The age of major monocot groups inferred from 800+ rbcL sequences

Abstract

Phylogenetic research on monocots has been extraordinarily active over the past years. With the familial interre-lationships being sufficiently understood, the question of divergence times and crown node ages of major lineages comes into focus. In this study we present the first attempt to estimate crown and stem node ages for most orders and families of monocots, based on rbcL sequence data and comprehensive taxon sampling. From our analysis it is obvious that considerable monocot diversification took place during the Early Cretaceous, with most families already present at the Cretaceous–Tertiary boundary. Araceae, Arecaceae and Orchidaceae are among the oldest families with crown node ages reaching back into the Early Cretaceous. We comment on possible error sources and the neces-sity for methodological improvement in molecular dating.

Full text

Botanical Journal of the Linnean Society
, 2004,
146
, 385– 398. With 1 figure
© 2004 The Linnean Society of London,
Botanical Journal of the Linnean Society,
2004,
146
, 385– 398
385
Blackwell Science, LtdOxford, UKBOJBotanical Journal of the Linnean Society0024-4074The Linnean Society of London, 2004? 2004
1464
385398
Original Article
AGE OF MONOCOT GROUPS
T. J ANSSEN and K. BREMER
*Corresponding author. Postal address from 01/2005:
Albrecht-von-Haller-Institute für Pflanzenwissenschaften,
Abt. Spezielle Botanik, Untere Karspüle 2, 37073 Göttingen,
Germany. E-mail: janssen@mnhn.fr
The age of major monocot groups inferred from 800
+
rbcL
sequences
THOMAS JANSSEN
1
* and KÅRE BREMER
2
1
Museum National d’Histoire Naturelle, Département de Systématique et Evolution, USM 0602 Taxinomie
et collections, 16 rue Buffon, 75005 Paris, France
2
Department of Systematic Botany, Evolutionary Biology Centre, Norbyvägen 18D, SE-752 36 Uppsala,
Sweden
Received August 2003; accepted for publication June 2004
Phylogenetic research on monocots has been extraordinarily active over the past years. With the familial interre-
lationships being sufficiently understood, the question of divergence times and crown node ages of major lineages
comes into focus. In this study we present the first attempt to estimate crown and stem node ages for most orders
and families of monocots, based on
rbcL
sequence data and comprehensive taxon sampling. From our analysis it is
obvious that considerable monocot diversification took place during the Early Cretaceous, with most families already
present at the Cretaceous–Tertiary boundary. Araceae, Arecaceae and Orchidaceae are among the oldest families
with crown node ages reaching back into the Early Cretaceous. We comment on possible error sources and the neces-
sity for methodological improvement in molecular dating. © 2004 The Linnean Society of London,
Botanical Jour-
nal of the Linnean Society
, 2004,
146
, 385–398.
ADDITIONAL KEYWORDS:
Cretaceous – dating – fossils – NPRS – phylogeny – Tertiary.
INTRODUCTION
Monocots constitute an angiosperm clade of out-
standing economic and ecological importance. The
Angiosperm Phylogeny Group (APG II, 2003) recog-
nizes 81 families in ten orders (with two families
unplaced to order). Distribution of monocots is world-
wide, with some families predominantly in open tem-
perate habitats (e.g. grasses, Poaceae) and others with
important species diversity in the tropics (e.g. orchids,
Orchidaceae and palms, Arecaceae).
Phylogenetic research on monocots has received
much interest recently, fostered by three international
congresses. Chase
et al
. (2000) proposed a first phylo-
genetic tree, including all orders, based on a large data
set comprising three DNA regions (
rbcL
,
atpB
and 18S
rDNA) but with comparatively limited taxon sam-
pling. Meanwhile, a detailed treatment is available for
several orders (Les, Cleland & Waycott, 1997; Fay
et al
., 2000; Kress
et al
., 2001; Vinnersten & Bremer,
2001; Bremer, 2002; Caddick
et al
., 2002a, b). With the
results of the available studies taken together, we
have at present a rather well supported phylogenetic
hypothesis for the whole clade that is fairly well
resolved down to family level. This provides us with a
solid basis for research on age, biogeography and
evolution of this group.
Only a few studies are available to date on age
inference including monocots. Wikström, Savolainen
& Chase (2001) analysed a data set of 560
angiosperms (
rbcL
,
atpB
and 18S rDNA) using
nonparametric rate smoothing (Sanderson, 1997).
Bremer (2000) determined the ages of major monocot
lineages, calculating mean branch lengths in a phylog-
eny of
rbcL
sequences. Both studies also addressed
the issue of age calibration using evidence from the
fossil record. Givnish
et al
. (2000) estimated ages for
several groups of commelinids using
ndhF
sequences.
Recently, some detailed studies comparing different
dating methods have become available for Liliales
(Vinnersten & Bremer, 2001) and Poales (Bremer,
2002). These studies notwithstanding, the node ages

386
T. JANSSEN and K. BREMER
© 2004 The Linnean Society of London,
Botanical Journal of the Linnean Society,
2004,
146
, 385– 398
of most parts of the monocot tree are still rather
imperfectly known.
Rate heterogeneities, age calibration, the underlying
phylogenetic hypothesis and taxon sampling are major
sources of error in phylogenetic dating with DNA
sequences. Error sources related to sequence evolution
have been extensively discussed (review by Sanderson,
1998). It is well known that increased taxon sampling
may increase the accuracy of phylogenetic reconstruc-
tion (Källersjö
et al
., 1998; Rydin & Källersjö, 2002;
Zwickl & Hillis, 2002). The influence of taxon sampling
on dating has not received much interest (but see
Baldwin & Sanderson, 1998) and is still insufficiently
known, but improved taxon sampling has been shown
to reduce error in branch length estimates (Sanderson,
1990) and to circumvent putative biases introduced by
the overrepresentation of certain groups of herbaceous
plants, which are prone to accelerated evolutionary
rates (Wikström
et al
., 2001).
Herein, we present an extended study on the age of
monocots. We use all monocot
rbcL
sequences of suffi-
cient quality currently available from GenBank to
infer divergence times of major lineages and crown
node ages for most monocot families. Sequences from
all families (
sensu
APG II, 2003) of monocots, with the
exception of the achlorophyllous Corsiaceae and Triu-
ridaceae, have been included. Within the now enlarged
(APG II, 2003) families Asparagaceae, Alliaceae and
Xanthorrhoeaceae, all of the formerly recognized
Asparagales families (
sensu
APG, 1998) are repre-
sented by at least one sequence. Dating analysis
undertaken in this study is thus based on an extensive
data set comprising more than 800
rbcL
sequences.
Our age estimates are based on a well supported con-
sensus phylogeny derived from several detailed stud-
ies. We provide the largest taxon sampling currently
available under the assumption that this is as benefi-
cial for dating attempts as it has been shown to be for
phylogenetic reconstruction.
Nonparametric rate smoothing (NPRS; Sanderson,
1997) may cope with rate heterogeneities, which are
more likely to occur in trees with an extensive and
diverse taxon sampling. Furthermore, rate changes
among adjacent parts of the tree are presumably
smaller if more taxa are sampled. This agrees with the
underlying assumption of NPRS that no abrupt rate
changes occur in the tree. For calibration, we follow
Bremer (2000) who used eight reference fossils to esti-
mate the split between
Acorus
and the other monocots
at 134 Mya.
MATERIAL AND METHODS
D
ATA
SET
We downloaded all Liliopsida (monocot)
rbcL
sequences available from GenBank in August 2002
with the aim of sampling as many genera as possible.
Each genus is represented by a single sequence in the
data matrix. If multiple sequences of the same quality
were available for one genus, the most recent was cho-
sen unless there was an older version of higher quality
with respect to sequence length and percentage of
ambiguous bases. Sequences with species names have
been preferred to sequences of ambiguously deter-
mined or undetermined species. If these selection cri-
teria failed, one sequence was chosen arbitrarily, that
is the ‘most common’ species or simply the first item in
the list. Following selection, sequences with long poly-
N stretches (more than 150 ambiguous bases per
sequence) and sequences of insufficient length (less
than 75% of the total length of the data matrix) were
discarded.
The resulting data set consists of the
rbcL
sequences of 878 genera from 77 families represent-
ing all ten orders of monocots (Table 1). Family
assignments have been adopted from APG II (2003).
The matrix comprises a total of 1343 included
sequence positions. Of these, 760 are parsimony-
informative and 24 are single or double gap charac-
ters resulting from putative errors in some sequences.
The alignment was carried out by hand and involved
no indels.
rbcL
VERSUS
THREE
GENES
Our dating is based on
rbcL
sequences only. To assess
the historical information provided by this single gene
we compared it with that of the data matrix compris-
ing three genes (including
rbcL
) by Chase
et al
. (2000).
Their tree was published with parsimony branch
lengths. We also obtained parsimony branch lengths
for a
rbcL
-gene-only tree comprising exactly the same
taxa as the three-gene-tree using PAUP (Swofford,
2001), with settings as described in Chase
et al
. (2000)
and their tree as a topological constraint. Dating of
their tree was then performed with the three-gene-
tree’s branch lengths as well as with the
rbcL
-only-
tree’s branch lengths. For this comparative purpose,
no correction was made for multiple substitutions.
Dating was performed using nonparametric rate
smoothing (see below; Sanderson, 1997, 1999).
Ages estimated with
rbcL
only were 0–20 Myr (in
six instances 20–24 Myr) older or 0–10 Myr (in two
cases 13–14 Myr) younger than with the three-gene
data set. Compared with the three-gene data set there
was no overall tendency for generally younger or
generally older age estimates with
rbcL
only.
Our data set with its single gene comprises fewer
characters per taxon compared with a multigene data
set, making it less suitable for phylogenetic recon-
struction. However, considering the similarity of ages
obtained with all three genes and with
rbcL
only, we

