Gene trees: A powerful tool for exploring the evolutionary biology of species and speciation ALAN R. TEMPLETON,* STEPHANIE D. MASKAS† andMITCHELL B. CRUZAN‡ *Department of Biology, Campus Box 1137, Washington University, St Louis, MS 63130-4899, and Departments of †Botany and ‡Ecology and Evolutionary Biology, University of Tennessee, Knoxville, TN 37996, USA Abstract Evolutionary trees can be constructed from the haplotypes observed with molecular surveys of sequence or restriction site variation. Such gene trees can be constructed regardless of whether or not all of the individual specimens came from one or many species. Hence, these gene trees can straddle the species/population interface, thereby providing a powerful tool for studying the meaning of species and the process of specia- tion. We illustrate how historical approaches using gene trees can be used to separate the effects of population structure from population history, in order to rigorously test the species status of a group, and to test hypotheses about the process of speciation. A worked example of species status in the Piriqueta caroliniana complex is presented. Species status is evaluated under the cohesion species concept that defines a species as an evolutionary lineage with boundaries arising from the forces that create reproductive communities. Such forces are collectively called cohesion mechanisms and consist of two main sub- types: (i) genetic exchangeability, and (ii) ecological interchangeability. To make this def- inition operational, populations that behave as separate evolutionary lineages are first identified. A method is reviewed for inferring lineages using explicit statistical criteria from geographic overlays upon gene trees. Once lineages have been identified, the next step is to use the cohesion mechanisms to identify candidate traits that should contribute to genetic exchangeability and/or ecological interchangeability. The cohesion species are then identified by performing overlays upon gene trees in order to identify significant transitions in the candidate traits. Cohesion species are recognized only when statisti- cally significant reproductive/ecological transitions occur that are concordant with the lineages defined earlier. This data-rich method of recognizing species automatically gen- erates much information about the biogeography, population structure, historical events, and ecology and/or reproductive biology of the group under study. In turn, this infor- mation provides much insight into the process of speciation. It also makes the criteria, data, methods of analysis and degree of support for the species inference completely explicit, thereby avoiding confusion, inconsistency and artificial controversies that plague much of the literature on species concepts. Keywords: chloroplast DNA, coalescence, cohesion species concept, haplotype tree, phylogeog- raphy, Piriqueta caroliniana, speciation, species. Received 14 June 1999; accepted 20 August 1999 Plant Species Biology (2000) 15, 211–222 Introduction The concept of species is central to much of evolutionary biology, ecology and conservation biology. In practice, species are almost always defined by morphological criteria. Such a species criterion suffers from the deficien- cies exemplified by sibling species (two or more species identified by reproductive criteria, but that are morpho- logically indistinguishable) and polytypic species (a single reproductive community that has morphologically distinct subpopulatons). The dominance of morphol- ogical data stems from the fact that they are by far the most abundant, and often the only data available to © 2000 The Society for the Study of Species Biology Correspondence: Alan R. Templeton (Email: temple_a@biology.wustl.edu).