ArticlePDF Available

A taxonomic revision of the eusporangiate fern family Marattiaceae, with description of a new genus Ptisana

Authors:

Abstract and Figures

Recent phylogenetic investigations of marattioid ferns (Marattiaceae) have generated a refined understanding of genus- and species-level relationships within the family. A taxonomic revision of marattioid ferns is presented here. This new taxonomy is congruent with the current understanding of the phylogeny of marattioid ferns based on both molecular and morphological data. Marattiaceae is recognized here in the broad and most commonly used sense, encompassing all extant genera. Marattia as traditionally defined is paraphyletic, and is split here into three morphologically distinct and monophyletic genera: Marattia s.str. is restricted to the Neotropics and Hawaii; Eupodium, a distinctive Neotropical genus that has not been recognized by recent authors, is resurrected; and a newly described genus Ptisana comprises the Paleotropical species. Angiopteris is circumscribed broadly to include Archangiopteris, Macroglossum, Protomarattia, and Protangiopteris. Neither Danaea nor Christensenia require changes to their circumscriptions. As Marattia s.l. is undergoing the most drastic change in this taxonomy, species-level detail is given for Marattia s.str. and the segregated genera Eupodium and Ptisana. New combinations are made in Eupodium and Ptisana. Considerable monographic work still remains to be done, and is of particular importance given the threatened status of many presently poorly understood species in this family.
Content may be subject to copyright.
737
Murdock • Taxonomic revision of MarattiaceaeTAXON 57 (3) • August 2008: 737–755
INTRODUCTION
Marattioid ferns (marattioids) are an ancient lineage
of eusporangiate ferns with extant species essentially
restricted to the tropical regions of the world. Marat-
tioids are an early diverging lineage of the ferns, related
to leptosporangiate ferns (Polypodiopsida), Equisetales,
Ophioglossales, and Psilotales (Smith & al, 2006). The
interrelationships of these lineages are controversial and
generally poorly resolved in molecular phylogenies (Roth-
well, 1996, 1999; Pryer & al., 2001, 2004; Wikström &
Pryer, 2005; Rothwell & Nixon, 2006; Schuettpelz & al.,
2006). Extant marattioids can be diagnosed morphologi-
cally on the basis of multiple distinctive characters, sev-
eral of which have been used to diagnose putative fossil
relatives. These characters include: large, thick-walled
sporangia that are fully or partially fused into synangia,
with each sporangium producing large quantities of small
spores; complex polycyclic stem vasculature; leaves ba-
sally flanked by paired, starchy, vascularized auricles (of-
ten called “stipules”); leaves possessing prominent fleshy
swellings (pulvini) at nodes and/or bases of segments;
extensive mucilage canal systems throughout the tissues
of the plant; cyclocytic multicyclic stomata; first leaves
of sporophytes emerging through the upper surface of the
gametophyte; and rod-shaped projections on mesophyll
cells. Despite numerous publications on various aspects of
the unique morphology of marattioids, the delimitation of
extant taxa is still poorly understood (Holttum, 1978; Hill
& Camus, 1986; Rolleri & al., 1991; Murdock, 2008).
Fossil marattioid ferns are first known from the Car-
boniferous, reaching their peak of diversity in the Permian
(Hill & Camus, 1986). Forms more directly comparable
to extant lineages first appeared in the Mesozoic, includ-
ing some fossil taxa that have been treated in the modern
genera Marattia and Angiopteris (Harris, 1961; Van Cittert,
1966). Marattioid ferns are lacking from the Tertiary fossil
record, with the possible exception of Danaea coloradensis
Knowlt. from the Eocene of Colorado (but see MacGini-
tie, 1969). Modern marattioids are generally thought to be
relictual representatives of a lineage that was at one time
more dominant and widespread, and are morphologically
and genetically isolated from other fern lineages (Mamay,
1950; Millay, 1977; Taylor & Taylor, 1993; Collinson, 2001).
Authors have expressed this phylogenetic isolation taxo-
nomically by creating names for marattioid ferns at every
major suprageneric rank from division to family (e.g., divi-
sion Marattiophyta Heintze, Cormofyt. Fylog.: 22. 1927;
Class Marattiopsida Doweld, Tent. Syst. Pl. Vasc.: vii. 2001;
subclass Marattiidae Klinge, Fl. Est-Liv-Churland 1: 93.
1882; order Marattiales Link, Hort. Berol. 2: 148. 1833)
(Murdock & al. 2006; Smith & al., 2006).
For extant marattioid taxa, the most common prac-
tice has been to treat all genera within a single family
A taxonomic revision of the eusporangiate fern family Marattiaceae, with
description of a new genus Ptisana
Andrew G. Murdock
Department of Integrative Biology, 1001 Valley Life Sciences Bldg., University of California Berkeley,
Berkeley, California 94720-2465, U.S.A. murdock@berkeley.edu
Recent phylogenetic investigations of marattioid ferns (Marattiaceae) have generated a refined understanding
of genus- and species-level relationships within the family. A taxonomic revision of marattioid ferns is pre-
sented here. This new taxonomy is congruent with the current understanding of the phylogeny of marattioid
ferns based on both molecular and morphological data. Marattiaceae is recognized here in the broad and most
commonly used sense, encompassing all extant genera. Marattia as traditionally defined is paraphyletic, and
is split here into three morphologically distinct and monophyletic genera: Marattia s.str. is restricted to the
Neotropics and Hawaii; Eupodium, a distinctive Neotropical genus that has not been recognized by recent
authors, is resurrected; and a newly described genus Ptisana comprises the Paleotropical species. Angiopteris
is circumscribed broadly to include Archangiopteris, Macroglossum, Protomarattia, and Protangiopteris.
Neither Danaea nor Christensenia require changes to their circumscriptions. As Marattia s.l. is undergoing the
most drastic change in this taxonomy, species-level detail is given for Marattia s.str. and the segregated genera
Eupodium and Ptisana. New combinations are made in Eupodium and Ptisana. Considerable monographic
work still remains to be done, and is of particular importance given the threatened status of many presently
poorly understood species in this family.
KEYWORDS: Angiopteris, Christensenia, Danaea, Eupodium, Marattia, Marattiaceae, Ptisana
738
TAXON 57 (3) • August 2008: 737–755Murdock • Taxonomic revision of Marattiaceae
Marattiaceae (Rolleri & al., 2003; Smith & al., 2006).
Numerous conflicting classification systems have been
proposed for taxa traditionally treated within Marat-
tiaceae s.l., but there has been a general consenus that
these taxa as a whole constitute an evolutionary group,
even when authors have chosen to dissect Marattiaceae
into several smaller families. Morphological, molecu-
lar, and combined phylogenetic analyses have consis-
tently supported the monophyly of marattioid ferns (Hill
& Camus, 1986; Stevenson & Loconte, 1996; Pryer &
al., 2001, 2004; Wikström & Pryer, 2005; Rothwell &
Nixon, 2006; Schuettpelz & al., 2006; Qiu & al., 2007;
Murdock, 2008). Amongst those who have split the group
into smaller families, there has been very little agreement.
The most common ways that authors have chosen to split
Marattiaceae are (1) to segregate Danaea into its own
family, Danaeaceae (Presl, 1845; de Vriese & Harting,
1853) or (2) to recognize four families Danaeaceae, Mar-
attiaceae, Angiopteridaceae, and Christenseniaceae (or
Kaulfussiaceae) (Reimers, 1954; Ching, 1958a; Pichi Ser-
molli, 1957, 1969, 1970, 1977; Devi, 1981; Doweld, 2001).
Some have broadened the circumscription of Danaeaceae
to include Christensenia (e.g., Bierhorst, 1971; Takhtajan,
1986) or Macroglossum (Nakai, 1937). Others have rec-
ognized two families, Marattiaceae and Angiopteridaceae
(e.g., Christensen, 1938; Stevenson & Loconte, 1996),
with Marattiaceae including Danaea and Christensenia.
Campbell (1940) treated Christensenia in its own family
but left the remaining genera in Marattiaceae. Additional
distinct familial taxonomies exist beyond these examples
(e.g., Campbell, 1895), serving to indicate the general lack
of consensus on how to classify marattioid ferns.
