Studies on Pacific Ferns Part I. Nomenclature Changes and Distributions of Some Species of Hymenophyllum, Arthropteris, Microlepia, Oleandra, and Adiantum

G. BROWNLIE!

DURING A PERIOD OF STUDY at the Herbarium by Brackenridge states "sari few on upper half of the Royal Botanic Gardens at Kew the author of the fronds-two-valved for ~ of its length, spent some time examining collections of ferns obovate, lips entire,-receptacle short, included," from the South Pacific area. It was quickly while Baker describes his species as "sari one apparent that in most groups there is some to a pinna near the base on the upper side­ confusion in interpretation of the various indusilum with a campanulate rube and sub­ species, and the following paper is an attempt orbicular entire lips." These two descriptions to clear up some of these.points. are not dissimil ar, and a comparison of Baker's type material with the illustrations of Bracken­ GENUS Hymenophyllum ridge's specimen tends to suppOrt the idea that these two are in fact the same species. Baker's Differences between the specimens examined statement that his material was glabrous is not and the descriptions appearing in Copeland's supported by an examination of the type, which revision of the genus were noted in some small has hairs on the stipe, lower rachis, and rhizome species of H ymenophyllum (sens . lat.) . Con­ similar to those illustrated by Brackenridge. The sequently a detailed examination of this group difference in size between the two forms was undertaken, with particular reference to described is bridged by later collections from those species possessing fronds with serrated Fiji , and Baker 's specimen seems to be a large margins which Copeland distinguished as Mer­ example only. ingium and Hymenophyllum. The following ob­ Specimens examined (all from Fiji): Me­ servations, which are the result of these investi­ Gillivray unnumbered (Kew, Baker's type ); gations, amount to a further revision of this Parks 20040 (Kew, B.M.2); A. C. Smith 1368 group for the South Pacific region, excluding (Kew) ; RM.S. Herald Bot. No. 138 (Kew); Australia and New Zealand. The present author RM.S. Herald Bot. No. 286 (Kew) . prefers to retain, in the meantime, the generic name Hymenophyllum for all the species con­ Hymenophyllum pseudotunbridgense Watts. cerned, as he is not convinced of the validity of Hymenophyllum pseudotunbridgense Watts. Copeland's distinguishing characters. Proc. Linn. Soc. N.S.W. 39, 766, 1915. H ymenophyllum macgillivrayi (Bak.) Capel. FIJI Philipp. J. Sci. 64, 60, 1937. Hymenophyllum affine Brack. In his section on H . macgillivrayi Copeland Hymenophyllum affine Brack. Expl. Exped. quotes Baker's description of Trichomane.r mac­ 16, 265, 1854. gillivrayi, and supplements this with a further Trichomanes macgillivrayi Bak. Ann. Bot. description based on a collection by J. Horne 5, 195, 1891. 1877-78, illustrating this on plate 38. This The original description of this Fijian species illustration perfectly matches further Horne material at Kew which is distinct from Baker's 1 Botany Department, University of Canterbury, T. macgillivrayi. The involucre is longer and Christchurch, New Zealand. Manuscript received April 20, 1959. 2 British Museum (Natural History).

