Sonen gtatces eee evans es Oey dete | Ne bel ree Sy EM wth EOS Cer rer. oA rs - “4 ; pate tat te botieg 5 padeietnte nuns Ye ety : : ete a ea eae , Soa ~ % T ne =e Vehel. wv - = 0 “ re pane ? Ms nM z . c Soe te . «i : crits Recent re bp tetelytleienen dignema es ye rs et tera Pebeatey A ribs pe ~ Aan de athe Rete Ane . y Wivirtaheteen ks *. “ Recut 4 ~ os Ppt ne Pa De: oD Hy An he Ne _ ile is 4 3 ay = ety ae . %, Ay UAT. THE NATURAL HISTORY - OF JUAN FERNANDEZ - AND EASTER ISLAND E: EDITED BY DR. CARL SKOTTSBERG £ : VOL. Il ee BOTANY 2 : PART IV #3 =f 13. A. ZAHLBRUCKNER: Die Flechten der Osterinsel, nebst einem Nach- 3 = trag zu der Flechtenflora von Juan Fernandez. 14. JOHS. BOYE PETERSEN: Marine Cyanophyceae from Easter Island. 15. L. ROMELL: Basidiomycetes from Juan Fernandez. 16. KARL KEISSLER: Ascomyceten, Fungi imperfecti und Urediacen von Juan Fernandez. * 17. CARL SKOTTSBERG: The Vegetation of Easter Island. 3 18. CARL SKOTTSBERG: Pollinationsbiologie und Samenverbreitung auf EE den Juan Fernandez-Inseln. = 19. K. KEISSLER: Nachtrag zur Pilzflora von frist Fernandez. UPPSALA 1928 ALMQVIST & WIKSELLS BOKTRYCKERI-A. a ee 13. Die Flechten der Osterinsel, nebst einem Nachtrag zu der Flechtenflora von Juan Fernandez. Von Dr. A. ZAHLBRUCKNER. Unsere Kentnisse iiber die Flechtenflora der Osteninseln beschranken sich beinahe auf eine kurze, man darf sagen belanglose Notiz von F. H. KNOWL- TON, Botanic. Gazette, vol XII (1888), p. 94—95. Im Jahre 1885 wurde seitens der Vereinigten Staaten Nordamerikas der Dampfer» Mohican» nach den Oster- inseln entsendet zur Aufbringung der bekannten grossen Steinidole fiir das U. S. National Museum. Die Expeditionsteilnehmer haben wahrend ihres dortigen Aufenthaltes ihre Aufmerksamkeit den Flechten zugewendet und drei Arten derselben mitgebracht, welche dann von H. WILLEY bestimmt wurden. Als Resultate dieser Bestimmung wurden genannt: 1. Usnea barbata (L.); 2. Eine sterile Parmela aus der Verwandtschalt der Parmelia laevigata; 3. Physcia siellaris. ’ Ad 1. Was WILLEY unter Usnea barbata (L.) gemeint hat, ist schwer zu sagen. Usnea barbata im Sinne der alteren Autoren ist ein Sammelname fiir sehr verschiedene Arten und umfasst aufrechte strauchige und hangende Spezies. Selbst dariiber, was der Lichen barbatus des LINNE ist, zu entscheiden, stosst auf Schwierigkeiten, da nach R. HOWE Jr.’s Untersuchungen das Origi- nalstiick LINNE’s identisch ist mit Uszea articulata, er aber diese nebst Lichen barbatus beschreibt, letzterer aber in seinem Herbar nicht vorhanden ist.’ Die Aufsammlungen SKOTTSBERG’s lassen vermuten, dass keine Uswea vom plicata- Typus gemeint sei, eher eine vom /orida-Typus und dann konnte WILLEY vielleicht Usnea Steinert vor sich gehabt haben. Immerhin bleibt dies nur Ver- mutung und die Frage kann endgiiltig erst nach Einsichtsnahme in das von der amerikanischen Expedition mitgebrachte Stiick gelost werden. Ad 2. Eine Parmelia aus der Verwandtschaft der Parmelia laevigata sah ich nicht, aber das Vorkommen einer solchen ist leicht moglich. Ad 3. Pkhyscia stellaris sah ich von den Osterinseln nicht, halte ihr Vor- kommen daselbst nicht fiir unwahrscheinlich. 1 Vergl. Howe, R. jr. in Bull. Torrey Bot. Club, vol. XLI, 1918, p. 876. 2Q—2610. The Nat. Hist. of Juan Fernandez and Easter Isl. Vol. II. 450 A. ZAHLBRUCKNER Ausserdem wird in der Arbeit von F. FUENTES, Resena botdnica sobre la isla de Pascua, Publ. del Inst. Centr. Meteor. y Geofis. de Chile, N:o 4, San- tiago 1913, eine Flechte, Usuea xanthopoga Ny. fir die Osterinsel angegeben. Ob tatsdchlich diese Art, welche mit der unten aufgefiihrten U. Stezverz nichts zu tun hat, dort vorkommt, oder ob die Bestimmung ganz falsch ist, lasst sich ohne Untersuchung des Materials nicht sagen. Das von C. SKOTTSBERG aufgesammelte Material umfasst 22 Flechtenarten. Es sind hauptsachlich auf vulkanischem Gestein lebende Flechten. Unter diesen ist mit den meisten Arten die Gattung Aze/lza vertreten, deren Arten gerne vulkanische Gesteine besiedeln, wie dies auch der Fall ist auf Juan Fernandez und den Hawaiischen Inseln. Auffallend ist die Armut an Strauch- und Blatt- flechten, die dem bewahrten Sammler gewiss nicht entgangen sind — und ebenso diejenige an Flechten mit blaugriinen Algenkomponenten, welche nur durch Heppia Guepini vertreten sind. In der letzteren Beziehung steht die Oster- insel in einem scharfen Gegensatz zu den ebenfalls vulkanischen Samoainseln, wo sie sich mit nahezu 11% an der Zusammensetzung der Flechtenvegetatiou beteiligen.* Arthoniaceae. Arthonia fuscescens Fee. Fér, Essai Cryptog. Ecorc. Offic. (1824), p. 56, tab. XIII, fig. 8 et Suppl. (1837), p. 41, tab. XI, fig. 14; Mt. Arc., in Mémoir. Soc. Phys. et Hist. Nat. Geneve, vol. XXIX, no. 8 (1887), p. 35; Wiey., Synops. Gen. Arthonia (1890), p. 13; A. ZAUEBR Catal. ich aUniv.,.vol, Tiapesac. Im Krater des Rano Kau, auf der Rinde von Sophora toromzro und Lrous- sonetia papyrifera. Diese Art unterscheidet sich von der habituell sehr ahnlichen Arthonza vubella (Fée) Nyl. durch das etwas braunlich gefarbte Lager und die wesent- lich kleineren Sporen. Graphidaceae. Opegrapha (sect. Ewopegrapha) paschalis A. Zahlbr. nov. spec. Thallus epilithicus, crustaceus, uniformis, tenuis, ochraceo-viridescens, made- factus viridis, KHO —, Ca Cl, O,—, opacus, granuloso-inaequalis, continuus vel passim minute areolatus, areolis planis, fissuris tenuissimis separatis, sorediis et isidiis non praeditus, in margine linea obscuriore non cinctus, madefactus non odorus, fere homoeomericus, gonidiis copiosis, viridibus, globosis vel sub- globosis, 12—15 », latis, congestis, membrana tenui, sed distincta cinctis; hyphae medullares non amylaceae. Apothecia sessilia, nigra, nuda, linearia, simplicia vel 2—3 furcata, ramis stellatim dispositis, in toto usque 2 mm longis et ad 0,2 mm latis, ad apicem rotundatis vel acutatis; discus rimaeformis, angustissimus; excipulum fuligineum, 1 Vergl. A. ZAHLBRUCKNER: Die Flechten der Samoa-Inseln (Denkschr. math.naturw. Kl. Akad. Wiss. Wien, Bd. LXXXI, 1907, p. 2—3. DIE FLECHTEN DER OSTERINSEL 451 integrum, inferne incrassatum, labiis erectis, superne acutatis et inflexis; hyme- nium superne sordidulum, caeterum decolor, pulverulento-inspersum, I cupreum; paraphyses filiformes, dense ramosae, eseptatae, ad apicem non latiores, parum distinctae; asci ellipsoidei, ad basin breviter caudati, ad apicem rotundati et ibidem membrana modice incrassata cincti, recti vel leviter curvati, facile liberi, 75—85y. longi et 16—18y, lati, 8-spori; sporae in ascis verticales, triseriales, decolores, dactyloideo-fusiformes, ad apices angustato-rotundatae, rectae vel subrectae 8 —10—(12) loculares, loculis disciformibus, septis tenuibus, membrana leviter incrassata cinctae, 35—40y. longae et 7,5—9 py. latae. Im Krater des Rano Kau, auf vulkanischem Gerdll. Ich bin mir dariitber nicht sicher, ob die Gonidien dem Trentepohliatypus angehoren; die helle Farbe und die + kugeligen, nicht zu Ketten aneinander gefiigten Zellen sprechen dagegen. Allerdings ist zu beachten, dass die Trente- pohliagonidien im dicht gefiigtem Lager oft ausblassen, was ich bei tropischen Graphideen mit dickerem Lager zu sehen oft Gelegenheit hatte und dass die Zellen durch sich dazwischen drangende Hyphen oft aus dem Verbande gerissen werden. Es lag nur ein Stiickchen der Flechte vor und somit konnte ich den Sachverhalt nicht eingehender untersuchen. Eine Opegrapha mit Palmellaceen- gonidien wurde bisher meines Wissens nicht beobachtet. Sollten nun dennoch die Gonidien nicht zu Zventepohiia gehoren, dann wirft sich die Frage ab, ob die Apothezien nicht auf dem Lager einer fremden Flechte aufsitzen und der Organismus als Flechtenparasit zu betrachten sei oder ob etwa analog der Gattung A//arthonia der Vertreter einer neuen Gattung (oder Untergattung) vorlage. Graphis lineola Ach. Acu., Lichenogr. Univ. (1810), p. 264; A. Zanipr., Catal. Lich. Univ., vol. Ms ps} sir 6. Im Krater des Rano Kau, auf der Rinde von Sophora toromiro und Brous- sonetia papyrifera. Ein tropisches, beziehungsweise subtropisches Element der Flechtenflora der Osterinsel. Diploschistaceae. Diploschistes anactinus A. Zahlbr. A. ZAuHLBR. in Hedwigia, vol. XXXI (1892), p. 14 et Catal. Lich. Univ., vol. II, p. 657. — Urceolaria anactina Nyl., Lich. Japon. (1890), p. 39. Im Krater des Rano Kau, im Gerdlle; Strandfelsen von Mataveri; Lava- ebene in Hanga Ho Orno (La Pérouse Bay), auf Lavablocken. Diese Flechte ist bisher nur aus Japan angegeben. Sie ist aber wahrschein- lich auf vulkanischen Inseln weit verbreitet. Diploschistes scruposus (L.) Norm. f. argillosus Dalla Torre et Sarnth. Datta, Torre et Sarntu., Die Flecht. Tirol (1892), p. 298; A. ZAHLBR., Catal. Lich. Univ., vol. ll, p. 669. — Urceolaria scruposa f. argillosa Ach., Method. Lich. (1803), p. 148. 452 A. ZAHLBRUCKNER Cortex Ca Cl,O, bene rubescit, medulla autem non coloratur; medulla I lutescens vel dilute coerulescens. Aussenseite des Rano Kau, auf dem Erdboden. Heppiaceae. Heppia Guepini Ny]. Nyt. apud Hue in Revue de Botan., vol. V, (1886), p. 18; A. Zautpr., Catal. Lich. Univ., vol. Ill, p. 197. — Endocarpon Guepint Del. apud Dusy, Botanic. Gallic., vol. II (1840), p. 594. Hanga Ho Orno (La Pérouse Bay), auf Lavablocken in der Ebene. Stimmt mit der europaischen Pflanze vollkommen iberein. Lecideaceae. Lecidea (sect. Au/ecidea) paschalis A. Zahlbr. nov. spec. Thallus epilithicus, crustaceus, uniformis, expansus, substrato adnatus, subtartareus, tenuis, pallide ochraceus vel ochraceo-albescens, opacus, KHO stramineo-flaveus, Ca Cl,O, flavens, KHO + Ca Cl,O, flavus, areolato-diffractus vel areolatus, areolis planis, parvis, usque I mm latis, subangulosis, fissuris valde tenuibus separatis, sorediis et isidiis non praeditus, in margine sat bene limitatus et linea tenui sordidescente vel obscurata cinctus; gonidia cystococ- coidea, globosa, 7—9u. lata; hyphae medullares non amylaceae. Apothecia immersa, nigra vel fusco-nigra, opaca, minuta, ad 0,5 mm lata, leviter convexa, passim crebra, margine conspicuo non praedita; excipulum dimidiatum, tenue, sordidescens, molle, gonidia non includens, non paraplecten- chymaticum; hymenium superne anguste nigrescens, caeterum decolor et purum, 60—70y. altum, I coeruleum; hypothecium pallide lutescens, molle, ex hyphis conglutinatis, verticalibus, in hymenium transeuntibus formatum; paraphyses filiformes, conglutinatae, simplices, eseptatae, ad apicem nigro-capitatae; asci hymenio paulum breviores, clavati, superne late rotundati, 8-spori; sporae in ascis subuni- vel biseriales, decolores, simplices, late ovales vel late ellipsoideae, membrana tenui cinctae, demum pallide fuscescentes, 8 —10y. longae et 4— 5. latae. Pycnoconidia non visa. Strandfelsen von Mataveri. Das Hymenium wird haufig von dunklen Torula-Hyphen (wohl einem para- sitischen Pilz angehorig) durchlaufen. Cladoniaceae. Cladonia pityrea (Flk.) Fr. var. sorediosa Wain. Warnio, Clad., vol. II (1894), p. 355. DIE FLECHTEN DER OSTERINSEL 453 Kratersee von Rano Kau, c. 110 m, auf Humusboden zwischen Moosen. Thallus teils aus schmalen Bechern teils aus zugespitzten Podetien, die an der Spitze 2—3 Aste tragen, zusammengesetzt. Acarosporaceae, Acarospora (sect. Ewacarospora) Skottsbergii A. Zahlbr. nov. spec. Thallus epilithicus, sat late effusus, e maculis minutis, 0,1:—0,2 mm latis, quasi minute flocculosis, rotundatis vel irregularibus, crebris, sed non approxi- matis, plus minus dispersis, fere planis, utplurimum leviter convexis formatus, albus, opacus, KHO stramineo-flaveus, Ca Cl,O0,—, KHO + Ca Cl, O, —, sorediis et isidiis destitutus, in margine ab hypothallo cinereo-nigricante circumdatus. Apothecia rara, dispersa vel approximata, usque 0,8 mm lata, inter ma- culas thallinas ad hypothallum sedentia, umbrino-fusca, opaca, convexa, ad basin leviter constricta; margo disco concolor, madefactus paulum obscurior, depressus, in sectione transversali extus rufo-fuscus, intus decolor et gonidia cysticoccoidea includens, non paraplectenchymaticus; hypothecium pallidum; hymenium superne non obscuratum, decolor et purum, ad 150. altum, I coeruleum; paraphyses capillares, simplices, eseptatae, ad apicem haud latiores, leviter conglutinatae; asci hymenio subaequilongi, obrapiformes, ad apicem rotundati et membrana ibidem modice incrassata cincti, myriospori; sporae decolores, globulosae, mem brana tenuissima cinctae, exiguae, ad I u. latae. Pycnoconidia non visa. Gipfel des Vaintu Rova, auf vulkanischem Tuff. Die Art ist durch die Bildung des Thallus und durch die auf dem Hypo- thallus sitzenden Apothezien sehr auffallig und namentlich in letzterem Belange von den tibrigen Arten der Gattung sehr verschieden. Parmeliaceae. Parmelia reticulata Tay]. Tayi. apud Macx., Flora Hibern., vol. II, (1836), p. 148; Du Rietz in Bot. Notiser (1924), p. 333. — Parmelia cetrata Ach. var. sorediifera_Wain., Etud. Lich. Brésil, vol. I (1890), p. 40. Vaintu Rova, c. 300 m und im Krater des Rano Kau, c. 120 m, in beiden Fallen auf Gestein lebend und steril. Parmelia conspersa (Ehrh.) Ach. var. lusitana ((Nyl.) Lett. Lett. in Hedwigia, vol. LV (1914), p. 44; STuR. in Oest. Bot. Zeitschr., vol. LXVI (1916), p. 277. — Parmelia lusitana Ny\. in Flora, vol. LXIV (1881), p. 449 et Lich. Pyren. Orient. (1891), p. 5. Medulla KHO lutescens, thallus isidiis increbris obsitus. Hanga Ho Orno (La Pérouse Bay), auf Strandfelsen einige Meter uber dem Meere und bei Mataveri, auf Felsen, stets steril. 454 A. ZAHLBRUCKNER Stimmt mit der siideuropaischen Flechte vollig iberein, wenn man davon absieht, dass der Thallus etwas lebhafter gelb gefarbt ist. Parmelia caperata (L.) Ach. Hanga Ho Orno (La Pérouse Bay), auf vulkanischen Blécken etwas ent- fernt vom Strande, in sterilen Stiicken. Auch diese Flechte zeigt eine lebhaftere Farbe als die in Mittel- und Sud- europa lebenden Individuen. Ganz ebenso gefarbt fand ich diese Art in der Aufsammlung J. F. ROCK’s auf den Hawaiischen Inseln, wo sie ebenfalls auf vul- kanischem Gestein vorkommt und sehr haufig auftritt. Usneaceae. Usnea Steineri A. Zahlbr. var. tincta A. Zahlbr. A. ZaAuiLBr. in Denkschrift d. math.-naturh. Klasse der K. Akad. Wiss. Wien, vol. LX XXIII (1909), p. 183; Howe jr. in Bryologist, vol. XVIII (1915), p. 62. Krater des Rano Kau, c. 120 m, auf Gerdll und auf dem Gipfel des Katiki, c. 412 m; in beiden Fallen steril. Die Usnea Steinert A. Zahlbr. und ihre Varietaten gehoren den tropischen und subtropischen Regionen an und vertreten dort die nordlichere Usnea flo- vida (L.). Caloplacaceae. Caloplaca rubina A. Zahlbr. A. ZAHLBR. apud Sxkottsse., The Natur. Hist. Juan Fernandez and East. Isl., vol. II (1924), p. 306. Hanga Ho Orno (La Pérouse Bay), auf Steinen einige Meter vom Meer. Caloplaca (sect. Gasfarrinia) lucens A. Zahlbr. A. ZAHLBR. in Deutsch. Siidpolarexped. 1901—1903, vol. VIII (1906), p. 29. — Lecanora elegans (Link.) Th. Fr. f. ducens Nyl. apud Crompie in Journ. Linn. Soc. London, Bot., vol. XV (1876), p. 184. — Placodium lucens Ny\., Lich. Nov. Zeland. (1888), p. 145. Auf den Strandfelsen bei Mataveri. Buelliaceae. Buellia stellulata Mudd. f. protothallina Wainio. Warnio, Etud. Lich. Brésil, vol. I (1890), p. 174. — Lecidea stellata ‘Yay. f. protothallina Krph. in Flora, vol. LIX (1876), p. 267. Gipfel des Vaintu Rova; im Krater des Rano Raraku, auf vulkanischem Gestein; Strandfelsen von Mataveri; Hanga Ho Orno (La Pérouse Bay), Lavaebene auf Bloécken; im Krater der Rano Kau, auf Gerdll. — Diese Flechte ist auf dem vulkanischen Gestein der Osterinsel sehr haufig, ich fand sie auf zahlreichen Stiicken der Aufsammlung. DIE FLECHTEN DER OSTERINSEL 4 un on Buellia fernandeziana A. Zahlbr. A. ZAHLBR. in Kgl. Svensk. Vetensk.-Akad. Handb., vol. LVII no. 6, (1917), P- 53: Thallus epilithicus, crustaceus, uniformis, tartareus, subisabellino-albidus, opacus, KHO e flavo sanguineus (crystallos aciculares, longiusculos, substellatim- aggregatos effundens), Ca Cl,O,—, KHO + Ca Cl,O, non mutatas, irregulariter expansus, demum (imprimis centrum versus) crassiusculus, usque I mm altus, in superficie inaequalis, rimoso-areolatus, areolis angulosis, 0.3—1I,5 mm latis, utplurimum leviter convexis, rare subplanis, rimis tenuibus separatis, areolis marginalibus hinc inde subverruculosis et ultimis primum separatis vel dispersis, demum confluentibus, in margine haud bene limitatus et linea obscuriore non cinctus, sorediis et isidiis destitutus; stratum corticale angustum, 30—40 m altum, ex hyphis dense intricatis et inspersis formatum; stratum gonidiale infra stratum corticale situm, subcontinuum, tenue, gonidiis cystococcoideis, globosis, laete viridibus, 8—12¥. latis; medulla crassiuscula, alba, KHO e flavo sanguinea, Ca Cl,O, non tincta, ex hyphis non amylaceis, dense contextis formata. Apothecia lecideina, rotunda, dispersa vel approximata, parva, 0,4—0,8 mm in diam., primum arcte adpressa, demum sessilia vel subsessilia, convexa, nigra, plus minus lutoso-pulverulenta, demum subnuda, opaca; margo proprius primum tenuissimus, vix prominulus mox depressus; excipulum dimidiatum, _fusco-nigrum, inferne cum hypothecio crasso, fusco-nigro, KHO —, NO; —, con- fluens; hypothecium versus hymenium dilute olivaceo-fuscum et ibidem non bene limitatum, inferne pedatim productum; hymenium superne sordide fuscum, pulverulentum, KHO solutionem dilute olivaceo-lutescentem effundens, caeterum decolor, purum, I intense violaceo-coeruleum; paraphyses filiformes, conglutinatae, sed distinctae, simplices, eseptatae, ad apicem modice latiores, subclavatae; asci hymenio subaequilongi, clavati vel oblongo-clavati, ad apices rotundati et ibidem membrana primum bene incrassata cincti, 8-spori; sporae in ascis sub- uniseriales, rare biseriales, fuscae, obtusato-ellipsoideae, utrinque Jate rotundatae, uniseptatae, ad septa non constrictae, membrana tenui, septo non crassiore, 8,5—12u. longae et 6—7 wu. latae. Pycnoconidia non visa. Strandfelsen von Mataveri und von Hanga Piko. Die urspriingliche Beschreibung konnte auf Grund des _ vorliegenden reicheren Materials erganzt und erweitert werden. Insbesondere trat die Be- reifung der Apothezien deutlich zutage, wahrend sie an den Stiicken von Juan Fernandez weniger bemerkbar war. Buellia (Euduellia) halophiloides A. Zahlbr. nov. spec. (vide infra, p. 459). var. pruinosula A. Zahlbr. nov. var. A typo differt apotheciis leviter caesio-pruinosis. Strandfelsen von Mataveri, haufig. 456 A. ZAHLBRUCKNER Buellia (Euéuellia) paschalis A. Zahlbr. nov. spec. Thallus epilithicus, tartareus, plagas irregulares et confluentes formans, tenuis, ochraceo-albidus, opacus, KHO e flavo sanguineus, Ca Cl,O, —, KHO + Ca Cl,O, in roseum vergens, rimoso-areolatus, areolis versus ambitum thalli dispersis, subrotundatis, congestis, protothallo tenui cinereo-nigricanti insidentibus, minutis, 0,z—0,3 mm latis, centrum thalli versus confluentibus, paulum magis turgescentibus, convexis vel verruculiformibus, subirregularibus, isidiis et sore- diis non praeditus. Apothecia lecideina, rotunda, parva, usque 0,8 mm lata, dispersa, sessilia, nigra, Opaca, mox convexa, margine primum valde tenui et vix prominulo, demum depresso cincta; excipulum fuligineum, dimidiatum; hypothecium crassum, obscure rufo-fuscum, versus hymenium parum dilutius; hymenium superne anguste et bene limitate fuscum, caeterum decolor, purum, 60—62¥, altum, I intense coeruleum; paraphyses filiformes, strictae, simplices, eseptatae, conglu- tinatae, ad apicem modice clavatae; asci clavati, superne rotundati et ibidem primum membrana modice incrassata cincti, 8-spori; sporae in ascis subuni- seriales, fuscae, breves, late ellipsoideae, utrinque bene rotundatae, uniseptatae, ad septa non constrictae, septo et membrana tenui, 8—gy. longae et 5—6u. latae. Pycnoconidia non visa. Strandfelsen von Mataveri. Buellia Glaziouana Mill. Arg. Mixy. Arc.! in Flora, vol. LXIII (1880), p. 19; Warn., Etud. Lich. Brésil, vol. I (1890), p. 169. — Lecédea Glaziouana Krph.! in Flora, vol. LIX (1876). Dacsige Thallus crustaceus, epilithicus, substrato arcte adnatus, tenuis, 0,:—0,6 mm altus, uniformis, rosulas irregulares, minores vel majores, in margine linea valde tenui et obscurata cinctas, confluentes formans, demum sat late expansus, laevis, Stramineus, nitidulus, KHO vix paulum magis flavens, Ca Cl,O,—, KHO4+ Ca Cl,O,—, minute areolatus vel areolato-rimosus, areolis minus angulosis, O,r-0,6 mm latis, planis, rimis pertenuibus limitatis, sorediis et isidiis non praeditus; medulla tenuis, alba, ex hyphis amyloideis formata. Apothecia lecideina, dispersa, primum fere innata, demum adpressa vel adpresso-sessilia, ad basin leviter constricta, minuta, usque 0,6 mm lata, rotunda, e concaviusculo plana vel planiuscula, nigra, opaca, madefacta parum in fusce- scentem vergentia; margo tenuis, parum prominulus, integer, primum thallo fere concolor, demum obscuratus, in sectione extus anguste obscuratus, intus palli- dus, ex hyphis subradiantibus, sat crassis, leptodermaticis formatus, gonidia nulla includens, excipulum fuscum, ad latera hymenii angustum, superne inflexum, infra hymenium crassius, plus minus lentiforme; hymenium superne fuscum, NO, —-, non inspersum, caeterum decolor, purum, 60—8o0y,. altum, I e coeruleo demum obscuratum; paraphyses filiformes, strictae, ad 1,5, crassae, simplices, eseptatae, ad apicem septato-clavatae et fuscae, haud arcte conglutinatae; asci hymenio subaequilongi, oblongo-clavati, ad apicem rotundati et membrana in- crassata cincti, 8-spori; sporae in ascis bi- vel triseriales, fuscae, ovales vel DIE FLECHTEN DER OSTERINSEL 457 late ellipsoideae, utrinque late rotundatae, rectae, rare leviter curvulae, uni- septatae, ad septa non constrictae, septo et membrana tenui, 9—14 yp. longae et 7—8 yp. latae. Conceptacula pycnoconidiorum in tuberculis thallinis minutis immersa, vertice minuto, convexo et fusconigricante, globosa, perifulcrium (excepto ver- tice) pallidum; fulcra exobasidialia, basidiis anguste lageniformibus, pycnoconidiis paulum brevioribus; pycnoconidia bacillaria, recta, utrinque retusa, 7—8 p. longa et ad 1 w lata. Strandfelsen von Mataveri und Hanga Piko und im Krater des Rano Kau, auf dem unteren Gerdlle; Rano Raraku, auf harten Tuffstiicken; Hanga Ho Orno (La Pérouse Bay), auf der Ebene. Eine Farbenreaktion, hervorgerufen durch Chlorkalk nach Hinzuftigung von Kalilauge, konnte ich weder am Urstiicke, noch an den von MULLER ARG. bestimmten oder den Stiicken von der Osterinsel beobachten. Die Medullar- schicht des Lagers hingegen farbt sich mit Iod blau, unter dem Mikroskope sieht man, dass die ganze Markschicht nicht gleichmassig, sondern nur in ein- zelnen Partien gefarbt erscheint. Rinodina (J/ischeblastia) Pérousii A. Zahlbr. nov. spec. Thallus epilithicus, tenuis, subtartareus, glauco-albidus, opacus, KHO —, Ca ClhO,—, KHO + Ca Cl,O,—, squamulato-areolatus, areolis plus minus dispersis vel rarius subcontiguis, substrato adpressis, sat irregularibus, parvis, usque I mm latis, planis vel planiusculis, sorediis et isidiis destitutus, hypothallo distincto non insidens. Apothecia lecanorina, rotunda, parva, usque 0,8 mm lata, adpresso-sessilia; discus niger, fere opacus, e plano leviter convexus; margo thallinus tenuis, thallo concolor, integer, discum non superans; hypothecium decolor ex hyphis in- tricatis formatum; hymenium superne anguste rufofuscum, caeterum decolor, dilucidum, purum, 95—110y. altum, I intense coeruleum; paraphyses filiformes, strictae, bene limitatae, simplices, eseptatae, ad apicem clavatae et ibidem tenuissime et increbrae septatae, conglutinatae; asci ellipsoidei vel ovales, superne rotundati et membrana sat bene incrassata cincti, 8-spori; sporae in ascis bi- vel triseriales, primum fumosae, demum fuscae, ellipsoideae, utrinque rotundatae, rectae, polari-diblastae, loculis depresso-cordatis, circa '/4 longitudinis sporarum metientes, ad septum non constrictae, 17—19u. longae et 9—1I0u. latae. Hanga Ho Orno (La Pérouse Bay), am Strande, aber vom Ufer etwas entfernt, auf vulkanischem Gestein. Diese neue Art, die ich nur auf einem Stiick der Aufsammlung zwischen Caloplaca rubina fand, gehért dem Formenkreis der Rzxodina discolor (Hepp) an, von welcher sie sich aber durch eine andere Ausbildung des Thallus, durch die deutlich berandeten Apothezien schon 4Ausserlich leicht unterscheiden asst; ausserdem sind die Sporen kleiner. Physciaceae. Pyxine enteroxantha Ny]. Nvyt.! in Flora, vol. LVI (1873), p. 196, Lich. Japon. (1890), p. 35 et Lich. Pyren. Orient. (1891), p. 74. 458 A. ZAHLBRUCKNER f. subpruinosa A. Zahlbr. nov. f. Lobi marginales thalli tantum insulatim pruinosi, thallus caeterum nudus. Hanga Ho Orno (La Pérouse Bay), auf Strandfelsen einige Meter iiber dem Wasser, in sterilen Stiicken gesammelt. Physcia picta (Sw.) Nyl. Lavaebene in Hanga Ho Orno (La Pérouse Bay), Strandfelsen von Hanga Piko und Mataverl, auf vulkanischen Blocken, Rano Raraku, auf Tuff, stets steril. Scheint eine der haufigsten Flechten der dunklen Lavablocke zu sein. Auch in der von ROCK und FAURIE auf den Hawaiischen Inseln ge- machten Aufsammlungen fand ich die Art haufig, mit Vorliebe auf den dunklen, einer grossen Erwarmung ausgesetzten vulkanischen Blocken. Anaptychia speciosa (Wulf.) Wainio f. sorediifera A. Zahlbr. nov. comb. Physcia speciosa f. sorediifera Mill. Arg. in Flora, vol. LXVII (1884), p. 688. Im Krater des Rano Kau, c. 125 m, itiber Moosen in kiimmerlichen Exemplaren; Vaintu Rova, c. 300 m, auf dem Erdboden. Nachtrag zur Flechtenflora von Juan Fernandez. Buellia (sect. Avéuec/lia) masafuerana A. Zahlbr. nov. spec. Thallus epilithicus, crustaceus, uniformis, tenuis, tartareus, effusus, isabel- lino-albidus, opacus, KHO flavens, Ca Cl,O, et KHO+ Ca Cl,O, non mu- tatus, areolatus, areolis minutis, ad 0,2 mm latis, angulosis, pianis, fissuris tenuissimis separatis, sorediis et isidiis destitutus, in margine linea nigra non cinctus. ° Apothecia lecideina, sat crebra, dispersa vel approximata, rotundata, usque I mm lata, nigra, opaca et demum nitidula, sessilia, primum plana, margine tenui, vix prominulo, mox convexa margine depresso; excipulum fuligineum, integrum; hypothecium crassum, obscure fuscum; hymenium superne anguste nigricans, non inspersum, caeterum decolor et purum, usque 60¥y. altum, I vio- laceo-coeruleum; paraphyses filiformes, bene limitatae, simplices, eseptatae, ad apicem capitatae, modice conglutinatae; asci anguste clavati, superne rotundati et membrana paulum incrassata cincti, 8 spori; sporae in ascis subbiseriales, fuscae, ellipsoideae vel + ovales, utrinque rotundatae, rectae, uniseptatae, ad septum non constrictae, septo et membrana tenui, 12—15 u. longae et 6—7 y, latae. Pycnoconidia non visa. Masafuera: Block im unteren Teil des Vacas-Tals, auf vulkanischem Gestein. Der Buellia halophila Mill. Arg. zunachst stehend, durch das dunkle Hypo- thezium und die gewolbten Apothezien verschieden. DIE FLECHTEN DER OSTERINSEL 459 Buellia (sect. “uéue//za) barrilensis A. Zahlbr. nov. spec. Thallus crustaceus, uniformis, expansus, tenuis, tartareus, pallide isabellinus vel straminescens, opacus, KHO parum magis flavens; Ca Cl,O, —, KHO + Ca Cl,O, —, verruculosus, verruculis confertis, rotundatis vel subirregularibus, laevigatis, sorediis et isidiis non praeditus, in margine linea nigra angustaque cinctus; medulla alba, hyphis non amyloideis. Apothecia sat crebra, dispersa vel approximata, nigra, opaca, parva, 0,4— 0.7 mm lata, rotunda, mox convexa, margine depresso; discus non pruinosus; excipulum fusconigrum, dimidiatum, tantum ad latera hymenii evolutum; hypo- thecium fuscescens, molle, ex hyphis intricatis formatum; hymenium superne sat latiuscule et sat bene determinate nigricans, granuloso-pulverulentum, caeterum decolor et purum, ad 80». altum, I violaceo-coeruleum, demum sordide obscu- ratum; paraphyses capillari-filiformes, simplices, eseptatae, ad apicem capitato- clavatae, conglutinatae; asci ellipsoideo clavati, superne rotundati et membrana incrassata cincti, recti, hymenio subaequilongi; sporae in ascis bi- vel triseriales, fuscidulae, ellipsoideae vel ovali-ellipsoideae, utrinque bene rotundatae, rectae, uniseptatae, septo et membrana tenui, ad septum non constrictae, I14—I7 uv. longae et 6—8 », late. Pycnoconidia non visa. Masafuera: Cordon de Barril, 1,000 —1,300 m, auf vulkanischem Gestein. In der Artenreihe mit gelbem Lager kommt die als neu angesprochene Art der Suellia Glaziouana zunachst und lasst sich von dieser durch das warzige k6ornige Lager leicht unterscheiden; von Awellza concinna var. oceanica trennt sie die fehlende Chlorkalkreaktion des Thallus und das helle Hypo- thezium. Buellia (sect. Auéuel/za) halophiloides A. Zahlbr. nov. spec. Thallus epilithicus, tenuis, substrato arcte adpressus, rosulas rotundatus, vel irregulares, ad 1,5 cm lates, confluentes formans, tenuis, tartareus, isabellino albidus, opacus, KHO — vel dilute flavescens, Ca Cl,O, —, KHO + Ca Cl,0, —, tenuissime rimoso-areolatus, areolis parvis, usque 0,5 mm latis, demum leviter convexis, in margine protothallo distincta non circumdatus; medulla alba, ex hyphis non amylaceis formata. Apothecia crebra, dispersa vel approximata, lecideina, primum adpressa, demum sessilia, nigra, opaca (detrita nitidula), primum plana et margine proprio angusto, vix perspicuo cincta, demum convexa, margine depresso, usque 0,8 mm lata; excipulum dimidiatum, fuligineum; hypothecium olivaceo-fuscum, molle; hymenium superne anguste nigrescens, non inspersum, caeterum decolor et purum, 60—65. altum, I violaceo-coeruleum; paraphyses filiformes, eseptatae, ad apicem clavato-capitatae, conglutinatae; asci ellipsoideo- vel subovali-clavati, superne rotundati et membrana incrassata cincti, 8-spori; sporae in ascis biseriales, fuscae, ellipsoideae vel subovales, rectae, utrinque sat late rotundatae, uniseptatae, septo tenui, ad septa non vel nonnihil levissime constrictae, mem- brana tenui cinctae, g—1I1 wu longae et 4—} u. latae. Pycnoconidia non visa. 460 A. ZAHLBRUCKNER Masatierra: Bahia del Padre, auf Strandfelsen. . Eine Varietat derselben Art kommt auf der Osterinsel vor, s. oben S. 455. A Buella halophila Mill. Arg., cui proxima, differt imprimis areolis thalli non planis, apotheciis demum convexis. Species affines: I. Areolae thalli convexae, apothecia convexa BL. halophiloides. Il. Areolae thalli planae. 1. Apothecia plana; hypothecium pallidum. L. halophita. 2. Apothecia convexa; hypothecium obscurun. L. masafuerana. Ferner sei bemerkt, dass die Lokalangabe ftir Ca/oplaca clandestina A. Zahlbr., Die Flechten der Juan Fernandez-Inseln, S. 395, falsch angegeben wurde. »Masatierra» ist zu ersetzen durch: Masafuera: Quebrada del Blindado, auf Rinde von Dendroseris gigantea. Printed **]2 1926. 14. Marine Cyanophyceae from Easter Island. By JOHS; BOYE “PETERSEN, In some of the samples of marine algae from Easter Island investigated by Dr. F. BORGESEN!' Cyanophyceae occurred. Dr. BORGESEN asked me to undertake the investigation of this group, and I accepted his request with pleasure. As no Cyanophyceae have hitherto been recorded from Easter Is- land, I presume that even the most sketchy list may be of interest, and I have therefore attempted to determine the often incompletely developed spe- cimens found in the samples. Only in 3 samples preserved in formalin and in one dried sample I found determinable Cyanophyceae. All the samples were collected at Hanga Piko in the litoral zone (June 1917). Fam. Oscillatoriaceae. Lyngbya C. Ag. L. sordida (Zanard.) Gom. Gomont, Monogr. des Oscillariées II, Ann. Sc. nat. Waser I .116,, D5 alee, pl. Il, fig. 21. This species occurred in the dried sample, growing in Zonaria variegata. Area of distribution: Atlantic coast of Europe, Mediterranean Sea, West Indies, Friendly Islands (Pacific Ocean). Hydrocoleum Kiitz. H. comoides (Harv.) Gom. Gomonn, Monoproil} Ipne., ‘Dieasyp335) plox lh figiigi—s: Among Hypnuea Espert | found a form with trichomes measuring 18—20 p. in diameter. The apex was attenuated and often provided with a flat calyptra. The trichomes were thus exactly in accordance with those of //. comozdes, and ' The natural history of Juan Fernandez and Easter Island edited by Dr. Carl Skottsberg. fel. Il, p:. 24°77. 462 JOHS. BOYE PETERSEN the sheaths were comparatively distinct; but I have not been able to detect in the deficient material any sheath containing more than one trichome. Nevertheless I feel convinced that the form observed is identical with the named species. Area of distribution: The shores of the Bermuda Islands, Guadeloupe and Australia. Fam. Scytonemaceae. Microchaete Thuret. M. vitiensis Askenasy. Borner et Franautt, Revision III, Ann. sc. nat. 7’ sér. T. V, 1887, p. 85. The plant formed a red layer (c. 5 mm. in diameter) in Zonarza varie- gata. The filaments were 8—g y, in diam., the trichomes 6—7. The cells are quadrate or slightly shorter than the breadth. The heterocysts are only basal. Area of distribution: Pacific Ocean. Fam. Rivulariaceae. Calothrix Ag. C. crustacea Thuret. BorneET et THureEtT, Notes algologiques I, p. 13, pl. 4; BorNnEet et FLAHAULT, meyision Fable. i di p--3s50; I found this species among Sfhacelaria sp. in a sample chiefly containing Hypnea Esperi, and the Calothrix must undoubtedly have grown on the stone, being removed from this together with the larger algae. The diameter of the threads was 17,6 —23, of the trichomes 11—17y.. Numerous heterocysts were present, basal as well as intercalary. As the apices of the trichomes unfortunately were wanting in all the observed specimens, a confusion with C. rectangularis S. et G. (SETCHELL and GARDNER, The marine algae etc. I, Univ. Calif. Publ. Vol. 8, No. 1, p. 100, pl. 6, fig. 21) was not excluded, the distinguishing character of the two species being exactly the length of the apical hairs. However, as the observed Ca/othrix was provided with numerous heterocysts of different form, these being sometimes cylindrical, sometimes depressed, almost disc-shaped, and as the latter form evidently does not occur in C. rvectangulavts judging from the description and the figures in SETCHELL and GARDNER |. c., I am of the opinion that the found specimens ought to be identified with C. crustacea. In one respect only they apparently differred from this species: the threads were not swollen at the base but on the con- trary slightly attenuated towards the base. Area of distribution: Atlantic Ocean, Mediterranean Sea, Pacific Ocean. MARINE CYANOPHYCEAE FROM EASTER ISLAND 462 Brachytrichia Zanard. B. maculans Gom. Gomont in Schmidt: Flora of Koh Chang IV, Bot. Tidsskr. Vol. 24, p. 127, pl. V, fig. 5—7. — Corzins and Hervey, The algae of Bermuda. Proc. Am. Ac. mets and Sc. Vol. 53, No. 1, 1917, p..28. The specimens found formed a flat, compact thallus, attached to a stone. The thallus was 0,24 mm. thick and about 2 cm. in diameter. It consisted of numerous erect, parallel double-threads. Every sheath contains one trichome that in the majority of cases is bent in the middle, having the two parts parallel and each of them ending in a hair. The species has formerly been found growing between tide-marks, and it is also found at Easter Island under similar conditions. Area of distribution: Koh Chang (Gulf of Siam), Bermuda. Conse- quently this species is now recorded from the three Oceans. Printed '*/2 1926. t . ; 7 ee #5, HAs: AUS Ae ARMAS AyIaaM~ i oe ayers ae a net Beet able eae a Mie tebe Jes ‘it gions more’ thes one: 4e gos brx,- nc th- aay io¢errTed is ide sani a : “mon anal ve AE ee tat Tealebel": sath | ¥T bgnaiAebe tes 36(balie ae wrbt mn da Bide, MP Ghat shut te sagia ott vavaalt hoe 2ut1s0D — 5 Seq .gtor .t of 2 ov ae sq0ta-8 of trathowtin acellactt foeqnios 22 ¢ Bespin bao) eosniooeael ie basrienos 3 ravetenib qintg> Qcinade-bracdaid? ant sed <"—e sisaton 2No entakiog dtasd2 “tev. i abnaiizaldyeb tsifeiag abate Owl stp Gobred Sibiige eebbeesh ated “ez overs lo aes asad Ying) zalt esivoqe odT sled wai gothes made le doses amo vetesl to edo uate 4) i} ine eobrecobit myseciod eno) Abutnryl (maim to Wad} S080.) Goa: rosso tisib 7 7 | Aue {CY avwik oft mien) babioge? ‘pom iat cared aL | eg aye) wey = a 2 - . = : al = in Ss =~ - 7 7 ‘ Ee Sr. = © ae ee es a @ Sarg ; ] =3-4— =F, etted-esi-. wie C2hiny i T- os ; wt & i< : i restangetor?: oe al eee Sd ee > ae : Ualv.. Lat. Fag Ves. fic. t, yen, : ~—- => Nectria spec. Masatierra: an Rinde von Myrceugenta fernandeziana, Quebrada Villagra, feck FO17. Steril, daher sichere und nahere Bestimmung nicht méglich. Hypocrea rufa Fr. Masatierra: an morschem Holz, Cordén del Centinela, 1. IV. 1917; Quebrada del Juanango, 9. IV. 1917. Nur schlechte Schlauche, Sporen nicht gesehen. 1 Cfr. Centralblatt fiir Bakteriologie. XXXIX., 1913, pag. 625. 480 KARL KEISSLER Dothideales. ?Phyllachora graminis Fuck. Masafuera: an Bblattern von Agvostis masafuerana Pilg., Las Torres, eG nls. is, 54> Nie rOn 7. Sphaertales. Corynelia tropica (Auersw. et Rbh.) Starb. Masafuera: Cordon del Barril, ca. 500 m. s. m., an Friichten von Drzmys Wintert Forst., 1. III. 1917. Perithecien aussen schwach warzig, Sporen fast rund, ca. 9X7 pw. Ur- spriinglich auf Blattern von Podocarpus salgna aus Chile angegeben (RABH., Fungi eur. nr. 1261), hier an Friichten von Drzmys, offenbar dieselbe Art. Trematosphaeria Robinsoniae nov. spec. Peritheciis ligno decorticato semi-immersis, globosis, dein applanatis, ma- jusculis (ca. O.;—1 mm diametro), atris laxe gregariis, interdum 2—3 confluen- tibus, papilla minuta ostiolatis, carbonaceis. Ascis cylindraceis, apice rotundatis, tenuiter tunicatis, mox collabentibus, 8-sporis, ?...longis, 11 w latis, paraphysibus filiformibus numercsis flexuosis ascis multo longioribus ca. 0.5 4 crassis circumdatis. Sporis late-ellipticis, oblique-monostichis, dilute-brunneis, 3-septatis (septis non constrictis), 4-guttu- latis, muco non obvallatis, ca. 12.5% 7—9 ». Hymenio Jodis ope primum coe- ruleo, demum dilute-viridi. Masatierra: ad lignum decorticatum Rodznsonzae gracilis DC., Portezuelo de Villagra, Goo,m: Ss. ni, 22,71) 19017. Didymellina Rhaphithamni nov. spec. Maculis albidis, subcircularibus, plerumque brunneo-cinctis, ca. 0.3—0.8 cm diametro. Peritheciis epiphyllis, gregariis, interdum + confluentibus (2 vel pluribus concrescentibus), stromate destitutis, globosis, atris, poro pertusis, primum im- mersis, demum emergentibus, contextu pseudoparenchymatico, ca. 150—200 p. diametro, membrana perithecili ca. 15-25 crassa. Ascis cylindricis, apice applanatis ibique incrassatis, breviter stipitatis, bulla basali instructis, 8-sporis, non fasciculatis, ca. 60X8.5—9 uw, J—. Sporis hyalinis, anguste oblongis, subacutiusculis, rectis, medio 1-septatis (cellulis aequalibus), non constrictis, eguttulatis, irregulariter distichis, ca. 18— 19.5 3.5 %. Paraphysibus nullis. Masatierra: in foliis vivis Rhaphithamnt venusti (Phil.) Skottsb., Villagra, A Oor7: ASCOMYCETEN, FUNGI IMPERFECTI UND UREDINEEN 451 Die vorliegende neu beschriebene Art verhalt sich wie eine Dzdymella Sacc. (also parallelstehende Schlauche), hat aber keine Paraphysen, Jodreaktion fehlt, gehort daher in die von HOHNEL in Annal. mycol., vol. XVI (1918), p. 66 (nr. 207) aufgestellte Gattung Dzdymellina, deren Berechtigung allerdings spater von PETRAK [vgl. I. c., vol. XXII (1924), p. 16 (nr. 305)] angezweifelt wurde. Trabutia Escalloniae (P. Henn.) Theiss. et Syd. in Ann. mycol. vol. XIII (1915), p. 351. Phyllachora Escalloniae P. Henn. (nec Pat.) apud REHM in Hedw., Bd. XI (1901), p. (104). Masatierra: an lebenden Blattern von Fscallonia Calcottiae H. et A., Portezuelo de Villagra, 500—550 m, 2. IV. 1917. (Det. H. Sypow.) Bisher nur aus Argentinien bekannt. Skottsbergiella Petrak nov. gen. Stroma ausgebreitet, ziemlich scharf begrenzte, durch die stark pustel- formig vorgewolbten Perithezien grobkérnig rauhe Krusten bildend, sich im Periderm und Rindenparenchym entwickelnd, von parenchymatischen, oben fast opak schwarzbraun gefarbtem, weiter innen fast hyalinem Gewebe. Perithezien ziemlich gross, dicht herdenweise einschichtig beisammenstehend, oben fest mit der deckenden Stromakruste verwachsen, mit dieser halbkugelig vorspringend, oben plotzlich in das zylindrisch kegelf6rmige, ziemlich stark verlangerte, aussen fest mit hdckerigen oder leistenformigen Stromamassen verwachsene, dadurch meist sehr unregelmassig, oft verschieden kantig und gefurcht erscheinende Ostiolum verjiingt. Asci sehr zahlreich, zartwandig, 8-sporig, mit kurzem, sehr zartem Stiel. Sporen Janglich oder langlich spindelformig, seltener fast zylin- drisch, hyalin, zweizellig, beidendig mit einem kurzen, hyalinen, dornformigen, verganglichen Anhangsel, mittelgross. Pseudoparaphysen ziemlich zahlreich, bald stark verschrumpfend und verschleimend. Skottsbergiella diaporthoides Petrak nov, spec. Stroma ausgebreitet, im Umrisse meist ziemlich regelmassig elliptische bis ca. 35 mm lange und bis 18 mm breite, ziemlich scharf begrenzte, durch die sehr dicht stehenden, stark pustelformig vorspringenden Perithezien grobkornig rauhe, schwarzbraune Krusten bildend. Das Stromagewebe entwickelt sich der Hauptsache nach im Periderm und Rindenparenchym, dessen Zellen besonders in den obersten Schichten zuerst stark vergréssert und schliesslich fast ganz zerstOrt werden, wobei ihre Wande durch mehr oder weniger dickes, parenchymatisches, aus unregelmdssig rundlich eckigen, dunkel, oft fast opak schwarzbraunen, ca. 5—I0 yw. grossen, etwas dickwandigen Zellen bestehendes Stromagewebe ver- drangt werden, welches eine wabig grosszellige Struktur zeigt, weil das Lumen der Zellen durch — wahrscheinlich sekundar eingelagertes — aus unregelmassig eckigen, bis ca. 15 ». grossen, ziemlich diinnwandigen, fast hyalinen oder nur sehr hell gelbbraunlich gefarbten Zellen bestehendes Stromagewebe ausgefiillt 31—2610. The Nat. Hist. of Juan Fernandes and Easter Isl, Vol. II. 482 KARL KEISSLER wird. Weiter unten wird das Stromagewebe ziemlich undeutlich, die Zellen des Substrates sind hier nur wenig vergrossert und bilden mit dem Gewebe des Pilzes eine hyaline oder subhyaline, ziemlich grobkornig kriimelige Masse. Nach unten hin wird das Stroma durch das Kambium begrenzt, dessen Zellen hypertrophisch verdickte Wande haben und in den innersten Schichten durch ein hyalines oder sehr hell gelb-braunlich gefarbtes, aus rundlich eckigen oder ganz unregelmdssigen, ziemlich diinnwandigen, ca. 4—6 p. grossen Zellen be- stehendes Stromagewebe ausgeftllt werden. Die herdenweise, einschichtig bei- sammenstehenden Perithezien ragen scheinbar bis zur Halfte oder bis zu zwei Drittelu ihrer Héhe vor, in Wirklichkeit entwickeln sie sich in grossen mehr oder minder rundlichen Hohlungen des Stromas, deren untere Halfte eingesenkt bleibt, wahrend die obere, stark vorgewolbte Halfte von einer ca. 180—300 p. dicken Stromakruste ttberzogen wird, welche der Hauptsache nach aus dem stromatisch infizierten Periderm besteht, das nur am Grunde zwischen den Peri- thezien rissig unterbrochen, am Scheitel von dem dicken Miindungskegel durch- brochen wird, an den Seiten aber stets erhalten bleibt und hier stellenweise oft noch sehr deutlich die kaum oder nur sehr wenig infizierten Aussenwande der obersten Zellschichte erkennen lasst. Perithezien rundlich, meist etwas unregel- massig, ca. 600-800 » im Durchmesser, oben plotzlich in das zylindrisch- kegelformige, an der Spitze meist ziemlich stark verjiingte, stumpf abgerundete, unten ca. 100—160 ». dicke, bis ca. 500 wp. lange, von einem ca. 75 ». weiten, reich mit vorwarts gerichteten, fadigen, subhyalinen Periphysen ausgestatteten Miindungskanal durchbohrte Ostiolum verjiingt, welches aussen mit machtigen Stromamassen fest und vollstandig verwachsen ist und dadurch viel dicker, oft ganz unregelmassig und an den Seiten oft stark gefurcht erscheint. Perithezien- membran hautig, unten ca. 25—30 y dick, in der oberen Halfte allmahlich an Starke zunehmend und am Grunde des Ostiolums bis zu 100 w dick werdend, sich in der unteren Halfte leicht von der Innenflache der Stromahodhlung ab- losend, in der oberen Halfte, besonders am Grunde des Ostiolums fest mit dem Stroma verwachsend und allmahlich in das Gewebe desselben tbergehend, unten aus wenigen, oben aus immer zahlreicher werdenden Lagen von ziemlich stark zusammengepressten, diinnwandigen, durchscheinend schwarzbraunen, ge- streckten, ca. 10—25 y. langen, 3—5 p. breiten Zellen bestehend, welche sich innen rasch entfarben, fast hyalin werden, oder nur sehr hell gelblich gefarbt, inhaitsreich und oft sehr undeutlich sind, weshalb die innersten Schichten der Membran oft fast konzentrisch hyphig gebaut zu sein scheinen. Asci sehr zahl- reich, keulig oder fast langlich spindelformig, oben fast gestutzt abgerundet, kaum oder schwach, unten meist stark verjiingt und in einen kirzeren oder langeren sehr zarten und verganglichen Stiel iibergehend, 8-sporig, zartwandig, p. sp. ca. 68—75/7.5—11 .. Sporen zwei-, seltener undeutlich dreireihig, langlich- spindelformig, selten fast zylindrisch, beidendig schwach, seltener etwas starker verjiingt, stumpf abgerundet und mit einem hyalinen, verganglichen, schwer sichtbaren, zylindrischen, gegen die Spitze hin schwach verjiingten, stumpfen, ca. 5—7 ». langen, I.s—I.8 ». dicken, geraden oder schwach gekriimmten An- hangsel versehen, ungefahr in der Mitte mit einer Querwand, an dieser nicht oder nur schwach eingeschniirt, in jeder Zelle meist mit I—2 grdsseren aber meist sehr undeutlichen Oeltrépfchen und undeutlich kornigem Plasma, 15—23 p. ASCOMYCETEN, FUNGI IMPERFECTI UND UREDINEEN 483 lang, 3.5—5 pu, selten bis 6 p breit. Pseudoparaphysen ziemlich zahlreich, aber sehr verganglich, bis auf undeutliche Reste vollig verschrumpft und verschleimt. Die Gattung Skottsbergiel/la stimmt in Bezug auf den Bau der Fruchtschicht vollig mit Dzaporthe iiberein, lasst sich aber von dieser Gattung leicht und sicher durch das eigenartige Stroma unterscheiden. Masatierra: an lebenden Zweigen von Coprosma triflorum Benth. & Hook. fil., Quebrada Laura, Apr. 1917. Hypoxylon marginatum Mont. Masafuera: an abgestorbenen Asten von Myrceugenia Schultzei Joh., Quebrada del Mono, 12. II. 1917. Fehlt in JOHOW’s Flora von Juan Fernandez, ist aber schon fiir Chile angegeben. Das von MONTAGNE in Gay, Hist. Chile, vol. 7 (1850), p. 438 beschriebene //. Bertervzd macht nach der Abbildung (Atl. Crypt. nr. 9 fig. 3) eigentlich den Eindruck einer Poronza. Hypoxylon fuscum Fr. Masatierra: auf Holz, Quebrada Damajuana, 30. XII. 1916. Hypoxylon spec. Masafuera: an Zweigen von MLyrceugenta Schultzet Joh., zwischen Sanchez sung’ Toltén, 25. 11. 1917. Xylaria tenuissima Fr.’ Masatierra: an Zweigen, Quebrada de la Damajuana, 30. XII. 1916; an Asten von Myrceugenia fernandesiana Joh., Cordon del Centinela, 5. IV. 1917. Von MONTAGNE in GAy, Hist. Chile, Bot. vol. 7 (1850), p. 432, fiir Chile angegeben. Fungi imperfecti. Sphaeropsidales. Phyllosticta myrticola Sacc., Syll. fung. vol. 22/2 (1913), p. 867. Depazea myrticola Klotzsch in N. Act. Nat. Cur., vol. XIX (1843), p. 242; RABENH. in KLoTzscuH, Herb. viv. mycol. nr. 1146 (1847) ct in Bot. Zeit., Bd. 6 (1848), p. 293; Mont. in Gay, Hist. Chile, vol. 7 (1850), p. 490; JoHow, Estud. Fl. Juan Fernand. (1896), p. 193. 484 KARL KEISSLER Masatierra: an lebenden Blattern von J/yrceugenia fernandeziana Jon., Quebrada de la Damajuana, 30. XII. 1916. Liegt zwar steril vor, diirfte aber der obengenannte Pilz sein. Urspriing- lich aus Chile beschrieben, in JoHOW’s Flora auf gleicher Nahrpflanze, wie hier vorliegend, fiir Juan Fernandez angefuhrt. Phyllosticta Winterii Speg. Masatierra: an lebenden Blattern von Drzmys Wantert Forst., Puerto In- gles) 16,01), 1017- Rortezuelo des V illagtas=2-.1Ve 1917; Von SPEGAZZINI auf gleicher Nahrpflanze vom Feuerlande beschrieben. Sporen zylindrisch abgerundet, ohne Oeltropfen, ca. 3.50.5 , also etwas von den Angaben bei SPEGAZZINI abweichend. Die von demselben Autor auf Blattern der gleichen Nahrpflanze beschriebene Phoma Drimydzs (ebenfalls aus Feuerland) diirfte vielleicht gleichfalls eine Phy//losticta sein, die aber spezifisch gut verschieden erscheint. Disculina Myrceugeniae nov. spec. Pustulis quasi callosis, brunneis, hypophyllis, + rotundatis, disciformibus, perspicue limitatis, ca. O..—I mm diametro, poro nigro subrotundo vel rimoso dehiscentibus; stromate(?) subepidermali, nigro, tenui; loculo uno lentiforme, primum epidermide tecto, deinde illam rumpente, circum conidiophoris praedito. Conidiophoris brevibus, hyalinis, simplicibus. Conidiis hyalinis, acicula- ribus, utrinque acutatis, leviter curvatis, aseptatis, eguttulatis, ca. 16X 1.8 w. Masatierra: ad folia subviva Myrceugeniae fernandesianae prope Pangal, ° Thos de OM Es liegt hier ein Vertreter der von HOHNEL in Fragm. Mykol. nr. 988 (vgl. Sitzungsber. Ak. Wiss. Wien, m. n. kl. Abt. 1, Bd. 125 (1916), p. 78 aufge- stellten Gattung Dzsculina vor, als deren Typus Cryptosporium Neesi Corda zu betrachten ist. Sie entspricht gleichsam einer Dzscu/a mit grossen, (meist) gekriimmten (cylindrisch-)spindelformigen Konidien. Fly phomycetes. Botrytis Bassiana Bals. Masatierra: an toten Kafern unter Rinde, Plazuela del Yunque IV. 1917. Botrytis cinerea Bon. Masafuera: an Friichten von Av7stotelia Maqui L’Heér., Quebrada Chozas, 2O- EON 7: ASCOMYCETEN, FUNGI IMPERFECTI UND UREDINEEN 485 Uredineae. Uredo australis Diet. et Neg. Masafuera: auf der Unterseite lebender Blatter von Luphrasia formosis- stma Skottsb., auf der Hochebene, 1200 m,.7. III. 1917. Bisher aus den Anden in Chile auf Luphr. andicola und chrysantha (leg. NEGER) bekannt. Uredo invisa Speg.? Masatierra: Auf lebenden Blattern (unterseits) von Solanum fernandezt- anum Phil. Valle Colonial, Quebrada del Monte Maderugo, 345 m, 14. I. 1917. Wenn auch Solanum fernandesianum auf Juan Fernandez endemisch ist, so scheint doch der auf dieser Nahrpflanze auftretende Pilz ganz gut auf obige Art zu passen, welche von SPEGAZZINI aus Argentinien und Uruguay auf S. sisymbritfolium beschrieben wurde. Sporen blassbraun, rund, warzig, ca. 23— 26.5 |. Uredo Robinsoniae nov. spec. Soris hypophyllis, sine maculis, dense aggregatis, totum fere folium occu- pantibus, minutis, rotundatis vel ellipticis, primum epidermide tectis, dein ea fissa cinctis, albidis, ca 0.5 cm (0.8—I cm) diametro; sporis subglobosis (inter- dum subangulatis), ellipticis (rarius subpyriformibus), albidis, dense verruculosis, ca. 23—32X17.5—26.5 uy. metientibus, episporio ca. 3.5 . crasso. Masatierra: in foliis vivis Rodznsoniae macrocephalae DC. prope El Gentinelan(300° m, 20:-Tll- 1917) et’ Puerto Frances (250 m, 13. XIT. 1o16); in foliis vivis juvenilibus Rodznsontae Gayanae DC., prope Portezuelo de Villagra, Gas550°m, 2°1V.- 1917. Fir die auf Juan Fernandez endemische Gattung Lodznsonza ist allem Anschein nach bisher keine Uredineen-Art beschrieben worden. Die Exemplare des Pilzes auf R. Gavana stimmen in allem mit jenem auf 2. macrocephala iiberein, nur sind die Sporenlager mehr rund und grodsser (ca. 0.8—I cm), wahrend diejenigen auf den Blattern der letztgenannten Art mehr elliptisch und kleiner (ca. 0.5 cm) sind. Die auf den Blattern von 2. macrocephala unterseits, oft nach der Oberseite durchdringenden, grau-griinlichen Flecken, die haufig unabhangig von der Uredinee erscheinen und nur manchmal in die Sporenlager zufallig sich hineinerstrecken, scheinen den Anfangen eines anderen Pilzes an- zugehoren. In denselben sieht man manchmal dickere schwarze Hyphenstrange. Uromyces Cestri Monv. Aecidium Cestrt Mont. Masatierra: Auf der Unterseite der Blatter von Cestrum Pargut (Aecidien), Puetto Ingles, 20. 1. ‘1917- 486 KARL KEISSLER Nach Sypow, Monogr. Ured. II (1909), p. 29 gehort Aectdinm Cestr7 in den Formenkreis von Uvromyces Cestvz. Bereits fiir Masatierra auf gleicher Nahrpflanze angegeben. Uromyces Dendroseridis nov. spec. Soris teleutosporis hypophyllis (interdum maculis minutissimis, in pagina superiore foliorum vix visibilibus, brunneolis insidentibus), subrotundis, parvis, ca. 03—04 cm diametro, plerumque sparsis, interdum aggregatis vel subcon- fluentibus, brunneis, pulverulentis, nudis, limbo angusto epidermidis fissae cir- cumdatis. Teleutosporis subglobosis, pallide-brunneis, subtiliter spinulosis, ca. 31—33.5X23—27 ww metientibus; episporio crasso, ca. 5.s—7 , membrana exteriore brunnea, ca. I y. crassa, interiore flavida 4.5—6 uw crassa. Masatierra: in foliis vivis Dezdroseridis micranthae Hk. et Arn. prope Portezuelo de Villagra ca. 560 m, 27. III. 1917. Fiir die auf Juan Fernandez endemische Gattung Dendroserzs scheint bisher keine Uredinee bekannt worden zu sein. Weissliche Flecken, und schwarze Punkte, die gelegentlich auf den Blattern auftreten, sind auf Frassspuren und Exkremente einer 7/rzps-Art zuriickzufihren. Kuehneola andicola Diet. Uredo andicola Diet. et Neg. Masafuera: Auf den Blattern von Rubus geocdes Sm., auf der Hochebene, 1200 m, 14. II. 1917. Uredo-Stadium. Bisher aus den Anden von Chile (leg. NEGER; SyDow, Ured., nr. 41 a) und aus dem Feuerlande (leg. SKOTTSBERG; VESTERGREN, Microm. rar. sel., nr. 1725) bekannt. Printed ‘*/6 1927. 17. The Vegetation of Easter Island, By CARL SKOTTSBERG. With Plates 28—33. In Vol. I, p. 4—8, some notes on the geological structure of Easter Island were published; I refer to these notes and to the literature quoted in the same place. Besides, reference should be made to W. KNOCHE'’s paper »Algunas observaciones a la geomorfologia de la Isla de Pascua» (Inst. Centr. Meteor. y Geofis. de Chile, No. 4, p. 150-154; also with German text). Some of his place names are not in correspondance with those on the more recent Chilean map published by me I. c. KNOCHE’s » Volcan Rana Hana Hana» is = Maunga Terevaka and Rano Aroi, while his »Terevaca» seems to be the same as Cerro Otu on the Chilean map. His »Puukiteke» is Maunga (or Puu) Katiki; »Puu» is used for »hill» in many Hawaiian names. Easter Island scenery is very monotonous (pl. 28, 29). The island is a low, undulating basaltic plateau of practically the same geological structure everywhere and covered by the same kind of vegetation. A considerable number of cones, of many different colours and with a varying petrographic structure, rise above the plains, but their influence on the distribution of the flora is slight if any. Even the highest mountain, Terevaka (530 m) is not sufficiently elevated to call forth any distinct vertical regions. A greater in- terest is attached to the three »ranos», the comparatively deep craters con- taining a pool of water, but only one of them, Rano Kao, is deep enough (about 200 m) to act as a secluded valley and to offer better conditions for arboreous vegetation. The hydrographic features are remarkable. While other volcanic South Sea islands exhibit signs of a very substantial erosion, with deep ravines and narrow ridges, there is not a single valley or ravine in Easter Island with ex- ception of the grooves on the steep crater rims. There is no stream anywhere on the island. Still, the rainfall is considerable, heavier than in Juan Fernandez, where running water is plentiful and erosion very active. This difference be- tween islands of roughly the same structure and with the same type of climate has long attracted the interest of visiting naturalists. Even if we remember that Easter Island is rather low and that, as to its original morphology, it must have been different fromJuan Fernandez, this cannot explain why not the higher land in the northwest and northeast parts has been cut up into ridges and valleys carry- 488 CARL SKOTTSBERG ing water at least after the heavier rains. The explanation generally offered is that the rocks are so porous that all the water rapidly disappears beneath the surface and collects at various depths or drains off through underground chan- nels, eventually reaching the sea through caves in the coast rocks. Evapora- tion, stimulated by frequent and fresh winds, is considerable. Consequently, the original outline of the cones and craters has been retained so that the topography presents a very youthful appearance, but from this we cannot conclude that the island is very much younger than other Pacific islands. Ex- tensive talus formations in the craters, especially in Rano Kao, bear witness of long ages of slow erosion, and the destructive work of the sea is quite im- pressive. KNOCHE suggests that the island may have been twice as large as it is now. The northernmost of the parasitic cones of Katiki has been worn away so that only half of it is left, and the same is the case with an- other crater west of Anakena. And the small rocky islets off the shore on the south side of Rano Kao must have formed part of the southwestern head- land. The volcanic forces have come to rest very long ago, and no tradition exists of any kind of volcanic activity. The soil on the lava plains is stony and shallow, and bare rock face is exposed in many places. Cracks and accumulated boulders arrest the finer soil and have a more luxuriant vegetation. The soil colour varies with that of the mother rock from a deep brown to red and yellow of several hues. While the basalts, although they are porous and honeycombed, seem to be very hard, and the boulders have sharp angles, the tufas are loose and easily disinte- grated, whence follows that the slopes of the cones are covered with vegeta- tion. In places a loose volcanic sand is found. Where there is sufficient soil this is of good quality and quite fertile when properly cultivated, and in pre- historic and early historic times extensive plantations existed, supporting a population of several thousand people. The injurious effects of the strong winds were lessened through the construction of enclosures. Various kinds were described and figured in my »Notes», p. 13, 14. Insect life is very poor. Earthworms occur, but so far only an intro- duced species has been discovered. The climate. The island lies in the zone of the trade winds (Mataveri, lat. 27° 10! S, long. 109° 26' E) which put their stamp on the weather. As all the land is of a very moderate height, there is no perceptible difference between the different sides of the island, and the climate is probably very uniform all over the place. The rainfall is greater on the more elevated hills. In 1911, a well equipped meteorological station was erected in Mataveri, and a smaller one on the rim of Rano Kao. Besides, there was a pluviometer on Otu (Terevaca of KNOCHE). All these places are situated in the south- west part of the island. The observations were run from May 1, IgII, to April 30, 1912, and published in the Reports of the Instituto Central Meteoro- légico y Geofisico de Chile (No. 4, 1913). The observations were continued in THE VEGETATION OF EASTER ISLAND 489 Mataveri during Ig912 (same series, Nos. 6, 8 and 10, 1914) and 1913 (No. 13, 1914, and Nos. 15 and 16, 1915), but after that year the item »Isla de Pascua» unfortunately disappears from the reports. Temperature I1912—-13. Mataveri, alt. 30 m. med. med. abs. abs. med . max. min. max. max. WADE). » His tf MID 26.3 19.4 28.4 16.8 Bebe retis), ./f23c7 A Br 19.7 31.0 17.2 Mareh, 05) (5i.6 2255 26.3 19.8 20.5 17.5 prays 2. ea) (2x 24.6 18.2 27.0 13.5 (EIS Fearne apes) 1c) 2915 77, 26.4 12.7 mers 2 17 7 20.8 WS <4 23.6 12.2 abyss. sas. cortaceum, 1 (rare) Paspalum Forsterianum, 1 Oxalis corniculata, I F, Digitaria sanguinalis, 4 G Euphorbia serpens, 2 Erythraea spec.? (seedlings), Mosses between the bunch grasses and filling small fissures: abundant Bryum argenteum var. lanatum and Campylopus saxicola; less common Cam- pylopus introflexus and Wersza spec. On the rocks and boulders: Physcza picta-Luellia-Ass. Physcia picta (white), Buellia Glastouana (straw-coloured, black fruit), ZB. stellulata f{. protothallina (greyish-white, black fruit), Dzploschistes anactinus (large mouse-grey crusts), Heppza Guepint (squamulose, bronze-coloured crusts), Parmelia caperata (greyish green). Slopes of Maunga Kattki, Fune 16, 1917. A. Lower slopes, below 300 m. Deeper soil than on the plain, of a reddish brown colour. Stones hidden under a dense grass cover. Sporobolus elongatus-Lragrostis elongata-Ass. F, Sporobolus elongatus, 3 Digitaria sanguinalis, 2 Dichelachne cf. crinita, 2 Ophioglossum coriaceum, 1 Axonopus paschalis, \ » reticulatum, 1 (rare) F, Lragrostis elongata, 5 G Euphorbia hirta, 1. Mosses not collected; »same species as in Hanga Ho Orno». B. Higher slopes, 300—400 m. Moister. Axonopus paschalis-Kyllinga brevifolta-Ass. F, Axonopus paschalts, 3 F, Kyliinga brevitolia, 3 Erigeron linifolius, 2 Digitaria sanguinalis, I Dantonia paschalis, 1 (rare) Ophioglossum cortaceum, 1 G A WMontia-like plant, sterile, material lost. Mosses as before. Stipa horridula was observed in this association forming a couple of small societies on the seaward slope. 1 Previously referred to C. echinatus. 32—2610. The. Nat. Hist. of Juan Fernandez and Easter Isl. Vol. II. 498 CARL SKOTTSBERG C. The shallow crater depression, a circular saucer-shaped hollow, about 75 m across and 5 to 6 m deep. Moist, sandy soil. Ay/linga brevifolia-Ass. F, Ayllinga brevifolia, 5 Erigeron linifolius, 1 Axonopus paschalts, 3 F, Agrostis retrofracta, 1 Dichelachne sciurea, \ Oxalis corniculata, 1. Eragrostis elongata, 1 On outcrops of rock along the crater rim: Auella stellulata f. protothal- lina (light grey, black fruit), and Usuea Stetnert var. tincta (yellowish). Crater of Rano Kao, inside, higher slopes, c. 275 m. Fune 23,1917. Spo- robolus elongatius-Ass. F, Sporobolus elongatus, 4 Digitaria horizontalis, 1 F, Avxonopus paschalis, 2 Microlepia strigosa, 1 Kyllinga brevifolia, 2 F, Cynodon dactylon, 2 Bromus unioloides, 1 Oxalis corniculata, I G Mosses: Philonotis laxissima (small carpets), Wersza flavipes. LOCAL EDAPHIC COMMUNITIES IN CREVICES CONTAINING WATER Fissure on the slope of Vaintu Rova, c. 300 m. Fune 16, 1917. Kyllinga brevifolia-Ass. F, Microlepia strigosa, 2 F, Kyllinga brevifolta, 4 Polypodium phymatodes, 2 Asplenium obtusatum, 1 Cyperus vegetus, I Nasturtium sarmentosum, \ Doodia paschalis, i (rare) G Mosses and lichens: /iss¢dens pascuanus (mats), Papillaria pascuana. Anaptychia speciosa {. soredifera, Parmelia reticulata (whitish). — Hepaticae immixed, not determined. Deep, dark crevice on Tea-tea, c. 290 m. In the bottom a spring. On the vertical walls Polypodium phymatodes-Ass. F, Polypodium phymatodes, 3 Doodia paschals, \ Microlepia strigosa, 2 Sonchus oleraceus, I. LOCAL COMMUNITIES ON THE TALUS SLOPES IN THE CRATERS Rano Kao, Fune 22—2}3, 1917. On the inner slopes, three kinds of habitats may be distinguished: the outcrops of rocks with lichen associations, patches of finer material covered with the Sporobolus-Ass. (see above) and the extensive accumulations of larger THE VEGETATION OF EASTER ISLAND 499 boulders fallen from the walls. The upper parts of this talus are barren, the less coarse particles having been washed down, and the vegetation is the same kind of broken J/zcrolepia-Ass. so often observed in similar places, with the addition of Polypodium phymatodes. Lower down, elev. 120—130 m, there must be a good deal of soil among the boulders. Here are old plantations of bananas (Pl. 30), naturalized specimens of Aroussonetia, Cordyline etc. and planted trees of Melza azedarach, Acacia sp. and others. As a result of the nature of the substratum, there is no closed plant cover. The vegetation is a complex of association fragments: carpets of J/zcrolepia and other ferns, with some trees, luxuriant moss-mats, especially in depressions between the boulders, and a mosaic of crustaceous lichens on the stones. This is the richest vege- tation found in the island. Microlepia strigosa-Ass. with scattered trees. Pl. 31. S Broussonetia papyrifera, {, not over 2.5 m high Triumfetta semitriloba, 1, rare F, Sophora toromiro, one tree, 1.7 m F, Microlepia strigosa, 4 Dryopterts gongylodes, i Polypodium phymatodes, 2 > parasitica, 1 Ageratum conyzoides, I Solanum Insulae-Paschalis, 1 Caesalpinia bonduc, a trailing Verbena litoralis, 1. shrub, I, rare On the bark of Broussonetia and Sophora: Arthonia fuscescens and Gra- phis lineola. Vittaria elongata-Moss Ass. F, Vittaria elongata, 1 G Campylopus dicranodontioides, C. saxicola, Kabronia macroblepharoides, Macromitrium spec., Papillaria pascuana, Rhacopilum cuspidigerum, . Frullania spec., Lejeunia spec. Anaptychia spectosa {. soreditfera, Cladonia pityrea f. soredtosa, Par- melita reticulata, all whitish. Lichen Ass. F, Usnea Steinert var. tincta (yellowish). G Suellia Glaziouana (straw-coloured, black fruit), B. ste/dulata f. proto- thallina (light grey, black fruit), 2. spec. (black), Dzploschtstes anactinus (mouse- grey), Opegrapha paschalis (greenish grey, black fruit), Parmelia reticulata (whitish). Rano Arot, Fune 25, 1917. Rano Aroi is a shallow crater without arboreous vegetation. The lower talus slopes, elev. c. 400—420 m, are composed of smaller stones and boulders, and the vegetation more evenly distributed than in R. Kao. Just as in this place, three kinds of associations are distinguished. 500 CARL SKOTTSBERG Dichelachne sciurea-Microlepia strigosa Ass. F, Dechelachne sciurea, 3 Hypochoeris radicata, 1 F, Microlepia strigosa, 3 Sonchus oleraceus, I Polypodium phymatodes, 2 Verbena litoralis, 1 Cyperus polystachyus, 1, rare F, Calystegia sepium fia, I Doodia paschalis, 1, rare Dryopteris Espinosat, 1, ge- Dryopteris parasitica, 1 nerally under stones. Elaphoglossum tahitense, 1, G Cryptogams not collected. gregarious Vittaria elongata-Moss Ass. F, Vittaria elongata, 2 G Campylopus adicranodontioides, Papillarta pascuana. The lichens were not collected. The third large crater, Rano Raraku, is wide open and the soft tufa forms a dry, sandy soil covered with Sporvobolus-Ass. The rocks are devoid of macroscopic vegetation, but on nodules of hard stone Auellza Glazzouana, B. stellulata and Physcia picta (badly developed) were observed. THE CRATER LAKES Rano Kao, Fune 23, 1917. The lake, almost circular in outline, measures between 600 and 700 m across. No outlet is known. Its surface lies about 110 m above sea level. Its depth is not known but supposed to exceed 100 m. Round the edge is a zone of shallow water filled by a matted, elastic network of Polygonum rhizomes, across which a firmer reed zone is reached. Together they form the mar- ginal swamp. The reed belt leads over to the peat bog, which is supra-aquatic. Evidently this bog should be classed with the high-moor formation. No Sphagnum has been found; the peat-forming moss is a species of Campylopus, endemic in Easter Island. The surface of this bog looks flat but it is not quite even. Drier, almost inconspicuously raised patches are found. The bog forms a continuous cover over the lake, leaving only the margin free. It is possible that the principal reason why there is open water round the margin is the constant afflux of rain-water from the steep slopes. I. Outer swamp zone, 5 to7 m wide. Polygonum acuminatum-Ass. Plate 32. F, Polygonum acuminatum, 5. 2. Reed-swamp zone, of very variable width, sometimes much wider than the pure Polygonum belt. Scirpus riparius-Ass. Plate 30, 32. S Scirpus riparius var. paschalis, ¥, Dryopteris gongylodes, 2 3 to 4 m high, 5 Kyllinga brevifolia, 2 F, Polygonum acuminatum, 5 G Campylopus hygrophilus, 5. THE VEGETATION OF EASTER ISLAND 501 Campylopus lygrophilus, an endemic species, is submerged and fills the numerous water-holes. The Scpetum reappears round the irregular pools of open water scattered over the bog (distinctly visible on plate 33), and is accom- panied by the same Campylopus, which forms a pure association in the water, corresponding to the hydrophilous Spagna in peat bog depressions elsewhere. I do not know if the water-holes are primary or secondary. 3. The peat bog. Campylopus turficola-Ass. Plate 33. A sordid green peat cover of unknown thickness, floating on the water. The depth of this peat must be considerable. It is wet, but of course not covered with water, and it is firm enough to carry the weight of a man, but there are many treacherous places. Even cattle are reported to have crossed the bog safely. F, Agrostis retrofracta, 1 F, Vittaria elongata, 1 Axonopus paschalis, 1 G Campylopus turficola, 5. Asplenium adiantoides var. squa- mulosum, I On the slightly raised areas Campylopus turficola is replaced by a form called f. depauperata. Cladonia pityrea f. sorediosa is immixed. Rano Aroi, Fune 25, 1917. Alt. c. 400 m. The crater has gentle slopes and is shallow. The lake is probably not deep; it is overgrown with a vegetation less differentiated than in R. Kao. The Sczrpetum does not form a continuous belt. 1. Scirpus riparius-Kyllinga brevifolia-Ass. F, Scirpus riparius var. paschalis, Dryopteris gongylodes, 1 in patches, 5 Polygonum acuminatum, 1 F, Kyllinga brevifolia, 5 G Mosses not collected. Cyperus vegetus, 2 2. Campylopus-Ass. with fairly abundant Agrostis retrofracta. No collections made. Rano Raraku, Fune 29, 1917. The crater floor is wide and the lake probably shallow. The water is open with a wide fringe of Scirpus riparius var. paschalis and Polygonum acumina- tum. 1 do not know why the Campylopus cover is absent here, unless some chemical property of the water is responsible, but this is difficult to believe. The open water was not accessible. CONCLUDING REMARKS In order to study the plant communities of Easter Island, poor as they are, more time is needed than was at our disposal, but I hope that at least the main features and the principal associations have been described with fair 502 CARL SKOTTSBERG accuracy. An exception must be made for the top region of Terevaka, the more important as it rises above all the other hills. We ascended this moun- tain ina gale of wind and in a deluge of rain, making note-taking impossible, and we were soon forced to retrace our steps, especially as illness made me little capable of any serious undertaking. We did not see any other vascular plants than such as we had found below, but there was a notable increase of mosses, mixed with a species of C/adonia with podetia. As we resolved to return no collecting was done. Unfortunately our ship, the »Baquedano» of the Chilean Navy, left two days later without giving us a chance to visit the mountain a second time. Explanation of Plates. 28 A. The country E of Hanga Roa and Mataveri, with a grove of planted trees. Photo. C. Skottsberg. 28 B. Maunga Katiki with parasitic cones: Vaintu Rova, Tea-tea and the half cone mentioned on p. 488. In the foreground the stony grassland at Hanga Ho Orno. Photo. C. Skottsberg. 29. View of the island from the north slope of Rano Kao. In the foreground dense Sforo- bolus-Ass. Photo. K. Backstrém. - 30. In the crater of Rano Kao, steep slope above the lake with plantations of bananas among the huge boulders, and a glimpse of the Sczrfetum along the lake edge. The tree in the center 1s Melia azedarach,; to the left of this, Avundo donax is observed. Photo. C. Skottsberg. 31. Same locality as pl. 30. Left, a small specimen of Sofhora toromiro, right, a fern mat, mainly MMicrolepia strigosa,; to the left of the big boulder a frond of Polypodium phy- matodes, Photo. C. Skottsberg. 32. Crater lake of Rano Kao, from the north shore. In front, Polygonum acuminatum, be- hind, Sczrfus rifarius var. paschalis. Photo. C. Skottsberg. View from the upper slope of Rano Kao across the crater lake. The Campylopfus bog with water-holes. The small dark dots are tufts of Asplenium adiantoides var. squa- mulosum. Photo. K. Backstrém. Ww ww Printed ?"/6 1927. Nat. Hist. Fuan Fernandez aud Easter Isl. Vol. II. PLATE 28. “¢ ‘ aes ne De le Yeon Ta ag oP. Sang ip : oe ; m* or Nat. Hist. Fuan Fernandes and Easter Isl. Vol. 11. PEATE. 20) Jif Vol. ster 1s/, 10 > and I Fuan Fernande: ELISA Nat. om Pein. "7 | /sZ. and Faster Fuan Fernandes XY FTtst. Nat. Nat. Hist. Fuan Fernande and Faster /sl. Paore JT. a4 JO: PLATE Ee ol, SZEK LSi= aol} “a ernandes and / Nat. Hist. Xuan F 18. Pollinationsbiologie und Samenverbreitung auf den Juan Fernandez-Inseln. Von CARL SKOTTSBERG. Wahrend unsres Aufenthalts auf den Juan Fernandez-Inseln nahmen flo- ristische und soziologische Untersuchungen unsere Krafte in vollem Mass in Anspruch. Fiir bliitenbiologische Fragen blieb nur wenig Zeit brig; gelegent- lich wurden aber auch auf diesem Gebiet einige Beobachtungen gemacht, die ich unten zusammengestellt habe. Sie sind nachtraglich durch einige Unter- suchungen an Spiritusmaterial erganzt vorden. Phanologische Daten haben wir die ganze Zeit gesammelt — wenn auch nicht in dem anfanglich geplantem Umfang —, weil geniigende Kenntnis vom Gang des Bliihens fiir ein gutes Verstandnis des pflanzengeographischen Florencharakters nicht unwichtig ist. Die Zeit unsres Aufenthalts, 1/12 1916—®°/4 I917 reichte zwar nicht aus, um ein vollstandiges Bild von der Phanologie zu geben; erganzt durch meine fru- heren Beobachtungen im Monat August 1908 und durch zerstreute Angaben bei anderen Verfassern diirfte aber die hier verdffentlichte Zusammenstellung nicht ganz ohne Wert sein. Das Klima der Inseln ist warmtemperiert und méassig regnerisch, die atmosphirischen Verhaltnisse recht giinstig fiir ein reiches Insektenleben. Das Wetter ist nur mdssig windig, selten sehr stiirmig. Frost kommt wahrschein- lich vor in den hdchsten Teilen von Masafuera, aber nicht in tieferen Lagen und nicht auf Masatierra. In der vegetativen Tatigkeit der Gewachse ist eine winterliche Ruhezeit wenigstens in den meisten Fallen recht ausgepragt. Pollinationsbiologie. Die Bliitenbesucher. Von bliitenbesuchenden Insekten ist nur eine ganz kleine Anzahl bekannt. Sehen wir von den Coleopteren und den noch unwichtigeren Gruppen ab, so kom- men nur Schmetterlinge und Fliegen in Frage, denn unter den Hymenopteren scheinen Biene und Hummeln ganzlich zu fehlen. Unsere Ausbeute von Dip- teren ist noch nicht bearbeitet worden, dagegen findet sich in diesem Werk, 504 CARL SKOTTSBERG Bd III, s. 255 ff. eine von CHR. AURIVILLIUS gemachte Zusammenstellung der gesammelten Schmetterlinge. Unter diesen sind nur wenige regelmassige Blu- menbesucher, und es befindet sich darunter ein einziger Tagfalter, Pyramezs carye, welcher ausserdem wahrscheinlich erst in spaterer Zeit eingeschleppt wurde. Er ist auf Masatierra haufig und ein fleissiger Besucher verschiedener Unkrduter in der basalen Region, widmet aber auch einigen der einheimischen Arten seine Aufmerksamkeit. Schwarmer fehlen, dagegen sind unter den 8 bekannten Noctuiden ein paar Blumenbesucher, von welchen die wichtige Syngrapha gammoides besonders erwahnenswert ist. Auch einige Eulen der Gattungen Lycophotta und Rachipleura wurden verhaltnismassig haufig beob- achtet, sonst macht aber die Grossschmetterlingfauna einen armlichen Eindruck. Man ist etwas erstaunt, eine so grosse Anzahl von Pflanzen mit lebhaft ge- farbten Bliiten zu finden, darunter auch didzische Arten. , 4 Es kommt aber eine weitere Besucherklasse in Betracht, die Trochiliden der Gattung Lustephanus; es sind zwei Arten, der endemische £. fernandensis und E. galeritus, welcher wahrend des Winters auch auf dem gegentberliegenden Festland vorkommt. Die Kolibris gehéren auf Masatierra zu den bedeutungs- vollen Bestauber. Friiher kam £. fernandensts — denn £. Leyboldi ist hochst- wahrscheinlich mit ihm identisch — ebenfalls auf Masafuera vor, wo ich im Fig. 1—2. K6pfe der Trochiliden von Juan Fernandez. 1. Lustephanus fernandensis §. EE. galeritus Q. Nat. Gr. August 1908 mehrere beobachtete. Im 1917 bekamen wir keinen einzigen zu sehen; er scheint hier ausgestorben, was tbrigens keinen, der das Schicksal Masafueras kennt, iiberraschen wird. Fig. 1 und 2 stellen Umrisse der Kopfe der beiden Trochiliden dar in natiirlicher Grosse. Die Exemplare im Zool. Museum, Gothenburg, standen mir zu Verfiigung. Sie sind aber nicht ausgestopft und ich habe deshalb nur den Schnabel und das Federkleid an dessen Wurzel in allen Einzelheiten dar- stellen konnen. Korper- und Schnabellange der untersuchten Exemplare, in mm. K6rper (ohne Schnabel) Schnabel Eustephanus fernandensis o&..... 119 17 fre 118 16 Cure 120 16.5 One ; 106 : 17 Doo oo % 109 17 Eustephanus galeritus Cie Sa. Xe 88 15 OF: 7 oO 6 gO 14,5 Wenigstens £. fernandensis hat einen geniigend langen Schnabel um die langrohrigsten Arten ausbeuten zu kénnen. Nach Mitteilung von Prof. Dr. E. POLLINATIONSBIOLOGIE UND SAMENVERBREITUNG AUF DEN JUAN FERNANDEZ-INSELN 505 LONNBERG kann die Zunge noch wenigstens I cm ausgestreckt werden. Ob dies beim Nektarsaugen geschieht, ist mir nicht bekannt. Die fortschreitende Vernichtung der urspriinglichen Vegetation iibt natiir- lich einen verhangnisvollen Einfluss auf die Fauna aus. Fiir den direkten Ver- lust von Honig und Pollen als Nahrung wird aber ein gewisser Ersatz dadurch geboten, dass viele adventive Pflanzen sich eingefunden haben. Fiir die ein- heimischen bedeuten sie vielleicht einen Konkurrenz um die Aufmerksamkeit der Blumenbesucher. Die Unkrauter bewohnen im allgemeinen die unterste, fast stets sonnige Region, und hier findet man auch Schmetterlinge und Fliegen lebhaft tatig, wahrend die finstere Waldregion und die Hochkimme, wo viele schonbliihende endemische Arten zu Hause sind, nur wenig von Insekten be- lebt werden. Haufigere entomophile Unkrauter sind Szlene gallica, Brassica napus, Matthiola incana, Kuta chalepensis, Aristotelia maqui, Lythrum hyssopi- folia, Ugni Molinae, Oenothera mollissima, Cynoglossum creticum, Verbena Lito- ralis, Physalis peruviana, Dipsacus silvestris, Gnaphalium cheiranthifolium, Lrigeron canadensis, Stlybum marianum, Centaurea melitensis. Auch die ornithophilen Arten haben einen Konkurrenten erhalten in Lobelia tupa, deren leuchtend rote Bliiten von Lustephanus galeritus besucht werden. Es ergibt sich demnach, dass auf dem Gebiet der Fremdbestaubung durch Tiere die Verhaltnisse fiir die einheimische Flora nicht sehr giinstig sind. Unten werden die Arten innerhalb jeder Gruppe in Ubereinstimmung mit meiner Arbeit »The Phanerogams etc.», s. 95 ff. aufgezahlt. Anemophile Arten. Gramineae. 13 Arten. Bei Bromus masafueranus wurde Kleistogamie konstatiert. Cyperaceae. 14 Arten. Palmae. Fuanta australis. WDidzisch. Die Bliiten sind I. c. 108 abge- bildet worden. Die grossen und schweren Staubbeutel sind sehr leicht beweg- lich, weil die Faden oben stark verdiinnt sind. Die miannlichen Bliiten sind wohlriechend, die weiblichen haben keine Duft. Juncaceae. 6 Arten. Piperaceae. Peperomia, 4 Arten. Es liegen keine Untersuchungen vor, vermutlich ist aber die Gattung anemophil. Urticaceae. Uvtica, 2 Arten, von welchen ich nur U. fevnandeziana aus eigener Erfahrung kenne. Den Bau der Bliitenknauel habe ich in »Studien» S. 20 behandelt. Die Bliitenkonstruktion ist die fiir die Gattung typische. — Boehmeria excelsa. Monozisch. — Parietaria debilis. Nicht untersucht. Chenopodiaceae. Chenopodium, 3 Arten, strauchig bis baumformig; gy- nomondozisch. Die Bliiten wurden in »Phanerogams» 120 abgebildet. — Sadz- cornia peruviana. Wie andere Arten dieser Gattung. Rosaceae. Margyricarpus setosus subsp. digynus und 2 Acaena-Arten. Vom Sanguisorba-Typus. Euphorbiaceae. Dysopsis glechomoides subsp. hirsuta. Monozisch. Die kleinen unscheinbaren Bliiten sitzen axillar in 2- oder 3-bliitigen Meonochasien, 506 CARL SKOTTSBERG jedes Geschlecht fiir sich oder beide zusammen. Nektar fehlt, und die Expo- sition der Bliiten lasst uns mit Sicherheit auf Anemophilie schliessen. Wahrend die 2 Bliite fast sitzend ist (Stiel 2—3 mm) erhebt sich die ¢ iiber die Blatter auf einem einen bis mehrere cm langen, leicht beweglichen Stiel. Empetraceae. mpetrum rubrum. Als grosse Seltenheit auf dem Gipfel von Masafuera. Nur steril gefunden. Gunneraceae. Die Bliitenverhaltnisse der drei Ganznera-Arten habe ich schon eingehend (Studien, 10 ff.) geschildert. Die ausgepragte Protandrie (es kommen auch rein weibliche Bliiten vor) spricht fiir Fremdbestaubung. Geitono- gamie wird aber leicht eintreffen konnen, weil sich die Bliiten des sehr grossen und dichten Bliitenstandes sukzessiv entwickeln. Die Staubbeutel sind sehr kurz gestielt und nicht beweglich, die Narben sind aber lang und biirsten- formig. Nach SCHNEGG (Flora, Bd. 90, I902) sollen G. chzlensts und ei- nige andere Arten apogam sein. Die Arten von Juan Fernandez sind in die- ser Hinsicht nicht untersucht worden. Zwischen G. pelfata und bracteata fand ich aber eine ganze Serie von Zwischenformen, die ich als Bastarde und deren Nachkommen gedeutet habe. Halorrhagidaceae. Die drei Halorrhagis-Arten schliessen sich Gunnera an, sind jedoch viel »typischer» anemophil. Die Bliitenstande sind locker, die Bliiten gestielt und die Antheren leicht beweglich auf langen Faden. Be- schreibungen und Abbildungen findet man in »The Phanerogams» 153—157. Plantaginaceae. Plantago fernandezia. Bekannter, sehr charakteristischer Schopfbaum. Die grossen, langgestielten Ahren sind axillar und treten zwischen den Blattern des machtigen Schopfes nickend hervor. Bau und Entwickelung der Bliite sind fir die Gattung typisch. P. truncata subsp. Skottsbergi ist eine einjahrige Pflanze von sehr alltaglichem Aussehen., Rubiaceae. NVertera depressa. Wird nach der Literatur zu urteilen als anemophil betrachtet. — Coprosma triflorum und pyrifolium. NNahrscheinlich sind alle Coprosma-Arten anemophil, die unsrigen jedenfalls sehr ausgepragt, was die in » The Phanerogams» 173 mitgeteilten Abbildungen wohl zur Geniige zeigen. Der Pollen staubt bei leisester Erschiitterung der Zweige. Entomophile Arten., Iridaceae. Libertia formosa. Die Bliitenstande sind mit ihren zahlreichen, bis 25 mm messenden, milchweissen Bliiten sehr augenfallig. In der Knospe stehen die Griffel im Zentrum zusammen, biegen sich spater in scharfem Win- kel zwischen den Staubfaden heraus, so dass die Antheren von den Narben entfernt werden. Nektar wird von dem verdickten Nagel der inneren Tepalen ausgeschieden. Als Besucher wurden Pyramezs carye und einige Fliegen beob- achtet. (Santalaceae. Santalum fernandezianum. Ausgestorben. Bliiten nach JoHow (Flora S. 132) klein und weiss.) Aizoaceae. Tetragonia expansa. Die kleinen gelben Bliiten fihren Nektar und werden vielleicht von Fliegen besucht. Caryophyllaceae. Spergularia confertiflora mit den Varietaten polyphylla POLLINATIONSBIOLOGIE UND SAMENVERBREITUNG AUF DEN JUAN FERNANDEZ-INSELN 507 und glaverrima, die vielleicht ebensogut als Arten aufgefasst werden k6énnen. Bliiten weiss oder zart rosa, etwa 8 mm gross. In der Kultur autogam mit normaler Samenproduktion. — Paronychia chilensis. Die winzig kleinen, apeta- len und versteckt sitzenden Bliten sind wahrscheinlich autogam. Ranunculaceae. Ranunculus caprarum. Nur mit Friichten bekannt. Von der Grosse des Rezeptakels und der Achanien zu urteilen dirfte diese statt- liche Spezies grossbliitig sein. Die Farbe ist natirlich unbekannt geblieben. Berberidaceae. Aerberts corymbosa. Die zwei ausseren Kreise sind griin, die beiden inneren orangegelb mit rotlichen Nektarien, und die Bliiten auch sonst typisch gebaut. In allen noch geschlossenen, aber sonst voll ausgebilde- ten Bliiten, die ich untersuchte, waren die Staubbeutel reif, einige hatten schon ihre Bewegung nach innen ausgefihrt, sind also gereizt worden, und Pollen war am Rand des Narbenkopfes abgelagert. Kleine Kafer, die in den offenen Bliiten umher- kriechen, dringen wahrscheinlich auch in die geschlossene Bliite hinein. In biologischen Werken, z. B. bei KIRCHNER, S. 136—138, wurde die Be- staubung von Serderts geschildert. Als Keimflache fiir den Pollen wird die ober- flache der grossen Narbenscheibe, » die eigent- liche Narbe», angegeben; ihr Rand soll mit steifen, klebrigen Haaren besetzt sein und hier sollen die Pollenkérner nicht kei- men kodnnen, so dass Autogamie verhindert wird. Bei B. corymbosa ist die ganze Ober- Fig. 3—4. Berberis corymbosa. flache der Narbe papillds, und die Papillen 3. Langsschnitt durch das Gynaceum, SOLS. : . ‘ 4. Langsschnitt durch die papillése Zone setzen sich langs der Wand des offenen, am Seitenrand des Griffelkopfes, x 180. engen Griffelkanals bis zur Ovarialhohlung fort (Fig. 3). Die marginalen Papillen (Fig. 4) sind bedeutend langer, sonst aber den iibrigen gleich und kaum als »steife Haare» zu bezeichnen. Ihre Funktion ist mir nicht klar geworden. Wenn beim Abladen des Pollens der eigenen Beutel alles von dieser Randzone aufgefangen wird und hier nicht keimen kann, so wiirde man wohl diese Zone als Schutz gegen Autogamie auffassen konnen. Bei 2. corymbosa wird aber auch etwas von dem eigenen Pol- len oberhalb der Marginalzone abgestreift. — 4. masafuerana. Bliiten unbekannt. Magnoliaceae. Drimys Winteri var. confertifolia gehort zu den augen- falligsten und haufigsten Baumen. Die doldenahnlichen Bliitenstande stehen angehauft an den Zweigenden, sind lang gestielt und von den Laubblattern umgeben und bilden ziemlich dichte, von weitem sichtbare Strausse. Der Durch- messer der Einzelbliite betragt etwa 2 cm. JOHOW’s Angabe, dass Drzmys eine »Mmuy escasa produccion de flores» hat (l.c. S. 255) kann ich nicht bestatigen. Der Bliitenbau stimmt im allgemeinen mit dem von mir fir eine kultivierte Form beschriebenen (Feuerl. Bliiten S. 45). iiberein. Staubblatter und Griffel bilden einen erst kompakten, spater durch das Auseinanderspreizen der Staub- blatter gelockerten Kegel. Ich zahlte 8 oder 9 Tepalen, 32 bis 37 Staubblat- 508 CARL SKOTTSBERG ter, aber nur 5 bis 6 (sogar 4) Griffel. Autogamie ist wegen Protogynie, Stellung der Narbe an der Innenseite des Griffels und Bewegung der Staub- blatter.erschwert und kaum zu erwarten. Dyimys hat nektarlose Pollenblumen. Besucher wurden nicht beobachtet. Reife Friichte wurden nur in geringer Zahl gefunden. Lactoridaceae. Lactoris fernandeztana. Dieser seltene Strauch kommt im feuchten, schattigen Wald der steilen Abhange vor. Die dicht zweizeilig beblatterten Zweige breiten sich schirmartig aus, sind weich und leicht beweg- lich. Die Hohe des Strauches ist selten mehr als #/; bis 1 m. Von oben sieht man von den Bliten gewohnlich gar nichts, denn die axillaren, 2- bis 4-bliitigen oder auf eine Einzelbliite reduzierten Monochasien werden von den Blattern Fig. 5—10. Lactoris fernandeziana. 5. Weibliche Blite. 6. Zwitterblite, verbliiht. 7. Junge Zwitterblite mit nur einem Staub- blattkreis (der innere abortiert). 8. Pollentetraden. 9—10. Staubblatt, Innen- und Aussenseite. 5—7, 9 u. 10 X 15, 8 X 180. verdeckt. Die Bliiten hangen auf kurzem, 3—4 mm langem, haarfeinem Stiel. Die Tepalen sind hautig und graugriin, die sitzenden Staubbeutel braunlich mit gelben Thezen. In der Mitte erheben sich die drei griinen balgabnlichen Fruchtblatter mit ziemlich grossen braungriinen Narben. Die von mir unter- suchten Exemplare waren alle gynomond6zisch, und zwar sind von den Bliiten eines Monochasiums gewohlich eine der sekundaren Bliiten rein weiblich (fig. 5) die anderen (fig. 6) zwitterig. Die Staubblatter der Q Bliite sind nur als win- zige Hocker an der Ansatzflache angedeutet. Es kommt vor, dass in der Zwitterbliite der innere Staubblattkreis unterdriickt ist; einen solchen Fall stellt Fig. 7 dar. HEMSLEY (Taf. 59) bildet auch eine mannliche Bliite ab, die der Zwitterbliite sehr ahnlich aussieht, aber kleinere Fruchtblatter hat. Ich kenne seule. «ow POLLINATIONSBIOLOGIE UND SAMENVERBREITUNG AUF DEN JUAN FERNANDEZ-INSELN 5°9 aus eigener Erfahrung keine ¢ Bliten. Die 9 sind etwas kleiner als die Q. Die Pollenkérner bilden bekanntlich Tetraden; diese sind sehr klein, nur 36— 39 uw, ihre Oberflache glatt (Fig. 8). Wie die Bestaubung stattfindet ist mit nicht bekannt. Dass sie statt- findet, weiss ich, denn ich fand Pollentetraden auf den Narben. Die Sa- menbildung ist sehr gut, ja es scheint dass fast alle Samenanlagen zu Samen werden. Anpassungen an einer bestimmten Bestéubungsart fehlen, und ich kann nicht sagen wie die Tetraden nach den Narben gebracht werden. Auto- gamie scheint ausgeschlossen. Die Bliiten sind protogynisch, Staubbeutel und Narben kommen gar nicht mit einander in Beriihrung, und fallt Pollen heraus, so trifft er nicht die nach innen gerichteten Narben. Als Bestauber wiirde man vielleicht winzig kleine Kafer oder Fliegen erwarten. Die Bliiten sind nektarlos. Die Moglichkeit von Windbestaubung ist nicht ganz von der Hand zu weisen, aber fast alle Merkmale, die wir sonst bei anemophilen Arten fin- den, fehlen; héchstens kann man anfihren, dass die Narbenpapillen ziemlich lang sind und der Pollen klein und glatt. Die Antheren (Fig. 9, 10) sitzen ohne Faden hinter den Tepalen halb versteckt, die Bliitenstiele sind zwar diinn, aber sehr kurz, und die Bliiten nehmen ja eine versteckte Lage ein, die man kaum bei einer anemophilen Pflanze erwartet. Die Bliiten gehoren einem ziem- lich primitiven Typus. Die reichliche Samenproduktion ist eher iiberraschend. Ob Lactorzs méglicherweise apogam ist, wissen wir noch nicht. Cruciferae. Drei Cardamzne-Arten mit kleinen weissen Bliiten. Nicht un- tersucht. Saxifragaceae. Lscallonta Callcottiae, ein haufiger Strauch oder kleiner Baum mit auffalligen, angenehm duftenden Bliitenrispen. Kelch braunlich kar- min, Kronblatter schon karmin, 14 bis 15 mm lang, Faden und Griffel rosa, Beutel goldgelb, Narbe grasgriin. Schon vor dem Offnen der Krone entleeren die Beutel ihren Staub am Rand der breiten Narbenscheibe, die aber noch nicht bestaubungsfahig ist; es herrscht also Protandrie. Erst spater erreicht die Narbe ihre volle Entwickelung, und es sammelt sich in der zentralen Grube reichlich Flissigkeit. Die langen festen Nagel der Kronblatter bilden eine 8—1o mm lange und 3 mm weite Rohre, deren Eingang von Beuteln und Narbe gesperrt wird. Honig wird tropfenweise von dem fleischigen Wulst des Bliitenbodens erzeugt. £. Callcottiae hat den Charakter einer Tagfalterblume und wird auch fleissig von Pyramezs carye und Syngrapha gammoides besucht und bestaubt. KNUTH (III: 1, S. 330) wiedergibt eine Angabe von LUDWIG (Bot. Centralbl. Bd. 69, 1897), dass JOHOW in seiner Flora Kolibris als Bestauber festgestellt hat. Das kann ich aber nicht finden. JOHOW schreibt |. c. 254: »En efecto, no puede negarse que en varias plantas del archipiélago, v. g. en el Rhaphi- thamnus, la Escallonia i la Myrceugenia fernandeziana, las flores tienen una posicion i estructura tales que las referidas aves pueden, al buscar insectos en ellas, contribuir a su fecundacion, i es cierto que los picaflores abundan, sobre todo en Masatierra, de una manera extraordinaria». Damit hat er nicht ge- sagt, dass er Besuche von Kolibris beobachtet hat, LUDWIG hat diese Stelle unrichtig iibersetzt. Nun ist aber Rhaphithamnus typisch, und Lscallonta wahr- scheinlich fakultativ ornithophil, wie ich unten naher erdrtern werde. Leider 510 CARL SKOTTSBERG enthalten meine Aufzeichnungen keine Angabe ob die Nektarproduktion stark gesteigert werden kann, so dass die Rohre grossere Mengen von Fliissigkeit enthalt, wie man es bei einer Vogelblume gerne findet. Die Kronblatter sind nicht papillo6s, bei starker Vergr6sserung zeigt aber die Kutikula von der Epi- dermis des Nagels eine dichte Langsstreifung, die durch winzige Leisten und Falten hervorgerufen wird, welche gegen die Spreite zu unregelmassiger ver- laufen, bis sie endlich, wo die Zellen + isodiametrisch werden, ein Gewirr von Schlangenlinien darstellen. Das erinnert an die als Anpassung zur Steigerung der Adhasion gedeutete Struktur, welche PORSCH beschreibt. Aber bei /sca/- fonia findet man diese Langsstreifung auch an der Epidermis der Aussenseite. Rosaceae. Rubus geoides. Vegi. Feuerl. Bliiten 40. Auf Masafuera machte ich keine Beobachtungen uber die bliitenbiologischen Verhaltnisse bei dieser Art. Fig. r1—12. Colletia sfhartioides. 11. Blite auf mannlichem, 12 auf Zwitterstadium. x 6. Rutaceae. Fagara mayu. Didzischer Waldbaum. Die kleinen, wenig auffalligen Bliitten sind zu ziemlich grossen Rispen gesammelt, ich habe aber nur von den mannlichen lebendes Material gehabt, die weiblichen sah ich nur in Frucht- und sehr friihem Knospenstadium. Die braunlichen oder braun- gelben Perigonblatter und die Entwickelung der Blumen sehr zeitig im Frihjahr oder Spatwinter lassen vielleicht Anemophilie in Frage kommen. Wir finden aber, besonders deutlich bei ¢, einen wulstigen Diskus, und Entomophilie auf einer niedrigen Stufe wird ftir andere, ahnlich gebaute Rutaceenbliiten ange- nommen. Rhamnaceae. Colletia spartiotdes. Die Bliten sind von Gay (II, 34) be- schrieben worden; offenbar hat er aber keine ganz entwickelten gefunden, denn er glaubt, dass die Kelchzipfel stets zusammenneigen. In der Natur fand ich nur sterile oder fruchttragende Stdcke. Ich habe aber diese Art seit Jahren in Kultur, und ein Strauch hat im Oktober 1927 zum ersten Mal gebliiht. Die kleinen einzelnen oder ein paar beisammen sitzenden Bliiten sind + hangend. Ihre Form geht aus Fig. 11 und 12 hervor; sie sind finf- oder seltener vier- POLLINATIONSBIOLOGIE UND SAMENVERBREITUNG AUF DEN JUAN FERNANDEZ-INSELN 511 zahlig. Der Tubus ist 4 bis 5 mm hoch, die zuriickgeschlagenen Kelchzipfel etwa 2 mm lang. Die Grundfarbe ist. weiss, mehr oder minder kraftig mit rosa angestrichen. Die Geschlechtsteile ragen aus, sind etwa gleich lang oder der Griffel meistens ein wenig langer als die Staubblatter. Da die Bliiten homogam sind, tritt Selbstbestaubung leicht ein, wenigstens war dies bei der kultivierten Pflanze der Fall. Friichte wurden nicht gebildet. In der Natur werden wohl Falter die Bliitten besuchen. Der fiir die Gattung charakteristische Diskusring (in den Figuren als dunklere Linie sichtbar) sondert in der durch ihn gebildeten Rinne reichlich Nektar ab. Flacourtiaceae. Azara fernandeztana. Zweige dorsiventral, flach ausge- breitet, die Bliitendolden an den Sprossoberseiten exponiert und sehr in die Augen fallend. Die gelbgriinen Perigonblatter spielen dabei keine Rolle gegen die zahlreichen, dicht stehenden goldgelben Staubblatter, aus deren Mitte der einsame Griffel aufragt. Die angenehm duftende Bliite ist homogam, aber An- theren und Narbe kommen wohl kaum mit einander in direkter Beriihrung. Nach REICHE, (Rio Maule, 39) wird A. celastrifolia von der Honigbiene eifrig gesucht, in seiner Arbeit iiber die Gattung Azara, 504, erzahlt er, dass vor jedem Kelchblatt eine winzig kleine Schuppe steht, dass er aber keine unzweideutige Nektarproduktion beobachten konnte. Die reichliche Behaarung des Bliitenbodens spricht eher gegen Ausscheidung von Nektar, und REICHE halt es ftir sehr wahrscheinlich, dass die Bienen nur Pollen sammeln. Wie man es verstehen soll, dass er in einer anderen Arbeit (Grundziige, 149) unter den »sehr honigreichen Bliiten» auch Azara aufzahlt, weiss ich nicht. Bei A. fer- nandestana habe ich jedenfalls keinen Honig entdecken kénnen, halte vielmehr die Bliite fiir eine recht typische Pollenblume. Die Bliitenachse ist mit steifen Haaren dicht bedeckt. Die schuppenformigen Gebilde habe ich vergebens ge- sucht, dagegen findet man zuweilen kleine Staminodien. Besucher wurden nicht wahrgenommen. Myrtaceae. Ugni Selkirkiz. Bliiten + hangend, etwa 8 mm gross, weiss mit farbigen Staubblattern, die Faden karminrot, die Beutel rétlich gelb, beide etwa I.2 mm lang. Schon in der noch geschlossenen Bliite fangen die An- theren an, sich zu entleeren, die Narbe ist zu dieser Zeit noch nicht reif. Nektar wird zwischen den Staubblattern und dem Griffel erzeugt. Die eben gedffnete Bliite has einen an Platanthera erinnernden, sehr angenehmen Ge- ruch, welcher aber spater verloren geht. Keine Besucher wurden wahrge- nommen, dagegen beobachteten wir Falter als Besucher der hier nur verwil- derten Art U. Molinae. — Myrteola nummularia. Ganz kleine, weisse Bliiten, die nicht untersucht wurden. — J/yrceugenia fernandeziana. Dieser haufigste Waldbaum von Masatierra ist mit seinen dichten weissen Bliitenrispen eine auffallige Erscheinung. Wie bei so vielen anderen Myrtaceen sind die Blu- menblatter als Schauapparate ohne Bedeutung; ihre Rolle wurde von den 60 bis 70 Staubblattern mit ihren weissen Faden und rétlichen Beuteln iiber- nommen. Die Bliite ist stark protogynisch, der Griffel ragt schon aus der Knospe heraus, wahrend samtliche Staubblatter noch eingeknickt stehen. Sie richten sich auf und biegen aus, bis der Durchmesser des Adroeceums etwa 14 mm ist. Die Narbe ist inzwischen gewelkt. Besucher haben wir nie gesehen. Der Bliitenboden ist zottig, Honig konnte ich nicht entdecken, so dass wohl diese 512 CARL SKOTTSBERG Mirceugenta Pollenblumen hat. Andere chilenische Arten werden nach JoHow (Bestaubungsbiol. II. S. 38) von Hummeln und Bienen gesucht und bei J/. obtusa fand er Honig. KNUTH’s Behauptung dass die IZyrceugenta-Arten von Juan Fernandez durch Trochiliden bestaubt werden geht auf das oben erwahnte Referat von LuDWwiIG zuriick. — Myrceugenta Schulzet, der Waldbaum von Masafuera, ist biologisch etwas verschieden von der anderen Inselart. Die Bliiten sind bis 15 mm gross und ganz weiss, indem auch die Beutel weisslich sind. Laut Beobachtungen an Ort und Stelle sind die Bliiten dieser Spezies eher protandrisch. Die zahlreichen (bis 110 wurden gezahlt) 8—10 mm langen Staubblatter biegen sich nach answarts und lassen den 7 mm langen Griffel Fig. 13—14. Eryngium bupleuroides. 13. Bliite auf 9 Stadium. 14. Blite auf % Stadium, den grossen, kapillar festgehaltenen Honigtropfen zeigend. 8. frei. Trotzdem dass sie so viel zahlreicher sind, nehmen sie auf dem Bliten- boden eine schmalere Zone ein, so dass ein breiterer, wenig behaarter bis kahler Diskus entsteht. Nektar konnte ich aber nicht konstatieren, dagegen sind oberflachliche Oldriisen vorhanden. Umbelliferae. Lryngium bupleurordes. Die bis 2 cm breiten Kopfe sitzen auf langen, festen Stielen und fallen trotz ihren gelblich- und braunlich-griinen Farben auf. Ausgepragte Protogynie herrscht (Fig. 13) und Selbstbestaubung kann nur ausnahmsweise stattfinden. Honig wird so reichlich erzeugt, dass er einen grossen, gewolbten Tropfen, der zwischen den langen Staubfaden kapillar festgehalten wird (Fig. 14), bildet. Als Besucher haben sowohl JOHOW als der Verfasser Fliegen beobachtet, und zwar in grosser Anzahl. Gelegentlich ist Eryngium auch ornithophil; s. unten. — £. zvaccessum hat grossere Kopfe, bis 3.7 cm breit, durch ihre dunkelviolette, fast schwarzliche Farbe auffallend. Ich fand aber nur fast verbliihte Pflanzen. — £. sarcophyllum von Masafuera hat POLLINATIONSBIOLOGIE UND SAMENVERBREITUNG AUF DEN JUAN FERNANDEZ-INSELN 513 weisslich griine, nur I cm grosse Képfe. Nicht naher untersucht. — Ajpzum fernandezianum. Bliiten wie bei anderen Arten dieser Gattung. Nicht un- tersucht. Ericaceae. Pernettya rigida. Die gegen die Sprossspitzen gesammelten, achselstandigen Bliitenglocken sind + einseitig gewendet und werden dadurch recht auffallend. In ihrem Bau stimmen sie mit anderen Pernettya-Bliiten uber- ein (vgl. z. B. Feuerl. Bliiten, 37 und 41). Gegen die weisse Farbe der Krone heben sich die braunen Staubbeutel und die karminrote Narbe ab. Die Augen- falligkeit wird bei der eigentiimlichen, von A. DE CANDOLLE als Art aufgefassten P. racemulosa (Prodr. VII, S. 588), die von HEMSLEY, JOHOW etc. zu P. r7gzda gestellt wird, dadurch gesteigert, dass statt Einzelbliiten mehrblutige Trauben in den Blattachseln sitzen.. Sonst hat man keinen Unterschied zwischen den beiden Formen finden kénnen. Beide sind gynodidzisch. Die Zwitterbliten sind 5.5—7 mm lang, ihre Staubblatter messen 2.5—3.5, im allgemeinen wenig- stens 3 mm, der Griffel 1.s—2.5 mm. Die weiblichen Bliiten sind durchschnitt- lich etwas kleiner, 4.5—6 mm, ihre Staubblatter 1.6—2 mm; sie sind fast immer steril oder es kommen Ubergangsformen vor, welche einige entwickelte Pollen- k6érner enthalten. Der Griffel ist 2—2.4 mm. lang. Als Besucher haben wir wiederholt Pyramets carye und Syngrapha gam- moides beobachtet. 33—2610. The Nat. Hist. of Juan Fernandez and Easter Isl. Vol. II. 514 CARL SKOTTSBERG Convolvulaceae. Dzchondra repens (vielleicht eingeschleppt; winzige griin- liche Blumen) und Calystegia tugurtorum (weiss, grossbliitig) habe ich nicht untersuchen konnen. Boraginaceae. Se/kirkia Lerterot. Sehr kleiner Schopfbaum mit weissen Bliiten. Die Bestéaubungsverhaltnisse sind mir nicht bekannt. Fig. 16—22. Cuminia fernandezia. 16. Zwitterblite von einer kultivierten Pflanze, X 5. 17. Korolle und 18 Gyndaceum einer Zwitterbliite von einer wilden Pflanze, x 5; 19 dieselbe Blite, Teil der Korolle von innen gesehen, mit 2 Staubblattern, X 15. 20. Verkiimmertes Gyndceum einer Zwitterblite (kultiviertes Ex.), X 15. 21..Weibliche Blite einer kultivierten Pflanze, X 5, und 22 Teil der Korolle von innen, mit 2 Staminodien, X I5. Labiatae. Cwminia fernandezia und ertantha nehme ich an dieser Stelle auf, trotzdem ich Kolibris (ein einziges Mal) aber keine Insekten als Besucher beobachtete. Ich muss diese Arten wegen der Struktur der Bliite, die ja ganz wie bei einer langrdhrigen Hummel- oder Falterblume ist, als nur gelegentlich ornithophil auffassen. Beide sind kleine Baume mit achselstandigen, ziemlich lockeren Dichasien (s. Fig. 15). In »The Phanerogams», 164—-165 habe ich POLLINATIONSBIOLOGIE UND SAMENVERBREITUNG AUF DEN JUAN FERNANDEZ-INSELN 515 den Bliitenbau kurz geschildert. Unten werden etwas ausfiihrlichere Daten mitgeteilt. Beide Arten sind sexuell dimorph, was ja tibrigens bei den Labiaten eine haufige Erscheinung ist, mit zwitterigen (I. c. mannlich genannt) und weiblichen Bliiten. Letztere sind kurzkronig mit herausragendem Griffel, also leicht zu erkennen. In der Ausbildung des Griffels sind beide Formen gleich, bei der kurzkronigen die Staubbeutel sehr verkiimmert und ohne Pollen. Wie ich l.c. bemerkte, habe ich beobachtet, dass die »mdannlichen» Bliiten friih abfallen; deshalb habe ich sie als steril auf- gefasst. Es kommt aber vor, dass sie Friichte ausbilden, und wberhaupt fallen die Bliiten bei Cwsmznza sehr leicht ab. Gynomondzie ist nach meinen Feldnotizen die Regel, nach derselben Quelle kommt aber auch Gynodidzie vor bei C. eriantha. Ein jetzt (August 1927) im hiesigen Garten bliihendes Baumchen von C. fevnandesia hat zwei bliihende Zweige, einer ist Q, der andere %, aber auch mit Uber- gangstypen. S\ 7 C. fernandezia. & (Fig. 16—19). Gesamtlange der W/Z Bliite 12 bis 15 mm, Lange der Krone I1.s5 bis 14 mm, NY Breite tiber den Saum 4—5 mm, iiber den glockenformigen NZ Teil etwa 3 mm. Unten verengt sich die Krone zu einer 4,5—6.5 mm langen und I—I.z2 mm breiten Rodhre. Die Farbe dieses untersten Teiles ist weiss, sonst sch6n blau- violett oder rotlichviolett. Die Stellung der weissen Staub- Ni blatter geht aus Fig. 19 hervor; ihre Beutel, die sich \ vor der Narbenreife offnen, sind etwa I mm gross. Das Gynaceum, weiss mit hellvioletten Griffelzweigen, misst 7 bis g mm. Bei der eben erwahnten kultivierten Pflanze fand ich eine dusserlich normale Zwitterbliite, die ein ver- kiimmertes Gynaceum hatte (Fig. 20). C. fernandesia 8. Gesamtlange der Bliite 7 bis 9.5 PRE Ore ee mm, Korolle 6.;—8, Griffel 7.5—8.5; kleine Staminodien _ dezia. 2 (Fig. 21, 22). Uberginge mit einem oder zwei fertilen eT oF Fae Staubblattern sind nicht selten. haarte Zone zu zeigen. C. eriantha % (Fig. 24, 25). Gesamtlainge der Bliite ee ae 10 bis 14 mm, Krone 9—13 mn, sehr hell lila bis fast : weiss, die enge Rohre 4—5 mm lang. Staubblatter und Griffel wie bei der Vorigen. C. eriantha @ (Fig. 27, 28). Gesamtlainge der Bliite 8—g mm, Krone 7—8.5 mm, die enge Rohre scharf abgesetzt, etwa 3 mm lang. Staminodien sehr klein (in einem Fall wurden grossere, aber sterile Beutel gesehen, Fig. 26), Griffel wie bei ¥. Bei beiden Arten scheidet der kleine fleischige, griine Diskus unterhalb der Klausen Nektar aus, welcher den verengten Teil der Krone bis in die Sperrzone fillt. Der mittlere Teil der Krone ist namlich mit langen, farblosen, steifen und scharfspitzigen, ein- bis zwei-, selten dreizelligen Haaren bekleidet (Fig. 23), im Zentrum einen engen Kanal ftr den Griffel frei lassend. Die as EEZA SS S> = Sos 516 CARL SKOTTSBERG Haare sind nach innen gerichtet, sperren wohl trotzdem den Zutritt fiir kleine kriechende Insekten, k6nnen aber das Einftihren eines Insektenriissels nicht ver- hindern. Der Griffel liegt am Saum der Oberseite der Bliite an; als sich die Fig. 24—29. Cuminia eriantha. 24. Krone einer Zwitterbliite, X 5 und 25 Teil der Krone von innen mit 3 Staubblattern, x 15. 26. Weibliche Bliite mit herausragendem Griffel und Ubergangsgebilden zwischen Staubblattern und Staminodien, X 15. 27. Krone einer weiblichen Bliite, x 5 und 28 Teil der Krone von innen, mit 2 Staminodien, X 15. 29. Petaloides Staminodium einer sonst normalen, viermannigen Zwitterblite, % 15. Staubbeutel 6ffnen, sind die Narbenzweige noch an einander gepresst, weichen aber bald auseinander. Der untere ist etwas linger. Zum Schluss mag hier erwahnt werden, dass ich einmal in einer Zwit- terbliite von C. evzantha an der Insertionsstelle des sonst stets ganzlich fehlenden fiinften Staubblattes eine Bildung fand, die man wohl am ehesten als petaloides Staminodium deuten wird (Fig. 209). POLLINATIONSBIOLOGIE UND SAMENVERBREITUNG AUF DEN JUAN FERNANDEZ-INSELN 5.7 Solanaceae. Solanum fernandezianum ist habituell und auch im Bau der Blitte S. tberosum ziemlich ahnlich und wie dieses nektarlos. Die lebhaft violette, 15—16 mm breite Krone ist recht augenfallig, und der goldgelbe Antherenkegel hebt sich scharf ab. Schon vor der vollstandigen Entfaltung der Krone sind die Spitzlocher der Antheren ge6ffnet, und Pollen tritt beim Erschiitterung heraus; die Bliite ist homogam und Selbstbestaubung muss, weil die Narbe kaum herausragt, leicht eintreten konnen (Fig. 30). Die Bestaubung bei S. ¢uberosum und verwandten Arten wurde neuerdings von E, WERTH eingehend untersucht. Formen mit deutlicher Zygomorphie, lang herausragender Narbe und Heterantherie zeigen eine Steigerung der Anpassungsmerkmale, wahrend andere, darunter .S. Zde- ie rosum selbst, auf dem Wege zur OT gM a hese Vervollkommung des_ Bestaub- Oe pen Pee upgsmechanismus __ »phylogene- 5 tisch gedacht (einstweilen) stehen geblieben sind» (WERTH, Kar- toffelbl. 149). S. fernandeztanum steht in dieser Hinsicht noch ein wenig tiefer durch die Stellung der Narbe auf gleichem Niveau mit den Antherenlochern. Zygo- morphie ist aber angedeutet durch die etwas grossere Lange des untersten Staubfadens (2 mm, die wubrigen 1.75 mm; alle An- theren messen 3 mm) und durch | ‘ die S-formige Kriimmung des Fig. pond: Solanum Sernandezianum. ; ‘ 2 j $ 30. Geschlechtsteile von der Seite gesehen. 31. Gyna- Griffels, die manchmal ziemlich ceum mit stark gekriimmtem Griffel. x 10. stark ist (Fig. 31), ferner durch die nicht ganz gleichformige Verteilung der Behaarung des Griffels. Besucher haben wir nie wahrgenommen, der Samenansatz ist aber sehr reichlich. Ich habe diese Art in Kultur gehabt. Ohne Fremdbestéubung wurden Samen gebildet. — S. Kobinsontanum und masafueranum sind zwei Morellen mit Bliten von dem Typus des S. xzgrum. Sie wurden nicht untersucht. Scrophulariaceae. M7csmulus parviflorus var. externus. Kleine gelbe, wenig auffallige Blumen, Krone 10—12 mm lang, kaum zweimal so lang als der Kelch. Nicht untersucht; keine Besucher beobachtet. — Euphrasia formosissima. Farbige Abbildungen finden sich auf Taf. 20 in »The Phanerogams». Der Bliiten- mechanismus ist ganz derselbe wie bei allogamen Formen von £. officinalis und Verwandten. Besucher habe ich nicht beobachtet. Rubiaceae. Oldenlandia thesiifolia. ‘Kleine hellviolette, heterostylische Bliiten. Nicht untersucht. — Galum masafueranum. Winzige gelbgriine Bliiten wie bei so vielen anderen Arten. Nicht untersucht. Campanulaceae. — Wahlenbergia Larrainit, fernandeziana, Grahamae und Masafuerae. Typisch und ausgepragt protandrisch mit Abladen des Pollens auf den oberen, verdickten und behaarten Teil des Griffels. Alle diese Arten sind weissbliitig und gehéren zu den auffalligeren Erscheinungen der Flora 518 CARL SKOTTSBERG (farbige Abbildungen |. c. Taf. 20); besonders gilt dies W. Grahamae. Taf. 15, Fig. 2 zeigt die vorziigliche Exposition ihrer weiten Glocken. Bei dieser Art wurde als Besucher Pyramezs carye beobachtet. W. Berterot weicht sowohl durch ihre schoOn karminrote Farbe (1. c. Taf. 20, Fig. 8) als durch ihre Blitengestalt von den anderen Arten erheblich ab, indem die Krone sich gegen die Mtindung stark verengt und hier nur 2.;—3 mm weit ist. Der Griffel, welcher bei den anderen Arten bedeutend kiirzer als die Krone ist, ist bei W. Lerterot in zweckmassiger Weise so stark verlangert, dass die Narben in der Miindung stehen, wo sie mit dem Kopf eines besuchenden Falters unbedingt in Kontakt kommen. Welche die Bestauber sind, ist mir unbekannt geblieben. Sie mussen eine Rissellange von 10—12 mm haben. Die anderen Arten k6nnen von kurzriisseligeren Insekten ausgebeutet werden. In jeder der fiinf episepalen Gruben im Bliitenboden trifft man einen grossen Nektartropfen. Lobeliaceae. Lobelia anceps. Bliten tief blauviolett bis fast weiss. In Chile nach REICHE (Best. chil. Camp.) haufig autogam. Compositae. Lagenophora Harioti. Die Farbe der Randbliiten wechselt von weiss bis rosa und wird mit zunehmendem Alter tiefer rot. Sonst nicht untersucht. —- /vzgeron. Keine der fiinf Arten zeigt in ihrem Blitenbau etwas Bemerkungswertes. Bei allen sind die Randbliiten weiss, die Scheiben- bliiten gelb. Nahere Beschreibung und Abbildungen in » The Phanerogams», 182— 187. — Guaphalium spiciforme. Nicht untersucht. -— Adbrotanella crassipes. Die winzig kleinen Kopfe sitzen versteckt in den Blattrosetten. Die Bliten sind braunlichrot. Weiteres |. c. 189, mit Abbildungen. Robinsonia. Diese Gattung von kleinen Schopfbaumen ist insofern be- merkenswert, dass samtliche Arten streng didzisch sind. Bei 2. macrocephala sind die ¢ Bliiten noch nicht bekannt; die 2 haben schwarzrote Ligulae und gelbe Scheibenbliiten mit rétlichen Zipfeln. Bei R. Gayana sind die ¢ gelb, die 2 gelbgriin, bei RX. thurifera die & schmutzig gelb (Randbl.) und gelbgriin (Scheibenbl.), die Q gelbgriin, bei R. eventa die ¢ goldgelb (Randbl.) und gelb- griin (Scheibenbl.), die 2 gelb, bei R. Masafuerae S und 2 gelb (Randbl.) und gelbgriin (Scheibenbl.) und bei R. gracilis die & gelb, die & gelb (Randbl.) und gelbgriin (Scheibenbl.). Die ziemlich kleinen Kopfe sind zu reichen Bliiten- standen vereinigt, die oberhalb des Blattschopfes exponiert werden und auch durch Honigduft ausgezeichnet sind. Fremdbestaubung ist erforderlich und findet wohl auch haufig statt, Friichte werden wenigstens bei 2. macrocephala und 2. Gayana reichlich gebildet, dagegen fand ich bei 2. M/asafuerae, evenia und gracilis die meisten Achanien taub. JoHOW (Flora 255) hat Dipteren als Besucher beobachtet. In »The Phanerogams», 194 und 197 wurden samtliche Arten abgebildet. Rhetinodendron Berterzi. WRispen gross, aus mehreren hundert Kopfen bestehend, wie bei Aodznsonza terminal am Ende der Zweige, aber wegen der triiberen Farbe nicht sehr auffallig, Auch Ahetinodendron ist didzisch. Kopfe wurden |. c. 198 abgebildet (2 etwas grodsser als gewéhnlich, ¢ eher etwas klein), Aussenhiille aus wenigen, 1—2 mm grossen Schuppen, Innenhiille eng zylindrisch, stets mit nur drei Blattern, 5—6 mm hoch und 2 mm breit. Vor jedem Blatt sitzt eine Bliite, das Kopfchen ist also stets dreibliitig. Die mann- POLLINATIONSBIOLOGIE UND SAMENVERBREITUNG AUF DEN JUAN FERNANDEZ INSELN 519 liche Bliite (Fig. 32) ist gelbgriin, hat eine 4 mm lange Rohre mit 2.5 mm langen, 0,6 mm breiten, zuriickgerollten Zipfeln. Aus der Miindung ragt der griinlichgelbe Staubblatttubus 3 mm heraus. Das Konnektiv ist braunviolett. Die Griffelzweige (Fig. 33) wachsen hier nicht aus der Spitze des Tubus hervor, sondern werden durch die seitliche Aufschlitzung der Antherenrohre frei. Manch- mal bleibt dabei Pollen in der offenen Rohre liegen. Das untere Ende des Griffels ist leicht angeschwollen und sondert Nektar ab. Die weibliche Bliite (Fig. 34—36) ist nur etwa 5 mm lang, die Zipfel kurz, gerade und leicht gelb- Seo) = ESS =~ —— Yo D ae LIGA ESS Fig. 32—36. Rhetinodendron Ber terit. 32. Mannliche Blitte, X 10. 33. Griffel derselben, X 20. 34. Weibliche Bliite, x 10. 35. Sta- minodien und 36 Griffelbasis derselben, < 20. lich getont, die Bliite sonst rein griin. Die Honigduft ist bei beiden Geschlechtern deutlich, Besucher habe ich aber nie gesehen. Vielleicht kommt hier auch eine Ubertragung des Pollens durch den Wind in Frage. Darauf konnten die lang herausragende, leicht bewegliche Antherenrohre und die hervorgestreckten, grossen Narben der ¢ Bliite deuten, ferner die reichbliitigen, leicht bewegten Rispen und die dusserst bescheidene Farbung der Kopfchen. Centaurodendron dracaenoides. Die Bliiten dieses sehr seltenen Schopf- baumes sind noch unbekannt. Dendroseris. Diese Schopfbaumgattung zerfallt von bliitenbiologischem Gesichtspunkt in zwei Gruppen, eine die Eudendroseris-Arten, die andere die Phoenicoseris- und Rea-Arten umfassend. Der Bliitenstand ist bei allen terminal. 520 CARL SKOTTSBERG Eudendroseris. Kopfe nicht sehr zahlreich, wohl selten iiber 20, dafiir aber um so grosser, bis 5 cm breit und leuchtend orangegelb (The Phanerogams, Taf. 20, Fig. 11, 12). Die Achse ist nicht steif genug um die Képfe iiber das Niveau der Blattschopfe emporzuheben (vgl. |. c., Taf 18, 19). Hierher gehéren D. macrophylla, macrantha, marginata und /itoralis, alle von sehr beschrankter Verbreitung. Keine Besucher sind beobachtet worden. Ausfthrliche Beschreibung des Bliitenbaus 1. c, 201I—204. — Phoenicoseris und Rea. Kopfchen kleiner, Fig, 37. Blitenstand von Dendroseris micrantha auf Masatierra, nach toy) Photographie des Verf. 10 bis 20 mm breit, aber sehr zahlreich bis viele hundert, bei DY. »zeantea, wo der Bliitenstand tiber 1 m hoch sein kann, noch zahlreicher. Die Bliitenstande stehen bei allen aufrecht und vorziiglich exponiert, die Einzelképfe sind aber mehr oder weniger hangend, wie Fig. 37 zeigt. Die Bliitenfarbe ist bei allen weisslich (rein weiss bis leicht gelblich). Hierher gehdren die hapaxanthischen D. pinnata und regia (bei dieser die Bliiten nicht bekannt) und die polla- kanthischen D. neriifolia, micrantha, pruinata und gigantea. Nur D. micrantha ist haufig; bei dieser wurden Dipteren als Besucher festgestellt. Die Blumen duften schwach nach Nektar und starker nach frischen Schnittbohnen. POLLINATIONSBIOLOGIE UND SAMENVERBREITUNG AUF DEN JUAN FERNANDEZ-INSELN 521 Farbe und Augenfilligkeit der Blumen bei den entomophilen Arten. Unten habe ich die entomophilen Arten nach Blittenfarbe und Augen- falligkeit zusammengestellt. Bliitenstand, je nach den Verhaltnissen. benutzt, wo I unscheinbar bis wenig auffallend, 2 auffallend bedeutet. Die Verbreitung wurde auch Augenfialligkeit to iS) i) Se) Weiss Libertia forniosa*. .. . Santalum fernandezianum . Spergularia confertiflora Drimys Winteri var. . . Cardamine chenopodiifolia » flaccida .. » Kiwsselit= str Rubus pedides . |. < awa . Colletia spartioides. . .. Wismig Selkirkive ss. < Myrteola nummularia. . . Myrceugenia fernandeziana » Sonulzeieem sr Apium fernandezianum . . Pernettya rigida ... Calystegia tuguriorum . . Selkirkiay Bertenoiys se.) Solanum Robinsonianum . » masafueranum Euphrasia formosissima . . Wahlenbergia Larrainii . . » fernandeziana » Grahamae . » Masafuerae . Lagenophora Harioti . . Erigeron fruticosus. ... » luteoviridis . » Ingae? ys iAweree » turricola » rupicolat) FS f Gnaphalium spiciforme . . Dendroseris pinnata ... » regia = > neriifolia .. » micrantha . » pruinata... » gigantea . . |. Letztere bezieht sich auf die Einzelbliite oder den Es wurde eine dreigradige Skala massig auffallend und 3 sehr mitgeteilt. Masatierra Masafuera ive + 4- + +) ee + + + + . + oo ie a ee + a . = ale = oS - . + ene + + a8 4 oie + hse ae a5 . 2s sie a = . =e Aric _ ee + - + - + + sat, “ — “+ hs + ae + — - an + ; + . + Oy: PORE cee -F 423 + 522 CARL SKOTTSBERG Augenfalligkeit Masatierra Masafuera Gelb I WietracOniagexpan sauce le aay re ee ee + + 2 Bexbenis icorymbosaape ey, nei arieaien ot. + 2 » inasatueranasieee| sweetie). =f 2 Sophora*femmandeziana ~ ilsae/iiisitie . < == 3 » MMA SALWera tabs. {sats Pein eth ele tte aia 3 Azaratemandezianag lei Anse nae. + Minulits spaivinlonusovar oon wo eee aie + 2 RopinsomiaaGayana. |. a). haces ee 4+ 3 » churiferages« Jncccee oars + 2 » EVENT Aaa ee men aCe oh 2 » Masatnerdewe o 085 tame ia ee + 2 » sracilisg yee td: pane oa 3 Dendroseris macrophyllaene se wee + 3 » MMA EIONE Sud og Gh oboe o os of 3 » mateinaltal Sree 5 Soca ws al 3 » Titoralli sitet cdot ge nme et 8 + Grin (gelbgriin, griinweiss) I Earonyelna chilensisy stark ae ate a th at I Ractonsttermandezianay- « 5 .6. cernuus® | Heleocharis maculosa | Oreobolus obtusangulus Cladium scirpoideum | 7 . . . . | Uncinia brevicaulis POLLINATIONSBIOLOGIE UND SAMENVERBREITUNG AUF DEN JUAN FERNANDEZ-INSELN 541 Monate VIIT) 1) XS | RL | RUD Da | DVeew | Vio |) Vil Mittl, Monatstemp., °C. . . Y253] 13.9 | 13.5 15.4 || 17.7 [AXG!0|| 19/5) 16.9 | V7.2) 15.5) 13.9) 12.7 Niederschlag, mm... . . | 108.0] 67.8| 40.4] 40.1 | 14.0/ 18.6] 31.1 | 59.2 | 76.5 | 201.4] 242.1| 170.7 Uncinia Douglasii ; costata phleoides .... tenuis Carex Banksii . . ) Berteroniana .... ama australisit-.~ses-- re: Greigia Berteroi Ochagavia elegans. .... = oy a == | | Enza sainsularis’ 2° 2". | =m =_ PMREUS IM Bria tus~.... ecvestanwes ee 1 | capillanis*.—. sees | | | IGUCLISAE ee soe | sre Dombeyanus.... | OT . | planifolius }-.0 3. 5 | | ibentiasrormosa 2 5 5. ne Toe Peperomia® Berteroana margaritiftera » Skottsbergii fernandeziana Urtica Masafuerae® » fernandeziana Boehmeria excelsa . Parietaria humilis ® Phrygilanthus Berteroi | Santalum fernandezianum ’ | Chenopodium Sanctae Clarae | > Crusoeanum . nesodendron . Salicornia peruviana. ... Tetragonia expansa 42 CARL SKOTTSBERG OL i Monate VIIL) DX) ea eet XI It ll | IDL5| Wig| aac VII | ie | | | sd Mittl. Monatstemp., °C. . . 12.3] 13.3 3.5 | 15.4| 17.7 | 19.0| 19.5 | 18.9| 17.2 15.51 13\.0|unles7 eo . | | P| | | Niederschlag, mm. . . . . | 108.0] 67.8] 40.4| 40.1 | 14.0 | 18.6 | 31.1 | 5§9.2| 76.5 | 201.4)242.1] 170.7 | | | | | LLL “ | | | Spergularia confertiflora® | Paronychia chilensis . . . Ranunculus caprarum .. .- Berberis corymbosa ...- . masafuerana Drimys Winteri var. Lactoris fernandeziana. . . Cardamine chenopodiifolia : = ap GEE eo o> | flaccida® | | Krusselii! /-s=-scer. | Ge | | | Escallona Callcottiae ... | 5 ees ] | Rubus, geoides}™- RENE ee a ee masafuerana. . a Eryngium bupleuroides inaccessum ... UE _ sarcophyllum Apium fernmandezianum . . gy qe —— Pernettya rigid Dichondra repens ..... B® e——-— - Calystegia tuguriorum 2 = Scisrkia Berteroi— +.» +. -- EC Rhaphithamnus venustus'® . OOO —E—E—ES Cuminia fernandezia. . . . a) eniantha.....-o---- sven 288 PS Solanum fernandezianUm . _aaaaaggmppepeeeeeeeseeesrseesrrrprensrpnmererene Robinsonianum § masafueranum . . 2 eE-— — Nicotiana cordifolia .... SETS I rr Mimulus externus ..... 908 eee - - Euphrasia formosissima . . 288 Ee —— - — Plantago fernandezia. .. . SS ——Ee Skottsbergii#. . . ee Oldenlandia thesiifolia. . . ee Nertera depressa‘ Coprosma triflorum ... . | I ee pyrifolium'® .. EEE — Galium masafueranum.. . DY —E Wahlenbergia Larrainii . . — | | oo 544 CARL SKOTTSBERG Monate WAG DOC NSE) S00) a TL 3) SULT 9) Wa ae 9) Vel aie Mittl. Monatstemp., °C. . . 12.3] 13-3 | 13-5 | 15.4| 17-7| 19.0| 19.5 | 18.9] 17.2| 15.5 13.9 12.7 rT. 5 | Ps | / Niederschlag, mm. . . . . | 108.0) 67.8| 40.4| 40.1 | 14.0] 18.6 | 31.1 | 59.2 | 76.5 | 201.4] 242.1) 170.7 Wahlenbergia fernandeziana . Grahamae Masafuerae. . Berteroi Lobelia anceps® Lagenophora Harioti Erigeron fruticosus luteoviridis » Ingae . turricola rupicola Gnaphalium spiciforme . . . Abrotanella crassipes . ... Robinsonia macrocephala . . » Gayanal-mitmienrets » thurificraaeeeeane EVENI1Acseeoe aes Masafuerae » Sracilissa: Rhetinodendron Berterii. . . | ge Centaurodendron dracaenoides | Dm mm ee — — — + | Dendroseris macrophylla . . | = 2 == | | | » MACTAntha wren ——+—= B 8 | | | | | ; | | | | | | marginata ... —— om o | | = | | . . | | | litoralis se che 28 | | | | > PWINATAy ce 4 ce. fe _— —_—_—e a | | | | : | | » regia | | | | ~ e | TLE TOA eee | —_ al Ne = | | imoueerebondney C4 a —__ — j | | Pe pruinata ... ... | [a —___ — — | | . | gigantea .... -.s a = Try POLLINATIONSBIOLOGIE UND SAMENVERBREITUNG AUF DEN JUAN FERNANDEZ-INSELN 545 Anmerkung en. 1) Friichte schon am 17. Febr. (1917). 2) Nur von Puiipps blahend gefunden (November). 3) Mit Bliiten und Friichten wahrend des ganzen Jahres (beob. Aug., Dez.—April). 4) Junge griine Friichte schon Mitte Januar, wahrend der Bliitezeit; sie reifen erst im nachsten Jahr und zwei Generationen werden auf demselben Baum gefunden. Die reifen sind korallrot. 5) Von GERMAIN Okt. 1854 mit Bliten und Friichten gesammelt. 6) Wahrscheinlich ephemar. 7) Nach JoHow’s Angaben. 8) Wahrscheinlich das ganze Jahr bliihend und fruchtend. 9) Blaht im Spatwinter und Frihjahr. Wahrend der Monate December—April gab es nur Friichte; erst Mitte April fand ich die Blitenstande angelegt, die wahrscheinlich im August oder September aufblihen. 10) Im Februar, Marz und August fand ich nur Friichte. Ist im September von GuajARDo blihend gesammelt worden. 11) Bliht im Herbst und Winter (Marz nach Gay). Die Friichte reifen erst im folgenden Sommer (Jan., Febr.). 12) Nach Jonow mit Bliiten um Neujahrzeit. 13) Mit Frichten das ganze Jahr gefunden. 14) Annuell. 