Donald R. Farrar

A community-derived classification for extant lycophytes and ferns

Phylogeny has long informed pteridophyte classification. As our ability to infer evolutionary trees has improved, classifications aimed at recognizing natural groups have become increasingly predictive and stable. Here, we provide a modern, comprehensive classification for lycophytes and ferns, down to the genus level, utilizing a community‐based approach. We use monophyly as the primary criterion for the recognition of taxa, but also aim to preserve existing taxa and circumscriptions that are both widely accepted and consistent with our understanding of pteridophyte phylogeny. In total, this classification treats an estimated 11 916 species in 337 genera, 51 families, 14 orders, and two classes. This classification is not intended as the final word on lycophyte and fern taxonomy, but rather a summary statement of current hypotheses, derived from the best available data and shaped by those most familiar with the plants in question. We hope that it will serve as a resource for those wanting references to the recent literature on pteridophyte phylogeny and classification, a framework for guiding future investigations, and a stimulus to further discourse.

Significance of gametophyte form in long-distance colonization by tropical, epiphytic ferns

Gametophyte morphology of tropical epiphytic ferns may confer an advantage for establishment on islands. Most tropical, epiphytic ferns belong to five families: Hymenophyllaceae, Grammitidaceae, Vittariaceae, Polypodiaceae, and Elaphoglossaceae. Gametophytes of these families are long-lived and clone-forming. In addition, most Hymenophyllaceae, Grammitidaceae, and Vittariaceae produce dispersible gemmae. Each of these characteristics increases opportunity for outbreeding, and when island floras are statistically compared with floras of adjacent mainlands, island floras are found to be rich in epiphytic species possessing gemmae (Hymenophyllaceae, Grammitidaceae, and Vittariaceae), and depauperate in epiphytic species lacking gemmae (Polypodiaceae and Elaphoglossaceae). We propose that gametophytic gemmae significantly aid long-distance colonization of outbreeding species because gemmae 1) allow gametophytes to exploit available niches through dispersal of gemmae, and through clonal expansion and persistence of the resulting gametophyte, and 2) facilitate sexual reproduction by providing the opportunity for sperm and antheridiogen transfer when gametophytes are distant, and by providing a new source of tissue for antheridia formation.

Species and Evolution in Asexually Reproducing Independent Fern Gametophytes

Vegetative reproduction by the gametophyte generation has allowed a number of fern species to persist beyond their normal geographic range. In at least two of these, Trichomanes and Vittaria in the eastern United States, the sporophyte stage of the life cycle has been eliminated completely, perhaps for ten million years or more. Despite this long period of reliance on vegetative reproduction, genetic diversity in the Appalachian Vittaria gametophytes, as measured by starch gel enzyme electrophoresis, remains comparable to that of sexual plants, although it is largely partitioned into monomorphic populations. Widely separated populations are not particularly di- vergent; the taxon as a whole appears equally as cohesive as sexual species, suggesting that factors other than gene flow are responsible for long term maintenance of species integrity. The Appalachian Vittaria gametophytes and similarly derived taxa merit recognition as distinct species. Sexually reproducing pteridophytes alter- nate a diploid, spore-producing plant, the spo- rophyte, with a haploid, gamete-producing plant, the gametophyte. In most taxa, the ga- metophyte generation is short-lived and incon- spicuous relative to the sporophyte plants, which are generally perennial. Life spans of

Significance of Form in Fern Gametophytes: Clonal, Gemmiferous Gametophytes of Callistopteris baueriana (Hymenophyllaceae)

Gametophytes of Callistopteris baueriana (Endl.) Copeland exhibit an unreported and unique morphology for the Hymenophyllaceae, combining characters that traditionally have been thought to be specific to either Hymenophyllum s. 1. or Trichomanes s. 1. We describe here the mature gametophyte morphology of C. baueriana from Hawaii. To explore gametophyte character evolution in the Hymenophyllaceae, gametophyte characters of selected genera within the family were analyzed phylogenetically. In Hawaii, gametophytes of C. baueriana were associated with sporophyte production or grew independently, often in extensive, dense mats. Well-developed sporophytes occurred at high elevations. At drier midelevations, sporophytes were dwarfed. Sporophytes were absent at the lowest elevations and gametophytes grew independently in large bryophyte-like mats. The sporophytes of C. baueriana seemed unable to survive and reproduce in dry conditions, whereas gametophytes survived and maintained their presence by vegetative growth and the production of gemmae. The gametophytes had a branched, ribbon-like thallus and were long-lived and clone-forming. Generally, a meristematic branch grew along the substrate while subordinate branches grew perpendicularly to the substrate and terminated in the production of gemmae. The gemmae were largely uniseriate, spindle-shaped, and branched or unbranched. Each sat between apical protuberances of a bulbous gemmifer. At maturity the gemma abscised from the gemmifer, possibly with the assistance of pressure from the apical protuberances and swollen dorsal side of the gemmifer. Antheridia, produced marginally on small thalli or on gemmae, were composed of at least five cells, and averaged 76 μm in diameter. Archegonia were formed ventrally on lateral pads and had necks composed of four tiers of cells. Parsimony analysis of gametophyte characters of selected genera within the Hymenophyllaceae resulted in a single, most parsimonious tree, with Callistopteris sister to Trichomanes s. s. and Vandenboschia, the two most derived taxa in the analysis. Mecodium was sister to the clade containing Callistopteris, Trichomanes s. s. and Vandenboschia. Hymenophyllum s. s. was the basal-most taxon. The tree was consistent with hypotheses of gametophyte character evolution within the Hymenophyllaceae, proposing that gametophyte growth habit composed partly or wholly of filaments, one-dimensional gemmae, elongate, flask-shaped gemmifers, archegonia with necks of four tiers of cells, and small antheridia of five jacket cells are derived characters within the family. A thalloid growth habit, two-dimensional gemmae, bulbous gemmifers, archegonial necks with six to 10 tiers of cells, and large complex antheridia represent ancestral characters.