242 Pacific Ferns, I-BROWNLIE 243 narrower in the Horne material and the recep­ Hymenophyllum dimid iatum Mett. tacle is markedly extruded. Copeland has thus Hymenophyllum dimidiatum Mett. Kuhn. confused two quite distinct Fijian species under Linn. 35, 393, 1868. the name Hymenophyllum macgillivrayi. A com­ parison of the Horne material with species from This species is somewhat larger than the surrounding areas shows that it is almost iden­ preceding one and is distinct in the possession tical with H. pseudotunbridgense from Queens­ of a serrated involucre. An examination of the land, and consequently this name should be type material shows this fern is not glabrous, given to the form illustrated by Copeland as H . as appeared in the original description, bur has macgillivrayi. hairs on the stipe and rachis. In fact, no glabrous Specimens examined (all at Kew) : Queens­ species was noted among the groups discussed land, Bailey unnumbered; Fiji, Horne 983 in in this paper. parr, 1076, and unnumbered; A. C. Smith 5713. Specimens examined: Deplanche in Herb. Lenormand (Caen) ; Vieillard 2258 (Kew, Hymenophyllum denticulatum Sw. B.M.). Hymenoph yllum denticulatum Sw. Schrad. Hymenophyllum piliferum C. Chr. Journ. 1800/2, 10, 1801. Hymenophyllum piliferum C. Chr. Vjschr. This species is quite distinct from all other Naturf. Ges. Zurich 70, 221, 1925. small Fijian members of the genus in the posses­ sion of a conspicuous wing along the whole Christensen separated this from H. dimid­ rachis. iatum on the basis of the hairs on stipe and rachis bur, as noted above, the type material of that species is not glabrous. However, it is felt NEW CALEDONIA that it should be retained as a valid species on Hymenophyllum subdimidiatum Ros. the grounds of its more open growth form and much smaller size. It is similar to H. s1tbdi­ Hymenophyllum subdimidiatum Ros. Meded. midiatum in form and size bur has the serrated Rijks-Herb, Leiden no. 11, 1, 1912. involucre of H. dimidiatum. Copeland included Hymenophyllum viride Ros. Copel. Philipp. specimens assigned to this by Christensen in his Sci. 64, 59, 1937. J. examples of H. dimidiatum and ignored Chris­ In his treatment of this species Copeland tensen's species completely . quotes Rosenstock 's description and mentions Specimens examined: Franc 1450 in Herb. C. the type material as Schlechter 14799, which he Christensen (B.M.); Rosenstock 133 (B.M.). states he has not seen, bur in the same paper he publishes the new species H. viride from Hymenophyllum pumilio Ros. material with the same collection number. The Hymenophyllum pumilio Ros. Fedde Reperr plate illustrating this species shows it to be the 9,72,1910. same as Rosenstock 's original description and This is a very reduced species somewhat sim­ the same as a cotype (Schlechter 14799 ) at ilar to H. minimum of New Zealand, but lack­ Kew. ing the spines found on the sides of the sorus It is doubtful whether or not this species is in that species. It cannot be confused with any distinct from H. pseudotunbridgense, for, al­ other New Caledonian fern. though no markedly extruded receptacles were noted, the involucre shape is the same and the general appearance of the plant is very similar. SAMOA However, until a greater amour of comparative Hymenophyllum praetervisum Christ. material is available it is perhaps better to re­ Hymenoph yllum praetervisum Christ. Engl. tain the species. Material examined: Schlechter 14799 (Kew) : Jahrb. 23, 338, 1896. Balansa 2701 (Kew, B.M.) . This is the only small species in this section 244 PACIFIC SCIENCE, Vol. XIV, July 1960 in Samoa and is quite distinct from the Fijian not immersed. Involucre suborbicular, cleft more and New Caledonian forms in having the sari than half way to the base. Lips entire, receptacle free from the pinnae and not in any way im­ extruded. mersed . Lord Howe Island, Mt, Gower, at a high al­ titude. C. Moore No.4 (Kew). LORD HOWE ISLAND This species is related to H. pseudotunbridg­ ense Watts, but is distinguished by the relatively H ymenophyllum moorei Bak. wide frond and very long, narrow segments. H ymenophyllum moor ei Bak. Syn. Fil. 464, The segments are narrower and longer than 1874. those of H . subdimidiatum Ros., to which it Hymenophyllum pumilum C. Moore, Capel. also appears to be related. Philipp. J. Sci. 64, 91,1937 (in part). The remainder of this paper consists of a varied assortment of ferns in which confusion Copeland uni ted the Australian H. pumilum exists either as to specific distinctiveness or as and the Lord Howe Island H. moor ei, disre­ to geographical range. garding the distinctions noted by Baker in the original descriptions. A comparison of both Artbropteris neocaledonica Copel. types at Kew shows that the sori in the Aus­ tralian material are terminal, not immersed at Arthropteris neocaledonica Capel. Sargentia the base, and have valves slightly serrated at 1,4, 1942. the apex, while the Lord Howe Island spec­ This has been distinguished from A. tenella imen has lateral sori, immersed at the base, and ( Forst. f.) J. Sm. by the possession of black with conspicuously toothed valves. The distinc­ instead of brown scales, and in having the posi­ tion in position of the sari is alone sufficient tion of the sorus well marked on the upper to retain the two as separate species. Cope­ surface. An examination of collections at Kew land's illustration (pI. 45) figures H . m oorei, shows the last character to be of doubtful value, not H . pumilum. but the distinctive colour of the scales is in fact Specimens examined: C. Moore unnumbered consistent for all New Caledonian material. This (Kew), Baker's type. contrasts very clearly with the light chestnut­ Hymenophyllum howense Brownlie n.sp. brown scales of New Zealand and Norfolk Is­ land specimens. Many of the collections from Rhizoma longe repens, sparsis et rufis pilis New Caledonia had previously been identified ornatum; stipes filiformis, 2 ad 4 em. longus, as A . tenella, and the geographical range of that ad basem pilosus; frondes usque ad 5 em. altae, species given in Christensen's "Index Filicum" 2 em. Iatae, lineares ovatae, rri-quadri-pinnarif­ includes that island, but the characteristic scale idae; segmenris usque ad 2 em. longis, angus tis colouring was observed on all specimens from linearibus, inconspicue serrulatis; sori in apici­ there. It seems best then to distinguish two bus segrnenrorum breviorum, non imrnersi, in­ species, geographically separated, Artbropteris dusia suborbiculari, labiis integris, receptaculo tenella in New Zealand and Norfolk Island, and exserto . A. neocaledonica confined to New Caledonia. Rhizome creeping, with scattered multicellu­ lar rugose hairs which sometimes occur in Mi crolepia speluncae (1.) Moore var. exserta clusters at the base of the stipes and at branch Mett. junctions. Stipe filiform, 2-4 em. long, with Davallia speluncae var. exserta Hk. and Bak. scattered hairs similar to those on the rhizome. Syn. Fil. 470, 1864. Fronds up to 5 em. long and 2 em. wide at the Microlepia exserta Metr . Kuhn Linn. 36, 148, widest point, linear-ovate in outline, flabellately 1869; C. Chr. Ind . Fil. 426, 1906. 3 or 4 pinnarifid, Divisions narrow, linear, up to 2 em. long, obscurely serrate. Sori mostly on An examination of material of M. speluncae upper half of frond , terminal on short segments, from the Pacific islands shows the distinguish- Pacific Ferns, I-BROWNLIE 245

