All components of the "fluid tissue" blood should be considered as functional units. This seems familiar to physicians for the blood cells. For the dissolved plasmaproteins most doctors are, however, not yet used to think in functional terms. We propose a functional classification of well-characterized plasmaproteins, and a list of correlations between clinical conditions and protein changes. The clinician and the director of the laboratory have special knowledge and accordingly different tasks. An outline of the possible distribution of their duties and for their collaboration is given. We propose the stepwise evaluation of laboratory tests requested which in its details is obviously quite subjective and debatable. Accordingly, tests are categorized for routine, for special questions, and for scientific investigation. Finally, the clinician ought to know possible sources of error and mistake. For detailed advice, he may ask the director of the laboratory. If these principles--details may be modified--are respected, doctors should be able to obtain the laboratory data they need for diagnosis and treatment with minimal inconvenience to the patient and lowest cost to the community.