The taxonomic disposition of family Lemnaceae has become controversial in recent years,
following phylogenetic studies that merged
Lemnaceae within a more broadly inclusive
Araceae (Cabrera et al. 2008; Cusimano et al.
2011; Nauheimer et al. 2012; Henriquez et al.
2014). Analyses using extensive plastid DNA
data indicate that the most recent common
ancestor of the core Araceae and the morphologically divergent genera Gymnostachys, Lysichiton, Orontium, and Symplocarpus also gave
rise to duckweeds. Thus, to preserve Araceae as a
monophyletic family, the broader community
(including the influential Angiosperm Phylogeny
Group; APG IV 2016) has opted to sink duckweeds within Araceae. However, we continue to
advocate the continued recognition of Lemnaceae as a distinct family for several reasons.
First, the available phylogenetic evidence
supporting the descent of Lemnaceae from
within Araceae cannot be considered comprehensive. Previous phylogenetic analyses of Araceae used only plastid sequence data (Cusimano
et al. 2011; Nauheimer et al. 2012), without
sampling any nuclear loci. Plastid phylogenies
for Araceae sensu lato consistently show an
alternative topology for duckweeds that places
Lemna as sibling to Landoltia + Wolffia + Wolffiella, rather than showing the topology
obtained in duckweed-specific studies, i.e., Landoltia sibling to Lemna + Wolffia + Wolffiella
(Les et al. 2002; Tippery et al. 2015). Numerous
examples exist of nuclear and plastid phylogenetic discordance across land plants (e.g., Tippery and Les 2011; Bruun-Lund et al. 2017), as
well as persistent taxonomic issues that have not
been resolved unambiguously using molecular
data (e.g., Mathews 2009; Wickett et al. 2014).
Moreover, the Araceae s.l. studies recovered
exceptionally long-branch lengths for all Lemnaceae genera and highlighted their morphological distinctness in fundamental traits such as
pollen structure and chromosome number
(Cusimano et al. 2011). Although maximum
likelihood and Bayesian phylogenetic methods
are less sensitive to long-branch artifacts (Gaut
and Lewis 1995; Philippe et al. 2005), there
remains evidence that long-branch lineages
should be evaluated with caution (Kück et al.
2012).
Secondly, even if the phylogeny depicted by
some authors (e.g., Cabrera et al. 2008; Cusimano et al. 2011) correctly depicts evolutionary
relationships, a morphologically more satisfying
taxonomy could be achieved by recognizing
family Orontiaceae R.Br. ex Lindl. (i.e., Gymnostachydoideae Bogner & Nicolson and Orontioideae Mayo, Bogner & P.C.Boyce) as sibling
to Lemnaceae + Araceae. Orontiaceae long have
been considered distinct from “true” Araceae
(sensu Mayo et al. 1997; Cusimano et al. 2011),
and they differ in some noteworthy features, such
as largely dimerous flowers, unidirectional tepal
and stamen development, and orthotropous ovule
orientation, although these features also are
homoplasic among Araceae s.l. genera (Grayum
1991; Buzgo 2001). The circumscription of
family Orontiaceae is not without precedent.
Brown (1810) initially published “Orontiaceae”
as a section of family “Aroideae” (i.e., Araceae),
and later, Lindley (1846) elevated Orontiaceae to
the level of “order,” at a level equivalent to
“order” Araceae. The International Code of
Nomenclature (ICN; Turland et al. 2018) Art.
18.2 allows such “orders” to be treated as validly
published families, if that was the intent of the
author. Given the equivalent presentation of
Araceae by Lindley, it appears quite clear that he
intended for Orontiaceae to be considered at the
rank of family also. Segregating Orontiaceae
from Araceae s.l. would allow a distinct lineage
to be understood more clearly in terms of its
morphological evolution and biogeography. In
contrast, sinking the undeniably distinct Lemnaceae within a broader Araceae stands to upset
the morphological uniformity of Araceae.
Finally, the amount of nomenclatural confusion would be reduced if the taxa of Orontiaceae
were removed from Araceae. The family Orontiaceae contains only nine species in four genera
(Gymnostachys: 1, Lysichiton: 2, Orontium: 1,
Symplocarpus: 5), whereas Lemnaceae comprise
37 species in five genera. The concept of family
Lemnaceae extends back nearly two centuries
(Gray 1821; IPNI 2018), and over that time, there
has been little confusion about what species
2 Tiny Plants with Enormous Potential: Phylogeny and Evolution …
23
following phylogenetic studies that merged
Lemnaceae within a more broadly inclusive
Araceae (Cabrera et al. 2008; Cusimano et al.
