NEWS

		CHANGES IN APE VERSION 5.8


NEW FEATURES

    o dist.topo() has a new option 'mc.cores'. With the default PH85
      distance, the gain of using several cores is visible with 200+
      trees. The internal code for the score distance has been
      improved and is twice faster with a single core, and 4-5 times
      faster with mc.cores = 2.

    o keep.tip() gains a '...' argument which is passed to drop.tip().
      The behaviour of the latter is changed so that node labels, if
      available, are used to create new tip labels when subtree = TRUE
      (see PR #90 on GitHub).

    o nj() can now handle distances stored in long vectors. There was
      previously an implicit limit to 46,341 observations (see PR #97
      on GitHub).

    o del.gaps(), del.colgapsonly(), and del.rowgapsonly() now accepts
      objects of class "AAbin". Besides, the class of the input object
      is always respected (it used to be "DNAbin" only even when it
      had several elements).

    o makeNodeLabel() is now generic with methods for the "phylo" and
      "multiPhylo" classes.

    o unroot() gains two options, 'collapse.singles' and
      'keep.root.edge', to handle some complicated situations (see
      issue #103 on GitHub).

    o read.GenBank() has a new argument 'type' to download AA
      sequences. The default is type = "DNA", and the other choice is
      type = "AA".


BUG FIXES

    o rtt() was fixed when dates include NAs (issue #73 on GitHub).

    o njs(), bionj(), bionjs(), fastme.bal(), and fastme.ols() now
      check that there are at least three observations in the distance
      matrix (issue #84 on GitHub).

    o nodelabels(, frame = "circle") used to ignore the argument
      'adj'. This is fixed (issue #85 on GitHub).

    o write.tree() ignored tree names. This bug was introduced in ape
      5.7 (see issue #101 on GitHub).

    o read.tree() now reads correctly trees with many nodes but a
      single tip such as "(());" (see issue #104 on GitHub).

    o bitsplits() failed with large trees. There are now checks on
      memory requirements; some explanations have been added in the
      help page.

    o comparePhylo() failed with two large trees. This is fixed.


OTHER CHANGES

    o .uncompressTipLabel() now returns an object of the same class
      than the input object (previously the returned object was always
      of class "multiPhylo").

    o The help page of is.binary() has been clarified that branch
      lengths are ignored (issue #83 on GitHub).

    o dist.nodes() has an improved internal C code with unlimited tree
      size. The code is three times faster with small trees; twice
      faster with ~1000 tips (the gain is less with bigger trees).

    o ace() now uses node labels, if available, in the output.

    o comparePhylo() now restores the graphical window in its previous
      state (i.e., after calling layout() if plot = TRUE).

    o read.tree() now gives a warning (still returning NULL) if there
      is (are) no semicolon(s) [end of tree(s)] in the Newick string.

    o write.tree() now gives a warning if the length of the node
      labels does not match the number of nodes given by '$Nnode'.



		CHANGES IN APE VERSION 5.7-1


BUG FIXES

    o Some very unbalanced trees resulted in a segfault (see issue #70
      on GitHub).


OTHER CHANGES

    o ape 5.7 introduced an implicit dependence on R >= 4.1. This has
      been corrected and this version should work with R >= 3.4 (and
      certainly with R >= 3.2).

    o All examples should now write files in the TMP directory (some
      still used to do this in the user's working directory).



		CHANGES IN APE VERSION 5.7


NEW FEATURES

    o The new function muscle5 calls MUSCLE5. There are two other new
      functions that depend on this program, efastats and letterconf,
      to compute summaries for a set of alternative alignments and
      display the confidence in each site in HTML, respectively.

    o print.LargeNumber() gains the options 'latex' (FALSE by default)
      and 'digits' (depends on 'latex').

    o drop.tip() and keep.tip() are now generic functions with methods
      for the classes "phylo" and "multiPhylo".


BUG FIXES

    o A bug was fixed in multi2di() and di2multi() for compressed
      "multiPhylo" objects causing an error if the object had to be
      reordered internally (see PR #51 on GitHub).

    o unroot() now accepts different "order" of the "phylo" object
      (see PR #52 on GitHub).

    o write.tree(), ladderize(), and balance() failed with very
      unbalanced trees (issues #53 and #54 on GitHub).

    o PGLS models used a wrong ordering of data if 'species' was a
      factor in corBrownian(, form = ~species) or other correlation
      structure functions.

    o A bug was fixed in the calculations of the variance of the TN93
      distance (dist.dna).

    o consensus() did not use postprocess.prop.part() and might
      scramble nodelabels (see PR #65 on GitHub).

    o is.binary.phylo() used to return TRUE with trees such as
      ((a:1):1,b:2,c:2):1;. This test now conforms with what is
      written in the documentation.


OTHER CHANGES

    o reconstruct() gains the options 'low_alpha = 0.0001' and
      'low_alpha = 1' to allow estimation of values of alpha greater
      than 1.

    o src/me.c and src/me.h are now OK with -Wstrict-prototypes flag.

    o makeNodeLabel(method = "md5sum") now uses the function digest
      from the package of the same name. The package tools has been
      removed from the Imports list.



		CHANGES IN APE VERSION 5.6-2


NEW FEATURES

    o plot.phylo() can now plot "tidy" trees with the option type =
      "t[idy]" (see: van der Ploeg, 2014, Drawing non-layered tidy
      trees in linear time. J. Software: Practice and Experience
      44:1467).


BUG FIXES

    o is.monophyletic() didn't work correctly with duplicated labels
      (issue #50 on GH).

    o birthdeath()'s 'if' conditions should be now all of length 1.



		CHANGES IN APE VERSION 5.6-1


BUG FIXES

    o axisPhylo() misplaced the scale with some trees drawn with
      plot(type = "fan").

    o The internal code of bitsplits() has been improved and can now
      handle all types of unrooted trees.



		CHANGES IN APE VERSION 5.6


NEW FEATURES

    o The new generic function degree returns the degrees of all nodes
      in a tree or network. There are methods for "phylo" and "evonet"
      objects.

    o The new function solveAmbiguousBases replaces R, Y, W, ... by A,
      G, C, or T using columnwise base frequencies.

    o There is a new method as.phylo.default() for any object
      inheriting the class "phylo".

    o read.dna() and read.FASTA() can now read gzipped files.

    o where() now works also with objects of class "AAbin". 'NULL' is
      now returned, instead of an error, if the sequence is shorter
      than the searched pattern (useful with lists with some short
      sequences; issue #28 on GH).

    o comparePhylo() has a new '...' argument for options passed to
      plot.phylo(), and two new options: 'commons' to specify whether
      to show the splits in common, and 'location' for the legend (PRs
      #41 and #42 on GH).

    o The code of dnds() has been rewritten which fixes some minor
      bugs, is more robust, and gains the options 'details' to print
      the numbers of TS and TV for each degeneracy category, and
      'return.category' to return the degeneracy categories of the
      original data (see ?dnds for details).


BUG FIXES

    o A bug was introduced in read.tree() in ape 5.5 when reading
      trees with quoted labels (issue #14 on GH).

    o nj() did not duplicate the distances if originally stored in a
      "dist" object (issue #12 on GH).

    o The double-centered matrix computed by pcoa() is now more
      strictly symmetric, thus avoiding complex eigenvalues in some
      cases (issue #17 on GH).

    o chronos() failed with discrete models when branch lengths were
      extremely heterogeneous leading to badly initial rate values.

    o multi2di() sometimes resulted in an ultrametric tree to become
      non-ultrametric (see issue #23 on GH).

    o read.nexus() failed to read correctly files with labels
      containing spaces in the TRANSLATION bloc (PR #29 on GH).

    o ace(, type = "d") now woks with n = 2.

    o drop.tip() used to remove the root edge of a tree if the option
      'root.edge = 0' was set. Since the latter is the default, the
      side-effect was to possibly turn a rooted tree into an unrooted
      one (see issue #32 on GH). The root edge is now removed only if
      the option 'rooted = TRUE' is used. The help page has been
      clarified on this point.

    o as.hclust.phylo() now returns the correct order of leaves so
      that plot() does not cross lines (PR #33 on GH).

    o unique.multiPhylo() failed with a single tree in the list or
      with an empty list (PR #38 on GH).

    o trans() failed with a matrix with a single codon. A similar bug
      affected as.character.AAbin().

    o rTraitDisc() now uses expm::expm unless this package is not
      installed in which case ape::matexpo is used.


OTHER CHANGES

    o multi2di() is faster as it avoids calling which() repeatedly.

    o dist.nodes() now gives an error with an explicit error message
      if the tree is too large (i.e., more than 46,340 tips + nodes).

    o The default values of the options 'edge.color', 'edge.width',
      and 'edge.lty' of plot.phylo() are now taken from the graphical
      parameters: par("fg"), par("lwd"), and par("lty"), respectively.

    o The visual aspect of circular trees (type = "f" in plot.phylo)
      has been improved.

    o ace() now accepts trees with some branch lengths equal to zero.

    o write.tree() now accepts simple lists of trees.

    o chronos(, model = "clock") is a short-cut to specifying model =
      "discrete" and a single category of rates in the settings.



		CHANGES IN APE VERSION 5.5


NEW FEATURES

    o The new functions rtopology and rmtopology generate random trees
      with equal frequencies of *labelled* topologies. rtree(,
      equiprob = TRUE) now gives equal frequencies of *unlabelled*
      topologies. These features are described in the new vignette
      "RandomTopologies".

    o dist.dna()'s internal code has been improved: it now accepts up
      to 2.1 billion sequences each up to 2.1 Gb (see CHANGES in APE
      VERSION 4.0 below).

    o nj()'s internal code has been improved and is slightly more
      efficient: combined with the improved dist.dna(), boot.phylo()
      with an NJ tree is now about 10% faster.

    o howmanytrees() can now compute (approximately) the number of
      binary, (un)rooted, labelled topologies beyond the standard
      resolution of numeric values (i.e., > 10^308) thanks to the new
      utility function LargeNumber.


BUG FIXES

    o bind.tree() sometimes returned badly conformed trees.

    o write.FASTA() made R crash if the output file was in a
      non-existing directory (thanks to Richel Bilderbeek for the
      report).

    o ltt.plot.coords() now handles correctly trees with singleton
      nodes. This also fixes functions calling this one (e.g.,
      ltt.plot).

    o estimate.dates() now works with non-bifurcating trees.

    o root(phy, outgroup, resolve.root = TRUE) failed to place
      correctly the new root if 'outgroup' was of longer two or
      more and paraphyletic (although monophyletic once 'phy' was
      unrooted).

    o pcoa() failed when a single axis was returned.


OTHER CHANGES

    o node- and tiplabels(pie = ) now accept data frames (or a
      single-column matrix/data frame).

    o boot.phylo() now returns integer(0) with a warning if there are
      less than 4 rows in the data (argument x).
      In case of unrooted trees, the first support value is now NA
      (it used to be B, the number of bootstrap replications).

    o prop.part() and is now faster thanks to some internal
      improvements. This also affects dist.topo() which is twice
      faster for small trees, and up to 30 times faster for trees with
      1000 tips.

    o bitsplits() now gives an error with rooted trees



		CHANGES IN APE VERSION 5.4-1


NEW FEATURES

    o rtree() gains several improvements:
      * The new option 'equiprob = TRUE' generates all topologies in
        equal proportions; 'equiprob = FALSE' generates unbalanced
        topologies in higher proportions (as in previous versions of
        ape). The latter is the default because several packages use
        rtree() with set.seed() in their examples.
      * Rooted trees with n = 1 tip can now be simulated.
      * Vector of tip labels supplied by the user are now checked and
        a warning is issued if its length does not match 'n'.
      * The internals of the code have been rewritten and are now more
	efficient (particularly if 'equiprob = TRUE').

    o multi2di() gains an option 'equiprob' similar to rtree().

    o The new function nexus2DNAbin is a helper to convert outputs
      from read.nexus.data().


BUG FIXES

    o is.monophyletic() might return the wrong answer if the tips were
      given as labels not alphabetically sorted.

    o boot.phylo() returned bootstrap values all equal to 0.



		CHANGES IN APE VERSION 5.4


NEW FEATURES

    o The new function getAnnotationsGenBank reads annotations from
      GenBank and returns them in data frame(s).

    o read.GenBank() gains two options: 'chunk.size' to set the number
      of sequences downloaded together, and 'quiet' to display the
      download progress (will also display the name of the temporary
      file where the sequences were saved). The options 'gene.names'
      has been removed.

    o The new convenience function latag2n (leading and traling
      alignment gaps to N) substitutes leading and trailing gaps in
      DNA alignments into N's. The gaps in the middle of the sequences
      are left unchanged.

    o There are new cbind() and rbind() methods for the class "AAbin".

    o seg.sites() gains two options: 'strict' and 'trailingGapsAsN' to
      give more flexibility in the presence of ambiguous bases and/or
      alignment gaps.

    o read.nexus.data() now handles polymorphisms in discrete
      characters (however with no distinction between ambiguous or
      true polymorphic characters).

    o trans() now supports genetic codes 1 to 6 (dnds() still supports
      only genetic codes 1 and 2).

    o chronos() has been improved with contributions by Santiago
      Claramunt:
      * Some errors were fixed in the calculation of the log-
        likelihood of the "discrete" models with more than one
        category of rates.
      * Calculations are now faster with "discrete" models.
      * A general convergence diagnostic has been added in the output.

      Some other contributions by Guillaume Louvel:
      * Improved algorithm to find initial dates.
      * There is now a parameter 'epsilon' to check convergence.
      * The number of iterations is output in the final results.

    o Phylogenetic correlation structures (functions with class
      "corPhyl") can now specify grouping factors via the arguments
      'form' and 'covariates' (which are no more ignored; see examples
      in ?ape::corClasses). This can be used, for instance, to specify
      species names to match the tree with the covariates.


BUG FIXES

    o A bug was fixed in pcoa() (thanks to Jesse Connell).

    o A bug was fixed in countBipartitions().

    o unroot() did not handle node labels correctly.

    o label2table() now keeps all levels in labels (not only the first
      three ones.

    o A bug was fixed in drop.tip(, root.edge = ).

    o rphylo() failed with time-dependent parameters.

    o is.monophyletic() now returns an error if some labels are not
      found in the tree (thanks to David Winter for the fix).

    o read.tree() and read.nexus() crashed R if the numbers of left
      and right parentheses in a Newick string were not equal: this is
      fixed and an error is now returned.


OTHER CHANGES

    o The option 'tbr' in fastme.bal() has been disabled.



		CHANGES IN APE VERSION 5.3


NEW FEATURES

    o The new function dnds calculates pairwise dN/dS ratios.

    o alview() has a new option 'showpos' to display the positions of
      the sites (can be a subset of these). Previously, only the
      position of the last site was printed.


