A Generic Classification of the Thelypteridaceae Author Fawcett, Susan Author Smith, Alan R. text 2021 BRIT Press Fort Worth, Texas, USA book 10.17348/jbrit.v15.i2.1206 14111022 CHRISTELLA Christella H. Lév., Flore de Kouy-tchéou 472. 1915.— TYPE : Christella parasitica (L.) H. Lév.[= Polypodium parasiticum L.]. Thelypteris subg. Cyclosoriopsis K. Iwats., Mem. Coll. Sci. Univ. Kyoto , ser. B , 31(1):28. 1964 .— TYPE : Thelypteris dentata (Forssk.) E.P. St. John [= Polypodium dentatum Forssk. ] = Christella dentata (Forssk.) Brownsey & Jermy For complete synonymy, see Holttum (1976b , 1977b , 1982 ), Lin et al.(2013) . Etymology .—Honoring Swiss pteridologist Hermann Christ, 1833–1933 ( Stewart et al. 1983 ). Plants terrestrial, medium-sized (> 40 cm ) to large (fronds to ca. 150 cm tall); rhizomes most often short- to long-creeping or suberect, occasionally erect; rhizome scales almost always lanceolate, brown, setulose on margins and surfaces; fronds monomorphic, occasionally weakly dimorphic with the fertile ones taller and longer-stiped, pinnate-pinnatifid, erect or arching; stipes stramineous to purplish, adaxially grooved, bearing scales at the base like those of rhizome apices; blades chartaceous to subcoriaceous, drying greenish, with proximal several 1–5(–10) pairs of pinna gradually reduced and almost always auricled at acroscopic base (the lowest not less than 2 cm long), less often blades truncate at base (e.g., C. moluccana , C. parasitica ); blade apices usually not conform and generally pinnatifid or pinnatisect, infrequently subconform; rachises generally quite hairy, hairs 0.1–1.0 mm long, rarely also scaly (e.g., C. crinipes ), lacking proliferous buds in axils of pinnae; pinnae usually alternate or becoming alternate distally, adaxially with a groove that is not continuous with the rachis groove, shallowly to often deeply pinnatifid; veins usually prominent on both sides, unbranched, 1–3 lowermost pairs from adjacent segments united at an obtuse angle below the sinus, and forming excurrent veins running to the sinus, rarely lowermost veins connivent at the sinus (e.g., C. conspersa ) or meeting the segment margins above the sinus (e.g., C.harveyi ), free vein ends reaching segment margins; aerophores absent at pinna bases, not swollen (pinnae abaxially sometimes with a slightly raised, darkened lunate ridge at attachment to rachis); indument abaxially of stipes, rachises, costae, veins, and often laminar tissue between veins of hyaline, acicular, spreading hairs (except C. evoluta , which is nearly glabrous), these short to long (0.1–1+ mm), generally unicellular, sometimes also with short capitate hairs, these usually with orangish to reddish, somewhat elongate (pear-shaped or clavate) glands on costules and veins, lacking sessile spherical glands and costal scales; indument adaxially of generally long (> 0.5 mm) hyaline, unicellular setae on stipes, rachises, and costae, sometimes also on veins, also often with scattered to rather dense, usually somewhat spreading (at least not closely appressed) hairs 0.2–0.3 mm between veins on most species; pustules absent on laminar tissue on both sides; sori medial or nearly so, not coalescent at maturity, indusiate (except C. nana ), indusia usually setose on margins and surfaces, and somewhat persistent; sporangia usually without setulae or glands on the capsules, but often each with a unicellular, orangish, elongate gland on the stalk (such glands absent in many African spp.); spores dark brown, with perispore variously ridged, rugose, or tuberculate, lacking narrow wings; x = 36 (14+ spp. counted), diploids and several different tetraploids ( C. dentata , C. parasitica , C. subpubescens ) known, as well as many interspecific hybrids (see below). Diagnosis .