The main features of the phylogeny program TNT are discussed. Windows versions have a menu interface, while Macintosh and Linux versions are command-driven. The program can analyze data sets with discrete (additive, non-additive, step-matrix) as well as continuous characters (evaluated with Farris optimization). Effective analysis of large data sets can be carried out in reasonable times, and a number of methods to help identifying wildcard taxa in the case of ambiguous data sets are implemented. A variety of methods for diagnosing trees and exploring character evolution is available in TNT, and publication-quality tree-diagrams can be saved as metafiles. Through the use of a number of native commands and a simple but powerful scripting language, TNT allows the user an enormous flexibility in phylogenetic analyses or simulations.

The Parsimony Ratchet 1 is presented as a new method performed. The number of characters to be sampled for for analysis of large data sets. The method can be easily reweighting in step 2 is determined by the user; I have implemented with existing phylogenetic software by genfound that between 5 and 25% of the characters provide erating batch command files. Such an approach has been good results in most cases. The performance of the implemented in the programs DADA (Nixon, 1998) and ratchet for large data sets is outstanding, and the results Winclada (Nixon, 1999). The Parsimony Ratchet has also of analyses of the 500 taxon seed plant rbcL data set been implemented in the most recent versions of NONA (Chase et al., 1993) are presented here. A separate analy-(Goloboff, 1998). These implementations of the ratchet sis of a three-gene data set for 567 taxa will be presented use the following steps: (1) Generate a starting tree (e.g., elsewhere (Soltis et al., in preparation) demonstrating a "Wagner" tree followed by some level of branch swapthe same extraordinary power. With the 500-taxon data ping or not). (2) Randomly select a subset of characters, set, shortest trees are typically found within 22 h (four each of which is given additional weight (e.g., add 1 to the runs of 200 iterations) on a 200-MHz Pentium Pro. These weight of each selected character). (3) Perform branch analyses indicate efficiency increases of 20؋-80؋ over swapping (e.g., "branch-breaking" or TBR) on the current "traditional methods" such as varying taxon order rantree using the reweighted matrix, keeping only one (or domly and holding few trees, followed by more complete few) tree. (4) Set all weights for the characters to the analyses of the best trees found, and thousands of times "original" weights (typically, equal weights). (5) Perform faster than nonstrategic searches with PAUP. Because the branch swapping (e.g., branch-breaking or TBR) on the ratchet samples many tree islands with fewer trees from current tree (from step 3) holding one (or few) tree. (6) each island, it provides much more accurate estimates Return to step 2. Steps 2-6 are considered to be one of the "true" consensus than collecting many trees from iteration, and typically, 50-200 or more iterations are few islands. With the ratchet, Goloboff's NONA, and existing computer hardware, data sets that were previously intractable or required months or years of analy-1 This method, the Parsimony Ratchet, was originally presented at sis with PAUP* can now be adequately analyzed in a few the Numerical Cladistics Symposium at the American Museum of hours or days.

The Parsimony Ratchet 1 is presented as a new method performed. The number of characters to be sampled for for analysis of large data sets. The method can be easily reweighting in step 2 is determined by the user; I have implemented with existing phylogenetic software by genfound that between 5 and 25% of the characters provide erating batch command files. Such an approach has been good results in most cases. The performance of the implemented in the programs DADA (Nixon, 1998) and ratchet for large data sets is outstanding, and the results Winclada (Nixon, 1999). The Parsimony Ratchet has also of analyses of the 500 taxon seed plant rbcL data set been implemented in the most recent versions of NONA (Chase et al., 1993) are presented here. A separate analy-(Goloboff, 1998). These implementations of the ratchet sis of a three-gene data set for 567 taxa will be presented use the following steps: (1) Generate a starting tree (e.g., elsewhere (Soltis et al., in preparation) demonstrating a "Wagner" tree followed by some level of branch swapthe same extraordinary power. With the 500-taxon data ping or not). (2) Randomly select a subset of characters, set, shortest trees are typically found within 22 h (four each of which is given additional weight (e.g., add 1 to the runs of 200 iterations) on a 200-MHz Pentium Pro. These weight of each selected character). (3) Perform branch analyses indicate efficiency increases of 20؋-80؋ over swapping (e.g., "branch-breaking" or TBR) on the current "traditional methods" such as varying taxon order rantree using the reweighted matrix, keeping only one (or domly and holding few trees, followed by more complete few) tree. (4) Set all weights for the characters to the analyses of the best trees found, and thousands of times "original" weights (typically, equal weights). (5) Perform faster than nonstrategic searches with PAUP. Because the branch swapping (e.g., branch-breaking or TBR) on the ratchet samples many tree islands with fewer trees from current tree (from step 3) holding one (or few) tree. (6) each island, it provides much more accurate estimates Return to step 2. Steps 2-6 are considered to be one of the "true" consensus than collecting many trees from iteration, and typically, 50-200 or more iterations are few islands. With the ratchet, Goloboff's NONA, and existing computer hardware, data sets that were previously intractable or required months or years of analy-1 This method, the Parsimony Ratchet, was originally presented at sis with PAUP* can now be adequately analyzed in a few the Numerical Cladistics Symposium at the American Museum of hours or days.

