Systematic Botany (2010), 35(1): pp. 121–139
© Copyright 2010 by the American Society of Plant Taxonomists
Systematics of the Aquatic Angiosperm Genus Myriophyllum (Haloragaceae)
Michael L. Moody1,2,4 and Donald H. Les3
1
University of Western Australia, School of Plant Biology, 35 Stirling Highway, Crawley, Western Australia 6009 Australia
2
Department of Environment and Conservation, Science Division, 17 Dick Perry Avenue, Kensington,
Western Australia 6151 Australia
3
University of Connecticut, Department of Ecology and Evolutionary Biology, Storrs, Connecticut 06269-3043 U. S. A.
4
Author for correspondence (miclmood@cyllene.uwa.edu.au)
Communicating Editor: Mark P. Simmons
Abstract—The angiosperm genus Myriophyllum (Haloragaceae) is among the most species-rich genera of aquatic core-eudicots. Myriophyllum
has a cosmopolitan distribution with its center of diversity in Australia (> 37 endemics). The widespread invasive species of the genus
(M. aquaticum, M. heterophyllum, and M. spicatum) have drawn attention from international natural resource managers. Myriophyllum species
are notoriously difficult to identify using vegetative morphology alone, which commonly is all that is available for these highly clonal plants.
The relationships among taxa have been difficult to determine with suspected parallelisms in sex expression, sepal and petal loss, and reduced
stamen number. A molecular phylogenetic approach was taken to examine relationships among taxa and to employ molecular markers for the
reliable identification of Myriophyllum species. This study included » 80% of the known Myriophyllum species. Both nrDNA ITS and cpDNA
matK and trnK data were used to examine phylogenetic relationships among species. The nrDNA ITS data proved highly variable and could
differentiate between all but one species pair examined. These analyses also uncovered multiple cryptic species among Australian complexes.
Phylogenetic results support major realignments in the subgeneric classification including a recombination for the rare monotypic genus
Meziella, which was nested within Myriophyllum. Here we present the new combinations and taxa Myriophyllum subgenus Meziella, sections
Pectinatum and Pelonastes, subsections Isophylleae and Nudiflorum with the new combination Myriophyllum trifidum to accommodate the
former monospecific genus Meziella.
Keywords—aquatic plant, Bayesian analysis, cryptic species, molecular phylogeny, matK, nrDNA ITS.
The characteristic morphology of Myriophyllum is a submerged stem with whorled or alternate, pectinate leaves. The
submerged stems have vascular lacunae (air chambers) that
allow the plants to transfer oxygen to the submerged roots as
well as to confer buoyancy. Stem size varies widely from compact mud-flat forms to some plants surfacing from a rooted
base at water depths of > 10 m. Most Myriophyllum have an
emergent inflorescence with flowers borne individually in the
axils of emergent leaves (bracts) and most species are monoecious with female flowers basal, male flowers distal, and hermaphrodite flowers often intermediate.
In such a large complex genus, infrageneric delimitation
has emphasized departures from this common morphological
theme (Schindler 1905; Orchard 1986). Some species display
dioecy (e.g. M. aquaticum, M. implicatum Orchard), whereas
others are monoecious but with separate male and female
emergent inflorescences on the same individual (e.g. “M. propinquum Alliance” species [Table 1], M. lophatum Orchard).
Strictly hermaphroditic flowers also are found (e.g. M. callitrichoides Orchard, M. mattogrossense Hoehne, M. muricatum Orchard) as are submerged flowers (e.g. M. dicoccum
F. Muell., M. farwellii Morong, M. humile (Raf.) Morong). Some
species are highly diminutive, possess linear leaves, often lack
differentiated submerged leaves and grow only in shallow
water or on mudflats (e.g. M. drummondii Benth., M. lophatum,
M. pedunculatum Hook. f., M. votschii Schindl.), whereas several display both the common submerged morphology as
well as a “mud flat form” (e.g. M. heterophyllum, M. humile,
M. pinnatum Britton, Sterns & Poggenb.).
Taxonomic History—Myriophyllum has been hypothesized
as distinct within Haloragaceae due to a combination of characteristics including its aquatic habit (also found in Meionectes
R. Br., Meziella Schindl., and Proserpinaca L.), propensity
towards monoecy (also found in Laurembergia P. J. Bergius)
and a fruit that splits at maturity into two or four individual
nutlets (not found elsewhere in the family; Schindler 1905;
Orchard 1986). Myriophyllum was found to be paraphyletic in
The angiosperm genus Myriophyllum L. (Haloragaceae R. Br.)
is among the most species-rich (» 68 spp.) of the aquatic coreeudicots (as defined by APG II 2003). These “watermilfoils”
have a world-wide distribution (except Antarctica) with a
center of diversity in Australia (42 spp.; 37 endemic). North
America (14 spp.; seven endemic) and Asia (16 spp.; eight
endemic) also have high continental diversity and share
seven common species (at least four due to introductions).
Myriophyllum is well-known for its invasive species. The
aggressive Eurasian M. spicatum L. (Eurasian watermilfoil)
and South American M. aquaticum (Vell.) Verdc. (parrotfeather)
are now established on most continents and listed as noxious
weeds in several U. S. A. states. The North American endemic
M. heterophyllum Michx. reportedly is naturalized in Europe
(Cirujano and Medina 1997; Wimmer 1997) and Asia (Yu et al.
2002), and also is considered to be invasive outside its endemic
range in the northeast and northwest United States (Les and
Mehrhoff 1999; Moody and Les 2002). Hybridization also has
been shown to play a role in North American invasions with
two hybrid lineages now recognized (M. spicatum × M. sibiricum and M. heterophyllum × M. laxum; Moody and Les 2002).
Infrageneric relationships in Myriophyllum have proven to
be particularly frustrating using morphology alone, as stated
by Meijden (1969, p. 303), “…. the species show a reticulate
affinity by parallelism, especially as regards reduction in
both vegetative and sexual organs, the usefulness of distinguishing infrageneric taxa is debatable and not advisable.”
Reliable morphological identification of Myriophyllum is
particularly difficult in the field. Plant identification is most
complex when reproductive structures are lacking, as is common among many aquatic taxa (Sculthorpe 1967; Cronk and
Fennesy 2001). Even Myriophyllum species that presumably
are distantly related are not easily differentiated when only
submerged vegetative characters are available, as often is the
case (Aiken 1981; Orchard 1986). Furthermore, common vegetative plasticity in Myriophyllum (e.g. submerged and emergent vegetative forms) compounds the problem.
121
122
SYSTEMATIC BOTANY
Table 1. Myriophyllum classification systems as interpreted from
Schindler (1905; column 1) and Orchard (1986; column 2). Orchard did not
present a formal classification at the subgeneric level. The Myriophyllum
“Alliances” follow the terminology of Orchard (1986). *Taxa sampled for
our analyses. Numbers following a species name refer to the number of
individuals sampled for these analyses (usually in column two), when
taxa only present in column 1 then included there. Not all taxa sampled
are included here (See Appendix 1 for further sampling details).
Genus Myriophyllum (Schindler 1905)
subgenus Eumyriophyllum
section Pentapteris
subsection Spirophyllum
M. gracile
*M. trachycarpum
*M. filiforme
subsection Pelonastes
*M. tillaeoides
*M. longibracteolatum
*M. pedunculatum
*M. votschii
*M. amphibium
subsection Spondylophyllum
*M. ussuriense
*M. robustum
*M. verticillatum
M. propinquum
*M. brasiliense
*M. spicatum
*M. verrucosum
*M. elatinoides
M. indicum
*M. latifolium
subsection Leiocarpium
*M. alterniflorum
M. muelleri
section Tessaronia
subsection Trachycarpaeum
M. tetrandrum
M. tuberculatum
M. intermedium
M. axilliflorum
subsection Spondylastrum
*M. heterophyllum
*M. hippuroides
*M. pinnatum (2)
M. sparsiflorum
subsection Ptilophyllum
*M. humile (3)
*M. laxum (8)
*M. tenellum (3)
subgenus Brachytheca
M. integrifolium
*M. drummondii
M. glomeratum
subgenus Dicarpum
M. mezianum
*M. dicoccum
Genus Myriophyllum L. (Orchard 1986)
M. aquaticum Alliance
*M. aquaticum (Vell.) Verdc. (3)
*M. robustum Hook. f. (1)
M. aquaticum Alliance Associates
*M. verticillatum L. (4)
*M. heterophyllum Michx. (9)
*M. hippuroides Nutt. (1)
M. salsugineum Alliance
*M. salsugineum Orchard (2)
*M. quitense Kunth (4)
*M. triphyllum Orchard (1)
*M. caput-medusae Orchard (2)
M. porcatum Orchard
*M. verrucosum Lindl. (3)
M. salsugineum Alliance Associates
*M. spicatum L. (7)
M. muelleri Sonder
*M. decussatum Orchard (1)
M. indicum Willd.
M. tetrandrum Roxb.
M. propinquum Alliance
M. propinquum A. Cunn.
*M. variifolium Hook. f. (8)
*M. simulans Orchard (4)
*M. alpinum Orchard (2)
*M. crispatum Orchard (6)
*M. ussuriense Maxim. (2)
M. propinquum Alliance Associates
*M. papillosum Orchard (2)
*M. latifolium F. Muell. (2)
M. muricatum Alliance
*M. muricatum Orchard (3)
M. tuberculatum Roxb.
M. muricatum Alliance Associates
*M. dicoccum F. Muell. (2)
*M. balladoniense Orchard (1)
M. striatum Alliance
M. implicatum Orchard
M. striatum Orchard
M. costatum Orchard
M. striatum Alliance Associates
M. gracile Benth.
*M. trachycarpum F. Muell. (2)
*M. filiforme Benth. (2)
*M. petraeum Orchard (2)
M. mezianum Alliance
*M. coronatum Meijden (1)
M. mezianum Schindl.
M. siamense (Craib) Tardieu
M. bonii Tardieu
M. integrifolium Alliance
*M. limnophilum Orchard (1)
M. integrifolium Hook. f.
*M. drummondii Benth. (1)
*M. echinatum Orchard (1)
M. integrifolium Alliance Associates
M. glomeratum Schindl.
M. callitrichoides Alliance
M. callitrichoides Orchard
M. pedunculatum Alliance
*M. amphibium Labill. (1)
*M. pedunculatum Hook. f. (4)
*M. tillaeoides Diels (1)
M. pedunculatum Alliance Associates
*M. lophatum Orchard (2)
M. austropygmaeum Orchard
*M. votschii Schindl. (1)
M. pygmaeum Mattf.
[Volume 35
regard to the monotypic Meziella, in recent phylogenetic analyses (Moody and Les 2007a), but the sampling of Myriophyllum
was not comprehensive and ITS data alone proved ambiguous. Meziella is similar in habit to Myriophyllum but possesses
hermaphrodite flowers (although described as functionally
monoecious) and while forming four nutlets, they do not
split at maturity due to a persistent exocarp (Orchard and
Keighery 1993). The other aquatic Haloragaceae (Proserpinaca
and Meionectes) are distinct, having perfect, two- or three-merous flowers with a nut and are only distantly related (Moody
and Les 2007a).
A tribal status for Myriophyllum within Haloragaceae
(Myriophylleae) was applied by Schindler (1905). Orchard
(1975) supported this ranking and suggested that an elevation to subfamily might be warranted. Schindler’s treatment
of Myriophyllum recognized only 36 species, whereas recent
treatments of the genus now distinguish about 68 (Orchard
1980, 1981, 1986; Aiken 1981; Yu et al. 2002). Schindler’s (1905)
classification included three subgenera, two sections, seven
subsections and two series. Both Meijden (1969) and Orchard
(1986) suspected Schindler’s classification to be artificial.
Orchard (1986) proposed seven distinct “Alliances” based on
a range of characters for each group (Table 1). None of these
treatments had an explicit phylogenetic basis.
Here we use Bayesian and parsimony analyses of multiple molecular data sets (nrDNA ITS and cpDNA trnK and
matK) to: 1) examine phylogenetic relationships among Myriophyllum and evaluate species limits and subgeneric “Alliances”
as currently proposed (Table 1); 2) establish a clade-based
infrageneric classification scheme; 3) define morphological
character states that delimit taxonomic groups within Myriophyllum and evaluate the potential parallelisms pertaining
to sexual dimorphism and reductions in floral morphology;
and 4) determine the utility of ITS as a de facto “barcode”
among the often difficult-to-identify watermilfoils.
Materials and Methods
Taxon Sampling—Forty-three known species and multiple unknown
or undescribed Myriophyllum taxa were sampled; Meziella trifida (Nees)
Schindl., Laurembergia repens (L.) Berg., Trihaloragis hexandrus (F. Muell.)
M. L. Moody & D. H. Les, Gonocarpus montanus (Hook. f.) Orchard, and
Haloragis digyna Labill. also were sampled, the latter four as outgroup
taxa (Appendix 1). Sampling included all Myriophyllum species known
to Europe, North America, and South America, eight of 16 Asian species, 28 of 42 Australian species (mostly lacking narrow endemics), and
lacked the two African endemic species. Most taxa were sampled from
multiple accessions and, when possible, from across a wide geographic
range (Appendix 1). Most taxa were collected in the field and NaCl-CTAB
preserved (Rogstad 1992) while some were sampled from herbarium
specimens (Appendix 1). The sampling covers most alliances proposed
by Orchard (1986) and all of Schindler’s (1905) subgenera and sections
(Table 1).
We examined sequence data from the nrDNA ITS (here forward
referred to as ‘ITS’) and cpDNA matK + trnK region (here forward referred
to as ‘cpDNA’). Three data sets were constructed: (1) ITS had 71 accessions
(several not included in cpDNA data sets) including: a) multiple cloned
copies from M. papillosum Orchard and M. sibiricum Komarov, which
had polymorphisms, making direct sequencing techniques impractical
and cloning techniques necessary (as described in Moody and Les 2002);
b) multiple accessions of M. crispatum Orchard, M. filiforme Benth., M.
heterophyllum, M. laxum Schuttlw. ex Chapm., M. muricatum, M. pedunculatum, M. quitense Kunth, M. sibiricum, M. simulans Orchard, and M. variifolium Hook. f., which showed variation in ITS, but not cpDNA, across their
geographic range; and c) accessions of M. amphibium Labill. and M. echinatum Orchard, which did not amplify for cpDNA; (2) cpDNA included
60 accessions; (3) combined data included 60 accessions that were represented in both ITS and cpDNA data sets, sometimes using only a single
2010]
MOODY AND LES: MYRIOPHYLLUM SYSTEMATICS
copy of ITS when multiple were found in a single taxon (in these cases ITS
haplotypes formed an exclusive lineage; see Fig. 1).