AGE OF MONOCOT GROUPS
387
© 2004 The Linnean Society of London,
Botanical Journal of the Linnean Society,
2004,
146
, 385– 398
Table 1.
Sampling and age estimates for families and orders of monocots. Orders are arranged according to major groups
recognized within monocots, viz. core monocots and commelinids. Orders are in the same sequence as in Fig. 1. Numbers
in parentheses after orders indicate the number of sampled families in relation to the number of families recognized in
this order according to APG II (2003). The first column gives the number of genera per family or order for which
rbcL
sequences were sampled. Crown and stem node age estimates are found in the second and third columns. Crown node age
estimates are absent for monogeneric families and for families for which data for only one genus were available
Taxon
Number
of
genera
sampled
Crown
node
age
(Mya)
Stem node
age (Mya)
Acoraceae 1 – 134
(calibration)
Alismatales (14/14) 78 128 131
Alismataceae 8 55 57
Aponogetonaceae 1 – 98
Araceae 29 117 128
Butomaceae 1 – 88
Cymodoceaceae 5 61 67
Hydrocharitaceae 16 75 88
Juncaginaceae 3 52 82
Limnocharitaceae 2 44 57
Posidoniaceae 1 – 67
Potamogetonaceae 5 23 47
Ruppiaceae 1 – 65
Scheuchzeriaceae 1 – 92
Tofieldiaceae 2 100 124
Zosteraceae 3 17 47
CORE
MONOCOTS
674 126 131
Petrosaviaceae 2 123 126
Dioscoreales (3/3) 14 123 124
Burmanniaceae 3 93 116
Dioscoreaceae 8 80 116
Nartheciaceae 3 76 123
Pandanales (4/5) 12 114 124
Cyclanthaceae 4 77 98
Pandanaceae 2 51 98
Stemonaceae 4 84 108
Velloziaceae 2 14 108
Liliales (9/10) 47 117 124
Alstroemeriaceae 3 30 76
Campynemataceae 2 73 117
Colchicaceae 9 44 76
Liliaceae 17 80 91
Luzuriagaceae 2 56 79
Melanthiaceae 10 97 109
Philesiaceae 2 60 76
Rhipogonaceae 1 – 76
Smilacaceae 1 – 90
Asparagales (14/14) 371 119 122
Alliaceae 55 87 91
Asparagaceae 63 89 91
Asteliaceae 3 92 104
Blandfordiaceae 1 – 100
Boryaceae 2 54 109
Doryanthaceae 1 – 107
Hypoxidaceae 6 78 100
Iridaceae 57 96 103
Ixioliriaceae 1 – 112
Lanariaceae 1 – 113
Orchidaceae 145 111 119
Tecophilaeaceae 8 87 108
Xanthorrhoeaceae 27 90 93
Xeronemataceae 1 – 100
COMMELINIDS
353 120 122
Arecaceae 120 110 120
Dasypogonaceae 3 100 119
Commelinales (5/5) 37 110 114
Commelinaceae 27 62 89
Haemodoraceae 2 81 98
Hanguanaceae 1 – 104
Philydraceae 2 47 110
Pontederiaceae 5 39 89
Zingiberales (8/8) 21 88 114
Cannaceae 1 – 68
Costaceae 4 47 79
Heliconiaceae 1 – 88
Lowiaceae 1 – 78
Marantaceae 4 57 68
Musaceae 3 61 87
Strelitziaceae 3 59 78
Zingiberaceae 4 26 79
Poales (18/18) 172 113 117
Anarthriaceae 3 55 96
Bromeliaceae 7 96 112
Centrolepidaceae 1 – 97
Cyperaceae 48 76 88
Ecdeiocoleaceae 2 73 89
Eriocaulaceae 6 58 105
Flagellariaceae 1 – 108
Hydatellaceae 1 – 87
Joinvilleaceae 1 – 90
Juncaceae 6 74 88
Mayacaceae 1 – 87
Poaceae 52 83 89
Rapateaceae 6 79 112
Restionaceae 29 74 96
Sparganiaceae 1 – 89
Thurniaceae 2 33 98
Typhaceae 1 – 89
Xyridaceae 4 87 105
Taxon
Number
of
genera
sampled
Crown
node
age
(Mya)
Stem node
age (Mya)

388
T. JANSSEN and K. BREMER
© 2004 The Linnean Society of London,
Botanical Journal of the Linnean Society,
2004,
146
, 385– 398
assume that the information content is still sufficient
for age inference. Restricting the data set to a single
gene does not markedly alter the historical informa-
tion in the data set, makes more taxa available, and
enables us to infer crown node ages for most families
as well as to address potential problems caused by a
limited taxon sampling.
O
BTAINING
A
TREE
WITH
BRANCH
LENGTHS
FOR
DATING
Phylogenetic reconstruction is not attempted in this
study. A topological backbone constraint tree was con-
structed, combining well supported clades from earlier
studies focused on phylogenetic reconstruction of var-
ious monocot groups and based on data sets including
two or more genes.
First, all orders and families were constrained to
be monophyletic. No constraints were applied within
families. Second, ordinal interrelationships were
adopted from APG II (2003). Third, interrelation-
ships of families within orders were constrained
using sufficiently supported nodes, that is with
bootstrap or jackknife frequencies of 85% or higher,
from Les
et al
. (1997; Alismatales), Fay
et al
. (2000;
Asparagales), Vinnersten & Bremer (2001; Liliales),
Caddick
et al
. (2002a, b; Dioscoreales, Pandanales),
Kress
et al
. (2001; Zingiberales) and Bremer (2002;
Poales). Branches with low support in those studies
were collapsed into polytomies in the constraint tree.
For Commelinales, there is no detailed study avail-
able to date. Some well supported nodes for this
order were adopted from Chase
et al
. (2000). The
constraint tree topology is available from the authors
upon request.
One completely resolved tree was then obtained by a
heuristic PAUP search (Swofford, 2001) with topolog-
ical constraints enforced as specified above, 1000 ran-
dom addition sequences holding five trees at each
addition step, using tree bisection reconnection (TBR)
branch swapping, and the MulTrees option not in
effect and hence, keeping only one tree per replicate.
One single most parsimonious tree, 13 533 steps long,
was found and saved as an input tree for subsequent
analyses.
Parsimony branch lengths corresponding to the
total number of observed substitutions per branch
tend to underestimate the number of changes, espe-
cially on long branches. Therefore, corrected branch
lengths were obtained with PAUP using a maximum
likelihood optimization under the GTR
+
G
model of
sequence evolution (
a
=
0.5). With our data set, esti-
mation of the shape parameter from the sequences is
not computationally feasible in a reasonable amount
of time. We chose, therefore, to adopt the default
parameter value from PAUP, arguing that the influ-
ence of the error in branch lengths thus introduced is
small compared with calibration uncertainty (Bremer,
Friis & Bremer, 2004). The branch length calculation
(yielding the number of expected substitutions per
site) resulted in some zero-length branches, which are
not tolerated by the algorithms implemented in the
r8s programme (Sanderson, 1999). Hence, the lengths
of 58 such branches in the tree were set to 0.000001
substitutions per site. Considering the number of
branches in the tree and the average number of sub-
stitutions per site, this modification should not influ-
ence the outcome of the analysis. The input tree with
branch lengths is available from the authors upon
request.
D
ATING
Using eight reference fossils, Bremer (2000) esti-
mated the age of the split between
Acorus
and all
other monocot lineages at 134 Mya. Later, it has been
shown that one of Bremer’s reference fossils (fossil B)
is not a
Pistia
of Araceae as assumed, but a different,
unknown plant (Stockey, 2003). Recalculation of the
mean change rates used by Bremer, and recalibration
without that fossil gives, however, the same result. We
used this dating of the crown group age of monocots to
calibrate our tree. Further age constraints using
reference fossils were not included, but consistency of
the inferred node ages with putative positions of all
reference fossils used by Bremer (2000) has been
checked.
The input tree with corrected branch lengths (see
above) was subjected to nonparametric rate smoothing
as implemented in the r8s programme (NPRS;
Sanderson, 1997, 1999). Searches were restarted
three times with a perturbation of the initial param-
eters to ensure the solution reached a stable optimum.
To check for the presence of multiple optima, optimi-
zations were also restarted three times from different
initial divergence time estimates.
We also tried the mean path length method using
the PATH program (MPL; Britton, 2002; Britton
et al
.,
2002). MPL tolerates zero-length branches but
requires a strictly bifurcating tree with branch
lengths in terms of whole numbers of substitutions,
observed or expected (corrected). Hence, we multiplied
the maximum likelihood-derived expected branch
lengths per site by the number of sites included in
the analysis to obtain (corrected) parsimony branch
lengths. Polytomies within the families were arbi-
trarily resolved, assigning zero-lengths to the
branches by branch-swapping using the tree itself as a
topological constraint and saving only one tree. The
MPL clock tests (see Britton
et al
., 2002) reported
significant deviations from the clock at about one third
of the nodes, including many nodes at the base of the

