THE Natural History Juan Fernandez AND Easter Island EDITED BY CARL SKOTTSBERG THE NATURAL HISTORY OF JUAN FERNANDEZ AND EASTER ISLAND EDITED BY DR. CARLSKOTTSBERG VOL. I GEOGRAPHY, GEOLOGY, ORIGIN OF ISLAND LIFE WITH 14 PLATES AWE UPPSALA 1920-1956 ALMQVIST&WIKSELLS BOKTRYCKERI AB 707782 Table of Contents. 1. Skottsberg, Carl. Notes on a visit to Easter Island 3 2. Hagerman, T. H. Beitrage zur Geologic der Juan Fernandez-Inseln .... 21 3. QuENSEL, P. Additional Comments on the Geology of the Juan Fernandez Islands 37 4. Skottsberg, C. A Geographical Sketch of the Juan Fernandez Islands ... 89 5. Derivation of the Flora and Fauna of Juan Fernandez and Easter Island 193 The Nat. Hist, of Juan Fernandez and Easter Isl. Vol. I. "\iica4a "Ho Orwo EASTER ISLAND after Ihe map of fhe Qrilean H\'dr. OHice -t:aooooo Ueiifhi 1/1 miners abox-e sea leve? . Obsrrvatiart spaf 310 /^i&oVo^uH I. Notes on a visit to Easter Island. ♦ ■■. ' By CARL SKOTTSBERG. With 14 plates, i map, and 3 text figures. While working on the Juan Fernandez Islands, our party obtained per- mission to accompany the Chilean corvette » General Baquedano» on her cruise to Easter Island, in 191 7. A short and very popular account of our visit appears in my book »Till Robinson-on och varldens ande» (1918). Although the purpose of our survey was purely biological, no scientist visiting the famous island can help taking a vivid interest in the archaeological remains, and we occasionally made a (ew observations. However, I have refrained from writing anything on the subject, as I had to wait for the publication of the results obtained by Mr. and Mrs. SCORESBY RouTLEDGE of the British »Mana» Ex- pedition. Last year Mrs. RoUTLEDGE published a most interesting account of their work on the island (The Mystery of Easter Island), where the ancient monuments of all kinds are amply described and illustrated. A second volume will follow, containing the detailed descriptions of the prehistoric remains. I willingly admit that this little paper will appear rather unnecessary since the British Expedition has explored the place with such a wonderful accuracy. But it is Mrs. RoUTLEDGE's excellent narrative which has induced me to collect a few notes and to use them as a basis for a discussion of some interesting points. I have also found it worth while to add a number of my photographs, which may be of some value. The »General Baquedano» sailed from Iquique on May 27th, 1917. She carried a Government commission presided over by Bishop Rafael Edwards, a prominent Chilean ecclesiastic, who went to continue his studies on native conditions and to distribute a large amount of materials, clothing etc. among the members of the little island colony. Capitan de fragata J. T. Merino was in command of the vessel, and he as well as the bishop and the officers of the ship did all in their power to assist us in our undertaking. After a rather uneventful cruise our vessel anchored in La Perouse Bay on June 15th, and the same day we made our first excursion along the north coast. We were bound for Hanga Roa (Cook's Bay) but were detained in La Perouse on account of adverse winds; finally we resolved to cross the island , t ARL SKOTTSBERO (,n liorseback, and arrived at Matavcri, the scat of tlie farm house, on the 19th. hi the meantime u e liad made some excursions in the northeastern part, where Mt Katiki was ascended. At Matavcri we were cordially received by Mr. ri:k( V i:i)MlNhS, the mana-er, and were invited to to take up our quarters in his house. ( juite naturally, the natives were in a state of great excitement ovc-r the arrival of the vessel with their much beloved bishop, the missionaries (two Capuchin l^rethrcn) and the many useful articles reported to be onboard; and conse(iuentl\- the\- wore rather unwilling to render us any assistance. It was only through the kind intervention of the bishop that we were able to get an important })ait of our scientific baggage, which had been left on the beach at La Perouse, transported to Mataveri. From our headquarters the district round Hanga Koa and Ilanga Piko etc. was visited and several trips under- taken to Kano Kao and Orongo, the famous stone village. Further, our work was exlendend to Kano Aroi and Mt. Terevaka, the highest mountain, and also to tlie south coast and to Rano Raraku, the image mountain. We had expected to remain at least three weeks on the island, and greatly regretted that we were ordered onboard already after a fortnight's stay, espe- cially as my ca{)acit\' for hard work had become reduced on account of illness. The >Ha«iucdano» left ICaster Island on July 1st. (IKNKK.M. NOTF.S ON TIIK GEOGKArilV OF EASTER ISLAND Tlic topographical features of the island are fairly well illustrated on the accom[)anying map, the result of Chilean Navy surveys of later years. The position of the observation spot in Ilanga Roa is given as Lat. 27° 08' 06" S., Long. 10/ 25' 54" W. Mrs. Roitledge's map is based on U. S. Hydro- graphic ()ft'ice chart no. i 1 19, from which it differs in the position of some of the mountains and in the geographical names. There are certain discrepancies between this map and the Chilean one. and the difference between the latter and the U S. chart are still more considerable, especially in the configuration of tlic northwestern part of the island. There has been some confusion in the placing of the names, but I take it for granted that all the names used by Mrs. R. art- properly s|)elt and righth' placed. 1 lie island is known to be wholly volcanic. There are no signs of recent action, save for a couple of tcj)id sj^rings below high water mark reported to, but never seen l)\ us. It is rather curious that both THOMSON (Smiths. Inst. Ann. kcp. iSS.^. Washington 1S91) and Agassiz (Mem. Mus. Comp. Zool. ( ambridgc 33, i9) should discuss volcanic eruptions and great earthquakes as a possible reason for the destruction of the megalithic monuments and for the (lisappearani e of the greater part of the population, as all signs of recent catastrophes are entirely absent. On the other hand, not a few craters are well preserved; someiinies thev are arranged on distinct lines suggesting lines of less resistance in the older, more or less horisontal basaltic beds, which form the bulk of the island. 1 lie tufas and ashes of the numerous cones present a great variation ot colour contributing to lessen the monotony of the scenery. The attention of the visitor is especially drawn to the three great A"^;/^. Rano NOTES ON A VISIT TO EASTER ISLAND 5 is the native name for a mountain which contains water. R. Kao forms the broad southwestern promontory. The greatest height of the rim was found by the writer to be 316 m., as the result of three aneroid observations (differences in temperature duly considered) at different occasions. Agassiz has i 327 feet or 403 m.; the Chilean map, 400 m. The crater lake measures, according to Cooke, 2085 feet across (634 m.); this figure may be the result of a careful observation, but seems too small. Its surface was found by me to be 120 m. above sea level; THOMSON and CooKE (Smiths. Inst. Ann. Rep. 1897, Wash- ington 1899), say, respectively, that it is 700 and 600 feet below the crater rim, thus according to their figures for the latter corresponding to a height of 190 or 160 m.; taking 316 m. as the starting point we get 106 or 134 m. I am sorry that we were unable to examine the thickness of the peat that covers the sheet of water save for some irregular pools, which do not appear to have decreased much in size since the photographs of the Albatross Ex- pedition were taken. No reliable figure for the depth of the lake has been obtained; according to CoOKE, Mr. Salmon tried to sound, but at a depth of 300 feet the line broke without having reached the bottom. I need not tell that according to the belief of the islanders, the pool in Rano Kao belongs to the category of famous lakes without a bottom. The lake is partly surrounded by stands of a very robust buUrush, an endemic variety of the widespread Scirpus riparius, called paschalis by Dr. KUKENTHAL. The country NE of Rano Kao is hilly, one of the cones being known as Punapau (Plate i), the seat of the hat quarry. The northwestern corner of the island is occupied by the highest mountain, the Terevaka (Plate 2), non seldom veiled by a bank of clouds. This name is not mentioned in Mrs. RouTLEDGE's book, where the entire high land in question is called Rano Aroi. But the latter name only applies to the crater on the southeast slope of the mountain. On the Chilean map appears Cerro Terevaka, separated from Rano Aroi (or Roi) by a shallow depression, and both names were recognized by the native Juan Tepano, who accompanied us to this place. The top of Terevaka did not present any marked signs of being a crater; the height was found to be 530 m., which I believe is nearer to the truth that the figure 770 put down on the Chilean map. I had expected to find some notable difference in the flora of the highland, but was rather disappointed. The cryptogams, mosses and lichens, however, played a much greater part here than in the lowlands, where they are of a very slight importance. Rano Aroi is a very modest and shallow rano and cannot at all be compared with the grand R. Kao. The height was found to be 425 m. The lake is overgrown with vegetation. There is a gap in the east wall through which, after prolonged rains, the water flows down to another pool, which empties itself into a long, narrow fissure, crossed by the track from La Perouse Bay to Hanga Roa. This fissure does not seem to have been eroded by water but suggests a volcanic origin. The land along the north coast, from the hills backing Anakena Cove to Katiki, is a rather flat basaltic plain, with occasional outcrops of hard rock and strewn with innumerable sharp-edged stones, partly hidden by the coarse grass and making walking disagreeable, more so for a person in a state of ill- TAkL SKOTISUKKG health. The slope of Katiki (another name not found on Mrs. RouTLEDGE's map, but frecpienlly used) is comparatively gentle. The top was found to be 412 m. higli (300 m. on the Chilean ma[) must be wrong); it presents a rather striking appearance, forming a shallow circular basin, perfectly dry and with a flat bottom 5-6 m. below the rim, which is 75—80 m. across. On the north slope is a succession of three cones, of which the northernmost is gradually eaten away through tlie action of the sea. The one nearest Katiki, Vaintu Rova, is of a light yellowish colour; the liight is 310 m. On the south slope wc came across a deep fissure, containing rain water and surrounded by a fine growth of ferns. The natives, of course, all know this rare watering-place, and I guess this is the well spoken of by Captain CooK in his second voyage: • Towards the eastern end of the island they met with a well whose water was |)erfectly fresh, being considerably above the level of the sea; but it was dirty, (Mving to the filthiness or cleanliness (call it what you will) of the natives, who never go to drink v.ithout washing themselves all over as soon as they have ilone . . . (the edition in Iweryman's Library, p. 163). North of Vaintu Rova stands the somewhat lower Tea-tea, the » white mountain*. SW of Katiki the famous image mountain, Rano Raraku, is situated, so ably described and illustrated by Mrs. ROUTLEDGE, who gained an intimate knowledge of this unique place. It is shown on Plate 5. Between this rano and the hills east of llanga Roa there is an extensive plain, only broken by a few higher eminences. The visitor, even if he be not a geographer, cannot fail to notice the absence of every trace of valley or ravine caused by the action of running water. It is almost with surprise that, one learns the figure for rainfall, i 218 mm., the average of 8 years' observations. This is, indeed, no small amount, surpassing that of Juan h\'rnandez, where erosion has modelled the entire island into a system of deej) valleys and sharp ridges, l^ut in Easter Island there is no stream, no brook; only in the crater lakes water is always found. The great .scarcity of water makes the high development of the ancient culture (juitc astonishing. The climate is warmer in Easter Island than in Juan Fer- nandez, the evaporation undoubtedly much greater, the winds at least equally tresibly also be the same ahu as no. 34 Punahoa of THOMSON 1. c, p. 505. lie also <^ives the total length of the single moai as 32 feet. Ilmvever, the entire structure is said to have a length of 175 feet and a width «)f S feet, which figures must be entirely incorrect if ahu Pare is meant. \W what kind of apparatus or devices the statues were transported from Kano Raraku to the coast, in some cases to rather inaccessible places, remains a m\ster\'. The natives possessed strong cordage, and Mrs. R. has made out that long lines were used occasionally, but veritable hawsers would have been needed to drag the statues along over the ground in the manner imagined by Thomson. We have seen that there is evidence against the island ever having produced good sized timber suitable for rollers. THOMSON thinks that, after a smooth road had been constructed, »the images were dragged by means of ropes made of indigenous hemp»; »seaweed and grass made excellent lubri- cants*. He could »clearly see how it was accomplished with a large force of able-bodietl men» (p. 49S). I must confess that I find it less easy to understand how the work was done, for the least obstacle would become a serious one; and the roads must have been made as smooth as a fioor in order to serve the puri)ose, the images being rather fragile. Mrs. R. has traced the few high- wa\-s leading from Rano Raraku to the coast; but if really the images were draggetl up to the numerous ahu all round the island, these roads cannot have sufficed, but an elaborate network of very smooth paths was required, of which all traces would have disappeared. It is true that seaweeds are plentiful, but there is no species of any considerable size and I fail to see how the quantities re(|nired could have been brought together. It is astonishing that no tradition on the means of transport survives. According to Mrs. R. the natives in- variabl\- offered one exi)lanation: that the images were transported by the aid of supernatural forces. < )n p. 4X6 Thonlsox discusses the possibilit)^ of a transport by sea. Near a grouj) of ahu he discovered a fine landing-place made by art, »admirably adapted to the landing of heavy weights*. hVom old drawings we know what the aboriginal canoes were like — not a single one, as far as I know, has been preserved to our watched. The ob.server must keep outside the t(nver. or on the top of it, not a very comfortable place. Really, the tower itselt would have been little more than a refuge in bad weather and during the night, but for such a jnupose a much simpler structure would indeed NOTES ON A VISIT TO EASTER ISLAND 13 suffice. Thus the importance of the tower with its lower apartment appears to stand in no reasonable proportion to the vast amount of labour required to build it. We should perhaps remember that the permanent dwelling-houses were much more fragile. The tower suggests some kind of fortress, with a chamber for stores or treasures: it would be easy to defend the entrance. Speaking of the narrow entrances to the Orongo houses, Thomson remarks (p. 483): »The low contracted entrances were used here as well as elsewhere for defence. Factional fights were common, and it was necessary that every house should be guarded against surprise and easily defended ». He adds: » Another reason might be found for making the openings as small as possible, in the absence of doors to shut out the storms». But, at least at Orongo, there were plenty of slabs suitable for doors if wanted. 2. Diagrams of old plantations; a seen from above; b three types of shelter, in section. I = bananas, 2 = Melia azedarach, 3 = Andropogon halepensis. Scale i : 230. On the other hand, an enemy could pull down the roof over the occupants. Also, we should expect to find a communication between the tower and the underground chamber, which we have not found. So, after all, this theory may not hold good. Perhaps prisoners were locked up in the towers, where they could be easily guarded till the hour arrived to put them to death. But it is also necessary to consider whether these buildings may not have had some relation to unkown rites. Did they have a ritual purpose, we do not need to wonder about their elaborate structure, as natives may invest any amount of labour in connection with religious or other ritual buildings. There is a description and figure of a similar building in La Pi^ROUSE's Voyage, reproduced by Stolfe in Ymer, 1883. He states that they are only found on the top of Rano Raraku, which is, of course, a mistake. They were oval in shape; close by was the underground chamber with its separate en- trance. In several respects, especially concerning the ahu, there are great dis- 14 CAF^L SKOTTSBERG crepancics bctw ccn tlic statements and illustrations of the old navigators and the results obtained by modern explorers. A critical examination of the old stories would be welcome, and it is to be expected that Mrs. RoUTLEDGE will undertake to scrutinize the entire literature. There are, e. g., in La PIiROUSE's and I'lNAKl's narratives designs of ahu which do not at all correspond to modern descriptions or photographs. In the vicinit)- ot Ilanga IIo Orno we saw many remains of native planta- tions. 'ihe\- are of several types. One, seen in fig. 2 a, is probably of a later date, as the material has been taken from an ahu, the front wall of l-ur V n. Tsv.) .iMrd-mcii)> on rock at Oron^o (hciolu ol the rock 1,6—1,7111.) b. Incised marks on door poost at Orongo. which fornix the back wall of the garden. Circular miniature gardens are represented in fig. 2 /'. The need of shelter and moisture is well unterstood. Mclia was said to be u.oun for the sake of the timber. Probably it is of recent intii)(iuction. •>K'»N(;o AND WW. lURI) cui/r Mrs. koriiKixiK (i,-votcd nuich time to the survev of the Orongo village, an(l as a detailed plan was made and every hou.se measured and described, I shall content mysell with a few short remarks. The last house (if I remember right) towards the gap of the crater rim. close to the sculptured n.cks, had one door-.post with incised carvings left. NOTES ON A VISIT TO EASTER ISLAND 15 vide fig. 3 b. This house was pointed out to us as the house of Ariki. The ariki was the chief of the Miru clan, the authority on the script (i. e. the »ariki-mau», vide R. p. 241; all Miru were also called ariki). Now, the same design was found on a skull in the possession of the schoolmaster, Mr. I. VlVES, and this skull was attributed to an ariki. Unfortunately, the owner did not want to part with his treasure. The design is unlike the one figured by Mrs. R. (fig. 96) of another Miru skull. The Bird Cult i? described, with full details, by Mrs. R. Special attention is paid to the rock carvings. I sketched a couple of the »bird-men» (fig. 3 a). Their meaning is not known with certainty. Mrs. R. believes them either to represent one of the egg-gods (they were spoken of as »Make-make») or made to immortalize the bird-men, the winners of the egg-race; she finds the latter explanation more probable. I have not been able to form an independent opinion. The same carvings are seen on a flat stone opposite Orongo, marking the place where the path descends into the crater of Rano Kao. All that is left of prehistoric remains, at least of the large ones, will remain on the island. Shortly before our visit a law was passed prohibiting the removal of statues etc., so that we had to abandon our idea of bringing home a small image presented to us by one of the residents. The »Mana» was just in time to rescue the small but unique statue from Motu Nui. • . • WOOD CARVINGS The famous wooden statues as well as other pieces of carving are gone from the island for ever. What is ofi"ered to passing visitors is not worth mentioning. The art is gone. One old moai-miro, in a very much decayed state, had been discovered in a cave after the departure of the »Mana». It was presented to Bishop EDWARDS. In 1908, while staying at Valparaiso, a Swedish captain, Mr. G. Karstrom, who had been shipwrecked on Easter Island many years before, presented me with two beautiful wooden images, one of which is in the Etnographical Museum in Stockholm; the other is owned by a private person. HOUSEHOLD GOODS, WEAPONS, ETC. Very little of this kind is now to be encountered. Sticks used for net- knitting are available, and so are baskets or rather bags made of bullrushes (figured by THOMSON on Plate 51). There are still some people skilled in the preparation of tapa cloth from the mahute and of strings from the hau-hau, and we had samples made for the collection. Curiously enough, Geiseler does not mention the latter plant, but states that all the cordage, fishing-nets etc. were made from the bullrushes. It is generally stated that the islanders never possessed any earthen-ware. Contrary to this, RUTLAND (Transactions New Zeal. Inst. 29, 1896) says that ,6 CARL sKonsr^KKG the earliest liiscoverers liad seen Mude carthen-\vare» on the island, a statement due to some misinterpretation. Of stone implements, besides the toki, and the stone adze and chisels, we t;x>t one lishiiook, very neatly wrought but unfor- tunately not complete, as the point is missing (Plate 14, fig. 4)- There is a drawing of one of these hooks in Thomson's report (Plate 58). Another curious article is tlie si)hcrical stone ball, Fig. 5 on Plate 14. It shows two holes which communicate so that a string can be passed through, and may have been worn as an ornament. No explanation was offered. Perhaps it is a fetish stonc». llioM.soN has described and figured many such stones, but none of them i)resent any likeness to this one. The object on Plate 14, Fig. 6 is not, as might be suspected, a broken spear-liead or ))iataa, but has been given its present shape on purpose. It fits well into the hand and may have been used as a knife or scrape. But if it was used with a handle, my explanation may not be satisfactory. Spear-heads are ct)mmonl\- found in the soil and also manufactured to satisfy the demand of visitors. Two, of an ordinar\' type and apparently old, are seen on Plate 14, h'igs. 7, S. According to THOMSON there were at least nine kinds, all with dilVerent names, a statement well needing the corroboration of Mrs. R. ORKIIN OF THE PEOPLE The histor\- of Faster Island is full of mystery, but I think that Mrs. R. has come pretty near the solution of some of the problems. She has drawn some imj)ortant conclusions from the legendary traditions still alive. A tale of two ditTerent races and two successive colonizations runs through the old legends. Tlie anthropological evidence seems to be in favour of a double origin, Melancsian and I'oKnesian. The comparative studies of the Bird Cult in the .Solomon Islands and I^aster Island (by PI. Balfolr, vide Mrs. R.) seem nothing less than convincing. The bird rei)resented in the numerous carvings, paintings etc. of ICaster Island is not the holy bird of this place, but the frigate bird, u<»rshi|)|)cil in the Solomon Islands. 'Phe bird figures were called penguins by Lkhm.wn (l-.ssai dune bibliograi)hie, Anthropos, 1907), which undoubtedly must be a mistake, especiall)' as penguins hardly ever visit these waters. it wi- sum up the results obtained, there is evidence that the Easter Is- landers ate ot a twofold origin and that, after the Melanesian immigration, a P(»l\-nesian immigration followed. l"he population now tends to assume a multicoloured aspect; there has been a late influence from Tahiti (so we were told) and various white men ha\'e contributed towards the »amelioration» of the race. The xouul; girl figured on Plates 12-13 ^^•'■'^ ^^^^ to be of » pure Easter Island race , but whether rejjrcsenting a IVlelanesian or Polynesian type, I am unai)le to trli. A critical examination of the language would be of interest. Many words are the same as in the Maori or other Polynesian tongues, such as viaunga (mountain), inaJiul, (paper mulberry), // (Cord^line), ciimara (sweet potato) a. o. A large vocai)ular\-, collected by Padre Roi SSEI,, was publi.shed in Santiago, NOTES ON A VISIT TO EASTER ISLAND 17 191 7, but it miist be used with much criticism, as it contains many Tahitian words and also corrupted EngHsh, French or Spanish; good examples are anio (agneau) and rnutone = sheep, himene (hymn) = to sing,' teperanate = serpent, tokini = stockings, tiaporo (diablo) = devil, viretute = virtue, given without reservation. A closer look reveals that the material is not at all so rich as the number of words would indicate, for the author has invented hundreds of ex- pressions for ideas wholly unfamiliar to the aboriginal soul, by combining the words and extending their meaning in a most improper manner, e. g. expres- sions for cabin, desert, doctrine, palace, river, saint, W. C. etc. etc. to quote a few obvious examples, of which scores could be given. This is, I believe, a common missionary method to enrich the language with ideas and expressions necessary for the translation of reHgious and other books, but otherwise never used by the natives. Concerning the name of the sweet potato, see below. It has been ad- vanced as indicating an American influence previous to the Columbian era. Rutland (1. c.) thinks that the ancient monuments bear witness of a constant communication between the island and Peru and Mexico: »from hence architects of Easter Island may have been derived ». CULTIVATED PLANTS If we knew the history of the cultivated plants, many a mystery related to the history of mankind would be solved. But, unfortunately, discussion often begins with the original home of the wild parents of these plants, and there it also ends. The first record of domesticated plants in Easter Island is that of RoG- GEVEEN, the discoverer of the island or, at least, the first white man to set his foot upon it. He makes the following statement on p. 120 (De Reis van Jacob Roggeveen. Worken uitgeven door de Linschoten-Vereenigung 4. 191 1): »en toegebragt worden alles wat sy hadden, bestaende en boomvrugten, aardgewasch en hoenderen», that is, tree-fruits, soil-fruits (rootcrops) and hens; and, farther down: »want na verloop van een kleynen tijd bragten sy eene menigte van suykerriet, hoenderen, ubaswortelen en bananas», that is sugar- cane, bananas and ubas-roots. But what is ubas? Most likely the same word as the Malesian ubi (uwi, huvvi), yams (Dioscorea alata), now called ufi in the island. All these plants are of Old World origin and have spread from the Indo-malayan region over the Pacific. According to Friderici, the same word, in a corrupted form, is current in South America: »Dieses Wort schlagt eine Briicke iiber den grossen Ozean: es gehort als op unter der Bezeichnung 'siisse Kartoffel' zum Sprachschatz der Chimu, des kiistenbewohnenden Kulturvolkes westlichen Siidamerikas* (Wiss. Ergebn. seiner amtl. Forschungsreise nach dem Bismarck Archipel im Jahre 1908). But the bridge in question seems to be weak. The word cumara is used for sweet potato [Ipomaea batatas or Batatas edulis) from New Zealand through Polynesia to Easter Island. According to Cheeseman (Manual of the New Zealand Plora) the Maori introduced the plant from Polynesia when they colonized the country (supposingly 1350 — 1400), and it was described by SOLANDER as Convolvulus chrysorhizus, now reduced to 2 — 20199. The Nat. Hist, of Juan Fernandez and Easter Isl. Vol. I. I 8 CARL SKOT'ISBI-RG a synonym of I. batatas. This plant is universally considered to be of Central American origin, although wild plants are not found nowadays, nor is the more precise locality known where they grew. We must consider whether the same species was not originally a native both in America and in the Polynesian legion or, whether the cultivated forms were not derived from more than one wild species, so that it is unnecessary to suppose that the sweet potato was intro- duced to Polynesia from the American coast. Yams is obtained from forms of several wild species characteristic of different continents. If this theory holds good, we should expect to find different names for the sweet potato on the two sides of the Pacific. But according to R. Lenz (Diccionario etimolojico. Santiago 1910), the word cumara is found in the Quichua language; it is not indicated as the principal name of the sweet potato, which is apichii, but nevertheless used, according to this author, for a »clase parecida* of the camote, thus for some form of the same plant. From this fact some people would conclude that, as the plant is American and called cumara by the Quichua, it was introduced to Polynesia under the same name long before the Columbian era. It is useless to discuss this matter any further till we know more of the history of the camote and also, whether the word cumara in Quichua really applies to the true sweet potato and, if such be the case, belongs to the original Quichua language or has been introduced through the ICuropeans. If old communications existed between America and Polynesia, many other proofs must be found. Much has been written about old land- bridges across the ocean, considered by some naturalists to be indispensable for the explanation of the distribution of animals and plants. But generally their existence was supposed to have ceased long before the age of Man. Only H.XLI.IKR (Uber friihere LandbriJcken, Pflanzen- und Volkerwanderungen zwi- schen Australasien und Amerika. Mededeel.'s Rijks Herb. Leiden 13, 191 2) gives them a longevity sufficent to let people march across. I am afraid that such bridges rest on a very unstable foundation. To return to the sweet potato, we have seen that it is not mentioned by Ro<;gevkkn as existing in Easter Island in 1722. CoOK and FoRSTER found it in cultivation. At that time also Broussonetia, Thespesia (also Triumfetta?) and toromiro were cultivated in addition to taro, bananas and sugar-cane. .According to tradition all of them were brought by Hotu Matua's party, the first settlers. The barahii mentioned by F. ViDAL GoRMAZ, Jeografia nautica, p. 177 (Anuario Midrogr. de la Marina de Chile, 7) is, to judge from the de- scripton, the same as the hau-hau. The calabash mentioned by THOMSON, p. 535 is Lagenaria vulgaris. At present, the following food-plants are cultiv- ated: sugar, wheat, Indian corn, taro, pineapple, yams, bananas, white mul- berry, figs, maniok, oranges, lemons, grapes, peaches, quince, plums, beans, sweet potatoes, tomatoes, melons, artichokes and lettuce, but several of these only on a very small scale and exclusively in the garden of Mataveri. Some tobacco is also grown. I do not know what Thomson means by the »two varieties of indigenous hemp», as there is no plant of this kind either in a cultivated or abandoned state. The cordage has always been prepared from the hau-hau. as far as I have been able to ascertain. Nor does Mrs. R. refer to any such plant, nor to the hau hau. NOTES ON A VISIT TO EASTER ISLAND 1 9 THE FUTURE OF THE ISLAND The power of resistance of the Easter Island people was definitely crushed through the Peruvian slave raids, and through missionaries and farmers they lost the strength which lies in the possession of an aboriginal culture. Their removal to Hanga Roa, where a village was formed, was very unlucky, as it meant giving up many small plantations and induced the people to lead a parasitic life, expecting everything from their new rulers. Although they have left so many wonderful monuments to bear witness of earlier busy days and of a people of warriors, they are now, with few exceptions, lazy beggars. In part this may be due to their pronounced feelings of animosity against the intruders, as they regard themselves as the true possessors of the island. It appears that ever since the establishment of a farming company the state of affairs has never been lucky, and Mrs. R. has an interesting tale to tell of an anxious time. I do not at all believe that the present manager is to blame, for we got the impression that he is as well liked as any white man in his position can expect to be. In Chile, nobody seems to have taken much notice of the distant colony till Bishop Edw.ards entered the field. During his first visit, in 1916, he informed himself of the state of things, and he returned in 1917 invested with powers to put everything right if he could. Among other things he wanted to take up war against the leprosy.^ Not quite 5 % of the population suffer from this disease; they are confined to a colony some distance from Hanga Roa. Apparently it is not very contagious, for the isolation is not quite effective. The surgeon of the »Baquedano», Dr. G. LONGO, examined almost every soul, but only one or two new cases were discovered. As acco- modations for the most advanced cases had been wanting, the vessel this time brought materials for the construction of a small hospital which was to be erected by the new »subdelegado» or governor. Captain MERINO carried in- structions to examine the claims against the company, and a meeting was held where the natives put forth their demands. I understood that the Company was said to have taken possession of more land than it was entitled to and that the natives wanted it to be restored. Officers went round with natives who indicated the seats of their former homes and fields, and parts of the land were measured. The scheme was, I think, that certain parts should be restored to the old owners, that the village should be abandoned, and that the natives should move into »the camp» in order to become selfsustaining. A certain amount of native labour should be granted to the manager at a fixed rate of pay. I have had ho chance to learn how far the realisation of this humanitary scheme has advanded; nor would I venture to foretell if it is likely to meet with success. EASTER ISLAND AS A FIELD OF PSEUDOGEOGRAPHICAL SPECULATION Finally, I shall make a few remarks in addition to what Mrs. R. tells us (p. 290) of the theosophists' views of Easter Island, which are based on errors ^ The surgeon of the »Mohican», Dr. Cooke, does not mention this malady as existing in the island in the year 1886. It was imported from Tahiti. 2 0 CARL SKOTTSBEKG re^^ardini,^ the cxistino^ monuments. Last year a small book appeared, entitled »lJet sunkne kontinent (Atlantis)», b\- a Xoruegian, C. SuND (Copenhagen I919), where also the supposed Pacific continent is spoken of. As might be expected, I'Lastcr Island forms an important item. With my permission, two of my photo- <,'raphs were reproduced. No doubt Mr. SUND regards himself as excused for his mistakes, for he has quoted various obscure authors; but it must be regretted that lie should not happen to draw from a single reliable source, not even from m\- popular dcscrif)ti()n, which was known to him. Mr. SUND tells us of the Iv^^vptian influence in Easter Island, of the enormous foundation walls and ruins of temples; almost ever\- mountain had sculptured designs of goods, fishes and p\'ramids, the cave paintings w:ere in the l^oltec or Egyptian style, etc. There are 300 tablets with script on the island (if it were but true!), waiting to be deciphered. On the mountain terraces are fortresses with walls up to 80 feet high. The pyramid is the architectonical principle, built as the Egj^ptiari one, even with the same kind of cement. All materials, bricks, glass, porcelain, ever>thing was known in I£aster Island; religion, symbols and habits were the same as in I\gypt, only, the culture of the island was older. There are fan- tastic groups of statues roundabout, gods of hard store with faces up to 25 feet high, in the highlands there are images on high stone pillars or staircase- like foundations, and with square hats of stone, most of them covered with script in a probably forgotten language. Round them are the remains of large walls and buildings, so they probably- stood in vast temple-yards. And so forth. No wonder that Mr. SUNI) draws the most surprising conclusions. Now, this must not be taken too seriously and will do no harm in scientific circles. The general reader, however, will get a rather curious idea of Easter Island. I dare say the place is remarkable enough in itself and need not be glorified bv such fantastic inventions. iMually, I wish to express my sincere gratitude for kind assistance to the Commander and Officers of the »General Baquedano», to Bishop RafaEL Ed- wards, Mr. I'KK( V I-:i)MiN])s, Mr. I. ViVES and Baron Erland NoRDENSKlOLD. Explanation of Plate 14. 1. StiMic ;u1/.:, no. U). i. 307; iu)t quite V-'- 2. .Stone ciiiscl, no. 10. 1. 320; alino.st ^i. V ■ >^ no. K). I. :;:i ; ,^ a,. 4. I'lsh hook of stone, no. iq. i :,2y, almost "/;. V Stone 1^11. no. 10. I. 500, not qu;?e nat. si/e. (\ Knite or scrape- no. 19. i. 31^, not ciuite nat si/e. 7. Speai-Iie.id, no. ly. i. 31 -,. •' „ ■'^- " . UO. lU. I. :;i J . -j.:. 'I'lie orii,'inaK in the .Museum. Ciothenbiii". Nat. Hist. Juan Fernandez and Easter hi. Vol. I. Plate i bo c o bC C .\'^//. I list. '7iui)i I'iDhVhIiz tvni luutir Isl. ]\>l. I. Platk 2. c (U C/2 (U hd" t c ^ r^ 2 > o p^ ■•n <\i c3 fe biO (-] C ■»-> g3 O ffi C Nat. Hist. Juan Fernandez and Easter IsL Vol. I. Plate 3. (U TD bJO n c^ rri p c; c (U •-^ J3 ^ .\(//. Hi si. '7im>/ l-ii)uvi(ii\: cDui l-.astd I si. I'd. I. Plate 4. m 0 c3 03 G 4> •'"' ^ c ^ ^ 1^ ^ '2, be *-i _•'-' o; ' > -C "^ cu 0 3 H "cJ: 13 >-( OJ c3 bJD > c3 -^ _s -1^ (U rZZ rO Vi-( OJ 0 Ji ^ 0 — < oj .s ^ (U c/J ^ 03 c 0 c3 m a trt rt t-c C/2 *>: Nat, Hist. Juan Fernandez and Easter Isl. Vol. I. Plate 5. .\(//. llist. '"fUiVi l-t))iiuhii:: auJ l.astiy I si. \'ol. I. Plate 6. / m Nat. Hist. Jtian Fcrnartdez and Easter Isl. Vol. I. Plate 7, \)uvuii:: a)id f.astir I si. I'o/. I. Plate io. Nat. Hist. Juan Fernandez and Easter Is/. ]^o/, I. Plate ii, Xd/. Hist. Juan Foiitvidcz ajid liastcr Is/. J 'of. I. Plate 12. Nat. Hist. Juan Fcj'nandez and Easter Isl. Vol. I. Plate 13. Photo by K. B'dckstrbm Same girl as in Plate 12. .\(i/. I list. '7iui// J-\)iiii)!(/i-: ivid Juis/ir Is/. I'ol. /. Pl.ATK 14. 2. Beitrage zur Geologic der Juan Fernandez-Inseln. Von TOR H. HAGERMAN. Mit 12 Textfiguren. Die Schwedische Pazifik-Expedition 1916 — 1917 unter der Leitung von Professor C. Skottsberg brachte unter anderem eine Gesteinsammlung von den Juan Fernandez- und Oster-Inseln zuriick, die dem Mineralogischen Institut der Hochschule zu Stockholm zur Bearbeitung iibergeben wurde. Bei der von mir vorgenommenen Untersuchung von etwa 50 Handstiicken von den Juan Fernandez-Inseln stand mir ausserdem Quensel's Material von einer friiheren Beschreibung^ desselben geologischen Gebietes zur Verfiigung. Die Juan Fernandez-Inseln liegen zwischen 33 und 34° S. Br., 660 km W. von Valparaiso. Die Inselgruppe besteht aus zwei grosseren Inseln, Masatierra, Flacheninhalt ca. 95 qkm und W. von derselben Masafuera, 85 qkm. Nahe der erstgenannten liegt eine kleine Insel St. Clara, 5 qkm. Der Gebirgsgrund dieser Inseln besteht ausschliesslich aus Effusivgesteinen. Eine exakte Altersbestimmung derselben kann kaum gemacht werden, da keine Sedimentgesteine vorhanden sind. QUENSEL^ nimmt an dass die vulkanischen Gesteine kaum alter als jungtertiar sein diirften, und es ist seitdem nichts be- kannt geworden, was fiir eine veranderte Auffassung sprache. Die Inselgruppe ist einer kraftigen Erosion ausgesetzt gewesen, sodass die urspriinglichen Vulkankegel nicht mehr zu erkennen sind. Besonders auf Masa- fuera, wo die Wasserscheide weit nach W. verschoben liegt, hat sich eine aus- gesprochene Canon-Landschaft gebildet. Die wilden Terrainformen sind deutlich aus den zahlreichen Photographien zu erkennen, von denen viele in Skottsberg's Reisebeschreibung^, wie auch in Vol. I und II dieses Werkes veroffentlicht worden sind. Masatierra. Die Untersuchung des Materials von Masatierra hat erneut bestatigt, was Quensel bereits hervorhebt, namlich, dass die Gesteine untereinander chemisch und mineralogisch nahe verwandt sind, und sich im wesentlichen nur strukturell voneinander unterscheiden. Sie konnen als verschiedene Erstarrungsformen ein 1 P. D. Quensel, Die Geologie der Juan-Fernandez-Inseln. Bull. Geol. Ups., Vol. XI, p. 253-290. 2 L. c. p. 256. 3 C. Skottsberg, Till Robinsonon och varldens ande. Stockholm 1918. 3—248. The Nat. Hist, of Juan Fernandez and Easter Isl. Vol. I. 2 2 TOR H. HAGERMAN unci (lesselbcn Magmas betrachtet werden; spaterhin sind allenthalben stellen- wcisc sckundare Veranderungen infolge hydrotliermaler Prozesse entstanden, die abweichende Ausbildunoen hervorgerufen haben. Die Gesteine sind im primaren Zustande durchwegs mehr oder weniger olivin- reiclie Hasalte mit der Zusammensetzung: Olivin, Pyroxen, Plagioklas, Magnetit, Ilnienit und oft ein wenig Glasbasis. Charakteristisch fur alle diese verschicdenen Teilen der Insel entnommenen Proben ist besonders die Zusammensetzung des Pyroxens. Dieser besteht aus einem Titanaugit, sofort erkennbar an seiner schwach rotvioletten Farbe und starken Dispersion der optischen Aciisen. 13och ist zu bemerken, dass derselbe nicht zu den extremsten Typen gehort. Audi die Felds{nite weisen eine konstante Zusammensetzung auf. Sie sind fast aile zwillingsgebildet nach dem Albit- oder Karlsbader-Gesetz. In Schnitten senkrecht zu M zeigen die Albitlamellen eine max.-Auslosung von 32 — 33°, in einigen iMnzclfallen diese W'erte mit hochst 2° variierend. Der Feldspat ist also ein Labrador von eincr Durchschnittzusammensetzung Ab42An58. Da die Hasalte also mineralogiscli einander nahe vervvandt sind, wurden sie liauptsaclilich nach der Struktur in folgende Typen eingeteilt: basaltische Laven, (teils dichte, teils grobkornig doleritische, teils schlackige) Tuftc, hydrothermale Um\vandlungS[)rodukte. Basaltische Laven. Diese (iesteine bilden, wie friihere Verfasser bereits betont haben, den Haupt- bestandteil der Insel. Auf Grund des vorliegenden Materials konnte man den X'erlauf der verschicdenen Lavastrome und deren Neigungsverhaltnisse nicht be- stinimen und so auch keine Klarheit iiber die P>uptionsstellen erlangen. Die Fundstiitten der dem Verfasser zur Untersuchung vorliegenden Hand- stucke sind ziemlich gleichmassig iiber die Insel verteilt. Von den zu den Laven gchorenden (iesteinsproben sind ungefahr ^/^ sehr porose und schlackige Typen, wahrend ^3 (6 Stck.) dichte, dabei gleichmassigere und feinkornige Gesteine darstellen. Die Struktur dieser Gesteine ist im allgemeinen hypokristallin porphyrisch. In einigen I-'allen, besonders bci den dichten Typen, kommt es vor, dass Glas- basis ganz fehlt. Mit Bczug auf die mincralogische Zusammensetzung der basaltischen Laven sind folgende Mineralien beobachtet worden : Olivin nebst dessen Umwandlungs- produkten, die vorgenannten Pyroxene und Plagioklase sowie ILrzmineralien (Mag- netit und Ilmenit). Die I'Linsprenglinge sind Olivine und ihre Umwandlungsprodukte sowie i'eldspat. Die Augitkorner konnen sich zuweilen der Grosse der Einsprenglinge naiiern. Die Olivineinsprenglinge erreichen ihre grosste durchschnittliche Ausdehnung, ca. 2 mm, in Proben, die von dem nordlichen Ufer der Padrebucht stammen. Gerade diese ICinsprenglinge zeigen meistens in einer scharf begrenzten Zone die von Oll.NSF.i.' fruher beschriebene Iddingsitumvvandlung. Aus Dunnschlififen der genannten (icsteine ist klar ersichtlich, wie die Umwandlung von den Ran- ' L. c. p. 260. BEITRAGE ZUR GEOLOGIE DER JUAN FERNANDEZ- INSELN 23 dern nach der Mitte zu ausgegangen ist und sich 0,02 — 0,03 mm in den Olivin hineinerstreckt hat. An manchen Stellen ist die Umwandlung langs der Spalt- risse vor sich gegangen, wahrend anderweitig die Durchgange merkwiirdigerweise vollkommen unveranderte Teile der OHvinkerne durchqueren. (Vergl. Fig. i.) In der Gesamterscheinung des obengenannten Praparates mochte der Verfasser die Iddingsitbildung als das Resultat einer von aussen kommenden chemischen Beeinflussung ansehen. WASHINGTON^ verweist die Iddingsitbildung bis auf mag- matischen Ursprung zurijck. Man braucht vielleicht nicht so weit zu gehen, da dieselbe ebensogut einer hydrothermalen Umwandlungsperiode zugeschrieben wer- den kann. Fig. I. An den Randern iddingsitumwandelte Olivine. Olivinbasalt von der Padrebucht. — Vergr. 56 x . Photo Hj. Olsson. In einem anderen Gestein von der Padrebucht, das etvvas feinkorniger ent- wickelt ist, treten auch Feldspateinsprenghnge von 0,5 — i mm Lange auf, neben Resten von Olivinen. Der Olivinumwandlungsprozess hat hier ein anderes Pro- dukt hervorgebracht, namUch gewohnHchen Serpentin. Dies ist auch bei den der Grundmasse angehorenden OHvinkornern der Fall. Dieselben sind vollkommen als Serpentin ausgeflossen. Auch der Augit scheint an den Randern etwas ange- grififen zu sein. Beinahe identisch entwickelt sind zwei andere feinkornige Basalte, der eine von der Mitte des siidlichen Ufers, am Fuss des Yunque, der andere von einem Gebirgsriicken (385 m ii. M.) SW. von Tres Puntas, W. von der Villagra- Bucht, herstammend. Iddingsit- und Serpentinumwandlungen der Olivine sind, wie ich besonders hervorheben mochte, niemals in ein und demselben Gestein gleichzeitig angetroffen worden. Die beiden erwahnten Gesteine von der Padrebucht und an beiden Seiten von Villagra zeigen eine eigentiimliche primare Struktur der gleichkornig ent- wickelten Grundmasse, indem die Plagioklase mit einer mittleren Ausdehnung 1 Italian petrogr. sketches. Journ. Geology. 4 (1896), p. 835 — 836. 24 TOR H. HAGERMAN von 0,2 mm sich vielfach radialstrahlig mit den Augitindividuen geordnet haben. 1st man erst einmal auf diese spharolitahnlichen Bildungen aufmerksam gewor- den, so findet man sie haufig hauptsachlich in den feinkornigsten Proben dieser Gesteine ausgebildet. Besonders schone Beispiele hiervon zeigt ein Handstiick, das der »ryramides einem Ikrggipfel ungefahr in der Mitte der Insel unvveit des SKIA'IKK-Denkmals, entnommen wurde. (Fig. 2.) lune sehr ahnliche Krscheinung ist von Reiter^ beschrieben worden. Er sclimolz 45% Albit, 45% Augit und 10% Magnetit zusammen. »Der Schliff V /r Pyroxen (-^^">PUPyramide». — Vergr. 250 x Zeichnung vom Vcrf. eincr durcli 7 Stundcn abgckiihlten Schmelze zeigt eine sphariodale Anordnung der Kristalle, in dem niagnetitreiche Kerne von Glaspartien mit einzelnen aus- gcschiedenen Augit- und Plagioklasleisten und Kristalliten umgeben sind. Die W'iederhokmg des X'crsuchcs bei 30-stundiger Abkiihlung ergab eine Schmelze mit kornig-pori)hyrisclicr Struktur. Zweifelsohne ist dies audi in dem vorliegenden l^'alle zutrcffeiid, indcm die s{)harolitrulirenden Laven einer raschen und ungestorten Abkuhlung ausgcset/.t gewesen sein durften. In Anschluss an dicse Spharolite seien hier die in Pig. 3 abgebildeten kreuz- formig licgenden Olivinkristalle erwahnt. Der DiinnschlifT entstammt einem etwas grobkornigen Gestein von Bahia Cumberland. Wie aus der Figur deutlich her- ^ U. H. Rkii F.R, Kxporimentellc Studien an Silikatschmelzen. Neues Jahrbuch. Beil. Bd. 22 1906 , p. 197. BEITRAGE ZUR GEOLOGIE DER JUAN FERNANDEZ-INSELN 25 vorgeht, handelt es sich urn eine skelettartige Ausbildung der Kristallindividuen. Eine gesetzmassige Verwachsung der verschiedenen Individuen habe ich nicht nachweisen konnen. Im Vaqueriatal tritt, wie aus der untenstehenden Photographic (Fig. 4) er- sichtlich, ein fast horizontal liegendes Gestein auf. Nach ihrer grobkristallinischen Struktur zu urteilen, sind diese Basalte moglicherweise als intrusiv aufzufassen> Das nur an zwei Seiten zugeschlagene Handstiick ist vorziiglich durch Schrump- fung unter rechtem Winkel zerkluftet. (Fig. 5.) U. d. M. zeigt dasselbe ein un- verandertes hochkristallinisches Aussehen. Reichlich albitlamellierte Feldspatleisten, durchschnittlich ca. 2 mm lang, bedin- gen mit Olivin- und Titanaugitkristallen eine ophitische Struktur. Ausser diesen Mineralien habe ich nebst Magnetit hier und da ein Biotitkorn gefunden. Moglicher- weise erstreckt sich dieser Basalt bis zur Cumberland Bay, wo eine ahnliche Aus- bildung von QUENSEL^ beschrieben wurde. Auch das Material Skottsberg's enthalt eine ahnliche Probe von dort, in losem Block gefunden. Der Mineralbestand der beiden letztgenannten Handstiicke ist der- selbe wie jener der Vaqueriaprobe nur mit dem Unterschied, dass kein Biotit vor- handen ist. Auf Grund seiner grobkristallinischen Struktur muss das obenerwahnte Gestein zu den doleritischen Basalten gerechnet werden. Feinkorniger, aber im iibrigen dem vorgenannten Gestein vollig gleich, ist der bei »Tres Puntas» genommene Basalt. Die Handstiicke bestehen aus langen, schmalen, dreiseitigen Prismen. Fig. 3 a. Zentrisch angeordnete Olivine. Vergr. 56 x . Verf. phot. Unter den schlackigen Laven weisen einige eine auffallende Analogic zu rezenter Oberflachenbildung auf. Besonders ist dies der Fall bei einem sehr po- rosen glasreichen Gestein vom Ufer s. von Yunque. Der vorerwahnte, auf dem Gipfel der »Pyramide» befindliche dichte Basalt hat ein schlackiges und glasiges Lavabett als Unterlage. In einigen anderen der schlackigen Gesteine sind die Locher mehr oder we- niger mit Opal, Chlorit, Serpentin und Calcit ausgefiillt. 1 Vergl. jedoch Quensel p. 263—264. 2 L. c. p. 263. 26 TOR H. HAGKRMAN Tuffe. Unter dem mitgebracliten Material befinden sich zwei Proben von ausge- sprochenen Tufien. Der eine, ein poroses, dichtes Gestein, stammt von El Puente, dem Istmus zwischen der Padrebucht und Carbajal und ist ein Palagonittuff mit einigen sporadischen Augit- und Magnetitkornern. Der andere Tuff stammt von dem nordlichen Ufer der Padrebucht, von wo einige umgevvandelte Olivinba- salte (s. S. 23) herriihren. Ausser Augit und Magnetit enthalt derselbe einige grossere vollkommen reine Olivinkorner in einem teilvveise kryptokristallinisch aus- sehenden Glase. Stellt man diese verschiedenen Bildungen aus der Nahe der Fi^. 3 b. Detail von Fig. 3 a.— Verg^r. 170 x. Verf. phot. Padrebucht zusammen, so gelangt man zu der Auffassung, dass dieses Gebiet frische Spuren vulkanischer Tatigkcit aufweist. Vergleicht man die obengenannten Tuffe mit den von OlKNSKl.' bescliriebenen roten Tuffen von der Cumberland Bay, so scheinen die letztgenannten nicht so empfindlich gegen Verwitterung zu sein, wie besonders die Palagonittuffe. Hydrothermale Bildungen. K\n aragonithaltiges (iestein vom Ufer gleich siidlich vom Yunquegipfel diirfte als liydrothermal umgewandelter Basalt angesehen vverden. (Fig. 6.) Das Hand- stuck ist ein von weissen Streifen durchzogenes scharfgriines Gestein, das u. d. M. 1 L. c. p. 266. BEITRAGE ZUR GEOLOGIE DER JUAN FERNANDEZ-INSELN 27 grosse Augitkristalle in einer vollig zerflossenen Serpentinmasse zeigt. Das Pra- parat ist von Aragonitbandern durchzogen. Dieses Gestein muss als ein stark umgebildeter Olivinfels bezeichnet werden. QUENSEL^ hat ganz frische Gesteine von letztgenanntem Typus angetroffen und beschrieben. Vielleicht kann die Ara- gonitbildung hier eine Andeutung geben, auf welche Weise der Olivin chemisch umgewandelt worden ist. Auf derselben Stelle wurde auch eine reine Kalksinterbildung gefunden, was darauf hinweist, dass diese Gegend in spaterer Zeit postvulkanischen Prozessen hydrothermaler Natur ausgesetzt gewesen ist. In Fig. 7 ist eine eigentumliche Bildung dargestellt, wie sie auf dem offenen Fig. 4. Das Vaqueriatal. Wasserfall iiber den saulenformig abgesonderten doleritischen Basalt. Plateau bei Puente vorkommt. Es sind lange, in einem »Sandfeld» aufrechtste- hende rohrahnliche Bildungen, Aragonit, Pyroxen, Magnetit sowie etwas Olivin in einem Zement von Karbonat enthalten. Die wahrscheinlichste Deutung dieser Phanomene ist wohl, dass mit Calciumkarbonat gesattigte thermale Gewasser iiber eine Vegetationsdecke geflossen sind, wobei Wurzeln etc. mit einer Kruste von oben angegebener Zusammensetzung iiberzogen wurden. Santa Clara. An dem Siidende von Masatierra liegt die kleine Insel Santa Clara. Von Santa Clara selbst ist keine Probe mitgebracht, dagegen von der kleinen Insel Morro de los alelies, die bei tiefstem Wasserstand mit der Hauptinsel zusammen- 1 L. c. p. 265. 28 TOR H. HAGERMAN hangt. Sowohl in Diinnschlifif wie in Handstiick zeigt dieses Gestein, das als fast vertikale Gange auftritt, eine voUige tjbereinstimmung mit einem der dichten Basalte von Masatierra vom Gipfel des Cerro Negro SO von Yunque, 190 m ii. d. M. Ur- spriingliche Plagioklaseinsprenglinge von bis zu 2 mm Lange sind oft so stark kaolinisiert, dass beim Schleifen nur die Hohlraume iibrig geblieben sind und dem Gestein ein falsches, schlackiges Aus- sehen verleihen. Hierbei ist interessant, dass der Feldspat der Grundmasse sich frisch beibehalten hat. Die Olivine sind natiirlich vollig in Serpentin umgewandelt. Das ganze Praparat ist von Ilme- nitskeletten durchwachsen. Masafuera. I'ig. 5. Handstiick voin Vaqueria K'ang. Verf. phot. Etwa 180 km vvestlich von Masatierra erhebt sich die Insel Masafuera. Abweichend von Masa- tierra in Bezug auf die einheitliche Mineralzusammensetzung der Gesteine Uefert Masafuera Beispiele petrographisch weit verschiedener Typen. Basalte mit den dazugehorenden Gangformen von ungefahr gleichem Mineralbestand wie die auf Masatierra vorkommenden gibt es zvvar auch hier, ausserdem finden sich aber auch an Erzmineralien stark iibersattigte Basalte, sowie den Trachytandesiten sich nahernde Gesteine. Am interessantesten ist jedoch das Vorkommen von reinen Alkaligesteinen, wie z. B. die von QUENSEL angefiihrten Natrontrachyte. Natrontrachyt. Leider ist das einzige mitgebrachte Handstiick dieses Gesteins von einem losen Block am Fuss der Steilwand von Tierras Blancas abgeschlagen. Skotts- HER(; hat indessen miindlich berichtet, dass zahlreiche Blocke desselben Gesteines in den Talusbildungen von Tierras Blancas vorkommen, und dass, soweit er ver- stehcn koniitc, dasselbe hellgraue Gestein den ganzen oberen Teil der Steilwand bildct; seiner Kartenskizze nach zu urteilen tritt dasselbe bereits 400 m ii. d. M. auf. Dies ist von Bedeutung fiir das Feststellen der Eruptionsfolge, die spater kurz erwalint wcrden soil. Das (iestein besteht aus gleichmassigen Kornern und ist sehr reich an Feld- spat. Die I^'eldsj)atleisten erreichen eine Lange von 0,3 — 0,4 mm. Sie sind gut parallclorientiert und verleihen dem Gestein eine trachytoidale Struktur. Da Albit- zwillinge nicht vorhanden sind und der Feldspat durch die Anlagerung der diinnen Individucn unscharfe Bcgrenzungen zeigt, konnte eine genaue Bestimmung des- selben nicht ausgefuhrt vverden. Die Lichtbrechung halt sich im allgemeinen etwas liber Kollolith (n 1,535), stellenweise ist das Relief jedoch ganz verschwunden. Urn eine nalierc Kcnntnis von den Feldspaten zu bekommen, ist eine Alkalibestim- mung des Gesteins ausgefuhrt worden. Diese ergab 3,45 % KgO und 7,34 % NagO. Dies wiirdc einem Gehalt von 20,44 % Ortoklas und 62,27 % Albit im Gestein ent- sprechen. Auf (irund der Lichtbrechungsverhaltnisse diirfte jedenfalls neben einem Kali-Xatronfeidspat auch cin saurer Plagioklas der Oligoklasreihe vorhanden sein. BEITRAGE ZUR GEOLOGIE DER JUAN FERNANDEZ-INSELN 29 Eine geringere Menge Pyroxen tritt ebenfalls auf. Die durchschnittliche Ausdehnung desselben ist 0,1 mm, die Farbe ist gelbbraun, die kristallographische Ausbildung schlecht entwickelt. Eine Ausloschung von c:x = 40° (ungef.) deutet auf Augit. Schliesslich war auch Magnetit vorhanden, der ofters fliessende Be- grenzung der graubraunen Glasbasis gegeniiber zeigt. ..— Aragonitband, von einem Augitindivi- duum iiberquert Fig. 6. Hydrothermal umgewandelter Olivinfels von dem Yunque. — Vergr. 12,5 x, Verf. phot. J 9 m i^fe' ^■^M ---^r '^:f^ y • « ■ . ' z , . . . . * Fig. 7. Sinterbildung von Puente. Massstab in Cm. Verf. phot. Der von QuENSEL^ beschriebene Natrontrachyt ist ungefahr 1200 m ii. d, M. gefunden worden und zeigt bei einem Vergleich einige Abweichungen. Das Hand- stuck scheint im Gegensatz zu dem obenbeschriebenen etwas verkieselt zu sein. U. d. M. bemerkt man sofort, dass der Pyroxen hier abweichend von dem vorer- wahnten farblos ist. QUENSEL bezeichnet denselben als Diopsid. Ferner ist die farbige Glasbasis nicht vorhanden. 1 L. c. p. 283. 30 TOR H. HAGERMAN Andesit. Von (lem Berggipfel »Las Torres >, 1 370 m ii. d. M., und in losen Blocken von dem Ik'rgrucken XO von der genannten Stelle 1 200 m ii. d. M. sind Hand- stiicke von einem feldspatreichcn aschgrauen Gestein mitgebracht worden. Obwohl etwas olivinreicher, stelien dieselben den von Olensel^ beschriebenen, einem 1 100 m ii. d. M. liegenden Niveau entnommenen Trachytandesiten sehr nahe. Als Kinsprenglinge koniinen Feldspat und Olivin vor, u. d. M. zeigt der erstgenannte eine Zusammensetzung von AbggAnga und erreicht eine Korngrosse von 3 — 4 mm. Die Olivine sind etwas kleiner. Sie werden im allgemeinen nur ca. i mm in Diameter und sind wenig verandert, nur an den Randern zeigt sich eine schwach gelbe Farbe, wo die Umwandlung begonnen hat. In der Grundmasse dominiert der Plagioklas mit einer Ausdehnung von ca. 0,1 mm. Die Pyroxene und Magnetite sind noch kleiner, im allgemeinen nur 0,05 mm. Die Magnetite sind vollig idio- morph. Die Klassifizicrung hiehergehorender Gesteine ist etwas unsicher. Die trachytoidale Struktur konnte auf einen gevvissen, nicht wahrnehmbaren Alkali- gelialt in der (irundmasse deuten. Geniigende Griinde, sie als Trachytandesite zu bezeiclinen, liegen jedoch nicht vor. Sehr iiiteressant ist ein bei Correspondencia (1420 m ii. d. M.) genommenes (iestein. Mikroskopisch zeigt dieses Handstiick ein aschgraues, porphyrisches Aussehcn, doch sind einige Partien bedeutend dunkler. Als Kinsprenglinge kom- men 1^'eldspat und Olivin vor. U. d. M. zeigt der erstgenannte eine Zusammen- setzung Ab42An58 und erreicht eine Korngrosse von durchschnittlich 2 mm bei einer niax.