Despite the overwhelming variety and abundance of invertebrate species the contribution of invertebrate studies to our understanding of the behavioral and physiological bases of learning has been minimal. Although anthropocentric biases may be responsible for the inordinate number of studies performed with the usual laboratory animals, clearly a vigorous extension of the comparative approach in the behavioral sciences is demanded. Several features of the comparative approach are outlined and suggestions are made which minimize any of its inherent difficulties. A multi-level and polythetic approach is proposed which considers multiple characteristics supplemented by evidence obtained at other levels to establish a meaningful behavioral taxonomy. Controlled systematic variation may be used to analyze the functional relations of the performance of different species on the same behavioral task. Another strategy compares simplified and complex versions of the same system in order to assess the quantitative and qualitative relationships between complexity and capacity. The structure of many invertebrate species renders them ideally suited to such analysis. Data on habituation obtained in Limulus polyphemus employing such a paradigm is discussed. The relationships between electrophysiological habituation and levels of neural complexity are clearly demonstrated. Future contributions from this preparation as well as from other invertebrate species are also discussed.