Murdock (2008) inferred the phylogeny of marattioid
ferns based on combined analysis of nucleotide sequence
data (plastid trnS-trnG intergenic spacer plus the trnG
intron, rps4-trnS intergenic spacer, atpB, and rbcL) and
morphology. The resulting phylogeny (Fig. 1) supports
several possible familial circumscriptions, most logically
amongst these being: (1) Marattiaceae s.l. (i.e., incl. all
genera); or (2) Danaeaceae (incl. only Danaea) and Mar-
attiaceae (incl. all other genera). Choosing between these
(and other possibilities) is entirely subjective; moreover,
dividing Marattiaceae into smaller families serves little
purpose other than to suggest that these lineages are of
relatively ancient origin. The best course is to maintain
Marattiaceae in the broad and most commonly used sense,
particularly in light of the distinctive suite of shared de-
rived characters that unites marattioids and distinguishes
them from all other extant plant lineages.
The topology produced by Murdock (2008) does not
correspond with any single previous generic taxonomy, in-
cluding the most recent treatment by Rolleri & al. (2003).
However, most of the well-supported clades correspond to
previously described genera. In this paper the genus-level
taxonomy of Marattiaceae is revised to be congruent with
the current understanding of the phylogeny and homol-
ogy of morphological characters. The circumscriptions
of Christensenia and Danaea remain unchanged and un-
controversial. Angiopteris is monophyletic only if circum-
scribed to include Archangiopteris and Macroglossum, a
result also found by the morphological cladistic analysis
of Marattiaceae produced by Hill & Camus (1986) (with
several new taxonomic combinations in Angiopteris made
later; Camus, 1989). Marattia in the traditional sense is
paraphyletic, and includes several lineages that are readily
diagnosable based on both morphology and geographic
distribution.
There are several options for resolving the paraphyly
of Marattia taxonomically. (1) The circumscription of
Marattia could be broadened to include the genus An-
giopteris. Since Angiopteris is morphologically distinct,
comprises over 200 named species, is unique in the family
in having spherical trilete spores, and because Marattia
and Angiopteris are thought from fossil evidence to have
diverged at least 160 mya (Hill & Camus, 1986; Liu & al.,
2000), this approach seems unreasonable. (2) Marattia
could be split into two genera, Marattia and Eupodium
(the name with priority for the second genus). Eupodium
was separated from Marattia entirely on the basis of the
stalked synangia found in the Marattia laevis/kaulfussii
species complex, and as such it seems inappropriate to
apply this name to the relatively species-rich Paleotro-
pical clade with sessile synangia and leaf morphology
unlike that of Eupodium. Additionally, both Eupodium
and the Paleotropical clade are highly supported on very
long branches (Murdock, 2008), but the grouping of the
two clades received somewhat decreased support. I find
neither of the preceding options acceptable, and so I am
choosing a third option: (3) split Marattia s.l. into three
morphologically distinct genera: Marattia Sw. (type:
Marattia alata Sw.), Eupodium J. Sm. (type: Eupodium
kaulfussii (J. Sm.) J. Sm.), and Ptisana Murdock, gen.
nov., all of which are described in detail below. None of
the available options serves nomenclatural stability espe-
cially well as all options require many new combinations
and the first option would require multiple new names
for the homonyms created by combining Angiopteris and
Marattia. If preserving nomenclatural stability is of para-
mount importance, one could propose a conserved type
for Marattia within the Paleotropical Marattia (Ptisana):
there are relatively few species in Marattia s.str., there are
multiple available generic names, and several previous
authors (notably C. Presl and de Vriese) construed the type
of Marattia to be the African M. fraxinea Sm, albeit erro-
neously. However, there is sufficient diversity in Marattia
s.str., as well as a long history of nomenclatural usage in
the Neotropics and Hawaii, such that this option is dif-
ficult to defend. Additionally, while retypifying Marattia
739
Murdock • Taxonomic revision of MarattiaceaeTAXON 57 (3) • August 2008: 737–755
Fig. 1. Consensus phylogeny based on nucleotide sequence data, coded indels, and morphology based on Murdock
(2008). Generic names are those used in the present revision, generic types indicated by a τ. Support values are shown
for the branches subtending recognized genera (above: Bayesian posterior probability, maximum parsimony bootstrap;
below: decay index). Bold lines indicate a Bayesian posterior probability 97%. Silhouettes are diagrammatic cross-sec-
tions of synangia typical for each genus. Lack of lateral fusion between sporangia in Angiopteris is indicated by a second
row of sporangia. Dehiscence pores are shown for Danaea, all other genera dehisce via slits.
740
TAXON 57 (3) • August 2008: 737–755Murdock • Taxonomic revision of Marattiaceae
may serve stability in one sense, it would certainly create
unnecessary confusion.
While this paper focuses on extant taxa, the revision
of generic circumscriptions in Marattiaceae will affect
interpretation of fossil affinities, nomenclature, and con-
clusions drawn from these fossils (e.g., divergence time
estimation using fossil calibration of phylogenetic trees,
trends in morphological evolution, and historical bio-
geographic patterns). While numerous fossils have been
ascribed to Marattia (or Marattiopsis Schimper), these
appear to be more closely allied with Ptisana ; in fact there
are no known fossils clearly referable to Marattia under
the concept of the genus presented here. Marattia hook-
eri Ettingsh. & Gard., an Eocene fossil, was described
as being most similar to Marattia kaulfussii (Gardner &
Ettingshausen, 1882), which is treated here in the genus
Eupodium. However, this species was based on sterile
and fragmentary fossils; in such a state it is nearly impos-
sible to differentiate between Marattia and Eupodium.
Moreover, it is unlikely that this plant belongs in Marat-
tiaceae at all, as the branching of the veins is most unlike
that seen in modern Eupodium or any other marattioid
genus. Further study of marattioid fossil material will be
required to produce an integrative taxonomy of extinct
and extant taxa.
Significant species-level monographic work remains
to be done in this family, and is of particular importance
given the threatened status of many presently poorly un-
derstood species in this family. Several species of Angiop-
teris, particularly those formerly placed by many authors
in Archangiopteris and Macroglossum, are threatened or
endangered in their native ranges (Fu & Jin, 1992; Hsu &
al., 2000; Kingston & al., 2004). Many marattioid species
are threatened by habitat destruction and by introduced
feral pigs that favor plants in this family as food sources
(Warshauer, 1979). Highly restricted endemics such as
Ptisana howeana of Lord Howe Island and Marattia doug-
lasii of Hawaii, both of which are seriously threatened
and very rare, are particularly susceptible and should be a
critical conservation concern (Pickard, 1983; Brownsey &
Smith-Dodsworth, 2000). Since our understanding of spe-
cies delimitation in this family is generally poor, establish-
ing conservation priorities is currently problematic and
will be greatly aided by future monographic studies.
TERMINOLOGY
Individual sporangia are grouped into sori, and the
sporangia in each sorus can be fused to form a synangium.
The opening of an individual sporangium is here called an
aperture. Sporangia in a synangium are separated laterally
by septa. Each half of a bilaterally symmetric synangium
is a valve. Individual sporangia may have labia formed
by differentiated cells surrounding the aperture that cre-
ate a subtle convex projection. Idioblasts are specialized
leaf cells that contain silica in the form of opal (used here
specifically in the sense of Rolleri & al., 2001b). Paraphy-
ses, used in the sense of Rolleri & al. (2000), refer to the
scales or uniseriate hairs attached to the receptacle below
a sorus, which are sometimes referred to collectively as an
“indusium” in older literature. False veins, also sometimes
called “recurrent veins,” “venuloids,” or “fibre bundles,
occur between true secondary leaf veins in Angiopteris
and occasionally in Marattia s.str. Pulvini are swollen re-
gions of leaves that occur at the base of segments, at nodes,
or along stipes. Auricles (also “stipules” or “trophopods”)
are paired appendages that occur at the base of stipes and
contain dormant vegetative buds.
TAXONOMIC CLASSIFICATION
OF MARATTIACEAE
Marattiaceae Kaulf., Enum. Filic.: 31. 1824; nom. cons.
prop. – Type: Marattia Sw., Prodr. 8.: 128. 1788.
= Danaeaceae C. Agardh, Aphor. Bot.: 117. 1822; nom.
rej. prop. – Type: Danaea Sm. in Mem. Acad. Sci.
Turin 5: 420, t. 9. 1793.
= Angiopteridaceae Fée ex J. Bommer in Bull. Soc.