BUG FIXES

    o mantel.test() returned NA if the second matrix had its diagonal
      with at least one NA.

    o Several bugs were fixed in plot.popsize(). It gains the options
      'xlab', 'ylab', and 'log' for more flexible plotting.


OTHER CHANGES

    o NAMESPACE has now explicit exports. The present version has 325
      exported objects. Most internal functions are not exported, as
      well as several methods.

    o checkLabel() is now documented.

    o The running times of write.tree() now scales with N (number of
      trees) instead of N^2: writing 2 million trees with 9 tips now
      takes 8 mins instead of 28 hrs with ape 5.2.



		CHANGES IN APE VERSION 5.2


NEW FEATURES

    o The new function keep.tip does the opposite operation than
      drop.tip().

    o The new function rDNAbin generates random DNA sequences.

    o The new generic function updateLabel changes some (or all)
      labels in an object giving two vectors 'old' and 'new' of
      labels. There are methods for the following classes: DNAbin,
      AAbin, phylo, evonet, data.frame, and matrix.

    o read.gff() has a new option 'GFF3' (TRUE by default) to set
      correctly the column names of GFF2 (aka GTF) files.

    o Xplor() has the new option from = "HOME" which is passed to
      Xplorefiles().

    o ace() can take state uncertainty for discrete characters into
      account: this should be coded with R's NA.

    o plotTreeTime() has a new option, color (TRUE by default).

    o as.phylo.formula() has a new option, collapse, to add a single
      node if a given taxonomic level has only one representative
      (contributed by Eric Marcon).


BUG FIXES

    o read.FASTA(, type = "AA") failed.

    o image.DNAbin() failed with a single sequence.

    o write.nexus.data() wrote the binary representation of DNA or AA
      sequences.

    o read.tree() failed to read correctly single-quoted node labels
      unless some tip labels were also single-quoted.

    o node.height(, clado.style = TRUE) returned the wrong vector (it
      used to return the same result than node.depth().

    o tiplabels() did not align labels properly if the tree was
      plotted with plot(, type = "fan", align.tip.label = TRUE).


OTHER CHANGES

    o print.DNAbin() now prints nicely the total number of bases.

    o drop.tip(tr, 1:Ntip(tr), trim.internal = FALSE) now returns a
      tree with all the terminal branches of tr removed, unless it has
      less than three nodes.



                CHANGES IN APE VERSION 5.1


NEW FEATURES

    o The new function write.FASTA is much more efficient than
      write.dna(, format = "fasta") for very big data set. It can also
      write "AAbin" objects into FASTA files.

    o The new function comparePhylo compares two trees with optional
      plot showing the differences.

    o There are three new methods as.AAbin() to convert from
      BioConductor for the classes "AAString", "AAStringSet", and
      "AAMultipleAlignment".

    o read.FASTA() can now read amino acid sequences; it has a new
      option type = "DNA" by default.

    o write.nexus.data() can now handle continuous and standard data
      thanks to a contribution by Thomas Guillerme.

    o muscle(), clustal(), and clustalomega() have new arguments to do
      tree-guided or progressive alignment, and a new option 'file' to
      save the alignment in a file (see ?details). These functions
      now handle "AAbin" sequences.

    o mantel.test() can now analyse asymmetric matrices following a
      suggestion by Andrzej Galecki.

    o as.prop.part.bitsplits() gains an option 'include.trivial =
      FALSE', and the generic as.prop.part() gains a '...' argument.


BUG FIXES

    o as.hclust.phylo() failed with two-tip trees.

    o drop.tip(, subtree = TRUE) failed when dropping a single tip.
      Thanks to Thomas Sibley for the fix.

    o nj() now checks that there are at least three observations.

    o chronos() better finds initial dates in difficult situations,
      particularly when the root age is unknown.

    o chronos() now exits with the current estimates if the plogLik is
      NA/NaN. Thanks to Daniel Lang for the fix.

    o plot.phylo(phy, type = "p", edge.color = x) failed if 'phy' had
      singleton nodes and 'x' was a vector of length > 1.

    o bind.tree() failed if the trees had unordered edges. Thanks to
      Veronika Boskova for the fix.

    o cbind.DNAbin() now checks that the rownames of each matrix are
      not duplicated (unless check.names = FALSE).


OTHER CHANGES

    o branching.times() now returns a zero-length vector and a warning
      if the tree has no branch length (it used to return an error).

    o igraph is now a suggested package.

    o The as.DNAbin() methods to convert from BioConductor are more
      efficient.



		CHANGES IN APE VERSION 5.0


NEW FEATURES

    o ape now supports all types of phylogenetic trees and nerworks
      (see details below).

    o The new functions read.evonet and write.evonet read and write
      files in Newick extended format (Cardona et al. 2008. BMC
      Bioinformatics 9:532).

    o Four new functions contributed by Franz Krah help to manipulate
      AA sequences: methods c(), as.list(), and as.matrix() for the
      class "AAbin", and as.AAbin() for lists.

    o The new function has.singles tests if a tree has single nodes:
      such trees can now be plotted with plot.phylo().

    o There are two new convenience functions cladewise() and
      postorder() which are short-cuts to reorder(, index.only =
      TRUE).

    o The new function Xplor shows all data files on the local machine
      in a Web browser with clickable links to the directories and
      files; it is based on the Xplorefiles() function.

    o image.DNAbin() and image.AAbin() gain two options, grid and
      show.base (or show.aa), to display a grid and the symbols of the
      bases (or amino acids), and three options to control the aspect
      of the latter: base.cex, base.font, and base.col (or aa.cex,
      aa.font, and aa.col).

    o image.AAbin() now shows the polymorphic positions.

    o Several internal improvements were made to pcoa(). biplot.pcoa()
      gains an optional argument 'main' to add a title.

    o tiplabels() has a new option 'offset' to position the labels
      with respect to the tips of the tree. This can be used in
      combination with 'adj' and works for all types of trees (except
      unrooted ones).

    o phydataplot(type = "mosaic", continuous = FALSE) allows to
      define the colours more flexibly thanks to a suggestion by
      Elizabeth Purdom; this can be done if the function passed with
      'funcol' returns a named vector. Also, the values are now sorted
      alphabetically in the legend.

    o phydataplot(type = "mosaic", legend = "none") does not display
      the legend.

    o read.tree() and read.nexus() can now read trees with singleton
      nodes (i.e., nodes of degree 2) thanks to Klaus.

    o read.tree() can now read labels with special characters within
      single quotes.

    o There are reorder() and Nedge() methods for objects of class
      "evonet", and a new generic as.evonet() with a method for the
      class "phylo".

    o LLT() gains a '...' argument (e.g., to use log = "y").

    o drop.tip() and extract.clade() gain an option 'collapse.singles'
      (TRUE by default).


BUG FIXES

    o CADM.*(... make.sym = FALSE) failed: this is now fixed.

    o phydataplot(x, , type = "mosaic") now works if x is a vector or
      a one-column matrix.

    o as.igraph.phylo() has been improved to work with igraph >= 1.0.

    o reorder.multiPhylo() failed with compressed lists.

    o arecompatible() returned FALSE when comparing two identical
      splits.

    o as.hclust.phylo() failed if internal branches near the root have
      length zero, so sorting the branching times failed due to
      rounding errors.

    o plot.phylo() failed when branch lengths were very long compared
      to the tip labels: use of the plotting space has been improved.


OTHER CHANGES

    o getMRCA() is now much faster (thanks to Joseph Brown and Klaus).

    o read.tree() and read.nexus() are now based on C code (thanks to
      Klaus) and should be several times faster.

    o The default read.tree(, comment.char = "#") has been changed to
      comment.char = "" so that extended Newick files can be read.

    o The internal code of prop.part() and of reorder.phylo() have
      been rewritten in C++ and should be several times faster.
      collapse.singles(), drop.tip(), and extract.clade() are also
      much faster.

    o dist.topo() is much faster with the default distance.

    o write.tree() and write.nexus() used to replace multiple
      underscores or dashes in labels by single ones: they are now
      left unchanged. The help page ?write.tree has been clarified on
      how tip and node labels are checked before being printed.



		CHANGES IN APE VERSION 4.1


NEW FEATURES

    o The new function read.fastq reads FASTQ files returning a
      "DNAbin" object with an attribute "QUAL".

    o The new functions Xplorefiles, bydir, and edit.file.extensions
      help to find files on the local disk.

    o read.FASTA() can now read connections thanks to a contribution
      by RJ Ewing: the help page shows how to read *.zip, *.gz, or
      *.bz2 files. This function can now read remote files through
      secure connections (with HTTPS or FTPS). These also work for
      read.dna(format = "fasta").

    o read.nexus() has a new option, force.multi (FALSE by default),
      to always return an object of class "multiPhylo" even if the
      file contains a single tree (to keep the tree name which is lost
      in the object of class "phylo").


BUG FIXES

    o read.FASTA() crashed when reading a file with no sequence or a
      compressed file not identified as such (see above about the
      support for connections).

    o read.GenBank() failed when at least one accession number was
      wrong: it now returns the sequences eventually read and prints a
      warning with the accession numbers not read.

    o extract.clade() failed with trees previously rooted with
      root(resolve.root = TRUE).

    o A bug was fixed in estimate.dates().

    o A bug was fixed in multi2di.multiPhylo().

    o drop.tip() failed when dropping all tips but one.

    o nj() now checks for infinite distances (it used to return a
      badly conformed tree).


OTHER CHANGES

    o chronos() now checks the presence of branch lengths in the tree.

    o print.DNAbin() now prints base frequencies until 10 million
      bases. The display of labels has been improved.



		CHANGES IN APE VERSION 4.0


NEW FEATURES

    o Two new functions, estimate.mu and estimate.dates, contributed
      by Bradley Jones, estimate the mutation rate and the dates for a
      rooted phylogenetic tree with dated tips.

    o The function plotTreeTime plots a non-ultrametric tree together
      with the dates of the tips.

    o The new function read.gff reads GFF files.

    o The new method all.equal.DNAbin compares two sets of DNA
      sequences mainly to compare alignments with a graphical display
      of the differences.

    o boot.phylo() has two new options: 'jumble' (TRUE by default) to
      randomize the order of the rows of the data matrix (and avoid
      artificially too large bootstrap values associated with very
      short branches), and 'mc.cores' to perform parallel bootstraps.

    o reconstruct() now computes ancestral states under directional
      models (Brownian with trend and OU; see: Royer-Carenzi, M. and
      Didier, G. 2016. A comparison of ancestral state reconstruction
      methods for quantitative characters. J Theor Biol, 404:126-142).

    o root() has the new option 'edgelabel' to treat node labels as
      edge labels so they are associated with the correct edges when
      using drawSupportOnEdges (edgelabel = FALSE by default).

    o read.GenBank() is faster and more flexible: it can read records
      such as contigs or scaffolds (based on a code available on
      GitHub for a while).

    o is.ultrametric() has a new argument, option, to change the
      criterion used to test ultrametricity.

    o .compressTipLabel() has a new option (ref = NULL) to constrain
      the order of the tips.


BUG FIXES

    o plot.phylo() used to fail with NAs in branch lengths: they are
      now ignored with a warning.

    o plot.phylo(, type = "unrooted", label.offset = 1) used to have
      no effect (unless lab4ut = "a" was used).

    o checkValidPhylo() failed with trees with a single node.

    o unroot() returned badly conformed trees in some situations.

    o richness.yule.test() failed when the speciation rate(s) is high:
      the code has been improved with better starting values for the
      optimization.

    o dist.topo() now checks that the trees have the same labels. The
      code is ~100 times faster for comparing 100 trees with 100
      labels.

    o as.character.AAbin() striped matrix rownames (thanks to Shaun
      Wilkinson for the report).


OTHER CHANGES

    o clustal(), clustalomega(), muscle(), and tcoffee() now give a
      more explicit error message when they fail to find the
      executable. The help page has been completed and gives some
      recommendations.

    o "DNAbin" objects larger than 2.1 billion bases (exactly 2^31-1)
      can now be handled by most functions in ape. The limitations are
      the same than in base R: 2^52 (~4.5*10^15) elements, and the
      numbers of rows and columns of matrices cannot exceed 2^31-1.
      read.dna() can now read FASTA files larger than 2.1 gigabytes.
      Two functions are still limited to 2.1 Gb: dist.dna and
      DNAbin2indel (the limit applies to the product of the numbers of
      sequences by the sequence length since they work with matrices).

    o As a side-effect of the above, several functions are slightly
      faster as several calls to .C have been replaced by .Call.

    o The following functions are now generic with methods for the
      classes "phylo" and "multiPhylo": di2multi, multi2di, Ntip,
      Nnode, Nedge, reorder, is.ultrametric, root, unroot, is.rooted,
      and is.binary (new function name, see below).

    o c.phylo() and c.multiPhylo() have now recursive = TRUE by
      default and try to return a list of single trees unless some
      objects are not trees or lists of trees.

    o image.DNAbin() and image.AAbin() now order rows from top to
      bottom.

    o read.GenBank() now uses HTTPS in place of HTTP.

    o ape now requires R 3.2.0 (or higher).

    o boot.phylo() now checks for duplicated labels.

    o checkAlignment() now prints the gap lengths that are on the
      borders of the alignment.


DEPRECATED & DEFUNCT

    o is.binary.tree() is now obsolete and will be removed soon. It
      currently calls is.binary().

    o The data sets landplants.newick and opsin were removed.



		CHANGES IN APE VERSION 3.5


NEW FEATURES

    o Eleven new functions are provided to work with the new class
      "AAbin":

        - trans() translates DNA to AA sequences;

        - complement() returns the reverse complement of DNA sequences;

        - Six methods to manipulate and display objects of class
          "AAbin": print, [, labels, as.character, as.phyDat, and
          image (alview() works also on AA sequences);

        - New generic as.AAbin() with a method for the class "character";

        - dist.aa() computes pairwise Hamming distances;

        - AAsubst() returns the indices of polymporphic sites (similar
          to seg.sites() for "DNAbin").

    o The new function checkAlignment performs diagnostics on a DNA
      alignment.

    o The new functions plotBreakLongEdges and drawSupportOnEdges help
      to plot and annotate trees.

    o del.colgapsonly() gains two options: threshold (1 by default) to
      delete columns with less than 100% of gaps, and freq.only (FALSE
      by default) to return only the number of gaps for each column.

    o The new function del.rowgapsonly does the same operation than
      del.colgapsonly but on the rows (with the same new options).


BUG FIXES

    o root() did not calculate the basal branch length correctly
      (thanks to Liam Revell for the report).

    o clustal() failed with long labels.

    o as.DNAbin.character() now works with uppercase letters as well.

    o prop.clades() now returns a sensible results whatever the value
      of its option 'rooted' (fixed by Klaus). The code is also faster.

    o root(resolve.root = TRUE) failed in some situations.


