Immunosuppressive agents are not by and large a unique class of compounds, but are rather a variety of chemical and biological substances that share the common property of inhibiting the immune response. A knowledge of the biochemical locus of action of a new agent enables a reasonable prediction of its effectiveness as an immunosuppressant, particularly if it is known to inhibit DNA, RNA, or protein synthesis. Yet in light of the need for highly selective agents, capable of affecting only immunity to a defined antigen without general cytocidal effects, or capable of inhibiting only one element of the immune response, empiricism and the testing of a variety of different agents have a distinct place in addition to careful deductive reasoning. Biological agents, such as antibodies or antisera, that work against the types of cells involved in immunity, also require testing in a variety of immunological systems and should not be assumed to work in only one of them. Thus, antilymphocyte serum has been ascertained to inhibit cellular immunity while leaving antibody production relatively unaffected, a feature that was not readily predictable a priori.