— Holttum (1976b) considered the essential characters of Christella to include the presence of an elongate, unicellular gland on the sporangial stalks (illustrated by Smith 1971 ), usually gradual reduction of auricled proximal pinnae (1–5 pairs), the universal presence of erect, acicular hairs on blades, including between the veins, and the absence of sessile spherical glands on the blades, as well as on the sporangial capsules. However, there are exceptions to some of these characters in particular species, e.g., the African species initially placed in sect. Pelazoneuron lack glands on the sporangial stalks, and C. parasitica lacks reduced proximal pinnae. Among related genera, Christella is distinguished from Pseudocyclosorus by the stipe base scales lanceolate, hairy on the margins and both sides, spreading (vs. stipe base scales often ovate, glabrous or with few hairs, appressed); generally anastomosing veins or pairs of veins that unite and produce an excurrent vein that runs to the sinus; the smaller number (or even absence) of reduced proximal pinnae at the base of the blades (vs. blade bases with as many as 12 pairs of abruptly reduced pinnae, the lowest < 5 mm and glanduliform); the presence of acicular hairs on and between the veins, both abaxially and adaxially (vs. hairs usually absent on veins and/or between veins, both sides of lamina); and the usual presence of an orangish, unicellular, clavate or tubular gland on each sporangial stalk as well as sometimes on the indusia and along the veins and costules abaxially. Amblovenatum differs from Christella in often bearing sulfur-yellow, sessile, spherical glands along the veins and costules, especially towards the tips of the segments. Biogeography and ecology .—The 66 known species of Christella are mostly restricted to the Old World tropics and subtropics, from Africa through India and southeast Asia, China , Japan , and Malesia , Melanesia, and Polynesia. The center of diversity of the genus is clearly China and mainland southeast Asia to India (nearly 40 spp.), with a diminished number of species in Malesia , Melanesia, and Polynesia (17 species; Holttum 1977b ), and even fewer in Africa (ca. four spp.; Holttum 1974a ) and the Neotropics (two native, one naturalized; Smith 1971 ). The two native species in the Americas are C. conspersa and the pantropical C. hispidula ; a third species, C. dentata , historically confused with C. hispidula , is very widely naturalized and is now one of the most commonly collected ferns in many parts of North and South America, from southeastern United States and Mexico to southern Brazil , Paraguay , and Bolivia ( Smith 1971 ; Strother & Smith 1970 ). A group of species in eastern Malesia and Melanesia seems to be confined to limestone or coral rocks (e.g., C. buwaldae , C. calcarea , C. gretheri , C. minima , C. moluccana , C. nana , C. perpubescens ; Chen et al. 2017:340–341 ) and were placed in Christella sect. Leptochristella by Holttum (1982) . Christella species are often widespread, locally common or even abundant, and sometimes weedy; the species generally occur at low to middle elevations, 0–1600(–2500) m. They are also quite variable morphologically, and this is reflected in the many heterotypic synonyms (> 10) for the most widespread species (see, e.g., Li et al. 2013 ; Lin et al. 2013 ). Species are commonly found along roadsides and trails, in ditches, ravines, and sometimes on limestone outcrops, often in slightly to heavily disturbed habitats. The ubiquitous C. dentata ( Fig. 2J ) is a common greenhouse weed. Circumscription of the genus herein is largely as treated in Holttum (1971 , 1976b , 1982 ), except that we exclude some African and all American species recognized as Christella