OTHER CHANGES

    o boot.phylo(phy, ....) has now by default rooted = is.rooted(phy).



		CHANGES IN APE VERSION 3.4


NEW FEATURES

    o The new function checkValidPhylo checks the internal structure
      of a "phylo" object.

    o The new function clustalomega calls Clustal-Omega from R.

    o The new function DNAbin2indel recodes DNA sequences with gaps
      as numeric values.

    o The new function alview prints an alignment in a user-friendly
      way in the console or a file.

    o Three new functions, label2table, stripLabel and
      abbreviateGenus, help to manage taxa labels.

    o phydataplot() now accepts style="boxplot" and style="dotchart",
      and has a new option 'scaling' when the data are on a very
      different scale than the branch lengths of the tree.

    o phydataplot() now accepts style="mosaic" to plot discrete or
      continuous variables in matrix form, and gains three options
      (width, continuous, funcol) that work with this option.

    o plot.phylo() has a new option (align.tip.label) to align the
      labels of the tips even when the tree is not ultrametric.


BUG FIXES

    o parafit() did not work with data frames or "dist" objects: this
      is now fixed.

    o plot.phylo() did not manage space correctly when show.tip.label
      = FALSE and some tip labels were long (thanks to David Bapst for
      the report and to Liam Revell for the fix).

    o collapse.singles() failed in some situations, returning trees
      that could make R crash. The function was completely re-written.

    o In clustal(), the default path and executable names were not
      correctly set under Windows.

    o balance() failed after reordering a tree (bug reported by
      G. Valiente in January 2008).

    o seg.sites() failed to find some segregating sites in some
      specific situations: gaps and ambiguous nucleotides are now
      handled correctly. The code is slightly faster.

    o rtt() didn't find the optimal root if located on a long branch.

    o plot.phylo(, type = "f") sometimes failed when colouring the
      edges of a tree with multichotomies (thanks to François
      Michonneau for the fix).

    o rphylo() now returns better conformed trees.

    o drop.tip() sometimes shuffled node labels.

    o edgelabels() did not put the labels in the correct place with
      "fan" trees (fixed by Klaus).

    o reorder.phylo(, index.only = TRUE) sometimes returned the tree
      if it has a single node.

    o as.hclust.phylo() returned node heights divided by two.

    o ladderize() did not work with trees not in "cladewise" postorder.

    o dist.dna(, model = "indelblock") has been fixed.

    o as.hclust.phylo failed when trees had ties in their branching
      times.

    o root() failed with some trees returned by phangorn.


OTHER CHANGES

    o Updated NAMESPACE file for r-devel.

    o root() gives a more explicit error message when the outgroup is
      not among the labels of the tree.



		CHANGES IN APE VERSION 3.3


NEW FEATURES

    o Two new functions contributed by Anthony Ives, binaryPGLMM and
      binaryPGLMM.sim, fit phylogenetic generalized linear mixed
      models to binary data and simulate such data.

    o The new function corphylo, contributed by Anthony Ives,
      calculates Pearson correlation coefficients for multiple
      continuous traits that may have phylogenetic signal.

    o The new function rphylo simulates trees under any time-dependent
      model of diversification using the method from Stadler (2011,
      Syst Biol 60: 676) which conditions the simulation on a fixed
      number of living species at present time.

    o The new function del.colgapsonly removes the columns of a DNA
      alignment that contain only gaps (useful when a small matrix is
      extracted from a large alignment).

    o There is a sort() method for objects of class "bitsplits".

    o plot.phylo(type = "fan", root.edge = TRUE) now draws the root
      edge. Besides, the root edge is now coloured like the other
      edges if they are all of the same colour (for all tree types).

    o dist.topo() now accepts lists of trees.


BUG FIXES

    o prop.clades() failed with compressed lists of trees.

    o read.tree() did not ignore comments inside straight brackets
      (contrary to what is written in ape's FAQ) potentially resulting
      in wrongly conformed "phylo" objects.

    o dbdTime(x = 0, ...., conditional = TRUE) did not return 0.

    o parafit() required a fix when F2.stat could not be calculated.

    o write.nexus.data() did not write "DATATYPE=" correctly so that
      the output NEXUS file could not be read by some other programs
      (thanks to Robin Cristofari for the fix).

    o seg.sites() made R crash when all sequences ended with N's.

    o root(phy, node = Ntip(phy) + 1, resolve.root = TRUE) now returns
      an explicit error.


OTHER CHANGES

    o plot.prop.part() has been improved: it now uses image()
      internally, and gains a new option (col). It is now much more
      efficient to display very large numbers of bipartitions.

    o as.character.DNAbin() and as.DNAbin.character() are much faster
      (contributed by Klaus Schliep).


DEPRECATED & DEFUNCT

    o node.height.clado() has been removed.



		CHANGES IN APE VERSION 3.2


NEW FEATURES

    o The new function reconstruct, contributed by Manuela
      Royer-Carenzi and Gilles Didier, does ancestral character state
      reconstruction using new algorithms based on matrix computations.

    o The new function vcv2phylo, contributed by Simon Blomberg,
      transforms a variance-covariance matrix into a phylogenetic tree
      (ultrametric or not).

    o The new function def helps to (re)define attributes for plotting
      and annotating trees using taxon names, labels, or other vectors
      of character strings.

    o There are four new as.DNAbin functions (S3 methods) to convert
      objects storing DNA sequences in BioConductor. The supported
      (S4) classes are: DNAString, DNAStringSet, DNAMultipleAlignment,
      and PairwiseAlignmentsSingleSubject (the examples show how to
      convert from the class DNAStringSetList).

    o The new function nodepath finds paths of nodes in a tree.

    o axisPhylo() gains two options: root.time and backward (similar
      to ltt.plot).

    o write.nexus.data() now accepts matrices (it accepted only lists
      in previous versions).


BUG FIXES

    o dist.dna() did not compute distances correctly if gamma = TRUE.

    o read.nexus() sometimes failed on files with no TRANSLATION block.

    o read.nexus() failed to read correctly tree names (thanks to
      Graham Gower for the fix).

    o multi2di() used to return trees with elements as doubles instead
      of integers (this was not a problem for most applications).

    o In the output of phymltest(), the model F84 and HKY were
      swapped (thanks to Luiz Max Fagundes de Carvalho for the fix).


OTHER CHANGES

    o plot.phylo() now saves the element 'root.time' of the plotted
      tree (if present) in the environment .PlotPhyloEnv.

    o write.nexus.data() now writes correctly all "DNAbin" objects
      (matrices and lists). The help page has been clarified.



		CHANGES IN APE VERSION 3.1-4


BUG FIXES

    o A bug was fixed in which.edge().


OTHER CHANGES

    o This file is no more distributed with the sources of ape and can
      be found at: http://ape-package.ird.fr/NEWS.



		CHANGES IN APE VERSION 3.1-3


NEW FEATURES

    o The new function rtt contributed by Rosemary McCloskey roots a
      tree with dated tips in the location most compatible with the
      assumption of a strict molecular clock.


BUG FIXES

    o root(, resolve.root = TRUE) did not manage node labels correctly
      in the case where 'outgroup' was already an outgroup of the tree.

    o read.dna(file, format = "fasta") works (again) with URIs (i.e.,
      file = "http://..." or "ftp://...").

    o rotate() made R crash if the tree was not in cladewise order.

    o which.edge() did not work correctly for trees not in cladewise
      order.

    o read.GenBank failed to read older GenBank records.



		CHANGES IN APE VERSION 3.1-2


NEW FEATURES

    o plot.phylo: the option 'open.angle' now works also when type =
      "radial". The option 'lab4ut' now works also for type = "radial"
      or "fan"; the default is now NULL since it will behave
      differently depending on 'type'.

    o ace(, type = "discrete") gains an option 'marginal' (FALSE by
      default). The default is now to do a joint reconstruction of
      ancestral states.


BUG FIXES

    o chronos: the gradients were not correctly computed when the age
      of the root was unknown.

    o prop.part() did not work with objects of class composite like
      c("chronos", "phylo"). (Thanks to Steve Walker for the fix.)

    o bionjs: a bug was fixed in the C code.

    o read.GenBank() failed if a line starts with "ORIGIN" in the
      description of the sequence (thanks to Sofia Sal Bregua for the
      fix).

    o cophyloplot() works better with length.line = 0.

    o The internal function clado.build used to return trees with the
      'edge' matrix stored as double instead of integer.


OTHER CHANGES

    o base.freq() now checks the class of the data.



		CHANGES IN APE VERSION 3.1-1


BUG FIXES

    o rTraitCont: a bug was introduced in the previous release.


OTHER CHANGES

    o ape now requires R 3.0.0 or higher.



		CHANGES IN APE VERSION 3.1


NEW FEATURES

    o Two new functions, bitsplits and countBipartitions, handle
      bipartitions (aka splits) more efficiently.

    o The new generic function as.prop.part helps to convert among
      classes of bipartitions.

    o Two new functions, phydataplot and ring, helps to graphically
      annotate trees. The help page has many examples.

    o The new function LLT draws the theoretical LTT-plot under
      specified values of speciation and extinction rates together
      with a prediction interval. The function has several options for
      flexible plotting.

    o predict.compar.gee() gains a new option, newdata, to predict
      values from new observations of the predictors. This works like
      most predict() methods.


BUG FIXES

    o root: the fix introduced in the previous version was not correct.

    o compar.gee: using a user-defined link failed.

    o plot.phylo: the tip labels were often outside the plotting area
      with direction = "downward".

    o unique.multiPhylo() did not work correctly with compressed lists.
      This function now returns a vector of integers giving the
      correspondance among similar trees.


OTHER CHANGES

    o add.scale.bar() now draws a longer bar by default.

    o In boot.phylo(), the data 'x' must be a matrix-like object
      (lists are no more accepted).

    o boot.phylo() uses by default (ie, if rooted = FALSE) the new
      function countBipartitions, and so should be much faster even
      with small sample sizes and especially if the number of
      bootstrap replicates is large.

    o plot.phylo(, type = "fan") now colours the arcs in the same way
      than the default type = "p".

    o All instances of DUP=FALSE in calls of .C have been removed.



		CHANGES IN APE VERSION 3.0-11


BUG FIXES

    o branching.times() now reorders the tree if needed.

    o root(, resolved.root = TRUE) sometimes misplaced the root when
      'outgroup' was of length one.

    o nodelabels() didn't work with versions of R older than 3.0.0.


OTHER CHANGES

    o The code of branching.times() is now ca. 3 times faster.



		CHANGES IN APE VERSION 3.0-10


NEW FEATURES

    o The new function rotateConstr rotates internal branches given a
      constraint on the order of the tips.

    o plot.phylo() has a new option, node.depth, to specify the depths
      of nodes when the tree is plotted without branch lengths.
      node.depth() has a new option, method, with the same effect.

    o ace() has a new option, use.eigen, used when type = "d" to avoid
      computing matrix exponentials. With other coding improvements,
      the function is now about five times faster.


BUG FIXES

    o node.height() and node.height.clado() returned wrong values.

    o pie charts drawn by nodelabels(), and other functions, used to
      draw a "3-o'clock" segment when only one proportion was equal to
      one (fixed by Klaus).

    o plot(phy, main = "title...") now works again.


OTHER CHANGES

    o Improved DESCRIPTION and NAMESPACE files.

    o The C routines are now registered.

    o The generic function as.igraph has been removed from ape as it
      is now defined in the package igraph.

    o node.height() has now an option 'clado.style' (FALSE by default).
      node.height.clado() will be removed soon.

    o The code of several functions has been improved.


DEPRECATED & DEFUNCT

    o mst() has been moved to pegas.



		CHANGES IN APE VERSION 3.0-9


NEW FEATURES

    o getMRCA() is now documented.


BUG FIXES

    o ace(, type = "d") now checks for zero or negative branch lengths.

    o ace(, type = "d") now catches the error if the SEs of the rates
      cannot be calculated (usually due to a poor model fit).

    o reorder.phylo() made R crash with badly conformed "phylo" trees.

    o drop.tip(phy, interactive = TRUE) did not work.

    o drop.tip() now returns NULL (with a warning) if all tips are
      dropped.

    o mixedFontLabel() failed when passed only one vector of
      labels. It also failed when labels included - or '. In some
      cases, spaces in labels were not treated correctly.

    o root(, resolve.root = TRUE) wrongly placed the new root node
      when the outgroup has more than one tip. This is fixed, and now
      the root node is placed at the MRCA of the outgroup (and not of
      the ingroup).

    o where() failed with matrices.



		CHANGES IN APE VERSION 3.0-8


NEW FEATURES

    o The new function ewLasso tests whether an incomplete set of
      distances uniquely determines the edge weights of a given
      unrooted topology using the 'Lasso' method by Dress et
      al. (2012, J. Math. Biol. 65:77).

    o ace() gains a new option 'use.expm' to use expm() from the
      package of the same name in place of matexpo().


BUG FIXES

    o read.dna(, "fasta") may add '\r' in labels: this is fixed.

    o prop.clades() returned wrong numbers when the tip labels of
      'phy' are not in the same order than the list of trees (thanks
      to Rupert Collins for the report).

    o CADM.post() displayed "1" on the diagonal of the matrix of
      Mantel p-values. It now displays "NA" on the diagonal,
      indicating that no test of significance is computed between a
      distance matrix and itself.

    o rtree(n, rooted = FALSE) returned trees with an 'edge' matrix
      stored as doubles instead of integers for n > 4.


OTHER CHANGES

    o The files CDAM.global.R and CDAM.post.R have been renamed
      CADM.global.R and CADM.post.R.

    o ace() has a new default for its option 'method': this is "REML"
      for continuous characters and "ML" for discrete ones.

    o ape does not import gee anymore so the latter doesn't need to
      be installed.



		CHANGES IN APE VERSION 3.0-7


NEW FEATURES

    o The new function chronos estimates chronograms by penalised
      likelihood and maximum likelihood with a completely reworked
      code and interface. There is a new function makeChronosCalib to
      set the calibration points easily. chronos() will eventually
      replace chronopl().

    o The new function 'where' searches patterns in DNA sequences.

    o pic() gains an option 'rescaled.tree = FALSE' to return the tree
      with its branch lengths rescaled for the PIC calculation.

    o clustal(), muscle(), and tcoffee() gain an option
      'original.ordering = TRUE' to ease the comparisons of
      alignments.

    o plot.phylo() has a new option, open.angle, used when plotting
      circular trees.

    o The new function read.FASTA reads FASTA files much faster and
      more efficiently. It is called internally by read.dna(, "fasta")
      or can be called directly.


