A multicentric study of clinical Staphylococcus isolates was performed by seven operative units working in different areas of Italy. Over a 6-month period, a total of 3,226 staphylococci, isolated from in- and outpatients, were identified and tested for antimicrobial susceptibility by a protocol agreed upon by all units. On the basis of their bacteriolytic-activity patterns and other conventional tests, the isolates were identified by lyogroups , which closely correlate with human Staphylococcus species. Lyogroup I (Staphylococcus aureus) and lyogroup III (Staphylococcus capitis) were the most and the least frequently isolated staphylococci, respectively. Significant differences depending on strain origin from in- or outpatients were only observed with lyogroup IV (i.e., novobiocin- resistant staphylococci), whose isolation from outpatients was three times greater than from inpatients. Lyogroup I was predominant among isolates from most clinical sources. Lyogroup IV predominated in strains isolated from the urinary tract; lyogroup V (Staphylococcus epidermidis) predominated in strains from blood, cerebrospinal fluid, and indwelling artificial devices; and lyogroup VI ( Staphylococcus hominis, Staphylococcus haemolyticus, and Staphylococcus warneri ) predominated in strains from bile and the male genital tract. The incidence of methicillin resistance within the different lyogroups varied from unit to unit, suggesting epidemiological differences among different hospitals and different geographical areas. On the whole, methicillin resistance was more frequent in coagulase-negative staphylococci than in S. aureus and ranged from 19% for lyogroups I and III to 30% for lyogroup II (Staphylococcus simulans). Laboratory testing with 18 additional antibiotics suggested the occurrence of some specific differences in susceptibility among the different lyogroups . The rate of organisms resistant to the various antibiotics was greater among methicillin-resistant than among methicillin -susceptible staphylococci; particularly marked differences occurred with cephalosporins, rifampin, gentamicin, and tobramycin. The results suggested an increasing spread in Italy, during the last few years, of staphylococcal resistance to methicillin and to many other antibiotics. Some questions about the actual reliability of laboratory tests for the determination of staphylococcal susceptibility to methicillin and other beta-lactam antibiotics were raised by parallel test performances in which both unsupplemented and 5% NaCl-supplemented Mueller-Hinton agars were used. The presence of NaCl heightened, on the whole, the number of resistant strains detected; however, a few isolates resistant in the unsupplemented medium and susceptible in the salt-supplemented medium were also encountered. This was true not only for methicillin but also for all other beta-lactam antibiotics tested except cefamandole. With cefamandole, the presence of 5% NaCl reduced the number of resistant strains detected.