2011; Nauheimer et al. 2012; Henriquez et al.
2014). Analyses using extensive plastid DNA
data indicate that the most recent common
ancestor of the core Araceae and the morphologically divergent genera Gymnostachys, Lysichiton, Orontium, and Symplocarpus also gave
rise to duckweeds. Thus, to preserve Araceae as a
monophyletic family, the broader community
(including the influential Angiosperm Phylogeny
Group; APG IV 2016) has opted to sink duckweeds within Araceae. However, we continue to
advocate the continued recognition of Lemnaceae as a distinct family for several reasons.
First, the available phylogenetic evidence
supporting the descent of Lemnaceae from
within Araceae cannot be considered comprehensive. Previous phylogenetic analyses of Araceae used only plastid sequence data (Cusimano
et al. 2011; Nauheimer et al. 2012), without
sampling any nuclear loci. Plastid phylogenies
for Araceae sensu lato consistently show an
alternative topology for duckweeds that places
Lemna as sibling to Landoltia + Wolffia + Wolffiella, rather than showing the topology
obtained in duckweed-specific studies, i.e., Landoltia sibling to Lemna + Wolffia + Wolffiella
(Les et al. 2002; Tippery et al. 2015). Numerous
examples exist of nuclear and plastid phylogenetic discordance across land plants (e.g., Tippery and Les 2011; Bruun-Lund et al. 2017), as
well as persistent taxonomic issues that have not
been resolved unambiguously using molecular
data (e.g., Mathews 2009; Wickett et al. 2014).
Moreover, the Araceae s.l. studies recovered
exceptionally long-branch lengths for all Lemnaceae genera and highlighted their morphological distinctness in fundamental traits such as
pollen structure and chromosome number
(Cusimano et al. 2011). Although maximum
likelihood and Bayesian phylogenetic methods
are less sensitive to long-branch artifacts (Gaut
and Lewis 1995; Philippe et al. 2005), there
remains evidence that long-branch lineages
should be evaluated with caution (Kück et al.
2012).
Secondly, even if the phylogeny depicted by
some authors (e.g., Cabrera et al. 2008; Cusimano et al. 2011) correctly depicts evolutionary
relationships, a morphologically more satisfying
taxonomy could be achieved by recognizing
family Orontiaceae R.Br. ex Lindl. (i.e., Gymnostachydoideae Bogner & Nicolson and Orontioideae Mayo, Bogner & P.C.Boyce) as sibling
to Lemnaceae + Araceae. Orontiaceae long have
been considered distinct from “true” Araceae
(sensu Mayo et al. 1997; Cusimano et al. 2011),
and they differ in some noteworthy features, such
as largely dimerous flowers, unidirectional tepal
and stamen development, and orthotropous ovule
orientation, although these features also are
homoplasic among Araceae s.l. genera (Grayum
1991; Buzgo 2001). The circumscription of
family Orontiaceae is not without precedent.
Brown (1810) initially published “Orontiaceae”
as a section of family “Aroideae” (i.e., Araceae),
and later, Lindley (1846) elevated Orontiaceae to
the level of “order,” at a level equivalent to
“order” Araceae. The International Code of
Nomenclature (ICN; Turland et al. 2018) Art.
18.2 allows such “orders” to be treated as validly
published families, if that was the intent of the
author. Given the equivalent presentation of
Araceae by Lindley, it appears quite clear that he
intended for Orontiaceae to be considered at the
rank of family also. Segregating Orontiaceae
from Araceae s.l. would allow a distinct lineage
to be understood more clearly in terms of its
morphological evolution and biogeography. In
contrast, sinking the undeniably distinct Lemnaceae within a broader Araceae stands to upset
the morphological uniformity of Araceae.
Finally, the amount of nomenclatural confusion would be reduced if the taxa of Orontiaceae
were removed from Araceae. The family Orontiaceae contains only nine species in four genera
(Gymnostachys: 1, Lysichiton: 2, Orontium: 1,
Symplocarpus: 5), whereas Lemnaceae comprise
37 species in five genera. The concept of family
Lemnaceae extends back nearly two centuries
(Gray 1821; IPNI 2018), and over that time, there
has been little confusion about what species
2 Tiny Plants with Enormous Potential: Phylogeny and Evolution …
23