BUG FIXES

    o drop.tip() shuffled node labels on some trees.

    o axisPhylo() now works correctly with circular trees, and gives a
      sensible error message when type = "r" or "u".


OTHER CHANGES

    o .compressTipLabel() is 10 times faster thanks to Joseph Brown.

    o base.freq() is now faster with lists.

    o as.matrix.DNAbin() should be faster and more efficient with
      lists; it now accepts vectors.



		CHANGES IN APE VERSION 3.0-6


NEW FEATURES

    o reorder.phylo() has a new order, "postorder", and a new option
      index.only = TRUE to return only the vector of indices (the tree
      is unmodified, see ?reorder.phylo for details).

    o The three new functions node.depth.edgelength, node.height, and
      node.height.clado make some internal code available from R. See
      ?node.depth (which was already documented) for details.


BUG FIXES

    o reorder(, "pruningwise") made R crash if the rows of the edge
      matrix are in random order: this is now fixed.

    o drop.tip() sometimes shuffled node labels (thanks to Rebecca
      Best for the report).

    o drop.tip(phy, "") returned a tree with zero-length tip labels:
      it now returns the tree unchanged (thanks to Brian Anacker for
      the report).

    o plot.phylo() made R crash if the tree has zero-length tip
      labels: it now returns NULL (thanks again to Brian Anacker).


OTHER CHANGES

    o dist.nodes() is now 6 to 10 times faster.

    o reorder(, "cladewise") is now faster. The change is not very
      visible for small trees (n < 1000) but this can be more than
      1000 faster for big trees (n >= 1e4).

    o The attribute "order" of the objects of class "phylo" is now
      strongly recommended, though not mandatory. Most functions in
      ape should return a tree with this attribute correctly set.

    o dbd() is now vectorized on both arguments 'x' (number of species
      in clade) and 't' (clade age) to make likelihood calculations
      easier and faster.



		CHANGES IN APE VERSION 3.0-5


BUG FIXES

    o ace() should better catch errors when SEs cannot be computed.


OTHER CHANGES

    o write.dna(format = "fasta") now conforms more closely to the
      FASTA standard thanks to François Michonneau.

    o print.DNAbin() does not print base compositions if there are more
      than one million nucleotides.



		CHANGES IN APE VERSION 3.0-4


BUG FIXES

    o read.dna() failed to read Phylip files if the first line used
      tabulations instead of white spaces.

    o read.dna() failed to read Phylip or Clustal files with less than
      10 nucleotides. (See other changes in this function below.)

OTHER CHANGES

    o read.dna() now requires at least one space (or tab) between the
      taxa names and the sequences (whatever the length of taxa
      names). write.dna() now follows the same rule.

    o The option 'seq.names' of read.dna has been removed.

    o The files ape-defunct.R and ape-defunct.Rd, which have not been
      modified for almost two years, have been removed.

    o The C code of bionj() has been reworked: it is more stable (by
      avoiding passing character strings), slightly faster (by about
      20%), and numerically more accurate.

    o The C code of fastme.*() has been slightly modified and should
      be more stable by avoiding passing character strings (the
      results are identical to the previous versions).

    o The file src/newick.c has been removed.



		CHANGES IN APE VERSION 3.0-3


BUG FIXES

    o birthdeath() now catches errors and warnings much better so that
      a result is returned in most cases.


OTHER CHANGES

    o Because of problems with character string manipulation in C, the
      examples in ?bionj and in ?fastme have been disallowed. In the
      meantime, these functions might be unstable. This will be solved
      for the next release.



		CHANGES IN APE VERSION 3.0-2


NEW FEATURES

    o The new function alex (alignment explorator) zooms in a DNA
      alignment and opens the result in a new window.


BUG FIXES

    o compute.brtime() did not completely randomized the order of the
      branching times.

    o write.nexus() did not work correctly with rooted trees (thanks
      to Matt Johnson for the fix).

    o mltt.plot(, backward = FALSE) did not set the x-axis correctly.

    o A bug was introduced in prop.clades() with ape 3.0. The help page
      has been clarified relative to the use of the option 'rooted'.

    o mantel.test() printed a useless warning message.

    o plot.phylo(, direction = "downward") ignored 'y.lim'.

    o is.monophyletic() did not work correctly if 'tips' was not stored
      as integers.

    o prop.part() could make R crash if the first tree had many
      multichotomies.

    o njs(), bionjs(), and mvrs() now return an error if 'fs < 1'.

    o SDM() did not work correctly. The code has also been generally
      improved.


OTHER CHANGES

    o The DESCRIPTION file has been updated.

    o The option 'original.data' of write.nexus() has been removed.

    o The files bionjs.c, mvr.c, mvrs.c, njs.c, triangMtd.c, and
      triangMtds.c have been improved which should fix some bugs in
      the corresponding functions.

    o dist.gene() now coerces input data frame as matrix resulting in
      much faster calculations (thanks to a suggestion by Markus
      Schlegel).



		CHANGES IN APE VERSION 3.0-1


NEW FEATURES

    o dist.dna() has two new models: "indel" and "indelblock".

    o bind.tree() now accepts 'position' > 0 when the trees have no
      banch length permitting to create a node in 'x' when grafting
      'y' (see ?bind.tree for details).


BUG FIXES

    o cophyloplot( , rotate = TRUE) made R hanged after a few clicks.
      Also the tree is no more plotted twice.

    o read.GenBank() has been updated to work with EFetch 2.0.

    o unroot() on trees in "pruningwise" order did not respect this
      attribute.



		CHANGES IN APE VERSION 3.0


NEW FEATURES

    o The three functions dyule, dbd, and dbdTime calculate the
      density probability (i.e., the distribution of the number of
      species) for the Yule, the constant and the time-dependent
      birth-beath models, respectively. These probabilities can be
      conditional on no extinction and/or on a log-scale.

    o plot.phylo() has a new option 'rotate.tree' to rotate unrooted,
      fan, or radial trees around the center of the plot.

    o boot.phylo() and prop.clades() have a new option rooted =
      FALSE. Note that the behaviour of prop.part() is unchanged.

    o edgelabels() has a new option 'date' to place labels on edges of
      chronograms using the time scale (suggestion by Rob Lanfear).


BUG FIXES

    o In chronopl(), the code setting the initial dates failed in
      complicated settings (several dates known within intervals).
      This has been generally improved and should result in faster
      and more efficient convergence even in simple settings.

    o mantel.test() sometimes returned P-values > 1 with the default
      two-tailed test.

    o extract.clade() sometimes shuffled some tip labels (thanks to
      Ludovic Mallet and Mahendra Mariadassou for the fix).

    o clustal() should now find by default the executable under Windows.


OTHER CHANGES

    o The code of yule() has been simplified and is now much faster for
      big trees.

    o The code of mantel.test() has been adjusted to be consistent
      with common permutation tests.

    o The C code of base.freq() has been improved and is now nearly 8
      times faster.

    o The option 'original.data' of write.nexus() is now deprecated and
      will be removed in a future release.

    o The code of is.ultrametric() has been improved and is now 3 times
      faster.

    o The code of vcv.phylo() has been improved and is now 10 or 30
      times faster for 100 or 1000 tips, respectively. Consequently,
      fitting models with PGLS will be faster overall.



		CHANGES IN APE VERSION 2.8


NEW FEATURES

    o Twelve new functions have been written by Andrei-Alin Popescu:
      additive, ultrametric, is.compatible, arecompatible, mvr, mvrs,
      njs, bionjs, SDM, treePop, triangMtd, triangMtd*.

    o A new class "bitsplits" has been created by Andrei-Alin Popescu
      to code splits (aka, bipartition).

    o There is a new generic function as.bitsplits with a method to
      convert from the class "prop.part" to the class "bitsplits".

    o The new function ltt.coplot plots on the same scales a tree and
      the derived LTT plot.

    o ltt.plot() has two new options: backward and tol. It can now
      handle non-ultrametic trees and its internal coding has been
      improved. The coordinates of the plot can now be computed with
      the new function ltt.plot.coords.


BUG FIXES

    o prop.part() crashed if some trees had some multichotomies.



		CHANGES IN APE VERSION 2.7-3


NEW FEATURES

    o The new function compute.brtime computes and sets branching times.

    o mantel.test() has a new argument 'alternative' which is
      "two-sided" by default. Previously, this test was one-tailed
      with no possibility to change.

    o ace() can now do REML estimation with continuous characters,
      giving better estimates of the variance of the Brownian motion
      process.


BUG FIXES

    o Branch lengths were wrongly updated with bind.tree(, where = <tip>,
      position = 0). (Thanks to Liam Revell for digging this bug out.)

    o Simulation of OU process with rTraitCont() did not work correctly.
      This now uses formula from Gillespie (1996) reduced to a BM
      process when alpha = 0 to avoid division by zero. The option
      'linear' has been removed.

    o Cross-validation in chronopl() did not work when 'age.max' was
      used.

    o consensus(, p = 0.5) could return an incorrect tree if some
      incompatible splits occur in 50% of the trees (especially with
      small number of trees).

    o c() with "multiPhylo" did not work correctly (thanks to Klaus
      Schliep for the fix).

    o root() failed in some cases with an outgroup made of several tips.
      The help page has been clarified a bit.



		CHANGES IN APE VERSION 2.7-2


NEW FEATURES

    o There is a new class "evonet" to code evolutionary networks, with
      a constructor function evonet(), a print() and a plot() methods,
      and four conversion methods to the classes "phylo", "networx",
      "network", and "igraph".

    o The new function rTraitMult does multivariate traits simulation
      with user-defined models.

    o plot.phylo() has a new option 'plot = TRUE'. If FALSE, the tree
      is not plotted but the graphical device is set and the
      coordinates are saved as usual.

    o diversity.contrast.test() gains a fourth version of the test with
      method = "logratio"; the literature citations have been clarified.

    o add.scale.bar() has two new options, 'lwd' and 'lcol', to modify
      the aspect of the bar.

    o boot.phylo() now displays a progress bar by default (can be off
      if 'quiet = TRUE').

    o There is a new predict() method for compar.gee().


BUG FIXES

    o bionj() made R crash if distances were too large. It now returns
      an error if at least one distance is greater than 100.

    o drop.tip() returned a wrong tree if 'tip' was of zero length.

    o read.nexus.data() failed with URLs.

    o boot.phylo() returned overestimated support values in the
      presence of identical or nearly identical sequences.


OTHER CHANGES

    o The data bird.families, bird.orders, cynipids, and woodmouse are
      now provided as .rda files.



		CHANGES IN APE VERSION 2.7-1


NEW FEATURES

    o The new function trex does tree exploration with multiple
      graphical devices.

    o The new function kronoviz plots several rooted (dated) trees on
      the scale scale.

    o identify.phylo() has a new option 'quiet' (FALSE by default).


BUG FIXES

    o A bug was introduced in read.nexus() in ape 2.7.

    o image.DNAbin() did not colour correctly the bases if there were
      some '-' and no 'N'.

    o .compressTipLabel() failed with a list with a single tree.

    o identify.phylo() returned a wrong answer when the x- and y-scales
      are very different.

    o write.nexus() failed with lists of trees with compressed labels.


OTHER CHANGES

    o identify.phylo() now returns NULL if the user right- (instead of
      left-) clicks (an error was returned previously).

    o read.nexus() should be robust to commands inserted in the TREES
      block.



		CHANGES IN APE VERSION 2.7


NEW FEATURES

    o There is a new image() method for "DNAbin" objects: it plots DNA
      alignments in a flexible and efficient way.

    o Two new functions as.network.phylo and as.igraph.phylo convert
      trees of class "phylo" into these respective network classes
      defined in the packages of the same names.

    o The three new functions clustal, muscle, and tcoffee perform
      nucleotide sequence alignment by calling the external programs
      of the same names.

    o Four new functions, diversity.contrast.test, mcconwaysims.test,
      richness.yule.test, and slowinskiguyer.test, implement various
      tests of diversification shifts using sister-clade comparisons.

    o base.freq() gains an option 'all' to count all the possible bases
      including the ambiguous ones (defaults to FALSE).

    o write.nexus() now writes tree names in the NEXUS file if given a
      list of trees with names.


BUG FIXES

    o prop.part() failed in some situations with unrooted trees.

    o read.nexus() shuffled node labels when a TRANSLATE block was
      present.

    o varCompPhylip() did not work if 'exec' was specified.

    o bind.tree() shuffled node labels when position > 0 and 'where'
      was not the root.


OTHER CHANGES

    o BaseProportion in src/dist_dna.c has been modified.

    o A number of functions in src/tree_build.c have been modified.

    o The matching representation has now only two columns as the third
      column was redundant.



		CHANGES IN APE VERSION 2.6-3


NEW FEATURES

    o rTraitCont() and rTraitDisc() gains a '...' argument used with
      user-defined models (suggestion by Gene Hunt).


BUG FIXES

    o as.hclust.phylo() now returns an error with unrooted trees.

    o as.hclust.phylo() failed with trees with node labels (thanks to
      Jinlong Zhang for pointing this bug out).

    o read.dna(, "fasta") failed if sequences were not all of the same
      length.

    o plot.phylo() did not recycle values of 'font', 'cex' and
      'tip.color' correctly when type = "fan" or "radial".

    o plot.phylo() ignored 'label.offset' when type = "radial", "fan", or
      "unrooted" with lab4ut = "axial" (the placement of tip labels still
      needs to be improved with lab4ut = "horizontal").


OTHER CHANGES

    o In drop.fossil() the default tol = 0 has been raised to 1e-8.

    o The help command ?phylo now points to the man page of read.tree()
      where this class is described. Similarly, ?matching points to the
      man page of as.matching().



		CHANGES IN APE VERSION 2.6-2


NEW FEATURES

    o Two new functions, pic.ortho and varCompPhylip, implements the
      orthonormal contrasts of Felsenstein (2008, Am Nat, 171:713). The
      second function requires Phylip to be installed on the computer.

    o bd.ext() has a new option conditional = TRUE to use probabilities
      conditioned on no extinction for the taxonomic data.


BUG FIXES

    o write.tree() failed to output correctly tree names.

    o dist.nodes() returned duplicated column(s) with unrooted and/or
      multichotomous trees.

    o mcmc.popsize() terminated unexpectedly if the progress bar was
      turned off.

    o prop.part(x) made R frozen if 'x' is of class "multiPhylo".

    o Compilation under Mandriva failed (thanks to Jos Käfer for the fix).

    o drop.tip() shuffled tip labels with subtree = TRUE or trim.internal
      = FALSE.

    o Objects returned by as.hclust.phylo() failed when analysed with
      cutree() or rect.hclust().

    o write.tree() did not output correctly node labels (thanks to Naim
      Matasci and Jeremy Beaulieu for the fix).

    o ace(type = "discrete") has been improved thanks to Naim Marasci and
      Jeremy Beaulieu.