DNA Extraction, PCR, and Sequencing—Total genomic DNA was
extracted from fresh, NaCl-CTAB preserved, and herbarium specimen
leaf material using a modified CTAB miniprep procedure (Doyle and
Doyle 1987) or the Qiagen DNeasy Plant Minikit (Qiagen Inc., Valencia,
California). Double-stranded DNAs were amplified using PCR following
the protocols and conditions of Moody et al. (2001) to amplify the ITS-1,
ITS-2, and 5.8S region of nuclear ribosomal DNA using the ITS4 and ITS5
primers or, in the case of several herbarium specimens, amplification was
conducted on smaller segments using ITS2 and ITS5 to amplify the ITS-1
region and ITS3 and ITS4 to amplify the ITS-2 region (White et al. 1990).
The cpDNA trnK introns and matK coding region were amplified using
the primers trnK-3914F and trnK-2R (Johnson and Soltis 1994). Several
additional primers were used to amplify trnK and matK from DNA of
herbarium leaf material including matK68F, matK1872R (Johnson and
Soltis 1994), matK900F (Moody and Les 2002), and trnKR and matK70R.
Cycle sequencing of ITS used combinations of the ITS2, ITS3, ITS4, and
Fig. 1. Phylogenetic relationships in Myriophyllum as indicated by
a majority rule consensus of 29,000 trees (after discarding burn-in) from
Bayesian analysis of ITS sequence data analyzed under a GTR + I + Γ
model. Harmonic mean –ln likelihood = 6,956.49. Dashed lines refer to
clades not resolved in the strict consensus of the parsimony analysis.
Numbers above branches refer to posterior probabilities and numbers
below branches are bootstrap support from parsimony analyses. When
numbers are lacking below branches parsimony did not resolve clades
or had bootstrap values < 50% supporting that node. Species accessions
followed by c1, c2, etc. refer to multiple copies cloned from individuals.
Species accessions followed by numbers refer to divergent genotypes.
EU = Europe, NA = North America, SA = South America, T = Tasmania,
WA = Western Australia. [* M. muricatum and M. dicoccum not included
in bootstrap analyses of ITS data, as discussed in text, thus bootstrap not
relevant].
123
ITS5 (White et al. 1990) primers. The trnK introns and matK region were
sequenced using trnK3914F, matK68F, matK1872R, matK900F, trnK360F,
trnK2R, matK70R, trnKR, and the Haloragaceae specific primer “trnk3F”
(Moody and Les 2007a). Sequences were obtained using Big Dye terminator technology on an ABI 3100 automated sequencer (Applied Biosystems,
Foster City, California).
Phylogenetic Analyses—Sequences were edited for polymorphisms
and/or variable sites using Sequencher 4.1.2 (Gene Code Corp., Ann
Arbor, Michigan) and aligned manually using a similarity (Simmons
2004) and event-based approach (Morrison 2006). Alignment minimized
indel events and sequence similarity was used to identify gap boundaries and optimize multiple alignment possibilities within gaps when necessary using MacClade 4.06 (Maddison and Maddison 2003). Combined
data had 1.42% (cpDNA 1.13%; ITS 2.53%) of sites coded as missing data.
The trnK3 intron could not be amplified for M. dicoccum and M. sp. nov.
542 and ITS 1 could not be amplified for M. dicoccum. Parsimony analyses were performed with PAUP* 4.0b10 (Swofford 2002) using heuristic
searches with random taxon addition sequences and tree bisection-reconnection with unordered, equally weighted characters and 1,000 analysis
replicates. Indels were treated as missing data and invariant indel regions
were removed. Standard measures of homoplasy, ensemble consistency
index (CI; Kluge and Farris 1969), ensemble rescaled consistency index
(RC), ensemble retention index (RI; Farris 1989), and level of internal support (bootstrap values; Felsenstein 1985) were calculated using PAUP*
4.0b10. Bootstrap analyses were conducted using 1,000 replicate heuristic
searches as above except for analyses of ITS where saved trees were limited to 500 for tree scores > 1,000.
Bayesian MCMC analyses (Yang and Rannala 1997) were performed
on each data set using MrBayes v3.1.2 (Huelsenbeck and Ronquist 2001).
The ITS and cpDNA trnK and matK regions (matK was treated with each
codon position as a separate data set using an individual best-fit model for
each) were each initially examined to determine the best-fit model using
MrModeltest v1.1b (Nylander 2003) as determined by the likelihood ratio
test (Felsenstein 1988). Bayesian analyses were performed with the best-fit
model for each character partition twice for 3.0 × 107 generations. Markov
Chain Monte Carlo was implemented with four heated chains and trees
were sampled every 1,000 generations. Trace plots of likelihood scores
were used to determine when stationarity was reached and compared
from independent runs to assess convergence. Consensus trees recovered
from each individual run were visually compared for topology and posterior probability and we also compared split frequencies using AWTY
(Nylander et al. 2008) to further determine consensus among duplicate
analyses of each of the three data sets analyzed. The first 1,000 trees were
discarded as burn-in, with the remaining trees used to generate a 50%
majority rule consensus tree where the percentage of the nodes recovered
represented each node’s posterior probability (PP). Nodal support was
determined using Bayesian PP ≥ 0.95 as the criterion for strong support.
Incongruence—Although data were combined regardless of outcome (see Discussion Below) incongruence was examined between ITS
(reduced to include only taxa in cpDNA) and cpDNA. Congruence of
data was tested using the incongruence length difference test (ILD; Farris
et al. 1995) as implemented in PAUP* 4.0b10. ITS and cpDNA data were
analyzed using 1,000 homogeneity replicates with heuristic searches as
described above under parsimony analysis. Incongruence also was determined visually by comparing tree topologies. Where incongruence was
detected, the conflicting branches were evaluated for relative support
given parsimony bootstrap and Bayesian posterior probabilities and were
discussed individually.
Analysis of Character Evolution—Character states were optimized on
the phylogenetic hypothesis resulting from our Bayesian analysis of combined data so relative branch lengths, as determined by Bayesian analysis,
could be incorporated for ML analysis of ancestral states. Two unidentifiable taxa (M. sp. “red 1” and M. sp. “red 2”) were trimmed from the tree
given our lack of floral characters for these taxa. Ancestral state optimization was performed using likelihood methodologies implemented in the
program Mesquite 1.04 (Maddison and Maddison 2004). We chose to use a
one-rate model following the observations of Mooers and Schluter (1999).
The ML model used for the analysis of the morphological data was Mk1
(Lewis 2001). Key characters traditionally associated with Myriophyllum
taxonomy (stamen number, sepals present/absent, and degree of sexual
dimorphism) were optimized. Character states were compiled from several literature sources (Schindler 1905; Meijden 1969; Orchard 1975, 1980,
1981, 1986; Aiken 1981; Yu et al. 2002) and herbarium specimens. In some
cases sexual dimorphism was reported to be somewhat labile with occasional cases of dioecy in primarily monoecious taxa or the reverse as well
as some inconsistency in the reporting of presence/absence of hermaphrodite flowers. In these cases taxa were coded for their primary character
124
SYSTEMATIC BOTANY
[Volume 35
state based on the most current treatments and observations from specimens used in these analyses.
Results
The ITS data set (71 accessions) consisted of 644 bp of aligned
sequence. There were 337 variable characters with 260 parsimony informative (including outgroup taxa). The cpDNA
data set (60 accessions) trnK 5’ intron had 735 bp aligned
sequence of which 203 sites were variable and 103 parsimony
informative. The matK data set had 1,503 bp aligned of which
414 sites were variable and 194 parsimony informative. The
trnK 3’ data set had 146 bp aligned of which 61 sites were variable and 33 parsimony informative. Data sets are available on
TreeBASE (study number S2323).
Parsimony Results—Parsimony analysis of the ITS data
resulted in 9,949 equally parsimonious trees in three islands
of 1,196 steps (CI = 0.492, RC = 0.386, RI = 0.785). When
M. dicoccum and M. muricatum were removed ITS data resulted
in 15,674 equally parsimonious trees of 1,110 steps (CI = 0.512,
RC = 0.405, RI = 0.791). Parsimony analysis of the cpDNA data
resulted in 520 equally parsimonious trees of 1,095 steps (CI =
0.742, RC = 0.658, RI = 0.886). The combined analysis resulted
in 128 equally parsimonious trees of 2,265 steps (CI = 0.612;
RC = 0.498, RI = 0.813). For cpDNA and combined analyses
parsimony results were comparable to Bayesian analyses,
but with less resolution. Bootstrap is not directly comparable
to Bayesian PP, but in general bootstrap support provides
lower values than for Bayesian PP (Cummings et al. 2003;
Douady et al. 2003) as was the case here. For ITS a long branch
for the clade including M. dicoccum and M. muricatum was
resolved and the placement of the clade was incongruent with
results from Bayesian analyses. The clade was placed sister to
M. subg. Brachytheca rather than within the clade (BS < 50%;
not shown). This is in disagreement with Bayesian analysis of
ITS and all other analyses of all data sets which agreed on the
position of this clade (Figs. 1–3). We attributed this incongruent result to long-branch attraction (Felsenstein 1978). Given
that most other relationships among taxa were not incongruent between parsimony and Bayesian analyses these taxa
were removed from the data set to perform bootstrap analyses of the ITS data. In general, ITS data provided lower resolution under parsimony than Bayesian analyses (Figs. 1–3).
Bayesian Results—Posterior probability distributions of
29,000 sampled trees were obtained for each Bayesian analysis
using a best-fit ML models with defined parameters (Table 2).
Visual comparison of the majority consensus trees from the two
separate runs for each data set disclosed no major discrepancies between tree topologies or PP nodal support. Comparisons
of split frequencies using AWTY (Nylander et al. 2008) also
supported consensus among multiple runs. Final trees represented the majority rule consensus of 29,000 trees, conservatively discarding the first 1,000 (one million generations)
as burn in (Figs. 1–3). ML parameters and likelihood scores
are presented in Table 2. Bayesian estimated branch lengths
are represented on phylograms based on consensus results
from analyses of each data set (Fig. 4). Bayesian results for ITS
were incongruent with parsimony analyses in the placement
of M. dicoccum and M. muricatum as discussed above.
Phylogenetic Relationships—All Bayesian analyses resolved
two major clades within Myriophyllum, M. subg. Brachytheca,
and M. subg. Myriophyllum (Figs. 1–3). Myriophyllum subg.
Meziella was resolved as sister to M. subg. Brachytheca in the
Fig. 2. Phylogenetic relationships in Myriophyllum as indicated by
a majority rule consensus of 29,000 trees (after discarding burn-in) from
Bayesian analysis of cpDNA (trnK 5’ intron, trnK 3’ intron and matK)
sequence data analyzed under models defined in Table 2. Harmonic mean
–ln likelihood = 10,634.40. Dashed lines refer to clades not resolved in the
strict consensus of the parsimony analysis. Numbers above branches refer
to posterior probabilities and numbers below branches are bootstrap support from parsimony analyses. When numbers are lacking below branches
parsimony did not resolve clades or had bootstrap values < 50% supporting that node. Numbers after species refer to divergent genotypes among
accessions fitting species descriptions. EU = Europe, NA = North America,
SA = South America, T = Tasmania.
combined and cpDNA analyses (Figs. 2, 3), whereas M. subg.
Meziella was part of a polytomy with M. subg. Brachytheca and
M. subg. Myriophyllum in ITS analyses (Fig. 1).
MYRIOPHYLLUM subg. MYRIOPHYLLUM—This subgenus had
two well-supported sister clades using cpDNA and combined data: 1) M. sect. Pectinatum had two South American
species (M. aquaticum and M. mattogrossense) and two undescribed taxa of unknown geographic origin and 2) M. sect.
Myriophyllum contained several geographically diverse
species. ITS data provided only weak support for M. sect.
Myriophyllum and resolved M. decussatum Orchard as sister to
the rest of M. sect. Myriophyllum (Fig. 1), whereas cpDNA and
combined analyses provide an alternative hypothesis regarding this taxon (Figs. 2, 3).
All analyses supported M. verticillatum L. and M. oguraense
Miki as a clade (Figs. 1–3) within M. sect. Myriophyllum. ITS
supported M. robustum Hook. f. as sister to this clade only
2010]
MOODY AND LES: MYRIOPHYLLUM SYSTEMATICS
125
Fig. 3. Phylogenetic relationships in Myriophyllum as indicated by a majority rule consensus tree of 29,000 trees (after discarding burn-in) from
Bayesian analysis of combined ITS and cpDNA sequence data analyzed under models as defined in Table 2. Harmonic mean –ln Likelihood = 17,532.97.
Dashed lines refer to clades not resolved in the strict consensus of the parsimony analysis. Numbers above branches refer to posterior probabilities and
numbers below branches are bootstrap support from parsimony analyses. When numbers are lacking below branches parsimony did not resolve clades or
had bootstrap values < 50% supporting that node. Numbers after species refer to unique genotypes among accessions fitting species descriptions. Cladebased classification system is represented along the right margin. EU = Europe, NA = North America, SA = South America, T = Tasmania.
126
SYSTEMATIC BOTANY
[Volume 35
Table 2. Average likelihood model parameters estimated for Bayesian analyses of each of the 6 data partitions in the combined data
(columns 2–6) and the ITS data with additional taxa (column 7) after the first 1,000 trees were discarded as burn in. Harmonic mean –ln likelihoods are:
combined–17532.97; cpDNA–10634.40; ITS–6956.49. The matK data were partitioned into first, second and third codon positions.
model
C>T
C>G
A>T
A>G
A>C
A
C
G
T
α
inv.
trnK5
trnK3
matK pos1
matK pos2
matK pos3
ITS
ITS
GTR + Γ
1.260
0.625
0.299
0.815
0.683
0.347
0.145
0.195
0.313
0.467
——-
GTR + Γ
1.713
1.317
0.155
1.605
1.464
0.311
0.156
0.184
0.349
1.201
——–
GTR + Γ
1.496
2.156
0.141
1.381
1.619
0.327
0.185
0.143
0.345
0.350
——-
GTR + I + Γ
1.203
1.202
0.286
1.445
1.538
0.301
0.152
0.140
0.407
1.264
0.142
GTR + Γ
1.169
0.849
0.326
1.081
1.104
0.295
0.216
0.173
0.317
0.455
——
GTR + I + Γ
4.174
0.598
2.077
2.163
0.996
0.207
0.305
0.285
0.204
0.848
0.597
GTR + I + Γ
4.271
0.410
1.953
1.614
0.723
0.212
0.325
0.269
0.194
1.097
0.291
under Bayesian analyses with weak support, whereas cpDNA
and combined data included M. robustum within M. subsect.
Myriophyllum. Relationships among members of M. subsect.
Myriophyllum generally were not well-supported. There was
strong support in all analyses for a clade including M. quitense
and M. triphyllum Orchard and a clade with three northern
hemisphere taxa (M. alterniflorum DC. sister to M. spicatum
and M. sibiricum).
