Athyrium bipinnatum K.Hori (Athyriaceae), a new cornopteroid fern from Japan

Abstract I describe Athyrium bipinnatumsp. nov. and discuss morphological differences between closely related species. The new species is endemic to Japan, occurring on the islands of Honshu, Kyushu, and Shikoku. Based on the criteria of the International union for conservation of nature and natural resources, this new species is here considered endangered.


Introduction
Cornopteris Nakai is a small Asian (Himalayas, East and Southeast Asia) genus of terrestrial ferns that used to be recognized by many pteridologists (e.g., Nakai 1930;Ito 1939;Ching 1945;Holttum 1958;Tagawa 1959;Kato 1977Kato , 1979, on the basis of fleshy stipes, corniculate leaf axes, and exindusiate sori. Regarding its phylogenetic relationships, Ching (1945) regarded it to be "a little offshoot of the exindusiate Diplazium." Based on the presence of J-shaped sori in some species of Cornopteris, as well as on stipe features and spinulose midribs of the laminae, Kato (1977) concluded that Cornopteris was actually more closely related to Athyrium than to Diplazium. Serizawa (1981) agreed with Kato's conclusion and subsumed Cornopteris in Athyrium. To corroborate his taxonomic decision of lumping the two genera, Serizawa (1981) also highlighted the existence of natural hybrids between them.
Cornopteroid species of the genus Athyrium are mainly distributed in the Himalayas, in East and Southeast Asia. The exact number of species is unknown. Kato (1979) recognized nine species of cornopteroid ferns, whereas Zhaorong and Kato (2013) recognized 16 species only in China.
The following is a brief history of the classification of one Athyrium species, namely Athyrium × christensenianum (Koidz.) Seriz., which is partly the focus of the present study. Athyrium × christensenianum was first described by Koidzumi (1924) as Diplazium christensenianum Koidz., from Jeju (Quelpaert)-Island, South Korea ( Fig. 1). Kato (1979) synonymized Cornopteris hakonensis Nakai, from Hakone, Japan (Nakai 1930, Fig. 2) under Cornopteris christensenianum. Subsequently, Serizawa (1981) transferred it to Athyrium and created the nothospecific name A. × christensenianum. The hybrid origin of this species has been pointed out by many authors (Kurita 1964, Hirabayashi 1970, Kato 1979, Park and Kato 2003. According to these authors (Kurita 1964, Hirabayashi 1970, Kato 1979, Park and Kato 2003, this is a triploid "species" that has resulted from a cross between diploid sexual Athyrium crenulatoserrulatum Makino and tetraploid sexual Athyrium decurrentialatum (Hook.) Copel. It remains unclear whether A. × christensenianum is merely a sterile hybrid or an independent fertile species. Kato (1979) categorized it as being of a "hybrid origin species" between A. crenulatoserrulatum Makino and A. decurrentialatum because it has intermediate morphological characteristics between these two species such as serrated segments, subcartilaginous margins of the blade, and elliptical sori. On the other hand, Serizawa (1981) treated this "species" as a sterile "hybrid." The present study agrees with the treatment of A. × christensenianum by Serizawa (1981).
Recent studies have reported complicated relationships in the A. × christensenianum complex. First, Hori and Murakami (2019) reported a tetraploid sexual cytotype of A. × christensenianum (as the independent species A. christensenianum). Subsequently, Hori (2019) reported that tetraploid sexual A. × christensenianum had one allele of A. crenulatoserrulatum and one of A. decurrentialatum, each in the biparental inherited nuclear DNA marker of the gene AK1. He also found that triploid A. × christensenianum had two alleles of A. crenulatoserrulatum and one allele of A. decurrentialatum. Therefore, he suggested two hypotheses: (1) tetraploid sexual A. × christensenianum originated from the hybridization of diploid sexual A. crenulatoserrulatum with an ancestral or extinct diploid A. decurrentialatum; and (2) triploid A. × christensenianum originated from the hybridization of diploid sexual A. crenulatoserrulatum and tetra- ploid sexual A. × christensenianum (Fig. 3). In the present study, the unclear taxonomy of A. × christensenianum was clarified by describing tetraploid sexual A. × christensenianum as a new species: Athyrium bipinnatum from Japan (Fig. 3).

