'^^5- OCCASIONAL PAPERS OF THE California Academy of Sciences No. 44, 154 pages, 50 figures, frontispiece November 8, 1963 GALAPAGOS ISLANDS A Unique Area for Scientific Livestigations A Symposium presented at the TENTH PACIFIC SCIENCE CONGRESS of the Pacific Science Association, held at the University of Hawaii, Honolulu, Hawaii, U.S.A., 21 August to 6 September 1961. Sponsored by the National Academy of Sciences, Bernice Pauahi Bishop Museum, and the University of Hawaii 5 f/iarine Bir ' ■ ical U L. I B R A R V ■^'*^N0V2 51963 WOODS ftOlE, IV!'^^^ San Francisco Published by the Academy 1963 Academy Bay, Isla Santa Cruz (Indefatigable). Sulivan Bay, Isla San Salvador (James). View from Isla Bartolome (Bartholomew). OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES No. 44, 154 pages, 50 figures, frontispiece November 8, 1963 GALAPAGOS ISLANDS: A Unique Area for Scientific Investigations A Symposium presented at the TENTH PACIFIC SCIENCE CONGRESS ofthe Pacific Science Association, held at the University of Hawaii, Honolulu, Hawaii, U.S.A., 21 August to 6 September 1961. Sponsored by the National Academy of Sciences, Bernice Pauahi bishop Museum, and the University of Hawaii Table of Contents PAGE Introduction. VICTOR Van Straelen 5 Bathymetry in the Galapagos Region. George Shumway AND Thomas E. Chase H The Climate of the Galapagos Islands. Leo Alpert 21 Archaeology in the Galapagos Islands. Thor Heyerdahl . . 45 Opportunities for Botanical Study on the Galapagos Islands. Henry K. Svenson 53 Biosystematic Studies on Galapagos Tomatoes. C. M. Rick 59 Composition and Relationship of the Terrestrial Faunas of Easter, Juan Fernandez, Desventuradas, and Galapagos Islands. Guillermo Kuschel 79 The Marine Shore-Fishes of the Galapagos Islands. Richard H. Rosenblatt and Boyd W. Walker 97 Evolutionary Patterns in Danvin's Finches. Robert I. Bowman 107 Protection and Conservation Problems on the Galapagos Islands. Misael AcosTA-SOLi's 141 Future Scientific Studies in the Galapagos Islands. Jean Dorst 147 -3 INTRODUCTION* Victor Van Straelen President Charles Darwin Foundation for the Galapagos Islands Brussels, Belgium The Galapagos Archipelago has been termed a "living laboratory of evo- lution." Today no other oceanic island harbors a greater number of endemic species of plants and animals than the Galapagos. Once the Hawaiian Islands had many more. Now most of them are gone— gone with the wind! Such a di- saster could have been prevented with proper management and without inter- fering with local economic interests. To those of us who are concerned with the historical development of knowledge, and more particularly of the biological sciences, it is striking that the most momentous turn in man's outlook on life and its forms was based on observations made in the tropics. From the eastern tropical Pacific area and some of the islands nearby, Charles Darwin drew his most significant conclu- sions; from the western Pacific area, in the Malayan Archipelago, Alfred Rus- sel Wallace came to nearly the same conclusions. The luxuriant plant and an- imal life of islands located in the oldest of all oceanic areas now existing, were incentives to answers on what appeared as insoluble enigmas. Wallace and Darwin became friends and in 1858 published simultaneous- ly, at a meeting of the Linnean Society of London, their views on what from thereon was to be known as the theory of descent and the theory of natural selection. These two great men followed their research in the same directions although on divergent subjects. Wallace's contribution culminated in a momen- tous book entitled Island Life. But in 1859 Darwin published his book "on the origin of species by means of natural selection, or the preservation of favoured races in the strug- gle for life." Itisthe result ofhis observations made during a voyage of near- ly five years on board the Beagle, from the 2nd of December, 1831, until the 29th of October, 1836. Darwin meditated on his observations for more than 20 years before he expressed openly his conclusions about them. He had visited many islands, among which I shall mention now only the Galapagos. * Presented at the TENTH PACIFIC SCIENCE CONGRESS of the Pacific Science Association, held at the University of Hawaii, Honolulu, Hawaii, U.S.A., 21 August to 6 September 1961, and sponsored by the NATIONAL ACADEMY OF SCIENCES, BERNICE Pauahi Bishop Museum, and the University of Hawaii. -5- 6 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers At a time when nearly all biologists accepted the constancy of species, Darwin convinced them of the contrary, not by a mere successful compilation of ideas, but by the presentation of new ideas and a wealth of new facts. Stirred by his observations on the Galapagos, Darwin collected a vast quan- tity of evidence and used it for the demonstration of organic evolution brought about by natural selection. Before that the idea of selection was only known to breeders who unconsciously practiced what we call today "artificial selec- tion." This year (1961) marks the 131st anniversary of the start of this mo- mentous adventure, and in September, precisely the 127th anniversary of Dar- win's contact with the Galapagos. He arrived there on the 15th of September, 1835, landing on Chatham Island (San Cristobal). Then on the 23rd on Charles Island (Floreana), on Albemarle (Isabela) on the 29th, on James (Santiago) the 8th of October, leaving the Galapagos forever on the 20th of October. His contact with the fauna of these islands was decisive for the perception of what would become the theory of descent or evolution. Darwin had an uncon- tested gift for new observation and for observing new facts by chance. This was the case with the species of birds he discovered on the Galapagos. For obvious reasons, Darwin's attention was drawn by the most conspic- uous elements of the Galapagos fauna: birds and reptiles. It is nearly always the case, vertebrates and phanerogams easily attract attention. But Darwin, a birdwatcher from his early boyhood, picked less conspicuous, but most in- teresting birds from the point of view of evolution, the finches (now known as Darwin's Finches), recently so masterly studied from an anatomical and ec- ological point of view by my friend Robert Bowman, Secretary for the Americas of the Charles Darwin Foundation. I shall dwell no longer on the spectacu- lar biota of Galapagos. Surely, by the irresponsible action of man, the flora and the fauna of the Galapagos are, in many respects, no more what they once were. What remains of the glorious pieces of architecture and art of ancient Greece and Rome? Were they not until recent times used as quarries providing building stone for the housing of barbarian invaders? But man can destroy beauty which can also be replaced by man, provided that he has the necessary genius. Never is he able to coin a new form of liv- ing being. He who takes that responsibility of destruction, even if he has no sense of doing so, cuts forever a link with a very remote past and an infinite chain of processes leading to an unforeseeable future. In Galapagos the native life forms and their associations are in mortal danger. Surely, we may already be grateful to the Ecuadorian Government for having responded favorably to the many calls for protection that scientific bodies have addressed to it for more than thirty years, such as the National Academy of Sciences of the United States, the Royal Society of London, the Academiedes Sciences de Paris, die Preussische Akademie des Wissenschaf- No. 44) VAN STRAELEN: GALAPAGOS SYMPOSIUM 7 ten, Academie royale de Belgique, and so on. UNESCO took on the problem from the very beginning of its inception as an international organization in the days when Sir Julian Huxley was its Director-General. In 1961 the General Assembly of the International Union for Biological Sciences met at Amsterdam, in the seat of the Dutch Royal Academy of Scien- ces, and on July 15 voted unanimously a resolution imploring the Ecuadorian Government to take steps for the rescue of the Galapagos biological treasures. On the 29th of September, 1961, the Xth Pacific Science Congress as- sembled at Honolulu voted unanimously a similar motion. Over 2750 scientists, coming from all areas surrounding the Pacific Ocean and from western Europe, participated in this action. But protection and conservation laws as enacted on two occasions for Galapagos by the Government of Ecuador are very difficult to enforce and they need understanding. Nature conservation is a technique that comes under ap- plied biology. This is the reason, among many others, why, under the sponsorship of UNESCO, the Charles Darwin Foundation for the Galapagos Isles was created in 1959, on the occasion of the centennary of the publication of "The Origin of Species." No more fitting memorial could be erected to Charles Darwin whose name stands alongside that of Isaac Newton. By mentioning Isaac Newton, I think of the rapid advances of the physi- cal sciences during the past 100 years, progress in understanding that far sur- passes the rate of increase of knowledge in the field of biology. A simple cen- sus of present living forms is still in its very beginnings. Perhaps no more than tv^enty-five per cent of lower plants and animals are known. Organisms, representing high ranking systematic divisions, are still discovered nowadays. Many organisms living in fresh or marine waters or between the grains of soil remain to be discovered. The behavior and life conditions, what incur present- day jargon we call ethology and ecology respectively, are known only approx- imately for a few vertebrates and insects. Why is this so? Because many branches of biology did not have the ben- efit of the strong cooperating organizations that made possible the incredible developments of physics in its broadest sense. Biologists are still the poor- est among the members ofthe scientist family. The future generations will be amazed by the neglect from which we are suffering. Does man not belong to that largest category among those that can be distinguished on earth and that are the living beings?. It is one ofthe aims of the Darwin Foundation to go into a close inves- tigation of the life communities existing on the Galapagos, provided this or- ganization obtains the resources needed. A very ambitious plan has been drawn up which includes many problems in the physical and geophysical sci- ences. Many years, and still more qualified people drawn from all nations, are 8 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers needed for its realization. Surely, as discoveries are made, the plan will be broadened. Everything points to the fact that speciation in the Galapagos is not restricted to vertebrates, but includes all groups of living beings present on the isles and in the adjoining seas. Until recently, investigations on the origin of diversification in marine organisms have been made on a very restricted number ofmollusks and fishes. There is much evidence that the seas bathing the shores of Galapagos will provide a most favorable opportunity for tackling these problems. The diver- sity of the physics of waters due to the confluence of cold and warm currents is surely at the origin of great differentiations in the environments. Variations in ecological factors are, evidently, responsible in the long run for morpho- logical variations, and as we all know, living beings until now, have been de- fined mainly by their shapes. Another aspect of investigations on sea-waters has recently been re- vealed. Up to now sea-water appeared to be a solution of mineral salts. There is much evidence that besides these salts a fair amount of organic matter is dissolved in the seas. Saturated and non-saturated hydrocarbons exist. In most of the waters of the old world this problem cannot safely be investigated on account of the numerous streams carrying enormous amounts of polluted waters into the oceans. This complication does not exist around Galapagos and therefore this region provides an ideal site for future investigations along this line. All over the world the bulk of life communities remain unexplored. Be- low the northern tropics, especially on the American continent, the wealth and variety of these communities is far greater than in any other part of the world at the same latitude. No continent possesses a greater wealth of plant and animal life than Central and South America. We know that until about a century ago, it was the same for the Caribbean islands, now covered with a cosmopolitan world of life that came in and destroyed biological communities as endemic as those of the Galapagos. On the South American continent life is far richer than in Africa. The Congo and Cameroon forests, the only green spots of importance still exist- ing in Africa, maybe mentioned as poor when they are compared with the Am- azonian forest in its broadest sense. The potentialities of these renewable resources are as yet unexplored. The Galapagos international undertaking cannot bear full fruit unless, first of all, it is considered by the Ecuadorian nation as its own enterprise. Without its understanding and its cooperation in earnest, nothing, in the long run, can be achieved. Besides the endless investigations we ought to con- sider the conservation techniques that are now well worked out in many coun- tries. The release of exotic species must be stopped all over the archipelago No. 44) VAN STRAELEN: GALAPAGOS SYMPOSIUM 9 and especially in those areas of Galapagos designated as wildlife reserves. In the latter, the exotics that turned feral must be controlled through exterm- ination or by removal. Also, the transplantation of native species from one island to another should be prohibited until the results of studies still to be done are known. Any defacement of the protected areas, due to irresponsible yachtsmen coming from abroad, must be stopped. I hope that the day will soon come when all our South American states will establish large natural reserves on their territories following the example of the Republic of Ecuador. A heavy responsibility rests on the shoulders of the South American re- publics with regard to the proper use and maintenance of the wealth of their territories. They are now confronted with the conservation problems and all that this means. Let them not follow the distressing examples of so many ter- ritories of the Old World. From my early childhood I had contacts with Latin America. It is only during the last ten years that I have been called upon to participate in certain scientific undertakings sponsored there by UNESCO. The philosophy of all the states of this continent is not unfamiliar to me. Expression of high ideals are easily discovered, even when they are hidden by a passing dark cloud, whose silver lining reveals the persistence of the source of enlightenment. Educational systems are an expression of the ideas of nations. Let the Latin American systems offer some chances for the understanding of nature and its conservation so as to make the gift of life more valuable and the men more worthy of the gift. BATHYMETRY IN THE GALAPAGOS REGION George Shumway U. S. Navy Electronics Laboratory San Diego, California and Thomas E. Chase U. S. Bureau of Commercial Fisheries Biological Laboratory San Diego, California Most of the bathymetric data presented here are from expeditions of Scripps Institution of Oceanography, based on sounding lines obtained while crossing the region during the course of other work (fig. 1). Especially use- ful sounding lines were obtained by Scripps expeditions SHELLBACK, DOL- PHIN, STEP-I, RISEPAC, and SWAN-SONG. Additional sounding lines have come from U.S. Navy ships participating inHIJUMP expedition. U.S. Navy Hy- drographic Office chart 1798 contains a number of isolated soundings in the vicinity of the islands which also were included. Because available data are not abundant and sounding information is randomly located, the charts pre- sented here must be considered only a preliminary synthesis. On all expeditions of Scripps Institution of Oceanography, echo sound- ings were obtained with AN/UQ-1 echo sounders and were recorded on Preci- sion Depth Recorders (except SHELLBACK, made before PDR's were used). The acoustic soundings were made with the assumption of an average sound speed in sea water of 4800 ft/sec. Corrections for true water sound speed have not been made. The remoteness of the Galapagos Islands from the important world trade routes and the incomplete status of the mapping of the ocean floors have left the bathymetry of the Galapagos region relatively unknown up to the present time. In the past decade sufficient data have accumulated to construct the charts presented here (figs. 2, 3, 5, 6, 7). A series of charts of the eastern tropical Pacific including the Galapagos region is being prepared by the U.S. Bureau of Commercial Fisheries and the Institute of Marine Resources, Uni- versity of California. * Presented at the TENTH PACIFIC SCIENCE CONGRESS of the Pacific Science Association, held at the University of Hawaii, Honolulu, Hawaii, U. S. A., 21 August to 6 September 1961, and sponsored by the NATIONAL ACADEMY OF SCIENCES, BERNICE Pauahi Bishop Museum, and the University of Hawaii. - 11 12 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers io*> I00< 95< 90« 85= 80< 75° Figure 1. Sounding lines in the east equatorial Pacific Ocean, from Scripps Institution of Oceanography and U. S. Navy cruises. Regional Setting The Galapagos Islands are a group of typical Pacific Basin basaltic volcanoes (MacDonald, 1949) about which there is little petrographic informa- tion. The volcanoes do not rise directly from the deep sea floor, but are perched on a platform elevated a thousand fathoms or more above the surrounding deep sea floor. This platform connects through a saddle on the east with the Car- negie Ridge (Shumway, 1957), and on the north it joins Cocos Ridge (Shumway, 1954). Typical fracture zone topography (Menard, 1955a, 1955b, 1960; Menard and Fisher, 1958) extends westward from the vicinity of Darwin Island for more than a thousand miles. In the past decade, largely through the efforts of H. W. Menard and R. L. Fisher of Scripps Institution of Oceanography, an overall picture of the bath- ymetry in the eastern Pacific Ocean has come to light. It reveals that the dom- inant topographic features, beyond the trenches at the continental margins. No. 44) SHUMWAY & CHASE: GALAPAGOS SYMPOSIUM 13 CREST OF EAST PACIFIC RISE (\\\\\1 FRACTURE ZONES RIDGES 40« 20« 20° 40« 160° 140° 120 Figure 2. Crest of East Pacific Rise and location of East Pacific fracture zones (after Menard, 1960, modified), are the East Pacific Rise, which is an elongated bulge of the sea floor ex- tending from the vicinity of Antarctica to the Gulf of Alaska, and a series of fracture zones running westward from the continents for distances up to 3000 14 CALIFORNIA ACADENiY OF SCIENCES (Occ. Papers miles (fig. 2) (Menard, 1960). The Galapagos Islands are located on the east- ern flank of the East Pacific Rise, and the Galapagos fracture zone cuts across the East Pacific Rise and apparently dies out on the western flank. The fact that Cocos Ridge meets the trend of the Galapagos fracture at an oblique an- gle does not seem to be unique, for the Clipperton fracture to the north has the Tehuantepec Ridge joining it obliquely, and the Easter fracture zone to the south has the Nasca Ridge joining it obliquely; all three ridges, Cocos, Tehuantepec, and Nasca have similar northeast-southwest trends. Thus the fracture zone and ridges associated with the Galapagos Islands are part of a regular system of topography in the eastern Pacific. Figure 3. Eastern portion of Galapagos Fracture Zone, showing location of transverse profiles. Galapagos Fracture Zone The existence of fracture zone topography west of Darwin Island was shown by Shumway (1954) on the basis of two echo sounding lines run across it by the SHELLBACK expedition; but when his manuscript was written in 1953, fracture zones had net assumed the importance geologically they now have. The matter was not pursued further at that time. Menard (1955) mentioned that on the basis of the limited information then available, a fracture zone running west from the region of the Galapagos Islands probably existed. By 1960 with more data available, Menard (1960) included a fracture zone in the vicinity of the Galapagos Islands on his bathymetric chart of the Pacific Ocean. The following description of the Galapagos fracture zone is preliminary and is based upon soundings obtained through 1961. These data consist pri- No. 44) SHUMWAY & CHASE: GALAPAGOS SYMPOSIUM 15 J STEP _U- J^J^, ^IkJIk SWAN SONG 96° W 95° W N 1 s 4 •M VN !■« ,-. oooru- SHELLBACK 1 IBOO- 1 y.A-^Vv— ^-— ^^ ^/. \ \ . — 1- 'l/V/" ~-^-\^vyi lil^ ^ - " 5-H 2-N I'N OOFM- EASTROPIC I90O- — ,-A /\— • — ^ ^ Z( ^'^ ■ lorw 4'N 3'N 2*N (• SHELLBACK A rtfl 1500- J J V iOOO- lV ^ — -^— ^ — ' '[lillll!l!lllllllil lililliillllllllillllllllllllllliilllllllllll IllHliliHillilliilii liliil liJll,!l,lll!llllllllllllllllli llllllillllll; ■iiiiUll lorw GALAPAGOS FRACTURE ZONE VERT EXAG < 20 Figure 4. Transverse profiles across eastern portion of Galapagos Fracture Zone. marily of transverse crossings of the zone. Large areas remain to be surveyed to ascertain the continuity of east-west trending ridges, troughs, and other structures (figs. 3, 4). For the sake of explicitness we refer to the zone of irregular topography running west from Darwin Island as the Galapagos fracture zone, and retain the established names Cocos Ridge and Carnegie Ridge for the related struc- tures to the east. This does not imply that the variously named features are structurally separate entities. Rather, it is probable that these features are structurally related and that they are part of the fracture zone. Galapagos Platform The volcanic islands of the Galapagos are perched on top of a platform whose top lies a thousand fathoms or more above the deep sea floor to the south (fig. 5). It does not seem appropriate to call this feature a ridge, for at the 700 fathom isobath it has a length of 270 sea-miles and a width of about 100 sea-miles, and isobathic contours along its southwestern slope have a marked convexity. To the east, the platform connects through a saddle with the western end of the Carnegie Ridge (figure 6 (Shumway, 1957). The low point of the saddle lies between 1200 fathoms and 1300 fathoms. 16 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers GALAPAGOS ISLANDS REGION Figure 5. Bathymetry in the vicinity of Galapagos Islands. On the south, southwest, and west sides, the Galapagos platform drops steeply to the deep-sea floor at more than 1800 fathoms depth. On the south- west side there is an elongate depression about 200 fathoms deeper than the sea floor farther to the southwest (fig. 5). Similar marginal depressions often are present around large seamounts, and a notable depression is found along the eastern and northeastern sides of the Hawaiian Ridge (Hamilton, 1957). The steep side of the platform lies close to the west and southwest sides of Fernandina Island and Isabella Island, with depths of 1700 fathoms being found within 6 or 8 sea-miles of the islands. On the northeast side of the platform the bottom drops off gradually to- ward the Panama Basin without complications. On the north and northwest sides, however, the topography is complex; it is in this region that the frac- ture zone, Galapagos platform, and Cocos Ridge meet (figs. 3, 5, 7). Unfor- tunately, not enough sounding data are available to allow the complex topo- graphy to be resolved in the detail desirable, A tongue of relatively deep water, i.e. greater than 1400 fathoms, lies between the north end of Isabella Island, Pinta Island, and Marchena Island. Darwin Island and Wolf Island are isolated from the Galapagos platform by about 60 sea-miles of water deeper than 1200 fathoms; in this respect they are not part of the Galapagos group. The upper surface of the Galapagos platform contains numerous irregu- larities that cannot be contoured adequately on the basis of available sound- ing data. No attempt has been made to depict contours shoaler than 700 fath- No. 44) SHUMWAY & CHASE: GALAPAGOS SYMPOSIUM 17 Figure 6. Bathymetry for Carnegie Ridge and the eastern end of the Galapagos Platform. oms (fig. 5). A very detailed survey would be necessary to determine the in- ter-island bathymetry. Cocos Ridge Cocos Ridge, running from the vicinity of Costa Rica to the region im- mediately north of the Galapagos Islands via Cocos Island was discussed by Shumway (1954). Since then additional sounding data which improve knowl- edge of the ridge have been obtained (fig. 7). Fisher's (1961) chart of the Middle America Trench shows that the northern end of Cocos Ridge is termin- ated by relatively deeper water at the eastern end of the Middle America Trench. The trench, however, does not extend east of the ridge. These facts suggest that the two structures have exerted some influence upon each other at their place of junction. A favorably directed sounding line, running in an east-west direction at about 5° N. latitude, was obtained by Scripps Institution of Oceanography ex- pedition DOLDRUMS. A bathymetric profile drawn from data obtained on this sounding line (fig. 8) reveals a domed central portion about 100 sea-miles in width flanked on the west by a steep escarpment about 500 fathoms high. To the east of the escarpment there is a steep-sided elevation, but apparently this is a local feature. The data suggest that the northwest flank of Cocos Ridge may be a zone of particular crustal instability where repeated volcanic activity has created 18 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers W> •W* COhw(IKJA<. rijMfA.tS 4 WSkt COCOS RIDGE Figure 7. Bathymetry for (^oros Ridge. a narrow ridge, portions of which rise higher than the main, central part of cos Island, is located along this northwestern flanking ridge. Co- SUMMARY The Galapagos Islands rise from an elongated platform ofeastwest trend which at the 700 fathom isobath is 270 sea-miles long and 100 sea-miles wide. South and west of this platform the sea floor drops steeply to depths greater than 1800 fathoms. The eastern end of the platform connects through a saddle with the Carnegie Ridge. On the northeast side of the platform the sea floor No. 44) SHUMWAY & CHASE: GALAPAGOS SYMPOSIUM 19 VERT EXAG X20 Figure 8. Bathymetric profile across Cocos Ridge along an east-west line at 5°N. drops off gradually toward the Panama Basin. To the north, the platform unites with the end of Cocos Ridge and with a fracture zone that extends westward from the vicinity of Darwin Island. The fracture zone that extends almost due west from Darwin Island con- tains topographic features typical of the other fracture zones which parallel it to the north. This distinctive topography, with high narrow ridges adjacent to deep troughs, extends at least 600 sea-miles west of Darwin Island; the zone is less well defined farther west, but there is evidence of it as far as 1600 miles west of Darwin Island. Literature Cited Fisher, R. L. 1961. Middle America Trench: topography and structure. Geological Society of America Bulletin, vol. 72, pp. 703-720. Hamilton, E. L. 1957. Marine geology of the southern Hawaiian Ridge. Geological Society of America Bulletin, vol. 68, pp. 1011-1026. MACDONALD, G. A. 1949. Hawaiian petrographic province. Geological Society of America Bulle- tin, vol. 60, pp. 1541-1596. Menard, H. W. 1955a. Deformation of the northeastern Pacific Basin and the west coast of North America. Geological Society of America Bullet in, vol. 66, pp. 1149-1198. Menard, H. W., and R. L. Fisher 1958. Clipperton Fracture Zone in the northeastern equatorial Pacific. Jour- nal of Geology, vol. 66, pp. 239-253. SHUMWAY, Geo. 1954. Carnegie Ridge and Cocos Ridge in the east equatorial Pacific. Jour- nal of Geology, vol. 62, pp. 573-586. 1957. Carnegie Ridge bathymetry. Deep-Sea Research, vol.4, pp. 250-253. THE CLIMATE OF THE GALAPAGOS ISLANDS* Leo Alpert Tropical and Desert Branch Earth Sciences Division Army Research Office, OCRD Washington, D. C Introduction Little information is available about the climate of the Galapagos Is- lands - and for that matter, about the entire Eastern Tropical Pacific Ocean Area - because of the scarcity of surface and upper air weather observations. Prior to World War II, the only weather data available were contained in the records of a few scientific expeditions that had visited the islands from time to time, and some sporadic observations of local inhabitants. Since the ex- peditions spent only a few weeks in any one locality and were primarily inter- ested in the fauna, flora, and geology of the islands rather than in the climate, no complete series of weather observations was obtained. In addition, no ser- ies of winds aloft or radiosonde observations had been made. The observations made by Wolf (1879, 1892, and 1895) during August to November, 1875, and May to July, 1878, have been the basis of all previous analyses of the weather and climate of the islands. The Galapagos expedition of the California Academy of Sciences spent one year, September 24, 1905, to September 25, 1906, in the archipelago, vis- iting each island at least once and some of the larger islands two or more times at different seasons of the year. Observations and notes concerning meteor- ological conditions during this period, and also the. botanical regions from which much valuable information can be inferred (Stewart, 1911), have gone almost unnoticed, not being utilized by either Knoch (1930) or Schott (1931, 1935, and 1938). During World War II, an airbase and weather station (0°27'S., 90°6'W., 10 feet msl) were established on Seymour Island (fig. 1). Seymour Island is small and "pear shaped," located just north of Santa Cruz Island, and sep- arated from it by a narrow channel less than one-half mile wide. Seymour is five miles long and three and one-half miles wide in its widest southern half. * Presented at the TENTH PACIFIC SCIENCE CONGRESS of the Pacific Science Association, held at the University of Hawaii, Honolulu, Hawaii, U. S. A., 21 August to 6 September 1961, and sponsored by the NATIONAL ACADEMY OF SCIENCES, BERNICE Pauahi Bishop Museum, and the University of Hawaii. -21- 22 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers 9 0'30' STATIONS- SEYMOUR MILES Figure 1. Topographic map of Seymour Island and Santa Cruz Island, The volcanic rock surface slopes gradually upward from a sandy beach in the west to a 200-foot cliff on the windward southeast. The highest peak on Santa Cruz Island (2,835 ft.) lies 12 miles SSW. of the former weatherstation on Seymour Island, and is the principal topographic feature in the immediate station vicinity. This mountain influences consider- ably the weather and climate of the whole of Santa Cruz Island. Half-hourly surface observations for aircraft operations were made from August, 1942, to August, 1945. They included only such elements as pres- sure, temperature, wind direction and speed, relative humidity, precipitation, sky condition, cloud type (amount and direction), and weather. No data were obtained on solar radiation, evaporation, transpiration, soil temperature and moisture, and dew. In addition to these surface observatioits, some aircraft observations were made during this period; two to four winds aloft observa- tions were made daily (Alpert, 1946), and, between February, 1943, and July, 1944, one radiosonde observation was made daily. When, in late 1945, the air- base on Seymour Island was turned over to Ecuador, the weather station was closed. No. 44) ALPERT: GALAPAGOS SYMPOSIUM 23 The only other series of weather observations of any consequence are those for the weather station (0° 54' S., 89° 30' W., 20 feet msl) on San Cristo- bal Island (figure 2). The observations cover the nine-year period, 1950 through 1958, and include only pressure, temperature, weather, precipitation, wind direction and speed, humidity, cloud type and amount, and storms. Plans are underway to establish observation stations on Seymour, Isa- bella, Florena, and Santa Cruz islands. A weather station is being planned for the Charles Darwin Research Station at Academy Bay on the south coast of Santa Cruz Island (figure 2). All of these weather stations are located near sea level and close to shore. No significant series of weather observations are available for the mountainous interior of any of the islands, where clima- tic conditions are known to differ considerably from those in the arid lowlands. Figure 2. Location of weather stations in the Galapagos Islands. The Dry Season Along the coast the year is divided climatologically into two distinct 24 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Figure 3. Patrol aircraft weather reports over the tropical eastern Pacific Ocean, No- vember 16, 1942. The intertropical convergence zone is shown by the dashed double line. The station model is outlined in Appendix A. (See Page 42) seasons, the rainy season lasting from January through April, and the dry season lasting the remainder of the year. During the dry season, the intertropical convergence zone is located No. 44) ALPERT: GALAPAGOS SYMPOSIUM 25 far north of the islands as shown in the example in figure 3, and exerts no direct influence on the climate. Divergency in the South Equatorial Current south of the islands causes upwelling of cold water which is carried past the southern islands. Figure 4 shows the August ocean surface temperature for the Eastern Pacific Ocean. A narrow belt of relatively cold water extends in an east-west direction between 5°S. and the Equator. Ocean surface tempera- tures increase rapidly north of the Equator. The southerly surface winds (see figure 3) are cooled in the lower layers by this cold water. Consequently, abnormally low air temperatures for an island station at sea level in the trop- ics, are recorded during the months of June through December. Table 1 shows the climatological data at Seymour Island. The mean daily maximum tempera- ture is 66° F. Figure 4. August ocean-surface isotherms for the eastern Pacific Ocean. The cooling of the lower atmosphere and subsidence in the South Paci- fic high pressure cell, shown in figure 5, contribute to the formation of a tem- perature inversion. In August, 1952, (Neiburger, 1958), radiosonde observa- tions from ships indicated that the height of the base of the inversion was at 600 meters among the southern islands at 1° South latitude, and increased to a height of 800 meters among the northern islands at the Equator (figure 6). Figure? shows two soundings; one on 31 March 1943, made during an extend- ed humid spell (lasting from 29 March to 2 April 1943, table 2) of the rainy season, and the other on 20 October 1943, during an extreme development of a temperature inversion in the dry season. Table 3 gives additional informa- 26 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers tion (relative humidity and mixing ratio) for these two soundings; and table 4 shows the winds aloft observations on these two days. This inversion is read- ily apparent in the dry season sounding of 20 October 1943, shown by the dashed line in figure 6. On the average, the inversion base is at 3,000 to 4,000 feet, and the top is at 5,000 to 6,000 feet. Compared to the rainy season, there is a marked "drying" of the entire air mass (lower specific humidity) especially immediately about the base of the inversion, which prevents moisture from diffusing aloft from the surface layers (table 3). The lifting and convective condensation levels increase in height and the air mass is stable for ordinary lifting and surface heating pro- cesses. Thus, convection and the formation of cumulus clouds, typical of the tropics, are dampened. Table 1. Clhnatological data at Seymour Island weather station. J V M A M J J A s O N D An. Average pressure 1000 mb+ (T 11.8 11.0 10.8 10.9 11.1 12.4 12.6 13.0 13.0 13.0 12.6 12.5 12.1 Mean Maximum temperature (2^ 86 86 88 87 86 83 81 81 80 81 81 83 Mean maximum temperature (2^ 72 75 75 75 73 71 69 67 66 67 68 70 Prevailing wind direction (2^ SE V. t; !•: SSE SSE SSE SSE SSE SSE SSE SSE Average wind speed (2^ 8 i 6 8 10 10 10 11 11 11 10 10 Average relative humidity (3_ 74 78 76 80 76 74 76 75 74 70 71 71 Total rainfall {X_ 0.81 1.39 1.06 0.67 T 0.01 0.01 0.01 T 0.01 T T 3.97 Prevailing charac- ter of rain- fall (4 RW RW RW RW RW RW-I. L L \. L L L Total numberdays with rainfall (J 8 9 6 6 4 4 9 8 ^ 2 4 6 73 Number of days with thunder- storms (5^ 0 <1 1 <1 0 0 0 0 0 0 0 0 Percent of time sky was overcast \1_ 23 9 12 13 9 26 29 28 30 23 29 24 (I- August 1942 - April 1944 (J- August 1942 - August 194J ) (2 - August 1942 - February 1945 (5^- January 1943 - August 194 5 (3^- January 1943- February 1945 No. 44) ALPERT: GALAPAGOS SYMPOSIUM 27 Table 2. Daily rciinfctll at Seymour Island weather station in March and April, 1^4 ^. Date Inches DATE Inches March 13, 1943 0.01 Aiiri 1, 1943 0.14 18 0.01 2 0.38 29, 0.41 13, 0.12 30, 0.33 23, 0.03 31, 0.11 24, 0.19 0.87 0.86 Stratiform clouds form below the base of the inversion in May (table 1), and are the typical cloud form from May through December. Low stratiform clouds are present throughout the day over the ocean, but over the islands, surface heating causes the clouds to break and dissipate during the afternoon. Generally, it remains clear over land until sunrise. The stratocumulus clouds form after sunrise, are most dense in the morning, and break in the afternoon. Occasionally, they may last throughout the day. The breaking of the strati- form deck over Seymour Island usually coincides with the shift in surface wind direction from the SSE. land breeze through W. to the WNW. sea breeze at the Seymour Island weather station. The sea breeze is especially well de- veloped because of the relatively large temperature difference between the Table 3- Radiosonde observations at Seymour Island weather station. 