The problem with which human ecology confronts medical librarians is that of inherent complexity combined with vagueness of definition and thought, associated with the risk of dispersing effort and funds ineffectually and too widely. Much of this paper is concerned with explaining why ecology must primarily be concerned with the "physiology of ecosystems," an ecosystem being an organized assemblage of populations of various living organisms plus all the nonliving components with which they interact or through which they exchange information. This is the broadest view of synecology. Much of the great complexity of an ecosystem is only apparent, because the integration of the components gives it statistical simplicities and, especially, some reliable homeostasis. In practice, fragmentary studies must be made, the lowliest of these being the autecology of an individual or a species such as man or the Norway rat. It is only in such studies that the term "environment" has meaning: the concept of environment disappears in the ecosystem, of which environments are simply component parts. For such fragmentary studies, it is recommended that the prefix "aut-" be retained, or that "bionomics" be used instead. Such studies should be made always with the intention of relating them to the ecosystem. An ecosystem becomes much more difficult to study if it is rapidly evolving. Ecosystems which include man as a dominant species are not only evolving, but contain whimsical and irrational human social elements lacking in animal groups. It should suffice to restrict the responsibilities of the medical librarian to catering for modern epidemiology (medical ecology). The range of disciplines involved is summarized and some practical hints added; the contributions of an active group (department, institute, or another) and cross-campus cooperation are stressed.