Abstract
The Podostemaceae are highly enigmatic plants which are restricted to submerged river-rock habitats. The availability of new material of nine taxa from continental Africa prompted this new study. Five species belonging to the genera Dicraeanthus, Leiothylax, Letestuella, Macropodiella, and Stonesia and another four species of the large genera Inversodicraea sensu stricto and Ledermanniella sensu stricto have been analysed for the first time. New anatomical and developmental data are described and illustrated by use of microtome sections and scanning electron microscopy. In parallel, phylogenetic analyses of all available sequence data of African Podostemaceae have been conducted using three plastid markers (matK, trnD-trnT, rpoB-trnC). Inversodicraea cf. bosii appears basal within the continental African clade. The remaining taxa are distributed in three, rather poorly supported, major clades which are consistent with their morphology: (1) the Inversodicraea clade is characterised by stem scales and contains members of the former Ledermanniella subg. Phyllosoma with either pollen-monads or dyads; (2) the Ledermanniella-Monad group consisting of Leiothylax, Letestuella, Macropodiella, Stonesia, and Ledermanniella species—all taxa without stem scales but showing pollen as single grains, with Monandriella linearifolia being basal to this clade; (3) the Ledermanniella-Dyad clade including Djinga, Dicraeanthus, and Ledermanniella species without stem scales but with pollen dyads. To reduce the polyphyly of Ledermanniella sensu lato (i.e. sensu C. Cusset) we propose restricting Ledermanniella to the species of the former subgenus Ledermanniella, resurrecting Monandriella as monotypic genus, and accepting the genus name Inversodicraea for members of Ledermanniella subg. Phyllosoma.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
Introduction
The Podostemaceae (“river-weeds”, including Tristichaceae of some authors; Malpighiales) are the largest family of strictly aquatic angiosperms (Cook and Rutishauser 2007). Molecular analyses indicate that Podostemaceae belong to the Malpighiales clade in the eurosids I group (Gustafsson et al. 2002; APG II 2003; Korotkova et al. 2009; Wurdack and Davis 2009) and might date back to the Campanian, about 76 mya (Davis et al. 2005). Three subfamilies can be recognised: Tristichoideae and Weddellinoideae as the two smaller podostemaceous subfamilies, and as third subfamily the large and morphologically distinct Podostemoideae with 42 genera and ca. 270 spp. (Engler 1928; Kita and Kato 2001; Cook and Rutishauser 2007).
In the past few years there has been a study focus on African, including Madagascan, Podostemoideae (Moline et al. 2007; Kita et al. 2008). They are classified in 16 genera and ca. 82 species. Most of these inconspicuous water plants are restricted to a single river or waterfall (Grob et al. 2007; Ameka et al. 2009). Many of the genera and species in Africa occur in Cameroon and Gabon (Engler 1928; Cusset 1987; Cook and Rutishauser 2007; Ghogue et al. 2009).
The genera present in this region are, among others, Dicraeanthus, Djinga, Inversodicraea, Ledermanniella, Leiothylax, Letestuella, Macropodiella, Monandriella, and Stonesia. Some of them, however, expand their range outside of Cameroon and Gabon (Table 1). Identification keys, drawings and short descriptions of all African Podostemaceae are provided in Rutishauser et al. (2007), in addition to photographs of selected specimens. Molecular data of African Podostemaceae have been analysed by, e.g., Kita et al. (2005), Moline et al. (2007), and Pfeifer et al. (2009).
Moline et al. (2007) used matK sequence and morphological data for eight African Podostemaceae species of the genera Dicraeanthus, Djinga, and Ledermanniella sensu lato (incl. Inversodicraea). Accordingly, all podostemoids studied from continental Africa form a clade that is sister to the Madagascan genera Endocaulos and Thelethylax. Moline et al. (2007) have already tentatively suggested revisions of the generic delimitations in the African taxa. For example, Ledermanniella sensu lato (incl. Inversodicraea and Monandriella), as described by Cusset (1974, 1984a, b), the largest podostemoid genus in Africa with ca. 46 spp., turned out to be an artificial and somewhat awkward genus. Its two subgenera Ledermanniella and Phyllosoma (as proposed by Cusset 1984a, b) were distinguishable by the absence versus presence of stem scales in addition to the usually compound foliage leaves (Moline et al. 2007; Rutishauser et al. 2007). Before Cusset’s revision (1974, 1984a, b) most Ledermanniella spp. belonged to the genus Inversodicraea Engler ex R.E. Fries. Among the species known before Cusset and included in this study (Tables 1, 4), Ledermanniella bifurcata, L. bowlingii, L. cristata, L. ledermannii, and L. pusilla were members of Inversodicraea, whereas Ledermanniella (with L. linearifolia), and Monandriella (with M. linearifolia, syn. Ledermanniella monandra) were monotypic genera.
The availability of new material of these rarely collected plants prompted our new analysis. Since the study by Moline et al. (2007), material of five additional species from small genera such as Dicraeanthus zehnderi, Leiothylax quangensis, Letestuella tisserantii, Macropodiella heteromorpha, and Stonesia ghoguei became available to us and has been analysed here. Four species from the large genus Ledermanniella sensu lato (incl. Inversodicraea, Monandriella) were also added: Inversodicraea cf. annithomae and I. cristata, Ledermanniella pusilla, and Monandriella linearifolia (syn. L. monandra). Moreover, we extended the number of chloroplast molecular markers, by sequencing matK, trnD-trnT, and rpoB-trnC for most of the taxa.
The Podostemaceae show many morphological peculiarities making this family a worthwhile, but often problematic study object. Some features may be seen as adaptations to seasonally submerged river rocks with flowering in air possible for only a short period after the rainy period. Many Podostemaceae have a flattened photosynthetic body which adheres to a hard substrate. It has been called a “thallus” because the conventional demarcation into stem, leaf, and root is often not obvious. Various botanists (Schnell 1998; Sehgal et al. 2002, 2007) regard this vegetative body as a unique architectural type. In contrast, we here adopt the classical root–shoot model (CRS model) with its structural categories roots, shoots (including stems and leaves), as already proposed by, e.g., Warming (1881, 1891, 1899), Jäger-Zürn (2000), and Cook and Rutishauser (2007). The term “stem” is applied to a cylindrical or slightly flattened photosynthetic body that develops exogenous leaves. We use the term “root” for a slightly or strongly flattened photosynthetic body when adhesive hairs and endogenous shoot buds are developed but no exogenous leaves.
The objectives of this study were:
-
1
To perform a molecular phylogenetic analysis of African podostemoids and to include morphological data across nine genera and 17 species.
-
2
To describe morphological idiosyncrasies of African Podostemoideae and to plot them on to the molecular tree in order to describe their evolutionary history. Special emphasis will be given to the infraspecific variability of stem scale morphology because this feature is traditionally taken as very important for species delimitation in Inversodicraea (syn. Ledermanniella subg. Phyllosoma). To better visualise morphological differences, members of four genera (Dicraeanthus, Inversodicraea, Leiothylax, and Letestuella) will be described and illustrated using microtome sections and scanning electron microscopy.