ing character of this variety, the extruded re­ apparent in dried examples and which is not ceptacle, to be present only in some material found in material from other areas. They also from Fiji, and in that area the degree of extru­ lack the definite lobing on the lower edge of the sion varies considerably. Other Fijian specimens pinnules which is found in New Caledonian do not have the receptacle extruded at all. In specimens. These form A. aneitense. A . fulvum view of the range noted in this character, and as represented by material from New Zealand the fact that it is not consistent, it is better and Norfolk Island has rugose hairs on the up­ to regard this as no more than a variety re­ per surface of the pinnae, a character not noted stricted to Fiji. in any material from New Caledonia ascribed to this species. A summary of this discussion is Oleandra neriiformis Cav. given here. Oleandra ciliata Kl. Kuhn Linn. 36, 126, Rugose hairs on upper surface of pinnae . A. 1869; C. Chr. Ind. Fil. 466, 1906. fulvum (N ew Zealand and Norfolk Island). Christensen has already reduced many of the Rugose hairs absent on upper surface: Pacific forms of Oleandra to synonyms of the Pinnae subglaucous beneath .._.. .._ . widespread O. neriiformis, and an examination ______A. aneitense (Aneirum) . of a cotype of this species at the British Museum Pinnae not subglaucous beneath...... of Natural History shows that it does not go out­ ______A. novae-caledoniae (New Caledonia). side the range of characters now accepted for that species. SUMMARY

Adiantum novae-caledoniae Keys. The small serrated-frond species of the genus Specimens of Adiantum from New Caledonia Hymenophyllum are revised for the South Pa­ seem to have been variously identified as A. cific islands region, and the morphological char­ novae-caledoniae Keys., A. aneitense Carr., and acters and distributions are clarified for one A : fulvum Raoul, without any clear idea as to species of Arthropteris and three of Adiantttm, the distinguishing characters, if any, of these while one species of Oleandra is reduced to three species. A detailed examination of material synonymity, and a species of Mi crolepia is re­ from Aneitum, New Caledonia, Norfolk Island, duced to varietal rank. and New Zealand shows that they are not only easily distinguished morphologically, but are ACKNOWLEDGMENTS geographically isolated. All material from New Caledonia lacks rugose hairs on the upper sur­ I wish to thank Mr. F. Ballard of the Herb­ face of the pinnae, and the pinnules are usually arium, Royal Botanic Gardens , Kew, for his markedly attenuated. These represent A. novae­ assistance with this work, and I acknowledge caledoniae. Specimens from Aneirum also are my indebtedness to the Council of the Univer­ glabrous on die upper surface and are sub­ sity of Canterbury for provision of the refresher glaucous beneath, a character which is still very leave during which it was undertaken.