THE NATURAL HISTORY OF JUAN FERNANDEZ AND EASTER ISLAND EDITED BY DR. CARL SKOTTSBERG VOL. II BOTANY WITH ii6 PLATES UPPSALA 1 9 20-195 3 ALMQVIST & WIKSELLS BOKTRYCKERI AB LIBRARY NEW YORK BOTANICAL GARDEN Utgivet med understod av Langmanska kulturfonden UPPSALA 1953 ALMQVIST & WIKSELLS BOKTRYCKERI AB Table of Contents. Christensen, Carl and Skottsberg, Carl. The Pteridophyta of the Juan Fernandez Islands i Christensen, Carl and Skottsberg, Carl. The Ferns of Easter Island 47 Fries, Rob. E. Die Myxomyceten der Juan Fernandez-Inseln 55 Fries, Thore C. E. Die Gasteromyceten der Juan Fernandez- und Osterinseln 59 Skottsberg, Carl. The Phanerogams of Easter Island 61 Munster Strom, K. Freshwater Algae from Juan Fernandez and Easter Island .... 85 Skottsberg, Carl. The Phanerogams of the Juan Fernandez Islands 95 Brotherus, V. F. Musci Insulae-Paschalis 241 Borgesen, F. Marine Algae from Easter Island 247 SjosTEDT, L. Gunnar. Ein neues Sargassum von der Osterinsel 311 Zahlbruckner, a. Die Flechten der Juan Fernandez-Inseln 315 Brotherus, V. F. The Musci of the Juan Fernandez Islands 409 Zahlbruckner, A. Die Flechten der Oster- insel nebst einem Beitrag zur Flechten- flora von Juan Fernandez 449 BoYE Petersen, Johs. Marine Cyano- phyceae from Easter Island 461 Romell, L. Bacidiomycetes from Juan Fer- nandez 465 Keissler, Karl. Ascomyceten, Fungi im- perfecti und Uredineen von Juan Fer- nandez 473 Skottsberg, Carl. The Vegetation of Easter Island 487 Skottsberg, Carl. Pollinationsbiologie und Samenverbreitung auf den Juan Fernan- dez-Inseln 503 Keissler, K. Nachtrag zur Pilzfjora von Juan Fernandez 549 Evans, Alexander W. The thallose Hepa- ticae of the Juan Fernandez Islands . . 551 Setchell, William Albert. The Codiums of the Juan Fernandez Islands 587 Levring, Tore. Die Meeresalgen der Juan Fernandez-Inseln 601 Skottsberg, C. Marine Algal Communities of the Juan Fernandez Islands, with re- marks on the Composition of the Flora 672 Herzog, Th. Die foliosen Lebermoose der Juan Fernandez-Inseln und der Osterinsel 697 Levring, Tore. Die Corallinaceen der Juan Fernandez-Inseln 753 Levring, Tore. Einige Corallinaceen von der Oster-Insel 759 Skottsberg, C. Additional Remarks to "Marine Algal Communities of the Juan Fernandez Islands" 761 Skottsberg, C. A Supplement to the Pte- ridophytes and Phanerogams of Juan Fernandez and Easter Island 763 Skottsberg, C. The Vegetation of the Juan Fernandez Islands 793 -A I THE NATURAL HISTORY OF JUAN FERNANDEZ AND EASTER ISLAND EDITED BY DR. CARL SKOTTSBERG VOL. II BOTANY PART I I. C. Christensen & C. SKOTTSBERG: The Pteridophyta of the Juan Fernandez Islands, 2. C. Christensen & C. SKOTTSBERG: The Fems of Easter Island. 3. Rob. E. Fries: Die Myxomyceten der Juan Fernandez-Inseln. UPPSALA 1920 ALMQVIST & WIKSELLS BOKTRYCKERI-A.-B. I. The Pteridophyta of the Juan Fernandez Islands. By CARL CHRISTENSKN and CARL SKOTTSBERG. Witli 7 text figures and 5 plates. This paper is based on material collected by the undersigned, assisted by Mrs, Skottsberg, 19 16— 17. The study of the collection was begun in Sant- iago, where, thanks to the kindness of Mr. F. FuENTES, I had ample oppor- tunity to see the important herbaria in the Museo Nacional and also, with the kind permission of Professor F. JOHOW, could study his herbarium in the In- stituto Pedagojico. In order to solve some important questions and to classify some species which remained doubtful, I asked the prominent pteridologist, Dr. C. ChriSTENSEN, for assistance, and he willingly offered himself to make a thorough study of the entire collection. I have the great pleasure herewith to thank him for all the trouble he has taken in preparing the descriptions of the new species, in adding numerous critical notes on the systematic position and affinity and in making the fine and very exact drawings. The localities in Masatierra are enumerated from E to W along the north and south sides of the island, respectively; for Masafuera, the localities are enumerated from N to S along the east coast and then round to the west coast, followed by the observations made in the high mountain region. The Spanish names are used, with the following abbrevations: B. = Bahi'a, bay; C. = Cordon, mountain ridge; Co = Cerro, mountain; Pta = Punta, cape; Pto = Puerto, cove; Q. = Quebrada, valley with steep sides, gorge, canyon; V. = Valle, valley. — Maps will accompany Vol. L Unless otherwise stated, all species have been observed by myself in all the places quoted. The numbers in brackets refer to our collection, of which specimens are kept in Stockholm (Riksmuseum), Gothenburg, Upsala etc., and in several foreign herbaria. A ! after the name of a collector signifies that I have seen the original specimen. An * before a latin name indicates a new addition to the flora of the group; if put before the name of one. of the islands, it s gnifies that the species was found for the first time in the island in question. The altitudes are in meters above sea level. I — 80206, The Nat. Hist, of Juan Fernandez and Easier Isl. .Vol. II. CARL CHRISTENSEN AND CARL SKOTTSBERG The remarks on fertility refer to my observations in 1908' and 1916— 17; where I have reason to beheve that a species occurs with ripe sporangia also at other seasons, these remarks are put in brackets. We have adopted the nomenclature followed by CHRISTENSEN in his » Index Filicum», where also explanations of the abbreviated book titles will be found. C. Skottsberz. Hymenophyllaceae, Trichoiiianes L. I. T. Philippianum : Smrm, Enum. pi. cr. vase. chil. 38 (1858); v. d. Bosch, Syn. Hymen, ed. Goddijn, Meded. Rijk's Herb. Leiden no. 17: 24 f. 13 (191 3); C. Chr. Ind. 647. — Syn. 7. dichotoinuni Philippi Anal. Univ. 1856: 169, Bot. Zeit. 1856: 650, Ann. sc. nat. II. 7: no; Hemsl. 