Syst. Biol. 52(1) 2003

**Schwarz et al.**

*Abstract.*—Allodapine bees have long been regarded as providing useful material for examining the origins of social behaviour, and previous studies have assumed that sociality arose within the Allodapini, and linked the evolution of sociality to a transition from mass provisioning to progressive provisioning of brood. Early phylogenetic studies of allodapines were based on morphological and life-history data, but critical aspects of these studies relied on small character sets where the polarity and coding of characters is problematic. We used nucleotide sequence data from one nuclear and two mitochondrial gene fragments to examine phylogenetic structure among nine allodapine genera. Our data set comprised 1506 nucleotide positions of which 402 were parsimony-informative. Maximum parsimony, log determinant, and maximum likelihood analyses produced highly similar phylogenetic topologies, and all analyses indicated that the tropical African genus *Macrogalea* was the sister group to all other allodapines. This finding conflicts with previous studies which suggest that *Compsomelissa + Halterapis* form the most basal group. Changing the basal node of the Allodapini has major consequences for understanding evolution in this tribe. Our results caste doubt on the previous hypotheses that progressive provisioning and caste-like social behaviour evolved among lineages leading to the extant allodapine taxa. Instead, our results suggest that mass provisioning in *Halterapis* is a derived feature and that social behavior is an ancestral trait for all allodapine lineages. The forms of social behaviour present in extant allodapines are likely to have resulted from a long evolutionary history and this may help explain the complexity of social traits found in many allodapine bees.

**Lee and Hugall**

*Abstract.*—In simultaneous analyses of multiple data partitions, the trees relevant when measuring support for a clade are the optimal tree, and the best tree lacking the clade (i.e., the most reasonable alternative). The parsimony-based method of partitioned branch support (PBS) "forces" each data set to arbitrate between the two relevant trees. This value is the amount each data set contributes to clade support in the combined analysis, and can be very different from support apparent in separate analyses. The approach used in PBS can also be employed in likelihood: a simultaneous analysis of all data retrieves the maximum likelihood tree, and the best tree without the clade of interest is also found. Each data set is fitted to the two trees and the log-likelihood difference calculated, giving "partitioned likelihood support" (PLS) for each data set. These calculations can be performed regardless of the complexity of the ML model adopted. The significance of PLS can be evaluated using a variety of resampling methods, such as the Kishino-Hasegawa test, the Shimodiara-Hasegawa test, or likelihood weights, although the appropriateness and assumptions of these tests remains debated.

**Danforth et al.**

*Abstract.*—We performed a phylogenetic analysis of the species, species groups, and subgenera within the predominantly eusocial lineage of *Lasioglossum* (the *Hemihalictus* series) based on three protein coding genes (one mitochondrial [COI] and two nuclear [EF-1_ and LW opsin]). The entire data set consisted of 3421 aligned nucleotide sites, 854 of which were parsimony informative. Analyses by equal weights parsimony, maximum likelihood, and Bayesian methods yielded good resolution among the 53 taxa/populations with strong bootstrap support and high posterior probabilities for most nodes. There was no significant incongruence among genes, and parsimony, ML, and Bayesian methods yielded congruent results. Using the resulting cladograms we mapped social behavior onto the tree for 42 of the taxa/populations in order to infer the likely history of social evolution within *Lasioglossum*. Our results indicate that eusociality has had a single origin within *Lasioglossum*. Within the predominantly eusocial clade, however, there have been multiple reversals from eusociality to solitary nesting, social polymorphism, or social parasitism. We count 6 losses of eusociality within the predominantly eusocial clade, suggesting that reversals from eusociality to solitary nesting, social parasitism, or social polymorphism may be more common in primitively eusocial Hymenoptera than previously anticipated. Our results support the view that eusociality is hard to evolve but is easily lost. This is potentially important for understanding the early evolution of the advanced eusocial insects, such as ants, termites and corbiculate bees.

**Johnson et al.**

*Abstract.*—Cospeciation generally increases the similarity between host and parasite phylogenies. Incongruence between host and parasite phylogenies has previously been explained in terms of host-switching, sorting, and duplication events. Here we describe an additional process, failure of the parasite to speciate in response to host speciation, which may be important in some host-parasite systems. "Failure to speciate" is likely to occur when gene flow among parasite populations is much higher than that of their hosts. We reconstructed trees from mitochondrial and nuclear DNA sequences for pigeons and doves (Aves: Columbiformes) and their feather lice in the genus *Columbicola* (Insecta: Phthiraptera). While comparisons of the trees from each group revealed a significant amount of cospeciation, there was also a significant degree of incongruence. Cophylogenetic analyses generally indicated that host-switching may be an important process in the history of this host-parasite association. Using terminal sister taxon comparisons, we also identified three apparent cases where the host has speciated, while the associated parasite has not. In two of these cases of "failure to speciate," these comparisons involve allopatric sister taxa of hosts whose lice also occur on hosts sympatric with both of the allopatric sisters. These additional hosts for generalist lice may promote gene flow between lice on the allopatric sister species. Relative rate comparisons for the mitochondrial COI gene indicate that molecular substitution occurs about 11 times faster in lice than in their avian hosts.

