In-stream and riparian habitat data were collected from 58 stations throughout Iowa between 1995 and 2001 for an ongoing study of interior river habitats. Out of 162 variables calculated for each station, data visualization software was used to select 37 variables that encompassed the majority of habitat variability and relationships between variables. This subset of variables was examined using univariate and multivariate methods, including Pearson correlations, linear regressions, ANOVAs, canonical correlations and a non-metric multidimensional scaling (MDS) ordination. Stream order significantly affected several riparian and stream size-related variables. Eleven in-stream, riparian and reach variables had significant regressions with drainage area. Only one variable, proportion of banks with canopy, had a significant regression with sinuosity. The axes of the 3-dimensional ordination used to plot habitat station similarities were multivariate expressions of stream size, rock versus wood cover and canopy cover. The ordination revealed that larger stations had decreasing levels of overall habitat heterogeneity. Canonical correlation analysis used to examine linkages between in-stream and riparian habitat variables revealed three significant canonical variables; one depicted a stage of the channel evolution model (CEM), the second linked the number of vegetation types to the number of cover types and flow classifications, the third canonical correlation reflected a process of riparian trees eventually becoming in-stream cover. Pervasive, intense agriculture with its associated channelization and sedimentation create a backdrop of profound alteration in shaping stream habitat in Iowa. The relationships we demonstrated among in-stream and riparian habitat characteristics in Iowa streams must be viewed in light of this backdrop of alteration.