31 March 1943 20 Ociober 194 3 Pres. Elev. Temp. Re ative Mixing Pres. Elev. Temp. Re iative Mixing (Mb) (Feet) (°C) Humidity Ratio (Mb) (Feet) (°C) Humidity Ratio Code Percent (Grams/ Kg) Code Percent (Grams/ Kg) 400 -17 4 (40-49) 1.0 400 -12 1 (0-19) 0.8 439 -12 O (20-29) 0.8 434 -11 1 (0-19) 0.6 488 20,000 - 7 4 (40-49) 2.5 490 20,000 - 4 1 (0-19) 1.0 521 - 3 6 (60-69) 3.3 532 1 1 (0-19) 1.1 608 5 6 (60-69) 6.0 552 1 3 (30-39) 2.6 692 9 7 (70-79) 7.3 635 8 1 (0-19) 1.7 710 10,000 10 6 (60-69 7.8 710 10,000 13 O (20-29) 3.0 701 14 7 (70-79) 9.2 795 20 1 (0-19) 2.6 852 5,000 18 7 (70-79) 10.9 852 5,000 21 3 (30-39) 5.4 862 19 7 (70-79) 11.2 890 16 7 (70-79) 10.4 934 22 7 (70-79) 17.2 928 20 7 (70-79) 10.3 1011 surface 27 74 17.2 1011 surface 21 72 11.3 28 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Table 4. Winds aloft observations at Seymour Island weather station. Elevation 31 MARCH 1943 20 October 1943 (FEET) DIRECTION (360°) SPEED (MPH) DIRECTION (360°) SPEED (MPH) 30,000 25,000 120 140 12 8 20,000 15,000 14,000 12,000 250 70 60 20 1 7 8 8 10,000 9,000 8,000 7,000 6,000 40 60 70 50 80 13 12 9 4 7 170 200 9 4 5,000 4,000 3,000 2,000 1,000 Surface 100 90 80 80 70 170 7 7 9 5 8 8 250 270 200 180 180 140 2 3 6 13 16 10 cold ocean surface and the heated land surface. Table 1 shows that because of the stable conditions, drizzle (identi- fied by the "L" in table 1) rather than the rain shower (identified by the "RW") is the prevailing rainfall type. Over land, the drizzle generally falls between daybreak and noon on four to nine days each month, but rarely reduces visi- bility below 3 miles. The drizzle yields only traces to a hundredth of an inch of rainfall per month on the lowlands during the dry season. Thus, the low- lands are dry during these months. The base of the stratiform clouds is generally between 2,000 to 2,500 feet, and the top 3,000 to 4,000 feet. The mountain slopes, particularly on the windward southern side from a height of about 800 feet upwards, receive pre- cipitation in the form of drizzle, fog, mist, and heavy dew. This form of pre- cipitation, known locally as the "garua, "frequently continues for periods of several days without a break, and the higher mountain slopes are more or less continuously enveloped in clouds and fog. The meteorological elements for two typical dry season days are shown in table 5. Features to note are the cool range of temperatures; stratocumulus overcast and drizzle forming in the morning and breaking before noon; the south-southeast land breeze changing to a west-northwest sea breeze; the No. 44) ALPERT: GALAPAGOS SYMPOSIUM 29 relatively high wind speeds throughout the day; the typical stratocumulus and altocumulus.clouds; the southerly direction from which the stratocumulus clouds are moving; and the relatively high ceiling and good visibility. The Rainy Season The rainy season lasts four months, January through April. By January, as shown by the surface pressure map of April (figure 5), a major shift in the general circulation has taken place as the intertropical convergence zone APRIL OCTOBER Figure 5. Mean sea level pressure pattern in April and October. 30 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Table 5- Two typical dry scasoti days at Seymour Ishnid iveather station. Airway Vikathkr Report Seymour Island 16 October 1942 CO Celling (T (Hundred Ft.) Sky Condition (2 (Hundred Ft.) o C a t- 1 CO c o tn c c p u~ m J2 O Temperature and Dew Point CF) c c o 10 COPS 1-sc-s 1835 20S 71/62 SE 8 COPS 1-sc-s 1935 E 22 0 70/61 SSE' 14 COPS 10-sc-s 2035 E 22 B 69/62 SSE 11 COPS 7-sc-s 2135 228 69/62 SE 9 COPS 2-sc-s 2235 22S 68/62 SSE 9 COPS F-ac-u 2-sc-s 2335 E 25' B 68/59 SE 7 COPS F-ac-u 7-sc-s (1, (2 , (3 , (^ - See Table 8. moves southward to a position a few degrees north of the islands. In figure 8 the intertropical convergence zone is located just north of the islands. The movement and characteristics of the intertropical convergence zone in the Eastern Tropical Pacific Ocean Area are little known. Weather observations from the islands can provide data for the study of this important climatic con- trol of the Eastern Tropical Pacific Ocean Area. No. 44) ALPERT: GALAPAGOS SYMPOSIUM 31 Table 5- Continued Airway Weather Report Seymour Island 17 October 1942 c o ^ > c _ --; o o 2 c J H £ .■^ tL, C o 3 •" u en W^ ■? -^ n .;; 4^ C v OJ t~-i bc 1J 5 1^ ^ " c: o •-" a .5 "^ ^ 2 Q c/) a,' . — ' c c *j 4-t £ ^ T) -a E 3 >■ D nJ V) c c H ^ 5 a,' a; H Q cs is Remarks (4^ Clouds (4 0035 i: 25 B 68/60 SSE 6 COPS 8-sc-s 0135 E 22 f") 68/60 SE " COPS 10-sc-s 0235 E 22 0 68/59 SSE 9 COPS 10-sc-s 0335 E 20 0 68/60 SSE 7 COPS 10-sc-s 0435 E 20 0 68/60 S g COPS 10-sc-s 0535 E 20 o L- 66/61 S 0 COPS VSBY 4 S L- 10-sc-s 0635 E 22 0 68/61 SSE 8 COPS 10-sc-s 0735 E 22 o 69/61 SSE 10 COPS 10-sc-s 0835 E 25 o 70/61 S 10 COPS 10-sc-s 0935 E 27 0 L- 70/61 S 9 COPS 10-sc-s 1035 E 28 o 73/61 SSE 10 COPS 10-sc-s 1135 E 26 H 74/62 WNW 9 COPS 9-sc-s 1235 E 27 B 76/62 NW 15 COPS 7-sc-s 1335 26S 77/63 WNW 10 CTPS 5-sc-s 1435 26S 76/63 NNW 14 CTPS 3-sc-s 1535 27S 79/61 SSE 14 COPS 2-sc-s 1635 22S 75/61 WNW 12 COPS 5-sc-s 1735 E 25 B20S 73/61 NW 5 COPS 9-sc-s 1835 E 22 O20S 71/63 SSE 9 COPS RINOVC 10-sc-s 1935 E 22 B 70/62 SSE 12 COPS 9-sc-s 2035 E 22 0 69/63 SSE 12 COPS BINOVC 10-sc-s 2135 E 22 0 69/64 SSE 10 COPS BINOVC 10-sc-s 2235 E 22 n 69/64 s 13 COPS BINOVC 10-sc-s 2335 E 22 f) 70/61 SSE 12 COPS BINOVC 10-sc-s With the southward movement of the intertropical convergence zone to a mean position of T" to 2°N. latitude in February and March, the mean monthly pressure (table 1) decreases, reaching a minimum in March. The prevailing surface wind direction shifts from SSE. to E. The average wind speed decreas- es to less than 8 mph., and calms become relatively frequent. Aloft the wind direction is easterly at all levels and the wind speed is the lowest of the year, averaging about 10 mph. (table 4). The islands are in the "Doldrums." 32 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers With the shift in surface wind direction, the upwelling of cold water to the south of the islands ceases, and truly tropical temperature conditions pre- vail. In March, the warmest month, the mean daily maximum temperature at Seymour Island is 88° F., and the mean daily minimum is 75° F., quite a change from the September temperatures. During these months, the dry and almost barren volcanic rock surface of Seymour Island is strongly heated causing lo- cal convective currents (figure 9). Dust devils are observed in the afternoon, some extending to 200-300 feet above the island. There is a noteworthy difference in the typical sounding for the rainy season, shown by the solid line, compared to the dry season, shown by the dashed line in figure 7. Temperature increases in the lower layer. Because of the heating of the lower layers and the decrease in subsidence from the outflow of the South Pacific high pressure cell, which has been displaced southward (figure 5), the inversion present during the previous dry season months is dissipated (table 3). Figure 6. The topography (hundreds of meters) of the base of the inversion during the cruise of the Horizon, June and July, 1952, (Shellback Expedition). No. 44) ALPERT: GALAPAGOS SYMPOSIUM 33 Convection now carries moisture aloft above the former (dry-season) in- version level, and a considerable increase in moisture is recorded at all lev- els as shown in table 3. The air mass is generally conditionally stable, and is readily made unstable by a moderate amount of either heating or mechani- cal lifting (such as is induced by the flow of air against the mountains). Table 6. Wean monthly rainfall at Seymour Island weather station. YEAR JAN. FEB. MAR. APR. MAY JUNE JULY AUG. SEP. OCT. NOV. DEC. ANNUAL 1942 T(l) T T T T 1943 0.03 1.40 0.07 0.07 T T 0.03 0.04 T 0.02 T T 3.26 1944 0.29 0.68 2.31 1.13 T 0.03 T T T T T T 4.44 1945 2.10 2.10 T na T T T T T (1) T =Trace Cumulus clouds first appear over the mountains in January, and almost daily during the afternoon in the rainy season, convectionally induced rain showers and, at times, thunderstorms form over the mountains. Figure 10 shows a cumulo-nimbus cloud over the mountains on Santa Cruz Island. Table 1 shows an average of one thunderstorm at Seymour Island in February. Heavy local showers may be recorded on the mountain slopes, which, in the opinion of the author, may receive more than 50 inches of rainfall annually. As the showers are carried away from the mountains and their sustaining upslope cur- rents, they dissipate, frequently before reachingthe leeward coast. Figure 11 shows a rain shower on Santa Cruz Island. When a shower is carried across Table 7. Mean annual rainfall at San Cristobal weather station, San Cristobal Island (00° 54' S. 89° 5'^' U', 6 meters). YEAR JAN. FEB. MAR. APR. MAY JUNE JULY AUG. SEP. OCT. NOV. DEC ANNUAL 1950 0.0 0.06 0.50 0.0 0.02 0.0 0.0 0.04 0.33 0.05 0.19 0.25 1.46 1951 3.94 5.04 1.34 3.24 0.60 0.67 2.97 0.29 0.43 0.27 0.09 0.29 19.23 1952 0.87 0.75 0.06 0.0 1.24 0.05 0.09 0.32 0.12 0.47 0.04 0.55 4.54 1953 7.05 19.17 3.39 18.03 6.67 0.09 0.10 0.50 0.29 0.23 0.24 0.22 56.08 1954 0.37 1.75 3.58 0.0 0.02 0.02 0.24 0.21 0.85 0.18 0.20 0.26 6.98 1955 0.33 7.69 3.43 0.47 0.08 0.11 0.26 0.29 0.27 0.18 0.27 0.18 13.58 1956 0.17 4.57 11.46 9.21 0.05 0.25 0.32 0.32 0.26 0.17 0.34 0.03 27.22 1957 0.14 10.16 13.62 7.01 1.61 0.25 0.21 0.37 0.18 0.43 0.58 2.95 37.57 1958 1.65 4.41 3.60 0.56 0.05 0.06 0.07 0.05 0.09 0.17 0.14 0.28 11.20 A\'G 1.62 5.96 4.54 4.29 1.16 0.17 0.48 0.26 0.23 0.24 0.23 0.55 19.81 34 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Seymour Island, rainfall may be recorded at the weather station. These show- ers on Seymour Island are shortlived, local, and cover a small area. At times, squalls form over the water during the night and produce a short, light shower when they move across the islands during the early morning. Thus, even in the so-called rainy season, the coastal lowlands are light- ly watered, and are arid. The total rainfall (tables 2 and 4) forthe four months of the rainy season at Seymour Island (according to the short three-year rec- ord) averages only 3.93 inches. The heaviest monthly rainfall recorded at the station was 2.31 inches in March, 1944. In contrast, 0.07 inches was recorded in March, 1943, and only a trace was recorded in March, 1945. It is surprising that the heavier rainfall recorded among the islands in the "EI Niiio" year 1943 (see below). 0 10 20 30 JOCTOBER ARCH 600 700 800 900 1000 TEMPERATURE CO Figure 7. Rainy (31 March 1943) and dry (20 October 1943) season sounding at Seymour Island weather station. No. 44) ALPERT: GALAPAGOS SYMPOSIUM 35 Figure 8, Patrol aircraft weather reports over the tropical eastern Pacific Ocean, March 8, 1943. The intertropical convergence zone is shown by the solid double line. The station model is outlined in Appendix A. (see Page 42)- 36 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Table 8. Tu'o typical rainy season days at Seymour Island weather station. Airway weather Report Seymour Island 14 February 1943 in c Ceiling (T (Hundred Ft.) Sky Condition (2 (Hundred Ft.) 1^ o n C8 u I.- [en) a o [0 > o c _o o 3 t^ w O Temperature and Dew Point (°F) n _o tj t- s c a: a _E -o OJ 0) D. m -a c Remarks (4, 5 Clouds (4 0035 E150 B/ 80/73 ENE 1 7-ac-e 0135 E150 B/ 79/74 NNE 1 7-ac-u 0235 E150 B/ 79/73 W 1 8-ac-u 0335 s/ 79/73 sw 2 150S 5-ac-u 0435 808 78/72 wsw 3 2-as-u 0535 SOS 78/72 s 1 3-as-u 0635 E 80 B 78/72 s 3 8-ac-u 0735 E120 B/30S 78/74 ssw 1 8-ac-se 1-cu-n 0835 E120 B/30S 81/74 sw 2 CTPS 6-ac-u 3- cu-n 0935 E120 B/30S 82/74 WNW 4 CTPS 3-ac-u 3- cu-n 1035 E120 B/30S 84/74 w 6 CTPS (1) (4 2-ac-u 4- cu-e 1135 E120 B/30S 83/73 NW 7 CTPS RW-S 2-ac-u 4-cu-e 1235 E120 B/30S 85/74 NNE 7 CTPS RW-SE (1) 2-ac-u 4- cu-e 1335 S/30S 86/73 ENE 11 CTPS RW-S (2) 1-ci- u 1-ac-u 2- cu-n 1435 E120 B/30S 86/72 E 8 CTPS RW-S (2) 6-ac-e 1- cu-e 1535 EllO O/30S 87/73 E 8 BINOVC (1) VSBY 6 S RW 8-ac-e 2-cu-e 1635 E120 O/30S 86/74 E 8 BINOVC (1) RW-S 8-ac-e 2- cu-e 1735 E130 B/30S 85/73 E 7 (1) RW-S 6-ac-u 2-cu-e 1835 E130 O/30S 84/72 E 6 BINOVC (3) 9-ac-e 1-cu-e 1935 E130 B/30S 83/70 SSE 12 COPS 6-ac-u 1- cu-s 2035 S/30S 82/71 SSE 12 COPS 150S 4-ac-e 1-cu-u 2135 S/30S 81/71 SSE 11 COPS 150S 1-ac-e 1- cu-u 2235 s/ 80/71 SE 12 COPS 160S 4-ac-se F-cu-u 2335 s/ _ 79/73 SSE 12 160S 5-as-u (T E=estimated. "^ (2 0=overcast, B=broken, S=scattered, C=clear. (1) Towering cumulus all quadrants. (3 Visibility unlimited entire period. (J BINOVC breaks in overcast COPS clouds over peaks to south, CTPS clouds topping peaks to south: Clouds; number of tenths of the layer, kind, direction from which the cloud is moving, (ac ^altocumulus, as =altostratus, ci=cirus, cu=cumulus, sc = stratocumulus) (2) Towering cumulus southwest quadrant. (3) Towering cumulus north quadrant. (4) Towering cumulus north semi-circle. (5) Towering cumulus northwest quadrant. No. 44) ALPERT: GALAPAGOS SYMPOSIUM 37 Table 8. Continued. Airway weather Report Seymour Island 15 February 1943 c 1-1 ) E Ceiling (T (Hundred Ft.) Sky Condition (2 (Hundred Ft.) Weather and or Obstruction to Visi( Temperature and Dew Point (°F) c o o u Q -□ c E D. m -a c Remarks ^, 5 Clouds (4^ 0035 ElOO 0/ 80/73 s 4 10-as-u 0135 -s/ 79/72 sw 4 5-as-u 0235 ElOO B/ 79/73 C 9-as-u 0335 ElOO B/ 79/73 NE 1 7-as-u 0435 ElOO B/ 78/73 E 2 7-as-u 0535 ElOO B/ 78-73 E 1 7-as-u 0635 8/ 78/74 SSE 2 lOOS 5-as-u 0735 E150 B/30S 78/74 ESE 5 COPS (1)(4 3-ac-u 3-cu-e 0835 E150 O/30S 79/73 SSE 7 COPS BINOVC (2) 2-ci-u 5-ac-s 3-cu-e 0935 E150 O/30S 81/73 SE 6 COPS BINOVC (4) 1-ci-u 6-ac-e 3-cu-s 1035 E150 O/30S 84/71 E 6 CTPS (4) 6-ac-ne 4-cu-e 1135 E150 O/30S 84/72 ESE 12 CTPS (4) 6-ac-ne 4-cu-e 1235 E150 O/30S 87/76 E lOE COPS RW-S 6-ac-n 4-cu-e 1335 E150 O/20S 84/75 E 18E COPS (3) RW-S 6-ac-n 4-cu-e 1435 E 25 0/B 82/72 ESE 11 COPS \^) rt«-3 4-as-n 6-cu-e 1535 E 25 0/B R- 81/74 ESE 11 COPS (3) 4-as-n 6-sc-e 1635 E120 0/25S 80/73 E 10 COPS 6-ac-u 4-sc-e 1735 E120 0/25S 80/73 E 10 COPS 6-ac-u 4-sc-e 1835 EllO O/30S 80/72 E 12 BINOVC 9-ac-u 1-sc-e 1935 EllO 0/ 79/73 E 6 BINOVC 10-ac-e 2035 EllO 0/ 79/72 E 6 10-ac-e 2135 EllO B/ 79/72 ESE 5 9-ac-u 2235 E180 B/ 79/72 SE 4 6-ac-e 2335 E180 B/ 78/74 ESE 1 6-ac-e is not reflected in the rainfall record at the Seymour Island weather station. A similar pattern of low rainfall, less than 20 inches, is evident in the raintall records at San Cristobal, shown in table 7. The rainfall and rainy season are very irregular. Virtually rainless years are not unknown on the lowlands. The years 1906 (Svenson, 1946), 1930 (Sven- son, 1946), and 1950 (table 7), were almost rainless. During the nine-year period, 1950-1958, San Crist6bal (table 7) recorded only 1.46 inches of rain- fall in 1950. During the four rainy season months of that year, no rainfall was 38 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Figure 9. Scattered cumulus clouds over Santa Cruz Island (background) on 7 March 1943. Arid surface of Seymour Island in foreground, Seymour Island airfield and former weather sta- tion are located south of the bay in center of photograph. recorded in January, 0.06 inches in February, 0,50 inches in March, and only a trace of rainfall in April, In contrast to the virtually rainless years that occur from time to time, such as 1950 at San Cristobal, unusually heavy rains may fall when the "El Nino" phenomenon is well developed. This was the case only three years later at San Cristobal, which recorded 56,8 inches of rainfall in 1954, Of this rainfall, 19.17 inches fell in February, and 18.3 inches in April. Even May, usually a dry month, recorded 6.67 inches of rainfall that year. Rainfall was also heavy at San Cristobal in 1957, amounting to 37.57 inches. The "El Nino" rainfall is known to have affected the island at least in 1891 (Agassiz, 1892), 1925 (Beebe, 1926), 1929 (Svenson, 1946), 1953 (table 7), and 1957 (table 7). It may have occurred there in other years. For example, the "El Nino" rainfall has been recorded on the west coast of South America in 1828, 1845, 1864, 1871, 1877-78, 1891, 1904, 1918, 1925-26, 1929, 1932, 1939, 1941, 1943, 1953, and 1957, The "El Nino" appears to be associated with a southward displacement of the intertropical convergence zone from its normal position so that the islands are fully under its influence. The pheno- No. 44) ALPERT: GALAPAGOS SYMPOSIUM 39 men on is little understood, and is worthy of further research, which would be aided considerably by surface and upper air observations from the islands. The meteorological elements of two typical rainy season days are shown on table 8. Noteworthy features of table 8 are the tropical range of tempera- tures; the rain showers forming at noon to the south over Santa Cruz mountains figure 11); the relatively low wind speeds; the typical cloudscape of low cumulus and altocumulus; the easterly direction from which the cumulus clouds are moving; the relatively high ceiling and good visibility. Rainfall and Vegetation Since there are no other places on the islands at which rainfall obser- vations have been made, the areal distribution of rainfall cannot be mapped reliably. However, the character of the vegetation on the islands is largely controlled by the rainfall. Thus, the areal distribution of the vegetation re- flects the areal rainfall distribution, which in turn is dependent to a great de- gree upon exposure and elevation. However, there are great differences in the elevation at which different vegetation regions begin and end on the different islands, and on the sides of a given island (Stewart, 1911). It seems likely that the size of the island and the degree of the slope are involved. Figure 10. Cumulonimbus cloud over the Santa Cruz mountains on 4 March 1943. No rain- fall was recorded at the Seymour Island weather station on this day. 40 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Four botanical regions, the dry transition, moist, and grassy, can be recognized above the strand vegetation, which forms a narrow belt along the shores of the islands in certain places. These botanical regions identify two rainfall zones, the dry zone and the moist zone. i Figure 11. Rainshower over Santa Cruz Island during the latter part of February, 1943. Runway of Seymour Island airfield in foreground. No rainfall was recorded at the Seymour Island weather station on this day. The Dry Zone From the dry coast, which receives from five to twenty inches of rain- fall a year, to the 400-1,000 foot level (and up to 1,500 feet on the north side of James and Santa Cruz islands), the growth is a dull-grey, sparse and scrub- by thorn-forest, consisting of scattered dwarf deciduous trees between which grow coarse grasses, low thorny bushes, and cacti which sometimes attain a height of thirty feet (figure -9). The dry zone covers the major portion of the surface of all the islands. Since Harrington, Bindloe, Culpepper, Hood, Tower, and Wenman islands, as well as Seymour Island, do not attain an elevation of 1,000 feet above sea level, they are arid, and support xerophytic plants typi- cal of the dry region vegetation. During the rainy season, the coastal vegetation has a deceptive light- green appearance from the distance. Closer inspection reveals that only the terminal twigs and end branches are in leaf; underneath, the woody stems are as bare as in the dry season. No. 44) ALPERT: GALAPAGOS SYMPOSIUM 41 The vegetation was green down to the water's edge during the "El Nino" period of 1891 and 1925. Agassiz (1892), who visited the islands dur- ing the rainy season months of February through April, 1891, wrote, "This year quite heavy rains extended to the very level of the sea, a somewhat un- usual state of things. I could not help contrast the green appearance of the island, covered as they were by a comparatively thick growth of bushes, shrubs and trees, with the description given of them by Darwin, who represented them in the height of the dry season in September, 1835, as the supreme expression of desolation and barrenness." The Moist Zone Between about 1,000 and 3,000 feet above sea level on the mountain slopes, the flora is more abundant and decidedly mesophytic in character. This is a result of the heavier rainfall in the rainy season, and the availability of moisture from drizzle, fog, mist (garua), and dew during the dry season, es- pecially in the southern (windward) slopes. Here the moist zone is 100-200 feet lower than on the northern (leeward) slopes. Between the dry-region ve- getation and the moist-region vegetation, a transition region is present being composed of a mixture of xerophytic plants from the dry region vegetation and the more hardy of the mesophytic plants from the. moist region. This moist region vegetation is characterized by forests. Some trees are two feet in diameter. The undergrowth is often dense and resembles that of the moist tropics, the rain-forest type being closely approached in places. During March, 1943, the lower boundary of the moist zone could be clearly seen by the author on the slopes of Santa Cruz Island because of the darker green color of the rain forest than of the thorn forest. Above 1,500 feet in certain places, the forests give way to meadows of long perennial grasses and ferns. Except on protected places, trees are al- most entirely lacking. The greater speed of the wind at the higher elevations combined with a somewhat smaller amount of precipitation is probably the rea- son for the absence of trees (Stewart, 1911). The drying power of the wind, and the effect of the wind on the struc- tural form of the plants is marked in the upper exposed parts of the islands. In addition, these lie near the tops of the mountains. These are often clear, while a few hundred feet below, the mountain side may be entirely enveloped in clouds and fog. The soil at the top of the mountains has been observed to be dry and dusty, whereas at the same time a little below the top, it was moist, or even muddy. Only on Albemarle, San Cristobal, and Santa Cruz islands is a grassy region well developed. 42 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Conclusions The available data indicate that the climate of the Galapagos Islands differs from that of most islands near the Equator. These differences are evi- dent in the surprisingly low air temperatures, the inversion and associated stable stratiform cloud and precipitation forms, the prevailing sparse rainfall of the lowlands, and the side extremes of rainless years and years of abun- dant rainfall when the "El Nino" phenomenon appears. Weather observations from the islands could contribute materially to the understanding of the climate of the tropical Eastern Pacific Ocean Area, and particularly of the intertropical convergence zone. Additional series of weather observations, including a more complete coverage of the climatic elements, are needed in order to present a better picture of the climate and to analyze the inter-relationship of the climate and the poorly known biota. Thus, weather observations and climatic studies should be part of the basic research program of the Charles Darwin Research Station. Appendix A Station Model for Synoptic Chart (The station model used in fig- ures 3 and 8 is outlined below) Hb (N) Ch Ht (N) Hb (N) Cm Ht (N) TT B V O DD— n WW GG ¥ Hb (N) Cl Ht (N) Fp W B Flight altitude in hundreds of feet. Cl Low cloud type, international symbol. Qvi Middle cloud type, international sym- bol. Ch High cloud type, international symbol. DD Wind direction at flight altitude. F Wind speed at flight altitude, a half barb equals a Beaufort Force of one unit. Fp Frontal passage: i marked wind shift; >X marked incident or end of turbulence; O marked temperature change (not with altitude); begin- ning orend of precipitation; change in cloud forms. GG Greenwich time of observation. Hb Height (in hundreds of feet) of base of low, middle, or high clouds above sea level. Hj Height (in hundreds of feet) of top of low, miiidle, or high clouds above sea level. N Amount (in tenths) of low, middle, or high clouds. TT Temperature in °C. at flight altitude. V Visibility at flight altitude, the min- imumvalues ofthe international code. WW State of current weather, international symbols. W State of past weather, international symbols. Note that Hg is reported when the plane is flown below the cloud deck, and Hy is reported when the plane is flown above cloud deck. When Hg is reported and Ht is not reported, N is entered in place of Ht; when Ht is reported and Hb is not reported, N is entered in place of Hb . I I No. 44) ALPERT: GALAPAGOS SYMPOSIUM 43 Literature Cited Agassiz, a. 1892. General sketch ofthe expedition ofthe Albatross from February to May, 1891. Bulletin of the Museum of Comparative Zoology, vol. 23, pp. 1-89. Alpert, L. 1945. The intertropical convergence zone of the eastern Pacific region (I). Bulletin of the Americ an Meteorological Society, vol.26, pp. 426-432. 1946a. The intertropical convergence zone of the eastern Pacific region (II). Bulletin of the American Meteorological Society, vol. 27, pp. 15-29. 1946b. The intertropical convergence zone of the eastern Pacific region (III). Bulletin of the American Meteorological Society, vol. 27, pp. 62-66. 1946c. Weather over the tropical eastern Pacific Ocean, 7 and 8 March, 1943. Bulletin of the American Meteorological Society, Vol.27, pp. 384-398. 1946d. Atmospheric cross-sections of the stratus zone of the tropical eastern Pacific Ocean. Transactions of the American Geophysical Union, vol. 27, pp. 800-812. 1948. Notes on the areal distribution of annual mean rainfall over the tropi- cal eastern Pacific Oc e an. Bulletin of the American Meteorological Society, vol. 29, pp. 38-41. Beebe, W. 1926. The Arcturus adventure, an account of the New York Zoological Soci- ety's first oce anographic expedition, xix + 439 pp., New York: G.P. Putnam's Sons. Darwin, C 1897. Journal of researches into the natural history and geology of the coun- tries visited during the voyage of H.M.S. Beagle round the world, under the command of Capt. FitzRoy, R.N.,x+ 519 pp., New York: D. Appleton and Company. Ecuador, Republica del 1945. Boletin Me teorologico , No. 2. Quito: Servicio Meteorologico del Ecuador. 1950. Boletin Meteorologico, No. 3. Quito: Servicio Meteorolo gico d el Ec uador. 1954. Boletin Meteorologico, No. 4. Quito: Servicio Meteorolo gico del Ec uador. 1951. Boletin Meteorologico dela Armada del Ecuador, No. 1. Quito: Servicio Meteorologic a de la Armada. 1953. Boletin Meteorologico dela Armada del Ecuador, No. 2. Quito: Servicio Meteorologic a de la Armada. 1955. Boletin Meteorologico dela Armada del Ecuador, No. 3. Quito: Servicio Meteorologic a de la Armada. 1959. Boletin Meteorologica, No.l. Quito: Direccion General de Me teorologia. Knoch, K. 1930. Klimakunde von Siidamerika. In Handbuch der Klimatologie, Band 2, Teil G, pp. 123-125. Berlin: Neiburger, M. 1944. Research paper No. 19, U.S. Weather Bureau, Washington. 1958. Final report, subtropical Pacific meteorological project. Dept. of Mete- orology, University of California, Los Angeles. 44 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers POSNER, G. S. 1957. Studies of ichthyology and oceanography off coastal Peru. The Peru current. Bulletin of the Bingham Oce anographic Collection, vol. 17, pp. 106-155. SCHOTT, G. 1931. Der Peru-Strom un seine ndrdlichen Nachbargebiete in normaler und anor- maler Ausbildung. Annals of Hydrographic and Maritime Meteorology, vol.59, pp. 161-159, 200-213, 240-253. 1935. Geographie des Indischen und Stillen Ozeane. Hamburg: Deutsche See- warte. 1938. Klimakunde der Siidse e-Inseln. In Handbuch der Klimatologie, Band 4, Teil T, pp. 37-40. Berlin: Stewart, A. 1911. A botanical survey of the Galapagos Islands. Proceedings of the Cali- fornia Academy of Sciences, ser. 4, vol. 1, pp. 7-288- SVENSON, H. K. 1946. Vegetation ofthe coast of Ecuador and Peru and its relation to the Gal- apagos Archipelago. American Journal of Botany, vol. 33, pp. 394-426. SVERDRUP, H. U. 1942. Oceanography for meteorologists, 246 pp. New York: Prentice-Hall, Inc. U. S. Hydrographic Office. 1944. Monthly surface temperature charts of the south Pacific Ocean. Hydro- graphic Office Miscellaneous Publications no. 10,532, Washington. WOLF, T. 1879a. Apuntes sobreelclima delas Islas Galapagos, segun las observacione s hechas en los meses de Agosto a Noviembre de 1875. Boletin del Observatorio Astronomico de Quito, No. 3. 1879b. Ein Besuch der Galap agos-Inseln . Sammlung von Vortragen, vol. 1: H eidelberg. 1892. Geografia y Geologia del Ecuador. Publicada per orden del Gobierno de la Republica. Leipzig, pp. 469-493. 1895. Die Galapagos-Inseln. Verhandlungen der Gesellschaft filr Erdkunde, vol. 22, pp. 246-265. ARCHAEOLOGY IN THE GALAPAGOS ISLANDS* Thor Heyerdahl Laigueglia, Italy Archaeological investigations of the Galapagos group have until recent- ly been neglected on the assumption that the area has been outside the range of aboriginal craft from either South America or Polynesia. It is noteworthy, however, that the Galapagos group was considered within the reach of abori- ginal craft from Peru and Ecuador by observers from the 16th to the 19th cen- tury who were personally familiar with guara-operated balsa rafts, whereas the confidence in this remarkably ingenious water vehicle disappeared with the craft itself at the turn of the present century. Discussions on the possibility of pre-Spanish visits to the Galapagos have all admittedly been biased by the writers' attitude toward balsa rafts. When Miguel Cabello de Balboa and Pedro Sarmiento de Gamboa inde- pendently recorded the 16th century Inca versions of Inca Tupac Yupanqui's enduring ocean voyage by balsa rafts to distant islands in the Pacific, they were both personally familiar with the type of rafts in question, which they also describe. Although Polynesia was still unknown to Europeans, Bishop de Berlanga had by then drifted so far out as to discover the Galapagos group, and Balboa suggested that these were perhaps the islands visited by the Inca's armada of rafts. Sarmiento de Gamboa, however, who was himself a keen navi- gator, inquired about the old sailing directions still preserved among some of the Peruvian raftsmen, and concluded that the inhabited islands known to the coastal raftsmen and only revisited by Inca Tupac's armada were in the South Pacific on a line west southwest from Callao and at a distance of about 600 leagues (2,400 miles). He was so confident in this specific position that he talked the Viceroy into organizing the first Mendana expedition which was ac- tually sailing straight into the waters immediately surrounding Easter Island when Mendana, to Gamboa's disgust, altered course and, instead, discovered other islands further away from Peru, first Melanesia, and, on a subsequent voyage, Polynesia. * Presented at the TENTH PACIFIC SCIENCE CONGRESS of the Pacific Science Association, held at the University of Hawaii, Honolulu, Hawaii, U.S. A., 21 August to 6 September 1961, and sponsored by the NATIONAL ACADEMY OF SCIENCES, BERNICE Pauahi Bishop Museum, and the University of Hawaii. -45- 46 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers The remarkable capacity of balsa sailing rafts and the expertness of their crews in navigation were unanimously praised by the contemporary chron- iclers, including Saamanos, Xeres, Andagoya, Oviedo, Zarate, Las Casas, Bal- boa, Gamboa, Inca Garcilasso, Benzoni,and Cobo. In 1619 the Dutch admiral Spilbergen had his whole fleet atPayta, Peru, supplied with dried fish from a sailing raft that had been out fishing for two months in the open ocean between Payta and the Galapagos group. The raft is illustrated with a native crew navigating with characteristic guarahoards sunk between the logs fore and aft. In 1680 the buccaneer Captain Sharp cruised in the local waters, trying as well to land in the Galapagos. His sailing vessel first followed the coast towards Peru, but turned into the open ocean off Punta Parina to avoid being detected by the Spaniards. Out there, where the impact of the Humboldt Cur- rent strikes out towards the Galapagos, and in the midst of what the buccan- eers describe as a very stiff off-shore gale, they encountered a merchant bal- sa raft under sail. Their own pilot advised them not to meddle with its native crew, "for it was very doubtful whether we should be able to come up with them or not..." We learn from the same early buccaneer record that these ab- original balsa rafts sail "excellently well," and that some are so big as to carry two hundred and fifty packs of meal from the valleys of Peru to Panama without wetting any of it. In 1736 two Spanish naval officers, Juan and Ulloa, made the first tech- nical study of the ingenious guara method which permitted the Indians to steer their rafts into the open ocean irrespective of the direction of the winds. Ar- chaeological specimens of guara, dating back to pre-Inca times, are still pre- served in desert graves from the Chimu area and as far south as Paracas and lea in south central Peru, and, ethnographically, guara were commonly in use in northern Peru and Ecuador during Juan and Ulloa's investigations in the Guayas region. They reported that Ecuadorian balsa rafts, from 75 to 90 feet long, with entire families onboard, and often a cargo of 20 to 25 tons, resisted the rapidity of the currents in the open ocean off Puna Island and northern Peru, and added: "...but the greatest singularity of this floating vehicle is, that it sails, tacks, and works as well in contrary winds as ships with a keel, and makes very little leeway. This advantage it derives from another method of steering than by a rudder; namely, by some boards, three or four yards in length, and half a yard in breadth, called guaras, which are placed vertically, both at the head and stern between the main beams, and by thrustingsome of these deep in the water, and raising others, they bear away, luff up, tack, lay to, and perform all the other motions of a regular ship. An invention hitherto unknown to the most intelligent nations of Europe..." (Juan and Ulloa, 1748, vol. 1, p. 264). Humboldt, Stevenson, and Paris continue to praise the amazing sailing abilities and seaworthiness of the balsa rafts surviving in the 19th century. No. 44) HEYERDAHL: GALAPAGOS SYMPOSIUM 47 and in 1832 Morrell reported seeing them fifty miles from land and able to "beat to windward like a pilot boat..." Skogman on his world cruise in 1854 reported that deep sea-going balsa rafts even visited the distant Galapagos group, and he met them at sea navigating with bipod masts and long guaras sunk between the logs fore and aft (Skogman, 1854, vol. 1, p. 164). About the turn of the century the balsa rafts disappeared, and the direct- ly associated and ingenious technique of guara navigation was ignored and forgotten. Archaeological guara were common, but often ignorantly labeled as agricultural tools, while writers who have realized it was a former naviga- tional device have judged the guara to be a kind of rudder or a usual center- board serving merely as a substitute to a keel to reduce the leeway of a raft. At this time the first scholarly discussions of possible pre-Spanish vis- its to the Galapagos began. Historians of Inca history from Markham in 1907 to Means in 1942 have been so impressed by the obviously historic aspect of Inca Tupac's ocean voyage that they believed his raft armada to have visited the Galapagos, since these were the nearest oceanic islands. Hutchinson (1875) had by then termed the balsa raft a "floating bundle of corkwood," and Means, although believing the Inca had reached the Galapagos, underesti- mated the raft which had taken him there, stating it was "obviously a type of boat that would awake nothing but scorn in the breasts of shipbuilders of al- most any other maritime people in the world. Lothrop (1932) made a more comprehensive study of the practical as- pects of such a voyage, but was misled by an erroneous 19th century source to believe that the Galapagos could never have been reached by balsa rafts. He referred to Byam (1850), an English traveller a century ago, who also saw a balsa beating against the wind off northern Peru, and who was told by his captain that these rafts could tack much closer into a contrary wind than a European whale-boat, but that they went slower through the water, and that in a few weeks they lost their buoyancy and had to be taken ashore to dry. From the latter statement Lothrop concluded that a balsa raft was unable to remain afloat at sea long enough to complete a voyage to the Galapagos, and he sug- gested that Tupac may rather have transported an army by sea and plundered the mainland to the north of Guayaquil. Hornell(1946) wrote: "Certainly no ordinary, untreated balsa raft could make a prolonged oversea voyage unless the Inca's seamen knew of an effec- tive method of treating its absorbent logs with some kind of waterproofing composition..." He found it likely, however, that the early Peruvians used some preparation of gum, resin or wax in some solvent to rub over the logs, and that this had helped the Inca rafts remain afloat to the Galapagos. However, the erroneous verdict of the balsa rafts had now spread into the general Pacific