Bot. Belgique 5: 345, 359. 1867 (asAngiopterideae”)
– Type: Angiopteris Hoffm. in Comment. Soc. Re-
giae Sci. Gott. 12: 29. 1796; nom. cons., non Adans.,
1763.
= Kaulfussiaceae Campb., Evol. Land Pl.: 333. 1940;
nom. illeg. (name based on an illegitimate generic
name) – Type: Kaulfussia Blume, Enum. Pl. Javae 2:
260. 1828; nom. illeg., non Dennst. (1818).
= Christenseniaceae Ching, Bull. Fan Mem. Inst. Biol.,
Bot. 10: 227. 1940 – Type: Christensenia Maxon in
Proc. Biol. Soc. Wash. 18: 239. 1905.
Habit terrestrial or occasionally epiphytic, rhizomes
erect, ascending, or creeping. Fronds monomorphic and
relatively long-lived or dimorphic with fertile fronds
short-lived; simple, palmately lobed, pedate, or once to
multiple times pinnately compound, with paired basal
auricles; stipes and auricles often with conspicuous scat-
tered lenticels; pulvini present at bases of fronds, at nodes
or bases of leaf segments, and occasionally along stipes at
leafless nodes; stipes winged or not; blade margins simple,
toothed, or irregularly crisped; veins free, simple, bifur-
cating, branching distally, or reticulate; false veins present
between true veins and extending from margin towards
midrib, or absent. Indument composed of multicellular
uniseriate hairs (simple or branched, glandular or not),
and scales (basifixed or peltate, sometimes raised on short
stalks). Sori lacking true indusia but often subtended by a
ring of paraphyses, superficial, sunken into the lamina, or
741
Murdock • Taxonomic revision of MarattiaceaeTAXON 57 (3) • August 2008: 737–755
stalked, elongate along veins and bilateral, or radial, spo-
rangia thick-walled, at least partially fused into synangia,
sporangia dehiscing via a slit or pore. Spores monolete or
trilete, occasionally alete or incipiently trilete, perispore
present, > 1,000 spores per sporangium. Prothalli large,
thalloid, photosynthetic, sporophytes emerging upwards
through the tissue of the gametophytes. Mycorrhizal as-
sociations common in stems, roots, and prothalli. Chro-
mosome number x = 40, 39.
KEY TO THE GENERA OF MARAT-
TIACEAE
1 Sori sunken into lamina, sporangia dehiscing via api-
cal pores; leaves generally strongly dimorphic, rarely
almost monomorphic; rachis nodes strongly pulvi-
noid, petiole bases slightly pulvinoid . . . .I. Danaea
1 Sori superficial or raised, rarely partially sunken, spo-
rangia dehiscing via lateral to apical slits; leaves ±
monomorphic; rachis nodes not or slightly pulvinoid,
petiole bases generally strongly pulvinoid . . . . . . . 2
2 Sori ± radially symmetric, scattered, dehiscing via ±
apical slits; blades palmately lobed to pedately com-
pound (appearing palmate); venation reticulate; scale
cells ± isodiametric; rhizomes creeping; stomata large
( > 150 μm), circular. . . . . . . . . . . .II. Christensenia
2 Sori bilaterally symmetric (rarely radially symmet-
ric), in regular rows, dehiscing via lateral slits; blades
pinnately compound; venation free; scale cells elon-
gate; rhizomes erect or creeping; stomata smaller
( < 100 μm), elliptical. . . . . . . . . . . . . . . . . . . . . . . . 3
3 Receptacles naked, synangia prominently stalked;
adaxial side of distal blade segments with awns scat-
tered along veins; plants bearing one frond at a time,
rarely two; restricted to the Neotropics . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .III. Eupodium
3 Receptacles bearing scales, uniseriate hairs, or na-
ked, synangia sessile to slightly raised (rarely short-
stalked); blades lacking awns; plants bearing multiple
fronds at a time; pantropical. . . . . . . . . . . . . . . . . . 4
4 Sporangial apertures not or subtly labiate; synangia
generally not sulcate along septa; venuloids lacking
between veins of terminal segments; terminal seg-
ments generally with a distinct suture at point of at-
tachment; Paleotropical. . . . . . . . . . . . . .IV. Ptisana
4 Sporangial apertures generally distinctly labiate;
synangia vertically sulcate along septa or cleft with
sporangia free above; venuloids present between
veins of terminal segments or lacking; terminal seg-
ments lacking a distinct suture at point of attachment;
pantropical. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
5 Sporangia fully fused into synangia (rarely partially
free) and vertically sulcate along septa, lacking an
apical patch of specialized cells; false veins lacking or
obscure; rhizomes erect; spores monolete; Neotropical
and Hawaii . . . . . . . . . . . . . . . . . . . . . . V. Marattia
5 Sporangia fused basally but free apically (rarely
partially fused), with a patch of specialized thick-
walled cells at the sporangial apex; false veins often
prominent, rarely lacking; rhizomes erect or creeping;
spores trilete; Paleotropical (occasionally naturalized
in the Neotropics) . . . . . . . . . . . . . . VI. Angiopteris
I. Danaea Sm. in Mém. Acad. Roy. Sci. Turin 5: 420, t.
9. 1793; nom. cons., non Danaa All. – Type: Danaea
nodosa (L.) Sm.,m. Acad. Roy. Sci. Turin 5: 420,
t. 9(11). 1793 Acrostichum nodosum L., Sp. Pl. 2:
1070. 1753.
= Heterodanaea C. Presl, Suppl. Tent. Pterid.: 38. 1845
– Type: Heterodanaea stenophylla (Kunze) C. Presl,
Suppl. Tent. Pterid.: 38. 1845 Danaea stenophylla
Kunze, Farrnkräuter 1(3): 55, t. 28. 1840.
Habit terrestrial or occasionally epiphytic, rhizomes
creeping, ascending, or erect. Sterile fronds blades 1-pin-
nate or simple, fronds with paired basal auricles; pulvini
present at nodes and often along stipes at leafless nodes;
stipes winged or not; margins simple, toothed, or irregu-
larly crisped; veins free, simple or bifurcating; sterile
fronds in some species proliferous near the apex. Fertile
fronds generally longer, narrower, and more erect than
sterile fronds, with smaller and thicker blades; densely
covered in sori, appearing almost acrostichoid in many
species. Indument: glandular hairs present on secondary
leaf veins (lacking elsewhere), non-glandular hairs scat-
tered, hairs simple or branched; scales peltate, flattened,
broad, irregular with lobed to fimbriate margins, raised
on ± centrally attached short stalks, mostly restricted to
costae, scale cells ± isodiametric (Fig. 2M). Idioblasts
absent. Sori elongate along veins, sporangia forming two
rows along veins, fully fused into synangia, each sporan-
gium dehiscing independently (but ± simultaneously) via
an apical pore nearest the inner axis. Spores monolete,
ellipsoid with a spinulose exospore, spines simple or oc-
casionally fused. (Fig. 2L–M).
Chromosome number. – 2n = 80, 160 (Walker, 1966;
Benko-Iseppon & Rodrigues da Fonseca, 2000). Triploids
possibly of hybrid origin have also been reported (Walker,
1966, 1973; Camus & Pérez García, 1995).
Distribution. – Widely distributed in the Neotropics
from Mexico to Brazil and the Caribbean.
Taxonomy. Danaea is a distinctive genus, restricted
to the Neotropics. Most species are strongly dimorphic,
and several species have erect or ascending rhizomes,
reminiscent of the arborescent architecture of some fossil
marattioids. Unlike other genera in Marattiaceae, Danaea
possesses conspicuous pulvini at nodes, with petiole bases
only slightly swollen, and has sporangia that dehisce via
742
TAXON 57 (3) • August 2008: 737–755Murdock • Taxonomic revision of Marattiaceae
Fig. 2. A, Ptisana smithii, deeply cut synangia with sporangia lacking labia; B, Ptisana howeana, sutures at base of pin-
nules, characteristic of Ptisana; C, abaxial leaf surface of Ptisana howeana, showing typical pinnule shape and vena-
tion of Ptisana; D, Eupodium laevis, stalked synangia characteristic of the genus; E, Eupodium kaulfussii, stalked radial
synangia; F, Eupodium laevis, abaxial leaf surface; G, Marattia alata synangia with labiate sporangia; HI, Marattia doug-
lasii, abaxial leaf surface, showing mature synangia and pulvini; J, Angiopteris chingii ( = Archangiopteris hokouensis),
parially free sporangia and elongate sorus; K, Angiopteris evecta, abaxial leaf surface with false veins; LM, Danaea el-
liptica, back-lit fertile frond showing apical sporangial dehiscence pore and sunken synangia; large peltate scales; NO,
Christensenia aesculifolia circular synangium; abaxial leaf surface showing reticulate venation, scattered synangia, and
distinctive whitish cast. Photos: H+I, John Game; O, Dennis Walker; all others by A. Murdock.