15) In der alpinen Re- gion von Masafuera noch Mitte Marz bliihend beobachtet. Das Klima von Juan Fernandez zeigt eine deutliche Periodizitat, nicht sehr auffallend in der Temperatur, aber um so mehr in der Verteilung der Niederschlage. Mai bis August ist Regenzeit, September und April vermitteln den Ubergang zum relativ trockenen Sommer. Eine Sommerdiirre, wie wir sie auf dem gegeniiberliegenden Festland regelmassig haben, existiert auf den Inseln nicht. Im allgemeinen steht die Periodizitat der floralen Phanomene in guter Ubereinstimmung mit dem Gang des Klimas. Die Pflanzen der waldlosen basalen Region entwickeln schon im Spatsommer und Herbst neue Bliitentriebe, bliihen machmal schon wahrend des Winters und tragen im Frihjahr und Som- mer reife Frichte. So verhalten sich Szpa, die Piptochaetium- und Polypogon- Arten, Danthonia, Scirpus nodosus, Libertia, Paronychia, Margyricarpus und Apium, und die meisten sind Birger der mittelchilenischen Flora. Diese Flora beherrscht die niedrigeren westlichen Teile von Masatierra und die benachbarte kleine Insel Santa Clara, Teile, die eine bedeutende Senkung in spater Zeit erlitten, welche sicher ein Trockenwerden des lokalen Klimas zu Folge hatte. Fiir diese Gegend gelten gar nicht die oben angefiihrten Niederschlags- zahlen, die im Kolonietal im Zentrum der Insel Masatierra beobachtet wurden. Charakteristisch fiir dieselbe Region ist Spergularia confertifiora, hier wachsen auch die annuellen Urtica Masafuerae und Plantago *Skottsbergiz, die schon vor dem Anfang des Sommers ihr Leben beendigt haben. Auch die Gunnera- Arten, welche aus deni Waldgebiet in die untere Region hinabsteigen, biiihen sehr friih. Unter den Baumen sind die Sophora-Arten zu erwahnen; sie tragen schon im Herbst Bliitensprosse, die sich aber erst im Friihjahr weiter ent- wickeln, und 4ahnlich verhalt sich Fagarva mayu. Die im April angelegten Bliitenstande entfalten sich nach winterlicher Ruhe im August, die Entwickelung wird also mitten in der Regenzeit abgebrochen, was wir vielleicht auf histo- rische Griinde zuriickfihren k6nnen. Der Regenwald der Taler und die Strauchgesellschaften der Abhange stehen systematisch und biologisch der valdivianischen Vegetation nahe, leider besitzen wir aber betreffs dieser nicht geniigende Daten, um einen Vergleich zu machen. 35—2610. The Nat. Hist. of Juan Fernandez and Easter Isl. Vol. II. 546 CARL SKOTTSBERG Die meisten Waldbaume und Straucher bliihen im Sommer, d. h. in der warmeren und trockeneren Zeit von November bis Januar: Yuania, Urtica fer- nandestana, Boehmeria, Berberts, Drimys, Lactoris, L:scallonia, Azara, Myre- eugenta, Halorrhagis, Pernettya, Khaphithamnus, Nicotiana, Coprosma, Wahlen- bergia; auch Ugni Selkirki und Colletta, welche sehr spat blihen, mogen hier Erwahnung finden. Die Krauter des feuchten Waldbodens, z. B. die Peperomza- Arten, Dysopszs und Solanum fernandeztanum \konnen wohl fast das ganze Jahr mit Bliiten gefunden werden, wahrend Arten, welche die offenen, lichten Stand- orte vorziehen, eine mehr beschrankte Blitezeit haben: Koehleria, Bromus, Cladium, Uncinia Douglasi, Carex Berteroniana und Ochagavia. Wenden wir uns jetzt den endemischen Schopf- und Kleinbaumen zu, von welchen die meisten die felsigen Abhange oberhalb 400 oder 500 m. besiedeln. Es sind Eryngium bupleuroides und inaccessum, Selkirkia, die Cumirnia-Arten, Plantago fernandesia, Evigeron fruticosus, die Robinsonia-Arten, Rhetenoden- dron, Centaurodendron und die Dendroseris-Arten. Mit einer einzigen Aus- nahme, Lryngium bupleuroides, welches von August bis April blihend gefunden wurde und ferner durch seine Haufigkeit zeigt, dass es sich in vollkommener Harmonie mit den jetzigen Verhaltnissen steht, machen die tbrigen den Ein- druck, sich in sehr labiler Lage zu befinden. Sie sind auch — Lryngzum macht eine Ausnahme — in der Kultur sehr empfindlich. Ihre Standorte zeich- nen sich, weil Nebel hier haufig auftritt, durch hohe atmospharische Feuchtig- keit aus. Von den montanen Arten bliht Avyug7um znaccessum im Vorsommer (November), Plantago fernandesia und Kobinsonia Gayana im Dezember, die iibrigen erst im Januar und Februar. Ein besonderes Interesse beanspruchen drei Compositenbaume, von welchen in gewissen Jahren kein Exemplar zu Blute kommt. So verhalt sich das dusserst seltene Cextaurodendron. Kein Exemplar bliihte im 1917 und ebensowenig hatten sie im 1916 gebliiht. Die beiden an- deren, Dendroseris pinnata und regia, sind hapaxanthisch. Wir sahen D. pn- nata auf allen Ausfliigen in die montane Region. Ein blihendes Baum- chen ist sehr auffallig und schon von weitem sichtbar, aber wir fanden wah- rend des Sommers 1916—17 nur ein einziges bliihendes Exemplar. Im August 1908 fand ich einige wenige, die wahrend des Sommers 1907—08 gebliht und nachher ihr Leben beendigt hatten. Von JL. vegza bliihte 1916—17 kein Exemplar. Ich meine also, dass wir auch in den phanologischen Verhaltnissen An- deutungen finden, dass die Flora von Juan Fernandez verschiedenen genetischen Elementen angehort, was ja ubrigens mit den systematischen und geographi- schen Befunden in vollem Einklang steht. POLLINATIONSBIOLOGIE UND SAMENVERBREITUNG AUF DEN JUAN FERNANDEZ-INSELN 547 Literaturverzeichnis. Gay, C., Historia fisica y politica de Chile. Botdnica. Paris 1845—52. Guppy, H. B., Observations of a naturalist in the Pacific II. London 1906. Hemstey, W. B, Report on the botany of Juan Fernandez and Masafuera. Rep. Challenger. Botany. Vol. 1. London 1885. Jouow, F., Estudios sobre la flora de las Islas de Juan Fernandez. Santiago 1896. ——, Zur Bestéubungsbiologie chilenischer Bliiten I—II. Verh. des deutschen wis- sensch. Vereins in Santiago. IV. 1go0o—o1. ——, Estudios de Biolojia Vejetal. Anales de la Universidad de Chile. Santiago IgIo. KIRCHNER, O. von, Blumen und Insekten. Leipzig 1911. Knutu, P., Handbuch der Bliitenbiologie. Leipzig 1898—1905. Merz, C., Bromeliaceae. Monogr. Phanerog. ed. C. p—E Canpotie. IX. Paris 1896. Porscu, O., Vogelblumenstudien I. Jahrb. fiir wissensch. Botanik LXIII. Leipzig 1924. Rercur, K., Die Vegetationsverhiltnisse am Unterlaufe des Rio Maule. Bot. Jahrb. fiir Systematik etc. 21. Leipzig 1896. ——., Beitriige zur Kenntnis der Gattung Azara. Ibid. ———, Zur Kenntnis der Bestéubung chilenischer Campanulaceen und Goodenia- ceen. Verh. des deutsch. wissensch. Vereins in Santiago. IV. 1902. ——, Grundziige der Pflanzenverbreitung in Chile. Die Vegetation der Erde VIII. Leipzig 1907. SKOTTSBERG, C., Feuerlaindische Bliiten. Wiss. Ergebn. der schwed. Siidpolar-Exp. IV: 2. Stockholm 1gos. ——, Studien iiber die Vegetation der Juan Fernandez-Inseln. K. Vetensk. Akad. Handl. 51:9. Uppsala 1914. ——, The Phanerogams of the Juan Fernandez Islands. Nat. Hist. of Juan Fern. and Easter Isl. II. Uppsala 1922. ——, Juan Fernandez and Hawaii. Bernice P. Bishop Museum. Bull. 16. Hono- lulu 1925. Wertu, E., Kurzer Uberblick iiber die Gesamtfrage der Ornithophilie. Bot. Jahrb. fiir Systematik etc. 53, Beibl. 116. Leipzig rors. —-—, Zum Verstindnis des Bestéubungsmechanismus der Kartoffelbliite. Angewandte Botanik VI. Berlin 1924. Printed ®), 7928. Berichtigung zu S. 511. Es zeigte sich spdter, dass auf dem erwahnten kultivierten Strauch von Colletia spartioides einige Fruchtknoten sich weiter ausbilden. Re ar i ae ae et 1c (ah en SATA RAGS Wty ORAL ay ie vena evtseke aonls” aiboibiuaHorr ss mn if is tae ne OP ie oe “al: Soe Ses tesi ai Shinmer, Sartore: Te Baars aS: i Madcon Fabhores Paps, ots rer eee 3. So Seay, oA oi aatna: ay ( oe Ae ; PA Berd! Baty it 5; ughiogh - Bae ii j vse at) x, Rave pe 10 Sas : i ae ibe 1 tf Sain yet aes ae ane (Piel anes ela uit tude wi : reir Hu pepcead e isd sibs kt “at Papper $) ch: eat figits p rF ‘oneth ee on . ta) ee PP is then ee res E. o ' at, ; oft i Soh hebtagarinlt is ab, aaledt OEE yf Patan ino! de eee r ‘ td - eh ioe rr Ae : an whee leach anges yt Oe ee cr ey: meciohin rst. i aise “rue ae eds ik. 417) ae ae 3 rire Pee? aes loderen, wzisue aie” al oainiiy aN 19. Nachtrag zur Pilzflora von Juan Fernandez. Von Dr. K. KEISSLER (Wien). Nachdem ich schon unter Nr. 16 dieser Serie ein Verzeichnis der von Prof. Dr. C. SKOTTSBERG gesammelten Ascomyzeten, Fungi imperfecti und Uredineen ver6ffentlicht hatte, wurden mir einige weitere Proben derselben Aufsammlung zugestellt, die ich unten behandelt habe. Darunter befand sich ein neuer Hymenomyzet, der von V. LITSCHAUER (Innsbruck) beschrieben wurde. Professor I. WEESE (Wien) hat drei Vectrza-Arten bestimmt. Hymenomycetes. Corticium subsphaerosporum Litsch. nov. spec. Parce et interrupte effusum, Ss e) pelliculare, molle, laxe adhaerens, S : 5 A Gc subtus et ambitu araneosum, in- S = terdum cherdulis rhizoideis fim- | by briatum; hymenium laeve, ex al- bido pallide cremeum, siccitate sub lente non contiguum sed porosum vel ut parvis glebulis b rimosum; basidia clavata, 10 — 12 X3,5— 4,5 2—4 sterigma- a tibus subulatis, rectis, 3—4 u longis; sporae obovatae-angulatae basi attenuatae, suboblique acu- tatae, 3—3,5(4)x 2—2,5(3) p, hyalinae, laeves, uniguttulatae, saepe 2— 4 adglutinatae; hyphae distinctae, regulares, tenuiter tuni- Corticium subsphaerosporum Litsch. nov. spec. . a. Basdien, b. Sporen, c. Hyphen, catae, laeves, hyalinae, septato- 2000/1 2500/1 500/1. nodosae, I,;—3 uv. Crassae, saepe ad septa ad 64. ampullatae, maximam partem parvulis aggregatis sphaeroideis cristallinis obscuratae. Ad petiolos Dicksoniae Berteroanae v. virgatae (Filic.). 550 K. KEISSLER Masafuera, Los Inocentes, g. III. 1917. Diese Art gehort innerhalb der Gattung Coréeczum in die Gruppe »Humi- cola», siehe Bourd. et Galz., Hym. de France, III, Bull. Soc. Myc. France, Tom. XXVII, p. 259 und steht am nachsten dem Corteceum sphaerosporum (Maire) v. Hoehn. et Litsch. Beitrg. HI, 1908, p. 25 (K. Akad. Wiss. Wien, Bd. CXVII, p. 1105) und dem Cortzctum suffocatum Bourd. et Galz., |. c., p. 262. Sie ist aber von diesen beiden Arten trotz grosser Ahnlichkeit im Aussehen und in der Struktur an den kleineren und anders gestalteten Sporen leicht zu unterscheiden. Pyrenomycetes. Fypocreates. Nectria (Cosmophora) Skottsbergii Weese nov. spec. ad. int. Masatierra: an der Rinde von MWyrceugenia fernandeziana, Quebreda del Juanango, 9. IV. 1917. Nectria cfr. flavolanata Berk. et Br. f. glabra. Masatierra: auf der abgefallenen Borke von Boehmeria excelsa, Pangal, Tele On. Nectria sanguinea (Bolt.) Fr. f. efespheria (Fr.). — Vgl. oben S. 479. Masatierra: auf moderndem Holz, Quebrada Damajuana, 30. XII. 1916. Sphaertales. Lasiosphaeria ovina Ces. et De Not. Masatierra: auf toten Zweigen, Cordon Salsipuedes, 13. I. 1917. Fungi imperfectt. Flyphomycetes. Chloridium viride Lk. Masatierra: auf moderndem Holz, Valle Colonial, Pico Central, 18. I. 1917. Printed '8/2 1928. VOEo IL saaeare HA OK A ee nioeg ae ATR ON Veh ae ¢ 1 eh 3 ies : of i ae - ; < * > , ry > hg > . z Pee 56) ates w. Evans: “The thallose Hepaticae of the Juan Fernandez Fae a tends 7 21 on ALBER T SercHRLL: The Codiums of the Juan Pesgabise s TS 3 Sry ; - y ? é -aupeag pes ie gm es ry CARL Saptrmtet- Marine Algal Communities of the Juan Fernandez Ee : ees with —— on the Soest of the Flora. as Tore LE ‘Die Dicallinncees: der ae Pe aa oe Laseln: 7 i See LEVRING: Einige Corallinaceen von der OsterInsel. ipa SKOTTSBERG: Additional Remarks t to Marine Algal Communities SS ee 20. The thallose Hepaticae of the Juan Fernandez Islands. By ALEXANDER W. EVANS. With six text figures. The thallose Hepaticae include the Marchantiales, the Jungermanniales Anacrogynae, and the Anthocerotales. The Marchantiales are represented in the Juan Fernandez islands by the Rebouliaceae and the Marchantiaceae; the Jungermanniales Anacrogynae by the Riccardiaceae, the Monocleaceae, and the Pelliaceae; and the Anthocerotales by the Anthocerotaceae, the only family of the order. The earliest collection of Hepaticae from the Juan Fernandez islands was made in 1830 by the Italian botanist, C. G. BERTERO. It includes several thallose species, the most conspicuous of which is W/archantia Berteroana Lehm. & Lindenb., which was described as new in 1834.’ This robust species, one of the largest members of the genus, is now known to be widely distributed in the higher latitudes of the Southern Hemisphere. In the following year MONTAGNE published a full account of BERTERO's cellular cryptogams under the title: »Prodromus Florae Fernandesianae. Pars prima, sistens enumerationem plantarum cellularium quas in Insula Juan Fer- nandez a Cl. Bertero collectas describi edique curavit C. Montagne, D. M.»’ In this paper he reported three additional thallose hepatics from the islands, namely: Fungermannia furcata L., F. pinnatifida Nees, and F. rhizoloba Schwaegr. The specimens upon which the first two records are based have not been available for study. It is hardly probable, however, that the plants referred to F. furcata represent Metsgeria furcata (L.) Dumort. of northern regions. It is much more likely that they would now be included under J/. decipiens (Massal.) Schiffn. & Gottsche, and the plants referred to F. pinnatifida perhaps represent Riccardia breviramosa (Steph.) Evans. The plants referred to 7. rhizoloba were made the type of Symphyogyna Hochstettert Nees & Mont. in 1836.2 They proved to be distinct from the true ¥. rizzoloba of Bourbon, now known as Symphyogyna rhizoloba (Schwaegr.) Nees. In 1850 MONTAGNE re- ported the cosmopolitan MJarchantia polymorpha L. from Juan Fernandez, on the basis of one of BERTERO’s specimens.* 1 In LEHMANN, Pug. Pl. 6:21, 49 (1834). 2 Ann. Sci. Nat. II, 3: 347—356; 4: 86—99 (1835). 8 Ibid. II. 5:68 (1836). ‘In Gay, Hist. Fis. Polit. Chile 7: 308 (1850). 36 — 30579. The Nat. Hist. of Juan Fernandez and Easter Isl. Vol. II. 552 ALEXANDER W. EVANS The next important contribution to our knowledge of Juan Fernandez Hepaticae was made by MITTEN in 1884 and constituted a part of HEMSLEY’'s Challenger Report.1_ Most of the specimens listed were found by E. G. REED, a collector sent to the islands by W. SAUNDERS; the others were collected by H. N. MOSELEY, botanist of the Expedition. In addition to Symphyogyna Hoch- stettert? and Marchantia Berteroana, MIiVTEN reported the thallose Rzccardia multifida (L.) S. F. Gray. A specimen bearing this name is in the Mitten Herbarium at the New York Botanical Garden and is listed in the present report under R. drevzramosa, which resembles the northern &. su/tzfida in many important respects. Nearly thirty years later, in 1911, STEPHANI published a report on the Hepaticae of Juan Fernandez, which was based on collections made by C. SKOTTSBERG, during the year 1908. This report is included in his memoir, »Die Lebermoose», which was issued as the second number of the »Botanische Ergebnisse der Schwedischen Expedition nach Patagonien und dem Feuerlande 1907 —1909.»" SKOTTSBERG’s collections are considerably larger than those made by the earlier explorers and include material from the island of Masa- fuera, as well as from the island of Masatierra, to which the earlier collectors had restricted their attention. STEPHANI reports fifty species of Hepaticae from the islands, of which the following nine are thallose: Aueura breviramosa Steph., Anthoceros Skottsberget Steph., Marchantia cephaloscypha Steph., Metzgeria chilensts Steph., Fossombronia fernandeziensts Steph., Reboulia hemisphaerica (L.} Raddi, Symphyogyna Hochstettert Nees & Mont., S. éxtegerrima Steph., and S. lanceolata Steph. Through the kindness of Dr. H. MOLLER, of the Natural History Museum (Riksmuseum) at Stockholm, and of Professor O. JUEL, of the University of Upsala, the writer has had the privilege of studying most of the specimens upon which STEPHANI’s reports were based. The results of these studies, which necessitate certain revisions and a few reductions to synonymy, are given in the descriptive part of the present paper. STEPHANI's additions, together with the earlier records, represent a net total of ten thallose Hepaticae for the islands. i In 1916—17 Dr. SKOTTSBERG, accompanied by his wife, Mrs. INGA SKOTTS- BERG, again visited the Juan Fernandez group and remained there for a period of five months. They explored both Masatierra and Masafuera thoroughly for Hepaticae and brought back a collection which is far more extensive and representative than all the earlier collections put together. The present report, which is based on the thallose species of this collection, increases the number of species now definitely known from the islands to twenty-seven. The collec- tion includes also a very few sterile specimens which are too incomplete for identification. When a species is marked with an asterisk (*) it represents an addition to the flora of Juan Fernandez; when an island is marked with an asterisk it means that the species in question is reported from that island for the first time. 1 Bot. Challenger Exped. 18: 84—8g9 (1884). ? Kungl. Svenska Vet. Akad. Handl. 469: 1—92. f. 1—35 (1911). THE THALLOSE HEPATICAE' OF THE JUAN FERNANDEZ ISLANDS eal ut v Marchantiales. Rebouliaceae. Plagiochasma Lehm. & Lindenb. *1, P. rupestre (Forst.) Steph. Bull. Herb. Boissier 6: 783 (1898); EVANs, Bull. Torrey Club 42: 277 f. 1—4 (1915). — Syn. Aytonia rupestre Forst. Char. Gen. Plant. 147 pl. 74 (1776). The species has a long synonymy. Damp or wet rocks, often in shade, below 500 m. (Fertile Dec.-Feb.). Masatierra: Pangal, waterfall at some distance from the water, 205 m. (no. 1); slope of Co. Damajuana, in forest (no. 2); Q. Damajuana, steep slope with forest, 345 m. (no. 3); B. Cumberland, vicinity of the colony, SKOTTSBERG, 1908 (no. 4); S. slope of Co. Chumacera (no. 5). Masafuera: Q. de las Vacas, under Calystegia (no. 6). Although P. rupestre is here reported from Juan Fernandez for the first time it was collected by SKOTTSBERG on Masatierra in 1908, as noted above, the earlier specimens growing in admixture with Reboulia hemisphaerica. The writer has recently published a description of this variable and widely distributed species, giving a full synonymy. Area of distribution: Mediterranean region of Europe, Asia and Africa; the Atlantic Islands; Abyssinia; South Africa; Australia and New Zealand; New Mexico and Arizona; Jamaica; Mexico; Brazil; Ecuador, south to Chile; Gala- pagos Islands; Juan Fernandez. Reboulia Raddi. 2. R. hemisphaerica (L.) Raddi, Opusc. Sci. Bologna 2: 357 (1818); STEPHANI, 5. — Syn. Marchantia hemisphaerica L. Sp. Pl. 1138 (1753). The species has a long synonymy. On rocks, below 500 m. (Fertile Feb.). Masatierra: B. Cumberland, vicinity of the colony, SKOTTSBERG, 1908 (no. 40, cited by STEPHANI, 5). Masafuera: Between Q. de las Casas and Q. del Blindado, on rocks among ferns (no. 7); cited also by STEPHANI, 5, from the vicinity of Q. de las Casas. Most writers assign a wide range of variability to A. hemzsphaerica and regard the genus Reboulia as monotypic. Area of distribution: much the same as that of Plagzochasma rupestre but even more extensive, reaching northern Europe, Asia and North America. ALEXANDER W. EVANS OL On AS Marchantiaceae. Lunularia {Mich.) Adans. *3. L. cruciata (L.) Dumort. Comm. Bot. 116 (1822). — Syn. Marchantia cruciata L. Sp. Pl. 1137 (1753). - Several other synonyms might be cited. Earth, rocks and decayed logs, mostly below 500 m. Masatierra: Q. Juanango, on ground in forest, 250 m. (no. 8). Masafuera: Q. de las Casas, on rock wall (no. 9), at entrance (no. g a), and in the outer part (no. 10); same locality, SKOTTSBERG, 1908 (no. 11); Q. de la Loberia, forest soil, 280 m. (no. 12); high ridge between Sanchez and Toltén, on the ground and on decayed logs, 515 m. (no. 13). As noted obove ZL. cructata occurred in SKOTTSBERG’s earlier collection but is here recorded from Juan Fernandez for the first time. The earlier spe- cimens were mixed with J/archantia foliacea. Probable all writers at the present time regard the genus Lunuular7a as monotypic. Area of distribution: Mediterranean region of Europe, Africa and Asia; Australia, Chile and Juan Fernandez; present also in many other parts of the world but usually appearing as if introduced and confined to greenhouses in cold climates; absent from most tropical districts. Marchantia {Marchant f.| L. 4. M. polymorpha L. Sp. Pl. 1137 (1753). On rocks, at low elevations. Masatierra: Co. Chumacera, S. slope on forested wet cliffs (no. 14); originally cited from Juan Fernandez by MONTAGNE, 308, no definite locality being mentioned; see also JOHOW 1896: 184. Area of distribution: almost cosmopolitan but especially abundant in northern regions; becoming rarer southward and sporadic in the tropics, often appearing introduced; in South Temperate and Antarctic regions now known from Chile, Kerguelen Island, the Crozet Islands, South Georgia and New Zea- land, as well as from Juan Fernandez. 5. M. Berteroana Lehm. & Lindenb. in LEHMANN, Pug. Plant. 6: 21 (1834); SCHIFFNER, Oesterr. Bot. Zeitschr. 46: 41 (1896); EVANS, Trans. Con- necticut Acad. 21: 246 (1917). — Syn. IZ. tabularis Nees, Naturg. Europ, Leberm. 4:71 (1838); I. cephaloscypha Steph. Hedwigia 22: 51 (1883). On soil and rocks, mostly at low elevations. (Fertile Dec. to Feb.). Masatierra: Pangal, on wet rocks (no. 15). *Masafuera: Q. de las Casas (no. 16); Q. del Blindado, on soil and stones, 440 m. (no. 16 a). THE THALLOSE HEPATICAE OF THE JUAN FERNANDEZ ISLANDS 555 The species was based on specimens collected by BERTERO in 1830 on Masatierra, no more definite locality being mentioned; it has since been col lected, presumably on the same island, by MOSELEY and by HASTINGS (see EVANS, 248). Area of distribution: Argentina, Chile, Juan Fernandez, and Falkland Islands; St. Helena and Cape Colony; Australia, Tasmania and New Zealand: apparently restricted to the Southern Hemisphere. *6. M. foliacea Mitt. in HOOKER, J. D., Flora Nov. Zeland. 2: 168 (1855). On soil and rocks, often in wet localities. (Fertile Dec. to Feb.). Masatierra: Pangal, waterfall and on stones in a spring, 205 m. (no. 17); ©. Damajuana, waterfall in the forest, 248 m. (no. 18). Masafuera: Q. del Mono, on the ground in a narrow gorge, 570 m. (no. 19); Q. de las Casas, SKOTTSBERG, 1908 (no. 27). Although this interesting species is here reported for the first time from Juan Fernandez, it was collected by SKOTTSBERG in 1908, as noted above, and listed by STEPHANI under the name JV. cephaloscypha (see STEPHANI, 5). According to a specimen in the herbarium of the University of Upsala, J/. foliacea was collected still earlier in 1908 by T. HALLE at Quicavi on the is- land of Chiloé. This specimen was listed by Stephani as J/. Notarzsa Lehm., a name which JV. foliacea antedates by two years. In the Mitten Herbarium (now belonging to the New York Botanical Garden) there are two sheets in the J/. folzacea cover with specimens from New Zea- land, one with specimens from Tasmania and Australia, and one with a specimen from the Solomon Islands. Among the New Zealand specimens are several collected by W. COLENSO and by J. JOLLIFFE, and it was from these that the original description was presumably drawn. Unfortunately these early specimens represent two species, making it necessary to designate one as the type. The writer therefore suggests that COLENSO’s No. 2070 be so designated. This species is the better developed of the two, it agrees best with MITTEN’s description, and it shows fairly well-developed female receptacles. The spec- imens from Juan Fernandez and Chiloé agree closely with this specimen and the same thing is true of a specimen from Tasmania collected by ARCHER. The specimens from Australia and the Solomon Jslands clearly belong elsewhere. Other New Zealand specimens in the Mitten Herbarium, which should be referred to the true JZ. foliacea, were collected by MOSSMAN and KERR, while several additional collectors are cited by Stephani in his Species Hepat- icarum. The present species belongs to the section Ch/amzdium, characterized by the flat or merely convex rays of the female receptacle. In this respect it stands in sharp contrast to VW. polymorpha and M. Berteroana, both of which have terete rays and belong to the section Astromarchantia. The receptacle of M. foliacea is usually flat and almost circular in outline, measuring about 1 cm. in diameter when well developed. The rays are normally nine but may be as few as seven or as many as eleven. At their broad and truncate extrem- ities they are decidedly leaf-like in appearance, as the specific name implies. 556 ALEXANDER W. EVANS The sinuses are narrow and extend about half way to the center of the discs. Except for the basal sinus, which is wider and deeper than the other the sinuses are subequal in width and depth. As in other species of the genus various irregularities are to be expected in the female receptacle, especially if the condi- tions have been unfavorable for its full development. The thallus is variable in size but is usually 3—6 cm. long and 5—12 mm. wide. The upper surface is pale to dark green and rarely shows signs of pigmentation. The lower surface, however, is pigmented with purple, sometimes throughout its entire extent and sometimes in the median portion only. The epidermal pores are of the cruciate type, the internal cells (normally four in number) approaching each other closely and leaving a very narrow opening. This type of pore is certainly implied by GOEBEL’s account,’ where he states that the opening can be reduced to a very narrow slit by means of the bounding cells, even if it can not be closed completely. STEPHANI, however, states that the opening is subquadrate, with oval bounding cells, showing that he may have had a different species before him when he drew up his description. The pores in JZ. foltacea are usually surrounded by six or seven rows of cells, each row, except the innermost of the upper series, being composed of four cells. The innermost row is usually composed of from six to twelve or four- teen cells. The compact ventral tissue of the thallus, so far as observed by the writer, is destitute of both slime cells and elongated sclerotic cells. GOEBEL states, however, that cells of the latter type are present and figures them clearly in both transverse and longitudinal view. In all probability the species is variable and fails to develop sclerotic cells under very moist environmental conditions, but the material available for study is too scanty to settle the question defi- nitely. The appendages of the median ventral scales are ovate to suborbicular in form and vary greatly in size, measuring 0.3—O9 mm. in length and 0.25— 0.6 mm. in width. The apices may be blunt, apiculate, acute or short-acuminate and the margins are either subentire or vaguely and irregularly sinuate-dentate; it is only in the rarest instances that a tooth is apiculate at the apex. The median cells of the appendages usually measure 50—60y in length and 35— 45» in width, while the marginal cells tend to be a little narrower. In most cases there is no sharp distinction between the median and marginal cells, but the latter sometimes extend at right angles to the margin and form an indistinct border. Additional features of the female receptacle which deserve mention are the stalk and the involucres. The stalk sometimes attains a height of 4—5 cm. and shows a broad dorsal band of green tissue with air-chambers; on the oppo- site side two furrows with rhizoids are present. The involucres are extremely delicate and the material studied is insufficient to give a wholly satisfactory idea of their features. It shows, however, that the free margins are rather densely ciliate, and that there is some indication of lobes in addition, but these lobes are apparently indifinite and separated by shallow sinuses. Mature spores are about 21, in diameter and show a yellowish or brownish color. A distinct ' Flora 96: 192, f. 143 (1906). THE THALLOSE HEPATICAE OF THE JUAN FERNANDEZ ISLANDS 557 wn marginal wing with a wavy edge is present at the periphery of the spherical face, and the latter bears a series of low flat ridges sometimes forming a vague reticulum. Male receptacles are not present in the American material, but the New Zealand specimens illustrate their characteristics. The stalks are about I cm. high and are destitute of green tissue with air-chambers. When well-developed the disc is about 0.5 cm. across and shows eight blunt lobes separated by shallow sinuses. When poorly developed the lobes are often very indistinct. The margins of the cupules are rather closely dentate to short-ciliate, the teeth varying from one to eight cells in length and from one to three cells in width at the base. In the apical portion on the outer surface a few scattered teeth are sometimes present but the surface is otherwise smooth. Probably the closest ally of J/. folzacea is the widely distributed J/. paleacea Bertol. of the Northern Hemisphere. The two species are very distinct, how- ever, in their cupules and involucres. In J/. faleacea both of these organs are deeply lobed and the lobes bear teeth or cilia, giving the effect of pinnate structures; in WZ. folzacea the cupules and involucres are toothed or ciliate but the lobes, even if present at all, are vague and a definitely pinnate appearance is not produced. On the whole J/. /folzacea is a little more robust than J/. paleacea, and the epidermal pores, although built up on the same cruciate type, tend to be slightly larger and to show more cells in the innermost row of the outer series. Whether LEHMANN’s JZ. WNotartsi is actually a synonym of J. foliacea is doubtful. The published descriptions would hardly lead to this conclusion; they point rather to J/. chenopoda L. At the same time the descriptions are not altogether decisive, and the only way in which the question could be de- finitely answered would be by the study of LEHMANN’s original specimens. Area of distribution: New Zealand and Tasmania; Chiloé and Juan Fernandez. Jungermanniales. Riccardiaceae. Riccardia S. F. Gray. #7. RR. fuegiensis Massal. Nuovo Gior. Bot. Ital. 17: 255 pl. 26 f. 34 (1885); EVANS, Trans. Connecticut Acad. 25: 113 (1921). Alpine heath, among grasses. Masafuera: the High Plateau, 1100 m., mixed with other bryophytes (no. 20). The determination is based on a few scattered stems, but these show ventral lamellae and are therefore referred to R. fwegéenszs, the only species in which these structures have as yet been described. Unfortunately the material from Masafuera is less robust than typical Chilean plants, the margins of the 553 ALEXANDER W. EVANS Fig. 1. Riccardia breviramosa (Steph.) Evans. A—C. Parts of male thalli, x 8. D. Transverse section of the main axis, X 100. E, F. Trans- verse sections of primary branches, X 100. G. Tip of a primary branch, with secondary branches, x 40. H Marginal part of main axis in transverse section, X 100. I. Transverse section of a metzgerioid branch, X 140. J. ‘Transverse section of a stolon, X 140. K. Ventral surface of a primary branch, X 10c. A—E, G. and I—K were drawn from no. 21; F and H, from no. 25. wings are less crispate, and the trigones are less pronounced. These slight differences, however, are merely quantitative and would hardly be sufficient to justify a specific separation, even if the specimens were more abundant. THE THALLOSE HEPATICAE OF THE JAUN FERNANDEZ ISLANDS 559 Area of distribution: South Chile, Magellan region to Fuegia and neigh- boring islands; Masafuera. 8. R. breviramosa (Steph.) Evans, comb. nov. — Syn. Aneura brevi- ramosa Steph. Kungl. Svenska Vet.-Akad. Handl. 46°: 6 f. 1 a (1911). — Fig. 1. On tree ferns, decayed logs and rocks, 700 m. and below. Masatierra: El Yunque, in moist woods, SKOTTSBERG 1908 (no. 1, cited by STEPHANI, 6); B. Cumberland, caves V and VI, mostly on the side and back walls (no. 21, 22, 23, 24) and in the old caves near landing (no 24 a); SW. slope of the Portezuelo ridge, ravine among Dicksonza, 550 m. (no. 25); C. Salsipuedes, Décksonia forest, 660 m. (no. 26); without exact locality, REED (Herb. Mitten, cited under Réccardius multifidus by MITTEN, 88). *Masafuera: E. of Las Inocentes, on Dicksonta, 700 m. (no. 27). This is apparently the most abundant species of Azccardia in Juan Fer- nandez. It was based on tio specimens from Masatierra and one from the vicinity of Port Stanley, Falkland Islands. One of the Masatierra specimens came from the vicinity of the colony and the other from El Yunque. The latter has been examined by the writer and includes specimens of the following species as well as the true Ameura breviramosa. STEPHANI’s figure represents the cross section of a primary branch and is exceedingly diagrammatic. The thalli of R. dreviramosa (Fig. 1, A—C) are prostrate and grow in depressed mats, sometimes in layers. They present a translucent appearance, and the color varies from dull green to brownish. Rhizoids are scantily pro- duced and are largely restricted to the stolons, so that the plants are only loosely attached to the substratum. The main axis is mostly 1—2 mm. wide and 0.3—0.35 mm. thick, and the living portion is 1—2 cm. long. The upper surface is plane or nearly so (Fig. 1, D), the lower surface is distinctly convex, and there is a gradual thinning out toward the margins, which are bounded by more or less crenulate wings one to three cells broad. In the median portion the axis is mostly eight to ten cells thick. The cells of the surface layer are scarcely longer than broad, averaging about 55. in length by 45. in width; along the margin they are distinctly smaller, measuring perhaps 32. in length. The interior cells are about 80». wide and 160—200 yp, long, so that there is a rather marked contrast in width when the surface-cells and the interior cells are compared. Except in size the cells show little differentiation, and the cell-walls are everywhere thin or only slightly thickened. The primary branches are rarely more than I mm apart and are some- times so close together that their bases are almost in contact. They spread obliquely to widely and sometimes give the thallus a regularly pinnate appear- ance. In rare cases a primary branch represents a new axis and continues its growth indefinitely. Usually it stops growing after attaining a length of 4—6 mm, and under such circumstances is both narrower and thinner than the axis, measuring perhaps 0.7—O9 mm. in width and O.2 mm. in thickness. At inter- vals of about 0.5 mm. on each side the primary branches give rise to short secondary branches, which are rarely more than 0.5 mm. in length, 0.3 mm. in 560 ALEXANDER W. EVANS width and O.12 mm. in thickness. Tertiary branches arise only when the sec- ondary branches are more vigorous. In their histological features the branches are essentially like the main axis, but the primary branches (Fig. 1, E) are only about six cells thick in the middle and the secondary branches only four or five cells. In most cases the marginal wings with their crenulations are more distinct on the branches (Fig. 1, G, K) than on the main axis, and the ventral slime papillae show a definite arrangement in two longitudinal rows. The branches show clearly also the gradual decrease in the size of the surface-cells in passing from the median portion to the margin. Although marginal wings and crenulations are distinctive of the species, they are not necessarily present on every vegetative axis. Sometimes the main axis (Fig. 1, H) or even a primary branch (Fig. 1, F) is rounded or bluntly pointed in cross section. Under such circumstances the surface-view does not show a wing or even a clearly differentiated row of marginal cells, and the thallus is bounded by a continuous line. Attention may be called also to two modifications which the secondary branches or the apices of the primary branches sometimes undergo. In one of these modifications the branches are only O.135—0.2 mm. in width and are flatter than ordinary branches. They present, in fact, a strongly metzgerioid appearance, the narrow costa being sharply marked off from the unistratose wings (Fig. 1, I). The other modification is equally narrow but is only slightly flattened and shows no differentiation into costa and wings (Fig. 1,J). Branches of this type may be interpreted as stolons; they seem to arise without any regularity (Fig. 1, B), although in some cases they may be present on one side of an axis but not on the other (Fig. 1, C). The inflorescence of XR. breviramosa seems to be invariably dioicous, as stated by Stephani. The male branches usually arise, either singly or in sub- opposite pairs, from the base of a primary branch, but in some cases they are borne directly on the main axis. The inflorescence, which is often more or less recurved, occupies the greater part of the branch and usually measures 0.4—0.6 mm. in length and 0.3 mm. in width. The antheridia in most cases number from four to sixteen, and the openings into the antheridial chambers are usually separated by two rows of cells. The wings are only one cell wide but are distinctly crenulate. As a rule they are erect or somewhat connivent, thus partially concealing the upper surface of the inflorescence. The female branches agree with the male branches in their position, being borne either on the main axis or at the base of a primary branch. The arche- gonia are usually abundant and may number sixteen or even more. The wings are divided up into irregular lobes, blunt or sharp at the apices and usually five to seven cells long. They are supplemented by similar outgrowths in connection with the archegonia. In rare cases a female branch may continue its growth beyond the inflorescence and form a simple or branched expansion of the usual vegetative type. The »calyptra» is about 0.6 mm. in diameter and may attain a length of about 2 mm. The corona is not distinct, but the entire surface of the calyptra is covered over with crowded and irregular projecting cells or groups of cells. THE THALLOSE HEPATICAE OF THE JUAN FERNANDEZ ISLANDS 561 The description just given is drawn from robust material, which naturally brings out the distinctive features of the species most clearly. Some of the specimens from Cumberland Bay, however, especially those collected on the wettest parts of the back wall, are greatly etiolated, the main axes are often only 0.4—0.6 mm. wide, the narrow primary branches are sometimes more than I mm. apart, and secondary branches are relatively infrequent. These specimens, however, show crenulate wings of the characteristic type, and the few sexual branches present are essentially like those on the more vigorous plants. There are two species of the Chilean mainland to which XR. éreviramosa is obviously related. One of these is A. autoica (Steph.) Evans, which is usu- ally dioicous in spite of its name, and the other is &. texerrzma (Steph.) Evans, which is normally autvicous. Both of these species have recently been described and figured by the writer.’ In 2. auéoica the prostrate axis is adherent to the substratum by means of rhizoids; the differentiation into various types of branches is usually far more evident than in 2. breviramosa; the cells are distinctly smaller, both the surface-cells and those of the interior averaging only 30. in width; the male inflorescence is more complex, the openings into the antheridial chambers being separated by three or four rows of cells; and the involucre of the female inflorescence, which is directed forward rather than upward, is ciliate but no lobate. In R. fexerrima the whole thallus is adherent to the substratum and is thinner and much more delicate than that of R. dreviramosa, although the interior cells average only 50, in width. The female inflorescence, more- over, is shorter, the archegonia are fewer, and the wings are narrower and subentire. The crenulate wings of R. dreviramosa will at once recall the very similar structures found in &. multifida (L.) S. F. Gray, a common and widely distrib- uted species of Europe, Asia and North America. It is not surprising, there- - fore, that MITTEN referred the scanty and sterile Masatierra material at his disposal to this northern plant. He stated, however, that his determination was merely provisional. On the whole R. multefida is narrower and thinner, the axis being only 0o.;—1 mm. wide and four to seven cells thick; and the branching tends to be more definitely bipinnate, the secondary branches being longer. The inflorescence, moreover, is autoicous, and the involucre of the female inflorescence is ciliate rather than lobate. Area of distribution. Falkland Islands; Juan Fernandez. *9. R. adglutinata nov. spec. — Fig. 2, A—C. Growing in depressed mats, dull green, often becoming tinged with yellowish or reddish brown in the older parts, usually opaque: thallus postrate throughout and more or less closely adherent to the substratum; axis flattened, sometimes wingless but often bordered by a wing one or two cells wide; branches close together or somewhat distant, sometimes equalling the axis in width but shorter and often narrower, irregularly once or twice pinnate, the ultimate branches often persisting as rudiments, marginal wing if present often two or three cells wide and sometimes crenulate: inflorescence autoicous: ¢ inflorescences borne singly or 1 Trans. Connecticut Acad. 25 159. 164 f. 7 and 8, A, B (1921). 562 ALEXANDER W. EVANS in groups of two or three, the wing crenulate, one cell wide, usually somewhat involute; antheridia usually four to twelve: @ inflorescence occupying a simple and very short branch; involucre composed of short cilia or lobes, one to three cells long and often more or less coalescent; archegonia borne single or in small groups; calyptra while young covered over with rounded celis or groups of cells and showing a low but distinct corona, afterwards becoming nearly smooth: cells of outer layer of capsule wall with numerous thickenings in the form of incomplete rings; cells of inner layer without local thickenings; spores greenish, minutely verrucwlose. On rocks and decayed logs, 600 m. and below. Masatierra: V. Anson, near Plazoleta, on decayed timber, 260 m. (no. 27); El Yunque, in moist woods, SKOTTSBERG 1908 (mixed with . dreviramosa, (no. 28a); V. Colonial, Q. Gutierrez, c. 300 m. (no. 29); B. Cumberland, on moist cliffs near the sea (no. 30); SW. slope of Co. Piramide, c. 600 m. (no. 31); Q. Juanango, in forest, c. 250 m. (no. 32, 33); Q. Juanango, waterfall (no. 34). Masafuera: innermost accessible part of ©. de las Casas, on trunks, c. 200 m. (no. 35); Q. de las Casas, on log in the stream (no. 36). No. 35 may be designated the type. The present species (Fig. 2, A, B) grows in depressed mats and is usually opaque. It is dark grecn in color but often shows a peculier yellowish or reddish brown pigmentation even in the younger portions. The axis and most of the branches cling closely to the substratum by means of crowded rhizoids, making it difficult to remove intact thalli. The axis is strongly flattened (fig. 2, C), measuring 2—4 mm. in width and only 0.3—0.35 mm. in thickness, while the length of the living portion is usually 1—2 cm. The upper surface is plane or nearly so, and the lower surface is somewhat convex. From the median portion, which is usually six to eight cells thick, the axis thins out gradually toward the margins; here an indistinct wing, one or two cells wide, is some- times distinguishable, but the axis is often wingless and shows no signs of a differentiated row of marginal cells. The surface-cells average about 45 X 30 p,, but the innermost cells measure 100—140 p. in length by about 80». in width. In cross section there is therefore a gradual increase in width in passing from the surface inward. The cells walls are thin throughout. Branches are given off by the axis at indefinite intervals (Fig. 2, A, B) and the thallus scarcely ever approaches a regularly pinnate condition. The primary branches are sometimes so close together that they overlap, but they may be separated by a distance of 1 mm. or more. In some cases they equal the axis in width; they tend, however, to be narrower and are sometimes only Os mm. wide. Except in unusual instances, where the primary branches rep- resent new thalli, they rarely exceed a Jength of 5 mm. In most cases sec- ondary and even tertiary branches are present, but these usually remain very short and may not develop beyond a rudimentary condition. Under such circumstances the higher axis shows one or more very short bilobed outgrowths, with rounded lobes and apical cells in the depressions between them. ‘The branches are nearly as thick as the main axis but tend to thin out more abruptly and are sometimes bounded by fairly distinct wings two or three cells THE THALLOSE HEPATICAE OF THE JUAN FERNANDEZ ISLANDS 563 wide. Marginal crenulations, formed by projecting cells, are occasionally to be observed, especially on the branch-rudiments. Even here, however, they are rarely conspicuous, and most of the branches are entire or nearly so. In some cases branches or even rather complicated branch-systems assume the appearance of stolons. Such branches, which are only 0.2-04 mm. wide, are slightly flattened, but the edges are rounded and wholly destitute of wings. The devel- opment of these stolon-like structures, which are never highly differentiated, seems to be induced when branches force their way under older thalli. The inflorescence of A. adglutinata is normally autoicous although it is difficult to demonstrate both antheridia and archegonia on every individual plant. The male branches usually occur singly, more rarely in groups of two or three, and arise directly from the main axis or from the base of a branch. The in- florescence, which apparently never proliferates, occupies nearly the whole of the branch and is sometimes recurved. It measures 0.3—0.7 mm. in length and about 0.3 mm. in width. The marginal wings, although but one cell wide, are more or less involute and distinctly crenulate; the antheridia number from four to twelve; and the openings into the antheridial chambers are separated by two rows of cells. The female branches are very short and are similar in position to the male branches. Sometimes branches of both sexes occur in close proximity, although there is nothing definite in their arrangement. The inflorescence occupies the whole of the branch, and no instances of proliferation have been observed. The involucre consists of a series of short cilia or narrow lobes, one to three cells long and often more or less coalescent. The archegonia occur singly or in small groups, six being the highest number seen. The massive »calyptra» measures 2.4 mm. in length and about 0.9 mm. in diameter at maturity and is then almost smooth. In younger stages the surface is covered over with rounded cells or groups of cells, and the apex bears a low but distinct corona. The material examined includes a few mature sporophytes in excellent condition for study. The oval capsule is about I 2 mm. long and 0.45 mm. wide, and the wall shows the usual two layers of cells. In the outer layer the cell-walls turned toward the median lines of the valves show numerous bands of thickening, which extend partially or wholly across the inner tangential walls and thus form incomplete rings; in the inner layer the cell-walls are thin throughout. The spores are greenish and 18—22 u, in diameter, while the slightly thickened walls are covered over with minute verruculae. The eleaters, which are brown and long-pointed at each end, are mostly 0.15--03 mm. long by 10—12u. wide and show a single, broad, loosely wound spiral. Gemmae are sometimes produced in considerable abundance on rather narrow vegetative branches. They conform to the characteristic two-celled type and average about 50 x 30, their walls being slightly thickened. When R. adglutinata is typically developed its dark and opaque thallus, often tinged with yellowish or reddish, is very different in appearance from the more or less translucent thallus of R. drevzramosa. Unfortunately, in poorly developed material, the opacity is sometimes less pronounced, and the plants then bear a certain superficial resemblance to the older species. In their cell- structure, moreover, R. adglutinata and R. breviramosa are a good deal alike; 564 ALEXANDER W. EVANS they both show a contrast in width when the exterior and interior cells are compared, the cell walls in both are thin and delicate, and the cell-measurements are not very different. There are, however, certain important differences between them, aside from those in texture and color. In &. dreviramosa, for example, the thalli often have a regularly pinnate appearance; the secondary branches rarely remain in a rudimentary condition but are usually simple; the margins of the branches are commonly distinctly crenulate; rhizoids are absent or spar- ingly developed; the inflorescence is dioecious, the female branch develops num- erous archegonia; and the lobes of the involucre are mostly five to seven cells long. In &. adglutinata, on the other hand, the branching is more irreg- ular; the secondary branches often start tertiary branches, but both frequently persist as rudiments; the margins are rarely crenulate, except in restricted regions; rhizoids are abundantly produced; the inflorescence is autoicous; the female branch bears few archegonia and sometimes but a single one; and the lobes or cilia of the involucre are only one to three cells long. The structure of the capsule in R. adglutinata is the same as that of R. multifida,, so far as the local thickenings of the cell walls are concerned, and the species agree further in their inflorescence. Otherwise they have little in common. As already noted under R. breviramosa, R. multifida is a narrower and thinner plant, far more regularly pinnate, and the wings of the branches are usually distinctly crenulate. Area of distribution: Endemic. #10, R. insularis Schiffn. Deutsche Siidpolar-Expedition 8: 66 pl. 6 f. 1—7 (1906). — Fig. 2, D—H. Wet rocks at low altitudes. Masatierra: B. Cumberland, caves V and VI, side and back walls (no. 37, 38, 39). It is with some hesitation that these specimens from wet rocks are referred to R. insularis, partly because Schiffner’s species was based on material from the far-distant islands of St. Paul and New Amsterdam, partly because the original specimens grew on the roots of ferns, and partly because bisexual branches were occasionally present on these specimens but have not been de- tected on the Masatierra plants. At the same time, except for this single structural difference, these plants agree so fully with SCHIFFNER’s detailed de- scription and careful figures of R. zusularis that a specific separation seems uncalled for. It may be remarked also that bisexual branches, even in species where they are sometimes found, are not necessarily present on every individual, so that too much stress should not be laid on their absence. The plants (Fig. 2, D) are a pale or dull green, turning yellowish or brownish with age, and are fully as translucent in appearance as KX. breviramosa. They form depressed mats, becoming loosely layered in time, and produce rhizoids very sparingly. The main axis, of which the living portion is I—2 cm. long, is strongly flattened, being only 0.15-.0.2 mm. thick although 1—2 mm. wide (Fig. 2, G). From the median portion, where the cells are in only four or five layers, it thins out very gradually toward the edges, which are bordered THE THALLOSE HEPATICAE OF THE JUAN FERNANDEZ ISLANDS 565 Rios 2: A—C. Riccardia adglutinata Evans. A, B. Parts of thalli, with male and female inflorescences, X 8. C. Transverse section. of the main axis, X 100. The figures were all drawn from the type specimen. D—H. Riccardia insularis Schifin, D. Part of a thallus, with male and female inflorescences, x 8. E, F. Branches with male and __ female inflorescences and a young sporophyte, X 8. G. Transverse section of a primary branch, % 100. H. Male branch, dorsal view, X 100. The figures: were all drawn from specimens col- lected at B. Cumberland. 566 ALEXANDER W. EVANS by a unistratose wing one to three cells wide. The cells of the superficial layer average about 60 X 42 y., becoming shorter toward the edges, the marginal cells being almost isodiametric. Some of these marginal cells project slightly, forming irregular and often scattered crenulations, but the wings are sometimes nearly or quite entire. The interior cells are mostly 160—200 y. in length and 80 p. in width, so that the contrast in size between the exterior and interior cells is pronounced in cross section. The walls are thin and delicate throughout. On the whole the branching is less regular than SCHIFFNER implies in his description of A. zzsularis, and it is sometimes difficult to draw a sharp line between the main axis and the primary branches. Uusually, however, the latter are only 0.5—1 cm. long, I—I.5 mm. wide and three or four cells thick in the middle. Their structure is essentially the same as that of the axis. In most cases the primary branches are simple, but sometimes short tertiary branches or their rudiments can be demonstrated. An occasional branch is modified into a stolon-like structure or becomes stolon-like in the outer part. Such a branch is only 0.2—0.3 mm. wide and is subterete and wingless. The median portion is usually five to seven cells thick, and the interior cells are only slightly larger than the surface-cells. In the Masatierra specimens the inflorescence seems to be constantly autoicous (Fig. 2, D, E), although male and female branches sometimes arise in close proximity. The male branches are numerous and are borne, singly or in groups of two or three, either on the main axis or on a primary branch. The inflorescence occupies the greater part of the branch (Fig. 2, H) and measures 0.6—I mm. in length by 0.3—0.4 mm. in width. It is bordered by a spreading crenulate wing, one cell wide, and usually develops from eight to sixteen antheridia, even more being present in some cases. The openings into the antheridial chambers are separated by two rows of cells, although three rows are sometimes present between the two longitudinal rows of antheridia. The female branches (Fig. 2, D—F) are similar in size and position to the male branches. The archegonia are usually numerous, from twelve to twenty-four being often present, and the involucre consists of a ciliolate wing, with cilia two or three cells long, supplemented by a few similar cilia among the archegonia. The »calyptra» may attain a length of 4 mm. and a diameter of 0.8 mm. at maturity but is often smaller, SCHIFFNER giving the length as 2 mm. and the diameter as 0.5 mm. The surface sometimes bears a few un- fertilized archegonia in the lower part and, in addition, scattered outgrowths, some of which resemble the involucral cilia. Otherwise the surface is smooth or nearly so and the apical corona is inconspicuous. The capsule is essentially like that of A. adglutinafa, in so far as the local thickenings in the cells of the wall are concerned. It agrees also in all respects with SCHIFFNER's description, even to the absence of half-annular thickenings in the marginal rows of the valve-cells. At maturity it measures I—!.2 mm. in length by 035--0.5 mm. in width. The yellowish spores are mostly 14—16 p. in diameter with slightly thickened minutely dotted walls; and the elaters, each with a loosely wound broad spiral, measure 0.3— 0.45 mm, by 12—I4, THE THALLOSE HEPATICAE OF THE JUAN FERNANDEZ ISLANDS 567 It will be seen that A. zzsu/aris resembles the two preceding species in size and general habit but is relatively thinner than either. It agrees with AX. breviramosa, furthermore, in its translucent texture, in its scanty development of rhizoids, and in most of the characters derived from the male inflorescence. It differs, however, in its monoicous inflorescence, in its ciliolate rather than lobate involucre, and in the more or less complete absence of marginal crenu- lations which its vegetative branches exhibit. With 2. adelutimata it agrees in being monoicous, and also in the nature and arrangement of the local thickenings in the capsule wall, but R. adglutinata can usually be at once distinguished by its opacity, by its abundant rhizoids and by its much shorter female branches Fig. 3. Riccardia variabilis Evans. ’ A—C. Parts of thalli of various types, X 8. D. Tramsverse section of a wide branch, X 100. E, F. Transverse sections of narrower branches, x 100. G. Transverse section of a stolon, x 100. H. Male branch, dorsal view. The figures were all drawn from the type specimen. with more irregular involucres. SCHIFFNER has already compared &. zvsularis with the closely related R. maltifiordes Schiffn., of Java, and also with two other species from the same island. Area of distribution: St. Paul and New Amsterdam; Masatierra. *11. R. variabilis nov. spec. — Fig. 3. Growing in thin, somewhat turf-like mats, dull green becoming brownish with age, more or less translucent: thallus in part prostrate and adherent to the substratum but the apices of the axis and its branches often ascending; axis usually subterete and wingless in the lower part, becoming flattened above and often showing an indistinct wing one to three cells wide; branches very 37—30579- The Nat. Hist. of Juan Fernandez and Easter Ish: Vobx El. 568 ALEXANDER W. EVANS irregular, the primary branches more or less scattered and (if photosynthetic) soon spreading away from the substratum and tending to broaden out into rounded expansions, sparingly subdivided, wings as in the axis, sometimes vaguely crenulate from projecting cells: inflorescence dioicous: & inflorescences arising singly, spreading to suberect; antheridia mostly four to ten; @ inflores- cence occupying a very short branch, the wing narrow and irregularly crenulate: sporophyte not seen. On decayed wood, 300 m.; also on soil. Masatierra: Q. Gutierrez, on decayed wood, 300 m. (no. 40), the type- specimen. The author would refer to the same species a sterile specimen from Concepcién, Chile, collected by R. THAXTER in November, 1905 (no. 67); this specimen grew on soil. The smaller size of R. variabilis, together with its irregular habit (Fig. 3, A—C) will at once distinguish it from any of its predecessors. The plants, however, are not very different in color from R. dreviramosa and R. insularis and show a similar translucent texture. In the more typical individuals, the axis is fairly distinct (Fig. 3, C) and is prostrate at the base, the apical portion curving away from the substratum. It is usually 3—5 mm. long and at first only 0.1—0.3 mm. wide, but sometimes becomes 1—I.5 mm. broad in the apical portion. The thickness is usually about 0.1 mm. throughout, thus showing a subterete condition at the base and gradually passing into a strongly flattened condition. In the subterete portion the edges are rounded and wingless; in the flattened portion a row of marginal cells or a wing two or three cells broad is often distinguishable, the wing being subentire or vaguely and irregularly crenulate from projecting cells. In the thickest portion of the axis the cells are in only four or five layers, and the cell-walls are everywhere thin and deli- cate. The surface cells have an average diameter of about 35 ». and are some- times nearly isodiametric especially along the margin; in many cases, however, they are longer than broad and may attain a length of 60—100 y.. The interior cells are mostly 120—200 », long and 60 »u. wide, the cross sections thus showing a sharp contrast in size between the cells. The branches, which are very irregular and sometimes hardly distinct from the main axis, usually arise at intervals of o.3—1 mm. At the same time photo- synthetic branches (Fig. 3, D—F) and stolons (Fig. 3, G) can be distinguished, although connected with each other by intermediate types. A typical photo- synthetic branch is narrowed and prostrate at the base and broadens out at the ascending apex into a blunt expansion 0.5—1I.5 mm. wide, thus resembling the apical portion of a main axis. It usually gives rise to a few similar but shorter secondary branches and the latter sometimes bear one or two very short tertiary branches or their rudiments. The cell-structure of the photosynthetic branches is essentially like that of the axis. The stolons are about the same as the basal prostrate portion of the axis or a photosynthetic branch. They are mostly 1—2 mm. broad, simple or very sparingly branched, wingless and but slightly flattened. Their cells average only 20. in width and show little contrast in cross section between the exterior and interior layers. On the stolons and on the prostrate portions of the other axes rhizoids are abundantly produced; elsewhere they are scanty or absent altogether. THE THALLOSE HEPATICAE OF THE JUAN FERNANDEZ ISLANDS 5609 The inflorescence of XR. variabilis seems to be definitely dioicous. The male branches (Fig. 3, B) are sparingly produced and arise singly from the axis or a primary branch, rarely from a secondary branch. The inflorescence (Fig. 3, H), which is usually without a stalk, is mostly 0.3—0.45 mm. in length and about 0.3 mm. wide. The antheridia in most cases number from six to ten and the openings into the antheridial chambers are separated by one or two rows of cells. The wing is crenulate, spreading to suberect, and only one cell wide. The female branch figured is the only one demonstrated. It is very short and shows a narrow and irregularly crenulate wing. The calyptra is 1.8 mm. in length by 0.45 mm. in width and is so old and battered that the corona is no longer distinguishable. The surface, however, still bears vestiges of small pro- jecting clusters of rounded cells. Among the species of Rzccardia previously reported from Chile, the closest ally of R. variabilis is evidently R. conimitra (Steph.) Evans, known from three stations along the western coast.’ In this species the thallus is about as large as that of R. variabilis and is almost as irregular in its branching. The branches show, moreover, a similar vague differentiation into photosynthetic branches and stolons, and the axes of various ranks are only four to six cells thick in the middle. Even the male inflorescences are much alike in the two species, except that in A. conimitra the upper surface is often roughened by bulging cells. There are, however, several important distinctions to be noted. In &. varzadbz/zs, for instance, a typical photosynthetic branch broadens out conspicuously toward the apex; there is little or no tendency for the branches to become: attenuate, except perhaps in the case of the stolons; and the external cells of the thallus are distinctly smaller than the interior cells, except again in the case of the stolons. In R. coniméttra, on the other hand, the photosynthetic branches do not broaden out to any extent; there is a strong tendency for the branches and even for the main axes to grow out into slender attenuate extensions; and the external cells of the upper surface are usually distinctly larger than the interior cells, the contrast being especially marked on the narrow branches. Another species of the Southern Hemisphere with which R. varzadbzlzs should be compared is Aveura subantarctica Kaalaas of the Crozet Islands,” known to the writer from description only. This species agrees with 2. varza- bilis in general habit, the primary axis is described as five cells thick, and a marked contrast in size is indicated between the external and internal cells. The inflorescence in A. subantarctica, however, is autoicous; the branches are described as attenuate, although the axis is said to be dilated as a rule at the apex; and the marginal wing, even when present, is said to be only one cell wide. The Crozet Island species is furthermore somewhat larger than X. varza- bilis, the length being given as 1I.5—2 cm. and the width as 0.5—0.6 mm. KAALAAS suspected that the specimens from Kerguelen Island, which MITTEN had reported under the name -4. mu/tifida, might possibly be the same as his new species, but he left the matter in uncertainty. Area of distribution: Chile (Concepcién); Masatierra. 1 Trans. Connecticut Acad. 25: 156 f. 6 (1921). 2 Nyt Mag. Naturv. 49: 87 f. 1 (1911). 570° ALEXANDER W. EVANS * 12. R. leptostachya nov. spec. — Fig. 4. Growing in velvety but somewhat depressed mats, dull green becoming brownish with age, more or less translucent: thallus consisting of a prostrate axis giving off ascending branches and prostrate stolons, the axis flattened and often with a wing one or two cells wide; ascending branches mostly simple but sometimes sparingly subdivided, more strongly flattened than the axis, the marginal wings being two or three cells wide and crenulate: inflorescence dioi- cous: ¢ inflorescences arising singly, not recurved, the wing -crenulate, erect or incurved; antheridia mostly ten to twenty: 2 inflorescence occupying a very short branch; involucre consisting of a series of short and irregular cilia or Fig. 4. Riccardia leptostachya Evans. A—C. Parts of thalli of various types, X 14. D. Transverse section of prostrate axis, X 100. E. Transverse section of an erect branch, % 100. F. Male branch, dorsal view, X 100. G. Fe- male branch with young sporophyte, X 50. The figures were all drawn from the type specimen. laciniae, three to five cells long; calyptra smooth or nearly so, the corona in- conspicuous: sporophyte not seen. On decayed wood, 260 m. and below. Masatierra: V. Anson, near Plazoleta, on decayed timber, 260 m. (no. 41); Q. Juanango, on logs in river bed (no. 42). No. 42 may be designated the type. In &. leptostachya the differentiation in the branches, which is rather vaguely indicated in X. variabilis, is more completely realized, although it still falls short of the definite differentiation found in certain other members of the genus. The thallus in other words, shows a prostrate axis with long-continued growth, giving rise to ascending photosynthetic branches and prostrate stolons, THE THALLOSE HEPATICAE OF THE JUAN FERNANDEZ ISLANDS 571 both with limited growth (Fig. 4, A—C). The axis itself, however, in many cases if not always, curves away from the substratum after a while and has its own growth brought to an end, and branches of a character intermediate between the extreme types are of frequent occurrence. In consequence of the ascending branches the plants form velvety mats, although the texture is too delicate to give them much rigidity. The color and degree of translucency are much as in &. variadzizs. The prostrate axis (Fig. 4, D) is mostly 2—4 mm. long, 0.2z—0.4 mm. wide and O.r5 mm, thick in the middle. It is therefore more or less distinctly flat- tened but usually turns an edge rather than a surface to the substratum. From the median portion, where the cells are in four or five layers, it thins out rather abruptly and is usually bounded on one or both sides by a unistratose wing one or two cells wide. The surface cells average about 55 X 40 p, becoming shorter toward the edges, the marginal cells being only about 35», long. The interior cells average 60» in width and are usually 120—140 p, long; the cross section thus shows a contrast in size, although this is less marked than in some of the preceding species. The cell-walls are thin and delicate. The photosynthetic branches in most cases arise from one edge of the axis and the stolons from the other, but stolons are occasionally borne on the photosynthetic branches in the basal part. The branches are sometimes so close together that they almost touch and sometimes as much as I mm. apart. The photosynthetic branches, when normally developed, are 1.5—3 mm. long, 0.3--0.6 mm. wide, and a little thinner than the axis. Their sides tend to be parallel and their apices are truncate or slightly emarginate. In many cases these branches are simple; in other cases they show a very few secondary branches, which are nearly as wide but usually much shorter. From the median region, which is only four cells thick, the branches gradually thin out toward the edges, where crenulate wings two or three cells wide are commonly devel- oped. The cell-structure is essentially the same as that of the axis (Fig. 4, E). The stolons are sometimes simple but often more or less branched, occasionally forming coralloid branch-systems. They are usually 0.1 mm. or less in width and show much the same structure as the stolons of X. varzadbilzs, the edges being windless and the contrast in width between the external and internal cells being slight. Rhizoids are fairly abundant on the stolons and other prostrate axes but are rare or absent elsewhere. The male and female branches are borne on distinct individuals. The male branches arise singly, usually from the prostrate axis (Fig. 4, B), more rarely from the base of an ascending branch (Fig. 4, C), still more rarely from a stolon. In most cases the inflorescence occupies the whole or the greater part of the branch, but occasionally a short flat proliferation is developed. The inflorescence is large, when the size of the plant is taken into account, the length being 0o.7—1.5 mm. and the width 0.25—035 mm. The structure, however, is remarkably simple (Fig. 4, F), the openings into the antheridial chambers being separated by a single row of cells. The antheridia are mostly ten to twenty, and the crenulate, erect or incurved wing, one cell high, scarcely bulges above the general surface. The female branch is borne on the prostrate axis or at the base of a 572 ALEXANDER W. EVANS photosynthetic branch (Fig. 4, A). It is very short, the number of archegonia being only two to four. The involucre (Fig. 4, G) consists of a series of irreg- ular laciniae or cilia mostly three to five cells long and often coalescent at the base. The calyptra, known only in a very young condition, is smooth or nearly so, and the corona is inconspicuous. Unfortunately the sporophyte is still unknown. In the preceding account 2. /eptostachya has been respeatedly compared with &. variabilis, and several points of agreement have been pointed out. The most important distinctions are perhaps the following: 2. leptostachya is on the whole a smaller plant; the photosynthetic branches do not broaden out toward the apex to any extent and are more definitely winged, the wings being distinctly crenulate; the male inflorescence is built up on a still more simple plan, the openings into the antheridial chambers being uniformly separated by single rows of cells; and the involucre of the female inflorescence is laciniate or ciliate rather than crenulate. There is no known species of the Chilean mainland with which R. /epfostachya has much in common. In &. aufozca, to be sure, there is a somewhat similar differentiation into axis, stolons and photo- synthetic branches, but this species is much more robust than XX. /eptostachya, it shows almost no contrast in width between the external and internal cells of the thallus, and the openings into the antheridial chambers are separated by three or four rows of cells. Area of distribution: Endemic. #13. R. nudimitra (Steph.) Evans, Trans. Connecticut Acad. 25: 177 f. It (1921). — Syn. Aneura nudimitra Steph. Sp. Hepat. 6: 35 (1917). On decayed wood. Masafuera: ravine near the Campo Correspondencia, 1150 m. (no. 43). Since R. xudimztra has recently been described and figured by the writer it will be sufficient here to compare it with R&R. /eptostachya, the only Juan Fer- nandez species that it resembles in any marked degree. In XX. nudzmitra the differentiation of the thallus into branches or axes of various types is still better marked than in &. /eftostachya, although the creeping stem still retains its power of curving away from the substratum and developing into a photosynthetic system similar to those formed by the erect or ascending branches. The turf- like appearance produced by the photosynthetic branches is very pronounced, and the branches are far more rigid and opaque than those of KX. leptostachya. The rigidity is due partially to the smaller size of the cells, the surface-cells averaging about 30 x 20», and the interior cells averaging only 25 » in width. The cell-walls, moreover, tend to be more strongly thickened, especially those of the interior cells, and a vaguely defined strand of sclerenchymatous tissue is sometimes discernible. This strand is more pronounced in the basal portions of the photosynthetic branch-systems and in the prostrate axis, where it is in the form of a compressed hollow cylinder with thin-walled cells inside. Area of distribution: South Chile; Masafuera. THE THALLOSE HEPATICAE OF THE JUAN FERNANDEZ ISLANDS 573 Metzgeria Raddi. *14. M. decrescens Steph. Bull. Herb. Boissier 7: 932 (1899); Evans, Proc. mmr. cicad. 58: 279 f. 2 (1923). On rocks in ravines and in beds of streams, 1000 m. and above. Masafuera: along a brook, 1000 m., SKOTTSBERG 1908 (no. 29, cited by STEPHANI, p. 10, as MZ. chilensis Steph.); ravine near the Campo Correspondencia, 1150 m. (nos. 44, 45); the High Plateau, 1100 m. (no. 46). The specimens here referred to M/. decrescens are by no means typical, and some of them diverge rather strikingly from the Chilean material recently reported upon by the writer. These divergent specimens, however, are con- nected by insensible intergradations with less aberrant forms, and the latter could hardly be distinguished from some of the less robust Chilean plants. There seems to be no reason, therefore, for attempting to separate the Juan Fernandez specimens specifically. The two plants that approach typical J/. decrescens most closely are no. 44 from the vicinity of the Correspondencia Camp and the plant from the 1908 collection, determined as J7. chzlensis by STEPHANI. In these the thalli are distinctly convex, as they should be, the marginal hairs are borne singly, and the thallus is otherwise naked, except for a few scattered hairs on the ventral surface of the costa. Ventral vegetative branches, however, are occa- sionally present, and the costae are relatively simple, the rows of cortical cells rarely exceeding four either dorsally or ventrally. Although no ventral branches except those with sexual organs have as yet been demonstrated in the Chilean material, it would be unwise to state definitely that they are constantly absent; and costae as simple as those of the Juan Fernandez plants are often to be found on poorly developed thalli or on some of the branches of robust thalli. No. 45 from the vicinity of the Correspondencia camp and the specimen from the High Plateau differ in two more important respects, many of the thalli being plane or nearly so and the ventral surface of the costa often bearing densely crowded hairs. Even here, however, a few branches are distinctly con- vex and parts of the costae are nearly or quite free from hairs. On the whole the plants with plane thalli present the appearance of being poorly developed, and yet their costae often approach in complexity those of robust Chilean specimens. It should be added that the Juan Fernandez plants often show the gradual transition between the unistratose wings and the costa, so characteristic of MW. decrescens. When the writer discussed this species on the basis of the Chilean ma- terial no female branches were found except a few too badly disintegrated to be described. No. 45, cited above, shows female branches in considerable abundance. They are more or less strongly concave and show the usual ob- cordate outline, although the apical sinus is sometimes very shallow. Their length is mostly 0.4;—0.; mm. and their width 0.;—0.6 mm. The marginal hairs, which are borne singly, are sometimes separated by one or two cells, but often a. series of hairs will adjoin one another, each representing the projection 574 ALEXANDER W. EVANS of a marginal cell. The narrow wings have a width of three to five cells and are naked, but the thickened median portion is often more or less densely pilose. The hairs are straight or nearly so. Area of distribution: South Chile; Masafuera. #15. M. decipiens (Massal.) Schiffn. & Gottsche, in SCHIFFNER, Forsch- ungsreise »Gazelle» 4*: 43 (1890); Evans, Proc. Amer. Acad. 58: 296 f. 6, 7 (1923). — MM. furcata B decipiens Massal. Nuovo Gior. Bot. Ital. 17: 256 pl. 28 f. 36 (1885); J. glaberrima Steph. Bull. Herb. Boissier 7: 939 (1899). On bark of trees. Masatierra: V. Colonial, Q. seca, on Fagara, 435 m. (no. 47); Q. de la Choza, on trunks, 250 m. (no, 48). Masafuera: Q. del Mono, on base of M/yrceugenia, 440 m, (no. 49). The characteristics of this widely distributed and variable antarctic species have recently been discussed by the writer. It is distinguished by its plane or slightly convex thallus, by the frequent occurrence of ventral branches, by its costa bounded both above and below by two rows of cortical cells, by a lack of hairs on the surface of the wings, by its marginal hairs usually borne singly and often scantily produced, by its dioicous inflorescence, and by a total lack of gemmae. It shows, moreover, a marked tendency to revert to more juvenile conditions, narrow unistratose conditions being not uncommon. The Juan Fer- nandez specimens agree closely with those from Chile; the material from Masa- tierra seems to be entirely sterile, but the Masafuera plants show both male and female branches in considerable abundance, a few of the latter bearing calyptras with well-advanced sporophytes. Area of distribution: South Chile; Australia; New Zealand; Masafuera. *16. M. multiformis nov. spec. — Fig. 5. Pale green, often turning whitish upon drying, scattered or growing in thin depressed mats, loosely adherent to the substratum or free; thallus pros- trate, sparingly dichotomous and frequently branching ventrally, plane or very slightly convex, well-developed thalli mostly 1—1.5 cm. long and 0.9—I.2 mm. wide, the forks usually I cm. or more apart but sometimes only 2—6 mm.; costa bounded both dorsally and ventrally by two rows of cortical cells, often poorly developed or absent altogether; wings mostly twelve to sixteen cells broad, the cells averaging about 43 X 37, thin-walled throughout or with minute and indistinct trigones; hairs restricted to the margin and ventral sur- face of costa, varying greatly in abundance and sometimes very sparingly pro- duced, marginal hairs borne singly, mostly 0.1.—0.3 mm. long and 12—18 p. wide, often slightly displaced to the ventral surface and sometimes branched at the tip and acting as rhizoids, straight or irregularly curved or contorted: inflorescence dioicous: g branches not seen: 2 branches sometimes fairly abundant, broadly obcordate, mostly 0 25—0.3 mm. long and 0.33;—0.45 mm. wide, slightly concave (from above), hairs abundant along the margin and representing the prolongation of marginal cells, otherwise scanty and restricted to the ventral surface of the THE THALLOSE HEPATICAE OF THE JUAN FERNANDEZ ISLANDS 57 on thickened median portion: calyptra and sporophyte not seen: gemmae marginal, borne on undifferentiated and sometimes ecostate thalli, the length variable but often about 1 cm., the width mostly 0.2z—0O.5; mm., or ten to sixteen cells; costa often absent but sometimes present in a more or less rudimentary form; hairs, if present, few and restricted to the margin. On rocks and ferns. Masatierra: Co. Centinela, on Alechnum in thickets, 550 m., and on trees, 530 m. {no. 50, 51); C. Salsipuedes, on Z7hyrsopterts in Dicksonia forest, 660 m. (no. 52). Masafuera: innermost accessible part of Q. de Jas Casas, on rocks, 200 m. (no. 53, 54); Q. del Blindado, on rocks in forest, 440 m. (no. 55); C. del Barril, on dead ferns (no. 56). No. 50 may be designated the type. As its name implies J/. multiformzs is an exceedingly variable species, and it owes its variability in great part to its strong tendency to assume re- versionary or juvenile forms. In this respect it resembles the northern J/. furcata (L.) Dumort. in certain of its manifestations. Although some of the thalli of M. multiformis may be costate throughout (Fig. 5, A), others show a well- developed costa only here and there and pass gradually or abruptly into a wholly ecostate condition (Fig. 5, B, C) or into a condition where the costa is only partially differentiated and indistinct. The ecostate thalli are usually narrow but this is by no means necessarily the case. Some of the examples studied, in fact, have been thirty cells across and have thus attained a width of I mm. or more. Thalli of this character occasionally fork but seem to show no indi- cations whatever of dorsiventrality; in other words the ventral surface lacks appendicular organs completely, not even slime-papillae being present. One peculiar manifestation of reversion is shown in Fig. 5, E, which represents a female branch with unfertilized archegonia renewing its growth as a vegetative thallus, The marginal gemmae of J/. mu/tiformis, which are sometimes produced in considerable abundance, arise in no definite order. The gemmiparous thalli show no evidence of being limited in their growth by the presence of the gemmae and are not morphologically differentiated in any way. They may be costate or ecostate (Fig. 5, F. G), and the thalli with costae sometimes produce female branches and gemmae at the same time, showing that neither type of organ inhibits the appearance of the other. The gemmae arise directly from marginal cells which have not undergone a preliminary division. They broaden out gradually from a stalk-like base and apparently never branch before becoming detached; in other respects they are much like ordinary vegetative thalli and may be distinctly or indistinctly costate or else entirely ecostate, one condition passing readily into another. As a rule an ecostate thallus gives rise to ecostate gemmae, but exceptions occur; and a costate thallus may bear ecostate and costate gemmae indiscriminately. The gemmae with costae sometimes show a few ventral slime-papillae and a slight displacement of the marginal hairs to the ventral surface, but these are the only evidences of dorsiventrality to be observed. Ecostate gemmae seem to be alike on both surfaces, and the same 576 ALEXANDER W. EVANS Oi! CRIT SQA Fig. 5. Metzgeria multiformis Evans. A. Part of a thallus, showing a distinct costa, X 50. B. Ecostate portion of a broad thallus, * 50. C. Part of a narrow thallus, showing a reversion from a costate to a ecostate condition, 50. D. Female branch, X 100. E. Female branch, showing a short vegetative proliferation, xX 100. F. Ecostate gemma still attached to parent thallus, X 100. G. Part of a gemma, showing change from a costate to an ecostate condition, X 100. B was drawn from a Masa- fuera specimen (no. 55), and the remaining figures from the type-specimen. thing is true of many costate gemmae, especially if the costae are rudimentary. The gemmae usually remain attached until they are 1 cm. or more in length; their width and the average measurements of their cells are somewhat less than in the case of ordinary thalli. No cases of germination have been observed. THE THALLOSE HEPATICAE OF THE JUAN FERNANDEZ ISLANDS 577 The relationship between J/. mu/ltiformts and the preceding species is remarkably close. Both are of about the same size, both are dioicous, both have a costa of the same type, and both are essentially alike in the distribution of their marginal and costal hairs. Both species, moreover, show a tendency toward reversion, although this tendency is even more strongly manifested in M. multiformis than in MW. decipiens. On the whole the texture of the thallus is more delicate in J7. mu/tiformzs than in WW. decipiens, the cell measurements averaging a trifle higher, but this distinction is of relatively slight significance. The most important distinction between the species has to do with gemmae: in JZ. multiformis these are usually present and may be fairly abundant; in M. decipiens gemmae never occur. It is of course difficult to establish the constancy of a negative character, such as the absence of gemmae. The state- ment just made, however, may be assumed to be correct, since it is based on a long series of specimens of J/. decipiens from Chile, where JZ. multiformis has not yet been found. Another gemmiparous species of Metzgeria, also closely allied to J. dec?- piens, is M. epiphylla Evans, recently described and figured from material collected in Chile." In this species the gemmiparous branches are often highly differentiated, terminating in blunt points with reduced wings and curving away from the substratum. The gemmae often arise very close together and are always small, ecostate bodies, rarely more than four cells across. These features will at once distinguish J/. epzphylla from MW. multiformis. Area of distribution: Endemic. *17. M. violacea (Ach.) Dumort. Recueil d’Obs. sur les Jung. 26 (1835); EvANs, Proc. Amer. Acad. 58: 306, f. 9 (1923). — Syn Fungermanunia violacea Ach.; WEBER & Monr, Beitr. Naturk. 1; 76, pl. 1, f. 1—-3 (1805). On bark of trees and similar substrata. Masafuera: Cordon del Barril, on Drzmys Wenterz (no. 57). This species has recently been discussed by the writer, so that only its more important features will be alluded to here. The costa in the vegetative thalli is almost constantly bounded both dorsally and ventrally by two rows of cortical cells, the species agreeing in this important respect with JZ. decipiens and J. multiformis. It is characterized by a bluish or purplish coloration appearing after death in dried herbarium specimens and by an abundant pro- duction of marginal and submarginal gemmae, borne on more or less specialized branches. The latter, if typically developed, curve away from the substratum and are further distinguished by their narrow. wings and by their quickly limited growth. Most of the Masafuera specimens do not show these extremes of modification, but a few of the gemmiparous branches are typical and indicate that the determination is correct. The coloration is clearly marked in many of the thalli, although the majority scarcely show it. Area of distribution: New Zealand; South Chile and the neighboring parts of Argentina; Masafuera. 1 Proc. Amer. Acad. 58: 303, f. 8 (1923). 578 ALEXANDER W EVANS Hymenophytum Dumort. 18. H. flabellatum (Labill.) Dumort. Recueil d’Obs. sur les Jung. 25 (1835); EvaANs, Bull. Torrey Club 52: 504 (1925). — Syn. Fungermannia flabel- lata Labill. Nov. Holland. Spec. Pl. 2: 109 (1806); Hymenophytum pedicellatum Steph. Kungl. Svenska Vet.-Akad. Handl. 46°: 11 f. 2a (1911); Symphyogyna entegervima Steph. op. cit. 13, f. 2e (1911). On rocks in ravines and in beds of streams. Masafuera: bank of a brook, 1000 m., SKOTTSBERG 1908 (no. 43, cited by STEPHANI, 13, as Symphyogyna integerrima; cited also, 14, as Symphyogyna sp.); Q. del Mono, narrow, dark ravine, 570 m. (no. 58); Q. de las Casas, moist rocks near the brook, 200 m. (no. 59); Q. del Blindado, moist rocks in forest, 440 m. (no. 60); brook just below Campo Correspondencia, 1120 m. (no. 61); the High Plateau, bed of small stream, 1100 m. (no. 62). Area of distribution: Australia; Tasmania; New Zealand; New Caledonia; Fiji Islands; Colombia; Chile; Masafuera. Symphyogyna Nees & Mont. 19. S. circinata Nees & Mont. Ann. Sci. Nat. Bot. II. 5: 18 (1836); EVANS, Trans. Connecticut Acad. 28: 338 (1927). — Syn. S. canaliculata Steph. Mém. Herb. Boissier 11: 336. (1900), in part; S. decumbens Steph. Kungl. Svenska Vet.-Akad. Handl. 46°: 12 (1911). On rocks and trees. Masatierra: without definite locality, BERTERO (1830), type; Pangal, water- fall, outside the rush, 205 m. (no. 63); Pangal, on stones in the spray, 205 m. (no. 64); high ridge south of Pangal, on tree ferns (no. 65); S. slope of Co. Chumacera, Boehmeria forest (no. 66). Masafuera: Q. de las Casas, SKOTTSBERG, 1908 (no. 42), cited by STEPHANI, p. 14, as Symphyogyna sp.; same locality, 1908 (no. 42 a); cave near beach, near ©. de las Casas (no. 67); Q. de las Vacas (no. 68). ~ Area of distribution: Chile; Juan Fernandez. 20. S. Hochstetteri Nees & Mont. Ann. Sci. Nat. Bot. Il. 5: 17 (1835); EVANS, Trans. Connecticut Acad. 27: 21 (1925). On rocks, stones, and shaded soil. Masatierra: without definite localities, BERTERO, 1830, and MOSELEY, Challenger Expedition. Porto Frances, on earth and stones in forest, 500 m. (no. 69); Q. Damajuana, waterfall, 248 m. (no. 70, 71); Anson Valley, near Plazoleta, on stones (no. 72); El Yunque, N. slope, on stones, 400—500 m. (no. 73); B. Cumberland, Pico Central, 390 m. (no. 74); B. Cumberland, caves IV, V, and VI, mostly on the walls (no. 75, 76, 77, 78, 79); Portezuelo, ravine on THE THALLOSE HEPATICAE OF THE JUAN FERNANDEZ ISLANDS 579 SW. slope, among Dicksonia, 550 m. (no. 80); Q. Gutierrez, 500 m. slope, among Dicksonia, 550 m. (no. 80); QO. Guiterrez, 500 m. (no. 81). *Masafuera: Q. del Blindado, moist rocks in forest, 440 m. (no. 82). Area of distribution: Juan Fernandez; Falkland Islands. 21. S. Hymenophyllum (Hook.) Mont. & Nees in G. L. & N. Syn. Hep. 480 (1846). — Syn. Fungermannia Hymenophyllum Hook. Musc. Exot. pl. 14 (1818); Symphyogyna chiloénsis Steph. Kungl. Svenska Vet.-Akad. Handl. 46°; teerout)s S. /amceclaza Steph. op. cit. 13. (191X). On earth, especially in shade. Masatierra: on bare earth, base of El] Yunque, SKOTTSBERG, 1908 (no. 45, type of S. /anceolata Steph.); ridge between Q. Piedra agujeriada and Q. Laura, 650 m. (no. 83); Portezuelo, ravine on SW. slope, among Decksonza, 550 m. (no. 83 a); C. Salsipuedes, Decksonza forest, 660 m. (no. 84). *Masafuera: vicinity of the Campo Correspondencia, 1100—1150 m. (no. 85, 86, 87, 88, 89). In the dendroid species of Symphyogyna, which are typified by S. Hymeno- phyllum, the differentiation of the thallus is usually much more clearly marked than in the decumbent species, to which S. czrcinata and S. Hochstetterz belong. The prostrate portion of the thallus consists of a wingless, terete rhizome, densely covered with rhizoids. This curves upward and forms an erect and more or less elongate stipe, destitute of rhizoids, which broadens out into the characteristic, broadly winged, aerial portion of the thallus. The latter is sometimes simple but usually branches by dichotomy from one to five times. In typical cases the dichotomies are close together, the branches of each rank are subequal in length, and the thallus forms (in the more complex examples) a flat branch system, spreading more or less horizontally and showing a sub- circular outline. The ultimate branches of the system are limited in growth and not infrequently develop apical wings. Between the complex branch-system just described and the unbranched winged thallus, which usually departs but slightly from a vertical position, all possible gradations may be demonstrated. The conditions may be further complicated by the phenomenon of reversion, which sometimes takes place when an aerial thallus or one of its young branches comes in contact with the substratum. Under these circumstances the wings narrow and disappear, and the thallus continues its growth as a terete rhizome with rhizoids. Aside from the dichotomous branching exhibited by the aerial thalli, the rhizomes and the bases of the stipe usually give rise very sparingly to intercalary branches. These grow out horizontally in the form of new rhizomes, which in turn develop new aerial thalli. It this way extensive colonies of such thalli may be formed, more or less intimately connected subterraneously. The conditions just described are strikingly paralleled by Hymenophytum flabellatum, as already pointed out by GOEBEL and other writers. The dendroid species of Symphyogyna are in need of careful study, not only in the herbarium but also in the field and under cultural conditions. When the descriptions given in the literature are compared it will be seen that some 580 ALEXANDER W. EVANS of the species, at least, are separated by unsatisfactory and often inconstant characters. This is particularly true of those drawn from the alar teeth and from the female involucres. In the writer’s experience the teeth, even on a single thallus, may vary greatly in both size and shape, and yet differences no greater than these variations have been utilized as differential characters. In the case of the female involucres, where the lobes may be few and broad or numerous and narrow, similar variations have been utilized. In all probability the result of more intensive study would be the reduction of certain widely recognized species to synonymy. Area of distribution: New Zealand; Australia; Tasmania; Chile; Juan Fernandez; probably tropical America. Pallavicinia S. F. Gray. *22. P. xiphoides (Tayl.) Steph. Mém. Herb. Boissier 11; 314 (1900). — Diplolaena xiphoides Tay\. Jour. Bot. 3: 569 (1844). — Fig. 6. On soil. Masafuera: Q. del Mono, on the ground, 570 m. (no. 90); Q. de las Casas, 200 m. (no. QI, 92). The specimens from Q@. del Mono agree in all essential respects with New Zealand specimens of P. azphordes in the Mitten Herbarium. Those from QO. de las Casas, however, tend to be more robust and to show a more gradual thinning of the thallus in passing from the midrib to the unistratose portions of the wings. At first sight these differences might appear to be specific in character, and yet they might equally well come within the normal range of variation of a single species. Since the specimens are completely sterile it seems advisable to adopt the second view. An individual plant of P. a7phozdes shows the usual differentiation into a wingless rhizome and a winged thallus with a distinct midrib. The rhizome, which is slightly flattened, is 0.183—0.35 mm. in width, and the surface may be either smooth or more or less densely radiculose. Except at the very base the rhizome is traversed by a slender strand of tracheid-like elements with brown walls; the strand is about 30 yu. in diameter and the individual cells about 5 pu. On the outside an epidermal layer of thin-walled cells is present. The cells average about 12. in width, and their slightly bulging walls are usually more or less tinged with brown. The rest of the rhizome is composed of from four to seven layers of cells, averaging about 20y, in diameter and having thin, colorless walls. The rhizome varies greatly in length; at one extreme it may be scarcely 1 mm. long and at the other may attain a length of I.5 cm. It gradually broadens out into the winged thallus, which soon curves or bends upward and assumes an ascending position. The winged thallus is either free from rhizoids throughout or sparingly radiculose in the basal portion. The length may be as much as 3.5 cm. and the width is 1—3 mm.; an individual thallus, however, rarely varies in width to any great extent. The thalli sometimes branch once or even twice by THE THALLOSE HEPATICAE OF THE JUAN FERNANDEZ ISLANDS 581 dichotomy but often do not branch at all by this method. In normal cases the growth is limited, although no examples of true apical wings have been observed. At the same time a reversionary process is not infrequent, whereby a thallus becomes narrower or may even continue its growth as a wingless rhizome. In addition to the forking of the winged thallus, intercalary branching may occur. The intercalary branches arise very sparingly either from the rhizome or from the ventral surface of the winged thallus in the lower part. They easily become detached, thus forming new independent plants and representing a type Fig. 6. Pallavicinia xiphoides (Tayl.) Steph. A. Transverse section of a midrib, showing strand of tracheidal tissue and parts below, X 225. B. Transition-region between midrib and unistratose portion of wing, X 225. C, D. Outer portions of the same wing in transverse sections, X 225. E—I. Marginal teeth, x 100. A—G were drawn from no. 91 and the remaining figures from a New Zealand specimen. of vegetative reproduction. In this way extensive mats of plants may be produced. The strand of tracheidal tissue in the rhizome extends into the midrib of the winged thallus, where it sometimes increases to 40 ». in width. In the case of a dichotomy the forking of the thallus is preceded by a forking of the strand; but, in the case of an. intercalary branch, the strand arises de ovo and is not continuous with the strand of the original thallus. This condition, which of course facilitates the separation of the branch, was first noted by LEITGEB? in the case of P. Lye/lit (Hook.) S. F. Gray and is clearly shown in CAMPBELL * Untersuch, aber Leberm. 3: 80 (1877). 5382 ALEXANDER W. EVANS and WILLIAMS’ figure of P. Zollingeri (Gottsche) Schiffn.’ It is found not only in the genus Val/avicintza but also in the related genera Symphyogyna and Hymenophytum. 3 The other tissues of the rhizome are likewise continuous with the midrib of the winged thallus. This midrib is distinctly convex ventrally and slightly so dorsally. It is bounded on each surface (Fig. 6, A) by a layer of green cells with bulging and slightly thickened walls, averaging about 13 y. in width. The cells between these surface-layers and the median strand are usually in five or six layers and average about 26», in diameter. They are pale or colorless and store starch in greater or less abundance, and their walls are either very thin throughout or slightly thickened in the outer layers. In typical cases the midrib merges very gradually into the wings, although the transition may be more abrupt. The unistratose portions of the wings vary in consequence from a width of only two or three cells (Fig. 6, C) to a width of twelve to fifteen cells, and marked differences in width are often to be observed in different parts of a single thallus (Cf. Fig. 6, D, with Fig. 6, C.) In cases where the transition is gradual, the internal cells in the transition- region are often distinguished by their large size (Fig. 6, B) and may attain a width of 60y. or more. These cells show very conspicuously in preparations cleared with potash, and the marked contrast in size between them and the surface cells can be easily demonstrated by focussing. Sometimes, however, as shown in Fig. 6, C, the distinction between superficial and internal cells tends to break down. The cells in the unistratose portion of the wings are variable in size but average about 30 in diameter. An occasional marginal cell may be longer and narrower, measuring perhaps 40X15 yp, but cells of this type are so in- constant that they have but little taxonomic significance. The walls of the alar cells are either thin or slightly and uniformly thickened. At fairly regular in- tervals along the margin the wings develop characteristic appendages in the form of slender teeth (Fig. 6, E—I). These teeth are mostly three or four cells long and two to four cells wide at the base. Except in the lower part, where a gradual transition into the unistratose wings is to be observed, the cells of the teeth are distinguished by their size, a length of g0>—120 y. not being un- common. The walls of the elongate cells are usually more or less thickened, although this feature it not shown in the figures. The teeth often lie parallel with the wing-margin or spread slightly, but they are sometimes curved inward and may be more or less appressed to the dorsal or ventral surface of the wings. GOEBEL” has shown that the marginal teeth of P. «7phozdes develop in very young segments and that they arch over the apical region and may even be more or less interwoven, thus forming a protective arrangement. This is supplemented by the abundant development of two-celled slime-papillae on both surfaces of the thallus. It should be noted, however, that the teeth themselves do not secrete slime and that none of the papillae are marginal in position. 1 Morph. study of some members of the genus Pa//avicinia fig. 2, B (1914). ? Flora 96: 184. f. 137 (1906). THE THALLOSE HEPATICAE OF THE JUAN FERNANDEZ ISLANDS 583 In this respect the species differs from P. Lye/iz and certain other members of the genus with »entire» thalli, in which marginal papillae are present. Area of uistribution: New Zealand; Masafuera. Monocleaceae. Monoclea Hook. *23. Monoclea Forsteri Hook. Musc. Exot. pl. 174 (1820). On moist rocks. Masatierra: Co. Damajuana, moist cliffs (no. 93); Q. Damajuana, water- fall, 248 m. (no. 94); Q. Gutierrez, 300 m. (no. 95); Cordén Salsipuedes, Drck- sonia forest, 660 m. (no. 96); Q. de la Choza, 250 m. (no. 97). Masafuera: Q. del Mono, soil in forest, 450 m. (no. 99); Q. de las Casas, damp rocks at bottom (no. 100). Area of distribution: New Zealand; Chile; Juan Fernandez. Pelliaceae. Noteroclada Tayl. *24. N. confluens Tayl. Jour. Bot. 3: 478 (1844), — Syn. Androcryphia porphyrorhiza Nees, G. L. & N. Syn. Hep. 470 (1846), in part; Moteroclada leucorhiza Spruce, Trans. & Proc. Bot. Soc. [Edinburgh] 15: 530 (1885). Masafuera: bed of small stream near the Campo Correspondencia 1100 m. (no. IOI). A full account of this interesting species has been published by SCHIFFNER.! Area of distribution: Mexico; Columbia—Chile; Brazil; Argentina; Masafuera; Falkland Islands. Fossombronia Raddi. 25. KF. fernandeziensis Steph. Kungl. Svenska Vet.-Akad. Handl. 46°: ES (LOe 1). Masatierra: vicinity of the colony, on earth, SKOTTSBERG 1908 (no. IT, cited by STEPHANI). Area of distribution: Endemic. 1 Osterreich. Bot. Zeitschr. 61: 325—332 (1911). 38 — 30570. The Nat. Hist. of Juan Fernandez and Easter Ist. Vol. II. 584 ALEXANDER. W. EVANS Anthocerotales. Anthocerotaceae. Anthoceros L. 26. Anthoceros Skottsbergii Steph. Kungl. Svenska Vet.-Akad. Handl. AG": (TOLL), in patt- On moist rocks and soil. Masatierra: B. Cumberland, vicinity of the colony, SKOTTSBERG, 1908 (no: .2); cited by STEPHANI, p. 91;-V.- Anson, near Plazoleta (no. 102) sem Cumberland, moist cliffs near the beach (no. 103, 104, 105); Portezuelo de Villagra, 590 m. (no. 106); along path to Portezuelo, on clay (no. 107). *Masafuera: Q. del Mono, narrow gorge on ground, 570 m. (no. 108); along path between QO de las Chozas and Q. de las Casas, 450 m. (No. Ico, 110); Q. de las Casas, innermost accessible part, 200 m. (no. III); ravine near the Campo Correspondencia, 1150 m. (no. I12). STEPHANI based his A. Skottsberg7z on two specimens, the first collected by SKOTTSBERG on the island of Masatierra and the seeond by the same col- lector on the island of Chiloé. These two specimens, however, belong to two different species, the Masatierra plant being characterized by pale yellow spores, which are not hispid, and the Chiloé plant by dark spores, which are distinctly hispid. In the writer’s opinion the name A. Skottsbergizz should be restricted to the Masatierra plant because, in the original description, the spores are described as »pallidae, leves», and because this plant is the one first mentioned. Unfortunately this opinion is at variance with the view held by STEPHANI. Five years after his original description was published he again proposed A. Skotisbergit as a new species’ but this time based it on the Chiloé plant alone and described the spores as »fuscae, hispidae». In this description he makes no mention of the Masatierra plant. His Species Hepaticarum, in fact, which is supposed to include the Hepaticae of the entire world, does not include a single representative of the Anthocerotales from Juan Fernandez. Since the original A. Skottsbergdi was an aggregate of two species it might be argued that the author of the species would be justified in restricting the name to either one of its component parts. In the present instance, however, the writer feels that he was not justified in restricting the name to the Chiloé plant, be- cause this plant does not agree with the original description. It would have been better if he had given the Chiloé plant a new name. The spores of A. Skottsberg7?, as here restricted, are pale yellow and are mostly 40—48 wu. in diameter, the average being about 45». The entire surface is covered over with very minute, punctulate markings, and the spherical sur- face bears in addition a series. of low ridges. These, in their most typical development, are in two concentric series. The outer series, which is near the periphery of the face, consists of about twelve ridges, in the form of crescents or horse-shoes, with their convexities directed inwards: the inner series of two or three longer ridges, forming a more or less definite, interrupted circle. But 1 Sp. Hepat. 5: 1000 (1916). THE THALLOSE HEPATICAE OF THE JUAN FERNANDEZ ISLANDS 585 the ridges are subject to a good deal of variation. Those of the outer series, for example, may be irregular in arrangement, some being nearer the periphery than others; they may vary in the degree of their curvature, some being slightly curved or almost straight; and their convexities may be turned in various di- rections, instead of being all directed inwards. In the inner series the ridges, instead of forming a circle, may have wavy outlines or turn their convexities inward. In some cases the ridges are fewer, and this reduction in number may be associated with further irregularities. The ridges on the spores of A. Skottsberga represent an unusual feature. STEPHANI, in this writings, makes no mention of spore-markings of this type in Anthoceros, and apparently the only species with which such markings have been associated is A. Pearsont M. A. Howe of western North America.’ In the northern species, however, the spores are only 36 vw. in diameter, and the ridges are fewer, shorter, and grouped together in the central portion of the spherical face, leaving a broad peripheral zone free from ridges. In their arrangement, moreover, the ridges show marked irregularity. The pale brownish pseudo-elaters of A. Skottsbergiz are mostly 50-—250y. in length and 6—12y, in diameter. They are often variously bent or contorted and are sometimes branched, their cells vary in number from one to four, and their slightly thickened outer walls are often wavy. ; Area of distribution: Endemic. Megaceros Campbell. *27. Megaceros fuegiensis Steph. Kungl. Svenska Vet.-Akad, Handl. 46°: gt (1911). On moist rocks and soil. Masatierra: Q. Damajuana, waterfall in forest, 248 m. (no. 113). Masafuera: ©. de las Casas, among stones, innermost accessible part, 248 m. (no. 114). The determination of these Juan Fernandez specimens as JZ. fuegzensis must be regarded as provisional. The genus Megaceros is in need of careful revision. Except in the case of a few Indo-Malayan representatives, which have been thoroughly described by SCHIFFNER, CAMPBELL, and GOEBEL, our know!l- edge of the species is exceedingly incomplete, and it is rarely possible to determine a specimen by means of the published descriptions alone. In the present case two courses were open: first, to make the specimens the type of a new species; or, second, to refer them to a published species, even if they did not agree with the descriptions in every detail. Since the specimens, although fertile, are not abundant, the second course seemed preferable. It should perhaps be noted that STEPHANI did not include M7. fwegiens7s in his Species Hepat- icarum, published in 1916. The only species that he lists from southern South America is MM. endiviaefolius (Mont.) Steph., reported simply from »Fretum magellanicum.» This species is more robust than the specimens listed above and is characterized by the lacerate and crispate margins of the thallus. 1 Bull. Torrey Club 25: 8 pl. 322, 323 (1898). 536 ALEXANDER W. EVANS In the Juan Fernandez specimens the thallus, which is about 3 cm. long, is irregularly lobed, the lobes being rounded and measuring 5—7 mm. in width. Sometimes the margin shows irregular sinuations or vague tooth-like outgrowths, those best developed being three or four cells long and three to eight cells wide at the base. Except for the outgrowths, which may be only one cell thick, the thallus lacks a unistratose border. As in all known species of Megaceros the thallus is solid and shows no intercellular spaces filled with slime. It is bounded above and below by a layer of small cells, about 20 ». in width, and the much larger internal cells are in about eight layers in the thickest portions. These internal cells are mostly 60—100¥y. in width and 50—70 p, in height. The cells are everywhere thin-walled and colorless, but some of the internal walls show a delicately pitted condition, the pits being irregular but showing a tendency to be narrowly fusiform and parallel; the walls thus acquire a reticulate appearance. The chloroplasts, in number and in distribution, agree with GOEBEL’s descriptions. The involucre, which sometimes flares at the mouth, is mostly 4—6 mm. long and 1 mm. wide. The capsule is 2—2.5 cm. long and about 0.5 mm. in diameter. The spores are pale yellowish green and average about 25 pw in greatest width. The whole surface is very minutely punctulate, and the spher- ical face’ bears in addition twelve to sixteen verrucae in the form of short cylinders or truncate cones, about 2, in diameter and 2. or less in height. The elaters are mostly one-, two-, or three-celled and are 60—200 ». long by 8, in diameter. The single spiral band is yellowish brown in color and about 8 y. in width. The spores of the Juan Fernandez specimens are almost exactly like those of the West Indian JW. vincentzanus (Lehm. & Lindenb.) Campbell and agree equally well with those of M. cristisporus Steph. of Costa Rica. STEPHANI, in fact, describes several other tropical American species with similar spores. M. vincentianus is distinguished by a minutely lacerate thalline margin, JZ. cristisporus by a shorter involucre, and the other species (according to the de- scriptions) by slight differences of one sort or another. In JZ. fuegzenszs, which STEPHANI records from Tierra del Fuego only, the spores are said to be 36u. in diameter, but the description agrees in most other respects with the Juan Fernandez plants. Area of distribution: Tierra del Fuego; Juan Fernandez. The following numbers represent sterile Anthocerotaceae, which are not in a condition to be determined: — Masatierra: waterfall in Pangal, 205 m. (no. 115, 116); V. Anson, near Plazoleta, on stones (no. I17). Masafuera: Q. de las Casas, innermost accessible part, 200 m. (no. 118); Q. de la Loberia, in water (no. 119). In all probability these specimens, if sporophytes were present, would be referable to one of the preceding species. Printed *4/11 1030. 21. The Codiums of the Juan Fernandez Islands. By WILLIAM ALBERT SETCHELL. With plates 34—48. The Codiums collected by Professor Dr. CARL SKOTTSBERG in the Juan Fernandez Islands, coming as they do, from a distinct as well as a remote marine geographic province, necessarily raise the question as to what members of this fairly extensive and varied genus have been found either in the islands themselves or on the western coasts of South America. So far as general data of surface temperature is concerned, both Masatierra and Masafuera lie between the isotheres of 15° and 20° C. and between the isocrymes of 10° and 15° C. Undoubtedly there are many disturbances of temperature due to currents, both general and local (wind currents particularly) and the usual inter- ference in shallow waters due to greater or less insolation and air temperature influence. There may be expected on the Juan Fernandez coasts, then, a variety of zonal expression, typical of 10—20° C. amplitude for outer waters, with tolerance of less regularity or constancy in other temperatures. From Juan Fernandez, there seems to have been reported only one Codium, »C. tomentosum», collected by H. N. MOSELEY during the visit of the Challenger Expedition. GEORGE DICKIE, who makes the report (Jour. Linn. Soc., Bot. 15:453, Dec. 1876), adds: — »a slightly branched variety 2—3 feet long». The name »/omentosum» has, of course, been widely, and often very loosely, applied, almost to each and every either strictly or nearly cylindrical dichotomous Codium. The note of DICKIE, however, calls attention to characters which indicate the possibility of this Challenger plant being associated rather with the series of forms clustering around C. decorticatum (Woodw.) M. A. Howe. Examination of the Challenger specimens (in Herb. Kew!) fully confirms this, and the exact identification will be considered later. There are very few references to Codium along the west shores of South America. The U. S. Exploring Expedition of 1838—1842, under command of Capt. WILKES, U. S. N., collected one at Valparaiso, Chile, which is recorded (see BAILEY and W. H. Harvey, U. S. Exploring Expedition, 17 (Algae): 165, 1874) under the »blanket» name of »Codium adhaerens». SVEDELIUS (in Sv. Exp. till Magellanslanderna, 3, no. 8: 299—304, pl. 16, fig. 1, pl. 17, figs. I4—19, 1900) lists three (3) species from Tierra del Fuego and West Pata- gonia, under the names: C. mucronatum var. californicum J. Ag., C. contractum Kjellm., and C. dimorphum Sved. Finally, MARSHALL A. HOoweE, in his 39—30579. The nat. Hist. of Juan Fernandez and Easter Isl. Vol. IT. 588 WILLIAM ALBERT SETCHELL »Marine Algae of Peru» (Mem. Torrey Bot. Club, 15: 43—46, figs. g—14, pls. 9, 10, 1914), has described and figured two (2) species of Codzum from Peru: — C. tomentosum var. peruvianum M. A. Howe and C. foveolatum M. A. Howe. O. C. SCHMIDT, in his contributions towards a knowledge of the genus Codzum (Bibliotheca Botanica, Heft. 91: 22, 1923) seems to regard the Pacific coasts of the Americas as inhabited by endemic species, outside of the very generally distributed C. fragile (Sur.) Hariot. SCHMIDT does not add any reference to species beyond those given above. It is to be assumed, then, that the plants referred to the various species by the various authors, as outlined above, are all that are known to occur on the Pacific Coast of South America and with which immediate comparison should be made. Certainly none of these species, with the sole exception of C. fragz/e, is at all closely related to those of the northern or to those of the tropical Pacific, nor do any of these seem, with the exception above noted, to indicate any intimate relationship with the New Zealand-Australian Codiums. There are some eleven (11) numbers of Codiums, collected by Professor Dr. CARL SKOTTSBERG on the islands of Masatierra and Masafuera. Three (3) of these are of the cushion type (Zy/ecodium) while the remaining eight (8) are elongated and dichotomously branched (Schzzocodium). The Tylecodiums are in excellent condition, except that they are sterile, but the Schizocodiums are almost entirely incomplete, not showing well the desirable distinctive differ- ences in habit. Subgenus Tylecodium Setchell. » (ulecodium>, in WO H:'S. "LUCAS, Proc..Linn. Soc:;"N“S: WarGezee 1935 (with suggestion by Lucas of preferable orthography!). The cushion Codiums differ not only in habit from the truly branched Codiums (Schzzocodium), but also in utricular morphology, their utricles being typically branched while those of the species of Schzzocodium are simple. Of Tylecodiums, the first South American collection was made at Valparaiso, Chile, by the U. S. Exploring Expedition, and was listed by Bailey and Harvey (loc. cit.) as Codium adhaerens; the second from Puerto Melinca, in the Guiatecas Islands of West Patagonia, named C. dimorphum by SVEDELIUS (loc. cit.). SKOTTSBERG made a third collection, represented by 3 numbers from Masa- tierra, completing the record of Tylecodiums for the west coast of South America, so far as known up to the time of writing. Specimens of all three collections have been examined and it seems best to regard them as represen- tative of at least two different species. Codium cerebriforme sp. nov. Plate 34, figures 1—3, and plate 39, figures 1—5. \ Thallus hemispherico-pulvinatus, cerebriformi-convolutus, 4 cm. hor.-diam., usque ad 2 cm. altus, colore?, infero per areas parvas adhaerens; lobis thalli e THE CODIUMS OF THE JUAN FERNANDEZ ISLANDS £50 basibus paucis pluribusve, elongatis et ipsis dactyliformi convolutoque sinuosis, verticaliter compressis et ipsos sinuoso arcte complectentibus, thallum totum cerebriformem efficientibus; utriculis mox dense ramosis, ramulis pyramidaliter oriendis positisque, macro-steno-physis, usque ad 650—725 ». longis et ad 60—75 wu. latis, apice fornicatis, nunc rotundis, nunc obtuse conicis, membrana apicali tenue usque ad 5 y, crassa, levi, sub apicibus plus minusve et curto tumescentibus et piliferis, pilis in verticillis irregularibus singulis usque ad plu- ribus positis; gametangiis nondum visis. Masatierra: No. 2, beach drift, Jan. 1917; No. 5, Bahia del Padre, lower littoral, Station 2 B, Jan. 25, 1917; No. 9, Santa Clara Strait, in 20—25 meters depth, on Nullipore bottom, Mar. 27, 1917. (Type specimen, No. 2, Herb. Univ. Calif. No. 549414). As may be seen from the figures (plate 34, figures 1—3), the habit, while varying somewhat in the three specimens available for study, still is cerebriform pulvinate through the radiating, sinuous, compressed lobes which are closely appressed together (especially before being handled), and the sinuous curves fitting into one another to form the brain-like habit. The differences, probably one of greater or less luxuriance in development, are chiefly in the matter of convexity of the pulvinus, No. 2 being the more rounded, while Nos. 5 and 9g are more depressed and flattened out. The last two approach the hahit of C. dimorphum (see SVEDELIUS, loc. cit.), but both from the figure as well as from the description of Svedelius, the character of the »lobes» seems different, those of C. dimorphum being shorter and more segregated from one another. In other words, there seems to be a discontinuity of the lobes of C. dimorphum and a greater elongated sinuosity in C. cerebriforme. The two species are undoubtedly closely related but, so far as the present acquaintance- ship with them suffices, presumably entirely distinct. Of the distinction between the two, this seems to be indicated not only by the variation in general habit, but is supported by the not too great, but still positive, differences in the utricular structure. The utricles of Codzum cerebriforme (see plate 39, figures I—5) are of the compound or ramose type. A single primary utricle, usually very long and increasing in diameter with age, gives off at successive heights along its sides, spirally arranged, secondary utricles or utricular branches, and this goes on until very complex groups (such as are partially represented in our figures 4 and 5 on plate 39) are built up. In such groups of 7ylecodium, the primary utricle may be recognized, at least early, by its greater length, greater diameter, and usually by its more profuse development of hairs (often fallen and represented only by their basal scars). In larger (and older) utricle groups, the earlier branches may assume the form and may possibly vie with the original primary utricle in dimensions and branching. Such primary, and even pseudoprimary, utricles are the ones to be described and measured as typical of the species. The later utricular branches are usually shorter, more slender, less modified at the apex and, especially, in the complexity of the hair-forming zones. They are ever to be considered as more juvenile and are, of course, much less typical of the species, often indeed much like those of nearly related species. 590 * WILLIAM ALBERT SETCHELL SVEDELIUS (loc. cit.) does not mention hairs on the utricles of his Codzum dimorphum, nor does he figure them. Examination of a portion of his type material fails to show hairs present even on the primary fruiting utricles. It may be considered, therefore, that hairs are lacking in C. dzmorphum Sved. In all three specimens of C. cerebriforme hairs (or hair-scars) are prominent on the primary and pseudo-primary utricles at a fairly early stage in the deve- lopment of the utricle fascicles (see plate 39, figures 2—5). This fact, together with the possibly less striking differences in habit, justifies the separation of the Juan Fernandez plants from those of Melinca in the Guiatecas Islands. The lack of the galeate thickened tip, so characteristic of the primary utricles of C. difforme, may further serve to distinguish them. <.1,, diani. Plate 37. Codium fernandezianum Setchell. 11. Fragment No. 3. 1 diam. (f ps7lophysalicum Setchell). 12 » No. 6. X ¥ diam. 33 > No. 10. X1 diam. (f. pszlophysalicum Setchell). THE CODIUMS OF THE JUAN FERNANDEZ ISLANDS SOQ Plate 38. Codium fernandezianum Setchell. 14. Fragment No. 11 A. X1 diam. TES » Nona < Te diam: Codium unilaterale {. Skottsbergianum Setchell. 16. Jype, No. 8. X06 diam. Plate 39. Codium cerebriforme Setchell. Utricles and utricle groups of Nos. 2 (figs. 1, 2), 5 (figs. 3, 4), and 9 (fig. 5). 95 diam. Plate 40. Codium dimorphum Sved. f.? Utricles from the Valparaiso specimen collected by the U. S. Expl, Exp’n, 1838—1842 (figs. 6—8). X95 diam. Plate Al. Codium fernandezianum Setchell. \ ) Utricles from No. 7 A (figs. 9—13), mostly stenophyse with hairs and gametangia. > 95 diam. (except that fig. 10 is 63.33 +). Plate 42. Codium fernandezianum Setchell. Utricles from No. 1 (figs. 14—18) and tip of one of No. 4B (fig. 19) showing mostly (rapidly matured?) utricles of differing dimensions, largely a brachyphyse form (with hairs and gametangia). X95 diam. Plate 435. Codium fernandezianum Setchell. Well developed, but younger (figs. 20—23) utricles, typically megaphyse, from No. 4 A (with hairs, except fig. 23, and gametangia). X95 diam. Plate 44. Codium fernandezianum Setchell. Utricle variation in No. 11 A (figs. 24—27, with hairs and gametangia), X95 diam. Plate 45. Codium fernandezianum Setchell. Various utricles of restricted size and mostly abnormal development from the short, lateral proliferations of No. 7 B (figs. 28—36, mostly with hairs). > 95 diam. ‘ 600 WILLIAM ALBERT SETCHELL Plate 46. Codium fernandezianum Setchell. Four (4) utricles, one (fig. 37) of the broader type (macro-eury-physe) the other three (figs 38—40) of the narrower type (macro-steno-physe), all from No. 6, (with hairs and game- tangia). X95 diam. Plate 47. Codium fernandezianum {. psilophysalicum Setchell. Broader (fig. 41) and narrower (fig. 42, 43) utricles from No. 3 (hairless or psilophyse, but with gametangia). 