-
3
To solve the mismatch at the generic rank in African Podostemaceae–Podostemoideae.
Methods
Taxon sampling
The objective of the sampling was to include representatives of important groups of African Podostemaceae to infer their infrageneric relationships. Information on the selected taxa is given in Table 1: Dicraeanthus africanus, D. zehnderi, Djinga felicis, Inversodicraea cf. annithomae, I. cf. bosii, I. cristata, I. ledermannii, Ledermanniella bifurcata, L. bowlingii, L. letouzeyi, L. linearifolia, L. pusilla, Monandriella linearifolia, Leiothylax quangensis, Letestuella tisserantii, Macropodiella heteromorpha, and Stonesia ghoguei. The monophyly of this African group has been tested by including matK data of all Podostemaceae taxa as used by Moline et al. (2007). Accordingly, the African clade received a bootstrap support of 91% (not shown). Because of availability, Podostemum ceratophyllum from Eastern USA was sequenced as outgroup for all three markers (Philbrick and Novelo 2004). Herbarium specimens, silica-gel dried samples (stems and/or roots) for molecular studies, and ethanol (70%)-fixed samples for morphological studies were collected for each species. Vouchers and wet plant material are housed in the National Herbarium of Cameroon (YA), the Herbarium of Zurich Universities (Z/ZT), and (for L. bowlingii) also the Ghana Herbarium (GC). Identification of taxa was not always easy. Two species of Inversodicraea could not be identified without doubt: Inversodicraea cf. bosii (closely resembling I. ledermannii) and Inversodicraea cf. annithomae (perhaps conspecific with Ledermanniella ntemensis Y.Kita, Koi, Rutish and M.Kato). We included from three taxa (Dicraeanthus zehnderi, Djinga felicis, and Inversodicraea ledermannii) two somewhat polymorphic accessions each, and from Inversodicraea cristata three accessions that finally turned out to belong to the same species (for D. felicis and I. ledermannii see Moline et al. 2007).
We sequenced three chloroplast markers for most of the African Podostemaceae in our analysis. Despite intensive, repeated efforts, unsuccessful amplifications led to a high proportion of 18.8% of missing data in the matrices (rpoB: 22/24, matK: 19/24, trnL intron: 15/24 sequences available; Table 1).
Laboratory
For scanning electron microscopy the dissected plant parts were critical-point dried and sputter coated with gold. The electron micrographs were taken with a Jeol scanning electron microscope at 20 kV. For microtome sections, specimens were embedded in Kulzer’s Technovit (2-hydroxyethyl methacrylate), and sectioned with a Microm HM 355 rotary microtome and conventional microtome knife types C and D. The mostly 7 μm thick sections were stained with ruthenium red and toluidine blue.
DNA extractions, amplifications, and cycle sequencing followed the procedures described in Moline et al. (2007). The primers used were: trnL intron: trnLUAAF (TabC) and trnLUAAR (TabD; Taberlet et al. 1991); rpoB: rpoB and trnCGCAR (Shaw et al. 2005), matK: Pod2Rm, Pod5Fm, Pod6Rm, Pod8Fm, Pod41Fm, PodAFm, and Podnew1Rm (Moline et al. 2007). EMBL accessions are given in Table 1.
Data analyses
The data matrix included 24 terminal taxa comprising 17 African species and P. ceratophyllum as outgroup, which is closely related to the African, Madagascan, and Asian members of Podostemoideae (Kita and Kato 2001; Moline et al. 2007). Sequences were aligned using ClustalX (Higgins and Sharp 1988) with the default settings and then manually adjusted in MacClade version 4.05 (Maddison and Maddison 2002). The aligned data set is available as online supplement. There are no conflicting nodes in the analyses of the individual data sets (not shown) with bootstrap values above 70% (Mason-Gamer and Kellogg 1996). Wiens (2006) basically recommended the inclusion of missing data rather than excluding taxa or genes from phylogenetic analyses. Therefore, we followed the total evidence approach and combined the three data sets into a single large one using a super matrix approach (Sanderson et al. 1998; Bininda-Emonds 2004; de Queiroz and Gatesy 2007). The manageable number of taxa enabled a maximum likelihood (ML, Felsenstein 1981) analysis to be performed using PAUP. ML combines the advantages of model-based methods for processing probabilities of data, and extensive tree searches, and often produces similar results to Bayesian inference or maximum parsimony (Leaché and Reeder 2002; Sober 2004). To determine the nucleotide substitution model that best fits the data, the hierarchical likelihood ratio test (AIC) was used in Modeltest3.0 (Posada and Crandall 1998). ML analyses specifying the GTR + I + G model were implemented as heuristic searches with a 100 random-addition-sequence, TBR, and steepest descent options. The same options were used for ML bootstrap analyses (100 replicates).
For character optimization, 23 morphological characters were listed in Appendix 1 and coded in Table 2. The analysis was carried out using Mesquite 1.12 (Maddison and Maddison 2007) applying the options “trace all characters” and “parsimony ancestral states reconstruction”. The ancestral character states were reconstructed on nodes of the ML tree (Fig. 6) and are listed in Table 3.
Results: molecular data
The selected optimum model of sequence evolution for this combined data set was the general time reversible (GTR + I + G) model (Rodriguez et al. 1990): unequal base frequencies (A = 0.3613, C = 0.1406, G = 0.1181, T = 0.3800) and six substitution types (A/C: 0.9246, A/G: 1.2361, A/T: 0.3773, C/G: 1.2094, C/T: 1.0104) gamma distribution of rates among sites with alpha shape parameter 0.8496, proportion of invariable sites 0.3963. The analysis using these values yielded one optimum ML tree with a log-likelihood score of –ln L = 9484.50 (Fig. 6).
There are three weakly supported groups (clades to some degree) in continental Africa:
-
1
the I nversodicraea group (abbreviated as I n in Fig. 6), with three Inversodicraea species (I. cf. annithomae, I. cristata, I. ledermannii) forming a clade and the fourth Inversodicraea species (i.e. I. cf. bosii) being sister to all taxa studied from continental Africa;
-
2
the second group has pollen monads (i.e. single pollen) and is referred to as the L edermanniella - M onad group (LM). It appears paraphyletic and consists of a clade with Ledermanniella bifurcata in addition to members of the genera Leiothylax, Letestuella, Macropodiella, and Stonesia, with M. linearifolia being basal to the clade; and
-
3
the remaining taxa show pollen dyads and accordingly, form the L edermanniella - D yad group (LD). It is actually a weakly supported clade. Besides the Ghanaian endemic Ledermanniella bowlingii there are three other species of subg. Ledermanniella, i.e. L. letouzeyi, L. linearifolia and L. pusilla. Moreover, all species of the genera Dicraeanthus and Djinga belong to this group.
