| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Systematics |
2Department of Botany, Stockholm University, SE-106 91 Stockholm, Sweden, and Molecular Systematics Laboratory, Swedish Museum of Natural History, Box 50007, SE-104 05 Stockholm, Sweden; 3Department of Palaeontology, The Natural History Museum, Cromwell Road, London, SW7 5BD, UK
Received for publication May 3, 2001. Accepted for publication October 11, 2001.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: lycopod • phylogeny • rbcL • resurrection plant • rubisco • Selaginellaceae • xerophyte
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). In these characteristics, the morphology of Selaginellaceae has changed little since the group was first encountered in the fossil record in the tropical wetland floras of the Carboniferous Period (Thomas, 1992, 1997
). Selaginellaceae are a cosmopolitan family with species capable of growing under a wide range of climate, soil, and light regimes. The group contains frost-tolerant, arctic-alpine species, delicate terrestrial rainforest species, and physiologically robust, drought-adapted xerophytes of desert scrub and heathland. Greatest diversity occurs in primary tropical rainforest, and it is conceivable that some elements of the group might have persisted in similar environments since the Late Paleozoic. Recently, we developed an outline phylogeny for living Selaginellaceae based on rbcL gene sequences (Korall, Kenrick, and Therrien, 1999
). Here, we build on this previous molecular work to develop a more detailed phylogenetic framework for the family.
Previous systematic studies divided Selaginellaceae into numerous groups (Spring, 1850
; Braun, 1857
; Baker, 1883
; Hieronymus, 1901
; Walton and Alston, 1938
; Jermy, 1986
; Soják, 1992
), and we follow the classification of Jermy (1986)
. Jermy recognized one genus (Selaginella Pal. Beauv.) containing five subgenera: Selaginella (2 species), Ericetorum Jermy (3 species), Tetragonostachys Jermy (
50 species), Stachygynandrum (Pal. Beauv.) Baker (600 species), and Heterostachys Baker (60 species). Our preliminary molecular phylogeny based on a representative sample of 18 species revealed that Selaginellaceae constitutes a monophyletic group and that the morphologically distinctive subgenus Selaginella is sister group to a clade comprising all other species (Korall, Kenrick, and Therrien, 1999
). In addition, the large subgenus Stachygynandrum was shown to be polyphyletic with some members paraphyletic to Tetragonostachys, Ericetorum, and Heterostachys. Tetragonostachys and Ericetorum are monophyletic, but monophyly of Heterostachys was untested.
Here we extend our phylogenetic data set of rbcL gene sequences from 18 to 62 species. Our sample includes a broad and representative selection of diversity, adding up to almost 10% of living species. In particular, we have focused on improving our sample of Stachygynandrum, choosing species from all continents and a wide range of environments. Taxon choice has been tailored to investigate the relationships of xerophytic species, and we have included four resurrection plants as well as other drought-tolerant forms. Taxa have also been chosen to reflect the morphological diversity within the group, focusing on those features that have been used by previous authors to define groups. In addition to characters of leaf, branch, and root, we have paid particular attention to stele morphology, sporangial arrangement, and megaspore wall structure. We have been careful to select characteristics that are observable in fossils, with the long-term aim of calibrating our phylogenetic tree.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
and Stefanovic, Rakotondrainibe, and Badré (1997)
where applicable; otherwise Reed (1965–1966)
is followed.
Choice of species
Ingroup
Choice of ingroup was based on previous taxonomic work (Baker, 1883
; Hieronymus, 1901
; Walton and Alston, 1938
; Jermy, 1986
), comparative morphology (emphasizing growth form, sporangial arrangement, and leaf, stele, and megaspore morphology), geographical distribution, and chromosome number. A total of 62 species were chosen and 44 of the sequences analysed here were not previously published (Table 1). Species recognition in Selaginellaceae is difficult and much effort was devoted to accurate identification. Where necessary and whenever possible, specimens were compared to type material, as indicated in Table 1. Most information on sporangial arrangement (Horner and Arnott, 1963
; Fraile and Riba, 1981
; Quansah, 1988
), chromosome numbers (Manton, 1950
; Tchermak-Woes and Dolezal-Janish, 1959
; Kuriachan, 1963
; Jermy, Jones, and Colden, 1967
; Löve and Löve, 1976
; Takamiya, 1993
), and some data on stele arrangement (Hieronymus, 1901
; Mickel and Hellwig, 1969
; Jermy, 1990
) were taken from the literature. Because species determination in Selaginellaceae is difficult and herbarium and cultivated specimens often are labelled incorrectly, data culled from the literature should be used with caution. Sporangial arrangement, rhizophore development, and stele form have, as far as possible, been verified through studies of herbarium material.
|
Fifty-three anisophyllous species were chosen from Stachygynandrum (including ten members of the series Articulatae Spring) and Heterostachys. All growth forms (e.g., creeping, erect, twining, rosettes) and a wide range of geographical distributions are covered (Table 1). Monostelic, bistelic, and polystelic species were included (Table 1). The steles of bistelic forms are predominantly terete, whereas those of monostelic forms are elliptical to strap-shaped. Selaginella exaltata has a unique "actino-plectostele" (Mickel and Hellwig, 1969
). Selaginella novae-hollandiae is quite variable in general morphology and widespread in Central and South America (Alston, Jermy, and Rankin, 1981
), and we suspect that it may represent a group of closely related species. Two morphotypes were included, one from Venezuela and one from Ecuador. Similarly, S. pallescens has two different growth forms, rosette (a so-called resurrection plant) and erect, and both forms were included in the analysis.