743
Murdock • Taxonomic revision of MarattiaceaeTAXON 57 (3) • August 2008: 737–755
an apical pore. The circumscription of the genus is not
controversial, but the unique morphology of this group
has at times prompted authors to recognize Danaea in a
separate family, Danaeaceae. However, because Maratti-
aceae has been proposed for conservation over the earlier
Danaeaceae (Murdock & al., 2006), when the two families
are combined, the name Marattiaceae must be used unless
this proposal is rejected in the future.
Species.Rolleri (2004) recognized 17 species, but
Christenhusz (in prep., pers. comm.) recognizes ~ 50 spe-
cies; there are ~ 60 currently published names. More detail
on the phylogeny of Danaea can be found in Christenhusz
& al. (2008).
II. Christensenia Maxon in Proc. Biol. Soc. Wash. 18:
239. 1905 Kaulfussia Blume, Enum. Pl. Javae 2:
260. 1828, non Dennst. in Fortsetz. Allg. Teutsch.
Gart.-Mag. 3(1): 30, 40. 1818, nec Nees, Horae Phys.
Berol.: 53, t. 11. 1820 – Type: Christensenia aescu-
lifolia (Blume) Maxon in Proc. Biol. Soc. Wash. 18:
239. 1905 Aspidium aesculifolium Blume, Enum. Pl.
Javae 2: 143. 1828 Kaulfussia aesculifolia (Blume)
Blume, Enum. Pl. Javae 2: 260. 1828.
= Macrostoma Griff. in Asiat. Res. 19(1): 110. 1836, pro
syn. Kaulfussia Blume.
Habit terrestrial, rhizomes creeping. Fronds: blades
palmately lobed to pedately compound; pulvini pres-
ent at bases of segments, absent on the stipes; margins
simple to broadly crenate; veins reticulate; stomata large,
raised, round, permanently open, densely distributed on
the abaxial leaf surfaces, visible to the naked eye. Indu-
ment: glandular hairs present, scattered uniseriate simple
hairs present, scales peltate, flattened, with lobed margins,
larger scales raised on short stalks, scale cells ± isodiamet-
ric and thick. Idioblasts absent. Sori ± in rows on either
side of leaf veins, often appearing scattered, sporangia
fully fused into ± circular sessile synangia, each spor-
nagium dehiscing independently but ± simultaneously
via a vertical slit on the inside of the synangium. Spores
monolete, ellipsoid, with long-spinulose exospores, spines
simple. (Fig. 2N–O).
Distribution. – Patchily distributed from northeast
India, Burma, China (Yunnan), Vietnam, Malaysia, In-
donesia, Philippines, Solomon Islands, Bismarck Archi-
pelago.
Chromosome number. 2n = 80, 160 (Braithwaite,
1977; Walker, 1979).
Taxonomy. – Pichi-Sermolli (1957) stated that Mac-
rostoma Griff. is a later homonym of Macrostoma Hed-
wig. However, Hedwig simply misspelled the earlier Mac-
rostema Pers., and clearly attributed the name to Persoon,
complete with page reference. Thus, Hedwig did not ef-
fectively publish of a new name, and Macrostoma Griff.
cannot be a later homonym (although it is confusingly
similar both to Macrostema Pers. and Macrostomium
Blume). However, Macrostoma Griff. was never validly
published (see Maxon, 1905; Pichi Sermolli, 1957).
Christensenia is the most unusual extant genus of
marattioid ferns because of its reticulate venation, pedate
blades, circular synangia, creeping rhizomes, and raised,
permanently open stomata. Because of these unique char-
acters, Christensenia has been placed by some authors
in its own family Christenseniaceae Ching. An earlier
name, Kaulfussiaceae Campbell, cannot be used because
it is based on an illegitimate generic name, Kaulfussia
Blume.
Species.Rolleri (1993), in the most recent treatment
of the genus, recognized two species and two varieties,
Christensenia aesculifolia (Blume) Maxon subsp. aescu-
lifolia, C. aesculifolia subsp. korthalsii (de Vriese) Rolleri,
and C. lobbiana (de Vriese) Rolleri. Christensenia is rare
in the wild and poorly represented in herbaria, and as such
it is difficult to determine whether the variation noted in
Rolleri’s revision is taxonomically informative or simply
a product of phenotypic plasticity.
III. Eupodium J. Sm., J. Bot. (Hooker) 4: 190. 1841
Type: Eupodium kaulfussii (J. Sm.) J. Sm. ex Hook.
Gen. Fil.: t. 118. 1842 Marattia kaulfussii J. Sm. in
Hook. Gen. Fil.: t. 26. 1839.
Habit terrestrial, rhizomes ± erect, each bearing only
one frond (rarely two). Fronds: blades 2- to 5-pinnate,
decreasingly divided distally; pulvini present at bases of
segments, smooth to tuberculate, absent on the stipes; pri-
mary pinnae opposite or subopposite, margins variously
toothed to deeply cut; veins free, generally branching;
adaxial vein surfaces bearing scattered awns concolorous
with veins, awns absent abaxially or sparse and smaller
than adaxial awns. Indument: glandular hairs absent, uni-
seriate simple or branching hairs occasional, scales peltate
but generally appearing basifixed due to extreme asym-
metry, larger scales raised on podia, scale cells elongate.
Idioblasts present, sparse. Sori borne on veins, ± medial;
sporangia fully fused into synangia (rarely partially free
above), synangia stalked, stalks to 3 mm long (Fig. 2DF);
synangial outer walls composed of elongate polygonal
cells creating a wrinkled or rugose appearance, generally
sulcate vertically along septa; paraphyses absent; synangia
bilateral and comprising two opposing rows of sporangia,
or radial; syanangia opening as a unit, shallowly cut each
with a central pad of tissue at apex of central axis, spo-
rangia dehiscing via a vertical slit on the inner surface
of each valve, apertures ± labiate; paraphyses lacking.
Spores monolete, ellipsoid, exospores spinulose, spines
simple. (Fig. 2DF).
While some species of Marattia s.str. have raised
or short-stalked synangia (e.g., Marattia cicutifolia), the
stalks in Eupodium are very prominent and vascularized.
744
TAXON 57 (3) • August 2008: 737–755Murdock • Taxonomic revision of Marattiaceae
Additionally, Eupodium is unique in the Marattiaceae in
having scattered awns along veins of pinnae (most promi-
nent near the ends of segments on adaxial side).
Distribution. – Widely distributed in the Neotropics
from Central America, the Caribbean, northwestern South
America, southern Brazil, and northeastern Argentina.
Chromosome number. – Unknown.
Species.Two species. Most recent treatments have
recognized only one morphologically variable species
under the name Marattia laevis Sm.; however molecular
and morphological investigations have revealed a disjunct
species in southern Brazil and northeastern Argentina,
recognized here as Eupodium kaulfussii (J. Sm.) J. Sm.
1. Eupodium laevis (Sm.) Murdock, comb. nov. Marat-
tia laevis Sm., Pl. Icon. Ined. 2: t. 47. 1790 Myrio-
theca laevis (Sm.) Poir. in Lam., Encycl. 4: 403. 1797
Discostegia laevis (Sm.) C. Presl in Abh. Böhm.
Ges. Wiss. 4: 272. 1845 Marattia alata Sw. var.
laevis (Sm.) Farw. in Amer. Midl. Naturalist 12: 308.
1931 – Type: St. Domingue (Hispaniola), Thiery 90
(holotype: LINN!).
= Marattia pittieri Maxon in Contr. U.S. Natl. Herb.
17: 421. 1914 – Type: Panama, Chiriquí, 1,625 m., 24
Mar 1911, Maxon 5704 (holotype: US!, photo BM!;
isotype: NY, P!).
