Human growth hormone is, in effect, defined by its activity in an in vivo bioassay and the standard used with it, growth being measured as the increase in body weight in hypophysectomised immature rats. The assay reflects the hormone's survival and metabolism in vivo, its cell-cell interactions, the activation and effects of its secondary hormones, such as GF1 and GF2, and various feedback mechanisms. Although it is insensitive, imprecise, easily influenced by contaminants TSH and vasopressin, it is the only practical assay that reflects all the in vivo properties of "hGH". The in vivo tibial epiphysis bioassay is more sensitive and precise, but the response reflects only the elongation of bone. Both these bioassays are well established. By contrast, in vitro receptor assays do not reflect in vivo properties; there may be different natural forms of receptor molecules, they may be altered during their extraction, and the measured response (like those of immunoassays) is not relevant to the biological action of the hormone. The validity of a bioassay depends on the use of a suitable standard. The collaborative study of the International Standard for human growth hormone (in 1984) revealed marked disparities between results with different assay methods. When a growth hormone protein (such as somatotropin, 191 amino acids) is produced in quantity, reproducibly, and with physicochemical properties consistently related to in vivo bioassay results, it may then be reasonable to use physico-chemical tests for control purposes. Many such tests require international reference materials for comparison purposes.