Relationships within
Tectaria
: the undivided leaves unraveled
Our phylogenetic tree shows that the
Tectaria
clade is composed of three clades all with strong support and each with a different geographical distribution: one is Palaeotropical (
Tectaria
I), one is Neotropical (
Tectaria II
) and one is South-East Asian (
Tectaria III
).
Tectaria
clade III includes the peculiar genus
Cionidium
, an endemic to New Caledonia, which was segregated because of its extramarginal (stalked) sori and lanceolate, non-clathrate scales. Even though Bower (1928) placed
Cionidium moorei
in
Deparia Hook. & Grev.
, it resembles
Tectaria
in most other characters, except the sori. Copeland (1947) therefore considered it most closely related to
T. seemannii (E.Fourn.) Copeland (1929: 359)
, another New Caledonian species. Our phylogenetic analyses places
Cionidium
with reasonable support in a clade comprised of the South-East Asian species
T. impressa (Fée) Holttum (1988: 483)
and
T. simonsii (Baker) Ching (1931: 32)
and the more widespread
T. devexa (Kunze) Copeland (1907: 415)
, which extends into Polynesia.
The Neotropical clade
Tectaria
I includes
T. trifoliata
(L.) Cavanilles (1802: 249), which is the type species of the genus. We also included two samples of
T. incisa
and one of its synonyms,
T. martinicensis (Spreng.) Copeland (1907: 410)
. This taxon is variable and taxonomically complex and probably includes more than one species. Indeed, the three specimens were not resolved together, indicating that this species is not monophyletic. A more detailed study of the
T. incisa
complex using morphological and molecular characters will be needed to tease this species complex apart and allow its biogeography to be discussed. Taxonomic study of this species complex is important because of its weedy and invasive nature and possibility of hybridisation with other taxa ( Wagner et al. 1978, Gordon & Thomas 1997).
Fadyenia prolifera Hooker
(in Hooker & Bauer 1840: t53-B) is a Caribbean species that was originally separated from
Tectaria
because of its simple proliferous leaves and peculiar lunate indusia ( Hooker & Bauer 1840). Later it was placed in
Tectaria
because of its anastomosing veins forming elongate areoles ( Tryon & Tryon 1981), a placement which was corroborated by molecular studies ( Schuettpelz & Pryer 2007). In our analyses,
T. prolifera (Hook.) Tryon & Tryon (1981: 136)
was found to be sister to
T. panamensis
, the type species of the segregate genus
Dictyoxiphium
. Since
Dracoglossum
View in CoL
has been excluded from
Tectaria
,
T. panamensis
and
T. prolifera
are the only Neotropical species in the genus with simple leaves. However, separating them from other
Tectaria spp.
at a generic level is not justified according to our results, because their lineage is deeply embedded in the Neotropical clade of
Tectaria
.
Tectaria panamensis
is also known to hybridize with
T. incisa
, resulting in sterile plants with intermediate leaf division ( Wagner et al. 1978). Because most juvenile sporophyte leaves are simple, it has been suggested that entire leaves in adults belonging to lineages with generally divided leaves are a form of paedomorphy or neoteny, simple leaves having evolved through reduction, as is also seen in
Marattiaceae
( Stidd 1974, Christenhusz et al. 2008, Christenhusz 2010a) and
Marsileaceae ( Pryer & Hearn 2009)
View in CoL
.
In
Tectaria
clade I, about four lineages can be recognized, although some with weak support only. The genus
Ctenitopsis
(here represented by
Tectaria fuscipes (Wall. ex Bedd.) Christensen 1931: 290
,
T. kusukusensis (Hayata) Lellinger 1968: 157
,
T. sagenioides (Mett.) Christenhusz 2010b: 58
and
T. subsageniacea (Christ) Christenhusz 2010b: 59
), has been segregated from
Tectaria
on the basis of partially anastomosing veins and absence of included veinlets ( Tardieu-Blot & Christensen 1938). However, in our analyses it forms a well supported clade that is deeply embedded in
Tectaria
. A second set of species with simple leaves showing paedomorphy is found in
Tectaria
clade I:
Tectaria singaporiana (Wall. ex Hook. & Grev.) Copeland (1917: 368)
, the type species of former segregate genus
Podopeltis
, is sister to
T. harlandii (Hook.) Kuo (2002: 173)
, the latter species previously known as
Hemigramma decurrens (Hook.) Copeland (1928: 404
; see Xing et al. 2013). Both species have simple leaves (although they are deeply lobed in
T. harlandii
) and sporangia with confluent sori along the veinlets. This is different from typical
Tectaria spp.
, which have sporangia in distinct round sori.
Another segregate genus with simple leaves is
Quercifilix
. This genus has dimorphic leaves with the fertile leaves having laminae that are much contracted and with sporangia placed densely along veinlets. However,
Quercifilix
(
T. zeilanica (Houtt.) Sledge 1972: 422
),
Podopeltis
(
T. singaporiana
) and
Hemigramma
(
T. harlandii
) all have the more or less anastomosing veins that
form copious
areoles typical of many
Tectaria
. Holttum (1988) did not recognize
Hemigramma
as a natural group, but showed that its species are similar to
Tectaria
. The condition of leaf dimorphism has arisen many times in ferns, in
Tectaria
alone at least eight times according to the results of our phylogenetic analysis.
Tectaria
also contains species with partial leaf dimorphism: the fertile leaves are otherwise similar to sterile ones, but are more contracted and/or have longer petioles. The segregate genera
Fadyenia
,
Hemigramma, Podolepis
and
Quercifilix
were all based on leaf dimorphism but are now included in
Tectaria
( Kramer et al. 1990, Smith 1995, Schuettpelz & Pryer 2007, Christenhusz et al. 2011), and it is more likely that they are derived through reduction and paedomorphy as has been suggested for other fern families (e.g. Asama 1960, Stidd 1974, Pryer & Hearn 2009, Christenhusz 2010a).