Most veterinary and milk hygiene laboratories identify streptococci and enterococci based on serological and biochemical tests. The analysis of 16S rDNA was suggested to be used for more exact identification; however, its use has not been considered so far in monitoring studies. The objective of the present study was to compare a conventional phenotypic method with restriction fragment length polymorphism analysis of 16S rDNA (16S rDNA RFLP) for identification of streptococci isolated from composite milk samples collected in connection with intramammary infection (IMI) in six Argentinean dairy farms. Composite milk samples (n = 1223) from cows belonging to six herds were collected for bacteriological analysis. Twelve reference strains and fifty streptococci or streptococcuslike isolates were identified to species level by the API 20 Strep system, conventional biochemical tests and 16S rDNA RFLP in a blind assay. The remaining streptococci or streptococcus-like isolates (n = 40) were identified to the species level both by 16S rDNA RFLP and conventional biochemical tests. As indicated by Kappa values, agreement between the 16S rDNA RFLP and the conventional scheme for identification of Streptococcus agalactiae, S. dysgalactiae, S. uberis, S. equinus and Enterococcus faecalis was 0.91, 0.73, 0.92, 0.81 and 0.85, respectively. Together with the less frequently isolated streptococcal species, the conventional scheme correctly identified 77 out of 90 isolates (85.5%). Thus, the use of 16S rDNA RFLP is considered valuable for monitoring studies due to its affordable cost for standard laboratories.