Results: vegetative structures
Leaf architecture
Leaves may be petiolate with a blade forked several times into narrow segments, e.g. in Dicraeanthus zehnderi and Inversodicraea cristata. The petioles of I. cristata are nearly cylindrical (Fig. 1a, d) whereas the terminal leaf segments are slightly flattened with chloroplasts restricted to the epidermis (Fig. 1f). Other African Podostemoideae such as Ledermanniella linearifolia and Leiothylax quangensis have entire linear or only once-forked leaves. Several African Podostemoideae have stipules which are subulate lateral appendages of the leaf base. For example, D. zehnderi (Fig. 3a–c) and L. quangensis (Fig. 4e) have stipules attached to the leaf base. In members of Inversodicraea (e.g. I. cristata) it is not obvious if there are stipules or not, because the stem is covered with many stipule-like scales in addition to the petiolate and compound leaves (Fig. 1a, b). Filled with silica bodies (Fig. 1c, e) these stem scales protect the shoots in the rushing water. The stem scales outnumber the compound leaves up to 20 times (Fig. 1d). The compound leaves contain rudimentary vascular bundles in the terminal segments. Inversodicraea cristata is usually characterised by cristate stem scales with five to seven teeth. However, a closer look reveals that the scales of I. cristata are more polymorphic. Scales along vegetative stem portions are broad (triangular) and provided with five to seven (rarely up to 10) teeth (Fig. 1a, b). Next to the terminal flowers, however, the stem scales of I. cristata are subulate with one or two teeth only (Fig. 2b). The uppermost foliage leaves are reduced to filamentous structures with a few forks. The uppermost stem scales and associated leaf bases are fused to some degree, forming a cup that surrounds and protects the floral buds, at least their basal portions (Fig. 2a, b).
Infraspecific variation of shoot size
Shoot length can vary substantially within one and the same species. For example, Dicraeanthus zehnderi usually has root-born shoots 10–60 cm long which are branched several times. However, there are also very short shoots with two leaves and a single terminal flower (not yet described in the taxonomic literature). Other podostemoids such as Leiothylax quangensis also differ strongly in the length of root-born shoots. The longest shoots are branched and up to 30 cm long, whereas others are 2 cm or less. Tiny shoots are reduced to flower buds, which are sessile on the upper root surface.
Shoot branching
Only a few African podostemoids show axillary branching as typical for most other angiosperms. For example, single flowers seem to occupy the leaf axils of long shoots in Letestuella tisserantii. Also vegetative lateral buds of Letestuella arise in leaf axils, or seemingly so (Fig. 5g).
In many African podostemoids stem branching is associated with double-sheathed leaves which occur toward the shoot tip after the formation of a set of normal, i.e. single-sheathed leaves. Double-sheathed leaves subtend two shoot buds, one on the left (ventral) side and one on the right (dorsal) side of their base, as shown in a longitudinal section of a branching shoot tip of Inversodicraea cristata (Fig. 1c). The two buds on opposite sides of a double-sheathed leaf may differ substantially in their size. For example, the double-sheathed leaf (L) of I. cristata shown in Fig. 1a, b appears as lateral appendage of an elongated stem which is covered with cristate scales. The dorsal side of the leaf base (i.e. its “dorsal sheath”) is occupied by a globular shoot bud (B) that seems to rest (Fig. 1b). In other samples of I. cristata the shoots (or shoot buds) in both sheaths of a double-sheathed leaf are more equal in size. A similar branching type is found in Dicraeanthus zehnderi where the leaves of long (“main”) shoots arise in a distichous pattern, or nearly so. Each leaf (i.e. L6–L8 in Fig. 3a, b) is again a double-sheathed leaf with the ventral sheath provided with two lateral stipules. On the dorsal side of the leaf bases we find two or three subulate stipules which belong to the “dorsal sheath” of a double-sheathed leaf (Fig. 3b, marked with asterisks). The ventral sheath again faces the main shoot (or a terminal flower) whereas the “dorsal sheath” subtends an inconspicuous shoot bud with outgrowing tiny leaves (L′ in Fig. 3a).
Stems in some African podostemoids such as Letestuella quangensis tend to be slightly flattened rather than being cylindrical or terete. Moreover, the vascular stem tissue is reduced and may consist of three to five rudimentary bundles arranged in a bow (Fig. 4g). Short shoots that occur along long shoots may consist of a sessile flower and one leaf (Fig. 4d). Two consecutive stem cross-sections at and below leaf insertion show the insertion of a short lateral shoot (SVS) in a non-axillary position on the right flank rather than on the dorsal side of the leaf base (Fig. 4e, f). The leaf L is again a double-sheathed leaf with the ventral (=main) sheath facing towards the stem whereas its “dorsal sheath” (including additional shoot bud) is displaced towards one leaf flank.
Root architecture
Narrow ribbon-like roots about two to four times as wide as thick are found in various African podostemoids. In Inversodicraea cristata the root ribbons form endogenous buds which turn into lateral roots (Fig. 1g). Much broader ribbons and crusts are found in, e.g., Dicraeanthus (Fig. 3h) and Letestuella (Fig. 5h). The dorsiventrally flattened crustose roots of, e.g., D. zehnderi are fixed to the rock by adhesive hairs. These crustose roots have a layer of vascular tissue or isolated vascular strands lacking obvious xylem and phloem, in contrast to a central cylinder as typical for roots in most other angiosperms (Fig. 3h). In the latter species and in Leiothylax it is difficult to distinguish the root tissue from the disk-like holdfast that surrounds the base of an established stem as its adhesive foot. Crustose roots give rise to daughter roots (lobes) exogenously while root-born shoots start as endogenous buds and break through the upper surface of the crustose roots (Fig. 5h).
Results: reproductive structures
Spathella as tubular or sac-like cover of floral buds
All Podostemoideae are characterised by a tubular or sac-like cover (“spathella”) that protects each floral bud. In African podostemoids such as Leiothylax and Letestuella these spathellas have a slightly pointed tip (Fig. 5b, c). The spathellas consist of ca. four cell layers and lack any vascular tissue. The outer spathella epidermis is papillate in, e.g., Dicraeanthus zehnderi (Fig. 3c, e) and smooth in other taxa. The spathella ruptures in order to release the flower before anthesis (Fig. 2a).
