Elbow osteoarthritis (OA) is the most common cause of forelimb lameness in the dog. The goal of nonsurgical and/or surgical management of the diseases that cause elbow OA is to slow the progression OA in the joint and reduce lameness in the patient. Nonsurgical and surgical management of these conditions frequently leads to unsatisfactory results. Current treatment alternatives for dogs with moderate to severe elbow OA include nonsurgical management, removing loose bodies and osteophytes from the joint, and arthrodesis. Similarly, these treatments frequently lead to a poor outcome. Improvements in implant design and surgical techniques have made total elbow arthroplasty a satisfactory treatment for arthritic disorders of the elbow in humans since the mid-1970's. In two separate evaluations, 91% of total elbow arthroplasty cases had excellent long-term (∼4 years) outcomes. It is possible that total elbow arthroplasty in the dog would also yield a good clinical outcome. The anatomy and mechanics of the elbow joint, unfortunately, are dramatically different between humans and dog. Thus, a total elbow arthroplasty system that can consistently provide a good result needs to be developed. Total elbow arthroplasty has been reported in the dog, the results from these reports, however, have been poor;In this thesis the results from an anatomical study of the dog elbow are reported. Those results were used to design an initial total elbow arthroplasty system. That system was studied in vivo in the normal dog and yielded inconsistent results. Modifications of the initial design were made and a two component, nonconstrained system was used in a second in vivo study in normal dogs. Using the modified system, 50% of the dogs studied had normal limb function in the operated limb one-year after surgery. Additional modifications were made and a final the total elbow arthroplasty system is proposed for use in client owned animals with naturally occurring elbow OA.