sect. Pelazoneuron by Holttum and allied either with Pseudocyclosorus (the African ones) or placed in Pelazoneuron (the American ones). Some of the species accepted in the Flora of China ( Lin et al. 2013 ; Li et al. 2013 ), e.g., C. molliusculus (Wall. ex Kuhn) Ching , C.parvifolius Ching , C. procurrens (Mett.) Copel. , and C. pygmaeus Ching & C. F. Zhang , are known from relatively small areas in China and rather few specimens. We have not seen specimens or even photos of types .We prefer to await a more comprehensive treatment of the genus through its wide range to better evaluate the utility and consistency of the characters being used to define them, as well as their relationships to broadly variable and more widely distributed members of Christella in China and elsewhere, e.g., C. parasitica , C. dentata , and C. subpubescens . Taxonomic and phylogenetic studies.— Christella s.s . , as defined here, including some of the African species of sect. Pelazoneuron discussed by Holttum (1974a) , but not the type , is a member of the christelloid clade, which also includes Sphaerostephanos , Pneumatopteris , Reholttumia , and a few smaller genera ( Fig. 1 ). In its redefined sense, the genus is sister to the clade including Pseudocyclosorus , Trigonospora , and Abacopteris , and this combined clade is sister to Amblovenatum s.s . A clade of African species, including C. chaseana , C . gueintziana , and C. microbasis , variously resolves as sister to Christella s.s. , or as sister to Pseudocyclosorus (Fawcett et al. in press). Species previously recognized as Christella are resolved in clades corresponding to Pelazoneuron and Menisciopsis . The taxonomy and relationships within Christella are still preliminary, because of their widespread and variable nature, propensity to hybridize with related species (and perhaps also with species in other genera), the widespread polyploidy in the genus, and their tendency toward weediness. For general characterization and treatments of Christella in Asia see Holttum’s seminal works (1971, 1976b) and the revision of Chinese taxa by Li et al. (2013) ; coverage for Pacific species is by Holttum (1977b) , and for Malesian taxa by Holttum (1982) . Holttum (1974a) also treated the African species, and those from adjacent islands. Some authors, for example Lin et al. (2013) and Li (2013) , have included Christella s.l. as part of a much larger genus Cyclosorus s.l . , which, thus defined, includes several other christelloid genera (especially Sphaerostephanos , Pneumatopteris sensu Holttum ), and Cyclosorus itself, which we recognize as a small genus of two species. However, Cyclosorus s.s . is significantly different in several morphological characters and falls outside the christelloid clade ( He & Zhang 2012 ; Almeida et al. 2016 ; Fawcett et al. in press). Definitions of Cyclosorus s.l . often exclude Pronephrium s.l. , which is now known to be polyphyletic, with two of its four segregates ( Menisciopsis and Grypothrix ) falling within the chingioid clade and the other two ( Pronephrium s.s . , and Abacopteris ) being within the christelloid clade and related to Christella ( Fig. 1 ). As a consequence of new phylogenetic information, and reconsideration of the morphology, we here choose to recognize Christella in a more restricted sense. Notes .—Hybrids between species of Christella abound, and many different infrageneric hybrids have been proposed ( Shieh & Tsai 1987 ; Wagner 1993 ; Fraser-Jenkins 1997): Thelypteris ×aculodentata Fraser-Jenk.