-Lange von 5 mm. Der Olivin ist vollig frisch, die Kinsprenglinge treten aber in zwei verschie- dencn ICntwicklungen auf. Dies steht im Zusammenhang damit, dass das Gestein, wie bercits erwahnt, nicht vollig homogen ist. Die dunkleren Partien erweisen sich bei mikroskopischer Untersuchung als bedeutend magnetit- und ilmenitreicher als der iibrige Teil des Gesteins. In den dunkleren Schlieren finden sich nun Olivine mit Magnetiteinschliissen vollgesteckt. Iksonders an den Randern ist der Magnetit so reichlich vorhanden, dass die Olivinkorner vollig opak sind. Sowohl aus diesem Grunde als auch in- folge der abgerundeten Form der Mineralkcirner scheint es, als ob diese Olivin- korner einer kraftigen Resorption ausgesetzt gewesen waren. Naheliegend ist nun, dass diese dunkleren Schlieren mit ihrem grosseren Kisengehalt Riickstande auf- geloster Hruchstucke sind, welche urspriinglich zu dem Typus gehorten, die einem schlackigen l^asalt vom (jipfel des Inocentas entsprechen, der spater beschrieben werden soil. In einer luitfernung von kaum i^/g mm von einem der erwahnten Olivin- korner treten Individuen des anderen Typus auf. Diese sind ganz einschlussfrei, vollig idiomorph mit scharfen Hegrenzungsflachen, erreichen einen Durchschnitt bis zu 4 mm und ents{)rechen vermutlich der intratellurischen Olivingeneration des Ilauptgesteins. Der Feldspat in der Grundmasse der helleren Schlieren tritt in Stengein von ca. 0,14 mm Lange auf und verleiht durch seine Parallelorientierung dem Gestein eine Fluidalstruktur. Die Zusammensetzung desselben ist Ab46An54. Im iibrigen enthalt die Grundmasse Magnetit und Pyroxen. Die (irundmasse der dunkleren Partien unterscheidet sich von der obenge- ^ L. c. p. 282. BEITRAGE ZUR GEOLOGIE DER JUAN FERNANDEZ-INSELN 3 1 nannten durch ihren Gehalt an Ilmenit, leicht erkennbar an seinem tafelformigen Habitus. Ausserdem kommen hier feine Nadeln vor, welche aus einem ziemlich stark lichtbrechenden Mineral bestehen. Dasselbe ist pleochroitisch von braun- gelber bis gelbgruner Farbe und weist parallele Ausloschung auf. Wegen der kleinen Dimensionen der Korner konnte eine sichere Bestimmung derselben nicht ausgefuhrt werden. Mit grosster Wahrscheinlichkeit liegt hier nur eine feinblattrige Ausbildung von Ilmenit vor. Basaltische Laven. Auf dem Uferplateau an der Ostseite der Insel bei dem Casastal steht ein feinkorniger Basalt mit porphyrischen Feldspat- und Olivineinsprenglingen an. Die Feldspate erreichen eine Lange von 0,9 mm und erweisen sich als Plagio- klase mit einer Zusammensetzung von Ab4oAn6o und stimmen also mit dem Feldspat der Masatierra-Basalten iiberein. Die von QuENSEL^ erwahnten, stark basischen Feldspatkerne habe ich nicht angetroffen. Die Olivinkorner erreichen in diesem Praparat eine Grosse von 0,3 — 0,4 mm und sind etwas iddingsitumgewandelt. Von den Mineralien in der Grundmasse werden die Plagioklase am grossten, 0,08 mm. Der Pyroxen ist dagegen so klein, dass eine nahere Bestimmung sich nicht aus- fiihren Hess. Er erscheint in kleinen, viereckigen, farblosen Kornern, meistens zusammen mit dem Magnetit. Diesem Gestein sehr nahe verwandt ist dasjenige, welches am Ufer des Mono- Tales ansteht. Makroskopisch sind die beiden Gesteine einander sehr ahnlich. U. d. M. tritt jedoch ein Unterschied auf, und zwar indem die Grundmasse des Monobasaltes hier bedeutend mehr Olivin enthalt, weshalb man dieses Gestein auch wegen der zahlreicheren Olivineinsprenglinge als einen Olivinbasalt bezeichnen muss, wahrend sich das erstgenannte den Feldspatbasalten nahert. In Quebrada del Ovalo, ungefahr 150 m ii. d. M., steht eine saulenformige Basaltkuppe an, die dem Tal seinen Namen gegeben hat. Sie ist oben horizon- talzerkliiftet, wahrend weiter unten eine prismatische Vertikalzerkliiftung ansetzt. Vorausgesetzt, dass die Kuppe aus ein und demselben Gesteine besteht, kann dies als ein gutes Beispiel der von Iddings^ dargetanen Veranderung in der Richtung fiir den kleinsten Schrumpfwiderstand innerhalb des erstarrten Gesteins angesehen werden. Die Handstiicke sind den unteren Teilen der 20 — 30 m hohen Saule ent- nommen und bestehen aus einem porphyrischen Olivinbasalt. Die Einsprenglinge sind Olivine von ca. 0,6 mm Durchmesser. Dieselben sind an den Randern dunkel- rot, kaum durchleuchtend, was wahrscheinlich durch einen Gehalt an freiem FcgOg verursacht wird. Untenstehende Photographic (Fig. 9) stellt einen dieser iibrigens sehr sparlich vorkommenden Einsprenglinge dar. Die Grundmasse ist der des vorstehend erwahnten Olivinbasalts voUig gleich, nur etwas grober. Der Plagioklas, ein Labrador, erreicht eine Lange von ca. 0,12 mm, das ganze Praparat ist parallelorientiert. Besonders hervorgehoben sei, dass das Olivin in der Grundmasse nicht rotpigmentiert ist. Im iibrigen findet sich Magnetit und der farblose Pyroxen. 1 S. p. 276. 2 J. P. Iddings, The columnar structure in the igneous rocks of Orange Mountain, N.J. Amer. Journ. 31 (1886), p. 321. 32 TOR H. HAGERMAN Der hochste Berg auf Masafuera ist der Inocentes. Von dem Gipfel, ca. 1 500 m ii. d. M., wurde ein rotlicher, sehr schlackiger porphyrischer Basalt mitgebracht, der sich besonders durch seinen hohen Gehalt an Erzmineralien auszeichnet. Unter dem Mikroskop erwies sich die Grundmasse als hyalopilitisch. Die Einsprenglinge bestehen hauptsachlich aus grossen, im ersten Augenblick voll- standig opaken Kornern von ca. 3 mm Durchmesser. Bei genauerer Untersuchung erweisen sich jedoch mehrere als stellenweise durchsichtig. Diese Teile besitzen Fig. 8 a. Das heterogene Ciestein von Correspondencia. Dunklere Partien oben und unten sichtbar. — Vergr. 10 x. Photo. E. Dahlstrom. die hohen Interferenzfarben des OHvins, dazu weisen die Korner auch im iibrigen den Habitus des Oiivins auf. Bei Beobachtung in konvergentem polarisiertem Licht bei einem Achsenaustritt wurde keine Krummung des Achsenbalkens beob- achtet, vveshalb der Achsenvvinkel nahezu 90° sein durfte. Dies deutet auf cine fur liasahe normale Zusammensetzung von OHvin mit verhaltnismassig niedrigem Eisengehalt. Kand. S. L.\M)E1<(;kkn fuhrte eine Eisen- und Titanbestimmung der Ge- steinsarten aus, vvobci folgende Werte erhalten wurden: 18,82% FeaOg und 3,59 % TiOg. Dies entspricht einem Gehalt von 6,81% Ilmenit und 1 5,44 % Magnetit. Diese Zahlen sind natiirlich zu hoch, da Fe auch in den geringen Mengen Olivin, Pyroxen und Glasbasis enlhalten ist, die sich in dem Gestein befinden. Es ist BEITRAGE ZUR GEOLOGIE DER JUAN FERNANDEZ-INSELN 33 jedoch besonders interessant, dass, obgleich das Magma eisenubersattigt war, das Eisen die Konstitution des Olivins nicht nennenswert beeinflusste. Von solchen magnetitijbersattigten Olivinen spricht Doss^ in einer Beschrei- bung von Basalten aus Syrien: »In diesen Gesteinen beherbergt der porphyrische Olivin eine derartige Menge von Magnetitkornern, dass dieselben meist die Halfte, zuweilen ungefahr Vs des ganzen Kristalldurchschnittes einzunehmen scheinen». An einer spateren Stelle schreibt er: »Das Extrem hiervon tritt dann ein, wenn der 6V\ Fig. 8 b. Detail von Fig. 8 a (bei — ). — Vergr. 210 x, Zu beachten ist die parallele Anordnung des Magnetits. Photo. E. Dahlstrom. Olivinkrystall einen breiten, vollig opaken, schwarzen Saum von Magneteismen besitzt». Das Phanomen stimmt mit dem vorliegenden voUkommen uberein, nur dass bier der entgegengesetzte Fall vorliegt, indem das ganze Korn voUstandig opak sein kann, auch in sehr diinnen Praparaten. Reiter2 hat sich mit der Zusammenschmelzung von Olivin und Magnetit beschaftigt und sagt: »Bei Abkuhlung tritt eine gewisse Ubersattigung ein. Vom Magnetit scheidet sich ein Teil ab, dann wird der ubersattigte Olivin ausge- schieden . . . 1 B. Doss, Die basaltischen Laven und Tuffe der Provins Hauran und vom Diret et-TuluI in Syrien. T. M. P. M. Bd. 7 (1886), p. 483—484. 2 L. c. p. 232. 34 lOR H. HAGKRMAN Wir erhalten hiemit also Magnetite in zonarer Anordnung in den nachher entstehentlen Olivinen ein^esclilossen; ...» Die umstehende Tliotographie (Fig. lo) zeigt, wie der Magnetit parallel an- geordnet ist, sodass die durchsichtigen Partien in Streifen auftreten. Unter den I^insprenglingen komnit, wenn auch selten, hier und da ein Feld- si)atindividiiiim \ or. dessen Zusamniensetzung ungefahr Ab4oAn6o ist; es handelt sich also um einen Labrador, unbedeutend abweichend von dem Feldspate der ubrigen (iesteinsarten. Die basischen Tlagioklase der Grundmasse zeigten zum Teil unscharfe Be- grenzungen. durcli die Anlagerung der diinneren Individuen verursacht. Die Pyroxene sind bedeutend kleiner und erreichen einen Durchschnitt von ca. 0,02—0.03 mm, wenn auch einzelne stengelige Individuen etwas grosser werden -0*^3^ r/iA Fi •'K- 9. Kotj)igmentierte Olivineinsprenglinge. Basalt von El Ovalo. — Vergr. 56 Photo. Hi. Olsson. konncn. \'on diesen kommen zvvei verschiedene Ausbildungen vor. Der eine (nor- male) ist mit den in den beschriebenen basaltischen Gesteinen auftretenden Py- roxen indentiscli und vermutlich als Diopsid zu betrachten, da er beinahe farblos und kaiun pleochroitisch bei einer Ausloschung von ca. 45° ist. Der andere Typus ist braimgelb und tritt besonders uin die Locher herum, niemals zusammen mit dem erstgenannten Tyjius atif. stellt aber wahrscheinlich nur eine Pigmentierung dcsselbeu dar. Auch hierin zeigt das Gestein eine grosse Analogic mit den vor- erwahnten. von Dos.s beschriebenen I^asalten. \\x sagt auf Seite 481: »Hier be- sitzen die in der Xahe der von Kalkspat ausgefiillten Hohlraume gelegenen Augite eine goldgelbc I'\-irbc; . . . . Im iibrigcn bcsteht die (irimdmasse aus kleinen, idiomorphen Magnetitkornern, Ilmenit luul Glasbasis. Der ganze (iebirgskamm soil aus diesem ausgesprochen basischen Gestein l;cstehen. I^ereits mikroskopisch zeigt ein anderes Handstuck, das vom Strandabhange nahe des Casatales abgeschlagen ist, eine auffallende Ahnlichkeit mit dem Ino- BEITRAGE ZUR GEOLOGIE DER JUAN FERNANDEZ-INSELN 35 centesgestein. (Laut Angabe von Skottsberg soil das ganze ostliche Ufer aus diesem Gesteinstypus bestehen, abwechselnd mit den obenerwahnten Feldspat- basalten.) Auch hier treten magnetitfuhrende Olivine auf, wenn auch sparlicher. Die Feldspate sind dieselben und besteht der einzige Unterschied darin, dass das Gestein glasreicher ist, weshalb der in dem letztgenannten Gestein vorkommende Pyroxen nicht mit Bestimmtheit hat wahrgenommen werden konnen. Nimmt man an, dass dieses Gestein einer rascheren Abkiihlung ausgesetzt gewesen ist als das vorgenannte, kann man dasselbe diesem gleichstellen und moglicherweise eine Andeutung iiber eine nicht unbedeutende Ausbreitung der fraglichen stark basischen Gesteinsart finden. Fig. 10. Magnetitiibersattigte Olivinsprenglinge im Basalt von Inocentes. Verf. phot. Umstehende Tabelle ist eine Zusammenstellung iiber die Veranderlichkeit der efifusiven Masafuera-Gesteine im Verhaltnis zum Niveau der Fundstatten. Die Gesteine sind nach abnehmender Basisitat geordnet und dabei das Material Skotts- BERG's wie auch dasjenige QuENSEL's beriicksichtigt. Zwar fallen die Lavabetten etwas nach NNO ab, doch diirfte mit Riicksicht auf den geringen Umfang der Insel die umstehende Tabelle ein gutes Bild der Eruptionsfolge gevvahren, wobei die altesten Gesteine beim Meeresniveau, also links beginnen. Vor allem verstosst der stark basische Basalt aus der Hohe von 1500 m ii. d. M. gegen die unter Zugrundlegung der Eruptionsfolge vorgebrachte Theorie, dass die Gesteine hier gravitativ differentiert sein sollen. Die Gesetzmassigkeit, die QUENSEL bei diesen Gesteinen gefunden zu haben glaubte — mit den sauren und alkalireichen als den jiingsten — scheint mir schwerlich mit diesen Beob- achtungen in Ubereinstimmung gebracht werden zu konnen. 36 TOR H. HAGERMAN 1 Natrontrachvte . . . ! i ! o ! i 1 1 X i 1 1 1 1 1 O O 1 Basanit 1 X j^^^^l^^FeUispat ^q i '\ 1 Ohvin- i oo O 1 i Er/iibersatti^ae Ha salte 0 i 1 O 13 14 1500 m ii. d. M. Die \'erteilung der Effusivgesteine der InseJ Masafuera. O bezeichnet Das Material Skottsberg's. X » » , QUENSEL's. Ks ist selbstverstandlich denkbar, dass die in den tieferen Teilen des Zufluss- kanales versunkcnen schweren Magmas zuletzt herausgestossen wurden, doch spricht gcgcn diese Annahme das bereits in einer Hohe von 400 m ii. d. M. auftretende alkalisciie Ciestein. Scliliesslich ist aus der Tabelle zu ersehen, dass aus einer Hohe von 500 — 1000 111 Liberhaupt kein Material untersucht ist. Die Masafuera Gesteine sind jcdenfalls in Minblick aiif den Mineralbestand so verschiedenartig, dass eine nahere L'ntcrsuchung der noch unbekannten Hohenlagen wichtige Aufschliisse ergeben durfte. /'»'j-/ ^7-intcd February 2^th, ig2^. Reprinted ivithout change April 13th, igjj4. 3- Additional Comments on the Geology of the Juan Fernandez Islands. By PERCY OUENSEL. Contents. Page Introduction ^^ Main Geological Features 40 Petrology of the Volcanic Formations 44 Masatierra 4^ Masafuera 56 Regional Relations. Tectonic Connections -74 Petrographic Connections 77 General Conclusions 79 Acknowledgements 82 Tables of Analyses 83 Bibliography 86 Introduction. The Juan Fernandez Islands consist of Masatierra, Masafuera and Santa Clara. In many publications the name Juan Fernandez has been used to denote Masatierra only, the other islands then being indicated by the names as above. The largest" island, Masatierra, situated 660 km from and nearly due west of Valparaiso, measures 95 square km. Masafuera, 170 km further westward, is 64 square km. The small island Santa Clara, close to Masatierra, is only 5 square km in area. During more or less casual visits to Juan Fernandez stray observations have been recorded on geological features of the islands. In all cases they refer to Masatierra. It may be of interest in this connection to give a summary recount thereof. The first samples of volcanic rocks of Masatierra were, as far as known, collected by Lord CocilCRANE in January 1823. He was returning to Europe after 5 years service as admiral in Chilean service during the war of independ- ence. The ship at his disposal, 'Colonel Allen', touched at Masatierra for two days. Mrs Maria Graham, a passenger on board, has in her diary given the following details, recorded by Thomas Sutcliffe in his book Cmsoniana: "Lord 3 — 516795 The Nat. Hist, of Juan Fernandez and Easter Isl. Vol. I. 3S IMKlV (JIKNSKL Cochcrane brought from the siiniiiiit (1.500 feet) a piece of black porous lava; and uiuler it lie foiiiul some dark iiardened clay full of cells, the inside of which appear sli^litly \ itritk-d. The ishuul seems chiefly composed of this porous lava; the strata of which, beini; crossed at ri^ht an^^les by a very compact black lava, dij) on tile eastern side of the island about 22' and on the west side l6°, pointing to the centre of tiie island as an apex" (l, j). 198). In 1S30 C\ Hi:kii:r<» j^ublished some observations under the title 'Notice sur rilistoire naturelle de file juan hV-rnandez'. With regard to geological ques- tions he sa\s: "je pense ciu'un geologue n'y trouverait cjue du basalte dans les etats, lueiue dans cehii de la j)lus parfaite decomposition; plusieurs blocs sont parsemes dune cristallisation j)articulicre, a laquelle on donne, je croix, le nom d()li\ine ... 11 n'\- a pas de trace de volcan ; les pierres qu'on prend pour de la hue, et dont (iuel(|ues-unes ressemblent assez aux scories on de la pierre j)once. ne sont, a mon avis, (jue du basalte decompose; on trouve aussi cette rociie sous forme sj)hericiue. et composee de couches concentriques ..." (2, p. 34;; compare in the latter respect h^ig. 10 on page 52 of this paper). .\. (Ai.ix i.Kic;!!, who accomj)anied Captain P. Parker Kixc; on the sur- \f\ing \()\ages of II. M.S. Adventure and Beagle on their first expedition 1S26 1S30, read before the (Geological Society of London on Jan. 5th, 1831, a statement on 'The geology of the island of Juan Fernandez'. In the Proceedings of that N'ear the following account of Caldcleugh's discourse is given: "The author could discover no trace of a volcano, said to exist here by former visitors; all the rocks, according to him, consist of basaltic greenstone and trap of various mineralogical structure, both amorphous and vesicular, together with trappean concretions, no other contained minerals being observable except olivine and chaux carbonatee metastaticpie. It is further mentioned that the basalt in parts is almost colunuiar, and in others has a [)eaked and serrated outline, the mass being, here and there, traversed by dykes. Owing to the peculiar character of this basalt, and especially from the great quantity of olivine, the author compares its age with that of the basalt of Bohemia, the Rhine, the Vivarrais and Beaulieu in Provence ' (3, p. 256, also i)uhlishe(l in the Phil. Mag. and Annals of Philosophy, \'ol. l.\, I S3 I, }). 220). Captain King recaj)itulates Caldcleugh's narrative, as given above, with the addition: "In captain llAi.i.s interesting journal, there is a list of geological and mmeralogical specimens, of which one from Alasafuera is named vesicular lava" I4. p. 304). '1 he uhimate destiny of these specimens is unknown. Members of tlie Dumont dl'rville expedition, when visiting Masatierra in 1S3S. collected and specified several different saiu})les of the lavas from the island '-^- P- ■'4'- 1 he material tor the new analyses of basalts from Masatierra, which L\(K- of a \olcanic centre seems therefore to be implied" (8, p. 176). 1 he sid) marine eruption nuist be taken as conclusive evidence that the iuHuediate nei^hl)f)iMhood of the Juan I'Y-rnandez Islands has been the seat of volcamc action within the last 11; years. A point of further interest is that the e.xplosion was simultaneous with violent earthcjuakes on the Chilean coast, as Darwin already obserxcd (see j). 75). Hruggen refers to some finther observations of sub-marine eruptions in the vicinity oi' the Juan l-'ernandez Islands, recorded by Fk. Goll in his paper 'J)ie lu-dbeben Chiles' (Munchener Geogr. Studien 1904, Nr. 14). The following denotements by GoM, are taken from Hriiggen (16, p. 332): ADDITIONAL COMMENTS ON THE GEOLOGY OF THE JUAN FERNANDEZ ISLANDS 43 Fig, 3. Quebrada del Varadero on the east coast of Masafuera. Photo C. Skottsberg. 44 PERrv (jrHNSKi, "'Se«4un CioW se prodiijo", cl 12 de febrero de 1839 una erupcion submarina \- marenioto a unas 120 kni al este de la Isla de ^las a Tierra. rLl mismo autor cita la obscrvacion si<;iiicnte hecha en un punto iin poco mas austral: 'l^^n octubre de 1867, se sintio un temblor submarino en 34 55' S y 7/° 38' W (unas lOO millas SIC de Juan I-'ernande/.V despues el buque navego durante dos boras por agua tie color bianco lechoso. habiendo mucho pescado muerto en la superficie,' Se trata probablemente del mismo fencHiieno que describe JoSK M. Po.MAR en la forma sigiiiente, aun(]ue dice (jue el [)unto se halla a 100 millas al S\V en \ ez de Si'', de Juan I-'ernantlez: 'ICn 1867, el capitan SlMPSON de la barca britanica C'oronella navegaba en el I'acitico con mucha calma y vientos contrarios, con e.\cej)ci()n de un hierte \ iento acompaiiado de siete temblores que se produjeron como a 100 millas al SW de la Isla de Juan Fernandez; durante dos horas navego |)or agua tan blanca como leche; sondeo, pero no toco fondo en 100 pies de |)rorim(latl, \ io muclios pescados muertos y una gran cantidad de pajaros por todas j)artes. Agregaba el capitan Simpson que si hubiera estado 10 millas mas adelante, el cluxiiie hubiera sido peor y hubiera causado averias al buque.'" Concerning the earthquake of Vallenar in the province of Atacama on the loth of Nov. 1922 Ikuggen cites I^AII.EV WlLUS' observations on the contem- poraneous volcanic activity on San Felix. I^ruggen concludes his opinion on the submarine explosions, given above, as follows: "A la teoria del origin de los tsunamis por erupciones submarinas podria objetarse (jue serian erupciones muy excepcionales, ya que consisterian en una sola o nuiy pocas explosiones que causan las pocas olas sismicas, apagandose luego la actividad. V.n realidad se tratara solamente de las primeras explosiones que abren la chimenea para la salida de la lava o de los gases y que tienen la fuerza suhciente para causar el tsunami. Tambien en otro sentido, las erupciones sub- marinas se distinguen de las de los volcanes de los Andes, que solo excepcional- mente entran en actividad durante los terremotos. Los volcanos submarines parecen estar en relacicm mas estrecha con los focos sismicos de los grandes terremotos chilenos" (16, p. 332). Petrology of the Volcanic Formations. The following descrij)tion of the rocks of Masatierra and Masafuera is based on specimens collected by myself in 1908 and by Skottsbkrc; in 191 7. In many cases the two collections supplement one another and help to elucidate to a certain degree the distribution of the somewhat varying types of the lava flows. 1 he specimens from Masatierra in both collections give conclusive evidence that this inland in the main is formed of a rather uniform series of basaltic lava beds, only diverging in respect of coarser or finer grain or of a higher or lower content ot oJiMiie. The specimens from Masafuera on the other hand indicate more obx lou-, dissimilarities in the com[)osition of the rocks at different levels. In the following the principal petrological features of the two islands will first be treated. I'lHlcr a later heading references will be made to resemblances in various respects to other volcanic islands of the eastern Pacific. ADDITIONAL COMMENTS ON THE GEOLOGY OF THE JUAN FERNANDEZ ISLANDS 45. Masatierra. (Lat. 33° 37i' S.. Long. 78° 50' W.) The two collections of specimens from Masatierra by Skottsberg and my- self have partly been taken from different localities. The central parts of the island are however sparsely represented. Somewhat more complete series of rock samples originate from the heights around Bahia Cumberland (Ensenada de San Juan Bautista), Puerto Frances and Portezuelo as well as from Bahia del Padre. From the adjacent small island Santa Clara there is only one specimen. It is not easy to single out the local distribution of the different lava beds on account of the inconnected localities from which specimens have been collected. Some characteristic features may however be found which indicate that certain types of lavas are restricted to localised areas. To some extent one may then draw conclusions regarding the sequence of the volcanic eruptions. The predominating rocks of Masatierra are olivine basalts, differing only in their content of olivine. Lava beds from around Puerto Frances are exceptionally rich in this mineral, the content of which can reach 40 vol.% of the rock (Fig. 6). Such rocks, with an extreme content of olivine, I named picrite basalts in my earlier paper (12, p. 265). Lacroix originally named such basalts 'picrite feldspatique' but later discarded this name, substituting for it the name oceanite', under which name he includes the basalts of the Juan Fernandez Islands with an exceptionally high content of olivine (14, p. 65). Since the name picrite basalt, as originally defined in my former paper, has later been adopted ^ I will retain this name for the rocks in question, with the name oceanite as synonym. From the extremely olivine-rich basalts around Puerto Frances there is every transition over intermediate types to feldspar basalts without any olivine at all or with only a very insignificant amount thereof. Such rocks, however, have a relatively limited distribution on Masatierra. In general one may say that types with a very high content of olivine are restricted to lower elevations, whereas higher up more normal olivine basalts predominate. A second feature of dissimilarity in the basalts is found in their texture. Many of the lava beds show a coarse-grained ophitic texture and may be classed as dolerites, and have been so named by Hagerman and myself in our previous papers. Such rocks are usual around Puerto Frances and Bahia Cumberland as well as at Vaqueria and Tres Puntas. Specimens from all these localities are in outward appearance very much alike. In general these rocks, in contrast to other lavas, are singularly fresh. Only the olivine often shows a dark brown rim, indicating an incipient alteration to iddingsite {c(. 12, p. 260). The doleritic basalts seem only to occur at lower horizons, where the lava beds generally attain their greatest thickness. All the specimens thereof at hand are from between sea level and 200 m. A sample in Skottsberg's collection from Bahia Cumberland (Fig. 7) and another from Vaqueria are both from 1 50 m above sea ' The name oceanite was first proposed by Lacroix in 1923 (in Mineralogie de Madagascar, Vol. Ill, p. 49). The name is given "k cause de leur abondance dans les iles du Pac-hque". ^ Holmes, Q. J. Geol. Soc, 172, 1916, p, 231; Washington, Am. J. of Sc. V, 1923. p. 471- rKKCV (llKN'SEl, ADDITIONAL COMMENTS ON THE GEOLOGY OF THE JUAN FERNANDEZ ISLANDS 47 Fig. 5. Basaltic lava beds. Western head of Bahia Cumberland, Masatierra. Photo C. Skottsberg. level. The two speciinens are in texture and mineral composition identical. A third sample from Tres Puntas is from 200 m. Specimens in my earlier collection were taken from the lower lava beds in the quebradas around Puerto Frances. As no scoriaceous or slaggy development is to be seen in the upper or lower parts of the doleritic beds, which have a very coarse-grained texture, they may well represent intrusions between previously consolidated lava flows, in accordance with what Daly has assumed to be the case with similar doleritic rocks in Hawaii (R. Daly, Differentiation in Hawaii. Journ. of Geology, Vol. 9, 191 1, p. 291). The doleritic basalts and the picrite basalts have much the same mineralogical composition. The only essential difference is the higher content of olivine in the picrite basalts. The other rock-forming constituents in both rocks are labradorite, a pleochroic Ti-augite and magnetite. In the picrite basalts (oceanites) from around Puerto Frances as well as in the olivine basalts in general the phenocrysts of olivine are often more or less altered to iddingsite. In some cases only insignificant rests of olivine are left; the iddingsite pseudomorphs, however, still retain the crystal habit of the olivine. To the petrogenetic problem of iddingsite as representing a deuteric mineral I will return later in connection with equivalent alterations in certain lava beds of Masafuera, where the 'iddingsitisation' has gone further and there gives the rocks a very characteristic aspect (see p. 60). In the picrite basalts from Puerto Frances inclusions of dunite occur. The large olivine crystals of this rock are singularly fresh, without an}^ signs of even 48 PF.RCV QU F.N SEL ^ M^^m^'- ^;^ v!i>;?r: > Fij,''. 6. Ficritc basalt oceanite). x 9. Puerto Frances, Masatierra. ig. 7- I ><'l('riti< basalt, x ^y. Bahia Cumberland, Masatierra. ADDITIONAL COMMENTS ON THE GEOLOGY OF THE JUAN FERNANDEIZ ISLANDS 49 a periferic alteration to iddingsite which is otherwise usual in the olivine of the surrounding basaltic lavas. At higher levels the basalts have a more normal character. Large phenocrysts of olivine, or olivine and feldspar, lie in a more or less glassy groundmass of augite, olivine and labradorite with abundant small grains of magnetite, llmenite in tabular form or in skeleton cry.stals is now often present. Most of these lavas are vesicular, scoriaceous or slaggy. The cavities are in many cases rimmed or filled with opal. The occurrence of basalts of this type is widespread up to the highest parts of the island, and they are without doubt the dominant lavas of Masatierra. Somewhat divergent lava beds seem, however, to predominate at intermediate horizons. All the specimens at hand of this type are holocrystalline rocks of an ash-grey colour, aphanitic and aphyric in texture (Fig. 8). They generally show a characteristic light zone of weathering and a tendency to develope a columnar struc- ture (Fig. 9). The fine-grained mineral assemblage consists of augite, labradorite and very abundant magnetite in small euhedral crystals, evenly dispersed throughout the rock. llmenite is also generally present in skeleton crystals. Stray small grains of olivine may in some cases be observed; in other specimens olivine is completely absent. In vesicular lavas of this type the cavities are again more or less filled with opal (Fig. 8 b). These lava beds, which represent the only specific feldspar basalts of Masa- tierra, are found at heights between 400 and 500 m (Cordon Chifladores 400 m, Portezuelo 500 m). At an elevation below 100 m at Punta Larga in the more western part of the island the same type has been found, but then in the form of a dike, which may signify a channel for the analogous lavas at higher levels. The very characteristic aspect of these rocks seems to indicate that they represent a definite epoch of intrusion, intermediate between the doleritic basalts and picrite basalts of the lower parts and the scoriaceous olivine basalts of the higher horizons. An analysis has been made of a very similar rock from Masafuera, which confirms its classification there as a feldspar basalt (see p. 66). Lacroix's four new analyses of basaltic lavas from Masatierra indicate that the analysed rocks are similar in composition. It is regrettable that the specimens all originate from Bahia Cumberland. Probably the members of the d'Urville expedition only brought back samples from that locality and these were there- fore the only specimens available in Paris for Lacroix's analyses. On the other hand the insignificant dip of the lava flows on Masatierra (i5' — 20°) may infer that the analysed rocks can be taken as representative in chemical composition for the basal basalts of the island. Microscopic determinations of corresponding specimens from the other localities point in the same direction. With the exception of the ultra-femic picrite basalts from Puerto Frances and the inclusions therein of dunitic rocks, and the light grey feldspar basalts of intermediate horizons, we may conclude that the dominant rocks of Masatierra consist of rather normal olivine basalts deviating principally in texture and in a varying content of olivine. so rKKcV (UENSKL I'"ij(. H a. X'esirular aphyric feldspar basalt cokminar structure . (see Fig-. 9 . 30. Portezuelo, Masatierra Fig. ' occurrences of superficial tuffs seem still to have evaded destruction by erosion. Wlu-n Kcnard says that "amongst the specimens collected at Juan I'"cnian(iez (Masaticrra) !)>• the Ciiallenger l^x|)edition in 1 875, we have not found any specimen which iniglit belong to any recent eruption, no tuffs, no volcanic ash are to be found and e\cr\thing seems to prove that they have been washed away by the waves and the atmosj)heric agencies" (8, j). 176), this last conclusion seems (juestionable. 1 lagerman refers to two specimens in Skottsberg's collections ADDITIONAL COMMENTS ON THE OEOLOGY OF THE JUAN FERNANDEZ ISLANDS 53 as representing recent tuffacious material. The one he describes as "ein poroses, dichtes Gestein von El Puente, dem Istmus zwischen der Padrebucht und Carbajal und ist ein Palagonittuff mit einigen sporadischen Augit- und Magnetitkornern. Der andere Tuff stammt von dem nordlichen Ufer der Padrebucht, von wo einige umgewandelte Olivinbasalte herriihren" (13, p. 26). As these specimens have been collected without any observations regarding their petrological position, one cannot draw any conclusive evidence as to their age, but the specimens from both the localities have every appearance of being recent pyroclastic sediments. I have previously described rocks from Bahia Cumberland, filling out the greater part of the bay, which I assumed to be of recent tuffaccous origin. The description was given as follows: "In den zentralen Teilen von Cumberland Hay liegen noch ziemlich machtige, meist lebhaft rot gefarbte Tufflager, die sehr stark umgewandelt sind. Bruchstucke von Olivinkristallen, lapilliartige Lavabruchstucke, P>zkorner und Glas liegen in einer Grundmasse, die aus einer weichen, mit Mes.ser schneidbaren, roten lateritahnlichen Lehmsubstanz besteht, die durch Verwitterung aus dem ursprlinglichen Tuffmaterial hervorgegangen sein durfte. Uberall in den Tuffsedimenten ist eine deutliche Lagerung sichtbar" (12, p. 266). It is over 40 years since I visited the locality and naturally I cannot now rely on any personal recollection. The inundation in connection with the volcanic disturbance of 1835, referred to above (p. 41), may have wrought such havoc, that superficial deposits could have been re-formed. But the composition of the formation, as well as my notes from the field, offer indications that the tuffaceous material of Bahia Cumber- land also represents pyroclastic sediments of recent volcanic origin. Even if only trivial remains of tuffaceous formations, indicative of late volcanic activity on the island, are left, the submarine explosion of 1835 confirms without doubt that the area in the immediate vicinity has at that time been subjected to disturbances of volcanic nature. Some rocks from around Bahia del Padre deserve special notice. Schulze and Pohlmann have observed the deviating nature of the rock assemblage and the latter has commented thereon as follows: "De suma importancia para esplicar la formacion jeolojica de Masatierra es la entrada a Bahia del Padre, situada en la parte suroeste de la isla. Aqui se observa debajo de las capas basalticas ya descritas un grand macizo de roca compacta verdosa, que es andesitica. Segun la opinion de Schulze' que, a me parecer, es correcta, esta roca verdosa es la mas antigua de la isla. P^n ningun otro punto, ni en Masatierra, ni en Masafuera se ha observado una configuracion jeolojica analoga a la mencionada" (11, p. 4). The rocks referred to by Pohlmann certainly show a divergent aspect, l^ut I do not think they can be taken as represehting rocks belonging to a more ancient formation than the lavas around, nor that they should be classified as andesites, as Pohlmann has assumed. I have as the result of microscopical examination of the rocks come to the conclusion that the deviating character is the result of post-volcanic alterations through thermal processes. In my earlier paper I have given in some detail the reasons for this conclusion (12, p. 266), and ' Ur. Schulze died before publishing his observations. 4 — 516795 The Nat. Hist, of Juan Fernandez and Easter Isl. Vol. I. 54 \'\\--like substance which was found to be newtonite. In intimate association with this mineral is nearly always found a vellowi^ii waw substance which shows ever\' resemblance to what mineralogi- call\- ma\- be si^nitied as bole (I''i<;. 13). It is very brittle with a conchoidal fraclme. in water it readil\' disintegrates into small angular fragments. Tliis bole mineral is also t'ound tilling cracks, or occurs in smaller masses in the cavities of the basalts around the baw in t!ie agglomeritic lavas of the vicinit)- large cavities are filled with hard compact magnesite. As earl\- as 1886 Darapsky described this mineral by the r.ame "( ilockenstein" and gave an analysis thereof, which shows it to be an exceptionall)- pure magnesite (9, p. I 13). Without doubt the magnesite is primarily derued tVom olivine, the decomposition being caused by the same processes as ha\e changed the basaltic lavas nearby. The whole asj^ect of the rocks from l^ahia del Padre, with magnesite, calcite, serpentine, chlorite, sca{)olite and pyrite as secondary minerals, seems without doubt to indicate that the lava beds in question have been subjected to alterations in connection with thermal processes during some intermediate phase of volcanic activity on the island. There seems no reason to classify them as andesites of an older formation, as is done by Schulze and Pohlmann. In a s[)ecimen collected by Skottsberg from the shore south of El Yunque, Ilagerman also tound evidence of a far-reaching decomposition. Under the heading "! Ixclrothermale I^ildungen" he gives the following description: "Das Handstiick ist ein von weissen .Streifen durchzogenes scharfgriines Gestein, das u. d. M. grosse Augitkristalle in einer vollig zeri^ossenen Serpentinmasse zeigt. Das Pra- ))arat ist von Aragonitbandern durchzogen. Dieses Gestein muss als ein stark umgebildetes ( )livinfels bezeichnct werdcn" (13, p. 27). The large olivine crystals of the duniti'. tound as inclusions in the picrite basalt at Puerto Prances, show, on the contrarx-, no signs of secondar\- alteration (see p. 47). There is therefore no doubt that the higiily decomposed '( )hvinfels' described by Hagerman has succumbed to a later decomposition of much the same nature as has been active around I)ahia (K-1 I'adre. Masafuera. I. at. ^y 52' S.. Lon^. pes of petrographic interest. They have, however, hitherto only been sinumanl\- described in the j)a|)ers by myself (l 2, p. 274) and b\' I lagerman 1 1 3. p. j8i. I he rocks which predominate at lower levels are mostly vesicular to slaggy basalts. The\- are well represented around the Ouebrada de las Casas. At higher levels the basalts consist of more compact lava beds. In contrast to the basalts ADDITIONAL COMMENTS ON THE GEOLOGY OF THE JUAN FERNANDEZ ISLANDS 57 QwebradxvNegrcb Q Saitdahto Q SandaJLo QuiebraSa. PcUo Chx>% this (juestion I will return under a concluding heading dealing with the chemical and petrological connections between the rock of the Juan I-'ernandez Islands and other volcanic islands of the eastern Pacific. ADDITIONAL COMMENTS ON THE GEOLOGY OF THE JUAN FERNANDEZ ISLANDS 67 Fig. •4. Olivine basalt, supersaturated with iron oxide, x 35. Vesicular flow-breccia. Los Inocentes ^elevation c. i 500 m), Masafuera. Fig. 25. Olivine phenocryst from specimen Fig. 24. x no. Specks and olivine as rests in a pseudomorph after olivine. Los Inocentes. cak^ <:i tUKiiLcred 6S PKKCV ()1-i:nski. I"ij4. 26. Disintegrated phenocryst of olivine, caused by high temperature oxidation. The rim around the olivine crystal is hematite. Within the olivine the light grains are mineral components, formed by exsolution. < 150. Los Inocentes, Masafuera. Photo P. Ramdohr. Pig. 27. I'art of disintegrated olivine |)heno( ryst, enlarged x 600. Against the dark background of olivine the exsolution proiUn t is seen in the form of small composite grains consisting of hematite light and a darker undefined component. Larger lij^ht grains see PI. II, P^ig. I. Photo P. Ramdohr. ADDITIONAL COMMENTS ON THE GEOLOGY OF THE JUAN FERNANDEZ ISLANDS 69 Fig. 28. High-temperature exsolution of spinel, enclosed in olivine phenocryst (PI. I, Fig. 2). Light lamellae, according to Ramdohr probably magnetite in groundmass of excess spinel, x 600. Photo P. Ramdohr. Fig. 29. Magnetite in groundmass of the lava bed as seen in PI. I, Fig. i pa:t-a!iy c!):inge(l to hematite (martite). x 300. Photo P. Ramdohr, 5 — 516795 The Nat. Hist, of Juan Fernandez and Easter Isl. Vol. I. ■JO I'KKCN (JIKN'SKI. An observation of importance with rej^ard to the distribution of different lava beds on Masafuera is tlie find of a deviatini^ type of olivine basalt from Los Inocentes. Accordin*; to Skottsber^'s observations such la\as probably occupy the hii^diest part of the island, representing^ elevations above I 420 m. The few samples brouj^ht back are hi<4hl\' scoriaceous flow breccias containing a high content of iron oxides (h'ig. 24). Hagerman has in the preceding publication of this series given a descrij)tion of these lavas which he characterizes as slaggy ohvine basalts with large olivine phenocrysts, supersaturated with magnetite (•3. V- ^S'^- 'l"he phenocrysts of olivine are under the microscope found to be almost opacjiie, due to the {precipitation of new-formed ore minerals. A varying amount of residual olivine is, however, nearly always to be observed in the form of specks or streaks (l^'ig. 25). \o signs of alteration are to be observed in this olivine. Optical determinations indicate that only a low content of about 8% FeO is present in the molecule. The groundmass consists of slender laths of labradorite, small grains of augite, magnetite, ilmenite and pseudobrookite in a dark brown glass matrix. To determine the mineral composition of the pseudomorphs after olivine Professor S. Gavklin and Dr Uv'i"KNlM)GAARl)'r kindly undertook to examine some polished sections of the rock. Professor P. Ramdoiir (Heidelberg) contempora- neously supervised a section for the same reason. It thereby became apparent that the seemingly opacpie constituent was not magnetite and that the mineral assemblage of the pseudomorphs was of a complicated nature. Professor Ramdohr has taken four photomicrographs thereof and kindly put them at my disposal. They are reproduced in Fig. 26 — 29 with Ramdohr's explanatory notes. In Fig. 26 the essential components can be observed. A rim of hematite is seen to encircle an i(liom()r|)hic crystal of olivine with specks of disintegrated minerals. In the enlarged microphotograph Fig. 27 these minerals are seen in the form of small lighter grains uniformly distributed against the dark background of olivine. One can now observe that the grains consist of two constituents. The one component is hematite. Repeated attempts have been made to determine the second com- ponent both in polished sections and with X-ray powder photographs. No con- clusive evidence regarding the true nature of this mineral has, however, been attained. The singular alteration of the olivine phenocrysts must in all probability be connected with the saiue processes as have controlled the formation of the deuteric iddingsite, though in the samj)lcs at hand this mineral is not extant. P:dwards seems to have described a very similar formation in the iddingsite-bearing basalts from two X'ictorian localities in Australia. After concluding that the iddingsite must have been formed before the ultimate consolidation of the lava flow, Fdwards says: "In some instances, however, the action has gone further, and a rim of iron oxid is formed on the outer margin of the iddingsite. Fventually all the original olivine vanishes, and the iddingsite, which had formed a rim about it, is completely rej)laced by magnetite ... It is essential for the formation of iddingsite that the magma should not only be rich in water vapour, but that it should have differentiated ADDITIONAL COMMENTS ON THE GEOLOGY OF THE JUAN FERNANDEZ ISLANDS 7 I in such a manner as to give rise to an iron-rich final fluid" (Am. Min. 23, 1938, p. 280). A photomicrograph in the text shows, according to Edwards, "a pseudo- morph of iron ore after iddingsite, itself a pseudomorph after olivine". These suppositions seem also referable to the very similar formation on Masafuera. The lava bed in question contains 33.86 % FeO according to deter- mination by Landergren (13, p. 33). This is more than double the content of FeO in any other basalts of the Juan Fernandez Islands. There seems little doubt that the pseudomorphs in question have formed under much the same conditions as advanced by Edwards. This is furthermore supported by the nearly identical appearance of the photomicrographs of the pseudomorphs from the Australian localities and from Masafuera, in both cases in connection with lavas containing iddingsite. Edwards assumes that, via an intermediate stage of iddingsite, the ulti- mately formed component is magnetite. As narrated above, this is not the case in the lava from Masafuera, where the pseudomorphs are found to be of a more complicated composition, formed without iddingsite as an intermediate phase. All the observations given above seem to indicate that the iddingsite-bearing basalt from elevations about i 000 m as well as the flow breccias from the highest elevations on Masafuera have been subjected to an automorphic re-mineralisation, prior to the final consolidation of the magma. This would suggest that the volcanic eruptions have been interposed by periods, during which the lava in a molten state has temporally stagnated in the volcanic vent under conditions which in connection with active volatile phases have led to the formation of such deuteric minerals as iddingsite and to the partial high temperature exsolution of the oli- vine phenocrysts. The conclusions which may be drawn regarding the distribution of the different lavas of Masafuera would be now that rocks of more normal basaltic composition, principally feldspar basalts, occupy the lower and intermediate eleva- tions of the island, whereas the highest parts consist of an olivine basalt, super- saturated with iron oxides. At some intermediate elevations lavas of a more alkaline character are found in the form of at least two beds of soda-trachyte, interposed between dominant flows of basaltic lavas. Horizontal distribution of different types of lava on Masafuera. in m X 100 o I 2 3 4 5 6 7 8 91011121314 1500 Olivine basalt supersaturated with magnetite Light grey phyric feldspar basalt Soda-trachyte Iddingsite-bearing phyric oli- vine basalt Dark phyric and aphyric feldspar- and olivine basalt 5*— 516795 X X X X X X X X v X I'F.KfV (JUFNSF.L Fi^^ 30. Picrite basalt (masafuerite;. Dike rock. Nat. size. Loberia vieja, Masafuera. } Ia<^erman has recorded in his paper in tabular form his conception of the horizontal distribution of the different lavas of Masafuera. This is reproduced on p. 71 with sotne slight corrections on the base of renewed examination. My former supposition that the basaltic lavas only occur up to an elevation of about I 000 m, from there on being succeeded by more alkaline rocks of soda-trachytic composition, is no longer in agreement with more recent observations, based on Skottsberg's new collections. Any thought of gravitative differentiation to explain the sequence of the volcanic rocks, which I tentatively proposed in my former publication, must in the light of later observations be discarded now. The numerous basaltic dikes, traversing the whole island in a VV^est-East direction, are worth special notice as representing rocks exceptionally rich in olivine. In this res})ect they exceed the most olivine-rich picrite basalts from Puerto I'Vances on Masatierra. l^OWKN has commented on these rocks as follows: "( )ne otiier rock may be mentioned in this connection. It is a picrite basalt from Juan PY'rnandc/., a dike, not a lava, but quenched so as to reveal the fact of its origin. In it is shown the highest amount of normative olivine (53 %) of any rock termed basalt by the author describing it. Great crystals of olivine lie in an aphanitic ground composed mainly of plagioclase and augite (Fig. 30 — 32). Some of the olivmc basalts of this island group are, locally at least, about as rich in olivine as this dike, but they have not been analyzed. Their high olivine content is invariably due to an increased amount of phenocrysts of olivine about i cm in diameter. Plainly these crystals were not in solution in the dike or flow material at the time of its intrusion or extrusion. This ADDITIONAL COMMENTS ON THE GEOLOGY OF THE JUAN FERNANDEZ ISLANDS 73 Fig. 31. Masafuerite dike, x 9. Loberia vieja, Masafuera. Fig. 32. Masafuerite dike, x 9. Analogous to sample Fig. 31 but with rims of iddingsitc all phenocrysts of olivine. Loberia vieja, Masafuera. 74 PERCY (UKNSEL fact does not prove that they were not in solution in that material at an earlier time. Hut if one finds the condition siiown by these basalts to be invariably true of all rocks rich in olivine, which have suffered quenching, one must conclude that large amounts of olivine occur in solution in magmatic liquids. A survey of igneous provinces leaves no question that such rocks do have this character, that is, thc\' always contain either all of their olivine or all in excess of a (juite small amount (apparently some 12 — 15 %) as relatively large pheno- crysts. rhe\-, therefore, force acceptance of the stated conclusion" (23, p. 163). The (juotation above refers to purely theoretical questions but indicates the extreme j)osition these rocks hold in petrographic classification with regard to the abnor- mally high content of olivine in basaltic lavas. JoilANNSEN has named these rocks masafuerite, with the following argument: "The picrite basalt from Masafuera of the islands of the Juan Fernandez group is a most extraordinary rock . . . While this particular rock occurs as a dike, on the adjacent island of Masatierra, for example at Puerto h'rances, there is a similar rock in the form of a lava flow with large olivine crystals in a groundmass containing more or less the same mineral. I am placing the rock among the hypabyssals on the basis of the occurrence on Masa- fuera. To all olivine-melabasalt dikes which contain more olivine than any other mineral and in addition carry basic plagioclase and augite, I should like to apply the name masafuerite" (24, p. 334). We already have the name picrite basalt for the lava flows of much the same composition on Masatierra, with Lacroix's name oceanite as synonym, given with the following definition: "les roches basaltiques porphyriques a olivine sont parfois extraordinairement riches en peridot; dans I'echantillon analyse, tons les grains de ce mineral se touchent, ils sont reunis par une petite quantite de plagues de labrador, englobant des microlites d'augite et des lames d'ilmenite" (14, p. 44). Johann.sen has restricted the name masafuerite to aschistic dike rocks of a picrite basaltic magma. These dikes may represent, at least in part, transmission channels for the upj)er basalt beds, supersaturated with iron oxides. A significant feature in this respect is that they contain numerous 'schlieren' of darker colour, due to abundant minute grains of magnetite. These streaks may indicate relics from a magmatic flow, subse(juently consf)lidated in the shape of the olivine basalt, supersaturated in iron oxides, which now forms the highest parts of the island. At a later period the ciiannels may then have been filled with the melanocratic magma, which now characterizes them as such singular rocks. Sequent intrusions of this nature might lead to phases of crystalHsation in accordance with l^owen's conception of these dikes as (jucnched rocks, referred to above. Regional Relations. Tectonic Connections. In the introductory lines I already noted that during recent years as well as in older reports speculations have been offered regarding connections in one or ADDITIONAL COMMENTS ON THE GEOLOGY OF THE JUAN FERNANDEZ ISLANDS 75 Other respect between the Juan Fernandez Islands and the volcanic islands of the eastern Pacific. Some quotations may be given. Charles Darwin already expressed views regarding geotectonic connections between the Juan Fernandez Islands and, on the one side, the South American continent, on the other, the Galapagos Islands. In his 'Observations on the volcanic islands and parts of South America' he says: "Some authors have remarked that volcanic islands occur scattered, though at very unequal distances, along the shores of the great continents, as if in some measure connected with them. In the case of Juan Fernandez, situated 330 miles from the coast of Chile, there was undoubtedly a connection between the volcanic forces acting under this island and under the continent as was shown during the earthquake of 1835, The islands, moreover, of some of the small volcanic groups, which border the continents, are placed in lines, related to those along which the adjoining shores of the continent trend; I may instance the lines of intersection at the Gala- pagos" (20, p. 144). In his paper 'Constitution lithologique des lies volcaniques de la Polynesie Australe' Lacroix gives expression to much the same trend of thought when he writes: "Particulierement interessantes sont les iles volcaniques qui se trouvent a une plus ou moins distance de I'Amerique du Sud, les iles Juan Fernandez, San Felix et San Ambrosio, et enfin Galapagos, puis au large de I'Amerique centrale, I'ile Clipperton. Bien que la connaissance de la lithologie de ces iles soit loin d'etre completement eclaircie, on peut a present assurer que leur laves different de celles des Cordilleres des Andes, c'est-a-dire de la serie circumpacifique et montrer qu'elles se rattachent a la serie intrapacifique" (14, p. 64). In 'La Face de la Terre' Emmanuel de Margerie refers to the same subject as follows: "Le Relay de la marine des Etats-Unis a signale, au large de Valparaiso 5.651 metres. A I'ouest de ces fosses sont situees les deux iles volcaniques anciennes de San Felix et San Ambrosio; au Sud de ces iles le croiseur Chilien Presidente Pinto a trouve, sur une etendue de 760 km, des profondeurs si faibles qu'il est probable qu'une crete sous-marine, orientee a peu pres N — S, s'allonge dans la direction de I'ile Juan Fernandez" (25, III, P-I359)- In his 'Description and Geology' of San Felix and San Ambrosio, Bailev Willis writes: "San Felix and San Ambrosio are volcanic islands in the South Pacific Ocean, San Felix being situated in latitude 26°: 5' south and longitude 80°;' west of Greenwich and San Ambrosio lying about 16 km to the east-south-east. They are about 500 miles west of Chanaral on the east coast of Chile, and the same distance due north of the group of Juan Fernandez and Mas-a-fuera. The South Pacific charts show several rocks or islets and some whose existence is recorded as doubtful, which, with the above-named islands, form an archipelago strewn on a narrow submarine ridge that extends along the meridian ol 80 degrees west from about 36 degrees south to 26 degrees south, the ridge being defined by the 2.000-meter contour line. Knowing that all these islands and islets are peaks of volcanoes, we may suspect that there are more of them than we can see; but this must remain an unverified guess until detailed soundings PEKCN OlENSEI. can be made. The deptli of the ocean in this region, which lies west of the Ricliartls Deej), varies from 4.000 to 5.000 meters. The islands, therefore, repre- sent tlie summits of volcanoes probably sixteen to eighteen thousand feet or more in height — that is to say, they compare with the volcanoes of the Andes, which are situated on the other side of the deep" (26, p. 365). In an interesting paf)er on the geology of Galapagos, Cocos and Easter Island, L. J. CliriUi has j)ublished some noteworthy remarks on the regional relations of the \olcanic islands of the eastern central Pacific. He writes as follows: "L'nder the east central Pacific there lies a vast area, the Albatross Plateau, under depths of less than 2.000 fathoms, though on all sides the depths exceed this figure. Xo islands rise from the central part of this plateau, but at each end is an archipelago that appears to have been built up on a set of intersecting fissures, the Martjuesas at the western end, and the Galapagos at the eastern. On or near its southern margin too there are several volcanic islands, including the Mangareva (Gambier) Archipelago, Pitcairn, Piaster, Sala-y-Gomez and the Juan I^'ernandez Islands. "It is suggested that the plateau constitutes a resistant block which has withstood lateral pressure that has been brought to bear on it from all sides, that around its margin it has become cracked and fissured, and that on the fissures volcanic islands have been erected. That these islands owe their origin to a common cause is suggesed by the similarity of their structure and geological history, so far it is known. . . . Petrographically, too, these islands resemble each other and differ from most of the other Pacific islands. The most striking charac- teristics of their rocks are the almost complete absence of nepheline-bearing types and the presence of virtual free silicia in many. "Cocos, St. Vd'ix and St. Ambrose islands are constituted in part of nephe- line rocks, and for this reason they are regarded as lying, not on the resistant block, but beyond its eastern margin. Petrographically they resemble the Society Islands and Austral Islands which lie to the west of the plateau. "It is thought that beyond the margins of the block the crust is more pliable and has yielded to pressure, with the formation of anticlines and synclines. \'olcanoes that have produced nepheline-bearing rocks have been erected on the anticlines. The folds have tended in the western area to migrate from southwest to northeast with a wave-like motion j)roved by the history of their coral reefs. I here is not sufficient evidence, however, to determine whether the folds which l)r()bably underlie Cocos, St. Felix and St. Ambrose islands have suffered a similar movement ' (27, |). 43). Ji .\\ Hiu(;(;k\ has recently in his book 'Fundamentos de la Geologia de Chile' discussed the geotectonic position of the Juan P'ernandez Islands. He writes: "klsta zona (la regicui situada al este de Llico, en Arauco) de dislocaciones tan extranas a la structura de la Cordillera de los Andes, coincide con la region donde una ancha loma submarina se desprende del continente. F:ncima de la loma se levantan las islas Juan I^'ernandez y mas al norte las de San Ambrosio y San Felix. Parece que se trata de una antigua Cordillera que se separo del actual conti- ADDITIONAL COMMENTS ON THE GEOLOGY OF THE JUAN FERNANDEZ ISLANDS 77 nente entre Magellanes y Arauco y que se hundio posteriormente. A este cordillera o simplemente zona continental, que llamaremos 'Tierra de Juan Fernandez', se debera el termino de las capas de la Quiriquina en el sur de Arauco y tambien el hecho que sedimentos marinos de Eoceno no se conozcan mas a sur. Pero en el Oligoceno, cuando la costa del piso de Navidad se extendio hasta la region de Ypiin (45° L. S.), se habia hundido una gran parte de la Tierra de Juan Fernandez, conservandose probablemente cierta extencion en la vecindad de las islas volcanicas, cuyas rocas se formaron solo mas tarde en erupciones posteriores. A juzgar por el grado de denudacion y en vista de las actividades volcanicas recientes en Mas a Tierra y San Felix, la parte volcanica de estas islas se habra formado en el Terciario superior, probablemente en el Plioceno, cuando existia todavia un resto de la antigua Tierra de Juan Fernandez, de la cual immigro la flora del Eoceno. Cuando mas tarde se hundio tambien este resto, sobresalian solamente las partes volcanicas, constituyendo las islas de Juan Fernandez, que Servian de refugio para la flora" (16, p. 59). The references now given suffice to show how the position of the Juan Fernandez Islands has from different points of view been geotectonically connected with other volcanic islands or groups of islands of the eastern Pacific. Petrographic Connections. In many papers of recent years, petrological and petrographical connections between the rocks of the Juan Fernandez Islands and those of other volcanic islands of the Pacific Ocean have also been the subject of discussion. The types of lava which in this respect have been of special interest are the extremely melanocratic picrite basalts (occanites of Lacroix) and their occasional combi- nations with more alkaline rocks. A short summary of the literature on this subject may be given first. We may conveniently begin with the islands of San Felix and San Ambro- sio, which geographically lie nearest. The distance is 760 km due north of Juan Fernandez. H. S. Washington has given a petrographical description of the rocks. I may quote some lines from his general conclusions: "It would appear from the specimens brought back by Willis that the lavas of San Felix volcano are, so far as known, only of two kinds — a decidedly sodic trachyte and a somewhat variable nepheline basanite, which seems to be highly vitreous. There is little doubt that the yellow tuff is derived from a nephelite basanite magma closely similar to that of the flows. The prominent characteristic of these two types of lava is their high content in alkalies, especially in soda, while high titanium and phosphorus appear to be other constant characters of minor but still considerable interest. This conclusion as to the generally highly sodic character of the San Felix lavas is subject to the limitations imposed by the absence of specimens from the lower flows and from various parts of the island. Such basaltic lavas, especially if highly vitreous, may appear megascopically to be very uniform and yet modally and chemically very diverse. It is, therelore. 