literature, and deprived archaeologists of any stimulus to investigate the arid and uninhabited Galapagos. General visitors to the group, rather than being alerted to the possibility of finding pre-Inca vestiges, denied 48 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers it absolutely. For example, vonHagen (1949, p. 178) was lead to assert: "What- ever islands the Inca sailed to, he did not sail to the Galapagos." He backed this assertion by citing presumably authoritative statements to the effect that the Andean seaboard dwellers were "majestically inept" in marine matters, and concluded that Inca landings in the Galapagos "was a manifest impossi- bility." Subsequent events have shown that the modern verdict on balsa rafts has been erroneous and directly misleading. Since 1947, five manned sailing rafts from Peru have passed the Galapagos, one to be picked up later drifting about in the doldrums, and four to end up in East, Central, and West Polynesia. One was actually heading on for Melanesia when picked up off Samoa. Of more importance still, renewed experiments with the guara technique carried out by Estrada, Reed, Skjolsvold, and the writer in a balsa raft off Ecuador in 1953, resulted in the rediscovery of the functional system of this exceedingly in- genous navigational invention, verifying all the discredited early records to the effect that, through a correct interplay between guara fore and aft, the bal- sa will turn around and tack along any chosen course regardless of wind direc- tion. Other experiments have shown that balsa rafts of green timber will re- tain perfect buoyancy for two years and probably more. Accordingly, the Galapagos are located far within the feasible range of aboriginal Peruvian and Ecuadorian navigation. With this knowledge in mind, an expedition organized to search for pos- sible archaeological sites was led to the Galapagos by the writer in 1953, with E. K. Reed and A. Skjolsvold as participating archaeologists. No attempt was made to accomplish an exhaustive survey of the group or any single is- land, and areas for investigation were selected according to apparent geo- graphical possibilities for aboriginal occupation combined with primitive land- ing facilities. Four pre-Spanish occupation areas were located on three different is- lands. The largest site was on the plateau above James Bay on Santiago Is- land, where eight different aboriginal camp sites were located. A mountain ridge separated these from another site at Buccaneer Bay on the same island. The two other sites were encountered respectively at Whale Bay on Santa Cruz and at Black Beach on Floreana. An additional prehistoric site was located at Cabo Colorado on Santa Cruz by Mr. J. C. Couffer and Mr. C. Hall subse- quent to the departure of our expedition. The combined sites yielded in all 1961 aboriginal ceramic sherds, re- presenting at least 131 pols, probably more. Of these, 44 pots were identifi- able with known ceramic wares from the coasts of Ecuador and northern Peru and 13 additional pots are probably identifiable with ware from the same area. The remaining 74 pots represent aboriginal ware of which 67 are unidentified merely because of insufficient characteristics in the limited material preserved, whereas 7 pots are unidentifiable in spite of striking characteristics in the No. 44) HEYERDAHL: GALAPAGOS SYMPOSIUM 49 remaining sherds. Some sites produced only Peruvian sherds, while others yielded both Peruvian and Ecuadorian material. The ceramic types from the North Coast of Peru were studied and identified by C. Evans and B. J. Meggers of the Smithsonian Institution. La Plata Molded ware is represented by 3 pots from two different local- ities in James Bay. San Juan Molded is represented by one pot from another locality in James Bay. Queneto Polished Plain is represented by 2 pots from two different localities in James Bay. Tiahuanacoid ware is represented by 3 pots from two different localities in James Bay. San Nicolas Molded is repre- sented by one pot from James Bay. Tomaval Plain is represented by at least 15 pots from James Bay, Buccaneer Bay, Whale Bay, and Black Beach. Another five pots from three sites were probable Tomaval Plain. Castillo Plain is re- presented by at least 10 pots from James Bay, Whale Bay, and Black Beach. The latter site also produced a Mochica-type clay whistle. Another 6 pots were probable Castillo Plain. The other identifiable pots were characteristic plainware of the Guayas area in Ecuador. The material is reported in detail by Skjolsvold and the writer in Memoir no. 12 of the Society for American Archaeology. With the exception of three pots of hitherto unknown non-European type, represented by 377 rim, handle, and body sherds of a very thin ware with thick, glossy red slip and complex form, no distinctly new types of ceramic were en- countered. In other words, the material collected, as such is in itself of scant scientific value. Its only importance is embodied in the fact that it has been left behind in the Galapagos Islands, from 600 to 1,000 miles from its identi- fiable mainland points of origins. Naturally then, the question arises: to what extent may some of these remains have found their way to these oceanic islands in post-Columbian times? It may be useful therefore to review very briefly the early history of the Ga- lapagos. The group was accidentally discovered by Europeans in 1535, when Bishop Tomas de Berlanga was caught by the strong off-shore set of the com- bined El Nino and Humboldt Current while sailing from Panama bound for Peru. A day was spent on one island and two on another in futile search of water, whereupon the Spaniards barely managed to tack hstck to Ecuador against the strong westbound currents. Coming from Panama, however, the Bishop and his party could hardly have brought aboriginal Peruvian or Ecuadorian ceram- ics to the Galapagos. A second visit to the group occurred in 1546, when Captain Diego de Rivadeneira stole a ship at Arica, present Chile, and set sail for Guatemala. He rediscovered the Galapagos, and a brief and futile search for water was made on one of the smaller islands, whereupon the ship immediately left the group without setting foot ashore on any of the other islands. Under these cir- cumstances this party could not have left the sherds under discussion. 50 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers We know that some extremely few other Spanish caravels sailed into the Galapagos Sea in the latter part of the 16th century, but it is also known that they made no use of the islands, which they found to be desert and without fruit or water. It is possible that some of these caravels carried some Indians on board, and that some of the latter went ashore with ceramic pots, some of which were broken there, but it is hardly possible that they carried a minimum of 131 aboriginal pots ashore, and broke them all. Nor would they even have brought along such variety of ware, representing widely separated geographi- cal regions and cultural epochs in aboriginal Peru and Ecuador. Although first referred to as Galapagos on a map by Ortelius in 1570, this remote group in the treacherous Humboldt Current remained Las Islas En- cantadas to the Spaniards, until the British buccaneers found it a convenient hide-out towards the end of the 17th century. As cited above, the buccaneer Captain Sharp, who described merchant balsa rafts carrying cargo between the valleys of Peru and Panama, attempted to call at the Galapagos in 1680, but the currents prevented him even from landing. Four years later, in 1684, Euro- peans got a brief foothold ashore for the first time, when a group including Cowley, Dampier, Davis, Wafer, Ringrose, and John Cook anchored for twelve days in James Bay on Santiago, while dividing their spoils. A British Museum manuscript by Cowley reflects the isolation of the group until then: "...wee sailed away to the Westwards to see if wee could find those Islands called theGalipoloes, which made the Spaniards Laugh at us telling us that they were inchanted Islands and that there was never any but Captaino Porialto that had ever seen them but could not come nearethem to Anchor at them, and they were but Shadowes and noe reall Islands." A curious incident is that this buccaneer party in 1684 stored a strange booty in James Bay, including eight tons of quince marmalade. The Viceroy of Peru detected their hide-out, all the large jars were destroyed, and count- less sherds of thick, wheel-made "Spanish jars," first noted by Colnett in 1798, are still found all over the local plateau. An interesting point is that some of these sherds were seen by us imbedded in the large black lava flow part of which covers a main portion of the local valley, thus showing that this major volcanic outburst on Santiago Island post-dates A.D. 1684. These pioneering British buccaneers were followed in 1700 by a French expedition under Beauchesne-Gouin which remained a month, whereas the Span- iards arrived to explore and map the group under Torres in 1789, a visit which is recorded to be the first of any consequence by a Spaniard since Berlanga's brief visit of discovery. There is, accordingly, no foundation for a hypothesis of post-European introduction of the aboriginal refuse in various sites in the Galapagos group. The identification by Evans and Meggers of the Peruvian ware shows that the local deposits consist of material dating back through Estero, La Plata, and Tomaval periods on the mainland, which means that refuse from at least two No. 44) HEYERDAHL: GALAPAGOS SYMPOSIUM 51 of the Galapagos sited are dateable to Coastal Tiahuanaco times. The discovery of sherds from a minimum of 131 aboriginal pots broken and left behind in the Galapagos implies a considerable human activity in precolonial times. It is quite obvious that our cursory survey failed to encoun- ter all sites, and only uncovered part of the material still available. Owing to the general scarcity of soil on the coastal cliffs, much of the refuse is also washed into the sea. It is also clear that we are dealing with repeated visits rather than permanent habitation, as the latter would have left thicker depos- its and a more homogenous ware. A local development would scarcely have succeeded in achieving an independent evolution in pottery that closely fol- lowed the mainland pattern from Castillo and Tomaval Plain ware through poly- chrome Tiahuanacoid, San Nicolas Molded, and finally the three characteris- tic types of Chimu blackware as represented by Queneto Polished Plain and San Juan and La Plata Molded. The refuse deposited represents ceramic types from the Guayas area of Ecuador down to the Casma Valley near the transition to the Central Coast of Peru, 1,000 miles away. To summarize: The use of the Galapagos Islands probably as a fishing outpost is not a practice of European origin, but the continuation of an abor- iginal pattern that appears to date at least as far back as the Coastal Tia- huanaco period in the Peruvian archaeological sequence. Literature Cited Byam, G. 1850. Wanderings in some of the western republics of America... London. HAGEN, V.W., VON 1949. Ecuador and the Galapagos Islands. University of Oklahoma Press, IX + 290 pp. HORNELL, J. 1946. How did the sweet potato reach Oceania? Journal of Linnaean Society of London, vol. 53, no. 348, pp. 41-62, figs. 1-2, 1 map. London. Hutchinson, T. J. 1875. Anthropology of prehistoric Peru. Journal of the Royal Anthropological Institute, vol. 4. London. Juan, G., and A. De Ulloa 1748. Relacion historica del viaje a la America Meridional... Vol. 1. Madrid. LOTHROP, S. K. 1932. Aboriginal navigation off the West Coast of South America. Royal An- thropological Institute, vol. 62,. London Skogman, C. 1854. F reg at ten Eugenics Resa Omkring Jorden Aren 1851-53. Vol. 1 . Stockholm. OPPORTUNITIES FOR BOTANICAL STUDY ON THE GALAPAGOS ISLANDS* i Henry K. Svenson U. S. Geological Survey Washington, D. C Establishment of a permanent biological station on the Galapagos Is- lands offers botanists an unusual opportunity for study of the ecology and systematics of flowering plants and lichens. Only a small part of the Galapa- gos Islands has been well explored. The flowering plants are known only in a provisional way. The lichens, worked on byLinder (1934), are present in pro- fusion but are comparatively little known, although as "orchilla moss" they were once used in dyes, and represented the only vegetable product of the Is- lands that was of any commerical value. The general ecology of the Islands has been treated in some detail by Stewart (1915), but the ecology of varia- tion within individual species is almost unknown. Much of the southern part of Indefatigable Island is accessible by foot from the Darwin Research Station. Even in the neighborhood of the Station the complex variability within individual species would be sufficient for the solu- tion of many fundamental problems that have been known since the time of Darwin. For example, careful field examination with a hand lens would allow differentiation of the various species of Cordia (Boraginaceae), the identifi- cation of which was based by Johnston (1935) primarily on the presence of simple, branched, or stellate hairs. Likewise, field examination of mature seeds or achenes of such genera as Altemanthera and Scalesia would further the understanding of fruiting parts in these difficult genera. It is to be hoped that plants can be experimentally grown from seeds, in order to solve many of the tangles associated with variable species in the Galapagos Islands. Such experiments could be carried on in Hawaii, California, or even in a little plot adjacent to the laboratory. Variability within individual species on the Galapagos Islands is more marked than in the same species on the mainland of Ecuador (Svenson, 1946, • Presented at the TENTH PACIFIC SCIENCE CONGRESS of the Pacific Science Association, held at the University of Hawaii, Honolulu, Hawaii, U. S. A., 21 August to 6 September 1961, and sponsored by the NATIONAL ACADEMY OF SCIENCES, BERNICE Pauahi bishop Museum, and the University of Hawaii. 1 Publication authorized by the Director of the U. S. Geological Survey. -53- 54 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers p. 142). The variability commonly takes the form of reduction in leaf surface, as may be seen in Croton (Svenson, 1935, plate 4), or the lack of spines in Acanthospermum microcarpum (Robinson, 1902, plate 1), or of leaves and car- pels as in Sidaspinosa. This variability within species is fundamental to the study of evolution of species in the Galapagos Islands, and probably uncon- sciously impressed itself on Darwin. Although Darwin had no extensive knowl- edge of South American plants and depended upon his friend. Hooker, for ident- ification of specimens, he was an admirable collector. He saw the great dif- ference in the vegetation of the Islands compared with what he had seen on the Peruvian coast. Excellent accounts are given in the Voyage of the Beagle, and references in the Origin of Species. 2 The letter which he wrote from the Galapagos Islands to his sister is unfortunately missing (Barlow, 1946). As an incentive to exploration it may be mentioned that Darwin collected a num- ber of species in the Islands, including small Compositae such as Elvira in- elegans, which have never been found again. Lack (1947) published a memorable account of Darwin's finches, the only group of birds outstanding as an example of adaptive radiation. As the diet of many of these birds is vegetative, the identification of seeds in bird crops should be one of the goals of field study in the Islands [See Bowman, 1961]. Lack states (page 17) that "Most species of Darwin's finches occur on a number of islands. In some cases the island populations differ sufficiently to justify division into subspecies, in other cases the differences are less marked, and yet in others, they are barely perceptible. They are not in general confined to individual islands." He notes that populations of finches on the smallest islands, such as Wenman and Tower, are the least variable; that those on the moderately small islands of Abingdon and Bindloe are somewhat more variable; and that the population on the larger island of James is more vari- able still. Whether such generalizations are true of plants is not known. The relatively small number of species of plants on the Galapagos Islands should make studies much easier than on the mainland, with its more complicated flora. Lack notes (page 115) that Darwin's realization that a species may be represented by different forms in different regions was one of the most impor- tant results of the voyage of the Beagle, since it led directly to the question- ing of the immutability of species. He states (page 125), "The apparent fixity of species is most striking and provides the basis for systematic zoology... Charles Darwin and many after him are wrong when they assert that the deter- mination of species is purely arbitrary." Dobzhansky (1941, page 365), simi- larly finds that, "The notion, entertained by some biologists unfamiliar with 2 In this footnote I extend my appreciation to the late Professor L. J. Henderson of Harvard University. His course on the history of science introduced me to the Origin of Species, Merz' History of European Thought in the Nineteenth Century, and his own book, The Fitness of the Environment, a background for Darwinian evolution. No. 44) SVENSON: GALAPAGOS SYMPOSIUM 55 the subject, that species are arbitrary units like all other systematic units, is unfounded." Zimmermann (1954, page 195) observes that Linnaeus passes in general as the representative of constancy in species from the beginning, and of an artificial system, but that is correct only of the younger Linnaeus. We may now turn our attention to a correlation between the vegetation of the Galapagos Islands, the South American mainland, and the Caribbean re- gion, and the means by which plants may have come to the Islands. Many Com- positae are tree-like on islands, but such a point of view can easily be over- emphasized. Compositae (Baccharis) and also cacti (Cereus) are just as large, if not larger, on the adjacent mainland of Ecuador. The Galapagos Islands, it is true, are a focal point of dry zones from the Caribbean to Peru. But between these dry zones intrudes the great tropical rain forest, which extends from Darien to Ecuador, with species that have little or no relationship to the Ga- lapagos flora. Overlooking of this simple fact is, I believe, the basis of much of Croizat's troubles (1958) in respect to the Galapagos vegetation. Species on the Galapagos Islands are more variable than those on the adjacent conti- nent, and Howell (1934, page 515) has fittingly called the Islands "Evolution's workshop and showcase." The number of species in the Islands is relatively small, which is all to the good, for more attention can then be paid to varia- tion in species, without being overwhelmed by their number. It does not mean that new species will not be discovered in the Galapagos Islands, for every expedition finds them. The studies should try to show how continental spe- cies have reached the Islands, in the manner of the excellent contribution made by Howard (1950) for Bimini Island in the Bahamas. As adequate meteorological observations are made on the Islands, there is the opportunity of correlating the climatic area occupied by individual spe- cies with the distribution of the same species on the mainland. For example, the yellow-flowered Piscidia carthagenensis Jacquin is one of the largest trees on the Islands. It was long considered to be the same as the widespread P. erythrina. Stewart knew it from Chatham and Santa Cruz Islands. Altitudi- nal distribution of this tree (Svenson, 1935, page 210) exceeds that of any other tree in the Islands, extending from near the coast at Academy Bay to over 2,000 feet elevation. Its greatest size is in the most humid region; above 1,000 feet elevation it is dwarfed. On the mainland, Piscidia carthagenensis is known only from western Ecuador, the Cartagena-Barranquilla region of Colombia, and the north coast of Venezuela. All of these areas have a dry climate. Thus the continental distribution of this tree demonstrates that even the moistest parts of the Galapagos Islands correspond in their vegetation to areas of dry climate on the mainland. Such continental areas in northern South America are shown on a map by Sorge (1930). From a general point of view, the climatic references in respect to vegetation given by Koppen (1931), Lauer (1951 and 1952), Tregurtha (1961), and especially Papadakis (1961), should also be consulted. 56 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers The best knov/n example of radiative adaptation of plants in the Gala- pagos Islands is Scalesia. Sixteen species are recognized by Howell (1941). The genus is related to Helianthus of western South America, but the exact relationship is not known. Also of extraordinary interest from the point of adaptive radiation is the genus Euphorbia, which was elaborated by A. Hass- ler (1929). All Galapagos species of Euphorbia are related to one another ex- cept E. viminea, which is known also from the Bahama Islands, and chiefly from Turks Island. Euphorbia viminea may thus be presumed to have come from the Bahamas. Nevertheless, the shrub is now common throughout the Ga- lapagos Islands, and it was treated by Robinson and Greenman (1895, page 136) as an example of unusual variation, with the notation, "Perhaps no spe- cies to be found on the different islands better illustrates the noteworthy ra- cial divergence in related forms than Euphorbia viminea." It was first collec- ted by Macrae in 1825 on Albemarle Island. The variation is, I think, mainly the difference between adult and juvenile foliage (Svenson, 1935, plate 2). This species is probably easily spread by proliferations developed in the leaf axils, and presumably has come from the Bahamas in connection with the salt and whaling industry. Turks Island was an important source of high-grade salt for eastern United States during the first half of the nineteenth century, especially for preservation of meat; and in early days, the salt was loaded directly into ships by means of wheelbarrows. My friend, Robert Cushman Murphy has informed me that he does not know of any definite records of whalers stopping at Turks Island for salt, but there is no reason to suppose that they did not do so. To show how easily plant fragments can spread, we may note the peculiar distribution of Eleocharis pachycarpa, a Chilean spe- cies which appeared at Port Jackson in Australia at a very early date. The well known anthropologist Herbert Spinden, then of the Brooklyn Museum, sug- gested a possible explanation to me. In order to avoid the hot journey around the Cape of Good Hope, ships transported sheep and cattle from England to Australia by way of Cape Horn, with stops at Buenos Aires and Santiago for pasture. Under such conditions, fragments of the Eleocharis could easily be picked up. In the Galapagos Islands, modern scientific investigation begins with the visit of David Douglas and John Scouler in January, 1825, but whaling op- erations were carried on at an earlier date. Woodes Rogers and other pirates were there as early as 1708. What plant introductions, if any, these early vis- itors made upon the arid shores of the Galapagos Islands is unknown. Dry areas of the Islands seem to have the more pronounced endemism. Darwin found the Galapagos finches to be characteristic of arid regions, and Howell mentions pockets in arid regions where endemism. is exceptionally well developed. Thus Howell (1941, page 237) states, "Not uncommonly the same ecologic factors affecting segregation are locally active in several, unrelated groups of plants, and because of the peculiar geologic history or critical geo- No. 44) SVENSON: GALAPAGOS SYMPOSIUM 57 graphic position of the particular district where they grow, an endemic area may emerge in which unrelated entities exhibit parallel responses to some specialized ecologic condition. Such an area seems to occur in the vicinity of Sullivan Bay on James Island where several remarkable and distinct species are found... another is that region including the Seymour Islands and adjacent Indefatigable." Stebbins (1952, page 34), in a general discussion of aridity as a stimu- lus to evolution, not specifically on the Galapagos Islands, mentions that, "In the dryer areas it is possible that more species originate, reach their cli- max, decline, and become extinct, than in more favorable regions. Reduction of leaf surface, development oftrichomes, scales and other coverings, ofsunk- en stomata, of deciduous leaves, of extensive root systems, of bulbs, storage roots, and other structures... all of these appear as frequent modifications of xerophytes." Many references have been made to the lack of certain families of plants in the Galapagos Islands, but to me the absence of the Capparida- ceae, so prominently represented as bushes and trees on the xerophytic coasts of Ecuador and Peru, is the most striking. From the foregoing review, opportunities for botanical research in the Galapagos Islands would appear to be outstanding. For much of the background, and for help in what I have said, I am especially indebted to John Thomas Howell of the California Academy of Sciences. Literature Cited BARLOW, Nora, ed. 1946. Charles Darwin and the Voyage of the Beagle, unpublished letters and notebooks. 279 pp. The Philosophical Library. New York. Croizat, L. 1958. Panbiogeographie. Vol. 1, 961 pp. The New World. Weldon & Wesley, Herts, England. DOBZHANSKY, T. 1941. Genetics and the origin of species. Vol. XVL 364 pp. Columbia Univer- sity Press, ed. 2. Hassler, a. 1939. On the limitation of the species within the Euphorbia-group Cheloneae Doiss. Botaniska Notiser, For., Ar. 1939, pp. 745-748. Henderson, L. J. 1913. The fitness of the environment. The MacMillan Company. New York. Howard, R. A. 1950. Vegetation on the Bimini Island group, Bahamas, B.W.I. Ecological Monographs, vol. 20, pp. 317-349. Howell, J.T. 1934. Cacti in the Galapagos Islands. Cactus and Succulent Journal, vol. 5, pp. 515-518, 531-532, figs. 1-9. 1941. The Genus Scalesia. Proceedings of the California Ac ademy of Sciences, ser. 4, vol. 22, pp. 221-271. 58 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Johnston, I. M. 1935. The Genus Cordia. Journal of the Arnold Arboretum, vol. 16, pp. 174-176. KOPPEN, W. 1931. Grundriss der Klimakunde, ed. 2. Walter de Gruyter & Co. Berlin. Lack, D. 1947. Darwin's finches. 208 pp. Cambridge Unitersity Press. Cambridge Eng- land. Lauer, W. 1951. Die Zahl der humiden bzw. ariden Monate (Isohygromenen) in Siidamer- ika u. Afrika. Erdkunde, Band V, Heft 4. 1952. Studien zur Klima und Vegetationskunde der Tropen. Bonner Geograph- ische Abhandlungen, Heft 9. LiNDER, D. 1934. Lichens of the Galapagos Islands. The Templeton Crocker Expedition of the California Academy of Sciences. Proceedings of the Cali- fornia Academy of Sciences, ser. 4, vol. 21, pp. 211-221. Merz, J. T. 1903. History of European thought in the Nineteenth Century. London. PAPADAKIS, J. 1961. Climatic tables for the world. Av. Cordoba 4564, Buenos Aires. Robinson, B. L. 1902. Flora of the Galapagos Islands. Proceedings of the American Academy of Arts and Sciences, vol. 38, no. 4, pp. 77-269. ROBINSON, B. L.,and J. M. Greenman. 1895. On the flora of the Galapagos Islands, as shown by the collection of Dr. G.Baur. American Journal of Science, ser. 3, vol. 50, pp. 135-149. SORGE, E. 1930. Die Trockengrenze Slidamerikas. Zeitschrift der Gesellschaft fiir Erd- kunde zu Berlin, pp. 277-287. Stebbins, L. 1952. Aridity as a stimulant to plant evolution. American Naturalist, vol. 86, pp. 33-44. Stewart, A. 1911. A botanical survey of the Galapagos Islands. Proceedings of the Cali- fornia Academy of Sciences, ser. 4, vol. 1, pp. 7-288. 1915. Some observations concerning the botanical conditions on the Galapagos Islands. Proceedings of the Wisconsin Academy of Sciences, vol. 18, pp. 272-340. SVENSON, H. K. 1935. Plants of the Astor Expedition (Galapagos and Cocos Islands). Ameri- can Journal of Botany, vol. 22, pp. 208-277. 1946. Vegetation of the coast of Ecuador and Peru and its relation to the Ga- lapagos Islands. American Journal of Botany, vol.33, pp. 394-498. TREGURTHA, G. T. 1961. The earth's problem climates. University of Wisconsin Press. Zimmermann, a. 1954. Evolution, die Geschichte ihrer Probleme und Erkentnisse. Freiberg. BIOSYSTEMATIC STUDIES ON GALAPAGOS TOMATOES* i Charles M. Rick University of California Davis, California Introduction The present investigations were undertaken in order to ascertain the natural relationships of the Galapagos tomatoes. These anomalous members of the genus Lycopersicon deserve attention for several reasons: (1) Systema- tic problems. Although the Galapagos tomatoes have been collected frequently and are therefore well documented in various herbaria of the world, numerous problems have persisted in their biosystematics. The first specimens were collected by Darwin in 1835 and many additional collections have been made by numerous expeditions since that time. Various forms of the Galapagos to- matoes have been classified into as many as four species and subspecies. Al- though they have attracted the attention of relatively few systematists, much disagreement can be found among these few treatments. The taxonomic situa- tion was reviewed and a regrouping of the material proposed by Rick (1956). (2) Potential contribution to the genetics and breeding of tomatoes. As a new source of germ plasm for such studies, the Galapagos tomatoes have already exceeded expectations. For reasons that will be briefly outlined below, these accessions prove to be unusually useful in enriching the genetic variation of cultivated tomatoes. (3) Adaptability to experimental approaches. All acces- sions so far obtained have proved amenable to culture. Although they do not thrive well under field conditions in the Central Valley of California, they survive there and grow luxuriantly in the greenhouse and in the field in other regions (for example, coastal California). The plants flower and fruit when grown in small containers, and controlled matings can easily be made. Seed dormancy posed an obstacle to our first investigations, but this problem was * Presented at the TENTH PACIFIC SCIENCE CONGRESS of the Pacific Science Association, held at the University of Hawaii, Honolulu, Hawaii, U. S. A., 21 August to 6 September 1961, and sponsored by the NATIONAL ACADEMY OF SCIENCES, BERNICE Pauahi Bishop Museum, and the University of Hawaii. ^ Partial support of this work by grant no. G-10704 of the National Science Foundation and GA AGR 5547 of the Rockerfeller Foiindation is gratefully acknowledged. -59- 60 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers eventually solved by the application of special seed treatments (Rick and Bow- man, 1961). These tomatoes therefore constitute one of the few genera in the Galapagos biota that can be subjected to a wide range of biosystematic ap- proaches: the nature of breeding systems, compatibility with other species, genetic differentiation between races, etc. (4) Relationships to the Galapagos flora in general. In any such investigations it is naturally hoped that the find- ings on a specific group might shed light on the larger problems— in this case, the evolution of the Galapagos flora. The present study is based on herbarium specimens and on living cul- tures established from 34 accessions collected by Alf Kastdalen, Zouzou Coray de Castro, Miguel Castro, Otis Barton, and Robert I. Bowman, and by the writer during a visit in 1956. The writer is greatly indebted to the afore- mentioned people for tomato accessions and for assistance in many other re- spects. Description of Races The Galapagos tomatoes are all low, spreading herbs with relatively diminutive plant parts. Although they can be treated in culture as shortcycle annuals, most plants encountered in the wild are perennials, continually pro- ducing new branches from the crown of the plant in the same fashion as most other wild species of Lycopersicon do in their native habitat. They abound in the arid, lowest zone of the islands, but have also been encountered occa- sionally in the middle, forested belt of the larger islands. Phenomenally drought-resistant, they continue to grow throughout the year, and are often the only mesophytic plants in leaf during the long dry period of the lowest zone (figures 1 and 2). All examined collections have twelve pairs of chromosomes, in common with other species of Lycopersicon. Although differentiated into many different biotypes, all truly native forms of the Galapagos tomatoes possess in common at least five morphologi- cal and physiological traits by which they can be distinguished from all other known species: (1) Yellow or orange fruit color; (2) Yellow-green foliage co- lor; (3) Minute seed size; (4) Consistent and severe seed dormancy; (5) Char- acteristic physiological growth requirements. Although poorly understood, the latter differ from those of any other tomato species we have cultured. The same traits were recognized in an earlier survey (Rick, 1956), which was based on herbarium specimens and living material from only three acces- sions. One of those collections, LA292, then identified as L. esculentum var. cerasiforme, has since been demonstrated to have been a cultigen, probably escaped from a nearby garden. No such form with large (>2cm.) red fruits has been demonstrated unequivocally to be native to the Galapagos Islands. For more details concerning the taxonomy and description of races the reader is referred to Rick (1956). No. 44) RICK: GALAPAGOS SYMPOSIUM 61 \ ..W *' Figure 1. Single plant of L. esculentum var. minor growing among lava rocks near the north shore of Jervis Island. Elevation 25 feet. Opuntia myriacantha in background. All other plants are dead or dormant, characteristic of vegetation in the lowest zone during the dry season. The commonest biotype is L. esculentum var. minor (L. cheesmanii vat. minor) hereafter designated var. minor. This form is known from at least six of the main islands and from many different sites on Albemarle and James islands. In addition to those characteristics that are constant in all Galapagos toma- toes, as noted above, this biotype exhibits short internodes, large accrescent calyx, dense hairiness, and highly dissected leaves. The consistent appear- ance of these four unique traits and other less tangible ones leaves little room to doubt the close genetic relationship of all accessions of this type. These traits are shown in figures 1 to 5 of Rick (1956), and the general as- pects in the wild are shown in figure 1. Var. minor has always been found at low elevations, sometimes within 100 feet of, and only several feet in eleva- tion above, the surf. The widespread distribution and proximity to the strand of this form hint that marine transport of some kind might have been respon- sible in part for its present distribution. The next most common entity is one that has most generally been class- ified as .a form of L. pimpinellifolium, hereafter designated as Gal. ppn. In contrast to the two preceding biotypes, it has longer internodes, less divided 62 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers leaves, more elongate corolla segments, fewer hairs, and smaller, spreading calyx. It corresponds to Rick's drawings (1956), figures 7 to 10, in all details except for the absence of pedicel articulation illustrated in the latter. Recent- ly, field studies have revealed that the monogenically-determined absence of joint is limited to the Academy Bay region of Indefatigable Island and that normal articulation prevails throughout most of the range of this type. It has been found on four of the islands, generally at higher points in the lower zone than var. minor and occasionally extending to the interior of the islands, as on Indefatigable. As a race this one is less clearly defined, various constant- ly breeding deviations have been seen in some of the collections as well as inter gradations with the following form in a few populations. Its typical habi- tat is shown in figure 2. The third main type is the typical form of Lycopersicon cheesmanii, the type locality for which is the north side of Indefatigable. Collections of liv- ing material have been obtained from this general area and also from nearby Seymour Island. In some respects typical L. cheesmanii is intermediate be- tween the two preceding forms. Its morphological affinities are closer to those Figure 2. Colony of Galapagos form of L. pimpinellifolium (LA430) growing among the lava boulders along the volcanic escarpment 1 km. NE of Academy Bay, Indefatigable Island. Note dormant condition of the trees (Piscidia erythrina) and absence of any other vegetation in leaf, characteristic of the lowest zone during the dry season. No. 44) RICK: GALAPAGOS SYMPOSIUM 63 of Gal. ppn., but it resembles var. minor in respect to its shorter internodes. It is intermediate in respect to density of epidermal hairs. Its foliage is some- what different than that of the two other forms, being less divided than either and having lateral segments orbicular in outline. A collection showing some resemblances to typical L. cheesmani has been collected from Essex Point on the southwestern extremity of Albemarle. The key morphological features of this biotype have been illustrated by Luckwill (1943). Various other biotypes of