NOTES ON TAXONOMY OF ATHYRIUM NIGRIPES (BLUME) T. MOORE COMPLEX, AN INSIGHT FROM MOUNTAINS IN JAVA

pulcherrimum and At. nitidulum with phenetic approach. The analysis formed four clusters; cluster I consist of the large sized and include type specimen of At. pulcherrimum and Asp. costale, cluster II for plants with standard form and include type specimen of At. nigripes , cluster III only for serrated plants and cluster IV contain the At. nitidulum form. Since Asp. costale clustered in group I, this species is more closely related to At. pulcherrimum than to At. nigripes nested in group II. Large size and degree of division indicate major clustering within the complex species.

The genus Athyrium is a medium sized terrestrial fern, mainly distributed in the northern hemisphere that in the tropics grow only in humid high altitude (Rothfles 2012, PPG I 2016, Wei et al 2018).This genus can be differentiated from other Athyriaceae genera by a combination of characters, i.e., confluent groove with projections, inequilateral pinna or pinnule base, entire scale margin, linear to horseshoe-shaped indusium (Liu et al 2011).Athyrium nigripes (Blume) T. Moore firstly published in Blume's Enumeratio Plantarum Javae (1828) as Aspidium nigripes, based on specimen collected from the wetland at a valley between Mount Burangrang and Mount Tangkuban Parahu in West Java.Blume placed this name under the bipinnate frond section.Aspidium costale, a species that later is considered related, consists of four varieties, described in the tripinnate section.Moore (1857), in his Index Filicum, transferred Aspidium nigripes to Athyrium, but in later pages, he reduced it into Athyrium tenuifrons var.b tenellum.Hooker (1860) reversed this synonymy and accept At. nigripes instead.In the same publication, he synonymized Blume's Aspidium costale as Asplenium gymnogrammoides, a name described by Klotzsch in an unpublished manuscript that was validated by Mettenius (1859) as explained in Sledge (1956).Hooker (1874), in his subsequent account, included both name as synonym of At. nigripes.Since then, Blume's four varieties of Aspidium costale is known as At.nigripes, as adopted by Raciborski (1898) in his Pteridophyte Flora of Buitenzorg (old name for Bogor) under Asplenium, and by Alderwerlt (1908Alderwerlt ( , 1917) ) in his accounts of Malayan fern flora.However, Backer and Posthumus (1939), in their fern flora of Java, recognized them as separate, following previous Hooker's (1860) treatment.Sledge (1956) explained that synonymizing Klotzsch Asplenium gymnogrammoides with Blume's Aspidium costale was an error which continue to persist in Indian flora accounts.He found that Ceylon type specimen of Aspl.gymnogrammoides, as also many other gatherings from Ceylon and India, is quite different with Javan specimen At. nigripes.He suggested that Indian Aspl.gymnogrammoides is partly a form of At. solenopteris.However, he concluded that this Blume's At. nigripes is a rather scarce plant and the materials abundantly available at Kew and British Museum Herbarium is belong to At. appendiculiferum Alderw. (1914), a name described from Mount Tandikat in Sumatra.On the other hand, At. nigripes also recognized in Flora of China (Wang et al 2013) which Fraser-Jenkins et al (2018) suggested that the adopted concept is a misconception for gathering that should belong to At. setiferum C.Chr (Christensen 1906), a species initially published without description (nomen nudum) as Allantodia tenella Wall.(herbarium catalogue, dated 1821) that then corrected by Hope (1899) as Asp.tenellum, based on specimen from Nepal.In summary, Blume's At. nigripes is presumed not exist neither in India nor in China, and a species described by Alderwerlt (1914) from Sumatra is likely the same entity.
Our initial examination also shown that Copeland's At. pulcherrimum (1913), described from a plant collected in Mt.Pangrango, Java, might be another form of large At.nigripes.Backer and Posthumus (1939)  These interchanging synonymies indicate a wide range of morphological variety, instigate user to apply names differently.We applied phenetic approach to a gathering collected from Java, in the attempt to elucidate relationship within the morphological range and contribute species delimitation exercise for the name described from the island, i.e.At. nigripes (Blume) T. Moore, Asp.costale Blume, At. pulcherrimum Copel. and At. nitidulum (Kunze) Rac.The last name was tagged by Rosenstock (dated 1915) to the type specimen of Asp.costale type C that is kept in Leiden herbarium, while At.nitidulum type specimen' whereabouts is unknown.We refer to Raciborski's speci-men and description (1898) in this study to clarify its relationship to the core At.nigripes.We confined area for investigation only to Java since there were very few specimens gathered from other location before us.Moreover, specimen examined by Blume, Raciborski, Backer and Posthumus that available in Bogor were mostly collected from Java.