[ Christella dentata × jaculosa ]; Thelypteris × dentiarida Fraser-Jenk. [ Christella arida × dentata ]; Thelypteris × gorkhalensis Fraser-Jenk. [ Christella arida × clarkei ]; Christella × intermedia (W.C. Shieh & J.L. Tsai) D.D. Palmer (= Thelypteris ×incesta W.H. Wagner ) [ Christella dentata × parasitica ]; Thelypteris ×inexpectata Fraser-Jenk. [ Christella dentata × evoluta ]; Cyclosorus × intermedius W.C. Shieh & J.L. Tsai [ Christella dentata × parasitica ]; Thelypteris × jaculodentata Fraser-Jenk. [ C. dentata × jaculosa ]; Christella × kumaunica Holttum [ Christella arida × procera ?]; Thelypteris × linii Fraser-Jenk. [ Christella clarkei × dentata ]; Thelypteris × nareshii Fraser-Jenk. [ Christella dentata × procera ]; Thelypteris × papilioides Fraser-Jenk. [ Christella papilio × procera ]; Thelypteris × parahispidula Fraser-Jenk. [ Christella hispidula × parasitica ]. Christella × altissima Holttum , type from Natal, Buchanan 103b ( K , isotype UC !) appears to be a sterile hybrid with malformed spores and involving C. dentata . Li et al. (2013) listed an additional 22 described species in Cyclosorus subg. Cyclosoriopsis , all with types from mainland China or Taiwan , that they consider putative hybrids, based on morphological intermediacy, lack of sori or no mature normal sori, and malformed spores. Hybrids within Christella can be expected whenever two species in the genus co-occur, but all suspected hybrids need substantiation by more than just morphological intermediacy and spore malformation: cytological, ecological, and nucleotide sequence evidence are also needed to confirm hybridity. A few intergeneric hybrids, mostly from the Paleotropics, are also suspected, with one parent a Christella : Thelypteris × nepalensis Fraser-Jenk. [ Christella dentata × Menisciopsis penangiana ]; Thelypteris × varievenulosa Viane [ Pneumatopteris afra × Christella hispidula ], from the Ivory Coast ( Viane 1985 ); another, given a hybrid generic name, is × Chrismatopteris holttumii Quansah & D. S. Edwards , from Ghana [ Pneumatopteris afra × Christella dentata ; Quansah & Edwards 1986 ]. Thelypteris × palmeri W.H. Wagner [ Christella dentata × Menisciopsis cyatheoides ]. Christella × wildemanii (Christ) J.P. Roux , was accorded hybrid status by Roux (2009) , but the putative parents were not indicated. By Holttum (1974a) and Viane (1985) , this was thought to be a hybrid involving Pneumatopteris afra , which is apparently most closely related to Abacopteris and Christella (Fawcett et al. in press). Viane (1985) believed the other parent might be a species of Holttum’s Christella sect. Pelazoneuron , which we consider members of the pseudocyclosoroid clade, since none of Holttum’s (1974a) African taxa are related to the type of sect. Pelazoneuron , P. patens . In the Neotropics, Smith (1971) postulated a possible origin of tetraploid Thelypteris kunthii (treated by us herein as a species of Pelazoneuron ) from an allopolyploid hybridization event between two diploids, Thelypteris hispidula (at the time treated as Thelypteris quadrangularis (Fée) Schelpe , now synonymized under Christella hispidula ) and T.ovata (treated by us herein as Pelazoneuron ovatum ). There is now some support for such a hypothesis based on topologies of discordant gene trees (Fawcett et al. in press). Perhaps reflecting their penchant to hybridize, members of Christella have also been the subject of artificial hybridization experimentation, maybe more so than within any other genus of ferns. Results of experiments, by Panigrahi & Manton (1958) and Ghatak & Manton (1971) , and conclusions derived from them, were summarized by Holttum (1976b: 295–297) . Crosses were made between known diploids (especially C. hispidula ) from many areas of Asia and Africa, and several tetraploids (e.g., C. dentata , parasitica ), also from multiple sources. Chromosome pairing in the resultant triploids was often n pairs and n univalents, probably indicating a shared genome by