Distribution. – Hispaniola, Cuba, Puerto Rico, Costa
Rica, Panama, Colombia, Ecuador, Peru, Bolivia, Ven-
ezuela. One collection has been found from Trinidad
(H. Prestoe, 1866, BM!), although this species was not
included in the flora of Proctor (1977).
This species is widespread and encompasses signifi-
cant morphological variation. Collections from Caribbean
islands tend to have tuberculate pulvini, numerous scales
at the base of the stipes, and moderately scaly blades.
Collections from Bolivia, Peru, Ecuador, and Venezula
tend to have smooth pulvini bearing long scales, stipe
bases that are glabrous to lightly scaly, and blades that
are generally densely scaly. Plants from Central America
fall somewhere in between these two forms, and there is
considerable variation within forms. Further investigation
may reveal other currently unrecognized taxa within this
species.
2. Eupodium kaulfussii (J. Sm.) J. Sm. in Hook., Gen.
Fil.: t. 118. 1842 Marattia kaulfussii J. Sm. in Hook.,
Gen. Fil.: t. 26. 1839 – Type: Brazil, 1816, Cunning-
ham s.n. (holotype: BM!).
= Eupodium kaulfussii (J. Sm.) J. Sm. var. macropteron
Fée, Crypt. Vasc. Brésil: 215. 1869 – Type: Brazil, Rio
de Janeiro, Glaziou 2273 (holotype: P!).
= Eupodium kaulfussii (J. Sm.) J. Sm. var. acuminatum
Fée, Crypt. Vasc. Brésil: 215. 1869 – Type: Brazil, Rio
de Janeiro, Glaziou 2274 (holotype: P!).
Distribution. – Southern Brazil (Rio de Janeiro, São
Paulo, Paraná, Santa Catarina, Rio Grande do Sul), north-
ern Argentina (Misiones).
This species is distinguished from the widespread
Eupodium laevis by presence of awns on the abaxial vein
surfaces (infrequent and smaller than those on the adax-
ial vein surfaces), scales ± absent from laminar surfaces,
scales sparse and small at bases of primary segments,
pulvini not tuberculate, synangia ranging from fully ra-
dial to bilateral (Fig. 2E) (see also Ponce & Smith, 1998),
and its generally smaller stature. The holotype has a mix-
ture of radial and bilateral synangia, smooth pulvini, and
both adaxial and abaxial awns. Despite the morphological
similarity between the two species of Eupodium, Murdock
(2008) found considerably more genetic distance between
these two forms than is typical of recognized species in
Marattiaceae, fully three times the number of base pair
differences than found between the highly morphologi-
cally distinct Marattia laxa of Mexico and Marattia doug-
lasii of Hawaii.
IV. Ptisana Murdock, gen. nov. – Type: Ptisana salicina
(Sm.) Murdock, comb. nov. Marattia salicina Sm.
in Rees, Cycl. 22. 1812 – Type: Norfolk Island (label
erroneously reads “New South Wales”, see Alston,
1936 and Pichi Sermolli, 1991), Molesworth s.n. (ho-
lotype: K!; isotype: LINN).
= Marattia sensu C. Presl, Suppl. Tent. Pterid.: 8. 1845,
non Sw., Prodr. 8: 128. 1788 – Type: Marattia fraxinea
Sm.
Filices terrestres; rhizomates erecta, omnino auricu-
lis veteribus obtecta; folia auriculis geminatis basalibus,
laminae bipinnatae usque ad quadripinnatae, segmen-
tis basaliter (cum pulvinis) tumidis; sporangia omnino
connata in synangiis, aperturis sine labiis; synangia
sessiles, profunde incisa, a squamis vel pilis uniseriatis
subtentis.
Ab Eupodio et Marattia aperturis sporangiorum sine
labiis, synangiis profunde incisis, parietibus externis tex-
turae colliculosae differt. Praeterea ab Eupodio synangiis
sessilibus a squamis vel pilis uniseriatis subtentis, multis
foliis laminisque aristis carentibus differt.
Habit terrestrial, rhizome ± erect, bearing multiple
fronds simultaneously. Fronds: blades 2- to 4-pinnate, ±
evenly divided throughout, but often with conspicuously
irregular division of some segments; pulvini present at
bases of segments, absent on the stipes; primary pinnae
generally opposite, terminal segments generally with a
prominent suture at point of attachment (occasionally
subtle or absent) (Fig. 2B), margins generally toothed,
occasionally crenulate or ± simple; veins free, simple or
bifurcating. Indument: glandular hairs absent, uniseriate
simple or branching hairs present, scales peltate but gener-
ally appearing basifixed due to extreme asymmetry, larger
745
Murdock • Taxonomic revision of MarattiaceaeTAXON 57 (3) • August 2008: 737–755
scales raised on podia, scale cells elongate. Idioblasts
present on abaxial laminae, solitary or lightly clustered,
of varying density; often present in chains along second-
ary veins (most easily visible in dried specimens). Sori
borne on veins, marginal to medial, sporangia fully fused
into synangia, synangia sessile (not raised or stalked);
synangial outer walls composed of small, isodiametric,
rounded cells creating a ± bead-like appearance, gener-
ally not sulcate along septa in mature synangia; synangia
subtended by nest-like rings of paraphyses or uniseriate
hairs of varying density; synangia bilateral, each com-
prising two opposing rows of sporangia; each synangium
opening as a unit, deeply cut with no central pad of tis-
sue; each sporangium dehiscing via a vertical slit on the
inner surface of each valve, apertures lacking labia or
slightly labiate. Spores monolete (rarely trilete or alete),
exospores granular to rugose; spinulose spores not re-
corded from this genus, but further investigation is re-
quired. (Fig. 2A–C).
Diagnosis.Differs from Eupodium and Marattia by
having sporangia without labiate apertures, synangia that
are deeply cut, valves generally lacking flaring lips, and
with a thickened beaded texture to outer walls of synangia.
Further differs from Eupodium by having sessile synangia
subtended by scales or uniseriate hairs, blades lacking
awns, and by bearing multiple fronds at one time.
Distribution. – Widely distributed in the Paleotropics
from Ascension Island to the Marquesas Islands, with the
highest diversity in New Guinea.
Chromosome number. – 2n = 78, 156 (Brownlie, 1961;
Ninan, 1956; Raj & Manickam, 1987; Manickam & Raj,
1988; Takamiya, 1995). One count of 2n = 80 (Manton,
1959) was published by Manton based on a count relayed
to her by T.G. Walker and lacks a validating photograph
or drawing; as all other counts for Ptisana are based on
x = 39, this count is possibly erroneous. Further investiga-
tion is required to determine if a base number of x = 39 is
a synapomorphy for the genus or any subdivision thereof.
There are no published chromosome counts for the puta-
tive sister genus Eupodium.
Etymology. – This generic name is derived from the
Latin feminine noun ptisana meaning pearl barley or a
beverage made from pearl barley. This word was in turn
derived from the Greek “πτισανη” with the same mean-
ing, a form of “πτισσειν” meaning to peel or crush. The
form of the synangium is reminiscent of pearl barley
(Fig. 2A–C). The English words “tisane” and the now
archaic “ptisan” are also similarly derived. In Latin this
word has also been spelled tisana or tipsana ; however the
chosen orthography ptisana seems most appropriate for a
pteridophyte genus.
Taxonomy. – Due to the complex history of the no-
menclature in Marattia s.l. and the numerous generic
names that have been applied to this group in the past,
proposing a new genus name is perhaps surprising and
merits explanation. Based on the arguments presented by
Pichi Sermolli (1957), this group of ferns would be called
Myriotheca Juss. However, the history of the name Myrio-
theca is more complicated than Pichi Sermolli realized.
The initial publication of Myriotheca in 1789 by Jussieu
(Gen. Pl.: 15, 447. 1789) included a description, but no
species were assigned to this genus. On page 447, in the
appendix to this work, Jussieu stated that Myriotheca was
congeneric with Caenopteris (currently recognized as a
synonym of Asplenium), rendering the name Myriotheca
invalidly published because it was not accepted by the au-
thor in the same publication and was placed in synonymy
with an older generic name. Poiret (in Lam., Encycl. 4:
403. 1798) assigned three species to Myriotheca : Marat-
tia alata, Marattia laevis, and Marattia fraxinea. Poiret
included the type of Marattia Sw., M. alata, in Myrio-
the ca, making Myriotheca superfluous upon publication,
and in any event a later homonym of Myriotheca Juss.
