Evaluation of periodontal therapy involves the use of several oral indices to describe the health status of hard and soft tissues. It was the objective of the present study to evaluate the reliability and reproducibility of some of these indices. A calibration and standardization session was designed to calibrate 10 examiners and a "gold standard" (an experienced examiner) in evaluating the following indices: the Volpe-Manhold calculus index (VM), the Lobene stain index (SI), a non-invasive modification of the Löe-Silness gingival index (GI), the papillary bleeding score (PBS) of Loesche, and the plaque index (PI) of Quigley-Hein as modified by Turesky. For each index, the average intraclass correlation was calculated between two subject visits. The highest intraclass correlation, 0.94, was found for PBS. The intraclass correlation for PI was 0.70 and for VM, 0.65. The lowest intraclass correlations were for stain, 0.47, and GI, 0.25. Intra-subject correlations between the 2 visits were good for all indices, but were best for PBS, followed by VM. PBS seems to be the most reliable index (both intra- and inter-examiner) for measuring the oral health status and is therefore recommended for use in clinical studies.