Capsular calcification was present clinically in 64 of 404 silicone gel breast implant capsules (15.8%) analyzed from 1981 to 1996. It presented as white-gray plaques on the inner surface of capsules in 62 of 64 capsules, and as massive heterotopic ossification in 2 capsules. Chi-squared analysis confirmed that calcification was related to the generation of the implant (i.e., year of manufacture; p < 0.001). All 28 first-generation implants (1963-1972, with Dacron patches) were clinically intact and all demonstrated extensive calcification. Their mean duration in situ was 17.6 years (range, 14-28 years). Thirty-four of the 348 second-generation implants (9.8%; 1973-1987) were associated with capsular calcification. Their mean duration in situ was 16.0 years (range, 13-22 years). Because all first-generation implants demonstrated calcification, they were compared with the second-generation implants that had been in place for the same duration (>14 years). Only 42% of these 81 second-generation implants demonstrated calcification, compared with 100% of the first-generation implants (p < 0.001). Thus, thicker first-generation implants with Dacron patches are more likely to calcify and the effect is not entirely due to their longevity. None of the 28 third-generation implants (1987-1991) demonstrated calcification. Their mean duration in situ was 4.2 years (range, 2-7 years). For second-generation implants, calcification was related to duration in situ (p < 0.001). None of the 294 implants in place for less than 11 years were associated with significant clinical calcification. The percentages of capsules with calcification were 13 to 14 years, 33%; 15 to 16 years, 45%; and 17 to 22 years, 57%. Calcification with second-generation implants was not associated with patches on the envelopes. Of the 34 second-generation implants with calcification, only two had patches (composed of silicone, not Dacron). Among second-generation implants, calcification was related to implant integrity. Of implants in place for more than 12 years, 52.5% of those implants that were ruptured showed calcification, but only 10.0% of intact implants demonstrated calcification (p < 0.001). Seventeen of the 64 calcified capsules were examined histologically. In all of these specimens, calcification existed in two forms: globular aggregates on the surface of the capsule (adjacent to the implant) and actual bone formation within the fibrous tissue of the capsule. All calcified capsules demonstrated both globular aggregates and true bone formation regardless of the implant generation, duration in situ, or integrity. Ultrastructural analysis was performed on four capsules from 2 women who had received first-generation Dow Corning gel implants 24 and 28 years previously, and on 2 capsules from one woman who had received Heyer-Schulte gel implants 21 years previously. These capsules were analyzed according to distribution, density, mineral nature, crystal phases, and elements within crystals by electron microscopy, energy-dispersive X-ray spectrometry, and electron diffraction. These analyses confirmed two types of calcification, each with hydroxyapatite crystals. In areas of heterotopic bone, crystals 40 x 10 nm were deposited in an orderly fashion on collagen fibers. In contrast, in areas of globular aggregates, spherulitic aggregates of much larger crystals were present, without any relationship to the collagen. Titanium was demonstrated in capsules of first-generation implants at areas of attachment of the Dacron patches. The calcification associated with saline implants revealed only one form of crystal: agglomerates, which were adherent to the elastomeric shell of the implants. A hypothesis is presented to explain the differences in calcification deposition properties between silicone gel-filled and saline-filled breast implants.