In fish, regulation of the growth hormone (GH) receptor has been mainly analyzed by binding studies, with some discrepancies in the results. The present work aims at determining whether circulating GH levels influence the measurement of hepatic GH-binding capacities. To do this, the effectiveness of the dissociation of the GH/GH receptor (GHR) complex by an MgCl2 treatment was assessed in rainbow trout, and data on GH-binding capacities under various physiological conditions were studied using different means of expression. Our results reveal that MgCl2 treatment dissociated the liver GH/GHR complex formed under in vitro conditions (85 +/- 23 vs. 361 +/- 16 fmol/g of liver; p < 0.001) but not in vivo, showing such treatment in trout is not applicable. A comparison of fasted ( 3 weeks) and fed fish revealed that GH-binding capacities, expressed as femtomoles per milligram protein or femtomoles per gram of liver, were similar in both fed and fasted fish. However, when changes in liver and body size were taken into account, the total GH-binding capacities were lower in the fasted fish (0.026 +/- 0.006 vs. 0.062 +/- 0.009 fmol/cm 3 liver; p < 0.05). One day after hypophysectomy or GH injection, changes in the plasma GH levels increased or decreased GH-binding capacities, respectively. Five days later, GH-binding capacities increased in GH-injected fish (527 +/- 38 vs. 399 +/- 38 fmol/g liver; p < 0.01). Our interpretation is that acute treatment modified GH-binding capacities through receptor occupancy and that GH stimulated the synthesis of its own receptor. On the other hand, long- term treatment through successive injections of GH lowered the total binding capacities (approx. 40%), which could result from receptor occupancy. We conclude that circulating GH levels strongly influence the measurement of GH-binding capacities in the liver, thereby limiting interpretation of the binding data and preventing accurate conclusions to be drawn on GHR regulation.