MYRIOPHYLLUM subg. BRACHYTHECA—This subgenus was
resolved in all analyses with M. sect. Pelonastes well-supported
as sister to the rest of the subgenus (Figs. 1–3). Myriophyllum
sect. Tessaronia branched next and was well-supported in
all Bayesian analyses including the Austral-Asian species
(M. dicoccum), Australian species (M. latifolium F. Muell., M.
muricatum) and the North American endemics, M. subsect.
Spondylastrum. The North American endemics formed a well-
supported clade in all analyses (Figs. 1–3), but relationships
among most of these taxa were generally not well-supported.
Sister to M. sect. Tessaronia was M. sect. Pentapteris, which
was well-supported in all analyses (Figs. 1–3) with two subsections M. subsect. Spirophyllum and M. subsect. Nudiflorum.
Combined and cpDNA resolved a sister relationship of
M. papillosum to the rest of M. subsect. Nudiflorum with weak
support, whereas ITS included M. papillosum as part of a
polytomy with other members of M. subsect. Nudiflorum.
Resolution within the “variifolium + lophatum” clade was not
well-supported and in some cases there was well-supported
incongruence between ITS and cpDNA. Taxa with characteristics of M. simulans Orchard were polyphyletic and
M. variifolium paraphyletic. Within M. sect. Pentapteris there is
consensus for three major clades using cpDNA and combined
data: 1) the “variifolium” clade; 2) the “lophatum” clade; and 3)
Fig. 4. Phylogenetic relationships in Myriophyllum as indicated by phylograms of majority rule consensus tree of 29,000 trees (after discarding burnin) from Bayesian analyses of: a) ITS data; b) cpDNA data; c) Combined ITS and cpDNA data. Branch lengths are proportional.
2010]
MOODY AND LES: MYRIOPHYLLUM SYSTEMATICS
M. subsect. Spirophyllum (Fig. 3). ITS data alone did not support all of these clades (Fig. 1).
Emphasis was placed on widespread sampling of accessions
from the Australian taxa in Orchard’s (1986) “M. propinquum
Alliance” (including M. alpinum, M. crispatum, M. simulans,
and M. variifolium) due to the high morphological variability
among most of these taxa. Divergent ITS and/or cpDNA genotypes were found among all these species, except M. alpinum.
Some taxa not strictly conforming to the morphological
description of previously recognized species were recognized
to have unique genotypes. In cases where there is strong support for independent lineages (thus supporting a hypothesis
of reproductive isolation) coupled with unique morphology
we make recommendations for new species.
Discussion
Data and Phylogenetic Analyses—Incongruence length
difference (ILD) tests found significant incongruence between
the ITS and cpDNA data. Both empirical and simulated data
(Dolphin et al. 2000; Yoder et al. 2001; Barker and Lutzoni
2002) have demonstrated that this test has suggested noncombinability of data when data performed better if partitions
were combined. The ILD test does appear to be conservative
with low susceptibility to type II error, thus a simple way to
initially examine data partitions if congruence is not rejected.
When congruence is rejected there can be a number of reasons
involved with rates of molecular evolution between data sets
(Dolphin et al. 2000; Darlu and Lecointre 2002; Barker and
Lutzoni 2002). Bayesian analyses with case appropriate ML
models for individual partitions of data can help lessen some
problems associated with combining molecular data sets
that evolve under different evolutionary models (Nylander
et al. 2004). This approach was used in these combined data
analyses.
Incongruence may also exist between the nuclear and plastid genomes as a consequence of hybridization or lineage
sorting. Recent hybridization among Myriophyllum species
has been documented genetically (Moody and Les 2002) and
has been suggested to occur among members of Orchard’s
(1986) “M. propinquum Alliance”. Ploidy level is not known for
most Myriophyllum, but some chromosome counts have been
completed (M. alterniflorum [2n = 14], M. quitense [2n = 42],
M. sibiricum [2n = 42], M. spicatum [2n = 42], M. ussuriense
[2n = 14, 21], and M. verticillatum [2n = 28]; Löve and Löve
1958; Löve and Ritchie 1966; Löve 1978; Ceska et al. 1986) with
an apparent base number of n = 7 and polyploids not uncommon. However, we do not find incongruence between data
sets regarding the polyploid taxa (2n = 28, 42) that would suggest an allopolyploid history nor did we find many instances
in which multiple divergent copies of ITS were identified, but
where they were found (e.g. M. sibiricum, M. papillosum) they
formed a monophyletic lineage. This does not preclude the
potential of allopolyploid origins for these taxa, as ITS can
homogenize over time in allopolyploids to one parental genotype (Wendel et al. 1995), but evidence of an allopolyploid
history does not manifest itself through incongruence regarding these known polyploid taxa.
Where incongruence was found in phylogenetic hypotheses between plastid and nrDNA data sets regarding specific
taxa (i.e. M. robustum, M. decussatum, M. petraeum) nodal support for the relationship in one or both data sets was generally
weak, thus making the result of incongruence ambiguous.
127
Some cases of incongruence had strong support for alternative hypotheses provided by each data set and these cases
usually involved relationships at tips of the phylogeny (i.e.
M. tillaeoides, M. hippuroides) among closely related species.
These cases may represent hybrid origins, but the data do not
discount the alternative hypotheses of lineage sorting, thus
additional data to test a hybrid hypothesis will be needed.
Where there was incongruence between nuclear and plastid
DNA phylogenies, each case will be discussed individually.
Long-branch attraction (Felsenstein 1978) involving M. dicoccum and M. muricatum (see Fig. 4) also was likely to be associated with at least some of the incongruence observed between
ITS and cpDNA data sets detected using ILD, a parsimony
based approach.
Myriophyllum Systematics—Neither Meijden and Caspers
(1971) nor Orchard (1986) adopted a formal subgeneric classification system for Myriophyllum and both authors regarded
Schindler’s (1905) subgeneric system to be largely artificial
(Meijden and Caspers 1971; Orchard 1986). Our findings also
often contradict Schindler’s classification. Although Orchard
(1986) recognized “Alliances,” his system was informal and
not based explicitly on phylogenetic hypotheses. Despite
the historical difficulties with subgeneric classification of
Myriophyllum, we believe that our phylogenetic results, which
strongly support several nested clades, allow us to formally
recognize three subgenera, five sections, and five subsections
(see Table 3 and taxonomic discussion below). Thus we propose a formal subgeneric classification, which conserves most
of Schindler’s (1905) nomenclature (with some major realignments) and we recognize types, absent from Schindler’s
(1905) treatment. The subgeneric names (Table 3) are used in
our ensuing discussion of phylogenetic relationships. In some
cases, where relationships among clades remained unresolved
(or required further sampling), we have used clade-based terminology instead of applying formal taxonomic categories.
Taxonomic Status of Meziella—The status of the monotypic Western Australian Declared Rare Flora (DRF) Meziella
was ambiguous until recently. Prior to its recent rediscovery
by Orchard and Keighery (1993) this taxon was known from
a single specimen collected in 1901 in southwest Western
Australia and thought to be extinct. Schindler (1905) erected
the genus, but prior to that the taxon had been included in
Gonocarpus (1844) and Haloragis (1846). Given their access to
limited plant material, both Schindler (1905) and Orchard
(1975) tentatively placed Meziella in a position near Haloragis.
With fresh flowering and fruiting material Orchard and
Keighery (1993) assessed a close relationship of Meziella with
Myriophyllum. Moody and Les (2007a) found Myriophyllum
to be paraphyletic in regard to Meziella and recommended a
more thorough sampling of Myriophyllum to further evaluate
the status of the former genus.
Meziella and Myriophyllum share a four-loculed ovary
forming four nutlets (pyrenes), which is distinctive among
Haloragaceae. In Myriophyllum the nutlets usually dehisce
at fruit maturity, while in Meziella they are indehiscent due
to a persistent hardened exocarp (Orchard and Keighery
1993); although the hardened exocarp is similar to that of
some Myriophyllum species (e.g. M. decussatum, M. muricatum) described as having “tardy” dehiscence (Orchard 1986,
personal observation). Meziella fruits also are distinctively
ornamented with the persistent sepals becoming long woody
spines and elongate spines forming on the exocarp; however,
extreme fruit ornamentation in the form of spines also is found
128
SYSTEMATIC BOTANY
Table 3. Subgeneric classification for Myriophyllum using phylogenetic hypotheses from combined nrDNA ITS and cpDNA trnK + matK
data (unless otherwise denoted) and species continental distribution.
Unknown taxa are those not sampled for molecular data and whose placement remains unclear given morphology alone. All known Myriophyllum
species are included. [* not included in molecular analyses but likely placement based on morphological evidence and/or affinities suggested by
Orchard (1986)]. Afr = Africa, Aus = Australia, Eu = Europe, NA = North
America, SA = South America. This comprehensive list is based on multiple sources (Australian Plant Names Index [APNI], http://www.anbg.
gov.au/cgi-bin/apni; Meijden 1969; Meijden and Caspers 1971; Orchard
1975, 1980, 1981, 1986; Aiken 1981; Yu et al. 2002).
Subgeneric classification
1) Myriophyllum subgenus
Myriophyllum
A. M. section Myriophyllum
1. M. subsection Myriophyllum
M. alterniflorum DC.
M. balladoniense Orchard
M. caput-medusae Orchard
M. decussatum Orchard
M. porcatum Orchard*
M. quitense Kunth
M. robustum Hook. f.
M. salsugineum Orchard
M. sibiricum Komarov
M. spicatum L.
M. triphyllum Orchard
M. verrucosum Lindl.
2. M. subsection Isophylleae
M. oguraense Miki
M. verticillatum L.
B. M. section Pectinatum
M. aquaticum (Vell.) Verdc.
M. mattogrossense Hoehne
2) M. subgenus Meziella
M. trifidum (Nees) M.L. Moody
& D.H. Les
3) M. subgenus Brachytheca
A. M. section Pelonastes
M. amphibium Labill.
M. pedunculatum Hook. f.
M. tillaeoides Diels
B. M. section Tessaronia
M. dicoccum F. Muell.
M. exasperatum D. Wang,
D. Yu & Z. Yu Li*
M. latifolium F. Muell.
M. muricatum Orchard
M. tuberculatum Roxb.*
1. M. subsection Spondylastrum
M. farwellii Morong
M. heterophyllum Michx.
M. hippuroides Nutt. ex Torr.
& A. Gray
M. humile (Raf.) Morong
M. laxum Schuttl. ex Chapm.
M. pinnatum Britton,
Sterns & Poggenb.
M. tenellum Bigelow
C. M. section Pentapteris
1. M. subsection Spirophyllum
M. bonii Tardieu*
M. coronatum Meijden
M. costatum Orchard*
M. filiforme Benth.
Continental distribution
Northern Hemisphere: Asia,
EU, NA
Aus
Aus
Aus
Aus
NA, SA
Aus
Aus
Northern Hemisphere: Asia,
EU, NA
All continents (except Antarctica);
endemic Asia, EU
Aus
Aus
Asia
Northern Hemisphere:
Asia, EU, NA
All continents (except Antarctica);
endemic SA
SA
Aus
Aus
Aus
Aus
Asia, Aus
Asia
Aus
Aus
Asia
NA
Asia, Eu, NA; endemic NA
NA
NA
NA
NA
NA
Asia
Aus
Aus
Aus
(Continued)
[Volume 35
Table 3.
Continued.
Subgeneric classification
M. implicatum Orchard*
M. mezianum Schindl.*
M. siamense (Craib) Tardieu*
M. striatum Orchard*
M. trachycarpum F. Muell.
2. M. subsection Nudiflorum
M. alpinum Orchard
M. austropygmaeum Orchard*
M. crispatum Orchard
M. drummondii Benth.
M. echinatum Orchard
M. gracile Benth.*
M. integrifolium Hook. f.*
M. lapidicola Orchard*
M. limnophilum Orchard
M. lophatum Orchard
M. papillosum Orchard
M. petraeum Orchard
M. propinquum A. Cunn.*
M. pygmaeum Mattf.*
M. simulans Orchard
M. ussuriense Maxim.
M. variifolium Hook. f.
M. votschii Schindl.
Unknown (not sampled)
M. artesium Halford & Hensham
M. axilliflorum Baker
M. callitrichoides Orchard
M. glomeratum Schindl.
M. indicum Willd.
M. muelleri Sonder
M. oliganthum (Wight & Arn.) F. Muell.
M. tetrandrum Roxb.
Continental distribution
Aus
Afr
Asia
Aus
Aus
Aus
Aus
Aus
Aus
Aus
Aus
Aus
Aus
Aus
Aus
Aus
Aus
Aus
Aus
Aus
Asia, NA
Aus
Aus
Aus
Afr.
Aus
Aus
Asia
Aus
Asia
Asia
among Myriophyllum species (e.g. M. coronatum Meijden,
M. muricatum). The characteristic trifid submerged leaves of
Meziella also occur in members of M. sect. Pelonastes, which
branches subsequent to Meziella in our molecular phylogeny
(Fig. 3). Both Myriophyllum pedunculatum and M. tillaeoides
primarily have linear leaves but also have some submerged
leaves that become trifid, a trait uncommon in Myriophyllum,
but shared with Meziella. Although other character states
found in Meziella are not common among Haloragaceae (e.g.
apiculate stamens, four stamens) or within Myriophyllum (e.g.
all hermaphrodite flowers), they all occur (but only simultaneously in M. mattogrossense) within some Myriophyllum species (Orchard and Keighery 1993).
Our data strongly support not only a close relationship of Meziella and Myriophyllum but further indicate
that Myriophyllum is paraphyletic with regard to Meziella.
Despite its lack of dehiscent nutlets, Meziella is resolved with
strong support by cpDNA and combined analyses as sister to
M. subg. Brachytheca (PP = 1.00; BS = 93, 89 respectively). ITS
data alone remain ambiguous as to the placement of Meziella
within Myriophyllum. Using ITS alone, Moody and Les (2007a)
found a weakly supported sister group relationship of Meziella
with Myriophyllum using Bayesian analysis, a relationship not
supported by parsimony analysis, which placed Meziella as
part of a polytomy with Myriophyllum species. That study
included a larger family wide sampling, but a less inclusive
sampling of Myriophyllum. When the expanded Myriophyllum
sampling is added to the Moody and Les (2007a) ITS data set
the results remain ambiguous (not shown). Given the strong
2010]
MOODY AND LES: MYRIOPHYLLUM SYSTEMATICS
phylogenetic evidence (including shared morphology) for the
transfer of Meziella to Myriophyllum, we accommodate this
taxon by the new combination Myriophyllum trifidum (Nees)
M. L. Moody and D. H. Les (=Meziella trifida (Nees) Schindler)
and recognition of a monotypic subgenus: Myriophyllum subg.
Meziella (Schindl.) M. L. Moody and D. H. Les (see taxonomic
revision below).