In addition to the ten articulate species, five possibly closely related nonarticulate species were chosen (S. australiensis, S. lyalli, S. myosurus, S. polymorpha, and S. sinensis). These possess at least one of the characteristics of articulate species (i.e., single, rarely few, basal megasporangium subtended by sterile sporophylls). Chromosome numbers are known for 24 of the species chosen for the analysis (Table 1), x = 8, 9, and 10 were represented, as well as 2n = 50–60 (S. martensii).
Outgroup
Two species of Isoetaceae, Isoetes melanopoda and I. lacustris, were included as outgroups. The sister-group relationship between Selaginellaceae and Isoetaceae has been confirmed in several studies, both morphological (Kenrick and Crane, 1997
) and molecular (Kranz and Huss, 1996
; Wikström and Kenrick, 1997
; Korall, Kenrick, and Therrien, 1999
).
DNA extraction, amplification, and sequencing
Total DNA was extracted from 34 specimens using the DNeasy Plant Mini Kit from Qiagen (Santa Clarita, California, USA). Total DNA from nine species, mainly Madagascan, was most kindly provided by S. Stefanovic (Department of Botany, University of Washington, Seattle, Washington, USA). The rbcL sequence of the rosette form of S. pallescens was kindly donated by James Therrien (Department of Botany, University of Kansas, Lawrence, Kansas, USA). DNA extracts were made from fresh material, from specimens dried in silica gel, or from herbarium specimens. The species included in the analysis and references to sequences taken from the literature are given in Table 1. Fragments corresponding to bases 18–1383 of the rbcL gene of Marchantia polymorpha (Ohyama et al., 1986
) were amplified using the polymerase chain reaction (PCR) and two primers (rbcL 1F corresponding to bases 1–17 and rbcL 1409R corresponding to bases 1384–1409) (for primer sequences see Korall, Kenrick, and Therrien, 1999
). Polymerase chain reaction was performed in 25-µL aliquots using Ready-To-Go PCR beads from Amersham Pharmacia Biotech (Uppsala, Sweden). The reactions were run in a Perkin-Elmer Thermal Cycler with one cycle of 95°C for 5 min and 30 cycles of 94°C for 30 sec, 50°C for 30 sec, and 72°C for 2 min. A second amplification, using product from the first PCR as template was occasionally necessary to obtain sufficient DNA for sequencing. Two different amplification strategies were used: high amounts of DNA template (15 µL) and only 15 PCR cycles, or nested PCR, where one internal primer was used in combination with one of the amplification primers. Nested PCR was the most successful method, and allowed us to obtain sequences from poor-quality or old material, such as S. pilifera collected in 1907. The Thermo Sequenase Fluorescent Sequencing Kit from Amersham Pharmacia Biotech (Uppsala, Sweden) was used to sequence double-stranded PCR products for the rbcL gene. Samples were electrophoresed on 6% Pharmacia Long Ranger acrylamide gels on an Amersham Pharmacia Biotech (Uppsala, Sweden) automated "ALF-express" sequencer. All species were sequenced in both directions using six different primers (Korall, Kenrick, and Therrien, 1999
). Sequences were assembled and edited using the Staden Package (Staden, 1996
). All sequences are deposited in EMBL.
A smaller partial PCR product was obtained for four species (S. willdenovii, S. myosurus, S. plana, and S. helvetica). In these cases an internal primer was used in combination with one of the amplification primers. The length of the resulting sequences varied between 677 and 985 base pairs.
Phylogenetic analysis
Visual alignment of the rbcL sequences was unproblematic because of the absence of insertions and deletions. The data matrix contained 1299 characters corresponding to bases 83–1382 of the rbcL gene of Marchantia polymorpha (Ohyama et al., 1986
). Parsimony analyses of the data were performed using PAUP* 4.0 (Swofford, 2000
). Analyses used the heuristic search option and the settings were random-sequence addition with 500 replicates, tree bisection-reconnection (TBR) branch swapping, collapse of zero length branches, and MULTREES on. An equal weighting scheme was employed with no transition-transversion bias (Albert and Mishler, 1992
). In all analyses, trees were rooted using both Isoetes species.