AGE OF MONOCOT GROUPS
389
© 2004 The Linnean Society of London,
Botanical Journal of the Linnean Society,
2004,
146
, 385– 398
tree. Hence, we proceeded with NPRS dating, which
permits rate changes throughout the tree. Penalized
likelihood (Sanderson, 2002) was not available as an
alternative, parametric dating approach because the
software failed to run with a data set of this size. Nec-
essary computation time for Bayesian estimation of
divergence times (Kishino, Thorne & Bruno, 2001;
Thorne & Kishino, 2002) with our data is not currently
feasible.
No confidence intervals are provided with our NPRS
analysis because calculating confidence intervals by
bootstrapping with r8s (NPRS) is not computationally
feasible with a data matrix of this size. However, our
MPL analysis reported confidence intervals less than
±
16 Myr. Earlier studies (Vinnersten & Bremer, 2001)
indicated that confidence intervals calculated by boot-
strapping with r8s and by the MPL procedure are sim-
ilar, so confidence intervals in this range may be
envisaged in interpreting the ages obtained by our r8s
analysis.
RESULTS
The
rbcL
matrix yielded a single most parsimonious
tree that was not fully resolved but showed some poly-
tomies, especially within clades that were represented
by a high number of rather closely related species
such as Arecaceae, Orchidaceae and some families in
Asparagales. At a deeper level, Dioscoreales and
Pandanales were in a trichotomy with the Liliales–
Asparagales–commelinids clade.
Taking into account the large taxon sampling, it is
not surprising to find considerable variation in total
branch lengths from the root to the terminals of the
tree, indicating different substitution rates. Notewor-
thy are comparably slow rates of molecular evolution
in Arecaceae (Wilson, Gaut & Clegg, 1990). Some taxa
show particularly long-terminal branches, attribut-
able to putative phylogenetic misplacement in the
case of
Xyris
,
Mayaca
,
Trithuria
and
Centrolepis
(see
also Bremer, 2002), and in
Burmannia
possibly
related to a specialized life history (Soltis
et al
., 2000).
Within Asparagaceae (Amaryllidaceae),
Hessea
,
Geth-
yllis
,
Strumaria
and
Traubia
display comparably long
branches. Meerow
et al
. (1999) also found elevated
terminal branch lengths for these taxa.
Xerolirion
(Asparagaceae, Laxmanniaceae) is another taxon with
an anomalously long branch.
Age estimates for crown and stem nodes at the
family level and above are shown in Figure 1 and in
Table 1. Disregarding a few younger clades in
Alismatales, all divergences are older than 65 Myr.
With the exception of Zingiberales (88 Myr), all orders
are older than 100 Myr. Most families diverged
between 65 Mya and 100 Mya. A crown group from the
Early Cretaceous (> 100 Mya) has been found for
Araceae, Petrosaviaceae, Orchidaceae, Arecaceae and
Dasypogonaceae.
DISCUSSION
P
HYLOGENY
In terms of taxon numbers, our tree is so far the most
comprehensive of monocots. However, phylogenetic
reconstruction was not a major goal of this study, and
resolution obtained for all unconstrained nodes has to
be considered with caution because no extensive
search has been conducted. Below we will discuss
topological deviations of our tree from in-depth studies
of limited groups. It is not our intention to discuss
alternative phylogenetic hypotheses, but to point out
possibly problematic nodes for which dating results
may change with the availability of better (multigene)
global phylogenetic trees. Chase
et al
. (2000) also pre-
sented an extensive family level hypothesis of monocot
phylogeny. In their tree, Petrosaviacae are sister to
Pandanales, although with low support, whereas it is
the sister group of all other core monocots in our anal-
ysis. Although Chase et al. (2000) did not sample Alis-
matales extensively, they also found Araceae and
Tofieldiaceae to be sister to Alismatales in that order.
Furthermore, they obtained a different resolution
within Zingiberales, although with low support and
incongruent with the extensive analysis by Kress et al.
(2001). Otherwise, there is no important conflict
between their tree and ours. Notably, we find exactly
the same topology for all unconstrained nodes in
Commelinales, Dioscoreales and Pandanales.
Except for the position of Ruppiaceae and Scheuchz-
eriaceae, our topology is congruent with that of Les
et al. (1997). However, deviations concern families
that have been placed with a low bootstrap support
only (40 and 11%, respectively).
The topology adopted by Bremer (2000) is also
largely congruent with our tree used for dating. As in
our analysis, and unlike Vinnersten & Bremer (2001),
Campynemataceae are sister to the rest of Liliales.
Orchids, however, cluster within the Boryaceae–
Lanariaceae clade, whereas this clade is sister to
Orchidaceae and the rest of Asparagales in our anal-
ysis. Bromeliaceae are not sister to Rapateaceae in
our tree, but to the graminoid clade including Erio-
caulaceae and Xyridaceae. Mayacaceae are sister to
Eriocaulaceae and Xyridaceae, whereas in our tree
this family is sister to Hydatellaceae and close to
Cyperaceae. The latter conflict may probably be
attributed to a branch attraction phenomenon
(Bremer, 2002).
We find the same interrelationship among the three
families of Dioscoreales as Caddick et al. (2002a, b).
However, in Pandanales, these authors have Vellozi-
aceae as sister of the rest and not just to Stemonaceae.

390 T. JANSSEN and K. BREMER
© 2004 The Linnean Society of London, Botanical Journal of the Linnean Society, 2004, 146, 385 –398
*
*
*
*
**
*
*
*
*
*
*
*
*
*
*
**
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*

AGE OF MONOCOT GROUPS 391
© 2004 The Linnean Society of London, Botanical Journal of the Linnean Society, 2004, 146, 385 –398
Vinnersten & Bremer (2001) presented a detailed
study of Liliales. In their tree, Campynemataceae are
sister to Melanthiaceae–Smilacaceae–Liliaceae, and
the Alstroemeriaceae–Luzuriagaceae–Colchicaceae
clade is sister to these. However, parsimony branch
lengths supporting the deep topology in Liliales are
comparably short. Disregarding the relationships
within the enlarged Alliaceae and Asparagaceae in
Asparagales (APG II, 2003), our tree is not in conflict
with that of Fay et al. (2000). Unconstrained nodes in
our tree correspond to polytomies in their consensus
tree. Relationships between Boryaceae and related
families received low support in the analysis by Fay
et al. (2000), and it is not surprising that we find
a different topology. The position at three basal
nodes in Zingiberales (Musaceae, Heliconiaceae and
Strelitziaceae–Lowiaceae) differs from that in the tree
by Kress et al. (2001), who only obtained a Bremer
support of one for this alternative topology. Finally,
comparing our Poales tree with that of Bremer (2002),
some unconstrained taxa end up in different positions.
This is true for Flagellariaceae and Centrolepidaceae,
although they remain within Bremer’s graminoid
clade. Restionaceae are monophyletic in our analysis,
not including Centrolepidaceae which are placed as a
sister group of Anarthriaceae–Restionaceae. Bremer
found a Bromeliaceae–Sparganiaceae–Typhaceae
clade sister to the cyperoid clade. However, the node
supporting this deep level relationship within Poales,
as well as the relationships within that clade of three
families, received low jackknife support and are hence,
not in strong conflict with Bromeliaceae being sister
to Xyridaceae–Eriocaulaceae and Sparganiaceae–
Typhaceae being sister to Hydatellaceae–Mayacaceae
and the cyperoid clade. Hydatellaceae and Maya-
caceae were excluded from Bremer’s analysis because
of their unstable positions. In our reconstruction,
those two families end up as sister groups, which
probably is an artefact.
INCREASED TAXON SAMPLING IN PHYLOGENETIC
INFERENCE AND DATING
It has been argued that phylogenetic reconstruction
profits, in terms of accuracy and support, from the
inclusion of more taxa in the analysis (Källersjö et al.,
1998; Soltis et al., 2000; Rydin & Källersjö, 2002;
Zwickl & Hillis, 2002). However, few indications are
available as to whether the same is true for molecular
dating. Extended sampling may level out biases intro-
duced by an overrepresentation of groups with high
substitution rates, such as grasses. Their exaggerating
effect on overall age estimates (especially at deeper
levels) should be cancelled out by the inclusion of slow
groups, such as palms. In dating, adding more genes to
the data set is not likely to be helpful because rate
variation appears to be systematic and lineage-
specific, such that different molecular markers are
affected in the same way and will thus not alleviate
differences among lineages (Sanderson & Doyle,
2001). Adding more taxa seems to remain the only
option to improve data sets for age inference. Hence,
the purpose of increased taxon sampling in this study
is not only to estimate crown node ages, but we also
expect it to increase precision of age estimates.
Even though the effect of uneven rates causing
different branch lengths, for example in groups with
an atypical life history such as aquatics or parasites,
cannot be entirely removed by adding taxa to the tree,
excessively long branches may be broken up and there
may be a ‘smoothing’ of rate differences between adja-
cent regions of the tree. This should, in turn, increase
correspondence with the underlying assumption in
nonparametric rate smoothing (Sanderson, 1997),
namely that rates are autocorrelated between lineages
and abrupt changes do not occur.
Sanderson (1990) noted that reduced taxon sam-
pling results in reduced branch length estimates,
since hidden homoplasy may give rise to additional
character state transformations when more taxa are
added to the phylogeny. Hence, overall branch lengths
would become longer, and optimization methods would
be likely to find older ages at deeper nodes. Indeed,
there seems to be a general tendency towards pushing
divergence times back in time with increased taxon
sampling.
DIVERGENCE TIMES
In calibrating our tree we followed Bremer (2000)
who, based on eight reference fossils, estimated the
split between Acorus and all other monocots at
134 Mya, the only reliable calibration available to this
date. According to our dating (Fig. 1), much diversifi-
cation took place before 100 Mya during the Early
Cretaceous. With a few exceptions, all family stem
Figure 1. Dated phylogenetic tree of monocot families obtained from analysis of 878 rbcL sequences. Thick bars show
inferred crown node ages (absent in families represented by a single sequence only). The number of rbcL sequences and
the stem and crown node ages for each family are given in Table 1. Asterisks represent nodes not included in the enforced
topological constraints compiled from well-supported clades in various published analyses (listed in the text); all other
branches represent well supported clades included in a constraint tree used in the analysis. The two vertical lines show
the mid-Cretaceous 100 Mya and the Cretaceous–Tertiary boundary 65 Mya.


