The number of Roman sarcophagi without explicit Christian iconography conventionally dated to the fourth century is not commensurate with any reasonable estimate of the number of well-to-do pagans. This article explores several possible explanations for the anomaly. One approach would be to attempt to correct the archaeological record by finding errors in the religious classification of monuments that exaggerate the Christian corpus, adding non-Christian sarcophagi that have escaped published inventories, or establishing a systematic misdating of pagan sarcophagi. Alternatively, the preserved monuments could be taken as an accurate reflection of original production, the shortfall implying some change in commemorative habits specific to non-Christians. The author concludes that neither of these theories is likely to reduce the pagan sarcophagus deficit substantially. Instead, the shortfall is ascribed mainly to differential rates of preservation. This hypothesis is consistent with certain medieval practices of reuse that suggest a higher probability of survival for antique sarcophagi bearing Christian imagery.