The '122 patent claims the compound risedronate, the active ingredient of P's osteoporosis drug Actonel(R). Risedronate is a member of a group of compounds referred to as bisphosphonates that are active in inhibiting bone resorption. The first two bisphosphonates studied for the treatment of metabolic bone diseases, etidronate (EHDP) and clodronate, had clinical problems that prevented their commercialization. P conducted extensive experimentation involving hundreds of different bisphosphonate compounds. P could not predict the efficacy or toxicity of the new compounds. P identified risedronate as a promising candidate. On December 6, 1985, the inventors applied for a patent on the compound. Risedronate is neither claimed nor disclosed in P's '406 patent. The '406 patent 'addresses the central problem seen in bisphosphonates at the time, namely that they inhibited bone mineralization, by teaching the use of a cyclic administrative regimen to achieve a separation of the benign effect of anti-resorption from the unwanted side effect of anti-mineralization in patients.' The '406 patent lists thirty-six polyphosphonate molecules as treatment candidates and eight preferred compounds for intermittent dosing, including 2-pyr EHDP. D announced to the world its plans to sell risedronate as a generic equivalent. D contends that the structural similarities between risedronate and 2-pyr EHDP render the '122 patent obvious. The court concluded that the '406 patent would not have led a person of ordinary skill in the art to identify 2-pyr EHDP as the lead compound. In light of the extremely unpredictable nature of bisphosphonates at the time of the invention, it found that a person of ordinary skill in the art would not have been motivated to make the specific molecular modifications to make risedronate. It held that that unexpected results of risedronate's potency and toxicity rebut a claim of obviousness. D appealed.