The evolution of criteria for success of dental implants reflects the evolution of knowledge of implant performance and biologic reactions to them. Most criteria are written in the interest of predictability of performance of an implant system. This approach is valuable to the clinician in selecting implants for patients, but the criteria for bone level changes may be unnecessarily stringent for monitoring performance in an individual patient. Some experience and difficulties related to radiographic monitoring are reviewed. The need for tempering the application of marginal bone criteria to clinical applications is illustrated by longitudinal data on marginal bone height changes.