Different approaches to circumscribe staminodial structures in the angiosperms are reviewed. The need for a morphological distinction between "true staminodes" (derived from stamens or homologous to stamens) and "pseudostaminodes" (nonhomologous to stamens) is emphasized. In phylogenetic studies the term "staminode" is often used uncritically, without knowledge of the true homology of these structures. Staminodes are either whole organs (outer tiers or whorls, namely petals, intermediate tiers, or organs within a tier), or partial organs. This article aims to discuss the shortcomings of the past and current approach of staminodes and proposes definitions of staminode types for use as characters in phylogenetic analyses. Staminodial structures should be classified according to their position and function in the flower. Both aspects are intricately linked and make the identification of staminodes sometimes problematic. Shifts in time (heterochrony) and space (heterotopy or homeosis) make that a regressing organ either aborts completely or becomes remodeled into something new. Petals are included in the definition of staminodes as they combine function and heterotopy. A hierarchical ordering of staminodial types is given and discussed. Three interdependent but possibly complementary functions are attached to the occurrence of staminodes: an attractive, nutritional, and structural function. The importance of staminodes for the evolution of the androecium and flower is demonstrated. The difficulty in unmasking pseudostaminodes, comprising receptacular disks, is demonstrated. The value and shortcomings of molecular-based interpretations of staminodes are discussed. It is shown that the decision to recognize a staminode from receptacular emergences often relies on unstable grounds and remains largely dependent on the acceptance of a given phylogenetic background.