69; Johow 1893:44, 1896: 173; non Kze 1847. Fig. I. Damp montane forests above 500 m., epiphytic on trunks of treeferns. (Fertile Jan. — April.) Masatierra: the high ridge between Q.^ Laura and Q. Piedra agujeriada, on Dicksonia, c. 650 m (no. 594); C. Centinela, on trunks of ferns, 580 m (no. 616); NE. slope of La Damajuana, on Blechmim cycadifolium c, 500 m (no. 351); C. Salsipuedes, Q. Helechos, on Dicksonia, 660 m (no. 283). T. dicJiotomiiDi Phil, was described from specimens collected by GERMAIN (!). No locality except >.Juan Fernandez» was stated. In Herb. Kew is a specimen collected by Reed(!) and labelled »Valdivia», probably in PHlUPPl's hand-writing (Hemsley, 1. c). It is possible that this locality is false: Reed made ex- tensive collections in both places and the localities may have been confounded. Until further investigations can be made, the species should be listed as endemic. T. Piiilippianuui is one of the most distinct species of the genus. It has been described and figured in details by VAN DEN BosCH, but his description and figure were not published until lately by GODDIJN. The numerous spe- cimens at hand correspond fairly well with this description, but the species grows much larger than his figure shows. The largest leaves are over 7 cm long including the 2 cm long stipe. The filiform, wide-creeping rhizome and the basal part of the stipes are sparsely furnished with reddish hairs, the plant otherwise being perfectly glabrous. The leaves are either scattered or fasciculate in small tufts of 2 — 7 on short branches from the rhizome. The lamina is exceedingly thin, yellowish green, the cells larger than in any other species of the genus, clearly seen * Botan. Ergebii. der Schwed. Exp. nach Pa'tagonien und deni Fcuerlande. IV. K. Svenska Vet.-Akad. Handl. Bd 51, Xo. 9 (1913). * For abbreviations, see p. i. THE PTERIDOPHYTA OF THE JUAX FERNANDEZ ISLANDS 3 through a weak lens. In shape and division the leaf varies considerably; generally it is oblanceolate, broadening upwards from the cuneate, shortly decurrent base; subdeltoid, deeply pinnatifid fronds, leaving a broad even wing to the midrib are, however, not uncommon. Segments erect, 2—2,; mm broad, obtuse or emarginate or deeply furcate at apex, very unequal in length, the larger ones often producing a secondary segment on the upper side near the base. Margins even, not at all thickened. Veins not regularly pinnatifid but rather repetitiously dichotomous, not atrofuscous as says BosCH, terminating in the emargination; spurious veins none. Sori rarely terminal on the main branches Fig. I. Trichomanes Philip/^ianutH Sturm. Plants showing different tvpes of furcation, na: size, and enlarged sorus. of the veins but generally confined to the basal secondar\- segments, deeply immersed; indusium turbinate with a widened entire mouth; receptacle long exserted but easily broken ott. Van den Bosch placed the species in his genus Goiiocormis, but it may be seriously doubted whether this position is natural. It looks ver\- much different from all other species of that group, all of which belong to the Old World. In general habit and size it resembles the Polynesian T. huniile Forst., but it is certainly not nearly related to it. B\- its cellular structure it is quite unique. Area of distribution: Probably endemic in Masatierra. 2. T. Ingae C. Chr. nov. spec. Syn. T. pyxidifenan L., Hemsl. 69; Johow 1893: 45, 1896: 173. EutricJioiuayies e grege T, pyxidiferi L., rhizomate intricato. filiformi; 4 CARL CHRISTENSEN AND CARL SKOTTSBERG foliis atroviridibus, planis, glabris, lanceolatis, maximis vix 5 cm longis et i cm latis, stipite i — 1,5 cm longo, e medio sursum anguste alato, rhachide tota alis sub 0,5 mm latis utrinque alata. Lamina bipinnatifida, pinnis 5 — 6-jugis, alternis, suberectis, infimis paulo abbreviatis, mediis 5 — 8 mm longis, inaequi- lateralibus, basi superiore segmentum soriferum gerentibus, parte exteriore in 2 — 4 segmenta sterilia contigua subtruncata vel leviter emarginata divisis. Mar- ginibus, praesertim versus apicem, minutissime et irregulariter crenato-dentatis. Venis distinctis, spuriis nullis. Cellulis parenchymaticis parvis, irregularibus, 4 — 6-angulatis, fere isodiametricis parietibus crassis, In'alinis. Soris singulis axillaribus, rarissime binis in pinna, tubo c\"lindrico anguste marginato, ore in- tegro vel levissime emarginato, paulo expanse; receptaculo exserto. Fig. 2. In moss mats on stones and on the ground, in dense forests, rare. (Fertile Aug.