A novel and highly reproducible amplified fragment length polymorphism (AFLP) typing approach was developed for typing of Enterococcus strains from the environment. Pooled and corresponding single faecal sample isolates were analysed to test the efficiency and coverage of dominant isolates for future sampling procedures. AFLP development was based on the selection of appropriate restriction enzymes and the design of adaptors and primers which was supported by in silico optimisation of selective bases using Enterococcus spp. genome data. Three optimal combinations of selective bases at the 3' end of the designed primers (i.e., CC, GG, CG) could be determined. AFLP fragment analysis using a capillary sequencer and intralane standardisation resulted in excellent methodical stability (> or =98% similarity for GG and > or =94% similarity for CC). Furthermore, the developed typing method was evaluated on 16 type trains of the genera Enterococcus and Streptococcus and 398 faecal isolates of cow pats from five alpine pastures in a karstic catchment area. Statistical analysis revealed a discrimination capacity of DI > or =0.95 (Simpson Diversity Index) and a reproducibility level of > or =94% similarity indicating the methods high typing capacity and robustness. Results of the comparative analysis of single and pooled faecal samples indicate that for a "strain to strain" based faecal source tracking, pooled faecal samples rather than single faecal samples are likely to be the most efficient sampling strategy for collecting the abundant corresponding strains.