Syst. Biol. 51(6) 2002

Mooers and Heard
Abstract.—Symposium Introduction.

Savolainen et al.
Abstract.—While the heritability of speciation rates and extinction risks are crucial parameters in models of macroevolution, little direct evidence has been available to assess the occurrence, strength, or generality of this heritability. We test for heritability using correlations between ancestral and descendent branch lengths in phylogenetic trees, an approach first applied to a bird phylogeny by Harvey and co-workers (Harvey et al., 1991 in Genes in Ecology. Blackwell Scientific, Oxford). We apply Harvey et al's test to some of the largest DNA sequence-based phylogenetic analyses published to date for plants, insects, fungi and bacteria: if one of two parent lineages splits first, and if this is the case for any heritable reason, then, on average we expect its daughter lineages to also split first. We also used a randomization procedure to assess significance of branch length heritability. Using maximum parsimony and maximum likelihood branch lengths and trees made ultra-metric after either non-parametric rate smoothing or by enforcing a molecular clock, we found a pattern for most clades consistent with heritable net cladogenesis. Heritability of cladogenesis may be a general phenomenon, detectable across a large number of lineages and a broad range of taxa.

Purvis and Agapow
Abstract.—We present two lines of evidence indicating that the degree of symmetry in phylogenetic topologies differs at different hierarchical levels. First, in a set of 61 phylogenies with superspecific taxa as their terminals, trees were on average more unbalanced (asymmetric) when the species-richnesses of terminals were considered than when they were not. Second, nodes with a given number of higher taxa descended from them were on average more unbalanced than nodes with the same number of species as descendants. We discuss possible reasons - some biological, some artefactual - for the differences. Whatever the reason, our results caution against treating species-level and higher-level phylogenies as equivalent when considering tree shape. The imbalance measure adopted here permits the use of trees that contain polytomies, facilitating a larger sample than has been achieved previously.

Chan and Moore
Abstract.—Prolific cladogenesis, adaptive radiation, species selection, key innovations, and mass extinctions are a few examples of biological phenomena that lead to differential diversification among lineages. Central to the study of differential diversification rates is the ability to distinguish chance variation from that which requires deterministic explanation. In order to detect diversification rate variation among lineages, we propose a number of methods that incorporate information on the topological distribution of species diversity from all internal nodes of a phylogenetic tree. These whole-tree methods (MII, ME, and MR) are explicitly connected to a null model of random diversification—the equal-rates Markov (ERM) random branching model—and an alternative model of differential diversification: MII is based on the product of individual nodal ERM probabilities; ME on the sum of individual nodal ERM probabilities, and MR on a transformation of ERM probabilities that corresponds to a formalized system that orders trees by their relative symmetry. These methods have been implemented in a freely available computer program, SymmeTREE, to detect clades with variable diversification rates and thereby allow the study of biological processes correlated with and possibly causal to diversification. Application of these methods to several published phylogenies demonstrates their ability to contend with relatively large, incompletely resolved trees. These topology-based methods do not require estimates of relative branch lengths, which should facilitate the analysis of phylogenies for which such data are unreliable or unavailable, such as supertrees.

Agapow and Purvis
Abstract.—We use simulations to compare the relative power of eight statistical tests to detect imbalance in phylogenies that is too great to be ascribed to an equal-rates Markov null model. Three of these tests have never had their power assessed before. Our simulations are the first to assess performance under scenarios in which lineages' speciation rates can evolve independently. In one of the scenarios explored, rates depend upon the value of an evolving trait, whereas in the other the probability that a species will speciate declines with the time since it last did so. The results indicate that the relative performance of the methods depends upon how the imbalance is generated: different types of process lead to different imbalance signatures - different patterns of imbalance at different depths in phylogeny - and the measures of tree shape differ in the depth of phylogeny at which they are most sensitive. Relative performance is also affected by tree size, but does not appear to depend greatly upon the degree of speciation rate variation among lineages. Two of the indices (Colless' index Ic and Shao and Sokal's Nbar) show reasonable performance throughout, while another (Shao and Sokal's B2) is never indicated to be a preferred method. Two tests that do not require completely resolved phylogenies - mean Ic and mean Ic10 - have reasonable power.