**Suchard et al.**

*Abstract.*—We propose a Bayesian method for testing molecular clock hypotheses for use with aligned sequence data from multiple taxa. Our method utilizes a non-reversible nucleotide substitution model to avoid the necessity of specifying either a known tree relating the taxa or an outgroup for rooting the tree. We employ reversible jump Markov chain Monte Carlo to sample from the posterior distribution of the phylogenetic model parameters and conduct hypothesis testing using Bayes factors, the ratio of the posterior to prior odds of competing models. Here, the Bayes factors reflect the relative support of the sequence data for equal rates of evolutionary change between taxa versus unequal rates, averaged over all possible phylogenetic parameters, including the tree and root position. As the molecular clock model is a restriction of the more general unequal rates model, we use the Savage-Dickey ratio to estimate the Bayes factors. The Savage-Dickey ratio provides a convenient approach to calculating Bayes factors in favor of sharp hypotheses. Critical to calculating the Savage-Dickey ratio is a determination of the prior induced on the modeling restrictions. We demonstrate our method on a well-studied mtDNA sequence dataset consisting of nine primates. We find strong support against a global molecular clock, but do find support for a local clock among the anthropoids. We provide mathematical derivations of the induced priors on branch length restrictions assuming equally likely trees. These derivations also have more general applicability to the examination of prior assumptions in Bayesian phylogenetics.

**Miller**

*Abstract.*—Systematists expect their hypotheses to be asymptotically precise. As the number of phylogenetically informative characters for a set of taxa increases, the relationships implied should stabilize on some topology. If true, this increasing stability should clearly manifest itself if an index of congruence is plotted against the accumulating number of characters. Continuous Jackknife Function (CJF) analysis is a new graphical method that portrays the extent to which available data converge on a specified phylogenetic hypothesis, the reference tree. The method removes characters with increasing probability, analyzes the rarefied data matrices phylogenetically, and scores the clades shared between each of the resulting trees and the reference tree. As more characters are removed, the number of shared clades must decrease, but the rate of decrease will depend on how decisively the data support the reference tree. Curves for stable phylogenies are clearly asymptotic with nearly 100% congruence for a substantial part of the curve. Less stable phylogenies lose congruent nodes quickly as characters are excluded, resulting in a more linear or even a sigmoidal relationship. Curves can be interpreted as predictors of whether the addition of new data of the same type is likely to alter the hypothesis under test. Continuous Jackknife Function analysis makes statistical assumptions about the collection of character data. To the extent that CJF curves are sensitive to violations of unbiased character collection, they will be misleading as predictors. Convergence of data on a reference tree does not guarantee historical accuracy, but it does predict that the accumulation of further data under the sampling model will not lead to rapid changes in the hypothesis.

**Rohlf and Bookstein**

*Abstract.*—Any change in shape of a configuration of landmark points in two or three dimensions includes a uniform component, a component that is a wholly linear (affine) transformation. The formulas for estimating this component have been standardized for 2D data but not for 3D data. We suggest estimating the component by way of the complementarity between the uniform component and the space of partial warps. The component can be estimated by regression in either one space or the other: regression on the partial warps, followed by their removal, or regression on a basis for the uniform component itself. Either of the new methods can be used for both two and three-dimensional landmark data and thus generalizes Bookstein's (1996) linearized Procrustes formula for estimating the uniform component in 2D.

