Four adult Merino sheep were used in the experiment, which was divided into four parts. For the 5 days before parts 1 and 3 saline was injected and for the 5 days before parts 2 and 4 growth hormone (GH; 4 mg day-1 subcutaneously) was injected. In parts 1 and 2 a primed continuous infusion of [6,6-2H2]glucose and either saline or GH, respectively, were infused for 5 h. The first 3 h was the control period. From 3 to 5 h insulin (0.5 mU kg-1 min-1) was infused. Coincident with the insulin infusion, normal glucose was also infused at a variable rate, dependent on the rapidly determined plasma glucose concentration, in order to keep the plasma glucose concentration constant. Parts 3 and 4 of the experiment were the same as parts 1 and 2, respectively, except for the following: the glucose isotope and saline or GH were infused for 7 h, from 3 to 7 h somatostatin (SRIF; 0.417 microgram kg-1 min-1) was infused, and from 5 to 7 h insulin was infused. Measurements of glucose turnover were made in the last 40 min of the control, insulin-only, SRIF-only and insulin-plus-SRIF infusion periods. Plasma insulin levels were reduced to below the level of detection by the SRIF infusion; under such conditions whole body glucose uptake should be entirely non-insulin mediated (NIMGU). Expressing glucose uptake as glucose metabolic clearance rate revealed that GH had no effect on NIMGU but significantly reduced the level of insulin-mediated glucose uptake (IMGU). Thus a reduction in the rate of NIMGU is probably not part of the mechanism by which GH repartitions glucose to sites of growth and milk production, whilst the present study confirms the antagonistic effect of GH on IMGU.