Cytosolic, detergent-solubilized and membrane-bound growth hormone (GH) receptors from rabbit adipose tissue and liver were tested for reactivity with a panel of monoclonal antibodies (MAbs). The cytosolic and detergent-solubilized forms of adipose tissue and liver GH receptors were identically reactive with four precipitating and two hormone-binding-site-directed MAbs. However, the membrane-bound form of the adipose receptor was 1000-fold less reactive with one binding-site-directed MAb (MAb 7) than the membrane-bound liver GH receptor. Reactivity with another inhibitory MAb (MAb 263) was identical for adipose tissue and liver membrane GH receptors. The relative potency of 22,000-Mr and 20,000-Mr forms of human GH was identical in assays with liver and adipose tissue membrane receptors. Thus, contrary to earlier suggestions, the discrepancy between the growth-promoting and insulin-like activities of 20,000-Mr human GH cannot be rationalized by a difference in the affinity of this hormone for 'somatogenic' and 'metabolic' receptors when the comparison is made in the same species. Cross-linking studies showed that the major GH-binding subunit of liver and adipose tissue GH receptors had the same Mr (54,000 +/- 5000, reduced). The ligand-binding subunits of liver and adipose tissue receptors are identical by several criteria, but one epitope on the adipose tissue receptor appears to be masked upon membrane insertion, possibly by close association with a tissue-specific component. Tissue specificity may be determined by association of a ubiquitous GH-binding subunit with tissue-specific membrane components, rather than by differences in amino acid sequence.