The Gordon Holmes spinocerebellar ataxia syndrome (GHS) is associated with idiopathic hypogonadotrophic hypogonadism (IHH). There are conflicting reports in the literature as to whether the primary neuroendocrine defect is of hypothalamic GnRH secretion, as with most causes of IHH, or of pituitary resistance to GnRH action. Because of the anatomical inaccessibility of the hypophyseal portal circulation, direct measurement of GnRH levels in human subjects is not possible. Previous investigators have attempted to unravel this problem through the use of GnRH stimulation tests and the limitations of this approach may explain the differing results obtained. We used the more physiological approach of treating a male GHS patient for four weeks with GnRH, 7-10 microg/pulse, delivered subcutaneously at 90 minute frequency via a portable minipump. This therapy failed to induce any rise in plasma gonadotrophin and testosterone concentrations. By contrast, eight weeks treatment with exogenous gonadotrophins maintained physiological plasma testosterone concentrations and induced testicular enlargement with induction of spermatogenesis. The data indicate that the primary endocrinopathy in GHS is of pituitary gonadotrophin secretion and not of hypothalamic GnRH. Moreover, the patient did not harbour any mutation of the GnRH receptor gene. Two clinical observations are consistent with progressive involution of gonadotrophic function, rather than a congenital gonadotrophin deficiency. First, the patient's development was arrested at early mid-puberty at the time of original presentation and, second, effective spermatogenesis was induced extremely rapidly during gonadotrophin treatment, suggesting prior exposure of the testes to FSH. Both spinocerebellar ataxia and pituitary dysfunction might thus have been in evolution since late childhood.