The Macroinvertebrate Community Index (MCI) is used by Regional Councils and other organisations in New Zealand for a range of purposes including State of Environment and consent monitoring in freshwaters. The index is designed to reflect human impacts on waterways, particularly organic pollution, and is calculated from tolerance values (TVs) assigned to freshwater invertebrate taxa. For streams with hard bed substrate, published TVs exist for 180 freshwater invertebrate taxa, with many other taxa not having TVs assigned. Regional Councils currently treat taxa that do not have assigned TVs in different ways; some councils exclude these taxa from their MCI calculations while others have developed TVs using professional judgement. Developing standard TVs for all freshwater invertebrate taxa is a key step towards ensuring national consistency in calculation and reporting of the MCI. Extensive testing of the revised TVs and resulting MCI site scores was beyond the scope of this study. We compiled a national-scale dataset of macroinvertebrate community data from over 1300 freshwater monitoring sites. There were insufficient data from streams with soft bed substrate to revise the version of the MCI used in soft-bottomed streams (MCI-sb) therefore these 122 sites were excluded. A total of 10548 samples were collected at the remaining 1266 hard-bottomed sites. Data were divided into two datasets; the full dataset with all sampling occasions per site and a reduced dataset, consisting of 50 random sub-samples of the full dataset, each sub-sample consisting of one sampling occasion per site. Data were further grouped into eight classes based on Climate and Source of Flow categories from the River Environment Classification (REC). To revise the MCI TVs we applied an iterative computational process developed by Chessman (2003), which has been previously applied to develop two other indices in New Zealand (MCI-sb and a wetland MCI). The Chessman method was run on all eight environmental classes on both the full and reduced datasets. Tolerance values were compared across the environmental classes and two different approaches to assigning singular TVs to taxa across the environmental classes were compared. This resulted in revised MCI TVs for 234 taxa. There were 12 taxa with insufficient data to generate revised TVs. All of the revised TVs reported here were assigned using an objective computational approach, whereas of the 180 original published scores, 133 were assigned by professional judgement. Further testing is required to determine whether the revised TVs are more or less sensitive to gradients of human impacts than the original TVs. Our preliminary analyses show that the revised TVs and resulting MCI site scores were correlated with existing TVs and MCI site scores, and also with catchment-scale measures of land use. Revised MCI site scores were generally, but not always, higher than original MCI scores. As such, while revised TVs and MCI site scores are likely to provide a sensitive indicator of human impacts on rivers and streams it may be necessary to also revise water quality categories. For example, the original values place 15% of sites in the ‘excellent’ water quality category (MCI > 120) while the revised values place 50% of sites in this category. We provide revised TVs for 234 taxa but note the following caveats: 1) Original and revised TVs must be used separately; they are not interchangeable or directly comparable. Back calculation of MCI site scores would be required for historical comparisons using the revised TVs. We recommend that revised MCIs are reported as MCI-2-hb. 2) How taxa without revised TVs are included or excluded from analyses should be reported to ensure transparency, especially when comparing between sites or over time. 3) A new water quality categorical scale may be required as MCI site scores based on revised MCI TVs are generally higher than those created using the original TVs 4) Further testing is required to validate the sensitivity (in relation to human impacts on waterways) of the MCI scores developed using the revised TVs. The data were collected between 1990 and 2012 from 1388 sites distributed nationally (map provided). Listing of data in tables, figures and appendices follow: Tables Table 2-1: Sites were categorised by combined REC Climate and Source of Flow classes. Table 2 2: Number and percentage of samples with MCI-hb site scores within the four water quality classes for MCI-hb as identified in Stark and Maxted (2007b. Table 2 3: Spearman rank correlations (Rs) between upstream land use type and MCI-hb site scores generated using the original and revised TVs. Table 3 1: Taranaki Regional Councils MCI categories of biological water quality conditions adapted for Taranaki streams. Figures Figure 2-1: Distribution of invertebrate sampling sites showing all sites and those with hard (HB sites) and soft (SB sites) bed substrate. Figure 2 2: Histograms of MCI-hb tolerance values for taxa a) as originally reported by Stark and Maxted (2007b) and b) the revised MCI-hb. Figure 2 3: Original MCI-hb tolerance values (as reported in Stark and Maxted 2007b) and revised MCI-hb tolerance values for 161 taxa for which both values exist. Figure 2 4: MCI-hb for all site visits (n = 10548) calculated from the original and revised MCI-hb taxa tolerance values. Figure 2 5: Frequency histograms for MCI site scores calculated using the original MCI-hb tolerance values in Stark and Maxted (2007b) and the revised values. Figure 2 6: Scatterplots of MCI-hb site scores calculated from revised and original MCI TVs and the proportion of upstream land in pastoral and native land use. Appendices: Appendix A Revised MCI-hb tolerance values for freshwater invertebrate taxa. Appendix B Taxa with existing MCI-hb tolerance values for which revised scores were not generated Appendix C. Original and revised taxa tolerance values across environmental classes. Appendix D. Distributions of tolerance values across eight environmental class for individual taxa from 50 datasets each generated with one random visit per site. Appendix E. Testing the influence of environmental class on generated TVs.