Myriophyllum (Higher Level Relationships)—A fundamental split of Myriophyllum into distinct clades is evident
here (Fig. 3) and we recognize three subgenera to accommodate them (Table 3). Schindler’s (1905) treatment also divided
Myriophyllum into three subgenera (Table 1), but the taxonomic alignments within his subgenera are not entirely supported by our analyses. Myriophyllum subg. Myriophyllum
(Fig. 3; Table 3) corresponds well to Schindler’s (1905)
M. subsect. Spondylophyllum + M. subsect. Leiocarpium (Table 1),
but his inclusion of M. propinquum A. Cunn. (then much more
broadly defined to include Australian taxa [in part M. crispatum, M. simulans, and M. variifolium]) and M. ussuriense Maxim.
as allied with the other taxa in M. subsect. Spondylophyllum,
is strongly discordant with our phylogenetic results. Our
realigned M. subg. Brachytheca (Fig. 3; Table 3) includes species distributed throughout the major sections of Schindler’s
(1905) Myriophyllum classification. Schindler’s subgeneric
classification heavily emphasized the presence or absence of
hermaphrodite flowers to distinguish his two most speciesrich sections, and relied primarily on floral morphology to
further delimit segregates. However, our phylogenetic results
coupled with ancestral character state optimization indicate
that several of the reproductive characteristics shared among
disparate Myriophyllum are a consequence of parallelism, convergence, or plesiomorphy rather than synapomorphy (Fig. 5).
This observation supports Meijden (1969) who concluded that
infrageneric classification of Myriophyllum has likely been
confounded by misleading patterns of character evolution.
Although Orchard (1986) also did not recognize the same
fundamental groupings of species within Myriophyllum
indicated by our phylogenetic analysis, his “M. aquaticum
Alliance” and “M. salsugineum Alliance” taken together
encompass all the taxa in M. subg. Myriophyllum. A notable exception is that he recognized M. heterophyllum and
M. hippuroides Nutt. ex Torr. & A. Gray as loosely allied to the
“M. aquaticum Alliance”, whereas our phylogenetic analyses
clearly support their inclusion in M. subg. Brachytheca, which
encompasses Orchard’s (1986) other seven “Alliances” (Table 1).
So far we have been unable to identify morphological attributes that are exclusive to either of these clades, a factor that
helps to explain their lack of recognition in earlier treatments
of the genus (Schindler 1905; Meijden 1969; Meijden and
Caspers 1971; Orchard 1986).
Myriophyllum Subgenus Myriophyllum—This subgenus
has a world-wide distribution. Most species are widespread
or relatively common in their range, with only few maintaining a narrow endemism (e.g. M. balladoniense Orchard,
M. decussatum, M. porcatum Orchard). This clade includes
all the naturally widespread northern hemisphere species
(M. alterniflorum, M. sibiricum, M. verticillatum) and all the
South American species (M. aquaticum, M. mattogrossense,
M. quitense). Also included are two wide-spread invasive species (M. aquaticum, M. spicatum) whose natural distributions
are limited, but which are widespread outside their native
range. Although no one morphological feature has been identified that distinguishes all members of this clade, most have
129
strictly whorled submerged leaves (opposite in M. decussatum), and the linear emergent leaves that are common among
the other subgenera are found only in M. balladoniense.
Myriophyllum Section Pectinatum—This clade is sister to
the rest of M. subg. Myriophyllum and is supported in all analyses. Morphologically, M. mattogrossense and M. aquaticum
share an emergent inflorescence with all leaves pectinate. This
characteristic is uncommon in Myriophyllum, although strictly
pectinate emergent leaves also are found in the next branching
M. subsect. Isophylleae, as well as in M. robustum and the distantly related M. alpinum. Myriophyllum aquaticum and M. mattogrossense also are the only species of Myriophyllum endemic
to South America (although there is some debate concerning
the status of M. quitense in North America; see below).
Myriophyllum mattogrossense has been poorly collected,
but is known to range widely across South America and
likely is more common than currently known (Orchard 1981;
Orchard and Kasselman 1992). Outside of some collections
from Ecuador (Orchard and Kasselman 1992), M. mattogrossense has been described as a submerged species with only its
inflorescence becoming emergent. In contrast, Myriophyllum
aquaticum is dioecious and frequently found in an emergent
form producing dense populations (the submersed leaves
often deteriorate when the plants grow in standing water).
This species, which has been sold commonly in the ornamental aquatic plant trade, is invasive on most continents.
Myriophyllum sect. Pectinatum also includes two taxa (M. sp.
“red 1” and M. sp. “red 2”) that cannot be identified using any
current taxonomic resource. These taxa were acquired from
the aquatic plant trade in the U. S. A. (Maine and Washington)
and Australia (Queensland) as “M. mattogrossense” or “M. propinquum”, often with the common name “feather plant”. The
taxa are most closely related to M. aquaticum but are highly
divergent in both ITS and cpDNA (Figs. 1–4). All nonindigenous Myriophyllum aquaticum specimens collected from geographically diverse locations in North America (Appendix 1)
have identical ITS genotypes (they are strictly female plants)
that are unique from these two taxa.
Although these M. aquaticum-like taxa clearly are distinct
at the molecular level, defining them morphologically has
remained elusive given that no sexual structures have yet
been observed for either. Unfortunately, attempts to induce
flowering so far have failed and the original locality data are
not available for either taxon. These plants can survive in
inundated soils with all leaves emergent, rigid, and pectinate.
Some emergent forms of Myriophyllum sp. “red 1” were examined in cultivation growing adjacent to M. aquaticum. These
cultivars closely resembled M. aquaticum but maintained phenotypes that were distinctly red and more compact in habit
than M. aquaticum, whereas submerged forms of these plants
were common to many Myriophyllum species and, unlike
M. aquaticum, they do not appear to shed leaves with inundation. The emergent forms of these plants resemble those
described for Ecuadorian Myriophyllum that have emergent
vegetative forms, but were included under M. mattogrossense
with a more inclusive definition of the species (Orchard and
Kasselman 1992). Further examination of these cultivated
plants in flower will be needed to compare to Ecuadorian taxa
and also to exclude the possibility that their identity might lay
among the more poorly known Asian species not included in
these analyses (e.g. M. tetrandrum Roxb. and M. exasperatum
D. Wang, D. Yu & Z. Yu Li), given that the Australian cultivar originated from the Asian plant trade. Although none
130
SYSTEMATIC BOTANY
[Volume 35
Fig. 5. Maximum likelihood ancestral character state reconstructions for Myriophyllum using the combined ITS and cpDNA phylogram from Fig. 4
(above) to account for branch lengths. Two taxa (M. sp. (red 1) and M. sp. (red 2)) were trimmed from trees due to lack of information regarding reproductive structures. The ancestral character state reconstructions are as follows: a) carpellate sepals: present = white, absent = black; b) staminate sepals:
present = white, absent = black; c) stamen number: 8 = white, 4 = black, £ 2 = grey; d) sexual dimorphism: monoecious = white, dioecious = grey, morphological hermaphrodite = black; e) sexual dimorphism: some hermaphrodite flowers present = white, hermaphrodite flowers usually absent = black.
Coloration of circles at nodes is proportional to the likelihood of representative character states being ancestral at the given node.
of the Asian species has been described to have the emergent vegetative form found in these taxa (Dong et al. 2002; Yu
et al. 2002).
Myriophyllum Section Myriophyllum—This clade is strongly
supported in combined and cpDNA analyses. All members
were part of the “M. salsugineum Alliance” of Orchard (1986)
except M. balladoniense for which the placement was considered uncertain and M. robustum, which was suspected to have
an affinity with M. aquaticum. Myriophyllum sect. Myriophyllum includes two sister clades, M. subsect. Myriophyllum and
M. subsect. Isophylleae. Although M. subsect. Isophylleae is supported by all analyses, Bayesian analyses of ITS resolve a sister group relationship with M. robustum (with a short branch
and weak support), which is not resolved by parsimony or
any analyses of cpDNA and combined data. Its affinity is
most likely with M. subsect. Myriophyllum (discussed below).
Kadono (1988) suspected a close relationship between the
two taxa of M. subsect. Isophylleae (M. oguraense and M. verticillatum), which share the characteristic of having pectinate
emergent leaves that are similar in shape to, but much smaller
2010]
MOODY AND LES: MYRIOPHYLLUM SYSTEMATICS
than, the submerged leaves. Also, both have an elongate
turion (vegetative overwintering bud). The primary characteristic that differentiates these taxa is the turion shape (club
shaped in M. verticillatum and linear in M. oguraense; Weber
and Nooden 1974; Kadono 1988). Until recently M. oguraense
was believed endemic to Japan (Miki 1934; Kadono 1988);
however, its recent discovery in China confirms a wider distribution in sympatry with its sister species M. verticillatum,
found at the limits of its range in China and Japan (Yu et al.
2002).
Myriophyllum verticillatum is found throughout the northern temperate zone. Several varieties of Myriophyllum verticillatum have been proposed in North America (Fassett 1940;
Fernald 1950) but these taxa later were regarded as artificial
(Aiken 1979, 1981) and no molecular variation was found
among a wide geographic sampling in the U. S. A. (Appendix
1). Vegetative forms of M. verticillatum have been mistaken
for M. heterophyllum, M. hippuroides, and M. spicatum in North
America, but its pectinate bracts and turion shape readily distinguish this species from all other North American taxa.
Myriophyllum subsect. Myriophyllum is here defined broadly
to include M. robustum and M. decussatum which conflicts
with results from Bayesian analyses of ITS (parsimony results
are ambiguous). In general, ITS provides poor support for
higher-level relationships among M. subg. Myriophyllum and
the conflict observed in Bayesian analyses is not strongly supported. In cpDNA and combined analyses M. decussatum is
sister to the rest of the members of this subsection. Orchard
(1986) considered M. decussatum aberrant within the genus in
many characteristics, but suggested it most likely was loosely
allied with M. salsugineum due to its emergent leaf shape. Our
results do not contradict that hypothesis. Morphologically,
the species of this subsection have distinctive ovate emergent leaves that are entire, lobed, or toothed with the notable exceptions of M. robustum (pectinate) and M. balladoniense
(linear). Many of the taxa in this group had been part of a
broadly defined M. elatinoides Gaudich. until Orchard’s (1980,
1986) revisions recognized several new species and clarified
the priority of M. quitense.
Although many relationships among species of M. subsect. Myriophyllum were not well-supported, two small clades
were resolved consistently (Figs. 1–3). A clade including
M. quitense (New World) and M. triphyllum (New Zealand)
was well-supported by all analyses. Myriophyllum quitense is
the only other formally recognized South American species
besides M. aquaticum and M. mattogrossense (Orchard 1981). It
ranges into Mexico (Retana 1983) with disjunct populations in
the Deschutes River, Oregon (Fernald 1919) and recently has
been recognized from New Brunswick and British Columbia,
Canada (Ceska et al. 1986; McAlpine et al. 2007). Both ITS
and cpDNA provide evidence of molecular differentiation
between the North American and South American populations of M. quitense (Fig. 4) but morphological variation is not
evident (Orchard 1981; personal observations). The distinct
molecular variation (Fig. 4) between South American and
North American taxa does lend support to the hypothesis
that North American M. quitense might represent relict populations (Ceska et al. 1986) rather than an introduction from
South America (Aiken 1981).
A clade of predominantly northern hemisphere temperate
taxa (M. alterniflorum, M. sibiricum, and M. spicatum) also is
well-supported (Figs. 1–3). These taxa have a northern hemisphere distribution, although M. spicatum has been introduced
131
to North America (Couch and Nelson 1985) and has become
notoriously invasive across the continent. The members of
this clade were hypothesized to be part of a polyploid complex including M. verticillatum (Aiken 1981). Although our
phylogenetic evidence excludes M. verticillatum from such a
scenario, the chromosome counts of M. alterniflorum (2n = 14),
M. sibiricum (2n = 42) and M. spicatum (2n = 42) support the
possibility that M. sibiricum and M. spicatum could have originated from a hexaploid descendent of M. alterniflorum, which
is sister to the latter two taxa.
The status of M. sibiricum (ex. M. exalbescens) has been of
some contention. Some viewed the taxon as distinct from
M. spicatum (Fasset 1940; Sculthorpe 1967; Aiken 1981),
whereas others have recognized it as merely a subspecies
(Patten 1954) or variety (Jepson 1925; Nichols 1975). Much of
the debate over taxonomic rank was clarified with the establishment that M. spicatum was introduced to North America
around 1942 (Couch and Nelson 1985), which occurred well
after Fernald’s (1919) description of M. exalbescens (= M. sibiricum). Work describing growth and turion formation in M.
sibiricum helped to identify morphological characters that differentiated M. sibiricum from M. spicatum (Aiken 1979, 1981;
Moody and Les 2007b) and the typification of M. sibiricum
(Aiken and Cronquist 1988) has clarified the nomenclature.
A geographically diverse sampling (Appendix 1; Moody
and Les 2002, 2007b) has indicated consistent molecular divergence between M. spicatum and M. sibiricum and
this observation is supported morphologically by their leaf
length/pinnae ratios (Moody and Les 2007b). European specimens of M. sibiricum also were examined for molecular and
morphological variation compared to North American taxa.
The DNA sampled from three M. sibiricum accessions from
Europe (Appendix 1) all shared three point mutations in ITS
that were unique from North American M. sibiricum; however,
the plants lacked clear morphological differentiation (e.g. in
leaf length/pinnae ratio) from North American M. sibiricum.
Segregation of North American and European M. sibiricum at
the species or infraspecific level might be appropriate if the
genetic variation remains consistent through a larger sampling across the species range.
All other species of M. subsect. Myriophyllum are Australian
endemics. Relationships among these taxa are not strongly
supported, despite their distinctness both morphologically
(Orchard 1980, 1981, 1986) and genetically (Fig. 4). These species have adapted to a wide range of aquatic habitats from
ephemeral water bodies (M. verrucosum Lindl., M. balladoniense) to slow moving streams (M. caput-medusae Orchard) and
permanent water bodies with high salinity (M. salsugineum).
Myriophyllum Subgenus Brachytheca—This clade is highly
diverse with many species retaining narrow endemism in comparison to M. subg. Myriophyllum. This subgenus is greatly
expanded from Schindler’s (1905) treatment including many
species from his M. subg. Pentapteris. The majority of species
in this clade are restricted to Australia, and several are considered to be of conservation concern (e.g. M. latifolium, M. petraeum Orchard, etc.). There is a clade of seven North American
endemics (Fig. 3) and some species that range into Asia,
Papua New Guinea, and New Zealand. There is a conspicuous absence of species from South America and Europe in
this clade. The diversity of vegetative forms found in M. subg.
Brachytheca is much greater than that found in M. subg.
Myriophyllum (i.e. minute linear-leaved plants to robust plants
displaying various degrees of heterophylly), a factor that
132
SYSTEMATIC BOTANY
correlates with the diversity of habitats where these plants
have evolved (i.e. deep-water lakes to shallow ephemeral
pools with a highly labile wet-season duration).
Myriophyllum Section Pelonastes—Myriophyllum pedunculatum and M. tillaeoides form a clade sister to the rest of M. subg.
