This trial was designed to test the effects of insulin, recombinant bovine somatotropin (rbST), and their interaction on milk protein and selected blood parameters in dairy cows. Eight Holstein cows (86 +/- 10 d in milk) were divided in two groups and used in two replicates of a Latin square design with four animals, four periods, and four treatments: 1) intravenous infusion of saline, 2) infusion of saline and subcutaneous administration of 40 mg of rbST per day, 3) intravenous infusion of 12 mg of insulin per day coupled with glucose infusion, and 4) rbST administration combined with insulin and glucose infusion. The glucose infusion rate was adjusted to maintain euglycemia. Each experimental period lasted 14 d: treatments were administered during the first 6 d, and no treatment was administered during the following 8-d resting phase. The average daily amount of glucose infusion needed to avoid hypoglycemia was 2.8 kg/cow when only insulin was infused as opposed to 2.2 kg/cow when both insulin and rbST were administered, indicating that either rbST causes a peripheral resistance to insulin or rbST increased liver gluconeogenesis or both. Data from the last 3 d of infusion were analyzed by using the SAS system for mixed models. Percent protein of milk tended to be lower (2.84 vs. 2.79%) and milk urea content was lower (16.6 vs. 14.8 mg/dl) during rbST administration, regardless of insulin infusion. Insulin infusion increased percent protein (2.78 vs. 2.85%) and percent casein (2.36 vs. 2.46%) and decreased milk urea content (17.1 vs. 14.3 mg/dl) regardless of rbST administration. For milk yield, protein yield, casein yield, lactose percent, and lactose yield, there were significant interactions between insulin and rbST administration. For example, casein yield averaged 1.17, 1.12, 1.20, and 1.28 kg/d for saline, insulin, rbST, and insulin combined with rbST, respectively. Similarly, there was a significant interaction between insulin and rbST on IGF-I levels, which were 122.5, 181.3, 342.3, and 492.2 ng/ml for saline, insulin, rbST, and insulin combined with rbST, respectively. In conclusion, these results clearly demonstrated that insulin interacts with bST in early lactation to improve milk protein synthesis and yield in dairy cows. These effects are probably mediated through a combination of bST nutrient mobilization, bST-induced gluconeogenesis, bST-induced insulin peripheral resistance, and bST/insulin synergism on insulin-like growth factor-I secretion and on mammary epithelial tissue.