A phylogenetic analysis of a combined data set for 560 angiosperms and seven outgroups based on three genes, 18s rDNA (1855 bp), rbcl. (1428 bp), and atpB (1450 bp) representing a total of 4733 bp is presented. Parsimony analysis was expedited by use of a new computer program, the RATCHET. Parsimony jackknifing was performed to assess the support of clades. The combination of three data sets for numerous species has resulted in the most highly resolved and strongly supported topology yet obtained for angiosperms. In contrast to previous analyses based on single genes, much of the spine of the tree and most of the larger clades receive jackknife support 250%. Some of the noneudicots form a grade followed by a strongly supported eudicot clade. The early-branching angiosperms are Amborellaceae, Nymphaeaceae, and a clade of Austrobaileyaceae, Illiciaceae, and Schisandraceae. The remaining noneudicots, except Ceratophyllaceae, form a weakly supported core eumagnoliid clade comprising six well-supported subclades: Chloranthaceae, monocots, WinteraceaeICanellaceae, Piperales, Laurales, and Magnoliales. Ceratophyllaceae are sister to the eudicots. Within the well-supported eudicot clade, the early-diverging eudicots (e.g. Proteales, Ranunculales, Trochodendraceae, Sabiaceae) form a grade, followed by the core eudicots, the monophyly of which is also strongly supported. The core eudicots comprise * Correspondence to 0. E. Soltis. 0024-4074/00/080381+81 $35.00/0 38 1 Q 2000 The Linnean Society of London 382 D. E. SOLTIS ETAL.six well-supported subclades: (1) Berberidopsidaceae/Aextoxicaceae; (2) Myrothamnaceae/ Gunneraceae; (3) Saxifragales, which are the sister to Vitaceae (including Lea) plus a strongly supported eurosid clade; (4) Santalales; (5) Caryophyllales, to which Dilleniaceae are sister; and (6) an asterid clade. The relationships among these six subclades of core eudicots do not receive strong support. This large data set has also helped place a number of enigmatic angiosperm families, including Podostemaceae, Aphloiaceae, and Ixerbaceae. This analysis further illustrates the tractability of large data sets and supports a recent, phylogenetically based, ordinal-level reclassification of the angiosperms based largely, but not exclusively, on molecular (DNA sequence) data.

Abstract— The goal of a phylogenetic species concept is to reveal the smallest units that are analysable by cladistic methods and interpretable as the result of phylogenctic history. We define species as the smallest aggregation of populations (sexual) or lineagcs (asexual) diagnosable by a unique combination of character states in comparable individuals (semaphoronts). A character state is an inherited attribute distributed among all comparable individuals (semaphoronts) of the same historical population, clade, or terminal lineage. This definition of species is character‐based and pattern oriented. Evolutionary explanations of phylogenetic species are consistent with contemporary explanations of processes of speciation, but require only the assumption of nested hierarchical pattern. We discuss the compatibility of the phylogenetic species concept with various biological needs for species and justify its use at the exclusion of alternative species concepts.