Brachytheca (Fig. 3). These species along with M. amphibium
(only sampled for ITS) and M. austropygmaeum Orchard
(not sampled) make up the core species of Orchard’s (1986)
“M. pedunculatum Alliance.” However, M. austropygmaeum
lacks sepals on its carpellate flowers (discussed later) and
likely does not belong here. The plants of this clade are small,
opposite-leaved perennials that usually form mats in ephemeral water holes. They retain sepals on their carpellate flowers, a characteristic found elsewhere in M. subg. Brachytheca
only among M. sect. Tessaronia (Fig. 5a). Orchard (1986) suggested a loose alliance of these taxa with other vegetatively
similar Myriophyllum (M. lophatum, M. pygmaeum Mattf.,
and M. votschii), some of which are only distantly related in
M. subsect. Nudiflorum (Fig. 3).
Currently there are three described subspecies of M. pedunculatum. Two genotypes of M. pedunculatum were recognized
in these analyses, here referred to as M. pedunculatum (T) [Les
643, Tasmania] and M. pedunculatum subsp. pedunculatum all
collected from New South Wales (Appendix 1). The Tasmanian
accession is sister to M. tillaeoides using cpDNA but sister to
other M. pedunculatum subsp. pedunculatum using ITS. The
smooth fruits of M. pedunculatum (T) are similar to M. pedunculatum subsp. novae-zelandiae Orchard, whereas the deep red
stigmas and vegetative features are more common to other
M. pedunculatum subspecies (subsp. longibracteolatum, subsp.
pedunculatum). Besides its geographic isolation (endemic to
Western Australia), M. tillaeoides differs from M. pedunculatum (including M. pedunculatum [T]) in lacking pedunculate
male flowers and by having sessile trifid submerged leaves.
The pattern of molecular incongruence and morphological
variation could have taxonomic implications regarding species delimitation in this group and potential hybrid origins of
species. The diminutive nature of the members in this clade
has made their morphological assessment difficult and it is
evident from these data that further sampling of M. pedunculatum populations will be necessary to assess the taxonomic
status of Tasmanian M. pedunculatum as well as phylogenetic
patterns throughout this clade.
Myriophyllum Section Tessaronia—This is a well-supported
clade that includes a geographically diverse assemblage of
Australian taxa and a subclade of endemic North American
species. The Australian taxa include M. latifolium, M. muricatum, and M. dicoccum, the latter two found primarily in northern Australia. Myriophyllum dicoccum also ranges into India
and Vietnam (Meijden and Caspers 1971; Yu et al. 2002) while
M. muricatum was split only recently from M. tuberculatum
Roxb. (Orchard 1986), a species known primarily from India
and Malaysia. Schindler (1905) implied a close link between
M. tuberculatum and North American endemics by placing
them in the same section (Tessaronia) and Meijden (1969) noted
the strong similarities of M. dicoccum and the North American
endemic M. humile, most notably the unusual submerged
flowers also shared with M. farwellii. A better sampling of the
Asian taxa from Schindler’s (1905) M. subsect. Trachycarpaeum
(Table 1) will be necessary to further explore the link between
Asian and North American taxa.
Myriophyllum dicoccum and M. muricatum form a wellsupported clade in all analyses but are highly divergent at the
[Volume 35
molecular level with long branches based on both ITS and
cpDNA (Fig. 4). As discussed earlier, the long branch of this
clade may be responsible for its anomalous placement by parsimony analysis of ITS data. Orchard (1986) considered these
two taxa closely allied (Table 1) based on a host of vegetative
characteristics including a robust habit, irregularly arranged
(i.e. both alternate and whorled) submerged leaves, and elongate, mostly entire leaves. These features also are characteristic of M. latifolium, some North American endemics, and Asian
species of Schindler’s (1905) M. sect. Tessaronia. The relationship of Myriophyllum latifolium sister to M. dicoccum and
M. muricatum is tentative due to relatively weak branch support and ambiguity given ITS alone (Figs. 1–3). This rare, narrow endemic of the coastal region of central and northern New
South Wales is distinct among members of this clade in being
dioecious, having carpellate flowers in fascicles, and leaves
that are strictly whorled. Orchard suggested that M. latifolium
was loosely allied with M. papillosum based mostly on their
shared fascicled flowers and large, flattened emergent leaves;
however, he did not consider the retention of sepals on carpellate flowers in M. latifolium (Fig. 5), whose loss apparently
is synapomorphic for M. sect. Pentapteris, which includes
M. papillosum.
Myriophyllum subsect. Spondylastrum, including all the North
American endemic taxa, is a well-supported but heteromorphic clade where relations among most taxa remain poorly
resolved (Figs. 1–3). One clade including M. heterophyllum,
M. hippuroides, and M. pinnatum is supported in all analyses
although relationships among these species are incongruent
between ITS and cpDNA. These taxa have the characteristic Myriophyllum submerged morphology. Myriophyllum heterophyllum and M. hippuroides both can become robust plants
with wide diameter stems and their submerged and emergent
leaves mostly appear to be whorled (but see England and
Tolbert [1964] regarding leaf initiation). Generally, these taxa
are delimited by the degree of dissection of their emergent
leaves and by their geographical (east-west North America)
disjunction. Myriophyllum heterophyllum also is vegetatively
plastic, whereas M. hippuroides is not.
Myriophyllum heterophyllum has three ITS genotypes
defined by point mutations and these are paraphyletic with
respect to M. hippuroides. Our cpDNA data revealed no variability in M. heterophyllum, but a recent study indicated at
least two major chloroplast lineages based on trnL–trnF data
(R. Thum et al. pers. comm.) that appear to have geographic
structure (midwest vs. southeast U. S. A.). Our cpDNA results
are incongruent in relation to ITS with M. hippuroides sister to
M. pinnatum-M. heterophyllum. Myriophyllum pinnatum is distinct from each of these taxa in vegetative form with scattered
leaves (rather than whorled) as is common to other endemic
North American taxa. Myriophyllum heterophyllum is known
to hybridize with M. laxum (Moody and Les 2002) and the
incongruence between data sets could be due to hybrid origins of species in this clade, although lineage sorting also is
a possibility.
Myriophyllum heterophyllum, M. laxum, and M. pinnatum are
sympatric in the southeastern U. S. A. The latter two are similar vegetatively, with submerged and emergent leaves that
usually are scattered or irregularly whorled. When reproductive structures are lacking, M. laxum can be mistaken for
M. pinnatum and hybrid M. heterophyllum × M. laxum. Although
the more robust stems of M. pinnatum generally differ from
the more “lax” stems of M. laxum, this characteristic can be
2010]
MOODY AND LES: MYRIOPHYLLUM SYSTEMATICS
plastic and difficult to evaluate. Moody and Les (2002) initially
described the invasive hybrids M. heterophyllum × M. laxum as
involving M. heterophyllum and M. pinnatum. However, collections of flowering M. laxum specimens in Florida [Moody 173,
174; Appendix 1] provided molecular evidence that M. laxum
(not M. pinnatum) was in fact the taxon involved in the hybrid
formed with M. heterophyllum. Several specimens from Florida
[Moody 53, 57, 58, 67, 69, 77; Appendix 1], were believed to
be M. pinnatum based on vegetative morphology, herbarium specimen location data, and distribution; however, they
proved to have the M. laxum genotype as confirmed by DNA
extracted from reproductive material. Myriophyllum pinnatum has emergent, deeply-lobed bracts and fruits with prominent tuberculate ridges, whereas M. laxum has much smaller,
often spatulate or obovate, unlobed bracts and fruits lacking
prominent tuberculate ridges. The phylogenetic placement
of the three remaining North American endemics is ambiguous within M. subsect. Spondylastrum. These taxa (M. farwellii,
M. humile, and M. tenellum Bigelow) all lack the differentiated
emergent growth form of the rest of M. sect. Tessaronia and
two are unusual in having submerged flowers (M. farwellii,
M. humile).
Although relationships among many of the North American
endemics remain unclear given these data, a comparably high
level of molecular differentiation is evident among the species (Fig. 4). All species have a unique ITS and cpDNA profile.
Although genotypic variation is found within some species
across their range (e.g. M. laxum and M. heterophyllum; Fig. 4),
the genotypes are distinct from other species in the clade. This
extent of variation has proven to be particularly helpful for
plant identification. Given that many of the taxa in this clade
display high plasticity and similarity in vegetative form, a
DNA marker (ITS) has been used to differentiate the northeast and northwest invasive M. heterophyllum and its hybrid
M. heterophyllum × M. laxum from native taxa for early detection of invasive species for state management agencies (e.g.
Maine and New Hampshire Department of Natural Resources
and Washington Department of Ecology).
Myriophyllum Section Pentapteris—This group is well
supported by both ITS and cpDNA. It is much reduced from
Schindler’s (1905) treatment and includes all members of
Schindler’s M. subsect. Spirophyllum but only few species
from M. subsect. Pelonastes and M. subsect. Spondylophyllum.
The loss of sepals on carpellate flowers is a synapomorphy
for M. sect. Pentapteris (Fig. 5a) within M. subg. Brachytheca.
This reduction also occurs in parallel among a few members
of M. subg. Myriophyllum (Fig. 5a). The loss of sepals in staminate flowers is uncommon in the genus but is found among
several members of M. sect. Pentapteris, including M. subsect,
Spirophyllum and some members of the “lophatum” clade
(Fig. 5b).
Myriophyllum subsect. Spirophyllum is inclusive of at least
three species from northern Australia, all of which have
asepalous staminate flowers (Figs. 3, 5b). This subsection
also likely includes the other members of Orchard’s (1986)
“M. striatum Alliance” and “M. mezianum Alliance” given
their shared morphology. The relationship between M. filiforme and M. trachycarpum F. Muell. was expected (Schindler
1905; Orchard 1986). Two accessions of M. filiforme with distinct cpDNA and ITS genotypes (Fig. 4) have been identified.
The individual accessions are geographically disjunct, but
further sampling will be needed to identify if this variation is
consistent and whether distinct morphological features can be
133
identified to correspond with these relatively highly divergent genotypes.
Myriophyllum coronatum also is part of M. subsect. Spirophyllum. It is known only from a single Australian locality
(Bronto Lake at the Northern-most tip in the Cape York peninsula, Queensland), but also occurs in Papua New Guinea.
This species has been allied to several east Asian species not
sampled here (M. mezianum Schindl., M. bonii Tardieu, and
M. siamense (Craib) Tardieu), all share reduced staminate
flowers (two petals, lack sepals, one stamen), dicarpic carpellate flowers (except the tetracarpic M. bonii), and a similar,
emergent, mat-like habit, which lacks pinnate leaves (Meijden
1969; Orchard 1986), a combination of characteristics that is
unique in the genus. Given the well-supported placement of
M. coronatum in M. subsect. Spirophyllum, these morphologically and geographically similar species likely belong here
as well. Meijden (1969) and Schindler (1905) hypothesized
a close relationship of M. coronatum with M. dicoccum given
their shared dicarpic carpellate flowers, but M. dicoccum differs substantially from M. coronatum in the presence of hermaphrodite flowers, tetramerous staminate flowers and the
retention of sepals on all flowers.
Myriophyllum subsect. Nudiflorum is well-supported. It has
an Austral-Asian distribution and its center of diversity is in
Australia including at least 12 endemics. It encompasses a
diverse assemblage of vegetative forms and includes all the
species sampled here from Orchard’s (1986) “M. propinquum
Alliance” (except M. latifolium), “M. integrifolium Alliance”,
and “M. pedunculatum Alliance Associates.” There are two
major clades within M. subsect. Nudiflorum and a weakly
supported sister taxon relationship of M. papillosum resulting from cpDNA and combined analyses. ITS provided little
resolution at higher levels in this clade, although some small
clades were supported.
In comparison to other Myriophyllum subsect. Nudiflorum
taxa, M. papillosum has robust vegetative features with wide
diameter stems and leaves frequently in whorls > 6. It also has
unusual reproductive features for the genus with male flowers in fascicles and carpellate flowers often densely clustered
in the axils of emergent leaves. Multiple copies of ITS were
found in M. papillosum, which were identified using cloning
techniques and together formed a clade (Fig. 1). The robust
nature and odd floral arrangements in combination with
multiple copies of the nrDNA ITS could indicate a polyploid
origin of this taxon. Chromosome counts have not yet been
performed for most Myriophyllum taxa and will be necessary
to evaluate this hypothesis.
The “lophatum” clade is well-supported under cpDNA
and combined analyses, whereas ITS does not resolve the
clade, rather the taxa form a polytomy with members of the
“variifolium” clade. The “lophatum” clade includes species
ranging widely in habit. There are small, mat-forming, opposite, and linear-leaved species of shallow ephemeral pools
(M. lophatum, M. votschii) and robust plants with pinnate
leaves that sometimes grow in deep waters (M. simulans 441,
M. alpinum Orchard).
Myriophyllum lophatum and M. votschii were proposed as
allies to Orchard’s (1986) “M. pedunculatum Alliance,” here
recognized as part of M. sect. Pelonastes. Myriophyllum lophatum grows in sympatry with M. pedunculatum in southeast
Australia while M. votschii grows in sympatry with M. pedunculatum in New Zealand. While the vegetative form of these
three species is nearly identical, as are most reproductive
134
SYSTEMATIC BOTANY
features, the lack of sepals along with molecular data in
M. lophatum and M. votschii allies them with M. sect. Pentapteris.
The other group of diminutive taxa, Orchard’s (1986) “M. integrifolium Alliance,” are closely allied to M. lophatum and
M. votschii and share floral features, but have alternate rather
than opposite leaves. The three endemic Western Australian
species of the alliance (M. drummondii, M. echinatum, and
M. limnophilum) were sampled for these analyses and found to
be monophyletic. Myriophyllum limnophilum and M. drummondii are the only species sampled that could not be distinguished
from each other using ITS. Given limited morphological variation (degree of leaf dimorphism and fruit ornamentation) the
status of these species will need further evaluation.
Myriophyllum simulans 441 is well-supported as part of
the “lophatum clade” while M. simulans 622 is not (Fig. 3;
Appendix 1). Both taxa fit the broad description attributed
to this species, but clearly are distinct from each other in our
analyses. Orchard (1986, 1990) described M. simulans as highly
variable with affinities to M. gracile Benth. and M. variifolium.
Only the latter has been sampled here, as the former is much
less common; living specimens could not be collected and
herbarium specimens provided poor quality DNA. The shape
and size of the staminate flower sepals and carpellate flower
bracteoles as well as emergent leaf shape (linear vs. lanceolate) readily distinguish the M. simulans genotypes sampled
here; however, a much broader sampling of these taxa will
be necessary to determine the consistency of these features.
Orchard (1986) regarded M. simulans as being an extremely
variable species with some “genetically fixed” isolates, a
hypothesis supported by our observations.
Another “lophatum” clade member, Myriophyllum sp. nov.