Martins and Housworth
Abstract.—This study explores the impact of phylogeny shape on the results of interspecific statistical analyses incorporating phylogenetic information. In most phylogenetic comparative methods (PCMs), the phylogeny can be represented as a relationship matrix, and the hierarchical nature of interspecific phylogenies translates into a distinctive block-like matrix that can be described by its eigenvectors (topology) and eigenvalues (branch lengths). Thus, differences in the eigenvectors and eigenvalues of different relationship matrices can be used to gauge the impact of possible phylogeny errors by comparing the actual phylogeny used in a PCM analysis with a second phylogenetic hypothesis that may be more accurate. For example, we can use the sum of inverse eigenvalues as a rough index to compare the impact of phylogenies with different branch lengths. Topological differences are better described by the eigenvectors. In general, phylogeny errors that involve deep splits in the phylogeny (e.g., moving a taxon across the base of the phylogeny) are likely to have much greater impact than will those involving small perturbations in the fine structure near the tips. Small perturbations, however, may have more of an impact if the phylogeny structure is highly dependent (with many recent splits near the tips of the tree). Unfortunately, the impact of any phylogeny difference on the results of a PCM depends on the details of the data being considered. Recommendations regarding the choice, design and statistical power of interspecific analyses are also made.

Pybus et al.
Abstract.—The relative positions of branching events in a phylogeny contain information about evolutionary and population dynamic processes. We provide new summary statistics of branching event times and describe how these statistics can be used to infer rates of species diversification from inter-species trees, or rates of population growth from intra-species trees. We also introduce a phylogenetic method for estimating the level of taxon sampling in a clade. Different evolutionary models and different sampling regimes can produce similar patterns of branching events, so it is important to consider explicitly the model assumptions involved when making evolutionary inferences. The phylogeny of the mosquito-borne flaviviruses serves as an illustrative case study, and the results suggest that there could be several thousand currently unidentified viruses in this clade.

Heard and Mooers
Abstract.—Current models of diversification with evolving speciation rates have trouble mimicking the extreme imbalance seen in estimated phylogenies. These models have not incorporated extinction. Here we report on a simple simulation model that includes heritable and evolving speciation rates coupled with mass extinctions. We show that random (but not selective) mass extinctions, coupled with evolving among-lineage variation in speciation rates, increase imbalance of post-recovery clades. These results identify random mass extinctions as plausible contributors to the imbalance of modern clades. Paleontological evidence suggests that mass extinctions are often random with respect to ecological and morphological traits, consistent with our simulations. In contrast, evidence that the current anthropogenic mass extinction is phylogenetically selective suggests the current extinction episode may be qualitatively different from past ones in the way it reshapes future biotas.

Webb and Pitman
Abstract.—The frequency distribution of numbers of species in taxonomic groups, where many species belong to a few very diverse higher taxa, is mirrored by that of the abundances of species in most communities, where many individuals belong to a few very abundant species. Various hypotheses mechanistically link a species' community abundance with the diversity of the higher-level taxon (genus, family, order) to which it belongs, but empirical data are equivocal about general trends in the relation between rank-taxon diversity and mean abundance. One reason for this inconclusive result may be the effect of the semi-subjective nature of rank-based classification. We assessed the relationship between clade diversity and mean species abundance for two diverse tropical tree communities, using both traditional, rank-based analysis, and two new phylogenetic analyses (based on the ratio of individuals to taxa at each node in the phylogeny). Both rank-based and phylogenetic analyses using taxonomic ranks above the species level as terminal taxa detected a trend associating common species with species-rich families. In contrast, phylogenetic analyses using species as terminal taxa could not distinguish the observed distribution of species abundances from a random distribution with respect to clade diversity. The difference between these results might be due to: i) the absence of a real phylogeny-wide relationship between clade abundance and diversity, ii) the influence of poor phylogenetic resolution within families in our phylogenies, or iii) insufficient sensitivity of our metrics to subtle tree-wide effects. Further development and application of phylogeny-based methods for testing abundance-diversity relationships is urged.

Harcourt-Brown
Abstract.—Studies of phylogenetic tree shape often concentrate on the balance of phylogenies of extant taxa. Paleontological phylogenies (that include extinct taxa) can contain additional useful information, and can directly document changes in tree shape through evolutionary time. Unfortunately, the inclusion of extinct taxa lowers the power of direct examinations of tree balance, because it increases the range of tree shapes expected under null models of evolution (with equal rates of speciation and extinction across lineages). A promising approach for the analysis of tree shape in paleontological phylogenies is to break the phylogeny down into time slices, examining the shape of the phylogeny of taxa alive at each time slice and changes in that shape between successive time slices. I illustrate this method using 57 time slices through a stratophenetic phylogeny of the Cretaceous planktonic foraminiferal superfamily Globotruncanacea. At three of 56 intervals between time slices (93 - 92.5 Ma, 89 - 88.5 Ma and 85.5 - 84 Ma), the group shows steep increases in imbalance. While none of these increases is statistically significant after Bonferroni correction, these points in the history of the Globotruncanacea are nevertheless identified as deserving of further macroevolutionary investigation. Interestingly, the 84 Ma timeslice coincides with a peak in species turnover for the superfamily. Time slices through phylogenies may prove useful for identifying periods of time when evolution was proceeding in a non-stochastic manner.