392 T. JANSSEN and K. BREMER
© 2004 The Linnean Society of London, Botanical Journal of the Linnean Society, 2004, 146, 385 –398
lineages were present by 65 Mya at the Cretaceous–
Tertiary boundary. Moreover, a considerable portion of
the family crown nodes dates back to the Late
Cretaceous.
Bremer (2000) generally obtained younger ages. We
found about three times as many Early Cretaceous
monocot lineages as did Bremer. He used parsimony
branch lengths, which tend to underestimate change
and may result in younger ages compared with maxi-
mum likelihood estimates. Furthermore, there may be
unknown methodological effects leading to younger
age estimates with mean-path-lengths or older age
estimates with nonparametric rate smoothing. Even
more importantly, sampling was much more limited
in Bremer’s study compared with this paper and,
as noted above, increased sampling may lead to
increased age estimates.
Our age estimates for Liliales are 10–50 Myr older
than those presented by Vinnersten & Bremer (2001).
This is probably related to the difference between our
study and that of Bremer (2000), since Vinnersten &
Bremer calibrated the Liliales crown node at 82 Mya
following Bremer (2000). We estimate this node at
117 Mya, and if the tree by Vinnersten & Bremer is
calibrated with our crown node age estimate for the
order, then age estimates within the order become
largely similar.
Within Poales, our results are remarkably congru-
ent with the age estimates of Bremer (2002). Most
deviations are within the expected confidence inter-
vals. Bromeliaceae are about 26 Myr older in our
analysis, which might be a result of their different
placement in the tree. This effect is more dramatic
for Centrolepidaceae, the stem node of which was
estimated to be at 97 Mya but around 45 Mya when
nested within Restionaceae. We find Cyperaceae
36 Myr older than did Bremer, a difference probably
due to differences in sampling. Bremer sampled
eight taxa of Cyperaceae compared with 48 in our
analysis.
It emerges from the above discussion that
increased sampling pushes divergence times back
in time. Although some theoretical considerations
(Sanderson, 1990) might explain such behaviour to a
certain extent, this tendency is worrying. The already
existing gap between molecular age estimates and the
ages obtained from the fossil record will be enhanced
unless we hit an upper limit for the ages at a certain
sampling density. The observed phenomenon might
also, however, be a systematic error and may point to
the necessity of methodological improvement in
molecular dating. If the observed tendency is not a
methodological artefact, then we must assume that
current age estimates are generally too young
because the present sampling is far from complete in
most families.
The crown nodes of Araceae, Petrosaviaceae,
Orchidaceae, Arecaceae and Dasypogonaceae have
been estimated to date back to the Early Creta-
ceous. Among these, Petrosaviaceae and Dasy-
pogonaceae are small families with a limited
representation in our data, and they occupy posi-
tions as sister to major clades in core monocots and
commelinids, respectively. In these cases the age
estimates are certainly enhanced by the phyloge-
netic positions and should be interpreted with
caution. It is not surprising to find Araceae and
Arecaceae among the old families, with macrofossils
being known from the Albian and the Campanian,
respectively (Herendeen & Crane, 1995). A crown
node age of 111 Myr for Orchidaceae is unexpected,
however. Traditionally, this family has been looked
at as a very specialized and hence, probably, a young
group. Considering the extensive sampling within
orchids (145 genera) and its firm phylogenetic
position at the base of the Asparagales, this age
estimate appears to be well supported. However, a
methodological bias due to the extended sampling
still cannot be excluded. If our age estimate turns
out to be true, the evolutionary history of this fam-
ily could be seen in a new light. Orchid diversity is
not necessarily due to a rapid and recent radiation,
and similar patterns, for example in palms, might be
hypothesized.
DIVERGENCE TIMES AND THE FOSSIL RECORD
Molecular age estimates generally predate fossil ages.
The fossil record of monocots is comparatively poor,
with few fossils attributable to families reaching back
beyond the Maastrichtian. Reasons for this may be
the herbaceous habit and widespread zoophilous
pollination (Herendeen & Crane, 1995). Monocot
pollen is known from the Early Cretaceous (125 Mya),
and fossils attributable to taxonomic groups date from
the Turonian (Arecaceae pollen, 90 Mya) and the
Santonian–Campanian boundary (Zingiberales fruits,
83 Mya; Bremer, 2000). The ages estimated here are
somewhat older; for monocots 9 Myr, for Arecaceae
10 Myr and for Zingiberales 5 Myr older.
From fossil evidence, a major radiation of
angiosperms is obvious in the mid-Cretaceous 130–90
Mya (Lidgard & Crane, 1990; Herendeen & Crane,
1995). With the diversification burst of angiosperms
setting in no earlier than about 115 Mya, there still
remains a considerable gap between the fossil record
and our molecular dating, with the majority of lin-
eages already present by that time. At this time we
cannot decide whether this is an indication of an
incomplete fossil record or an argument against cur-
rently available molecular dating methodologies.

AGE OF MONOCOT GROUPS 393
© 2004 The Linnean Society of London, Botanical Journal of the Linnean Society, 2004, 146, 385 –398
DIVERGENCE TIMES – ERROR SOURCES
Molecular dating may be affected by a variety of error
sources (Sanderson & Doyle, 2001), the majority of
which are not sufficiently theoretically understood.
Substitutional noise as a stochastic characteristic of
the substitution process as such imposes a lower
boundary on error and finds its expression in minimal
confidence intervals. Branch length estimation will be
affected by rate variation across sites and across lin-
eages, as well as by the optimization method used.
Rate variation across sites might lead to incorrect esti-
mates, especially at high rates (Sanderson & Doyle,
2001). According to these authors, the use of a gamma
distribution in the maximum likelihood model leads to
10–30 Myr younger ages compared with the assump-
tion of equal rates.
Rate variation across lineages is harder to deal
with. Clock-like behaviour over the whole topology
would be ideal for molecular dating, but with more lin-
eages being included in the analysis this assumption
becomes more and more unlikely. NPRS (Sanderson,
1997) is a dating method that deals with rate hetero-
geneities by minimizing rate changes between adja-
cent branches. The assumption of rate autocorrelation
appears to be reasonable, but there is also no a priori
reason to exclude the possibility of abrupt rate
changes between adjacent lineages. Until more work
has been done on the evolution of evolutionary rates,
dating will largely depend on the underlying hypoth-
eses and the optimization method used.
Molecular dating is unable to provide absolute
ages. The phylogenetic tree has to be placed in an
appropriate stratigraphic context to calibrate the
obtained relative ages. Calibration is a major error
source because it depends on the dating of the fossil
itself and on its attachment point to the phylogeny.
Hence, several reference fossils are desirable in large
trees comprising a variety of taxa and considerable
rate heterogeneity.
We found that, for a restricted taxon sampling, the
ages inferred from rbcL were only comparable with
those obtained with a three-gene matrix (rbcL + atpB
+ 18S rDNA). It has been argued that evolutionary
rates do not differ significantly among different cellu-
lar compartments (Sanderson & Doyle, 2001). If this is
true, the bias introduced using a single gene for dating
might be negligible.
Dating largely depends on the underlying topology
used to estimate branch lengths. Nodes marked with
an asterisk in Figure 1 have not been constrained
prior to phylogenetic reconstruction. Alternative
topologies, especially those congruent with earlier
studies (see above), are less parsimonious. However,
topological rearrangements at these nodes might
affect age inference. Branches are short at a deeper
level in our tree, and rearrangements at this level only
slightly affect branch length estimates. Rearrange-
ments at a higher level should only slightly affect
branch lengths within the rearranged clade. Under
the assumption that no major rearrangements may be
expected, the topological bias should be reduced to an
acceptable minimum. The absence of major conflict
(see above) between our topology and earlier recon-
structions renders this assumption reasonable.
Sampling biases, especially an overrepresentation
of rapidly evolving herbaceous lineages, may be
alleviated by including more and diverse taxa. Our
approach, to include information from all available
genera, is the only way to avoid a priori selection of
data, but it may not necessarily lead to a representa-
tive taxon sampling. Two families, Orchidaceae and
Arecaceae, may be over-represented in our data set.
Sanderson & Doyle (2001) pointed out that non-
clocklike behaviour of evolutionary rates might lead
to significant deviation among results obtained with
different dating methods. Different methods may
introduce systematic biases, which are generally hard
to detect. If our finding that increased sampling leads
to older age estimates is corroborated in the future,
then current dating methods need revision.
ACKNOWLEDGEMENTS
We thank J. Nylander for computer assistance and T.
Britton for running MPL analyses. The study was
supported by a Swedish Research Council grant to KB.
REFERENCES
APG. 1998. An ordinal classification for the families of
flowering plants. Annals of the Missouri Botanical Garden
85: 531–553.
APG II. 2003. An update of the Angiosperm Phylogeny Group
classification for the orders and families of flowering plants:
APG II. Botanical Journal of the Linnean Society 141: 399–
436.
Baldwin BG, Sanderson MJ. 1998. Age and rate of diversi-
fication of the Hawaiian silversword alliance (Compositae).
Proceedings of the National Academy of Sciences, USA 95:
9402–9406.
Bremer K. 2000. Early Cretaceous lineages of monocot
flowering plants. Proceedings of the National Academy of
Sciences, USA 97: 4707–4711.
Bremer K. 2002. Gondwanan evolution of the grass alliance of
families (Poales). Evolution 56: 1374–1387.
Bremer K, Friis EM, Bremer B. 2004. Molecular phyloge-
netic dating of asterid flowering plants shows Early Creta-
ceous diversification. Systematic Biology 53: 496–505.
Britton T. 2002. PATH. Computer program and documentation
available from http://www.math.uu.se/~tomb/PATH.html
Britton T, Oxelman B, Vinnersten A, Bremer K. 2002.
Phylogenetic dating with confidence intervals using mean