Support for individual clades was assessed using the decay index (Bremer, 1988
; Donoghue et al., 1992
) and bootstrap values (Felsenstein, 1985
). Decay indices were calculated using AutoDecay 4.0.2 (Eriksson, 1999
) and PAUP* 4.0 (Swofford, 2000
). PAUP* 4.0 settings used during decay analyses to find the tree length of constrained trees were: heuristic search with 200 replicates of random addition sequence, TBR branch swapping, collapse of zero length branches, random sequence addition with one replicate, and MULTREES off. Bootstrap values were calculated using PAUP* 4.0 by performing 30 000 replicates with the following options selected: heuristic search, TBR branch swapping, collapse of zero length branches, random sequence addition with one replicate, and MULTREES off.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
Clade B comprises only anisophyllous species (decay index 3; bootstrap 76%), and it contains several well-supported subclades (Fig. 1). There is strong support for a cosmopolitan clade (Asia, North America, Africa/Madagascar) of drought-adapted species (S. tamariscina/S. imbricata, decay index 43; bootstrap 100%). This predominantly xerophytic group is sister to a clade containing all other species: S. apoda/S. denticulata. Some of the subsequent nodes are weakly or moderately supported, leaving two strongly supported major nodes: A clade of southeast Asian and Australian species, S. longipinna/S. alopecuroides (decay index 4; bootstrap 97%), and a clade of European, Madagascan, and Asian species, S. plana/S. denticulata (decay index 17; bootstrap 100%). In addition, a poorly supported clade of Central and South American humid tropical species is found, S. pallescens/S. novae-hollandiae (decay index 1; bootstrap <50%). Species with dimorphic sporophylls (Heterostachys sensu Jermy, 1986
) form a polyphyletic assemblage. Neither the two geographic variants of S. novae-hollandiae nor the two morphologically different forms of S. pallescens were resolved as monophyletic.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Therrien and Haufler, 2000
). In previous systematic treatments, major divisions of the family have often been based on one or more morphological features, such as leaf dimorphism (isophylly/anisophylly), phyllotaxy, and growth form. The rbcL data suggest that several of these key characters have evolved iteratively, implying that these groups are polyphyletic or paraphyletic. It should be noted, however, that earlier classifications are in most cases not phylogenetically based systems and were thus not necessarily intended to reflect evolutionary relationships (exceptions include Soják, 1992
). Of the groups recognized by Jermy (1986)
, the rbcL tree indicates that Stachygynandrum and Heterostachys are polyphyletic. Whether Ericetorum is monophyletic or not cannot be deduced from this study. Of the groups recognized by Soják (1992)
the genera Selaginella (= subg. Selaginella of Jermy, 1986
) and Bryodesma (= subg. Tetragonostachys of Jermy, 1986
) are monophyletic, but the large genus Lycopodioides Boehm (= subgenera Ericetorum, Stachygynandrum, and Heterostachys of Jermy, 1986
) is paraphyletic to Bryodesma.
Perhaps the most surprising aspect of our analysis is the large number of new and well-supported groupings not previously recognized on morphological grounds. These clades range enormously in size from those containing perhaps a handful of species to others with many hundreds. Examples of these include S. remotifolia/S. fragilis, S. pygmaea/S. gracillima, S. sinensis/S. australiensis, S. tamariscina/S. imbricata, S. apoda/S. denticulata, S. longipinna/S. alopecuroides, and S. plana/S. denticulata. One of these groupings shares the distinctive ecological trait of being xerophytic (S. tamariscina/S. imbricata), but the other clades do not appear to be supported by any obvious characters. This may reflect a poor understanding of the comparative morphology, ecology or physiology of species in this large and diverse family. It may also be an artifact of limited sampling because the rbcL data sample 10% of living species diversity. There is certainly a need to reevaluate these species groups from other perspectives. The taxonomic status of various groups is discussed in more detail below.
New groups
Our previous molecular analysis (Korall, Kenrick, and Therrien, 1999
) identified a clade with a conspicuous morphological marker, which we have called the rhizophoric clade (Fig. 1). All species in this group bear a distinctive and unique root-like structure termed the rhizophore (Fig. 3C and F), which is lacking in the two species of subg. Selaginella. Another morphological character supporting the rhizophoric clade is the decussate arrangement of sporophylls. Bootstrap support for this clade is very low. Within the rhizophore bearing group, a basal dichotomy gives rise to two major clades: clade A (poorly supported) with both isophyllous and anisophyllous species and clade B (moderate support) with only anisophyllous species (Fig. 1). The exclusion of two species with exceptionally long branches (S. sinensis and S. australiensis), strengthens support for these three clades (see below "branch lengths, pseudogenes, and morphotypes").
|
, S. wallacei Hieron. in Webster and Steeves [1964]
), and we have observed this developmental pattern in an additional ten out of 50 species in this group. Observations of herbarium and living material show that dorsal rhizophores also occur in the resurrection plant S. lepidophylla and in S. myosurus.
The well-supported S. remotifolia/S. fragilis clade contains all of the articulate species sampled, except S. exaltata. In 1850, Spring incorporated all species with articulations into a widely recognized and predominantly South American group called Articulatae (Braun, 1865
; Hieronymus, 1901
; Walton and Alston, 1938
). Articulations are small swellings of the stem that occur below dichotomies. Somers (1978)
listed several additional morphological features found in this group. Distinguishing features include (1) dorsal rhizophores, (2) a single, basal megasporangium (rarely two) associated with several enlarged sterile sporophylls, (3) large megaspores, and (4) unique microsporangia. Monophyly of Articulatae is consistent with the strict consensus tree (Fig. 1), but has very low support. Although the articulations seem to be a synapomorphy of the group, many of the characteristics documented by Somers (1978)
have also been reported for nonarticulate species (e.g., the dorsal rhizophores). It seems likely, therefore, that a more detailed study of Articulatae and close relatives will show that some of the putative characteristics of the group are in fact synapomorphies of a series of more inclusive clades.