— Dec.) Fig. 2. Tricho))iaties Ingae C. Chr. a two fronds, nat. size; b fragment, X 2'/?, f sorus and d tip of sterile pinna, more enlarged. Masatierra, central part: \'. Anson, XK. slope of El Vunque, 300 m (no. 162), in the same place \\here it was found in 1908; XE. slope of La Damajuana, c. 450 m; V. Colonial, O Gutierrez (no. 68). — Found before only by MoSELEV and by Skottsberg, 1908. This new species, named in honour of Mrs. IXGA Skottsrerg, belongs to a number of closely related species which by most authors are united under the name T. pyxidiferum L. Convinced of this being a collective species that includes several valid ones, a number of which have already been described by VAN DEN Bosch, we have tried, but without success, to identify the Juan Fernandez form with one or other of the forms from the American continent. Our new species differs greatly from the true West Indian T. pyxidiferujii b}' its much less divided lamina and by its dark colour; in these characters it approaches some South Brazilian forms, especially T. cmarginatum Pr., but it differs from these as well as from all other forms known to us by its cellular structure and its uneven margins. The cells are proportionally very small, very unequal in size and seldom longer than broad; by this structure the leaf becomes much firmer than in all other forms without spurious veins. Seen TRE PTERIDOPHYTA OF THE JUAN FERNANDEZ ISLANDS 5 through a strong lens, the margins of the outer parts of the segments are irregularly toothed or crenate by outwards vaulted marginal cells, the external wall being thick and hyaline. A similar feature was not found in a large number of specimens of T. pyxidifeniui sens. lat. examined. The numerous specimens collected have all the same characteristic habit, with the outer seg- ments of the pinnae very close. The diagnosis above was based on the larger fronds; most of them are considerably smaller. Area of distribution: Endemic in Masatierra. 3. T. cxsectum Kunze, Anal. 47 tab, 29 f. 2 (1837); Hemsl. 69; Johow 1893: 44 f. 32, 1896: 173. — Syn. Hyntcnopliylluui fucifoniic Colla 1836: 33 tab. 63. In dense humid forests, epiphytic on trees and trunks of arboreous ferns, and in damp recesses of the narrow gorges, among moss, hanging down from the rocks. (Fertile Jan. — April.) Masatierra: C. Chifladores, above the innermost part of the Frances Valley, c. 500 m; ridge between Q. Laura and Q. Piedra agujeriada, on Dick- sonia 650 m (no. 595); C Centinela, on fern trees 560 m; El Pangal, wet forest in the gorge (no. 1130); B. Cumberland, in Cave VI, forming a beautiful pure carpet on the vertical walls (no. 598; also JoHOW); Q. Juanango, under fallen trunks (no. 608); Q. Villagra, mountain spur W. of El Yunque, c. 500 m. Masafuera: Q. del Mono, c. 475 m; Q. de las Casas, under overhanging rocks in the canyon (no. 450, also JOHOw); Q. del Blindado, 440 m; O. de las Vacas (Jo how). The nearest ally of this pretty species, that reaches 40 cm in length, is T. teneriiDi Spr., widely spread through tropical America. Area of distribution: South Chile, Valdivia to Chiloe; Juan Fernandez. Serpyllopsis v. d. Bosch 4. S. caespitosa (Gaud.) C. Chr. Arkiv for Bot. 10: 29 c. fig. (1910). — Syn. Hymenophylhan caespitosum Gaud. Ann. sc. nat. 5: 99 (1825); Freyc. Voy. Bot. 374 tab. 5 f. 2 (mala); TricJwvianes caespitosum Hook. Sp. Fil. i: 132 tab. 40 B; C. Chr. Ind. 637; Serpyllopsis antarctica v. d. Bosch Ned. Kr. Arch. 4: 377 (1859)- Var. fernandeziana C. Chr. et Skottsb. nov. var. This interesting moss-like fern was first detected in the islands by Skotts- BERG, 1908. In 1910 C. Christensen referred the specimens of the 1908 collection to var. densifolia (Phil.) C. Chr. 1. c. 30 [Hyin. densifoliiim Phil. Linn. 29: 108, 1857), which was collected in the Guaitecas Islands and presum- ably is not rare in the humid forests of the Chilean Archipelago. The specimens from Juan Fernandez were sterile; now we have a beautiful lot of specimens, some of which are fertile, and an examination of the sori has proved that we O CARL CHRISTENSEN AND CARL SKOTTSBERG have to do with a form different from dcnsifolia. It may be described as follows. Lamina lutco-viridis, pinnis plerumque ovatis secus costam parcissime pilosis, ut in var densifoHa; differt soris profunde immersis utrinque late foliaceo- alatis, fere ad medium bipartitis, labiis rotundatis, distincte dentatis, extus 2 — 4 cristis iongitudinalibus dentatis instructis, receptaculo ad maturitatem exserto. In the moss carpets on trunks and branches of trees, in the elevated forests of Masatierra, where it has not been found below 450 m; on Masafuera also in the heath of the 'highland, among rocks. (Fertile in Jan.) Masatierra: C. Chifladores, in the innermost part of the Frances valley, on Drhiiys Winteri c. 500 m (no. 1128); ridge between Q. Laura and Q. Piedra agujeriada c. 650 m (no. 1129); on the main crest of the island, above Pangal, on Dnuiys c. 800 m (no. 1127); V. Colonial, along the road to Portezuelo, on Dninys c. 450 m (no. 136); C. Salsipuedes, brushwood on the ridge, 615 m, and in O. Helechos, on Drintys, 660 m (no. 284). Masafuera: Heath above O. del Mono, c. 850 m (no. 421); rocks at Las Torres, 1350 -1370 m (no. 421 b — see below); S. slope of Los Inocentes, in Dicksoiiia-iorest on the trunks, c. 950 m; also near the highest summit, 1450 m in moss mats (no. 381); Q. Loberi'a, between blocks in the stream. Var. dcnsifolia has 1 — 8 sori terminal on the pinnae, the lips obscurely toothed and faintly crested, the wings on both sides much narrower. The typical form of the species, from the F"alkland Islands, has terminal sori with very short lips; it turns brown when dried and is much more hairy, with more oblong pinnae. Var, fernandeziana is generally small with only 2—5 pairs of pinnae which are close, often imbricated and clasped along the midrib. There is, however, also a form with more distant and more oblong pinnae (no. 421 b), a form corresponding to var. elongata Hook, of the type; the special characters are due to external conditions (moisture and shade). Area of distribution: South Chile to Fuegia, Falkland Islands; the var. fernandesiana endemic in Juan Fernandez. Hymenoglossum Presl 5. H. cruentum (Cav.) Pr., Hymenoph. 