Position of flower bud inside spathella
There is one flower per spathella, as typical for most podostemoid genera worldwide (Cook and Rutishauser 2007). Most African members have flower buds in an upside-down position inside the spathellas. For example, Leiothylax quangensis has inverted flower buds before anthesis. Thus, the closed spathella contains an elongate floral stalk (pedicel) with a hanging ovary, i.e. the two stigmas are pointing downwards (Fig. 4c). Similarly, the flower buds of all species of Dicraeanthus, Inversodicraea and Ledermanniella have an inverted position inside the spathella. Thus, both pedicel and ovary are observable in mid-level sections through spathellas (Figs. 3e, 4a). African Podostemoideae start with an erect to inclined floral bud position during early development, as shown for Leiothylax quangensis (Fig. 4d). The flower bud of most African podostemoids then turns to an upside-down position by continued pedicel elongation inside the enlarging spathella. Before anthesis the spathella ruptures and delivers the flower that soon becomes upright or nearly so on the elongating pedicel, as shown for I. cristata (Fig. 2a, b). Only a few African Podostemoideae have upright or slightly inclined ovaries throughout their whole development. For example, Letestuella tisserantii has an oblique ovary position (Fig. 5c). After spathella rupture the inclined ovary may become completely upright (Fig. 5a).
Flower structure
Various African members have a gynophore, i.e. a gynoecium stalk in addition to a floral stalk (e.g. Leiothylax, Fig. 4c). The androecium of African podostemoids consists either of a single stamen (Fig. 5a) or of a Y-shaped structure with a common foot (andropodium) carrying 2 (rarely 3) stamens (Fig. 2a). There are exceptions, for example Letestuella, which switch between flowers with a single stamen (Fig. 5a) and flowers with two stamens on an andropodium within one and the same plant. There are usually two subulate appendages (“tepals”) on either side of the androecium (Figs. 2a, 5a). Exceptionally, there may be a twin of tepals on one side, i.e. three tepals in total for a single flower, as observed once in Letestuella (Fig. 5a).
There are usually two stigmas overtopping the ovary in Podostemoideae, according to the two carpels involved in ovary formation. Most podostemoids have two free, linear stigmas, as typical for nearly all Ledermanniella spp. (Fig. 2a). Peculiar stigma shapes are found in Dicraeanthus and Letestuella. The stigmas are prominent and horn-like in Dicraeanthus (Fig. 3d) whereas they are short and club-shaped in Letestuella (Fig. 5a).
All African taxa covered by this study have unilocular ovaries lacking a septum (Figs. 4a, 5e). The ovules are borne on a slightly inflated central placenta that is connected with the ovary tip and the ovary base (Fig. 4c). The ovules are bitegmic and tenuinucellate (Figs. 2f, 4b).
Pollen: monads versus dyads
As usual for most Podostemoideae the pollen of African members is tricolpate (Fig. 3g). African Podostemoideae have monads or dyads, depending on the group. For example, monads are character for I. cristata (Fig. 2c), whereas Dicraenthus (Fig. 3f) has dyads, i.e. pollen grains that are dispersed in pairs. Loose dyads that easily decay into monads are found in Djinga felicis and Stonesia ghoguei.
Capsules: ribbed versus smooth
The ovary walls of all Podostemaceae have transversally elongate fibres in the inner epidermis and longitudinally elongate fibers in the hypodermis (Figs. 2d, e, 4a–c, 5f). The ovaries in African podostemoids can be globular as in Leiothylax (Fig. 4c) and Letestuella (Fig. 5a). Or they are ellipsoidal to spindle-shaped as in Inversodicraea cristata (Fig. 2a). The ovaries of anthetic flowers of most Podostemoideae do not have ribs. During capsule maturation after anthesis some podostemoids (including all members of Ledermanniella) acquire prominent ribs. These ribs of some groups become clearly visible on shrinking or shedding of the parenchyma of the young capsule wall. For example, Inversodicraea cristata has a nearly smooth ovary during anthesis (Fig. 2a), but a strongly ribbed capsule afterwards (Fig. 2d). Mature capsules are smooth (i.e. without ribs) in Leiothylax quangensis (Fig. 4a–c). Only slightly ribbed capsules are found in Letestuella tisserantii (Fig. 5d–f).
Discussion
Molecular phylogeny
Overall, our ML tree (Fig. 6) does not conflict with the clades found in the analysis by Moline et al. (2007), but is able to attribute our newly collected taxa to particular groups. In general, however, many clades show only low bootstrap values, which might be caused by the high proportion of missing data hampering our phylogenetic conclusions. Still, several nodes gain additional support by morphological evidence. An expanded phylogenetic analysis indicated a monophyletic group of all continental African Podostemoideae included here being sister to the two Madagascan genera Endocaulos and Thelethylax (Pfeifer et al. 2009, unpublished data), a clade characterised in some members by endogenous formation of flowers along stems, because of dedifferentiation of cortex tissue, and epiphyllous flowers. These are architectural peculiarities restricted to African podostemoids (Moline et al. 2007; Rutishauser et al. 2008; Pfeifer et al. 2009). In general, the groupings correlate with Ledermanniella subg. Phyllosoma, now attributed to Inversodicraea, and groups with pollen monads and dyads, respectively. Within the African clade, Inversodicraea cf. bosii occupies the most basal position, being sister to the lineage including the remaining taxa. Such a relationship, however, may be questioned because of high morphological similarities between Inversodicraea cf. bosii and I. ledermannii and preliminary molecular phylogenetic studies by Kita et al. (2005, 2008), where I. cf. bosii occupies a basal position within the Inversodicraea clade. The presence of stem scales for this entire group as proposed by Cusset (1984a) may be a further argument to consider the inclusion of I. cf. bosii. Another critical point in our tree may be the intermediate position of Monandriella between the Inversodicraea and the Ledermanniella-Monad group, making the latter group paraphyletic with regard to Ledermanniella-Dyad taxa and therefore suggesting a secondary evolution of pollen dyads. At this point we leave it open whether such a scenario, supported by the presence of shared exogenous root branching as homoplastic synapomorphy, or inclusion of Monandriella into the LM group is favoured. Despite bootstrap support <50%, the LD group consists of taxa with pollen dyads only, which may be seen as substantial morphological support. Within this clade, the Dicraenthus africanus–D. zehnderi lineage appears best supported by their typical stigma lobes (Tables 3, 5).
The need for a taxonomic rearrangement with splitting of the large and artificial genus Ledermanniella as defined by Cusset (1974, 1984a, 1987) is obvious. Irrespective of whether Inversodicraea (formerly Ledermanniella subg. Phyllosoma), the Ledermanniella-Monad group, and the Ledermanniella-Dyad group are monophyletic or possibly paraphyletic and weakly supported, the polyphyly of Ledermanniella sensu lato in its present (i.e. Cusset’s) circumscription is well documented. Hence, some groups appear more natural on returning to genera names used before 1974 (Tables 4, 5). Thus, we advocate:
-
1
resurrecting Monandriella Engler (1926, 1928) as a monotypic genus, with M. linearifolia Engler (syn. L. monandra);
-
2
restricting the genus name Inversodicraea Engler ex R.E. Fries (1914) to all members of the “scaly” Ledermanniella subg. Phyllosoma.
Because of the described morphological and molecular evidence we apply the same concept to Inversodicraea bosii. We further advocate:
-
3
accepting the genus name Ledermanniella Engler (1909) for all species of the former subgenus Ledermanniella, except for L. monandra (see above).