Phylogeny of the Vittariaceae: Convergent Simplification Leads to a Polyphyletic Vittaria

The fern family Vittariaceae contains nearly 100 species of tropical epiphytes with simple leaf morphology. Different interpretations of the limited number of morphological char- acters has led to controversy in the generic and subgeneric taxonomy of the family. A 1380 bp fragment of the chloroplast-encoded rbcL gene was amplified and sequenced from species repre- senting the genera and subgenera of the family. Parsimony analysis of the sequence data resulted in two most parsimonious trees which differ only in the position of the monotypic Ananthacorus. Each tree has two main clades which separate in a basal dichotomy. In the first principal clade, Ananthacorus appears either as sister to a clade containing Antrophyum ensiforme and A. bor- yanum or sister to a clade containing Vittaria lineata, V graminifolia, V. dimorpha, and V isoeti- folia. The other principal clade is made up primarily of species divided into two sister groups. One of these groups contains only Old-World species of Vittaria while the other contains New- World species of Vittaria corresponding to Benedicts subgenus Radiovittaria with Hecistopteris sister to the latter clade. The rbcL topology is congruent with the character-state distributions for several morphological characters: Spore shape, paraphysis terminal cell shape, gametophyte gem- ma development, and leaf arrangement on the rhizome. The fern family Vittariaceae contains approximately 100 species of tropical epiphytes with distinctive morphology (Tryon and Tryon, 1982). The leaves are entire in all but one species, and the sporangia, without indusia, occur along veins. This distinctive but simple morphology makes the familial limits uncontroversial but provides few characters useful for intrafamilial taxonomy. Both the paucity of characters and disagreement about their interpretation has led to disagreement about generic circumscription. The number of genera rang- es from five to ten depending on which characters are considered significant (Benedict, 1911; Copeland, 1947; Tryon and Tryon, 1982). The genera of the Vittariaceae are defined by most authors using a combi- nation of venation and soriation (see Fig. 1). In Vittaria J.E. Smith, the veins enclose a single rank of areolae between the costa and the margin of the leaf, and a single line of sporangia follows the commisural marginal vein (Fig. lb, c). Antrophyum Kaulf. has several ranks of areolae, and several soral lines lying over the veins between the costa and the margin. A third genus, Mono- gramma Schkuhr, is composed of exceedingly small plants (laminae less than 1 mm wide) with either a single vein or a simple vein loop, and with the sporangia restricted to one margin (Fig. lg). Most species are included in these three genera; the three remaining genera, Ananthacorus Underw. & Maxon, Anetium Splitg., and Hecistopteris J. Smith, are monotypic. The largest genus, Vittaria, is a pantropical group of about 50 species. The leaves are lanceolate to long-linear, with venation consisting of a midrib and

Antheridiogen production and response in Polypodiaceae species.

Antheridiogen chemicals secreted by living fern gametophytes have been shown to influence production of male gametangia and thus mating systems in a large number of terrestrial fern species. Antheridiogens have not previously been thought to be prevalent in the Polypodiaceae, a large family composed mostly of tropical epiphytes. This study presents bioassay methods more sensitive than previously used to detect antheridiogen and demonstrates that antheridiogens are also operative in the Polypodiaceae and in epiphytic species. Seven species in six genera (Campyloneurum angustifolium, C. phyllitidis, Lepisorus thunbergianus, Microgramma heterophylla, Phlebodium aureum, Phymatosorus scolopendria, and Polypodium pellucidum) were tested for the presence of an antheridiogen system. All species tested except P. aureum were induced to produce antheridia precociously by their own antheridiogen and by that of Pteridium aquilinum (APt). Phlebodium aureum responded to APt and promoted antheridium formation in Onoclea sensibilis but did not respond to its own antheridiogen. Spores of all species except P. aureum were induced to germinate in darkness by antheridiogen of the same species and by APt and to form antheridia in the dark, further enhancing the possibility of intergametophytic mating.