35 (1843); v. d. Bosch, Ned. Kr. Arch. 4: 385. — Syn. HymenopJiylltim cruentjuii Cav. Descr. 275 (1802) et auctt., Hemsl. 6-]-^ Johow 1893: 40 f 26, 1896: 169; C. Chr. Ind. 359. Plate I F"ig. i. In the montane forest belt, with moss on the soil or on trunks of trees, scarce. (Fertile Dec— Jan.) Masatierra, not rare in the central parts: V. Anson, NPl slope of La Damajuana c. 430 m (no. 353); slopes of El Yuncjue, 480 m; V. Colonial, C. Central 535 m; along the road to Portezuelo, c. 500 m (no. C)"]); north wall of Co Piramide, a narrow ledge 600 m; O. Villagra, mountain spur VV. of El Yunque, c. 500 m and SW. face of Co Piramide, among bruslnvood, 590 m. THE PTERIDOPH\TA OF THE JUAN FERNANDEZ ISLANDS 7 *Masafiiera: O. de las Casas, humid moss mats in the inner canyon, scarce (no. 473); C. del Barril, in moss on rocks, c. 750 m (no. 407). — New for this island. The specimens are small and sterile. In restoring the genus Hynicnoglossimi we follow VAX DEN Bo.SCH, the monographer of the famih-, contrary to all other pteridologists. Our reasons for doing so are not only the unique habit of the plant, but are based upon a very essential character, overlooked by Presl and all later authors. The margins of the frond, not the sterile parts alone, but also the valves of the indusium, are bordered by a thick cbeneous line. The veins terminate in the parenchyma within this line. Area of distribution: South Chile, Valdivia to Chiloc and Western Patagonia to 49°; Juan Fernandez. Hymenoph villi 111 Sm. 6. II. cuncatum Kze. Anal. 50 (1837); C. Chr. Ind. 359, Arkiv for Bot. 10: 21 (1910). — Syn. H. raruvi auctt. quoad plant, amer.; Hemsl. 6%\ Johow 1893: 42 f. 30, 1896: 171. H. polyanthos auctt. quoad plant, fernand.; Hemsl. 6t\ Johow 1S93: 42 (pro pa tc.- non f. 29), 1S96: 172 (pro parte.-). H. dichoto- iiium. Johow 1893: 44 (pro parte.-) f. 27, 1896: 173 (pro parte.-). Fig- 3- Probably fertile at any time of the year. Under this name whe unite all the smaller glabrous Hymenophylla of the islands, being convinced that they all belong to one species, although the ex- treme forms look very different. All forms agree in cellular structure and in shape of the indusium, which is broad, flat with a broad subtruncate or broadly cuneate base and with rounded or sometimes subacute valves. The indusium varies, as our figures show, in the same degree in the largest and smallest forms. KUNZE described the stipe as unwinged. We find among the very ex- tensive material several individuals which correspond perfectly to his description (no. 396); they have ovate or subdeltoid, middle-sized, yellowish-green fronds with wingless stipe and broad, semiorbicular indusia. Other plants agree with these in all respects but have the stipe winged to the base (no. 1145). ^^ e suppose that these are younger plants, while fronds without wings are older, having lost the wings. Large or small fronds occur with or without a winged stipe. If no. 396 is considered to be a typical, middle-sized form, the other specimens may be arranged as follows. l) Large plants, over 15 cm long incl. stipes, more decompound, with narrower, ovate-lanceolate lamina, often of a deeper green (no. 285\ Such plants have been referred to H. polyanthos Sw. by all authors, and probably also the more typical form was brought to this species by some. It is very difficult to find a single constant character by which the larger forms may be known from the variable H. polyanthos; the lobes are generally broader, the 8 CARL CHRISTENSEN AND CARL SKOTTSBERG valves of the indusium more rounded and flat, but certain forms of H. poly- antJios show essentially the same characters. The best mark is, perhaps, the glabrous base of the stipes in //. cunealum^ while most forms of //. polyaiitJios have the stipes pilose below. 2) Smaller plants, often much reduced and approaching //. ravuni R. Br. (nos. 146, 1 143), and referred to that species by all authors. The leaves are yellowish green, i — 3 cm long, the pinnae often more or less imbricate (no. 146). It must be admitted that some of these dwarf forms closely resemble H. rarum in most characters, but we find leaves with and without a winged stipe which otherwise are identical, and //. raniin has the stipe unwinged; further, all possible transitions between these dwarf forms and the larger ones are found. None of the specimens show the brownish colour of the typical H. rarum. Summarizing what has been said above, we find it convenient to distin- guish two types. 