ARTICLE Predicting Fish Assemblages from Substrate Variation in a Turbid River: Grab Samples Compared with Pole Probing

We tested two substrate quantification methods at 28 500-m-long sites in the Wabash River in 2008. We predicted boat electrofishing collections of fish using substrate characterization based on grab samples and pole probing. Our approach was to compare discrete scores with a nonparametric approach and use canonical correspondence analyses to analyze alternative substrate quantification methods to predict fish assemblage structure. There was no difference in the ranks of grab and pole samples. Both substrate quantification approaches provided reasonable predictions of fish assemblage variation. Mantel tests were used as an additional method of comparing the two substrate methods for concordance with fish assemblages. We suggest that the pole method is most cost effective because it does not require additional time for laboratory processing. The benthic habitat of stream ecosystems is a primary re-source for algae, invertebrate, and fish assemblages that utilize such habitats for particular life history stages. Benthic habitat provides cover and habitat for fish and aquatic insects (Gurtz and Wallace 1984; Matthews 1998). Quantifying substrates as ben-thic habitat is frequently required to understand fish assemblag