The response of canine lumbar vertebral bone to the application of a stiff transpedicular screw-plate implant from the third to the fifth lumbar vertebra, without arthrodesis, was investigated. Five groups of six dogs each were studied: dogs that had not had an operation (control group); dogs that had had a sham operation, with survival periods of three and six months; and dogs that had received an implant, with survival periods of three and six months. The results were the same in specimens from the control group and the group that had had a sham operation. In the groups that had received an implant, dual-photon absorptiometry revealed an insignificant decrease in bone-mineral content at the bypassed fourth lumbar segment (17 per cent at three months and 12 per cent at six months). When the data for the three and six-month intervals were pooled, the mean decrease in bone-mineral content of 14 per cent was significant. Histomorphic study yielded similar results; the mean decrease in bone-mineral content for the pooled three and six-month specimens was significant (16 per cent). Losses were similar for the ventral and dorsal columns. Histomorphometric analysis also showed a significant (13 per cent) loss of bone-mineral content at the adjacent caudad (sixth lumbar) vertebra for the pooled three and six-month groups. The results of the methods of analysis of loss of bone-mineral content correlated strongly. Under these experimental conditions, a stiff spinal implant caused loss of bone-mineral content of the bypassed vertebral segment, although the loss was less than anticipated and did not increase between three and six months.