7 8 I'KRCV (jrKNSRl. possible tluit earlier, lowermost flows are less sodic and more typically basaltic than the upper, which were the ones examined. " Washington continues: "In this {:)redominantly highly sodic character of the lavas San Felix appears to differ widel\' from other Pacific islands. At IMasafuera, it is true, both soda trach\-te (and nephcline basanite)', closely like those of San Felix, occur, but these are accompanied b\' basalt and [)icrite basalt, whereas at the neighbouring juan b'ernandez (Masatierra) the lavas appear to be, to judge from Quensel's description, only olivine basalt with neither trachyte nor basanite. Trachyte, also highl\' sodic, occurs at several other Pacific volcanic islands, as do also nephelite basanite and similar rocks high in soda; but at all of them the predominant lavas are more or less normal basalts or andesites; so that the general magmatic character is basaltic — that is to say, sodi-calcic, somewhat modified by distinctly sodic facies" (26, p. 382). Referring to the trachyte of San Felix Washington says that "in thin section the rock shows a somewhat peculiar texture, which resembles that of the trachyte of Masafuera described by Quensel, that of Puu Anahulu on the island of Hawaii and of the trachyte of Lahaina on Maui. The texture seems to be rather usual in the trachytes of the Intra-Pacific volcanic islands. Ill-defined laths of alkali feldspar make up most of the rock. Most of these are arranged irregularly, but here and there flow texture is evident" (26, p. 375). In several places in his paper 'La constitution lithologique des iles vol- canitjues de la Polynesie Australe' Lacroix compares the petrographic character ot the lavas of the Juan Fernandez Islands with those of other volcanic islands of the Pacific. Concerning the basalts of the Galapagos Islands he observes that they "offrent I'analogie la plus grande avec les basaltes de Masatierra et avec ceux des iles (iambier, c'est-a-dire avec les plus calciques des basaltes du I'acificjue et les plus pauvres en potasse" (14, p. 68). And further on in the same paper: "Les iles Juan P'ernandez se groupent au voisinage des iles Marquises et 1 on a vu (ju'il faut comparer ce que Ton sait des roches des iles Galapagos aux donnces concernant ceux des iles Gambier" (14, p. jy]. '1 hese conclusions of Lacroix refer to the normal basalts of Masatierra and are founded on the four new analyses of such basalts. The similarity in respect ot the Gambier Islands, however, goes a step further, as Lacroix describes from tiiere, associated with more normal basalts, typical oceanites (picrite basalts), which, as we have seen, also occur on Masatierra. A third author who has brought the Juan Fernandez rocks under discussion with reterence to chemical similarities with other volcanic islands of the Pacific is Conrad Hi kki. Under the title 'Chemismus und provinziale Verhaltnisse der jungcruptivcn (iesteine des pazifischen Oceans und seine Umrandung' he coor- dinates under the heading "Typus Hawaii" (in contrast to "Typus Tahiti"), the rocks of Hawaii, the Leeward group, Juan Fernandez and Samoa, remarking that the Juan I-'ernandez, San P'elix and San Ambrosio lavas are good representatives of the group, the basanite from Masafuera, however, showing a small deficiency ' This name now discarded see p. 63). ADDITIONAL COMMENTS ON THE GEOLOGY OF THE JUAN FERNANDEZ ISLANDS 79 in al — alk (28, p. 177). — This deficiency is now explained by the fact that the analysed specimen has attained an abnormal chemical composition through the formation of deuteric minerals (see p. 60). A second part of Chubb's paper on the geology of Galapagos, Cocos and Easter islands contains the "Petrology of the Galapagos Islands" by C. Richard- son. Under a concluding heading he says: "The Juan Fernandez Islands are the only islands on which both types of basalt (porphyric with dominant phenocrysts of olivine or with basic plagioclase) are found in addition to soda trachyte similar to that occurring in the Galapagos Archipelago. Although oceanites (and basanitic lavas) are also present, the Juan Fernandez Islands are petrologically closer to the Galapagos than any other islands" (27, p. 64). General Conclusions. A characteristic feature of the volcanic rocks of the Pacific is the universal predominance of sodium over potassium. Lacroix says that in this respect "toutes les roches etudiees presentent la commune caracteristique d'etre plus riche en soude qu'en potasse" (14, p. 55). BuRRi comes to the same conclusion, based on recalculations of all available analyses. He states that the Niggli value k is always under 0.4 and for the most typical regions of volcanic rocks of the intra- Pacific Ocean under 0.25 (28, p. 173). The value k in the four new analyses from Masatierra in no case exceeds 0.20 (average 0.17). And the light grey basalt from Masafuera has a still lower content of K2O (Niggli value k 0.08). Therefore, with regard to low percentage of potassium the basaltic lavas of both Masatierra and Masafuera must be con- sidered in this respect as representative for the basalts of the volcanic islands of the Pacific. In other respects the petrographic relationship between Juan Fernandez and other intra-pacific islands has been interpreted somewhat differently. The reason is, however, easy to explain. In some cases only the basaltic rocks of Masatierra have been taken into consideration, in other cases special notice has been given to the soda-trachytic lavas of Masafuera as indicating a casual presence of more alkaline rocks. Since no rocks of alkaline character occur on Masatierra, this island has petrographically been connected most closely with the Gambler and Marquesas Islands as, according to Lacroix, representing "les plus calciques des basaltes du Pacifique et les plus pauvres en potasse" (14, p. 68 and ']']\ In chemical composition the basaltic lavas of the Juan Fernandez Islands also show similarities with some of the basalts from the Hawaiian Islands. Several analyses of phyric and aphyric feldspar basalts from Kohala and Maunakea as well as from Kilauea, published by Washington are very similar to those of the basalts from Juan Fernandez (Am. J. of Sc. 5, 1923, p. 482—87 and 6, p. 341). It can be suggestive that together with both ancient and recent lavas of Kilauea, Washington also describes chrysophyric picrite basalts of much the same character as those from Masatierra and Masafuera. Another similarity can be given. WHITMAN CROSS has described, together with the normal basalts. 8o PKRCV QUENSEL soda-trachyte from Maui and Analuilu, on the first island in connection with a 'picritic basalt' [V. S. (ieol. Survey, Prof. Paper 88, 191 5, p. 26 — 28). This corre- sponds exactly to the rock assenibhif^e of Masafuera. Petro(^ra{)hic description of the Hawaiian basalts from the mentioned localities is, also in other respects, found to agree with both megascopic and microscopic features of the basalts of Juan Fernandez. Washington describes an aphyric basalt from Kohala as follows: "The type is a light grey, almost aphanitic lava, except that some rare, very small feldspar [)henocr)'sts may be present, and a few pheno- crysts of olivine are seen in most specimens . . . The texture is rough and trachytic, so that the rock would probably be considered an andesite or trachyte in the field. Most specimens are dense and very fine-grained or aphanitic, but vesicular forms may occur" (1. c, p. 485). This description might as well refer to the light grey basalts of Masafuera at elevations between i lOO and i 400 m, which I also, before an analysis was made, tentatively denoted as a trachy-andesite (12, p. 282). Although certain lavas of Hawaii evidently present similarities with the basalts from Juan I'ernandez, the general assemblage has, however, a different character. According to Washington "olivine-free labradorite basalts constitutes the most abundant type, followed in abundance by andesine basalt and then by oligoclase andesite' (Am. J. of Sc. 6, 1923, p. 355). The high percentage of andesine basalt and andesite denote a magmatic sequence differing from that of the non-alkaline lavas of more southern latitudes of the Pacific. It may, therefore, be advisible for the present to com[)ly with Lacroix when he says: "Les roches de cette ile (Hawaii) constitueraient une division speciale, ayant une originalite propre" (14, p. 76). If we take into consideration the assemblage of olivine and feldspar basalts and soda-trachytic lavas on Masafuera as a characteristic feature for this island it seems evident that, as Richardson already has assumed, the rocks of the (kila|)agos Archii)elago display the closest similarities. According to the analyses, j)ublished by Richardson, both the basalts and the soda-trachyte are in chemical composition very similar to equivalent rocks of Masafuera. Also soda-trachyte is f)f the same scarce occurrence in the Galapagos Islands as on Masafuera, the only sample being collected by Darwin on the Beagle voyage of 1835. Richardson says: "Juan T'ernandez are the only islands on which both types of basalt are found in addition to soda trachyte similar to that occurring in the Galapagos archipelago ' 1 27, j). 64). The low content of potassium is in common for the basalts from both island grouj)s. On the other hand we must evidently exclude any petrographical relationship between Juan I^'ernandez and San P'elix — San Ambrosio where the lavas have, as far as is known, a more j)ronounced alkaline composition, classified by Washington as soda-trachytes and nc))hcline basanites (1. c, p. 382). Richardson says: "The Juan I3 FrO , Mn( > . CaO , I'.aO. . M.^O Nii,,( ). K,() . IV>. . I' . . . . a .. . I 11 III 1\- y VI VII VIII 4r..2i 4().()H 47-S3 48.60 47.36 43-34 45-27 63.36 349 3.83 312 3-49 3-31 103 2.79 0.28 If). 03 i^'.75 1372 16.57 16.99 8.48 18.02 18.62 27^> 5-55 5-49 4.84 4.38 2.91 7-15 2.78 8..S6 7-34 6.90 7.29 8.32 10.99 7.38 1.02 0.23 024 0.1 1 0.18 0.18 0.13 0.07 0.09 10.53 10.78 10.33 10.82 9.01 0.01 6.62 5-03 6.34 1.68 0.55 4M 9.15 4.50 25.91 8.96 1.38 3.62 2.98 2.39 2.65 2.85 1.33 2.64 6.76 1.36 0.89 0.65 0.69 0-37 0.58 0.77 3-81 0,32 0.30 0.30 0.37 0.47 0.19 0.28 0.18 — 1 — ■ — — 0.14 — — 1 0.15 0.08 0.08 0.19 0.04 1 100.00 1 100.00 100.00 100.00 100.00 1 00.00 100.00 100.00 Molecular proportions (x 100) Sio, !•>.< \ !-•<•( ) .MnO (■;!( ). r.ao .Mk<> Na/) IV \ CrV'a I" ... CI . . Ill 76.94 4-36 15.72 i-73 12.33 0.32 18.84 16.24 5.84 1.44 0.22 [V 77.72 4.81 16.43 3-47 10.22 0.34 19.22 11.51 4.81 0.94 0.21 79.64 3-9' 13.72 3.44 9.60 0.15 18.42 22.69 3.85 0.69 0.21 80.92 4.36 16.25 3.03 10.15 0.25 19.29 1 1. 16 4.27 0.73 0.26 VI VII 78.85 4.13 16.67 2.74 11.58 0.25 16.06 0.01 16.17 4.60 0.39 0.33 0.79 0.23 72.16 1.29 8.32 1.82 15.30 0.18 8.97 64.26 2.14 0.62 0.13 0.23 75-37 3-48 17.68 4.48 10.27 o.io 11.30 22.22 4.26 0.82 0.28 0.09 0.54 105.49 0-35 18.27 1-74 1.42 0.13 2.99 3.42 10.90 4.04 0.13 0.1 1 ADDITIONAL COMMENTS ON THE GEOLOGY OF THE JUAN FERNANDEZ ISLANDS 85 Norms. Q C . or ab an ne di hy ol ap il . ml hm cm sal. fern II 8.08 18.14 23.48 6.76 21.78 10.46 0.74 6.62 4.01 56.39 4361 0.14 5-23 25.22 29.71 17-39 6.28 0.71 7.30 8.03 60.30 3971 III 0.40 3.84 20.19 24.81 19.32 16.84 0.71 5-93 7-97 49.24 50.77 IV V VI VII 4.25 4.06 22.39 31.29 16.00 7.50 0.87 6.62 7.02 61.99 38.01 0.66 2.17 24.12 32.49 7-43 19.20 I.I I 6.27 6.34 59-44 40.35 3-45 1 1.22 15-47 6.65 3-51 53-07 0.44 1.96 421 30.14 69.84 1-99 4.56 22.33 29.62 23.40 1.58 0.67 5.28 10.37 0.20 58.50 41.50 Quantitative system. I. Ill 3 4 Camptonose or :ab an II. III 4 4 dosodic or ab an III. III 4 4 dosodic or ab an IV. III 4 4 dosodic or ab an V. III 4 5 persodic or ab an VI. IV I 4 doferrous or ab an VII. III 4 4 dosodic or ab an VIII. I 2 4 Laurvikose or ab an 16.14:36.55 : 47.31 8.69:41.92:49.39 7.86 : 41.34: 50.80 7.03:38.78:54.19 3.69:41.03: 55.28 11.44:37.23: 51.33 8.07:39.51 : 52.42 25.91 : 65.88: 8.21 Niggli values. 4-49 0.79 22.48 57.15 7.12 3-43 0.45 0.53 2.78 0.86 91-93 8.05 si . . qz . al .. f m . c , . alk c/fm ti . . P •- k . mg. o. . . w . 103.7 -35-5 21.2 43-6 25.4 9.8 0.58 5.88 0.30 0.20 0.50 O.II 0.21 1 10.4 -22.4 23-3 41.2 27.3 8.2 0.66 6.83 0.30 0.16 0.40 0.24 0.40 III 105. 1 -18.9 17.8 51-9 24.3 6.0 0.47 5.16 0.28 0.15 0.58 0.17 0.41 IV 118. 7 -10.5 23.9 40.5 28.3 l-7> 0.70 6.40 0.38 0.15 0.40 0.22 0.37 110.7 23-4 47.0 22.6 7.0 0.48 5.80 0.66 0.08 0.48 0.16 0.32 VI VII 69.8 41.0 8.0 80.6 8.7 2.7 O.II 1.25 0.13 0.22 0.77 0.04 0.19 99-7 27.1 23-4 550 14.9 6.7 0.27 4.60 0.26 0.16 0.53 0.22 0.46 VIII 236.3 + 2.3 40.9 18.9 6.7 33-5 0.35 0.78 0.29 0.27 0.40 0.41 0.69 86 PKRCV QU KNSEL Bibliography. 1. 'V\i. Sr 1(1 iFKK. Criisoiiiana. Manchester 1843. 2. C. Hkrtkko. Notice siir I'Histoire naturelle de Tile Juan Fernandez. Annales des Sciences Naturelles, Paris ICS30, I'ome XXI, p. 345. 3. A. Cai.dc i.FiuJH. On the (leology of the Island of Juan Fernandez. Geol. Soc. of London, Proceedings, Vol. i, 1826 — 1833, p. 256. (Also published in Phil. .Mag. and Annals of Philosophy, Vol. IX, 1831, p. 220.) 4. P. Pakkkr Kin(;. Narrative of the surveying voyages of H. M. S. Adventure and Beagle. Vol. i. Proceedings of the first ex])edition 1826 — 1830, p. 304. 5. J. DiMoNi dTrviiik. \'oyage au Pole Sud et dans I'Oceanie. Histoire du Voyage, III, 1842. }) . 114. 6. I. (1ran(.k. (leologie, Mineralogie et (leographie physique du Voyage. Dumont d'Urville, Voyage au Pole Sud et dans I'Oceanie, 2^ Partie, 1854, p. 39. 7. L. Pi A IK. Zur Kenntnis der Insel Juan Fernandez. Verh. der Gesellschaft fiir Frdkunde /.u Berlin. Band XXIII, 1896, p. 221. - ]■ Bkm(;(;kn. Inindainentos de la (ieologia de Chile. Santiago de Chile 1950, p. 59, 17. 'P. II. Ti/Akii, H. N. MosKiKv, J. G. Puchanan and J. Murray. Narrative of the cruise of II. M.S. Challenger. Vol. I, 1885, p". 818. >8. T. SiK i.iKiK. The eartluiuake of Juan Fernandez as it occurred in the year I S3 5. London i.S3(). Sixteen years in Chile and Peru from 1822 to 1839. London i84i,p. 387 (with a sketch ot the sub-marine explosion). i'w. Darwin. ( ;colo,L;i( al observations on the volcanic islands of South America, visited during the voyage of H. M. S. P)eagle. London 1876, ]). 144. I'. Bar in. Pa( ificite, an anemousite basalt. Journ. Washington Acad, of Sc. Vol. -X X, I ()3o, p. 60. - - Mineral()gi( al Petrf)^ra])hy of the Pacific lavas. Am. J. of Sc. Vol. XXI, 1 93 I , ]). 3 So, 510. L. BowKN. The evolution of igneous rocks. Princeton 1928, p. 163. ADDITIONAL COMMENTS ON THE GEOLOGY OF IHE JUAN FERNANDEZ ISLANDS 87 24. A. JoHANNSEN. Petrography, 111, j). 334. 25. E. DE Margerie. La face de la 'Jerre. Ill, ]). 1359. 26. Bailey Willis and H. S. Washington. San Felix and San Ambrosio, their geo- logy and petrography. Bull. Geol. Soc. of America. Vol. 35, 1924, p. 365. 27. L. Chubb and C. Richardson. Geology of Galaprgos, Cocos and Kaster Islands. Bernice P. Bishop Museum, Honolulu. Bull. 100, 1933, p. 43, 47, 64. 28. C. BuRRi. Chemismus und provinciale Verhiiltnisse der jungeruptiven (iesteine des pacifischen Oceans und scine.'r I'mrandung. Schw. Min.-petrogr. Mitt. Band 6, 1926, p. 177. Unfortunately an interesting paper by Gorden A. Macdonald on the 'Hawaiian Petrographic Province', published in the Bull. (ieol. Soc. of America (60:2, 1949, p. 1588), in which comparisons with Juan Fernandez and other Central Pacific Islands are discussed, has evaded my attention until this paj)er was already in print. 4- A Geographical Sketch of the Juan Fernandez Islands. By C. SKOTTSBERG. The Juan Fernandez Islands were discovered on the 22nd November, 1574, by the Spanish navigator JUAN FERNANDEZ who called them Las Islas de Santa Cecilia. They consist of two islands, distant from each other, Masatierra with its satellite Santa Clara, and Masafuera. Masatierra lies 360 miles W of Valparaiso, Masafuera 92 miles W of Masatierra. According to the charts the position of the light in Cumberland Bay on Masatierra is 33°37'i5" S. and 78°49'5o" W., and of the summit of Masafuera, 33^46' S. and 8o°46' W. The islands are of volcanic origin and considered to be late Tertiary. They show no signs of recent activity, but a submarine eruption near Pta Bacalao in Masatierra is reported by Sutclifife to have occurred in 1835, and another E of this island by Goll in 1839 (Bruggen pp. 326, 332). Sutclifife (i, Plate p. 387) published a drawing of the eruption; the landscape is a pure flight of fancy. Some visitors have wanted to recognize a number of extinct craters. To this ques- tion I shall return later. When Ulloa thought that he saw flames bursting from the summit of Mt. Yunque, he certainly made a mistake. No geographer has, as far as I know, visited the islands, but many notes on their configuration and topography are found in the narratives of early naviga- tors as well as in the official reports to the Oficina Hidrografica in Valparaiso by the Commanders of surveying ships. Certain observations on the former dis- tribution of the forests were referred to in an earlier paper (Skottsberg 3). Many popular descriptions of the nature and life on Masatierra have appeared (see Bib- liography), some also paying attention to Masafuera. The latest, by JORGE Guz- man Parada, contains much useful material and will often be referred to here. Comments on some earlier descriptions and maps of Masatierra. The most interesting account of this island from the i8th century is found in Walter's narrative of Captain (later Lord) Anson's voyage. The illustrations are, even if not quite so accurate as the author thinks, vastly superior to the contemporaneous ones in Ulloa's work. Plate XIV is a prospect from ¥., including Santa Clara (called. Goat I.), the rock El Verdugo (Monkey Key) and part of the north coast of Masatierra, seen under almost right angle and with the con- spicuous mountains in correct position. Plate XV is a map, not bad in its main 6-537351 The Nat. Hist, of Juan Fernandez and Easter Isl. Vol. i. 90 C. SKOTTSBERG features; of the mountains only Mt. Vunque appears. The names on the map are Monkey Key, luist Bay (Pto' Frances), the Spout (a cascade not far from Pta Bacalao). West Hay (Pto Ingles) and Sugarloaf Bay (Vaqueria). Woods cover the east half; the treeless west half erroneously includes the still well wooded Villa- gra valleys. Plate W'l is a Special of Cumberland Bay, of which PI. XVII gives a good view, and PI. Will shows the Commodore's camp in the valley later named in commemoration of his visit. Masatierra was the rendezvous of Anson's scjuadron and brought salvation to the remnants of the crews, of which the greater [)art had fallen a victim to scorbut. The winter months of 1741 were spent here and the ships refitted. Am'omo 1)E Ui.i.o.v's narrative is accompanied by a panorama of the south side of Masatierra showing Mt. Yunque, Mt. Piramide, Co Negro and Damajuana, but other details cannot be identified. The map (Plate IV) is a rough sketch. Three bays have names, Puerto del Ingles, Englishman's harbour, very likely named to commemorate Selkirk as the cave called "Robinson's grotto" is found here, Puerto Grande de Juan P'ernandez (Cumberland Bay) and Puerto de Juan P'ernandez (Pto Frances). Some other (nameless) coves are indicated, e.g. Pangal. Three rivers empty in the harbours. I do not know the circumstances under which the survey by F^RANCISCO Am.\I)()R de Amava was made. It resulted in a map published in 1795 which has formed the basis of the charts still in use, but it may not have been known to Tho.mas Sutcijffe, whose book "Crusoniana" (1843) is accompanied by a map with more details than the older ones; with regard to the coast line it is inferior to Anson's. Sutcliffe was Governor of the islands in the 1830's, There are many names, but as I have not seen Amaya's original map I do not know which are new. Cumberland Bay is called Port of Juan Fernandez; W^est Bay, Ulloa's Puerto del Ingles, Selkirk Bay, and East Bay F'rench Bay. Sugar-loaf Bay (V^aqueria) is called Sandal Bay, an interesting name; perhaps most of the sandal- wood was obtained here in Sutcliffe's time. West of this place we find Desola- tion Bay, a well chosen name; now called Bahia Juanango. Herradura, undoubt- edly an old S{)anish name, is known now as Bahia del Padre; La Punta is Pta de la Isla. The east ca[)e, now Pta or Cabo Hueso de Ballena, is called Pta de Juanango. On the south coast we find Caravajal (Carvajal), Loberia, Villagra, Chamelo and Monkey I. These names are, however, misplaced. Sutcliffe's Carva- jal is Bahia Tierra Blanca, a name placed by him inland at the foot of the hills (where it belongs), the two bights on both sides of "Loberia", B. Chupones and B. X'lllagra, are nameless; the former is also called Tierras Amarillas on some charts, a name used by Sutcliffe for a tract of land back of his Tierras Blancas. X'illagra is located east instead of west of Mt. Yunque, and Chamelo used for the coast now called Playa Larga. The interior shows some topographical features; a mountain range can be followed from east cape to beyond the misplaced Yunque, and north of this is a short row of hills, corresponding to Cordon Central, which ' Abbrcridtions. I'>. r.ahia bay, C. - Cordon (range, ridge), Co = Cerro (mountain), L. = Loberia scaling grounds, M.-Morro small islet, rock), Pta = Punta (point, cape), Pto = Puerto (port, harbour, (). =Quebrada j-iarrow valley, gorge), V. = Valle (valley). A GEOGRAPHICAL SKETCH OF THE JUAN FERNANDEZ ISLANDS 9I separates "Anson's vale" from "Lonsdale" (now Valle Colonial). A name not found on any other map is "Kay's town", the settlement in Cumberland Bay. This name was given by Sutcliffe in commemoration of one JOMN Kay who, through his technical skill, greatly furthered the textile industry in England and whose biography appears in "Crusoniana", The Salsipuedes ridge and the ridge between Pto Ingles and Vaqueria are marked, while the conspicuous crest unit- ing Yunque and Salsipuedes has disappeared altogether. The topography of the western section is poor, only Tres Puntas placed in correct position. The name "Puente" is misplaced, but certainly refers to the elevated isthmus between Car- vajal and Herradura. Some later surveys and maps. From time to time the Chilean Hydrographic Office despatched a vessel to the islands as part of the work on a "derrotero" for the entire coast of the Re- public. The reports were published in the Anuario Hidrografico de la Marina de Chile. Lopez (1876) mainly repeats older statements with regard to distances, size of the islands, altitudes etc., ViEL (1878) concentrated his attention on the possibilities of making Masatierra productive, Vidal Gormaz (1881) little more than copied Lopez. The chart was not much improved. More information on the nature of the coast, the serviceableness of the harbours and anchorages, landmarks etc. are found in the compiled "Instrucciones nauticas" of 1896. For Cumberland Bay the original Spanish name Bahia San Juan Bautista is used, and some other early names are preferred, Bahia del Este, B. del Oeste, Pan de Azucar (Sugar-loaf, also Cerro Alto) and B. Pan de Azucar (Vaqueria), etc. Gunther's report of 1920 has little to add to the Instrucciones. A new chart had now been published and is reproduced in a very small scale. The distances between certain points indicated by Giinther agree rather well, with regard to the east section of Masatierra, with those on my map, while considerable dif- ference is noted in the length of the long, narrow western section, 12.96 km ac- cording to Giinther, 10.25 on my map, so that the total length between Pta de la Isla and Pta Hueso Ballena becomes 22.2 and 18.5 km, respectively. From American, French and English sources the well-known editor of geo- graphical and nautical works L. Friederichsen of Hamburg compiled a new map to accompany Ermel's popular account of his visit to Masatierra (1889). The central portion is much disfigured, but the general trend of the mountain ranges more or less correct, the details, however, erroneous in many cases. Most of the names used are Spanish. Some are still in use on the British and Cliilean charts, where, however, Punta is used for Cabo: C. del Padre, C. Tunquillar (Tinquillar), C. Lemos, Morro Juanango, C. de los Negros (now also called Pta Suroeste), B. de la Vaqueria, C. Salinas, Sal si puedes, C. San Carlos, C. Loberia, C. Bac- alao, C. Pescadores, C. Frances, Corrales de Molina (a series of hanging gorges E of Mt. Yunque), Morro Vifiillo, Bahia Chupones, C. O'Higgins. Some of the names on Friederichsen's map are now forgotten: Bahia de la Fe ( = B. Juanango), El Palillo (west head of Pangal), C. Madurgo (W of the east cape, here called 92 C. SKOTTSBERG C. Giiasabullena, a corruption of Hueso de Ballena), Morro Caletas ( = EI Verciii. Not endemic (4): Turdus magellanicus, luistephanus galeritus, Asio flam- meus, Pufhnus creatopus. II. South Pacific group. — 5 sp. a. l^ndemic (3): Pterodroma externa externa, cooki defilippiana and cooki masafuerae. i). Xot endemic (2): Fregetta grallaria, Pterodroma neglecta. The fust grouj) includes of more tropical birds jhiaeretcs, Cinclodes and Eu- stiphaiius, and of more temperate Turdus, Asio, CevcJuieis, Buteo and Puffiuus. Of the endemic species luistepliauus fernandcnsis is the most notable, in certain characters a unicpie t}'pe in the family Trochilidae. The second group is of par- ticular interest as including, beside the widespread frigate-bird, four species of Plcrodroina not breeding on the mainland, where, perhaps, a special race of /'. cooki breeds. The genus is essentially austral-circumpolar, as it were tricentric, w ith Tristan da Ounha representing the African sector. Cases like those of /^. neglecta and externa call for a common source and suggest that Pterodroma belongs to an Antarcto tertiary element which inhabited the coasts and islands of Antarctica n prjglacia! timjs. Oligochaeta. Mich Al.l.sKx (/untatus Simon. Patagonia and Fuegia. — Mt. A genus of 2 species, the second one from the Magellanian region. ^ Misumoiops Sjoestedti Berl. — Mt. An American, especially X. American genus. + (uiycjma Skottsbergi l^erl. — Mt. A S. American, especially Chilean genus of numerous species. (',. niaculaiipes Keyserl. Chile. — Mt. rOxysoma Delfini Simon. — Mt. A S. American genus. ^ Plulisca or)iata Berl. — Mt. A subantarctic-magellanian genus extending north into Chile. ^/V/. iugeus Herl.— Mt. ■r Lycosa I'CDiajidezi F. Cambr. — Mt. The genus is cosmopolitan. lii'oplirys quilpiioisis Simon. Centr. Chile. — Mt. The genus is known from Centr. and S. America, ICuroj^e, S. Africa and Japan. Of the 19 species enumerated 13 or perhaps 14 are endemic in the islands; 15 (11 or 12 endemic) are restricted to Masatierra, 3 (2 endemic) to Masafuera and a single Chilean species found on both islands. The only conclusion we can draw from these figures is that most likely only a minor part of the spiders occur- ring on the islands is known. It is quite possible that there is a marked differ- ence between the two islands, but it is not probable that Masafuera is so poor and that Santa Clara is devoid of spiders. Only a short visit was paid to this islet. The fauna makes the impression of being almost entirely neotropical or, at least, S. American, with the exception of the two species of Macrargus, which are of boreal parentage; of the doubtful /.^//^///^//r//;;/ nothing can be said. The {)resence of a southern, eventually Antarctic element is indicated by MecysinaucJic- )iiiis, possibly also by riiilisca, but so far there is no sign of a bicentric group. \\ hethcr it can be distinguished in subantarctic America I cannot tell. Acarina. TR.'u;.\Kl)n [268) enumerates 28 species, of which 2 are cosmopolitan, the re- mainder endemic. He points out that the collection, the first ever made in Juan Fer- nandez, undoubtedly represents onl}' a small part of the acarofauna; this is evident already from the fact that not a single sj^ecies came from Masafuera or Santa Clara. Whether the cosmopolitan sj)ecies are late arrivals or not is impossible to tell, but very likely they are. This would mean that the entire acarofauna is endemic, and new investigations will not change its independent character unless some species are discovered on the mainland. Of the 23 indigenous genera only one — probably DERIVATION OF THE FLORA AND FAUNA 297 a second will have to be described — is endemic and several have a wide distri- bution. The scant knowledge of this neglected group in these regions, particularly in Oceania, is to be regretted; it certainly does not yet lend itself to zoogeo- graphical speculations. Nevertheless it deserves to be mentioned that Euiergvs similis Trag. belongs to a genus hitherto recorded only from New Zealand and that Phyllhermannia deiitata Trag. is related to a neozelandic species; the genus is also found elsewhere. Pseudoscorpionidea (jp). + + Asterocherftes vittatus Beier. — Mt. The genus has its greatest resemblance to Thalassochernes Beier from New Zealand. + Chelanops insularis Beier. — Mt. + Ch. ktischeli Beier. — ^Mt. Related to a Chilean species. + Geogarypus bucculentus Beier. — Mt. 'V Parachernes kuscheli Beier. — Mt, Mf. + + Protoivirthius fernandezianus Beier.- — Mf. ^P. robustus Beier. — Mt. Neotropical elements are present, but species with their relatives in the Australian-Polynesian region are in dominance and part of the fauna shows not- able archaic characters (I.e. 205). Myriapoda. The very small and incomplete collection — no specimens were brought from Masafuera or Santa Clara — was studied by Verhoeff {2^4). In order to get some information on the distribution of the genera I asked Dr. Otto Sciiubart of Pirassununga, Brazil, for assistance, and he most liberally put his wide knowledge of this group at my disposal (letter, Aug. 27, 1954). Several changes had to be made in the nomenclature; the names used by Verhoeff, if different, have been put in brackets. Diplopoda. Brachyiulus pusilhis Leach (Microbrachyiulus litoralis Verh.). Indigenous in western Europe, adventitious in N. America and Argentina. Brachydesmus supertis Latzel. A European species, adventitious in N. America and Argentina. + Aulacode sinus insulanus (Verh.) Schubart (Semnosoma, Verh.). Endemic. A genus of 16 species, distributed over Chile and Argentina and belonging to the austral family Sphaerotrichopidae (S. Amer., S. Afr., Madag., Nossi Be, Austral., Tasm., N. Zeal., N. Caled., Hawaii). Nopoiulus veiiustus Meinert (pulchellus Leach). Widely distributed in Europe, introduced to N. America and Chile. Cylindroiulus frisius oceanicus Verh. Typical frishis {C. O^veni Bollman) introduced to N. America, Argentina (also in forma oceanicus), S. Africa and St. Paul's I. 298 C. SKOTTSBERG Chilopoda. + Xesoj^fop/ii/iis laticpllis (Attems) Schubart (Geophilus, Verb.). Endemic. The ^einis, wbicb has not been reported from S. America, includes after the latest revi- sion by Airr.MS (as subgenus of Geopliilus) i i species (i SAV. Austral., i N. Zeal, I X. Caled.. i An nam, 3 Jap. and 2 Eur.). r Xi'S(>o-i-()p/iiii(s biU-ckstrocini (Verb.) Schubart (Geophilus, Verb.). Endemic. Schizotoiia alaccr (Pocock) Silvestri. Chile, south to Fuegia, Argentina. A genus of 6 s[)ecies (3 Chile and Argent, to Patag., Eueg., i E. Austral., 1 N. Zeal., I Chatham Is.). Litliobio)>wyplia afyicaua Porat (Lamyctes insignis Pocock, insignisbaeckstroemi X'erh.). Widely distributed over Africa; also Tristan da Cunha, St. Paul's I., S.W. Australia and Hawaii. The genus very wide-ranging [N. Amer., W. Ind., S. Amer. (also Chile), Afr., \\. Ind., Austral., Tasm., N. Zeal., Chatham Is., N. Caled., Kermadec Is., Guam, Hawaii]. If we exclude the 4 species regarded, rightly I presume, as introduced with the human traffic, 5 species remain, 3 of these endemic in Juan Eernandez. This is indeed a very small number, but in spite of being so few, they tell a story of an austral-circumpolar, presumably Antarcto-tertiary element. CoUembola. Of the 8 species distinguished by SciloTT [216), the first ever collected in Juan PY'rnandez. 2 inhabit Chile, 3 are known from various parts of the world and 3 endemic. As long as so little is known about the distribution of this group it does not lend itself to zoogeographical speculations. The occurrence of widely dispersed boreal species in S. America and other parts of the south hemisphere (Australia, New Zealand etc.) is noteworthy, but whether their wdde range is due to the great age of CoUembola or a result of later dispersal is unknown. Thysanura [222, 2pj). ■ Is(>U'pis))ia luiucctois Silvestri. — Mt, Mf. The specific epithet refers to the intermediate position between IsolcpisDia and ll€tcrolcpis7>ia\ the species is com- pared with forms known from Africa and Australia. ' h KuscJielocliilis Ocliai:;aviac W'ygodz. — Mt. A monotypical endemic genus related to Alloinacliilis and XesoniacJiilis from Australia, but not, as far as known, to an American genus. Among the Invertel)rates treated above the endemic leech offers great in- terest. Of Arachnoidea the Pseudoscorpionidea include a remarkable Antarcto- tertiary element, whereas the true spiders, strangely enough, are quite disap- pointing in this respect, even more so than the centipedes. Insecta. In order to get an idea of the zoogeography of the island insects I asked a number of specialists for information on the general distribution of genera and DERIVATION OF THE FLORA AND FAUNA 299 Species. For their readiness to supply me with the necessary data I am much obhged to Dr. Olof Ahlberg, Stockholm (Thysanoptera), Dr. KjELL Ander, Linkoping (Orthoptera), Dr. Per Brinck, Lund (Coleoptera), Mr. Nils Bruce, Gardby (Coleoptera), Dr. Lars Brundin, Stockholm (Coleoptera), Mr. Felix Bryk, Stockholm (Lepidoptera), Dr. W. E. China, London (Hemiptera), Dr. K.-H. FoRSS- LUND, Stockholm (Trichoptera), Dr. G. J. Kerrich, London (Hymenoptera), Dr. K. Princis, Lund (Orthoptera), Mr. Bo Tjeder, Falun (Neuroptera) and Dr. B. P. UVAR(3V, London (Orthoptera). JOHOW (ijo) enumerates 26 species of insects from Juan Fernandez; some finds may, I presume, have escaped his notice, but probably not many, and it is evident that the entomofauna was very little known at that time. During our survey 19 16-17 a fair number of insects were collected and many novelties were described in vol. 3 of this work, but the collection gave the impression of being very fragmentary. The intense collecting undertaken in 195 i and 1952 by the Rev. Dr. GuiLLERMO KUSCHEL revealed, however, the existence in the islands of a surprisingly rich and varied insect world. As Dr. ALEXANDER, the wellknown specialist on Tipu- lidae, expresses himself (^.35): "Father Kuschel's collecting has completely revolutionized our knowledge of the insect fauna of the islands in many groups, in- cluding the crane-flies" — only 3 species were known, the number now amounts to 37. Until now only a part of Dr. Kusciiel's large material has been worked up by specialists, and I can only refer to what has been published (142, 208, 2g2, 2gj, jog, 314), but for some groups we now have sufficient data to form an opinion of the zoogeographical position of the islands as far as the insects go. At the end of 1954 Dr. KusCHEL joined my new expedition to the islands and brought back a third very large collection. When all his material has been studied, the insect fauna of Juan Fernandez will be better known than that of most isolated islands. At present about 340 indigenous species have been recorded, of which about 230(70 %) are regarded as endemic. Dr. KuscHEL (letter, Oct. 16, 1955) calculates that of a total of about 600 species collected by him, about 360 still await publication- Among them are 25-30 flies, probably over 50 butterflies, many endemic, at least 180 beetles (more than 120 weevils, of which 4 have been introduced accidentally, the remainder being endemic), and some 40 hymenopters. Orthoptera. Dermaptera [22j). + Euborellia annulipes (Lucas). — Mt, SC, Mf. The genus S. Amer., E. Afr., Orient, Ind., Ceylon, Tasm. Anisolabis Bormansi Scudd. Galapagos Is., Easter I.— Mt. A large genus of world-wide distribution. Saltatoria [j8, 22j). ^Hoplospyriuin Skottsbergi Chopard.— Mt. An American genus, the species related to species from N. America and Chile. Trimeroiropis ochraceipeimis Blanch. Chile.— Mt. The genus is American. 300 C. SKOTTSBERG Corrodentia. Isoptera. - Kiiloti-nfics i^^racilii^^Jicitus luiierson. — Mt. The only Termite known from Juan Fernandez. "The wing venation is close to that of Kalotermes broiini Froggatt from New Zealand" [8j. 393). Mallophaga [266). Puffimis creatopus and Ptcrodroma neglecta and exterua are infested with the same mallophagous parasites found on these and related birds in other regions. A new Halipcurus is mentioned but not described. Thysanoptera (7). Aeolothrips fasciatus L. Boreal. — Mt. -X- Physotlirips Skottsbcrgi Ahlb. — Mt. The genus is distributed over N. Amer- ica, luirope, W. Asia and Australia. -SencotJirips vieptiis Ahlb. — Mt. Perhaps nearest to a Californian species. The genus is otherwise confined to F^urope, where it is widely spread. rinips tabaci Lindem. X. America. — Mt. The genus known from N. America, luirasia, \. Africa and Australia. The two non-endemic species may have been introduced accidentally. Both were found in the spathe of Zantcdcschja acthiopica, cultivated and naturalized. Neuroptera (zcVf, g2, I2j). + + Couclioptcrella kusclieli Handsch. — Mt. + 6'. Diacidata Handsch. — Mt. (layoinyia falcata (Blanch.). Chile, Argentina. ^ — Mt, Mf. A small S. American genus. ' I Ic))U'yobiiis Siocstedti Xavas (M. fumosus l^>sb. -Peters., H. nigrinus F"sb.- Peters.). — Mt. An almost world-wide genus, absent from the S. hemisphere except for the Andean region (Colomb., Perii, l^oliv., Ikaz., Argent., Chile). -i //. Skottsbn-gi Xav.-is.— Mt, Mf. Trichoptera [21^. + Australoinyia inastUierra Schmid. Mt. The genus is known from Chile^ Patagonia and I^'alkland Is. ■\-A. luasafuera .Schmid. — Mf. I'ergcr Por/iri Xav. Centr. Ciiile. — Mt. A Chilean genus. Lepidoptera (tJ). A great number of genera and species will have to be added when Dr. Klsciikl's material has been determined. DERIVATION OF THE FLORA AND FAUNA 3OI Tineomorpha. Gelechidae. + + Apotheioeca synaphrista Meyr. — Mt. The genus is closely allied to the large and widespread Gelechia (N. and S. Amer., Galap., Palearct., Macaron., S. Afr., Australia). Oecophoridae. -\-Depressaria relegata Meyr. — Mt. Near a species from the Andes of Ecuador. The genus is Holarctic and also found in S. Africa. Endrosis lactella Schifif. — Mf. Widely spread, domestic. Tineidae. Monopis crocicapitella Clem. — Mt, Mf. In most parts of the globe, domestic. Pyralimorpha. Pyralidae. + Crambus fernandesellus Hamps. — Mt. A world-wide genus. Elasmopalpus lignosellus Zell. Centr. and S. Amer. — Mt. Ephestia kuehniella Zell. Widespread, domestic. + "^ Fernandocrambiis Baeckstj'oemi Kwx'w . — Mt, Mf. The genus nearly related to Crambus. + F. brunneus Auriv. — Mt. + F. fuscus Auriv. — Mt. + +yuania anmdata Auriv. — Mt. Similar in some ways to Ptochosiola Meyr. (S. Afr., Australia). Nomophila 7iociuella Schiff. Cosmopol. — Mt, Mf; adventitious. ■\-Pionea fiunipefinis (Warren) Hamps. — Mt. A world-wide genus. Scoparia Ragonoti Butl. — Mt, Mf. Chile. A very widespread genus. Geometfina. Geometridae. ^ Eupithecia halosydne Prout. — Mt. A widespread genus, but not found in Australasia. + E. (.?) znepta Prout.— Mt. -\-E. physocleora Prout. — Mt. ■\-Lobophora insulai'is Auriv. — Mt. An essentially Palaearctic genus. Tortricidae. + Crocidosema insulana Auriv. — Mt. A S. American genus. ■vEulia griseiceps Auriv. — Mt. The genus Holarctic, also in Hawaii; few else- where. ■\-E. Robinsoni Auriv. — Mt. + E. striolana Auriv. — Mt. Noctuina. Noctuidae. Copitarsia turbata Herr.-Sch. Venezuela, Colombia. — Mt. The genus in Mex., Centr. and S. Amer. (Argent., Chile). Feltia malefida Guen. Amer. (south to Chile); Macaronesia. — Mt. 30. C. SKOTTSBERG -V ~ Hoplotaysia inai^iia Auriv. — Mt. Related to Copitarsia. Li'ucauia inipuiicta (iucii. Chile.— ^It. A bipolar genus (Palaearct., N. Zeal.). -^ Lvcopltotia luuiks/roojii Auriv. — Mt. The genus widespread (Arct., Amer., luir., S. Afr., Madag., X. Zeal.). L. DitssiiDH Gucn. Chile, Patag. — Mt. Kac/nplusiii uu (iucn. I'atag.. Urug., Argent., Chile.— Mt. Sv>ii:-yapha i^-auiDioidts Hlanch. Chile. — Mt, Mf. A Palaearctic genus; Mex., S. Anicr. Rhopalocera. Pvrtv/ii/s cnrvi- Iluebn. Widespread in S. America and probably introduced in Juan bY'rnande/.. — Mt. Diptera [84, 142). Acroctyidac [208^. -V Opcodes knsclicli Sabr. — Mt. A temperate genus, recorded from all con- tinents, but only 2 species known from the mainland of S. America. AiitlioDiyzidae [12S]. Anthony zii cursor (Kieffer). Cosmopol., also S. Chile. — Mt. Calliplioridac {2-;-;) . L 'ally}itropyga huuteralis (Walker) Souza L. et Alb. (C. Selkirk! Enderl), Chile C<)ncej)ci6n). — Mt. Ml". raraludlia fuh'icrura (Desvoidy) Aub. et Baxt.— Mt, SC, Mf. Sarcojicsia cliloroi^astcr (Wiedem.) Arrib. Chile, I{^aster I. — SC, Mf. Sarc())ies}onii)na hicolor Souza L. et Alb. Chile (Santiago). — Mt, SC. A mono- t\-pical genus. Cecidoniyidae [84, 106). + ^- Tsadaria pallida iMulerl. — Mt, Mf. Related to Canipyloniyza Meigen. (. hloropuiac [2()g). Ilippilates ausiralis Sabrosky [//. (Cadrema) nietallicus P3nderl. non Beck. [^ II. flai'ipcs (Loew) Sabr.)]. I^xuad., Peru, Argent., Chile. — Mt, probably ad- ventitious. L uliiidae. Ciilcx intcrfor Dyar. — Mt. Mf. J) (die hop odidac [126). ^ Ilyilrophorus hu.ululi llarmston. — Mt. //. polioi^asUr (I'hil.) llarmston. Chile. — Mt, SC. rSy)npyc?ins fdiiandtrjrnsis llarmston. Near a Chilean sp. — Mt, Mf. liphydridac [28(/). Diniciocnia cars/a Iv.d. Wulp) Wirth. Argent., Uruguay.— Mt, SC. -V Discoceriiia fu))iipc)nus Wirth. — Mt. Near a Chilean sp. Ilyadina certa Crcsson. Chile. — Mt, Mf. Ilydrellia vulgaris Cresson. (iuatem., l^oliv., Chile. — Mt. DERIVATION OF THE FLORA AND FAUNA 303 + Scaiella a7igustipennis VVirth. — Mf. An almost Cosmopolitan genus. + 5. argeniifacies Wirth. — Mt. + S. brachyptera Wirth. — Mt. + 5. decemguitata Wirth. — Mt, SC, Mf. ^S. discalis Wirth. — Mt. + 5. fernandezensis Wirth. — Mt. + 5. kuscheli Wirth.— Mt, Mf. + S. lutea Wirth.— Mt. ■vS. niarginalis Wirth. — Mt. + 5. niasatierrensis Wirth. — Mt. ■vS. minima Wirth. — Mt. + 5. 7ia7iopiera Wirth.— Mt. + 5. pallida Wirth.— Mt. + S. pilimana Wirth.— Mf. + S. stenoptera Wirth. — Mt. + S. vittata Wirth. — Mt. + Scatophila fei'naftdesiana Wirth. — SC. vS. medifemur Wirth. Chile (Coquimbo). — Mt, SC. Heleidae [288). + Dasyhelea australis Wirth. — Mt, Mf. Near a Chilean sp., genus cosmop. + Forcipo7Jiyia te7ttiisqica77iipes Wirth. — Mt. A widespread genus (N. and S. Amer., Eur., Afr., Austral.); one species common to Paraguay and Australia. + F. sa7ictaeclarae Wirth. — Mt, SC. Helo77iyzidae (128). Blaesochaetophora picticoimis (Bigot) Henn. S. Chile. — Mt. Prosopa7ttru7n flavifro7is Tonn. et Mall. (Cnemospathis Baeckstroemi et Schoenemanni Enderl.). Chile, S. Africa, New Zealand. — Mt, Mf. Lo7ichaeidae (128). Lo7ichaea patago7iica Malloch. Chile. — Mt. Mnscidae (128). Austrocoe7tosia ig7wbilis (Stein) Hennig. Chile. — Mt. Ci'aspedockaeta Ii77ibi7iervis (Macq.) Hennig. S. Chile. — Mt, Mf. Delia platiira v. sa7icti-jacobi (Bigot) Hennig. Chile. — Mt, Mf. Euryo77i7na peregri7iU77t (Meigen) Hennig. Peru, Chile. — Mf. Fa7mia a7tthraci7ia (Walker) Hennig. Chile. — Mt. F. ca7ialicularis (L.) Cosmopol., also in Chile. — Mt, Mf. F. pu7ictive7ttris Malloch. S. Chile. — Mt. Fucellia i7tte7n7iedia Lundbeck (Egeria masatierrana et masafuerana Enderl.), Eur., Oceania. — Mt, Mf. Hydrotaea cya7ieive7itris Macq. Chile. — Mt, Mf. Li77i7iophora patag07tica Malloch. S. Chile, Patag. — Mt. ■\- NotoschoeTiomyza kuscheli Hennig. — Mt, Mf. Ophyra caerulea Macq. Centr. Chile to Fueg. — Mt. 304 C. SKOTTSBERG -\-ScliO€)W})iyzJ}ia eiudejii liennig. — Mf. + Sylliifniopliora lispofi'nfia Hennig. — Mt, Mf. MycetophiUdac [100]. -. lixicJiia furcilla h'reeni. — Mf. Near a Chilean species; world-wide genus. - Leia tiiallcolns Freem. — Mt. Allied to a species reported from Bolivia, Peru and Brazil; the genus world-wide. Macrocera fujicrca Freem. Chile.^ — Mt. A world-wide genus. -r Mvcetophila a)iij;ustifuyca P^ndcrl. — Mt, Mf. The genus world-wide. M. co)iiftra Freem. Chile. — Mf. J/, corjiuta Freem. Chile. — Mt. iM. fliU'oluuata hVeem. Chile.— Mt, Mf. M. (.') iusccta iM-eem. Chile.— Mt. M. spiiiosa Freem. Chile. — Mf. + J/. subfiiDWsa Freem. — Mt. k- Paraleia iiephrodops (r^nderl. s.n. Selkirkius) Freem. — Mt. /'. Jiubilipcuuis Walker. Chile. — Mf. The genus neotrop., Austral., Tasm. SciopJiila ocJircata Phil. Chile. — Mt. A w^orld-wide genus. Piophilidae [12S). Piophjla case? (L.). Cosmopol., domestic. — Mt, Mf. P. foi'colata Meigen. Cosmopol. — Mt. Phoridae [S4). -f Lioyella juajifcDiajidczica Enderl. — Mt, Mf. A F^uropean genus. P/iryjuidae (i2cS'). Plirync fuscipcuuis Macq. S. Chile.— Mt, Mf. l^latypizidac. Microsaiiia pallipes Meigen. Cosmopol. — Mf. Psycliodidat- (cS'y, 212). PsycJioda ci}ierea l^anks. Cosmopol., also Chile. — Mt. A widespread genus. \ l\ t>nuat}eyyc7isis Satchell. — Mt, Mf. Possibly = the following. vP. ))U)nitissima ICnderl. — Mt. /'. sci>tyi)ii Tonnoir. Widespread in temperate regions. — Mt, Mf. Say cop hagidac ( i* -f -f ) . Jfypopy^Q^ia nay'ia (Walker) Townsend. Chile.— Mt, SC, Mf. Scatopsidac [S^, roo). ■^ ^ Masatieyya fcyyugbiea luiderl. — Mt. Related to the European Rhaeboza linderl. I^'kkemax does not mention Masaiicyya. Scatopse fuscipes Meigen. — Mt, Mf. A world-wide genus. S. iioiata (L.). Cosmopol., introduced with the traffic. — Mt. Sciayidae [ioo\ + Brcidysia fnsca I^Veem. — Mt, Mf. A world-wide genus. DERIVATION OF THE FLORA AND FAUNA 305 + B. media Freem, — Mt, Mf. + Merianina kusckeli Freem. — Mt, Mf. Another species in Brazil. + Psiiosciara nitens Freem. — Mt. Sunuliidae (2g4). + Gigantodax kusckeli Wygodz.— Mt, Mf. An essentially Chilean-Patagonian genus with single species as far north as Mexico. Sphaeroceridae {84, 204). Archiborborus submaculatus Duda. S. Chile, Patag. — Mt. + + Gyretria binodatipes Enderl. — Mt. The genus is perhaps identical with Skottsbergia Enderl. and this is merged into Leptocera by RICHARDS who, however, does not list the two Gyretria species described by Enderlein, but not found in Dr. KuscilEL's material. + G. crassicosta Enderl. — Mf. Leptocera brachystoma (Stenhammar) Richards. Cosmopol., also in Chile. — Mt, Mf. The genus is widespread and well represented in S. America. + Z. cultellipennis (Enderl. ut Skottsbergia) Richards. — Mt. L. darivini Richards. Chile, Argent. — Mt, Mf. L. divergeiis Duda. Peru, Boliv., Chile, Argent. — Mt, SC. + Z. duplicata Richards. — Mt. + Z. ellipsipemiis Richards. — Mt. L. flavipes (Meigen) Richards. Eur., N. Afr.— Mt. L. mediospinosa Duda. Cosmopol. — Mt. L. pectinifera (Villen.) Richards. Eur., Falkl. Is.— Mt. L. pulchripes Duda. Argent., Parag., Urug. — Mt. + +Phthitia alexandri Richards. — Mt. The genus must be very u^diV Leptocera. + Pk. selkirki (Enderl. ut Pterodrepana) Richards. — Mt. + /%. venosa Enderl. — Mt. Syrphidae [96). Allograpta exotica (Wiedem.) (A. Skottsbergi Enderl.). Neotropical. An American genus of at least 16 species, the majority in S. America. — Mt, SC, Mf. ^■A. robinsoniana Enderl. — Mt, Mf. Melanostoma fenestratum (Macq.) Fluke. Chile. — Mt. An American genus. + J/. Lundbladi (Enderl. ut Carposcalis) Fluke. — Mt. Eristalis tenax (L.). Cosmopol, also in Chile. — Mt, Mf. Sterphus aurifrons Shannon. Chile. — Mt. Tachinidae (70). Lrtcamyia chilensis Aldrich. Chile. — Mt, SC, Mf. A S. American genus. + Pkanfasiosiphona kusckeli Cortes. — Mt. A Centr. American genus. Tendipedidae (288). Anatopynia vittigera Edw. S. Chile, Patag. — Mf. A world-wide genus. -^Clunio fuscipenitis Wirth. — Mf. A large, wide-ranging genus. + Hydrobaenus fernandezensis Wirth. — Mt, Mf. The genus world-wide. 20 — 557857 The Nat. Hist, of Juan Fernandez and Easter Isl. Vol. I 3o6 C. SKOTTSBKRG //. pnitontni (Goett^eb.) Coe. luigland. Chile, Patag. — Mt. //. pusilliis (luiton) Coc. luiglaiul, Kerguelen. — Mt. -^ l\)tio>io)uus acutus W'irth.— Mf. A genus of numerous species in southern S. America, few elsewhere (luir., \. Amer.). rP. t/isiish'/us W'irth.— Mt. Mf. P. kirfi-ri ((iarrett) lulwards. l^rit. Columb., Chile, Eur.— Mf. rP. kuscheli Wirtii.— Mf. /'. )iii^rinNs lulwartls. — S. Chile. — Mf. + /'. Si'lkirki W'irth.— Mt. Mf. 'Pit?iytars/(S ffar/pis (Meigen) Townes. \. Amer., S. Amer., also Chile, luir.— Mf. 'J'i/'iilii/iu- (7). l-lrioptcra pilipcs (Fabricius). Cosmoj)ol.— Mt, SC, Mf. + Lii/io?iiij (I)icranomyia) affabilis Alex. — Mt, Mf. A very large and wide- ranging genus. ■tL. a))ipliio)iis Alex. — Mt. -f A. axierasta Alex.— Mt. + A. Jiarpax Alex. — Mt. + A. kuscluliaiia Alex. — Mt. -tL. iiiasafuerae Alex. — Mf. + A. pidcsti'is Alex.— Mf. + A. selhiiki Alex.— :\lt. Mf. + A. siuardoi Alex. — Mt. L. tyilubeyculata Alex. S. Chile, Patag. — Mt, SC (an endemic variety). + A. I'oiatrix Alex. — Mt. + A. luucris Alex.— Mf. + A. yioiqucinia Alex. — Mt. -V MolophilHs ainpliacajitlius Alex. — Mt. A world-wide genus well developed on the mainland. + J/. ivitiinonis Alex. — Mt, Mf. + J/. apprcssospiiius Alex. — Mt. + J/. arcifirus Alex. — -Mf. ■V M . civiopus Alex. — Mt. + J/. iiffocajiiis /\lcx. — Mf. ■\M. (iistifurcus Alex.- Mt. + J/. filioliis Alex.— Mt. + J/. films Alex.— Mt. + J/. uiasafucrac Alex. — Mf. + J/. )}U(ltifi(ius Alex. — Mt. + J/. fNpfiuii/s Alex. — Mt. + J/. pcclijiij'crus Alex. — Mt. + J/. rcctispiiius Alex. — Mt. -f J/, sclkirkiainis (Knderl. ut Archimolophilus) Alex.— Mt. . -fJ/. .sVj'.r Alex.— Mf. DERIVATION OF THE FLORA AND FAUNA , 307 + M. tridens Alex.— Mt. ■vM. variants Alex.— Mf. -{■M. yunqiiensis Alex. — Mt. + Skan?to?toinyia kuscheli Alex. — Mf. An American genus, well represented also in Chile and extending north to Canada. ■\-Sh. masatierrae Alex. — Mt. ■\-Sh. selkirkiana Alex. — Mt. + Tipula baecksU'oemi Alex. — Mt. The genus world-wide. Coleoptera. As yet little has been published about the beetles collected by Dr. KUSCIIEL; to judge from what is known the number of island species no doubt will be multiplied. Anobiidae [igs\ Anobiuni punctatMin De Geer. Cosmopol, introduced. — Mt. Beside the typical species an endemic ssp. described by PiC (Mt). A. striatum 01. Cosmopol., introduced. — Mt. + Calymmaderus atronotatus Pic. — Mt, Mf. Near a Chilean species. Numerous species in N. and S. America. + ^ Masatierrum inipressipeime Pic. — Mt, Mf. A genus near Megorama Fall., a small N. American genus. Stegobium (Sitodrepa) paniceum (L.). Cosmopol., domestic. — Mf. Xyletomerus pubescens ssp. kuscheli Pic. — Mt, the ssp. endemic. A north American genus. X. pubescois Md,x. piceitarsis Pic (fumosus var., Pic). — Mt. The variety endemic. A nth ribiidae [133] . + + Opisolia lenis Jordan. — Mt. Related to Eucyclotropis Jordan (Centr. and S. Amer.). Bostrychidae [166). Neoterius pulvinatus Blanch. Chile. — Mt. A small genus reported from Peru and Chile. Prostepha7ius sulcicollis Fairm. et Germ. Chile. — Mt. Carabidae (5, 260). Bembidimn inconstaris Solier. Chile. — Mt. A world-wide genus. B. punctigerum Solier. Chile. — Mt. Laemosthenes complanatus Dejean (Pristonychus, Gory, Andrewes). Cosmopol., also Chile; introduced.— Mt, Mf. + Metius eurypterus Putzeys. — Mf. A S. American genus, mostly in the far south. M. flavipes Dejean. Chile.— Mt, SC. + M. kuscheli Straneo.— Mt. + J/. ovalipennis Straneo. — Mf. jo8 C. SKOTTSBERG + Plerostichus kusclicli Straneo. — Mt. A world-wide genus. + /V. selkirki Aiulrcwes.— Mt. SC. + /V. skottshfyi:;i Andrewes. — Mt. + /V. H'alkiri Andrewes. — Mt. + + Iraclnsivus hasalis Straneo. — Mt. The genus presumably endemic [260. ■3«). ■T '/'. bicolor Straneo. — Mt. + r. emdi'iii Straneo. — Mt. -f /". kuscluli Straneo.— Mt. + 7". (>i'ij/ipiN}iis Straneo.— Mt. f '/'. pallipcs (ierm.— Mt, Mf. + /'. pH)ntii:;ey Andrewes. — Mf. + 7. sericeus Andrewes. — Mt. + Trechisihus bacckstfoe)}ii (Andr.) Straneo. — Mf. An American genus. 7". fe»ioralis Germ. Chile. — Mt (end. ssp.), SC. + 7". kusclicli Jean n el. — Mt. r / 'ariopalpus crusoci Keed. — Mt, SC. ( Inysoinclidac [2S2, iS]. ^ vMiJiotula fcffiivnicziiVia l^echyne. — Mt. The genus is related to Hyp7io- pJiila (W. ICur. -Japan). \-M. kuschcU Hechyne.— Mt. fj/. iiitois W'eise. — Mt. Cioidac [166). ■ ( is hi))iaculatiis Germ.— Mt. A world-wide genus. + C fcy}ia}idczia)ins Lesne. — Mt. r C '. nifus (jerm. — Mt. i Icridac [21^;). Xccrohia rufipcs De (ieer. Cosmopol. — Mt, introduced. C Occific/Iidnc [2iS2). liriopis opposita (nicr. Chile. — ^It. The genus ranges from Vancouver I. to Patag. and I'"uegia. C(dydiidac [2oiA. ^ /[ycHonicrodcs uiasd/ucroisis Poj)e. — Mf. lieside the two Juan Fernandez species there is a third in New Zealand. 'r I\ ni(isa/icyrc>!S/s I'ope. — Mt. V Pyotoincrus iiisularis (ironvelle. — Mt. The genus N. and S. Amer., E. Ind., Japan, .Australia, \. Zeal, (numerous), .Samoa. 4 7'. j^cnudiiii Pope. — Mf. C yyptophagidac {j;6]. Cyyptopliai::us atomayioidcs (iron v. (Selkirki l^ruce'). Chile.— Mt. A world- ' Letter 19.9. 1954. DERIVATION OF THE FLORA AND FAUNA 309 wide genus, the subg. Mjiionomus, where the island species belong, in Eur., N. and Centr. Afr., Macaron., Centr. Asia, probably also N. Amer. + C. Skottsbergi Bruce. — Mt. + C. splendens Bruce. — Mt. + + Cryptosomatula longicornis Bruce. — Mf. + Loberosche?na convexum Bruce. — Mt. Beside the island species 3 in Bo- livia and 2 in Chile. + Z. discoideum Bruce. — Mt. Curculionidae [12). Dr. Kuschel, who specializes in this family, estimates the number of species collected by him to exceed one hundred. + + Arwletkrus gracilis Auriv. — Mt. + -v Apteronanus dendroseridis Auriv. — Mt. + A. (.^) gunnerae Auriv. — Mt. Arajnigus Fulleri Horn. A widespread noxious beetle. — Mt. + Caulophilus (.?) nigrirostris Auriv. — Mt. The genus in southern U.S.A., Centr. and S. America. It does not exist on Juan Fernandez (KuscilEL in litt.). + Cyphometopus masafuerae Auriv. — 'Mf. 3 species in Chile, where the island species most likely also occurs (Kuschel in litt.). ■^ -^Juanobia ruficeps Auriv. — Mt. + ^Juanorhinus Robinsoni Auriv. — SC. Otiorrhynchus rugosostriatus Goeze. W. and S. Europe. — Mt, accidentally introduced via Chile. + -v Pachy stylus dimidiaius Wollaston. — Mt. + P. nitidus Auriv. — Mt. + Pachy tragus crassirostris Wollaston. — Mt. + Pentarthrujn affijte Wollaston. — Mt. A widespread genus, found on many oceanic islands. + P. nigropiceum (Phil.) Auriv. — Mt. + P. nitidum Wollaston. — Mt. + /^. rufoclavatum Auriv. — Mt. Close to P. ^//V^/^' Broun from New Zealand. + +Platynanus arenarius Auriv. — Mt. + P. Baeckstroemi Auriv. — Mt. + P. hirsutissimus Auriv. — Mt. + P. quadratifer Auriv. — Mt. + P. sericatus Auriv. — Mt. + /^ Skottsbergi Auriv.— Mt. + Strongy/optei'us nitidirostris Auriv. — Mt. The genus is also found in Chile and New Caledonia. S. ovatus Boh. Chile.— Mt, Mf. Dermestidae (104). Dermestes vulpinus L. Cosmopol., introduced. — Mt. Dytiscidae (2^p, ii?)- Afiisomeria bisiriata Brulle. Chile. — Mt. The genus known from Tristan da Cunha. 3IO C. SKOTTSBERG ^Lancctcs Baeckstrocnii Zimmerm. — Mf. Tlie genus austral bicentric. RliaJitHs sigjiatiis ssp. knsclieli Guignot. — Mt, Mf. An endemic variety of a Chilean species. Latlnidiidae [ig6). r Co)ii)ioi)ius curtipe}uiis I'ic. — Mt. Near C. dijuidiaius Belon (Boliv., Chile). An essentially luiropcan genus. C. suhfasciatus Reitt. Chile.— Mt. Milasidae [joS). ^ rscudodiiurctiis Sclk'nki Flet. — Mt. An Argentine genus. Mycetophaoidae [ic)6). Mvit'/op/iajrus c/iili?isis Phil. Chile. — Mt. The genus in Kur., Asia, Afr. and Amer. Xitidididae [nj6, loj). rC)iips acuta Ciillogly.— Mt. A Chilean genus. 4-r. atrata (iillogly.— :\Tt. t L '. dii'vrsa Pic— Mt. + 6". fi'Diandczia (jillogly. — ?^It. •C. niucrojiis (iillogl)-. — Mt. Scarahaiidae (J'//). Aphodius iiranariiis L. Cosniopol., domestic. — Mt. Plfuropliarus cacsus Creutz. Reported from X. Amer., Chile, Eur., Orient, .Madagascar. — Mt, introduced. Scolytidac \^^o^)). (iiiathotriclius tortliyloidcs Schedl. Chile. — Mt. The genus in N. and S. America. IVilocotribus ivillei Schedl. Peru and Chile. — Mt. Sti ip //] 'li)iida e ( J* -r ) . ■ Atluta Kob'nisoui l^ernhauer. — Mt. Near a Brazilian species. A cosmopoli- tan genus of about 2000 sj)ecies. ■ liltusis sonirufa h'airni. et (ierm. — Mt. A genus of about 1 50 species, S. Amer. (also Chile), .Austral., N. Zealand. \Mid(>)i irusdi-aiius Bcrnhauer. — Mt. A cosmoj)olitan genus of about 500 sj)ccics. r Ocyusa l^acckstrocDu Bcrnhauer. — ?^It. About 40, mainly Palaearctic. IViilo)itlius nitidiponiis Solier. Chile. — Mt. A cosmopolitan genus of about (Soo species. I 7'rooop/ti(>fns Sko/Zs/xr^i;-/ Bernhauer. — Mt. A cosmopolitan genus of about 350 -^P- I I'DuiocJiilidac ( / 9"^ ). + r/ianodi-s/a crihraria (Blancli.). Includes Ph. auirulata Reitt.— Mt, SC. A Chilean genus. + /'//. robust a Pic. — Mt. + /V/. rarici^ata Germ.-Mt. DERIVATION OF THE FLORA AND FAUNA 3II Tenebrionidae [lO/f., ig6). Blapstinus punctulatus Solier. S. Amer., also Chile.— Mt, SC. Numerous spe- cies in N. and Centr. Amer., some in S. Amer. ^Enneboeus Baeckstroemi Pic— Mt. Near a species from Panama. The genus Mex.-Colomb., Tasmania. Nycterinus gracilipes Phil. Chile.— Mt. Numerous species in Chile. Hymenoptera. Aphelinidae (211). Aphelinus jucmidus Gahan. N. America. — Mt, Mf. Bethylidae [igi). + Cephalonomia skottsbergi Brues. — Mt. + +Lepidosternopsis kuscheliana Oglobin. — Mt. + Perisieroia maculicornis Oglobin. — SC. A widely distributed genus of about 25 species. + P. sanctae clarae Oglobin. — SC. Braconidae [i8y). + Apa?iteles evadite Nixon. — Mt, Mf. A wide-spread genus. ^■A. morroensis Nixon. — SC. Aphaereta minuta (Ns.). — Mt. + Opms kuscheli Nixon. — Mt, Mf. ^0. scabriventris Nixon. — Mt. Dryijiidae (igi). ■^Haplogonatopns insularis Oglobin.^ — Mt. A genus of 6 species (N. Amer,, J. Fern., Australia, Pacif. Is.). + + Idologonatopus nigrithorax Oglobin. — Mf. A genus related to the former. Elachertidae [211). + +Kus chelae her ius acrasia De Santis. — Mt, Mf. ■^ Pseudelachertus semijiavus De Santis. — Mt. The genus otherwise Aus- tralian. Encyrtidae [211). + Hemencyrtus kuscheli De Santis. — Mt, Mf. A neotropical genus. Eniedontidae (211). + Achrysochris bicarinata De Santis. — Mt, Mf. Eupai^acrias phytomyzae (Brethes) De Santis. Chile, Argentina. — Mt, introduced. + Omphalomorphella elachertiformis De Santis. Said to come near an Aus- tralian sp.— Mt, Mf. Eulophidae (211). ■V Diaulomyia calvaria De Santis. — Mt, Mf. Allied sp. in Australia. 312 C. SKOTTSBERG Formicidae [2SJ). Ponera trigoua Mayr var. opacior Forel. \. Amer., W. Ind., Chile, Argent. — Mt. The typical sfjecies in Ikazil. Probably spread with the traffic. Proiolepis obscura Mayr ssp. raga Forel. Melanesia. — Mt. The typical spe- cies Java, Australia, Hawaii, another variety N. Guinea and Melanesia. Tet}a))ionu))i guiiueiise (Fabricius). An African ant, now widely spread with the human traffic. — Mt. IcJniciDnonidae {2o6\ Rnicospilus purgatus Say. Temperate N. and S. America. — Mt, Mf. An al- most world-wide genus (Amer., luir., Afr., Austral., N. Zeal., Hawaii). -\- llcniiteles Bacckstroe})ii Roman. — Mt. An almost world-wide genus, less rich in the tropics. + //. luasafuerae Roman. — Mf. + flo/ocn'UDia {^) juaniaua Roman. — Mt. The genus is known from Europe. Mettlia (Paniscus) gerliugi Schrottky. Chile. — Mt. Range of genus very wide, including S. Amer., Falkl. Is., Rodriguez, Austral., N. Zeal. Stilpiius gagatcs Grav. var. Robi)iso7ii Roman. — Mf. The typical species in luirope, the genus also in N. America and Greenland. Tyiptognathus aequiciiicius Spin. Chile. — Mt. MyDiaridae [igo). Aiiagnis incaniatus Hal. Palaearctic. — Mf, undoubtedly introduced. + + CroimoDiytiiar fcrnandczi Oglobin. — Mt. ■V C. jniperfcctus Oglobin. — ]\It. + + Xcsopolynema caudatuui Oglobin. — Mt. P()ly}u))ia fiiscipes Hal. Palaearctic. — Mf, supposed to have been accidentally introduced. + V ScolopsoptcroJi kiiscJieli Oglobin. — Mt. Rhynchota. Heteroptera. A)itli()coyidac {2 1\ -^ Hucliivianiclla dcria Ik-rgroth.— Mt. Related to B. continua B. White from Madeira; other species reported from Tasmania and Hawaii. Lyctocoris catupcslris Vi\h\. Cosmopolitan, probably adventitious. — Mt. Lygacidac {2/, j6i). ' V Micrymcuus kuschcli Kormilev.— Mt. Most nearly related to Metagerra H. White from New Zealand (Kokmilfa). + J/. seclusus Ik-rgroth. — Mt. vXysius Bacckstrooni Hergroth.— Mt, Mf. An almost cosmopolitan genus witii numerous species in Xew Zealand and Melanesia, east to Samoa; greatest concentration in Hawaii. X. Pcrrcks/rorifii is closer to X. Ilutioni V>. White from Xew Zealand than to any American species (KoKMlLKv). + V Robnisowclioris tingitoidcs Kormilev. — Mt. P^orms a separate tribe. DERIVATION OF THE FLORA AND FAUNA 313 Miridae (50). + Derophthabna ferna7ideziana Carv. — Mt. A neotropical genus (Braz., Argent., Urug.). + +Kusche liana masatierrensis Carv. — Mt. Nabididae [21). Nabis (Reduviolus) pujictipennis Blanch. Chile. — Mt, Mf. A world-wide genus. Reduviidae (2g2, jio). Empicoris (Ploeariodes) rubrornaculaius (Blackb.). Almost cosmopolitan. — Mt, probably adventitious. + Metapterus additius Wygodz. — Mt. A wide-ranging genus (Amer., also Chile, S. Eur., N. Afr., W. Asia). ■vM. kuscheli Wygodz.— Mt. ^M. fnasatierrensis Wygodz. — Mt. Ploiaria chilensis (Phil.) Kuschel. — Mt, Mf. Almost cosmopolitan; also in Chile and probably adventitious in Juan Fernandez. Homoptera. Apkididae (communicated by Dr. Kuschel): 4 introduced species on garden plants. Cicadellidae (57). + +Evansiella kuscheli China. — Mt. Delphacidae (314). + Nesosydne sappho Fennah. — Mt. A genus known before from south and central Pacific islands including Hawaii, but never reported from America. -f-A^. minos Fennah. — Mt, Mf. + N. oreas Fennah. — Mt. + N. calypso Fennah. — Mt, Mf. + N. philoctetes Fennah.