The generic name Celanthera Thouin, possibly referable
to this genus, was to appear in a paper entitledSur un
nouveau genre des Plantes” in Mémoires de l’Académie
des Sciences in 1786; however Thouins paper was miss-
ing from the volume and was never published (Pichi Ser-
molli, 1957). Eumarattia Hook. & Baker (Syn. Fil.: 440.
1868), published at an unspecified infrageneric rank is il-
legitimate due to the usage of the prefix “eu-” (see Vienna
Code, Art. 21.3, McNeill & al., 2006). The subgeneric
name Mesosorus Rosenst. (1908) cannot be elevated to
the generic rank as this would be a later homonym of
Mesosorus Hassk. Mesocarpus Rosenst., published later
in the same year, was superfluous when published, as it
was needlessly proposed as a replacement name for Meso-
sorus. In summary, there is no available name for this
group at generic rank, nor is there an available subgeneric
name for elevation to the rank of genus; thus a new name
is needed.
Species.Twenty species and three varieties are cur-
rently recognized. The most commonly grown “Marattia
species in botanical gardens (e.g., M. salicina, M. attenu-
ata, M. oreades) are all in this new genus, and as a result
much of the scientific literature on Marattia is now refer-
able to Ptisana.
Species are listed below in order of date of publica-
tion. Combinations are made here for the best-defined
species that are most likely to withstand future taxo-
nomic revisions. Several of these species will likely be
placed in synonymy in the future, some will be removed
from synonymy upon closer inspection, and there are
an unknown number of species yet to be described. The
African Ptisana fraxinea complex, most fully reviewed
by Pichi Sermolli (1969), and the Malesian P. melane-
sica complex are the most in need of future revisionary
work.
746
TAXON 57 (3) • August 2008: 737–755Murdock • Taxonomic revision of Marattiaceae
1. Ptisana fraxinea (Sm.) Murdock comb. nov. Mar-
attia fraxinea Sm., Pl. Icon. Ined. 2: t. 48. 1790
Myriotheca fraxinea (Sm.) Poir. in Lam., Encycl. 4:
404. 1798 (as “frascinea”) – Type: Mauritius (“Île de
France), Thouin 91 (holotype: LINN, photo at BM!;
possible isotype: G).
a. Ptisana fraxinea (Sm.) Murdock var. fraxinea
= Myriotheca fraxinifolia Bory, Voy. Îles Afrique 1:
267. 1804 – Type: Réunion, Bory s.n. (holotype: P!;
isotype FI).
= Myriotheca sorbifolia Bory, Voy. Îles Afrique 1: 267.
1804 Marattia sorbifolia (Bory) Sw., Syn. Fil.: 168.
1806 – Type: Réunion, Bory s.n. (holotype: P!; iso-
types: B!, FI).
= Marattia oppositifolia Sm. in Rees, Cycl. 22. 1812 –
Type: Location and collector unknown (Herb. Smith
1644.4) (holotype: LINN).
= Marattia macrophylla de Vriese, Monogr. Marattiac.:
3, t. 3. 1853 – Type: Mauritius, Bojer s.n. (lectotype:
K!; lectotype designated by Pichi Sermolli, 1969).
= Marattia calliodous de Vriese, Monogr, Maratiac.: 6, t.
3, 4. 1853 – Type: Ceylon, Gardner s.n. (holotype: K!).
= Marattia cuneiformis de Vriese, Monogr, Maratiac.:
7, t. 3, 4. 1853 – Type: Réunion [“Île Bourbon”], col-
lector unknown (possible holotype: K!).
= Marattia boivini Mett. ex Kuhn, Filic. Afr.: 174. 1868
– Type: Madagascar, Boivin s.n. (possible holotype:
L!; istotype: P!).
= Marattia microcarpa Mett. ex Kuhn, Filic. Afr.: 175.
1868 – Type: Nosy-Be [Nossibeh], 1846–1848, Boivin
s.n. (lectotype: W, fragment and photo BM!; lectotype
designated by Pichi Sermolli, 1969).
= Marattia arguta Mett. ex Kuhn in Linnaea 36: 168.
1869 – Type: Location and collector unknown (ho-
lotype: B, destroyed). Note: this is likely a synonym
of M. fraxinea, but as the type has been destroyed
(Pichi Sermolli, 1969), this cannot be definitively
ascertained.
= Marattia odontosora Christ in J. Bot. (Morot.) 22 [Ser.
2, Vol. 2]: 19. 1909 – Type: Guinea, Tolo, 23 Dec 1905,
Caille 15629 (lectotype: P!, lectotype designated by
Pichi Sermolli, 1969).
Note. – In the extremes of their ranges, several forms
of P. fraxinea might be recognized as independent spe-
cies, however a range of intermediates exists between all
varieties and circumscription is problematic.
Distribution. – Africa, Madagascar and other islands
of the Indian Ocean, and continental India. While spe-
cies from as far away as the Marquesas Islands have been
called Marattia fraxinea, species from Asia and Oceania
are referable to other species of Ptisana.
b. Ptisana fraxinea (Sm.) Murdock var. salicifolia (Schrad.)
Murdock, comb. nov. Marattia salicifolia Schrad. in
Gött. Gel. Anz. 1818: 920. 1818 Marattia fraxinea
Sm. var. salicifolia (Schrad.) C. Chr. in Perrier, Cat. Pl.
Madag., Pterid.: 67. 1932 – Type: South Africa, Cape
of Good Hope, Hesse s.n. (holotype: LE).
= Marattia salicifolia Schrad. var. natalensis Kunze,
Farrnkräuter 1: 79. 1842 Marattia natalensis
(Kunze) C. Presl, Suppl. Tent. Pterid.: 9. 1845 Type:
South Africa, Natal, Mauritzburg, Krauss 257 (holo-
type: LZ, destroyed; isotypes: BM!, FI, K!).
= Marattia dregeana C. Presl, Suppl. Tent. Pterid.:
9. 1845 – Type: South Africa, Cape of Good Hope,
Drège s.n. (holotype: PRC ?; isotypes: BM!, K!, W).
Distribution. – Southern Africa and Madagascar.
Note. – Forms from the Cape region are very distinct,
but more northern populations grade into more typical
P. fraxinea.
c. Ptisana fraxinea (Sm.) Murdock var. robusta (Alston)
Murdock, comb. nov. Marattia robusta Alston in
Exell, Suppl. Cat. Vasc. Pl. S. Tomé: 8. 1956 – Type:
São Tomé, Monte Café, 1,200 m., Espírito Santo 191
(holotype: BM!; isotypes: COI!, LISC!).
Distribution. – São Tomé.
2. Ptisana salicina (Sm.) Murdock, comb. nov. Marattia
salicina Sm. in Rees, Cycl. 22. 1812 – Type: Norfolk
Island (label erroneously reads “New South Wales”),
Molesworth s.n. (holotype: K!; isotype: LINN).
= Marattia elegans Endl., Prodr. Fl. Norfolk.: 6. 1833
– Type: Norfolk Island (holotype: W, photo BM!; pos-
sible isotype: K!).
= Marattia stokesii E.D. Br. in Bull. Bernice P. Bishop
Mus. 89: 103, f. 18. 1931 – Type: Rapa, 10 Oct 1921,
Stokes 311-A–F (holotype: BISH!; isotypes: UC!,
US!).
= Marattia cincta Copel. in Bull. Bernice P. Bishop
Mus. 93: 6, t. 1. 1932 – Type: Society Islands, Raiatea,
Puorotepo, 1,610 m., 29 Jan 1931, Grant 5203 (holo-
type: BISH!; isotypes: MICH!, UC!).
= Marattia grantii Copel. in Bull. Bernice P. Bishop
Mus. 93: 6, t. 1. 1932 – Type: Society Islands, Tahaa,
Ohiri, 1515 ft., 25 Jan 1931, Grant 5172 (holotype:
BISH!; isotypes: MICH!, UC!).
Distribution. – Norfolk Island, New Zealand, New
Caledonia, Cook Islands, Austral Islands, Society Islands,
Marquesas Islands. Some collections from the Eungella
range of Australia may be this species.
Note. – This species is retained in a broad sense here,
but likely encompasses several distinct taxa, and is closely
related to P. smithii.