1. typicum. (//. cuncatuni Kze, H. poly ant Jios auctt.) — Fig. 3 a — d. One of the commonest ferns, found in all the forests in moss carpets on the soil, on stones or on trunks of trees, in the eastern and central parts of Masatierra, where it ranges from about 200 m above sea level to the highest parts reached, viz. the main ridge above Pangal, 795 m. It is also common among the brushwood on the rocky ridges and, of course, on Dicksonia. — Probably fertile at all seasons (Dec. — April, Aug.). Masatierra: numerous localities from Pto P'rances to Pto Ingles (nos. 46, 66, 70, 285, 300, 591, 1 144, 1 1 77; nos. 66 c and 70 form a transition to the following; both collected on rocks). Masafiiera: Q. del Mono, in forest 570 m (no. 1146); O. del B indado in forest (no. 440); subalpine and alpine heath, 850 — 1370 m (nos. 396,429, 1146, 1147); Q- Loberia, in moss on boulders in the woods, c. 300 m (no. 1145). 2. var. rarifornie C. Chr. et Skottsb. — Fig. 3 e — g. This is exactly the 77. rariwi of JOHO\v(!), often found on old trunks of Blechnuin cycadifolium, but also forming dense, almost moss- like patches on exposed ground. Leaves often i cm long only, generally deltoid in outline. Masatierra: At the road to Portezuelo, c. 450 m (no. 1143); on the sharp crest separating V. Anson from Q. Damajuana, on the peak, 365 m (no. 146). — f. imbricata; Fig. 3 h: Pto Ingles, the crest of the central ridge (no. 1178). Area of distribution: South Chile, Valdivia to Western Patagonia; Juan Fernandez. Tasmania? New Zealand.' South Africa? 7. H. caudiciilatum Mart. Ic. Crypt. Bras. 102 tab. 6-] (1834); Johow 1893: 41, 1896: 170; C. Chr. Arkiv for Bot. 10: 23 (1910). In the the elevated forests of the foggy region of Masatierra, very rare; apparently more frequent in Masafuera, found both in the woods and in the bottom of the deepest canj-ons. — (Fertile Dec. — Febr.) THE PTERIDOPIIVTA OF THE JUAN FERNANDEZ ISF.ANDS g Masaticrra: Just below the gap between l'",l Wituiuc and La Dama- jiiana, c. 500 m (no, 160); North face of Co Piraniidc, narrow ledge, c. Fig. 3. Hyiiienopliylhiiii aaieaiinn (Cav.) Pr. a typical ctineatiun (no. 596), with enlarged sori, indusium acute or rounded; in one are two sori; i^ large /i9/)'a«//ii:v-like frond with enlarged tip, showing acute indusia (no. 285); c blunt sorus from a plain like the former (no. 66); d fioly- a7ithos\\Vt frond with winged stipe (no. 1145); e—f rariforme with stipe and indusiam of dif- ferent types (no. 1143); g small rariforme (no. 146); h i. imbrkata (no. 1178). All plants nat. size. lO CARL CHRISTENSEN AND CARL SKOTTSBERG 600 111 (no. 360). — Only collected once before in Masatierra: Rkf,I)(!) without locality. *Masafuera: Q. del Mono, dense forest, 475 ni (no. 388); Q. de las Cliozas, in the forest; O. de las Casas, innermost part of the canyon, under stones below the cascade, 215 m (no. 465); 0. Loberia, under stones and in moss between huge blocks, 180—300 m (no. 481). — New for Masafuera, where fine, richly fertile specimens were collected in O. de las Casas. The specimens belong to the typical form. Area of distribution: South Chile, Valdivia to Western Patagonia (49° S. L.); Juan Fernandez. South Brazil. 8. II. fuciforme Sw. Syn. 148 (1806) et auctt.; C. Chr. Ind, 361 (where for » China austr.» read » Chile austr.»); Hemsl. 6-]\ Johow 1893: 41, f. 28, 1896: 170; C. Chr. Arkiv for Bot. 10: 24 (1910). — Syn. H. scinitcrcs Colla 1836: 32, tab. 61. Humid montane forests in rich soil or on decaying trunks, not observed below 480 m. (Fertile Dec. — ^Jan.) Masatierra: Slope of C. Chifladores, above the Frances valley, c. 500 m; V. Anson, slopes of El Yunque, 480 m; near the road to Portezuelo, c. 500 m (no. 138); north face of Co Piramide, with the preceding; Q- Villagra, mountain spur W. of El Yunque, c. 500 m and at the road-side in the highest part of the Villagra valley below the pass, c. 500 m (also JoHOW). This species, the largest of the genus, varies in cutting from tri- to quadri- pinnatifid. The less divided form is not, as we supposed in 1910, the only one found in the island. Specimens in the present collection are quite as much divided as specimens from Chile. Area of distribution: South Chile to Western Patagonia; Masatierra. 9. H. ferrugineum Colla 1836: 30; C. Chr. Ind. 361. — Syn. H. sub- iilisshmun Kze. Anal. 49 (1837); Hemsl. 6'^\ Johow 1893: 43 f. 31. 1896: 172; H. Bcrteroi Hook. Sp. fil. i: 93 tab. -^^^ C (1844); Gay 6: 534. Damp, shady forests of the highest mountain slopes, not found below 480 m; characteristic of the darkest corners in the fern groves. (Fertile in Jan.) Masatierra: C. Chifladores, woods in the highest part of the Frances valley, on Dicksonia c. 500 m; the ridge between Q. Laura ;ind O. Piedra agujeriada, c. 600 m and high up in this valley, 650 m; C. Centinela, under old stumps, 530—560 m; V. Anson, slopes of El Yunque, 480 m (no. 159, also P. Arredondo in Herb. JOHOWl); V. Colonial, north face of Co Piramide, with the preceding; C. Salsiquedes, Q. Helechos, on Dicksonia 660 m (no. -jG); Q. Villagra, high up near the pass, rare. *Masafiiera: O. de las Chozas, dark fern grove under Dicksonia, c. 700 ni (no. 535). — New for this island. With this characteristic species most authors unite //. Fraukliniac Colenso from New Zealand, to which it is so closely related that it is not unnatural to THE PTERIDOPHYTA OF THE JUAN FERNANDEZ ISLANDS T I unite them into one species. Specimens from Chile are scarcely to be distin- guished from others from New Zealand, often having distant pinnae, the lower ones being much abbreviated. All specimens from Juan Fernandez, the type localit}-, have closer segments and an ovate lanceolate lamina, with the lower