Nonpoint sources (NPS) of pollution can greatly impact water quality. One key component in effectively implementing a NPS pollution management program is the identification and assessment of human and animal sources of bacterial contamination. Fecal coliform bacteria, including Escherichia coli (E. coli), have extensively been used as an indicator of fecal pollution and the potential presence of other pathogenic microorganisms in water. E. coli is a common inhabitant of animal and human intestines and recent studies have shown that isolates from humans and various host animals (e.g. cattle, chickens, and pigs) may differ genetically and phenotypically. Analysis of E. coli isolates using genetic and biochemical tests may allow the original host animal to be identified, referred to as bacterial source tracking (BST). In the last few years there has been a tremendous amount of progress in the development and evaluation of BST methods, however, the scientific literature is full of conflicting reports. There is also significant controversy over the selection of the single, most useful BST method and the recent research suggests that multiple techniques may be the most successful. Molecular tools appear to hold the greatest promise for BST, providing the most conclusive characterization and level of discrimination for isolates. Antibiotic resistance analysis, a phenotypic characterization method, also has the potential to identify the human or animal origin of isolates. However, reference "libraries" of bacterial genetic fingerprints and antibiotic resistance profiles are needed to correctly identify the source of bacteria isolated from environmental water samples. Our laboratory is currently the lead on two Texas state-funded projects that include the development of E. coli bacterial source tracking libraries. Several thousand isolates are being characterized by automated ribotyping using the Qualicon RiboPrinter, repetitive sequence polymerase chain reaction (ERIC-PCR), pulsed field gel electrophoresis (PFGE) and antibiotic resistance analysis (ARA). Fingerprint patterns are being compared using multiple statistical tools to determine discrimination of sources and congruence of the different methods. The combination of ERIC-PCR, RiboPrinter ribotyping, PFGE, and ARA, will provide one of the most comprehensive BST libraries of environmental E. coli isolates. This multiphasic approach will allow the state of Texas to keep pace with developments in BST technology. By using standardized and automated methods, the libraries can be expanded through future projects and data shared with other investigators and regulatory agencies. Recent analysis of a model data set illustrated the need for cautious interpretation. For example, different similarity clusters were formed using the RiboPrinter software compared to BioNumerics bioinformatics software. A description of the approach taken for two BST projects in Texas and data analysis challenges are presented. Includes figures.