Up to 90% of patients with castrate-resistant prostate cancer develop bone metastases, and the majority of these men have received androgen deprivation therapy known to cause bone loss. Whether this treatment-induced change to the bone microenvironment affects disseminated tumour cells, potentially stimulating development of bone metastasis, remains to be determined. The objective of this study was to use an in vivo model mimicking androgen ablation to establish the effects of this intervention on disseminated prostate cancer cells in bone. We mimicked the effects of androgen deprivation on bone metastasis by castrating 12-week-old BALB/c nude mice that had disseminated, hormone-insensitive PC3 prostate cancer cells present in the long bones. Castration caused increased bone resorption and loss of bone volume, compared with sham operation. In addition, castration triggered growth of disseminated PC3 cells to form bone metastasis in 70% of animals. In contrast, only 10% of sham-operated animals had detectable long bone tumours. Weekly administration of 100 μg/kg zoledronic acid (ZOL) prevented castration-induced tumour growth in bone and increased bone volume, but did not eliminate the disseminated tumour cells. ZOL had no effect on tumour growth in the sham-operated animals, despite causing a significant increase in bone volume. This is the first demonstration that, in a model of prostate cancer bone metastasis, mimicking androgen ablation results in growth of disseminated tumour cells in bone through osteoclast-mediated mechanisms. We provide the first biological evidence supporting the administration of ZOL to prostate cancer patients at the time of androgen ablation to prevent subsequent relapse in bone.