Any comprehensive survey of the principles, structure and methodology of subject analysis would require a treatise of great length to do even partial justice to the subject. Such is here neither contemplated nor intended. Scope has been restricted to problems concerning library cataloging through subject headings. Observations herein concern American practices in devising subject headings for alphabetic card catalogs, irrespective of whether the finished products are assembled as portions of dictionary catalogs or are filed to form the subject sections of “divided” catalogs. The form of card used is assumed to be the “unit” type wherein each card includes identical and reasonably full descriptive and informative data for the communication covered. The term general catalog is taken to mean one designed for many users, to be inclusive of one or selected disciplines intensively or covering portions of all fields of knowledge rather than to connote any particular degree of specificity or level of analysis for individual communications. A searching criticism is directed to current tendencies in subject heading construction which are presently denoted as “direct and specific” in the hope that the full import and significance of the duality concept will emerge. It is suggested that the duality concept is applicable to all forms of subject analysis for all kinds of communications. It is recommended that research and experimental investigations be directed to effecting a marriage of the best features of subject‐headed unit cards with those of marginal punched cards. The “Uniterm” specie of coordinate indexing is reviewed critically because of its purported challenge to the major modes of subject analysis.