— Mt, Mf. + N. vulca7i Fennah. — Mt. ^Delphacodes kuscheli Fennah. — Mt. A widely distributed genus. ■vD. (Sogata) selkirki (Muir) Fennah. — Mt. Jassidae [21). -f + Alloproctus amandatus Bergroth. — Mf. Some zoogeographical statistics. Orthoptera. — Of the four species known 2 are endemic but of American affinity, one a Chilean species and one reported from Galapagos Is. and Easter I. A^^^r^/>/^r^.— Endemism strong, 4 species of 5, 2 forming an endemic genus, the fifth an American species. Relations presumably Andean. Lepidoptera.—0{ the 26 indigenous species 18 (69 %) are endemic, and there are four endemic genera. Of these, Apothetoeca and FernaTidocrambus are related to world-wide genera also represented in S. America, Hoplotarsia to an American 314 C. SKOTTSBERG genus; the systematic position of JuaJiia has not been stated. Together they include 6 species. Of tlie remaining 12, 7 belong to genera with a large to world- wide area including at least some part of America, and one belongs to an American genus. Kulia (3 sp.) and Lohopliora are essentially boreal. The 8 non-endemics are found in Chile or in some other part of S. America. The total absence of all indigenous Rhopalocera is remarkable. Diptera. — At present 157 named species belonging to 27 families have been reported. Nine or ten species at least have been introduced with the human traffic. 147 are thought to be indigenous and of these 94, 64%, are endemic. Con- sidering our insufficient knowledge of the dipterofauna of Chile, too much weight should not be laid on these figures, but even if quite a few of the insular en- demics will, in the future, be discovered on the mainland, I trust that a fair number will remain, sufficient to show the peculiar character of the fauna. Be- sides it can be foreseen that Dr. Kusc;ilEL's new material will bring to light some remarkable additions. The fauna is not a haphazard crowd of wind-drifted flies. It gets its stamp less from the few endemic genera — of 10 new genera proposed by Enderlein only 4 remain — than from the presence of six non-endemic, S. American or more wide-ranging genera with six or more species each, MolopJiilus with 19 (all end.), Scatella with 16 (all end.), Lii)wiiia with 13 (12 end.), Leptocera with 10 (3 end.), MycctopJiila with 7 (2 end.) and Podoiwmus with 6 (4 end.), together 71 species of which 56 (79%) are endemic. The Xeotropical-Chilean character of the fauna is obvious. This is what we ex[)ect (juite ai)art from what we may think about the history of the fauna, but the almost total absence of even a small austral-circumpolar or Pacific element is noteworthy; the only examples w^ould be Prosopantruni flavipes (austral- tricentric] and Inicellia jjitcDJicdia, said to be distributed over "Oceania". There are some striking cases of disjunction, suggesting bipolarity {IJoyella, Hydrobacniis, Podouoifius Kiejferi], but the distances will perhaps be lessened when the dis- tribution becomes better known. CoUoptoa. — I want to emphasize that of Dr. Klschel's collections ony 5 families have been worked out; it is to be regretted that no list of the Cur- culionids is available. On the other hand I believe that the beetles inhabiting the Chilean mainland are better known than the flies and some other insect groups so that the proportion between endemics and non-endemics will not undergo very great change in the future. The number of named s|)ccies hitherto reported from Juan Fernandez is 103, belonging to 19 families, [)erhaj)s little more than V'.i of the species found there. Eleven species are anthropochorous. Of the remaining 92 no less than 74 are endemic — 80 "o, only 20% having been found elsewhere. Future research will alter these figures, I suj)p()se, a number of island endemics will be stated to extend to Chile and vice versa, but on the other hand we have good reason to expect that practically all Curculionids collected but not yet described will j)rove to be endemic; of 22 indigenous species enumerated by AURIVILLIUS 21 were described as new. DERIVATION OF THE FLORA AND FAUNA 315 So far 49 indigenous genera are cited, of which lO are endemic; five of these are CurcuHonids. The non-endemic species are, with one exception — Xyletomerus piibescejis, a N. American species represented by 2 endemic varieties — also found in Chile. As yet no austral-bicentric or -tricentric species have been found. However, Pycnomerodes with 2 species in Juan Fernandez, i in New Zealand and none elsewhere, as well as Pyawmerus, Stroiigylopterus, Eleusis and Enneboeus suggest the existence of a small austral, possibly Antarctic element, even if the area in cases extends north of the Equator. Hymenoptera. — The 35 registered species, 5 of them adventitious, cannot represent but a minor portion of the fauna. Just as in all other insect groups endemism is strong, 23 species are endemic [']6.6%] and of the 26 genera 6. The affinities were, as a rule, not indicated by the authors; very likely they are, with some exceptions, with S. American forms. Haplogonatus is essentially south- ern, Prenolepis obscura is a southern, mainly Pacific ant. ApJielms jucu7idus and Stilpnus gagates are said to be boreal, but in these as in other similar cases the possibility of accidental introduction must be considered. Hemiptera. — The 21 indigenous species — there are 7 adventitious ones — are by no means a fair representation of the Rhynchota inhabiting the islands. Dr. Kuschel's collection contains twice as many species, more than half of them endemic; of the 21 named species 20 are restricted to Juan Fernandez. Where 3 of the 4 endemic genera have their relatives I cannot tell, but the fourth, Micry- menus, is most nearly allied to a genus in New Zealand. Buchaiianiella is quoted for Madeira, Tasmania and Hawaii, but not from America, Nysius, a world-wide genus, has a stronghold in New Zealand and in the Pacific, and the single is- land species stands nearer to a species endemic in New Zealand than to its American congeners. It seems likely that we have to do with an austral- antarctic element. Mollusca. The following synopsis is based on information supplied by Professor NlLS Odhner who put his unique knowledge of this group at my disposal. The additions and changes in his earlier list (i8g) are entirely due to him, and I thank him for invaluable assistance. Possibly the new material brought back by our 1954-55 survey will give additional taxonomic results, but they will not alter the zoogeographical position of the fauna. E7idodontidae . -f +A?nphidoxa helicophantoides Pfeifif. — Mt. The genus (only 2 species known) is related to Stephanoda. + A. marmorella Pfeiff. — Mt. + Charopa (Endodonta, s. lat.) involuta Odhner.— Mt. Numerous species, Polyn., N. Guin., Austral., N. Zeal. -f C. occulta Odhner. — Mt. -f C. skottsbergi Odhner. — Mf. 3l6 C. SKOTTSBERG ^Pioictuni conicion Odhner. — Mt. A genus of a rather small number of spe- cies, in various parts of the world. ^P. depressu))! Odhner. — Mt. ^- Radiodiscus inasafucrae ((Odhner) Pilsbry. — Mf. An American genus of few species, found in X. America (Arizona) and in S. America, south to Patagonia. ■^StepJuDwda iirctispira PfeitT. — Mt. About 30 species in western S. America (Chile, south to Fuegia). + .V. ceratoides Pfeiff.— Mt. + .V. quadrata b'erussac. — Mt. + 5. sclkirki \\. A. Smith.— Mt. -S. tcssellata Muehlf.— Mt, Mf. HilicJdae. Helix aspcrsa Muell. Cosmopol., introduced to Chile. — Mt. Li))iacida€. Agrioliiuax agrestis L. Very widespread, introduced to Chile. — Mt, Mf. Liuiax arhoruui Buch. -Chant. As the former. — Mt, Mf. Mi/ax gagatcs Draparn. As the former. — Mt, Mf. Sitcciiicidae (subfam. Succineinae). + Succiuea cuuiingi Reeve. — Mt. Succhiea is taken in its old sense; it has been split up, and the Juan Fernandez species belong to a group which must bear a diflerent name. It is reported from N. America, Panama, Galap. Is., St. Helena. S. Africa, Hawaii and Tahiti. + 6. fcr)ia7idi Reeve — Mt. .V. fragilis King. (syn. S. texta Odhner). Hawaii. — Mt. + .V. gayajia (D'Orbigny) (Odhner. — Mt. + .V. Diasafucrae Odhner. — Mf. + .S". pifigujs (Pfeiff.) Reeve. — Mf. + .S". seuiiglohosa Pfeiff.— Mt, SC. Tor)iatclli)iidae. ^ + I''ir?ia?idfzia buliniojdes Pfeiff. (inch consimilis Reeve). — Mt. + /'\ couifcra Reeve. — Mt. .' + /'". cyli)idrella Odhner. — Mt. Possibly identical with Torvatellina (F21asma- tina) tunita Anton, credited to Oi)aru I.* + /'". diapluDia King. — Mt. + /'\ expivisa Pilsbrw — Mt. + /'\ inorjiata Pilsbry. — Mt. ^ F. lo)iga Pilsbry. — Mt. + /'". pliilippiiUia Pilsbrw — Mt. + /'". splendid a Anton. — .Mt. f/'\ tryojii I'ilsbry. — Mt. + /'\ /('ils()?ii I'ilsJMy.— Mt. ' Rapa seems to be tlie name commonly used. DERIVATION OF THE FLORA AND FAUNA 317 + Tornatellina aperta Odhner. — Mt. A Pacific genus of over 50 species, ranging from E. Ind. and Japan over Micron., Polyn. (inch Hawaii), Melan. (Kermadec Is., N. Caled.) to Austral, and N. Zeal. T. bilamellata Anton. Recorded from Oparu I. + T. callosa Odhner. — Mt. + T. conica Anton. — Mt. + T. plicosa Odhner. — Mt. T. reclusiaia Petit. — Mf. According to Odhner probably identical with T. iiir- rita Anton. + T. trochiformis (Beck) Pfeiff. — Mt. T. trochlearis (Beck) Pfeiff. Oparu I. — Mt. + Tornatellinops minuta (Anton) Pilsbry et Cooke. — Mt. A Pacific genus of 22 species, reported from Japan, Philipp. Is., Polyn. and N. Zeal. Zonitidae. Hyalinia alliaria Miller. A widely distributed, anthropochorous species. — Mt. H. cellaria Miller. As the former.— Mt. Mf. Forty-six species are enumerated; of these 6 have been introduced through the human traffic. Of the remaining 40, 35 (87.5%) are supposed to be en- demic; 13 belong to the two endemic genera. The occurrence oi Succinea fragilis in Hawaii and Juan Fernandez and nowhere else in surprising, and 4 species are credited to Oparu (Rapa) Island, but the distribution is perhaps not too well known. The poverty of Masafuera, where only 5 species have been collected, 3 of them restricted to this island, is, I daresay, only apparent. Additional forms have been found later and still await study. On Santa Clara only empty shells of a Masatierran species were found. This islet seems entirely unfit for land-shells. Only two well-defined geographical groups are distinguished, to which a third of wider extension is added. I. American element. — 8 sp. Amphidoxa (2), Radiodiscus (i), Stephanoda (5). II. Pacific element. — ^23 sp. Charopa (3), Fernandezia (11), TornatelHna (8), Tornatellinops (i). III. Austral (or more widespread) element. — 9 sp. Punctum (2), Succinea (7, see above). Chapter IV. Continental and Oceanic islands. For a clear distinction between the two main kinds of islands WALLACE [2^8) is as a rule referred to as the leading authority. From a geographic-geological viewpoint an island, usually neovolcanic or coralline, which does not stand upon a continental shelf, is called oceanic. If situated on the shelf there is a strong 3l8 C. SKOTTSBERG j)ossibility that, at some period of its existence, it has formed part of the con- tinent. This is e.g. the case witli the Falkland Islands. A truly oceanic island lacks a continental basement of old, granitic or sedimentary rocks; at least, their presence has not been demonstrated. It is, on all sides, surrounded by deep water and a rise of perhaj:)s thousands of metres is required to bring it into contact with a continent. Mavr (/J^), however, argues that, from a biological viewpoint, every island, whether situated on a continental shelf or not, is oceanic which has received its entire living world across the open ocean — consequently it must be shown that every kind of organism present on the island has or once had the faculty of migrating across the sea and establishing itself, either the species ac- tually found or their ancestors. It goes without saying that the answer to the question "continental or oceanic?" should in the first place be looked for in the history of the oceans. With regard to the Pacific our knowledge of its origin and history is incomplete, and even if modern ()ceanograj)hical research has supplied a wealth of information on the hydrography, the nature of the sediments and so on, large parts of the southern Pacific are little known and soundings so few that we cannot form but a very general idea of the bathymetrical conditions and the configuration of the bottom. The northern half is of course far better known. As it is, we must admit that little or nothing has come to light that is opposed to the theory of the perma- nence of the Pacific Ocean. It is, with few exceptions, from the biologists' camp that the theory has been attacked, particularly by phytogeographers; the majority of zoologists seem to accept the conclusions arrived at by physiographers and geologists. It is easy to understand, however, that many biogeographers, struck by the [)er{)lexing disjunctions in the distribution of plants and animals, started to build bridges across wide expanses of sea, in cases with a generosity that led to absurdities. I have no reason to enter into details, our problem concerns Ant- arctica, southern South America and Juan P>rnandez, but even so it seems worth while to (juote a number of modern scientists, mainly geographers and geologists, who have expressed their opinion on the nature and history of the Pacific Ocean. Geotectonics of the Pacific Basin. H.MI.F.V Willis (2S4) thinks that a suboceanic pressure works against the con- tinents surrounding the Pacific, resulting in an expansion of the suboceanic mass and a deepening of the basin which, in its turn, has a displacing effect on the continental margin. lie summarizes p. 367-368: The consideration of the general facts of the geotectonics of the Pacific basin thus leads us to regard the great ocean as a dynamic realm, within which the peculiar char- acteristics of its rocks have facilitated the internal forces of the earth. The effects have i)een as a whole to deepen the basin in conse([uence of the expansion of the under- lying ro( ks. The expansion has in turn crushed the continental margins and raised the great cordilleras. Geologic studies of the mountain ranges have demonstrated that the actual orogenic j)criod began in the Jurassic or possibly somewhat earlier in the Mesozoic. Of the earlier periods we know but little, but the fragmentary records indicate that })eriods of orogenic activity alternated with those of (piiescence. DERIVATION OF THE FLORA AND FAUNA 3I9 Willis did not question the permanence of the basin; what interests us here more particularly is his belief in the instability of the marginal regions. In H. E. Gregory's view the Pacific basin inside the deep troughs is an old sink and it follows that all the islands within this sink are truly oceanic. Con- sidering geological evidence alone there has been no significant change in the position of Polynesian land masses since Pleistocene, most likely since early Ter- tiary time: "There is no geologic evidence that any Polynesian island stood in Jurassic or Cretaceous seas" (775. 1673). Still he thought that due regard should be taken to objections raised by other branches of science, and he did not extend the unaltered permanent basin outside the deep troughs. When, a little later, another prominent geologist, J. W. Gregory (116), expressed a different opinion, this attracted a good deal of attention. The Pacific had, he says, been claimed to have existed in its present shape and size throughout geological time, a hypo- thesis almost universally adopted by geophysicists and geologists, but from a biological viewpoint this theory did not satisfy. GREGORY was no believer in large-scale transmarine migration of either plants or animals and consequently inclined to consider the arguments put forth by the opponents to the permanence theory. He counted with a number of Pacific seas separated by stretches of land, and he looked upon the region where atolls serve as proofs of subsidence, a sub- sidence which gradually enlarged the basin until it reached its present size, as originally continental. Andrews [6) who was a firm believer in successful transoceanic migrations of all kinds of organisms and knew more about geology than most biogeogra- phers, shared H. E. Gregory's opinion: islands situated within the area bounded by the ocean deeps such as Hawaii, Marquesas, Society and Juan Fernandez, all differing in their geological structure from the continents, are oceanic. Parts of Andrews' interesting paper deserve to be quoted here. An examination of the continental blocks proper and the great western island arcs suggests that they have had similar histories, whatever great differences may other- wise exist between them. Japan, Eastern Australia, New Zealand, New Guinea, Fiji . . . may be taken as examples. Each has a foundation of ancient folded and metamor- phosed sediments, such as conglomerates, grits, quartzites, sandstones, slates, shales, and limestones, and each of these foundations has been subjected to marked plutonic intrusions of granitoid nature. Upon this foundation have been accumulated sediments similar to those mentioned above, together with lavas not only of basic but also of acid types. These, in turn, have been folded, overthrust, and invaded by plutonic rocks. This generalization is true even though as yet no consensus exists concerning the age, or ages, of the folded sediments and plutonic intrusives of the foundadon rocks. It would appear, however, that the foundation rocks of the island arcs which occur mar- ginally to the continents of Asia and Australia are not as old as the earliest members of the continental nuclei. This leads to the consideration of island arcs situated more centrally within the Pacific Ocean. P'or this purpose, these may be considered as in- cluding all the Pacific islands lying oceanward of the great island arcs mentioned above. The principal examples include the Hawaiian Islands, the Marquesas, Juan Fernandez, Easter Island, the Society Islands, the Cook Group, the Line Group, Micronesia, Samoa, and the Ladrone, Caroline, and Pellew groups; Tonga and the Hebrides occupy a peculiar position, mentioned below (p. 202—3). 320 SKOTTSBERG Field observations show that those islands have had histories which present marked diftercnces from those of the continents and their marginal island arcs. 'J'hus, they appear to be comjiosed almost entirely of volcanic material, mainly basic, together with "coral reef" formations, whereas granitoid intriisives and acid lavas are lacking, to- gether with the sediments invarial)ly associated with "continental" areas. And not only is this so, but the volcanic ejectamenta of these inner groups do not appear to con- tain fragments of granitoids and sediments such as might be expected from volcanos discharging through a foundation of rock formations such as compose the continents (P- 203). Another interesting feature is the ])eculiar topograj)hy of that portion of the Pacific floor which sej)arates the great island arcs and lands of "continental'' character from the more central grouj)s. Thus, on the American side of the Pacific, the "continental" lands are separated from the groups — such as Hawaii, the Marcjuesas, the Society Is- lands, and Juan I'crnandez — by a series of deep discontinuous ocean trenches, prac- tically collinear (j). 203). The question is whether these trenches are of quite the same nature and date from the same j^eriod as the deeper trenches arranged oceanward from the great western island arcs. It is unfortunate for the advocates of a "continental" origin of the Juan Fernandez flora that these islands are situated on the wrong side of the trench. However, the Galapagos Islands occupy a similar position, and still they have been claimed, on good grounds, once to have been united with Central America. In this connection another quotation from ANDREWS with regard to the New Hebrides and Tonga is of interest. Island groups which are difficult to place exactly in this scheme are the New Hebrides, Tonga, and possibly the Pellews and the Ladrones. A profound deep lies between New Caledonia and the New Hebrides, and this is suggestive of a noncon- tinental origin of the group. On the other hand, the occurrence of mineral deposits such as copj)er, iron, and nickel, of large kauri, fig, myrtaceous, and other trees, and of animals such as lizards, turtles, ducks, pigeons, and parrots, suggests that they may well have formed, at some earlier time, j)ortions of a continental margin which later be< amc involved in a j)Owerful movement within the marginal Pacific, resulting in the gradual submergence of these outer ))ortions, the present Hebrides, Tonga, and so on, being built upon su( h sinking area. This certainly is suggested for the New Hebrides and for Tonga, while the Pellew and Ladrone islands also have had complex histories, which would well repay close attention, in their structural, petrological and biological aspects (p. 203). An additional j)C)int of interest is the association of great ocean deeps with youthful volcanic zones, and inasnuu h as the trough and crest of an earth undulation are parts of the same strut ture, it is a legitimate inference that the great Pacific deeps or trenches are relatively youthful structures (p. 203). Andkkw.s summarized his idea of the Pacific basin in a number of points, which, with very slight verbal alterations, form the introduction to his paper on the origin of the Pacific insular floras (7. 613-14): 1. The continents bordering the Pacific have been larger, at various times, than they are at |)resent. 2. The great bordering island arcs of the Pacific — such as the Aleutians, Japan, the IMiilippines. the Netherlands Fast Indies, New (Guinea, Fiji, New Caledonia, and New Zealand— have been connected directly with the continental lands. Certain of these DERIVATION OF THE FLORA AND FAUNA 321 island groups — such as Fiji, New Caledonia, and New Zealand — appear to have been isolated at much earlier periods than others, such as New Guinea, the East Indies, and Japan. 3. The ancient borderlands of the continents have, in part at least, suffered un- dulatory submergence. Compensatory forms are the deep ocean trenches, on the one hand, and the mountain ranges of the continents, on the other. 4. The Pacific is a relatively deep and unstable area, whereas the Atlantic — with the exception of the broad intersected belt of activity directed toward the equator — has, on the whole, been relatively stable since the Palaeozoic. 5. The western area of the Pacific appears to possess a more complex structure than the eastern, owing to the earth's rotation, the width and weakness of the Pacific base, and the resistance opposed to this activity by the stable continental masses of Eurasia and the Australian— Sahul area. 6. The islands of the Pacific lying within the area bounded by the ocean trenches have not had continental histories, nor do they appear ever to have had actual and direct land connections with the continents. 7. The New Hebrides and Tongan Islands, and possibly also the Pellew and Mari- anne groups, appear to partake in some measure of the nature both of "continental" and "oceanic" islands. We shall have occasion to return to Andrews' opinion on the origin and his- tory of Pacific floras. Here it seems convenient to draw attention to the numerous submarine cones called "guyots" recently discovered in the sea between Hawaii and the Marianas and discussed by Hess (1J2). About 160 flat-topped peaks, presumably truncated volcanic islands, rise from 9000 to 150CO feet above the deap-sea bottom. In most cases their flattened summit was sounded in about 800 fathoms. Hess' working theory is that they were formed on land, sunk to their present level and levelled by sea action — this would mean that they stood with their summit at sea level long enough to be exposed to wave action. Thus, drowned reefs could be expected, but no such are reported: the guyots are, Hess thinks, very ancient structures dating from a "proterozoic episode of vul- canism", they are of pre-Cambrian age and consequently do not lend themselves to biological speculations. There may, however, have existed later islands, both Paleozoic and Mesozoic, but all have disappeared beneath the surface of the ocean, either built up again by reef-building organisms or sunk to a depth where these cannot live. The high volcanic islands are very young, perhaps not even Tertiary but Pleistocene or recent, because the rocks could rarely be proved to be of Ter- tiary age. We shall see by and by that the endemic insular floras and faunas cannot be anything like recent. "Oceanic islands", Hess continues, "are and have always been slowly sinking relative to sea level" as a consequence of the accumulation of bottom sediments causing the water level to rise. The red clay increases i cm in loooo years, the globigerina ooze the same amount in 5000 years. Thus millions of years are needed to account for even a very moderate submergence. Besides, raised shore-lines are, in many cases, unmistakable proofs of local emergence. The melting of the great inland ices ought to have had a greater influence. Finally, let us listen to one of the foremost authorities on geophysics, GUTEN- BERG (i2j). He finds (p. 7) that there is 21 -557857 The Nat. Hist, of Juan Fernandez and Easter Isl. Vol. I 02 2 t'. SKOTTSBERG growing accumulation of evidence that the Pacific basin shows unique features which are not duplicated in any other oceanic or continental area of the earth. 'J'here is no feature on the surface of the earth which compares in dimensions and importance with the Marshall line, within which the younger eruptive rocks are basaltic rather than andesistic. This discontinuity in the material of the crustal layers is called here the boundarv of the Pacitic l>asin. This is in conforniit}- with what I have quoted above from other sources. However, certain areas of the I'acific Ocean (near its borders, for example), at least part of the region between South America and the Easter Island rise, or between the Marianas and the Asiatic continent, show indications of continental layers. For the latter, petro- graphical and geophysical evidence agree. As seen on the niaj), (il TKXHKRC] goes a good way beyond the Juan Fer- nandez-San Anibrosio rise, but the blaster Island shield which, excepting the vi- cinity of this island, is covered by very deep water, belongs to the wide basaltic centre, where continental layers are lacking — in contrast to the Atlantic where "granitic la\-ers of the continents continue far out under the bottom . . . probably at least some continental rocks underlie its bottom throughout its area". Turning to tiie speculations of biologists I shall quote some representatives from the two op|)osite camps. Arldt [6') did not draw his conclusions merely from facts of distribution but compiled a wealth of geological, palaeontological, bathymetrical dates and so on, and constructeci a series of maps illustrating the distribution of land and sea through earth's history. A Cretaceous Oceania united South America with Australia + New Zealand, it disappeared during F^ogene and left the west coast of South America in the same position as to-day. F>om what he sa)-s about Juan Fernandez it appears that he regarded these islands as con- tinental 'see below p. 376), while still admitting the possibility of oversea migra- tion from the coast. Camphkll was for a long time a supporter of the land-bridge theorv. He regarded the Hawaiian Islands as formerly much larger and more closely connected with land masses to the southwest, having become isolated during early Tertiary time coincident with the ui)lift of the great Cordilleras (ples were crumpled up on the floor. The transfor- DERIVATION OF THE FLORA AND FAUNA 327 mation of the geanticline structures into rows of islands is explained by Du ToiT as follows; their crest could become deepened by crustal tension and broken into segments to form an island chain before vanishing . . . limbs were intermittently built up and destroyed during the Cretaceous-Tertiary through stretching in the direction of their length while they were still compressed by forces at right angles thereto (p. 293). Trying to apply these ideas to South America, which according to Du ToiT as well as WEGENER was pressing into the Pacific basin, the Juan Fernandez- Desventuradas-Merriam ridge could be compared to an advance-fold. But it is not convex to the Pacific, nor fronted by a fossa — this is situated on the wrong side and may well stand in causal relation to the upheaval of the Cordillera. To think that the submarine ridge emerging in the Juan Fernandez and Desventu- radas Islands is the easternmost advance-wave from a western borderland seems too phantastic. Du Toit's idea of the geological character of the ocean floor differs from Wegener's. Seismographic records, he says, scarcely bear out that the Pacific floor must be composed of basalt — the records could readily accord with a granitic layer up to about 10 km thick (p. 212). He was no believer in a more or less unlimited oversea migration of plants and animals, nor in land-bridges, and he critizises J. W. GREGORY and the bridge-constructing biogeographers: they are wrong, and the displacement hypothesis interprets otherwise. But when he speaks of the extensive "march into the ocean of crustal waves, thereby leaving their parent continents far in the rear" and of the "rhythmic intrusion, culminat- ing in the three migrations of the Cretaceo-Eocene, mid-Tertiary and late Ter- tiary" (p. 214), these advance-folds, when crumpled up from the ocean floor, were absolutely devoid of every sign of terrestrial life and without a trace left of the sial cover. I fail to see that they can solve any biogeographical problems — we have to fall back on oversea dispersal. Wegener's festoons were at least split off from the borders of a continent and left behind with their fauna and flora. With regard to Juan Fernandez we shall perhaps be able to find a less adventurous explanation of its history. Two years after the appearance of his book, Du ToiT summarized his theo- ries in a paper which I think it is worth while to quote [82. 75-76). The base- ment of the Melanesian islands is, he says, for the most part continental; the ocean floor consists of a relatively thin structure of sial underlain by sima, but this does not allow us to regard the sial as continental, because it may be a product of magmatic differentiation from the sima. He points to the parallelism between the great Tertiary folding-zones, most evident along the west coast of the Americas, and the trend of the coast line, and he thinks that the "compres- sive phases" were contemporaneous all around from New Zealand across Antarc- tica to South America. Coming back on the advance-folds he remarks that some of Gregory's hypothetical bridges or land-masses could well have been of this nature. The procedure is illustrated by a map showing the pressure direction and the formation of island arcs — except on the American side, where the sea is a blank. 328 C. SKOTTSBERG Xumerous biologists have found a solution of all or most of their difficulties to explain the present distribution of the organic world in the theory of conti- nental drift, combined with large-scale pole-wanderings. In view of this the opi- nions expressed by modern geologists and physiographers cannot be passed in silence. A symi)osium, arranged in 1950 (dj) offers an opportunity to get ac- quainted with their attitude. J. II. T. L'MlKiKOXK, TJic case for the cvust-substratiiui theory, pp. 67-71. Si Kss' terms sial and sima were petrographic. WEGENER attributed different physical properties to these types of rock; sial should be rigid but elastic, sima viscous. These statements lack foundation. Sima (basalt) has a higher melting- point, a})[)roximately I300°C, sial (granites) approximately 700°C. The crystal- line crust of the sima layer is at least as strong as the continental sial. Still the sial blocks were supposed to advance through the sima. Umbgrove concludes that continental drift is impossible at present. This granted, was it perhaps pos- sible in bygone times.' The answer is fetched from the Atlantic and Indian oceans with the intervening African continent; if the floor of the oceans originated as thought Wegener, the processes must have taken place during early Precambrian. U.Ml'.fiKovE asks if not the thick blanket of sediments would have been squeezed and [)iled up in front when America ploughed westward. He calculates that, considering the size of the westward drift, a plateau 200 km wide at sea level would have been formed in front; instead, "the continental slope is one of the steepest in the world and is fronted by deep-sea troughs" — here we have, how- ever, to consider the late upheaval of the Andes. And if, as Wegener's theory recjuires, the Atlantic originated in comparatively recent times, how are we to explain the enormous thickness of its bottom sediments, according to Hans Pet- ri'.KSsoN a maximum of loooo feet, "representing a time-span of 300 to 400 million years", which would bring us back to the Palaeozoic. If continental drift e\er occurred, UMliGROVE asserts, it took place some 3000 million years ago and C()nsc(iuently loses every shade of interest to the biogeographer. HakoM) Ji-.iikess, Mechanical aspects of continental drift and alteriiatii'e tJieories. ji;illrnandez, etc.) and Antarctica, Australia and New Zealand, vertical move- ments seem to otTer a less distant possibility. In a most cleverly written chapter Gooi) [log. 344-360) discussed land- bridge versus continental drift. If we cannot, he says, accept the former, nor put our trust in disj^ersal, a changing position of the continents is the only way out of the difficulties. This may be true, but we have seen that this theory does not help us to solve the problem concerning the oceanic islands. GORDON (//?) |)()ints out that the occurrence of a small but important subantarctic element in the Pacific, reaching north to the Hawaiian Islands where it is better displayed than in anv of the Polynesian or Melanesian groups, makes it impossible to deny both land bridges and the efficiency of transoceanic dispersal without providing for nngration with the help of shifting continents, (iooi) has run into an impissc f)vcr the Pacific islands like Hawaii. He has rejected the land- bridge hvpothcsis in fivour of ( outineutal drift.... Pnit he excludes continental drift so tar as the ishmds arc ( oik rrned, for he accepts them as truly oceanic, not continental frag- ments. \'et he will not ;i! i -pt oversi-as migration. Well, I can't see what explanation remains, if all the-e tlinr ;>!.■ e\( hided, but the plants are there (p. 148). W ri.l I {j- considerable and call for a modification of the theory. DERIVATION OF THE FLORA AND FAUNA 33I Chapter VI. Transoceanic migration. In some of my earlier papers I touclied upon the great problem of long- distance dispersal and the supposed efficiency of the transporting agents; see for instance 2ji. 20-30, where, however, only the flora was concerned. This time also the faunas are, to some extent at least, considered, and I shall quote a num- ber of authors, old and modern, who have expressed their opinion for or against overseas transport as the only possible means by which the isolated islands of the Pacific have received their indigenous flora and fauna. Advocates of large-scale overseas migration. Among earlier authors Engler [8^) and Grisebacii exercised great influence on their contemporaries. They divided the world into flora domains, regions and districts, characterized by a combination of certain important elements and by a greater or lesser degree of endemism, but still they never doubted the facility with which plants travelled across the oceans; progressive endemism was the unavoid- able corollary, but relict endemism was recognized as important. The attitude of this school is adequately expressed by GriSEBACH (j^j. 469): So merkwiirdig es auch sein mag, dass sogar einzelne Holzgewachse sich hier iiber das Stille Meer verbreitet haben, so ist ihre Wanderung doch aus der antarktischen Meeresstromung, den herrschenden Westwinden, oder durch Mithilfe der Seevogel, vielleicht auch durch alte Verkehrswege wohl hinlanglich zu erklaren, ohne dass die Annahme von Landverbindungen in der Vorwelt gerechtfertigt ware, die durch keine geologische Tatsache gestiitzt wird. By some the case of Krakatau was quoted as a proof that plants and animals are able to travel across water barriers; it is mentioned by Hayek [304), but with reservation : Wie die Besiedeliing einer Insel erfolgt, haben die oben angefiihrten Beobachtungen bei der Wiederbesiedelung des Krakatau gelehrt . . . Aber die Entfernung des Krakatau von den nachst gelegenen Inseln ist keine allzugrosse, sie betragt nur etwa 18—40 Km, also Entfernungen die auch durch die Flugtiere nicht allzuschwer iiberbriickt werden konnen (p. 251). When we have to deal with islands, separated by thousands of miles from all continents, the difficulties are of much greater magnitude, and he continues: Und doch miissen wir annehmen, dass auch die weit entfernt gelegenen Inseln ihre Pflanzendecke von den zunachst gelegenen Festlandern (und Inseln) erhalten haben, wenn auch vor undenklichen Zeiten und ganz allmahlich. Dafur spricht auch der Um- stand, dass die Flora dieser Inseln keineswegs von der ubrigen Flora der Erde grund- verschieden ist, sondern denselben Pflanzenfamilien angehort wie diese, demnach von derselben abstammen muss. The only exception known (at that time) was Juan Fernandez (Lactoridaceae). But he admits that there are grave difficulties: 332 C. SKOTTSBERG Selbst bei Inseln, die erst in relativ junger Zeit vom Festlande abgetrennt worden sind, ist ein weitcrer Aiistaiisch der Florenelemente zum mindesten wesentlich erschwert iind eine weitere /uwandcriing von Elementcn der Festlandsflora wenig wahrscheinlich. If tliis be true, lunv was immigration over tbousands of miles ever possible? Several writers who have {)aid special attention to Pacific problems occupy, more or less do^^matically, the same standpoint as Havek. Setciiell, with whom I had the j)ri\ile<;e to discuss this subject on various occasions, was already quoted {). 271 ; I shall add here what he says, in the same paper, about migration (2ig. 300). He found that 1 was "too narrow" in my allowances for migration possibilities; he belie\ed in "migration over very considerable breadth of barrier, whether of sea or land", and absence was not a result of failure to migrate successfully, but could be explained by obstacles to establishment. He regarded the oceanic islands as Tertiar)', but in his summary pp. 307-309 admitted the possibility of their being considerabl}' older, late Mesozoic or early to middle Tertiary, which would give time for extensive progressive evolution of endemic taxa and for the dying- out of their continental ancestors; or they had developed in other directions, making the relationships difficult or impossible to recognize. Geologists, however, refuse to give even the Hawaiian Islands a greater age than late Tertiary or even Pleistocene. It goes without saying that travel facilities are difi'erent in different cases; spore-plants are supposed to spread more freely than seed-plants, but even these are supposed to be quite capable. Thus Stebbins (ji'p. 537): '{"he seeds of j)lants may occasionally be transported over many hundreds of miles of ocean and may establish themselves on Oceanic islands like Hawaii, Juan Fernandez, St. Helena and the Canary Islands. r'l.oKiN has, he writes, shown that conifers of the south hemisphere have migrated freely from Australasia to South Ainerica and vice versa, w-hereas mammals are unable to pass and are absent from oceanic islands — but is it not customary to place them on a j)ar.' S'lEHl^.lxs' Antarctic connection does not include land-bridges, for "it existed for plants, but not for vertebrates" (but what about birds?). He looks for assistance in lost islands between Antarctica and New Zealand; on the o{)|)osite side the width of oj)en water is not so great, and seeds can still be carried from South America to Antarctica without much difficulty. As mentioned before, no botanist has greater confidence in long-distance dis- persal than h'()Siii;K(; : . . . transo(eani<; migration across at least 2500 miles without stej^ping-stones is not only a possibility hut a relatively common occurrence (99.867). Im)SI!I:i<(.s subject was the American element in the Hawaiian flora, but in order to exj)lain the j)resence f)f the dominant Australasian element we must count with still gieater distances. AxKl.RoD (7^) quoting F'o.SBERG takes a modified position. In case ot distances not exceeding some 200 or 300 miles there are no difficulties, "a coni|)]ete flora can transgress such a barrier without the loss of any significant floristic units '. A greater distance results in "waif assemblages", but many will find it impossible to regard e.g. the Hawaiian flora as a haphazard accumulation DERIVATION OF THE FLORA AND FAUNA 333 of waifs and their descendants. What AxELROD says about migration probabiHties during different geological epochs is of greater interest. Since plants are controlled largely by climate, and since climate has been changing during geologic time, it follows that plants comprising different communities have had different possibilities at different times . . . probabilities for long-distance migration were much higher for tropical plants in the Eocene than they are to-day. Temperate forest species had a m.uch higher probability from late Cretaceous to middle Tertiary, it is low now. Steppe plants had a higher probability during Pliocene than now. Desert spe- cies have a higher probability to-day than at any time before. His conclusions are drawn from the size and area of populations shifting with the extension of climatic regions. For my own part I have expressed my opinion on overseas migration in the Pacific on various occasions [2JI, 318, 248) and I am not going to repeat the dis- cussion here. My general conclusion was that the effect of transoceanic migration has been largely overestimated. GUPPY (121), who allowed birds, winds and currents to stock all oceanic islands with plants, arrived at the conclusion that this traffic was a thing of the past and that migration had practically ceased altogether. I expressed my doubts that it had ever been effective, in any case with regard to seed-plants. I have already remarked that Setctiell laid stress upon what he called the CEB (climatic-edaphic-biotic) factor complex. The main difficulty for the vagabond plants was not to cover the distance, be it ever so great, but to become a successful member of a community already established in the place where it happened to alight, and this difficulty increased as time went by; most surfaces of the earth, he says, are already stocked with closed vegetation, making it impossible for new arrivals to gain a foothold {2ig. 300). His ideas are clearly expressed in 218 (\). 874). As the islands have become more and more completely stocked each with its quota of plants and animals and have undergone various vicissitudes, particularly of elevation, erosion, etc. its hospitality to migrating germules necessarily has become less and less, the Biotic factor has become more complex and the Edaphic factor has also suffered change. In my view the result could just as well be the opposite, for these "vicis- situdes", emergence, erosion, volcanic activity and so forth create new soil, a more varied topography, a multitude of different habitats, all of which ought to give newcomers increased opportunities to get established. With Setchell, Andrews underlines the importance of CEB; genera expected to occur in Hawaii but absent "were not amenable to germination and survival after transport". Long before Setciiell, J. D. Hooker, Wallace and others had paid attention to the obstacles for the successful establishment of newcomers, WALLACE believed that St. Helena had become stocked with plants during early Tertiary time; later there was no room left, and the flora had changed so completely that no plant was recognized as an insular form of a continental species. We meet with Setchell's line of thinking in a recent paper by W. B. Taylor {26J. 572). In recent volcanic islands are many unstocked habitats to begin with, 334 ^- SKOTTSBERG but each new species would mean competition, and the entry of an additional species would be very difhcult and consequently of rare occurrence. This may be so, but tiiere are nian\- communities of a more open character than the forest, and even a closed forest is not like a tin packed with cigarettes; young secondary forest associations, steppes and savannas ought to offer good housing grounds to an intruder, suj)posing that he likes climate and soil. Experience shows that numerous aggressive plants brought by man, not only herbs but also trees and shrubs, tind suitable living conditions even in undisturbed natural communities. In Juan Fernandez I have had occasion to follow the invasion by Aristotelia maqui (chilensis) and to witness the fabulous ease with which it crowds out the native vegetation, and to observe I'i^Jii Moliiiac springing up on the ridges where the j)lant cover was o|)en and, from there, to enter the dense native brushwood. And they are only two of the many successful weeds, a third one, equally dan- gerous but of quite recent introduction, is Rubus nluiifolius. All three have fleshy fruits and are eaten by man and birds and propagate themselves rapidly. They are conunon on the opposite mainland (where, of course, the brambleberry was introduced from luirope), but man, no bird, carried them across to the islands. Similar examples are, I presume, oftered by almost all oceanic islands. I just hap- j)ene(l to read a book on Cape Verde Islands, where a naturalist tells us about La)ita)ia caniara sj^reading like wildfire and menacing the little there is left of natural savanna and steppe [326). Most zoologists favour the theory of long-distance dispersal. Mammals are, as a rule also bats, flying foxes and the like, excluded, but of birds some are able to cover very large distances, winged insects are carried ofl" to distant places where they never wanted to go, and so forth. I shall quote some zoogeographers who, with reference to the Pacific, have expressed their opinion on overseas transport. Pl-.KKINS, in his introduction to Fauna Haw-aiiensis (j^j. XLVl), wrote: All the islands being volcanic and having been built up from a great depth of ocean at various j)eriods, their entire fauna naturally originated from immigrants derived from other lands. These immigrants must have arrived either by flight, like the birds, or in drift Hke the flightless insects and jirobably the land Mollusca. Dritting logs were often regarded as an important kind of conveyance, but they come from North America and what they bring of animals, PERKINS says, would ser\e no |)urpose because it is unlikely that the passengers would become acclim- atized in ilawaii. The fauna must have come from the warmer parts of America, from Australia, Polynesia etc. "at rare intervals from the F^ocene until now". If we have to belie\e the geologists, no Ilawaii existed in the F^ocene — and how did those, who arrive "now", manage to become endemic genera and species? — non-endemic (lowering plants not brought by man are few. In some instances (ii LICK [Tig] admits the possibility of land connections, but Ilawaii, Juan hernandez, (ialaj)agos, St. Helena, etc. etc., have always been isolated. 'I he (juestion whether their fauna shows that "the ancestors possessed an almost inconceivable capacity for passing uninjured over vast stretches of ocean" is answered in the affirmative. It is significant that the Galapagos archipelago DERIVATION OF THE FLORA AND FAUNA 335 "was successfully reached by a giant tortoise"; already Wallace entertained the same idea — a strictly terrestrial animal crossing the ocean. Some biogeographers prefer one dispersal agent, some another, most have confidence in all, but it goes without saying that different types of plants and animals have availed themselves of different kinds of transport, I shall quote GULICK (iig.414) first. It is possible to go far toward a first diagnosis of the degree of a land's insularity by noting how exclusively it is peopled by types with a known capacity for colonizing across vast expanses of ocean. Our summary up to this point reveals very nearly which these forms may be. Quite a majority of them, both plants and animals, show characters that har- monize with wind-storm transportation. A respectable majority of the larger-seeded palms and some tough-lived earth-inhabiting invertebrates, suggest transportation by water or on drift-wood. Such seeds and invertebrate eggs as can withstand the digestive tracts of a bird, have a very substantial travelling radius by that means, easy 500 miles in the routine seasonal migrations, and possibly stretching in the extremest cases to almost transoceanic distances. . . . Dioecious plants and separate-sexed animals are statistically at a disadvan- tage, as compared to the reversed condition, because of their poorer chance of achieving fertilization. The ability to take a journey in a gravid condition helps the chances greatly. "Types with a known capacity for colonizing" — GULlCK proceeds from what should be proved, for their occurrence on isolated islands is in itself no proof of oceanity. Under his angle the great number of dioecious endemic phanerogams in Hawaii ought to have surprised him. It almost seems as if he believed that entire specimens with roots and all managed to reach a distant island and get established; surely, if only a male or an unfertilized female arrived) all was in vain until a mate of the opposite sex turned up; a pregnant female would of course do better (bye the bye, WALLACE tells a story of a pregnant boa constrictor arriving on a West Indian island with drift-wood and in good condition). I guess we can leave these chances aside, for plants spread by means of seeds, and whether wind-blown, epizoic or endozoic (provided they do not, as many assert, discharge their droppings soon after the departure), there is every chance that more than one seed of the same kind is brought; a single many-seeded berry is enough, and a bird picking drupes fills his stomach. A seed portion of a dioecious species gives, under ordinary conditions, 50 % of each sex. In the Hawaiian flora we find, GULiCK says (p. 418), "a preponderance of plants spread by wind-carried spores and minute seeds"; species with drupes and berries are, however, numerous. As an example of a presumably definite case of bird rather than wind carriage he mentions the Hawaiian species of Vaccinium, which he derives from North America. Their presence is most in- teresting, "as the distances involved must be very close to the extreme physio- logical maximum that land birds can traverse, and still carry fruit seeds in their droppings". To me it appears as a bad case of constipation. Besides, the Ha- waiian Vaccinia are not related to North] American groups but belong to a special section. Mayr, an extreme "oceanist" who refuses to admit land connections for either Fiji or New Caledonia, in his paper on the Pacific bird fauna [lyg] includes 336 C. SKOTTSBERG a general survey of the dispersal chances for other animals and also for plants. The special instance cited (p. 197) is not very convincing. liirds are excellent rivers and thus cai)able of ra])id and active spreading . . . capable of crossing considerable stretches of open sea to settle in new territories. There is abun- dant evidence of this, such as the resettlement of Krakatau Island, the recent arrival of Australian birds in New Zealand, and the colonization of untjuestionably oceanic islands. Not even the arrival of Australian birds in New Zealand brings conviction; the colonization of "unquestionably oceanic islands" certainly does, if we can prove it. Mavr continues p. 198: 'i'he possil)ility of transport by floats or in logs is not to be underestimated. Many tro])ical currents have a sj)ced of at least 2 knots, that is, about 50 miles a day, or 1000 miles in three weeks. It is probably not a great task for a wood-boring insect to survive 3 weeks in a drifting log. Air currents are, however, of uncomparably greater importance than sea currents. Even slight winds are of great influence on the distribution of floating and flying animals, as recent investigations have shown. It is astonishing how rich the "aerial plankton" is, even up to altitudes of 1000 meters and more. Normal winds would, of course, not account for the spreading of molluscs, flightless insects, and other small in- vertebrates. However, most of the islands, with which we are concerned, are situated within the zone of tropical hurricanes, the lifting force of which is quite extraordinary. . . . The fact that there are small molluscs and flightless insects on such typical oceanic islands as Kaster Island, juan Fernandez and Saint Helena is almost unassailable proof that such a method of disj:)ersal is a reality. Tropical hurricanes carry for hundreds and even thousands of miles. . . . I'he result of the recent surveys in the Hawaiian Islands, the Mar([uesas, and on Samoa indicate that there are indeed very few animals that cannot be transported across considerable stretches of the sea by winds, waves, other animals or man. On p. 201 Mavr adds some general remarks: The means of dispersal of most plants and animals are much more extensive than was formerly realized, and even rather irregular distributions can be explained without the hclj) of land bridges. Dispersal across the sea is, of course, most obvious for birds, and ornithologists were among the first who accepted the ideas of the permanency of continents and oceans. Most entomologists are also beginning to realize that they can solve most of their distribution difficulties without land bridges. The conchologist, hf)wcver, ])Ostu]ates even today continental connections between all or nearly all the is- hmds where land shells exist. As we shall see below ([). 350) Mayr declared himself unable to explain how land shells arc (lisj)crsecl. It seems that, also with regard to the birds, Mavr contradicts himself, for in the same paper (j). Kj.S) he asserts that most birds, particularly on tropical islands, |)recisely the islands we are discussing, are extraordinarily sedentary, and as an examj)le he mentions that of 265 species known from that part of New Guinea which is ()j)posite New Britain, a distance of 45 miles, only 80 occur on New Britain, and the situation in Western Papuan islands is even more conspicuous; he mentions two islands only 2 miles apart, with rather different fauna. "Literally hundreds of similar instances could be listed ... all of them indicating the sedentary habits." One is likely to remember Cirri'V's fruit-eating pigeons which were thought to be res|K)nsible for the dissemination across the Pacific of seeds too large for other birds. The "pigeons" are, now at least, restricted in range and of DERIVATION OF THE FLORA AND FAUNA 337 very little use on longer distances. I cannot help drawing the conclusion from this that the sedentary habit was acquired after the great colonization had taken place. Few phytogeographers have had greater faith in the capacity of wide-ranging marine birds to carry diaspores than Grisebach. In his discussion of bipolar species found in the far north and the far south but not at all in intermediate zones he selected Geiitiajia prosirata Haenke as the best example. Its distribu- tion is due, he says, to the wanderings of Diomedea exulans which, abweichend von der Lebensweise der meisten anderen Zugvogel, iiber beide Hemispharen, von Kap Horn bis zu den Kurilen und Kamtschatka, wandert und die Standorte jener Pflanze in der arktischen und antarktischen Flora in Verbindung setzt. Mil der Beute, die dieser Vogel verschlingt, kann er auch Samen von Pflanzen, welche, mil den Fliissen ins Meer gespult, in den Magen der Fische ubergehen, in einzelnen Fallen ausstreuen, so dass sie an fernen Kiisten aus seinem Dunger aufkeimen [32^. 469). I have not come across any comments on this bold theory. It is difficult to take it seriously, but to Grisebach the only gap in his argumentation was that nobody had happened to witness such an event. If he is wrong, he asks, why is there no trace of this Gentiana in the Andes, where it would thrive just as well as in the Alps and in the mountains of Asia? To this should be re- marked that G. prosirata is a polymorphous species of wide range and that it does occur in the Andes from Colombia to Chile, suggesting that it has mi- grated south along the mountains without the assistance of the albatross. Wallace (^/c^'. 259) tells us, on the authority of MoSELEY, naturalist to the "Challenger" expedition, of the great albatross breeding on Marion Island in the midst of dense, low herbage; I can add that this bird also breeds on South Georgia and on some other southern islands, but as far as I know they do not shift breeding places, and even if they did, they do not go on shore between the breeding seasons. TAYLOR, in his important paper on Macquarie Island [26 j), tells us about a giant petrel which was captured, tagged and released on this island and shot on South Georgia, 8000 km away, four months later, but these birds are often seen on land where they attack the penguin chickens; this was at least the case on Paulet Island in the Antarctic. Whether they aid in the dispersal of diaspores is unknown.^ Taylor quotes an observation, made on Macquarie, that seeds were found adhering to the feet of an albatross. These birds, when building their nests, regurgitate an oily fluid which makes seeds stick to their feet. Macquarie Island was ice-covered during the Glacial epoch and the plants, perhaps with the exception of some cryptogams, must have arrived since the retreat of the ice. The vascular flora consists of 35 species, all except 4 occurring in the New Zealand subantarctic area — the 3 species with a claim to be regarded as endemic should be reinvestigated — while those 4 species are found in subantarctic South America, from where they are derived. All Mac- ^ According to Taylor (p. 570) the two truly Antarctic phanerogams, Deschampsia ant- arctica and Colobanthus crassifolius, are very rare in the Antarctic and reproduce only vegeta- tively. I do not know where he obtained this information. They are scarce but have been reported from many localities along the coast and adjacent islands of Palmer Land between lat. 62 and 68 and, in favourable situations at least, both of them flower and produce ripe seeds — see my paper in Botan. Tidsskrift vol. 51, 1954- 22 - 557857 The Nat. Hist, of Juan Fernandez and Easter Isl. Vol. I 338 C. SKOTTSBERG quarie plants have, Tavlor states, propagules suited to bird transport. It is sur- prising that Wkrhi [jiy), who made a detailed study of the Kerguelen flora, asserts that not one of the flowering plants possesses any special dispersal mech- anism for either wind, water or bird carriage. Still, the two islands have some 9 species in common ( -,v). Taylor's conclusion that, "if long-distance dispersal has occurred on Macquarie Island, then it could well have occurred elsewhere" is certainly correct; we know, for one thing, many wide-spread sea-side plants and a number of widely dispersed aquatic species, possibly transported by migratory birds. South Georgia is in much the same situation as Macquarie, but still rather heavily glaciated thanks to its great altitude, and the possibility that any higher plants survived the Glacial epoch is very small indeed, whereas intlications that many mosses and lichens date from preglacial times are strong. The vascular flora is poorer than on Macquarie, and there are no endemic species. When Ta\i,()R accuses me of having argued against all overseas migration also n this case he must have misunderstood me. I expressly took this possibility into account in the paper he quotes (226). W'ali.ack calls attention to sea birds breeding on islands in the tropics; Phaeioji makes its nests on the Hawaiian Islands in 4000 ft. altitude and also in the highland of Tahiti, and such birds would account for the similarity of the mountain floras. In reality these floras have practically nothing in common. Miss GiUHS (?.?