Polycystic ovary syndrome (PCOS) is characterized by oligo/anovulation, hyperandrogenism, and in many cases, polycystic ovaries. Women with PCOS often exhibit persistently increased gonadotropin-releasing hormone (GnRH) pulse frequency compared to normal cycling women, contributing to enhanced luteinizing hormone (LH) secretion and increased androgen production [1]. Administration of progesterone (P) decreases GnRH pulse frequency in normal women, but the GnRH pulse generator in women with PCOS is relatively insensitive to P negative feedback [2]. Treatment with the androgen-receptor antagonist flutamide restores GnRH pulse generator sensitivity to P inhibition in women with PCOS [3]. Hyperandrogenemia in adolescence can represent a forerunner of adult PCOS, though the etiology of PCOS remains unclear. As a group, adolescent girls with HA display impaired GnRH pulse generator sensitivity to progesterone inhibition (i.e., reduced P-sensitivity). However, P-sensitivity is varied in HA girls, despite comparable levels of plasma free testosterone (T) [4] . The androgen receptor (AR) gene contains a polymorphic trinucleotide (CAG) repeat sequence located in the N-terminal transactivation domain of the AR [5] , and the number of CAG repeats influences AR activity. In vitro, expression of AR alleles with shorter CAG repeat lengths is associated with increased AR activity [6]. We measured AR CAG repeat length in a subgroup of subjects previously shown to have varied GnRH pulse generator sensitivity to P-mediated slowing [4] , hypothesizing that shorter AR CAG repeat length would be associated with reduced P-sensitivity in HA girls.