more limited distribution have also been dis- covered. In most cases these have been encountered in only one population or in an otherwise very narrowly restricted region. For the most part they can be described satisfactorily as combinations of characteristics of the preced- ing three main biotypes. Population Structure The flower parts of Galapagos tomatoes, like those of the cultivated L. esculentum, are disposed to promote self-pollination. After being shed into the anther tube, the pollen drifts downward to the stigma, which is situated at the mouth of the tube or is exserted slightly beyond it. The flowers are small and arrayed much less conspicuously than those of the highly cross-pollinated species. No inherent barriers to self-pollination exist; a very high proportion of the artificially selfed flowers set fruit; and under the insect-free condi- tions of a screened and fumigated greenhouse fruits are set freely, even with- out agitation of the flowers. Advantage was taken of every opportunity in the wild to observe the vi- sits of insects to flowers of Galapagos tomatoes. Such observations were made in some twelve native populations, the period of observation varying be- tween two hours to several days per population. The only insect visits observed were those to a large colony of Gal. ppn. (LA430) in the vicinity of Academy Bay in the talus of a volcanic escarpment along the trail to the highlands (fig- ure 2). Here the only known native Gala'pagos bee, Xylocopa darwinii Cocker- ell (Hurd, 1958), was observed to make sporadic visits to the tomato flowers. In contrast, the amount of bee activity in nearby flowering trees of Piscidia was intense. The bees were unquestionably visiting the tomato flowers for the purpose of collecting pollen because they displayed the usual vector ac- tivity in such acts as grasping the anther tube with their legs and vibrating it rapidly with a high-pitched buzz. In a period of 2 1/2 hours spent in this pop- ulation, only four bees were seen visiting tomato flowers— an activity of re- markably lower level than that observed in any tomato species on the adjoin- ing mainland (Rick, 1950) and far less than that seen in neighboring species of Piscidia, Momordica, and Cryptocarpus. It was not surprising therefore to find the genetic structure of most pop- ulations to be exceedingly uniform. With one exception the plants generally showed remarkable agreement for such genetically stable characters as flower 64 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers shape, inflorescence structure, fruit color, and anthocyanin pigmentation. In the population LA430, 75 plants were systematically scored for ten morpho- logical characters and only two were found to deviate in gross morphology from a single prevailing phenotype. One plant had a normally developed pedi- cel articulation, in contrast to the aforementioned jointless (/2) condition, and another had small seedless fruits, presumably resulting from some type of gen- etic sterility. Opportunity for similar studies was found in two other popula- tions of the same biotype within the same general area. Extreme uniformity was again observed in 19 plants constituting the total of one of these popula- tions (LA439) found on the cliffs bounding the west side of Academy Bay. In the third population (LA432), encountered 1 km. east of LA430, 163 plants were examined and 16 were found to deviate in one of several well defined traits chiefly of the fruits (figure 3). With the following exception, all the other wild populations studied showed the same degree of uniformity. The exceptional population (LA438) was one encountered along the coast of Albemarle approximately 5 km. southwest of Villamil. This area was also exceptional for the sympatric existence of both var. minor and Gal. ppn. Here, one isolated colony of about 20 plants was found to show great varia- tion between plants in what seemed to be various combinations of the traits of these two biotypes. Of 13 plants that were examined in detail, five were typical var. minor, the rest being "hybrid" in respect to the presence of some Gal. ppn. traits. Among the latter, five different types could be distinguished, several plants each being found for some of the types. Plants of Gal. ppn. were found several hundred feet away in lower places. The distributional map of races (figure 5) suggests that a situation sim- ilar to that for LA438 might exist on the slopes of the crater of Narborough. Max. minor and Gal. ppn. coexist there and an intermediate phenotype, closely approximating typical L. cheesmanii has also been found. Since population studies have not been made there, the existence of variable populations is un- certain. The same remarkable degree of uniformity is characteristic of progenies grown in culture from single wild plants. In no case, even among offspring from the variable population on Albemarle, was any genetic variation detect- able. These populations, mostly of no'more than 20 plants apiece, have been grown from over 50 parent plants from many populations. The degree of uni- formity typical of this material is illustrated by seedling cultures in figure 4. In contrast, the level of variability encountered in single-plant progenies of other wild tomato species is very much greater. From 20 years' experience in progeny testing tomato lines, I can state with confidence that the uniformity of Galapagos tomatoes is matched only by highly inbred lines of the cultivated L. esculentum. Another interesting aspect of population structure of the Galapagos to- matoes is the fixation of various recessive genes. One of the first examples No. 44) RICK: GALAPAGOS SYMPOSIUM 65 Figure 3. Fruit samples from population LA432 of Galapagos form of L. pimpinellifolium east of Academy Bay, Indefatigable Island. One cluster was taken from each of six different plants, some showing marked morphological deviations from the normal type. Typical form with jointless f/^j pedicels in upper right. Deviant with normal pedicel joints in middle left. Deviant with small fruits at lower left. Deviant with strong anthocyanin pigmentation (atv) in lower right. (1/2 Reduction) 66 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers is the aforementioned jointless trait, governed by a single recessive gene, /2. Since the time of the earlier work (Rick, 1956), field studies have shown that this gene is completely fixed in several populations in the vicinity of Aca- demy Bay. The yellowish foliage color of many populations of L. pimpinelli- folium is at least partly due to the presence of the lutescent gene /I. A com- plete lack of anthocyanin, characteristic of an entire population of Gal. ppn. near Wreck Bay, Chatham, was ascertained to be determined by ag. Although opportunity has not been found to investigate all of the encountered variations, at least one mutant trait encountered segregating within a population has been found to be determined by a previously unknown gene: high anthocyanin con- tent of fruits and foliage (figure 3) from the LA434 population segregates in this fashion. In backcrosses of L. esculentum, the purple flush of the fruit has not been recovered, but the recessive homozygotes can be clearly identified by the intense anthocyanin coloring of leaves under cool growing conditions. This new gene is provisionally named atroviolaceum (atv). The very high level of inbreeding that must exist in these populations provides ideal conditions for the fixation of such genes and the rapid differ- entiation of races. It is very likely that the observed monogenic variations arose from mutations in the Galapagos material and it is tempting to suppose that at least some of them have been fixed purely at random without reflecting selective advantage. Distribution All of the foregoing evidence reveals an extremely high level of uni- formity within populations of the Galapagos tomatoes. With minimal excep- tions among the whole group, it is therefore valid to equate a single popula- tion with a single genotype, thereby simplifying the visualization of geographic distribution. Figure 5 illustrates graphically the distribution for variation in four key morphological characters. The presence and intensity of expression of these characters is symbolized by Anderson's (1949) familiar ideograms. According to the scheme adopted. Gal. ppn. and var. minor are the contrasted extremes, the latter being represented in figure 5 by a solid circle with three long appendages and the former by an empty circle without appendages. Only those populations are included that have been grown in culture or for which adequate herbarium specimens exist. It is evident from this summary of geographic distribution that var. minor is widespread throughout the archipelago, lacking in only four of the larger islands, possibly for reasons of inadequate collections. Gal. ppn. appears on at least four of the main islands and is as widely distributed as the preceding Figure 4. Typical pattern of variation in seedling progenies. Each family, appearing to the right and above its respective number, is the progeny of a single plant collected in the wild. Family 471 and 474 are typical L. cheesmanii. Family 472 and 473 are L. esculentum var. minor. (1/4 Reduction) No. 44) RICK: GALAPAGOS SYMPOSIUM 67 [470 [4.0 [4743 SPa^w C473: ;;; 68 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers form in respect to longitude. Typical L. cheesmanii, represented in figure 5 by a half-solid circle and single vertical appendage, is distributed to three and j>erhaps four larger islands. Such widespread distribution of single biotypes contrasts strikingly with the narrow endemism of many other components of the Galapagos biota. To il- lustrate this contrast, figure 6 has been prepared to show the distribution of Scalesia species according to Howell's (1941) monograph. In this familiar ex- ample of Galapagos endemism, all but five of the nineteen well distinguished species are restricted to single islands. It is safe to speculate that at least part of this difference in distribution reflects differences in dispersal ability of the two genera. As already stated, the very close approach of var. minor to the strand of several islands suggests that this form might occasionally be distrib- uted by the sea. Experiments with seed germination, presented in the next sec- tion, offer additional suggestions as to mechanisms of interisland dispersal. Another hint of successful dispersive ability of the Galapagos tomatoes is given by evidence of their recent invasion of new habitats. Some popula- Hoirs 3rd order leof divisions Intemode length Colyx lorge ft Decrescent L pimpinel- li folium noneO noneo long no L esculentum vor minor mony^ much n short o yes '~^ ^7 m^ .^ Figure 5. Geographic distribution of races of Galapagos tomatoes. All islands are shown except the northernmost two, Culpepper and Wenman, from which tomatoes have not been re- ported. The population phenotypes are indicated by ideograms according to Anderson (1949). Galapagos L. pimpinellifolium is represented by an open circle; L. esculentum var. minor, by a solid circle with three long appendages; and typical L. cheesmanii, by a half-solid circle with a single vertical appendage. No. 44) RICK: GALAPAGOS SYMPOSIUM 69 Figure 6. Geographic distribution of species of Scalesia (after Howell, 1941). All islands are shown except the northernmost two, Culpepper and Wenman. Species are designated by the following symbols: af - S. affinis, as - aspera, at - atractyloides. B - Baurii, co - cordata, Cr - Crockeri, Da - Darwinii, di - divisa, g - gummifera. He - Helleri, Ho - Hopkinsii, in - incisa, m - micTocephala, p - pedunculata, V\-- var. typica, P2 - var. Svensoni, P3 - var. parviflora, P4 - vaT.indurata, P5 - var. ^/7osa, t- retroflexa, Sn- Snodgrassii, St. - Stewartii, v-villosa, S. Snodgras- sii has been collected from Wenman Island and an unidentified species from Culpepper Island. tions— for example var. minor on Bartholomew Island— have been found growing in crevices of lava flows in which the pristine condition of ripple marks at- tests very recent volcanic activity. Also, native tomatoes have been collected from areas where volcanic activity has been observed in historic times. Such populations have been collected from at least six different places, including the interior of the large central crater of Narborough Island, which suffered a violent eruption in 1825. The Galapagos tomatoes are clearly well adapted to migrate and colonize new areas. They are nevertheless endemic in the sense that the whole group is restricted to the Galapagos Islands. Seed Dormancy The problem of seed dormancy will not be presented in detail here, since this subject and the development of effective means of overcoming thedorman- 70 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers cy are treated by Rick and Bowman (1961). In summary it was found that all seeds of any races of Galapagos tomatoes thus far acquired show a severe dormancy, which is not corrected by prolonged storage. Of the many different treatments applied, the only ones effective in improving germination were meth- ods that partially or wholly removed the seed coats. The behavior of the seeds in those tests suggests that the inhibition is simply mechanical: once the seed coat is partly or entirely removed, the radical emerges rapidly. The simp- lest and most effective treatment was found to be the exposure of seeds to strong solutions of sodium hypochlorite. In a search for mechanisms that might permit natural establishment of the Galapagos tomatoes, it was found that passage of the seeds through the digestive tract of the giant Galapagos tor- toise, Testudo elephantopus porteri resulted in improvement of germination of war. minor from 1 to 85 per cent and of Gal. ppn. from 1 to 11 percent. Two to four weeks were required for passage through the tortoise gut. If this is a na- tural mechanism for establishment, it would be effective not only in permitting germination but also in expediting dispersal by virtue of the long digestive period. Dependence upon animal digestion poses some interesting evolutionary problems. The selection and establishment of such a dependency would seem- ingly call for much trial and error and would likely require a long period for establishment. If animals play a major role in breaking the tomato seed dormancy, the question might be asked: to what extent are they responsible for the present distribution of Galapagos tomatoes? The role of the tortoises is problematic: if they were responsible to any large degree for dispersal of the native toma- toes, the endemic distribution of the tortoises themselves would be contradic- tory. Beebe (1922) has ascertained that the tortoises can survive and swim in sea water. On the other hand, safe landings on the hazardous Galapagos shores are another matter. Yet, death or severe injury to a tortoise at landing would not necessarily preclude establishment of the tomatoes. This problem is ob- viously complex and has many facets; nevertheless, the possibility of inter- island dispersal by tortoises as a rare event cannot be precluded at the pres- ent state of knowledge. Other possible agents are marine or such terrestrial bird species that can migrate from one island to another. The only likely avian agents, however, would be those with a mild digestive action— for example, the Galapagos mocker— for tomato seeds do not tolerate much grinding by gizzards. The two native species of iguana might also be implicated. A treatment of the tomato distribution would not be complete without consideration of the effects of possible early changes in the configuration of the archipelago. The ocean floor is so shallow in the Galapagos area that only slight lowerings of the water level or elevations of the land masses would re- sult in land connections between various islands. Under such conditions wide- spread dispersal of the tomato races would not have presented serious diffi- culties. No. 44) RICK: GALAPAGOS SYMPOSIUM 71 Compatibility Tests Various accessions of Galapagos tomatoes have been subjected to rou- tine tests of cross-compatibility among themselves and between them and other species of tomatoes. These tests have been conducted in the spring and fall seasons under standard greenhouse conditions at Davis, California. Five to ten plants of each accession were used for the matings and six or more flowers were pollinated in each combination tested. The results of these hybridiza- tion experiments are presented in graphic form in figure 7. The tests consis- tently revealed absence of severe barriers to crossing between the two most widespread races from the Galapagos, var. minor and Gal. ppn. and L. esculen- tum, L. pimpinelli folium (typical form from the mainland), L. hirsutum, L. min- ^ 0 S t^^CEss '0^ L. esculentum L. pimpine lifolium L minutum L. chilense Figure 7. Compatibility polygon indicating the degree of compatibility between Galapagos forms and other tomatoes. The width of connecting lines indicates the relative amount of hy- brid seed produced by the hybridizations, the widest lines being equivalent to complete com- patibility as measured by the yield of selfs or sibs within the race. A dotted line indicates failure of hybridization. Circled numbers refer to specific accessions. All accessions that are intercompatible and behave similarly in all test crosses to Galapagos races are enclosed in larger circles. 72 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers utum, and Solanumpennellii [which in these and other tests (Rick, 1960) shows a much closer genetic relationship with Lycopersicon than Solanurn\. Seeds produced by these crosses yielded F^ hybrids that were vigorous in all com- binations and fertile in all save combinations with S. pennellii. On the other hand, very severe barriers prevent crossing with L. chilense and L. peruvian- um, none of the hundreds of attempted crosses yielding hybrid progeny. In addition to the aforementioned tests, thirteen other Galapagos acces- sions (seven of Gb\. ppn., three oivar. minor, two ofL. cheesmanii, and one of the intermediate type from Albemarle) were hybridized with L. esculentum. These crosses likewise produced normal yields of seeds, and the hybrids and backcross generations to L. esculentum grown therefrom displayed normal vig- or and complete fertility. Although these additional collections were not sys- tematically tested against each other and the other tomato species, it is a foregone conclusion that they would manifest the same compatibility relations as the two accessions (LA166 and 317) that were tested more extensively. The Galapagos tomatoes therefore comprise an interfertile group of races that hy- bridize freely with L. esculentum and other species of the L. esculentum com- plex. The complete compatibility between Galapagos tomatoes and members of the L. esculentum complex was entirely unexpected because the latest tax- onomic treatments (MuUer, 1949; Luckwill, 1943) placed the former in the same subsection as L. peruvianum and L. chilense. From the standpoint of tomato genetics and breeding this finding was a pleasant surprise, for it meant that the whole gamut of variation in Galapagos tomatoes is available for trans- fer to L. esculentum by applying the appropriate breeding techniques. Although it is not the main objective of this paper to consider applications in the area of genetics and breeding, some of the findings are illuminating from the stand- point of systematics and phylogeny. To date a number of interesting charac- ters have been transferred by backcrossing from Galapagos races to the gar- den tomato. Backcrosses from all of 15 different accessions, including the main races, have recovered the gene B, which diverts synthesis of the fruit carotenoids entirely to beta-carotene. The presence of this gene accounts for the yellow and orange fruit color in all accessions of Galapagos tomatoes. Other monogenic traits that have been encountered are the aforementioned /2, ag, l-^, and atv. Of this group of genes, ag and l-^ behave normally in back- crosses to L. esculentum, But /2 induces unexpected pleiotropic effects on floral parts, and the effects of atv seem to become diluted in the sense that the intense anthocyanin pigmentation develops on the herbage but not on fruits of backcross derivatives. Still other characters have been encountered in the derivatives, which were unknown in the wild parent. Examples of such "sur- prise" characters are elongate fruits, a thick, leathery calyx, and a pedicel joint that has normal morphology yet fails to abscise. The mode of inheritance of these characters has not yet been ascertained, but the limited available No. 44) RICK: GALAPAGOS SYMPOSIUM 73 data suggest simple genetic determination probably by recessive genes. Un- doubtedly such latent characters owe their appearance to a complementary in- teraction between genes from the wild parent and the genotype of the culti- vated tomato. Since extensive breeding tests are required to reveal these la- tent characters, only a slight beginning has been made in the survey of this source of new germ plasm. Natural Relationships According to all available evidence the races of Galapagos tomatoes are closely related among themselves. All genuinely native accessions pos- sess in common at least five unique traits, which are, except for B, unknown in any other wild tomatoes. No barrier, whether relating to incompatibility, vigor or fertility of F-j^or later generations, has been found to their interbreed- ing. The same freedom of hybridization has been revealed between them and L. esculentum and closely related species. If all the facts are taken into ac- count, separation of the group into different species can scarcely be justified. The taxonomic status of the whole group, on the other hand, is problematic. By morphological criteria alone, it should be recognized as a species. Yet the genetic evidence points to such a close relationship with L. esculentum that a rank above subspecies would not be justified: geographic isolation may be the only factor presenting merging and intergradation with forms of the latter species. Clearly then, although changes in the nomenclature are needed, any decision must be to some extent arbitrary. Since the taxonomy of the group is not the primary objective and complete taxonomic documentation not appro- priate here, no systematic revision is currently proposed. The above conclusions differ from those reached in an earlier study (Rick, 1956) based on herbarium material and experiments with living material of three accessions. The three were treated as L. esculentum vai. minor (LA- 317), L. pimpinellijolium (LA166), and L. esculentum var. cerasifoTme (LA- 292). That study suffered the shortcomings of inadequate living collections and lack of first-hand experience in the native habitat. In the meanwhile the opportunity to study living plants in many populations in the Galapagos and in experimental cultures has shown unequivocally that LA292, though correct- ly classified in 1956, is not native but likely excaped from nearby gardens in the same fashion as many other cultigens in the Galapagos. When this item is removed from the scene, homogeneity is evident between the two remaining collections and all other known truly native tomatoes. Origin of the Galapagos Tomatoes The conclusions reached in the earlier survey (Rick, 1956) concerning the origin of Galapagos tomatoes are largely fortified by the new evidence. In respect to the elimination of LA292 from consideration as a native form. 74 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers the picture is, in fact, simplified. These conclusions can be briefly summar- ized as follows. The closest approach to the Galapagos tomatoes in terms of morphology and genetic compatibility is found informs of L. pimpinellifolium from coastal Ecuador and Peru, although the former are well distinguished by their aforementioned exotic traits. Of these unusual features, orange fruit col- or (B), the accrescent calyx, and extreme hairiness of vegetative parts found in var. minor also appear in L. hirsutum, also native to the Ecuadorean coast. Whatever differentiation occurred in the origin of the Galapagos elements, it was expedited by their high level of self-pollination. The problem of origin has two aspects: determination of the ancestral forms from the mainland and differentiation of races on the Galapagos. In com- mon with most endemic groups of the archipelago, the two events might not have been independent; differentiation of races might have begun before mi- gration; although this process appears less likely than the migration of a sin- gle stem line. •; Any attempt to specify ancestral forms is fraught with the usual prob- lems of gaps in the record: absence of a reliable fossil record, uncertainties about the geological history, and others. If the new data contribute in any way, they reinforce the implication of L. pimpinellifolium and L. hirsutum or some extinct closely related forms. The former provides the closest approximation in total morphology to the Galapagos tomatoes; the latter could have furnished some of the exotic traits. The B gene previously found to be exclusive in L. hirsutum, LA166 and LA317 has been detected in every living collection from the Galapagos, and it seems safe to extrapolate that it exists in all Galapagos tomatoes. All accessions of var. minor, including living and herbarium mater- ial, possess in common the aforementioned accrescent calyx and excessive hairiness. In addition, one of the latent traits revealed in the new studies— a thick leathery calyx -points again to L. hirsutum. This new character is iden- tical in appearance with fleshy calyx (fl), a monogenic trait bred into L. escu- lentum from L. hirsutum (Butler, 1952), although a genetic test of identity has not yet been made. Taken together, the data suggest a relationship between the Galapagos tomatoes and L. hirsutum. Little more is revealed, however, and it is highly uncertain how they came to share these genes. Since L. hirsu- tum and L. pimpinellifolium differ extensively in a great welter of other morph- ological and physiological characters, it is tempting to suppose that hybridi- zations between the progenitors of the two led to introgression of a few genes from the former into the latter to form the stem line of the Galapagos tomatoes. It was pointed out in the early survey that self-pollination could have played a key role in the differentiation of the group. The recently acquired evidence from progeny tests, population variability, and activity of insect vectors reinforces the conclusion that the Galapagos tomatoes are very highly self-pollinated. Such a reproductive system guarantees a maximum opportun- No. 44) RICK: GALAPAGOS SYMPOSIUM 75 ity for rapid differentiation of biotypes and for fixation of genes. It could also account for the anomalous frequency of fixed monogenic abnormalities appear- ing in some or all individuals of certain populations. Nothing remotely com- parable to the fixation in whole populations of such genes as / l-i , or ag is known in any other wild tomatoes. Although guessing at the adaptive value of such characters is always hazardous, it is tempting to suppose that the sur- vival in similar environments of genes that on the one hand control loss (ag) and, on the other, intensification (atv) of anthocyanin is a random process. Also it is difficult superficially to perceive a selective advantage for partial loss of chlorophyll f/]^) or complete loss of pedicel articulation C/o). Genetic variation within populations was found at only very low levels in the intensively studied examples. Otherwise the only evidence of appre- ciable genetic variability was found inLA438 on the coast of Albemarle. Even in this instance, however, the variants were not products of immediate genetic segregation, for progenies of the tested plants bred true. The responsible gen- etic segregation, if any, must have occurred earlier, possibly many genera- tions earlier. The unique variation in this population is accompanied by an- other unique feature: the sympatric occurrence of var. minor and Gal. ppn. in the vicinity of LA438. This remarkable coincidence strongly suggests that the observed variation stemmed from introgression between the latter two races. The mode of derivation of the various races of Galapagos tomatoes re- mains highly conjectural. The high rate of self-fertilization and the widespread and occasional sympatric distribution of the three dominant races suggest that these races originated by natural selection from an ancient, variable stem line. Segregants from the introgression of L. pimpinellifolium and L. hirsutum and/or natural mutation might have provided the variable milieu from which these successful races emerged. From their original sites they could have spread together or separately throughout the archipelago. The aspects of dis- persal and colonization of these races are discussed in the sections on dis- tribution and seed dormancy. The other, rarer races might have originated in the same fashion or by introgression from sporadic hybridization as suggested by LA438. If the proposed origin of the Galapagos tomatoes in L. pimpinellifolium and L. hirsutum is correct, attention would be directed to the immediate main- land of Ecuador and north coast of Peru. Although forms of L. hirsutum occur elsewhere, the ones most compatible with the L.escw/erz^ww complex are found in the above region. In this respect the tomatoes provide another example of the many ecological and botanical ties between these continental areas and the Galapagos outlined by Svenson (1946). Transport to the islands from the adjacent coast is favored by several oceanographic features. In the first place, the prevailing Humboldt Current sweeps northwestward alongthe coast of Peru, and Ecuador whence it veers westward toward the Galapagos. In the second place, the ocean floor rises to form the Carnegie Ridge, which extends from 76 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers the Galapagos toward continental Ecuador, being separated from the latter by a narrow but deep trench (Shumway, 1954). If this ridge emerged above sea level at an appropriate time it might have aided the migrations of tomatoes and other forms of life to the Galapagos. Summary The Galapagos tomatoes comprise a group of closely related races pos- sessing in common five exotic morphological and physiological traits, in which they differ from most other species of the genus Lycopersicon. The three com- monest races are relatively well distributed throughout the archipelago and account for nearly all individuals of known populations. All observations on floral structure, pollination mechanisms in nature and in culture, population structure, and progeny tests point to a very high level of self-pollination. Appreciable genetic variability between individuals was discovered in one wild population, although the individuals bred true for their deviating phenotypes in progeny tests. Otherwise the extent of genetic variation between plants in a single population is so low that most individuals conform to a single phenotype. Various recessive genes of well distinguished phenotype are fixed throughout certain populations and appear in a small pro- portion of individuals of others. Considerable genetic differentiation is evi- dent between populations but is not much greater between islands than between populations on the same island. The inbreeding consequent to automatic self- pollination readily accounts for such population characteristics. Seeds of Galapagos tomatoes are minute and seldom germinate without treatments that remove or soften the seed coats. Passage through the digest- ive tract of the Galapagos tortoise improves germination, revealing a likely mechanism of natural dispersal and establishment. All tested Galapagos accessions are completely inter-compatible and cross-compatible with L. esculentum, L. pimpinellifolium, and other closely related entities. All such combinations tested yield viable, completely fer- tile F^ , Fo, and BC hybrids. Like other members of the L. esculentum com- plex, they are separated from L. peruvianum and L. chilense by severe com- patibility barriers. In common with all other known species, the Galapagos tomatoes have twelve pairs of chromosomes. According to the available evidence, all genuinely native Galapagos to- matoes should be treated as sub-specific forms of a single species of the L. esculentum complex. While they are approached most closely in morphology by L. pimpinellifolium, certain other traits are to be found in L. hirsutum, both of these species being native to the adjacent mainland. On the basis of these and other facts, suggestions are presented to account for the origin and differ- entiation of the Galapagos tomatoes. Rapid evolution of diverse races was undoubtedly promoted by the strictly autogamous breeding system. Dependence No. 44) RICK: GALAPAGOS SYMPOSIUM 77 upon animal digestion for establishment might account for the widespread dis- persal of certain races. Literature Cited Anderson, E. 1949. Introgressive hybridization. Wiley, New York. 109 pp. BEEBE, W. 1924. Gala'pagos: world's end. G.P.Putnam's Sons, London. 443 pp. Butler, L. 1952. The linkage map of the tomato. Journal of Heredity, vol. 43, pp. 25-35. Howell, J. T. 1941. The genua Scalesia. Proceedings of the California Ac ademy of Sc iences, 4th ser., vol.22, pp. 221-271. HURD, P. D., Jr. 1958. The carpenter bees of the eastern P acific Oc ean islands. Journal of the Kansas Entomological Society, vol.31, pp. 249-255. LUCKWILL, L. C. 1943. The genus Lycopersicon; an historical, biological, and taxonomic sur- vey of the wild and cultivated tomatoes. Aberdeen University Stud- ies, 120 pp. Muller, C. H. 1940. A revision of the genus Lycopersicon. United States Department of Agri- culture Miscellaneous Publications, 382 pp. RICK, C. M. 1950. Pollination relations of Lycopersicon esculentum in native and foreign regions. Evolution, vol. 4, pp. 110-122. 1956. Genetic and systematic studies on accessions of Lycopersicon from the Galapagos Islands. American Journal of Botany, vol. 43, pp. 687-696. 1960. Hybridization between Lycopersicon esculentum and Solanumpennellii: phylogenetic and cytogenetic significance. Proceedings of the Na- tional Academy of Sciences, vol. 46, pp. 78-82. RICK, C. M., AND R. L Bowman 1961. Galapagos tomatoes and tortoises. Evolution, vol. 15, pp. 407-417. Shumway, G. 1954. Carnegie Ridge and Cocos Ridge in the east equatorial Pacific. Journal of Geology, vol. 62, PP. 573-586. SVENSON, H. K. 1946. Vegetation of the coast of Ecuador and Peru and its relation to the Ga- lapagos Islands. I. Geographical relations of the flora. American Journal of Botany, vol. 33, pp. 394-426. COMPOSITION AND RELATIONSHIP OF THE TERRESTRIAL FAUNAS OF EASTER, JUAN FERNANDEZ, DESVENTURADAS, AND GALAPAGOS ISLANDS * Guillermo Kuschel Centra de Investigaciones Zoologicas Universidad de Chile Santiago, Chile South America has a great number of offshore islands, particularly in the fjord region of southern Chile, but there are also other truly oceanic is- lands lying far from the mainland and composed entirely of volcanic material. In this paper I shall attempt to present a general account of the composition of the terrestrial faunas of the oceanic islands, and to show their relationships with the faunas of other regions, before discussing the origin and possible antiquity of their older elements. Easter Island, Juan Fernandez, Desventura- das, and the Galapagos will be considered here, but the isolated Salay Gomez and the Cocos and Malpelo islands will not be dealt with because they are not sufficiently well known. In table 1 is shown the location of the islands to be considered, together with other basic data (see also fig. 1). Easter Island This remote island has a low and uniform topography relieved by a few craters which have no human record of volcanic activity. It is fairly arid, hav- ing no streams, lakes, or swampland, and showing surface water only in the depths of three of the craters. Its climate is warm-temperate, with its rainfall evenly distributed through the year (Cfa in the Koppen-Geiger classification). It is the only island of the four with a native human population. Man and his domestic animals, particularly sheep, have been largely responsible for the impoverishment of the flora and for this reason Easter Island has only 31 species of flowering plants. Skottsberg, in 1928, was therefore able to say ♦ Presented at the TENTH PACIFIC SCIENCE CONGRESS of the Pacific Science Association, held at the University of Hawaii, Honolulu, Hawaii, U. S. A., 21 August to 6 September 1961, and sponsored by the NATIONAL ACADEMY OF SCIENCES, BERNICE Pauahi Bishop Museum, and the University of Hawaii. -79- 80 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers ^ -© PASCUA (Easter) Lat. 27° 10' Long. 109° 26' Area 118 km^ Alt. 530 m Rainfall 1149 mm Temp. 20.4 C GALAPAGOS Lat. 00°00' Long. 89° 00' Area 7643 km^ Alt. ISOO m Rainfall 665 mm Temp. 23.6 C PANAMA 10 COLOMBIA ECUADOR - 5 - 10 - 15 PERU DESVENTURADAS Lat. 26° 19' Long. 79° 47' Area 7 km ^ Alt. 478 m Rainfall + 600 mm Temp. 17.3 C JUAN FERNANDEZ Lat. 33° 44' Long. 78° 50' Area 185 km Alt. 1500 m Rainfall 1152 mm Temp. 15.3 C - 20 25 30 35 CHILE - 40 45 50 - 55 No. 44) KUSCHEL: GALAPAGOS SYMPOSIUM 81 Table 1. Data on four island groups of the southeastern Pacific Ocean. PASCUA JUAN FERNANDEZ DESVENTLIRADAS GALAPAGOS (Easterls.) Latitude 27° 10' S 33° 37' S 26° 21' S 00° 00' Longitude 109° 26' W 78° 52' W 79° 47' W 89° 00' Area (km ) 118 185 7 7643 Max. altitude (m) 530 1500 478 1500 Ann. Temp. C. 20.4 15.3 17.3 23.6 Rainfall (mm) 1149.3 1152.2 600 665 Distance from continent (km ) 3760 666 859 950 in all fairness, "there does not exist another Island of the size of Easter and with such a fine climate where the native flora is so poor." Although we would not expect a rich fauna on Easter Island it is quite clear that the fauna has become further impoverished and it is possible to point to certain features which, in this respect, distinguish Easter Island from others under consideration: (1) The cosmopolitan or wide-spread element in its fauna is very high, (2) The Indo-Malayan, or Pacific, element is also appreciable. (3) Of the 79 terrestrial species of animal life only six are endemic and the position of even some of these is still debatable. These endemic species are: a) Chrysopa skottsbergi Esben Petersen, 1924 (Neuroptera) There is nothing published on its relationships with other species but it was most likely introduced from South America. b) Lipsana insulae-paschalis Enderlein, 1940 (Diptera) For this species Enderlein created a genus but gave no account at all of its relation to other genera. He also raised a new sub-family which he put in the Lonchaeidae. c) Bidessus skottsbergi Zimmermann, 1924 (Coleoptera: Dytiscidae) This is said to be extremely close to two Australian species. d) P acindonus bryani (Swezey, 1921) (Coleoptera: Curculionidae: Cossoninae) The genus Pacindonus Kuschel is of Indo-Madagassian origin but is widely distributed through the Pacific. It has a dozen species or so, and is not represented in the Americas or in the cold-temperate regions. Figure 1. Schematic position ofthe islands. Vertical line: South American con- tinent, with the parallels. Horizontal line: the equator. Arrows: main relationships of the terrestrial fauna. 