/), discussing the origin of the montane flora of Fiji, refused to regard birds as capable agents; wind may have been more efficient. No modern zoologist has tried to defend the theory of unlimited overseas migration with greater zeal than Zimmerman. In his Introduction to "Insects of Hawaii" we read: There is no evidence whatsoever to support the contention that they (i.e. the Ha- waiian Is.) are of continental origin or character, or that they were ever joined together in an elongate subcontinental land-mass or even in a continuous subaerial mountain range {2()8. 6). And, in opposition to certain other biologists he refuses to regard the islands as old, they are at most Pliocene and no part of them older than five million years; most of the lava is younger, the bulk of the land Pleistocene. Pie is opposed to my ideas but he thinks that the explanation offered by him will, partially at least, reconcile the differences between us. In an earlier paper (Amer. Naturalist 'j^y 1942), to which he refers, he spoke of former high islands, other than those found on maps, which once existed; once more the "stepping-stones routes" are called to life. Atolls are the remnant of many of them, or reefs like among the Leeward islands of Hawaii, and such preexisting islands would account for the immigration from all directions. He does not call for jumping of thousands of miles of open sea, but rather for series of shorter over-water steps. I am afraid that we need some substantial refurnishing of the Pacific basin to supi)ly a sufficient number of intermediate stations. Not all of these routes were, he says, available at the same time, and this would explain the apparent difference in age of various sections of the biota (pp. 51, 52). Most of the roads were cut off in Pliocene and DERIVATION OF THE FLORA AND FAUNA 339 early Pleistocene, some before Pliocene — when no land existed where stands the Hawaiian chain, if geologists are right. UsiNGER (^/j), describing the distribu- tion of Heteroptera, also looks for convenient stepping-stones. Divergent opinio7is on the means of transport. The dispersal agents universally recognized as important are air currents, especially monsoon and trade winds, and cyclonic storms, ocean currents, birds, and last not least man, who with his domestic animals and goods has become more and more important, whether he brings plants, seeds and animals to extend their range — and many of them become naturalized — or carries diaspores on his body and his belongings unaware. It is as a rule not difficult to find out where we have to do with human action, but we shall limit ourselves here to a discus- sion of natural factors of distribution. This subject has been treated by innumer- able writers in biology and a wealth of material was compiled by RiDLEY (20^), a firm believer in the great value of all kinds of dispersal mechanisms, some of which are, of course, very, wonderful. WULFF (2^1), in his chapter "Natural factors for distribution" is more critical. Of animals only the birds deserve to be men- tioned, but the plumage is no good for carrying diaspores any considerable dis- tance, especially over the sea, and the extreme marine birds, the strongest flyers, have no contact with land outside their own breeding-places. WULFF reduces the part taken by birds to almost nothing, but I believe that we have good reason to count with the migratory birds in certain cases. Water transport is responsible for the diffusion of litoral halophytes but rarely for migrations of inland species. Wind is important only on short distances, at least for seed-plants; special devices do not help very much. Even Ridley concluded that winged or plumed diaspores are not carried very far; spore-plants are more easily spread. Wulff remarks that according to Bentham Leguminosae and Labiatae hold their ground just as well if not better than Compositae. Altogether, if dispersal by natural factors had the significance ascribed to it, the vegetation of the globe, within a certain climatic zone, would be homogeneous and the sporeplants at least ought to be cosmopolitan, but they are "localized in definite areas, their distribution paralleling that of flowering plants" (p. 128). GoOD observes [log) that we have no proofs that species equipped with special dispersal mechanisms are more widely distri- buted than others. That certain plants with such devices show very wide areas whereas others without them are rare and local means nothing, because the reverse is also true; climatic and edaphic factors should always be taken into account. The relative value of the dispersal agents is put to the test when we deal with oceanic islands. Setchell (21'/) was inclined to give considerable credit to migratory birds. The occurrence of identical species of flov.ering plants in Arizona and Argentina and in California and Chile could be explained by bird transport, and bipolarity had originated in the same way. Bird transport helped him to understand the remarkable disjunct areas of arborescent Compositae and Lobe- liaceae; their birth place was in the Antarctic, and they had been carried by birds to New Zealand, Australia, Malaysia, Polynesia, Hawaii, South America and the high African mountains — we can add Saint Helena, Juan Fernandez and Des- 340 C. SKOTTSBERG venturadas. In this case it is not the question of identical species and rarely of j^enera. It is not improbable that the secret of their origin and early history lies hidden in the far soutii, but tiiis is all we can say. Hemslev (727. 66) when dis- cussing the endemic Coni[)ositae of Saint Helena and Juan Fernandez said that "wind seems at first to be the most probable agent"; still he doubted its efficiency. In Setciiell's view storms seem to ofler more than a possibility in many in- stances, particularly tropical cyclones and vertical thrombs able to carry even lieavy diaspores to a great altitude. In another paper (218) he points out that plant distribution in the Pacific has been from west to east against the prevailing winds and currents and, in the case of Hawaii, has given much better results than the expected east-west route; the "frequent cyclonic storms" are responsible for this anomal}' together with adverse biotic factors, but I fail to see why they shouldn't ofter the same difficulties for diaspores coming from the west. Andrews (^.615) paid s})ecial attention to the occurrence of scattered "Ant- arctic " genera and species in and around the Pacific and combines their distri- bution with the direction of ocean currents: In the South Pacific the westerly current sweeps by Australia, New Zealand, Tahiti, and the whole of the west const of South America, where it is joined by the cold uprising water along the South American coast. This gives rise to the north-moving Peruvian C'urrent which sweeps by Juan Fernandez, Peru, Central America, and Mexico, whence there is a detlection westerly toward the Hawaiian Islands and the tropics. This knowledge of the general circulation within the Pacific appears to throw a flood of light on the occurrence of the Australian, New Zealand and western South American elements in the Hawaiian flora, such as... (2 5 genera are enumerated). The influence of ocean currents is suggested i)articularly in the peculiar distribution oi Acaena, Gunnera, Acrfcra, On-obolus, Sautalinn, Sophora, and so on. Juan hY-rnandez lies outside the Peruvian current, but also the outer island, Masa- fuera, is reached by drift-wood. Its origin has not been investigated. A look at a current-chart shows that Andrews' reasoning has its weak points; besides, I cannot see that the genera he mentions are thalassochorous. I would recommend the reader to take a look at the many distribution maps pub- lished in a paper read on the same occasion when Andrews presented his argu- mentation ( iVeS). \\ ith regard to the cyclonic storms several authors have, as we have seen, em[)hasized their j)r()minent role in the violent dispersal of both plants and ani- mals; they are, in fact, considered to be the only imaginable force by which larger objects are trans|)()rte(l, and it is useless to deny that such events have taken j)lace and still take })lace, even if it is difficult to find definite proofs that the transport did lead to the establishment of an immigrant from afar. Most authors who have taken refuge in c>'clones have, however, expressed themselves in general terms without a clear idea of the extension of the cyclonic belts and the trend of the cyclones. In two j)apers \'lsiiER has summarized his studies on cyclonic storms in the Pacific. Three chief centres of origin are distinguished [322): (i) Western N.Pacific, originating some distance east of the Philippines in lat. 8° to 25°; (2) Western DERIVATION OF THE FLORA AND FAUNA 34I S. Pacific, particularly between Australia and Samoa in lat. io° to 25°; (3) Eastern N. Pacific off the west coast of Mexico and Central America. Occasionally tropical storms develop near Hawaii and over Australia. The normal course of (i) is WNW, recurring NE, and of (2) WSW, recurring SE. To what extent plant distribution runs parallel to cyclone tracks has, as far as I am aware, not been investigated. As ViSHER says, most of the cyclones originate over the sea "well out in the ocean" (p. 87). They hit many of the Polynesian islands, and possibly collect diaspores on one and deliver them on another, but a frequent dispersal of species in this way does not appear very probable. I wonder whether there is in Hawaii, with its 90 % endemics, a single flowering plant likely to have been borne there by a cyclonic storm. ViSHER is, however, opposed to land connections in the Pacific, with one exception: "it is known that Australia was formerly connected with Asia by way of the East Indies and New Caledonia" (j^J. 77). With regard to Polynesia he points to the west-east hurricanes and their colonizing power. Sea carriage also comes into the picture, violent cloudbursts may accompany the storm, brooks are transformed into swift rivers carrying plant material, tree trunks and soil, forming rafts, which are washed out into the ocean. A recent paper by Bergeron (j^^) gives a somewhat different aspect. His map shows the two areas in the Pacific north and south of the equator, where hurricanes arise, and how they move. The Hawaiian islands lie, as a rule, outside the tracks. The direction north of the equator is NW or WNW all through the hurricane belt, and there is no sign of an easterly direction enabling plants and animals to be carried all over Polynesia as far north and east as Hawaii, as ViSllER supposed. South of the equator the trend is S and SE; the east of the Pacific is not reached and Juan Fernandez lies, in longitude as well as in latitude, away from any cyclonic belt. It has often been stated that spore-plants, theoretically at least, have much greater facilities to colonize on long distances, but it has been shown that in reality their capacity is more limited than was formerly assumed. COPELAND (68. 165-166) expressed his opinion on the diffusion of fern spores: Fern spores are carried across water by the wind — ten miles of water is no barrier at all to their spread. One hundred miles may be one hundred times as great a barrier, because the spores must hit a target, a suitable place to germinate and grow. Still, ferns spread readily across seas this wide. A thousand miles makes the obstacle again one hundred times as great ... the limited viability of the spores, the chance of falling or being washed out of the air, and the chance of very different climate at such dis- tance, increase it materially. Ferns rarely jump a thousand miles of ocean. Still a number of species are believed to have crossed the south Atlantic, and I believe that three genera, Plagiogyria, Coniogramtne and Loxogra?nme, have flown the north Pacific from Japan to Mexico, each in one single instance. It is not exacdy impossible that direct colonization has occurred between Chile, New Zealand, Tasmania, the Cape, and Tristan da Cunha. It should be mentioned that CllRlST (59), with his unique knowledge of the distribution of ferns, pointed to the insignificance of spore dispersal as an explanation of the origin of widely disjunct areas. 342 C. SKOTTSBERG The fact that Irmsciikr tried to prove that the distribution of plants strongly su|:)ports \\'K(;fnkr's hypothesis does not lessen the value of what he says about the limited capacity of plants to migrate. It is small in the flowering plants [14.3. 291): Dass bei den Hliitenptlanzen die J^eforderung der Samen und Friichte durch Wind, Wasser und I'iere ganz wescntlich eingeschrankt werden muss und fiir geschlossene i-'orniationen auf grossen Kntfernungen hin nicht in Frage kommt, ist heute von den rtlan/.engeograi)hcn allgeniein anerkannt. Irmsciikr is too optimistic, we have seen that there are phytogeographers to whom overseas migration is not only possible and undoubtedly happens, on rare occasions at least, but rather of quite common occurrence. And with regard to s{)ore-plants their distribution should, if this be correct, show quite different distribution patterns than they actually do. The bryophytes are no exception to the rule. iRMSCiil'.R remarks that already in 1903 SlEPllANl denied that liverworts are able to make long and successful jumps. Attention should be paid to Domin's valuable j)aper [yd] in which he brings together numerous facts illustrating the same definite distribution patterns in this as in other groups, and Miss FULFORl) (e.g. 10 ;] has arrived at the same conclusion. We know that the spores, in many cases at least, are extremely sensitive to changed conditions and lose their viability very ra[)idly when exposed to the air — a promising field for experiments. Regard- ing mosses I refer to Herzog's work [i2g) where he speaks against the belief in the importance and great range of dispersal through the air. We find, IRMSCIIER says, the same disjunctions, the same part areas (Teilareale) in angiosperms, gymnosperms, ferns and bryophytes, and he continues p. 292 : Dass diese vier in ihren Verbreitungsmitteln so verschiedenen Pflanzengruppen die- seli)cn \'erhrcitungszuge ihrer Disjunktelemente ergeben, zeigt wohl einwandfrei, dass hierfur der "Wind" ebenso wie andere iiussere Krafte als Ursache abzulehnen sind. Wiiren sic in aiisschlaggcbender Weise an der Ausbreitung beteiligt, miisste die Besie- dclung entsj)rechend der Verschiedenheit der Friichte, Samen und Sporen ebenso ver- schiedcnartig ausgetallen scin, d. h. in den einzelnen Gruppen dieser biologischen Ver- schiedenheit entsprcchcnde charakteristische Merkmale zeigen. Dies ist aber nicht der i''all. Der alien vicr druppen ge m e i n sam en hochdisjunkten Ausbildung so vieler Areale nuiss vichnehr cine andere Ursache zu (irunde liegcn . . . This conniion cause was, in Irmsciier's opinion, continental displacement in the sense of Wi;(;ener. Those who disagree with him will have to look for vertical movements, emergence and submergence of land. Lichens, fungi etc. were not included in Irmsc:her's discussion. A survey of their distribution patterns is something to be asked for. Lichens are said to de- |)en(l on their vegetative reproduction bodies more than on spores. It is maintained that, with certain exceptions, terrestrial animals spread less easily than plants. 1 have consulted a number of zoogeographers in order to learn their opinion on the mode of transport likely to be used by invertebrates in their sup[K)sed ocean voyages. Birds etc. are left aside here. Numerous insects, butterflies, moths, flies, hymenoptera, grashoppers, cock- roaches, as well as spiders, myriapods, etc. follow man from land to land, from island to island. This is, I daresay, the only safe way for such animals to get DERIVATION OF THE FLORA AND FAUNA 343 abroad. For those, and they are in overwhelming majority, which are not an- thropochorous, the chance to cover large distances ought to be very small. How- ever, Zimmerman, in his admirable introduction to the Insects of Hawaii, surveys one order after the other and finds nothing that speaks against his belief in the permanent isolation of the Hawaiian as well as all other oceanic islands, and consequently concludes that, man-borne species excepted, all the ancestors of the Hawaiian insects were carried there by natural agents. Dispersal with the aid of birds is of slight importance, but he remarks that sea birds nest in forests on the islands; I do not think that this means very much, because they are stationary (comp. above, p. 337). Marine drift is probably the least successful of all meth- ods, he thinks. Thus, the bulk of the insect fauna was and is wind-borne, a traffic going on without interruption. Zimmerman refers to experiments clearly showing that both winged and unwinged insects are carried by air currents to great heights, 14,000-15,000 feet (p. 58). These are largely abnormal conditions and due to cyclonic storms, which account for the dispersal all over the mid- Pacific. The result is, as expected, a disharmonious fauna, where large groups common to all continents are lacking: "they have been eliminated by the selective agents of oversea dispersal". It would be interesting to know why all represent- atives of large and otherwise widely distributed insect groups are excluded from the passenger list. Besides, would not disharmony result even if land connections once existed.^ — the islands have remained isolated for a long time, perhaps millions of years, while migration, favoured by climatic and edaphic changes affecting the general character and composition of the vegetation has continued over land on the continents. HiNTON [dj) who, as we have seen, was opposed against both continental drift and land bridges, believed that wind-borne and raft-borne transportation across the oceans must have been of common occurrence. If this be so, why did no snakes, frogs or gymnosperm cones ever get aboard the rafts and arrive at distant islands.^ Rafts formed by large, uprooted trees are observed in big rivers like the Amazon or Orinoco, and I guess that an analysis of their composition would reveal the presence of a rather varied fauna. WALLACE'S Boa constrictor was referred to above (p. 335). Snakes and giant spiders are often found hidden in banana trunks imported to Europe from tropical America. On the other hand, the chances for the formation of substantial rafts are small within the tropical Pacific, where no big rivers empty. W/iat kind of invertebrates are likely to withstand transoceariic migration? HiNTON has an answer ready: The chance of accidental dispersal varies according to the nature of the group. Colonization of the Hawaiian and other islands, always far removed from any conti- nent, provides us with absolute proof of the kind of animals and plants that can with- stand long distance wind or raft transportation across the oceans. This sounds quite simple, but really is a very complicated problem, to which sufficient attention has not been paid. The possibilities vary according to 344 C. SKOTTSBERG size, flight capacity, habits, mode of reproduction, sensitiveness to changes of miUeu, and so forth, and I am afraid that we have httle knowledge, founded on facts, in most cases. liarth-iroDus have been carried all over the world with the human traffic. It is noteworthy that as a rule no truly indigenous species are found on oceanic islands, where, if land connections had existed, they could be expected, and even if the transport of eggs or living animals were effected only by means of rafts stocked with earth and plant material, as some believe, they ought to be present, but, as far as I am aware, nobody has witnessed such a transport. The presence of endemic leeches on Samoa and Juan Fernandez (Masafuera) can be understood only if the leeches are carried about on birds acting as hosts, otherwise I cannot see how they would be able to survive; they are very sensitive to exposure. To these in particular I should like to apply what GULICK, without referring to any special group of animals, wrote (z/^. 405): How is it possible at all for creatures that would die almost at touch of sea water to precede man by a million years on islands standing solitary in mid-ocean? Are their remote homes really the left-over fragments of ancient inter-continental land bridges, or are these creatures prima facie evidence that their ancestors possessed an almost incon- ceivable capacity for passing uninjured over vast stretches of open ocean? The extremes of hypothesis that have been proposed in response to this dilemma show us how difficult it has been to find a solution. Freshwater criistaceaiis occur on many islands, both traffic-borne and indigenous species; only the raft theory would account for their spread. It is supposed that cocoons of spiders are transported by wind, webs are torn loose with cocoons attached and carried up into the air, where a storm takes care of them. Theoretically this is not impossible, but whether the contents stand a journey of thousands of miles is doubtful. Adventitious species are found on Pacific islands, but the bulk of the spider faunas is indigenous and endemism is high. Hkki.anI), pointing to the general distribution of genera and the high spe- cific endemism, is in favour of former land connections [2j. 1052). L'isolation doit etre assez ancien pour ce que cette fauna ait pu acquerir les carac- tcres d'endemisme (ju'elle presente. Un botaniste a fixe vers le Pliocene cet isolement, mais jc serais porte a croire (pi'il est i)lutc)t ])lus ancien, en me basant sur la lenteur de revolution des Araignees. AI.\\R, who rejects all land connections, remarks (27^.214): Considering the haj)ha7,ard manner by which these oceanic islands receive their popu- lations, it is rather astonishing how similar the faunas of the various islands are. Ber- i.ANi), on the basis of the distribution of spiders, has come to the conclusion that the fauna of all Polynesia is so uniform as to suggest that these islands are but fragments of a single land mass. This view is similar to Pilsbry's, founded on Mollusca. Actually, this j)aradox of tlie similarity of the fiiunas of oceanic islands is solved in quite dif- ferent manner. Of all the j)Ossil)lc families, genera and species of the Papuan Region that are theoretically in a position to colonize, only a small fraction will eventually avail themselves of the opportunity. DERIVATION OF THE FLORA AND FAUNA 345 This is quite true, but will not the result become much the same with land connections? The main source of the fauna is the same, the similarity is a con- sequence, and local endemism is a result of isolation. Acarids present the same problem. MUMFORD [183) remarks that is is very difficult to compile lists of species for the Pacific islands and that, at present, no safe conclusion can be drawn with regard to the distribution of genera. This is true, but we know that numerous indigenous species occur and that local en- demism seems to be high. How these extremely delicate little animals would be able to stand long overseas voyages is difficult to imagine. They are plentiful in humid forest soils on Juan Fernandez and all the species except two adventitious ones are endemic. Little is known of their relationships. If washed down from the hill-sides and carried out into the sea, they will die — only wind transport remains. Pseudoscorpio7is may be more resistant to both salt water and desicca- tion, but their pronounced endemism bears witness of long isolation. All the false scorpions recorded from Juan Fernandez are endemic, and there is one endemic genus. They were unknown when MuMFORD (I.e. 246) wrote: As Chamberlain points out, it is doubtful whether anything like true insular endemism occurs in most species of false scorpions because of the ease with which they are distrib- uted. They are well adapted to be carried about by man. Among the Myria- poda are many local endemic species. These creatures are, according to my own experience, rather tough and might be able to spread by the same methods as earth-worms and land crustaceans. It is not very probable that they are blown from continents to distant islands. Cosmopolitan forms are probably adventitious. If this applies also to Collembola I do not know; very wide-spread species may be so ancient that they have attained their distribution when the map of the globe was quite different from the present one. It is unlikely that they are able to migrate overseas. Lindsay [i6y. 7 1 9) writes : The primitively wingless Collembola seem to constitute better material for distribution studies than any other insect order, because migration by flight is imp'ossible and the delicate integument makes it very unlikely that the insect could be carried any appre- ciable distance by the sea. We can safely add that, if blown out over the ocean, they would soon perish. Zimmerman suspects that none of the 32 species recorded for Hawaii is indige- nous there; thus, all are supposed to have been imported by the traffic. How- ever, 3 species are supposed to be endemic in Juan Fernandez. Two species of the Hawaiian Thysanuva are "possibly endemic" (Zimmerman) but perhaps adven- titious. These creatures do not appear to be fit for long-distance dispersal, and the two species known from Juan Fernandez are endemic. One of them belongs to a monotypical genus with Australian affinities, a disjunction not easily bridged over without land connections. Getting to the true insects, their mode of dispersal certainly varies a great deal. Whereas butterflies, moths, flies, hymenoptera etc. are known to be storm- driven and eventually carried far, heavy beetles, even if they be properly winged. 346 C. SKOTTSBERG are unfit for long journeys; whether they are transported by waves and currents and able to stand immersion in salt water during weeks and months I do not know. The endemic Sightless insects have, just as the flightless birds, given cause to much s[)eculati()n. They are supposed to descend from winged species; arrived, it is said, on an oceanic island, they had the choice of losing their wings or being blown oft' the island and lost altogether. ZIMMERMAN does not favour this view, for they may as well have lost their power of flight on the mother continent, which did not prevent them to be carried off by a hurricane. Mayr agrees with him, comp. quotation p. 336. Very well, but is it possible to imagine a flightless rail carried a thousand miles across an ocean.' There are numerous Ortlioptcra on Hawaii, most of them endemic species or even belonging to endemic genera. So far only 4 species have been reported from Juan Fernandez, two of them endemic. I presume that Zimmerman regards these insects as normally wind-borne. The relations of the Hawaiian Gryllidae are with Indo Pacific forms, so they have had a long way to go. The four species of tciDiitcs found in Hawaii are adventitious, which should indicate that they are unable to reach oceanic islands without human assistance, but the single species discovered in Juan Fernandez (Masatierra) is endemic. Either did it, or its ancestor, arrive over land, or a colony was carried in a floating log, which may seem un- likely, or a storm brought a winged, swarming couple which founded a new colony — even less probable. The termites are a very ancient order and date, it is said, back to the Mesozoic at least. Mallophaga are spread with their bird hosts. All the endemic Hawaiian species live on the Drepanididae, and the marine birds here and on Juan Fernandez are infested with widespread forms. Tliysanoptera seem to be easily' spread with human traffic. Most of the 90 species found in Hawaii are adventitious, few indigenous. Two of the 4 species in Juan h'ernandez are endemic. How these delicate insects manage to get about and to reach oceanic islands I cannot tell. Tlie Hawaiian islands have a rich and peculiar fauna of Xcuroptera, some "among the most aberrant of all" (ZIMMERMAN p. "j6). Of the five species re- corded for Juan Fernandez 4 are endemic, one of them belonging to an endemic genus. Wind drift must be taken into account, but it cannot be very effective. The situation remains tlie same when we get to the Lepidoptera, about lOOO species in Hawaii, of which 85 % have not been found elsewhere. Wind drift or immature stages (eggs, etc.) carried with plant material are the only possibilities, but they cannot be very great. vSo far only 26 indigenous species are known from Juan I^'ernandez, 70% of them endemic; when the list of Dr. Kusciiel's collection, which contains over 50 species, has been published, the figures will undergo alteration. In his survey of the Pacific lepidoptera Swezev [262) includes the Galapagos Islands, but does not mention Juan Fernandez. The Pacific islands were, he says p. 319, j)opulated from the Malayan and Oriental regions, and the fauna arrived in the main by accident, winds, ty[)hoons etc, or with plant material brought by currents. A comparable numerical development of species per genus has taken place in no other islands than the Hawaiian, and the author infers that they have a more ancient fauna, descended from ancestors that arrived at a more remote DERIVATION OF THE FLORA AND FAUNA 347 time. This does not accord with Zimmerman's statement that no primitive family is represented in Hawaii. Probably the Hawaiian Diptera have been less collected; Zimmerman, who is responsible for the figures quoted, indicates about 400 species, of which 60 % are endemic. In relation to their small size the Juan Fernandez Islands seem to be richer with about i 50 indigenous species (64 % endemic). Our knowledge of the dipterous fauna is in the main due to Dr. KuscilEl/s collec- tions; the 25—30 species not yet reported on will raise the total number. Over 1600 species of Colcopiera are reported from Hawaii, of which about 'j^ % are endemic, and there are numerous endemic genera. Numerous species found elsewhere are adventitious. In relation to its size, Juan Fernandez cannot be called poor (see above p. 307, etc.), and endemism is just as high here. I presume that Zimmerman and others regard wind as the principal dispersal agent, though not for all kinds of beetles, because we have to do with many different types of animals and of habits and habitats. It is difficult to imagine how a flightless beetle would be able to keep afloat in the air for thousands of miles; he must have had a great need of the numerous "stepping stones" postulated by ZIMMER- MAN. It should perhaps be mentioned that GULICK (77^.414) wrote that "it can hardly be doubted that some carrion-feeding insects have been distributed by adhering to sea birds". I doubt that this ever happened, but as he had just dis- cussed the transport of "invertebrate eggs" in the digestive tract of birds, I sup- pose that he means eggs of necrophilous flies or beetles which, brought across the sea, were deposited on another carrion and thus became established on an oceanic island. The taeevils, a most important and interesting feature in isolated island faunas, are often dependent on definite host plants, and are thought to sail along on logs as stowaways. Evidently the Curculionidae have become something of a stumbling-block. UsiNGER (2yj) came to the conclusion that they must possess some unknown special means of dispersal. Brinck, who discussed the coleopterous fauna of Tristan da Cunha (316. 97—104), another isolated volcanic group of islands where geologists failed to discover any traces of land connections, states that the fauna contains endemic elements and ofl"ers examples of remarkable disjunc- tions. It must have originated from extinct faunas of neighbouring continents if not of submerged lands. The only natural agent capable of transportation is the wind, but Brinck is convinced that "at present no beetles are invading the islands by natural means" (p. 103), and the reason is not adverse conditions, for several species, introduced with the human traffic, have become naturalized. The unavoid- able conclusion is that dispersal agents, man excepted, are insufficient — and they were the same in the past. The original beetle fauna has survived from pregla- cial time, an hypothesis that nobody would feel inclined to reject, but it does not help us to understand by what means it was able ever to arrive. The weevils of Hawaii have had a long way to come; according to ZIMMERMAN the ancestors as a rule came from the south Pacific or Indo-Pacific regions. Generally he regards also the peculiar genera to have originated in Hawaii or, eventually, in one of the lost islands serving as intermediate stations, but among the weevils are some that defy all explanations: 348 C. SKOTTSBERG Xt'sotflcus is evidently a relict endemic genus of four closely allied species, and there aj)pears to be nothing like it elsewhere . . . Oodemus with its 58 species ... is the largest genus of the Hawaiian Cossoninae . . . together with its close ally Atiotherus (3 species) endemic, and I know of no genus or group of genera from any region from which is might have come. It is an anomaly. Such cases brini^ us back to times long before the formation of the present Hawaiian chain, and similar cases are found also in Juan Fernandez, e.g. the endemic tribe yidviorJuui of .Vl KI\ILL11 s. Other examples are offered by many other insect groups. And, leaving them aside for a moment, is not the endemic Hawaiian bird family Drepanididae another anomaly.^ Whereas Brvan [j^g. 188) finds a Malayan origin most acceptable, GULICK writes (7/^.420): The history of Hawaiian land birds must have begun with the arrival of some form of troi)i(al American honey creeper, which became in due time the progenitor of all 18 genera and 40 sj)ecies of the Drepanididae. Me seems to have forgotten that the islands are claimed to have cHved out of the ocean in late Pliocene and Pleistocene times. I lyinoioptcra are plentiful in Hawaii, about 600 native species, among which endemics are numerous, and the ancestors are supposed to have come from the south and southwest Pacific, in exceptional cases from Asia and the Orient. This order is as yet little known in Juan Fernandez, see above p. 315. Wind drift seems to be the only possible mode of transport unless infected larvae of butterflies etc. arrived with drift-wood, which does not seem very probable. Ants are easily spread with the traffic. Of the 3 species known in Juan Fernandez only one appears to have arrived without human assistance. Among the Heteroptera in Hawaii, over 200 species and 80 % endemic, ZlM- MKRMAN pays special attention to the genus Xysius, which has its greatest known diversity in those islands. All are endemic and include the most divergent of all Xysius species. The ancestors are supposed to have come from the south and west Pacific. The genus is, according to UsiNGER (2'/j) common in the Australian and Oriental regions, extending through Melanesia to Fiji anci Samoa without a single re[)resentative east of this line, but important in the Hawaiian chain; the author seems to have overlooked its occurrence in Juan Fernandez. Xysius is su()posed to have reached Hawaii by a circuitous route over open water and the Leeward Hawaiian islands. This route is indicated by a submarine ridge of con- siderable depth and may once have been interrupted by island peaks such as Wake Island; thence it is followed to the Marianas and Caroline Islands and even- tually to the rich Papuan and Australian regions, l^ut in other cases it is less eas)- to construct a suitable route: "The presence of twenty very unique genera in Hawaii and their absence from old, high islands along the very route they are said to have traveled is inexi)lical)le by present theories" (I.e. 315). Why not presume that all related genera have died out.-, which is the easiest explanation. Xysius is a widespread genus, well developed also in New Zealand, and the single Juan I^'ernandez species is claimed to be related to another from New Zea- land. We know numerous examples of the same kind in other animal groups and particularly among the plants we have called Antarcto-tertiary. DERIVATION OF THE FLORA AND FAUNA 349 The Homoptera present the same picture in Hawaii, about 350 native spe- cies, most of them endemic and suggesting a southwest Pacific ancestry. Of the species reported from Juan Fernandez only one-third have been described, all endemic and as far as we know with their relatives in the south and central Pacific. No order characteristic of insular faunas has aroused greater interest among biogeographers than the Pacific larid molluscs, nowhere more wonderfully devel- oped than in the Hawaiian chain. In the discussion of land connections they occupy a central position. Wallace found that the wide distribution of the land snails is "by no means so easy to explain as that of the insects"; the chances have been "rare and exceptional", possibly eggs stuck to the feet of aquatic birds, or the animals themselves were storm-carried, "attached to leaves and twigs" — this would be the only means by which viviparous forms could be transported. GULICK [iig] treats the land-shell problem at some length. Speaking of Easter Island he asserts that "hurricanes spread gravid land-snails as dust over almost as great distance as plant seeds can be blown", but he gives no facts to support this very positive statement. The land-shells of Juan Fernandez and Saint Helena are then remarked upon. At least three elements can have derived their ancestry only from Polynesia, fully 3400 miles away, unless Easter Island served as a way station. The archaic complexion of the snail fauna is not necessarily very significant, as younger continental forms do not for the most part yield minute, easily wind-blown species. Archaic — exactly, malacologists emphasize that more modern types do not occur on oceanic islands unless brought by man, and this has, naturally enough, been used as an argument in favour of early land connections before the modern types existed. In Gulick's view the size, not the age, decides. Large, softskinned creatures invariably make a poor showing . . . large helices ex- amplify this disability so excellently that their failure to arrive is a sort of negative criterium for insularity (p. 414). But neither are all continental species large, nor all insular ones small, and the oceanic snail fauna includes many forms that cannot spread like dust. GULiCK remarks on the genus Partula that "its 120 geographically restricted species mostly weigh too much and are too tender to fit easily into theories of transport by air or sea", and this makes him take into account the possibilities of "land ridges" to facilitate transport. "It is evident", he says p. 419, "that the vast diversification is a proof of the great local antiquity of these families, and hence of a considerable antiquity of their island habitat." This is not true of the islands as they appear now^ and as they have stood for probably millions of years. Land snails, just as weevils, are dispersed by some unknown method, UsiN- GER thinks (I.e. 315), while Mayr, as already quoted, regarded the presence of small molluscs on oceanic islands as a proof of the efficiency of hurricanes, but later on expressed himself as follows. 350 C. SKOTTSBERG It seems to me that the wide acceptance of land bridges by conchologists is chiefly due to three reasons: (i) our almost complete ignorance of the means of disi)ersal of snails, (2) our lack of knowledge of the speed of speciation in snails, and (3) faulty classi- fication, i)articularly generic classification. A. Gulick has already directed attention to the ])resence of snails on most oceanic islands. 'J'hey were unquestionably carried there by some unknown means of transjiortation. . . . To me it seems incomparably simpler to assume a still unknown method of transportation than a land bridge that is unsupported by any other fact. Simple, no doubt, but we cannot get away from the problem by an "ignoramus". It is eas\' to under.stand that the presence of fairly large forms of land-shells on distant islands has caused a good deal of trouble. ZIMMERMAN tried to find a wa\' out of the difficulties. It has been said that large snails such at the Hawaiian achatinellids and amastrids are particularly unsuited for overseas distribution. However, if we approach the problem differentlv, different conclusions may be reached. If, as I believe, the large Hawaiian snails have evolved from small or minute ancestors, then the argument based upon their large si/e loses its weight. However, if small snails can be distributed overseas, then what is to j)revent eggs or tiny, immature specimens of large species from being similarly transported? (I.e. 61). In passing, ZiMMKRMAX quotes H. B. 13.\KKR, who thought that land-shells are carried along by migratory birds. After these speculations it is refreshing to read Hrvax (^o. 9): The j)resence of certain kinds of plants and animals found in Hawaii and related to species in the southwest Pacific is hard to explain by any known means of drift, either over the sea or through the air. Land snails constitute one such group. These mollusks, whi( h breathe directly from the air, would drown in water, particularly salt water; yet they must have moisture. They cannot stand long exposure to the sun, but live on the leaves and trunks of forest ])lants or beneath fallen leaves and trash on the ground. How did their ancestors reach Hawaii if they could neither swim nor drift? ( ^f^poju'iits ai^^aiiist the doi^iua of colonization across tJic oceans. Many biogeographcrs have arrived at the conclusion that the natural dispersal agents cannot be made responsible for the distribution of all kinds of biota across very wide expanses of open water. Most of the authors are botanists, which is sur- prising because the chances should be greater for seeds and spores than for eggs or individuals of delicate creatures. To ask for land bridges, or for extension of continental margins later submerged l)ut leaving behind land fragments, is to refuse to accept overseas migration as the only possible means of colonization. Conse- (piently, a number of authors have already been quoted in the chapter on the history of the Pacific basin, (iooD, Ikmsciikk, WTij-f, Cami'BELL, etc., as well as my own contributions to the discussion, the latest in 195 1 [2^8). I shall add here that I never disclaimed every possibility of migration over wide expanses of water; see 226 and 2:12. X'arious writers have compared isolated peaks on continents with oceanic islands and have stated that under present conditions an exchange of biota is improbable. \\\x SlKKMS expressed his opinion when dealing with the Malaysian mountain flora, in very plain terms [2^8)\ too many facts DERIVATION OF THE FLORA AND FAUNA 35I show the impossibility of attributing any importance worth mentioning to long- distance dispersal; the theory is, when it comes to migrations of floras, "not worth a straw", and Reiche (20j), referring to the history of the Chilean flora, called it "eine Kette von Unwahrscheinlichkeiten". Gordon (iij) was strongly inclined do deny ist value and believed that plants, and more so plant associa- tions, advance slowly over land and do not jump thousands of miles; he could not, however, help paying some attention to "Nature's great Krakatau experi- ment", of which enough has been said. Cain (42) thinks that "migration is usually not a random matter" (p. 162) and that long-distance dispersal rarely has resulted in migration and establishment, nor does it explain the discontinuous areas. In reality he belongs to the trust- worthy opponents. One of the reasons advanced by him is, however, not con- clusive: The phenomenon of local races (subspecific endemics) is entirely opposed to the idea of long-distance dispersal, for such variation depends upon isolation which would not exist if long-distance dispersal were generally effective (p. 161). Those who are in favour of the theory emphasize the haphazard character of the procedure; success may follow once and never more. If the immigrant belongs to a polymorphous species, with intermediate forms between the sub- species or varieties, it may, isolated as it remains from the rest of the population, stand out as a separate insular taxon. Among the zoologists few, mainly malacologists, are in favour of land con- nections, but some entomologists agree with them. Berland was already quoted. As a rule a connection South America-Antarctica-Australia (or New Zealand) is asked for, but to Enderlein this bridge was not sufficient: Die zahlreichen endemischen Gattungen (9) und Arten (2) zeigen, dass die Juan Fernandez-Inseln ein Refugium fiir die Reste der Faunen unfangreicherer untergegan- gener Gebiete darstellen, die nicht mit dem neotropischen Gebiet in Verbindung gestan- den haben (Enderlein S4. 643). Did Enderlein dream of a submerged Pacific continent.? Many of the island flies are also native of the mainland of Chile, others have their relations there, and if the islands were formerly connected with some other land, it was with South America. Besides, the majority of Enderlein's new genera have been reduced to synonymy by later authors, his speculations have little weight, and Wygodzinsky (2^^. 81), referring to Gigantodax kiiscJieli, arrived at a different conclusion. Gigaiitodax is an exclusively South American genus. Chapter VII. Biological characteristics of isolated islands. In his classical "Lecture on insular floras" [138) J. D. Hooker formulated, in very lucid terms, the special features of island floras and his opinion on their evolution. As examples he chose Macaronesia, St. Helena, Ascension and Ker- guelen Island. Their peculiarities were stated under five items. 352 C. SKOTTSBKRG 1. /;/ all casts ciDisidired flcrislic relaliois exist beiwcen the island and one niot/ier coiitinejit. 2. rite floras of all the islands in question are more temperate in character tha)i that of the mother cojitiiioit on the same latitude. This ina\' be true in most cases thanks to the influence of the surrounding ocean; it holds <;o()tl for Juan P'ernandez. 3. All these islands show many biological peculiarities by which they are distifii^uished. Tlie chstin<^uishino characteristics are mainly expressed in endemism. HoOKER referred the etidemics to two cate<^()ries, such as do not show affinity to the plants on the mother continent, and such as, even if belongini^ to endemic genera, are related to continental ones. If we turn to the Pacific where, for obvious reasons, only the high volcanic, well watered islands are considered, we find that all of them are distinguished by numerous, in man\' cases also very remarkable endemics. Hawaii stand out above the others and so do Juan Fernandez and Desventuradas; the floras of Tahiti, Samoa, Marcjuesas etc., as well as of Micronesia, are less independent. The degree of s|)atial isolation is not conclusive; the flora of Juan Fernandez is more peculiar than that of Marquesas which are situated much farther away from any continent, and this holds good not only for the angiosperms but also for the ferns. F^n- (lemism in angi()sj)erms is 69% in the former and 50 in the latter; of the ferns about 30 "o in each, but only Juan P^rnandez has an endemic genus. 4. the i!;e}ieral rule is that the species also fotaid in the mother continent are the most abitndant, the peculiar species are rarer, the peculiar j^enera of contineiital affinity rarer stilf but the plajits with 710 affijiity elseivJiere are often i>ery commoii. This is, I supj)ose, true of the islands examined by HOOKKR, with the excej)tion of St. Helena before the arrival of man, but not of Hawaii, nor of Juan l''ernandez. Some of the continental species — Libertia and a few grasses — are abundant, while others are rare, all according to the supply of suitable habitats; atuong the j)cculiar species of continental affinity are many quite common ones, e.g. the endemic species of Acaena, Driniys, Dysopsis, Es- callojiia, (iuniicra, Myrceui^enia (the leading forest tree on Masafuera), Pernettya, Rhafhithannius, ( ]<^n/\ flrijrcroji fruticosus, Uncijiia l)ouglasii\ and of the peculiar genera allied to South yVmerican ones, Xot/iomyrcia is the leading forest tree on Masatierra, where Ochaga/da is also common. To these may be added such common endemics as lioehmeria and the species of /v/^^'-<'7;'rt: and Cop rosm a, unde^v the assumption that related species of lloehnieria, and re{)resentatives of Coprosma and i-agara, once belonged to the neotropical flora. The plants of no affinity in the luother countr\-, that is the nearest continent, are as a rule very local, few are common and many extremely rare. Fossil)])- I have misunderstood llooKKK here. When we say that a species is abundant in a country we mean that it is copious; if we call it common, it is widely spread; if we call it rare, it has been rejjorted from a small number of localities only; and if we use "sj)ecies" in plural we mean the same thing. Hut what if HooKHK with "abundant" and "common" wanted to say that these DERIVATION OF THE FLORA AND FAUNA 353 Species were the most numerous in the flora, and that the "rare" ones were few in number? If this was what HoOKER meant, the result will be: Species also found on the continent: 46; endemic species allied to South American or other conti- nental species: 71; endemic genera, related to continental genera: 5 (6 species); endemic genera not related to continental ones: 12 (24 species). Still, this is per- haps to give a wrong interpretation to Hooker's words. There are very few native non-endemics on St. Helena, and they are of course not in the majority in the Canary Islands, nor are, in the latter place, the isolated endemic genera and species very many. In this case HoOKER says "plants", not "genera" and "species". If, to this group, as represented in Juan Fernandez, we add isolated endemic species, not related to any species in the mother continent, although the genera occur there, this group comprises 54 plants, or more than ^/g of the angio- sperms, while group 2 is reduced from 71 to 40. The figures would be: 46, 40, 6, 54, thus conforming much better to Hooker's rule. 5. Indigenous annual plants are extremely rare or altogether abse7it. Here "rare" must mean few species, and Therophytes are very few in Juan Fernandez and some of the registered species perhaps not originally native. How were plants transported to distant islands.^^ Hooker's answer is: either across the sea or over submerged bridges, and he adds: "the naturalist, who takes nothing for granted, finds insuperable obstacles to the ready acceptance of either". The situation is still the same 90 years after Hooker. Hooker regarded the isolated island plants as "relics of a far more ancient vegetation than now prevails on the mother continent", but he most certainly never wanted to say that the continental flora was altogether younger than the insular but that species, now restricted to the island, formerly occurred on the mother continent, having become replaced there by younger species. He based his opinion on the fact that Macaronesian relicts had been found as fossils in Tertiary deposits on the continent. Time has not permitted me to collect modern data, and I can only suppose that some of the old determinations still hold good. The vegetation of Europe has undergone great changes "within the lifetime of these Atlantic island species"; they once grew in Europe, but were driven out from there to be preserved on the islands, which they had reached "when condi- tions may have been very different from what they are now". The theory of a continental Macaronesia, including Madeira and the Azores, goes back to Forbes' theory of the former connection between the British Islands and the mainland, definitely proved ages ago. FORBES went further and revived the old idea of a lost Atlantis, still favoured by many. It is interesting to follow Hooker's discussion with his friend Darwin on island problems. Darwin believed in the efficiency of dispersal agents to carry plants and animals across wide expanses of sea, and his arguments made such a deep impression on HoOKER that he became almost convinced. Still, he hesi- tated, and certain serious difficulties prevented him from fully accepting Darwin's ideas. The composition of the flora of the Azores was not what we had reason to expect from the direction of winds and currents. The Macaronesian Ornis is almost the same as in Europe and undoubtedly came from there, but the flora 23 - 557857 The Nat. Hist, of Juan Fernandez and Easter Isl. Vol. I 354 SKOTTSBERG is considerably diftcrcnt; thus it may be argued that "the birds and plants do not come under the same category". Darwin replied that the migration of birds is continuous and frecjuent, and the individuals surviving and breeding, they keej) uj) the specific type, and do not give origin to local varieties; whilst the transport of seeds is casual and rare, and very few surviving, these not being crossed by the original stock, in the process of time give rise to varieties, etc., and do not j)crj)etuate the continental races (p. lo). This is the situation in a nut-shell, and Darwin's arguments are repeated by scores of biogeographers to this very day. Also St. Helena, Ascension and Kerguelen made HoOKER hesitate: They St. Helena and Ascension] have no land birds, but an African vegetation; and though nearly midway between Africa and America, they have scarcely a single American tvpe of flowering ]:)lants; and Kerguelen's Land has a flora of whose ele- ments most have emigrated not from the nearest land, but from the most distant (p. lo). 11i:msi,i:v [12'j. 59) remarks that Hooker seems to have forgotten the Compositae in St. Helena, most of them showing American affinity. Kerguelen's nearest land is Antarctica, but not a single flowering plant is known from the coast south of Kerguelen. Africa as a mother country — it goes without saying that subtropical or warm temperate plants cannot endure a subantarctic climate, and only the most distant lands, Tierra del Fuego and the Falkland Islands, were, thanks to the strong and constant west wind drift, re- garded as a mother country. Fven if South Georgia served as an intermediate station, the distances are very great; besides, we do not look west for the an- cestors of the peculiar endemics in the Kerguelen area. 'V\Tere the capacity of the dis{)ersal agents appeared to be inadequate, HoOKER was strongly inclined to look for better land connections. The existence of identical Macaronesian species on Madeira and the Canary Islands can, he says, hardly be explained without the help of intermediate masses of land, as the Salvages (supj)osing them to have been larger) . . . the only conceivable means of interisland transport . . . and if intermediate islands are granted (and Mr. Darwin freely admits these), why not continents? He must have found that the distance between Madeira and the Canaries is too large to permit direct transport of diaspores under present wind and current conditions; in the Kerguelen case they are, at least, favourable. Nevertheless, later on Hooker's faith in transoceanic migration was not as steadfast as before; the case of Kerguelen troubled him (.i""^?^): Turning to the natural agents of disi)ersion, winds are no doubt the most power- lul, and sufficient to account for the transport of Cryptogamic spores; these, almost throughout the year, blow from Fuegia to Kerguelen Island, and in the opposite direction only for very short i)eriods, but appear quite insufficient to transport seeds over 4000 miles (j). 13). Various j)henomenons . . . common to . . . Kerguelen, the Crozets and Marion, favour the supposition of these all having been peopled with land plants from South .\merica by intermediate tracts of land that have now disai)peared; in other words, DERIVATION OF THE FLORA AND FAUNA 355 that those islands constitute the wrecks of either an ancient continent or an archipelago which formerly extended further westwards ... (p. 15). Gulick's paper "Biological peculiarities of oceanic islands" does not, con- trary to its title, contain a review of the special characteristics of island biota; his object was to expound and defend the theory of permanent isolation of is- lands like the Hawaiian, Galapagos, Juan Fernandez, St. Helena, etc., which are said to offer irrefutable proofs of true oceanity. Of this enough has been said already; I shall return to Hooker's five points, to which others may be added. Endemism. — The occurrence of numerous genera and species restricted to oceanic islands has caused much discussion. "Reichtum an Endemismen ist iiber- haupt der hervorragendste Charakterzug der Insel-Floren", Hayek wrote [304). It is, however, equally pronounced in continental districts like the Cape region, south- western Australia, western China, California or Chile, where local concentrations of endemics are found. If insular endemics show distant affinity only or, in extreme cases, no af- finity at all, to continental taxa, they are looked upon as relicts; as the islands are geologically young, the endemics have not evolved there but must have im- migrated from some mother country, where they have become extinct. They may, however, have undergone some change after their arrival to the island. There is also a possibility that the continental progenitor has, in its turn, changed in a different direction, making its descendants so unlike that their relations are obscured. Species only slightly different from continental ones are much more numerous than the relicts; they are supposed to have originated in the islands and give examples of so-called progressive endemism. As Ciiristensen {60. 