95 diam. Plate 48. Codium fernandezianum {. pstlophysalicum Setchell. Utricles of various sizes (figs. 44—49) from No. 10 (all hairless or psilophyse). X95 diam. Printed "le “7p37: Nat. Hist. Fuan Fernandes and Faster Isl. Vol. | UU Hy] ‘MILLIMETER, | Codium ilo. 9 Juan Mernandez ekottebearg ha “AALAIWNITIIN MILL IMETER. GCodium No. 2 WKOUTED STE UaR FL2rTNnani3szs AL. ee POCO OCC Ce Lereepeaeen 3 i Codium Wo. 5 Juan Fernand; 8KOGGE tc sPE th ie ee qty TT) HT tl i nT] i yadda | | | | MILLIMETER. Nat. Hast. Fuan Fernandes and Easter Isl. Vol. I. PEAT ESE MILLIMETER, | Nat. Host. Fuan Fernandes and Easter Isl. Vol. I, PLAT] 36 \ i | ") ee | qd °° 3 & O Oh = &-D 9 Q a= 8 yp iin iv at Se: at : een? a . rf - im i : : a 4 ie! : ,' ' » r Witt 7" " ame) v iS (uae el i cite ia RT i he: Le) Nat. Hist. Fuan Fernandes and Easter Ist. Vol. 1. PLATE 43. 22 20 21 Nat. Hest. Fuan Fernandes and Easter [sl. Vol. 11. PLATE 44. at Paes - LALWRAN foe v4 ia tis te Oe “e : ce: » 4 Ph abe Nat. Hist. Fuan Fernandes and Easter sl. Vol. 11. PLATE 45. 2 ~——s | 33 32 \- spe avant \\ Ao MA oe asides) sea Ana ‘ oe , > ion \ ! \ \ \ 8g Nat. Hist. Fuan Fernandes and Easter Isl. Vol. 11. PLATE 46. Nat. Hist. Fuan Fernandes and taster Ist. Vol. 11. ELATE) 47. Nat. Hist. Fuan Fernandes and taster Ist. Vol. 17. PLATE 48. Ly A? | 46 Y 48 49 Dae 397T—30579- 1a) an g - é Pa jid . whee ; eo 7 sy - 7 AS Sa we Dat Usha, win’. al ic Squirt i. m oy ye | : 22. Die Meeresalgen der Juan Fernandez-Inseln. Von TORE LEVRING. Mit Tafeln 49—53. Vorwort. Es war meine Absicht, meine Sammlung von Meeresalgen selbst zu be- arbeiten. Ein kleiner Teil der Ausbeute von 1908 habe ich auf Helgoland im Laboratorium von Professor Dr. P. KucKUCK wahrend des Sommers Ig11 be- stimmt. Mit der bedeutend grésseren Sammlung von 1917 konnte ich lange nichts anfangen, brachte schliesslich das gesamte Material nach Amerika mit, wo ich 1934—35 in New Haven, Conn. wohnte. Fast jede Woche konnte ich ein paar Tage in New York zubringen, wo Dr. M. A. HOWE mir einen Arbeits- platz im Institut des Botanischen Gartens zu Verfiigung stellte. Selbst hatte er die Absicht, die Kalkalgen zu bearbeiten. Spater hatte ich Gelegenheit, zwei Wochen im Botanischen Institut der Berkeley-Universitat zu arbeiten, wo ich die sehr vollstandige algologische Bibliothek und die grossen Sammlungen Pro- fessor SETCHELLS benutzen konnte. Durch seine Vermittlung wurden auch viele mikroskopische Praparate hergestellt. Ich bin ihm ftr seine vielseitige Unterstiitzung zu grossem Dank verpflichtet. Leider wurde es mir aber un- moglich, vor meiner Riickkehr nach Schweden die Arbeit abzuschliessen. Die meisten Braun- und Griinalgen waren mehr oder weniger fertig bearbeitet, der erdosste Teil der Rotalgen aber nicht. Meine Absicht, die Arbeit spater fortzu- setzen, habe ich nicht verwirklichen k6nnen. Glicklicherweise fand ich in Dr. TORE LEVRING einen ausgezeichneten Mitarbeiter. Das gesamte Material nebst meinen Aufzeichnungen und Praparaten wurde ihm zu Verfiigung ge- stellt. Ich sage ihm meinen herzlichen Dank fir sein Entgegenkommen und fiir die grosse Mie, die er sich gegeben hat, eine erschopfende Darstellung dieser interessanten Algenflora zu geben. Leider fehlen im Verzeichnis unten die Corallinaceen. Nach Dr. HOWES Tod wurden sie von Mme PAUL LEMOINE in Paris zur Bestimmung tber- nommen. Das weitere Schicksal dieser Sammlung ist mir aus wohlbekannten Griinden unbekannt geblieben. Goteborg im Mai 1941. C. Skottsberg. 40— 41438. The Nat. Hist. of Juan Fernandez and Eastern Isl. Vol. II. 602 TORE LEVRING Als Prof. Dr. C. SKOTTSBERG mich bat, ich modchte die die von ihm be- gonnene Bearbeitung der marinen Algen von Juan Fernandez wbernehmen, habe ich dies mit grosstem Interesse getan. Das Material, das grodsstenteils in Alkohol aufbewahrt ist, war, als ich es erhielt, schon aufsortiert und mehrere Arten waren mit kiirzeren oder langeren Notizen versehen, was mir die Arbeit in verschiedener Hinsicht erleichtert hat. Einige neue Arten hatte Prof. SKOTTS- BERG im Manuskript aufgestellt. Unsere Kenntnis der marinen Algenflora von Juan Fernandez war bisher sehr ltickenhaft. Es liegen nur einige altere Angaben von MONTAGNE (1835, 1852 a), der 3 Arten angibt, vor. Spater hat DICKIE (1877) die von MOSELEY wahrend der »Challenger»-Expedition gemachte Sammlung bearbeitet. Sein Verzeichnis enthalt 22 Arten, von denen aber mehrere unrichtig bestimmt sind, und das Material ist im Herbarium Kew aufbewahrt. Prof. SKOTTSBERG hat friiher einmal diese Sammlung durchgesehen und Praparate, die ich untersucht habe, davon gemacht. Wegen des Krieges war es mir nicht moglich, selbst das ganze Material zu leihen. Das folgende systematische Verzeichnis enthalt g2 Arten. Dazu kommen noch 4, die nur der Gattung nach bestimmt werden konnten. Es sind die Sammlungen SKOTTSBERGS aus den Jahren 1908 und 1917, einschliesslich der alteren Angaben MONTAGNES und DICKIES. Von einigen Arten DICKIES gab es aber kein Material in Kew. Von zweien habe ich auch kein anderes Material gesehen. Weil ich mich tber die Richtigkeit dieser Angaben nicht aussern kann, habe ich sie nur als Fussnoten angefuhrt. Ausser diesem Material habe ich eine kleine Sammlung von Prof. W. R. TAYLOR, Ann Arbor, Mich., leihweise erhalten, woftir ich ihm bestens danke. Sie enthalt die Ausbeute der HASSLER-Expedition und eine von WALDO L. SCHMITT im Jahre 1926 gemachte Sammlung. Von den Sammlungen SKOTTSBERGS ist nur die Funde von Splachnidium (SKOTTSBERG 1920), Codium (SETCHELL 1937) und Distromium (LEVRING 1940) verOffentlicht, ferner zum Teil die Corallina- ceen von 1908 (LEMOINE 1920). Die Kalkalgen habe ich nicht beriicksichtigen k6onnen (s. oben). Von den 92 Arten sind 20 Chlorophyceen, 25 Phaeophyceen und 47 Rhodophyceen. Die Juan-Fernandez-Inseln liegen etwa 650 km westwarts von Mittelchile. Langs der siidamerikanischen Westktiste zieht der kalte Humboldt-Strom nord- warts. Nach neueren Untersuchungen liegen die Inseln ausserhalb dieses Stro- mes in dem warmen sog. Mentor-Strom. Die Temperatur des Oberflachen- wassers diirfte 15—20° sein. Die Temperaturverhaltnisse sind ungefahr die- selben wie an der peruanischen Kite (vgl. SKOTTSBERG 1936). Es ware vielleicht eine grosse Ubereinstimmung der Flora der Inseln und derjenigen von Chile und jedenfalls Peru zu erwarten. Dies ist aber gar nicht den Fall. Die Zusammensetzung der Flora ist hdchst interessant. Von den 92 Arten sind 31 (=etwa 34 %) endemisch. Eine (Chaetomorpha firma) ist nur noch bei San Ambrosio (nérdlich von Juan Fernandez) gefunden. Etwa 19 (= 21 %) sind kosmopolitisch oder subkosmopolitisch und etwa 13 (= 14 %) sind in den tropischen und warmen Meeren der Erde weitverbreitet. Von den ubrigen sind 9 (= 10 %) australisch-neuseelandische Arten, wovon einige auch in Chile und Peru vorkommen, 7 (= 8 %) sind Peru-Chile-Arten (1 davon auch an der Westkiiste Nordamerikas). Weiter sind 6 (= 6,5 %) kalifornisch, 1 nord- DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 603 und mittelatlantisch, 2 westindisch, I aus den Gegenden des siidlichen Siid- amerika, I aus Japan und Ozeanien. Die letzten diirften zum Teil Arten sein, die bloss wegen mangelhafter Kenntnis bisher nur in diesen Gebieten gefunden sind. Interessant ist auch Sflachnidium mit dem Endophyt Codzolum Kuckuckiz, die eine siidhemispharisch-temperierte Verbreitung haben, aber in Siidamerika fehlen. Von diesen verschiedenen Gruppen ist also die endemische die grésste, was wohl durch die isolierte Lage der Inseln zu erklaren ist. Die Flora zeigt eine deutliche Verwandtschaft mit der neuseelandischen und australischen. Be- deutend geringer sind die Beziehungen zu der Flora Perus and Chiles, beson- ders wenn man bedenkt, dass die fiir diese Kiisten so bezeichnenden riesengros- sen Laminariaceen (Lessonza, Macrocystis) ganz fehlen. Der tropische Charakter der Flora ist bedeutend starker als derjenige der peruanischen. Es ist offenbar, dass die Griinde der Zusammensetzung der Flora zum grossen Teil historischer Natur sind; sie steht wohl mit dem Vorkommen ehemaliger Kiistenverbindungen im Zusammenhang, Die pflanzengeographische Zusammensetzung der marinen Algenflora zeigt auch eine interessante Ubereinstimmung mit der terrestrischen Flora. Statzonenverzeichnis der Sammlung SKOTTSBERGS. 1908. St. 36. MZasatierra, Bahia Cumberland, 22. Aug. Litoral, Blockufer mit Sand: a obere-innere Zone, bei Niederwasser trocken; b untere-dussere Zone, von der Brandung bespilt. St. 37. Masatierra, Dredschen ausserhalb Bahia Cumberland, 24. Aug. 30—35 m. Sand mit Steinen. St. 38. Zasatierra, Puerto Ingles, dstliche Landzunge, 25. Aug., Litoral, Felsenufer mit Tum- peln. a und b wie St. 36. St. 39. Masatierra, Bahia del Padre, 26. Aug. Litoral, Felsenufer mit Blécken: a und b wie St. 36. Tiimpel in beiden. Die Grenze zwischen a und b wurde dort gezogen, wo Splachnidium zu dominieren beginnt. St. 40. Masafuera, 28. Aug.: a litorales Blockufer, bei Niederwasser trocken; b angetriebene Algen aus dem unteren Teil der Litoral- und von der Sublitoralregion. 1917. St. 3. Masatierra, Sidkiiste am Fuss von El Yunque, 7. Jan, Felsen, auch bei Niederwasser Wellenspritzer. St. 2. Masatierra, Bahia dél Padre, 15. und 26. Jan. Felsen und grosse Blécke. a und b wie St. 36, c angetriebene. St. 3. Masatierra, Bahia del Padre, 15 Jan. Aussere, stark exponierte Felsen. Entspricht der unteren Litoral- und oberen Sublitoralregion. St. 4. Die Strasse zwischen Masatierra und Santa Clara, 27. Marz. Dredschen 20—25 m. Boden mit Lithothamnien, Steinen and SchalJen. St. 5. Masatierra, nicht weit von St. 4, dicht ausserhalb Carbajal, 27. Marz. Derselbe Boden, mit Kalksand. St. 6. Masatierra, ausserhalb Bahia Cumberland, 31. Marz, Dredschen 30—40 m. Steinen, Sand, Schalenkies mit Fragmenten von Corallinaceen. 7. Masatierra, nahe St. 6, 60>—70 m. Sand mit Steinen. 8. Masatierra, ausserhalb Pangal, 10. April. Dredschen 10—15 m, Sand mit Steinen. St. 9. Masatierra, ausserhalb Punta Carlos, 10. April. Dredschen 20—30 m. Felsboden. 0. Masatierva, auf einer Boje in der Bucht, 3. Jan. Mit Algen und Lepadiden. 604 TORE LEVRING St. 11. Masafuera, Bahia Tolten, am Nordende der Insel, 3. Febr. 30—40 m. St. 12. Masafwera, die ausserst stark exponierte Westkiste ausserhalb Loberia, 16. Febr. Ange- triebene sublitorale Algen. St. 13. Masafuera, ausserhalb Sanchez, 14. Marz. Mit Hummerkérben heraufgeholt. Chlorophyceae. Ordn. Chlorococcales. Fam. Chlorococcaceae. Codiolum Braun. Codiolum Kuckuckii Skottsb. et Levr. nov. sp. — Fig. 1 A. Cellula in Sfplachnidio rugoso endophytica, pyriformis, longe stipitata, ad c:a 90 ». longa, corpore pyriformi-ovoideo 20-—30 y, longo et 10—15(—18) p. crasso, stipite filiformi + obliquo obtuso vel acuto apice interdum incrassato, omnino cavo. Diese Art wurde schon von KUCKUCK (1929 S. 77) besprochen und ab- gebildet. Sie kommt endophytisch und wie es scheint konstant in Splachntz- dium rugosum vor. Man findet sie dort besonders in den Sprossspitzen, aber auch in jungen Konzeptakeln. Sie ist mit dem verjiingten Teil in den Thallus der Wirtspflanze tief eingesenkt. Von den umgebenden Geweben ist sie schon durch ihre Farbe getrennt. ICUCKUCK (l.c.) schreibt: »Es liegt hier ein ty- pisches Codzolum vor, das an seinen Wirt ganz ahnlich gebunden ist und in ihm mit derselben Regelmassigkeit anzutreffen ist, wie das sonderbare Codzolum Petrocelidis, das fast keinem Thallus von Petrocelzs fehlt...» Die in Splachut- dium vorkommende Codzolum-Art ist schon friiher in der Literatur behandelt worden (MITCHELL and WHITTING 1892 S. 3; ROE 1916 S. 401). Von diesen Verfassern wird das Codzolum als »Initial-(Apical-)zellen» von Splachnidium angesehen und auch gut abgebildet. SKOTTSBERG (1920 S. 281) lehnte diese Angaben ab und meinte, dass sie eine endophytische Alge sei. KUCKUCK hat dann endlich das Ratsel der Initialzellen von Splachnidium gelost. Von allen bekannten Arten der Gattung unterscheidet sich C. Kuckuckit durch den hohlen Stiel; sonst ist es ihnen ausserlich sehr ahnlich. Leider ist der Zellinhalt schlecht erhalten. Was gewisse Codzo/um-Arten betrifft, haben sie sich als Entwicklungsstadien von Uvosfora erwiesen (vgl. JORDE 1933 S. 10; LEVRING 1937 S. 16). Dies trifft fiir C. Awckucki zweifelsohne nicht zu. Codiolum Kuckuckii wird bis etwa 90». lang. Der obere, birnformige Teil ist etwa 20—30 » lang, 10—15(—18) yp breit und lang gestielt. Der hohle Stiel ist manchmal unten etwas erweitert. Der Form geht gut aus meinen Figuren und aus denen von SKOTTSBERG (I. c.) und KUCKUCK (I. ¢.; auch in OLTMANNS 1922 S. 52) hervor. In Splachnidium diirfte sie nie fehlen. Biunidio Giese ot), 2.b. 38.430 bp: Geogr. Verbr.: Siidafrika, Ostindien, Australien, Tasmanien, Neuseeland, Juan Fernandez, San Felix. DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 605 Ordn. Ulothricales. Fam. Ulvaceae. Enteromorpha Link. Enteromorpha intestinalis (L.) Link; Ahlner 1877 S. 15; J. Ag. 1882 S. 131; Setchell and Gardner 1920 S. 252. Das untersuchte Material stimmt am besten mit f. gexawzna Ahlner wber- ein. Es gibt auch Ubergangsformen zu f. cornucopiae (Lyngb.) Ahlner. Die Exemplare sind bis etwa 8 cm hoch, die Zellen etwa 9—16 yu. im Durchmesser und der Thallus bis etwa 35 vu. dick. Fertil. Fundorte: St. 1, 2b, 36a, 38a; DICKIE (1877 S. 453, leg. MOSELEY). Geogr. Verbr.: Weit verbreitet in kalten und temperierten Meeren. iva. L. Ulva rigida C. Ag. 1823 S. 410; J. Ag. 1882 S. 168; Setchell and Gardner 1920 S. 269; U. lactuca var. rigida Le Jolis 1863 S. 38; Phycoserts rigida Kitz. hah, 6 Taf. 23. Die meisten Exemplare sind ziemlich klein, rasenbildend und aus mehr oder weniger zerrissenen und gekrauselten Thallusscheiben bestehend. Der Thallus ist bis etwa 100 p dick, die Zellen in Flachenansicht 9—15 p. im Durch- messer und im Querschnitt 2—3 mal so hoch. — An exponierten Lokalitaten in der Litoralregion gewachsen. Fundorte: St. 1, 3, 12, 36a, b, 39b, 40a; SCHMITT Nr. 241 (8.12. 1926). Geogr. Verbr.: Weit verbreitet. Ulva linza L.; Setchell and Gardner 1920 S. 262; Levring 1937 5S. 18; Enteromorpha linza (L.) J. Ag. 1882 S. 134. Das Material stimmt in allen Teilen mit dem von mir an den skandi- navischen Kiisten gesehenen iiberein. Die Exemplare sind bis etwa II cm lang, 2 cm breit und ziemlich stark gekraduselt. Uber das Verhaltnis der Art zu gewissen Lyxteromorpha-Formen siehe meine friihere Darstellung (LEVRING 1937). —- Im unteren Teil der Litoralregion. Fundorte> St. 2’b;'c) 10. Geogr. Verbr.: Wahrscheinlich weitverbreitet. Ulva nematoidea Bory 1828 S. 190; Montagne 1844—46 S. 6; Phycoseris lobata Kiitz. 1849 S. 477; Tab. 6 Taf. 27; Ulva fasctata costaia Howe 1914 S. 20 Taf. 1, Taf. 2 Fig. 1o—23. — Fig. 1 B—D, Taf. 49 Fig. I. Von Ktrzinc (1849 S. 477) wird U. nematoidea zu U. fasciata gefuhrt. MONTAGNE (lI. c.) hebt aber hervor, dass die fiir Chile beschriebene U. xema- totdea von U. fasciata, die zuerst im Mittelmeer gefunden ist, wohl getrennt ist. Meines Erachtens spricht auch alles fiir die Richtigkeit dieser Auffassung. Im Herb. AGARDH in Lund liegt ein Exemplar von U. xematozdea (Nr. 14478), 606 TORE LEVRING das von D’URVILLE wahrend der Forschungsreise mit »La Coquille» in Chile erbeutet und von Bory bestimmt worden ist und somit zum Originalmaterial Borys gehért. Der Thallus ist bis zur Basis in lange, schmale, ziemlich schwach undulierte Zipfel geteilt. Ich meine nun, dass man U. (Phycoserzs) lobata (fir Chile beschrieben) und U. fasciata costata (ftir Peru beschrieben) auch zu U. xematoidea stellen kann. Wenn man ein groésseres Material von verschiedenen Lokalitaten unter- suchte, wiirde man bestimmt alle Ubergiange finden. U. /odata ist etwas breiter und weniger tief geteilt als die typische Form. Von SETCHELL und GARDNER (1920 S. 268; Phyc. Bor. Amer. Nr. 863) ist diese fiir Kalifornien angegeben, wenn sie nun wirklich mit der siidamerikanischen Form identisch ist. U. fasciata costata ist weniger verzweigt als die typische Form, aber starker unduliert. In der Sammlung von Juan Fernandez liegen nur ein paar Exemplare vor. Sie lassen sich hier gut einreihen. Das grdsste und am besten entwickelte ist etwa 12 cm lang und stellt etwa eine Zwischenform zwischen der typischen U. nematoidea und U. fasctata costata dar — reich zerteilt mit langen, stark undulierten Zipfeln. Anatomisch stimmen diese gut iiberein. In Flachenansicht liegen die Zellen ohne Ordnung. Sie sind g—20 yp. im Durchmesser. Im Querschnitt sind die Zellen am Rande oder in den jiingeren Teilen des Thallus fast quadratisch, sonst etwas vertikal ausgezogen, bis 3mal so hoch wie breit. Der Thallus ist 40—90 p. dick. U. fasciata ist fiir verschiedene Teile der Welt angegeben. Die typische Art gehért dem Mittelmeergebiet. Was anderen solchen Formen angeht, so stehen sie wohl der echten U. fasciata mehr oder weniger nahe. Gegenwartig finde ich es aber am richtigsten, die in Chile, Peru und Juan Fernandez vor- kommenden Formen unter U. zematoidea zusammenzufihren. Sie gehoren ohne Zweifel zusammen und diirften nur von dem Standort beeinflusste Formen sein. Mit U. fasciata dirfte sie verwandt, jedoch nicht identisch sein. — Auf expo- nierten Felsen im unteren Teil der Litoralregion gefunden. PIMGo rues St.12 G3. Geogr. Verbr.: Chile, Peru, Juan Fernandez; Kalifornien? Ulva reticulata Forsk.; J. Ag. 1882 S. 166; Okamura Icon. 2 S. 182 Taf. 100; Phycoseris reticulata (Forsk.) Kitz. 1849 S. 478; Tab. 6 Taf. 29. — Fig. 1 E—F. Das Material besteht nur aus ein paar Exemplaren. Das grosste ist etwa 36 cm lang, ungefahr ‘/2mal so breit und unregelmassig zerrissen, mit zahl- reichen grossen und unzahligen kleinen Lochern. Der Thallus ist etwa 60 up. dick. In Flachenansicht sind die Zellen etwa 15—20 (selten bis 30) p. in Durch- messer. Im Querschnitt sind die fast quadratisch bis etwas rektangular. Das Material stimmt gut mit der Beschreibung und dem Herbarmaterial (u. a. von dem Originallokal, Rotes Meer), das ich gesehen habe, uberein. Ulva reticulata erinnert sehr an U. fenestrata Post. et Rupr. von der nordlichen Pazifikkiiste Amerikas (vgl. SETCHELL and GARDNER 1920 S. 267). Die Zellen sind aber hier kleiner, etwa 8—15 yp, im Durchmesser in vertikaler Richtung etwas ausgezogen, was bei U. reticulata kaum der Fall ist. DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 607 nanauannnnts oo LOL Fig. 1. A Coditolum Kuckuckit. B—D Ulva nematoidea. B—C Querschnitte (C Mitte des Thallus); D Flachenansicht. E—F Ulva reticulata. E Querschitt; F Flachenansicht. — A X 670; B—C xX 395; D—F xX 250. Eine andere nahe verwandte Art ist U. pertusa von Japan (KJELLMAN 1897 S. 4). KJELLMAN hebt gewisse Unterschiede zwischen den beiden Arten hervor, und von WEBER VAN BOSSE (1913 S. 51) werden beide fiir die ma- layische Region angegeben. Vor allem scheint der Thallus bei U. pertusa durch die Entwicklung der Locher nie netzformig wie bei U. retecu/a zu werden. Ich bin aber nicht ganz davon iberzeugt, dass sie artverschieden sind. Nach WITTROCK diirfte U. Jacinulata (Kiitz.) Wittr. (WITTR. et NORDST. Alg. exsicc. Nr. 433) ein jiingeres Stadium von U. retzculata sein. Die Zellen sind aber etwas kleiner als bei der typischen Form. Fundorte: St. 1; CROCKER Nr. 9825, Herb. Berkeley (Masatierra, Cum- berland Bay, 5—20 fathoms, Jan. 1931). Geogr. Verbr.: Rotes Meer, Indischer Ozean, Japan, Australien, West- kiiste von Nordamerika, Juan Fernandez. Ordn. Cladophorales. Fam. Boodleaceae. Microdictyon Dcne. Microdictyon japonicum Setchell 1925 S. 107; 1929 S. 528. Von dieser Art liegt ein ziemlich reiches Material vor. Die Bestimmung SKOTTSBERGS wurde von Prof. SETCHELL bestatigt. Er meint auch, dass die 608 TORE LEVRING 5 von BORGESEN (1924 S. 253) fiir die Osterinsel als 17. wmbilicatum angefihrte Art zu JZ. japonicum gehort. Die Exemplare sind etwa 1—3 cm hoch. Die Zellen der Hauptfaden sind 150—200 », dick und bis 500—1000 ws. lang, allmahlich kiirzer werdend; die der ausseren Verzweigungen sind etwa 40—60 yp, breit und 100—120 y, lang. — Auf Lithothamniev, Steinen usw. im unteren Teil der Litoralregion und in der Sublitoralregion bis in 45 m Tiefe gesammelt. Fundorte: St. 2, 4,5, 8, 13, 39 b; SeHmitT Nr. 133 (Santag@ieer 11.12. 1926, dét. SETCHELL), Nr. 149 (Cartbajal Bay; 15.12: 1926, WdebesmrE CHELL); Geogr. Verbr.: Japan, Malayische Inseln, Polynesien, Osterinsel, Juan Fernandez. Fam. Cladophoraceae. Chaetomorpha Kitz. Chaetomorpha linum (Mill.) Kitz. 1845 S. 204; Hauck 1885 S. 439; Hamel 1931 S. 30. Von dieser Art liegen mehrere Exemplare vor. Sie waren normal und bildeten loseliegende, verworrene Knauel. Die Faden sind etwa 130—240 p dick und die Zellen (1—)1,s—3mal so lang wie dick, zylindrisch. Fund ortes Stz 2:¢, (4,10; Ti; 12; Hassier-Exp. Nrs 67. Geogr. Verbr.: Temperierte und warme Meere.! Chaetomorpha aerea (Dillw.) Kiitz. 1849 S. 379; Tab. 3 Taf. 59; Hauck 1885 S. 438; Hamel 1931 S. 28. Zu dieser stelle ich einige kleinere Exemplare, die aber normal entwickelt sind. Von DICKIE (1877 S. 453) wird sie auch angegeben. Ich habe Material von seiner Sammlung gesehen. An den von mir untersuchten Exemplaren sind die Faden unten etwa 90 ». dick, oben allmahlich dicker bis etwa 280 wy. Die Zellen werden oft tonnenformig und sind 1—2mal so lang wie breit. Fundorte: St. 36; DICKIE (I. c., leg. MOSELEY). Geogr. Verbr.: Temperierte und warme Meere. Chaetomorpha antennina (Bory) Kitz. 1849 S. 379; Vickers 1908 S. 19 Taf. 7; Setchell and Gardner 1920 S. 203; Borgesen 1940 S. 37; Conferva an- tennina Bory; Howe 1914 S. 37; Chaetomorpha pacifica Kitz.; Howe 1. c.; Ch. pacifica (Lyon) Wille; Howe |. c.; Ch. media (C. Ag.) Kiitz.; Borgesen 1925 S. 37. Es dirfte jetzt sichergestellt sein, dass man die angefiihrten Synonyme zusammenstellen kann. Die Art ist etwas variabel. Sie wird etwa 4—10 cm 1 Chaetomorpha tortuosa Kitz. wird von Dickte (1877 S. 453) fir Juan Fernandez ange- geben (MosgELey). In Kew gibt es aber kein Material. Ich kann mich deshalb tiber die Richtig- keit dieser Angabe nicht dusseren. Die Art wird sonst fiir den Nordatlantik und das Mittelmeer zitiert. Nach BORGESEN (1925 S. 45) ist die mediterrane und kanarische Form von der echten Conferva tortuosa Dillw. spezifisch verschieden, DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 609 hoch. Die Faden sind etwa 450—600(- 900) yw. dick und die Zellen 1—4mal so lang. Gewohnlich sind sie etwa 500 y. dick; u. a. die Form von Mexiko (Ch. pacifica Lyon) war bis goo pw dick. Wie BORGESEN (1940 I. c) hervor- gehoben hat, ist die Lange der Basalzelle auch sehr variabel bei verschiedenen Formen, die sich aber nicht trennen lassen. Sie ist 3—4 bis 15—16 mm lang, gewohnlich 5—9 mm. Unten ist sie etwa 150 yu, oben 300—450 wu. dick. Die Exemplare von Juan Fernandez sind alle ziemlich klein und etwas dicker als normal. Sie sind bis 6 cm hoch. Die Faden sind gewohnlich 400— 650 w. dick, an den gréssten Exemplaren bis g00(—1000) ».. Die Zellen sind 1—3mal so lang wie breit. Die Basalzelle ist 5—8 mm lang, unten etwa 200 u, oben etwa 400 uw. dick. Ich finde es am besten, diese Form, obwohl es ziem- lich grobe Exemplare gibt, zu Ch. antennina zu stellen. Sie gehdrt zweifels- ohne zu dem Formenkreis dieser Art. Pundorte: St..2c;,\4, 36a, 39b. Geogr. Verbr.: Réunion, Westindien, Kanarische Inseln, Brasilien, Indien, Ceylon, Java, Hawaii, Westkiiste von Mexiko, Juan Fernandez. Chaetomorpha firma Levr. nov. sp. — Fig. 2 A—B. Caespites densi, filis crassis, saltem inferne rigidis, ad 6 cm longis, inferne 0,3 mm, superne I,; mm crassis. Cellulae cylindricae usque vage fusiformes, latitudine (I—)1,;—3-plo longiores. Cellula basalis circ. (1—)1,3;—2,5(— 3) cm longa. Die Art bildet auf Steinen und Lithothamnien dichte Biischel. Die Faden sind grob und, jedenfalls unten, steif. Sie werden bis 6 cm lang, unten etwa 0,3—I mm, oben allmahlich bis etwa 1,5 mm dick. Die Zellen sind zylindrisch bis schwach tonnenformig und (1I—)I,;—3mal so lang wie breit. Besonders be- zeichnend ist die grosse Basalzelle. Sie ist (1—)1,;—2,5(—3) cm lang, oben etwa I mm, unten 0,3—0,45 mm breit. Sie ist mit Rhizoiden versehen, und die Zellwande sind verstarkt, 25—40 uw. dick. Diese neue Art diirfte am nachsten mit Ch. robusta (Aresch.) Papenf., die in Siidafrika vorkommt und friiher als Ch. clavata wohlbekannt war, verwandt sein. Wie aber PAPENFUSS (1940 S. 200) gezeigt hat, ist die westindische Ch. clavata von C. AGARDH viel kleiner als die siidafrikanische, und die Arten sind auch sonst wohlgetrennt. Eine andere naheverwandte Art ist Ch. Dar- wind (Hook.) Kitz. von Australien, Tasmanien und Neuseeland. Bei Cz. ro- busta und Ch. Darwini finden wir aber nicht die riesengrosse Basalzelle, die fiir Ch. firma so bezeichnend ist. Weiter sind die Zellen bei der ersten starker tonnenformig und die Faden werden nach der Spitze hin auffallender dicker. Von Juan Fernandez liegt von dieser neuen Art ein gutes Material vor. Ich habe weiter einige Exemplare von der Insel San Ambrosio nordlich von Juan Fernandez gesehen, die mit ihr gut tibereinstimmen. Fir Chile gibt MONTAGNE (1852a S. 380) Ch. clavata an. Ob sie mit Ch. robusta, firma oder einer anderen Art identisch ist, lasst sich nur durch die Untersuchung der Exemplare feststellen. Fundorte: St. 3, 4. Geogr. Verbr.: Juan Fernandez, San Ambrosio. 610 TORE LEVRING Fig. 2, A—B Chaetomorpha firma. A Zwei Faden; B Basalteil. C—F Cladophora perpusilla. C Unterer, E—F Oberer, Teil. — A nat. Gr; B x 7; C, E—F X 65; D x asa: Cladophora Kiitz. Cladophora fascicularis (Mert.) Kitz. 1843 S. 268; Tab. 3 Taf. 90; Vickers 1908 S. 18 Taf. 13; Howe 1914 S. 34. Das Material ist sehr gering. Ein Exemplar (St. 8) scheint ein normales Aussehen zu haben. Die Verzweigung stimmt besonders gut mit den Figuren VICKERS (l.c.) iiberein. Die Zellen sind in den obersten Zweigen etwa 60— 80 ». dick und etwa 3—4mal so lang. Unten sind sie bis etwa 280 p, dick und etwa 4mal so lang. Von zwei anderen Einsammlungen ist die Ubereinstim- mung weniger gut. Eine (St. 5) hat weniger typische Verzweigung und die Zellen sind oben 70—100 », dick und 5—7(—9)mal so lang, unten bis 280 yp. dick und 1omal so lang. Ein Exemplar (St. 11) hat etwa normale Verzweigung und die Zellen sind oben 80—100 », dick und (3—)4—6mal so lang, unten bis 280 ». dick und 1omal so lang. Wie aber mehrere Verfasser hervorheben, ist die Art ziemlich variabel und umfasst zweifelsohne verschiedene Formen, die sich kaum trennen lassen. Nach Howe (I. c.) stimmt CZ. fascecularis von Peru weniger gut mit der westindischen Form iiberein. Die Juan Fernandez-Formen DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 611 diirften mehr an die von Peru und Chile (vgl. MONTAGNE 1852 S. 381) er- innern. Fundorte? St. 5, 6,’ 11. Geogr. Verbr.: Westindien, Brasilien, Kanarische Inseln, Rotes Meer, Mauritius, Juan Fernandez, Chile, Peru. Cladophora perpusilla Skottsb. et Levr. nov. sp. — Fig. 2 C—F. Thallus caespitosus, c:a 5—6 mm altus, dense ramosus, ramis pseudo- dichotomis vel alternis, raro oppositis, c:a 30—45 », basi ad 60. crassis. Rami ultimi 15—20(—25) ». crassi, obtusi. Cellulae 2,;—5-plo, inferne saepe ad I0-plo longiores quam latae, membranis tenuibus. Bildet kleine, etwa 5—7 mm hohe, dicht verzweigte Biischel. Faden unten bis etwa 60 ». dick, Hauptzweige fast dichotomisch oder seitlich, seltener oppo- niert verzweigt, etwa 30—45 ». dick. An den Verzweigungsstellen sind die Zellen oft in einer bezeichnenden Weise verwachsen, was man aber auch bei gewissen anderen Arten finden kann. Endverzweigungen gewohnlich einseitig gereiht, 15—20(—25) u. dick, etwas gebogen, mit abgerundeter Spitze. Die Zellen sind 2,;—5mal so lang wie breit. In den unteren Teilen werden die Zellen oft langer, bis etwa 10mal so lang. Zellwand diinn. Steht wahrscheinlich C/. mauritiana Kiitz., Cl. inserta Dickie u. a. nahe. Diese sind aber grésser und grober. Epiphytisch auf Splachnidium und Scytothamnus. Fertil. Fundort: St. 39 b. Geogr. Verbr.: Juan Fernandez. Ordn. Chaetophorales. Fam. Chaetophoraceae. Ectochaete (Huber) Wille. Ectochaete pacifica Levr. nov. sp. — Fig. 3 A—E. Thallus e filis endophyticis, in membranis hospitis (Chaetomorpha firma et Dictyota phlyctaenodes) repentibus, irregulariter ramosis, in partibus vetustioribus confluentibus compositus. Cellulae (7—)8—-12(—14) crassae, I—2(—3)-plo lon- giores; chromatophorum parietale, pyrenoideo unico (raro 2) instructum. Ich habe die Art reichlich in den dusseren Zellwanden von Chaetomorpha firma und Dictyota phlyctaenodes gefunden. Der Thallus besteht aus verzWeig- ten, endophytischen, bald zusammenfliessenden Faden, die ein fast zusammen- hangendes, gewodhnlich einschichtiges, sich stark verbreitendes Gewebe bilden. Gelegentlich kénnen sich in den Alteren Teilen einzelne Faden auch epiphytisch entwickeln. Die Zellen sind fast zylindrisch bis etwas kugelartig abgerundet, etwa (7—)8—12(—14) ». dick und 1—2(—3)mal so lang. Sie enthalten einen parietalen Chromatophor, welcher in der Regel nicht die ganze Zelle ausfullt und ein einziges Pyrenoid (selten zwei) enthalt. Hyaline Haare kommen reich- lich vor. Sporangien werden durch schwache Vergrosserung gewohniicher Thalluszellen gebildet. 612 TORE LEVRING Ohne Zweifel steht die Art E. leptochaete (Huber) Wille nahe. Bei dieser gibt es aber (I—)2—3 Pyrenoide. Weiter bildet /. pactfica kaum ein so dicht geschlossenes, pseudoparenchymatisches Gewebe, wie /epfiochaeze es tun kann. Punderte: St. 3; 4. Geogr. Verbr.: Juan Fernandez. Ectochaete ramosa Levr. nov. sp. — Fig. 3 F—K. Thallus e filis endophyticis, in membranis hospitis (Polyszphonta abscissa) repentibus, irregulariter ramosis compositus. Cellulae cylindricae, (7—)g—15 (—18) yw. crassae, I—3-plo longiores; chromatophorum parietale pyrenoideis instructum. Lebt in den Zellwainden von Polyszphonia abscissa. Der Thallus besteht aus endophytischen, verzweigten Faden. Ein parenchymatisches Gewebe wird nicht gebildet, sondern die Faden sind fast immer frei voneinander. Die Zellen sind fast zylindrisch, (7—)9—15(18) ». dick, 1—3mal so lang und enthalten einen parietalen Chromatophor mit 1—3 (in alteren Zellen gewohnlich 2—3) Pyre- noiden. Hyaline Haare sind vorhanden. Die Sporangien werden durch schwache Vergrosserung gewohnlicher Zellen gebildet. Diese Art ist der vorigen sehr ahnlich; sie unterscheiden sich aber in einigen Einzelheiten. Bei £. paczfica enthalten die Zellen in der Regel nur ein Pyrenoid. Weiter ist 4. ramosa etwas grosser als sowohl £. facifica als EL. leptochaete, und ihre Faden diirften sich in der Regel nicht pseudoparenchy- matisch zusammenschliessen. — In Polyszphonia abscissa zusammen mit £z/o- cladia viridis gefunden. Letztere wuchs dicht unter der Aussenmembran, wah- rend Lctochaete erst etwas tiefer zu finden war. Fundiort: St. 1. Geogr. Verbr.: Juan Fernandez. Entocladia Reinke. Entocladia viridis Reinke 1879 S. 476; Hauck 1885 S. 462; Kylin 1938 S. 67; Endoderma viridis (Reinke) Lagerheim 1883 S. 74. Kam endophytisch in Polysiphonia abscissa reichlich vor und ist der europaischen Form vollkommen gleich. Die Zellen sind 3—7 vy. breit und 1— 5mal so lang. Fertil. Pug e.o.et. St... 1 Geogr. Verbr.: Weit verbreitet. Ordn. Szphonales. Fam. Halicystidaceae. Halicystis Aresch. Halicystis pyriformis Levr. nov. sp. — Fig. 3 L—P. Cellula pyriformis, ad c:a 8 mm longa et 4 mm crassa, breviter stipitata, chromatophoro rotundato—fere oblongo, vage irregulari, 5—8 y. diametro, pyre- noideo instructo. DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 613 Fig. 3. A—E E£ctochaete pacifica. A—B, E in Flachenansicht; C—D von der Seite. F—K Ectochaete ramosa. YF—H, K in Flachenansicht; I von der Seite. L—P Hatlcystis pyriformis. L—O Habitus; P Chromatophoren. — A—K, P x 445; L—O X 2?/s. Bisher waren drei Hatcystis-Arten bekannt. //. ovalzs Aresch., die ty- pische Art der Gattung, kommt in Mittel- und Nordwesteuropa und an der Westkiiste von Nordamerika vor. Von MuRRAy (1893 S. 47), KUCKUCK (1907 S. 139) u.a. ist sie gut beschrieben worden. Die Zellen sind bis etwa 1 cm gross, oval oder kugelformig. Die Chromatophoren enthalten aber kein Pyre- noid. In diesem Zusammenhang verdienen auch die Untersuchungen KORN- MANNS (1938) iiber den Entwicklungskreis von HZ. ovalzs und Derbesza marina erwahnt zu werden. Eine andere europaische Art ist 7. farvula Schmitz. Sie wurde auch von Murray (l.c.) behandelt und von FELDMANN (1937 S. 76) eingehend beschrie- ben. Nach ihm kommt sie im Mittelmeer, an der baskischen Kiiste, den Kana- rischen Inseln und den Antillen vor. Wie bei der vorigen sind die Zellen sphdrisch bis oval, bis 5 mm gross. Die Chromatophoren sind 2—4mal so lang wie breit und enthalten ein Pyrenoid. Endlich ist 7. Osterhouti fir Bermuda von L. R. und A. H. BLINKS (1930) beschrieben worden. Diese Art wird bis iiber 3 cm gross. Die Chromatophoren sind mehr oder weniger langgestreckt und enthalten ein bis mehrere Pyrenoide. Das Halicystis-Material von Juan Fernandez ist sehr sparlich. Zweifels- ohne gehort es aber zu einer bisher unbekannten Art. 614 TORE LEVRING Die Zelle ist birnformig, bis 8 mm lang und 4 mm breit. Sie ist mit einem kurzen Stiel an der Unterlage (Corallinaceenkrusten) befestigt. Die Chromatophoren sind abgerundet bis schwach langgestreckt, etwas unregel- madssig und etwa 5—8 y. im Durchmesser, mit einem Pyrenoid. Von den oben erwahnten Arten unterscheidet sie sich vor allem im Bau und Aussehen der Chromatophoren. Bei /7. ovals fehlen die Pyrenoide. Die Chromatophoren sind bei den beiden anderen deutlich langgestreckt, bei /7. pyriformis fast abgerundet—kreisformig. 7. Osterhouti ist ausserdem noch viel grosser, An der Grenze zwischen Litoral- und Sublitoralregion gesammelt. Harn doit: Stes. . Geogr. Verbr.: Juan Fernandez. Fam. © oldiaceaie. Codium Stackh. Das Material dieser Gattung ist in einer besonderen Arbeit von W. A. SETCHELL (1937) beschrieben worden. Ich fihre deshalb hier nur die ange- troffenen Arten und ihre Fundorte an. Codium cerebriforme Setchell 1937 S. 588 Taf. 34 Fig. 1—3; Taf. 39 Fig. 1—5. Bund Ortersots 2c, 13594,030ib. Geogr. Verbr.: Juan Fernandez. Codium fernandezianum Setchell 1937 S. 592 Taf. 35—38 Fig. 5—15, Taf. 41—48 Fig. 9—49; C. tomentosum Dickie (non Huds.) 1877 S. 453. fe Wypicum Setchell' lc. S: 593: Bamidorte sot. 2.6. Lis, 12, 30.), 4o.b. Geogr. Verbr.: Juan Fernandez. f. pstlophysalicum Setchell |. c. S. 593. Mumedorte: St. 2c. 6. Geogr Verbr.: Juan heqnandez. Codium unilaterale Setchell et Gardner 1924 S. 710. f. Skottsbergzanum Setchell 1937 S. 596 Taf. 38 Fig. 16. Pumcdor te: ot. 12, 40 b: Geogr. Verbr.: Juan Fernandez (Hauptform von Kalifornien). DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 615 Phaeophyceae. Ordn. Lctocarpales. Fam. Eetocarpaceae, Ectocarpus (Lyngb.) Kjellm. Ectocarpus confervoides (Roth) Le Jolis 1863 S. 75. f. pygmaeus (Aresch.) Kjelim. 1890 S. 77; Saunders 1898 S. 154 Taf. 15 Fig. 5—9; Setchell and Gardner 1925 S. 415; Phyc. Bor. Amer. Nr. 525. Auf Lepas und Chondriella habe ich einen Lctocarpus gefunden, der mit dieser Form gut tbereinstimmt. Er bildet einen einige mm hohen Filz. Die Zellen der Faden sind etwa 12—30 yp dick und (1,;—)2—3(—5)mal so lang. Die plurilokularen Sporangien sind 20— 30 », dick und 60—100 y. lang. Pundioir tex Sts 10} 36:2. Geogr. Verbr.: Atlantischer Ozean, Mittelmeer, Siidafrika, Australien, Westkiiste von Amerika, Subantarkt. Inseln. Ectocarpus Mitchellae Harv. 1852 S. 142 Taf. 12G; Borgesen 1914 S. 162; 1939S. 75; £. virescens Thur.; Sauvageau 1896 S. 17; Borgesen 1926S. 18. Wie BORGESEN gezeigt hat, ist E. virescens mit L. Mitchellae identisch. Von Juan Fernandez liegt ein ziemlich reiches Material vor. Die Exemplare sind bis etwa 5 cm hoch und haben ein ganz normales Aussehen. Die Haupt- achsen sind etwa 40—60 y. dick und die plurilokularen Sporangien etwa 65— 100 y. lang und (15—)20—25(—30) » dick. Unilokulare habe ich nur selten gesehen; sie sind 43—60 p, lang und 20—25 yp, dick. Bemerkenswert ist, dass man bei dieser Art Haare mit weniger ausgepragter, basaler Zuwachszone fin- den kann. — In der Litoralregion auf Steinen und Sand in Tuimpeln und epi- phytisch auf Splachnidium, Codium cerebriforme, Corallina sp. u. a. Rairndiortes St52:bv3 7 ion 30a. Geogr. Verbr.: In den meisten warmeren Meeren. Ectocarpus granulosus (Smith) C. Ag.; Hauck 1885 S. 332; Kuckuck in Oltmanns 1922 S. 8; Setchell and Gardner 1925 S. 426. Stimmt mit der atlantischen Form sehr gut iiberein. Die meisten Exem- plare sind etwa 2 cm hoch, die Hauptachsen 45—100(—120) vu. dick und die plurilokularen Sporangien 50—80 y. lang und 30—45 p. breit. — In der Litoral- region auf Steinen oder auf Scytoszphon und Lepas gefunden. Fiundort esSt. 1op36'b: Geogr. Verb.: Nordatlantik, Mittelmeer, Kap der guten Hoffnung, West- kiiste von Nordamerika, Juan Fernandez, Neuseeland. Ectocarpus arabicus Fig. et De Not.; Borgesen 1937 S. 7; &. contferus Bors: 11914. /S.. 1645°1937 'S. 5 Auf Codium cerebriforme habe ich einige etwa 11,5 cm hohe Bischel von einem Lcfocarpus gefunden, der mit £. avadicus sehr gut iibereinstimmt. Die 616 TORE LEVRING Zellen sind 30-45 yp dick, ‘/2—4mal so lang, und enthalten mehrere kleine, scheibenformige Chromatophoren. Die Verzweigung und die Form der pluri- lokularen Sporangien, die 35—45 p. dick und 75-120 ». lang sind, sind sehr bezeichnend. Mit ihren basalen Rhizoiden dringt die Pflanze etwas in die lockere Unterlage ein. Wie BORGESEN (1937) gezeigt hat, soll die Art £. ara- dicus heissen. unidoris St. 3: Geogr. Verbr.: Westindien, Rotes Meer, Siidindien, Juan Fernandez. Ectocarpus pusillus Griff; Sauvageau 1895; Borgesen 1926 S. 30; Acinetospora pusilla (Griff.) Born. Stimmt mit der europdischen Form in allen Teilen tiberein. Es liegt nur ein sparliches Material vor. Die Exemplare, die epiphytisch auf Chaetomorpha wachsen, sind ein paar mm hoch. Die aufrechten Faden sind 20— 0 yp. (ge- wohnlich etwa 25—30 p) dick und die Zellen 1—4mal so lang. Die Ver- zweigung ist ganz normal. Die plurilokularen Sporangien sind etwa 30—50 p. dick und 60—100(—120) y. lang. Punidomt: St 30s): Geogr. Verbr.: England, Frankreich, Adriatisches Meer, Kanarische Inseln, Juan Fernandez. Ectocarpus minutissimus Skottsb. et Levr. nov. sp. — Fig. 4. Pars basalis e filis repentibus orta, cellulis reticulatis, 7—12(—15) u. crassis, 2—3-plo longioribus. Fila erecta simplicia, ad 0,;—1 mm longa, e cellulis cy- lindricis, 12—16 y. crassis, 2—4-plo longioribus quam latis formata. Chromato- phorum discoideum. Sporangia plurilocularia lateralia, stipite 1—2cellulari suf- fulta, vel terminalia, 4o—60(—75) p. longa, 18—24 u. crassa. Der Basalteil besteht aus kriechenden, verzweigten Faden, die auf der Unterlage (Dictyota phlyctaenodes) ein Netzwerk bilden. Eine zusammenhangende Basalscheibe kommt aber nicht zustande. Die Zellen dieser Faden sind etwa 7-—12(—15) yu. dick und 2—3mal so lang. Daraus wachsen aufrechte, gewohn- lich einfache Zweige, mit Zellen, welche zylindrisch, 12—16 pw. dick und 2— 4mal so lang sind, auf. Die meisten dieser Zweige sind ziemlich kurz, andere werden bis 0,;—I mm hoch und kénnen haarformig ausgezogen sein, in welchem Falle die Zellen an der Basis kiirzer und gegen die Spitze hin langer sind. Die Zellen enthalten mehrere kleine, scheibenformige Chromatophoren. Die plurilokularen Sporangien sind eiformig—konisch, 40—60(—75) p. lang und 18—24 , dick, und sitzen an den Zweigen seitlich, in der Regel auf einem 1— 2-zelligen Stiel, oder terminal. Von dieser Art liegt ein ziemlich reiches Mate- rial vor. Sie ahnelt 4. Batterszz, welcher aber in allen Teilen groésser ist. Fundort: St. 4. Geogr. Verbr.: Juan Fernandez. Ectocarpus cylindricus Saunders 1898 S. 150; Setchell and Gardner 1925 S. 432, f. typzcus Setchell and Gardner 1922 S. 415. Auf Microdictyon japonicum habe ich einen Ectocarpus gefunden, der mit dieser Art gut iibereinstimmt. Die Buschel sind bis etwa 1,5; mm hoch. Die DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN O17 Fig. 4. “ctocarpus minutisstmus. — A X 395; B—D xX 150. Zellen der aufrechten Faden sind etwa 18—28 y» breit und 2—3 (in den mitt- leren Teilen 0,;—1) mal so lang. Die plurilokularen Sporangien sind etwa 60— 95 w lang und 30—45 wu. breit. Pundort: St. 39 b. Geogr. Verbr.: Kalifornien, Juan Fernandez. f. codtophitlus Setchell et Gardner 1922 S. 415; 1925 S. 433. Diese Form kam auf Codium fernandesianum vor. Sie bildet kleine, ein paar mm hohe Biischel. Die endophytischen Faden dringen tief in die Wirts- pflanze ein. Die aufrechten Faden sind nur unten verzweigt und hier 25— 33 p. dick, nach oben allmahlich diinner werdend. Die Zellen sind etwa I—2,s, im oberen Teil bis 4mal so lang wie breit. Die plurilokularen Sporangien sitzen an der Basis der freien Faden und sind wenigstens 100, lang und 45 y. breit. Poamcdort: ot. 2 C= 12. Geogr. Verbr.: Westkiiste von U.S. A., Juan Fernandez. Ectocarpus hemisphericus Saunders 1898 S. 151; Setchell and Gardner 1925 S. 434. Zu diesem stelle ich einen Lctocarpus, welchen ich auf Splachnzdium und Laurencia gefunden habe. Die Ubereinstimmung mit dem von SAUNDERS ge- sammelten Exsickatexemplar (Phyc. Bor. Amer. Nr. 528) ist gut. Die Bischel sind bis etwa 2 mm hoch. Die Faden sind unten 20—27 yp. dick, oben allmah- lich diinner werdend. Die Zellen sind 1—3mal so lang wie breit. Die pluri- 41 — 41438. The Nat. Hist. of Juan Fernandez and Easter Isl. Vol. II. 618 TORE LEVRING lokularen Sporangien sind etwa 15—34(— 40) p. dick und 50—go0 yw. lang. In seiner Beschreibung gibt SAUNDERS die Breite mit 14~ 20 p, an. Nach seinen Exemplaren zu urteilen ist sie ebenso gross wie bei meinen. Die Art diirfte der vorigen sehr nahe stehen. Pundort: ot 30 -b: Geogr. Verbr.: Kalifornien, Juan Fernandez. Geminocarpus Skottsb. Geminocarpus geminatus (Hook. f. et Harv.) Skottsb. 1907 S. 13; Ectocarpus geminatus Hook. f. et Harv. 1845 S. 251; 1847 S. 469; Askenasy 1885 5.16) lati 5, Figg, OF 760: Von dieser Art wurden in zwei Proben einige Fragmente gefunden, einige zusammen mit Sfphacelaria cirrosa, andere auf Codium cerebriforme und Gra- teloupia subsimplex. Das Aussehen ist normal. Sie trugen plurilokulare Spor- angien. Fundorte:>St3, 30:b; 40. Geogr. Verbr.: Siidliches Siidamerika, Falkland-Inseln, Stidgeorgien, Kerguelen, Victorialand. Compsonema Kuck. Compsonema ramulosum Setchell et Gardner 1922 S. 362 Taf. 39 Figs—120 p. crassa, di-, raro tri- (tetra-)stromatica. Sori lineis pilorum concentricis, alternis + inconspicuis, separati. Sporangia 90—120 ». longa, 60—75 p crassa. Indusium conspicuum. Thallus bis 20 cm hoch, wiederholt zerschlitzt, mit keil—facherformigen Endsegmenten, 90—120 ». dick, zweischichtig. An einigen Exemplaren habe ich gefunden, dass um die Sori Teilungen eintreten konnen, so dass er 3(—4)- schichtig werden kann. Auch im untersten Teil konnen ahnliche Teilungen vorkommen. Die Sori sind von Haarleisten getrennt. Jede zweite Haarleiste ist aber mehr oder weniger undeutlich, so dass oft fast Doppelzonen von Sori gebildet werden. Sie gehort also unter die Padina pavonia-Gruppe von HAUCK (1877 S. 42). Zuweilen k6nnen die Haarleisten alternierend an der Ober- und Unterseite des Thallus entwickelt werden. Die Sporangien sind go—1I20 p. lang und 60—75 p. dick. Die Sori sind von einem deutlichen Indusium bedeckt. Von dieser neuen Art liegt ein grosses Material vor. Die Zonenanordnung der Sori und Haarleisten geht am besten aus Fig. 5 A hervor. Von den ubri- gen zweischichtigen Arten — P. distromatica und australis (HAUCK |. c. S. 43, 44), P. Sanctae Crucis (BORGESEN 1914 S. 201), P. japonica (YAMADA 1931 S. 67) — ist sie durch die Zonenanordnung der Sori und Haarleisten und das Vorkommen eines Indusiums wohlgetrennt. Eine Verwechslung mit den mehr- schichtigen Arten ist wohl kaum moglich. — Vom unteren Teil der Litoral- region bis in etwa 40 m Tiefe angetroffen. Mit Sporangien vom Januar bis April gesammelt. Bu ndiorte:.Stsi2ic, 3, 4,/6, 8; 41, 12,13; 40 b;» HASSLER exp yiNieRor Geogr. Verbr.: Juan Fernandez. Dictyota Lamour. Dictyota phlyctaenodes Montagne 1852a S. 200; 1852b S. 314. — Bic, Shh at 50; Pig. 1. Diese Art, die in der Sammlung reichlich vertreten ist, wurde von MON- TAGNE (Il. c.) nach Material von Juan Fernandez beschrieben. Sie wird bis etwa 20 cm hoch. Der Thallus ist unregelmassig dichotomisch reich verzweigt und wird bis 1—1,5 cm breit, in den jiingeren Teilen 0,;—1 cm, und 110—180 up. dick. Unten ist der Thallus mit einem kraftigen Rhizoidenlager bekleidet. Die Sprossspitzen sind stumpf und ausgerandet (Fig. 5 F—G). Die Farbe ist oliv- braun. Von MONTAGNE wurde sie mit Antheridien (?) gefunden. Selbst habe ich nur Tetrasporenpflanzen gesehen. Die Fortpflanzungsorgane sind in unregel- DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 621 : st a= wars - 7 ee (ini Hil uti Fig. 5. A—E Padina fernandeziana. A Thalluszipfel; B—E Querschnitte: C untere Teil mit Rhizoiden, D—E durch Sporangiensorus. F—I Dzictyota philyctaenodes. F—G Thallusspitze; H Querschnitt; I Poren. — A X 0; B—E xX 50; F—GX7; H X115; I X 305. massigen, mehr oder weniger undeutlichen, zusammenfliessenden Sori geordnet. Die Tetrasporangien sind 70—100 ». im Durchmesser. Die Art scheint mit der siidafrikanischen D. naevosa (Suhr) J. Ag., zu welcher DICKIE (1877 S. 452) Exemplare aus Juan Fernandez fihrte, sehr nahe verwandt zu sein. Bei dieser sind aber die Sori sehr deutlich, was bei denjenigen Exemplaren, die‘ich von D. phlyctaenodes gesehen habe, nicht der Fall war.. Ferner ist fiir diese das Rhizoidenlager an den unteren Thallusteilen bezeichnend. Auch die schénen Tiipfelungen (Fig. 5 I), die in den Wanden der Zentralzellen vorkommen (vgl. auch SCHMIDT 1938 S. 211), verdienen be- merkt zu werden, obwohl sie nicht fiir diese Art allein bezeichnend sind. D. phlyctaenodes scheint die einzige Déctyota-Art der Inseln zu sein. Sie kommt in der Sublitoralregion vor (20—45 m). Fundorte: ‘St. 2c, 4, 5, 8, 9, 11, 12, 13, 40b; Scumirr Nr.. 146 (Car- bajal Bay, 15.12. 1926); HASSLER-Exp. Nr. 66; MONTAGNE (I.c., leg. BERTERO); DICKIE (l.c., leg. MOSELEY). Geogr. Verbr.: Juan Fernandez. Glossophora J. Ag. Glossophora Kunthii (C. Ag.) J. Ag. 1887 S. 110; Howe 1914 S. 72; Zonaria Kunthii C. Ag. 1821 Taf. 15; Dictyota Kunthit Kitz. Tab. 9 Taf. 30:1; Glossophora Harveyi J. Ag. \.c. S. 111. — Fig. 6—7. Die beiden von J. AGARDH (l.c.) unterschiedenen G/. Kunthi von Peru und G/. Harveyi von Neuseeland und den Chatham Inseln, lassen sich nicht trennen. Ich habe das Originalmaterial im Herb. AGARDH untersucht und bin dabei zu dieser Auffassung gekommen. Die von J. AGARDH angefuhrten, art- 622 TORE LEVRING 4 * & OC ' at “ fs oY, heey a wo a wom TA \ < - $ XT ‘® | Tnunoninne A F ae = AVARULULUUCUUNAS TUS RS eee oniole le i hd 1S Sere COTE oe ) TAUUEDUINUNOTNNONACal OOS8085088800 Ssac¢ = Fig. 6. Glossophora Kunthit. A Querschnitte: A—B jiingere, E unterste Teile, D mit der Basis eines Sporophylls; F Pore; G Thallusspitze. — A—E x 150; F x 395; GX 40. trennenden Merkmale sind nur Unterschiede zwischen 4lteren und jiingeren Exemplaren oder sogar Thallusteilen. Ebenso diirfte es sich mit der von FARLOW (1902) beschriebenen G/. galapagensis verhalten. Leider habe ich von dieser kein Material gesehen. Die Art wird bis 30 cm (die Exemplare von Juan Fernandez bis 20 cm) hoch und ist dichotomisch verzweigt.. Gute Habitusbilder findet man bei C. AGARDH (I. c.), KUTZING (l.c.) und HOweE (l.c.). Der Thallus wird etwa 4— 10 mm breit und 150—300 y, dick. In den jiingeren Teilen besteht er aus einem zentralen Lager von grossen Zellen, einem einschichtigen Kortikallager von kleinen, dichtgefiillten Zellen und einem dazwischenliegenden Lager von Zellen intermediarer Grosse (Fig. 6 A—C). Mit zunehmendem Alter wird das intermediare Lager wenigstens zweischichtig (Fig. 6D). Im untersten Teil der Pflanze wird das Kortikallager oft 2—3-schichtig oder sogar mehr, weil sich hier das subkortikale und das kortikale Lager kaum trennen lassen. Die Zell- wande der zentralen Zellen werden mit zunehmendem Alter auch dicker. Poren (Fig. 6 F) kommen hier vor. Man findet auch kleine Gruppen von Haaren. DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 622 Fig. 7. Glossophora Kunthii. Sporophyll. Anlegung: A—D, H im Flachenansicht, E—G von der Seite; 1 Spitze; K Habitus; L Querschnitt mit Sporangien. — A—G X 230; H—I, L x 135; K x 18, Sie werden von den Kortikalzellen entwickelt (Fig. 6 A). Der Zuwachs des Thallus findet ganz wie bei Dzctyofa mittels einer Scheitelzelle statt. Von Fortpflanzungsorganen sind nur Tetrasporen bekannt. Fiir die Gattung besonders eigenartig ist, dass sie sich in kleinen Sporophyllen, von welchen die Thallusoberflache dicht iiberzogen ist, entwickeln. Vereinzelt findet man sie auch am Thallus selbst. Die Sporophylle werden auf der Thallusoberflache in der Weise angelegt, dass sich eine Kortikalzelle vergrossert und inhalts- reicher wird. Diese teilt sich in zwei Zellen, wovon die apikale zur Scheitel- zelle des Sporophylls wird und die andere sich sofort weiter teilt. Die ganze Entwicklung geht aus Fig. 7 A—H hervor. Das reife Sporophyll ist lanzettlich bis zungenformig, gelegentlich gegabelt, etwa 2—4(—5) mm lang, 0,4—0,6 mm breit und 90—-120 y», dick. Es besteht wie der Hauptthallus aus einem Zentral- lager von grossen Zellen und einem kleinzelligen Kortikallager. Das subkor- tikale Lager fehlt aber. Der Zuwachs geschieht auch hier mittels einer Scheitel- zelle. Die Tetrasporangien werden an der Oberflache entwickelt und sind 70— go v. in Durchmesser. — Im unteren Teil der Litoral- und obersten Teil der Sublitoralregion im Januar und August gesammelt. Pind onte- 2c. 2,25 b- Geogr. Verbr.: Neuseeland, Chatham-Inseln, Nord-Chile, Peru, Juan Fernandez. 624 TORE LEVRING Ordn. Chordariales. Fam. Chordariaceae. Papenfussiella Kylin. Papenfussiella Moseleyi Levr. nov. sp.; Chordaria hippuroides Dickie 1877 S. 452 (non Bory). — Fig. 8. Frons 10—20 cm alta, sparse ramosa, ramis longis. Axis centralis circ. 0,8—1,3 mm crassus. Fila perpherica 0,;—1I mm longa, 15—18 y. crassa, superne et inferne angustata, e cellulis cylindricis, 1-plo, superne ad 3-plo longioribus quam latis constructa. Sporangia unilocularia 70—90(—110) y. longa et 25— 30 w crassa. In der Sammlung MOSELEYS von Juan Fernandez liegen im Herb. Kew einige Exemplare von einer Chordariacee, die von DICKIE als Chordaria hip- puroides bestimmt wurde. Ich habe einige Proben des Materials untersucht und es stellte sich dabei sofort heraus, dass die Pflanze zu der von KYLIN (1940 S. 17) aufgestellten Gattung Papenfussiella gehort und eine neue Att ist, die sich den schon bekannten Arten dieser Gattung nahe anschliesst. Der Thallus ist 10—15(—20) cm hoch und sparlich mit langen Seitenasten versehen. Die Zentralachse ist etwa 0,8—1I,3 mm dick, die Assimilationsfaden O,;—I mm lang und 15—18 p, dick, nach oben und an der Basis diinner. Die Zellen sind zylindrisch oder an den Querwanden etwas eingeschniirt, etwa ein- mal, nach oben bis 3mal so lang wie breit. Unilokulare Sporangien 70—g90 (—110) w lang und 26—30 », dick. Am meisten ahnelt die Art P. evaczis, laxa und tristanenszs (die beiden ersten von Siidafrika, die letzte von Tristan da Cunha). Es liegen aber Unter- schiede in Einzelheiten vor. Leider stand mir nur getrocknetes Material zur Verftigung, was eine eingehende Untersuchung unméglich macht. Fundort: DICKIE (l.c., leg. MOSELEY). Geogr. Verbr.: Juan Fernandez. Fam. Splachnidiaceae. Splachnidium Grev. Splachnidium rugosum (L.) Grev.; Harv. 1858 Taf. 14; Kitz. Tab. 10 Date wo. tSkottsbere 1920/)S..277;.Kuckuck 10920):S:.77- Die Exemplare sind bis etwa 10 cm hoch und fertil; die Art erreicht bei Juan Fernandez nicht die Grosse der siidafrikanischen Pflanze, stimmt aber gut mit der neuseelandischen iiberein. Sie bildet grosse Bestande auf Steinen und Felsen besonders im unteren Teil der Litoralregion. Die systematische Stellung dieser interessanten Pflanze war lange unsicher und wurde erst durch die Untersuchungen SKOTTSBERGS (lI. c.) und KUCKUCKS (l.c.) klargelegt. Sie fithren sie als eine eigene Familie zu Chordarzales, eine Auffassung, der sich KYLIN (1940 S. 55) anschliesst. Ich bin derselben Mei- DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 62 un Fig. 8. Papenfussiella Moseleyz. A Habitus; B Assimilationsfaden mit Sporangien; C peripherer Zentralfaden aus der Sprossspitze. A X053) B—€ X 265. nung. Die viel beschriebenen »Initialzellen» haben sich als eine endophytische Griinalge, Codiolum Kuckuckii, erwiesen (vgl. S. 604). Fundorte: 2b, 3, 39b; HASSLER-Exp. Nr. 68, 69. Geogr. Verbr.: Siidafrika, St. Paul’s Insel, Ostindien, Australien, Tas- manien, Neuseeland, Juan Fernandez, San Felix. Ordn. Punctariales. Fam. Encoeliaceae. Colpomenia Derb. et Sol. Colpomenia sinuosa (Roth) Derb. et Sol. 1856 S. 11; Vickers 1908 S. 40 Taf. 22; Setchell and Gardner 1925 S. 539; Hydroclathrus sinuosus (Roth) Zanard.; Hauck 1885 S. 393. f. typica Setchell et Gardner (I. c.) S. 540. f. deformans Setchell et Gardner 1903 S. 242; 1925 S. 542; Scytosiphon bullosus Saunders 1898 S. 163; Asperococcus echinatus Dickie (non Mert.) 1877 S452: 626 TORE LEVRING Es liegen mehrere gut entwickelte Exemplare, die am nachsten mit f. 4- fica ibereinstimmen, vor. Die von DICKIE (l.c.) als Asperococcus echinatus an- gefiihrte Form ist mit f. deformans identisch. Seine Exemplare im Herb. Kew sind gross und unregelmassig schlauchformig. In der Sammlung SKOTTSBERGS gibt es einige Exemplare, die wahrscheinlich junge Individuen dieser Form sind. — Wurde vom unteren Teil der Litoralregion bis in 25 Tiefe gesammelt. Fertil. Fundorte: St. 2b, ¢;.3, 4, 10, 12, 38D; 40; DICKIE \lhe:, alsa coccus sinuosus und echinatus, leg. MOSELEY); HASSLER Nr. 64; SCHMITT Nr. 240 (8.12. 1926). Geogr. Verbr.: Weit verbreitet in warmeren Meeren. Hydroclathrus Bory. Hydroclathrus clathratus (Bory) Howe 1920 S. 590; /. cancellatus Bory; Harvey 1859 Taf. 98; Mitchell 1893 S. 53 Taf. 15; Vickers T9@8 "Sees ilaite22. Von dieser weitverbreiteten Art habe ich nur ein einziges Exemplar ge- sehen. Fundort: SCHMITT Nr. 147 (Carbajal Bay, 15.12.1926, 15—20 fa.). Geogr. Verbr.: Weitverbreitet in warmeren Meeren. Tlea Fries. Ilea fascia (Miull.) Fries; Kjellm. 1883 S. 319; Setchell and Gardner 1925 S. 535; Phyllitis fascia (Mill.) Kitz. 1843 S. 142 Taf. 24: III; Petalonza fascia (Mull.) O. Kuntze 1898 S. 419. Das Material ist ziemlich reichlich. Es ist im Januar und August im un- teren Teil der Litoralregion gesammelt. Die Exemplare sind bis etwa 7 cm hoch, etwa 0,5—1,3 cm breit und etwa 120—180u. dick. Diese weitverbreitete Art ist ziemlich variabel (vgl. SETCHELL und GARDNER l.c.; usw.). So kommen Formen mit oder ohne Haargruben vor. Die Exemplare von Juan Fernandez tragen kleine Haargruben, die an den im Januar gesammelten Exemplaren am besten entwickelt sind. Reich fertil im August, weniger im Januar. Fundorte: St. 10, 36b. Geogr. Verbr.: Kosmopolitisch. Scytosiphon (C. Ag.) Thur. Scytosiphon lomentaria (Lyngb.) J. Ag.; Setchell and Gardner 1925 S. 531; Chorda lomentaria Lyngb. 1819 S. 74; Harv. 1871 Taf. 285. Die meisten Exemplare sind im Januar bei St. 10 gesammelt; sie sind bis etwa 20 cm lang und ganz normal entwickelt. Dazu kommen noch einige Exemplare von anderen Lokalitaéten, die aber kleiner oder etwas deformiert sind. Sie sind aber fertil. — Kommt im unteren Teil der Litoralregion vor. DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 627 Fundorte: St. 2b, 10, 39b. Geogr. Verbr.: Weit verbreitet; an der siidamerikanischen Westkiiste von Chiloé siidwarts; Falkland-Inseln, Siidgeorgien. Fam. Chnoosporaceae. Scytothamnus Hook. f. et Harv. Scytothamnus australis (J. Ag.) Hook. f. et Harv. 1845 S. 531; Kiitz. Tab. 8 Taf. 12; Skottsberg 1907 S. 48; Chordarta australis J. Ag. 1841 S. 47. — Fig. 9g. Uber die systematische Stellung dieser Gattung sind verschiedene An- schauungen vorgebracht worden. Von J. AGARDH (1848 S. 63) wurde sie zu Chordariaceae gestellt. KJELLMAN (in ENGLER und PRANTL 1897 S. 214), DE TONI (1894 S. 454), SKOTTSBERG (1907 S. 47) und KUCKUCK (1929 S. 87) fuhren sie zu Dzctyosiphonaceae. Ein anatomisches Studium zeigt, dass die Annahme AGARDHS kaum richtig sein kann, weil sie nicht aus zusammen- geklebten Zellfaden aufgebaut ist. Ebenso verhalt es sich mit der Annahme der anderen Verfasser. Nach der jetzigen Begrenzung der D7ctyoszphonales (vgl. SETCHELL und GARDNER 1925 S. 586; KYLIN 1933 S. 93) ist fiir diese Ordnung ein Thallusaufbau durch die Tatigkeit einer Scheitelzelle und das Vorkommen von interkalaren Teilungen bezeichnend. Ein Studium von jungen Sprossspitzen zeigt deutlich das Fehlen einer Scheitelzelle. SKOTTSBERG (1921 S. 34) gibt fiir S. fasceculatus (Hook. f. et Harv.) Cotton und S. Azrsutus Skottsb. das Vorkommen von Scheitelzellen an. Die Zugehorigkeit dieser Arten zu Scytothamnus scheint aber nicht gesichert. Soweit ich finden kann, lasst sich Scytothamnus australis gut zu Puncta- riales stellen. Bezeichnend fiir diese Ordnung (vgl. KYLIN |. c.) ist ein Thallus- aufbau durch interkalare Teilungen und das Vorhandensein von interkalaren Langswanden. Aus Fig. 9B geht das Aussehen einer Sprossspitze hervor. Mit zunehmendem Alter werden die zentralen Teile des Thallus locker und das Gewebe ist hier nicht mehr kompakt. Man kann dies schon ein kleines Stick von der Spitze entfernt sehen. Das Langenwachstum scheint grésstenteils in einer meristematischen Zone an der Sprossspitze stattzufinden. Sie ist durch zahlreiche interkalare Teilungen gekennzeichnet. Die Gattung erinnert darin an Chnoospora. An Scytothamnus werden, wie SKOTTSBERG zuerst geschildert hat (1921, S. 34), Haargruben bisweilen entwickelt (Fig. 9 A). Sie haben ein ahnliches Aussehen wie bei anderen Phaeophyceen. Bemerkenswert ist aber, dass sie fehlen kénnen. An dem Material von Juan Fernandez sind sie gut und reich- lich entwickelt; den Pflanzen von Chiloé fehlten sie (SKOTTSBERG I. c.). Unter dem von mir untersuchten Herbarmaterial gibt es Exemplare, wo die Haar- gruben fehlen oder sparlich sind. So kommen z. B. unter dem Originalmaterial J. AGARDHS von Neuseeland Exemplare mit und ohne Haargruben vor. Ob das Vorhandensein oder Fehlen der Haargruben verschiedenen Entwicklungsstadien oder Formen entspricht bleibt unentschieden. Die Fortpflanzungsorgane sind unilokulare Sporangien, die in die Rinde eingesenkt sitzen. 628 TORE LEVRING Fig. 9. Scytothamnus australis, A Haargrube; B Thallusspitze im Langsschnitt; C Langs- schnitt mit Sporangien. — A—B X 330; C X 210. Am nachsten diirfte Scytothamnus mit Chnoospora verwandt sein, eine Vermutung, die schon von BORGESEN (1924 S. 263) und SETCHELL und GARD- NER (Il. c. S. 552) ausgesprochen worden ist. Bei beiden kommt eine meriste- matische Zone an den Sprossspitzen vor, die aber bei Chwoospora noch starker begrenzt ist (vgl. LEVRING 1938 S. 23). Die zentrale Markschicht ist bei dieser auch von festerem Bau als bei Scytothamnus. Beide haben Haargruben von ahnlichem Aussehen. Ich fiihre sie deshalb zu der von SETCHELL und GARD- NER (Il. c.) aufgestellten Familie Chnoosporaceae, die unter den Punctarzales gut abgegrenzt sein diirfte. Es liegt ein reiches Material vor, im Januar, Februar und August gesammelt und fertil. ScyZothamunus bildet im unteren Teil der Litoralregion grosse Bestande. Fundorte: St. 2b, 12, 39 b, 40; HASSLER-Exp. Nr. 70. Geogr. Verbr.: Australien, Neuseeland, Chile, Juan Fernandez. Rhodophyceae. Ordn. Porphyridiales. Fam. Porphyridiaceae. Goniotrichum Kitz. Goniotrichum alsidii (Zanard.) Howe 1914 S. 75; G. elegans (Chauv.) Le Jolis 1863 S. 103; Rosenv. 1909 S. 75. DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 629 Kommt auf verschiedenen Algen epiphytisch vor. Normal entwickelt. Riumdorte: Sti24.05.00,) 11; 124 3oa).b, Geogr. Verbr.: Wahrscheinlich kosmopolitisch. Goniotrichum cornu cervi (Reinsch) Hauck 1885 S. 5109. Wie dem vorigen zusammen mit anderen Mikroepiphyten angetroffen; scheint aber etwas seltener als G. a/szd7zz zu sein. mrimmarerLe: ot; Ayo, IT, I. ao Geogr. Verbr.: Mittelmeer, Nordatlantik, Westkiiste von Amerika. Ordn. Langiales. Fam. Bangiaceae. Erythrocladia Rosenv. Erythrocladia irregularis Rosenv. 1909 S. 72; Kylin 1925 S. 9. Von dieser habe ich vereinzelte Exemplare zusammen mit der folgenden, die gewohnlicher ist, auf Chaetomorpha antennina gefunden. Fertil. Bundort St. 36a. Geogr. Verbr.: Europa, Westkiiste von Nordamerika, Juan Fernandez. Erythrocladia subintegra Rosenv. 1909 S. 73; Kylin 1925 S. 9. Diese Art habe ich auf verschiedenen Algen gefunden. Wie die vorige stimmt sie durchaus mit der europaischen Form iberein. Fertil. Eoumrdorte: St §, 36 a | Geogr. Verbr.: Wahrscheinlich kosmopolitisch. Erythrotrichia Aresch. Erythrotrichia carnea (Dillw.) J. Ag.; Rosenv. 1909 S. 67. Scheint gut mir der europaischen Form ibereinzustimmen. Die Faden sind etwa 16—25 y dick und die Zellen ungefahr ebenso lang wie breit, zuweilen etwas kiirzer. Fertil. Hund or ter gst. 4, 3, 130, 30la,b. Geogr. Verbr.: Wahrscheinlich kosmopolitisch. Porphyra C. Ag. Porphyra umbilicalis (L) J. Ag. 1883 S. 67. Von dieser liegt ein etwa 10 cm hohes fertiles Exemplar vor. Auf Chondrt- ella habe ich einige kaum 1 cm hohe Exemplare gefunden, die wahrscheinlich auch zu derselben Art gehoren. Pundort: St 36 a. Geogr. Verbr.: Nordl. und siidl. Meere. TORE LEVRING 630 ans Se = wf a So Ss rN DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 631 Ordn. Nematlionales. Fam. Chantransiaceae. Chantransia (DC.) Schmitz. Chantransia catenulata (Howe) De Toni 1924 S. 44; Acrochaetium cate- nulatum Howe 1914 S. 84 Taf. 31 Fig. 12—18; Kylin und Skottsberg 1919 S. 3. Auf verschiedenen Algen habe ich vereinzelte, kleinere Exemplare gesehen. Die Zellen sind etwa 6—g yw (ausserste Zellen bis 4,5 p.) und die Sporangien 6—g pv. in Durchmesser, also etwas kleiner als normal. Wahrscheinlich waren aber die Exemplare jung. Fundorte: Sti 1, 4) 11, 12, 364, 39D: Geogr. Verbr.: Peru, Feuerland, Juan Fernandez. Chantransia Thuretii (Born.) Kylin 1907 S. 119; Rosenv. 1909 S. 104. Von dieser habe ich einige wenige Bischel zusammen mit anderen Mikro- epiphyten auf Dzctyota und Splachnidium gefunden. Sie gehort zu der von ROSENVINGE (lI. c.) aufgestellten f. agama und stimmt gut mit der europaischen Form iiberein. Die Faden sind etwa 2 mm hoch und (7—)8—1I ». breit. Die Zellen sind 2.;—6mal so lang wie breit und die Sporangien g—II uw. breit und 15—22 ». lang. Fundorte: St. 4, 39b. Geogr. Verbr.: Europa, Ostkiiste von Nordamerika, Kalifornien, Juan Fernandez. Chantransia pacifica (Kylin) Levr. nov. comb.; Acrochaetium pacificum Kylin 1925 S. 11; Rhodochorton pacificum (Kylin) Drew 1928 S. 169. Auf verschiedenen Algen und Sertularien fand ich eine Chantransia, die mit dieser Art gut wbereinstimmt. Die Faden sind bis etwa 0,6 mm hoch. Die Zellen sind 4,;5—6 4 breit und 2—4mal so lang, die Sporangien 5—7 »v. breit und 8—I11 ». lang. Fundorte: St. 5, 36a. Geogr. Verbr.: Westkiiste von Nordamerika. Chantransia fernandeziana Levr. nov. sp. — Fig. 10 I—R. Thallus c:a 1—1,5 mm altus. Spora germinans processum decumbentem cuneiformen in cortice hospitis (Déctyota phlyctaenodes) endophyticum et fila erecta emittens. Fila erecta radialiter ramosa, cellulis g—14 wu. latis, (2—)3— 5-plo longioribus. Chromatophorum parietale pyrenoideo instructum. Mono- sporangia sessilia, raro pedicellata vel terminalia, in parte superiore ramorum insidentia, oblonga, 11—15 vu. lata, 17—21 wu longa. Fig. 10. A—H Chantransia grandis. A Habitus; B—F Basalteil (B auf Scyfothamnus, C—D auf Splachnidium, F aut Codium); G—H Zellen aufrechter Zweige. I—R Chantransia_fernan- deziana. \ Habitus; K—L Zellen aufrechter Zweige; M—R Basalteil. A x 55; B—F, 1 X 120; G—H, K—R X 200. 632 TORE LEVRING Ch. fernandesiana kam ziemlich reichlich zusammen mit anderen Mikro- epiphyten vor. Der Thallus ist I—1,; mm hoch. Der Basalteil besteht aus der urspriinglichen Spore, von welcher kiirzere, rhizoidenartige Faden in die Rindenschicht der Wirtspflanze (Ductyota phlyctaenodes) Nineindringen. Diese Spore ist etwa 15—22 , in Durchmesser, hat ziemlich dicke Zellwande und bleibt lange erkenntlich. Erst an alteren Exemplaren wird sie undeutlich. Von diesem Basalteil gehen aufrechte Faden aus, die ziemlich regelmassig allseitig verzweigt sind. Ihre Zellen sind zylindrisch, g—14 yw. dick, 3—5(im untersten Teil bis 2)mal so lang. Sie enthalten einen parietalen Chromatophor mit einem Pyrenoid, das aber ofters undeutlich war. Die Monosporangien sind eiformig, II—I5 yu. breit und 17—21 yp, lang, und sitzen an der Oberseite der oberen Seitenzweige ungestielt, seltener auf einem einzelligen Stiel oder, wenn der Zweig kurz ist, terminal. Besonders bezeichnend ftir die neue Art ist der Basalteil mit seiner lange erkenntlichen Keimspore (Fig. 10 M-- R). Bei der Keimung wird eine schlauch- artige Verlangerung gebildet, die in die Wirtspflanze hineindringt. In der Regel werden einige kurze Verzweigungen gebildet. Der ganze endophytische Teil besteht indessen nur aus einigen wenigen Zellen, was besonders an Acrochae- tium opetigenum (BORGESEN 1915—20 S. 38) erinnert. Sie unterscheiden sich aber in Gréssenverhaltnissen und anderen Einzelheiten voneinander. Fundort: St.4. Geogr. Verbr.: Juan Fernandez. Chantransia grandis Levr. nov. sp. — Fig. 10 A—H. Thallus caespitosus, 5—g mm altus. Pars basalis e filis endophyticis cellulis irregularibus, 11—20 yp, crassis, 2—6-plo longioribus consistens. Fila erecta numerosa, radialiter ramosa, e cellulis cylindricis, (10—)15—20(—25) w crassis, 3—5-plo longis formata. Sporangia pedicellata, singula vel bina in superiore latere ramulorum insidentia, ovoidea, 14-20 »p. lata, 20—28 y. longa. Sie bildet etwa 5—9 mm hohe Biischel. Der Basalteil besteht aus endo- phytischen Faden, die in die Unterlage eindringen und sich dort verasteln. Besonders auf Codiuwm, das ja ziemlich locker gebaut ist, wird der Basalteil gut entwickelt. Die Zellen dieser Faden sind mehr oder weniger unregelmassig, 11—20 ». dick und 2—6mal so lang. Die aufrechten Faden sind reichlich und allseitig verzweigt. Die unteren Zweige sind lang und der Mutterachse gleich, die oberen werden allmahlich kiirzer und tragen an ihrer Oberseite Mono- sporangien, einzeln oder zwei zusammen auf einem einzelligen Stiel. Die Zellen sind zylindrisch, etwa 15—20 y. (im obersten Teil bis 10 ») dick und 3—5mal so lang. Im untersten Teil kénnen sie ausnahmsweise bis 25 py, dick werden. Sie enthalten einen parietalen, gelappten Chromatophor mit 1— 2 (selten bis 4) Pyrenoiden. Die Monosporangien sind eiformig, 14—20 y. breit und 20— 28 pw lang. Ich habe die Art auf Splachnidium rugosum, Scytothamnus australis und Codium fernandesianum gefunden. Es liegt ein ziemlich reiches Material vor. Am nachsten scheint sie mit den von DREW (1928) beschriebenen Ahodo- chorton magnificum, elegans und rhizoideum verwandt zu sein. Sie unter- DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 633 scheiden sich aber deutlich in Gr6ssenverhaltnissen und Einzelheiten in der Verzweigung. Fundorte: St. 12, 39.b. Geogr. Verbr.: Juan Fernandez. Chantransia Collinsiana (Borgs.) De Toni 1924 S. 69; S855 2 So Hi e oF 2, eae (7 J) 7 00 (} EME (MT) 0%o Ips 4 OH 0, ILMOCSG UGA a= ah 2 S's = PP Te ES B98 920 0 We of ae af if TODOS GHZ £0 Lo fe Ep Fig. 14. Chondriella pusilla, A Thallusspitze im Langsschnitt; B Querschnitt; C Dito mit Tetrasporen; D—E Habitus zystokarpientragender Exemplare. — A X 190; B—C X 315; D—E 1,2. Die reifen Zystokarpien stellen Knollchen dar, die fast immer einzeln nahe den Thallusspitzen sitzen. Ihr Aussehen geht aus Fig. 14 D—E und 15 F hervor. Im Innern liegt der reife Gonimoblast. Das Nahrgewebe ist jetzt als eine unbedeutende Faserschicht, wovon der Gonimoblast umgeben ist, und als undeutliche Faserbiindel im Innern desselben vorhanden. Eine besondere Miin- dung in den Zystokarpien scheint nicht vorzukommen. Die Karposporen wer- den durch Zerbrechen oder Auflosung der Zystokarpienwand frei. Mannliche Pflanzen habe ich nicht gesehen. Die Tetrasporangien, die paarig geteilt sind, werden in der Rindenschicht entwickelt (Fig. 14 C). Sie kommen in bald zusammenfliessenden Flecken vor, so dass sie fast iiber den ganzen Thallus zerstreut sind. Die tetrasporangien- fihrenden Teile der Rinde sind im Verhialtnis zu den sterilen Teilen etwas aufgelockert. Anatomisch erinnert Chondriella stark an die Gigartinaceen, scheint aber weniger von dem normalen Springbrunnentypus abgeleitet zu sein. Die langs- verlaufenden, zentralen Faden sind bei Chondriel/a jedenfalls in den jungeren Teilen deutlich (Fig. 14 A). Obwohl ich an meinem Material die Prokarpien- und Gonimoblastentwicklung nicht in allen Einzelheiten genau verfolgen konnte, gehen aber die Hauptziige aus der obigen Darstellung und aus den Figuren hervor. Das Aussehen der Prokarpien stimmt mit den Befunden bei den Gigar- tinaceen (vel. KYLIN 1937 5.254). In der Gonimoblastentwicklung ist sie auch den Gigartinaceen, wo sich die verschiedenen Gattungen in Einzelheiten unter- scheiden, ahnlich. Der grésste Unterschied liegt in der Anordnung der Tetra- 642 TORE LEVRING Fig. 15. Chondriella pusilla. A, D—E Prokarpien; B Junges Zystokarp; C Auxiliarzelle mit jungen Gonimoblastfaden; F Reifes Zystokarp; G Grenzschicht zwischen Gonimoblast und vegetativem Gewebe mit Verbindungsfaden. — A, D—E xX 785; B X 105; C X 500; F X 50; Gx<205- sporangien, die bei beiden Familien paarig geteilt sind. Bei den Gigartinaceen sind sie in Gruppen im Thallus eingeschlossen, bei Chondrie/la dagegen iber die Thallusoberflache verstreut und in der Aussenrinde eingelagert. Chondriella pusilla Levr. nov. sp. — Fig. 14—15, Taf. 51, Fig. 5. Frons cartilaginea, circ. I—I,5 cm alta, 250—425 wu. crassa et O,5—I mm lata. Cystocarpia 0,6—0,7 mm diametro. Tetrasporangia 30—40y. longa, IO— 14 p. crassa. Der Thallus ist fleischig—knorpelig, seitlich und gabelig verzweigt, zu- weilen auch proliferierend, etwa I—1,5 cm hoch, mehr oder weniger abgeflacht, etwa 250—425 y. dick und 0,;—1 mm breit. Die Zystokarpien sind 0,6—0,7 mm im Durchmesser, die Tetrasporangien 10—14 p dick und 30—4o yp, lang. — Bildet dichte Rasen auf Felsen in der Litoralregion. Pundorte: St) 36;.38\a,) 3014, b: Geogr. Verbr.: Juan Fernandez. DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 643 Ordn. Rhodymeniales. Fam. Rhodymeniaceae. Gloioderma J. Ag. Gloioderma fruticulosum (Harv.) De Toni 1900 S. 496; Kylin 1931 S. 7; Florea fruticulosa Harv. 1860 Taf. 156. Von dieser Art liegen ein paar jiingere Exemplare vor. Sie sind bis etwa 5 cm hoch und steril. Die Ubereinstimmung mit einigen der Exemplare HARVEYS, die ich gesehen habe, ist gut. hanmcdort: St. 5. Geogr. Verbr.: Australien, Juan Fernandez. Gloioderma subdichotomum Levr. nov. sp..— Fig. 16, Taf. 53, Fig. 1. Frons 2—4 cm alta, pseudodichotome ramosa, 2—4 mm lata, inferne sensim attenuata, 350—450 ». crassa. Tetrasporangia 28—38 p. crassa, 35—42 yp. longa. Der Thallus ist etwa 2—4 cm hoch, einigemal fast regelmassig dicho- tomisch verzweigt und etwas gallertig. Er ist etwa 2—4 mm breit und wird gegen die Basis allmahlich schmaler, deutlich abgeflacht und etwa 350—450 up. dick. Anatomisch stimmt G. swbdichotomum mit den anderen Arten der Gattung gut iiberein. Das Innengewebe ist grosszellig ohne Rhizoiden, die Aussenrinde sehr kleinzellig mit deutlicher Fadenstruktur. Leider habe ich nur Tetrasporen- pflanzen gefunden. Die Tetrasporangien sind paarig geteilt, in die Aussenrinde eingesenkt und tiber die Thallusoberflache verstreut. Sie sind etwa 28—38 yp. breit und 35—42 » lang. umd ore: St, 03: Geogr. Verbr.: Juan Fernandez. Smcece Poca > 2223009 e009049 Fig. 16. Gloiderma subdichotomum. A Querschnitt mit Tetrasporen; B Langsschnitt.— xX 150. 644 TORE LEVRING Botryocladia Kylin. Botryocladia fernandeziana Levr. nov. sp. — Fig. 17. Frons ad c:a 2 cm alta, e stipite 0,4—0,6 mm crasso et vesiculis strato interiore e cellulis magnis, exteriore e cellulis parvis reticulatim dispositis for- mata, vesiculis ad circ. 15 (20) mm longis et 5 mm crassis, radialiter dispositis. Glandulae paucae in cellulis singulis insidentia. Bis etwa 2 cm hoch. Der stengelformige Teil, der an Steinen, Litho- thamnien usw. mit einer kleinen Haftscheibe befestigt ist, ist etwa O,4—0,6 mm dick, wenig verzweigt und tragt allseitig Blasen, die eiformig—zylindrisch, bis Fig. 17. Botryocladia fernandeziana. A—D Habitus (D mit Zystokarpien); E, K Flachen- ansicht; F Querschnitt des Stieles; G Driisenzellen von der Innenseite gesehen; H Querschnitt mit Driisenzellen, I mit Tetrasporen. — A—D nat. Gr.; E, H—K x 170; F X 45; G X 100. etwa 15 (20) mm lang und 5 mm breit sind. Der Stengel ist solid. Die Innen- schicht der Blasenwande besteht aus grossen Zellen; die dussere, kleinzellige Schicht ist unvollstandig und fillt an der Oberflache die Zwischenrdume zwischen den grossen Zellen netzartig aus. Wie aus Fig. 17 E und K hervorgeht, kann diese Schicht verschieden gut entwickelt sein. Driisenzellen habe ich nur sehr vereinzelt gefunden. Sie sitzen an der Innenseite einzelner Wandzellen in Gruppen von wenigen Stiick (Fig. 17 G—H). Ich habe die Pflanze sowohl mit Zystokarpien als mit Tetrasporangien gefunden. Von anderen Arten scheint sie gut getrennt zu sein. Am nachsten durfte sie mit B. microphysa (Hauck) Kylin und pyrzformis (Borgs.) Kylin verwandt sein. Die Blasenwand ist bei diesen ahnlich gebaut. 2. pyriformzs ist aber viel grésser, die Blasen sind birnformig und die Drisenzellen auch etwas ver- schieden (vgl. BORGESEN I910 S. 187; 1915—20 S. 400). Bei B. microphysa DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 645 besteht der Thallus in der Regel nur aus einer einzigen gestielten, 2—6 mm langen Blase (vgl. KUCKUCK 1912 S. 209). Eine gewisse Ahnlichkeit zeigt sie auch mit B. Skottsbergit'. Die Blasen sind aber bei dieser kleiner und das aussere, kleinzellige Lager ist vollstandig. Pusddortie ots Geogr. Verbr.: Juan Fernandez. Rhodymenia Grev. Rhodymenia cuneifolia (Hook. f. et Harv.) Taylor 1939 S. 147; Phy/lo- phora cuneifolia Hook. f. et Harv. 1845 S. 260; 1847 S. 486; Cotton 1915 ep rg) Dak. .7. Von dieser Art liegen einige Stiicke alterer Exemplare vor. Die Art wurde von SKOTTSBERG in KYLIN und SKOTTSBERG (1919 S. 19) zu Kz. coral- lina (Bory) Grev. gestellt. Nach TAyLor diirfte diese 2%. corallina zwar mit Rh. cunetfolia, aber nicht mit der von Bory identisch sein. Letztere hat KYLIN (1931 S. 21 Taf. 8) abgebildet. Die von HOWE (1914 S. 124 Taf. 50—51) be- sprochene RA. corallina dirfte weder mit der Art BORYS noch mit RA. cunet- folia identisch sein, sondern ist wahrscheinlich als eine besondere Art anzu- sehen. — Steril in 30—45 m Tiefe gesammelt. Hundorte:,St.,5,:6. Geogr. Verbr.: Falklandsinseln, Chile, Juan Fernandez. Rhodymenia australis Sond. 1845 S. 56; Harv. 1860 Taf. 146; De Toni et Forti 1923 S. 29; Acropeltis australis (Sond.) J. Ag. 1852 S. 609; Kutz. Tab. 19 Taf. 34. Von dieser liegen ein paar Exemplare vor, welche mit anderen von mir gesehenen gut iibereinstimmen. Sie sind bis etwa 8 cm hoch und wahrschein- lich ziemlich jung. Die Verzweigung ist wiederholt fast regelmassig dichotomisch mit abgerundeten Thallusspitzen. In den unteren Teilen ist der Thallus etwa 125—200 uw. dick. Leider sind die Exemplare steril. Mit Rk. linearis J. Ag. (vgl. KYLIN 1931 S. 20 Taf. 6 Fig. 16) diirfte sie nahe verwandt sein. Habituell ist sie der von Howe (1914 S. 125 Taf. 50—51) als Rk. corallina (vgl. oben) abgebildeten Form sehr ahnlich. Zu be- merken ist aber, dass diese Art in allen Teilen viel grOsser ist. | Von einer Rhodymenia-Art habe ich einige Fragmente mit Tetrasporen und Zystokarpien gefunden. Sie zeigen eine gewisse Ubereinstimmung mit Rh. australis. Der Thallus ist etwas gekrausselt. Ob sie eine Form dieser oder eine neue Art darstellt lasst sich kaum entscheiden, weil das Material zu gering ist. — In der Sublitoralregion gefunden. © unser Ges. Se. 5 52. Geogr. Verbr.: Australien, Juan Fernandez. 1 B. Skottsbergii (Borgs.) Levr. nov. comb. — Chrysymenia Skottsbergii Borgesen 1924. S. 307. Der anatomische Bau zeigt deutlich, dass sie zu Botryocladia gestellt werden muss. 646 TORE LEVRING Dendrymenia Skottsb. Dendrymenia flabellifolia (Bory) Skottsb. 1923 S. 16; Sphaerococcus Jlabellifolius Bory 1828 S. 174 Taf. 17; Rhodymenia flabellifolia (Bory) Mon- tagne 1844——467 5. 105; hlowe) Told Ss. 124A siate 40: Es liegen einige bis etwa 10 cm hohe, offenbar junge Exemplare vor, die ich zu dieser Art stelle. Sowohl habituell als anatomisch ist die Ubereinstim- mung mit den Exemplaren im Herb. AGARDH, das u. a. Originalmaterial Borys enthalt, gut. Samtliche Exemplare waren steril. Rimndionrts St -40ia: Geogr. Verbr.: Westkiiste von Siidamerika. Ordn. Ceramiales. Fam. Ceramiaceae. Antithamnion Nag. Antithamnion minutissimum Levr. nov. sp. — Fig. 18. Frons ad 4 mm alta, sparse ramosa, cellulae ramorum pinnis oppositis, latere superiore pectinatis, ornatae; ramuli ultimi apice rotundati—acuti; cellulae ramorum (20—)40—-70 w. crassae, 2—2,5-plo longiores; glanduale sat multae; tetrasporangia pedicellata, 17—20 ». crassa et 25-30 p. longa. Der Thallus wird bis etwa 4 mm _ hoch und ist mit etwas verzweigten Hauptzweigen versehen. Jede Zelle der Hauptzweige tragt zwei opponierte Kurztriebe, die 100—-170 ». lang sind. Gelegentlich kann ein Kurztrieb durch einen Langtrieb ersetzt werden. Die Zellen der Hauptzweige sind 40—70 4, im oberen Teil bis etwa 20 p dick, und 2—2,s5mal so lang. An der Basis sind die Kurztriebe 9—14 ». dick. Die Zellen sind 2—2,smal so lang wie breit. An der Oberseite tragen die Kurztriebe Seitenaste, die abgerundet bis spitz endigen. Ihre Zellen sind 4—6 y, dick und 2—3mal so lang. Altere Seitenaste konnen ahnlich wie die Kurztriebe an der Oberseite ein paar Astchen tragen. Driisen- zellen kommen an den Kurztrieben und Seitenasten ziemlich reichlich vor und sind verhaltnismassig gross; an jedem Kurztrieb findet man etwa 1-3. Die Anzahl kann aber stark gesteigert sein (Fig. 18 D). Die Tetrasporagien sitzen kurz gestielt an den Kurztrieben und _ sind 17—20 y. breit und 25—30 » lang. Das Aussehen der Spermatangienstande geht aus Fig. 18C hervor. Die Karpogonaste entwickeln sich aus der untersten Zelle der Kurztriebe in der ftir Avtzthamnion normalen Weise (Fig. 18 E). Ebenso verhalt es sich mit der Gonimoblastentwicklung (Fig. 18 F—G). Die reifen Gonimoblasten bestehen aus einigen Gonimoloben (Fig. 18 H), die bis etwa 70 ». in Durchmesser sind. Die Verzweigung ist bei dieser Art sehr regelmadssig. Besonders durch ihre Kleinheit ist sie von den iibrigen Arten gut getrennt. Sie wurde in 30— 45 m Tiefe epiphytisch auf Rhodymenia gefunden. Eaumnidiownt: “St. 5. Geogr. Verbr.: Juan Fernandez. DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 647 Fig. 18. Antithamnion minutissimum. A Oberer, B unterer Teil eines Sprosses; C Sperma- tangienstande; D Zweig mit zahlreichen Driisenzellen; E Prokarptum; F—H Gonimoblast- entwicklung. — A—B, H x 155; C—D X 265; E—G xX 445. Antithamnion sp. Von einer Art dieser Gattung habe ich einigemal junge, sterile, bis etwa I cm hohe Exemplare gesehen. Weil sie noch zu unentwickelt sind, ist eine sichere Bestimmung nicht mdéglich. Wahrscheinlich gehdren sie zu dA. occeden- tale (KYLIN 1925 S. 47) oder einer mit dieser naheverwandten, neuen Art. Pundorte: St. %5}. 8: Pleonosporium Nag. Pleonosporium globuliferum Levr. nov. sp. — Fig. 19. Frons 2—3 cm alta, mollis, fila inferne 140—170 wu crassa, vage ramosa, cellulae (2—)4—5-plo longiores quam latae, rhizoideis singulis praedita, superne regulariter et radiatim ramosa, sat cylindrica, cellulae 50—8o pu. crassae, 2— 3-plo longiores. Ramuli inferiores simplices, superiores ramosi. Polysporangia glo- bosa, 65—95 . diametro, in latere superiore insidentia. Antheridia 30—40 p crassa et 70—100 wv. longa. Der Thallus bildet kleine, 2—3 cm hohe, weiche Biischel. Die Haupt- zweige sind unten 140—170 ». dick, mit Zellen, die (2—)4—5mal so lang wie dick sind. Unten sind sie wenig verzweigt, tragt hier und da abwartslaufende 648 TORE LEVRING ao RS SEO eS ‘Gs , On SO ( Ap C/V kf <> d/ C4 ara’ IG, ey / J J? y G SF A GS 4 Fig. 19. Pleonosporium globuliferum. A Spross mit Sporangien, B mit Spermatangienstanden; C Habitus; D Unterer Teil mit Rhizoiden; E Sporangium. — A X 30; B, D—E X 100; C X 6. Rhizoide (Fig. 19 D). Eine zusammenhangende Rinde wird aber nicht gebildet. Oben ist die Verzweigung regelmassig. Die Seitenzweige sind ziemlich gleich- massig dick, mit Zellen, die etwa 50—8o0 ». dick und 2—3mal so lang sind. Die unteren Seitenzweige sind oft einfach, die oberen in der Regel mit Seiten- asten, die 30—40 ». dick sind, versehen. Die Zweige enden stumpf. Die Poly- sporangien sind kugelig, 65—95 yu. im Durchmesser, ungestielt und sitzen vor- zugsweise an der Oberseite der Seitenzweige. Die Antheridien sind 30—40 wp dick und 70—100». lang und wie die Sporangien geordnet. Weibliche Pflanzen habe ich leider nicht gesehen. Epiphytisch auf anderen Algen. Von anderen nahestehenden Arten unterscheidet sie sich besonders durch die spharischen Polysporangien. Mun dott: St5135 . Geogr. Verbr.: Juan Fernandez. Ceramium (Roth) Lyngb. Ceramium rubrum (Huds.) C. Ag.; J. Ag. 1894 S. 37. Die Exemplare sind 3—7 cm hoch und gehoren einer Form an, die mehr oder weniger reichlich mit kurzen Seitenasten versehen ist. Sie stimmt mit ahnlichen, europaischen Formen gut iberein. Fundorte: St. 10; HASSLER-Exp. Nr. 61; SCHMITT Nr. 150 (Carbajal Bay, 15.12. 1926, 15—20 fa.). Geogr. Verbr.: Fast kosmopolitisch. DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 649 Centroceras Kiitz. Centroceras clavulatum (C. Ag.) Mont.; J. Ag. 1851 S. 148; Howe 1914 S. 158; Ceramium clavulatum C. Ag. Von diesem liegt ein ziemlich reiches Material vor. Die Exemplare haben ein normales Aussehen. Fundorte: St)'2'c, 3) 5; HASSLER-Exp. Nr..72; SCHMITT Nr. 142. Geogr. Verb.: In warmeren Meeren weit verbreitet. Fam. Delesseriaceae. Hypoglossum Kutz. Hypoglossum parvulum Levr. nov. sp. — Fig. 20. Fronus membranacea, ad circ. 4 cm aita et 3 mm lata, a costa prolifera. Costa subecorticata. Folia ad circ. 8 mm longa, subovalia-lanceolata. Sori tetrasporangis multis, oblongis praediti. Der Thallus ist sehr diinn, bis etwa 4 cm hoch und 3 mm breit, und von der Mittelrippe reichlich proliferierend. Mittelrippe schwach berindet. Pro- lifikationen bis etwa 8 mm lang, fast oval bis lanzettlich, mit der grdssten Breite gewodhnlich oberhalb der Mitte des Blattes, gegen die Basis allmahlich verjiingt. Sori mit zahlreichen Tetrasporangien, etwas langgestreckt an beiden Seiten der Mittelrippe, in der Regel etwas unterhalb der Mitte des Blattes. Antheridien oder Zystokarpien habe ich nicht gesehen. Fig. 20. AHypoglossum parvulum. A, C, D Habitus 4lterer Thallusteile, B junger Pflanze; E Thallusspitze. — A—D X 5,5; E X 445. 43-—41438. The Nat. Hist. of Juan Fernandez and Easter Isl. Vol. II. 650 TORE LEVRING Der Habitus dieser Art geht aus den Figuren hervor. Am niachsten ist sie mit H/. attenuatum Gardner verwandt. Bei diesem haben aber die Prolifika- tionen eine breite Basis und werden gegen die Spitze hin allmahlich schmaler (vgl. GARDNER 1927 S. 104). HZ. parvulum dirfte auch mit H/. tenuzfolinm (Harv.) J. Ag. verwandt sein, das aber grdsser und habituell ziemlich gut ge- trennt ist. Es liegt nur ein sparliches Material vor. H/. parvulum wachst epiphytisch auf verschiedenen anderen Algen, wie Corallina, Fernandosiphonia und Clado- phora, und auf Aydrozoen. In der Regel habe ich nur vereinzelte Exemplare gefunden. Pun dome: St. 5,6, 11, i27 13: Geogr. Verbr.: Juan Fernandez. Phycodrys Kitz. Phycodrys quercifolia (Bory) Skottsberg 1922 S. 427; 1923 S. 35; Kylin 1929 S. 8 Taf. 3 Fig. 8; Delesseria quercifolia Bory 1828 S. 186 Taf. 18. In der Sammlung liegen einige bis 4 cm hohe, sterile Fragment vor, die hoéchstwahrscheinlich zu dieser Art geh6ren und junge Exemplare darstellen. Biwndierit: St. 5. Geogr. Verbr.: Feuerland, Falkland, Siidgeorgien, Graham Land, Vic- toria Land, Neuseeland, Juan Fernandez. Schizoseris Kylin. Schizoseris Davisii (Hook. f. et Harv.) Kylin 1924 S. 68; 1929 S. Io Taf. 7 Fig. 16; Taylor 1939 S. 156; Delesseria Davisit Hook. f. et Harv. 1845 S252 se45— 47.9. 1044 Dati 759 Kutz: Bab. 16 Tat. 18: Von dieser Art liegen einige bis etwa 5 cm hohe, sterile Fragmente vor. Es scheinen junge Exemplare zu sein. Die Ubereinstimmung ist gut. KyundiontbessSts 5.56: Geogr. Verbr.: Neuseeland, Falkland-Inseln, Feuerland, Juan Fernandez. Hymenena Grev. Hymenena decumbens Levr. nov. sp. — Fig. 21. Frons tenuis membranacea, ad circ. 2,5 cm longa, subtus rhizoideis prae- dita, laciniata vel pseudodichotome ramosa, lacinis 5—8 mm latis, obtusis. Venae multae, submicroscopicae. Sori sparsi, rotundati, circ. 350—500 p. dia- metro. Cystocarpia rotundata, 600-850 yp. diametro. Der Thallus ist diinn hautig, bis etwa 2,5 cm lang (oder vielleicht etwas mehr), niederliegend, an der Unterseite mit mehrzelligen, bis 1,5(—2) mm langen Rhizoiden versehen, die an Steinen, Lithothamnien usw. haften, und gezipfelt oder fast dichotomisch verzweigt. Die Thalluszipfel sind 5—8 mm breit und enden stumpf. Der Thallus ist mit zahlreichen parallel verlaufenden, verzweigten DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 651 arteer sea: eels: RY BTA Fig. 21. Hymenena decumbens. A, C Unterseite mit Rhizoiden; B Oberseite mit Sori; D, F—H Habitus; E Thallusrand. — A x 25; B—C x 6; D, F—H X 0,85; E x 170. Nerven versehen, die fast mikroskopisch klein sind. Sporangiensori tber die Thallusoberflache verstreut, etwa 350—500 y. im Durchmesser. Zystokarpien rund, etwa 600—850 yw. im Durchmesser. Von dieser Art liegt nur ein ausserst sparliches Material aus grosserer Tiefe vor. Besonders durch das Vorkommen von Rhizoiden ist sie von den ubrigen Arten gut getrennt. Hom dientet)St. 5)013: Geoor V erbr.y Juan Pernandez: Fam. Dasyaceae. Heterosiphonia Mont. Heterosiphonia subsecundata (Suhr) Falkenberg 1901 S. 641 Taf. 18 Fig. 20; Dasya subsecundata Suhr 1840 S. 280; Setchell and Gardner 1937 Rat 21-2). subsecuuda. Waryey 1847.45. 67: Val. 27:.). Ag. 1863S. 1st. Wie SETCHELL und GARDNER (1937 S. 84) gezeigt haben, ist die kali- fornische Form, die man frither zu H. swésecundata gestellt hat, eine besondere Art, H. erecta. Bei dieser gibt es vier Perizentralzellen, bei jener sieben. Ich habe sowohl Material von H. erecta als von H. subsecundata (Originalmaterial von Valparaiso, Herb. AGARDH Nr. 43613) untersucht und kann nur die An- gaben SETCHELLS und GARDNERS bestatigen. In dem Juan Fernandez-Material liegt eine Heterosiphonia vor, die un- 652 TORE LEVRING zweifelhaft zu 77. subsecundata gehort. Sie hat sieben Perizentralzellen und die Grossenverhaltnisse sind dieselben wie bei dem Original. Bei meinen Exem- plaren, die nur bis etwa 1 cm hoch und wahrscheinlich jung sind, sind die Glieder verhaltnismassig langer als bei dem Original. Zwischen je zwei Zweigen gibt es in der Regel zwei Glieder, bei dem Original etwa zwei bis vier. Fwmdotteswn St.voy ir Geogr. Verbr.: Chile, Juan Fernandez. Fam. Rhodomelaceae. Polysiphonia Grev. Polysiphonia abscissa Hook. f. et Harv. 1845 S. 266; 1847 S. 480 Taf. 183; Kitz. Tab. 13 Taf. 70; Howe 1914 S. 140; P. microcarpa Hook. f. and Harv. 1845S. 265; 1847 S: 479. Tafiire2: Von dieser Art liegt ein ziemlich reiches Material vor, das mit der Be- schreibung und den Exemplaren, die ich gesehen habe, gut tbereinstimmt. Wie Howe (l.c.) hervorgehoben hat, ist die Lange der Glieder etwas variabel, bei meinen Pflanzen etwa 1—3mal so lang wie breit. Sie messen bis etwa 7 cm und tragen Tetrasporangien oder Zystokarpien. Die Astzellen werden in einer Linksspirale angelegt und die neuen Seiten- triebe entwickeln sich direkt aus Astzellen. Anatomisch steht sie also P. 2- ceolata nahe, unterscheidet sich aber durch die Form der Zystokarpien. Andere Arten wie P. zxcompta Harv. und P. tuticorinens?s Borgs. scheinen von ihr weniger scharf getrennt zu sein. — Gefunden in der Litoral- und Sublitoral- region auf Steinen, Felsen und epiphytisch auf anderen Algen. Funidorte: 1, 2c, 4,5, 8, 10, 11, 12, 39a,b; Scumirr Nr.148 (Garbayal Bay, 15.12.1926, 15—20 Fa.). Geogr. Verbr.: Uruguay, Peru, Juan Fernandez, Feuerland, Falkland, Kerguelen, Neuseeland, Tasmanien, Graham Land.! Pterosiphonia Falkenb. Pterosiphonia pusilla Levr. nov. sp.; Polysiphonia parasitica var. reptans Dickie 1877 S: 452. — Fig.i22=-23: Frons ad 1,5(—2) cm alta, rhizoideis sat multis praedita. Rami 90—140 (—200) p. crassi, pinnatim ramosi; ramuli 2 articulis distantes, 60—80(—100) yp. crassi, eodem modo ramosi. Articuli diametro 0,5;—1,5(—2)-plo longiores, sipho- nibus 4, corticatis. Cystocarpia in ramulis sessilia, 110—120 yw crassa, 135— 145 wu. longa. Thallus bis etwa 1,5(2) cm lang, ganz oder nur mit den alteren Teilen kriechend und an der Unterlage (anderen Algen) mit Rhizoiden befestigt. Die 1 Polysiphonia villum J. Ag. (jetzt = Lophosiphonia villum Setch. et Gardn.) ist von Dickie (1887 S. 452) fiir Juan Fernandez angegeben (MoseLry). Im Herb. Kew gibt es aber von dieser Art kein Material, weshalb ich mich tiber die Richtigkeit der Bestimmung nicht dussern kann. Sie wurde sonst an der Westkiiste Nordamerikas angetroffen. DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 653 Fig. 22. Pterosiphonia pusilla. A Sprossspitze; B Habitus; C—D Querschnitte; E Zweig mit Prokarpien und Zystokarp; F Alterer Teil mit Rindenzellen; G—K Prokarpien- und Gonimo- blastentwicklung. — az Auxiliarzelle; bstz basale sterile Zelle; kp Karpogonast; Istz laterale sterile Zellen; tz Tragzelle. — A X 265; B, EX 45; C—D X 155; F X 100; G—K X 445. 654 TORE LEVRING roo ] TT wa Sasee: a {7 coer] tt ty W Fig. 23. Ptervosiphonia pusilla. Niederliegende Form. A—H Habitus; I Querschnitt eines alteren Teiles. — A—H X 45; | X 155. Hauptzweige, die 90—140(—200) p. dick sind, sind federartig verzweigt und zwar so, dass fast regelmassig jedes zweite Segment einen Seitenzweig ab- wechselnd nach rechts und links abgibt. Die Seitenzweige sind 60—80(—100) p. dick und in der Regel ahnlich wie die Mutterachsen verzweigt. Die Glieder sind 0,5—I,5(—2)mal so lang wie breit. Die Perizentralzellen sind vier. Auf- rechte Sprosse schwach berindet, kriechende etwas starker. Zystokarpien seit- lich an den Seitenzweigen, etwa 110—-120 yp. breit und 135—145 p. lang. Die neue Art ist sehr variabel. Dies hangt damit zusammen, bis zu welchem Grade die Pflanze an der Unterlage festgewachsen ist. An Exem- plaren, wo die Hauptzweige grésstenteils frei sind, ist die Verzweigung sehr regelmassig. Die Hauptzweige sind hier stielrund, selten mehr als 100 p dick und die Rinde ist sehr schwach entwickelt. Rindenzellen werden aber frith ab- geschieden und die Rinde wird immer vorzugsweise an den beiden lateralen Kanten des Thallus entwickelt. Dieses Aussehen hat die Pflanze, wenn sie frei unter anderen Fadenalgen wachst, wo die Rhizoidenbildung nicht befordert wird. Rhizoide werden offenbar leicht entwickelt. An der normalen Form treten sie verstreut auf. An geeigneten Unterlagen, wie grosseren Algen, wird die ganze Pflanze niederliegend (Fig. 23) und mit zahlreichen kiirzeren oder langeren Rhizoiden von verschiedenem Aussehen, die an der Unterseite ent- wickelt werden, daran befestigt. Im Zusammenhang damit bekommt der Thallus ein mehr oder weniger abgedndertes Aussehen. Oft wird die Rindenentwick- lung stark gefordert, aber stets an den beiden lateralen Kanten. Zystokarpien- tragende Exemplare habe ich nur ein paar und zwar unter den freiwachsenden gesehen, tetrasporen- oder antheridientragende keine. ee DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 655 Die Prokarpien- und Gonimoblastentwicklung findet in der bei den Rhodo- melaceen gewohnlichen Weise statt. Trichoblasten werden, wenn die Pflanze fertil werden soll, angelegt. Sie sind aber immer sehr kurz (Fig. 22 G). Das befruchtungsreife Prokarp besteht aus der fertilen Perizentralzelle, die als Trag- zelle dient, einem vierzelligen Karpogonast, zwei lateralen sterilen Zellen und einer basalen sterilen Zelle (Fig. 22 H—]). Bemerkenswert ist, dass die Art nur vier Perizentralzellen hat. Sie stimmt darin mit den folgenden zwei Arten tiberein. Bisher waren nur Pteros7phonia- Arten mit wenigstens fiinf Perizentralzellen bekannt. Die Art kommt in der Sublitoralregion vor. Die von DICKIE (I. c.) als Polysiphonia parasitica var. reptans angefihrte Form gehodrt nach seinem Ma- terial, das ich gesehen habe, auch hierher. Pumdorte: ow 4, 5,0, TT; 12+ DICKIE (lc, leg: MOSELEY): Geogr. Verbr.: Juan Fernandez. Pterosiphonia Skottsbergii Levr. nov. sp. — Fig. 24. Frons 2—3,5 cm alta, cylindrica. Rami (150—)175—250 yw, ad basin ad 350 w crassi, pinnatim ramosi; ramuli 2 articulis distantes, 70—120 », crassi, simplices. Articuli diametro 1—1,5-plo longiores, siphonibus 4, subecorticatis. Cystocarpia 150—160 yp, crassa, 195—210 wp longa. Thallus 2—3,5 cm hoch, stielrund, unten mittels Rhizoiden an der Unter- lage haftend. Die Hauptzweige, die (150—)175—250 p, im untersten Teil bis 350 w dick sind, sind federartig verzweigt und zwar so, dass fast jedes zweite Segment einen Seitenzweig abwechselnd nach rechts oder links abgibt. Die Seitenzweige sind 70—-120 », dick, schwach gekriimmt und in der Regel unver- zweigt. Hie und da wachst ein Seitenzweig zu einem Hauptspross aus. Die Glieder sind 1—1,5mal so lang wie breit. Die Anzahl der Perizentralzellen ist vier. Rindenzellen werden schon friih abgeschieden. Die Rinde bleibt jedoch A Hak spitze. — A X 25; B—D X 45; D X 100. 656 TORE LEVRING immer ziemlich schwach, scheint fast gleichformig rings um den Thallus ent- wickelt und also nicht an den lateralen Kanten besonders gef6rdert. Zysto- karpien seitlich an den Zweigen, 150—160 yp, breit und 195—210 y, lang. Diese Art ist nach Verzweigung und anatomischem Aufbau der vorigen sehr ahnlich, nur in allen Teilen viel grésser. Ubergangsformen habe ich nicht gesehen, obwohl sie zusammen gefunden worden sind. Die Seitenzweige sind bei Pt. Skottsberga im Gegensatz zu denen bei Pt. pumila in der Regel unver- zweigt. Auch ist die Rinde an den lateralen Kanten nicht gefordert. Die Neigung, Rhizoide zu bilden, scheint hier zu fehlen. An den Sprossspitzen kann man finden, dass die Seitensprosse nicht immer ganz bilateral angelegt werden, sondern etwas verschoben, doch nicht so stark, dass der Spross ra- diar wird. Es liegt nur ein geringes Material vor, das in 10—15 m Tiefe gesammelt war und grdsstenteils steril ist. Nur ein Exemplar habe ich mit Zystokarpien gefunden. Fundonrt: St. 3: Geogr. Verbr.: Juan Fernandez. Pterosiphonia disticha Levr. nov. sp. — Fig. 25. Frons ad c:a 2,5 cm longa, vage compressa, c:a 0,s;—0,8 mm Crassa, ramis multis distichis, alternantibus, c:a 0,7—1,3 mm longis praedita. Ramuli pinnis 5—7 distichis ornati. Siphones 4; cortex crassus. Ramuli ultimi distiche ex- euntes alterni, bi—tripinnati compressi. Der Thallus bis etwa 2,5 cm lang, seitlich verzweigt, etwas abgeflacht. Hauptzweige etwa 0,5—0,8 mm breit, dicht mit zweiseitig geordneten, alter- Fig, 25. Ptevosiphonia disticha. A Habitus; B Sprossspitze; C Seitenzweig; D Querschnitt eines alteren, E eines jumeeren’Teiles, — A x 16;.B)X.265; -C xX 95; D X 45: EioaauGy: DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 6 7 ul nierenden, etwa 0,7—I,3 mm langen Seitenzweigen besetzt. Vereinzelt konnen sich die Seitenzweige weiter entwickeln und neue Hauptzweige bilden. Die Seitenzweige (Fig. 25 C) tragen in der Regel 5—7 kurze Endzweige. Die An- zahl der Perizentralzellen ist vier. Die Rinde ist sehr gut entwickelt. Sie ent- steht friih, so dass man die Perizentralzellen auch in den jiingeren Teilen in Flachenansicht nicht sehen kann. Von dieser Art liegt nur ein dusserst sparliches Material vor, dass leider steril ist. Aus dem anatomischen Bau ergibt sich sofort, dass sie zu Ptero- siphonia gehort. Sie erinnert stark an /¢. clocophylla (C. Ag.) Falkenb. und complanata (Clem.) Falkenb. Weil gute anatomische Trennungsmerkmale vor- handen sind, habe ich sie aber als neu beschrieben. /7. désttcha hat nur vier Perizentralzellen, die anderen 5 oder 6. Bisher waren iiberhaupt keine P¢ero- siphonia-Arten mit nur 4 Perizentralzellen bekannt. Nur in Berindung und Ha- bitus erinnert sie an die beiden genannten Arten. Die Exemplare sind bis 2,5 cm hoch. Vielleicht wird die Art noch grdsser. — Epiphytisch auf Law- vencia clavata gefunden. Pun dor t:).St...40ib. Geogr. Verbr.: Juan Fernandez. Pterosiphonia dendroidea (Mont.) Falkenb. 1901 S. 268; Howe 1914 S. 144; Polysiphonia dendroidea Mont. 1837 S. 353; 1839 S. 16 Taf. 5:1; Kiitz. Tab. 13 Taf. 25; Prerosiphonia spinifera Kitz. |. c. Taf. 24; Pt. Bartlin- giana Kiitz. |. c. Taf. 24; Pt. calliptera Kitz. |. c. Taf. 25. Von dieser Art liegt ein ziemlich reiches Material vor. Die Exemplare sind bis 6, ausnahmsweise bis 10 cm hoch, und sind den oben angefihrten Formen von KUTzinG mehr oder weniger ahnlich. Die meisten Exemplare sind etwas diinner als fiir die Art normal. Ich habe nur tetrasporentragende und sterile Exemplare gefunden. Fundorte: St. 4, 5, 8, 11, 12; SCHMITT Nr. 144 (9.12. 1926). Geogr. Verbr.: Westkiiste von Amerika von Chile bis U.S.A. Stromatocarpus Falkenb. Stromatocarpus microscopicus Levr. nov. sp. — Fig. 26—27, Taf. 53, Bic. 3,4. Frondes pulvinulos ad 1,7 mm diametro formantes, filamentis basalibus endophyticis hostem (Prerosiphonia pusilla et Skottsbergii) penetrantibus, mono- siphoneis, sparse ramosis. Tetrasporangia serie singulari in stichidiis cylindricis, curvatis evoluta; antheridia oblongo-ellipsoidea, 30—60 p crassa, I10—I50 longa; cystocarpia 175 y. crassa, 250 ». vel ultra longa. Pili nulli. Die Art lebt parasitisch auf Prerosiphonia pusilla. Einigemal habe ich sie auch auf Pr. Skottsbergii gefunden. Sie bildet kleine Biischel. Die Tetra- sporenpflanzen werden bis etwa 1,7 mm, die Geschlechtspflanzen kaum I mm hoch. Die Basis besteht aus einem System von monosiphonen Faden, die in die Wirtspflanze hineindringen und sich dort endophytisch entwickeln. Daraus erheben sich eine Anzahl aufrechter, freier, polysiphoner Zweige, die (30—)45 —75(—110) ». dick sind. Der Thallus ist monopodial und radiar gebaut. Die 658 TORE LEVRING Fig. 26. Stromatocarpus microscopicus. A Spross mit Prokarpien und Zystokarpien, B mit Tetrasporen, C mit Spermatangienstinden; D Anlagen von Spermatangienstianden; E Keim- pflanze. — A—C X 170; D x 450; E X 265. Astzellen werden in einer Linksspirale etwa mit der Divergenz 1/4 angelegt und zwar so, dass von fast jedem Segment eine Astzelle abgegliedert wird. Die Perizentralzellen sind 5 bis 7. Die erste wird wie bei Polyszphonia unter der Astzelle abgegliedert. An einer sterilen Sprossspitze fehlen Trichoblasten ganz. Die weitere Entwicklung der Astzellen ist bei den verschieden Pflanzen ver- schieden. Rinde fehlt. Bei den Tetrasporenpflanzen (Fig. 26 B) werden die Astzellen als polysiphone Seitenglieder entwickelt. Deshalb sind diese Pflanzen sehr reich verzweigt. Die Seitenzweige, die immer eingekriimmt sind, sind sehr deutlich spiralig geordnet. Die Tetrasporangien werden in den Seitenzweigen ahnlich wie bei Polystphonia entwickelt. Jedes Segment bildet also ein Sporangium. Bei den Geschlechtspflanzen entwickelt sich fast jede Astzelle als fertiler Tricho- blast. Nur sehr selten wird ein normaler Seitenzweig entwickelt. Deshalb sind die Geschlechtspflanzen viel weniger verzweigt als die Tetrasporenpflanzen und auch etwas kleiner. An ihren Zweigen sitzen die Antheridien bzw. Prokarpien (Zystokarpien) spiralig geordnet., Die Spermatangienstande, die sich also aus Trichoblasten entwickeln, haben wie bei Polyszphonia usw. einen zweizelligen Stiel, dessen obere Zelle indessen im Gegensatz zu den anderen keinen Seitenast tragt (Fig. 26 D). Ausnahmsweise DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN =kp \ \ -\-Istz mn \ Fig. 27. Stromatocarpus micvroscopicus. Prokarpienentwicklung. — bstz basale sterile Zelle; kp Karpogonast; Istz laterale sterile Zellen; tz Tragzelle. — x 585. ist der Stiel einzellig. Gelegentlich konnen auch Langswande auftreten. Die Spermatangienstande sind ausgezogen ellipsoidisch oder fast zylindrisch, etwas eingekriimmt, 30—60 uv. dick und 110—150 uw. lang. Die Prokarpien- und Gonimoblastentwicklung findet in gewohnlicher Weise statt und braucht deshalb nicht eingehender erortert zu werden. Die zweit- unterste Zelle der Trichoblasten stellt die fertile Zentralzelle dar (Fig. 27 A—B). Der obere Teil der Trichoblasten wird nur schwach entwickelt oder stirbt bald ab. Die fertile Zentralzelle spaltet 5 Perizentralzellen ab. Das _befruchtungs- reife Prokarp besteht aus der fertilen Perizentralzelle (Tragzelle), einem vierzel- ligen Karpogonast, zwei lateralen und einer basalen sterilen Zelle (Fig. 27 C—E). Nach der Befruchtung werden zuerst die Auxiliarzelle und von den drei sterilen Zellen je eine neue Zelle abgeschieden. Die Fruchthille scheint anfangs schwach entwickelt zu sein. Die reifen Zystokarpien sind 175 wu dick und 250 wu. lang, k6nnen aber wahrscheinlich noch etwas grOsser werden. Die Gattung Stromatocarpus wurde von FALKENBERG (1901 S. 185) mit der Art S¢r. parasttica von Siidafrika aufgestellt. Dass die oben beschriebene Art auch zu dieser Gattung gehoOrt, diirfte keinem Zweifel unterliegen, obwohl die Figuren FALKENBERGS nicht besonders lehrreich sind. Von S¢r. parasitica unterscheidet sie sich in mehreren Punkten: Grdsse, Anzahl der Perizentral- zellen usw. Von SETCHELL (1923 S. 395) wurde St. Gardnert von Kalifornien beschrieben, aber nicht abgebildet. Aus der Beschreibung geht aber deutlich hervor, dass die neue Art nicht damit identisch sein kann. U. a. hat sie wie Str. parasitica 4 Perizentralzellen. Was die systematische Stellung der Gattung betrifft, so ist sie wohl als ein reduzierter Typ aufzufassen und scheint gewisse verwandtschaftliche Be- ziehungen zu Pterosiphonia zu haben. Bezeichnend fiir beide ist das Fehlen der Trichoblasten. Pverosiphonia ist aber im Gegensatz zu Stromatocarpus, welcher radiar ist, bilateral gebaut. Gelegentlich kann an Pteros7phonia Arten eine spiralige Anordnung schwach angedeutet sein (vgl. S. 656; FALKENBERG l.c. S. 262). Wie SUNESON (1940 S. 322) fir Pterosiphonia parasitica gezeigt 660 TORE LEVRING hat, sind die Antheridien hier spiralig geordnet. Sie entsprechen Trichoblasten, wo die zwei untersten Zellen steril sind. Die zweitunterste tragt keinen Seiten- ast. Die Ubereinstimmung mit Szromatocarpus in diesem Punkt ist also auf- fallend. Der radiare Bau macht offenbar einen primitiven Eindruck und deutet auf Polysiphonia-ahnliche Formen. — Wurde in 10—15 m Tiefe angetroffen. Pundort: St: 3: Geogr. Verbr.: Juan Fernandez. Fernandosiphonia Levr. nov. gen. Frons filiformis, cylindrica, dense et iteratim unilateraliter ramosa, mono- podialis, trichoblastiis spiraliter insidentibus, caducis. Antheridia et cystocarpia in ramulis unilateralia. Tetrasporangia in serie singulari evoluta. Thallus fadenformig, stielrund, reich verzweigt, monopodial gebaut, die Seitenglieder in einer linkslaufigen Spirale angeordnet; sie werden teils als Trichoblasten, die verzweigt und hinfallig sind, teils als polysiphone Seiten- zweige entwickelt. Die Seitenzweige werden aber nur auf der Innenseite der Mutterachse entwickelt, so dass die Verzweigung wiederholt einseitig wird. Ahnlich werden die Antheridien und Zystokarpien nur auf der Innenseite ihrer Mutterachse entwickelt. Von Tetrasporangien wird nur eines in jedem Seg- ment erzeugt. Fernandosiphonia unilateralis Levr. nov. sp. — Fig. 28, Taf. 53, Fig. 2. Frons caespitosa, circ. 5—10 cm alta, robusta, ramis distinctis, vage ra- mosis, inferne 0,75—1,25(—I,s) mm crassis, superne sensim attenuatis. Ramulli multi, circ. 5—10 mm longi, squarrosi, repetite unilateraliter ramosi, articulis 4-siphoniis, corticatis, diametro 1'/:—'/s-plo longioribus. Cystocarpia globoso- ovoidea. Bildet etwa 5—10 cm hohe, kraftige Biischel. Die Hauptzweige, die einigemal seitlich verzweigt sind, sind in ihrer ganzen Lange deutlich, unten etwa 0,75—I,25(—I,s) mm dick, nach oben allmahlich diinner, nahe der Spitze etwa 0,35—0,5 mm. Die Hauptzweige tragen zahlreiche, etwa 5—10 mm lange, abstehende, oft schwach gekriimmte Seitenzweige, die oft mehr oder wenig deutlich einseitig geordnet sind. Sie tragen an ihrer Innenseite eine Reihe (etwa 4—7) etwa I—1I.5 mm lange Endverzweigungen, die in ahnlicher Weise auf ihrer Innenseite Seitenadste tragen. Die Glieder sind '/2—'/smal so lang wie breit und bestehen aus 4 Perizentralzellen. Der Thallus ist bis hoch hin- auf berindet. Die Zystokarpien sind eiformig. Die Sprossspitze ist der bei Polyszphonza in mehreren Hinsichten ahnlich (Fig. 28 D). Der Thallusaufbau ist monopodial. Die Thallusspitze scheint primar radiar zu sein. Sie ist etwas nach innen gekriimmt. Die Astzellen werden in einer Linksspirale etwa mit der Divergenz 2/5 angelegt. Jedes Seg- ment besteht aus vier Perizentralzellen. Die erste scheint wie bei Polyszphonia unter der Astzelle abgegliedert zu werden. Die Astzellen entwickeln sich zu monosiphonen Trichoblasten oder polysiphonen Seitentrieben. Die Trichoblasten sind verzweigt und hinfallig. Ihre unterste Zelle ist unverzweigt, die nachste DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 661 IN 4 i ageutiys Fig. 28. Fernandosiphonia unilateralis, A _Spross mit Spermatangienstanden, B mit Zysto- karpien, E mit Tetrasporen; C_Jingerer, F alterer Teil mit Rindenzellen; D Sterile Spross- spitze; G Querschnitt jéngerer, H alterer Teil. — A—C, E, H X 45; D X 290; F X 25; G X 170. bildet wie bei Polys¢fhonza einen Seitenast, der nach rechts gerichtet ist. Die polysiphonen Seitenglieder werden nur auf der Innenseite der Mutterachse ent- wickelt, im Gegensatz zu den Trichoblasten, die also spiralig, allseitig geordnet sind. An einem ausgewachsenen Spross sitzen somit die Seitenglieder auf der Innenseite gereiht. Sie kénnen aber etwas nach der rechten oder linken Flanke der Mutterachse verschoben sein. Diese einseitige Verzweigung wiederholt sich einigemal. Nur die Langsprosse scheinen allseitig (wahrscheinlich durch Ver- schiebungen) verzweigt zu sein. Sie tragen ihre Seitensprosse oft einseitig, jedenfalls in den oberen Teilen. Die Seitensprosse sind wiederholt und deutlich einseitig verzweigt. Diese hier beschriebene Form stimmt mit S¢reélocladia darin wberein, dass die Seitenzweige auf der Innenseite der Mutterachse entwickelt werden. Bei dieser Gattung fehlen aber die Trichoblasten und die Astzellen sind nicht spiralig angeordnet. Sie werden nur auf der Innenseite des Sprosses, der also primar dorsiventral ist, entwickelt. Die Spermatangienstainde entwickeln sich in den oberen Teilen der letzten Verzweigungen. Sie entsprechen wie bei Polyszphonia Trichoblasten. Sie haben einen zweizelligen Stiel, dessen unterste Zelle unverzweigt ist und dessen zweite 662 TORE LEVRING Fig. 29. Fernandosiphonia unilateralis. A, C—D_ Prokarpien und Gonimoblastentwicklung; B Anlage eines Spermatangienstandes. — az Auxiliarzelle; bstz basale sterile Zelle; kp Karpo- gonast; Istz laterale sterile Zellen; tz Tragzelle. — A, C—D X 610; B X 395. Zelle ahnlich wie die sterilen Trichoblasten an ihrer rechten Seite einen Seitenast tragt (Fig. 29 B). Bemerkenswert ist, dass fast ausschliesslich nur diejenigen Trichoblasten, die auf der Innenseite der Mutterachse sitzen, als Spermatangien- stande enwickelt werden (Fig. 28 A). Die Prokarpien werden wie bei den meisten Rhodomelaceen an den Tricho- blasten entwickelt, deren zweitunterste Zelle fertil ist. Die Prokarpienentwick- lung scheint vom gewohnlichen Typus zu sein (Fig. 29 A, C—D). Die fertile Zentralzelle spaltet ihre fertile Perizentralzelle nach innen ab. Diese, die als Tragzelle dient, spaltet eine laterale sterile Zelle, die sich spater in zwei gliedert, einen Karpogonast, der vierzellig ist und dessen drei unterste Zellen zweikernig sind, sowie eine basale sterile Zelle ab. Nach der Befruchtung gliedert die fertile Perizentralzelle nach oben eine Auxiliarzelle ab. Die drei sterilen Zellen erzeugen je eine neue Zelle. Die Zystokarpienhille wachst heran und die Karposporen werden gebildet. Wie bei den Spermatangienstanden, werden die Zystokarpien auch nur auf der Innenseite der Mutterachsen gebildet (Fig. 28 B). Die Tetrasporangien werden in den jiingeren Teilen der Verzweigungen in ahnlicher Weise wie bei Polysiphonia gebildet. Jedes fertile Segment erzeugt also nur ein Sporangium. Diejenigen Endverzweigungen, die Sporangien tragen, sind oft besonders reich verzweigt, so dass sie fast kleine Bischel an den Mutterachsen bilden. Von dieser Art liegt ein reichliches Material vor. Sie wurde auf Coral- linaceen und anderen Algen angetroffen. Fundorte: St. 2c, 11, 12; SCHMITT Nr. 148 (Carbajal Bay, 15.12. 1926, 15—20 Fa.). Geoor. Verbr.: Juan Fernandez. ee Saas —ee DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 663 Fig. 30. Dipterosiphonia parva. A—B Habitus; C—E Querschnitte; F Seitenzweig. — A—B X98; C—E >< 1fo;) EX 170. Dipterosiphonia Schmitz et Falkenb. Dipterosiphonia parva (Dickie) Skottsb. et Levr. nov. comb.; Pélota parva Dickie 1877 S. 452. — Fig. 30. Von dieser Art liegt nur ein sehr sparliches Material vor, das aber gut mit dem Original, das ich auch untersucht habe, tbereinstimmt. Die Art wurde von DICKIE nach Material von Juan Fernandez aufgestellt. Der anato- mische Bau zeigt aber, dass sie zu Dipterosiphonia gestellt werden muss. Der Thallus ist kriechend, mit Rhizoiden an der Unterlage (anderen Algen) befestigt und (nach DICKIE) bis etwa 2 cm hoch. Die Hauptzweige sind fast regelmassig verzweigt. Sie tragen in der fiir die Gattung bezeich- nenden Weise links und rechts abwechselnde, abstehende, federartig verzweigte Langtriebe, die bisweilen auch unverzweigt sein k6dnnen, und dorsale Kurz- triebe, die niederliegend sind. Auf der Unterseite werden Rhizoide entwickelt. Die Perizentralzellen sind 8—10. Die Glieder sind 1—'/2mal so lang wie breit. Die Hauptzweige sind etwa 150—240y. dick. Fortpflanzungsorgane sind nicht bekannt. D. parva scheint D. heteroclada (J. Ag.) Falkenb. von Neuseeland ziem- lich nahe zu stehen. Bei dieser ist die Anzahl der Perizentralzellen 6—8 und die Glieder sind etwa '/2mal so lang wie breit. Ferner sind die Zweige nicht so stark abstehend wie es bei J. farva der Fall ist. — Epiphytisch auf an- deren Algen. 664 TORE LEVRING Funderte: St. 11; Dickie (1. ¢, leg. MOSELEY); ScHmirr Nigga (8.12. 1926). Geogr. Verbr.: Juan Fernandez. Laurencia Lamour. Laurencia obtusa (Huds.) Lamour.; Yamada 1931 S. 222. Von dieser Art fiihrt YAMADA (l.c.) mehrere Formen an. In meinem Material liegen ein paar, etwa 3—4 cm hohe Exemplare vor, die ich mit ge- wissem Zo6gern zu dieser Art stelle. Einige Exemplare sind ziemlich sparlich, eines aber sehr dicht verzweigt. Die Zweige sind bis etwa 1 mm dick. Die Zellwande sind wenigsten in den 4lteren Teilen ziemlich dick. Wenn man der Art weite Grenzen gibt, lassen sich diese Formen wahrscheinlich hier ein- reihen; das Material ist aber zu gering, um sichere Schliisse zu erlauben. Fundorte.St-.12, 30ib, Geogr. Verbr.: Weit verbreitet in warmeren Meeren. Laurencia clavata Sonder 1852 S. 694; Yamada 1931 S. 228; Chondria clavata (Sond.) Harv. 1862 Taf. 189; Corynecladia clavata (Sond.) J. Ag. 1876 5043. Von dieser Art liegen einige bis etwa 7 cm hohe Exemplare vor. Ich habe sie mit Material im Herb. AGARDH (u. a. Exemplare von HARVEY) ver- glichen. Die Ubereinstimmung ist gut, nur sind sie etwas kleiner als die austra- lischen. “Fertil. Bundort? St.-40 b: Geogr. Verbr.: Australien, Malayische Inseln, Juan Fernandez. Genus incertae sedis. Hildenbrandia Nardo. Hildenbrandia prototypus Nardo; Hauck 1885 S. 38; HH. rosea Kiitz. 1843 S. 384. Schon entwickelt mit Tetrasporen. Boundiort?2 Sp, 2)ai Geogr. Verbr.: Weit verbreitet. DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 665 Literaturverzeichnis. AGARDH, C. A., Icones algarum ineditae II. Lund 1821. —-— Species algarum rite cognitae I. Lund 1823. AGARDH, J. G., In historiam algarum symbolae. — Linnaea 15. Halle 1841. —-— Species, genera et ordines algarum, Vol. I (1848), II: 1 (1851), Il: 2 (1852), II: 3 (1863). Lund 1848—63. —-— Epicricis systematis floridearum. Lund 1876. —— Till algernas systematik, Afd. 2. — Lunds Univ. Arsskrift 17. Lund 1882. —— Till algernas systematik, Afd. 5. — Ibidem 23, 1887. —-— Analecta algologica, Contin. Il. — Ibidem 30, 1894. AHLNER, K., Bidrag till kinnedomen om de svenska formerna af algslagtet Entero- morpha. Akad. afhandl. Stockholm 1877. Askenasy, E., Algen. —— Die Forschungsreise S. M. S. Gazelle, IV Theil, Bot. Berlin 1888. Bunks, L. R., and A. H., Two genera of algae new to Bermuda. — Bull. Torrey Bot. Club, Vol. 57. New York 1930. BOrGESEN, F., Some new or little known West Indian Florideae. II. — Bot. Tids- skrift, Bd. 30. Kobenhavn 1gto. —— Two crustaceous brown algae from the Danish West Indies. —- Nuova Nota- risia, S. 25. Padova 1912. —-— The marine algae of the Danish West Indies, Vol. I: 1 (1913), I: 2 (1914), II (1915—20). Kobenhavn 1913—20. —-— Marine algae trom Easter Island. —- The Natural History of Juan Fernandez and Easter Island, ed. C. Skorrsperc, Vel. II]. Uppsala 1924. ——— Marine algae from the Canary Islands I—III. — K. Danske Vidensk. Selsk., Biol. Meddel., Bd. 5, Kobenhavn 1925 (I); 6, 1926 (II); 6, 1927 (III: 1); 8, 1929 (IIl:.2); 9, 1930 (LI: 3). —— Contributions to a south Indian marine algal flora I. — Journ. Indian bot. soc., Vol. 16. Madras 1937. —— Some marine algae from Mauritius. I. Chlorophyceae. — K. Danske Vidensk. Selsk., Biol. Meddel., Bd. 15. Kobenhavn 1940. Bory DE Satnt-VINCENT, J. B., Cryptogamie. — Voyage autor du monde... sur... La Coquille, par M. L. I. Duperrey.' Paris: 1828. Couns, F. S., Hoxpen, I., SercHeLtt, W. A., Phycotheca Boreali-Americana, Fasc. 1—46. Malden 1895—r1919. Corron, A. D., Cryptogams from the Falkland Islands collected by Mrs. Vallentin. Journ. Linn. soc., Bot., Vol. 43. London 1915. DecalsneE, J., Mémoire sur les Corallines ou Polypiers calciferes. —- Ann. sc. nat., Bot, S.i2 PoiS>s ~Paris #842. Dersts, A., et Sorter, A., Mémoire sur qulques pointes de la physiologie des algues. Paris 1856. De Tont, J. B., Sylloge algarum, Vol. I, Patavii 1889; III, 1895; IV, 1897— 1903; VI, 1924. 44— 41438. The Nat. Hist. of Juan Fernandez and Easter Isl. Vel. II. 666 TORE LEVRING De Ton, G. B., e Forti, A., Alghe di Australia, Tasmania e Nuova Zelanda. — Mem. Istit. Veneto sc., lett. ed arti, -Vol. 29. Venezia 10922. Dickiz, G., Notes on algae collected by H.N. Moseley, M.A., of H.M.S. »Chal- lenger»®. — Journ. Linn. soc., Bot., Vol. 15. London 1877. Drew, Karu_een M., A revision of the genera Chantransia, Rhodochorton, and Acrochaetium. —— Univ. Calif. Publ., Bot., Vol. 14. Berkeley 1928. ENGLER, A., und PRranttL, K., Die natiirlichen Pflanzenfamilien. Erste Aufl. Bd. I: 2. Leipzig 1897. : FALKENBERG, P., Die Rhodomelaceen des Golfes von Neapel und der angrenzenden Meeresabschnitte. —- Fauna and Flora des Golfes von Neapel, 26, Mono- graphie. Berlin rgot. Fartow, W. G., Thallophytes and Musci of the Galapagos Islands. — Proc. Amer. Acad. arts and sc., Vol. 38. Boston 1902. FELDMANN, J., Remarques sur les genres Gelidium Lamour., Gelidiopsis Schmitz et Echinocaulon (Kiitz.) emend. — ‘Trav. cryptogam. dédies a L. Mangin. Pariser o sie ——— Une Nemalionale a carpotétraspores: Helminthocladia Hudsoni (C. Ag.) J. Ag. — Bull. soc.'@hist= nat.) l’Afrique du- Nord; *T. “30: Alger*1o30; FELDMANN, J., et Hamer, G., Observations sur quelques Gélidiacées. — Rev. génér. bot;; ©. 46. Paris@rg24e —-=' Flondées*de France; V1? *«Gélidiales?-—* Rev.valgologs 1.0%. sRarissuaga: GARDNER, N. L., New Rhodophyceae from the Pacific Coast of Nort America. VI. Univ: Calif.cPubl= Bot., “Vol:+r4se Berkeley’'192 7. GREVILLE, R. K., Scottish Cryptogamic Flora. Edinburgh 1823—29. HameL, .G., Chlorophycées de France. — Rev. algolog., T. 6: Paris 1931. Harvey, W. H., Nereis Boreali-Americana. I. Washington 1852. —— Phycologia Australica, Vol. 1 (1858), 2 (1859), 3 (1860), 4 (1862), 5 (1863). London 1858—63. ——- Phycologia Britannica, Vol. 1— 4. London 1871. Hauck, F., Die Meeresalgen Deutschlands und Oesterreichs. — RapeNuorst, Krypto- gamenflora von Deutschland, 2. Leipzig 1885. ——— Ueber einige von J. M. Hildebrandt im Rothen Meere und Indischen Ocean gesammelte Algen. IV. Padina Adans. — Hedwigia, Bd. 26. Dresden 1887. Hooker, J. D., and Harvey, W. H., Algae Antarcticae. —- London Journ. bot., Vol. 4. London 1845 (a). = Algae Novae Zelandiae. —— Ibidem, 1845 (b). —— Algae. — J. D. Hooker, The Cryptogamic Botany of the Antarctic voyage. London 1845 (c). =r Aigaes—— J]. D. Hooxer, The Botany oft the«Antarctic) voyages elec: antarctica. London 1847. Howe, M. A., The marine algae of Peru. — Mem. Torrey Bot. Club, Vol. 15. New York 1914. —— Algae. — Britton, N. L., and Mituspaucn, C: F., The Bahama flora. New York 1920. Jorpe, INGERIp, Untersuchungen iiber den Lebenszyklus von Urospora Aresch. und Codiolum A. Braun. — Nyt Magazin for Naturvidensk., Bd. 73. Oslo HIS) 6 ic KjeLLMAN, F. R., Norra Ishafvets algflora. —- Vegaexpeditionens vetenskapliga iakt- tagelser, 3. Stockholm 1883 (In Ubersetzung: The algae of the Arctic Sea. — K. Svenska Vet.-Akad. Handl., Bd. 20. Stockholm 1883.) -—— Handbok i Skandinaviens hafsalgflora. I Fucoideae. Stockholm 1890. ——— Marina Chlorophyceer fran Japan. — Bihang till K. Svenska Vet.-Akad. handl., Bd. 23. Stockholm 1897. DIE MEERESALGEN DER JUAN FERNANDEZ-INSELN 667 KornMann, P., Zur Entwicklungsgeschichte von Derbesia und Halicystis. — Planta, Bd. 28. Berlin 1938. Kuckuck, P., Uber den Bau und die Fortpflanzung von Halicystis Areschoug und Valonia Ginnani. — Bot. Zeitung, Jahrg. 65. Leipzig 1907. —— Beitrige zur Kenntnis der Meeresalgen. 13. — Wissensch. Meeresunters., N.F. Abt. Helgoland, Bd. 5. Oldenburg i Gr. 1912. ——— Fragmente einer Monographie der Phaeosporeen. Nach dem Tode des Ver- fassers herausgegeben von W. NieNnBuRG. — Wissensch. Meeresunters., N. F. Abt. Helgoland, Bd. 17. Kiel und Leipzig 1929. Kuntze, O., Revisio generum plantarum secundum leges nomenclaturae internatio- nales, Vol. III: 2. Leipzig 1898. Ktrzinc, F. T., Phycologia generalis. Leipzig 1843. ——— Phycologia germanica. Nordhausen 1845. —— Tabulae phycologicae. 1—19. Nordhausen 1845—69. —— Species algarum. Leipzig 1849. Kyun, H., Studien iiber die Algenflora der schwedischen Westkiiste. Dissert. Upsala 1907. ——-— Studien iiber die Delesseriaceen. —- Lunds Univ. Arsskrift, N. F. Avd. 2, Bd 20. Lund 1924. —— The marine red algae in the vicinity of the Biological Station at Friday Harbor Wash. — Ibidem Bd 21, 1925. , —— Die Delesseriaceen Neu-Seelands. — Ibidem Bd 25, 1929. —— Uber die Entwicklungsgeschichte der Florideen. — Ibidem Bd 26, 1930. —— Die Florideenordnung Rhodymeniales. —— Ibidem Bd 27, 1931. —— Uber die Entwicklungsgeschichte der Phaeophyceen. — Ibidem Bd 29, 1933. —— Zur Kenntnis der Entwicklungsgeschichte einiger Phaeophyceen. — Ibidem Bd 30, 1934. —— Anatomie der Rhodophyceen. — Linspauers Handbuch der Pflanzenanatomie. ireBea: VI 2B. ¢g.. Berlim\1937: —— Die Phaeophyceenordnung Chordariales. — Lunds Univ. Arsskrift, N. F. Avd. 2 Bd 36. Lund rg 4o. Kyun, H., und Sxorrssgerc, C., Zur Kenntnis der subantarktischen und antark- tischen Meeresalgen. II Rhodophyceen. — Wissensch. Ergebn. d. schwed. Siidpolarexp. 1901-03, Bd IV: 15. Stockholm 1919. Laceruem, G., Bidrag till Sveriges algflora. — Ofversigt af Vet.-Akad. Forhandl. Stockholm 1883. Le Joris, A., Liste des algues marines de Cherbourg. — Mém. soc. imp. sc. nat. Cherbourg, 10. Paris et Cherbourg 1863. Levrinc, T., Zur Kenntnis der Algenflora der norwegischen Westkiiste. —- Lunds Univ. Arsskrift, N.F. Avd. 2 Bd 33. Lund 1937. —— Verzeichnis einiger Chlorophyceen und Phaeophyceen von Siidafrika. — Ibidem Bd 34, 1938. —— Die Phaeophyceengattungen Chlanidophora, Distromium und Syringoderma. — K. Fysiogr. Sallsk. Forhandl., Bd 10. Lund 1g4o. Lynceyre, H. C., Tentamen Hydrophytologiae Danicae. Kobenhayn 1819. MitcHert, Marcarer O., On the structure of Hydroclathrus Bory. — Phycol. me- moirs ed. by G. Murray. London 1893. MITCHELL, MarcaReT O., and Wuirtinc, Frances G., On Splachnidium rugosum, Grev., the type of a new order of algae. — Ibidem, 1892. Monvtacne, C., Prodromus Florae Fernandesianae. Pars prima. — Ann. sc. nat., Bot 39.) 2. Vs 2. Pas) 7835. —-— Centurie de plantes cellulaires exotiques nouvelles. —- Ibidem, S. 2 T. 8, 1337. 668 TORE LEVRING MonvTacGne, C., Cryptogamie. — Voyage dans L’Amérique méridionale par M. Alcide d’Orbigny. Botanique, Sertum Patagonicum et Flora Boliviensis. Paris 1839. —— Cryptogames cellulaires. —- Voyage autour du monde exécuté pendant les années 1836 et 1837 sur la corvette La Bonite, T. 1. Paris 1844—,46. ——, Algas. — Gay, C., Historia fisica y politica de Chile, Botanica, T. 8. Paris 1852 (a). —— Diagnoses phycologicae. — Ann sc. nat., Bot., S. 3 T. 18. Paris 1852 (b). Murray, G., On Halicystis and Valonia. — Phycol. mem. ed. by G. Murray. London 1893. Oxamura, K., Contributions to the knowledge of the marine algae of Japan. III. == .Bot. magaz.,/ Vol. 13.. Tokyo) 189@: —— Icones of Japanese algae, Vol. 1-7. Tokyo 1907—1934. ——-— The distribution of marine algae in Pacific waters. —- Records of oceanogr. works in Japan, Vol. 4. Tokyo 1932. OLTMANNS, F., Morphologie und Biologie der Algen. 