Spore Retention and Release from Overwintering Fern Fronds

The need for ecological data on ferns is becoming increasingly apparent as more effort is directed toward an understanding of the significance of their morphological and physiological diversity (Wagner, 1973). If detailed ecological studies of fern sporophytes have to date been too few, such studies on the gametophyte generation are almost non-existent. The reproductive cycle of ferns has been known for over a century, and more than a thousand articles on fern gametophytes have been published, half of these in the last quarter century (Miller, 1968; Nif, 1975; Nayar & Kaur, 1971). However, nearly all data on gametophyte growth and sexual reproduction have been based on laboratory observations. Several factors have contributed to this paucity of information on gametophyte ecology, but perhaps the most important has been a widely held notion that gametophytes cannot be found in nature, or if found, cannot be identified. Several recent studies have indicated to the contrary, that in situ gametophyte studies not only are feasible, but that they are essential for the integration of existing laboratory data into studies on the natural history of ferns (Cousens, 1973; Holbrook-Walker & Lloyd, 1973; Lloyd, 1974; Farrar & Gooch, 1975). To investigate further the feasibility of studying fern gametophytes in nature, we have begun a long-term observational study of fern reproduction in Woodman Hollow, a relatively isolated canyon in central Iowa, in which 13 species and 11 genera of ferns occur (see Table 1). This study is designed to answer the following questions. When are spores available for germination? When and where does reproduction occur and how is it influenced by microand macroclimates? When and by what breeding systems are sporophytes produced? Does sexual reproduction occur in nature on a regular basis for all species? Results of the first year of study (Farrar & Gooch, 1975) indicate that the data needed to answer these and other questions will be forthcoming. Here we report some unexpected data relevant to the question of when spores are available for germination. Observations made on the time of spore maturation and first release during the growing season gave results which were similar to those of Hill and Wagner (1974) for pteridophytes in Michigan. Differences found in the two studies were no greater than might be expected due to differences in latitude, climate, habitat, and seasonal variation. Our observations also support their estimate that most spores of a given species are released during a period of about two weeks. However, a two week period of maximum release, if taken as a guide to the duration of spore release, may be very misleading. Our observations at Woodman Hollow indicate that for most species, significant quantities of spores are retained on the fronds after the initial release period and may be dispersed during a much longer period. Only in Botrychium virginianum and Osmunda claytoniana were essentially all

Gemmae: A Role in Sexual Reproduction in the Fern Genus Vittaria

Gemmae are generally defined as vegetative propagules. In the shoestring ferns, Vittaria, gemmae grown in the presence of mature gametophyte plants or on medium containing gibberellic acid produce antheridia in lieu of vegetative growth. This suggests that antheridial differentiation in Vittaria is controlled by a chemical antheridogen system similar to those described in other fern genera. In natural populations of Vittaria gametophytes composed primarily of long-lived individuals, gemmae may provide the only source of tissue susceptible to antheridogen action and may have evolved in response to that condition.

trnL-F is a powerful marker for DNA identification of field vittarioid gametophytes (Pteridaceae)

BACKGROUND AND AIMS The gametophyte phase of ferns plays an important role in habitat selection, dispersal, adaptation and evolution. However, ecological studies on fern gametophytes have been impeded due to the difficulty of species identification of free-living gametophytes. DNA barcoding provides an alternative approach to identifying fern gametophytes but is rarely applied to field studies. In this study, an example of field vittarioid gametophyte identification using DNA barcoding, which has not been done before, is given. METHODS A combination of distance-based and tree-based approaches was performed to evaluate the discriminating power of three candidate barcodes (matK, rbcL and trnL-F) on 16 vittarioid sporophytes. Sequences of the trnL-F region were generated from 15 fern gametophyte populations by tissue-direct PCR and were compared against the sporophyte dataset, using BLAST. KEY RESULTS trnL-F earns highest primer universality and discriminatory ability scores, whereas PCR success rates were very low for matK and rbcL regions (10·8 % and 41·3 %, respectively). BLAST analyses showed that all the sampled field gametophytes could be successfully identified to species level. Three gametophyte populations were also discovered to be living beyond the known occurrence of their sporophyte counterparts. CONCLUSIONS This study demonstrates that DNA barcoding (i.e. reference databasing, tissue-direct PCR and molecular analysis), especially the trnL-F region, is an efficient tool to identify field gametophytes, and has considerable potential in exploring the ecology of fern gametophytes.

The sporophyte-less filmy fern of eastern North America Trichomanes intricatum (Hymenophyllaceae) has the chloroplast genome of an Asian species.

Trichomanes intricatum, the sporophyte-less filmy fern of the eastern United States, has been considered to be a species whose sporophyte generation has become extinct or is possibly still present among the many species of Trichomanes s.l. in the new world tropics but unable to grow in a temperate climate. A close relationship to Asian species has heretofore not been considered. Comparison of rbcL and rps4-trnS sequences to species of Trichomanes s.l. reveals that T. intricatum shares its chloroplast genome with Crepidomanes schmidtianum of eastern Asia. Because C. schmidtianum is a sterile triploid and the ploidy level of T. intricatum is unknown, several scenarios leading to their sharing of these maternally inherited genes must be explored.