**Hormiga et al.**

*Abstract.*—Spiders of the recently described linyphiid genus *Orsonwelles* (Araneae, Linyphiidae) are one of the most conspicuous groups of terrestrial arthropods of Hawaiian native forests. There are thirteen known *Orsonwelles* species and all are single island endemics. This radiation provides an excellent example of insular gigantism. We have reconstructed the cladistic relationships of *Orsonwelles* species using a combination of morphological and molecular characters (both mitochondrial and nuclear sequences) under a parsimony framework. We explored and quantified the contribution of different character partitions and their sensitivity to changes in the traditional parameters (gap, transition and transversion costs). The character data show a strong phylogenetic signal, robust to parameter changes. The monophyly of the genus *Orsonwelles* is strongly supported. The parsimony analysis of all character evidence combined recovers a clade with of all the non-Kauai *Orsonwelles* species; the species from Kauai form a paraphyletic assemblage with respect to the latter former clade. The biogeographic pattern of the Hawaiian *Orsonwelles* species is consistent with colonization by island progression but alternative explanations for our data exist. Although the geographic origin of the radiation remains unknown, it appears that the ancestral colonizing species arrived first on Kauai (or to an older island). The ambiguity in the area cladogram (namely, post-Oahu colonization) is not derived from conflicting or unresolved phylogenetic signal among *Orsonwelles* species, but rather from the number of taxa in the youngest islands. Speciation in *Orsonwelles* occurs within islands more often (eight out of the twelve cladogenetic events) than between islands. A molecular clock was rejected for the sequence data. Divergence times were estimated by using the nonparametric rate smoothing method of Sanderson (1997) and the available geological data for calibration. The results suggest that the oldest divergences of *Orsonwelles* spiders (on Kauai) go back to about four million years.

**Jordan et al.**

*Abstract.*—Damselflies of the endemic Hawaiian genus *Megalagrion* have radiated into a wide variety of habitats and are an excellent system for the study of adaptive radiation. Past phylogenetic analysis based on morphological characters has been difficult. Here, we examine relationships among 56 individuals from 20 of the 23 described species using ML and Bayesian phylogenetic analysis of mitochondrial (1287 bp) and nuclear (1039 bp) DNA sequence data sets. Models of evolution were chosen using the Akaike Information Criterion. Problems with distant outgroups were accommodated by constraining the best ML ingroup topology but allowing the outgroups to attach to any ingroup branch in a bootstrap analysis. No strong contradictions are present between either data partition and the combined data set. Areas of disagreement are mainly confined to clades that are strongly supported by the mtDNA and weakly supported by the EF-1_ data due to lack of changes. However, the combined analysis results in a unique tree. Correlation between Bayesian posterior probabilities and bootstrap percentages decreases in concert with decreasing information in the data partitions. In cases where nodes are supported by single characters bootstrap proportions are dramatically reduced compared to posterior probabilities. Two speciation patterns are evident from the phylogenetic analysis. First, most speciation is inter-island, and occurs as members of established ecological guilds colonize new volcanoes after they emerge from the sea. Second, there are several instances of rapid radiation into a variety of specialized habitats, in one case entirely within the island of Kauai. Application of a local clock procedure to the mitochondrial DNA topology suggests that two of these radiations correspond to the development of habitat on the islands of Kauai and Oahu. About 4.0 Ma species simultaneously moved into fast streams and plant leaf axils on Kauai, and about 1.5 My later another group moved simultaneously to seeps and terrestrial habitats on Oahu. Results from the local clock analysis also strongly suggest that *Megalagrion* arrived in Hawaii about 10 Ma, well before the emergence of Kauai. Interestingly, date estimates were more sensitive to the particular node that was fixed in time than to the model of local branch evolution used. We propose a general model for the development of endemic damselfly species on Hawaiian Islands and document five potential cases of hybridization (*M. xanthomelas x M. pacificum, M. eudytum x M. vagabundum, M. orobates x M. oresitrophum, M. nesiotes x M. oahuense,* and *M. mauka x M. paludicola*).

**Gascuel et al.**

*Abstract.*—We develop a recurrence relation that counts the number of Tandem Duplication Trees (either rooted or unrooted) that are consistent with a set of n tandemly repeated sequences generated under the standard unequal recombination (or crossover) model of tandem duplications. We find that the number of rooted duplication trees is exactly twice the number of unrooted trees, which means, on average, only two positions for a root on a duplication tree are possible. Using the recurrence we can tabulate these numbers for small values of n. Further we develop an asymptotic formula, that for large n, provides estimates for these numbers. These numbers give a priori probabilities for phylogenies of the repeated sequences to be duplication trees. This extends earlier studies where exhaustive counts of the numbers for small n were obtained. One application showed the significance of finding that most maximum parsimony trees constructed from repeat sequences from Human immunoglobins and T-cell receptors were tandem duplication trees. Those findings provided strong support to the proposed mechanisms of tandem gene duplication. The recurrence relation also suggests effcient algorithms to recognize duplication trees and to generate random duplication trees for simulation. A linear-time recognition algorithm is detailed.