Rats bearing mammosomatotropic tumors have raised insulin but lowered glucose concentrations. To determine if growth hormone (GH) secreted by these tumors causes insulin antagonism, pancreatic suppression tests utilizing infusions (per kg per min) of glucose (8 mg), insulin (200 ng) and somatostatin (1.4 μg) for 130 min were performed. Although the steady state plasma glucose and insulin levels (mean of 90, 100, 110, 120 and 130 min samples) were similar in 8 control and 13 tumor-bearing rats, the decrease from the already depressed basal glucose concentration (mmoles/l±SE) in the tumor animals was less than in the controls (0.90 ± 0.30 vs. 2.56 ± 0.040, p < 0.005). Since the interpretation of these results was not entirely clear, glucose and insulin-glucose tolerance tests were performed. The glucose disappearance rates (%/min ± SE) in the glucose tolerance test were lower in 17 tumor rats (2.00 ± 0.13) compared to 17 control animals (2.51 ± 0.22). This difference just missed statistical significance (t = 2.00, value of 2.04 necessary for p = 0.05). The decrease occurred in the presence of increased insulin (nmoles/l ×16 min) levels (4.29 ± 0.38 vs. 2.58 ± 0.29, p < 0.005) suggesting insulin antagonism. The glucose disappearance rates (%/min ± SE) in the insulin-glucose tolerance test were less in 12 tumor-bearing rats compared to 11 control animals (2.80 ± 0.29 vs. 4.12 ± 0.35, p < 0.02). Thus, these GH-secreting tumors cause insulin antagonism in vivo. Freshly isolated hepatocytes from these tumor-bearing animals manifest decreased insulin binding and action (Diabetologia 25: 60, 1983). In the present study, however, insulin binding and action (net glucose-C¹⁴ incorporation into glycogen) were normal after the hepatocytes were cultured for two days. This suggests that the changes induced by GH in vivo that lead to insulin antagonism are short-lived.