2. Aufl., Bd. 2. Jena 1922. Papenruss, G. F., Notes on South African Marine Algae. I. — Bot. Notiser. Lund 1940. REINKE, J., Zwei parasitische Algen. — Bot. Zeitung, Jahrg. 37. Leipzig 1879. Ror, Maeet Lewis, A contribution to our knowledge of Splachnidium. — Bot. Gazette, Vol. 62. Chicago 1916. ROsENVINGE, L. K., The marine algae of Denmark, Voli: 1. — Ke Danske Vidensk. Selsk. Skrifter, 7. Rekke, Naturv. og Mathem. Afd., VII, 1. Koben- havn 190g. SAUNDERS, D. A., Phycological Memoirs. — Proc. Calif. Acad. Sc., S. 3, Bot., Vol. 1. San Francisco 1808. SAUVAGEAU, C., Note sur 1’»Ectocarpus pusillus» Griffiths. — Journ. bot., T. 9. Paris 1895. —-— Sur l’»Ectocarpus virescens» Thuret et ses deux sortes de sporanges pluri- loculaires. —=_Journ: bots, 1. 19: Pansaisgo6: is Sur quelques Myrionémacées. — Ann. sc. nat., Bot., S. 8, T. 5. Paris 1897. —— Remarques*sur les sphacélariacées. — Journ. bot. Paris 1900—14. Scumipt, O. C., Uber die Tiipfelung in den Zentralzellen der Dictyta dichotoma. —Hedwigia, Bde 72 ~ Dresden °1.93'8: SETCHELL, W.-A:;. Parasitic forideae. Il.-——-Univ., -Calif.. Publ., Bot.;agVolwae: Berkeley 1923. ———- J NOteSsoneMicrodictyon. 1. —=,Ibidem= Volii3. 10925: —— The genus Microdictyon. — Ibidem Vol. 14. 1929. —— The Codiums of the Juan Fernandez Islands. — The Natural History of Juan Fernandez and Easter Island, ed. by C. Skorrsperc, Vol. II. Uppsala 1937- SETCHELL, W. A., and Garpner, N. L., Algae of northwestern America. — Univ. Galt} Publ.,; Bot., Vol: 1. Berkeley 1903. —- The marine algae of the pacific coast of North America. II Chlorophyceae, III Melanophyceae. — Ibidem Vol. 8, 1920, 1925. —-— Phycological contributions. II to VI. — Ibidem Vol. 7, 1922. —— The marine algae (In Expedition of the California Academy of Science to the» Gulf -of California in 1g20); ==)Proc. CalifsiAcad) Se_, » » wi is » » 92 setze 93 » » » 6 » » » I 9 > 20 670 TORE LEVRING Tafelerklirungen. Taf. 49. 1. Ulva nematoidea Bory. Original im Herb. AGARDH. %/s nat. Gr. Padina fernandeziana Skottsb, et Levr. Original, leg. Skotrsperc. Etwa */10 nat. Gr. NV Taf. 50. 1. Dictyota phlyctaenodes Mont., leg. SKOTTSBERG. Etwa */s nat. Gr. 2. Liagora brachyclada Dene, leg. SKOTTSBERG. Etwa nat. Gr. Tat./51. Gelidium pseudointricatum Skottsb. et Levr. Original, leg. SKOTTSBERG. X 2. 2. Gelidium pseudointricatum var. altius Levy. Original, leg. SkoTTSBERG. Etwa nat. Gr. 3. Grateloupia subsimplex Levy. Original, leg. SkorrsBeRG. Etwa nat. Gr. 4. Cryptonemia prolifera Levr. Original, leg. SkOTTSBERG. “/10 nat. Gr. 5. Chondriella pusilla Levy. Original, leg. SkorrsBerG. Etwa nat. Gr. Taf. 52. 1. Nemastoma foliacea Levr. Original, leg. Scumirr. 7/10 nat. Gr, 2. Gymnogongrus furcellatus (C. Ag.) J. Ag. Original im Herb. AGArDH. $/10 nat. Gr. 3. Gymnogongrus furcellatus f. minor C. Ag. Hassiter-Exp. Nr. 71. $/10 nat. Gr. Gat 53- 1. Gloioderma subdichotomum Levy. Original, leg. SkorrsperG. Etwa nat. Gr. 2. Fernandosiphonia unilateralis Levr. Original, leg. SkovrsperG. Etwa nat. Gr. 3—4. Stromatocarpus microscopicus Levr. Original, leg. SkorrsBeRG: 3 miannliche, 4 tetra- sporentragende Pflanze. X 70. Printed 4/12 1041. Nat. Hist. Ffuan Fernandez and Easter Isl. Vol. I. PLATE 49. Nat. Hist. Fuan Fernandez and Easter Ist. Vol. 11. PLATE, 50: Nat. Hist. Fuan Fernandez and Easter Isl. Vol. I. PLATE 5 ce) _ Nat. Hist. Fuan Fernandez and Easter Isl. Vol. 11. PLATE 52 Nat. Hist. Ffuan Fernandes and Faster Ist. Vol. 1/1. PLATE &4 aM 9 3. Marine Algal Communities of the Juan Fernandez Is- lands, with remarks on the Composition of the Flora. By C. SKOTTSBERG, With Plate 54. During my two visits to Juan Fernandez observations were made on the mode of occurrence and composition of the various communities of marine algae. A publication had to be postponed until the species had been determined; for authors’ names, descriptions and figures, see Dr. T. LEVRING’s paper (this volume, no. 22). Unfortunately, the Corallinaceae of 1917, which were handed over to Mme PAvtL LEMOINE of Paris, could not be included. They contain several crust- aceous and fruticulose forms, but it is impossible to tell when their names will become available. The 1908 collection included, besides one or two Corallinae, two species of Lithophyllum, L. fernandezstanum Lem. and L. spec., possibly identical with J/elobesta pustulata Lam. in DICKIE’s paper (see LEMOINE in K. Svenska Vetenskapsakademiens Handl., Bd 61: 4, 1920). DICKIE listed Coradllina subulata Ell. and C. »Berteriti Mont.» (?= C. Berteroana Mont. ex HARVEY, Ner. austr. II, p. 103; Chile, BERTERO). I think that the same species are in my collection, but I cannot tell if DICKIE’s determinations are correct. Seaweed collecting along the shores of the Juan Fernandez Islands is no easy task. Most of the coast of Masatierra is constantly fringed by breakers, the south shore being exposed to very strong surf because the prevailing wind direction is SW—SE. Here landing is possible in Carbajal bay, at least in calm weather. The greater part of the shore is inaccessible from land. The only place where I was able to collect littoral algae was a small terrace with tidepools at the foot of Cerro Negro. The north coast is decidedly less difficult. There are several bays or coves which can be visited from land or sea, from west to east Bahia del Padre, Juanango, La Vaqueria, Puerto Ingles, Bahia Cumberland and Puerto Frances. I found the best collecting grounds in Padre bay, where the tidal region is wide and exhibits a very marked zonation, and in Puerto Ingles at the landing place where, however, the swell has to be watched closely. As a rule the coast is very steep, so that the littoral region 672 C. SKOTTSBERG becomes confined to a very narrow belt. The storm belt is indeed wide enough, but too boisterous for most non-crustaceous algae. A Durvillea would perhaps find suitable stations, but this as well as the other large Phaeophyceae, so con- spicuous along the rocky shores of south Chile, are absent. Another factor which helps to explain the poverty of the littoral flora is the nature of the rock. In the basal region of Masatierra it is to a great extent too easily disintegrated to resist the force of the breakers, and well developed communities were observed only where hard rock was exposed. MHere crustaceous Corallines will form a cover even in the stongest surf, and small fruticulose species become abundant. Masafuera is still more exposed than Masatierra and the littoral flora remarkably poor as far as my experience goes. The upper sublittoral region is, as a rule, inaccessible either from land or sea, and dredging offers considerable difficulties because of the steepness of the submarine shelf. Generally it is not possible to work in water less than 25 or 30 m deep without getting too close to the dangerous beach. The best results were obtained in the shallow strait between Masatierra and Santa Clara. Almost all my dredgings were made in water 20 m deep or more. Our knowledge of the region between low-water mark and 10 m_is very scanty. Whether any con- siderable number of species which do not belong to the lower littoral nor descend to the depths where my dredgings were made, are at home in the intermediate region, is impossible to tell, but hardly probable. Anyhow, we have every reason to believe that future investigations will reveal the existence of many algae which have escaped notice. No tidal observations have been made in these islands. We found the greatest difference between high and low tide to be about 4 ft. (= 1.2 m). The terminology is the same as used in my paper on Subantarctic algal communities (K. Svenska Vetenskapsakademiens Handlingar, ser. 3, vol. 19, 1941). Leading species, as observed directly in the littoral region, or occurring in abundance in a catch from the sublittoral, are in fat types. Microscopic en- dophytes are not included. ine littonalene econ St. 1. South coast of Masatierra, foot of Cerro Negro. 7.1.1917. Beach terrace of hard basalt, gently sloping inward. a. Deeper pools, fed by the surf through channels also in calm weather. Linteromorpha-Polysiphonia Ass. 1. Upper fringe. Enteromorpha intestinalis, (/va rigida. 2. Lower fringe. Ulva reticulata. — Chantransia catenulata (on Ulva), Poly- siphonia abscissa. b. Very shallow pools with tepid water, not washed. Cadlothrix vivipara Ass.” c. Upper sublittoral. Inaccessible. Cover of crustaceous Corallines. ‘ Dr. JoHS. Boye PETERSEN, Copenhagen, had the kindness to determine the Cya- nophyceae. MARINE ALGAL COMMUNITIES OF THE JUAN FERNANDEZ ISLANDS 6 ~I io) St. 36. Masatierra, Bahia Cumberland. 22.8. 1908. a. Upper littoral, dry during low tide. Boulders with coarse black sand between. On rock: Exteromorpha-Chondriella-Gelidium Ass. Chaetomorpha aérea, Enteromorpha intestinalis, Ulva rigida. — Compsonema ramulosum (epiph.), Ectocarpus confervoides \epiph.). — Chantransta catenulata and pacifica (epiph.), Chondriella pusilla, /rythrotrichia carnea (epiph.), Gelidium pseudointricatum, G. puszllum, Gigartina Lessoni, G. spec., Porphyra umbiticalis. On sand between the boulders, protected from the heavy surf, but as a rule gently washed: Lxteromorpha-Chaetomorpha antennina Ass. Chaetomorpha antennina, Enteromorpha intestinalis, (/va rigida. b. Lower littoral, washed by the surf at low water. Uva rigida Ass. Ulva rigida. — Lctocarpus granulosus, Ilea fascia. St. 38. Masatierra, Puerto Ingles, rocks at the small cove where landing is effected in good weather. 25.8. 1908. a. Upper littoral, laid dry for a couple of hours, but often sprayed. L7fe- romorpha-Chondriella-Gelidium Ass. Enteromorpha intestinalis. -—- Chondriella pusilla, Gelidium pseudoin- tricatum var. a/tus. b. Lower littoral, washed also at low tide. Glossophora Ass. Colpomentia sinuosa, Glossophora Kunthii.. — Lzthothammnion \Lithophyllum?) spec. St. 39. 26.8. 1908. St. 2—3. 15.1.+26.1.1917. Masatierra, Bahia del Padre. Wide rock beach with large boulders. Less exposed. a. Upper littoral (St. 39 a, St. 2a+b). Vertical extension 4—5 dm. 1. Upper belt. Dry at low tide and not washed in calm weather. Hydro- coleus Ass., making the rocks very slippery. Hydrocoleus lyngbyaceus. — Aii/denbrandia prototypus. 2. Middle belt. Dry during approximately 2 hours. xteromorpha-Chond- riella-Gelidium-Lithophyllum Ass. Enteromorpha intestinalis, (va /izza. — Chondriella pusilla, Gelidium pseudointricatum. — Lithophyllum fernandezianum. In pools: Codtum-Ectocarpus-Lithophyllum Ass. Codium fernandezianum, J/icrodictyon japonicum. — Ectocarpus Mitchellae, Sphacelaria cirrhosa (epiph.). — Corallina sp., Erythrotrichia carnea (epiph.), Gontotrichum alsidii (epiph.), Lithophyllum (Melobesia) sp. (epiph.), Polysiphonta abscissa. — Lithophyllum fernandeztanum. 3. Lower belt, transitional to b. Continuation of 2, with addition of bushy colonies of Scytothamnus and of some other species. The uppermost tufts of Splachnidium are found here. Chaetomorpha antennina. Enteromorpha intestinalis (on Scytothamnus), Ulva vigida. — Colpomenia sinuosa, Scytosiphon lomentaria, Scytothamnus australis. — Chondriella pusilla, Gelidium pseudointricatum, Gigartina spec. b. Lower littoral (St. 39 b, St. 3). Uncovered during 1—17/2 hours, but as a rule always washed by the surf; dry only when the sea is exceptionally smooth (see plate 54). 674 C. SKOTTSBERG Chaetomorpha firma-Splachnidium-Lithophyllum Ass. Chaetomorpha firma, Cladophora perpusilla (on Splachnidium), Codium cerebriforme, Ulva nematodea, U. rigida. — Colpomenia sinuosa, Ectocarpus arabicus (on Codium), E. hemisphaericus (on Splachnidium, etc.), /. Mitchellae (on Codium and Splachnidium), E. pusillus (on Chaetomorpha), Geminocarpus geminatus (on Corallina and Codium), Padina fernandeziana (large tufts), Scyzo- siphow lomentaria, Splachnidium rugosum. — Chantransia grandis and Thi retii (on Splachnidium), Corallina spec., Laurencia obtusa. — Lithophyllum fer- nandezianum. More or less confined to the pools and uppermost sublittoral: Codium- Lithothamnion (Lithophyllum?) Ass. Codium cerebriforme, C. fernandezianum, //alicysiis pyriformis (scarce, on the coralline crust), I/%crodictyon japonicum. — Ectocarpus arabicus (on Codium), E. cylindricus (on Microdictyon), :. Mitchellae (on Codium, etc.), Geminocarpus geminatus (on Codium), Glossophora Kunthii, Padina fernandeziana, S//ace- lavia cirrhosa (epiph.). — Centroceras clavulatum, Chantransia catenulata, grandis and Thuretii (epiph.), Corallina spec., Erythrotrichia carnea (epiph.), Gondotrichum cornu cervt and alstdit (epiph.), Polysiphonia abscissa. — Crusts of Lithothamnion (Lithophyllum?) spec. St. 40 a. Masafuera, landing place at entrance to Quebrada de las Casas. 28.8. 1908. Coast cliffs, exposed to the surf, as a rule washed or sprayed also at low tide. In crevices coarse black sand. Ulva rigida-Grateloupia Ass. Ulva rigida. — Geminocarpus geminatus (on Grateloupia), Scytothamnus australis, — Dendrymenia flabellifolia (young, very small), Gelrdella_ spec., Grateloupia subsimplex. The sublittoral region. St. 10. Masatierra, Bahia Cumberland. 3.1.1917. On a buoy in the har- bour, always remaining in the same relation to the water level. Uppermost sublittoral. a. Upper narrow zone at water’s edge. Scytosiphon Ass. Ectocarpus granulosus and Mitchellae (on Scyéosphon), Scytosiphon lo- mentaria. b. Lower zone. Ulva linza-Lepas anatifera (Ass.’) Ulva linza. — Colpomenia sinuosa (very small spms), Lctocarpus confervoides (on stalk of Lepas), E. granulosus and Mitchellae (on shell of Lepas), Lea fascia, Scytosiphon lomentaria (scarce). — Ceramium rubrum and Polysiphonia abscissa (both on small Spongzae). St. 8. Masatierra, in front of Pangal. 10.4.1917. 10—25 m. Sand with stones, supporting the algae. Dzctyotaceae-Codium-Rkhodomelaceae Ass. ™ The zoological collection from St. 10 includes Lefas fectinata Spengler (1 spm), Z. aza- tifera L. (several spms) and Balanus psittacus Darw. (see this work, vol. III, p. 483). MARINE ALGAL COMMUNITIES OF THE JUAN FERNANDEZ ISLANDS 675 Cladophora fascicularts, Codium fernandezianum, Microdiciyon japonicum. — Dictyota phlyctaenodes, Distromium Skottsbergi, Padina fernandeziana, Sf/ace- laria cirrhosa (epiph.). — | ot Codium fernandeztanum . . 4 Sard aes is ROUMSEPRONTA QUSCISS@ os. sw oe | — Eiypoglossum parvulum . . .. « «0.| + — : -- Pterosiphonia dendroidea. . .....\ + + — EESTI lige Pas ae ee pe gee ET ~ - Wirerodictyon japonicum ... » . » » »-|) e+ = - -- QUPOUNCTIIA SIMUOSE 2 2. eo is ws — a _ Chaetomorpha linum ......-.. - = Cladophora fascicularis .... 6...) + -- Fernandosiphonia unilateralis . . . 2) — = x ase WeGeOra,UTaChyCluda , -. i. sa me as — a3 = : 7 SOUT CAI | —— _ ate = = i » PLP UELOME See eee a =o = EL JETS ALLICLIUL. ecw = oe ch se et i - { == = IRMCELUFIL CUIFHOSA. . 5» = ss. «| + -- — { — = = CAV RLOMEINIL, PYOLWUCTA.. . . . » oi» & -- a= = ! 4 =a Heterosiphonia subsecundata .....) + = = a re = = ws es Hymenena decumbens ........-\ — = 4 = ae ae. sate as PAI AGITILULTIE DUCLIGUIIL = ee ee ve 2 -- _- — — — —- a = fimoaymenta cuneifolia.. ..... . «=: .| — = » GUSLG IES Oe en oA -— — - = = SAUER LS a ae a eee ee = = : - -- — — case in Juan Fernandez we do not know. The beach drift should originate from the lower littoral and the upper sublittoral which are more affected by the surf than the deeper water. Nothing forces us to conclude that any of the four species only found drifted, Codimm unilaterale, Laurencia clavata, Plocamium pacificum, and Pterosiphonia disticha, are confined to an intermediate zone, and nothing suggests that a considerable change in the composition of the flora takes place in this unknown region. The lower limit of algal growth is not known; at 30 m and beyond there are still numerous species. The picture given by the various dredgings is rather uniform, and with the material at hand only a single community can be distinguished, apart from a couple of local communities observed on a floating buoy. Within this com- munity two associations are distinguishable. 680 C. SKOTTSBERG Dictyotaceae-Codium-RKhodomelaceae Ass. On hard bottom, rock, stones and balls of Lzthothamnia, but also on shells, and coarse calcareous sand (shell fragments, etc.). One or more species of Dvzctyotaceae are always present, of Chlorophyceae Codium and often also Mecredictyon, and of RKhodophyceae several Rhodomelaceae besides many accessory species of other families, foliaceous or fruticulose. Most forms are rather small-sized; Padma and Codium fernande- stanum seem to be the largest. The composition is seen from Table II. Endo- phytes, microepiphytes and species known from a single locality have not been included. Lithothamnion (Lithophyllum) Ass. Where the surf is violent, a more or less pure cover of crustaceous Corallines can be seen to extend below low-water mark. One of the species is probably ZL. fernandeszanum. In deeper water the crust may form the bottom synusia of the Dectyotaceae-Codium-RKhodomelaceae Ass. Unfortunately, the species have not been determined. Scytosiphon lomentaria Ass. Observed once, colonizing a floating buoy of iron, and with a lower zone of Ulva linza Ass. (See above, St. 10.) Formations. In my paper on the algal communities of southern South America, etc. I tried to distinguish a number of physiognomic formations, represented in the cold and temperate waters of both hemispheres. The »Upper littoral formation of drought-resistant algae» includes in Juan Fernandez the //ydrocoleus(-Hilden- brandia) Ass. and the Lxteromorpha-Chondriella-Gelidium Ass. The »Middle- lower formation of surf resistant Rhodophyceae> seems to be very feebly de- veloped. No J/rzdaea, a genus of considerable importance along the opposite mainland of Chile, has been found, and species of Gzgartina seem to be rare and without influence on the physiognomy. The U/va-Grateloupia community mentioned under St. 40 a, but not sufficiently studied, occupies approximately the same position as an /ridaea Ass. The corresponding formation on less exposed coast is, in the Subantarctic region, developed as the complex community I called the Adenxocystis-Chloro- piyceae Ass. The dominant species are middle-sized to small brown and green algae. The equivalent to this seems to be the Chaetomorpha firma-Splachnidium Ass., which because of the greater variety of fairly large brown algae (Codfo- menia, Scytosiphon, Padina fernandesiana, and others) approaches the northern I‘ucus formation. The »Surf formation of large Phaeophyceae», with the Austral-Subantarctic Durvillea Ass., is wanting in Juan Fernandez, and the same is the case with other formations of large brown algae. Lessonza, Desmarestia and Macrocysits, leading genera of the mainland, are all absent. Pieces of Jacrocystis were collected off Bahia Cumberland by my zoological companion, Dr. K. BACKSTROM, which shows that drift from the American coast may reach the islands. Equally well developed along the coast of South America and around the islands is the »Formation of Crustaceous Corallines», where two associations can MARINE ALGAL COMMUNITIES OF THE JUAN FERNANDEZ ISLANDS 681 be distinguished, one forming a continuous cover on rock and the other en- crusting loose stones and shells and giving rise to calcareous sand. With regard to the »Formation of deep water Rhodophyceae? it is evident that in Juan Fernandez waters Dzctyofaceae are much more conspicuous than on the coast of Chile, but from the Peruvian coast several species are known. The presence of characteristic green algae such as Codium and Microdictyon makes the Juan Fernandez sublittoral formation even more mixed in appearance. Remarks on the geographical distribution and affinities of the Marine Flora. The Cyanophyceae and Corallinaceae have not been included; of the former, only 2 species were collected separately. The difference between Masa- tierra and Masafuera is, at least in part, due to deficient knowledge of Masa- fuera. Species of wide distribution, which are known from the west coast of South America, are marked also in the Peru-Chile column, etc. — Nomenclature and sequence of species are as in Dr. LEVRING’s paper. To the Subcosmopolitan group are referred widely distributed forms, mainly in Temperate regions, or at least not more or less confined to warmer seas. The list contains 93 species. To these should be added 2 species identified as to genus, and a few which could not be classified; further 2 species of Coral lina, Lithophyllum fernandezianum and L. (Melobesia?) sp., and one or several other species of crustaceous Corallines, bringing the total number to a little over 100 species. In a preliminary note on the marine flora (Nordiska Naturforskarmotet in Helsingfors, 1936) I pointed out the profound difference existing between the islands and the continent, so striking even to the casual observer, who will at once note the absence of the large brown algae, Durvillea, Lessonta, and Macro- cystts, of which the latter two run along the coast as far as Peru, MJacrocysizs reappearing on the coast of California. I remarked that the general likeness is perhaps greater with the flora of Peru (and northern Chile) than with the coast flora farther south, as the Subtropical element represented in the Peruvian flora is even stronger in Juan Fernandez, and that the insular endemism is remarkably high; further, that there is a small Australian—Neozeelandic element which does not belong to the circumpolar group of species known also from southern Chile and Magellania. Finally, that a couple of Subantarctic species, also found in Fuegia, the Falklands, etc., reach Juan Fernandez. My conclusion was that, taking into account the fact that the Humboldt current does not reach the is- lands, they are surrounded by warmer water, but that historical causes are partly responsible for the composition of the insular flora. ; Of the 93 species in Table III, 56 are confined to Masatierra, and only 8 to Masafuera; 29 were collected in both islands. Even if some difference exists between them, future investigations will very likely show that many species known at present only from Masatierra, occur also at Masafuera. No littoral 45—41438. The Nat. Hist. of Juan Fernandez and Easter Isl. Vol. II. 682 (G5 Masatierra +- ao +++ 4+ 4 ++ 4+ Masafuera Chlorophyceae. Codiolum Kuckuckti . . Lnteromorpha intestinalis WOE VELIEO. 5095 OC by isa ily. the » mematoidea . . DRELICLLAL ae Microdictyon japonicum Chaetomorpha linum. . » aerea . . » antennina » VEALLE Bee Cladophora fascicularis . » perpusilla . Ectochaete pacifica . . . » POUL “Be Entocladia viridis . . . Halicystis pyriformis . . Codium cerebriforme . Sernandezianum untlaterale fia . Phaeophyceae. Ectocarpus confervoides . » Mitchellae . » granulosus . arabicus .. . » pusillis ... » MUNULISSTNUS . » cylindricus . . » hemisphaericus Geminocarpus geminatus Compsonema ramulosum Myrionema strangulans Ralfsia expansa . SKOTTSBERG Table II. ac: eae: Pee aecea| oe eee a ea) Z z i} i Zi Remarks rei eal MAKE Sia Nee on 2" Nea ene a + As Splachnidium. 35 ar Se ot st = hoa ests + eka ert | + aa | West Pacific. ae ear T Bt ie ate + ot 1 Endemic, J. Fern. & S. Ambrosio. | Endemic. Endemic. Endemic. + AF a Endemic. | Endemic. | Endemic. +? * The typical species, Calif. + | seve) ES cain Pia | als | a Lr ayabe | + Atlantic only; north to | England. Endemic. a5 | sh Circump.-Subantarct. 3P 4s | + | + | Also Falkl. Is. + | Masatierra + + + +++ 4+ ++t+e4+t++ | + + -+- MARINE ALGAL COMMUNITIES OF THE JUAN FERNANDEZ ISLANDS 683 a2\alels8 {2 aHEACip 2 2 I ae ee Remarks Sa 2 ee aS = oe (oe ea Be I ee a Rc: = a & oe ee ee ee ee Sphacelaria cirrhosa . . 6 + +) + + Halopteris filtcina .....- +. ? . Distromium Skottsbergit .. . Endemic. Padina fernandeziama ... - Endemic, J. Fern. & S. Felix. Dictyota phlyctaenodes ... . Endemic. Glossophora Kunthiti.... . | = North to Galapagos ?comp. LEVRING, p.622). Papenfussiella Moseleyi. . . . Endemic. Splachnidium rugosum... . + S. Felix, S. Afr.—E. Indies. Colpomenia sinuosa ....- +)+)+)+ Hydroclathrus clathratus .. .« fe), dE || 22 || Se LSA TSS De OPS EE: BC. DO (© oe ae ee Scytosiphon lomentaria... +) + as || Se jek Scytothamnus australis... . +7) + “Chitoe- Rhodophyceae. Goniotrichum alsidit .....\| + + - > cornu cervi. ..-\ + — Erythrocladia irregularis. . . + | Europe. » subintegra . ..| + Erythrotrichia carnea ....| + Porphyra umbilicalis . . 2. ~| + | = Chantransia catenulata... . | + » SUL ELT ee oct -+- (Eur, Ath oN. “Amer. » ACTICO eee ot | + » jfernandeziana . .| Endemic. » WHUNMES eee ow 2 | Endemic. > collinsiama ... | | West Indies. Liagora brachyclada .... - Endemic. Gelidium pustllum..... - \ alae » pseudointricatum | ?3 3 Endemic in J. F-.? Grateloupia subsimplexr. .. - Endemic. Cryptonemia prolifera ....- Endemic. Nemastoma foliaceum .... Endemic. Plocamium pacificum..... + Gymnogongrus furcellatus . .« ae ea | Gigartina Lessonit.... + +) = _ North to Galapagos. 684 C. SKOTTSBERG (Cont. re S a | a S = | 8 oP a |e le 2/< = | @ | Ns Me; Remarks Z a, + | — | Chondriella pusilla. ..... | Endemic. + — | Gloioderma fruticulosum .. . =e — | + » subdichotomum . . Endemic. + | — Botryocladia fernandeziana . . Endemic. + | — | Rhodymenia cuneifolia ... . tay 4S. Chile; Subantarct. + | + » QUST USO + — | + | Dendrymenia flabellifolia . . . aE North to Galapagos. + — | Antithamnion minutissinum . | | Endemic. — | + | Pleonosporium globuliferum . . | Endemic. a8 — | Ceramtum rubvum...... | _ > a =F = [Rees Centroceras Clavulatum.... a5 F Sr + | + | Hypoglossum parvulum ... | | Endemic. 1 | = | PAycodrys GuUEercijouia <2). 7. 435'/[ 234 5 Circump.-subantarct. til, Samy ll MOGHETOSC HES page SZ7zohe wea e ys ce Spel) ae |° Probably north to Val- paraiso." + | + | Hymenena decumbens . | Endemic. + | + | Heterosiphonia subsecundata . | st + | + | Polysiphonia abstissa.... . SE |. ak Circump.-subantarct. + | + | Pterosiphonia pusilla. ....| | Endemic. +} — » Shottsbergiit. . . | | Endemic. =>) Les } WMGAUE oo oe Endemic. + | + » dendroidea ... + a= + | — | Stromatocarpus microscopicus . | Endemic. + | + | Fernandosiphonia unilateralis , | Endemic. + | + | Dipterosiphonia parva .... | Endemic. ap || Se |) LL GUS. 8G ae | ar | ate | ==. 1 sete » AEST De Pea | | ai | Malaya + | — | Hildenbrandia prototypus. . . | =e Ge! | |7 Fuegia. algae have been collected in Santa Clara, which has been visited by very few naturalists. My experience is that in this case collecting is, as a rule, almost impossible. We have no reason to expect anything that does not occur also around the nearby western end of Masatierra. Only 30 of the species found at Juan Fernandez are known from the coast of Peru and Chile; in addition, two, Schzzoseris Davisti and Hildenbrandia proto- * See KYLIN, Die Delesseriaceen Neu-Seelands, K. Fysiogr. Sallsk. Handl. N. F. 40; 2, p. Io. MARINE ALGAL COMMUNITIES OF THE JUAN FERNANDEZ ISLANDS 685 typus, occur in the Magellanian region, and if the Peruvian record for //alopteris filicina is correct, and Geltdium pseudointricatum is MONTAGNE's iutricatum, the total number is 34, or 367/2 %; thus, the independency of the insular flora in its relations to the continental is well brought out. No less than 20 of the 34 species are widely distributed, several of them subcosmopolitan. LEVRING gave a brief synopsis of the geographical elements (p. 602—603), and under each species carefully indicated its distribution, enabling us to make the following subdivision as a basis for a discussion. Group 1. Subcosmopolitan, or of wide distribution in temperate regions. — 20 sp. = 21'/2 %. Enteromorpha intestinalis, Ulva rigida, WU. linza, Chaetomorpha linum, Ch. aerea, Entocladia viridis, Ectocarpus confervoides, E. granulosus, Myrionema strangulans, Spha- celaria cirrhosa, Llea fascia, Scytosiphon lomentaria, Goniotrichum alsidii, G. cornu cervi, Erythrocladia irregularis, E. subintegra, Erythrotrichia carnea, Porphyra umbilicalis, Ceramium rubrum, fildenbrandia prototypus. Group 2. Subtropical, + widely distributed in warmer seas. — 13 sp. = 14 %. a. Found also on the mainland coast. — 7 sp. = 7'/2 %. Chaetomorpha antennina, Cladophora fascicularis, Halopteris filicina(?), Colpomenia sinuosa, Hlydroclathrus clathratus, Gelidium pusillum, Chaetoceras clavulatum. 6. Not found on the mainland coast. — 6 sp. = 67/2 %. Ulva reticulata, ctocarpus Mitchellae, EF. arabicus, E. pusillus, Ralfsia expansa, Laurencia obtusa. Group 3. Peruvian-Chilenian. — 7 sp.=7'%/2 %. Ulva nematoidea, Chantransia catenulata, Gigartina Lessonit, Rhodymenia cunei- folia, Dendrymenia flabellata, Heterosiphonia subsecundata, Fterosiphonia dendroidea (also Calif.). Group 4. Californian. — 7 sp. = 77/2 %. Codium unilaterale, Ectocarpus cylindricus, E.. hemisphaericus, Compsonema remosis- simum, Chantransia pacifica, Ch. Thuretii*, Plocamium pacificum. Group 5. West Pacific, including Australia, New Zealand, Oceania, Malaya, Japan. — 9 sp.=—9Q'/2 %. a. Extending to the west coast of South America. — 3 sp. Glossophora Kunthii, Scytothamnus australis, Gymnogongrus furcellatus. 6. Not found on the west coast of South America. — 4 sp. Microdictyon japonicum, Gloioderma fruticulosum, Rhodymenia australis, Laurencia clavata. c. Austral-circumpolar, not reaching the coast of South America. — 2 sp. Codiolum Kuckuckii, Splachnidium rugosum. * Atlantic, and Californian. Should perhaps find a better place in group I. 686 C. SKOTTSBERG Group 6. West Indian. — 1 sp. Chantransia Collinsiana.* Group 7. Subantarctic (Antarctic)-circumpolar. —- 4 sp. = 4 %. Geminocarpus geminatus, Phycodrys quercfolia, Schizoseris Davisi, FPolysiphonia abscissa (north to Peru). Group 8. Endemic. — 32 sp. = 3472 %. To utilize fully the endemic species for a discussion on the history of the flora, we must know their systematic position within the genera to which they belong. As it is, little is known about the genetic relation, in most cases at least. Therefore we have to confine ourselves to a subdivision according the general distribution of the genera. a. Widely distributed, in cases subcosmopolitan genera, not exclusively or mainly found in warmer seas. — 7 sp. Cladophora perpusilla, Ectochaete pacifica, E. ramosa, Ectocarpus minutissimus, Chantransia fernandeziana, Ch. grandis, Antithamnion minutissimum. 6. Genera with a + wide distribution in warmer seas. — 22 sp. Chaetomorpha firma, Codium cerebriforme, C. fernandezianum, Dictyota phlyctaenodes, Padina fernandeziana, Liagora brachyclada, Gelidium pseudointricatum (?), Grateloupia subsimplex, Cryptonemia prolifera, Nemastoma foliaceum, Botryocladia fernandeziana, Pleonosporium globuliferum, Pterosiphonia pusilla, P. Skottsbergu, P. disticha, Stromato- carpus microscopicus, Dipterosiphonia parva. Yo this group I attach Halicystes pyri- formis (3 Atlantic species, one of these also on the Pacific coast of N. Amer.), Dzsi/ro- mium Skottsbergii (a second species from Japan), Papenfussiella Moseleyi (most species South African), and the two endemic genera Chondriella (1 sp.) and Fernandosiphonia (Gasp. c. Genera with their main centre in Australian waters. — 3 sp. Gloioderma subdichotomum, Hypoglossum parvulum, Hymenena decumbens. We can disregard groups 1 and 8 a, which, with our present knowledge, cannot tell us anything about their history. Taking together groups 3 and 4 we get a West American group of 14 species(= 15 %), a remarkably small number. Even if future researches will add to this group we have little reason to believe that the proportion between this element and the non-American will undergo much change. Group 2, 13 sp.(= 14 %) is of a more or less distinctly subtropical char- acter, and the same may be said of group 8 b, 22 sp. (237/2 %), and of group 5, although Splachnidium with its obligate parasite reaches the cool waters of New Zealand. I think we can bring together all three to constitute a Subtropical element, 54 species, or 58 %, which may be said to give the Juan Fernandez flora its special characteristic features. We have seen that a Subtropical element * Gelidiella sp. is perhaps identical with a West Indian species (LEVRING p. 636). We cannot draw any conclusions from our present knowledge of the distribution of microepiphytes like L£ctocarpus pusillus or species of Chantransia, etc., which are likely to have been over- looked especially in regions where little collecting has been done. MARINE ALGAL COMMUNITIES OF THE JUAN FERNANDEZ ISLANDS 687 is found also on the shores of the continent, where 10 of the 54 species occur, and in the Peruvian flora are several non-insular species of the same character, such as Chaetomorpha cartilaginea, Caulerpa flagelliformis, Sargassum pacificum, Spatoglossum crispatum, Neurocarpus Cokeri, Gelidium spp., and perhaps some additional Rhodophyceae (HOWE, The Marine Algae of Peru, Mem. Torr. Bot. Club XV, 1914). Just as group 5c, group 7 is Circumpolar, but of a much more southerly distribution, and with the exception of Sc/zsoser7s, the species have been recorded from the Antarctic region proper. Whereas Polysiphonia abscissa ranges north to Peru, the other three appear to be confined to Magellanian waters: our know- ledge of the marine flora of Chile is, however, far from complete. Before entering on a discussion of the history of the marine flora of Juan Fernandez, it is necessary to pay attention to the fauna. It appears from Dr. L. PLATE’s brief summary (Zur Kenntnis der Insel Juan Fernandez; Verh. d. Gesellsch. fiir Erdkunde Berlin, XXIII, 1896) that — as we had every reason to expect — the parallelism in distribution is obvious. PLATE found that the littoral fauna of Masatierra — he did not visit Masa- fuera — differs not only from the fauna of Valparaiso bay, but also from the entire fauna along the stretch Arica—Cape Horn, a fauna of unmistakable uni- formity. As an example PLATE mentions the Echinoderms. The islands are poor in species as compared with the coast, but most of them are endemic. The continental forms occurring at the islands are either North Chilean or Peru- vian, or have a wider range along the coast down to the latitude of Valparaiso or even farther south. Examples taken from various other groups are also quoted; the Mollusca are said to be »sehr lehrreich>. PLATE continues: » Mir ist kein einziges Beispiel bekannt, dass eine fiir die Breite von Chiloé oder die Magalhaens-Strasse charakteristische Art auch auf Juan Fernandez vorkame, und soweit ich die Frage bis jetzt zu tibersehen vermag, zwingen die Tatsachen zu der Annahme, dass die Kiistenfauna der Insel von Peru oder Nord-Chile her eingewandert ist» (p. 224), and finally: »Aus dem Gesagten ergiebt sich dem- nach, dass die Fauna der Gezeiten- und Seichtwasserzone von Juan Fernandez erstens durch ihren Reichtum an endemischen Formen und zweitens dadurch angezeichnet ist, dass, wo Ubereinstimmungen mit der kontinentalen Fauna sich nachweisen lassen, diese auf eine Einwanderung von Nord-Chile oder Peru her hindeuten» (p. 226). There are, however, species which have other relations. Subantarctic forms, not ranging north to North Chile or Peru, are known from Juan Fernandez, also among the animals. Of non-endemic species, some have a quite different type of distribution. Two cases are mentioned by PLATE, the famous »langosta>, Palinurus frontalis (now called Fasus Lalandet), and Littorina mauritiana. Neither of these occur on the coast of America. Fasus is Austral- circumpolar and has very much the same distribution as Splachnidium (my group 5 Cc); just as this, it reaches the Desventuradas Islands. Lzé¢or7va is Indo- pacific (Mauritius, Australia, Oceania, E. Asia) and corresponds to my group 5 b. Consequently we have, also in the Fauna, a non-American element for which a 688 C. SKOTTSBERG South American origin is improbable, but PLATE seems to have paid too little attention to these exceptions from the rule. From the papers in the zoological volume of this work (vol. III) some in- formation on these questions may be had. W. FISCHER (p. 45) reports the first Gephyroid from Juan Fernandez, an endemic species of the genus Aspzdosiphon. It shows some relation to a species from Panama. There are many Sipunculidae known from Central America and the northwest coast of South America, but not from Chile. The Juan Fernan- dez species should be referred to a Subtropical element, represented also in America. RENDAHL's report on the Fishes (p. 49) lists 40 species stated to occur at the islands, but a major part of our collections was lost. The endemic element includes 20 species, of which one represents an endemic genus. The others belong to Subtropical or Tropical genera, some of which occur also in the At- lantic, and correspond to my group 8b; some are found also on the coast of Chile and Peru. Six genera have their home in Australian waters, including New Zealand, and of these 3 do not reach the continent; 1 is Indo-Pacific. I would include these under 8c. Of the non-endemic species, 16 are continental, mostly of a Subtropical character, and all except 3 found on the Chilean coast; of these, I is known from the Galapagos Islands, and 2 from Peru. Among these 16 fishes are representatives of groups 2, 3 (probably also 4), 5 a and 5c. Of the remaining 4, 2 belong to Australia and New Zealand (group 5 b), 1 is African-Subantarctic, and 1 from the Tropical Atlantic. Of the 15 species known also from the coast of Chile or Peru, no less than 9 have not been found elsewhere and constitute a stronger American element than we find among the algae (22"/2 % against 15 %). Among the Juan Fernandez fishes, there seem to be none corresponding to groups ft or 7, Subcosmopolitan or Subantarctic- antarctic; otherwise, the parallelism with the marine algae seems complete. Of marine Amphipoda our collection contains only 4 species (CHILTON, p. 81); 2 are »widely distributed in all seas», 1 was known from Subantarctic America and the coast of Chile and observed only on drifting Macrocysizs, thus not a regular member of the insular fauna, the fourth was identified with a species known before from Brazil. Marcus’ paper on the &ryozoa (p. 93) is of interest from our viewpoint, though our present knowledge of the Chilean and Peruvian faunas is very limited. Among the 19 species recorded from Juan Fernandez 2 are endemic; on the other hand, 8 are Cosmopolitan or have, at least, a wide range in the Atlantic and Pacific, all of them reaching high southern latitudes; they also belong to the Magellanian fauna. Two species so far only found in the Atlantic, including the Arctic Sea and Magellanian localities, may be added here, 2 are Southwest Pacific (Tasm., N. Zeal., etc.), 4 are Subantarctic-circumpolar, and 1 Indo-Pacific. We have no special West American element, but otherwise a parallelism with the distribution of the Algae seems pretty distinct. There are numerous species common to Juan Fernandez and the Magellanian region, though most of these have a wide range, and we have some more strictly Subantarctic forms, just MARINE ALGAL COMMUNITIES OF ‘JHE JUAN FERNANDEZ ISLANDS 689 as among the Algae. MARCUS states that PLATE’s conclusions are not supported by the Aryozea as regards the difference between the islands and the continent, and he emphasizes that there is a »weitgehende Ubereinstimmung» in the com- position of the American Subantarctic and the Juan Fernandez faunas. Only 1 species of the A/cyonarza has been described (MOLANDER, p. 129); it belongs to a genus known before from Malaisia, the Philippine Is., Australia and Oceania, but not from America, a distribution as in group 5 b. Of the 3 Ascidia (HARTMEYER, p. 131), I (endemic) species belongs to a Circum-austral genus, I is Subantarctic-Chilean (gr. 7), 1 Californian (gr. 4). The 2 Pycnogo- nidia (LOMAN, p. 137) are endemic; one belongs to a monotypical genus, the other to a genus well represented on the Pacific coast of North America, but not extending to South America. Three », namentlich unter den von mir aus Bolivien mitgebrachten Pflanzen, bei ¥. co/o- vata als eigene Formen unterzubringen. Das Blattzellnetz wird zur Erkennung der Art eher als die Blattform und der Habitus dienen konnen. Meist ist man ja auf die Blattmerkmale angewiesen, da die Pflanze recht selten Perianthien tragt. Ist dies aber der Fall, so bieten wohl die Involucralblatter das wert- vollste Merkmal. Am schwankendsten ist der Habitus und die Farbung, die vom blassen Griinlichbraun uber Braun bis ins Karmin und tiefste Schwarzrot gehen kann und stark von den Lichtverhaltnissen des Standortes abhangig sein diirfte. DIE FOLIOSEN LEBERMOOSE DER JUAN FERNANDEZ-INSELN UND DER OSTERINSEL 7O!I Geographische Verbreitung: Falklandsinseln, Andenkette von Feuer- land bis Mittelamerika (Costa Rica), Juan Fernandez, Neuseeland, Tasmanien, Ostaustralien und subantarktische Inselwelt, Siidafrika. *3, J. grandiflora (L. et G.) Spruce. — Syn. Fungermannia grandiflora eect G. Syu. Hep. p. 673. Auf Erde und Steinen. Masatierra: Am Pfad nach Portezuelo, Wald (no. 199 pp.). Auffallenderweise ist diese Art von Juan Fernandez nur sehr sparlich mit- gebracht worden, wahrend sie im Kustengebiet Westpatagoniens ganz allgemein verbreitet und haufig zu sein scheint. Geographische Verbreitung: Feuerland, Westpatagonien, Siidchile, Masatierra. J. spec. Masafuera: Las Torres, an Felsen, 1370 m (no. 257 pp.). Die Sparlichkeit und Sterilitat des Materials gestattet keine nahere Bestim- mung, doch scheinen die vorliegenden Proben wenigstens von den beiden vor- hergehenden Arten verschieden zu sein. Anastrophyllum Spruce. *4. A. leucocephalum (Tayl.) Spr. — Syn. $ungermannia leucocephala avis Journ. of Bot., 1846; p.. 272. An Felsen. Masafuera: Las Torres, an feuchten Felsen, 1330 m (no. 193 pp.). Trotz gewisser Abweichungen diurften die vorliegenden Pflanzen, deren Blatter nur bis zur Halfte eingeschnitten sind, zu dieser Art geh6ren. Die Form und Art der Zuspitzung der Lappen und der lange Dorn am Vorderrand_ spre- chen dafiir. Wahrscheinlich handelt es sich bei den Juan Fernandez-Pflanzen um etwas abweichende Hungerformen, wie es deren von einer ganzen Reihe von Arten zu geben scheint. Wenn man diese Parallelitat des Auftretens nicht wiederholt beobachtet hat, kann man leicht dazu verleitet werden, sie als eigene Arten zu beschreiben. Ich kenne solche Formen, die bei manchen sehr charak- teristischen Typen als getreue Miniaturausgaben der normal entwickelten Pflanzen erkannt werden kénnen, von folgenden Arten: Lophocolea carinato-bifida, L. striatella, L. attenuata, Adelanthus sphalerus, Famesoniella colorata, Lepicolea ochroleuca und Lepidolaena magellanica (s. Bemerkung bei Adelanthus spha- lerus!). Geographische Verbreitung: Feuerland, Peru, Venezuela, Masafuera. Vielleicht gehért auch A. cuspidatum St. aus Bolivia hierher. Die »spinula hamata» am basalen Vorderrand des Blattes entspricht hier der »lacinia magna» bei A. leucocephalum. Die tiefe Bucht zwischen den Lappen, deren feine Zu- spitzung und das Blattzellnetz wiirde sehr gut zu dA. /eucocephalum passen. 702 TH. HERZOG Acrobolbus Nees. #5. A. excisus (Mitt.) Schiffn. — Syn. Marsupidium excisum Mitt. 1877; Gymnanthe crystallina Mass. 1885; Marsupidium crystallinum Mass. 1889; 7yl- manthus crystallinus St. 1900. Auf Heideboden und an Felsen, fast immer andern Lebermoosen beige- mengt; nur in Hohenlagen. Masafuera: Feuchte Heide bei Correspondencia, ca 1,350 m (no. 91); Las Torres, an Felsen, 1,370 m (no. 93 pp., 193 pp., 257 pp.). Obwohl die Pflanze auf Juan Fernandez bisher nur steril beobachtet wurde, ist sie doch leicht an ihrer dichten Bekleidung mit grossen, glasperlenahnlichen Papillen zu erkennen. Sie stimmt darin weitgehend mit Lophozia antarctica Angstr. iiberein, die sich aber auch steril sofort durch ihre grossen Amphi- gastrien unterscheidet. Schon MASSALONGO (12) hat die Pflanze als dem (ihm wohl nicht zuganglichen) Acrobolbus excisus (Mitt.) nahestehend erkannt. Geographische Verbreitung: Kerguelen, Feuerland, Magellanstrasse, Westpatagonien (Gusinde), Masafuera. Anastrepta (Lindb.) Schiffn. *6, A. bifida St. — Syn. Plagiochila bifida St. 1886; Anastrophylum 6r- fidum St. 1900. Auf Heideboden und an Felsen, nur in hoheren Lagen. Masatierra: Am Pfad nach Portezuelo, an trockenen Tuffwanden im Wald (no. 107, 199). Masafuera: Heide nahe dem Correspondencia-Camp, 1,100 m (no. 122, 124 pp.); im Bett eines kleinen Flusses auf dem Hochplateau, ca 1,100 (no. 121 pp.); Hochplateau, auf der Heide zwischen Gras, ca 1,100 m (no. 338 pp.). Diese merkwiirdige kleine Pflanze nimmt fast eine Mittelstellung zwischen Anastrepta und Plagitochila ein. Kein Wunder, dass STEPHANI sie nach einander in verschiedenen Gattungen unterbrachte! Die Blattstellung mit ihrem an der Basis zuruckgeschlagenen Vorderrand lasst im sterilen Zustand leicht eine Ver- wechslung mit Plagztochila zu. Das Perianth lasst aber sofort den Irrtum er- kennen. Nach ihm konnte man zunachst an Axastrophyllum denken. In der Synonymik findet diese Merkmalskombination denn auch ihren entsprechenden Ausdruck. In »Species Hepaticarum» II, S. 191 schildert STEPHANI ausfihrlich, wie er erst allmahlich durch vergleichende Beobachtungen an Axastrepta orca- densis seinem Irrtum auf die Spur gekommen ist. Geographische Verbreitung: Magellanstrasse, Siidchile bei Valdivia (Hosseus), Juan Fernandez. DIE FOLIOSEN LEBERMOOSE DER JUAN FERNANDEZ-INSELN UND DER OSTERINSEL ~I 1) Os Plagiochila Dum. *7, P. Gayana G. Auf Erde, Steinen und Baumwurzeln. Masatierra: Puerto Frances, auf Erde und Steinen im Wald 500 m (no. 66); héchster Punkt im Frances-Tal, am Grund eines Baumstammes, 500 m (no. 69); Nordhang des Cerro Piramide, 600 m (no. 385 pp.). Die Auffindung dieser tiberaus eleganten, zierlichen Art auf Juan Fernandez war deswegen von besondrer Bedeutung, weil bis heute iiber sie eine bedauer- liche Unklarheit herrschte. In den meisten Herbarien befinden sich von ihr nur Falschbestimmungen, d. h. es werden unter ihrem Namen andre Arten auf- bewahrt. Am ehesten kann man zu ihr noch Pflanzen aus dem Herbar Warn- storf (»Fretum magellanicum») rechnen, die allerdings als forma subintegra zu unterscheiden waren. In »Species Hepaticarum» spricht sich STEPHANI dar- uber schon aus. In Ermangelung des nicht auffindbaren Originals musste er sich an GOTTSCHES Diagnose und Zeichnung halten. Er bemerkt dazu, die Pflanze sei bisher nicht wieder gefunden worden und die Exemplare in DUSEN’s Sammlungen gehoérten nicht hierher. Davon konnte ich mich an einem mir freundlichst aus dem Berliner Botanischen Museum iiberlassenen Beleg (»ad lacum Llanquihue in terra») tberzeugen. Umso wertvoller war die Auffindung der Art an mehreren Stellen auf Masatierra. Die vorliegenden, wohlentwickel- ten und mit den charakteristischen Perianthien versehenen Exemplare stimmen ausgezeichnet mit der von STEPHANI in Anlehnung an GOTTSCHE gegebenen Beschreibung und Abbildung itberein, so dass uber die Identitat kein Zweifel bestehen kann. Vielleicht stammt das Original von der benachbarten Kite. Das Zitat »Chile (Gay)» lasst keine genauere Bestimmung zu. Geographische Verbreitung: Siidchile, Masatierra. 8. P. fasciata St. Auf Erde und Steinen. Masatierra: Am Pfad nach Portezuelo, auf feuchter, lehmiger Erde, ca 450 m (no. 24); auf einem schmalen Vorsprung mit Baumen an senkrechter Felsklippe, Portezuelo, ca 600 m (no. 84, 389 pp.); Nordhang des Cerro Piramide, 600 m (no. 87); Riicken zwischen den Talern von Piedra Agujeriada und Laura, 600 m (no. 366 pp.). Masafuera: Heide bei Correspondencia, an Steinen, ca 1,350 m (no. 25). Auch diese hiibsche, sehr schlanke Art gehért zu der Gruppe der Klein- zelligen, unter denen H. Cart (2) die 3 Sektionen Flexicaules, Equitantes und Angulatae unterschieden hat. Er bringt hier P. fasctata, obwohl er keine Herbar- belege gesehen hat, wohl mit Recht in der Verwandtschaft von ?. /lexicaulis unter, bei welcher die schrage Stutzung der Blattspitze mit ihren gewohnlich 2 grossen Endzahnen charakteristisch erscheint. Beziehungen bestehen aber auch augenscheinlich zu P. Gayana G. und P. subquinguespina Herz. Geographische Verbreitung: Chiloé und Juan Fernandez. 704 TH. HERZOG *9, P. Hyadesiana B. et M. Miss. sc. 1889, p. 214. Wohl vorwiegend an Felsen und Steinen. Masatierra: Portezuelo de Villagra, in Rissen senkrechter Felsklippen, 590 m (no. 31). Masafuera: Innerster zuganglicher Teil des Casas-Cafion, ca 200 m (no. 38); Bett eines kleinen Flusses beim Correspondencia-Camp, ca 1,100 m (no. 39); Cordon del Barril, 650 m (no. 43). In ihrer gut entwickelten Form ist P. //yadestana eine stark verzweigte, verhaltnismassig schlanke, aber doch etwas starre, kleinblattrige Pflanze von ge- wohnlich_ schwarzlich-griiner Farbung. Besonders charakteristisch erscheint mir die schmal keilformig zulaufende Blattbasis, das engmaschige Zellnetz und die unregelmassigen, groben aber sparlichen Randzahne, die sich auf den oberen Blattteil beschranken. Das Juan Fernandezmaterial enthalt in no. 31 und 43 schmachtige, fast unverzweigte, wohl jugendlich unentwickelte Formen. Geographische Verbreitung: Feuerland, Sudchile, Juan Fernandez. *10. P. deformifolia St. — Syn. P. Schwabet Herz. Beih. Bot. Centrbl. Xe Osos Abts). An Felsen. Masafuera: Schlucht beim Correspondencia-Camp, 1,150 m (no. 37). Die vorliegenden Exemplare stimmen mit dem Typus (leg. SKOTTSBERG, Exped. suec. 1907—09, no. 279, Patagonia occid., Canal Messier, Port Hale) bestens uberein, sind aber auch identisch mit meiner P. Schwadbez. P. deformt- folia ist aber von STEPHANI unrichtig beschrieben und abgebildet worden, so dass ich die SCHWABE’sche Pflanze wohl verkennen musste. Erstens sind die Perianthe nicht, wie STEPHANI sie beschreibt, »ore truncato, repando», sondern feinstachelig gezahnt. Zweitens sind die oberen Blattzellen nicht 18 X18 » gross, sondern 12X14 — (hdchstens) 16 , woraus sich ein ganz andres Zellnetzbild ergibt. Bei meiner P. Schewadbez sind die oberen Blattzellen 10—12 » im Durch- messer, manche bis 14 oder 12X16 », also ganz ahnlich, praktisch wberein- stimmend. Stellung, Grdésse und Randgliederung der Blatter sind ganz die gleichen. Auch die Perianthmiindung stimmt in ihrer feinen Zahnelung ganz . uberein. Geographische Verbreitung: Westpatagonien, Juan Fernandez. D> oD ten) ) #11, P. chiloénsis St. Sv. Vet.-Ak. Handl. 1900, vol. 26, Bihang, p. 27. An Rinde und Felsen. Masafuera: Casas-Cafion, feuchte Felsen nahe dem Bach, ca 290 m (no. 45 pp.). Diese auf dem _ siidchilenischen Festland ziemlich haufige Art gehort auch zu der Sippe der kleinzelligen Plagiochilen ohne Eckverdickungen, zeichnet sich jedoch durch ihre langlich-eiférmigen Blatter mit oft gestutzter Spitze und reich- lich scharf gezahntem Hinterrand aus. Die Blattzahne sind stets 2—3-zellig, DIE FOLIOSEN LEBERMOOSE DER JUAN FERNANDEZ-INSELN UND DER OSTERINSEL 7°5 lang und schmal, fast dornig, aber aus ganz kurzen Zellen aufgebaut. Die vor- liegenden Exemplare sind nicht ganz typisch. Geographische Verbreitung: Siidchile, Chiloé, Juan Fernandez. #172. P. rectangulata St. 1. c. Auf Erde und Rinde, nur in Hochlagen. Masafuera: Bett eines kleinen Flusses beim Correspondencia-Camp, 1,100 m (no. 34); Hochland nérdlich des Correspondencia-Camps, ca 1,100 m (no. 35); Osthang des Innocentesberges, im Dzcksonta-Wald an Drimys, ca 1,000 m (no. 46). Die vorliegenden Exemplare stimmen zwar in ihrer charakteristisch breiten Blattform mit sehr schmaler Basis und grober, sparlicher Randzahnung mit dem Typus gut wtberein, weichen jedoch durch etwas engeres Blattzellnetz einiger- massen ab. Angesichts des sehr plastischen Verhaltens vieler Plagiochilen méchte ich doch glauben, dass sie sich bei P. rectangulata am ehesten einreihen lassen. Geographische Verbreitung: Westpatagonien, Juan Fernandez. *13. P. remotidens St. — Syn. P. amdbigua de Not. Acad. Torino, 185 vol. 16, pag. 213 (non Ldbg. et Hpe.). Auf Heideboden der Hohenlagen. Masafuera: Feuchte Heide oberhalb Correspondencia, ca 1,350 m (no. 28, QI pp.). Die Juan Fernandez-Pflanzen stimmen mit den CUNNINGHAM’schen Exem- plaren (Herb. Uppsala) gut iiberein, wahrend die DUSEN’schen Belege (no. 313) sich durch die beiderseits stark herablaufenden Blattrander der P. eguttans G. nahern. Fiir beide Arten sind die im oberen Blatteil sehr kleinen, derben Zellen ohne Eckverdickungen charakteristisch. STEPHANI bringt sie wegen ihrer etwas abweichenden Blattformen in zwei verschiedene Gruppen, doch sind Uber- gange deutlich zu beobachten. Geographische Verbreitung: Magellanstrasse, Sidchile, Juan Fer- nandez. 14. P. fusco-brunnea St. (Fig. 1, a—h). Hauptsachlich an Steinen und Felsen, selten an Baumrinde. Masatierra: Quebrada Damajuana, steiler Waldabhang, 345 m (no. 76); Felsgrat auf der Vorderseite des C. Pirdmide, 600 m (no. 68, fo. subintegra Herz.); Portezuelo de Villagra, steile Felswand, 600 m (no. 85). Masafuera: Quebrada de las Casas, Felswand (no. 56); Quebrada del Mono, enge Schlucht, an Dzcksonia, 570 m (no. 73); Quebrada del Blindado, feuchte Felsen im Wald, 440 m (no. 70) und an einem Décksonia-Stamm, 440 m (no. 62, fo. subintegra), Quebrada de la Loberia, Waldboden und Steine, 280 m (no. 77); Hochplateau, Lophosoria-Heide, ca 1,100 m (no. 79); Heide oberhalb Corres- pondencia, feuchter Boden, ca 1,400 m (no. 60, 63); Hochland noérdlich vom 706 TH. HERZOG Fig. 1. Blatter von Plagtochila fuscobrunnea St. (a—h), P. chilensts St. (i, k) und P. rubescens Ldbg. ()}: a= no. 6, b= no. 86, ¢ = no. 56, d= no, 63, e = no. 73, f= no. 53) = — meme h = no. 6 pp., i (leg. CLAUDE JOSEPH), k (leg. HOSSEUuS), | (Typus von P. 7abescens), alle 10/1. Correspondencia-Camp, Farnheide, ca 1,200 m (no. 86, fo. major); Bett eines kleinen Flusses beim Correspondencia-Camp, ca 1,100 m (no. 81, 335 pp.). In den von STEPHANI (20) beschriebenen und abgebildeten Exemplaren scheint zwar eine gute Art vorzuliegen. Dieser Eindruck geht jedoch, wenn man so reichliches Material untersuchen kann, wie es hier in der SKOTTSBERG- schen Sammiung vorliegt, immer mehr verloren. Man kann eine beinahe licken- lose Reihe zu »P. chilenszs St.» (Fig. 1 i, k)’ hin verfolgen, so dass man fast geneigt ware, die letzten Ubergangsformen zu P. chzlensis selbst zu rechnen (Stengel einer solchen Pflanze, zwischen Schzstochila Berteroana no. 6 einge- sprengt, kénnte man z. B. fast ebenso gut bei P. chz/enszs unterbringen!). Der Hauptunterschied zwischen beiden liegt in den im allgemeinen kirzeren und breiteren Blattern und deren sparlicheren Randzahnung, bei den extremen Formen von P. fuscobrunnea in der fast vollig fehlenden Zahnung (fo. subzntegra). Immer aber sind die Blattzahne unregelmassiger und grober als bei »P. chzlen- s7s> und namentlich an den Blattohren, die auch keine so deutliche Crista for- men, sehr zuriickgebildet. Das andre Extrem dagegen nahert sich sowohl in der schmaleren Blattform, wie auch der basalen Zahnung des Ventralrandes der P. »chilensis» schon ganz betrachtlich. Im ubrigen kommen bei P. fuscobrunnea sowohl breit- wie schmalblattrige Formen vor. Merkwiurdigerweise sind bei den letzteren, die in der Form mehr Anklange an P. »chilenszs» zeigen, die Rand- zahne oft besonders stark zuriickgebildet. Eine Zusammenstellung der verschie- denen beobachteten Blattformen mag das Gesagte belegen. Am lehrreichsten ist hier no. 63, bei der am gleichen Stengel typische /wscobrunnea- und ebenso * Zur Benennung von P. »chilensis», die ich der bisherigen Nomenklatur entsprechend hier im Text beibehalten habe, vergl. den Abschnitt iiber P. rubescens L. et L. DIE FOLIOSEN LEBERMOOSE DER JUAN FERNANDEZ-INSELN UND DER OSTERINSEL 707 typische chz/enszs-Blatter vorkommen! Auch in der R6étung der abgestorbenen Blattzellen stimmt P. fuscobrunnea mit P. chilensis vollig iiberein. Ich moéchte danach P. fuscobrunnea fiir eine Kleinart oder, vielleicht besser, da solche Formen bisher auf dem chilenischen Festland nicht beobachtet wur- den, ftir eine insulare Rasse der P. »chzlenszs» halten. Dass aber auch noch festlandische Formen die Inselgruppe erreichen, beweist die erwahnte no. 6 pp., die man nach ihrer stark dornigen Bewehrung auch ebensogut als kurzblattrige Form der P. »chzlensis» bestimmen k6nnte. Geographische Verbreitung: Endemisch auf Juan Fernandez. P. rubescens L. et L. (Fig. 1 ]). Nach Einsichtnahme in die Originale (Hb. Lindenberg no. 778 und 779) bin ich beinahe sicher, dass diese Art identisch ist mit P. chz/enszs St. Die Unterschiede sind minim. Im Formenkreis von P. chzlens7s entspricht in ihrem ausseren Bild P. rudescens sogar gerade jener Form, die wir als die typische P. chilensis anzusehen pflegen. Der Vorderrand des Blattes von P. rubescens ist zwar etwas mehr gezahnelt, wahrend er bei P. chzlenszs fast ganzrandig ist, doch tragt er auch bei dieser gelegentlich ein par Zahnchen. Die Blaurotfarbung der toten Zellen ist fiir die ganze Sektion » Chzlenses» CARL (2) charakteristisch und findet sich ebenso schén bei P. vubescens (daher wohl der Name!), desgleichen der Firnisglanz der Pflanze. Grodsse, Proportionen, Blattzuschnitt, die Haufung langer Dornzahne an der zu einer niederen Crista zusammenschliessenden Basis des postikalen Blattrandes, alles ist genau wie bei P. chzlenszs! — Ich habe sogar den Verdacht, dass P. rudescens aus der gleichen Gegend (Chile) stammt, da wir bei alteren Autoren oft geographische Ungenauigkeiten begegnen.’ Auch fehlt die Pflanze in SPRUCE’s »Hepaticae Andinae et Amazonicae». Bei ihrer auffallenden Erscheinung kénnte sie SPRUCE kaum tbersehen haben, wenn sie in den tropischen Anden vorkame. Sie ist auch spater nicht dort gesam- melt worden, wahrend die als P. chzlenszs beschriebene Pflanze aus dem siid- lichen Chile in jeder Sammlung erscheint und dort oft in grésster Uppigkeit entfaltet ist. Viele der aus Siidchile stammenden Herbarexemplare von P. cht lensis sind der P. rubescens im Herbar Lindenberg so ahnlich, dass sie, ohne Verdacht zu erregen, mit den Originalen der P. rudescens vertauscht werden kénnten. Ich halte die beiden daher fiir identisch. Der Name P. chzlenszs St. ware also zu streichen und durch den 4lteren, P. rubescens L. et L., zu er- setzen! * REICHE betont dies ausdriicklich in seinem Kapitel »Geschichte der botanischen Er- forschung Chiles» (15). Und unter die unzuverlassigen Autoren rechnet er auch HOOKER mit den Worten »Angaben... solcher Botaniker, welche wie HOOKER und ARNOTT, Chile nicht aus eigener Anschauung kannten und iiber die geographische Lage der von ihnen verzeich- neten Standorte oft nicht unterrichtet waren, sind haufig falsch oder unsicher». Zwar bezieht sich diese Bemerkung auf die Anwendung des Namens *Chile». Es diirfte aber mit »Peru» nicht anders sein, zumal die alten Botaniker fiir die geographische Verbreitung ihrer Pflanzen weniger Interesse zeigten und daher oft in der Mitteilung der Fundorte nachlassig waren. LINDENBERG’s Angabe »Peruvia (Herb. Hookeri)» weist schon auf diese Unsicherheit hin. Auch STEPHANI driickt seinen Zweifel durch die Fassung »Peruvia leg. ?» aus. 708 TH. HERZOG *15. P. pudetensis 5t. Auf Baumrinde oder Waldboden. Masatierra: Cordon Centinela, Blechnum cycadifolium-Ugni-Assoziation 600 —700 m (no. 80); Cerro Pirdmide, im Wald, 600 m (no. 83). Die Pflanze liegt in sterilen und 3 Sprossen vor und scheint nicht voll ent- wickelt zu sein, da viele der Blatter Hemmungsformen mit vereinfachter Rand- gliederung und betontem Hervortreten der beiden grossen Apikalzahne sind. Bucu’s Ansicht, dass die Art zu der Sektion » Chzlenses» CARL gehore, glaube ich unterschreiben zu ko6nnen. Zu beachten sind fir die richtige Bestimmung die ziemlich weiten und schon in der Blattmitte etwas gestreckten Zellen mit den deutlichen, dreieckig verdickten Zellecken. Geographische Verbreitung: Sidchile, Chiloé, Juan Fernandez. *16. P. homomalla St. (forma minus spinosa, Fig. 2, g—r). Auf Heideboden oder an Felsen, meist in Gesellschaft anderer magellani- scher Florenelemente. Masafuera: Cordon del Barril, ca 600 m (no. 72, 386 pp.); Hochland nérd- lich vom Correspondencia-Camp, ca 1,100 m (no. 74); Heide bei Las Torres und Correspondencia, bis 1,400 m (no. 75); Las Torres, an Felsen, 1,370 m (no. 93 pp-, 257 pp.). Im Herbar Stockholm liegen, von DUSEN (no. 251) und von HALLE und SKOTTSBERG (no. 279) gesammelt, zwei Proben unter Plagiochila heteromalla L. et L., von denen die eine handschriftlich von STEPHANI als »vera» bezeich- net wird (Fig. 2, g—n). Wenn es schon merkwiirdig anmutet, dass STEPHANI gewissermassen zwei verschiedene eteromalla anerkennt, von denen eine die »vera», die andere eben »unrichtig» ist, so wird die Sache noch unklarer da- durch, dass gerade diese angeblich echte »vera» gar nicht mit dem Original von P. heteromalla ibereinstimmt. Ich konnte mich davon durch Vergleichung mit den Typusexemplaren der P. heteromalla aus Peru im Herbar Lindenberg (Fig. 2, a—f) tberzeugen, nachdem ich dies schon aus der Beschreibung und Abbildung von LINDENBERG geschlossen hatte. Die dussere Ahnlichkeit der Pflanzen ist zwar so betrachtlich, dass eine Verwechslung schon im Bereich der Moglichkeit liegt. Andrerseits aber sind die Unterschiede sehr klar zu fassen. Wahrend namlich der Typus eine sehr dicht und fein dornig gezahnte Perianth- mindung besitzt, zeigen die patagonischen »vera»-Exemplare einen beinahe un- versehrten, nur mit entfernten, ganz kurzen (einzelligen) stumpfen Vorspriingen versehenen Mundsaum, den man bestenfalls als »repandus» oder »repando-den- tatus» bezeichnen kann. Die Unterschiede in der Randgliederung der Stengel- blatter und Floralblatter sind weniger auffallend, immerhin sind die Blatter bei der patagonischen Pflanze durchwegs gréber gezahnt als bei der peruanischen. Der allgemeine Blattumriss und die Grodsse der Blatter beider sind recht ahn- lich, ferner auch, wie schon gesagt, der Habitus, der durch die scharf einseits- wendigen, fast sparrig zuriickgebogenen Blatter zustande kommt. Zunachst nun stellten sich mir die Dinge so dar, als miisse auf dieser ab- DIE FOLIOSEN LEBERMOOSE DER JUAN FERNANDEZ-INSELN UND DER OSTERINSEL 7°9 Fig. 2. a—f: Plagiochila heteromalla Typus! (Hb. Lindenberg, Peru, 1832), a 2 Blatter Io/1; — b Blattzahn, ca 130/1; — c basales Zellnetz, ca 130/1; — d Floralblatt, 101; — e Perianth, ca 18/1; — f Zahne von der Perianthmiindung, ca 130/11. g—n P. homomalla St. = P. hetero- malla »vera» (DUSEN, no. 251): g 3 Blatter 10/1; — h Blattzahn, ca 130/1; — 1 basales Zell- netz, ca 130'1; — k 2 Involucralblatter 10/1; — 1 Perianth, 101; — m Perianthmiindung, 18/1; —n Randzellen der Perianthmiindung, ca 130/1. o—r: P. homomalla St. (SKOTTSBERG, Pto. Simpson: o 3 Blatter, 10/1; — p Zellnetz oben, ca 1301; — q Zellnetz unten, ca 1301; — r Perianth, Io’I. weichenden »vera»-Pflanze eine neue Art begriindet werden. Denn, dass sie sich nicht mit P. heferomalla vereinigen liesse, war augenscheinlich. Da kamen mir aber Exemplare der P. homomalla St. aus dem Herbar Stockholm (Fig. 2, o—r) LO TH. HERZOG zu Gesicht, die zwar nicht von der Typuspflanze selbst stammten, aber mit der Beschreibung und Abbildung STEPHANIS so vollkommen tbereinstimmten, dass sie unzweifelhaft hierher gehdren mussten. Sie sind auch mit dem gleichen Fundortsvermerk wie das Original versehen, nur der Sammler ist ein andrer. Sie stammen namlich von »Pto. Simpson, an einer feuchten Felswand», dem Fundort, den STEPHANI an erster Stelle in seiner Originalbeschreibung nennt und der demnach dem Original entsprechen dirfte, wenn die Belege auch nicht auf den Entdecker, CUNNINGHAM, zuriickgehen. Ich méochte sie daher ohne weiteres als Cotypus anerkennen. Damit ist nun erstlich die Unterbringung der P. heteromalla »vera» sicher- gestellt, die eben nichts andres als P. omomatlla ist, dann aber auch die Fest- legzung der vorliegenden Juan Fernandeznummern moglich geworden. Diese stimmen namlich trotz ihrer gewissen Merkmalsabwandlungen in allen wesent- lichen Punkten mit P. homomalla iberein, so dass ich nicht zogre, sie dieser STEPHANI'schen Art zuzuordnen. Die einzigen, aber sicher nur standortlich be- dingten Abweichungen finden sich im Grad der Blattrandzahnung, der beim Typus aus ziemlich regelmassigen, groben Zahnen besteht, bei manchen extre- men Formen der Juan Fernandez-Pflanzen jedoch mehr oder weniger abge- schwacht wird. Das ganz ibereinstimmende Blattzellnetz und die zu beob- achtenden Uberginge in der Zahnung, auch ihr gemeinsamer Habitus lassen ohne weiteres die Zusammengehorigkeit aller dieser Formen deutlich erkennen. Fir die Tracht massgebend ist die stets braunliche Farbe, der sehr kraftige Stengel und die derben, fast muschelschalig gewolbten, einseitswendigen Blatter, die eine gedunsene Beblatterung hervorbringen. Zur STEPHANI’schen Beschreibung von P. homomalla ware noch zu be- merken, dass die Erwahnung des Perianthes vollkommen fehlt, wahrend doch gerade das authentische Material von Pto. Simpson sehr schone Perianthien ent- halt. Solche Flichtigkeiten kommen eben bei STEPHANI leider 6fters vor und zwar nicht nur — wie hier — in seinen Nachtragen, wo sehr haufig Alters- und Ubermiidungserscheinungen zu beobachten sind, sondern auch, wie das friiher z. B. fiir Madotheca chilensis (9, 10, 11) erwahnt wurde, in den 4lteren Bear- beitungen. Das ebenfalls als P. homomalla bezeichnete Exemplar von »Isla Pacheco, in einer kleinen Lagune mit Lepicolea algoides» im Herbar Uppsala gehort aber sicher nicht hierher. Ob etwa der Beleg fiir diese P?. homomalla im Herbar Stephani verblieben ist, kann ich zur Zeit nicht nachpriifen. Bei der Vertei- lung von Mischrasen an verschiedene Herbarien sind ja haufig Irrtiimer unter- laufen. Es ist sogar vorgekommen, dass STEPHANI selbst gelegentlich unrichtige Belege ftir sich zuriickbehalten hat, die nun blinden Verehrern von Originalen als »Typen» gelten! Des ferneren bleibt fraglich, ob die im Herbar Lindenberg aufbewahrten Stiicke von P. heteromalla »\eg. POEPPIG, Chile» tatsachlich aus Chile stammen. Ihre ZugehGrigkeit zu der peruanischen P. heteromalla (von KUNTZE an LINDEN- BERG geschickt und die Typuspflanze darstellend) kann allerdings nicht bezwei- felt werden. Nur die Herkunft bleibt fraglich (vgl. Anmerkung 2!). DIE FOLIOSEN LEBERMOOSE DER JUAN FERNANDEZ-INSELN UND DER OSTERINSEL 711 17. P. Neesiana Ldbg. An Felsen und Baumen. Masatierra: Schmaler Baumfleck an senkrechter Felswand, Portezuelo, ca 600 m (no. 44); Quebrada de la Damajuana, Wald (no. 52) Yunque, an Holz, 400—500 m (no. 55). Samtliche Exemplare zeigen ausgezeichnete Ubereinstimmung mit dem Typus! ; Nordhang des Geographische Verbreitung: Westpatagonien bis Magellanstrasse, Chiloé, Stidchile, Juan Fernandez. tow Ps rinparia St, Auf Erde und feuchten Felsen. Masatierra: Quebrada Gutierrez, ca 300 m (no. 41), Nordhang des C. Pira- mide, 600 m (no. 54). Masafuera: Quebrada del Mono, Waldboden, 475 m (no. 40, 42); Casas- Cafion, feuchte Felsen im innersten Teil, ca 200 m (no. 45 pp.). Die vorliegenden Exemplare sind vollig identisch mit der Typuspflanze im Herbar des Riksmuseums in Stockholm. Ich kann jedoch kaum einen wesent- lichen Unterschied gegentiber P. Neestana finden, es seien denn die nur ein wenig grdsseren Blattzellen und der gelegentlich breitere Zuschnitt des Blattes. Da jedoch die Abbildung in STEPHANI’s Icones durch auffallend breite Ent- wicklung des ventralen Blattrandes hervorsticht, so médchte ich die Art nicht ohne weiteres einziehen. Eher ware daraus zu schliessen, dass — wie so oft — als Typus ein falscher Beleg aufbewahrt wurde. In diesem Fall miissten die hier der P. riparta zugerechneten Nummern zu P. Neeszana gezogen werden. Ich personlich neige der letzteren Alternative zu, bin aber, bei der Sparlichkeit des Vergleichsmaterials, ausserstande, dariiber eine endgultige Entscheidung zu treffen. Geographische Verbreitung: Westpatagonien, Juan Fernandez. 19. P. squarrosa St. — Syn. P. fernandesiensis St. In Gebiischen und im Decksonza-Wald. Masatierra: Auf dem Bergriicken zwischen Quebrada Agujeriada und QO. Laura, 650 m (no. 48, 49, 61, 71 c. per.); Cordon Centinela, Buschwerk, 530 m (no. 67), daselbst in der Blechnum cycadifolium-Ugnt-Assoziation, 600—700 m (no. 80 pp.); Bergriicken siidlich von Pangal, 795 m (no. 65; no. 29, fo. mznor); Salsipuedes, Dzcksonta-Wald, 660 m (no. 58). Masafuera: Cordon del Barril, an Déchsonza, goo—1,000 m (no. 64). Diese prachtvolle Art, die sich oft durch einen schénen Goldglanz aus- zeichnet, scheint auf Juan Fernandez besonders verbreitet zu sein. Sie ist in der vorliegenden Sammlung durch reichliches und bestentwickeltes Material vertreten. Nach Beschreibung und Zeichnung von STEPHANI (20) ist P. fernandesiensis St. vollig identisch mit P. sguarrosa St. Geographische Verbreitung: Westpatagonien und Juan Fernandez. 712 TH. HERZOG 20. P. robusta St. — Syn. 7. aveua St) Suppl.* Spee Hep; ava caulis Vayl.? An feuchten Felsen. Masafuera: Casas-Canon, im innersten zuganglichen Teil, an Felsen nahe dem Wasser, ca 200 m (no. 50). Die vorliegenden Exemplare passen am besten zu dem, was STEPHANI unter seiner P. arguta versteht. Durch vergleichende Anschauung kann man sich jedoch davon uberzeugen, dass ihre Unterscheidung von P. robusta nur auf sehr schwankenden Merkmalen beruht. So méochte ich P. argufa St. nur als Synonym von P. robusta St. gelten lassen. Es ist mir sogar sehr wahrschein- lich, dass auch P. vobusta selbst nichts andres als eine Form oder Varietat der wohl sehr vielgestaltigen P. duricaulis Tayl. sei. Alle derartigen Fragen harren noch der Entscheidung des Monographen. Auch H. CARL (2) hat einstweilen nur eine vorbereitende Gliederung des gewaltigen Stoffes erreichen kénnen. Sie wird aber bei einer vollstandigen synoptischen Bearbeitung der Gattung einmal gute Dienste leisten. Geographische Verbreitung: Feuerland, Westpatagonien bis in den argentinischen Nationalpark von Nahuelhuapi, Juan Fernandez. Tylunanthus Mitt.’ 27. 0A. dimbatus: Still ions3s ati): Auf Erde zwischen andern Lebermoosen herdenweise. Masatierra: Bergriicken siidlich von Pangal, 6stlich von Yunque, 795 m (no. 95, 298 pp.). Geographische Verbreitung: Westpatagonien, Juan Fernandez. 22. TT. ‘silvatieus St. (Fig: 3,/e): Auf Erde zwischen andern Lebermoosen oder ititber Farnwurzeln und Farn- stammen. Masafuera: Bett eines kleinen Flusses beim Correspondencia-Camp, 1,100 m (no. 94, 97); kleine Schlucht beim Correspondencia-Camp, 1,150 m (no. 96). Diese Art ist durch ihre etwa keilformige Blattgestalt gut charakterisiert. Geographische Verbreitung: Endemisch. *23. TT. densiretis Herz. n. sp. (Fig. 3, a—d). Sterilis; statura 7. sz/vaticum parum superans, laxe caespitosus vel grega- rius, pallide viridis, terricolus, aliis muscis intermixtus. Caulis simplex, erectus, ad 12 mm longus, cum foliis ad 3,5 mm latus, basi scapiformis, foliis squami- formibus remotis adspersus, superne distiche foliatus, rhizoidibus nullis nisi ba- salibus. Folia contigua, plano-expansa, ca 1,7 mm longa, ad 1,2 mm lata, sub- * Die bisherige Schreibweise » 7y/imanthus» beruht, wie K. MULLER festgestellt hat, auf einem Druckfehler! DIE FOLIOSEN LEBERMOOSE DER JUAN FERNANDEZ-INSELN UND DER OSTERINSEL ffs a Fig. 3. a—d: Tylunanthus densiretis Herz. n. sp.: a Stengel, 10/1; — b Blatt, 18/1; — c Zahne der Blattspitze, ca 130/1; — d Zellnetz unten, ca 130/1. — e—g: Zellnetze vom Blattrand bei e T. stlvaticus, — £ T. limbatus, — g T. patagonicus (alles 130/1). ovata, apice leviter emarginata, angusta basi inserta, superne irregulariter Centi- culata, margine dorsali basi anguste inflexo, margine ventrali e basi truncata alte arcuato; cellulae apicales diametro ca 18 , basales ad 40 » longae, 25 latae, parietibus strictis, trigonis nullis. Masafuera: Kleine Schlucht beim Correspondencia-Camp, ca 1,140 m (no. 92, 73 pp.) Die Art unterscheidet sich von den ubrigen Arten der Gattung auf Juan Fernandez einmal durch die fast eiformige Gestalt ihrer Blatter, dann aber auch durch die undeutliche Ausgliederung der beiden Blattlappen und das dichtere Zellnetz. Bei 7. sz/vaticus (Fig. 3, e) sind die Blatter langer und mehr keil- formig-spatelig, auch kommt meist die Zweilappigkeit deutlich zum Ausdruck, das Zellnetz ist, namentlich im unteren Teil des Blattes, wesentlich lockerer. Bei 7. “imbatus (Fig. 3, f) sind die breit inserierten, plump geformten Blatter, ihr an der Basis sch6n gerundeter ventraler Blattrand und das lockere Zellnetz kennzeichnend, weniger deutlich tritt die Differenzierung der Randzellen, auf die sich der Name bezieht, hervor. Immerhin kann man an den meisten Blattern eine starkere Wandverdickung der Randreihe beobachten. Ich habe zum Ver- gleich die Zellnetze dieser beiden Arten und des 7. patagonicus (Fig. 3, g) in der Figur wiedergegeben. Brauchbar zur Bestimmung der verschiedenen Arten scheint mir auch die Figur 9 von STEPHANI (19). Eine gewisse Abnlichkeit in der Blattform hat unsre Art auch mit Adelanthus unciformis, der aber habituell durch die seitlich dem Stengel angelegten und einseitswendigen Blatter auf den ersten Blick zu unterscheiden ist. 47— 41438. The Nat. Hist. of Juan Fernandez and Easter Isl. Vol. II. 714 TH. HERZOG Mylia S. F. Gray. 24. M. repens (Mitt.) Herz. — Syn. W. pseudorepens Herz. n. sp. in herb., Letoscyphus fernandestensis St. Masafuera: Bett eines kleinen Flusses beim Correspondencia-Camp, ca 1,100! m (no. 330) 333,64654 460.367): Die Angaben STEPHANI’s iiber die Zellgrésse (40 » oben, 40—50 p unten) decken sich nicht mit meinem Befund an Belegen, die STEPHANI selbst als Leioscyphus vepens bezeichnet und die auch tatsachlich in Blattform und -stellung, sowie in den Amphigastrien vollstandig zu seiner Beschreibung und Abbildung passen. Die Blattzellen messen namlich, ebenso wie bei den vorliegenden Juan Fernandez-Pflanzen, nur etwa 30 » in der Blattmitte. Im Gegensatz dazu finde ich bei JZ. fuegzenszs, die in der Blattform wiederum der STEPHANI'schen Be- schreibung entspricht, Zellen von 36-—40 y statt 30 pp, wie angegeben wird. Die JMZ7yfa-Arten Westpatagoniens bediirfen wohl noch einer kritischen Durch- arbeitung. Zur Zeit stehen ihrer Bestimmung noch deswegen grosse Schwierig- keiten im Wege, weil bei ihrer grossen Ahnlichkeit mit gewissen Ch7loscyphus- Arten, besonders Ch. zxtegrifolius, nur fertiles Material — und dies ist selten — Sicherheit gibt. Geographische Verbreitung: Neuseeland, Magellanstrasse, Juan Fer- nandez. 25. M. fuscovirens (Tayl.)? An Farnstammen (und auf Erde?). Masafuera: Quebrada del Mono, enge Schlucht, an Dzcksonza, 570 m (no. 322). Die Bestimmung ist lediglich durch Vergleichung mit Exemplaren durch- gefuhrt, die ich selbst nicht als ganz tbereinstimmend mit der STEPHANI’schen Beschreibung finde. Die Juan Fernandez-Pflanze zeigt namlich eine grob ge- zahnte Perianthmiindung, die zwar zur Originalbeschreibung TAYLOR’s (s. STE- PHANI Spec. Hep. II, s. 27) »sublaciniatum» ausgezeichnet passt. STEPHANI behauptet aber, das Originalexemplar habe eine vollig ganzrandige Perianth- mundung. Ob hier nicht der Irrtum bei STEPHANI liegt, der vielleicht gar nicht das wirkliche Original in der Hand hatte? Geographische Verbreitung: Magellanstrasse, Mittel- und Siidchile, Juan Fernandez. *26. M. ligulata (St.) Herz. Auf Erde. Masatierra: Sidwestseite von Portezuelo de Villagra, auf der Erde zwischen Gebiisch, 600 m (no. 250); SW-hang des C. Pirdmide, 600 m (no. 272); Que- brada de la Damajuana (no. 261). Die Juan Fernandez-Pflanzen stimmen vollkommen mit den im Universitats- herbar von Uppsala liegenden Exemplaren iiberein. Geographische Verbreitung: Westpatagonien, Juan Fernandez. DIE FOLIOSEN LEBERMOOSE DER JUAN FERNANDEZ-INSELN UND DER OSTERINSEL 725 Lophocolea Dum. Diese in der Subantarktis, Westpatagoniens einerseits, Ostaustraliens, Tas- maniens und Neuseelands andrerseits so ungeheuer artenreich entfaltete Gattung kénnte an Handlichkeit ausserordentlich gewinnen, wenn man sie in einige na- tiirliche Gruppen zerlegte. Die Einteilungsversuche STEPHANI's (20) haben das Blattzellnetz und die Perianthform zu wenig bericksichtigt, als dass sie zu einer befriedigenden natiirlichen Gliederung hatten fiihren konnen. Allerdings sind wir bei der z. T. recht schwierigen Fassung der einzelnen Arten noch recht weit von der Erfillung dieses Wunsches entfernt. Mir selbst war es unmdglich, im Rahmen dieser Arbeit die monographische Behandlung dieser riesigen Gat- tung auf mich zu nehmen. Ich war daher auch bei einer ganzen Anzahl von Nummern mit sterilen Pflanzen nicht imstande, ihre Bestimmung durchzufihren. Es sei daher nur soviel aus der reichhaltigen Sammlung angefihrt, als ich glaube verantworten zu kénnen. Die hierbei geausserten kritischen Bemerkungen mégen vielleicht einem spateren Bearbeiter willkommende Hinweise geben. Mehr wollen sie nicht sein. 27. L. rotundifolia St. An Felsen und Rinde zwischen andern Lebermoosen. Masafuera: Cordon del Barril, an Drimys (no. 249); Las Torres, an Felsen 1,370 m (no. 93, 193 pp., reichlich und mit Perianthien); daselbst (no. 257, sehr sparlich, aber voll ausgebildet). Ich habe den Verdacht, dass diese Art nur eine Xeromorphose von ZL. Puc- cioana darstellt. Auf alle Falle sind beide sehr nahe mit einander verwandt. Das Original von ZL. Puccioana hat viel schwachere Eckverdickungen an den Blattzellen und ist grossblattriger als das von L. rotundifolia. Von letzterer, die bisher nur steril beschrieben war, enthalt das Juan Fernandez-Material schone Perianthien, die eine fast vollige Identitat mit denen von L. Puccicana zeigen. Nur ist die Zahnung der Miindung etwas grdber. Auch das Involucrum ist nahezu vollig iibereinstimmend und zeigt die feine und dichte mamill6se Zahne- lung wie bei ZL. Puccroana. Geographische Verbreitung: Falklandsinseln, Feuerland, Westpata- gonien, Juan Fernandez. 28. L. fernandeziensis St. Auf Fallholz und an Felsen. Masatierra: Anson-Tal, bei Plazoleta, auf totem Holz, 260 m (no. 244); Quebrada de la Damajuana, Walder, (no. 260); Valle Colonial, Quebrada seca, an Drimys, 435 m (no. 312); Cordon Salsipuedes, Felsspalten, 600 m (no. 243). Masafuera: Quebrada del Mono, enge Schlucht, an Dzcksonia, 570 m (no. 246). L. fernandestensis ist von STEPHANI nur als steril beschrieben und abge- bildet. Es ist daher schwierig, ein sichres Urteil tiber die Art zu gewinnen. 716 TH. HERZOG Nach der Beschreibung dirften als bezeichnend folgende Merkmale gelten: auf- gerichtete fast flache Blatter, ihre einfache Gestalt mit gestutzter Spitze und abgerundeten Ecken, die verhaltnismassig kleinen (27% 27 ») Zellen, die am Rand etwas kerbig vorspringen, und die verhaltnismassig grossen und stark zerschlitzten, langzipfligen Amphigastrien. Alle diese Eigenschaften ko6nnen bei unserer Art beobachtet werden, ebenso die Ausbildung der d Bracteolen, die STEPHANI zwar abbildet, aber nicht beschreibt. Nur sind die Amphigastrien meist noch etwas weiter zerschlitzt, fast wie bei ZL. Kvauseana (nach der Abbil- dung!), zu der unsere Proben in den vegetativen Teilen sehr nahe Beziehungen zu haben scheinen. Da ich aber auch von dieser die Typuspflanze nicht ge- sehen habe, so bin ich nicht imstande daruber zu urteilen, wie weit die Art abzuandern vermag. Die von L. Arauseana bekannten Perianthien, bezw. deren Involucrum, wird namlich mit sehr stark gezahnten Randern beschrieben und abgebildet, wahrend die von mir friher, in der Sammlung HosseEus (8) als Z. Krauseana bestimmten Pflanzen durchwegs schwachere Zahnung aufwiesen. Da aber auch in der Beschreibung von L. Avauseana gesagt wird, dass eines der Involucralblatter auch ungezahnt und abgestumpft sein kann (»saepe altero sub- integerrimo obtusato»), so ist anzunehmen, dass dies gelegentlich (oder vielleicht sogar haufiger —- wenn die STEPHANI’sche Typenpflanze ein Extrem bilden sollte —) auch im ganzen Involucralkreis vorkommen kann. Diese Uberlegung hat mich l.c. veranlasst, die HOSSEUS’schen Pflanzen zu L. AKvauseana zu ziehen. Nun stimmen die hier in unserer Sammlung vorliegenden Pflanzen, die — wie gesagt — nach den vegetativen Teilen ohne weiteres als L. fernandeziensis be- stimmt werden missten, auch sehr weitgehend, und namentlich im Perianth und Involucrum mit meinen HOSSEUS’schen ZL. Avauseana-Proben wuberein, so dass ich glauben méchte, dass diese beiden Arten nur die Extreme einer und der- selben Art darstellen, die dann den alteren Namen L. Avauseana fihren misste. Ohne jedoch die Originale gesehen zu haben, mochte ich meine Ansicht oder Vermutung nicht in der Form einer Behauptung aussprechen. Schon die unter dem Namen L. fernandeziensis zusammengefassten Pflanzen dieser Sammlung sind so wuberaus plastisch in allen Teilen — die Wandelbarkeit betrifft sowohl Form wie Grosse der Blatter, allgemeinen Habitus, Grosse und Gliederungsgrad der Amphogastrien — und stimmt mit dem, was ich bei meinen vermuteten ZL. Kyrauseana-Formen der HOSSEUS’schen Sammlung gesehen habe, so gut tberein, dass ich nicht an ihrer Identitat zweifeln kann. Die einzige Frage bleibt einst- weilen, ob ich berechtigt war, die HOSSEUsS’schen Pflanzen trotz ihrer gewissen Abweichungen zu ZL. Krauseana zu rechnen. Die grosse Variabilitat dieses Formenkreises wird wtbrigens leichter verstandlich, wenn man die weite Ver- breitung der hierher gehérigen Pflanzen (unter Einschluss von L. fernandesiensis) betrachtet. Geographisehe Verbreitung: Chile, Juan Fernandez. 29. L. pallidevirens (Tayl.) St. Auf Waldboden und Steinen. Masatierra: In Dickichten auf dem hohen Felskamm tuber Pangal, 795 m (no. 262, 298 pp.); SW-Seite des C. Pirdmide, 600 m (no. 267). DIE FOLIOSEN LEBERMOOSE DER JUAN FERNANDEZ-INSELN UND DER OSTERINSEL ie iy Masafuera: Quebrada del Blindado, Waldboden und Steine, 440 m (no. 206). Dank ihrer ungeschlachten Gestalt mit den ungeteilten, grosszelligen Blattern und verhaltnismassig kleinen Amphigastrien ist die Art trotz ihres betrachtlichen Formenreichtums und meist sterilen Auftretens doch ziemlich leicht zu erkennen. In den Herbarien herrscht einigermassen Ubereinstimmung zwischen den als Z, pallidevirens bezeichneten Pflanzen. Geographische Verbreitung: Falklandsinseln, Feuerland, Siid- und Westpatagonien, Chiloé, Siidchile und Juan Fernandez. ao. 6. “attenuata: St. — Syn. £..cornuia St. Auf Erde, am Grund der Baume und auf Fallholz. Masatierra: Wald am Centinela-Riicken, auf Fallholz, 530 m (no. 266, 326 pp., fo. drevzfolia); Quebrada de la Damajuana (no. 287 fo. drevifolia und 392); Cumberland Bay, Pico Central, 390 m (no. 258 fo. cornuta); Valle Colo- nial, Quebrada seca, an alten Striinken und Stammen, 435 m (no. 259 fo. cor- nuta (St.)); Salsipuedes, Dzcksonia-Wald, 660 m (no. 277, 314, fo. drevifolia); Villagra, Quebrada Choza, Baumstamme und Striinke, 250 m (no. 284). Masafuera: Quebrada del Blindado an einem JDecksonza-Stamm, 440 m (no. 280 fo. brevifolia). L. attenuata St. und L. cornuta St. wird man kaum spezifisch trennen k6n- nen. Die als L. cornuta St. bestimmbaren Formen diirften bestenfalls als eigene Form der ZL. attenuata St. anzusehen sein. Geographische Verbreitung: Westpatagonien, Sidchile, Juan Fer- nandez. 31. L. textilis Tayl. Auf feuchter Erde. Masafuera: Casas-Cafion, feuchter Boden zwischen Asplentum magellanicum (no. 265, forma sterilis inevoluta), Durch ihre grossen durchsichtigen Blattzellen recht auffallende und nur mit L. attenuata und L. campanulata zu vergleichende Art. Die vorliegenden Ex- emplare sind leider als vdllig steril und unentwickelt nur mit Vorsicht hier unterzubringen. Geographische Verbreitung: Falklandsinseln, Feuerland, Magellan strasse, Westpatagonien, Juan Fernandez. 32. L. divergenticiliata St. An feuchten Felsen und Fallholz. Masatierra: Anson-Tal, bei Plazoleta, auf Fallholz, 260 m (no. 13 pp.); Nordhang des Cerro Pirdmide, an einem Baumfleck, 600 m (no. 268); Siidhang des Cerro Chumacera, an feuchten Felsen (no. 315). Masafuera: Casas-Cafion, feuchte Felsen, ca 200 m (no. 270); daselbst, in feuchten Felsnischen (no. 293); Quebrada de las Vacas, unter Cadystegza (no. 309). 718 TH. HERZOG Eine durch die Blattform mit meist 3 feinen, spreizenden Dornzipfeln leicht kenntliche Art! Geographische Verbreitung: Feuerland, Westpatagonien, Siidchile, Juan Fernandez. 33. L. muricata Nees. — Syn. L. hzrtifolia Tayl. J. of Bot. 1846, p. 366; L. horridula Sande Lac. Syn. Hep. javan., p. 30 (Fig. 4, k). Meist auf gefallenem Holz oder an Farnstammen. Masafuera: Im Bett eines kleinen Flusses beim Correspondencia-Camp, am Strunk einer Lophosoria, 1,100 m (no. 234 pp.). Die auf dem Festland haufige und uberhaupt in der Siidhemisphare weit verbreitete Art scheint im Inselgebiet von Juan Fernandez selten zu sein. Sie wird fast vollig vertreten von der folgenden ihr zwar nahestehenden, aber doch wohl geniigend unterschiedenen neuen Art L. submuricata Herz. Geographische Verbreitung: Ostaustralien, Neuseeland, Westpatago- nien und Sidchile bis Valparaiso, Juan Fernandez, Sudbrasilien, Cuba, Suidafrika, Mozambique, Insel Bourbon, Ostindien, Java,. Neuguinea. #24. L. submuricata Herz. n. sp. (Fig. 4, a—i). Monoica, gracillima, late caespitosa, viridissima, aspectu L. murzcatae com- paranda, valde variabilis, lignicola vel saxicola. Caulis repens, ad 2 cm longus, rhizoidibus longis affixus, dite inordinatim ramosus, Folia variabilia, nunc — in caulibus ramisque sterilibus — remotiuscula, subrecte patentia, planissima, e basi angustiore subrectangulata vel subovata, 4, ad 7/; biloba, supra medium + dense muriculato-dentata, nunc — in plantis bene evolutis, fertilibus — densiuscule imbricata, sub angulo ca 60° patentia, convexa, subfalcato-ovata, brevius biloba, immo subtruncata angulisque breviter apiculatis, marginibus minus denticulatis, rarius subintegerrimis, ad 0,9 mm longa, 0,5 mm lata; reti cellularum densa, opaca, cellulis apicalibus diametro 1c—12 y, laevibus, basalibus parum diversis, ubique chlorophyllosis, parietibus validis, trigonis nullis. Amphigastria caulina libera, parva, ca 150——200 » longa, ad basin fere bifida, laciniis basi 2 cellulas, superne I cellulam latis, margine utrinque lacinula altera piliformi armata, supra basin angulata vel obtuse denticulata. Folia subinvolucralia et involucralia ma- jora, subcomosa, magis prorsus spectantia, caulinis angustiora, subconvoluta, similia, amphigastrium florale caulinis duplo vel triplo majus, lanceolatum, ad medium bilobum, lobis anguste lanceolatis, acuminatis, conniventibus, margine parce fimbriolato-dentatis, sinu angusto, acuto. Perianthia longe emersa, ad 2 mm longa, oblongo-ovoidea, sub ore trilobo parum constricta, obtuse tricarinata, lobis late triangulatis, breviter bifidis, muriculato-dentatis, facie nunc laevissima, nunc (ut in ZL. muricata) muriculato-hispidula. — Androecia in caule vel ramis pro- priis terminalia, longe spicata, bracteolis 8—10-jugis, acute bilobis, apicibus squatrulosis, margine muriculatis, basi involutis, globoso-saccatis, monandris, antheridiis magnis, globosis, diametro ca 160 1. An Farnstammen, Fallholz, auf Erde und Steinen. DIE FOLIOSEN LEBERMOOSE DER JUAN FERNANDEZ-INSELN UND DER OSTERINSEL 719 Fig. 4. Lophocolea submuricata Herz. n. sp. — a St. blatt, 37 — b St. blatt mit Amphi- gastrium, 37/1; — c Zahne von einem Blattlappen, ca 130/1; — d St. amphigastrium, ca 130 I- —e St. stiick mit einem Perianth, 181; — f 2 Perianthe, 18:1; — g Subinvolucralamphigast; rium, ca 130/1; — h Involucralamphigastrium, ca 70/1; — 1© Ahre, 18/1. — k Lofhocolea muricata, St. amphigastrium, ca 70/1. Aa Masatierra: Dickichte auf dem Centinela-Riicken, 530 m (no. 276). Masafuera: Casas-Tal, an Steinen (no. 275); Quebrada del Blindado, an einem Dvecksonia-Stamm, 440 m (no. 133 pp., 280 pp.); Hoéhenriicken zwischen Sanchez und Toltén auf Fallholz und Erde, 215 m (no. 281, 308). Diese Art steht Z. muricata Nees unzweifelhaft nahe, besonders in jenen Formen, die ein bestacheltes Perianth besitzen. Hier, aber auch bei Pflanzen mit kahlem Perianth, finden sich auch auf den Involucralblattern da und dort einzelne Zellen zu kurzen Stacheln ausgewachsen; doch sind die Stengelblatter bei allen Formen vdllig glatt, wahrend sie bei ZL. muricafa, abgesehen von den bedeckten Stellen, auf der ganzen Flache bestachelt sind. Sehr auffallig ist die starke Variabilitat der Stengelblatter, deren mwrécafa-ahnliche Form und Rand- bewehrung am schénsten bei den sterilen Sprossen zu beobachten ist. Aber auch dicht beblatterte fertile Sprosse kénnen solche Blatter hervorbringen (s. Fig. 4), obwohl sie vorzugsweise in der allgemeinen Form vereinfacht, meist 720 TH. HERZOG nur kurz eingeschnitten oder gar quergestutzt sind und ihre beiden Lappen nur durch aufgesetzte Eckzahne andeuten, auch oft nur undeutlich Randzahnung zeigen. Der Hauptunterschied gegeniber ZL. muricata liegt in den Stengel- amphigastrien, die bei ZL. murzcata deutlich fiederig gewimpert sind, wahrend sie bei L. submuricata nur je einen fransenartigen Seitenzahn besitzen. Ferner sind bei £. muricata die Involucralblatter mit dem Amphigastrium verwachsen. Nach der Beschreibung von STEPHANI (20) hatte bei der Bestimmung der vorliegenden Pflanze ev. auch ZL. sfzxosa G. in Betracht kommen konnen. Ein Blick auf die Originalzeichnung GOTTSCHE’s, die mir aus dem Botanischen Mu- seum in Berlin entgegenkommenderweise zur Ansicht tberlassen wurde, zeigt jedoch sofort den wesentlichen Unterschied. Bei L. sfzuosa besteht die Rand- gliederung der Blatter aus fast zilienartigen langen Dornzahnen, wahrend an ihrer Stelle bei ZL. swbmurvicata nur kleine, wie dem Rand aufgesetzte Zahnchen zu beobachten sind. Diese kénnen sogar sehr sparlich werden und beinahe ver- schwinden, auch ist die Zweilappigkeit des Blattes bei ZL. sfzvosa viel starker ausgepragt, wahrend sie im aussersten Fall bei L. submuricata durch 2 kleine Eckzahne mit dazwischen liegender breiter, fast gestutzter Bucht ersetzt werden konnen. Sehr iiberraschend ist die grosse Ahnlichkeit unserer Art mit der vom atlantischen Kiistengebiet Europas und den Mittelmeerlandern bis zu den Kana- rischen Inseln verbreiteten ZL. fragrans Tayl. (= L. hirticalyx). Die Uberein- stimmung beider im Bau des Perianthes ist geradezu verbliffend. Sie erstreckt sich hier bis auf die labile Ausbildung der Stacheln, die in manchen Fallen ganzlich fehlen, in andern ansehnlich entwickelt sein konnen. In der Rand- gliederung der auch in ihrer Kleinzelligkeit und dem Chlorophyllreichtum ahn- lichen Blatter zeigt sich indessen bei genauerer Vergleichung der Unterschied, dass bei L. fragrans grobere Zahne gebildet werden, wahrend man bei L. suo- muricata nur von Zahnchen sprechen kann und diese selbst auch fehlen konnen. Ebenso ist bei ZL. fragrans die Zweilappigkeit des Blattes entschiedener. Die bei L. submuricata so oft anzutreffenden fast quergestutzten Blatter mit auf- gesetzten Eckdérnchen gibt es bei LZ. fragrans nicht. Bei dieser sind die beiden Lappen immer gut geschieden. Auch in den Amphigastrien liegen gewisse kleine Unterschiede. Alles in allem aber kénnen die engen Beziehungen zwischen diesen raumlich so weit getrennten Arten nicht tbersehen werden. Dieselben werden allerdings bei einer genaueren Umschau in den floristischen Beziehungen des atlantischen Europas zu Siidamerika weniger befremdlich erscheinen. Die erst kurzlich in Irland entdeckte Zelaranea nemoides stellt schon eine solche interessante Disjunktion zwischen Siidamerika und Westeuropa dar. Aber auch weitere Belege ahnlicher Art lassen sich aufzahlen. So fand ich z. B. in einer Sammlung des Pater GUSINDE aus Westpatagonien vollkommen typische J/y/za cuneifolia, die man bisher auf Westeuropa und die Azoren beschrankt glaubte, und auch der von Feuerland tiber die ganze Andenkette bis Mittelamerika ver- breitete Adelanthus unciformis steht dem irlandischen A. dugortiensis so nahe, dass man im Zweifel sein kann, ob die beiden Arten nicht tiberhaupt identisch sind. Ahnliche Zusammenhange diirften sich bei einer vergleichenden Betrach- DIE FOLIOSEN LEBERMOOSE DER JUAN FERNANDEZ-INSELN UND DER OSTERINSEL 721 tung der beiden Floren noch in grdsserer Zahl aufdecken lassen. Die bisher bekannt gewordenen Beispiele beweisen jedenfalls, dass wir an die Méglichkeit historischer Beziehungen zwischen den heute so weit getrennten Gebieten den- ken miussen. Chiloscyphus Corda. 35. Ch. integrifolius L. et L. — Syn. Ch. brevistipulus St. 1. c. 1911. Auf Walderde, Steinen und Fallholz. Masatierra: SW-Hang des C. Piramide, 600 m (no. 147); Valle Colonial, Quebrada seca, auf Fallholz und an alten Stammen, 435 m (no. 143, 146, 149, 153); Cordon Salsipuedes, Gestriippwald, 615 m (no. 151, 152). Masafuera: Quebrada de la Loberia, Waldboden und Steine, 280 m (no. 156, 157); Chozas, Lumawald (no. 145); Quebrada del Mono, enge Schlucht, 370 m (no. 142 pp.); Hodhenriicken zwischen Sanchez und Toltén, auf Fallholz, 515 m (no. 144, 148 pp.). Da sich die kleinen Stipulae des Ch. brevistipulus St. auch an Exemplaren finden, die die charakteristischen Involucren und Perianthien des Ch. zutegrifolius tragen, kann ich den Artwert des STEPHANI'schen Ch. brevestipulus nicht an- erkennen. Die angegebenen Unterschiede im Zellnetz zwischen den _ beiden Arten liegen innerhalb der Variationsbreite des Ch. mtegrifolius. Vielleicht lasst sich aber Ch. brevistipulus als lokale Inselform von Ch. zxtegrifolius auffassen, da diese Juan Fernandez-Pflanzen alle im Habitus etwas von den mir bekannten festlandischen Formen abweichen. Geographische Verbreitung: Magellanstrasse bis Mittelchile, Juan Fernandez. Saccogyna Dum. *326. S. squarristipula Herz. |. c. (8). Meist an Rinde, am Grund der Baum- u. besonders Farnstamme und auf Fallholz. Masatierra: Riicken zwischen den Talern von Piedra Agujeriada und Laura, 620 m (no. 289, 331); Dickichte am Centinelariicken, 530 m (no. 279 pp.); Wald am Centinelariicken an Baumen (no. 325, fo. odtusafa) auf Fallholz, 530 m (no. 326, 339); Cerro Damajuana, auf Rinde in feuchter Lage (no. 294); Héhen- ricken siidlich Pangal, an Baumfarnen (no. 324); Cumberland Bay, Pico Central, 390 m (no. 323); Salsipuedes, Dzcksonta-Wald, 660 m (no. 328). Am ehesten diirfte diese Art mit S. jugata Mitt. (= Chiloscyphus granulatus Schiffn.) verglichen werden, ohne jedoch identisch zu sein, wie man etwa nach der Umrisszeichnung in STEPHANI’s Icones glauben kénnte. Die sehr korrekte Abbildung von JV. Schiffuer (»Uber exotische Lebermoose...», Nov. Act. K. Peop. Carol... D: Ak. d_ Nat... Bd 1X... ne, 2. .Halle-13093. 5S. 263, Tafel 10) widerlegt jedoch eine solche Annahme klar. Geographische Verbreitung: Siidchile, Juan Fernandez. 722 TH. HERZOG Marsupidium Mitt. #27. M. piliferum St. — Syn. 7ylmanthus bisetulus St. Sp. Hep. Suppl. 1922, S. 246. Anscheinend nur in Hohenlagen. Masatierra: Hoher Felskamm siidlich Pangal, dstlich von Yunque, 795 m (no. 186, 298 pp.). Masafuera: Cordén del Barril, auf Drzmys (no. 108). Eine Gegenuberstellung mit der Beschreibung und Abbildung von 7y/- manthus bisetulus St. aus Neucaledonien ergibt auch ohne Einsicht ins Original vollige Identitat. STEPHANI hatte bei der Veroffentlichung von /7ylzmanthus bisetulus (1922) offenbar sein Marsupidium piliferum von 1908 vergessen. Die spatere Gattungszuordnung zu /ylzmanthus war ein Fehlgriff, der allerdings die Neuaufstellung einer Art von 7ylzmanthus begreiflich macht. Denn in dieser Sippe steht die Art allerdings recht isoliert. Im wbrigen ist der Nachweis von Marsupidium piliferum (als » 7. bisetulus») auf Neucaledonien zwar interessant, aber nicht weiter tiberraschend, da J. pzliferum ausser bei Corral in Chile (leg. KRAUSE) auch schon von Neu Sidwales (leg: F. v. MULLER) bekannt war. Sie gehdrt also zu jenem Disjunktelement zwischen Australien und Siidamerika, das, in sehr vielen Fallen schon belegt, in der hoheren Pflanzenwelt fast nur in Gattungsdisjunktion ausgepragt ist, bei den Moosen sich aber auch weitgehend auf Artidentitat erstreckt. Die Blattgestalt ist bei J/. pzlzferum ganz einzigartig, indem bei fast vollig kreisrundem Umriss am apikalen Rand ganz unvermittelt 2 derbe, lang krallen- formige Borsten (wegen ihrer Derbheit eher Borsten als Haare zu nennen!) ent- springen. Sie krimmen sich nach der Stengelunterseite und sind schon an un- entwickelten Blattern fast vollig ausgebildet. Geographische Verbreitung: Neucaledonien, Ostaustralien, Siidchile, Juan Fernandez. Adelanthus Mitt. 38. A. sphalerus Hook. et Tayl. fo. nov. mzxor Herz. Differt a forma typica follis parvis, biseriatim divaricatis. An Felsen in hoheren Lagen. Masatierra: Hoher Felskamm siidlich Pangal, 795 m (no. 298 pp.). Von vielen der in der Gebirgsheide der Inseln vorkommenden Arten sind Zwergformen zu beobachten, die vielleicht als Hungerformen des nahrstoffarmen Bodens zu betrachten sind. So liegen z. B. von diesem gleichen Standort noch solche Kimmerformen von Y¥amesontella colorata, Lepicolea ochroleuca und Lepzr- dolaena magellanica vor. Man konnte sie vielleicht alle unerwahnt lassen, wie ich das bei diesen drei letzteren getan habe, weil bei ihnen die Zusammenhange unschwer zu erkennen sind. Bei der vorliegenden Form von Adelanthus spha- ferus empfiehlt sich jedoch eine besondere Hervorhebung, weil sie habituell nicht nur durch ihre winzigen Dimensionen, sondern auch durch die nach beiden DIE FOLIOSEN LEBERMOOSE DER JUAN FERNANDEZ-INSELN UND DER OSTERINSEL 123 Seiten fast horizontal abstehenden, aber senkrecht zum Stengel gedrehten Blatter einen so abweichenden Habitus erhalt, dass die ZugehGrigkeit zu unserer Art schwer zu erkennen ist. A. sphalerus und A. unciformis stehen sich zwar sehr nahe, unterschei- den sich aber durch den Blattzuschnitt, der bei A. unciformzs immer eine Ab- rundung der Blattspitze mit gleichmassiger Randzahnung zeigt, wahrend bei 4. sphalerus die Blatter breit zugespitzt sind und unregelmassigere (an der Spitze selbst 2 grdéssere) Zahne tragen. Geographische Verbreitung der Gesamtart: Feuerland, Magellan- strasse, Westpatagonien, Juan Fernandez. Bazzania S. F. Gray. Die Gattung wurde Frl. M. FULFORD zur Bearbeitung iiberlassen. Bestim- mungen wurden wahrend der Drucklegung dem Herausgeber mitgeteilt; s. Nach- trag. In meinem Material fand ich nur Beimischungen (Masatierra, Cord6én Centinela). Lepidozia Dum. #39. L. sejuncta (Angstr.) St. var. nov. drevzseta Herz. Differt a typo statura minore, caulis corticis cellulis ampliatis, hyalinis necnon foliorum setis saepius 3 brevioribus, strictioribus. — An species propria? Auf Erde oder Baumwurzeln. Masafuera: Quebrada del Mono, enge Schlucht, 370 m (no. 142 pp.). Geographische Verbreitung der typischen Form: Tasmanien, Siid- brasilien, Mexico, Antillen, Azoren, Kamerun, Angola, Siidafrika. *40. L. bicuspidata Mass. Auf Erde? Masatierra: Riicken zwischen den Talern von Piedra Agujeriada und Laura, 650 m (no. 236 pp., zwischen L. d7sticha). Durch die schmal zweizipfligen Blatter, deren Zipfel aber flachig und nicht borstig sind, sehr charakteristische Art. Geographische Verbreitung: Feuerland, Westpatagonien, Juan Fer- nandez. “41. L. pseudozoopsis Herz. n. sp. (Fig. 5). Sterilis, parva, albescens, corticola. Caulis ca 10 mm longus, parum ramo- sus, cellulis corticalibus Zoopsidis instar maximis, turgidis, hyalinis, ca 10c— 120 1 metientibus, in sectione transversa 8, fasciculum centralem viridulum cingentibus obtectus, vitreo-pellucidus. Folia caulina distantia, patula, rarius approximata, subtransverse inserta, hyalina, vitrea, chlorophyllo parco, parva, 0,2 mm longa et lata, ad basin fere trifida, laciniis 5 cellulas longis, 1 cellulam latis, disco basali 1 cellulam alto, 6 cellulas lato, cellulis laciniarum infimis 724 TH. HERZOG maxime turgidis, 70 » longis, 40 » latis, sequentibus sensim minoribus, ultimis minutis, saepius obsoletis, emarcidis. Amphigastria caulina minima, appressa, bisetula, setulis turgidis, 2—3 cellulas longis, disco nullo, ex area caulis parce rhizoidifera enascentia. — Cetera ignota. Auf Fallholz, Farnen etc. sparlich zwischen andern Moosen. Masatierra: Centinela-Riicken, auf Fallholz, Farnen etc., 530 m (no. 227). Die Blattzellen enthalten nur sehr sparlich kleine Chlorophyllk6rner, die sich immer an den Zellgrenzen finden. Das Lumen bleibt im wesentlichen vollig chlorophyllfrei. Reicher an Chlorophyll sind die engen Zellen des Zentral- Fig. 5. Lepidozia pseudozoopsis Herz. n. sp. — a Stengelstiick, 37:1; — b St. blatt, ca 13017, —c St. amphigastrium, ca 130/1; — d Stengelquerschnitt, ca 75/1. stranges in Stengel und Asten, wo man gelegentlich auch in dem toten, ge- trockneten Material kleinste feine K6rnchen durch das ganze Zellumen beob- achten kann; auch hier finden sie sich aber vorzugsweise nahe den Querwanden, ahnlich wie wir das von den schmalen Blattzipfeln der Lepzdozzen, Trichocoleen etc. kennen. 42. L. fernandeziensis St. — Syn. L. effusiseta St. 1922. Auf Erde und Baumwurzeln(?), besonders im Dzcksonza-Wald. Masatierra: Riicken zwischeu den Talern von Piedra Agujeriada und Laura, Dicksonia-Wald, 650 m (no. 216, 220, 226); hoher Ricken uber Pangal, D7ck- sonta-Wald, 580 m (no. 210, 218); Felskamm siidl. Pangal, 6stlich Yunque, 795 m (no. 211, 273); Dickichte am Centinelariicken, 530 m (no. 214); SW- Seite des Cerro Pirdamide, 600 m (no. 215, 224); Salsipuedes, Dzcksonza-Wald, 660 m (no. 217). Geographische Verbreitung: Endemisch. STEPHANI hat 1. c. (19) diese Art rite beschrieben und abgebildet. Man ist daher mit Recht erstaunt, in seinen »Species Hepaticarum» Suppl. 1922 eine L. fernandesziensis St. n. sp. (sic!) zu finden, deren Beschreibung und Abbildung in den Icones sich weit von seiner alten L. fernandeziensis (1911) unterscheidet. Hier (20) hat aber STEPHANI auch eine L. effuszseta n. sp. aufgestellt und deren DIE FOLIOSEN LEBERMOOSE DER JUAN FERNANDEZ-INSELN UND DER OSTERINSEL 725 Beschreibung wie Abbildung passt ausgezeichnet auf die L, fernandesiensis St. 1911. Ebenso weist die Fundortsangabe »Juan Fernandez (leg. SKOTTSBERG): auf diese Art, wahrend in der Ver6ffentlichung von 1911 (19) eine L. effusiseta St. fehlt. Es unterliegt daher fir mich keinem Zweifel, dass STEPHANI infolge Vergesslichkeit spater eine andre ihm neu erschienene Art mit dem bereits vergebenen Namen »fernandeziensis» belegte, seine eigene alte L. fernandeziensis aber nicht wieder erkannte und sie unter dem Namen L. effusiseta St. nochmals als neu beschrieb. Sollte jedoch keine Verwechslung vorliegen, sondern eine absichtliche Neubenennung der alten Art, so widersprache dies allen Gepflogen- heiten und namentlich den Nomenklaturgesetzen, da die alte L. fernandeziensis vollgultig publiziert war. Wenn STEPHANI als ihr Autor sich berechtigt glaubte, aus irgend einem Grund eine Namensadnderung vorzunehmen, so hatte er zum mindesten das Synonym hinzufiigen miissen und den gleichen Namen nicht wieder fur eine neue Art verwenden diirfen, da hierdurch nur Verwirrung ge- stiftet werden konnte. Ich ersetze daher den falschen Namen der spdteren L. Sernandeztensis St. durch L. erronea Herz. (nom. nov.), wobei ich allerdings, da ich die Art nur aus Beschreibung und Abbildung kenne, den Typus jedoch nicht gesehen habe, offen lassen muss, ob es sich itiberhaupt um eine eigene Art handelt. Ich médchte eher glauben, dass sie in den Formenkreis von Z. Facquemonty St. gehort. L. fernandeziensis St. schliesst sich wohl am engsten an L. plumulosa L. et L. an, von der sie sich jedoch durch wesentlich geringere Masse und steifer abstehende Blatter unterscheidet. Bei kraftigen Blattern kommen _allerdings Falle vor, wo statt der normalen 4 Borstenzipfel deren 5—6 auftreten kénnen, so wie sie fir L. p/umulosa durchwegs charakteristisch sind. Noch wichtiger allerdings fiir die Erkennung von ZL. plumulosa scheint mir, dass auch die Am- phigastrien fast stets 5—6-zipflig und fast ebenso gross wie die Seitenblatter sind. Ferner sind bei ZL. plumulosa die Borstenzipfel gegen den Basaldiskus deutlich gewinkelt und fast der Stammchenachse parallel aufgerichtet, wahrend sie bei L. fernandeziensis gleichmassig abstehen. Es kommt dadurch ein wesent- lich andrer Habitus zustande. L. fernandeziensis ist gewohnlich tiberaus zierlich und ihre Sprosse oft nur einfach und sehr regelmassig kurz fiederastig. Flagel- lenartige Verlangerung der Aste ist recht selten. Von L. fernandeziensis kenne ich noch Fundorte vom chilenischen Fest- land, von wo ich sie friiher als LZ. effuszseta verOffentlicht habe (6, 9). Die in den Thermen von Puyuhuapi bekannt gewordene Form ist etwas abweichend. Als nachste Verwandte unsrer Art kommt ZL. Liudenbergi in Betracht. Viel- leicht handelt es sich tberhaupt nur um eine geographische Rasse dieser in Neuseeland allgemein verbreiteten Art.(?) 43. L. plumulosa L. et L. Auf Erde und Farnstriinken. Masatierra: Héhenriicken iiber Pangal, am Stamm eines Baumfarnes (no. 222, fo. simplificata). 726 TH. HERZOG Bei der vorliegenden Form finden wir fast nur 4-borstige Blatter und Am- phigastrien, die Hauptsprosse also in ihrer Entwicklungshohe etwa den Seiten- sprossen der normalen Form entsprechend. Unsre Pflanze kann trotzdem als voll- entwickelt gelten, da sie 2 Involukren ausgebildet hat. Diese liefern das sicherste Erkennungsmerkmal, namlich fast kreisrunde und nur schwach randgegliederte Hiillblatter, die LINDENBERG in seiner Lepzdozza-Monographie sehr gut beschreibt. Merkwiirdigerweise hat STEPHANI in seinen Species Hepaticarum davon keine Notiz genommen. Er beschreibt nur kurz die Perianthien, ibergeht aber die so charakteristischen Merkmale der Involucralblatter. Man kann diese Beobachtung bei STEPHANI haufig machen, dass er Teile einer Pflanze, die von den 4lteren Autoren schon ausgezeichnet beschrieben wurden, einfach ignoriert. Z. B. bei Madotheca chilensis L. et L. und MW. gracilenta Tayl., wo wir in den »Species Hepaticarum» jeden Hinweis auf Perianth und Involucrum vermissen. Uber- haupt finde ich in vielen Fallen selbst die kurzen Diagnosen in der Synopsis Hepaticarum korrekter und pragnanter als die wortreichen Beschreibungen in den »Species Hepaticarum». Als weiteres Kennzeichen von L. plumulosa er- wahne ich die im trockenen Zustand fast schneckenformig eingerollten Spross- und Astspitzen, sowie ihre etwas abgeplattete Querschnittsform und weiche Beblatterung. Geographische Verbreitung: Falklandsinseln, Feuerland, Westpata- gonien, Siidchile, Juan Fernandez, Neuseeland (Siidinsel), Aucklandinseln, Anti- podeninseln. (Die Angabe iiber ein Vorkommen auf Gronland beruht vermut- lich auf einem Irrtum). #44. L. fragillima Herz. n. sp. (Fig. 6). Dioica videtur; late denseque caespitosa, tamen gracillima, confervoidea, fusca, opaca, terricola. Caulis capillaceus, decumbens, dite ramosus, irregulariter bipinnatus, cum foliis ca 0,35 mm latus. Folia caulina ca 0,1 mm longa, pun- genti-patula, contigua, subtransverse inserta, ad basin fere 4-partita, laciniis por- rectis, rarius parum incurvis, basi 2 cellulas latis, in setam 2—3 cellulas longam sensim attenuatis, cellulis rectangularibus sat validis, cuticula striolata, disco basali 1—2 cellulas alto, ramalia 3-partita, ubique fragillima. Amphigastria cau- lina ad basin fere 3- vel 2-partita, lacinis convergentibus, ramea vulgo bipartita, cruribus inaequilongis. — Inflorescentia 2 in caule ramisve rhizomaticis aphyllis gemmiformis, foliis involucralibus ad medium fere 3- vel 2-fidis, laciniis apice angustissimis, extus hamatis, tenerrimis, hyalinis, laevissimis, laciniis tantum leviter striolatis, margine varie fimbriolatis (juvenilibus subintegerrimis). — Cetera desunt, androecia haud visa. Auf Erde, in schattigen Hohlen. Masatierra: Centinela-Riicken, Erdhohle uater Dicksonia (no. 232). Die ausserordentliche Briichigkeit der Blatter macht sich gleich dadurch bemerkbar, dass schon das Entwirren der in einander verflochteten Sprosse ge- nugt, um in einem Praparat tausende von Bruchstiicken erscheinen zu lassen. Hierbei sind keine besonderen Bruchstellen praformiert, sondern die Briichigkeit DIE FOLIOSEN LEBERMOOSE DER JUAN FERNANDEZ-INSELN UND DER OSTERINSEL 127 Q = d Fig. 6. Lepidozia fragillima Herz. n. sp. — a St. stiick, 37/1: — b St. blatt. ca 130/1; — c Blattzipfel, 260/1; — d 2 St. amphigastrien, 130/1; — e Aststtick, ca 130/11; — f Ausschnitt aus einem Praparat mit abgebrochenen B. stucken, ca 1301; — g junge Involucralblatter, ca 70/1; — h ausgewachsene Inv.-blatter, ca 70/1. erstreckt sich uber das ganze Blatt, so dass ganz unregelmassige Bruchstiicke entstehen. In der Abbildung ist ein winziger Ausschnitt aus einem solchen Praparat wiedergegeben. Wenn unsre Pflanze nicht wohlausgebildete 2 Knospen triige, konnte man versucht sein, in ihr eine reduzierte Hohlenform irgendeiner der bekannten pata- gonischen Lepidozien zu vermuten. MHierftr kénnte eventuell ZL. diversifolia St. in Betracht kommen, doch finde ich bei unsrer Art die Involucralblatter viel starker zerschlitzt. Auch an ZL. capillarzs und L. nemordes kénnte man denken. Als nachste Verwandte méchte ich aber ZL. verrucosa St. und L. serpens Spr. ansehen. Von ihnen ist aber unsre Art unterschieden durch die fast bis zur Basis 4-zipfligen Blatter mit schmalen, dusserst briichigen Abschnitten. Auch sind die Zellen der Blattzipfel langer gestreckt und nicht eigentlich warzig, son. dern mit gestrichelter Cuticula versehen. 45. L. disticha St. Auf Walderde und Fallholz. Masatierra: Cordén de los Chifladores, auf Erde (no. 231); Centinela- Riicken, auf Fallholz, 530 m (no. 233, 235 pp.); Rucken zwischen den Talern von Piedra Agujeriada und Laura, 650 m (no. 236); Cumberland Bay, auf dem Boden der Gefangnis-Hohle 2, innersten Teil (no. 219); Valle Colonial, Quebrada seca, 435 m (no. 238). Geographische Verbreitung: Endemisch. 728 TH. HERZOG 46. L. Jaquemonti St. Auf Waldboden. Masatierra: Am Pfad nach Portezuelo, trockene Tuffwand im Wald (no. 212). Masafuera: Hohenriticken Ostlich von Inocentes, Dzcksonza-Wald, 750 m (no: 221). Von der in der Blattform ahnlichen und in Chile weit verbreiteten Z. chordulifera Yay\. unterschieden durch die stumpf graugriinen, lockeren Rasen und die lockere Beblatterung. Die im trockenen Zustand fast kugelschalig hohlen Blatter geben den Stengeln ein Aussehen, als seien hintereinander kleine K6rn- chen aufgereiht. Sprosse und Aste liefern daher ein fast perlschnurartiges Bild. Bei L. chordulifera ist die Beblatterung geschlossen, die Blatter ausserdem etwas derber und an der Basis mehr randgegliedert. Geographische Verbreitung: Feuerland, Magellanstrasse, Chiloé, Juan Fernandez. Herberta S. F. Gray. 47. H. runcinata (Tayl.) Herz. — Syn. Sendtnera runcinata Tay). J. of Bot. 1846, p. 372, Schisma runcinatum St. An feuchten Felsen. Masafuera: Las Torres, an Felsen, 1350—1370 m (no. 184, 23 pp., 257 pp.). Die vorliegenden Exemplare sind etwas schmachtig ausgebildet, die riick- wartsgerichteten Fransen des basalen Blattrandes daher nicht so deutlich und kraftig wie bei typischen Exemplaren. Trotzdem ist die Hierhergehorigkeit un- serer Nummern nicht zu bezweifeln. Es handelt sich eben auch hier wohl um Kummerformen, wie ich sie schon wiederholt erwahnt habe. Bemerkenswert ist die gelegentlich sehr undeutliche Abgrenzung der vitta von den Randzellen, doch liegt dies mehr an der Gleichheit der Farbe, als an mangelnder Diffe- renzierung in der Zellform. Auch hier sind die Vittazellen deutlich langgestreckt. Ganz wubereinstimmende Exemplare besitze ich auch vom _ festlandischen Chile aus der Sammlung Hosseus, die aber in meiner Publikation (8) nicht an- gefuhrt wurden, weil ihre Bestimmung zur Zeit der Drucklegung noch nicht sichergestellt war. Lepicolea Dum. 48. L. ochroleuca (Spreng.) Spruce. Auf Felsen und Heideboden. Masafuera: Hochlandsheide bei Correspondencia, 1,200—1,350 m (no. 203, 206); Gipfel des Cerro Inocentes, 1,480 m (no. 205, 207). var. seriata Herz. 1. c. (7). Masatierra: Portezuelo de Villagra, Westseite an hohen Felsen, ca 600 m (no. 202), und an einer Felswand, 590 m (no. 204). Masafuera: Las Torres, an Felsen, 1,350 m (no. 201); Heide bei Corres- pondencia, 1,350 m (no. 208). : DIE FOLIOSEN LEBERMOOSE DER JUAN FERNANDEZ-INSELN UND DER OSTERINSEL 729 Eine der verbreitetsten Arten Westpatagoniens, die sich als sehr formen- reich erweist. In ihren Kreis gehort wohl auch die von mir (7) als LZ. serzata fiir selbstandig gehaltene Form, die sich durch sehr deutlich gereihte Astblatter und einfachere Blatter auszeichnet. Diese entbehren namlich so gut wie voll- standig der schmalen, borstenformigen Anhangsel an den 4 Blattlappen, die bei der normalen Form gewohnlich unparig auftreten. Geographische Verbreitung: Feuerland, Westpatagonien, Siidchile, Juan Fernandez, Sidbrasilien, Sidafrika (Tafelberg). Lepidolaena Dum. 49. L. magellanica (Lam.) Schiffn. — Syn. Fungermannia magellanica Lamarck, Encycl. bot. III, p. 284; Polyotus magellanicus G. Syn. Hep. p. 248. Auf feuchtem Heideboden. Masafuera: Feuchte Heide bei Correspondencia, ca 1,350 m (no. 290). Auffallenderweise ist dieses schéne, auf dem Festland von etwa 44° siid- warts wohl allgemein verbreitete, auf Feuerland sogar gemeine Lebermoos auf Juan Fernandez nur sparlich vertreten. Hier tritt es anscheinend nur in den héchsten Lagen der Gebirgsheide auf, die in ihrem gesamten Vegetations- charakter wohl am meisten an die Lebensverhaltnisse der siidpatagonischen Gebiete erinnert. Sie stellt also einen jener haufigen Falle dar, wo die magel- lanischen Arten bei ihrem Vordringen nach Norden sich immer mehr in die hoheren Gebirgslagen zuriickziehen, wie wir dies auch von einer grossen Zahl hoherer Pflanzen kennen. Geographische Verbreitung: Ostaustralien, Tasmanien, Neuseeland, Feuerland, Westpatagonien, Sidchile, Juan Fernandez. Trichocolea Dum. 50. T. opposita St. Auf Baumstriinken und Waldboden. Masatierra: Valle Colonial, Quebrada Gutierrez, Wald, ca 300 m (no. 387, 391); Nordhang des Yunque, auf Baumstiimpfen, 400—500 m (n. 388). Ich habe den Verdacht, dass 7. caolzta St. nichts andres ist. Die Original- exemplare in den Sammlungen Stockholm und Uppsala sind kaum von einander zu unterscheiden. Form und Anheftungsweise der Blatter sind bei 777chocolea besonders stark abhangig vom Langenwachstum der Sprosse. Ich glaube, dass alle vermeintlichen Unterschiede damit zusammenhangen und dass in dieser Gattung viel zu viel Arten beschrieben wurden. Man kann neben fast quer- gestellten (an kurzen, gestauchten Sprosstiicken) sehr stark herablaufende, schief oder fast lings inserierte Blatter (an rascher gestreckten Abschnitten) beim gleichen Spross beobachten. Damit im Zusammenhang stehen auch gewiss die Verwachsungsverhaltnisse von Blattern und Amphigastrien, wovon man sich be- sonders bei der nachst folgenden Art iiberzeugen kann. Auf diesen sehr 48 — 41438. The Nat. Hist of Juan Fernandez and Eas‘er Isl. Vol. II. 730 TH. HERZOG schwankenden Merkmalen k6nnen daher nicht wohl Artunterschiede begriindet werden. Die Art diirfte der neuseelandischen 7. australis nahe stehen. Geographische Verbreitung: Endemisch. 51. T. verticillata St. An Felsen, Baumwurzeln, Baumfarnstammen und in der Farnheide an- scheinend verbreitet. Masatierra: Steile Bergwand gerade unter Portezuelo de Villagra, ca 500 m (no. 375); Baumfleck an der senkrechten Felswand, Portezuelo, 600 m (no. 376, 389); Nordhang des C. Pirdmide, 600 m (no. 377, 383, 385); Valle Colonial, Tuffbank in der Maqui-Assoziation, 208 m (no. 378); Puerto Ingles, Wald auf dem Zentralriicken, ca 500 m (no. 384). Masafuera: Gipfel des Inocentes, Heide, 1,480 m (no. 380); Las Torres, in feuchten Felsnischen, ca 1,500 m (no. 379); Kleines Flussbett auf dem Hoch- plateau, 1,100 m (no. 381); Farnheide itiber Chozas, 850 m (no. 382); Cordon del Barril, 650 m (no. 386); Quebrada del Blindado, an Dzcksonza-Stamm, 435 m (no. 390). Diese in Siidchile und durch ganz Westpatagonien weit verbreitete Art scheint auch auf Juan Fernandez ganz allgemein und haufig vorzukommen. Dem entspricht ihr betrachtlicher Formenreichtum der sich zunachst in dem ver- schiedenartigen Wuchs ausdriickt. Wir finden sowohl zarte, kleine, wie kraf- tige, mehr polsterartig wachsende Formen, die letzteren besonders in der Berg- heide der Inseln. Was STEPHANI (19) als 7. decrescens beschrieben hat, ist offenbar nichts andres als eine der vielen Formen von 7. werticillata, die sich im allgemeinen durch sehr breite und niedere Blatter mit lang herablaufendem Antikalrand und leicht zuriickgekrimmtem Hinterrand auszeichnet. Die Amphi- gastrien sollen nach STEPHANI beiderseits mit den Blattern verwachsen sein, man findet aber fast haufiger nur einseitig verwachsene oder ganz freie Amphi- gastrien am gleichen Stengel — auch die Festlandsformen verhalten sich so — und auch die Blattform, die STEPHANI fiir seine 7. decrescens abbildet, kann man zwischen den normalen, lang herablaufenden Blattern hier und dort beob- achten. Sie hangt bei den Trichocoleen wohl allgemein von der Intensitat des Langenwachstums des Sprosses ab. Die Originale, die ich sowohl aus Stock- holm, wie aus Uppsala erhalten habe, kann ich von der festlandischen 7. veri- cillata keinesfalls trennen. Eine endgiltige Klarung setzt jedoch ausgedehntere Studien in der sehr verworrenen Systematik der Gattung voraus. Geographische Verbreitung: Westpatagonien, Siidchile bis in den argentinischen Nationalpark von Nahuelhuapi, Chiloé, Juan Fernandez. Schistochila Dum. 52. Sch. Berteroana (Hook.) St. Auf feuchtem Waldboden und schattigen Felsen, besonders im D7cksonta- Wald. DIE FOLIOSEN LEBERMOOSE DER JUAN FERNANDEZ-INSELN UND DER OSTERINSEL hak Masatierra: Riicken zwischen den Talern von Piedra Agujeriada und Laura, 650 m (no. 18, mit Prerygophyllum tenuinerve); Hohenriicken iiber Pangal, 795 m (no. 11, fo. robusta); Quebrada de la Damajuana, Wald (no. 16, 20); Nordhang des C. Pirdmide, 600 m (no. 6); SW-Hang des Portezuelo-Riickens, im Dzck- sonia-reichen Wald, 550 m (no. 12); Valle Colonial, Quebrada Gutierrez, 300 m (no. 391 pp.); Salsipuedes, im Dzcksonia-Wald ziemlich gemein, 660 m (no. 10); Quebrada Juanango, auf Waldboden, 250 m (no. 22). Masafuera: Casas-Cafion, feuchter Boden zwischen Asplentum magellanicum (no. 265 pp.) und im hintersten zuganglichen Teil an Felsen, ca 200 m (no. 15): Quebrada del Mono, auf dem Boden einer engen Schlucht, 570 m (no. 2) und im Wald auf Erde und Steinen, 475 m (no. 7); schmales Bachbett beim Cor- respondencia-Camp, 1,100 m (no. 14). Es ist sehr auffallig, dass STEPHANI dieses wohl schoénste und grésste Le- bermoos von Juan Fernandez in seiner Publikation (19) gar nicht erwahnt, ob- wohl in diesem Material umfangreiche Anteile aus Juan Fernandez stammen. Es ist doch undenkbar, dass dieses haufige und sicher in die Augen fallende Moos dem Sammler damals entgangen sein sollte. Geographische Verbreitung: Siudchile und Juan Fernandez. 53. Sch. Skottsbergii St. Auf Fallholz im Sumpfwald. Masatierra: Anson-Tal, bei Plazoleta, auf Fallholz, 260 m (no. 282 pp.). Masafuera: Bach gerade unterhalb Correspondencia-Camp, 1,100 m (no. 9, 136 pp.). Die Art steht der Sch. stratosa am nachsten und unterscheidet sich haupt- sachlich durch den ganzrandigen Oberlappen. Nach dem Original zu urteilen, das ich gesehen habe, sind STEPHANI’s Blattmasse ubertrieben. Geographische Verbreitung: Endemisch. *54. Sch. pachyla (Tayl.) St. — Syn. Fungermannia pachyla Tay)l. J. of Bot. 1844, p. 456; Gottschea pachyla Syn. Hep., p. 621; Schzstochila crassiretis Seouppl opec. Hep: 1924, 5. 480. Nur in Hohenlagen. Masafuera: Las Torres, Bergwand, 1,370 m (no. 23, mit Lepzcolea ochro- leuca, Lepidolaena magellanica, Herberta runcinata, Harpalejeunea setifera und Tsotachts sp.). Durch Vergleichung der Originale konnte ich mich von der Identitat der Sch. crassiretis St. und Sch. pachyla iberzeugen. Es handelt sich nur um un- wesentliche, gradweise Unterschiede, die zur Trennung der beiden nicht aus- reichen. Die von STEPHANI als Sch. pachyla bestimmten Exemplare von SKOTTS- BERG, no. 148, bilden ein Mittelglied in der Reihe pachyla-crassiretis und stehen der Sch. crassiretis naher als dem Typus von Sch. pachyla (Tayl.). Diesen selbst habe ich gesehen. Fiir beide charakteristisch sind die gleichgrossen Ober- und Unterlappen, ihre scharfe Zuspitzung und tiefbuchtige Trennung; iiberdies ist AZ TH. HERZOG der Unterlappen breit zuriickgerollt und beide sind mit wenigen grossen, durch stumpfe Buchten getrennten Zahnen versehen. Die Amphigastrien sind beim Typus tiefer zerteilt und haben nur wenige Randzilien; bei Sch. crasszretis sind sie breiter und weniger tief geteilt und tragen kraftigere Zilien. In der Be- schreibung von STEPHANI werden aber gerade fir die pachyla-Amphigastrien >ubique ciliis longis remotis armata» angegeben! Geographische Verbreitung: Magellanstrasse, Westpatagonien (als Sch. crassiretis), Juan Fernandez. *35. Sch. splachnophylla (Tayl.) St. — Syn. Fungermannia splachnophyla Tayl. J, of Bot. 1884, p. 455, Gottschea splachnophyla Syn. Hep., S. 621. Nur in Hohenlagen. Masafuera: Bett eines schmalen Flusses auf dem Hochplateau nahe dem Correspondencia-Camp, ca 1,100 m (no. 1, 3, 4, 5, 8, 17, 21). Die Juan Fernandez-Pflanzen von Sch. splachnophylla haben fast immer spar- lich stumpfgezahnte Oberlappen, wahrend neuseelandisches Material vollig ganz- randig ist, und auch STEPHANI beschreibt die Blatter der Pflanze von der Isla Desolacion als ganzrandig. Bei Schzstochi/a trifft man aber solche Schwankungen in der Randgliederung haufig. Auch die Amphigastrien sind sehr wechselnd in der Form. Die Art, die nur in grésseren Hohenlagen vorkommt, scheint auf Masafuera beschrankt zu sein. Sie ist in Sidamerika zu dem magellanischen Florenelement zu rechnen (s. die Bemerkungen bei Lep7dolaena magellanica!). Viele der hier- her gehérigen Arten sind auch in Ostaustralien, Tasmanien und Neuseeland verbreitet. Geographische Verbreitung: Neuseeland, Feuerland, Magellanstrasse, Isla Desolacién, Juan Fernandez. Balantiopsis Mitt. 56. B. cancellata (Nees) St. — Syn. Prelidium cancellatum Nees, Syn. Hep, paz5u (Mies 7272): Auf Erde an Bachufern. Masafuera: Bachschlucht beim Correspondencia-Camp, 1,140 m (no. 125); m hintersten zuganglichen Teil des Casas-Cafion, 200 m (no. 132). Die Art ist urspriinglich von NEES als Pe/7dium aufgestellt worden, STE- PHANI hat sie dann mit Recht, wie aus seiner Abbildung und Beschreibung hervorgeht, zu Lalantiopsis gezogen. In der Originalbeschreibung von NEES ist iubrigens ein schwer verstandlicher Fehler eingeschlichen, indem er lobulus dor- salis und ventralis miteinander verwechselt. Die eigentiimliche Querreihung der gestreckten Blattzellen (Fig. 7, g) ist dagegen gut hervorgehoben, stellt allerdings keine Besonderheit dieser einen Art, der sie den Namen »cancel/ata» eingetragen hat, dar, sondern findet sich auch noch bei andern Balantiopsisarten, wie z. B. B. erinacea und chilensis, deren gegenseitige Abgrenzung ohne Einsichtnahme Fig. 7. Balantiopsis hians Herz. n. sp. — a Habitus einer Pfllanze mit Marsupium, 10/1; — b Stengelstiick von der Dorsalseite, 18/1; — c 2 Stengelblatter, 18/1; — d Stengelblatt mit ausgebreiteten Lappen, 18/1; — e Zellnetz in der Mitte, 130/1; — f St. amphigastrium, 18/1; — g Balantiopsts cancellata, Zelinetz, 130/1. in die Originale freilich noch nicht moglich ist. Fur 4. cancellata scheint nach Abbildung und Beschreibung von STEPHANI namentlich das Gréssenverhaltnis von Ober- und Unterlappen charakteristisch zu sein. Der Oberlappen ist nam- lich betrachtlich kleiner als der Unterlappen, seine Stellung zum Stengel mehr parallel, die Form schmaler (»quadratisch» ist ‘nicht gut ausgedriickt) als bei 2. erinacea und chilensis, bei denen er sehr breit und mehr gerundet, mit strahlig angeordneten Randdornen, ist. Auch ist die ganze Pflanze wesentlich schmach- tiger, besonders die Blattmasse kleiner. Die J. F.