82 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Table 2. Climatological data for Easter Island. LATITUDE 27° 10' S LONGITUDE 109° 26* W ALTITUDE 41 m Mean max. Mean Temp. Mean m in. Relative Rainfall temp. C. C. temp. C. Humid it y mm . January 27.0 23.1 19.1 78 104.9 February 28.2 23.7 19.6 77 78.6 March 27.4 23.1 19.3 77 100.9 April 25.5 21.5 17.8 76 120.7 May 23.4 19.9 16.9 80 114.7 June 21.9 18.3 15.1 81 116.5 July 21.4 17.8 14.6 83 88.6 August 21.5 17.8 14.7 83 85.7 September 22.1 18.1 14.5 82 75.7 October 23.3 19.1 15.2 81 70.6 November 24.2 20.2 16.3 83 90.5 Dec ember 25.5 21.8 17.9 85 101.9 Annual 24.3 20.4 16.8 81 1,149.3 Years of observation 34 34 34 28 47 e) Tetragnatha paschae Berland, 1924 (Araneae) The genus is of world-wide distribution and this species may pos- sibly occur in other Pacific Islands. f) Melampus pascus Odhner, 1922 (Mollusca) This is of Oriental origin. Of the six endemic species so far known, three are of Oriental origin, one of Australian origin, and the origin of two is as yet undetermined. So far, then, there is no proven South American element in the fauna, and Easter Is- land must be considered as one of the Pacific islands which have the Indo- Malayan element as the strongest in their faunas. Juan Fernandez This is really a small archipelago comprising the islands of Masatierra, Santa Clara, and Masafuera, and Santa Clara may be counted a continuation of Masatierra, for it is separated from the western tip of the latter only by a narrow channel. Both Masatierra and Masafuera have a dense but varied cover of vegetation and this is particularly well developed around the higher hills of Masatierra. Only 170 km. (92 miles) separate Masatierra and Masafuera, yet each has its own endemisms. Masafuera tops Masatierra by 600 metres, No. 44) KUSCHEL: GALAPAGOS SYMPOSIUM 83 Table 3- Composition of the terrestrial fauna of Easter Island. ENDEMICS S. AMERICA ORIENTAL AUSTR- NEOZ. WIDE SPREAD UNDEFINED Oligochae ta .. • ■ 1 • • Isopoda • • • • • • .. 2 Myriapoda •• •• •• 2 1 Insecta (4) (-) (8) (1) (42) (2) Coll embola ,. .. 1 Odonata . . .. .. 1 Blattariae .. 3 1 Orthoptera .. 1 ■ • Embioptera .. 1 • • Dermaptera .. • • 1 Thysanoptera ■ • .. 1 Psocoptera • • .. • • 1 Hemiptera .. • • 9 Ne uroptera 1 .. 1 Lepidoptera .. 2 4 Diptera 1 .. 8 Coleop tera 2 1 8 Hymenoptera •• 7 1 Araneae 1 o 4 2 Mollusca 1 1 3 •• ■• Vprtebrata •• O •• •• -• Total 6 13 4 52 7 Table 4. Relationships of the fauna of Easter Island. S. AMERICA AUSTRALIA ORIENTAL- POLYNESIAN WIDE SPREAD UNDEFINED Number of species Percentage •• 2 2.53 16 20.25 52 65.82 9 11.39 its peaks are covered with snow in winter, and light falls of snow are common in summer. It is on these summits that the interesting Magellanic element of its flora is found. Both islands have some permanent streams but they lack lakes and swampland. There are no volcanic craters aDthough an eruption a century and a quarter ago (1835) took place off the extreme south-east of Cumberland Bay on Masatierra close to the coast. Juan Fernandez has a warm temperate climate with rain throughout the year, but with considerably more precipitation in winter than summer (Csb2 84 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Table 5- Climatological data of Juan Fernandez (Masatierra). LATITUDE 33° 37' S LONGITUDE 78° 52' W ALTITUDE 6 m. Mean max. Mean temp. Mean m i n . Rel ative Rainfall temp. C. c. temp. C . Humidity mm. January 21,7 18.4 15.2 74 24.9 February 22.0 18.9 15.6 73 30.1 March 21.0 18.3 15.1 74 39.8 April 19.7 16.8 13.4 77 82.4 May 17.8 15.2 12.0 79 149.0 June 16.0 13.7 10.5 78 160.3 July 15.1 12.9 9.8 80 142.0 August 14.7 12.3 9.2 79 113.8 September 15.2 12.4 9.4 77 76.8 October 16.2 13.3 10.2 76 54.5 November 18.0 15.0 11.7 74 34.2 Dec ember 20.0 17.0 14.0 73 26.4 Annual 17.9 15.3 12.2 76 1,152.2 Years of observation 35 35 35 35 48 in the Koppen-Geiger classification). The lower-lying areas at some distance from the hills have a very dry summer and an arid soil that applies to the ex- treme west of Masatierra and to Santa Clara. The flora includes 147 species of flowering plants. There are no native species of amphibians, reptiles, fresh-water fish, or mammals, but there are nine species of land-birds, all of which are related to the Chilean fauna. According toSkottsberg (1956) there are 147 species of flowering plants on Juan Fernandez, and of these 101 are endemic. He divided the 147 species into 6 elements as below: Andine Chilean element Subantartic Magellanian element Neotropical element Pacific element Atlantic-S. African element Eu-Fernandezian element It is most interesting to note here that the incidence of endemism, on a percentage basis, is very similar in the insects and flowering plants. It is not yet possible in the case of the insects to achieve the same precision of class- ification by origin which Skottsberg made for the plants, but the literature shows clearly enough that the southern Chilean element predominates strong- 69 46.9% 15 10.2% 19 12.9% 26 17.7% 6 4.1% 12 8.2% No. 44) KUSCHEL: GALAPAGOS SYMPOSIUM 85 Table 6. Composition of the terrestrial Arthropod fauna of Juan Fernandez. ENDEMIC ENDEMIC NON-ENDEMIC UNDEFINED TOTAL NUMBER GENERA SPECIES SPECIES SPECIES OF SPECIES Isopoda 1 8 3 , , 11 Myriapoda .. 3 6 9 Insecta (54) (440) (170) (77) (687) CoUembola . 3 5 4 12 Thysanura . 2 .. 2 Orthoptera . 2 1 • 1 4 Dermap tera . .• 1 1 Isoptera . 1 .. .. 1 P so cop tera . .. .. 4 4 Thysanopt era . 2 2 2 6 Hemip tera 8 32 7 9 48 Neuroptera 1 4 1 5 Trichoptera .. 2 1 3 Lepidoptera 12 63 23 2 88 Diptera 2 102 73 18 193 Coleoptera 20 191 38 6 235 Hymenoptera 1 1 38 15 31 84 Arane ae • • 13 6 .. 19 Pseudoscorpionida 4 10 .. 1 11 A cari 1 26 2 • • 28 ly. Second in importance, as in the case of the plants, comes the Pacific ele- ment, mainly Indo-Malayan. Desventuradas These consist of the two tiny islands of San Felix and San Ambrosio and the islet called Gonzalez. This group lies 777 km. (420 miles) to the north of Juan Fernandez and 859 km. (464 miles) from the nearest point of the mainland. San Ambrosio is larger, measuring roughly 4 km. long by 1 km. wide. Its coastline is very steeply countered and its upper parts constitute a pla- teau, lying between 300 and 478 metres above sea level, which is dissected by a few small and shallow ravines. Surface water and filtrations through the walls are totally absent except immediately after rain and for this reason the island is uninhabitable. There are no meteorological data for the Desventur- adas but it is known that there is abundant rain in winter, while from October to March no rain falls and the summit is quite free from low clouds. The prevailing wind is from the south-south-east. The vegetation is halophytic and though abundant on the plateau it is very scarce on the cliffs of the island. The island has only 19 native species of phanerogams and one moss, and there are no ferns at all. Terrestrial vertebrates are non-existent, 86 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Table 7. Percen tage of endemisms of Insecta and Phanerogamae. ENDEMIC NON-ENDEMIC UNDEFINED TOTAL NUMBER Insec ta Phanerogamae 64.43 68.7 24.35 21.3 11.22 687 147 either native or introduced, but there are seven species of sea birds which nest on the island and there is one land-bird which is probably only the Juan Fernandez Sparrow Hawk (Falco sparverius fernandensis Chapman) which is most likely a recent immigrant. San Ambrosio is one of the very few islands which has so far avoided invasionby any terrestrial vertebrates or by any weeds and has totally escaped the effects of fire. San Felix Island is only a little, way to the west of San Ambrosio. It is smaller, lower (170 m. summit), and is less steeply sloped, it is much more arid and its vegetation is very sparse. It has some eight species of phanero- gams, two of which are endemic. Table 8. Composition of the terrestrial Arthropod fauna of San Ambrosio. ENDEMIC ENDEMIC NON-ENDEMIC UNDEFINED TOTAL NUMBER GENERA SPECIES SPECIES SPECIES OF SPECIES Isopoda 2 • • •• 0 Chilopoda • • .. •• 2 2 Insecta (4) (16) (21) (37) (74) Collembola 1 1 Thysanura 1 .. 1 2 Orthop tera .. 1 1 •• 2 P socoptera .. .. • • 3 3 Thysanoptera .. .. 1 1 Hemiptera 1 4 4 9 Lepidoptera .. 21 21 Dip tera 1 2 7 3 12 Coleop tera 3 9 6 1 16 Hymenoptera •• 3 4 7 Arane ae .. 5 2 y Pseudoscorpionida 2 .. 2 Acari • • (20) (20) Gamasides 3 3 Uropodina .. 1 1 Trombidiformes 5 5 Acaridiae 1 1 Oribatei •• 10 10 No. 44) KUSCHEL: GALAPAGOS SYMPOSIUM 87 A point worth noting in table 8 is that the lepidopterous species out- number both the Diptera and the Coleoptera. As the first intensive collecting of the small animal life took place only at the end of 1960, much of the material yet remains to be determined by spe- cialists. Although the flora of the island only slightly resembles that of Juan Fernandez, the fauna is much more closely related. The percentage of endem- ism seems. to be appreciably lower than on Juan Fernandez, but there are not- able examples of endemism in both species and genera. Our knowledge to date enables us to distinguish four endemic genera in the Insecta, a genus of Calli- phoridae related to Callyntropyga of Juan Fernandez, two genera of Curculion- idae (Cossoninae) found only on Thamnosens and related to the Juan Fernan- dez fauna, and a carabid of problematical relationships, but certainly with no closely related genera in Juan Fernandez or on the Chilean mainland. A few other genera, formerly considered endemic in Juan Fernandez, are also repre- sented on San Ambrosio, although not by the same species. It cannot be doubted that the terrestrial fauna of the Islas Desventuradas is closely related to that of Juan Fernandez unlike the flora which is only distantly related as a whole. Galapagos These islands form an archipelago situated on the Equator some 950 km. (510) miles) from the nearest South American coast. They comprise 15 largish islands together with a host of smaller islands. Their total area is more than 40 times greater than Juan Fernandez and they offer a wider variety of envi- ronments, with arid and moist areas and with an altogether richer vegetation as they have nearly four times the number of plant species. The terrestrial vertebrate fauna is comparably well known but the same cannot be said for the invertebrates. Information on the invertebrates is very widely scattered, so that it is extremely difficult to obtain a concise overall picture of the Galapagos fauna, but I have been able to collate all known pa- pers on the Coleoptera and this is given in table 11 where it is compared with data for Juan Fernandez. Within any geographical zone there is a clear enough observable ratio of flowering plants to insects. The ratio varies with the incidence of endem- ism. On the basis of this ratio and the number of plants known to exist in the Galapagos Islands one would expect to find at least 550 to 800 species of Coleoptera, but up to now there are recorded only 190 species; that is to say, between a third and a quarter of the expected number. Moreover, the usual ratio of Carabidae to Curculionidae is roughly 1:5 and it is extremely strange that more Carabidae than Curculionidae are recorded for the Galapagos. From these facts we can only deduce that the Galapagos fauna has been selectively and incompletely collected. 88 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Table 9. Climatological data of San Cristobal (Chatham I.}, Galapagos. LATITUDE 00° 54* S LONGITUDE 08° 37' W ALTITUDE 2.7 Me an max. Me an temp. Mean min. Relative Rainfall temp. C . C. temp. C . Humidity mm. January 28.7 24.8 21.9 81 5.5 February 29.8 25.8 22.9 82 190.1 March 29.8 25.9 22.6 83 241.5 April 30.0 25.9 22.8 84 141.8 May 28.6 24.9 22.2 80 14.9 June 27.8 23.8 21.4 77 5.2 July 26.4 22.7 20.7 78 6.8 August 25.6 21.7 19.6 79 8.4 September 25.0 21.1 19.0 79 6.0 October 25.5 21.7 19.2 76 6.8 November 25.8 22.2 19.9 75 10.8 December 27.1 23.2 20.9 77 26.6 Annual 27.5 23.6 21.1 79 664.4 Years of observation 3 7 3 5 3 There is also a universally observed ratio between the endemism of phanagerogams and of the fauna, this being particularly true of the insects. From the table comparing the Coleoptera of the Galapagos and Juan Fernan- dez we can see that the incidence of endemism of Coleoptera species is about the same for both groups of islands, but for plants it is about 48 per cent for the Galapagos and is 68.7 per cent for Juan Fernandez; this seems to point to wrong interpretations in plant or Coleoptera studies somewhere, and the neces- sity for an intensive collection of invertebrates in the Galapagos is very ap- parent. The Gala'pagos are further from the mainland than are the Desventuradas and Juan Fernandez and both flora and fauna are typically Neotropical, with few exceptions. A high percentage of the species of the Galapagos fauna has been found, by many authors, to be most closely related to the faunas of Cen- tral America, Mexico, and the Caribbean islands; but this is due more than anything to our ignorance of much of the fauna between Panama and northern Peru. Of the fauna I have personally been able to examine (Curculionidae), I am quite certain that the most closely related mainland species are those which live immediately opposite the Galapagos. There are some remarkable examples of Galapagos species being extremely closely related to species on Puna Island in the mouth of the Golfo de Guayaquil. No. 44) KUSCHEL: GALAPAGOS SYMPOSIUM 89 Table 10. Climatological data of Seymour Island, Galapagos. Mean max. Mean temp. Mean min. Rainfall temp. C. c. temp. C. mm. J anuary 29.4 2 5.9 22.5 16.5 February 30.0 26.8 23.6 29.7 March 30.1 27.0 23.8 16.5 April 30.0 26.7 23.5 20.6 May 29.2 25.7 22.2 1.0 June 28.6 25.1 21.7 0.2 July 27.1 24.0 20.8 0.2 August 27.1 23.3 19.7 0.2 September 26.8 23.1 19.5 0.0 October 27.1 23.3 19.6 0.0 November 27.5 23.8 20.1 0.0 December 28.2 24.6 21.0 0.0 Annual 28.3 24.8 21.4 85.6 Years of observat ion 5 5 5 5 The occurrence ofendemism in species and plants in the Galapagos is much lower than in Juan Fernandez. The Origin of the Faunas of Juan Fernandez, Desventuradas, and Galapagos As has been mentioned earlier, all these islands have today a purely volcanic bedrock and their degree of erosion has led geologists unanimously to believe that the islands cannot possibly be older than the Pliocene; that is to say, that they are not much more than ten million years old. We have to ask if this relatively short time, by geological standards, is sufficient to account for the clear difference observed between the faunas of the islands. Geographically the Galapagos, Desventuradas, and Juan Fernandez are roughly the same distance from the South American coast and the line joining them is not far out of parallel with the line of the Andes. This makes one wonder if they might not have a common tectonic origin. It is also worthy of note that petrographic studies have shown very close similarity between the rocks of Masafuera and the Galapagos. On the origin of life in these islands much has been written but the opin- ions expressed have differed greatly, expecially with regard to the Galapagos. There are those who passionately maintain that winds, sea currents, and birds 90 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers are much more effective dispersal agents than is commonly conceded. That such agents do play a part in the repopulation of islands is not doubted, but as far as these particular islands are concerned it is easy enough to point to a series of components of the fauna which could not have been so transported, at least not across the natural barriers as they are today. I shall attempt to tackle this problem, beginning with Juan Fernandez and the Desventuradas. Geological studies have shown that the Chilean coast south of Valpa- raiso had more or less the same coastline in the late Cretaceous as at the present. During the Eocene, however, we know that an important movement occurred between the 38th and 45th parallels, as this region completely lacks Eocene sedimentary rocks, and the next marine sediments to appear are of late Oligocene age, beginning with Navidad strata. This means that during the Eocene and most of the Oligocene there was formed a continuous ridge of land which cut off the present coast from the sea. This unstable mass was called "Land of Juan Fernandez" by Briiggen (1950) and this distinguished geolo- gist presumed that this land extended obliquely northwards to Juan Fernan- dez, embracing the Desventuradas, for between these islands there is today a submarine chain nowhere deeper than 1.430 m. It is difficult to reconcile the proposed recent origin of the islands with our certain knowledge that Juan Fernandez and the Desventuradas contain an abundance of primitive elements in their flora and fauna and it is easier to ac- cept Brilggen's hypothesis of the islands' origin. That is that, far from being of Pliocene or Pleistocene age, the islands are the relics of an older exten- sive land-mass which had direct connection with the continent or at least was very much closer to it than are the islands today, so permitting dispersal of the flora and fauna until the very late Tertiary when it is supposed a final subsidence isolated the present-day islands which are basically volcanic cones. On the basis of Briiggen's hypothesis, which is founded on some geolo- gical facts, Skottsberg was able to understand and explain the Eocene flora of Juan Fernandez and also the high incidence of endemisms in genera and species. The occurrence of a subantarctic flora on the peaks of Masafuera, however, is still not satisfactorily explained for the ocean currents cannot be considered a likely means of transporting plants to island peaks, the present- day winds do not lie in the right direction, and we do not now observe bird mi- grations between southern Chile and Juan Fernandez. On the other hand, it is interesting to note that the subantarctic element in the flora is not paralleled in the fauna. As far as the terrestrial fauna is concerned, with its remarkable degree of generic and specific endemism and its high percentage of species related to those in southern Chile, the hypothetical "Land of Juan Fernandez" accounts completely for its presence in the islands. Nevertheless, I would like to draw attention to two facts of importance: No. 44) KUSCHEL: GALAPAGOS SYMPOSIUM 91 (1) The number of species of Coleoptera in the leaf litter is startlingly low despite the favorable conditions for development of such a faunal group. Most of the elements characteristic of southern Chile are missing, this being especially true of the Pselaphidae, Melandryidae, and Curculionidae(Crypto- rhynchinae.) (2) Among the flying insects, such as the Neuroptera and especially the Diptera, a large number of species are common to Juan Fernandez and south- ern Chile, or at least extremely closely interrelated, but many of these are found only in forests or their surrounds and not near the Chilean ports from which ships set sail for the islands. This makes it almost impossible for these species to have been transported to the islands by man 's agency. While on the subject of flying insects, I might here interpolate some mention of certain birds which can scarcely have reached the islands under their own power in conditions as they are today; I have in mind such species as Spizitornis fer- nandezianus, Cinclodes oustaleti baeckstroemi, and Aphrastura masafuerae. The first point, concerning the Coleoptera fauna of the leaf litter, seems to suggest that the Eocene fauna of the leaf litter lacked this element or, and perhaps more likely, that the "Land of Juan Fernandez" was not entirely con- tinuous but interrupted here and there sufficiently to prevent the dispersal of those species belonging exclusively to the soil. We can hope to get nearer the truth of this matter once the hypogeous fauna has been carefully studied. There is another geological fact worth mentioning here. Along the whole length of coast from the 5th parallel (Payta) in northern Peru to just beyond the 32nd parallel (to the North of Valparaiso), there are found no marine sedi- ments from the Eocene to the early Pliocene. This means that throughout this considerable period the coastline must have lain in what is now the Pacific Ocean, but we know nothing about this addition to the continental margin, which disappeared only in the late Pliocene. It is worth repeating here that the flora of San Felix and San Ambrosio is most closely related to that of Ata- cama, the mainland region immediately opposite the islands, and I might add that it is extremely difficult to account for this by means of the usual agents of dispersion, for neither atmospheric nor sea currents run in the required di- rection and birds do not now migrate from Atacama towards the islands. On the origin of life on the Galapagos, if I may move on to this last group of islands, much has been written and one might almost say that the number of opinions equals the number of authors. If one compares its flora and fauna with that of Juan Fernandez, it is immediately apparent that its in- cidence of generic and specific endemisms is proportionally lower, and also, that its species do not display, as a whole, so many primitive features. On these grounds we do not expect such an early origin of life in the Galapagos as in Juan Fernandez and the Desventuradas. As there is an ample evidence of important geological changes throughout the whole of the Tertiary along 92 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Table 11. The Coleoptera o( Galapagos and J nan Fernandez compared. Galapagos Juan Fernandez Genera Endemic genera Species Endemic species Genera Endemic genera Species Endemic species Cicindelidae 1 .. 2 2 ,, .. ., ,, Carabidae 8 ... 23 23 7 1 22 18 Dytiscidae 4 •• 4 1 3 1 3 2 Gyrinidae 1 .. 1 1 ,, , , .. . . Hydrophilidae 3 1 4 2 , , .. .. ., Limnebiidae 1 .. 1 1 ,, , , .. .. Staphylinidae 2 .. 2 1 16 7 20 16 Ptiliidae r. .. , , .. 3 , , 5 .. Scaphidiidae .. .. , , , , 1 , , 1 Histeridae 2 .. 4 2 1 , , 1 .. Passalidae 1 .. 1 ,, ,, .. .. .. Scarabaeidae 3 1 5 4 IC) .. 2(?) . , Trogidae 1 .. 1 .« ,, .. ., , , Cantharidae 1 1 , . ,, , , , , , , Lycidae 1 .. 1 .. .. , , .. .. Meloidae 1 .. 1 .. .. . . -- .. Mordellidae 1 .. 1 1 .. .. .. Tenebrionidae 9 2 40 37 3 ,. 3 1 Alleculidae 3 .. 4 3 .. ,. .. .. Monotomidae 1 .. 1 1 .. .. .. Oedemeridae 2 .. 5 4 .. ,, .. .. Nitidulidae 2 .. 2 1 1 ,, 5 5 Cucujidae 1 .. 1 1 .. .. .. .. Cryptophagidae 1 .. .. .. 4 2 8 7 Languriidae 1 .. 1 , , .. .. .. .. Cisidae .. .. .. .. 1 3 3 Lathridiidae .. .. .. ., 4 ,, 4 4 Colydiidae 1 .. 1 .. 2 ., 4 4 Mycetophagidae • • .. ,, , , 1 1 .. Elateridae 7 .. 14 12 1 1 .. Melasidae .. .. ,, , , 1 1 1 Buprestidae 2 • • 3 3 .. .. •• • • Ostomidae 2 .. 2 1 1 1 3 3 Cleridae 2 .. 2 1 1 1 .. • • Dasytidae 1 1 1 1 • • .. .. .. Dermestidae 1 .. 2 • • 1 .. 1 .. Anobiidae 3 .. 4 4 5 1 8 2 Bostrichidae (+Lyctidae) 3 .. 3 •• 2 .. 2 • • Coccinellidae 5 .. 5 3 2 .. 3 1 Cerambycidae 12 .. 18 14 • • • • •• •• Chrysomelidae 4 1 5 5 1 1 3 3 Bruchidae 2 .. 2 2 •• • • •• •• Anthribidae 1 ,, 1 1 1 1 1 1 Curculionidae 9 .. 17 16 12 6 126 120 Scolytidae 1 .. 1 1 2 • • 2 •• Platypodidae 1 •• 1 1 •• •• •• •• Total 107 6 190 152 79 21 235 191 No. 44) KUSCHEL: GALAPAGOS SYMPOSIUM 93 Table 11. Continued Galapagos Juan Fernandez Genera Endemic genera Species Endemic species Genera Endemic genera Species Endemic species Families (total) 39 27 Families with endemic subspec ies 31 16 Percentage of genera 94.4 5.6 73.5 26.5 Percentage of species 20 80 18.74 81.26 the rest of the South American coast, we might expect there to have been sim- ilar disturbances along the coast of Ecuador and Colombia and, in fact, there is evidence of changes in short level in some fossiliferous raised beaches of late Tertiary age. The two submarine ridges, the Cocos Ridge which runs from Costa Rica to the north of the Galapagos, and the Carnegie Ridge which runs from Ecuador to the same islands, suggest a possible former union or Closer proximity of islands and mainland. Shumway, however, discounts the former ridge by saying "the apparent absence of truncated seamounts on Cocos Ridge is evidence against a former emergence." Of the other submarine feature he says "on the basis of the bathymetry of Carnegie Ridge and the geological history of Colombia and Ecuador, it is speculated that the easternmost por- tion of Carnegie Ridge may once have been part of the western borderland for the early Tertiary geosyncline which existed along the present coastal low- land of western Ecuador. This would have shortened the sea route to the Ga- lapagos Islands and possibly facilitated the rafting of plants and animals to the Islands." The foregoing shows that there is yet no geological evidence of a land connection between the Galapagos and the continent in recent times and, in fact, dates any such connection as early Tertiary at the latest. This, however, hardly does more than deepen the mystery as the very close relationships be- tween animal and plant species on the Galapagos and the continent are most striking and unquestionable, while other species on both mainland and islands are identical. If the origin of life in the Galapagos is placed as far back as the early Tertiary, then it seems impossible to understand the much lower de- gree of speciation and specialization of the terrestrial organisms on the Gala- pagos in comparison with those on Juan Fernandez, especially as the more favorable climate produces a higher turnover of genetic material. One would therefore expect to find an even greater difference in species between the Ga- lapagos and the mainland than between Juan Fernandez and the continent. If 94 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers I might be permitted to express a personal view, I would say without much hesitation after having studied my own group of the fauna of all the islands, and making the considerable assumption that one can extrapolate from this group, that the Galapagos' fauna is considerably younger than that of Juan Fernandez and the Desventuradas. I would also say that the fauna of these last two remote groups of islands mostly dates back to the Eocene and part of of the Oligocene, while the Galapagos fauna, including the terrestrial verte- brates, might go back only to the Pliocene or, even, to the end of the Plio- cene and to the Pleistocene. Summary After a brief description of Easter Island, Juan Fernandez, the Desven- turadas, and the Galapagos, there follows a general account of the terrestrial faunas and their relationships with other biogeographic regions. All the base- ment rock now visible on the islands is volcanic and young in geological time, probably of the late Pliocene. The incidence of endemisms in species, genera, and even higher sys- tematic categories is considerably more in Juan Fernandez and the Desven- turadas than in the Galapagos and it is concluded that most of the life on these former groups is of more ancient origin than in the Galapagos. Geological evi- dence suggests that during the Eocene there existed between parallels 38 and 45 along what is now the coast of Chile, an extension of the land surface which probably embraced Juan Fernandez and the Desventuradas. The struc- ture of the basement rocks of these islands consisting of basalts and tuffs, also the small degree of erosion and denudation of the islands, do not indicate an age earlier than Pliocene for these two groups of islands. It is concluded that the last remnants of Briiggen's "Land of Juan Fernandez" can only have disappeared finally beneath the sea in very recent times and this supposition is sufficient to explain the presence of the younger elements of the flora and fauna on these islands. The date of the origin of life on the Galapagos is still an open field for speculation for, while sea bottom soundings have provided no evidence for the presence of emerged land of any sort between the continent and islands dur- ing the late Tertiary, the greater part of the Galapagos fauna, if not all of it, is relatively young. No. 44) KUSCHEL: GALAPAGOS SYMPOSIUM 95 Literature Cited Bruggen, J. 1950. Fundamentos de la Geologia de Chile. Santiago, ed. 2, pp. 1-510, il- lustr. and maps. Shumway, G. 1954. Carnegie Ridge and Cocos Ridge in the east equatorial Pacific. Jour- nal of Geology, vol. 62, pp. 573-586. SKOTTSBERG, C. ED. 1920-1956. History of Juan Fernandez and Easter Island, 3 vols., 688 pp. Upp- sala, Almquisl & Wiksells Boktryckeri-A.-B. SKOTTSBERG, C. 1949. Die Flora der Desventuradas Inseln (SanFelix und San Ambrosio), Goete- borgs Kungl. Vetenskapoch Vitterh els-Sam ha ell esHandlingar-Femte Foeljden, Ser. B, vol. 6, p. 3-88, 1937. (Spanish edition in Boletin Museo Nacional de Historia Natural, Santiago, vol. 24, pp. 1-64. 1945. The Juan Fernandez and Desventuradas islands, in Plants and Plant Science in Latin America, Waltham, Massachusetts, p. 150-153. 1951. Weitere Beitraege zur Flora der Insel San Ambrosio (Islas Desventur- adas, Chile), Arkiv for Botanik, ser. 2, vol. 1, no. 12, pp. 453-459. 1957. The vegetation of the Juan Fernandez and Desventuradas Islands. Pro- ceedings of the 8th Pacific Science Congress, vol.4, pp.l81-185» THE MARINE SHORE-FISHES OF THE GALAPAGOS ISLANDS* Richard H. Rosenblatt Scripps Institution of Oceanography La Jolla, California and Boyd W. Walker Department of Zoology University of California Los Angeles, California Introduction The Galapagos Islands have long been of interest to biologists as a na- tural laboratory for the study of the effects of isolation on the evolution of ter- restrial organisms. It has perhaps not been so clearly understood that the 650 miles of deep water separating the Galapagos from the South \merican main- land forms a barrier to the dispersal of shore-dwelling marine organisms as well. An analysis of the fish fauna indicates that the barrier has been effec- tive. The material presented here is based on a manuscript checklist of Gal- apagos fishes, compiled from a critical review of the literature, a re-examin- ation of much of the previously existing material, and records from recent col- lections. Owing to the changes necessary, our analysis is quite different from any based on published lists (Snodgrass and Heller, 1905; Fowler, 1938). The Environment The Galapagos Archipelago is a group of oceanic islands, consisting of 13 principal islands and a host of islets and rocks, lying some 650 miles west of Ecuador. The main portion of the archipelago is located between the equa- tor and 1°5' South latitude. Two small islands, Culpepper and Wenman, are separated by some 80 miles from the main body of the group. These, and a group of three small islands lying some 30 miles from the main group, are sep- arated from the major islands by deep water. The largest island, Albemarle, is about 80 miles long by 30 miles wide, but the others are considerably smaller. • Presented at the TENTH PACIFIC SCIENCE CONGRESS of the Pacific Science Association, held at the University of Hawaii, Honolulu, Hawaii, U.S.A., 21 August to 6 September 1961, and sponsored by the NATIONAL ACADEMY OF SCIENCES, BERNICE Pauahi Bishop Museum, and the University of Hawaii. -97- 98 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers The Galapagos are volcanic in origin, and geological opinion seems to favor the interpretation that they have never been connected with the main- land. Fossils of Pliocene age have been found, but the group is probably con- siderably older. As might be suspected from their volcanic origin, the Galapagos Islands are characterized by rocky shores, although sand and coral-gravel beaches are present. The bottom sediments at moderate depths are predominantly coarse. The Albatross and Velero station records indicate that the bottom is sand, rocks, or coral at almost all stations. Only once was mud encountered, at 70- 80 fathoms off Daphne Minor Island (Townsend, 1901, Fraser, 1943). The Galapagos Islands lie in the South Equatorial Current, which is composed mainly of cold Peru Current water, but with a component of warm water from the Equatorial Countercurrent to the north (fig. 1). The heterogeneous devia- tion of the waters bathing the Galapagos causes great variability in tempera- ture. Differences of 5° C. (Beebe, 1924) and 11° C. (Garth, 1946) have been reported for the two sides of Albemarle Island. At irregular intervals, the so- called "El Niiio" years, the Peru Current is deflected far to the west of South America. Warm water from Central America then sweeps far to the south and causes extensive warming, causing fish kills along the Peruvian coast. At these times, the tropical component in the waters surrounding the Galapagos must be greatly increased (Schott, 1931; Posner, 1957). Fourteen such El Nino years have been recorded since 1791, the most recent being in 1958. It is unfortunate that no systematic oceanographic work has been done at the Galapagos Islands. Little or nothing is known of short- or long-term fluc- tuations in temperature, and nothing is known of variability from island to is- land. Composition of the Fish Fauna The marine fish fauna of the Galapagos Islands is in large part typical of the eastern tropical Pacific faunal region, but it is characterized by a high degree of endemism among the shore species. Twenty-three per cent of the shore forms are confined to these islands. This endemism, plus significant elements from the transitional fauna between thePanamic and Chilean faunas, and from the western Pacific, distinguish the Galapagos fauna as a separate subunit of the Panamic faunal province. The fish fauna of the eastern tropical Pacific region (American Pacific Warm-Water Region of Ekman, 1953) is characterized by a high degree of en- demism at the species level. With the exception of the circumtropical species (Briggs, 1961), most of which are pelagic, almost all of the species found in this area are limited to it. Owing to the influence ofthe cold Peru and California currents, the ex- tent of the tropical regions is much curtailed on the western side of the Amer- icas. The limits of the tropical fauna are at about 25° N. latitude on the outer No. 44) ROSENBLATT & WALKER: GALAPAGOS SYMPOSIUM 99 T 10= 0° 10= ^ WINTER .NORTH EQUATORIAL CURRENT COST/i RICA \TO^I EQUATORIAL COUNTERCURRENT ^ ^ SOUTH EQUATORIAL CURRENT GALAPAGOS 'V ^^ PERU 120= ^ SUMMER 10' >7 ■^EQUATORIAL COUNTERCURRENT J" J ^;A\ t^.. COSTA RICA 0= SOUTH EQUATORIAL CURRENT GALAPAGOS'^'-' 10' ^ %PERU Figure 1. Current systems in the vicinity of the Galapagos Islands. (After Schott,1931 and Cromwell and Bennett, 1959). coast of Baja California and about 5°S. on the Peruvian coast. This is in marked contrast to the situation in the western Atlantic, where the boundaries of the tropical region are at about 35° N. and 35° S. The eastern tropical Pacific fauna is relatively depauperate in compar- ison with other tropical regions. Probably only the west African fauna is poor- er (Ekman, 1953, p. 56). This fauna is, however, characterized by a great de- velopment of fishes characteristic of muddy or sandy bottoms. The families Engraulididae (anchovies), Ariidae (marine catfishes), and Sciaenidae (croak- ers) are particularly well developed. In contrast, the fishes primarily adapted 100 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers to coral reefs are very poorly represented. This impoverishment is especially pronounced in the Labridae (wrasses), Chaetodontidae (butterflyfishes), Scar- idae (parrotfishes), and Acanthuridae (surgeonfishes). The relationships of the fishes of the eastern tropical Pacific lie main- ly with the western Atlantic fauna. A very large number of genera are limited to the New World tropics. This relationship