In a 4-year controlled, prospective trial, histomorphometric analysis was used to compare the tissue-level skeletal effects of fluoride therapy in 43 postmenopausal women (75 mg NaF/day) with those of 35 matching placebo subjects; all subjects received 1500 mg/day elemental calcium supplement. In addition to an initial, baseline biopsy, a second biopsy was obtained after 6, 18, 30 or 48 months. Measurements were made on a third biopsy obtained from 8 subjects following at least 72 months of fluoride therapy. The change in cancellous bone volume or trabecular thickness in fluoride-treated subjects was not different from a change in placebo-treated subjects. However, paired analysis in the fluoride-treated subjects indicated that bone volume was increased between the first and second biopsies (p < 0.005). Both osteoid length and width were significantly increased in fluoride compared with placebo subjects; however, only the osteoid surface increased linearly (r = 0.63, p < 0.001). The mineral apposition rate and relative tetracycline-covered bone surface were not different between fluoride and placebo treatment, although they were decreased in both groups in the second biopsy. The tetracycline-covered bone surface returned to normal in the third biopsy. Definitive evidence for osteomalacia is a prolonged mineralization lag time, which following fluoride treatment was found to be increased 9-fold in the second biopsy and 4-fold in the third biopsy. Further evidence for osteomalacia was increased osteoid thickness by 6 months, evidence of focal areas of interstitial mineralization defects, and broad tetracycline labels of low fluorescence intensity. In the third biopsies, osteoclastic resorption was observed beneath osteoid seams. Fluoride therapy increased the cortical width compared with placebo treatment (p < 0.02), and increased the osteoid surface in Haversian canals, but did not change the osteoid width, resorption surface or cortical porosity. After an initial rise, serum fluoride levels remained constant, and the urine values fell slightly. The bone fluoride concentration rose throughout the treatment period, and was correlated with the change in osteoid-covered bone surface (r = 0.56, p < 0.001). Although we found definitive evidence for osteomalacia, the cause of the osteomalacia was not determined in this study. On the other hand, the presence of bone resorption beneath unmineralized osteoid and of osteocyte halos is suggestive of hyperparathyroidism. Thus, it is possible that the strong stimulus for bone formation brought about by fluoride therapy resulted in relative calcium deficiency.