		CHANGES IN APE VERSION 2.6-1


NEW FEATURES

    o The new function speciesTree calculates the species tree from a set
      of gene trees. Several methods are available including maximum tree
      and shallowest divergence tree.


BUG FIXES

    o A bug introduced in write.tree() with ape 2.6 has been fixed.

    o as.list.DNAbin() did not work correctly with vectors.

    o as.hclust.phylo() failed with trees with node labels (thanks to
      Filipe Vieira for the fix).



		CHANGES IN APE VERSION 2.6


NEW FEATURES

    o The new functions rlineage and rbdtree simulate phylogenies under
      any user-defined time-dependent speciation-extinction model. They
      use continuous time algorithms.

    o The new function drop.fossil removes the extinct species from a
      phylogeny.

    o The new function bd.time fits a user-defined time-dependent
      birth-death model. It is a generalization of yule.time() taking
      extinction into account.

    o The new function MPR does most parsimonious reconstruction of
      discrete characters.

    o The new function Ftab computes the contingency table of base
      frequencies from a pair of sequences.

    o There is now an 'as.list' method for the class "DNAbin".

    o dist.dna() can compute the number of transitions or transversions
      with the option model = "Ts" or model = "Tv", respectively.

    o [node|tip|edge]labels() gain three options with default values to
      control the aspect of thermometers: horiz = TRUE, width = NULL,
      and height = NULL.

    o compar.gee() has been improved with the new option 'corStruct' as an
      alternative to 'phy' to specify the correlation structure, and
      calculation of the QIC (Pan 2001, Biometrics). The display of the
      results has also been improved.

    o read.GenBank() has a new option 'gene.names' to return the name of
      the gene (FALSE by default).


BUG FIXES

    o extract.clade() sometimes shuffled the tip labels.

    o plot.phylo(type = "unrooted") did not force asp = 1 (thanks to Klaus
      Schliep for the fix)

    o dist.dna(model = "logdet") used to divide distances by 4. The
      documentation has been clarified on the formulae used.


OTHER CHANGES

    o rTraitCont(model = "OU") has an option 'linear = TRUE' to possibly
      change the parameterisation (see ?rTraitCont for details).

    o pic() now returns a vector with the node labels of the tree (if
      available) as names.

    o write.tree() and read.tree() have been substantially improved thanks
      to contributions by Klaus Schliep.



		CHANGES IN APE VERSION 2.5-3


NEW FEATURES

    o The new function mixedFontLabel helps to make labels with bits of
      text to be plotted in different fonts.

    o There are now replacement operators for [, [[, and $ for the class
      "multiPhylo" (i.e., TREES[11:20] <- rmtree(10, 100)). They possibly
      check that the tip labels are the same in all trees.

    o Objects of class "multiPhylo" can be built with c(): there are
      methods for the classes "phylo" and "multiPhylo".

    o The internal functions .compressTipLabel and .uncompressTipLabel are
      now documented.


BUG FIXES

    o bind.tree(x, y, where, position = 0) did not work correctly if 'y'
      was a single-edge tree and 'where' was a tip.

    o rTraitCont() did not use the square-root of branch lengths when
      simulating a Brownian motion model.



		CHANGES IN APE VERSION 2.5-2


NEW FEATURES

    o There is now a print method for results from ace().

    o There is a labels() method for objects of class "DNAbin".

    o read.dna() has a new option 'as.matrix' to possibly force sequences
      in a FASTA file to be stored in a matrix (see ?read.dna for details).


BUG FIXES

    o as.phylo.hclust() used to multiply edge lengths by 2.

    o A minor bug was fixed in rTraitDisc().

    o ace() sometimes failed (parameter value was NaN and the optimisation
      failed).


DEPRECATED & DEFUNCT

    o evolve.phylo() and plot.ancestral() have been removed.

    o chronogram(), ratogram(), and NPRS.criterion() have been removed.


OTHER CHANGES

    o nj() has been improved and is now about 30% faster.

    o The default option 'drop' of [.DNAbin has been changed to FALSE to
      avoid dropping rownames when selecting a single sequence.

    o print.DNAbin() has been changed to summary.DNAbin() which has been
      removed.



		CHANGES IN APE VERSION 2.5-1


NEW FEATURES

    o The new function stree generates trees with regular shapes.

    o It is now possible to bind two trees with x + y (see ?bind.tree for
      details).

    o drop.tip(), extract.clade(), root(), and bind.tree() now have an
      'interactive' option to make the operation on a plotted tree.

    o cophyloplot() gains two new arguments 'lwd' and 'lty' for the
      association links; they are recycled like 'col' (which wasn't before).


BUG FIXES

    o rTraitDisc() did not use its 'freq' argument correctly (it was
      multiplied with the rate matrix column-wise instead of row-wise).

    o [node|tip|edge]labels(thermo = ) used to draw empty thermometers
      with NA values. Nothing is drawn now like with 'text' or 'pch'.
      The same bug occurred with the 'pie' option.

    o A bug was fixed in compar.ou() and the help page was clarified.

    o bind.tree() has been rewritten fixing several bugs and making it
      more efficient.

    o plot.phylo(type = "p") sometimes failed to colour correctly the
      vertical lines representing the nodes.

    o plot.phylo(direction = "l", x.lim = 30) failed to plot the branches
      in the correct direction though the tip labels were displayed
      correctly.


OTHER CHANGES

    o The c, cbind, and rbind methods for "DNAbin" objetcs now check that
      the sequences are correctly stored (in a list for c, in a matrix
      for the two other functions).



		CHANGES IN APE VERSION 2.5


NEW FEATURES

    o The new function parafit by Pierre Legendre tests for the
      coevolution between hosts and parasites. It has a companion
      function, pcoa, that does principal coordinate decomposition.
      The latter has a biplot method.

    o The new function lmorigin by Pierre Legendre performs multiple
      regression through the origin with testing by permutation.

    o The new functions rTraitCont and rTraitDisc simulate continuous and
      discrete traits under a wide range of evolutionary models.

    o The new function delta.plot does a delta plot following Holland et
      al. (2002, Mol. Biol. Evol. 12:2051).

    o The new function edges draws additional branches between any nodes
      and/or tips on a plotted tree.

    o The new function fancyarrows enhances arrows from graphics with
      triangle and harpoon heads; it can be called from edges().

    o add.scale.bar() has a new option 'ask' to draw interactively.

    o The branch length score replaces the geodesic distance in dist.topo.

    o Three new data sets are included: the gopher-lice data (gopher.D),
      SO2 air pollution in 41 US cities (lmorigin.ex1, from Sokal &
      Rohlf 1995), and some host-parasite specificity data
      (lmorigin.ex2, from Legendre & Desdevises 2009).


BUG FIXES

    o add.scale.bar() drew the bar outside the plotting region with the
      default options with unrooted or radial trees.

    o dist.topo() made R stuck when the trees had different sizes (thanks
      to Otto Cordero for the fix).


OTHER CHANGES

    o The geodesic distance has been replaced by the branch length score
      in dist.topo().



		CHANGES IN APE VERSION 2.4-1


NEW FEATURES

    o rtree() and rcoal() now accept a numeric vector for the 'br'
      argument.

    o vcv() is a new generic function with methods for the classes "phylo"
      and "corPhyl" so that it is possible to calculate the var-cov matrix
      for "transformation models". vcv.phylo() can still be used for trees
      of class "phylo"; its argument 'cor' has been renamed 'corr'.


BUG FIXES

    o bind.tree() failed when 'y' had no root edge.

    o read.nexus() shuffled tip labels when the trees have no branch
      lengths and there is a TRANSLATE block.

    o read.nexus() does not try to translate node labels if there is a
      translation table in the NEXUS file. See ?read.nexus for a
      clarification on this behaviour.

    o plot.multiPhylo() crashed R when plotting a list of trees with
      compressed tip labels.

    o write.nexus() did not translate the taxa names when asked for.

    o plot.phylo(type = "fan") did not rotate the tip labels correctly
      when the tree has branch lengths.

    o ace(type = "continuous", method = "ML") now avoids sigma² being
      negative (which resulted in an error).

    o nj() crashed with NA/NaN in the distance matrix: an error in now
      returned.



		CHANGES IN APE VERSION 2.4


NEW FEATURES

    o base.freq() has a new option 'freq' to return the counts; the
      default is still to return the proportions.


BUG FIXES

    o seg.sites() did not handle ambiguous nucleotides correctly: they
      are now ignored.

    o plot(phy, root.edge = TRUE) failed if there was no $root.edge in
      the tree: the argument is now ignored.

    o add.scale.bar() failed when 'x' and 'y' were given (thanks to Janet
      Young for the fix).


OTHER CHANGES

    o Trying to plot a tree with a single tip now returns NULL with a
      warning (it returned an error previously).

    o The way lines representing nodes are coloured in phylograms has
      been modified (as well as their widths and types) following some
      users' request; this is only for dichotomous nodes.

    o The argument 'adj' in [node][tip][edge]labels() now works when
      using 'pie' or 'thermo'.

    o A more informative message error is now returned by dist.dna() when
      'model' is badly specified (partial matching of this argument is
      done now).

    o Deprecated functions are now listed in a help page: see
      help("ape-defunct") with the quotes.


DEPRECATED & DEFUNCT

    o The functions heterozygosity, nuc.div, theta.h, theta.k and
      theta.s have been moved from ape to pegas.

    o The functions mlphylo, DNAmodel and sh.test have been removed.



		CHANGES IN APE VERSION 2.3-3


BUG FIXES

    o add.scale.bar() always drew a horizontal bar.

    o zoom() shuffled tips with unrooted trees.

    o write.nexus() failed to write correctly trees with a "TipLabel"
      attribute.

    o rcoal() failed to compute branch lengths with very large n.

    o A small bug was fixed in compar.cheverud() (thanks to Michael
      Phelan for the fix).

    o seg.sites() failed when passing a vector.

    o drop.tip() sometimes shuffled tip labels.

    o root() shuffled node labels with 'resolve.root = TRUE'.



		CHANGES IN APE VERSION 2.3-2


BUG FIXES

    o all.equal.phylo() did not compare unrooted trees correctly.

    o dist.topo(... method = "PH85") did not treat unrooted trees
      correctly (thanks to Tim Wallstrom for the fix).

    o root() sometimes failed to test for the monophyly of the
      outgroup correctly.

    o extract.clade() sometimes included too many edges.

    o vcv.phylo() did not work correctly when the tree is in
      "pruningwise" order.

    o nj() did not handle correctly distance matrices with many 0's.
      The code has also been significantly improved: 7, 70, 160 times
      faster with n = 100, 500, 1000, respectively.



		CHANGES IN APE VERSION 2.3-1


NEW FEATURES

    o The new function is.monophyletic tests the monophyly of a group.

    o There is now a c() method for lists of class "DNAbin".

    o yule.cov() now fits the null model, and its help page has been
      corrected with respect to this change.

    o drop.tip() has a new option 'rooted' to force (or not) a tree
      to be treated as (un)rooted.


BUG FIXES

    o dist.gene() failed on most occasions with the default
      pairwise.deletion = FALSE.

    o read.tree() failed to read correctly the tree name(s).

    o boot.phylo() now treats correctly data frames.

    o del.gaps() did not copy the rownames of a matrix.

    o A small bug was fixed in CDAM.global().

    o ace() failed with large data sets. Thanks to Rich FitzJohn for
      the fix. With other improvements, this function is now about 6
      times faster.

    o write.tree() failed with objects of class "multiPhylo".

    o drop.tip(, subtree = TRUE) sometimes shuffled tip labels.


OTHER CHANGES

    o [.multiPhylo and [.DNAbin now respect the original class.

    o Instances of the form class(phy) == "phylo" have been replaced
      by inherits(phy, "phylo").

    o rcoal() is now faster.


DEPRECATED & DEFUNCT

    o klastorin() has been removed.



		CHANGES IN APE VERSION 2.3


NEW FEATURES

    o The new functions CADM.global and CADM.post, contributed by
      Pierre Legendre, test the congruence among several distance
      matrices.

    o The new function yule.time fits a user-defined time-dependent
      Yule model by maximum likelihood.

    o The new function makeNodeLabel creates and/or modifies node
      labels in a flexible way.

    o read.tree() and write.tree() have been modified so that they can
      handle individual tree names.

    o plot.phylo() has a new argument 'edge.lty' that specifies the
      types of lines used for the edges (plain, dotted, dashed, ...)

    o phymltest() has been updated to work with PhyML 3.0.1.


BUG FIXES

    o drop.tip() shuffled tip labels in some cases.

    o drop.tip() did not handle node.label correctly.

    o is.ultrametric() now checks the ordering of the edge matrix.

    o ace() sometimes returned negative values of likelihoods of
      ancestral states (thanks to Dan Rabosky for solving this long
      lasting bug).


OTHER CHANGES

    o The data set xenarthra has been removed.



		CHANGES IN APE VERSION 2.2-4

BUG FIXES

    o The bug fix in read.nexus() in version 2.2-3 was wrong: this is
      now fixed. (Thanks to Peter Wragg for the fix!)

    o A warning message occurred for no reason with ace(method="GLS").


OTHER CHANGES

    o There is now a general help page displayed with '?ape'.



		CHANGES IN APE VERSION 2.2-3


NEW FEATURES

    o The new function extract.clade extracts a clade from a tree by
      specifying a node number or label.

    o fastme.bal() has two new options 'spr' and 'tbr' to perform tree
      operations of the same names.

    o dist.dna() can now return the number of site differences by
      specifying model="N".


BUG FIXES

    o chronopl() did not work with CV = TRUE.

    o read.nexus() did not work correctly in some situations (trees on
      multiple lines with different numbers of lines and/or with
      comments inserted within the trees).

    o ltt.plot(), ltt.lines(), and mltt.plot() did not count correctly
      the number of lineages with non-binary trees.


OTHER CHANGES

    o ape has now a namespace.

    o drop.tip() has been improved: it should be much faster and work
      better in some cases (e.g., see the example in ?zoom).



		CHANGES IN APE VERSION 2.2-2


NEW FEATURES

    o dist.gene() has been substantially improved and gains an option
      'pairwise.deletion'.

    o cbind.DNAbin() has a new option 'fill.with.gaps' and is now
      more flexible.


BUG FIXES

    o prop.part() failed with a single tree with the default option
     'check.labels = TRUE'.

   o summary.DNAbin() failed to display correctly the summary of
     sequence lengths with lists of sequences of 10,000 bases or more
     (because summary.default uses 4 significant digits by default).

   o read.nexus() failed to read a file with a single tree with line
     breaks in the Newick string.

   o del.gaps() returned a list of empty sequences when there were no
     gaps.