These proposals for the taxonomic changes are summarised in Table 4, including all authorities. A morphological comparison of the studied genera (including the resurrected genera Inversodicraea and Monandriella) is given in Table 5. Thus, we solved, at least to some extent, the unsatisfactory situation of the generic classification. Still, Ledermanniella sensu stricto remains polyphyletic because L. bifurcata, L. bowlingii, L. letouzeyi, L. linearifolia, and L. pusilla are scattered across the Ledermanniella-Monad and Ledermanniella-Dyad groups. Future taxonomic changes should be implemented as soon as more data confirm our results.
Morphological characters and hypotheses about their evolutionary history by plotting them on to the molecular cladogram
Morphological evolution in African podostemoids and the meaning of morphological data for taxonomy in this group will be discussed below. The genera involved are listed in Table 5, covering the most pertinent morphological characters for their delimitation (Fig. 6; Tables 2, 3).
-
Pollen monads versus dyads (=character 3 in Appendix 1, Tables 2, 3) Most Podostemoideae genera are characterised by monads or dyads (Cook and Rutishauser 2007). This indicates that monad versus dyad is useful as a diagnostic character. As typical for all Asian and many American podostemoids, the tricolpate pollen of several African genera (including Ledermanniella pro parte, Saxicolella, Thelethylax) are in firm (i.e. non-decaying) dyads (Bezuidenhout 1964; Ameka et al. 2002, 2003; Jäger-Zürn et al. 2006; Grob et al. 2007; Moline et al. 2007). In Ledermanniella s.str. and in Inversodicraea, there are taxa characterised by dyads and taxa lacking them. Our new molecular data indicate that in Ledermanniella s.str.there is a weakly supported clade, i.e. the Ledermanniella -Monad (LM) group, with pollen as single grains (monads). The Ledermanniella -Dyad (LD) group, however, consists of taxa with pollen dispersed in dyads. Moreover, it seems that monad genera such as Leiothylax, Letestuella, Macropodiella, and S. ghoguei (with a few dyads in addition to the usual monads) belong to the LM group, whereas the dyad genus Dicraeanthus, and Djinga felicis (with a few monads in addition to the usual dyads) are part of the LD group (Ghogue et al. 2009; Pfeifer et al. 2009).
-
Position of flower buds in spathella (=character 4 in Appendix 1, Tables 2, 3) Flower position (or, more exactly, ovary position) inside the spathella is an important character distinguishing the African and Madagascan podostemoid genera and species from other groups (Cusset 1987; Cook and Rutishauser 2007; Grob et al. 2007; Moline et al. 2007; Rutishauser et al. 2007). All podostemoid members in America and Australasia have erect (or only slightly inclined) flower buds inside the spathella. This pattern is found (as reversal) in a few podostemoid species from Africa and Madagascar, e.g. in Djinga and Letestuella (Cook and Rutishauser 2007; Ghogue et al. 2009). Most African Podostemoideae have inverted flower buds inside the spathellas (Ameka et al. 2003), a character state which is found at all internal nodes except for Djinga (Table 3). Thus, completely or partially inverted flower buds are a unique feature (synapomorphy) of several African and Madagascan Podostemoideae. It seems that podostemoids with inverted flowers evolved only once from podostemoids with erect flowers (Moline et al. 2007). The African genus Saxicolella (with upright flower buds) is not included in this analysis. Preliminary molecular data, however, indicate a basal position of Saxicolella with regard to all African Podostemoideae; thus Saxicolella may have retained erect floral buds as plesiomorphic character state (Ameka et al. 2002; Kelly et al. 2009).
-
Ribbed versus smooth capsules (=characters 8 and 9 in Appendix 1, Tables 2, 3) Mature capsules of African podostemoids are ellipsoid to spindle-shaped or (rarely) globose. They are smooth or more often ribbed. Most African taxa have three ribs per valve, running the entire length of the capsule. The dominance of three ribs per valve can be clearly interpreted as a plesiomorphic character state of African Podostemaceae (Table 3). Only Endocaulos, Paleodicraeia, and Stonesia are known to have an increased rib number per valve, with some but not all of the five to seven ribs running along the entire length of the capsule valve (Pfeifer et al. 2009).
-
Shoots with non-axillary branching and double-sheathed leaves (=character 13 in Appendix 1, Tables 2, 3) Axillary branching as typical for most angiosperms occurs only rarely in Podostemoideae, e.g. in Letestuella tisserantii and Saxicolella submersa from Africa (Ameka et al. 2002). In many Podostemoideae from Africa and America (e.g. Ledermanniella bowlingii, Podostemum ceratophyllum) there are leaves with two sheaths that are inserted laterally and opposite each other, which is the predominant character in our group. Such leaves have been called double-sheathed or “dithecous” by Warming (e.g. 1891) and others (Moline et al. 2006, 2007). The occurrence of double-sheathed leaves among conventional (i.e. single-sheathed) leaves allows the stem to branch by a peculiar process that may be called “bifurcation”. As long as a stem forms single-sheathed leaves, it grows in a monopodial manner. A double-sheathed leaf then appears in a terminal (or nearly terminal) position giving rise to new shoot modules (daughter shoots) in each sheath (as shown for Inversodicraea cristata, Fig. 1c). Imaichi (2007) observed double-sheathed leaves in Dicraeanthus africanus inside the lateral short shoots whereas the long shoots (with lengths up to 30 cm) are unbranched and lack double-sheathed leaves.
-
Flower position along shoots (=character 14 in Appendix 1, Tables 2, 3) Flowers in non-American Podostemoideae are either solitary in terminal position of short shoots, which do not contain double-sheathed leaves (e.g. Saxicolella amicorum, Thelethylax spp., and most Asian members), or they occur in clusters, positionally associated with double-sheathed leaves, in addition to leaves (bracts) with one sheath, as in species of the African genera Dicraeanthus, Inversodicraea, Ledermanniella, Saxicolella, Sphaerothylax, besides Zeylanidium subulatum as Asian exception (Jäger-Zürn 1999, 2000; Ameka et al. 2002, 2003; Imaichi et al. 2005; Moline et al. 2007). Positional association of flowers with double-sheathed leaves is also known from American podostemoids which are basal to all Asian, Madagascan, and African members of this subfamily (Rutishauser and Grubert 2000). It seems that various African podostemoids that switched to endogenous formation of flower buds from inside the stem cortex lack double-sheathed leaves. Within our group, an endogenous origin of flowers in the stem cortex evolved once in the Ledermanniella-Dyad group, but seems to be lost in Ledermanniella linearifolia. Endogenous floral bud formation along the stem was found in L. letouzeyi and Stonesia ghoguei (Moline et al. 2007; Rutishauser et al. 2007; Pfeifer et al. 2009).