The South American articulate species (S. sericea/S. fragilis) form a well-supported subclade that is sister to a clade containing the only two Old World articulate species sampled (i.e., S. kraussiana, S. remotifolia; Somers, 1978
). These two species are very similar in morphology and growth form and have been considered geographic variants of a single species (S. kraussiana; Baker, 1884
). Differences in stele and leaf shape (P. Korall, personal observations; Hieronymus, 1901
; Somers, 1978
) together with 91 base-pair differences in the rbcL sequences support their treatment as distinct species. Whereas S. remotifolia is strictly confined to the Old World (parts of east and southeast Asia), the distribution of S. kraussiana spans both New World and Old World (parts of Africa, South America). Alston, Jermy, and Rankin (1981)
point out that the general assumption is that S. kraussiana has been introduced into South America by humans. In addition to S. kraussiana there is at least one other Articulatae species in Africa. Selaginella grallipes Alston differs markedly in gross morphology from the creeping S. kraussiana and S. remotifolia and more closely resembles the scandent South American S. exaltata (Moran and Smith, 2001
). Including S. grallipes in the analysis might help elucidate the weakly supported relationships between basal Articulatae species and close relatives.
The strongly supported S. pygmaea/S. gracillima clade does not possess any obvious unique morphological synapomorphies. Morphologically, the three Madagascan species (S. lyallii, S. moratii, S. polymorpha) and Ericetorum (S. gracillima, S. uliginosa, S. pygmaea) are quite dissimilar. Ericetorum contains small, isophyllous plants that inhabit temperate heathland in Australia and South Africa. Selaginella pygmaea and S. gracillima are annuals. In contrast, S. lyallii, S. polymorpha, and S. moratii are tall plants (up to 80 cm height; Stefanovic, Rakotondrainibe, and Badré, 1997
) with frondose branches in the upper part, anisophyllous, and tropical. The rbcL sequences of the three Madagascan species differ in only one or two base pairs, and although the species are very similar in morphology, Stefanovic, Rakotondrainibe, and Badré (1997)
consider them to be different species.
The very large pantropical S. apoda/S. denticulata clade contains almost half of the 62 species included in the analysis. All have dimorphic leaves and are members of either Stachygynandrum or Heterostachys. Most species are confined to the humid tropics. One exception is the sister species to the rest of the group, S. apoda, which grows in moist, temperate environments in North America. It is likely that this clade includes many unsampled species from the humid tropics of southeast Asia and South America. The restricted geographic distribution and the relatively small differences in branch lengths (Fig. 2) among species within some subclades is consistent with comparatively recent speciation events.
|
; Baker, 1883
; Hieronymus, 1901
; Thomas and Quansah, 1991
), but our study shows that isophylly has a complex evolutionary history. Anisophylly is characteristic of the subgenera Heterostachys and Stachygynandrum (Fig. 3D–F), whereas Tetragonostachys (Fig. 3A–C), Selaginella, and Ericetorum are isophyllous (Fig. 2). Outgroup comparison with Isoetaceae indicates that leaves of one size is the plesiomorphic state for the family. The isophyllous states of Tetragonostachys and Ericetorum are most parsimoniously interpreted as reversals, but because most basal nodes have low support the possibility that isophylly is a retained plesiomorphic state and that anisophylly has originated several times cannot be ruled out.
A resupinate strobilus, one that exhibits sporophyll dimorphism with the smaller sporophylls borne in the same plane as the larger vegetative leaves, was used to characterize the polyphyletic subg. Heterostachys. Five resupinate species were included in the analysis (Fig. 1). Although they all appear in clade B, the rbcL data indicate that the evolution of resupinate strobili involved parallelisms and/or reversals. In a future study, it would be interesting to include S. bemarahensis S. Stefanovic and Rakotondrainibe and S. marinii S. Stefanovic and Rakotondrainibe. These species have resupinate strobili, but the rhizophores are dorsal (Stefanovic and Rakotondrainibe, 1996
), which implies a phylogenetic position closer to Articulatae and Tetragonostachys.
Stele morphology has been used as a basis for recognizing groups within Selaginellaceae (Hieronymus, 1901
), and this characteristic is documented for most species (Fig. 2, Table 1). It is clear from Fig. 2 that the bistelic condition has evolved at least once among the articulate species and that the polystelic condition has arisen independently at least four times, in S. lyallii and S. polymorpha, in S. articulata, in S. acanthostachys and in S. willdenovii and S. plana. A solenostele is reported as characteristic of Ericetorum (Jermy, 1986
).
