A comparative study was performed to examine the respective accuracy of 16S rDNA sequencing and of the commercial biochemical assay ID32 STAPH (bioMérieux, Marcy l'Etoile, France) in the identification of 232 staphylococcal samples representing 20 species and subspecies isolated from 367 dogs. Notable differences in species distribution were observed by comparing genotypic and phenotypic data. Partial sequencing of 16S rDNA resulted in an unambiguous identification of 226 (97.4%) of the isolates, whereas the phenotypic approach resulted in a correct diagnosis of 162 (69.8%) of the isolates. Statistical agreement between genotypic and phenotypic identification of staphylococci was substantial (Kappa coefficient of 0.6-0.8) for Staphylococcus aureus, S. hominis, S. warneri, S. cohnii subsp. urealyticus, and S. simulans, and "almost perfect" (Kappa coefficient of 0.8-1) for S. intermedius, S. epidermidis, S. equorum, S. haemolyticus, S. sciuri, and S. kloosi. No agreement above that expected by chance (Kappa coefficient=0) was observed for S. schleiferi subsp. coagulans, which was either confounded with S. intermedius and S. capitis, or categorized as unacceptable by the biochemical assay. Given the growing importance of this pathogen in veterinary medicine and its frequent misidentification with related staphylococci, a PCR-RFLP approach producing a S. schleiferi-specific restriction profile was developed. This fast and reliable assay represents a valuable tool in assisting in the monitoring of this pathogen.