3. Ptisana attenuata (Labill.) Murdock, comb. nov.
Marattia attenuata Labill., Sert. Austro-Caledon.: 9,
t. 13, 14. 1824 Marattia acuminata Willd. ex Kaulf,
747
Murdock • Taxonomic revision of MarattiaceaeTAXON 57 (3) • August 2008: 737–755
Enum. Filic.: 33, t. 1, f. 2. 1824 – Type: New Caledonia,
Labillardiere s.n. (holotype: P!; isotype: B!, BM!).
= Marattia rolandi-principis Rosenst. in Repert. Spec.
Nov. Regni. Veg. 10: 162. 1911 – Type: New Caledo-
nia, Plateau de Dogny, La Toa, 900 m, 2 Feb 1911,
Franc 1441 (holotype: P!; isotypes: B!, BISH!, BM!,
K!, L!, UC!, US!); possible isotypes [Franc 121 &
121a, same date and location]: BM!, GH!, K!, L!, NY!,
P!, PR!).
= Marattia smithii Mett. ex Kuhn forma soluta Comp-
ton in J. Linn. Soc., Bot. 45: 456. 1922 – Type: New
Caledonia, Ignambi, 1,500 ft., 4 Aug 1914, Compton
1674 (holotype: BM!).
= Marattia cooperi T. Moore in Gartenfl. 13: 154. 1864
– Type: New Caledonia, ex hort Veitch (possible ho-
lotype: K!); known only from cultivation.
Distribution. – New Caledonia.
Note. – Ptisana attenuata is treated here in a broad
sense, and likely encompasses several distinct taxa. Field
study of Ptisana in New Caledonia will be required to
clarify the taxonomy of this group. Soil chemistry and
altitude likely play a role in the variability of this species.
M. rolandi-principis, only known from two collections,
appears to be a juvenile but precociously fertile form of
the typical M. attenuata.
4. Ptisana sambucina (Blume) Murdock, comb. nov.
Marattia sambucina Blume, Enum. Pl. Javae: 256. 1828
– Type: Java, Blume s.n. (holotype: L!, photo BM!).
= Marattia brooksii Copel. in Philipp. J. Sci., C, 7: 59.
1912 – Type: Sarawak, Mt. Poe, 900 m, Brooks 138 (ho-
lotype: MICH!; isotype: BM!; possible isotype: P!).
= Marattia rugulosa Alderw. in Bull. Jard. Bot. Buiten-
zorg, sér. 2, 23: 17. 1916 – Type: Borneo, Amai Ambit,
Hallier 3219 (holotype: BO; isotypes: BM!, K!, L!).
Distribution. – Malesia, Vietnam.
Note. – Closely related to the P. melanesica complex
and to P. sylvatica.
5. Ptisana sylvatica (Blume) Murdock, comb. nov.
Marattia sylvatica Blume, Enum. Pl. Javae: 256. 1828
(often spelled “silvatica”) – Type: Java, Salak Moun-
tains, Blume s.n. (holotype: L!, photo BM!).
= Marattia caudata Copel. in Philipp. J. Sci., C. 9: 227.
1914 – Type: Sumatra, Lebong Tandai, Benkoelen,
Jun 1912, Brooks 41 (holotype: MICH!; isotypes:
BM!, P!).
= Marattia caudiformis Alderw. in Bull. Jard. Bot. Buit-
enzorg, sér. 2, 28: 33. 1918 – Type: Sumatra, Lebong
Simpang, Benkoelen, Aug 1917, Brooks 309.S (holo-
type: not located; isotype: BM!).
Distribution. – Malesia.
Note. – Closely related to the P. melanesica complex
and to P. sambucina.
6. Ptisana pellucida (C. Presl) Murdock, comb. nov.
Marattia pellucida C. Presl, Suppl. Tent. Pterid.: 10.
1845 – Type: Philippines, Luzon, Camarines Prov-
ince, Cuming 177 (possible holotype: L!; isotypes:
BM!, P!, US!).
= Marattia haenkeana C. Presl, Suppl. Tent. Pterid.: 11.
1845 – Type: Philippines, Luzon, Haenke s.n. (holo-
type: PR or PRC?). Note: relevant type material not
located, synonymy follows Copeland (1958).
= Marattia vestita Christ in Philipp. J. Sci., C, 2: 185.
1907 – Type: Philippines, Mindanao, Davao District,
Mt. Apo, Apr 1904, Copeland 1179 (holotype: MICH!;
isotype: US!).
Distribution. – Philippines, Malaysia.
7. Ptisana mertensiana (C. Presl) Murdock, comb. nov.
Gymnotheca mertensiana C. Presl, Epimel. Bot.: 2, t.
1. 1849 Marattia mertensiana (C. Presl) C. Chr., In-
dex Filic.: 415. 1906 – Type: Caroline Islands, Mertens
s.n. (possible holotype: W; isotype: BM!).
Distribution. – Micronesia.
8. Ptisana ternatea (de Vriese) Murdock, comb. nov.
Marattia ternatea de Vriese, Monogr. Marattiac.:
4. 1853 – Type: Ternate, “sylvis montis,” Aug 1821,
Reinwardt s.n. (holotype: L!, photo, BM!; possible
isotype: BM!).
= Marattia andaiensis Alderw. in Bull. Jard. Bot.
Buitenzorg, sér. 2, 23: 16. 1916 – Type: New Guinea,
Andai, Teijsmann s.n. (holotype: BO, photo BM!; iso-
type: BM!)
= Marattia papuana Alderw. in Bull. Jard. Bot. Buiten-
zorg, sér. 2, 23: 17. 1916 – Type: New Guinea, Con-
stantinhafen, Hollrung 613 (holotype: BO, photo BM!;
isotype: BM!)
= Marattia teysmanniana Alderw. in Bull. Jard. Bot.
Buitenzorg, sér. 2, 23: 18. 1916 – Type: Ternate, Tei-
jsmann 5208 HB (holotype: BO, photo BM!; isotypes:
BM!, U!)
Distribution. – Moluccas, New Guinea, Philippines.
9. Ptisana purpurascens (de Vriese) Murdock, comb.
nov. Marattia purpurascens de Vriese, Monogr.
Marattiac.: 7, t. 3, 4. 1853 Marattia fraxinea Sm. var.
purpurascens (de Vriese) Baker, Cat. ferns: 18. 1868 –
Type: Ascension Island, 1843, Hooker s.n. (lectotype:
K!, lectotype designated by Pichi Sermolli, 1969).
= Marattia elegans Endl. var. erecta Houlston & T.
Moore in Gard. Mag. Bot. 3: 146. 1851 – Type: In hort.,
from Ascension Island, (holotype: not located).
= Marattia ascensionis J. Sm., Cult. Ferns: 79. 1857
– Type: In hort. Kew, from Ascension Island, Wren
s.n. (lectotype: BM!; lectotype designated by Pichi
Sermolli, 1969).
748
TAXON 57 (3) • August 2008: 737–755Murdock • Taxonomic revision of Marattiaceae
= Marattia ascensionis J. Sm. var. cristata J. Sm., Cult.
Ferns: 79. 1857 – Type: In hort. Kew, from Ascension
Island, Wren s.n. (lectotype: BM!; lectotype desig-
nated by Pichi Sermolli, 1969).
Distribution. – Ascension Island.
Note. Closely related to P. fraxinea.
10. Ptisana smithii (Mett. ex Kuhn) Murdock, comb.
nov. Marattia smithii Mett. ex Kuhn in Verh. K.K.
Zool.-Bot. Ges. Wien 19: 584. 1869 – Lectotype desig-
nated here (chosen among isosyntypes, type material
at B destroyed): Aneiteum, New Hebrides, Dec 1858,
Herus 5 (lectotype: P!; isolectotype: GH!).
Distribution. – Vanuatu, Fiji, Samoa, Tonga.
Note. – Closely related to P. salicina.
11. Ptisana melanesica (Kuhn) Murdock, comb. nov.
Marattia melanesica Kuhn in Engler, Forschungsr.
Gazelle 4: 15, t. 3. 1889 – Type: New Hanover [Papua
New Guinea], “partes interiors insulae in silvis mon-
tanis altit. 200 m,” 24 Jul 1875, Naumann s.n. (holo-
type: B, destroyed).
= Marattia burkei Baker in Gard. Chron., ser. III, 22:
425, f. 129. 1897 – Type: “Hort. Veitch, Burke s.n.