-Pflanze, die in der Sammlung nur sparlich vertreten ist, passt jedenfalls zur Beschreibung und Abbildung gut, nur sind die Amphigastrien etwas weniger stark gegliedert. Geographische Verbreitung: Siudchile, Juan Fernandez. 57. B. chilensis St. In feuchten Waldschluchten in grossen, oft rétlich angehauchten, flachen Rasen. Masafuera: Im innersten Casas-Cafion, Felswand nahe am Wasser (no. 134, mit Plagiochila robusta); Schlucht beim Correspondencia-Camp, 1,150 m (no. 135). Zu dieser Art rechne ich, obwohl keine vollstandige Ubereinstimmung mit der Beschreibung und Abbildung herrscht, die vorliegenden Pflanzen, die ich urspriinglich als B. erz7zacea bestimmt hatte, seit ich die angeblich zu 4. ch7lenszs St. gehdrenden Exemplare von DUSEN, no. 420 (hb. HTALMAR MOLLER) gesehen habe. In diesen ist namlich, entgegen ihrer Bestimmung als 4. chzlenszs, die echte (Ss TH. HERZOG B. erinacea zu erkennen und von ihnen unterscheiden sich die Juan Fernandez- Pflanzen starker als von der Beschreibung der B. chzlenszs. Fir B. erinacea ist namlich charakteristisch die »carina conjunctionis minima», die weniger starke Randgliederung des Unterlappens und die relativ bedeutende Grosse der Amphi- gastrien. Bei den DUSEN’schen Pflanzen sind nun Ober- und Unterlappen so tief von einander getrennt, dass es kaum moglich ist, sie zusammenhangend vom Stengel abzulésen. Der verbindende dusserst schmale Steg kann praktisch als null angesehen werden. Die Juan Fernandez-Pflanzen haben aber immer eine deutliche carina conjunctionis, wenn sie auch gelegentlich recht kurz werden kann, zeigen aber eine starkere Randgliederung und namentlich eine so weit- gehende gabelige bis fiederige Zerspaltung der Amphigastrienzipfel, dass sie eher den Amphigastrien von B. chz/enszs entsprechen. Freilich sind die Amphigastrien nie breiter als lang, was durch die Beschreibung verlangt wiirde, im einzelnen aber schwanken sie in den Massen stark. Nach den aus dem Herbar Uppsala zur Ansicht erhaltenen Belegen von &. fragzlzs St. konnte es sich bei unseren Pflanzen auch um diese Art handeln. Die Belege passen aber schlecht zur Originalbeschreibung, so dass hier keine Sicherheit zu erhalten ist. Fur B. chz- /ensts spricht mehr der stark zuriickgebogene Ventralrand des Unterlappens, der der Anheftung des Blattes eine charakteristische Note verleiht. Auf alle Falle konnte ich mich davon iiberzeugen, dass die Juan Fernandez- Pflanzen identisch sind mit den von mir friiher (8, 9) als B. cancellata bezeich- neten Exemplaren aus den Sammlungen von C. C. HossEus aus Siidchile und H. G. SCHWABE aus Westpatagonien. Diese Angaben waren also zu berich- tigen, obwohl eine genaue Artbezeichnung nach Lage der Dinge noch nicht moglich ist. Auch hier, wie in vielen Fallen, bedarf die Gruppe noch einer genauen vergleichenden Durcharbeitung. Geographische Verbreitung: Westpatagonien, Siidchile, Juan Fer- nandez. 58. B. purpurata Mitt.: Auf feuchter Erde. Masafuera: Bett eines kleinen Flusses beim Correspondencia-Camp, I,100 m (no. 123). Die Bestimmung ist in Anbetracht der Sparlichkeit des Materials unsicher. Geographische Verbreitung: Siidchile, Juan Fernandez. #50:5) Boohians: Hera naizsps (Mies 7;. Fig? 's C: Von diesem liegen einige Exemplare vor, die mit den Beschreibungen FosLiEs gute Ubereinstimmung zeigen. Die Krusten sind etwa 125—300 yp dick. Die Sporangienkonzeptakeln sind bis etwa 200 » im Durchmesser. Wie FOSLIE hervorgehoben hat, gibt es keinen bestimmten Unterschied zwischen Hypothallium 5O— 41438. The Nat. Hist. of Juan Fernandez and Easter Is!. Vel. II. 754 TORE LEVRING C5 (Sj CIS SS (3) coma) == ea= === aes? SSS S055 WO) wi Jos ON Os == =s Ss SSO ess) aj i) sc ee GCOO ase. ee SORES ae 000 eeu ORS Se I} = Sa Q is) aoe so Sses3 Q. O eae eae Seo) sace MO00 6000 Baea aaaed OO00G m eet iF Ss S Fig. 1. a—b Lithophyllum discoideum, Querschnitte. a@ mit Tetrasporangienkonzeptakeln. c L. decipiens, Querschnitt. d—e L. bexrmudense, d oberer, e unterer Teil eines Querschnittes. —aix 50; OX 325, c—e< 190. und Perithallium. Die Basalzellen sind etwas unregelmassig, 8—12 » breit und oft etwas langer als breit. Die Zellen der vertikalen Reihen sind 6—g yw breit und 0,7—I,5-mal so lang. Fertil. Sie dirfte mit 2. Vendor sehr nahe verwandt sein. PMG Ot. Sty. Geogr. Verbr.: Kalifornien; Westindien; Juan Fernandez; Subantarkt. Siidamerika. Lithophyllum fernandezianum [Lemoine 1920 5. 9. Diese Art ist fiir die Insel yon LEMOINE (lI. c.) neubeschrieben worden. Selbst habe ich das Material nicht gesehen. Wo es z. Z. steckt, ist nicht bekannt. Fundorts St. 30:2. Geogr. Verbr.: Juan Fernandez. Lithophyllum discoideum Foslie 1900 S. 73; 1906 S. 22; 1908 S. 10 Vaf. 2 Fig. 2—4; Pseudolithophyllum discoideum Lemoine 1913 S. 46. — Fig. 1 A, B. Von dieser liegen einige gut entwickelte Exemplare vor, die mit der Be- schreibung FOSLIES gut iibereinstimmen. Die Krusten sind bis einige mm dick und sind an Steinen befestigt. Das Hypothallium ist verhaltnismassig schwach DIE CORALLINACEEN DER JUAN FERNANDEZ-INSELN 755 Fig. 2. Lithophyllum pustulatum. a Querschnitt durch ein mannliches Konzeptakel, 6 durch ein steriles Thallusstiick, ¢ durch ein weibliches Konzeptakel mit Gonimoblast, @ durch ein Sporangienkonzeptakel. — a X 375, 6—d X 220. entwickelt. Dagegen ist das Perithallium sehr kraftig. Die Zellen der letzteren sind etwa 6—8 » dick und etwa (1—)I,5—2-mal so lang. Die Sporangienkon- zeptakeln sind etwa 200—250(—300) » im Durchmesser. Mit Tetrasporen. Fandort: St. 8. Geogr. Verbr.: Juan Fernandez; Subantarkt. Sidamerika; Falkland-Inseln; Sud-Georgien. Lithophyllum pustulatum (Lamour.) Foslie 1904 S. 3; 1905 S. 117. — Fig. 2. Es liegt ein ziemlich reiches Material vor, das mit dieser weitverbreiteten Art gut ibereinstimmt. Sie bildet auf Pada bis etwa I cm grosse, zusammen- 750 TORE LEVRING fliessende Krusten, die bis 200 » dick sind. In ihrer grdéssten Ausdehnung sind sie 2(—3)-schichtig, nur in der Nahe der Konzeptakeln mehrschichtig. Die Entwicklung der Fortpflanzungsorgane (Fig. 2) scheint mit der von ZL. eapansum (vel. SUNESON 1937 S. 19) tibereinzustimmen. Dagegen diirfen die Bilder NICHOLS (1909) kaum richtig sein. Das von LEMOINE 1920, S. 10, angefthrte LZ. (Vermatolithon) spec., das mit L. pustulatum verglichen wird, stellt wahrscheinlich diese Art dar. FPuncdor tes St. o3) yr Geosr, Verb. Atlant: Mittelmeer Stiller Ozean: Lithophyllum bermudense Foslie et Howe 1906 S. 132. — Fig. 1 D, E. Die Ubereinstimmung mit dieser Art, die bisher nur fir Bermuda und Florida angegeben ist, scheint sehr gut zu sein. Die Krusten sind warzig und bis ein paar mm dick. Das Hypothallium besteht aus einer einzigen Zellreihe, deren Zellen 9—15 breit, 25—-50 ». lang sind. In dem Perithallium sind die Zellen (6—)g—15(—18) » breit, 15—-45 » lang und in deutlichen Langs- und Querreihen angeordnet. Die Poren treten auch sehr deutlich hervor. Fumdomies St. (4,13: Geogr. Verbr.: Bermuda; Florida; Juan Fernandez. Corallina L. Corallina chilensis Decaisne in Harvey 1847 5S. 103; Areschoug in J. Agardh 1852 S. 565; Manza 1940 S. 276; C. officinalis var. chilensis Kitz. 1858 Tat 66; -Yendo’19019S. 718 Taf. 54 Fig: 1 und Tafi56 Fies15; 14037 sme: Diese Form steht C. officinalis sehr nahe. Von YENDO (I. c.) wird sie auch zu dieser letzten Art gefihrt, er meint aber, sie sei doch etwas verschieden. Ich folge hier MANZA (I. c.) und fihre sie als selbstandige Art an. Die untersuchten Exemplare sind bis etwa 4 cm hoch und stimmen gut mit der Beschreibung und den Figuren dieser Art uberein. Fertil. Eiicorte:. its 26 by Geogr. Verbr.: Westktiste von Siidamerika; Juan Fernandez; Japan. Corallina Cuvieri Lamour.; Areschoug in J. Agardh 1852 S. 572; Yendo 1905 55 35; Manza. 1940 S. 270. f. Cuvzert Aresch. |. c.; Kitz. 1858 Taf. 70a—g. haspaig A“eschia linc; Kitz. els.c. atyyia-—b. fasuoulara vAcesch. Wlics ekiutzey. G Latss70 ha: Die angefiihrten Formen gehen in einander tber. Wahrscheinlich gehoren mehrere der Arten, die z. B. KtU1rzinG (I. c.) von den australischen Gewassern abgebildet hat, auch zu dieser Art. Es liegt ein reiches Material vor, das alle Ubergange zwischen den ver- schiedenen Formen zeigt. Von der normalen f. Cauvzerz scheinen die anderen nur vom Standort beeinflusste Formen zu sein. Fertil. Fundiontre wSthelas— bale .163; ig Gib Bo bidge: Geogr. Verbr.: Australien; Juan Fernandez. DIE CORALLINACEEN DER JUAN FERNANDEZ-INSELN fli Literaturverzeichnis. AGARDH, J. G., Species, genera et ordines algarum, Vol. II]: 2. Lund 1852. Fosuiz, M., On some Lithothamnia. — Kgl. Norske Vidensk. Selsk. Skr. Trondhjem 1897. —— Calcareous algae from Fuegia. — Svenska exped. till Magellanslinderna, Bd 3, Stockholm 1goo. —— Algologiske Notiser. — Kgl. Norske Vidensk. Selsk. Skr. ‘Trondhjem 1904. —— Remarks on northern Lithothamnia. —— Ibidem 1go05. —— Algologiske Notiser II. — Ibidem 1906. —— Antarctic and subantarctic Corallinaceae. — Wiss. Erg. d. schwed. Siidpolarexp. 1g01—o3, Bd. IV: 1, Stockholm 1908. —— Contributions to a monograph of the Lithothamnia. — After the author’s death collected and edited by HeENRIkK Printz. Trondhjem 1929. Fosuiz, M., and Howe, M. A., New American coralline algae. —— Bull. New York Bot. Garden, Vol. 4, New York 1906. Harvey, W. H., Nereis Australis. London 1847. Ktrzinc, F. T., Tabulae phycologicae, Bd. 8, Nordhausen 1858. LeMoINE, Paut, Mélobesiées. Révision des Mélobesiées Antarctiques. —— Deuxieme Expédition Antarctique Francaise tg0$8—r10. Paris 1913. —— Les Mélobesiées. — Bot. Erg. d. schwed. Exp. nach Patagonien und dem Feuerlande 1907—o9, VII. — K. Svenska Vet.-akad. Handl., Bd 61, Stock- holm 1920. Levrinc, T., Die Meeresalgen der Juan Fernandez-Inseln. — The Natural History of Juan Fernandez and Easter Island, ed. by C. Skottsberg, Vol. Hl. Uppsala 1941. Manza, A. V., A revision of the genera of Articulated Corallines. — Philippine Journ. Sc., Vol. 71, Manila 1940. SETCHELL, W. A., and Garpner, N. L., Marine Algae of the Revillagigedo Islands Expedition in 1925. — Proc. California acad. sc., Ser. 4, Vol. 19, San Fransisco 1930. Yenpo, K., Corallinae verae of Port Renfrew. — Minnesota bot. stud., Vol. 2, Minneapolis rgor. —— A revised list of Corallinae. — Journ. Coll. Sc. Imp. Univ. Tokyo, Vol. 20, Tokyo 1905. Printed "7's 1943. a ia aay Pa Ty sn bind eae sana die we wor We tek init , ibyer vi ioe el a ; we seer Slee ae) Inouwe ph ikebo snot eli 2 Livieoyet ine |e + sotieink init der voO fi re . é ‘ beecein mike Wimen Longe gin Hitt fi ‘lida ; we ’ ed i SBithe = : oo i ny M sialen ae vatatl oka Sya rie Iai, ". 7 Jorshet wihischeintich (tee Ate ae ve din Tor kl - f petienl 2ectbid- Mi RIsney (Abr og af ae cee Sen > oth lal nltnndioliol domed aCe “i L | derreterei barrel | opoivir ee taney ed odes ive Beary wnt Baga te m be 4 fina ky rot eiitbner |) sie aeior dyeotr/ ole flout, azo" , ets : inet mitreinetion eh erorl Prae uc 2a v oA bid - “PP ea” Soe Aber ToA Se Bawls 16 let eka Reve Men Catat il Mair ea tow raynled iors, 1: Ab hib het lipol,2sorktirn sll, vail) REAR FO fh Ly ARLE ON OF ei brat wo i Me) od Ye Ftd fy] aH Bow RL war’ se lav, . noboit Eau Ged: iss het “° ; RH ints At, HH . (Uaitt. btdb-ao 4 saleqaigs cA Jivy @o+++s0. Shoat ra . es : Rare) bys 10 TT He SD a ES bha gee “Melt 4a ead heaoih Ane ae KOOP Wea 7). Vt rHeie 3 wd ee a Set * ite’ Matyi agane fata filets eee nie : WAY The deers ahi es abiaer st Peay ieee ARTA oyl Gord O aoe abopiat “tdhbphas iva st GAL te gaat Ao eis ck A alban): bam hie “I 7° Se, at ow “bestia ork > phot ‘nts ott m. ee ope! Oa ‘ ® < fete © iipie Aa nGeanniteS Wales HUT 46. RP) SERRA vale é 4 ee” ; - Fant erlocueyngely ae See “ot eae sitinun LV S iyheth WAT lly) eee se athe id tar Malvas at QOS ss. X oes. Mai Me 1046 >. 270). , ; .enot oglolly a iS EP al Spi a ae) *: a / ies 0.14. Fabs 9 gr angie aes) bo i c : are i. 4 am: 2 j ; FE eedadeta Mapsels. 1. ¢.; Ailts, hc. kalo Jobe pee angetobpeay J Mansirets ech ame itt eliander ober) Vi shiseh ain © i" be Aen: abie A Rowe eat, 3) eon. dow —e 1" at full, ald e TbesenArh pelil Gina a le Me Bevaiey f ee rn etl Fergie e ¢ > ee 52 26. Einige Corallinaceen von der Oster-Insel. Von TORE. LEVKING, In der Arbeit von BORGESEN (1924) itiber die Meeresalgen von der Oster- Insel sind die Corallinaceen von Mme PAUL LEMOINE bearbeitet worden. In der von ihr nicht bearbeiteten Corallinaceen-sammlung von den Juan Fernandez- Inseln (vgl. LEVRING 1943 S. 753), fand ich aber auch unbestimmtes Material von der Oster-Insel. Ausser’ Lithothamnion siamense, Gontolithon myriocarpon und Porolithon onkodes gibt es noch drei Arten, die in dem Verzeichnis BORGESENS nicht auf- genommen worden sind. Lithothamnion mesomorphum Foslie Igo! S. 5. var. ornatum Foslie et Howe 1906 S. 129 Taf. 80 Fig. 2, Taf. go Fig. 2; Pemome in Borgesen 19o15—20 S. 155. Sowohl die Hauptform als var. ovzazum ist von West Indien bekannt. Die Ubereinstimmung des Oster-Inselmaterials mit var. ovvadum ist so gut, dass ihre Identitat nicht bezweifelt werden kann. Der Thallus ist etwa 150—300 » dick, nimmt also in dieser Hinsicht etwa eine Zwischenstellung zwischen den Angaben FOSLIES und HOwWES (I. c.) und denen von LEMOINE (I. c.) ein. Das Hypothallium ist ziemlich locker gebaut und die Zellen etwa 7—12 yp breit und 15—25(—30) » lang. Das Perithallium ist fast nur auf der Oberseite des Thallus entwickelt und die Zellen abgerundet oder etwas in der Lange gezogen, etwa 4—6 y breit und 5—9 p» lang. Die Konzeptakeln sind etwa halbkugelf6rmig, 300—500 », im Durchmesser, zerstreut oder zu Gruppen vereinigt. Pumdort: Hangs Piko (litoral, */c 1917). Geogr. Verbr.: Westindien; Florida; Osterinsel. Porolithon craspedium Foslie 1909 S. 43; Lemoine 1911 S. 163; Létho- phyllum craspedium Foslie 1900 S. 26. Von dieser Art liegen mehrere wohlentwickelte Exemplare vor. Sowohl in habitueller als in anatomischer Hinsicht scheint sie ganz typisch entwickelt zu sein. Fundort: Hanga Piko (litoral, 76 1917). Geogr. Verbr.: Indischer Ozean; Stiller Ozean (Gilbert-Inseln, Funafuti, Oster-Insel). j 760 TORE LEVRING Jania rubens Lamour.; Manza 1940 S. 270. Nach MANZA dirfte die Mehrzahl der Fanza-Arten, die friiher beschrieben worden sind, unter ¥. vwbexs zusammengefihrt werden konnen. Das untersuchte © Material lasst sich zweifelsohne auch gut in diese Formenkreis einreihen. Die Exemplare sind bis ein paar cm hoch und mit anderen Algen mehr oder weniger zusammengeflochten. Fertil. Von BORGESEN (1924 S. 292) wird ¥. ¢enella Kitz. angegeben. Wahr- scheinlich ist sie mit der von mir untersuchte Pflanze identisch. Fundort: Hanga Piko (litoral, "5/6 1917). Geogr. Verbr.: ‘Weitverbreitet. Literaturverzeichnis. BorGEseN, F., The marine algae of the Danish West Indies, Vol. Il. Kobenhavyn LO 5=—20) -——- Marine algae from Easter Island. ——- The Natural History of Juan Fernandez and Easter Island, ed. by C. Skorrsperc, Vol. II. Uppsala 1924. Fostiz, M., New or critical calcareous algae. — Kgl. Norske Vidensk. Selsk. Skr. 1899. ‘Trondhjem rgoo. == New Melobesiae. — lbidem;nopo;ifager; —— Algologiske Notiser VI. — Ibidem 1909, 1909. Fosiiz, M., and Howe, M. A., New American Coralline Algae. —- Bull. New York Bot. Garden, Vol. 4, New York 1906. LEMOINE, PAUL, Structure anatornique des Mélobesiées. Application a la classification. — Ann. l'Institut Océanogr. Monaco, T. 2, Paris 1g1t. LEvRING, T., Die Corallinaceen der Juan Fernandez-Inseln. —- The Natural History of Juan Fernandez and Easter Island, ed. by C. Skorrsperc, Vol. Il. Uppsala 1943. Manza, A. V., A revision of the genera of Articulated Corallines. — Philippine journ. sc., Vol. 71, Manila 1940. Printed "is 1043. 27. Additional Remarks to »Marine Algal Communities of the Juan Fernandez Islands» (Nr. 23). By CesKODTSBERG, In my paper on the communities formed by Marine Algae in the Juan Fernandez Islands the Coralline algae were quoted as Corvadlina sp., Lithotham- nion (or Lithophyllum), and Lithophyllum fernandezianum. The collection was in the hands of Mme Paul LEMOINE (comp. LEVRING, Nr. 25 of this publica- tion, p. 753) and my efforts to get into contact with her had proved fruitless. Finally, the Foreign Office in Stockholm kindly undertook, through the Swedish Consulate General in Paris, to make a search for the collection which, in spite of the absence of Mme LEMOINE, who is reported to have joined the fugitives from the capital, was found in the Laboratoire de Cryptogamie, taken care of and returned to me in perfect condition. I am greatly indebted to Mr. S. GRAFSTROM of the Foreign Office, and to Consul General R. NORDLING in Paris for their kind assistance. The collection was handed over to Dr. T. LEVRING, who promptly under- took to identify the species, thus enabling me to make the following important additions and corrections to my paper. P. 673. St. 36 b, Lower littoral: Coradllina chilensis and Cuvierz. — St. 39a 2, and St. 2, in pools: Corallina Cuviert; Lithophylum pustulatum (on Padina). P. 674. St.39b and St. 3, Lower littoral: Corallina Curvteri and Lithophyllum pustulatum (on Padina). P. 675. St. 8. Bottom layer, on stones: Lzthophylum Ass. of L. bermudense (small, warty crusts) and L. discoideum (very large, smooth clumps). — St. 4. Lithothamnium Ass. of L. validum and Lithophyllum bermudense. — St. 13, Sublittoral: Corallina Cuvier?. — St. 6, Sublittoral: Lzthothamnium validum. P. 676. St. 5. Ass. of Lithothamnion validum. — St. 11. Lithophyllum pustulatum (on Padina). — St. 7. On a large stone, thin crusts of Lzthophyllum decipiens. P. 680. Associations and Formations. The leading species of crustaceous Corallines seem to be Lithothamnium validum and Lithophyllum discoideum, which form a deep-water association on stones, pebbles and old shells which become thickly encrusted. With these is also Lethophyllum bermu- dense. The other association of the »Formation of Crustaceous Corallines» is 5O+— 41438. 762 C. SKOTTSBERG developed as a thin cover on bedrock in pools and below low-water mark in the surf zone. It is very difficult to get samples of this crust. Probably Lztho- phyllum decipiens is one of the leading species; whether, as I supposed, L. fer- nandesianum is another (comp. LEVRING l. c. p. 754), seems uncertain. P. 681 etc., Geographical distribution. The total number of species, Corallines included (with exclusion of the doubtful Lzthophyllum fernandestanum) is 100. The following additions should be made to the geographical groups; the figures have been corrected accordingly. Group I. 20 sp.= 20%. — Group 2. 14 sp.=—14%. @, 7 Spe 6, add Lithophyllum pustulatum; 7 sp.= 7 %. — Group 3. Add Corallina chi- Jensis (also reported from Japan); 8 sp. = 8 %. — Group 4. Add Lithophyllum decipiens (also West Indies) and Lzthothamnion validum; 9 sp. = 9 %. — Group 5. 10 sp.=I10 %. a, 3 sp.=3 %; 6, add Corallina Cuvieri; 5 sp. = 5 %. — Group 6. Add Lithophyllum bermudense; 2 sp.= 2 %. — Group 7. Add Litho- phyllum discoideum; 5 sp.=5 %. The increase in groups 5 (West Pacific) and 7 (Subantarctic) is of interest. Printed ™/5 1943. THE NATURAL HISTORY OF JUAN FERNANDEZ AND EASTER’ ISLAND ELLE DEY DR CARL SKOFFSBELAG VOL. I BOTANY PART VI 28. C. SKOTTSBERG: A Supplement to the Pteridophytes ae Phan- erogams of Juan Fernandez and Easter Island. 29. C. SKOTTSBERG: The Vegetation of the Juan Fernandez Islands. UPPSALAV Tass ALMQVIST & WIKSELLS BOKTRYCKERI AB 28. A Supplement to the Pteridophytes and Phanerogams of Juan Fernandez and Easter Island (Nos. 1, 2, 5 and 7). By C. SKOTTSBERG. With Plates 55—57 and 1 Text figure. In order to bring the taxonomical papers on the Ferns (nos. 1 and 2) and Flowering Plants (nos. 5 and 7), published in this volume in 1920 and 1922, up to date, the following notes were compiled. Some collections have been made in Juan Fernandez after our visit which I have been allowed to study and which gave occasion to a number of small papers. In the New York Herbarium I was able to examine the plants collected by J. CHAPIN in Easter Island in 1935. The species brought by the Franco-Belgian Mission to that place in 1934 have been quoted, where there was reason to do so (''M. Fr.-B.’), from the papers by A. GUILLAUMIN, and others. In some instances I have found Additions and Corrections to Vol. LPP iG. P. 790 add after Coriandrum, 76. Sanicula crassicaulis Poepp. (Syn. S. liberta Cham. et Schlechtend. quoad plantam chilensem). - 794 line ro from foot, read 1741. - 796 line 3 from top, read forest. - 831 under Drimys add: Rev. G. Kuscuet informed me that according to the in- habitants there is a marked difference in the quality of the wood between Masa- tierran and Masafueran specimens of the canelo. P. 889, Explanation of Fig. 32, read Photo 17/4 1917. The photograph taken by Dr. BAckKstrom will be published in Vol. I, no. 4. P. 931, Explanation of Fig. 41, read Pasto. P. 955, Explanation of Plate 79:1, read Photo 29.8. 1908. he] ae (Baye Wy 28. A Supplement to the Pteridophytes and Phanerogams of Juan Fernandez and Easter Island (Nos. 1, 2, 5 and 7). By C_ SS OTTSBERG. With Plates 55—57 and 1 Text figure. In order to bring the taxonomical papers on the Ferns (nos. I and 2) and Flowering Plants (nos. 5 and 7), published in this volume in 1920 and 1922, up to date, the following notes were compiled. Some collections have been made in Juan Fernandez after our visit which I have been allowed to study and which gave occasion to a number of small papers. In the New York Herbarium I was able to examine the plants collected by J. CHAPIN in Easter Island in 1935. The species brought by the Franco-Belgian Mission to that place in 1934 have been quoted, where there was reason to do so ('M. Fr.-B.”), from the papers by A. GUILLAUMIN, and others. In some instances I have found it necessary to take up again questions of a taxonomical nature, which eventu- ally resulted in a change of name or position, or both. Species already enumerated in this volume, under the same or under another name, are cited under their old number. New additions to the flora are marked by an asterisk. 1. C. CHRISTENSEN and C. SKOTTSBERG. The Pteridophyta of the Juan Fernandez Islands. 11. Hymenophyllum rugosum C. Chr. & Skottsb. 1.13. Add Masafuera: above Q. del Varadero, c. 950m. 14. Hymenophyllum plicatum Kaulf. 1824. H. dzchotomum, C. CHR. & SKOTTSB. I. 14 non Cav. — H. dichotomum Cav. is probably a Philippine species, quite distinct from the Chilean one (CHRISTENSEN, Ind. Fil. suppl. III). Add Masafuera: Q. Loberia in forest, 280 m. Dicksonia externa nov. spec. . Bertertana (Colla) Hook. var. virgata C. Chr. & Skottsb. 1.17. The differences between D. Berteriana of Masatierra and var. virgata of Masafuera are so conspicuous that it seems more correct to regard them as specifically distinct. 51—516710. The Nat. Hist. of Juan Fernandez and Easter Isl. Vol. 11. 764 C. SKOTTSBERG 21. Polystichum Berterianum (Colla) C. Chr.; C. CHr. & SKOTTSB. 1. 20. = Add Masafuera: Las Torres, 1 370 m. 24. Asplenium obliquum Forst. var. chondrophyllum (Bert. ex Colla) Mett.; CG. (Cur. & SKOTTSB:, 1.22, SKOTISB.- $2361 -«- EGOSER 32-25 2 Add Masatierra: Pangal, in a cave 50m s.m., BEHM in Herb. Bock! — Masafuera: ©. del Mono, 15. m s. m., WAGENKNECHT. 25. Asplenium macrosorum Bert. ex Colla; C. CHR. & SKOTTSB. 1. 22. Add Masafuera: SE. slope of Los Inocentes, 950 m. 27. Asplenium dareodies Desv. 1811. A. magellanicum Kaulf. 1824; C. CHR. & SKOTTSB. 1.24. One of these unwelcome name-changes, the species, a common fern of south Chile and Fuegia, having been known under the younger name for more than one hundred years, when WEATHERBY dug up the older name, which had never been used. 28. Blechnum auriculatum Cav. C. CHR. & SKOTTSB. 1.25. After a renewed examination of the large material in Herb. Stockholm I find it impossible to segregate any good varieties; comp. LOOSER g. With regard to remotum, we pointed out this l. c., and neither has farvulum, mentioned as “forma parvula’, any taxonomic standing. J have not seen specimens of B#. zz bulum Presl (var. t7bulum Capurro). 29. Blechnum valdiviense C. Chr.; C. CHR. & SKOTTSB. 1.26. B. blechnoides (Bory) Keyserling; LOOSER 9g. 54. A troublesome case of nomenclature. The oldest name is Lomarza blechnoides Bory 1828 (Chile, Concepcion), but this combination is antedated by L. dlechnordes Desv. 1827, quite another species, doubtfully indentified with L. aurzculatum by CHRISTENSEN. In 1873, KEYSERLING made the new combination Blechnum blech- noides (Bory) Keyserl., but this name should not be applied to BoRY’s plant which, as testified by LOOSER I. c., is very different from the plant KEYSERLING had before him; I have not seen KEYSERLING’s paper, of which there doesn’t seem to be a copy in any Scandinavian library. If, according to the Rules, 4. d/echnoides Keyserl. is the valid name for L. dlechnoides Bory, there is something wrong with these rules. To make the case more interesting, there is another B. dlechnoides (Lag. ex. Sw.) C. Chr., based on Asplenium blechnoides Lag. ex. Sw. 1806. SWARTZ, in 1810, called it B. unilaterale, but CHRISTENSEN, regardless of the fact that the combination 4. dlech- noides already existed, used it for LAGASCA’s Asplenzum and called BORY’s Lomaria B. valdiviense. Would it not be possible to convince those who forced these rules upon us that Blechnum blechnoides is a splendid example of a confusing name? LOOSER regards the Juan Fernandez plant as an endemic variety, var. Jernandeztanum, stated to be more delicate than the continental form and to have a poorly developed endlobe. This may be true in some instances, but some of my specimens are so like coast forms that I find it difficult or im- possible to separate them. A SUPPLEMENT TO THE PTERIDOPHYTES AND PHANEROGAMS OF JUAN FERNANDEZ 765 38. Pteris semiadnata Phil. C. CHR. & SKOTTSB. 1. 34. Add Masatierra: summit quebrada of El Yunque, TENZ no. 2 (Herb. Ham- burg; Skottsb. 5. 158). 39. Pteris Berteroana Ag. C. CHR. & SKOTTSB. I. 34. Add Masafuera: ©. Angosta, in the gorge. 42. Polypodium intermedium Colla; C. CHR. & SKOTTSB. 1. 37. I. typicum var. basicompositum n. var. A P. intermedio typico differt pinnis basalibus pinnatis. Masatierra: NE. slope of El Yunque (no. 154 B). Accidentally omitted 1. c. — No. 318, mentioned p. 38 as an aberrant rock form corresponding to var. basicompositum of subsp. masafueranum is better referred to this in spite of the locality in Masatierra, as already suggested l.c. 43. Polypodium Masafuerae Phil. Linnaea 29 (1857) 107. P. pychnocarpum C. Chr., C. CHR. & SKOTTSB. 1.41. See I. M. JOHNSTON in Contrib. Gray Herb. 85 (1929) 14 and LOOSER 6, 7. (BAKER, On the Seychelles Fern Flora (Trans. Irish Acad. 25, 1875, 516) cites Polypodium serrulatum (Sw.) Mett. from Juan Fernandez, a statement repeated by CHRISTENSEN and also by DIELS in Natiirl. Pflznfam. CHRISTENSEN must have forgotten this reference when we compiled our joint paper, and the species is not mentioned by JOHOW or LOOSER. MAXON l.c. showed that the name is illegitimate and replaced it by P. duale Maxon. This, the true Acrostichum serrulatum Sw., is widely distributed in tropical America and also known from West Africa, Madagascar and Mauritius. Its occurrence in the Seychelles is confirmed by typical specimens collected by Dr. JOHN ERIKSSON during the Swedish Deep Sea Expedition 1948 (Herb. Gotob.). On my request Mr. F. BAL- LARD went over the material at Kew and found a specimen mounted with several others from various places and labelled in W. J. HOOKER’s hand “Juan Fernandez, Capt. Woodd”’. This name is not in the list of collectors in Juan Fernandez published by JOHOW, but we can hardly expect this list to be complete. Mr. BALLARD, in arranging that I received the sheet on loan, expressed his serious doubt that the specimen said to have come from Juan Fernandez really belongs to P. duale, while it bears great resemblance to P. myosuroides Sw., and to judge from what I have seen of this (good material in Herb. Stockholm) I think that it belongs to this species, whereas other specimens on the Kew sheet clearly represent duale. MAXON credits myosuroides to Jamaica only, and its occurrence in Juan Fernandez is surprising. Mixed with the Woopp plant are a few Bryo- phytes. Thinking that they could perhaps give a clue to the origin of the fern, I asked Dr. TH. HERzOG’s opinion; I thank him for liberal assistance. The sample contained a species of Heréerta, not identical with the species known from Juan Fernandez, but belonging to a tropical group, and a Dicranacea of the genus Holodontium; this latter moss is, according to Dr. HERMAN PERSSON, quite different from all Dicranaceae found in either Juan Fernandez or the W est 766 C. SKOTTSBERG Indies. The genus is circumpolar-subantarctic. Under these circumstances I have thought it better not to include ?. duale among the ferns known with certainty to occur in our islands.| *Polypodium trichomanoides Sw. Cited for Juan Fernandez in HOOKER & BAKER, Syn. Fil. 326 but not by CHRISTENSEN, nor by LOOSER. Also in this case Mr. BALLARD kindly offered his assistance, and I obtained as loan a sheet from the Kew Herbarium containing the following lots: South America (PEARCE), Brit. Guiana (IM THURN), Brazil (BLANCHET), and Juan Fernandez, DOUGLAS (4 pieces), labelled by W. J. HOOKER. All are typical ¢rzchomanoides. DOUGLAS collected in Masatierra, and we have to take the statement for good, even if this little fern has escaped all other collectors. Dr. HERZOG recognized the fragments of two Hepaticae stuck in the tuft of the fern as two species of Plagzochila, probably filipendula St. and Lechlert St. Neither is registered by HERZOG in his treatment of the Hepaticae in this volume, but they, occur in Chile and West Patagonia and may very well await their discovery in the islands. Area of distribution: Tropical America, Juan Fernandez. 49. Ophioglossum fernandezianum C. Chr.; C. CHR. & SKOTTSB. 1. 44. For affinities, see CLAUSEN l.c. 153, 154. Lycopodium L. In his strange Handbook of the Lycopods, NESSEL 1. 80 gives the following information on their occurrence in Juan Fernandez: “Die Insel Juan Fernandez’’ — this name is sometimes used instead of Masatierra, but Lycopods have been found only in Masafuera — “‘weist nur 7 Arten auf, wovon 2 endemisch sind: “U. saururus; — L. scariosum, Gayanum, Skottsbergii, clavatum, magellanicum, Cunningham’. In his enumeration of the species, Urostachys saururus is quoted for “Juan Fernandez, Cumberland Bay, Herb. Stockholm!”, Lycopodium scariosum is not credited to these islands, L. Gayanum is called L. scartosum var. Gayanum, and L. Skottsbergii appears as L. scariosum var. Gayanum {. Skottsbergn. L. clavatum is not cited, but, quite correctly, ZL. magellanicum (our spm 482), and, besides, a var. Cunninghami, HOCHSTEDT 1/4 1913. L. scartosum var. Gayanum f. Skottsbergat was described in NESSEL 2.172. The specimen of L. saururus referred to by NESSEL as being in the Stockholm herbarium was searched for in vain. No so-called Uvostachys is credited to Juan Fernandez by HERTER. 50. L. magellanicum Sw. C. CHR. & SKOYTTSB. I. 45. The form from Masafuera is very robust, larger than any we had seen from the South, and possibly approaches BAKER’s var. Cunninghamit, BAKER 2. 25. There are many habitat forms of L. magellanicum, but only one form in Masa- fuera, and as long as nothing is known of its alleged taxonomic value, I cannot find any good reason to give it a special name. A SUPPLEMENT TO THE PTERIDOPHYTES AND PHANEROGAMS OF JUAN FERNANDEZ 767 51. L. scariosum Forst. -L.: Gayanum Remy, C. CHR. & SKOTTSB. I. 45. Var. Gayanum f. Skottsbergia Herter ex Nessel was based on my material from 1917, while the winter material, collected in the same place in 1908, was called v. Gayanum. In NESSEL 1. 345, nos. 381, 416, 417 and 534 were cited for f. Skottsbergiz, in 2.172 no. 534 is not mentioned. Where he got nos. 381 and 417 I cannot understand, because 381 in my collection (and the numbers are of course the same in all herbaria) is Serpyl/opsis caespitosa and no. 417 Evigeron rupicola. In 1949 HERTER (1. 41) raised f. Skottsbergz to specific rank. Strangely enough, in his paper 1950 (2.95) L. Skottsbergii from J. Fernandez figures under Ser. Magellanica, where it does not belong. In Herb. Stockholm are three sheets of a stunted form of ZL. magellanicum, collected by me in South Georgia, 1902, and labelled LZ. Skottsbergii Herter by HERTER himself. This was another MSS name, published without description in a paper by ROIVAINEN (Anal. Vanamo 6, 1936); in NESSEL 1 (1936) 316 the same specimens were called ZL. magellanicum var. pusillum Werter, and a description of this variety was published in NESSEL 2 (1940) 170. It was raised to specific rank under the name of ZL. Uvani Herter in 1949 (1.42). L. Skottsbergit Herter 1949 (J. Fernandez), not 1936 (South Georgia) must have been classified with Magellanica by an unintentional mistake. BAKER united Gayanum of Chile with scarzosum of New Zealand, regarding the former as intermediate between scarzosum genuinum and var. Fussieuz (Desv.) Baker 2.29 from the tropical Andes and Brazil. HERTER treats all as separate species. Even with regard to the large-leaved Fusszeuz with its long peduncle carrying from I to 5 spikes, the right to specific rank could be disputed. When working on my collection CHRISTENSEN left Z. Gayanum standing, and I had no objection, but I believe now that BAKER’s attitude is more correct. I have compared Gayanum and scartosum without being able to discover any tangible differences. Measurements of numerous distichous leaves from several individuals andestaken from! branches: of ‘the Same) type) gave “4.5—-6"XK 13-3) mm” for scariosum and 5—6 X 1.1—1.4 mm for Gayanum. The adnate basal portion, 2.5—3.5 mm long, is alike in both; the same is true of the linear-filiform leaves and of the sporophylls and spores. The sporophylls are 3.5;—4 mm long and I.5—2 mm wide. NESSEL calls the sporophylls “bewimpert” in both var. Gayanum and f. Skottsbergi. “Bewimpert’’ would mean ciliate along the edge. The scariose margin is more or less uneven, but not ciliate. L. scartosum forms another connecting link between South America and New Zealand. 2. C. CHRISTENSEN and C. SKOTTSBERG: The Ferns of Easter Island. 1. Asplenium adiantoides (L.) C. Chr. var. squamulosum C. Chr. & Skottsb. 2A Add Orongo, in a ruined stone hut, CHAPIN no. 1033! 768 C. SKOTTSBERG 5. Dryopteris gongylodes (Schkuhr) OK. C. CHR. & SKOTTSB. 2. 50. SCHKUHR’s spelling of the specific name was “goggilodus” and “goggylodus’”, corrected to gongylodes by LINK. Recently FOSBERG (I. c.) claimed that, according to the Rules, we must return to the original spelling; he made the new combina- tion goggilodus (Schkuhr) Fosberg, in spite of the undeniable fact that “goggilo- dus’? means nothing, whereas gougylodes expresses what SCHKUHR undoubtedly wanted to express. FOSBERG admits that the former version is an orthographic error but sees no reason to believe that it was unintentional. The conclusion would be that SCHKUHR deliberately coined a senseless epithet: thus, the error was intentional, for the author must have known how a Greek word ought to be latinized. In my opinion it is a clear example of an unintentional error. 6. Dryopteris dentata (Forsk.) C. Chr. D. parasitica (L.) OK., C. CHR. & SKOTTSB. 2. 50. Add Orongo, in a ruined stone-hut, CHAPIN no. 1032! *Davallia solida (Forst.) Sw. Rano Kao (M. Fr. -B.) Area of distribution: Malaya, Queensland, Polynesia. 7. Elaphoglossum Skottsbergii Krajina. /. tahitense, C. CHR. & SKOTTSB. 2.50 non Brack. According to KRAJINA l.c. 61 the Easter Island species differs materially from the Tahitian /. tahztense, of which he had seen the type, but the distin- guishing characteristics are not indicated. Area of distribution: Endemic. *Polystichum Fuentesii Espinosa |. c. 153. Along the road between Hanga Roa and Rano Raraku, FUENTES I9QI1I. This is the fern named P. aculeatum (L.) Schott by ESPINOSA in 1913, excluded from our list as a doubtful record. Its position in the aculeatum assem- blage remains to be investigated. Area of distribution: Endemic. 8. Microlepia strigosa (Thunb.) Presl. C. CHR. & SKOTTSB. 2. 51. Add Tongariki, CHAPIN no. 1050! 9. Polypodium scolopendria Burm. f. P. phymatodes, C. CHR. & SKOTTSB. Dale (\Polypodium fossum (Moore) Baker was credited to Easter Island by CHRIS- TENSEN, Ind. Fil. (1906) 528, as an endemic species. He must have forgotten all about it when we prepared our list. Recently Mr. LOOSER was able to show (10) that the locality “Easter Island’ is false, the original statement “‘Eastern Archi- pelago”’ having been interpreted wrongly. There can be little doubt that the species came from some part of Indonesia.| A SUPPLEMENT TO THE PTERIDOPHYTES AND PHANEROGAMS OF JUAN FERNANDEZ 769 12. Ophioglossum coriaceum A. Cunn. C. CHR. & SKOTTSB 2. 52. Referred to O. lusitanicum LL. as a subspecies by Clausen |. c. 161. Add to the Area of Distribution: N. Chile. New Caledonia. *Psilotum nudum (L.) Griseb. Rano Aroi (M. Fr.-B., as P. triquetrum Sw.). Area of distribution: Trop. and subtrop. regions. Its occurence in remote Easter Island is of interest. 5. CARL SKOTTSBERG: The Phanerogams of Easter Island. Lsopecies undoubtedly or probably indigencus. 3. Axonopus paschalis Pilger. SKOTTSB. 2. 63. Also collected by CHAPIN (no. 1002!) in Rano Kao. 6. Agrostis avenacea J. F. Gmel. A. filiformis (Forst.) Spreng., SKOTTSB. 2. 65. 11. Cyperus polystachyos Rottb. SKOTTSB. 2. 60. Found also by CHAPIN in Rano Aroi, nos. 1045! 1046! 15. Juncus plebeius Rk. Br. SKOTTSB. 2. 70. Add Mataveri, CHAPIN no. 1043! 17. Chenopodium ambiguum k. Br. SKOTTSB. 2. 71. Add Apina iti (M. Fr.-B.) 19. Tetragonia expansa Murr. SKOTTSB. 2. 72. Add Moto nui (M. Fr.-B.) 21. Caesalpinia bonduc (L.) Roxb. SKOTTSB. 2. 72. Add Rano Raraku (M. Fr.-B.); crater rim of Rano Kao, 145 m, CHAPIN no. 1026! 22. Sophora toromiro (Phil.) Skottsb. 2. 73. On the verge of extinction already 40 years ago, but still observed in Rano Kdosby CHAPIN in) 1935, no: 1025)! 27. Erythraea australis R. Br. SKOTTSB. 2. 76. Add Ana Hooka (M. Fr.-B.) Add to Area of distribution: New Caledonia, Fiji. 30. Lycium carolinianum Walt. var. sandwicense (Gray) C. L. Hitchcock Ie. 243. 2. sanadwicense Gtay,, SKOLISB: 2.76. 2 Plants now, occurring ina-wild- or semi-wild state but re- ported to have been intentionally introduced by theearly native immigrants. Triumfetta semitriloba L. SKOTTSB. 2. 78. “T. aff. rhomboidea Jacq.’ of GUILLAUMIN refers to this species. C. SKOTTSBERG ~s ~I 12) Sapindus saponaria L. SKOTTSB. 2. 78. Rano Kao (M. Fr.-B.); same locality, 145 m, CHAPIN no. 1026! Thespesia populnea (L.) Correa. SKOTTSB. 2. 78. Hanga Roa (M. Fr.-B.), most likely a recent reintroduction. Solanum Forsteri Seem. Fl. Vitiensis 174; HEMSLEY, Fl. of Tonga 187. S. msulae-paschalis Bitter ex SKOTTSB. 2. 78. BITTER was of the opinion that the Morella from Easter Island was the species FORSTER, Fl. Ins. Austr. Prodr. 18, called S. xzgrum, but it escaped him that it had been described by SEEMANN as S. Forsterz. Add Anakena (M. Fr.-B.); crater rim of Rano Kao, CHAPIN no. 1028! Area of distribution: Polynesia. 3. Species unintentionally introduced with the humantraffic alten 22: Some of the plants enumerated below were reported already by the American Albatross Expedition but no other locality than “Easter Island” indicated. *Apluda pilosa Rich. — Hanga Roa (M. Fr.-B.) *Chloris gayana Kth. — Mataveri, CHAPIN no. 1041! *Briza minor L. — Rano Kao and Rano Aroi (M. Fr.-B.) *Ceratochloa untoloides DC. Moeroa and Ana ahoha (M. Fr.-B.) Sesyrinchium micranthum Cav. SKOTTSB. 2.80. — Hanga Roa and Rano Kao (M. Fr.-B.); Rano Raraku, in an artificial cave in quarry, CHAPIN no. 1052! Portulaca oleracea LL. SKOTTSB. 2. 80. — Rano Aroi and Moto nui (M. Fr.-B.) Polycarpon tetraphyllum L. SKOTTSB. 2. 81. — Hanga Roa (M. Fr.-B.) *Boussingaultia gracilis Heim. — Toki (M. Fr.-B.) *Crotalaria striata DC. — Moeroa (M. Fr.-B.) *Teucaena glauca Benth. — Rano Kao (M. Fr.-B.). Very likely introduced intentionally in recent times. *Oxalis repens Thunb. — Hanga Roa, CHAPIN no. 1047! *Apium ammt Urb. — Hanga ho orno (Baie La Pérouse, M. Fr.-B.) *Convoluulus parvifiorus (Vahl? author’s name omitted by GUILLAUMIN). — Rano Kao (M. Fr.-B.) *Calystegia sepium (L.) R. Br. — Crater rim of Rano Kao, CHAPIN no. 1030! *Verbena bonariensis LL. — Rano Kao (M. Fr.-B.) “Verbena litoralis H. B. K. — Inside of Rano Kao on grassy slope, CHAPIN no. 1024! * Taraxacum spec. — Rano Kao, 300 m, CHAPIN no. 1021! 7. CARL SKOTTSBERG: The Phanerogams of the Juan Fernan- dez Islands. I~ indigenous spiecre's: 1. Stipa neesiana Trin. et Rupr. l.c. incl. var. fernandeziana etc. S. fer- nandeziana Phil. non Steud., S. Skottsbergi Pilger, SKOTTSB. 3. 96. — Plates 56, 57. In his Synopsis of the South American S#pae HITCHCOCK recognized 2 A SUPPLEMENT TO THE PTERIDOPHYTES AND PHANEROGAMS OF JUAN FERNANDEZ I fog (0 species as indigenous in Juan Fernandez, macrathera Phil. and fernandeztana (Trin. et Rupr.) Steud. In the key they are distinguished from each other and from their nearest relatives as follows. Pubescense on fruit! extending, at least in lines, nearly to the summit. Fruit 6—7 mm long mucronata Pubescense on fruit confined to the callus? and lower part of the body. Fruit 5—10 mm long. Fruit about 10 mm long Terminal segment of awn 5—7 cm long, flexuous macrathera ley sa se VSitaiont fernandeziana Fruit 7—8 mm long neestana As stated in 3.97 I came to the conclusion that there is only one true Stipa (excl. Nassella) in the islands, originally known as Sépa neestana Trin. et Rupr. y fernandeziana Trin. et Rupr., referred to by Hitchcock as S. fernande- siana (Trin. et Rupr.) Steudel. Certainly STEUDEL believed that what he had before him was the variety of TRINIUS and RUPRECHT, of which he had not seen the type, but as PILGER has shown (Skottsb. 1.22), STEUDEL’s fernandeziana is = Piptochaetium bicolor (Vahl) Presl, whereas the var. fernandeziana is a true Stipa. Of this HircHcock examined both the type and my material, but he never saw the plant on which STEUDEL based his combination. Another name for the island Stpa is S. Skottsbergi Pilger ex. SKOTTSB. 1.22, coined to replace feruandeziana Steud. There was already a S. fernande- stana Phil. 1873, which, in spite of being a later homonym, was retained by PILGER who at that time counted with two closely related species in the islands, and so did, as we have seen, HITCHCOCK who, however, rejected PHILIPPI's name as illegitimate according to the Rules and took up S. macrathera Phil. to replace it, because these two species were regarded as identical by HITCH- cock. I have seen the type of S. fernandeziana Phil. It differs from all other specimens from the islands that I have examined in the longer empty glumes and longer awn, but I have shown (3.97) that it represents an extreme form, of which only the type is known, that it is united with the normal state by transitions and that it cannot be regarded as specifically distinct. The type of S. macrathera Phil., which I have not seen, came from the vicinity ef Santiago. Of the three closely related species mucronata H. B. K., neestana and fer- nandeziana | have examined the material in Herb. Stockholm, 42 sheets (excluding duplicate sheets). This material is small, but it is distributed over the entire range of these species and gives, I believe, a fair idea of the variation; I have, however, not seen Mexican specimens of mucronata. Typical mucronata should, according to HITCHCOCK 223, have glumes 10— 12 mm long, a small lemma of only 5—6 (6—7 in the key, say 5—7) mn, dark brown, villous on the lower half, glabrate and papillate-rroughened in the Il body of lemma. stipe of lemma. I Gh C. SKOTTSBERG Table I. Stpa mucronata, mm. ea ere ————EEEEEEE——————————_—_—__ | Length of Pu- Length of} lemma, Country, name on label, collector pees ona ae ar peust ee blade glume crown, excl. awn Colomb. mucronata. Sneidern 2127. . (+) 12 ot 6.s—7 30-535 Ecuador » Acosta Solis 10327 . = 2 gated 6) 655 WS 35 Onl ns See » » Asplund 6021. . . ae 1ul LO) e.57) Ge » » » BIZ? oye! « (Ge) II, 10 5.5 3540 » » » 6852) Omi (+) 9,8 5 22—27 » neestana Holmgren 527 ... = 12, 10.5 6 Somes) | DSe 3 Peru mucronata Macbride 378 ... . air F:sG RS Bye ae IO WP Shorr s}5y |) Eo » » » T2ESVA. es he = E25b SWS § | -45' 20) 20—25 | 55:5 » » Nsplund s0253)c-) ieee ar Ii, “0 6 30—40 Bolivia» Mandon 1276 .... — 12, {mus 6 40o—-45 | 55:6 » » Buchtrenys, nel i ee — 9.5, 8.5 6 33-4001 a5 5e07 » » >» S5OBS.. be aae = Fg, ¢12 6.5 40—45 » » Hammarlund 179... == Nels » SC) 5-5= 0:5 | 3O=-S5 eases » » ASplunG es izes sit .« = 10.5, 9.5 5-5 27— 30 » » > GIO Hitt ols — 10.5, 10 | 6 30— 35 » » » 1KOPXO\ ay G OLES = Tessa 6 35—4e » » » BGIGS: Masie ts. -_— ith Teh 6 2530 » » Sys sii Seba: ee. (+) Tie 1Os5 6.5 35—40 Argent. » Buchtienisems “6700 are 1s) SO G57. 40—45 | 55:9 Chile pratensis Philippi 749... .. == LO, a0Or |eOss=— 7; 40—45 | 55: Io > amethystina Wechlert270) 0) se (+) E10) {i} 10357 4047 55 etl upper, abruptly contracted at summit into a smooth, cylindrical, whitish crown, and an awn of 35—50 mm, twice geniculate, the third segment straight or nearly so. None of the specimens examined, whether cited by HITCHCOCK or not, have a dark brown lemma, it is white or whitish in all except two, where it is very light brown. The contraction into a crown may be rather abrupt, but just as often it is not (see pl. 55:1—11). The size of the anthers often provides a good character in grasses as in other plants, but they are so easily caducous that they are rarely observed in dried specimens. In the two cases where mature anthers were seen they measured 2.5—2.7 mm. MACBRIDE 1283 differs from ordinary mucronata in the lemma being glabrous except at base, as in ~eeszana (pl. 55: 5). The specimens from Argentina and Chile approach xzeeszava in this respect. The awn is rarely over 40 mm long in mucronata. — Table I. The characteristic features of zeeszana are, if we follow TRINIUS and RUPRECHT, the subequal, about 14 mm long glumes, the 6—7 mm long lemma which is villous at the base only, the pubescence extending in a dorsal line half way up, and the long awn, 50 mm or more. Of the specimens I have seen only one A SUPPLEMENT TO THE PTERIDOPHYTES AND PHANEROGAMS OF JUAN FERNANDEZ 7 Table II. Spa neestana, mm. ~ |Length off = Pu- j|Length of lemma, Country, name on label, collector eee ine beer acRE EAP Pipl MIeUTe blade glume | crown, excl. awn | Chile weestana Skottsb. 1017 —_ eet 2 7 | 35—40 BOL ere | » fernandeziana Werderm. 331 | Cell ae toe |S) eae 50-55 le 50013 | | » bicolor Philippi 669 . + 13,, 12 hs) TH Oe 56:14 | » meestana Giinther & Buchtien s. n. ss 12) Liye Oo ese AC a HO 56: 15 | » amethystina Philippi 503 + 1 soleiccalieaero) 55-60 | 56: 16 | J. Fern. zeestana fern. Bertero 800 ate 13; 12 i =o 45—50 BOR] aI » fernandeziana Skottsb. 447 + HEE EN 8 55—60 56: 18 | » » » 19 -- TAS, 2p Oe— 10k | 160-65 56: 19 | » » > 28 3 + TAs 82) On O's 5 ——O0 56: 20 | » » > 244 GC) Dobe een LOE 5 —_Go 56: 21 | » » 171 +- Has tz 9) —=10) || - COE 5022 Urug. meestana Herter 135 = 5, heh yp Ob SKHiS > Yee cee an Boliv. » Asplund 117 (+) Wa 3) 10") 5000 | Argent. » Paredi-G019,. =: «. < .)| ~ (5) IA ts | 7-5 805) 05-70 » » Mandon 1276 . — 13), U7) 1 () s——=1100) 9! Néfo\— as Rye oul >» » Hammarlund 309 + 19, 18 10 ai S§—ee || ve 2 Chile manicata Middleton s. n. = 16, 14 g9 —I0 | 45—55 E726 Argentina zeeszana Gillies s. n. — 20, 18 Qp—10: || 7O--7,5 Be 27) Uruguay zeestana Schroeder 15307 ar its}, Wy | GC) =e) 7ST NES 57: 28 > > Tweedie s.n. . : — 22, 20 | 9 —10 | 80—g0 | 57:29 | J. Fern. fernandeziana Philippi, 1864. | 21,/ 16 Op > Leneo5—— 05 | Brazil » Malmes2o3\aycmr i = AT. WG) || ke) a | SOE) ||_-G 772 Sy had a lemma as short as 7 mm, in most they are 8 or 9; HITCHCOCK has 7—8. The length of the glumes varies even more, and so does the awn (see Table II). We could perhaps group the material in three lots: 1) “typical” xeeszana: empty glumes I11—13 and 10—12 mano, respectively, body of lemma 7—8.5, awn asa rule not over 50 mm; 2) empty glumes 13—15 and 11—14, body of lemma 8—10, awn attaining 60 or 65 mm; 3) spikelets distinctly larger, empty glumes 16—22 and 14—20, body of lemma 9—1I1, awn to 70, 80 or 90 mm, or even longer. Lot 2 would correspond to var. fernandesiana, but among plants from the islands BERTERO 800 p. p. and SKOTTSB. 447 belong to lot 1, and PHILIPPI (coll. 1864) to lot 3. Trintus & RUPRECHT’s statement “‘foliis planiusculis cum nodis glabris’’ does not hold good; all according to the conditions the blade is flat or convolute, and as a rule hirsute at least on inside, and the nodes are more or less hairy. They are, perhaps, almost glabrous in the type specimen, labelled ‘Juan Fernandez, Prescott’. I have never heard of him as a collector in these islands, and he is not in JoHOW’s list. My specimens were identified 77:4 C. SKOTTSBERG with the type of var. fernandeziana by HITCHCOCK. Lot 3 ought to correspond to S. macrathera Phil. and fernandeztana Phil 1873 non Steud 1854. But I do not see how we could segregate species or even tolerably good varieties. The Juan Fernandez form is variable enough to cover the entire range of variation, BERTERO no. 800 and the type of ferzandezzana Phil. being the extremes. Well developed anthers were observed in WERDERMANN no. 331, measuring 2 mm in length. Cleistogamous flowers with only one fertile but small anther were de- scribed. te2o72 Pubescence of leaves and nodes, length of blade and degree of convolution vary a great deal, and so does the length of awn. Under S. xeeszana taken in a broad sense, S. manicata Desv. and longiflora Steud. seem to fall. 2. Piptochaetium bicolor (Vahl) Desv. SKOTTSB. 3. 98. PARODI (2) referred the island plant to var. typzcum. Area of distribution: Argentina, Uruguay, Chile, Juan Fernandez. 3. Stipa (Nassella) laevissima (Phil.) Speg. Piptochaetium laevisstmum Phil.; SKOPISB? 3.08, PE ARODI .'c 13. Chusquea fernandeziana Phil. SKOTTSB. 3. 104, 8. 361. Add Masatierra: above Plazoleta del Yunque, west side, 300 m, with old panicles, BOCK no. 75! 20. Cladium scirpoideum (Steud.) Benth. et Hook. fil. SKOTTSB. 3. 106, 5. U5 SieS: 3Ol- Add Masatierra: summit region of El Yunque, TENZ no. 26!; ““Ostseite iiber Camote hinunter, 120 m (lowest station known), Herb. Bock! Hesperogreigia Skottsb. 6. 200. 29. H. Berteroi (Skottsb.) Skottsb. 1. c.; 8. 362. Gregia Berteroi, SKOTTSB.- 3. 109. Add Masatierra: Corral Remolino, Herb. BOCK no. 48! Inspired by a communication from Dr. LYMAN B. SMITH, LOOSER (11. 299) judges Hesperogreigia to be identical with Gvezgza, but he does not quote my paper on these genera. I think I have given good reasons for the generic separation from Greigza as typified by G. sphacelata (R. & P.) Regel, but I have never claimed it to be endemic in Juan Fernandez. To Hesperogrezgia belongs G. Landbeckit Phil. (H. Landbeckit (Phil.) Skottsb.), perhaps also other species which I had no opportunity to examine. 30. Ochagavia elegans Phil. SKOTTSB. 3. 110. HARMS l.c. 109 keeps Ochagavia and Rhodostachys as separate genera, whereas MEZ l.c. 204 unites them under the former name. I discussed this question with HARMS, who wrote: “Die Gattung Ochagavia Phil. scheint mir gut zu sein und sollte zzcht mit Rhodostachys vereinigt werden.’ MEZ remarks that O. elegans is not in cultivation, but I raised it from seeds in 1918 and again in 1919. It is a vigorous grower in a temperate house and cuttings were A SUPPLEMENT TO THE PTERIDOPHYTES AND PHANEROGAMS OF JUAN FERNANDEZ = 775 distributed to several botanical gardens, but I had to wait until April, 1944, before a specimen flowered. Fruits were not obtained; QO. is ornithogamous. 31. Luzula masafuerana n. sp. L. racemosa Desv. subsp. 7zsularzs Skottsb. 3.110; L. mmsularis Skottsb. 4. 89, nomen. I discussed the Masafueran Luzu/a at some length in 3. 111—112, expressing my conviction that ZL. racemosa Desv. coll. (type from Mexico) cannot be upheld in its present extent. In 1925 I used the name ZL. zzsu/aris. Ind. Kew. Suppl. 5, 1921, lists another species under the same name: “L. insularis Rouy: L. campestris var. insularis Briq.’’ This quotation is misleading, because ROUY never claimed his zzsu/arzs to be a separate species. Under L. campestris we find on p. 265: “Race. — L. insularis Rouy’’, and the expression “Capsule un peu plus courte que dans le type’ also shows that he did not regard it as a species. L. insularis Rouy would not, I believe, invalidate Z. zzsu/arzs Skottsb., but in order to avoid all possibilities of a mistake, I have found it better to propose a new name. 34. Juncus procerus E. Mey. F¥. acutus, JOHOW 150, SKOTTSB. 3. 112. During his revision of certain species of Fuzcus, Dr. H. WEIMARCK also studied my material from Juan Fernandez and stated that it belongs to ¥. pro- cerus (not published). Area of distribution: South Chile, Juan Fernandez. 37. Libertia formosa Graham’. Skottsb. 3. 113. This name has always been used for the large-flowered species from South Chile (Valdivia, Chiloe, etc.), with very large, reduced and bract-like upper leaves supporting sessile, compact and pseudo-umbellate inflorescences, until GUNCKEL lc. revived Strumaria chilensis Molina and coined the new combination Libertia chilensis (Mol.) Gunckel to replace ZL. formosa, a combination also used by DIELS Lacuna BO. There is another, similar species from the same region, formerly referred to the Neozeelandic L. zazozdes and later identified with ZL. e/egans Poepp. The main differences appear to be: formosa: bracts and bracteoles large, the primary bracts as much as 6 or 7 cm long; lateral inflorescences sessile, dense, or the lowermost pedunculate (even the lower 2 or 3 more or less pedunculate in the postfloral stage), peduncles and pedicels rather stout; sepals 6—8 mm long, petals 13—-16 mm, anthers renee, ok * elegans: bracts and bracteoles smaller, primary bracts 2—4 cm long; lateral inflorescences long pedunculate already in bud, less dense, peduncles and pedicels slender; sepals 4—5 mm long, petals 7—10 mm, anthers 2 mm +. There can be no doubt that MOLINA’s Strumaria (a South African genus of Amaryllidaceae), on which HERBERT based his Choerododia (1837), is a Lz bertia, and GUNCKEL wanted to prove that MOLINA had ZL. formosa in mind. | shall quote his description |. c. 130. * ROBERT GRAHAM, not MARIA G., as JOHOW thought. 776 C. SKOTTSBERG “17. Il Thekel. Strumaria Chilensis, getta da una radice fibrosa molte folgie lunghe due piedi strette apuntate lisce, d'un bel verde; tra le quali sorge un fusto diritto sugoso alto da cinque o sei piedi, guernito di tre o quattro folgie piu piccole amplessicauli e alterne. I fioro corimbosi sono composti di tre gran petali bianchi alternati con altri tre assai piu piccoli a punte rosse. I] frutto é una capsula triangolare di tre celle.” This is not a bad description, though the number of stamens is not stated, whereas the Latin diagnose on p. 284 is useless: “‘foliis linearibus; flor. altern. majoribus; struma libera.’’ And the species figures under Hexandria! I shall not deny that MOLINA may have had ZL. formosa before him, but I do not see that in the absence of a “type” this could be proved. GUNCKEL refers to Anal. Univ. de Chile 22 (1863), where PHILIPPI deals with certain plants described by MOLINA. He suggests that Strumaria chilensis was = Libertia ixtoides of GAY (not of SPRENGEL), but later (ibid. 1867) he changed his opinion, because he found that FEUILLFE’s plate (see below) does not represent L. zazozdes, but formosa. MOLINA, however, does not mention this plate. It is not easy to tell what FEUILLEE’s “Bermudiana Narcisso-Leucoij flore’, (Il. c. 9, pl. IV) is; the specimen figured has a depauperate inflorescence, but the description indicates 128 and 5 X1.5 mm for the petals and sepals, respectively, and GAY 31 cites it under zxo/des. Under these circumstances it seems better not to reject the wellknown name formosa given to an ornamental plant which has been in cultivation more than a hundred years. The type species of Lzbertia Spreng. is L. zazoides Spreng., based on Szsy- rinchium txtodes Forster from New Zealand, but when SPRENGEL published the new combination (1825, |. c. 168) he cited as place of origin both New Zealand and Chile. He did not tell where he got his Chilean specimens, but the name con- tinued to be applied also to a Chilean plant believed to be conspecific with z1- zoides from New Zealand. In 1862, when KLATT published his monograph of Libertia (2) he did not mention this at all; in its place we find “L. vestiozdes Spreng., Syst. Veg. I. 168, where, of course, no such name is found: “‘restioides’’, a senseless name in this case, is either a slip of the pen, perhaps a contamina- tion by Szsyrinchium restiordes Spreng. on the preceding page, or just a mis- print. Specimens from KLATT’s herbarium are in the Nat. Hist. Museum in Stockholm; no label has “‘restioides’’, all “‘ixioides’, written by KLATT — even a devoted lover of mis-spelt names will, I presume, admit that this is a case of unintentional error. Index Kewensis quotes “‘vestioides, Klatt = ixioides.’’ The story does not end here: when J. D. HOOKER (I. c. 274) quoted KLATT, the name turned out as vestzoides! KLATY refers to REICHENBACH who figured a specimen from New Zealand, and to SWEET’s L. grandiflora, also from New Zealand and regarded by some as a closely related species, by others as a variety. He did not keep the Chilean zxvo7des distinct, nor did he quote FEUILLEE. L. elegans Poepp. was regarded as belonging to formosa. L. ixtoides reappeared in KLATT 3.530, pl. 68, II, based on a specimen which had been wrongly labelled “Brasilia”. As synonyms were listed “L. res- tioides Hook. Fl. Novae Zelandiae 252, Reichenbach Icon. bot. exot. tab 157, A SUPPLEMENT TO THE PTERIDOPHYTES AND PHANEROGAMS OF JUAN FERNANDEZ phyla) and Klatt, Linnaea XXXI. 383°, but HOOKER, in 1853, could not very well mention vestzo7des, which dates from 1862, and REICHENBACH figured the true trtoides from New Zealand. BAKER 1.153 and 2.116 thinks that KLATT’s illus- tration represents formosa. To me it looks more like the Chilean Zxzo/des, listed by BAKER as L. elegans Poepp., syn. Roterbe elegans (Poepp.) Steud. nom., Plantae Lechlerianae 569, which I have seen. I am, however, not at all sure that this is a correct interpretation of POEPPIG’s e/egans. His description |. c. 1 “L. caule subramoso ... Caules ... leviter flexuosi, ... ramulo uno alterove abbreviato... Fasciculi florum alterni in caulis flexuris, interdum ramulos capi- tatim terminantes ... Flores magni ...’’ could just as well apply to formosa. POEP- PIG’s specimen is supposed to be in Leningrad and consequently out of reach. If it belongs to formosa, txtoides of GAY and others has no valid name. ZL. chz- /ensis Klotsch is a manuscript name, mentioned by Baker, for this species, but it cannot be taken up because of L. chzlens7s Gunckel. The citation “L. chi- lensis, Klotsch, ex Baker’? in Index Kewensis is misleading. GRAHAM I. c. 383 described two species, L. crassa and L. formosa, and as crassa precedes formosa it could claim priority, but they are united under /or- mosa by every author after GRAHAM. He redescribed and illustrated formosa in Bot. Magazine, without even mentioning the other species, and I don’t sup- pose that we are obliged to change the name. Both were raised from seeds col- lected in South Chile; LINDLEY, Bot. Register, indicates Chiloé as the place where the seeds of formosa were gathered. ZL. crassa is said to be more robust and to have larger flowers (L. grandiflora Philippi non Sweet), and the outer edge of the equitant leaves is said to be minutely serrulate in crassa and smooth in formosa. As far as my experience goes even the stoutest specimens, with leaves as wide and flowers as large as in crassa, have a perfectly smooth keel. BAKER I. 153 mentions crassa as “Var. L. crassa Grah.”’ In his monograph of Szsyrinchium (1. 96, 97) KLATY lists under no. 39 S. sesstlifiorum Poepp., “Ins. Juan Fernandez leg. Bertero nr. 1316, mis. Delessert 1831, Herb. Reg. Berol. nr. 108’. Gay cites Valdivia and Concepcion, but not Juan Fernandez, BAKER ‘‘Chile and Juan Fernandez’, and DIELS 1. c. 479 fol- lowed him. STEUDEL, Nomencl. ed. 2. 595 quotes BERTERO 1316 under “S. Fer- nandezianum Steud. Hochst.’’, which is a nomen nudum. Neither of these names was mentioned by JOHOW. According to KLATT 2.382 PoEppic’s S. sesszliflorum is = Libertia coer- ulescens Kunth et Bouché, and the same is, according to BAKER l. c. 153, the case with the specimens quoted by HooKER and ARNOTT in the Botany of Captain Beechey’s Voyage 47 under S. sessz/iflorum. Also POEPPIG |. c. 2 refers to these specimens in addition to those collected by himself in Chile. He did not visit Juan Fernandez, nor does he quote specimens from there; KLATT seems to have been the first to do so. A specimen formerly belonging to KLATT’s herbarium, now in Stockholm, is labelled “‘coerulescens Kth & Bouche S. ses- siliforum Péppig leg. Poppig & Philippi in Chile’. The identification is correct, and the identity with PoEPPIG’s Sésyrinchium seems well established thanks to 778 C. SKOTTSBERG his reference to BEECHEY’s plant. Under ZLzdertza its name has to be L. sess7lz- fora (Poepp.), certainly another unwelcome change of a long established name. It remained to find out what the mysterious Szsyrzuchium from Juan Fer- nandez was. Thanks to the kindness of Dr. CHARLES BAEHNI I have been able to examine a sheet of BERTERO 1316 from Herb. DELESSERT, Geneva. It is labelled “1316 Sisyrinchium? ex insula Juan Fernandez misit D® Larrain NY 1828 D. Bertero’, and is, as could be foreseen, typical Lzbertza formosa Grah. Unless BERTERO 1316 was a mixture of two very different things it is difficult to un- derstand how KLATT could identify a large- and white-flowered species with the small- and blue-flowered coerulescens. 41. Peperomia fernandeziana Miq. SKOTTSB. 3.115, 5. 158. Add Masatierra: near summit of El Yunque, TENZ no. 82! 53. Spergularia confertiflora Steud. Plate 57: 31—33. SKOTTSB. 3. 121. Since my paper was published, Mrs. ROSSBACH has monographed Spergu- farta from North and South America. L. conxfertzfiora is considered to be en- demic in Juan Fernandez, var. folyphylla (Phil.) Skottsb. is merged into the spe- cies, and var. glaberrima referred to S. medza (L.) Presl. She overlooked that what passes as genuine confertzfiora has been collected also on San Ambrose Island. I still think that we have better to retain polyphyl/a as a variety or form (Plate 57:33), and even if I am prepared to admit that ¢/aderrima stands just as near, or even nearer to medza than to confertiflora, | find it impossible to unite them. L. c. 122 I mentioned S. remotifora Steud., comparing it with a lax-flowered form of conxfertifiora, while Mrs. ROSSBACH brings it to w7/osa (Pers.) Camb., a closely allied species. Area of distribution: Juan Fernandez. San Ambrosio. The type of my var. glaberrima is no. 411, of which Mrs. ROSSBACH exam- ined a number of specimens. She remarks that it differs somewhat from ordinary S. media. While I agree that it should be removed from confertefiora, 1 cannot simply refer it to medza. It differs from this by its stout, woody caudex and branches, the total lack of pubescence, smaller, mucronulate leaves (4—7 mm long by 0.6—2 mm wide), shorter stipules (1.5—2.5 mm), smaller flowers with few stamens, and smaller, minutely tuberculate, generally unwinged seeds; in media the seeds are smooth or very slightly roughened, as a rule winged and 0.8—I mm large. I know S. medza from the continent and also from the Falkland Islands. It is not possible to leave the plant growing on Masafuera as a variety of confertiflora, nor is it advisable to let it disappear under the Cosmopolitan media. Consequently, it is segregated here as a species. I would have preferred to use the varietal name for it, but as one cannot feel sure that an older glader- vima will not turn up some day under Spergularia, or Spergula, with which the former is united by many, I have found it safer to give it new name. *Spergularia masafuerana n. sp. S. confertiflora Steud. var. glaberrima Skottsb. 3.123; S. media, ROSSBACH 1.121, 2.240. — Plate 57: 34. A SUPPLEMENT TO THE PTERIDOPHYTES AND PHANEROGAMS OF JUAN FERNANDEZ 779 Perennis laete viridis glaberrima, caudice sat valido, radice ad 3 mm crasso. Caules numerosi, suberecti vel decumbentes, abbreviati, ramosi internodiis 2—4 mm longis. Folia linearia, carnosa, margine angusto scarioso, subobtusa usque acuta et mucronulata, 4—6.5 mm longa, 0.6—1 mm lata, ex axillis gemmifera. Stipulae triangulares, acuminatae, subintegrae, 1.;—2.5 mm longae. Flores ob caulem usque ad apicem foliosum pseudoaxillares, dichasia et monochasia foliosa formantes, pedicellis statu fructifero ad 5 mm longis erectis. Sepala ovata — ovatolanceolata, obtusa, 3—3.5 mm longa, 1.2—1.6 mm lata, margine scarioso 2.8 mm longa, 0.6 mm. Capsula calycem + aequilonga, 3—3.5 mm longa, diam. 1.5—1.8 mm. ad 0.3 mm lato. Petala sepalis sat breviora, alba, ovata, 2.5 I—1I.3 mm lata. Stamina vulgo 2 vel 3, 2.3—2.5 mm longa, anthera 0.5 Semina matura fusca, complanata, late obovata, minutissime tuberculata, exalata (sed vide infra), 0.5;—0.6 mm longa et 0.4—9.5 mm lata. ’ Typus ex ins. Masafuera num. 411 in Herb. Holm. Area of distribution Endemic. Seeds of all forms, taken from my herbarium specimens, were sown in a Temperate house. They germinated well and developed normally, but at a rather different rate: confertifiora, sown 21/6 1921, fruting April, 1922 (no mature seeds) v.polyphylla, » .26/6 1921, » February, 1922 (few » * 2.) masafuerana, » 20/6 1921, » November, 1921 (numerous ) In cultivation, the three forms remained perfectly distinguishable from each other, but grew to become more luxuriant than the wild plants. A comparison is made in Table III. Whereas no winged seeds were observed in wild-growing specimens, they turned out to be quite common in cultivated masafuerana. Twelve capsules, 6 from one potted plant and 3 each from two, were examined. See Table IV. 58. Drimys confertifolia Phil. D. Wznterz Forst. var. confertifolta (Phil.) Johow; SKOTTSB. 3. 127. Restored to specific rank by Dr. A. C. SMITH |.c. The most important character seems to be the number of ovules, (16—)18—26 in confertifolia, g—18 in Wxterz. 63. Escallonia Callcottiae Hook. et Arn. SKOTTSB. 3.129, 5.158. Add Masatierra: summit of El Yunque, TENZ no. 49! (TENZ l.c., pt. 6. 9.) 64. Rubus geoides Sm. SKOTTSB. 3. 130. Add to Area of distribution: Chile, Cord. Nahuelbuta. 65. Margyricarpus digynus (Bitt. ut subsp.) n. sp. J/. setosus R. et P. subsp. digynus Bitter ex SKOTTSB. 1.9; id. 3. 130. Il have gone over a large material of J/. setosus or, as it is now called, JV. pinnatus (Lam.) OK. from its entire range on the continent. The insular form stands apart from all forms from the mainland in leaf size etc., as set forth in 52-—-516710. The Nat. Hist. of Juan Fernandez and Easter Isl. Vol. II. C. SKOTTSBERG Table III. A comparison between my material of wild-growing Spergularia and cultivated plants raised from seeds secured from that material. Measurements in mm. 1 colour; 2 pubescence; 3 length of stipules; 4 leaves, length and width; 5 sepals, idem; 6 petals, idem; 7 number of stamens; 8 length of filament; 9 anther, length and width; 10 length of capsule; 11 seeds, length and width. confertiflora v. polyphylla masafuerana Wild |... Cultiv. Wild Cultiv. Wild | Cultiv. I Dark green Dark green Light green | 2 Densely glandular Densely glandular None BO) gee 0 | 5 —6 Dig 6 I.5—2.5 3—4 {5 —20(—30)}] 10 —30 5 SOs see 20 4 —6.s\—10)| 10 —25 5 lo.s— 0.7—1)] o5— Lz C—O O:5— Is o<—I(—2) | I —2 weal ey bAtaa |) 4— 45 | 35— 45 | Wire Beas 1c Aaa : fe I — is | I3— 2 I.2— I.s5 I.2—I.6 | I.s— 2 5 (355 Ae 5 32—- 4 5 2.5—2.8 a 3 latyes 235 95 —— BG 4a 12EF PAS aes) es a) Ce 1G gee lee Uh 5 ao Ast 5 Bila A= 40 Zia 20s 8 2.4— 2.5 1.8— 2 2 — 2.5 1.8— 2 2.3—2.5 B= {O.s— 0.7 0..— 0.6 0.5—0.6 7 Ge 0.4 Oe Oy Ors O24 10 4 — 6 | Sia ele 4k ee 3 3s 3s Aes f{0.6— 0.7 0.6— 0.7 ‘| O.<—0.6 II (co. 0.5 Of OPE On Os5 Table Ive Winged and wingless seeds in cultivated S. masafuerana. Koh: 43 4 5 6 i 8 Oy:.| 10. |. .11,) |) Sezai | | | | | | | | | Winged . . | 5 28 22 | 18 thee 8 | 28 | 20 | 49 | GA\| | 3X0) | fe) 224 Wingless .| 48 2 MO) Se ||) 49 85 | 75 A128 | 98 | 109 | 756 \"53 | 36 | 62 | 49 | 62 | 57 | 113] 95 | 92 | 130 | 128) fegummeee BITTER’s description, and in the constant occurrence of 2 carpels. I regard it as specifically distinct. OF. I am unable to draw a limit between the plant growing on Masatierra and var. zsulae-exterioris Bitter ex SKOTTSB. 1.245 from Masafuera. Both seem to differ from typical australzs in anther size, having anthers only 0.25 mm long: and 0.4 mm wide, whereas they are 0.5 X 0.75 according to BITTER in the latter, Acaena ovalifolia Ruiz et Pav. subsp. australis Bitter. SKOTTSB. 3. 133. A SUPPLEMENT TO THE PTERIDOPHYTES AND PHANEROGAMS OF JUAN FERNANDEZ 781 but a much larger material from both islands is required in order to know if a special insular form or variety should be distinguished. Acaena argentea Kuiz et Pav. X Margyricarpus digynus; J/argyracaena Skottsbergu Bitter ex SKOTTSB. 3. 134. A comparative anatomical study of Margyracaena and its supposed parents by Miss ASTRID RINDE (Mrs. WALLERIUS) has confirmed the opinion on its origin expressed by BITTER and the author. 71. Fagara externa Skottsb. 3. 143, 5.170. Add Masafuera: C. Ruiz S. (fr. Oct. 1927! Herb. Looser). 72. Dysopsis hirsuta (Muell. Arg. ut subsp.) n. sp. D. glechomoides (Rich.) Muell. Arg. subsp., SKOTTSB. 3. 144. The insular form differs from both mainland forms (which, I believe, ought to be recognized as species) in so many characters that I have raised it to specific rank. *Callitriche marginata Torrey var. Lechleri Hegelm. C. stagualis, SKOTTSB. 4.220. In 1922 this was published as C. stagnalis on the authority of the late prof. G. SAMUELSSON and regarded as adventitious. Later SAMUELSSON renamed my material C. marginata, a North American species, var. Lechleri Hegelm. The variety is widespread in Chile where, undoubtedly, it is indigenous, and it may well have reached Masatierra without the aid of human traffic. 73. Empetrum rubrum Vahl. SKOTTSB. 3. 145. R. Goop suggests |]. c. 511 that my plant from Masafuera, and one specimen from Fuegia, belong to rubrum subsp. Eamesz (Fern. et Wieg.) R. Good, other- wise known only from E. Quebec, Labrador and New Foundland, a form with pubescent stem, glabrous leaves and pale red, translucent berries, differing from both rubrum var. atropurpureum (Fern. et Wieg.) R. Good and the Arctic xigrum f. purpureum (Raf.) R. Good. The principal difference between xigrum and rubrum is, according to GooD, that the former has glabrous, the latter pubescent shoots, a rule not without exception. Of the Masafueran form, only a single shrub has been found, without flower or fruit. The leaves are small as in Lameszz, the blade 2.;—3 X I—I.2 mm, but leaves just as small are observed in plants from South America, e. g. in specimens collected by BANKS and SOLANDER in Fuegia; they should be glabrous in Lameszz, but have a tomentose margin, with very long, curly hairs almost as in f. fa/klandicum, in the specimen from Mas- afuera. As long as the fruit is unknown, a comparison cannot be made with either Eames or any other northern form; at present I can see no good reason to link it with +wérum from the far north rather than with Subantarctic forms. 74. Colletia spartioides Bert. ex COLLA. SKOTTSB. 3.145 and 5.158; BENZ. L c.555 4s Add Masatierra: between B. Cumberland and Pto Frances, TENZ no. 31! 482 C. SKOTTSBERG Nothomyrcia Kausel 1. 147. N. fernandeziana (Hook. et Arn.) Kausel |. c. 148. Wyrceugenia fernandeziana (Hook. et Arn.) Berg. SKOTTSB. 3. 147. 83. Halorrhagis asperrima Skottsb. 3. 153. Of this two sets of specimens were collected, no. 442 with smooth fruits or occasionally with a trace of wings, and no. 457 with 4-winged fruits; single cases were, however, almost devoid of wings. Seeds of both numbers were sown Jan. 3, 1921, and 1o plants of each potted separately May 11. Of no. 442 about 2000 ripe fruits were gathered Jan. 22, 1922; all were quite smooth except a dozen, where from one to four small wings had developed. No. 457 gave 307 fruits; most had a trace of wings, but only 16 were typically 4-winged. 84. Halorrhagis masatierrana Skottsb. 3.155. The fruit seems to vary a good deal in size in different individuals; see also SKOTTSB. 5.155. Unfortunately, I did not manage to raise this species from seed. 85 Halorrhagis masafuerana Skottsb. 3. 156. Of this I had three sets of specimens, nos. 1416 with very smooth, quite wingless fruits, no. 441, equally wingless, but faintly rugose-tuberculate, and 485, where here and there a wing had been developed. Of all three fruits were sown, but no. 485 did not germinate. Nos. 441 and 1216 were sown the same day as H{. asperrima, and of the former 5, of the latter 8 vigorous plants were potted separately June 13. Ripe fruits were collected Jan. 23, 1922. Of both numbers there were single fruits faintly tuberculate-rugose transversally. In habit, leaf shape, colour, pubescence, etc. both species have proved constant. *Centella triflora (Ruiz et Pav.) Nannf. SKOTTSB. 5. 169. Masatierra: Bock! We have no good reason to suspect this to be a late introduction; perhaps it is, but just as well it may have been overlooked. Area of distribution: Chile, 36°—44°, Juan Fernandez. Authentic material of ““Uredo Hydrocotyles Montagne’ is in Herb. Uppsala (Herb. ELIAS FRIES), two intact leaves of the host plant, resembling 7. modesta Cham. et Schlecht. and labelled ‘““Bertero J. Fernandez herb. Montagne’. If the locality is correct, a species of Hydrocotyle occurs or, in any case, once occurred on Masatierra. 86. Eryngium bupleuroides Hook. et Arn. SKOTTSB. 3. 157. Add Masatierra: summit of El Yunque according to TENZ l.c. 6: 9. Eryngium bupleuroides X inaccessum. F. fernandezianum Skottsb. 1. 17, 3. 160. — On my request one of my students made a comparative anatomical study of E. fernandezianum and its supposed parents (alcohol material of all available). The results, which have not been published, confirm my view that it is of hybrid origin. A SUPPLEMENT TO THE PTERIDOPHYTES AND PHANEROGAMS OF JUAN FERNANDEZ 783 g1. Pernettya rigida DC. SKOTTSB. 3. 161. Through oversight P. vacemulosa DC. was not mentioned l.c., although | had excellent material of this ‘species’. The diagnose says ‘‘pedunculis lateralibus ex parte ramorum inferiore ortis ramosis multibracteatis.’’ SLEUMER |. c. 649 refers it to szg7da without comment, but KAUSEL 2.168 adds the remark “en muchas muestras se observa pseudo-racimos, en los cuales las hojas adyacentes se ven reducidas al tamano de una bractea’. The formation of these racemes takes place not only in the lower portion of the shoot. In SKOTTSB. no. 77 the top of a branch gradually passes into a terminal raceme, in no. 27 pro parte such racemes replace lateral, solitary flowers. PP. vacemulosa is a morphological aberration without taxonomic value. 93. Calystegia tuguriorum R. Br. SKOTTSB. 3. 162. STAPF l.c. treated this species at some length. He illustrates the New Zealand plant, leaving undecided whether or not the Chilean C. Hantelmannz Phil. is conspecific. The woody stem in this cannot, I suppose, prevent us from uniting them, if no other differences are found; besides, I am not sure that it is strictly herbaceous in New Zealand. I have not seen the fruit of Maztelmannt. PHILIPPI’s description fits the New Zealand plant well; however, he calls the corolla lobes short, while STAPF describes the corolla as divided to 1/4—1/2. The specimens I had occasion to see showed considerable variation in this respect. 94. Selkirkia Berteroi (Colla) Hemsl. SKOTTSB. 3. 163, 8. 362. Add Masatierra: south face of Co. Damajuana, fl. 26/10 1935, BOCK 110! 95. Rhapithamnus venustus (Phil.) B. L. Robinson in Proc. Amer. Acad. 51 (1916) 531. R. venustus (Phil.) Skottsb. 3.163. — I had not observed that the new combination had already been made by ROBINSON. Cuminia Colla. Skottsbergiella Epling; Fohowza Epling et Looser, see SKOTTSB. II. 97. C. eriantha Benth. SKOTTSB. 3. 165. EPLING 2.169 suppressed C. erzantha, but his discussion shows that he did not consult my papers, and he does not quote them. I have studied this genus in nature, and even if slightly hirtellous specimens of C. fernandezia were seen occasionally, I did not find it difficult to distinguish two species. JOHOW was quite positive on this point: ‘‘Reunir, como lo cree posible Hemsley, la C. fer- nandezia i la C. eriantha en una sola especie seria a mi juicio un error gra- visimo”’ (l.c. 81). BENTHAM observed only the smaller female flowers; also note Heures. in DC. Prodr, Xi 258. EPLING’s statement with regard to the mode of occurrence of Cwminia is not in accordance with the localities I have cited. 98. Solanum fernandezianum Phil. SKOTTSB. 3. 165, 8. 362. Add Masatierra: Plazuela del Yunque, fl. 23/11 1935, BOCK no. 3! 7 84 C. SKOTTSBERG 101. Nicotiana cordifolia Phil. SKOTTSB. 3. 168. Add Masafuera: Q. del Ovalo and Q. del Mono (P. ARAVENA 1941), the first finds north of Casas Valley (GOODSPEED |. c. 125). 102. Mimulus glabratus H.B.K. var. externus (Skottsb.) Skottsb. J/. parviflorus Lindl. var. externa Skottsb. 3. 168. GRANT l.c. 194 reduced parvifiorus to glabratus as a variety. The island form differs from both in the upper leaves, which are distinctly petiolate, from parvifiorus in being perennial. Stem pubescent, finally glabrate, petioles and leaf base with patent hairs along edge, blade puberulous on both faces. Pedicels short, 5—1O mm, pubescent. 104. Plantago fernandezia Bert. SKOTTSB. 3.171, 5.158. Add Masatierra: near summit of El Yunque, TENZ no. 61! 108. Coprosma Hookeri (G. Don) W.R. B. Oliver l.c. 141. C. tr7florum (Hook. et Arn.) Benth., et Hook. f. SKOTTSB. 3.172, a combination not in ac- cordance with the Rules. 109. Coprosma pyrifolia (Hook. et Arn.) Skottsb. 3. 173. Feminine gender now generally accepted. 113. Wahlenbergia Grahamae Hemsl. SKOTTSB. 3.178, 8. 362. Add Masatierra, Co Central, 400—450 m, fl. 28/12 1934, BOCK no. 59! 115. Wahlenbergia Berteroi Hook. et Arn. SKOTTSB. 3.179, 5. 158, 8. 362. Add Masatierra: Pangal, fl. 6/12 1935, BOCK no. 125! Pto Frances, TENZ [Cy Pty Sind. Robinsonia DC. In my treatment of this genus I criticized the tendency to split genera when nothing is gained in the direction of a better understanding of their components. I pointed out that Aodzmsonza includes two different entities, of which one is monotypic (RX. macrocephala); they are related to each other while differing from all other genera with the exception of Rhetinodendron. Twenty years have passed since this was written; the splitting process which, in numerous instances, means going back to an earlier generic conception, has continued, and I am convinced that if a new attempt is made to monograph the Compositae, Rodzzsonza will not be left standing in its present sense. 2. macrocephala will either become the sole member of that genus, or be excluded under a new generic name. Fortunately it is possible to take the latter course. DE CANDOLLE, I (1833) 333, founded the genus Rodzxsonza on two species discovered by BERTERO, but they were not described by him but by DECAISNE, as R. Gayana and R. thurifera \.c. 2 (1834) 28, where also R. macrocephala, a species not collected by BERTERO, was described. That DE CANDOLLE did not have macrocephala before him is evident from his generic description: “ Achenia glabra ... Pappus 1-serialis pilosus scaber caducissimus,’ while DECAISNE has A SUPPLEMENT TO THE PTERIDOPHYTES AND PHANEROGAMS OF JUAN FERNANDEZ 785 “Pappus uniserialis nunc persistens pilis in tubum coalitis nunc basi frangens caducissimus,” and DE CANDOLLE 2 (1837) 448 ““Pappus... setis nunc liberis nunc in vaginam concretis.’’ At the same time he figured Gayana and thurifera in DELESSERT’s Icon. select. plant. IV. As type of the genus Rodzusonza s. str. I propose R&. Gayana, the commonest species. KX. macrocephala will form a separate, monotypical genus. 120; Rh. oracilis, Dcne, SKOTTSB. 3. 190; 5. 156,:.6.. 362. Add Masatierra: Corral Remolino, 400 m, ¢ fl. 8/12 1935, Herb. Bock no. 82! summit of El Yunque, TENZ l.c. 6:9, and nos. 35! 39! Symphyochaeta (DC. ut sect.) n. gen. Dioica. Capitula 2 paniculam pyramidalem formantia, calyculo e bracteis paucis minutis instructa. Squamae involucri rigidae, usque apicem in cupulam concretae, maturitate fructuum irregulariter dehiscentem. Pappus & setis ad medium arcte coalitis vaginam membranaceam persistentem formans, corollam item persistentem amplectens. Achaenium maturum lageniforme, corolla duplum vel ultra longius, nervoso-costatum, transverse rugosum, parce sed perdistincte setosum. Icon. SKOTTSB. 3.194. — Arbuscula habitu Rodzvsoniae, sed folia reticulo perdenso pellucido venularum a Robinsoniis veris discrepantia (v. JoHOW 56, fig. 2). 125. S. macrocephala (Dcne) n. comb. Rodinsonia macrocephala Decne, SKOTTSB. 3. IQT. 131. Rhetinodendron Berterii (Dcne) Hemsl. SKOTTSB. 3. 197, 5. 160. Add Masatierra, south side of Portezuelo de Villagra, c. 400 m, BOCK no. 71! summit of El Yunque, TENZ no. 4o! (a very large-leaved form). 132. Centaurodendron dracaenoides Johow. BOCK; SKOTTSB. 3.198, 5.160, 8.364, where the flowers are described and illustrated. Add Masatierra: Co Damajuana, c. 500 m, fl. 8/12 1935, WEBER in Herb. Bock no. 111! Summit of El Yunque, TENZ nos. 36! 41! *VYunquea Skottsb. 5.163, 8.363; LOOSER 4. *Y. Tenzii Skottsb. Masatierra: summit of El Yunque, TENZ 1922, no. 42! ANDAUR, 925 m, Aug. 1934, Herb. Bock no. 40! YUNGE & GROSSE 1935, Herb. LOOSER n. 3162! Young leaves covered on both sides (more thickly on the lower) with a soft white, caducous wool. Marginal teeth end in a large capitate secretory gland. Dendroseris D. Don. SKOTTSB. 3.199. What I have said about Rodinsonia and its components applies also to Dendroseris which, according to my views, could be divided but just as well left in its accepted sense. WODEHOUSE’s pollen studies made me reconsider the alternatives. Of the three subgenera recognized by me, WODEHOUSE examined 786 C. SKOTTSBERG representatives of two, Kea and Phoenicoserzs. He found the pollen structure so different that he wrote: “These differences are such as to suggest that the two plants may be genetically unrelated’. The investigation was continued and ex- tended by G. ERDTMAN who had a much larger material at his disposal, includ- ing all four species of Ludendroserts, Phoentcoseris pinnata, and three species of Rea representing its two sections. Ludendroseris showed, as could be ex- pected, its own pollen type, different from the kinds described by WODEHOUSE. To my surprise the pollen in D. gzgantea, of the monotypical sect. Schzzoglos- sum, not only differs from that of Aurea, but is very like the pollen of Phoenz- coserts, from which, in all other respects, Y. g7gantea is far removed. I think that ERDTMAN’s statement shows that the three subgenera are connected; it goes without saying that, contrary to WODEHOUSE’s suggestion, they are genetically related. With a narrow genus concept it is easy to break up Dezdroserzs even without palynological assistance, and in spite of palynological evidence, D. gzgan- tea must form its own genus. Dr. ERDTMAN has been kind enough to supply the following palynological description and key to Dendroseris s. lat. Pollen grains slightly flattened (diameter including spines 30—55 1), with 3 equatorial aper- tures situated in small areas (apertural lacunae) sligthly depressed under the general surface. These lacunae communicate with 6 ab-apertural lacunae (ab-poral, WODEHOUSE), 3 in each hemisphere. Non-lacunar parts of surface covered with a more or less undulating spiniferous layer (‘‘sculptine”’; in certain species provided with 6 distinct para-apertural depressions without or with a few spines). I. Longest axis (spines included) 47—55 1. Ab-apertural lacunae usually smaller than apertural. Thin rods connect the spiniferous layer with the inner part of the exine. . Hudendroseris A. Longest axis 50—55 1. Spines (or, as in the following, the solid apices of more or less spinoid excrescences) about 4 p. 1. Sculptine surface only slightly undulating (spines less numerous than in 2) marginata 2. Sculptine surface distinctly undulating (spines more numerous than in 1) macrantha, macrophylla B. Longest axis about 47 ». Spines 3—4p. Ab-apertural lacunae distinctly smaller than Td PNG REVAL r cet orice ae fiag Stok. Soe" Sg SAE oy Se ak, Seige co ees de> 4) 5e) ee ne II. Longest axis 31—37 ». Ab-apertural lacunae usually not smaller than apertural. No rods like those mentioned under I. A. Spines less than 1.5». No distinct apertural depressions. Longest axis 31—32 p Rea sect. Eurea (micrantha, pruinata) B. Spines 2—4p. 6 distinct para-apertural depressions. Longest axis 34—37 p. 1. Spines about 2p, situated on top of distinct rounded protuberances Rea sect. Schizoglossum (gigantea) 2. Spines about 3—4 p, situated on top of feeble protuberances Phoenicoserts (pinnata) Dendroseris D. Don. ). subgen. Ludendroseris Skottsb. 3.200. Capitula majuscula usque 5 cm lata flosculis aurantiacis. Receptaculum nudum foveolatum foveolis autem fibrilloso-marginatis. Achaenia suberosa, alata, valde dimorpha, marginalibus trigono-alatis a dorso, ceteris a lateribus complanatis. Pappus brevissimus caducissimus. Arbusculae trunco cavo ramoso, foliis rosu- latis simplicibus molliter coriaceis. Typus: D. macrophylla D. Don. A SUPPLEMENT TO THE PTERIDOPHYTES AND PHANEROGAMS OF JUAN FERNANDEZ 787 | | | Achenes, x 10, of Phoenicoseris pinnata (a—d), Rea micrantha (e—h), and Hesperoseris gigantea (i—l); a, b, e, f, i, marginal, front view, c, d, g, h, k, 1 central, side view. 135. D. marginata (Bert. ex Decne) Hook. et Arn. SKOTTSB. 3. 202, 8. 363. Not D. macrophylla v. marginata Johow 259. Add Masatierra: south face of Co Damajuana, c. 500 m, fl.-fr. Oct. 1935, Herb. Bock no. 109! Pico Central, c. 500 m, fr. Dec. 1934, BOCK no. 87! 136. D. litoralis Skottsb. 3.204, 8.363. D. macrophylla v. marginata Joh. |.c. Add Masatierra: Morro Vinillo, fl. 12/11 1935, Herb. Bock! In flower and fruit characters this was shown to differ from its three con- geners (3.204, fig. 37). This also applies to the pollen. Phoenicoseris nov. gen. Dendroseris subgen. Phoentcoserts Skottsb. 3.201. Flores lactei. Receptaculum nudum foveolatum foveolis fibrilloso-marginatis. Achaenia dimorpha, irregulariter suberoso-tuberculata, pappo aequilongo caduco. Arbores palmiformes hapaxanthi trunco simplici cavo, foliis pinnatis. Typus: PA. pinnata (Bert. ex Dene) Skottsb. — Text figs. a—d. Structura pollinis ab illa Dendroseridis nec non Reae diversa, illae autem Hesperoseridis similis. *Phoenicoseris Berteriana (Dcne) n. comb. SKOTTSB. 5.166, 8.363. Den- droseris pinnata var. imsignis Bert., Johow; Skottsb. 3.205. Add Masatierra: foot of El Yunque, 350—400 m (a surprisingly low alti- tude), BOCK no. 86! near summit, ANDAUR, Herb. BOCK no. 41! TENZ no. 46! A. GROSSE in Herb. LOOSER! 138. Ph. regia (Skottsb.) n. comb. Dendroseris regia Skottsb 3.205. 788 C. SKOTTSBERG Rea Bert. ex Dene pro parte. Dendroseris subgen. Rea (Bert.) Skottsb. 3.201. Capitula 1—1.5 cm lata, flosculis lacteis. Receptaculum nudum. Achaenia obovoidea-oblageniformia, dimorpha, a dorso vel a latere compressa, valde nervoso- costata, pappo achaenio breviore, caduco. Arbusculae usque arbores trunco ra- moso solido, foliis rosulatis simplicibus tenuiter membranaceis. Typus: 2. mz- craniha Weney — lexteiseme——i. 139. Rea neriifolia Dene. Dendroserts nerifolia Hook. et Arn. SKOTTSB. 3.206. 140. Rea micrantha Bert., Decne. Dendroseris micrantha Hook. et Arn. SKOTTSB. 3.206. 141. Rea pruinata (Johow) n. comb. Dendroseris prumata Skottsb. 3.207, Bah7O;° 8: 3037 Add Masatierra: Pto Ingles, b.-fl. 25—26/12 1927, LOOSER nos. 488! 489! south side of Portezuelo de Villagra, 300 m, BOCK no. 65! Hesperoseris nov. gen. Dendroseris, Johow; LY. subgen. Rea sect. Schisoglossum Skottsb. 3. 201. Capitula parva anguste cylindracea receptaculo nudo flosculis lacteis. Achae- nia (matura haud visa) illis generis Reae similia, pappo achaenio multum longi- ore, ut videtur longius persistente, setis numerosissimis. Ligulae usque basin 5-divisae. Arbor trunco ramoso solido, foliis simplicibus papyraceis. Typus: /7. gigantea (Johow) Skottsb. — Text figs. i—l. The allotment of 8 important characters between the four dendroseroid gen- era is shown in Table V. iabler va SONA POSES” sor eB NS Mia Ser Coe A eit ceo) KO oem Sona +e kp ig Ex Alege 40) ICT UGDSIFUS, 6 SB Sy I AR eI eye WICSPCOSETIS. 5 JO Dy tar (Cc ot) ae P Explanation. Stem: S solid, s hollow; B branching, b simple. Leaves: L simple, 1 pinnate. Receptacle: R fibrillose, r naked. Corolla: C orange, c white. Ziguwle: T entire, t deeply 5-split. Achene: A winged, a unwinged. Paffuws: P as long as or longer, p much shorter than the achene. The statistics given in 3.211 have undergone a slight change. The num- ber of genera is 87, of which 16 or 18.4 % are endemic, the number of species 146, IOI or 70% being endemic. The percentage of species found in both is- lands is only 18.6. ELS Limtito.diuicied fs pieces’: 1. Paspalum adistichum LL. SKOTTSB. 3.214. — The identity of P. fernande- stanum Steud. with P. distichum was stated by R. ACEVEDO DE VARGAS. A SUPPLEMENT TO THE PTERIDOPHYTES AND PHANEROGAMZ OF JUAN FERNANDEZ 789 4. Phalaris angusta Nees. Ph. intermedia Bosc. var. angusta (Nees) Chap.; Ph. intermedia, Skottsb. 3.214. — PARODI stated 1.81 that PA. angusta is widely spread in Chile and differs from the North American P%. intermedia. *Phalaris amethystina Trin. Parodi 1.78. — Masatierra: CUMING 1855, REED 1873, according to PARODI. *Briza stricta (Hook.) Steud. — Masatierra: DOUGLAS no. 68! (Herb. Stockholm, det. E. ASPLUND). 15. Vulpia dertonensis (All.) Volk. Festuca, SKOTTSB. 3.215. — This is JoHOw's F. sciurotdes. If, as some authors think, /*. dertonens7s All. is = F. bromoides L., the correct name is Vudpia bromoides (L.) S. F. Gray. 16. Vulpia myuros (L.) C. C. Gmel. Festuca sciurotdes Roth, SKOTTSB. 3.215. — This is JOHOW’s F. muralts. 24. Rumex pulcher LL. SKOTTSB. 3.216. — To this belongs SKOTTSB. no. 1205, listed as R. conglomeratus Murr. (as testified by RECHINGER I. c. 47). “humex Jolrosus Rech. fil. |. c.42. Masafuera: ©. de las Casas: 1908, SKOTTSB. no. 472, as “R. conglomeratus forma?’’; testified by RECHINGER l.c. In the absence of ripe fruit the determination remains uncertain, but it is not conglomeratus. R. foltosus is known from two widely distant localities, Argen- tina: Catamarca, and Chile: Punta Arenas. RECHINGER regards no. 472 as not quite typical. 25. Rumex acetosella L. SKOTTSB. 3.216. Add Masafuera: B. Toltén, Q. de las Casas. LOVE (in herb.) referred no. 184 to R. angiocarpus Murb. with a query, but as ripe fruit is lacking, I don’t suppose that a safe determination is possible. 29. Chenopodium multifidum LL. SKOTTSB. 3.217. Add Masatierra: Pto Ingles. *Phytolacca dioica L. SKOTTSB. 5.168. Masatierra: TENZ no. 38! Det. H. WALTER. * Rubus ulmifolius Schott. LOOSER 2. Masatierra: Plazuela del Yunque, LOOSER 1927. A very dangerous weed. 53. Medicago denticulata Willd. SKOTTSB. 3.219. Exclude V. Colonial, no. 200, which is JW. arabica. 54. Medicago arabica (L.) All. M. maculata Willd., SKOTTSB. 3.219. Add Masatierra: V. Colonial (no. 200). 66. Luphorbia peplus L. SKOTTSB. 3.220. Add Masatierra: V. Anson, LOoosErR! 68. Aristotelia maqui L/Hér. SKOTTSB. 3.220. Add Masafuera: wooded ravine, 1600 ft, CHAPIN no. 1068! (Herb. New York). “A tree 25 ft high, diam. ’ Io in.’’ Unfortunately, the maqui has become well established also in Masafuera. 790 C. SKOTTSBERG *Crinodendron dependens (Ruiz et Pav.) OK. Masatierra: Bock 1932! (Herb. Hamburg). Most likely of recent non-accidental introduction. Cortandrum sativum L. SKOTTSB. 5.170. Masatierra: B. Cumberland, LOOSER ! 80. Apium tenudtfolium (Moench) Thell. A. Zactntatum DC., SKOTTSB. 3.222. Add Masatierra: B. Cumberland near the cemetery, LOOSER! 88. Convolvulus arvensis L. Skottsb. 3.222. Add Masatierra: Pto Ingles. *Artemisia absinthium V.. SKOTTSB. 5.171. Masatierra: B. Cumberland near the colony, LOOSER! Total number of adventitious species reported is 137. Bibliography. ACEVEDO DE VARGAS, REBECA, Contribucién al conocimiento del género Paspalum em Chile. Bol. Mus. Nac. de Hist. Nat. Santiago. XXII. 1944. Baker, J. G. (1), Systema Iridacearum. Journ. Linn. Soc. Bot. XVI. 1877. —-— (2), Handbook of the Fern allies. London 1887. ——— (3), Handbook of the Irideae. London 18092. Bock, C., Descripcién de la inflorescencia del Centaurodendron dracaenoides. Revista Univers. (Univ. catol. de Chile) XXI:1. 1936. Canpo.ie, A. P. pe (1), Genres nouveaux appartenant a la famille des Composées- ou Synanthérées. Archives de Botanique (GuitLEMIN) II. 1933. —— (2), Prodromus syst. nat. regni vegetabilis VI. 1837. CHRISTENSEN, C. & Skotrsperc, C. (1), The Pteridophyta of the Juan Fernandez. Islands. Nat. Hist. Il: 1. 1920. —— & —— (2), The Ferns of Easter Island. Ibid. CiausEN, R. T., A Monograph of the Ophioglossaceae. Mem. Torrey Bot. Club: 19. 1938. Decaisne, J. (1), Note sur un nouveau genre de Chicoriacées, recueilli par M. Ber- tero dans Vile de Juan Fernandez. Archives de Botanique (GuiLLEemmn) I. 1833.. —-- (2), Monographie des genres Balbisia et Robinsonia, de la famille des Composées.. AMM. ASCl Nat. 2%) Ser 1H 1S8izi4. Dieis, L., Iridaceae. Natiirl. Pflanzenfamilien 2. ed. 15a. 1930. Don, D., Descriptive Catalogue of the Compositae in the herbarium of Dr. Gillies. The Philosophical Magazine (London) XI. 1832. Epitnc, C. (1), Synopsis of the South American Labiatae. Fedde’s Report. Beihefte- Ba UXXXY 1935: —— (2), The Labiatae of Chile. Revista Univers. (Univ. Cat6l. de Chile) 22. 1937.. Espinosa, M. R., Un nuevo helecho de Pascua. Rev. Chil. Hist. Nat. 38. 1934. FEUILLEE, L., Histoire des Plantes médécinales... en usage aux... Perou & Chily. journ:=@bserv-" phys. mathém? et botan: Iss. “Paris 1725. FosBerc, F. R., Miscellaneous Notes on Hawaiian Plants 2. Occas. Pap. B. P. Bishop: Mus. Honolulu XVI: 15. 1942. Gay, C., Historia fisica y polftica de Chile. Botanica VI. 1853. Goop, R. D’O., The Genus Empetrum L. Journ. Linn. Soc. Bot. XLVII. 1927. GoopspreD, IT. H., Geographical distribution of the Chilean members of the Genus: Nicotiana. Revista Univers. (Univ. Catol. de Chile) 26:2. 1941. Granam, R. (1), Description of several New and Rare Plants... The Edinb. New Philos. Journ. XV. 1833. A SUPPLEMENT TO THE PTERIDOPHYTES AND PHANEROGAMS OF JUAN FERNANDEZ 791 Granam, R., (2), Libertia formosa. Curtis’s Botan. Mag. 61, tab. 3294. 1834. Grant, A. L., A Monograph of the Genus Mimulus. Ann. Missouri Bot. Garden Ay Ta 4; Guittaumin, A., Camus, A. et Tarpiort-BLor (Mme), Plantes vasculaires récoltées a Ile de Paques par la Mission Franco-Belge. Bull. Mus. Nat. d’Hist. Nat. Paris. Boa SER O11 0136: GUNCKEL, H., Notas botdnicas. Rev. Chil. Hist. Nat. XXX. 1928. Harms, H., Bromeliaceae. Natiirl. Pflanzenfamilien. 2. ed. 15 a. 1930. Herter, W. (1), Index Lycopodiorum. Montevideo 1949. —— (2), Systema Lycopodiorum. Rev. Sudamer. de Botanica. VIII: 4. 1950. Hircucock, A. S., Synopsis of the South American species of Stipa. Contrib. U. S. Nat. Herbarium 24:7. 1925. Hircucock, C. L., A monographic study of the genus Lycium of the western hemi- sphere. Ann. Missouri Bot. Garden XIX. 1932. Hooker, J. D.; Handbook of the New Zealand Flora. London 1867. Hooker, W. J. & Arnort, G. A. W., Contributions towards a flora of South America and the islands of the Pacific. Companion to the Botanical Magazine I. 1835. Jonow, F., Estudios sobre la Flora de las Islas de Juan Fernandez. Santiago 1896. KausEL, EBerHarpD (1), Notas mirtolégicas. Lilloa XHI. 1947. —-— (2), Comentario sobre las Ericadceas y Epacridaceas chilenas. Revista Univers. (Univ. Caté6l: de Chile) XXXIV: 1. 19409. Kiarr, F. W. (1), Monographia generis Sisyrinchium. Linnaea XXXI. 186r. —— (2), Berichtigungen und Nachtriige zu der Monographia generis Sisyrinchium. Nebst Monographie der Gattung Libertia. Linnaea XXXI. 1862. —-— (3), Irideae in Martius, Flora Brasiliensis Fasc. LIII. 1871. Krayina, V., Seven new Polynesian species of Elaphoglossum from the group E. conforme (Swartz) Schott. Studia Botan. Ceckoslov. I: 2. 1938. Linpey, J., Libertia formosa. Botan. Register 19, tab. 1630. 1833. Looser, G. (1), Excursiones a Juan Fernandez. Revista Univers. (Univ. Catél. de Chile) XII: 4 1927. ——— (2), La zarzamora (Rubus ulmifolius Schott) en Juan Fernandez. Rey. Chil. Hist. INat. XXX, 1927. —— (3), Ensayo sobre la distribucién geografica de los helechos chilenos. Rev. de Hist. y Geogr. LXX. 1932. —— (4) Nuevos hallazgos de Yunquea Tenzii Skottsberg. Botan. Miscel. IV. Revista Univers. (Univ. Catél. de Chile). XX: 4—5. 1935. —— (5), Algunos helechos sudamericanos coleccionados por Haenke y descritos por Pres]. Comun. del Museo de Concepcién I: 7. 1936. —— (6), Sobre algunos Helechos Chilenos descritos por el Dr. R. A. Phillippi. Revista Univers. (Univ. Catol. de Chile) XXII. 1937. —— (7), Nueva localidad cerca de Antofagasta del helecho Polypodium Masafuerae Philippi. Ibid, XXVIII: 1. 1943. —— (8), Sinopsis de los Asplenium de Chile. Lilloa X. 1944. —— (9), Los Blechnum (Filices) de Chile. Revista Univers. (Univ. Catol. de Chile) POOKIN: 2. 1047. —-— (10), Sobre un helecho atribuido a la isla de Pascua. Ibid. XXXIII: 1. 1948. —— (11), Comparacion de las bromeliaceas de Chile con las de Argentina y Pert. Revs Childe HistyGeorr mM. ure. 1948: Maxon, W. R., Notes upon Polypodium duale and its allies. Contrib. U. S. Nat. Herbarium 17: 4. 1914. MeEz, C., Bromeliaceae. Das Pflanzenreich IV: 32. Leipzig 1935. Mouina, I., Saggio sulla storia naturale del Chili. Ed. II. 1810. NesseL, H. (1), Die Barlappgewichse. Jena 1939. —— (2), Beitriige zur Kenntnis der Lycopodiaceen. Rev. Sudamer. de Botan. VI. 1940. 792 C. SKOTTSBERG Oxiver, W. R. B., The Genus Coprosma. B. P. Bishop Mus. Bull. 132. 1935. Paropl, L. R. (1) El género Phalaris en Chile. Rev. Argent. de Agron. 6: 2. 1939- —— (2), Revision de las Gramineas australes americanas de género Piptochaetum. Rev. Museo de La Plata, Botan. VI. 1944. Pouipr!, R. A., Plantarum novarum chilensium centuria VI. Linnaea XXX. 1859. Porppic, E., Fragmentum Synopseos Plantarum phanerogamarum in Chile lectarum. Leipzig 1833. RECHINGER FIL., K. A., Die siid- und zentralamerikanischen Arten der Gattung Rumex. Arkiv for Botanik 26 A:3. 1933. ReEICHENBACH, H. G. L., Iconographia botanica exotica. Leipzig 1827—32. Rinpe, Astrip, Margyracaena Skottsbergii Bitt. Meddel. G6teb. Botan. Tradg. XIII. 1939. RosspacH, Ruru P. (1), Spergularia in North and South America. Rhodora 42. 1940. —— (2), El género Spergularia (Caryophyllaceae) en Chile. Darwiniana 6:2 1943. Rouy, G., Floresde Erance XiMGiiome: SKOTTSBERG, C. (1), Studien iiber die Vegetation der Juan Fernandez-Inseln. K. Sven- ska Vet. akad. Handl. 51:9. 1914. —- (2), The Phanerogams of Easter Island. Nat. Hist. IH:2 1922. —— (3), The Phanerogams of the Juan Fernandez Islands. Ibid. —-— (4), Einige Bemerkungen iiber die Alpinen Gefiasspflanzen von Masafuera (Juan Fernandez-Inseln). Festschr. Carl Schréter. Verdéffentl. Geobot. Inst. Riibel 3. Lo2Ge —-— (s), Notes on some recent collections made in the Islands of Juan Fernandez. Meddel. G6teb. Botan. Traidg. IV. 1929. —— (6), Greigia Berteroi Skottsb. and its Systematic Position. Ibid. XI. 1936. —— (7), Die Flora der Desventuradas-Inseln (San Felix und San Ambrosio). Géteb. K. Vetenski 8 Vitterhyesamhy Handle sersuge Bisheng 271. —— (8), On Mr. C. Bock’s collection of plants from Masatierra (Juan Fernandez), with remarks on the flowers of Centaurodendron. Meddel. G6teb. Botan. Tradg. Xs 1938: -—— (9), Peperomia berteroana Miq. and P. tristanensis Christoph. Ibid. XVI. 1946. —— (zo), The Genus Peperomia in Chile. Ibid. XVII. 1947. —-— (11), An unnecessary name-change. Svensk Botan. Tidskr. 45. 1951. SLEUMER, H., Revision der Gattung Pernettya Gaud. Notizbl. Bot. Gart. u. Mus. Ber- lin-Dahlem. XII N. 115. 1935. SmitH, A. C., The American species of Drimys. Journ. Arnold Arbor. XXIV:1. 1943. SPRENGEL, C., Caroli Linnaeai Systema Vegetabilium. Ed. XVI, curante — —. I. Gét- tingen 1825. Stapr, O., Calystegia tuguriorum. Curtis’s Botan. Magazine CLII, t. 9136. 1928. STEUDEL, E. G., Synopsis plantarum glumacearum I. Stuttgart 1855. Tenz, O., Nach der Robinsoninsel. Andina III. Valparaiso 1924. Trintus, C. B., & Ruprecut, F. J., Gramina agrostidea II]. Mém. Acad. Impér. des Sciences Saint Pétersbourg Ser. 6, Sci. nat. V. 1849. WopeEuHousE, R. P., Pollen Grains. New York 1935. Explanation of Plates. Pl. 55, figs. 1—11: “fruits” of Stija mucronata H. B. K. For key, see Table I, p. 772, — 6 x. Pl. 56, figs. 12—22 and Pl. 57, figs. 23—30: “fruits” of Szifa meestana Trin. et Rupr. For key, see Table II, p. 773 — 6 x. Pl. 57. figs. 31—34: seeds of Sfergularia, 25 x. 31, 32 S. confertifiora, 33 var. polyphylla, 34 S. masafuerana. Printed 7/11 1951. Nat. Hist. Fuan Fernandez and Easter Isl. Vol. 17. PEATE ek. Photo M. Ferm. aes ~ a SSowe , FAW Nayege ee i “AHO Ys Pie epee 56. PLATE Vol. LT. /s/. and Easter Fuan Fernandez Nat. H1st. Photo M. FERM. Wd ; - i a ! ny Nat. Hist. Fuan Fernandez and Easter Isl. Vol. 1/1. PLATE wt a | Photo M. FER. ue a) a ate ~ : ’ he a bow eo oy 29. The Vegetation of the Juan Fernandez Islands. By C. SKOTTSBERG. With Plates 58—116 and 2 Maps. Thirtyfive years have passed since the modest research expedition known as “The Swedish Pacific Expedition 1916—17"’ was brought to a close. While the collections were taken care of for taxonomic study by myself and a considerable number of collaborators, a description of the plant communities had to wait until these studies had been finished. In 1942, the last of the series of papers dealing with the Cryptogams, which are of paramount physiognomical significance in these islands, was published. Long before that I had begun to arrange my field notes, but for various reasons not to be dwelt upon here I was prevented from returning to this matter until September, 1951. It cannot be denied that this long delay has had an unfavourable influence on the quality of my paper; it should have been written when impressions were still fresh and every detail in my diary and field notes stood out with perfect clearness. This memoir completes the Botany volume, and I use this occasion to pay a tribute of gratitude to my wife and companion, Mrs. INGA SKOTTSBERG, who devoted herself to our work with extraordinary energy and never failing interest. After our return to the mainland from a visit to Easter Island, the third member of our party, Mr. (now Dr.) K. BACKSTROM, paid a second visit to Masatierra in August, 1917. In order to get an idea of the winter stage of the trees and shrubs I had asked him to collect suitable material and to preserve it in alcohol, and this he did to my entire satisfaction. I am also indebted to him for the use of his negatives for reproduction. During our permanence in the islands sketch maps of the distribution of the principal plant communities were made, using the available but rather unsatis- factory charts and maps with such corrections as we were able to furnish. In 1952, the Chilean Air Force undertook an aérial survey of Masatierra and Santa Clara on my request and most generously put the result at my disposal, allowing me to improve the older maps. I want to express my deepest gratitude for this invaluable assistance. In this connection I also tender my very best thanks to the Swedish Minister in Chile, Mr. FOLKE WENNERBERG, and to Mr. BERTIL FRODIN of Santiago, who carried out the negotiations with the Chilean author- ities. The final compilation of the vegetation maps is due to the skill of the Cartographic Institute, Stockholm. If not otherwise stated, the photographs were taken by the author, who also is responsible for the drawings. Riksmuseum, Stockholm, September 1952. 53 — 516796. The Nat. Hist. of Juan Fernandez and Easter Isl. Vol. II. 794 C. SKOTTSBERG Introduction. In his book ‘“‘Crusoniana’’ THOMAS SUTCLIFFE indicates 1572 as the year when the islands were discovered by the Spanish navigator JUAN FERNANDEZ. Generally 1563 is quoted, but in cases also 1574. Be this as it may, man’s first contact with the virgin soil had its consequences, and when, some time after the discovery, Fernandez settled on Masatierra, he brought goats. A number of food- | plants must have been introduced by him and undoubtedly also some weeds. His stay on the island is said to have been short, but about 1580 he was succeeded by a small Jesuit colony which was abandoned in 1596. Cumberland Bay was, I presume, the scene of these as well as of most later events. In 1616 Masatierra was visited by the navigators SCHOUTEN and LE MAIRE, who found that not only goats but also pigs were plentiful (Johow 2 p. 8). L’HEREMITE, who spent several days in Cumberland Bay with his squadron ir 1624, mentions the abundance of sandalwood, the utility of certain trees for ship- building purposes, the occurrence of quince, etc. Forty years later Father DIEGO DE ROSALES planned to colonize the island and collected information on some of its products, e.g. the sandalwood, the chonta palm etc.; evidently these trees had already been cut and the timber exported, at least the sandalwood. The following period, aptly named the era of the buccaneers, must have meant a serious encroach on the forests within easy reach of Cumberland Bay. On re- peated occasions sailors were left behind and managed to make a living without much help. In 1686, several members of Captain DAVIES’s crew settled on the island. They were taken away by Captain STRONG, who also visited Masafuera, where goats were found to be plentiful. Follow the adventures of DAMPIER's men, among them ALEXANDER SELKIRK, whose life on Masatierra has been described so often that we need not dwell upon it here, all the more as his destructive influence on the vegetation had little significance. It should be men- tioned that in the meantime the Spaniards had introduced dogs on both islands in order to exterminate the goats, which were a source of fresh meat welcome to the British buccaneers. Even if this attempt did not meet with anything like a complete success, their number was greatly reduced at least on Masatierra, as testified by Anson’s men. GEORGE ANSON’s narrative, edited and published by his chaplain, RICHARD WALTER, is accompanied by maps and illustrations and gives valuable information on the status of the vegetation in 1740. Maps and views show forests down to the sea on the north side of the eastern half of Masatierra. Pl. XVII represents a view of Cumberland Bay and surroundings; there is little forest left right round the bay on the lower slopes, but both W. and E. of the harbour there are trees all over the place — the west side of Cerro Centinela, nowadays completely tree- less, is presented as well wooded. The section W. of Juanango is indicated as barren land; most likely it has never been wooded. On the large map of Masa- tierra and Santa Clara (called Goat Island), also the latter is wooded, but it is hard to believe that there were other trees than scattered specimens of Dendro- servis litoralis and Rea pruinata, which must have been exterminated by the goats THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 795 with the exception of a few stunted specimens growing out of reach on the coast cliffs, where they still linger. The forest of Masatierra was composed of many kinds of trees, and Walter particularly mentions three, a myrtle, a ‘“‘pimiento”’, and a cabbage tree; this, the palm ¥uanza, was scarce, at least in places of easy access. The myrtle was not, as might be expected, Myrceugenia (Nothomyrcia) Jernandeztana, which forms the bulk of the forest, but according to Johow (p. 11) Fagara mayu, whereas the pimiento was Nothomyrcia, the luma of the present inhabitants. /agara, now called naranjillo, was the largest tree observed by Walter, who tells us that it measured over 40 feet in height and furnished all the lumber needed by the expedition. In the open country, whether cleared or originally treeless, introduced plants such as turnips, radishes, sorrel (Rumex aceto- sella) and wild oats (Avena barbata) were abundant, Avena covering many acres. Anson brought seeds of several fruit trees and vegetables, but they never spread much outside the Bay. He established his headquarters in the beautiful valley later named for him, according to my belief the place where, before him, Selkirk had his base camp. The valley floor, watered by a stream, must have been well cleared and remains an open meadow, stocked with introduced herbs and grasses. Anson also paid a visit to Masafuera. Earlier visitors had described this island as barren, but Anson’s picture of the northeast side shows that save for the high coast cliffs, the slopes were forested. The treeless highland was not in sight. Dec. 1742—Jan. 1743 ANTONIO DE ULLOA lay at anchor in Cumberland Bay. His description of the vegetation is brief and his illustrations inferior to Anson’s, but it may be worth while to quote a couple of passages. P. 284 we read: “En sus Canadas hay mucho Monte, y Arboledas de Maderas muy buenas, entre las cuales se encuentran Arboles de Pmzenta semejante a la de Chiapa en Nueva Espana. En los Llanos, y Colinas crece mucho una Paja parecida a la de Avena, tanto que si entre una, y otra hay alguna diferencia, que haga distintas las es- pecies, es muy poca; y crece tanto, que sobrepuja en altura con mucho a la de un Hombre.”’ And p. 285: “El Territorio de esta Isla se compone de Montanas de mas que mediana altura, y sus Faldas correspondientes acia la parte de Norte son las que estan pobladas de Arboledas, cuyas Maderas, como ya se advirtio, son proprias para muchos usos; las que corren a la parte del Sur no tienen Arboles, sino es en las Quebradas, que forman las Colinas de los Cerros; y es sin duda, porque la fuerza de los Vientos Swrves continuos no los dexa crecer; pero si la Avena, 0 Paja semejante a ella, con la cual se cubren todos viciosa- mente. Entre los Arboles no hay ningunos Frutales ...’ On the map (PI. IV), however, the entire island, with the exception of the extreme western promontory, is covered with trees. Santa Clara is barren. After the publication of Juan and Ulloa’s work in 1748 and until Mrs. Graham published her journal in 1824, no visitor’s report on Masatierra has come to my knowledge. Johow (p. 14) refers to some notes on both islands published in 1767 in a book written by an anonymous author and which I have not been able to consult. It is of interest to learn that Masafuera is described as covered with fine forests. In Masatierra, great changes must have taken place in the meantime. The Spaniards having taken formal possession of the island, a penal colony was 796 C. SKOTTSBERG established in 1750 which, with short interruptions, existed until 1814, and a number of free colonists also settled in Cumberland Bay. In this as well as in the valleys of Puerto Frances and Puerto Ingles the lower slopes were cleared by forests fires; in 1816 the Cumberland (Colonial) valley was again ravaged by fire (Johow p. 17), and the number of introduced plants must have increased. No foreign vessel ventured to enter the harbour, but as the Spaniards did not hold Masa- fuera, which was left uninhabited and lacks harbours, this island was visited by the enemy on certain occasions. G. W. ANDERSON and HAWKESWORTH report on the visits to Masafuera in 1765 by Commodore (later Admiral) ByRON and by Captain CARTERET. I shall quote some of their remarks on the scenery and vegetation. a) From Byron’s voyage. “The valleys have a beautiful verdure, and their sides are full of trees from the top to the bottom” (Anderson p. 235). “This island is very high, and the greater part of it covered with wood; but towards the north, where I lay, some spots seemed to have been cleared, upon which numbers of goats were feeding, and they had a green and pleasant apperance’’ (Hawkesworth p. 86). b) From Carteret’s voyage Anderson publishes a drawing of the north side, with Buque Varado well in sight, and the valley below Cerro Verde, all rather exact. He speaks of the “impenetrable thickness of the woods’, and tells that “we saw severel leaves of the mountain cabbage’ — presumably Dendroseris macrophylla. Hawkesworth, who published the same picture, “NW side of Masafuero’’, relates Carteret’s vivid description of the sudden rise of the streams after a heavy rain: ’I enquired the Lieutenant after the people on shore, and he told me, that the violent rain that had fallen in the night, had suddenly brought down such torrents of water through the hollow or gully where they had taken up their station, that they were in the utmost danger of being swept away before it, and though with great difficulty they saved themselves, several of the casks were intirely lost’ (p. 325). The existence of extensive forests during the latter half of the 18th century is well testified. It has been the privilege of our time to destroy them. In January 1823 Mrs. MARIA GRAHAM spent a few days on Masatierra, and l.c. 347—353 she gives some notes on the geography and botany of this island. The native plants of interest were presented to her by her companion, lord COCH- RANE; among those he brought from Portezuelo she mentions “giant fuscia” (fuchsia), ‘‘andromedas, and myrtles’, but especially ‘“‘a lovely monopetalous flowering shrub; the leaves are thickset, shiny. green, the flower and berry of the richest purple’ — undoubtedly Rhapithamuus venustus. The ““andromedas”’ were of course Pernettya rigida, the myrtles very likely Uguz Se/kirkiz, but nothing even distantly resembling a fuchsia has ever been collected in the islands. Appar- ently she never collected either the large-flowered Wahlenbergia, named for her by Hemsley, or Escallonia Callcottiae (Mrs. Graham later became Mrs. Callcott). She did not extend her walks beyond the colony and the Anson valley. The seat of the former Spanish settlement had lost its native flora, but introduced fruit trees and naturalized vegetables were plentiful, near the sea the ground was covered with radishes and wild oats, and along the trail to Anson valley were THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 197 many European plants. Anson’s camp-site was easily recognized. A document, discovered by Mrs. Graham in the Secretariat of the Capitania general in Santiago and published by her as an Appendix, is of some interest. It includes a list, compiled at the end of the 18th century, of useful plants of Chile, in which the following trees, forming the forest of Masatierra, are mentioned (Latin names supplied by Johow p. 15): Drimys Winteri var. confertifolia (=D. confertifolia), Suania australis, Eugenia (= Nothomyrcia) fernandesziana, Sophora tetraptera (= S. fernandeziana), Berberis corymbosa, Zanthoxylon (=Fagara) mayu, Psychotria (= Coprosma) pyrifolia, Robinsonia gayana and thurifera, Santalum fernandezianum, and Rhaphithamnus longiflorus (= Rh. venustus), the majority not mentioned in earlier reports. The first botanist to visit Masatierra was DAVID DOUGLAS, accompanied by Dr. SCOULER, a geologist, in 1824. On their way round Cape Hoorn and north along the coast of Chile Masafuera came in sight, “appearing like a dark bare rock. We passed near enough to ascertain that it was far from being fertile, though a little verdure might be descried in the valleys, with some stunted trees on the hills, and a few goats browsing on the rocky clefts (I. c. 84). No landing could be effected, and they bore away for Masatierra, where they made a short stay. All Douglas has to say about the vegetation is that the hills were clothed with verdure to their summits. Only few genera are mentioned, but Douglas used his time well and collected 70 species of vascular plants. Of fundamental importance was the research carried out by CARLOS BERTERO in 1830. He was an indefatigable field worker and a keen observer, who trav- elled all over the island, and he was the first visitor who paid attention to the cryptogams. Had he not met with an untimely death — after a collecting trip to Tahiti he left this island for Chile, but was never heard of — he would have been able to give the first real description of the vegetation belts. The little we know about his excursions is contained in a letter published by Guillemin (I. c.) and referred to in some detail by Hemsley l.c. p. 4—6 and by Johow 2. 21—23. Bertero gives a good idea of the composition of the forest, pointing out the pro- found difference between Masatierra and the opposite coast of the mainland, while the resemblance with south Chile was obvious. The principal constituents of the forest were Drimys, Fagara and Nothomyrcia. No living Santalum was found but many pieces of wood more or less deeply buried in the soil. The native palm was not uncommon, but of difficult access: ‘‘Un palmier habite les pentes des montagnes les plus élévées ... on pourrait en tirer une grand partie s'il était facile de se le procurer.’’ Many species, most of them discovered by Bert- ero, are mentioned, and in cases remarks added on their mode of occurrence and properties, but little is said about the character of the different parts of the island; Bertero only wanted to give the reader an idea of his most notable dis- coveries. During Bertero’s sojourn on Masatierra, Captain PHILIP PARKER KING, com- manding H. M.S. Adventure, arrived in Cumberland Bay and spent a week at the island. King tells that myrtletrees abounded but that no living sandalwood was found, while plenty old wood was collected. The commonest herbaceous plant 798 C. SKOTTSBERG was ‘“‘a species of oat, which grows very luxuriantly, and towards the westward covers the ground for many miles.’’ Bertero accompanied King and assisted him in collecting. King remarks l.c. 305 that Bertero considered the vegetation to be ‘‘very little like the Chilean, but to partake more of that of California’; in his report Bertero says that he had found some plants also known from Cali- fornia. After his visit to Masatierra in 1832 CLAUDE GAY published an account of its flora and fauna, justly characterized by Johow (2. 44) as “mui fantdstica i llena de inexactitudes’”. A visit in June, 1838, by DUMONT D’URVILLE and JAC- QUINOT in command of a French expedition added nothing to our knowledge of the vegetation. Although Jacquinot believed that no naturalist had visited the place before, he only spent 21/, hours on shore and contented himself with the impression that “‘les productions de cette ile me paraissent étre tout a fait ana- logues a celles de la céte de Chili’ (Il. c. 336). The surroundings of the harbour with their fertile alluvial soil were almost entirely covered with Rhaphanus sa- tivus, but at some distance inland groups of trees were seen. End of October, 1854, FILIBERTO GERMAIN collected on the islands and made some interesting discoveries on Masafuera, where HUGH CUMING had been the first botanist, in 1830. Germain’s collections gave, as it were, birth to R. Ne PHILIPPI’S review of the flora of the archipelago (l.c.). His own short visit to Masatierra in 1864 (4 days) added few novelties but nevertheless is of particular interest, because the maqui (Ar7stotelia maqut) was for the first time stated to occur on Masatierra, very likely brought there on purpose — the berries are edible — and rapidly spread by the thrush. Its introduction into the defenseless native forest was the greatest calamity registered in the history of the vegetation, as well expressed by Johow 2. 