149) pointed out, another alternative leading to the establishment of endemic species should be considered. On the continent, from where a plant found its way to an isolated island, opportunities for crossing with other species often exist, eventually leading to the disappearance of the original taxon with its special characteristics. Its island offshoot does not share its fate but remains true to the original type. The island form did not originate through a genetic change of the continental species: it represents the surviving species and is, as it were, a relict. This does not apply to the pteridophytes. Crosses are extremely rare, the fern species represent, in a high degree, pure lines, whence it follows that insular en- demics are much rarer than among the phanerogams. This is true, but it is usually explained as a result of the enormous spore production and the facility with which they spread. Opinions about the true nature of systematically isolated taxa vary. GUPPY {122) regarded them as either highly specialized products of the islands, "the first of their race", or modified forms of allied continental genera, the majority of which had passed away, "the last of their race" and probably doomed; to him the islands appealed "more as registers of past floral conditions in the continents than as representing their present state" — this in accordance with Hooker's views. The Age-and-area theory of WiLLlS [286) claims that wides are older than endemics, a rule with few exceptions; in another paper [283) he states that "insular endemic genera are as a rule young beginners, not relics". I have dis- 350 C. SKOTTSBERG cussed his theories in an earher paper [jji) to which I refer. RlDLEV {20j) called the island endemics "newborn s[)ecies ", admittinf^ that all did not fall within this cate^ors- but were "epibiotics, relics at the end of their species life . . . unable to reach another suitable spot for their |j;rowth". It is noteworthy that so many of the "e[)ibiotics" are Composites, famous among diffusionists for their alleged effective dispersal mechanisms; in RiDi.KV's eyes they are, perhaps, pseudo-relicts. Cain (^2) takes more or less the same position: "the relic nature of an endemic should never be accepted without some form of positive evidence" (p. 227); proofs are hard to tind, no island cases as clear as Giiikgo or Mctascquoia are known. An endemic inhabiting a strikingly limited area may be a young species that had no time to spread, or it may be too stenotoj)ic, but others are what Cain (p. 230) calls "senescent"; such species occup\- a small area, are relatively constant, ecologically of narrow^ amplitude and show low competitive ability. They are unable to "penetrate the prevailing habitats that are dominated by the typical vegetation of the region" but behave just as stenotoj)ic young beginners. If indeed old and senescent they ought to siiow some primitive characteristics. According to Cain senescent species con- stitute "an anomalous element in the flora of a given region"; this may be true, for in many instances they survive from an earlier climatic period and are barely able to hold their ground under the changed conditions. Nevertheless there are cases when such anomalous species form the typical vegetation of a certain habi- tat and where nobody would dream of regarding them as young beginners; the '^ Robi)iso)iia assemblage" in Masatierra offers a good example (see 2ji). Xuiubcr of species per genus.— \Nq know that island floras contain a fair num- ber of monotypical genera, many genera that are large elsewhere but repre- sented on a given island by a single species, and few with many species, so that tlie numerical relation between species and genus approaches i and does not exceed 2. and this has been regarded as a good proof that the island is truly oceanic and has been peopled -accidentally by waifs and strays. This rule is not without exceptions, among which the Ilawaiian flora is the most striking. Fos- i:Kk(;, who contributed a cha|)ter on the higher flora to Zimmerman's book (2^8), indicates ^^) g'^'t the ratio 6.2 : i, but scores of well-marked species have been de- scribed since i 8oden gewinnt, geht zum Beispiel aus dem Buche von j'.cws hers or: "The mcgatherm hygrophilous forest of the tropics is probably the most an( ient t\ pc of habitat . . . and most recent of all (life-forms) is the annual type" (II. 321). The investigation of Ji:ki"RKV and ToKRKV [144) led to the conclusion that "the origin of the herbaceous type in dicotyledons is from woody or arboreal forms", and S 1 1 x K\\ i:i.L, in his monograph of Chacnaciis [2^g\ Compositae, 33 sp.), states that, "within a family or genus, woody perennial species are more j)rimitive than herbaceous annual s|)ecies '. An evolution in the reversed direction is, however, postulated by HlKl iix.sox [140. I. 4) in the Polycarpicae. He regards the herbaceous Kanunculaccae as the most j)rimitive; a woody structure is, in this order, secondarw The current oj)inion is that the woody Magnoliaceae are among the oldest living angiosperms. The rosette trees. '\\\(i common t\'|)e of this interesting growth-form is char- acterized by a candelabrum like mode of branching, the two or three (rarely more) innovations situated on practically the same level at the base of a terminal in- DERIVATION OF THE FLORA AND FAUNA 361 florescence; in some cases cauliflory is observed. A much more uncommon type is the unbranched stem terminated by a tuft of large leaves (if entire = Grise- BACH's Clavija form, j^j. I. ii), either pollacanthic with lateral inflorescences or hapaxanthic when, after a number of years, a terminal inflorescence ends the life of the individual; this type is the most unusual of all. All kinds have in com- mon the very short internodes and the short-lived leaves, a self-evident condition for the formation of a compact tuft. A reduction of the rosette tree to herbaceous state would result in a compact caudex multiceps or a single basal rosette; an hapaxanthic tree would become a therophyte. These growth forms are by no means restricted to island habitats but are also found in all continents, they belong to many different genera and families and are characteristic of such dicotylous families as Araliaceae, Caricaceae, Theo- phrastaceae etc.; typical examples are found in Compositae (e.g. Espeletia, species of Senecio), Epacridaceae [Dracophyllum], Lecythidaceae (Grias), Meliaceae (Ca- rapa), Rutaceae [Spatkelia], Sapindaceae, and so forth. Among the monocotyle- dons the vast majority of palms belong here, the Pandanaceae should of course be mentioned, and well-known examples are scattered through the Liliiflorae (spe- cies of Aloe, Cordyline, species of Yucca, Dracaena, Fourcroya etc.), Ravenala in the Musaceae, and Piiya Rainiondii in Bromeliaceae. Hapaxanthic trees are few; Krause [162) called attention to the interesting rutaceous Sohnreya excelsa Krause from Amazonas. Thus, even if rosette-trees are more or less widely distributed over the globe in warmer regions, it is a fact that they are a particularly conspicuous feature in island floras. In the Pacific they are very plentiful in the Hawaiian Islands, espe- cially among the Compositae and Lobeliaceae, also the hapaxanthic type repre- sented; they occur in the Galapagos Islands (Grisebach II. 512), we find many in New Caledonia, they constitute a large proportion of the poor Juan Fernandez flora (see Chapter IX), and they are scattered over Oceania. In the Atlantic they are numerous in Macaronesia (e.g. Campa7iula Vidalii on the Azores, MusscJiia on Ma- deira, species of Aeoniuni, Dracaena, EcJiiiwi, Melanoselinum, Sonchus, Sinapide7i- dru7n, etc.). I suppose that some are found also on the islands of the Indian Ocean, but I have no reliable information. I may be entirely wrong, but I have the idea that this is an old-fashioned growth form, a relict element in island floras, and that the high volcanic islands of the Pacific and the Atlantic, which are so like each other in geology and topography, are, in their present shape, of approximately the same age and date back to a period of land submergence and great volcanic activity. Perhaps also the "inland islands", the high volcanic mountains of Africa, deserve to be mentioned in this connection, famous as they are for their magni- ficent tree Lobelias and Senecios. Rosette trees seem to favour open, sunny situations; this is certainly the case in Juan Fernandez, where they are definitely adapted to such habitats. Single specimens of a few species are sometimes found growing in the shade, with the result that the internodes become much longer, the dense tuft dissolves and the formation of flowers is suppressed. ScillMPER believed that they are adapted to a very windy climate and fit to withstand the pressure of strong winds better 362 C. SKOTTSBERG than trees of a more ordinary type, but this theory has been refuted [22^), and recent observations in the field have not made me alter my opinion. Ab:se)ice of quadrupeds. The only mammals regarded as possibly native in isolated islands are bats, but many islands do not have any. Storm-drift has been postulated, but if the distance be very great, the animals might be without food too long unless Nature provides a fair supply of air-born insects. To other land mammals wide stretches o{ open water are an absolute obstacle and their absence is regardeil as one of the safest proofs of the permanent isolation of oceanic is- lands, just as the presence of endemic foxes in the Falkland Islands (now extinct) supports the theory of an earlier connection with South America. However, we cannot know that indigenous mammals never existed on islands, graduall}' having become extinct when the submergence of land had proceeded and grazing grounds became smaller and smaller. It would appear that this theory is contradicted b)' the fact that herbivorous mammals have been introduced by man to many islands and do make a living there also when not tended, but if allowed to naturalize and multi{)ly unrestrainedly in virgin surroundings, their ravages would perhaps prove catastrophal to the native vegetation and, as a con- secjuence thereof, to themselves. They would die out and leave the flora to re- cover. We must remember that the islands are small, most of them very small, that herbixorous animals need space and that pasture lands as they exist now are a result of cleared forest soil and introduction of innumerable alien weeds or cultigens. Another explanation of the absence of mammals is that the islands were cut off so early that mammals had not yet taken possession of the earth or were not universally distributed and perhaps not within reach. But these are mere wild specuhitions. Xo fossil remains have been discovered, nor can they be expected on purely volcanic islands. We cannot attack this problem with a hope of success as long as we know little or nothing of the geograj)hical history of islands and archi|)elagc)es. Hut wc have better remember that continental islands such as New Zealand, New Caledonia, etc. are in the same precarious position with regard to mammals as the Hawaiian Islands. -Xative reptiles and amphibians are also absent. CamI'BKIJ. (./^) mentions in j)assing that tiicre arc half a dozen lizards in Hawaii, but all are species wide- spread in the .South Pacific; most likely they were introduced with the early human immigrants or j)erha])s later. That there are no frogs or toads would indicate, Cami'I'.kij, thinks, that the archi])elago became isolated before the modern kinds of these animals had been developed. lUit, as MlMFORD points out, many con- tinental islands lack all lower vertebrates (/cV-,\ 24eredsamkeit, womit die Abstam- uiung flcr Vegetation ozeanischer Inseln von den Kontinenten vertheidigt zu werden ptlegt, kann die Thatsache nicht verdunkeln, dass in solchen Fallen [the genus Scalesia et( . arc mentioned die Organisationen nicht anzugeben sind, aus deren Variation man sie hervorgegangen \orstellen mochte. Die nahe Verwandtschaft hingegen, welche zwischen vielen endemischen Frzeugnissen des Archipels und denen der amerikanischen Floren unleugbar besteht, kann aus dem Pildungsgesetz der riiumlichen Analogien ebenso wohl, als aus ei?iem genctis( hen /usanmienhang abgeleitet werden. Und warum sollte iiberhaupt das Festland vor den Inseln den \'orzug selbstiindig entstandener Organisationen gehabt haben, deren crste Erzeugung in den friihesten Perioden der Erdgeschichte jeder Moglich- keit einer \'ariation vorausging? waruin sollte sich nicht spiiter und an verschiedenen ( )rten sich wiederholt haben, was urs])riingli( :h moglich war und wovon nur die Bedin- gungen ein noch ungelostes Kiithsel geblieben sind? (II. 512—513). Such ideas have little more than historical interest. Drudk, another leading authority on plant geograph)-, was more in accordance with modern thought: . . . dass die Flora der Inseln nicht nur als Transformationen der jetzt lebenden Kon- tinentaltloren erfasst werden darf, sondern dass auf vielen Inseln unzweifelhaft eine Weiter- DERIVATION OF THE FLORA AND FAUNA 365 entwickelung alter, vielleicht den Charakter einer alteren tertiaren Periode repriisentieren- der Stammfloren stattgefunden hat, welche sich hier im Schiitze der Abgeschiedenheit fern von dem Einfluss kontinentaler Umwalzungserscheiniingen sich erhalten konnten [jos. 128). Drude agreed with Hooker that extinct species, known as fossils in Europe, were the ancestors of species now endemic in Macaronesia. With regard to the Pacific islands several authors have discussed the question to what extent endemism has been of a progressive kind and if not only species but also isolated genera have evolved on the islands from a limited number of un- known ancestors. Most of them, in spite of their firm belief that the islands are, geologi- cally spoken, very recent, regard the insular biota as a local product; the first arrivals had had time to give rise to new genera, those that came a little later became new species, still more recent ones varieties of a continental species, and such as arrived in our era have not had time to change but are expected to do so, because isolation in a new and strange environment makes them adapt themselves by changing their genetic structure. Hayek (jo^. 252) expressed this very clearly: Dass diese eingewanderten Elemente infolge ihrer Isolierung eigene Entwickelungs- richtungen einschliigen, die sich in einer oft auftallend grossen Zahl von Endemismen aussern, ist ja selbstverstandlich. Among botanists of the latest decennia Andrews (6) may be chosen as a representative of the school of "rapid adaptive radiation". If a newcomer belongs to a primary form of a virile genus such as Acacia, Coprosma, etc. it may be expected rapidly to become differentiated into varieties and species. As Bentham pointed out long ago, the geographic station of a waif or colonist imposes variations upon it almost from the moment of its arrival. Eucalypts planted in New Zealand, California, and other places present marked differences from the forms the same species possess in Australia. In the second ])lace, if the plant assemblage into which the waif or colonist arrives be a result of long-continued struggle for existence such as occurs commonly in Holarctica and the cosmopolitan tropics, then the opportunity for the development of new forms is remote, unless the new arrival itself is a plastic form, and a grand example of the survival of the fittest. If, on the other hand, the newcomer belongs to an agressive species in its own continental setting, then it has, all other things being equal, an excellent chance of survival and of differentiating into new forms (p. 617). Andrews, himself an Australian, speaking on the Hawaiian flora, mentioned Acacia, Coprosma and other genera of Australian or southwest Pacific origin which developed new, endemic forms in Hawaii, where they take a prominent part in the vegetation. But when it comes to Eucalypts planted in California, nobody will, I suppose, consider this example as a proof of the origin of new taxa under the influence of new surroundings. There is, he continues, a difference in the physical character of islands; some did not encourage the colonist to vary and evolve, others did. If the island ... be very small and of negligible relief, it again has but little op- portunity for differentiation, and, furthermore, if the island be even large and high, but the time be short . . . the response will have been but slight (i.e.). 366 SKOTTSBERG If, on the other hand, the island group lies in the tropics and is large in area, the islands of the group being close together, if the vertical relief be very great indeed . . . if the {)recijMtation be very variable ... if the soil be rich, but variable in porosity; if the j)lant assemblages into which the waifs or colonists entered are not the end result of severe plant competition; then the stage is set for the rapid differentiation of primary types of agressive genera. 'i'he evidence available suggests that these genera never existed as such on the land from where their immediate predecessors were derived, but that virile types, of the families concerned, arrived as waifs from Malaysia, Australia, New^ Zealand, Central and North America. Once they found themselves removed from their former severe competition with other plants, they gave rise to the vigorous, endemic Hawaiian genera (p. 6i8). What .Vndki-.ws depicts is an island which has reached maturity. It has risen from the bottom of the ocean to an altitude where the moisture of the trade winds is condensed, where there are leeward and windward slopes, different habitats and a rich soil. This island looks back on a very long history, but life begins to arrive long before the island has come to rest. Aerial plankton will bring microscopic green and bluegreen algae, bacteria, spores of all kinds and also air-borne seeds, and some drift may be washed up on the shores. But to begin with there is little inimidit)' and hardly anything we can call soil, we need water before even the most j)rimitive organisms can exist, so that the island becomes fit to receive its first settlers, microscopic algae, then mosses and lichens and mycelia of fungi, to form soil where the first seeds can germinate and start to form an incipient vege- tation cover and an abode for a soil fauna and flora. Lava cracks in cooling, some water ma)' stand in the fissures which form the starting point for further develop- ment, a spectacle we have before our eyes where streams of lava are still formed. Of higher plants, ferns are likely to be among the first to get established — GUPPY even spoke of the "era of ferns"; a halophytic Asplcniuui is the only living thing observed on the far-fiung reef Sala y Gomez between Easter Island and South America. An island in a comparatively recent stage, where there is plenty space for new settlers, let it be that little comfort is as yet offered, would, we should think, otter good opportunities for "virile and agressive" immigrants to get a foothold and to become the ancestors of the most ancient element in the flora, but this is not what Andki.ws sa\s. In order to start an evolution of new species and genera a ver\' great vertical relief, a variable preci[)itation and a rich soil are the conditions, l3ut in oriler t(; get a rich soil cover we must have a closed vegetation cover, also forest. Wa I.I.AC!; [2yS. 295), sj)eaking of St. Helena, very rightly said that "no soil could be retained unless protected by the vegetation to which it in great part owed its origin", and the same is true everywhere. If the change in environment, the new living conditions, are the cause of variation, why did they not act until the island was already more or less stocked with plants.- Were there no virile species among the earliest immigrants which took possession of the land and formed the oldest element of the flora.' Uas it lost among the later arrived aggressive newcomers.- It is calculated that about 90% of the Hawaiian angiosperms are en- demic, most of them belonging to endemic genera or to species very different from their continental congeners. \)o some of the most remarkable monotypical genera DERIVATION OF THE FLORA AND FAUNA 367 represent true relics, descendants from the earliest settlers, whereas the other are examples of progressive endemism? This is often said, but it is not, as far as I can see, what Andrews means. In the opinion of zoologists still infected with Lamarckian ideas environment is the direct cause of new hereditary characters. Time after time we are told that as soon as a "germ" happens to land on an island, it gives rise to something new. However, this does not, Mayr says, imply that every little island is turned into a centre of evolution {lyg. 216). The small and usually rather isolated islands of Polynesia have not only not been new centres of evolution, like the Galapagos or Hawaiian Islands, but, on the contrary, there is good evidence that many of them are "traps". Species that reach these islands are doomed to extinction. This is peculiar. There are many small islands stocked with both non-endemic and endemic species, many of the latter stenotopic, it is true, but quite able to hold their own as long the environment remains unchanged. It is a truism that, if a newcomer lands on an island where, for climatic or other reasons it cannot live, it is doomed to disappear pretty quickly, but there is no reason why, once established and able to reproduce itself, it would become extinct as long as the habitat does not undergo any change for the worse. It is generally known among ornithologists, Mayr says, that island birds are very vulnerable. He continues: The recent considerations of Sewall Wright have given us a possible key to this curi- ous phenomenon. Apparently in these isolated populations there is more gene loss than gene mutation. The species are therefore adjusted to an exceedingly narrow limit of environmental conditions. They are unable to respond to any major change of conditions and must die if such a change occurs, or are crowded out if competitors arrive. We cannot be sure that mutations, should they be induced, would do them any good and we need not assume a "gene loss" to explain why organisms unable to escape to a more favourable habitat are bound to become exterminated as a result of "major changes of conditions". Zimmerman {2^8) is one of the prominent defenders of the idea that, in oceanic islands, as exemplified by the Hawaiian chain, a small number of immigrants has given rise to a comparatively rich fauna and flora; the proportion between genera and species attains figures expected under continental conditions but certainly not in oceanic islands. He admits that everything did not necessarily happen on the present islands as we behold them, for differentiation may have begun on some distant land and proceeded in the course of migration, with Hawaii as the terminus. He picked out, as an example, a large curculionid genus ranging from Australia to Micronesia and east to Marquesas; most islands or archipelagoes have their own endemic species, which have developed on their respective islands. He thinks that, if within a varying population, one pregnant female, not carrying the gene consti- tution of the entire population, gets isolated on an oceanic island, she may stand out as distinct from the average — and if this sequence of events "be accompanied by conditions conducive to isolation and survival, rapid and diversified speciation may follow" (p. 125). This process is repeated on island after island, and "the 368 C. SKOTTSBERG intensity of divergence will be increased". It is not clear to me why the intensity would be increased. Now, if sufficient time has elapsed and the original sources lia\e been eroded down and perhaps become sterile atolls, their faunas will have been exterminated and on the newer islands segregates without obvious ancestral relation will be left. This development explains why Hawaii has so many isolated endemic t\'pes. The living world in Hawaii is older than the rock — in a way, he sa)-s. ( )uite true, but it is not true that this possibility has, as he says, been entirely o\erl()oke(l in previous discussions; I think that 1 have, on repeated occasions, exj)ressed myself ver\' clearly on this point, even if 1 do not agree with ZIMMERMAN when it comes to exj)lain why and how it happened. The biota as we know it today is in part the ultimate product of a progressional devclopnient which has moved and evolved along great insular archipelagos over periods of time much longer than the ages rec^uired for the development of the main Hawaiian Islands and their contemporary biota. Various genera and stem forms of groups of species iiKu have evolved in islands — now atolls such as some of the leeward Hawaiian chain, the great Micronesian archi])elagos, the Line Islands — which form the approach to Hawaii. However, some of the genera and the bulk of species known today have originated on our present main islands (p. 125) ... in contemporary Hawaii there are preserved rem- nants of a biota which has in ])art develoj)ed by unique methods and in which are preserved forms whicli are the end ])roducts of species chains that carry back, through archi- ])elagos now worn away, to geological ages indeterminant (p. 126). In few words, we have to do with relict as well as progressive endemism — nobody objects to that. A genus may be an ancient relic, while the actual species are the result of more recent, progressive differentiation. On the other hand, there is no reason why not a species could be immensely old without having undergone any j)crceptible change. Zimmi:rm.\x does not hesitate to conjure up all the sunk archipelagos he needs, if only land connections arc left out of the discussion. Once more he describes his vision on p. 127 which I shall permit myself to quote, with the obvious risk of tiring out the reader. I believe that the great atoll chains of the Pacific may hold some of the now hidden (lues to the stories of the magnificent biological development of Polynesia. Many of the peculiar endemic groups of the Hawaiian and southeastern Polynesian islands owe their existence, if not their very origin, to ancient high islands of the one-time splendid archi- l)clagos marked by (lusters or coral reefs. Surviving lines of middle Tertiary and of |)erhaj)s even older continental faunas nuiy have had their germ plasm filtered down through su( cessivcly changing generations which have passed successfully through island maturity and degradation to atoll formation and have carried over to new high islands in different archipelagos. Thus, some supposedly old types such as certain land molluscs could have maintained themselves (but evolving) in insular isolation through long periods ot time while tlieir continental j)rogenitors became extinct or restricted under continental conditions. I cannot think of what kind of higher organisms would have passed success- fully through island degradation down to atoll stage; they must have left for new high islands long before their abode became uninhabitable. V.ven if, as ZIMMERMAN thinks, much of the evolution took place during migration from island to island, DERIVATION OF THE FLORA AND FAUNA 369 specific segregation was mainly effected after arriving at the final station, in this case Hawaii. This happened yesterday or the day before, geologically spoken: "the rate of erosion is such that these main islands could not have stood here as they are longer than from a period late in the Tertiary" (p. 121). "Explosive speciation" set in during late Pliocene and must have increased during "the great Pleistocene erosion which has left such a spectacular and rugged topography in its wake. New land open to colonization is conducive to speciation" (p. 122). This late and rapid differentiation is illustrated by ZIMMERMAN for the land snails pp. 98-101 : Helicidae. 59 species developed from one, or possibly two original immigrant stocks. Pupillidae. Possibly 4 ancestral species gave rise to the 86 Hawaiian forms. Cochliocopidae-Cochliocopinae. One immigrant of Cochliocopa stock could have given rise to the 142 forms. C.-Amastrinae. 294 forms apparently developed from one basic stock. Tornatellinidae-Tornatellininae. 117 forms derived from 4 or fewer ancestral forms. T.-Achaiinellinae. It appears certain that this subfamily had its origin and development in the Hawaiian area and all of the 215 forms may have been derived from a common tornatellinid ancestor. FOSBERG, in a chapter contributed to Zimmerman's book, tried to fix the number of ancestors of the Hawaiian angiosperms. His method is quite simple: if the species, few or many, of a certain genus present the appearance of a more or less homogeneous group, only a single ancestor is made responsible for the segregation; if subgenera or sections are distinguished, we must count with the same number of ancestors as of taxonomic groups within the genus. The method seems a little too easy; possibly we are confronted with a rather complicated question, the solution of which I am not going to attempt. Setchell (^218) uttered some sensible words on migration and endemism: Where endemism of the degree of ordinal or family endemism occurs on oceanic islands, we may feel strongly inclined to believe that evolution of such degree took place on the continental area which was the source of the original migration and not on the island where now found, the original becoming later extinct, leaving the migrant as an endemic. The same is true of generic endemism or even specific endemism of a strong type, that is when representing an isolated or aberrant species under the genus (p. 874). . . . To assume that insular conditions originate new forms is to overlook what has taken place on continents (p. 875). This is, however, what so many authors do. They claim that oceanic islands follow their special laws, that a plant or an animal which happens to land far away, will, as it were, lose its balance; hidden factors, repressed as long as they lived on their fatherland under "severe competition", are set free and allow them to develop their inherent possibilities, they are not subjected to any struggle for existence in their new environment. Says ZIMMERMAN: "The environment, of course, plays an all-important part in the development of species ... it is generally agreed that profound changes have been effected on organisms by environment" (p. 187). 24 - 557857 The Nat. Hist, of Juan Fernandez and Easter Isl. Vol. I 370 SKOTTSBERG This is pure Lamarckian langua^^^c. \Vc may say, of course, that change of environ- ment may cause mutations, but our experience tends to prove that only a fraction of I °o are \aluable, the remainder, if not deleterious, at least indifferent. Mutations, gene losses. Inbridization, polyploidy and so forth, all may have their share, but they are only ripples on the surface. Xo theory has been able to penetrate to the nucleus of the j)r()blem. Anyhow we have no reason to think that, in its pro- duction of species and genera, families and orders. Nature has followed other lines in islands than in continents. A prominent Swedish geneticist, after a life-time's s|)eculati()n on the causes generally accepted as responsible for the "origin of species", rejected all of them; unfortunately he threw the egg away with the shell and con- vinced himself that there had been no evolution at all. He did not, however, revert to an omnipotent creator but invented a new and entirely revolutionary theory; if there ever was a stillborn one, it was this (iS'6). I guess we can take it for granted that no peculiar, outstanding types were created on young volcanic islands, whatever their present faunistic and floristic status may be like. Their basic stock is much older than the rocks and inseparably connected with the great continental faunas and floras. Wallace, however, made a distinction between island and continental history. He emphasized that evolution has recjuired an enormously long time to produce the present status, that from the Cretaceous until now nothing of a revolutionary character had happened, families, genera antl in cases even species still living date from early Tertiary at least — but on islands the great period of creation was repeated during the last epochs. l^'inally I shall quote some selected passages from Stebhixs' book on variation and evolution Uiy). The ditterentiation of orders and families of flowering plants through the action of natural selection under ])resent conditions is well-nigh impossible. . . . All the trends leading to the difterentiation of families of flowering plants probably took place simul- taneously and at a relatively early stage of angiosperm evolution. For instance, both distributional and ])aleontological evidence indicates that the Com])Ositae, the most highly sj)e( iali/.ed family of dicotyledons, already existed in the latter part of the Cretaceous ])eriol)rcitung iiber Land angenommen werden. Von anderen Landtieren findet sich auf Juan Fernandez v.. W. noch der 'J'auscndfiissler Gecphilus latkollis. Endlich sind sie auch von cinem ( )ligo( haeten erreicht worden, der Ocnerodrilinen Kerria saltensis, ^\^ dMQ\\ in Chile vorkommt. Hier wiire eine jiingere, iiberseeische Einwanderung denkbar. Aber Kc) ) id ist au< h alt genug, dass sie auf dem Landwege nach Juan Fernandez gekommen scin kann. Auch die Flora der Inseln zeigt interessante Beziehungen (])p. 324-325). As could be expected, Arldt's south Pacific bridge was supported mainly by tile land molluscs, and they made him extend his bridge across the equator to ilawaii. There is in the Juan P'crnandez flora some perplexing affinities with Hawaii that scctn to defy all attempts to an explanation. The occurrence of two (not only one) endemic species of XcsogeopJiilus (formerly a subgenus of 6'r^////7//'j) is in- teresting, Kcrria is, j)erhaps, less important. Other striking cases of disjunction were already referred to in the lists of evertebrates above. It is a pity that Arldt over- looked the leech Xi'sophilaoiioji, but also from what he says about the butterflies it is evident that the numerous zoological papers forming vol. II of "The Natural History of Juan F'ernandez and blaster Island" had escaped his notice. The leech would, I sup|)ose, have furnished him with one of the most eloquent evidences of former land connections. Xor did he know the botanical volume of this work, DERIVATION OF THE FLORA AND FAUNA 377 for he gives a summary of my 19 14 paper (22^). He agrees on the great age of the "Palaeopacilic element"; second comes the small neotropical, third the large Chilean group: "hier liegt also wohl zum grossen Teil spate, iiberseeische Einwanderung vor". This was, he found, even more true of the subantarctic element, "das iiberhaupt keine endemische Arten aufzuweisen hat" — very few were known then, but later some endemic species were discovered. The question to which extent Juan Fernandez presents the biological pecu- liarities regarded as characteristic of oceanic islands — see Chapter VII — will now be answered. We have seen that endemism is very high among the phanerogams and that the various kinds distinguished by HoOKER are represented; we have even a primitive endemic family, Lactoridaceae, we have a proportionately large number of peculiar genera, some of them quite isolated, particularly among the Compositae {Centaur odendro7i, Dendroseris, Hesperosei'is, Phoenicoseris, Rea, Rhe- tinodendron^ Robinsonia, SympJiyochaeta, Yunqueci) but also in other families, Cuniiiiia (Labiatae), Juania (Palmae), Megalacfme and Podophorns (Gramineae); further there are a few endemic genera closely related to South American ones, Nothomyrcia (Myrtaceae), OcJiagavia (Bromeliaceae), and Selkirkia (Boraginaceae). As among the genera, so we find species of a strong character in Chenopodium, Coprosma, Eryngium, EupJirasia, Fagara, Peperomia, Plaiitago, Sanialum, Urtka, Wahleiibergia, well-marked but not very aberrant species in Berberis, Boehmeria, Car ex, Cladium, Colletia, Erjgeron, Escallonia, Gunner a, Halorrhagis, Hespero- greigia, Ranunculus, R/iaphitkaninus, Solanum and Ugni, and many not very different from their continental congeners in Abrotanella, Acaena, Apium, Azara, Cardami7te, Cliusquea, Drimys, Dysopsis, Galium, Luzula, Margyricarpus, Myrc- eugenia, Perneitya, Pkrygilanthus, Sopkora, Spergularia and Uncinia. Species undoubtedly native but also found elsewhere do not number more than 46, and in several cases their citizenship is open to question. Endemism among the ferns is not so high, but there is one very aberrant genus [Thyrsopteris] and several peculiar species. Bryophytes and lichens will not be considered; they were not included in HoOKER's paper and I have not had occasion to compare them with other island floras. There are no conifers, a single leguminous genus [Sopkora] with two species, and no orchids. The proportion genus : species is i : 1.65. Mammals, batrachians, reptiles and fresh-water fishes are absent. The earth-worms are supposed to be adventitious with the possible exception of Kerria saltensis. In all these respects Juan Fernandez agrees with the character attributed to oceanic islands. The high percentage of woody plants was emphasized by SiNNOTT and Bailey, but as the literature on which they based their figures is quite out of date, a new table was prepared. The object of Sinnott and Bailey was, as we have seen above, to show that the woody plants increase in number with the rising degree of endemism, and that the endemic genera were trees or shrubs and formed a more ancient element than the herbaceous plants which during former epochs were few in 378 C. SKOTTSBERG Table VL Percentage of \v()ocl\- and licrbaceous species in Juan Fernandez. Indi.utMious species .... Not eiitleniic ICndeinie, ^enus not eiuleniic . . 71 Als;) tlie <'enus endemic . . Total \N'()()( ly% Herbaceous % 147 68 46.3 79 1 53-7 46 5 10.9 41* 89.1 71 37 52.1 34* 47-9 30 27 90.0 3 10. 0 Four species suffruticose. comparison, and the Juan hVrnandez Islands were regarded as supporting their h>-|)othesis; this, to judge from tlie table, they certainly do. The five woody non-endemic species are liiiipctnDu nibnun and Saliconiia fruticosa (low, erect shrubs), and Myrtcola unnimularia, Riibiis gcoidcs and Calystegia tuguriorimi (trailing). The rosette tree form is observed in 16 genera with together 31 species, belonging to si.x families, Boraginaceae, Bromeliaceae, Chenopodiaceae, Compositae, Plantaginaceae and Umbelliferae; the palm Juania is of course excluded. As I have paid special attention to them in another paper (-^J/), where they were well illustrated, no more will be said here. Annual and biennial herbs, the therophytes of Raunklkr, are not completely lacking, but they are ver\' few: Cardamiiie chenopodiifolia, C/iae/otropis (2 species), Par'utaria, Plauiaoo tyuncata, I'etrao^oma and Urtica Masafuerae. In 2^1 I listed Cliartntropis among the hemicryj)tophytes; they give the impression of lasting more than one year antl are found green at all seasons, and the same may be true of Parictaria. I am not at all sure that PUxutago iruncala is native. The Caydanu}ic has been seen twice, last time in 1872, the Crtica not since 1854, when it was discoxercd. It may be that it is an ephemerous plant and disappears in early spring, a season when very few botanists have visited the islands. Thus, the "oceanic peculiarities" are all there, but possibly some of them can be exj)lained otherwise. iMulemism of a very high degree within a small area is no monopoly of isolated islands; it will be sufficient to mention the Cape fiorii or southwestern Australia. The large proj)ortion of woody plants can be understood if the islands became isolated before the myriads of herbs, particularly the annuals of Central Chile, had evolved. (iRlSKr,.\(ll's "Clavija" and related life forms are not confuied to oceanic islands; if my interpretation of this mor- phological t\'pe as an evidence of anti(juity is correct, we can understand why it takes such a pronuncnt part in old island fioras. The reason why therophytes are almost wanting is not climatic. This is amply |)roved by the innumerable anniuil weeds introduced with the traffic and thriving only too well. The climate is of a modified Mediterranean tyj)e, and from a purely climatic viewpoint we should ex{)ect a large percentage of native annuals and biennials; this question was discussed at some length in ^-r/. 827-830. Chile has hundreds of endemic DERIVATION OF THE FLORA AND FAUNA 379 therophytes, but they have not spread to the islands although many have special dispersal mechanisms, and they were not available at the time when the supposed connection with the mainland existed before the final uplift of the Andes. The existence of many aberrant genera and species and, above all, the marked difference between the phanerogams of Masatierra and Masafuera shows that there is no exchange between the two islands in spite of the very moderate distance; even the 360 miles separating Masatierra from Chile should, in the eyes of the diffusionists, amount to little if sufficient time be granted. The Mar- quesas Islands, situated much farther away from the continent, have a less peculiar angiospermous flora than Juan Fernandez; the very opposite ought to be expected if overseas migration had played the dominant role. The absence of the flora of Central Chile speaks against the efficiency of the natural dispersal agents. These circumstances are in favour of the opinion that the volcanic islands arose, not from the depths of the ocean, but on a piece of land formerly connected with South America and not sunk until the newborn islands, now reduced to ruins, had become a refuge for the ancient continental fauna and flora. It is also true that several large and widespread families are lacking, such as leguminous plants (with one exception) and lilies, well developed in Chile, but they belong to the modern Chilean flora. The plant world of oceanic islands is described as a haphazard collection of waifs and strays, and this is said to explain why so few genera contain more than a couple of species. But would not the result be the same if the actual islands originated through volcanic activity on a sinking land.? Chance would decide what took possession of the new soil, and different sets find a refuge on Masatierra and Masafuera. We know that of the mammals introduced by man the goat thrives and multiplies since 400 years and quickly became naturalized. There were no goats on the mainland when the land-bridge existed, but there may of course have been some primitive mammals; if any of them reached the islands, they have disappeared long ago. The islands, as we see them, appear never to have offered great possibilities for the subsistence of a mammalian fauna. They are very small and as there was litde open land there cannot have been any grazing grounds worth mentioning before man altered the landscape. Even after the ground has been cleared on all the lower slopes, pasture is miserable, and one valley after the other has been turned into a desert by the ravages of sheep and cattle. If left to run wild and multiply, the final result can be foreseen. Carnivorous animals need a prey, and there was none. The same is true of snakes which should thrive well now since the domestic rats and mice have been introduced. Chile's mammalian fauna is poor and nobody knows if the huemul, the pudu, the Chilean rodents and small marsupials would be able to make a living on Juan Fernandez. Reptiles are poorly represented in Chile, there is not a single tortoise, very few snakes and a dozen lizards. Amphibians are few, but toads and frogs occur. Among the invertebrates are several orders, the chances of which successfufly to get transported across the sea are very doubtful or, as far as we can see, none at all. 38o C. SKOTTSBERG The insular peculiarities as displayed in Juan Fernandez (and in other similar cases) do not, I think, permit us to take a definite position against the hypo- thesis of former land connections. Chapter X The Chilean coast line and the history of the Andes. I^^rom a look at the map we easily get the impression that the trend of the South American west coast is a j)roduct of the rise of the Andes, because this enormous uplift must have been compensated by the submergence of old border lands and by the formation of a deep trench, and that these movements, which certainly were of very great magnitude, may have extended its effects west as far as to the region where we find the Juan Fernandez and Desventuradas Islands forming the ex{)osed summits of a submarine ridge. I shall call this ridge the Chaigneau Ridge after the Chilean navy officer who was the first to survey it. flic Cliaigticau Ridge. — The two archipelagoes lie within the 2000 m line, San I'elix and San Ambrosio on a plateau rising above the 400 m curve and extending a long way toward Juan Fernandez, as seen from CllAlGNEAU's table ( -T^) which is reproduced here with the soundings rearranged from N. to S. accord- ing to latitude and with the addition of some figures from the latest chart. The Meniivn Ridge. — 160 miles NW. of San Felix-San Ambrosio another ridge, called the Mcrriam Ridge, was discovered during the U.S. "Carnegie" campaign 1928-29 (//). It extends between 25° 3'. 2 S., 82° 20' W., and 24° 54' S., "^i" 13' \\'.; the depths found were 1445 and 1260 m, respectively. The bank rises 3000 m above the bottom and is only 10 miles wide. Along the most elevated part i 1 86, 1188 and 1168 m were found. From the latter spot a series of soundings was taken SIv of the ridge, showing the rapid increase of the depth: 3 miles 1260 m, 9 miles 2751 m, 20 miles 3620 m, and 32 miles 4115 m;the Merriam ridge is separated from the Chaigneau ridge by deep water. Toward W'XW. the sloj)e is more gradual until a depth of 3000 m is reached. North of the Merriam ridge, in 21^40' S., 8i°4o' W., approximately, a sudden rise, bounded by the 2000 m curve and surrounded by deep water, has been discovered (see map). Mere the bottom rises to 972 m below the surface. I do not know if this remarkable j)lace has a name. The Caniegie Kidfi^e. — During the cruise of the "Carnegie" two soundings about 100 miles (jff the coast of Fcuador, lat. l°32' S., long. 82° 16', gave I 5 1 5 and 1454 m, respective!)-, indicating a rise of 1800 m above the bottom, but before the entire distance along the coast has been surveyed we do not know if a series of ridges, jirobably much less well-marked than the Chaigneau-Merriam ridge, can be traced all the way between lat. 35° and the equator. Between these ridges and the coast is the deep trench (greatest depth 7635 m), and south of the latitude of Juan Fernandez depths exceeding 5000 m are still found, but farther south the 4000 m curve is soon reached and the trench disappears. Older maps, two of which were reproduced in 22y. 44, 45, show DERIVATION OF THE FLORA AND FAUNA 381 Table VII . The Chaigneau Ridge. r = rock, s = sand. Depth in m Bottom S. lat. W. long. Remarks 38 . — . 26° 14' 79° 56' 14 km NE. San Felix 80 — 26° 15' 80° 4'.5 4.5 km N. San Felix 69 — 26° i7'.5 80° 8'.5 4 km W. San Felix 179 — 26° i8'.5 80° 8' 5 km W. San Felix 220 r 26° i8'.5 79° 57' Between San Felix and San Ambrosio 400 s 26° i8'.5 79° 49'.3 5 km NE. San Ambrosio 250 s 26° 19' 79° 52' 3 km N. San Ambrosio 150 s 26° 19' 79° 51' 3 km N. San Ambrosio 215 s 26° 19' 79° 49'.2 4.5 km NE. San Ambrosio 182 — 26° 25' 79° 56'.5 6 km S. San Ambrosio 148 — 26° 26' 79° 52' 8 km S. San Ambrosio 105 s 26° 30' 79° 49' 15.5 km S. San Ambrosio 400 26° 36'.5 80° 8'.5 33 km S. San Felix 550 27° 49'.5 80° 14' c. 165 km S. San Felix 675 28° 3'.5 80° 16' c. 200 km S. San Felix 660 28°33'.5 80° 11' c. 260 km S. San Felix 660 29° 14' 80° 15' c. 340 km S. San Felix 1300 30° 49' 80° 24'.3 1430 31° 26' 80° 10' 1800 — 32° 45'.5 80° 18' Between Masatierra and Masafuera, about i ° N. 1800 — 34° 34' 80° 36' About 100 km SSE. Masafuera Juan Fernandez situated on a "lobe" with somewhat shallower water, less than 3000 m deep, extending NW. from the coast and suggesting a possible former connection, but recent charts are less unambiguous. The 2000 m curve makes, however, a bulge around lat. 38°, where a depth of only 1238 m is indicated while deeper water, 1400-1500 m, is met with near the coast. Soundings be- tween the islands and the line where the trough stops are too few to be of much value, but in all probability the connection, if it did exist, should be looked for farther south. This is, as we shall see, also the opinion of the geologists. They do not hesitate to regard the banks just described as an extension from the continent. Nobody will, I suppose, argue that every marked rise of the ocean floor is a sign of sunken land; the majority of oceanographers prefer to call them independent products of volcanic action. Some, perhaps most of them, never reached the surface, some have done so, but were broken down, but many are still above the water and form the Pacific islands. But even the advocates of the permanence of this largest of basins admit that its margins are zones of considerable disturbance. If it can be proved, or at least be made probable, that 382 <-"• SKOTTSBERG the deep trench following the trend of the Andes, is a consequence of the gigan- tic mountain-builchng processes, tlie submarine ridges west of the trench can- not hel|> to get iniphcated. Tiiis idea is by no means new, it has been expressed by many: "the width of Soutli America may well be a good deal less now than before the Andes were uplifted", as G()(JI) says (lop. 349), but some of these writers did not tr\' to penetrate the complicated geographical-geological history of til is region of great tectonic disturbance. This is true of myself, when I tried to describe what I imagined having occurred (22/. 43), but it does not apply to Ikms(IIi:k (//i') who took pains to inform himself of the history of the Andes as told b\' geologists. That they arose in a geosyncline is proved by the Jurassic and Cretaceous beds now elevated thousands of meters and covering the older eruptives. To the east of the depression land had existed since the Permian, to the west was a Tacihc land mass of hypothetical width; one opinion regarded the Coast Range as belonging to this land. Irmscher, who took \\T:c;ener's sitle. did not ask for any large-scale subsidence correlated with the uplift of the Andes, because the resistance of the sialic crust to the westward drift of South America was sufticient to account for uplift and folding. Consequently, he was unwilling to accept Pexck's intrusion theory; if, in the future, it should appear essential to accept a land mass, it could be nothing more than a narrow strip which, perha{)s, had been connected with California (p. 45). l'i.N( K stellte (lie Aiisbildiint( dor ozeanischen Tiefen am Rande des sudamerika- nischen Kontincntcs, also das \'ersinken angrenzender 'j'eile des Pazifiks, der aufwarts bewegten andinen Scholle gegeniiber und schliesst, dass die niichstliegende Erklarung liir die \'olumenanderungen unter der festen Kruste in dort stattfindenden Massenver- schicbungen zii suchcn ist. Wenn Massen aus der pazifischen Region in die andine iihertrcten, so nuiss in crstcrer die Kruste nachsinken (p. 51). The cause of the uplift and folding was, according to Penck, a result of the intrusion of the andesitic magma, which lifted the mountains but, Irmscher remarks, this could not be the only source of the tangential pressure: i)ie ine( hanis( hen Ursachcn des l-altenvorganges sind zweifellos anderer Natur, und (lie .Magniaintnision ist genau so eine Uirkung derselben wie die P'allung. Denn es steht test, dass die ( ichirgshildung niit dem Kmpordringen des Magmas synchron ist und die Intrusion somit gleic hzcitig niit der angcnommenen annuihlichen Lostrennung Siidamerikas von .Atrika. Where does the (■iiaigneau-Merriam bank come in.- \Ve(;ener left it unexplained, according to 1)1 Tori it was an "advance foltl". l^EKRV and SiN(;i:w.\i,i) (j^), in their review of the tectonic history of South America, describe the develoj)ment in the following terms. Some students regard the Cordillera de la Costa in Chile as remnants of an an- ( ient massif, the bulk of \\hi( h has been downfaultcd beneath the waters of the Pa- cific. It (onsists of (rystalline ro( ks both igneous and metamorphic and these have (onimonly been assumed to be of great age — even Archean. This inference of antiquity rests, not upon their known rehitions, but upon the geosynclinal nature of the Andean seas which are clearly ej)icontinental and not shelf seas. DERIVATION OF THE FLORA AND FAUNA 383 The sediments in this geosynchne are compressed between the Brazihan massif and another massif in the west, the vestiges of which should be looked for in the coast range. Anyhow, "that there was land to the west of the Western Andes cannot be doubted". These authors date the Concepcion-Arauco series to the older Miocene and they do not. regard the flora as a coast flora. Let us now turn to Bruggen, author of a modern handbook on the geology of Chile (jji). To begin with I shall allow myself to quote Florin's summary (^5. 4-6) of Bruggen's earlier writings. Until Middle Tertiary times the great Andes and the Coastal Range were ... a continuous upfolded mountain chain subjected to powerful denudation. . . . The principal uplift of the Andes in Chile occurred in the Middle Cretaceous, and altered the palaeogeographical features of the Andean region considerably. The old geosynclinal had been turned into a continental area, at the western verge of which the border of the Pacific Ocean at the end of the Cretaceous occupied approximately the same line as to-day. Marine deposits of Danian ... as well as Palaeocene age are lacking, and at the beginning of the Tertiary period the continent probably extended further to the west. In the Eocene and Oligocene, respectively, subsidences took place and the ocean encroached more and more on the land. The Concepci6n-Arauco coal measures are coastal deposits, which have been called the Concepcion Series by Bruggen. This series, about 400 m. thick, rests unconformably on marine strata of Upper Cretaceous (Senonian) age. The shales containing fossil plants occur in conjunction with intercalated coal seems. . . . The base of this section is of marine origin, and in addition marine layers are intercalated here and there in its middle part, which is otherwise generally built up of freshwater deposits. According to Bruggen the Concepci6n Series is overlain by the deposits of the marine Navidad Series, which is upper Oligocene or Lower Miocene in age. Bruggen came to the conclusion that the Concepci(3n Series belongs to the Eocene, basing this on stratigraphical as well as on zoo-palaeontological evidence. The sediments of this series were according to him deposited on a broad, slowly sinking coastal plain, and subsequently subjected to considerable tilting and faulting, probably in the Miocene. Berry regarded both the Concepcion and the Navidad series as belonging to the Lower Miocene or possibly Upper Oligocene, but according to Bruggen this dating holds good for the latter only, while the former is much older, and whereas BrCggen thinks that the Eocene coal flora was deposited in extensive coastal swamps, Berry regarded it as neither limnic nor littoral, but inhabiting a lowland area away from the coast. FLORIN found that the plant remains were laid down in the vicinity of the sea and that they are too well preserved to have been transported any great distance. If it is true that the coast-line, at the end of the Cretaceous, occupied the same position as to-day, one is inclined to believe that the palaeogeography was different before the great uplift occurred in the Middle Cretaceous, when a large scale subsidence ought to have taken place. The oscillations along the coast of central Chile during later times could hardly have involved the area where the submarine ridges are found, so that the possibility to link them to that part of the continent as late as that is small. Consequently attention has been directed farther south, as already suggested. West Patagonia is a region of considerable and late subsidence; the longitudinal valley of central Chile 284 C. SKOTTSBERG disappeared under water to form the long series of the Patagonian channels, the Andean valle\-s became fiords and the broad dissected fringe of islands and skerries also give evidence to what has happened. The weight of the inland ice during the perioils of glaciation must, however, also be taken into account. The development of the coastal region, as told by BrCggen 1950, is ex- })lained by facts which, if they have been correctly interpreted, open wide per- spectives to the biologist, even if serious difficulties still have to be overcome. HkiciCKN begins by stating that "el mar del Eoceno" ended somewhere in the latitude of Arauco (38°), because a continental mass, "la Tierra de Juan Fernandez" still existed (p. 50), and pp. 56-59 he relates the history of this land. North of Rio Maulh'n (about 42°) is a zone of dislocations foreign to the struc- ture of the Andes, and this zone coincides with the direction of a broad sub- marine ridge which branches off from the continent; on this ridge are situated the Juan PY^nandez Islands and, farther north, San Ambrosio and San Felix. Taking the 2000 m cur\e as a boundary, the ridge extends south to the Magellan Straits; we observe e.g. in the island Diego de Almagro the same northwest direction that we tind in the Tertiary deposits of Parga and other places in the zone north of Rio Maulh'n. To this must be remarked that the 2000 m line sur- rounds the Chaigneau ridge and that in order to unite it with West Patagonia the 3000 m curve has to be used. This is also seen from I^rCggen's map, prob- ably copietl from Si pax. The absence of marine sediments of Eocene age shows (p. 59) that the Juan Fernandez land was, at that time, united with the continent, but that, during the Oligocene, subsidence set in is evident from the extension of the marine Xavidad series south to 45°, and this was, as we have heard, referred to Upper Oligocene or Power Miocene. Also after the separation the Juan Fernandez land continued to exist until finally, presumably with the late Tertiary uplift of the yVikIcs, the last rest disappeared, but not before considerable magma ejections had given birth to the two archipelagoes. To judge from the degree of denudation and in view of the recent volcanic activity close to Masa- tierra and on San h'elix^ the islands are young, probably Pliocene, ( iiando cxistt'a todavia im resto de la antigua Tierra de Juan Fernandez, de la cual inm/xro la flora del lioccno? Ciiando mas tarde se hundi6 tambien este resto, sobre- salian solanicntc las i)artc's volcanicas, constituyendo las islas actuales de Juan Fernandez, (lue Servian j)ara rcfugif) de la flora. This is the process as I have described it (.^27. 43) and BrCgcjKN is of the same oj)inion from the geologist's viewpoint. And if we go back to Hooker's lecture, we shall [wmX that the same idea, ai)i)lied to a different region, was famil- iar with him. 1 This refers to the siil,>inarinc eruption in 1835 "^r>fsis Berry is a species of Podocarpus, and Sequoia cJiilensis Fngclhardt p. p. another. These, together with a fern described by HallE (Lygo- ({n())i, ;i2] are important additions to the Arauco flora. F^LORIN, in accordance with l^KrciCl.N, ref(,'rs it to the I'>)cene and characterizes it as follows. The ( oinpo.sition of the fossil c:onifer vegetation, and the distributional aspects, of it> < onstitucnts, iii(!i( ;itc that it derives from a \varni-teiuj)crate or sul)troj)ical rain-forest, more particularly a lowland i)od()c.arj)-evcrgreen dicotylous broad-leaved tree forest, growint,' on tin- ( oastal plain or j)erhaps partly on low hills not far from the coast. The ( limatc was ])r()hal)ly ( harac tcri/.ed hy great humidity and rather uniform temperature, it was trostless, and wanner than the i)rescnt climate of the same district (p. 26). The possibility that plant material from the uplands had been carried down and become nnxcd with material frf)m the coastal plain is contradicted by the state ol preservation which is the same in all cases (p. 26). 