pinnae not or very slightly shortened. //. acrug'uwsuui (Thouars) Carm. from Tristan d'Acunha is another near relative, but much smaller than either (see liemsl. Chall. Rep. Hot. I. 2, tab. 38). A fourth related form grows in the caves of Amsterdam Island in the Southern Indian Ocean. It was, we believe without sufficient reason, referred to //. obtusum Hook, et Am. b}' Hkmslev (1. c. 271), a Hawaiian species; more properly it might be reduced to H. acnighwsnui as a variety. It is an interesting fact that four so closely related forms are scattered in remote localities over such a large area of the Southern Hemisphere. They are, we should guess, daughter forms of a single .species that in old times was widely dispersed over the Antarctic continent. It must not be forgotten, however, that some of the forais ma}- be derixates of the common tropical H. cilia turn. Area of distribution: South Chile, Valdivia to Fuegia; Juan Fernandez. Xew Zealand. *io. H. falklandiciim Baker, Syn. fil. 68 (1867); C. Chr. Arkiv for Bot. 10: 25 f. 3 (1910). Fig. 4 g— h- A new addition to the flora, growing both in the moss carpets of the forest floor and on trees, as well as in the alpine heath. (Fertile Feb. — March.) Masafuera: O. del Mono, in forest, 570 m (no. 389); 0. del Blindado, in forest, 440 m (no. 398); 0- de las Casas, far into the canyon, with Hepatics, c. 200 m (no. 474); alpine heath near the Correspondencia Camp and at Las Torres, 1 100 — 1370 in (nos. 1142, 1 176). Widely distributed through Subantarctic America and very likely not dif- ferent from H. Menziesii Pr. Hym. 51 from Staaten Island. Still Presl de- scribed his species as having obovate, subglobose, pedicellate sori, while H. falklandiciim has ovate-lanceolate valves of the indusium. The specimens from Masafuera are, as could be expected, larger than those from the heaths of Fuegia and the Falkland Islands and also of a more vivid colour, but in essential characters they do not differ. The largest fronds measure nearly 10 cm in length and are of the narrow lanceolate shape shown in fig. 3 a — b quoted above. The distinctly unilateral pinnae bear from i to 4 toothed segments on the upper side. The sori are single, or, in larger fronds 2 to each pinna, distincth" stipitate, with ovate-lanceolate, obtuse and quite entire valves, which turn blackish brown with age. By its narrow lanceolate fronds and especially by the long dark brown valves very difterent from all forms of H. peltatiaii. A specimen of the 1908 collection from Masafuera was 1. c. referred to H. pcltatinn. This is incorrect; probably the bad specimen should be referred to the next species rather than to H. falklandiciivi. 12 CARL CHRISTIiNSEN AND CARL SKOTTSBERG Area of distribution: Southern West Patagonia, Fuegia, Falkland Islands; South Georgia; Masafuera. *ii. H. riigosuni C. Chr. et Skottsb. nov. spec. Syn. H. Uinbyidgensc of Gay and Phil.? Hemsl. 6g; Johow 1893: 44, 1896: 172. Fig. 4 a-f. Leptocionmm rhizomate filiformi, longe repente; foliis remotis, stipitibus erectis, filiformibus, glabris, omnino e.xalatis, saepe flexuosis, 3 — 4 cm longis. Fig. 4. a—f Hyi/ietiopkyllitin rugosum C. Chr. et Skottsb. a well developed frond (no. 630); b leaf recalling H. tunbridgense (no. 423); c f. lanceolata (no. 1175); d fragment of a, X 4; e sorus, X lo; f piima of c, X 4; g — /' H. falklandictiin Bak.; g two pinnae (no. 389), X4; h sorus, X 10. All entire fronds in nat. size. Lamina formae typicae ovato-deltoidea vel ovato-lanceolata, 5 — 6 cm longa, 3 — 4 cm lata, brunnescente, perfecte plana, saepe curvata, subtus papillis bre- vissimis ad rachidem venasque rugosa, tripinnatifida; rhachide flexuosa, alata, alis denticulatis. Pinnis alternis, circ. 10-jugis, infimis plus minusve abbreviatis vel subaequalibus, aequaliter pinnatifidis vel basi inferiore paulo excisis, supre- mis minoribus distinctius inaequilateralibus. Pinnulis obliquis, inaequilateralibus, in segmenta 3 — 4 divisis, marginibus ubique acute denticulatis. Venis distinc- tissimis, nigriscentibus. Soris i — 8 pro pinna, praesertim in segmentis lateris superioris pinnarum insidentibus. Indusiis liberis nee immersis nee stipitatis, THE PTERIDOPHYTA OF THE JUAN FERNANDEZ ISLANDS 1 3 fere ad basin in valvas diias ovato-elongatas, brunncas vel nigrescentes, apice distincte et acute dentatas divisis; receptaculo incliiso. In the higher woods and humid heaths of the highlands, apparently not ver}- rare. (Fertile Feb. — April.) Masaticrra: I'"Je\ated slopes of O. Piedra agujeriada, on fern-trees, 650 ni (no, 592); C. Centinela, dark humid forests, c. 560 m (no. 1140); the longitud. ridge, above Pangal, on fern trees, c. 800 m; Q. Villagra, mountain spur \V. of El Yunque, c. 500 m (no. 630). Masafiiera: Woods on tlie Sanchez plain, 515 m (no, 1137); heath of the Correspondencia Camp, in the ravines, 1 100 — 1200 m (nos. 423, 1 139); C. del Barril, c. 750 m; O, del Blindado, in the forest, on ferns, 440 m (no. 397); near the summit of Los Inocentes, c. 1400 m (no. 1138). f. lanceolaia: minor, lamina lanceolata, 3 — 5 cm longa, raro supra i cm lata, alis rachidis vix denticulatis; habitu, magnitudine et pinnis omnibus inae- quilateralibus H. falklandico sat similis, sed differt et cum f. t)-pica congruit soris non stipitatis nee non valvis distincte denticulatis. — Fig. 4 c. Masafuera, at the