MATERIAL AND METHODS
We examined 95 specimens of related At. nigripes taxa that available in Bogor Herbarium.However, only 43 specimens, that each represent-ing single OTU, could be incorporated into the matrix.Some of them consist of two sheets of young and adult leaves.Only the adult specimen included to the matrix.These specimens are mostly collected from Mount Gede-Pangrango, a high mountain consists of two tops of about 3000 m asl that one with active volcanic crater, and the saddle that connect the two tops lies at about 2500 mm altitude.We excluded specimens that distinctly different such as those with winding frond or specimens that extremely incomplete.We also incorporated data derived from description of three species in their protologue whilst consulted image of the specimen type, i.e.At. pulcherrimum, Asp.costale, Asp.nigripes.For At. nitidulum, we use Raciborski's specimen and description (1898) instead of Kunze's (1848), since the later examined specimens is unknown and its description is fairly general.
We evaluate over 30 morphological characters but only 19 that has distinct variation across the heaps and thus scored, as presented in table 1 (matrix available as supplementary file).The quantitative character "length of stalk of basal pinnae" were log10-transformed to stabilized variation before scoring.The clustering analysis were perform using UPGMA applied in PAUP (Swofford 2003) with total character difference as the basis for distance measure and ties were broken randomly.

RESULT AND DISCUSSION
Variation within the complex All examined specimens has common characters, i.e., pale stipe scale that might be darker toward rhizome, bipinnate to tripinnate (at least at basal pinnae) frond, basal pinnae widest, axis with groove, pinnae rachis or costae bearing fleshy projection or "spine" arise from one side of the groove, similar but longer projection also present on costule (midrib).These projections often not present at basal and adjacent intersection, mostly more prominent toward apex.General shape of lamina is ovate to oblong-lanceolate with decent stipe length, apices mostly acuminate.
The range of variation in this complex is wide and mostly continues, lies in sizes, degree of indumentum covering, degree of incision, arrangement of leaflets and indusium.Small plant might be fertile with rather bipinnatifid appearance, while larger size tends to be tripinnate at basal pinnae with overlapping or non-overlapping pinnules.There are plants with contracted lobes or pinnules on leaflets that widely separated.This type of plant is collected from open area such that around a volcanic crater or baren space on the top of a mountain.Other has distinctly larger basal lobe, found in relatively shaded-dry area in high altitude.The complex obviously affected by environmental condition which expressed in their plastic feature.
We categorized the variation based on general appearance as follow:

