BACKGROUND: The arcuate nucleus, when damaged in young rats, can lead to pathological changes in adults, such as osteoporosis. Ovariectomized rats suffer from osteoporosis at eight weeks following surgery and the number of β -endorphin immunoreactive neurons in the arcuate nucleus of the hypothalamus is significantly decreased. OBJECTIVE: To establish a rat model of osteoporosis using ovariectomy and to explore changes in the number of β -endorphin neurons and to correlate any such change with serum hormone levels in response to exercise or rest. DESIGN, TIME AND SETTING: The completely randomized block design, neural morphology study was performed at the Key Laboratory of Physiology, Guangdong Medical College, China between March 2004 and January 2005. MATERIALS: Sixteen healthy female rats were selected for ovariectomy. METHODS: Following model establishment, rats were assigned to either rest or exercise groups and each rat was housed individually. Rats in the exercise group underwent an exercise regimen using a treadmill. MAIN OUTCOME MEASURES: Eight weeks following exercise, radioirnmunoassay was performed to detect serum growth hormone, estrogen and osteocalcin levels. Immunohistochemistry was used to measure changes in the number of β -endorphin neurons in the arcuate nucleus of the hypothalamus. Changes in bone metabolism were assessed using bone histomorphometry. RESULTS: In the exercise group, the β -endorphin immunoreactive neurons were high in number, darkly stained, and the nucleus was not obvious. In the rest group, the β-endorphin immunoreactive neurons were low in number and lightly stained. The number of β-endorphin immunoreactive neurons in the exercise group was higher compared with the rest group (t = 2.83, P < 0.05). Estrogen levels were similar between the re.st and exercise groups (P > 0.05). Serum osteocalcin and growth hormone levels were significantly higher in the exercise group compared with the rest group (t = 2.78, 2.32, P < 0.05). Compared with the rest group, the percentage of trabecular bone area, trabecular thickness, and trabecular number were significantly increased, but trabecular separation was significantly reduced (t=-2.48, 2.57, 2.32, 3.06, P < 0.05) in the exercise group. In the exercise group, the trabeculae of the tibia were arranged regularly and were high in number. In the rest group, the trabeculae of the tibia were organized in a disorderly manner and were low in number, with many fat particles. CONCLUSION: Exercise promotes bone growth and delays osteoporosis by inducing an increase in the number of β -endorphin immunoreactive neurons in the arcuate nucleus of the hypothalamus and by increasing serum growth hormone and osteocalcin levels.