542 [Les 542; Appendix 1] also would fit broadly into the definition of M. simulans, but is divergent in emergent leaf features,
with extremely long (> 28mm), many toothed, linear emergent leaves that are strictly whorled, a combination of features
not found among other closely allied Myriophyllum and will
be formally recognized (Moody, pers. obs.). Combined and
cpDNA data strongly support the affinity of M. simulans 441
and Myriophyllum sp. nov. 542 to M. alpinum, but also provide
evidence that all three taxa are divergent genetically (Fig. 4).
The “variifolium” clade is well-supported in both cpDNA
and combined analyses, but not using ITS alone in which
members of the group are part of a polytomy within the
“lophatum” clade. The Australian members of this clade were
sampled more intensively given their notoriously challenging
taxonomic history. Orchard (1986) described two new species
from this group (M. crispatum and M. simulans) suggesting that local “genetically fixed” lineages were still possible
within M. simulans and M. variifolium. Without reproductive features the plants in this group are often impossible to
distinguish with certainty. However, our analyses have distinguished unique genetic lineages, and in some cases, corresponding morphotypes.
The Australian M. crispatum and Asian M. ussuriense were
well-supported sister taxa in all analyses. The cpDNA and
combined analyses resolved this clade as the sister group to
the rest of the “variifolium” clade. Myriophyllum ussuriense
is the only known Asian member of the “variifolium” clade.
Lability in sexual expression appears to be common among
the taxa of this clade, with the normally monoecious M. crispatum, M. simulans, and M. variifolium also having some individuals with either strictly carpellate or staminate-flowered
stems (Orchard 1986; personal observation) and the normally
[Volume 35
dioecious M. ussuriense recently described as having some
monoecious individuals (Ueno and Kadono 2001). Meijden
and Caspers (1971) described M. ussuriense as being nearly
indistinguishable from M. propinquum, but their evaluation
occurred before Orchard’s revisions (1981, 1986) in which the
new Australian species were recognized (including M. crispatum) and M. propinquum no longer was considered to occur
naturally in Australia. Myriophyllum crispatum is unique in
this clade by its dense indumentum (crisped hairs) on the
stem and base of some of the leaves. This characteristic is
common on emergent stems of plants in southeast Australia,
but notably absent in the disjunct Western Australia and
Queensland taxa. Myriophyllum crispatum was resolved as
a single lineage when multiple accessions were examined,
although ITS distinguished the disjunct west and southeast
populations.
Inclusion of the poorly known M. petraeum, a declared priority species (Department of Environment and Conservation
[DEC], Washington), in the “variifolium” clade is supported
by cpDNA and combined data. The alternate, linear leaves
and lack of sepals on staminate flowers in M. petraeum also
characterize members of the “lophatum” clade, but its columnar fruits are similar to those found in M. variifolium and M.
propinquum. Orchard (1986) associated M. petraeum with his
“M. striatum Alliance” whose members sampled here are far
removed in M. subsect. Spirophyllum.
All other taxa in the “variifolium” clade, which includes
only the members of Orchard’s (1986) “M. propinquum
Alliance” in exclusion of M. alpinum, are similar morphologically and have strongly supported relationships in all
analyses. Orchard (1986) regarded “M. propinquum s. s.” as
occurring only in New Zealand, and transferred all Australian
taxa described formerly under the broader definition of M.
propinquum to M. crispatum, M. simulans, or M. variifolium.
Myriophyllum simulans and M. crispatum differ from M. propinquum and M. variifolium in their fruit structure; however,
the distinction between M. variifolium and M. propinquum is
based on their geographic disjunction and differences in morphology, specifically “greater size in all parts” for the former
(Orchard 1986, pg. 203). All accessions sampled here were collected in Australia and fell within the size range of M. variifolium. We were not able to attain specimens of M. propinquum
from New Zealand for these analyses, and have followed
the geographic and morphological distinctions of Orchard
(1980, 1986) until sampling of New Zealand material can be
conducted.
Myriophyllum propinquum and M. variifolium are unique
from other members of the “variifolium” clade in having distinctive green/yellow columnar-shaped fruits, whereas other
members of the clade have fruits with a distinctly rounded
base, red coloration and tuberculate ornamentation. We
detected two distinct genotypes (M. variifolium 444 and 457)
having the M. propinquum/M. variifolium fruit type (Fig. 3;
Appendix 1) in both ITS and cpDNA corresponding to multiple accessions (Appendix 1). They do not form a clade under
any analyses, but are paraphyletic regarding M. simulans 622
and M. sp. 425 (discussed below). Neither genotype was consistent morphologically with the species description of M. propinquum and no morphological synapomorphies have been
identified to distinguish the unique genotypes. Additional
sampling across the geographic distribution of M. variifolium
will be necessary to evaluate the need for further taxonomic
revisions.
2010]
MOODY AND LES: MYRIOPHYLLUM SYSTEMATICS
Two other taxa included here are allied closely to M. variifolium (M. simulans 622 and M. sp. 425). Myriophyllum simulans 622 was discussed in relation to M. simulans 441 above.
Myriophyllum sp. 425 has characteristics of both M. gracile
and M. simulans but does not strictly follow the description
of either. Its position within the clade is incongruent with
respect to ITS and cpDNA data analyses, aligning as sister to
either M. simulans 622 (cpDNA) or M. variifolium 457 (ITS), but
distinct genotypically from each. Thus, it likely is not a recent
hybrid between the two, although a deeper hybrid history
is possible. Myriophyllum sp. 425 was collected well outside
the documented range of M. gracile but given the morphological similarities between these taxa it will be necessary to
sample the latter to establish if the taxa are conspecific. Two
individuals with genotypes matching Myriophyllum sp. nov.
476 [Moody 476, Les 653] also have been identified (Appendix
1). This taxon is sister to the four species of the “variifolium”
clade discussed above (Fig. 3) and is morphologically distinct
from all. It will be described in a forthcoming publication.
Our molecular analyses clearly show that substantial taxonomic revision is necessary among the “variifolium” clade
species. Although several genotypes are consistent across
multiple accessions that can be defined under current species
definitions (e.g. M. crispatum, M. variifolium 444), some taxa
with unique genotypes were detected, which in some cases
also possessed distinctive morphological characteristics that
were not previously documented. Thus, we suspect these taxa
to represent multiple cryptic species. Further DNA sampling
of accessions of previously described species (e.g. M. gracile, M. propinquum) along with wider geographic sampling
of highly variable species (M. simulans, M. variifolium) will
be necessary to clarify the extent of taxonomically significant
variation among the “variifolium” clade.
Character Evolution—Meijden’s (1969) suggestion that
infrageneric relationships in Myriophyllum cannot be elucidated by morphology alone has some merit. Molecular data
have proven informative in this regard by helping to identify
plesiomorphy and/or parallelism as well as homology among
Myriophyllum character states. Our phylogenetic results indicate that some character states, which are homoplasious
when all Myriophyllum taxa are considered, can also be useful
to delimit some clades. For example, the loss of sepals defines
M. sect. Pentapteris within M. subg. Brachytheca, despite
being of labile occurrence in M. subg. Myriophyllum (Fig. 5b).
Similarly, reduction in stamen number is diagnostic for
M. subsect. Spondylastrum, even though other reductions are
found sporadically throughout the genus (Fig. 5c). However,
this character needs to be evaluated further as some taxa
described as having four stamens have also been reported to
have eight (e.g. M. callitrichoides [Orchard 1986], M. farwellii
[Aiken 1981]).
Sexual dimorphism has been described as somewhat labile
within Myriophyllum species, and some species described as
monoecious also have sporadically occurring dioecious individuals (e.g. M. crispatum, M. variifolium, M. simulans, etc.)
or, as is the case with M. ussuriense, a dioecious species with
sporadic occurrence of monoecious individuals (Ueno and
Kadono 2001; Ceska et al. 1986; Orchard 1981, 1986). Many
monoecious species also have intermediate hermaphrodite
flowers and reporting of hermaphrodite flowers in monoecious taxa has been inconsistent. For example Schindler (1905)
described M. trachycarpum and M. filiforme as having first
hermaphrodite flowers then becoming unisexual, whereas
135
Orchard (1986) described these species as lacking hermaphrodite flowers. Our ancestral character state optimization
supports the hypothesis that degree of sexual dimorphism
is particularly labile in the genus (Fig. 5d, e) and the use of
this character for subgeneric classification (e.g. Schindler
1905) has given rise to artificial groups. Reduction in floral
features is particularly common among aquatic plants as is
the tendency towards monoecy or dioecy (Sculthorpe 1967).
The high level of homoplasy associated with such features in
our phylogenetic analyses helps to explain why it has been
so difficult to define subgeneric relationships in Myriophyllum
strictly on the basis of morphology.
DNA “Barcoding” Myriophyllum—We have shown that ITS
sequence data can provide reliable markers (i.e. “DNA barcoding”; Chase et al. 2005) capable of distinguishing among all
Myriophyllum species sampled so far (except M. drummondii
vs. M. limnophilum of southwest Western Australia) as well as
invasive hybrid taxa in North America (Moody and Les 2002,
2007b). Particular caution must be exercised when hybrid
taxa are concerned, because ITS can become homogenized to
either parent of a hybrid with long term introgression (Wendel
et al. 1995). Given the lack of information regarding fertility,
the degree of introgression is not yet known among the invasive hybrid watermilfoil populations. Thus, relying strictly on
ITS data could fail to differentiate such hybrids from “pure”
parental genotypes.
The reliability of ITS to effectively discern among North
American taxa has been confirmed by sampling across a wide
geographic area for most species (Appendix 1; unpublished
data). In Myriophyllum, where identification on the basis of
morphology can be particularly difficult (especially when
reproductive characters are lacking), DNA-based identification can provide a useful resource for aquatic plant management programs (Moody et al. 2008). For example, this
approach is able to distinguish native from invasive taxa for
early detection (and removal before establishment) because
DNA methods allow identification of even small vegetative
fragments. Several North American agencies with aquatic
plant management concerns (i.e. Maine, Minnesota, New
Hampshire, and Wisconsin Departments of Natural Resources
and Washington State Department of Ecology) already have
incorporated these methods into their management programs
(Moody, unpublished data).
Taxonomic Revision
Myriophyllum trifidum (Nees) M. L. Moody & D. H. Les,
comb. nov., Gonocarpus trifidus Nees, Plantae Preissianae
1: 159. 1844. Haloragis trifida (Nees) Walpers, Repertorium
Botanices Systematicae 5: 672. 1846. Meziella trifida
(Nees) Schindl. Das Pflanzenreich 23: 61. 1905.—TYPE:
AUSTRALIA. “In turfosis humidis ad lacum haud procul
ab oppidulo Albany (Plantagenet) m. Octobri 1840 Herb.
Preiss No. 2401” (holotype: LE!; isotype MEL!).
Myriophyllum subgenus Meziella (Schindl.) M. L. Moody
& D. H. Les, comb. et stat. nov. Meziella Schindl. Das
Pflanzenreich 23: 61. 1905.—TYPE: M. trifidum (Nees)
M. L. Moody & D. H. Les.
Diagnosis: Leaves alternate. Emergent leaves linear, submerged leaves often trifid. Hermaphrodite flowers. Sepals
persistent on fruit becoming elongate spines. Mericarps not
separating freely at maturity.
136
SYSTEMATIC BOTANY
Myriophyllum subgenus Myriophyllum—TYPE: M. spicatum L. Species Plantarum 2: 992. 1753.
This subgenus includes in part members of Schindler’s (1905)
M. subg. Eumyriophyllum [nom. inval.]. This includes eight of 19
members of M. sect. Pentapteris Schindl. and all species from
M. subsect. Spondyllophyllum Schindl. and M. subsect. Leiocarpium Schindl. All the core species of Orchard’s “M. salsugineum Alliance” and “M. aquaticum Alliance” are in this group.
Diagnosis: This is a heteromorphic group in both vegetative
and reproductive morphology. In general species have submerged leaves pectinate and whorled (if present). Emergent
leaves are usually whorled (sometimes becoming alternate or
subwhorled only towards the apex), but sometimes opposite
(M. decussatum) or mostly alternate (M. alterniflorum, M. balladoniense). Emergent leaves pectinate, entire, or toothed, generally ovate, obovate (linear for M. balladoniense). Monoecious,
dioecious, or all flowers hermaphrodite.
Myriophyllum section Myriophyllum—TYPE: M. spicatum L.
Myriophyllum subsect. Spondylophyllum Schindl. Das Pflanzenreich 23: 86. 1905.
This section includes six of the ten members of Myriophyllum
subsect. Spondylophyllum and both species from M. subsect.
Leiocarpium.
Diagnosis: Same as for M. subgenus Myriophyllum.
Myriophyllum subsection Myriophyllum—TYPE: M. spicatum L.
Myriophyllum series Anisophylleae Schindl. Das Pflanzenreich
23: 89. 1905.
This subsection closely follows M. series Anisophylleae, but
does not include M. propinquum.
Diagnosis: Submerged leaves, if present, all whorled.
Emergent leaves usually entire or toothed at least towards
apex (pectinate in M. robustum). Monoecious (all hermaphrodite flowers in M. robustum).
Myriophyllum subsection Isophylleae (Schindl.) M. L.
Moody & D. H. Les, comb. et stat. nov. Myriophyllum
series Isophylleae Schindl. Das Pflanzenreich 23: 86. 1905.—
TYPE: M. verticillatum L. Species Plantarum 2: 992. 1753.
This subsection closely follows M. series Isophylleae Schindl.,
but does not include M. aquaticum (ex. M. brasiliense).
Diagnosis: Submerged leaves whorled and pectinate, emergent leaves pectinate becoming distinctively smaller towards
apex. Elongate turions. Monoecious with intermediate hermaphrodite flowers.
Myriophyllum section Pectinatum M. L. Moody & D. H.
Les, sect. nov.—TYPE: M. aquaticum (Vell.) Verdc. Kew
Bulletin 28: 36. 1973.
Folia omnia verticillata; difert a M. subsect. Myriophyllum
folia omnia emersa pectinata et difert a M. subsect. Isophylleae
folia emersa comparate non redacta de foliis submersis.
Diagnosis: Submerged and emergent leaves whorled and
pectinate. Emergent leaves pectinate and not highly reduced
in relation to submerged leaves. Plants dioecious or flowers
all hermaphrodite.
Myriophyllum subgenus Brachytheca Schindl. Das Pflanzenreich. 23: 102. 1905.—TYPE: M. variifolium Hook. f.
Hooker’s Icones Plantarum 3: 289. 1840.
This subgenus is more broadly defined than the original circumscription (three spp.) now including all species from M.
subsect. Spirophyllum Schindl. and M. sect. Tessaronia Schindl.
[Volume 35
Diagnosis: This is a heteromorphic group in both vegetative and reproductive morphology. Submerged leaves alternate, opposite, or whorled; pectinate, trifid, linear, or lacking.
Emergent leaves linear, terete, lanceolate, oblanceolate, ovate,
obovate, or pectinate (M. alpinum). Monoecious or dioecious
(all hermaphrodite flowers in M. muricatum).