(illustration: Gard. Chron., ser. III, 22: f. 129 (p. 435).
1897.; possible type: K!, “Hort. Veitch. 11/97. Collected
by the late Daniel Burke.”). Baker, in the protologue,
surmised that this species might have been collected
in Colombia. However, the illustration in the original
description, and a possible type specimen at Kew are
referable to P. melanesica ; the distinctive sutures of
Ptisana are apparent in both the illustration and speci-
men. Burke also collected in the Moluccas and New
Guinea on three separate occasions between 1854 and
his death in 1897 in Amboina (Veitch, 1906); he likely
collected this fern on one of these occasions.
= Marattia werneri Rosenst. in Repert. Spec. Nov.
Regni. Veg. 5: 44. 1908 – Type: New Guinea, Gelu
mountains, 1,700 m, Jul 1907, Werner 78 (specimens
labeled 49 also distributed as isotypes) (holotype: L!;
isotypes: BISH!, BM!, GH!, NY!, US!, UC!).
= Marattia kingii Copel. in Philipp. J. Sci., C, 6: 66.
1911 – Type: New Guinea, Goodenough Bay, 1,200 m,
King 214 (holotype: MICH!).
= Marattia tafaensis C. Chr. in Brittonia 2: 267, t. 1.
1937 – Type: New Guinea, Mt. Tafa, 2,100 m, 26 May
1933, Brass 4064 (holotype: GH!; isotypes: BM!,
MICH!, NY!, UC!).
= Marattia coronata Copel. in Philipp. J. Sci., C, 76: 23.
1941 – Type: Dutch New Guinea (Irian Jaya), Iden-
burg River, 1,800 m, Brass 12116 (holotype: MICH!;
isotypes: BM!, GH!, L!, UC!).
Distribution. – New Guinea and surrounding island
groups.
Note. – This group is distinctive within the genus for
having very finely divided pinnules with few but relatively
large synangia, with some collections having synangia
larger than their associated pinnules. While there are likely
several distinct taxa subsumed under this name, segregate
species as described do not appear distinct; some “spe-
cies” can be found within the same collection or even on
the same frond. Of the subsumed species, M. werneri and
M. tafaensis are the most distinctive and may warrant
taxonomic recognition, however more study is needed.
Closely related to P. costulisora with intermediates often
found, and part of a larger group of species related to
P. sambucina.
12. Ptisana obesa (Christ) Murdock, comb. nov. Marat-
tia obesa Christ in Nova Guinea 8, Bot. 1: 163. 1909
Type: New Guinea, “Noord-rivier près de Alkmaar”,
7 Sep 1907, Versteeg 1729 (holotype: L!; isotypes:
BM!, BO, K!, U!, US!).
= Marattia megaptera Copel. in Philipp. J. Sci., C, 76:
24. 1941 – Type: Dutch New Guinea (Irian Jaya), Iden-
burg River, 750 m., Brass 13668 (holotype: MICH!;
isotypes: GH!, L!, UC!).
Distribution. – New Guinea.
Note. – Questionably distinct from the Australian
P. oreades. Marattia megaptera appears to be a particu-
larly large-pinnuled specimen of this species.
13. Ptisana squamosa (Christ) Murdock, comb. nov.
Marattia squamosa Christ in Nova Guinea 8, Bot. 1:
163. 1909 – Type: New Guinea, “Resirug près Noord-
rivier à 350 m.,” 28 Aug 1907, Versteeg 1681 (holo-
type: K!; isotypes: BM!; U!).
Distribution. – New Guinea.
14. Ptisana grandifolia (Copel.) Murdock, comb. nov.
Marattia grandifolia Copel. in Philipp. J. Sci., C, 6:
66. 1911 – Type: New Guinea, King 305 (holotype:
MICH!; King 451, BM!, MICH!, appears to be identi-
cal to the type material).
Distribution. – New Guinea.
15. Ptisana novoguineensis (Rosenst.) Murdock, comb.
nov. Marattia novoguineensis Rosenst. in Repert.
Spec. Nov. Regni. Veg. 10: 342. 1912 – Type: New
Guinea, Kaiserwilhelmland, Logaueng, 300 m., Mar
1909, Bamler “L 48 [88] (holotype: P!; isotypes:
BM!, L!, NY!, PR!, UC!).
Distribution. – New Guinea.
16. Ptisana oreades (Domin) Murdock, comb. nov.
Marattia oreades Domin in Biblioth. Bot. 85: 219
220, t. 5 [caption to figure misspelledoreodoxa”].
1913 – Type: Australia, Queensland, Bellenden Ker,
749
Murdock • Taxonomic revision of MarattiaceaeTAXON 57 (3) • August 2008: 737–755
900–1,000 m, Jul 1909, Domin 540 (lectotype: PR,
designated by (and photo of type in) Pichi Sermolli,
1991).
Distribution. – Tropical Australia.
Note. – Questionably distinct from P. obesa from New
Guinea.
17. Ptisana howeana (W.R.B. Oliver) Murdock, comb.
nov. Marattia fraxinea Sm. var. howeana W.R.B.
Oliver in Trans. & Proc. New Zealand Inst. 49: 125.
1917. Marattia howeana (W.R.B. Oliver) P.S. Green
in Kew Bull. 43: 655. 1988 – Type: Lord Howe Is-
land, Deep Creek, 15 Nov 1913, Oliver s.n. (holotype:
WELT!, photo BM!; isotype: P!).
Distribution. – Lord Howe Island.
18. Ptisana rigida (Alderw.) Murdock, comb. nov. Mar-
attia rigida Alderw. in Nova Guinea 14: 32. 1924 –
Type: New Guinea, Doorman Summit, 1420 m., 7
Oct 1920, Lam 1552 (holotype: BO; isotypes: BM!,
K!, L!, U!, US!).
= Marattia koordersii Alderw., Malayan Ferns Fern
Allies, Suppl. 1: 529. 1917 – Type: Celebes, S.H.
Koorders 16982B (holotype: BO; isotype: BM!).
Distribution. – New Guinea, Sulawesi.
19. Ptisana platybasis (Copel.) Murdock, comb. nov.
Marattia platybasis Copel. in J. Arnold Arbor. 10:
174. 1929 – Type: New Guinea, Ihu, Vailala River,
19 Feb 1926, Brass 1005 (holotype: UC!; isotypes:
GH!, MICH!).
Distribution. – New Guinea.
20. Ptisana costulisora (Alston) Murdock, comb. nov.
Marattia costulisora Alston in J. Bot. 78: 228. 1940
Type: New Guinea, Morobe District, Mt. Sarawaket,
8,000–9,000 ft., 31 Mar 1937, Clemens 6157 (holotype
B!, photo BM!; isotypes: BM!, L!).
= Marattia brassii Copel. in Philipp. J. Sci., C, 76: 23.
1941 – Type: Dutch New Guinea (Irian Jaya), Iden-
burg River, 1,800 m, Brass 12117 (holoype: GH!, iso-
types: BM!, L!, UC!).
Distribution. – New Guinea and surrounding island
groups.
Note. – The type of M. brassii Copel. has submar-
ginal sori rather than the nearly medial sori of the type
of P. costulisora, however all intermediate states can be
found in other collections. Closely related to P. melanesica
complex with some plants having irregularly divided 2- to
3-pinnate fronds.
Species incertae sedis: These species are referable to
Ptisana, but type material has not been obtained
for study and no new combinations are made here:
Marattia dolichocarpa Alderw., Marattia paleolata
Alderw., Marattia boninensis Nakai, and Marattia
tuyamae Nakai.
V. Marattia Sw., Prodr. 8.: 128. 1788 Myriotheca sensu
Poir. in Lam., Encycl. 4: 403. 1798, non Juss., Gen.
Pl.: 15, 447. 1789 [ Caenopteris Berg.] Discostegia
C. Presl, Suppl. Tent. Pterid.: 11. 1845 – Type: Marat-
tia alata Sw., Prodr. 8: 128. 1788. Myriotheca sensu
Poir. is automatically typified on Marattia alata Sw.,
because Poiret included this, the type of Marattia Sw.,
in Myriotheca. Discostegia is also automatically typi-
fied on Marattia alata Sw., because Presl included
it in Discostegia (as well as the type of Eupodium
J. Sm.).
= Gymnotheca C. Presl, non Decne., Suppl. T