108—10g. In spite of being cut for fuel it is gaining ground and has spread to Masafuera. Another weed, also discovered by Philippi in 1864, was Acaena argentea, which has become a serious pest in Masatierra, thanks to its barbed achenes carried far and wide by man and beasts and not only transforming the ‘‘avenales’” into an Acaenetum, but also encroaching upon the native grass communities and the open rocky ridges several hundred m above the sea. It may be said to its credit that, in some measure at least, it helps to check erosion and that it forms an interesting hybrid with Margyrz- carpus digynus. \ believe that it was introduced to Masafuera with the convict settlement in 1909, for I did not see it there in 1908 but found it well established in 1917. It has not been reported from Santa Clara. In November, 1875, H. M.S. Challenger remained 3 days in Cumberland Bay. In spite of the short time the naturalist, Mr. H. N. MOSELEy, brought to- gether a considerable collection of plants. His report, copied by Hemsley l.c., contains little information on the general character of the vegetation, but is not without interest. Of importance is the statement that “everywhere for the first few hundred feet, trees are absent, the wood having been all felled,’ — this applies to Cumberland Bay — and when ascending the trail to Portezuelo, Mose- ley observed that “‘the first tree was met with at about 700 feet altitude, all below had been cut down;” thus, already 75 years ago, the lower boundary of THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 799 the native forest lay at about 200 m above sea level. Remarkably enough the maqui is not mentioned; 30 years later the same trail led through a dense macal before the ascent of the ridge begins. The maqui had not yet spread enough to make itself conspicuous, and Moseley did not collect it. Ywanza was said to have become ‘‘almost exterminated’; this is true of the mountains encircling Cumber- land Bay, but not of the island in general. In 1885, the first, but unfortunately not the last tourist excursion to Masa- tierra was arranged. It gave rise to two publications; one, by E. T. CAVIEDEZ, a newspaper man, is deemed by Johow to be superior to most other popular narratives on this subject, but it lacks all scientific interest. The other, by one ALEXANDER ERMEL, a German merchant, is, according to the same source, of no value; worse than that, it is full of nonsense. This is not the place to deal with the author’s geographical speculations, but a number of his botanical “observa- tions’’ should not be passed unnoticed. The forest is said to be formed by 3 species of trees, Hagara making up 90%, Drimys and Coprosma pyrifolia 5% each; the dominating tree, Mothomyrcia, is not even mentioned except as one of the several shrubs enumerated and including all the other forest trees. We are told that the Latin name of the native palm is Fabaea spectabilis (the only palm on the mainland of Chile), that it once formed forests but only succeeds under the shade of other trees, that the pith(!) is edible and the cluster of fruits so heavy that one person can hardly lift it, and that the fruit is superior in taste to hazel nuts and pine seeds — the mesocarp is edible, but quite insipid. The particulars referring to Gunnera, ragaria, Acaena etc. are equally absurd. FEDERICO JOHOW, who visited the Islands in 1891—92 and 1895 (together 21/, months) and in 1896 published a comprehensive account of fundamental importance, is the first scientist who attempted to distinguish definite plant communities (pp. 246—255 and 257—262). On Masatierra two forest types were distinguished: a) “las selvas altas i sombrias que llenan el interior de las grandes quebradas’’, and b) “los bosques que cubren la serrania de Masatierra, rejion formada por un gran numero de crestas, cumbres i precipicios que no dan cabida al crecimiento de Arboles grandes i en los cuales el sol penetra facilmente por entre la vejetacion.’’ This latter type includes a much greater variety of species, and with this he also classified “‘los claros i orillas de las selvas altas i las cercanias de los riachuelos que corren en el fondo de las quebradas’’; they deserve, however, to be treated separately. Lists of the trees and shrubs of the two types are furnished, with indication of the frequency from social to rare. On Masafuera only one kind of real forest was distinguished, covering the valley slopes and corresponding to type a) from Masatierra, but besides this another, very different kind, occurring in depressions on the highland and formed mainly by treeferns. The climbing plants, epiphytes, and herbs and grasses of the forest floor are discussed at some length; the latter, mainly ferns, were grouped according to their light-demands. No other cryptogams were considered. Johow further mentions that the exposed ridges of Masatierra are covered with low brushwood with Pernettya rigida and Halorrh- agis alata (=masatierrana) as leading species. A second group of communities characterizes the lower, treeless western Soo C. SKOTTSBERG section of Masatierra and the country that has lost its forest cover through hu- man action. With regard to the mountainous, still well wooded east and central part, Johow is convinced that the slopes toward the sea-shore, which plunge into the water in the shape of a more or less high cliff wall, were always destitute of trees, and that, considering the scanty rainfall of this basal region, there was a maritime timberline in about 100 m altitude. It goes without saying that there was never a forest but possibly a few dwarfed trees in cracks on the vertical barrancas — Kea pruinata is still found in such situations — but a fringe above the top of the wall may also have been barren and covered, as Johow says, by some kind of “steppe” or “‘meadow’’, and he gives a list of species which would have formed this community, evidently a mixture of several sociations. Very likely he drew his conclusions from observations made along the south side of the island. A third chapter is devoted to the vegetation of the sea-shore. Johow points out that topography and soil conditions are very unfavourable to plant life in general, and that only few species are able to thrive. All are scattered colonists; the only halophyte forming closed stands is Salscornia peruviana. From his list p- 259 I would like to exclude Spa bicolor (=neestana), Piptochaetium (= Stipa or Nassella) laevissimum and Ochagavia elegans, and, of the ferns, Adzantum and Blechnum australe (=auriculatum), for even if they are found near the sea in cases, they are much more common and much better developed inland. Instead, some species should be added, as will be seen later. Chapter 4 contains a general description of the vegetation of the narrow west- ern part of Masatierra and of Santa Clara. The list of species does not differ materially from the one just mentioned and is a mixture of inland and typical seaside plants. Chapter 5 is a very short and incomplete description of what Johow terms the “estepa de helechos’’ (fern-steppe) of the highland of Masafuera. As Johow never went more than half-way to the highest ridges, not one of the many Subalp- ine and Alpine species was seen by him. His species list would reflect the vegetation near the timberline, if Dzzmys had not been excluded. About a year after my first visit to the islands in 1908, professor OTTO BURGER, a former resident of Chile, published a book on Juan Fernandez, its geography, history and natural products, which were familiar to him because he had visited Masatierra. His description of the vegetation did not further our knowledge of the plant life and contains some erroneous statements. In company of my friend Cand. phil. (later Prof. Dr.) PERCY QUENSEL, a prominent geologist who has contributed important papers on the geology and petrography of Juan Fernandez, I] spent 4 days on Masatierra and 2 days on Masafuera in August, 1908. Some notes on the distribution of the vegetation and on the last living Savtalum were published in 1910 (z, 2) and in my paper of 1914 (3) pp. 55——-71 were devoted to an analysis of some of the plant com- munities, based on a survey of a number of different localities, the first attempt of its kind, because Johow’s descriptions do not refer to any particular spot; and for the first time the cryptogams were, to some extent at least, taken into THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS Sol account. Attention was also paid to the ecological factors, and the reasons set forth why the flora should not be called subtropical, as Johow did, but warm temperate. Special attention was given to bud protection. Schimper’s theory on the biological significance of the rosette-tree life-form was discussed and re- futed. After our return from the expedition of 1916—17, a number of preliminary reports were published (y—7), and in 1920 the botanical volume of the present work began to appear in parts. With regard to purely taxonomical papers I refer to the Bibliography. An additional general report of the expedition and its results appeared in 1924 (zo). A number of small papers on the ecology and sociology of the vegetation were published from time to time, and some of them are briefly summarized here. In 1925 (zz) I gave an account of the Alpine flora of Masafuera, its relations, growth forms and geographical elements. In a paper read before the Ithaca botanical congress in 1926 (z3) I expressed my opinion that some means ought to be found to make an analysis of exotic vegetation more intelligible to a reader not at all familiar with the flora. I suggested using a system of wellknown habit-forms with which the species mentioned in describing the composition of a community could be compared. The method was illustrated by the principal communities of Juan Fernandez; all the vascular plants were referred to a habit type represented by some wellknown life-form or genus, with the leaf size added according to Raunkiaer. I don’t suppose that anybody paid much attention to this proposal, and I shall not use this method here, because most of the species are illustrated. An abstract of a lecture on the plant com- munities has also been published (z6) and, finally, a condensed account dealing with the same subject (20). On Masatierra we made a detailed survey of the sociations of Bryophytes found in the old convict caves, with special regard to the light conditions (77). I shall conclude this summary by calling attention to recent publications by other authors. E. L. BRANCHI gave a popular account of the geography with an appendix on the vegetation, entirely based on the writings of Johow and Mo- seley. G. LOOSER reported on his two visits to Masatierra, and gave a general account of the plant life. He made the rather unpleasant discovery of Audus ulmifolius Schott, a most unwelcome new arrival: “Esta es la unica novedad que he descubierto en Juan Fernandez ... debe haber llegado en los ultimos 10 afios ... Es poco abundante, pero mucho temo llegue a ser un enemigo terrible para la flora indigena’’ (p. 391). Some time ago Mr. Looser told me that this serious pest was reported to have increased very fast. Unfortunately, the fruit is edible. Finally, Miss FILOMENA RAMIREZ condemned the illegal traffic in palm- wood, living treeferns etc. which threatens the native forests with destruction. 802 C. SKOTTSBERG P Botanical excursions in 1916—17. Masatierra. (Headquarters established in the factory of Recart y Doniez, Bahia Cumberland, Dec. 1, 1916.) » » » » (» » » » » » » 1916, Dec. 3. Tis 12. Excursion to Portezuelo de Villagra. — 9 hrs. P. M. Via Valle Anson to north face of Cerro Damajuana, alt. 250m. Excursion to Quebrada Gutierrez, a gully at the foot of the Portezuelo ridge, alt. 450 m. Lower slopes logged, ox-teams dragging out heavy luma trunks. — 8 hrs. Along trail to Puerto Ingles to crest of Cordon Salsipuedes, which was followed to about 600 m alt.; thence across Quebrada Salsipuedes to Loma de los Mufiozes and into a small hanging gully, belonging to the drainage system of Valle Ingles. Finest Dzcksontetum seen in Masatierra, 600-650 m alt. Return across Cordon Salsipuedes. — Io hrs. P. M. Along the coast trail to entrance to the Pangal gorge. Survey of the badly eroded slopes near the settlement; reconnaissance of the con- vict caves. Along trail to Puerto Frances and back. — 11 1/2 hrs. P. M. To Portezuelo de Villagra. — 6 hrs. Ascent of ridge W. of Pangal valley to about 360 m alt. — 6 hrs. Up Valle Anson and climb of a ridge between La Damajuana and El Yunque to about 575 m. — 10 £/2 hrs. Cordén Salsipuedes; survey of small hanging gully (fourth S. W. of Punta San Carlos), thence across ridge to head of Quebrada Salsipuedes, about 500 m alt. — .6.1/2,hrs. P. M. E. slope of Salsipuedes to 3d gully S. W. of Punta San Carlos, visited by me in 1908, then with a patch of fairly good forest, now ruined and invaded by weeds. P. M. To Portezuelo de Villagra. Survey of cliff walls on both sides of pass. Eryngium inaccessum discovered, rediscovery of E. fernandezianum first found in 1908 and which turned out to be a cross between duplewroides and inaccessum. Two plants, apparently the same I found in 1908. — 5 hrs. P.M. To Portezuelo in continuation of the work the day before. — 6 1/2 hrs. P. M. To Valle Anson and Plazoleta del Yunque. Adventitious flora around the settlement in Valle Colonial. A. M. Analysis of the vegetation in one of the old convict caves. P.M. Woods of the Damajuana ridge, alt. about 435 m. — 5 hrs. Survey of Pangal gorge to the cascade, alt. 200 m. Chenopodium crusoeanum discovered. — 8 hrs. Across Portezuelo to the Villagra valley, where camp was established in about 175 m alt. Excursion up a dry, wooded gully to about 400m alt. — g hrs. Along the treeless south slope of the main ridge, west to Punta Larga and back. — 10 1/2 hrs. P. M. To foot of Cerro Chumacera and the most westerly forest patches on Masatierra. To Plano del Yunque, Cerro Negro, c. 190 m alt., and to the shore along foot of El Yungques — 6/2 hrs: Return across Portezuelo to the settlement. Across Portezuelo and survey of wet rainforest covering the high slopes of Villa- gra, 450-550 m alt. — 9 1/2 hrs. Headquarters moved from Casa Recart y Doniez to a small house in the sett- lement.) 1917, Jan. 13. » » » » » » » » » » » » » » » » » » 15. 16. 18. 19. 10. II. 12. 13% 14. ite 16. Lye 18. 19. THE VEGETATION OF .THE JUAN FERNANDEZ ISLANDS $03 Across Cordon Salsipuedes to the Dicksonia gully above Puerto Ingles (see Dec. 8), which was followed down the steep slope, returning to the settlement. — 9 3/4 hrs. By boat to Bahia del Padre; survey of the bay and of the Puente. Camped on the beach. The bay is one of the few good places for collecting marine algae. Returned to Bahfa Cumberland. To Quebrada del Monte Maderugo (lower slopes logged); ascent of Cordén Cen- tral and Pico Central, c. 570m alt. — 8 '/a hrs. By boat to Puerto Ingles; ascent of the central ridge of the valley to about 575 m alt. — 7 hrs. — Camped for the night in the so-called Robinson cave. Survey of coast rocks; excursion into the west branch of the valley. P. M. return by boat to Bahfa Cumberland. To Portezuelo and north face of Cerro Piramide. — 9 hrs. Interior of Quebrada del Monte Maderugo to foot of precipice, about 380 m. By boat to Santa Clara and Morro del Spartan (de los Alelies), returning to Bahia del Padre, where we spent the night. Return to Bahfa Cumberland. Ascent of the Damajuana ridge to N. E. cliff wall, about 500 m alt. Primeval forest with old, thick-stemmed lumas. — 8 hrs. To Portezuelo. Discovery of male plant of Rodbénsonia thurifera. — 4 1/2 hrs. Left for Masafuera, Masafuera. Arrived under the island; landing at the deserted penal settlement in Casas im- possible. Anchored in Ensenada Toltén. Vegetation of coast slopes. Tried to reach Quebrada Sanchez, from where Casas can be reached scrambling along the coast rocks, but were beaten back. Landing impossible also in Toltén. Heavy breakers all along the coast, landing impossible; left for Masatierra, where we arrived Feb. 6, sailing again next day for Masafuera. Arrived in front of Quebrada de las Casas, where landing was effected and head- quarters established in one of the buildings of the former penal settlement. P. M. excursion south along coast cliffs and into the Vacas canyon. Halorrhagis masa- fuerana discovered. Day spent in the interior of Casas canyon. Berberis masafuerana discovered. Up the zigzag trail (built in 1909 by the convicts and now partly damaged by landslides) to the upper forested slopes of the Mono and Chozas valleys, alt. about 700 m. -—— 6 hrs. P. M. excursion into the Vacas canyon. To the sources of Quebrada Pastos on edge of Plano de la Mona, where camp was established at a small stream in about 1130 m altitude. Ascent of Las Tor- res, about 1370 m alt. — 11 hrs. — New species discovered 14-15 Feb.: Euph- rasta formostssima, Robinsonia Masafuerae, Phoenicoseris regia, Galium masafue- ranum, Evrigeron luteoviridis, Ingae and turricola, Bromus masafueranus and Agrostis masafuerana, several species known before from south Chile and Sub- antarctic America were also found. Ascent of Cerro Correspondencia and along Cordén Atravesado to Cordon del Barril, alt. about 1400 m. Return to camp and descent to Casas. — to hrs. South along foot of coast barranca and round the island to the western shore, where we camped for the night on the Loberia beach. — 6 hours. Climb into Quebrada de la Loberfa Vieja to foot of precipice, about 190 m alt. Uncinia costata and Chenopodium nesodendron discovered. Return along the beach to; Casas; —— ro. hrs. P. M. Along the beach north to Quebrada del Mono. Ascent along zigzag trail leading to Papal; survey of Quebrada del Blindado and Cordoén del Barril to about 730 m. 804 19 > » » » » 17, Keb; D2 Vee 1917 March 18 » » » » » » » 19 25% oo™NI C. SKOTTSBERG Survey of the forests in the Mono and Chozas valleys. — 6 hrs. P. M. South along the coast to entrance to the Varadero canyon, where we camped for the night. Climbed the south canyon wall to 400 m alt., then along a ridge toward Ino- centes, but finding it impossible to force the impenetrable fern forest before dark, we turned back after having reached about 950 m alt., made our descent into Varadero and returned to Casas. — II 1/2 hrs. Short visit to Casas canyon. Across upper Chozas valley to Plano de la Mona, descending to about 500 m along the ridge separating the quebradas of Sanchez and Toltén and returning by the same route to Casas. Solanum masafueranum discovered. — Io 1/2 hrs. A.M. To the innermost accessible part of Casas canyon. P. M. Lichenological excursion along coast barranca to Quebrada del Mono. To Papal, and ascent of the Barril ridge to Cordén Atravesado. — 10 1/2 hrs. Survey of the woods in the Mono and Chozas valleys continued. — 9 hrs. Short excursion to outher section of the Vacas canyon. To the high plain above Chozas and across to the precipice forming the inac- cessible west wall of the island; along “Avenida de las Cabras’”’ to above Buque (Barco) Varado. Ranunculus caprarum discovered. Return to Casas. — 97/2 hrs. Survey of Casas canyon continued. Same route as on March 5 to edge of west precipice and back; north to Cerro Verde. — 11 1/2 hrs. Along the beach south to Playa Ancha; from there ascent of the Inocentes- Vacas ridge to summit of island, about 1500 m alt. Discovery of Adrotanella crasstpes and of additional Magellanic plants. — 11 1/2 hrs. Short visit to Casas canyon. A. M. South along the coast to Quebrada del Varadero. Survey of Casas canyon finished. A. M. Survey of Papal and the Blindado ridges finished, ascending to about 725 m. — 5 hrs. P. M. North along coast to the Ovalo, Pasto and Sanchez valleys. — 4 hrs. Left by boat for Ensenada Toltén; anchored for the night. Sailed for Masatierra. Masatierra. . Arrived in Bahia Cumberland. . P. M. Excursion to Portezuelo. A. M. A rain-gauge and max. and min. thermometers put up at Portezuelo, and read every two days; these visits not mentioned below unless combined with biological observations. . To Portezuelo and back. — 5 */2 hrs. . Excursion to Cerro Centinela and Rabanal. . To Valle Anson and Plazoleta del Yunque. — 5 hrs. . Work around Portezuelo continued. — 6 hrs. . Survey of convict caves continued. . Ascent of the ridge between Quebrada de la Piedra Agujereada and Q. Laura to about 650 m alt., thence descending to about 350 m, E. across Laura and down the ridge between this valley and Q. Pesca de los Viejos, W. from there and back over Centinela. — 11 hrs. . Work in caves finished. . A. M. To Portezuelo. — 4 hrs. 9. By boat to Quebrada Juanango, ascent of ridge separating this valley from Va- queria and down into this gully, returning by boat to B. Cumberland. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 805 1917 April 11. Ascent of Cordén Centinela to the point where it branches off from the main ridge, alt. about 780 m. Rediscovery of Bertero’s “ 77/landsia’, now Hesperogreigia Berterot, — 1o hrs. 12. To Portezuelo and upper slopes of Villagra. — 6 1/2 hrs. 14. To Portezuelo. — 4 hrs. 17. Along trail towards Puerto Frances, up Cordén Chifladores and west branch of Frances Valley to 500 m alt. — I1 hrs. 20. A. M. Portezuelo and north wall of Cerro Piramide. — 5 ?/2 hrs. 21. P. M. Cordon Salsipuedes. — 5 hrs. | 24. Across Portezuelo and down into east part of Villagra; ascent of one of the buttresses of Yunque to foot of a vertical wall, about 500 m alt., returning same route. —/< hrs. 30. Sailed for Valparaiso. CHAPTER I. Environmental factors. Temperature and rainfall. The single meteorological station is situated in the settlement in Bahia Cum- berland where observations were begun in 1901, but the gaps are many, and no dates after 1942 have been available. From 1901 to 1922 inclusive the sta- tion was situated 10 m above sea level, and from 1929 at 6 m, the Subdelegado civil being in charge. The difference in elevation being insignificant, the two series have been treated as one. From June, 1923, to end of 1928, the altitude was 345 m; the locality, not indicated in the Anuario, was, I suppose, the Cent- inela Radio Station. Our only aneroid reading, 355 m, is probably a little too high. The dates available, incomplete as they are, give a good idea of the climate of the basal region on the north side of central Masatierra; see Table I. In 1914 (Skottsb. 3) I published figures for 7 years (1901I—1907) but com- plete for 5 only: mean annual temp. 15°.6C., means of daily max. and min. 18°.8 and 13°.5, average rainfall 1081 mm. Whereas the temperature figures are about the same as in the table below, the difference in rainfall is considerable. This is not surprising, because the variation from year to year was very great during the period 1901—1942, from 317.7 in I9OI to I 697.7 in 1922, both figures being very exceptional, as shown by the graph, Fig. 1. These observations tell us nothing about the climate in the montane region where the rainforest and mountain scrub cover the slopes and ridges, but some conclusions may be drawn from the short series obtained in 345 m altitude, Table II, The figures from the two stations are not directly comparable, because the observations were not taken simultaneously; nevertheless, the difference in temp- erature is clear enough. No reliable conclusions can be drawn from the observed difference in the amount of precipitation. As was mentioned above we put up a small weather station in the Porte- zuelo pass, where readings were taken every second day between March 25 and April 30. The elevation generally given is 550 m, while the average of our 806 C. SKOTTSBERG Table I. Masatierra, Bahia Cumberland, 33° 37’.5 S., 78° 49’.7 W. 10 (6, 1929-42) m above sea-level. 1901-1942, based on 24 complete annual series. ee ——— | Temperature, °C. Rainfall | Mean daily extr. | Mean Mm Days Max. Min. Pijans . 260". - 18.7 21.7 15-0 18.2 4 WHE be tcteretere = 19.0 22.4 16.0 2g SG) 6 Marchi.< <1 18.4 217, 15-5 32.4 8 / yor ailleiecacio oe 16.9 19.8 14.0 81.8 II | May .....%.- «| 15-4 18.1 12.8 146.8 15 (ajune.os2 25. | 13-4 16.0 Ti.2 162.1 17 bi? orate omic | 12.9 14.9 10.2 139.3 15 PUES ROMO CLO 12.4 15.0 9.3 111.6 14 | Sept. ...... 12.6 15-5 9-7 70.5 10 I OLous Bigtece on 13.4 16.2 10.2 44.1 8 IENIOWAGeeclee re 15.1 eros T1252 31.7 6 | IDIXO; Gidioo co 17.6 20.5 13.8 23.5 5 Ste aee ere 15.4 | 18.4 | 12.7 889.71 | TO? 1 21 353-0: 24 = 889.9 2 24OLO | 245 = 27.0 Absolute extremes: 34.0 (March, 1906) 3.3 (June, 1929) Note. Means of annual maxima and minima 20.7 Ba Table II. Masatierra, 345 m above sea-level, 1924-1928; 5 years. Fi ° < | IRS, sododus | Temperature, °C. Rainfall Mean daily extr. Mean Mm Days Max. Min. 16.0 19.8 13.5 209.0 sae) 16.7 20.7 14.0 50.8 9 16.1 10.7 13.6 53.0 13 14.4 17.6 11.8 85.2 15 12.6 15.3 10.3 152.8 17 Tye 13.8 (SH) 158.5 19 10.5 13.0 8.2 214.1 20 9-5 12.5 Te. 87.2 14 10.0 1322 7.5 115.9 14 10.8 14.5 8.4 62.3 I2 12.5 16.0 9.4 34-9 9 14.3 LEO 11.0 37-5 8 Tae 16.2 10.3 | 1081.21 160? 2 800: 5 = I60. Maple Wile 2\ prt, tol 7. a Absolute extremes: 25.5 (Jan., 1925) 2.0 (Oct., 1923) Note. Means of annual maxima and minima 24.9 5:7 | | Meteorological station Portezuelo Means of eds ainfall, Mean temp. | Hn max. min | 16.8 19.1 14.4 73.5 II.6 18.4 8.6 86.1 THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 807 See Os 04-05) (06 11 12 13,144 15°. 16--17° 18 19". 20" 21+ 22) 29°30! 31" 332) 33° 34 35) Seu s7i esos? Fig. 1. Masatierra, Meteorological station in Bahfa Cumberland. Variation in rainfall 1Ig01I—1939. Figures in left margin denote millimetres. No observations 1907—I9Io and 1923—1928. numerous aneroid readings is 590. For comparison between the two stations see Table III. Due to the much cooler nights in the mountains, the average tempe- rature is much lower; the difference in day temperature is very small. With the character of the vegetation in mind I was surprised that the difference in rain- fall was not much greater. Anybody who has lived for some time on Masatierra will tell you that it rains much more in the mountains than near the coast, and our experience is the same; we could see the rain clouds gather over the high ridges while the settlkement had sunshine, and a longer series of observations would, I presume, show a much more pronounced difference between the coast- land and the mountain range. It goes without saying that when the moisture- laden air current from SE to SW meets the high coast and is forced to rise, condensation is at its greatest over the south border of the mountain range. There is another climatic factor of very great importance, dense fog fre- quently covering the mountains above 400—500 m. Between March 25 and April 30 the Portezuelo pass was visited 18 times, and 9 times mist was recorded. The photograph, Fig. 2, of the east part of the south side of the island which rises precipitously from the sea, shows the sharp lower boundary of the cloud- bank; this is a common sight. Returning to the temperature, observations were taken during every excur- sion jointly with the aneroid readings, a condition for working out the altitude. These observations are given in Table IV. We have no exact knowledge of the climate of the lower, barren western part of Masatierra. The backbone of the island, which rises abruptly to 500— g00 m above the south coast from Cape Guasaballena to Cerro Pirdmide, and 808 C. SKOTTSBERG SS BAAic Ar aiGieaie ae AEs AGS AEE eg PUA I OE EAA GEE Fig. 2. South coast of Masatierra from the summit of Cerro Negro, looking East toward El Verdugo (the small detached rock), showing sharp lower limit of fog belt. — Photo 7/1 1917. culminates in the summit aptly named El Yunque (The Anvil), 927 m according to the sketch map published by Johow 2, crosses the wide, central part and then, gradually losing altitude, runs along the NW coast to Bahia del Padre. With the prevailing winds condensation will rarely or not at all take place over the range W. of Juanango. As was pointed out in the Introduction, this section appears always to have been devoid of trees, and the character of the soil and vegetation testifies that the rainfall is scanty. Santa Clara belongs to the same dry district. I shall return to these questions in Chapter. IV. Of the climate of Masafuera very little is known. The island is uninhab- ited; a convict settlement was established in 1909 but abandoned after few years; later, it was occupied a second time for a short period. No climatic rec- ords, if taken, have been published, and in any case they would have told us nothing of the conditions except in a single protected spot close to the beach. As the island is much higher than Masatierra, about 1500 m according to our observations (1800 or even 2000 m are indicated on the charts, but these fig- ures are much exaggerated), climatic vegetation belts are well developed; the highland above approximately 1000 m probably enjoys a cool temperate climate, and if the information obtained from fishermen visiting the island is correct, snow has been observed on the summits, indicating that frosts, unknown on Masatierra, occur. In view of the total absence of regular observations our temperature read- ings for February-March, 1917, may offer some interest. Table V. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 809 Table IV. Temperature observations on Masatierra, 3.12. 1916-24.4.1917 °C. First and last reading for each day refer-to headquarters at sea-level. Date Hour "IC, Alt., m Hour “(6x Hour Ge 3.12 8.30 a.m. 19.0 590 12 noon 13.0 5.30 p.m. 19.5 6.12 2 p.m 15.0 145 I p.m. 14-0 » 250 4 p.m. 12.0 6 p.m. 14.0 7 hie Qa; m: 15.0 455 12 noon 12.5 » 350 Cc. 2 pian, 13.0 5 p.m. 17.0 8.12 6.30 a.m. 17.0 275 8 a.m. 13-5 » 615 I2 noon 15-0 » 670 LE (p. a: 13-0 » 660 G: 2) pei: 12.0 5 p.m. 17.0 E302 7a aly TL. 16.5 220 IO a.m. 17.0 » 45 Ne Ging 190), 19.8 » 300 12 noon 13-5 » 45 75 joe stole 17-0 6.30 p.m. 15.2 14.12 rsp. m. 19.0 590 3 p.m 11.0 7a Dati LO 15.12 1 p.m. 18.0 590 4 p.m 13.0 » 590 5.30 p.m 11.0 af JO), Va 16.0 17.12 2) p.m. 17.5 3605 5 p.m 13.0 7.15 p.m. 15.0 18.12 8 (asa 16.0 380 COG iy ian) 12.0 | » 470 Coy eran 11.5 » 540 G: I pya, Teley » 575 €; 2 p.m 11.0 | » 220 4230) p= mi 14.0 5.30 p.m. 16.0 20.12 Tt Ya). 17.3 435 I p.m 14.5 » 460 3 p.m 267, 6.30 p.m. 16.0 24.12 I p.m. 17.0 320 ir a.m 17.0 » 590 I2 noon 13.0 6 p.m. 15.0 25.12 12 noon 16.0 440 2 ps: 12.5 » 590 1.30 p.m. 12.8 6.30 p.m. 15.8 28.12 VIE G2), 1001 17.5 220) a I2 noon at Gyo 4.30 p.m. 15.0 30.12 I p.m. 18.0 345 By joosnal 13.0 5.30 p.m. 16.0 itil 220 1 Sanam Wz » 200 Typ et 14.8 6 p.m. 15.0 4.1 8.30 a.m, 107/572 590 IO a.m. Ties » 175 I2 noon 16.0 » 175 6 p.m. 14.8 5-1 175 7 p.m. asa » 385 9 a.m. 13.0 » 175 5-30 p.m. 14.0 6.1 175 12 noon 19.0 wel 175 10.30 a.m. 14.8 » 5 2.30 Prin. 18.0 » 190 3.30 p.m. 21.0 » FS 6.30 p.m. 14.4 8.1 175, I pm: 16.6 » 590 3 p.m. 13.8 5 p.m. 18.4 10.1 Io a.m. 21.0 250 iat) fe) 0001 19.2 » 550 I2 noon 20.4 » 590 WAG) 10), 0081, 17.2 54 — 516796. The Nat. Hist. of Juan Fernandez and Easter Isl. Vol II. 810 Table IV (forts.). C. SKOTTSBERG Date 14.4 16.4 18.4 20.4 22.4 24.4 » 26.4 28.4 30.4 Hour 9.45 a.m. ga.m. 1 9 a.m. I2 noon 10.30 a.m. MOWAn ta, 10.30 a.m. 7.30 a.m. Qyalem: 7.30 a.m. | 23.0 23.0 2088 22.0 22.0 20.0 18.7 18.0 15.8 1 Puerto Ingles, at sea-level 2 Vaqueria, at sea-level Alt., m Hour as Hour res So nnn nnn UE EEEEyEIEEEEESEIEE SSEEESEET 510 I p.m 18.0 420 2.30 p.m 19.0 590 5 p.m 16.0 6.30 p.m. 19.0 610 ¥2 noon 20.9 605 I p.m 16.5 575 2 p.m 17.0 6 p.m. 21.0 395 Tk Gp 20015 20.0 540 I p.m 22.0 oe) 3 p.m 22.9 520 4 p.m 18.4 5 p.m. 22.0 200 12 noon 20.3 3 p.m. 25.8 590 11.30 a.m. 12.5 590 3 p.m. 16.5 6 p.m. 18.0 235 2p.m. 17-6 390 Avon. 16.0 5 p.m. 18.5 440 12 noon 16.0 530 2) pi. 15.0 6.30 p.m. 18.0 220 10.30 a.m. 17.9 590 Lis 3 Olay. 14.0 4.30 21.9 590 10.30 a.m. 12.8 5 p.m. 167/73 590 8.30 a. m. 12.0 590 8.30 a.m. 10.8 270 12 noon 18.2 5 p.m. 20.0 590 8.30 a.m. 15.2 590 Il a.m. 14.2 505 I p.m. 15.2 2.30 p.m. 17.3 590 8.30 a.m. 11.4 590 8.30 a.m. 11.8 { 305 8.30 a.m. 14.3 6.30 p.m. 18.0 590 8.30 a.m. 13.0 590 8.30 a. m. 11.8 225 12 noon 16.8 4.30 p.m. 19.8 590 8.30 a. m. 11.0 600 12 noon 16.2 780 3 p.m. 15.4 7 ep amoE 16.0 5990 8 a.m. 13.5 590 8 a.m. 13.0 590 8 a.m. 12.0 590 8 a.m. 12.2 590 8.15 a.m. 10.1 590 8 a.m. 11.5 590 8.30 a.m. Tt72 415 I p.m. 12.8 4.30 p.m. 15.5 590 8.30 a.m. 11.6 590 8 a.m. 7.8 590 8.30 a.m. 12.5 THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 811 Table V. Temperature observations on Masafuera, 11.2.-14.3. 1917. °C. First and last reading for each day refer to head-quarters at sea-level. Date Hour re Alt., m Hour 2G; Hour ai II.2 10 a.m. 23.6 215 2 p.m. 16.0 6 p.m. 19.0 [2.2 agate vant 23.0 715 25). 10: 16.0 6 p.m. 18.3 14.2 8 a.m. 20.0 650 Cas Oman: 22.5 » 840 c. 12 noon 15.5 » I 100 C. 3) p.m. 16.0 » I 130 3.30 p.m. 16.0 » I 370 4 p.m. 12.0 » I 130 Tepes TLE » » 9 p.m. 8.0 15.2 » oy Glesiank, 14.2 » I 350 ' Qvaran: 13.0 » I 400 10 a.m. 15.0 » I 375 Tayo 17.0 » I 130 12 noon 14.5 5 p.m. 20.6 72 190 9 a.m. 20.0 » 275 10 a.m. 19.0 5 p.m. 22.0 19.2 8 a.m. 20.6 350 18.0 » 435 16.0 » 725 C2) p.m. 20.3 6pm. 20.5 20.2 12 noon 20.0 570 15 ).53 » 475 15.0 6 p.m. 17.5 222 6.30 a.m. 18.5 740 c. 12 noon 12.5 » 950 Cy 2) pam: 11.0 6 p.m. 19.0 25.2 7.30 a.m. 18.5 990 10.3 » I 130 CaO) asm 10.0 » I 240 10.0 » I 130 C43) pem: 12.5 7 pea. 18.7 1.3 8 a.m. 18.0 190 9 a.m. 1720 » 375 Io a.m. 16.3 » 710 Iie Gly eal, 15.0 » 830 13.3 » 920 I2 noon 16.8 » 980 14.2 » I 100 af yo, 1004 14.2 » I 350 2 p.m. 12.0 » I 360 3 p.m. 12.0 » 820 4 p.m. TE » 725 5 p.m. 13.0 » 550 15.0 7230.) .1111. 17.0 3.3 9a.m 19.0 550 10.30 a.m. 15.0 » 525 3 p.m. 17.0 5 p.m. 18.7 5.3 8 a.m 17.0 I 130 Il a.m. 15.0 » I 200 It JO. 106l. 11.0 5 p.m. 19.0 Wes 7 a.m 19.1 585 15.0 » I 000 c. 10 a.m 11.8 » I 250 c. 12 noon 13.0 » I 130 c. 4 p.m. 9.2 » 585 c. 5 p.m. 12.8 6.30 p.m. 17-4 812 C. SKOTTSBERG Table V. (forts.) Date Hour nC: Alt; sm Hour on Hour He, 9.3 7 a.m. 14.0 475 CG. 8-41 18.1 » 550 18.0 » 630 | 18.0 » 1500 2 p.m. 8.9 14.3 Gi tpaaits 18.7 22 8 a.m. 18.0 » 670 9 a.m. 19.0 » 725 10 a.m. 17.0 12 noon 22.0 Johow has little to say about the climate of the islands; at the time when he published his book no dates were available as far as I am aware, but as he classifies the forests as “‘subtropicales stempreverdes” it is evident that he regar- ded the climate as subtropical. From the meteorological observations published at the time of my first visit (1908) I concluded that “‘varm-temperate’’ was a more correct expression (3. 52). It is in good accordance with the classification of KOpPEN |. c. On his map central Chile along the foot of the Cordillera belongs to a region of “Warm gemassigte Regenklimate mit trockenem Sommer’’; to this type belongs the climate of Masatierra, even if the summer is not as dry as on the mainland. With regard to Masafuera it is possible that the highland climate approaches the Temperate climate of south Chile with rain at all seasons, cul- minating during the winter half-year. Winds. The islands are situated near the south limit of the SE trade wind belt, where the influence of winds from S to W also makes itself felt, while winds from other directions are of rare occurrence. In 1914 (Skottsb. 3. 49) I published a table for the years 1901—-1906, based on observations three times a day made in Bahia Cumberland; of this a summary is given in Table VI. Table VI. Wind directions in % of all observations 1901—1906, calms included (a) and excluded (b). ] | | N INNE| NE|ENE| E ESE SE ‘SSE S ‘ssw, sw WSW WwW WNW NW NNW Calm] auier oO 0.4 oO 0:5 ,4|50-L | 2320 20:4 1373 Sona :8 520" 4 | 7.69) 1.6) Sea fe) | II | Jo) |) Tuaae fo) 0.4 fo) 0.6) 40: | 15:3)/23.0] 15.0] rOoNl 14a) S37au| SON) cS mleoea Ome il Winds from SE, SSE and S 47.3 %, from SSW, SW, WSW and W 34.4%, together 81.7%, leaving only 7% for other directions, or, if calm days are ex- cluded, 53,3, 38,7 and 7.7%, respectively. Winds from SSE prevail from August to March, with SE and S—SW coming second; SW overweighs from May to July, bringing heavy rain over the mountains, and from April to July winds from W—NW are more frequent than at other times, but as a rule Bahia Cumberland THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 813 offers safe anchorage. Very likely wind conditions are the same around Masa- fuera, where the higher south half, to judge from the character of the vegetation, receives a very large amount of precipitation. Bahia Toltén west of the North cape generally offers some shelter for small craft, even if landing is not always possible and anchorage unsafe with winds north of SW. The West coast is re- ported always to be inaccessible. As a rule vessels anchor outside Quebrada de las Casas on the East coast, from where it is easy to reach the interior of the island, but as winds east of S are frequent, landing is rarely comfortable and often hazardous or quite impossible. Light conditions. Some observations were made on the strength of light, measured with an ac- tinometer based on the principle of exposing photographic paper and comparing with a standard shade. If the light at midday in the open is set = 1, the follow- ing figures were obtained. Primeval rain-forest 1/20—1/45, under a closed roof of Dzcksonza sinking to 1/100; in the bottom of a narrow gorge, filled with Dzcksonza 1/420 was found. Where ) THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 8 Table XV. Soil temperatures, °C. Masatierra, Bahia Cumberland. An asterisk indicates that the bedrock was struck. Temperature of the air I m above the ground. 1—4 lower montane, 5—10 upper montane forest, II scrub, 12—16 deforested and eroded areas near sea-level. On®'| & Init} Depth in Cm No.) Date | Hour | Air | naked} plant = | earth | cover | 10 II 12 | 15 17 20 25 30 35 40 50 | pee. | 4.20'p. | 18.5 20.0 16.7 2 | 30.3 I.00 p. | 17.0 16.0 15.3 15.6 3 » > | 18.2 | 2212 19.4 18.3 4| 8.4 10.15 a. | 16.0) 13.8 16.1 » | 20.4 | 12.00 n. | 16.0 14.1 15.9 *16.2 5 bee 11.30 a. 13.6 | 12.8 BE Ate 6 | 31.3 E.00 p. |} 15.2 14.9 £555 a | 8.4. 9.30 a. | 14.8 13.3 14.7 14.5 8 | 2 | 9.45 a. | 13.2 gL 13.5 Pc-4-, [e9-00.a. | 52-3 12.4 14.2 14.2 9 | 20.4 11.00 a. | 11.1 | BEE -zTSS IO | 14.4 9.00 a. | 12.8 13.8 *15.8 » | 20.4 g.00 a. | 13.0 13.3 *15.4 ex | 35-3 It.oo a. | 14.2 13.8 Fr4I5 » 2.4. | 8.30 a. | 13.3 E27 *13.6 » | 8.4 8.30 a. | 12.3 12.8 *14.0 » | 14.4 9.00 a. | 12.6 525 *14.0 » | 20.4 8.15 a. | II.o 10.3 *13.4 r2 | 12.52. | 3.00 p. | 15-5 16.0 } 17.5 FEFE 13 | 30.12.| 11.00 a. | 18.0] 38.6 | 27.0 *20.0 14 | 11.1. | 3.00 p.| 22.0] 34.0 | 25.0 *23.5 > | Tea. | 3-20 p. | 22.5 b 43tS i] 2535 19.9 BS | te 4ul.13.45 p>) 20.2,| 29.0 | *26.5 16 8.4. | II.ooa. | 19.8 | | 24.2 | *18.9 1. Dense Aristofelia-Soc. (macal) in Valle Anson, c. 250 m. 2. Plazoleta del Yunque, c. 220 m. Wet forest with Dicksonia and Gunnera. 3. Same locality, on cleared ground covered with weeds. 4. Nothomyrcia-Aristotelia forest along trail from the colony to Portezuelo, c. 250 m. Subsoil below mon nw Io. EIT. 12. 13. 14. 15. 16. 20 cm very hard. . Same kind of forest, with abundant Thyrsopteris, on the same trail, c. 350 m. . Steep wooded slope below Portezuelo, 490 m. . Brushwood about 50 m below the pass. . Wet forest with Dicksonia and Gunnera on the Villagra side of the pass, c. 550 m. . 50 m below loc. 8. Thin forest, dense undergrowth of ferns. Ledge below the Selkirk memorial tablet. Meteorological observation post at Portezuelo, 590 m. Scrub on rocky soil. Steep slope near the colony, facing E, c. 30 m. Sand and gravel with Ugni Molinae — Blech- num auriculatum-Soc. r Near loc. 12, c. 20 m. Sand with patches of Acaena argentea. Bright sunshine. Mobile sand below the old prison caves in Bahia Cumberland, c. 20 m. Bright sunshine. Same place, a little lower down; sand packed hard. Steep slope near the colony, of the same kind as loc. 12 and I 3. Bright sunshine. 824 C. SKOTTSBERG With regard to Masafuera, Quensel concludes that “rocks of more normal basaltic composition, principally feldspar basalts, occupy the lower and intermediate elevations of the island’, up to approximately 1000 m; these rocks are mostly vesicular to slaggy, but more compact higher up, “whereas the highest parts consist of an olivine basalt, supersaturated with iron oxides” (l. c. 71). This last- mentioned kind seems to be restricted to the summit; |. c. 66 the lavas between 1100 and 1420 m are classified as feldspar basalt. Anyhow, the fact that normal olivine basalts form the base of the island helps us to understand why the topo- graphy is so different from the configuration of Masatierra. Volcanic soils have a reputation for being rich in mineral nourishment; basalts give rise to fertile soils. Quensel has published a number of rock analyses from both islands to which I refer (I. c. 83—84). Of the samples of the humus layer collected by us those from the lower levels of Masatierra were lost; only four remain, two from each island. I wanted to get them analysed in the same way as the rock samples in order to be able to compare them, but this met with in- superable difficulties and I had to content myself with such dates as could be supplied by the Bureau of Agricultural Chemistry in Stockholm (Experimentfaltet), which I have to thank for the figures under 1 to 4 in Table XIV. I have added figures for 3 rock samples according to Quensel (II, VI and VII). With these the figures for the soil samples are not comparable; the methods of analysis are different. The soils are acid to moderately acid and, in comparison with the mother rock, poor in Ca, K and P. But these figures tell us nothing about the mineral contents of the underlying clay; they refer to the humus layer. Soil temperatures. The temperature of the soil was observed in 16 places on Masatierra from near sea level to the Portezuelo pass. The apparatus used, con- structed by the late Professor A. HAMBERG, was a sharp-pointed iron-tube, into which the thermometer was lowered. The wall of the tube was perforated on level with the mercury, and the tube was closed with a stopper. The greatest depth reached was 50cm. Table XV. The figures for stations 12—16 are high. The soil is poor, thin and exposed to the sun, it dries quickly and does not allow the seeds of forest species to germinate. The new soil is taken possession of by the much less exacting introduced species or, if exposed to erosion, will remain a desert. At loc. 4, where a stream crosses the trail, the temperature of the water was taken on three occasions. Table XVI. Table XVI. Date Hour Air Water FOr, | 10.00 a. 16.5 13.0 Spi II.00 a. 17.9 13.8 20.4. | I2.00n. | 16.0 12.1 The stream comes from the Portezuelo ridge and is considerably cooler than the soil in the same altitude, c. 250 m. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS oo i) ut CHAPTER II. The Life-forms. In my first report on the flora (3. 53—54) I tried to construct a biological spectrum, using the method invented by C. RAUNKIAER, in order to find out whether or not, in the case of Juan Fernandez, the spectrum was an expression of one of his plant climates. I came to the result that this was not the case; the spectrum seemed to be anomalous, and I tried to find the reason. Our 1916—17 survey added a great number of endemic species and several non-endemics not found before in the islands. Further, I had an opportunity to study the morphology and behaviour of practically every vascular species in the field, enabling me to refer them to the life-form classes with greater exactitude (Table XVII). Abbrevi- ations are according to Raunkiaer with the slight difference that Me stands for Mesophanerophyte — there are no Megaphanerophytes in the islands; both are taken together by Raunkiaer as MM — and that Mi instead of M is used for Microphanerophyte. On the basis of this material the spectrum set forth in Table XVIII was constructed, and compared with a “normal spectrum’’ based on 1000 species instead of 400 (Raunkiaer 5). The main difference between the two lies in the larger percentage of N and Ch, the smaller percentage of Th and the absence of G in the islands. As we shall see later, typical H are not so well represented as the figure seems to indicate. Compared with my old figures the main difference lies in the rise of Ch from 13 to 25.3 % and the fall of H from 30 to 22%. The rise of Ch is caused partly by the removal of a number of H to Ch, partly by the character of the flora above the timber-line in Masafuera, which was little known before 1917: of the 21 species restricted to this region 14 (66 2/3 %) are Ch and 7 (33 1/3 %) H. If we want to compare the plant climate and spectrum of Juan Fernandez with other climates and spectra it is recommended to treat Masatierra (incl. Santa Clara) and Masafuera separately. The result is accounted for in Table XIX. With regard to Masatierra, the figures greatly exceeding the corresponding ones in the normal spectrum are the figures for N and Ch, but as the Ch do not reach 20%, they need not occupy us here, while the Ph deserve special attention. The percentage approaches the figures known from the Ph climate of the tropical zone with not too small rainfall (Raunkiaer 2. 79). Thus we have reason to com- pare the spectrum for Masatierra with some tropical spectra published by Raunkiaer l.c., and this I have done in Table XX. Unfortunately, Raunkiaer’s figures were based on old lists, but even if later additions had been considered it is hardly probable that the spectra would have turned out very different. The principal difference between Masatierra and these tropical islands lies in the higher percentage of H and the low percentage of Th, but H are far from reaching a figure characteristic of a Hemicryptophyte climate (50% and over). To judge from the spectrum, Masatierra is expected to have a subtropical climate, but we have seen that it is warm-temperate, and this may be concluded also from the cultivated plants (Johow z). Hardy things like apples, pears, cherries, raspberries, 55—516796. The Nat. Hist. of Juan Fernandez and Easter Isl. Vol. II. . 826 C. SKOTTSBERG Table XVII. List of the Phanerogams of Juan Fernandez indicating life-form and leaf-size class after Raunkiaer’s method. Species’ restricted to the highland of Masa- fuera are marked by *. | | *Abrotanella crassipes *Acaena maSsafuerana .......... *Agrostis masafuerana........-- Apium fernandezianum ........ Azara fernandeziana..........- Berberis corymbosa » masafueranaeicmst). eh Boehmeria: excelsa Ons ON C8) OG Bromus fernandezianus * » masafueranus Callitriche marginata.......... Calystegia tuguriorum ......... Cardamine chenopodiifolia » flaccida » Kruesselil CAPO O'O OFO COnraO Carex berteroniana Neate? snip) Banksii Centaurodendron dracaenoides. . Centellaitriflorayeint et) «0-2 > Sanctae Clarae... | Chusquea fernandeziana Cladium scirpoideum Colletia spartioides............ | Coprosma Hookeri » pyrifola lajistaelle «66 je ee Cuminia erliantha » fErMANGdEeZaerease eo ones > ovalifolia subsp. australis | Chenopodium crusoeanum...... | » nesodendron..... | Cyperus eragrostis............. | » TELEXAIS) 4.04 PIE - Manthoniascollinaeere sce eciecias : » macrantha eOOW oo » macrophylla » TMALSIMALA I. cela he - | Dichondra repens Drimys confertifolia | Dysopsis hirsuta truncata subsp. Skotts- Sophora fernandeziana......... Mi mi | TT ae a a oe aoa 2 Th | mi > MmASaieralid |. « AREER DIS fs: aie/e.s spake i bast © alee Stipa, IAaevissima, «assed. . waned a *Ranunculus caprarum ......... Ch | me | PRE MCOSIAUIA s o.6 <.crsid'eia = slate | H 1 | | Rea micrantha.........+...+-- Mi | me Symphyochaeta macrocephala ..| Mi | me | BMHIETIILONIAL 5 cies wie e+ 0 nisin ole « Mi me Petiaponia expansa-<:.... <2. <1 Th | mi | | Pe PLUM ACA se coe oles viele ele ol N me Trisetum chromostachyum ..... H n | Rhaphithamnus venustus ...... Mi | mi WOfaeit Seid cit 7 A es Arena. N n | Rhetinodendron Berteril....... Mi | me | *Uncinia brevicaulis............ H i) | Robinsonia evenia............- | Mi~| me | D COSA TTT rae {ico farement » Payallal. cel. oe ale Ie Mi | mi > DIGRESS 46 HSAneeonro. H n D RACH IS Gag acemayoe ee Mi | mi “gah PHIEOIMES eaters ets) < = = H n | > Masafuerae ........ Lee ct) ini =): 0 ie CERES Ve Seats ee Sis H n > ERUEHETA, -ai-te/etalanlaete Mi. | me, |) Urtica fermmandeziana ~~. -....~. N mi “Pvt ses CEE oe pe odanaodcooe Ch | n WWhnWehinteme np oscceccdeee | Th »| cae Salicornia peruviana........... [,Gh |. 1 Wahlenbergia Berteroi ........ Ch | n | Santalum fernandezianum...... | Me | mi D fernandeziana ...| N |n | SG@ILDHS) CEMNUUS «14. - -\- «Ge oie - | geal | l » Grahamae ...... N ~ [pant > MOGOSUS: Wicks cetera ote ota |H In » arrainii) 110s = 1S ergy eee) Selkkirkia Berterits:.%|< . 2 sajsujaste INE. \ji-m03 D Masafuerae ..... }Ch jn | Solanum fernandezianum ...... Ne |lome WandueashenZiieanie ee sitet | Mi | ma | Table. XVII. The biological spectrum of Juan Fernandez compared to the normal spectrum. Life-forms in % of total number of species Number of) Ph species | s | = [mm| M| N|ca| H | G [HH] Te |O%) | | | Juan Fernandez ..... 146 | — | 0.7 | 6.8 pee 24 |25.3| 22 | — | 1.4 | 3-4| 48 | |The normal spectrum I 000 2 | PiukSoelas 15 | 9» |\i26 4 | Pow l3 | 46 | Table XIX. Biological spectra of Masatierra and Masafuera compared. Life-forms in % of total number of species | Number of | Ph i | | | y Ee | She) | Mee | Me | ON | ch | # | G | HH] ane) Piiasatierra, «aod. - «as | 100 a NPR IR es [25 ) |. p22. | els ee ene WRASASUIET A 5 3 ota «4. 73 ai et 6S 5.2 h20.5| 37] 26 | SE 2) a es | | | 3 |The normal spectrum| t1ooo | 2] 3{|8 |18 |15 | 9| 26| | ZTE | 2 40 828 C. SKOTTSBERG Table XX. Biological spectra of Masatierra, Seychelle Is. and West Indies. eee ee ree —————————————————————————————— | | Life-forms in % of total number of species ESheteelotes of| Ph | species | ¢ | £.|mMM| M | N |ch-| B.|.G.| HH] Th | | Seychelle Is. ........ | DGS na |\eme okt | 3 me) 23 | 24 6) TZ 3 2 16 61 St Thomas and St Jan | | (West Indies) ..... 904 2 I al 23 ulpe3On| 2 9 3 I TAe, Jn |Masatierra .......... Too — I 8 21, 25 7 22 == 2 4 55 The normal spectrum, I 000 2 3 Si ee el 9 | 26 4 es Sy [res Table XXI. Biological spectra of Masatierra and Macaronesia. Life-forms in % of total number of species Number of species Ph Ch H re HH Th SHE WAZOLES creole os eotel ot fat et terete 435 6 a 33 AEs | 5 46 Madenarey ay hove cnet tcl cle setter seer onctets 448 20 II 24 2 3 39 Canaty, US tees tetettereret liar | I 302 22 07 | = || 9 Sr Shel 5 2 39 SHINO So, duo Bann Olio oldloso,.b'a0 | 400 9 13 32 II 2 33 MaSatientas oa0 co cc chee ot Too 55 17 22 a= 2 4 The normal spectrum........ I 000 | 46 9 26 4at nr Wee 13 Table XXII. Biological spectra of Juan Fernandez. Life-forms in % of total number of species | Number of ow Ph Ch H HH Th | Indigenous species........ | 146 48 2563) | | 22 1.4 3.4 Adventitious species ...... | 87 | 5-7 2.3 29.9 4.1 60.9 JNM SHOES 5 oo Sogo uppoooe 2238 32.2 LOL, 25 Mes} 2 Table XXIII. Biological spectra for Masafuera. | Life-forms in % of total number of species | |Number of species '/' ¢ | eam) mw | N | ch) | G | HH] Th |) Pp ones tibelltater-nypere ico: 2 268 | 7728 Sul ee bial 23) — | —] 5.8] 46 Highland yO OO 1 OOOO 34 |e —= || oF! ——| 14.7 56 | 23.5 — — | aT | The normal spectrum . I 000 2 SP les 18 Se On 20 4 2A\ES 46 12 sp., I here rather M. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 829 Table XXIV. Leaf-size classes. Leaf-size in % of total number of species | | Number of |_ | | | | species | 44 ] |) ees eee | ma lentes | | : | BUIRSDEGIES: = craters cients 0.5 - + +2 146 O58 30:7 32.2 0327, 277, 2 Forest specieS-..........-..-. 125 8 33.6 37.6 15.2 3.2 2.4 Ape | PY ie. oh wa clean 3 'e 21 ro Pee ))|| 5 _ | gooseberries etc. are grown with more or less good result, peaches, quince, chestnut, figs, grapes, artichokes and maize do well, Eucalypins globulus, Pinus radiata, Albizzia lophanta are cultivated, Zantedeschia aethiopica, Ricinus communis, Pelar- gonium spp., Tropaeolum majus etc. have become naturalized, but Cztrus fruits have failed. Nevertheless, the very high Ph percentage ought to correspond to a subtropical climate with sufficient precipitation. On the other hand, the islands lie within the zone of winter-rains, and we have already seen that there is con- siderable resemblance between the climates of Masatierra and of the Macaronesian Islands. The biological spectra, taken from Raunkiaer (6) are given in Table XXI and compared with the spectrum for Masatierra; I have added the spectrum for the Mediterranean island of Samos after Raunkiaer 2.55. Compared with the resemblance in climate, the difference in spectra is much greater than we have reason to expect: the preponderance of Ph in Juan Fernandez stands out, as well as the low percentage of Th. Macaronesia has a Therophyte climate, less evident in Samos, where, on the other hand, G are numerous. Even if the summer is much less dry in Masatierra, the insignificant number of Th is surprising. L.c. I referred to Raunkiaer’s statement that if a spectrum is a true expression of the climate it remains unaltered even if we add alien species which have become naturalized and which do not depend on cultivated soil or on the vicinity of settlements for their existance. I reckoned with 36 neophytes regarded as naturalized. Their number has now grown to 87, which have spread outside Valle Colonial on Masatierra, many of them also found on Masafuera, where some species never observed on Masatierra occur. Table XXII shows the life- forms in % of the total number of indigenous, adventitious and indigenous + ad- ventitious species. I believe that the last of these spectra, which approaches the spectra for Madeira and the Canary islands, is a more correct expression of the climate than the first. This, then, must be anomalous, and we have to find the reason for this anomaly. L.c. I held the isolated endemics responsible. They have no relatives in South America, whereas several point toward New Zealand and Oceania; together they form part of Engler’s ‘‘Altozeanisches Element”. In Juan Fernandez it is represented by 51 species, 24 of these belonging to 12 of the 16 endemic genera. Their biological spectrum shows 9.8 % Me, 27.4 % Mi, 43.1 % N, 11.7% Ch, and 7.8% H. The Ph% is 80.3. It is this element more than any other that puts its stamp on the spectrum and invites us to pay attention to 830 C. SKOTTSBERG the historical factor, i.e. the origin and evolution of this insular flora, a problem not discussed in this paper. It remains to cast a glance on the spectrum for Masafuera. We learn from Table XIX that it differs not little from the spectrum for Masatierra in the percent- age of Ph, which is much lower, and in the percentage of Ch, which is much higher. It is evident that this spectrum cannot apply to the vegetation below the timber-line. Nor is it possible to get a correct idea of the relation between the life-forms in the highland only from the fact that of the 21 species restricted to this region 2/3 are Ch and 1/3 H. Many species extend their range through both regions. For these reasons it seems advisable to treat them separately; see Table XXIII. In the forest region Ph dominate in spite of the small number of rosette-trees, which are so important on Masatierra. The region above the timber-line has a Chamaephyte climate of oceanic type. As a complement to his life-forms Raunkiaer later (z) introduced his leaf-size classes, lepto-, nano-, micro-, meso-, macro-, and megaphyllous. The allotment of the flowering plants of Juan Fernandez is found in Table XXIV. CHAPTER III. Morphological and biological remarks on the Vascular Plants. THE LEADING FOREST-TREES. The dominant trees of Masatierra are Wothomyrcia fernandeziana, Drimys confertifolia, and fagara mayu, of Masafuera Myrceugenta Schulzei and Drimys; Fagara externa, the vicarious species of Masafuera, does not take a leading part in the formation of the forest. Second to these, but common on both islands, comes Coprosma pyrifolia. Nothomyrcia fernandesiana (Hook. et Arn.) Kausel, “Luma’’. Myrtaceae. The dominant tree all through the forested section, extending from near the sea (at least in bygone times) to the highest ridges wherever a suitable habitat is found, but not reported from the summit of El Yunque; ranging west to Quebrada Juanango on the north side and to the foot of the precipice of Cerro Chumacera on the south side. In good soil 15—20 or even 25 m (Johow) tall with a trunk diameter of 25—30, in very old specimens as much as 50 or 60 cm, the largest measured by me 80 cm at the base and 71.5 cm 1.5 m above the ground. Isolated specimens with a broad, very compact head (Fig. 6). Trunk often angular in section and several-stemmed trees not uncommon (Fig. 6). Bark smooth, light gray with large lenticels, thin (I—2 mm), as a rule with numerous epiphytes, among which a purple /7x//ania is conspicuous. Leaves (Skottsb. 3, Pl. 7:2) stiff chartaceous, dark green, often covered with the sooty black fungus Lzmacinia scoriadea, according to Johow (s.n. L. fernandeziana) very destructive, but I did not get the impression that it ever kills the trees. Buds naked, pubescent. The THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 831 Fig. 6. Two solitary trees of Nothomyrcia fernandeziana in a deforested valley. — Photo K. Backstr6m. young shoots, which develop in late summer, start with 2—3 pair of reduced, almost scale-like leaves, becoming separated by prolonged internodes. Young leaves soft, crimson-coloured. The terminal and axillary inflorescences are developed in great profusion from August to November. The apex of the vegetative shoots not producing a panicle appears to die at the end of the growth season. Myrceugenia Schulzei Johow, ‘“‘Luma’’. Myrtaceae. Replaces Nothomyrcza in Masafuera and extends from about 200 to about 800 m above sea-level. A fairly large tree (Pl. 107:2), reaching a height of 15—20 m in favourable situations and a stem thickness of 40—50 cm (60 according to Johow). Bark light gray, squamulose, 2—3 mm thick. Leaves of very variable size and thickness, to 6 cm long and +glabrous in the closed forest, barely 1.5—2.5 cm long, thicker, firmer and covered underneath with a fluffy, golden and caducous tomentum in open, sunny situations. Buds naked but densely hirsute (Fig. 8 a) as are the young branches; observed resting in August, 1908, and in Feb.—March, 1917, so that it seems that the autumn and winter is the growth season. Flowers during the summer, much later than othomyrcia (Pl. 107: 2). Drimys confertifolia Phil., “Canelo’’. Winteraceae. One of the commonest trees in both islands, ranging through the entire wooded section of Masatierra, where it reaches the summit and also clings to the exposed ridges where few other trees will grow; on Masafuera rarely descending below 500 m and as a rule forming the tree limit in 1000—1100 m with 1200 as the highest station observed. As Johow remarks it is less abundant on this island. A medium-sized tree, 8—12m 832 C. SKOTTSBERG ; : . athe p oe he “ ‘ ee f Odean = or Se OER was oth ma ALN ee Ge Nae eis at EE a AE BS me Fig. 7. About 2 1/2 m wide base of Fagara mayu along trail to Valle Anson, Masatierra. Length of carpenter's rule placed above the base 52 cm. — Photo 28/12 1916. tall but rarely exceeding 10 m, with a straight, stout trunk 1/,—1 m thick, some- times divided from the base which in the fog region attains a very large size, 2m across or even more (Pl. 96:1). Pl. 58:1 shows ordinary specimens with few main branches and a dense head. Bark smooth, thin (1.5—2 mm), light gray when young, later darker, thicker, cracked and more or less covered with bryophytes and lichens. Young leaves tender, upper surface more or less crimson- coloured and contrasting with the glaucous bloom of the lower side; mature leaves thin leathery, bright green above, lighter than the luma leaves and making the canelo trees stand out well even from a considerable distance. The pseudoterm- inal and axillary, umbel-like panicles are well exposed and sprinkle the head with white blotches. They develop from early spring to summer and are protected in bud by a set of reduced leaves. The terminal bud of the vegetative-floral shoots appears, as a rule at least, to die, and axillary innovations grow out during the summer, ending in a bud covered by scalelike leaves which remain reduced and very small when growth starts anew (Fig. 8 b; comp. Drimys Winter7, Raun- kiaer z. 35, fig. 15 A). In D. confertefolia the innovation begins with a few very short internodes between the reduced leaves, succeeded by prolonged internodes with normal leaves, followed by short internodes causing the leaves to become crowded toward the tip of the branch; hence the specific epithet. I have seen THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 833 cases where the axillary buds of an innovation developed in April and gave rise to proleptic shoots and even few-flowered umbels. It also happens that the tip of a flowering branch continues to grow and forms a set of crowded leaves, but after April no new ones were produced, and I doubt that the very small apical bud makes a fresh start in the spring. Shoots not having flowered were at rest in August (Fig. 8 b). Fagara mayu (Bert.) Engl., “Naranjillo”. Rutaceae. A common tree all through the wooded country, extending to at least 600 m above sea-level and going west to Juanango and Cerro Chumacera. The largest tree on this island, up to 25 or perhaps 30 m tall; the three largest trunks seen measured 87, 95 and 110 cm in diam. 1.5 m above the ground. The first-mentioned is shown on Pl. 59:1. In old trees the base is wide and the roots, spread horizontally and continued in clumsy buttresses, help to support the trunk much as in so many large trees in tropical forests (Fig. 7). Head shaped as a more or less regular, inverted cone (see Skottsb. 3, Pl. I:1). Bark thin, 1—2 mm, gray, rough and ringed by broad lenticels. Wood yellow, hard and durable. Leaves pinnate, firm, the young ones very resinous. Buds naked but weli protected by a coat of resin (Fig. 8d). Specimens in full bloom were collected in August, 1917 (Fig. 8c; none were found at the same time of the year in 1908), and mature fruits together with female buds — the species is dioecious — in April. Vegetative buds had started life in August and continued to develop all through the summer and as late as April; possibly there is no definite period of rest. With the new growth the leaflets of the older leaves are gradually lost, leaving the rhachis standing. Fagara externa Skottsb., ““Naranjillo”. Lowest and highest stations observed, 280 and 515 m. A medium-sized tree, probably not over 20 m tall, much more slender than its Masatierran congener and less frequent. Only known with fruit, collected in Feb.—March, 1917. Coprosma pyrifolia (Hook. et Arn.) Skottsb., ‘‘Peralillo”. Rubiaceae. In both islands, an important tree on Masatierra, ranging next to luma, canelo and naranjillo, from near the sea to at least 650 m and west to Cerro Chumacera. Scattered on Masafuera, but probably occurring in all wooded valleys and forming small pure stands in some places outside the luma forest, e. g. at the entrance to Que- brada de las Casas (Pl. 109:1), Along the edge of the highland it attains 950 m, and one very small plant was observed in the Alpine heath. A medium-sized tree; mature specimens according to Bertero 4.5;—6 m high on Masatierra, rarely over 10m according to Johow, who figures a fine specimen on pl. VII — the explana- tion says 7m high with a trunk I90 cm in circumference, but measurements on the plate give other results: with a height of 7 m, the diameter was about 27 cm, with a diameter of about 58 cm, the height would have been 15 m. We saw the largest specimens on Masafuera, about 12 m as estimated. The smallest fruiting plant was only 2 m tall. Bark about 3 mm thick, light grayish brown when young, later darker to almost black and checkered by prominent cork-ridges, mostly without epiphytes. Leaves soft coriaceous, very bright green and glossy in sunny, open places. Shoots continue their growth through the summer; Fig. 8 e shows an apex in the beginning of April, when the terminal bud, protected by the 834 C. SKOTTSBERG Fig. 8. a Myrceugenia Schulzei in February. Apex of shoot with pedicels left in the axils. 2x.— 6 Drimys confertifolia in August. Nat. size. — ¢ Fagara mayu in August with & flower-buds ready to burst (other trees were already in flower). Nat. size. — d the same, terminal bud. 3 1/2x. — e Coprosma pyrifolia, top of branch in April. Nat. size. —_/ the same; two young leaves, which will not develop further, have appeared. Nat. size. — g Coprosma Hookeri, apex of shoot in August, with axillary bud. 2.4 the same, apex in March. Nat. size. — 7, & two apices of same, showing a reduced (i) and a normal (k) first leaf-pair. 2x. 7 Escallonita Callcottiae, apex of shoot in August. 2x. stipules, has come to rest. In Fig. 8 f a pair of reduced leaves has been formed; they will remain in this state. No flowering specimens were collected on Masa- tierra, but ripe fruits in April; on Masafuera with flowers in February and March. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 83 on TREES OF SECONDARY IMPORTANCE, BUT MORE OR LESS COMMON. Coprosma Hookeri (G. Don) W.R. B. Oliver, “Olivillo”. Only on Masatierra, common in all the woods from about 300 m; west to Cerro Chumacera. Not unfrequent in the brushwood along the ridges. Johow calls it a shrub (“estatura frutescente’’), but it is a small to medium-sized tree, occasionally as much as 7 or 8 m tall but generally smaller and branched from near the base (Pl. 62: 2 and 63:1, where the arboreous habit is distinctly seen). Trunk rarely over 20 cm in diam.; often elliptic in section; a sample in our collection measures 19x12 cm. Bark smooth, light gray. Leaves thin coriaceous, of 2 years’ duration. Buds effectively protected by the stipular sheath (Figs. 8 g—k). The new growth starts in early spring (Fig. 8 g) and continues until late summer (Fig. 8 h). The first leaf pair is of normal size (Fig. 8 k) or reduced (Fig. 8 i). Flowers Nov.—Feb. Boehmeria excelsa (Bert. ex Steud.) Wedd., ‘Manzano’. Urticaceae. Only on Masatierra, where it is common along the streams in the valley bottoms, mixed with luma, peralillo etc. or forming pure stands (Skottsb. 2, Pl. 7); not observed above 400 m, but ranging west to the south slope of Cerro Chumacera. Maxi- mum height 5—8 m (Looser), 10 m (Johow); generally not over 7 or 8 m; trunk short, bulky, more or less irregular in section, to 50 cm thick, often sprouting from the base. In many-stemmed specimens the trunks may be pressed together and more or less twisted round each other, and in the furrows formed in this way a dense felt of aerial roots is developed, but their function has not been investigated. From the lower limbs roots will go down into the soil. Bark dark grayish brown, densely warty, 2—3 mm thick. The wood is very light, soft and full of sap; a bored stem will produce a stream of a faint yellowish, tasteless liquid which continues to flow for some minutes. Leaves of the common urticaceous type, coriaceous, tomentose, unlike everything else in the island flora (Skottsb. 3, Pl. 7:1). Buds naked; the caducous stipules will offer protection only at a very early period, but the involute, densely hirsute leaves protect each other well (Fig. 9 d). Sprouting in August; the young shoots will bear flower clusters from November on; in April still growing and all kinds of stages found. Pl. 64: 2. Rhaphithamnus venustus (Phil.) B. L. Robins., ‘Juan Bueno’. Verbenaceae. On both islands, common in the forests of Masatierra, especially on the higher humid slopes and ranging west to the south precipice of Cerro Chumacera. Much less frequent on Masafuera, observed from about 440 to 515 m. A middle-sized tree, 6—8 m tall with trunk to 40 cm in diam. Pl. 59:2 illustrates an unusually large specimen, a good 1om tall with the distance to the lowest limb 3 m, and a flattened trunk 46 and 26 cm in diam., respectively, 1.5 m above the ground. Bark often covered with foliaceous lichens. Branchlets slender, pendent (PI. 64:1), exposing the dark lilac-coloured flowers. Leaves small, firm, dark green, often attacked by Limacinia. The tip of a resting shoot apex in August is seen in Fig. 9 a; bud naked, but densely hirsute, as are the young leaves. In Nov.—Dec. the flowers appear; inflorescence a 2-flowered dichasium, ending in a needle-like spine, which, however, is not always developed (Fig. 9 b). Below the inflorescence is a serial accessory leaf-bud. As a rule growth of the innovations is arrested in 836 C. SKOTTSBERG Fig. 9. a Rhapithamnus venustus, resting apex in August. Nat. size. — d-part of a flowering branch in December. Nat. size. — ¢ part of a long-shoot at end of March. Nat. size. — @ Boehmeria excelsa, open bud in August. 2 x. — e Cuminia fernandezia, apex of shoot at end of March; a-a and b-b are leaves belonging to the same pair. Nat. size. — f, g the same 9/4 1917, accessory bud (f) and bud with a pair of reduced leaves (g). 2x. — A Sophora fernan- deziana, top of annual shoot. — 7 the same, dwarf floral branchlet, both 9/4 1917. — & the same, top of branch in August; ab axillary, tb terminal bud, fl inflorescense. All nat. size. — 7 Azara fernandeziana, node with axillary bud in August. Nat. size. — m the same, bud. 2x. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 837 March, but they produce new leaves and flowers as late as in April or have stopped growing and end in a bud. Other branches of the same order are several dm long and carry axillary spines and accessory buds, and a second bud, barely visible in the leaf axil, may be present (Fig. 9 c); these spines, which bear 1—2 pair of minute scales, will not carry flowers. In the upper axils no spines had been formed. There is a difference between long vegetative ‘‘prolongation’’ shoots and short vegetative-floral shoots, Seeds were sown in the Botanical Garden, Goteborg,’ 7.8. 1918, and a few ger- minated in March, 191g. In October, 1924, two plants remained, but one died soon after without having flowered; nor has the other, still alive in 1952, flowered. Escallonia Callcottiae Hook. et Arn. Saxifragaceae. Endemic on Masatierra, where it is common in light forest and scrub all over the wooded region and ascends to the summit of El Yunque; also frequent on the windswept ridges and scattered in the barren grass-lands on the south slope of the island, west to Punta ated. Pls."62:1, 63:1. A small’ tree, branched from near the base and’to™4 or 5 m tall, but much stunted and shrub-like in dry, exposed situations. Leaves firm chartaceous. Panicles terminal on vegetative-floral shoots; flowering season extending from late winter or early spring to April, flowers pink. In August buds were closed or opening, giving rise to a generation of branches which will flower during the summer (Fig. 8 1), but in other cases dormant buds were noted which had produced a single, very much reduced leaf and were covered by regular budscales, formed by the leaf-base; a miniature limb is sometimes developed. In one instance I noted the following sequence of foliar organs (August, 1917): 4—5 budscales, 2 or 3 leaves of reduced size, then larger leaves or, between these and the reduced ones, another set of scales: the bud had started to grow, had become dormant for a while and then made a fresh start. Growing apices were still observed in January, but in the middle of April growth appeared to have come to a standstill. It often happens that the apex dies without producing a panicle. In the case of /sca//onza as well as in many other the lack of observa- tions during the months of May to July is to be regretted. Seeds were sown 25.7. 1918 and 2 seedlings obtained in the beginning of Septem- ber; the plants flowered for the first time end of September, 1921. Easily propagated by cuttings. Cuminta fernandezia Colla. Labiatae. Endemic on Masatierra, not uncommon in the montane region, but never abundant; seen once only below 300 m. Ascends to the summit of El Yunque. A small tree, probably rarely over 2 or 3 m tall and branched from near the base; main stem to 20 cm in diam. (Pl. 65). Old bark grayish black, thin, cracked and fibrose. Leaves thin, bright green. Branches collected in August had, as a rule, lost the leaves of a long series of nodes, which, the lower ones excepted, bore remnants of dichasia. Leaf scars very promin- ent, ebony white in contrast to the smooth, brown cortex, every node with a minute pubescent bud, accessorial in the flowering zone and hidden at the base of the old peduncle (Fig. 9 f). The shoots continue to grow and to produce leaves * All observations on the cultivation of plants from Juan Fernandez refer to this garden. 838 C. SKOTTSBERG and small but very numerous violet flowers all through the summer and end in a terminal, naked bud. It happens that the apex of a branch which has flowered profusely and produced a generation of lateral, vegetative-floral branches, dies at the end of the second season. Fig. 9 e shows a fruiting branch end of March: I flowered in December and continued to grow and to produce flowers, II are lateral branches of which one has reached fruiting stage while the other soon stopped growing. Buds in the lower axils of II may form a pair of small, reduced leaves before coming to rest (Fig. 9 g). Seeds sown 23.3.1918 germinated during April and May, and one plant flowered for the first time in the beginning of September, 1921. Cwmznia does well in a tem- perate house and flowers and fruits freely. Santalum fernandezianum F. Phil., ““Sandalo”. The sandalwood, now extinct, must have been rather abundant during bygone ages, until it became a victim of reckless exploitation. In 1908 I saw the last — as far as we know — tree on Masatierra in the interior of Puerto (Quebrada) Ingles; see Skottsb. z for partic- ulars. I shall add here that the bark is dark brownish gray, divided in rectang- ular pieces and to 12 mm thick. Sections of a stump of a dead limb measure 13.5X10.5 cm and show distinct rings, each with a lighter and a darker zone; about 54 were counted. No rings could be distinguished in other Juan Fernandez trees. The buds were described and illustrated in Skottsb. 3. 56, fig. 9. The last tree on Masatierra died before 1916, I was told. Place-names such as Quebrada del Sandalo, Q. del Sandalito and Rodado del Sandalo show that a species of Santalum was found all along the east side of Masafuera, where it was extermin- ated a very long time ago. Johow, who devotes 5 quarto pages to a detailed and interesting relation of the history of S. fernandeztanum, remarks that in view of the inaccessibility of a large part of Masafuera it is quite possible that numerous living trees still exist on this island. However, Masafuera is much less inaccessible than the interior of Masatierra, the forest patches are of very limited extent and easily explored, and no trace of a living specimen has ever been discovered. Fuania australis (Mart.) Dr., “Chonta’’. Palmae. The chonta palm (a name given to many species of palms in the Andes) is confined to the woods of Masatierra, where it ranges from Puerto Frances to Puerto Ingles and Villagra, extending from about 200 m to the top of El Yunque; it cannot be called uncommon on the higher montane slopes and ridges from about 500 m. Being so utterly unlike all other trees in habit it can be traced quite easily with a pair of field-glasses from a long distance. In old days it must have been much more frequent, but as the “heart”, the apex with the undeveloped leaves, is edible (“col de palma’’) and the wood appreciated for making walking-sticks and for cabinet work, it is becoming extinct in places of easy access, especially in the mountains around Bahia Cumberland. That it has been protected by law for over 50 years seems to make no difference. In less accessible places one is still -—— or was, at least, in 1917 — able to come across small groves of a number of tall specimens and afresh growth of vigorous young plants (Pl. 96:2). There is no bird on the island likely to disperse the spherical drupes which are coral red and measure THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 839 1.8 cm across, with a thin, sweetish and mealy mesocarp and produced in exuberance. They require one year to ripen so that fruit-stands with green and with red drupes are found simultaneously on the same tree. Ywuanza is dioecious. As mature the drupes stay on for a long time and fall to the ground near the mother tree; if shaken by a strong wind they may of course be hurled some little distance. More than once we saw a tree surrounded by a growth of baby chontas. Fuanza is a stately, straight-stemmed palm up to 10 or 12, by exception as much as 15 m tall with a stem diameter of 15—20 cm; on a fallen, rather small specimen, only 7 m long, 21 cm was noted, another fallen trunk was 12 m long and 16 cm in diameter. In fully developed specimens the trunk has become cleansed from the dead leaves right to the top; a young tree, 4.5 m tall, still bore numerous dead leaves. Bark bright dark green, polished, ringed with whitish leaf scars. Leaves to 3.5m long; they have.been described in detail by Johow. Spathe 1—2 m long. No observations were made on the seasonal development. A specimen planted in Mr. Charpentier’s garden in 1903 was 5 m tall in April, 1917, and estimated by the owner to be at least 20 years old. According to the same source 2 leaves are developed annually. If this is true, the flowering specimen illustrated on Pl. 60:2 would be well over 100 years and the fruiting tree on Pl. 61 not much younger. Seen in full bloom during January; fruit is found throughout the year. The large yellowish-white male panicles are rather showy (Pl. 60:1) and fragrant. The female flowers are almost scentless. Numerous mature seeds were sown, but none germinated. LOCAL TREE SPECIES OF SLIGHT IMPORTANCE. Sophora fernandeztana (Phil.) Skottsb., ‘“Mayu-monte’’. Papilionaceae. Scatte- red through the lower montane forests of Masatierra, west to Quebrada Juanango and extending from about 200 to about 450 m, always very local. A small to middle-sized tree; Johow p. 88 says “‘alcanza una altura de 5 metros’, but on p. 246 it is mentioned among the species that reach a maximum height of 10 m; specimens as tall as that are, I believe, rare, and the trunk rarely surpasses 20 or 25 cm in thickness. Bark about 3 mm thick, dark brownish gray, when old squam- ose. Wood very hard and durable, which made Johow remark that, unfortunately, it can be used for small objects only; his opinion that the island species is ident- ical with the common mainland species may serve as an excuse. (All indigenous trees and shrubs of Juan Fernandez are endemic.) In April, Sophora was clad in bright green foliage developed during the summer, the vegetative long-shoots ending in a naked bud, covered with a dense, white to rusty brown, velvety tomentum; below these shoots were floral dwarf-shoots bearing 1 or 2 leaves and ending in a raceme covered with the same kind of tomentum, each flower-bud enclosed in a boat-shaped bract (Figs. 9 h, i). In August, the leaflets are shed, leaving the rhachis standing, while new leaves are unfolded; fig. 9 k shows that racemes are developed in the uppermost axils. Growth continues all through the summer. Sophora flowers in early spring, presumably in September—October; in December only young pods were observed. Ripe seeds were collected in the beginning of April. The golden yellow flowers are quite showy. 840 C. SKOTTSBERG Seeds sown in 1918 and again in 1923. Plants still alive in 1952, but they have not flowered. Sophora masafuerana (Phil.) Skottsb., ‘““Mayu-monte’’. Endemic on Masafuera, on the ledges of the cliff walls in the deep canyons, not common and often beyond reach. Smaller than its Masatierran congener, with thin limbs, trunk prostrate—ascending. Bark dark brown, almost black. Seasonal development probably as in the former. With ripe seeds in the middle of February, Cuminia ertantha Benth. As C. fernandezia, but much more scarce and not seen by us below 500m. A small tree, but in cases a little taller than the latter and the short main stem to 35 cm in diameter. Seasonal development as in this. Azara fernandezsiana C. Gay. Flacourtiaceae. Only on Masatierra, where it is scattered through the montane forest belt; very rare below 350 m, highest altitude noted 570 m. Not observed west of Cordon Salsipuedes. Only a solitary or a few specimens found in each place. Maximum height 10 m (Johow), trunk to 25 cm in diameter. Bark similar to that in Hagara mayu, wood light brown. Leaves firm membranous, glabrous, to 6 or 7 cm long. Branches collected in the beginning of August, 1917, bore open, naked buds with involute, hirsute leaves (Fig. 9 m), while others observed a fortnight later in 1908 were sprouting. In full growth and bloom during December, 1916; a flowering specimen with the golden yellow tassels contrasting against the bright green, lustrous foliage is a very attractive sight. Growth continues all through the summer, and fruit was found in April. A tree observed on April 6 had come to rest, but a young specimen nearby was still producing new leaves. * * * I shall add some remarks on the maqui, Avzstotelia maqui L’Her. Introduced from Chile, it has come to play an unfortunate role on Masatierra, where it is crowding out the native forest. In most of the valley bottoms the maqui forms a more or less wide fringe, extending up the slopes. Morphologically it is a shrub, vigorously sprouting from the stem base and resembling the common European hazel in habit (Fig. 10), but the individual secondary stems are of arbor- eous type, so that if we call it a many-stemmed, small tree, a few m tall, this is not far from the truth. Under the canopy of the maqui no native tree will germinate, whereas the maqui springs up almost anywhere, if the soil is not too dry; consequently, it appears under the closed cover of the luma forest. Of course it does not kill the other trees, but because they are prevented from reproducing, the native forest is gradually transformed into a pure macal. Dispersal is provi- ded for by the numerous thrushes; this was proved by examining the stomach contents. It is used as firewood, but the population is small and the growth of the maqui extremely rapid, and when logged, dozens of saplings from the base immediately fill the gap (Pl. 63:2). Scattered specimens were found in the higher mountains of Masatierra, while it was scarce on Masafuera where, however, several robust plants were met with in the interior of one of the valleys, so that there is no reason to believe that this island will be spared. In April the buds, which have true budscales, were resting. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS S41 Fig. 10. Dense stand of Avistotelia magui (a macal) along trail to Valle Anson, Masatierra. Photo 30/3 1917. THE ROSETTE TREES. This interesting growth-form is particularly well represented in Juan Fernandez where no less than 25 species occur, representing I2 genera, of which 9 are endemic, and 4 families: Compositae (9 genera, 20 species), Umbelliferae (1 gen., 3 sp.), Plantaginaceae (1 sp.), and Broneliaceae (1 sp.); 5 species approach the rosette type closely, but will be treated separately. Buds never protected by bud scales. As a rule the rosette trees are not found in the closed forest; they may be met with in open glades and concentrate in the narrow belt where the forest is replaced by scrub at the foot of the steep rocky ridges. Two morphological types are distinguished: a) the candelabrum type; below the terminal inflorescence 2 or 3 innovations are developed, resulting in a pseudo-dichotomic — trichotomic mode of branching (all genera except 3 or 4); b) the palmiform type, either pollac- anthic with axillary inflorescences (Plantago) or hapaxanthic with terminal in- florescence (Phoentcoseris); stem simple. For Yuxquea and Hesperogreigea, see below. a. The Candelabrum type. Centaurodendron dracaenoides Johow. Compositae. Confined to the elevated ridges encircling Bahia Cumberland, from Cerro Damajuana to north of Portezuelo de Villagra and ascending to the summit of El Yunque; not seen below 500 m. 56 — 516796. The Nat. Hist. of Juan Fernandez and Easter Isl. Vol. II. $42 C. SKOTTSBERG Fig. 11. Vungquea Tenzit on the summit of El Yunque. — Photo Augusto Grosse 17/8 1934. Very rare. A small tree, 2—3, seldom 4 m tall, with a broad hemispherical head: main stem up to 9g cm thick, bark ringed with conspicuous leaf scars. A tall specimen, which was well over 3 m high, had 4 generations of branches and had flowered thrice; the innovations require several years before producing the large terminal panicle. In August, 1908, as well as during our permanence from December to May, 1916—17, only remnants of old inflorescences were seen, and I am pretty sure that not a single individual flowered that summer; whether the same was the case during the summer 1908—og I cannot tell. Johow had no better luck, and a flowering plant was seen for the first time in December, 1935; for partic- ulars, see Skottsb. 79. 363—365, fig. 1. Flowers dark blue. Cenxtaurodendren is an extreme rosulate type (comp. Skottsb. 3, Pl. 5:2); on the top of a branch I counted 25 leaves crowded over a distance of 30 mm, the average length of the internodes being about 1 mm. Bud forming a closed cone, seen with leaves unfolding in August (Skottsb. 3. 57, fig. 10). Yunquea Lenzi Skottsb. Compositae. Collected twice in the only locality known, the summit of El Yunque. All material preserved is a few large leaves, the largest 60 cm long; for particulars see Skottsb. 77. 163—166, figs. 14, 15. The label indicates that it is a 3m tall tree of the type of a Dendroseris, but whether the collector means Vendroseris s. str. or, and this is quite possible, Phoenzcoseris, we do not know; the photograph fig. 11 shows a palmiform plant which has not yet flowered in spite of having attained a height of 2.;—3 m. If of candelabrum type it ought to have flowered and branched before reaching this size. The morphology of THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 843 the terminal bud is unknown, whether as in its supposed nearest relative, Centaurodendron, or not, but from this it differs by the woolly, long petiolate leaves. Robinsonia gayana Dene, ‘“Resino’’. Compositae. The vernacular name refers to the glassy, aromatic resin exsuded by the stem and formerly together with the resin from X&. ¢thurifera used as a remedy for headache. Endemic on Masa- tierra, by far the commonest Rodzsonta and fortunately the only species found by the casual visitor. Frequent in the montane brushwood, descending to about 200 m above sea-level and often growing in crevices on the cliff walls, ranging west to the slope of Cerro Chumacera (Pl. 66:1; see also Skottsb. 2, Pl. 10: 1—2). Rarely as much as 4m tall, with the main stem several cm thick. Bark gray, I—2 mm thick, ringed by very dense leaf scars; a thin, fibrous wood cylinder surrounds the soft, whitish pith. Leaves soft, thin leathery as in all species. Branches collected in August, 1908, bore 2 innovations at the base of the old corymb, which dated from 1907 or 1906, as it may take two years or more before a shoot flowers. These innovations had a 2.5—3 cm long stem which had shed all its leaves except the apical tuft, ending as in all these species in a characteristic, pointed leaf cone, as illustrated in Skottsb. 3. 59, fig. 12. As a rule 2 innovations are formed, one in the axil of each of the two uppermost leaves, but sometimes only one is developed. The panicles appear in early summer; in December the innovations were small, sessile and 1—2 cm long, or they had already begun to lose the lower leaves, so that the apical tuft was borne on a short, naked piece of stem. Fig. 12 a shows the status at the end of March; | ends in a corymb from 1915—16; II is 2 years old and ends in a fruiting corymb which flowered in December, 1916, unless II has developed in a single season; most of the leaves have been shed, and all will soon fall. Of the 2 innovations (III) one is favoured. Fig. 12 b was drawn from nature 25.3. 1917; the earliest leaves, dotted in the figure, were of reduced size and already dead (see also III in 12a). The genus is dioecious; the corymbs, in spite of being well exposed, are not very conspicuous, the flowers being yellow or dirty yellow in d and greenish yellow in 2. In order to get an idea of the rate of growth a leaf, a, just detached from the closed cone, was marked with a notch on March 25 and the develop- ment of the shoot followed for some weeks: 27.3, 6 detached, ¢ with point free; 31.3, also ¢ free; 2.4, d with the apex free; 8.4 no perceptible change; 12.4 a half free; 14.4 a@ free, apex of e free; 20.4 e free, f with top of point free — the rate of growth is slowing down. Seeds were sown 26.7. 1918 and germinated in October, but the young seedlings died during the winter. Robinsonia evenia Phil. Endemic on Masatierra; not uncommon in the upper montane region but very seldom seen below 500 m. Epiphytic on Dzcksonza and Blechnum cycadifolium; see this volume, no. 7, p. 193. Fig. 13 shows two young plants growing on an old thick Alechnum trunk. 2—3, occasionally as much as 4m tall with a main stem up to 7.5 cm in diameter. Cortex light brownish gray, I—2 mm thick. Morphology and mode of growth as in &. gayana, but the innovations, commonly 2, may attain a length of 3—4 cm already in December; 844 C. SKOTTSBERG 4 A) i) 1 Ae Al ORs Wh va lq ny Mie Gf Ne \ Veen GTA rf iy Mdyh Fig. 12. a Robinsonia gayana, apex of branch at end of March 1917; diagrammatic. I with old terminal inflorescense from 1915—16; II fruiting branches; III innovations, of which the larger one is supposed to flower in Nov.—Dec. the same year. — 6 the same, innovation 31/3 1917, showing withered reduced leaves at base. 1/2 nat. size. — c juvenile leaf of Robinsonia eventia, nat. size. — d, e two seedlings of Robinsonia gracilis, epiphytic on Dicksonia. Nat. size. — f top of fruiting branch of Symfhyochaeta macrocephala, showing position of innovations; diagrammatic. the first internode is long, as much as 3 cm and carries a leaf of reduced size. The first leaves on the main stem differ considerably from the later ones, with which they are united by transitions; they are succulent and densely hirsute, and the limb is oval. The difference is seen in the specimen to the right in Fig. 13, the left-hand one having shed all the juvenile leaves. One juvenile leaf is shown in Fig. 12 c. The definite leaves are perfectly glabrous as in other species but distinctly thicker, an adaptation to the epiphytic life. Flowers Dec.—Feb. Robinsonia Masafuerae Skottsb. Endemic on Masafuera, where it is scattered in the highland between 650 and 1370 m, preferring the fern-beds along the gullies. Hardly surpassing 2m in height, with a short main stem. A well-developed THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 845 a Fig. 13. Two young plants of Rodimsonia evenia growing as epiphytes on Llechnum cycadifolium ; the right-hand specimen with juvenile, hirsute lower leaves. Upper slopes of Villagra, Masa- tierra. — Photo Io0/I 1917. specimen — these are very rare, most were damaged by goats and the species was undoubtedly more common before this pest was introduced — was 1.7m tall with the main stem 17 cm long to the first dichotomy and 7 cm thick, the 846 C. SKOTTSBERG branches 0.92 and 1.12m long and 4.6 and 2.7 cm thick, respectively, with subumbellate innovations, 7 and 6 in number, respectively. This is the plant illustrated on Pl. 17:1 of no. 7 in this volume; it was an exception from the rule, also observed by this species, that only 2 or 3 innovations are formed. End of February they were in full growth. The nearest relative is R. eventa, and like this it occurs as an epiphyte on Dzcksonia and also on Lophosoria, but as a rule it is not epiphytic. Flowers Jan.—Feb. Sterile specimens were seen in the forest, an unsuitable habitat for a Rodzmsonza; stem thin, with long internodes, leaves long, flaccid. Robinsonia thurifera Dene, “Resino macho” (=male r.). Pl. 67. Endemic on Masatierra and very rare, only found on the slope of Cordon Chifladores above Puerto Frances, 350——450 m, and near Portezuelo de Villagra, 550—600 m. Largest of the genus, up to 5 or even 6 m tall with a trunk 5—10 cm thick; the largest tree seen by us had a stem 1.53 m long and 18.5 cm thick. Large specimens had g generations of branches. Often trichotomous. Flowers later than any other tree or shrub in the islands; bloom at its peak in March—April, when also the innovations are well developed. On March 15 a comparison was made with FR. gayana (also see above); in R. thurifera, the flowering shoots still bore all their leaves, only the oldest ones turning yellow, and the innovations were less advanced. They did not, as in &. gayana, begin with a number of re- duced leaves. A single epiphytic plant was observed, growing on Alechnum cycadtfolium. Robinsonia gracilis Dene, “‘Resinillo’’. Endemic on Masatierra and inhabiting the same kind of stations as the former, but more common and ranging from the ridge between the Piedra Agujereada and Laura valleys to Cord6én Salsipuedes. Not seen by us below 500 m, but reported from 400 and 450 (Skottsb. 79. 362). Once numerous seedlings were found growing on Dicksonza (Fig. 12 d, e), and on one occasion a flowering specimen on Blechnum cycadifolium. It is possible that the large plant mentioned below had germinated on the base of a Dicksonza, against which its stem was tightly pressed, later sending roots into the soil, but in other cases I had no reason to suspect that it had started as an epiphyte. Mostly of lower stature than its congeners and easily recognized by the small, narrow, rather stiff leaves, different in shape from the juvenile ones seen in Fig. 12. Pl. 68: 1 shows a typical specimen, 1—2 m tall with a main stem 20—50 cm long and 5—8 cm thick. A plant 1.6m tall had 16 generations of branches. One exceptionally large specimen measured 4 m, with the trunk 17 cm thick. Young bark greenish, later gray. Seen with resting leaf buds of the usual type in August, with flowering heads from the middle of December to the end of January. The leaves stay on longer than in the other species, but are gradually shed up to the corymb before the winter. Towards the end of December the innovations, 2—4 in number, are about 1 cm long; in April they measured several cm and the lower leaves had been shed; the first few are of smaller size and short-lived. The apex was still growing and new leaves formed. Symphyochaeta macrocephala (Dene) Skottsb., “Incensio’ (a corruption for incienso: the resin was once used not only as medicine, but also as incense). Also THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 847 in its mode of occurrence this remarkable species differs from Rodzsonza, from which I segregated it (23). It is very rare and grows on the low hills and ridges below the forest at 300—350 m above sea-level, ranging from Puerto Frances to Pangal. The habit is the same as in Rodinsonia, but more straggling (Pl. 68:2). According to Johow it grows to a height of 3—5 m, but we found no specimens surpassing 2 m. Bark about 1 mm thick, gray with brown leaf scars; wood cylinder 1 mm thick, filled with a dense, farinaceous, yellowish pith. On Dec. 13, when we came across Symphyochaeta for the first time, the male plants were long past flowering and the female on the decline; there was no sign of any gemmae which, in Rodzxsonza, are well advanced during the flowering season. Fig. 12 f shows a fruiting branch at the end of March, when the involucre was bursting open and the achenes ripe. The branch is leafless but bears 3 innovations (II), one very much smaller than the other two, but they are not placed in the axils at the base of the ped- uncle but toward the base of the branch. They cannot very well reach maturity and produce an inflorescence already next spring, but will need at least a year and a half before this happens. On Pl. 68:2 several long, vegetative shoots are seen, very likely 2—3 years old. Seeds sown 22.3. 1918 gave 4 seedlings in the beginning of May, and 1.10 the tallest plant was 50 cm high, but unfortunately all died during the winter. Rhetinodendron Berterit (Decne) Hemsl., ““Resino hembra” (female r.), “‘incensio’”’. Endemic in the elevated central part of Masatierra, in open brushwood and forest along the ridges, rarely observed below 500 m but found down to 400 m on the south side, where it extends west to Cerro Chumacera. Also on the summit of El Yunque, where unusually large-leaved plants occur (Skottsb. 74. 159, fig. 4). Johow states p. 55 that Rhetznodendron often grows as an epiphyte on Decksonia, and on p. 250 that “a veces se encuentra tambien en el suelo pero con preferencia, i no solamente en su primera edad sino tambien como arbol crecido, habita los troncos de la Dicksonia berteroana ...’’; a photograph taken in the forest on the south side of Portezuelo (Pl. XVI) shows a tall Dicksonza, and the explanation p. 287 tells that a specimen of Ahetinodendron grows on the base of the fern. The long- stemmed supposed epiphyte is distinctly seen, but unfortunately the lower portion of the trunk is completely hidden; still we have no reason to doubt the correctness of Johow’s statement. But was it Rhetzxodendron? There is little difference, when seen from a distance, between this and Rodzusonia evenia, which, as we have seen above, is a regular epiphyte, but Johow tells (p. 60) that he failed to find this species. On the photograph is a much smaller plant, probably of the same kind, attached to the fern trunk a little below the top, but not referred to by Johow. His statements made us pay careful attention to the question of epiphytism; we came across 2 seedlings on the trunk of Zhyrsopteris, but were unable to find an adult, epiphytic specimen. Neger makes the following curious statement (I. c. p. 458): “Auf der Robinson-Insel Juan Fernandez siedelt sich die sonst terrestrisch lebende baumartige Komposite Retinodendron Berterii nur auf dem Baumfarn Dicksonia Berteroana an’) — but Rhetinodendron is only found on Masatierra. Height according to Johow 2—4m, but trees 5 m tall were seen more than 848 C. SKOTTSBERG Fig. 14. Old, not flowering plant of Rhetinodendron Berterii near Portezuelo de Villagra, Masa- tierra, c. 550 m above sea-level. — Photo 12/4 1917. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 849 Fig. 15. Dendroseris litoralis in Mr. Charpentier’s garden, Masatierra; a young, unbranched, not yet flowered plant, 1 1/2 years old. — Photo 14/4 1917. once, and one unusually large plant attained 8 m with a long main stem 8 cm thick. Internodes of the flowering shoot longer than in Roézvsonza and leafy during the flowering season Dec.—February. Bark very thin, dark brown, wood and pith as in Symphyochaeta. Innovations as in Robimsonza, 2—4 in number. On a branch collected in August, 1917, the innovations had the first internode up to 3 cm long, immediately followed by shorter ones, 3—5 mm, all leafless; the large terminal bud, cone-shaped as in Rodzusonia, was 5.4—7.5 cm long and stil] resting. In Fig. 14 is an old specimen in April, showing no sign of having flowered during the previous season, and a branch may require several years before it produces the terminal panicle. Fruiting trees in early autumn (April) bear still growing innovations. Rhetinodendron is dioecious; heads very numerous, but smaller and much more inconspicuous than in Rodznsonia. The four species of Dendroseris s. str. (Skottsb. 23) are nearly related to each other and of much the same general appearance. The young shoot is solid, with a wood cylinder 5 mm thick or more, and a soft pith, the older ones hollow, with a thin diaphragm across the nodes. Leaves large, soft coriaceous, bright green and glabrous, the limb revolute in bud, protected by the broad sheath of the older leaves. Heads large, involucre glaucous, flowers bright orange yellow. C. SKOTTSBERG Fig. 16. Large, several years old Dendroseris litoralis in the Goteborg Botan. Garden. Exyngium bupleurotdes in front. — Photo 6/11 1933. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 851 Fig. 17. Diagrammatic sketch of a flowering shoot of Dendroseris litoralis cultivated in the Goteborg Botan. Garden. Main axis dotted. Dendroseris litoralis Skottsb. Only known from Morro Vinillo, a rock off the south coast of Masatierra, and from the coast cliffs of Santa Clara with the barely detached islet Morro del Spartan (or de los Alelies), Pl. 103. To 1.5—2 m tall with a main trunk about 10 cm thick. The cultivated plant in Fig. 15 was I 1/2 years old and showed no sign to flower; another, 5 years old and illustrated on Pl. 19 (no. 7, this volume), had flowered for the third time. Bark smooth, bright green and lustrous the first year, then turning brown. In bloom from Novy.-Dec. to Feb.-March. Seeds sown 27.3. 1918 gave seedlings 20.4; sown again (of the same portion) 28.3. 1919, the first seedlings were obtained 15.5 (Fig. 19 a). One year old plants match the specimen in Fig. 15. All suffered badly from mildew which frustrated all our efforts to get rid of it, and no specimen lived to, produce flowers. Some years later fresh seeds were sent from Masatierra and numerous new plants raised which, although attacked by the fungus, grew well; one is seen in Fig. 16. It was cut back and in this way induced to branch, the fastgrowing branches supplying cuttings, evidently an easy mode of propagation in Dendroseris. One cutting, potted in 1936 and transplanted 852 C. SKOTTSBERG Fig. 18. Dendroseris macrantha, Portezuelo de Villagra, Masatierra. — Photo 22/1 1917. free in the temperate house in 1938, was very fast-growing and had to be cut back; it formed several branches and flowered in July, 1940. The height was 3 m, the length of the main stem 2.3 m, with a thickness of 9.6 cm at the base and 6.1 cm at the top. Fig. 17 is a diagrammatic sketch of a flowering branch. The internodes are much longer than in nature and the innovations, 6 in this case, twice or three times as many. One of these was potted 15.3. 1941 and flowered from 1.6 to 7.7. 1942. With its orange-coloured head the size of a dandelion, and glaucous involucre, it is quite . attractive and deserves the attention of botanical gardens. Dendroseris marginata (Bert. ex Dene) Hook. et Arn. Endemic on Masatierra, very rare, at present only known from a few localities in the centre of the island, 450—600 m above sea-level. Very like DY. /toralzs in all respects; the specimen illustrated on Pl. 18 (this vol., no. 7), was 1.25 m tall with a stem diameter of 3.5 cm. Dendroseris macrantha (Bert. ex Dene) Skottsb. Masatierra, only known from the vicinity of Portezuelo de Villagra and probably on the verge of extinction. A fair-sized tree as observed by Bertero (see p. 202 of this volume), but two small plants growing among shrubs was all we were able to discover (Fig. 18). Agrees with the former in all essential characteristics. Dendroseris macrophylla D.-Don. Endemic on Masafuera, scattered on rock ledges in the canyons and on lofty coast cliffs; highest station noted about 700 m. Most specimens seen were 2 or 3 m tall with trunk not over 5 cm thick (Skottsb. 3, Pl. 4:2), but in better soil D. macrophylla surpasses its congeners, attaining a THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 853 height of 6 m with the trunk at least 15 cm thick. Bark very thin, only 0.5—1I mm, brownish gray. With flowers from January to March or April. The 3 species of Kea, reestablished as an independent genus, are nearly related to each other. The stem is solid and woody throughout, the flower-heads much smaller than in Denxdroseris s. str. and of different shape, and the flowers milky white. All are confined to Masatierra. Rea micrantha Bert. ex Dene. Quite common all over the wooded country and of physiognomic importance in the montane forest and brushwood between 400 and 600 m, extending west to the slope of Cerro Chumacera, but also thriving on the lower barren slopes and open spaces in the macal belt, where it is a dwarf tree rarely over 2 m tall with much smaller leaves and panicles. In the forest it grows to a height of 6—8 m with a trunk of I—1.5 m and 20—30 cm thick. On large trees about 10 generations of branches were counted. Young bark smooth, grayish or yellowish green, turning brownish gray, with distinct leaf scars; old bark dark gray, 4—5 mm thick, divided into rectangular pieces and often covered with mosses and lichens. On Pl. 1:2, Skottsb. 3, is a fair-sized tree in August, every branch ending in an expanded, sterile rosette. The terminal bud forms a pointed, closed cone; the young leaves are pubescent, the old ones glabrous and papyraceous. The panicles begin to unfold in December and come into bloom end of January; flowering lasts through February and March, in higher altitude until the middle of April. Mature achenes are found from end of March. Pl. 69 shows a good-sized tree in full bloom. With its large, dense panicles of innumerable white heads it offered a beautiful sight. Johow’s pl. VI, drawn by an artist, is a poor representation of this fine species; besides, the shape of the leaf is not correct. The flowering shoot loses its leaves during the summer; the innovations, 2 or 3 in number, are situated low down on the shoot as in Symphyochaeta (Fig. 19 ¢); Fig. 19 d shows one season’s growth. Long branches not having reached flowering stage must be several years old; comp. the plates cited above. At the timber-line in one of the hanging gulches on Cordoén Salsipuedes I came across a large specimen (21.4. 1917) which had flowered profusely and then died. Seeds sown 27.3. 1918 gave a few seedlings in the beginning of May (Fig. 19 b). They grew quite well for a couple of years, forming numerous axillary buds on the stem, which shed its leaves as usual. Also in nature additional innovations are observed, causing the branching to become irregular. All the cultivated plants died without having flowered. Rea pruinata (Johow) Skottsb. A seaside species, scattered in the dry ravines, west to Bahia Chupones and also found on Santa Clara. Very near the former, but much smaller and often only 1—1.5 m tall. Flowers earlier, found in full bloom already in the beginning of January. A few stunted specimens are seen on FleAto3-T: Rea nertifolia Dene. Confined to the quebradas between Bahia Cumberland and Puerto Frances according to Johow, but in spite of a careful search we were unable to find it. On p. 70 Johow gives a detailed description: 2—4 m tall with a main stem 10—30 cm thick, leaves stiff coriaceous, not crowded toward the branch tips, evidently lasting longer than in the other species. SKOTTSBERG 854 G. Fig. 19. a seedling of Dendroseris litoralis. Nat. size. — 6 seedling of Rea micrantha. Nat. size. — ¢ flowering shoot of Rea micrantha in the end of March, diagrammatic. — d stem of an innovation in the end of March; a not lignified, bark green, developed in late summer; 4 lignified, bark brown, probably developed in early spring or in the preceding autumn. Nat. size. — e Eryngium bupleuroides in April; diagrammatic. —_/ seedling of Phoenicoseris pinnata, epiphytic on Llechnum cycadifolium. Nat. size. — g the same, juvenile leaf of a slightly older plant. Nat. size. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 8: Ar wal Fig. 20. Eryngium bupleuroides (a) and sarcophyllum (b) in cultivation and flowering for the first time in the Goteborg Botan. Garden. — Photo 17/10 1920 (a) and 3/3 1922 (b). Hesperoseris gigantea (Johow) Skottsb. Endemic on Masafuera; scattered on the cliffs in the valleys and through the forest region; lowest and highest stations observed, 200 and 880 m. A small to middle-sized tree, the largest seen well over 8 m with a solid trunk 28.5 cm thick 1 1/2 m above the ground. Leaves larger and thinner than in Rea, the juvenile ones very large (p. 208). Pl. 106: 2. Found in full bloom in February; the enormous panicle with its innumerable creamy white heads makes a fine show. The innovations with a leafless stem about 4 cm long, crowned by a growing rosette; often one takes the lead and continues the mother shoot, resulting in a habit different from that of fea. Eryngium bupleuroides Hook. et Arn. Endemic on Masatierra, not uncommon in open woods and along the high ridges from Cordon Centinela to the ridge be- tween Vaqueria and Juanango and to the south ‘face of Cerro Chumacera, but rarely observed below 400 m (down to 200 m according to Johow); ascends to the summit of El Yunque. A typical candelabrum tree, up to 3 m tall, very densely and regularly branched and forming a hemispherical canopy (Pl. 70). Main stem short, 0.5—1 m long, or even shorter, with a maximum diameter of 10 cm. Young bark light 856 C. SKOTTSBERG brown with prominent white leaf scars, old bark blackish, 5—7 mm thick, deeply and coarsely furrowed. Leaves coriaceous. Collected in August 1908 and 1917 with flowers, fruit and growing innovations (Skottsb. 3. 58, fig. 11) which, to judge from observations made in 1916, flower in the beginning of December and bear a new set of innovations in the uppermost axils; in the middle of the month stronger ones had a leafy stem 9 cm long with internodes of 1—1.5 cm, while the fruiting mother shoot was losing its last leaves. Innovations of this size will flower later during the summer, so that 2 generations of branches are formed each season and a third during the winter. Fig. 19 e illustrates a branch examined April 24, 1917: I, now in fruit, is supposed to have flowered in January; of the three branches marked II, one is quite small, but two are in full bloom and bear growing buds (III). Seeds sown 22.3. 1918 gave 7 seedling in the beginning of May, of which the first flowered in October, 1920, and produced fruit; see Fig. 20 a, where 2 strong innovations are seen at the base of the peduncle. Grows well in a temperate house and is easily propagated by cuttings. Eryngium inaccessum Skottsb. Only known from the precipitous cliffs of Portezuelo de Villagra, Masatierra, almost 600 m above sea-level. A miniature tree with the same morphology as the former, but very slow-growing and hardly much taller than 1/2 m, with a main stem 20 cm long and 2—2.5 cm thick. Bark dull brown, ringed like an earthworm (PI. 12:2, this vol. no. 7). Leaves stiff cori- aceous, with pungent teeth; for particulars, see |. c. 158. The material secured (25.12. 1916) was past flowering, with immature fruit. Two innovations seem to be the rule, but in cases only one is developed; they were short and sessile at the end of December. Bud protection as in the former. The leaves are not shed as early as in this but stay on long after they have become brown and dry. Lryngium sarcophyllum Hook. et Arn. Confined to the coast cliffs of Masa- fuera, rare. Johow calls it a shrub or subshrub, but its architecture is exactly the same as in the former two; see Skottsb. 3, pl. 5:5, where an entire specimen, collected 28.8. 1908, is figured. A miniature tree not over 40 or 50 cm tall; main stem 10—20 cm, about I cm thick, more or less clothed with old leaves. Leaves linear, very succulent, resembling the leaves of a AVezuza or Mesembryanthemum. The plant figured l.c. is regularly dichotomous, with the old, dry head of the main stem still preserved, and the two vegetative-floral branches each with two inno- vations. To judge from our observations in February 1917, these innovations probably flowered in Jan.—Feb. 1909, and as the flowering specimens seen in Feb.—March 1917 had well developed innovations, three generations of fruiting branches may be formed during one year. Seeds sown 21.9. 1918 rested until March rgrg, when a few germinated; others, sown 27.3. 1919, produced seedlings in May the same year. Most of them died soon, but one flowered in February 1920. This plant is seen in Fig. 20 b. There is one axillary shoot at the base of the peduncle, another in the axil of the lowest leaf still left, a third at the point where the series of longer internodes begins, and several sprouting from the base of the stem and forming a dense tussock. Nothing like this was observed in nature. Unfortunately the specimen soon died and the seeds did not germinate. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 85 ~I b. The Palmiform type. Phoenicoseris pinnata (Bert. ex Decne) Skottsb. Endemic on Masatierra, not uncommon on the steep slopes more or less covered with low trees and shrubs, rarely seen below 400 and ascending to about 800 m; ranges from the ridge be- tween the Piedra Agujereada and Laura valleys to the central ridge in Valle Ingles; also on the upper slopes of Villagra. A small tree, hardly over 4 m tall, the largest specimen seen by us with the hollow trunk 10 cm in diameter. Bark about I mm thick, light brown to gray, densely ringed with prominent leaf scars; old trunks with numerous crustaceous lichens. Leaves large, membranous, pinnate, with deeply forked pinnae (Skottsb. zz. 167, fig. 16). This and its congeners are the only monocarpic ligneous plants in the islands; with regard to Yunquea and Hesperogreigia, see these. We do not know how long it takes before the formation of the large panicle ends the life-cycle; probably many years. In 1908 one dead tree was found which had flowered in 1907; it is seen on Pl. 10, Skottsb. 2, accompanied by two younger specimens; another, still sterile, tall and presumably old, is shown on Fig. 21. An indication or, as it were, a proof that a long time passes before the flower-stage is reached, is that extremely few specimens flower in the same year, and the panicle is so conspicuous even from a long distance that it cannot fail to be noticed. Johow saw a single tree in bloom during his stay in the islands, and exactly the same happened to us. Flowers Dec.—Jan. In August 1908 | found the limit between the numerous mature leaves and the resting term- inal bud quite sharp, in April 1917 young leaves were unfolding in rapid suc- cession. Branched specimens are, as stated by Johow, very rare; two cases were noticed by us, one with 2, the other with 3 short branches at the apex. Neither of these plants had flowered. I presume that the apex had been damaged, stimul- ating dormant axillary buds to develop; to make sure, it would have been neces- sary to cut the trees. Seedlings (once an adult spm) were found on Dicksonia or Blechnum cycadifolium; one is seen in Fig. 19 f, and a juvenile leaf in 19 g. Phoenicoseris berteriana (Dene) Skottsb. Only in the upper wet montane forests of central Masatierra, also collected on the summit of El Yunque, very rare. A little-known species, of the same habit as the former, but smaller. Found with flowers once only, by Bertero. Leaves larger than in the former (Skottsb. 74. 167, fig. 17), the juvenile ones simple with only the reniform end-lobe developed. Phoentcoseris regia Skottsb. Masafuera, in the highland above the forest, from about 750 to about 1400 m above sea-level, often associated with Dzcksonia externa and occasionally epiphytic. Probably more frequent in former times, now scarce and more or less confined to places not easily accessible to the goats. Habit as in the former, but more robust (Pl. 17: 2, this vol. no. 7); trunk to 1.7 m tall (largest measured 2 m) and 13 cm thick; bark reddish brown, very thin, wood cylinder thin, and of the broad pith only the diaphragms remain. Juvenile leaves to 1 1/2 m long, adult to 65 cm. Seedlings as in the former. Flowers unknown; see |. c. 206. Flantago fernandezia Bert. Endemic on Masatierra, only observed above 500 m in the brushwood of the high ridges around Bahia Cumberland and on the summit of El Yunque. Rare. Stem prostrate-ascending-erect, rooting at the lower nodes, 57 — 516796. The Nat. Hist. of Juan Fernandez and Easter Isl. Vol. II. 858 C. SKOTTSBERG Fig. 21. Phoenicoserts pinnata in the brushwood on Cordén Salsipuedes, Masatierra; c. 2m tall. — Photo 8/12 1916. leafless below the terminal rosette, longest measured 2.5 m long and 2—2.5 cm thick. Bark grayish brown, about 1 mm thick, with dense leaf scars. Leaves soft coriaceous. Growth unlimited, but probably arrested during the winter months, as a conspicuous terminal bud was observed in August. Flowers Dec.—March. PE P7 Th Hesperogreigia Berterot (Skottsb.) Skottsb. Bromeliaceae. Endemic on Masa- tierra, found once only by us, in the wet fern-forest above Pangal, alt. about 660 m, none of the few plants seen showing any sign of having flowered. Habit of Plaxtago fernandezia; stem about 1 m long and about 3 cm thick, covered with the sheaths of old leaves, rooting along the trailing base, ascending-erect and ending in a THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 559 Fig. 22. Evigeron fruticosus at edge of forest, Quebrada de las Chozas, Masafuera. — Photo 3/3 1927. rosette of long, ribbon-like leaves. The inflorescence was discovered in January, 1935, and described by me (78), but as it had been detached from the plant, I could not decide if it is terminal and the species monocarpic, which seems more likely, or lateral. It was in fruit. c. Other dwarf or miniature trees. Erigeron fruticosus DC. Masatierra, from near the sea to the high mountains, in rocky or stony soil, not uncommon; not seen in the eastern part of the island but extending west to Cerro Tres Puntas and Bahia Chupones. Masafuera, scat- tered on the barrancas and ridges to about 750 m. Stem 1/2 to 1 m long and 6.5 cm thick, branching rather irregular, but sometimes in the fashion of a typical, miniature candelabrum tree, forked below the terminal inflorescence (Fig. 22); in other cases with several alternating lateral branches. Leaves more or less crowded toward the end of the branches, but not forming a dense rosette. With flowers from November to March. Fig. 23 a shows the top of a branch in the beginning of March, with a terminal, pubescent bud, perhaps ready to resume activity after a period of rest; in April I observed numerous shoots in full growth. On the other hand, material collected in August 1908 corresponds to Fig. 23 a, while branches secured at about the same time in 1917 had started to grow. 860 C. SKOTTSBERG Fig. 23. a apex of shoot of Evigeron fruticosus in August. Nat. size. — 6 innovation of Cheno- podium nesodendron in February. Nat. size. — c seedling of the same. Nat. size. Selkirkia Berterot (Colla) Hemsl. Boraginaceae. Confined to Masatierra and only on the high ridges around Bahia Cumberland between 400 and 600 m above sea-level; very few individuals seen in each of the few localities. “A shrub, 1—2 m tall’ (Johow), but Se/kzrkza is a miniature tree, seldom over I m high, with a short main stem 1—4 cm thick; bark dull brown, longitudinally striate. Flowers early, past flowering in middle December. Inflorescence terminal on branches two years old and several dm long with internodes of about I cm, second years’ growth still leaf-bearing and with a number of open, silky buds. This stage corresponds to Hemsley’s Pl. 57. End of January the innovations were about 1 dm long and leafless up to the terminal rosette, surrounded by a number of withered leaves. Chenopodium Sanctae Clarae Johow. Endemic on Santa Clara, almost confined to the small islet Morro del Spartan. A miniature tree 1/2 to 1 m tall with a main stem about 30 cm long and “‘as a thick as a man’s arm’’ according to Johow, I cannot remember having seen it as thick as that. Leaves crowded toward the end of the branches; panicle terminal. Specimens collected 26.1. 1917 bore old dry panicles. Collected by Johow with flowers in June. Chenopodium crusoeanum Skottsb. Discovered in a single locality on Masatierra, growing on the perpendicular canyon wall in Pangal, 220 m above sea-level. Of the same habit as the former but smaller, barely 1/2 m tall; stem and primary branches about 1 cm thick. With flowers and fruit 1.1. 1917. Pl. 11, this vol. no. 7. Chenopodium nesodendron Skottsb. Endemic on Masafuera, on open, grassy slopes in the forest region at about 500 m, and in a gorge on the west side at 300 m above sea-level. Very rare, undoubtedly due to the ravages of the goats. A dwarf tree with a distinct main stem to 2.8 m long and 11 mm thick near the THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 861 Fig. 24. Three years’ old plant of Chenopodium nesodendron, grown in the Goteborg Botanical Garden. — Photo 3/3 1922. 