'I he rhiiileufu flora. — The fossiliferous beds of Rio Pichileufu are situated in 41 s. lat. about 30 miles east of Lake .Xahuelhuapi in a treeless steppe country. DERIVATION OF THE FLORA AND FAUNA 387 Berry {2^) distinguished more than 130 species, most of them referred to still living genera and belonging to 48 families and 21 orders. The general character is subtropical, and the following families may be mentioned: Anacardiaceae, An- nonaceae, Apocynaceae, Asclepiadaceae, Bignoniaceae, Burseraceae, Caesalpinia- ceae, Celastraceae, Cochlospermaceae, Erythroxylaceae, Euphorbiaceae, Flacourt- iaceae, Icacinaceae, Lauraceae, Loganiaceae, Meliaceae, Mimosaceae, Monimiaceae, Moraceae, Myristicaceae, Myrtaceae, Nyctaginiaceae, Rubiaceae, Rutaceae, Sapinda- ceae, Sapotaceae, Sterculiaeae, Styracaceae, Symplocaceae, Vitaceae. Of conifers we find Araucaria pichileufuensis, Fiizroya tertiaria, Libocedrus prechilensis and 2 Podocarpus\ further, there is a species of Zaniia and Ginkgo patagonica, and of ferns 3 species, one of them a Dicksonia. Araucaria and Libocedrtis are, according to Florin, correctly named (pj), Ginkgo should be called Ginkgoites\ Fitzroya belongs to Podocarpus, and one of the Podocarpus sp. belongs to Acmopyle of PiLGER. There is no trace of Fagaceae and Berry referred the flora to Low^er Mio- cene and regarded it as contemporaneous with the Arauco flora; they have 20 species in common, the general character is the same and is said to bear witness of the same climate. The relief of the Andes was low, no rain-shadow existed, prevailing westerly winds carried abundant moisture across the country, there was rain forest where now we have dry grass-land. Still, there is a difference between Arauco and Pichileufu. Berry [zf] pointed out that the present South Chilean rain forest flora is not represented in the Arauco flora whereas the Pichileufu beds contain such Chilean genera as Azara, Berberis, Maytenus and Myrceugenia and, in addition, the following Antarcto-tertiary genera: Drimys, Embothrium, Eucrypliia, Laurelia, Libocedrus and Lontatia — provided that the determinations are correct. Nevertheless the age is supposed to be the same. Lower Miocene according to Berry, Eocene according to Florin, thus older than the Araucaria-Nothofagus beds of Magallanes. It is surprising that, if the two floras are of exactly the same age, the advancing Antarctic flora had not found its way to the coast of Chile; Pichileufu ought to be younger, but perhaps still Eocene, a period of very great length. The Chalia flora. — Of considerable interest was the discovery, in Santa Cruz Territory in the valley of Rio Chalia about 51° s. lat., of a fossil flora similar to the Arauco and Pichileufu floras and proving that the subtropical vegetation had extended far south. Araucaria and Nothofagus are absent, the only conifer found, Fitzroya tertiaria, is, as shown by FLORIN, a Podocarpus. Of angiosperm families Anacardiaceae, Annonaceae, Bignoniaceae, Lauraceae, Monimiaceae, Myr- taceae, Sterculiaceae etc. are represented, of Chilean genera Laurelia and Peumus may be mentioned. The age is early Miocene according to BERRY (jj^). Eocene according to Frenguelli [337), the climate warm temperate. Berry {313) regarded all the fossil floras containing an abundance of Faga- ceae [Nothofagus, according to DUSEN also Fagus, which is questionable) as of approximately the same age and older than the Concepcion-Arauco series. The Nirihuao flora. — Three localities close together on Nirihuao river near Lake Nahuelhuapi. Some ferns, among them Alsophila australis, also known from Seymour Island, further Zamia, Araucaria Nathorstii, Fagus (?) and one species 388 C. SKOTTSBERG o{ Xo/Ziofcii^iis. A^e proposed b\' HKkRV (jJ4) Lowest Miocene or Upper Oligocene, then \-oun<^er than the Ma^allancs flora. r/if Mdi^iil/cvns fiord. — DusKX [yQ) distin^uislied two plant-bearing horizons, an upper Ayaucafia horizon and a lower Xothofagus horizon. Of the remaining dicots'lons genera none were identified with living ones; DUSEX preferred to call them liscalliuiiipJirUiDii. I lydnvioeiphylliDu etc. There is no obviously tropical eletnent, it is a temperate flora. DlSKX regarded the two horizons as distinctly different in age, dating the upper to Lower ^Miocene, the lower to, perhaps, Oligocene. Hkrrv. who doubted the correctness of this distinction, regarded them as older than the Concepcion-Arauco flora, which seems improbable. riie Snuioin- flora. — To judge from DusKX's description (cVo) the tropical element is not conspicuous, whereas the actual South Chilean forest flora is well represented: Araucaria (nearly related to A. araiicaua), Drimys, Xotfiofagus., L'aldcluria, lAUirelia, I.oDiatia, all supposed to be of Antarctic parentage; lean see little reason for I^KKR\'s assertion that the Seymour flora contains "a large element of subtropical or warm temperate types like those found to-day in south- ern Hrazil" together with "another large element of forms suggestive of the existing temperate flora of Southern Chile and Patagonia"; the former was sub- tropical and coastal, the latter temj^erate and montane, washed down, DuSEN thought, from the mountains and embedded together with the leaves of the low- land trees. The age was estimated to be Upper Eocene. Florin's discovery of a species of Aonopyle [I^liyllites sj)., DlSEx) is of particular interest [338]. We have seen that Berry considered the Xotfiofagus beds to be older than the Arauco-Pichileufu deposits. y\ll the local fossil floras of Patagonia are, he says [333], older than the marine Patagonian transgression and undoubtedly pre- Miocene. P'rexcueli,! distinguished three epochs: (a) Late Cretaceous to early lv)cene, with a tropical flora, known from the Chali'a beds; (b) an intermediate period with subtropical and temperate types [Xotfiofagus) mixed; to this he would, 1 suppose, refer the Seymour flora; (c) the youngest epoch, Miocene-Pliocene, evident!)' extending into Pleistocene: a temperate flora, now ranging along both sides of the southern Andes and characterized by the dominance of A^otfiofagus and ot a number of conifers. The more or less corresponding development of the Andes was according to I5ERRV (2S): (i) I'^ocene-early Miocene: low relief, no high continuous mountains, followed by (2) a period of great uplift; (3) late Miocene to early Pliocene: mature erosion, low relief; (4) late Pliocene to Pleis- tocene: extensive uplift, beginning of the formation of the Chilean and Peruvian deeps, where earlier there was land; (5) submergence of the coastal plain. Finally, but not mentioned by HlRRV, the series of glacial and interglacial periods, a most important factor of disturbance. 'I his is tlie background against which we have to discuss the history of the Juan I'ernandez flora. DERIVATION OF THE FLORA AND FAUNA 389 Chapter XII. Antarctica as a source of the present circumpolar floras. Much has been written on this subject and it is not necessary to review the entire hterature, which has been done already by several authors, but the Antarctic problems are so important when it comes to an analysis of Juan Fernan- dez that they cannot be passed in silence. J. D. Hooker was the first to survey all the lands scattered around the Antarctic, Tierra del Fuego and the Falkland Islands, Kerguelen Island, Tasmania, New Zealand and its subantarctic dependen- cies; he was struck by the discontinuous distribution of many genera, families or even species— Hemsley, i2y (a), and the author [228) have given lists of such genera and species — he drew the consequences, although nothing was known then about the vanished flora of the large, ice-covered continent, nor of the palaeo- geography of the adjacent zone, and the idea that Antarctica had formerly ex- tended farther north and that the sporadic southern islands eventually were remnants of larger land masses entered his mind. Since that, the various subant- arctic and austral floras have become very well known, and the cases of remark- able disjunctions have multiplied. To those who adhere strictly to the hypothesis of long-distance dispersal across the oceans this means nothing more than further proofs that they are right, for the west-wind drift explains everything. Fortunately, the land bridge between South America and Antarctica rests on solid foundation, geographical as well as geological. In his important paper of 1929 [137) HOLTE- DAHL has shown that the old idea of land connection between Tierra del Fuego and Graham land (Palmer peninsula) by way of the Burdwood Bank, Shag Rocks, South Georgia and the South Sandwich and South Orkney Islands, which had been doubted by some, holds good: we have to do with a mountain range, a continuation of the South American Andes, bordered by deep water which, on the Pacific side, has the character of an abysmal trench but which exhibits old sediments to such an extent that we are forced to postulate land where there is now deep sea. The South Sandwich Islands, being entirely neovolcanic, have the appearance of an "oceanic" archipelago, but they were built up during late Ter- tiary times over an older foundation — a parallel to the history of Juan Fernandez and of many other islands. It is . . . quite evident that the South Shetland land mass has once had several times the width that it has to-day. . . . With their large amount of terrigenous, clastic sediments etc., the South Orkneys and also South Georgia agree with the folded ranges of the continent. In fact, in order to explain these masses of sediments we must necessarily assume land to have been present where there is now deep sea. We need not assume that all the links of the Scotia or, as it is also called. South Antillean Arc were united at the same time, let it be that this is possible or even probable; if so, there was no communication by water between the At- lantic and the Pacific, which undoutedly must have had its consequences to the water circulation, a question I am quite unfit to discuss. To judge from the compo- 390 C. SKOTTSBERG sition of the local subantarctic floras, rich in herbaceous plants belonging to many different families and orders, land communication must have persisted to middle or even late Tertiary times. That the connection goes far back is shown by the occurrence of a Gondwana flora on both sides of the Drake passage (Graham land, Falkland Islands), and the Cretaceous rocks of South Georgia prove, as HOLTEDAIIL says, the existence of a land mass of considerable size where there is now sea. Joyce (<5-f) remarks that "there is good evidence that the Scotia Arc with its extension into West Antarctica has persisted as a structural feature since Lower Palaeozoic times". The Scotia passage offers one of the migration routes over land that we are in need of, but another passage, the iMacquarie route between East Antarctica and Australia-New Zealand is required to make the trans-antarctic route complete. There are intermediate islands, Macquarie, Auckland and Campbell Islands, with a subantarctic flora suggesting former connections, and there are tracts with shal- lower water between Tasmania and the continent, but for want of geological evi- dence this bridge is hypothetical, and many authors prefer to speak of submerged intermediate islands sufficiently close to facilitate the spread of organisms able to cross moderate water barriers. As AxELROD says [14.. 183): Archijielagoes of only slightly greater extent than those now present could account for the continuity of the Antarcto-Tertiary Flora in all these regions during the early and middle Cenozoic. Florin, referring to the present and former distribution of conifers, expressed himself in similar terms. Antarctica has ])layed an important role in the development and distribution of the southern group of conifers. The data related to its distributions considered in this paper seem most readily explained by assuming land connections, or at least much closer proximity between Antarctica and the adjacent southern ends of South America, Australia, New Zealand and South Africa (95.92), and after the discovery of a Tertiary species q{ Acuwpylc in Patagonia and a sec- ond fossil form in the TLocene of Seymour Island, he wrote [33S. 136): Fur cine ehemalige Verkniipfung der australischen Region mit der Antarktis spricht auch die Verhreitung der Gattung Acmopyle. Dass also die Antarktis in diesem Falle als cine alte Vermittlerin zwischen der australischen Region und Siidamerika gedient hat, muss meines Erachtens angenommen werden. However, one of the supposed links, Macquarie Island, does not, it seems, possess any plants, perhaps not even mosses or lichens, dating back to the height of the Glacial lY^riod, fcjr this island was, at least during maximum glaciation, entirely ice-covered and must, Tavlor says, have received its present plant world, very poor it is true, in postglacial time from the north and across a considerable stretch of open sea [263). In Antarctica proper the situation is, with regard to lichens and mosses, different. Dahe [^2. 231), basing his opinion on the discovery of numerous endemic lichens and of a few mosses not very far from the south pole, concludes that part of the flora survived the glaciations. Here, where high DERIVATION OF THE FLORA AND FAUNA 39I mountain ranges are found right along the coast, the inland ice, even during maximum glaciation, cannot have covered everything. The flora may not have been as rich in species as it is now but perhaps a little more varied than one has been inclined to think. Many species are also found in the subantarctic zone, and further research work will, perhaps, reduce the number of endemic species. Florin spoke, as we have seen, of a "proximity to South Africa". This is where most biogeographers hesitate, in spite of such eloquent facts as the distri- bution of Restionaceae, Proteaceae and other families, and the occurrence of a subgenus of Gunnera on African soil. HoOKER, it is true, had a vision of a larger Kerguelen land, but this was still a long way off from Africa. Perhaps GULICK {^iig) ought to be mentioned here; he opposed the continental nature of isolated islands, but he was tempted to exempt what he called "continental outsiders", "Kerguelen, Crozet, St. Paul and two or three more"; the sea lacked the deepness of a typical ocean, and sediments occurred on Kerguelen; for these reasons he admitted a possible former existence of a "northward lobe of the Antarctic con- tinent". The lichenologist C. W. DODGE has taken up this question; the lichen flora of "Kerguelia" presents features of great antiquity as well as of prolonged isolation (5 endemic genera), and the angiosperms include such aberrant types as Pringlea and Lyallia\ it should be remembered that Werth was opposed to over- seas dispersal. Kerguelen is volcanic, but old, the oldest lavas dating from late Mesozoic or early Tertiary times, and on them fluviatile sediments and, on top of these, Oligocene strata rest. Erosion broke down the island during Miocene- Pliocene, but renewed volcanic activity followed from the end of Pliocene into Pleistocene. The Gaussberg-Kerguelen ridge connects Antarctica with Kerguelen + Heard Island; an elevation of 400 fathoms would be sufficient to unite the two islands, a rise of 100 fathoms would transform the Crozet group into a single island, and DODGE supposes that there is a submarine connection between Ker- guelen and the Crozet swell. Kerguelia in its prime would include all the islands, also Marion and Prince Edward. The great difficulty, the extension to South Africa, remains. So much seems to be certain that, if this bridge did exist, separation took place early, long before the other Antarctic connections were broken off. I think that the majority of phytogeographers agree with MERRILL who, in his last work [306), wrote that there is no reason whatever to doubt the validity of this ancient Antarctic route of migration of various families and genera of plants; certainly, no experienced phytogeo- grapher would question the validity of this route, for it is as thoroughly established as its more evident equivalent by what is now the Arctic region (p. 178). Most botanists have drawn their conclusions from the present distribution of the plants and this is, as a rule, all they can do, because few have left any traces of their distribution in earlier epochs. Nevertheless we have no good reason to doubt the important part taken by Antarctica in the history of the south hemisphere, but the proofs that such was the case. Berry emphasizes [28. 34), 392 C. SKOTTSBERG must rest on palaeobotanical evidence, and it happens that the fossil records are at variance with current ideas. The occurrence of Ayaiicaria, Driniys, Laiwe/ia, XoiJiofagus etc. in Tertiary deposits on Seymour Island would lead us to infer that they are of Antarctic origin and have radiated from there, but Araucaria once iiad a world-wide distribution, Drhuys belongs to an order — Magnoliales — of Holarctic range, the same is true of Fagaceae, and even Laurclia is, Berrv points out, open to doubt, in spite of the fact that the Monimiaceae are a south- ern family. Many other genera, called Antarctic on the strength of their modern distribution, are known as fossils in the north temperate and Arctic zones. '' Araji- caria stands as a per[)etual warning against forgetting that the past is the key to the {:)resent", l^KRRV wrote (I.e. 36). A Holarctic genus may have reached New Zealand or Australia as well as Patagonia from the north, never having used an Antarctic route, and without leaving a trace of its wanderings. On the other hand, the little we know about the preglacial vegetation of Antarctica is sufficient to prove that this large land mass, just as every other part of the globe, was inhab- ited by a rich and varied flora, that it may have been a primary centre of evolution, that, in other instances, it served as a secondary centre and that it was a much-trodden road between America and Australia-New Zealand. Miss GlHHS appears to have been one of the very few experienced phyto- gcogra[)hers who refused to regard Antarctica either as a centre or as a migra- tion route over land; it had always been surrounded on all sides by water and no other agents than "the wild west wind" [106. 103) and a pole-ward north-west wind, coming from Asia, were needed to explain every distribution pattern. The southern focus of development was not Antarctica but the mountains of New (kiinea. The highland of Tasmania, the subject of her survey, had received nothing from the south, all the so-called Antarctic plants, genera lika Ahroiajtclla, Astdia, C'arp/ia, Colobajiihus, Coprosnia, Driinys, Gaimardia, Gunner a, Lage?io- pliora, Xotliofagus, Orcobolus and so on, had come from New Guinea, and from there they had radiated to Polynesia, Hawaii, Juan P^ernandez, Tierra del PTiego and, I presume, Antarctica. Had she lived to hear of the discovery of A7v//^<7^//i' in New (luinea, where more species have been found than anywhere else, and in .\ew Caledonia, and of the rich development of the Winteraceae in New Guinea, she wf)ul(l have regarded such finds as a forcible proof of the correctness of her opinion. Many /ooge()gra{)hers have looked with much suspicion at the Antarctic continent as a centre of radiation. Sl.Ml'Sox, in his review of the theories involving Antarctica in the distribution of vertebrates (^^,?), concluded that dispersal had been, in all cases, from north to south; not even the Scotia Arc had ever been used as a route of migration. His reasoning is logical and often conclusive. The invertebrates are, however, left aside. To quote part of his summary (p. T^y): There is no known bioti( fact that demands an Antarctic land-migration route for its explanation and there is none that it more simply explained by that hypothesis than by any other. The affinities of the southern faunas as a whole are what would he exj)ected from the present northern connection known, or with considerably prob- ability inferred, to have existed at ai)propriate times in the past. 'J'here are certain DERIVATION OF THE FLORA AND FAUNA 393 troublesome anomalies and exceptions in the evidence, but none of these can be ade- quately explained by postulating an Antarctic connection. The general weight of evi- dence is against such a connection. In scientific theory the best-supported and most nearly self-sufficient hypothesis should be preferred and unnecessary additional hypotheses should be rejected or held in abeyance. On this basis the Antarctic migration route hypothesis remains simply a hypothesis with no proper place in scientific thinking. To this I shall make a few remarks. If Simpson had said "no fact involving the vertebrates" instead of "no biotic fact" — very well, let us assume that the routes across from and to either America or New Zealand were impassable to mammals, reptiles, amphibia and flightless land-birds, birds with good flight capa- city would have found little difficulty to cross, and many biotic facts are known that clearly speak in favour of an Antafctic migration route for invertebrates and plants. Simpson must have thought that either did the Antarctic continent never possess a fauna of land vertebrates, or, if it did have one, it had evolved inde- pendently of all other faunas and disappeared without leaving a single trace. Only penguins are known in a fossil state in Antarctic Tertiary deposits. Until fossil land vertebrates are discovered, the question of the former existence of an Antarctic fauna of terrestrial vertebrates must be left open. Among the invertebrates are many examples of a discontinuous distribution most readily understood if Antarctica is taken into account. Several were men- tioned in the chapter devoted to the composition of the fauna of Juan Fernandez, a few more may be quoted here. Berland, dealing with the Pacific spider fauna (2j): Nous avons tire de notre etude cette notion importante que la liaison entre I'Aus- tralie et I'Amerique a eu lieu par une terre antarctique dont les temoins restent ac- tuellement, mais ni par la Nouvelle Caledonie, ni par la Nouvelle Zelande, ni par le groupe Samoa-Tonga-Fiji, et, par voie de consequence, qu'elle n'a pas eu lieu par le centre du Pacifique (p. 1053), Berlioz (2^), with examples off'ered by the distribution and affinities of beetles, states that the group of Buprestidae, forming "le noyau essentiel" in the Bupres- tid fauna of Australia, has mainly South American affinities (several genera), and that la faune des Lucanides d'Australie et de Papouasie presente avec celles de I'Amerique du Sud surtout de la region andine et patagonienne, des affinites aussi etroites que curieuses. Finally Lindsay (idy), calling attention to the Subantarctic Collembola, extremely delicate creatures "supposed not to be carried any appreciable distance either by wind or sea, thus being important proofs of former land connections", mentions a genus of 3 species of which one is Fuegian, one recorded from the Scotia Arc, and one found in New Zealand. And other similar examples may be found. 394 C. SKOTTSBERG Chapter XIII. The history of Juan Fernandez — a tentative sketch. In my first [)aper on tlie Botany of Juan Fernandez (22/^), where only the vascular plants, as known at that time, were included, I expressed my view on the history of the flora in the following words: Frcilich haben wir keine Ahnung davon, wie schnell Arten oder (lattiingen ent" stehen, aber wir konnen uns kaum denkcn, dass in der kurzen Zeit, die seit der Ent" stehung der jetzigen Inseln verHossen ist, sich Typen wie Lactoris oder Rohinsojiia aiis "Keimen" entwickelten, die nach den Inseln gebracht wiirden, um sich in ungestorter Isolierung umzutormcn . . . Ich bin der Meinung, dass das alte Klement nicht auf Masa- ticrra oder Masafucra entstand, sondern alter ist als die jetzigen Inseln, und dass es wenig wahrschcinlich ist, dass die alten, endemischen (oder andere, eng vervvandte) (iattungcn und Arten von Juan Fernandez, noch nachdem die Inseln gebildet waren, die vielen vermeintlichen Ursprungsorte bewohnten, und dass Veriinderungen in der I'flanzenwelt von Polynesian, Neuseeland, Chile u.s.w. in (puirtiirer Zeit die isolierte Stel- lung bewirkt haben. Ich glaube also, dass in vor- und friihtertiarer Zeit grossere Ent- wickelungszentra existierten, und dass ihre Flora nunmehr als ein altpazifischer Rest fortlebt. Auch eine Resttlora wird sich aus vielen Familien und (iattungen aber ver- ba! tnismiissig wenigen Arten zusammensetzen. At that time our knowledge of the flora was incomplete; much fresh and new material was added in 19 16-17. I had fixed my attention on what I called the "Old Pacific element", but did not venture to look for an exact site of an evolu- tion centre, though the possibility of Antarctica as an important source of genera and families had been pointed out before; the first sign of the presence of there- tofore unknown, subantarctic flora in Masafuera had been observed, but I had no reason to link it with the Old Pacific types. During the 19 16-17 survey a rather strong Antarctic-bicentric group took shape, and in my 1925 sketch of the history of the flora (.?.?/) not only was this, but also the Old Pacific plants claimed to have "reached Juan F'ernandez over South America, where they have disappeared" (p. 31). My object this time is to see if we can approach these problems in other than general terms. Our starting point is the "Tierra de Juan P'ernandez" of BrCckien, forming a westward extension or lobe of South America, reaching the actual site of Juan PY'rnandez and Desventuradas Islands, as indicated by the bathymetrical conditions, something in keeping with the peninsula of Lower California and separated from the coast by a broad, toward the south gradually narrowed bay. However, if we remember that the deep trough is supposed to have originated with the final uj)lift of the Andes, the j)resent coast line ought to be recent all along, but this is not in accordance with the opinion that the P^ocene Arauco flora was a coast flora. The conclusion would be that the great depths were initiated already during Cretaceous times, getting dec|)er and deeper with the successive periods of uplift. The Juan PY'rnandez land or peninsula formed part of the neotropical P2ocene flora region which extended from Venezuela and Brazil to south Chile and east across the mountains, during this era of low relief, as shown by the fossiliferous beds DERIVATION OF THE FLORA AND FAUNA 395 of Pichileufii. As the Concepcion-Arauco flora it was a subtropical rain forest flora with podocarps, tree ferns, evergreen dicotyledonous trees and lianas, be- longing to some 30 tropical families; the species were, with two exceptions, referred by Berry to still existing genera — if the determinations are reliable. This is perhaps more than we can expect; all we can say is that Berry was a man with a wide experience of both fossil and living tropical plants and that undoubtedly many of the families listed and perhaps also a fair number of the genera are correctly placed. Two are found in the present flora of Juan Fernandez. To Azai-a celastriniformis and tertiaria — and the fossil does suggest Azara — BERRY remarks {28. 107): As a recent form occurs on the island of Juan Fernandez one can predicate a con- siderable antiquity 'for the genus, which is more than verified by the present fossil forms. His Berberis corynibosiflora is of still greater interest: I have seen leaves of all the South American species and the most similar is Berberis corymbosa Hook, et Arn. of Juan Fernandez (i.e. 75). I have compared his illustrations with the island species and I am willing to testify to the striking similarity between them. A revision of these most important fossil floras, with application of modern technique, is eagerly longed for. Anyhow, the neotropical character of the old flora has been safely estab- lished, and Arauco and Pichileufu agree in their general composition; 20 species occur in both. If this flora extended to Juan Fernandez, this land must have been sufficiently high to force the prevailing westerly winds to unload part of their moisture and to give rise to altitudinal belts. The flora most likely had its special distinctive marks. In view of the very large area it inhabits, it cannot have been uniform, and different floristic provinces showed special features and had their own endemics. If anything still survives in identical or very similar form can only be a subject of conjecture and is not demonstrated by leaf impressions. Two of the genera reported from the mainland, Azara and Myrceugenia, still occur in Chile and Juan Fernandez, and I am inclined to believe that the endemic element in the insular forest flora dates back to early Tertiary times, genera like Podo- phorus, Megalachne, Jiiaiiia, Ochagavia, Nothornyrcia and Selkirkia, and species of Chusquea, Hesperogreigia, Urtica, Phrygilanthus, Cheiiopodinin, Colletia, Dys- vpsis, Ugni, Eryngium, RJiaphithamnus, Solanum and Nicoiiana. Centaurodeiidron, Yunquea and the four endemic Cichoriaceous genera stand apart. They are montane and we have no clue at all to their history, but w^e can take it for granted that they are not "new beginners", but old relicts, without any near relatives anywhere. Whether we assume that they arose in the islands and have left no marks in the continent, or derive the four dendroseroid genera, which form a natural group, from Antarctic ancestors, a possibility certainly not ofl"ered by Ceiitaiirodendron and Yunquea, we are victims of wild speculation. Few endemic ferns belong to the neotropical element, Trichomanes bigae, Dryopteris maequalifolia, Asple7t2uni macrosorum and stellaium, Pellaea chilensis, 396 C. SKOTTSBERG Polvpodhini interniednnji and Ophioglossiim ferjiiutdezianmn, but PrZ/^rrt' and OpJiio- glossum do not grow in the forest, the former inhabiting the dry coast chffs, the hitter tlie open grass kind, and they may have their own history. How and when the non-endemic South American species reached Juan Fer- nandez is hard to tell. They are temperate and, with the exception of il/j'r/^^f^/rt )ni))n)nilaria, herbaceous. Myrteola is the only member of this group that extends south to the subantarctic zone. Of the others Dantlionia, Koeleria, Stipa, Piptochae- tiii))i, lUfocliaris, JuJiciis procerus (also douibeyanus and iiubricatusi), LUiertia, Pipiroj}iia fcniandcziaua, Parictaria and Miniulus show, in their mode of occur- rence, every sign of being indigenous. Advocates of transoceanic dispersal would not hesitate to call them "late arrivals which have not had time to change", and a direct transport is not altogether impossible. It would be interesting to know if a grass land existed when Great Juan Fernandez was connected with the mainland, but unfortunately we do not even know the extension of the Stipeiuvi when the early voyagers reported on the vegetation and already found the introduced Avena barbata in dominance in the treeless western part of Masatierra. A remark made by 1^rC(;gex deserves to be quoted in this connection. A current coming from the south swept past the shore of Great Juan Fernandez. When, during the Na- \idad transgression, separation from Chile occurred, . . . cl mar del polo sur entro en comunicacion con el mar que banaba las costas de Chile Central y cl primer antecesor de la corriente de Humboldt llevo las aguas inas frescas hacia el norte, dando principio a la gran zona desertica. HrCG(;kn seems to have forgotten that at that time Antarctica was covered, not by an inland ice, but by luxuriant vegetation, and that the sea cannot have been cold; however, there must have been an uppwelling of cold water, and the dry climate of the basal belt may have been as unfavourable for tree growth as it is now. Nevertheless 1 cannot believe that the present steppe-like communities date back to early or even middle Tertiary time. On the other hand it seems (luite unlikely that the s[)ecies of Stipa, PiptocJiaetiu))i, DantJionia, etc. were in- troduced with the traffic while, in this respect, Chaetoiropis (the endemic nature of i h. inibcrbis (juestionable), the two Cypcrus, jf uncus capillaceus, Paronychia, Ccnlclla. Ilcdyotis and Plantago iruncata are under suspicion. Of the ferns found elsewhere Polypodium lanceolalurn is pantropical and old enough to have belonged to the ancient flora, and this may be true also of 7'richouiaues exsccfuiu, Hy))ieiiopJiyllum spp., Adianiuni chilcnse, Ptcris chiknsis and scmiad)iata. and lilaplioglossui)i\ 1 have suggested that the latter was carried directly to Masatierra by a northerly storm, and this could have been the case also with Polypodium Masafucrac, observed a single time lOO years ago and never again. P. tric/ioiuauoides remains doubtful in spite of the specimens still extant and labelled Juan hY'rnandez {2^(j. 766). The presence of a large, presumably boreal element is not difficult to explain, for it extends all along the Andes to the far south. Many of the species are endemic, Agrostis juasafucraua, 2 sj). of Spergularia, 2 Berberis, Cardamine Krues- DERIVATION OF THE FLORA AND FAUNA 397 selii, Galium masafuerajtuni and 6 Erigeron. The Boreal character of these genera is recognized. This element extended south along the precursors of the late Tertiary mountains; the species of Agrostisi^^ Spergularia, Berberis, Cardamine and Ga- lium have their closest relatives either in the tropical Andes [Berberis] or in Chile. Erigerofi deserves special attention on account of the large number of Andean species nearly related to each other but less so to the insular group, which shows a remarkable differentiation: E.fruticosus and its cognate luteoviridis, the three herbaceous rosette herbs, and the peculiar E. rupicola of the coast rocks; of these fruticosus is found on both islands and the other species endemic on Masafuera. Nine non-endemic species, all found in Chile, also belong here, Trisctum, Carex Banksii, Pa^'onychia, Cardami^ie flaccid a, Callitriche, Rubus, Empeirum, Calysiegia and Gnaphalium, and, among the ferns, perhaps Cysiopteris. Rubus geoides forms together with R. radicans an isolated section but has a more southerly distribution but all may have reached Juan Fernandez from South Chile. Finally, Cuminia remains to be accounted for. Whether w^e link it with the palaeotropical Prasioideae, which seems to be the best way, or with Bystropogon, it stands out as an isolated relict genus. I have distinguished a large Antarcto-tertiary element, over 40 % of the angiosperms and 60 % of the ferns. Among the former there are three or, if Lac- toris is kept aside, two groups; one of them (i) is still represented in South America, the other (2) not. In (i) tw^o types can be distinguished, (a) not confined to subantarctic or alpine habitats and demanding a milder climate; to this lot I refer Uncinia Douglasii and costata, Drimys, Phrygila7itkus, Escallonia, Margyri- carpus, Sophora, Guiuicra, Apium, Pernettya, and possibly Planiago fernandezia. All have relatives in Chile. To these are added the species also occurring on the mainland: Danthonia, Koeleria, Juncus, Libertia, Acaena ovalifolia (indigenous?)^ Centella and Nertera\ the two grasses are, however, only tentatively referred to this element. The occurrence of an Antarctic element in the Eocene flora of the mainland has been demonstrated. Berry lists Araucaria, Libocedrus, Drimys, Embothrium, Laurelia and Eucryphia from the Pichileufu beds; even if the "magnolia stock" is of Boreal origin this does not exclude the possibility that the Winteraceae radiated from Antarctica, a parallel case to Fagaceae and Nothofagus. If we follow COPELANI) many of the ferns also belong to i a: Hymenoglossum, five species of Hymeiwphyllum, Lopkosoria, Polystichum, Bleclnium, Hypolepis, Histiopteris, Gleichenia pedalis (indigenous.-) and Lycopodium scariosum. Group 1 b includes the so-called Subantarctic-Magellanian element inhabiting Fuegia, the Falkland Islands, etc. ranging north along the Andes and belonging to a well-knowm circumpolar assemblage of genera and species. Here we find Oreobolus, Uncijiia brevicaulis, phleoides and tenuis, and Lagenophora hirsuta, further three endemic species, Acaeiia masafuerana, Abrotaiiella crassipes, both with near relatives in West Patagonia-Fuegia, finally Agrostis masafuerana, if its relation to the bicentric A. magellanica is confirmed after monographic treat- ment. I brought it to the boreal group. The following pteridophytes are attached 398 C. SKOTTSBERG here: Serpyllopsis, lIy))uniopliyllui)i falklandicuni, Polypodium (Grammitis) niagel- latiiciiiu, GlcicJioria quadripartiia, and Lycopodium inag€lla7naivi. Did this Antarctic element, mostly not endemic, extend to Great Juan Fer- nandez or did it arrive after the separation from the mainland took place, per- haps even after that the volcanic islands had been formed? The same question was raised when we discussed the non-endemic neotropical-temperate Stipa, Pipto- cliactiu)n, Myrtio/a, Rn/uis, and so forth; the Boreal group is, as we have seen, also inxolved. Is it probable that also this flora dates back to early or middle Tertiary times: This seems unlikely. The subantarctic species are, with the excep- tion of GrauDJiitis, restricted to the highland of Masafuera, but may have occurred also on Masatierra when the islands stood higher. Either we must assume that the land connection with the continent persisted much longer than is otherwise probable, or those species have immigrated across the water in late Pliocene or in postglacial time. I have suggested this on repeated occasions {2JI, 2^2, 340). However, we must not forget that we have to do, not with stray colonists, but with plant communities composed by flowering plants, ferns, bryophytes and lichens. Unfortunately nobody beheld the vegetation before introduced species, AuthoxiDitJniui odoratiiDi and Rumex acctosella, had invaded the highland and changed the entire aspect beyond recognition. To distinguish, among the bryophytes, the old element which undoubtedly must have formed an important part of the subtropical forest flora is more than I can undertake; we do not know if the many endemic species are relicts or not; in fact, we do not even know if they are endemic until the opposite mainland has been well explored. So much can be said that species with a pronounced tropical distribution are few, but if we add the Chilean species extending through the Valdivian and Magellanian forest zones, this South American group makes u[) about 40 "o of the mosses and 30% of the hepatics, endemic species of Ameri- can affinity included. Perhaps half a dozen mosses, not counting the few that accompany man wherever he goes, and some liverworts, have a wide distribution outside America. The dominant element is Antarctic; this was emphasized by such authorities as Cakdoi' and Hkrz()(;: 56% of the mosses and 68% of the hej)atics were referred to the Antarcto-tertiary element. When it comes to distinguishing corresponding groups among the lichens we move on very unsafe ground, but there are indications that, beside a large South .American element, we also ha\e an Antarctic group to which no less than 58 species were referred. They are austral- or subantarctic bicentric or tricentric, but many of them range north into lower latitudes. A conspicuous part is formed by Stictaceae. The great bulk of the family is by no means southern, Lobaria, Sticla and PscudocypIuUaria are frequent in tropical, subtropical and temperate- f)ceanic climates throughout, but the austral-circumpolar species are so many that we cannot exclude Antarctica as a possible source. We shall {)roceed to group 2. It comprises the genera or species which are foreign to the South American flora and have their relations in Australasia. JoilOW recogni/x'd very few; he had, just as several later authors, no other explanation to offer than that they had arrived across the Pacific from Australia, New Zealand, DERIVATION OF THE FLORA AND FAUNA 399 the East Indies, etc. without reaching the coast of Chile. They are much more numerous than JoilOW thought, 28 species: Cladium, Carex bertero7iiana, Pepero- mia berteroana, margaritifera and Skottsbergii, Boehmeria, Santalum, Ranmuulns, Fagara (2), Halorrhagis (3), Euphrasia, Coprosma (2), Waklenbergia (5), Robinso- 7iia (5), Syniphyochaeta and Rhctmodendron, the three last genera endemic. Objec- tions may be raised against including Carex and Euphrasia; the section to which Carex bertero7iiana was referred by KCkenthal is almost confined to New Zealand and barely represented in Australia, Tasmania and Norfolk Island, but one little known Chilean species is included, and Euphrasia formosissima is distantly related to E. perpusilla of South Chile. The species of Wahlenbergia are puzzling, but I have given my reasons for bringing them here as representing an African sector. The most eloquent members are, perhaps, Santalum, Ranunculus, Halorrhagis and Coprosma. This element is conspicuous also among the ferns: the extremely old Thyr- sopteris, ArtJiroptcris, entirely unfamiliar with the neotropical flora, Dickso7tia, Blechnum ScJiottii and Pteris berteroana. I never looked in earnest for a direct road across the south Pacific from Australasia to Juan Fernandez, a route which ought to have had South America as its terminus. I preferred to think that the group in question reached the islands over the Scotia bridge and South America where, however, it had become extinct. To prove this we must turn to palaeontological evidence. The Eocene beds on the mainland contain leaves of many different plants, and it is not impossible that a revision of the material will contribute to a solution of the problem. In the lists published by Berry two items call for attention, Cyatheoides tJiyrsopt- eroides in the Arauco flora, and Coprosma from Pichileufu, but the material is sterile. It is true that, to judge from Berry's illustrations, Cyatheoides suggests Thyrsopteris, but the author later [28. 57) compared it with h\s Dickso7iia patago- nica, which was found with sori and undoubtedly belongs to the Cyatheaceae. Thyrsopteris-like fossils have been reported from various places in the north hemisphere. He described 2 species of Coprosma, based on leaf impressions which, as far as I can see, tell us little about their systematic position. To C. spathulifolia he remarks: These tiny leaves have occasioned a good deal of trouble, as the South American representatives of the genus are not similar to the fossil. . . . The Chilean species are not closely similar. . . . and to C. incerta, a most appropriate name: . . . they are so much like the endemic species of Coprosma of the Juan Fernandez Islands and several forms from the Hawaiian Islands that I feel constrained so to identify them, at least tentatively. . . . I cannot find that they agree better with Coprosma than with many other genera. When Berry gives the distribution of the genus as "from the Malayan archipelago through the Pacific islands to Chile" he includes Juan Fernandez under Chile where, politically, the islands belong, for there are no species on the mainland. 400 C. SKOTTSBERG Jt is very easy to construct a liypothetical passage by which the "" Coprosnia group" readied Juan Fernandez without crossing the Antarctic or encroaching very much upon the surface of the Pacific Ocean. If we have reason to think that South America extended farther west I cannot see why this wasn't tiie case also with West Patagonia and Tierra del h\iego, a region which undoubtedly has undergone con- siderable submergence; we have to count with a wide Scotia bridge and an exten- sion of Palmer (Graham) Land, where the geographical-geological situation is the same as in South America and where the uplift of the mighty "Antarctandes" ought to have been accompanied by submergence of the fore-land. Plants and animals could have travelled by a circuitous route from the New Zealand region over West Antarctica to Juan P^rnandez without finding their way east to what is n(nv Chile. This south Pacific path was suggested above when I tried to divide the angiosperms according to their supposed primary sources; the species involved are enumerated under \: 3, forming a group "as far as known without continental American affinities, either suggesting an ancient Antarcto-Pacific track east from Australasia without reaching America, or having arrived along the road over the Scotia Arc without leaving any traces in the present American flora"' (p. 269). This idea of a South Pacific track is not new. It was postulated by Arldt as a South Pacific bridge and it finds an expression in Croizat's South Pacific base-line, which, however, if it existed, hardly permits us to draw such far-reaching conclusions as he did. In the case of Juan Fernandez only 13 genera are concerned, belonging to 11 families, the ferns not included, but they make up 20% of the angiosperms, and others may have existed that disappeared later. It remains to see if, among the cryptogams, a '' Coprosnia group" can be recognized. In IlKRZ()(;'s paper on the Hepaticae [130) a single species is indicated as restricted to Xew Zealand and Juan PY^nandez, Palla2>iciuia xiphoidcs, and two endemic s{)ecies are said to have their nearest allies not in South America but in Xew Zealand. \o case equal to Pallainci)iia is found among the mosses, but several endemic species are considered to be related, not to American ones, but to s{)ecies inhabiting the south-west Pacific region. With regard to tlie faioia I shall confine myself to some general remarks. The few land-birds are of neotropical origin. Of the seven species, three, liusttpluvius i:;aleyilus, the owl and the thrush, occur in identical forms on the mainland; the remainder are either endemic varieties or endemic species. The most divergent is linstipluvms fcniaiidoisis. (ioiri sen's idea that it originated in the islands as a mutation of /:". o^alcritus is contradicted by the fact that they are not at all closely related but even brought to different genera by some ornitho- logists. The former is a relict, the latter i)erhaj)s a late immigrant. Of the breeding sea birds the genus l^teroihoiiia forms an austral-circumpolar element and may, in j)reglacial time, have inhabited the coasts of Antarctica and adjacent islands. Little can be said as }'et about the invertebrates. The endemic leech, Ncso- philac))i())i, is an important case of non-American ancestry, and the only terrestrial amphi[)od is bicentric. The spider fauna is an appendix to the fauna of South America, but with special features; there is no endemic genus, but specific endem- DERIVATION OF THE FLORA AND FAUNA 40I ism is high. In the fauna of the mainland Antarctic affinities have been stated to occur (Berland 2j. 1044): The New Caledonia, New Zealand group has affinities with the Malaysian region and still more with Australia. But the small islands situated south of New Zealand, namely the Campbell, Auckland and Macquarie Islands, are different; they present, rather abundantly, a group of spiders, Cybaeinae, relatives of which are found in the extreme south of South America; these spiders are not present in New Zealand, but are allied to Australian and Tasmanian species. This is an interesting observation, for it is known that connections between East Antarctica and lands to the north have been looked for both with New Zea- land and over Tasmania with East Australia. Among the millipedes Aulacodesimis and Nesogeophilus are austral genera and the species endemic; Schizotaenia alacer is known from Chile, but the genus is eminently austral-bicentric. For the same reason the endemic genus of Thysa- nura merits to be noticed. A very great number of insects have been reported from Juan Fernandez, most of them endemic, also many of the genera. The majority has been described only recently and very often nothing was said about their relations; where they were stated they are, as a rule, to be found in South America. Isolated forms are plentiful and bear witness of a long history. For the single termite an Antarctic ancestry is postulated. Most of the butterflies collected have not yet been de- scribed. Diptera are numerous and largely allied to American genera or species, and the non-endemic forms mostly Chilean. I have not been able to get a proper insight into the distribution of the many genera found elsewhere. Little can be said about the beetles until Dr. Kuschel's material has been described. Two presumably austral-bicentric cases are noticed, Pycnoinerodes and Eleusis. As in so many oceanic islands there is in Juan Fernandez a remarkable display of endemic wingless Curculionids, living on the endemic plants of South American or Antarctic parentage, and examples of strict specialization are known. Host and lodger look back upon a long common history, but whether this implies a common original ancestral home or adaptation in the islands I cannot tell. Of Hymenoptera, Haplo- gonatus, Prenolepis and, perhaps, Metelia show Antarctic connections. The antiquity of the endemic land shells cannot be disputed. The Tornatellids are an ancient Pacific group and their presence in Juan Fernandez as well as in other isolated islands and archipelagoes has been considered a proof of former land connections. Their display in Hawaii is unparalleled, Germain summarized his opinion on the evolution of the Hawaiian fauna in the following terms which, mutatis mutandis, apply also to Juan P^ernandez [lOj. 995). S'il est bien ainsi, cette famille primordiale [the ancestors of Achatinellidae, Amas- tridae, Leptachatinidae and Tornatellinidae] doit avoir une tres ancienne origine et remonter au Paleozoique. . . . le peuplement malacologique de I'archipel des Hawaii est fort ancien et doit remonter a des temps primaires. II n'a put se faire, comme le preuvent les developpements precedents sur la repartition des genres et des especes, que sur une aire contenue, ce qui exclut la possibilite de considerer les iles Hawaii comme le resultat de I'activite des volcans sous-marins. 26 — 557857 The Nat. Hist, of Juan Fernandez and EaJter Isl. Vol. I 4o: C. SKOTTSBERG Finally, let us try to reconstruct the history of our islands, beginning with the time when there existed a "Tierra de Juan Fernandez" in BrCggen's sense. It must have become isolated and reduced in size rather early. We do not know if the fauna included vertebrates other than birds; if it did they did not survive the long volcanic period. Unfortunately we know too little of their history in Chile, when they hrst appeared in modern forms, and so forth. The absence of all gymnosperms is difficult to explain. The Eocene flora of Chile contained several, Araucaria, Libocedrus and Podocarpus, all still living there and accompanied by I'itzrova, SaxcgotJiaca, Pilgerodoidroji and Dacrydhim, and even if no close land connection existed, some of them ought not to have had much greater difficulties to get transported across the water barrier than some of the angiosperms found on the islands. If, on the other hand, a land bridge existed, I can see no obvious reason why conifers did not use it or, if they did, why they didn't take possession of the new volcanic soil. Introduced araucarias, pines and cypresses do well on Masatierra. The only possibility, remote perhaps, could be that they had not been able to spread as far as to Juan Fernandez when the connection was cut. Their absence gives us no clue to the time when this happened. Another element in the Chilean flora would seem to come to our rescue, the XotJiofagiis flora. In my sketch of 1925 [2JI. 33) I expressed the idea that "the connections between the islands and the main land were severed before the south Chilean flora assumed its present composition, and also before the advancing XotJiofiii^us flora reached these latitudes", and BrCcgen is of the same opinion. He states that tropical South America is the ancestral home of much of the Chilean forest flora, and adds: Pero, la actual flora de Chile central contiene, ademas, una mezcia con una flora dc (lima mas fresco, (jue despues de la separaci6n de la Tierra de Juan Fernandez inmif^ro y (jue se caracteriza por los generos Nothofagus, Araucaria, etc. Araiicaria was found in the Pichileufu beds which are supposed to be Eocene, whereas the southern beeches appear in this latitude considerably later. In the Magellanian region Dl SKN, as we have seen, distinguished two horizons, a lower, Oligocene, with Xoiliofagus, and an upper. Lower Miocene, \m\\\\ Araucaria. Accord- ing t(j Hi when also an earthquake occurred, and other earthquakes were registered in 1809, 1822 and 1835 (27 j). As I said, I believe that Masatierra and Masafuera represent two separate centres; the distance between them is 92 nautical miles and the sea is deep. Whether they are of exactly the same age and became extinct at the same time is difficult to tell; to judge by the topography, Masafuera makes an impression of being much less eroded, and, as a consequence, younger, but the difference is, I think, mainly due to differences in the basalt; the petrographical structure is not quite the same. It is supposed that the eruptions began during Pliocene and lasted a very long time and that part of the submarine ridge was still above water when the main eruption centres became extinct. Otherwise the result would have been two lifeless islands without a sign of the old endemic biota. The flora and fauna were inherited from the sinking land. The process is easily observed in many volcanic islands and nowhere to greater advantage than in the island of Hawaii. As soon as the lava has cooled down, plants get established, microscopic algae, modest tufts of mosses and particularly lichens of the genus Stereocmilon [34-1), but a conditio sine qua 7wn is that moisture is available, that erosion sets in and soil is formed. Even under very favourable conditions and with the sources for repopulation next door, it will take a long time before a closed vege- tation cover gets established. With regard to Juan Fernandez, nothing much could happen until the islands had been built up to a considerable altitude, undoubtedly greater than now, when streams rushed down the mountain slopes and started to excavate valleys, for no plant cover, not to speak of forest growth, could get established until erosion and abrasion had done part of their w^ork. This means that a good deal of the fundament, now hundreds of metres below the surface of the ocean, was still exposed and retained a portion of the original flora and fauna, which became the principal source of the flora and fauna of the 404 C. SKOTTSBERG volcanic soil. It goes without saying that chance played a dominant role and that the fragmentary character of the island world is easy to understand. It is also possible to understand why different species happened to become isolated on Masatierra and Masafuera. According to Joiiow Masafuera was populated through overseas transport from Masatierra, and this explained why the former was so much poorer, but only 50% of the vascular plants found on Masafuera were known to him. Other reasons for the dissimilarities between the islands are differences in the topography, particularly in altitude, Masatierra is 915, Masafuera 1570 m high. If we could lower Masafuera 650 metres, the entire highland region wnth its special flora and fauna would disappear. The question of the origin of the alpine flora is, as we have seen, difficult to answer. Did it exist in the islands before the separation from Chile took place, and were there any habitats where it could thrive.- Many of the species are of Antarctic origin and immigrated to the extreme south of America, with or perhaps after the NoiJiofagus flora, where they abound in the bogs of the rainy zone and in the mountains above the timberline. And if, as was explained above, the NotJiofagus flora never had an opportunity to spread to Juan Fernandez over land, nor were those Magellanian plants able to come. There are several montane plants in Masafuera which un- doubtedly date back to early times, but they are of different origin: the species of liriireron, Euphrasia fonnosissi))ia, McgalacJme masafueraiia, PJwenicoseris regia^ ]\anu7iculus capranuH and Robbisoiiia Masafuerae, all of them peculiar endemics, and of these Ranuuculus and JMegalacJine are found only along the high ridge above 1300 m. luen if some of them are of Antarctic ancestry, their history is another. When it comes to the Magellanian group, Orcoboliis, LageiiopJiora^ (ileichcnia, and so forth, it is difficult to exclude the possibility that they be glacial or postglacial immigrants. Carcx Baiiksii, limpeirnni ruhruni — only a single plant seen — (niliuui juasafueranimi, (iiiapJialiujii spiciforinc, Myrtcola yiuDiDiulai'ia and Rubus gcoides are, as was already told, not of Antarctic origin, but may have extended far south after the recession of the ice and accompanied Orcobohis, etc. And we cannot refuse to admit that various Chilean ferns, bryophytes and lichens, belonging to the forest, were transported across from South Chile, for even if the {:)revailing winds are westerly, storms from other directions occur, and there has been j^ilenty time. .Still, the floristic difference between the two islands is a warning, not to put too much faith in the efficiency of the natural agents, among which, in this case, the wind stands foremost. The biological differences between Juan Fernandez and the mainland increased during the Ice Age. West Patagonia and l^\iegia were covered by inland ice [idf); if small, ice-free refuges occurred has not been definitely stated, but is not alto- getlier improbable; their part taken in the repopulation of the country was, however, of minor importance. With the north-south trend of the Cordillera, the road toward north lay open, and the big island of Chiloe was not ice-covered, allowing the subantarctic fiora and fauna to survive, perhaps also some of the hardy trees and shrubs. The high crests of the Andes in Central Chile were covered by glaciers descending into the valleys; scjueezed between the mountains and the coast a migration back and forth went on during the successive interglacial and glacial DERIVATION OF THE FLORA AND FAUNA 405 periods. At times the subantarctic bog and heath must have occupied considerable areas north of their present range, but during all this shifting to and fro many species may have been lost, some of them surviving on Masafuera. The glaciers did not come down to the coast of Central Chile, but the climate was cold and numerous ancient stenotopic types, some of them surviving on Juan Fernandez, perished. The influence of the glacial periods cannot have been very destructive on the islands. Some plant species reduced their range and became rare or extinct; several, apparently with a very narrow physiological amplitude, are on the verge of extinction today. In Masafuera, the timberline was lowered, I suppose, and the forest patches in the valleys came down toward the sea. Masatierra has no climatic upper timberline. In a depression on the summit of the highest peak. El Yunque, Driniys, JiiaJiia, Cuminia, Escallonia, Rhethiodendron, Dicksonia, etc. etc., luxuriate more than anywhere else, thanks to the constant humidity. Part 2. EASTER ISLAND. Chapter XIV. Composition, distribution and relationships of the Flora. I. Angiospermae. Gramineae. PaspaluDi L. More than 200; trop. to temp., most numerous in Amer. forstcria)un)i Muegge. X. Caled. scrobiculatnui L. var. orbicitlare (Forst.) Domin. X. Guin., Austral., N. Caled. ^ Polyn. Axouopus l^eauv. About 75, the majority in X. and S. Amer. paschalis Pilger. Related, according to I^ILGKR, to A. scoparius (Fluegge) Pilger, a S. American species. Siipa L.i horridula Pilger. Related species in S. Amer. and Austral. Sporoboliis R. Hr. Over lOO, trop.-subtrop., most numerous in Amer. elou(^atus R. Br. [as indicus (L.) R. Br. in 2jo\. S. As., Malays., Austral., X. Zeal. Calauiai:;rostis Adans. About 150, mostly temp., north and south. rctyofracta (Willd.) Link. (Agrostis W'illd., A. filitbrmis (Forst.) Spreng.) Aus- tral., 'i'asm., X. Zeal., Polyn., Hawaii. DichclacJuic ICndl. 2, Austral., X. Zeal. sc/Nna (R. \^v.) Hook. fil. Austral., X. Zeal. nauthojiia DC. paschalis Pilger. Perhaps most nearly related to /). cJiilciisis 13esv. I'j-(li^ras/is Host. ()\er 200, trop.-subtrop. i-loiigata Jacfj. \\. Ind. to Malays, and Polyn. Cyperaceae. C)pcri(s L. cragrostis Lam. \^ery likely introduced, and this may be the case also with the other species. ^ The area of genera and species also found in Juan Fernandez was indicated in Pt. i and is not repeated liere. DERIVATION OF THE FLORA AND FAUNA 407 polystachyus Rottb. Trop.