-
Stem scales (=characters 19 and 20 in Appendix 1, Tables 2, 3) All members of Inversodicraea (ca. 20 spp.) are characterised by stem scales with entire or toothed margins (Cusset 1984a, 1987). These stem scales contain many silica bodies and may serve as exoskeleton. They resemble the scaly leaves of the Brazilian podostemoid Diamantina and the stem scales of Weddellina, which forms its own subfamily (Rutishauser et al. 2005; Koi and Kato 2007). The stem scales in members of Inversodicraea are usually much smaller and more numerous than the compound leaves, which are also present. Compound leaves are arranged in a distichous or helical pattern, whereas stem scales are inserted irregularly around the stem. Scale morphology is traditionally taken as important to distinguish the species of Inversodicraea (Cusset 1984a, 1987; Rutishauser et al. 2007). Scale morphology, however, may vary substantially within a species. For example, scales of I. ledermannii are tridentate, more rarely entire or quadridentate (Moline et al. 2007). Inversodicraea cf. annithomae has elliptical to oboval scales. Their margin is smooth or slightly to strongly dentate with up to ten teeth (Cusset 1984a). The scales of I. cristata are even more polymorphic. Scales along vegetative stem portions are triangular and provided with several teeth. Next to the terminal flowers, however, the stem scales of I. cristata are more similar to those of I. ledermannii, being subulate with one or two teeth only.
-
Ribbon-like to crustose roots (=characters 21–23 in Appendix 1, Tables 2, 3) Roots with their tips covered by a cap (calyptra) are found in the outgroup taxa such as Podostemum ceratophyllum from North America (Philbrick and Novelo 2004; Moline et al. 2006; Jäger-Zürn 2008). Most African Podostemoideae, however, lack root-caps completely (for Saxicolella and Thelethylax as exceptions, see Ameka et al. 2002, Grob et al. 2007). Cap-less ribbon-like roots (width usually 6 mm or less) are found in many African and Madagascan Podostemaceae, e.g. Ledermanniella bowlingii (Ameka et al. 2003) and Inversodicraea ledermannii (Moline et al. 2007). They show endogenous formation of both lateral roots and shoots along the margins. In various African members there are crustose roots (width > 10 mm) with endogenous shoots arising from the upper surface. Crustose roots branch into daughter (or lateral) lobes by exogenous branching. Such lobed crusts occur in some species of Dicraeanthus, Inversodicraea, Ledermanniella, Macropodiella, and Sphaerothylax (Rutishauser et al. 2007). Crustose roots may have evolved once or twice within African podostemoids from groups having ribbon-like roots (Moline et al. 2007). In our phylogeny they probably evolved in the core Ledermanniella-Monad and the Ledermanniella-Dyad group (node h in Fig. 6; Table 3), although with some plasticity. The broad root ribbons (6 mm wide or more) with their exogenous lobes of Djinga felicis (Ghogue et al. 2009) and Stonesia ghoguei (Pfeifer et al. 2009) may be viewed as intermediates between narrow ribbons and broad crusts. According to Koi et al. (2006), who studied Asian podostemoids, crustose roots evolved as homoplasies twice (e.g. Hydrobryum spp., Zeylanidium olivaceum) from Asian members with narrow ribbon-like roots.
Conclusions and outlook
Our results from analysis of new material of rarely collected African Podostemaceae reveal the morphological differentiation and interpret the enigmatic features in a phylogenetic framework. Similar to Moline et al. (2007) we infer that the generic delimitations in some African taxa require revision. In a first step we implemented some taxonomic rearrangements, while others are awaiting new molecular data. The inclusion of additional molecular markers and additional African and non-African members of Podostemoideae in molecular analyses is upcoming (cf. Kita et al. 2005, 2008; Kelly et al. 2009; Brad Ruhfel, in preparation). As far as we are aware, African Podostemaceae occupy either very restricted areas or are widespread. New field collections of African Podostemaceae are needed to evaluate which of the “local species” (e.g. Dicraeanthus zehnderi, Inversodicraea annithomae, I. bosii) belong to polymorphic but more widespread taxa. Morphological and developmental analyses of various African podostemoids are on the way (Konrad Huber, in preparation). These studies will promote our knowledge of this interesting plant group by combining anatomical data and molecular systematics. There is an ongoing multidisciplinary project on Podostemaceae in Cameroon (Ghogue et al. 2009), including botanical fieldwork, ethnobotanical surveys, and studies on biogeography and conservation biology.
References
Ameka GK, Pfeifer E, Rutishauser R (2002) Developmental morphology of Saxicolella amicorum and S. submersa (Podostemaceae: Podostemoideae) from Ghana. Bot J Linn Soc Lond 139:255–273
Ameka GK, Clerk CG, Pfeifer E et al (2003) Developmental morphology of Ledermanniella bowlingii (Podostemaceae) from Ghana. Plant Syst Evol 237:165–183
Ameka GK, Ghogue J-P, Rutishauser R (2009) Podostemaceae research in Africa—state of the art and open questions. In: Proceedings XVIII AETFAT congress Yaoundé, Cameroon (in press)
APG II (Bremer B, Bremer K, Chase MW et al) (2003) An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants. Bot J Linn Soc 141:399–436
Bezuidenhout A (1964) The pollen of the African Podostemaceae. Pollen Spores 6:463–478
Bininda-Emonds ORP (2004) The evolution of supertrees. Trends Ecol Evol 19:315–323
Cook CDK, Rutishauser R (2007) Podostemaceae. In: Kubitzki K (ed) The families and genera of vascular plants, vol 9. Springer, Berlin, pp 304–344
Cusset C (1974) Contribution à l’étude des Podostemaceae. 4. Les genres Ledermanniella, Monandriella et Inversodicraea. Adansonia Sér 2 14(2):271–275
Cusset C (1984a) Contribution à l’étude des Podostemaceae. 7. Ledermanniella Engl. sous-genre Phyllosoma C. Cusset. Bull Mus Natl Hist Nat Paris Sér 5 (Sect B Adansonia) 4:361–390
Cusset C (1984b) Contribution à l’étude des Podostemaceae. 8. Ledermanniella Engl. sous-genre Ledermanniella. Bull Mus Natl Hist Nat Paris Sér 6 (Sect B Adansonia) 3:249–278
Cusset C (1987) Podostemaceae and Tristichaceae. In: Satabié B, Morat Ph (eds) Flore du Cameroun, vol 30. Ceper, Yaoundé, pp 51–99