Several studies have shown that distinctive patterns of sporangial arrangement exist in Selaginella (e.g., Horner and Arnott, 1963
; Fraile and Riba, 1981
; Quansah, 1988
). Major trends that have been recognized include the organization of sporangia into rows (two rows of microsporangia and two rows of megasporangia) and the restriction of megasporangia to a zone at the base of the strobilus. The only pattern that seems always to be consistent within species is the presence of a single (rarely two) basal megasporangium (Fraile and Riba, 1981
; Quansah, 1988
). This arrangement has been used as a feature unifying the articulate species (Somers, 1978
), but is also found in other taxa. This was recognized by Hieronymus (1901)
who included sporangial arrangement as a diagnostic character in his classification by dividing the large subgenus Heterophyllum (corresponding to Stachygynandrum and Heterostachys together) into Oligomacrosporangiatae (few megasporangia per strobilus) and Pleiomacrosporangiatae (many megasporangia per strobilus). Included in our analysis, in addition to the articulate species, are S. myosurus, S. australiensis, S. lyallii, and S. polymorpha and S. sinensis. Our results show that species with a single basal megasporangium are all found in clade A, but that this characteristic does not uniquely circumscribe a distinct clade.
Branch lengths, pseudogenes, and morphotypes
With 566 phylogenetically informative characters for 62 taxa, the rbcL gene has a very high level of variation in Selaginellaceae, and this is distributed unequally within the family (Fig. 2). The branches in clade A are longer (more substitutions) than those in clade B, and this phenomenon has the potential of causing analytical problems. Long branches can cause spurious groupings (Felsenstein, 1978
), but this is mainly a problem when they are connected by short internodes. This pattern is not seen in our tree (Fig. 2) in which internodes are of a similar length to terminal branches for the most part. The branches leading to S. sinensis and S. australiensis are exceptional even for Selaginellaceae and have led to unstable topologies. Under some ingroup/outgroup combinations, these two species have even grouped with seed plants in the Gnetales, also a group with an exceptionally long branch in rbcL data sets. Because of this problem we acknowledge the possibility of spurious groupings, but we suspect that this is not a major problem because many groupings within the tree are comparatively stable and well supported, both in terms of bootstrap support and decay index. Furthermore, if S. sinensis and S. australiensis are excluded from the analyses, the same topology is found and support for the basal nodes increases. The support for the rhizophoric clade, clade A, and clade B, rise from bootstrap values/decay indices of <50%/2 to 94%/12, <50%/2 to 85%/7, and <50%/3 to 91%/4, respectively. In view of the branch lengths, it is particularly important to test the rbcL gene tree through the development of a parallel data set based on additional genes.
Our laboratory data indicate the possible existence within Selaginellaceae of a pseudogene that is shorter than the functional rbcL gene. Double PCR products were amplified in six of the included species (S. acanthostachys, S. arizonica, S. novae-hollandiae [from Venezuela], S. remotifolia, S. rupincola, S. sellowii), and multiple products were observed in one species (S. plana). Pseudogenes of rbcL have previously been found in both chlorophyllous and parasitic angiosperms and in red algae (see Sennblad, Endress, and Bremer, 1998
, and references therein). Because no attempts to sequence the shorter fragments have been made, we cannot exclude the possibility that the fragments are amplifications of unrelated parts of the genome.
The rbcL sequences revealed problems in the species delimitation of S. novae-hollandiae. Two morphotypes of S. novae-hollandiae were included in the analysis. These specimens collected from Venezuela and Ecuador were not resolved as a monophyletic group (Fig. 1) and the rbcL sequences differ in 27 base pairs. Analyzing the data set with a constraint that forced the two taxa together produced trees five steps longer than the most parsimonious trees. These results are consistent with the presence of more than one species in S. novae-hollandiae. Similarly, the rbcL data show that the rosette and erect forms of S. pallescens differ in nine base pairs and that they are closely related but not a monophyletic group (Fig. 1). In this case, a constraint yields trees only two steps longer than the most parsimonious trees. This might also constitute grounds for considering the separation of rosette and erect forms at the species level. Interspecific variation in the rbcL gene of Selaginellaceae has not previously been examined, but studies on ferns show that within-species variation is low: only 0–2 base-pair differences are commonly found between individuals considered to belong to the same species (Hauk, 1995
; Yatabe, Takamiya, and Murakami, 1998
). Our results indicate that species diversity may be underestimated in the family and reflect the need of a thorough revision of the family at the species level.
Origins of Selaginellaceae
The earliest fossil evidence of Selaginellaceae comes from the tropical wetland floras of the Carboniferous Period (Visean Epoch, 333–350 million years ago [Ma]; Rowe, 1988
). By the Late Carboniferous (290–323 Ma) branching stems that bore minute leaves of two distinct sizes were widespread in coal measure floras (Thomas, 1992, 1997
). These observations demonstrate that the characteristic of leaf dimorphism appeared early on and that the family was an important component of the humid tropical floras of the Late Paleozoic. Selaginellaceae are predominantly forest-floor-dwelling species, and the possession of leaves of two distinct sizes is most probably an adaptation to poor light quality (Hébant and Lee, 1984
). The evolution of this characteristic broadly coincides with earliest evidence for closed-canopy forest, which was caused by the rise of tree ferns during the Westphalian, some 303–320 Ma (DiMichele et al., 1992
). The ecological association between Selaginellaceae and coal measure forests and the similarity of modern species to fossils raises the possibility that some living tropical clades may be relicts of these ancient Carboniferous groups (Korall, Kenrick, and Therrien, 1999
).