862 C. SKOTTSBERG base; bark green to straw-coloured. Branching irregular, from near the base; leaves short-lived, thus more or less crowded toward the end of the branches, were the internodes are shortened (fig. 23 b, 24). With ripe seeds in the beginning of March. Seeds sown 23.3. 1918 germinated after a month; a 2 months’ old seedling is seen in Fig. 23 c. A year later they had grown to miniature trees, and in March 1920 budding panicles appeared, but they soon faded and fell. In January 1922 one plant flowered, and the central panicle gave ripe seeds (Fig. 24), whereupon the plant died, and the seeds appear not to have germinated. Another series of plants was raised from seeds sown 28.3. 1919, but all were lost before coming into bloom. Urtica fernandeziana (Rich.) Ross. Both islands, very rare on Masatierra and only reported from the forests on both sides on Portezuelo de Villagra, less uncom- mon on Masafuera in some of the wooded valleys between 400 and 800 m and also observed in the highland to about 1350 m. Johow calls this insular, non-stinging nettle a shrub or subshrub, but it has the habit of a miniature tree with a max- imum height of 1.7 m and a short main trunk as much as 3 cm thick, and I doubt that it regenerates from the base. Bark long remaining green, cork developed on older branches, which are woody throughout. Pl. 72:2. With flowers and small axillary shoots in August, 1908 (Skottsb. 3, Pl. 6:5); with flowers and fruit in February 1917. TALL SHRUBS (Mi—N). Chusquea fernandeziana Phil., ““Colihue’’. Gramineae. Only in the forests of Ma- satierra, from the wet gullies back of Puerto Frances to Puerto Ingles and Cerro Alto, rare below 400 m (descending to 160 m on the bank of a stream in Villagra), not uncommon between 500 and 800 m. The slender, spreading canes of the island bamboo grow to a length of several m with a diameter of at least 3.5 cm, forming light green cascades on the steep slopes (Pl. 74:1). Neither in 1908 nor in 1917 flowers or old panicles could be detected. As customary among the bamboos, flowers are produced at long intervals; known from the type material only, collected in November, 1864, until fresh material was obtained in 1935 (Skottsb. TO 305): Berberts corymbosa Hook. et Arn., ‘““Michay”’. Endemic on Masatierra, not uncommon in the forests and brushwood from 200 to 600 m above sea-level and ranging west to Quebrada Juanango, but never abundant and of slight physiog- nomic importance. Habit of our common European barberry, the erect to slightly curved long-shoots up to 2.5 m long and 1.5—4 cm thick (in one instance as much as 15 cm). Young bark dull cinnamon brown, older grayish brown and furnished with long, narrow, anastomosing ridges. Wood yellow, very hard and heavy. One of the very few native ligneous plants with typical budscales. Johow says that Berberis stands leafless during July and August. Specimens collected 23.8. 1908 were bursting into leaf and had young racemes; in the beginning of December the flowering season approaches its close, and ripe berries were found toward the end of the month. End of March many shrubs were losing the leaves of the older dwarf-shoots, while the long-shoots developed during the summer showed no sign THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 863 of shedding, and in April other shrubs were still foliated, but as I had no oppor- tunity to make observations during the winter months, and as Johow visited the island in June, it is probable that Leréberzs offers the only example of a deciduous species in these islands. Berberis masafuerana Skottsb. Only known from two of the gorges in Masa- fuera, growing on precipitous cliffs; a single seedling was found in the bottom of the Casas valley. A little-known species, nearly related to the former, but more slender. Bark cinnamon brown to dark violet. Flower and fruit unknown. Colletia spartiotdes Bert. ex Colla. Rhamnaceae. Endemic on Masatierra, growing on the steep barrancas in some of the valleys from Puerto Frances to Puerto Ingles, very scarce. Lowest and highest stations noted, 220 and 500 m. A several-stemmed shrub or short-stem tree with I—2 m long branches, straggling and pendent from the rocks. Bark green, when young finely pubescent, later corky. Leaves thin, soon caducous. Fig. 25 d represents a leafy branch in January, ending in a spine, pungent but quite innocent as compared with the armament of the continental species, and bearing axillary, simple spines or spine-systems, with leaves as a rule reduced to the minute stipules. This branch will not continue its growth unless a serial accessory bud is developed below the spine, as the figure shows; a second accessory branch may be formed below the first. The solitary flowers replace spines or are borne on minute accessory shoots. Fruits from January on, flowers in the autumn. Seeds sown 7.8. 1918 germinated next spring, but the seedlings died during the following winter. A new trial gave better results, and one specimen, planted free in a bed in the succulent house, grew to a large size and flowered in October, 1927. No seeds were produced. Still alive in 1952. Pernettya rigida DC., “Murtillo”. Ericaceae. On both islands, on Masatierra a dominant member of the scrub above the forest, in cases forming low impenetr- able thickets and extending along the rocky ridges down to about 150 m above sea-level, but scarce in the western barren country. Scattered in the highland of Masafuera, locally abundant but rare below 500 m; on the slopes of Los Ino- centes to about 1400 m. A much-branched shrub with small and stiff, coriaceous leaves and 1—2 mM ttall, occasionally a short-stem tree up to 4 m high with a low main stem, 5—-1O cm thick near the base. Regenerates with long, vigorous shoots from the caudex. Plants examined in August 1908 and 1917 were resting, most apices being dead; the apex in Fig. 25 a is supposed not to develop further. Axillary buds covered by about 9 regular bud-scales. In the beginning of De- cember many plants were in full bloom on the sunny exposed ridges, while still in bud at the edge of the forest. Axillary innovations, quite conspicuous on ac- count of their crimson colour, were then developing below the vegetative-floral region; at the base they are surrounded by persistant bud-scales, followed by a few reduced leaves. Flowering continues through Jan.—March (PI. 72: 1); ripe, dark red berries in March—April; by this time’vegetative growth has ceased, and most of the apices are dead. Stronger shoots will start to flower in November, the bud- scales functioning as bracts around and above the base of the pedicel, in rare in- stances supporting lateral flowers (P. racemulosa DC.) 864 C. SKOTTSBERG Fig. 25. a top of branch of Pernettya rigida in August. 2x. — 6, c two branches of Ugmi Sel- kirkii in the end of January. Nat. size. — d@ growing shoot of Colletia spartiotdes in January; pubescence of stem left out. 2x. — e Resting apex of Halorrhagis masatierrana in August. Nat. size. — f, g seedlings of Halorrhagis asperrima. Nat. size. — h seedling of Halorrhagis masafuerana, Nat. size. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 865 Seeds sown 25.3.1918 gave a few seedlings during the first weeks of May, but the young plants soon died. Ugni Selkirkii (Hook. et Arn.) Berg, “Murtillo’. Myrtaceae. Endemic on Masatierra, not uncommon along the elevated ridges but not observed by us much lower than at 450 or 500 m. A densely branched shrub, 1—2 m high, rarely a short-stem tree as much as 3m tall and with a main stem a few cm in diam- eter. Leaves almost glabrous, small, stiff coriaceous, with yellow transparent dots. The monopodial shoot begins with a pair of smaller leaves, followed by I—4 pairs of regular size and ends in a naked, hirsute terminal bud. Second year's growth will repeat the same sequence (Fig. 25 b), bear solitary flowers in January from the lowest nodes and then continue its growth, or remain sterile, and third year’s growth carry the first flowers. Generally only 2 (Fig. 25 c), but sometimes 4 or even 6 flowers are found on a shoot. Innovations developing with the flowers. Fruiting branches, collected in August 1908, were resting. Nicotiana cordifolia Phil. Endemic on Masafuera, scattered along the rocky shore and at the entrance to some of the canyons, very scarce, but not so rare as supposed by earlier visitors. A straggling shrub or short-stem tree with term- inal and axillary panicles borne on the annual shoot; colour of corolla a brown- ish purple. Large specimens 1.5—2 m tall (Pl. 73), branched from near the base, the short main stem 5—5.5 cm thick. Leaves large, gray tomentose. Material collected 28.8. 1908 (Skottsb. 3, pl. 6:4) in late bloom and with numerous in- novations in various stages, of which those in the upper axils are large, with a stem 10—15 cm long, and will provide for the ramification. The most advanced ones had small flowerbuds. In full bloom and with ripe capsules in Feb.—March 1917. Possibly the flowering season extends over most of the year. Numerous seeds were sown 27.3. 1918 and 2 seedlings obtained after seven weeks; growth was rapid and both flowered in November but died during the winter. Sown a second time 28.3. 1919 with better result; the first plants flowered in October and bore ripe seeds in November. Two stout specimens were still alive in October 1924, but died during the winter, and propagation by cuttings failed. SMALL TO DWARFISH SHRUBS, AS A RULE UNDER I M (N-Ch). Lactoris fernandesiana Phil. Lactoridaceae. Endemic on Masatierra and con- fined to the wet fogswept forest belt from Valle Frances to Valle Ingles above 500 m, everywhere very scarce though not quite so rare as has been supposed. A densely branched, more or less hemispherical shrub or, perhaps more correctly, a miniature short-stem tree, because a very short but always distinct main stem is present; maximum thickness 1.5 cm. Pl. 75:1. Tallest individual obser- ved 1.15 m. Sprouting in August, the tender glabrous blades folded along the midrib and covered by the hyaline ochrea of the next oldest leaf (Fig. 26 a). Terminal and axillary, 1—3-flowered monochasia from November to January (Figs. 26 b, c); flowers either $ or 2 (also 5, Hemsley, Pl. 59), the axillary ones with a dorsal bracteole. Ripe seeds end of April. Margyricarpus digynus (Bitter) Skottsb. Rosaceae. Endemic on Masatierra, not uncommon and sometimes abundant along the barren stony ridges below and 866 C. SKOTTSBERG Fig. 26. Lactoris fernandeziana: a top of branch in August; 4 terminal and ¢ axillary inflores- Gense,, 2°. above the forest, also in the grass-lands on the south side, where it extends west to Punta Larga. An intricately branched dwarf shrub with prostrate-ascending main branches, 10—20 cm tall or more. Internodes very short, leaves short-lived and crowded toward the branch-tips. New leaves and flowers from August to late summer. Halorrhagis masatierrana Skottsb. Endemic on Masatierra, in the same kind of stations as Margyricarpus, rather common and ranging from Rabanal to Jua- nango on the north, and to Bahia Chupones on the south side of the island, and from 150 to 600 m above sea-level. A small suffruticose shrub, 20—50 cm tall, with terminal, foliated panicles of inconspicuous flowers from Nov. on, the flowering branches dying after the dispersal of the fruits. With prolonged, growing innovations in March, with resting apices in August (Fig. 25 e). Buds naked. FHlalorrhagis asperrima Skottsb. Endemic on Masafuera, scattered in some of the canyons in dry, stony ground, but also found along the edge of the great western barranca in about 1200m altitude. Of the same habit as the former; with flowers and fruits in February. Seeds sown 27.4.1918 germinated after 2 weeks and the plants flowered and fruited in the autumn. Of easy cultivation. Two seedling are seen in Fig. 25 f, g. FHlalorrhagis masafuerana Skottsb. Endemic on Masafuera, scattered along the stony beach and coast-cliffs and also found on canyon walls near the en- trance. Of the same habit as the former. With flowers and fruits in February. Seeds sown 29.3.1918 germinated after 2 weeks, and some plants flowered al- ready in June and produced ripe fruit. Of easy cultivation. A seedling is seen in Fig. 25 h. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 867 Empetrum rubrum Vahl. A single sterile shrub found on the summit of Los Inocentes, Masafuera, about 1500 m above sea-level. Wahlenbergia Grahamae Hemsl. Campanulaceae. Only in the mountains of Masatierra around Bahia Cumberland, also on the Villagra side, among grasses, ferns and shrubs in open situations (Pl. 15:2, this vol. no. 7). A straggling, ir- regularly branched shrub, with densely foliated vegetative-floral shoots, the distal part of which dies after fruiting, leaving the woody stem standing. Regenerates from the moderately thick xylopodium. Innovations develop from a few (mostly I or 2) axils below the terminal inflorescence; they were quite small in March. Old branches will reach a length of 50 cm and a thickness of 7 or 8 mm. Flowers Dec.—Jan. (1. c. Pl. 20: 10). Wahlenbergia fernandeziana (A. DC.) Skottsb. With the former, but occasion- ally descending to near sea-level. A small erect shrub, otherwise as the former. As in this, not all branches belonging to the same generation flower their first summer. Flowers Jan.—March (I. c. Pl. 20:9). Wahlenbergia Larrainit (Bert. ex Colla) Skottsb. Masatierra, on dry, exposed ridges from Puerto Frances to Vaqueria, rare. Lowest and highest stations observed, 75 and 465 m above sea-level. Habit and mode of growth as in the former. With flowers Dec.—Jan. (1. c. Pl. ‘20: 9). Lrigeron luteoviridis Skottsb. Endemic in the high country of Masafuera, 1200— 1350 m, rare. Differs from /. fruticosus in lacking a distinct main stem; branches from the top of the very long, woody tap-root. Vegetative-floral shoots 20-30 cm long, woody below; innovations from the base with the first internodes long and soon leafless, then shorter to very short and the leaves forming a dense terminal bunch. Buds naked, open. With flowers in Feb.—March and fruit in March. Luphrasta formosisstma Skottsb. Only on Masafuera, not uncommon in the mountain grassland and heath from about 750 to abont 1400 m. A peculiar shrubby species, in exceptional cases as much as 70 cm tall, but generally much smaller. The morphology was described on p. 170; see also Pl. 15:1. Ochagavia elegans Phil. Bromeliaceae. Endemic on Masatierra, on the rocky ridges from near sea-level to the high crests, forming extensive patches on the steep, sun-baked cliff walls (Pl. 97), ranging west to the north precipice of Cerro Tres Puntas and the islet of Juanango. Observed once growing epiphytically on a tree in the forest. Prostrate-ascending with very long, trailing, finally lignified branches, densely clothed with curved, thick and glaucous, clasping and pungent leaves. Inflorescence terminal, sessile in the centre of the rosulate leaves, quite showy (Fig. 27 b). Flowers a deep lilac pink; Dec.—Feb. Seeds were sown 22.3. 1918 and 27.3. 1919 und seedlings obtained after 1—2 months. Of easy cultivation and a vigorous grower, also easily propagated by cuttings. Planted free in a bed in the succulent-house, it stubbornly refused to flower until a potted cutting finally produced an inflorescence in April 1944. It looked like a dwarf tree; the shoot died without ripening the seeds, but regenerated from the stem base (Fig. 27 a). 868 G. SKOTTSBERG Fig. 27. Ochagavia elegans flowering in the Goteborg Botan. Garden. — Photo T. Nathorst- Windahl, April 1944. MORE OR LESS PROSTRATE AND SUFFRUTICOSE CHAMAEPHYTES. Acaena ovalifolia Ruiz et Pav. subsp. australis Bitter. Rosaceae. In open places in the forest region of both islands, to 600 m above sea-level on Masa- tierra and in the Alpine grass-land of Masafuera, where it ascends to about 1350 m; everywhere very scarce. Matted, suffruticose. With flowers and fruit Dec.—Feb. Acaena masafuerana Bitter. Only in the Alpine region of Masafuera, 1100— 1350 m, very local. Suffruticose, forming low carpets (Pl. 114). With flowers and fruit Feb.—March. Lxigeron rupicola Phil. Endemic on Masafuera, not uncommon on the coast cliffs. A dwarf subshrub with a stout, woody tap-root; loosely branched from the base; branches slender, decumbent-ascending, to 20 cm long. Seen in bloom from end of February until the middle of March. Spergularia confertifiora Steud. All islands, on the dry sunny cliffs and slopes near the sea, ranging west on Masatierra to Bahia del Padre and Punta Larga. A small suffruticose, densely leafy perennial, erect to spreading. Probably with flowers and capsules at any time. With regard to the behaviour of this and the following under cultivation, see p. 779 of this volume. Spergularia masafuerana Skottsb. Only observed in the stony grass-land at the entrance to the Casas valley, Masafuera. As the former. Myrteola nummularia (Poir.) Berg. Myrtaceae. Only in the Alpine region of Masafuera, 1100—1400 m, rather scarce. A trailing, dorsiventral, small-leaved dwarf shrub, with the habit of Vaccinium oxycoccus. With the last flowers and with mipewbeties in Eeb—March, Pl 115-2) Rubus geotdes Sm. Masafuera, with Myrteola and Acaena masafuerana, not uncommon. A trailing, dorsiventral dwarf shrub. With fruit in February. THE VEGETATION OF ‘THE JUAN FERNANDEZ ISLANDS 869 Woopy CLIMBERS, EPIPHYTES and PARASITES. The almost complete absence of these life-forms in Juan Fernandez is note- worthy, the more so as they are more or less well represented on the Chilean mainland. Calystegia tuguriorum ®. Br. Only found in three of the canyons in Masa- fuera, not far from their entrance, trailing at the foot of the barrancas and pend- ent from the ledges. There are few trees or shrubs in the places where Caly- stegia grows and it was not seen climbing. The prostrate, slender and densely foliated branches are several m long and wind around each other, forming an entangled network. I have never seen it in flower or fruit. Robinsonia evenia is perhaps not an obligate epiphyte, even if it comes nearer to this type than any other species in the islands; see above p. 843. Not few of the other rosette trees will, as we have seen, germinate on the trunks of tree-ferns, and the same is the case with some other ligneous species, but they cannot be classified as facultative epiphytes. This is, however, true of a couple of the herbs, as will be noted later. The others may start as epiphytes, later sending roots into the soil. Phrygilanthus Berterot Phil. (Loranthaceae) is the only parasite (hemiparasitic) on trunks of trees (Vothomyrcia); it must be very rare, and to our regret we failed to find it. Luphrasia formosissima is, of course, a root-parasite (on Per- nettya), and the same was perhaps true of the extinct Santa/um. HERBACEOUS PLANTS. Phanerophytes. Gunnera peltata Phil., ‘““Pangue’’ (Panque and Panke are corruptions and should be avoided). Gunneraceae. Endemic on Masatierra, common all through the forest belt and ranging west to Cerro Tres Puntas, preferably in humid valleys, where it attains enormous size, extending from the highest ridges, where there is suffi- cient moisture, and descending to about 100 m above sea-level in Villagra, forming a fringe along the streams in the grass-land (PI. 85). The largest herb in the islands, often branching from the base (observed already on seedlings); stem horizontal and rooting, curved-erect, raised to 2 or 3 m, exceptionally as much as 5 m, 15—30cm in diameter, with a terminal rosette of leaves and with extra- axillary inflorescences (Pl 76; see also Skottsb. 3, pl. 2). With regard to the morphology and development, see Skottsb. 72. In the terminal bud the leaves are protected by the so-called ligules and by the mucus produced by cauline glands. Leaves (the small juvenile ones excepted) peltate, limb soft chartaceous with succulent ribs, to 1.9m across, petiole to 85 cm long and 3 cm thick, widening to a sheath 20 cm broad at base. Flowers from late winter to February, fruit (small, coral red drupes) from January on. Inflorescence to 1 m long. G. peltata is sensitive to drought, and a cut leaf will wither in a couple of hours. Ermel (p. 87) gave a fantastic description of the giant pangue, telling that the leaves are 25 feet tall, that the limb can bear the weight of man and that the petioles, which are soft and very juicy, are used as fence-posts! 870 C. SKOTTSBERG Gunnera Masafuerae Skottsb. Replaces the former on Masafuera, growing along the water (PI. 111) and on the wet walls of the barrancas in the canyons (Pl. 109:2; Skottsb. 3, pl. 3:2); abundant in the fog-swept highland, especially along the water-courses and in depressions in the fernbeds, forming a tapestry on the steep cliff walls of the gorges and on the great western barranca, to about 1400 m above sea-level. Habit as the former, but of lower stature; rhizome ascending, branched below, raising the terminal rosette to about 1/2 m above the ground (total length up to 1.5 m), 20—30 cm thick. Leaves larger than in the other species; blade not peltate, of the same texture as in the former, to nearly 3m wide (largest measured 2.87 m), petiole softer and more juicy, with an acidulous, refreshing taste, to 1.7 m long and 12 cm wide at base. Inflorescence pendent, to 1.75 m long and 25 cm wide. Flowering season as in the former, but no ripe drupes were found before our departure from the island in the mid- dle of March. Gunnera bracteata Steud. Endemic on Masatierra, less common than feltata and apparently confined to the elevated ridges from Rabanal to Quebrada Jua- nango, a characteristic member of the scrub in more or less open, windy situa- tions above 400 m. Smaller than fe/fata, stem rarely over 1 m long (to 1.25 m) and 10 cm thick, buds almost overflowing with mucilage. Leaves smaller, blade 350—75 cm wide, not peltate (see, however, p. 148), firm coriaceous, dark green; petiole about 50 cm long. Inflorescence about 40 cm long. Flowering season as in the former; with ripe drupes in April. Solanum fernandezianum Phil. Endemic in the forests of Masatierra, rare and jocal, extending west to Valle Ingles; lowest and highest stations observed 235 and 550 m. A stately, profusely branched species of the “uéerosum group, with bright violet flowers. The aerial shoot system lasts more than one year. Ber- tero speaks of small, bitter tubers, but neither we nor Johow were able to find them — compare Bitter |. c. 381; if they do exist, this species is the only represent- ative of the Geophytes in Juan Fernandez. Observed in flower already in late winter; fruits in April. The life-cycle needs further study. Seeds sown 29.3.1918 germinated after 3 weeks, the plants flowered already in July and bore fruit in the autumn, whereupon they died. Some cuttings had been taken, one of which survived, flowered in July 1919, fruited in October and then died. Sown again 7.3.1919; a few seedlings appeared after 2 1/2 months and gave flowering plants from July to September; they did not set fruit but lived through the winter and were propagated by cuttings in March 1920; the cuttings flowered in May, but died the same year without fruiting. In no case tubers were formed; in cultivation the species behaved as an annual. Solanum masafueranum Bitter et Skottsb. A very characteristic Morea (PI. 14, this vol. no. 7), found in two places in the forest belt of Masafuera. A large, apparently pluriennial herb as much as 1.75 m tall, with stem and branches of 1. order to 6 mm thick and more or less lignified. Collected with flowers and young fruit Feb.—March. Solanum robinsonianum Bitter. All islands, apparently scarce; an endemic microspecies belonging to the xzgrum assemblage, but this anthropophilous THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 871 group has not been well studied in Chile, and I doubt that S. vobznsonianum belonged to the flora of the islands before the arrival of man — an analogous case is Zaraxacum fernandesianum Dahlst. which, without doubt, has been in- troduced recently. Herbaceous Chamaephytes. Bromus fernandesianus (Phil.) Skottsb. Common in the wooded parts of Masa- tierra, in open glades and among brushwood from about 250 m to the high rid- ges and down to about 200m on the Villagra side, often seen in abundance on the steep slopes, deeply rooted in rock fissures and hanging down over the cliffs. Less common on Masafuera but frequent on the canyon walls and luxur- iating in wet moss carpets; ascends to about 1350 m. A slender cespitose grass Feb. Carex berteroniana Steud. Endemic; fairly common on Masatierra from Puerto with culms to 1/2 m long or more, geniculate, branching. Flowers Nov. Frances to Quebrada Juanango and in the forested country on the south side of the island, generally in open spaces in the woods, but also on rock ledges and in the mountain scrub, between 200 and 600 m above sea-level. Apparently much more local on Masafuera, on the canyon walls and in the forest, not observed above 450m. A large, densely cespitose species with culms to I m tall. With flowering spikes from September to January; ripe fruits collected in February. Seeds sown 26.7 1918 germinated in the middle of August; flowers and fruit in April 1920. Of easy cultivation. Paronychia chilensis DC. Caryophyllaceae. Barren slopes on the north side of Masatierra, rare and perhaps a recent arrival. A small, bushy perennial with decumbent branches, silvery white from the large, lustrous and thinly papyraceous stipules. With fruit in December. Ranunculus caprarum Skottsb. Endemic on Masafuera and only observed on top of the high western barranca at 1200—1300 m altitude, very rare. A stout rhizomatous, rosulate herb (p. 124, fig. 8) with scape to 75 cm high; the few individuals found bore almost ripe achenes and leafy basal innovations. All had been damaged by goats. Cardamine flaccida Cham. et Schlecht. In wet soil and along the streams in the forested region of Masatierra, but not observed above 300 m. A tender and fragile, more or less decumbent herb, rooting from the lower nodes. With flowers and pods Dec.—April. Cardamine Kruesselti Johow. Endemic on Masafuera. In some of the valleys, rare. A small, 10—15 cm tall, rosulate herb with a tuberous vertical rhizome. Apium fernandezianum Johow, “Panul’’. Endemic in the dry western part of Masatierra near the sea, local and very scarce. A characteristic species (Skottsb. 3, pl. 5:6), low, rosulate with up to 1m long stolons rooting at the nodes and giving rise to new plants. With flowers Aug.—Jan.; ripe fruit collected in January. Sown in March 1g20, flowered and fruited in September; easily propagated by cuttings. 872 C. SKOTTSBERG Centella triflora (Ruiz and Pav.) Nannf. Umbelliferae. A small rosulate herb; SEGnO 710 2: Dichondra repens Forst. Convolvulaceae. In both islands along streams not far from the sea, rare. See p. 162. A small, stoloniferous, matted herb. Oldenlandia thestifolia K.Schum. Rubiaceae. Undoubtedly of recent intro- duction and only seen once on a stream-bank in the settlement on Masatierra. A tiny cespitose herb with stems rooting from the lower nodes. Lobelia alata Labill. var. rupincola (Bert.) Wimmer. (L. anceps Thunb. p. p., listed under this name by Hemsley, Johow and myself; L. rupincola Bert. MS., Colla.) Also in Chile. On both islands, on moist rocks near the sea, on the can- yon walls where water trickles down and at the cascades. The flowering shoot dies after fruiting, but propagates with stolons from the basal rosette, so that the species cannot, as has been done, pass as an annual. With flowers Aug.—Jan. Wahlenbergia Berterot Hook. et Arn. Endemic on Masatierra and Santa Clara; dry, rocky or sandy places near the sea, rare. Vegetative-floral shoots of one year’s duration from a large lactiferous subterranean bulb. Flowers Dec.— Jan., ripe capsules end of January; innovations well developed in February. Flowers a beautiful pink (Pl. 20:8 in this vol. no. 7). Wahlenbergia Masafuerae (Phil.) Skottsb. Endemic on Masafuera and not uncommon on the steep coast rocks and canyon walls. Morphology as in the former. Flowers Dec.—Feb. Erigeron Ingae Skottsb. Endemic in the highland of Masafuera, 1350—1400 m. Rhizome long, woody, with subterranean stolons ending in a rosette; flowering shoots 30—4o cm long, erect (Pl. 16:1, this vol. no. 7). Fig. 28 a shows an inno- vation in March. With flowers and fruit Feb.—March. Evigeron turricola Skottsb. With the former, rare. A cespitose, rosulate herb with monocephalous scapes. Otherwise as the former. Gnaphalium spiciforme Schulz Bip. In the Alpine region of Masafuera, 1300— 1350 m, rare. With achenes in Feb.—March. Lagenophora Harioti Franch. Compositae. In the Alpine region of Masafuera, 830-1350 m, in company with Acaexa masafuerana, mosses etc. A minute ro- sulate, laxly cespitose herb with short runners ending in a rosette. Scape about 5 cm long, monocephalous. With flowers and fruit Feb.—March. Abrotanella crassipes Skottsb. Compositae. Endemic on the summit of Masa- fuera, 1300-1500 m. Forming round, flattened, compact cushions. With fruit in March. Oreobolus obtusangulus Gaud. Cyperaceae. With the former, and of the same habit. With fruits in March. Dysopsis hirsuta (Muell. Arg.) Skottsb. Euphorbiaceae. Endemic on Masa- tierra, common in wet, shady woods between 350 and 600 m. PI. 78:1. A frag- ile, matted, hirsute perennial, irregularly branched from the rhizome; gemmae in the soil surface, with growing apices in April (Fig. 28 b). With flowers and fruit Dec.—Feb. Mimulus glabratus H. B. K. var. eaxternus Skottsb. The variety endemic on Masafuera. In wet places under overhanging rocks and along the stream in some THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 873 of the canyons. A loosely branched, decumbent-erect perennial, forming more or less extensive patches. Seen with flowers and capsules in February. Galium masafueranum Skottsb. Moss carpets in the Alpine heath of Masa- fuera, from 1100 to about 1400 m. A tender and very fragile, trailing herb; with flowers and fruit Feb.—March (p. 174, fig. 25). Nertera granadensis (Mutis ex L. fil.) Druce (N. depressa Banks et Sol.). Rubiaceae. In wet moss mats, scarce, from 200 to 1350 m. A tender, much bran- ched, prostrate herb. With flowers and fruit feb.—March and most likely at any time from spring to winter. Salicornia peruviana Kunth. All islands, one of the very few halophytes, local but abundant where found. Suffruticose, 3—4 dm tall. With flowers Jan.—Feb. Hemicryptophytes. A typical Hemicryptophyte in Raunkiaer’s sense has monocarpic vegetative- floral shoots and survives the unfavourable season in the form of one or several buds situated just beneath the soil surface or at the surface and then protected by dead leaves. Except in the highland of Masafuera there is no unfavourable season in Juan Fernandez during which either cold or dry weather would cause the plants to become dormant. The difference in temperature between summer and winter is slight, frost is unknown, and even if most of the rain falls during the winter months, the summer is by no means dry. Thus the herbs and grasses stand, as a rule, fresh and green the year round, even if the flowering shoots are monocarpic. The innovations do not pass any season as closed, well-protected buds; they will, perhaps, arrest their growth for some time in a more or less advanced stage. Few species are typical hemicryptophytes; most of them approach the chamaephytes in their mode of living. Stipa neesiana Trin. et Rupr. (S. fernandeziana Phil.). Both islands, common in the natural grass-lands as far west as Punta Larga and also found in open spaces in the wooded section of Masatierra; more abundant on the lower, treeless slopes of Masafuera, where it is an important constituent of the grass-land. A coarse, cespitose grass with the dead leaves long persistent; culms to 1 m tall. Observed in flower July—Dec., in fruit Dec.—Feb. Stipa (Nassella) /aevissima (Phil.) Speg. (Piptochaetium, Phil.). Masatierra, on the lower treeless slopes from Bahia Cumberland to Punta Larga and Que- brada Juanango, abundant; also here and there on the stony ridges to about 600 m above sea-level. A leading grass in the basal region of Masafuera; see p. 98 for particulars. A typical “‘tunicate’’ grass, with filiform leaves as long as the culms, to 50cm or more; new growth more or less concealed among the mass of old, yellow sheaths. With flowers July—Nov. and fruits Dec.—Jan. Piptochaetium bicolor (Vahl) Desv. Masatierra, on open soil in the central part of the island, not common, but rather abundant on the dry lower slopes of Villagra. A tunicate grass with narrow convolute leaves 10—20 cm long and culms up to 1 m. Flowering and fruiting season as in the former. Polypogon imberbis (Phil.) Johow. Endemic; on both islands near the coast, 58 — 516796. The Nat. Hist. of Juan Fernandez and Easter Isl. Vol. II. 874 C. SKOTTSBERG f \ Fig. 28. a Erigeron Ingae, end of stolon in February. Nat size. — 6 Dysopsis hirsuta, part of base with innovation in the beginning of April. 2x. — c, d Peperomia berteroana in the end of March, part of rhizome with innovation. The line indicates the soil surface. Nat. size. — e Peperomia margaritifera in the beginning of April, base of stem with innovations. Nat. size. — f the same, innovation. 2x.— g Libertia formosa, base of plant with outermost leaves removed, showing two innovations. I-4 leaves, 7 prophyll, a and 6 buds. Nat. size. — #4 the same: a seen from the other side. 2x. — 7 diagrammatic cross section of base of Lzbertia. 1, 2 leaves, pr prophyll. but rare on Masatierra; not uncommon on the rocky beaches of Masafuera from Ensenada Toltén to the south coast. With flowers Dec.—March and fruit Feb.— March. Polypogon chilensis (Kunth) Pilger. Masatierra, on barren slopes from Valle Colonial to Quebrada Juanango, not uncommon on the Villagra side. A larger THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 875 species than the former; culms as much as 2 m tall according to Johow. With flowers Jan.—April. Agrostis masafuerana Pilger. Endemic in the Alpine region of Masafuera, 1300-1350 m, rare. Moderately cespitose; found with flowers in February. Danthonia collina Phil. Masatierra, on dry, stony ridges with thin soil; both sides of the main range in the central part of the island, 200—400 m, scarce though locally abundant. A tunicate grass forming hard, compact clumps; leaves rigid, 10—20 cm long, persistent, culms 50 cm tall or more. With flowers Dec.— Jan. and fruit in January. Koeleria micrathera (Desv.) Griseb. Masatierra, very rare and only seen by us in a single locality near Portezuelo de Villagra. Densely tufted, leaves about 30 cm long, culms about 50 cm. With flowers in January. Bromus masafueranus Pilger et Skottsb. Endemic on Masafuera, seen in a single place among rocks about 1350 m above sea-level. A very conspicuous, densely tufted grass, forming large tussocks 1/, m high with culms to 75 cm tall. With flowers and fruit in February. Cyperus eragrostis Lam. (C. vegetus Willd.) Both islands; in moist places, mostly near the sea. A robust species with flowers and fruit Jan.—Feb. Scirpus nodosus Rottb. Both islands; scattered on Masatierra from the coast to the elevated ridges and rather common on the south slopes, ranging west to Bahia Chupones. Frequent along the shore on Masafuera. A rigid rush, to 60 cm tall. With flowers from late winter to February. Scirpus cernuus Vahl. Both islands, forming dense, low-growing mats in marshy places along the shore and on stream banks in some of the valleys. With flowers and fruits Dec.—Feb. Heleocharis maculosa (Vahl) Kunth. subsp. fuscopurpurea (Steud.) Kuekenth. Known from two places on Masatierra, growing along running water. Small, mat- forming. With flowers and fruits Dec.—Jan. Cladium scirpotdeum (Steud.) Benth. et Hook. fil. Endemic on Masatierra, ranging through the forest country from Pangal to Quebrada Juanango and from about 200 m or occasionally less to the summit of El Yunque, a dominant feat- ure of the cascades. A large Viucentia with very long and narrow, ribbon-like leaves pendent from the rocks; the leaves will reach a length of 2 m in such places (Pl. 78). With flowers Nov.—April. Uncinta Douglasii Boott. Endemic, common on both islands in the forests and on the barrancas of the gorges. A slender, soft and graceful sedge with leaves and culms ?/,—1 m long. With flowers and fruits Dec.—Feb. A closely related species is U. costata Kuekenth., known from a single locality on Masafuera. Three non-endemic species belong to the Alpine region of Masafuera: U. senuzs Poepp., a small plant not over 20 cm high and scattered from 900 to I500 m, U. brevicaulis Thouars, tufted, 20—30 cm tall, 1100—1400 m, scarce, and U. phle- oides Pers., in a single locality at 1130 m, a robust, densely tufted species (PI. 115:1). All with fruits in Feb.—March. Carex Banksti Boott. Masafuera, known from a single locality, about 1350 876 C. SKOTTSBERG m. A 30—40 cm tall sedge, quite conspicuous thanks to the long, dark brown, pendulous 2 spikes. In fruit in February. Lusula masafuerana Skottsb. Endemic in the highland of Masafuera, 1100— 1400 m, rare. Robust and densely tufted, with culms to 50 cm long. With ripe seeds in February. Funcus procerus ¥. Mey. (see p. 775). Masatierra, on barren slopes near the sea and on the elevated ridges (Pl. go: 2), west to Bahia Chupones and Vaqueria; rare. Largest of the rushes in Juan Fernandez, to 1 m tall. With flowers in January and fruits in April. Funcus tmbricatus Laharpe. Masatierra, grassy slopes near the sea, scarce. With flowers and fruits Dec.—Jan. Funcus dombeyanus C. Gay. Scattered in the grassland of Masatierra (J. microcephalus Kunth according to Johow; see p. 112), observed by us in a single locality. With flowers and fruits in January. The remaining 2 species, ¥. capzi- laceus Lam. and planifolius RK. Br., were found by us for the first time and only near the settlement in Masatierra; they are, I presume, recent arrivals. Libertia formosa Graham. Iridaceae. Both islands, in open situations on erass-covered, stony slopes and ridges from about 200 to 600 m on Masatierra and from the coast to about 1400 m on Masafuera. One of the most conspicuous herbs, forming large, bright green tufts. Leaves narrow, sword-shaped, equitant, flowers white. Buds from the lower axils, tightly embraced by the persistent leaf bases (Fig. 28 g—i). With flowers Nov.—Jan. and fruits in February. Seeds sown 25.7.1918 germinated in September and the plants flowered end of April 1920. Long known in cultivation, originating from Chile. Peperomia berteroana Miq. Piperaceae. Endemic. Not uncommon in the wet, shady montane forests of Masatierra above 400 m; in Masafuera on the moist walls in the deep canyons, apparently rare and not seen with flowers. A succulent species with a thick rhizome and as much as 50 or 60 cm tall, vegetative-floral shoots; robust stems to 12 mm thick. Leaves short-lived, accumulated toward the end of the shoot (PI. 79: 2). Gemmae in the soil surface with apex covered by redu- ced, scale-like leaves (Fig. 28 c, d), resembling the winter-buds of a certain type of hemicryptophyte, but the flowering shoots continue to grow and to produce new leaves and spikes at any season — a kind of combination of a nanophanero- phyte and a hemicryptophyte. For particulars, see Skottsb. 2z, 22, with illu- strations. Peperomia margaritifera Bert. ex Hook. Endemic on Masatierra, in wet soil in two of the well-watered valleys (p. 114), rare. Habit of the former, but much less succulent. Rhizome short, incrassate, shoots to 40 cm long, with num- erous small gemmae (as many as 9 observed) in a circle immediately under the soil surface, beginning with a few short internodes with minute, scalelike leaves; a prolonged internode follows, ending in a naked apical bud (Fig. 28 e, f). In April the vegetative-floral shoots continued to form new spikes and leaves, and very likely growth goes on during the winter as in the former; see also Skottsb. 22. Peperomia Skottsbergii C. DC. Replaces P. margaritifera on Masafuera, where THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 877 Quebrada de las Casas is the only locality known so far (Pl. 10, this vol. no. mena). Peperomia fernandeztana Mig. On both islands, but also on the mainland of Chile; not uncommon in the humid forests of Masatierra, ranging west to Valle Ingles, on Masafuera also under overhanging rocks in the canyons (PI. 113). Terrestrial and epiphytic. A moderately succulent species with a long rhizome and to 65 cm tall, branched vegetative-floral shoots. Seasonal development as in the other species. For particulars, see Skottsb. 22. Callitriche marginata Torr. var. Lechleri Hegelm. Found once only, in a stream in the settlement in Bahia Cumberland. Probably a recent addition to the flora. Therophytes. Only 5 species of Therophytes have been recorded from Juan Fernandez, a surprisingly small number in view of the host of spring annuals known from the mainland: U7rtica Masafuerae Phil., Parzetaria debilis Forst. var. fernandeziana (Steud.) Skottsb.,* Zetragonia expansa Murr., Cardamine chenopodiifolia Pers., and Plantago truncata Cham. subsp. Skottsbergit Pilger. Cardamine and Plantago are confined to Masatierra (Plaxtago also on Santa Clara), Urtica endemic on Masa- fuera, the remaining two found on both islands, Ze¢ragonza also on Santa Clara. Cardamine has been observed twice, in 1854 and 1872, Uriéica only in 1854. THE PTERIDOPHYTES. Ferns play a leading réle in the forest and scrub, in the undergrowth and as epiphytes; they form veritable forests in both islands and put their stamp on the Subalpine heath of Masafuera. Of Fern Allies 2 species of Lycopodium occur in the high country in Masafuera, but no Eguise‘um has been recorded from Juan Fernandez. Ophioglossum fernandezianum C. Chr. is confined to Masatierra. Tree-ferns. Dicksonia berteroana (Colla) Hook. Endemic on Masatierra, scattered or forming small groves in the wooded valleys as far west as Valle Ingles and de- scending to about 200 m above sea-level in well-watered gullies; much more frequent above 400 m, ascending to the summit of El Yunque and locally abundant. For particulars, see p. 17. Trunk not seldom to 4 or 5 m tall; largest specimen measured 6 m high with a trunk 25 cm in diameter. Pls. 87, 93: 1, 96: 1. In valleys easily accessible from the settlement in Bahia Cumberland hundreds of specimens have been dug up by the inhabitants and sold to visitors, who try to plant them in their gardens on the mainland, probably without success. Dicksonia externa Skottsb. Endemic on Masafuera, where it replaces the former. Scattered in the forest from about 450 m, forming extensive closed stands on the fog-swept slopes around Los Inocentes and thin groves in the grass-land * P. debilis typicae arcte affinis, sed differt statura minore et debiliore nec non fructibus minoribus, I-1.15 mm longis et 0.7-0.75 mm latis; vide Skottsb. in Géteb. Vet. & Vitt. Samh. Handl. 5:6. 13-14, fig. 5. The same form was described by C. Gay as Fy2irea humifusa (Chile). 878 C. SKOTTSBERG north of Inocentes (Pls. 105, 108: 2; comp. p. 18). More robust than the former, with the rather clumsy trunk as a rule 2—3 m long and often with one or a few branches. The specimen figured by Johow (pl. XVIII), measured 4.5 m (p. 287), but much taller ones are occasionally seen; thus the largest measured by us, illus- trated on Pl. 2 (this vol. no. 7), was 6.5 m tall and undoubtedly very old, as the trunk measured about 1.5 m across the enlarged base and 67 cm below the first living limb; there were 6 branches, 2 of them dead. An average-sized frond has a pet- iole about 50 cm long and 4.5 cm broad at base and a blade of about 1,3 X1 m. Blechnum cycadifolium (Colla) Sturm. Endemic and common on both islands, particularly abundant on the windswept ridges, where it is one of the leading plants in the scrub; extends from near sea-level to the most elevated crests in Masatierra, ranging west to Bahia Chupones. Less frequent on Masafuera but ascending above the timber-line to near the summit; see also p. 28. Johow 2 p. 244 says “apenas arborescente’, but in my opinion it grows to become a true tree-fern; the trunk, often low and curved, will be well over 1 m tall in the forest and 15—20 cm thick. The rigid, dark green and lustrous fronds give to this plant a striking Cycas-like appearance (Pl. 4, this vol. no. 7; pls. 58:2, 84). Thyrsopterts elegans Kze. Endemic. Masatierra, not uncommon in the mont- ane woods from Rabanal to Valle Ingles; descends on the Villagra side to about 175—250 m above sea-level. Masafuera, apparently confined to the highland, where it reaches 1100 m; see also p. 16. Ascends from a clumsy, prostrate rhi- zome, passing into an erect trunk raising the canopy of magnificent fronds well over man’s height (Pl. 81). A frond lasts 2 years. The trunk may reach I m, but is as a rule shorter; thickness about 20 cm. Very large fronds (Masatierra) with stipe 1.7—2.5 m and, blade 1.3—1.75 m, as wide or a little narrower. A small fertile frond is figured in Skottsb. 3, pl. 7:4. The sporangia-bearing seg- ment developes during the summer and reaches maturity in March. Lophosoria quadripinnita (J. F. Gmel.) C. Chr. Cyatheaceae. A form of this widespread species is one of the commonest ferns in the islands, ranging from the ravines in the deforested valleys to the highest wooded ridges in Masatierra and a dominant on the highland of Masafuera. Johow states p. 154 (Alsophila pruinata Kze) that the trunk attains a length of 2 m and a thickness of 15-20 cm. We have never seen specimens of this size, and 1 am convinced that they must have been exceptional. Even if a trunk is developed, it is as a rule much shorter, considerably less than 1 m. The fronds are very large, to 3 or sometimes 4 m long. The stunted form growing in the Subalpine-Alpine region on Masa- fuera is seen on Pl. 115: 1. Dryopterts naequalifolia (Colla) C. Chr. Endemic and common in the forests of both islands, and not rare in many places on the deforested slopes. The erect trunk generally remains quite short but will reach a height of 1/, or perhaps 1 m under favourable conditions. The fronds attain a length of 2 m (Pl. 3, this volsno..7). THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 879 Scandent ferns. Arthropteris altescandens (Colla) J. Sm. Endemic; both islands, but rare on Masafuera; not uncomon in the lower forest belt on Masatierra from Valle Fran- ces to Quebrada Juanango, but not observed at a greater altitude than 400 m. Trailing over stones and fallen trunks and winding in the trees (Pl. 95: 1, see also Pl. 9, Skottsb. 2). Blechnum Schottit (Colla) C. Chr. Endemic and not uncommon in the forests of Masatierra, also in the upper montane region. Seems to be rare on Masafuera. Fronds of much firmer texture than in the former, dark green and longer, to 50 cm, exceptionally large ones 90 cm long. Skottsb. 3, Pl. 7: 3; pl. 80. Epiphytic ferns. Numerous ferns, most of them belonging to the Hymenophyllaceae, which will be treated separately, clothe the moss-covered trunks of the trees but are just as often found on the forest floor. Among the Polypodiaceae, Asplenium dareotdes Desv. (magellanicum Kaulf.) and Blechnum valdiviense C. Chr. are epiphytic on trees and tree-ferns, but also terrestrial. The small Polypodium Billardieri (Willd.) C. Chr., a very rare species in these islands, was seen growing in moss on Drimys on Masatierra, and on a moss-covered rock on Masafuera; the common P. /az- ceolatum L., often met with in both islands though more plentiful on Masatierra, is rarely terrestrial; on Pl. 82 it is seen covering the branches of a fallen luma, its slender, wiry rhizomes interlaced with the thick rhizomes of P. zxtermedium. The fronds are coriaceous in the former, which is a typical xerophyte; in cases they are as much as 25—35 cm long, but generally much smaller. Polypodium intermedium Colla (P. translucens Kze). With regard to the polymorphism of this endemic species, see pp. 37—41, fig. 6, and p. 765. On Masatierra it generally grows on Dyimys, on Masafuera (subsp. masafueranum C. Chr. et Skottsb.) on Myrceugenia. A couple of small plants were once found on each of the islands, growing on moss-covered rocks near the sea (p. 38, fig. 6 d and p. 765); their fronds were stiff and coriaceous, whereas the epiphytes have herbaceous fronds with the net of veins clearly visible when the frond is held against the light. Plants of the typical species are seen on Pl. 82. Thanks to the succulent, water-storing rhizome (Fig. 29) they are able to withstand long periods of drought. The colour of the rhizome is a light grayish green and the numerous scales are reddish brown. The rhizome is flattened dorsiventrally, in section e. g. 15 X10 and 20X5 mm. Johow has published a very good illustration of this fern (2, Pl. XVIII, P. translucens). In order to test the drought resistance a simple experiment was arranged 25.3. 17. The apical portion of five plants (I—V) of the typical form were detached from the substratum, carefully cleansed from mosses and lichens and the cut end sealed with grease. They were hung on the veranda protected from the rain, but exposed to full light. They were weighed every 48 hours until 18.4, when the intervals were prolonged, and the experiment was finished 29.4 because we had to prepare to leave the island. The results are accounted for in Table XXV. 880 C. SKOTTSBERG 9) N 9. End of rhizome of Polyfodium intermedium typicum, the orbicular frond scars surrounded by a dense whorl of narrow paleae. Nat. size. — Photo Mrs. M. Ferm. I. Rhizome 24.5 cm, 1 frond, and a short branch with another; the branch was constricted at the base which explains why its frond withered first of all. II. Rhizome 8.5 cm, with 2 fronds. 1OUE » ToGo ese eae LONG. IVs » 17 2} . e3itronds: Whe » 19 es » 2 ) Thanks to its water storage a detached rhizome is able to keep the fronds turgescent for many days, in cases for a month or more. When the experiment had been finished, the dry fronds were removed from rhizomes H, IV and V and the rhizome and fronds weighed separately. The result was: II. Rhizome 1.7, fronds 1.8 g IV. 2 95, *% 458 V. » 3.O. 72 yon A fresh frond weighs about 3 times as much as a dry one; in V, the 3 fronds should have weighed 5.1 g, the rhizome 20.7 g, while, 29/4 it weighed 8.9 g; in IV, where the fronds were much larger, their fresh weight would have been approximately 13.5 g, the rhizome 28.6 g; 20/4 it weighed only 9.5 g. In II, the fronds would have weighed 5.4 g fresh, the rhizome 9.6, reduced to 1.7 29/4. The loss of water of the rhizome was 11.8 g in V, 19.1 g in IV, and 7.9 in Il, or 57, 67 and 82%, respectively. | Ferns of the field strata. The abundance of ferns is a well-known feature in the physiognomy of the plant communities and this applies not only to the wooded country but also to the areas where the forest has disappeared. Of the arborescent species, Blechnum cycadifolium, Lophosorta and Dryopteris may form large beds, and several herb- aceous species are still able to compete with the aggressive army of weeds now filling the outer section of the valley bottoms and the lower, sun-baked slopes. The most abundant of all is Blechnum auriculatum Cav., luxuriating in the closed THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 881 Table XXV. Polypodium intermedium. Weight in grams. 25.3- | 27-3. | 29.3. | 31.3. | 2-4. | 4-4- 6.4. 8.4. | 10.4.] 12.4. | 14.4.| 18.4.| 22.4. | 29.4. I | 36.4 | 29.1| 26.7| 24.8) 22.7 | 21.4 | 20.3 | 19.2 spe Ps Yen Fes Gey ea all tay fem ce of 7) en/na nee CY Mis) L560} L-5,| Lost GiOi a 7Oul 6:62 | Ser 5.3 4.8 4.6 4.5 | 4.0 a7, 3.5 MI) £2.9 | 9:9) 8.8) 8.4 78 | 72 | 6.8 6.4 6.0 5-7 5-4} 5.1% 4.6 An3° MtAe.© | 33.2) 30.3 | 28:1 | 25.6'|23.6 | 2x.8" | 20.2 | 18.7%) 17.7 | 17:2 | 16.19%). 14.8 | 4.0 Namie Sl 627.4'| 619-5 | 18°3)| 17.01)'15-.9 [15-217 |'54.557| 13-8) | 13:4 | 13:0) || 12.278) rr 2 106 1 Frond on branchlet dead since several days, the one on the main rhizome still green and turgescent. 2 Of all eleven fronds only the one on rhizome I still quite fresh. 3 Fronds beginning to droop. 4 Both fronds yellow and dead. 5 The single frond fading. 6 Frond dead. 7 Oldest frond withered. 8 Also next oldest frond drooping. ® Youngest frond beginning to turn yellow. 10 All fronds dead. 11 Fronds beginning to fade. 12 Oldest frond dead. 13 All fronds dead. forest and widely spread in smaller and more stunted forms in the barren country, where it ranges west to Punta Larga and Santa Clara I., and it is just as common on Masafuera. The shade-form is illustrated on Pl. 83. Equally widespread, from near sea level to the high mountains and extending west as far as Bahia Chu- pones, is Polystichum berterianum (Colla) C. Chr. with a wide-creeping rhizome and fronds 1/; to 1 m high; in Masafuera it reaches the Alpine region, 1300-1400 m. Hrstiopteris incisa (Thunb.) J. Sm. has the same large distribution as the for- mer and is a characteristic fern of recently burnt areas. It is a stately fern, large specimens attaining the height of man. Hypolepis rugosula (Labill.) J. Sm. (Skottsb. 17. 369, fig. 4), also common, but less widespread, belongs to the lower region in Masatierra and was not seen above the timber-line on Masafuera. Another very common fern is Adzantum chilense Kaulf., found in the forest as well as on the banks of streams and roads in company with Hypolepis, Polystichum, Blech- num auriculatum etc. It is one of the few species inhabiting the coast cliffs in the three islands. The special seaside fern is Asplenium obliquum Forst. (Pl. 77: 2) in its endemic, but not too well-marked variety chondrophyllum (Bert. MS). Mett., a succulent halophyte found in the three islands. The few stations where the small, densely tomentose, xerophilous Pe//aea chilensis Fée (endemic, both islands) has been found are, with one exception, dry rocks of the coast. Another drought- resistant fern is Gletchenta pedalis (Kaulf.) Spr., only found on Masatierra on the mountain-spurs surrounding Bahia Cumberland. As this conspicuous species is wanting in collections made before 1854, it is suspected to be a late arrival from south Chile. Of the ferns inhabiting the humid forest and well-watered gulches three large species are met with all through the wooded country in both islands, lechnum chilense (Kaulf.) Mett, Polystichum vestitum (Forst.) Pres] and Pterts berteroana Ag., the last-mentioned endemic. Blechnum chilense is a social fern with fronds 50—70 cm long; it seems to prefer open glades and banks of streams, ranging west to Cerro Chumacera and Bahia Chupones in Masatierra. In Masafuera it also inhabits the Alpine region (Pls. 93, 112). Polystzchum vestitum is not uncom- 882 G. SKOTTSBERG mon in the closed forest and on the banks of well-watered ravines from Valle Frances to Villagra and ascends to the lofty ridges on Masafuera; in favourable situations the fronds will reach a length of 1 m. Pterzs berteroana, one of the finest ferns in the islands, with fronds to 1 m tall and more, thrives in the deep shade and is found on many places on both islands (PI. 113). A much rarer, but just as large fern is Prerzs semztadnata Phil., restricted to the fog region of central Masatierra above 500 m and reported from the forest patch on the summit of El Yunque; on Masafuera it also occurs in the Alpine region, where it is stunted. Pteris chilensis Desv. is of smaller stature than the other two, scattered through the wooded country of Masatierra but rare on Masafuera. Among these large and conspicuous ferns, the endemic Alechnum longicauda C. Chr. stands first. It is restricted to the canyons in Masafuera, where it adorns the wet cliff walls; see p. 28—29 and Pl: 5, this ‘vol. no. 1. Among the small species of the forest floor, Asplentum dareotdes Desv. and macrosorum Bert. ex. Colla (1 c., Pl. 1: 2) have been mentioned already. They occur on both islands, where the latter is endemic. The same holds good of the elegant A. stellatum Colla with frond tips proliferous. Alechnum valdiviense C. Chr., also found in the Alpine region of Masafuera, is easily known by its reproduction by subterranean runners and by its very dark colour. As we have seen above, a considerable number of ferns inhabits the high- lands of Masafuera. Like some of the Hymenophylla, the rare Glezchenia quadrt- partita (Poir.) Moore is restricted to high altitudes on Masafuera and was not seen below 1150 m. The two club-mosses, Lycopodium scartosum Forst. (gayanum Remy) and magellanicum Sw. have the same distribution as Gletchenia quadrt- partita. Filmy ferns. In Juan Fernandez, the Hymenophyllaceae are represented by 15 species, //y- menoglossum cruentum (Cav.) Presl, 10 species of Hymenophyllum, 3 of Tricho- manes, and Serpyllopsis caespitosa (Gaud.) C. Chr., all of them being more or less strictly confined to moist, shady places in the montane region. //ymenoglossum (Pl. 1:1, this vol. no. 1) is scarce on Masatierra and rare on Masafuera. Of Ay- menophyllum, cuneatum Kze (p. 9, fig. 3) and plicatum Kaulf. (dichotomum auctt. non Cay.) are common and physiognomically important; less frequent, but not very rare on Masatierra, is 7. ferrugineum Colla. The endemic //. rugosum C. Chr. et Skottsb. (p. 12, fig. 4), known by its bronze-green colour, is scattered over both islands, AY. tortwosum Hook. et Grev., once reported from Masatierra, is rare on Masafuera, H. falklandicum Bak., pectinatum Cav. and secundum Hook. et Grev. only found on Masafuera, H. secundum in a single locality, about 1350 m above sea-level (fectixatum recorded once for Masatierra, but never found again), The largest species in Juan Fernandez, and probably one of the largest in the genus, is 7. fuciforme Sw., restricted to the montane forests in the eastern half of Masatierra (Fig. 30); longest frond measured, 82 cm. Similar to this, but THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 553 5 PP : Vege ae 4 Fig. 30. Aymenophyllum fuciforme, large plant in the forest of Villagra, Masatierra, c. 550 m above sea-level. — Photo 10/1 1917. not as large, is HW. caudiculatum Mart. which is much rarer on Masatierra than on Masafuera. Trichomanes exsectum Kze is the only species of this genus inhabiting both islands; a fine carpet is seen on Pl. 83. The small 7. /ugae C. Chr. (p. 4, fig. 2, endemic) and 7. phzlippianum Sturm (p. 3, fig. 1) have only been recorded from Masatierra. Serpyllopsis caespitosa, a rather variable plant, is represented in the islands by var. fernandeziana C. Chr. et Skottsb. Many filmy ferns are terrestrial and epiphytic; such is the case with Aymeno- glossum, Hymenophyllum cuneatum, falklandicum, ferrugineum, pectinatum, plicatum and rugosum, Trichomanes exsectum and Serpyllopsis. The latter is nearly always epiphytic, on Drimys and Dicksonia, and the trunks of tree-ferns seem to be pre- ferred by most species. Zrichomanes philippianum, endemic on Masatierra, has not been found growing on the ground but exclusively on the trunks of Dzck- sonia berteroana and Blechnum cycadifolium. In his “‘Biologie der Pflanzen’, p. 220, NEGER wrote: “In extremer Ausbildung finden wir ‘hygrophile Struktur’ bei den stets von dampfgesattigter Luft umge- benen Hautfarnen (Hymenophyllum), welche in tiefem Waldschatten der ewig feuchten tropischen und subtropischen Regenwalder gedeihen. Nattrlich sind diese Gewichse im hdchsten Grad emfindlich gegen jede auch nur voruberge- hende Austrocknung.”” With these words the author wanted to emphasize that the filmy ferns are unable to stand desiccation, even if of short duration. On 884 C. SKOTTSBERG our many visits to Portezuelo de Villagra we noted that after a spell of dry, sunny weather the carpets of Hymenophyllum along the trail had shrivelled, with the pinnae and rhachis convolute, while after a drizzle and in misty weather they had regained turgescence the next day, and it became evident that they did not de- pend, or at least not only depend, on water taken up from the substratum but were able to absorb moisture with their entire frond. In order to state the rate at which turgescence is restored under different conditions, I made a few simple experiments. Hymenophyllum fuciforme. A specimen with one large frond was taken into the laboratory 25/3 at 1 p.m., without removing the earth and moss. At g p. m. the frond was dry and shrivelled. It was left like that to 4/4; when at 4 p.m., now very dry, the earth and moss was well moistened and the plant placed under a bell-glass on moistened filter-paper. At 9 p.m. 5/4 little effect was noticed, but *after 12 hours turgescence was getting restored. The frond was then detached; it weighed 4.52 g. Left to dry to 6/4 at noon, when it weighed 4 g, it was sprayed, and placed under a bell-glass at 4 p.m.; after 2 hours it had straightened the rhachis and unfolded the pinnae, the weight then being 5 g. Flymenophyllum cuneatum. a) A tuft with moss and earth was placed in the laboratory 25/3 at 1 p.m. and left to dry. At 9 p.m. the fronds began to wither. When completely dry, the plant was dipped in water 4/4 at 4.45 p.m., taking care that the substratum was not mois- tened. Already at 5.30 the frond segments began to expand, the rachis still being rolled up, but half an hour later the plant looked perfectly normal. b) Six detached fronds, well dried so that they could be powdered by rubbing them between your fingers, were weighed and put on a piece of wetted filter-paper under a bell-glass 4/4 at 4.10 p.m., and the increase in weight noted: 4/4 4.10 p.m. 6.38 @; 5.30 p.m. 0.48 g; 8.30 p.m. 0.5 g; &/4 7.15 a. m..0.63 2; 8: 3ompemonO g; 6/4 8.30 a.m. 0.7 g; 8.30 p.m. 0.7 g. Turgescence was then completely restored. flymenophyllum plicatum. A tuft with earth and moss was left to dry in the labo- ratory 27/3 at 2 pm. At 8 p.m. all the fronds were rolled up and the plant left like this until 4/4 at 4.30 p.m., when it was sprayed with rain-water. Already after 5 minutes some effect was noticeable, and at 5.30 p.m. the fronds were once more turgescent. Left once more to dry until 6/4, when at 9.35 a.m. it was again sprayed and put under a bell-glass. Restored to turgescence at 10.30 a.m. Trichomanes exsectum. A detached frond weighing 0.5 g was left to dry in the laboratory 7/4 at 2 p.m. and 13/4 at 3 p.m. put under a bell-glass with wetted filter-paper, but without being in contact with this. Weight 13/4 at 3 p.m. 0.28 g, 13/4 9 p.m. 0.36 g, 14/4 9 a.m. 0.52 g and at g p.m. 0.52 g; turgescence then completely restored. It should be mentioned that, under the microscope, the cells had a normal appearance; this also applies to the 3 species of A/ymenophyllum. These simple experiments will show, I believe, that filmy ferns behave like bryophytes and lichens and are able to stand desiccation for many days — very likely much longer — without sustaining damage, and that direct absorption by the frond from a film of water or from the humid atmosphere is a more import- ant source of moisture than water taken up from the substratum. Water absorption by the leaves in higher plants has been investigated by several authors. A synopsis was published by K. Werze in 1924 (Flora N. F. 17); his experi- ments with plants from Temperate regions led him to the conclusion that such absorption certainly occurs, but that it lacks biological significance. Recently (Oikos, Suppl. I, THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 885 1951) INGA ARVIDSON showed that in plants of Swedish “‘alvar’’ soil leaves are able to absorb sufficient water from the air (rain, dew and “‘fog-water’’) to counterbalance the loss caused by transpiration and that, during a drought, this is quite important. With regard to ferns, V. B. Wittrock studied the behaviour of a number of species cultivated in the open (Acta Horti Bergiani I, 1891) if allowed to dry and then put in a moist chamber or dipped in water. In some, e. g. Polypodium vulgare and species of Asplenium, turgescence was restored, while others, e. g. Adiantum capillus Venerts, were killed if allowed to become quite dry, being unable to replace the loss of water through frond absorption. B. Wantin (Sy. Mosskulturféren:s kvartalsskr., 5, 1943) showed that Polypodium vulgare and Phyllitis scolopendrium are unable to cover their need of water by root absorption alone and that frond absorption is necessary. In Adiantum, the cuticle is waxy and no film of water formed. I regret having neglected to study A. chilense. In the filmy ferns, which appear not to have been studied before, absorption of atmospheric water by the entire, monostromatic frond seems to be quite normal and of paramount importance. CHAPTER IV. The Vegetation. INTRODUCTION. My analysis of the plant communities of Juan Fernandez does not meet the de- mands of modern plant sociology. The use of meter quadrates was not general at the time of our survey, and their serviceableness, quite satisfactory in certain types of plant cover, e.g. in grass-land and heath, is small or none in a dense primeval forest rich in species of ligneous plants and offering all kinds of obstacles. The slopes are mostly steep and very hard-worked, and to get the rare trees and shrubs (which are just as characteristic of a certain type of forest as are the dominants) included would have necessitated a large number of analyses. This our limited time did not permit. Where it was convenient, a square 10 X IO m was selected, a list made of all the species of vascular plants, their rela- tive importance estimated, the stratification observed, and samples taken of the cryptogams, terrestrial and epiphytic. It happened that this procedure was too time-absorbing and that I had to restrict the note-taking to a simple nameclist, which, however, will serve to give the reader a general idea of the composition of the vegetation. Rock face communities formed by cryptogams, particularly lichens, are an important feature in Juan Fernandez, where exposed cliffs are met with everywhere; I am no lichenologist, and even if I tried to secure samples of as many species as possible, many must have escaped my attention. The same applies to the synusiae of the ground stratum in forest, scrub and heath, and to the communities of epiphytes. I want the reader to understand that I am quite aware of these deficiencies. The nomenclature follows, with regard to the Phanerogams, my paper on the flora (no. 7 in this volume, 1922, and the Supplement, no. 28) and, for the Pteridophytes, CHRISTENSEN and SKOTTSBERG and (no. 1, 1920, no. 28); for the Mosses V. F. BROTHERUS (no. 8, 1924), for the thallose Hepaticae A. W. EVANS 886 C. SKOTTSBERG (no. 20, 1930), for the foliose Hepaticae Th. HERZOG (no. 24, 1942), and for the Lichens A. ZAHLBRUCKNER (no. II, 1924, with a few additions in no. 13, 1926). Later, lichenologists engaged in critical studies have, in a number of instances, diverged from Zahlbruckner’s opinion; the genus Pseudocyphellaria, not recognized by him, has been taken up again, and a number of specimens of this and other genera have been revised: see G. Degelius in Acta Phytogeogr. Suec. 7 (1935), p. 52, I. Mackenzie Lamb in Lilloa 13 (1947), A. H. Magnusson in Meddel. Goteb. Bot. Tradg. 14 (1940), pp. 18, 34 and in K. Sv. Vetenskaps.-akad. Handl. Ser. 3, vol. 7:4 (1929), p. 86, G. A. Malme in Arkiv for botanik 26 A: 13 (1935), p. 5, J. Motyka, Lich. gener. Usnea Studium Monogr., Lwow 1936—38, and R. Santesson in Arkiv for Botanik 30 A: 10 (1942), p. 16 and ibid. 31 A: 7 (1944), pp. 14,15. I owe this information to Dr. SANTESSON. Prof. G. E. DU RIETZ kindly supplied some corrections. Stratification. Four layers were distinguished: IV, Tree layer, above 6 m, subdivided in an upper, IV: 2, and a lower, IV: 1, with the limit between them at 12 m; III, Shrub layer, 6—1.8 m; II, Field layer, composed of 3 strata, upper (II: 3), 1.8—o.6 m, middle (II: 2), 0.6—o.2 m, and lower (II: 1), under 0.2 m; Bottom or Ground layer (I), mostly formed by Cryptogams. Degree of covering. The relative role played by each species in a de- fined stratum is indicated with the Hult-Sernander scale from I to 5; with regard to the forest communities, | now would have preferred to simplify this scale; the figures given are, of course, not a result of measuring, but of estimation. The Plant Communities of Masatierra. REMARKS ON THE TOPOGRAPHY. A geographical sketch will be found in vol. I, where also the geology has been described (no. 3, by P. D. QUENSEL). Here I shall confine myself to a few introductory remarks. Masatierra is a deeply eroded, volcanic, Tertiary mountain range, extending I—W and built up by basalt beds of varying thickness; agglomerate banks and tuff layers also occur, but no recent volcanic products. In the eastern half the strata dip gently N, the crest, about 500 to 800 m high, follows the south coast- line and falls abruptly into the sea, forming a magnificent escarpment, cut by shallow, almost vertical hanging gulches, separated by narrow buttresses and more or less inaccessible. With the prevailing wind directions, the clouds discharge their moisture on the high ridges and feed the series of parallel wooded valleys trending north. West of the bold summit, El Yunque, the main ridge traverses the broadest part of the island, with peaks rising to 600—700 m; here erosion has cut deep back from both sides, leaving a narrow, precipitous rock wall standing, crossed by the Portezuelo pass and separating the two principal valleys of Masatierra, the Cumberland valley system and the Villagra valley. The main ridge or back- bone now turns northwest, at first cut by rather deep valleys going north and then following close to the north shore, which forms a high unbroken, inaccessible THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 887 escarpment; the westernmost forest patches are found here. The strata dip southeast, a direction consequently taken by the valleys, which run down to the south coast without reaching sea-level, as abrasion has been more effective than erosion. The reason for this is that the ridge west of Cerro Tres Puntas rapidly loses height, reaching its lowest point between the Padre and Carbajal coves, separated by the so-called Puente, only about 50 m high. Thus, the long western promontory does not force the winds up to an altitude sufficient to cool the air and cause the moisture to condense; rainfall is small, and the shallow gullies are dry most of the year. This is the barren, treeless section of Masatierra. In the eastern half of the island the lower slopes and the distal part of the valley bottoms also are barren; down toward the sea they may have been treeless always — this was, as we have seen, Johow’s opinion — but their poverty is otherwise, as shown by historical documents, due to man’s destructive influence (Figs. 32—34, Pls. 86, 90:1). GENERAL DISTRIBUTION OF THE PLANT COMMUNITIES. The principal plant communities may be grouped under three headings, ever- green dicotyledonous forest (“rain forest’), evergreen brushwood and scrub, and natural grass-land. To these we shall add the vegetation along the streams and waterfalls, the communities on rock faces, and the halophytic vegetation of the shore. Communities of hydrophytes are absent; there is no lake or pond in either island, and hardly a true water-plant in the streams, all of them small; there is nothing that could be called a river. Where the forest has been cleared, it has been replaced by secondary types of vegetation, the macal, formed by almost pure stands of Avzstotelia maqui, and the weed-fields, formed by a mixture of indigenous and adventitious herbs and grasses, or lacking all native species. No climatic upper timber-line exists in Masatierra; even the summit of El Yunque harbours a small forest grove. Nevertheless, one easily gets the impression that a timber-line is more or less distinctly developed at the foot of the backbone of east and central Masatierra, but it is a topographical or, as it were, edaphic line, following the foot of the perpendicular rock wall where the forest is forced to dissoive — see e.g. Figs. 31 and 35 and Pl. 88. The position of this line varies; 500—600 m is perhaps a common level, but it lies lower or higher according to the circumstances. Nobody who has followed the trail from the settlement up the wooded spur in the background of the valley to Portezuelo de Villagra has failed to observe that the composition of the forest changes with increasing altitude and that there is a sequence of belts, between which, however, no sharp limits can be drawn. Valleys and ridges from Puerto Frances to Pangal. Starting from the small cove up the valley, we cross the macal at about 200 m above sea-level; it is not pure, as some members of the original luma (Vothomyrcia) forest are still left. As we climb Cord6én Chifladores, the maqui falls back, and 888 C. SKOTTSBERG Fig. 31. Upper slopes of Valle Frances, Masatierra, showing forest boundary. — Photo 17/4 1917. in about 300 m we enter the virgin luma forest of the lower montane type, which as a rule extends to 400 or 450 m, where the region of frequent fogs begins and tree ferns appear in greater number; in this particular spot not much D7cksonia was seen below 500m (St. 11), but a little higher up it became plentiful, and Weymouthia mollis hung in festoons from the twigs of the lumas, testifying that we had reached the upper montane forest of the foggy belt. This forest fills the upper gulches of Valle Frances (Fig. 31). Where the forest abuts on the escarpment, and on the narrow ledges candelabrum trees, i.e. what I have called “‘the Rodinsonia assemblage’, appear, and the crest of the Chifladores ridge is covered with scrub. Quebrada de ia Pesca de los Viejos is, just as the valleys to the west, very barren in the outer half and covered with weeds, but a few solitary lumas linger on the slopes, and there is not much maqui. The interior is well forested, and the stream carries water at least during the winter months. The two following valleys, Quebrada de Laura (Fig. 32) and Q. de la Piedra Agujereada (Fig. 33) offer exactly the same aspect, the streams are dry during the summer months, except perhaps after an occasional deluge, and the valley ends in a coast cliff. Ascending the ridge between the Laura and Piedra Agujereada valleys, we soon strike the forest; at about 420 m Dicksonia is scattered, taking the lead at about 600 m, where it grows to twice a man’s height and where ¥xawza is more abundant than in any other place visited by us (St. 12). Here the forest extends over the ridge THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 889 Fig. some water in the stream-bed. Left, patches of Araena argentea. August 1917. 32. Quebrada Laura, Masatierra, looking North toward the sea. Marks of erosion on slopes; Photo K. Backstrom (St. 18). The gullies of Rabanal so named on account of the rdbanos (Raphanus sativus) once abundant in this place) end on a broad alluvial plain, where, in 1917, Silybum marianum \P\. 86) was in dominance. The lower slopes of the gullies were filled with maqui (Fg. 34), but higher up native trees appear; there was some Maqui everywhere, and the forest is degraded (Pls. 69, 80). Probably this valley was once wooded right down toward the sea. Cordon Centinela between Rabanal and Pangal can be followed up to the main ridge. Toward the sea the slopes and crest are very barren, but as we reach an altitude of 300 m, good forest is met with, at 400 m the lower montane type is well developed, Dicksonza begins to appear, and about 500 m the ridge is covered with forest, low and rather thin on the very crest, but abounding in large ferns and with several representatives or the Rodzzsonia assemblage (St. 15, 20), and on the flanks, which are very steep, there is virgin forest which, at about 550 m, has the character of a typical fern forest with dominant Dzcksonza and an abundance of Weymouthia and filmy ferns. The occurrence of Byssocaulon niveum in the fern forest is noteworthy. At about 780 m we reach the main range, covered with the kind of scrub inhabiting all the elevated, exposed ridges, a community char- acterized by Ugui Selkirkit, Pernettya, Gunnera bracteata, Blechnum cycadifolium etc. (St. 30, Pl. 92:2). 59—516796. The Nat. Hist. of Juan Fernandez and Easter Isl. Vol. II. 890 C. SKOTTSBERG Fig. 33. Quebrada de la Piedra Agujereada, Masatierra, looking South. Deforested slopes with a few scattered lumas. — Photo 17/4 1917. The entrance to the Pangal gorge is, at about 100 m, filled with degraded luma-maqui forest with Boehmeria excelsa along the stream. A little farther in, ferns become frequent in the undergrowth, and above 200 m good-sized Dcksonia appears in the bottom of the rapidly narrowing gulch; Arthropferis is a common scandent fern here (Pl. 87). Further advance is soon checked by a perpendicular cliff, over which the water comes down in a cascade, fringed with Cladzum and Gunnera peltata (Pl. 78, St. 38). The canyon walls are too steep to allow anything like a forest to exist, but solitary small trees and shrubs and patches of herbs and grasses have colonized every ledge and nook where there is sufficient soil to offer a foothold; landslides had left their mark here. Above the waterfall the val- ley is V-shaped and filled with forest, but the limit between the lower and upper montane type was not determined here (Pl. 89:1). From Cordén Escarpado to Cordén Salsipuedes. The deeply eroded valley system facing Bahia Cumberland drains the most elevated part of the main range. It consists of 3 valleys, Quebrada de la Damajuana, Valle Anson and Valle Colonial. The crest of Cordén Escarpado has low brushwood (St. 32), with a mixture of native species and weeds in the undergrowth. In the Damajuana gulch the macal begins at about 150 m above sea-level and fills the valley to the foot of the steep slope of Cordén Damajuana, which is very barren toward the sea (St. 46), but the inner part of the valley from about 200 or 250 m is well THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 8g! Fig. 34. El Rabanal, Masatierra, looking South. Macal in the gullies, higher up luma forest. Valley flat covered with weeds, S7/yéum martanum (dead) abundant. — Photo 5/3 1917. forested with large Boehmeriae on the stream banks (St. 2). At 250 m is a small waterfall. In order to gain a higher altitude it is necessary to follow the ridge which abuts on the Damajuana dome. The maqui disappears at about 300—350 m, the slopes bear dense and dark luma forest, toward 500 m Lcksonia becomes frequent, and the crest of the ridge is covered with mountain scrub with tall Ugenit Selkirkii and a fine representation of the Rodzxsonia flora (St. 24). Through a wide macal a trail leads into the Anson valley to an old clearing about 250 m above sea-level, called Plazoleta (Plazuela, “small square’) del Yunque; to the west, nearby, towers Pico Central (Pl. 89: 2). Well watered and surrounded by tall forest with planted Eucalypts at the edge, this sheltered spot guards memories from bygone times of warfare; Anson had his camp and garden here, and before him, Selkirk is supposed to have built his cottage in this hidden corner. The trail crosses the valley floor and soon ascends the slope of El Yunque. Dzck- sonia and tall-stemmed Alechnum cycadifolium appear and increase toward 400 m, the scandent ABlechnum Schottii is abundant, and at about 475 m the forest assumes the character of the upper montane type, with 7hyrsopieris, giant Hymenophyllum Juciforme and plentiful Weymouthia. As we approach the precipice of the main ridge, leading to the summit of E] Yunque, some rosette-trees appear, including one of the rarest species (St. 16). The cliff wall, down which the water rushes after a rain, is covered with Gunnera peltala and luxuriant Cladium scirpotdeum. 892 C. SKOTTSBERG The distal part of the Colonial valley between Cordén Central and Salsipuedes has lost almost every trace of its native flora; and only small groves of more or less natural forest are left on both sides of the spur which leads to Portezuelo, while the macal occupies large areas. From Quebrada del Monte Maderugo the forest extends up the west slope of Cordén Central and the face of Cerro Piramide. The pure macal changes above 200 m to a mixed luma-maqui stand, and in 400 m a dense forest is well developed on the steep slope (St. 6); toward the summit of the ridge, 540—550, the Rodzzsonia assemblage appears, particularly rich on the east slope of the crest; farther west, at the foot of the wall of Cerro Piramide, the same kind of vegetation was met with already at 380 m (St. 25). The crest of Cordén Central is a knife-edge, rocky and dry, culminating in Pico Central, 575 m, and covered with light mountain scrub alternating with patches of brush- wood (St. 31). West of Quebrada del Monte Maderugo is Q. Gutierrez (Pl. 84), into which a trail branches off from the path to Portezuelo. Old lower montane forest begins at about 300 m (PI. 93), invaded by maqui along the trail (St. 3), and the same type continues up to the ridge for another 150 m, where it ap- proaches the upper montane type. Portezuelo de Villagra. Thanks to the trail through this spectacular gate-way (Pl. 91:1), which offers the only passage over the main ridge, the steep rock-walls on both sides have become accessible. There is no place where the Robinsonia assemblage can be studied more conveniently or where it is better represented — two species, Denxdroserts macrantha and Eryngium inaccessum, have only been found here — and the same is true of the upper montane forest, partic- ularly on the Villagra side (St. 13); facing Cumberland Bay, a very narrow ledge near the Selkirk memorial tablet, covered with a patch of rich brushwood, can be reached (St. 21). The precipitous southwest face of Cerro Pirdmide is tapestried with verdure, an impenetrable tissue of brushwood, drenched in rain and mist and exposed to the full force of the wind, the influence of which is well demon- strated; the difference between the windward and leeward sides of the almost vertical buttresses is seen on Pl. 91:2. The narrow ridge of Cerro Piramide is covered with mountain scrub of the Ugni Selkirkit-Blechnum cycadifolium type. The Salsipuedes ridges. Rising abruptly above the floor of Valle Colonial, Cordén Salsipuedes (“get out, if you can’’) separates this valley from Puerto Ingles, which cannot by reached on foot along the shore. The ridge terminates in the escarpment of Punta San Carlos. On its east side is a series of: small, very steep hanging valleys, where no water comes down during the summer months (Pl. 90:1). The two valleys nearer the sea which were wooded 20 years before our visit, were now treeless, but the other still had forest, a fringe of macal below, about 300m native trees coming in, and above 400 m, just under the ridge, comparatively pure luma forest of a poor lower montane type (St. 7). The ridge is barren toward the sea and covered with degraded grass-land, but a little higher up native shrubs begin to appear, and about 600 m we enter a low brushwood (St. 26), alternating with open dwarf scrub and grass in dry, stony places (St. 27, 28). The ridge can be followed to about 740 m by climbing the serrated edge (St. 29). Between this ridge and the next, for which the name THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 893 Loma de los Munozes was given, lies Quebrada Salsipuedes, in its uppermost part filled with a fern forest, which goes over the ridge, where Blechnum cycadi- folium dominates in the undergrowth (St. 17, Pl. 58:2). Members of the Rodznsonia assemblage are found wherever the forest dissolves on the steep slopes as well as in the light brushwood. Beyond the Loma lies a hanging valley, facing Valle Ingles and named by us Quebrada de los Helechos; an almost pure Decksonzetum fills the bottom of this interesting gulch (St. 19) from about 660 down to 575 m, gradually passing into a forest of the ordinary upper montane type, which extends down to at least 450 m; luma and canelo attain unusually large dimensions, and the rare Pteris semiadnata \uxuriates. The gradient is very steep. There is no stream in this gulch. From Puerto Ingles to Quebrada del Juanango. The large and deep twin valley running south from Puerto Ingles once contained fine virgin forest; here is where the last living Savfa/uwm was discovered. The distal part of the flat valley floor has no forest left except the small Boehmeria grove observed by me in 1908 (2, pl. 7) and found intact in 1917. There are some shrubs of Cestrum parqui L'Hér. near the sea. Arzstotelia is less plentiful in this valley than in Cumberland Bay, but all open spaces are invaded by weeds (St. 50). From the central ridge both branch valleys are overlooked. The west branch has almost no forest left below 200 m, a few patches of luma and naranjillo and, remarkably enough because they are so easy to get hold of, small groups of Fuania. Walking up this valley, one has to cross the most extensive beds of Acaena argentea seen anywhere. Higher up the forest on the slopes is in much better condition (Pl. 88), and from about 380 m it passes into brushwood, a little further on Robzvsonia gayana appears, and at 475 m, more or less, low and thin brushwood follows the ridge. It is fairly rich in species, among them the rare Colletia spartioides. It was surprising to find Rea pruinata here, so far from the coast. Between Puerto Ingles and La Vaqueria towers the dome-shaped Cerro Alto, with its 627 m a conspicuous landmark. On the very steep inland faces are patches of forest. It is to be regretted that we neglected this mountain. A herd of semi-wild cattle roams the Vaqueria valley (hence its name). Toward the sea is degraded grass-land with stray patches of luma, with their compact heads looking as if they had been trimmed (Fig. 6), and higher up a thinned, dry forest. The sharp ridge separating Vaqueria from the Juanango gulch (Fig. 35) is more or less covered with scrub (St. 33), where a few rosette-trees come down to about 300m, a remarkably low altitude. Juanango has not been eroded to sea-level, but ends in a barranca, over which the stream comes down. The lowest part is barren and full of weeds (St. 51), but a little higher up are remnants of what I believe represents native grass-land (St. 45), and at 215 m above sea- level an open stand of lumas and Boehmeria follows the stream (St. 9), passing into the closed luma forest (St. 4) of the typical lower montane type on the slopes. From Quebrada Juanango to Bahia del Padre. The coast along this reach is a high, vertical cliff wall, falling abruptly into the sea. At the foot of Cerro Chumacera, which looks like a gigantic slab standing on end, are two 894 C. SKOTTSBERG Fig. 35. Dividing ridge between La Vaqueria and Quebrada del Juanango, Masatierra. Luma forest. — Photo 9/4 1917. small patches of luma forest, the last ones in this direction, and on the almost vertical face of Cerro Tres Puntas is a large carpet of Ochagavia elegans, the finest growth of this lithophyte found on the island, but from here to Punta de la Isla nothing but the poorest Avena barbata-grass-land is to.be seen. Bahia del Padre receives frequent visits from the langust-fishers and as a consequence is full of weeds, much more numerous than the scattered native litoral plants — with the exception of Sadicornia peruviana, which is abundant (Pl. 102). From the narrow, gently sloping beach a short climb leads to the Puente, a flat tuff bed with a field of mobile sand (St. 52). From Cabo Guasaballena along the south side of the island to Cerro Negro. Little can be added to what has been said already about this forbidding coast. The elevated summit ridge is, as it were, supported by a series of buttresses, separated by shallow ravines, where stunted trees, shrubs and ferns, among which the large leaves of Gunnera are conspicuous, cling to the rock. The high ridge is drenched in rain, and from the crest cascades tumble down into the abyss. The dome of El Yunque dominates the scenery, but it is often hidden from sight by fog (Fig. 2). At the foot of El Yunque, Cerro Negro rises like a bastion (Pl. 92, alt. about 180 m, one reading only). West of this hill stretches the Plano del Yunque; from here toward Punta Larga is pasture- land, where Acaena argentea and numerous other weeds are widely spread. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 895 The valleys of Villagra. This wide and well-watered valley system harbours some of the best upper montane forest left on the island; precipitation is, as far as we were able to judge, higher here than on the north slopes. Some maqui is found in the ravines, but there is no macal of any significance. Toward the sea, from the foot of El Yunque to the spur separating Quebrada de la Choza from the Chumacera valley, the forest ends abruptly in a sharp line, 220—225 m above sea-level (PI. 85:1). I regard this line as a climatic seaward timber-line, because I find it difficult to believe that under the present climatic conditions the forest ever went farther down. The vegetation below the forest _is xerophilous grassland (St. 43) with stony tracts with open, dwarfed scrub (St. 44): Most of the rain falls over the ridges surrounding Villagra Bay, and only along the water-courses is a closed fringe of bright green Guzmnera peltata, well set off against the yellowish grass and the silver gray Acaena (Pl. 85:2). Wind is another important factor, likely to force the forest back, and the edge is wind-trimmed. Our camp was situated on a low spur between two streams almost over- grown with Gunznera (Pl. 76:2), about 175 m above the sea. On the spur, be- tween the camp and the edge of the forest, low shrubs formed a scrub (Pl. 91), with Escallonia, Cuminia fernandesia, Margyricarpus, Robinsonia gayana, Thyrsop- teris, Dryopteris naequalifolia, Polystichun bertertanum, Blechnum Schott, Bromus fernandesianus and Peperomia berteroana, a very unusual assemblage at such a low altitude, for several of them an absolute record. The Choza forest is of typical lower montane type, tall-sstemmed, dark and with poorly developed undergrowth (St. 5, Pls. 94: 2, 95:1). It ascends to about 400 m, where it meets the precipice of the ridge; water trickles down over the rocks, where a few rosette-trees and patches of C/adium, typical of such localities, have gained a foothold. In the main Villagra valley, maqui extends down toc. 160 m. From about 400 m the forest assumes the character of the upper montane type (St. 13). Dzcksonia, Thyrsopteris and Weymouthia are common, below the pass Lacfor7s is in its prime, and the vegetation passes into the cliff forest already mentioned. On the west flank of El Yunque a small hanging valley is accessible to about 500 m above sea-level, where a vertical cliff wall checks further advance. The forest had been thinned, I don’t know for what purpose; there were numerous fallen trees and many dead ones still standing, and the ground had been grazed. Above approximately 400 m undamaged Dicksonia-Weymouthza-forest was met with, with thick-stemmed bamboo, Lactoris and seven species of Hymenophyllaceae (St. 14). From Villagra to Punta Larga. Proceeding west from Villagra, the last forest patches are encountered at the foot of Cerro Chumacera, two small groves of a poor luma forest with some members of the Rodézmsonia assemblage (St. 23), and below, at the foot of a projecting cliff, where water comes down also during the dry season, a small Boehmeria grove (St. 10). These groves are seen in Fig. 36. West from here the country is treeless grass-land, nearer to Vil- lagra mainly native Stpetum (Pl. 100), farther west large areas of Avena barbata; Acaena is common, but becomes scarce toward Punta Larga. Upon the whole, I got the impression that native species play a much greater réle in the composition of the grass-land on this side than on the north side of the island, where intro- 896 C. SKOTTSBERG Fig. 36. Most westerly forest groves on Masatierra at the foot of Cerro Chumacera. Below, a stand of Boehmeria excelsa, above two small groves of Mothomyrcia fernandeztana. — Photo 6/1 1917. duced species dominate. All the way cattle tracks run up and down the valleys and over the ridges. Referring to what was said above about the regional distribution of the forest types and of the individual species (Chapter III), I shall summarize the conclusions in a list of the species, divided into 3 groups. 1. Ranging over the entire wooded area; those marked with a * also reported from the summit of El Yunque. *Drimys, Fagara mayu, */uania, Nothomyrcia, Coprosma Hookeri and pyrifolia, * Escallonia, Rhaphithamnus, Pernettya, Berberis corymbosa, Colletia, Erigeron fruticosus, Robinsonia gayana, Rea micrantha, Halorrhagis masatierrana, Wahlenbergia Berterot (2) and fernandeziana, Mar gyricarpus, Solanum fernandezianum, Gunnera peltata, Bromus fernandeztanus, Carex berteroniana, *Cladium, Uncinia Douglasit, Libertia, Peperomia berteroana and margaritifera, Acaena ovalifolia, *Blechnum cycadifolium, auriculatum, chilense, valdiviense and Schottit, * Dicksonia berteroana, Lophosonia, Dryopteris, Polystichum berterianum and vestitum, Hypolepis rugosula, Adiantum chilense, Pteris berteroana and chilensis, Histiopteris, Polypodium lanceolatum, Gleichenia pedalts, Hymenophyllum cuneatum and plicatum. 2. Rarely or not ascending into the upper montane belt. Boehmeria, Sophora fernandeziana, Symphyochaeta, Rea pruinata, Dendroseris litoralis, Solanum robinsonianum, Wahlenbergia Larrainii, Lobelia, Arthropteris, Pellaea chilensis. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 897 3. More or less strictly confined to the upper montane belt, rarely seen below 400 m. Cuminia fernandeziana and *eriantha, Azara fernandesziana, *Ugni Selkirkit, Chusquea, Wahlenbergia Grahamae, Eryngium *bupleuroides and inaccessum, Plantago fernandezta, Robin- sonia thurifera, evenia and * gracilis, *Rhetinodendron, * Centaurodendron, Dendroseris macrantha and marginata, Phoenicoseris pinnata and *berteriana, Selkirkia, Hesperogreigia, Lactorts, Urtica fernandeziana, Gunnera bracteata, Peperomia fernandeztana, Dysopsis, Thyrsopterts, Asplenium dareovides, macrosorum and stellatum, *Pteris semiadnata, Polypodium Billardiert and intermedium, and the majority of the Hymenophyllaceae. THE SUMMIT OF EL YUNQUE. On Johow’s map the height is given as 927 m, but other figures are also quoted (Skottsb. z5. 156). The climb to the summit is difficult, which may be guessed already from the photographs, irrespectively of the side from where they were taken; the only approach is along the eastern spur, but even this route looks rather formidable, for it is a knife-edge ridge with very little foothold and with a precipice on either side of several hundred m. Besides, the rock is not reliable. VICUNA MACKENNA relates the first ascent made in 1795 by two convicts, who gained their freedom thanks to this feat. The author describes the summit as covered with pangue (Guznera), canelo (Drimys), chonta (Fuania), and murtillo, a name applied to the introduced Ugni Molinae (perhaps also to the endemic U. Selkirkii) and Pernettya; most likely U. Se/kirkiz is meant, perhaps also Pernettya. The ascent had not been repeated since, and we certainly would have made an attempt, if this had been possible, but none of us was an expert climber. Some of the younger fishermen are excellent mountaineers, but as it proved impossible to persuade anybody to assist, we had to give up. In 1921 a colonist who arrived in Masatierra after our visit, UBERLINDO ANDAUR by name, managed to make the ascent, and on February 7, 1922, he guided the late Dr. OrTro TENZ, an amateur botanist, to the summit, from where he brought an interesting collection, which was handed over to me for identification (zg). An account of this excursion was published in a Chilean newspaper (Tenz z) in 1922, and in 1924 he gave a new tale of it with more particulars (2); as this Chilean periodical had a small circulation only and seems to be rare in European libra- ries, I shall give a summary here. The ridge leading to the E. face of the mountain is more or less covered by scrub with Ugni Selkirkid and Blechnum cycadifolium as \eading species; scattered canelos occur, and this vegetation continues up the very steep barranca. Tenz mentions tall Blechnum and trunks of Dendroseris, presumably D. margznata or macrantha; his collection contains only two rosette trees to be thought of, Phoenicoseris berteriana and Yunguea Tenzii, both sterile. The former inhabits very wet, shaded places and the latter was found in the forest (Fig. 11), and from what he says in his newspaper article about a “Dendroseris’’, supposed to be new, this was Yuxqguea and not a species growing on the exposed ridge. On the rim around the wooded gulch was a large sedge, undoubtedly Cladium scirpoideum, which is in the collection. Other species quoted from the highest 898 C. SKOTTSBERG crest are short-stemmed Blechnum cycadifolium, Eryngium bupleuroides and Robin- sonia gracilis. The small ravine on the top faces south and is filled with rain forest ot an extreme type. Almost daily shrouded in the trade clouds, there is perhaps no wetter place on Masatierra. I quote Tenz 2, pt. 6.2. “Dichtes Pflanzenwuchs gedeiht auf dem schwarzen etwas sumpfigen Boden der Mulde. Bis hierhin drangen noch nicht die neueren Besteiger und wohl auch nicht die beiden ersten.. Fast alle Baume des jungfraulichen Waldes wachsen nicht senkrecht zum Boden, sondern schmiegen sich diesem eine Strecke weit an, der feuchten, lebenspendenden Seebrise nach Stiden entgegen, um dann bogig aufzusteigen. Zwischen alten, ehrwiirdigen Canelos, zottig mit Epiphyten bewachsen, recken sich machtige Dendroseris' empor. Glanzend griine Chontapalmen strecken ihren schlanken Leib dem Lichte entgegen. Aus ihren Wipfeln leuchten die gedrangten Rispen der knallroten Friichte. Durch das zierliche Blattwerk hoher Baumfarne,? das sich gegen den Himmel wie Filigran abhebt, schliipfen Sonnen- strahlen, werden auf den rotbraunschuppigen Farnstammen zuruckgeworfen und erzeugen im Halbdunkel des Urwaldes jenen Goldton, den wir an den Werken des Meisters Rembrandt bewundern. Das Unterholz bilden Robinsonien,? Rhetino- dendren,* grossblattrige Formen der rotbliihenden, endemischen Escallonia und Straucher der ebenfalls endemischen Lippenblitlergattung Cuminia® mit blauen Bliiten. Zierliche Farne der Gattung Litobrochia® bedecken den Boden.” In his newspaper article Tenz also mentions a plant with a trunk 2 m tall and oval, smooth leaves, evidently Cextaurodendron, of which there are unusually large leaves in the collection. THE STATIONS. The figures in brackets correspond to the numbers in my note-books. The Forest region. Physiognomically, the forests of Juan Fernandez have much in common with the Macaronesian laurineous forest as well as with the MJ/etroszderos-forest in the Hawaiian Islands, but at the same time it contains many elements linking them to the Valdivian rain-forest in south Chile, from which they differ by the total absence of Nothofagus as well as of woody climbers. The abundance of ferns, bryophytes and foliaceous lichens is the same, and most of the cryptogams are also found in south Chile. A conspicuous feature of the upper forest are the numerous rosette trees, almost without correspondence in the flora of the mainland. * Vunguea Tensitz. 2 Dicksonia berteroana. 3 Only large-leaved 7. graczlis in the collection. 4 Leaves twice the ordinary size. 5 Only C. ertantha in the collection. ° Pterts (Litobrochia) semzadnata. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 899 The Lower montane forest. 1. (2 E.) Nothomyrcta-Rea micrantha-Sociation. Puerto Frances, Loma del In- cienso, c. 360m, 13/12 1916. Low, thin forest on dry, stony soil. Ill I Rea micranta 3 Cladonia aggregata Nothomyrcia fernandeztana 2 > cocctfera Symphvochaeta macrocephala 1 Cora pavonia Diploschistes scruposus cop. Il: 3 . = Parmelia perlata 7 2 07° , *, y ry ~ . Wahlenbergia Larraini 1 Stereocaulon proximum II: 2 Epiphytes Bromus fernandezianus 3 Pseudocyphellaria aurata Carex berteroniana 2 Usnea dasypogoides Polystichum berterianum 2 Il: 1 Asplenium stellatum 1 greg. 2. (1 B.) Nothomyrcia-Drimys-Sociation. Quebrada de la Damajuana, c. 250 m. 6/12 1916. Well-watered ravine on steep slope. Black, wet vegetable mould. IV:2 Blechnum chilense i Fagara mayu 1 Polystichum berterianum 1 » vestitum 1 IV:2 Uncinia Douglasit 1 Nothomyrcia fernandeziana 4 - Drimys confertifolia 2 Boehmeria excelsa 1 Juania australis 1 (one tree Sophora fernandeziana 1 (one tree) Chiloscyphus integrifolius Ill Lophocolea fernandeztensts Madotheca chilensts cop. Schistochila berteroana Strepsilejeunea acuminata cop. Rigodium toxarium Thutdium Valdiviae Arthropteris altescandens, climbing on J/yrceugenia II: 3 Dryopteris inaequalifolia 2 Pteris berteroana 2 Blechnum cycadifolium 1% Epiphytes Lophocolea attenuata Lopholejeunia spinosa Mylia ligulata II: 2 Plagiochila neestana - > a Blechnum auriculatum 3 Radula microloba Adiantum chilense 1 Leptogium moluccanum 3. (1 D.) Nothomyrcia-Sociation. Interior of Valle Colonial, Quebrada Gutierrez, 7/12 1916. Deep black, tough vegetable mould. Pls. 74:1, 93. Typical lower montane forest, surveyed from 300 to 450 m, where the steep rock wall is met with. Not analyzed in detail. At 350—450 m some new elements appear, solitary trees of Rhaphithamnus, Fuania, Rea micrantha, and Azara, further Chusquea; Gunnera peltata is scattered, and,over the stones trails Arthropteris (abundant) and Blechnum Schottii (much rarer), rising high up in the trees. Evi- dently the upper montane type is approached. goo C. SKOTTSBERG On soil and decayed logs Epiphytes on Nothomyrcia and Drimys Lamprophyllum splendidisstmum Chaphidostegium aberrans Stereodon Lechlert Thamnium Ingae Eriopus leptoloma Chacopilum fernandezianum Rhynchostegium complanum Rigodium toxartum Madotheca subsquarrvosa cop. Monoclea forstert PALE Plagiochila riparia f ; : é phellaria cinnamomea Riccardia adglutinata cop. PSEU OEY fC ULT A CHEETA » variabilis Epiphytes on /wania Schistochila berteroana Dppioniaeaboiebile cn Symphyogyna Hochstettert cop. Bacidia delapsans Trichocolea opposite Pertusaria Skottsbergit Thelotrema lepadinum 4. (38.) Nothomyrcia-Sociation. Quebrada del Juanango, on steep slope facing south, c. 250m; gradient 30°. 9/4 1917. Wier Lunularia criuctata Nothomyrcia fernandeziana 4 Madotheca chilensts cop. Fagara mayu 1 Radula microloba Riccardia adglutinata II Schistochila berteroana ry mente a ° . . ci Nothomyrcia 2 Pannaria fuegiensis on dead timber Boehmeria excelsa \ Coprosma pyrifolia Epiphytes Fagara 1 Polypodium intermedium Rhaphithamnus venustus 1 » lanceolatum Gre Macromit Cun Sp. Rhaphidostegium caespitosum 5 Dryopteris tnaegualifolia eal 5 ni pi d : I Rigodium toxarium Polystichum berterianum 2 » vestitum 2 Riccardia adglutinata Ista Leptogium cyanescens Lobaria crenulata Uncinia Douglasit 4 Nephroma cellulosum IME ¢ Parmelia pertusa Asplenium dareoides 1 Pseudocyphellaria aurata flymenophyllum cuneatum 1 y berteroana » cinnamomea I » Juliginosa Rhaphidostegium aberrans » Guilleminit Rigodium toxarium Psoroma sphinctrinum Thamnium Ingae 5. (7.) Nothomyrcia-Sociation. Villagra, Quebrada de la Choza, 4/1 1917. Gentle slope, c. 250 m. Dark brown, plastic vegetable mould, with cover of dead leaves and twigs. Pls. 94: 2, 95:1. IV: 2 IV:1 Nothomyrcia fernandeziana 5 Nothomyrcia 4 Fagara mayu i Boehmeria excelsa 1 Drimys confertifolia 1 Fagara 1 THE VEGETATION OF THE JUAN III—II: 3 Nothomyrcia 3 Arthropteris altescandens 1 scandent Il: 3 Dryopteris inaequalifolia 1 Pteris berteroana 1 II: 2 Polystichum vestitum 1% Uncinia Douglasti 1 igs i Arthropteris 2 greg. Blechnum valdiviense 2 I Fissidens maschalanthus stones and soil Tsopterygium fernandeztanum FERNANDEZ ISLANDS gol Rhacopilum fernandezianum Stereodon Lechleri Madotheca chilensts Monoclea Forsteri Epiphytes Lamprophyllum splendidisstmum Rhacopilum fernandezianum Rigodium toxartum Stereodon Lechleri Thamnium Ingae Lophocolea attenuata Metzgeria decipiens Catillaria endochroma Lopadium leucoxanthum Parmelia soredica Pseudocyphellaria Durvillet 6. (12.) Nothomyrcia-Sociation. Interior of Valle Colonial, slope of Cordon Central, Quebrada del Monte Maderugo, c. 390 m, 18/1 1917. Gradient 30. Deep vegetable mould, mixed and covered with withered leaves, twigs etc. Logged. IV: 2 Drimys confertifolia 1 Fagara mayu 1 Blechnum Schottti 1% Halorrhagis masatierrana \ Uncinia Douglasti 1 ES: Arthropteris 2 Hymenophyllum cuneatum 2 Asplenium macrosorum 1 rare Dysopsis hirsuta 1 I On decaying wood, and stem-bases Ver Nothomyrcia fernandeztana 5 Boehmeria excelsa 1 Coprosma Hookeri 1 » pyrifolia i Ill Nothomyrcia 4 GR ‘ Tsopterygium fernandezstanum cop. ristotelia maqui 2 Pterygophylum tenuinerve Rhacopilum fernandeztanum Rigodium toxarium Lepidoszia disticha Lophocolea attenuata Saccogyna sqguarristipula Symphyogyna Hochstettert cop. greg. Arthropteris altescandens 1 scandent Rhaphithamnus venustus 1 Ih3 Nothomyrcia 2 Blechnum Schottit 1 scandent Dicksonia berteroana \ E2 Epiphytes Polystichum berterianum 2 Polypodium lanceolatum » vestitum 2 Mosses and lichens not recorded A little higher up the Weymouthia facies begins, and many new species appear, e.g. Azara, Cuminia eriantha and fernandezia, Rea micrantha, Thyrsop- teris, Asplenium stellatum, Pteris berteroana, and as epiphytes Peperomia fernan- deztana and Polypodium intermedium. 7. (4.) Nothomyrcia-Drimys-Sociation. Cordon Salsipuedes, hanging valley facing Valle Colonial, called by us Quebrada Seca, without a sign of running wa- go2 C. SKOTTSBERG ter during the summer, 435 m. 20/12 1916. Gradient 35°. Soil thin, sepia brown, crumbling and sandy under pressure, interwoven with roots and decaying leaves and twigs. IV: 2 Epiphytes Fagara mayu \ (poor) Hymenophyllum cuneatum r » plicatum IV: 1 Polypodium intermedium rare Nothomyrcia fernandeziana 4 » lanceolatum Drimys confertifolia 2 : Moss-lichen societies on trunks Rhaphithamnus venustus 2 Coprosma Hookeri i On Nothomyrcia » pyrifolia 1 Macromitrium fernandezianum Rea micrantha 1 Rhacopilum fernandezianum Juania australis 1 (one tree) Rigodium toxarium Ill Lopholejeunea spinosa Nothomyrcia 4 Leptogium callithaninion Drimys 1 » tremelloides Gunnera peltata I Lophosoria quadripinnata 1 Rhaphithamnus 1 J ; Lopholejeunia spinosa Metzgeria decipiens cop. On Fagara II: 3 Catillaria intermixta Polystichum berterianum 3 Graphis intricata Dichsonia berteroana 1 Thelotrema lepadinum Il: 2 On Drimys Polystichum berterianum 3 LODO EE ea Blechnum chilense 2 greg. Arthonia complanata Adiantum chilense 1 Megalospora versicolor Libertia formosa 1% Psoroma sphinctrinum Rhetinodendron Berterit (one seedling) Pyrenula aspistea Uncinia Douglasii 1 greg. = sag On Juanita It Aphanolejeunea diaphana Dysopsis hirsuta 2 Frullania stipatiloba cop. Hymenophyllum cuneatum 1 Lejeunea reticulata 2 plicatum 1 Bacidia delapsans Parmelia pertusa ] ‘(mostly around stem-bases) Om Coprosma Hookert Rhaphidostegium brachycladulum » caespitosum Frullania stipatiloba Fissidens crassicuspes » leptochaete Isopterygium fernandezianum Rigodium arborescens Lopholejeunea spinosa » toxarium Leptogium cyanescens Stereodon Lechleri Microphiale lutea Thamnium Ingae Parmelia laevigata ‘ H od » microsticta Chiloscyphus integrifolius cop. Parmeliella nigrocincta Lepidozia disticha Sticta laciniata Lophocolea attenuata Thelotrema lepadinum THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 9°03 On Rea micrantha Pseudocyphellaria aurata Rhaphidostegium sp. » berteroana Stereodon Lechlert » cinnamomea ; . » Richardt Frullania crassa B as soroma pholidotum cop. Madotheca subsquarrosa f Medes Pp » sphinctrinum Catillaria intermixta Graphis Dumastit Leptogium moluccanum Pyrenula mammillana Sticta lineariloba Lopadium leucoxanthum Boechmeria excelsa is found scattered in the forests of the lower slopes and becomes abundant along the watercourses in the bottom of the valleys, where it will take the lead, in cases to the exclusion of other trees. The pure Boehme- rietum is damp and dark. Pl. 64:2. 8. (2 D.) Boehmeria-Sociation. Exterior part of Pangal, below 200 m, 11/12 1916. ; LVi2 Il: 3-2 Fagara mayu 1 Histiopteris incisa 1 n Lophosoria quadripinnata 1 DV 1 Boehmeria excelsa 5 Nothomyrcia fernandeziana 1 II: 1, on stream bank Cardamine flaccida 1 greg. Scirpus cernuus 1 greg. Mosses and lichens not recorded. 9. (37 B.) Boehmeria-Nothomyrcia-Sociation. Quebrada del Juanango, 215 m. 9/4 1917. Not completely analyzed. IV: 1 Fagara mayu 1} III (open stand) Boehmeria excelsa 2 Drimys confertifolia 2 Nothomyrcia fernandeziana 2 Sophora fernandeziana 1 II: 2 Dryopteris itnaequalifolia greg. Polystichum berterianum greg. > vestitum greg. Uncinia Douglasti greg. Blechnum chilense » Schottit Bromus fernandezianus Cyperus eragrostis Histiopteris incisa Hypolepis rugosula Polypogon crinitus DES: Adiantum chilense Asplenium dareoides Arthropteris altescandens Blechnum valdiviense Cardamine flaccida Scirpus cernuus I Rhaphidostegium aberrans Rhacopilum fernandezianum Riccardia leptostachys on log Cladium scirpoideum Strepsilejeunea acuminata on stone 10. (8 C.) Boehmeria-Sociation. Foot of Cerro Chumacera, south side of is- land, a little below St. 23, 6/1 1917. Altitude probably between 350 and 400 m. Clayey, wet soil, watered by a cascade from the rock-wall of columnar basalt, and much tramped by cattle. Fig. 36. go4 C. SKOTTSBERG Vii Megs Boehmeria excelsa 5 Adiantum chilense 1 Blechnum chilense \ IT: 3 Pteris chilensts \ Histiopterts incisa 1 Pterts berteroana 1% hie Upper montaniewonesite 11. (40.) Drimys-Nothomyrcia-Sociation. Slope of Cordon Chifladores, interior of Valle Frances, west branch, 17/4 1917. C. 500 m, on very steep slope; stand somewhat open. Dark chocolate-brown soil, subsoil a loose, yellow, much weath- ered basalt. Wire Ie2 Drimys confertifolia 4 Juania 2 Nothomyrcia. fernandestana 2 Uncinita Douglasti 2 Hymenophyllum fuciforme 1 Wa Coprosma pyrifolia 2-1 Il: 1 Drimys 2-1 Dysopsts hirsuta 4 greg. Juanita australis 2-1 Asplenium stellatum 2 greg. Azara fernandeziana 1 Hlymenophyllum cuneatum 2 Coprosma Hlookeri % Trichomanes exsectum 2 Asplenium dareotdes 1 - » macrosorum rare Nothomyrcia 3 Blechnum valdiviense 1 greg. Dicksonia berteroana 2 Rhaphithamnus venustus 1 I Robinsonia evenia 1, on Dicksonia (on soil and decaying wood) Il: 3 Rigodium toxarium = ; Thamnium Ingae cop. Coprosma Hookert 3 e z Dryopteris inaegualifolia 3 Madotheca chilensis Nothomyrcia 3 Plagiochila gayana Polystichum vestitum 3 Symphvogyna Hochstettert Pteris berteroana 2 d Lemmopsis polychidiotdes Blechnum chilense » cycadifolium 1 Epiphytes Dicksonia i Serpyllopsis caespitosa v. fernandeziana Gunnera peltata 1 Stereodon Lechleri Juania i Bazzania peruviana Madotheca chilensis Radula decora Lactoris fernandeziana 1 Polystichum berterianum 1 Pseudocyphellaria argyracea Psoroma cephalodinum Sticta lineariloba This is a transition type; a little higher up D¢zcksonia increased and among the epiphytes were Hymenophyllum ferrugineum and Weymouthia mollis. 12. (35.) Drimys-Nothomyrcia- Weymouthia-Sociation. Ridge between Que- brada de la Piedra Agujereada and Q. Laura, a little below the crest, c. 650m, 5/4 1917. Stand very dense, dark. Deep, humid reddish-brown soil. Pl. 96. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS go5 IV: 2 Hymenophyllum ferrugineum Drimys confertifolia 4, very thick £ plicatum Nothomyrcia fernandesztana 4 a PULOSUM Trichomanes exsectum Mig Dicranoloma fernandeztanum Drimys 3 Nothomyrcia 3 Juania australis 2 Coprosma Hookert \ > pyrifolia 1 Lepyrodon tomentosus Macromitrium fernandezianum Pterygophyllum tenuinerve Ptychomnium falcatulum Stereodon Lechlert Ill Weymouthia mollis cop. Juania 3-4 Bazzania cerina Dicksonia berteroana 3 » peruviana Rhaphithamnus venustus 2 Lepidozia disticha Blechnum Schottii scandent » sernandeztensis cop. Coprosma Hookert 1 Plagiochila squarrosa Robinsonia evenia 1 on Dicksonta Saccogyna squarristipula ‘ Schtstochila berteroana cop. 3 Dryopterts inaequalifolia 2 Histiopteris incisa 1 Hypolepis rugosula \ Juania 1 Pteris berteroana 1 Phoentcoseris bertertana 1 rare Pseudocyphellaria subvariabilis Psoroma angustisectum Sphaerophorus melanocarpus Sticta laciniata » — latifrons » lineariloba Epiphyllous Hefaticae II: 2 Blechnum Schotttt 3 On /uania Siphonolejyeunea nudicalycina cop. lies. On 7richomanes Asplenium macrosorum 1 rare Aphanolejzeunea asperrima Cololejeunea Skottsbergit Microlejeunea sp. Dicksonia, one spm on Drimys Siphonolejeunea nudicalycina Epiphytes and on decayed wood 13. (9.) Nothomyrcia-Dicksonia- Weymouthia-Sociation. Villagra, a little below Portezuelo, 510 m, IO/I 1917. Very steep slope, gradient 40—45°, subsoil a soft, brick-red “clay-stone’’, looking like laterite (what it is not), vegetable mould mixed with leaves, pieces of wood etc, accumulated in hollows and fissures. Trunks of trees inclined-ascending. Vea Cuminia eriantha 1 Dicksonia berteroana 2 Escallonia Callcottiae 1 Drimys confertifolia 2 Gunnera peltata 1 Nothomyrcia fernandeziana 2 Pernettya rigida 1 Azara fernandeziana 1 Rhaphithamnus 1% Coprosma pyrifolia i Rhetinodendron Berterti \ Rea micrantha 1 Robinsonta evenia 1 on Dicksonia Rhaphithamnus venustus 1 » thurifera 1 I Thyrsopteris elegans 1 Berberis corymbosa 1 Blechnum cycadifolium % » Schottit 1 scandent Coprosma Hookeri i Iie 3 Dryopteris inaegualifolia 2 Gunnera peltata 2 60—516796. The Nat. Hist. of Juan Fernandez and Easter Isl. Vol. II. 906 C. SKOTTSBERG Hypolepis rugosula 2 Bromus fernandezianus 1 Polystichum berterianum 2 Hlalorrhagis masatierrana 1 Thyrsopteris 2 Histiopteris incisa 1 Blechnum cycadifolium 1 Hymenophyllum fuciforme 1% Lactoris fernandeziana 1 Peperomia berteroana 1 Lophosoria quadripinnata 1 Nothomyrcta 1% Pteris berteroana 1 Il: 1 Acaena ovalifolia 1 Dysopsis hirsuta 1 greg. I:2 Blechnum valdiviense Blechnum chilense 3 flymenophyllum plicatum 1 » Schottit 3 Mosses and lichens not collected. Weymouthia abundant. 14. (40 B.) Nothomyrcia-Dicksonta- Weymouthia-Sociation. South-west face of El Yunque, on steep buttress, c. 450—500 m, 24/4 1917. Relative frequency not estimated. VERT Epiphytes Drimys confertifolia (list incomplete) Nothomyrcia fernandeziana HHymenoglossum cruentum Asara fernandestana rare flymenophyllum cuneatum > Serrugineum Ill bb pes » Suciforme Dicksonia berteroana cop. 53 plicatum Chusquea fernandesiana, very stout 5 rugosum Cuminia fernandezia Polypodium Billardieri v. magellanicum Lophosoria quadripinnata Trichomanes-exsectuamn Rhetinodendron Bertertt : ; ; ; ; Lepyrodon tmplexus Robinsonia evenia on Dicksonia ay p Lopidium concinnum cop. II: 3 Rigodium toxarium Berberts corymbosa Thamnium Caroli ‘ “sf . tip . . 3lechnum cycadifolium Weymouthia mollis cop. Lactorts fernandeztana rare Madotheca chilensis II: 2 Coenogonium velutinum Asplenium macrosorum Pseudocyphellaria endochrysea Peperomia berteroana Sticta lineariloba Ground cover not collected. In the immediate vicinity of Portezuelo, 590—600 m above sea-level, the following epiphytic lichens were collected on bark. On Berberis corymbosa: Arthonia berberina and cytisi, Pvrenula Kunthii and mam- millana; on Kobinsonia gracilis: Catillaria leucochlora, Leptogium moluccanum, Pannaria hilaris, Psoroma pholidotum and sphinctrinum; on Robinsonia thurifera: Catillaria leuco- chlora, Pannaria hilaris, Pseudocyphellaria aurata and fragillima, Psoroma pholidotum and sphinctrinum, Pyrenula mammillana. 15. (33.) Nothomyrcia-Drimys-Weymouthia-Sociation. Cordén Centinela, c. 530m, 28/3 1917. Low, closed stand on steep slope a little below the crest. Pie Nothomyrcia fernandeziana 3 Coprosma Hookeri 3 Fagara mayu 2 Drimys confertifolia 3 THE VEGETATION OF THE JUAN Il Blechnum Schottit 2 scandent Thyrsopteris elegans 2 Coprosma 1 Dicksonia berteroana \ Drimys 1 Fagara \ Gunnera peltata 1 Nothomyrcia 1 Rhaphithamnus venustus = INP) Blechnum cycadifolium 2 Dryopteris inaequalifolia Polystichum bertertanum to N lige Blechnum Schottit 4 » chilense 1 Polystichum vestitum 1 FERNANDEZ ISLANDS QO7 1 (mostly on decayed wood and around stem-bases) Hymenophyllum ferrugineum Isopterygium fernandeztanum Baszania peruviana Lepidoszia disticha » pseudozoopsis Lophocolea attenuata Saccogyna crassistipula Epiphytes Polypodium intermedium Weymouthia mollis cop. Lepidozia plumulosa on Dicksonia Saccogyna crasststipula Lichens not collected. 16. (3 B.) Nothomyrcia-Dicksonia- Veymouthia-Sociation. Valle Anson; steep rock wall below Portezuelo del Yunque, c. 450—550 m, 18/12 1916. Relative frequency not estimated. IW Drimys conferttfolia Nothomyrcia fernandestana AWE Th Azara fernandeziana Blechnum Schottii scandent Juanita australis Ill Chusquea fernandeziana Dicksonia berteroana cop. Phoenicoseris pinnata Robinsonia evenia on Thyrsopteris Thyrsopteris elegans Mee) Berberis corymbosa Pernettya rigida Phoenicoseris berteriana rare Rhetinodendron Berterii Ne2 Bromus fernandesianus Cladium scirpoideum Ochagavia elegans 11) gs Asplenium macrosorum Asplenium stellatum Dysopsis hirsuta cop. Hymenophyllum caudiculatum » Sfuciforme I Plagiochasma rupestre Plagiochila neesiana Symphyogyna Hochstettert Epiphytes, the majority on WVothomyrcia Hymenoglossum cruentunt Hymenophyllum cuneatum » ferrugineum » plicatum Polypodium intermedium Dicranoloma Menstestt Tsopterygium fernandezstanum Lepyrodon lagurus Lopidium concinnum Macromitrium fernandeztanum Rigodium toxarium Stereodon Lechleri Weymouthia mollis cop. Trichocolea opposita Pyrenula mammillana cop. Sticta lineariloba 908 C. SKOTTSBERG 17. (2 B.) Drimys-Blechnum cycadifolium-Sociation. Near crest of Cordon Salsipuedes, c. 670 m, 8/12 1916. Soil covered with stems and dead fronds of the tree-ferns, full of pits. Pl. 58:2. IV:2 +1 neg Drimys confertifolia 5 Blechnum chilense 2 ssi Dryopteris inaequalifolia 2 Blechnum cycadtfolium 4 Dicksonia berteroana 2 Gunnera peltata 2 Blechnum Schottii 1 scandent on Dzck- Dicranoloma Billardieri sonia Ptychomnium fernandezianum Escallonia Callcottiae 1 Usnea cf. dasypoga cop. Epiphytes on Drimys (list very incomplete) Polypodium intermedium The absence of Weymouthia is noteworthy, but a consequence of the lack of Nothomyrcia and other trees on which it likes to grow. As a rule it seems to avoid Drimys. 18. (36.) Drimys-Dicksonia-Sociation. On ridge between Quebrada de la Piedra Agujereada and Q. Laura, c. 650 m, 5/4 1917. Close to St. 12. Soil covered with fern debris, decaying logs etc. IV: 1—Ill Drimys confertifolia 2 Coprosma Hookeri 1 » pyrifolia i Juania australis 1 ll Dicksonia berteroana 5 Blechnum Schottit 2 scandent Robinsonia evenia 2 on Dicksonia Cuminia eriantha 1 Phoenicoseris pinnata 1 Rhaphithamnus venustus 1 Robinsonia gracilis 1 1 123 Dryopteris tnaequalifolia 3 Blechnum cycadifolium 2 II: 2 Blechnum Schottii 3 Polystichum berterianum 2 Wea Blechnum valdiviense 2 I Dicranoloma fernandezianum Distichopyllum subelimbatum Pterygophyllum tenuinerve Rigodium toxarium Stereodon Lechleri Lepidozia bicuspidata » disticha Plagiochila squarrosa cop. Saccogyna squarristipula Schistochila berteroana cop. Symphyogyna hymenophyllum cop. Sphaerophorus melanocarpus Epiphytes (mostly on Drzmys) Serpyllopsis caespitosa v. fernandeziana Polypodium Billardiert v. magellanicum » intermedium » lanceolatum Dicranoloma fernandezianum Lepyrodon tomentosus Lopidium concinnum Macromitrium fernandeztanum Pterygophyllum tenuinerve Ptychomnium falcatulum Rigodium toxarium Stereodon Lechleri Weymouthia mollis Bazzania cerina » peruviana Lopholejeunea spinosa Madotheca chilensts cop. Leptogium moluccanum Pseudocyphellaria chloroleuca » Sragillima » subvariabilis THE VEGETATION OF THE JUAN Sticta laciniata cop. latifrons » lineariloba On Dicksonia Hymenophyllum cuneatum Trichomanes exsectum » philippianum FERNANDEZ ISLANDS 9°09 Lepidozia fernandeztensts Epiphyllous Hepaticae on 77ichomanes exsectum : Aphanolejeunea asperrima Cololejeunea Skottsbergit Siphonolejeunea nudicalycina 19. (11.) Drimys-Dicksonia- Weymouthia-Sociation. Small hanging gulch on Cord6n Salsipuedes (Loma de los Mufiozes), facing Puerto Ingles, 13/1 1917, 660 m. Peat formed by decayed Dzcksonza and covered with fallen stems and dead fronds of the tree-ferns. Steep slope; no running water. IV:2 Drimys confertifolia 1 LVeT Coprosma Hookeri i Drimys 1 Ot Dicksonia berteroana 4-5 Blechnum Schottii 2 scandent on Dick- Sonia Blechnum cycadifolium 1 Cuminia eriantha \ Nothomyrcia fernandeziana 1 Phoentcoseris pinnata 1 on Dicksonia Robinsonia evenia 1 on Dicksonia Thyrsopteris elegans Il: 3 Blechnum cycadifolium 2 » chilense 1 Dryopteris inaegualifolia i Gunnera peltata 1 II: 2 Blechnum Schottii 3 Polystichum berterianum 1 > vestitum 1 I:-r Blechnum valdiviense 2 Asplenium macrosorum 1 I Lepyrodon implexus Pterygophyllum tenuinerve Rhaphidostegium aberrans Rhizogonium Novae Hollandiae Lophocolea attenuata Monoclea Forstert cop. Schistochila berteroana Symphyogyna hymenophyllum Epiphytes (mainly on Drimys) Polypodium intermedium » lanceolatum Serpyllopsis caespitosa v. fernandeziana Dicranoloma fernandezianum Tsopterygium fernandezianum Rigodium toxarium (festoons) Weymouthia mollis on Nothomyrcia Bazzania peruviana cop. Pseudocyphellaria gilva Psoroma pholidotum On Dicksonia: Blechnum valdiviense Hymenophyllum ferrugineum Trichomanes philippianum Dicranoloma capillifolium Distichophyllum subelimbatum Tsopterygium fernandezianum Lamprophyllum splendidissimum Rigodium toxarium Chiloscyphus integrtfolius Lepidozia fernandeziensis Cop. Riccardia breviramosa on petioles Saccogyna squarristipula On Thyrsopteris Metzgeria multiformis on petioles 20. (32.) Drimys-Blechnum cycadifolium-Sociation. Cordén Centinela, on the ridge, c. 540 m, 28/3 1917. Thin soil on crumbling, weathered basalt. A transition type, with undergrowth including elements of the scrub. g1o C. SKOTTSBERG Ill I Campylopus aberrans Drimys confertifolia 3-4 ; Ptychomnium subaciculare Rea micrantha 2 Coprosma Hookeri \ Pseudocyphellaria mougeotiana Fagara mayu 1 Nothomyrcia fernandeztana 1 Robinsonia gracilis (few reaching this Rhaphidostegium aberrans Epiphytes (mostly on Drzmys) ) ; : Statin Bazzania cerina cop. Chiloscyphus integrifolius WB s ‘ Frullania stipatiloba Blechnum cycadifolium 4 Lepidozia fernandeztensis Drimys 3 Metzgerta multiformts Coprosma Hookert 2 Saccogyna squarristipula Pernettva rigida 2 Ugnt Selkirkit 2 Escallonia Callcottiae 1 Phoentcoserts pinnata 1 Robinsonia gayana 1 > gracilis i On petioles of Blechnum cycadifolium Thyrsopteris elegans 1 Leptogium callithamnion » moluccanum » phyllocarpum Sticta laciniata Lophocolea submuricata Wee Metzgeria multiformts . st On Rea micrantha Blechnum Schottit 2 Polystichum berterianum 2 Cattllaria intermixta Dicksonia berteroana \ Leptogium callithamnium » phyllocarpum Ik 1 Pseudocyphellaria aurata Hymenophyllum cuneatum 3 greg. Psoroma sphinctrinum The brushwood along the foot of the preciprces,anidyon fie rock ledges Above the closed forest cover the dark gray basalt ridges rise precipitously in steps; at the foot of the escarpment the forest dissolves; shrubs and low trees colonize the narrow ledges and fissures where some soil is found. The sun heats the rock, and in the middle of a clear day the temperature may be as high as near the coast, and showers are frequent. Here the sun- and light-loving rosette trees develop their full splendour. 