-subtrop., wide-spread. cylindrostachys Boeck. As the former. brevifolhis (Rottb.) Hassk. Pantrop. Scirpus L. ripariiis Presl. N. Amer., Calif, to S. Amer., south to Fueg., Falkland. I can- not find that var. paschalis Kiiekenth. deserves to be distinguished. Juncaceae. J uncus L. plebeius R. Br. Colomb.-Urug., Austral, Tasm., N. Zeal. Piperaceae. Peperomia Ruiz et Pav. rejlexa Dietr. Pantrop., very widespread, north to Hawaii and south to N. Zeal.; also on Rapa and possibly Pitcairn. Chenopodiaceae. Chenopodiiim ambiguiun R. Br. Austral., Tasm., N. Zeal. Polygonaceae. Polygo7imn L. About 150, world-wide. acuminatum H. B. K. W. Ind., Centr. and S. Amer., Galap. Is., trop. and S. Afr., Orient. Nyctaginiaceae. Boerhaavia L. About 20. Pantrop. diffusa L. Widespread in the Pacific and a common weed in the Old and New World; probably of aboriginal introduction. Aizoaceae. Tetragonia L. expansa Murr. Cruciferae. Nasturtium R. Br. About 50. Widespread, mostly temp. sarmentosum (Sol.) O. E. Sch, Austral.-Polyn., Hawaii. Leguminosae. Caesalpiiiia L. At least 90-100; pantrop. bonduc (L.) Roxb. Trop. As.-Polyn. Soph or a L. toromiro (Phil.) Skottsb. Nearly related to the species from J. Fern. Euphorbiaceae. Euphorbia L. 1 500-1600; world-wide. hirta L. Indomal.-Polyn., often adventitious. serpe7is L. As the former, also common in trop. Amer. 408 C. SKOTTSBERG Umbelliferae. ApiuDi L. prostration Labill. Austral-circuiii[)., incl. A. atistrale Thouars. Primulaceae. Satiiolus L. 9 (12?), I cosmop.. 2 (5?) X. Amer., 3 southern S. Amer., i S. Afr., I W. Austral., and tlie following. repejis (Forst.) Pers. Austral circump. Gentianaceae. Erythraea Neck. (Centaurium Hill.) 30-40, subtrop.-temp. australis R. Br. Austral., \. Caled., Fiji. Convolvulaceae. Calystegia R. l^r. 7-8, temp.-subtrop. stpiuin (L.) R. Br. forma. All continents, also reported from Australia, but possibly introduced in the s. hemisph. Ipoiuaca L. pes caprae (L.) Roth. Pantrop. Solanaceae. LyciuDi L. About 100, most numerous in S. Amer. carolhiiamiui Walt. var. scDuiuiccnsc (Gray) L. C. Hitchc. Rapa, Hawaii. The flora is extremely poor, not much richer than in the low coral islands: 16 families, 26 genera and 31 species, and I am not at all sure that all of them are indigenous and did exist here before man appeared on the scene; some may have been accidentally or purposely introduced by the aborigines, by the American whalers and in modern times. With regard to the "endemic" Solamini Insulae PascJialis Bitter, see 2^c)\ it was used as medicine. P'our species are endemic (13 %), but three of them belong to large grass genera needing monographic study. FoRSiKK mentions {347), beside some cultigens, only Panictmi filifonue Jacq. (-- Digitaria sanguinalis), a common weed, SJicfficldia ( = Samolus) rcpens, Aveim filiformis ( -- Calamagrostis retrofracta) and Solanioii nigruvi; in his journal (j^. As., Austral. co)ispcrsa (I^^hrh.) Ach. var. lusitaiia (Xyl.). S. Eur.; the species cosmop. Usnea Wigg. subtonilosa (Zbr.) Motyka [344). Masafuera. Described as U. Steineri var. by Zaiilhrl (KNKR, who also distinguished var. tincta Zbr. and quoted it for Easter Island; this is called l'. inicta by MoTYKA, who records it for S. Amer. only. Caloplaca Th. Vx. rub'Dia Zbr. J. Fern. lucois (Xyl.) Zbr. Tata*^., Falkl., S. Georgia. Biullia De Not. stcllulata (Tayl.) Mudd. Cosmop. fcvjiaiideziana Zbr. J. Fern. Iialopliiloides Zbr. var. The typical sp. J. Fern. paschalis Zbr. glazjouajia M. Arg. Brazil. R'niod'nia (S. Gray) ^lass. About 300; very widely distributed. Perousii Zbr. /V.ivV/r (Fr.) Xyl. entcroxantJia Xyl. forma. S.W. luir., Japan. Pliyscia (Schreb.) Vainio. picta (Sw.) Xyl. Wide-spread trop.-subtrop. Anaptycliia Koerb. spcciosa (W'ulf.) Mass. Widely distributed; in Amer. south to Fueg, These 23 sj)ecies. 5 regarded as endemic, represent, I am sure, only a minor j)art of the lichen flora and do not lend themselves to geographical speculations. There are several strange cases of disjunction serving, J daresay, to illustrate our insufficient knowledge of the distribution of lichens. V. Fungi. Our collection contained a single s[)ecics, lu)i>istella piisilla Eloyd, known before from Australia [102). DERIVATION OF THE FLORA AND FAUNA 417 Chapter XV. Composition, distribution and relationships of the Fauna. Indigenous vertebrates, birds excepted, lacking. The principal occupation in the island is farming, cattle and sheep are plentiful and roam over the island which, with the exception of outlying rocks, has lost its primitiveness. As a con- sequence of the changes in the plant cover, particularly the extermination of the indigenous trees, also the fauna was impoverished, while through the introduction of useful plants, numerous weeds and all kinds of goods many foreign insects and other invertebrates made their appearance, as the lists below will show. As little research work has been done hitherto, many more species will probably be found, indigenous as well as introduced. Aves (775). Sterna lunata Peale. Molucc, Polyn., Fiji, Hawaii. Ajious stolidus (L.) imicoior Nordmann. Sala y Gomez. The typical species trop.-subtrop., but not observed on the coast of America. Procelsterna coerulea (Benn.) skottsber^gi Loennb. Typical coerulea on Christmas I., 4 other subspecies scattered over the Pacific. Gygis alba (Sparrm.) j'oyana Matthews. With the typical species wide-spread trop. Pterodroma heraldica Salvin paschae Loennb. The typical species S.W. Pacific. Sula cyanops (Sundev.). Trop. seas throughout the world. According to the natives some other sea birds occur, but there are no land birds. Oligochaeta [i8i). Pheretina califoriiica (Kbg). Introduced. Reported from Calif., Mex., Madeira and Lower Egypt. Araneida {^22). Scytodes liigtibris Thorell. Burma, N. Caled., perhaps all over Oceania. Pholciis phalangioides Fuessli. Eur., now spread over a large part of the globe. Theridium tepidariorum C. Koch. Cosmop. + Tetragnata Paschae Berland. A large cosmop. genus. Corinna cetrata Simon. N. Caled. Hasarius Adansoni Audouin. Cosmop. Plexippus Paykulli Audouin. Cosmop. Possibly all the spiders are adventitious (Berland I.e.). Two species were determined as to genus only. 27 ~ 557857 The Nat. His*, of Juan Fernandez and Easter Isl. Vol. I 41 8 C. SKOTTSBERG Myriapoda {2'/^. PacJivnicriitni fen-ug'nieuni Latz. Wide-spread in luir., undoubtedly introduced, probably from Chile. OytJio))i()ypJia gracilis Koch, Latz. Apparently wide-spread; introduced. A third sj^ecies, belonging to Lajiiyctes, could not be named. Collembola (216). liiitoDiobyya Diultifaciala (Tullb.). N. and S. Amer., Eur., X. As., N. Zeal, J. Viixw. Introduced. Embioptera [222). Oligotojfia \\)sscl€ri (Krauss). Ceylon, Sumatra, Java. Insecta. Odonata (225). PciDitala flai'cscois Fabr. Amer., Afr., As., Austral. Orthoptefa-Dermaptera (225, 20 1). I am indebted to Dr. Princis for information on the nomenclature and dis- tribution. Anisolabis Boniiansi Scudd. Galap. Is., Masatierra. OiiycJiostylus uotulaius (Stal; Allacta, 22^. 297). Formosa, Malays., X. Guin., X. Caled., Samoa, Tahiti, Marquesas, Hawaii. Introduced from Tahiti? Periplancta Australasiac (Fabr.). Probably originally African, now cosmop.; the genus Afr. -Orient. Diploptcra punctata (I^schtz, D. dytiscoides, 22j^. 297). Ind. Ceylon, Burma, Malays., .Austral., Samoa, .Marcjuesas, Hawaii; the genus Oriental. Accidentally introduced.' McUviozostcria pliilpotti {^\\d.\\). X. Zeal. An Australian genus, represented on some Pacific islands. lUatclla raga Hcb. X. Amer.; Asia.' The genus probably Oriental. Probably adventitious. Ofthoptera-Saltatoria (j'V). (iyylliis occaiiicus Lc (iuillou. Malaysia and Japan to Polyn. CllOTARD regards its presence on luister Island as a j)r()of of the facility with which certain insects are transpcjrted large distances; nothing will prevent animals, he says, to be carried across the Pacific fr(jm yXustralia. In this case, however, I guess that man has been the agent. Thysanoptera [j). I laplotlirips usitatus Pagn. \ar. incrinis Ahlb. The typical species in Hawaii. DERIVATION OF THE FLORA AND FAUNA 419 Neuroptera (92). Chrysopa laiiata Banks. Wide-spread in S. Amer. and also found on Hawaii. A wide-ranging genus of several hundred sp. -t- Chr. Skottsbergi Esben-Peters. Lepidoptera (zj). Agrotis ypsiloii Rott. Cosmop. Cirphis Loreyi Dup. Widely distributed. Achaea melicerta Drury. On all islands in the Pacific and Indian oceans. Phytometra chalcytes Esp. Eur., As., most islands of the Pacific and Indian oceans. It is not probable that any of these Noctuidae are indigenous. Diptera [84). Sarconesia chlorogaster (Wied.). Chile, J. Fern. Introduced. + + Lip Sana insulae-paschalis Enderl. Leptocera (Coprophila) ferruginaia (Stenh.) var. insulae-paschalis Enderl. The typical species, known from Eur., Ind. and S. Amer., lives in horse-dung and was spread with the horse. The variety was picked from the carcass of a sheep and has, Enderlein remarks (I.e. 679), perhaps developed after the arrival in Easter Island where, however, there are many horses. Coleoptera. Cii rculionidae (/.?) . Aramigus Fulleri Horn. A noxious beetle, probably of N. Amer. origin, in- troduced into many countries and on some isolated islands. + Pentarthruni paschale Auriv. Areocerus fasciculates Deg. Cosmop., introduced. Dytiscidae [2gg). ■V Bides sus Skottsbergi ZxvnmQvni. The occurrence of an endemic aquatic beetle in the crater lake of Rano Kao among the endemic hygrophilous mosses is of interest. Elateridae [308). Siinodactylus Delfijii Fleut. Chile.-^ Austral., N. Guin., N. Brit., Solomon Is., Hawaii. Staphylinidae [2^). Pliilonthus longicorjiis Steph. Cosmop. Hymenoptera. Formicidae {283). Ponera trigona Mayr. var. opacior Forel. N. Amer., W. Ind., Chile. The typical species in Brazil, a subspecies in Austral. Cardiocondyla nuda Mayr. subsp. minuta Forel. Hawaii. The typical species Ind., Ceylon, Austral., N. Guin. One subsp. is Mediterranean. 27* -557857 420 C. SKOTTSBERG Tetramorium guiuecnse Fabr. Probably of African origin, now pantrop.; in hothouses in the temp, region. T. siniil/iuiioH F. Smith. As the former. Plagiolepis niactai'ishi Wheeler. Formosa, Hawaii, Society Is. Prenolepis hourbojiica Forel subsp. Skottsbergi Wheeler. The typical species known from Chagos, Nicobar and Seychelle Is., and E. Afr., Pemba I. Other subspecies in Ind., Comoro Is., E. As., Philipp. Is. and Hawaii. Ants are easily carried about by man, but it seems likely that Piaster Island also has indigenous forms. Vespidae (206). Polistes Jiebraeus F. \i. Afr., Madag., Ind., China; Tahiti? Hemiptera (21). Clerada apicicoiiiis Sign. Wide-spread, introduced. Redui'iolus capsiformis Germ. As the former. Mollusca {i8g). Umax arboruDi Bruch-Chant. Cosmop., introduced. Milax gagates Drap. As the former. Melampiis philippi Kuester. Peru. + J/. pascus Odhner. The genus distributed over the Pacific from Hawaii to N. Caled. S. Amer. Tor7iatellbwps nnprcssa Mouss. (Syn. Pacificella variabilis Odhner I.e.). Dis- tributed from P^iji to Easter Island. Perhaps introduced with living plant material. The fauna, as known hitherto, presents the same picture of extreme poverty as the flora, and even if future researches will double the number of species and reveal the occurrence of groups not yet recorded, a considerable portion will con- sist of late immigrants. The known endemics are very few and one or two of them questionable, and our experience from the old list of Juan P'ernandez Diptera (S4) bodes no good for the single endemic genus. Altogether half a dozen en- demic species and some endemic forms of lower category have been described, and of the species found elsewhere some are, perhaps, indigenous. An example of remarkable discontinuous distribution is offered by Melanozosteria pJiilpotti, New Zealand and ICaster Island ; possibly it will be discovered in intermediate stations, but such stations are difficult to find in other cases where Piaster Island is the terminus: Anisolabus BovDiausi, Galapagos Islands -f Juan P^rnandez, C/iry- sopa lajiata. South America and Hawaii, Poiicra trigo)ia, America, and llaplotJirips 7U)tatus, Hawaii. A direct overseas transport is not very probable, and I cannot tell if these animals are likely to have been introduced with the traffic. DERIVATION OF THE FLORA AND FAUNA 421 Chapter XVI. The biogeographical history of Easter Island. The composition of the present fauna and flora does not help us to throw any Hght on the earlier history of Easter Island, and we do not know what they were like before the arrival of aboriginal man many centuries ago. We know that the island became densely populated, that the natural resources, evidently poor, were exploited, the soil cultivated wherever this was possible and a number of useful plants introduced from other parts of Polynesia; tradition tells that the first colonists arrived from Rapa, but other opinions have also been expressed. ROGGE- VEEN, the discoverer of the island in 1722, did not bring a naturalist, but to judge by his narrative the island must have looked much the same as when Sparrman and the FORSTERS, who came with CoOK in 1774, made the first bio- logical observations. Forster collected and cited a few species (346) and men- tions, in his narrative (j^/), "Mimosa" [SopJwra toromiro) and Apiurn, which he had observed before in New Zealand. If there had been other indigenous trees, they had disappeared; Broussonetia papyrifera, Thespesia populnea and very likely also Triunifetta semitriloba had been introduced but were scarce. For wood the natives depended on SopJiora, and most of this was gone already. FORSTER found the place very barren, but on p. 578 he speaks of a hillock covered with toromiro, and later on another similar hill is mentioned (p. 592), but all the trees were low, not over 9 or 10 feet, the main trunk of the biggest as thick as a man's thigh. No wonder that the single canoe seen was a patchwork of pieces 2 or 3 feet long, and so was the paddle. The population did not exceed 700. When Thompson and Cooke [343) visited the island in 1886, groups of trees were ob- served in some places: In other parts of the island may be seen, in places in considerable numbers, a hardwood tree, more properly bush or brush, called by the natives toromiro. These must have flourished well at one time, but are now all, or nearly all, dead and decaying by reason of being stripped of their bark by the flocks of sheep which roam at will all over the island. None of the trees are, perhaps, over 10 feet in height, nor their trunks more than 2 or 3 inches in diameter (p. 705). The last specimens of toromiro are restricted to the inside of the crater Rano Kao. Easter Island was made a national park in order to protect the unique stone monuments, and is a bird sanctuary, but otherwise nature is not preserved but the land grazed over without restriction as far as I am aware. Among the many isolated islands of the Pacific, Easter occupies a rather unique position. Oceanic islands belong to two main categories, high volcanic and low coralline; only the former are of greater biological interest and possess the standard set of "peculiarities" described by Hooker, Wallace and others. Easter Island seems to form a type by itself. It is volcanic and cannot be called low, for the highest mountain is 530 m high and some of the craters reach an altitude 422 C. SKOTTSBERG •of 300-400 m, sufficient, one would think, to create a humid montane belt with fairly luxuriant arboreous vegetation, but of tliis there is nothing, in any case nothing left. Rains are frecjuent, and the amount of precipitation is not small, but evaporation, favoured by high temperatures and the strong S.E. trade wind, is great and most of the water rapidly disappears underground. The climax vege- tation is an oceanic steppe-like meadow or grass heath, as some would prefer to call it (J42]. The flora docs not present many of the characteristics of oceanic islands. There are no endemic genera, no peculiar endemic species, no prepon- derance of woody {)lants; Sof^Jiora is the only tree and Lychnn the only shrub; the flora is herbaceous, comj^rising few therophytes but many annual weeds. The ratio species : genus is 1.2 : i. With the exception of Gra7uin€ae, which dominate, and Cvpcraccac (some of these perhaps not indigenous) most large and world-wide families are absent, even Compositae; there are no conifers, no orchids but a fair number of ferns: in these respects the island conforms to typical oceanic islands. On the other hand, luister Island has little in common with the low islands, atolls or other coralline structures with no rock foundation exposed, where en- and Henderson (Elizabeth), an islet of raised coral said to be only 25 m high ( lY f ) but nevertheless the home of an endemic Stvitaliwi (compare Laysan of the Leeward Hawaiian Islands with S. cllipiicuui var. laysanense). The Marquesas flora is considered to be well known and the same may be true of the flora of the otiicr islands, even if no complete lists have been published; I suppose that all the novelties have been described, but my figures for wide-spread species are, perhaps, too low. There is a difference between my figures and those given by Brown, because varieties are counted by him as units equal to species, which e.\[)lains why his figures for the endemics are so high. The largest families are Rubiaceae (36), Cyperaceae (25), Compositae (19), lui[)horbiaceae (16), Gramineae (13), Leguminosae (12), and Piperaceae (10). Other large and imj)ortant families, such as Araliaceae, Cruciferae, Ericaceae, Malvaceae, Myrtaceae, Orchidaceae, Sapindaceae, etc., are represented by fewer species. We iiave every reason to believe that the flora has suffered losses after man had taken })ossessic)n of the soil. The t(jtal number of presumably indigenous species — many of aboriginal introduction and not few later arrivals have become naturalized — is 282, of which 156 are endemic within the area. The genera are 145, of which only 3 are endemic according to Brown. The ratio species : genus is almost 2:1. No genus is very large, the Largest is Psychotria with 12 sjjecies, and 10 have from 5 to 10 species each. The distribution of the species and the number of local endemics are indi- cated in Table VIII. The figures do not pretend to be exact. A large proportif)n of endenfics and of woody, arboreous or fruticose species — suffruticose excluded — are characteristic of oceanic floras of considerable anti- c}uity. Of the 156 endemic species 113 (72.4%) are woody, of the 126 found elsewhere 69 (54.8 %). The herbaceous species are, with very few exceptions, peren- nial. Systematically isolated tyi)es are few, and even the Marquesas Islands can- not, in this respect, be compared with either Hawaii or Juan P^ernandez. DERIVATION OF THE FLORA AND FAUNA 425 Table VIII. . Distribution of angiosperms in Southeastern Polynesia. Number of Number of % Woody species species endemics endemics Number % Marquesas 151 79 52.3 100 66.2 Tuamotu 44 3 7.0 23 52.3 Austral 80 12 15.0 53 66.7 Rapa 89 44 49.4 62 69.6 Mangareva 28 3 10.7 17 60.9 Pitcairn 26 2 7.7 19 73.1 Henderson 21 3 14.3 i^ 71 ^^ Total 282 15,6 55.3 182 64.5 Brown (jj) regards southeastern Polynesia, with the Tuamotus in the centre, as an old, submerged region: Affinities point to the Tuamotuan region as one of the ancient mid-oceanic centers of origin for a large part of the dicotyledonous flora of southeastern Polynesia (HI. 6); and, speaking of the distribution of Fitchia, he writes (I.e. 364): The grouping and affinity of these allied species strongly suggest the Tuamotuan region as the center of origin at a rather remote period, possibly at the dawn of the Tertiary or somewhat earlier, when it may be assumed that high (pre-Tuamotuan) islands existed in place of the low (Tuamotuan) atolls of the present. Within the Marquesas archipelago the sea is shallow: Apparently, an emergence of 100 meters would cause land to appear in six places; an additional emergence of 300 meters would unite or bring into close contact all land areas of the archipelago. . . . Botanical evidences, outlined in an earlier paper,^ indicate that the islands were at one time 1000 to 2000 meters higher than at pre- sent . . . (I.e. I. 17). The floristic affinities of this region is with Malaysia-Melanesia-Australia; there is no neotropical element in spite of the prevailing direction of winds and currents. Brown, who could be expected also to look toward America, remarks: The Cichorieae, to which Fitchia belongs, are best represented in Europe and America, pointing to a more remote American center of origin for the pre-Tuamotuan ancestral stock. This brings up the Dendroseris-Thamnosei'is "^xohX&va.TX^^ iowx <\^wdiXQ's^xQ\(\ genera form a very natural group, Thamnoseris of Desventuradas Islands stands apart, and so does Fitchia. If they are, at least distantly, related, and isolated from all other Cichoriaceous genera, then the possibility of an Antarcto-tertiary ancestry should be considered. I have, however, referred the Dendroseris assem- blage to an ancient neotropical element, absent from the present continental flora, ^ Proceed. 2d Pan-Pacif. Sci. Congr., Vol. 2, 1923. 426 C. SKOTTSBERG and TJiamuoscris finds no better place, but to derive FitcJiia from an American source seems little inviting. Another solution is, perhaps, in sight. Professor GUN- XAR KRnr.MAX kindly told me that, to judge by the pollen morphology, Fitchia may have to be removed from the Cichoriuui subfamily where J. D. HoOKER placed it next to Dcudroscyis and where it has remained. There is in the Pacific Ocean no island of the size, geology and altitude of l^aster Island with such an extremely poor flora and with a subtropical climate favourable for plant growth, but nor is there an island as isolated as this, and the conclusion will be that poverty is a result of isolation — even if man is re- sponsible for the disappearance of part of the flora, it cannot have been rich; the Marquesas Is., which have been inhabited longer, I believe, and formerly had a large native population, still preserve a fairly rich and varied angiospermic flora, half of which is endemic. The distances are too great to be overcome except on very rare occasions. The nearest land is to the west, the small most easterly islets of the Alangareva (Gambler) group, but winds (S.E. trade-wind) and currents are unfavourable for transport from \V., and Piaster Island appears to lie away from the cyclonic tracks. Beach drift is responsible for the arrival of several s{)ecies, Ipoinaca, Caesalpijiia, Cheuopodium, Tciragonia, Eryt/iraea, Apium, Samo- lus, LyciuDi and perhaps some grasses and species of Cypcrus, altogether about ^'3 of the angiosperms. Storms bring light diaspores, but it is noteworthy that CoDipositac are absent. I can find no special adaptations for bird carriage, but the j)ossibility of rare cases of epizoic transport cannot be excluded. However that may be, Piaster is a good example of an island peopled by "waifs and strays". Affinities are, as we have seen, with Malaysia-Australia or pantropical, whereas the well-marked east Polynesian flora has contributed nothing, not even its leading famil}' Rubiaceae, rich in drupe-fruited forms. Sophora torouiiro is allied to vS". ''tetra- ptcrd^ of Raivavae and Rapa; I cannot tell if this is the true /^/r^//rr^, a native of New Zealand, but I do not think it is, and as Brown's description (III. 120) shf)ws, it differs much from toroniiro, which comes very close to S. masafuerana. Neither is of American ancestry: sect. Edivardsia is austral-circumpolar and gene- rallv regarded to be of Antarctic origin or, at least, history. With the exception of Lyciian carolinianiwi var. sa7idvicense, supposed to belong to the beach drift, there is, if FitcJiia is definitely excluded, no American element in the fiora of southeastern Polynesia, nor is it expected there. It is, as we have seen, found in I^aster Island. Of the 3 endemic grasses, Stipa was ten- tatively brought to the palaeotropical element, Axonopus to the neotropical. Dan- tlionia is an austral-circum{)olar, tricentric genus. Three American, not endemic si)ecies. Cypcrus rraj^rostis, Scirpus riparius and Polygomim acumiiiatmn, remain to be accounted for. \'{ I^^aster Island once had a richer flora is an open question. According to newspa[)er reports a palynological survey of the swamp in the crater of Rano Kao was planned for IllA I.rdaiii/s recent survey. The thickness of the loose, water-soaked Cauipylopus peat was not measured by me; it is a somewhat dan- gerous quagmire which cannot be bored with the usual methods, but samples may DERIVATION OF THE FLORA AND FAUNA 427 be dug out from different depths, and if pollen of species not growing on the island are found, some light will be thrown on the history of the flora. The map accompanying this paper was prepared by the Oceanographical Institute in Goteborg. I am greatly indebted to the Director, Professor Hans Pettersson, and to Dr. Borje Kullenberg for valuable assistance. May, 1956. Additions. P. 251. Usnea Gaudichaudii Motyka, known from the "espinal" of Central and North Chile, is quoted for Juan Fernandez (Masatierra) by MOTYKA 34^. 600 as found by Bertero, 1830. As its occurrence there seems little probable — Bertero collected also near Valparaiso, etc. — I have excluded the species from my list. To Chapters IV and XVI. A recent paper by R. FURON, "Importance paleogeographique des mouve- ments de subsidence du Pacific Central" (Rev. gen. des Sciences 62, 1955), should be noted here. His object is expressed in the following terms: Constatant combien les biogeographes manquent de documentation geologique, il nous a paru utile de regrouper les notions acquises au cours de ces dernieres annees, notions qui eclairent fort bien I'historie du Pacifique depuis le Cretace (p. 307). Whereas the Galapagos Islands show, he says, a purely oceanic type of rocks, andesitic basalt and tuff are found on Easter Island and andesite and trachyte on Pitcairn. Contrary to what was told above, corals of Cretaceous age were dredged in a depth of 2000 m on one of the Central Pacific guyots. The deep borings through atolls, the latest on Eniwetok in 1953, have penetrated through coral formations dating from Pleistocene to Eocene to the bedrock of basalt, indicating a subsidence since the end of the Cretaceous of 2000 m. Perhaps other parts of the Pacific would give still greater figures. To Chapter V. In the Proceedings of the Cotterwood Natur. Field Club 31, 1955, T. A. Sprague gives an account of the Drift Theory of Du ToiT and finds that, from a botanist's viewpoint, this theory "offers the best explanation hitherto brought forward of the major problems of biogeography". With reference to the physical side of the drift process he quotes Holmes' "Principles of Physical Geography" (1944). In the case of Juan Fernandez, which certainly is one of the minor pro- blems, I cannot see that the drift theory offers an acceptable solution. To Chapter VII, p. 360. An important paper by H. M0LIIOLM Hansen, "Life forms as age indica- tors", Ringkjobing 1956, confirms the opinions of SiNNOTT & Bailev and others. Bibliography. r. Ar.r.AVKs, H. dfs. De speciebiis generis lichenum Cladoniae ex Insulis Tristan da Ciinba. Results Norwcg. Kxj). to..., no. 4. 1940. Adams, Lk Rov. The theory of isolation as apj)lied to plants. Science 22, 1905. Ahi.I'.krc, O. Thysanoptera from Juan Fernandez and Easter Island. Nat. Hist. Juan I'ern. Ill, 1922. Alkxandkk, C. r. Tipulidae (l)iptera). (insect. J. Fern. 5.) Rev. Chil. de Entomol. 2, 1952. Andrfavks, H. T. Coleoptera-Carabidae of the Juan Fernandez Islands. Nat. Hist. Juan Fern. Ill, i 93 i . Andrkws, F. C. Origin of the Pacific Insular Floras. Proceed. 6th Pacif. Sci. Congr. 1939, IV. Berkeley 1940. The structure of the Pacific Pasin. Ibid. VI. Ari.dt, Th. Die Entwickelung der Kontinente und ihrer Febewelt. Leipzig 1907, 2d ed. (vol. I only), Berlin 1938. Arnf.ll, S. List of Hepaticae collected in Marion Island by Mr. K. W. Rand. Svensk bot. tidskr. 47, 1953. Hepaticae of Chile and Argentina collected by R. Santesson. Ibid. 49, 1955. An. I', J. P. Preliminary results of Ocean magnetic observations on the Carnegie . . . October 1928 to January 1929. Terrestrial Magnetism and Atmospheric Electri- city, March 1929. 12. AuRixir.Lius, Chr. Coleoptera-Curculionidae von Juan Fernandez und der OsterTnsel. Nat. Hist. Juan Yern. Ill, 1926. 13. AuRiviFijis, CiiR., Prout, I>. B. cv Mfvrick, E. Lepidoptera von Juan Fernandez und der ( )stcr-Insel. Ibid. 1922. 14. AxKFRoi), D.J. Variables affecting the probabilities of dispersal. Proceed, of the Symposium on the South Atlantic Basin, 1949. Bull. Amer. Mus. Nat. Hist. 99, 1952. 15. A theory of Angiosi)erm evolution. Evolution \T, 1952. 16. Bai.F()i;r-1')R()\vnk, J. A(|uati(-, Coleoptera of Oceania. Occ. Pap. B. P. Bishop Mus. Honolulu XVIII: 7, 1945. 17. Bariram, E. V>. Mosses of Chile and Argentina mainly collected by R. Santesson. Svensk bot. tidskr. 46, 1952. 18. Bfchvnf, J. Alticidae (Coleojitera). (insect. J. I'ern. i i .) Rev. Chil. de Entomol. 2, I () 5 2 . 19. P)FiKR, M. I'seudoscorpione von Juan Fernandez. Rev. Chil. de Entomol. 4, 1955. 20. P)FN'iHAM, (j. Notes on the classification, history and geographical distribution of the Compositae. Journ. Linn. Soc. P)Ot. 13, 1873. 21. Bfrgrdih, E. Hemiptera from Juan I'crnandez and Easter Island. Nat. Hist. Juan Fern. Ill, 1924. 22. Bfrlaxi), L. Araignees de I'ile de Pa(|ues ct dcs iles Juan Fernandez. Nat. Hist. Juan Fern. Ill, 1924. 23. • Remar(|ues sur la rej)artition et les affinites des Araignees du Pacifique. Proceed. 3d Pacif. Sci. Congr. 1926. Tokyo 1928. DERIVATION OF THE FLORA AND FAUNA 429 24. Berlioz, J. Quelqiies considerations entomologiques et conclusions siir le role bio- geographique de la «Ligne de Wallace*; Compte rendu Soc. de Biogeogr. 30^ annee, no. 258, 1953. 25. Bernhauer, M. Coleoptera-Staphylinidae von den Juan Fernandez-Inseln und der Osterinsel. Nat. Hist. Juan Fern. Ill, 192 1. 26. Berry, E. W. The origin and distribution of the family Myrtaceae. Bot. Gaz. 59, 19 15. 27. The Flora of the Concepcion-Arauco coal measures of Chile. Johns Hopkins Univ. Stud., Geol. 4. Baltimore 1922. 28. Tertiary Flora from the Rio Pichileufu, Argentina. Geol. Soc. of Amer. Spec. Pap. No. 12. Baltimore 1938. 29. Berry, E. W. & Singewald, J. T. jr. The tectonic history of Western South America. Proceed. 3d Pacif. Sci. Congr. 1926. Tokyo 1928. 30. Bews, J. W. The World's Grasses. London 1929. 31. Bitter, G. Die Gattung Acaena. Bibl. Botan. 74, 1910-11. 32. BouLY DE Lesdain, M. Lichens recueillis en 1930 dans les iles Kerguelen, Saint-Paul et Amsterdam. Ann. Crypt, exot. 4, 1931. ;^;^. Brotherus, V. F. Musci Insulae-Paschalis. Nat. Hist. Juan Fern. II, 1924. 34. The Musci of the Juan Fernandez Islands. Ibid. 35. Brown, F. B. H. Flora of Southeastern Polynesia. I (1931), II (1931), III (1935). Bernice P. Bishop Mus. Bull. 84, 89, 130. Honolulu. 36. Bruce, N. Coleoptera-Cryptofagidae von Juan Fernandez. Nat. Hist. Juan Fern. Ill, 1940. 37. Brues, C. T. a new species of parasitic Hymenoptera from Juan Fernandez. Ibid. 1924. 38. Burtt, B. L. & Hill, A. W. The genera Gaultheria and Pernettya in New Zealand, Tasmania and Australia. Linn. Soc. Journ. Bot. 49:611, 1935. 39. Bryan, E. H. jr. A summary of the Hawaiian birds. Proceed. 6th Pacif. Sci. Congr. 1939, IV. Berkeley 1940. 40. The Hawaiian Chain. Honolulu 1954. 41. Burger, O. Die Robinson-Insel. Leipzig 1909. 42. Cain, S. A. Foundations of Plant Geography. New York and London 1944. 43. Camp, W. Distribution patterns of modern plants and the problems of ancient dis- persals. Ecol. Monogr. 17, 1947. 44. Campbell, D. H. The Derivation of the Flora of Hawaii. Leland Stanford jr. Univ. Publ. 1 919. 45. — ■ — Australian element in the Hawaiian Flora. Amer. Journ. of Bot. 15, 1928. 46. The Flora of the Hawaiian Islands. Quarterly Rev. Biol. 8, 1933. 47. Continental drift and plant distribution. Science 95, 1942. 48. Cardot, J. La Flore bryologique des Terres Magellaniques, de la Georgie du Sud et de i'Antarctide. Wiss. Ergebn. schwed. Sudpolar-Exp. 1 901-1903. IV: 8, 1908. 49. Cardot, J. & Brothf;rus, V. F. Les Mousses. Bot. Ergebn. schwed. Exp. nach Pata- gonien und dem Feuerlande 1907-1909. K. Sv. Vet.-akad. Handl. 63: 10, 1923. 50. Carvalho, J. C. M. Miridae (Hemiptera). (Insect. J. Fern. 3.) Rev. Chil. de Entomol. 2, 1952. 51. Catalogue of Birds of the British Museum (Natural History) V (1881), XVI (1892). 52. Chaigneau, J. F. Cordon submarino paralelo a las costas de Chile entre las islas Juan Fernandez y San Ambrosio. Anuario hidrogr. Marina de Chile 22, 1900. 53. Chaney, R. W. Tertiary centres and migration routes. Ecol. Monogr. 17, 1947. 54. Cheeseman, T. F. The vascular flora of Macquarie Island. Scient. Reports Australas. Antarct. Exp. 1911-14, VII. 1919. 55. Manual of the New Zealand Flora. 2d ed. Wellington 1925. 56. Chilton, Ch. A small collection of Amphipoda from Juan Fernandez. Nat. Hist. Juan Fern. Ill, 192 1. 57. China, W. E. Homoptera-Cicadellidae. Rev. Chil. de Entomol. 4, 1955. 28 - 557857 The Nat. Hist, of Juan Fernandez and Easter Isl. Vol. I 43© C. SKOTTSBERG 58. Chopard, L. Gryllides de Juan Fernandez et de Tile de Paques. Nat. Hist. Juan P>rn. Ill, 1924. 59. Christ, H. Geographic dcr Fame. Jena 1910. 60. Christenskn, C. Demonstration of C. Skottsberg's fern collection from Juan Fernan- dez. Botan. Tidsskr. 37, 1920. 61. The Pteridophytes of Tristan da Cunha. Res. Norweg. Exp. 1 937-1 938, no. 6. Oslo 1940. 62. Chrisik.nskn, C. &: SKorisuERG, C. The Pteridophyta of the Juan P>rnandez Islands. Nat. Hist. Juan Fern. II, 1920. 63. The Ferns of Easter Island. Ibid. 64. CL.AUSEN, R. T. A monograph of the Ophioglossaceae. Mem. Torrey Bot. Club, 10, 1938. 65. Continental Drift. A Symposium on the Theory of — . The Advancement of Science VHI: 29, 195 I. 66. CoPELANi), E. B. Genera Hymenophyllacearum. Philipp. Journ. Sci. 67, 1938. 67. Fern evolution in Antarctica. Ibid. 70, 1939. 68. Antarctica as the source of existing ferns. Proceed. 6th Pacif. Sci. Congr. 1939, IV. Berkeley 1940. 69. Genera Filicum. Annal. Crypt, et Phytopath. 5. Waltham, Mass. 1947. 70. CoRrEs, R. Tachynidae (Diptera). (insect. J. Fern. 9.) Rev. Chil. de Entomol. 2, 1952. 71. Croizat, L. Manual of Phyto-Geography. The Hague 1952. 72. Dahl, E. On different types of unglaciated areas during the ice ages and their sig- nificance to phytogeography. New Phytologist 45, 1916. 73. Darp.ishire, O. V. The Lichens. Wiss. Ergebn. schwed. Siidpolar-Exp. 1901-1903, IV: I. Stockholm 191 2. 74. Dodge, C. W. Lichens and Lichen Parasites. Reports Brit. Austral. N. Zeal. Antarct. Research Exp. 1929-31. Ser. B. VII. Adelaide 1948. 75. DoMiN, K. Monograi)hische Ubersicht der Gattung Centella L. Bot. Jahrb. fiir System. etc. ed. by A. Engler, 41, 1908. 76. Grundzuge der })flanzengeographischen Verbreitung und Gliederung der Leber- moose. Mem. Soc. Roy. de Sciences de Boheme 1923. Prague 1924. 77. Du RiETZ, G. E. Two new species of Euphrasia from the Philippines and their phyto- gcographical significance. Svensk bot. tidskr. 25, 1931. 78. Problems of bipolar plant distribution. Acta Phytogeogr. Suec. XIII, 1940. 79. Dlsex, p. Uber die tertiare Flora der Magellanslander. Wiss. Ergebn. schwed. Exp. 1 89 5- 1 897, I, 1907. 80. Uber die tertiare Flora der ScymourTnsel. Wiss. Ergebn. schwed. Siidpolar- Exp. 1901-1903, III, 1910. 81. Du ToiT, L. Our Wandering Continents. Edinburgh and London 1937. 82. Observations on the evolution of the Pacific Ocean. Proceed. 6th Pacif. Sci. Congr. 1939, I- Berkeley 1940. 83. Emerson, E. A. A new 'J'ermite from the Juan Fernandez Islands. Nat. Hist. Juan Fern. Ill, i 924. 84. Enderi.ein, G. Die DijUcrenfauna der fuan P'ernandez-Inseln und der Oster-Insel. Ibid. 1938. 85. Engi.ek, a. Versuch einer Entwickelungsgeschichte dcr Pflanztnwelt seit der Tertiar- periode. II. 1882. 86. EN(;i.ER'r. S. La Tierra de Hotu Matua. Valparaiso 1948. 87. Epi.ing, C. Synopsis of the South American Labiatae. Fedde, Repert. Beih. 85, 1935. 88. • The Labiatae of Chile. Rev. Universitaria 22, 1937. 89. — • — ■ The distribution of the American Labiatae. Proceed. 6th Pacif. Sci. Congr. 1939, IV. Berkeley 1940. 90. ■ Distribuci6n geogrdfica y parentesco de las Labiadas de la America del Sur. Rev. Universitaria 25, 1940. DERIVATION OF THE FLORA AND FAUNA 43 1 91. Ermel, a. Eine Reise nach der Robinson-Insel. Hamburg 1889. 92. Esben-Petersen, p. More Neuroptera from Juan Eernandez and Easter Island. Nat. Hist. Juan Fern. Ill, 1924. 93. Evans, A. W. The Thallose Hepaticae of the Juan Fernandez Islands. Ibid. II, 1930. 94. Fassett, N. C. Callitriche in the New World. Rhodora 53, 1951. 95. Florin, R. The Tertiary fossil conifers of South Chile and their phytogeographical significance. K. Sv. Vet.-akad. Handl. ser. 3, 19, 1940. 96. Fluke, C. L. Syrphidae. (insect. J. Fern. 18.) Rev. Chil. de Entomol. 4, 1955. 97. FosBERG, F. R. Derivation of the Flora of the Hawaiian Islands. See Zimmerman, E. C. 98. Lignes biogeographiques dans I'ouest du Pacifique. Compte rendu Soc. de Bio- geogr. 29 annee, no. 256, 1952. 99. The American element in the Hawaiian Flora. Proceed. 7th Intern. Botan. Congr. 1950. Stockholm 1953. 100. Freeman, P. Mycetophilidae, Sciaridae, Cecidomyidae and Scatopsidae. (insect. J. Fern. 13.) Rev. Chil. de Entomol. 3, 1953. loi. Fries, R. E. Die Myxomyceten der Juan Fernandez-lnseln. Nat. Hist. Juan Fern. II, 1920. 102. Fries, Th. jr. Die Gasteromyceten der Juan Fernandez- und Osterinseln. Ibid. 1922. 103. FuLFORD, M. Some distribution patterns of South American Leafy Hepaticae. Proceed. 7th Intern. Botan. Congr. 1950. Stockholm 1953. 104. Gebien, H. Coleoptera-Tenebrionidae von Juan Fernandez. Nat. Hist. Juan Fern. Ill, 192 I. T05. Germain, L. L'origine et revolution de la faune de Hawaii. Proceed. 3d Pacif. Sci. Congr. 1926. Tokyo 1928. 106. GiBBS, L. Notes on Phytogeography and Flora of the mountain summit plateau of Tasmania. Journ. of Ecology 8, 1920. 107. GiLLOGLY, L. R. Coleoptera-Nitidulidae. (insect. J. Fern. 24.) Rev. Chil. de Entomol. 4, 1955- 108. Good, R. See Continental Drift. 109. The Geography of the Flowering Plants. 2d ed. London 1953. no. GooDALL, J. D., Johnson, A. W., Philippi, R. A. Las Aves de Chile. I, 1946. II, 195 I. Buenos Aires. III. GooDSPEED, T. H. Plant Hunters in the Andes. New York 1941. 112. The genus Nicotiana. Waltham, Mass. 1954- 113. Gordon, H. D. The problem of sub-antarctic plant distribution. Report Austral, and New Zeal. Ass. Advancem. Sci, 27, 1949- 114. Grant, Adele L. A monograph of the genus Mimulus. Ann. Miss. Botan. Card. XI, 1924. 115. Gregory, H. E. Types of Pacific Islands. Proceed. 3d Pacif. Sci. Congr. 1926, 2. Tokyo 1928. 116. Gregory, J. W. Theories of the origin of the Pacific. Proceed. Geol. Soc. London 86, 1930. 117. GuiGNOT, F. Dytiscidae (Coleoptera). (Insect. J. Fern. 10.) Rev. Chil. de Entomol. 2, 1952. 118. GuiLLAUMiN, A. Les regions floristiques du Pacific d'apr^s leur endemisme et la repartition de quelques plantes phanerogames. Proceed. 3d Pacif. Sci. Congr. 1926. Tokyo 1928. 119. GuLiCK, A. Biological peculiarities of oceanic islands. Quart. Review Biol. 7, 1932. 120 GuNDERSEN, A. Families of Dicotyledons. Waltham, Mass. 1950. 121. GUPPY, H. B. Observations of a naturalist in the Pacific between 1896 and 1899. 11. London 1906. 122. The island and the continent. Journ. of Ecol. 7, 1919. 432 C. SKOTTSBERG 123. GuTKNBKRG, B. Geophvsical and geological observations in the Pacific area and tectonic hypotheses. Proceed. 7th Pacif. Sci. Congr. 1949, II. Wellington 1953. 124. GoKTSCH, W. Die Robinson-Insel jiian Fernandez iind ihre biogeographischen Pro- bleme. Phoenix 19, 1933. 125. Handschin, C. Neuroptera. (Insect. J. Fern. 15.) Rev. Chil. de Fntomol. 4, 1955. 126. Harmstox. F. C. r)olichoj)odidae. (Insect. J. Fern. 17.) Ibid. 127. Hkmsi.f.v, \V. B. (a) Rej)ort on the Present State of Knowledge of Various Insular Floras, (b) Report on the Botany of Juan Fernandez and Masafuera. Rep, Sci. Res. H. M. S. Challenger. Bot. I, 1885. 128. Hf.nnig, W. Phryneidae, Helomyzidae, Lonchaeidae, Piophilidae, Anthomyzidae, Muscidae. (insect. J. Fern. 16.) Rev. Chil. de Entomol. 4, 1955. 129. Hf.rzog, Th. (ieographie der Moose. Jena 1926. 130. — - — Die foliosen Lebermoose der Juan Fernandez-Inseln und der Oster-Insel. Nat, Hist. Juan Fern. II, 1942, 131. Zur Bryophytenrtora Chiles. Revue Bryol. et Lichenol. XXIII, 1954, 132. Hfss, H. H. Drowned ancient islands of the Pacific. Amer. Journ. Sci. 244, 1946. 133. Hkkfx, C. M. Polypodiacearum argentinarum catalogus. Rev. Mus. de La Plata 15, I 908. I 34, Hill, A. W. Antarctica and j)roblems of geographical distribution. Proceed, 5th Intern, Bot, Congr. 1926. Ithaca 1929. 135. HiNTON, H. E. See Continental Drift. 136. HoDGSOX, E. Amy & Saixsburv, G, O. K. Bryophytes collected by G. E. Du Rietz on the Antipodes Islands. Svensk bot, tidskr. 42, 1948, 137. HoLTFDAHL, O. On the geology and physiography of some Antarctic and Sub-Ant- arctic islands, Scient. Res. Norweg. Antarct. Exp. No. 3, Ed. by Det norske videnskaj)sakademi. Oslo 1929. 138. Hooker, J. I). Lecture on Insular Floras before the British x\ssociation at Notting- ham, 1866. Also in Gard. Chronicle, Jan. 1867. 139. Hue, A.-M. Lichens. Deuxieme Exj)ed. Antarct. Fran^aise 1908-1910. Paris 1915. 140. HuTCHhxsoN, J. The Families of Flowering Plants. I. 1926. II. 1934. 141. Huxley, J. Species formation and geographical isolation. Proceed. Linnean Soc, Lond. 1938. 142. IxsEcros. Los Insectos de las islas Juan Fernandez. Ed. by Dr. G, Kuschel, 1-12. Rev. Chil. de Entomol, 2,1951. — 13,14, vol. 3, 1953. — i 5-27, vol, 4, 1955. 143. Irmschek, E. Pflanzenverbreitung und Entwickelung der Kontinente. I-II. Mitteil, Hamburg. Inst, allgem. Botan. 5, 1922; 8, 1929. 144. Jeffrey, E. C. rnandez und der Oster-Insel. Nat. Hist. Juan Fern. Ill, 192 i . 182. Montagne, C. Lfquenes, in Gay, Hist. fis. y polit. de Chile. Botanica 8. Paris 1852. 183. MuMFORD, E. P. The present status of studies of faun al distribution with reference to Oceanic Islands. Proceed. 6th Pacif. Sci. Congr. 1939, IV. Berkeley 1940. 184. MiJLLER, J. Lichens. Miss. Scient. du Cap Horn 1882-83. V. Paris 1889. 185. Navas, L. Neuropteres des lies Juan Fernandez et de I'lle de Paques. Nat. Hist. Juan Fern. Ill, 1921. 186. NiLSSoN, H. Syntetische Artbildung. Lund 1953. 187. Nixon, G. E. J. Hymenoptera-Braconidae. (insect. J. Fern. 26.) Rev. Chil. de Ento- mol. 4, 1955. 188. Nybelin, O. Nesophilaemon n. g. fiir Philaemon skottsbergi L. Johansson. Zool. Anzeiger 142, 1943. 434 C. SKOTTSBERG 189. Odhnkr, X. Hj. Mollusca from Juan Fernandez and Easter Island. Nat. Hist. Juan Fern. Ill, 1922. Addenda, ibid. 1926. 190. Oglobin, a. Mymaridac (Hymenoptera). (Insect. J. Fern. 1 2.) Rev. Chil. de Entoniol 2, 1952. 191. Bethylidae y Dryinidae (Hymenoptera). (Insect. J. Fern. 14.) Ibid. 3, 1953 192. Oi.ivER, W. R. B. The genus Coprosma. Bull. Bishop Mus. Honolulu 132, 1935. 193. Pkrkins, R. C. L. Introduction, being a review of the land-fauna of Hawaiia Fauna Hawaiiensis I. Cambridge 19 13. 194. Pfeiffkk, H. Oreobolus R. Br., eine merkwiirdige Cyperaceengattung. Fedde, Repert 23, 1927. 195. Pic, M. Coleoptera-Anobiida de Juan Fernandez. Nat. Hist. Juan Fern. HI, 1924 196. Coleoptera-Clavicornia et autres de Juan Fernandez. Ibid. 197. PiLGF.R, R. Uber einige Gramineae der Skottsbergschen Sammlung von Juan Fer nandez. Fedde, Repert. 16, 1920. 198. *;*. Das System der Gramineae unter Ausschluss der Bambusoideae. Bot Jahrb. fur System, etc. ed. by A. Engler, 76, 1954. 199. Pi, .ATE, L. Zur Kenntnis der Insel Juan Fernandez. Verhandl. Ges. fiir Erdkunde Berlin XXIII, 1896. 200. Pope, R. D. Coleoptera-Colydiidae. (insect. J. Fern. 25.) Rev. Chil. de Entomol. 4, 1955- 201. Princis, K. Uber einige neue bzw. wenig bekannte Blattarien. Arkiv f. Zool. 41, 1948. 202. Ramiu), B. Historia da flora do planalto riograndense. Anais Bot. Barbosa Rodriguez V, 1953- 203. Reiche, K. (irundziige der Pflanzenverbreitung in Chile. Engler & Drude, Vege- tation der Erde 7. Leipzig 1907. 204. Richards, O. W. Sphaeroceridae. (insect. J. P>rn. 21.) Rev. Chil. de Entomol. 4, ^955- 205. Ridley, H. N. The Dispersal of Plants throughout the World. London 1930. 206. Roman, A. Ichneumoniden von Juan Fernandez. Nat, Hist. Juan Fern. HI, 1924. 207. RoMELL, L. Bacidiomycetes from Juan Fernandez, Ibid. II, 1928. 208. Sahroskv, C. W. A new species of Ogcodes from the Juan Fernandez islands. Rev. Chil. de ?>ntomol. 1, 1951. 209. ■ Chloropidae. (insect. J. I^ern. 19.) Ibid. 4, 1955. 210. Sainshurv, G. O. K. a Handbook of the New Zealand Mosses. Bull. Roy. Soc. of N. Zeal. no. 5, 1955. 211. Saxtis, L. de. Hymenoptera-Eulophidae, Entedontidae (Chalcidoidea I). (Insect. J. Fern. 27.) Rev. Chil. de F'.ntomol. 4, 1955. 212. Satchell, G. H. Psychodidae (Diptera). (insect. J. Fern. 8.) Rev. Chil. de Entomol. 2, 1952. 213. Schenkling, S. Coleoptera-Cleridae von Juan Fernandez. Nat. Hist. Juan Fern. IH^ 1931- 214. ScHMii), F. Trichoptera. (insect. J. Fern. 4.) Rev. Chil. de Entomol. 2, 1952, 215. Schmidt, A. Coleoptera-Scarabaeidae von Juan Fernandez. Nat. Hist. Juan Fern. HI, 1931- 216. ScHf)Tr, H. Collembola aus den Juan FernandezTnseln und der Osterinsel. Nat. Hist. Juan Fern. Ill, 1921. 217. Setchell, W. A. Les migrations des oiseaux et la dissemination des plantes. Compte- rendu sommaire Soc. de Biogeogr. 22, 1926. 218. — — ■ Migration and endemism with reference to Pacific Insular floras. Proceed. 3d Pacif. Sci. Congr. 1926. Tokyo 1928. 219. Pacific Insular floras and Pacific paleogeography. Amer. Naturalist 69, 1935. 220. Seward, A. C. A study in contrasts. The present and past distribution of certain ferns. Linn. Soc. Journ. of Bot. 46, 1922. DERIVATION OF THE FLORA AND FAUNA 435 22 1. Seward, A. C. & Conway, Verona. A phytogeographical problem: fossil plants from the Kerguelen Archipelago. Ann. of Bot. 48, 1934. 222. SiLVESTRi, F. Thysanura and Embioptera. Nat. Hist. Juan Fern. Ill, 1924. 223. Simpson, G. G. Antarctica as a faimal migration route. Proceed. 6th Pacif. Sci. Congr. 1939. Berkeley 1940. 224. SiNNOTT, E. W. & Bailey, I. W. The origin and dispersal of Herbaceous Angio- sperms. Ann. of Bot. 28, 1914. 225. SjosTEDT, Y. Odonata. Orthoptera. Nat. Hist. Juan Fern. Ill, 1924. 226. Skottsberg, C. Die Gefasspflanzen Sudgeorgiens. Wiss. Ergebn. Schwed. Sudpolar- Exp. IV. Stockholm 1905. 227. Studien uber die Vegetation der Juan Fernandez-Inseln. K. Sv. Vet.-akad. Handl. 51, 1914. 228. Notes on the relations between the floras of Subantarctic America and New Zealand. Plant World 18, 19 15. 229. The Phanerogams of the Juan Fernandez Islands. Nat. Hist. Juan Fern. II, 192 2. 230. — — The Phanerogams of Easter Island. Ibid. 231. Juan Fernandez and Hawaii. A phytogeographical discussion. Bull. Bishop Mus. 16. Honolulu 1925. 232. — — Einige Bemerkungen uber die alpinen Gefasspflanzen von Masafuera. Veroff". Geobot. Inst. Riibel 3. Zurich 1925. 233- Einige Pflanzen von der Oster-Insel. Acta Horti Gotob. 3, 1927. 234. Remarks on the relative independency of Pacific Floras. Proceed. 3d Pacif. Sci. Congr. 1926. Tokyo 1928. 235. Pollinationsbiologie und Samenverbreitung auf den Juan Fernandez-Inseln. Nat. Hist. Juan Fern. II, 1928. 236. The geographical distribution of the sandalwoods and its significance. Proceed. 4th Pacif. Sci. Congr. 1929. III. Batavia 1930. 237. Notes on some recent collections made in the Island of Juan Fernandez. Acta Horti Gotob. 4, 1929. 238. Marine algal Communities of the Juan Fernandez Islands. Nat. Hist. Juan Fern. II, 1941. 239- I-e peuplement des iles pacifiques du Chili. Soc. de Biogeogr. IV, 1934. 240. • — ■ — Greigia Berteroi Skottsb. and its systematic position. Acta Horti Gotob. 11, 1936. 241. Die Flora der Desventuradas-Inseln. Goteb. K. Vet. o. Vitterh. Samh. Handl. 5:6 foljden, ser. B, bd 5:6, 1937. 242. Geographical isolation as a factor in species formation. Proceed. Linn. Soc. of Lond., Sess. 150, 1938. 243. On Mr. C. Bock's collection of plants from Masatierra (juan Fernandez) with remarks on the flower of Centaurodendron. Acta Horti Gotob. 12, 1938. 244. Hawaiian Vascular Plants. Ibid. 15, 1944. 245. • Peperomia berteroana Miq. and P. tristanensis Christoph., an interesting case of disjunction. Ibid. 16, 1947. 246. Fine kleine Pflanzensammlung von San Ambrosio. Ibid. 17, 1947- 247. The genus Peperomia in Chile. Ibid. 248. Biogeografiska Stillahavsproblem. Statens naturv. forskn.-rad, arsbok 4 (i949~ 50). Stockholm 195 i. 249. A supplement to the Pteridophytes and Phanerogams of Juan Fernandez and Easter Island. Nat. Hist. Juan Fern. II, 195 i. 250. On the supposed occurrence of Blechnum longicauda C. Chr. in Brazil. Svensk bot. tidskr. 48, 1954. 251. The Vegetation of the Juan Fernandez Islands.Nat.Hist.JuanFern.il, 1953. 436 C. SKOTTSBERG 252. Slkumer, H. Revision der Oattung Pernettya. Notizbl. Botan. Gartens Berlin 12, 1935- 253. Smith, A. C. Taxonomic notes on the old world species of Winteraceae. journ Arnold Arbor. 24, 1943. 254. The American species of Drimys. Ibid. 255. Sorz.A LopFs, H. iv Alhuqufrque, I). Calliphoridae, Sarcophagidae. (Insect. J. Fern. 22.) Rev. Chil. de Entomol. 4, 1955. 256. Si'AKRE, B. Review of Croizat, Manual of Phytogeography. Rev. Universitaria 38, 1053- 257. SrEP.HiNS, G. L. JR. Additional evidence for a holarctic dispersal of flowering plants in the Mesozoic era. Proceed. 6th Pacif. Sci. Congr. 1939, IV. Berkeley 1941. 258. SiEEMs, C. G. G. J. VAN. On the origin of the Malaysian mountain flora. Bull. Jard. bot. lUiitenzorg, Ser. 3, 13-14, 1935-36. 259. S'locKWELE, P. A revision of the genus Stenactis. Contrib. Dudley Herbar. 1940. 260. Straneo, S. L. c^' Jeaxnef,, R. Goleoptera-Carabidae. (insect. J. Fern. 23.) Rev. Chil. de Entomol. 4, 1955. 261'. SvENsoN, H. Monogra])hic studies in the genus Eleocharis V. Rhodora 41, 1939. 262. SwEZEV, O. H. Distribution of Lepidoptera in Pacific Island Groups. Proceed. 6th Pacif. Sci. Congr. 1939, IV. Berkeley 1940. 263. Taveor, \V, B. An example of long distance dispersal. Ecology 35, 1954. 264. Theriot, J. Mousses recoltees dans Pile Mas a Tierra (Juan Fernandez) en 1927, par M. Gualterio Looser. Rev. Chil. de Hist. Nat. 31, 1927. 265. Mousses de Pile de Paques. Revue bryol. et lichenol. Ser. II: 10, 1937. 266. Thompson, G. B. Anoplura from Juan Fernandez. Nat. Hist. Juan P'ern. Ill, 1940. 267. Tonnoir, a. L. Australian Mycetophilidae. Proceed. Linn. Soc. N. S. Wales 54, 1929. 268. 'Prac.ardh, I. Acarina from the Juan P^rnandez Islands. Nat. Hist. Juan Fern. Ill, 1931. 269. 'PiRMEE, J. M. Repartition geographique des Eryngium II. Bull. Mus. Nat. d'Hist. Nat. Paris 1949. 270. 'PuvAMA, 'P. On Santalum boninense and the distribution of the species of Santalum. Journ. Japan. Bot. 15, 1939. 271. • On genus Haloragis and Micronesian si)ecies. Ibid. 16, 1940. 272. Ump.grove, J. H. 'P. See Continental Drift. 273. IsiNCER, R. L. Distribution of the Heteroj)tera from Oceania. Proceed. 6th Pacif. Sci. C'ongr. 1939, IV. Berkeley 1940. 274. Vkrhoekf, K. L'ber Myriapoden von Juan P>rnandez und der Osterinsel. Nat. Hist. Juan Fern. Ill, 1924. 275. \'n)AL (ioRMAZ, F. Geografia naiitica de la Republica de Chile. Anuario Hidrogrdf. de la Marina de Chile 7, 1881. 276. \'iERiE\ppER, F. Uber echten und falschen Vikarismus. Osterr. botan. Zeitschr. 68, 1919. 277. Wahri'.erc;, R. Einige terrestre Isoi)oden von den Juan Fernandez-Inseln. Nat. Hist. Juan I'ern. Ill, 192 2. 278. Waeeace, a. R. Island Life. London 1880. 3d ed. 191 1. 279. Weeks, L. (i. Palcogeography of South America. Bull. Geol. Soc. of Amer. 59, 1948. 280. Wegener, A. Die Entstehung der Kontinente und Ozeane. Braunschweig. 3d ed. 1922, 5th ed. 1936. 281. Weinl\r( K, H. Studies in Juncaceae with special reference to the species in P^thiopia and the Cape. Svensk bot. tidskr. 40, 1946. 282. Weise, J. Coleoptera-Chrysomelidae and Coccinellidae von Juan Fernandez. Nat. Hist. Juan Fern. Ill, 1924. 283. Wheeler, W. M. Formicidae from Piaster Island and Juan P'ernandez. Ibid. DERIVATION OF THE FLORA AND FAUNA 437 284. Willis, B. Geotectonics of the Pacific. Proceed. 3d Pacif. Sci. Congr. 1926. Tokyo 1928. 285. Willis, J. C. Endemic genera of plants in their relation to others. Annals of Bot. 35, 192 I . 286. Age and Area. A Study in the Origin and Distribution of Plants. 1922. 287. Winkler, H. Geographic, in Manual of Pteridology, ed. by F. Verdoorn. The Hague 1938. 288. WiRTH, W. W. Heleidae and Tendipedidae. (insect. J. Fern. 7.) Rev. Chil. de Entomol. 2, 1952. 289. Ephydridae. (insect. J. Fern. 20.) Rev. Chil. de Entomol. 4, 1955. 290. WooLDRiDGE. See Continental Drift. 291. WuLFF, E. V. An Introduction to Historical Plant Geography. Waltham, Mass. 1943. 292. Wygodzinsky, P. Contribuci6n al conocimiento del genero Metapterus Costa de las Americas y de Juan Fernandez. Rev. Chil. de Entomol. i, 195 i. 293. Thysanura from Juan Fernandez Island. Ibid. 294. Simuliidae (Diptera). (insect. J. Fern. 6.) Rev. Chil. de Entomol. 2, 1952, 295. Zahlbruckner, a. Die Flechten. Bot. Ergebn. schwed. Exped. nach Patagonien und dem Feuerlande. VI. K. Sv. Vet.-akad. Handl. 57:6, 1917. 296. Die Flechten der Juan Fernandez-Inseln. Nat. Hist. Juan Fern. II, 1924. 297. Die Flechten der Osterinsel, nebst einem Nachtrag zur Flechtenflora von Juan Fernandez. Ibid. 1926. 298. Zimmerman, E. C. Insects of Hawaii. I. Introduction. Honolulu 1948. 299. ZiMMERMANN, A. Coleoptera-Dytiscidae von Juan Fernandez und der Osterinsel. Nat. Hist. Juan Fern. Ill, 1924. 300. Bryan, W. A. in "El Mercurio", Valparaiso, Apr. 23, 1920. 301. Gourlay, W. B. in "The South Pacific Mail", Valparaiso, Feb. 2, 1928. 302. QuENSEL, P. Die Geologie der Juan Fernandez-Inseln. Bull. Geol. Inst. Upsala 11, 1913- 303. Branchi, E. C. La Isla de Robinson. Valparaiso 1922. 304. Hayek, A. Allgemeine Pflanzengeographie. Berlin 1926. 305. Drude, O. Handbuch der Pflanzengeographie. Stuttgart 1890. 306. Merrill, E. D. The Botany of Cook's Voyages. Chronica Botanica 14: 5/6. Waltham, Mass. 1954. 307. Hillebrand, W. The Flora of the Hawaiian Islands. Heidelberg 1888. 308. Fleutiaux, E. Coleoptera-Serricornia de Juan Fernandez et de Pile de Paques. Nat. Hist. Juan Fern. Ill, 1924. 309. ScHEDL, K. E. Chilenische Borkenkafer. Rev. Chil. de Entomol. 4, 1955. 310. Wygodzinsky, P. Reduviidae y Cimicidae (Hemiptera). (Insect. J. Fern. 2.) Rev. Chil. de Entomol. 2, 1952. 311. Arwidsson, Th. Einige parasitische Pilze aus Juan P'ernandez und der Osterinsel. Svensk bot. tidskr. 34, 1940. 312. Halle, T. G. A fossil fertile Lygodium from the Tertiary of South Chile. Ibid. 313. Malta, N. Die Gattung Zygodon Hook, et Tayl. Latv. Univ. Botan. Darza Darbi n. I. Riga 1926. 314. Fennah, R. G. Homoptera-Delphacidae. Proceed. Entomol. Soc. of London, Ser. B 24, 1955- 315. Berry, E. W. Fossil plants from Chubut Territory collected by the Scarritt Patagonian Expedition. Amer. Mus. Novit. 536, 1932. 316. Brinck, p. Coleoptera of Tristan da Cunha. Results Norweg. Scient. Exped. 1937-38 no. 17. Oslo 1948, 317. Werth, E. Die Vegetation der subantarktischen Inseln Kerguelen, Possession und Heard-Eiland II. Deutsche Sudpolar-Exped. 1901-02. Bd. VIII. 1911. 318. Skottsberg, C. The Flora of the Hawaiian Islands and the History of the Pacific Basin. Proceed. 6th Pacif. Sci. Congr. 1939, IV. Berkeley 1940. 438 C. SKOTTSBERG 31Q. Stebbins, G. L. jr. Variation and Evolution in Plants. New York 1950. 320. LiNDHERG, H. De bortglomda oarna. Helsingfors 1955. 321. GiBBS, Lilian. A Contribution to the Montane Flora of Fiji. Journ. Linn. Soc. Bot. 39, 19C9. 322. ViSHKR, S. S. Tropical cyclones of the Pacific. Bernice P. Bishop Mus. Bull. 20, 1925. 323. Tropical cyclones and the dispersal of life from island to island in the Pacific. Amer. Naturalist 59, 1925. 324. Bergeron, T. The problem of tropical hurricanes. Journ. R. Meteor. Soc. 80: 344. 1950. 325. Grisebach, a. Die Vegetation der Erde. 2d ed. Leipzig. I, 1872, 11, 1884. 326. Skottsberg, C. On Scirpus nodosus Rottb. Acta Soc. pro Fauna et Flora Fenn. 72, 1956. 327. Santesson, R. The South American Cladinae. Arkiv for Bot. 30, 1942. — The South American Menegazziae. Ibid. 329. • Contributions to the Lichen Flora of South America. Ibid. 31, 1944. 330. JoHOw, F. Ueber die Resultate der Expedition nach den Islas Desventuradas (San Ambrosio und San Felix). Verhandl. d. deutsch, wissensch. Vereins zu San- tiago IIL Valparaiso 1898. 331. Bruggen, J. Fundamentos de la Geologia de Chile. Santiago 1950. 332. Martin, C. Landeskunde von Chile. Hamburg 1909. ;^^;^. Stephani, F. Bot. Ergebn. schwed. Exped. nach Patagonien und dem Feuerlande. 2. Die Lebermoose. K. Sv. Vet.-akad. Handl. 46, 191 1. 334. I^ERRV, E. W. Tertiary fossil plants from the Argentine Republic. Proceed. U.S. Nation. Mus. 73, 1928. 335. A Miocene flora from Patagonia. Johns Hopkins Univ. Stud., Geol. no. 6, 1925. Not seen, only a preliminary notice in Proceed. Nat. Acad, of Sci. 11: 7, 1925. 336. Frenguelli, J. Recientes progresos en el conocimiento de la geologia y paleo- geograffa de Patagonia. Rev. Mus. de Eva Peron (La Plata), N.S. IV, Geol., 1953- 337. l>a Flora f6sil de la regi6n del alto Rio Chalia en Santa Cruz (Argentina). Notas del Mus. de Eva Per6n (La Plata) XVL 98, 1953. 338. Florin, R. Die heutige und friihere Verbreitung der Gattung Acmopyle Pilger. Svensk bot. tidskr. 34, 1940. 339. Hooker, J. D. Observations on the Botany of Kerguelen Island. Philos. Trans. Roy. Soc. of London 168, 1879. 340. Skottsberg, C. Antarctic Plants in Polynesia. Essays in Geobotany in honor of W. A. Setchell. Berkeley 1936. 341. Plant succession on recent lava flows in the island of Hawaii. Goteb. K. Vet. o. Vitterh. Samh. Handl., 6:e foljden, ser. B, bd 1:8, 1941. 342. The Vegetation of Easter Island. Nat. Hist. Juan Fern. II, 1927. 343. Cooke, (i. H. Te pito te henua, known as Rapa Nui. Rep. Smithson. Inst. 1897. Rep. U.S. National Mus. Part I. Washington 1899. 344. MoTVKA, J. Lichenum generis Usnea studium monographicum. Lw6w 1938. 345. Brigham, W. T. An index to the islands of the Pacific ocean. Mem. Bishop Museum Vol. I: 2. Honolulu 1900. 346. Forster, (r. Florulae insularum australium Prodromus. Goettingen 1776. 347. A Voyage round the world . . . commanded by Capt. James Cook 1772, 3, 4 and 5. Vol. I. London 1777. 348. De plantis esculentis insularum oceani australis Commentatio botanica. Berlin 1786. Contents. Part I. The Juan Fernandez Islands. Chapter I. Composition, distribution and relationships of the Flora 193 Chapter II. Sources of the island flora as judged by the total distribution of the geographical elements, with special reference to the composition of the Chilean flora 256 Chapter III. Composition, distribution and relationships of the Fauna 292 Chapter IV, Continental and Oceanic islands 317 Chapter V. The Pacific Ocean and Continental Drift 325 Chapter VI. Transoceanic migration 331 Chapter VII. Biological characteristics of isolated islands 35 1 Chapter VIII, Evolution in Oceanic islands 363 Chapter IX. Juan Fernandez — oceanic or continental? 372 Chapter X. The Chilean coast line and the history of the Andes 380 Chapter XI. The Tertiary floras of Chile and Patagonia 386 Chapter XII. Antarctica as a source of the present circumpolar floras 389 Chapter XIII. The history of Juan Fernandez — a tentative sketch 394 Part 2. Easter Island. Chapter XIV. Composition, distribution and relationships of the Flora 406 Chapter XV. Composition, distribution and relationships of the Fauna 417 Chapter XVI. The biogeographical history of Easter Island 421 Bibliography 428 Printed October 26th, 1956. / P^L ^»' \-C QH 198 J8S$5 v.l BioMed Skottsberg, Carl Johan Frederik The natural history of Juan Fernandez and Easter Island. PLEASE DO NOT REMOVE CARDS OR SLIPS FROM THIS POCKET UNIVERSITY OF TORONTO LIBRARY