Correspondencia Camp, in the ravine, with the type, 1 160 m (no. 1 175). This new species is to H. tunbridgcnsc what H. fcxlklaudicuvi is to H. peltaUivi. It is very variable in size and shape of the lamina, but all forms have the same characteristic brown or dark brown indusia with elongated, toothed valves, similar in outline to those of //./<:7//^/^?;/^//V/cw, but never stipitate and alwa\-s dentate (comp. Fig. e and //). The larger, ovate-deltoid fronds recall H. dicJiotovium, but they are quite plane and the stipe is wingless; the sori are very different, the mature sporangia never exserted. Middle-sized leaves with short, ovate-oblong lamina with all pinnae unequal-sided (Fig. b) can scarcely be distinguished from // tunbridgense when sterile, but leaves with mature sori at once show the difference; true H. tiDibridgoisc never has so elongated valves, which never are brown. There is little doubt of such forms having been referred to H. tunbridgense, a species scarcely occurring in the islands. The f. la}iceolata is very similar to H. Jalklajidicuin in general habit, but is easily known by its sessile indusia and toothed valves. The collection contains many intermediate forms between the extremes. Area of distribution: Endemic. 12. H. pectinatum Cav, Descr. 275 (1802); Johow 1S93: 42, 1896: 170. (Fertile, March.) Masatierra: Reed (IHerb. Kew) without exact locality. *Masafuera: In the highest part of the island, under rocks etc. in the alpine heath; the Correspondencia Camp, in the ravine, ii6om (no. 1141); Las Torres, in wet moss under rocks, rare, 1370 m (no, 476); S. slope of Los Ino- centes, in a fern grove, 950 m, on Dicksonia, and in the heath near the summit, c. 1400 m (no. 378). — New for this island. The specimens are very poorly fructiferous; the largest ones have most of the secondary segments deeply cleft, thus somewhat resembling H. secundum. 14 CARL CHRISTENSHN AND CARL SKOTTSBERG Area of distribution: South Chile, Valdivia to Fuegia; Juan Fernandez. *I3. H. secundum Hook et Grev. Ic. Fil. tab. 133 (1S29); C. Chr. Ind. 367. (Poorly fertile, Feb.) Masafuera: Las Torres, edge of the western iirecipice, 1370 m (no. 475). New to the islands. The specimens are small but otherwise typical. By the entire valves of the indusium surely to be distinguished from //. rugosmn. Area of distribution: South Chile, Valdivia to Fuegia; Masafuera. 14. H. dichotomum Cav. Descr. 276 (1802); Hook. Sp. P'il. i: 98 tab. 36 A et auctt.; Hemsl, Gy, Johow 1893: 40 (pro parte?, certe non f. 27), 1896: 169 (pro parte); C. Chr. Arkiv for Bot. 10: 27 (1910). — Syn. H. nigricans Colla 1836: 32 tab. 62. H. polycmtJios Johow 1893: 42 (pro parte.?) f. 29, 1896; 171 (pro parte.'') Probably fertile at almost any season. One of the commonest species, found on the low, dry slopes of western Masatierra as well as in the central and eastern parts, where it is frequently ir^et with in the wooded region from its lower limit to the highest ridges; also common in Masafuera. It grows on the soil, on stones or on trees, both fallen and living ones. Masatierra: eastern and central parts, common (nos. 45, 218, 571, 593, 1 133, 1 134); the dry western part: ridge SW of Co Tres Puntas, 300-350 m, under stones (no. 1132), and in the bottom of dried-up ravines (no. 1 132 b). Masafuera: Q. del Mono, on trees, 475 m (no. 390); Q, de las Casas, in the canyon (no. 451); Q. del Blindado, common in the forest 440 m; Z>/r/'j>-6';/M- grove on the lower slope of Los Inocentes, above O. del Varadero, c. 950 m; Q. Loberia, between blocks in the streem, 180 m (no. 480). This common species varies considerably in size and shape of the frond, measuring from 2 — 3 to 20 cm in len^^th, in shape broadly ovate to lanceo- late, but all forms have the very characteristic plicate frond of a dark green colour. Young leaves have distinct brown, crisped wings along the stipe, much narrower, however, than those of H. tortuostaii; these wings are shed early so that older leaves have an unwinged stipe. The marginal teeth are much shorter than in H. toi/nosum. The sori are confined to the upper third of the frond (which is often more or less contracted), where they are numerous. The ovate-subacute, frequently apiculate valves of the indusium are entire and faintly or not at all spinose on their back, as described and figured by HoOKER. The large brown sporangia are exserted in the mature sorus. The smaller forms have the leaves very much contorted and look peculiar when compared with the larger ones. We think that H. mgricans Colla is this species; at least Bertero's spe- cimens in Herb. Kew(!) under this name clearly belong here and not to H. tortuositm to which Gay referred Colla's species. Evidently this species has been largely confounded with others. It is perhaps the most common of the genus in the islands, but, nevertheless, JOHOW only quotes other collectors and not himself, this because he misunderstood the species, for his fig. 29, 1893, is not »H. polyanthos», but //. dicJiotoiiiuDi, THE rrr.RinopiiYTA of rnv. juan fkrnanoez islands 15 while, on the other hand, his fl<,^ 27 »II. (Hchotonuim* illustrates a form of //. ciincatmn. Area of distribution: South Cliile, Vaiciivia to Western Patagonia; Juan I'Y'rnandez. 15. II. tortiiosum Hook, et Grev. Ic. Fil. tab. 129 (1829); Henisl. 