Biogeography of Hanging Gardens of the Colorado Plateau

Abstract

The lush plant growth of hanging gardens attracted the attention of early explorers, botanists, and pioneers of the Colorado Plateau, a region dominated by bare rock and xeric vegetation. Hanging gardens are small, hydrophytic, herbaceous plant communities (Malanson 1980, Welsh and Toft 1975, Tuhy and MacMahon 1988, Welsh 1989a, Heil et al. 1991) on canyon walls in the entrenched drainage system of the Colorado Plateau. The microclimate of larger hanging gardens contrasts sharply with the surrounding high desert: water is abundant, soils are moist, and canyon walls offer periodic shade. This study was conducted to expand our current knowledge of hanging gardens with baseline data from a comparative ecological approach. Hanging gardens, in seven National Park Service units on the Colorado Plateau, were sampled for vascular plant diversity, distribution, abundance, levels of endemism, and biogeographic origin of endemic taxa. At the consumer level, insects, other arthropods, and small mammals were collected on selected hanging gardens to determine species diversity, distribution, and relative abundance. Since the publication of MacArthur and Wilson's (1967) Theory of Island Biogeography, many attempts have been made to use the positive correlation between island size and species richness to conserve species diversity in fragmented and natural terrestrial habitat islands (see Shafer 1990 for a review). An initial impetus for this study was the island-like appearance of hanging garden habitats. Theoretical work on the species richness­area relationship have prompted many empirical studies testing support for one of three alternate hypotheses: (1) MacArthur and Wilson's (1963, 1967) equilibrium hypothesis, (2) Williams' (1943) (and later Connor and McCoy 1979) habitat diversity hypothesis, and (3) Arrhenius' (1921) passive sampling hypothesis. This study describes the hanging-garden vascular plant species-area relationship and tests support for these alternate hypotheses. The immigration and extinction curves that MacArthur and Wilson (1967) incorporated into their equilibrium hypothesis were recently put into a spatially structured, multispecies metapopulation model (Gotelli 1991) as an alternative to an original metapopulation model and Hanski's (1982) core­satellite species metapopulation model. Gotelli (1991) then introduced a fourth metapopulation model which incorporated both propagule rain and the rescue effect. The last model, as well as the original metapopulation model, and the equilibrium hypothesis (MacArthur and Wilson 1967) based model, predict a unimodal distribution on species occurrence histograms (Gotelli 1991). The core­satellite model predicts a bimodal distribution on species-occurrence histograms. This study was primarily designed to collect distribution and abundance data for hanging-garden plant and bee species as a test of the core-satellite prediction.