If one assumes that cat-scratch disease is an entity caused by a specific virus or other living or nonliving agent, the reservoir or source of this etiologic agent still remains unknown. The assumption that an infection has occurred may not be justified from the available findings, since this disease has been transmitted to only one human volunteer1 and three or four monkeys* and has not been serially passed in any other animals or other media. Serological tests suggesting a relation of the disease to the psittacosis-lympho-granuloma virus group are also inconclusive. In addition, what were originally thought to be virus inclusion particles in epithelioid cells in lymph node granulomata of cat-scratch disease have been shown to be particles present in many other viral and nonviral granulomatous diseases.2 However, in order to start an investigation of this disease one might well assume the existence of an etio-logic agent and