394 T. JANSSEN and K. BREMER
© 2004 The Linnean Society of London, Botanical Journal of the Linnean Society, 2004, 146, 385 –398
path lengths. Molecular Phylogenetics and Evolution 24: 58–
65.
Caddick LR, Rudall PJ, Wilkin P, Hedderson TAJ, Chase
MW. 2002a. Phylogenetics of Dioscoreales based on com-
bined analyses of morphological and molecular data. Botan-
ical Journal of the Linnean Society 138: 123–144.
Caddick LR, Wilkin P, Rudall PJ, Hedderson TAJ, Chase
MW. 2002b. Yams reclassified: a recircumscription of
Dioscoreaceae and Dioscoreales. Taxon 51: 103–114.
Chase MW, Soltis DE, Soltis PS, Rudall PJ, Fay MF,
Hahn WH, Sullivan S, Joseph J, Molvray M, Kores PJ,
Givnish TJ, Sytsma KJ, Pires JC. 2000. Higher-level sys-
tematics of the monocotyledons: An assessment of current
knowledge and a new classification. In: Wilson KL, Morrison
DA, eds. Monocots: systematics and evolution. Melbourne:
CSIRO. 3–16.
Fay MF, Rudall PJ, Sullivan S, Stobart KL, de Bruijn AY,
Reeves G, Qamaruze-Zaman F, Hong W-P, Joseph J,
Hahn WJ, Conran JG, Chase MW. 2000. Phylogenetic
studies of Asparagales based on four plastid DNA regions.
In: Wilson KL, Morrison DA, eds. Monocots: systematics and
evolution. Melbourne: CSIRO, 360–371.
Givnish TJ, Evans TM, Zjhra ML, Patterson TB, Berry
PE, Sytsma KJ. 2000. Molecular evolution, adaptive radi-
ation, and geographic diversification in the amphiatlantic
family Rapateaceae: evidence from ndhF sequences and mor-
phology. Evolution 54: 1915–1937.
Herendeen PS, Crane PR. 1995. The fossil history of the
monocotyledons. In: Rudall PJ, Cribb PJ, Cutler DF,
Humphries CJ, eds. Monocotyledons: systematics and evolu-
tion. Kew: Royal Botanic Gardens, 1–21.
Källersjö M, Farris JS, Chase MW, Bremer B, Fay MF,
Humphries CJ, Petersen G, Seberg O, Bremer K. 1998.
Simultaneous parsimony jackknife analysis of 2538 rbcL
DNA sequences reveals support for major clades of green
plants, land plants, seed plants and flowering plants. Plant
Systematics and Evolution 213: 259–287.
Kishino H, Thorne JL, Bruno WJ. 2001. Performance of a
divergence time estimation method under a probabilistic
model of rate evolution. Molecular Biology and Evolution 18:
352–361.
Kress WJ, Prince LM, Hahn WJ, Zimmer EA. 2001. Unrav-
elling the evolutionary radiation of the families of the
Zingiberales using morphological and molecular evidence.
Systematic Biology 50: 926–944.
Les DH, Cleland MA, Waycott M. 1997. Phylogenetic
studies in Alismatidae, II: Evolution of marine angiosperms
(seagrasses) and hydrophily. Systematic Botany 22: 443–
463.
Lidgard S, Crane PR. 1990. Angiosperm diversification and
Cretaceous floristic trends: a comparison of palynofloras and
leaf macrofloras. Paleobiology 16: 77–93.
Meerow AW, Fay MF, Guy CL, Li Q-B, Zaman FQ, Chase
MW. 1999. Systematics of Amaryllidaceae based on cladistic
analysis of plastid rbcL and trnL-F sequence data. American
Journal of Botany 86: 1325–1345.
Rydin C, Källersjö M. 2002. Taxon sampling and seed plant
phylogeny. Cladistics 18: 485–513.
Sanderson MJ. 1990. Estimating rates of speciation and
evolution: a bias due to homoplasy. Cladistics 6: 387–391.
Sanderson MJ. 1997. A nonparametric approach to estimat-
ing divergence times in the absence of rate constancy. Molec-
ular Biology and Evolution 14: 1218–1231.
Sanderson MJ. 1998. Estimating rate and time in molecu-
lar phylogenies: Beyond the molecular clock? In: Soltis DE,
Soltis PS, Doyle JJ, eds. Molecular systematics of plants II:
DNA sequencing. Boston: Kluwer Academic Press, 242–
264.
Sanderson MJ. 1999. r8s 0.20. Computer program and
documentation available from http://phylo.ucdavis.edu/r8s/
r8s.html
Sanderson MJ. 2002. Estimating absolute rates of molecular
evolution and divergence times: a penalized likelihood
approach. Molecular Biology and Evolution 19: 101–109.
Sanderson MJ, Doyle JA. 2001. Sources of error and confi-
dence intervals in estimating the age of angiosperms from
rbcL and 18S rDNA data. American Journal of Botany 88:
1499–1516.
Soltis DE, Soltis PS, Chase MW, Mort ME, Albach DC,
Zanis M, Savolainen V, Hahn WH, Hoot SB, Fay MF,
Axtell M, Swenson SM, Prince LM, Kress WJ, Nixon
KC, Farris JS. 2000. Angiosperm phylogeny inferred from
18S rDNA, rbcL, and atpB sequences. Botanical Journal of
the Linnean Society 133: 381–461.
Stockey RA. 2003. The fossil record of basal monocots.
Abstract in Monocots III. Los Angeles, California: Rancho
Santa Ana Botanic Garden, 89.
Swofford DL. 2001. PAUP: phylogenetic analysis using par-
simony (*and other methods). Version 4.0b6. Sunderland:
Sinauer Associates.
Thorne JL, Kishino H. 2002. Divergence time and evolution-
ary rate estimation with multilocus data. Systematic Biology
51: 689–702.
Vinnersten A, Bremer K. 2001. Age and biogeography of
major clades in Liliales. American Journal of Botany 88:
1695–1703.
Wikström N, Savolainen V, Chase MW. 2001. Evolution of
the angiosperms: calibrating the family tree. Proceedings of
the Royal Society of London B 268: 2211–2220.
Wilson MA, Gaut B, Clegg MT. 1990. Chloroplast DNA
evolves slowly in the palm family (Arecaceae). Molecular
Biology and Evolution 7: 303–314.
Zwickl DJ, Hillis DM. 2002. Increased taxon sampling
greatly reduces phylogenetic error. Systematic Biology 51:
588–598.
APPENDIX
GenBank accession numbers to the rbcL sequences for
all genera sampled.
INCERTAE SEDIS
Dasypogonaceae: Calectasia (AJ286557), Dasypogon
(AF206758), Kingia (AJ404840), Petrosaviaceae: Japo-
nolirion (AF206784), Petrosavia (AF206806).