Davis CC, Webb CO, Wurdack KJ et al (2005) Explosive radiation of Malpighiales supports a mid-Cretaceous origin of modern tropical rain forests. Am Nat 165:E36–E65
De Queiroz A, Gatesy J (2007) The supermatrix approach to systematics. Trends Ecol Evol 22:34–41
Engler A (1909) Podostemonaceae Africanae III. Ledermanniella nov. gen. Bot Jahrb Syst 43:378–381
Engler A (1926) Podostemonaceae Africanae IV. Monandriella nov. gen. Bot Jahrb Syst 60:451–467
Engler A (1928) Reihe Podostemales. In: Engler A, Prantl K (eds) Die natürlichen Pflanzenfamilien. Engelmann, Leipzig, pp 1–68, 483–484
Felsenstein J (1981) Evolutionary trees from DNA sequences: a maximum likelihood approach. J Mol Evol 17:368–376
Fries RE (1914) Podostemonaceae. Wissenschaftliche Ergebnisse der Schwedischen Rhodesia-Kongo-Expedition 1911–1912, vol 1. Aftonbladets, Stockholm, pp 56–58
Ghogue JP, Imaïchi R, Kita Y et al (2009) River ecology and distribution of Podostemaceae (river-weeds) in Cameroon, especially its West Province. In: Proceedings XVIII AETFAT congress Yaoundé, Cameroon (in press)
Ghogue JP, Ameka GK, Grob V et al (2009b) Enigmatic morphology of Djinga felicis (Podostemaceae–Podostemoideae), a badly known endemic from northwestern Cameroon. Bot J Linn Soc 160:64–81
Grob V, Pfeifer E, Rutishauser R (2007) Morphology, development and regeneration of Thelethylax minutiflora, a Madagascan riverweed (Podostemaceae). Phyton 47:205–229
Gustafsson MHG, Bittrich V, Stevens P (2002) Phylogeny of Clusiaceae based on rbcL sequences. Int J Plant Sci 163:1045–1054
Higgins DG, Sharp PM (1988) CLUSTAL: a package for performing multiple sequence alignment on a microcomputer. Gene 73:237–244
Imaichi R (2007) Elaborate floral development of African Dicraeanthus africanus. In: Abstracts XVIII AETFAT congress Yaoundé, Cameroon, p 137
Imaichi R, Hiyama Y, Kato M (2005) Leaf development in absence of a shoot apical meristem in Zeylanidium subulatum (Podostemaceae): evolutionary implications. Ann Bot 96:51–58
Jäger-Zürn I (1999) Developmental morphology of the shoot system of Podostemum subulatum (Podostemaceae–Podostemoideae). Beitr Biol Pflanzen 71:281–334
Jäger-Zürn I (2000) The unusual ramification mode of Sphaerothylax abyssinica (Wedd.) Warm. (Podostemaceae–Podostemoideae). Flora 195:200–227
Jäger-Zürn I (2008) Morphological analysis of shoots and roots in Thelethylax minutiflora and T. insolata (Podostemaceae–Podostemoideae): taxonomic and evolutionary implications. Bot Jahrb Syst 127:245–272
Jäger-Zürn I, Novelo AR, Philbrick CT (2006) Microspore development in Podostemaceae–Podostemoideae, with implications on the characterization of the subfamilies. Plant Syst Evol 256:209–216
Kelly LJ, Ameka GK, Chase MW (2009) DNA barcoding of African Podostemaceae (river-weeds): a test of proposed barcode regions. Plant Syst Evol (in press)
Kita Y, Kato M (2001) Infrafamilial phylogeny of the aquatic angiosperm Podostemaceae inferred from the nucleotide sequence of the matK gene. Plant Biol 3:156–163
Kita Y, Imaichi R, Ghogue J-P et al (2005) Molecular phylogeny of Cameroonian and Ghanaian Podostemaceae. In: Abstract poster, international botanical congress, Vienna
Kita Y, Koi S, Rutishauser R et al (2008) A new species of Ledermanniella (Podostemaceae) from Cameroon. Acta Phytotaxon Geobot 59:223–227
Koi S, Kato M (2007) Developmental morphology of the shoot in Weddellina squamulosa and implications for shoot evolution in the Podostemaceae. Ann Bot 99:1121–1130
Koi S, Fujinami R, Kubo N et al (2006) Comparative anatomy of root meristem and root cap in some species of Podostemaceae and the evolution of root dorsiventrality. Am J Bot 93:682–692
Korotkova N, Schneider JV, Quandt D et al (2009) Phylogeny of the eudicot order Malpighiales—analysis of a recalcitrant clade with sequences of the petD group II intron. Plant Syst Evol 257(3–4):189–203
Leaché AD, Reeder TW (2002) Molecular systematics of the Eastern fence lizard (Sceloporus undulatus): a comparison of parsimony, likelihood, and Bayesian approaches. Syst Biol 51:44–68
Lebrun JP, Stork AL (1991). Enumeration des plantes à fleurs d’Afrique tropicale, 1. Généralités et Annonaceae à Pandaceae. Conservatoire et Jardin botaniques de Geneve, Geneva
Maddison DR, Maddison WP (2002) MacClade. Version 4. Sinauer, Sunderland
Maddison WP, Maddison DR (2007) Mesquite: a modular system for evolutionary analysis. Version 1.12, pp 315–319. http://mesquiteproject.org
Mason-Gamer RJ, Kellogg EA (1996) Testing for phylogenetic conflict among molecular data sets in the tribe Triticeae (Gramineae). Syst Biol 45:524–545
Moline PM, Don L, Philbrick CT et al (2006) Comparative morphology and molecular systematics of Podostemum (including Crenias)–American river-weeds (Podostemaceae). Bot Jahrb Syst 126:427–476
Moline PM, Thiv M, Ameka GK et al (2007) Comparative morphology and molecular systematics of African Podostemaceae–Podostemoideae, with emphasis on Dicraeanthus and Ledermanniella from Cameroon. Int J Plant Sci 168:159–180
Pfeifer E, Grob V, Thiv M et al (2009) Stonesia ghoguei—peculiar morphology of a new Cameroonian species (Podostemaceae–Podostemoideae). Novon 19:102–116
Philbrick CT, Novelo RA (2004) Monograph of Podostemum (Podostemaceae). Syst Bot Monogr 70:1–106
Posada D, Crandall KA (1998) MODELTEST: testing the model of DNA substitution. Bioinformatics 14:817–818
Rodriguez F, Rodriguez JF, Oliver A et al (1990) The general stochastic model of nucleotide substitution. J Theor Biol 142:485–501
Rutishauser R, Grubert M (2000) Developmental morphology of Apinagia multibranchiata (Podostemaceae) from the Venezuelan Guyanas. Bot J Linn Soc 132:299–323
Rutishauser R, Pfeifer E, Novelo RA et al (2005) Diamantina lombardii—an odd Brazilian member of the Podostemaceae. Flora 200:245–255
Rutishauser R, Pfeifer E, Bernhard A (2007) Podostemaceae of Africa and Madagascar: keys to genera and species, including genera descriptions, illustrations to all species known, synonyms, and literature list. Revised version 20-07-2007 (first version 2004). http://www.systbot.uzh.ch/podostemaceae
Rutishauser R, Grob V, Pfeifer E (2008) Plants are used to having identity crises. In: Minelli A, Fusco G (eds) Key themes in evolutionary developmental biology. Cambridge University Press, Cambridge, pp 194–213