carries through on a suprageneric level as well. The blennioid family Chaenopsidae (Stephens, MS) is found on- ly in this region, as is the trachinoid family Dactyloscopidae (stargazers. In addition, the atherinid (silverside) subfamily Atherinopsinae (of Jordan and Hubbs, 1919) is restricted to the New World, as is the gobiesocid (clingfish) subfamily Gobiesocinae. The clinid subfamily Labrisominae has a similar distribution, except for two obviously derivative forms found in west Africa, This basic unity of the American fish fauna (which led Ekman, 1953, p. 30, to term it the Atlanto-East Pacific Fauna) is due to the presence of a Tertiary Central American water gap (Durham and Allison, 1960). During the existence of this connection the faunas on the two sides of the Americas must have been very similar, although probably not identical. The differences which we see now are due in large part to differentiation since the destruction of the water gap, and some migration into the eastern Pacific by Indo-West Pacific species. The Galapagos ichthyofauna is relatively large as compared with that of the other oceanic islands of the eastern Pacific. This enrichment is es- pecially noteworthy in the Serranidae and in certain families that are charac- teristic of sandy shores, such as the Gerridae, Pomadasyidae, and Sciaenidae. We record 269 species from 88 families. Seventy-five percent are shore forms and twenty-five per cent are pelagic or coastal pelagic. The few deep sea forms that have been taken near the Galapagos Islands are not considered. Most of the species found on the Galapagos (60 per cent) are found at other localities in the eastern Pacific, but their ranges do not extend else- where. By far the largest segment of these (53 percent of the total fauna) are eastern tropical Pacific endemics, and most are forms ranging widely through- out the region. This large segment would seem to determine the proper affini- ties of the fauna. Eight per cent are pantropic. There is small but notable representation (six per cent of the total fauna) of forms characteristic of the transitional fauna between the Panamic and Chilean regions. The presence of these forms indicates the profound effect of the Peru Current on these es- sentially equatorial islands. Less than two per cent of the species are in common with the Atlantic and these are also found in other eastern tropical Pacific localities. Only five species (about two per cent) are found only on the Galapagos and other eastern tropical Pacific offshore islands. Endemism Eighteen per cent of the species are endemic but only one of these (a flying fish of doubtful validity) is pelagic. The endemism is largely confined No. 44) ROSENBLATT & WALKER: GALAPAGOS SYMPOSIUM 101 to the shore fishes; twenty-three per cent of these forms are peculiar to the Galapagos. Further, endemism is not evenly distributed between families. Over a third of the families with endemic species have two or more such forms. And the number of endemics does not correlate with the family representation. The Serranidae (basses) and the Carangidae (jacks) are the largest families; each has 17 representatives. There are but three endemic serranids, and all of the carangids have been taken elsewhere. The pomacentrids (demoiselles) are the next most speciose group with nine representatives, but only one of these is limited to the Galapagos group. There are seven labrids (wrasses), but all have been taken elsewhere. On the other hand, three of the five sciae- nids (croakers) are endemic, as are five of the nine pomadasyids (grunts) and clinids (klipfishes). This difference in degree of differentiation correlates with a differen- tial in vagility. Carangids are strong swimmers, and many live pelagically. Serranid larvae are well suited to pelagic life (Smith, 1959), and at least some have special modifications for it. Pomacentrids also have a pelagic prejuven- ile (Hubbs, 1958) stage in their life history. There is no evidence that the labrids have a long larval period, but many of the forms are crevice dwellers and nibblers, and thus would seem eminently suited to rafting. The sciaenids and pomadasyids are, on the other hand, mostly charac- teristic of shallow waters in muddy or sandy bays and along open sandy beach- es. Larvae of these groups are almost never taken in offshore collections, and very small juveniles are found schooling near the bottom. Clinids are small, demersal fishes of rocky reefs. The adults are sed- entary, their eggs are usually demersal, and the larval period is short. Forms with this type of life history are poor candidates for transport by currents. The one clinid which seems to be completely undifferentiated (Labrisomus multi- porosus) has the widest geographic distribution of any Pacific species in the genus, and also seems to have a longer larval life than is usual for the group (Hubbs, 1953). Most of the endemic species of the Galapagos are strongly differentia- ted from their congeners. It is thus difficult to pinpoint any mainland species as a sibling or possible ancestor. It is possible, however, to determine the group (usually generic) to which the Galapagos forms are most closely related. When the data was analyzed in this fashion we find that of the 46 Galapagos endemics, 34 are most closely related to eastern tropical Pacific species. Five are related to forms characteristic of the Peruvian-Chilean warm temper- ate, two are derived from Indo-West Pacific groups, and one is allied to a west Atlantic species. The relationships of three were too questionable to in- clude. This is in general accord with the overall composition of the Galapagos ichthyofauna. 102 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Analysis of Faunal Relationships The Galapagos are not particularly close faunistically to the other east- ern Pacific oceanic islands, Cocos, Clipperton, and the Islas Revillagigedo. It is true that their faunas have certain common characteristics as opposed to that of the adjacent mainland areas. They are rich (for this area) in pelagic types and in Indo-West Pacific species, relatively rich in rocky shore forms, and poor in sandy shore and muddy bottom forms—that is, they are unbalanced. In addition the endemics tend to show certain parallel modifications. We do not feel that these similarities imply any real relationship however. The rela- tive faunal imbalance reflects two factors. One is the deep water between the islands and the mainland, which acts as a filter bridge which excludes a large number of species. The other is the nature of the insular environment, where deep waters are found close to the shore, and the bottom tends to be mostly rocky. In the same way the parallel morphological modifications toward more terete bodies, longer fins, and an increase in number and length of gill-rakers can be ascribed to adaptation to the island environment. Snodgrass and Heller (1905) listed seven forms which occurred at two or more of the islands, and termed them "Eastern Pacific Insular species." Further collecting indicates that the ranges of four: Prionurus laticlavius, Melichthys radula, Halichoeres nicholsi, and Lutjanus viridis, extend to the mainland. Pomacentrus leucorus is restricted to the Revillagigedo Islands and is replaced at Isla del Coco and the Galapagos by Pomacentrus beebei (Loren P. Woods, personal communication). Pomacentrus arcifrons and Apogon atradorsata still are known only from the Galapagos and Isla del Coco. The presence of these three species on Cocos and the Galapagos indi- cates that there must have been some faunal transfer in the recent past. It is almost certain that these species originated on Cocos Island, since transport is possible only south from Cocos. The 15 species derived from the warm temperate fauna of Peru and Chile represent an element not found on the other islands. The presence of these species and the five endemics with their affinities in this area is related to the Peru Current which flows from the mainland toward the Galapagos. Of the 24 Indo-West Pacific forms in the Galapagos fauna (9 per cent) only two have not been found at other localities in the eastern tropical Paci- fic. However, one of these is a burrowing eel, otherwise known only from the type locality, and the other is a pelagic flying fish which likely has a wider eastern Pacific distribution. Of the remaining 22, 4 are found only on the other oceanic islands and 18 are found on the mainland as well. Although the abso- lute numbers of Indo-Pacific forms are about the same at the Galapagos and the mainland, these species form a more conspicuous element in the depau- perate Galapagos fauna. As might be expected, most of these Indo-West Pacific forms are well No. 44) ROSENBLATT & WALKER: GALAPAGOS SYMPOSIUM 103 adapted to transport by currents. Of the 24 forms, 7 have long pelagic larval stages and 12 are pelagic as adults as well. Five others are well suited to rafting, by accompanying floating debris. The two remaining forms are a labrid (wrasse) and a scarid (parrot fish). Little is known of the early life history of these groups, but both the young and adults are nibblers and known to frequent areas of algal growth. Thus they might be good candidates for trans- port by rafting. These conclusions are in general agreement with those of Briggs (1961) and Hubbs and Rosenblatt (manuscript). The evidence indicates that there has been little, if any, direct inva- sion of the Galapagos Islands from the west. Only two currents impinge on the Galapagos Islands. The flow of the South Equatorial Current is to the west. The recently discovered Cromwell Current flows from west to east, but it is an undercurrent and may well be too deep to be of any use to the larvae or young of offshore forms (Knauss, 1960, and personal communication). Clipper- ton and Cocos islands, which are far to the north in the path of the east-flow- ing Equatorial Countercurrent, have several Indo-Pacific species which have not been taken elsewhere in the eastern Pacific. If the Cromwell Current were indeed carrying trans-Pacific migrants, it might be expected that the Galapa- gos Islands would have a similar number of such forms not found elsewhere. This expectation is heightened by the prevailing current which flows strongly away from the mainland. The opposite is, of course, the case. All but two of the western Pacific migrants found at the Galapagos occur at other eastern Pacific localities. The distributional data and the known current patterns in the eastern Pacific thus indicate that the Galapagos received its Indo-Pacific elements secondarily from the mainland and Clipperton and Cocos islands. During De- cember, January, and February there is a southwest flow of water from the "Panama Bight" into the South Equatorial Current (Cromwell and Bennett, 1959). The possibility for transport would be much greater during El Nino years, when the flow of warm water from the north is much greater (Schott, 1931; Posner, 1957). Garth (1946) distinguished a Gulf of California element in the brachy- uran crab fauna. To explain the uniquely known occurrence at the Galapagos, and in the Gulf of California, of 11 species of crabs, he found it necessary to invoke a drastic alteration in the current systems of the Pacific. More recent- ly (Garth, 1960), this was considered part of a general eastern Pacific island relationship. As we have indicated earlier, we find no such relationships in the fishes. Further, we see no reason to expect them. The juggling of ocean currents to explain distribution patterns is as fraught with hazards as the erec- tion of land bridges, and should be approached with equal circumspection. It seems probable that intensive collecting efforts along the mainland coast would reduce considerably the number of species known only from the Gulf and the Galapagos. 104 CALIFORNIA ACADEMY OF SCIENCES (Occ Papers Another concept in need of re-examination is that of the Galapagos ori- gin of certain Central American species. The burden of Garth's argument was that a species now occurring at Galapagos and Central America, but without an Atlantic analogue, must have evolved at the Galapagos. Its absence from the Atlantic is evidence that it did not invade the Central American mainland until after the closure of the Central American water gap. However, the postulate that the western Atlantic and eastern Pacific had completely common faunas during the existence of a connection between the Atlantic and Pacific is not compelling. Certainly the connection was a shallow one, and probably muddy (Schuchert, 1935). Such a connection might be of little use to a species restricted to the rocky shore habitat. And these are the species which should be particularly suited to life at the Galapagos, where rocky habitat predominates. The fauna of the Gulf of California indicates that continuity of coast- line in itself does not indicate faunal homogeneity. Certainly the Gulf is in communication with the rest of the Pacific and yet there are endemic species of fishes and crabs, and a number of Panamic species are missing (Walker, 1960; Garth, 1960). One could also point to the existence of a large number of Central American species of fishes which have no Atlantic analogs and which are not found at the Galapagos or any other oceanic island. If they did not use the Central American water gap, or if they evolved subsequent to its closure, may this also not be true of those species which are found at the Galapagos? This line of reasoning does not prove that the Galapagos have not been the center of origin of certain Central American species, but we be- lieve it indicates that the evidence for such an origin is not strong. Literature Cited Beebe, William 1924. Galapagos: world's end. G.P.Putnam's Sons, New York, xxi+443 pp., frontis, pis. 1-8. Briggs, John C. 1961. The East Pacific Barrier and the distribution of marine shore fishes. Evolution, vol. 15, no. 4, pp. 545-554. Cromwell, T., and E. B. Bennett 1959. Surface drift charts for the Eastern Tropical P acific Ocean. Inter-Amer- ican Tropical Tuna Commiss ion, Bulletin, vol. 3, no. 5, pp. 217-235 (English): pp. 236-238 (Spanish). No. 44) ROSENBLATT & WALKER: GALAPAGOS SYMPOSIUM 105 Durham, J. W., and E.G. Allison I960. The geologic history of Baja California and its marine faunas. System- atic Zoology, vol. 9, no. 2, pp. 47-91. Ekman, Sven 1953. Zoogeography of the sea. Sidgwick and Jackson, Ltd., London, xiv + 417 pp. Fowler, Henry W. 1938. The fishes of the George Vanderbilt South Pacific Expedition, 1937. Academy of Natural Science of Philadelphia, Monograph no. 2, v + 342 pp., pis. 1-12. Fraser, C. McLean 1943. General account ofthe scientific work of the Velero III in the eastern Pacific, 1931-41. Part 111. A ten-year list of the Velero III col- lecting stations. Allan Hancock Pacific Expeditions, vol. 1, no. 3, pp. 259-431, charts 1-115. Garth, John S. 1946. Distribution studies of Galapagos Brachyura. Allan Hancock Pacific Expeditions, vol. 5, no. 11, pp. 603-638, charts 1-10. 1960. Distribution and affinities of the brachyuran Crustacea. Systematic Zoology, vol. 9, no. 3, pp. 100-123. HuBBS, Carl L. 1958. Dikellorhynchus and Kanazawaichthys: nominal fish genera interpreted as based on prejuveniles of Malacanthus and Antennarius, respec- tively, Copeia, no. 4, pp. 282-285. HuBBS, Clark 1953. Revision of the eastern Pacific fishes of the clinid genus Labrisomus. Zoologica, New York Zoological Society, vol. 3 8, pt. 3, pp. 113-136. Jordan, D. S., and C. L. Hubbs 1919. Studies in ichthyology: A monographic review of the family Atherinidae or silversides. Leland Stanford Junior University Publications, University Series, pp. 1-87, 12 pis. Knauss, J. 1960. Measurements of the Cromwell Current. Deep Sea Research, vol. 6, no. 4, pp. 265-286. Posner, G. S. 1957. The Peru Current. Bulletin of the Bingham Oceanographic Collection, vol. 16, no. 2, pp. 106-155. Schott, G. 1931. Der Peru-Strom und seine nbrdlichen, Nachbargebie te in normaler und anorm aler ■ Ausbildung. Annalen der Hydrographie und Marltimen Meteorologie, vol. 59, pp. 161-169, 200-213, 240-257, pis. 17-22. Schuchert, C. 1935. Historical geology of the Antillean-Caribbean region or the lands bor- dering the Gulf of Mexico and the Caribbean Sea. John Wiley, New York, 811 pp. Smith, C. L., Jr. 1959. A revision of the American groupers {Epinephelus and related genera). Microfilm-Xerox copy of Ph.D. thesis. University of Michigan, xiv + 563 pp. 106 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Snodgrass, R. E., and E. Heller 1905. Shore fishes of the Re villagigedo, Clipperton, Cocos and Galapagos Is- lands. Proceedings ofthe Washington Academy of Sciences, vol. 6, pp. 333-427. TOWNSEND, C. H. 1901. Dredging and other records ofthe United States Fish Commission Steam- er Albatross, with bibliography relative to the work of the vessel. United States Commission of Fish and Fisheries, Commissioners Report for 1900, pp. 387-562, pi. 1-7. WALKER, BOYD W. 1960. The distribution and affinities of the marine fish fauna of the Gulf of California. Systematic Zoology, vol. 9, no. 3, pp. 123-133. EVOLUTIONARY PATTERNS IN DARWIN'S FINCHES ♦ Robert I. Bowman Department of Biology San Francisco State College San Francisco, California Introduction Darwin's finches of the Galapagos Islands l represent one of the best known groups of Galapagoan animals. Indeed, they are famous out of all pro- portion to their size because of their remarkable degree of adaptive radiation in feeding habits and associated structures — perhaps one of the best exam- ples of this phenomenon in the class Aves (fig. 1). Furthermore, many of the species show such an unusual range of variation in structure that to the cas- ual observer there would seem to be a nearly continuous intergradation of forms (fig. 2). It was this very fact which greatly impressed the young Charles Darwin when, in 1835, he first encountered these birds in their native habitat. In his "Journal of Researches," second edition (1845), he wrote: "Seeing this gradation and diversity of structure in one small, intimately related group of birds, one might really fancy that from an original paucity of birds in this archipelago, one species had been taken and modified for different ends." Patterns of evolution in Darwin's finches are, fundamentally, adapta- tions to the environment. In the past it has been customary to explain evolu- tion as largely or entirely determined by factors of the external environment acting through selection. Until fairly recently much less attention has been paid to the inherent properties and potentialities of peculiar genetic systems, which are the limiting factors of the organism's internal environment (Mayr, 1945; White, 1948). Surely, all evolutionary patterns result from the opportun- istic interaction of factors of the external environment and genetic systems. * Presented at the TENTH PACIFIC SCIENCE CONGRESS of the Pacific Science Association, held at the University of Hawaii, Honolulu, Hawaii, U.S. A., 21 August to 6 September 1961, and sponsored by the NATIONAL ACADEMY OF SCIENCES, BERNICE Pauahi Bishop Museum, and the University of Hawaii. 1 One species of Darwin's finch, the honeycreeper-finch (Pinaroloxias inornata), resides only on Cocos Island, Costa Rica. It is, unquestionably, a member of the Geospizinae subfamily of fringillid finches. (See Swarth, 1931.) - 107- 108 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Figure 1. The pattern of adaptive radiation in Darwin's finches. Figure 2. Variations in shape of bill in 14 species of Geospizinae. A, Geospiza magni- rostris (Tower Island); B, Geospiza magnirostris (James Island); C, Geospiza conirostris (Hood Island); D, Geospiza fortis (Albemarle Island); E, Geospiza conirostris (Gardner-near-Hood Is- land); F, Geospiza foTtis (Chatham Island); G, Geospiza /or//s (Charles Island); H, Geospiza ful- iginosa (Abingdon Island); I, Geospiza difficilis (Indefatigable Island); J, Certhidea olivacea (Indefatigable Island); K, Pinaroloxias inornata (Cocos Island); L, Platyspiza crassirostris (Charles Island); M, Camarhynchus psittacula (James Island); N, Camarhynchus psittacula(Bind- loe Island); O, Cactospiza pallida (Chatham Island); P, Camarhynchus pauper (Charles Island); Q, Camarhynchus psittacula (Albemarle Island); R, Camarhynchus parvulus (Indefatigable Is- land); S, Cactospiza pallida (James Island); T, Geospiza scandens (Abingdon Island); U, Geo- spiza scandens (James Island); V, Geospiza fuliginosa (Chatham Island), After Swarth, 1931. No. 44) BOWMAN: GALAPAGOS SYMPOSIUM 109 A O B O c O D O E O M no CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers In Darwin's finches our study of evolutionary patterns is, necessarily, restricted in fact to the observable end-products of a long unrecorded history: There is, of course, considerable basis in theory for a divergence of opinion regarding the significance of the facts and on the nature and importance of the biological processes involved; for example, the meaning of intraspecific variation in bill structure, the importance of the chance factor ("random gen- etic drift") in accounting for minor interisland differences, and the role of in- terspecific "competition" in shaping the feeding niches of sympatric "sib- bling" species. It is not surprising, therefore, that several authors have differed in their biological interpretation of this remarkable group of birds (Lowe, 1930, and Stresemann, 1936; Swarth, 1931, and Lack, 1945, 1947; and Bowman, 1961). We must await the results of many more studies, and especially those of an ecological and ethological nature, before our understanding of the total biol- ogy of Darwin's finches is anywhere nearly complete. The remarks which follow are concerned mainly with the results of evo- lution; that is, the adaptations by which the Galapagos finches appear to have "solved" the problems of survival in their particular insular environment. Some of the views presented here have been discussed more fully elsewhere (Bowman, 1961). ^^^^--^^^___^ C^^C^ ZONE WIDTH AS A PERCENT OF TOTAL LENGTH OF TRANSECT FROM ACADEMY BAY TO HIGHEST PEAK ALTITUDINAL LIMITS OF ZONES ALONG TRANSECT (IN FEET ABOVE SEA LEVEL) Figure 3. Altitudinal zonation of the vegetation on the south side of Indefatigable Island, from Academy Bay to the top. Diagrammatic. No. 44) BOWMAN: GALAPAGOS SYMPOSIUM 111 The Galapagos Environment Basic to my discussion is an understanding of the Galapagos environ- ment, about which the following facts are pertinent. The various islands are remarkably multiform, superficial appearances to the contrary. They differ in their geologic age (Chubb, 1933), size and ele- vation, and distance from each other (table 1). The climatic conditions in Gal- apagos are somewhat anomalous for the tropical Pacific Ocean largely owing to the modifying effects of the cool Humboldt Current, which surrounds the is- lands. Whereas there are marked differences in average precipitation both sea- sonally and yearly, mean daily air temperatures at sea-level vary only two or three degrees throughout the year (Alpert, 1946). Correlated with interisland differences in geologic age, climatic expos- ure, and time of last volcanic activity, are the interisland dissimilarities in species composition, growth-form, and relative abundance of the vegetation (table 1), as well as the physical character of the substratum (color, amount of soil, and mulch). As is characteristic of most oceanic islands, the flora and fauna are depauperate in the sense that certain of the typical mainland groups are ab- sent. With regard to the plants, an individual community is not rich in species, but because of the local diversity in topography, soil, and moisture, there are many different plant communities within a single island and between the sev- eral islands (Howell, 1942; Stewart, 1911, 1915). In response to the climatol- ogical and pedological features associated with an increase in elevation, there results an altitudinal zonation of the vegetation on the higher islands. This phenomenon is most clearly and elaborately demonstrated on the south side of Indefatigable Island from Academy Bay inland (fig. 3). Close to shore there is an "Arid Coastal Zone" dominated by cacti, deciduous shrubs, and dwarf trees (fig. 4); the general aspect is light grey in color, except for a pale green cast during the rainy season. At slightly higher elevations there is an open forest growth called the "Transition Zone" (fig. 5); this is a region of inter- gradation of plants typical of the higher and lower regions. At still greater elevations a non-deciduous forest prevails, dominated by the tree-composite Scalesia pedunculata and a dense undergrowth of shrubs (fig. 6). The xero- phytic nature of the vegetation of the "Scalesia Forest Zone" is most appar- ent during periods of drought (Svenson, 1946). Above the Scalesia forest two plant species, Psidium galapageium and Zanthoxylum fagara, retain their tree stature and constitute the "Brown Zone." Above the Brown Zone on Indefat- igable Island there occurs a uniqueshrub formation termed the "Miconia Belt," about 6 to 10 feet high, composed principally of the fern Pteridium sp. and the endemic melanostome shrub Miconia robinsoniana (fig. 8). The highest peaks are treeless and densely vegetated with low growing herbs and ferns, constituting the "Upland Zone" (fig. 7). 112 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Table 1. Distribution of 13 species of Geospizinae and comparative data on 16 main islands of the Galapagos Archipelago. Species of Geospizinae c o -o W) c IS < to E aj < c c M C to o c re U 5 a U I- ttj a, a. u a U re 0 c C 0 0 -a CO M a c OJ E a > C u 0 re 3 0 E 'Ti 0 c re E c Geospiza magnirostris foTtis fuliginosa difficilis scandens conirostris X X X X X X X X X X X X X X X X X Xa X X ..f X X X ••g X Xb X Xb X X X X ..c X X X X X Xh X X X X X X X X X X X X X Xi X X X X X ..d X X X ..d Xe X Platyspiza crassirostris X X X X X X X X X X Camarhynchus psittacula pauper parvulus X X X X X X X X X X ..j X • • X X •• X X X X X X Xk X X X • • XI Cactospiza pallida heliobates ■■ X X •• •• ..m X • • X •• X X X ..n X X •• Certhidea olivacea X X X X X X X X .V X X X X X X X Total number of species per island.... q 10 7 7 9 •7 4 9 3 10 10 9 9 8 4 5 Relative size of island 8 1 11 ( 6 5 15 13 9 2 4 14 3 11 12 16 Highest eleva- tion (ft.) ° o C<1 o o in in o in 00 in 1 — 1 o o r— 1 0 in ro CM 0 in in 0 in 0 in in en CO CM CM m 0 CM 0 0 0 0 CM 0 CM 0 m CO Vegetation P zones 1-3 1-4 1 1 1-4 1-4 1 1-2 1 1-4 1-4 1 1-4 1 1 1 Total number vascular plants 119 329 48 47 319 306 (' 103 79 193 224 42 80 52 22 14 Relative size of flora 6 1 1 12 0 J 16 / 9 5 4 13 8 10 14 15 Distance from Indefatigable Island s 75 17 10 54 31 42 162 6 55 . • 12 15 59 Mo 58 139 No. 44) BOWMAN: GALAPAGOS SYMPOSIUM 113 One individual collected in September, 1957. Compare with Lack (1945, pp. 9-10) and Swarth (1931, PP- 146-147). b Compare Swarth (1931, pp. 149-150, 206( and Lack (1945, pp. 8-9). See Gifford (1919, p. 227) and Swarth (1931, p. 138, 164). d Reported from island but probably not a permanent resident. ^ See Swarth (1931, p. 174). See Lack (1945, pp. 14-15). ^ See Lack (1945, p. 14). h , . , . Now considered extinct. Numerous individuals collected by author in September, 1957. See alsoGifford (1919, p. 238). Now considered extinct on this island. See Lack (1945, p. 17), Rothschild and Hartert (1899, p. 167), and Gifford (1919, p. 246). The wTiter observed this species on Narborough Island during September, 1957. The only pre- vious record is that of Snodgrass and Heller (1904, p. 286) based on a single specimen. Swarth (193L P- 223) refers this specimen to the form "af/inis". See Gifford (1919, p. 250). One specimen in collection of the California Academy of Sciences (see Swarth, 1931, p. 249). See Gifford (1919, p. 254). Elevations from United States Hydrographic Office Map no. 1798, 11th ed., 1946, except for Seymour Island. From Stewart (1911, 1915): 1, dry; 2, transition; 3, moist; 4, grassy. '^ From Stewart (1911, p. 237). n P The smaller the number, the larger (relatively) the flora. Distances measured in English miles using the American Geographical Society map N.A.-17, 1949 edition. Darwin's finches reside on all of the main islands of the Galapagos group, although the number of species represented and their relative abundance differ from island to island (table 1). In addition to the 13 species of finches, there is a spotty occurrence of mainland groups of land birds, including the following resident species: the Yellow Warbler (Dendroica petechia aureola), the thrasher-like Mockingbirds (four island species of the endemic genus Ne- somimus), the Martin (Progne modesta), three flycatchers (two species of Py- Tocephalus and one species of Myiarchus), acuc\ioo(Coccyzus melacoryphus), and an endemic dove (Nesopelia galapagoensis). Two North American mi- grants, the Bobolink (Dolichonyx oryzivoms) and the Barn Swallow (Hirundo erythrogaster), are regular winter visitants to Galapagos (Swarth, 1931). 114 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Figure 4. Vegetation of the Arid Coastal Zone near Academy Bay, Indefatigable Island. (Photo courtesy R. Freund.) No. 44) BOWMAN: GALAPAGOS SYMPOSIUM 115 Figure 5. Vegetation of the Transition Zone, two miles north of Academy Bay. 116 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Figure 6. Vegetation of the Scalesia Forest Zone, six miles north of Academy Bay. Figure 7. Vegetation of the Upland Zone, highlands north of Academy Bay. (Photo cour- tesy R. Freund.) No. 44) BOWMAN: GALAPAGOS SYMPOSIUM 117 Concerning predators on the finches, the following three species of ver- tebrates are the most important (fig. 9): a hawk (Buteo galapagoensis), the Short-eared Owl ( Asio galapagoensis), and a colubrid snake (Dromicus biser- ialis). Evolutionary Patterns The major patterns of differentiation in Darwin's finches appear to be concerned in the main with adaptations for food-getting. Guided by selection, the exploitation of the constitutional and ecological opportunities has resulted in an impressive array of food-niche specializations. The adjustments of the organism to the different ways of feeding have involved not only the basic feeding mechanism (bill, tongue, palate, jaw muscles, stomach, and intestine) but also correlated features (plumage, heart, and behavior). a. Structural pattern in the bill. Darwin's finches are most readily distinguished on the basis of the relative size and shape of the bill. The 14 species may be grouped into six genera as follows: Geospiza (6 species), Camarhynchus (3 species), Cactospiza (2 species), Platyspiza (1 species), Certhidea (1 species), and Pinaroloxias (1 species). Structural features and mechanical potentialities of the six bill types are summarized in table 2. In brief, the seed crushing bill of Geospiza, which may be likened to a heavy duty linesman's pliers (fig. 10), is capable of its greatest adduction at the base. In Camarhynchus, the bill resembles a high leverage diagonal pliers with high cutting potential at the tip, and is used to cut into tough woody tissues wherein insect larvae are to be found. The elongated bill of Cacto- spiza shows similarities in structure to a pair of long-chain-nose pliers and serves both for tip-biting and probing while excavating in woody tissues for insects. The genus Platyspiza is characterized by a bill that basically is similar to a parrot-head gripping pliers, with crushing potentialities more or less equally distributed along its length. The bill of Certhidea, somewhat analagous to a needle-nose pliers, resembles in great detail the bills of cer- tain parulid warblers, and is suited for probing crevices in search of small in- sect food. The bill of Pinaroloxias is decurved and slender and resembles grossly a pair of curved needle-nose pliers. It is well suited for procuring soft foods such as nectar and insects, and for piercing fleshy fruits. Species differences in the bills of Geospiza concern the absolute size and depth-to-length ratio, which features determine the mechanical potential- ities of the bill. The interspecific differences may more readily be apprecia- ted if we compare four sympatric species of Geospiza from Indefatigable Is- land (fig. 11). Especially instructive here is the nature of the individual var- iation in size of bill in G. fortis, which in its smallest version is but a trifle larger than G. fuliginosa, and in its largest version only slightly smaller than G. magnirostris. But despite this impressive size difference within a single 118 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers No. 44) BOWMAN: GALAPAGOS SYMPOSIUM 119 Table 2. Bill features in six genera of Geospizinae. GENUS BILL CHARACTERISTICS* BILL CAPABILITIES Geospiza Culmen convex. Strong biting at tip. Gonys essentially straight. Very strong crushing at base. Bill large to small, conical elongate. Camarhynchus Culmen convex (usually strongly so). Very strong biting at tip. Gonys convex (slightly to strongly). Strong crushing at base. Bill short, laterally compressed. Cactospiza Culmen slightly convex. Strong biting at tip. Gonys slightly convex. Strong probing. Bill relatively slender. Platyspiza Culmen strongly convex Very strong crushing along Gonys straight. entire length. Bill short, broad, and deep. Very strong biting at tip. Pinaroloxias Culmen curved. Weak probing. Gonys straight. Grasping tip. Bill slender, sharp-pointed, decurved. Certhidea Culmen straight proximally, slightly Grasping tip. curved distally. ** Strong probing. Gonys straight or slightly convex. Bill small, slender, acute. * After Swarth, 1931. See fig. 10. ** After Ridgway, 1896 [1897} population of G. fortis, the same relative shape of bill prevails, as evidenced by the paralleling of the culmen and gonys in the three sizes of bill. Within Camarhynchus and Cactospiza, species and individual differen- ces in the bill involve the absolute size, depth-to-length ratio, and, more prominently, the curvature of the culmen and gonys. (See fig. 2, and compare M, N, and Q for Camarhynchus psittacula, and O and S ioT Cactospiza pallida. ) b. Structural pattern in the jaw musculature. The potentialities of the bills are realized only when the upper and lower mandibles are set in mo- tion by the jaw muscles. The relative size and position of the muscles affect the action of the bill. Without entering into a discussion of the kinetics of the avian bill, or a detailed description of the individual muscles (see Bow- man, 1961, for details), suffice it to say that within the Geospizinae the great est variation in the jaw musculature concerns the size of certain of the "ad- ductor" groups; namely, M. adductor mandihulae extemus and Mm. pterygoi- deus dorsalis et ventralis (see figs. 12 and 13). Figure 8. Vegetation of the Miconia Belt, eight miles north of Academy Bay. (Photo cour- tesy R. Freund.) 120 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers vfvr-r^-M-^^ ~ tiWi ; >^ ■ No. 44) BOWMAN: GALAPAGOS SYMPOSIUM 121 GEOSPIZA HEAVY DUTY LINESMAN'S PLIERS PLATYSPIZA PARROT-HEAD GRIPPING PLIERS CAMARHYNCHUS HIGH LEVERAGE DIAGONAL PLIERS PINAROLOXIAS CURVED NEEDLE NOSE PLIERS CACTOSPIZA LONG CHAIN NOSE PLIERS CERTHIDEA NEEDLE NOSE PLIERS Figure 10. Comparison of shapes of bill in six genera of Geospizinae with shapes of six kinds of pliers (cf. table 2). To illustrate the nature of the differences, let us consider three sib- ling species of Geospiza on Indefatigable Island. The large lateral and ven- tral muscle complexes (nos. 1, 2, 3 and 5, 6, 7, 8, respectively, in figs. 12 and 13) become disproportionately larger as we proceed from the small- to the medium- to the large-billed species (i.e., G. fuliginosa, G. fortis, G. magni- Tostris). In parallel fashion we observe a disproportionate increase in size of these muscles as we proceed from the small- to the large-billed species of Camarhynchus (fig. 13). c. The pattern of feeding. On the basis of their diets, the 13 species of Galapagos finch may be grouped into four categories as follows: 1. Almost exclusively herbivorous: Platyspiza crassirostris 2. Chiefly gramnivorous (with some insects): 6 species of Geospiza 3. Chiefly insectivorous (with some seeds): 3 species of Camarhyn- chus and 2 species of Cactospiza. 4. Almost exclusively insectivorous: Certhidea olivacea. Thus there are six different kinds of seed-eating niches occupied by six species of Camarhynchus , Cactospiza, and Certhidea, in addition to the herbivore niche occupied by Platyspiza crassirostris. Figure 9. Three important predators on Galapagos finches. a. Buteo galapagoensis; b. Asio galapagoensis; c. Dromicus biserialis. 122 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers large magnirostris small magnirostris large fort is culm en medium fort is fuliginosa small for lis scan dens gonys Figure 11. Bill profiles of four species of Geospiza from Indefatigable Island. A more detailed analysis of the diets and the foraging locations provides us with important points of difference in the feeding niches. The nine resi- dent species of finch on Indefatigable Island have been most thoroughly stud- ied in this regard and data on these are presented in table 3 and figures 14, 15 and 16. d. Explanation of patterns described. Species of Geospiza represent several lines of evolution toward the solution of the problem of efficient ex- ploitation of the seed resources of the Galapagos environment. The great ar- ray of sparrow-like crushing bills is a clear reflection of the abundance of No. 44) BOWMAN: GALAPAGOS SYMPOSIUM 123 Geospiza magnirostris Geospiza fortis Geospiza fuliginosa Figure 12. Jaw muscles of Geospiza magnirostris, G. fortis, and G. fuliginosa: above, lateral view of superficial muscles; below, ventral view of superficiar(left half) and deep (right half) muscles. 