the parents, as well as spore malformation (abortion). Other tetraploids used in attempted hybridizations with diploid C. hispidula failed, suggesting that the genome of the latter was not part of the makeup of the tetraploids. The authors concluded that some tetraploids in this group (placed by them in Cyclosorus ) were allotetraploid and shared a genome with C. hispidula . Additional conclusions related to whether certain characters, i.e., ones frequently used in distinguishing species of Christella (and by extension, other thelypteroids), behaved as recessive or dominant in hybridizations, and the recessive characters included: 1) erect rhizomes (vs. creeping); 2) gradual reduction of proximal pinnae (vs. no reduction); 3) absence (vs. presence) of thick glandular hairs abaxially on laminae; and 4) verrucose perispore ornamentation (vs. irregular ridges). These findings have implications for erecting a usable taxonomy. In light of these discoveries, it is not surprising that Christella is especially confusing in the field and in the herbarium. Constituent species and infrageneric taxa (generally following Holttum 1982 ; Li et al. 2013 ).—* Christella acuminata (Houtt.) Lév. ; C. adenopelta Holttum ; ** C . arida (D. Don) Holttum ; C. balansae (Ching) Holttum ; ** C. boninensis (Kodama ex Koidz.) Holttum ; C. burmanica (Ching) Holttum ; C. buwaldae (Holttum) Holttum ; * C. calcarea D.Glenny , sp. ined.; C. callensii (Alston) Holttum ; ** C. calvescens (Ching) Holttum ; C. carolinensis (Hosok.) Holttum ; * C. chaseana (Schelpe) Holttum ; C. clarkei (Bedd.) Holttum ; * C. conspersa (Schrad.) Á . LÖve & D. LÖve; ** C. crinipes (Hook.) Holttum ; C. cuneatus Ching ex K.H. Shing ; C. cylindrothrix (Rosenst.) Holttum ; * C. dentata (Forssk.) Brownsey & Jermy ( Fig. 2J ); * C. ensifera (Tagawa) Holttum ; C. euphlebia (Ching) Holttum ; * C. evoluta (C.B. Clarke & Baker) Holttum ; C. fukienensis (Ching) Holttum ; * C. gretheri (W.H. Wagner) Holttum ; * C. guamensis Holttum ; * C . gueintziana (Mett.) Holttum [several authors, e.g., Schelpe (1964) , Holttum (1974a) , and Roux (2009 , 2013), have used the altered spelling “ gueinziana ”, but this is impermissible]; C. gustavii (Bedd.) Holttum [placement in Christella disputed by Li et al. 2013 ]; * C. harveyi (Mett.) Holttum ; C. harveyi var. connivens Holttum ; * C. hispidula (Decne.) Holttum ; C. hokouensis (Ching) Holttum [placement in Christella disputed by Li et al.]; * C. jaculosa (Christ) Holttum ; C. jinghongensis (Ching ex K.H. Shing) A.R. Sm. & S.E. Fawc. ; C. kendujharensis S.K. Behera & S.K. Barik. ; C. kumaunica Holttum ; * C. latipinna (Benth.) H. Lév. ; C. lebeufii (Baker) Holttum ; C. microbasis (Baker) Holttum ; C. minima Holttum ; C. modesta Holttum ; C. moluccana M. Kato ; C . mutifrons ( C . Chr.) Holttum; C. namburensis (Bedd.) Holttum ; C. nana Holttum ; ** C. nanxiensis (Ching ex K.H. Shing) A.R. Sm. & S.E. Fawc. ; C. pacifica Holttum ; * C. papilio (Hope) Holttum ; C. papyracea (Bedd.) Holttum ; * C. parasitica (L.) Lév.; C. peekelii (Alderw.) Holttum ; C. perpubescens (Alston) Holttum ; C. procera (D. Don) Mazumdar , with heterotypic synonym C. appendiculata (C. Presl) Holttum (= Thelypteris appendiculoides Fraser-Jenk. ); C. prolixa (Willd.) Holttum ; C . pseudogueintziana (Bonap.) Alston; C. rupicola (Hosok.) Holttum ; ** C . scaberul a (Ching) Holttum; C. semisagittata (Roxb.) Holttum ; C. shimenensis (K.H. Shing & C.M. Zhang) A.R. Sm. & S.E. Fawc. ; * C. siamensis (Tagawa & K. Iwats.) Holttum ; C. sledgei (Fraser-Jenk.) Ranil ; C. subdentata Holttum ; ** C. subelata (Baker) Holttum ; C . subjunc ta (Baker) Holttum; * C. subpubescens (Blume) Holttum ; C. taprobanica (Panigrahi) Holttum ; C. timorensis Holttum ; C. wulingshanensis (C.M. Zhang) A.R. Sm. & S.E. Fawc. ; C. zeylanica (Fée) Holttum. Excluded names.