Why do parasites harm their host? The persistence of this question in the history of the life sciences rests partly on a seeming biological paradox. In effect, although the annihilation of the host by a parasite could be depicted as a crude example of "the survival of the fittest," situations where the host dies rapidly often amount to a pyrrhic victory because parasitic colonies harboured within it may die as well before the transmission stage. So why would natural selection favour high virulence if this results in both the host's and the pathogen's deaths? From the last quarter of the nineteenth century to the late 1970s, it was often considered that, all else being equal, pathogen's evolution towards harmlessness was the expected outcome of long-term biological associations, as it would ensure the survival of both species. Frequently dismissed today as naively adaptationist, this perspective was however widely defended by some of those who pioneered an ecological, even evolutionary view of disease in the early-twentieth century, and was often ascribed to American bacteriologist and comparative pathologist Theobald Smith (1859-1934). Since the early 1980s, the mathematical model of the trade-off--based on the idea that pathogens face several compromises between their mode of transmission, the level of virulence, and the cost of resistance--has challenged the assumed tendency towards harmlessness and has gained currency. This paper first analyzes Theobald Smith's conception of disease, his experimental work on Texas cattle fever, and his formulation of the "law of declining virulence." The following sections then trace the legacy of this model of disease evolution from circa 1900 to its widespread acceptance in the mid-twentieth century and until its downfall in the 1980s. Particular attention is given to the case of the myxoma virus and how it acted as an empirical confirmation of Smith's model in the 1950s. Finally, the paper examines both significant empirical and theoretical challenges to the avirulence model. The present study not only fills a gap in the history of disease transmission and ecology but also sheds light on the intermingled relationship between bacteriology, evolutionary biology, and public health in the past century.