OTHER CHANGES

    o phymltest() has been updated for PhyML 3.0 and gains an option
      'append', whereas the option 'path2exec' has been removed.

    o rbind.DNAbin() and cbind.DNAbin() now accept a single matrix
      which is returned unchanged (instead of an error).

    o The data sets bird.orders and bird.families are now stored as
      Newick strings; i.e., the command data(bird.orders) calls
      read.tree().



		CHANGES IN APE VERSION 2.2-1


NEW FEATURES

    o The new function makeLabel() helps to modify labels of trees,
      lists of trees, or DNA sequences, with several utilities to
      truncate and/or make them unique, substituting some
      characters, and so on.

    o The new function del.gaps() removes insertion gaps ("-") in a
      set of DNA sequences.

    o read.dna() can now read Clustal files (*.aln).


BUG FIXES

    o root() failed with 'resolve.root = TRUE' when the root was
      already the specified root.

    o Several bugs were fixed in mlphylo().

    o collapsed.singles() did not propagate the 'Nnode' and
      'node.labels' elements (thanks to Elizabeth Purdom for the fix).

    o read.nexus() failed to remove correctly the comments within
      trees.

    o read.nexus() failed to read a file with a single tree and no
      translation of tip labels.

    o read.nexus() failed to place correctly tip labels when reading
      a single tree with no edge lengths.

    o A bug was fixed in sh.test().


OTHER CHANGES

    o unique.multiPhylo() is faster thanks to a suggestion by Vladimir
      Minin.

    o The option 'check.labels' of consensus() and prop.part() is now
      TRUE by default.

    o write.dna() now does not truncate names to 10 characters with
      the Phylip formats.



		CHANGES IN APE VERSION 2.2


NEW FEATURES

    o Four new functions have been written by Damien de Vienne for the
      graphical exploration of large trees (cophyloplot, subtrees,
      subtreeplot), and to return the graphical coordinates of tree
      (without plotting).

    o The new functions corPagel and corBlomberg implement the Pagel's
      "lambda" and Blomberg et al.'s "ACDC" correlation structures.

    o chronopl() has been improved and gains several options: see its
      help page for details.

    o boot.phylo() has now an option 'trees' to possibly return the
      bootstraped trees (the default is FALSE).

    o prop.part() has been improved and should now be faster in all
      situations.


BUG FIXES

    o read.dna() failed if "?" occurred in the first 10 sites of the
      first sequence.

    o The x/y aspect of the plot is now respected when plotting a
      circular tree (type = "r" or "f").

    o Drawing the tip labels sometimes failed when plotting circular
      trees.

    o zoom() failed when tip labels were used instead of their numbers
      (thanks to Yan Wong for the fix).

    o drop.tip() failed with some trees (fixed by Yan Wong).

    o seg.sites() failed with a list.

    o consensus() failed in some cases. The function has been improved
      as well and is faster.



		CHANGES IN APE VERSION 2.1-3


BUG FIXES

    o A bug in read.nexus() made the Windows R-GUI crash.

    o An error was fixed in the computation of ancestral character
      states by generalized least squares in ace().

    o di2multi() did not modify node labels correctly.

    o multi2di() failed if the tree had its attribute "order" set to
      "cladewise".



		CHANGES IN APE VERSION 2.1-2


NEW FEATURES

    o There three new methods for the "multiPhylo" class: str, $,
      and [[.

    o root() gains the options 'node' and 'resolve.root'
      (FALSE by default) as well as its code being improved.

    o mltt.plot() has now an option 'log' used in the same way
      than in plot.default().


BUG FIXES

    o mltt.plot() failed to display the legend with an unnamed
      list of trees.

    o nodelabels() with pies now correcly uses the argument
      'cex' to draw symbols of different sizes (which has
      worked already for thermometers).

    o read.nexus() generally failed to read very big files.


OTHER CHANGES

    o The argument 'family' of compar.gee() can now be a function
      as well as a character string.

    o read.tree() and read.nexus() now return an unnamed list if
      'tree.names = NULL'.

    o read.nexus() now returns a modified object of class "multiPhylo"
      when there is a TRANSLATE block in the NEXUS file: the individual
      trees have no 'tip.label' vector, but the list has a 'TipLabel'
      attribute. The new methods '$' and '[[' set these elements
      correctly when extracting trees.



		CHANGES IN APE VERSION 2.1-1


NEW FEATURES

    o The new function rmtree generates lists of random trees.

    o rcoal() now generates a genuine coalescent tree by default
      (thanks to Vladimir Minin for the code).


BUG FIXES

    o nuc.div() returned an incorrect value with the default
      pairwise.deletion = FALSE.


OTHER CHANGES

    o The internal codes of bionj(), fastme.bal(), and fastme.ols()
      have been improved so that they are stabler and faster.

    o R packages used by ape are now loaded silently; lattice and gee
      are loaded only when needed.



		CHANGES IN APE VERSION 2.1


NEW FEATURES

    o The new function identify.phylo identifies clades on a plotted
      tree using the mouse.

    o It is now possible to subset a list of trees (object of class
      "multiPhylo") with "[" while keeping its class correct.

    o The new function as.DNAbin.alignment converts DNA sequences
      stored in the "alignment" format of the package seqinr into
      an object of class "DNAbin".

    o The new function weight.taxo2 helps to build similarity matrices
      given two taxonomic levels (usually called by other functions).

    o write.tree() can now take a list of trees (class "multiPhylo")
      as its main argument.

    o plot.correlogram() and plot.correlogramList() have been
      improved, and gain several options (see the help page for
      details). A legend is now plotted by default.


BUG FIXES

    o dist.dna() returned some incorrect values with `model = "JC69"'
      and `pairwise.deletion = TRUE'. This affected only the
      distances involving sequences with missing values. (Thanks
      to Bruno Toupance for digging this bug out.)

    o write.tree() failed with some trees: this is fixed by removing
      the `multi.line' option (trees are now always printed on a
      single line).

    o read.nexus() did not correctly detect trees with multiple root
      edges (see OTHER CHANGES).


OTHER CHANGES

    o The code of mlphylo() has been almost entirely rewritten, and
      should be much stabler. The options have been also greatly
      simplified (see ?mlphylo and ?DNAmodel for details).

    o The internal function nTips has been renamed klastorin_nTips.

    o The code of is.ultrametric() contained redundancies and has
      been cleaned-up.

    o The code of Moran.I() and of correlogram.formula() have been
      improved.

    o read.tree() and read.nexus() now return an error when trying to
      read a tree with multiple root edges (see BUG FIXES). The
      correction applied in previous version did not work in all
      situations.

    o The class c("multi.tree", "phylo") has been renamed
      "multiPhylo".


DOCUMENTATION

    o There is now a vignette in ape: see vignette("MoranI", "ape").


DEPRECATED & DEFUNCT

    o as.matching() and as.phylo.matching() do not support branch
      lengths.

    o correlogram.phylo() and discrete.dist() have been removed.



		CHANGES IN APE VERSION 2.0-2


NEW FEATURES

    o The new function matexpo computes the exponential of a square
      matrix.

    o The new function unique.multi.tree removes duplicate trees from
      a list.

    o yule() has a new option `use.root.edge = FALSE' that specifies
      to ignore, by default, the root edge of the tree if it exists.


BUG FIXES

    o which.edge() failed when the index of a single terminal edge was
      looked for.

    o In diversi.time(), the values returned for model C were
      incorrect.

    o A bug was fixed in yule() that affected the calculation of the
      likelihood in the presence of ties in the branching times.

    o There was a bug in the C function mat_expo4x4 affecting the
      calculations of the transition probabilities for models HKY and
      GTR in mlphylo().

    o A small bug was fixed in as.matrix.DNAbin (thanks to James
      Bullard).

    o rtree() did not `shuffle' the tip labels by default, so only a
      limited number of labelled topologies could be generated.



		CHANGES IN APE VERSION 2.0-1


NEW FEATURES

    o The three new functions bionj, fastme.ols, and fastme.bal
      perform phylogeny estimation by the BIONJ and fastME methods in
      OLS and balanced versions. This is a port to R of previous
      previous programs done by Vincent Lefort.

    o The new function chronoMPL performs molecular dating with the
      mean path lengths method of Britton et al. (2002, Mol. Phyl.
      Evol. 24: 58).

    o The new function rotate, contributed by Christoph Heibl, swaps
      two clades connected to the same node. It works also with
      multichotomous nodes.

    o The new `method' as.matrix.DNAbin() may be used to convert
      easily DNA sequences stored in a list into a matrix while
      keeping the names and the class.


BUG FIXES

    o chronopl() failed when some branch lengths were equal to zero:
      an error message is now returned.

    o di2multi() failed when there was a series of consecutive edges
      to remove.



		CHANGES IN APE VERSION 1.10-2


NEW FEATURES

    o plot.phylo() can now plot circular trees: the option is type =
      "fan" or type = "f" (to avoid the ambiguity with type = "c").

    o prop.part() has a new option `check.labels = FALSE' which allows
      to considerably speed-up the calculations of bipartitions. As a
      consequence, calculations of bootstrap values with boot.phylo()
      should be much faster.


BUG FIXES

    o read.GenBank() did not return correctly the list of species as
      from ape 1.10: this is fixed in this version

    o Applying as.phylo() on a tree of class "phylo" failed: the
      object is now returned unchanged.



		CHANGES IN APE VERSION 1.10-1


NEW FEATURES

    o The three new functions Ntip, Nnode, and Nedge return, for a
      given tree, the number of tips, nodes, or edges, respectively.


BUG FIXES

    o read.nexus() did not set correctly the class of the returned
      object when reading multiple trees.

    o mllt.plot() failed with objects of class c("multi.tree",
      "phylo").

    o unroot() did not work correctly in most cases.

    o reorder.phylo() made R freeze in some occasions.

    o Plotting a tree in pruningwise order failed.

    o When plotting an unrooted tree, the tip labels where not all
      correctly positioned if the option `cex' was used.



		CHANGES IN APE VERSION 1.10


NEW FEATURES

    o Five new `method' functions have been introduced to manipulate
      DNA sequences in binary format (see below).

    o Three new functions have been introduced to convert between the
      new binary and the character formats.

    o The new function as.alignment converts DNA sequences stored as
      single characters into the class "alignment" used by the package
      seqinr.

    o read.dna() and read.GenBank() have a new argument `as.character'
      controlling whether the sequences are returned in binary format
      or as character.


BUG FIXES

    o root() failed when the tree had node labels: this is fixed.

    o plot.phylo() did not correctly set the limits on the y-axis with
      the default setting: this is fixed.

    o dist.dna() returned a wrong result for the LogDet, paralinear,
      and BH87 models with `pairwise.deletion = TRUE'.


OTHER CHANGES

    o DNA sequences are now internally stored in a binary format. See
      the document "A Bit-Level Coding Scheme for Nucleotides" for the
      details. Most functions analyzing DNA functions have been
      modified accordingly and are now much faster (dist.dna is now
      ca. 60 times faster).



		CHANGES IN APE VERSION 1.9-4


BUG FIXES

    o A bug was fixed in edgelabels().

    o as.phylo.hclust() did not work correctly when the object of
      class "hclust" has its labels set to NULL: the returned tree has
      now its tip labels set to "1", "2", ...

    o consensus could fail if some tip labels are a subset of others
      (e.g., "a" and "a_1"): this is now fixed.

    o mlphylo() failed in most cases if some branch lengths of the
      initial tree were greater than one: an error message is now
      issued.

    o mlphylo() failed in most cases when estimating the proportion of
      invariants: this is fixed.



      		CHANGES IN APE VERSION 1.9-3


NEW FEATURES

    o The new function edgelabels adds labels on the edge of the tree
      in the same way than nodelabels or tiplabels.


BUG FIXES

    o multi2di() did not handle correctly branch lengths with the
      default option `random = TRUE': this is now fixed.

    o A bug was fixed in nuc.div() when using pairwise deletions.

    o A bug occurred in the analysis of bipartitions with large
      numbers of large trees, with consequences on prop.part,
      prop.clades, and boot.phylo.

    o The calculation of the Billera-Holmes-Vogtmann distance in
      dist.topo was wrong: this has been fixed.



		CHANGES IN APE VERSION 1.9-2


NEW FEATURES

    o The new function ladderize reorganizes the internal structure of
      a tree to plot them left- or right-ladderized.

    o The new function dist.nodes computes the patristic distances
      between all nodes, internal and terminal, of a tree. It replaces
      the option `full = TRUE' of cophenetic.phylo (see below).


BUG FIXES

    o A bug was fixed in old2new.phylo().

    o Some bugs were fixed in chronopl().

    o The edge colours were not correctly displayed by plot.phylo
      (thank you to Li-San Wang for the fix).

    o cophenetic.phylo() failed with multichotomous trees: this is
      fixed.


OTHER CHANGES

    o read.dna() now returns the sequences in a matrix if they are
      aligned (interleaved or sequential format). Sequences in FASTA
      format are still returned in a list.

    o The option `full' of cophenetic.phylo() has been removed because
      it could not be used from the generic.


DEPRECATED & DEFUNCT

    o rotate() has been removed; this function did not work correctly
      since ape 1.9.



		CHANGES IN APE VERSION 1.9-1


BUG FIXES

    o Trees with a single tip were not read correctly in R as the
      element `Nnode' was not set: this is fixed.

    o unroot() did not set correctly the number of nodes of the
      unrooted tree in most cases.

    o read.GenBank() failed when fetching very long sequences,
      particularly of the BX-series.

    o A bug was introduced in read.tree() with ape 1.9: it has been
      fixed



		CHANGES IN APE VERSION 1.9


NEW FEATURES

    o There are two new print `methods' for trees of class "phylo" and
      lists of trees of class "multi.tree", so that they are now
      displayed in a compact and informative way.

    o There are two new functions, old2new.phylo and new2old.phylo,
      for converting between the old and new coding of the class
      "phylo".

    o dist.dna() has three new models: Barry and Hartigan ("BH87"),
      LogDet ("logdet"), and paralinear ("paralin").

    o compute.brlen() has been extended: several methods are now
      available to compute branch lengths.

    o write.dna() can now handle matrices as well as lists.


BUG FIXES

    o cophenetic.phylo() sometimes returned a wrong result with
      multichotomous trees: this is fixed.

    o rotate() failed when a single tip was specified: the tree is now
      returned unchanged.

    o ace() did not return the correct index matrix with custom
      models: this is fixed.

    o multi2di() did not work correctly when resolving multichotomies
      randomly: the topology was always the same, only the arrangement
      of clades was randomized: this is fixed. This function now
      accepts trees with no branch lengths.

    o The output of diversi.gof() was blurred by useless prints when a
      user distribution was specified. This has been corrected, and
      the help page of this function has been expanded.


