Reports of methicillin-resistant Staphylococcus aureus (MRSA) in animals have become more frequent in recent years. This paper documents the recovery of MRSA from animals with respiratory, urinary tract or wound infection and from animals subjected to surgical procedures following treatment in one veterinary hospital and 16 private veterinary clinics in different geographical locations throughout Ireland. MRSA was recovered from 25 animals comprising 14 dogs, eight horses, one cat, one rabbit and a seal, and also from 10 attendant veterinary personnel. Clinical susceptibility testing suggested that the 35 isolates fell into two different groups. One group of isolates (Group 1) was resistant to one or more of the following classes of antimicrobials: macrolides, lincosamines, tetracyclines and/or fluoroquinolones. The second group (Group 2) was resistant to macrolides, aminoglycosides, tetracyclines and trimethoprim/sulphamethoxazole and variably resistant to fluoroquinolones, lincosamines and rifampicin. One isolate in Group 2 was susceptible to trimethoprim. Epidemiological typing by antibiogram-resistogram (AR) typing, biotyping and by chromosomal DNA restriction fragment length polymorphism analysis using SmaI digestion followed by pulsed field gel electrophoresis (PFGE), confirmed these two major clusters. PFGE analysis showed that most isolates from non-equine animals were indistinguishable from each other and from the isolates from personnel caring for these animals. MRSA was isolated from eight horses which attended six different veterinary practices before referral to an equine veterinary hospital. Isolates from the eight horses and from their attendant personnel had PFGE patterns that were indistinguishable and were unlike the patterns obtained from the other isolates. Comparison of PFGE patterns of isolates from veterinary sources with patterns from MRSA recovered in human hospitals showed that the most frequently occurring pattern of MRSA from non-equine animals was indistinguishable from the predominant pattern obtained from the most prevalent MRSA strain in the human population in Ireland. However, the patterns of the isolates from horses were unlike any patterns previously reported in Irish studies of human isolates. This study shows that transmission of two strains of MRSA is occurring in veterinary practices in Ireland and that one strain may have arisen from human hospitals. The source of the second strain remains to be determined.