The purpose of this investigation was to develop a self-administered questionnaire to measure sexual desire. In the development phase, items were generated and pilot-tested with 24 subjects. Based on these data, items were deleted, added, or modified. Next, in Study One, the Sexual Desire Inventory (SDI) was administered to 197 females and 117 males. Factor analyses revealed that the SDI was multifactorial; however, none of the generated factor solutions up to five factors yielded a good fit. Interpretation of the factors led to revisions of the SDI. It was hypothesized that sexual desire might consist of two related dimensions; dyadic sexual desire and solitary sexual desire. Items on the SDI were modified to measure these two dimensions, and the revised SDI was administered to 249 females and 131 males. Factor analysis supported the presence of these two dimensions. Internal consistency estimates using Cronbach's alpha revealed coefficients of .86 for dyadic sexual desire and .96 for solitary sexual desire, providing preliminary evidence for the reliability of the SDI. The implications of these findings are discussed.