A new potential therapeutic agent for postmenopausal osteoporosis, raloxifene, previously known as keoxifene, was evaluated by x-ray densitometry and more traditional techniques in quantitating the short-term (4-5 weeks) effects of ovariectomy on bones from 6-month-old rats. A Hologic QDR 1000/W and, to a limited extent, a Lunar DPXL, was used to quantitate ovariectomy, estrogen replacement, and raloxifene effects on vertebrae, femora, and tibiae. Both instruments performed well with precisions of 1.6% (Hologic) and 0.9% (Lunar) for anesthetized rats, which improved to 0.4% (Hologic) and 0.5% (Lunar) when the same rats were frozen. The lumbar vertebrae L1-4 showed a 12% decrease in bone mineral density 4 weeks after ovariectomy, compared with a 9% decrease for femora. Tibiae were also examined, but edge-detection problems prevented reproducible analysis of this site in vivo. The decrease in bone mineral density postovariectomy, especially for femora, was found to include both an increase in the projected area and a slight but not significant decrease in the bone mineral content of L1-4 and femora. These changes in density parameters of femora were supported by a decrease in dry weight and volume and a marginal increase in the second moment of inertia I for the identical femora examined ex vivo. Examination of individual lumbar vertebrae L1-5 suggested that the bone mineral density of L3 changes most dramatically in response to ovariectomy, but present techniques lack the spatial resolution and precision to quantitate bone changes reliably in individual vertebrae. 17 beta-Estradiol administered at 100 micrograms/kg/day subcutaneously inhibited ovariectomy effects on L1-4 bone mineral density, femoral moment of inertia, dry weight, and volume and to a lesser extent, femoral bone mineral density. A nonsteroidal compound, raloxifene HCl, at 1 mg/kg/day per os, had bone effects and effects on body weight that were largely indistinguishable from those of 17 beta-estradiol; however, raloxifene did not produce the uterotrophic effects observed with estrogen. The half-maximal efficacious dose of raloxifene on L1-4 bone mineral density was between 0.1 and 1.0 mg/kg/day per os. These data show that dual-energy x-ray absorptiometry compares favorably with traditional methods in quantitating bone changes caused by ovariectomy in small rodents, that L1-4 is a more sensitive region than whole femora in evaluating the effect of estrogen deficiency on bone loss, and the raloxifene may have promise as a treatment for conditions characterized by excessive bone loss after ovariectomy.