To investigate the impact of human papillomavirus (HPV) on the epidemiology of oral squamous cell carcinomas (OSCCs) in the United States, we assessed differences in patient characteristics, incidence, and survival between potentially HPV-related and HPV-unrelated OSCC sites. Data from nine Surveillance, Epidemiology, and End Results program registries (1973 to 2004) were used to classify OSCCs by anatomic site as potentially HPV-related (n = 17,625) or HPV-unrelated (n = 28,144). Joinpoint regression and age-period-cohort models were used to assess incidence trends. Life-table analyses were used to compare 2-year overall survival for HPV-related and HPV-unrelated OSCCs. HPV-related OSCCs were diagnosed at younger ages than HPV-unrelated OSCCs (mean ages at diagnosis, 61.0 and 63.8 years, respectively; P < .001). Incidence increased significantly for HPV-related OSCC from 1973 to 2004 (annual percentage change [APC] = 0.80; P < .001), particularly among white men and at younger ages. By contrast, incidence for HPV-unrelated OSCC was stable through 1982 (APC = 0.82; P = .186) and declined significantly during 1983 to 2004 (APC = -1.85; P < .001). When treated with radiation, improvements in 2-year survival across calendar periods were more pronounced for HPV-related OSCCs (absolute increase in survival from 1973 through 1982 to 1993 through 2004 for localized, regional, and distant stages = 9.9%, 23.1%, and 18.6%, respectively) than HPV-unrelated OSCCs (5.6%, 3.1%, and 9.9%, respectively). During 1993 to 2004, for all stages treated with radiation, patients with HPV-related OSCCs had significantly higher survival rates than those with HPV-unrelated OSCCs. The proportion of OSCCs that are potentially HPV-related increased in the United States from 1973 to 2004, perhaps as a result of changing sexual behaviors. Recent improvements in survival with radiotherapy may be due in part to a shift in the etiology of OSCCs.