AGE OF MONOCOT GROUPS 395
© 2004 The Linnean Society of London, Botanical Journal of the Linnean Society, 2004, 146, 385 –398
ACORALES
Acoraceae: Acorus (M91625).
ALISMATALES
Alismataceae: Alisma (L08759), Baldellia (U80677),
Damasonium (U80678), Echinodorus (U80679), Luro-
nium (U80680), Ranalisma (U80681), Sagittaria
(L08767), Wiesneria (U80682), Aponogetonaceae:
Aponogeton (U80684), Araceae: Amorphophallus
(AF497060), Anchomanes (AF497108), Anthurium
(AJ005627), Ariopsis (L10255), Arisaema (AF497109),
Dieffenbachia (AJ005631), Filarum (AF497110),
Gonatopus (AF497111), Gymnostachys (M91629),
Hapaline (AF497112), Landoltia (AY034223), Lasia
(L10250), Lemna (AY034234), Montrichardia (L10248),
Orontium (AJ005632), Peltandra (AJ005628), Philoden-
dron (AJ005623), Pistia (M96963), Pseudodracontium
(AF497106), Scindapsus (AJ005625), Spathiphyllum
(AJ235807), Spirodela (AY034222), Symplocarpus
(L10247), Typhonium (AF497113), Wolffia (AY034254),
Wolffiella (AY034244), Xanthosoma (AJ007543),
Zamioculcas (AJ005624), Zantedeschia (AF065474),
Butomaceae: Butomus (U80685), Cymodoceaceae:
Amphibolis (U80686), Cymodocea (U80687), Halodule
(U80689), Syringodium (U03727), Thalassodendron
(U80692), Hydrocharitaceae: Apalanthe (U80693),
Blyxa (U80694), Egeria (AB004887), Elodea
(AB004888), Enhalus (AB004889), Halophila (U80698),
Hydrilla (AB004891), Hydrocharis (U80701), Lagarosi-
phon (AB004893), Limnobium (AB004894), Najas
(U03731), Nechamandra (U80706), Ottelia (AB004895),
Stratiotes (AB004896), Thalassia (AB004897), Vallisne-
ria (U03726), Juncaginaceae: Cycnogeton (U80713),
Lilaea (U80715), Triglochin (U80714), Limnocharita-
ceae: Hydrocleys (AB004900), Limnocharis (U80717),
Posidoniaceae: Posidonia (U80718), Potamogetonaceae:
Coleogeton (U80727), Groenlandia (U80720), Lepilaena
(U80729), Potamogeton (L08765), Zannichellia
(U03725), Ruppiaceae: Ruppia (U03729), Scheuchzeri-
aceae: Scheuchzeria (U03728), Tofieldiaceae: Pleea
(AJ131774), Tofieldia (AJ286562), Zosteraceae: Hetero-
zostera (U80730), Phyllospadix (U80731), Zostera
(AY077964).
ARECALES
Arecaceae: Acrocomia (AY044625), Aiphanes
(AJ404831), Allagoptera (AJ404828), Ammandra
(AJ404838), Aphandra (AJ404837), Archontophoenix
(AF449156), Areca (AJ404819), Arenga (AJ404788),
Asterogyne (AJ404833), Astrocaryum (AY012510),
Attalea (AJ404829), Bactris (AY044627), Balaka
(AJ404814), Barcella (AY044630), Beccariophoenix
(AJ404826), Bentinckia (AY012499), Borassodendron
(AJ404768), Borassus (AY012469), Brassiophoenix
(AJ404815), Burretio (AY012500), Butia (AY044632),
Calamus (AJ404775), Calyptrocalyx (AY012501), Calyp-
tronoma (AJ404832), Caryota (AJ404790), Ceroxylon
(AJ404781), Chamaedorea (AF206748), Chamaerops
(AJ404754), Chambeyronia (AY012489), Chelyocarpus
(AY012457), Chuniophoenix (AJ404764), Clinostigma
(AF449157), Coccothrinax (AJ404751), Cocos
(AY012507), Corypha (AJ404762), Cryosophila
(AJ404747), Cyphophoenix (AJ404821), Cyrtostachys
(AJ404810), Desmoncus (AY044628), Dictyocaryum
(AY012479), Dictyosperma (AY012503), Drymophloeus
(AY012494), Dypsis (AY012486), Elaeis (AJ404830), cf.
Eremospatha (AJ404773), Eugeissona (AJ404774),
Euterpe (AJ404802), Gastrococos (AY044629), Gaussia
(AJ404784), Geonoma (AJ404834), Gronophyllum
(AJ404816), Guihaia (AJ404755), Hedyscepe
(AJ404807), Howea (AY012492), Hydriastele
(AJ404817), Hyophorbe (AJ404785), Hyospathe
(AJ404804), Hyphaene (AY012470), Iguanura
(AJ404820), Iriartea (AF233088), Itaya (AJ404748),
Johannesteijsmannia (AJ404758), Kentiopsis
(AJ404809), Kerriodoxa (AJ404765), Laccospadix
(AJ404812), Lemurophoenix (AJ404801), Leopoldinia
(AJ404798), Licuala (AY012462), Linospadix
(AF449158), Livistona (AJ404757), Lodoicea
(AJ404769), Lytocaryum (AY044633), Manicaria
(AJ404797), Marojejya (AJ404825), Masoala
(AJ404824), Mauritia (AJ404777), Metroxylon
(AF233089), Nannorrhops (AJ404763), Nenga
(AJ404818), Neonicholsonia (AJ404803), Nypa
(AJ404778), Oenocarpus (AY044624), Oncocalamus
(AJ404776), Oncosperma (AY012505), Orania
(AJ404796), Oraniopsis (AJ404782), Orbignya
(AY012508), Phoenix (M81814), Phytelephas
(AJ404835), Podococcus (AF233090), Prestoea
(AY012487), Pritchardiopsis (AY012464), Pseudophoe-
nix (AJ404779), Ptychosperma (AY012495), Ravenea
(AY012475), Reinhardtia (AJ404799), Rhapidophyl-
lum (AJ404753), Rhapis (AJ404756), Rhopalostylis
(AJ404808), Roscheria (AJ404822), Roystonea
(AY012488), Sabal (AJ404766), Salacca (AY012472),
Satranala (AJ404771), Scheelea (AY044636), Schippia
(AJ404749), Sclerosperma (AJ404823), Serenoa
(AJ404760), Socratea (AY012480), Syagrus (AY044634),
Synechanthus (AJ404786), Thrinax (AJ404750),
Trachycarpus (AJ404752), Trithrinax (AJ404745), Veit-
chia (AJ404813), Voanioa (AY044635), Wallichia
(AJ404792), Washingtonia (AY012465), Wendlandiella
(AY012477), Wettinia (AJ404794).
ASPARAGALES
Alliaceae: Acis (Z77256), Agapanthus (Z69221), Allium
(Z69204), Amaryllis (Z69219), Boophane (AF116945), cf.
Brunsvigia (AF116946), Caliphruria (AF116947),
Calostemma (AF116948), Clivia (L05032), Crinum
(AF116951), Cryptostephanus (AF116952), Cyrtanthus
(AF116953), Eucharis (AF116954), Eucrosia
(AF116955), Eustephia (AF116956), Galanthus
(Z69218), Gethyllis (AF116957), Gilliesia (Z69208),
Griffinia (AF116958), Habranthus (AF116959), Hae-
manthus (AF116960), Hannonia (AF116961), Hessea
(AF116962), Hieronymiella (AF116963), Hippeastrum
(AF206776), Hymenocallis (AF116965), Ipheion
(Z69201), Ismene (AF116966), Leptochiton (AF116970),
Leucocoryne (Z69199), Lycoris (AB034753), Milula
(AF116991), Narcissus (AF116972), Nerine (AF116973),
Nothoscordum (Z69202), Pamianthe (AF116974), Pan-
cratium (AF116975), Paramongaia (AF116976), Phae-

