic and environmental variation within the Park (Fig. 1). Each of the 1988-89 sampling sites was 1 square kilometer as defined by Universal Transverse Mercator coordinates on USGS topographic maps. Field experience in 1987 indicated that sampling sites must be at least this size to represent the range of small-scale patchiness and microhabitats adequately. Although standard community-sampling procedures call for use of sampling sites that are homogeneous in structure and composition (Gauch, 1982), I chose to maximize sampling of habitat diversity in the 1988-89 sites for two reasons. First, this study was designed to inventory species occurrences across a very large area; thus, site homogeneity was ignored to maximize broadscale coverage. Second, because I observed in 1987 that species diversity was much higher along ecotones, I included ecotones in many of the sites. Nevertheless, broad habitat-type characterizations were still possible at each site.

This particular design resulted in a certain bias. Vegetational types or ecological space defined by position on the topographic-elevational gradient were not sampled in direct proportion to their frequency of occurrence in Glacier National Park. Instead, some of the rarer habitats were overrepresented as I attempted to sample rare species and to increase coverage of ecological space along temperature and moisture gradients. The primary disadvantage of this approach is in statistical analysis of the data; the perceived rarity of a species may not be indicative of its true rarity in Glacier National Park.

Sampling

Butterflies were censused in three separate 50 x 50-m subplots in each site by netting for 20 min and releasing. Subplots were chosen to represent as much of the variation included within the larger plot as possible. The sampling periods were established empirically by plotting the number of species recorded against time. The average time at which the species-effort curve flattened out (i.e., no more species added) was considered to be the optimal sampling time in each subplot. Presence-absence data were recorded rather than abundance values, because collection of abundance data is so time-consuming that the number of samples would have been reduced dramatically. Also, species richness (S) is the simplest, most practical, and most objective measure of species diversity (Peet, 1974).

Species Richness Through Time

Earlier species lists were compared with the results of this study to determine whether species diversity had changed over time. The primary species list was that of Kohler (1980), which is based on data compiled from university collections at Montana State University and the University of Montana, private collections, the Glacier National Park collection, records from natural history museums, and a review of the published literature. John S. Garth, a Glacier National Park naturalist, also compiled species lists of butterflies during the summers of 1935, 1949, and 1950. The nomenclature used in the Appendix is standardized using Scott (1986). Because the data on species diversity were recorded as presence or absence, the significance of changes in species occurrences over time was tested by the use of simple contingency tables and G-tests (Sokal and Rohlf, 1981).

Previous Section -- Introduction
Table of Contents
Next Section -- Results