Myriophyllum section Pentapteris DC. Prodromus 3: 68.
1828.—TYPE: M. variifolium Hook. f.
This section is much reduced from Schindler’s (1905)
treatment and includes M. subsect. Spirophyllum but only
few species from M. subsect. Pelonastes and M. subsect.
Spondylophyllum.
Diagnosis: Monoecious or dioecious, lacking hermaphrodite flowers. Lacking sepals in carpellate flowers and often
in staminate flowers. Usually eight stamens (also four, two,
or one).
Myriophyllum subsection Spirophyllum Schindl. Das
Pflanzenreich 23: 82. 1905.—TYPE: M. trachycarpum
F. Muell. Fragmenta Phytographiae Australiae 2: 87. 1861.
This subsection varies from Schindler’s (1905) description
by the inclusion of M. coronatum. The placement of M. gracile
is as yet unknown.
Diagnosis: Leaves alternate or opposite becoming distally
alternate. Leaves linear, lanceolate, oblanceolate, ovate, obovate, lower-most submerged leaves becoming pectinate in
some species. Monoecious. Carpellate and staminate flowers
lack sepals. Stamens eight or one.
Myriophyllum subsection Nudiflorum M. L. Moody &
D. H. Les, subsect. nov.—TYPE: M. variifolium Hook. f.
Hooker’s Icones Plantarum 3: 289. 1840.
Folia verticillata, alterna, vel sparsa. Folia submersa linearia ad pectinata. Flores unisexuales, plantae monoeciae vel
dioeciae. Flores feminei sepalis, petalis et staminibus nullis.
This subsection includes most members of Orchard’s M.
propinquum Alliance and M. integrifolium Alliance and elements from across Schindler’s classification.
Diagnosis: Vegetatively heteromorphic with leaves whorled,
opposite or alternate. Submerged leaves linear or pectinate.
Emergent leaves linear, ovate, obovate, lanceolate, oblanceolate, or pectinate. Plants monoecious or dioecious, lacking
hermaphrodite flowers. Carpellate flowers lack a perianth.
Staminate flowers with perianth present (sepals sometimes
lacking) with (2, 4) or 8 stamens.
Myriophyllum section Tessaronia Schindl. Das Pflanzenreich 23: 95. 1905.—TYPE: M. heterophyllum Michx.
Flora Boreali-Americana 2: 191. 1803.
This section closely follows that of Schindler (1905).
Diagnosis: Submerged leaves whorled (subwhorled) or
alternate, pectinate or scale-like (M. tenellum). Emergent leaves
(when present) ovate, obovate, lanceolate, oblanceolate; often
lobed, toothed, or serrate. Plants monoecious usually with
hermaphrodite flowers intermediate on the inflorescence.
Sepals present on carpellate and staminate flowers. Stamens
four (eight).
Myriophyllum subsection Spondylastrum Schindl. Das
Pflanzenreich 23: 98. 1905.—TYPE: M. heterophyllum Michx.
This subsection follows Schindler (1905) with the addition
of all members of M. subsect. Ptilophylllum Schindl.
Diagnosis: Plants monoecious with hermaphrodite flowers
intermediate on the inflorescence. Sepals present on carpellate
2010]
MOODY AND LES: MYRIOPHYLLUM SYSTEMATICS
and staminate flowers. Stamens four. Endemic to North
America.
Myriophyllum section Pelonastes (Hook. f.) M. L. Moody &
D. H. Les, comb. et stat. nov. Pelonastes Hook. f. Hooker’s
Journal of Botany and Kew Garden Miscellany 474.
1847.—TYPE: M. tillaeoides Diels, Botanische Jahrbucher
fur Systematik 35: 448. 1904. Myriophyllum subsect.
Pelonastes (Hook. f.) Schindl. Das Pflanzenreich 23: 83.
1905.
This section closely follows M. subsect. Pelonastes but
excludes M. votschii.
Diagnosis: Small plants. Leaves opposite, linear, lanceolate,
ovate with lower-most submerged leaves sometimes trifid.
Sepals present on staminate and carpellate flowers. Stamens
eight.
Acknowledgments. The authors are especially grateful to S. Jacobs
(NSW) for extensive field assistance in Australia; also to R. Bayer,
R. Cranfield, G. Crow, C. B. Hellquist, Y. Kadono, C. Martine, N. Ritter,
J. Rourke, A.-M. Schwarz, and G. Towler for assistance and/or collections
in the field. We thank P. Lewis, L. Lewis, C. Jones, the editor, and two anonymous reviewers for helpful discussion and/or comments on the manuscript; and the staff of the following herbaria for their generosity in loaning
specimens: C, CAL, CANB, CBG, HBG, MEL, NSW, PERTH, R, UBC, UPS,
USCH. This research was funded by NSF DDIG 0309123, NEBC Graduate
Student Research Award Program, Karling Graduate Student Research
award (BSA), William R. Anderson Student Research Grant (ASPT), and
the University of Connecticut Bamford Endowment Fund.
Literature Cited
Aiken, S. G. 1979. North American species of Myriophyllum (Haloragaceae).
Ph. D. thesis. Minneapolis: University of Minnesota.
Aiken, S. G. 1981. A conspectus of Myriophyllum (Haloragaceae) in North
America. Brittonia 33: 57–69.
Aiken, S. G. and A. Cronquist. 1988. Lectotypification of Myriophyllum
sibiricum Komarov (Haloragaceae). Taxon 37: 958–966.
Angiosperm Phylogeny Group II. 2003. An updated classification of the
angiosperms. Botanical Journal of the Linnean Society 141: 399–436.
Barker, F. K. and F. M. Lutzoni. 2002. The utility of the incongruence length
difference test. Systematic Biology 51: 625–637.
Ceska, O., A. Ceska, and P. D. Warrington. 1986. Myriophyllum quitense and
Myriophyllum ussuriense (Haloragaceae) in British Columbia, Canada.
Brittonia 38: 73–81.
Chase, M. W., N. Salamin, M. Wilkinson, J. M. Dunwell, R. P. Kesanakurthi,
N. Haidar, and V. Savolainen. 2005. Land plants and DNA barcodes:
short-term and long-term goals. Philosophical Transactions of the Royal
Society Britain 260: 1889–1895.
Cirujano, S. and L. Medina. 1997. Myriophyllum heterophyllum Michx.
(Haloragaceae), naturalized in Spain. Anales del Jardín Botánico de
Madrid 55: 164–165.
Couch, R. and E. Nelson. 1985. Myriophyllum spicatum in North America.
Pp. 8–18 in Proceedings of the first international symposium on watermilfoil (Myriophyllum spicatum) and related Haloragaceae species, July 23–24,
1985, Vancouver, British Columbia, ed. L. W. J. Anderson Vicksburg,
The Aquatic Plant Management Society, Inc.
Cronk, J. K. and M. S. Fennessy. 2001. Wetland plants: biology and ecology.
Boca Raton, Florida: Lewis Publishers.
Cummings, M. P., S. A. Handley, D. S. Myers, D. L. Reed, A. Rokas, and
K. Winka. 2003. Comparing bootstrap and posterior probability values in the four-taxon case. Systematic Biology 52: 477–487.
Darlu, P. and G. Lecointre. 2002. When does the incongruence length difference test fail? Molecular Biology and Evolution 19: 432–437.
Dolphin, K., R. Belshaw, D. L. C. Orme, and D. L. J. Quicke. 2000. Noise
and incongruence: Interpreting results of the incongruence length
difference test. Molecular Phylogenetics and Evolution 17: 401–406.
Dong, W., Y. Dan, and L. Zhen-Yu. 2002. Myriophyllum exasperatum
(Haloragaceae) a new species from China. Annales Botanici Fennici 39:
267–271.
Douady, C. J., F. Delsuc, Y. Boucher, W. F. Doolittle, and E. J. P. Douzery.
2003. Comparison of Bayesian and maximum likelihood bootstrap
measures of phylogenetic reliability. Molecular Biology and Evolution
20: 248–254.
137
Doyle, J. J. and J. L. Doyle. 1987. A rapid DNA isolation procedure for small
quantities of fresh leaf tissue. Phytochemical Bulletin of the Botanical
Society of America 19: 11–15.
England, W. H. and R. J. Tolbert. 1964. A seasonal study of the vegetative
shoot apex of Myriophyllum heterophyllum. American Journal of Botany
51: 349–353.
Farris, J. S. 1989. The retention index and the rescaled consistency index.
Cladistics 5: 417–419.
Farris, J. S., M. Källersjö, and A. G. Kluge. 1995. Testing significance of
incongruence. Cladistics 10: 315–319.
Fasset, N. C. 1940. A manual of aquatic plants. New York: McGraw-Hill,
Inc.
Felsenstein, J. 1978. Cases in which parsimony or compatibility methods
will be positively misleading. Systematic Zoology 27: 401–410.
Felsenstein, J. 1985. Confidence limits on phylogenies: an approach using
the bootstrap. Evolution 39: 783–791.
Felsenstein, J. 1988. Phylogenies from molecular sequences: inference and
reliability. Annual Review of Genetics 22: 521–565.
Fernald, M. L. 1919. Two new Myriophyllum and a species new to the
United States. Rhodora 21: 121–124.
Fernald, M. L. 1950. Gray’s manual of botany. Ed. 8. New York: American
Book Co.
Huelsenbeck, J. P. and F. Ronquist. 2001. MRBAYES: Bayesian inference of
phylogenetic trees. Bioinformatics 17: 754–755.
Jepson, W. L. 1925. Manual of flowering plants of California. Berkeley:
Associate Student Store, University of California.
Johnson, L. A. and D. E. Soltis. 1994. matK DNA sequences and phylogenetic reconstruction in Saxifragaceae s. str. Systematic Botany 19:
143–156.
Kadono, Y. 1988. Germination of the turion of Myriophyllum oguraense
Miki. Aquatic Botany 31: 355–360.
Kluge, A. G. and J. S. Farris. 1969. Quantitative phyletics and evolution of
Anurans. Systematic Zoology 18: 1–32.
Les, D. H. and L. J. Mehrhoff. 1999. Introduction of nonindigenous aquatic
vascular plants in southern New England: a historical perspective.
Biological Invasions 1: 281–300.
Lewis, P. O. 2001. A likelihood approach to estimating phylogeny from discrete morphological character data. Systematic Biology 50: 913–925.
Löve, A. 1978. IOPB Chromosome number reports LXI. Taxon 27: 519–535.
Löve, A. and D. Löve. 1958. The American element in the flora of the
British Isles. Botaniska Notiser 3: 373–388.
Löve, A. and J. G. Ritchie. 1966. Chromosome numbers from central and
northern Canada. Canadian Journal of Botany 44: 429–439.
Maddison, D. R. and W. P. Maddison. 2003. MacClade: Analysis of phylogeny and character evolution. Version 4.06. Sunderland: Sinauer
Associates.
Maddison , W. P. and D. R. Maddison . 2004 . Mesquite : a modular
system for evolutionary analyses. Version 1.04 . Website http://
mesquiteproject.org [Accessed October 2006].
McAlpine, D. F., G. Bishop, O. Ceska, M. L. Moody, and A. Ceska. 2007.
Andean watermilfoil, Myriophyllum quitense (Haloragaceae), in the
Saint John River estuary system, New Brunswick, Canada. Rhodora
109: 101–107.
Meijden, R. V.-D. 1969. An annotated key to the South-East Asiatic,
Malesian, Mascarene, and African species of Myriophyllum
(Haloragaceae). Blumea 17: 304–311.
Meijden, R. V.-D. and N. Caspers. 1971. Haloragaceae. Pp. 239–263 in
Flora Malesiana Vol. 7, ed. C. G. G. J van Steenis. Djakarta: NoordhoffKolff N. V.
Miki, S. 1934. On fresh water plants new to Japan. Botanical Magazine
(Tokyo) 48: 335–336.
Moody, M. L. and D. H. Les. 2002. Evidence of hybridity in invasive
watermilfoil (Myriophyllum) populations. Proceedings of the National
Academy of Sciences USA 99: 14867–14871.
Moody, M. L. and D. H. Les. 2007a. Phylogenetic systematics and character evolution in the angiosperm family Haloragaceae. American
Journal of Botany 94: 2005–2025.
Moody, M. L. and D. H. Les. 2007b. Distribution and composition of invasive hybrid watermilfoil (Myriophyllum spicatum × M. sibiricum) populations in North America. Biological Invasions 9: 559–570.
Moody, M. L., L. Hufford, D. E. Soltis, and P. S. Soltis. 2001. Phylogenetic
relationships of Loasaceae subfamily Gronovioideae inferred
from matK and ITS sequence data. American Journal of Botany 88:
326–336.
Moody, M. L., D. H. Les, and J. M. DiTomaso. 2008. The role of plant systematics in invasive aquatic plant management. Journal of Aquatic
Plant Management 46: 7–15.
138
SYSTEMATIC BOTANY
Mooers, A. O. and D. Schluter. 1999. Reconstructing ancestor states
with maximum likelihood: support for one- and two-rate models.
Systematic Biology 48: 623–633.
Morrison, D. A. 2006. Multiple sequence alignment for phylogenetic purposes. Australian Systematic Botany 19: 479–539.
Nichols, S. A. 1975. Identification and management of Eurasian watermilfoil in Wisconsin. Wisconsin Academy of Arts 63: 116–128.
Nylander, J. A. A. 2003. MrModeltest 1.1b. Department of Systematic
Zoology, EBC, Uppsala University.
Nylander, J. A. A., F. Ronquist, J. P. Huelsenbeck, and J. L. Nieves-Aldrey.
2004. Bayesian phylogenetic analysis of combined data. Systematic
Biology 53: 47–67.
Nylander, J. A. A., J. C. Wilgenbusch, D. L. Warren, and D. L. Swofford.
2008. AWTY (are we there yet?): a system for graphical exploration
of MCMC convergence in Bayesian phylogenetics. Bioinformatics 24:
581–583.
Orchard, A. E. 1975. Taxonomic revisions in the family Haloragaceae.
1. The genera Haloragis, Haloragodendron, Glischrocaryon, Meziella and
Gonocarpus. Bulletin of the Auckland Institute and Museum 10: 1–299.
Orchard, A. E. 1980. Haloragaceae in Australasia. 1. New Zealand: A revision of the genus and a synopsis of the family. Brunonia 2: 247–287.
Orchard, A. E. 1981. A revision of South American Myriophyllum
(Haloragaceae) and its repercussions on some Australian and North
American species. Brunonia 4: 27–65.
Orchard, A. E. 1986. Myriophyllum (Haloragaceae) in Australasia. II. The
Australian species. Brunonia 8: 173–291.
Orchard, A. E. 1990. Haloragaceae. Pp. 5–85 in Flora of Australia Vol. 18, ed.
A. S. George. Canberra: AGPS Press.
Orchard, A. E. and C. Kasselmann. 1992. Notes on Myriophyllum mattogrossense (Haloragaceae). Nordic Journal of Botany 12: 81–84.
