Abstract The purpose of this study was to evaluate the reliability of differential dark‐field microscopy (DDFM) of subgingival bacteria to correctly predict the recurrence of periodontitis in treated patients on 2 different recall schemes. The results are based on data from 51 patients on 3‐month recalls (C group) and 39 patients (T group) who received periodontal prophylaxes according to a customized regimen based on DDFM of subgingival bacterial morphotypes. No significant differences were detected in the rate of recurrence of periodontitis between the 2 groups, although the incidence of disease recurrence tended to be greater in the C group. The proportions of spirochetes and motile rods observed at base line, or the average of the values observed at base line, 3 months and 6 months, were compared between subjects with and without disease recurrences. Significantly elevated proportions of these bacterial morphotypes were observed in patients of the T group who demonstrated evidence of disease recurrence during the first year as compared to subjects without disease recurrence. No significant differences could be detected in bacterial morphotypes between patients resistant to disease recurrence and those who developed recurrences at some time during the 3‐year period covered by this report.The ability of DDFM of subgingival bacterial morphotypes to predict future disease occurrence was tested for 1‐ and 3‐year periods, in both the C and T groups. The test appears to be reliable only for subjects on the experimental maintenance regimen (group T) and for prediction of disease recurrence for the initial 1‐year period. The reliability of the test to predict disease recurrence over a 3‐year period fails off significantly in the T group. The test was not reliable in predicting disease recurrence in the C group, presumably because of the disrupting effect of regularly scheduled tri‐monthly prophylaxes on the composition of the subgingival microbiota.