Sanderson MJ, Purvis A, Henze C (1998) Phylogenetic supertrees: assembling the trees of life. Tree 13:105–109
Schnell R (1998) Anatomie des Podostémacées. In: Carlquist S, Cutler DF, Fink S, Ozenda P, Roth I, Ziegler H (eds) Encyclopedia of plant anatomy: extreme adaptations in angiospermous hydrophytes. Borntraeger, Berlin, pp 197–283
Sehgal A, Sethi M, Mohan Ram HY (2002) Origin, structure and interpretation of thallus in Hydrobryopsis sessilis (Podostemaceae). Int J Plant Sci 163:891–905
Sehgal A, Khurana JP, Sethi M et al (2007) Organ identity of the thalloid plant body of Griffithella hookeriana and Polypleurum stylosum—Podostemoideae (Podostemaceae). Plant Syst Evol 267:93–104
Shaw J, Lickey E, Beck J et al (2005) The tortoise and the hare II: relative utility of 21 noncoding chloroplast DNA. Am J Bot 92:142–166
Sober E (2004) The contest between parsimony and likelihood. Syst Biol 53:644–653
Taberlet P, Gielly L, Pautou G et al (1991) Universal primers for amplification of three non-coding regions of chloroplast DNA. Plant Mol Biol 17:1105–1109
Warming E (1881) Familien Podostemaceae I. Kgl Danske Vidensk Selsk Nat Math Afd 2:1–34
Warming E (1891) Familien Podostemaceae IV. Kgl Danske Vidensk Selsk Nat Math Afd 7:133–179
Warming E (1899) Familien Podostemaceae V. Kgl Danske Vidensk Selsk Nat Math Afd 9:105–154
Wiens JJ (2006) Missing data and the design of phylogenetic analyses. J Biomed Inform 39:34–42
Wurdack KJ, Davis CC (2009) Malpighiales phylogenetics: gaining ground on one of the most recalcitrant clades in the angiosperm tree of life. Am J Bot 96:1551–1570
Acknowledgments
We would like to express our thanks to G. Ameka (Legon, Ghana) for his help with the field work, U. Jauch (Institute of Plant Biology, University of Zurich) for his technical assistance with scanning electron microscopy, K. Esfeld (University of Bern) for her DNA laboratory work, and P. K. Endress (University of Zurich) as reviewer for his precious comments on the manuscript. This paper is part of a research project supported by the Swiss National Science Foundation (grant No. 3100AO-105974 to R. Rutishauser).
Author information
Authors and Affiliations
Corresponding author
Electronic supplementary material
This file is unfortunately not in the Publisher's archive anymore: Supplementary doc S1 (NEX 69 kb)
Appendix 1: Morphological characters
Appendix 1: Morphological characters
Morphological characters and characters states used for African Podostemaceae modified and supplemented after Moline et al. (2007) and Philbrick and Novelo (2004).
-
1.
Maximum stamen number: 0 = two (or three but never more), 1 = one.
-
2.
Third tepal on andropodium: 0 = absent, 1 = present.
-
3.
Pollen union: 0 = monads, 1 = dyads, write (01) if loose dyads present or a mixture of monads to dyads with a ratio from 1:4 to 4:1 (as found in Djinga and Stonesia pro parte).
-
4.
Ovary position of flower buds in spathella: 0 = erect, 1 = oblique-inclined, 2 = inverted (hanging). Important here is the maximum degree of floral inversion in the spathella. Certain African taxa with completely inverted floral buds such as Inversodicraea ledermannii (Moline et al. 2007) show partial reversal before anthesis to an oblique-inclined position inside the elongating but still closed spathella.
-
5.
Ovary and capsule shape: 0 = spindle-shaped (fusiform) to obovoid and ellipsoidal (at least twice as long as thick), but not flattened, 1 = globose to subglobose (less than twice as long as thick), ellipsoidal and laterally flattened.
-
6.
Gynophore: 0 = absent, 1 = present (at least 1/6 of ovary length).
-
7.
Shape of stigma lobes: 0 = each lobe entire, linear, 1 = lobes branched (multilobed), 2 = each lobe entire, conical (as typical for Dicraeanthus), 3 = each lobe entire, club-shaped.
-
8.
Ribs of mature capsule: 0 = absent, 1 = present.
-
9.
Number of ribs per capsule valve: 0 = none observable, 1 = three ribs per valve, 2 = five or more ribs per valve.
-
10.
Septa in ovary and capsule: 0 = absent (unilocular ovary), 1 = present (bilocular or trilocular ovary).
-
11.
Valve symmetry: 0 = two equal valves (isolobous), 1 = two unequal valves, one persistent.
-
12.
Vegetative stem length: 0 = lacking or <1 cm, 1 = 1–10 cm, 2 = 10–30 cm.
-
13.
Double-sheathed leaves: 0 = absent (or not yet observed), 1 = present.
-
14.
Flower bud origin along stem: 0 = exogenous, usually protected by leaf sheaths in addition to spathella, 1 = endogenous origin in stem cortex (with or without preformed keyholes). Use (01) when intermediacy obvious, as in Djinga.
-
15.
Epiphylly: Additional flowers arising from the clefts of (on) forked leaves: 0 = no epiphyllous flowers observable, 1 = epiphyllous flowers usually observable (as in Ledermanniella letouzeyi and Stonesia ghoguei), (01 if epiphyllous flowers only rarely seen such as in special form of L. bowlingii, shown in Ameka et al. 2009).
-
16.
Maximum leaf length: 0 = <1 cm, 1 = 1–5 cm, 2 = always > 6 cm (often > 20 cm).
-
17.
Blade division of foliage leaves: 0 = forked (at least once, usually several times), 1 = simple and usually filamentous.
-
18.
Stipules: 0 = absent (sheath only), 1 = present (any type, usually two in lateral position, some leaves within same shoot may lack them).
-
19.
Stem scales in addition to foliage leaves (as typical for Inversodicraea): 0 = absent, 1 = present.
-
20.
Shape of stem scales (if present at all): 0 = subulate (entire or with 1 or 2 lateral teeth), 1 = broad and dentate to cristate (with up to 10 teeth). Notice: I. cristata has subulate stem scales just below flowers (Fig. 2b).
-
21.
Root shape: 0 = narrow to broad ribbons with outgrowth of endogenous shoots (flowers) only along flanks, 1 = crustose (disk-like) with outgrowth of endogenous shoots (flowers) on upper surface.
-
22.
Root cap: 0 = absent, 1 = present (at least rudimentary).
-
23.
Root branching: 0 = endogenous, 1 = exogenous outgrowth of daughter roots (lobes).
Rights and permissions
About this article
Cite this article
Thiv, M., Ghogue, JP., Grob, V. et al. How to get off the mismatch at the generic rank in African Podostemaceae?. Plant Syst Evol 283, 57–77 (2009). https://doi.org/10.1007/s00606-009-0214-4
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00606-009-0214-4