If elements of Selaginellaceae have persisted in tropical wetlands since the Carboniferous, we would expect to resolve species from the humid tropics as a basal grade or clade within the family, which would be paraphyletic or basal to xerophytic and temperate clades. The results of our molecular analysis do not, however, support this interpretation. Results document two major clades of predominantly humid tropical species, the S. pallescens/S. denticulata clade and a clade containing Articulatae and its sister species, S. myosurus. The S. pallescens/S. denticulata clade is further divided into strongly supported Southeast Asian–Australian (S. longipinna/S. alopecuroides) and weakly supported neotropical elements (S. pallescens/S. novae-hollandiae). Several of the basal clades in Selaginellaceae are either temperate, xerophytic, or contain significant proportions of species that fit these categories. Subgenus Selaginella contains the temperate/arctic alpine S. selaginoides. In clade A, the S. pygmaea/S. australiensis clade is probably temperate, and the Tetragonostachys plus S. lepidophylla clade is xerophytic. Likewise, in clade B, the S. tamariscina/S. imbricata clade contains drought-adapted species, and the next branch up is S. apoda, a temperate North American species. These data show a comparatively derived position for both major clades of humid tropical species. The cladogram topology therefore indicates that species and species groups of the modern humid tropics are unlikely to be relicts of Carboniferous cladogenesis, but are probably of more recent origin.
How can this interpretation be reconciled with the early appearance of species with dimorphic leaves in the tropical wetlands of the Carboniferous Period? The relationships of the early fossil species are currently very poorly understood. This is mainly because many are compression fossils, and they do not contain much information on anatomical or reproductive structures. They are therefore relatively information poor. It is possible that some of these early species would branch from the stem group of the rhizophoric clade or perhaps even from the Selaginellaceae stem group. In either case, these species would not be expected to possess the distinctive rhizophore, and in this context it is interesting to note that rhizophores have not been observed in any Paleozoic fossils.
Multiple origins of resurrection plants
Although most species diversity in Selaginellaceae occurs in primary tropical rainforest, a substantial number of species are able to withstand seasonal drought or even live in very arid parts of the world. The moss-like xerophytes (Fig. 3A–C) in Tetragonostachys have been shown to be a monophyletic group (Korall, Kenrick, and Therrien, 1999
; Therrien, Haufler, and Korall, 1999
; Therrien and Haufler, 2000
). Physiologically, species in this group are able to withstand prolonged desiccation, and morphologically they exhibit adaptations to reducing water loss by possessing small leaves with thick cuticles. In Stachygynandrum, at least ten species have a rosette of branches and during drought the branches curl inwards, forming a ball. These are called "resurrection plants" and the name alludes to their ability to revive rapidly by uncurling their branches to reform the rosette habit under more favorable moisture conditions. We included four of these ten species in our analysis in order to investigate their relationships to each other and to Tetragonostachys. Results show that resurrection plants have evolved at least three times in Selaginellaceae. The resurrection plant S. lepidophylla is sister to the moss-like xerophytes of the Tetragonostachys clade. Selaginella lepidophylla also has the distinctive rhizophore development that characterizes the more inclusive dorsal rhizophoric clade. The resurrection plants S. tamariscina (eastern Asia) and S. pilifera (southern USA, Mexico) are found in a strongly supported clade with other non-resurrection Asian (S. stauntoniana) and Madagascan species (S. digitata, S. helioclada, S. imbricata). All species in this clade grow in seasonally dry areas. In contrast to other xerophytes, the rosette form of the resurrection plant S. pallescens is nested in a humid tropical clade. The rbcL data therefore support diverse relationships for resurrection plants, indicating that this response to arid conditions has evolved iteratively in Selaginellaceae. Incurling of branches in a less pronounced fashion is also found in other drought-adapted species such as some species of Tetragonostachys (e.g., S. arizonica; Tryon, 1955
; see Fig. 3A–C), and the non-rosette-forming species of the S. tamariscina/S. imbricata clade (S. stauntoniana and S. imbricata; P. Korall, personal observations; S. digitata and S. helioclada; Stefanovic, Rakotondrainibe, and Badré, 1997
). There is evidence therefore in the morphology of other species that the resurrection plants' adaptations to aridity may just be an extreme example of a more general but less dramatic trend of branch incurling that is common in the family.
