1. We investigated the effects of a wildfire on stream physical, chemical and biological characteristics in a Mediterranean climate, comparing stream community structure and consumer resource use in burned versus unburned catchments in Santa Barbara County, CA, U.S.A. 2. Canopy cover was lower and water temperature was higher in streams draining basins where the riparian vegetation burned than in streams in unburned basins or burned basins where riparian vegetation remained intact. Stream flow and suspended sediment concentrations during large post-fire storms and wet season nutrient levels were higher in burned than unburned catchments, with increased sedimentation after flood peaks. 3. A year after fires, algal levels were highest in streams where riparian vegetation burned and lowest in streams in burned basins where the riparian canopy remained intact. In contrast, streams in burned basins had lower particulate organic matter, detritivore and predator levels than unburned basins, regardless of whether riparian vegetation burned. Where present, southern California steelhead trout (Oncorhynchus mykiss) were extirpated from burned basins. 4. Algivore densities were high in streams with burned riparian vegetation for two post-fire years before declining to unburned stream levels. Shredder densities rebounded in streams in burned basins with intact riparian vegetation, but remained low for 4 years where riparian vegetation burned. Predatory invertebrate densities increased at sites where trout were eliminated by wildfire. 5. Hydrogen stable isotope analysis indicated that the diets of most invertebrate taxa in streams with burned riparian vegetation a year after fires were comprised of a higher proportion of algal material than riparian detritus relative to invertebrates in streams with intact riparian vegetation. 6. Wildfire impacts on stream food webs are determined, in part, by fire severity in the riparian zone. Streams with burned riparian canopies supported algal-based food webs and streams with intact riparian canopies sustained detrital-based food webs. Fire affected basal resources (nutrients, light, allochthonous inputs) with bottom-up effects on primary producers and consumers, but top-down effects were decoupled at the trophic link between invertebrate predators and primary consumers.

1. Effects of copper on species composition and production of benthic insects in an oligotrophic stream dosed at low concentrations {2.5-15 ^g 1"' Cu,; approximately 12-75 ng I"' Cu-+) were determined. Dosing was initially in autumn-early winter when peak densities of many species occur. It was resumed the following summer near the time of egg hatch of most species and continued through the remaining aquatic stages of univoltine and multivoltine taxa.2. Declines in population density of species representing all major orders {Ephemeroptera. Plecoptera, Coleoptera. Trichoptera and Diptera) occurred at 5 and/or 10 /^g 1"' Cu^. Differences in sensitivity among families existed in the Plecoptera, Trichoptera and Diptera. These differences were apparently related to the trophic preferences of taxa, with herbivores and detritivores more sensitive than predators.3. Equivalent percentage declines in annual production occurred for herbivores/detritivores and predators exposed to copper, suggesting an adjustment in total predator production in response to the reduction in prey abundance.4. Possible examples of reiease of benthic insects from competition with less tolerant species were rare. Of the thirty-seven most abundant taxa, only the predators Rhyacophila vaccua and Rhyacophila angelita were more abundant [than in the control] at copper concentrations (5 and 10//gl"' Cuj) causing population declines of other species. 5. Species richness and percentage similarity were sensitive communitylevel indices of copper toxicity, whereas Brillouin's diversity was insensitive to differences among the control and copper-treated stream sections.

Arundo donax (giant reed) is invasive in Mediterranean, sub-, and tropical riparian systems worldwide. The armored scale Rhizaspidiotus donacis is approved for biocontrol in North America, but an adventive population was recently discovered in southern California. We documented this population’s distribution, phylogeny, phenology, potential host spillover to Phragmites spp., and potential for parasitism by a common biocontrol parasitoid of citrus scale. The adventive scale was found within a single watershed and is genetically closest to Iberian scale genotypes. Rhizaspidiotus donacis developed on Phragmites haplotypes but at much lower densities than Arundo. The adventive population is univoltine, producing crawlers from March-June. Aphytis melinus parasitoids exhibited sustained interest in R. donacis during choice and no-choice trials and oviposition resulted in a small second generation. Rhizaspidiotus donacis appears limited in distribution by its univoltinism and sessile adult females. This presents challenges for broad biocontrol implementation but allows for targeted application. The genetic differentiation between imported biocontrol samples and adventive populations presents an opportunity for exploring benefits of hybrids and/or alternative genotypes where establishment has been difficult. While unlikely to occur in situ, spillover to vulnerable endemic Phragmites or deleterious parasitoid effects on scale biocontrol agents warrants consideration when planning use of R. donacis.