Orchard, A. E. and G. J. Keighery. 1993. The status, ecology and relationships of Meziella (Haloragaceae). Nuytsia 9: 111–117.
Patten, B. C. 1954. The status of some American species of Myriophyllum as
revealed by discovery of intergrade material between M. exalbescens
Fern. and M. spicatum in New Jersey. Rhodora 56: 213–225.
Retana, A. N. 1983. Registros nuevos de plantas acuaticas mexicanas I:
Myriophyllum quitense H. B. K. (Haloragaceae). Boletín de la Sociedad
Botánica de México 45: 147–149.
Rogstad, S. H. 1992. Saturated NaCl-CTAB solution as a means of field
preservation of leaves for DNA analyses. Taxon 41: 701–708.
Schindler, A. K. 1905. Halorrhagaceae. Pp. 1–133 in Das Pflanzenreich,
IV. ed. H. G. A. Engler. Leipzig: Wilhelm Engelmann.
Sculthorpe, C. D. 1967. The biology of aquatic vascular plants. London:
Edward Arnold Ltd.
Simmons, M. P. 2004. Independence of alignment and tree search.
Molecular Phylogenetics and Evolution 31: 874–879.
Swofford, D. L. 2002. Phylogenetic analysis using parsimony (*and other
methods), Version 4.0b10. Sunderland: Sinauer Associates.
Ueno, S. and Y. Kadono. 2001. Monoecious plants of Myriophyllum ussuriense
(Regel) Maxim. in Japan. Journal of Plant Research 114: 375–376.
Weber, J. A. and L. D. Nooden. 1974. Turion formation and germination
in Myriophyllum verticillatum: Phenology and its interpretation. The
Michigan Botanist 13: 151–158.
Wendel, J. F., A. Schnabel, and T. Seelanan. 1995. Bidirectional interlocus
concerted evolution following allopolyploid speciation in cotton
(Gossypium). Proceedings of the National Academy of Sciences USA 92:
280–284.
White, T. J., S. L. Bruns, and J. Taylor. 1990. Amplification and direct
sequencing of fungal ribosomal RNA genes for phylogenetics. Pp. 315–
322 in PCR protocols: a guide to methods and applications. eds. M. Innis,
D. Gelfand, J. Sninsky, and T. J. White. San Diego: Academic Press.
Wimmer, W. 1997. Myriophyllum heterophyllum Michaux in Niedersachsen
und Bremen sowie seine Bestimmung im vegetativen Zustand. Flora
Rundbriefe 31: 23–31.
Yang, Z. and B. Rannala. 1997. Bayesian phylogenetic inference using
DNA sequences: a Markov Chain Monte Carlo method. Molecular
Biology and Evolution 14: 717–724.
Yoder, A. D., J. A. Irwin, and B. A. Payseur. 2001. Failure of the ILD to
determine data combinability for slow loris phylogeny. Systematic
Biology 50: 408–424.
Yu, D., W. Dong, L. Zhen-Yu, and M. M. Funston. 2002. Taxonomic revision of the genus Myriophyllum (Haloragaceae) in China. Rhodora 104:
396–421.
Appendix 1. Accession data for taxa included in these phylogenetic
analyses. When multiple collection numbers are listed multiple specimens
were sampled and duplicate sequences for ITS were recovered (if num-
[Volume 35
bers follow taxon name, multiple genotypes were recovered from different
populations). Multiple collection numbers after a name refer to multiple
locations within the general geographic location listed. GenBank numbers follow consecutively: ITS, trnK 5’ intron, matK, trnK 3’ intron. When
GenBank number is replaced with “—-“, then no sequence was generated.
When a single GenBank number is listed after a taxon it is for ITS. * refers
to DNA extracted directly from an herbarium specimen. HPBG = Harold
Porter Botanical Garden, RBGC = Royal Botanical Garden Canberra.
Gonocarpus montanus, Moody 448 (CONN), Australia, New South Wales;
Moody 449b (CONN), Australia, New South Wales, EF178770, EF178952,
EF179044, EF178860; Haloragis digyna, Moody 411 (CONN), Australia,
Western Australia, EF178747, EF178929, EF179021, EF178838; Laurembergia
repens, Rourke (Cult.) HPBG, South Africa; *Williams 113 (GHPG), South
Africa, EF178735, EF178917, EF179009, EF178827; Myriophyllum alpinum,
Moody 449, 453 (CONN), Australia, New South Wales, EF178720, EF178902,
EF178994, EF178812; M. alterniflorum, Moody 109a, 111 (CONN), U. S. A.,
Wisconsin, EF178704, EF178886, EF178978, EF178797; M. amphibium,
*Orchard 5295 (NSW), Australia, Tasmania, FJ870941; M. aquaticum, Moody
(cult.), U. S. A., University of Connecticut; Moody 51 (CONN), U. S. A.,
California; Moody 56 (CONN), U. S. A., Florida, EF178727, EF178909,
EF179001, EF178819; M. balladoniense, Moody 389 (CONN), Australia,
Western Australia, EF178708, EF178890, EF178982, EF178801; M. caputmedusae, Moody 443, 446 (CONN), Australia, New South Wales, EF178703,
EF178885, EF178977, EF178796; M. coronatum, *Paijmans 3039 (CANB),
Australia, Queensland, EF178717, EF178899, EF178991, EF178809;
M. crispatum, Moody 433, 437, 445 (CONN), Australia, New South Wales;
Les 660 (CONN), Australia, New South Wales, EF178721, EF178903,
EF178995, EF178813; M. crispatum (WA), Moody 413, 418 (CONN), Australia,
Western Australia, FJ870942; M. decussatum, *Stretch s. n. (PERTH), Australia,
Western Australia, FJ870943, FJ861354, FJ870930, FJ861337; M. dicoccum,
Jacobs 8252, 8259 (NSW), Australia, Northern Territory, AY335976,
AY336006, AY335982,——–; M. drummondii, Moody 409 (CONN), Australia,
Western Australia, EF178725, EF178907, EF178999, EF178817; M. echinatum, *Keighery 689 (PERTH), Australia, Western Australia, FJ870944; M.
farwellii, Moody 97 (CONN), U. S. A., Minnesota; Moody 106 (CONN), U. S.
A., Wisconsin; Callahan s. n. (CONN), U. S. A., New Hampshire, EF178731,
EF178913, EF179005, EF178823; M. filiforme (1), *Wilson 1810 (NSW),
Australia, Northern Territory, EF178716, EF178898, EF178990, EF178808;
M. filiforme (2), Jacobs 8021 (CONN), Australia, Western Australia, FJ870945,
FJ861351, EF178990, FJ861334; M. heterophyllum (1), Moody 101 (CONN), U.
S. A., Minnesota; Moody 150 (CONN), U. S. A., Maine, AF513824, EF178915,
EF179007, EF178825; M. heterophyllum (2), Moody H2 (CONN), U. S. A.,
Connecticut; Moody 149 (CONN), U. S. A., Maine, AF513823; M. heterophyllum (3), Moody 105 (CONN), U. S. A., Wisconsin; Moody 141 (CONN), U. S.
A., Rhode Island; Moody 143 (CONN), U. S. A., Massachusetts; Moody 176
(CONN), U. S. A., South Carolina; Moody 178 (CONN), U. S. A., Oregon,
AF513822; M. hippuroides, Moody 179 (CONN), U. S. A., Oregon, FJ870946,
FJ861364, FJ870939, FJ861347; M. humile, Moody 141 (CONN), U. S. A.,
Rhode Island; Moody CT10 (CONN), U. S. A., Connecticut; Gerber s. n.
(CONN), U. S. A., Wisconsin, FJ870947, FJ861363, FJ870938, FJ861346;
M. latifolium, *Orchard 4793 (NSW), Australia, New South Wales; *Jacobs
6706 (NSW), Australia, New South Wales, EF178729, EF178911, EF179003,
EF178821; M. laxum (1), Moody 57, 58, 67, 77, 173, 174 (CONN), U. S. A.,
Florida, EF178732, EF178914, EF179006, EF178824; M. laxum (2), Moody
170 (CONN), U. S. A., South Carolina; Moody 172 (CONN), U. S. A., North
Carolina, FJ870948, FJ861365, FJ870940, FJ861348; M. limnophilum, Moody
417 (CONN), Australia, Western Australia, FJ870949, FJ861358, FJ870933,
FJ861341; M. lophatum, Moody 455, 456 (CONN), Australia, New South
Wales, EF178718, EF178900, EF178992, EF178810; M. mattogrossense, Ritter
2314 (LPB), Bolivia, Carrasco, EF178728, EF178910, EF179002, EF178820;
M. muricatum (1), Jacobs 8557 (NSW), Australia, Queensland, FJ870963,
FJ861362, FJ870937, FJ861345; M. muricatum (2), Jacobs 8577 (NSW),
Australia, Queensland, FJ870964; M. muricatum (3), Jacobs 8604 (NSW),
Australia, Queensland, FJ870965; M. oguraense, Kadono s. n. (HYO), Japan,
Hyogo, EF178705, EF178887, EF178979, EF178798; M. papillosum, Moody
424 (CONN), Australia, New South Wales; Les 614 (CONN), Australia,
New South Wales, (FJ870950, FJ870951, FJ870952), EF178906, EF178998,
EF178816; M. pedunculatum subsp. pedunculatum, Moody 452, 467 (CONN),
Australia, New South Wales; Les 652 (CONN), Australia, New South
Wales, EF178711, EF178893, EF178985, EF178804; M. pedunculatum (T), Les
643 (CONN), Australia, Tasmania, FJ870953, FJ861357, EF178985, FJ861340;
M. petraeum, *Archer 1564 (NSW), Australia, Western Australia; *Brown
1123 (PERTH), Australia, Western Australia, EF178712, EF178894,
EF178986, EF178805; M. pinnatum, Moody 511 (CONN), U. S. A., Connecticut; NASC (GH), U. S. A., Missouri, FJ870966, FJ890500, FJ890498,
FJ890499; M. quitense (NA), Moody 180, 183 (CONN), U. S. A., Oregon,
2010]
MOODY AND LES: MYRIOPHYLLUM SYSTEMATICS
EF178700, EF178882, EF178974, EF178793; M. quitense (SA), *Goodall 83 (C),
Chile; Ritter 3939 (LPB), Bolivia, FJ870954, FJ861352, EF178974, FJ861335;
M. robustum, Schwarz s. n. (CONN), New Zealand (Cult.), FJ870955,
FJ861353, FJ870929, FJ861336; M. salsugineum, Moody 412 (CONN),
Australia, Victoria; Les (Cult.), Australia, University of Tasmania,
EF178701, EF178883, EF178975, EF178794; M. sibiricum, Moody 82 (CONN),
U. S. A., California; Moody 99 (CONN), U. S. A., Minnesota; Moody 181
(CONN), U. S. A., Oregon; Moody 125 (CONN), U. S. A., Wisconsin; Moody
212 (CONN), U. S. A., Colorado, DQ786014-DQ86018, EF178706, EF178888,
EF178980; M. sibiricum (Eu.), *Ulvinen 6124 (C), Finland; *Fenskild 8291 (C),
Finland; *Hamen 61 (C), Denmark, FJ870956, FJ861350, FJ870928, FJ861333;
M. simulans 441, Moody 440, 441 (CONN), Australia, Victoria, EF178719,
EF178901, EF178993, EF178811; M. simulans 622, Les 622, 646 (CONN),
Australia, New South Wales, EF178722, EF178904, EF178996, EF178814;
M. spicatum, Moody 79 (CONN), U. S. A., Florida; Moody 86 (CONN), U. S.
A., California; Moody 117 (CONN), U. S. A., Wisconsin; Moody 134 (CONN),
U. S. A., Connecticut; Moody 159 (CONN), U. S. A., Indiana; Moody 164
(CONN), U. S. A., Minnesota; Moody 185 (CONN), U. S. A., Oregon;
EF178702, EF178884, EF178976, EF178795; M. tenellum, Moody 110 (CONN),
U. S. A., Wisconsin; Moody 93 (CONN), U. S. A., Minnesota; Callahan s.n.
(CONN), U. S. A., New Hampshire, EF178730, EF178912, EF179004,
EF178822; M. tillaeoides, Moody 415 (CONN), Australia, Western Australia,
EF178710, EF178892, EF178984, EF178803; M. trachycarpum, Jacobs 8843
(NSW), Australia, Northern Territory; Martine 863 (CONN), Australia,
Western Australia, EF178715, EF178897, EF178989, EF178807; M. trifidum,
139
Moody 404, 405, 410 (CONN), Australia, Western Australia, EF178734,
EF178916, EF179008, EF178826; M. triphyllum, Glenny 7455 (PDD), New
Zealand, FJ870957, FJ861349, FJ870927, FJ861332; M. ussuriense, Kadono
“Kasai-1” (HYO), Japan, Hyogo; Kadono “Kasai-2” (HYO), Japan, Hyogo,
EF178726, EF178908, EF179000, EF178818; M. variifolium 444, Moody 444,
469, 473 (CONN), Australia, New South Wales; Les 628 (CONN), Australia,
FJ870958, FJ861360, FJ870935, FJ861343; M. variifolium 457, Moody 450, 457,
460 (CONN), Australia, Victoria, FJ870959, FJ861361, FJ870936, FJ861344;
M. verrucosum, Moody 427 (CONN), Australia, New South Wales; Les 601
(CONN), Australia, New South Wales, Jacobs 8777 (NSW), Australia,
Northern Territory, EF178707, EF178889, EF178981, EF178800; M. verticillatum, Moody 83 (CONN), U. S. A., California; Moody 100 (CONN), U. S. A.,
Minnesota; Moody 147 (CONN), U. S. A., Maine; Moody 177 (CONN), U. S. A.,
Oregon, EF178709, EF178891, EF178983, EF178802; M. votschii, *Rixon 31
(NSW), New Zealand EF178714, EF178896, EF178988, EF178806; M. sp.
425, Moody 425 (CONN), Australia, New South Wales, FJ870960, FJ861359,
FJ870934, FJ861342; M. sp. (red 1), (Cultivated in U. S. A., Maine pet store
as ‘M. mattogrossense’); (Cultivated in U. S. A., Washington pet store as
‘M. propinquum’), FJ870962, FJ861356, FJ870932, FJ861339; M. sp. (red 2),
Jacobs 8547 (CONN) Cultivated in Australia, FJ870961, FJ861355, FJ870931,
FJ861338; M. sp. nov. 476, Moody 476 (CONN), Australia, New South
Wales, EF178723, EF178905, EF178997, EF178815; M. sp. nov. 542, Les 542
(CONN), Australia, New South Wales, EF178713, EF178895, EF178987,
—–. Trihaloragis hexandrus, *Bright 93 (PERTH), Australia, WA; *Lepschi
3360 (PERTH), Australia, WA EF178759, EF178941, EF179033, EF178849.
Michael Moody