OTHER CHANGES

    o The internal structure of the class "phylo" has been changed:
      see the document "Definition of Formats for Coding Phylogenetic
      Trees in R" for the details. In addition, the code of most
      functions has been improved.

    o Several functions have been improved by replacing some R codes
      by C codes: pic, plot.phylo, and reorder.phylo.

    o There is now a citation information: see citation("ape") in R.

    o write.tree() now does not add extra 0's to branch lengths so
      that 1.23 is printed "1.23" by default, not "1.2300000000".

    o The syntax of bind.tree() has been simplified. This function now
      accepts trees with no branch lengths, and handles correctly node
      labels.

    o The option `as.numeric' of mrca() has been removed.

    o The unused options `format' and `rooted' of read.tree() have
      been removed.

    o The unused option `format' of write.tree() has been removed.

    o The use of node.depth() has been simplified.



		CHANGES IN APE VERSION 1.8-5


NEW FEATURES

    o Two new functions read.nexus.data() and write.nexus.data(),
      contributed by Johan Nylander, allow to read and write molecular
      sequences in NEXUS files.

    o The new function reorder.phylo() reorders the internal structure
      of a tree of class "phylo". It is used as the generic, e.g.,
      reorder(tr).

    o read.tree() and read.nexus() can now read trees with a single
      edge.

    o The new data set `cynipids' supplies a set of protein sequences
      in NEXUS format.


BUG FIXES

    o The code of all.equal.phylo() has been completely rewritten
      (thanks to Benoît Durand) which fixes several bugs.

    o read.tree() and read.nexus() now checks the labels of the tree
      to remove or substitute any characters that are illegal in the
      Newick format (parentheses, etc.)

    o A negative P-value could be returned by mantel.test(): this is
      now fixed.



		CHANGES IN APE VERSION 1.8-4


NEW FEATURES

    o The new function sh.test() computes the Shimodaira-
      Hasegawa test.

    o The new function collapse.singles() removes the nodes with a
      single descendant from a tree.

    o plot.phylo() has a new argument `tip.color' to specify the
      colours of the tips.

    o mlphylo() has now an option `quiet' to control the display of
      the progress of the analysis (the default is FALSE).


BUG FIXES

    o read.dna() did not read correctly sequences in sequential format
      with leading alignment gaps "-": this is fixed.

    o ace() returned a list with no class so that the generic
      functions (anova, logLik, ...) could not be used directly. This
      is fixed as ace() now returns an object of class "ace".

    o anova.ace() had a small bug when computing the number of degrees
      of freedom: this is fixed.

    o mlphylo() did not work when the sequences were in a matrix or
      a data frame: this is fixed.

    o rtree() did not work correctly when trying to simulate an
      unrooted tree with two tips: an error message is now issued.


OTHER CHANGES

    o The algorithm of rtree() has been changed: it is now about 40,
      100, and 130 times faster for 10, 100, and 1000 tips,
      respectively.



		CHANGES IN APE VERSION 1.8-3


NEW FEATURES

    o There are four new `method' functions to be used with the
      results of ace(): logLik(), deviance(), AIC(), and anova().

    o The plot method of phymltest has two new arguments: `main' to
      change the title, and `col' to control the colour of the
      segments showing the AIC values.

    o ace() has a new argument `ip' that gives the initial values used
      in the ML estimation with discrete characters (see the examples
      in ?ace). This function now returns a matrix giving the indices
      of the estimated rates when analysing discrete characters.

    o nodelabels() and tiplabels() have a new argument `pie' to
      represent proportions, with any number of categories, as
      piecharts. The use of the option `thermo' has been improved:
      there is now no limitation on the number of categories.


BUG FIXES

    o mlphylo() did not work with more than two partitions: this is
      fixed.

    o root() failed if the proposed outgroup was already an outgroup
      in the tree: this is fixed.

    o The `col' argument in nodelabels() and tiplabels() was not
      correctly passed when `text' was used: this is fixed.

    o Two bugs were fixed in mlphylo(): parameters were not always
      correctly output, and the estimation failed in some cases.

    o plot.phylo() was stuck when given a tree with a single tip: this
      is fixed and a message error is now returned.

    o An error was corrected in the help page of gammaStat regarding
      the calculation of P-values.

    o Using gls() could crash R when the number of species in the tree
      and in the variables were different: this is fixed.



    		CHANGES IN APE VERSION 1.8-2


NEW FEATURES

    o The new function mlphylo() fits a phylogenetic tree by maximum
      likelihood from DNA sequences. Its companion function DNAmodel()
      is used to define the substitution model which may include
      partitioning. There are methods for logLik(), deviance(), and
      AIC(), and the summary() method has been extended to display in
      a friendly way the results of this model fitting. Currently, the
      functionality is limited to estimating the substitution and
      associated parameters and computing the likelihood.

    o The new function drop1.compar.gee (used as, e.g., drop1(m))
      tests for single effects in GEE-based comparative method. A
      warning message is printed if there is not enough degrees of
      freedom.


BUG FIXES

    o An error message was sometimes issued by plot.multi.tree(),
      though with no consequence.



		CHANGES IN APE VERSION 1.8-1


NEW FEATURES

    o There is a new plot method for lists of trees (objects of class
      "multi.tree"): it calls plot.phylo() internally and is
      documented on the same help page.


BUG FIXES

    o A bug was fixed in the C code that analyzes bipartitions: this
      has impact on several functions like prop.part, prop.clades,
      boot.phylo, or consensus.

    o root() did not work correctly when the specified outgroup had
      more than one element: this is fixed.

    o dist.dna() sometimes returned a warning inappropriately: this
      has been corrected.

    o If the distance object given to nj() had no rownames, nj()
      returned a tree with no tip labels: it now returns tips labelled
      "1", "2", ..., corresponding to the row numbers.


OTHER CHANGES

    o nj() has been slightly changed so that tips with a zero distance
      are first aggregated with zero-lengthed branches; the usual NJ
      procedure is then performed on a distance matrix without 0's.



		CHANGES IN APE VERSION 1.8


NEW FEATURES

    o The new function chronopl() estimates dates using the penalized
      likelihood method by Sanderson (2002; Mol. Biol. Evol., 19:101).

    o The new function consensus() calculates the consensus tree of a
      list of trees.

    o The new function evolve.phylo() simulates the evolution of
      continuous characters along a phylogeny under a Brownian model.

    o The new plot method for objects of class "ancestral" displays a
      tree together with ancestral values, as returned by the above
      function.

    o The new function as.phylo.formula() returns a phylogeny from a
      set of nested taxonomic variables given as a formula.

    o The new function read.caic() reads trees in CAIC format.

    o The new function tiplabels() allows to add labels to the tips
      of a tree using text or plotting symbols in a flexible way.

    o The new function unroot() unroots a phylogeny.

    o multi2di() has a new option, `random', which specifies whether
      to resolve the multichotomies randomly (the default) or not.

    o prop.part() now returns an object of class "prop.part" for which
      there are print (to display a partition in a more friendly way)
      and summary (to extract the numbers) methods.

    o plot.phylo() has a new option, `show.tip.label', specifying
      whether to print the labels of the tips. The default is TRUE.

    o The code of nj() has been replaced by a faster C code: it is now
      about 10, 25, and 40 times faster for 50, 100, and 200 taxa,
      respectively.

    o write.nexus() now writes whether a tree is rooted or not.


BUG FIXES

    o Two bugs have been fixed in root(): unrooted trees are now
      handled corretly, and node labels are now output normally.

    o A bug was fixed in phymltest(): the executable couldn't be found
      in some cases.

    o Three bugs have been fixed in ace(): computing the likelihood of
      ancestral states of discrete characters failed, custom models
      did not work, and the function failed with a null gradient (a
      warning message is now returned; this latter bug was also
      present in yule.cov() as well and is now fixed).

    o pic() hanged out when missing data were present: an error is now
      returned.

    o A small bug was fixed in dist.dna() where the gamma correction
      was not always correctly dispatched.

    o plot.phylo() plotted correctly the root edge only when the tree
      was plotted rightwards: this works now for all directions.


OTHER CHANGES

    o dist.taxo() has been renamed as weight.taxo().

    o dist.phylo() has been replaced by the method cophenetic.phylo().

    o Various error and warning messages have been improved.



		CHANGES IN APE VERSION 1.7
NEW FEATURES

    o The new function ace() estimates ancestral character states for
      continuous characters (with ML, GLS, and contrasts methods), and
      discrete characters (with ML only) for any number of states.

    o The new function compar.ou() fits the Ornstein-Uhlenbeck model
      of directional evolution for continuous characters. The user
      specifies the node(s) of the tree where the character optimum
      changes.

    o The new function is.rooted() tests whether a tree (of class
      "phylo") is rooted.

    o The new function rcoal() generates random ultrametric trees with
      the possibility to specify the function that generates the
      inter-nodes distances.

    o The new function mrca() gives for all pairs of tips in a tree
      (and optionally nodes too) the most recent common ancestor.

    o nodelabels() has a new option `thermo' to plot proportions (up
      to three classes) on the nodes of a tree.

    o rtree() has been improved: it can now generate rooted or
      unrooted trees, and the mathematical function that generates the
      branch lengths may be specified by the user. The tip labels may
      be given directly in the call to rtree. The limit cases (n = 2,
      3) are now handled correctly.

    o dist.topo() has a new argument `method' with two choices: "PH85"
      for Penny and Henny's method (already available before and now
      the default), and "BHV01" for the geometric distance by Billera
      et al. (2001, Adv. Appl. Math. 27:733).

    o write.tree() has a new option, `digits', which specifies the
      number of digits to be printed in the Newick tree. By default
      digits = 10. The numbers are now always printed in decimal form
      (i.e., 1.0e-1 is now avoided).

    o dist.dna() can now compute the raw distances between pairs of
      DNA sequences by specifying model = "raw".

    o dist.phylo() has a new option `full' to possibly compute the
      distances among all tips and nodes of the tree. The default is
      `full = FALSE'.


BUG FIXES

    o Several bugs were fixed in all.equal.phylo().

    o dist.dna() did not handle correctly gaps ("-") in alignments:
      they are now considered as missing data.

    o rotate() did not work if the tips were not ordered: this is
      fixed.

    o mantel.test() returned NA in some special cases: this is fixed
      and the function has been improved and is now faster.

    o A bug was fixed in diversi.gof() where the calculation of A² was
      incorrect.

    o cherry() did not work correctly under some OSs (mainly Linux):
      this is fixed.

    o is.binary.tree() has been modified so that it works with both
      rooted and unrooted trees.

    o The documentation of theta.s() was not correct: this has been
      fixed.

    o plot.mst() did not work correctly: this is fixed.



		CHANGES IN APE VERSION 1.6


NEW FEATURES

    o The new function dist.topo() computes the topological distances
      between two trees.

    o The new function boot.phylo() performs a bootstrap analysis on
      phylogeny estimation.

    o The new functions prop.part() and prop.clades() analyse
      bipartitions from a series of trees.


OTHER CHANGES

    o read.GenBank() now uses the EFetch utility of NCBI instead of
      the usual Web interface: it is now much faster (e.g., 12 times
      faster to retrieve 8 sequences, 37 times for 60 sequences).


BUG FIXES

    o Several bugs were fixed in read.dna().

    o Several bugs were fixed in diversi.time().

    o is.binary.tree() did not work correctly if the tree has no edge
      lengths: this is fixed.

    o drop.tip() did not correctly propagated the `node.label' of a
      tree: this is fixed.



 		CHANGES IN APE VERSION 1.5


NEW FEATURES

    o Two new functions, as.matching.phylo() and as.phylo.matching(),
      convert objects between the classes "phylo" and "matching". The
      latter implements the representation of binary trees introduced by
      Diaconis and Holmes (1998; PNAS 95:14600). The generic function
      as.matching() has been introduced as well.

    o Two new functions, multi2di() and di2multi(), allow to resolve
      and collapse multichotomies with branches of length zero.

    o The new function nuc.div() computes the nucleotide diversity
      from a sample a DNA sequences.

    o dist.dna() has been completely rewritten with a much faster
      (particularly for large data sets) C code. Eight models are
      available: JC69, K80, F81, K81, F84, T92, TN93, and GG95 (the
      option `method' has been renamed `model'). Computation of variance
      is available for all models. A gamma-correction is possible for
      JC69, K80, F81, and TN93. There is a new option, pairwise.deletion,
      to remove sites with missing data on a pairwise basis. The option
      `GCcontent' has been removed.

    o read.GenBank() has a new option (species.names) which specifies
      whether to return the species names of the organisms in addition
      to the accession numbers of the sequences (this is the default
      behaviour).

    o write.nexus() can now write several trees in the same NEXUS file.

    o drop.tip() has a new option `root.edge' that allows to specify the
      new root edge if internal branches are trimmed.


BUG FIXES

    o as.phylo.hclust() failed if some labels had parentheses: this
      is fixed.

    o Several bugs were fixed in all.equal.phylo(). This function now
      returns the logical TRUE if the trees are identical but with
      different representations (a report was printed previously).

    o read.GenBank() did not correctly handle ambiguous base codes:
      this is fixed.


OTHER CHANGES

    o birthdeath() now returns an object of class "birthdeath" for
      which there is a print method.



		CHANGES IN APE VERSION 1.4


NEW FEATURES

    o The new function nj() performs phylogeny estimation with the
      neighbor-joining method of Saitou and Nei (1987; Mol. Biol.
      Evol., 4:406).

    o The new function which.edge() identifies the edges of a tree
      that belong to a group specified as a set of tips.

    o The new function as.phylo.phylog() converts an object of class
      "phylog" (from the package ade4) into an object of class
      "phylo".

    o The new function axisPhylo() draws axes on the side of a
      phylogeny plot.

    o The new function howmanytrees() calculates the number of trees
      in different cases and giving a number of tips.

    o write.tree() has a new option `multi.line' (TRUE by default) to
      write a Newick tree on several lines rather than on a single
      line.

    o The functionalities of zoom() have been extended. Several
      subtrees can be visualized at the same time, and they are marked
      on the main tree with colors. The context of the subtrees can be
      marked with the option `subtree' (see below).

    o drop.tip() has a new option `subtree' (FALSE by default) which
      specifies whether to output in the tree how many tips have been
      deleted and where.

    o The arguments of add.scale.bar() have been redefined and have
      now default values (see ?add.scale.bar for details). This
      function now works even if the plotted tree has no edge length.

    o plot.phylo() can now plot radial trees, but this does not take
      edge lengths into account.

    o In plot.phylo() with `type = "phylogram"', if the values of
      `edge.color' and `edge.width' are identical for sister-branches,
      they are propagated to the vertical line that link them.


