Macroinvertebrates were collected from 19 sites on 14 streams on the island of Oahu and from 9 sites on 7 streams on the island of Kauai to evaluate associations between macroinvertebrate assemblages and environmental variables and to determine whether or not it would be feasible, in future studies, to develop macroinvertebrate metrics that would indicate stream quality based on the macroinvertebrate assemblages and/or components of the assemblages. The purpose of applying rapid bioassessment techniques is to identify stream quality problems and to document changes in stream quality. Samples were collected at 10 sites in 1999, 3 sites in 2000, and 5 sites in 2003 on Oahu and at 9 sites on Kauai in 2003. Additionally, multiple year and multiple reach samples were collected at 1 site on Oahu. Macroinvertebrates were collected primarily from boulder/cobble riffles or from the fastest flowing habitat when riffles were absent. Although most streams in Hawaii originate in mountainous, forested areas, the lower reaches often drain urban, agricultural, or mixed land-use areas. The macroinvertebrate community data were used to identify metrics that could best differentiate between sites according to levels of environmental impairment. Environmental assessments were conducted using land-use/land-cover data, bed-sediment and fish-tissue contaminant data, and reach-level environmental data using a calibration set of 15 sites. The final scores of the environmental assessments were used to classify the sites into three categories of impairment: mild, moderate or severe. A number of invertebrate metrics were then tested and calibrated to the environmental assessments scores. The individual metrics that were the best at discerning environmental assessments among the sites were combined into a multimetric benthic index of biotic integrity (BIBI). These metrics were: total invertebrate abundance, taxa richness, insect relative abundance, amphipod abundance, crayfish presence or absence, and native mountain shrimp presence or absence. Because this index is in the preliminary stage of development and additional 'pristine' sites need to be sampled and assessed to develop a more robust measure of biotic integrity, the index will be referred to as a Preliminary Hawaiian Benthic Index of Biotic Integrity (P-HBIBI). The P-HBIBI scores were then classified into three categories of impairment: mild, moderate, or severe. The P-HBIBI was then used to assess the remaining sites and classify them into impairment categories. The P-HBIBI was correlated (r2 = 0.72; p < 0.005) with a reduced environmental assessment determined without contaminants data. The results of this study suggest that the development of a reliable Hawaiian benthic index of biotic integrity (HBIBI), based on macroinvertebrate assemblages, is feasible; however, a much larger sample size, including more samples from 'pristine' sites and from the other islands, would be required.