396 T. JANSSEN and K. BREMER
© 2004 The Linnean Society of London, Botanical Journal of the Linnean Society, 2004, 146, 385 –398
dranassa (AF116977), Proiphys (AF116978), Rauhia
(AF116979), Rhodophiala (AF116980), Scadoxus
(AF116981), Solaria (Z69207), Sprekelia (AF116982),
Stenomesson (Z69217), Sternbergia (AF116984), Stru-
maria (AF116985), Traubia (AF116986), Tristagma
(Z69206), Tulbaghia (Z69203), Ungernia (AF116987),
Vagaria (AF116988), Worsleya (AF116989), Zephyran-
thes (AF116990), Asparagaceae: Acanthocarpus
(Z77296), Agave (AF206729), Albuca (Z69223),
Androstephium (AJ311060), Anemarrhena (Z77251),
Anthericum (Z69225), Aphyllanthes (Z77259), Arthropo-
dium (Z69233), Asparagus (L05028), Aspidistra
(Z77269), Behnia (Z69226), Bessera (Z69215), Bowiea
(Z69237), Brodiaea (Z69210), Calibanus (Z77276),
Camassia (Z69238), Campylandra (AB029835),
Chamaexeros (Z77298), Chlorogalum (Z69228), Chloro-
phytum (L05031), Comospermum (Z73679), Convallaria
(D28334), Danae (L05034), Dasylirion (AB029847),
Dichelostemma (AJ311063), Disporopsis (D17373), Dra-
caena (AB029848), Echeandia (Z69212), Eriospermum
(Z77277), Hemiphylacus (Z73688), Herreria (Z69230),
Heteropolygonatum (AB029831), Hosta (L10253), Hya-
cinthus (AF116995), Ledebouria (L05038), Leucocrinum
(Z77252), Liriope (Z77271), Lomandra (L05039), Maian-
themum (D17378), Muilla (Z69213), Muscari (Z77278),
Myrsiphyllum (Z77260), Nolina (L05030), Ophiopogon
(AB029840), Ornithogalum (Z69224), Paradisea
(Z69229), Peliosanthes (AB029843), Polianthes
(Z69227), Polygonatum (AB029827), Reineckea
(AB029834), Rohdea (AB029836), Ruscus (Z77274),
Sansevieria (Z73698), Scilla (D28161), Semele (Z77275),
Smilacina (D17380), Sowerbaea (Z69234), Thysanotus
(Z69236), Tricalistra (AB029839), Triteleia (Z69198),
Tupistra (AB029838), Whiteheadia (Z77279), Xerolirion
(Z77299), Asteliaceae: Astelia (Z77261), Collospermum
(Y14986), Milligania (Z73693), Blandfordiaceae: Bland-
fordia (Z73694), Boryaceae: Alania (Y14982), Borya
(Z77262), Doryanthaceae: Doryanthes (Z73697), Hypoxi-
daceae: Curculigo (Z73701), Empodium (Y14987),
Hypoxis (Z73702), Pauridia (Y14991), Rhodohypoxis
(Z77280), Spiloxene (Z77281), Iridaceae: Alophia
(AJ309678), Anomatheca (Z73703), Aristea (Z77282),
Babiana (AJ309673), Belamcanda (AJ307078), Bobar-
tia (AJ307079), Calydorea (AJ309682), Chasmanthe
(AJ309660), Cipura (AJ309681), Crocus (AJ309668),
Cypella (AJ309683), Dietes (AJ307080), Diplarrhena
(AJ309686), Eleutherine (Z77283), Ennealophus
(AJ309684), Ferraria (AJ307081), Freesia (Z77284), Gal-
axia (AJ309685), Geissorhiza (AJ309676), Gelasine
(AJ309674), Geosiris (Z77285), Gladiolus (Z77286),
Gynandriris (AJ309698), Herbertia (AJ309692), Hesper-
antha (AJ309656), Hesperoxiphion (AJ309677), Home-
ria (AJ309691), Iris (L05037), Isophysis (Z77287), Ixia
(Z77288), Klattia (AJ309667), Lapeirousia (AJ309665),
Libertia (AJ309687), Micranthus (AJ309662), Moraea
(AJ307084), Neomarica (AJ309679), Nivenia (Z77289),
Olsynium (AJ309688), Orthrosanthus (L10249), Pa r-
danthopsis (AJ309696), Patersonia (AJ277879), Pillan-
sia (AJ309671), Radinosiphon (AJ309661), Romulea
(AJ309659), Savannosiphon (AJ309664), Schizostylis
(AJ309657), Sisyrinchium (Z77290), Solenomelus
(AJ309689), Sparaxis (AJ309669), Syringodea
(AJ309670), Thereianthus (AJ309663), Tigridia
(AJ309680), Trimezia (AJ309672), Tritonia (AJ309675),
Tritoniopsis (AJ309658), Watsonia (AJ309666), Witse-
nia (AJ277880), Ixioliriaceae: Ixiolirion (Z73704),
Lanariaceae: Lanaria (Z77313), Orchidaceae: Acanthep-
hippium (AF074100), Acianthus (AF074101), Acineta
(AF074102), Aeranthes (AF074104), Altensteinia
(AF074105), Ancistrochilus (AF264152), Angraecum
(AF074106), Ansellia (AF074107), Anthogonium
(AF264153), Aplectrum (AF074108), Apostasia
(Z73705), Arethusa (AF074109), Arpophyllum
(AF074110), Arundina (AF074111), Basiphyllaea
(AF264155), Bifrenaria (AF074112), Bletia (AF264156),
Bletilla (AF074114), Bothriochilus (AF264158), Bulbo-
phyllum (AF074115), Cadetia (D58406), Caladenia
(AF074116), Calanthe (AF264159), Calochilus
(AF074118), Calopogon (AF074119), Calypso
(AF074120), Catasetum (AF074121), Cattleya
(AF074122), Cephalanthera (AF074123), Chiloglottis
(AF074124), Chloraea (AF074125), Chysis (AF074126),
Cleisostoma (AF074130), Cleistes (AF074127), Clem-
atepistephium (AF074131), Coelia (AF074132), Coelog-
yne (AF074133), Collabium (AF264163), Coryanthes
(AF074134), Corybas (AF074135), Corymborkis
(AF074136), Cranichis (AF074137), Cryptarrhena
(AF074138), Cryptocentrum (AF074139), Cryptostylis
(AF074140), Cymbidium (AF074141), Cypripedium
(AF074142), Cyrtopodium (AF074143), Dendrobium
(AF074145), Dendrochilum (AF264164), Diaphananthe
(AF074147), Diceratostele (AF074148), Dichaea
(AF074149), Dilochia (AF264165), Dilomilis
(AF074150), Diplocaulobium (D58409), Disa
(AF274006), Diuris (AF074152), Dressleria (AF074153),
Duckeella (AF074154), Earina (AF074155), Eleorchis
(AF264166), Elleanthus (AF074156), Encyclia
(AF074157), Epigeneium (D58410), Epipactis (Z73707),
Epistephium (AF074162), Eria (AF074164), Eriaxis
(AF074165), Eriochilus (AF074166), Eriopsis
(AF074167), Erythrorchis (AF074168), Eulophia
(AF074170), Flickingeria (D58411), Galeandra
(AF074171), Glomera (AF074172), Glossodia
(AF074173), Goodyera (AF074174), Govenia
(AF074175), Grammatophyllum (AF074176), Habe-
naria (AF074177), Hexalectris (AF264168), Houlletia
(AF074178), Huntleya (AF074179), Isotria (AF074180),
Kegeliella (AF074181), Koellensteinia (AF074182),
Liparis (AF074183), Listera (AF074184), Lycaste
(AF074185), Lycomormium (AF074186), Lyperanthus
(AF074187), Malaxis (AF074188), Masdevallia
(AF074189), Maxillaria (AF074190), Megastylis
(AF074191), Meiracyllium (AF074192), Mexipedium
(AF074193), Microtis (AF074194), Mischobulbum
(AF264169), Monophyllorchis (AF074195), Mormodes
(AF074196), Neofinetia (AF074197), Neomoorea
(AF074198), Nephelaphyllum (AF264170), Nervilia
(AF074199), Neuwiedia (AF074200), Oncidium
(AF074201), Ophrys (AF074202), Orthoceras
(AF074204), Pa chyplectron (AF074205), Palmorchis
(AF074206), Paphiopedilum (AF074207), Phaius
(AF074210), Phalaenopsis (AF074211), Phragmipe-
dium (AF074212), Phreatia (AF074214), Platanthera
(AF074215), Platythelys (AF074216), Pleione
(AF264173), Pleurothallis (AF074217), Podochilus
(AF074218), Pogonia (AF074219), Polystachya
(AF074222), Ponthieva (AF074223), Pseuderia
(D58412), Pterostylis (AF074224), Satyrium

