Until the 1950s, American plant ecology was dominated by the community-unit theory – that plants grow together in definite communities which constitute the proper subject matter for ecological research. Only H. A. Gleason proposed the alternative "individualistic hypothesis". In the 1950s the nature of the plant community was re-examined in a number of field studies. John Curtis led a re-assessment of ecological theory. This paper provides a historical analysis of aspects of his work. <br><br> Born in 1913, Curtis did his doctorate at the Univ. of Wisconsin, under Benjamin Duggar, receiving a fine training in physiological research. In 1941, he made a career shift toward community ecology. Dubious of the validity of the concept of the plant community, Curtis began an intensive investigation of the vegetation of Wisconsin. American ecology was in an insecure position, isolated from the mainstream of biological science. Curtis’s ambition was reform – to establish ecology as "a science rather than an art". The improvement of research methodology was a major concern. <br><br> Curtis and his colleagues found that the best way to arrange the data from their study stands was into a sequence of continuous variation, each dominant gradually peaking in frequency along a continuum. There were no distinct "associations" of species. By the 1970s, the continuum, which Curtis presented as a vindication of Gleason, was accepted as a generally valid description of mature vegetation.