BUG FIXES

    o Repeated calls to as.phylo.hclust() or as.hclust.phylo() made R
      crashing. This is fixed.

    o In plot.phylo(), the options `edge.color' and `edge.width' are
      now properly recycled; their default values are now "black" and
      1, respectively.

    o A bug has been fixed in write.nexus().


OTHER CHANGES

    o The function node.depth.edgelength() has been removed and
      replaced by a C code.



		CHANGES IN APE VERSION 1.3-1


NEW FEATURES

    o The new function nodelabels() allows to add labels to the nodes
      of a tree using text or plotting symbols in a flexible way.

    o In plot.phylo() the arguments `x.lim' and `y.lim' can now be two
      numeric values specifying the lower and upper limits on the x-
      and y-axes. This allows to leave some space on any side of the
      tree. If a single value is given, this is taken as the upper
      limit (as before).



		CHANGES IN APE VERSION 1.3


NEW FEATURES

    o The new function phymltest() calls the software PHYML and fits
      28 models of DNA sequence evolution. There are a print method to
      display likelihood and AIC values, a summary method to compute
      the hierarchical likelihood ratio tests, and a plot method to
      display graphically the AIC values of each model.

    o The new function yule.cov() fits the Yule model with covariates,
      a model where the speciation rate is affected by several species
      traits through a generalized linear model. The parameters are
      estimated by maximum likelihood.

    o Three new functions, corBrownian(), corGrafen(), and
      corMartins(), compute the expected correlation structures among
      species given a phylogeny under different models of evolution.
      These can be used for GLS comparative phylogenetic methods (see
      the examples). There are coef() and corMatrix() methods and an
      Initialize.corPhyl() function associated.

    o The new function compar.cheverud() implements Cheverud et al.'s
      (1985; Evolution 39:1335) phylogenetic comparative method.

    o The new function varcomp() estimates variance components; it has
      a plot method.

    o Two new functions, panel.superpose.correlogram() and
      plot.correlogramList(), allow to plot several phylogenetic
      correlograms.

    o The new function node.leafnumber() computes the number of leaves
      of a subtree defined by a particular node.

    o The new function node.sons() gets all tags of son nodes from a
      given parent node.

    o The new function compute.brlen() computes the branch lengths of
      a tree according to a specified method.

    o plot.phylo() has three new options: "cex" controls the size of
      the (tip and node) labels (thus it is no more needed to change
      the global graphical parameter), "direction" which allows to
      plot the tree rightwards, leftwards, upwards, or downwards, and
      "y.lim" which sets the upper limit on the y-axis.


BUG FIXES

    o Some functions which try to match tip labels and names of
      additional data (e.g. vector) are likely to fail if there are
      typing or syntax errors. If both series of names do not perfectly
      match, they are ignored and a warning message is now issued.
      These functions are bd.ext, compar.gee, pic. Their help pages
      have been clarified on this point.



		CHANGES IN APE VERSION 1.2-7


NEW FEATURES

    o The new function root() reroots a phylogenetic tree with respect
      to a specified outgroup.

    o The new function rotate() rotates an internal branch of a tree.

    o In plot.phylo(), the new argument "lab4ut" (labels for unrooted
      trees) controls the display of the tip labels in unrooted trees.
      This display has been greatly improved: the tip labels are now not
      expected to overlap with the tree (particularly if lab4ut =
      "axial"). In all cases, combining appropriate values of "lab4ut"
      and the font size (via "par(cex = )") should result in readable
      unrooted trees. See ?plot.phylo for some examples.

    o In drop.tip(), the argument `tip' can now be numeric or character.


BUG FIXES

    o drop.tip() did not work correctly with trees with no branch
      lengths: this is fixed.

    o A bug in plot.phylo(..., type = "unrooted") made some trees being
      plotted with some line crossings: this is now fixed.



		CHANGES IN APE VERSION 1.2-6


NEW FEATURES

    o Six new functions (Moran.I, correlogram.formula, discrete.dist,
      correlogram.phylo, dist.taxo, plot.correlogram) have been added
      to implement comparative methods with an autocorrelation approach.

    o A new data set describing some life history traits of Carnivores
      has been included.


BUG FIXES

    o A fix was made on mcmc.popsize() to conform to R 2.0.0.


OTHER CHANGES

    o When plotting a tree with plot.phylo(), the new default of the
      option `label.offset' is now 0, so the labels are always visible.



		CHANGES IN APE VERSION 1.2-5


NEW FEATURES

    o The new function bd.ext() fits a birth-death model with combined
      phylogenetic and taxonomic data, and estimates the corresponding
      speciation and extinction rates.


OTHER CHANGES

    o The package gee is no more required by ape but only suggested
      since only the function compar.gee() calls gee.



		CHANGES IN APE VERSION 1.2-4


NEW FEATURES

    o Four new functions (mcmc.popsize, extract.popsize, plot.popsize,
      and lines.popsize) implementing a new approach for inferring the
      demographic history from genealogies using a reversible jump
      MCMC have been introduced.

    o The unit of time in the skyline plot and in the new plots can
      now be chosen to be actual years, rather than substitutions.



		CHANGES IN APE VERSION 1.2-3


NEW FEATURES

    o The new function rtree() generates a random binary tree with or
      without branch lengths.

    o Two new functions for drawing lineages-through-time (LTT) plots
      are provided: ltt.lines() adds a LTT curve to an existing plot,
      and mltt.plot() does a multiple LTT plot giving several trees as
      arguments (see `?ltt.plot' for details).


BUG FIXES

    o Some taxon names made R crashing when calling as.phylo.hclust():
      this is fixed.

    o dist.dna() returned an error with two identical DNA sequences
      (only using the Jukes-Cantor method returned 0): this is fixed.


OTHER CHANGES

    o The function dist.phylo() has been re-written using a different
      algorithm: it is now about four times faster.

    o The code of branching.times() has been improved: it is now about
      twice faster.



		CHANGES IN APE VERSION 1.2-2


NEW FEATURES

    o The new function seg.sites() finds the segregating sites in a
      sample of DNA sequences.


BUG FIXES

    o A bug introduced in read.tree() and in read.nexus() with version
      1.2-1 was fixed.

    o A few errors were corrected and a few examples were added in the
      help pages.



		CHANGES IN APE VERSION 1.2-1


NEW FEATURES

    o plot.phylo() can now draw the edge of the root of a tree if it
      has one (see the new option `root.edge', its default is FALSE).


BUG FIXES

    o A bug was fixed in read.nexus(): files with semicolons inside
      comment blocks were not read correctly.

    o The behaviour of read.tree() and read.nexus() was corrected so
      that tree files with badly represented root edges (e.g., with
      an extra pair of parentheses, see the help pages for details)
      are now correctly represented in the object of class "phylo";
      a warning message is now issued.



		CHANGES IN APE VERSION 1.2


NEW FEATURES

    o plot.phylo() has been completely re-written and offers several
      new functionalities. Three types of trees can now be drawn:
      phylogram (as previously), cladogram, and unrooted tree; in
      all three types the branch lengths can be drawn using the edge
      lengths of the phylogeny or not (e.g., if the latter is absent).
      The vertical position of the nodes can be adjusted with two
      choices (see option `node.pos'). The code has been re-structured,
      and two new functions (potentially useful for developpers) are
      documented separately: node.depth.edgelength() and node.depth();
      see the respective help pages for details.

    o The new function zoom() allows to explore very large trees by
      focusing on a small portion of it.

    o The new function yule() fits by maximum likelihood the Yule model
      (birth-only process) to a phylogenetic tree.

    o Support for writing DNA sequences in FASTA format has been
      introduced in write.dna() (support for reading sequences in
      this format was introduced in read.dna() in version 1.1-2).
      The function has been completely re-written, fixing some bugs
      (see below); the default behaviour is no more to display the
      sequences on the standard output. Several options have been
      introduced to control the sequence printing in a flexible
      way. The help page has been extended.

    o A new data set is included: a supertree of bats in NEXUS format.


BUG FIXES

    o In theta.s(), the default of the option `variance' has
      been changed to `FALSE' (as was indicated in the help page).

    o Several bugs were fixed in the code of all.equal.phylo().

    o Several bugs were fixed in write.dna(), particularly this
      function did not work with `format = "interleaved"'.

    o Various errors were corrected in the help pages.


OTHER CHANGES

    o The argument names of as.hclust.phylo() have been changed
      from "(phy)" to "(x, ...)" to conform to the definition of
      the corresponding generic function.

    o gamma.stat() has been renamed gammaStat() to avoid confusion
      since gamma() is a generic function.



		CHANGES IN APE VERSION 1.1-3


BUG FIXES

    o base.freq() previously did not return a value of 0 for
      bases absent in the data (e.g., a vector of length 3 was
      returned if one base was absent). This is now fixed (a
      vector of length 4 is always returned).

    o Several bugs were fixed in read.nexus(), including that this
      function did not work in this absence of a "TRANSLATE"
      command in the NEXUS file, and that the commands were
      case-sensitive.



		CHANGES IN APE VERSION 1.1-2


NEW FEATURES

    o The Tamura and Nei (1993) model of DNA distance is now implemented
      in dist.dna(): five models are now available in this function.

    o A new data set is included: a set of 15 sequences of the
      cytochrome b mitochondrial gene of the woodmouse (Apodemus
      sylvaticus).


BUG FIXES

    o A bug in read.nexus() was fixed.

    o read.dna() previously did not work correctly in most cases.
      The function has been completely re-written and its help page
      has been considerably extended (see ?read.dna for details).
      Underscores (_) in taxon names are no more replaced with
      spaces (this behaviour was undocumented).

    o A bug was fixed in write.dna().



		CHANGES IN APE VERSION 1.1-1


BUG FIXES

    o A bug in read.tree() introduced in APE 1.1 was fixed.

    o A bug in compar.gee() resulted in an error when trying to fit
      a model with `family = "binomial"'. This is now fixed.



 		CHANGES IN APE VERSION 1.1


NEW FEATURES

    o The Klastorin (1982) method as suggested by Misawa and Tajima
      (2000, Mol. Biol. Evol. 17:1879-1884) for classifying genes
      on the basis of phylogenetic trees has been implemented (see
      the function klastorin()).

    o Functions have been added to convert APE's "phylo" objects in
      "hclust" cluster objects and vice versa (see the help page of
      as.phylo for details).

    o Three new functions, ratogram(), chronogram() and NPRS.criterion(),
      are introduced for the estimation of absolute evolutionary rates
      (ratogram) and dated clock-like trees (chronogram) from
      phylogenetic trees using the non-parametric rate smoothing approach
      by MJ Sanderson (1997, Mol. Biol. Evol. 14:1218-1231).

    o A summary method is now provided printing a summary information on a
      phylogenetic tree with, for instance, `summary(tree)'.

    o The behaviour of read.tree() was changed so that all spaces and
      tabulations in tree files are now ignored. Consequently, spaces in tip
      labels are no more allowed. Another side effect is that read.nexus()
      now does not replace the underscores (_) in tip labels with spaces
      (this behaviour was undocumented).

    o The function plot.phylo() has a new option (`underscore') which
      specifies whether the underscores in tip labels should be written on
      the plot as such or replaced with spaces (the default).

    o The function birthdeath() now computes 95% confidence intervals of
      the estimated parameters using profile likelihood.

    o Three new data sets are included: a gene tree estimated from 36
      landplant rbcL sequences, a gene tree estimated from 32 opsin
      sequences, and a gene tree for 50 BRCA1 mammalian sequences.


BUG FIXES

    o A bug was fixed in dist.gene() where nothing was returned.

    o A bug in plot.mst() was fixed.

    o A bug in vcv.phylo() resulted in false correlations when the
      option `cor = TRUE' was used (now fixed).



		CHANGES IN APE VERSION 1.0


NEW FEATURES

    o Two new functions, read.dna() and write.dna(), read/write in a file
      DNA sequences in interleaved or in sequential format.

    o Two new functions, read.nexus() and write.nexus(), read/write trees
      in a NEXUS file.

    o The new function bind.tree() allows to bind two trees together,
      possibly handling root edges to give internal branches.

    o The new function drop.tip() removes the tips in a phylogenetic tree,
      and trims (or not) the corresponding internal branches.

    o The new function is.ultrametric() tests if a tree is ultrametric.

    o The function plot.phylo() has more functionalities such as drawing the
      branches with different colours and/or different widths, showing the
      node labels, controling the position and font of the labels, rotating
      the labels, and controling the space around the plot.

    o The function read.tree() can now read trees with no branch length,
      such as "(a,b),c);". Consequently, the element `edge.length' in
      objects of class "phylo" is now optional.

    o The function write.tree() has a new default behaviour: if the default
      for the option `file' is used (i.e. file = ""), then a variable of
      mode character containing the tree in Newick format is returned which
      can thus be assigned (e.g., tree <- write.tree(phy)).

    o The function read.tree() has a new argument `text' which allows
      to read the tree in a variable of mode character.

    o A new data set is included: the phylogenetic relationships among
      the orders of birds from Sibley and Ahlquist (1990).



		CHANGES IN APE VERSION 0.2-1


BUG FIXES

    o Several bugs were fixed in the help pages.



		CHANGES IN APE VERSION 0.2


NEW FEATURES

    o The function write.tree() writes phylogenetic trees (objects of class
      "phylo") in an ASCII file using the Newick parenthetic format.

    o The function birthdeath() fits a birth-death model to branching times
      by maximum likelihood, and estimates the corresponding speciation and
      extinction rates.

    o The function scale.bar() adds a scale bar to a plot of a phylogenetic
      tree.

    o The function is.binary.tree() tests whether a phylogeny is binary.

    o Two generic functions, coalescent.intervals() and collapsed.intervals(),
      as well as some methods are introduced.

    o Several functions, including some generics and methods, for computing
      skyline plot estimates (classic and generalized) of effective
      population size through time are introduced and replace the function
      skyline.plot() in version 0.1.

    o Two data sets are now included: the phylogenetic relationships among
      the families of birds from Sibley and Ahlquist (1990), and an
      estimated clock-like phylogeny of HIV sequences sampled in the
      Democratic Republic of Congo.


DEPRECATED & DEFUNCT

    o The function skyline.plot() in ape 0.1 has been deprecated and
      replaced by more elaborate functions (see above).


BUG FIXES

    o Two important bugs were fixed in plot.phylo(): phylogenies with
      multichotomies not at the root or not with only terminal branches,
      and phylogenies with a single node (i.e. only terminal branches)
      did not plot. These trees should be plotted correctly now.

    o Several bugs were fixed in diversi.time() in the computation of
      AICs and LRTs.

    o Various errors were corrected in the help pages.