Smallpox, caused by variola virus (VARV), is a devastating human disease that affected millions worldwide until the virus was eradicated in the 1970 s. Subsequent cessation of vaccination has resulted in an immunologically naive human population that would be at risk should VARV be used as an agent of bioterrorism. The development of antivirals and improved vaccines to counter this threat would be facilitated by the development of animal models using authentic VARV. Towards this end, cynomolgus macaques were identified as adequate hosts for VARV, developing ordinary or hemorrhagic smallpox in a dose-dependent fashion. To further refine this model, we performed a serial sampling study on macaques exposed to doses of VARV strain Harper calibrated to induce ordinary or hemorrhagic disease. Several key differences were noted between these models. In the ordinary smallpox model, lymphoid and myeloid hyperplasias were consistently found whereas lymphocytolysis and hematopoietic necrosis developed in hemorrhagic smallpox. Viral antigen accumulation, as assessed immunohistochemically, was mild and transient in the ordinary smallpox model. In contrast, in the hemorrhagic model antigen distribution was widespread and included tissues and cells not involved in the ordinary model. Hemorrhagic smallpox developed only in the presence of secondary bacterial infections - an observation also commonly noted in historical reports of human smallpox. Together, our results support the macaque model as an excellent surrogate for human smallpox in terms of disease onset, acute disease course, and gross and histopathological lesions.