|
|
FOOTNOTES |
|---|
4 Author for reprint requests (petra.korall{at}botan.su.se
)
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Alston A. H. G. A. C. Jermy J. M. Rankin 1981 The genus Selaginella in tropical South America. Bulletin of the British Museum (Natural History) Botany Series 9: 233-330
Baker J. G. 1883 A synopsis of the genus Selaginella. Journal of Botany 21: 1-5, 42–46, 80–84, 97–100, 141–145, 210–213, 240–244
Baker J. G. 1884 A synopsis of the genus Selaginella. Journal of Botany 22: 23-26, 86–90, 110–113, 243–247, 275–278, 295–300, 373–377
Braun A. 1857 Appendix plantarum novarum et minus cognitarum. In Horto regio botanico Berolinensi coluntur: 12–22
Braun A. 1865 Über die Gattung Selaginella. (Also titled: Beiträge zur Kenntniss der Gattung Selaginella). Monatsberichte der Koniglich preussischen Akademie der Wissenschaften zu Berlin 30: 185-209
Bremer K. 1988 The limits of amino acid sequence data in angiosperm phylogenetic reconstruction. Evolution 42: 795-803[CrossRef][ISI]
DiMichele W. A. R. W. Hook R. Beerbower J. A. Boy R. A. Gastaldo N. Hotton III T. M. Phillips S. E. Scheckler W. A. Shear H.-D. Sues 1992 Paleozoic terrestrial ecosystems. In A. K. Behrensmeyer, J. D. Damuth, W. A. DiMichele, R. Potts, H.-D. Sues, and S. L. Wing [eds.], Terrestrial ecosystems through time: evolutionary paleoecology of terrestrial plants and animals, 205–325. University of Chicago Press, Chicago, Illinois, USA
Donoghue M. J. R. G. Olmstead J. F. Smith J. D. Palmer 1992 Phylogenetic relationships of Dipsacales based on rbcL sequences. Annals of the Missouri Botanical Garden 79: 333-345
Eriksson T. 1999 AutoDecay version 4.0.2. Bergius Foundation, Royal Swedish Academy of Sciences, Stockholm, Sweden
Felsenstein J. 1978 Cases in which parsimony or compatibility methods will be positively misleading. Systematic Zoology 27: 401-410[CrossRef][ISI]
Felsenstein J. 1985 Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783-791[CrossRef][ISI]
Fraile E. R. Riba 1981 Distribucion esporangial en estrobilos de especies de Selaginella. Boletin de la Sociedad Botanica de Mexico 41: 33-40
Hauk W. D. 1995 A molecular assessment of relationships among cryptic species of Botrychium subgenus Botrychium (Ophioglossaceae). American Fern Journal 85: 375-394[CrossRef][ISI]
Hébant C. D. W. Lee 1984 Ultrastructural basis and developmental control of blue iridescence in Selaginella leaves. American Journal of Botany 71: 216-219[CrossRef][ISI]
Hieronymus G. 1901 Selaginellaceae. In A. Engler and K. Prantl [eds.], Die Natürlichen Pflanzenfamilien, vol. 1, part. 4, 621–716. W. Engelmann, Leipzig, Germany
Horner H. T. H. J. Arnott 1963 Sporangial arrangement in North American species of Selaginella. Botanical Gazette 124: 371-383[CrossRef]
Jermy A. C. 1986 Subgeneric names in Selaginella. Fern Gazette 13: 117-118
Jermy A. C. 1990 Selaginellaceae. In K. Kubitzki [ed.], The families and genera of vascular plants, vol. 1, 39–45. K. U. Kramer and P. S. Green [eds.], Pteridophytes and Gymnosperms. Springer, Berlin, Germany
Jermy A. C. K. Jones C. Colden 1967 Cytomorphological variation in Selaginella. Journal of the Linnean Society of London, Botany 60: 147-158
Kenrick P. P. R. Crane 1997 The origin and early diversification of land plants: a cladistic study. Smithsonian Institution Press, Washington, D.C., USA
Korall P. P. Kenrick J. P. Therrien 1999 Phylogeny of Selaginellaceae: evaluation of generic/subgeneric relationships based on rbcL gene sequences. International Journal of Plant Sciences 160: 585-594[CrossRef]
Kranz H. D. V. A. R. Huss 1996 Molecular evolution of pteridophytes and their relationships to seed plants: evidence from complete 18S rRNA gene sequences. Plant Systematics and Evolution 202: 1-11[CrossRef][ISI]
Kuriachan P. I. 1963 Cytology of the genus Selaginella. Cytologia 28: 376-380[ISI]
Löve A. D. Löve 1976 IOPB chromosome number reports LIII. Taxon 25: 483-500
Manhart J. R. 1994 Phylogenetic analysis of green plant rbcL sequences. Molecular Phylogenetics and Evolution 3: 114-127[CrossRef][Medline]
Manton I. 1950 Problems of cytology and evolution in the Pteridophyta. Cambridge University Press, Cambridge, UK
Mickel J. T. R. I. Hellwig 1969 Actino-plectostely, a complex new stelar pattern in Selaginella. American Fern Journal 59: 123-134[CrossRef]
Moran R. C. A. R. Smith 2001 Phytogeographic relationships between neotropical and African-Madagascan pteridophytes. Brittonia 53: 304-351[ISI]
Musselman L. J. K. D. Heafner 1997 Monograph of southeastern Isoetes. Association of Southeastern Biologists Bulletin 44: 119 (Abstract)
