Analysis of Taxonomic Congruence among Morphological, Ecological, and Biogeographic Data Sets for the Leptopodomorpha (Hemiptera)

Abstract

Schuh, R. T., andJ. T. Polhemus (Department of Entomology, American Museum of Natural History, Central Park West at 79th Street, New York, New York 10024 and 3115 S. York, Englewood, Colorado 80110) 1980. Analysis of taxonomic congruence among morphological, ecological and biogeographic data sets for the Leptopodomorpha (Hemiptera). Syst. Zool. 29:1-26.-A data set of 47 morphological characters derived from the literature and original observation is prepared for the Leptopodomorpha. An argument is presented for considering the Leptopodomorpha as a monophyletic group and the Nepomorpha as its sister group. The data set is analyzed by cladistic (Wagner) and phenetic (UPGMA) methods in an effort to arrive at a most natural classification. Predictivity and stability are measured by comparing classifications based on a complete and partial taxon set as well as a random bipartition of the character set. Cladistic analysis produces more stable and predictive classifications with greater consensus and component information. The minimum length Wagner tree has greater information content than the UPGMA phenogram by virtue of its higher cophenetic correlation coefficient and its more parsimonious description of the character data. The Wagner tree is compared with four existing classifications for the Leptopodomorpha in the form of networks and trees. One of the published schemes is represented by the same network as the most parsimonious tree, but contains additional homoplasies as a rooted tree. The remaining published schemes are represented by different networks and describe the data less accurately (in more steps) than the most parsimonious tree. Wagner, UPGMA, and published results are examined for congruence with available ecological and distributional data. The minimum length Wagner tree and a published scheme of Polhemus show maximum congruence. A classification isomorphic with the Wagner tree is proposed as the one which should be used for future studies in the Leptopodomorpha. Objective criteria for evaluating classifications and properties of cladistic and phenetic methods are discussed. [Cladistics; Wagner; phenetics; UPGMA; classification; naturalness; prediction; stability; Leptopodomorpha.] Earlier claims by pheneticists (cf. Sokal and Sneath, 1963, and contained references) that their methods would make taxonomy logical and objective and by Hennig (1966) that the phylogenetic system of classification should serve as the general reference system for biology precipitated an extended debate about the goals and methods of biological classifi-cation. Much of the debate has not served to resolve the fundamental issue of What is the best classification for some or all organisms and how might it be determined? Rather, it has often concentrated on terminology without reference to concepts (cf. Mayr, 1974; Ashlock, 1979 [see Wiley, 1979]) or on motivational concerns as, for example, expressed by Bock (1977) or Ashlock (1979) who wish to make maximum use of evolutionary theory in biological classification, or Sokal and Sneath (1963) and Sneath and Sokal (1973) who prefer not to interpret classification in a phylogenetic sense (Farris, 1977). Some have wondered if we can or should arrive at a best classification (Frelin and Vuilleumier, 1979) and others have claimed simply that there is no best classification (Johnson, 1970; Key, 1974). Classifications are based on the analysis of data, which can be in the form of discrete characters or distance information. Quantitative techniques are available for analyzing both types of data (Farris, 1970, 1972; Sneath and Sokal, 1973). The interpretations we give to results are primarily related not to the way we col-