1. M. adductor mandibulae externus superficialis; 2. M. adductor mandibulae ex- ternus medialis; 3. M. adductor mandibulae externus profundus; 4. M. adductor mandibulae pos- terior; 5. M. pterygoideus dorsalis lateralis; 6. M. pterygoideus dorsalis medialis; 7. M. ptery- goideus ventralis lateralis; 8, M. pterygoideus ventralis medialis; 9, M. depressor mandibulae; 10, M. pseudotemporalis profundus; 11, M. pseudotemporalis superficialis; 12, M. retractor pal- atini; 13, Lig. jugomand. art.; 14, Lig. jugomand. ext.; 15, Proc. palato-max.; 16, Proc. transpal.; 17, pterygoid; 18, rhampbotheca. seeds differing in their size, hardness, and location. A characteristic adapta- tion of angiosperms in arid and semi-arid regions is a drought-resistant seed (Stebbins, 1952), which condition is well developed in Galapagos plants (Hook- er, 1847, pp. 256-257). The adaptive trends in the Geospiza series proceeding from G. fuligi- nosa to G. fortis to G. magnirostris are as follows: Trend 1. more terrestrial in foraging habits 2. consumption of larger seeds 3. consumption of harder seeds Data presented in fig. 14 fig. 15 fig. 15 124 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Trend Data presented in 4. consumption of fewer insects fig. 14 5. bill larger and thicker-based fig. 10 6. "adductor" muscles relatively larger fig. 12 Three species of Geospiza; namely, G. difficilis, G. scandens, and G. conirostns, differ from the foregoing species in having elongated bills; that is, bills which are longer relative to their basal depth (see fig. 2, I,U,C, respec- tively). The elongate bill permits the bird to seize food that might be more difficult to obtain were the bill attenuate, because interference of the eyes with the substrate is thus substantially diminished. For example, G. scandens is prone to probe the thick pear-like fruit of Opuntia cactus for moderately hard seeds, and to insert the bill tip into the spine clusters to procure sugary secretions at the extra-floral nectaries. In G. conirostric and G. difficilis the elongated bill permits the birds to reach seeds and insects in crevices and beneath leaf litter, with a minimum of interference to the eyes. Platyspiza crassirostris Camarhynchus psittacula Camarhynchus parvulus Figure 13. Jaw muscles of Platyspiza crassirostris, Camarhynchus psittacula, and Cam- arhynchus parvulus; above, lateral view of superficial muscles; below, ventral view of super- ficial muscles; below, ventral view of superficial (left half) and deep (right half) muscles. Num- bers as in figure 12. No. 44) BOWMAN: GALAPAGOS SYMPOSIUM 125 S I S I S I FEEDING STATION C porvuSus S, 60 S40 2 30 uj20 = 10 ®®® '^ ft .-^ _^ 1%^ ii Isi S I S I N I FEEDING STATION Plafyspiza g 60 f^50 ® ® © IV 40 ^30 s 10 1- n Cactospiza (?. magnirostns (? /'or/zs G fuhginosa u 60 ^50 ® ® ® ^40^ 2 30 ■:■ ? 10 :•; 1- u V 4 %m I _i>:i C psittacula r« 1 ■E-i';- S I S I S L FEEDING STATION (^ ground (B) trunks, large branches fC) leaves , twigs 1® i S seeds I insects fl nectar L leaves, flowers. seeds Figure 14. The principal feeding stations and the extent of their occupancy by nine spe- cies of Geospizinae on Indefatigable Island. The large-, medium-, and small-billed individuals of Geospiza fortis on Indefatigable Island (fig. 11) are known to differ in their diets, with the larger forms taking progressively harder seeds than the smaller forms (see Bowman, 1961, p. 60. In Camarhynchus, Cactospiza, and Certhidea, the "problem" of extract- ing insects, which are concealed beneath woody tissues during the daylight hours to escape from predators and dessication, has been "solved" through the evolution of powerful tip-biting bills (Camarhynchus), probing bills (Cacto- spiza), and forceps-like bills (Certhidea). The adaptive trends in the Camarhynchus series from C. parvulus to C. psittacula on Indefatigable Island are as follows: Trend Data presented in 1. more arboreal in foraging habits fig. 14 2. consumption of larger insects larvae fig. 16 3. consumption of fewer seeds fig. 14 126 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Trend Data presented in 4. bill larger and thicker fig. 13 5. "adductor" muscles relatively larger fig. 13 If we compare all the "insectivorous" finches on Indefatigable Island in the sequence Certhidea olivacea--Camarhynchus parvulus—Camarhynchus psittacula--Cactospiza pallida, we find that the smaller-billed species feed on smaller insects (larvae) than do the larger-billed species (fig. 16.) Concom- itantly, Certhidea ioTSiges typically on the leaves and terminal twigs of bushes and trees, whereas the larger-billed and smaller-billed species of Camarhyn- c>?>ws forage typically on the larger and smaller branches of trees, respectively, with Cactospiza showing a distinct preference for large branches and trunks in its search for insect booty (see fig. 14). Geospiza fuliginosa Geospiza scandens Geospiza forfis Geospiza magnirostris A B C D E F G H I SIZE OF SEED Geospiza fuliginosa A B C D E SEED HARDNESS Geospizo scandens Geospiza fortis Geospiza mognirostris Figure 15. Analysis of the diets of four species of Geospiza from Indefatigable Island on the basis of size (above) and hardness (below) of seeds consumed. e. Structural pattern in the digestive tract. Invertebrates, general- ly, it is well known that herbivorous species possess relatively longer intest- ines than do carnivorous and nectivorous species— a condition correlated with the greater east of digestion and assimilation of proteins and simple sugars as compared to starches and celluloses. A trend of this kind may be demon- strated in species of Galapagos finches on Indefatigable Island (fig. 17) by comparing the relative lengths of the intestine (i.e., length of intestine over cube root of body weight) with the relative amounts of cellulose-rich food in the diets. No. 44) BOWMAN: GALAPAGOS SYMPOSIUM 127 Certhtdea olrvacea A B C 0 SIZE OF LARVA Camorhynchus porvulus Comarhynchus psittaculo Cactospizo pallido Figure 16. Analysis of the diets of four species of insectivorous finches from Indefati- gable Island on the basis of size of larvae consumed. Whereas the gradual shortening of the intestine in the morphological series Geospiza magnirostris--G. fortis—G. fuliginosa is correlated with a re- duction in starch- and cellulose-rich seeds and an increase in protein-rich in- sects (fig. 14), in G. scandens the still shorter intestine is correlated with a seed diet supplemented with sugar-rich nectar and comparatively more insects. The intestine of Platyspiza is conspicuously longer than that of Geospiza magnirostris owing to the fact that the former species consumes, in addition to seeds, large quantities of buds, leaves, and flowers, which items are taken much less frequently by the latter species. Platyspiza rarely feeds on insects. In the insectivorous series of finches; namely, Camarhynchus parvulus— C. psittacula--Cactospiza pallida-- Certhidea olivacea, there is a regular re- duction in the relative amounts of plant food in the diets, tending toward com- plete insectivorousness in Certhidea. Correlated with this shift in diet is the gradual reduction in length of intestine (see also fig. 14). f. Foraging activity and heart size. Recent studies by Hartman (1954, 1955) and Norris and Williamson (1955) have shown, among other things, that species differences in heart weight may be a reflection of certain inherent physiological adaptations, and particularly as this relates to activity. It has been assumed that in any one species a particular heart size is an adaptation of the circulatory system to the sum total of the various physiological condi- tions resulting from physical exertion, heat production, and environmental in- fluences. Since the major differences between species of Galapagos finches concern adaptations for food-getting, and since food-getting constitutes a major part of the daytime activity of the birds, some correlation between heart size and type of foraging activity might be expected. Platyspiza crassirostris, the largest (heaviest) finch, has the smallest heart ratio (table 4). Correlated with this is the fact that Platyspiza is most commonly seen sitting quietly in bushes or trees feeding on berries, leaf buds, or flowers. Platyspiza show less activity in food-getting than does any other species of finch. 128 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Table 3- Food habits of the Geospizinae, P^CJOD HABITS OF GEMS FOOD HABITS OF SPFCIFS SPECIES UITH SIMILAR BILL Geospiza: mainly seeds, and magnirostris; small variety of Coccothraustes coccothraustes occasionally exposed insects. hard seeds ( Fringillidae) fortis: large variety of moder- Melanospiza richardsoni ately hard seeds (Fringillidae) fuliginosa: large variety of Passerina cyanea soft seeds (Fringillidae) scandens: small variety of Tangavius aeneus moderately hard seeds; pref- (Icteridae) erences for fruits and nectar of Opuntia cactus difficilist poorly known but Lophosphingus griseo-cristatus presumably similar to fuli- (Fringillidae) ginosa, but possibly includ- ing more insects conirostris: poorly known but Saltator albicollis presumably soft to hard (Fringillidae) seeds and soft tissues of Opuntia cactus Camarhynckius: mainly con- psiltacula: moderate variety of Psittiparus gularis cealed insects excavated largish insects; fewsoft seeds ( Paradoxornithidae) from woody tissues, and occasionally seeds pauper: poorly known, but pre- Similar to above sumably intermediate between psiltacula and parvulus parvulus: large variety of Parus inornata smallish insects and moder- (Paridae) ate amount oi soft seeds Cactospiza: mainly conceal- pallida: small variety of larg- Tachyphonus coronatus ed insects removed by furrowing with bill and by ish insects; soft fruits (Traupidae) means of a "tool" held in heliobates: poorly known but Similar to above bill {pallida only); occa- presumably mainly insects sionally soft fruits and from mangroves seeds Platyspiza: buds, leaves, crassirostris: same as for Structural equivalent not known flowers, fleshy fruits, soft genus to hard seeds Certhidea; insects exclus- olivacea; same as for genus Basileuterus belli ively, and mainly small (Parulidae) exposed forms Pinaroloxiasr presumably inornata: same as for genus Coereba flaveola insects, nectar, and ("Coerebidae") some fruit No. 44) BOWMAN: GALAPAGOS SYMPOSIUM 129 LjJ CO LU U. O X I- z Ld LlI > UJ 1 — 1 1 n 1 r- 1 1 1 0 80 - 70 ■ • no"* ,x- 60 • • i^ 0^ .^^ 50 ■ 40 • 7 C5- 0)- 30 • 1— — 1 1 H 1 1- 1 — 1 1 10 20 30 40 50 60 70 80 90 PERCENTAGE OF CELLULOSE-RICH FOOD IN DIET Figure 17. Relationship between intestinal length and diet in nine species of Geospizinae from Indefatigable Island. Camarhynchus psittacula and C. parvulus have very similar methods of foraging, but there are some fairly obvious differences. In general, C parvulus forages on the small branches and terminal twigs of trees and bushes, where its behavior resembles that of certain parids. The larger C. psittacula tends to forage more on the larger branches of trees, where its twisting actions with the bill appear to be more vigorous than similar actions in the smaller C par- vulus, and undoubtedly of greater absolute strength. In other words, the field observations do not indicate any important differences in intensity of foraging activity, but do demonstrate certain significant differences in habitat selec- tion. The difference in heart ratio between C parvulus (.664) and C. psitta- cula (.632) is prohahly the result oi the diiierences in general body size (fig. 18). Certhidea olivacea is not only the smallest but also the most active of the Galapagos finches. Its relatively large heart (largest heart ratio) is surely a reflection of this high level of activity. Its erratic aerial flights, frequent wing flitting, and constant changing of position while foraging in the foliage, are all manifestations of its relatively greater metabolic rate. Certhidea is al- most the perfect antithesis of Platyspiza with regard to activity and relative heart size. 130 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers 8 9 10 15 20 25 30 35 BODY WEIGHT IN GRAMS Figure 18. Average heart weight plotted against average body weight for nine species of Galapagos finches from Indefatigable Island (Logarithmic scale.) Cactospiza pallida and Geospiza scandens have the same heart ratio (.690), which value is second largest in the group (table 4). Not only do these two species have the same body size and superficially similar bills (fig. 1), but also both may forage on the Opuntia cactus. Because of these similarities, both were considered originally to be members of the same genus ("Cactomis"). The practice of foraging on the trunks of trees (or cactus) requires an excep- tionally high expenditure of energy, merely in maintaining ahold, let alone in climbing about or in excavating with the bill. This might account for the rela- tively large and identical heart ratio for these species. The three species of "ground-finch," Geospiza magnirostris, G. fortis, and G. fuliginosa, have heart ratios of .674, .647, and .594, respectively. This series constitutes a major exception to the general principle that smaller birds have relatively larger hearts than do larger birds (Hartman, 1955, p. 223). In No. 44) BOWMAN: GALAPAGOS SYMPOSIUM 131 Table 4. Heart ratio for nine species of Galapagos finches from Indefatigable Island, SPECIES HEART RATIO* SPECIES HEART RATIO* Geospiza magnirostris fort is fuliginosa .... scandens Cactospiza pallida ... .674 .647 .594 .690 .690 Cer thidea olivacea.. Camarhynchus parvulus .... psittacula .. Platyspiza crassirosiris .697 .664 .632 .538 * Mean heart weight expressed as a per cent of mean body weight attempting to account for the differences, it should first be pointed out that all three species spend about 50 per cent of their time foraging at ground level where scratching with the feet in search of seeds demands the greatest phys- ical exertion of any foraging activity typical of these three species. Field ob- servations indicate that G. magnirostris is somewhat more wary than the two smaller species, and head movements are much slower and more deliberate in G. magnirostris. The latter face prompted me to examine the relative size and weight of the head in these three species. Whereas in all finches other than G. magnirostris, G. fortis, and G. fuliginosa, the skull and lower jaw make up 23 to 34 percent of the total skeletal weight, in the three species of Geospiza under consideration the skull and lower jaw comprise 46.8, 38.9, and 27.8 per cent, respectively, of the total skeletal weight. In view of these rather striking differences in head size, it would seem to be reasonable to imagine that the differences in heart ratio between the lar- ger and smaller species oiGeospiza(G. scawfiews excepted) might be correlated, at least in part, with the relatively greater energy demands for the function- ing of the very large jaw musculature, and also for maintaining the posture of a much heavier head (see fig. 12). The differences in the heart ratios between Geospiza magnirostris, G. fortis, and G. fuliginosa are attributed mainly to differences in the relative size differences in the skull and jaw musculature. Observations on the flight of species of Geospiza in the wild and in cap- tivity indicate that G. scandens has the strongest and most rapid flight as well as the greatest agility, whereas G. fuliginosa has the weakest type of flight. It is probably not fortuitous that, of the four species of Geospiza on Indefatigable Island, G. scandens has the largest and G. fuliginosa the small- est heart ratio. 132 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers 100 111 o -^ INSECTIVOROUS ABOVE GROUND FORAGING !'■ ■ i I y ■'■'*■ I ■ ■ ■ ••■■■■■■•■■•■ lasaaaaflaaaaB aaaaaaaaaaaaa aaaaaaaaaaaaa -----^aaaaaea - - - - a a a :•:•:■:• ground level foraging :•:•:•:■:■:■:•: • a ■ ■ ■ ■ ■ ari ■ ■••••■- ■ ■ a a a ■ I I a • a I a a a a I a a a L, a ■ ■ aji • a a a a a ■ a a ■ a a a r a ■ a a a a a • ■ a a a L a • a a a a ■ a a a a a a ~ ■ ■ ■ ■ iti'n i'l ■■■ i laaaaaaaaaaaa aaaaaaaaaaaa laaaaaaaaaaaa GEOSPIZA PLATYSPIZA CAMARHYNCHUS CACTOSPIZA CERTHIDEA < LU < 2 V%l^t^T ® ® ! ® © ® ® Figure 19. Correlation of male plumage condition, foraging level, and diet in five genera of Geospizinae. g. Adaptive variation in plumage and bill. (1) Plumage coloration. Most previous workers on the finches have claimed that the plumage colorations are of little or no adaptive value. I do not believe that the available evidence supports this view. Because of the dietary differences between the finch species, it is not surprising that we should find differences in foraging level (fig. 14). For ex- ample, the seed-eaters of the genus Geospiza spend about 50 percent of their time on the ground in search of seeds. In the arid coastal zone, where these species are most common, there is a preponderance of dark colored lava. Cor- related with these conditions of behavior and environment are the adult male plumages that may be fully black, partially black, or non-black and essential- ly like that of the females (fig. 19). The insect-eaters of the genus Camarhynchus, as well as the fruit-and- bud eaters of the genus Platyspiza, find their food chiefly on the branches and No. 44) BOWMAN: GALAPAGOS SYMPOSIUM 133 among the foliage of trees and shrubs where the dominant colors are green and grey. Occasionally the birds will forage for seeds on the ground. Thus, these birds, Camarhynchus and Platyspiza are brought into contact with backgrounds of strong color contrast, namely, grey-green and black. The plumages of the adult males of Camarhynchus and Platyspiza are usually black over head and breast, black on the head only, or completely grey brown like the females (fig. 19). Those individuals with black on the anterior parts of the body are surprisingly difficult to see either on the ground or in the trees, because of the visually disruptive effect produced by this kind of marking. The remaining species of finches of the genera Cactospiza and Certhi- dea rarely forage at ground level. Their plumages are various shades of olive and grey, but never black, thus matching the dominant hues in the trees and shrubs where they forage. The selective force in the case of plumage coloration is predation by hawks, owls, and snakes (fig. 9). There are numerous examples of geographic variation in plumage color- ation in the Galapagos finches. The most striking of these is to be found in the genus Geospiza. In this group there may be a succession of plumages in the male beginning with no black, followed by increasing amounts of black over head and breast, to complete blackness (see fig. 19). This gradual de- velopment is at least partly associated with age. At each annual molt an in- creasing number of black feathers may appear, but in addition there are some plumage types that are genetically fixed. Some birds may appear to acquire a partially black plumage precociously, and retain this stage in subsequent years. Other males never seem to acquire any black plumage whatsoever. Further- more, the relative frequency of the various plumage types in adult males seems to be somewhat different from island to island (Lack, 1945; Swarth, 1931), al- though an intensive field study of this situation is still wanting. Because there are differences in the dominant background color between islands, we may assume that each plumage type has its own selective advantage in cer- tain kinds of environments, and that the specific frequency of occurrence of each type may change from season to season, depending upon the age compo- sition and sex ratio of the population, as well as upon the local feeding hab- its of the birds. In other words, the non-black, the partially black, and the fully black male plumages adapt the population as a whole to the extremes of background coloration. The need for such camouflage is greatest at the end of the dry season in the arid coastal zone. (2) Bill structure. It seems clear that a species living on two islands, differing in their food supply, is going to show geographic variation in its feeding harits. And such is known to be the case on Galapagos. Let me illus- trate this point with one example. The bill of the grosbeak-finch (Geospiza magnirostris) from Tower Island is almost the largest for the species; on In- 134 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers defatigable Island the bill is of somewhat smaller size. An examination of the skulls from both islands shows that there are differences in the overall rug- gedness, in the size of processes for muscle attachment, and in relationships of bones of the skull (fig. 20). The functional significance of these inter-is- land differences in morphology may be summarized as follows: the finches on Tower Island are capable of more powerful adduction than the finches on In- defatigable Island. Indeed, on Tower Island relatively more of the available seeds are large in size and hard shelled, than are those on Indefatigable Is- land, to judge from the known facts about the vegetation on the two islands. The anatomical differences in the head region between these two island pop- ulations of Geospiza magnirostris are of the same general character as those Tower Island Indefatigable Island Figure 20. Comparison of the skulls of Geospiza magnirostris from Tower and Indefatig- able islands in lateral profile (left) and posterior profile (right). differences between large-, medium-, and small-billed individuals oi Geospiza fortis (fig. 11), about whose functional significance there is no doubt. Addi- tional examples are discussed elsewhere (Bowman, 1961). Adaptive Radiation in Darwin's Finches Within this one insular sub-familial group of songbirds, the Geospizinae, we find biological equivalents of no less than seven continental familial groups, (see fig. 1). This pattern of adaptive radiation has been shaped largely by the nature of the Galapagos environment and by the genetic constitution of the birds. No. 44) BOWMAN: GALAPAGOS SYMPOSIUM 135 Origin of the Geospizinae. Although Galapagos probably has been in existence since Tertiary times (Shumway, 1954), we have no notion as to the exact time of entry by the ancestors of the finches, or of any other terrestrial group. There seems to be little doubt that the Galapagos (and Cocos) islands are truly "oceanic" in origin inasmuch as they are composed almost entirely of basaltic lava, and the deeps between the archipelago and the mainland are so great as to preclude a former continuous land connection (see Shumway, 1954). The general inaccessibility (or unsuitability) of the islands to terres- trial colonists from the adjacent mainland of South America is indicated by the seemingly random make-up of the biota, especially well seen in the fam- ilies of vascular plants and insects (Hooker, 1847; and Van Dyke, 1953). In view of the remarkable uniformity in the internal anatomy, plumage, song, nest, and egg in the 14 species of Geospizinae, it is reasonably certain that Darwin's finches constitute a monophyletic group of birds. But the pre- cise nature of the ancestral type is not quite so obvious as some writers would seem to think (see Simpson, et al., 1957, p. 446). Avian systematists are of the opinion that several New World families of songbirds are phylogenetically closely related, including the sparrows (Fringillidae), thetroupials (Icteridae), the tanagers (Thraupidae), the warblers (Parulidae), and the honeycreepers (Coerebidae). I think it is significant, therefore, that we find among Darwin's finches ecological and morphological counterparts of these mainland families (see fig. 1). And in addition, we find equivalents of two other mainland fami- lies, the parrotbills (Paradoxornithidae) and the titmice (Paridae), which are not generally considered to be closely allied to the previously mentioned groups. Since only five of the 14 species of Geospizinae are clearly identifiable as "finch" types, I see no reason to assume a priori that a "finch" origin of the group is any more likely than a "warbler" origin, etc. (see table 3). Also, we are not obliged to assume that the geospizine ancestors came from the ad- jacent coast of South America (Ecuador) simply because that region is the most proximate. To the evidence marshalled by Swarth (1934), showing a close af- finity between elements of the Galapagos and Caribbean avifaunae, may be added the example of Geospiza fortis, which shows a striking resemblance to Melanospiza richardsoni of St. Lucia Island, Lesser Antilles. The Galapagos may have been no less remote in effect to colonists from the Caribbean than from coastal Ecuador when one considers the possibility of chance rafting in mid-Tertiary times by means of an ocean current system flowing through breach- es in the Isthmus of Panama (Swarth, 1934; Vinton, 1951). In summary it may be said that we have no precise notion about the na- ture of the ancestors of the Geospizinae; nor do we know the place of origin of the colonial stock other than that it was from America. 136 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Conditions on the Galapagos Islands in Colonial times. We can only speculate on the environmental conditions as they existed when the ancestors of Darwin's finches arrived in the Galapagos Archipelago. It would appear to this writer, however, that there is little justification for the unilateral assump- tion that once the ancestral geospizines gained access to the islands, they entered into a land of remarkable ecological opportunity (Simpson, etal., 1957, p. 445), of abundant foods and varied living quarters, unmarred by the presence of competitive neighbors, and with complete freedom from enemies (Lack, 1947, p. 114). It is just as likely, I believe, that ecological opportunity was very limited when the ancestors of Darwin's finches first arrived in their newly found environment, and that they evolved together with the floral and other faunal elements of Galapagos, including the predators. The various islands appear to be of different geological age (Chubb, 1933), which means that the number of islands, their size and height above sea level, probably were different at various times in the past. Present-day differences in the flora and fauna of the islands— islands in some cases sep- arated by only a few miles of ocean— are, indeed, striking (see table 1), and surely are a reflection of differences which existed to a greater or lesser ex- tent in the past. Constitutional make-up of the geospizine ancestors. Whether the or- iginal colonization of Galapagos included many or few individuals, or if there was more than one invasion by the ancestral type, we shall never know. Lack of information on these and other matters makes it difficult to assess the ef- fect of random genetic drift in the evolution of the Geospizinae. To be sure, the degree of heterozygosity in the genetic environment of the founding fore- bears was determined by the number of the invading colonists and their indi- vidual hereditary constitution. If we assume that the founders were few in number and derived entirely from a genetically depauperate "peripheral" main- land population, then one may envision some form of disharmony in the genetic environment of the colonists (Mayr, 1954). The effects of a "genetic bottle- neck" may have been manifest at several times subsequent to the initial col- onization. For example, interisland invasions, volcanic eruptions, and period- ic droughts are some of the possible causes of major population reductions. The most favorable structure for rapid evolution is that of a large or medium sized population divided into many small sub-units or colonies which are largely isolated from each other, but can interchange genes through occa- sional migration between them (Wright, 1940). Such a population permits new gene combinations to become established in the individual sub-units both through natural selection ("anti-chance" factor) and through random genetic drift ("chance factor"), without the swamping effect which occurs in large populations. At the same time, migration between colonies prevents their stag- nation, and allows the population as a whole to draw upon a large supply of genes (Stebbins, 1952, p. 35). No. 44) BOWMAN: GALAPAGOS SYMPOSIUM 137 One might ask, is it possible to distinguish the results of natural se- lection from those of random genetic drift? What criteria are we to use? In order for two small insular populations to drift apart genetically, there must be identical, or very similar, environmental conditions on the two islands. At the present time we know that selection pressures on the finches differ in kind and degree from island to island as a consequence of differences in food sup- ply (seeds, insects) and predation pressure (hawks, owls, snakes), and it is reasonable to think that such differences have existed to a greater or lesser extent in the past. All the structural and behavioral features of the Geospizinae studied up to now have an adaptive explanation. Even certain minor inter-populational differences in plumage coloration and bill dimensions can reasonably be ex- plained by selection; there is no need to invoke chance factors. This is not to say that random genetic drift did not play some role in the evolution of the Geospizinae through interaction with selection, but rather, that convincing evidence has not yet been marshalled in support of it. The constitutional limitations of the founders of Darwin's finches are suggested by the apparent gaps in the picture of adaptive radiation. For ex- ample, genetic factors (other than "Sewall Wright effect") might explain the absence of "lark-finches" and "shrike-finches" in the geospizines, since alaudids and laniids are not closely related to the "parulid-thraupid-icterid- coerebid-fringillid" complex, to which the geospizines seem to be allied. It is likely that the lark and shrike niches are available on certain of the larger and higher islands of Galapagos. It might be suggested that the absence of a specialized nectar feeder ("honeycreeper-finch") on the Galapagos is due to the genetic inability of the Geospizinae to evolve in that direction. But this appears not to be the case. Rather, on Galapagos there is no continuous and plentiful supply of nectar to support this kind of feeder, in contrast to condi- tions on Cocos Island where a "honeycreeper-finch" (Pinaroloxias inornata) has evolved in a lush tropical rain forest. We should hardly expect the geospizines to have given rise to "mock- ingbird-finches", "swallow-finches", "flycatcher-finches", and "cuckoo- finches" since had mockingbirds, swallows, etc. , been neighbors of the newly evolving geospizines, these niches would have been occupied and unavail- able for exploitation by the geospizines. Also, it should be noted that these mainland types are, presumably, remotely related to the ancestral geospizine stock, and therefore we may suppose that even if the mainland mockingbirds, swallows, etc. were absent from Galapagos, geospizine counterparts might not have evolved. With unlimited ecological opportunity, time and the consti- tutional make-up of the geospizine ancestors are the over-riding restrictions shaping the evolutionary destiny of the finches. The pre-adaptational potential of the ancestral geospizines was consid- erable, to judge from the results of adaptive radiation that we see today (fig. 138 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers 1). This genetic potential was released in the insular environment of Galapa- gos where rather sudden changes in the feeding mechanism and associated be- havior were necessitated by frequent shifts in the conditions of the environ- ment resulting from repeated inter-island invasions. One character, possibly of a pre-adaptive nature, occurring in all geospizines and also in several close- ly related mainland groups, is the presence of partially pneumatized window areas in the cranial vault. These stress-resistant regions may have permitted more rapid readjustments in the head than would otherwise have been possible, as new feeding opportunities became available (Bowman, 1961, p. 261). The origin of new species of Galapagos finches. When the ancestral geospizines arrived in Galapagos, they were confronted with an environment different in most of its ecological aspects (food resources, competitive neigh- bors, and the like) from the one just vacated. Following the original estab- lishment of a colony on one of the islands, the birds soon moved about the ar- chipelago establishing footholds on other islands. During such periods of population dispersal, the birds were confronted with somewhat different food conditions (among other things), and being more or less isolated by ocean from other populations, island ecotypes soon evolved. Ultimately, selection piled up sufficient genetic differences so that when the incipient species on island "A" (where it had evolved) moved to island "B," it remained reproductive ly isolated from all other closely related birds present, and thereafter behaved as a full species. In this manner the various species of Geospizinae were probably evolved. Island races of the same species are a reflection of subtle inter-island differences in feeding niches. For example, the Geospizamagnirostris "format" might have been evolved on an island where large hard seeds were the primary food available. Once this basic grosbeak bill appeared, its size could be further modified by selection to better accomodate the slightly different-sized hard seeds prevailing on other islands, thus effecting the most efficient bill- mechanism commensurate with the available food resources. Conclusion The patterns of evolution discussed above are largely concerned with biological adjustments in Darwin's finches to their food plants and animals (such as powerful bills for crushing hard seeds and cutting into tough woody tissues)! to predators (such as plumage coloration for concealment), and to the physical environment (such as relatively small hearts in response to warm and constant air temperatures). Each pattern involved not only the whole or- ganism, but the organism plus its environment. It is the totality of these func- tional systems that has been subjected to selection. No. 44) BOWMAN: GALAPAGOS SYMPOSIUM 139 Literature Cited Alpert, L. 1946. Notes on the weather and climate of Seymour Island, Galapagos Archi- pelago. Bulletin of the American Meteorological Society, vol. 27, pp. 200-209. BOWMAN, R. I. 1961. Morphological differentiation and adaptation in the Galapagos finches. University of California Publications in Zoology, vol.58, pp. 1-VII, 1-302. Chubb, L. J. 1933. Geology of Galapagos, Cocos, and Easter islands. Bernice P. Bishop Museum Bulletin, p. 110. DARWIN, C. 1845. Journal of researches into the natural history and geology of the coun- tries visited during the voyage of H.M.S. Beagle round the world, under the command of Capt. FitzRoy, R. N., 2d rev. ed.; London; John Murray. GIFFORD, E. W. 1919. Field notes on the land birds of the Galapagos Islands and of Cocos Island, Costa Rica. Proceedings of the California Academy of Sci- ences, ser. 4, vol. 21, pp. 189-258. Hartman, F. a. 1954. Cardiac and pectoral muscles of trochilids. Auk, vol. 72, pp. 267-469. 1955. Heart weights in birds. Condor, vol. 57, pp. 221-238. Hooker, J. D. 1847. On the vegetation of the Galapagos Archipelago, as compared with that of some other tropical islands and of the continent of America. Transactions of the Linnaean Society, vol. 20, pp. 235-262. Howell, J. T. 1942. Up under the equator. Sierra Club Bulletin, vol. 27, pp. 79-82. Lack, D. 1945. The Galapagos finches (Geo spizinae ): A study in variation. Occasional Papers of the California Academy of Sciences,, no. 21, pp. 1-159. 1947. Darwin's finches. Cambridge University Press. 208 pp. Lowe, P. R. 1930. Hybridisation in birds in its possible relation to the evolution of the species. Bulletin of the British Ornithological Club, vol. 50, pp. 22-29. MAYR, E. 1954. Change of genetic environment and evolution. In, Evolution as a process, pp. 157-180; Editors: J. Huxley, A. C. Hardy, E. B. Ford; London: George Allen and Unwin. NORRis, R. A., AND F.S.L. Williamson 1955. Variation and relative heart size of certain passerines with increase in altitude. Wilson Bulletin, vol. 67, pp. 78-83. RIDGWAY, R. 1896 [T897]. Birds of the Galapagos Archipelago. Proceedings of the U.nited States National Museum, vol. 19, pp. 459-670. 140 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Rothschild, W., and E. Hartert 1899. A review of the ornithology of the Galapagos Islands. With notes on the Webster-H arris expedition. Novites Zoologicae, vol. 6, pp. 85-205. Shumway, G. 1954. Carnegie Ridge and Cocos Ridge in the east equatorial Pacific. Jour- nal of Geology, vol. 62, pp. 573-586. Simpson, G. G., C S. Pittendrigh, and L. H. Tiffany 1957. Life: an introduction to biology. New York: Harcourt, Brace and Co. 845 pp. Snodgrass, R. E., and E. Heller 1904. Papers from the Hopkins-Stanford Galapagos expedition, 1898-1899. XVI. Birds. Proceedings of the Washington Academy of Sciences, vol. 5, pp. 231-372. Stebbins, G. L. 1952. Aridity as a stimulus to plant evolution. American Naturalist, vol. 86, pp. 33-44. Stewart, A. 1911. A botanical survey of the Galapagos Islands, Proceedings of the Cali- fornia Academy of Sciences, ser. 4, vol. 1, pp. 7-288. 