The term "species" has been a central tenet of biological belief since the early days of biology. But the concepts attached to the term have varied and often were not defined rigorously. The purpose of this paper is to investigate the biological species concept (BSC) : to consider its theoretical aspects, how one would actually delimit a biological species in nature, whether such species exist in nature, and whether the concept is of any unique value to the study of evolution. The classical definition of the BSC is partitioned into its essential components, and some of their aspects and problems are discussed. Three fundamental operations necessary for the delimitation of biological species in nature are described in detail. These are operations based on criteria of: (1) geographic contiguity, (2) phenetic similarity, and (3) interbreeding. Two properties of interbreeding, connectedness and success of reproduction, are defined and discussed. A flow chart for recognizing biological species is constructed from the definition as given by Mayr. Each step involves one of the three operations mentioned above. Reasons are given for including each step, as well as the inherent difficulties of each. It can be seen that most steps are either largely or entirely phenetic, even in theory. The necessary phenetic steps are termed "phenetic bottlenecks." To test the flow chart, we assume the unrealistic but optimal situation of total knowledge about the interbreeding relations among sampled organisms. The phenetic bottlenecks remain in this optimal case, and the degree of reliance on phenetic information for the delimitation of biological species increases as we depart from the optimal situation and make it more realistic. The BSC is found to be arbitrary (sensu Simpson) when attempts are made to apply it to actual data in nature, and not only because arbitrary phenetic decisions are a necessary part of the delimitation of biological species in nature. On asking some essential questions about the value of the BSC to taxonomy and evolution, we find that the BSC is not necessary for practical taxonomy, is neither necessary nor especially useful for evolutionary taxonomy, nor is it a unique or heuristic concept necessary for generating hypotheses in evolutionary theory. Most of the important evolutionary principles commonly associated with the BSC could just as easily be applied to localized biological populations, often resulting in deeper insight into evolution. Having decided that the BSC is neither operational nor heuristic nor of any practical value, we conclude that the phenetic species as normally described is the desirable species concept to be associated with the taxonomic category "species," and that the localized biological population may be the most useful unit for evolutionary study.