69; Johow 1893: 43, 1896: 172. Apparent!)' rare, but easily confounded with the former. (I'^erlile Feb.) Masaticrra: Reki)(!) in Herb. Kew. *Masafuera: Q. del Blindado, in woods, 440 m (no. 1136); heath at the upper course of Q. del Mono, c. 850 m (no. 1135). — New for this island. To this species we refer two sets of specimens from Masafuera. They differ from the common //. dichotonnim in the rather persistent, broad, crisp, brown wings of the stipe, the non-plicate lamina, which is more or less undulate with much larger marginal teeth. The segments are closer and the larger l)iiinae have the pendent tips found in most specimens from the continent. The valves of the indusium are rather acute and nearly entire, thus approaching H. ih'c/ioioiniiJH. H. tortuosiim was also collected in Masatierra by Philippi and DowN'iON, t. Hemslev. We have not seen the former; the latter (Herb. Kew!) belongs to dicliotoDiinn. Area of distribution: South Chile, Valdivia to Fuegia, Falkland Islands; Juan Fernandez. \H. renifonne Hook, is mentioned by PllILlPPl (Anal. Univ. 1856: 160, Bot. Zeit. 1856: 631) who undoubtedly based his statement on Gay 6: 535. It has not been found by any of the known collectors, and Gay does not tell where lie got his information as to its occurrence in Juan Fernandez. As far as we know there are no specimens in any herbarium from this place, so that it must be excluded from the list. H. dicranotridnnn (Pr.) Sadeb. (Syn. H. chiloense Hook,)- is listed by JoilOW 1893: 44 and 1896: 172, as gathered by Reed in Masatierra. The determination is probably incorrect. In Herb. Kew.(!) are specimens of H. dentatuin Cav. (Syn. H. Bridgesii Hook.) labelled (by PuiLlPPl) Juan P^ernandez oct. 1872 Reed. As no author lists this species, the locality remains doubtful.] Cyatheaceae. Thyrsopteris Kunze 16. T. elegans Kze Linnaea 9: 507 (1834), Farrnkr. tab. i, et auctt.; Hemsl. 66\ Johow 1893: 22 f. 4, 1896: 155. — Syn. Panicularia Berteri Colla 1836: 35 tab. 64. Characteristic of the upper montane forests in Masatierra, generally not descending below 400 m; belonging to the highland heath in Masafuera, where 1 6 CARL CHRISTENSEN AND CARL SKOTTSBERG it is rather poorly developed. The fertile pinnae are developed during the summer; in the beginning of December we collected fronds with very young sori and still in the end of January no ripe ones were observed; in the later half of March the sporangia were ripe and the spores disseminated. Masatierra, not rare in the central part: C. Centinela c. 530 m; V. Anson, slopes of El Yunque, 480 m; C. Central, c. 500 m; along the road to Portezuelo in several places (lowest station observed 208 m, at a small stream), especially near the pass, c. 500 m (no. 14); C. Salsipuedes, open brushwood on the narrow ridge, 615, 625 m; Q. Salsipuedes, hght woods, c. 500 m, numerous; Q. Helechos, not rare in Z)/V/^j^i?///V?-groves; Pto Ingles, on the cential ridge, 470 m; Q. Villagra, frequent in the humid woods below the pass (no. 277), but also seen on a low ridge just below the lower limit of trees, c. 200 m. Masafuera: Q. de las Casas (Jonow); subalpine heath, in the shallow fern-clad beds of the streams, not rare from the Sanchez plain to the Corres- pondencia ravine, 1000 — iioo m (no. 542); C. del Barril, 900 — 1000 m, not rare; on the ridge forming the south wall of Q. de las Vacas, not rare above 1000 m, Dickso?iia groves on the lower S. slope of Los Inocentes, etc. etc. The specimens from the two islands are fully identical. Area of distribution: Endemic. Lophosoria Presl 17. L. quadripiniiata (Gmel.) C. Chr. — Syn. Alsophila quadripinnata (Gmel.) C. Chr. Ind. 47 cum syn.; A. pruinata (Sw.) Klf. et auctt.; Hemsl. 6&^ Johow 1893: 20 f. 2, 1896: 154. One of the most common ferns, growing in widely different stations, and ranging from the ravines of the treeless basal region to the highest woods and heaths. Found with ripe sori in the autumn and winter. Masatierra: In all the quebradas between Pto Frances and Pto Ingles, common both in forested and in treeless tracts; numerous ravines near the colony, representing the last trace of indigenous vegetation; also on the elevated ridges separating V. Colonial from Villagra, 500—625 m (no. 590). Masafuera: Forming extensive beds in the subalpine and alpine region, ranging nearly to the highest summits (1350 — 1400 m), also on the highest parts of the great western precipice; not uncommon in the forests and along the streams of the canyons etc., e. g. in Q. de las Casas (no. 458). This interesting species which, as BowER has proved, is generically distinct from Alsophila, is widely spread through tropical and subtropical America, in several rather distinct forms. The insular form has been identified with the one from South Chile (Polypodium cinereum Cav.). Still it seems to differ from this in its dense cover of wool on the lower surface of the midrib, the Chilean form being decidedly less woolly. Further, the specimens from the two islands differ from each other in pubescence; if this character be constant, each island is inhabited by its special form. The form from Masafuera differs from other ones especially by the very dense rusty brown wool on the midribs of the THE PTI':R1D0P1I\ TA OT Till: JUAN I'l; RNANDi;/. ISLANDS \ -j secondary and tertiary pinnules. Besides the woolly hairs, the form from Masa- tierra shows numerous rather ri