1915. Some observations concerning the botanical conditions on the Galapa- gos Islands. Transactions of the Wisconsin Academy of Sciences, Arts and Letters, vol. 18, pp. 272-340. Stresemann, E. 1936. Z ur Frage der Artibildung in der Gattung Geospiza. Orgaan der Club van Nederlandische Vogelkundigan, vol. 9, pp. 13-21. SVENSON, H. K. 1946. Vegetation of the coast of Ecuador and Peru and its relation to the Ga- lap a-gos Archipelago. American Journal of Botany, vol.33, pp. 394- 426. Swarth, H. S. 1931. The avifauna of the Galapagos Islands. Occasional Papers of the Cali- fornia Academy of Sciences, no. 18, pp. 1-299. 1934. The bird fauna of the Galapagos Islands in relation to species forma- tion. Biological Reviews, vol. 9, pp. 213-234. VAN Dyke, E.G. 1953. The Coleoptera of the Galapagos Islands. Occasional Papers of the California Academy of Sciences, no. 22, pp. 1-181. Vinton, K. W. 1951. Origin of life on the Galapagos Islands. American Journal of Science, vol. 249, pp. 356-376. White, M. J. D. 1948. Animal cytology and evolution. Cambridge University Press. Wright, S. 1940. Breeding structures of populations in relation to speciation. American Naturalist, vol. 74, pp. 232-248. PROTECTION AND CONSERVATION PROBLEMS ON THE GALAPAGOS ISLANDS* Misael Acosta-Solfs Instituto Ecuatoriano de Ciencias Naturales Quito, Ecuador The Galapagos Archipelago is a jewel of nature which, after Darwin's studies in 1859, has attracted the attention of scientists the world over. In 1934 several sections of Galapagos were declared an official reserve. At that time, the Government of Ecuador authorized a survey by the engineer Frederick Paez. He was encouraged by a group of professors at the Ecuador- ian Central University, and by many other persons fond of nature, including the learned professor Jonah Guerrero. Unfortunately, the executive decree of 1934 did not have the expected results because there was no means for its enforcement. The scientific inter- est of the world bloomed again in 1935 because of the centenary of Darwin's visit to Galapagos. In 1937 the Ecuadorian Government organized its first National Scien- tific Commission. The members of this commission, which included this au- thor, and helped by the President in charge at that time. General Albert Enqi- quez, took a trip to Galapagos in a warship of the National Navy, Cotopaxi Cannoneer. On its return, a complete report was prepared which pointed to the necessity of preserving all nature in the Archipelago and especially its rich fisheries resources. But practically no real protection resulted. Then this au- thor suggested the need for establishing a biological station so that scientif- ic investigations on the Galapagos biota could be carried out and at the same time some protection afforded the native biota. Only after 23 years, because of interest demonstrated by the Government of Ecuador, UNESCO, and the In- ternational Union for the Conservation of Nature, has the establishment of a biological station on Galapagos become a reality. This research center has been named in honor of Charles Darwin, the great British naturalist, and au- thor of the "Origin of Species." * Presented at the TENTH PACIFIC SCIENCE CONGRESS of the Pacific Science Association, held at the University of Hawaii, Honolulu, Hawaii, U.S.A., 21 August to 6 September 1961, and sponsored by the NATIONAL ACADEMY OF SCIENCES, BERNICE Pauahi Bishop Museum, and the University of Hawaii. -141- 142 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Galapagos is one of the most interesting scientific reserves in Ecua- dor, and, indeed, in the world. To preserve Galapagos as a National reserve means the protection and conservation of its resources. Its protection is indispensable not only for my country, but also for the world. For many years scientists and scientific in- stitutions have been unsuccessful in coordinating their plans, but since 1957, as a result of the studies and reports of two biologists, Dr. Robert I. Bowman of the United States of America, and Dr. I. Eibl-Eibesfeldt of Austria, plans moved quickly ahead for the establishment of a biological station at Academy Bay on the south side of Indefatigable (Santa Cruz) Island. Sponsors of the 1957 reconnaissance were the International Union for the Conservation of Na- ture, UNESCO, the International Committee for Bird Protection, the New York Zoological Society, the Conservation Foundation, and Life Magazine. In 1959, and coincident with the "Darwin Centennial," the Charles Darwin Foundation for the Galapagos Isles was founded by Professor Victor VanStraelen, in Bel- gium. The principal purpose of this international scientific organization is to conserve the indigenous biota of Galapagos and to promote scientific research of a fundamental nature in the Galapagos region. Now we must ask, does the Galapagos Archipelago need only protection or does it also need better methods of conservation? Distinction Between "Protection" and "Conservation" We need to distinguish between "protection" and "conservation," since these terms are often confused. "Protection" is a word generally used by naturalists who are not prim- arily concerned with economic benefit or any other utilitarian exploitation. Their point of view is mainly scientific and esthetic. They are concerned with spiritual enjoyment for themselves and future generations. "Conservation" is a practical and positive word with economic connotations, that is used by those who look upon nature as a resource for exploitation, but who use mod- ern principles of management to preserve the resources of today so that some will remain for the future. Thus the purposes and objectives of the protec- tionists and conservationists are rather different, technically speaking. It is not possible to separate their activities because the protectionists also tend to preserve the fauna, flora, and minerals; the conservationists, in turn, sup- port the protection of certain species of plants and animals and kinds of min- erals. The conservationists tend to preserve definite species of animals, for example, because they constitute an economic resource, that is to say, a na- tural resources which must not be exhausted and thereby harm those people who make their living by its exploitation. These people are not interested in a species as such, but in its quantity. On the other hand, the protectionists are preoccupied, fundamentally, with species which are disappearing or which are No. 44) ACOSTA-SOLIS: GALAPAGOS SYMPOSIUM 143 close to extinction. The protectionist is not interested in the species for its economic utility or for personal profit, but for science. He looks at species as objects for scientific study. The main difference between protectionists and conservationists is in the application of their own concepts to the native biota and to the introduced or exotic one. The protectionists are almost entirely preoccupied with the protection of the native biota that forms the natural life of a specific environ- ment, because invasions by exotic species always produce a disturbance of the biological equilibrium of the environment where the invasion has taken place. But the latter does not disturb the conservationists, especially if they are hunters or fishermen. The protectionist, using scientific means, tries to prevent the invasion of species which will disturb the natural biological en- vironment. He does not oppose the introduction and acclimatization of exotic species, provided they are retained in some special gardens. The introduction of exotic species in large areas or national parks can form the basis of a new economic resource deserving of the conservationist's attention. For the protectionist, any species deserves protection when it is threat- ened with extinction. The conservationist is interested only in the species of economic value as exploitable resources of nature. Thus insofar as Galapagos is concerned, it must be borne in mind that both concepts, protection and conservation, are applicable to the biology and pedology of the different islands of the archipelago. Some Protectionist Suggestions The protectionist suggestions given by this author since 1937 and sub- sequently encouraged by UNESCO and lUCN reports are, in short, as follows: 1. To determine those islands or areas to be set aside as reserves for Galapagos wildlife where effective protection can be enforced. In the areas declared as "reserves," all hunting, agriculture, and human settlement will be prohibited by law. Visits by tourists to these reserves will have to be strictly controlled; only the scientific excursions under competent leadership will be permitted; scientific collections should always be allowed but only under strict regulation in reserve areas. 2. The areas or islands declared as "reserves" will be those which have rare species, or species in danger of extinction, or those that have scen- eries which are worthy of protection as a living museum of nature. Reserves must be without settlements. In the areas where there are villages, the "re- serves" should not be established and that is why we cannot speak about "reserves" in Chatham (San Cristobal) and Charles (Floreana) islands; also in the southern part of Albemarle (Isabela) Island. On the contrary, on Inde- fatigable Island (Isla Santa Cruz) there is a special situation. The western side still possesses a relatively undisturbed habitat with good-sized colonies 144 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers of giant tortoises, in need of protection. And in the eastern end of the same island there are agricultural lands which have not yet been farmed and where human population is scarce. In this case the best thing to do is to delimit the areas: those for protection, and those for settling and agriculture. In this way there would be less interference between the two functions, the natural biol- ogy of the giant tortoises, and the use of the natural edaphic resources, all managed under conservationist principles. 3. Besides the legal and technical provisions suggested since 1937 in behalf of protection of the Galapagos biota, it will be necessary to instigate a campaign of protectionist publicity all over Ecuador, in the grade schools, high schools, etc., and enlist their cooperation in the governmental project, so that they will visit the Archipelago and be spokesmen for its protection, and by their actions set a good example. On the other hand, the campaign of conservationist education will have to be taught to the settlers who live in Galapagos, through lectures, illustrated pamphlets, and motion pictures. In this way, the settlers will learn how to be responsible to surrounding nature and to recognize the value of the native biota. Some Conservationist Suggestions The conservationist must bear in mind the following points when deal- ing with the economic resources of Galapagos. 1. Recognition of the good agricultural lands in Galapagos. According to our experience, the islands that have agricultural lands with areas or high strips at 500 meters above sea-level are the large ones, that is to say, where humidity and rains have permitted the formation of a thicker soil stratum, where there is a lush herbaceous, shrubby, and arboreal vegetation. We have some examples of these on Chatham (San Cristobal) and Charles (Floreana) islands, on the eastern side of Indefatigable (Santa Cruz) Island, and on the south of Albemarle (Isabela) Island. 2. After the pedological and agricultural studies, we must classify the lands according to agricultural capacities, that is to say, to delimit them on a special map. It is true that this work would take a long time, but on the other hand this would show us the best use of the lands for grazing, planta- tions, forests, horticulture, etc. Good land use is essential for maintenance of continuous productivity of agricultural crops as well as for wildlife. 3. Careful planning of colonization and agriculture is absolutely neces- sary for wise land use on Galapagos. The productivity of Galapagos is strictly limited and therefore Galapagos colonization must be meticulously controlled. Galapagos Fauna that Must be Protected Immediately According to our personal observations in 1937, supplemented by the findings of later investigators, the following groups of animals must be pro- No. 44) ACOSTA-SOLIS: GALAPAGOS SYMPOSIUM 145 tected by legal decrees and also by actual enforcement. 1. The gigantic tortoises on Hood (Espanola), Duncan (Pinzon), and Abingdon (Pinta) islands, and, if not too late, the ones in the interior of Cha- tham (San Cristobal) Island. There are numerous torotises on Indefatigable (Santa Cruz) and Albemarle (Isabela) islands, which are also in great danger of extinction because of the relentless and uncontrolled hunting. 2. The land iguana needs protection, especially the small colony that exists on Indefatigable Island. If the western portion of the island is set aside as a biological reservation, and if we enforce the protective laws, then the little colony can be saved and even increased in its own habitat. 3. The fur seals of Galapagos must be protected from hunters. 4. Birds of all species deserve protection and care from the commercial hunters. Penguins and flightless cormorants are not abundant and need full protection. The flamingos are not very common according to several reports, and every effort should be made to fully protect this beautiful species in Ga- lapagos. Special Areas In Galapagos we can distinguish several islands or special habitats for certain groups and species of fauna and flora. In a protectionist sense the following islands or areas may be mentioned as being of great biological im- portance: A. Special areas for bird protection: 1. Hood (Espanola) Island is the nesting site of the Galapagos al- batross and where blue-footed boobies, frigate-birds, red-billed tropic-birds, and Ion-billed mockingbirds are concentrated. The gigantic tortoise is very rare or possibly extinct on this island. One of the longest coral-sand beaches in all of Galapagos is situated on the north shore of the island, adjacent to Gardner- near-Hood Island. 2. Tower (Espaiiola) Island has a large number of nesting red-footed and masked boobies, frigate-birds, and one of the largest examples of Darwin's finches, Geospiza magnirostris. Besides, in this island there is much beauti- ful scenery, such as the central crater lake with its mangrove swamps and the picturesque Darwin Bay. 3. Narbo rough (Fernandina) Island constitutes the largest home of the Galapagos penguin and flightless cormorant. Here there are large colonies of land and sea iguanas, and the native rat (Nesoryzomys narboroughi). The scenery of the central crater of this island is magnificent. There are hot sul- phur springs and very fresh lava flows, all of which are of much interest to geologists and volcanologists. In the fresh water that formerly occurred in the central crater, a fish has been collected, hitherto unknown for the Galapa- gos region. The gigantic tortoises still occur on the south side of the island. 146 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Narborough Island is very important to scientists because of its primitive na- ture and the fact that the biota has not been spoiled with the exotic plants and animals that have found their way to all the islands with permanent human hab- itations such as Chatham (San Cristobal), Charles (Floreana), south Albemarle (Isabela), and Indefatigable (Santa Cruz) islands. 4. Culpepper rocks, forming the most northerly point in the Gala- pagos Archipelago, is the appropriate site for the large concentration of sea- birds, including boobies, frigate-birds, terns, and petrels. B. Special areas for the protection of the gigantic tortoises: 1. Indefatigable (Santa Cruz) Island is by far the most interesting of the reserves for the gigantic tortoises. These large reptiles have been dis- appearing because of abuse from colonists, hunters, and collectors. The west- ern half of Indefatigable Island must be set aside as a reserve in which no agriculture will be permitted. 2. Duncan (Pinzon) Island has some giant tortoises still remain- ing on its south side. If we keep out the hunters and collectors we should be able to restore this small colony to its former level of abundance. 3. Abingdon (Pinta) Island still likely has a few tortoises. Full protection should be afforded those few remaining animals. 4. Albemarle (Isabela) Island has many tortoises and is second only to Indefatigable Island as an important reservoir of these animals. From Perry Isthmus to the north end of the island should be kept as a reserve. Much could be done to teach the colonists on the southern part of Isabela Island about the necessity of protecting these chelonians. C. Special areas for the protection of land iguanas: 1. Harrington (Santa Fe) Island is the home of thousands of land iguanas and feral goats. Land iguanas have survived the inroads on their num- bers made by commercial hunters. From now on hunting must be curtailed and the civil authorities must assist in the patrol work. 2. Plaza Island, located at the east end of Indefatigable (Santa Cruz) Island, constitutes the typical habitat of a pink-colored iguana. Be- cause of their small numbers there is great danger of extermination by hunt- ers. This island also harbors sea lions, petrels, tropic-birds, and because of certain unusual geological aspects, it constitutes a very important area in need of reserve status. 3. Narborough (Fernandina) Island, in addition to being a reserve for birds, must be considered as a reserve for the marine and land iguanas; at present, the colony of land iguanas is very large and great herds of marine iguanas may be found along the outer shores. But I repeat, they need official and scientific control. FUTURE SCIENTIFIC STUDIES IN THE GALAPAGOS ISLANDS* Jean Dorst Museum National D'Histoire Naturelle, Paris, France The Galapagos Islands rank among the most celebrated places in the world, since they provide one of the most clearcut natural experiments in or- ganic evolution. Every biologist, once in his life, should go on a pilgrimage to the spot where one of the greatest achievements of science was born. It was in the Galapagos that young Charles Darwin found his inspira- tion. In July, 1837, not later than two years after his visit to Galapagos, Dar- win wrote: "I opened my first notebook on Transmutation of Species. Had been greatly struck from about month of previous March on character of South American fossils and species on Galapagos archipelago. These facts origin (especially latter) of all my views." These prospects are not only of historical and retrospective value. The islands are still today "evolution's workshop and showcase," and they have maintained the same importance up to the present time, where a great deal of research remains to be done in various fields. Of course, evolution is observable anywhere in the world, but the trends and the laws which govern the phenomenon often are hidden by complexity and by the innumerable factors involved. On the contrary the simplification of the Galapagos ecosystem makes it much more apparent in these islands. Numerous studies in systematics have already been done on the Gala- pagos. Vertebrates have been the most thoroughly studied, except the smaller reptiles and the endemic rats. Nevertheless most of the investigations on in- vertebrates are outdated or even completely missing. The inventory of all liv- ing creatures, from the smallest to the giant tortoises, must be undertaken by priority with all the modern concepts in mind. This will form a basis for fu- ture studies on the Galapagos fauna and especially those dealing with the biological aspects. Studies in systematics will also shed new light on the problems of speciation and on evolution in progress. Owing to the laws which govern evolution among invertebrates, we may assume that a better knowledge * Presented at the TENTH PACIFIC SCIENCE CONGRESS of the Pacific Science Association, held at the University of Hawaii, Honolulu, Hawaii, U.S.A., 21 August to 6 September 1961, and sponsored by the NATIONAL ACADEMY OF SCIENCES, BERNICE Pauahi Bishop Museum, and the University of Hawaii. - 147 148 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers of the microfauna could lead to a better understanding of many biological prob- lems. In recent years, scientists have brought new techniques to a high degree of perfection, especially in the field of genetics, chromosome analysis, bio- chemical tests like chromatography, and even micromorphology. Many Galapa- gos animals could be tested by such techniques and the results may give clues to their interrelationships and on the modes of evolution in each group. This applies especially to genetical investigations. The geneticists have now achieved a great deal of information on the mechanism of heredity. New methods and large scale experiments have been undertaken. But we may wonder why no one has tried to apply these techniques to the Galapagos ani- mals or plants — a wonderful field for experimentation owing to the simplifica- tion of the problems involved as mentioned above. However, evolutionary studies are only one aspect of the investigations to be undertaken in the Galapagos, where biological researches of all types might be successful. One of the many fields of investigation could deal with behavior, some very peculiar trends being observable on these islands. The earlier natural- ists reported on strange behavior patterns to be observed among Galapagos birds and reptiles. Development of these patterns are, in large part, under the influence of the very particular conditions prevailing in the islands, where the number of species is smaller than anywhere else, and where few enemies are present, the mammals being almost completely absent. The splitting of spe- cies into local populations, owing to isolation on islands with no communica- tion, has a role which can not be minimized. A great deal of observation con- firms that the trends apparent in behavior have paralleled those observed in morphological diversification. One of the best examples is to be found among the well known Darwin's finches or Geospizinae, studied from the morphological point of view by David Lack (Darwin's Finches, London, 1947) and recently by Robert Bowman (Univ. Calif. Publ. Zool. 58, 1961). The evolution of these fringillids shows a per- fect adaptive radiation in morphology and anatomy in relation to diversity of habitats and ecological niches. They also show a very elaborate evolution in behavior patterns, and especially in vocalizations. In the last few months Dr. Bowman has undertaken a thorough study of these aspects with modern tech- niques. But some other aspects might be investigated with success and such researches have to be extended to other birds like the Galapagos mockingbirds (Nesomimus) differentiated into several geographical races. Such studies, which in recent times brought so many interesting facts to our knowledge, must be extended to all Galapagos birds and also to the reptiles, especially the marine iguana ( Amblyrhynchus cristatus), one of the most interesting rep- tiles in the world. This lizard, with numerous adaptations to conditions in the sea, from which it never straggles far, shows very strong tendencies to gre- gariousness, especially along the shores of Narborough Island, where colonies No. 44) DORST: GALAPAGOS SYMPOSIUM 149 of many hundreds may be observed; but anywhere else this iguana is more or less gregarious and single individuals are met very seldom. This social be- havior may be very interesting to investigate carefully in relation to ecologi- cal factors. Of course, these chosen examples are only a few among the many which could be studied in the Galapagos. Furthermore one of the main tasks for scientists in these islands is to study the Galapagos ecosystem. This aspect of investigation has a particular importance in the light of conservation of nature and wildlife. Nature is badly threatened in the Galapagos as a consequence of human impact. When they were discovered in 1535 by the Bishop of Panama, Tomas de Berlanga, they were in a virgin condition. Since this not so remote time, various people set- tled, destroyed the habitats, and killed the animals, sometimes for food, some- times for "fun." Moreover, domestic animals, and especially goats and pigs, have been introduced and soon became feral. They fundamentally modified the balance of nature and contributed in a very large measure to depletion of stocks of endemic species and disruption of natural habitats. Therefore many species are on the verge of extinction and some must already be considered to have vanished completely. To preserve what is left we must undertake a thorough ecological study of the environment in the Galapagos. The pedological analysis is one of the first tasks as it will constitute the basis for all further work. It will also be a wonderful field of investigation in itself, since the Galapagos, purely volcanic in origin, offer various stages of transformation and colonization of lava flows by microorganisms and plants. All stages from pure mineral soil to top-soil, where cultivation is possible (in some very limited areas), are present in these islands. A detailed study of the soils, in relation to the geologic history of the islands and chronology of the various lava flows and volcanic eruptions, might lead to discovery of some very important facts and to a better understanding of the evolution of land of volcanic origin. By the way, it could also give a basis for better land use, avoiding utilization for agricultural purposes of land that is better kept in its natural state. The second stage of such a broad ecological survey is a thorough study of structure of vegetation. Provided a complete systematic list of all plant species living in the Galapagos is available, it would be a fascinating work to describe the various plant communities and to follow their evolution. These communities range in type from real desert biomes with a highly characteristic spaced distribution of cacti and spiny bushes, with large barren areas in be- tween, to grasslands, and dense moist forests, among which the lofty Scale- sia trees, an endemic genus of Compositae, are the most conspicuous plants. Many forms of plants could be studied in relation to the different life zones. Besides knowledge of their relationship — especially in Scalesia, wild toma- toes and cacti — could lead to very interesting conclusions on plant evolution. These studies, in relation to climate and soil distribution, must be synthesized 150 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers and then summarized in a detailed vegetation map covering the whole Archi- pelago, a great need for all scientists working in this part of the world. Like animals, plants have been modified to a large extent by the impact of species introduced by man, voluntarily or not, and which compete with en- demics and often compel them to modify their response. We have at our dis- posal numerous data collected by earlier botanists. Comparison with the con- temporary situation could be of the greatest interest for a historical approach to an understandingof any fluctuations in the make-up of the flora, and of course to preserve some of the most typical, and still untouched, habitats. With this pedological and botanical background zoologists could then undertake their various investigations on terrestrial communities. Soil com- munities are of primary interest and, as their elements probably show marked differences from what is known in other places, the complete study of the sys- tem could be of the greatest interest. Soil ecology yet remains so little under- stood over the world that this particular aspect of investigations in the Gala- pagos Islands could give some clues to several problems as yet unsolved. If we take into consideration the upper levels of the ecosystem, we may imagine that here also we could bring new facts to light. There are few mammals and most of the terrestrial vertebrates are birds or reptiles. Leaving aside the birds, the place of tortoises, land iguanas, and smaller lizards in the ecosys- ten is to be investigated thoroughly, and, besides being of interest in itself, this may give a good idea of what life was like in former times. In the Gala- pagos we are faced with a "fossil biocenosis." Many adaptations, some very strict, may be found between reptiles and plants. Recently Rick and Bowman (Evolution, 15, pp. 407-417, 1961) emphasized the fact that the seeds of na- tive tomatoes, subject to a dormancy of indefinite duration, may be activated in the digestive tract of the giant tortoises. The passage through the gut, re- quiring from one to three weeks and even longer, produces a marked improve- ment in speed and percentage of germination. Thus tortoises may be the main agents for breaking the dormancy and also for actually spreading the seeds through their droppings — a very important adaptive device. Similar examples of symbiosis might be found in many other aspects of the Galapagos biota, which, from the biological point of view, constitutes a distinct world in itself, evolved in isolation from the rest of the world. To get information on the evolution of the balance of nature in the Ga- lapagos, much routine work is to be undertaken immediately under the direct guidance of the Darwin Research Station, beginning with censuses of all threat- ened species in well defined areas or along carefully chosen linear transects. Periodically, i.e., several times a year, reptiles and birds must be numbered. These figures will provide data on their seasonal fluctuations, the dynamics of their populations, very important basic facts that must be known if we want to succeed in protecting them efficiently. No. 44) DORST: GALAPAGOS SYMPOSIUM 151 Evolution of plant communities and of habitats will be watched within quadrates established in different areas, surrounded by fences and even rat- proof walls; periodical surveys will be undertaken to see how the vegetation cover is modified under natural conditions. Comparison with unprotected areas will measure the importance of the impact of feral animals on the plants. This work is of primary importance as a background for all types of researches in the Galapagos, and also as a basis for the recommending of practical mea- sures to safeguard the wildlife. If research on terrestrial biomes is of such great importance for the fu- ture of this invaluable natural heritage, we must not however forget that the seas surrounding the archipelago also constitute a unique field for investiga- tions by oceanographers. This part of the Pacific is probably one of the most complex of all marine sectors. Several currents of various origins meet here, carrying waters of very different characteristics. The Humboldt Current, flow- ing from the East, brings cool Antarctic waters, producing oceanographic con- ditions seemingly paradoxical for islands situated on the equator; it mingles with warm eastward flowing waters brought by the Equatorial counter-current and subsurface Cromwell Current, and by a current coming from the Gulf of Pan- ama. From this odd situation results a juxtaposition of warm and cold areas, often very distinct and visible like a mosaic of different colors, and an inter- mingling of marine faunas of various origins. Fur seals and penguins are liv- ing side by side with flying fishes and tropic-birds; the same may be observed among marine invertebrates. Distribution of marine plants and animals within the whole archipelago must be carefully investigated in relation to oceanographic conditions. This aspect of the researches will probably bring to light some very interesting facts such as a narrow range of adaptations of various organisms to certain physical factors of sea waters. These investigations could also be of the greatest significance to the economic development of Galapagos. Seas surrounding the archipelago are nu- tritionally very rich, and many fishes, the size of which ranges from the small herring to the big albacore, swarm in these waters. A local fishing industry has already been established on a very small scale, but with encouraging re- sults. We must bear in mind that cropping of marine products is the only na- tural resource of direct economic importance in the Galapagos where agricul- ture will never be successful, except within very limited areas, owing to water shortages and untillable land. Efforts to promote agriculture as a resource to attract settlers or to encourage a major expansion of cultivation, would mean poor land use with concomitant destruction of habitats suitable to wildlife, that could never be restored. But if we want to save these islands, so famous to scientists, we must give to settlers a higher standard of living — the best way to discourage cropping of tortoises and destruction of habitats, as is true everywhere in the world. We are convinced that economic development of the 152 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Galapagos Islands must consist of a rational exploitation of the sea, a contri- bution of great value to the many Ecuadorians who suffer from an acute pro- tein shortage. Such exploitation must be based on scientific facts if we are to manage properly the fishery so as to avoid over utilization and consequent depletion of the resource. Our research station may assume this tesponsibility in cooperation with several other organizations interested in oceanographic researches in this part of the Pacific Ocean, such as the Inter-American Tro- pical Tuna Commission and FAO Laboratories. Thus we should integrate into the plan of economic development of Ecuador, which is a point of very great importance both on technical and psychological grounds. The program of investigation is therefore unlimited. For all these rea- sons it was urgent to establish a research station in these islands. This has now been done thanks to the action of the Charles Darwin Foundation for the Galapagos Isles, which organization was founded in 1959 and is governed by an Executive Council including representatives from all nations interested in these researches. With funds raised in several countries, the construction of a field station has been achieved at Academy Bay on Indefatigable Island (figs. 1 and 2). The station is already in operation and material improvements continue to be made. The main purpose of the Darwin Foundation is to offer scientists of all disciplines and nationalities living accomodations and work- Figure 1. General maintenance building of the Charles Darwin Research Station, Academy Bay, Indefatigable Island, Galapagos. Photo, courtesy R. Leveque, December, 1961. No. 44) DORST: GALAPAGOS SYMPOSIUM 153 «^ t9>fM^^' Figure 2. Main laboratory building of the Charles Darwin Research Station, Academy Bay, Indefatigable Island, Galapagos. Photo, courtesy R. Leveque, December, 1961. ing facilities in a well equipped field laboratory. Basic equipment for scien- tific research, especially in the field of biology, will be available to all visit- ing scientists approved by the Foundation. A small research vessel for trans- portation and oceanographic studies will soon be available for use within the limits of the archipelago. The Darwin Research Station is run by its own scientific and technical staff which cooperates fully with investigators working under Foundation au- spices. Routine scientific work is done by the staff, including meteorological and oceanographic measurements, censuses of animals, studies on changes in habitats, etc. Of course, particular scientific investigations will be undertaken by visiting specialists. But our main objective is to make the station open to visiting scientists working on special projects from many branches of science, from geology to oceanography to terrestrial ecology. It is the intention of the Foundation to co- operate with all scientific organizations throughout the world that are inter- ested in Galapagos research. It must be remembered, of course, that we are the guests of the people of Ecuador, its government and institutions, which have given enthusiastic support to our efforts, and with which we are cooperating in all our activities. 154 CALIFORNIA ACADEMY OF SCIENCES (Occ. Papers Collaboration with the scientific authorities is especially important and it will be a great accomplishment indeed when young Ecuadorian scientists study together with foreign specialists in our Station, a real international crossroad. Moreover, we hope that the Ecuadorian government will consult with the Foundation in matters concerning the conservation of nature in Galapagos. All legal measures must be taken by Ecuador, for it is not our intention to in- terfere in any manner whatsoever with its sovereignty. We are willing, however, to answer official requests for advice, on the basis of scientific investiga- tions made under Foundation auspices. Our Foundation is probably unique in its class. It is an international institution whose aim is to advance our scientific knowledge specifically in the Galapagos region and to conserve the remarkable biota inhabiting it. The evolutionary theories of Charles Darwin, which belong to mankind as a whole, completely justify such an organization. On the other hand, the invaluable natural inheritance of Galapagos wildlife was threatened by men of all nation- alities; it is logical, therefore, that scientists and conservationists from all nations collaborate with the Ecuadorians to study and save what is left of these islands for the benefit of generations to come. MBl. WHOI LIBRARY UH ITFV J L