21. (34.) Drimys-Nothomyrcia-Thyrsopteris-Sociation. Narrow ledge on north face of Cerro Piramide, c. 590 m, 31/3 1917. Upper margin of forest, where the rosette trees are concentrated to open spaces between the trees; perhaps better described as a complex of two sociations. III, over 3 m III, less than 3 m Drimys confertifolia 4 Thyrsopterts elegans 4 Nothomyrcia fernandeziana 4 Berberis corymbosa 2 Coprosma Hookeri 2 Blechnum Schottit 1 scandent Rhaphithamnus venustus 2 Centaurodendron dracaenoides 1 Aristotelia maqui 1 Chusquea fernandeziana 1 Juanita australis \ Cuminia fernandeszia 1 on ar . 5 Rea micrantha 1 Dicksonia berteroana 1% THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS gil Erynetum bupleuroides i Escallonia Callcottiae 1 Gunnera bracteata \ » peltata 1 » bracteata x peltata 1 Pernettya rigida 1 Phoenicoserts pinnata 1 Rhetinodendron Berterit 1 Robinsonia gayana 1 Ing Polystichum berterianum 3 Dryopteris tnaequalifolia 2 Escallonia 2 Plantago fernandezia 2 Carex berteroniana 1 Lophosoria quadripinnata \ Selkirkia Berteroit 1 Uncinia Douglasit 1 II: 2 Blechnum Schotttt 4 Polystichum berterianum 2 Blechnum chilense \ Bromus fernandezianus 1 Flalorrhagis masatierrana \ flymenophyllum fuctforme i Libertia formosa 1 Peperomia berteroana 1 Polystichum vestitum 1 QI 5 Blechnum auriculatum 2 Dysopsis hirsuta 1 greg. 22. (14.) Drimys-Ugni Selkirkit-Blechnum face of Cerro Pirdmide, steep rock wall, c. 590 to the prevailing winds than St. 21, and, like forest and scrub. Ill Drimys confertifolia 2 Coprosma Hookeri 1 Robinsonia thurifera 1 (Ne 8 Blechnum cycadtfolium 3-4 Uent Selkirkit 3-4 Pernettya rigida 2-3 Gunnera bracteata 2 Centaurodendron dracaenoides 1 Coprosma 1 Dendroseris macrantha 1 (2 spms) Escallonia Callcottiae 1 Asplenium stellatum 1 Elaphoglossum Lindenti \ (2 spms Hymenoglossum cruentum 1 Hymenophyllum caudiculatum 1 » cuneatum 1 » ferrugineum 1 » plicatum \ Ochagavia elegans 1 Peperomia fernandeziana 1 I Rigodium toxarium Stereodon Lechleri Thamnium Caroli Lophocoleu divergenticiliata Plagiochila fasciata » fuscobrunnea » gayana », riparia Schistochila berteroana Epiphytes Polypodium intermedium Lepyrodon implexus Macromitrium fernandezianum Rigodium toxartum Stereodon Lechleri Weymouthia mollis Madotheca chilensis cop. Plagtochila pudetensis » neestana Trichocolea vertictllata Sticta lineartloba cycadifolium-Sociation. Southwest m, 22/1 1917, much more exposed this, composed of elements from Gunnera peltata 1 Rea micrantha 1 Phoenicoseris pinnata 1 Robinsonia gayana 1 » gracitis 1 on Blechnum cycadifolium Nie 2 Blechnum chilense 2-3 Polystichum berterianum 2 Blechnum Schottit \ Halorrhagis masatierrana 1 eT Bromus fernandezianus 3 gi2 Hymenoglossum cruentum 1% Hymenophyllum plicatum 1 I Dicranoloma capillifolium lsopterygium fernandezianum Rigodium toxarium Chiloscyphus integrifolius Lepidozia fernandeztensis Cop. Mylia ligulata Riccardia breviramosa Nephroma antarcticum Stereocaulon proximum -. SKOTTSBERG Epiphytes (mainly on Drimys) Macromttrium saxatile Khynchostegium complanum Bazzania peruviana Frullania stipatiloba Lophocolea pallidevirens cop. Nephroma antarcticum > cellulosum Parmelia pilosella Pseudocyphellaria gilva 23. (8 B.) Nothomyrcia-Blechnum cycadifolium-Sociation. South side of island, foot of Cerro Chumacera, 6/1 1917. Altitude not known, probably 350— 400 m. This is the most westerly forest on Masatierra’s south side; thinned along the upper margin where the bed-rock is exposed and the rosette trees become conspicuous. IV:1 Cuminia fernandezia 1 Nothomyrcia fernandeziana 4 Rea micrantha \ Fagara mayu 1 os J II: 3 III Blechnum cycadifolium 2 Aristotelia maqui I Robinsonia gayana 2 Coprosma Hookert i Eryngium bupleuroides 1 » pyrifolia i Rhetinodendron Berterit 1 I shall give two examples of the colonies of rosette trees often found on otherwise barren rock-faces above the timber-line. 24. (14 E.) Mixed brushwood. NE face of Cerro Damajuana, 520 m, 29/1 1917. Bryophytes not collected. III—II: 3 Robinsonia gayana Chusquea fernandeziana Selkirkia Berterot Colletia spartioides II: 2 Dendroseris marginata Eryngium bupleuroides Pernettya rigida I Plantago fernandezia Pannaria rubiginosa var. Phoenicoseris pinnata Pseudocyphellaria aurata Rea micrantha > berteroana Wahlenbergia fernandeziana 25. (14 B.) Mixed brushwood. Valle Colonial, Quebrada del Monte Maderugo, 24/1 1917, c. 380 m. Steep rock wall facing W. Bryophytes not collected. IV: 1 Cuminia fernandezia Juanita australis Eryngium bupleuroides Escallonia Callcottiae Rea micrantha Rhaphithamnus venustus Sophora fernandeziana III Berberis corymbosa Coprosma Hookeri THE VEGETATION OF THE JUAN Diag Colletia spartioides Robinsonia gayana Selkirkia Berterot Wey Bromus fernandesianus Halorrhagis masatierrana FERNANDEZ ISLANDS 913 Ochagavia elegans Peperomia berteroana » Sernandeziana I Porina depressula 26. (10.) Rea micrantha-Drimys-Nothomyrcia-Sociation. Cordén Salsipuedes, steep, stony ridge, interrupted by rock ledges; gradient 30°. A thin cover of vegetable mould on ledges. II—II: 3 Rea micrantha 3 Coprosma Hookeri 2 Drimys confertifolia 2 Nothomyrcia fernandeziana 2 Aristotelia maqui I Cuminia fernandezia 1 kagara mayu |} Phoenicoseris pinnata 1 Rhaphithamnus venustus 1 Il: 3 Blechnum cycadifolium 1 Escallonia Callcottiae 1 Lophosoria quadripinnata 1 Robinsonia gayana 1 Thyrsopteris elegans 1 H2 Polystichum bertertanum 2 Bromus fernandeztanus 1 greg. Adiantum chilense \ 13/1 1917, 625 m. IME Hymenophyllum cuneatum 1 » plicatum 1 I Dicranoloma Billardieri Rigodium arborescens » robustum Cladonia pycnoclada Pseudocyphellaria mougeotiana Stereocaulon proximum Epiphytes (mostly on Drzmys) Macromitrium saxatile Ptychomnium subaciculare Rigodium toxartum Weymouthia mollis on Nothomyrcia Catillaria leucochlora on Cumintia Pseudocyphellaria Durvillet » gilva » Richardi Psoroma pholidotum Sticta lineariloba The brushwood and scrub on the exposed ridges. 27. (10 B.) Pernettya-Escallonia-Sociation. Cordén Salsipuedes, on ridge, alter- nating with St. 26, 13/1 1917. Very exposed, soil thin and stony. II: 3—2 Pernettya rigida 4 Escallonia Calicottiae 2 Coprosma Hookeri 1 Robinsonia gayana 1 II: 2 Bromus fernandezianus 2 greg. Polystichum berterianum 2 greg. fHalorrhagis masatierrana 1 Libertia formosa 1 IVES. Acaena argentea 3 I Campylopus aberrans Dicranoloma Billardieri cop. Rhacomitrium symphyodontum Rigodium toxarium Cladonia coccifera » pycnoclada Pseudocyphellaria mougeotiana gI4 C. SKOTTSBERG 28. (2.) Pernettya-Gunnera-Sociation. Cordon Salsipuedes, knife-edge ridge I—2 m wide, 615 m, 8/12 1916. Soil stony, very shallow, trees dependent on fissures. Ill Cladonia furcata Drimys confertifolia 1 Stereocaulon ramulosum Escallonia Callcottiae \ Epiphytes Fagara mayu | Nothomyrcia fernandeztana 1 Phoenicoserts pinnata 1% Rigodium toxarium Dicranoloma Billardtert Ptychomnium subaciculare Serpyllopsis caespitosa v. fernandeziana HI 23 Gunnera peltata Pernettya rigida Phoenicoseris pinnata 2 Blechnum cycadifolium 1 No Chiloscyphus tntegrifolius Frullania stipatiloba Radula decora No Coprosma Hookeri 1 Megalospora versicolor Gunnera bracteata 1 Psoroma pholidotum cop. Robinsonia gayana | Pseudocyphellaria gilva Thyrsopterts elegans 1 > chloroleuca » Durvillet He » endochrysea Bromus fernandezianus 3-4 Polystichum berterianum 2 Blechnum Schottit % » Richardt cop. Stereocaulon ramulosum Flalorrhagis masatierrana 1 On Phoentcoseris Margyricarpus digynus \ Frullania magellanica Polystichum vestitum % : F Arthopyrenia Cinchonae I Pseudocyphellaria berteroana Dicranoloma Billardtert Pyrenula mammillana Ptychomnium subaciculare Rhacopilum asstmile Rigodium arborescens Pseudocyphellaria berteroana > toxarium cop. Psoroma pholidotum On Margyricarpus 29. (2 C.) Pernettya-Robinsonia gayana-Sociation. Cordén Salsipuedes, on sharp edge of ridge, c. 700 m, 8/12 1916. NE) 8%} MWe Pernettya rigida 4 Blechnum auriculatum 1 Robinsonia gayana 3 Gnaphalium stachydifolium 1 Eryngium bupleuroides 1 Phoenicoseris pinnata I Robinsonia gracilis 1 Campylopus aberrans Dicranoloma Billardiert 2 3} . - . II: 2 Ptychomitrium fernandesianum Gnaphalium cheiranthifolium 1 . oc ae : Epiphytes Margyricarpus digynus 1 Leptogium moluccanum Psoroma pholidotum 30. (39.) Ugni Selkirkii-Blechnum cycadifolium-Sociation. Main ridge at junc- tion with Cordon Centinela, c. 780 m, 11/4 1917. Pl. 92:2. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 915 II: 3 Ugni Selkirkit 5 Blechnum cycadifolium 3 Pernettya rigida 2 Chusquea fernandestana \ Coprosma pyrifolia i Dicksonia berteroana \ Drimys confertifolia 1 Gunnera bracteata 1 Robinsonia gracilis 1 iWleg Polystichum bertertanum 2 I Dicranoloma Billardiert cop. Ptychomnium subaciculare Thysanomitrium Richardt Adelanthus sphalerus Lepidoszia fragillima » Jernandesziensts Cop. Marsupidium piliferum Lophocolea pallidevirens Plagiochila pudetensis Schistochila berteroana Tylunanthus limbatus Cladonia didyma Psoroma angustisectum Epiphytes (mainly on tree-ferns) Hymenophyllum cuneatum » plicatum » rugosum Polypodium Billardiert v. magellanicum Serpyllopsis caespitosa v. fernandeziana Dicranoloma capillifolium Distichophyllum subelimbatum Ptychomnium subaciculare Bazzanita cerina Lepidozia plumulosa » Sernandeztensts » Sragillima Marsupidium piliferum on Drimys Metzgeria multiformis Plagiochila pudetensts on Drimys » SQUArTOSA Saccogyna squarristipula Parmeliella pycnophora var. on bark Pseudocyphellaria berteroana cop. on Coprosma Sticta laciniata on bark 31. (12 B.) Pernettya-Ugni Selkirkiz-Sociation, Valle Colonial, Cordén Central, Wes: Pernettya rigida Blechnum cycadifolium Escallonia Caltcottiae Robinsonia gayana » gractlts Ugni Selkirkit Wee Anthoxanthum odoratum Erigeron fruticosus 18/1 1917, c. 575 m. Crest of ridge, 1—4 m wide; dry, stony soil, with open spaces invaded by weeds. Relative frequency not estimated; no mosses or lichens collected. Gleichenia pedalis Halorrhagis masatterrana Libertia formosa Polystichum berterianum Wahlenbergia fernandeztana » Grahamae II: 1 Acaena argentea Bromus fernandezianus Rumex acetosella The scrub on the lower and drier ridges is poorer and more invaded by weeds, and species common in the grass-lands occur. Whether or not the forest once extended over these ridges is hard to tell, but not impossible. On the south side of the island communities of small shrubs alternate with grass- and weed-land; they will be treated together. 32. (3.) Pernettya-Escallonia-Sociatiop. Bahia Cumberland, Cordon Escarpado, on the peak, c. 360 m, 17/12 1916. Dry, barren and stony ridge separating Que- brada del Minero from Pangal. 916 I Pernettya rigida 3 Escallonia Callcottiae 2 Coprosma Flookert 1 Drimys confertifolia 1% i Dipsacus silvestris I Polystichum bertertanum 1 greg. Danthonia collina 1 Gunnera peltata 1 Stipa neestana 1 iW ibein Acaena argentea 2 Rumex acetosella 2 Blechnum auriculatum % (2—3 m) C. SKOTTSBERG Vulpia bromoides Under a stone Hymenophyllum cuneatum » plicatum I Campylopus aberrans » polytrichoides Rhaphidostegium caespitosum Cora pavonia Lecidea inactiva cop. Parmelia pilosella Pertusaria hadrocarpa Pseudocyphellaria mougeotiana Stereocaulon proximum 33. (38 B.) Pernettya-Escallonia-Sociation. Ridge separating La Vaqueria from Quebrada del Juanango, c. 300 m, 9/4 1917. Relative frequency not estimated. L3 Escallonia Callcottiae, common Pernettya rigida, common Cuminia fernandezia Eryngium bupleuroides Rea micrantha Robinsonia gayana II: 2 Anthoxanthum odoratum Evrigeron fruticosus Flalorrhagis masatierrana Ochagavia elegans Wahlenbergia Larrainii Cryptogams not collected. Perpendicular rock faces, particularly on the more elevated ridges where there is sufficient moisture, are covered with communities of mosses and lichens and colonized by scattered chasmophytes. 34. (5.) Ochagavia-moss-lichen complex. Cordén Salsipuedes, rock face in Quebrada Seca above the forest described St. 7, 20/12 1916, c. 450 m. Pl. 97. lig ——2 Robinsonia gayana 1 Wahlenbergia Larrainii 1 I Big Ochagavia elegans, large bed Bromus fernandezianus 1 Halorrhagis masatierrana 1 Pernettya rigida \ I Amphidium cyathicarpum Thysanomitrium Richardi frullania stipatiloba Buellia sp. Caloplaca Selkirkii cop. » subcerina Lecidea cyanosarca > inactiva Pertusaria hadrocarpa Porina fernandeziana » rufocarpella Ramatlina linearis Schismatomma accedens » melancholicum 35. (34 B.) Community of Bryophytes and Lichens. Portezuelo de Villagra, c. 590 m, 3/12 1916, 21/1 and 30/4 1917. Vertical wall near the Selkirk Memorial tablet, facing NE. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS QI7 Campylopus introflexus Lecidea cyanosarca Ditrichum affine » enteroleuca Fissidens pycnotylus > inactiva Rhacomttrium symphyodontum > viridans Thysanomitrium Richardt Parmelia caperata Tortula scabrinervis . laevigatula » pilosella Pertusaria hadrocarpa Pseudocyphellaria argyracea Psoroma sphinctrinum Ramalina linearis Frullania stipatiloba Jamesoniella colorata Lepicolea ochroleuca Plagiochila hyadesiana Blastenia fernandeziana Schismatomma accedens Buellia sp. cop. » melancholicum Caloplaca rubina Stereocaulon proximum » Selkirkii cop. » ramulosum Catillaria theobromina Sticta lineariloba Cladonia gracilis cop. Usnea dasypogoides 36. (5 B.) Community of Bryophytes and Lichens. Portezuelo de Villagra, c. 600 m, 25/12 1916. Vertical wall facing SW. Anacolia subsessilis Lepicolea ochroleuca Dicranoloma Billardieri Ditrichum affine Fissidens pycnotylus Rigodium toxarium Thysanomitrium Richardt Weymouthia mollis (poor) Lecidea enteroleuca Parmelia laevigatula Pertusaria hadrocarpa Pseudocyphellaria argyracea Stereocaulon ramulosum Theloschistes flavicans 37. (7 B.) Ramalina linearis-community. South side of island, Loma de las Cabras, rock pinnacle, 380 m, 5/1 1917. Sea-birds often perch on this rock, which was spotted with their droppings. Ramatlina linearis, covering Lecidea avium the west face, otherwise scattered Pertusaria hadrocarpa Parmelia perlata, forming a fringe » melanospora around the base Physcia picta Gyalecta jenensts Stereocaulon proximum in fissures The Vegetation at the streams and cascades. Innumerable rills come down from the ridges in the wooded country, where most of the valleys from Frances to Juanango carry some water most of the time; the current is as a rule swift and the water leaps over the thresholds formed by harder layers of basalt, which are covered with luxuriant verdure. 38. (6.) The canyon of Pangal, at the first waterfall counted from below, 200 m, 1/1 1917. Pl. 98. a. Gunnera peltata-Cladium-Fissidens rigidulus-Sociation. In the spray on the side of the running water and probably more or less inundated after heavy rains. 918 Iisyg Gunnera peltata 3 II: N Cladium scirpotdeum 4 Polypogon crinitus 2 greg. Blechnum chilense 1 Bromus fernandesztanus Carex berterontana 1 Rumex cf. sanguineus I Bieri Cardamine flaccida 3 Scirpus cernuus \ c. SKOTTSBERG I Fissidens rigidulus cop. Philonotis glabrata Ptergygcophyllum anomalum Thamnium proboscideum Anthoceros sp. Marchantia foliacea Plagiochasma rupestre Symphyogyna circinnata b. Cardamine flaccida-Philonotis krauseana-Sociation. In the stream. West Asplenium dareotdes 3 Cardamine flaccida 2 I Philonotis krauseana cop. Sctaromium pachyloma, floated Lophocolea sp. Marchantia foliacea Megaceros fuegiensts cop. Symphyogyna circinnata 39. (1 C.) Cardamine flaccida-Fissidens rigidulus-Sociation. Quebrada de la Damajuana, margin of small waterfall, c. 250 m, 6/12 1916. 2 Uncinia Douglasti 2 Il: 1 Cardamine flaccida 3 Asplenium dareotdes 1 I fissidens rigidulus cop. Thamnium crassinervium Megaceros fuegiensts Marchantia foliacea Monoclea Forstert Symphogyna Hockstettert Leptogium moluccanum 40. (37 C.) Cardamine flaccida-Fissidens rigidulus-Sociation. Quebrada del Juanango, c. 200 m, 9/4 1917. Boulders at foot moss-cushions between the boulders. Wes Peperomia berteroana 1 rare » margaritifera \ rare Il: 1 Cardamine flaccida 2 Scirpus cernuus 1 41. (8 D.) Fissidens rigidulus-Sociation. rock wall above the Boehmeria-forest (St. 10), Wee) Brassica napus I Plantago major 1 of waterfall washed by the stream; I Fissidens rigidulus cop. Rigodium toxarium Sciaromium pachyloma Riccardia adglutinata On decaying log Rhacopilum fernandeztanum Riccardia leptostachya South slope of Cerro Chumacera, 6/1 1917. I fissidens rigidulus cop. Rigodium toxarium Sciaromium pachyloma THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 919 Lophocolea divergentictliata Plagiochasma rupestre » sp. Symphyogyna circinnata Marchantia polymorpha 42. Gunnera peltata-Sociation. Villagra, below the lower timber-line, forming a fringe along the stream. 5/1 1917. Pls. 76:2, 85:2. 143 Gunnera peltata 5 At the streams in Bahia Cumberland, especially in the settlement, many more species have been found; of these Vasturtium aquaticum was introduced on pur- pose. Callitriche marginata var. Lechleri (p. 781), indigenous in Chile, Fuxcus capillaceus and plantfolius, Oldenlandia thestifolta, all not found before 1916, and Heleocharis maculosa and Polygonum hydropiperoides, which were known before, seem to be restricted to the inhabited valley and are supposed to be recent additions to the flora, but may have been brought by visiting birds (e. g. Vanellus cayennensis, Johow 2.238), if not with the traffic. The Grass-lands. In the native grass-land of these islands two species of Spa dominate, ac- companied by other xerophilous species, and it lies near at hand to refer this type to the steppe formation. The climate is no steppe climate, the vegetation has little to do with a true steppe or prairie, whereas a comparison with the Mediterranean steppe is more acceptable. Similar steppe communities are found in central Chile, where the climate gets its character from the winter-rains and the very dry summer. To judge from the available meteorological observations in Juan Fernandez, the difference in precipitation between summer and winter is less pronounced, but we know from experience that the basal region on the south side of central Masatierra, and the entire west section, must receive much less rain than the montane region. In its physiognomy the insular grass-land differs materially from the open country in central Chile in the absence of the xerophilous chaparral shrubs and spring annuals and also in the total lack of Cactaceae. There are numerous annuals it is true, but all of them are aliens. In stony places we also find a few stunted shrubs belonging to species common along the barren ridges in the forested country. 43. Stipa laevissima-Sociation. Villagra, c. 150 m, 4—6/1 1917. On gentle slope; open, sandy-gravelly soil. Pl. 100. i<2 Piptochaetium bicolor 1 Stipa laevissima 4 Polypogon chilensis \ > wmeestana 1-2 Scirpus nodosus 1 Cyperus eragrostis 1, rare eas Flalorrhagis masatierrana 1, rare Juncus procerus 1, rare >» dombeyanus 1 Libertia formosa 1 Blechnum auriculatum 1% Spergularia confertifiora 1 920 C. SKOTTSBERG As far as our experience goes, a community as pure as this only covers small areas; as a rule some of the weeds listed below are also present, sometimes in abundance. Acaena argentea, Avena barbata, Aira caryophyllea, Anthoxanthum odoratum, Bromus unio- loides, Cynoglossum creticum,’ Dipsacus silvestris, Geranium core-core, Hordeum murinum, Hypo. chaeris glabra, Gnaphalium cheiranthifolium, Lolium multiflorum, Lythrum hyssopifolia, Malva parviflora, Medicago denticulata, Melilotus indicus, Ranunculus muricatus, Rumex acetosella, crispus and pulcher, Silene gallica, Silybum marianum, Sonchus oleraceus, Verbena litoralis, Vulpia bromoides. 44. (8.) Dwarf scrub—grass-land complex, 150—200 m, 4—6/1 1917. Along the south side of the island are many places with outcrops of rock and with thin stony soil, particularly on the ridges between Villagra and Punta Larga. The list below is not an analysis of a definite area, but a summary of our notes. Ligneous or suffruticose species Lythrum hyssopifolia ike Vulpia bromoides Blechnum cycadifolium I Cuminia fernandezia rare Cryptogams Erigeron fruticosus rare h on hard basalt, e on earth Escallonia Callcottiae ; : Anacolta subsessilis var. Halorrhagis masatierrana Margyricarpus digynus . Ca y WE om CUT Rea pruinata h Grimmia phyllorhizans Wahlenbergia fernandeziana rare hokiyeh i i he us AIEEE e Rhacopilum fernandezianum 10 Jamesoniella colorata Acaena argentea greg. Spergularia confertifiora Anaptychia hypoleuca h Blastenia fernandeziana cop. Herbs and grasses Buellia masafuerana II: 2 h » stellulata Blechnum chilense SP: h Caloplaca rubina > Selkirkit cop. Cladonia aggregata » cocctfera » pityrea Cora pavonia Lecanora atra Lecidea avium Bromus unioloides Gnaphalium cheiranthifolium Histiopteris tncisa Hordeum murinum oo om O Juncus imbricatus Libertia formosa Lolium multiflorum Piptochaetium bicolor - - » J af Polystichum berterianum, common h Dd peabaag Toll RS: Scirpus nodosus Pertusaria hadrocarpa h » melanospora Cop. Stipa laevissima Physcta sp. Placopsis fuscidula Ramalina linearts > neestana II: 1 Adiantum chilense Pseudocy phellaria mougeotiana Blechnum auriculatum common e Stereocaulon ramulosum Hymenophyllum plicatum rare e Sticta Weitgelit * C. creticum Mill., an older name for C. fictum Ait. in my plant catalogue. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS g21 On the north side of Masatierra much less is left of what I regard as native grass-land; the barren slopes, formerly at least in part forested, are now waste-land. 45. (37.) Stipa laevissima-Sociation. Quebrada del Juanango, on gentle slope near the sea above the low barranca, 9/4 1917. Dry, sandy soil. II: 2 Polypogon chilensis 1 Stipa laevisstma 5 Scirpus nodosus 1 Bromus unioloides 2 Setaria imberbis 1 Stipa neesiana 2 Uncinia Douglasti 1 Carex berteroniana 1 A little higher up the composition changes, Carex and Uxcimza are more frequent, but at the same time weeds become much more numerous; see St. 50. The Waste-lands on the north side of Masatierra. Waste-land extends from one end of the island to the other, but no vege- tation analyses were made west of Bahia Cumberland. At best we find a degraded grass-land, where S7pa laevissima and neestana occur, in cases also Pipfochaetium bicolor and Polypogon chilensis, anda stray specimen of “scallonia, Margyricar- pus, Paronychia, Funcus imbricatus, Spergularta and Wahlenbergia Berterot or Lar- rainit may be seen here and there. The flats and lower slopes near the settle- ment which are covered with alluvial soil, look more like a desert than anything else, subject as the are to the eroding forces of the water after every heavy rain (Fig. 5) when the earth carried off colours the sea in the bay, and erosion phe- nomena are observed in most of the denuded valleys near the sea (Figs. 32, 33), Traces of solifluction were distinct enough in Valle Colonial; an example is given inyotay7z. 46. (1.) Barren ridge north of Cerro Damajuana, c. 150 m, 6/12 1917. Dry, rocky and sandy soil; vascular plants all scattered or forming patches; relative frequency not estimated. II: 2 I Bromus unioloides e on earth s on stones Gnaphalium cheiranthifolium Oenothera mollissima LBuellia sp. Cora pavonia S e e Diploschistes scruposus II: 1 z s Acaena argentea Anthoxanthum odoratum Lecanora atra Parmelia caperata es » cetrata Blechnum aurtculatum Ss > conspersa Danthonia collina e » microsticta Hymenophyllum plicatum Rumex acetosella S S » savatilts e Peltigera rufescens s Pertusaria hadrocarpa e Stereocaulon proximum 47. (2 F.) Ugni Molinae-Blechnum auriculatum-Sociation. Valle Colonial, c. 30 m, 12/12 1917. Steep micro-terraced talus slope; soil orange-red, sandy-clayey; terrace surface covered with washed-out gravel, edge firm, held together by Blech- num auriculatum. 61 — 516796. The Nat. Hist. of Juan Fernandez and Easter Isl. Vol. II. Q22 C. SKOTTSBERG II: 2-3 II: 1 Ugni Molinae 1-2 Blechnum auriculatum 4-5 Acaena argentea 2 I]: 2 Rumex acetosella 2 Anthoxanthum odoratum 2-3 Micropsis nana 1 Escallonia Callcottiae 1 Blechnum cycadifolium 1 48. Weeds observed by us in Valle Colonial. A few garden escapes and weeds reported by earlier collectors, but not found again, have been excluded. Acaena argentea, Aira caryophyllea, Amaranthus chlorostachys and deflexus, Ammi vis- naga, Anthemis cotula, Anthoxanthum odoratum, Apium laciniatum, Artemisia absinthium (re- cently recorded), Avena barbata, Bidens leucantha, Briza minor, Bromus unioloides, Centaurea melitensis, Cichorium intybus, Chenopodium multifidum and murale, Convolvulus arvensis, Cyno- glossum creticum, Dipsacus silvestris, Euphorbia peplus, Geranium core-core, Gnaphalium chei- ranthifolium and spicatum, Hordeum murinum, Hypochaeris glabra, Lobelia tupa, Lolium multi florum, Lythrum hyssopifolia, Malva nicaeensis, Marrubium vulgare, Medicago arabica and lupulina, Melilotus indicus, Micropsis nana, Modiola caroliniana, Oenothera mollissima, Oxalis corniculata, Physalis peruviana, Plantago lanceolata and major, Polygonum aviculare, Ranunculus muricatus, Rubus ulmifolius (recently recorded), Rumex acetosella, crispus and pulcher, Ruta chalepensis, Setaria imberbis, Silene gallica, Silybum marianum, Sisymbrium officinale, Sonchus oleraceus, Stellaria media, Taraxacum fernandezianum, Verbena litoralis, Vulpia bromoides, Xanthium spinosum. 49. Weeds of the valley floor in Puerto Ingles, 19/1 1917. Acaena argentea, Aira caryophyllea, Anthoxanthum odoratum, Avena barbata, Bromus unioloides, Chenopodium multifidum and murale, Convolvulus arvensis, Cynoglossum creticum, Euphorbia peplus, Geranium core-core, Gnaphalium cheiranthifolium, Lolium multiflorum, Malva parviflora, Marrubium vulgare, Rumex acetosella, crispus and pulcher, Silene gallica, Silybum marianum, Sonchus oleraceus. 50. Weeds noted in Quebrada del Juanango, 9/4 1917. Acaena argentea, Anthoxanthum odoratum, Avena barbata, Bromus unioloides, Dipsacus silvestris, Erigeron canadensis, Geranium core-core, Hordeum murinum, Malva parviflora, Medi- cago denticulata, Physalis peruviana, Rumex conglomeratus and crispus, Ruta chalepensis, Se- taria imberbis, Silybum marianum, Sonchus oleraceus, Verbena litoralis. 51. Weed flora on the talus strip in Bahia del Padre, 15/1 1917. Amblyopappus pusillus, Anthoxanthum odoratum, Apium laciniatum, Avena barbata, Bras- sica rapa and napus, Centaurea melitensis, Chenopodium multifidum and murale, Erigeron cana- densis and crispus (linifolius), Hordeum murinum, Matthiola incana (Pl. 101: 2), Medicago den- ticulata, Melilotus indicus, Raphanus sativus, Rumex crispus, Silybum marianum, Sonchus ole- raceus. The sand-field on El Puente (see above p. 894) had islands of vegetation on the margins, abundant Chenopodium multifidum and scattered Avena barbata-patches, Brassica rapa, Hor- deum murinum, Matthiola incana, Melilotus indicus, Silybum marianum and Sonchus oleraceus. The Vegetation of the sea-shore. Most of the north coast of Masatierra, and practically the entire south coast, is bordered by a vertical abrasion cliff, with little or no talus at the foot, washed by a never-resting surf; around the bays the beach is shingles, boulders or gravel. The nature of the rock is little favourable for the establishment of plant commun- ities; it disintegrates readily under the force of the breakers and as a rule lacks THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 923 a lichen cover, with the exception of the hard and very resistant basalt dykes which are so conspicuous here. Unfortunately landing is difficult, and very often quite impossible where the barranca plunges abruptly into the sea, and there are long reaches of this kind of coast. In little caves washed by the spray a few ferns have found refuge, among them the halophytic Asplenium obliquum (P\. 77:2). In cases it is accompanied by Adiantum chilense and Blechnum auriculatum, and, where freshwater oozes out, Lodelia alata. Where there is a strip of beach, a fringe of Salzcornia may be seen. 52. (13.) Lobelia alata-colony. Landing place in Puerto Ingles, a small cove, separated from the bay, where the surf generally is too strong, by a projecting spur of rock, forming the roof of the so-called Robinson’s cave. Steep, slightly overhanging cliff, where water trickles down. 20/1 1917. Pl. 1or:t. Ikiz Asplenium obliquum var. 1, fissures Rea pruinata 1 Spergularia confertifiora 1 Polypogon crinitus I ; I Rumex crispus I Bryum fernandezianum, carpets Il: 1 Lobelia alata v. rupincola 4 greg. The rock above is covered with lichens: Lecanora saxicola, Parmelia caperata and Caloplaca elegans var. australis. 53. (11 B.) Lichen communities on exposed coast rocks, north headland of Bahia del Padre, 15/1 1917. Lowest exposed strata little resistant to erosion, surf heavy and surge zone without vegetation, but the storm zone has a black coating, perhaps a Verrucarza, of which I was unable to get samples on my only visit to this place. On the harder basalt dyke, the spray zone, well developed above 31/. to 4 m above high-water level, has a cover of crustaceous lichens; the single fruticulose species, Ramalina linearis, did not, as far as I remember, occupy a definite zone. The list does not at all pretend to be complete. Blastenia fernandeziana cop. Gyalecta jenensts Buellia halophiloides Lecanora atra » siphoniatula > dispersa » stellulata cop. 2 Ingae > spp. Microglaena fernandeziana Caloplaca elegans v. australis cop. Parmelia conspersa var. > Selkirkit Pertusaria melanospora cop. Dirina limitata Verrucaria microspora 54. Salicornia peruviana-Sociation. Bahia del Padre, at the landing-place, forming a pure salt-meadow, 15/1 1917. Pl. 102. II: 2 Salicornia peruviana 5 A rill crosses the Salicornietum. Sczrpus cernuus forms a fringe near the out- let. Zetragonia expansa was seen in Bahia del Padre by Johow. 924 C. SKOTTSBERG On the slope leading up to the Puente (PI. 102: 1) are scattered individuals of Apium fernandezianum (type locality), Plantago truncata ssp. Skottsbergi and Spergularia confertifiora. The rock, which easily disintegrates, is almost free from lichens, but Pxellia halophila, Caloplaca elegans v. australis, C. Selkirkit and Lecanora conspersa were collected on harder surfaces. The Vegetation of Santa Clara. This barren island, where no running water is found, belongs to the same climatic district as the lower west section of Masatierra but enjoys (Johow 2 p. 261) frequent rains during the winter months giving birth to a cover of herbs and grasses which support a herd of goats. In summer the island looks like a desert. Of the native flora little or nothing is left. The coast is bordered by a high cliff and accessible in a single place on the leeward side, where a small detached rock or islet, El Morro del Spartan (also called M. de los Alelies; aleli= Matthiola incana, sown here by fishermen), offers some shelter and permits landing on either side of the channel in good weather. The interior of Santa Clara is a sandy, dissected table-land averaging perhaps 100 m in altitude but the highest hill is 375 m according to Johow’s map. -At the time of our visit most of the annuals were gone, and some may have escaped notice. 55. (14 C.) Avena barbata-Sociation. Table-land of Santa Clara, opposite Morro del Spartan, 26/1 1917. M3 Rumex pulcher Rea pruniata 1 rare Salicornia peruviana Sonchus oleraceus Il:2 Silene gallica Avena barbata abundant Centaurea melitensis Ik: 1 Chenopodium murale Amblyopappus pusillus Erigeron canadensis Malva parviflora Hordeum murinum Medicago denticulata Matthiola incana Plantago truncata ssp. Skottsbergit Melilotus indicus Spergularia confertifiora Rumex crispus Wahlenbergia Berterot rare In a narrow ravine in the coast cliff with some moisture and inaccessible to the goats grew a few specimens of Dendroseris litoralis and Rea pruinata. It must have been in a place of this kind Johow saw Adzantum chilense, Asplenium obliquum and Blechnum auriculatum. Morro del Spartan consists of the same soft volcanic rock as in Bahia del Padre and on Santa Clara and is, as in all these places, cut through by thin vertical dykes of hard, dark basalt; it has become just as invaded by weeds as the island itself, but because the goats are unable to cross the stony channel, where the tidal current is strong, a few native plants, which otherwise would have disap- peared long ago, are still to be seen and give to this isolated spot a rather pe- culiar appearance; the vegetation bears a distant resemblance to pictures of open Campos in Brazil. The soft rock is devoid of lichens. Pl. 103. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 925 56. (14 D.) Dendroseris Litoral?s-Sociation. Morro del Spartan, 26/1 1917. Dry, sandy and naked soil. 1} Matthiola incana Dendroserts litoralis 1-2 Rumex crispus Rea pruinata 1 z pulcher Solanum robinsonianum 1% IT: 2 Sonchus oleraceus Avena barbata abundant Wahlenbergia Berterot \ Centaurea melitensis S : Tier Chenopodium murale Chenopodium Sanctae Clarae 1 Amblyopappus pusillus Erigeron canadensis Spergularia confertifiora 1% Gnaphalium cheiranthifolium Tetragonia expansa 1 Hordeum murinum The hard basalt dykes are covered with lichens. Acarospora xanthophana Microglaena fernandeztana Blastenia fernandeztana Parmelia cetrata Buellia concinna var. » comspersa var. » stellulata Pertusaria melanospora cop. Caloplaca elegans v. australis » Shottsbergit » rubina Cop. The Plant Communities of Masafuera. REMARKS ON THE TOPOGRAPHY. Masafuera rises abruptly out of the depths of the ocean, a solid, approx- imately rectangular block, tilted NE and with the longer axis directed N—S, an elevated tableland bordered by very steep to almost vertical, shore-bounded escarpments, which along the west side rise to over 1000 m, culminating in Los Inocentes and Cordén Atravesado. The coast cliffs on the east side are much lower, but just as steep. The longitudinal ridge follows the western border of the block and is very frequently hidden in thick fog; precipitation is high and the west barranca furrowed by hanging ravines, where a cascade tumbles down after every downpour. On the east side the ridge is bordered by the nar- row Plano de la Mona, where the streams which have eroded the deep valleys opening on the east coast have their sources. As the south half of the island is considerably higher than the northern half, rising to about 1500 m, precip- itation is much heavier here and the eroding forces very powerful; this is where we find a magnificent series of canyons, narrow chasms many hundred metres deep, with vertical walls. The largest two, the Casas and Vacas quebradas, have been cut back almost right through the island; it is just possible to find a passage round the origin of the Casas valley, whereas the knife-edge crest forming the back wall of Vacas is impassable. Between the two valleys lies Cordén del Bar- ril, broad and gently sloping toward the sea in its lower part, upwards narrowing to a thin ridge with an abyss to 900 m deep on either side. The table-land 926 C. SKOTTSBERG north of Quebrada de las Casas receives less rain because the dominating SE —SW winds discharge most of their moisture over the higher country to the south. It is dissected by a series of parallel valleys not forming canyons, and the ridges separating them are gently sloping, rather wide plains. The coast cliffs are borded by talus and it is possible to make a circuit of the island along the shore from Quebrada Sanchez round the east and south side and north along the west cliff-wall to the end of Playa Larga. At Loberia Vieja are a number of old huts from the time of the convict settlement. At the place called Buque Varado is a small promontory formed by alluvial deposits; it is inaccessible ex- cept perhaps with a very calm sea, an extremely rare event on this side of the island, and even then landing must be hazardous. The talus of Toltén, a shal- low bay at the northern end of the island, frequented by the langust-fishers, is covered with grassland. Unfortunately it is hardly possible to proceed from here either up to the table-land or along the beach. The regular landing place is a rock at the south head-land of Quebrada de las Casas, but one may have to watch for an opportunity. The jetty once built for the settlement was destroyed by the sea. Plates 104:2, 109, 110 and 116:1 and text figs. 38—41 will give an idea of the topography; for further information I refer to Pt. I, no. 4.’ GENERAL DISTRIBUTION OF THE PLANT COMMUNITIES. Contrary to Masatierra, Masafuera reaches a height sufficient to give birth to a climatic timber-line and, above this, to a series of heath-communities lacking in the other island. Besides, we find the same formations as on Masatierra, the evergreen forest and the grass-land, while a scrub comparable to the scrub on Masatierra is very poorly represented. From the reports of the early visitors we can come to no other conclusion than that the island was well wooded as late as in the middle of the 18th century and that the forest went farther down than now, perhaps to near sea-level in some valleys. When and under which circumstances destruction set in we cannot tell, but I think that we are not mistaken if we put the ravage in connection with the search for sandal-wood, of which pieces have been found buried in the soil in many places; names like Quebrada del Sdndalo, Q. del Sandalito and Rodado del Sandalo bear witness of the one-time large distribution of a species of Saxtalum. It must have become extinct before 1830, when Cuming visited the island as the first collector of plants. No scientist ever saw it and we shall never know whether it was identical with S. fernandezianum or belonged to a vicarious species or variety. Unfortunately, none of the few naturalists who visited Masafuera gave a description of the state of the forests. The island remained uninhabited until 1909, when a convict settle- ment with its base near the entrance to Quebrada de las Casas was established, but it was abandoned after a few years. The condition of the forest was much the same in I917 as in 1908, when I visited the island for the first time, al- though the encroach made after 1909 was visible in many places. Some forest had been burnt or cut by the inmates of the penal colony, huts of native tim- * To be published in 1953. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 9 N ~ Fig. 37. Ruined forest in Quebrada de las Chozas, Masafuera. — Photo 3/3 1917. ber had been built in some of the valleys (Pl. 104:1), and in some sites aband- oned potato fields and garden plots, where a new weed flora was well established, were seen. Fig. 37 shows a grove, intact in 1908 but now completely ruined. And once destroyed, the forest does not come back; native ferns, Axthoxanthum and other introduced herbs and grasses rapidly invade the clearings, and tree seedlings may fall a victim to the innumerable goats. Erosion had begun its work in many places. The speed with which this procedure goes on was illu- strated along the zig-zag roads built on the steep walls of the Casas valley from the base station to the dwelling sites of the convicts; parts of the road-track had already been carried off by land-slides. A few years after our visit in 1917 the settlement was occupied a second time, but soon abandoned, and probably more damage was done. The island is now part of the Juan Fernandez National Park and will, I suppose, remain uninhabited. There can be little doubt that ever since the Spanish settlement on Masatierra was founded, people from there have gone to Masafuera just as the fishermen still do, roaming the ridges and table-land to look for goats and setting fire to the grass. Maqui was seen by us in some of the valleys, but a macal like those in Masatierra was observed only in Quebrada del Blindado, where there was a thick grove at about 400 m above sea-level: here the luma forest (IZyrceugenia Schulzez) begins. In the deepest valleys, Casas, Vacas and Varadero, scattered trees appear a 928 C. SKOTTSBERG Fig. 38. Lower slopes of Cordon del Barril, Masafuera. Anthoranihum, Lophosoria etc.; lowest altitude for Wyrceugenia Schulzez. — Photo K. Backstrém, Feb. 1917. few hundred m from the sea in an altitude of less than 150 m above sea-level, but on the almost perpendicular walls there is no room for a forest. On the slopes of the upper, V-shaped parts of the valleys are groves of luma. On the gently sloping table-land between the Casas and Vacas valleys are three small gullies not eroded down to the shore, each with a patch of forest, and on the plain be- tween the Cabreros and Blindado gullies called El Papal lumas are scattered in the grass-land at 300 m (Fig. 38). In Blindado the first trees were found at about 330 m, 50m higher up were small patches of forest, and above 425 ma narrow fringe of good forest follows the stream to 700 m, where it breaks up into smaller groves, the last observed in 825 m. In all the valleys north of Casas conditions are about the same. In the Mono and Chozas valleys the first lumas appear at 200m, groves of some extension at 350 and good forest of luma and naranjillo (Fagara externa) at 450 m (Pls. 106:1, 107), extending to 715-740 m (Fig. 39). From c. 550 m canelos (Drimys) appear. The largest forest still in- tact covers the ridge between Quebrada Sanchez and Q. Larga and the north slope of this valley almost to the top of the barranca back of Toltén bay (PI. 108:1). Turning to the west side of the island, strips of forest are seen in the hanging gulches, all apparently inaccessible without special precautions and ar- rangements, with exception of a gorge at the north end of Playa Larga, where the forest descends to about 250 m and solitary trees 50 m lower down (St. 57). On the steep south slope of the island, where the ridges radiating from Los Inocentes come down, patches of forest are seen in the gullies above Vicente Porras and Rodado del Sandalo. The position of the timber-line varies, but as a rule the last forest patches are met with between 700 and 750m, even if scattered groups of lumas and THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS Q29 “5S Sa Fig. 39. Timber-line in Quebrada del Mono, Masafuera, looking East. Almost pure Anthoxranthum Soc.; in the centre a patch of Lofhosoria. Left on slope a few Dicksoniae. — Photo K. Back- strom, Feb. 1917. canelos are found higher up, and this is particularly the case with the latter. Single canelos occur in the Dicksonia-Lophosoria belt up to 1000 or 1100 m, and one solitary tree, about 5 m tall, was found on the bank of a small ravine at 1130m. The record is held by the peralillo (Coprosma pyrifolia), of which a 50 cm tall male plant was found at 1200 m above sea-level on the Plano de la Mona above Buque Varado. It does not seem possible to fix a line between a lower and an upper montane belt on this island, because the trees and shrubs characterizing the upper woods on Masatierra are absent from Masafuera, the flora of which is poor in comparison. Rea micrantha is, as it were, replaced by Hesperoseris, but this is much less common. The only arboreous species serving as an indicator of an upper, more humid forest belt is Dicksonia externa. Conditions are, however, different in the north and south part of the island; only in the latter a closed, extensive Dicksonia forest with scattered trees (only Drzmys, as far as I have seen) is developed, and here it extends up the slopes of Los Inocentes to at least 1200 m. North of Casas we find Dzcksonza in all the wet, wooded gulches, as a rule not under 550m, but occasionally down to 440, as for instance in Quebrada del Mono. A very thick stand is seen on PI. 108:2. On Masafuera we find almost nothing that corresponds to the high-montane brushwood which is such a characteristic feature in the scenery on Masatierra and 930 C. SKOTTSBERG Fig. 40. Looking down the shallow gully of Quebrada del Pasto on Plano de la Mona, Masa- fuera, from a place near the Correspondenia camp, c. 1150 m. Anthoxranthum in foreground, behind the grass a thick bed of Lofhosorta, which forms patches on the slopes. — Photo 5/3 1917. the home of the numerous rosette-trees, the “Aobzusonia assemblage’. Geology and topography are different, and above all, Masafuera has only four species with this growth-form, //esperoserzs, already mentioned, Dendroseris macrophylla, Phoentcoserts regia and Robinsonia Masafuerae. Dendroseris grows on the walls of the deep canyons down to 200m above sea-level, and also in the ravines on the table-land, Phoentcoserzs and Robinsonia in the fern-beds along the ridges from 650 to 1350 m, always scattered, generally rare and badly damaged by the goats wherever accessible. Nor is the scrub, one of the most important communities on Masatierra, prop- erly developed; the ridges are broader and grass-covered, the shrubs scattered, and only two of the leading species are found on Masafuera, Pernettya rigida and Blechnum cycadifolium. The former may be met with as far down as a little below 500 m, and it increases as we approach the table-land, but nowhere does it form thickets at all comparable to the Masatierran scrub. Above the timber-line a peculiar plant community puts its stamp on the landscape, Johow’s “‘Estepa de helechos” or Fern-Steppe. It has nothing in com- mon with a steppe in the current sense of this term, but it cannot be denied that, physiognomically, it has some resemblance to a low savanna, a kind of fern savanna, with scattered Décksonda and an undergrowth of lower ferns, par- THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 93! Fig. 41. Looking down Quebrada del Pastro, Masafuera, from its source toward the sea. Alpine heath with abundant Lofhosorta. — Photo 53 1917. ticularly Lophosoria, patches of Gunnera and abundant Axthoxanthum odoraium, see Pl. 105: 1. Physiologically, it is of course no savanna at all, but a heath, in- vaded by Axthoxanthum. At about 1100 m Dicksonza falls back, but small groups and single specimens of low stature are found also higher up. This is what I have distinguished as a Subalpine belt, and it is well developed in the south section of the table-land, but north of Quebrada del Ovalo little Décksonia is found. The Lophosoria-Anthoxanthum-Gunnera community is a complex and should perhaps be regarded as a mosaic of three sociations, each with its own dominant, with different demands on soil moisture, Axthoxanthum forming pure carpets on the drier soil and Gunzzera preferring more humid depression or places where water trickles down (Figs. 40, 41). It forms pure societies on wet canyon walls and on the high western barranca (Pl. 109:2). In the high country east and south of Los Inocentes, clouded over most of the time and supposed to receive more rain than the country north of the Barril ridge, the closed Dzcksonza forest ascends high up; seen from the summit it seems to form an unbroken cover over the ridges and gulches, leaving little room for a ‘savanna’ of the kind described. The transition from a Subalpine to an Alpine region is almost imperceptible. Some Alpine species are occasionally found at lower altitudes, but the majority make their appearance in about 1100 m, a level selected as a suitable lower limit of an Alpine belt, but it must not be forgotten that the Lophosoria-Anthox- anthum-Gunnera complex still reigns supreme, and that the Alpine species do 932 C. SKOTTSBERG Fig. 42. Stony Alpine heath near Cerro Correspondencia, Masafuera, c. 1350 m above sea-level. — Photo 14/2 1917. not make themselves conspicuous before we reach the stony slopes and rocky crags higher up, approximately above 1300 m where, in addition to the Magell- anian element, a number of noteworthy endemics make their appearance, with bryophytes and lichens forming closed carpets (Pls. 114, 115, text fig. 42.) Grass-land occupies much larger areas on Masafuera than any other kind of vegetation, covering the entire basal region and ascending across the dissected table-land to its western border. On the talus deposits at the foot of the cliffs along the east side, on the valley slopes of the lower montane belt and the plains between the canyons, the grass-land has same flora as the native grass-land of Masatierra, but some of the species (Piptochaetium, Danthonia etc.) do not occur on Masafuera. Of foreign grasses Anthoxanthum is by far the most abund- ant, taking the part of Avena barbata, which is of little importance. Of other weeds Acaena was only seen in two places, but it may have spread since, and most other adventitious species, many dating from the time of the settlement, are more or less confined to the vicinity of the abandoned dwelling sites. The abundance of Anxthoxanthum on the highland was mentioned above; Rumex ace- tosella comes second. It is hard to imagine what the upland country looked like before the arrival of these intruders. If we could exterminate them, the result would be extensive desert-likes areas, and I do not think that these were ever covered by either Lophosoria or Gunnera, because I fail to see how the weeds ever managed to germinate in the deep shade of the fern fronds or pangue lea- ves, much less being able to crowd them out unless the country was burnt over. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 933 Possibly this is what happened. I should add that the goats do not seem to touch either the ferns or the pangue leaves. Perhaps some of the cespitose na- tive grasses and sedges and mat-forming herbs were more widely spread; other species, more palatable to the goats, may have become extinct. Phoenicoseris regia, Robinsonia Masafuerae and Ranunculus caprarum are on the verge of ex- tinction. The vegetation along the streams and cascades is characterized by the ab- sence of the leading species in Masatierra, Cladium scirpoideum and Cardamine flaccida. Gunnera peltata is replaced by G. Masafuerae and Cladium, in a way, by Blechnum longicauda. Mimulus glabratus, not found on Masatierra, has to be added. A few hydrophilous bryophytes are common. This vegetation is at its best in the deep canyons from Quebrada de las Casas south to Q. del Varadero. These magnificent gorges constitute one of the most interesting feat- ures of Masafuera. In the U-shaped part trees and shrubs have colonized the ledges and ravines on the vertical walls; grasses, ferns and other herbs of the forest floor cover the rock faces. I have not tried to co-ordinate these sociation fragments with the forest communities but prefer to treat the vegetation of the canyons under a separate heading. THE STATIONS. The Forest region. The luma-forest of Masafuera differs from the Masatierran forest in some respects. Vothomyrcia is replaced by Myrceugenia Schulzsez, which has a lighter colour, Fagara externa does not attain the large dimensions of /’. mayu, Sophora masafuerana was not seen in the forest and is much less common. Cofrosma pyrifolia, Drimys and Rhaphithamnus are shared by both types. As the species often forming a dense undergrowth on Masatierra are absent and, with one excep- tion (Berberis corymbosa), not represented by vicarious species (2. masafuerana is rare and was not observed in the forest), the Masafueran forest is less dense and very monotonous. Ferns and other cryptogams are abundant, but several of the conspicuous species are wanting. The delicate Dysopszs, quite important in the dark, humid woods of Masatierra, is absent. The scarcity of rosette trees was already pointed out. With increasing altitude the epiphytes and the filmy ferns increase in number, and Weymouthia makes its appearance. The lower montane forest. 57. (20.) Myrceugenta-Coprosma pyrifolia-Sociation. Quebrada de la Loberia Vieja, c. 280 m, 17/2 1917. Thin, very stony soil mixed with twigs and dried leaves. IV: 2-1 Coprosma pyrifolia 2 Myrceugenia Schulzet 4 Fagara externa I Ill (rarely over 2 m) Il: 3 Myrceugenia 4 Coprosma 4 Gunnera Masafuerae 1 934 in 1917 I]: 3-2 Pteris berteroana 3 ie Uncinia Douglasit 2 Bromus fernandezianus 1 Carex berteroniana 1 Gunnera I Libertia formosa 1 Rumex pulcher I Sonchus oleraceus Uncinia costata 1 Il: 1 moss-cushions and around boulders Adiantum chilense \ Asplenium dareoides 1 Blechnum auriculatum 1 > valdiviense 1 Hymenophyllum caudiculatum 1 » cuneatum 1 » plicatum \ ] Fissidens fernandezianus Rigodium toxarium Chiloscyphus integrifolius Lunularia cruciata Plagtochila fuscobrunnea 58. (21.) Myrceugenia-Fagara externa-Sociation. Quebrada del Blindado, 19/2 FASS. 1 SOL aS iN Ot. .55,7.. V3.2 Myrceugenia Schulzet 4 IV: 1 Coprosma pyrifolia 3 Myrceugenia 2 Dicksonia externa 1 Fagara externa \ Rhaphithamnus venustus 1 Ill fagara 3 Lophosoria quadripinnata 3 Myrceugenia 3 Coprosma 2 Dicksonia Flesperoseris gigantea 1 Nes Myrceugenia 2 Carex berteroniana 1 Gunnera Masafuerae \ Halorrhagis asperrima \ SKOTTSBERG Leptogium moluccanum Pertusaria hadrocarpa Pseudocyphellaria fragillima Epiphytes, mainly on Myrceugenia Polypodium intermedium ssp. masafue- ranum Rhaphidostegium caespitosum Rigodium toxarium Zygodon sp. Chiloscyphus integrifolius Frullania chilensis cop. » lobulata Radula decora » microloba Strepsilejeunea macroloba Arthopyrenia adnexa Bacidia subluteola Caloplaca subcerina Catillaria melastegia Leplogium cyanescens » moluccanum Lobaria crenulata Parmelia cetrata Parmeliella nigrocincta Pseudocyphellaria fragillima Thelotrema lepadinum cop. Histiopteris incisa 1 Hypolepis rugosula 1 Pteris berteroana 1 II 2 Polystichum vestitum 2 Uncinia Douglasii 2 Gunnera I Blechnum longicauda 1 Dryopteris inaequalifolia x Polystichum berterianum 1 Pterts semiadnata 1 rtica fernandeziana 1 Diss Asplenium dareoides 2 greg. Blechnum auriculatum Hymenophyllum plicatum 1 I Amphidium cyathicarpum Bartramia aristata Eriopus leptoloma Fissidens maschalanthus TOL7, 427 5snne Plsvie86:A1; TOP: THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS Porothamnium fasciculatum Rhacopilum assimile Rhaphidostegium aberrans Hymenophytum flabellatum Lophocolea attenuata > pallidevirens » submuricata Marchantia berteroana Metzgeria decipiens » multiformis Mylia sp. Plagiochila fuscobrunnea Symphyogyna Hochstetteri Epiphytes On Myrceugenia Polypodium intermedium ssp. masafue- ranum vatrr. Hymenophyllum spp. Neckera rotundata Lamprophyllum splendidissimum Rhynchostegium complanum Rigodium toxartum Metzgeria decipiens Parmelia cetrarioides » nilgherrensts Pertusaria letoplaca var. On Hesperoseris Bacidia endoleuca Caloplaca clandestina Catillaria melastegia Lecanora albellina var. On Dicksonia Asplenium dareoides Hymenophyllum cuneatum » Jalklandicum » plicatum » rugosum » tortuosum Trichomanes exsectum Rhaphidostegium aberrans Lophocolea attenuata Plagiochila fuscobrunnea Trichocolea verticillata 935 59. (23.) Myrceugenia-Fagara externa-Sociation. Quebrada del Mono, 20/2 Soil as before. IV:2 Myrceugenia Schulzet 4 fagara externa 1 IV: 1 Myrceugenia 4 Fagara 2 Coprosma pyrifolia Ill Myrceugenia 3 Coprosma 1 Dicksonia externa \ Drimys confertifolia 1 Fagara 1 Rhaphithamnus venustus 1 II: 3 Dryopteris inaequalifolia 4 Myrceugenia 2 Histiopteris incisa 1 Urtica fernandeziana 1 J WE. Polystichum vestitum 2 Peperomia fernandeziana 1 Uncinia Douglasii 1 Wes Asplenium dareoides 3 greg. Blechnum valdiviense 3 Hymenophyllum caudiculatum 1 > plicatum 1 Trichomanes exsectum 1 I Hypnodendron microstictum Porothamnium fasciculatum Rigodium arborescens » toxarium Madotheca chilensis Flagiochila riparia Saccogyna sguarristipula Schistochila berteroana Epiphytes, mainly on Myrceugenia Polypodium intermedium ssp. masa- fueranum vat. Polypodium lanceolatum Hymenophyllum plicatum Trichomanes exsectum Lepyrodon parvulus Macromitrium Masafuerae 936 C. SKOTTSBERG Rhacopilum assimtile Pseudocyphellaria aurata Rhaphidostegium aberrans Thelotrema lepadinum Rigodium toxarium festoons Usnea dasypogotdes Metzgeria decipiens Epiphyllous on Aymenophyllum caudiculatum Microphiale lutea Parmelia soredica Lejeunea reticulata On Cordon del Barril, where solitary trees of Drzmys and Coprosma pyrifolia ascend to at least 1ooo m, the following epiphytes were collected on Drzmys., Eucamptodon perichaetialis Nephroma antarcticum Leptostomum Menziestt Parmelia pertusa Rhizogonium mntoides » ptlosella var. Ulota fernandeziana » revoluta Zygodon pentastichum Pseudocyphellaria chloroleuca » nitida Lophocolea rotundifolia Marsupidium piliferum Metzgeria violacea , Psoroma sphinctrinum 60. (24.) Myrceugenia-Fagara externa-Sociation. The plain between Quebrada Sanchez and Q. Larga, 25/2 1914, 515 m. Dry, granular vegetable mould with few stones but covered with dead leaves. Cryptogams on stone-blocks, pieces of bark and decaying logs. Pl. 108:1. DVe2 I Myrceugenia Schulset 5 Campylopus subareodictyon IV Rhacopilum fernandezianum fest ; : Rhaphidostegium aberrans Myrceugenia i Rigodium artorescens Ill Chiloscyphus integrifolius Myrceugenia 5 Lophocolea submuricata Fagara externa 2 Lunularia cruciata “7 5 > mye, Coprosma pyrifolia I Radula decora Rhaphithamnus venustus 1 2 Dusenit II: 3 Epiphytes, mainly on J/yrceugenia Myrceugenia 3 Polypodium intermedium ssp. masa- Pteris berteroana 1 Jueranum vatr. Solanum masafueranum 1 Polypodium lanceolatum Il: 2 Leptodon Smiuthit Lepyrodon parvulus Macromitrium sp. Weymouthia mollis festoons Blechnum auriculatum 2 Myrceugenia 2 Adiantum chilense 1 Pteris chilensis 1 Lopholejeunea spinosa Uncinia Douglasii 1 Radula decora It Bacidia endoleuca : ‘ Catillaria leucochlora Myrceugenia 2 : : Pertusaria polycarpa var. Pseudocyphellaria cinnamomea » chloroleuca Ramalina usnea Asplenium dareoides 1 greg. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 937 Upper montane (Dicksonia) forest. 61. (26 C.) Myrceugenia-Dicksonia externa-Sociation. Quebrada de las Chozas 3/3 1917, c. 700m. Ravine with dry, stony streambed. Analysis incomplete. Wen MIB Myrceugenta Schulzet Asplenium dareotdes Drimys confertifolia » macrosorum 1 » stellatum Blechnum valdiviense Dicksonia externa 5 I I: [3 On stones Dryopteris inaequalifolia Lophosoria quadripinnata Blechnum longicauda Gunnera Masafuerae Epiphytes on Dzcksonia Urtica fernandeziana Porothamnium fasciculatum Rigodium arborescens Hymenophyllum ferrugineum 2 Pseudocyphellaria chloroleuca Peperomia fernandeziana 62. (22.) Dicksonia externa-Sociation. Quebrada del Mono 20/2 1917, c. 550 m. Damp and dark, narrow ravine with stony streambed. Cover of dead fern fronds. Moss-carpet on stones. Pl. 108: 2. Ill Hymenophytum flabellatum Dicksonia externa 5, 4—5 m tall Lepidozia sejuncta Gunnera Masafuerae 1 Lophocolea attenuata » submuricata II: 3 Marchantia foliacea Blechnum longicauda 2 Pallavicinia xtphoides greg. Polystichum vestitum 2 Solenostoma crassulum Il:2 Cladonia fimbriata var. Blechnum chilense 1 Epiphytes » longicauda \ H IL ; Polystichum vestitum 1 VAG LIE SEALE » falklandicum Il: 1 Polystichum berterianum shade form Blechnum valdiviense 2, in deepest shade Asplenium dareoides 1 Polystichum vestitum 1 Campylopus sp. lsopterygium sp. Rhaphidostegium aberrans I Mylia cf. fuscovirens greg. Fisstdens maschalanthus Plagiochila fuscobrunnea Anthoceros Skottsbergti greg. Chiloscyphus integrifolius 63. (23 B.) Dicksonia externa-Sociation. South-east slope of Los Inocentes 22/2 1917, c.950 m. Soil covered with decaying trunks and fronds of Dzcksonza. INS 3 Il Drimys confertifolia, one tree Dicksonia externa 5 62 — 516796. The Nat. Hist. of Juan Fernandez and Easter Isl. Vol. II. 938 C. SKOTTSBERG Il: 1 Epiphytes on Dicksonia Asplenium macrosorum 1 rare Hymenophyllum pectinatum I » plicatum 2 rugosSuUmM Fissidens maschalanthus : : Serpyllopsis caespitosa v. fernandeziana Rhaphidostegium aberrans pyllop Y4 Se fernandestana “piphytes on Drimys 5 ; , Dein J Riccardia breviramosa Dicranoloma capillifolium Macromitrium saxatile cop. Epiphyllous Hefaticae on Hymenophyllum spp. Zygodon pentastichum Brachiolejeunea spruceana Rigodium toxartum festoons Cololejeunea Shottsbergii Weymouthia mollis » Caliph Fulbose Bazszania cerina Harpalejeunea oxyota Lejeunea reticulata Cora pavonia - ; : : is Siphonolejeunea nudicalycina Nephroma antarcticum Parmelia revoluta Sphaerophorus melanocarpus Usnea subtorulosa 64. (23 C.) Dicksonia externa-Blechnum cycadifolum-Sociation. A lower, less dense stand near St. 63, with burnt patches; soil a thick, spongy fern-peat, under- mined by the burrows of the Masafueran fardela (Oestrelata externa Salv.). Anal- ysis incomplete. Il Phoenicoserts regia \ (much damaged) Dicksonia externa 4 Robinsonia Masafuerae 1% Thyrsopteris elegans 1 II: 3 Urtica fernandeziana burnt place Blechnum cycadtfolium 2 Dryopteris tnaegualifolia i I Histiopterts incisa 1 Marchantia polymorpha, covering Lophosoria quadripinnata \ burnt places ihe Vegetation of the deep can yoms: Our survey of the interior of the canyons was concentrated on Quebrada de las Casas, together with ©. de las Vacas the finest example in the island of this valley type. The streambed consists of round, polished stones and large boulders; it fills the valley floor, which in the narrowest places does not exceed 10 or I5 m in width, whereas the vertical walls are several hundred m high (Pls. 109:2, 110). The rock at the foot of the cliff may offer a passage, now on one side, now on the other, or the streambed itself is the only practicable road. After a spell of dry weather only pools of crystal clear water are left, but after a downpour in the mountains the scenery rapidly changes, the stream overflows and the canyon is closed. The air is humid and the light very subdued, the morning sun illuminating the depths of the gorge for a short while only. Along the stream Guznera luxuriates as nowhere else (Pl. 111). The vegetation on the walls is a mixture of trees and shrubs, clinging to the shelves and crevices, grasses and herbs, particularly ferns, and wet carpets of bryophytes in various shades of THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 939 brown and green. The outer, dry and grass-covered slopes not far from the entrance are not considered here (PI. 109: 1). 65. (40 C.) »Cliff forest» on the south wall of the Casas canyon, about 1 km from the entrance, 11/2 1914. A detailed analysis of the nearly vertical wall could not be attempted; species forming conspicuous patches are marked greg. Ill Fragaria chiloensis Myrceugenia Schulzet 2 Libertia formosa greg. Coprosma pyrifolia 1 Peperomia Skottsbergit Dendroseris macrophylla i Polystichum vestitum Hesferoseris gigantea 1 Sonchus oleraceus Sophora masafuerana 1% Uncinia Douglas Wahlenbergia Masafuerae Is 3 Gunnera Masafuerae greg. ues Lophosoria quadripinnata Adiantum chilense Pterts berteroana Asplentum dareoides » obliquum v. chondrophyllum II: 2 Blechnum valdiviense Blechnum chilense Flymenoglossum cruentum > longicauda Hymenophyllum plicatum > Schott greg. Lobelia alata v. rupincola Bromus fernandezianus greg. Nertera granadensts var. Carex berteroniana greg. Oxalis corniculata Cystopterts fragilis f:a, rare Stellaria media Evrigeron fruticosus Bryophytes and lichens not recorded. 66. (40 D.) Same locality and date; under overhanging rocks. 3 Peperomia fernandeziana Dryopteris inaequalifolia 2 Skottsbergit II: 2 2-1 I Adiantum chilense Trichomanes exsectum greg. Asplenium obliquum v. chondrophyllum Anthoceros Skottsbergii greg. » stellatum Metzgeria multiformis greg. Cystopteris fragilis var. Pallavicinia xiphoides greg. 67. (28.) »Cliff forest» on north wall of Quebrada de las Casas, 6/3 1917. Relatively dry locality with few and narrow ledges. Ver Erigeron fruticosus 1 Myrceugenia Schulzei 1 Halorrhagis asperrima 1 Sophora fernandeziana 1 Libertia formosa 1 greg. Peperomia berteroana 1 Ill » Skottsbergii 1 Coprosma pyrifolia 1 Polystichum berterianum 1 greg. Dendroseris macrophylla 1% » vestitum 1 I:3 Uncinia Douglasiz greg. Dryopteris tnaequalifolia 1 ey Asplenium dareoides 1 greg. Il: 2 » Stellatum 1 Blechnum Schottii 1 greg. Blechnum auriculatum 1 Bromus fernandeztanus 1 greg. » . wvaldiviense 1 greg. 940 C. SKOTTSBERG Parictaria debilis v. fernandeziana \ Fissidens maschalanthus Wahlenbergia Masafuerae 1 Pinnatella macrosticta Rhacopilum fernandezianum I Trichomanes exsectum Lunularia cruciata 68. (29). Gunnera Masafuerae-Blechnum longicauda-Sociation. Slightly over- hanging cliff at the first waterfall in Quebrada de las Casas, c. 200 m above sea-level, 6/3 1917. This community covers considerable areas of the canyon walls. It forms a fine tapestry. List of cryptogams incomplete. Ill Isr Coprosma pyrifolia 1 Asplenium dareoides 3 Blechnum valdiviense 3 Cystopteris fragilis var. 1 Gunnera Masafuerae 2 Mimulus glabratus v. externus Pteris berteroana 1 II: 3 I Il: 2 a na Trichomanes exsectum 4 Blechnum longicauda 2 ices Fissidens rigidulus » chilense 1 Py: ‘ hah 9 Dryopteris inaequalifolia 1 orothamnium fasciculatum Peperomia Skottsbergit 1 Monoclea Forsteri Lophocolea dtvergenticiliata 69. (30.) Gunnera Masafuerae-Blechnum longicauda-Sociation. Quebrada de las Casas, wet cliff wall at the first waterfall, 215 m above sea-level and c. 2 km from entrance, 11/2 and 6/3 1917. The fall is framed by this community. Il: 3 I Gunnera Masafuerae 2 Hymenophyllum falklandicum 1 Coprosma pyrifolia 1 i oa aie. Porothamnium fasciculatum greg. Dryopteris tnaequalifolia i Lamprophyllum splendidissimum Il: 2 Anthoceros Skottsbergit Blechnum longicauda 4 Balantiopsis cancellata Bromus fernandezianus 1 » chilensts greg. Carex berteroniana 1 Lophocolea textilis Uncinia Douglasii 1 Megaceros fuegtensis Metzgeria multiformis greg. Monoclea Forsteri Pallavicinia xiphoides greg. Slechnum valdiviense 1 greg. Plagiochila chiloensis II: 1 Asplenium dareoides 2 Lobelia alata v. rupincola 1 » hyadesiana greg Mimulus glabratus v. externus 1 » riparia Parietaria debilis v. fernandeziana 1 » robusta Schistochtila berteroana 70. (30 B.) Community of Bryophytes. Same locality and dates as St. 69, but exposed to the current. The canyon is blocked by a threshold a few metres high, over which the water plunges; at the foot are huge boulders which, together with the ledges, are more or less covered with mosses and Hepaticae. After a period of dry weather there is very little water in the stream, but after heavy rains the banks and boulders are washed by the torrent. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 94! A. On the banks and boulders. Dendroligotrichum dendroides greg. Hypopterygium Thoutnt thaphidostegium caespitosoides Rhizogonium mntioides Hymenophytum flabellatum Megaceros fuegiensts Metzgeria multiformis greg. Pallavicinia xiphoides greg. Plagtochila chiloensts Schistochila berteroana Peltigera polydactyla Under the overhanging side of the boulders Hymenophyllum caudiculatum Hymenophyllum plicatum On a fallen trunk, stuck between boulders, a cover of Aiccardia adglutinata. B. On the vertical face of the threshold. Marginal zone, periodically dry Cyptodon crassinervis Porothamnium fasciculatum Lecanora masafuerana Central zone, nearly always washed Oncophorus fuegianus Sciaromium pachyloma greg. 71. (30 C.) “Cliff forest’ on the canyon wall of Quebrada Angosta, c. 1 km from the entrance and a little below the first threshold, 12/3 1917. This is the narrowest of the deep canyons, only about 5 m wide. Walls vertical, streambed occupying the entire width of the floor. In dry weather only pools left here and there. Degree of covering not estimated and no bryophytes or lichens collected. HiIl: 3 Coprosma pyrifolia Dendroseris macrophylla Myrceugenia Schulzet Sophora masafuerana NBs) Dryopteris inaequalifolia Gunnera Masafuerae Nicotiana cordifolia l52 Agrostis stolonifera Blechnum chilense » longicauda Bromus fernandezianus Erigeron fruticosus Halorrhagis asperrima Hypolepis rugosula Libertia formosa Peperomia berteroana Polypogon crinitus Polystichum vestitum Pteris berteroana Il: 1 Adiantum chilense Asplenium dareoides Blechnum auriculatum » valdiviense Erigeron rupicola Mimulus glabratus v. externus Parietaria debilis v. fernandeziana Petroselinum crispum Wahlenbergia Masafuerae 72. (20 B.) Quebrada de la Loberia Vieja. Very high, vertical waterfall in the bottom of the gorge, c. 190 m above sea-level, 17/2 1917. There was no water at the time of our visit. A. Gunnera Masafuerae-Blechnum longicauda-Sociation. On the banks, sprayed when the jet of water comes down. Ing I2 Gunnera Masafuerae 1 Blechnum longicauda, a \arge patch Histiopteris incisa 1 942 C. SKOTTSBERG I Anthoceros Skottsbergit Metzgeria decrescens greg. Riccardia nudimitra Symphyogyna hymenophyllum greg. Peperomia berteroana 1 Polypogon crinitus I Verbena litoralis I Men Mimulus glabratus v. externus 2 Erigeron rupicola 1 B. Sociation of hygrophilous Bryophytes on the perpendicular wall, washed by the cascade. Pterygophyllum obscurum Fissidens rigidulus Sciaromium pachyloma Philonotis vagans var. » krauseana 1 DEASS 73. (40 E.) Calystegia tuguriorum-Sociation. North branch of Quebrada de las Vacas, wet ‘barranca not far from entrance, 10/2 + 13/2 1917, Calystegia forming a pure society at the foot of the cliff (Pl. 77:1). Width of canyon about IO m. Tz3 cor Cystopteris fragilis var. 1 Lobelia alata v. rupincola \ Mimulus glabratus v. externus 1 Dendroseris macrophylla 1 Nicotiana cordifolia 1 iis Calystegia tuguriorum 5 Bromus fernandeztanus 1 Hypolepis rugosula 1 Peperomia berteroana 1 > fernandesziana 1 Pteris berteroana \ I Bryum fernandezianum Eustichia Poeppigi Gymnostomum calcareum Lophocolea divergenticiliata Plagiochasma rupestre The Vegetation above the timber-line. I pointed out above that there is no continuous timber-line on Masafuera; the wooded belt consists of separate stands filling the valleys between approx- imately 300 and 700 m, in cases beginning lower down and extending higher up. Along the intervening ridges the basal grass-land stands in open communication with the Subalpine table-land. The Subalpine fern-grass heath. 74. (40 F.) Dicksonia externa-Lophosoria-Gunnera Masafuerae-Anthoxanthum complex (p. 931). Quebrada del Mono, a little above the uppermost luma-grove, c. 720 m above sea-level, 12/2 1917. Bryophytes and lichens not recorded. Pl. 105. Ill Pernettya rigida 1 Dicksonia externa 1 ll 251 Wis: Lophosoria quadripinnata 3 greg. Gunnera Masafuerae 2 greg. Ilo Anthoxanthum odoratum 5 greg. Rumex acetosella 3 Blechnum auriculatum 2 I Nertera granadensts var. 1 THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 943 75. (25.) Dicksonia externa-Lophosoria-Sociation. Table-land above Quebrada del Pasto, c. 1100 m, 25/2 1917. Open patches with stony soil, where the major- ity of the bryophytes and lichens were collected. HI-Il: 3 Dicksonia externa 1 Phoenicoseris regia 1 rare Robinsonia Masafuerae \ rare 10 Ga} Lophosoria quadripinnata 5 Pernettya rigtda 1 Thyrsopteris elegans \ Mr Blechnum cycadifolium 1 tsi Hymenophyllum cuneatum 1 » falklandicum 1 I Bartramia sp. Campylopus areodictyon Dendroligotrichum dendrotdes cop. Polytrichadelphus magellanicus Rigodium hylocomioides Thysanomitrium Richardit . FHymenophytum flabellatum Jamesoniella colorata Plagiochila fuscobrunnea cop. » homomalla » rectangulata Cladonia gracilis Lecanora polytropa Leptogium Menztestt Peltigera polydactyla Pseudocyphellaria argyracea » endochrysea Stereocaulon paschale v. magellanicum Epiphytes on Dicksonia Ocellularia subdenticulata Pseudocyphellaria chloroleuca Epiphyllous on Dendroligotrichum Strepsilejeunea squarrosula 76. (26.) Pernettya-Anthoxanthum-Sociation. Near St. 75, 25/2 1917. iMNle a Anthoxanthum odoratum 5 greg. Blechnum auriculatum 4 Pernettya rigida 2 Blechnum chilense 1 Gunnera Masafuerae 1 Lophosoria quadripartita 1 Polystichum berterianum 1% WIS it Euphrasia formosissima 1% ] Nertera granadensis var. 1 Campylopus areodictyon Ptychomnium ptychocarpum Thysanomitrium Richardit Riccardia fuegiensis Cladonia aggregata » coccifera » gracilis » pityrea Cora pavonia Enterostigma Skottsbergii cop. on stones Lecanora coarctata Parmelia saxatilis on stones Placopsis cf. fuscidula var. » gelida var. Pseudocyphellaria mougeotiana Stereocaulon paschale v. magellanicum » ramulosum 77. (19.) Lophosoria-Blechnum chilense-Sociation. Correspondencia Camp, on bank of small stream, 1130 m, 15/2 1917. The stream has cut a ravine where water is always found. Pl. 112. IW/2 ai Drimys confertifolia, one tree 3 Lophosoria quadripinnata 4 Blechnum chilense 3 Phoenicoservis regia one spm Robinsonia Masafuerae one spm Pernettya rigida 1 IIE 16 Flymenophyllum cuneatum 1 ‘ 944 I Nertera granadensts var. 1 Polytrichadelphus magellanicus C. SKOTTSBERG On stipes of Lophosoria- Lophocolea muricata 78. (19 B.) Gleichenia quadripartita-moss-Sociation. Same locality and date. The vertical walls of the ravine; a narrow strip below St. 77, shaded by over- hanging ferns on the bank. M:2 Gleichenia guadripartita 2 Blechnum chilense 1 Gleichenia cf. litoralis 1 rare Polystichum vestitum 1 Pteris semtadnata 1 sr Acaena ovalifolia ssp. australis 1 Hymenophyllum rugosum 1 greg. I Nertera granadensis var. 1 Rubus geoides i Dendroligotrichum dendroides cop. Dicranoloma nigricaule Fissidens maschalanthus Rhaphidostegium Masafuerae cop. Rhizogonium mnioides 79. (27.) Same locality. Aryophyte-communities on the walls of the ravine and in the bottom on the rock bed and boulders, 5/3 1917. A. Under St. 78; and’ above B. Bartramia aristata greg. Dendroligotrichum dendroides greg. Dicranella costata Ditrichum affine Hypnodendron microstictum Philonotis scabrifolia greg. Polytrichadelphus magellanicus greg. Anastrepta bifida Balantiopsis hians Myla repens Plagiochila fuscobrunnea B. Under A and downward to the rock bottom of the ravine; very damp locality, surprisingly rich in species. Hymenophyllum rugosum Bartramia artstata Didymodon linearis Ditrichum affine Acrobolbus excisus Anastrepta bifida Anthoceros Skottsbergii Balantiopsis hians Hymenophytum flabellatum greg. Metzgeria decrescens greg. Mylia repens Noteroclada confluens greg. Plagiochila hyadesiana » rectangulata Schistochila berteroana » Skottsbergit » Splachnophylla Symphyogyna hymenophyllum Trichocolea verticillata Tylunanthus silvaticus C. Upper surface of boulders, as a rule above the water. Hymenophyllum plicatum Rhacomitrium subnigritum greg. » symphyodontum Anastrepta bifida Balantiopsis hians D. Lateral surfaces of boulders, washed when the stream is filled. Balantiopsis chilensis greg. Distichophyllum fernandezianum Plagiochila deformifolia Baeomyces chilensts Placopsis gelida var. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 945 The Alpine fern-grass heath. As a suitable boundary between a Subalpine and Alpine belt the 1100 m line was suggested; this is where an assemblage of species belonging to what | have called the Magellanian element makes its appearance: Acaena masafuerana, Galiunm masafueranum, Gleichenta quadripartita, Lagenophora Hartott, Luzula ma- safuerana, Myrteola nummularia, Rubus geoides, Uncinia brevicaulis and phleordes. Some are occasionally found in a lower altitude; thus, the lowest record is 550 m for Euphrasia formosissima (which does not belong to the Magellanian group), 750 m for Lycopodium scariosum, 830 m for Lagenophora, and 925 for Uncinia tenuis. Stations 75 and 77—79 are situated on this line; note Glezchenza, Rubus geotdes and a number of Austral bryophytes. On the other hand, not few of the Alpines were not observed below 1300 m. Pl. 114, 115. 80. (17.) Acaena masafuerana-Anthoxanthum-Sociation. Near summit of Cerro Correspondencia, c. 1350 m, 14—15/2 1917. Soil stony and strewn with angular boulders; subsoil moist, clayey-stony. Pl. 114, text fig. 42. II: 2 Polytrichadelphus magellanicus Anthoxanthum odoratum 2-3 Ptychomnium ptychocarpum greg. Gleichenia quadripartita 2 Rhacomitrium lanuginosum Agrostis masafuerana 1 » symphyodontum greg. Blechnum chilense 1 Thysanomitrium Richardiz, often as a Erigeron luteoviridis 1 dark fringe around boulders Libertia formosa 1 Zygodon pentastichum Lophosoria quadripinnata 1 Pernettya rigida 1 Acrobolbus excisus Frullania magellanica stones Il: 1 Jamesoniella colorata greg. Lepicolea ochroleuca greg. Lepidolaena magellanica Plagiochila fasciata stones > homomalla greg. » remotidens Mylia repens Solenostoma crassulum Acaena masafuerana 2 greg. Lycopodium scaritosum 2 Erigeron turricola \ Euphrasia formosissima 1 Gnaphalium spiciforme 1 Lagenophora Harioti 1 Luzula masafuerana 1 Lycopodium magellanicum 1 Cladonia aggregata » Scartosum 1 » coccifera I » gracilis > pyxidata greg. Lecanora atra stones » polytropa stones Lecidea leucozonata stones Placopsis gelida var. greg. Pseudocyphellaria endochrysea Breutelia Masafuerae Stereocaulon paschale v. magellanicum Ditrichum longisetum stones > proximum v. compactius Leptodontium fernandezianum Nertera granadensis 3 greg. Galium masafueranum 1 Hymenophyllum falklandicum 1 Rubus geoides 1 Serphyllopsis caespitosa v. fernandeztana 81. (18.) Acaena masafuerana-Anthoxanthum-Sociation. Same locality and date. Level, much less stony soil; dense carpet of flowering plants, leaving little room for cryptogams. 946 C. SKOTTSBERG II: 2 I Anthoxanthum odoratum 3 Nertera granadensis var. 1 Carex Banksii 1 Rubus geoides 1% D 544 , , “4 OF, e 5 . Pernettya rigida \ Dendroligotrichum dendroides IN Polytrichadelphus magellanicus Acaena masafuerana § Rhacomitrium loriforme Uncinia tenuis 2 Pseudocyphellaria mougeotiana Euphrasia formostssima 1 82. (16.) Open Anthoxanthum-dcaexa masafuerana-Sociation. West edge of highland near Las Torres, c. 1350 m, 14/2 1917. Soil sandy-clayey, light red- dish brown, mixed with stones; surface strewn with stones and boulders. Flower- ing plants scattered; solitary moss-cushions, lichens on stones. Soil probably saturated with water after heavy rains; signs of solifluction. II: 2 I Anthoxanthum odoratum, fringing the Ditrichum longisetum stones boulders Rhacomitrium lanuginosum Erigeron luteoviridts 1 M symphyodontum Ike2 Placopsis gelida var. Stereocaulon proximum v. compactius Acaena masafuerana \ p p Gnaphalium spiciforme 1 83. (15 A.) Lophosoria-Pteris semiadnata-Anthoxanthum-Sociation. Stony slope under summit of Las Torres (St. 84), 14/2 1917. II: 3-2 Acaena masafuerana 1 greg. Lophosoria quadripinnata 5 » ovalifolia ssp. australis 1 Pteris semtadnata 3 greg. Blechnum valdiviense 1 Iso I Anthoxanthum odoratum 4 Gleichenia quadrip~artita 2 greg. flistiopteris incisa 1 Polystichum vestitum 1 Cryptogams not recorded; see 84. Nertera granadensts var. 5 greg. Lagenophora Harioti 1 Il: 1 Rumex acetosella 2 84. (15 B.) Gletchenia quadripartita-Acaena masafuerana-Sociation. Summit of Las Torres, c. 1370 m, 14/2 1917. Rock ledges with patches and small col- onies of vascular plants; luxuriant moss carpets where water trickles down. irs Luzula masafuerana 1 Phoenicoseris regia 1 rare Uncinia brevicaults 1 Robinsonia Masafuerae 1 rare Urtica fernandeziana 1 3 =2 I: 1 Bromus masafueranus 1% Acaena masafuerana 3 greg. Agrostis masafuerana 1 II: 2 Asplenium dareoides 1 Gleichenta quadripartita 4 Cerastium caespitosum I Anthoxanthum odoratum 1 Hymenophyllum cuneatum 1 Bromus fernandezianus 1 » Jalklandicum 1 Carex Banksii 1 » pectinatum 1 rare Erigeron Ingae 1 Lagenophora Hartoti 1 THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 947 Rumex acetosella I Stellaria media I Uncinia tenuis 1 I Nertera granadensis var. 1 Rubus geotdes \ Serpyllopsis caespitosa v. fernandeziana I Bartramia artstata » patens greg. Ditrichum longisetum Psilopilum antarcticum Rhacocarpus Humboldtit Rhacomitrium striatipilum Zygodon pentastichum Acrobolbus excisus greg. Anastrophyllum leucocephalum Harpalejeunea setifera Flerberta runcinata greg. Tsotachis sp. Jamesoniella colorata greg. Lepidolaena magellanica Lepicolea ochroleuca greg. Lophocolea rotundifolia greg. Plagtochila fuscobrunnea » homomalla Schistochila pachyla Trichocolea verticillata Nephroma antarcticum Parmelia saxatilis Peltigera polydactyla Placopsis gelida var. Stereocaulon ramulosum 85. (30 D.) Alpine heath on summit ridge of Los Inocentes, 1 300—1 500 m, 9/3 1917. Patches of heath between exposed bed-rock. Above 1 300 m the vege- tation has the character of a scrub, similar to the scrub formation of the ridges on Masatierra, with Pernettya rigida and Blechnum cycadifolium as \eading species, but above 1 400 the heath was more open and cushion-plants (Aérotanella, Oreo- bolus) became conspicuous, continuing to the summit. Unfortunately, thick fog forced us to return without having had time to analyse suitable sample-plots. The following species were noted above I 350 m. Abrotanella crassifes, common above I 400 m Acaena masafuerana, also down to 1100m Blechnum cycadifolium, to 1 m tall Empetrum rubrum, a single plant on the summit Evigeron Ingae v. Innocentium Galium masafueranum Gleichenia quadripartita Flistiopterts incisa Hymenophyllum pectinatum » rugosum Lagenophora Harioti Lophosoria guadripinnata Myrteola nummularia, also down to I10oo m Oreobolus obtusangulus, above 1400 m Pernettya rigida, common Rubus geoides Uncinia tenuts, also down to IIoo m Ptychomnium ptychocarpum Rhacomitrium lanuginosum Zygodon pentastichum Lepicolea ochroleuca Trtchocolea verticillata Cladonia pycnoclada Usnea subtorulosa from c. 1000 m Beside the bryophytes and lichens mentioned under the stations, the fol- Rhaphidostegium aberrans Thuidium Masafuerae Radula hastata lowing species are more or less common in the heath formation; the list is of course far from complete, as numerous species must have escaped a non-spe- cialist. Lichens on stone Buellia barrilensis Caloplaca Selkirkit 948 high west barranca Spa laevissima and neesiana put their stamp on the scenery; the country looks like a dry steppe, yellow during the summer and probably not much greener at other seasons, because the new blades do not make themselves very conspicuous in the dense tussock of the dead ones. Pl. 116:1. A few other native species may occur in this community, and some weeds are rarely Gc. Lecanora atra coarctata Lecidea cyanosarca Massalongia carnosa Parmelia abstrusa » cetrata » cons persa » filosella var. revoluta Pertusaria hadrocarpa Psoroma vulcanicum SKOTTSBERG Rhizocarpon microspermum » obscuratum Usnea florida Lichens on earth Placopsis fuscidula, very abundant Pseudocyphellaria Freycinetit » gilva Sphaerophorus melanocarpus On dead mosses Toninia bullata The Grass-land of the Basal region. Along the east side of the island and on the talus slopes at the foot of the absent. 86. (14 G.) Szpa-Sociation. Rocky slope near entrance to Quebrada de las Vacas, 10/2 1917. I-II: 3 Nicotiana cordifolia 1 Is Stipa laevissima 4 » meesiana 4 Anthoxanthum odoratum 3 Bromus unioloides 3 Hlalorrhagis asperrima 1 87. As St. 86. Epilithic mosses and lichens on the valley floor near entrance to Quebrada de las Casas, 10/3 1917. Anacolia subsessilis var. Campylopus aberrans Ceratodon purpureus Eustichia Poeppigti Funaria hygrometrica Ptychomnium fernandezianum Rigodium arborescens Trichostomum brachydontium 88. (14 F.) Degraded Stipa laevissima-Sociation. Ensenada Toltén, on stony talus slope near the sea, 2/2 1917. II :2 Stipa laevissima 5 Anthoxanthum odoratum 3 Avena barbata 2 Chenopodium multifidum 1 greg. Halorrhagis masafuerana 1 Peperomia berteroana 1 rare On loose stones Buellia masafuerana » stellulata cop. Lecanora atra Pertusaria melanocarpa cop. Rhizocarpon geographicum Acarospora smaragdula var. Caloplaca rubina cop. Candelariella vitellina cop. Lecanora polytropa Lecidea enteroleuca cop. Parmelia caperata Bromus unioloides 1 Chenopodium murale | Physalis peruviana I Rumex cripus I Marrubium vulgare I THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 949 Polystichum berterianum 1, between Aira caryophyllea 1 stones Galinsoga parviflora I Scirpus nodosus 1 Rumex acetosella I Sonchus oleraceus I Silene gallica 1 Spergularia confertifiora 1 Il: 1 Wahlenbergia Masafuerae 1 Oxalis corniculata 2 Adiantum chilense 1 between stones 89. In addition I give a list of the weeds noted in the vicinity of the aban- doned settlement in Quebrada de las Casas, Feb.—March 1917. Acaena argentea, Aira caryophyllea, Anethum graveolens, Anthemis cotula, Bromus unioloides, Centaurea melitensis, Cerastium caespitosum, C. glomeratum, Cestrum parqui, Chenopodium ambrosioides, Cichorium intybus, Cotula australis, Erigeron canadensis, Gnapha- lium spicatum and stachydifollum, Hordeum murinum, Marrubium vulgare, Matthiola incana, Medicago denticulata and sativa, Melilotus indicus, Melissa officinalis, Mentha pulegium, Oenothera mollissima, Oxalis corniculata, Physalis peruviana, Raphanus sativus, Rumex acetosella, conglomeratus, crispus and pulcher, Senecio vulgaris, Silene gallica, Silybum marianum, Solanum _ argenteum, Sonchus oleraceus, Stellaria media, Taraxacum fernandezianum, Trifolium repens, Verbascum virgatum, Verbena litoralis. This is by no means a complete list of all introduced species found on the island. The Vegetation of the sea-shore. The short description of the geography of Masafuera (p. 926) shows that the conditions for the development of littoral plant communities are very much better here than on Masatierra. The accumulation of talus material along the coast has been considerable; it is coarse and consists of stones of all sizes, and even if the surf washes out most of the finer particles and there is little chance for a salt-meadow to become established, some soil is left between the stones, where colonists have gained a foot-hold. Small caves of the same kind as on Masa- tierra are also found. The basalt rock is hard and generally more or less covered with crustaceous lichens. 90. (14 H.) Colonists on the coast rocks on both sides of the entrance to the Casas valley, 10/2 1917. II: 3-2 Ruta chalepensis Bahia ambrosioides, locally abundant Scirpus nodosus Libertia formosa Il-1 II: 2 Blechnum aurtculatum Erigeron rupicola Lobelia alata v. rupincola Pellaea chilensis Spergularia confertifiora 2 masafuerana Tetragonia expansa Wahlenbergia Masafuerae Adiantum chilense Asplenium obliquum v. chondrophyllum Bromus unioloides Cyperus eragrostis Polystichum berterianum Pteris berteroana 91. (14 1.) Adiantum chilense-Sociation. Small cave near landing-place, Que- brada de las Casas, 10/2 1917. Plate 116: 2. 950° c Lean Adiantum chilense 4 Asplenium obliguum v. chondrophyllum 3 Oxalis corniculata 2 Blechnum auriculatum 1 SKOTTSBERG ] Bryum fernandezianum funaria hygrometrica Symphyogyna circinata greg. 92. (19 B.) Colonists on coast cliffs between Quebrada Angosta and Punta Negra, 16/2 1917. Asplenium obliqguum v. chondrophyllum, common Carex berteroniana Erigeron rupicola Eryngium sarcophyllum Libertia formosa Lobelia alata v. rupincola, wet places Nicotiana cordifolia 93. (19 C.) Scirpus nodosus-Sociation. Tierras Blancas, on a narrow talus stip at foot of cliff, 16/2 1917. II: 3 Nicotiana cordifolia, many shrubs II: Libertia formosa 2 Scirpus nodosus 2 No 94. (31.) Degraded Scirpus nodosus-Sociation. Playa Ancha, 12/3 1917. Level beach with sandy-clayey soil, strewn with stones; along the water a wall of rounded boulders. II: 2 Chenopodium multifidum 2 Sonchus oleraceus 2 Scirpus nodosus 1 La Polypogon crinitus 3 Anagallis arvensis 2 Spergularia confertifiora 2 Tetragonia expansa 1 I Lecidea icterica on earth 95. (23 D.) Littoral communities of lichens at south bluff of Casas near land- ing. Exposed to surf and spray. 24/2 1917. A. Lowest zone, washed by the sea also in calm weather. No lichens or algae. B. A narrow zone 3—4 m above high-water, protected by huge stone-blocks, but sprayed in stormy weather. Buellia stellulata cop. Caloplaca Selkirkit Pertusaria melanospora Rhizocarpon geographicum C. A little above and inside B, rarely reached by the spray. Caloplaca elegans v. australis Lecanora atra » dispersa Parmelia conspersa var. Pertusaria hadrocarpa » melanospora Cop. 96. (26 C.) Lichen community on vertical dyke in coast cliff near Quebrada del Mono, 27/2 1917. Buellia stellulata cop. Caloplaca elegans vy. australis cop. 2 orthoclada large patches » Selkirkit Lecanora atra Pertusaria melanospora Rhizocarpon geographicum THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS Qg5I Bibliography. Anales del Servicio Meteorolojico de Chile. ANDERSON, G. W. A new, authentic, and complete collection of voyages round the world. London. No date. Anonymus. Compendio de Historia Jeografica, Nautica 1 Civil de Chile. Bologna 1776. Anson, G. A voyage round the world in the years 1740——44. Compiled by Richard Walter. London 1749. BerTERO, C. J. Notice sur |’Histoire naturelle de ile Juan Fernandez. Ann. des Sci. mat. XXII. 1830. Bitter, G. Solana nova vel minus cognita V. Fedde, Repertorium XI. rg12. Brianck, E. Handbuch der Bodenlehre. III. Berlin 1930. BrancuHl, E. C. La Isla de Robinson. Valparaiso 1922. Burcer, O. Die Robinson-Insel. Leipzig 1909. CHRISTENSEN, C. & SkorrsperG, C. The Pteridophyta of the Juan Fernandez Islands. The Natural History of Juan Fernandez and Easter Island Il: 1. 1920. Douctas, D. A sketch of a journey to the North-western parts of the continent of North America during the years 1824, 5, 6, and 7. W. J. Hooker’s Companion to the Botanical Magazine II. 1836. Dumont d’UrvitLE, J. Voyage au Podle sud et dans ]’Océanie III. Paris 1842. Ermet, A. Eine Reise nach der Robinson-Crusoe-Insel. Hamburg 1889. Gay, C. Apercu sur les recherches d’histoire naturelle faites dans |’Amérique du sud, et principalement dans le Chili, pendant les années 1830 et 1831. Ann. des Sci. nat. XXVIII. GrauHaM, Marta. Journal of a residence in Chile during the year 1822. London 1824. HawkeEswortu, J. An account of the voyages undertaken ... for making discoveries in the Southern Hemisphere. I. London 1783. Hemsiey, W. B. Report on the Botany of Juan Fernandez and Masafuera. Report H. M.S. Challenger. Botany I. 1885. Jouow, F. (1) Las plantas de cultivo en Juan Fernandez. Anal. Univ. de Chile 84. 1893. -—— (2) Estudios sobre la Flora de las Islas de Juan Fernandez. Santiago 1896. Juan, G. y Uttoa, A. pe. Relacion historica del viage a la America meridional. Pt. 2, vol. II]. Madrid 1748. Kopren, W. Klassifikation der Klimate nach Temperatur, Niederschlag und Jahreslauf. Petermann’s Mitteil. 64. 1918. Looser, G. Excursiones a Juan Fernandez. Revista Universitaria (Univ. Cat6l. de Chile) XII: 4. 1927. Mosetey, H. N. Notes by a naturalist on the “Challenger’’. London 1879. Narrative of the Surveying Voyages of His Majesty's ships Adventure and Beagle. I. London 1839. Proceedings of the first expedition, 1826—1830, under the com- mand of Captain P. Parker KING. NeEceEr, F. W. Biologie der Pflanzen. Stuttgart 1913. Puitipp1, R. A. Observaciones sobre la Flora de Juan Fernandez. Anal. Univ. de Chile 1856. (German translation: ‘‘Bemerkungen iiber die Flora der Insel Juan Fernandez’’. Bot. Zeitung 1856.) QuENSEL, P. Additional Comments on the Geology of the Juan Fernandez Islands. Nat. list. le 4, 1Qi52) ; 952 C. SKOTTSBERG Raunki&r, C. (1) Planterigets Livsformer og deres Betydning for Geografien. Copen- hagen 1907. —_— (2) Livsformernes Statistik som Grundlag for biologisk Plantegeografi. Bot. Tids- skr. 29. 1908. — — (3) Livsformen hos Planter paa ny Jord. Mém. Acad. Sci. et Lettres Copenh. Ser. 7, VIII. 1909. —-— (4) Om Bladstorrelsens Anvendelse i den biologiske Plantegeografi. Bot. Tidsskr. | 1926: —_— (s) ber das biologische Normalspektrum. K. Danske Vidensk. Selsk. Biol. Meddelelser. I: 4. 1918. —— (6) The life-form spectrum of some Atlantic islands. Botan. Studier 4. Copen- hagen 1936. SxoTrsBerG, C. (1) Juan Fernandez-darnas sandeltrid. Sv. Botan. Tidskr. 4. 1910. —— (2) Vegetationsbilder von den Juan Fernandez-Inseln. Vegetationsbilder, heraus- geg. von G. Karsten und H. Scuenck. 8. Reihe, H. 2. Jena 1910. —-— (3) Studien iiber die Vegetation der Juan Fernandez-Inseln. K. Sy. Vetensk. akad. Handl. 51:9. 1914. —-— (4) Den svenska Pacific-Expeditionen 1916—17. Ymer 37. 1917. —— (s) Die schwedische Expedition nach den chilenischen Inseln im Grossen Ozean. Petermanns Mitteil. 64. 1918. —w— (6) The islands of Juan Fernandez. Geogr. Review V:5. 1918. —— (7) Den svenske Ekspedition til de chilenske Stille-Havs-Oer. Geogr. Tidsskr. 25,54 1919: —w— (8) The Phanerogams of the Juan Fernandez Islands. Nat. Hist. II: 2. 1922. —— (9g) Nagra i Goéteborgs botaniska tridgard odlade Juan Fernandez-arter. Meddel. Goteb. Botan. Tradg. I. 1922. —— (10) Svenska Pacificexpeditionen 1916—17. Ymer 44. 1924. In English in Geogr. Annaler 6. 1924. ——— (11) Einige Bemerkungen iiber die alpinen Gefiasspflanzen von Masafuera (Juan Fernandez-Inseln). Veréffentl. des Geobot. Institutes Riibel in Ziirich. (Festschrift Carl Schréter.) 1925. —— (r2) Zur Organographie von Gunnera. Sv. Botan. Tidskr. 22. 1928. —— (13) Plant Communities of the Juan Fernandez Islands. Proceed. Intern. Congress of Plant Sciences, Ithaca, N. Y. 1926. I. 1929. —-— (14) Notes on some recent collections made in the Islands of Juan Fernandez. Meddel. Goéteb. Botan. Tridg. IV (1928). 1929. —— (15) En mirkvardig 6flora. Naturens liv i ord och bild. Ny foljd. 2. Stock- holm 1931. —— (16) Juan Fernandez-6arnas vaixtsamhiéllen. Memoranda Soc. pro Fauna et Flora Fenn. 7. 1032. —— (17) Notes on the vegetation in the Cumberland Bay Caves, Masatierra, Juan Fernandez Islands. Ecology XVI. 1935. —— (18) Greigia Berteroi Skottsb. and its systematic position. Meddel. Géteb. Botan. Tradg. XI. 1936. — (19) On Mr. C. Bock’s collection of plants from Masatierra (Juan Fernandez), with remarks on the flowers of Centaurodendron. Ibid. XII. 1938. —-— (20) The Juan Fernandez and Desventuradas Islands. Plants and Plant Sciences in Latin America. Chronica Botanica Co. Waltham, Mass. 1945. ) The Genus Peperomia in Chile. Ibid. XVII. 1947. ) A Supplement to the Pteridophytes and Phanerogams of Juan Fernandez and Easter Island. Nat. Hist. II: 6. 1951. THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 953 Surcuirre, TH. (1) Sixteen years in Chile and Peru from 1822 to 1839. London 1841. —w— (2) Crusoniana, or Truth versus Fiction, elucidated in the history of the Island of Juan Fernandez. Manchester 1843. Tenz, O. (1) Una ascencién emocionante a la cumbre del monte Yunque. El Mer- curio 6/3 1922. Valparaiso. —-— (2) Nach der Robinsoninsel. Andina 3: 3—6. Valparaiso 1924. Water, R. A voyage round the world in the years 1740——44 by George Anson. Lon- don 1749. VicunA Mackenna, B. Juan Fernandez. Historia verdadera de la isla de Robinson Crusoe. Santiago 1883. Explanation of Plates. Plate 58. 1. Drimys confertifolia at upper edge of Nothomyrcia-forest on slope of a hanging valley on Cordon Salsipuedes, Masatierra, c. 400 m above sea-level. Right a young Dyrzmys in flower. Mrs. S. — Photo 8.12. 1916. 2. Forest of Drimys with dense undergrowth of Blechnum cycadifolium and Dicksonia berteroana, c. 670 m above sea-level on Cordon Salsipuedes. PEDRO GUTIERREZ standing among Blech- num. — Photo 8.12. 1916. Plate 59: 1. Large specimen of -agara mayu on trail to Plazoleta, Masatierra. Mrs. S. at foot of trunk. — Photo 28.12. 1916. 2. Very large specimen of Rhaphithamnus venustus in primeval forest on the Villagra-side of Portezuelo, Masatierra, c. 550 m above sea-level. The trunk is covered with foliaceous lichens. Mrs. S. at foot of tree. — Photo 10.1. 1917. Plate 60. Juania australis in the interior of Valle Ingles, Masatierra, c. 350 m above sea-level. 1 male, 2 female specimen, both in full bloom. Mrs. S. — Photo 19.1. 1917. Plate 61. Fruiting tree of /wania australis. Same place and date as Plate 60. Left, Gunnera peltata FHlistiopteris, Polystichum, etc.; right, Aristotelia. Plate 62. 1. Escallonia Callcottiae in flower on slope of Cerro Damajuana, Masatierra. — Photo 29.1 IQI7. 2. Coprosma Hookeri in brush-wood on Cordén Salsipuedes, Masatierra. — Photo 21.4. 1917. Plate 63. 1. Escallonia Callcottiae (left) and Cofrosma Hookeri (right) on slope of Cerro Damajuana, Masatierra. — Photo 29.1. 1917. 2. Young growth of A77zstotelia, Cordén Salsipuedes, Masatierra. — Photo 21.4. 1917. Plate 64. 1. Flowering branch of Rhaphithamnus venustus, along trail to Plazoleta, Masatierra. — Photo 28.12. 1916. 2. Very stout specimen of Boehmeria excelsa; same place and date. Mrs. S. 63 — 516796. The Nat. Hist. of Juan Fernandez and Easter Isl. Vol. 11. 954 C. SKOTTSBERG Ny N N nN Plate 65. Cuminia fernandezia in flower. Small tree in thin forest of Rabanal, Masatierra. — Photo 28.3. 1917. Plate 66. Robinsonia gayana, old 2 specimen growing out of a crack in the steep wall a little below Portezuelo de Villagra, Masatierra, c. 550 m above sea-level. — Photo 10.1. 1917. Robinsonia evenia 2, same place, c. 575 m. — Photo 3.12. 1916. Plate 67, Robinsonia thurifera at Portezuelo de Villagra, Masatierra. 1 d, 2 9. — Photo 12.4. 1917. Plate 68. Robinsonia gracilis, a 2.1 m tall specimen on Cordon Centinela, Masatierra, c. 530 m above sea-level. — Photo 28.3. 1917. Symphyochaeta macrocephala 2 on Loma del Incienso, Valle Frances, Masatierra. — Photo 1352: 1916; Plate 69. Good-sized tree of Rea micrantha in flower in damaged forest of El Rabanal, Masatierra. Aristotelia maqui in upper left corner. — Photo 28.3. 1917. Plate 70. Eryngium bupleuroides in flower near Portezuelo de Villagra, Masatierra. — Photo Io.1. 1917. Plate 71. Plantago fernandezia in flower near Portezuelo de Villagra, Masatierra. — Photo 3.12. 1916 @) wande25212% 1916! (2): Plate 72. Pernettya rigida in flower along trail from the settlement in Bahia Cumberland to Porte- zuelo de Villagra, Masatierra. — Photo 22.1. 1917. Urtica fernandeziana in forest on the high slope of Quebrada del Mono, Masafuera. — Photo 7.3. 1917. Plate 73. Large specimen of (Vicotiana cordifolia at Playa Ancha, Masafuera. PEDRO GUTIERREZ. — Photo23-2: 5O17- Plate 74. Chusquea fernandeziana in Quebrada Gutierrez, Masatierra, c. 300 m above sea-level. — Photo 7.12. 1916. Festoons of Weymouthia mollis on Nothomyrcia fernandeziana in the upper montane forest on the steep slope below Portezuelo del Yunque, Masatierra, c. 500 m above sea-level. — Photo 18.12. 1916. Plate 75. Lactoris fernandeziana in the Villagra valley below Portezuelo, Masatierra, c. 550 m above sea-level; height of specimen 1 m. — Photo 1o.1. 1917. Gunnera bracteata in scrub at Portezuelo de Villagra, Masatierra, c. 580 m above sea-level, with Pernettya, Coprosma Hookeri, Blechnum chilense and cycadtfolium, etc. — Photo 30.4. 1917. NE N N THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 955 Plate 76. Gunnera peltata on Masatierra. Large plant in El Pangal. GUILLERMO ARREDONDO. — Photo Aug. 1908. Gunneretum in dry stream-bed of Villagra, c. 170 m above sea-level. Mrs. S. — Photo 6.1. 1917. Plate 77. Calystegia tuguriorum in the canyon of Quebrada de las Vacas, Masafuera. — Photo 13.2. 1917. Asplenium obliquum var. chondrophyllum on coast cliffs at Puerto Ingles, Masatierra. — Photo 20.1. 1917. Plate 78. Cladium scirpoideum at a waterfall in the Pangal canyon, Masatierra. — Photo 1.1. 1917. Plate 79. Dysopsis hirsuta in dense forest near Portezuelo de Villagra, Masatierra. — Photo 27.3. 1917. Peperomia berteroana in forest on SW slope of El Yunque, Masatierra. — Photo 24.4 1917. Plate 80. Blechnum Schottii on Drimys in the forest of El Rabanal, Masatierra. — Photo kK. Backstrom, Aug. 1917. Plate 81. Thyrsopteris elegans in the upper montane forest of Villagra, Masatierra, c. 500 m; Mrs. S.; left Gunnera peltata, right Blechnum Schottiit. — Photo 12.4. 1917. Plate 82. Polypodium lanceolatum (fronds entire) and intermedium (scarce, frond pinnatifid) as epi- phytes in forest above Valle Anson, Masatierra. — Photo 18.12. 1916. Plate 83. Trichomanes exsectum on the wall of prison cave VI, Bahia Cumberland, Masatierra. Left, Blechnum auriculatum. — Photo 7.4. 1917. Plate 84. Macal in Quebrada Gutierrez, Masatierra. Where bed-rock exposed a patch of native ferns, Blechnum cycadifolium and, right and left of this, B. Schottiz. Under B. cycadifolium in centre, a tuft of Cladium scirpotdeum. — Photo 7.12. 1916. Plate 85. . Wooded slope of Villagra, Masatierra, showing boundary between forest and grass-land. Ravine in the west part of Villagra below the forest, looking NW toward the coast. Tuss- ock grass-land (.StiJetum) and fringe of Gunnera along stream-bed. — Photo 6.1. 1917. Plate 86. Almost pure society of Sz/ybum marianum in El] Rabanal, Masatierra. — Photo K. Back- strom Aug. 1917. Plate 87. Forest with good-sized Dicksonia berteroana in the bottom of El Pangal, Masatierra; K. BACKSTROM at base of a canelo (Drimys). — Photo 1.1. 1917. 956 C. SKOTTSBERG N N = iS) N tN Plate 88. Virgin forest covering the west slope of the west quebrada, Valle Ingles, Masatierra. On the spur in centre, groups of /zania. — Photo 20.1. 1917. Plate 89. Upper montane forest of El Pangal, Masatierra, above the precipice. Left an old, dead Drimys, and a Juanita; another tall chonta in centre. — Photo K. Backstrém Aug. 1917. Plazoleta del Yunque, Masatierra, with Pico Central. At foot of ridge a dense macal. The flat is covered with weeds. Mrs. S. — Photo 30.3. 1917. Plate 90: SE slope of Cordén Salsipuedes, Masatierra. — Photo 30.12. 1916. Crest of central ridge in Valle Ingles, Masatierra, with open scrub. Left Rodéimsonia gayana, right Juncus procerus and Silybum marianum, Mrs. S. — Photo 19.1. 1917. Plate 91. Brush-wood at foot of precipitous rock-wall on north side of Portezuelo de Villagra, Masa- tierra, facing NE. Rich Robinsonia flora (R. gavana, Phoenicoserts pinnata, Eryngium bu- pleuroides, etc.). — Photo 27.3. 1917. SW face of Cerro Piramide at Portezuelo de Villagra, Masatierra, showing influence of wind; note difference between leeward and windward sides of ridges. — Photo 10.1. 1917. Plate 92. El] Yunque seen from our Villagra camp, Masatierra. Thick fog fills the wooded gully ex- plored by Dr. TENz. The hill near the beach is Cerro Negro. — Photo K. Backstrém 6.1. 1917. Main ridge E. of El Yunque, Masatierra, seen from top of Cord6én Centinela, c. 780 m above sea-level, looking E. Scrub covers the ridge. — Photo 11.4. 1917. Plate 93. Lower montane Nothomyrcia forest in Quebrada Gutierrez, Masatierra. — Photo 7.12. 1916. Blechnum chilense (centre); Mrs. S. at foot of a tall Dicksonia berteroana; left Mr. BACKSTROM. Young /uania australis. Plate 94. Comparatively dry Mothomyrcia forest in El Rabanal, Masatierra. — Photo K. Backstrém, Aug. 1917. Dry Nothomyrcia forest in Quebrada de la Choza, Villagra, Masatierra, c. 300 m. — Photo K. Backstrém 4.1. 1917. Plate 95. Same locality as on Plate 94:2, but forest more open. The scandent fern is Arthropteris altescandens. — Photo K. Backstrém 4.1. 1917. General view of forest cover in Quebrada del Monte Maderugo, Masatierra. — Photo 2.12. 1916, Plate 96. Upper montane forest on ridge between Quebrada de la Piedra Agujereada and Q. Laura, Masatierra. — Photo C. S. 5.4. 1917. A very old tree of Drimys confertifolia with Dicksonia berteroana as epiphyte. In front a trunk of /wania australis. P. GUTIERREZ. A fine growth of young /uanza. N THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 957 Plate’ 97. Cliffs on Cordén Salsipuedes, Masatierra, more or less covered with moss-lichen communi- ties, with flowering plants as colonists. A large patch of Ochagavia elegans in centre. — Photo 8.12. 1916, Plate 98. Vegetation fringing a waterfall in El Pangal, Masatierra. Cladium sctrpordeum and Gun- nera peltata dominating; Llechnum chilense. Mrs. S. — Photo 1.1. 1917. Plate 99. Gnarled short-stem tree of Cuminia fernandezia (centre) and a large plant of Robinsonia gayana, a little below the forest in Quebrada de la Choza, Villagra, Masatierra, c. 200 m above sea-level. Among the ferns Dryofpterts tnaequalifolia (extreme left), Blechnum Schottit and Polystichum berterianum can be distinguished. — Photo K. Backstrém 6.1. 1917. Plate 100. Grass-land (Stifetum) of Villagra, Masatierra, with scattered Gunnera peltata. — Photo 6.1. IQI7. Plate 101. Lichen cover on seaside wall of ‘‘Robinson’s cave’, Puerto Ingles, Masatierra. — Photo ZONeOi7- Matthiola incana on the Puente, Bahia del Padre, Masatierra. — Photo K. Backstr6m, Jan. IQI7. Plate 102. Shore vegetation in Bahia del Padre, Masatierra. El Puente in the background. Grass- and weed-covered talus slope; Sadzcornietum along beach. Mrs. S. at camp-site; left our sleeping-bags. — Photo 26.1. 1917. Pure Salicornia peruviana-Soc. Same locality. — Photo 15.1. 1917. Plate 103. Santa Clara, Morro del Spartan (de los Alelfes). — Photo 26.1. 1917. “Campos’’-like community with Dendroseris litoralis and Rea pruinata, the latter past flow- ering. Mrs. S. Dendroseris litoralis in full flower. Mr. MUENA. Plate 104. Abandoned dwelling-site in Quebrada de las Chozas, Masafuera, c. 550 m above sea-level, in a grove of Myrceugenia Schulzet. The grass is Anthoxvanthum odoratum. Mrs. S. — Photo Bolo: Partly wooded slopes of Quebrada Inocentes, Masafuera. — Photo 9.3. 1917. Plate 105. Dicksonia externa—Lophosoria—Anthoxanthum-Soc. on slopes of Cordén del Barril, Masafuera. —— Photo 1.3: 1917. Subalpine heath above Quebrada del Mono, Masafuera. A complex of Lofhosorta-beds, patches of Guanera Masafuerae and carpets of Anthoxanthum odoratum. Dicksonia externa scattered. — Photo 12.2. 1917. Plate 106. Myrceugenia-Fagara externaSoc. extending up the side-gullies in Quebrada del Mono, Masafuera. In the forest an abandoned house from the time of the penal settlement. — Photon 12.2107 Hesperoseris gigantea in Quebrada de las Chozas, Masafuera, — Photo 14.2. 1917. 958 C. SKOTTSBERG NO N N Plate 107. A grove of old Myrceugenia Schulzet in Quebrada del Mono, Masafuera. Left a dead tree. — Photo 12.2. 1917. Young A/yrceugenia in flower at edge of forest. Mrs. S. Same locality and date. Plate 108. Virgin I/yrceugenia forest between Quebrada de Sanchez and Q. Larga, Masafuera, c. 515 m above sea-level. — Photo 26.2. 1917. Dense grove of Dicksonia externa in a gorge in Quebrada del Mono, Masafuera, c. 550 m above sea-level. — Photo 20.2. 1917. Plate 109. Entrance to the canyon, Quebrada de las Casas, Masafuera. The trees are Coprosma pyri- joita.—* Photomty.27 1907. North wall in the innermost part of the Casas canyon, seen from the ridge of Cordon del Barril, c. 1000 m above sea-level. A tapestry of Gumnera Masafuerae and groups of Myrc- eugenia Schulzet. —- Photo 1.3. 1917. Plate 110. Looking down the canyon of Quebrada de las Casas, Masafuera, from the second threshold. Width of canyon less than 10 metres. Pools of water left in stream-bed. — Photo 6.3. 1917. Plate 111. 1,2. Gunnera Masafuerae in flower along stream in the Casas canyon, Masafuera. Mrs. S. — N No Photo 23.2. 1917. Plate 112. Ravine of a small stream at Correspondencia camp, highland of Masafuera, c. 1130 m above sea-level (St. 77—79). Last tree of Dsimys observed. Lophosoria and Blechnum chilense hanging over the bank of the brook. — Photo 14.2. 1917. Plate 113. Vegetation on slightly overhanging canyon wall in Quebrada de las Casas, Masafuera. Peperomia fernandeziana in centre, P. Skottsbergii at extreme right. The large fern is Prerts berteroana. — Photo 23.2. 1917. Plate 114. A carpet of Acaena masafuerana and scattered Lagenophora Harioti (small rosettes) near summit of Cerro Correspondencia, Masafuera. — Photo 14.2. 1917. Plate 115. Uncinia phleoides in a bed of Lophosoria near the Correspondencia camp, Masafuera, c. 1130 m above sea-level. — Photo 5.3. 1917. Myrteola nummularia in the Alpine heath of Plano de la Mona, Masafuera, c. 1250 m. ——EeMOtO! 5:3. 1OL7- Plate 116. Grass-land on the lower slope of Cordén del Barril, Masafuera, c. 125 m above sea-level. The building is a stone watch-tower from the time of the penal settlement. — Photo K. Backstrom Feb. 1917. Cave with tufts of Asplenium obliguum var. chondrophylium on the shore between the Casas and Vacas valleys, Masafuera, — Photo 13.2. 1917. ; THE VEGETATION OF THE JUAN FERNANDEZ ISLANDS 959 Additions. The last sheets of this paper were going through the press when my Chilean friend, Mr. G. Looser, sent me a recently published book by JorGe GuzMAN Parapa, ““Cum- bres ocednicas’’, Santiago (no date). It deals with the geography and history of our islands and is based on the author’s visits and on extensive studies of the literature. I shall quote some of his statements. September 1574, seems to be the exact date of discovery. Of special interest to us are the recent events influencing the state of things with regard to nature protection. Rubus ulmifolius, the zarzamora of the Chileans, was introduced to Masatierra on pur- pose by Mr. DesipERIO CHARPENTIER, to be used as living fences. The author tells (p. 160) that it has taken possession of good agricultural soil and constitutes a serious problem. Experts have calculated with 300 hectares of arable soil on the islands, and still there are persons who dream of an intensification of agriculture and cattle- and sheep-breeding. The precious chonta palm is said to be on the verge of extinction (p. 214—215); it was still plentiful in some places in 1917. At that time there were no sheep on the islands, but lately Masatierra has been converted into a sheep-farm with 5000 sheep (1948) running wild all over the place (p. 178). Unfortunately, pasture runs low in March and April, but the author is confident that by introducing new kinds of grasses this difficulty will be overcome quite easily. As destroyers of native floras sheep enjoy very high reputation. The human population of Masatierra has been trebled since 1917 and counted about 650 in 1944 (p. 37). On Santa Clara are about 600 wild goats and 500 sheep, but no settlement. On Masafuera, the penal settlement was in operation from 1909 to 1913, and again from 1927 to 1930. At present about 70 free colonists live on this islands. Viewed against the fact that the archipelago was made a national park on Jan. 16, 1935, these news are very strange and alarming. The law expressly prohibits the erection of human habitations on Masafuera, and not even collecting of specimens of the flora and fauna is permitted, unless with special license and for scientific use only. The author declares himself a staunch friend of the protection of the living world, but nevertheless seems to regard with satisfaction the plan to intensify the tourist traffic, to build hotels, etc. From what he tells us in his interesting book we can arrive at no other conclusion than that the islands are doomed, an irreparable loss to the entire scientific world. g60 C, SKOTTSBERG Contents. FOre WOT co eek cca cs veie dos deg bots aie sieges clegetar Sigua’ +o cre ouelere,cleltdoya ie eet sen ote 793 Botanical EXCursioms ... 60... ..cq mec cies wre re shee che oie ele erie nei lel bre ee eee 802 Chapter I. Environmental factors .........- 0.5 eee cece cece cece cece e eee ences 805 Temperature and rainfall 2.200.200 op eee cir ne en te ences os cons foiesns + oem 805 sec ra eee meh Me tie Sn aN te srind bamicto bitched 0.010 009000008 812 Tight Conditions 20 ee eee ene ee oe eee ee cee ee eae eae oleate ee 813 The climate of Masatierra as compared with the climate of the coast zone of Chile oo. ccc ce ne coc ne ce nee tiene tele sa sie 3 ne 6 oo RDO NERRe SS aeeNe cet alent eee 813 [he climate of Masatierra compared to the climate of Macaronesia.......... 819 Soils. os. visi ules Side Gis aie wie « ates SoacsiySeaie thous Gecheene, -aephae Gaede RCRIREERS ACSC ka eee 819 Chapter Le Whe. Mife-forisy 37.2%. ect «crcl pleiel aaa ele athe heey ete 825 Chapter III. Morphological and biological remarks on the vascular plants ...... 830 The leading’ fOrestirees ea Me. ererereainsee aan ae (oe he ee eae $30 Trees: ofiisecondary: importance s2.25. bs eR RRe is. Bao. ee 835 Local. tree, species... . idan s de)-nayaoae deo. ates ee Sbeiiee tte ae eee 839 PE YrOSCEEELEFE ES! o.. | 5.2 oo 5, s eleeduece Sig eshecaal® © chek hae S(CREIOIRO > CLI IO eee oop Aud Tall SHrabs wc 5. wage els Ss cols e oh epee orm sous es yhtiny eee lereucte tan iene enne ete 862 SinmallSWra bse. 5 cece eee oases Gea woes im ad epee, sm ays aeeley seeieEY ha ete ee eer 866 Herbaceous plants ss. NAR 2AM. Sn PI on laseis che ears oe aintere ale hana ee ae 869 ThesPteridophytes:isivtcts ssi a aik SERIE SE Se peace ee 877 Chapter IV, The. vegetation: ..\. s...:a 9-12 -seiys =i a. Bi be erase 0 aie © te eee 885 UntrOoGUctiON .. 5» oa Sere eee tee 898 Tes fOrESt TECUOMY os ie: 5 '- ove ons) wins sues Slap ap stap epee pe TeRe rok ore nee ee 898 Whe orass-lamd sys cies sels’ ope oe Sack aoe tee ok ia nel ana th 919 ihe vegetation ‘of the ‘sea-shore . >. st ar. eee ere ciao eee g22 The veretationol Santa ‘Claraco.n ae ED oes ete cree oie ol ccs 924 @he bioksreeie Pesos a lee paale teehee een 926 APIVEMStATIONS 2. oui ie patie alee so cet cele Seite bagate ee came eee «an eee ep 933 Ane” TOLESE TE GION . 6.0. os ess «2 pioisin cieteuere TRI feeer Cee 933 ihe’ vecetation’ above “the (timber-line* 2. 2.0. ee eels oaths ieee 942 Mheigrass:land sof: the: basal region 42a. 2 tie. is en eee 948 The, yegetation ‘oft the seaishone-cac.e-\. ne. bepeeet Gtr adel a: eeeee 949 Biblrography 2.0 00.00. ee Eee wb eee ve ns Cd ete re g5t Kxplanation™ Of plates. 0... 6... 0.5 05 ce bees wae to oc-008 sila tee erent 953 AN GGTHONS: 6 seis ieee deals ogee. ace siecle eleepenanede ees nesevaye’ «cafe crea 959 Printed ™'/2 1053. ae i oe l aster Isl. ra i S be vé Fernan FuUaAH “4 Fist. Nat. 7 fe Easter Isl. Vol AL N ‘ fi \ om wed J a A x a. Ma i’ » : Ny “ [shi ead and kaste / Fuan Fernandes ~y Htst. Nat. Vol, IT PLATE 6 Nat. Hist. Fuan Fernandes and Easter Isl. Nat. Hist. Fuan Fernandez and Easter Isl. Vol. [1 PLATE 62 63 PLATE Nat. Fiis?. N Nat. Hist. Fuan Fernandez and Faster Isl. Vol. [1 PLATE 64 Nat. Hist. Ffuan Fernandez and Easter Ist. Vol. Il PLATE 65 Nat. Hirst. Fuan Fernandez and Easter Isl. Vol. Il PLATE -67 Fel Fuan Nat. Hust. Fernandes Fuan ~ Nat. Fest. PLATE 69 Vath, THE aster Isl. Z LE y Feerna A TUAK “Yy Nat. Hizst. Nat. Hist. Fuan Fernandez and Easter Isl. Vol. [1 Hts PLATE Vol. Lf ler, Isl. 7 Eas Nat. Hest. Fuan Fernandez and Easter Isl. Vol. II PLATE 73 Suan Fernan CY Nat. Fizst. Nat. Hist. Fuan Fernandes and kaster [sl. Vol. [1 PLATE 74 aw. , Ve 5 “4, pe On Vy Nat. Hist. Fuan Fe Easter Asta Ol PLATE SAS 9 70 PLATE bo dL - | tek and kaster Fernandes yy Tuan Fizst.! Nat. Nat. Hist. Fuan Fernandez and Easter LS Ol ed. PLATE 77 — , j ae LLa SLE} > Nek . - * NIN, . i > ae eS eal. Aa ah <7, Ss Wie ee < a) ‘ Stef A Das , h ee ig 3 = * ‘ 3 fy . ¥ rea oe " io - ‘ ome, Kins . is . . rr 4 UN # Minas ty 4h.4 . eu, Fee. ' ‘ ' ‘ ww Ae ~ ey Ys ; ” i . = x ~ ‘aon. SiN ite ay AN Fuan ~y Nat. First. Nat. Hist. Fuan Fernandez and Easter Isl. Vol. Il PLATE 80 Cates a We Suan Fernua FISTS Nat. PEAT] Isl. Vol, Il and Easte y Fernandes Fuan Nat. Fizst. d Easter Isl. Vol. [I PLATE 83 S . ie x * 4 > “ft unr) * ie ~ nf PLATE 84 Vole TT ter Isl. tas: and Vi t. Fuan Fernandes 7S Jag Nat. TUQAI Yy Nat. Hist. 4 : x 4 | 5 a 5 Vol. Lf and Faster Isl. n lernande Fu First. Nat. Nat. Hist. Ffuan Fernanda md Easter Isl. Vol. [1 PLATE 87 7 ~ ~ s by “ ~S ~~ ~ ~i ~ ~ ~ ~ nl ~ ~ ~ ~ } é ¥ Htst. Nat. Nat. Hist. Fuan Ferna and Easter Isl. Vol. IT PLATE 89 Nat. Hist. Fuan Fernandez and Easter Isl. Vol. [1 PLATE Qt Nat. Hist. ¥uan Fernandez and Easter Isl. Vol. II PLATE OR g2 PLATE LY "7 | Is/. and Easter Fernandes TUQAT ~~“ Nat. Hist. N N Nat. Hist. Fuan Fernandez and Easter Jsl. wv = ns S - N ; Nat. Hist. Fuan Ye taster Isl. Vot, Li PLATE Q7 Nat. Fist. Nat. I] ™ .~ — Tsz. and kaster Fernandes yy FuUaAn lag; Se Nat. Nat. Hist. Juan Fernandez and Easter Isl. Vol. I PLATE I01 Fuan and fa. Nat. Hist. Fuan Fernand and Easter Isl. Vol. If PLATE log } [¢ N Fernande FuUan ~ Nat. Hzst. Vat. Hist. Fuan fe 1 Easter Isl. Vol. [1 PLATE - 105 N = 107 PLATI Vol. 1 Easter Isl. — d TuUaH Elist. 7 Nat. LiL ry | Isl. > and Easter Fernande Fuan TLESER Ff Nat. eudive- 4 oe aed od De > 109 PLATE Hist. Nat. > © []/ ~ .~ bars, Is f and Easter Fernandez Fuan Nat. H1st. Nat. Hzst. Fuan Fernandez a 1a Easter Isl. Vol. Il PEA aie Nat. Hist. Fuan Fernandez and Easter Isl. Vol. [1 Nat. Hist. fuan Fernandez and Easter Isl. Vol. [1 PLATE, 5 Nat. Hist. Fuan Fernandez and Easter Isl. Vol. I PLATE I1I4 Nat. Hist. Fuan Fernandes and Easter Isl. Vol. fH PLATE oas 2 Nat. Hist. Fuan Fernandez and Easter Isl. Vol. I PEATE 110 los Vi ances Ne) Bal Sketch maps of the Juan Fernandez Islands with approximate distribution of the principal plant communities By C. Skottsberg 1917 Scale 1:50000 Distance between curves apprasimately 40 m. MASATIERRA B, = Bahia C.=Cordén Co = Cerro Ens. = Ensenada L.=Loma Masatlerra; Lower montane forest (Nothomyreia, Drimys, Fagara may, Boehmeria, Coprosma) Uppor montane forest ¢ same, Cuminia, Azars, Gunnera, tvee-ferns, Hymenophyllaceae ‘and Bryophytos) Brushwood and scrub (Pernettys, Ugai, Blechnum eycadifolium) fon the high ridges, rock fece carpets Scallered groups or specimens of Nothomyreia, Notive (Stipa, Nassella, Pipto: haetium) and Becondary (Avena) grassland, waste land home of the nia assemblage" Masafuera: M.= Morro Lower montane forest (Myrcougenia, Drimys, extorna, Coprosma py P.= Puna Pro = Puerto Q.= Quebrada Upper montane forest (Dicksonia externa) Rock face carpets (Gunners Masafuorae, ferns, grass) Subalpine grass heath complex with dominant Anthoranthum (intro- duced), Lophosoria and Gunnera; Jocally Dicksonia seattered Pil} Patches of heath with numerous Alpine and Subantarctic species; moss and lichen mats. Native (Nassella, Stipa) and sec: ‘ondary (Anthoxanthum) grass-land Maqui forest (Aristotelia) topographic maps exist. As told on p. 79), the Chilean Air Farce, on my request, undertook to sur~ April 1932 Masarierra was photographed from the air and the result put at my dispoml. Large yor of the island were, however, hidden by clouds, but some important corrections of che current chart could be Bade For unforteen reasons the survey had to be discontinued after the first and only Qight. Mr, Bertil Frodin, who Jot teco invited to join, wok a fine series of Kodachrome pictares and these be most liberally placed at my disposal Aramber of them will be published in Vol. 1, no. 4. They form 4 moxt valuable addition 10 our lange set of photo ppb Without bi ‘ould not have ventured to draw « new map. I want w emphatize thar is docs not in deuail. Basenvial corrections of the very unsatidactory map of Santa Clara haye been made | Ne cuteral fit for we was obtained from Matafuera. My map is & sketch made by our party in 1917 and later wed the Chilean chart. A new map ix urgently needed. P. Negros M. Juanango & ~3) La “o, "%, a P. Aguabuena Los Chamelos?o Ingles Q. Helechos Q. Salsipuedes oP Bacalao P. Pescadores Q. Piedra Agujereada S= Z SSF Hucso Ballena El VerdugoD 2 3 Scale 1:50000 MASAFUERA ién_-Cabo Norte in = "FQ. Sanchez /} \1Q-Negra 4Q, Sandalito U .Q.Sindalo Q. Pasto =F Choro Dofia Maria Scale 1: 50000 ry ¥ igi} | ie t MM i } e P ? ] ’ 3 ® ¢ 4 a c fi | i} {MM 5 5749 5 0023 8 ll — _ $1) | > , a Caen si e s a: = gue ~~ 88 : ‘ PS ae Uae : 7 : 7 ? ae ~~ > - a - a - o SS fe = =~ 7 * .~'* ~_ = ty , : ms < oy ae =e, ad n — ¥ "c : 5 a5 ¥ = Fang or ona # ;