de Queiroz
Abstract.—Key innovations have often been invoked to explain the exceptional diversification of particular groups. However, there are few convincing examples of traits that are repeatedly and consistently associated with increased diversification. The paucity of such cases may reflect the contingent nature of the diversifying effect of key traits. These contingencies can be viewed as statistical interactions between the trait and at least three kinds of factors: (1) other taxa; (2) other traits of the group itself; (3) the physical environment. I describe tentative examples in each of these categories: (1) a dampening of the diversification of clades with image-forming eyes by groups that earlier evolved such eyes; (2) an effect of growth form (woody or herbaceous) on the diversifying effect of biotic seed dispersal in angiosperms; (3) an effect of atmospheric CO-2 level on the diversifying effect of C4 photosynthesis in monocots. These examples suggest the need for more complex analyses of the relationship between possible key traits and diversification. They also suggest that radiations may be predictable given certain circumstances, thus supporting a view of evolution as both predictable and contingent. I argue that, ironically, a certain degree of predictability may be critical to arguments for evolutionary contingency.

Drovetski
Abstract.—The phylogeny of grouse (Aves: Tetraoninae) was reconstructed using four non-coding loci: two were W-linked, one was autosomal, and one was the mitochondrial control region (CR). The rapidly evolving CR provided resolution throughout the tree, whereas the slowly evolving nuclear loci failed to resolve deeper nodes. The tree based on all four loci combined was almost identical to the CR tree and did not improve resolution or bootstrap support. The stemminess and imbalance of the trees were good determinants of the quality of the phylogenetic signal. The skewness of the tree score distribution (g1) behaved contrary to prediction; loci that had a more symmetric tree score distribution produced trees that had greater stemminess and balance. The quality of the phylogenetic signal was related to the evolutionary rate. Four clades of grouse were discovered. Two of these clades corresponded to currently recognized genera Bonasa and Lagopus. Bonasa was the sister to other grouse and Lagopus was the sister to the other two non-Bonasa clades. The third clade included Falcipennis, Tetrao, and Lyrurus. The fourth clade included the genera Centrocercus, Dendragapus, and Tympanuchus. The data supported recognition of Falcipennis canadensis franklinii and Dendragapus obscurus fuliginosus as species.

Pereira et al.
Abstract.—The Cracidae is one of the most endangered and distinctive bird families in the Neotropics, yet the higher relationships among taxa remain uncertain. Here the molecular phylogeny of its eleven genera is inferred using 10,678 analyzable sites (5,412 from seven different mitochondrial segments and 5,266 sites from four nuclear genes). We performed combinability tests to check conflicts in phylogenetic signals of separate genes and genomes. Phylogenetic analysis showed that the unrooted tree ((curassows, horned guan),(guans, chachalacas)) was favored by most data partitions, and that different data partitions provided support for different parts of the tree. In particular, the concatenated mtDNA genes resolved shallower nodes, whereas the combined nuclear sequences resolved the basal connections among the major clades of curassows, horned guan, chachalacas and guans. Therefore we decided that for the Cracidae all data should be combined for phylogenetic analysis. Maximum parsimony, maximum likelihood and Bayesian analyses of this large data set produced similar trees. The maximum parsimony tree indicated that guans are the sister group to (horned guan, (curassows, chachalacas)), whereas the maximum likelihood and Bayesian analysis recovered a tree where the horned guan is a sister clade to curassows, and these two clades had the chachalacas as sister group. Parametric bootstrapping showed that alternative trees previously proposed for the cracid genera are significantly less likely than our estimate of their relationships. A likelihood ratio test of the hypothesis of a molecular clock for cracid mtDNA sequences using the optimal ML topology did not reject rate constancy of substitutions through time. We estimated cracids to have originated between 64 and 90 MYA, with a mean estimate of 76 MYA. Diversification of the genera occurred approximately 41 to 3 MYA, corresponding with periods of global climate change and other Earth history events that likely promoted divergences of higher level taxa.

Bertelli et al.
Abstract.—A cladistic analysis of the tinamous, including the 47 currently recognized species and some distinct subspecies, scored for 80 integumentary characters from adult and natal plumage, ramphoteca (corneum sheath of bill), and podoteca (horny scales of legs) is presented. For the adult plumage (50 characters), we studied feather pigmentation patterns from different pterylae (feather tracts); a criterion of overlap of basic pigmentation elements was used to assign costs to the transformation between the states in most of these characters, in such a way that transformations between more similar conditions were less costly. The consensus was almost fully resolved, and about 50% of its groups were relatively well supported. Since the only outgroup that could be used provided a poor root, two possible rootings of the ingroup subtree were considered; in either case, only one of the two traditional subfamilies (the steppe tinamous) was recovered, and the other (the forest tinamous) appeared as paraphyletic. The results of the present analysis are compared to those from an osteological data set, using a strict supertree technique. The combined tree has a large number of nodes, indicating a high degree of congruence between the two data sets.