Top-down impacts on creosotebush herbivores in a spatially and temporally complex environment

I studied top-down effects on creosotebush (Larrea tridentata) insect herbivores by protecting them from bird and arthropod predation. The purpose of my study was to investigate experimentally the differential and/or interactive effects of two predator guilds on herbivore densities. I performed my experiments on 24 creosotebushes at each of three sites within the Jornada Long-Term Ecological Research Site in the Chihuahuan Desert of southwestern New Mexico, USA. At each site the experiment consisted of: six creosotebushes from which birds were excluded with nylon mesh cages; six creosotebushes from which arthropod predators were removed by hand or aspirator; six creosotebushes from which both birds and arthropod predators were removed; and six control creosotebushes from which neither birds nor arthropod predators were removed. I conducted nondestructive nocturnal visual censuses of herbivores on each creosotebush at the beginning of the experiments in mid-May, 6 wk after the start of the experiment in late June, and 12 wk after the start of the experiment in early August. I conducted these experiments and censuses in 1993 and again in 1994. In both years the herbivore densities became significantly higher in experimental than in control creosotebushes. The effects of bird and arthropod predation on herbivore densities were additive in 1993, but they were compensatory in 1994. In 1994 arthropod predator densities became lower in creosotebushes from which birds had been removed than in creosotebushes from which birds had not been removed, but this result did not obtain in 1993. These results may be due to a combination of factors including: avian and arthropod predation on herbivores, ''intraguild'' predation of birds on arthropod predators, and competition within the herbivore community. The relative numerical impacts of the predator-removal experiments varied among seasons and among sites within either year, but temporal and spatial variation in predator impacts did not correlate strongly with known gradients of climatic or bottom-up heterogeneity in this system. The results of this study confirm the important direct and cumulative effects of multiple predator guilds, even against a complex background of temporal and spatial heterogeneity.