Materials and methods
First, the morphological characteristics of a tetraploid sexual specimen of Athyrium bipinnatum (Hori 2974) and the sexual specimens (Hori 2975(Hori , 2976 described by Hori (2019) were examined. Two type specimens of A. christensenianum, deposited at KYO (as Diplazium christensenianum; Fig. 1) and TI (as Cornopteris hakonensis; Fig. 2), were also examined. Furthermore, based on morphological characteristics, more specimens of A. bipinnatum were identified at MAK and MBK, and from online images at PE herbarium (PE: http://pe.ibcas.ac.cn/en/), Taiwan Forestry Research Institute herbarium (TAIF: http:// taif.tfri.gov.tw/search.php), Collection Database of Specimens and Materials (TNS: http://db.kahaku.go.jp/webmuseum/), and from those in JSTOR Global Plants (https:// plants.jstor.org/) and the Global Biodiversity Information Facility (GBIF: https://www. gbif.org). Additional samples were also collected in the field and cultivated samples were taken from Koishikawa Botanical Garden of Tokyo University. For the conservation assessment, the area of occupancy (AOO) and extent of occurrence (EOO) were estimated using GeoCAT (Bachman et al. 2011), with the default settings for grid size applied. The morphology of scales and spores was observed using a KEYENCE VHX-D500.
Etymology. Athyrium bipinnatum is named as a new species with bipinnate blades, which distinguish it from A. × christensenianum. This new species could be one of the parents of A. × christensenianum (Hori 2019

Distribution and ecology.
Athyrium bipinnatum is known from the western part of Honshu, Shikoku, and Kyushu in Japan (Fig. 5). It was observed to grow on soil under planted coniferous forest containing Cryptomeria japonica at ca. 600-1100 m alt. This species is endemic to Japan. Conservation status. IUCN Red List Category. Based on estimates from GeoCAT, the EOO of A. bipinnatum was 53,068 km 2 ; however, this estimate should be smaller because this EOO contains an area of sea around Japan. Wild deer are known to eat A. bipinnatum (pers. obs.), so individual numbers within populations are currently decreasing and population reduction is occurring now. The known AOO of A. bipinnatum is 44 km 2 . Based on the IUCN criteria (IUCN 2001(IUCN , 2012, A. bipinnatum falls into the Endangered (EN) category. A formal evaluation of endangerment can be summarized by the following IUCN hierarchical alphanumeric coding system of criteria and subcriteria: EN A1abc+A2+C1+C2a(i).

Discussion
Athyrium × christensenianum was first described by Koidzumi (1924) as Diplazium christensenianum Koidz., from Jeju (Quelpaert)-Island, South Korea (Fig. 1). Subsequently, Nakai (1930) described Cornopteris hakonensis Nakai, from Hakone, Japan (Fig. 2). I examined type material of both taxa, and found that they represent the same hybrid in having large size of stipes, blades, 2-pinnate pinnatifid pinnae in the middle of the blades. Athyrium bipinnatum is distinguished from them in having shorter stipes, smaller blades, and by its 2-pinnate pinnae in the middle of the blades.
Contrary to the findings of the present study, Park and Kato (2003) reported regular-shaped and germinative spores in their description of A. × christensenianum. Here, regular-shaped spores could not been found from dozens of A. × christensenianum herbarium specimens; however, regular-shaped spores were found from several specimens of A. bipinnatum, A. crenulatoserrulatum, and A. decurrentialatum. Therefore, it is possible to speculate that Park and Kato (2003) perhaps confused the materials of these different species when they examined spore shape.  In summary, the previously unclear taxonomy of A. × christensenianum was clarified in the present study by describing tetraploid sexual "A. × christensenianum" as the new species A. bipinnatum. Importantly, the conservation status analysis reported here suggests that A. bipinnatum may be endangered; thus, this study has relevance to the conservation of cornopteroid ferns.