AGE OF MONOCOT GROUPS 397
© 2004 The Linnean Society of London, Botanical Journal of the Linnean Society, 2004, 146, 385 –398
(AF074226), Selenipedium (AF074227), Sobralia
(AF074228), Spathoglottis (AF264175), Spiranthes
(AF074229), Stanhopea (AF074230), Stellilabium
(AF074231), Tainia (AF264176), Thelymitra
(AF074232), Thunia (AF074233), Tipularia (AF074234),
Trichotosia (AF074235), Triphora (AF074236), Tropidia
(AF074237), Vanilla (AF074239), Xerorchis (AF074244),
Xylobium (AF074245), Zygopetalum (AF074246), Teco-
philaeaceae: Conanthera (Z77311), Cyanastrum
(U41572), Cyanella (Z77312), Kabuyea (Y17336), Odon-
tostomum (Z77314), Tecophilaea (Z73709), Walleria
(Y17338), Zephyra (Y17340), Xanthorrhoeaceae: Aloe
(L05029), Asphodeline (Z73681), Asphodelus (Z73682),
Astroloba (Z73683), Bulbine (Z73684), Bulbinella
(Z73685), Caesia (Z77297), Eremurus (Z73686), Eustre-
phus (AF116996), Gasteria (Z73687), Geitonoplesium
(AF116997), Haworthia (L05035), Hemerocallis
(L05036), Herpolirion (Z77303), Jodrellia (Y17335),
Johnsonia (Z77304), Kniphofia (Z73689), Lomatophyl-
lum (Z73690), Pasithea (Z77305), Phormium (Z69232),
Poellnitzia (Z73691), Simethis (Z69231), Stawellia
(Z77306), Stypandra (Z77307), Trachyandra (Z73692),
Tricoryne (Z77308), Xanthorrhoea (Z73710), Xerone-
mataceae: Xeronema (Z69235).
COMMELINALES
Commelinaceae: Amischotolype (AF312239), Aneilema
(AF312252), Anthericopsis (AF312259), Belosynapsis
(AF312257), Buforrestia (AF036886), Callisia
(AF312248), Cochliostema (AF312244), Coleotrype
(AF312243), Commelina (L05033), Cyanotis
(AF312241), Dichorisandra (AF312242), Elasis
(AF312251), Floscopa (AF312255), Geogenanthus
(AF312261), Gibasis (AF312250), Murdannia
(AF312256), Palisota (AF312240), Pollia (AF312262),
Polyspatha (AF312263), Rhopalephora (AF312264),
Siderasis (AF312254), Stanfieldiella (AF312265),
Thyrsanthemum (AF312246), Tinantia (AF312260),
Tradescantia (L05041), Tripogandra (AF312249), Wel-
denia (AF312245), Haemodoraceae: Anigozanthos
(AJ286556), Wachendorfia (AF312266), Hanguanaceae:
Hanguana (AJ404842), Philydraceae: cf. Helmholtzia
(AF206774), Philydrella (AF206808), Pontederiaceae:
Eichhornia (U41574), Heteranthera (U41581), Hydro-
thrix (U41582), Monochoria (U41588), Pontederia
(U41593).
DIOSCOREALES
Burmanniaceae: Burmannia (AF206742), Geomitra
(AF307488), Gymnosiphon (AF307489), Dioscoreaceae:
Avetra (AF307476), Dioscorea (D28327), Nanarepenta
(AF307473), Rajania (AF307472), Stenomeris
(AF307475), Tacca (AJ286561), Tamus (AF307474),
Trichopus (AF307477), Nartheciaceae: Aletris (M. W.
Chase, pers. comm.), Lophiola (AJ417897), Narthecium
(AJ286560).
LILIALES
Alstroemeriaceae: Alstroemeria (Z77254), Bomarea
(Z77255), Leontochir (AY120369), Campynemataceae:
Campynema (Z77264), Campynemanthe (AJ276349),
Colchicaceae: Androcymbium (Z77265), Burchardia
(Z77266), Colchicum (L12673), Disporum (D17376), Glo-
riosa (D28867), Iphigenia (AJ417893), Petermannia
(Z77267), Tripladenia (Z77268), Uvularia (AB009952),
Liliaceae: Amana (AB024385), Calochortus (Z77263),
Cardiocrinum (AB034918), Clintonia (AB056856),
Erythronium (D28156), Fritillaria (Z77293), Gagea
(AB034752), Lilium (AB034926), Lloydia (Z77294),
Medeola (D28158), Nomocharis (Z77295), Notholirion
(AB034919), Prosartes (D17374), Scoliopus (D28163),
Streptopus (D17381), Tricyrtis (D17382), Tulipa
(Z77292), Luzuriagaceae: Drymophila (AJ276346),
Luzuriaga (Z77300), Melanthiaceae: Amianthium
(AJ417895), Chamaelirium (AJ276347), Daiswa
(D28155), Heloniopsis (AJ417894), Kinugasa (D28157),
Melanthium (AJ276348), Paris (D28159), Trillium
(AB018845), Veratrum (D28168), Xerophyllum
(AJ276350), Philesiaceae: Lapageria (Z77301), Philesia
(Z77302), Rhipogonaceae: Rhipogonum (Z77309), Smila-
caceae: Smilax (Z77310).
PANDANALES
Cyclanthaceae: Carludovica (AF197596), Cyclanthus
(AY007660), Ludovia (L10251), Sphaeradenia
(AJ235808), Pandanaceae: Freycinetia (AF206770),
Pandanus (M91632), Stemonaceae: Croomia (D28154),
Pentastemona (AF307490), Stemona (AJ131948), Sti-
choneurone (AF307492), Velloziaceae: Barbacenia
(AJ131946), Vellozia (L19970).
POALES
Anarthriaceae: Anarthria (AF148760), Hopkinsia
(AF148777), Lyginia (AF148787), Bromeliaceae: Aech-
mea (L19978), Ananas (L19977), Catopsis (L19976),
Glomeropitcairnia (L19975), Hechtia (L19974), Puya
(L19973), Tillandsia (L19971), Centrolepidaceae:
Centrolepis (AF148766), Cyperaceae: Abildgaardia
(Y12985), Actinoscirpus (Y12953), Alinula (AJ278290),
Ascolepis (Y13003), Becquerelia (Y12948), Blysmus
(AJ404700), Bolboschoenus (Y12996), Bulbostylis
(Y12992), Carex (Y12999), Carpha (AF307909), Caustis
(Y12976), Chorizandra (AJ419939), Chrysithrix
(AJ419938), Cladium (Y12988), Coleochloa (Y12975),
Courtoisina (AY040590), Cyperus (Y13016), Desmo-
schoenus (AJ404701), Eleocharis (Y13012), Eriophorum
(Y12951), Ficinia (Y12963), Fimbristylis (Y13009),
Fuirena (Y12971), Gahnia (Y12973), Hellmuthia
(Y13000), Hypolytrum (Y12956), Isolepis (Y12962),
Kobresia (U49232), Kyllinga (Y12979), Kyllingiella
(AY040592), Lepironia (Y12957), Lipocarpha (Y12991),
Mapania (Y12955), Mesomelaena (Y12959), Nemum
(Y12945), Oreobolus (Y12972), Oxycaryum (Y13006),
Pleurostachys (Y12989), Pycreus (Y13005), Remirea
(AY040593), Rhynchospora (AF206818), Schoenoplectus
(Y12943), Schoenus (Y12983), Scirpoides (Y13001),
Scleria (Y12968), Scripus (AJ297509), Sphaerocyperus
(AJ404699), Trichophorum (Y12969), Ecdeiocoleaceae:
Ecdeiocolea (AJ286559), Georgeantha (AF148772),
Eriocaulaceae: Eriocaulon (L10252), Lachnocaulon
(Y13019), Mesanthemum (AJ419941), Paepalanthus

398 T. JANSSEN and K. BREMER
© 2004 The Linnean Society of London, Botanical Journal of the Linnean Society, 2004, 146, 385 –398
(AJ419942), Syngonanthus (AJ419943), Tonina
(AF036878), Flagellariaceae: Flagellaria (L12678),
Hydatellaceae: Trithuria (AF458076), Joinvilleaceae:
Joinvillea (L01471), Juncaceae: Distichia (AJ419944),
Juncus (L12681), Luzula (AJ419945), Marsipposper-
mum (AJ419946), Oxychloe (Y12978), Rostkovia
(AJ419947), Mayacaceae: Mayaca (AJ419948), Poaceae:
Amphipogon (U88403), Anomochloa (AF021875), Aris-
tida (U31359), Arundo (U13226), Avena (L15300), Bam-
busa (M91626), Bromus (Z49836), Cenchrus (L14622),
Centropodia (U31100), Chasmanthium (U31101),
Chusquea (U13227), Cyperochloa (U88404), Danthonia
(U31102), Elymus (Z49837), Elytrophorus (U88405),
Enneapogon (U31103), Eragrostis (U31104), Eremium
(Z49840), Eriachne (AF352580), Guaduella (AF164778),
Gynerium (U31105), Hordeum (Z49842), Hyparrhenia
(U31436), Karroochloa (U31437), Leersia (U13228),
Leymus (Z49843), Lithachne (U13231), Merxmuellera
(U31438), Moliniopsis (U31439), Monachather
(U31379), Neurachne (X55827), Oryza (D00207),
Pennisetum (L14623), Peridictyon (Z49845), Pharus
(AJ419950), Phragmites (U29900), Phyllostachys
(U13230), Plinthanthesis (U31440), Pseudoroegneria
(Z49846), Puccinellia (L14621), Puelia (AF164780),
Rytidosperma (U31441), Setaria (X79900), Stipa
(U31442), Stipagrostis (U31378), Streptochaeta
(AJ419949), Styppeiochloa (U88406), Thysanolaena
(U31380), Tristachya (U31381), Triticum (D00206),
Zea (11990232), Zizania (L05043), Rapateaceae:
Cephalostemon (AF036884), Kunhardtia (AF036883),
Rapatea (AF460969), Schoenocephalium (AF460970),
Spathanthus (AF460971), Stegolepis (L19972), Restion-
aceae: Acion (AF148762), Alexgeorgea (AF148759),
Baloskion (AF148764), Calorophus (AF148765), Chaet-
anthus (AF148782), Chordifex (AF148789), Coleocarya
(AF148769), Dapsilanthus (AF148780), Desmocladus
(AF148770), Dielsia (AF148771), Elegia (L12675),
Empodisma (AF148775), Eurychorda (AF148790),
Guringalia (AF148763), Harperia (AF148776), Kulinia
(AF148778), Lepidobolus (AF148779), Leptocarpus
(AF307924), Lepyrodia (AF148785), Loxocarya
(AF148786), Meeboldina (AF148783), Melanostachya
(AF148788), Restio (AJ419951), Saropsis (AF148791),
Sporadanthus (AF148793), Taraxis (AF148794), Trem-
ulina (AF148792), Tyrbastes (AF148795), Winifredia
(AF148796), Sparganiaceae: Sparganium (M91633),
Thurniaceae: Prionium (U49223), Thurnia (AF036881),
Typhaceae: Typha (M91634), Xyridaceae: Abolboda
(AJ419952), Aratitiyopea (AF461418), Orectanthe
(AF036880), Xyris (AJ286563).
ZINGIBERALES
Cannaceae: Canna (AF378774), Costaceae: Costus
(AF243510), Dimerocostus (AF243838), Monocostus
(AF243839), Tapeinochilos (AF243840), Heliconiaceae:
Heliconia (L05451), Lowiaceae: Orchidantha
(AF243841), Marantaceae: Calathea (AF243842),
Maranta (AF378768), Marantochloa (AF378769),
Pleiostachya (AF378781), Musaceae: Ensete
(AF243843), Musa (AF378770), Musella (AF243844),
Strelitziaceae: Phenakospermum (AF243845), Ravenala
(L05459), Strelitzia (AF243846), Zingiberaceae: Globba
(AF243847), Hedychium (AF243848), Riedelia
(AF243849), Zingiber (AF243850).