The putative negative feedback effects of IGF-I and IGF-II on GH secretion were tested by intracerebroventricular (icv) and intrapituitary administration to sheep. Over two consecutive days, serial jugular blood samples were taken at 10 min intervals for 6 h from ewes (n = 3/group) fitted with indwelling stainless steel cannulae into the lateral or third cerebral ventricles. The sheep were injected (icv) with either vehicle or purified ovine IGF-I (2, 4 or 8 micrograms). IGF-I injection had no effect on plasma GH secretion. Serial blood samples were taken from a second group of nine ewes in which ovine or recombinant human (rh) IGF-I was infused (2.5 micrograms/h for 2 h) into the third ventricle; once again, IGF-I failed to affect the episodic pattern of GH secretion. Three ewes fitted with indwelling stainless steel cannulae placed in the anterior pituitary gland were consecutively infused with either ovine or rhIGF-I (2.5 micrograms/h for 2 h) or vehicle. Plasma GH concentrations were suppressed in 3/3 sheep from 1-1.5 h after the commencement of infusion and GH levels remained low for the remainder of the sampling period. In another group of five ewes synergistic effects of IGF-I and IGF-II on GH secretion were tested by icv infusion of rhIGF-I, rhIGF-II, or rhIGF-I+rhIGF-II (5 micrograms/h for 2 h) or vehicle (sterile 10 mM HCl/saline). Each sheep received each treatment in a randomised design. Infusion (icv) of IGF-I and IGF-II alone or in combination failed to alter GH secretion.(ABSTRACT TRUNCATED AT 250 WORDS)