Phenetic analysis of the Athyrium nigripes complex in Java
Our analysis shows that the gathering with 9 initial categories were polarized into 4 major groups (Fig. 1).Group I is plants with large sizes, tripinnate or at least basal pinnae with pinnately arranged pinnules (Fig. 2).Leaflet ranged from contracted to broad and overlaps.However, type specimen of At. pulcherrimum and Asp.costale (type C) (Fig. 3) belong to this group but clustered in its own branch.Group II (Fig. 4) consists of plants with medium sizes, basal pinnae deeply lobed but immediately shift to shallower incision at or before middle pinnae.Type specimen of At. nigripes nest in this group.Group III (Fig. 5) is a cluster of plants with distinctive serrated margin and rather longer lamina compare to medium sized plants.Some of these specimens were once identified as At.gymnogrammoides.Group IV (Fig. 6) is also an exclusive cluster of specimens identified as At.nitidulum (sensu Raciborski) which distinct it their hair covering and rather dark stipe/rhizome scales.
This exercise confirms some of the assumption made by different authors.However, further investigation with broader taxon sampling and additional characters are indispensable.Our clustering demonstrates that, Asp.costale (type C) and At.pulcherrimum most likely are the same species.Both has distinctively larger and separated basal pinnule or lobe.As mentioned in the account, Blume separated Asp.nigripes (basionym of At. nigripes) from Asp. costale by their degree of division, i.e., bipinnate vs tripinnate.Nevertheless, there are notable continues variation between the two species.Asp.costale has four varieties which only type C and D specimen detected in the online resource (Fig. 3).Member of group I is belong to these varieties as they fit the description to some extent.Hence, large and more-divided version of At. nigripes might be included in this group.Raciborski suggested that this "nigripes-costale" group is highly variable that need decent investigation.Whenever both species treated as the same species, Asp.costale and At.pulcherrimum would be the synonyms for At.nigripes.
In addition, the type specimen of At. nigripes that nest in cluster II is a rather distinct specimen.It has shorter pinnae with oblique and narrow lobes, while most of the other specimen in the cluster tend to have lobes and leaflets inserted in almost horizontal line.Sledge (1956) suggested that this species might be a quite scarce plant, while Fraser-Jenkins (2018) believed that they are juvenile plant develop swiftly to fertile specimen.Despite of its placement in the distance tree, this rare form might or might not be different from other standard bipinnate form.Recollecting specimen of this type from its provenance would be fundamental in confirming At. nigripes and group II identity.
The serrated group in the third cluster is most similar to type specimen of At. gymnogrammoides from Srilanka (Fig. 5).However, since the type specimen only contain an incomplete and small portion of a plant, we cannot confidently make any decent comparison.This third group distinctive in its bipinnate form with serrated margin, rather rounded and cuneate-truncate pinnules with shallow lobes.Such characters separated them from Asp. costale.However, the cluster is the immediate sister to nigripes-standard form cluster which make them can be treated as one.
The fourth cluster contains At. nitidulum sensu Raciborski (1898) (Fig 6).This species is recognized in different flora with divergent concept.The description of Allantodia nitidula (Kunze, 1848) mentions its elegant look and glabrous rachis, separated it from hairy Allantodia paludosa, a name that later treated as synonym of At. nigripes in Backer and Posthumus's Fern Flora of Java (1939).Alderwerelt (1908) discriminate the two based on indusium, i.e., horseshoe-shaped in At. nitidulum present only near apex, both horseshoeshaped and double (back-to-back) indusium in At. nigripes.This combination of characters might present interchangeably, thus unreliable as sole identification hint.Nevertheless, uniting this cluster with sister cluster II and III is a plausible treatment with such limited observed variation.
The analysis clustered the complex into four groups.Group I comprise of specimen with large size and prominent division.Group III is immediate cluster to group II, and group IV is clustered next to both.We might treat all group as different species, or merge group II and III that resulted in three species, or gather Group II, III and IV as one species separated from group I.There is not enough confidence to delimit either alternative, due to inadequate samples, especially in the group III and IV, and the absence of additional specimen of At. nigripes type.However, this exercise demonstrate that size and degree of division might indicate different entities within this complex.

CONCLUSION
This study depicts the agreement and separation of continues variation within the At.nigripes complex.Large form with higher leave division is separated from the bipinnate form.Thus, Asp.cos-tale is more closely related to At. pulcherrimum compare to At. nigripes.Although the bipinnate form comprise of three clusters, there is not enough evidence to differentiate them because of the absence of Asp.nigripes and Al.nitidula type specimen, and the limited amount of sampled material.For further clarification, it is essential to gather more specimen, include more characters, e.g., cytology, spore ornamentation, molecular sequences, and employ different approaches.

Fig. 3 .
Fig. 3. Comparison of the types nested in Group I, from left to right: At. pulcherrimum, Asp.costale Type C and Type D

Fig. 5 .
Fig. 5. Member of Group III with serrated margin (left) and type specimen of At. gymnogrammoides (right).

Fig. 6 .
Fig. 6.At. nitidulum of the group IV (left); a closer image of basal pinna rachis with short hairs on its surface (right).

Table 1 .
erroneously included this name under At.macorcarpum together with the simply pinnate A. puncticaule.The use of names in relation to At. nigripes and Asp.costale is summarized in the following table: Names adopted for Athyrium nigripes and its presumed close relatives in

Table 1 .
Morphological character observed and used in clustering analysis 1Short acuminate apex = almost acute; Long acuminate = gradually narrowed half way toward apex 2 Decent space = as in standard form, gap slightly less than half wide of pinnule lobe