— Dryopteris albociliata Copel. [= C. parasitica ]; Dryopteris assamica Rosenst. [= C. cylindrothrix ]; Cyclosorus benguetense Copel. [= C. hispidula ]; Polystichum benoitianum Gaudich. [= Sphaerostephanos benoitianus (Gaudich.) Holttum—Holttum thought it possibly related to Strophocaulon invisum (G. Forst.) S.E. Fawc. & A.R. Sm. ]; Nephrodium biauritum Bedd. [= C . lebeuffii ]; Dryopteris contigua Rosenst. [= C. hispidula ]; Nephrodium didymosorum Parish ex Bedd. [= C. parasitica ]; Dryopteris euaensis Copel. [= C. harveyi ]; Nephrodium eurostotrichum Baker = Christella distans (Hook.) Holttum , incertae sedis ; Cyclosorus falcatulus (Christ) Copel. [= C. hispidula ]; Dryopteris hirtopilosa Rosenst. [= C. hispidula ]; Cyclosorus jerdonii Ching (excluded from Christella by Holttum 1976b: 336 ); Dryopteris meeboldii Rosenst. [= C. malabariensis ]; Dryopteris mindanaensis Christ [= C. dentata ]; C. molliuscula (Wall. ex Kuhn) Holttum [= Pseudocyclosorus canus (Baker) Holttum & Grimes ]; C. multiauriculata Punetha [= C. dentata ]; Thelypteris novae-hiberniae Holttum [= C. harveyi ]; Polypodium nymphale Forst. [= C. dentata ]; Aspidium obliquatum Mett. [= C. prolixa ]; Aspidium procurrens Mett. [= C. parasitica ]; Dryopteris pseudoamboinensis Rosenst. [= C. subpubescens ]; Nephrodium quadrangulare Fée [= C. hispidula ]; Dryopteris repandula Alderw. [= C. hispidula ]; Cyclosorus subaridus Tatew. & Tagawa ex Tagawa [= Christella jaculosa ]; Dryopteris submollis Alderw. [= C. dentata ]; Dryopteris sumatrana Alderw. [= C. subpubescens ] Excluded species.— Three Hawaiian endemic species treated in Christella by Holttum (1977b) and Palmer (2003) , Christella boydiae , C. cyatheoides and C. wailele are here transferred to the genus Menisciopsis . Based on a photo of the type of Christella burundensis Pic.Serm. , and notes on the specimen by Holttum, this is not a Christella , and may be Pelazoneuron kunthii (Desv.) A.R. Sm. & S.E. Fawc. ; it has free veins and the proximal pinnae are not reduced. If that species, it is likely introduced and not native to Burundi . Incertae sedis.— Christella afzelii (C. Chr.) Holttum. Based on an image of the type , this appears to be bipinnate; veins are free and sometimes forked. It is unlikely to be a Christella . Several Afro-Madagascan species (represented by Pneumatopteris humbertii and Christella distans in Fawcett et al. in press) are more closely related to Pseudocyclosorus and Trigonospora than to Christella s.s . Holttum (1976b) considered Christella distans the most aberrant species in Christella , in part because of the swollen aerophores; we agree, and its taxonomic placement remains uncertain. Two other species, Christella friesii (Brause) Holttum and C. guineensis (Christ) Holttum , may also belong to the Pseudocyclorosus + Trigonospora clade. Based on the original description ( Rakotondrainibe & Jouy 2011 ) and images of the type ( P !) of Christella darainensis Rakotondr. , we note that the veins are free with the proximal one of a pair running to the sinus, the adjacent vein meeting the margin above the sinus, the rhizome is erect, and the stipe base scales are ovate and glabrous, unlike those of Christella , suggesting it may be more closely related to Trigonospora . The African Thelypteridaceae have been the most difficult taxonomically and have been subject to the greatest taxonomic upheaval as a result of our phylogenetic study (see Menisorus and Pneumatopteris for further discussion). This morphologically variable, and poorly studied species group cannot be confidently placed; additional molecular sampling and study of herbarium material will be necessary before taxonomic changes are made.