The present paper on the design of clinical trials of periodontal therapy first addresses the issue of the etiology of periodontal disease. It is suggested that most if not all forms of destructive periodontal disease are caused by microorganisms and that there are different forms of disease with different microbial etiologies. The progressive nature of destructive periodontal disease is subsequently discussed and it is emphasized that, in a given patient, periodontal sites which show signs of inflammation and attachment loss may not over a period of several months and years show further sign of attachment loss. The present methods of assessing periodontal disease do not allow us to discriminate between potentially active and inactive sites in untreated patients. The significance and variability of indicators of periodontal disease such as bleeding on probing, probing pocket depth and probing attachment level measurements are discussed. The errors inherent in the various measurements are analyzed and suggestions are presented describing how alterations in any of the above parameters could be identified and presented in a clinical trial. Of concern for the statistical analysis of clinical data of periodontal disease is the definition of the "experimental unit". For a number of years, the "experimental unit" in periodontal trials was the patient. It is clear, however, that different sites within the same individual show different patterns of disease progression and lesion morphology and often respond differently to periodontal therapy. Statistical analyses must consequently be designed which recognize differences in site-to-site infection and lesion morphology within a common host. Until such analyses are available, the investigator should be wary of pooling data within the same individual, since such pooling may obscure meaningful alternatives which may take place in individual periodontal sites. Some goals of periodontal therapy are subsequently identified. 4 goals are discussed more in detail, namely: to establish conditions which will allow the patient to maintain a dentition without further breakdown of the periodontium; to reduce pocket depth to establish an anatomy in the dentogingival region which with proper maintainance care will prevent the re-establishment of the subgingival infection; to gain attachment as a result of treatment; to assess the effect of a certain chemotherapeutic agent on periodontal disease.