Ohyama K. H. Fukuzawa T. Kohchi H. Shirai T. Sano S. Sano K. Umesono Y. Shiki M. Takeuchi Z. Chang S. Aota H. Inokuchi H. Ozeki 1986 Chloroplast gene organization deduced from the complete sequence of the liverwort Marchantia polymorpha chloroplast DNA. Nature 322: 572-574[CrossRef][ISI]
Quansah N. 1988 Sporangial distribution patterns in the strobili of African and Madagascan Selaginella. Annals of Botany 61: 243-247
Reed C. F. 1965–1966 Index Selaginellarum. Memórias da Sociedade Broteriana 18: 1-287
Rowe N. P. 1988 A herbaceous lycophyte from the Lower Carboniferous Drybrook Sandstone of the Forest of Dean, Gloucestershire. Palaeontology 31: 69-83[ISI]
Sennblad B. M. E. Endress B. Bremer 1998 Morphology and molecular data in phylogenetic fraternity—the tribe Wrightieae (Apocynaceae) revisited. American Journal of Botany 85: 1143-1158[Abstract]
Soják J. 1992 Generische Problematik der Selaginellaceae. Preslia 64: 151-158
Somers P. 1978 A systematic survey of the Articulatae series of the genus Selaginella and monographic treatment of the S. sulcata group (sensu str.). Ph.D. dissertation, University of Tennessee, Knoxville, Tennessee, USA
Spring A. F. 1850 Monographie de la famille des Lycopodiacées, Second partie. Mémoires de l'académie Royale des Sciences, des Lettres et des Beaux-arts de Belgique 24: 1-358
Staden R. 1996 The Staden sequence analysis package. Molecular Biotechnology 5: 233-241[ISI][Medline]
Stefanovic S. F. Rakotondrainibe 1996 New taxa and a new rank of Selaginella (Selaginellaceae) from Madagascar and the Comoros. Novon 6: 203-209
Stefanovic S. F. Rakotondrainibe F. Badré 1997 Famille 14. Sélaginellacées. In P. Morat [ed.], Flore de Madagascar et des Comores, 68. Muséum National d'Histoire Naturelle, Paris, France
Swofford D. 2000 PAUP*: phylogenetic analysis using parsimony (and other methods) beta version 4.0b4a. Sinauer, Sunderland, Massachusetts, USA
Takamiya M. 1993 Comparative karyomorphology and interrelationships of Selaginella in Japan. Journal of Plant Research 106: 149-166[CrossRef][ISI]
Tchermak-Woes E. R. Dolezal-Janish 1959 Über die karyologische Anatomie einiger Pteridophyten sowie auffallende Unterschiede im Kernvolumen bei Cyrtomium falcatum. Österreichische botanische Zeitschrift 116: 315-324
Therrien J. P. C. H. Haufler 2000 Phylogeny and biogeography of Selaginella subg. Tetragonostachys based on nuclear ribosomal ITS sequence data. American Journal of Botany 87: 98 (Abstract)
Therrien J. P. C. H. Haufler P. Korall 1999 Phylogeny and biogeography of Selaginella subg. Tetragonostachys. XVI International Botanical Congress, abstracts 120. Missouri Botanical Garden, St. Louis, Missouri, USA
Thomas B. A. 1992 Paleozoic herbaceous lycopsids and the beginnings of extant Lycopodium sens. lat. and Selaginella sens. lat. Annals of the Missouri Botanical Garden 79: 623-631[CrossRef][ISI]
Thomas B. A. 1997 Upper Carboniferous herbaceous lycopsids. Review of Palaeobotany and Palynology 95: 129-153[CrossRef]
Thomas B. A. N. Quansah 1991 The palaeobotanical case for dividing Selaginella (Selaginellaceae: Pteridophyta). Fern Gazette 14: 59-64
Tryon R. M. 1955 Selaginella rupestris and its allies. Annals of the Missouri Botanical Garden 42: 1-99
Walton J. A. H. G. Alston 1938 Lycopodiinae. In F. Verdoorn [ed.], Manual of pteridology, 500–506. Martinus Nijhoff, The Hague, The Netherlands
Webster T. R. T. A. Steeves 1963 Morphology and development of the root of Selaginella densa Rydb. Phytomorphology 13: 367-376
Webster T. R. T. A. Steeves 1964 Developmental morphology of the root of Selaginella kraussiana A. Br. and Selaginella wallacei Hieron. Canadian Journal of Botany 42: 1665-1676
Wikström N. P. Kenrick 1997 Phylogeny of Lycopodiaceae (Lycopsida) and the relationships of Phylloglossum drummondii Kunze based on rbcL sequences. International Journal of Plant Sciences 158: 862-871[CrossRef]
Yatabe Y. M. Takamiya N. Murakami 1998 Variation in the rbcL sequence of Stegnogramma pozoi subsp. mollissima (Thelypteridaceae) in Japan. Journal of Plant Research 111: 557-564[CrossRef][ISI]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
80% are thicker. Support values with arrows pointing to nodes denote values on these nodes when S. sinensis and S. australiensis are excluded from the analysis. All species not labeled with subgenus (first column) are classified in subg. Stachygynandrum. Note that series Articulatae is found within Stachygynandrum. 1The rosette form of S. pallescens. 2The erect form of S. pallescens. 3Specimen collected in Venezuela. 4Specimen collected in Ecuador
